Current state of knowledge of effects of offshore renewable energy generation devices on marine mammals and research requirements

5 Consequences of Impacts

In Section 5.1 we consider the possible cascade of effects - from individual impacts to population vital parameters. In Section 5.2 we consider modelling frameworks that can use such information to predict population level consequences.

5.1 Individual

The potential proximate impacts outlined in Section 4 can have various consequences for individual animals within a population depending on different factors at the time of exposure. These include species, sex, condition, age, sociality, behaviour and season. Impacts considered at the individual level may be manifest at the population level if responses are sufficient to produce effects on any of the vital rates (namely reproduction or survival, two major drivers of population trajectories and abundance) or on permanent emigration rates. Assessing such impacts at the individual level will therefore be considered in the context of the ultimate effects on the population in the short, medium or long term, recognising that there may also be transient effects and welfare issues for individual animals.

The pathways that the three main proximate impacts, namely noise, physical contact and habitat alteration may take to potentially affect either survival or reproduction or both are shown in Figure 1. The key points within each are discussed below.

5.1.1 Noise

5.1.1.1 Direct effects on survival and reproduction

Although there are no specific studies on the effects of noise on marine mammal vital rates, particularly reproduction, much attention has focused on the direct effects of noise on the auditory system (Southall et al. 2007), particularly following the mass strandings of beaked whales in the Bahamas, Canary Islands and Mediterranean, that, in some cases, have clearly been associated with the use of midrange tactical military sonar (Wartzok et al. 2005; D'Amico et al. 2009; Filadelfo et al. 2009). Follow-up studies have also considered the non-auditory physiological effects of sound. These include resonance effects in which air spaces and gas-filled tissues could be theoretically driven into resonance by acoustic energy, (Finneran 2012) and rectified diffusion in which microscopic bubble nuclei are formed in the presence of high-intensity sound (Crum et al. 2005). Such direct physical effects, generally resulting from very loud impulse sounds in specific frequency ranges have the potential to be individually fatal (Nedwell et al. 2007b). Extremely loud noises exceeding 220 dBre 1 µPa are likely to occur only at close range during piling activity.

Studies on sheep and mice have reported hearing damage in the foetus of pregnant females exposed to noise during gestation (Griffiths et al. 1994; Pierson 1996) although the strongest responses were reported in the studies which exposed the ewes to intense broadband noise (120 dB sound pressure level for 16 h). In humans, a study on the effects of noise exposure to the foetus of mothers that worked while pregnant, in noise conditions ranging from 65 to 95 dBA-8h (Lalande, Hetu & Lambert 1986) found a three-fold increase in the risk of having a high-frequency hearing loss in the children whose mothers were exposed to noise in the range of 85 to 95 dB Leq, and a significant increase in the risk of hearing loss at a frequency of 4000 Hz when these exposures involved a strong component of low-frequency noise. Although it is difficult to directly compare these studies with the exposure levels to underwater sound that are likely to occur as a result of OREG activity the results of these and other studies on noise-induced hearing loss in the foetus (Pierson 1996) suggest this is a potential direct effect risk. Further studies on auditory brainstem responses ( ABR) in seal pups of females potentially exposed to OREG construction and operation noise are warranted. Noise attenuation differences between species may mean that the received levels between mice, sheep and marine mammals are different and that higher exposures are required before the same effect is seen but permanent threshold shifts ( PTS) in pups and juveniles could have consequences for their longevity, survival and reproduction as they may be unable to forage well or reproduce successfully due to impairment in hearing abilities necessary to navigate and mate.

Research gap
Title	Code	Details	Status	Reporting date
Auditory brain stem responses in seal pups of females exposed to OREG construction and operation noise	ABRPuP	There is some evidence from studies in mice, sheep and humans that foetuses exposed to noise during gestation might be at risk of some hearing loss. This effect could be investigated in seal pups (using ABR response measures) from females exposed to OREG noise.	Not funded	NA

Ear infections are known to cause hearing damage in harbour seals (Ketten et al. 2011). An additional pre-existing impact to consider, particularly relevant to marine mammals in some regions, is the impact of exposure to polychlorinated biphenyls ( PCBs) on hearing (direct ototoxicity). Crofton and colleagues (Crofton & Rice 1999; Crofton et al. 2000; Lasky et al. 2002) demonstrated the ability of PCBs to disrupt the development of the cochlea in rats, by disrupting thyroid function. A relationship between thyroid hormone disruption and PCBs in seals has been well established in a number of studies (Brouwer, Reijnders & Koeman 1989; Tabuchi et al. 2006) and harbour seals in some regions in Scotland (particularly on those on the southwest coast such as Islay and Jura where OREG developments are planned) have high levels of PCBs in their blubber (Hall & Thomas 2007). However, no studies on the ototoxic effects of PCBs in seals prior to OREG developments have been carried out. Further studies on ABR responses in seals in relation to PCB exposure and age are needed to investigate this, particularly to establish any existing negative impacts prior to OREG construction and to determine the potential for exacerbating effects.

Research gap
Title	Code	Details	Status	Reporting date
Ototoxic effects of PCB exposure in seals	PCB	Previous studies have shown that harbour seals in particular captured on Islay and Jura have very high levels of PCBs in their blubber. PCBs have the potential to cause cochlear damage during development. Further work on hearing loss in these animals in relation to their age and PCB exposure levels would determine if they have pre-existing damage caused by these pollutants.	Not funded	NA

5.1.1.2 Indirect effects on survival and reproduction

If noise exposure is below some critical energy flux density limit, there will be a temporary loss of hearing known as a temporary threshold shift ( TTS) in hearing sensitivity (Wartzok et al. 2005). This is because the hair cells in the ear will eventually return to their normal shape. However, if the sound exceeds a higher limit the hair cells in the cochlea become damaged and die. This is known as the permanent threshold shift ( PTS). Threshold limits for these effects vary among species and individuals so need to be characterized statistically.

Wind- OREG construction noise is probably the most widely recognised potential impact with particular emphasis on piling noise. Pile driving produces high sound source levels and the potential impacts are of most concern (Dolman & Simmonds 2010). In addition, there is some evidence that other non-auditory effects of noise exposure may need to be considered as loud noise exposure may directly affect reproduction in mammals, largely mediated through the neuroendocrine system ( Figure 1).

The hypothalamic pituitary axis ( HPA) is the integral part of the endocrine system responsible for maintaining homeostasis. The neural core consists of the hypothalamus which contains neurosecretory neurons that synthesise hormones such as dopamine and corticotropin-releasing hormone and the pituitary gland that produced adrenocorticotropin. The adrenal gland makes up the third arm which secretes catecholamines (epinephrine and norepinephrine, also known as adrenalin) and steroid hormones such as cortisol, corticosterone and aldosterone (the corticosteroids). Production of these proteins increases after a wide variety of stressful stimulus to allow the animal to respond quickly, the so-called 'fight or flight' response. The physiological changes are collectively known as the 'stress' response. The ability of an animal to return to and maintain a balanced state is central to the impact of noise on individual reproduction and survival.

The connections between the physiological, cellular, and genetic processes involved in individual responses to noise and their effects on behaviour and fitness suggests an integrative framework is necessary when assessing how and why animals are affected (Kight & Swaddle 2011). Given the cascade of interlinked effects of stressful noise in a receptor like a marine mammal it's not possible to find a single hormone or protein response marker. To address this issue a number of corticosteroid, other hormone and protein markers to assess stress need to be developed validated and dose-response curves in relation to noise established (Wartzok et al. 2005). Wartzok et al. (2005) also recommend the "development of a sampling package that could take blood samples on a controlled basis and stabilize hormones for later analysis or process samples "on-board" for corticosteroids at various stages of a controlled exposure experiment would be invaluable for determining the stress that the sound is producing". SMRU and scientists from the University of Tokyo in conjunction with a company called Little Leonardo in Japan are currently in the process of developing a remote blood sampling device for phocid seals and early results from the studies on the captive animals at SMRU are very promising (Takei and Hall pers. comm.).

Research gap
Title	Code	Details	Status	Reporting date
Hormone and protein markers in marine mammals in relation to noise exposure	HORM	Establishment of hormone and protein markers and noise associated dose-response relationships for the key marine mammal species found in Scottish waters. Captive studies in harbour and grey seals could determine the variability in a range of potential markers in blood, faeces, urine and skin samples taken from animals exposed to various sound sources and levels.	Not funded	NA

There are a number of studies which suggest that noise stress is particularly damaging to females, probably due to their size, hormone expression and the costs of reproductive investment (Kight & Swaddle 2011). Rasmussen et al. (2009) found that mice exposed to construction noise produced a significantly higher number of stillborn mice compared to the control group when exposure occurred early in gestation. In particular the average litter size of the mice exposed to 70-90 dBA (1 h daily exposure to a 6 min continuous loop of concrete saw cutting with a dominant energy between 2 and 8 kHz) during the peri-implantation period was significantly smaller (p=0.005) than that of controls (5.8 mean litter size in exposed compared to 10.2 mean litter size in unexposed). Zakem & Alliston (1974) exposed mice to 83 and 95 dBA of noise intermittently during gestation also reported increased preimplantation mortality, decreased litter size, embryo size and weight among the exposed offspring. Kimmel, Cook &Staples (1976) exposed pregnant mice to 100 dB of noise on various days during the first two weeks of gestation and reported significantly increased resorption rates and decreased numbers of live foetuses per litter in each of the treated groups of animals. Meyer, Aldrich & Easterly (1989) reviewed these and a number of other studies on rats and mice (Nawrot, Cook & Staples 1980; Nawrot, Cook & Hamm 1981; Cook, Nawrot & Hamm 1982) finding that embryotoxic studies of noise supported an exposure effect although the data were quite inconsistent among the various exposure conditions. However, many did report statistically significant effects that included intrauterine growth retardation, foetal mortality, increased litter resorptions and teratogenesis (malformations in the embryo). Although the specific mechanisms for these effects were often unreported some studies have suggested that it is likely that increased catecholamines in blood cause a reduction in blood flow to the uterus and ovaries resulting in foetal death (Esquivel, Castro-Vazquez & Rosner 1974). Other studies suggest an immune-endocrine disequilibrium as a result of increased glucocorticoids, resulting in an unsustained pregnancy (Arck et al. 1995).

However, although mice have a similar hearing range and audiogram to those of pinnipeds and odontocetes, particularly at the high frequencies (mouse range 4 kHz to 45 kHz at 70 dB sound pressure level ( SPL), odontocetes have best frequency hearing at between 10 and 100 kHz at better than 70 dB re 1 µPa, for pinnipeds between 1 and 30 kHz at better than 70 dB re 1 µPa) it is difficult to determine how comparable the data from these model species would be for pinnipeds and odontocete cetaceans. Comparing the levels of noise to some reference, such as the noise required to produce a TTS might be a useful benchmark. The noise exposure in the reproductive effects experiments were often at or around the level causing TTS in mice. Studies of bottlenose dolphin using pure tones and octave band noise ranging from 1 s for pure tones up to 54 min for octave band noise at frequencies between 3 and 75 kHz found the observed minimum intensity at which TTS was observed to range from 160 to 194 dB re 1 µPa. This reference level may then provide some guidance for the magnitude of reproductive effects taking into account the difference between SPL sensitivities in air and in water. However, we would guard against using this as some reference threshold, given the difference in species sensitivities, reproductive strategies (single births compared to litters of offspring) etc. it is not clear how relevant or transferrable the results of the studies on laboratory animal models are for assessing the direct risks of underwater noise exposure to marine mammal reproduction. In addition the laboratory animals were unable to escape the noise. Although the most likely response to loud impulsive noises such as pile driving would be for the marine mammals to move out of the area when the noise becomes intolerable, effects may occur at levels within the zone of toleration. Animals may tolerate chronic the noise, perhaps because of productive foraging habitat, with potential consequences for future reproductive success.

5.1.1.3 Displacement, avoidance and permanent emigration.

There is a significant body of data showing that whales avoid underwater sounds starting at 110-120 dB re 1 µPa. Beluga whales for example fled icebreaker noise at received levels of 94-105 dB rms re 1 µPa returning in 1-2 days to the area where the received noise was 120 dB re 1 µPa (Finley et al. 1990). Kastelein et al. (1997) reported captive harbour porpoise avoided exposure to high frequency pingers with a source level of 103-117 dB re 1 µPa at 1 m and received levels of 78-90 dB rms re 1 µPa. In captive bottlenose dolphins Finneran & Schlundt (2004) found that the probability of adverse reactions to noise increased with increasing received levels from 160 to 200 dB rms re 1 µPa at 1 m. Seals have a lower maximum hearing frequency and sensitivity than odontocete cetaceans with a high frequency cut off between 30 and 60 kHz. However, there is a paucity of information and data on the response of phocid seals to known levels of noise such as piling. Some data exists on the startle response (Gotz & Janik 2011) and on the effect of seismic air guns (Gordon et al. 2003). This Research Gap ( DECC2) is considered in section 5.1.1.1.6, (Nedwell et al. 2007b) suggest that animals will show strong avoidance to levels of 90 dB _ht and above and slight reactions at around 75 dB _ht.

Any avoidance behaviour or displacement of marine mammals by OREG noise could have cascade effects on reproduction and survival (as shown in Figure 1), again largely mediated through the hypothalamic-pituitary axis and the stress response. The two primary responses that could affect vital rates are changes in the ability of animals to forage or their ability to find mates. In addition once displaced, they may not return to their original habitat, resulting in increases in permanent emigration and effects on meta-population. At what level the different species of marine mammal receptors are likely to move away will depend on their foraging and communication methods, cetaceans may be more sensitive to impacts using echolocation to find prey than phocid seals that rely on other senses, particularly their vibrissae for detecting prey (Dehnhardt et al. 2001).

Alternative terminology for the stress response, within the context of effects of noise on the life cycle of an animal, has been proposed by McEwen & Wingfield (McEwen & Wingfield 2003; McEwen & Wingfield 2010) where various levels of allostatic state (equivalent to the 'health' concept with the PCOD framework) are used to categorise individuals. Allostasis refers to the physiological and behavioural mechanisms used to support the stability of the physiological systems that sustain life. This structure is used in the NRC Report on Marine Mammals and Ocean Noise (Wartzok et al. 2005) to investigate impacts of noise on energy requirements. This translates into a useful framework for modelling an individual's energy balance were an animal's allostatic load results from its need to obtain enough energy for normal activities on a seasonal basis such as moulting, mating and lactating. Animals can adapt to extra demands on this within limits, however if resources in the environment are insufficient, or other challenges such as disease increase the allostatic load the animal may be unable to cope and will develop a physiological dysfunctions that may lead to death.

Figure 1. Potential pathways to effects on vital rates

Studies are also being funded by the Strategic Environment Defense Program ( SERDP) and the Office of Naval Research ( ONR) in the US to determine the effect of noise on behaviour and condition in cetaceans and then how condition relates to reproductive state. One study (entitled "Behavioural Ecology of Cetaceans: The Relationship of Body Condition with Behaviour and Reproductive Status") is being carried out by SMRU and it addresses the issue of how noise (killer whale playbacks) affect condition and ultimately reproductive state in beaked whales and humpback whales.

Thus if the habitat noise is above a tolerance threshold that produces an aversion response animals are likely to move away and may be unable to find good alternative foraging, breeding, moulting or resting areas. This will have knock-on effects on their nutritional state, body condition, immune function and reproductive ability (Barber, Crooks & Fristrup 2010). Determining at what level this is likely to occur and the number of receptors in the populations whose reproductive output or survival probability is reduced is likely to be difficult.

However, it is now recognised that since (particularly when conducting appropriate assessments within the context of impacts on SACs) risk assessments are a requirement for consent, a pragmatic approach to this uncertainty should be taken. It is thus important to determine the most appropriate individual displacement or permanent emigration dose-response relationships to use for the various potentially impacted marine mammal receptors.

One example of a pragmatic approach that has been taken is a recent framework for assessing the effects of pile-driving noise from wind- OREG construction on a harbour seal populations (Thompson et al. in prep) where the effect of PTS on vital rates was modelled. The noise distribution from the piling activities and the seals' behaviour were used to estimate received noise levels. They then used a combination of dose response for TTS and PTS (Nedwell et al. 2007b; Southall et al. 2007) and information on response of harbour porpoise to piling noise (Brandt et al., 2011) in the absence of data for seals, to estimate the total sound exposure, likelihood of hearing damage and level of displacement. They also assumed that individuals experiencing PTS would be subjected to an additional mortality risk and that the consequences of behavioural displacement would be that a proportion of pregnant females would not be able to sustain their pregnancy due to poor foraging, energy imbalance and allostasis overload, resulting in stillbirth, abortion or reduced pre or post weaning survival of pups. Each of these stages had associated uncertainty and this was reflected in the final model used to estimate the potential effect on the population (see Section 5.2).

Noise induced hearing loss, TTS or PTS may have an impact on survival as individuals may be unable to detect and thus avoid predators. Studies suggest that a decrease in hearing sensitivity could increase the risk of predation for harbour seals (Deecke, Slater & Ford 2002) in areas where killer whales are likely seal predators (such as Shetland and Orkney).

5.1.2 Physical contact

Clearly it is important that individual marine mammals are not in direct contact with, for example tidal OREG devices that could result in their death. Furthermore, non-fatal collisions that result in major trauma could subsequently become infected and lead to death through septicaemia. More minor injury may affect immune function where their immune system is activated and up-regulated, a process that can be energetically costly (Lochmiller & Deerenberg 2000). This again may lead to allostatic overload and may temporarily impair the reproductive capacity, depending on age and life history stage of the injured animal.

Collision risk models are being evaluated as part of the MSS/001/11 Research Project (see Section 4.2.2.1.) and further work on collision outcomes are being carried out by SMRU with support from Marine Scotland and SNH. Entanglement may also be an issue for an individual that could result in death directly or indirectly through injury and infection. Continued collaboration with the Scottish Marine Animal Stranding Scheme ( SMASS) will assist in determining the likelihood of these effects being important at either the individual or population level.

5.1.3 Disturbance effects

Disturbance over a prolonged period, either for seals at haulout sites or for all marine mammals during foraging at sea, could result in a chronic stress response. Whilst continual elevated circulating corticosteroid levels are not beneficial to the animal, in the short term they are the appropriate physiological response to a perceived stressor. If the stressor is of short duration and magnitude this is clearly how the animal copes in order to eventually return to a homeostatic state. Disturbance effects may be considered analogous to anti-predator responses. Mammals and birds have evolved to cope with repeated and often intense threat from predation. Occasional or sporadic disturbance may be within the scope of marine mammals' abilities to deal with the existing challenges these populations face in their natural environment. However, in chronic stress such as may be induced by continual disturbance (or indeed perhaps continual low level noise) animals may begin to exhibit adrenal hypertrophy and hyperplasia (Reber et al. 2007). Rats exposed to chronic stress often exhibit adrenal enlargement and increased basal plasma corticosteroid levels, despite normal plasma adrenocorticotropic hormone levels, suggesting that chronic stress also affects the peripheral limb of the HPA axis as ACTH is produced in the pituitary (Ulrich-Lai et al. 2006; Reber et al. 2007). However, additional studies of chronic or prolonged stress (particularly in chronic fatigue syndrome in humans) have found the opposite, that the adrenal gland may become dampened with low levels of cortisol (hypocortisolism) being produced, possibly causes by preceding long periods of high HPA axis activity (Fries et al. 2005). This may have knock-on physiological effects on individual animal health and reproductive capacity. The role of cortisol in the stress response is only one of many downstream physiological actions it has. It is important in immune function (Costa-Pinto & Palermo-Neto 2010) and reproduction (Jensen Pena, Monk & Champagne 2012) and phocid seals have some of the highest circulating levels of cortisol of any mammal possibility related to their diving physiology, although this is only speculation. However, it is clear that decreased cortisol production by the adrenals following chronic stress may have other unforeseen consequences for an individual's physiology and allostatic state.

A MASTS (Marine Alliance for Science and Technology for Scotland) funded PhD study at SMRU starting in October will be addressing some of the energetic issues and costs associated with disturbance in harbour seals. However, this will also provide an opportunity to further our understanding of the HPA axis and its response to disturbance with the confines of the captive situation.

Research gap
Title	Code	Details	Status	Reporting date
Effects of disturbance on the hypothalamic pituitary adrenal axis.	DIS	Repeated disturbance of animals (seals from haulout sites for example) may have impacts on their ability to respond normally to novel stressors (through adrenal fatigue). Before that stage is reached it is important to understand how the HPA responds to disturbance. A PhD study just starting will investigate the energetic costs of disturbance and additional research into the effects on the HPA by monitoring hormone and protein markers in excreta could be included.	Funded by MASTS	2016

5.2 Population

Effects on individuals will be transmitted through to affect the sizes of populations by altering overall fecundity and the survival probabilities of particular groups of animals. Displacement of animals may, temporarily or permanently change the local population that they belong to and the connectedness of different sub-populations.

The primary population scale effects are likely to be changes in abundance, which could lead to, local or widespread, extinctions. Long-term displacement away from areas, or reductions in abundance, could reduce the connections between local populations. In turn, that could reduce gene flow, increase the vulnerability to localised events and make it more difficult for the subsequent re‑colonisation of areas.

The remainder of this section will consider focus on four issues:

• the scaling up from estimates of individual effects to changes in abundance;
• how conservation targets are currently set for marine mammals;
• the balancing of different environmental goals; and
• the specific problem of the implications of working within an area where abundance is already declining.

5.2.1 Extrapolating to population level effects

If the baseline population dynamics of a species are well understood, and the size and distribution of effects of an OREG development on individuals can be estimated (see Section 5.1), then computer simulation should be able to estimate likely consequences for a development. Three approaches to investigating existing or anticipated changes are shown in the box below:

Regression Analysis simply takes the numbers of individuals observed and looks for trends or patterns through time. It is straightforward to do, and does not attempt to investigate detailed processes within the population or changes in its structure. It can therefore work with small datasets, but is limited in what it can discover about causes. This method was used to demonstrate that the reductions in numbers of harbour seals counted in eastern Scotland were not merely due to measurement error but indicated real declines (Lonergan et al. 2007). The impact of simple effects, such as across the board reductions in survival, can be considered by comparing them to estimates of current trends in abundance.

State-space models divide the population into classes, age or stage groups, and model the process of animals moving between them as well as how observations are made. The grey seal population model (Newman et al. 2009b) is an example of this. A similar approach is taken in a Moray Firth harbour seal model ( SCOS 2010). Once a model is fitted to historical data, the implications of particular changes in vital rates can be investigated by simulation. The two limitations on this approach are the amounts of data required to estimate the connections between the various groups of animals and the computational complexity of the model fitting.

Instead of considering classes of animals as behaving in the same way, Individual-based models keep track of each individual separately. If enough information is available, that should allow the models to capture the effects of individual variation. However, for marine mammal populations, sufficient information is seldom available to take advantage of this potential. In practice these, even more than state space models, need to make sweeping generalisations to cover gaps in knowledge.

5.2.1.1 PCOD/ PCAD

The recent development of PCOD (Population Consequences of Disturbance) models or a subset of PCOD referred to as Population Consequences of Acoustic Disturbance ( PCAD) dealing specifically with acoustic disturbance effects is an attempt to use whatever information is available in a formal model framework to estimate population consequences. Such models may incorporate both state-space and individual-based approaches. They also incorporate dose-response functions of potential OREG stressors.

PCOD is a simple population model where assumptions are made about the relationships between observed behavioural responses and vital rates of individuals. Such a modelling framework does not avoid the requirement for information on particular demographic parameters or on the types of dose-response relationships that allow us to link behavioural responses and population processes. At best such a system can be seen as a sensitivity analysis tool to allow us to assess the potential effects of different environmental disturbances. At the very least it provides a method to identify the most critical data gaps. A description of the PCOD methodology, data requirements and methods for dealing with data deficiencies are presented in Lusseau et al (2012). Clearly for any such model the data requirements will be determined by the particular species of interest and the specific device/interaction that is under consideration. Lusseau et al. (2012) describe a PCOD framework for modelling the population consequences of disturbance developed by the US Office of Naval Research working group on PCAD (Anon 2012). This combines all aspects of the internal state of an individual that might affect its fitness into a single term denoted as "Health". This is a catchall for such factors as energetic state or resistance to disease, etc. The effects of the health of the animal on its vital rates are then calculated, where "Vital rates" refers to all the components of individual fitness (probability of survival, fecundity, growth rate etc. as well as secondary factors such as offspring survival).

The data requirements for such studies are extensive. As a minimum, there needs to be some form of estimate of the initial structure of the population of interest, its vital rates and information on the important components of the "health" state of the population. In addition, for any particular disturbance factor, the model requires a set of parameterised dose-response models that ultimately predict changes in specific population vital rates. Such models have a large number of parameters and thus have significant data requirements to ensure that prediction uncertainty is kept within 'managerially useful' bounds. A sensitivity analysis of such models will indicate the most pressing data requirements.

Harwood & King (2012) describe how the Interim and Full PCOD frameworks will require estimates of ten critical sets of information.

1. The sound field produced during construction and operation of a particular offshore renewables development (with associated uncertainty).

2. The sound levels that are likely to cause Temporary Threshold Shift ( TTS). This information should, preferably, be in the form of a dose-response relationship, with associated uncertainty, for each priority species.

3. The sound levels that are likely to cause Permanent Threshold Shift ( PTS), preferably in the form of a dose-response relationship, with associated uncertainty, for each priority species.

4. The sound levels that are likely to result in a "significant" behavioural response (defined as one that is likely to impair an individual's ability to survive, breed, reproduce, or raise young, or that is likely to result in that individual being displaced from an area for a longer period than normal). These should, preferably, be in the form of a dose-response relationship, with associated uncertainty, for each priority species.

5. The number of animals likely to be exposed to sound levels that could result in PTS, TTS or a "significant" behavioural response during one day of construction or operation of an offshore wind farm. The number of animals of each species (with associated uncertainty) that are likely to collide with or become entrapped in a marine renewables device or be exposed to sound levels that could result in a "significant" behavioural response during one day of construction or operation of this device.

6. The number of animals (with associated uncertainty) that are likely to be exposed to sound levels likely to result in PTS, TTS or a "significant" behavioural response over the entire course of construction of an offshore wind farm.

7. The potential effect of experiencing PTS at a specified frequency on the vital rates (probability of survival for one year, probability of giving birth) for an individual of each species, by age/stage class ( e.g. adult males, adult females, calves, juveniles), with associated uncertainty;

8. A mathematical function linking the number of days on which an individual experiences TTS or a "significant" behavioural response and its vital rates (probability of survival for one year, probability of giving birth), with associated uncertainty, for the different age classes of each priority species.

9. The current population size and population history for each Management Unit ( MU) of the five priority species, with associated uncertainty.

10. The key demographic parameters (adult survival, calf survival, juvenile survival, annual probability of pupping/calving, age at first pupping/calving, longevity) for each species, in each MU (if parameters are likely to vary between MUs) with an indication of likely levels of variation between years.

Where there are insufficient data (or even no data) to adequately parameterise these models Lusseau et al. (2012) describe a process of incorporating expert advice. This expert panel approach (Marcot et al. 2012) has the attraction of proffering provisional scientific advice on time-critical issues. However considerable caution must be used in panel composition, questionnaire construction and subsequent analysis to avoid significant output bias.

A preliminary expert opinion elicitation exercise is being carried out as part of the preliminary work for the ORJIP process. The results of this exercise will be used to define the important remaining information requirements for developing a series of PCOD models. Developing and implementing PCOD models has been identified as a major research requirement under the ORJIP programme and forms Project 2: Evidence gathering for a Population Consequences of Acoustic Disturbance ( PCAD) model to predict impacts to marine mammals from underwater noise.

This section of the current review will be further developed when the results of this exercise are available. Inevitably this will lead to an increase in and changes to the

5.2.2 Marine mammal conservation targets

The environmental monitoring and mitigation requirements imposed on developers are designed in part to reduce the impact of marine renewables on species of conservation concern. Conservation legislation is usually written in terms of population or stock management, usually with specific management targets such as maintaining favourable conservation status. It is becoming increasingly clear that regulators need to consider the population consequences of any management decisions in order to fulfil their requirements under national and international legislation.

Where lethal takes are involved, it is a "relatively" simple task to assess the likely impacts on a population. However in a case such as the development of a new marine renewables industry where novel and as yet untested technologies are being deployed in the real world such relatively easy methods are not suitable. We are not yet sure of the extent or intensity of any effects, we are not even certain that there will be any directly injurious interactions.

The EU Habitats Directive, now supplemented by the Marine Strategy Framework Directive ( MSFD), sets the requirements for conservation management of marine mammals in European waters. These provide very little guidance beyond a requirement that populations be "favourable" in size, range and prospects. The only fixed reference point they give are that things must be no worse than when the Habitats Directive came into force; 1994 in the UK. The interpretation of the directives has largely been devolved to EU member states, though with a threat of facing legal proceedings if their efforts are not considered adequate (Lonergan 2011; Lonergan 2012).

Internationally, the main method used to set allowable "takes" of marine mammals is the Potential Biological Removal ( PBR). This was developed within the US Marine Mammal Protection Act and allows a small proportion of any population to be removed (Wade 1998). While it is often presented as an having objectively determined target population size, use of PBR will tend to drive a population towards a proportion of its carrying capacity that is implicitly determined by the details of how density dependence affects individuals' chances of survival and reproduction. This approach is currently used by the Scottish Government in determining applications for licences to shoot seals around fish farms and set nets. It is also used to determine whether the number of animals predicted to suffer fatal interactions with renewables (or other industry) is acceptable. The International Whaling Commission ( IWC) uses a more detailed calculation to set notional catch limits for large cetaceans, though it has not actually agreed to any commercial hunting (Lonergan 2011).

The setting of conservation goals is clearly a societal and political decision. While that decision lies outside science, it does depend on knowledge of the implications and balances of risks and benefits of potential courses of action. Choosing appropriate targets therefore needs to be an iterative process with scientists being able to provide progressively more detailed information as the range of potentially acceptable options is narrowed.

Research gap
Title	Code	Details	Status	Reporting date
UK Marine mammal conservation targets.	MMCT	Consideration of the cumulative effects of multiple impacts on neighbouring and connected populations is complicated. The IWC has a well-developed simulation framework for examining this. Work is currently being funded by JNCC to look at how similar methods could be applied to sub-populations of seals and small cetaceans within European waters.	Funded by JNCC, On-going	Sept 2013