Once the completed Seabird 2000 recording forms  had been collated, logged and checked they were then ready for data entry.  From very early on in the project it was decided that the data would be collected and disseminated according to the principles and standards of the National Biodiversity Network (NBN). The NBN is a UK wide initiative to share information about wildlife via the internet (for more information see http://www.nbn.org.uk/). For data entry, the Recorder 2000 software was used and was customized to allow the data to be entered directly from the survey sheets into a relational database. In total, Seabird 2000 surveyors recorded over 25,000 individual species counts, which took the equivalent of one person two years to enter onto the database.

 

Validation and checking of the data was required to ensure that it had been accurately copied from the Seabird 2000 recording forms. The customised 'add-in' for Recorder 2000 provided validation routines during data entry to ensure that, for example, grid references were correct, dates were within the survey period specified etc. These validation routines were able to identify some but not all errors in data entry. Therefore, once the data had been entered and had satisfied the validation routines, data-entry personnel then checked and ticked off the entries presented in two data-trees in Recorder 2000 .  The Location Tree contains information about each site and its constituent sub-sites. The Observations Tree contains information collected on each visit to a sub-site, including records of species present, counts, habitat types and predators present.

 

A frequent source of error in the final analysis of the Seabird 2000 data was caused by duplicate counts of the same species at the same sub-site. The Seabird 2000 recording forms enabled surveyors to enter repeated counts of a species at a sub-site as 'recounts', if they superseded previous counts because they were of better quality. However, some surveyors failed to flag recounts on the forms. Such errors were mostly identified at the data entry stage, but some duplicates were mistakenly entered on the database. All duplicate counts entered on the database for the same species at the same sub-site, that were not flagged as recounts, were individually checked to remove any errors. 

 

Information in the Seabird 2000 database is held in 120 interrelated tables. Once all the data had been entered and was ready for analysis, an automatic routine moved the data into a single spreadsheet. Summaries of the analyses performed on this spreadsheet are given in a separate chapter for each species.

 

 

Within each species' account on this website, the population size and distribution of each species is summarised at the following scales: colony, administrative area and country.

 

 

Each species chapter presents a map of Britain and Ireland showing the position and size of each colony. In most cases, a 'colony' equates to a single Seabird 2000 sub-site. However, to improve clarity, the counts from adjacent sub-sites in areas of high seabird density or in island groups, have been combined and presented as a single colony. Within all species accounts (except, Leach's and European Storm-Petrels, Manx Shearwaters, Arctic and Great Skuas and Black Guillemots) the data for Shetland and Orkney presented on maps showing distribution and abundance have been summarised by the geographical areas shown in Figure 3.  However the insert maps of Shetland and Orkney on the same figures present data at the sub-site/colony level.

 

 

Each chapter presents a table containing the total population size of the particular species in each administrative area within Great Britain, Northern Ireland and the Republic of Ireland, plus totals for each of the Isle of Man and the Channel Islands. The administrative boundaries used in this book to summarise Seabird 2000 data are shown Table 1 and Figure 3. However, these administrative areas no longer used by local and national governments within the UK and were replaced in 1996 by new county and unitary authority boundaries. We considered the old boundaries (Table 1) to be more appropriate for summarising seabird distribution in the UK since the new boundaries are based upon head of population and thus tend to be small in urban areas and large in rural areas. Furthermore, the boundaries used in this book are much more comparable to those used to summarise data from the previous two censuses (Cramp et al., 1974; Lloyd et al., 1991) and still provide a spatial scale that is relevant to conservation GOs and NGOs throughout the UK.

 

The administrative areas of Caithness, Sutherland and Ross & Cromarty in northern Scotland include coastlines adjacent to both the North Atlantic and the North Sea. In order to examine any differences in the seabird populations breeding along the two coastlines, coastal counts from the SCR Census and Seabird 2000 were divided into i) north coast Caithness and east coast Caithness, separated at Duncansby Head Lighthouse (OS grid ref. ND406733); ii) northwest coast Sutherland and east coast Sutherland; iii) west coast Ross & Cromarty and east coast Ross & Cromarty (this division was not possible for Operation Seafarer counts).

 

 

Separate population totals are given for Britain (including Isle of Man and the Channel Isles) and for Ireland (i.e. Northern Ireland and the Republic of Ireland combined). This division is relevant to the frames of the EU Birds Directive (1979) within which, the importance of bird populations are assessed in the contexts of Britain and Ireland.

 

 

 

Most counts were conducted in 1969 and 1970, with some counts included from 1968 and 1971.  Coverage was confined to coastal colonies but was thought to be complete apart from gaps in Foula, Shetland and in Caithness. The survey was initially summarised by Cramp et al. (1974). The data was then entered on to the SCR database and summarised by Lloyd et al. (1991) in order to compare directly with the results of the SCR Census and to take account of changing administrative boundaries between the two censuses. They also applied more consistent criteria to select counts from locations where repeated counts had been conducted. Therefore the totals summarised in Lloyd et al. (1991) for Operation Seafarer  are slightly different from those in Cramp et al. (1974). In this book we have used the totals for Operation Seafarer given in Lloyd et al. (1991). The Operation Seafarer totals for administrative areas and countries are tabulated along-side those from Seabird 2000 in each species account; except Black Guillemots, Storm-petrels and Manx Shearwaters, since census methods for these species were not comparable between the two censuses (see each species account for details).

 

Seabird Colony Register Census (SCR)

 

Most counts were conducted during 1985 and 1987, but counts were used from 1988 if no previous counts were available (see Lloyd et al.1991). Some data on urban gulls, terns, skuas and black guillemots were collected outwith 1985-88 and were included in the results of the SCR Census (Lloyd et al. 1991) and are detailed below. Data from the SCR Database for the years 1985-88 were imported into the Recorder 2000 database so it could be viewed and analysed alongside the Seabird 2000 dataset and others (see below). The SCR data contained duplicate counts from the same sites but conducted in different years. A single count was selected for each site according to the criteria used by Lloyd et al. (1991) and listed in Table 2.

 

SCR Census totals for administrative areas and countries are tabulated alongside those from Operation Seafarer and Seabird 2000 in each species chapter; except storm-petrels, Manx Shearwaters and Black Guillemots in the Republic of Ireland, since SCR census methods for these species were not comparable with those of Seabird 2000 (see chapter on Census Methods).. As in Operation Seafarer, the SCR Census did not attempt to survey storm-petrels and Manx Shearwaters. Some colonies were visited in one or both of the censuses and best guesses of the size of breeding populations, usually of orders of magnitude, were made. These were listed in Lloyd et al. (1991) along with a list of other known colonies based on information given in Cramp et al. (1974) gathered from visits to colonies during Operation Seafarer and from historical records.

 

Apart from Black Guillemots and urban gulls (see below), the task of completing a comprehensive comparison of counts from the SCR and Seabird 2000 censuses for every single colony of every species in Britain and Ireland proved beyond the scope and timetable for producing this book. During the SCR Census around 3,300 coastal sites and 700 inland sites in Britain and Ireland were recorded on the SCR Database. The main problems with comparing these to sites surveyed during Seabird 2000 were three-fold: 1) Coastal sites in the SCR were of widely varying size from stretches of coastline of less than 1km to single large islands or even groups of islands. 2) SCR sites did not necessarily take account of important boundaries such as SSSI's or SPA's (first designated only in 1982). 3) 77% of coastal sites were defined by only by a single, central OS grid reference, so it was impossible to assess where the site actually started and ended. 4) Some stretches of coast, particularly those containing large numbers of seabirds were counted in more than one year during the SCR Census, but changes in surveyor often mean that different areas were surveyed in each year.

 

Hence presentation of change at the colony level was restricted to the largest and most important colonies of each species. For most species, these top colonies were selected on the basis of SCR Census counts and colony definitions. For a given species, the counts for each colony from the SCR were listed separately for Britain and for Ireland, in descending order of size, and those colonies containing a certain percentage (see individual species' accounts) of the British or All-Ireland population were selected as 'top colonies'. For those species  that are aggregated in fewer colonies (e.g. Atlantic Puffin), a higher percentage of the respective populations were represented by the 'top colonies' than for those species that have a more dispersed distribution (e.g. European Shag). Colonies of. gulls tend to be more mobile between years, with some colonies disappearing while new ones form. In order to ensure that important new gull colonies were selected as top sites, large colonies of gulls were selected not just on the basis of their size during the SCR, but were also of their size during Seabird 2000. Top sites of Great Cormorants were selected in a similar way. Within each species chapter, the size of each top colony during Seabird 2000 and the SCR Census are tabulated and their location marked on a map, which also denotes the percentage change in numbers between the two censuses.

 

 

No inland colonies were surveyed during Operation Seafarer, all totals listed alongside SCR and Seabird 2000 counts in this book are for coastal colonies only. Counts from inland colonies were contained in the SCR data imported into the Recorder 2000 database. Coverage of inland sites during the SCR census was incomplete and so was not included in the administrative area and country totals reported in Lloyd et al. (1991). Inland sites were defined by Lloyd et al. (1991) as being within a 1km OS grid square that did not contain any coastline, except in Orkney and Shetland, where all sites were considered as coastal. However, it was not stated in the SCR database whether or not a site was considered inland or coastal. Thus, in order to separate out coastal sites from inland sites in both the Seabird 2000 and SCR data, the following criteria were applied. Using GIS, sites were defined as coastal if they were situated within 5km of the mean high-water mark (OS Boundary-Line ™ © crown copyright). This 5km coastal buffer zone was based on a frequency histogram of the number of colonies and their distance from the coast (Figure 4). For distances greater than 5km, there was no discernable pattern in the frequency distribution, suggesting that there is little if any coastal influence on the position of colonies situated more than 5km from the coast and that these are effectively 'inland'. The 5km buffer was not applied to colonies in Orkney, Shetland and the Western Isles, where all colonies were considered coastal.

International Context

 

Within each species' account there is a table giving a breakdown of the international distribution of the species, based on the latest population estimates for each country or region in which the species is known to breed. Lloyd et al. (1991) presented international population estimates, which they used to put the results of the SCR Census into a global context. Lloyd et al. (1991) drew mainly on population estimates published by the International Council for Bird Preservation (Croxall et al. 1984, Croxall 1991). While no more recent figures exist for some areas, notably parts of northeast Asia, there have been more recent initiatives to survey and collate estimates of seabird populations in other regions. For example, the European Bird Census Council (Birdlife International/EBCC 2000, Hagemeijer & Blair 1997); and reviews of North American population estimates published in the Birds of North America, by the American Ornithologist's Union.

 

Where possible, population estimates are broken down by sub-species. The population estimates from Seabird 2000 for Britain (including Isle of Man and the Channel Islands) and Ireland are expressed separately as a percentage of the world population and of the relevant biogeographic population. The biogeographic population is used to assess the international significance of colonies and is one of the criteria used to designate sites as Special Protection Areas under Article 4 of the EU Birds Directive (1979). The separate contexts of Britain and Ireland arise from the fact that these are the relevant frames of reference under the Directive. The definition of the biogeographic population of each species is prescribed in Stroud et al. (2000). 

 

In estimating the sizes of the global and biogeographic populations of breeding seabirds a minimum and a maximum estimate were derived from the best information available from each individual country in which the species breeds. The former is simply the total of the minimum estimates for each country and the latter the total of the maximum estimates. In cases where the published population estimate is not given as a range then clearly the (single) estimate was used in deriving the relevant international total estimate. No rounding conventions were adopted as many of the published minimum and maximum estimates for individual countries have already been rounded; while some of these may be of specious accuracy, to apply a further, arbitrary rounding convention would compound any error associated with them.  However, and again to avoid seeming accurate in the absence of it, the total minimum and maximum global and biogeographic estimates were rounded to the next lowest order of magnitude (Table 3).

 

In determining the biogeographical importance of breeding seabird populations in Britain and Ireland, again a minimum and a maximum value were derived. The former was computed by dividing the minimum total British or Irish population estimated from the present survey by the maximum biogeographical estimate, and the latter was derived by dividing the maximum British or Irish estimate by the minimum biogeographical estimate. The global importance of British and Irish populations were assessed in similar fashion. Minimum and maximum values for biogeographical and global importance were rounded to only one decimal place, again to avoid spurious accuracy.