Grey Seal Pup Production

Methods

Indicator metric and data collection

Grey seal (Halichoerus grypus) pups are counted at major breeding sites (‘colonies’) usually using aerial photography survey methods (Figure a); where these are not possible, ground counts or boat-based counts may be used instead (e.g. Stringell et al., 2014). Multiple counts per colony are generated, spread across the breeding season (September to February, depending on colony location).

Most Contracting Parties count pups several times during the pupping season and the maximum count is taken as a minimum estimate of ‘pup production’, that is, the total number of pups produced over the season as a whole. The maximum pup count will always under-estimate pup production because pups are born and leave the colony at different times. In the United Kingdom, pup counts are converted to total pup production at each colony. This is carried out using an established statistical model that describes how the number of pups at the site varies over the season (SCOS, 2014). Both types of estimate (i.e. maximum count and modelled pup production) are used in this assessment. From calculations across Scottish colonies, maximum pup counts represent 0.82 (95% confidence interval, 0.81–0.83) of total pup production. Hereafter, both types of estimate are referred to collectively as ‘pup production’.

The assessment in the Celtic Seas did not include Ireland because the density of Irish data available was not sufficient to enable a comprehensive assessment of change in abundance. In Ireland, the status of the grey seal is considered to be ‘Favourable’ given the current knowledge of population size, distribution, ecology and prevailing pressures on these species (NPWS, 2013).

Colony survey frequency varies by ‘Assessment Unit’ (AU) (Figure A) and ranges from annual to about five-yearly. Details of seal monitoring programmes for each AU are given in Table a.

Spatial scope: assessment units (AUs)

Despite their ability to travel long distances, individual mature grey seals of both sexes are usually faithful to particular breeding sites, and may return to within 10–100 m of individual breeding locations (Pomeroy et al. 2000). The Greater North Sea and UK parts of the Celtic Seas were divided into 21 AUs  to provide a description of changes in pup production at a ‘local’ scale (Figure b). Direct impacts on pup production, such as disturbance, persecution and food availability may be more detectable at these smaller scales.

It is important to consider that AUs do not represent demographically independent populations. Additionally, survey coverage and monitoring effort is higher where breeding grey seals are most abundant. In many AUs monitoring is undertaken in specific areas by local organisations and does not form part of synoptic surveys. See Table a for details of monitoring programmes in each AU.

The AUs are the same as those used for the harbour seal abundance and distribution assessment under the indicator on Seal Abundance and Distribution . The assessment units in the Greater North Sea are broadly similar to those previously defined as OSPAR EcoQO sub-units.

Baselines

Seals have been hunted both illegally and legally for a long time and it is not possible to know the undisturbed state, nor the current carrying capacity that could be attained alongside protection from illegal hunting. Time series data for grey seal pup production do not provide an indication of a time when seal populations were not impacted or what the values would have been. It is therefore not possible to identify a baseline representing un-impacted conditions. The method adopted for this assessment assesses grey seal pup production data using both a fixed baseline and a rolling baseline with associated assessment values.

A fixed-baseline year of 1992 was used, which is the baseline year used by some European Union Member States for seals under the Habitats Directive (Council Directive 92/43/EEC). For AUs where grey seal pup production data for 1992 were not available, the start of the time series was used as the baseline.

Bearing in mind that the arbitrarily assigned baseline does not necessarily reflect a state without impacts, it is not possible to assess the status of seals in relation to the concept of ‘favourable conservation status’ under the European Union Habitats Directive. Instead, trends in the seal population are assessed. A short-term rate-based assessment value was also adopted that uses a rolling baseline (Method 1; OSPAR, 2012). Here, the average annual rate of population change over consecutive six-year periods (the reporting cycle for the European Union Marine Strategy Framework Directive; MSFD) was assessed against the next period. The rolling baseline provides a means to indicate change in population size compared with the previous six-year assessment period, rather than relying solely on an historical fixed baseline, which probably reflects a point in time when the population is already depleted due to anthropogenic pressures.

A potential issue with this type of quantitative trend thresholds, known as ‘shifting baselines’ is that each successive assessment uses a different starting point as the basis for comparison. This could result in a substantial cumulative decrease occurring over more than one six-year assessment period not being flagged as a problem, because in each six-year period the rate of decline remained below the assessment value (OSPAR, 2012).

Use of the two types of baseline and associated assessment values seeks to provide an indicator that would warn against both a slow but long-term steady decline (the problem of ‘shifting baselines’ associated with only having a rolling baseline) and against a recovery followed by a subsequent decline (potentially missed with a fixed baseline set below reference conditions).

Assessment values

OSPAR Intermediate Assessment 2017 Indicator Assessment values are not to be considered as equivalent to proposed European Union MSFD criteria threshold values, however they can be used for the purposes of their MSFD obligations by those Contracting Parties that wish to do so.

Indicator assessment values were set as a percentage deviation from the baseline value (Method 3; OSPAR, 2012).

Associated with the above baselines, two assessment values were used to assess grey seal pup production in each AU:

• Assessment value 1:No decline in grey seal pup production of >1% per year in the previous six-year period (a decline of approximately 6% over six years).
• Assessment value 2:No decline in grey seal pup production of >25% since the fixed baseline in 1992 (or closest year).

For assessment value 2: in cases where a shorter timescale than the period to 1992 is assessed, a 25% decline since the baseline is therefore not equivalent to a 25% decline in those AUs where data do extend to 1992. For example, a 25% decline since 2003 describes a more rapid contraction in population size than a 25% decline since 1992.

Using the two assessment values together would provide a means to identify a need for management measures to promote population growth (i.e. recovery) and to identify subsequent slowing of population growth as carrying capacity is approached.

Assessing change in pup production

Pup production estimates from all colonies within a given AU were summed to generate an overall assessment for that area. Quantitative assessments of trends in pup production were only completed for those AUs where three or more annual data points were available.

In addition to the assessment of the Wadden Sea (AU17), a separate assessment was conducted on the Dutch part of the Wadden Sea, where the time series began in 2001, compared to 2008 for other parts of the Wadden Sea. Earlier data exist, both for Dutch and German waters, but these were collected using different methods and so are difficult to compare with data collected subsequently.

Grey seal pup production estimates were assessed in relation to a baseline set at 1992 (or the start of the time series where this was after 1992), and by calculating the average rate of annual population change within the last reporting round (2009–2014).

To determine the change in pup production since the baseline year, generalised linear models or generalised additive models were fitted to the sum of pup production data within an AU with a quasi-Poisson error distribution and log link using all available annual survey data for the period 1992–2014. The percentage change in pup numbers since the baseline year (Formula A; Δ_abundance) and 80% confidence intervals were calculated from fitted values. Although no formal hypothesis testing was conducted, 80% confidence intervals were calculated to reflect the choice to set the significance level, α, equal to 0.20 or 20%.

To estimate the annual change in pup production over the previous six-year reporting round, a linear trend was fitted to the sum of all available data in each AU for the reporting round 2009–2014. GLMs were fitted to the sum of pup production data within an AU with a quasi-Poisson error distribution and log link. Annual growth rate (%) and 80% confidence intervals were estimated for each of the AUs. Although no formal hypothesis testing was conducted, 80% confidence intervals were calculated to reflect the choice to set the significance level, α, equal to 0.20 or 20%.

Indicator metrics and associated confidence intervals were assessed against two thresholds: the average annual growth rate (%) and 80% confidence intervals within the last reporting round were less than -1% per year, and the percentage change in abundance since baseline year (Δ_baseline) and 80% confidence intervals were less than -25%. If the 80% confidence intervals encompassed the threshold, the assessment was classified as ‘inconclusive’.

Statistical power of detecting change in pup production

There are many reasons why pup counts could vary, aside from representing true changes in pup production, These include variation in weather and a recent disturbance at a colony. It is therefore advisable to examine the variability in survey counts and to incorporate this variability into trend estimates. The International Council for the Exploration of the Sea (ICES) Working Group for Marine Mammal Ecology (WGMME) (ICES, 2014) provided general advice on the need to understand the statistical power of monitoring programmes. In the present context, statistical power is the percentage confidence in not missing a significant decline. Statistical power depends on the sample size (number of surveys), the level of statistical significance set (α-level), variance in the counts, and the magnitude of the trend, that is, -1% and -25%. The ICES WGMME (ICES, 2014) recommended that monitoring should achieve a minimum of 80% power – which equates to a 20% chance of making a Type II error (i.e. the frequency with which a true decline would not be detected). The same group also recommended that the threshold for detection of a statistically ‘significant’ trend be relaxed from the traditional α=0.05 to α=0.20. The α parameter, or statistical significance level, equates to the probability of concluding that a statistically significant trend exists when it does not (Type I error). A α value of 0.2 and power of 80% means there is equal probability of making an incorrect conclusion (either Type I or Type II error) about the detection of a trend.

Current monitoring programmes vary in the level of statistical power achievable. To carry out a full study of retrospective power to detect changes in the observed population trends, detailed information about the between- and within-year variability in all survey counts would be necessary. A full assessment of power was not undertaken here but confidence intervals (CIs)  provide a measure of confidence in the assessment. Simply said, they describe the frequency with which the true, unobservable, population parameter (here, the mean count) could be expected to fall within the intervals described by an upper and lower confidence limit. Where the confidence intervals encompass the assessment value the data do not provide conclusive evidence for the calculated value being above or below the assessment value.

It was not possible to make a quantitative assessment in all 21 Assessment Units (AUs) for several reasons. One being that grey seals may not breed in sufficient numbers in all AUs to warrant regular monitoring; in such cases the assessment area was considered ‘not applicable’ to this assessment (Figure 1). The remaining assessment gaps were due to a lack of sufficient annual data points to perform a quantitative assessment of the time series. Infrequent monitoring may be because there are relatively few animals breeding there and / or may be related to the difficulty in assessing breeding sites during the autumn and winter. Continued support for the long-term seal monitoring programmes in place within Contracting Parties will ensure that sufficient data are collected to provide quantitative assessments of grey seal pup production for those areas. The aggregation of data across national programmes in the North-East Atlantic provides a possibility for meta-analyses that place regional / national population trends into the wider regional context.

Pup production when used on its own is not a meaningful indicator of population condition.An ideal measure of population condition would include information on female fecundity and pup survival. Such demographic parameters are only available where long-term and detailed monitoring of colonies exists (e.g. Isle of May in Scotland). Another good indicator of population condition would be changes in body condition (e.g. blubber thickness).

From the present data it is not possible to exclude human influence on pup production. Even without pressure from human activities it is clear that pup production cannot keep increasing indefinitely and will reduce as the natural carrying capacity is reached. Further demographic studies could help predict the natural carrying capacity for the number of grey seals in the North-East Atlantic and provide an indication of likely future trends in pup production, with and without impacts from human pressures.

A demographic link between AUs is possible, but the extent of movement of adults between colonies and how this influences local pup production is at present poorly understood. Further international studies on the movements of individual seals could help better understand this and even prevent unintended disturbance of these inter-colony exchanges.

ICES 2014. Report of the working group on Marine Mammal Ecology (WGMME). 10-13 March, 2014 Woods Hole, Massachusetts, USA. ICES CM 2014/ACOM: 27, 230pp

NPWS (2013). The Status of Protected EU Habitats and Species in Ireland Overview Volume 1. Unpublished Report National Parks & Wildlife Services. Department of Arts, Heritage and the Gaeltacht Dublin Ireland.)

OSPAR, 2012. MSFD Advice Manual and Background Document on Biodiversity. Version 3.2. OSPAR Biodiversity Series. Available at: http://www.ospar.org/documents/dbase/publications/p00581/p00581_advice%20document%20d1_d2_d4_d6_biodiversity.pdf

Pomeroy, P. P., Twiss, S. D. and Redman, P. 2000. Philopatry, site fidelity and local kin associations within grey seal breeding colonies. Ethology: 899–919.

SCOS. 2014. Scientific advice on matters related to the management of seal populations, 2014. Currently available on the SCOS website at http://www.smru.st-and.ac.uk/documents/2259.pdf

Stringell, T., Millar, C., Sanderson, W., Westcott, S. and McMath, M. 2014. When aerial surveys will not do: grey seal pup production in cryptic habitats of Wales. Journal of the Marine Biological Association of the United Kingdom. 94: 1155-1159.

Brasseur, S., van Polanen Petel, T., Gerrodette, T., Meesters, E., Reijnders, P., Aarts, G. 2015. Rapid recovery of Dutch grey seal colonies fuelled by immigration. Marine Mammal Science. 31: 405-426.

SCOS. 2014. Scientific advice on matters related to the management of seal populations, 2014. Currently available on the SCOS website at http://www.smru.st-and.ac.uk/documents/2259.pdf