Scottish Marine and Freshwater Science Volume 5 Number 9: Strategic Surveys of Seabirds off the West Coast of Lewis to Determine Use of Seaspace in Areas of Potential Marine Renewable Energy Developments

4. Results

Desk study

4.1 The only published surveys of seabirds to have taken place to the west of Lewis are reported in Webb et al. 1990. These surveys consisted of boat-based surveys from chartered boats and ships of opportunity, visual aerial surveys at 4 and 9 km from the shore and a series of winter point counts in Loch Roag during the winter. Surveys mainly took place between 1986 and 1989. These showed the area to be of importance for divers (mainly in the winter, especially in Loch Roag); fulmars, European storm-petrels (around the Butt of Lewis in summer); gannets (mainly around the Butt of Lewis in summer); cormorants and shags (especially in Loch Roag); eiders Somateria mollissima and red-breasted mergansers Mergus serrator (especially in Loch Roag); lesser black-backed gulls, herring gulls and great black-backed gulls (mostly in offshore areas and around the Butt of Lewis); kittiwake (mostly around the Butt of Lewis); guillemots, razorbills and puffins (mainly around the Butt of Lewis, but concentrations of puffins around Gallan Head ca. 20km south-west of Loch Roag); and black guillemots (Loch Roag and adjacent open coastline and around the Butt of Lewis). They identified the area around the Butt of Lewis to be of particular importance for moulting auks and other seabirds during the late summer and early autumn.

4.2 WWT Consulting undertook a visual aerial survey covering an area around North and South Uist on 30 ^th July 2007. During the survey a wide range of seabird species were seen including; Arctic skuas, great skuas Stercorarius skua, black guillemots, fulmars, gannets, Manx shearwaters Puffinus puffinus, kittiwakes and auk species. The majority of birds were seen in small numbers (less than 50) with only gannets recorded in large numbers (950).

4.3 Annual aerial surveys have been carried out by JNCC between March 2003 and March 2007 focusing on the inshore areas on the west coast of the Outer Hebrides, including the west coast of Harris, North Uist, Benbecula, South Uist and Barra and thus south of the study area. These surveys were targeted at seaducks and divers and have revealed regular wintering populations of great northern divers, eiders, common scoters Melanitta nigra, red-breasted mergansers and long-tailed ducks Clangula hyemalis ( JNCC 2010b).

Digital aerial survey results

Survey dates

4.4 Table 1 shows the dates of digital video aerial surveys and their associated survey conditions. The May 2012 and February 2013 surveys were due to take place on the 2 ^nd of those months, however a technical fault with the aircraft on both dates prevented this so they were postponed to the next suitable weather opportunities on 29 ^th May and 12 ^th February respectively. The ground counters were already in place or had started counts on both occasions so these were completed in the absence of synchronised aerial surveys.

Table 1 - Survey dates, times and conditions

Date of survey	Start time	Time on survey	Wind Beaufort scale	Cloud	Sea state
19th April 2012	11:50	3hr 3min	Force 3-4 easterly	Scattered	3-4
29th May 2012	14:46	2hr 59min	Force 5 north easterly	No cloud	3
18th June 2012	11:21	2hr 45min	Force 1 northerly	Scattered	1
26th July 2012	10:41	2hr 39min	Force 2 westerly	Overcast	2
22 nd September 2012	11:39	2hr 43min	Force 3 Westerly	Overcast	2 - 3
17 th December 2012	12:04	2hr 35min	Force 3 South Easterly	Overcast	2- 3
12th February 2013	12:10	3hr 2min	Force 3 Easterly	Scattered	5

Transect Counts

4.5 Counts of each species recorded in the survey strips during each survey are presented in Tables 5 and 6 (Appendix II).

4.6 Locations of records are provided in Figures 5 to 47 (Appendix I). This also shows information on behaviour and flight direction.

April

4.7 In April, a total of 803 birds were recorded. Fulmars were the most numerous species, with a count of 318 birds. A further 80 records, mainly of sitting birds, were identified to the species group fulmar/gull species, and 46 birds were recorded as fulmar/small gull (excluding little gull). Fulmars were recorded throughout the survey area with a larger concentration in the south west. This included approximately 160 individuals that were recorded within the Bernera Wave Farm development area and its 2km buffer ( BWF).

4.8 Large auks (guillemots or razorbills) were the next most abundant species, with 88 recorded. Sixty five birds were recorded as auk species (guillemots, razorbills, black guillemots, puffins or little auks), 35 as guillemots, 10 as puffins and six as razorbills. Seven birds were identified as small auk (little auk/puffin) and fifteen birds were assigned the category of auk/small gull (auk species, black-headed gulls, little gulls, common gulls or kittiwakes).

4.9 Auks were recorded throughout the survey area with between 20 and 30 individuals recorded within the BWF. This included guillemots and razorbills with the other individuals not identified to species. The distribution showed greater auk numbers at the north end of the island consistent with the findings of Webb et al. (1990) from March to June .

4.10 There were 44 records of gannets and 30 records of kittiwakes during the April flight with approximately a third of these records occurring within the BWF.

4.11 One great northern diver was identified, along with six unidentified diver species. One Iceland gull was recorded with two herring gulls and one lesser black-backed gull. One black-backed gull species (lesser or great black-backed gull) was recorded with one small gull species and two small gulls (excluding little gull) and three gull species.

4.12 There was also a count of four great skuas, two greylag geese Anser anser, two eiders, one small wader species and three unidentified storm petrels. Seventeen birds were given the category 'small bird'.

4.13 Cetaceans and seals were recorded during the April survey; the highest count was of six harbour porpoises. Four white-beaked dolphins were recorded, as was one seal species. Half of the harbour porpoise records along with all of the white-beaked dolphin records occurred in the BWF.

May

4.14 In May the total number of birds counted was lower, with 347 records. Gannets were the most numerous bird recorded. Their numbers had increased to a count of 91, over twice the April numbers. Fulmar numbers had decreased by over 200 individuals since the April flight to a count of 84, in addition to five fulmar/gull species. This decrease may reflect increased nest attendance or a stochastic event in distribution. Fulmars should have returned from their pre egg laying exodus by the time of survey as apparently occupied sites ( AOS) and eggs had been recorded at other colonies by third week of May ( e.g. Aiton et al. 2012). 26 gannets were recorded within the BWF with individuals flying and sitting. Fulmar numbers in the BWF were small with 5 individuals recorded flying through the area. High numbers of fulmars were recorded in the wave areas of interest close into shore. These are likely to be breeding birds located around nest sites on the cliffs.

4.15 Forty auk species were recorded, with 12 large auks, one small auk, 22 puffins and one guillemot. One puffin was recorded within the BWF with three other records of auk species. The fewer auks recorded may reflect peak colony attendance during May.

4.16 Eider numbers increased in May with 25 records along with two unidentified duck species.

4.17 Sixteen common gulls and one great black-backed gull were recorded in May, these species were absent from the previous month's survey. Three herring gulls were counted with two kittiwakes, one gull species and one small gull species. Three common gulls were recorded flying through the wave areas of interest with one herring gull recorded loafing on the land on the edge of the development area.

4.18 Three Arctic/common terns were recorded in May, with 15 unidentified tern species plus one tern/small gull species.

4.19 Ten greylag geese were recorded as well as one shag, two cormorants, five wader species and three small birds.

4.20 No cetaceans or seals were recorded in May.

June

4.21 In June total bird numbers recorded had increased from May, with a total count of 516 individuals.

4.22 Fulmars were again the most numerous bird with 105 being counted along with 25 fulmar/gull species, one gull species, three small gull species and one fulmar/small gull (excluding little gull). Compared to the previous month few individuals were recorded in the coastal wave areas of interest. Approximately 10 individuals were recorded within the BWF.

4.23 Eighty one large auks were recorded, with 78 auk species, 32 puffins, three guillemots and two small auks. One shearwater/auk species was also noted. Five puffins and one guillemot were recorded in the BWF with small numbers of unidentified auk species recorded here and also in the coastal wave area of interest.

4.24 Fewer gannets were recorded in June than May, with a count of 69, 19 of which occurred in the BWF. Twelve eiders were counted in June with two common scoters and three unidentified duck species.

4.25 Kittiwake numbers had increased since May, with a count of nine. Seven herring gulls were recorded along with four black-headed gulls, which were recorded for the first time this month. Two common gulls, one great black-backed gull, one grey gull species (herring gull or common gull), one lesser black-backed gull and one gull species were also recorded.

4.26 Thirteen unidentified tern species were recorded, three Arctic/common terns and one tern/small gull species.

4.27 Two red-throated divers were recorded in June along with one unidentified diver species and two large auk/throated divers.

4.28 Four unidentified wader species and an oystercatcher Haematopus ostralegus were recorded. Of those bird objects that could not be identified to family, six 'small birds' and one 'big bird' were recorded.

4.29 Cetaceans, seals and sharks were all recorded in June with 15 harbour porpoises, one minke whale, five unidentified seals, two grey seals, one small cetacean/seal species and one basking shark identified. Four of the harbour porpoises, the minke whale, basking shark and one of the grey seals were all recorded swimming through the BWF. Only one record was made close to shore with a harbour porpoise recorded on the edge of the wave area of interest.

July

4.30 In July total bird numbers increased again, with 541 individuals being recorded.

4.31 Gannet numbers in July were almost three times the June count, with 199 individuals counted. 61 were recorded within the BWF with a number of sitting birds observed in the centre of the development area.

4.32 Fulmar numbers had increased slightly from June's count with 114 identified plus six fulmar/gull species. 16 were recorded within the BWF.

4.33 Large auk numbers dropped slightly from June with a count of 63, auk species numbers dropped by more than half from the previous month with a count of 30, however more auks were identified to species, with two black guillemots and 13 guillemots recorded. Fifteen puffins and four razorbills were identified. Three small auks were also noted. Single puffin, razorbill and guillemot records were made within the BWF with another single puffin record from the inshore wave areas of interest. Small numbers of unidentified auks were recorded in all three wave areas.

4.34 Greylag geese records had increased in July with a count of 16.

4.35 Eight kittiwakes were recorded, along with three lesser black-backed gulls, two herring gulls, one great black-backed gull, one common gull, one large gull species and one unidentified gull species.

4.36 Two Arctic/common terns were recorded and three unidentified tern species, a red-throated diver, a shag, a great skua and one unidentified skua species.

4.37 Fifteen unidentified storm petrels were counted in July, along with one unidentified duck species, one small wader species, one shearwater species and fourteen small birds.

4.38 Cetaceans, sharks and other marine mammals were more abundant this month, with four Risso's dolphins, three white-beaked dolphins, two common dolphins a basking shark, a grey seal and two unidentified seal species recorded.

September

4.39 September saw an increase in bird numbers from July with 639 individuals recorded.

4.40 Gannets were the most numerous species with 247 individuals recorded, 48 of which occurred within the BWF. Very small numbers were also recorded in the coastal wave areas of interest.

4.41 Greylag geese were at their highest number with 116 individuals recorded. All of these records were recorded over land however.

4.42 Fulmar numbers dropped significantly with only 28 individuals recorded. Approximately two thirds of these records occurred within the BWF with no records occurring in the coastal areas of interest. Greater numbers were displayed in the south west of the survey area in contrast to Webb et al. (1990) which showed fewer numbers for September to November when compared to other areas.

4.43 Only small numbers of auks were identified to species in September with 11 guillemots, three razorbills and one puffin recorded. A further 52 auks were recorded with 28 identified as large auks and 24 recorded as unidentified auk species. A single guillemot and small numbers of large and unidentified auks were recorded in the BWF with single records of guillemot and large auk species recorded in the coastal wave areas of interest.

4.44 Gulls occurred in small numbers with four kittiwakes, three common gulls, two lesser black-backed gulls and a single great black-backed gull identified. The only gulls recorded within the BWF were two large gulls that were not identified to species.

4.45 Diver and cormorant records were all confined to the areas adjacent to the shore with three red-throated divers and individual shag and cormorant records. Only one red-throated diver was recorded within the more northerly coastal wave area of interest.

4.46 Basking sharks reached their maximum count during September with six individuals recorded. Marine mammals occurred in smaller numbers with a single harbour porpoise a common seal and two grey seals recorded.

December

4.47 Fulmar numbers increased dramatically during December with 133 individuals recorded. These were largely concentrated in the southwest of the survey area with 37 occurring within the BWF. This south westerly distribution of fulmar is absent from the findings of Webb et al. (1990) but both sets appear to show a higher distribution around the northern tip of Lewis. Fewer individuals were seen close into the coast with only one recorded within the coastal development areas.

4.48 A total of 92 auks were recorded with two puffins and two guillemots recorded. 82 individuals were recorded as unidentified auk species with a further six identified as large auks. Both guillemot records and one of the puffin records were made within the BWF along with approximately 20 individuals that could not be identified to species.

4.49 Gull numbers increased from the previous month with 22 kittiwakes, six herring gulls, nine lesser black-backed gulls and two great black-backed gulls recorded. Unidentified gull numbers also increased with 27 unidentified gull species and 17 large gulls recorded. Five of the kittiwake records along with five unidentified gulls were recorded within the BWF. Very small numbers were recorded in the coastal areas with the majority of the records occurring either over land or directly adjacent to the coast.

4.50 Divers occurred in their highest numbers with one red-throated diver recorded and 26 unidentified diver species recorded. The unidentified divers occurred in three groups all of which occurred in the coastal wave areas of interest.

4.51 Gannets occurred in their lowest numbers with only two individuals recorded. Cetacean numbers were also greatly reduced with a single harbour porpoise recorded in the southwest of the survey area.

February

4.52 Fulmar numbers continued to increase with 186 recorded during February. As with the December survey, large numbers occurred in the southwest of the block with 122 individuals recorded within the BWF.

4.53 Auk numbers remained high with a total of 177 individuals recorded. This included five puffins, one guillemot, two small auks, 13 large auks and 156 unidentified individuals. Approximately 30 records occurred within the BWF and included the guillemot record and two of the puffin records.

4.54 Gull numbers continued to increase with 121 kittiwakes, 14 herring gulls, 11 great black-backed gulls and single lesser black-backed gull and black-headed gull records made. Three of the great black-backed gulls were recorded within the BWF along with approximately 30 kittiwake records.

4.55 Diver numbers reduced during February with only three unidentified divers recorded. 13 eiders were recorded around the mouth of Loch Roag and two common scoters were identified in the north of the survey area. These and two of the diver records were made within the northern most wave area of interest.

4.56 Gannet and marine mammal numbers remained low with only five gannets recorded and again a single harbour porpoise record.

Abundance estimates

4.57 Table 6 (Appendix II), presents point abundance estimates for each species in the study area from design-based analysis of encounter rates within the strip transects following Buckland et al. (2001).

4.58 The abundances estimated for all species and species groups are approximately ten times the number of records recorded, as the strip transect survey covered 10% of the study area.

4.59 For those species, where only a single record was made, this was simply multiplied by 10 to obtain an estimate. Compared with breeding population counts from seabird colonies within foraging range (from the Seabird Monitoring Programme ( http://jncc.defra.gov.uk/smp/) the estimated abundances of species recorded were low. This may be attributed to low relative importance of the area for these seabird species, particularly for those from more distant colonies having larger foraging ranges where the study area occupies a much smaller proportion of foraging area available. However it may also incorporate stochastic or sampling methodological issues such as colony attendance, chance foraging out of the area and availability of diving birds, which is discussed below (Comparison of ground-based and digital aerial survey).

4.60 A maximum of 3,161 fulmars were estimated to be in the study area ( Table 6) which was just 3% of the fulmars breeding within the colonies within foraging range on the Flannan Islands, St Kilda, the Monach Isles, North Rona, Sula Sgeir, the Shiant Isles, Handa and Cape Wrath.

4.61 Gannet abundance estimates peaked at 2,464 in September ( Table 6), which again is around 1.7% of the 146,574 gannets breeding on the colonies at Flannan Isles, St Kilda and North Rona & Sula Sgeir.

4.62 Around 254,000 auks were counted on colonies within foraging range of the study area during Seabird 2000, of which the peak count of 2,051 auks in April was less than one percent.

4.63 In December 269 divers were estimated in the study area. If more than 170 of these were red-throated divers this would represent over 1% of the British wintering population (Musgrove et al. 2011). The ground counts in December recorded higher numbers of great northern divers than red-throated divers (see ground-based count results below) in which case if more than 25 were great-northern divers, this would exceed the overwintering 1% threshold for this species. In April 69 divers were estimated, which based on observed species ratios could have been great northern divers, so again this would exceed the overwintering 1% level. The estimates of unidentified diver species in June July, December and February exceed the overwintering 1% level for great northern diver and black throated diver (1% = 6) though obviously the actual numbers of these species present are unknown.

Ground-based count results for northwest coast of Lewis

4.64 Table 2 presents the dates that ground counts were undertaken. Ground counts were conducted on the same days as aerial surveys on 19th April, 18th June, 26th July, 22 ^nd September and 17 ^th December 2012. In May the ground count was undertaken on 2 ^nd May, when the original aerial survey was planned, but the aerial survey was postponed until 29 ^th May due to a technical fault. Similarly in February 2013 the ground count was started on 2 ^nd February, when the aerial count was due to take place but on transit the survey aircraft had to abort due to a technical fault and did not complete the survey until 12 ^th February.

4.65 Numbers of birds recorded from each vantage point on each date are shown in Table 7 (Appendix II). Numbers of marine mammals and basking sharks are presented in Table 8 (Appendix II).

4.66 Locations of records are presented in Figures 48 to 84 (Appendix I).

April

4.67 In April, fulmars were the most commonly recorded bird with 25 counted, nine of which were at Port of Ness. Common gulls were the second most abundant bird, with 15 recorded. Eleven great northern divers were seen, 10 shags , nine herring gulls, eight black-headed gulls and eight guillemots were also recorded.

4.68 There were six records each of black guillemots, razorbills and lesser black-backed gulls . Four great black-backed gulls and four large auks (guillemots or razorbills) were also seen. Three gannets were recorded. There were single records for red-throated diver , great skua , Iceland gull Larus glaucoides, unidentified gull species and a cormorant . A total of 122 birds of 19 species were recorded in April.

4.69 In April sea state ranged between 3 and 5 during the VP counts so it is possible some, especially cetacean species were not detected.

May

4.70 In May, fulmar numbers had increased with a count of 37; again the highest number was at the Port of Ness (15). Great northern diver numbers had also increased with 20 being recorded. Razorbills and guillemots had also become more numerous with a count of 17 and 14 respectively. Shag numbers had a small increase since the previous month with 13 recorded.

4.71 Nine black guillemots were seen, and nine puffins were also counted, this species was not recorded at all in April. Five red-breasted mergansers were seen, all from Bragar. Five gannets were also counted. There were two records each of red-throated divers, kittiwakes , black-headed gulls, great black-backed gulls and cormorants. Large auk species, lesser black-backed gull and Slavonian grebe Podiceps auritus had single records each. A total of 145 birds of 18 species were recorded in May.

June

4.72 In June, bird activity had generally increased with 245 birds of 21 species being recorded. Guillemots were the most abundant with a count of 85, the majority being recorded from Port of Ness and Butt of Lewis East. Fulmar numbers had increased again with a count of 52. Twenty six black guillemots were recorded, the highest count (11) coming from the Butt of Lewis West.

4.73 Gannets were recorded around the coast, with a count of 19. Thirteen shags were recorded in June, as were eight Arctic terns Sterna paradisaea. Six large auk species and six razorbills were seen. Common gulls were recorded again in June, with a count of four. There were three records each of red-throated divers, great northern divers, and herring gulls and two records of Arctic skuas, great black-backed gulls and puffins. There were single records of black-headed gulls, black-throated divers, cormorants, common terns, lesser black-backed gulls and great skua.

4.74 One unidentified dolphin species, two grey seals and four harbour porpoises were recorded from vantage points in June.

July

4.75 In July, bird activity had generally decreased with 111 birds of 18 species being recorded. The most abundant species in July was shag with 37 records. There were 22 records of black guillemots, 12 records of guillemots and six records of fulmars. Razorbills, great northern divers and gannets had five records each. There were counts of four individuals for red-throated divers and Arctic terns, and three gannets were recorded. There were single records for Arctic skua, cormorants, great black-backed gulls, large auk species, kittiwakes, great skuas and European storm-petrels.

4.76 In the July counts a basking shark, five grey seals and a harbour porpoise were recorded.

September

4.77 A total of 315 birds were recorded during the September count with 10 different species recorded. The most abundant species were gannets with 161 individuals recorded. All of these records were made from the count points in the north west with a number of individuals occurring within or just outside of coastal wave areas of interest.

4.78 Shags were the next most abundant species with 99 individuals recorded with most records occurring around the Butt of Lewis.

4.79 All of the other species recorded occurred in much lower numbers with 17 herring gulls, 13 great back-backed gulls, nine common gulls, six cormorants, four guillemots, three razorbills, two red-throated divers and a single black guillemot record.

4.80 A single basking shark record was also made along with eight grey seals. No cetaceans were recorded during this survey.

December

4.81 Total bird number decreased during the December survey with 247 individuals recorded. Eleven different species were recorded with fulmars the most abundant species. 83 individuals were recorded with all of the records occurring in the northwest of the survey block around the Butt of Lewis.

4.82 Eiders were the second most abundant species with 62 individuals recorded from the Melbost vantage points. These were recorded in two separate groups close in to the shore.

4.83 Gulls were recorded in slightly higher numbers with 33 herring gulls, 18 great black-backed gulls and four common gulls and kittiwakes seen.

4.84 Ten great northern divers, three red-throated divers and 26 shags were recorded within inshore waters with very low numbers of auks recorded in the northwest of the survey area. This included a single guillemot record and three black guillemot records.

4.85 No cetacean, mammal or shark records were made during December.

February

4.86 Numbers of birds more than doubled between the December and February counts with a total of 655 individuals recorded, comprising 12 species. This included a large count of 430 fulmars with all but one individual seen around the Butt of Lewis. Herring gull numbers continued to increase with 103 birds recorded. Other gulls seen during this survey included 11 kittiwakes and ten great black-backed gulls.

4.87 Shags were recorded throughout the survey area with 45 individuals recorded. Forty-five guillemots were also seen, all from the Port of Ness vantage point.

4.88 All other species occurred in very low numbers with five great northern divers, two razorbills and single records of red-throated divers, gannets, eiders and black guillemots. In February sea state ranged between 3 and 5 during the VP counts so it is possible some, especially cetacean species were not detected.

Ground-based counts of Loch Roag

4.89 The timetable of planned survey visits proved not possible to achieve due to weather constraints. A total of six survey visits were achieved, in April, June, July and October 2012 and January and February 2013 ( Table 9, Appendix II). High quality survey data can only be obtained in calm conditions with good visibility but the summer and autumn of 2012 was notable for long periods of unsettled weather. As a result it proved not possible to undertake a survey visit in May. The planned late August/early September visit also proved impossible to achieve due to weather constraints but was eventually completed on 2 ^nd October.

4.90 The April, June and October visits were all completed in a single day in excellent (April and June) or very good (October) survey conditions and achieved good coverage of the whole survey area ( Figure 85, Appendix I). The July visit had to be abandoned twice due to deteriorating weather conditions. Despite attempting July survey work on three dates, weather problems resulted in there being no coverage of the north-west part of the survey area in July ( Figure 86, Appendix I).

4.91 The January and February 2013 surveys were achieved in generally excellent survey conditions and good coverage ( Figure 87, Appendix I) though spread over two consecutive days each time due to insufficient day length.

4.92 Time constraints meant that it was not possible to achieve complete coverage of the survey area as Figures 85, 86 and 87 (Appendix I) show. The large number of small islands and the indented nature of the coastline meant that even if a 1km square had some visual coverage it was common for a proportion to be out of sight. However, despite these limitations, and with the exception of the north-western parts of the survey area in July referred to above, a very high level of coverage was achieved and it is considered likely that at least 75% of birds on the water in the survey area were recorded on each visit. The corresponding figure for flying birds is likely to be even greater, and is considered likely to exceed 90% detection of birds flying in the survey area. The coverage and detection of hauled out seal in Loch Roag is also considered likely to be at least 75%, however the detection of seals is dependent on hauling-out behaviour which is typically entrained to the tidal cycle. Put simply, seals are most likely to be hauled out within two to three hours either side of low tide and so counts made within this period were more likely to encounter hauled out seals.

Waterbirds

4.93 Thirty eight waterbird species were recorded during the six survey visits ( Tables 10 to 12, Appendix II) and both species of seal ( Table 13, Appendix II). Details of the numbers seen on each visit in each survey sector are presented for most of these species individually in Tables 14 to 44 (Appendix II).

4.94 On the basis of these results, inner Loch Roag appears to have a relatively high value for six bird species of high conservation importance, namely the three diver species, Slavonian grebes, Arctic terns and common terns. These six species are all on Annex 1 of the EU Birds Directive. Cormorants, eiders and black guillemots, although of lower conservation value, occur in regionally important (>1%) numbers and are therefore also of importance. These nine species are considered to be the priority bird species and each of these is discussed in more detail in the Discussion. Figures 88 to 94 (Appendix I) show the recorded numbers and distributions for red-throated divers, black-throated divers, great northern divers, Slavonian grebe, cormorants, shags and black guillemots.

4.95 The commonest bird species recorded were eiders, shags, herring gulls, great northern divers, black guillemots and greylag geese with over 30 individuals recorded on average for each over the six visits ( Table 11, Appendix II). Herring gulls, greylag geese and shags are all common breeding species in Lewis and the numbers using inner Loch Roag are of low importance and with respect to the regional populations. Greylag Geese were particularly abundant in July, and at this time of year a high proportion of the adults seen were accompanied by dependent goslings. A total of 78 goslings in an estimated 18 broods were counted in July. Great northern divers were recorded in greatest numbers in the winter and are regularly occurring winter visitors. The Scottish wintering population is estimated at 1,000-3,000 (Forrester et al. 2007) and the biogeographic wintering population 5,000 (Wetlands International 2012) so the 80 and 95 recorded in January and February 2013 respectively represent nationally and internationally important numbers.

4.96 Fulmars, red-breasted mergansers, common gulls, great black-backed gulls, razorbills and gannets were all regularly recorded in more moderate numbers. These are all common species in the Western Isles and it appears that the numbers using inner Loch Roag are of low importance with respect to the regional populations. Five of the 11 razorbills seen in July were accompanied by dependent young. Smaller numbers of black-headed gulls, kittiwakes, little grebes Tachybaptus ruficollis, Manx shearwaters, lesser black-backed gulls and a glaucous gull Larus hyperboreus were recorded. With the exception of glaucous gull (a regular migrant species) the other species were recorded in very low numbers in respect to regional populations.

4.97 Small numbers of mallards Anas platyrhynchos and teals Anas crecca, and one pair of shelducks Tadorna tadorna were seen on most visits; all these were likely to be breeding locally. These are all common breeding species in the Western Isles. Five common scoters were seen in April; given the time of year and lack of local breeding records, these were likely to be passage birds. The single common scoter seen in June, the height of the breeding season was unexpected and was probably a non-breeding bird, though raises the possibility that this species may breed locally. Small numbers of long-tailed ducks (up to three) and goldeneyes (five) were seen in January and/or February 2013, again representing small relative proportions of these wintering ducks.

4.98 Recording wader species was not an aim of these surveys. Nevertheless it is relevant to report that no significant concentrations of waders were located, nor were significant areas of suitable habitats noted that are likely to be attractive to large numbers of waders. Small numbers (<10 of any species) of oystercatchers, curlews Numenius arquata, ringed plovers Charadrius hiaticula, common sandpipers Actitis hypoleucos, lapwings Vanellus vanellus and greenshanks Tringa nebularia were seen on the July visit, and all these birds are likely to have been locally breeding birds. The only notable record was five greenshanks seen feeding on intertidal habitat in Loch Cean Hulavig, in the extreme south east part of Loch Roag.

Marine mammals

4.99 No cetacean species (or basking sharks) were seen during the course of the survey work of Loch Roag

4.100 Small numbers of harbour seals (maximum total 12 individuals) were recorded on all visits ( Table 45, Appendix II and Figure 95, Appendix I). The great majority of records were from the sheltered south east part of the survey area in Loch Hulavig. The most important haul-out site was at Eilean an Tighe ( NB220/300), a small island in the southern part of Loch Hulavig, where up to nine common seals were seen hauled out together. Counting was undertaken irrespective of the tidal conditions. However, counting of the southern parts of the survey area on April and July visits coincided with the low tide period (when seals are more likely to be hauled out) and this may explain why higher numbers were recorded on these survey visits.

4.101 Grey seals were scarce for all surveys except the February 2013 survey when 83 were recorded in the West Great Bernera count section ( Table 46, Appendix II, Figure 95, Appendix I).

Density surface modelling of digital aerial survey data

4.102 Sufficient data were collected to undertake density surface modelling for fulmars, gannets and all auk species combined for each month. Table 47 (Appendix II) presents the density surface model statistics and abundance and density estimates obtained.

4.103 Comparison between Tables 6 and 47 shows the abundances estimated from DSM were similar to estimates from the design-based analyses which is to some degree to be expected given the extent and intensity of survey coverage and the inclusion of only x and y as covariates.

4.104 The resulting density surface maps are shown in Figures 96 to 137 (Appendix I). Accompanying each map is a plot of relative coefficient of variation ( CV) - a measure of confidence in the density estimate for each grid cell estimated via Bayesian results as per Wood (2006). Note that legend scales vary between species maps.

Fulmars

4.105 Table 32 shows estimated fulmar numbers decreased considerably in the survey area from 3,197 birds in April to around 1,000 through May (1,159), June (976) and July (1,121) before decreasing to just 302 in September and back up again to 1,295 in December and 1,690 in February 2013.

4.106 The model diagnostics ( Table 47, Appendix II), density surface and CV maps for fulmars in April showed a reasonable fit with the observations, though with a relatively high multimodal smooth (estimated degrees of freedom 7.4). Highest densities of up to 18 birds km ^-2 were predicted in the south west of the survey area and around the Butt of Lewis and lowest densities in central areas ( Figures 96 & 97, Appendix I). In May, June and July highest estimated densities were close to shore, particularly around Loch Roag and the Butt of Lewis, but with generally lower densities than recorded in April. Note that the high estimated densities east of the Butt of Lewis may be an artefact of this area being the extreme east of the study area and the data showing a correlation with increasing longitude east. This is shown by high CVs in the corresponding CV maps as the modelled estimates here differ from actual densities recorded.

4.107 Few fulmars were recorded in September and the predicted surface reflects the higher densities recorded further offshore than during May to July, particularly in the south west. More 'flexible' multimodal models were fitted to the December and February (2013) data with higher overall densities predicted, especially in the south west and north east of the survey area in both months.

Gannets

4.108 The estimated number of gannets in April was 415, increasing to 932 in May, decreasing again to 684 in June before increasing greatly to an estimated 2,035 in July and 2,571 in September. Estimated numbers then decreased again through 887 in December to 50 in February 2013.

4.109 Of the few gannets recorded in April the density surface model reflected slightly higher densities recorded in the south west and north of the survey area. In May, observations were relatively evenly distributed across the survey area, approaching a seemingly more random distribution, so only a weakly fitting model could be fitted (probability p that the model had no effect >0.5), with predicted densities around 0.9 birds km ^-2 across the study area, just marginally higher in the north and the south and lower in central areas.

4.110 In June there was a general arc of observations from south of Loch Roag extending to more offshore areas in the centre of the survey area then a broad band of lower concentrations towards the shore in the north. This lead to a relatively flexible model fit (estimated degrees of freedom 7.0) being selected with highest densities up to 3 birds km ^-2 in the south and central northern areas. In July the high recorded numbers were well distributed through the survey area, with a slight trend of higher densities (up to 3 birds km ^-2) in the south west smoothing to lower densities in the north east.

4.111 In September gannets were even more widely distributed such that the x, y model with very low estimated degrees of freedom (<1) only accounted for 0.8% of the deviance. The predicted densities were in the range 2-3 birds km ^-2 throughout the study area, with a non significant trend ( p>0.1) of slightly higher densities to the north.

4.112 Only two gannets were recorded in December and five in February (2013) so the density surface modelling maps are only provided for completeness and provide some reference for comparing confidence with the other maps.

Auks

4.113 The highest numbers of auks were recorded and estimated in April (2,077), then fewer in May (760) more in June (1,813) and July (1,348) before dropping to the lowest numbers in September (676) before increasing through the winter with 887 predicted for December and 1,765 in February (2013).

4.114 In April the highest auk densities of up to 7 birds km ^-2 occurred around the north end of Lewis with lower densities in the centre of the study area and higher again in the south. This pattern of distribution was similar in May though with lower numbers and estimated densities only up to 3 birds km ^-2 in northern waters and to the south. In June a more flexible model was fitted (estimated degrees of freedom 5.3) predicting highest densities around Loch Roag (up to 6 birds km ^-2), decreasing through the central part of the study area and increasing again towards the north.

4.115 In July auks were more widely distributed such that the model fit was not significant and accounted for less than 1% of the deviance. Auks were predicted at densities around 1.3 birds km ^-2 across the study area. In September the highest densities of around 1 bird km ^-2 were predicted radiating out from the observations around Loch Roag but again densities were fairly evenly predicted across the study area reflecting the otherwise wide distribution of the observations and relatively poor model fit (just failing significance at the 5% level).

In December and February (2013) auk observations and predicted densities displayed a similar pattern of distribution to April and May 2012, with significant model fits predicting higher densities in the north and south of the study area (up to 2.2 birds km ^-2 in December and 4.2 birds km ^-2 in February).

Comparison of ground-based and digital aerial survey

4.116 Figures 138 to 142 (Appendix I) show the aerial survey and ground count records made during the synchronised surveys. Only sitting birds are shown as flying birds were generally not recorded during the ground counts.

4.117 From these figures it can be seen that very few records were made in areas of overlapping coverage and no records were exactly coincident in space or time, so using one method to check identification of records from the other is impossible. However a number of general observations were made.

4.118 The ground counts recorded larger numbers of diver species (red-throated divers, black-throated divers and great northern divers), shags and black guillemots overall and given the differences in areas covered higher densities of all auk species.

4.119 Figures 143 to 151 (Appendix I) show the results of detailed analyses of the densities of auk species. These plots attempt to identify patterns in the distribution of auks ('depth' axis) counted from the vantage points and to what extent that pattern may be effected by declining detectability from the vantage point (horizontal axis). The auk densities from aerial surveys in the respective area and season are provided in blue to the right for comparison along the distance to coast gradient. Figure 143 shows that in the breeding season guillemots were recorded at a density of 7 birds km ^-2 within 0.5km of the VPs (and coast) around the Butt of Lewis. Very few guillemots were recorded in the corresponding aerial survey transect segments with a peak density of 0.28 birds km ^-2 occurring 1.5km from the coast, however this is based on few records. More records were made of large auk sp. (guillemot or razorbill) or unidentified auk species. A density of 1.5 birds km ^-2 was estimated for large auk sp. within 0.5km of the coast, with no other auk records. If all of these were actually guillemots, this represents 21% of the density estimated within 0.5km of the VPs. VP records decreased rapidly with both distance from vantage point and distance from coast so as few guillemot records were made from the aerial surveys conclusions cannot be drawn as to the relative affects of decreasing detectability and increasing distance from coast for this species.

4.120 Figure 144 shows the estimated densities for razorbills again in the breeding season months around the Butt of Lewis. These show lower densities recorded from the VPs than guillemots, with a very obvious near shore distribution. Peak density was 1.3 birds km ^-2 up to 0.5km from the VPs and this did not decline greatly up to 1km from the VP along the coast (up to 0.5km out), however densities beyond 0.5km from the coast were very low. This suggests detectability of razorbills was good out to at least 1km. Given the similarities in structure and plumage between razorbills and guillemots this would also suggest that detectability of guillemots should also have been good out to 1km and so that species too had a predominantly near shore (within 0.5km) distribution in this area. No razorbills were recorded in the corresponding aerial survey transect segments so no direct comparisons could be made. However, summing peak guillemot and razorbill densities within 0.5km of VPs gives an overall density of guillemots and razorbills of 8.3 birds km ^-2 compared to 1.5 birds km ^-2 large auk sp. (guillemots or razorbills) estimated from aerial surveys in this coastal distance band (no other auks were recorded in this band during these months). Thus the aerial survey estimated guillemot or razorbill density was only 18% of that recorded from the VPs. In fact the aerial surveys only recorded 3 guillemots or razorbills compared to 32 from the VPs in this coastal band (9.4%). Possible explanations include by chance the aerial survey transects not capturing as well used areas and/or availability bias, where the rapid aerial surveys cannot detect birds that are diving. If the latter, this suggests maybe a tenth of birds of these species may be available for any instantaneous snapshot.

4.121 Figure 144 shows the estimated densities of black guillemots from the April to July VP counts around the Butt of Lewis. These show a similar pattern to razorbills, with highest densities within 0.5km of the coast and no steep decline in detectability out to 1km suggesting a true near shore distribution. A count of 17 birds gave a peak density of 4.4 birds km ^-2. No black guillemots (or unidentified auk species) were recorded in the corresponding aerial survey transect segments again suggesting either chance differences in areas covered excluded them from the aerial survey coverage, or the high proportion of time spent in dives by this species made them unavailable for recording by digital aerial survey.

4.122 Very few puffins were recorded, with a peak of 4 recorded 1km from the vantage points producing a peak density of 0.8 birds km ^-2. Lower numbers were recorded up to 2km from the VPs (and coast). With such small sample sizes patterns cannot be confidently determined, however the low densities of puffins estimated between 0.5 and 1.5km from the coast from aerial survey records suggests this species which in contrast to guillemot, razorbill and black guillemot does not breed on the north of Lewis may have a more widespread coastal distribution here.

4.123 Figures 147 to 151 (Appendix I) show the densities of records from the west coast vantage points (from Aird Dell south) and corresponding aerial survey transects. In contrast to Figure 143, Figure 147 shows that the 29 guillemots recorded were widely distributed from 0.5 to 2km from the coast (and vantage points) with the peak density of 0.6 birds km ^-2 occurring 2km from the coast. Though it is impossible to tell if this is an undercount due to distance from vantage point, it does at least show that the species can be regularly recorded at this distance which provides greater confidence in the distribution patterns discussed for birds around the Butt of Lewis above. Very few guillemots were recorded from aerial surveys for comparison with more records of large auks or unidentified auk species. The former of these showed an increase in densities further from the coast with a peak density of 0.45birds km ^-2 at 2km, a similar pattern to that produced from the VP estimates.

4.124 Like guillemots, razorbills also showed a more offshore peak in densities compared to around the Butt of Lewis with a peak of 1.1 birds km ^-2 at 1km from the coast ( Figure 148). This is higher than the peak density for guillemots, which is the opposite of around the Butt of Lewis colonies. No razorbills were recorded in the corresponding aerial survey transects so direct comparisons cannot be made. Records were made of large auk species (guillemots or razorbills) but comparisons are complicated by the recording of unidentified auk species also, which could include guillemots and razorbills. Thus comparison will be made of all auk densities combined following black guillemot and puffin below.

4.125 As with VP count data from around the Butt of Lewis, Figure 149 shows highest densities of black guillemots up to 1 bird km ^-2 within 0.5km of the VPs (and coast) from the west coast VPs with much lower densities further offshore. Again, no black guillemots were recorded from the corresponding aerial survey transects so distribution comparisons cannot be made, however of note there were no unidentified auk species recorded either which suggests that the black guillemots (of which 17 were counted from VPs within 0.5km of the coast) were not available for recording in the aerial survey transects either due to location or diving (there were also no records of unidentified small birds which is a category that would 'catch' potential confusion with e.g. a small gull).

4.126 Eight puffins were recorded from west coast VPs compared to 13 from around the north. Again the peak density of 0.4 birds km ^-2 occurred between 0.5 and 1km from the coast (and VPs) suggesting a less near shore distribution compared to black guillemot. Few puffins were recorded in aerial surveys with a peak density of 0.08 birds km ^-2 between 1km and 1.5km from the coast.

4.127 Figure 151 shows a comparison of auk densities derived from vantage points records with those from aerial surveys along the west coast section. For the west coast vantage points comparison of densities at the level of auk species is required due to the complication of unidentified auks recorded during the aerial surveys. The figure shows the VP counts consistently recorded higher densities of auks in all distance bands from the coast with a peak density of 1.4 birds km ^-2 at 0.5 to 1km from the coast compared to 0.7 birds km ^-2 from aerial surveys in the same band.

4.128 In contrast to these breeding season densities, aerial survey counts between September and February yielded higher densities of auks along the west coast than the vantage points, both overall and in all distance from coast bands apart from 0km to 0.5km ( Figure 152). The majority of VP records were of black guillemots, which to some extent explains the predominantly near shore highest densities recorded from the VPs while the peak in the 1km to 1.5km distance band suggests detection was reasonable at least out to this distance. A possible explanation for this reversal in densities is that some VPs were not covered in the winter months (see Table 2, Appendix II) and that the sea state was a 4 or 5 during the February 2013 surveys which would have made auk detection much more difficult from the shore as opposed to from the aerial survey video frames. Overall sample sizes were also small in these months so chance differences between the two timings or locations may have had a greater effect. A similar comparison for winter months around the Butt of Lewis was not practical due to very few records from VPs during these months, again perhaps due to poor sea state.

4.129 Very few records of shags and divers were made during the aerial surveys so the analyses completed above for auks were not completed for these species. It is likely that the fewer aerial survey records are due at least partly to the increased availability of these species for vantage point counts over approximately 30 minute counts compared to the near instantaneous recording during the digital surveys.