Puffin - 2003-07

Common offshore around western Pembrokeshire between late March and late July each year, but rarely seen outside this period, they congregate in large rafts close inshore near their breeding colonies in the spring and early summer months and socialise around their breeding colonies each evening.

Puffins are mostly confined to offshore islands free of predators, particularly rats. They nest in burrows on steep slopes in sometimes very dense colonies.  The principal colonies are on Skomer Island, including Middleholm, and Skokholm with small populations on the Bishops near Ramsey, on St Margaret’s Island and at Stackpole Head where they nest in sheer cliff sites.

Numbers in the past were clearly very much higher than they are now and for example it was estimated that 500,000 to 700,000 pairs nested on Grassholm in the  1890’s,  a figure disputed by modern analysis which suggested that around 200,000 pairs would be more realistic.  Whatever the actual figure, the density was clearly much higher than in today’s colonies and on Grassholm there are still remnants of the collapsed burrows testifying to this. This colony was more or less deserted during the 1920’s and it has been suggested that the birds relocated to Skomer and Skokholm, although it is not clear how that assumption was made. There were also apparently big colonies on Caldey and Ramsey in historic times, before rats devastated them.

Mathew (1894) made the observation that they were undoubtedly the most common bird in Pembrokeshire with huge colonies on Skomer stating “ there is scarcely a yard of ground free of them” and Lockley et al (1949) thought that around 50,000 pairs were breeding on Skomer and 20,000 on Skokholm. There were clearly many more birds present than there are today. 

The population estimates which have been made since the 1950’s on both Skomer and Skokholm have been consistent in suggesting that around 9,000 to 10,500 pairs may breed there.  Estimates are however subject to substantial inaccuracies because it is not possible to count burrows, as both rabbits and shearwaters breed on both islands in abundance, and estimates have tended to be done by extrapolation of detailed studies in one small part of the islands.

Consequently, in an attempt to provide greater consistency in counting and to provide comparative long term data of population trends, all counts of birds since the 1980’s have concentrated on the maximum attendance by adult birds in the spring. The spring counts in 2003-2007 of around 10,800 individual birds on Skomer and 4,800 on Skokholm  suggests a rather similar breeding population to the 1985-87 Seabird Colony Survey estimate of a total breeding population of  about 10,600 pairs  on the two islands.

Stephen Sutcliffe

Fieldwork 2003-07 (based on 490 tetrads)

Red = breeding confirmed = 4

Orange = breeding probable = 2

Yellow = breeding possible = 0

Total tetrads in which registered = 6 (1.2%)

Razorbill - 2003-07

Razorbills nest mostly in loose colonies scattered around broken cliffs in cracks and holes and less commonly on small ledges. In some places just a few birds might breed in suitable sites, so their distribution is slightly wider around the county than the gregarious Guillemot.

 There are many fewer Razorbills than either Guillemots or Puffins but they occupy many more breeding sites around the coastline, especially along the south coast of the county and between Dinas Head and Newport along the north coast.  Their distribution appears to have hardly changed during the last century.

 Razorbills are not easy to count as they nest in scattered, often well concealed, small groups.  In the late 1970s (but for some reason not on Skokholm until the late 1980) the count methodology was changed.  Now all individuals attending the colony (i.e. observed on land), are counted, rather than using the older method of estimating “apparently occupied sites”.   Accordingly population comparisons have since been made using a standard conversion factor of 0.67 across the UK, i.e. for every three birds counted there are two nest sites on average. Even so, because of the problems of seeing birds easily at some sites (where sometimes the birds are very evident and sometimes they hide very effectively, which is possibly weather related), and because of the simple variation in attendance from day to day, the census counts are at best reasonable estimates. 

 Because of these census variables, data needs to be accumulated over a long period of time to assess population changes.  Information for Razorbills at their main breding sites has been collected for more than 40 years. The counts during the Seabird Colony register in 1985 – 1988 found a population of around 6,600 individuals in the county, half of them on Skomer. Seabird 2000 counts between 1998 and 2002 showed a significant increase to c. 9,000 birds during the period 2003–07, with almost 5,000 of them on Skomer.  This is the highest known population level.

Steve Sutcliffe

 

Fieldwork 2003-07 (based on 490 tetrads)

Red = breeding confirmed = 27

Orange = breeding probable = 1

Yellow = possible breeding = not included

Total tetrads in which registered = 28 (5.7%)

Guillemot - 2003-07

Guillemots nest in dense, noisy colonies with many birds crowding together.  Mainly nesting on ledges on near vertical cliffs, they are also present on the top of sea stacks in a few sites.

Guillemots are found in very large numbers on the south coast at Castlemartin, especially at Elegug Stacks, where the well known colonies provide a thrilling and easily observed wildlife spectacle. Skomer has the largest island colony, but there are quite significant colonies on St Margaret’s Island, on Skokholm, and on Ramsey.  A small number breed on Grassholm and on the north coast of the county there are scattered colonies of a few hundred birds.

The overall population has been increasing steadily for 40 years and although there have been minor “blips” in the rate of increase this trend is echoed at all colonies.  The main study colonies and the detailed whole island counts on Skomer, suggest an average annual rate of increase of 5.7%, the colony increasing from 2,400 birds in 1970 to 17,700 in 2008.  In the period between the two atlases of 1984–88 and 2003-07, the overall county population grew from around 16,000 to over 30,000 birds.

The number of Guillemots is currently at its highest level since counting seabirds became a regular  annual priority for conservationists and ornithologists, but there is some evidence from comments made by Lockley et al (1949) that pre war populations were higher, perhaps much higher, on the basis of some photographs of colonies. It was surmised that during the war years the high level of marine pollution caused the deaths of many thousands of Guillemots.

Guillemots are very vulnerable to marine pollution incidents as they spend most of their time "rafting" (floating on the surface) and shallow-diving. The Sea Empress incident at the entrance to Milford Haven in February 1996, followed by the Erika incident off the French coast in January 2000, killed many birds from the Pembrokeshire colonies. The Sea Empress spill killed mainly adults and there was a subsequent 60% decline in breeding numbers on St Margaret’s Island. The effects at other coastal and the major island colonies was less clear but subsequent research has shown that such incidents do manifest themselves in reduced survival rates in the years following the incident.  However, because there are so many immature birds trying to establish themselves in the colonies the effect has been masked.  (Votier et al 2005). The Erika incident mainly affected immature birds less than 6 years old and there was no discernable subsequent effect on numbers at any of the colonies.

Steve Sutcliffe

Fieldwork 2003-07 (based on 490 tetrads)

Red = breeding confirmed = 17

Yellow = breeding possible = not included

Total tetrads in which registered = 17 (3.5%

Great Black-backed Gull - 2003-07

Great black-backs are top predators on the seabird islands, taking eggs and chicks of other species, and are regular predators of Puffins, Manx Shearwaters and Rabbits. They nest in single pairs or loose colonies, the former preferring prominent places on the top of rocky outcrops, while those in colonies will select dense vegetation on the top of the islands. Over 90% of the Pembrokeshire pairs breed on the offshore islands with rest as scattered individual pairs around most of the county coastline.

Numbers of these gulls have fluctuated widely.  Apparently very small numbers were present in the county in the late 19th Century (Harrison & Hurrell, 1933) but by 1949 they could be found all around the coastline and especially on the islands, where Lockley et al (1949) found 310 pairs on Skomer, Middleholm, Skokholm, Grassholm and St Margaret’s combined. Davis (1958) reported between 490 and 520 pairs in the county following an all Wales survey and they continued to increase with 542 pairs recorded in 1969’s Operation Seafarer (Cramp et al 1974). 

Control measures in the 1960s and 1970s on Skomer and Skokholm, followed by a botulism outbreak in the early 1980’s, reduced numbers to a low of about 140 pairs in the mid-1980s.  There has been a steady but slow recovery since then and the population had risen to around 300 pairs in 2000 but has subsequently remained stable, except on Skomer where it has declined by about a third. The mean breeding success recorded on the Pembrokeshire islands in the ten years to 2005 was 1.15 to 1.29 chicks per pair, by far the highest success rate in the UK during this period. In consequence there has been a steady recovery since the 1980’s and the population in the county had risen to around 300 pairs in 2000 and has subsequently remained stable at around this number.

Steve Sutcliffe

 

Fieldwork 2003-07 (based on 490 tetrads)

Red = breeding confirmed = 21

Orange = breeding probable = 9

Yellow = breeding possible = not included

Total tetrads in which found = 30 (6.1%)

Lesser Black-backed Gull - 2003-07

Formerly the main colonies were confined to the islands of Skomer and Skokholm with a few on Caldey, St Margaret’s and Ramsey and occasionally on the mainland cliffs. Mathew (1894) knew of colonies of 20 – 30 pairs but there were around 2,000 by the middle of the century (Lockley et al 1949). The population increased dramatically from the early 1950’s, to peak at around 20,000 pairs on Skomer and 5,000 on Skokholm in the mid-1980’s. 

This increase was fuelled largely by a ready source of food in the form small fish that were a discarded by-catch of the scampi fishery in the nearby Smalls grounds.  New legislation introduced in the mid-1980’s changed the mesh size of the nets so that fewer small fish were caught and this food supply was lost to the gulls in a very short period. The result has been almost very low breeding success, including total failure in some years, from 1989 to the present day and a resultant steady decline in the number of gulls. Numbers of Lesser Black-backed Gulls breeding on the mainland coast have increased in several places in the last 20 years, with a gradual spread in distribution during this time, but the population here is small.

Over 95% of Lesser Black-backed Gulls in the county breed on Skomer, Skokholm and Caldey Islands.  They are found in large colonies wherever there is good cover for chicks and prefer a bracken/gorse/thick grass habitat.  In 2007 the county population was around half of the peak at c. 13,000 pairs and continuing to decline because there are so few immature birds entering the breeding population.  

Steve Sutcliffe

Fieldwork 2003-07 (based on 490 tetrads)

Red = breeding confirmed = 30

Orange = breeding probable = 6

Yellow = breeding possible = not included

Total tetrads in which registered = 36 (7.3%)

Kittiwake - 2003-07

Kittiwakes nest in dense colonies on ledges of vertical cliffs.  Although now almost totally confined to the islands of Skomer, St Margaret’s, Grassholm and Ramsey, they were formerly very common coastal breeding gulls, especially along the Castlemartin coastline to Stackpole Head. Kittiwakes have never nested on Skokholm.

These changes in the breeding colony sizes and distribution have been dramatic and largely remain unexplained, for whilst the Skomer population has remained relatively stable at around 2,200 pairs, other colonies have disappeared completely. In the period following the Seabird survey of 1985-87  the thriving colony on St Margaret’s Island reduced from well over 300 pairs to just 6 in 2000, and then increased again to over 300 in 2009.  This quite rapid change must include some birds which have moved from other breeding sites. Along the Castlemartin peninsula a population of around 700 pairs during 1984-88 had declined to about 30 pairs by 2007.

Overall the Kittiwake population has declined significantly in the county in recent times.  Between 1969-1970 and 1985-1987 surveys the county population increased from 3,037 to 3,935 pairs but had subsequently declined to around 3,100 pairs in 2000 and had probably a similar population total in the period 2003 to 2007.

The breeding success on Skomer is variable with productivity ranging between 0.30 and 1.01 chicks per pair but generally near to 0.5.  It is a low level of success and there is a “feeling” that, as at many other colonies around Britain in the first decade of the 2000’s, Pembrokeshire colonies have suffered breeding failure and a rapidly declining population is just about “hanging on”.

Steve Sutcliffe

 

Fieldwork 2003-07 (based on 490 tetrads)

Red = breeding confirmed = 6 (1.2%)

Shag - 2003-07

Shags are essentially inshore feeding birds that prefer rocky coastlines.  Their nests are well-distributed around Pembrokeshire, with concentrations on the offshore islands. 

Despite being relatively common around the coastline, breeding Shags are not very easy to count accurately because they nest in deep crevices and so can be almost invisible.  Also some sites are not commonly visited or counted. The population has fluctuated considerably in the last fifty years with a low point during the late 1980’s and early 1990’s. The reasons for that decline are unknown but it was seen in all areas of the county. The Sea Empress oil spill in 1996 caused the death of a small number of Shags, mainly around the entrance to Milford Haven and around the south coast of the county.  The counts in 2008 were the highest since the early 1970’s and reflect high breeding success during recent years.

Only about 200 pairs currently breed in the county. The overall impression is, however, of a small but vibrant population which is doing well in the first decade of this century.

Steve Sutcliffe.

 

Fieldwork 2003-07 (based on 490 tetrads)

Red = breeding confirmed = 26

Orange = breeding probable = 3

Yellow = breeding possible

Total tetrads in which registered = 29 (5.9%)

Cormorant - 2003-07

The largest colony of Cormorants is on St Margaret’s Island and this is one of the best studied in the UK, with almost continuous census data for 50 years and over 5,000 chicks ringed since the mid 1960’s.  There are  smaller colonies on Thorne Island (it relocated from Sheep Island in the early 1990’s), on the Mew Stone, Skomer, at times on Stack Rocks off Little Haven, on the Green Scar at Solva, around Dinas Island and at Cemaes Head.  Occasionally pairs might nest at scattered mainland sites.  There are many coastal drying out sites (including jetties etc in the Milford Haven Waterway) where sometimes good numbers of birds can be seen.

The Cormorant population fluctuates from year to year, as probably in poor springs some birds fail to breed. Overall counts show there was a tendency to a decrease during the 1990’s but a modest increase again in the early 2000’s.  The total population in the county in 1984-88 survey was around 300 pairs and a census following the Sea Empress oil spill (1996) found 360 pairs, with 180 of these on St Margaret’s Island.  The population during the most recent years appears to have been stable and may have increased slightly.

The highest county population was almost certainly in the early 1970’s, when around 450 - 500 pairs probably bred, 330 of them on St Margaret’s, but the reasons for the subsequent decrease are poorly understood.  Breeding success varies from year to year but there are few years of substantial failure. Ringing has shown that considerable numbers are shot on rivers or killed in coastal fishing nets during the winter period. It seems likely that this is a major cause of population changes.

Steve Sutcliffe

 

Fieldwork 2003-07 (based on 490 tetrads)

Red = breeding confirmed = 13

Orange = breeding probable = 2

Yellow = breeding possible

Total tetrads in which registered = 15 (3.1%)

Storm Petrel - 2003-07

Storm Petrels breed only on the offshore islands, preferring rocky scree slopes and boulder areas where they can find small crevices and burrows to nest in. Sometimes you can smell their slightly fishy odour near their nest chamber but the best clue that they are present is their soft purring and clicky song. On Skokholm they nest in the herringbone style walls, in the boulder scree in the bays and especially in the “quarry” – a large boulder rockfall area on the north western corner of the island.

They are a very difficult species to census and a huge amount of effort has gone into establishing a reliable method of gaining consistent data. Nationally the use of a song playback system, listening for responses from the birds to taped calls played to the burrows, has now been adopted but there are still questions over interpretation of the results.  In consequence the historical data, (pre 1980’s) may not be directly comparable recent census results. Until a consistent methodology is applied, the information available is a “best guess” at a particular time.

In the first decade of the 2000’s,  the best information available suggests populations of up to 100 pairs on Ramsey and North Bishop, perhaps around 150 pairs on Skomer and 2,000 pairs on Skokholm. The estimates on Skomer suggest a recent increase although this may mainly relate to new sites being located.  On Skokholm the population has certainly declined as many known nest sites in the wall systems have been deserted. The latest census attempt in 2003 estimated the population to be only 1,011 apparently occupied sites on the whole island, compared with varying estimates of 5,000 to 7,000 in 1969, 3,000 to 4,000 in 1995 and in 2001 about 2,000 pairs.

There have been no recent studies of Storm Petrels on the Pembrokeshire islands to understand their survival rates or their breeding success, so the reasons for the declines are unknown.

Because of the importance of the Storm Petrel population on the Pembrokeshire Islands (in European and indeed World terms) these island have been designated a Special Protection Area for them (under the EC Directive on the conservation of wild birds (79/409/EEC)

Steve Sutcliffe.

Fieldwork 2003-07

Red = breeding confirmed = 3

Total tetrads in which registered = 3

Manx Shearwater - 2003-07

The Manx Shearwater colonies of the Pembrokeshire islands total around 50% of the world population, with around 120,000 pairs on Skomer, 45,000 pairs on Skokholm and 4,000 pairs on Ramsey. Because of the importance of this Shearwater (in European and indeed World terms) these island have been designated a Special Protection Area for them (under the EC Directive on the conservation of wild birds (79/409/EEC)

They are very difficult to census accurately and methods have changed over the years, from simple estimates based on counts of birds seen at night, estimates of burrow densities through capture/recapture methods derived from known numbers of ringed birds, through to the present day estimates based on counts of burrows and the responses of adult birds to tape recordings of their calls. Each method has been refined and compared but throughout there has been strong evidence of a continuing slow increase in the populations in the last fifty years, perhaps with an indication of a reverse of this in the early 2000’s.

On Ramsey Island the small population (perhaps 1,500 – 2,000 pairs) has risen rapidly since the Brown Rats were removed in 2000. In 2008 the population was estimated as 4,000 pairs.

Steve Sutcliffe.

Fieldwork 2003-07

Red = breeding confirmed = 3

Total tetrads in which registered = 3

Fulmar - 2003-07

Fulmars are now one of the commonest breeding seabirds around the coast of the UK but only a century ago were virtually unheard of as far south as Pembrokeshire. Their spectacular spread around the coast of the British Isles in the last century  and a half is one of the best documented of any bird species. The Seabird 2000 survey estimated that there were just under 538,000 apparently occupied sites in Britain and Ireland.

In Pembrokeshire Fulmars can now be found virtually everywhere around the coastline. They nest on sheer cliff faces on small ledges in small groups and sometimes as individual pairs.  Near Saundersfoot they nest behind the tangled roots of cliff edge trees, on the predator free islands sometimes on wide accessible ledges but always they try to find as much height as they can.  Only where the cliffs are comprised of low sloping sandstone or are exposed fully to the glare of the sun are they absent. They are fairly easy to count as they occupy their sites for many months each year, and are only totally absent in September and October when they are at sea moulting.

The number of occupied sites increased steadily in the county from the first known breeding site at Flimston in 1940, to what seems to be a discernable peak of near to 2,500 sites in the mid to late 1990’s.  Annual counts are very variable as many of the sites are occupied by non breeding or prospecting pairs.

Between the counts from 1985-88 and the Seabird 2000 counts, Fulmar numbers in the county jumped from 1,409 sites to almost exactly 2,500 sites. However these totals  mask the changes from 1996, when detailed annual counts on Ramsey, Skomer, Skokholm, Castlemartin, St Margaret’s and Caldey, suggest that the trend in subsequent years has been at best stability and more probably the population is showing a tendency to a steady decline. The distribution maps also show a decline in tetrad occupancy

Steve Sutcliffe

Fieldwork 2003-07 (based on 490 tetrads)

Red = breeding confirmed = 61

Orange = breeding probable = 6

Yellow = breeding possible

Total tetrads in which registered = 67 (13.7%)

Pembrokeshire Breeding Bird Atlas 2003-07

The Pembrokeshire Breeding Bird Atlas 2003-07 provides an update to the 1994 Birds of Pembrokeshire, covering the breeding species.

The Atlas of Breeding Birds of Pembrokeshire 2003-2007 is available as a print-on-demand publication by clicking here

THE MAKING OF THE ATLAS

Introduction

The county of Pembrokeshire is some 158,000 hectares (613 square miles) in extent, with its southern, western and northern boundaries all coastal. The climate is maritime, with strong oceanic influences found also in the type and distribution of fauna and flora.

A review of all the species of birds recorded in the county in all seasons is to be found in the "Birds of Pembrokeshire" by Jack Donovan and Graham Rees, published in 1994 and including a breeding bird atlas based on fieldwork done in 1884-88. This important publication provided a baseline on which to compare mapped distributions and population estimates of different species within the county.

The populations of any animals that can walk, fly, swim or crawl will change in numbers and distribution over a period of time, so that any atlas will become out of date over a period of years, decades or centuries. 

Biodiversity is the buzzword, and the need to know what species are where has become important, indeed a requirement, for planning and conservation purposes.

Trends towards milder and windier winters, and towards cloudier, wetter summers, may herald longer term climate change.  But there is already anecdotal evidence, as well as indications from country-wide surveys, that there are other changes happening in the natural world.

At a local level, it was decided that the "Birds of Pembrokeshire" should be updated.  This breeding birds atlas is first stage in that process.  It does, indeed, show that there have been changes, both losses and gains, within the 15 to 20 years between the two fieldwork periods

Methodology

On the 2nd September 2002 the Pembrokeshire Bird Group convened a meeting at “The Patch”, Furzey Park, Haverfordwest, to discuss the desirability of producing a new avifauna for Pembrokeshire.  All interested parties were invited to attend and those interested but unable to attend were encouraged to communicate their views.

The meeting agreed to go ahead with such a project and that it should encompass a breeding birds survey using a tetrad (2km x 2km) grid. To this end an Avifauna Committee was elected, comprised of Graham Rees (chair), Annie and Bob Haycock, Jane Hodges, Trevor Price, and Mike Young-Powell.

At their first meeting, the committee decided that the breeding survey should be the first aspect to be addressed. It was to run from 2003 to 2007 and to take the same form as the 1984-88 survey so that the two would be directly comparable.  The 1984-88 survey was the first in the county to use as fine a scale as the tetrad grid.  It was also the first attempt made to estimate the size of the breeding population of each species in the county.

Recording forms and accompanying instructions (see appendix IV) were printed and distributed in time for field work to begin in 2003. 

Data from completed recording forms, representing some 30,000 records from 490 tetrads, were entered on computer by a small team comprising John and Marion Best, Annie and Bob Haycock, Fiona and Trevor Price, using the computer software package MapMate. A summary of the records from the 1984-88 survey was similarly entered, so that comparable maps could be generated.

The methods used for atlas fieldwork followed those of The Atlas of Breeding Birds in Britain and Ireland (Sharrock 1976) but recorded at the tetrad level rather than hectad (10km x 10km) level. For our local atlas however, adjustments have been made in as much as the islands of Grassholm, Caldey, Skokholm and Ramsey have been regarded as occupying one tetrad each, while Skomer and Middleholm have been treated as a composite tetrad. As in that first national atlas, the final maps are populated with small dots to indicate that a species was present during the breeding season, medium-sized dots to show that it probably bred, and large dots to show that breeding was confirmed, as shown in figure 1:

The example map above shows the data entry record card in MapMate that helps generate distribution maps. Records entered this way gradually build up a growing series of distribution dots on a map, Each record entered represents a breeding registration for a particular species in a particular tetrad. The species in this case is House Martin and the tetrad is SN04L (SN0442 Trwyn y Bwa).

The more visits that are made to a particular tetrad, the more information there is behind each dot providing an eventual final outcome - in this case the largest dot showing confirmed breeding.  In other tetrads medium dots represent probable breeding (e.g. bird showing territorial behaviour or visiting potential nest site, and probably went on to breed successfully) while the small dots show that the species was present in a tetrad at some time during the survey, but no evidence was found to indicate that it was doing more than feeding or resting there.

Another team of volunteers, comprising Graham Rees, Bob Haycock, Jane Hodges, Steve Sutcliffe, Paddy Jenks and Richard Dobbins wrote the individual species accounts.  Annie Haycock then assembled the maps and texts into this volume.

The map below shows the tetrads covered in the 1984-88 atlas (in yellow) and 12 extra tetrads (blue dots) covered during the 2003-07 period.  These were all peripheral tetrads, either coastal or along the county boundary, and including only a small amount of land.

Population estimates

The 2003-07 Pembrokeshire breeding survey was deliberately undertaken using the same methodology as the previous 1984-88 survey, so that the two were directly comparable. It should, therefore prove useful in assessing biodiversity, species distribution changes and perhaps in evaluating the effects of changes in weather patterns over the period.

However, there are limitations to the information gathered by this kind of survey.  It is basically a presence or absence survey, with some additional value in terms of proving whether or not species was breeding throughout the area.  Nonetheless, it does give a reasonable indication of the spatial distribution of each species, and whether a species is localised, is widespread occurring in most tetrads, or is widespread but scattered.

The results take no account of time spent in the field in each tetrad, or of observer bias or competence. Figure 2 shows the number of species recorded in each tetrad in each atlas period (all species are included, whether or not they showed signs of breeding).  In the later period there appears to have been a shift away from the north, with more species recorded in the south-east and on the St. David's peninsula.  While some of this shift may be genuine, some of it is likely to be observer bias as people inevitably record closer to their home areas unless directed to do otherwise.

People were not asked to count birds whilst recording in tetrads during the main 2003-07 survey.  The limited population data collected as part of the BAP breeding survey applied to nature reserves and other special areas, and so was not directly applicable to the county as a whole. 

For the 1984-88 atlas, population estimates were compiled from survey results combined with extensive personal experience (of the authors) within the county (Donovan & Rees 1994). These estimates provided a baseline used to inform population estimates for the 2003-07 atlas. For example, an increase in the number of tetrads recorded for a species was assumed to have the same proportional increase in population. For a few species, e.g. Yellowhammer, anecdotal evidence indicated that has there has been a thinning out of the population, while for others, such as Chough, long term surveillance of nest sites showed there has been an increase in the breeding density, at least in some areas.  This approach therefore has limitations.

It will be noticed in the accounts for many species, that the results of the 1988-91 National Atlas  (Gibbons et al) have been used in testing the original estimates made following the pioneering 1984-88 local atlas.  For a few species, this has resulted in a revision of the original local population estimate and is explained in the individual species text.

The BTO/JNCC/RSPB Breeding Bird Survey (BBS), which came into being in 1994, also provides a useful tool for estimating populations. This survey is carried out nationally, and provides indices of the populations of birds in Britain in summer.  These indices are based on both the changes in numbers of each species counted along two parallel 1km transects in a 1km square, and on changes in the percentage of squares in which they occur. 

For most species in this atlas, a new population size has been estimated, based on a combination of data from the 1984-88 fieldwork, the 2003-07 distribution maps, the results from the 1988-91 National Atlas, and the BBS indices for Wales.  The current National Atlas (fieldwork in progress at the time of this publication), backed by extensive information from other bird research and survey, will undoubtedly become useful in further refining these population estimates in due course. This new information will be considered in the more comprehensive Pembrokeshire Avifauna.

For other species, notably seabirds, chough and peregrine, more specific data are collected annually and this allows a more detailed assessment of population size to be made. Seabirds on the Islands and the Castlemartin coast are counted annually for the Seabird Monitoring Programme administered by the JNCC (www.jncc.gov.uk/page-1550).  For all these species, there is more discussion about the population changes in the species accounts.

Density of recorded species in each atlas period

1984-88 density map. The smallest dots represent 2 species, the largest represent 86 species.  The counts include non-breeding species.

2003-07 density map. The smallest dots represent 2 species, the largest represent 81 species. The counts include non-breeding species.

The average number of species per tetrad is 35.

INTERPRETING THE SPECIES ACCOUNTS

For those species of particular conservation concern, the designation of red or amber-listing, or UK BAP or LBAP is given in the title line.  Further information about these designations is given in appendix III.

A brief introduction is given for each species, followed by comments about changes in distribution and population, and a discussion of problems with calculating such changes.

Maps are not reproduced for all species, where the breeding distribution is so limited that it can readily be expressed in the text, e.g. Gannet. For some scarcer and legally protected (Schedule I of the Wildlife and Countryside Act) species, maps are provided at a 10 kilometre square level.

For the majority of species, a table shows the total number of tetrads in which the species was found, plus the numbers of tetrads that registered “confirmed”, “probable” or “possible” breeding during each atlas period.  Note that for some species, for example rooks and other colonial nesting species, birds foraging in fields or in flight do not give any indication of the location of their nest sites, which may be some distance away, and therefore such "possible" breeding records have been excluded.  For other species, for example skylark, the distribution was based largely on birds singing on more than one date in the same place to show they are holding territory. In these instances the maps show mostly probable breeding.