Census Methods

Co-ordination of Seabird 2000

In May 1998, the Joint Nature Conservation Committee (JNCC) appointed Ian Mitchell as Seabird 2000 Co-ordinator. He was responsible for the planning and completion of survey work, collation of data and final analysis and production of this book. The Seabird 2000 co-ordinator implemented the recommendations of a steering group comprised of members affiliated to each of the Seabird 2000 Partners (Table 1). In the Republic of Ireland, Seabird 2000 was co-ordinated by Stephen Newton of BirdWatch Ireland.

The co-ordination of fieldwork in Britain, Northern Ireland, the Channel Islands and the Isle of Man was achieved through a team of 45 regional co-ordinators (see Acknowledgements). Of these, 17 were area staff of the Seabird 2000 partners, Scottish Natural Heritage (SNH), Natural England (formally English Nature), Countryside Council for Wales (CCW) and the Royal Society for the Protection of Birds (RSPB); the remainder were volunteer enthusiasts, often County Bird Recorders or Regional Representatives for the British Trust for Ornithology. The role of regional co-ordinators was to i) recruit surveyors; ii) assign surveyors to colonies within their region; iii) ensure all colonies in the region were surveyed; iv) issue surveyors with survey instructions, recording forms and guidelines on health and safety. In the Republic of Ireland virtually all fieldwork was co-ordinated centrally by Stephen Newton with assistance from Oscar Merne of National Parks & Wildlife (NPW), Dúchas the Heritage Service. the fieldwork was conducted by BirdWatch Ireland volunteers and staff, and NPW Regional Management and Research staff.

More than 1,000 surveyors contributed to Seabird 2000 (see Acknowledgements). Most volunteers were recruited by the regional co-ordinators and by Ian Mitchell via the media, posters, leaflets and a biannual newsletter. Small grants were awarded to volunteers by the Seabird Group to cover incidental expenses (e.g. petrol, boat hire etc).

Duration

Most counts were conducted between 1998 and 2002. The original plan was to complete field work in 2001, but in March 2001, foot and mouth disease broke out in England and soon spread throughout the country and into Wales, southern Scotland, Northern Ireland and in Louth, Republic of Ireland. Restrictions aimed at containing the disease and preventing its spread into northern Scotland and most of Ireland severely hindered Seabird 2000 fieldwork in 2001, and many surveys had to be postponed until 2002. Fieldwork in 1998 consisted of a Black Guillemot survey of eastern Ireland (Louth to Waterford), a Black Guillemot survey of the Yell Sound (Shetland) and surveys of cliff-nesting seabirds on the Scottish islands of Sule Skerry, Sule Stack (both in Orkney), Sula Sgeir, North Rona, Mingulay, Berneray and the Flannan Isles (all in the Western Isles). Counts conducted in 2003 of Black Guillemots in northwest Skye and of Atlantic Puffins on the Farne Islands (Northumberland) were also included.

In each year counts were conducted between late March and late July, depending on the species (see below).

Coverage

Barton (1997) assessed the resources required to undertake a complete census of breeding seabirds in Britain and Ireland. Using data from the SCR Census (1985-88), he identified potential problems and the best strategy to obtain accurate population estimates for all 25 species of seabird breeding in Britain and Ireland. He suggested that strategies be deployed for groups of species with similar survey methodology. For certain species groups (e.g. storm-petrels, inland gulls) and certain geographical areas (NW Scotland and W. Ireland) where complete survey coverage would be difficult to achieve, he suggested sampling strategies that would provide an unbiased estimate of the total population of Britain and Ireland. The sampling strategies involved selecting at random and surveying a certain percentage of either colonies, or 20km OS grid squares, or both (see Barton 1987) 

Given the various scenarios presented by Barton (1997), the Seabird 2000 Steering Group decided that Seabird 2000 should aim to count all colonies in Britain and Ireland. This presented two major challenges for the co-ordinators: 1) finding all existing colonies, and 2) ensuring that no major gaps were left. To ensure that all colonies were located, a list of all known colonies was extracted from the Seabird Colony Register (SCR) Database, administered by JNCC. The SCR Database contains records of colonies throughout Britain and Ireland, including the results of the two preceding national censuses (see Introduction). Regional co-ordinators were sent a list of all colonies extracted from the SCR Database within their region, containing details of the most recent counts of each species at each site. The list was divided into i) cliff-nesters (i.e. Northern Fulmar, European Shag, Black-legged Kittiwake, Great Black-backed Gull, Common Guillemot, Razorbill and Atlantic Puffin); ii) Great Cormorant; iii) inland/urban nesting gulls; iv) Manx Shearwater, Leach's Storm-petrel and European Storm-petrel; and v) Black Guillemots. To ensure that no major gaps were left, colonies of cliff-nesters, Great Cormorants, inland gulls and Black Guillemots were prioritised in terms of size (cf. Barton 1987). In 1999, all colonies were targeted that contained 5% of the British or Irish population of one or more species. The remaining colonies were then divided into high and low priority, and co-ordinators were asked to ensure that all high priority colonies were at least counted. For each species, colonies were listed in order of size. The high priority colonies were those that contained the top of 95% of the remaining British or Irish population of one or more species, or the top 90% of the British or Irish population of European Shag.  Regional co-ordinators were asked to provide information on additional sites that may not have been recorded in the SCR Database. They were also asked to survey stretches of coastline on either side of the listed coastal colonies in order to take into account the expansion of colonies and the formation of new ones.  Details of the coverage achieved is given in each of the species accounts on this website.

Count methods

The counting methods implemented in Seabird 2000 were identical to those outlined in the Seabird Monitoring Handbook (Walsh et al. 1995) and to those of Gilbert et al. (1998) for the two species of storm-petrel. A slimmed-down synopsis of the methods detailed in the Handbook was included in the counting instructions provided to all counters.

Assessment of count quality

Table 2 summarises the prescribed seasonal and diurnal timings and units for counting each species. Each species chapter gives an assessment of the overall count quality i.e. the proportion of counts conducted as follows: 1) at the recommended date and time; 2) at the recommended date, but outside the recommended time period; 3) at the recommended date, but time not given; 4. date not given; 5. outside the recommended date and time.

Weather conditions

All surveyors were instructed to record weather conditions on Seabird 2000 recording forms. They were instructed to conduct counts, when possible, during optimum weather conditions. Weather conditions can greatly affect the attendance of some cliff-nesting birds at colonies (e.g. Common Guillemots, Razorbills, Black Guillemots, Northern Fulmars), so counts should be conducted within the conditions specified below to ensure comparability of counts made in different years and at different colonies. The species accounts on this website give an assessment of the overall quality of weather conditions in which counts were conducted i.e. the proportion of counts conducted in either 'ideal' and 'poor' weather conditions. Ideal weather conditions were defined as follows:

1) wind speed equal to or less than Beaufort force 4; 2) sea state less than or equal to force 4; 3) good visibility ; 4) either dry, or if rain is light and/or discontinuous.

Surveyors were instructed to avoid flush counts, or any other disturbance of colonies during wet weather, in order to prevent eggs and chicks becoming irreversibly chilled.

Recording counts

Surveyors were asked to enter their counts on recording forms. The forms consisted of three separate sheets: Sheet A recorded information about the site (eg. site name, six figure OS grid reference, habitat type); Sheet B recorded information about each visit to the site (i.e. date, time, count method used, weather conditions); Sheet C recorded the counts obtained on each visit to the site (i.e. species name, count, count unit, count accuracy). The structure of the forms were designed to be compatible with the new system of data capture designed for Seabird 2000.

Defining sites

The list of sites provided to Regional Co-ordinators at the beginning of the census was based on the definitions given by contributors to the SCR Database, including the previous two national censuses. Over the years, colonies had been counted on the same stretch of coastline by different observers who had divided the coast up quite differently and thus, direct comparisons between different years was often not possible at the scale at which observers had collected the data. In view of this, the Seabird 2000 Steering Group recommended that a priority of Seabird 2000 should be to standardise the division of the coastline of Britain and Ireland and produce a definitive list of colonies to provide a basis for future censuses and monitoring.

Stretches of coastline were divided into 'sites', which were usually based on previous count sections defined in the SCR Database. In the UK, the National Co-ordinator provided counters with 1:10,000 scale maps on which the boundaries of each site were clearly marked. Also marked on the maps of sites in the UK were the boundaries of areas designated for protection (e.g. Sites of Special Scientific Interest, Local and National Nature Reserves, Special Protection Areas), plus boundaries of independent nature reserves (eg. RSPB, local Wildlife Trusts etc.) (see Figure 1). In the Republic of Ireland 1:50,000 Discovery series maps were used together with six inch to the mile or GIS produced maps that were available for some SPAs.

In coastal areas, surveyors were instructed to divide sites into easily defined count sections or 'sub-sites', around 1km in length and provide separate counts for each sub-site. Surveyors were instructed that no sub-site should cross one of the existing boundaries marked on the field maps (eg. SPA or RSPB reserve) and each sub-site should be chosen so that a different observer, visiting the area years later, could identify the section exactly, by using the above 1:10,000 or 1:50,000 maps.  The aim of this mapping approach was to ensure that counts could be ascribed to a specific area along a stretch of coastline. This will be extremely useful to the end-users of the Seabird 2000 database, e.g. the UK Country Conservation Agencies who may only be interested in obtaining data for stretches of coastline which fall within legislative boundaries, such as SPA's. Likewise, local Wildlife Trusts may require information relating solely to one of their nature reserves, which may lie within a larger seabird colony.

The six figure OS grid references of the start and finish point of each sub-site were entered onto sheet A of the recording forms. Inland colonies of gulls, terns and Great Cormorants were defined by a single six figure grid reference denoting the centre of the colony. A separate set of forms was completed for each site. The information collected on each visit to each sub-site was recorded separately on Sheets B and C.

Black Guillemots tend to occur in small groups, scattered along long stretches of coastline.  In view of this, sub-sites of black guillemot surveys were defined on OS 1:50,000 or 1:25,000 scale maps. Sections of surveyed coastline were divided into sub-sections of 1km or less only where Black Guillemots were actually present. Sections of coastline surveyed but not found to contain Black Guillemots were not sub-divided into 1km long sub-sites, and a zero count was entered.

Other information collected

Surveyors were also asked to record the presence of potential predators of seabirds and give details of the evidence of their presence (eg. sightings, scats, prints, regurgitated pellets). Predation is an important factor driving changes in population size and distribution of certain seabirds (see chapter on Causes of Seabird Population Change). It is particularly important to determine the presence of mammalian predators introduced on islands that contain ground nesting seabirds (eg. Atlantic Puffins, storm-petrels, Manx Shearwaters), since these predators can have devastating effects on such populations.

Habitat classifications as used in the BTO's Breeding Bird Survey were used to define habitat type for each sub-site. Up to four habitat codes were permitted per sub-site.

Duplicate counts

Any analysis that summarises count data at a spatial scale must have only one count for a species at a location. In practice, surveyors may revisit a sub-site to recount a species if the previous count was not reliable, due to sub-optimal timing, weather conditions or some other factor. Surveyors were instructed to record on Sheet B insert link whether or not a count for a particular species on a particular visit should be added to previous counts or replace them.

Table 1: Seabird 2000 Steering Group (1998-2002)

Table 2: Summary of prescribed counting periods and units, used to assess count quality

Figure 1: Example of Coastal Site Mapping