Marine Bird Breeding Success / Failure

Indicators of breeding failure were constructed using time series of annual mean breeding success (number of chicks fledged per pair, clutch or nest) of marine bird species at colonies throughout the North-East Atlantic (total counts or survey plots). A separate assessment was conducted for each species in each OSPAR Region. An assessment was also conducted for each sub-division of the Greater North Sea, as shown in Figure a, and the Norwegian parts of the Arctic Waters. Separate assessments were not conducted for each Arctic sub-division because there were too few colonies included in each sub-division.

The breeding failure assessments for each species were constructed from time series of annual estimates of breeding success at a sample of colonies. The time series used covered the periods 1991–2015 for the Greater North Sea and Celtic Seas and 2010–2015 for the Norwegian parts of Arctic Waters. Not all colonies in the sample will have been observed every year in the time series, therefore (prior to the analysis) missing annual observations were predicted using models (for methodology see Trend Analysis).

Species-Specific Assessment of Breeding Failure

Rationale for Assessment Values

The assessment values applied for this indicator were developed by Cook et al. (2014a) and were agreed to be technically appropriate by the Joint OSPAR / ICES / HELCOM Working Group on Marine Birds (ICES 2015). The rationale and values are described in the following sections.

OSPAR Intermediate Assessment (IA) 2017 Indicator Assessment values are not necessarily to be considered as equivalent to proposed EU Marine Strategy Framework Directive (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.

Parameter / metric

‘Annual colony failure rate’, that is, the percentage of colonies failing per year, per species (from Cook et al. 2014a).

‘Failure’ is defined as when a colony fails to produce any young or produces very few young during a single breeding season. Failure is quantified by Cook et al. (2014a) as when the mean annual productivity of a breeding colony is 0.1 or less chicks per pair, clutch or nest. As failure could be interpreted as an unusual deviation from ‘normal’ levels of breeding success, the precise assessment value below which a colony is defined as failing may be different at some colonies, even for the same species. The assessment value used for determining failure can be adjusted according to knowledge of the colony in question. Ideally, the assessment value should be obtained from any clear response to important environmental factors such as low food availability (e.g. Cury et al. 2011). The assessment value of 0.1 chicks per pair, clutch or nest should be used as the default assessment value, unless there is good evidence to show that ‘failure’ of some species in some areas should be set at a different level (ICES 2015).

Trend analysis

The assessments for each species are constructed from a time series of annual estimates of breeding success at a sample of colonies. Not all colonies in the sample will have been observed every year in the time series. Missing annual observations were predicted using a Generalised Linear Model (GLM) framework with a binomial error structure (after Cook et al. 2014a and 2014b). Breeding success for each colony in each year was calculated, and where this value was 0.1 or less chicks per pair, clutch or nest, the colony was assessed as having failed in that year. Breeding success or failure was modelled in relation to year and site, because not all sites were covered in all years. The coefficient for each year was then taken to represent the probability of breeding failure occurring at any given site within that calendar year. The parameter ‘Year’ was included in the GLM as a fixed effect factor, rather than a random effect, so that the coefficients would not be constrained to follow a normal distribution.

To minimise the impact of differences in sampling rate, and to ensure that breeding success was likely to be representative of the colony as a whole, minimum criteria were set for inclusion of data within the model. Only those colonies at which a mean of ten nests were monitored for at least three years were considered. To prevent any individual site having an undue influence over the value of the coefficients, a jack-knife statistical approach was used, dropping each site from the model in turn. Models were run for each species in each sub-region in turn. The final indicator value presented for each species, in each sub-region, in each year is the mean probability of breeding failure calculated from each run of the jack-knife analysis.

Assessment

In summary, ‘breeding failure’ is when the annual mean breeding success of a colony is 0.1 or less chicks per pair, clutch or nest, and ‘annual colony failure rate’ is the percentage of colonies of species that are experiencing breeding failure in a single breeding season.

Breeding failure for each species was assessed in two stages:

1. Assessing if colony failure is widespread - the annual colony failure rate (the percentage of colonies experiencing breeding failure) can be considered to be ‘widespread’ if it exceeds the assessment values set for each species (see below).

2. Assessing the frequency of widespread breeding failure occurrence. Widespread breeding failure will be considered to be frequently occurring if it occurs in four years or more out of six.

Assessing if colony failure rate is widespread

The annual colony failure rate (the percentage of colonies failing) of each species was assessed against one of the two upper assessment values below, depending on the species. This was done for each region:

a) terns: mean percentage of colonies failing per year, over the preceding 15 years;

b) all species except terns: 5% of colonies failing per year.

The aim of the assessment values for annual colony failure rate is to identify widespread breeding failures and to differentiate large-scale impacts from local problems, where only a small proportion of colonies fail per year. The above assessment values were taken from Cook et al. (2014a), who tested various assessment values for each species.

A different assessment value was applied to the breeding failure rate of terns because they often desert colonies, sometimes before laying eggs, in response to local disturbances or impacts on food supply (Shealer and Kress 1991; Holt 1994; Cook et al. 2011). The assessment value for terns is designed to identify years of unusually high rates of breeding colony failure.

A fixed assessment value of 5% was appropriate for all other species that do not tend to desert colonies all at once. Cook et al. (2014a) found that years in which colony failure rate is more than 5% are much rarer in other species and they therefore provide a good indicator that pressures may be impacting on the population.

Assessing the frequency of widespread colony failure

For each species, the frequency of widespread colony failure was assessed over the most recent six-year period from 2010 to 2015, inclusive. A six-year period was chosen because it equals the length of the EU’s MSFD reporting cycle. To carry out an assessment of a species over a certain time period, there needed to be some observed data of breeding success at some colonies during the most recent year in the period (2015).

One or two years of widespread colony failure were considered as ‘acceptable’, given the wide range of possible natural and anthropogenic factors that could cause breeding failure in some species. The cumulative effect of widespread colony failure in four or more years out of six, was considered most likely to have a significant impact on recruitment into the regional population (Cook et al. 2014a). Low recruitment could lead to declines in population size and to affect the assessments of the OSPAR indicator on marine bird abundance.

Breeding failure is a ‘life history’ strategy for some species such as Arctic tern which, if conditions are sub-optimal, will desert a colony all at the same time, rather than remaining and trying, but failing, to raise young. Therefore, the definition of failure used in this indicator assessment (i.e. <0.1 chicks fledged per pair, clutch or nest) may provide an overly pessimistic indication of breeding performance in such species. However, this approach of assessing the frequency of widespread failure could reduce the chance of false negative assessments being made (ICES 2015).

Species Selection and Aggregation (functional groups)

There were sufficient time series data to conduct assessments of 16 species in the Norwegian parts of Arctic Waters, 20 in the Greater North Sea and 20 in the Celtic Seas. Overall, 25 different species were assessed in at least one region.

Some species were omitted from the regional assessments because the time series did not contain an estimate of breeding success in 2015 and the model was unable to predict missing values beyond the length of the time series. These included data from the Netherlands on common eider (Somateria mollissima), Eurasian spoonbill (Platalea leucorodia), pied avocet (Recurvirostra avosetta), and oystercatcher (Haematopus ostralegus) and from roseate tern (Sterna dougallii) colonies in Ireland (Celtic Seas). However, the breeding failure rates for these species in the years when data were available are presented in Figure b26 to Figure b29 for information.

Table a presents the five marine bird functional groups. The species assessed in each region were all seabirds from two of the five functional groups:

a) Surface feeders, which forage on small fish, zooplankton and other invertebrates at or within the surface layer (within 1 m to 2 m of the surface);

b) Water column feeders, which actively dive below the surface to a broad range of depths to feed on fish and invertebrates (for example, squid, zooplankton) in the water column or close to the seabed.

The 25 different species assessed and the functional groups to which they were assigned, are presented in Table b.

^adenotes species that were omitted from the regional assessments because the time series did not contain an estimate of breeding success in 2015.

Arctic Waters

The Arctic sub-region contains the highest concentrations of breeding seabirds in the North-East Atlantic. For the first time, this Indicator Assessment includes data from monitoring of seabird productivity along the Norwegian coasts (including Svalbard and Jan Mayen) of the Norwegian and Barents Seas. In future it would be beneficial if other Contracting Parties in the region could make their monitoring data available in a similar way.

Greater North Sea

Most countries in the Greater North Sea collect seabird productivity data and undertake some monitoring of a few breeding waterbird species (Table c). The main gap in data is in the Skagerrak and Kattegat where breeding success is measured along the Norwegian coast, possibly along the Swedish coast, but not along the Danish coast. There is a coordinated scheme of annual monitoring of breeding success within the Wadden Sea (Netherlands, Denmark and Germany) that was initiated in 2009 (monitoring in the Netherlands part of the Wadden Sea started in 2005). However, data collected by this scheme since 2012 in Denmark, Lower Saxony and Schleswig-Holstein (Germany) were not available and therefore these areas were not included in this assessment. Continued monitoring in these areas should mean they will be included in future assessments of this indicator.

Bay of Biscay and Iberian Coast

In the Bay of Biscay and Iberian Coast, monitoring of productivity in France and Spain has created time series of data suitable for constructing this Indicator Assessment, but in Spain this is restricted to a single species – the European shag (Table c). It is uncertain what productivity monitoring is carried out along the Portuguese mainland coast.

Wider Atlantic

Currently, the Azores, in partnership with the other Macaronesian archipelagos outside the OSPAR Maritime Area (Madeira and the Canary Islands) have been working to establish a common monitoring plan for breeding seabirds within the framework of an EU co-financed project, MISTIC Seas. If the plan is implemented, it would enable data on seabird breeding success (and other parameters) to be collected and used for future assessment cycles.

Assessment Methods: Knowledge Gaps

The ICES / OSPAR / HELCOM Joint Working Group on Marine Birds (JWGBird) developed this indicator assessment but has acknowledged some limitations (ICES 2015). The assessment methods for the marine bird breeding success / failure indicator currently focus on the extreme events of almost no chicks being produced by a colony, on average, per year. In doing so, they fail to identify other years where poor breeding success could still have significant negative impacts on the population in the longer term.

However, it is not straightforward to categorise annual breeding success as ‘good’ or ‘poor’. The reason breeding has not been directly assessed as ‘good’ or ‘poor’ in this indicator is because the number of chicks that need to be produced each year to sustain a population or cause it to grow, varies substantially as other demographic parameters (e.g. survival rates) also vary in space and time. Information on demographics such as survival rate, age at first breeding and immature survival rates are more resource demanding to measure owing to the need to monitor individual birds from year to year. For well-studied species and at a few intensively studied sites these data do exist.

A possible step forward towards setting accurate and objective targets for annual breeding success rates would be to collate an inventory of ongoing monitoring of survival rates in the North-East Atlantic and conduct a review of published estimates. Once survival estimates and other demographics have been collated, some simple population modelling could be undertaken to produce some preliminary estimates of the levels of breeding success required to sustain or increase the population.

Cook, A.S.C.P., Calbrade, N.A., Austin, G.E. and Burton, N.H.K. 2011. Determining foraging use of the Dee estuary by common terns from the recent declining colony at the Shotton Lagoons and Reedbeds SSSI. BTO report to CCW, Thetford.

Cook, A.S.C.P., Robinson, R.A. and Ross-Smith, V.H. 2014a. Development of MSFD Indicators, Baselines and Target for Seabird Breeding Failure Occurrence in the UK (2012), JNCC Report 539, ISSN 0963 8901.

Cook A.S.C.P., Dadam, D., Mitchell, I., Ross-Smith, V.H. and Robinson, R.A. 2014b. Indicators of seabird reproductive performance demonstrate the impact of commercial fisheries on seabird populations in the North Sea. Ecological Indicators 38: 1–11.

Cury, P.M., Boyd, I.L., Bonhommeau, S., Anker-Nilssen, T., Crawford, R.J.M., Furness, R.W., Mills, J.A., Murphy, E.J., Österblom H., Paleczny, M., Piatt, P.F., Roux, J.-P., Shannon, L. and Sydeman, W.J. 2011. Global seabird response to forage fish depletion – one-third for the birds. Science 334: 1703–1706.

Holt, D.W. 1994. Effects of short-eared owls on common tern colony desertion, reproduction, and mortality. Colonial Waterbirds 17: 1-6.

ICES. 2015. Report of the Joint ICES/OSPAR Working Group on Seabirds (JWGBird), 17–21 November 2014, Copenhagen, Denmark. ICES CM 2014/ACOM: 30. 115 pp.

ICES. 2015. Report of the Joint ICES/OSPAR/HELCOM Working Group on Seabirds (JWGBIRD), 9–13 November 2015, Copenhagen, Denmark.u117234.42pp.

Shealer, D.A. and Kress, S.W. 1991. Nocturnal abandonment response to black-crowned night-heron disturbance in a common tern colony. Colonial Waterbirds, 14, 51-56.