Scottish Marine and Freshwater Science Volume 5 Number 2: A Protocol for Implementing the Interim Population Consequences of Disturbance (PCoD) Approach...

Appendix 1: Expert Elicitation Process and Results

Summary

We approached 150 leading experts in the field of marine mammal science. These experts were chosen because they actively work in the field of marine mammal population biology, on the impacts of noise on marine mammal hearing, or on the effects of disturbance on marine mammals. The criteria for the initial selection of experts were:

• Published on the population biology of one or more of the five priority species (or a closely related species) over the last 5 years
• Published on the impacts of noise on the hearing of one or more of the five priority species (or a closely related species) over the last 5 years
• Published on the effects of disturbance on one or more of the five priority species (or a closely related species) over the last 5 years

We also approached members of advisory groups that are involved with the conservation of any of the priority species, or closely related species ( e.g. IUCN Seal Specialist Group, IUCN Cetacean Specialist Group, Society of Marine Mammalogists Conservation Committee, relevant ICES Working Group, and the UK Special Committee on Seals).

Forty-one experts (27%) completed the PCOD questionnaire, some for multiple species. The total number of questionnaires completed is shown in Figure A1.1.

Figure A1.1. The number of questionnaires completed for each species (inclusive of all age classes) in the first round of expert elicitation for the interim PCoD framework.

Analytical Approach - Non-Technical Summary

Expert elicitation seeks to accumulate opinions from experts about parameters for which there are currently few or no data. In this process each expert provides estimates of, and some indications of uncertainty surrounding, the parameters of interest.

In order to generate the most accurate and robust results, we used a 4-step method, based on the approach developed by Spiers-Bridge et al. (2010) for each question in our questionnaire (as recommended by Burgman et al. 2011). Using this established method, experts are asked for their estimates of the lowest and highest realistic value for the parameter in question, their best estimate of the parameter and their level of confidence that the interval they provided contains the true value.

Each expert was solicited for best estimates that reflected his or her understanding of two sets of parameters:

• the potential effect of hearing damage (a permanent shift in the threshold for hearing - PTS - in a specified frequency range) on survival and, for mature females, on the probability of giving birth and;
• three parameters that determined the relationship between the number of days of disturbance an individual might experience in a year and its survival and, for mature females, the probability of giving birth. Figure A1.2, below, shows how these parameters determine the shape of the relationship between disturbance and survival or fertility. Parameter A defines the maximum effect of disturbance on survival (we assumed that the maximum effect of disturbance would be to reduce the probability of giving birth to zero), parameter B defines the amount of disturbance an individual can tolerate before it has any effect on survival or fertility, and C defines how many days of disturbance are required to have the maximum effect on survival or fertility.

Figure A1.2. Hypothetical relationship between days of disturbance and the probability of survival.

Experts were also prompted to provide information on the confidence/uncertainty associated with their estimates. The uncertainty in the expert's estimates were provided by a) a range that 'bounds' the estimate and b) a level of confidence about their estimate. Statistical models that were consistent with each expert's best estimate and associated uncertainty were generated using some of the approaches described in Albert et al. (2010), Garthwaite et al. (2005), and Genest & Zidek (1986). For example, their best guess and bounds could define a triangular Probability Density Function ( PDF), or lead to a particular Beta distribution. In this way, each expert's opinion was encapsulated in an individualised statistical distribution. All of the experts' distributions were then combined into an overall PDF, which was used to provide a representative sample of parameter values or as a set of random draws, as required. As discussed in Appendix 2, for each iteration of the model, the software selects a set of parameter values at random from these statistical distributions derived from the results of the expert elicitation process. This is equivalent to soliciting the opinions of one 'virtual' expert for each iteration and at least 500 random draws were conducted for each development scenario.

Analytical Approach - Technical Summary

Two main outputs are required here: univariate distributions for single parameters ( e.g. the effect of PTS at 1-2 kHz on the probability of survival), and bi- or trivariate distributions that characterise curves ( e.g. the relationship between the number of days of disturbance and the probability of giving birth). These will be used in a Monte-Carlo fashion to provide inputs to the PCOD protocol ( i.e. random draws will be required from these distributions).

For basic univariate treatments the experts' estimates were used to fit Gamma, Beta, truncated Normal, Uniform or Triangular PDFs. For multivariate treatments, copula-based simulation methods were used (Iman & Conover, 1982). These allow a multivariate distribution of correlated variables to be defined using arbitrary marginal distributions and a separate correlation structure. The general process in these two cases is described below.

In the univariate case, where the parameter has a 0-1 bounded domain:

• The expert's estimates were used to define Beta or Triangular PDFs.
• The individual distributions were combined as a weighted sum, subsequently normalised, to give a collected PDF for the parameter.

In the bi- and trivariate cases, where the parameters are a mix of 0-1 bounded and non-negative domains:

• The experts' estimates were used to define Beta or Triangular distributions for the 0-1 bounded parameters.
• The experts' estimates were used to define Gamma, Triangular or truncated Normal distributions for the 0- infinity bounded parameters.
• Correlation matrices were calculated for the sets of parameters.
• Collectively these distributions and matrices define the margins and correlation structure for the parameters, allowing copula methods to be applied. Correlation was provided by a truncated multivariate Normal distribution.
• Random draws from each expert's multivariate distribution, which is a mix of estimated marginal distributions and governing correlation structure, could then be made.
• Experts' multivariate distributions were sampled intensively, in proportion with their confidence, to build an overarching two- or three-dimensional PDF.

Results

Harbour Seals - Mature Females

1. What do you think would be the effect of experiencing PTS at around 1-2 kHz on the survival of a mature female?

Most experts felt that missing an octave wide band (1-2 kHz) in an animal's hearing would affect survival by less than 10%, but a few felt that it might have a larger effect (around 50%).

2. Do you think experiencing PTS at around 2-10 kHz will affect the survival of a mature female?

Most experts felt that missing a 2.5 octave wide band (2-10 kHz) in an animal's hearing would affect survival by less than 20%, but a few felt that it might have a larger effect (~50%).

3. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the probability that an adult female will give birth in the future?

Some experts suggested that PTS may reduce a female's ability to hear males calling during the breeding season and therefore there will be a reduced opportunity to find mates, leading to missed breeding opportunities. In addition, foraging efficiency might be reduced, leading to reduced body condition which may impact on fecundity and fertility. Expert opinion leant towards a change of less than 30% in fertility, but a wide range of values was suggested.

4. Disturbance curves: Are you able to indicate the minimum and maximum number of days of disturbance that a mature female can tolerate before it has any effect on the probability of giving birth and what do you think is the minimum and maximum number of days of disturbance that would be necessary to have the maximum effect, in which the probability of giving birth is reduced to zero?

Experts felt that disturbance which lasted for more than 20 days may result in reduced foraging efficiency and lead to reduced body condition and fertility. Some experts felt a female may not be able to encounter males during oestrus if she was displaced to unfamiliar locations as a result of chronic noise (disturbance) resulting in missed mating opportunities. Most experts believe that a substantial amount of disturbance (150-280 days) was necessary to reduce the probability of giving birth down to 0.5, but some thought that this could occur at relatively low levels of disturbance (<50 days).

Harbour Seals - Juveniles

1. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the survival of a juvenile?

Most experts felt that missing an octave wide band at lower frequencies (1-2 kHz) or experiencing PTS in slightly higher frequencies (2-10 kHz) would not influence mortality, but some thought that animals would have a reduced ability to detect predators or other hazards, or reduced foraging efficiency leading to a decrease in survival of more than 20 %.

2. Disturbance curves: Are you able to indicate (i) the minimum and maximum number of days of disturbance that a juvenile can tolerate before survival is affected (ii) the lowest and highest plausible values for the maximum effect of disturbance on calf survival and (iii) the minimum and maximum number of days of disturbance that would cause this maximum effect.

Experts felt disturbance would affect juvenile survival due to the disruption of foraging (reduced time in prime foraging habitat) and problems associated with displacement to unfamiliar areas which would affect feeding behaviours and resting at known haul-out sites. One expert thought if a juvenile animal was disturbed, as well as experiencing a reduced ability to forage, the animal would be less able to detect predators. Most experts thought that about 50 days of disturbance would be required to have any effect on survival, and that the maximum effect of disturbance on survival would be to reduction to around 0.6. However, some experts thought that relatively small levels of disturbance (<50 days) could result in certain death.

Harbour Seals - Pups

1. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the survival of a pup?

Experts felt that PTS may cause separation between mother and pup or interfere with mother-pup vocal exchanges, resulting in starvation if separation was prolonged. Pups may also be unable to detect hazards in air (predators) and water (vessels, turbines) or other anthropogenic impacts. Most experts thought that the reduction in survival would be less than 30%.

2. Disturbance curves: Are you able to indicate (i) the minimum and maximum number of days of disturbance that a mother-calf pair can tolerate before pup survival is affected (ii) the lowest and highest plausible values for the maximum effect of disturbance on calf survival and (iii) the minimum and maximum number of days of disturbance that would cause this maximum effect.

Experts felt that disturbance would affect pup survival through reduced lactation opportunities if adult females spend more time alert and moving as a consequence of being disturbed. Experts also thought pup survival would be affected if mother and pup became separated when the mother moved to a new location as a result of disturbance and there would be interference with the mother-pup bond through vocal exchanges. Most experts thought that pups could tolerate around 20 days of disturbance before it had some effect on survival, but some believed this much disturbance could result in certain death. Others believed that much longer periods of disturbance were required to have the maximum effect on survival, and that the maximum reduction in survival ranged between 5 and 70%.

Grey Seals - Mature Females

1. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the survival of a mature female?

Experts thought PTS may result in reduced forging efficiency through reduced ability to detect prey using passive acoustics ( i.e. active listening), and thus poor condition in grey seal adult females. It may also lead to a reduced ability to detect predators and anthropogenic threats. Predicted effects on survival varied widely.

2. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the probability that an adult female will give birth in the future?

There were only a small number of responses for this question. Most experts thought that PTS may affect foraging efficiency leading to poor body condition and could result in a reduction in fertility of around 40%.

3. Disturbance curves: Are you able to indicate the minimum and maximum number of days of disturbance that a mature female can tolerate before it has any effect on the probability of giving birth and what do you think is the minimum and maximum number of days of disturbance that would be necessary to have the maximum effect where the probability of giving birth is reduced to 0?

Experts felt disturbance may result in reduced foraging efficiency which could affect fertility and interfere with mating opportunities due to habitat displacement. One expert said 'changes in the probability of giving birth as a result of disturbance may occur due to reduced energy intake, resulting in reduced allocation of energy to foetus, and subsequently early abortion due to reduced energy intake'. There was broad agreement that animals could tolerate a small number of days of disturbance before it had any effect on fertility. However, some experts believed that 50 days of disturbance would reduce the probability of giving birth by 50%, whereas others thought that around 100 days would only reduce the probability of giving birth to 0.7 ( i.e. a reduction of 0.3) .

Grey Seals - Juveniles

1. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the survival of a juvenile?

There were only a small number of responses for this species . Experts thought PTS may result in inability to detect and avoid hazards in the environment, and a reduction in ability to detect prey using passive acoustics. This could result in a reduction in survival of around 50%.

2. Disturbance curves: Are you able to indicate (i) the minimum and maximum number of days of disturbance that a juvenile can tolerate before survival is affected (ii) the lowest and highest plausible values for the maximum effect of disturbance on calf survival and (iii) the minimum and maximum number of days of disturbance that would cause this maximum effect.

Experts felt that between 10 and 100 days of disturbance would affect juvenile survival due to reduced foraging efficiency, leading to reduced body condition and even starvation. One expert said 'if animals were disturbed by a noise source and moved away from an area of residence then it would be exposure to predators that would likely increase mortality rather than any direct effect of noise'. Another expert felt that disturbance would result in a juvenile animal's inability to hear and avoid vessels and other potentially traumatic devices in the environment which would affect survival. There was considerable uncertainty about the maximum effect of disturbance on survival and the number of days of disturbance necessary to cause this, although there was most support for a maximum reduction of 50-60% in the probability of survival.

Harbour Porpoise - Mature Females

Note: Although some experts said that 1-2 kHz / 2-10kHz is at the lower end of the effective hearing range of porpoises, and well below the frequency range they use for echolocation, other experts pointed out that noise that was sufficiently loud to cause a TTS of 40dB (the criterion for PTS), was also likely to affect hearing at higher frequencies to some degree. Impairment at these higher frequencies would have effects on the likelihood of survival in harbour porpoises.

1. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the survival of a mature female?

Experts thought that PTS may reduce an animal's ability to detect predators, prey and hazards in the environment (particularly fishing gear), and affect communication. Most experts believed the effect on survival would be less than 20%, but some thought it could be as high as 40%.

2. Do you think experiencing PTS at around 1-2 kHz will affect the probability that an adult female will give birth in the future?

Experts thought PTS could result in reduced ability to communicate, reduced foraging efficiency and therefore reduced body condition, although, in general, the effects on fertility were likely to be less than 20%. Results for 2-10 kHz were similar to those for PTS at 1-2 kHz, although there was some support for larger effects on fertility (up to 40% or more).

3. Disturbance curves: Are you able to indicate the minimum and maximum number of days of disturbance that a mature female can tolerate before it has any effect on the probability of giving birth and what do you think is the minimum and maximum number of days of disturbance that would be necessary to have the maximum effect where the probability of giving birth is reduced to zero?

Most experts felt that disturbance lasting more than 50-100 days may result in reduced foraging efficiency which could affect fertility, or induce pregnancy failure, and interfere with mating opportunities due to habitat displacement. Experts also highlighted that 'elevated stress levels as a result of being displaced from a known location may impact fecundity'. The maximum effect on the probability of giving birth was thought to be a 50% reduction.

Harbour Porpoise - Juveniles

1. Do you think experiencing PTS at around 1-2 kHz will affect the survival of a juvenile?

Experts thought that PTS at 1-2 kHz or 2-10 kHz in juvenile harbour porpoise might reduce their ability to detect predators, their foraging efficiency and their ability to avoid anthropogenic hazards. Most predicted that the effect on survival would be less than 20%, but some felt this could be as high as 40% for 1-2 kHz and 60% for 2-10 kHz.

2. Disturbance curves: Are you able to indicate (i) the minimum and maximum number of days of disturbance that a juvenile can tolerate before survival is affected (ii) the lowest and highest plausible values for the maximum effect of disturbance on calf survival and (iii) the minimum and maximum number of days of disturbance that would cause this maximum effect.

Most experts felt disturbance lasting 20-100 days would affect juvenile survival due to reduced foraging efficiency, leading to reduced body condition and even starvation. Others felt 100-200 days disturbance would be required. Inexperienced animals were thought to be more likely to be disturbed and therefore experience a reduction in foraging success. Juvenile animals were thought to have very limited energy stores and therefore more susceptible to starvation. If animals were displaced to an unfamiliar area they have difficulties in finding "shelter" as well as food. Most believed that continuing disturbance could reduce survival by a maximum of 60%, but one expert believed that any disturbance lasting more than 20 days would result in certain death.

Harbour Porpoise - Calves

1. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the survival of a calf?

Experts thought PTS at 1-2 kHz may affect the ability of mothers and calves to stay in contact, which could have consequences for survival, and that the calf may be unable to detect predators or anthropogenic hazards. Most believed that the effects on survival would be less than10%, but some thought that they could be as high as 50%. Results for 2-10 kHz were similar to those for PTS at 1-2 kHz. However, one expert noted that 'If PTS is assumed to begin at 40 dB of TTS the hearing of the calf will also be impaired at higher frequencies, which in turn will increase the chance of losing contact to its mother (especially after disturbance).

2. Disturbance curves: Are you able to indicate (i) the minimum and maximum number of days of disturbance that a mother-calf pair can tolerate before calf survival is affected (ii) the lowest and highest plausible values for the maximum effect of disturbance on calf survival and (iii) the minimum and maximum number of days of disturbance that would cause this maximum effect.

Experts felt that 10-50 days of disturbance would affect calf survival if mothers and their calves were separated as a result of the mothers' response to disturbance, which would affect amount of milk transferred to the calf. This reduced food intake for the calf, coupled with an increased activity budget (following their mother's response to sound) could affect calf survival. Opinions were divided, however, on the maximum effects of disturbance on survival. Some believed that this was likely to be a reduction in survival of 80-100% and was likely to occur after 10-40 days of disturbance, whereas others believed that the maximum effect was likely to be a reduction in survival of 5-50%.

Bottlenose Dolphins - Mature Females

1. Do you think experiencing PTS at around 1-2 kHz will affect the survival of a mature female?

Some experts thought PTS at 1-2 kHz for bottlenose dolphins may result in a reduction in an animal's ability to obtain prey (reduction in foraging efficiency) which would lead to poor body condition, but the predicted effects were less than 5%. Experts thought that PTS at 2-10 kHz for bottlenose dolphins may result in a reduction in an animal's ability to obtain prey (reduction in foraging efficiency) which would lead to poor body condition, in an inability to hear conspecific whistles which would disrupt social behaviour and reduce the distance over which social cohesion is maintained, but most thought the effect of this on survival would be small.

2. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the probability that an adult female will give birth in the future?

Experts thought PTS at 1-2 kHz would affect an animal's ability to capture prey and thus reduce foraging efficiency, reduce body condition and impact on fecundity such as pregnancy failure. Most thought that the effect on fertility would be less than 10%, and one expert thought there would be no long term impact from small PTS at these frequencies. Experts thought PTS at 2-10 kHz could affect an animal's ability to capture prey and thus reduce foraging efficiency, reduce body condition and have an impact on fertility, but the effect would be less than 20%. One expert thought that 'Given a small PTS, it would have little consequence in the long term. The TTS preceding it, which could last for days/weeks, would be the more severe issue if it affected reproductive effort by impeding social communication'.

3. Disturbance curves: Are you able to indicate the minimum and maximum number of days of disturbance that a mature female can tolerate before it has any effect on the probability of giving birth and what do you think is the minimum and maximum number of days of disturbance that would be necessary to have the maximum effect where the probability of giving birth is reduced to zero?

Experts felt disturbance could result in reduced foraging efficiency and displacement from critical foraging area that would place a severe strain on a female's energy budget; this might affect fertility or result in pregnancy failure. Experts also highlighted that 'elevated stress levels as a result of being displaced from a known location may impact fecundity'. There was wide variation in the number of days of disturbance that experts believed a female could tolerate before it would have any effect on fertility and in the number of days of disturbance required to reduce the probability of giving birth down to 0.5.

Bottlenose Dolphins - Juveniles

1. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the survival of a juvenile?

Experts thought that PTS at 1-2 kHz for bottlenose dolphin juveniles may result in a reduction in the animal's ability to obtain prey and thus a reduction in foraging efficiency. One expert was not sure, but felt that some level of precaution should be applied, whereas another thought a small PTS at these frequencies would not affect survival. Experts thought that PTS at 2-10 kHz for bottlenose dolphin juveniles may result in a reduction in the animal's ability to obtain prey (through its effect on sensory biology - echolocation) and thus a reduction in foraging efficiency. There were potential social effects if it reduced the animal's ability to detect conspecific whistles. Most thought that the effect would be less than 10%, and one expert thought the effect would be negligible.

2. Disturbance curves: Are you able to indicate (i) the minimum and maximum number of days of disturbance that a juvenile can tolerate before survival is affected (ii) the lowest and highest plausible values for the maximum effect of disturbance on calf survival and (iii) the minimum and maximum number of days of disturbance that would cause this maximum effect.

Experts felt that disturbance in excess of 50 days could affect juvenile survival due to reduced foraging efficiency and increased stress levels, leading to reduced body condition. One expert said disturbance may disrupt a juvenile learning foraging behaviours and will disrupt social interactions. Most experts believed that the maximum effect of disturbance would be to reduce survival by less than 10%.

Bottlenose Dolphins - Calves

1. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the survival of a calf?

Experts thought that PTS at 1-2 kHz for bottlenose dolphin calves may result in an increased risk of losing contact with their mother, which could lead to starvation. One expert thought a small PTS at these frequencies would not affect survival. Experts thought that PTS at 2-10 kHz for bottlenose dolphin calves may result in an increased risk of losing contact with their mother which could lead to starvation. One expert thought a small PTS at these frequencies would not affect survival but that the TTS that precedes it is more likely a factor, depending on the magnitude and duration of the shift.

2. Disturbance curves: Are you able to indicate (i) the minimum and maximum number of days of disturbance that a mother-calf pair can tolerate before calf survival is affected (ii) the lowest and highest plausible values for the maximum effect of disturbance on calf survival and (iii) the minimum and maximum number of days of disturbance that would cause this maximum effect.

Experts felt that disturbance could affect calf survival if it exceeded 30-50 days, because it could result in mothers becoming separated from their calves and this could affect the amount of milk transferred from the mother to her calf. One expert said 'in instances where masking or a threshold shift may also occur, the reuniting of a separated mother-calf pair may be impeded'. Opinions were divided on the maximum effect of disturbance on survival: some suggested that this was likely to be around 80%, whereas others thought it would be around 30%. One expert thought that < 10 days of disturbance would result in the probability of survival dropping to 10%.

Minke Whales - Mature Females

1. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the survival of a mature female?

Experts thought that PTS may affect minke whale mature female survival if it impairs the animal's ability to detect acoustic indications of predators and prey fields for foraging. This would lead to poor condition and ultimately affect survival. Predicted effects varied from a 90% reduction in survival probability to no effect, with most support for a reduction of less than 20%.

2. Do you think experiencing PTS at around 1-2 kHz or 2-10 kHz will affect the probability that an adult female will give birth in the future?

Experts thought PTS at 1-2 kHz may affect fertility in mature female minke whales if it impacts on mating opportunities or if reduced foraging means reduced body condition which affected pregnancy, specifically 'Disturbance of feeding activity will reduce the body condition of the female which will lower the amount of energy the female can invest in the foetus. If her body condition becomes very poor the only way for the female to [compensate] is to reduce investment in the foetus. Predicted effects varied widely between a 5% and a 90% reduction in probability of giving birth, although there was most support for a reduction of less than 40%. Opinions about the effect of PTS at 2-10 kHz were more consistent than for 1-2 kHz, with most experts suggesting a reduction in the probability of giving birth of less than 20%.

3. Disturbance curves: Are you able to indicate the minimum and maximum number of days of disturbance that a mature female can tolerate before it has any effect on the probability of giving birth and what do you think is the minimum and maximum number of days of disturbance that would be necessary to have the maximum effect where the probability of giving birth is reduced to zero?

Experts felt disturbance may result in reduced feeding and an increase in energetic costs of movement and therefore a reduction in body condition and elevated stress levels, if disturbance exceeded for 10-100 days it could affect fertility. One expert suggested that 'reduced body condition will have a negative effect on foetus growth which will reduce the size of the calf at birth and consequently calf survival. If the reduction in energy investment in foetus is large enough then foetus would be aborted'.

Discussion

The results from this round of expert elicitation have provided us with a preliminary set of parameter values for the relationships between injury (onset of PTS), disturbance and vital rates for the five priority species that we can use for a preliminary implementation of the PCOD protocol. However, it is clear from the expert elicitation literature and from face-to-face discussions with Professor Mark Burgman (Director of the Australian Centre of Excellence for Risk Analysis, University of Melbourne) that we would obtain more robust and reliable estimates of these parameters by adopting the Delphi process (as used, for example, by MacMillan & Marshall (2006) in a study of the environmental requirements of capercaillie in Scotland). We therefore consulted the same panel of experts again, showed them the results from the first round of the elicitation and asked them if they would like to modify their original responses. Any modified responses were then used in the final analysis, instead of the experts' original responses.

General Comments On The Pcod Questionnaire

Here we provide a short summary of the points and concerns raised by experts as part of elicitation process. A number of experts provided feedback on what they felt was either missing from the questionnaire, or was not properly assessed.

The most common concern raised was the treatment of the degree of PTS onset, or lack thereof. Experts rightly pointed out that PTS is not synonymous with profound or total hearing loss and that a PTS of less than 3 dB would probably be imperceptible to an animal as loss, whereas a PTS of 60 dB could have significant impacts. Experts felt the more relevant effect is the magnitude and duration of the preceding TTS. We certainly agree with this point in the general sense and we have already acknowledged that the impact will differ depending on the degree of threshold shift in the PCoD background information that was provided alongside the questionnaire. For the purpose of the questionnaire, however, it was decided that it would be too difficult to provide a range of threshold shifts or, given the uncertainty surrounding PTS onset, try and force an agreement on what level of threshold shift would cause a significant impact or not. In the context of environmental impact assessments and currently applied noise exposure criteria, we wanted to know the potential effects on an animal exposed to an SEL that is 20dB or more above the threshold for TTS onset ( i.e. the level recommended by Finneran & Jenkins (2012) for calculating the likely onset of PTS); we fully admit that currently we have no idea what kind of shift in hearing threshold this will actually induce and this does highlight that further research in this area is needed.

Another point raised was that the absence of frequency bands for PTS made it difficult to provide a likely impact. For example, some experts suggested that impacts for harbour porpoise are likely to be far more severe for higher frequency sources and that there were unlikely to be many acoustic signals of interest in the 1-10 kHz region. Although some experts said that 1-2 kHz / 2-10kHz is at the lower end of the effective hearing range of porpoises, and well below the frequency range they use for echolocation, other experts pointed out that noise that was sufficiently loud to cause a TTS of 40dB (the criterion for PTS), was also likely to affect hearing at higher frequencies to some degree. Impairment at these higher frequencies would have effects on the likelihood of survival of some species.

A few experts also noted their dissatisfaction with the definition of disturbance and felt exposure time needed to be better defined for them to be able to judge the effects on the animals. It is correct, for example, that a 1-hour exposure would have a different effect than a 300 day exposure. We have, however, discussed the difficulties in presenting a number of different scenarios and we treated disturbance in number of days, rather than number of hours, as this was thought to be a time period that was salient to the animal in terms of measuring population level effects over a number of years.

One expert felt that the probability of giving birth was not a reasonable metric for assessing the effect of PTS on an animal's vital rates. They felt that adult females with hearing difficulties may still reproduce but that it would be less likely that they would raise the calf to maturity. We feel our approach still accounts for this, whether the calf/pup is born and dies or is not born at all, we still see a reduction in surviving calves/pups that are recruited into the population.

Some experts also raised the point that disturbance and PTS may lead to elevated stress levels which may also incur a fitness cost and have population level impacts. We certainly agree with this and the effect of stress on individuals and populations does warrant further investigation. However, given that estimates of stress are not provided in developers ES chapters it was out with the scope of this approach, at least for now.

Finally, one expert was concerned that the lack of information on parameters asked in the questionnaire made the results of the expert elicitation very questionable. They felt there was a risk of giving regulators the wrong impression of certainty and they hoped that this uncertainty would be reflected in the outcome of the study. We have indeed discussed the different measures of uncertainty inherent in this interim approach in the full report and have highlighted the need for further research.

List of the Experts who Participated in the PCoD Questionnaire

Expert	Affiliation
Ailsa Hall	SMRU, University of St Andrews
Ben Wilson	SAMS
Carol Sparling	SMRU Marine Ltd, University of St Andrews
Clive McMahon	Charles Darwin University, Australia
Cormac Booth	SMRU Marine Ltd, University of St Andrews
Darlene Ketten	WHOI, USA
Dave Thompson	SMRU, University of St Andrews
David Lusseau	University of Aberdeen
Dorian Hauser	National Marine Mammal Foundation, US
Elizabeth Slooten	University of Otago, New Zealand
Enrico Pirotta	University of Aberdeen
Fredrick Christiansen	University of Aberdeen
Frances Gulland	Marine Mammal Centre, Sausalito, US
Garry Stenson	Fisheries and Oceans Canada
Gordon Hastie	SMRU, University of St Andrews
Isabelle Charrier	CNRS, France
Jacob Nabe Nielsen	Arhus University, Denmark
Jacob Tougaard	Arhus University, Denmark
John Harwood	SMRU Marine Ltd, University of St Andrews
John Terhune	University of New Brunswick, Canada
Jonas Teilmann	Arhus University, Denmark
Klaus Lucke	IMARES, Holland
Leslie New	Marine Mammal Commission, US
Lisa Schwarz	University of California, Santa Cruz, US
Luke Rendell	SMRU, University of St Andrews
Marla Holt	NWFSC, US
Mike Hammill	Fisheries and Oceans Canada
Mike Weise	Office Naval Research, US
Patrick Miller	SMRU, University of St Andrews
Paul Nachtigall	University of Hawaii
Paul Thompson	University of Aberdeen
Petter Kvadsheim	Norwegian Defense Research Establishment ( FFI)
Philip Hamilton	New England Aquarium, US
Rob Harcourt	University of Macquarie, Australia
Rob Schick	CREEM, University of St Andrews
Rob Williams	Oceans Initiative
Sophie Smout	SMRU, University of St Andrews
Stephanie Watwood	NUWC Newport, US
Tim Ragen	Marine Mammal Commission, US
Ulf Lindstrøm	Institute of Marine Research, Norway
Veronique Lesage	Fisheries and Oceans Canada

Appendix 1 - Literature Cited

Albert, I., Donnet, S., Guihenneuc-Jouyaux, C., Low-Choy, S., Mengersen, K and Rousseau, J. (2012).Combining Expert Opinions in Prior Elicitation. Bayesian Anal. 7(3), 503-532.

Burgman, M. A., McBride, M., Ashton, R., Speirs-Bridge, A., Flander, L., Wintle, B., Fidler, F., et al. (2011).Expert status and performance. PloS one, 6(7), e22998.doi:10.1371/journal.pone.0022998

Finneran, J. J., & Jenkins, A. K. (2012). Criteria and Thresholds for U . S . Navy Acoustic and Explosive Effects Analysis. SPAWAR Marine Mammal Program.

Garthwaite, P.H., Kadane, J.B. and O'Hagan, A. (2005).Statistical Methods for Eliciting Probability Distributions. Journal of the American Statistical Association. 100(470), 680-701. DOI:10.1198/016214505000000105

Genest, C. and Zidek, J.V. (1986).Combining Probability Distributions: A Critique and an Annotated Bibliography.Statist. 1(1), 114-135.

Iman, R. L. & Conover, W. J. (1982).A distribution-free approach to inducing rank correlation among input variables. Communications in Statistics - Simulation and Computation.11(3).

MacMillan, D. C., & Marshall, K. (2006). The Delphi process - an expert-based approach to ecological modelling in data-poor environments. Animal Conservation, 9(1), 11-19.

Speirs-Bridge, A., Fidler, F., McBride, M., Flander, L., Cumming, G., & Burgman, M. (2010). Reducing overconfidence in the interval judgments of experts. Risk Analysis, 30(3), 512-23. doi:10.1111/j.1539-6924.2009.01337.x