NOT TO BE CITED WITHOUT PRIOR REFERENCE TO THE AUTHORS

ICES Annual Science Conference 2011 ICES CM 2011/S:09 Theme Session S: Extracting energy from waves and tides – what are the consequences for ecosystems, physical processes and other sea users?

Sources of underwater noise and disturbance arising from marine renewable developments

Jonathan Side and Robert Beharie

International Centre for Island Technology/Institute of Petroleum Engineering, Heriot‐Watt University, Orkney Campus, The Old Academy, Stromness, Orkney. KW16 3AW Tel: +44 (0)1856 850605; Fax +44 (0)1856 851349; email: [email protected]

ABSTRACT

Anthropogenic underwater noise has received increasing attention in recent years and is one of the Marine Strategy Framework Directive’s descriptors of good environmental status. Wave and tidal energy projects pose new challenges and have required a variety of installation methods where the disturbance arising from sources of underwater noise are only now being considered. The paper reviews the sources of underwater noise associated with the development of this new sector and the methods and studies conducted to date to determine noise levels. Sources of underwater noise arising from the deployment of marine renewables devices are compared with background and other sources of anthropogenic noise in the marine environment. We examine the models used for transmission loss and the present thresholds that are applied to the disturbance of marine mammals in the vicinity, highlighting those sources of most concern and suggesting methods of mitigation.

KEYWORDS: underwater noise, wave energy, tidal energy, marine mammals, environmental disturbance, Sonobuoy

INTRODUCTION

The European Marine Strategy Framework Directive (MSFD) seeks the maintenance of good environmental status within the marine environment of the European regional seas. It aims to provide a coherent strategy for achieving this by 2020, which involves assessment, the establishment of targets for a range of qualitative descriptors which are designed to characterise a desired condition of good environmental status, and monitoring programmes to provide ongoing assessments of environmental status against targets. There are many other elements to the strategy including measures enabling the spatial management of all marine activities, and thus the possibility of creating much more extensive

Page 1 marine protected areas (MPAs) than those presently designated in coastal waters under the earlier Birds (Special Protection Areas) and Habitats Directives (Special Areas of Conservation). Table 1 provides a summary of the qualitative descriptors being used, many of which could have a bearing on the development of offshore wind and wave and tidal energy projects. It is interesting to note that under earlier legal frameworks concerned with reducing and eliminating pollution, the definition of pollution would not include the removal of energy, marine pollution being defined as:

“The introduction by man, directly or indirectly, of substances or energy into the marine environment (including estuaries) resulting in such deleterious effects as hazards to human health, harm to living resources and to marine ecosystems, damage to amenities or interference with other legitimate uses of the sea.”

The MSFD provides a more extensive definition of pollution, specifically referencing for example underwater noise, but nonetheless the removal of energy is excluded from its terms:

‘Pollution’ means the direct or indirect introduction into the marine environment, as a result of human activity, of substances or energy, including human‐induced marine underwater noise, which results or is likely to result in deleterious effects such as harm to living resources and marine ecosystems, including loss of biodiversity, hazards to human health, the hindering of marine activities, including fishing, tourism and recreation and other legitimate uses of the sea, impairment of the quality for use of sea water and reduction of amenities or, in general, impairment of the sustainable use of marine goods and services;

Table 1: Annex I Qualitative descriptors for determining good environmental status

1. Biological diversity is maintained. The quality and occurrence of habitats and the distribution and abundance of species are in line with prevailing physiographic, geographic and climatic conditions. 2. Non-indigenous species introduced by human activities are at levels that do not adversely alter the ecosystems. 3. Populations of all commercially exploited fish and shellfish are within safe biological limits, exhibiting a population age and size distribution that is indicative of a healthy stock. 4. All elements of the marine food webs, to the extent that they are known, occur at normal abundance and diversity and levels capable of ensuring the long-term abundance of the species and the retention of their full reproductive capacity. 5. Human-induced eutrophication is minimised, especially adverse effects thereof, such as losses in biodiversity, ecosystem degradation, harmful algae blooms and oxygen deficiency in bottom waters. 6. Sea-floor integrity is at a level that ensures that the structure and functions of the ecosystems are safeguarded and benthic ecosystems, in particular, are not adversely affected. 7. Permanent alteration of hydrographical conditions does not adversely affect marine ecosystems. 8. Concentrations of contaminants are at levels not giving rise to pollution effects. 9. Contaminants in fish and other seafood for human consumption do not exceed levels established by Community legislation or other relevant standards. 10. Properties and quantities of marine litter do not cause harm to the coastal and marine environment. 11. Introduction of energy, including underwater noise, is at levels that do not adversely affect the marine environment.

Member States shall consider each of the qualitative descriptors listed in this Annex in order to identify those descriptors which are to be used to determine good environmental status for that marine region or sub-region.

Source Annex 1, European Marine Strategy Framework Directive 2008/56/EC

Page 2 As by definition the Precautionary Principle only applies to pollution, the large scale extraction of wave and tidal energy cannot be considered under this or anti‐pollution measures. Instead, however, the over‐arching objective of achieving targets of good environmental status in the MSFD means that, in this context, potential impacts from marine renewables developments on many of the eleven Qualitative Descriptors (Table 1) should be considered, and obviously Descriptors 6 and 7. Last year the Council Decision on criteria and methodological standards for determining good environmental status for each of the descriptors was published (2010/477/EU). For the purposes of this paper only the criteria and methods for underwater noise are discussed further, and Table 2 sets out the requirements for measurement of underwater noise contained in the Council Decision.

Table 2: Requirements for Descriptor 11: Introduction of energy, including underwater noise, is at levels that do not adversely affect the marine environment.

Together with underwater noise, which is highlighted throughout Directive 2008/56/EC, other forms of energy input have the potential to impact on components of marine ecosystems, such as thermal energy, electromagnetic fields and light. Additional scientific and technical progress is still required to support the further development of criteria related to this descriptor, including in relation to impacts of introduction of energy on marine life, relevant noise and frequency levels (which may need to be adapted, where appropriate, subject to the requirement of regional cooperation). At the current stage, the main orientations for the measurement of underwater noise have been identified as a first priority in relation to assessment and monitoring, subject to further development, including in relation to mapping. Anthropogenic sounds may be of short duration (e.g. impulsive such as from seismic surveys and piling for wind farms and platforms, as well as explosions) or be long lasting (e.g. continuous such as dredging, shipping and energy installations) affecting organisms in different ways. Most commercial activities entailing high level noise levels affecting relatively broad areas are executed under regulated conditions subject to a license. This creates the opportunity for coordinating coherent requirements for measuring such loud impulsive sounds.

11.1. Distribution in time and place of loud, low and mid frequency impulsive sounds — Proportion of days and their distribution within a calendar year over areas of a determined surface, as well as their spatial distribution, in which anthropogenic sound sources exceed levels that are likely to entail significant impact on marine animals measured as Sound Exposure Level (in dB re 1μPa 2 .s) or as peak sound pressure level (in dB re 1μPa peak) at one metre, measured over the frequency band 10 Hz to 10 kHz.

11.2. Continuous low frequency sound — Trends in the ambient noise level within the 1/3 octave bands 63 and 125 Hz (centre frequency) (re 1μΡa RMS; average noise level in these octave bands over a year) measured by observation stations and/or with the use of models if appropriate.

Source: EN L 232/24 Official Journal of the European Union 2.9.2010

It can be readily seen that there is a distinction maintained between continuous underwater low frequency noise such as that arising from shipping and other relatively constant sources, and impulsive sounds such as those arising from pile driving in the offshore industry. Of course some sources such as chain noise from heavy chains used in mooring systems laid on bedrock may exhibit impulsive noise as well as continuous jangling.

SOURCES OF UNDERWATER NOISE IN THE MARINE RENEWABLES SECTOR

Scotland has seen major initiatives for the development of wave and tidal energy capture, including in Orkney the establishment of European Marine Energy Centre (EMEC) with both wave and tidal test sites that can be used by developers; as well as now the issuing of lease rights to the seabed for major multi‐

Page 3 device wave and tidal farms (see Figure 1). Many of these initiatives encouraging the development of wave and tidal technologies are discussed elsewhere in this Conference (see for example Davies et al., 2011).

Figure 1 Showing the arrangements for lease rights put in place by the UK Crown Estate for wave (green) and tidal (orange) field developments. All developments will be subject to license approval by Marine Scotland.

There are a number of sources of noise from the installation, maintenance and operation of wave and tidal devices. No measurements, to our knowledge, have yet been published of the continuous operation of energy convertors, but we surmise that although this represents a continuous input of underwater sound, in general we would anticipate that for reasons of efficiency the noise levels associated with power generation would be low relative to other sources of noise continuously pervading the marine environment. Put simply the greater the noise during operations, the greater the opportunity for designing greater efficiency into the design of a marine energy convertor. A range of the presently deployed wave and tidal devices either on test or close to full‐scale testing in Orkney and elsewhere are shown in Figure 2 (tidal devices) and Figure 3 (wave devices). It can be seen that wave devices tend to be free floating with a mooring, and tidal devices are mostly fixed to the seabed. The exceptions of course, in these figures, are the Scotrenewables floating tidal turbine which is moored, and Aquamarine’s wave device, Oyster, which is fixed to the seabed.

But while noise levels from operational devices have still to be studied, many other sources of noise that are likely to arise as a result of installation and maintenance have been studied for similar

Page 4 applications (Nedwell et al., 2003) or now for marine renewable energy device installation (Nedwell and Brooker, 2008; Beharie and Side 2011a, and 2011b). For tidal devices it is the operations associated with the installation of the devices, and possibly subsequent maintenance activities, which are most likely to generate noise levels of concern to the regulator.

Figure 2 A selection of tidal devices:top panel (left to right) MCT Seagen in Strangford Lough, Open Hydro at the Fall of Warness, Scotrenewables SRT250 leaving H&W Belfast for Fall of Warness,Orkney;bottom Panel (left to right) SMD Hydrovision TidEl, Lunar Energy, Hammerfest Strøm.

Figure 3 A selection of wave devices: top left Pelamis P2 off Orkney, Bottom left Wavegen’s Limpet on Islay, Aquamarine’s Oyster off Orkney, and to the right the OPT Power Buoy.

Page 5

Figure 4 shows some typical installations to date, of the Marine Current Turbines Seagen device in Strangford Lough, of the drilling for installation of Aquamarine’s Oyster, of the drilling for the Voith tidal turbine, and for the installation of anchor blocks and chain for the Scotrenewables floating tidal turbine.

Figure 4 A selection of installation methods: top left Seacore jack‐up barge during the installation of Aquamarine Oyster 2 off Orkney, bottom left Rambiz installation of MCT Seagen in Strangford Lough, top right Voe Viking workboat, used for the installation of moorings for the Scotrenewables SRT250 in the Fall of Warness, Orkney. Bottom right the North Sea Giant commencing drilling for the monopile foundation for the Voith berth at the Fall of Warness.

As maintenance will be required we can also postulate that once operational large wave and tidal farms will be regularly visited by vessels for this purpose. For some technologies, like the Pelamis and Scotrenewables turbine this will simply be a matter of unhooking one device and replacing it by another, for others, particularly seabed mounted technologies with no surface piercing column to provide access to the turbine, rather more elaborate vessels may be required for intervention.

Page 6 Table 3 Source levels from anthropogenic underwater noise for various activities that may be involved in marine renewables developments.

Activity/Source Reported levels / Estimate Reference Pile driving (4.0-4.7m 243-257 dB re 1μPa at 1m (peak to Nedwell et al. (2007) diameter piles) peak)

Pile driving (1.8m 226 - 250 dB re 1μPa at 1m (peak to Bailey et al. (2010) diameter piles) peak)

185-196 dB re 1μPa at 100m (rms) Pile driving (2.4m Caltrans (2001) diameter piles) 197-207 dB re1μPa at 100m (peak to peak)

Concrete block and 167-173 dB re 1μPa at 1m (peak to ICIT measurement in the Fall of anchor chain placement peak) Warness (Beharie and Side, 2011a) on bedrock

Lower range from measured data in the Fall of Warness (Beharie and Dynamic Positioning 167-190 dB re 1μPa at 1m (rms) Side, 2011b). Higher value Drillships presumably for much larger vessels from NRC (2003).

Larger vessels 180-190 dB re 1μPa at 1m (rms) OSPAR Commission (2009)

Pile Drilling 160-180 dB re 1μPa at 1m (rms) ICIT, Nedwell & Brooker (2008)

Small work-boats (with OSPAR Commission (2009) and 160-180 dB re 1μPa at 1m (rms) thrusters) and ships ICIT (Beharie and Side, 2011a)

165-175 dB re 1μPa at 1m (rms) OSPAR Commission (2009) – probably includes pile drilling for installation and also vessel activity. Wave and tidal devices ICIT estimate excluding <160 dB re 1μPa at 1m (rms) during installation and vessel activity device operations

We consider it unlikely that pile driving as opposed to pile drilling is likely to be a favoured means of fixture for tidal and wave devices (though obviously it has been a favoured technique for some coastal and offshore wind developments). The reasons for this are that for most tidal sites the seabed will be bedrock or a thin veneer of sediment over bedrock making pile driving impractical. For wave devices offshore it is likely that conventional anchors will be used, and for those inshore, again the presence of wave swept bedrock will require pile drilling rather than pile driving.

Following concerns over the avoidance behaviour of a number of marine mammals, in particular of harbour porpoises (Phocoena phocoena) during pile driving at the German Alpha Ventus field, a number of NGOs (e.g. the Whale and Dolphin Conservation Society and Friends of the Earth, Germany) have called for the replacement of pile driving by pile drilling technology (Brensing, 2010).

Page 7

Their investigations suggested predicted sound pressure levels of 161dB using Herrenknecht jack‐up pile drilling platforms; the lack of a cited reference pressure level for underwater noise, however, leaves some doubt over this source. The only previous scientific evaluation of sound pressure levels from pin pile drilling appears to be that of Nedwell and Brooker (2008). This work was a COWRIE commissioned study undertaken during the drilling of pin piles for the SeaGen tidal turbine device in Strangford Lough. From a fit to the sound pressure level measurements taken at varying distances from the drilling operation, Nedwell and Brooker indicate a source sound pressure level of 162 dB re 1μPa at 1m. This operation used a Seacore / Wirth B5 pile top drill with a drill bit size of 1.15m and corresponds to rms pressure levels of 125Pa at 1m. By comparison the rms pressure source levels that have been inferred from measurements during pile driving operations are of the order of 2.5‐3 x 106Pa.

MEASUREMENT DIFFICULTIES

Historically noise measurements have been made from a silently drifting vessel, which in calm seas with little tidal movement can stand off some distance from the hydrophone, but which, in the presence of waves, will inevitably itself generate noise from the slapping of water on the vessel hull. If one fixes a hydrophone in place (in some manner to the seabed, for example) then it will provide a measure of the flow of water particles past it. Indeed in waves it will mimic a pressure sensor which is, of course, often used for the measurement of the passage of waves at a fixed point. The solution to this which we have sought is the deployment of a free drifting spar buoy, designed to minimize noise from the wave action on the buoy structure itself, but which can be deployed and remotely monitored from some distance. In tidal streams it is free to drift with the water flow.

Figure 5 The Hydata Sonobuoy

The Hydata Sonobuoy is designed for remote sensing of calibrated acoustic data in marine environments. It is designed to remain in a stable vertical position in waves thereby minimising spurious noises generated by hydrodynamic flows against the hydrophone. Able to be remotely controlled from a control station up to 2km away via its 54 Megabit 2.4GHz wireless link, the system

Page 8 can also transmit data back to the control station. Range and speed of the data communication link depends on sea‐state and antenna type on the shore station. Typical data transmission performance is 54Mbps <1km to 1Mbps at 2km. The weight of the Sonobuoy is approximately 37kg which enables one to two personnel to deploy and retrieve it from a small vessel such as a RIB allowing quick and effective positioning within tidal flows. The ability to readily retrieve the Sonobuoy in fast flowing tidal channels is of particular benefit for risk assessment evaluations.

Recordings are presently taken at 44.1kHz sample rate and a 16 bit depth giving a theoretical maximum dynamic range of 96dB and a frequency detection rate of up to 22kHz. The pre‐amplified, cylindrical hydrophone is designed for the detection of infrasonic, audible, and ultrasonic sounds with a linear frequency range of 0.016 ‐ 44kHz and a sensitivity of ‐185dB, re 1V/µPa.

A continuous dataset of one second positional coordinates of the Sonobuoy are taken using an on‐ board Globalsat GPS data logger mounted on the antenna mast. This unit is equipped with inbuilt ‘European Geostationary Navigation Overlay Service’ (EGNOS) calibration which provides a positioning accuracy to within 1.5m (ESA, 2011). It is fitted with light reflecting panels to enhance laser range finding measurements if required.

Based on readily available technology the Sonobuoy has the possibility of easily being adapted for a variety of other data acquisition purposes with the addition or replacement of complementary measuring instrumentation.

MEASUREMENT AND PREDICTION OF TRANSMISSION LOSS FOR ZONES OF EFFECT

In general the description of sound transmission loss from a sound source underwater and the corresponding zone of effect for a vulnerable target species requires:

1. The determination of the sound pressure level of the sound source (usually for continuous sounds in rms dB re 1μPa at 1m, or e.g. peak to peak for impulsive sounds ). 2. The determination of background levels in the area occupied by the target species. 3. The setting of appropriate thresholds of concern for the target species. 4. A model of underwater sound attenuation, which describes transmission loss appropriately for the area under consideration. 5. The determination of the zone within which such thresholds are exceeded or the distance required before background noise levels are likely to mask any signal from the sound source.

In particular in Orkney there are concerns, which have been raised by the statutory conservation agency (Scottish Natural Heritage) over the welfare of marine mammals, and in particular at the EMEC Fall of Warness tidal test site, of seals. These are focused, particularly in the summer months during pupping, on the influence of transmitted noise on seal haul‐outs on the nearby Seal Skerry and adjacent coast (see Figure 6). The species is the harbour seal (Phoca vitulina) and haul‐outs are some 1.4km from the nearest berth at the EMEC site. The swimming activity of pups is believed to be limited to water depths less than 5m.

Page 9

Figure 6 Showing the location of the berths for tidal device testing at the EMEC Fall of Warness site, and the proximity to Seal Skerry and the adjacent coast

BACKGROUND NOISE

A measure of the ambient noise environment is therefore particularly useful in assessing the influence of noise from point source maritime activities. In shallow coastal regions background noise levels can vary from 90‐155dB re 1μPa (Nedwell et al., 2003). In recent studies in the Moray Firth background noise levels ranged from 104‐121dB re 1μPa (Bailey et al., in press) and in the Strangford Lough study from 115‐125dB re 1μPa (Nedwell and Brooker, 2008). Maritime traffic can have a significant influence on these levels and thus in the vicinity of construction vessel traffic for the Beatrice Wind Farm in the Moray Firth Bailey et al. record background noise levels increasing to 138dB re 1μPa. A comparison of the power spectrum levels of background noise for both the Moray Firth and Strangford Lough indicated comparatively elevated background noise levels for the main channel in the Fall of Warness (Wilson and Carter, 2008). More recent work at the Fall of Warness site suggests background levels in the main channel of between 112dB and 127dB re 1 μPa with a mean of 122dB (Beharie and Side, 2011b). At times of vessel activity though, this increases and mean values of 124dB and 126dB have been recorded. Even where vessel activity is present mean values closer inshore towards areas used as seal haul out are lower, and values in the main channel dependent on tide.

In many recordings it is possible to detect Acoustic Deterrent Devices (seal scarers) being used by fish farms many kilometers distant.

Page 10 THRESHOLDS OF CONCERN FOR TARGET SPECIES

In addition to the level of background noise there are a number of thresholds that have gained acceptance in the scientific literature when considering the effects of underwater noise on vulnerable species:

1. Auditory injury or permanent threshold shift in hearing (PTS) 2. Temporary threshold shift in hearing (TTS) 3. Behavioural disturbance thresholds (BHT) – sometimes ranked as minor or major. 4. Hearing Threshold (sometimes “ht”) or auditory threshold for the species concerned

Generally the latter, auditory thresholds, are used to analyse measured data to determine perceived noise levels for the species concerned. This mirrors the approach employed with human perception of noise levels (see Mohl, 1968 for an example of auditory sensitivity in the harbour seal).

Suggested rms levels for PTS and TTS in pinnipeds of 218dB and 212dB re 1μPa (Southall et al., 2007) exceed any that are forecast from the continuous source sound pressure levels suggested above (see Table 3), and this is equally true for peak broadband thresholds with the assumption that pile driving will not be used. Put simply, auditory impairment in the harbour seal could not occur even immediately adjacent to a seabed pile drilling operation. It is worth making comparison at this stage with pile driving activities. For example Bailey et al. (in press) conclude that for pinnipeds PTS onset would occur within a 20m zone of the pile driving operation for the Beatrice Wind Farm in the Moray Firth and TTS onset within a 40m zone.

Historically the behavioural disturbance threshold proposed by the US National Marine Fisheries Service (NMFS) for the lower limit of auditory damage (180dB re 1μPa) has been used, and this is shown in Figure 7.

Other work (Harris et al., 2001) has suggested Minor Disturbance and Major Disturbance thresholds of 160 and 200 dB re 1μPa (usually considered as peak to peak, not rms). Bailey et al. note that these disturbance thresholds are really precautionary: few field studies have been conducted, and the longer term effects of the behavioural responses have not been determined. Again it is important to remember that high‐energy, short bursts from pile driving, underwater explosives and seismic sound sources are not directly comparable with continuous noise sources. In the graphical outputs of this paper the US NMFS threshold of (180dB re 1μPa) has been used, and does provide the basis for the determination of safety zones in California and Sakalin (see, for a review of international safety standards in this respect, Compton et al., 2007). The 160dB Minor Disturbance threshold is included in Figure 7 as a precautionary reference point only.

MODELS OF TRANSMISSION LOSS

There are a number of commercially available software packages, as well as very well supported open source software, for modeling sound attenuation and transmission loss. Generally these fall into one of the following categories:

• Ray tracing models such as Bellhop • Normal mode models such as Kraken

Page 11 • Parabolic Equation models such as PECan • Wavenumber integration models such as Scooter.

Most have enjoyed a wide application for oceanic noise modelling applications, where there is stratification in deep water, to which many are particularly suited, and where the source levels and frequency spectra are well known (e.g. naval applications and seismic surveys). There is also now an acoustic toolbox with MATLAB routines available as open source code from the Ocean Acoustics Library which is supported by the US Office of Naval Research.

There is one example of the use of the normal mode program Kraken to model the Pelamis in its installed positions off Portugal (Patricio et al., 2009) which was extended by Barrio (2009) to examine perceived levels of sound for the auditory thresholds of the harbour porpoise (Phocoena phocoena). Unfortunately in both these studies the assumed broadband source levels of 175dBrms and 170dBrms from the Pelamis array, while producing observable outputs from Kraken seem unrealistically high. For these, a hypothetical source was generated using pure tones of 200, 400, 600 and 800Hz superimposed on a continuous broadband noise spectrum (200 to 1000Hz).

In practice it is unlikely that either the frequency or source sound pressure levels for many operational devices, and the sources of noise arising from their installation methods will be known, and thus a fit of actual field data to simple spreading models tends to be favoured, particularly as such waters are well mixed.

In an unbounded medium sound waves will spread spherically and the intensity will decrease with distance from the source. In such purely geometrical cases sound attenuation is described by a simple spherical spreading relationship for transmission loss: P TL = log20 10 rP )( Here P is the source sound pressure (at 1m) and P(r) is the sound pressure at distance r. The transmission loss at any distance can thus be calculated and it can be readily shown that:

TL= log20 10 r which is referred to as the spherical spreading (SS) law (r is the distance from source). If the geometry approximates to a channel (with horizontal extent much greater than depth then a cylindrical spreading (CS) relationship for wave propagation geometry is suggested:

TL = log10 10 r In addition to these purely theoretic geometrical considerations there are many other factors influencing the propagation of underwater sound waves. Generally sound waves from a source close to the seabed or surface will travel along multiple sound paths before reaching a single point at distance. Multi‐path propagation is common where a source is located relatively close to a boundary (sea surface or seabed) and when the depth is small in relation to the horizontal propagation distance. In this case while some sound waves may follow a path directly from the source to the receiver others will be reflected from the surface and seabed many times resulting in constructive and destructive interference, with the received sound pressure level being reduced as a consequence of reflection losses.

Page 12 In these circumstances the smoothness of sea surface, and physical nature of the seabed and its topography, are critical to the received sound pressure levels at any point. Because of constructive and destructive interference it is possible for the resulting sound field to contain an alternating series of sound pressure maxima and minima. This reflective phenomenon is sometimes referred to as the Lloyd mirror effect. Importantly it should be remembered that sound can also be transmitted through bedrock at the seabed, adding a further consideration to this complex mix.

Frequently the literature suggests an intermediate form (IS) of the above transmission loss equations to take into account these effects:

TL = log15 10 r which, with the other spreading forms discussed is shown graphically in Figure 7.

Figure 7 Showing 3 forms of the transmission loss model (SS ‐ Spherical Spreading; CS – Cylindrical Spreading; IS – the intermediate form) and attenuation to background levels with distance from a 170dB source.

Thus in the literature we find a variety of formulations for shallow water transmission loss, many of this general form, but with additional terms where models are fitted to data. Importantly in many shallow water studies, where multi‐path propagation occurs, the level of transmission loss observed has required the use of values greater than 20 (greater than that used to describe spherical spreading). These appear frequently where piling operations are involved in shallow coastal waters, and where peak to peak or impulse metrics are analysed, rather than rms dB, see for example Malme et al. (1986) and Nedwell et al. (2003). In general for work over much longer distances transmission loss formulae may also include an absorption coefficient (α) as formulated:

= log10 + αrrNTL

Page 13

The above was used by Bailey et al. (2010) to describe the sound propagation from pile driving in the Moray Firth during the installation of offshore wind turbines, where peak to peak background sound pressure levels were not reached until some 10km distant.

In general geometric spreading models of this kind are used to fit field data, and from Figure 7 it is evident that it is important to understand transmission losses in the Fall of Warness with seal haul‐outs and pupping areas within 1.4km of development activities. The 170dB re 1μPa at 1m source level used in this Figure is certainly within the range of possible values derived from Table 3.

MEASURED VALUES

The only other study which in some way corresponds to drilling activities in the Fall of Warness is for pile drilling during the installation of the Marine Current Turbine’s SeaGen, in Strangford Lough (Nedwell and Brooker, 2008). The fit to measured data from this is shown in Figure 8 and suggests a source level at 1m for the drilling of 162dB and transmission losses of an intermediate spreading form of 16log10r.

Figure 8 Showing fit of field data ranges to empirically derived model of sound attenuation from pile drilling for the SeaGen tidal turbine in Strangford Lough. (Source: from Nedwell and Brooker, 2008).

This, and a more conservative 170dB, with transmission losses of 15log10r has been used in predictive assessments for the Environmental Statements that each developer on the test sites at the Fall of Warness must produce. A recent study, however, has enabled more confidence in these approaches.

Page 14 The installation of the anchor blocks and moorings for the Scotrenewables tidal turbine provided an opportunity for the measurement of sound pressure levels from the workboat. The Voe Viking workboat used for the installation of the anchor blocks, chains and moorings has three 800hp engines (with 1700mm diameter propellers) and is 26m in length. The vessel complies with the MCA Code of Practice for the Safety of Small Workboats & Pilot Boats which requires that a surface noise level of 65dB (A) should not be exceeded.

Typical but larger twin engine work boats have recorded underwater noise levels of 159dB (re 1μPa at 1m) and for the environmental statement it was estimated that the addition of one further engine would result in a combined sound pressure level from all three of 161dB (re 1μPa at 1m). The Environmental Statement noted that the dominant frequencies for workboats are in the 400‐650Hz range, and suggested a worst case prediction of a SPL of 156dB rms at 1m for a single thruster, and a SPL of 161dB rms at 1m where all three thrusters were operating simultaneously.

Figure 9 shows the power spectral density of three 1s samples of background noise from recordings made on the site. All seem to exhibit somewhat raised levels between 50Hz and 500Hz suggesting that engine room noise from the Voe Viking is detectable in all samples.

Figure 9 Comparison of three recordings of background noise covering the range of estimated levels.

From measurements made using the Sonobuoy during the installation it was uncertain as to when 3 thrusters were simultaneously working and there was no obvious way of detecting the number of thrusters in use at any one time from the hydrophone recordings. Figure 10 shows a recording which captures 5 pulses of thruster noise in close succession. The duration of thruster pulses was highly variable sometimes being less than a second and on other occasions several tens of seconds. It can be seen from the spectrogram that the dominant frequencies are less than 1kHz. Figure 11 shows the power spectral densities for comparison with those of background noise in Figure 9.

Obviously the levels are higher than those for the background noise in Figure 9, and noticeably so in the 100Hz to 650Hz frequency range, which is around 15dB above background levels. Nonetheless it also suggests that engine noise (even with no thrusters operational) was contributing to background noise

Page 15 levels during the study. The variability across the frequency range is evident even in recordings with the same sound pressure level (of 131 dB rms re 1μPa).

Figure 10 Spectrogram showing the intermittency of noise from the vessel thrusters

From the data we have developed a first attempt at a model for noise propagation at the Fall of Warness, which determines transmission loss with distance, and which we hope to confirm shortly with new studies on the site.

Figure 11 Comparison of 3 extracts of thrusters noise from the Voe Viking

A number of assumptions were made in doing this, and each thus provides a caveat for the robustness of the resulting model.

Page 16 1. As it was impossible to determine whether all thrusters were operating on full power at any point only the maxima in any sample were selected. This was helped by listening to the recordings with spectrograms and dBrms plots of the time series. 2. Given that the buoy was drifting and the Voe Viking also moving we were limited to samples where distances could be reliably estimated.

Given these constraints to the reliability of the method it was nonetheless possible to isolate 85 samples, which appeared to correspond to maximum levels when thrusters were in use. These data are plotted in Figure 12 along with the fitted model. The model parameters determined by linear regression suggest a sound pressure level for the thrusters of 162 dB rms re 1μPa at 1m. This is consistent with the estimation made in the Environmental Statement, though perhaps a little higher than anticipated, probably owing to the selection of maximum recorded values.

Figure12 Best fit noise propagation model with distance from the thruster source

Using the transmission loss model empirically derived from samples of the recording of the Voe Viking’s thrusters (of 14.4Log10r) we were then able to determine likely source sound pressure levels for the placement of the anchor blocks and chain noise.

• For the placement of the anchor block itself, the recorded peak to peak sound pressure levels was equivalent to 167dB at 1m.

• Greater sound pressure levels were associated with the chain noise recorded during this period, equivalent to a maximum of 173dB at 1m (peak to peak).

• The recorded sound pressure levels associated with the placement of the mooring’s secondary clump weight were lower, with peak to peak values always being less than 154dB at 1m.

In this work we noted that while the intermediate case of transmission loss spreading appears to be an appropriate selection, it is likely that changes in seabed topography across the Fall of Warness will mean that transmission loss will vary depending on the direction in which measurements are taken. At the Conference we hope to report on a recently completed study which for another Fall of Warness

Page 17 site, and this time for pile drilling, suggested source levels of 167dB and a transmission loss of 16.6Log10r from a linear model of fitted data (Beharie and Side, 2011b).

SOUNDFIELDS FROM WAVE DEVICE ARRAYS

While much of this paper has been concerned with noise from installation of tidal energy devices we turn our attention finally to another underwater noise related concern of relevance to wave and tidal devices in arrays. The question posed is whether the soundfield encountered by an animal moving into a large multi‐device array might be such that it causes confusion and aberrant behavior as a consequence. So, for example, if we consider a single source of broadband noise from a wave device, close to the surface, and trace rays from this, as modelled in Figure 13, we can see that as the wave passes over the device the pattern of the ray trace changes.

Figure 13 Fixed noise source with regular wave moving from right to left, note the formation of a focus, by the wave crest, in the ray trace as the source approaches the wave trough.

As the wavecrest passes over the source, the rays are focused by the crest to a point of convergence. The reflection from the surface wave causes the wavefront to fold over on itself, with a focal point and caustics which diverge from this to form a fan. This phenomenon, where wave focusing gives rise to rapidly fluctuating signal arrivals with high intensity, has been observed in studies of underwater acoustic communications (Preisig and Deane, 2004) and has been experimentally created in tank tests by Tindle et al. (2009). For a single source and regular waves it is possible to use a wavefront model to generate the changing soundfield with the passage of the waves. If one takes the angles at source in one of the traces above we can determine for any horizontal section the depth of each ray as a function of its angle at source, and thus the eigenrays for any given point (receiver) can be determined. This approach can be extended to find the ray travel time along the path of each ray and hence to reconstruct the signal at any point on the horizontal section. For simple pulsed signals at 200kHz, in regular waves, under the controlled conditions of a test facility, Tindle et al. (2009) have shown an excellent correlation between measured and modelled data. While the approach is an appropriate one for arrays of wave devices the situation in reality is far more complex. There are a number of other variables that need to be considered.

Page 18 If we consider the noise source: firstly, it is unlikely to be stationary, but itself will move in response to the waves. For some devices (such as the Pelamis) it would be inappropriate to consider a single source, multiple sources would be necessary, again all moving in response to the passage of the wave. In terms of the noise sources the model would also need to be scaled up to a multi‐device array. Finally we would expect the source levels to vary with wave height, and possibly the broadband spectra as well.

If we consider the waves: firstly the waves are unlikely to be regular, but instead will be describable by some general spectrum such as Bretschneider or Jonswap. It is very possible to generate sequences of random waves from these, but equally likely that the features of each wave will change as energy is extracted from it by the device.

Finally consider the difficulties in empirically validating such a model at sea!

It is undeniable then that an animal moving close to the surface in an array of wave energy convertors will experience intensified wave focusing events. Some experimental data suggests that this may be by as much as 10dB (Preisig and Deane, 2004). In such situations it may be that approaches (described in earlier section of this paper), which determine source levels and transmission losses to well defined zones of disturbance are less appropriate, though obviously source levels will be important in defining the extent of the changing soundfield. Our approach now is to develop a wavefront model which is able to characterise in time the motion of the wave and noise source for regular, and random waves generated from a known spectrum. This should at least provide some indication of the possible changing patterns of the soundfield, and the geographical extent over which such phenomena occur for given source levels.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the support of Scotrenewables, and of the EPSRC Supergen Marine (Workstream 10), and Scottish Funding Council awards for AMRECS, and Hebridean Futures.

Page 19 REFERENCES

Barrio A (2009). Modelling underwater acoustic noise as a tool for coastal management. Masters Thesis, Faculdade de Ciências e Tecnologia da Universidade do Algarve. 107pp.

Bailey H, Senior, B, Simmons D, Rusin J, Picken G, and Thompson P (2010). Assessing underwater noise levels during pile‐driving at an offshore windfarm and its potential effects on marine mammals. Marine Pollution Bulletin 60, 888‐897

Beharie R and Side J (2011a). Acoustic Environmental Monitoring: North‐west mooring leg installation for the Scotrenewables SRT250 tidal turbine. Report No. 2010/04/SR, ICIT, Heriot‐Watt University, Orkney Campus.

Beharie R and Side J (2011b). Acoustic Environmental Monitoring: Foundation drilling operations for the Voith tidal turbine. Report No. 2011/08/AQ, ICIT, Heriot‐Watt University, Orkney Campus.

Brensing K (2010). No piling – No problem. Poster presentation by the Whale and Dolphin Conservation Society on behalf of WDCS, IFAW, NABU, DEEPWAVE and Freunde Der Erde. [accessed Nov 2010] http://www.bsh.de/de/Das_BSH/Veranstaltungen/Cetacean_Society/Poster_NGOs.pdf

Caltrans (2001). Fisheries Impact Assessment – Oakland Bay Bridge East Span Seismic Safety Project. PIPD EA 012081. Caltrans, San Francisco. 51 pp

Compton RA, Goodwin L, Handy R and Abbot V (2007). critical examination of worldwide guidelines for minimising the disturbance to marine mammals during seismic surveys. Marine Policy 32(3), 255‐262.

Davies I M, Robertson M, Jones G, Watret R, and McKie J (2011). Test centre to commercial arrays: Progress towards wave and tidal energy in Scotland. ICES Annual Science Conference 2011, ICES CM 2011/S:12. Gdansk, Poland.

ESA (2011). The European Space Agency: The European Geostationary Navigation Overlay Service (EGNOS) http://www.esa.int/esaNA/GGG63950NDC_egnos_0.html [accessed August 2011].

Harris R E, Miller G W and Richardson W J (2001). Seal responses to airgun sounds during summer seismic surveys in the Alaskan Beaufort Sea. Marine Mammal Science 17, 795‐812.

Kastak D and Schusterman R J (1998). Low frequency amphibious hearing in pinnipeds: Methods, measurements, noise and ecology. Journal of the Acoustical Society of America, 103(4), 2216‐ 2228.

Malme C I, Smith C W and Miles P R (1986). Characterisation of geophysical acoustic survey sounds. Minerals Management Service. Los Angeles, California. Report No OCS/MMS‐86/0032.

Mohl B (1968). Auditory senstivity of the common seal in air and water. Journal of Auditory Research, 8, 27‐38.

Page 20 Nedwell J R, Langworthy J and Howell D (2003). Assessment of sub‐sea acoustic noise and vibration from offshore wind turbines and its impact on marine wildlife; initial measurements of underwater noise during construction of offshore windfarms, and comparison with background noise. A report commissioned by COWRIE (Subacoustech Report No 544R0423).

Nedwell J R and Brooker A G (2008). Measurement and assessment of background underwater noise and its comparison with noise from pin pile drilling operations during installation of the SeaGen tidal turbine device, Strangford Lough. A report commissioned by COWRIE (Subacoustech Report No 724R0120). 33pp

NRC (2003). Ocean Noise and Marine Mammals. National Research Council, National Academies Press. 192 pp.

OSPAR Commission (2009). Overview of the impacts of anthropogenic underwater sound in the marine environment. Report 441/2009, OSPAR Commission, London. 133 pp.

Patricio S, Soares C and Sarmento A (2009). Underwater Noise Modelling of Wave Energy Devices. Proceedings of the 8th European Wave and Tidal Energy Conference, Uppsala, Sweden, 2009. pp 1020‐1028

Preisig J C and Deane G B (2004). Surface wave focussing and acoustic communications in the surf zone. J. Acoust. Soc. Am. 116 (4), PP2067‐2080

Southall B L, Bowles A E, Ellison W T, Finneran J J, Gentry R L, Greene Jnr C R, Kastak D, Ketten D R, Miller J H, Nachtigall P E, Richardson W J, Thomas J A, Tyak P L (2007). Marine mammal noise exposure criteria; initial scientific recommendation. Aquatic Mammals 33, 411‐521.

Tindle C T, Deane G B and Preisig J C (2009). Reflection of underwater sound from surface waves. J. Acoust. Soc. Am. 125 (1), pp66‐72

Urick R J (1983). Principles of Underwater Sound. McGraw‐Hill, New York.

Wilson B and Carter C (2008). Acoustic monitoring of the EMEC Fall of Warness tidal‐stream test site. SAMS Research Services Ltd., Oban.

Page 21