Marine Bird Abundance

Overview

This indicator assessment begins by constructing a time series of annual estimates of ‘relative breeding abundance’ or ‘relative non-breeding abundance’ of each species in each OSPAR Region or sub-division thereof (see section on Spatial Aggregation and Figure a). For the majority of species, data are available for breeding abundance or non-breeding abundance; both are available for only five species in the Arctic Waters, 14 species in the Greater North Sea and just one species in the Celtic Seas. Relative abundance is the number of adult birds or breeding pairs estimated each year as a proportion of a baseline (see section below on parameter/metric). Each species-specific estimate of annual relative abundance is compared against an assessment value. This indicator assessment is based on the proportion of species in a region or sub-division of that region that have exceeded assessment values for relative abundance in the most recent year of the time series. Separate assessments were carried out for relative breeding abundance and relative non-breeding abundance.

This indicator assessment is confined to three OSPAR Regions: Arctic Waters, the Greater North Sea and the Celtic Seas. The assessment of Arctic Waters is based on data from Norway only, but includes both breeding and non-breeding abundance data from both the Norwegian Sea and Barents Sea (excluding Jan Mayen and Svalbard). The assessment of the Greater North Sea includes both breeding and non-breeding abundance data from all Contracting Parties, except France. The assessment of the Celtic Seas is based on breeding abundance data from Ireland and the United Kingdom and on non-breeding abundance data from the United Kingdom only. No assessments were possible for the Bay of Biscay and Iberian Coast and for the Wider Atlantic because no data were provided to this assessment by France, Spain and Portugal.

Species-Specific Indicators of Relative Abundance

Data acquisition

In 2016, the following data were requested from Contracting Parties on counts of breeding and non-breeding adult birds collected during the period 1980–2015:

1. Breeding seabird colonies and breeding waterbirds nesting close to the coast and using the marine environment (e.g. for food): counts of breeding pairs (preferably, or failing that adult birds) per species per colony per year.
2. Wintering and migrating waterbirds: numbers of birds per species per site per year that are counted from land or from the air. These data mostly comprise maximum or single counts conducted in January (N.B. within the Arctic Circle counts are conducted in March when there is sufficient daylight to do so). Data from the Wadden Sea Trilateral (Germany, Netherlands and Denmark) Monitoring and Assessment Programme and from the United Kingdom Wetland Bird Survey also comprised a mean of counts conducted throughout a one-year period (from July in one year to June in the next).

Most data refer to individual colonies or sites rather than large stretches of coastline. Exceptions are the Wadden Sea, the Dutch Delta and for non-breeding data in the United Kingdom, where colonies have been aggregated to regions and sub-divisions.

Spatial aggregation

Separate indicator assessments were conducted for each species in each OSPAR Region. An assessment was also conducted for each sub-division of the Greater North Sea, and the mainland part of the Norwegian Arctic Waters (Figure a). These smaller sub-divisions may help to interpret the assessment results. The sub-divisional boundaries are based on a coarse assessment of the main oceanographic features such as currents and depth, and some relatively clear-cut differences in seabird / waterbird community structure and population trends (ICES, 2013c,d).

Data used in this Assessment

When all data from participating Contracting Parties had been collated, it became clear that prior to 1992, the number of annual counts per site was much lower in some areas, with many sites not being surveyed at all. Prior to 1992, a large amount of data would need to be interpolated and confidence in the accuracy of trends in abundance during this period would be very low. For this reason, the data used in this indicator assessment were restricted to 1992 onwards. The Arctic Waters and Celtic Seas regional scale assessments used data collected during the period 1992–2015. In the Greater North Sea, regional assessments were only possible for the period 1992–2014, because more recent data were not available for the Wadden Sea.

In the Wadden Sea (Denmark, Germany, Netherlands), the status of non-breeding waterbirds is best indicated by year-to-year change in the mean of counts conducted throughout the year (i.e. from July in one year to June in the next) (Blew et al., 2016). These year-round counts are also conducted in the United Kingdom and in the Netherlands (outside of the Wadden Sea), but are not carried out elsewhere in the Greater North Sea. All Contracting Parties supplied maximum or single counts conducted in January (N.B. within the Arctic Circle, counts are conducted in March when there is sufficient daylight to do so). The January counts were used for the indicator assessment in the Greater North Sea region as a whole and separately, in each Greater North Sea sub-division, apart from the southern North Sea (sub-division d in Figure a), where mean year-round counts from the United Kingdom, Netherlands, Germany and Denmark were used (year-round counts were not available for Belgium). Year-round counts for the Wadden Sea (Denmark, Germany, Netherlands) were only available up to July 2013- June 2014 (i.e. 2013/14). Hence, the indicator assessments for each of the Greater North Sea sub-divisions were restricted to the period 1992/93 – 2013/14.

Trend Analysis

This indicator assessment requires for each species an annual estimate of either breeding or non-breeding abundance (depending on species) per site or colony. Not all the colonies and sites in the data provided had been observed every year. Missing annual observations were interpolated from the observed data using statistical models. The minimum number of years of counts for a particular species required for a colony or site to be included in the analysis was set at two for all species except northern fulmar (Fulmarus glacialis), which was set at a minimum of five years (ICES, 2010, 2011).

Some Contracting Parties provided data that had no missing values because these had already been interpolated prior to submission, using tried and tested methods:

1. United Kingdom non-breeding abundance, all species (for methodology see Underhill et al., 1994).

2. Denmark, Germany and the Netherlands in the Wadden Sea (Greater North Sea sub-division d) non-breeding abundance, all species (for methodology see Blew and Südbeck, 2005; Blew et al., 2016; see also www.waddensea-infogate.org/Migratory%20Birds/Migratory_birds.html).

3. Denmark, Germany and the Netherlands in the Wadden Sea (Greater North Sea sub-division d) breeding abundance, all species (for methodology see Koffijberg et al., 2015; see also www.waddensea-infogate.org/Breeding%20Birds/Breeding_birds.html).

For all other data submitted by Contracting Parties, missing annual observations were interpolated using a modified chain method, first developed by Thomas (1993). This method estimates values of missing observations based on information in other years and sites. The advantage of this method is that it allows for site-specific variation, thereby avoiding the conventional assumption that changes in abundance at different sites occur synchronously. A further advantage of this approach is that it can easily incorporate counts of whole colonies and counts from smaller plots within the same colonies that are monitored more frequently than the whole colony. The Thomas estimation method has been used to construct trends in abundance for earlier iterations of this indicator and its OSPAR Ecological Quality Objective (EcoQO) predecessor. Details of the method are as reported by ICES (2008: Annex 3).

Applying Regional Weightings to Abundance Trends

Not all the colonies or sites in an assessment unit will be monitored and present in the dataset. The proportion of a population that is monitored varies between species and between countries. There is a resultant bias, in that those countries where few sites are monitored are under-represented in the trends for a given assessment unit, compared to those countries where a larger proportion of sites are monitored.

To remove such bias, the annual estimates of breeding and non-breeding abundance in each country were weighted according to the size of the population in that country. Each Contracting Party provided recent estimates of total abundance for each species along their coastline within each of the assessment units in Figure a. To apply a regional weighting, each annual estimate of abundance in each assessment unit was divided by a proportion p, where pis the proportion of the total population that is present within the sites or colonies that are included in the data provided. The total number of adult birds or pairs in an assessment unit were, in most instances, taken from national censuses.

As an example, the weighted annual breeding abundance of a species in the Celtic Seas region, y_III in year j, was calculated from annual estimates of abundance in each constituent country, i.e. the United Kingdom Celtic Seas coast (y_UK) and Ireland (y_IRE) as follows:

y_IIIj= (y_UKj / p_UK) + (y_IREj / p_IRE)

where p_UK and p_IRE are the proportions of the respective populations in the United Kingdom (UK) and Ireland (IRE) that are contained in the sample of colonies that were monitored in each constituent country in the Celtic Seas region.

Regional weightings were not necessary where data contained all the colonies or sites in that area. In Belgium, all the breeding and non-breeding sites in the country are monitored, so no weightings were required for these data. Regional weighting was also not necessary for estimates of annual abundance that had been interpolated before submission. These included trends in breeding and non-breeding abundance in the Dutch, German and Danish Wadden Sea, and trends in non-breeding abundance in the United Kingdom.

Parameter/Metric

To compare the state of species that have very different population sizes, estimates of abundance (measured in numbers of adult birds or breeding pairs) were converted to an indicator metric that uses a single scale (i.e. a proportion) for all species. The indicator metric is relative abundance, which is annual abundance expressed as a percentage of the baseline:

relative abundance = annual abundance / baseline abundance

Baselines

It is preferable to set baselines objectively, using one of the methods recommended by ICES (2015). But most Contracting Parties did not provide baselines in 2016. As an alternative, relative abundance was calculated using a baseline equal to the abundance at the start of the time series (i.e. 1992).

Species Selection and Aggregation (Functional Groups)

There were sufficient data to construct species-specific indicators of relative breeding abundance for the following 32 species: Arctic skua, Arctic tern, Atlantic puffin, black guillemot, black-headed gull, black-legged kittiwake, common eider, common guillemot, common gull, common tern, Eurasian spoonbill, European shag, great black-backed gull, great cormorant, great skua, herring gull, Kentish plover, lesser black-backed gull, little tern, mallard, Mediterranean gull, mute swan, northern fulmar, northern gannet, oystercatcher, pied avocet, razorbill, red-breasted merganser, ringed plover, roseate tern, Sandwich tern, and shelduck.

There were sufficient data to construct species-specific indicators of relative non-breeding abundance for the following 57 species: barnacle goose, bar-tailed godwit, Bewick’s swan, black guillemot, black-headed gull, black-tailed godwit, Brent goose, Canada goose, common eider, common gull, common merganser, common scoter, coot, dunlin, Eurasian curlew, Eurasian spoonbill, European shag, goldeneye, golden plover, great black-backed gull, great cormorant, great crested grebe, greater scaup, Greenland white-fronted goose, greenshank, grey plover, herring gull, king eider, lapwing, long-tailed duck, mallard, mute swan, oystercatcher, pied avocet, pintail, pochard, purple sandpiper, red-breasted merganser, red knot, red-necked grebe, redshank, red-throated diver, ringed plover, ruff, sanderling, shelduck, shoveler, Slavonian grebe, smew, spotted redshank, teal, tufted duck, turnstone, velvet scoter, whimbrel,whooper swan and wigeon.

Species were assigned to the functional groups given in Table a. The species assessed and the functional groups to which they were assigned, are given in Table b.

​​​​​​Species-specific assessment values

This assessment uses two different assessment values that are designed to reflect the resilience of different species to declines in their population (see ICES, 2008, 2010, 2011). It is desirable for the annual relative abundance of a species to be above, either:

• 0.8 (i.e. 80% of the baseline) – for species that lay one egg; or
• 0.7 (i.e. 70% of the baseline) – for species that lay more than one egg.

The reason for having two assessment values is because species that lay only one egg are expected to recover more slowly from declines in population size than species that can potentially produce more than one chick per year. If relative abundance is below the appropriate assessment value, it is considered to be in ‘poor’ status and further research and / or management is recommended, depending on what is appropriate.

An upper assessment value of 1.3 (i.e. 130% of the baseline) was applied to the annual relative abundance of all species. This upper assessment value is used to identify potentially disruptive increases in some species that might impact on other species (ICES, 2008). For instance, large predatory seabird species have benefited from the provision of food from fishery discards. The increase in numbers of species such as great black-backed gull and great skua have, in some areas, led to declines in species such as kittiwake that they prey on. However, this has the potential to wrongly identify a species as having a detrimental impact on other species when in fact it is in recovery to levels in excess of the baseline (ICES, 2010, 2011, 2013b). As a result, this upper assessment value is not used as an indicator of status and is only used to provide a trigger for further research and / or management, if increases in one species are likely to result in decreases in others.

Integration of species-specific assessments

The status of marine bird communities was assessed by calculating the proportion of species exceeding the lower assessment values, as previously used for breeding seabirds in the EcoQO on seabird population trends as an index of community health (ICES, 2008), according to the following integration rule:

Changes in abundance of marine birds should exceed species-specific assessment values in 75% or more of species that are assessed.

Humphreys et al. (2012) also recommended a value of 75% for non-breeding shorebirds and coastal breeding waterbirds in the United Kingdom because it is comparable to the assessment values used for shorebirds by the WeBS Alerts system.

In this indicator assessment, relative breeding abundance and relative non-breeding abundance were assessed separately. This is because most species in the breeding assessment are seabirds that use the wider marine environment and most species in the non-breeding assessment use intertidal and inshore areas. The breeding and non-breeding assessments therefore indicate impacts from different suites of pressures, operating in different parts of the marine environment.

To provide greater insight into the likely impacts operating on relative breeding abundance and on relative non-breeding abundance, species-specific assessments were integrated at different spatial scales: for each OSPAR Region and for each sub-division of the Greater North Sea and Arctic Waters. Within each Region, species-specific assessments were also integrated for each functional group (Table a).

Expansion of the indicator to other OSPAR Regions

Of the 21 marine bird species breeding in the Bay of Biscay and Iberian Coast region, nine occur in very low numbers, and with no monitoring data collected for two (Cory’s shearwater, Calonectris diomedea and band-rumped storm-petrel, Oceanodroma castro) (ICES, 2008). Data quality for six of the 10 remaining species was assessed as ‘good’, three were assessed as ‘sparse’, and data quality for one (little tern, Sternula albifrons) breeding in Portugal was unknown (ICES, 2008).

The Azores are currently working with the other Macaronesian archipelagos (Madeira and the Canary Islands) to establish a common monitoring plan for breeding seabirds within the framework of the MISTIC Seas Project (European Union funded). This plan will be implemented, with the support of a recently granted European Union Project MISTIC SEAS II, and will enable the collection of data on seabird abundance.

Inclusion of at-sea data

Data on seabirds at sea, collected from boats or planes, were not included in the abundance indicator. However, these data need to be included in the future to obtain reliable results on trends in species that occur in substantial numbers offshore. At present, several Contracting Parties carry out, or plan to carry out, national at-sea monitoring programmes. Elsewhere, there are either no at-sea surveys or those that do exist are very limited in spatial and temporal coverage. Overall, coordination of surveys (e.g. regarding timing) between countries is lacking. There is a need to develop a concept for joint survey efforts delivering the necessary data basis for the abundance indicator work, to implement this concept in the frame of national survey programmes in future years, and to develop a methodological approach for aggregating and analysing the data.

Baselines

Most Contracting Parties did not provide baselines in 2016. As an alternative, relative breeding abundance was calculated using a baseline equal to the abundance in the second year of the time series (i.e. 1992). ICES (2015) recommended that it is preferable to set baselines objectively, using one of the methods (a) or (b), below.

• 'Historical reference’: where abundance is known at a point in the past long before the time series began, but where the reasons for subsequent change are unknown. Historical population estimates should be used as baselines if they were recorded: before known human impacts; and / or before other major declines in population; or at known plateaus in population trends, following increases and peaks in population size.
• ‘Reference level’: the population size that could be expected if human impacts were negligible (this can be derived from known population sizes either historically or from within available time series). Use of the highest known population estimate when the population has decreased in size, as a result of human impacts (e.g. periods of severe contamination) or following stochastic natural impacts (e.g. severe weather events). Use of recent population estimates (e.g. previous five-year mean) when a species has been colonising.

ICES. 2008. Report of the Workshop on Seabird Ecological Quality Indicator, 8–9 March 2008, Lisbon, Portugal. ICES CM 2008/LRC:06. 60 pp.

ICES. 2010. Report of the Working Group on Seabird Ecology (WGSE). 15–19 March 2010, Copenhagen. ICES CM 2010/SSGEF:10. 81 pp.

ICES. 2011. Report of the Working Group on Seabird Ecology (WGSE). 1–4 November 2011, Madeira, Portugal. ICES CM 2011/SSGEF:07. 77 pp.

ICES. 2012. ICES advice on EcoQO for seabird populations in OSPAR regions II and III. In Report of the ICES Advisory Committee, 2012. ICES Advice 2012, Book 1, Section 1.5.5.1. Also available as a separate advice sheet at: http://www.ices.dk/sites/pub/Publication%20Reports/Advice/2012/Special%20Requests/OSPAR_EcoQO_for_seabird_populations.pdf

ICES. 2013a. OSPAR request on ecological quality objective for seabird populations in OSPAR Region III (Celtic seas). In Report of the ICES Advisory Committee, 2013. ICES Advice 2013, Book 1, Section 1.5.6.1. Also available as a separate advice sheet at: http://www.ices.dk/sites/pub/Publication%20Reports/Advice/2013/Special%20requests/OSPAR_EcoQO_region_III.pdf

ICES. 2013b. Report of the ICES Ad hoc Group on Seabird Ecology (AGSE), 28–29 November 2012, ICES Headquarters, Copenhagen, Denmark. ICES CM 2012/ACOM:82. 30 pp.

ICES. 2013c. Report of the Joint ICES/OSPAR Expert Group on Seabirds (WGBIRD), 22–23 October 2013, ICES Headquarters , Copenhagen, Denmark. ICES CM 2013/ACOM:78. 24 pp.

ICES. 2013d. OSPAR request on an update of the ecological quality objective (EcoQO) on seabird population trends. In Report of the ICES Advisory Committee, 2013. ICES Advice 2013, Book 1, Section 1.5.6.9.

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

Blew, J., Günther, K., Hälterlein, B., Kleefstra, R., Laursen,K., Scheiffarth, G. 2016. Trends of Migratory and Wintering Waterbirds in the Wadden Sea 1987/1988 - 2013/2014. Wadden Sea Ecosystem No. 37. Common Wadden Sea Secretariat, Joint Monitoring Group of Migratory Birds in the Wadden Sea,Wilhelmshaven, Germany.

Blew, J. and Südbeck, P. (Eds.) 2005. Migratory Waterbirds in the Wadden Sea 1980 – 2000. Wadden Sea Ecosystem No. 20. Common Wadden Sea Secretariat, Trilateral Monitoring and Assessment Group, Joint Monitoring Group of Migratory Birds in the Wadden Sea, Wilhelmshaven, Germany.

Humphreys E M, Risely K, Austin G E, Johnston A and Burton N.H.K. 2012. Development of MSFD Indicators, Baselines and Targets for Population Size and Distribution of Marine Birds in the UK. BTO Research Report No. 626.

Koffijberg, K., K. Laursen, B. Hälterlein, G. Reichert, J. Frikke & L. Soldaat 2015. Trends of Breeding Birds in the Wadden Sea 1991 - 2013. Wadden Sea Ecosystem No. 35. Common Wadden Sea Secretariat, Joint Monitoring Group of Breeding Birds in the Wadden Sea, Wilhelmshaven, Germany.

Thomas, G.E. (1993) Estimating annual total heron population counts. Appl. Statistics 42: 473-486.

Underhill, L.G. & Prŷs-Jones, R. (1994) Index numbers for waterbird populations. I. Review and methodology.  J. Appl. Ecol., 31, 463-480.