Updated Technical Report summarising information on marine mammals which occur in the Moray Firth

Transcription

1 Updated Technical Report summarising information on marine mammals which occur in the Moray Firth Project Name: Reference: Project Manager: BOWL EIA Work Marine Mammals RPS Kate Grellier Drafted by: Kate Grellier QA by: Professor John Harwood Date: Thursday, 16 February 2012 VAT reg. No. GB SMRU LIMITED is a limited company registered in Scotland, Registered Number: Registered Office: 5 Atholl Crescent, Edinburgh EH3 8EJ

2 Table of contents Table of contents... 1 List of Figures... 3 List of Tables Introduction Designations and legislation European Protected Species (cetaceans) Seals Natura 2000 sites Data collection methods and sources Cetaceans Surveys Habitat association modelling Density estimates Passive acoustic monitoring (PAM) Seals Surveys Telemetry Species accounts Bottlenose dolphin Distribution Seasonal variation Abundance Density The Moray Firth SAC Links between the Moray Firth SAC and the BOWL site Harbour porpoise Distribution Seasonal variation Abundance/density Minke whale Distribution Seasonal variation Abundance Common dolphin Distribution Seasonal variation Abundance White-beaked dolphin Distribution

3 4.5.2 Seasonal variation Abundance In the southern Moray Firth (in relation to the cable route) Risso s dolphin Distribution Seasonal variation Abundance In the southern Moray Firth (in relation to the cable route) Other cetacean species In the southern Moray Firth (in relation to the cable route) Harbour seal Distribution Seasonal variation on land Abundance The Dornoch Firth and Morrich More SAC Links between the Dornoch Firth and Morrich More SAC and the BOWL site Loch Fleet National Nature Reserve (NNR) In the southern Moray Firth (in relation to the cable route) Mortality from corkscrew injuries Grey seal Distribution Seasonal variation on land Abundance Grey seal SACs Links between grey seal SACs and the BOWL site In the southern Moray Firth (in relation to the cable route) References Appendices

4 List of Figures Figure 1. A map of the Moray Firth (limit shown by the dashed line) showing the bathymetry and the outer boundary of the Moray Firth SAC (taken from Thompson et al. 2010a and adapted) Figure 2. Sightings of bottlenose dolphins around the UK ( ; taken from Reid et al. 2003) Figure 3. Sightings of dolphins from the datasets shown in Table 2 (taken from Thompson and Brookes 2011) Figure 4. Map of track lines from the University of Aberdeen s August and September 2010 aerial surveys of the outer Moray Firth (taken from Thompson and Brookes 2011) Figure 5. Map of the survey track followed on the BOWL boat surveys (taken from Thompson and Brookes 2011) Figure 6. Sightings of dolphins made during the University of Aberdeen s 2010 aerial surveys of the outer Moray Firth (taken from Thompson and Brookes 2011) Figure 7. Sightings of dolphins made during the BOWL boat surveys between April and October 2010 (taken from Thompson and Brookes 2011) Figure 8. Prediction of the dolphin species composition within each 4x4km grid cell using all data (taken from Thompson and Brookes 2011) Figure 9. Prediction of the dolphin species composition within each 4x4km grid cell using all data except the BOWL boat survey data (taken from Thompson and Brookes 2011) Figure 10. Prediction of the likelihood that dolphins encountered in each 4x4km grid cell are likely to be bottlenose dolphins. Data are as for Figure 9, but presented as bottlenose dolphins (black portion of pie chart) vs. all other species (taken from Thompson and Brookes 2011) Figure 11. Predicted number of bottlenose dolphins in each 4x4km grid cell (taken from Thompson 2011a) Figure 12. Spatial variation in the occurrence of dolphins in April October of 2009 and The pie charts show the proportion of days that dolphins were detected on PODs at each sampling location (taken from Thompson and Brookes 2011) Figure 13. Frequency histogram showing the number of hours per day that dolphins were detected on PODs deployed at the Beatrice Demonstrator site from (left) and within the BOWL site from (right) (taken from Thompson and Brookes 2011). This Figure only includes those days on which any animals were detected Figure 14. CRRU survey tracks extending from Lossiemouth (in the west) to Fraserburgh (in the east; taken from Thompson et al. 2010a) Figure 15. Dolphin sightings made by the CRRU between 2001 and 2008 (taken from Thompson et al. 2010a) Figure 16. Bar chart showing the distribution of bottlenose dolphin encounters across the survey area (from Lossiemouth to Fraserburgh). The dark grey bars represent the number of visits to each 1km section of the coast (left y axis). The encounter rate (right y axis) is shown in pale grey (taken from Culloch and Robinson 2008) Figure 17. Monthly values for the % of days that dolphins (black circles) and porpoises (red squares) were detected by PODs deployed within the BOWL area (taken from Thompson and Brookes 2011) Figure 18. Frequency histogram of encounter rate for bottlenose dolphins across each survey month (+/ 95% confidence intervals; taken from Culloch and Robinson 2008)

5 Figure 19. Map of the Moray Firth showing the location of the two POD sites at Lossiemouth (left) and Spey Bay (right) from 2009 to 2011 (taken from Thompson 2011) Figure 20. Monthly variation in the median number of hours that dolphins were detected at the Spey Bay (left) and Lossiemouth (right) POD sites from 2009 to 2011 (taken from Thompson 2011) Figure 21. The median number of hours per day that dolphins were detected at the Spey Bay (left) and Lossiemouth (right) POD sites (taken from Thompson 2011) Figure 22. Annual abundance estimates (+/ 95% confidence intervals) of dolphins using the southern Moray Firth coast showing the degree of inter annual variability (plotted using data from Culloch and Robinson 2008) Figure 23. The results of the classification of whistle events in the EAR data using the whistle classifier. BND=events classified as bottlenose dolphins (white) and OTHER=events classified as other species (grey). N.B. The scale of the y axis for the D01 EAR is different to the EARs deployed on the BOWL and MORL sites (taken from Booth et al. 2011) Figure 24. Sightings of harbour porpoises around the UK ( ; taken from Reid et al. 2003) Figure 25. The predicted number of harbour porpoises in each 4x4km cell. Values are based upon measures of relative abundance derived from habitat association modelling, scaled according to estimates of absolute abundance from aerial line transect surveys and extrapolated to other areas according to predicted relative abundance (taken from Thompson and Brookes 2011) Figure 26. Spatial variation in the occurrence of porpoises in April October 2009 and The Figure shows the proportion of days that porpoises were detected on PODs at each sampling location (taken from Thompson and Brookes 2011) Figure 27. Fine scale spatial variation in the occurrence of porpoises in and around the BOWL (and MORL) development site. Data are from April October 2009 and Pie charts for each sampling site represent the median proportion of time that porpoises were detected each day (taken from Thompson and Brookes 2011) Figure 28. Frequency histogram showing the number of hours per day that porpoises were detected on PODs deployed at the Beatrice Demonstrator site between 2005 and 2010 (left; taken from Thompson and Brookes 2011) and within the BOWL site between 2009 and 2011 (right; taken from Thompson and Brookes 2011). This Figure only includes those days on which there was at least one detection Figure 29. Harbour porpoise (circle) and minke whale (triangle) sightings made by the CRRU between 2001 and 2008 (taken from Thompson et al. 2010a) Figure 30. Figure showing which PODs were included in the analysis of seasonal trends in detections of porpoises in the Moray Firth SAC ( ), along the southern Moray Firth coast ( ) and in the outer Moray Firth ( ; taken from Thompson et al. 2010a) Figure 31. Monthly variation in the median number of hours per day that porpoises were detected on PODs within the BOWL development area (taken from Thompson and Brookes 2011) Figure 32. Monthly variation in the median number of hours that porpoises were detected at the Spey Bay (left) and Lossiemouth (right) POD sites (taken from Thompson 2011) Figure 33. The median number of hours per day that porpoises were detected at the Spey Bay (left) and Lossiemouth (right) POD sites (taken from Thompson 2011) Figure 34. Sightings of minke whales around the UK ( ; taken from Reid et al. 2003)

6 Figure 35. Map of survey tracks from the University of Aberdeen s 2009 boat based surveys (taken from Thompson and Brookes 2011) Figure 36. Sightings of minke whales made during the University of Aberdeen s 2009 boat based surveys (track lines are shown in Figure 35; taken from Thompson et al. 2010a) Figure 37. Sightings of minke whales made during the University of Aberdeen s 2010 aerial surveys of the outer Moray Firth (track lines are shown in Figure 4; University of Aberdeen unpublished data) Figure 38. The number of minke whale sightings made each month during the BOWL boat surveys (solid bars; BOWL unpublished data). It should be noted that although the number of sightings has not been effort corrected, the amount of survey effort in each month is similar. Where two surveys were carried out in a single month (April 2010 and January 2011), as opposed to the usual one, the mean number of sightings was calculated. Months in which no surveys were carried out (November 2009, January 2010 and November 2010) are shown by the patterned bars Figure 39. The number of minke whale encounters per km of survey effort between the months of May and October 2001 to 2006 (taken from Robinson et al. 2009) Figure 40. Minke whale sightings per unit effort in months when the cold water current, or the warm water plume, were dominant (taken from Tetley et al. 2008). The mean, interquartile range and range are shown Figure 41. Sightings of common dolphins around the UK ( ; taken from Reid et al. 2003) Figure 42. The distribution of common dolphin sightings recorded during surveys carried out between February and November in 2001 to 2009 by the CRRU (area covered shown by the shaded boxes) and WDCS (area covered by all the boxes; taken from Robinson et al. 2010) Figure 43. Sightings of white beaked dolphins around the UK ( ; taken from Reid et al. 2003) Figure 44. Sightings of Risso s dolphins around the UK ( ; taken from Reid et al. 2003) Figure 45. Sightings of fin, humpback, killer and long finned pilot whales around the UK ( ; taken from Reid et al. 2003) Figure 46. The extent of the Dornoch Firth and Morrich More harbour seal SAC shown by the hashed area (information accessible via the NBN Gateway) Figure 47. The number and distribution of harbour seals counted during SMRU thermal imaging surveys between August 2007 and 2009 (taken from Duck and Thompson 2009) Figure 48. Foraging areas used by radio tagged harbour seals from the Dornoch Firth in Summer (left) and Winter (right). The shading is related to the number of different individuals whose foraging areas overlapped each 1km square (taken from Thompson et al. 1996) Figure 49. Tracks of individual harbour seals captured in the Dornoch Firth and Loch Fleet in 2004 and 2005 (left) and density of foraging locations at sea (right) (taken from Sharples et al. 2008) Figure 50. Comparison of adult female foraging locations in 1989 (n=5) and 2009 (n=5; taken from Cordes et al. 2011). The foraging locations of each individual are represented by a different symbol (key relating symbols to individuals is shown at the bottom right of each map). The solid lines show the 50% contours for individual foraging areas as calculated by 5

7 Kernel analysis. The Dornoch Firth and Morrich More SAC is shown by the shaded area. The Loch Fleet NNR is shown by the hashed area Figure 51. Number of harbour seals at different haulout sites in the Moray Firth during moult (August) surveys (produced using data from Table 2 in Duck et al. 2010). When multiple surveys of haulout sites were carried out in one year (2004, 2005, 2006, 2007, 2008) the mean count for the site is presented. Dashed lines indicate a hypothetical trend between counts (due to a lack of data for the intervening years) Figure 52. Trends in abundance of harbour seals within the Dornoch Firth (filled triangles) and at the nearby Loch Fleet NNR (open circles; taken from Cordes et al. 2011) Figure 53. Predicted numbers of harbour seals from the Dornoch Firth SAC and Loch Fleet NNR in different 4x4km grid cells across the Moray Firth (taken from Bailey and Thompson 2011) Figure 54. Tracks of 17 harbour seals (2 of which were tagged at the Dornoch Firth and Morrich More SAC and Loch Fleet NNR, the other 15 were tagged in the northern half of Orkney, including the Sanday SAC; left) and 15 female harbour seal pups (tagged on Sanday in Orkney; right) fitted with SRDL tags (taken from SMRU Ltd 2011). Only tracks (colour coded by individual seal) of seals which entered the Pentland Firth and Orkney Strategic Area at least once are shown, but they give an idea of how widely both young and adult harbour seals range Figure 55. The number and distribution of grey seals counted during SMRU thermal imaging surveys of the Moray Firth in August 2007 and 2009 (taken from Duck and Thompson 2009) Figure 56. The estimated usage of the British marine environment by grey seals (taken from Matthiopoulos et al. 2004) Figure 57. Estimated grey seal total (at sea and hauled out) usage around the proposed MORL/BOWL development sites. White contours show standard deviation from mean usage as a measure of uncertainty Figure 58. Estimated grey seal at sea usage around the proposed MORL/BOWL development sites. White contours show standard deviation from mean usage as a measure of uncertainty Figure 59. Estimated grey seal hauled out usage around the proposed MORL/BOWL development sites. White contours show standard deviation from mean usage as a measure of uncertainty Figure 60. Maximum monthly counts of grey seals at inner Moray Firth haulout sites from January 1988 to August 1990 (taken from Thompson et al. 1996) Figure 61. Grey seal pup production estimates for North Sea colonies from 1960 to 2009 (produced using data from Duck and Morris 2010) and for the Duncansby Head to Helmsdale area from 2005 to 2009 (produced using data from Duck and Mackey 2006; Duck and Mackey 2007; Duck and Mackey 2008; Duck 2009; Duck and Morris 2010). The Duncansby Head to Helmsdale colony was counted by boat prior to 2005, but only single counts were made and therefore no pup production estimates are available. Note that the surveys described here do not account for the small groups of seals breeding in caves along the Sutherland and Caithness coasts Figure 62. Mean daily locations of five grey seals satellite tagged in the Moray Firth illustrating the long range movements made by four of the five individuals (taken from Thompson et al. 1996)

8 Figure 63. Foraging areas of three satellite tagged grey seals (taken from Thompson et al. 1996) Figure 64. The extent of grey seal pup (n=39) movements from the breeding sites where they were tagged (taken from Russell 2011). The tracks are colour coded by tagging location (see legend). The solid black line shows a 100km buffer zone around the BOWL and MORL wind farm sites Figure 65. Tracks of grey seals (aged one year and above; n=65) which, at least once while they were tagged, entered a 100km buffer zone around the proposed BOWL and MORL wind farm sites. Each colour represents a different individual (taken from Russell 2011) Figure 66. Tracks of grey seals (aged one year and above; n=65) which, at least once while they were tagged, entered a 100km buffer zone around the proposed BOWL and MORL wind farm sites. Each colour represents a different individual. This Figure shows the same information as Figure 65 but is magnified to show the Moray Firth in more detail (taken from Russell 2011) List of Tables Table 1. Marine mammal species occurring in the Moray Firth Table 2. The datasets reviewed and collected by the University of Aberdeen (taken from Thompson and Brookes 2011) Table 3. Abundance estimates for the Scottish East Coast bottlenose dolphin population Table 4. Density estimates for dolphins in each of the University of Aberdeen 2010 aerial survey areas (Figure 4; Thompson and Brookes 2011) and for bottlenose dolphins in SCANS II Block J (Moray Firth, Orkney and Shetland; SCANS II 2008) Table 5. Abundance estimates (from mark recapture analysis of photographs) for the Moray Firth SAC core study area for the last decade Table 6. Abundance and density estimates for harbour porpoises in the north western North Sea and Moray Firth Table 7. Abundance estimates for minke whales in the north western North Sea Table 8. Abundance estimates for white beaked dolphins in the north western North Sea. 54 Table 9. Summary of telemetry data for harbour seals tagged in the Dornoch Firth and Morrich More SAC and Loch Fleet NNR (Thompson et al. 1994; Sharples et al. 2008; Cordes et al. 2011)

9 1 Introduction The Moray Firth is a roughly triangular inlet of the North Sea situated north and east of Inverness. It is the largest Firth in Scotland stretching from Duncansby Head in the north to Fraserburgh in the east and the Beauly Firth in the west. The Firth has more than 800km of coastline, much of which is cliffs. A number of rivers flow into the Moray Firth including the Ness, the Findhorn and the Spey. There are three main inlets in the Firth, the Beauly, Cromarty and Dornoch Firths. The Moray Firth is one of the most reliable places in the UK for observing marine mammals close to shore. It also contains the Beatrice oil field (58 08 N 3 06 W) and is the planned site for two deep water wind farms. It supports a fishing industry much of which focuses on scallops (Pectinidae) and Norway lobsters (Nephrops norvegicus). Much of the inner Moray Firth (Figure 42) is designated as a Special Area of Conservation (SAC; Figure 1) for bottlenose dolphins and sandbanks which are slightly covered by sea water all the time. The Moray Firth is home to two cetacean species which are present year round (bottlenose dolphin and harbour porpoise), one species which is present seasonally (minke whale), and seven other species whose presence is either unknown or occasional (Reid et al. 2003). It is also home to two seal species which are present year round (harbour seal and grey seal). These marine mammal species are listed in Table 1 along with information on their biology and designations. Duncansby Head Dunbeath Loch Fleet Dornoch Firth Helmsdale Spey Bay Cromarty Firth Findhorn Lossiemouth Fraserburgh Ardersier Inverness Firth Beauly Firth Figure 1. A map of the Moray Firth (limit shown by the dashed line) showing the bathymetry and the outer boundary of the Moray Firth SAC (taken from Thompson et al. 2010a and adapted). 8

10 Common name Latin name EU Habitats Directive SAC designations EPS Moray Firth specific information on: Annex II 1 Annex IV 2 Annex V 3 within the Moray Firth Abundance Distribution Presence Bottlenose dolphin Tursiops truncatus 193 (95% PI= ) 4 Coastal Year round Harbour porpoise Phocoena phocoena 10,254 (CV=0.36) 5 Coastal and offshore Year round Minke whale Balaenoptera acutorostrata 835 (CV=1.02) 5 Coastal and offshore Seasonal Common dolphin Delphinus delphis Unknown Coastal and offshore Unknown White beaked dolphin Lagenorhynchus albirostris 682 (CV=0.86) 5 Offshore Unknown Risso s dolphin Grampus griseus Unknown Offshore Unknown Fin whale Balaenoptera physalus Unknown Coastal and offshore Occasional Humpback whale Megaptera novaeangliae Unknown Coastal and offshore Occasional Killer whale Orcinus orca Unknown Coastal and offshore Occasional Long finned pilot whale Globicephala melas Unknown Coastal and offshore Occasional Harbour (or common) seal Phoca vitulina ~1,000 6 Coastal and offshore Year round Grey seal Halichoerus grypus 1,098 7 Coastal and offshore Year round Table 1. Marine mammal species occurring in the Moray Firth. 1 Species requiring designation of Special Areas of Conservation 2 Species in need of strict protection 3 Species whose taking from the wild can be restricted by European law 4 The best available estimate of the number of bottlenose dolphins in the Scottish East Coast population (Cheney et al. 2011a) 5 Estimate of the number of animals in SCANS II Block J (Moray Firth, Orkney and Shetland; andrews.ac.uk/scans2/inner finalreport.html) 6 The number of harbour seals counted between Montrose and Cape Wrath (SCOS Main Advice 2009) 7 The 2008 grey seal pup production estimate for Duncansby Head to Helmsdale (Duck and Morris 2010) 9

11 2 Designations and legislation 2.1 European Protected Species (cetaceans) All cetaceans are European Protected Species (EPS) meaning that they are protected by the EU Habitats Directive and are listed in Annex IV (species of community interest in need of strict protection). This Directive was translated into law in Scotland under the Conservation (Natural Habitats, &c.) (Scotland) Regulations These Regulations supersede the Wildlife and Countryside Act 1981 which first offered protection to cetaceans within UK waters within the 12nmile limit. The Conservation (Natural Habitats, &c.) Amendment (Scotland) Regulations 2007 further strengthen the 1994 Act. Regulation 39 states that it is an offence to deliberately or recklessly capture, injure or kill a wild animal of an EPS. It is also an offence to disturb an EPS 8. Outwith the 12nmile limit, the same protection is afforded to EPS by the Offshore Marine Conservation (Natural Habitats, &c.) Regulations In addition, cetaceans are listed as UK Biodiversity Action Plan priority species Seals Both seal species which occur in Scotland are protected under Part 6 of the Marine (Scotland) Act which prohibits the taking of seals except under licence. This Act supersedes all existing seal legislation e.g. the Conservation of Seals Act 1970 and the Conservation of Seals (Scotland) Order Scottish Ministers may grant a licence authorising the killing or taking of seals under certain circumstances (e.g. for the protection of fisheries or aquaculture activities, or for scientific or welfare reasons). In addition, it is now an offence to disturb seals at designated haulout sites in Scotland. Although not afforded the protection given to EPS, both harbour and grey seals are listed on Annex II of the EU Habitats Directive. This listing means that the presence of these species can result in the designation of SACs. Harbour seals are also listed as a UK Biodiversity Action Plan priority species Natura 2000 sites In addition to affording protection at a species level, European legislation also requires Member States to protect important habitats. This has led to the establishment of a network of sites that contribute to the protection of the habitats and species listed under Annexes I and II of the Directive. Bottlenose dolphins, harbour porpoises, harbour seals and grey seals are listed under Annex II, which means that the presence of these species can result in the designation of SACs (see Section and Section for those designated in the Moray Firth). SACs, combined with Special Protection Areas (SPAs; designated for birds under the EU Birds Directive), form a network of European protected sites known as Natura 2000 sites. SAC sites are chosen on the basis that they will make a significant contribution to species or habitat conservation. Where SACs have been established, care must be taken to (1) avoid deterioration of the habitats of the qualifying species or significant disturbance to the qualifying species, thus ensuring that the integrity of the site is maintained and the site makes an appropriate contribution to achieving favourable conservation status for each of the qualifying features and (2) ensure for the qualifying species that the following are established then maintained in the long term: Population of the species as a viable component of the site; distribution of the species within site; distribution and extent of habitats supporting the

12 species; structure, function and supporting processes of habitats supporting the species; no significant disturbance of the species. 3 Data collection methods and sources 3.1 Cetaceans Surveys Much of the survey data on cetaceans in the Moray Firth has been collected by the University of Aberdeen (UoA) through previous and ongoing work carried out in relation to the Beatrice Demonstrator Project and assessments of the impact of seismic surveys. There are two main datasets, one collected using boat based line transect surveys and one collected using aerial linetransect surveys. Additional survey data were available from the boat based seabird and marine mammal surveys that have been carried out by the Institute of Estuarine and Coastal Studies (IECS) 12 on behalf of BOWL (hereafter referred to as BOWL boat surveys, and the resulting data as BOWL boat survey data ). The University of Aberdeen data were collected during April October while the BOWL boat surveys were carried out year round. Each of the datasets was collected using broadly similar line transect methods and effort data were collected in the form of transect distance surveyed. Location, species and number of animals sighted were recorded, but the number and experience of observers did vary between surveys. No deviation from the survey track line was made when animals were sighted Habitat association modelling In general, habitat characteristics (such as depth, slope, sediment type etc.) can be used to predict the distribution and density of species in areas for which there are a lack of data (as long as there are enough data collected in other areas). These methods are as applicable to seals as they are to cetaceans Density estimates Standard procedures available in the program Distance are often used to calculate density and abundance Passive acoustic monitoring (PAM) Passive acoustic monitoring (PAM) is an increasingly useful tool for providing fine scale spatial data on cetacean distribution and temporal trends in occurrence within key areas. POrpoise Detectors (PODs) 13 continuously monitor within the khz range for possible cetacean echolocation clicks and record the centre frequency, frequency trend, duration, intensity and bandwidth of each detected click. They were originally designed to study harbour porpoises but can be programmed to detect a range of species. It has been estimated that harbour porpoises can be detected at distances of approximately 200m, while bottlenose dolphins can be detected up to 1200m away. Battery power limits the life of each device to several months; at this point devices will

13 require servicing. An accompanying software program is used to post process the recovered data, detect characteristic click trains and remove noises from other similar sources such as boat sonar. Resulting data on the number of cetacean click trains recorded in each minute can be used to determine the presence or absence of target species in the area where the POD was deployed, and the timing and duration of encounters with target species (Thompson and Brookes 2011 see Appendix 1). It should be noted that these detections could be of dolphins of any species PODs can differentiate between dolphin and porpoise clicks, but not clicks made by different dolphin species. Other devices, such as Ecological Acoustic Recorders (EARs) 14, comprise hydrophones (underwater microphones) that monitor over the frequencies used by the different species in the area of interest and suitably broadband recording systems. Unlike PODs, these devices can be used to detect other noises made by cetaceans, such as whistles. The frequency spectrum of recorded whistles can be examined (using whistle detectors and whistle contour classifiers) and the whistles attributed to species (see Section 4.1.6). Towed hydrophone arrays are also commonly used on surveys, depending on the nature of the study. 3.2 Seals Surveys Harbour seals Over the last 20 years, counts of harbour seals have been carried out at haulout sites throughout the Moray Firth by the University of Aberdeen ( ) and SMRU ( ) during both the breeding season (mid June to mid July) and moult (August). Harbour seals tend to spend longer at haulout sites during the moult and this is when the greatest and most consistent numbers are found ashore (Duck et al. 2008). Counts have either been made from land using a telescope or from the air using either thermal imaging or conventional photography. Where seals haul out onto sandbanks and are relatively easy to locate, surveys are normally carried out using a fixed wing aircraft and hand held oblique digital photography. Where they haul out onto rocky and seaweed shores and are well camouflaged, they are surveyed by helicopter using a thermal imaging camera. Moult surveys of harbour seals around the Scottish coast are usually carried out by SMRU on an approximately five yearly cycle, but the Moray Firth and Firth of Tay are surveyed annually (Duck et al. 2010). However, since the decline in harbour seal numbers in the North Sea (Lonergan et al. 2007), surveys in other areas of particular interest or importance have been carried out more often than every five years. Not all individuals in the population are counted during surveys because at any one time a proportion will be at sea. The survey counts are normally presented as minimum estimates of population size. Telemetry based, mark recapture estimates suggest that approximately 60 70% of the population are counted during the moult surveys, leading to an estimate for the total British population of 40,000 46,000 animals (SCOS Main Advice 2010) Grey seals Every year between September and December, SMRU conducts aerial surveys of the major grey seal breeding colonies in Scotland to determine the number of pups born. Normally, these main breeding

14 colonies are surveyed between four and six times during the breeding season (at approximately 10 to 12 day intervals) but a number of smaller, or more difficult to survey, colonies are surveyed three times during the breeding season. Approximately 40 additional colonies are surveyed once during the breeding season, on a two to four year rotation. Routine searches are also made for new colonies. Surveys are carried out using a light, twin engine survey modified aircraft (with the exception of the South Ronaldsay and Shetland colonies which are surveyed from the ground). A large format camera is mounted in the floor of the plane and takes high resolution images of the areas used by breeding seals. Numbers of pups born (pup production) at the regularly surveyed colonies is estimated from counts derived from the aerial photographs using a model of the birth process and development of pups. A lognormal distribution is fitted to colonies surveyed four or more times and a normal distribution to colonies surveyed three times (Duck and Morris 2010). SMRU reports pup production for the whole of the UK annually through the Special Committee on Seals (SCOS) which formulates scientific advice on matters related to the management of seal populations to government on behalf of the Natural Environment Research Council (NERC). SCOS reports are available to the general public via the SCOS section of the SMRU website 15. Grey seals are also counted during the harbour seal moult surveys carried out by SMRU (see Section ) Telemetry The distribution of UK seals at sea has predominantly been studied using VHF radio, SRDL (Satellite Relay Data Logger) and GPS phone tags developed by SMRU. Triangulation is used to locate animals fitted with VHF radio tags; SRDL and GPS phone tags collect and relay information that can be used to determine the location of individuals as well as individual dive information. Seals are captured on or around their haulout sites using a variety of methods. All activity relating to catching and working with seals in the UK is licensed by the Home Office in accordance with the Animals (Scientific Procedures) Act The primary method of tag attachment is by gluing the tag with rapid setting epoxy resin onto the fur on the dorsal neck region. The tag will certainly detach during the annual moult (for grey seals January March; for harbour seals August September). However typical tag deployment is seldom greater than six months. These tags transmit data on seal locations with the duration of data varying between individual deployments andrews.ac.uk/pageset.aspx?psr=411 13

15 4 Species accounts 4.1 Bottlenose dolphin Bottlenose dolphins have a worldwide distribution and occur in both tropical and temperate seas in both hemispheres. Recent genetic, morphologic and physiologic studies suggest that revision of the genus may be necessary to acknowledge differences between forms from different oceans, as well as differences between forms in inshore vs. offshore habitats within ocean basins (Wells and Scott 2009). Along the Atlantic seaboard of Europe bottlenose dolphins are locally common off the coasts of Spain, Portugal, France, Ireland, Wales and Scotland (especially the Moray Firth; Figure 2). They also occur further offshore in deep waters of the North Atlantic (Reid et al. 2003). Figure 2. Sightings of bottlenose dolphins around the UK ( ; taken from Reid et al. 2003) Distribution Bottlenose dolphins using the Moray Firth range as far afield as the Firths of Forth and Tay, and sometimes even as far south as the Tyne Estuary (Wilson et al. 2004; Thompson et al. 2011). The bottlenose dolphins throughout this range are referred to as the Scottish East Coast population. 14

16 There is also evidence that bottlenose dolphins which are regularly photographed and identified off the west coast of Scotland occasionally visit the Moray Firth (Cheney et al. 2011a) Visual survey data Available data from existing cetacean surveys in the Moray Firth were reviewed by the University of Aberdeen using information from peer reviewed journals and the grey literature and unpublished data collected by various groups (Table 2; Thompson et al. 2010a). These data represent observations made over a period of 30 years, from 1980 to 2010 (Table 2), although coverage of the outer Moray Firth is patchy both spatially and temporally. Almost all bottlenose dolphin sightings were within 15km of the coast in the inner part of the Moray Firth SAC or the coastal strip along the southern Moray Firth coast. There were a few records of bottlenose dolphins in the outer Moray Firth most sightings of dolphins in these offshore waters were of common, white beaked or Risso s dolphins (Figure 3). Two types of visual surveys were carried out over the BOWL site in 2010 aerial surveys conducted by the University of Aberdeen (Figure 4) and the BOWL boat surveys (Figure 5). All bottlenose dolphins encountered during the University of Aberdeen aerial surveys were located in the inner Moray Firth or within 10km of the southern Moray Firth coast (Figure 6). No bottlenose dolphins were encountered offshore. All offshore dolphins were other species (common, whitebeaked and Risso s dolphins) or could not be identified to species (two sightings). The BOWL bird and marine mammal boat surveys commenced in October 2009 and were carried out as part of a 21 month programme to support the BOWL OWF environmental impact assessment (EIA). Ten dolphin encounters were recorded during the BOWL boat surveys (Figure 5), four of which were identified as bottlenose dolphins (Figure 7). Apart from a couple of encounters, neither of which was within the BOWL site, this is the only time during the last 30 years that bottlenose dolphins have been encountered outwith the coastal strip in the inner part of the Moray Firth SAC or along the southern Moray Firth coast. While it is possible that bottlenose dolphins occur at the BOWL site, all other data collected over the last 30 years suggest that this is unlikely. In addition, only one encounter with any of the other dolphin species more commonly seen offshore was recorded by the BOWL boat survey observers. This suggests that the BOWL boat survey records of bottlenose dolphins may be a result of species mis identification. Unfortunately the data were not collected in a way that could later be verified (e.g. good quality photographs, broadband acoustic recordings). Additionally, without good quality photographs of dorsal fins for photo ID purposes, it is not possible to ascertain whether these animals were bottlenose dolphins from the Moray Firth SAC. The importance of using experienced observers to ensure data are of a high standard cannot be overstated. Collecting data in a way that can later be verified is also important (Grellier 2010). There were insufficient data to produce habitat association models for each individual dolphin species, however the University of Aberdeen were able to use these survey data in classification trees to assess the likely species identity of dolphins that may be encountered in different parts of the Moray Firth, given its habitat characteristics (Thompson and Brookes 2011). This presence only method does not account for effort so, used alone, it cannot provide a prediction of the number of animals that might be found in an area. However, the approach does show the likely species composition in different areas if dolphins were present. The analysis was run twice, once with all of the data (Figure 8) and once excluding the BOWL boat survey data which, given the offshore location of the survey area, contained an atypically large number of bottlenose dolphin sightings relative to sightings of other species (Figure 9). Including the series of encounters made during the BOWL boat surveys meant that the model predicted a higher likelihood that dolphins encountered in this 15

17 specific offshore area are likely to be bottlenose dolphins (Figure 8; Figure 9). Given uncertainties over the reliability of species identification from the BOWL boat surveys, and supporting evidence from acoustic work (see Section 4.1.6), predictions from the second model which exclude these data provide a more robust picture of the likely species composition of groups of dolphins encountered in different parts of the Moray Firth (Figure 9). The results suggest that any dolphins encountered along the coastal strip of the Moray Firth are most likely to be bottlenose dolphins, but those encountered in offshore areas are, in general, more likely to be other species (Figure 9). Data on the likely presence of bottlenose dolphins vs. other dolphin species are presented separately in Figure 10. By combining information on the likely species composition in different areas if dolphins were present (Figure 9 and Figure 10) with the density estimate of dolphins per km 2 (i.e dolphins per 4x4km grid cell; see Table 4), the number of bottlenose dolphins in each 4x4km grid cell was predicted (Figure 11; Thompson 2011a see Appendix 2). As stated in the report, it must be recognised that this density distribution remains very conservative when focussing on impacts in offshore areas and along the northern coast of the Moray Firth SAC, as this approach tends to underestimate the number of animals occurring in the inner Moray Firth and along the southern Moray Firth coast. i.e. whilst the total number of animals represented in this Figure (213) appear reasonable given current estimates of population size, these animals are predicted to be much more widely distributed outside their core areas than would be expected given other data on the number of animals typically found in the inner part of the Moray Firth SAC and along the Moray coast (e.g. Bailey and Thompson 2009; Thompson et al. 2011). Dataset Year BOWL boat survey 2010 JNCC Seabirds at Sea JNCC seismic MMO MORL 2010 Crown Estate UoA AFEN UoA boat 2009 UoA aerial 2010 UoA SAC UoA Photo ID Table 2. The datasets reviewed and collected by the University of Aberdeen (taken from Thompson and Brookes 2011). 16 Data collected as part of a project funded by the Atlantic Frontier Environmental Network, a grouping of oil and gas companies and UK Government departments which aims to ensure sound management and regulation of oil and gas activities in the Atlantic Frontier. 16

18 Figure 3. Sightings of dolphins from the datasets shown in Table 2 (taken from Thompson and Brookes 2011). B A Figure 4. Map of track lines from the University of Aberdeen s August and September 2010 aerial surveys of the outer Moray Firth (taken from Thompson and Brookes 2011). 17

19 Figure 5. Map of the survey track followed on the BOWL boat surveys (taken from Thompson and Brookes 2011). Figure 6. Sightings of dolphins made during the University of Aberdeen s 2010 aerial surveys of the outer Moray Firth (taken from Thompson and Brookes 2011). 18

20 Figure 7. Sightings of dolphins made during the BOWL boat surveys between April and October 2010 (taken from Thompson and Brookes 2011). Figure 8. Prediction of the dolphin species composition within each 4x4km grid cell using all data (taken from Thompson and Brookes 2011). 19

21 Figure 9. Prediction of the dolphin species composition within each 4x4km grid cell using all data except the BOWL boat survey data (taken from Thompson and Brookes 2011). 20

22 Figure 10. Prediction of the likelihood that dolphins encountered in each 4x4km grid cell are likely to be bottlenose dolphins. Data are as for Figure 9, but presented as bottlenose dolphins (black portion of pie chart) vs. all other species (taken from Thompson and Brookes 2011). 21

23 Figure 11. Predicted number of bottlenose dolphins in each 4x4km grid cell (taken from Thompson 2011a) Acoustic survey data An extensive array of PODs was deployed across the Moray Firth by the University of Aberdeen during their DECC funded study in 2009 (data recovered from 56 deployments) and 2010 (data recovered from 60 deployments; Thompson et al. 2010a; Thompson and Brookes 2011) and their MORL/BOWL funded studies in 2009 to 2011 (data recovered from 24 deployments; Thompson and Brookes 2011; Figure 12). Dolphins were detected on all PODs at least once during their deployments, but the proportion of days on which they were detected varied considerably (Figure 12). It is not possible to use click characteristics to determine which species of dolphins have been detected, and it is likely that detections in different areas represent different species. Dolphins were detected regularly in the inner Moray Firth and along the southern Moray Firth coast, but detections were less frequent in the central part of the Moray Firth. However, dolphin detections increased again at more offshore locations, including those around the BOWL site (Figure 12). The longest time series of passive acoustic monitoring data was available from the Beatrice Demonstrator site, where devices were deployed by the University of Aberdeen between August 2005 and December After a break in studies during 2008, devices were again deployed at this site in May 2009 and data collection is anticipated to continue until at least Autumn There have been some gaps in the time series due either to equipment loss or failure but these data provide a unique opportunity to explore longer term temporal change in the occurrence of dolphins at an offshore site (Thompson and Brookes 2011). Overall, dolphins were detected only rarely (<6% of deployment days). On those days that dolphins were detected, they were recorded for a median of 1 hour (IQ range=1 1; Figure 13). 22

24 Passive acoustic monitoring data are also available from two sites within the BOWL development area for a period of almost two years, and from three additional sites for the final nine months of the study (Thompson and Brookes 2011). Inspection of these data indicates that dolphin detections were low (Figure 17) and, on the days that they were present, they were generally detected for only 1 or 2 hours a day (Figure 13). Broadband sound recordings were made (using EARs) at the BOWL and MORL sites in 2010 with the purpose of identifying whether the dolphin activity recorded by the PODs was due to the presence of bottlenose dolphins or other dolphin species. Twenty two whistle classification events were recorded over 88 days, but none were attributed to bottlenose dolphins (Figure 23). These results are detailed in Section and support previous evidence that bottlenose dolphins are generally not present at the BOWL site, at least during the July October sampling period used in the acoustic study. Chanonry Point Macduff Figure 12. Spatial variation in the occurrence of dolphins in April October of 2009 and The pie charts show the proportion of days that dolphins were detected on PODs at each sampling location (taken from Thompson and Brookes 2011). 23

25 Figure 13. Frequency histogram showing the number of hours per day that dolphins were detected on PODs deployed at the Beatrice Demonstrator site from (left) and within the BOWL site from (right) (taken from Thompson and Brookes 2011). This Figure only includes those days on which any animals were detected In the southern Moray Firth (in relation to the cable route) The University of Aberdeen s POD data show that dolphins are present along the southern Moray Firth coast (from Chanonry Point in the inner Moray Firth along the coast to the east of Macduff) on a high proportion of days ( 75% of days at half of the POD deployment sites), at least between April and October (Figure 12). Visual survey data (see Table 2 for data sources) show that dolphins sighted along this coast were almost exclusively bottlenose dolphins (Figure 3; Figure 6). When the likely species composition in different areas of the Moray Firth was assessed, results showed that any dolphins encountered along the coastal strip of the Moray Firth were most likely to be bottlenose dolphins while those encountered in offshore areas were more likely to be other species (Thompson and Brookes 2011; Figure 9; Figure 10). Additional surveys (to those listed in Table 2) have been carried out between May and October along the southern shore of the outer Moray Firth (between Lossiemouth and Fraserburgh) by the Cetacean Research and Rescue Unit 17 (CRRU; Figure 14). The location of dolphin sightings made between 2001 and 2008 are shown in Figure 15. Four dolphin species were seen but 94% of dolphin sightings made were of bottlenose dolphins (Robinson et al. 2007). However, bottlenose dolphins were only encountered on the survey route closest to the shore (Figure 14; Figure 15) in shallow waters rarely exceeding 25m depth (Robinson et al. 2007). Furthermore, they appeared to have a preference (see peak in encounter rate; Figure 16) for Spey Bay, at the western end of the survey area directly adjacent to the SAC (Culloch and Robinson 2008; Figure 16). Knowing which months these sightings occurred would aid understanding of why Spey Bay appears to be a preferred area. The preference may be food related: the River Spey supports a spawning population of Atlantic salmon, a known prey of bottlenose dolphins in this area (Harding Hill 1993; Santos et al. 2001)

26 Spey Bay Whitehills Lossiemouth Fraserburgh Figure 14. CRRU survey tracks extending from Lossiemouth (in the west) to Fraserburgh (in the east; taken from Thompson et al. 2010a). Figure 15. Dolphin sightings made by the CRRU between 2001 and 2008 (taken from Thompson et al. 2010a). 25

27 Figure 16. Bar chart showing the distribution of bottlenose dolphin encounters across the survey area (from Lossiemouth to Fraserburgh). The dark grey bars represent the number of visits to each 1km section of the coast (left y axis). The encounter rate (right y axis) is shown in pale grey (taken from Culloch and Robinson 2008) Seasonal variation Visual and acoustic (POD) data from the Scottish East Coast Work carried out by Wilson et al. (1997) in the inner Moray Firth (Figure 42) in the early 1990s showed that although bottlenose dolphins were sighted in all months of the year, there were consistent seasonal fluctuations in the number of individuals present numbers were low in Winter and Spring, and peaked in Summer and Autumn. Area use also changed with season the outer part of the inner Moray Firth was used for most of the year while areas closer to the head of the Firth were only used seasonally. The issue of reduced dolphin presence in Winter was re visited by Cheney et al. (2011b), who found that the pattern documented in the 1990s appeared to have been conserved to the present time. Furthermore, other areas of high dolphin occurrence in the outer Moray Firth (Figure 42) and eastern coasts to the south (Spey Bay, Aberdeen and St Andrews Bay) also appeared to be used less in Winter. No new areas were discovered that were used by dolphins in Winter but not in Summer. It should be noted, however, that the power of this study to detect significant new areas of use was low, particularly in offshore areas. Dolphins continued to use their entire known Summer range in Winter, but with apparently lower rates of occupancy. The most obvious explanation is that the population increases its range in Winter to other areas that are, as yet, unknown. Without some indication of where those areas might be, it has proved extremely difficult to target sufficient search effort to find them (Cheney et al. 2011b). There is another possibility: dolphins maintain their Summer range but change their behaviour so that they are harder to detect. The most obvious way for this to occur would be by increasing their group sizes. This would produce lower visual sightings rates and lower rates of dolphin positive days on the POD recorders. There is some evidence that dolphins increase their group sizes in Winter but the mark recapture estimates of dolphin numbers in the inner Moray Firth in Winter suggest that this factor alone cannot entirely explain the apparent reduced occupancy rates (between around 25% (2007) and 45% (2006) fewer individual dolphins used the SAC in Winter compared with Summer; Cheney et al. 2011b) Acoustic (POD) data from the BOWL site Passive acoustic monitoring data are available from two sites within the BOWL development area for a period of almost two years, and from three additional sites for the final nine months of the study (Thompson and Brookes 2011). Dolphin detections (circles) remained relatively low throughout the year, with no obvious seasonal pattern (Figure 17). 26

28 100 % days detected each month J A S O N D J F M A M J J A S O N D J F M Figure 17. Monthly values for the % of days that dolphins (black circles) and porpoises (red squares) were detected by PODs deployed within the BOWL area (taken from Thompson and Brookes 2011) In the southern Moray Firth (in relation to the cable route) Bottlenose dolphins use their entire known Summer range in Winter but, with apparently lower rates of occupancy. For example, Spey Bay, an area of high dolphin occurrence, appeared to be used less in Winter (Cheney et al. 2011b). Bottlenose dolphins were sighted in all of the months in which CRRU surveys were carried out (May to October; Figure 18), with no significant difference in sightings rates between months (Culloch and Robinson 2008). However, acoustic detections of dolphins on PODs deployed along the southern Moray Firth coast (Figure 30) declined from mid July to November (Thompson et al. 2010a). Year round POD data were available from two sites along the southern Moray Firth coast from the Summer of 2009 to the Spring (Spey Bay; n=628 days) or Summer (Lossiemouth; n=741 days) of 2011 (Figure 19; Thompson 2011). Dolphins were detected on 65% of days at Spey Bay and 63% of days at Lossiemouth. Visual sightings in these areas suggest that most of these detections were of bottlenose dolphins (Figure 15). There appeared to be seasonal patterns in dolphin detections at Spey Bay, with peaks in Summer and early Winter in both years (although it is not possible to tell if this pattern is real without a statistical analysis); this was less evident at Lossiemouth (Figure 20). Median values for the number of hours per day that dolphins were detected were broadly similar at the two sites (Figure 21). Typically, dolphins were detected for more hours per day at these coastal sites (Figure 21) than at the offshore sites (Figure 13). 27

29 Figure 18. Frequency histogram of encounter rate for bottlenose dolphins across each survey month (+/ 95% confidence intervals; taken from Culloch and Robinson 2008). Lossiemouth Spey Bay Figure 19. Map of the Moray Firth showing the location of the two POD sites at Lossiemouth (left) and Spey Bay (right) from 2009 to 2011 (taken from Thompson 2011). 28

30 12 12 Median No of Hours Detected Per Day Median No of Hours Detected Per Day J A S O N D J F M A M J J A S O N D J F M A M J J J A S O N D J F M A M J J A S O N D J F M A M J J Figure 20. Monthly variation in the median number of hours that dolphins were detected at the Spey Bay (left) and Lossiemouth (right) POD sites from 2009 to 2011 (taken from Thompson 2011) Frequency Frequency Number of hours detected Number of hours detected Figure 21. The median number of hours per day that dolphins were detected at the Spey Bay (left) and Lossiemouth (right) POD sites (taken from Thompson 2011) Abundance The size of the Scottish East Coast bottlenose dolphin population (and various subsets thereof) has been estimated regularly since the University of Aberdeen research started on a rigorous basis in Some estimates (e.g. Wilson et al. 1999; Cheney et al. 2011a) are of the whole population, while others (Corkrey et al. 2008; Culloch and Robinson 2008; Quick 2006) use data collected in only part of the population s range. The SCANS II estimate (SCANS II 2008) is based on data collected over an area wider than the population s range (including Orkney and Shetland). The available estimates are summarised in Table 3. The 2006 estimate of 193 animals (95% PI ; Cheney et al. 2011a) is considered to be the best available estimate of the number of bottlenose dolphins in the Scottish East Coast population. Although it is not the most recent estimate, it is more precise than the most recent (2007) one (Cheney et al. 2011a). Furthermore, the 2006 estimate was calculated using data from every research group that carries out photo identification (photo ID) work on bottlenose dolphins off the Scottish East Coast. This estimate is higher than the estimate of 129 animals (95% CI ) calculated for 1992 data from the Moray Firth only (Wilson et al. 1999). However, it is important not to over interpret the significance of this difference because of the difference in the area over which the data were collected (Moray Firth vs. Scottish East Coast) and the methodology used for 29

31 estimation (1992=the software program CAPTURE 18 ; 2006=multi site Bayesian mark recapture analysis). Nevertheless, the difference does suggest that not all of the animals in the East Coast population use the Moray Firth (noted by Thompson et al. 2006; Thompson et al. 2009; Cheney et al. 2011a). It is worth noting that while the SCANS II estimate for bottlenose dolphins in survey Block J covers a wider area (Moray Firth, Orkney and Shetland), it is of the same order of magnitude (100s of animals) as the other estimates (Table 3; SCANS II 2008). Culloch and Robinson (2008) estimated the abundance of bottlenose dolphins in the southern outer Moray Firth in The estimates ranged from 61 (95% CI=48 74) in 2004 to 108 (95% CI=87 129) in These estimates suggest that the southern outer Moray Firth is an important area for a large percentage of the population, with a substantial number of individuals using the area during each of the survey years. Quick (2006) estimated the number of bottlenose dolphins in the St Andrews Bay area in 2003/2004 using two different methods. The best estimates were 89 (using the software program CAPTURE) and 112 (using Bayesian mark recapture analysis) individuals. Both estimates suggest that a substantial proportion of the Scottish East Coast population uses the St Andrews Bay area, at least in Summer. Although there is no apparent trend in abundance, there is no doubt that the Scottish East Coast bottlenose dolphin population is small and isolated. Year Area Estimate 95% Confidence 1 or Reference Probability 2 Interval (CI or PI) 1992 Moray Firth Wilson et al SCANS II Block J (Moray Firth, ,888 1 SCANS II 2008 Orkney and Shetland) 2006 Scottish East Coast Cheney et al. 2011a Table 3. Abundance estimates for the Scottish East Coast bottlenose dolphin population In the southern Moray Firth (in relation to the cable route) Bottlenose dolphin abundance in the southern Moray Firth was estimated annually between 2001 and 2004 using mark recapture analysis (Figure 22). Although estimates were variable (e.g. 108 individuals in 2003 compared to 61 individuals in 2004; Culloch and Robinson 2008), between 32 and 56% of the population, which numbered 193 individuals in 2006 (Cheney et al. 2011a; Table 3), used the southern Moray Firth coast in each of the survey years. This suggests that that the southern Moray Firth is an important area for this population. The median number of hours per day that dolphins were detected on PODs deployed in the southern Moray Firth was 1 (interquartile range 0.5 2; Thompson et al. 2010a)

32 Abundance estimate (+/ 95% CI) Year Figure 22. Annual abundance estimates (+/ 95% confidence intervals) of dolphins using the southern Moray Firth coast showing the degree of inter annual variability (plotted using data from Culloch and Robinson 2008) Density The SCANS II density estimate for bottlenose dolphins in Block J (Moray Firth, Orkney and Shetland; bottlenose dolphins per km 2 ) is similar to the densities estimated by the University of Aberdeen in their offshore aerial survey blocks (Block A: 0.012; Block B: dolphins per km 2 ; Table 4). Dolphin density was greatest around the coast (0.259 dolphins per km 2 ; Table 4). Multiplication of the density estimate from the University of Aberdeen s aerial survey Block B (0.018) by the area of the BOWL site (121km 2 ) gives an estimate of two dolphins in the BOWL site (Thompson and Brookes 2011). It was only possible to use these density estimates to estimate the combined abundance of all dolphin species. Nevertheless, viewed in conjunction with results from the classification tree (Figure 8; Figure 9), these analyses suggest that the numbers of any species of dolphin, and particularly bottlenose dolphin, in the vicinity of the proposed wind farm site are likely to be low. Area All areas surveyed Block A Block B Coast SCANS II Block J (Moray Firth, Orkney and Density (animals per km 2 ) Coefficient of variation (CV) Reference Thompson and Brookes 2011; see Figure 4 for areas SCANS II 2008 Shetland) Table 4. Density estimates for dolphins in each of the University of Aberdeen 2010 aerial survey areas (Figure 4; Thompson and Brookes 2011) and for bottlenose dolphins in SCANS II Block J (Moray Firth, Orkney and Shetland; SCANS II 2008). 31

33 4.1.5 The Moray Firth SAC The Moray Firth Special Area of Conservation (SAC) is one of two areas in the UK that have been proposed as part of the Natura 2000 series to conserve bottlenose dolphins (the other is in Cardigan Bay, Wales). The SAC extends from the inner firths to Helmsdale on the north coast and Lossiemouth on the south coast (Figure 1), and includes areas that are regularly used by the population of bottlenose dolphins occurring along the East Coast of Scotland. As a result of this designation, SNH has a responsibility to report on the condition of the SAC for the conservation status of the bottlenose dolphin population every six years. In 2004, SNH entered into a Memorandum of Agreement with the University of Aberdeen to support their photo ID studies and use these data to report on the condition of the SAC. Bottlenose dolphin abundance estimates for the Moray Firth SAC for the years can be found in two site condition monitoring reports (Thompson et al. 2006; Thompson et al. 2009; Table 5). A third site condition monitoring report is currently in preparation and draft estimates were made available by the University of Aberdeen (cited as University of Aberdeen unpublished data in Table 5). Annual estimates of the number of dolphins using a core study area (where bottlenose dolphins are most frequently sighted) within the SAC show considerable variability from year to year (Table 5). Using estimates of total population size (Table 3; Wilson et al. 1999; Cheney et al. 2011a), these data indicate that a high proportion of this population of bottlenose dolphins use the SAC. The current condition status assessment of the population is Unfavourable (recovering) (Thompson et al. 2006; Thompson et al. 2009). Previous work showed that there was a reduction in the use of the SAC by dolphins during the late 1990s, followed by a slight increase during the previous reporting period (Thompson et al. 2006). Additional data collected since 2004 further highlights the degree of inter annual variability in the number of dolphins using the corestudy area within the SAC, complicating the assessment of condition status within the 3 year reporting windows. Year Area Estimate 95% Confidence Interval Reference 2001 SAC University of Aberdeen unpublished data 2002 SAC Thompson et al SAC Thompson et al SAC Thompson et al SAC Thompson et al SAC Thompson et al SAC Thompson et al SAC University of Aberdeen unpublished data 2009 SAC University of Aberdeen unpublished data 2010 SAC University of Aberdeen unpublished data Table 5. Abundance estimates (from mark recapture analysis of photographs) for the Moray Firth SAC core study area for the last decade. Acoustic monitoring using PODs was carried out by the University of Aberdeen to support efforts to monitor the amount of time that dolphins spend in the core study area used for estimating abundance within the SAC (Thompson et al. 2009). Data from these devices have highlighted that dolphins were present in this area on almost all days during July and August, although there was also inter annual variation in the amount of time they spent in the area each day. This method promises to provide a useful indication of the level of use of the SAC that complements photo ID based abundance estimates. 32

34 4.1.6 Links between the Moray Firth SAC and the BOWL site Information on the movement of bottlenose dolphin between the Moray Firth SAC and other areas have been summarised in Section and Section (Thompson et al. 2011). Photo ID information has been used to link bottlenose dolphins using the SAC with those found off the Firths of Forth and Tay, however there are, to date, no photo ID data linking bottlenose dolphins from the Moray Firth SAC to the BOWL site. With few visual observations of bottlenose dolphins at the BOWL site, and no photographs that can be used to confirm species or individual identity (or SAC linkage), another method was used to explore whether this link existed. A new whistle classifier was constructed to distinguish bottlenose dolphins from the other dolphin species that might be encountered at the BOWL (and MORL) site, because it is not possible to differentiate between dolphin species using echolocation clicks/pod data (Booth et al. 2011). Recordings of five dolphin species (bottlenose, Risso s, white beaked, white sided and common dolphins) were sourced from around Scotland to train the whistle classifier. Two EAR deployments (E17 and E21) were made in 2010 in order to collect acoustic data from dolphins at the BOWL site; three more (A20, A22 and E16) were made at the neighbouring MORL site (Figure 23). Twenty two whistle classification events were recorded over 88 days, but none was attributed to bottlenose dolphins (Figure 23). These results support previous evidence that bottlenose dolphins are generally not present at the BOWL (or MORL) site, at least during the July October sampling period used in this study. For comparison, a sixth deployment was made in order to collect acoustic data from a core area that is regularly used by bottlenose dolphins within the Moray Firth SAC (D01; Figure 23). Twenty eight whistle classification events were recorded over 25 days, 79% of which were attributed to bottlenose dolphins. 33

35 Figure 23. The results of the classification of whistle events in the EAR data using the whistle classifier. BND=events classified as bottlenose dolphins (white) and OTHER=events classified as other species (grey). N.B. The scale of the y axis for the D01 EAR is different to the EARs deployed on the BOWL and MORL sites (taken from Booth et al. 2011). 4.2 Harbour porpoise Harbour porpoises occur primarily in temperate waters of the North Pacific and North Atlantic. They are the most numerous of the cetaceans found in north western European continental shelf waters (Reid et al. 2003). They are mainly confined to shelf waters although sightings have been made in deep water (Figure 24). 34

36 Figure 24. Sightings of harbour porpoises around the UK ( ; taken from Reid et al. 2003) Distribution Available data from existing cetacean surveys in the Moray Firth were reviewed by the University of Aberdeen using information from peer reviewed journals and the grey literature and unpublished data collected by various groups (Thompson et al. 2010a). These data represent observations made over a period of 30 years, from 1980 to 2010 (Table 2), although coverage of the outer Moray Firth is patchy in both space and time. Harbour porpoises were the most commonly encountered species in almost all studies, being seen throughout inshore and offshore waters of the Moray Firth. Habitat association modelling was based on data from five different surveys (BOWL, MORL, UoA boat, UoA aerial and UoA SAC; Table 2; Thompson and Brookes 2011). Datasets were adjusted to remove data from those cells where no habitat data were available. The results of this model were used to predict spatial variation in the relative abundance of porpoises across the Moray Firth and then scaled to absolute abundance using the density estimates obtained from the aerial line transect survey (Table 6; Figure 25). The predicted number of porpoises in each 4x4km cell (up to around 20 on some cells within the BOWL site) was based upon depth, and proportions of sand and gravelly sand within each cell, standardised for a constant unit of effort. 35

37 A second assessment of broad scale spatial variation in the occurrence of harbour porpoises across the Moray Firth was made based on data from arrays of PODs deployed during the University of Aberdeen s DECC funded study in 2009 and 2010 (Thompson et al. 2010a; Figure 26). Harbour porpoises were detected at least once on every POD during their deployments. Detection rates were generally high: porpoises were present at offshore sites on almost all sampling days (Figure 26). The lowest detection rates occurred in those coastal areas where dolphins occurred more commonly (Figure 12). Finer scale information on variation in the median number of hours that porpoises were detected each day around the BOWL and MORL sites in 2009 and 2010 is shown in Figure 27. Porpoises were present for around a quarter of each day at 14 sites, half the day at 9 sites and three quarters of the day at 3 sites. The longest time series of passive acoustic monitoring data was available from the Beatrice Demonstrator site, where devices were deployed by the University of Aberdeen between August 2005 and December After a break in studies during 2008, devices were again deployed at this site in May 2009 and data collection is anticipated to continue until at least Autumn There have been some gaps in the time series due to equipment loss or failure, but these data provide a unique opportunity to explore longer term temporal change in the occurrence of porpoises at an offshore site (Thompson and Brookes 2011). Overall, porpoises were detected on most (>93%) days that PODs were deployed at this site. On those days that porpoises were detected, they were recorded for a median of 4 hours (IQ range=2 7; Figure 28). Passive acoustic monitoring data are available from two sites within the BOWL development area for a period of almost two years ( ), and from three additional sites for the final nine months of the study (Thompson and Brookes 2011). Inspection of these data indicates that porpoises were present in the area on an almost daily basis (Figure 17) and were present for many hours each day (Figure 28). 36

38 Buckie Figure 25. The predicted number of harbour porpoises in each 4x4km cell. Values are based upon measures of relative abundance derived from habitat association modelling, scaled according to estimates of absolute abundance from aerial line transect surveys and extrapolated to other areas according to predicted relative abundance (taken from Thompson and Brookes 2011). 37

39 Figure 26. Spatial variation in the occurrence of porpoises in April October 2009 and The Figure shows the proportion of days that porpoises were detected on PODs at each sampling location (taken from Thompson and Brookes 2011). Figure 27. Fine scale spatial variation in the occurrence of porpoises in and around the BOWL (and MORL) development site. Data are from April October 2009 and Pie charts for each sampling 38

40 site represent the median proportion of time that porpoises were detected each day (taken from Thompson and Brookes 2011) Frequency Number of hours detected Figure 28. Frequency histogram showing the number of hours per day that porpoises were detected on PODs deployed at the Beatrice Demonstrator site between 2005 and 2010 (left; taken from Thompson and Brookes 2011) and within the BOWL site between 2009 and 2011 (right; taken from Thompson and Brookes 2011). This Figure only includes those days on which there was at least one detection In the southern Moray Firth (in relation to the cable route) The University of Aberdeen s POD data show that harbour porpoises are present along the southern Moray Firth coast on a high proportion of days ( 75% of days at 65% of the POD deployment sites), at least between April and October (Figure 26). They are also predicted (using measures of relative abundance derived from habitat association modelling; Thompson and Brookes 2011) to be present in large numbers along this coastline (generally >10 porpoises per 4x4km cell; Figure 25). The exception to this appears to be off Buckie (east Spey Bay), where the predicted number of porpoises per cell was lower (<5 per 4x4km cell). Additional surveys (to those listed in Table 2) have been carried out between May and October along the southern shore of the outer Moray Firth (between Lossiemouth and Fraserburgh) by the CRRU (Figure 14). Porpoises were the most commonly sighted species on these surveys; sightings made between 2001 and 2008 are shown in Figure

41 Figure 29. Harbour porpoise (circle) and minke whale (triangle) sightings made by the CRRU between 2001 and 2008 (taken from Thompson et al. 2010a) Seasonal variation In the Moray Firth A preliminary investigation of changes in the seasonal occurrence of porpoises was carried out by the University of Aberdeen who pooled POD data from (1) the outer Moray Firth (which includes offshore parts of the Moray Firth SAC) (2) the southern Moray Firth coast and (3) coastal sites within the Moray Firth SAC (Figure 30; Thompson et al. 2010a). Porpoises were detected at all sites but were most consistently detected in the outer Moray Firth. Seasonal patterns within the Moray Firth SAC and along the southern Moray Firth coast appeared to be similar with a tendency for porpoise detections to increase through the sampling period (the majority of PODs were deployed in mid July and recovered in November). The median number of hours per day that porpoises were detected was similar in the two coastal areas (southern Moray Firth coast=2.5; Moray Firth SAC=2). There was no apparent seasonal trend in the outer Moray Firth. Porpoises were consistently detected at all sites in the outer Moray Firth, typically for 5 8 hours per day (Thompson et al. 2010a). 40

42 Figure 30. Figure showing which PODs were included in the analysis of seasonal trends in detections of porpoises in the Moray Firth SAC ( ), along the southern Moray Firth coast ( ) and in the outer Moray Firth ( ; taken from Thompson et al. 2010a) At the BOWL site Passive acoustic monitoring (POD) data are available from two sites within the BOWL development area for a period of almost two years ( ), and from three additional sites for the final nine months of the study (Thompson and Brookes 2011; see Section 4.2.1). Inspection of these data indicates that porpoises were present in the area on an almost daily basis (Figure 17). However, the median number of hours that porpoises were detected appeared to vary seasonally, with peaks in the Winter and late Summer (Figure 31). Statistical analysis of this pattern has yet to be carried out. 41

43 Median No of Hours Detected Per Day J A S O N D J F M A M J J A S O N D J F M Figure 31. Monthly variation in the median number of hours per day that porpoises were detected on PODs within the BOWL development area (taken from Thompson and Brookes 2011) In the southern Moray Firth (in relation to the cable route) Harbour porpoises were sighted along the southern Moray Firth coast in all months between May and October (Robinson et al. 2007). However, there is evidence for an increase in visual encounters (from per km in May to per km in October; Robinson et al. 2007) and acoustic detections (from mid July to November; Thompson et al. 2010a) through the year. Neonatal calves were typically observed by the CRRU between May and July, consistent with the known calving period for this species in the North Sea (Lockyer 1995). The observed seasonal increase in porpoise encounters may be due to a movement inshore by lactating females and their calves, followed thereafter by males (Robinson et al. 2007). Year round POD data were available from two sites along the southern Moray Firth coast from the Summer of 2009 to the Spring (Spey Bay; n=628 days) or Summer (Lossiemouth; n=741 days) of 2011 (Figure 19; Thompson 2011). Porpoises were detected on 52% of days at Spey Bay and 51% of days at Lossiemouth. There appeared to be seasonal patterns in porpoise detections at Spey Bay with a peak in late Summer/Autumn in both years (although it is not possible to tell if this pattern is real without a statistical analysis); this was less evident at Lossiemouth (Figure 32). Median values for the number of hours per day that porpoises were detected were broadly similar at the two sites (Figure 33). Typically, porpoises were detected for fewer hours per day at these coastal sites than at the offshore sites (Figure 28). 42

44 12 12 Median No of Hours Detected Per Day Median No of Hours Detected Per Day J A S O N D J F M A M J J A S O N D J F M A M J J J A S O N D J F M A M J J A S O N D J F M A M J J Figure 32. Monthly variation in the median number of hours that porpoises were detected at the Spey Bay (left) and Lossiemouth (right) POD sites (taken from Thompson 2011) Frequency Frequency Number of hours detected Number of hours detected Figure 33. The median number of hours per day that porpoises were detected at the Spey Bay (left) and Lossiemouth (right) POD sites (taken from Thompson 2011) Abundance/density Porpoise abundance was estimated for the Moray Firth, Orkney and Shetland (Block J) during both the SCANS (1994) and SCANS II (2005) surveys (Table 6). The SCANS density estimate (of porpoises per km 2 ) is similar to the density estimated by the University of Aberdeen in the furthest offshore of their aerial survey blocks (Block B: porpoises per km 2 ; Figure 4). However, the SCANS II density estimate (0.274 porpoises per km 2 ) is lower than both the SCANS and University of Aberdeen Block B estimates. Multiplication of the density estimate from the University of Aberdeen s aerial survey Block B (0.812) by the area of the BOWL site (121km 2 ) gives an estimate of the number of porpoises present in the BOWL site of 98 individuals (Thompson and Brookes 2011). 43

45 Year Area Abundance CV Density CV Reference (animals per km 2 ) 1994 SCANS Block J (Moray 24, Hammond et al Firth, Orkney and Shetland) 2005 SCANS II Block J (Moray Firth, Orkney and Shetland) 10, SCANS II UoA aerial all areas surveyed UoA aerial Block A UoA aerial Block B UoA aerial Coast NA NA Thompson and Brookes 2011; see Figure 4 for areas Table 6. Abundance and density estimates for harbour porpoises in the north western North Sea and Moray Firth In the southern Moray Firth (in relation to the cable route) Porpoise abundance has not been estimated in the southern Moray Firth but there is information on encounter rates (Robinson et al. 2007) that can be used as a proxy for density. Encounter rate was lowest on the inshore survey route (0.077 porpoises per km effort) and higher on the three outer routes (between 0.22 and 0.24 porpoises per km effort). 4.3 Minke whale Minke whales are extensively distributed in the northern and southern hemispheres in tropical, temperate and polar seas (Reid et al. 2003). They occur along the Atlantic seaboard of Britain and Ireland and also throughout the northern and central North Sea as far south as the Yorkshire coast (Figure 34) Distribution Available data from existing cetacean surveys in the Moray Firth were reviewed by the University of Aberdeen using information from peer reviewed journals and the grey literature and unpublished data collected by various groups (Thompson et al. 2010a). These data represent observations made over a period of 30 years, from 1980 to 2010 (Table 2), although coverage of the outer Moray Firth is patchy in both space and time. Minke whales were the second most commonly sighted species in offshore waters after harbour porpoises, although there was some evidence that this may be a relatively recent situation as there were comparatively few minke whale sightings in earlier datasets (Thompson et al. 2010a). This finding is backed up by sightings made during the University of Aberdeen s boat (Figure 36) and aerial (Figure 37) surveys when the majority of minke whales sightings occurred during the offshore, rather than the coastal, transects. 44

46 Figure 34. Sightings of minke whales around the UK ( ; taken from Reid et al. 2003). 45

47 Figure 35. Map of survey tracks from the University of Aberdeen s 2009 boat based surveys (taken from Thompson and Brookes 2011). Figure 36. Sightings of minke whales made during the University of Aberdeen s 2009 boat based surveys (track lines are shown in Figure 35; taken from Thompson et al. 2010a). 46

48 Figure 37. Sightings of minke whales made during the University of Aberdeen s 2010 aerial surveys of the outer Moray Firth (track lines are shown in Figure 4; University of Aberdeen unpublished data) In the southern Moray Firth (in relation to the cable route) Minke whales are found in coastal as well as offshore waters. Sightings made during surveys (additional to those listed in Table 2) carried out between May and October along the southern shore of the outer Moray Firth (between Lossiemouth and Fraserburgh) by the CRRU (Figure 14) are shown in Figure 29. Minke whales were the second most commonly sighted species on these surveys after harbour porpoises. Minke whales were sighted along the coast from the east side of Spey Bay to Fraserburgh. While they were encountered on all four CRRU survey routes (Figure 14), corrections for effort revealed a considerably higher abundance on the three outer routes (Robinson et al. 2007). Using southern outer Moray Firth data from May to October 2001 to 2006, Robinson et al. (2009) found a strong spatial preference by minke whales for water depths between 20 and 50m, steep slopes (>60 ), a northerly facing aspect and sandy gravel sediment type. Lesser sandeels (Ammodytes marinus), key minke whale prey in Scottish waters (Pierce et al. 2004), require sediments of coarse sand and fine gravel for burrowing and protection and the arrival of whales in the study area each year appears to be synchronised with the emergence of sandeels into the water column to feed (Robinson et al. 2009). 47

49 4.3.2 Seasonal variation Minke whales are present seasonally in the Moray Firth. Most sightings have been made between May and September, with few records between October and April (Reid et al. 2003). This finding is supported by the BOWL boat survey data which were collected over the BOWL site in : the first minke whale sightings of the year were made in April and the last in October (Figure 38). In 2010 (the only full calendar year of data collection) the number of sightings peaked in July Number of sightings O N D J F M A M J J A S O N D J F M A M J J Figure 38. The number of minke whale sightings made each month during the BOWL boat surveys (solid bars; BOWL unpublished data). It should be noted that although the number of sightings has not been effort corrected, the amount of survey effort in each month is similar. Where two surveys were carried out in a single month (April 2010 and January 2011), as opposed to the usual one, the mean number of sightings was calculated. Months in which no surveys were carried out (November 2009, January 2010 and November 2010) are shown by the patterned bars In the southern Moray Firth (in relation to the cable route) Minke whales are typically recorded along the south Moray Firth coast from mid June onwards with the number of encounters remaining fairly constant from July to October thereafter (Robinson et al. 2007; Robinson et al. 2009; Figure 39). They are absent during the winter/spring months. Minke whale sightings in the southern Moray Firth have been found to be correlated with oceanographic features (Tetley et al. 2008). A cold water current and a warm water plume appear to dominate the Moray Firth region. Encounter rates are significantly higher during warm plume events than when the cold current is dominant (Figure 40). Levels of phytoplankton biomass also appear to be substantially greater during warm water plume events. Tetley et al. (2008) hypothesise that the highest minke whale encounter rates are associated with the presence of targeted prey species that are attracted by high densities of phytoplankton. 48

50 Figure 39. The number of minke whale encounters per km of survey effort between the months of May and October 2001 to 2006 (taken from Robinson et al. 2009). Figure 40. Minke whale sightings per unit effort in months when the cold water current, or the warm water plume, were dominant (taken from Tetley et al. 2008). The mean, interquartile range and range are shown. 49

51 4.3.3 Abundance Minke whale abundance was estimated for the north western North Sea (Block D) during the SCANS (1994) survey and the Moray Firth, Orkney and Shetland (Block J) during the SCANS II (2005) survey (Table 7). The density estimates from the two surveys are not significantly different. Year Area Estimate CV Density (animals CV Reference per km 2 ) 1994 SCANS Block D (north western North Sea) 2, Hammond et al SCANS II Block J (Moray Firth, Orkney and Shetland) SCANS II 2008 Table 7. Abundance estimates for minke whales in the north western North Sea In the southern Moray Firth (in relation to the cable route) Minke whale abundance has not been estimated in the southern Moray Firth; however, encounter rates (Robinson et al. 2007) were lower on the inshore survey route (0.011 minke whales per km effort) than on the three outer routes (between and minke whales per km effort). 4.4 Common dolphin Common dolphins are among the most abundant cetaceans in the world s warm temperate and tropical waters (Reid et al. 2003). In the North Atlantic all individuals appear to be D. delphis, the short beaked common dolphin, which is mainly distributed south of around 60 N. They are found in continental shelf waters off western British and Irish coasts although they have been observed occasionally in the North Sea (Figure 41), mainly between June and September (Reid et al. 2003) Distribution Common dolphins are more regularly sighted off the UK s west coast (where they occur both around the coast and in offshore waters) than in the North Sea (Figure 41). The few sightings that have been made in the Moray Firth have also been made both around the coast (University of Aberdeen unpublished data) and offshore (Figure 41), predominantly on the north side of the Moray Firth (Figure 3). 50

52 Figure 41. Sightings of common dolphins around the UK ( ; taken from Reid et al. 2003) In the southern Moray Firth (in relation to the cable route) Common dolphin sightings occur both around the southern Moray Firth coast and offshore (Figure 42) in water depths of 51 to 209m (mean depth / 42.2m) and at a distance from shore of 5 to 32km (mean distance / 8.0km; Robinson et al. 2010). 51

53 Figure 42. The distribution of common dolphin sightings recorded during surveys carried out between February and November in 2001 to 2009 by the CRRU (area covered shown by the shaded boxes) and WDCS (area covered by all the boxes; taken from Robinson et al. 2010) Seasonal variation There have been too few common dolphin sightings in the Moray Firth to be able to assess seasonal variation In the southern Moray Firth (in relation to the cable route) Common dolphins were recorded in the southern Moray Firth between May and August in 2006 to 2009, with newborn calves observed in June and July (Robinson et al. 2010). This increase in sightings in recent years (since 2006) may be due to rising sea temperatures: common dolphins are normally found in waters warmer than the northern North Sea, but appear to be able to react rapidly to changes in water temperature and adjust their distribution accordingly (Robinson et al. 2010; MacLeod et al. 2008) Abundance Abundance of common dolphins in the Moray Firth is likely to be low; there have been too few sightings to estimate abundance. 52

54 In the southern Moray Firth (in relation to the cable route) There are no abundance estimates for common dolphins using the southern Moray Firth. Group sizes recorded during the CRRU/WDCS surveys (see Figure 42) ranged from 2 to over 450 individuals (n=13; Robinson et al. 2010). 4.5 White beaked dolphin White beaked dolphins are restricted to temperate and sub Arctic seas in the North Atlantic. They occur over a large part of the northern European continental shelf but are recorded most frequently in the western part of the central and northern North Sea and off northern and western Scotland (Reid et al. 2003) Distribution Most of the white beaked dolphin sightings in the Moray Firth have been made offshore (Figure 3; Figure 43). They are the most commonly sighted dolphin species in the outer Moray Firth and there have also been occasional sightings in the inner Moray Firth (Figure 3). This species occurs closer to the coast in other areas e.g. Aberdeenshire (Figure 43; Weir et al. 2007). Figure 43. Sightings of white beaked dolphins around the UK ( ; taken from Reid et al. 2003). 53

55 4.5.2 Seasonal variation There have been too few white beaked dolphin sightings in the Moray Firth to be able to assess seasonal variation Abundance White beaked dolphin abundance was estimated for the north western North Sea (Block D) during the SCANS (1994) survey and the Moray Firth, Orkney and Shetland (Block J) during the SCANS II (2005) survey (Table 8). The density estimates from the two surveys are very similar. Furthermore, the SCANS II density estimate is very similar to the University of Aberdeen 2010 aerial survey density estimate for dolphins obtained for Block B (0.018 individuals per km 2 ; Table 4). Year Area Estimate CV Density (animals CV Reference per km 2 ) 1994 SCANS Block D (north western North Sea) 1, Hammond et al SCANS II Block J (Moray Firth, Orkney and Shetland) SCANS II 2008 Table 8. Abundance estimates for white beaked dolphins in the north western North Sea In the southern Moray Firth (in relation to the cable route) White beaked dolphin sightings off the southern Moray Firth coast are relatively rare (see Figure 43; no sightings recorded by the CRRU), although there have been several sightings further offshore (Figure 3). It is possible that the species is present during Winter (when fewer bottlenose dolphins are sighted around the coast) but no dedicated visual surveys of the southern Moray Firth coast have been carried out between October and May (Robinson et al. 2007). 4.6 Risso s dolphin Risso s dolphins occur in virtually all of the world s oceans between 60 S and 60 N although the species does not appear to be common anywhere (Reid et al. 2003). In north west Europe they appear to be a continental shelf species. Most sightings are from western Scotland, particularly around the Outer Hebrides (Figure 44). Sightings in the northern North Sea were obtained primarily in July and August although some animals were present off north east Scotland and Shetland in Winter (Reid et al. 2003) Distribution Of the few Risso s dolphin sightings in the Moray Firth, most have been made offshore (Figure 3; Figure 44). 54

56 Figure 44. Sightings of Risso s dolphins around the UK ( ; taken from Reid et al. 2003) Seasonal variation There have been too few Risso s dolphin sightings in the Moray Firth to be able to assess seasonal variation Abundance Abundance of Risso s dolphins in the Moray Firth is likely to be low; there have been too few sightings to estimate abundance In the southern Moray Firth (in relation to the cable route) Risso s dolphin sightings in the Moray Firth mainly occur offshore (Figure 3; Figure 44). There were five sightings off the southern Moray Firth coast during CRRU surveys carried out between May and October 2001 to All five sightings were made during the month of September, between the 20 and 50m isobaths (Robinson et al. 2007). 55

57 4.7 Other cetacean species Four other cetacean species (fin, humpback, killer and long finned pilot whale) have occasionally been sighted in the Moray Firth (Table 1). All four are Annex IV and European Protected species (Table 1) and are widely distributed in the North Atlantic (Roman and Palumbi 2003; Foote et al. 2009; Fullard et al. 2000; see also They have been sighted both around the coast and offshore in the Moray Firth (Figure 45; University of Aberdeen unpublished data) but their frequency of occurrence is low (i.e. once every few years) and there have been too few sightings to estimate abundance. Figure 45. Sightings of fin, humpback, killer and long finned pilot whales around the UK ( ; taken from Reid et al. 2003) In the southern Moray Firth (in relation to the cable route) Killer whales (6 sightings), pilot whales (3 sightings), humpback whales (1 sighting) and white sided dolphins (Lagenorhynchus acutus; 1 sighting) have all been encountered intermittently off the southern Moray Firth coast during the Summer months (Robinson et al. 2007; see Figure 14 for survey route). 56

58 4.8 Harbour seal Britain is home to 30% of the population of the European harbour seal P. v. vitulina, and Scotland holds 84% of the British population. Harbour seals are present in the Moray Firth all year round and use intertidal haulout sites to rest between foraging trips, to breed (June/July) and to moult (August/September). They are an Annex II species for which part of the Dornoch Firth has been designated as an SAC (Figure 46; see Section 4.8.4). Compared with the 1990s, major declines have been observed in Shetland (50%), Orkney (67%), the Outer Hebrides (35%), the Moray Firth (40%) and the Firth of Tay (85%; SCOS Main Advice 2010). Other populations appear to be stable (Strathclyde and the west coast of the Highland region) or increasing (English East Coast). Figure 46. The extent of the Dornoch Firth and Morrich More harbour seal SAC shown by the hashed area (information accessible via the NBN Gateway 19 ) Distribution On land during the moult (August) Figure 47 shows the distribution of seals counted during the latest SMRU moult surveys in the Moray Firth and adjacent areas. The largest numbers of animals were hauled out at Findhorn, Ardersier, in the Beauly, Cromarty and Dornoch Firths, and at Loch Fleet &srcDsKey=GA

59 Findhorn Figure 47. The number and distribution of harbour seals counted during SMRU thermal imaging surveys between August 2007 and 2009 (taken from Duck and Thompson 2009) At sea Telemetry data provide the best estimate of year round distribution. The distribution of harbour seals when they are at sea in the Moray Firth has been examined in a number of studies using VHF radio, SRDL and GPS phone telemetry. In the late 1980s/early 1990s Thompson et al. (1996) caught 21 harbour seals and fitted them with VHF radio tags. All 21 seals foraged within 60km of their haulout sites. They showed seasonal variation in their foraging areas (Figure 48) which was related to changes in their terrestrial distribution. When compared to tracks from five grey seals, which were tagged as part of the same study, there was overlap in the foraging areas used by both species in more inshore areas (Thompson et al. 1996). In , ten harbour seals were captured in the Dornoch Firth and Loch Fleet and fitted with SRDL tags (Sharples et al. 2008). The majority of foraging occurred to the east and north east of the haulout sites (Figure 49). The densest areas of foraging were between 30 and 70km from haulout sites. Only two animals switched haulout site, one travelled approximately 50km to haul out in the Beauly Firth while the other travelled more than 190km to haul out on Stronsay in Orkney (Figure 49), just south of the harbour seal SAC on Sanday. Cordes et al. (2011) compared foraging areas of breeding females caught in the Dornoch Firth and at Loch Fleet in 1989 and 2009 using a combination of VHF radio (1989) and GPS phone (2009) telemetry. They found that females foraged in broadly similar areas during both time periods (Figure 50). 58

60 Figure 48. Foraging areas used by radio tagged harbour seals from the Dornoch Firth in Summer (left) and Winter (right). The shading is related to the number of different individuals whose foraging areas overlapped each 1km square (taken from Thompson et al. 1996). Figure 49. Tracks of individual harbour seals captured in the Dornoch Firth and Loch Fleet in 2004 and 2005 (left) and density of foraging locations at sea (right) (taken from Sharples et al. 2008). 59

61 Figure 50. Comparison of adult female foraging locations in 1989 (n=5) and 2009 (n=5; taken from Cordes et al. 2011). The foraging locations of each individual are represented by a different symbol (key relating symbols to individuals is shown at the bottom right of each map). The solid lines show the 50% contours for individual foraging areas as calculated by Kernel analysis. The Dornoch Firth and Morrich More SAC is shown by the shaded area. The Loch Fleet NNR is shown by the hashed area Seasonal variation on land Harbour seals are present at haulout sites in the inner Moray Firth throughout the year although the number of animals at these sites peaks during June, July and August (the breeding and moulting seasons; Thompson et al. 1996). Their distribution within this area is known to vary seasonally and between years and is likely influenced by proximity to foraging area (outside the breeding season) and site characteristics (during the breeding season; Thompson et al. 1996) Abundance Numbers of harbour seals in many areas around the UK have declined since the 1990s. Numbers in Shetland have declined by 50%, Orkney by 67%, the Outer Hebrides by 35%, the Moray Firth by 40% and the Firth of Tay by 85%. Numbers in other areas (Strathclyde, the west coast of the Highland region and the English East Coast) do no show consistent declines (SCOS Main Advice 2010). 60

62 Nearly 25,000 harbour seals were counted in the whole of Britain during the most recent available SMRU moult counts of which 84% were in Scotland and approximately 4% (<1,000) in the Moray Firth (SCOS Main Advice 2010). Moult surveys provide the best estimates of population size On land during the breeding season (June/July) Breeding season counts in the inner Moray Firth have been carried out since the late 1980s by the University of Aberdeen and SMRU (Duck at al. 2010). Counts of were usual in the inner Moray Firth in the late 1980s (after the phocine distemper virus outbreak in 1988) and these increased to around 1000 animals in Since then breeding season numbers have decreased steadily: by only around 500 animals were counted in the inner Moray Firth. Numbers increased by 27% in 2009 to 671 animals (Duck at al. 2010) On land during the moult (August) Figure 51 shows trends in the number of harbour seals counted at the different haulout sites in the Moray Firth during annual moult (August) surveys carried out over the last 15 years by the University of Aberdeen and SMRU. Numbers in the Dornoch Firth have decreased over this period. Cordes et al. (2011) showed that there has been a shift in distribution between the Dornoch Firth and Loch Fleet. In 1988, all mother pup pairs counted in the two estuaries were located at haulout sites within the Dornoch Firth but by 2008 the newly developed site at Loch Fleet accounted for 37% of mother pup pairs. 700 Number of harbour seals counted Year Loch Fleet to Dunbeath Loch Fleet Dornoch Firth (SAC) Cromarty Firth Beauly Firth Ardersier Findhorn Figure 51. Number of harbour seals at different haulout sites in the Moray Firth during moult (August) surveys (produced using data from Table 2 in Duck et al. 2010). When multiple surveys of haulout sites were carried out in one year (2004, 2005, 2006, 2007, 2008) the mean count for the site is presented. Dashed lines indicate a hypothetical trend between counts (due to a lack of data for the intervening years). 61

63 4.8.4 The Dornoch Firth and Morrich More SAC Selection of SACs for harbour seals was primarily based on numbers counted during their annual moult, in August. Additional surveys confirmed that the candidate SACs were also used for breeding. There are eight harbour seal SACs designated in Scotland: Yell Sound Coast and Mousa (Shetland), Sanday (Orkney), Ascrib, Isay and Dunvegan (Skye), Eileanan agus Sgeiran Lios mor (Isles and Skerries of Lismore) and South East Islay Skerries (Strathclyde), Dornoch Firth and Morrich More (Moray Firth), and the Firth of Tay and Eden Estuary (Tayside and Fife). The marine component of the Dornoch Firth and Morrich More SAC extends from Bonar Bridge to the mouth of the estuary between Dornoch Point on the north shore and to the west of Portmahomack on the south shore (Figure 46; SNH 2006). As well as harbour seals, the estuary supports a second qualifying Annex II species, the otter (Lutra lutra). Trends in abundance of harbour seals within the Dornoch Firth (and at the nearby Loch Fleet National Nature Reserve/NNR) are shown in Figure 52. Numbers in the Dornoch Firth are now around one third of what they were 20 years ago, while numbers at sites in Loch Fleet, which was not used by harbour seals 20 years ago, have increased (though not by an equivalent number). The condition of the Dornoch Firth and Morrich More SAC has been assessed three times during the last reporting cycle. There were 405 seals in 2000, 220 seals in 2002 (although this is considered an undercount because the survey was undertaken more than two hours after low tide), and 290 seals in 2003 (SNH 2005). These data, along with previous counts made in 1992 (662), 1994 (542) and 1997 (593), indicate that the number of harbour seals within the SAC during the moulting season has decreased over the reporting cycle. The SAC is considered to be Unfavourable (recovering) (SNH 2005) and a management plan 20 is now in place which is addressing the main reason believed to be behind the decline (shooting of seals mainly to protect salmon and sea trout fisheries). Figure 52. Trends in abundance of harbour seals within the Dornoch Firth (filled triangles) and at the nearby Loch Fleet NNR (open circles; taken from Cordes et al. 2011)

64 4.8.5 Links between the Dornoch Firth and Morrich More SAC and the BOWL site Over the last two decades, 37 individual harbour seals from the Dornoch Firth and Morrich More SAC and the nearby Loch Fleet NNR have been tracked using a variety of techniques (VHF radio, SRDL and GPS phone telemetry) (Thompson et al. 1994; Sharples et al. 2008; Cordes et al. 2011; see Section ; Table 9). These data were used to underpin predictions of use of the BOWL (and MORL) sites by harbour seals using habitat association modelling (Bailey and Thompson 2011). Habitat characteristics used included depth, slope, distance to nearest haulout and sediment type. Different error structures for each of these technologies (VHF radio, SRDL and GPS phone telemetry) required the development of a novel Bayesian state space approach to integrate these data into a single modelling framework (Bailey and Thompson 2011). Existing procedures were then used to predict habitat usage (Aarts et al. 2008) and estimate how many harbour seals from the Dornoch Firth and Morrich More SAC were likely to use habitats within the development areas (Matthiopoulos et al. 2004). Analysis of the combined VHF radio, SRDL and GPS phone telemetry dataset demonstrated that harbour seals from the Dornoch Firth and Morrich More SAC were dispersed widely across the Moray Firth, particularly over offshore sandbanks, and are likely to spend time foraging within the BOWL (and MORL) sites (Bailey and Thompson 2011). Although there is variability in the importance of different parts of the BOWL and MORL sites, some 4x4km grid squares in this region might be expected to hold up to eight seals (based on the highest levels of abundance seen over the last two decades), representing a density approaching 0.5 individuals per km 2 (Figure 53). As well as the link with the Dornoch Firth and Morrich More SAC, animals from other harbour seal SACs, e.g. that in Orkney (see Section 4.8.4), may also use the BOWL site. Figure 54 shows tracks of harbour seals tagged with SRDL tags, mostly in Orkney. Although this Figure shows only tracks of seals which entered the Pentland Firth and Orkney Strategic Area at least once (SMRU Ltd 2011), it gives an idea of how widely harbour seals range and highlights the possibility that animals from other SACs may also use the Moray Firth. Tag type Deployment years Number of tags VHF radio SRDL GPS phone Table 9. Summary of telemetry data for harbour seals tagged in the Dornoch Firth and Morrich More SAC and Loch Fleet NNR (Thompson et al. 1994; Sharples et al. 2008; Cordes et al. 2011). 63

65 Spey Bay Figure 53. Predicted numbers of harbour seals from the Dornoch Firth SAC and Loch Fleet NNR in different 4x4km grid cells across the Moray Firth (taken from Bailey and Thompson 2011). Figure 54. Tracks of 17 harbour seals (2 of which were tagged at the Dornoch Firth and Morrich More SAC and Loch Fleet NNR, the other 15 were tagged in the northern half of Orkney, including the Sanday SAC; left) and 15 female harbour seal pups (tagged on Sanday in Orkney; right) fitted with SRDL tags (taken from SMRU Ltd 2011). Only tracks (colour coded by individual seal) of seals which entered the Pentland Firth and Orkney Strategic Area at least once are shown, but they give an idea of how widely both young and adult harbour seals range. 64

66 4.8.6 Loch Fleet National Nature Reserve (NNR) Photo ID studies of harbour seals using the Loch Fleet NNR offer the opportunity to assess individual measures of fitness such as reproduction rates, survival and phenology (biological cycles and events) by observing the behaviour of individuals through time (Thompson and Wheeler 2008). Loch Fleet is the closest harbour seal breeding site to the BOWL site, and has become increasingly important relative to the Dornoch Firth SAC over the last 20 years (Figure 52; Cordes at el. 2011). Photo ID has been carried out at this site during the breeding season for the last six years (University of Aberdeen unpublished data) In the southern Moray Firth (in relation to the cable route) Findhorn is the main harbour seal haulout on the southern Moray Firth coast (Figure 47). Numbers have varied over the years and the site is currently used by around 100 animals during the August moult (Figure 51). The site is also used for pupping. The area around Spey Bay appears to be used by at least some of the animals tagged in the Dornoch Firth and Loch Fleet for foraging (Figure 53). It should also be noted that there is evidence of movement of harbour seals between the Sanday (Orkney) and the Dornoch Firth and Morrich More SACs and the southern Moray Firth coast (Figure 54; Section 4.8.5) Mortality from corkscrew injuries In 2009 a previously unidentified source of anthropogenic mortality was identified in harbour and grey seals in Scotland (Thompson et al. 2010b). In 2010, severely damaged seal carcasses were found on beaches in eastern Scotland (St Andrews Bay, Tay and Eden Estuaries and Firth of Forth), along the North Norfolk coast in England (centred on the Blakeney Point Nature Reserve) and within and around Strangford Lough in Northern Ireland (Thompson et al. 2010b). All the seals had a characteristic wound consisting of a single smooth edged cut that started at the head and spiralled around the body. In most cases the resulting spiral strip of skin and blubber was detached from the underlying tissue. In each case the wound would have been fatal (Thompson et al. 2010b). The extremely neat edge to the wound strongly suggested the effects of a blade with a smooth edge applied with considerable force, while the spiral shape was consistent with rotation about the longitudinal axis of the animal. The injuries were consistent with the seals being drawn through a ducted propeller such as a Kort nozzle or some types of Azimuth thruster (Thompson et al. 2010b). Such systems are common to a wide range of ships including tugs, self propelled barges and rigs, various types of offshore support vessels and research boats. All other proposed explanations of the injuries, including Greenland shark predation, are difficult to reconcile with the observations and, based on the evidence to date, seem very unlikely to have been the cause of these mortalities (Thompson et al. 2010b). At present the population consequences of these mortalities is unknown (SCOS Main Advice 2010). 4.9 Grey seal About 45% of the world population of grey seals is found in Britain and over 90% of British grey seals breed in Scotland (SCOS Main Advice 2008). Although grey seal pup production has increased steadily since the 1960s, when dedicated surveys began, and continues to increase rapidly in the North Sea, there is clear evidence that growth is now levelling off in Orkney and the Hebrides. Grey seals are present in the Moray Firth year round, hauling out at intertidal sites between foraging trips and breeding on beaches (or in caves) above the high water mark along the Helmsdale coastline in Autumn. They are an Annex II species and are protected by means of a network of SACs (see Section 4.9.4). 65

67 4.9.1 Distribution On land Figure 55 gives an idea of the distribution of grey seals on land around the Moray Firth in Summer and shows that the greatest numbers of grey seals haul out in the Dornoch Firth at this time of year. In Winter the majority of grey seals hauling out in the Moray Firth can be found along the Helmsdale coast where they breed. Duncansby Head Brora Helmsdale Dornoch Firth Findhorn Figure 55. The number and distribution of grey seals counted during SMRU thermal imaging surveys of the Moray Firth in August 2007 and 2009 (taken from Duck and Thompson 2009) At sea In UK waters One of the best ways to estimate how grey seals use the marine environment is to use telemetry data (e.g. Matthiopoulos et al. 2004). As a result of modelling and interpolating satellite telemetry (from 110 tagged individuals captured at major haulout sites during and observed from May to September i.e. outside the breeding and moulting seasons) and haulout survey data around Britain, we can see that grey seal usage is primarily concentrated: (i) off the northern coasts of the British Isles; (ii) closer to the coast than might be expected purely on the basis of accessibility from the haulout sites; and (iii) in a limited number of marine hot spots e.g. the Pentland Firth and, to a lesser extent, the Moray Firth (Figure 56). 66

68 Figure 56. The estimated usage of the British marine environment by grey seals (taken from Matthiopoulos et al. 2004) In the Moray Firth The density of grey seals in the Moray Firth was estimated by Jones and Matthiopoulos (2011 see Appendix 3) using telemetry and count data. As for the other species for which density has been estimated (bottlenose dolphin, harbour porpoise and harbour seal), a 4x4km 2 grid was used. Maps of estimated total, at sea, and hauled out usage in a study area surrounding the proposed MORL and BOWL wind farm developments are presented (although it is anticipated that use of the at sea densities will be most appropriate when assessing the potential impact of construction). Figure 57 shows the spatial usage of grey seals around the proposed MORL/BOWL proposed development sites with white contour lines denoting standard deviation as a measure of uncertainty around estimated usage. Highest usage is located in the Dornoch Firth (and also in the Pentland Firth and around the Orkney Islands). Possible offshore foraging patches can also be seen (denoted in orange and yellow). 67

69 Figure 58 shows estimated at sea usage. Total and at sea usage display similar characteristics although at sea usage is 28% lower than total usage due to the removal of hauled out usage. Figure 59 shows estimated grey seal hauled out usage. The highest hauled out usage occurs in the Dornoch Firth (and at some of the islands in the Pentland Firth, e.g. Stroma, and around Orkney, e.g. Stronsay). Figure 57. Estimated grey seal total (at sea and hauled out) usage around the proposed MORL/BOWL development sites. White contours show standard deviation from mean usage as a measure of uncertainty. 68

70 Figure 58. Estimated grey seal at sea usage around the proposed MORL/BOWL development sites. White contours show standard deviation from mean usage as a measure of uncertainty. Figure 59. Estimated grey seal hauled out usage around the proposed MORL/BOWL development sites. White contours show standard deviation from mean usage as a measure of uncertainty. 69

71 4.9.2 Seasonal variation on land The number of grey seals counted at inter tidal haulout sites within the inner Moray Firth appears to be highest during the Summer, at least historically (Figure 60). Only a few grey seals (tens) use these sites during the Winter. Figure 60. Maximum monthly counts of grey seals at inner Moray Firth haulout sites from January 1988 to August 1990 (taken from Thompson et al. 1996) Abundance Each year, SMRU conducts aerial surveys of the major grey seal breeding colonies in Britain to determine the number of pups born (pup production). Pup production surveys have been carried out by SMRU in the North Sea since the 1960s (SCOS Main Advice 2008). The total number of seals associated with surveyed sites can then be estimated by applying a population model to the estimates of pup production (SCOS Main Advice 2008). The latest SMRU survey data available for the Moray Firth were collected in 2009 when pup production was estimated for the coastline between Duncansby Head and Helmsdale (Figure 55; SCOS Main Advice 2010). Figure 61 shows the trend in pup production which has been occurring at the regularly surveyed North Sea colonies (Isle of May, Fast Castle, Farne Islands, Donna Nook, Blakeney Point and Horsey in east Norfolk) since the 1980s. Pup production at Helmsdale (surveyed less regularly than other North Sea sites) has remained constant over this period and now accounts for less than 15% of grey seal pup production in the North Sea (Figure 61). The typical abundance of grey seals on land during August (i.e. outwith the breeding season) is shown in Figure 55. The greatest numbers of animals (low hundreds) haul out in the Dornoch Firth and to the north of Brora. 70

72 Pup production estimate North Sea colonies Duncansby Head to Helmsdale Figure 61. Grey seal pup production estimates for North Sea colonies from 1960 to 2009 (produced using data from Duck and Morris 2010) and for the Duncansby Head to Helmsdale area from 2005 to 2009 (produced using data from Duck and Mackey 2006; Duck and Mackey 2007; Duck and Mackey 2008; Duck 2009; Duck and Morris 2010). The Duncansby Head to Helmsdale colony was counted by boat prior to 2005, but only single counts were made and therefore no pup production estimates are available. Note that the surveys described here do not account for the small groups of seals breeding in caves along the Sutherland and Caithness coasts Grey seal SACs Grey seal SAC designation was based on the numbers of pups born at individual breeding colonies. Candidate SACs had to contribute a reasonable proportion of pups to the total and had to be evenly distributed across the geographic range of grey seal breeding colonies. The largest grey seal breeding colonies were selected as SACs using the most up to date pup production data (SMRU Ltd 2011). There are six grey seal SACs in Scotland: the Treshnish Isles (Strathclyde), the Monach Isles (Outer Hebrides), North Rona (Outer Hebrides), Faray and Holm of Faray (Orkney), the Isle of May (Firth of Forth) and the Berwickshire and North Northumberland Coast (which crosses the border between Scotland and England on the east coast) Links between grey seal SACs and the BOWL site Published studies relating to grey seal movements at sea show that, while grey seals often forage close to shore in areas local to the sites they are using to haul out, they also make long distance movements (e.g. McConnell et al. 1999). For example, Figure 62 shows the mean daily locations of five grey seals satellite tagged in the Moray Firth illustrating the long range movements (from the Moray Firth to haulout sites km away in Orkney, the Firth of Forth and at the Farne Islands) made by four of the five tagged seals (Thompson et al. 1996). This highlights the interchange that occurs between areas for grey seals. Even the three grey seals which foraged within the Moray Firth travelled up to 145km from haulout sites to do so (Figure 63). SMRU has been deploying telemetry tags on grey seals since 1988 (see Section 3.2.2). BOWL and MORL commissioned work examined these data to determine the extent of movement of (1) pups from breeding sites and (2) animals aged one year and above (Russell 2011). All telemetry locations were cleaned according to SMRU protocol (Russell et al. 2011) and, where appropriate, corrected for 71

73 positional error using a linear Gaussian state space Kalman filter (Royer and Lutcavage 2008; Jones et al. 2011). Grey seal pups showed considerable inter individual variation in the extent of movements they made upon departing from breeding colonies (Figure 64). For example, pups tagged at the Isle of May made movements out into the North Sea both close to and far from shore; some travelled great distances, moving as far afield as Norway and The Netherlands (Figure 64). Although there have not been any telemetry deployments on pups at the Helmsdale breeding colony, it is likely that their movements will be consistent with those of pups which were tagged at the breeding colonies shown in Figure 64. For grey seals aged one year and above, a buffer zone was generated which extended 100km from the boundary of the potential BOWL and MORL wind farm development sites (Russell 2011). Data were presented if a location was recorded inside the buffer zone (Figure 65; Figure 66). Grey seals aged one year and above ranged widely e.g. the individual which travelled to the Faroe Islands. There is a high probability that grey seals using the Moray Firth and/or the BOWL site have hauled out, or will haul out, at some point at one or more of the six Scottish grey seal SACs (Figure 65; Figure 66; see Section 4.9.4). Figure 62. Mean daily locations of five grey seals satellite tagged in the Moray Firth illustrating the long range movements made by four of the five individuals (taken from Thompson et al. 1996). 72

74 Figure 63. Foraging areas of three satellite tagged grey seals (taken from Thompson et al. 1996). 73

75 Figure 64. The extent of grey seal pup (n=39) movements from the breeding sites where they were tagged (taken from Russell 2011). The tracks are colour coded by tagging location (see legend). The solid black line shows a 100km buffer zone around the BOWL and MORL wind farm sites. 74

76 Figure 65. Tracks of grey seals (aged one year and above; n=65) which, at least once while they were tagged, entered a 100km buffer zone around the proposed BOWL and MORL wind farm sites. Each colour represents a different individual (taken from Russell 2011). 75

77 Figure 66. Tracks of grey seals (aged one year and above; n=65) which, at least once while they were tagged, entered a 100km buffer zone around the proposed BOWL and MORL wind farm sites. Each colour represents a different individual. This Figure shows the same information as Figure 65 but is magnified to show the Moray Firth in more detail (taken from Russell 2011). 76

78 4.9.6 In the southern Moray Firth (in relation to the cable route) There are several grey seal haulouts on the southern Moray Firth coast, the largest of which (at least in August) is at Findhorn (Figure 55), but there are no grey seal breeding colonies. The outer Moray Firth appears to be used primarily by grey seals transiting between sites to the south of the Moray Firth and sites in Orkney (Figure 66). 77

79 5 References Bailey, H. and Thompson, P.M Using marine mammal habitat modelling to identify priority conservation zones within a marine protected area. Marine Ecology Progress Series 378: Bailey, H. and Thompson, P Technical report on harbour seal telemetry and habitat model. SMRU Ltd/University of Aberdeen report to MORL and BOWL. Booth, C., Caillat, M., Gillespie, D. and Grellier, K Developing acoustic methods for determining the likelihood that dolphins using the MORL and BOWL wind farm sites are bottlenose dolphins. SMRU Ltd report to MORL and BOWL. Cheney, B., Costa, M., Culloch, R., Durban, J.W., Elwen, S., Islas, V., Eisfeld, S.M., Evans, P.G.H., Hammond, P.S., Ingram, S., Janik, V.M., O Brien, J., Pope, A., Phillips, C., Quick, N., Robinson, K.P., Thompson, P.M., Weir, C.R. and Wilson, B. 2011a. Multi site mark recapture estimate of the abundance amd movement rates of bottlenose dolphins. In Thompson, P.M., Cheney, B., Ingram, S., Stevick, P., Wilson, B. and Hammond, P.S. (Eds) Distribution, abundance and population structure of bottlenose dolphins in Scottish waters. Scottish Government and Scottish Natural Heritage funded report. Scottish Natural Heritage Commissioned Report No Cheney, B., Barton, T.R., Candido, A.T., Elwen, S., Hammond, P.S., Ingram, S., Phillips, C., Pope, A., Thompson, P.M., Weir, C.R. and Wilson, B. 2011b. Winter distribution of bottlenose dolphins using the Moray Firth Special Area of Conservation. In Thompson, P.M., Cheney, B., Ingram, S., Stevick, P., Wilson, B. and Hammond, P.S. (Eds) Distribution, abundance and population structure of bottlenose dolphins in Scottish waters. Scottish Government and Scottish Natural Heritage funded report. Scottish Natural Heritage Commissioned Report No Cordes, L.S., Duck, C.D., Mackay, B.L., Hall, A.J. and Thompson, P.M Long term patterns in harbour seal site use and the consequences for managing protected areas. Animal Conservation 14: no. doi: /j x. Corkrey, R., Brooks, S., Lusseau, D., Parsons, K., Durban, J.W., Hammond, P.S. and Thompson, P.M A Bayesian capture recapture population model with simultaneous estimation of heterogeneity. Journal of the American Statistical Association 103: Culloch, R.M. and Robinson, K.P Bottlenose dolphins using coastal regions adjacent to a Special Area of Conservation in north east Scotland. Journal of the Marine Biological Association of the United Kingdom 88(6): Duck, C.D. and Mackey, B.L Grey seal pup production in Britain in SCOS briefing paper 06/01. In Scientific Advice on Matters Related to the Management of Seal Populations: 2006 which can be downloaded from andrews.ac.uk/pageset.aspx?psr=411 Duck, C.D. and Mackey, B.L Grey seal pup production in Britain in SCOS briefing paper 07/01. In Scientific Advice on Matters Related to the Management of Seal Populations: 2007 which can be downloaded from andrews.ac.uk/pageset.aspx?psr=411. Duck, C.D. and Mackey, B.L Grey seal pup production in Britain in SCOS briefing paper 08/01. In Scientific Advice on Matters Related to the Management of Seal Populations: 2008 which can be downloaded from andrews.ac.uk/pageset.aspx?psr=411. Duck, C.D., Thompson, D. and Mackey, B.L The status of British common seal populations in SCOS briefing paper 08/03. In Scientific Advice on Matters Related to the Management of Seal 78

80 Populations: 2008 which can be downloaded from Duck, C.D Grey seal pup production in Great Britain and Ireland in SCOS Briefing paper 09/01. In Scientific Advice on Matters Related to the Management of Seal Populations: 2009 which can be downloaded from andrews.ac.uk/pageset.aspx?psr=411. Duck, C.D. and Thompson, D The status of British common seal populations in SCOS briefing paper 09/03. In Scientific Advice on Matters Related to the Management of Seal Populations: 2009 which can be downloaded from Duck, C.D. and Morris, C.D Grey seal pup production in Britain in In Scientific Advice on Matters Related to the Management of Seal Populations: Duck, C.D., Morris, C.D. and Thompson, D The status of British harbour seals populations in In Scientific Advice on Matters Related to the Management of Seal Populations: Foote, A.D., Newton, J., Piertney, S.B., Willerslev, E. and Gilbert, M.T.P Ecological, morphological and genetic divergence of sympatric North Atlantic killer whale populations. Molecular Ecology 18(24): Fullard, K.J., Early, G., Heide Jorgensen, M.P., Bloch, D., Rosing Asvid, A. and Amos, W Population structure of long finned pilot whales in the North Atlantic: a correlation with sea surface temperature? Molecular Ecology 9(7): Grellier, K Proposed approaches to marine mammal data gathering to be adopted within the Beatrice Offshore Wind Farm site. SMRU Ltd report to BOWL. Hammond, P.S., Berggren, P., Benke, H., Borchers, D.L., Collet, A., Heide Jorgensen, M.P., Heimlich, S., Hiby, A.R., Leopold, M.F. and Oien, N Abundance of harbour porpoise and other cetaceans in the North Sea and adjacent waters. Journal of Applied Ecology 39: Harding Hill, R The Moray Firth Review. Scottish Natural Heritage, Northwest Region, Scotland, UK. Jones, E.L., McConnell, B.J., Duck, C., Morris, C. and Matthiopoulos, J SCOS Briefing paper. Jones, E. and Matthiopoulos, J Grey seal usage maps for MORL/BOWL developments. SMRU Ltd report to MORL and BOWL. Lockyer, C Investigation of aspects of the life history of the harbour porpoise in British waters. In: Bjørge, A. and Donovan, G.P. (eds.) Biology of the Phocoenids. International Whaling Commission, Cambridge, UK. Special Issue 16: Lonergan, M., Duck, C.D., Thompson, D., Mackey, B.L., Cunningham, L. and Boyd, I.L Using sparse survey data to investigate the declining abundance of British harbour seals. Journal of Zoology 271: MacLeod, C.D., Weir, C.R., Santos, M.B. and Dunn, T.E Temperature based summer habitat partitioning between white beaked and common dolphins around the United Kingdom and Republic of Ireland. Journal of the Marine Biological Association of the UK 88(6):

81 Matthiopoulos, J., McConnell, B., Duck, C. and Fedak, M Using satellite telemetry and aerial counts to estimate space use by grey seals around the British Isles. Journal of Applied Ecology 41: McConnell, B.J., Fedak, M.A., Lovell, P. and Hammond P.S Movements and foraging areas of grey seals in the North Sea. Journal of Applied Ecology 36: Pierce, G.J., Santos, M.B., Reid, R.J., Patterson, I.A.P. and Ross, H.M Diet of minke whales in Scottish (UK) waters with notes on strandings of this species in Scotland J. Mar. Biol. Ass. UK 84: Quick, N.J Vocal behaviour and abundance of bottlenose dolphins in St Andrews Bay. PhD thesis, University of St Andrews. Reid, J.B., Evans, P.G.H. and Northridge, S.P Atlas of cetacean distribution in north west European waters. Joint Nature Conservation Committee, Peterborough. Robinson, K.P., Baumgartner, N., Eisfeld, S.M., Clark, N.M., Culloch, R.M., Haskins, G.N., Zapponi, L., Whaley, A.R., Weare, J.S. and Tetley, M.J The summer distribution and occurrence of cetaceans in the coastal waters of the outer southern Moray Firth in northeast Scotland (UK). Lutra 50(11): Robinson, K.P., Tetley, M.J. and Mitchelson Jacob, E.G The distribution and habitat preference of coastally occurring minke whales in the outer southern Moray Firth, northeast Scotland. J. Coast. Conserv. 13: Robinson, K.P., Eisfeld, S.M., Costa, M. and Simmonds, M.P Short beaked common dolphin occurrence in the Moray Firth, north east Scotland. Marine Biodiversity Records 3: e55. Roman, J. and Palumbi, S.R Whales before whaling in the North Atlantic. Science 301(5632): Royer, F. and Lutcavage, M Filtering and interpreting location errors in satellite telemetry of marine animals. Journal of Experimental Marine Biology and Ecology 359: Russell, D Grey seal tracks. SMRU Ltd/SMRU report to MORL and BOWL. Russell, D.J.F., Matthiopoulos, J. and McConnell, B.J SCOS Briefing paper. Santos, M.B., Pierce, G.J., Reid, R.J., Patterson, I.A.P., Ross, H.M. and Mente, E Stomach contents of bottlenose dolphins in Scottish waters. Journal of the Marine Biological Association of the UK 81: SCANS II Small cetaceans in the European Atlantic and North Sea (SCANS II). Final report to the European Commission LIFE Nature programme on project LIFE04NAT/GB/ SCOS Main Advice In Scientific Advice on Matters Related to the Management of Seal Populations: 2008 which can be downloaded from SCOS Main Advice In Scientific Advice on Matters Related to the Management of Seal Populations: 2009 which can be downloaded from 80

82 SCOS Main Advice In Scientific Advice on Matters Related to the Management of Seal Populations: 2010 which can be downloaded from SMRU Ltd Utilisation of space by grey and harbour seals in the Pentland Firth and Orkney waters. Scottish Natural Heritage Commissioned Report No SNH SNH site condition monitoring report SNH Dornoch Firth and Morrich More Special Area of Conservation. Tetley, M.J., Mitchelson Jacob, E.G. and Robinson, K.P The summer distribution of coastal minke whales in the southern outer Moray Firth, NE Scotland, in relation to co occurring mesoscale oceanographic features. Remote Sensing of Environment 112: Thompson, P.M., McConnell, B.J., Tollit, D.J., Mackay, A., Hunter, C. and Racey, P.A Comparative distribution, movements and diet of harbour and grey seals from the Moray Firth, N.E. Scotland. Journal of Applied Ecology 33: Thompson, P.M., Lusseau, D., Corkrey, R. and Hammond, P.S Moray Firth bottlenose dolphin monitoring strategy options. Scottish Natural Heritage Commissioned Report No. 079 (ROAME No. F02AA409). Thompson, P.M., Corkrey, R., Lusseau, D., Lusseau, S., Quick, N., Durban, J.W., Parsons, K.M. and Hammond, P.S An assessment of the current condition of the Moray Firth bottlenose dolphin population. Scottish Natural Heritage Commissioned Report No Thompson, P.M. and Wheeler, H Photo ID Based estimates of reproductive patterns in female harbour seals. Marine Mammal Science 24: Thompson, P.M., Cheney, B., Candido, A.T. and Hammond, P.S Site Condition Monitoring of bottlenose dolphins within the Moray Firth Special Area of Conservation: Interim report Scottish Natural Heritage Commissioned Report. Thompson, P., Brookes, K., Cheney, B., Candido, A., Bates, H., Richardson, N. and Barton, T. 2010a. Assessing the potential impact of oil and gas exploration operations on cetaceans in the Moray Firth. First year report for DECC, Scottish Government, COWRIE and Oil & Gas UK. University of Aberdeen. 65pp. Thompson, D., Bexton, S., Brownlow, A., Wood, D., Patterson, T., Pye, K., Lonergan, M. and Milne, R. 2010b. Report on recent seal mortalities in UK waters caused by extensive lacerations. SMRU Report which can be downloaded from andrews.ac.uk/newsitem.aspx?ni=308. Thompson, P.M., Cheney, B., Ingram, S., Stevick, P., Wilson, B. and Hammond, P.S. (Eds) Distribution, abundance and population structure of bottlenose dolphins in Scottish waters. Scottish Government and Scottish Natural Heritage funded report. Scottish Natural Heritage Commissioned Report No Thompson, P. and Brookes, K Technical report on pre consent marine mammal data gathering at the MORL and BOWL wind farm sites. University of Aberdeen report to MORL and BOWL. Thompson, P Passive acoustic monitoring of cetacean activity on the southern Moray Firth coast. Unpublished University of Aberdeen report to BOWL. 81

83 Thompson, P. 2011a. Bottlenose dolphin densities across the Moray Firth. Unpublished University of Aberdeen report to BOWL. Weir, C.R., Stockin, K.A. and Pierce, G.J Spatial and temporal trends in the distribution of harbour porpoises, white beaked dolphins and minke whales off Aberdeenshire (UK), north western North Sea. Journal of the Marine Biological Association of the UK 87: Wells, R.S. and Scott, M.D Common bottlenose dolphin Tursiops truncatus. In Perrin, W.F., Wursig, B. and Thewissen, J.G.M (Eds.) Encyclopedia of Marine Mammals. Second edition. Elsevier, London. Wilson, B., Thompson, P.M. and Hammond, P.S Habitat use by bottlenose dolphins: Seasonal distribution and stratified movement patterns in the Moray Firth, Scotland. J. Appl. Ecol. 34: Wilson, B., Hammond, P.S. and Thompson, P.M Estimating size and assessing trends in a coastal bottlenose dolphin population. Ecological Applications 9(1): Wilson, B., Reid, R.J., Grellier, K., Thompson, P.M. and Hammond, P.S Considering the temporal when managing the spatial: a population range expansion impacts protected areas based management for bottlenose dolphins. Animal Conservation 7:

84 6 Appendices Appendix 1: Thompson, P. and Brookes, K Technical report on pre consent marine mammal data gathering at the MORL and BOWL wind farm sites. University of Aberdeen report to MORL and BOWL. Appendix 2: Appendix 3: Thompson, P. 2011a. Bottlenose dolphin densities across the Moray Firth. Unpublished University of Aberdeen report to BOWL. Jones, E. and Matthiopoulos, J Grey seal usage maps for MORL/BOWL developments. SMRU Ltd report to MORL and BOWL. 83

85 Bottlenose dolphin densities across the Moray Firth Paul Thompson, 11 th October 2011 Recent discussions on assessments of the impact of noise on marine mammals have highlighted how one can overlay information on predicted noise levels over available data on animal density. Our technical report on pre-consent marine mammal data gathering (Thompson & Brookes 2011) used habitat association analyses to produce density maps for harbour porpoise (Fig 3.8) and harbour seal (Annex 1, Fig 13). However, there are several reasons why it is more difficult, and arguably less appropriate, to produce equivalent maps for bottlenose dolphins. First, the dolphin distribution is known to be much patchier, with high use areas on the coast, and survey techniques have differed in these high and low density areas. Second, these animals are highly mobile, so an average density based on studies over a large area (eg. SCANS II data) are of more limited use when considering their likely presence in relatively small areas of interest around a windfram site. In general, it will be more appropriate to use metrics which account for temporal patterns of occurrence when considering bottlenose dolphin use of these smaller areas. Work is underway through the DECC project to achieve this using a combination of PAM data and classification tree analysis. In the meantime, however, there remains a demand for the best available data on bottlenose dolphin densities to incorporate into noise assessment for the BOWL and MORL EIAs. One option is to use the Hammond et al s (In prep) SCANS II estimate for Area J (0.011 animals/km 2 ) as an average for the whole region, but we know that this underestimates the use of the coastal areas known to be important to the East coast of Scotland Bottlenose Dolphin population. As an alternative, I suggest that we use our estimate of (0.066 animals/km 2 ) for the density of all dolphin species in the Moray Firth (Table 3.7), and use the classification tree analyses (Fig 3.16) to account for spatial variation in the density of bottlenose dolphins in different parts of the Moray Firth. The resulting density map is provided below, and a data file with the estimated number of bottlenose dolphins in each 4x4km square across the Moray Firth is attached. In using this density distribution, it must be recognised that this remains very conservative when focussing on impacts in offshore areas and along the northern coast of the Moray Firth SAC, as this approach tends to underestimate the number of animals occurring in the inner Moray Firth and along the southern Moray Firth coast. Ie. Whilst the total number of animals represented on this map (213) apprear reasonable given current estimates of population size, these animals are predicted to be much more widely distributed outisde their core areas than we d expect given other data on the number of animals typically found in the

86 inner part of the Moray Firth SAC and along the Moray coast (Bailey & Thompson 2009; Thompson et al 2011). Figure 1. Predicted number of bottlenose dolphins in each 4x4 km cell. Baseline dolphin density across the whole area is taken to be per km 2 (ie per 4x4km cell). Data from the classification tree are then used to estimate the probability of dolphins in a particular cell being a bottlenose dolphin. Data are provided in file: Predicted Bottlenose Dolphin density.xlsx References: Bailey, H. & Thompson, P.M. (2009) Using marine mammal habitat modelling to identify priority conservation zones within a marine protected area. Marine Ecology Progress Series, 378, Hammond, P.S., Macleod, K., Berggren, P., Borchers, D.L., Burt, M.L., Cañadas, A., Desportes, G., Donovan, G.P., Gilles, A., Gillespie, D., Gordon, J., Hiby, L., Kuklik, I., Leaper, R., Lehnert, K., Leopold, M., Lovell, P., Øien, N., Paxton, C.G.M., Ridoux, V., Rogan, E., Samarra, F., Scheidat, M., Sequeira, M., Siebert, U., Skov, H., Swift, R., Tasker, M.L., Teilmann, J., Van Canneyt, O. & Vázquez, J.A. (In prep). Abundance of harbour porpoise and other cetaceans in European Atlantic shelf waters. Thompson, P.M., Cheney, B., Ingram, S., Stevick, P., Wilson, B. & Hammond, P.S. (Eds) (2011b). Distribution, abundance and population structure of bottlenose dolphins in Scottish waters. Scottish Natural Heritage & Scottish Government Funded Report. SNH Commissioned Report No 354

87 TECHNICAL REPORT ON PRE-CONSENT MARINE MAMMAL DATA GATHERING AT THE MORL AND BOWL WIND FARM SITES Paul Thompson & Kate Brookes University of Aberdeen, Institute of Biological & Environmental Sciences, Lighthouse Field Station, Cromarty, Ross-shire IV11 8YJ. Report to MORL & BOWL 17th June 2011 Figures revised to include correct BOWL boundary, 18 th January

88 2

89 EXECUTIVE OVERVIEW & KEY FINDINGS The Moray Firth supports breeding populations of both harbour seals and grey seals, and sightings of at least fifteen species of cetaceans have been recorded in these waters. Recent work carried out for DECC, Marine Scotland, Oil & Gas UK and COWRIE has greatly enhanced our understanding of the occurrence of cetaceans in the vicinity of the MORL and BOWL windfarm sites. Nevertheless, following a review of the marine mammal data requirements for these offshore windfarm EIAs, it was agreed that there were several requirements either for additional data collection or more detailed analysis of existing data. The resulting work programme was developed in discussion with the developers and regulatory bodies, and has been conducted by the University of Aberdeen and SMRU Ltd at the University of St Andrews. Overall, the work programme has aimed to collect additional data that, in combination with existing data sources, can be used in Environmental Statements to address the following three objectives: Objective 1: To characterise the sites with respect to the marine mammal species present; detail seasonality and year-to-year variability in occurrence. Objective 2: To assess the density of animals at the proposed sites. Objective 3: To assess the likelihood of exchange between local SACs and the proposed wind farm sites. This main report describes the work carried out to address Objectives 1 and 2, and two additional Annexes describe work carried out by SMRU Ltd under Objective 3. Key Findings from these studies were: Objective 1- Characterisation & seasonal patterns Broad-scale deployments of echolocation detectors (C-PODs) in 2009 and 2010 showed that spatial variation in the occurrence of both porpoises and dolphins across the Moray Firth was consistent between years (Figure 3.11). 3

90 The combined C-POD data set from 2009 and 2010 indicates that harbour porpoises were widely dispersed across the Moray Firth (Figure 3.12a), with porpoises typically detected on every day at sites in the vicinity of the MORL and BOWL windfarm sites. In contrast, C-PODs detected dolphins regularly in the inner Moray Firth and along the southern Moray Firth coast, less frequently in the central Moray Firth, and at intermediate levels in the vicinity of the MORL & BOWL sites (Figure 3.12b). Field comparisons confirmed that data from analogue (T-POD) and digital (C- POD) echolocation detectors were comparable (Figure 3.18), thereby providing a long time-series of data ( ) that showed similar levels of occurrence of both dolphins and porpoises in the vicinity of the Beatrice demonstrator site over time (Figure 3.20). More intensive deployments of C-PODS on the BOWL (Section 3.6.3) and MORL (Section 3.6.4) sites have been made since May 2009, demonstrating that both porpoises and dolphins use these areas throughout the year. Seasonal fluctuations in the median number of hours per day that porpoises were detected in the BOWL (Figure 3.24) and MORL (Figure 3.27) sites. Whilst C-PODs provide important information on spatial and temporal variation in the occurrence of dolphins, they cannot currently be used to identify different dolphin species. Visual sightings of cetaceans from summer (April-October) boatbased and aerial surveys that were made over and around the BOWL and MORL sites in 2010 were collated, and maps of dolphin sightings from each survey team presented to provide information on the species of dolphins detected in the Outer Moray Firth. Aerial surveys using observers from NERI and WWT Consulting detected Risso s dolphins, common dolphins and white-beaked dolphins in offshore waters, but bottlenose dolphins were only sighted within 10km of the southern Moray Firth coast (Figure 3.3). Boat based surveys conducted by Natural Power detected Risso s dolphins, common dolphins and white-beaked dolphins and a single 4

91 bottlenose dolphin over the MORL site (Figure 3.4). Boat based surveys over the BOWL site conducted by IECS reported a single encounter with a group of common dolphins, four encounters with groups of bottlenose dolphins, but no sightings of Risso s or white-beaked dolphins (Figure 3.5). A previous review of all historic data from the Outer Moray Firth found that all previous reported sightings around the BOWL and MORL sites were of Risso s dolphins, common dolphins and white-beaked dolphins (Thompson et al. 2010). These data highlight the difficulty of interpreting data collected from surveyors with different levels of experience, particularly where survey teams are selected for the primary purpose of detecting seabirds. Nevertheless, even if all visual species identifications are accepted, the majority of encounters with dolphins over the MORL and BOWL sites are likely to be species other than bottlenose dolphins. Objective 2 Density estimates Harbour porpoises were by far the most commonly encountered species in the Outer Moray Firth. Line-transect aerial surveys conducted in August and September were used to estimate the density of this species in different sample blocks in the Outer Moray Firth. Density within the sample block covering the MORL and BOWL sites was 0.81 porpoises per km 2 (Table 3.5), which is comparable with the highest density areas recorded in European waters during the broad-scale SCANS and SCANS II surveys. Spatial variation in the relative density of harbour porpoises across the Moray Firth was determined using over 1000 sightings from combined boat and aerial surveys in a habitat association model. Predicted values from the resulting GAMM were then scaled by the measure of absolute density from the aerial survey to predict the likely number of porpoises in each 4 x 4 km grid cell across the Moray Firth (Figure 3.8). These modelled distributions will underpin future assessments of the impact of disturbance and displacement for porpoises. There were too few sighting of dolphins to produce density estimates for individual species, but a combined estimate for all species showed that densities were low in offshore areas, but higher along the coast (Table 3.7). Presence only data from all 5

92 available surveys conducted since 1982 were used within a classification tree to assess spatial variation in the likely species composition of dolphins encountered across the Moray Firth (Figure 3.15). These data, along with those from broader scale SCANS II data and regional estimates of bottlenose dolphin population size, will be used to underpin future assessments of the impact of disturbance and displacement for dolphins. VHF, Satellite and GPS-GSM data collected over two decades from 37 tagged harbour seals were integrated to model habitat association and predict relative densities of foraging seals across the Moray Firth by scaling to total population size (Annex 1, Figure 11). These modelled distributions will underpin future assessments of the impact of disturbance and displacement for harbour seals. Objective 3 Connectivity between the MORL & BOWL sites & SAC s Broad band acoustic recorders were deployed in the MORL and BOWL sites between July and October 2010, and whistle classification software developed by SMRU Ltd to assess the likelihood that dolphins in this area were bottlenose dolphin (that may be from the Moray Firth SAC) rather than other species such as Risso s dolphin, common dolphin or white-beaked dolphin. Comparable data were collected at a site within the Moray Firth SAC. Dolphins were recorded on eight occasions within the MORL and BOWL sites, and all encounters were classified at species other than bottlenose dolphins. In contrast, 79% of encounters within the SAC were of bottlenose dolphins (Annex II, Figure 7). Analysis of the combined VHF, Satellite and GPS-GSM telemetry dataset demonstrated that some harbour seals from the Dornoch Firth & Morrich More SAC are likely to spend time foraging within the MORL and BOWL sites (Annex I, Figure 11). 6

93 1. Background The Moray Firth supports breeding populations of both harbour seals and grey seals, and sightings of at least fifteen species of cetaceans have been recorded in these waters (Grellier & Lacey 2010). Whilst there has been a long history of research on marine mammals in this area, this has tended to focus on coastal and inshore areas. This work has lead to estimates of distribution, population size and trends for both harbour seals and bottlenose dolphins. Some studies of marine mammals in the vicinity of the Smith Bank were carried out as part of the Beatrice Demonstrator project but, in general, much less is known about the distribution and abundance of marine mammals in more offshore areas within the Moray Firth. Recently, this lack of data has been highlighted in Appropriate Assessments for oil and gas activities, particularly in relation to the potential use of offshore areas by bottlenose dolphins that inhabit the Moray Firth SAC. To address this, the Department of Energy & Climate Change (DECC) funded work in 2009 and 2010 to review existing data on cetacean distribution in the outer Moray Firth, and collect additional data to support the management of oil and gas activities in the area. Work carried out in 2009 and 2010 under this DECC funded project greatly enhanced the data available to assess other developments in the outer Moray Firth (Thompson et al. 2010a). Nevertheless, following a review of the marine mammal data requirements for BOWL s environmental impact assessments (Grellier & Lacey 2010), it was agreed that there were several requirements either for additional data collection or more detailed analysis of existing data. Not only are these data required to support consenting for both the BOWL and MORL developments, but they will also be needed to provide a more robust baseline for subsequent monitoring programmes during the construction phase of these projects. The resulting work programme was developed in discussion with the developers and regulatory bodies. The work has been co-ordinated by the University of Aberdeen, with key studies being carried out under sub-contract to SMRU Ltd at the University of St Andrews. Overall, the work programme has aimed to collect 7

94 additional data that, in combination with existing data sources, can be used in Environmental Statements to address the following three objectives: 1. To characterise the sites with respect to the marine mammal species present; detail seasonality and year-to-year variability in occurrence. 2. To assess the density of animals at the proposed sites. 3. To assess the likelihood of exchange between local SACs and the proposed wind farm sites. The key requirement under Objective 1 was additional data on cetacean distribution and occurrence. In particular, information on seasonal and inter-annual variation in the occurrence of key species (eg. harbour porpoises and dolphins) within the development areas was required. These data can then be used to complement existing visual data (eg. Reid et al. 2003; Thompson et al. 2010a) and the data that will result from the boat-based seabird and marine mammal surveys being conducted by BOWL and MORL. The agreed approach for these additional studies was to extend passive acoustic monitoring studies that were initiated during the Beatrice Demonstrator project (Bailey et al. 2010; Thompson et al. 2010b) and which had been further developed through the DECC funded project (Thompson et al. 2010a). The key requirement under Objective 2 was for robust region-specific density estimates of cetaceans in and around the BOWL and MORL sites. Such data are required for EPS licences to estimate the number of animals that may be disturbed, where the use of region-specific data is likely to be more appropriate than using the broader scale density estimates available through SCANS (Hammond et al. 2002) or SCANS-II (SCANS-II 2008). The precise area of interest for these density data will depend upon the results of concurrent noise modelling studies, making it difficult to pre-define suitable survey areas. However, during August & September 2010, DECC funded the University of Aberdeen to conduct an intensive series of aerial linetransect surveys across two 25 x 25 km survey blocks in the central Moray Firth. One of these sites covered the whole BOWL site and a large part of the MORL site. The agreed approach for these studies was to use the high quality data from these aerial surveys within habitat association models (see e.g. Bailey & Thompson 2009), and 8

95 predict the density of cetaceans within the development sites and their surrounding waters. Subsequently it was agreed with MORL and BOWL to explore the potential for integrating available data from boat-based surveys into these habitat association analyses. The key requirements under Objective 3 were for data to assess the likely connectivity between the BOWL and MORL development sites and marine mammal SACs in the region. The two species of concern in this respect are bottlenose dolphins using the Moray Firth SAC (Thompson et al. 2011) and harbour seals using the Dornoch Firth and Morrich More SAC (Cordes et al. 2011). Bottlenose dolphins. Previous studies using echolocation detectors (C-PODS) had shown relatively high levels of dolphin activity in the Outer Moray Firth (Thompson et al. 2010a). But used alone, C-PODS cannot discriminate between the bottlenose dolphins that might be using the Moray Firth SAC and the other species that are potentially using this area (primarily common, white-beaked and Risso s dolphin). However, SMRU Ltd has developed new approaches which can use broadband recordings to better discriminate between different dolphin species. Given the limited coverage of visual surveys in offshore areas, it was agreed that these passive acoustic techniques using broadband sound recordings would provide the greatest opportunities for assessing the probability that dolphins detected in the Outer Moray Firth were likely to be bottlenose dolphins. Harbour seals. Over the last two decades, over 37 individual harbour seals from the Dornoch Firth and Morrich More SAC and the nearby Loch Fleet NNR have been tracked using a variety of techniques (VHF, Satellite and GSM telemetry) (Thompson et al. 1994; Sharples et al. 2008; Cordes et al. 2011). It was agreed that the most appropriate method for addressing this objective for harbour seals was to use these existing data to underpin predictions of use of the BOWL & MORL sites using habitat association modelling. However, the different error structures for each of these technologies required the development of a novel Bayesian state-space approach to integrate these data into a single modelling framework. Existing procedures could then be used to predict habitat usage (Aarts et al., 2008) and estimate how many 9

96 harbour seals from the Dornoch Firth and Morrich More SAC are likely to use habitats within the development areas (Matthiopoulos et al., 2004). In the main body of this report, we describe the work carried out to address Objectives 1 and 2. Two additional reports cover work carried out by SMRU Ltd under Objective 3 are presented as annexes to this report; one assessing linkage with the bottlenose dolphin SAC and one assessing linkage with the harbour seal SAC. 10

97 2. Methodology 2.1 Visual surveys Data sources Primary data sets for this part of the study were collected by the University of Aberdeen through previous and ongoing surveys carried out in relation to the Beatrice Demonstrator Project and assessments of the impact of seismic surveys. These included two data sets that were collected using boat-based line transect surveys, and one that was collected using aerial line-transect survey. All these data were collected during the summer months (April-October). Additional data on sightings of harbour porpoises and dolphins of all species were also made available from the boat-based seabird and marine mammal surveys that were carried out during April to October 2010 by Natural Power (on behalf of MORL) and IECS (on behalf of BOWL). Each of the datasets used broadly similar survey methods. All used line-transect methods and collected effort data in the format of transect distance surveyed. All recorded the location, species and number of animals sighted, although the number and experience of observers varied between surveys. No deviation from the survey track line was made when animals were sighted AU Boat surveys within the Moray Firth SAC (2004, 2005) These surveys were carried out to provide baseline data from the Moray Firth SAC for the Beatrice Demonstrator project. As outlined in Bailey & Thompson (2009), they aimed to assess habitat associations of bottlenose dolphins and harbour porpoises along the survey route and model their relative abundance across the Moray Firth SAC to identify hotspot areas for these species. Line transect surveys were designed to provide representative coverage across the Moray Firth SAC (Figure 2.1) during the summers of 2004 and 2005 (Table 2.1). Surveys were conducted using a 8.5 m Newhaven Sea Warrior, and the single observer recorded sightings of any marine mammals from the top of the wheelhouse 11

98 at approximately 3.5 m above sea level. Total survey distance was 1628 km, and survey speed was approximately 7 knots. Table 2.1. Days of survey effort carried out during the University of Aberdeen s boatbased surveys within the Moray Firth SAC. Year Month # Survey Days 2004 August September October April May June July 1 Figure 2.1. Map of the survey tracks used during the University of Aberdeen s boatbased surveys within the Moray Firth SAC. 12

99 AU Boat surveys in Outer Moray Firth (2009) These surveys were carried out in the summer of 2009, to collect data for DECC to support their assessment of proposed oil and gas exploration within the Moray Firth (Thompson et al. 2010a). Surveys covered a large geographical area at relatively low resolution, with the aim of determining which species were present within the offshore waters of the Moray Firth at a broad scale. Three different vessels were used; two fishing vessels and a converted lifeboat. Observation height varied between vessels, but was at least 5 m above sea level. Survey speed was approximately 8 knots. Two observers were on watch at all times, each scanning one side, in the 180 arc ahead of the boat. Total survey distance was 1671 km. Table 2.2. Days of survey effort carried out during the 2009 boat-based surveys. Year Month # Survey Days 2009 June August September October 2 Figure 2.2. Map of total effort during the 2009 Univ of Aberdeen boat based surveys. 13

100 AU Aerial surveys in Outer Moray Firth (2010) Aerial surveys were carried out in the summer of 2010 by the University of Aberdeen. The primary reason for these surveys was to estimate the density of harbour porpoises in two 25x25km survey blocks as part of the programme of work investigating impacts of seismic surveys on cetaceans. In addition, surveys were designed to compare the occurrence of different dolphin species along the north and south coasts of the Moray Firth to support assessments of connectivity with the Moray Firth SAC (Figure 2.3). Surveys followed the line-transect procedures used for SCANS and SCANS-II and used experienced aerial surveyors from NERI and WWT Consulting Ltd. A Partenavia 68 aircraft, fitted with bubble windows was flown at a speed of 100 knots, at 600 ft above sea level. Two observers scanned the area on either side of the aircraft. A total of 5664 km of transect were surveyed during five survey days in August and eight in September Surveys that were incomplete, or carried out in poor sighting conditions were excluded from analysis, giving a final total effort of 4784 km. Further details of survey protocols are given in section Figure 2.3. Map of survey tracks used during the 2010 aerial surveys 14

101 Natural Power surveys of MORL site (2010) These surveys were carried out in 2010 as part of a two year programme of bird and marine mammal surveys to support the MORL ES. Monthly surveys began in April Only data collected up to October 2010 are presented here to allow comparison with University of Aberdeen data. In total, 3015 km of survey effort was carried out during this period. A variety of survey vessels were used, but all had survey platforms at least 5m above sea level and travelled at approximately 10 knots on a series of standard transects (Fig 2.4). Table 2.3. Days of survey effort carried out during the 2010 Natural Power surveys Year Month # Survey Days 2010 April May June July August September October 4 Figure 2.4. Map of survey tracks used for the MORL April to October 2010 surveys 15

102 IECS Boat surveys of BOWL site (2010) These surveys were carried out in 2010 as part of a two year programme of bird and marine mammal surveys to support the BOWL ES. Monthly surveys began in November 2009, but only data collected between April and October 2010 are incorporated into this analysis to allow comparison with University of Aberdeen data. Total survey effort during this period was 1390 km. The survey vessel was a converted lifeboat, with an observation height of approximately 3 m above sea level. Table 2.4. Days of survey effort carried out at during the 2010 IECS surveys Year Month # Survey Days 2010 April May June July August September October 0 Figure 2.5. Map of survey tracks used for the BOWL April to October 2010 surveys 16

103 2.1.2 Habitat association modelling To assess habitat association of cetaceans, survey and habitat data were summarised across a 4x4 km grid. Based upon earlier analyses of data from within the SAC (Bailey & Thompson 2009), four habitat variables were assessed: depth, sediment type, slope and distance to the coast (Figure 2.6). For depth, slope and distance to the coast, a mean value for every 4x4 km grid cell was calculated using BGS data available through SeaZone. Sediment was processed to give the proportion of sand and gravelly sand sediments within each cell, on the basis that sand eels prefer habitat with high proportions of these sediments (Holland et al., 2005) and it is reasonable to assume that porpoises would seek out these areas when foraging. For some cells that were surveyed in the inner Moray Firth, BGS habitat data were not available, and data from these cells were therefore removed from the analysis. To ensure a good estimate of the proportion of sand and gravelly sand within the cell, any cell with less than 50% data coverage was removed. The slope variable was highly right skewed with one observation much larger than the others. This observation was removed, giving a range of slopes between 0 and Some coastal cells had mean depth values that were above sea level, and a minimum mean depth of 5 m was therefore used to ensure that most of the cell was available to porpoises. Depth was also right skewed, so a maximum depth of 73.5 m was used. Survey effort and sightings data were also split into the same 4x4 km grid cells. Where multiple surveys covered the same cells, the data were treated separately, leading to some cells being included in the analysis up to four times. Cells were not included if they contained less than 1 km of effort. In total 429 cell observations were included in the model, from 241 unique cells (Figure 2.7). Each observation was coded to reflect the data collection method; either aerial or boat based, to allow models to account for the potential difference in sighting rate between these methods. 17

104 Figure 2.6. Habitat variables (depth, slope, distance to coast and sediment type) summarised over a 4x4 km grid. Figure 2.7 The total amount of survey effort in each 4 x 4 km cell. 18

105 Harbour Porpoise model There were over 1000 sightings of harbour porpoises from the combined surveys, but relatively few sightings of different dolphin species (see results). Whilst this provided good opportunities for modelling harbour porpoise habitat associations, there were insufficient data to model individual species distribution for other species. For the porpoise models, data exploration indicated that depth and distance from the coast were highly collinear, so distance from the coast was removed from the model. Porpoises are known to occur at a large range of distances from the coast, and it is more likely that they respond to depth in terms of available food. Models were constructed with a count of animals in each 4x4 km grid cell for each dataset, along with a value for each habitat variable. The log of the total transect length within each grid cell was used as an offset variable. The relationship between porpoise numbers and depth was non-linear, so generalised additive mixed models (GAMM) were used. These models allow nonlinear relationships and also account for the pseudoreplication caused by including some cells more than once. The initial model included depth, the proportion of the sediment that was sand or gravelly sand, slope and the log of effort as an offset. Cell identity was included as a random effect. Models were weighted by the ratio of effort to the maximum value of effort, thereby allowing cells with more effort to have more influence on the estimated values from the model. Initial models were found to be overdispersed when using a Poisson distribution and the final models therefore used a negative binomial distribution. Model selection was based on AIC (Akaike, 1974), which gives information on the accuracy of the model, taking into account its complexity. Model selection aims to minimise the AIC score. Analysis was carried out in R version and the mgcv package (Wood, 2008) was used for GAMM analyses. 19

106 Dolphin model Although there were insufficient data to produce habitat association models for individual dolphin species, we were able to use these survey data in classification trees (De'ath & Fabricius 2000) to assess the likely species identify of dolphins that may be encountered in different parts of the Moray Firth. In particular, we developed this analysis to assess the likelihood that any dolphins encountered in offshore areas (see results in section 3.4) were bottlenose dolphins. Habitat and other spatial variables (eg distance to coast) were recorded at each of the locations where visual sightings of different dolphin species had been made. Classification trees were then developed by repeatedly splitting the dataset in two, until most animals were assigned to a unique species group on the basis of these different spatial variables. The resulting tree could then be used to predict the proportion of each species that might be expected in an area given its habitat characteristics. These presence only methods do not account for effort so, used alone, they cannot provide a prediction of the number of animals that might be found in an area. However, this approach does tell us that, if dolphins were present, this is the likely species composition that we would find in different areas. We used available sightings of dolphins that were identified to species within the Moray Firth between 1982 and Several datasets described in Thompson et al. (2010a) were used, in combination with the four datasets described in section Datasets without counts of animals or with poor locational precision were excluded from the analysis. In total, eight datasets were included (Table 2.5). Individual dolphin sightings were used in the classification tree. Each sighting was assigned the habitat values averaged over the 4x4 km grid cell that it was seen within. These were the same habitat values used in the porpoise model, described in section The tree was built using R version and the tree package (Ripley, 2010). Four habitat variables; depth, distance to coast, slope and sediment type, were included, as well as the X and Y coordinates of the middle of the grid cell. The tree was weighted by the count of animals in each sighting. 20

107 Table 2.5. The number of sightings and count of dolphins used from each of the datasets included in the analysis. JNCC Seabirds at Sea data include data from the RSPB surveys in 1982 and Dataset Year Number of dolphin sightings Number of animals recorded BOWL JNCC Seabirds at Sea JNCC seismic MMO MORL Crown Estate UoA AFEN UoA 2009 boat UoA 2010 aerial UoA SAC UoA Photo-ID The analysis was run twice, once with all of the data, and once excluding data collected by IECS over the BOWL site which, given its offshore location, contained an atypically large number of bottlenose dolphin sightings relative to sightings of other species (see results). Predictions were then made from the output of each analysis, on the basis of the habitat characteristics of cells. 21

108 Estimation of density from line-transect aerial surveys A key part of the DECC funded assessment of the impacts of seismic exploration involved an estimation of changes in cetacean density (primarily harbour porpoises) at an impact and control site, before and during a proposed seismic survey in September These two survey blocks were each 25 x 25km, with the control block (Figure 2.8, Block B) covering a large part of the BOWL and MORL development areas. B A Figure 2.8. A map of the Moray Firth showing the position of the aerial survey blocks in relation to the location of the MORL and BOWL sites. 22

109 Figure 2.9. Map showing the total aerial survey effort used in the calculations of density and the regions used to estimate abundance (pink = coastal; yellow = central Moray Firth; blue = outer Moray Firth). As outlined in section , aerial surveys were carried out during August and September of 2010, from a Partenavia 68 aircraft fitted with observer bubble windows. Within the two offshore blocks, parallel north/south transect lines spaced at 4 km were flown on each survey. An offset of 1 km was used and the starting position was selected randomly so that during the course of the survey period, the blocks were covered at 1 km spacing (Figure 2.9). On the coastal transects, the aeroplane flew parallel to the coast at a distance of 1 km offshore, returning on a parallel transect 5km offshore. The two blocks and transect were surveyed 9 times, the north coast route 6 times and the south coast survey route 5 times. The aim of these surveys was to use standard procedures available in the program Distance (Thomas et al. 2009) to calculate density and abundance. Data from the whole 45-day survey period were pooled to provide estimates of density both across the entire survey area and in different sub-areas. 23

110 Observers followed protocols developed for SCANS and SCANS-II aerial surveys to collect data. Observations were made out of different sides of the aeroplane, and the two observers each recorded sightings into separate voice recorders. Time, species, number of animals and the declination angle to the sighting were recorded as a minimum. GPS data were recorded automatically every five seconds and these data were subsequently interpolated to give the location of the aeroplane when the sighting was made. The horizontal distance from the trackline to the sighting was later calculated from the declination angle and used to calculate the position of each animal seen. Environmental variables were recorded by a third observer and included Beaufort sea state and glare intensity. A subjective measure of sighting conditions was recorded as four levels; poor, moderate, good and excellent. These levels related to the likelihood that a porpoise would be observed if it were present, and took into consideration all variables that might influence observers ability to detect animals. All data collected under poor sighting conditions were removed prior to analysis using Distance. One of the key assumptions of Distance analysis is that all animals on the track line are detected i.e. g(0), the detection probability on the track line, is equal to 1 (Thomas et al. 2009). Data collected in studies such as this fail to meet this assumption in two ways, and this must be accounted for when fitting a detection function to the data. The first problem is that observers were unable to view the sea areas directly below the aeroplane. This blind sector extended through the closest 20, which is equivalent to the closest 66 m to the aeroplane. To account for this, data were left truncated at 66 m, meaning that the program did not try to fit a detection function to this area. The second failure of this assumption occurs because marine mammals spend a proportion of their time under water, and are therefore not available for detection at all times, even when they are within the area being surveyed. To account for this, the probability of detecting an animal on the track line, g(0) is estimated and used as a multiplier when estimating density and abundance. Given the much larger dataset available from the SCANS-II aerial surveys of the North Sea in 2005, we used their value of 0.45 for g(0) for harbour porpoises (Hammond et al. In prep). This value was calculated using the racetrack method 24

111 where the aeroplane circles back around a sighting to determine the re-sighting rate (Hiby, 1999). Environmental covariates that may have affected detection were included when modelling the detection function. Four covariates were tested; observer identity, sea state, sighting conditions and glare intensity. These were added using a forward stepwise selection procedure based on AIC. Observer identity and sighting conditions were retained in the detection function as they contributed to a lower AIC. For porpoises, the same detection function was used throughout all analyses and was estimated using the entire dataset. There were insufficient sightings to estimate detection functions for different dolphin species. On the assumption that the detection of the different dolphin species likely to occur in this region is similar, we produced a single detection function for all dolphins, and used this to provide an estimate of density and abundance for all pooled dolphins of all species. Ongoing work is exploring whether this dataset can be used to produce robust estimates of density of the individual dolphin species. In the meantime, the output from the classification tree analysis provides an indication of the likely species composition of dolphins in different parts of the Moray Firth. Density estimates for porpoises and dolphins were calculated both for the entire survey area and for the sub-areas areas separately. To provide an estimate of the total number of individual porpoises and dolphins in the region, we stratified the region into three areas each represented by one of the main sampling areas used for our aerial surveys; a coastal strip within 5km of land, a central Moray zone and an outer Moray Firth zone (Figure 2.9). 25

112 2.2 Passive acoustic monitoring As highlighted previously, cetaceans spend most of their time underwater, and are often difficult to detect even when at the surface. They do, however, regularly vocalise, and this has meant that passive acoustic monitoring studies have been increasingly used to provide fine-scale spatial data on cetacean distribution and temporal trends in occurrence within key areas (Clark & Clapham, 2004; Verfuss et al., 2007; Van Parijs et al. 2009). The development of automated devices that can remotely record cetacean echolocation clicks for periods of up to 6 months has proved particularly important for supporting assessments of the impact of different anthropogenic activities including fisheries by-catch and marine renewables (Thompson et al. 2010b). These Timing POrpoise Detectors (T-PODS) were originally designed to study harbour porpoises (Thomsen et al., 2005), but can be programmed to detect echolocation clicks from a range of other species (Philpott et al., 2007). For harbour porpoises, it has been estimated that animals can be detected within a distance of approximately 200m around the T-POD (Tougaard et al., 2006), whereas field studies indicate that bottlenose dolphins can be detected at distances up to 1200m (Philpott et al., 2007; Bailey et al. 2010). In 2009, production of T-PODs ceased, and these were replaced with a new digital device, the C-POD ( The University of Aberdeen have been conducting passive acoustic studies of cetaceans in the Moray Firth since This has involved studies using both T- PODs and C-PODs, in both coastal and offshore waters. Whilst some data from the Smith Bank have been published (eg. Bailey et al. 2010; Thompson et al. 2010b), integration of these data with additional unpublished and new data now provides an opportunity to explore temporal patterns of occurrence of harbour porpoises and dolphins on the Smith Bank over the last 5 years. The following sections outline the different data sets available for these analyses, and describe the device characteristics and analysis methods for T-PODs and C- PODs. 26

113 2.2.1 Data sources Beatrice Demonstrator study ( ) A key objective of the Beatrice Demonstrator project was to develop and/or validate methods that could be used for assessing changes in the occurrence of cetaceans in response to offshore wind turbine construction. As a result, some studies were conducted in inshore waters, where visual observations could be used to validate acoustic detections on T-PODs (Bailey et al. 2010). In addition, data were collected at other sites in the Moray Firth between August 2005 and December First, at a site near the Beatrice Demonstrator turbines, and secondly, at a site 40km to the south near Lossiemouth (Figure 2.10). Our original aim was to use this second location as a control site. In practice the identification of suitable control sites was constrained by the limited information available at this time on cetacean distribution in the Outer Moray Firth, and uncertainties over the scale of the potential impact (see discussion in Thompson et al. 2010b). Data from August to October of 2005, 2006 & 2007 have previously been published in Thompson et al. (2010b) but, until the present study, there has been no analysis of the full dataset to explore long-term variation in acoustic detections at this site. Because earlier studies during the Beatrice Demonstrator project were based upon T-PODS (Bailey et al. 2010), and more recent work has been conducted using C-PODS (Thompson et al 2010b), any investigation of these long-term patterns first required a comparison of performance of these two different devices. To address this, we deployed both a C-POD and a T-POD at 14 of the moorings used during the 2010 field season (see Section ). In each case, the two devices were cable tied side by side at the same position on the mooring. 27

114 Figure Sites used for passive acoustic monitoring during the Beatrice Demonstrator study, From Bailey et al SNH & SEERAD Studies ( ) Following the Beatrice Demonstrator Project, further passive acoustic monitoring in the Moray Firth was conducted as part of a broader scale SNH and SEERAD funded study of the distribution and abundance of bottlenose dolphins in Scottish coastal waters (Thompson et al. 2011b). No additional data were collected from the windfarm development areas. Nevertheless, these studies continued the time-series of data at the Lossiemouth site, and extended this to other sites along the southern Moray Firth coast (Figure 2.11) which could potentially support ES work related to cable installations. Almost all of these data were collected using T-PODs, but the newly developed C-PODs were deployed at three sites in the final year of the study. 28

115 N W E Helmsdale # # Beatrice Brora# # Block 17/3 (3) Tarbat Ness # # Block 17/3 (1) # Block 17/7 Sutors # Lossiemouth # # # Spey Bay Chanonry Culbin # # Nairn # Kessock S # Cruden Bay # Stonehaven # Arbroath # Fife Ness Kilometers Figure Sites used for passive acoustic monitoring during SNH and SEERAD funded studies From Thompson et al. 2011b DECC Study ( ) DECC funded studies in the 2009 and 2010 led to the deployment of an extensive array of C-POD monitoring across the Moray Firth. In 2009, deployments were made for the period May-Nov (Thompson et al. 2010a). In 2010, deployments were made for the period July-Dec. In both years, these studies involved deployments at multiple sites within both the BOWL and MORL development areas (Figure 2.12). All primary deployments were made using C-PODs. For a subset of moorings in 2010, we paired a C-POD with a T-POD to provide data for a comparison of detection rates for the two devices. 29

116 a) 2009 b) 2010 Figure Sites used for passive acoustic monitoring during DECC funded studies in a) 2009 and b) MORL/BOWL funded studies ( ) Between 2009 and 2011, most acoustic data collected through these other studies were collected during the period July-November. To complement these data and assess seasonal patterns of occurrence, BOWL and MORL contracted the University to make additional deployments within their development areas at other times of year. 30

117 In the first of these winters (2009/10), deployments were made at two sites within both the MORL and BOWL area, although one device from the MORL area was lost. In the second winter (2010/11), deployments were made at 15 sites within the MORL area and 6 sites within the BOWL area. Locations of all the PAM sites within and in the vicinity of the MORL and BOWL development areas are shown in Figure Figure Sites at which T-PODs or C-PODs have been deployed within or close to the MORL and BOWL development areas T-PODs T-PODs incorporate a hydrophone, analogue processor and digital timing system that automatically logs the start and end of each echolocation click to 10 s resolution. In every minute, the T-POD runs 6 successive scans within different userdefined frequencies, logging detections for periods of up to 5 months. An accompanying software program is used to post-process the recovered data, detect characteristic click trains, and remove noises from other sources such as boat sonar 31

118 (see for details). Resulting data on the number of cetacean click trains recorded in each minute can be used to determine the presence or absence of target species in different time periods, or to identify the timing and duration of encounters with target species. In our studies, we used Version 4 and Version 5 T-PODs to detect echolocation click trains. Following guidelines for use in areas where both harbour porpoises and bottlenose dolphins might be detected, T-PODs were configured to detect clicks from dolphins and porpoises on alternate channels. For dolphins we set a target frequency of 50 khz and a reference frequency of 70 khz on three of the channels. For porpoises, we set a target frequency of 130 khz and reference of 92 khz on the three other channels. All data were processed using version 8.24 of the manufacturer s software (version 4.1 train filter). This train detection algorithm in the T-POD software assigns trains into several different categories. We used the category CET ALL, which combined both the high probability click trains (CET HI) and less distinctive trains (CET LO), following the recommendation of the manufacturer ( and previous assessments of performance for detecting harbour porpoises (Thomsen et al., 2005). T-POD data were subsequently used to determine those hours in which dolphins and porpoises had been detected at each site on each day. Validation studies in the inner Moray Firth had previously shown that false porpoise detections sometimes occurred within a series of dolphin detections, even when dolphins were confirmed to be the only cetacean present in the area. In areas where dolphins are common and porpoises are rare, this can artificially inflate the occurrence of porpoises. In such areas, this problem can be avoided by only considering porpoise detections as positive if they occurred during a sampling period in which no dolphin clicks were detected. In practice, this was not an issue in the Outer Moray Firth as dolphin detections were extremely rare C-PODs In 2008, production of the T-POD ceased following the development of a digital echolocation detector, the C-POD ( A V.0 C-POD was produced 32

119 during 2008, and the first V.1 C-POD units were available in June The C-POD continuously monitors within the range of khz for possible cetacean clicks, and records the centre frequency, frequency trend, duration, intensity, and bandwidth of each click. As with T-PODs, these data are then post-processed to differentiate between dolphins, porpoises and other high frequency sounds such as boat sonar. The output indicates the level of confidence in classification of the detection as a cetacean echolocation click by classing each as CetHi, CetMod or CetLow. Prior to deployment in the Outer Moray Firth, all new C-PODs were first bench tested using an artificial high frequency noise source. Short trial deployments of 1-2 days were then made in the mouth of the Cromarty Firth, an area which is used by bottlenose dolphins on a daily basis during summer, to ensure that they were detecting dolphin echolocation clicks. Once recovered, data were downloaded and analysed using V1.054 of the C-POD train filter to identify detections of harbour porpoises and dolphins. In these analyses, we used only CetHi and CetMod detections to estimate the number of hours that either porpoises or dolphins were detected at each sampling site on each day. Because of greater levels of fishing effort in the outer Moray Firth, we modified the mooring design used previously in inshore areas. In this study, we used moorings with a single riser from a 100 kg or 150 kg weight, and a larger surface Dhan buoy with radar reflector and flag (Figure 2.14). As in previous studies, PODs were attached to the riser at a height of approximately 2-6 m above the seabed. In 2009, offshore deployments were made from FV Rois Mhairi and some recoveries were made from FV Alba, MV Topcat and MV Solstice. In 2010 and 2011 all deployments and recoveries were made from MV Solstice. 33

120 Figure Single riser mooring design with Dhan buoy used to suspend PoDs in the water column 34

121 3. Results 3.1 Cetacean sightings during different visual survey programmes Overall, there were over 1000 encounters with a total of seven different species of cetacean during the visual survey programmes outlined in section 2.1 (Table 3.1). Table 3.1. Number of sightings of cetaceans recorded during each of the different visual survey programmes. Species AU SAC AU Boat AU Aerial MORL BOWL Harbour porpoise Bottlenose Dolphin White-beaked dolphin Risso s Dolphin Common Dolphin Unidentified Dolphin Killer Whale Minke Whale Maps presenting raw data on the distribution of all sightings of harbour porpoises and dolphins are shown in sections in Figures In both these figures and in Table 3.1, we include all those sightings where the different survey teams were confident about species identification. 35

122 Figure 3.1. Sightings of a) dolphins and b) harbour porpoises made during the University of Aberdeen surveys within the Moray Firth SAC. All sightings of dolphins on these surveys were reported as bottlenose dolphins. 36

123 Figure 3.2. Sightings of a) dolphins and b) harbour porpoises made during the University of Aberdeen s 2009 boat based surveys in the Outer Moray Firth. 37

124 Figure 3.3. Sightings of a) dolphins and b) harbour porpoises made during the University of Aberdeen s 2010 aerial surveys of the Outer Moray Firth. 38

125 Figure 3.4. Sightings of a) dolphins and b) harbour porpoises made during Natural Power boat surveys of the MORL site between April and October of

126 Figure 3.5. Sightings of a) dolphins and b) harbour porpoises made during IECS boat surveys of the BOWL site between April and October

127 3.2 Modelling harbour porpoise habitat association & distribution Models were based on over 1000 sightings of porpoises from the five different surveys (Table 3.2) Table 3.2. Total effort and number of sightings of animals used from each dataset once datasets were adjusted to remove data from those cells where no habitat data were available. Dataset Total effort (km) used in models Total porpoises used in models UoA SAC UoA 2009 boat UoA 2010 aerial BOWL MORL The model with the lowest AIC (1739) excluded slope and then method, but a 2D smoother (the GAM equivalent of an interaction term) for depth and proportion of sand and gravelly sand was included. The final model therefore contained only this 2D smoother and effort as an offset in the fixed effects, and cell identity in the random effects. The r 2 of this model was The random effects of the model showed that there was a relatively strong correlation, of 0.69 between observations from the same cell. This was calculated as: where is variance of the random intercept and is variance of the residual term (Zuur et al., 2009). In this case, a=0.710 and b=

128 Table 3.3. Results of the GAMM of porpoise counts. Parametric coefficients Estimate Standard error t-value p-value Intercept <0.001 Smooth terms Estimated df Reference df F p-value te(depth,psndgrvsnd) <0.001 The shape of the 2D smoother (Figure 3.6) produced by the final GAMM of porpoise numbers shows that few animals were sighted in shallow or deep waters, but more were found at intermediate depths of around 40 m to 50 m. At these depths, increases in the proportion of sand and gravelly sand lead to an increase in the probability of sightings. The peak in porpoise sightings in deep water with low proportions of sand and gravelly sand is a result of very few observations with these habitat characteristics, and any predictions for deeper water areas are therefore extremely uncertain. Figure 3.6. Two dimensional smoother used in the porpoise habitat association model to described the relationship with both depth and the proportion of the sediment made up of sand and gravelly sand. 42

129 The results of this model were then used to predict spatial variation in the relative abundance of porpoises across the Moray Firth. The predicted number of porpoises in each 4x4 km cell was based upon the depth and proportion of sand and gravelly sand within that cell, and standardised for a constant unit of effort. Figure 3.7a shows the predicted number of porpoises encountered in different parts of the Moray Firth for a standard 1km of survey effort, and Figure 3.7b shows the standard error of this prediction for each cell. Although cell identify was included as a random effect, model validation plots indicted there was still some spatial correlation in residuals. This means that predicted densities for cells outside the main survey area (see Figure 2.7) are the most uncertain (Figure 3.7b). This is particularly so for deep water areas due to the interaction with depth and sediment type (Figure 3.6), and the lack of survey effort in waters deeper than 80m. In this report, we therefore do not make predictions for any cells where water depth is greater than 120m, and the higher uncertainty for the cells with depths in the range m should be recognised when interpreting these data. These values for the relative abundance of porpoises were subsequently scaled to absolute abundance using the density estimates obtained from the aerial line transect survey (see section 3.5). This was based on the highest quality data from the 98 4 x 4 km cells that overlapped the two 25 x 25km survey blocks (Figure 2.8). Using the estimated density value for each of these two blocks, we calculated the total number of porpoises that were predicted to be within these 98 cells. These animals were then re-distributed across the 98 cells according to each cells predicted measure of relative abundance from the habitat association modelling (Figure 3.7a). The resulting values provide an indication of the number of porpoises likely to be present in each 4 x 4 km cell (Figure 3.8). 43

130 Figure 3.7. a) Predictions of the number of harbour porpoise within each cell, given 1 km of effort. b) Standard errors around predicted values in each cell. 44

131 Figure 3.8. The predicted number of harbour porpoises in each cell. Values are based upon measures of relative abundance derived from the habitat association modelling (Fig 3.7), scaled according to estimates of absolute abundance from aerial line transect surveys (Table 3.4), and extrapolated to other areas according to predicted relative abundance (Fig 3.7). 45

132 3.3 Assessment of spatial variation in cetacean occurrence using passive acoustic monitoring data. This assessment of broad scale spatial variation in the occurrence of harbour porpoises and dolphins across the Moray Firth was based on data from the arrays of C-PODS deployed during the DECC funded study in 2009 and 2010 (Figure 2.12). The primary period of data collection in both years was between July and October, and data were recovered from 56 of 64 devices (88%) in 2009 and 60 of 68 devices (88%) in There were slight differences in both the spatial pattern and temporal coverage between years because of changes in the study design and patterns of equipment loss or failure (see Figure 2.12). Nevertheless, these passive acoustic monitoring data show a consistent pattern in both years. Both dolphins and porpoises were detected on all PODS at least once during their deployments, but the number of days on which they were detected varied considerably. Currently, it is not possible to use these click characteristics to determine which species of dolphins have been detected on the PODs, and it is likely that detections in different areas represent different species. Dolphins were detected regularly in the inner Moray Firth and along the southern Moray Firth coast, but detections were less frequent in the central part of the Moray Firth. However, dolphin detections increased again at more offshore locations, including those around the windfarm sites (Figure 3.9). In contrast, harbour porpoises were detected more commonly throughout the whole study area, with the lowest detection rates in those coastal areas where dolphins occurred more commonly (Figure 3.10). 46

133 Figure 3.9. Proportion of days that dolphins were detected a) in 2009 and b) in 2010 at each PAM site. Figures are updated versions of those presented in Thompson et al. 2010a and Thompson et al. 2011a. 47

134 Figure Proportion of days that porpoises were detected a) in 2009 and b) in 2010 at each PAM site. Figures are updated versions of those presented in Thompson et al. 2010a and Thompson et al. 2011a. 48

135 a) Porpoises 100 % days detected (2010) % days detected (2009) Mean hrs/day detected (2010) Mean hrs/day detected (2009) b) Dolphins % days detected (2010) % days detected (2009) Mean hrs/day detected (2010) Mean hrs/day detected (2009) Figure Comparison of the percentage of days that a) porpoises and b) dolphins were detected at 33 sites that were monitored in both 2009 and Also presented are the mean number of hours that animals were present on those days on which detections were made. Data are from August and September only. Figures are taken from Thompson et al. 2011a. 49

136 A comparison of inter-annual consistency in spatial variation in occurrence was made using data from 33 sites that were used in both 2009 and Sampling periods differed slightly between years, but data from August and September were available from all sites. Figure 3.11 shows that there was a significant relationship between the percentage of days detected and the average number of hours that animals were detected on each of those days for both dolphins and porpoises. Given this finding, we pooled data from both 2009 and 2010 to provide an overall summary of spatial variation in the occurrence of porpoises and dolphins across the wider Moray Firth (Figure 3.12). At offshore sites, porpoises were present on almost all sampling days (Figure 3.12). To provide finer scale information on variation in the occurrence of porpoises around the windfarm sites, we therefore estimated the median number of hours per day that porpoises were detected at each of the offshore sites in and around the BOWL and MORL development areas (Figure 3.13). 50

137 Figure Spatial variation in the occurrence of a) porpoise and b) dolphins in the summers (April-Oct) of 2009 and Figures show the proportion of days that animals were detected on C-PODs at each sampling location, using pooled data from Thompson et al. 2010a and 2011a. 51

138 Figure Fine-scale spatial variation in the occurrence of porpoises in and around the BOWL and MORL development sites. Data are from the summers (April-Oct ) of 2009 and Pie-charts for each sampling site represent the median number of hours that porpoises were detected each day during the sampling period (April Oct of 2009 and 2010). 3.4 Using classification tress to model spatial variation in the occurrence of different dolphin species Over 1000 sightings dolphins were used in the analyses, although most of these were from surveys conducted in coastal areas (Table 3.4, Figure 3.14). Table 3.4 The number of sightings and counts of animals of each of the four species of dolphin included in the analysis Species Number of Number of sightings animals Bottlenose dolphin Common dolphin Risso s dolphin 4 6 White beaked dolphin

139 Figure Sightings of dolphins from all data sources used in the classification tree. The classification tree that included the full dataset used all six variables available to determine classes and had 23 terminal nodes. The results from this tree suggest that any dolphins encountered along the coastal strip are most likely to be bottlenose dolphins, but those encountered in offshore areas are, in general, more likely to be other species (Figure 3.15a). However, including the series of encounters during the IECS/BOWL surveys meant that the model predicted a higher likelihood that dolphins encountered in this specific offshore area are likely to be bottlenose dolphins. The tree which excluded the IECS/BOWL data had 21 terminal nodes and used depth, slope, distance to coast, sediment type and latitude. Given uncertainties over the reliability of species identification from the IECS surveys, and supporting evidence from acoustic work (see Annex II), we suggest that predictions from this model provide the more robust picture of likely species composition of groups of dolphins encountered in different parts of the Moray Firth (Figure 3.15b). Data on the likely presence of bottlenose dolphins are also presented separately in Figure

140 Figure 3.15a Prediction of the dolphin species composition within each4x4 km grid cell, using all data. 54

141 Figure 3.15b. Prediction of the dolphin species composition within each4x4 km grid cell, using all data except for the IECS/BOWL dataset 55

142 Figure Prediction of the likelihood that dolphins encountered in each 4x4 km grid cell are likely to be bottlenose dolphins. Data are as for Figure 3.15b, but presented as bottlenose dolphin (black portion of pie chart) vs all other species. 56

143 3.5 Estimation of density Harbour porpoise density There were 230 sightings of harbour porpoises, representing 350 individuals, during the aerial line transect surveys shown in Figure Density estimates were made both for the entire survey area, and for sub-areas (Table 3.5). Combining data from all areas, the density was estimated to be 0.64 porpoises per km 2. When analysed separately, these data indicated that densities were highest in the survey block that included the BOWL and MORL development sites, where densities were estimated to reach 0.81 porpoises per km 2. These estimates indicate that the BOWL and MORL development areas contained approximately 100 and 420 individual harbour porpoises respectively during this period (Table 3.6). Figure Locations of sightings of harbour porpoise made during the aerial line transect surveys in August and September

144 Table 3.5. Estimates of porpoise density (individuals per km 2 ) in each of the survey areas. Area All surveyed areas Porpoise density Coefficient of variation 95% confidence range Equivalent number of animals Block A Block B Coast Table 3.6. Estimates of the number of individual porpoises present in the BOWL and MORL sites are based on data from Block B (see Table 3.5). Site Area (km²) Number of porpoises 95% confidence range BOWL Site MORL Site

145 3.5.2 Dolphin density Relatively few dolphins were recorded during the aerial surveys (30 sighting of 90 individuals). The resulting CV s of these estimates were relatively high compared with our porpoise estimate, but similar to those from for estimates density estimates for white-beaked dolphins (CV = 0.96) and bottlenose dolphins (CV = 0.87) in area J (Moray Firth, Orkney & Shetland) during SCANS II. It was only possible to use these density estimates (Table 3.7) to estimate the combined abundance of all dolphin species (Table 3.8). Nevertheless, viewed in conjunction with results from the classification tree (Figure 3.15 and 3.16), these analyses highlight that the numbers of any species of dolphin, and particularly bottlenose dolphin, are likely to be low in the vicinity of the proposed windfarms. This is especially so given that estimates are likely to be positively biased given our use of a g(0) for harbour porpoises; a species that is more difficult to detect than dolphins. Furthermore, estimates of the total numbers of animals within the coastal strip are also likely to be high because the surveys were conducted over parts of the Moray Firth that are known to be used regularly by bottlenose dolphins. SCANS II was unable to estimate abundance of common dolphin and Risso s dolphin in this area, but the density estimates for white-beaked dolphin ( individuals per km 2 ) and bottlenose dolphins (0.011 individuals per km 2 ) for area J are similar to estimates obtained in this study. 59

146 Area All surveyed areas Table 3.7. Estimates of dolphin density (individuals per km 2 ) in each of the survey areas. Dolphin density Coefficient of variation 95% confidence range Equivalent number of animals Block A Block B Coast Table 3.8 Estimates of the number of individuals present in different regions (see 2.9) within the Moray Firth based on the estimated density in the sample blocks within each of those regions (see Table 3.7). Site Area (km²) Number of Dolphins 95% confidence range BOWL Site MORL Site Coastal Strip Central Moray Firth Outer Moray Firth

147 3.6 Temporal patterns of acoustic detections within the MORL and BOWL development sites Comparability of data from T-PODs and C-PODs Data were recovered from nine of the Outer Moray Firth sites at which a both a C- POD and T-POD had been deployed during Data were available from both devices for between 20 and 101 days depending upon the site. Porpoises were detected regularly at all nine sites. Comparison of the number of detection positive hours each day indicated that there was a significant relationship between the detection rates on C-PODs and T-PODs both for all sites combined (Figure 3.18a) and specifically for the Beatrice Demonstrator site where there had been a time series of data using both types of device (Figure 3.18b). a) b) Detection +ve Hours/Day (T-POD) Detection +ve Hours/Day (T-POD) Detection +ve Hours/Day (C-POD) Detection +ve Hours/Day (C-POD) 24 Figure Comparison of the number of hours per day that porpoises were detected on paired T-PODs and C-PODs in 2010; a) for all nine paired devices (see Table 3.7) b) data for the Beatrice Demonstrator site. The line shown on each figure represents a 1:1 relationship. Overall, the average difference in the number of hours each day that porpoises were detected was close to zero (x = , SD = 2.33), suggesting that there was no consistent bias when using one or other device (Figure 3.19a). Detection rates for dolphins were much lower, preventing a more detailed comparison of the number of 61

148 hours that dolphins were detected on each device in each day. However, the average difference in the number of hours each day that dolphins were detected was also close to zero (x = , SD = 0.63; Figure 3.19b). a) b) Frequency Frequency Difference in Detection +ve Hours Per Day Difference in Detection +ve Hours Per Day 3 4 Figure Differences in the number of hours that a) porpoises and b) dolphins were detected on the nine matched pairs of T-PODs and C-PODs (see Table 3.7 for sample sizes). Table 3.9. Comparison of the mean numbers of hours per day that dolphins and porpoises were detected on the T-PODs and C-PODs that were deployed together at each of nine sites in the summer of Site Dolphins Harbour Porpoise X Hrs/day detected (SE) X Hrs/day detected (SE) Median Hrs/day (IQ) N (days) T-POD C-POD T-POD C-POD T-POD C-POD E (0.04) 0.11 (0.04) 7.60 (0.40) 7.80 (0.36) 7 (5-10) 8 (5-10) 81 E (0.09) 0.10 (0.03) 6.13 (0.35) 5.47 (0.27) 6 (4-8) 5 (4-7) 100 A (0.01) 0.15 (0.04) 5.97 (0.25) 6.33 (0.28) 6 (4-8) 6 (4-8) 101 D (0.05) 0.10 (0.03) 6.99 (0.33) 6.05 (0.32) 7 (5-8) 6 (4-8) 83 E (0.14) 0.25 (0.10) 5.50 (1.69) 5.10 (1.66) 1.5 (0-14) 1 (0-13) 20 E (0.04) 0.34 (0.09) 7.27 (0.47) 7.15 (0.49) 6 (5-10) 7 (4-10) 67 A (0.02) 0.13 (0.05) 4.85 (0.47) 8.69 (0.58) 5 (2-7) 8.5 (5-11) 48 A (0.07) 0.34 (0.06) 6.66 (0.35) 6.86 (0.33) 6 (4-9) 7 (5-9) 100 E (0.03) 0.29 (0.05) 3.43 (0.24) 4.87 (0.28) 3 (2-5) 5 (3-7) 97 62

149 Temporal variability in T-POD and C-POD detections at the Beatrice Demonstrator site. The longest time-series of passive acoustic monitoring data was available from the Beatrice Demonstrator site, where devices were deployed between August 2005 and December After a break in studies during 2008, devices were again deployed at this site in May 2009 and data collection is anticipated to continue until at least autumn There have been some gaps in the time-series due either to equipment loss or failure (Table 3.10), but these data provide a unique opportunity to explore longer-term temporal change in the occurrence of dolphins and porpoises at an offshore site. Table The number of T-PODs and C-PODS deployed and successfully recovered at the Beatrice Demonstrator site in each month of , Months blocked in black are those where a single device has been deployed but not yet recovered T C T C T C T C T C T C T C Jan 2 1 Feb 2 1 Mar Apr May 2 1 Jun Jul Aug Sep Oct Nov Dec Overall, porpoises were detected on most (> 93%) days that T-PODs or C-PODs were deployed at this site, whereas dolphins were detected only rarely (< 6% of deployment days) and this pattern was consistent across all five years in which data were collected (Figure 3.20). On those days that porpoises were detected, they were 63

150 recorded for a median of 4 hours (IQ range = 2-7), whereas on those days that dolphins were detected, they were recorded for a median of one hour (IQ range = 1-1) (Fig 3.21). The median number of hours that porpoises were detected on each day was also consistent across years (Figure 3.22). 100 % Days Dectected Figure Annual values for the % of days that porpoises (squares) and dolphins (circles) were detected at the PAM site near the Beatrice Demonstrator. See Fig 2.10 for the site location and Table 3.8 for sample sizes Frequency Frequency Number of hours per day Number of hours per day Figure Frequency histograms for the number of hours that a) porpoises and b) dolphins were detected on those days in which there was at least one detection. (Data are from the entire period ). 64

151 14 No of hours/day detected Figure Annual estimates in the median number of hours per day (with IQ ranges) that porpoises were detected at the PAM site near the Beatrice Demonstrator Seasonal variability in C-POD detections within the BOWL site. Passive acoustic monitoring data are available from two sites within the BOWL development area for a period of almost two years, and from three additional sites for the final nine months of the study (Table 3.11). Inspection of these data indicates that porpoises were present in the area on an almost daily basis, whereas dolphin detections remained much lower throughout the year (Figure 3.23). However, the median number of hours that porpoises are detected does appear to vary seasonally, with peaks in the winter and late summer (Figure 3.24). Not only were porpoises detected almost daily, but they were typically present for many hours each day. In contrast, dolphins were generally detected form only one or two hours a day, even on those few days that they were detected (Figure 3.25). 65

152 Table The number of sites within the BOWL and MORL development areas at which C-POD data were collected in each month of 2009, 2010 and Numbers in brackets represent devices deployed but not yet recovered. BOWL MORL Jan (6) Feb (6) Mar (6) Apr 2 1 (14) May (14) Jun (14) Jul Aug Sep Oct Nov Dec (6) 100 % days detected each month A J S O N D J F M A M J J A S O N D J F M Figure Monthly values for the % of days that porpoises (squares) and dolphins (circles) were detected at site within the BOWL development area. See Table 3.8 for sample sizes. 66

153 Median No of Hours Detected Per Day J A S O N D J F M A M J J A S O N D J F M Figure Monthly variation in the median number of hours per day that porpoises were detected on C-PODs within the BOWL development area. Sample sizes are provided in Table 3.8. a) Porpoises b) Dolphins Frequency Frequency Number of hours detected Number of hours detected Figure Frequency histograms showing the number of hours that a) porpoises and b) dolphins were detected on C-PODs from the BOWL site. Data are from , and only include those days on which any animals were detected. 67

154 Seasonal variability in C-POD detections within the MORL site. Due to equipment loss, there were no complete records from any single site within the MORL development area, although data were collected from at least one site in each month of the study (Table 3.11). However, extensive additional data will be available from the current deployments and further DECC-funded work planned for the latter half of Porpoises again appear to be present in the area on an almost daily basis, whereas dolphin detections remain low throughout the year (Figure 3.26). Seasonal patterns in the median number of hours that porpoises are detected remain less clear at this stage, and further evaluation will be undertaken once additional data are recovered. Nevertheless, it is clear that porpoises are typically present in the area throughout the year, for several hours a day (Figure 3.27 & 3.28). In contrast dolphins were typically only detected for one or two hours on those days that they were recorded on the site (Figure 3.28). 100 % Days detected each month A M J J A S O N D J F M A M J J A S O N D J F M A M J Figure Monthly values for the % of days that porpoises (squares) and dolphins (circles) were detected at site within the MORL development area. See Table 3.9 for sample sizes. Data from 8 sites for Dec 2010 to March 2011 are included, and additional devices will be collecting data until June

155 Median No of Hours Detected Per Day M J J A S O N D J F M A M J J A S O N D J F M A M J Figure Monthly variation in the median number of hours per day that porpoises were detected on C-PODs within the MORL development area. Sample sizes are provided in Table 3.8. Additional data for the period Dec 2010 to June 2011are currently being collected. a) Porpoises b) Dolphins Frequency Frequency Number of hours detected Number of hours detected Figure Frequency histograms showing the number of hours that a) porpoises and b) dolphins were detected on C-PODs from the MORL site. Data are from , and only include those days on which any animals were detected. 69

156 4. Acknowledgements Much of this work builds upon data collected through recent studies conducted for DECC, COWRIE, Oil & Gas UK and Marine Scotland. We would like to thank them and all the funding bodies that have supported the collection of other external data used in this report. Thanks also to the other organisations that provided data for the integrated analyses of historical sightings and to the many individuals who have collected survey data at sea over the years. Particular thanks go to Tim Barton, Barbara Cheney, Nick Richardson, Helen Bates and Bill Ruck for all their work on recent PAM studies, and to Rasmus Nielson, Gareth Bradbury and other staff at WWT Consulting Ltd for their skilled work during the 2010 aerial survey programme. 5. References Akaike, H. (1974) A new look at the statistical model identification. IEEE Transactions on Automatic Control, AC19, Aarts, G., MacKenzie, M., McConnell, B., Fedak, M., & Matthiopoulos, J. (2008) Estimating space-use and habitat preference from wildlife telemetry data. Ecography, 31, Bailey, H., Senior, B., Simmons, D., Rusin, J., Picken, G., & Thompson, P.M. (2010) Assessing underwater noise levels during pile-driving at an offshore windfarm and its potential effects on marine mammals. Marine Pollution Bulletin, 60, Bailey, H. & Thompson, P.M. (2009) Using marine mammal habitat modelling to identify priority conservation zones within a marine protected area. Marine Ecology Progress Series, 378, Camphuysen, C.J., Fox, A.D., Leopold, M.F. & Petersen, I.K. (2004) Towards standardised seabirds at sea census techniques in connection with environmental impact assessment for offshore wind farms in the UK: a comparison of ship and aerial sampling methods for marine birds, and their applicability to offshore wind farm assessments. COWRIE BAM Clark, C.W. & Clapham, P.J. (2004) Acoustic monitoring on a humpback whale (Megaptera novaeangliae) feeding ground shows continual singing into late spring. Proceedings of the Royal Society of London Series B-Biological Sciences, 271, Cordes, L.S., Duck, C.D., Mackey, B.L., Hall, A.J., & Thompson, P.M. (2011) Longterm patterns in harbour seal site-use and the consequences for managing protected areas. Animal Conservation. DOI: /j x 70

157 De'ath, G, & Fabricius, K.E. (2000). Classification and regression trees: a powerful yet simple technique for ecological data analysis. Ecology 81: Grellier, K. and Lacey, C. (2010). Review of existing published marine mammal data for the Moray Firth, SMRU Ltd: 25 pp. Hammond, P.S., Berggren, P., Benke, H., Borchers, D.L., Collet, A., Heide- Jørgensen, M.P., Heimlich, S., Hiby, A.R., Leopold, M.F., & Øien, N. (2002) Abundance of harbour porpoise and other cetaceans in the North Sea and adjacent waters. Journal of Applied Ecology, 39, Hammond, P.S., Macleod, K., Berggren, P., Borchers, D.L., Burt, M.L., Cañadas, A., Desportes, G., Donovan, G.P., Gilles, A., Gillespie, D., Gordon, J., Hiby, L., Kuklik, I., Leaper, R., Lehnert, K., Leopold, M., Lovell, P., Øien, N., Paxton, C.G.M., Ridoux, V., Rogan, E., Samarra, F., Scheidat, M., Sequeira, M., Siebert, U., Skov, H., Swift, R., Tasker, M.L., Teilmann, J., Van Canneyt, O. & Vázquez, J.A. (In prep). Abundance of harbour porpoise and other cetaceans in European Atlantic shelf waters. Hexter, R. (2009a) High Resolution Video Survey of Seabirds and Mammals in the Rhyl Flats Area. Report to COWRIE. ISBN: Hexter, R (2009b) High Resolution Video Survey of Seabirds and Mammals in the Norfolk Area. Report to COWRIE. ISBN: Hiby, A.R. (1999) The objective identification of duplicate sightings in aerial survey for porpoise. In: Garner, G.W., Amstrup, S.C., Laake, J.L., Manly, B.F.J., McDonald, L.L., Robertson, D.G. (eds) Marine mammal survey and assessment methods. Balkema, Rotterdam, p Holland, G.J., Greenstreet, S.P.R., Gibb, I.M., Fraser, H.M. & Robertson, M.R. (2005) Identifying sandeel Ammodytes marinus sediment habitat preferences in the marine environment. Marine Ecology Progress Series, 303, Matthiopoulos, J., McConnell, B., Duck, C., & Fedak, M. (2004) Using satellite telemetry and aerial counts to estimate space use by grey seals around the British Isles. Journal of Applied Ecology, 41, Philpott, E., Englund, A., Ingram, S., & Rogan, E. (2007) Using T-PODs to investigate the echolocation of coastal bottlenose dolphins. Journal of the Marine Biological Association of the United Kingdom, 87, Reid, J B, Evans, P G H, and Northridge, S P (2003) Atlas of Cetacean distribution in north-west European waters. Joint Nature Conservation Committee, Peterborough. Ripley, B. (2010) tree: Classification and regression trees. R package version

158 Sharples, R. J., Matthiopoulos, J. and Hammond, P. S. (2008). Distribution and movements of harbour seals around the coast of Britain: Outer Hebrides, Shetland, Orkney, the Moray Firth, St Andrews Bay, The Wash and the Thames, Sea Mammal Research Unit, University of St Andrews: 65 pp. Thomas, L., Buckland, S.T., Rexstad, E.A., Laake, J.L., Strindberg, S., Hedley, S.L., Bishop, J.R.B., Marques, T. A. & Burnham, K.P. (2009) Distance software: design and analysis of distance sampling surveys for estimating population size. Journal of Applied Ecology, 47, Thompson, P., Brookes, K., Cheney, B., Cândido, A., Bates, H., Richardson, N. & Barton, T. (2010a) Assessing the potential impacts of oil and gas exploration operations on cetaceans in the Moray Firth. First year report to DECC, Scottish Government, COWRIE and Oil & Gas UK. Thompson, P., Brookes, K., Cheney, B., Bates, H., Richardson, N. & Barton, T. (2011a) Assessing the potential impact of oil and gas exploration operations on cetaceans in the Moray Firth. Second year report to DECC, Scottish Government, COWRIE and Oil & Gas UK. Thompson, P.M., Cheney, B., Ingram, S., Stevick, P., Wilson, B. & Hammond, P.S. (Eds) (2011b). Distribution, abundance and population structure of bottlenose dolphins in Scottish waters. Scottish Natural Heritage & Scottish Government Funded Report. SNH Commissioned Report No 354 Thompson, P.M., Lusseau, D., Barton, T., Simmons, D., Rusin, J., & Bailey, H. (2010b) Assessing the responses of coastal cetaceans to the construction of offshore wind turbines. Marine Pollution Bulletin, 60, Thompson, P.M., Miller, D., Cooper, R., & Hammond, P.S. (1994) Changes in the distribution and activity of female harbor seals during the breeding season - implications for their lactation strategy and mating patterns. Journal of Animal Ecology, 63, Thomsen, F., van Elk, N., Brock, V., & Piper, W. (2005) On the performance of automated porpoise-click-detectors in experiments with captive harbour porpoises (Phocoena phocoena) (L). Journal of the Acoustical Society of America, 118, Tougaard, J., Poulsen, L., Amundin, M., Larsen, F., Rye, J., & Teilman, J Detection function of T-PODS and estimation of porpoise densities. In Proceedings of the Workshop on Static Acoustic Monitoring of Cetaceans (eds R. Leeney & N. Treganza), Vol. Special Issue No. 46. European Cetacean Society. Van Parijs, S.M., Clark, C.W., Sousa-Lima, R.S., Parks, S.E., Rankin, S., Risch, D., & Van Opzeeland, I.C. (2009) Management and research applications of real-time and archival passive acoustic sensors over varying temporal and spatial scales. Marine Ecology Progress Series, 395,

159 Verfuss, U.K., Honnef, C.G., Meding, A., Dahne, M., Mundry, R., & Benke, H. (2007) Geographic and seasonal variation of harbour porpoise (Phocoena phocoena) presence in the German Baltic Sea revealed by passive acoustic monitoring. Journal of the Marine Biological Association of the United Kingdom, 87, Wood, S.N. (2008) Fast stable direct fitting and smoothness selection for generalized additive models. Journal of the Royal Statistical Society (B), 70, Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A., Smith, G.M. (2009) Mixed effects models and extensions in ecology with R. Springer, New York. 73

160 74

161 ANNEX I TECHNICAL REPORT ON HARBOUR SEAL TELEMETRY AND HABITAT MODEL Helen Bailey 1 & Paul Thompson 2 1. SMRU Ltd, New Technology Centre, North Haugh, St Andrews, Fife, KY16 9SR 2. University of Aberdeen, Institute of Biological & Environmental Sciences, Lighthouse Field Station, Cromarty, Ross-shire IV11 8YJ. Report to MORL & BOWL 75

162 76

163 1. Background Harbour seals are resident in the Moray Firth throughout the year, breeding and resting on inter-tidal sandbanks in the inner Moray Firth (Thompson et al. 1996), and making regular foraging trips into the central and outer Moray Firth (Thompson et al. 1998). Although there are a few non-breeding haul-out sites along the outer Moray Firth coast (see Grellier & Lacey 2010), most of this population is found at haul-out sites within the inner firths. The closest known harbour seal breeding site to the MORL and BOWL windfarm sites is in the Loch Fleet National Nature Reserve (NNR), and the next nearest is in the Dornoch Firth (Figure 1). In the early 1990 s, the Moray Firth harbour seal population was estimated to contain approximately 1650 individuals (Thompson et al 1997). Although this formed a relatively small proportion of the UK population, it did represent the largest breeding population on the east coast of Scotland. Within the Moray Firth, over half the population was found breeding in the Dornoch Firth (Thompson et al 1997) and, as a result, harbour seals are one of the key features that led to the designation of this area as the Dornoch Firth and Morrich More Special Area of Conservation (SAC). A series of research projects during the late 1980 s and 1990 s resulted in the Moray Firth population becoming one of the most intensively studied harbour seal populations in the world. As a result, there is a wide-range of published studies on different aspects of their ecology, including work on foraging and diving behaviour (Thompson et al. 1998; Tollit et al. 1998), diet (Pierce et al. 1991; Tollit & Thompson 1996), female reproductive biology (Thompson et al. 1994; Gardiner et al. 1996; Thompson & Wheeler 2008), male vocalizations and display behaviour (Van Parijs et al. 1997; 1999), the impacts of disease and parasite burdens (Thompson et al. 1998; 2002) and interactions with salmonid fisheries (Middlemas et al. 2005; Butler et al. 2008). Regular annual surveys in both the June/July pupping season and August moult were also carried out to explore how observed variations in natural environmental conditions human impacts such as shooting influenced population dynamics (Thompson et al. 2007). These annual surveys were conducted by the University of Aberdeen between 1987 and 2004, and have since been integrated into the NERC Sea Mammal Research Unit s broader scale monitoring programme for 77

164 UK harbour seals. Broad-scale surveys across Scotland have revealed that harbour seals have declined significantly in most areas (Lonergan et al. 2008; Scottish Government 2011). The pattern of population change is markedly different to that seen in areas affected by mass mortalities from the 1988 and 2002 phocine distemper virus outbreaks (Harkonen et al. 2006), and the factors driving harbour seal declines in Scottish waters remain unclear (Lonergan et al. 2008). Within the Moray Firth, shooting by fisheries managers has clearly contributed to observed declines (Thompson et al. 2007). Marine Scotland now limit the number of harbour seals that can be shot each year through the Moray Firth Seal Management Plan (Butler et al. 2008). However, despite extensive research on other aspects of their biology, limited understanding of variation in key demographic parameters such as reproductive rates and survival has constrained our ability to model recovery rates or assess the key drivers of population dynamics in this or any other harbour seal population worldwide. This is largely a result of the harbour seal s reproductive behaviour, because mothers and pups move readily in and out of the water (Boness & Bowen 1996) and it is therefore difficult to collect demographic data from these species compared with pinnipeds such as grey seals that stay ashore during the breeding season. A key requirement for the MORL and BOWL EIA s is an assessment of the connectivity between the proposed windfarm sites and protected species in locals SACs. In the case of harbour seals, this requires information both on the origin of those seals that may be encountered on the windfarm sites, and the extent to which far-scale effects such as construction noise may overlap with other areas used by harbour seals from the Dornoch Firth and Morrich More SAC. Over the last 20 years, several different studies have used tracking devices to study the foraging movements of harbour seals from the Dornoch Firth and Loch Fleet (Thompson et al. 1996, 1997, 1998; Sharples et al. 2008; Cordes et al. 2011). Compared with most sites, the foraging areas of Moray Firth harbour seals are therefore well characterised, and it was not considered necessary at this stage to conduct additional tracking studies. Instead, the key requirement has been to use the different data sets within a common statistical framework that provides an integrated picture of the foraging distribution of harbour seals from these two breeding sites. The primary challenge in achieving this 78

165 is that technological developments over the last 20 years mean that different studies have used a variety of techniques (VHF telemetry, satellite telemetry & GPS-GSM technology), each with different levels of accuracy and temporal resolution. In this report, we describe how we use a Bayesian State Space Modelling (SSM) approach to integrate tracking data from multiple tag types and standardize position estimates while accounting for location error. We then use the standardized tracking data set to predict habitat usage and estimate the absolute number of harbour seals using different parts of the Moray Firth by scaling by the population size estimated from haul-out counts. As further background for these assessments, we present the latest information on abundance trends in the Dornoch Firth and Loch Fleet. 2. Methodology 2.1 Analysis of Telemetry data Tracking data were available from 37 individual seals that were captured in either Loch Fleet or the Dornoch Firth (Figure 1) and tagged between 1989 and 2009 (Table 1). Seals were captured under licence using either hand nets or beach seine nets, and then sedated while measurements were taken and tags glued to their hair on the head or neck. Capture and handling techniques are described in Thompson et al. (1992). Table 1. Summary of harbour seal telemetry data in the Moray Firth, Scotland. Tag type Deployment years Number of tags Mean duration Sex ratio (Male:Female) (days) VHF :9 Argos satellite :5 GPS GSM :5 Total/Mean :1 79

166 Figure 1. A map showing the location of harbour seal haul-out sites in the Dornoch Firth and Loch Fleet (taken from Cordes et al. 2011) VHF telemetry Between 1989 and 1991, 21 VHF radio tags were attached to harbour seals as part of a Scottish Office funded project on harbour seal foraging ecology (Table 1). Subsequent tracking of these individuals was designed to collect one position per day for six days per week. Radio-fixes were made from coastal vantage points with a three-element Yagi aerial using the null average method (Springer 1979). The accuracy of fixes was estimated using a test transmitter, and the standard deviation of the error between estimated and true bearings used to produce 95% confidence limits for fixes on radio-tagged seals (Thompson & Miller 1990) Satellite telemetry As part of the SEA programme, eleven Sea Mammal Research Unit (SMRU) satellite relay data loggers (SRDLs) were attached to harbour seals in the Moray Firth (Scotland) between 2004 and 2007 (Table 1). These SRDLs transmit data via the Argos system (McConnell et al. 1999). Service Argos allocates all positions to seven location classes, which describe the quality of those locations. Unfortunately, many marine animal tracking studies typically result in lower accuracy positions, and location errors may be several kilometers (Costa et al. 2010). 80

167 2.1.2 GPS GSM telemetry GPS GSM tags combine a GPS (Global Positioning System) sensor with a mobile phone GSM (Global System for Mobile Communications) modem to relay data ashore (McConnell et al. 2004). In 2009, GPS GSM tags were attached to five harbour seals in the Moray Firth as part of a study carried out for Marine Scotland (Cordes et al. 2011) (Table 1). These tags are able to produce much more frequent locations, providing a mean of 37 GPS positions per day compared to 10 Argos positions per day. They are also higher accuracy than Argos locations (Costa et al. 2010). The mean error of GPS positions within a stationary test was 40 m (Hazel 2009). This is approximately four times greater than the best Argos location quality. Hazel (2009) reported no appreciable directional bias in GPS error, and no significant difference between the latitudinal and longitudinal components of the linear error. Nevertheless, occasional errors may arise, and a 10 km h -1 speed filter was therefore applied to the tracks (Costa et al. 2010) State Space Modelling The state-space modelling approach was based on models developed for use with satellite telemetry data (Jonsen et al. 2007, Bailey et al. 2008). This provides a statistical framework for integrating error in the location estimates with a process model of the movement. For the satellite telemetry data, this model was applied to all of the raw Argos satellite positions to obtain daily position estimates (Jonsen et al. 2007, Bailey et al. 2008). For the GPS GSM data, since the rare extreme values had been removed using the speed filter, the SSM error structure was modified from the t-distributions that had been used for each Argos location class (Jonsen et al. 2005) to a single normal distribution. The accuracy of GPS positions is higher when locations are derived from at least 6 satellites (mean = 32 m, SD = 36.9 m) (Hazel 2009), which was the case for the majority of locations from the GPS GSM tagged seals. This estimate of error was therefore incorporated into the SSM. 81

168 For the VHF telemetry data, the SSM error structure was modified in a similar manner to that for the GPS data. A single normal error distribution was used and the parameters based on the error distribution of the 95% confidence limits for fixes. This resulted in a mean linear error of 1.66 km (SD = 0.93 km). However, the mean number of VHF positions per day was less than one at This led to high uncertainty in the output SSM daily positions and we therefore only retained those daily positions that had a corresponding VHF location to ensure that there were no spurious SSM locations. 2.2 Habitat association modelling The 95% credible limits for each SSM position were used to estimate the uncertainty in all positions (Figures 2 and 3). Characterisation of these uncertainties was important for determining the scale at which movement can be related to underlying habitat variables (Patterson et al. 2010). The uncertainty in the SSM positions derived from the GPS tracks was very small because of the high frequency and accuracy of the positions, and was below the resolution of the available environmental data. A suitable grid size for averaging the environmental data was therefore chosen based on the mean width of the 95% credible limits for the Argos and VHF derived SSM positions. Based upon these criteria, a grid size of 4 x 4 km was applied to the environmental data and associated with the seal positions in the habitat analysis. Grid cells within 2 km of a haulout site were removed to reduce bias towards locations were hauled out on land or resting in the water in inshore haul-out areas. (Thompson et al. 1998). Two methods were used to model seal occurrence and habitat preference. The first method used a presence-absence approach within each of the 4 x 4 km grid cells. Any cell that contained at least one seal SSM position was coded as 1 for seal presence. Based on the average travel speed and foraging trip duration (Thompson et al. 1998), all of the grid cells within the Moray Firth were considered available habitat. Cells containing no locations were therefore coded as 0 for seal absence. A generalised additive model (GAM) was applied with a binomial error distribution and logit link function. The environmental variables considered to be likely 82

169 explanatory variables of seal occurrence were water depth, seabed slope, distance to the nearest haulout, and seabed sediment type. The first three of these were treated as continuous variables and the last as a categorical variable, where the most common sediment type (sand, marine sediment) was used as the reference level. Visual inspection of distributions was used to determine whether transformations of the variables were necessary or supported the removal of any outliers. Variance inflation factors were used to test for collinearity between the explanatory environmental variables. These were all less than 3, indicating there was no significant collinearity (Zuur et al. 2009). The smoother terms for the continuous variables were derived using penalized regression splines with a shrinkage term so that, for large levels of smoothing, a smoother could have 0 degrees of freedom and be effectively removed from the model (Wood 2006). The model was applied using the R software package (R Development Core Team 2008) and contributed package mgcv (Wood 2006). The GAM output was visually checked for spatial correlation by plotting the residuals against the spatial coordinates. There were no obvious clusters of negative or positive residuals, and no clear clusters of large residuals indicating that spatial correlation was not significant (Zuur et al. 2009). The second method used a case/control approach where random control points were generated to represent habitat availability. This gave a measure of habitat preference, which was defined as the ratio of the use of a habitat over its availability (Aarts et al. 2008). Control points were generated using the equation for accessibility calculated by Matthiopoulos et al. (2004) as d -1.98, where d is the distance from the haulout in units of 5 km. Since we were using grid cells of 4 km, this was modified accordingly to (0.8*d) Twice the number of control points as seal locations were selected so that habitat availability would be sufficiently approximated (Aarts et al. 2008). Each seal and control location was associated with environmental data in the nearest 4 x 4 km grid cell, thus taking the uncertainty in the SSM seal positions into account. The same environmental variables were used in this method and the presence-absence GAM. Initially, a generalized additive mixed model (GAMM) was applied with a binomial error distribution and logit link function. A random effect term was included to account for the correlation within individual tracks. However, the model would not converge, even after increasing the number of iterations and raising the number of control points up to five times the number of seal locations. A 83

170 generalized linear mixed model (GLMM) had similar issues and a generalized estimating equation (GEE) model was therefore applied instead. This approach has the advantage that GEEs are less analytically complex and model convergence is more likely. The correlation among pairs of seal locations is also likely to differ from the correlation among available control points (Fieberg et al. 2010). GEEs have the advantage that their parameter estimates and empirical standard errors are robust to misspecification of the correlation structure (Hardin & Hilbe 2003), and also provide a population averaged inference rather than subject specific (Fieberg et al. 2009). A GEE model was therefore applied with five times the number of control points as seal positions to ensure accurate representation of available habitat (Koper & Manseau 2009) and an independence working correlation to avoid biased regression parameter estimators (see Craiu et al. 2008). This GEE model provided an estimate of foraging habitat preference. Since this can vary between seasons and sexes, this was repeated using only data from the summer breeding period (April to July). The model was performed using the contributed R package geepack version (Yan & Fine 2004). 2.3 Harbour Seal Abundance on land and at sea Estimates of the size of the Moray Firth harbour seal population were taken from Thompson et al (1997). This population estimate was based upon breeding season counts at haul-out sites which were then scaled to total population size using independently collected data on the proportion of animals that were likely to be in the water at the time of these counts. Data on trends in abundance at haul-out sites across the Moray Firth were based on recent analysis of the time series of annual surveys conducted in the Dornoch Firth and Loch Fleet (Cordes et al. 2011). To estimate absolute numbers of harbour seals using different parts of the Moray Firth, we combined these data with the output from the presence-absence GAM. Predictions from the presence-absence GAM resulted in a probability of seal occurrence in each of the 4 x 4km cells across the Moray Firth. The total number of seals in the population was then dispersed across this density surface in relation to 84

171 the predicted importance of this cell, thereby providing an estimate of the number of seals likely to be occurring in each cell at any one moment in time. Currently we do not formally incorporate uncertainty into this estimate. Instead, the estimate is conservative in two ways. First, we used the average population estimate of 1653 from 1993 (from Thompson et al. (1997), when the population was at a peak compared with current numbers (see results). Second, we assumed that all seals might be foraging at sea at the same time. However, a sub-set of the population are hauled out on every low tide through the year, and many animals typically remain around haul-out sites for several days between offshore foraging trips. As a result it is likely that the number of seals at sea is typically only 60-90% of the total population depending both upon season and the age and status of individual seals (Thompson et al. 1998). 3. Results 3.1 SSM locations The SSM most probable daily locations derived from the seal telemetry data showed a high degree of overlap between the three tag types (Figure 2), indicating consistency in habitat use between tagging methods and over the 20 year period. The majority of locations occurred near the haulout sites where the seals were tagged in the Dornoch Firth and Loch Fleet. There was also a high number around and to the north of Tarbet Ness (Figure 3), which has previously been identified as foraging habitat (Thompson et al. 1996, Tollit et al. 1998). The greatest dispersion was shown in the Argos satellite positions which extended into the NE part of the Moray Firth. 85

172 Figure 2: a) Daily seal SSM locations derived from Argos satellite (red), GPS GSM (green), and VHF (blue) positions (circles). 86

173 Figure 3: Number of daily SSM harbour seal locations within 4 x 4 km grid cells (colour coding based on the quantile distribution) with the proposed BOWL and MORL wind farm sites overlaid. 3.2 Presence-absence GAM The results of the presence-absence GAM showed that depth, slope and distance to nearest haulout were significantly related to the probability of harbour seal presence, but sediment type was not (Table 2). The probability of seal occurrence was highest at intermediate depths (approximately m) and decreased with increasing seabed slope (Figure 4). The probability of seal presence was highest within 30 km of the nearest haulout and declined rapidly beyond 100 km distance. Predicted 87

174 probabilities of seal presence and densities were in the inner Moray Firth, near the coast and in the northeastern part of the Moray Firth, including the MORL and BOWL sites (Figure 5). Table 2: Results of GAM for probability of seal presence in relation to square root of water depth, square root of seabed slope, distance to nearest haulout and seabed sediment type (reference level: sand, marine sediment). Smoother term: edf Chi-square P value Depth < 0.001* Slope < 0.001* Distance to nearest haulout * Parametric coefficients: Estimate Z value P value Intercept < 0.001* Sediment Sand, marine, gravelly Rock or gravel Sand, marine, muddy or mud, marine, sandy Overall deviance explained 35.2% edf, estimated degrees of freedom * denotes statistical significance at 5% level 88

175 Figure 4: GAM smoothing curves for square root of water depth (m), square root of seabed slope (degrees), and distance to nearest haulout (km) in relation to probability of seal presence. 89

176 a) b) Figure 5: a) Seal presence from SSM daily positions in 4 x 4 km grid cells shown in red, b) GAM predicted probabilities of seal presence (white cells indicate no data), The BOWL (green) and MORL (blue) sites are overlaid. 90

177 3.3 Summer presence/absence GAM Depth and slope were significantly related to the probability of harbour seal presence, but distance to nearest haulout and sediment type were not (Table 3). The probability of seal occurrence was highest at intermediate depths (approximately m) and decreased with increasing seabed slope (Figure 6). The probability of seal presence was highest within 30 km of the nearest haulout and then remained relatively constant beyond this except for a slight drop at distances greater than 100 km. The predicted probabilities of seal presence and densities were lower in the NE part of the Moray Firth during the summer breeding period (Figure 7). Table 3: Results of GAM for summer (April to July) seal presence in relation to square root of water depth, square root of seabed slope, distance to nearest haulout and seabed sediment type (reference level: sand, marine sediment). Overall Smoother term: edf Chi-square P value deviance explained Depth < 0.001* Slope < 0.001* Distance to nearest haulout Parametric coefficients: Estimate Z value P value Intercept < 0.001* Sediment Sand, marine, gravelly Rock or gravel Sand, marine, muddy or mud, marine, sandy % edf, estimated degrees of freedom * denotes statistical significance at 5% level 91

178 Figure 6: GAM smoothing curves for square root of water depth (m), square root of seabed slope (degrees), and distance to nearest haulout (km) in relation to probability of seal presence during the summer breeding period (April to July). 92

179 a) b) Figure 7: a) Seal presence from SSM daily positions during summer (April to July) in 4 x 4 km grid cells shown in red, b) GAM predicted probabilities of seal presence (white cells indicate no data). 93

180 3.4. Case/control GEE model The results of the case/control GEE model indicated that seal foraging habitat preference is significantly related to sediment type, depth, slope and distance to nearest haulout (Table 4). Seals significantly preferred sand, marine, muddy sediment over sand, marine sediment and had lower preference for gravel, sandy, marine and gravel marine sediment than sand, marine sediment. Seals preferred mid-water depths, shallow slopes and farther distances from the haulouts (compared to the distribution of control points). Foraging habitat preference was highest in the northeastern part of the Moray Firth and also in small areas of the southeastern part (Figure 8). Table 4: Results of GEE for seal habitat preference in relation to square root of water depth, square root of seabed slope, logarithm (to the base 10) of distance to nearest haulout and seabed sediment type (reference level: sand, marine sediment). * denotes statistical significance at 5% level Term: Estimate Standard Wald Error Statistic P-value Intercept < 0.001* Depth < 0.001* Depth < 0.001* Slope < 0.001* Distance to nearest haulout < 0.001* Sediment Sand, marine, gravelly Gravel, sandy marine * Gravel, marine * Sand, marine, muddy * Mud, marine, sandy

181 a) b) Figure 8: a) Map of seal SSM daily positions and control points, b) GEE predicted values of seal habitat preference (white cells indicate no data). The BOWL (green) and MORL (blue) sites are overlaid. 95

182 3.5 Summer case/control GEE model The results of the case/control GEE model for the summer breeding period (April to July) indicated that seal foraging habitat preference is significantly related to sediment type, depth, slope and distance to nearest haulout (Table 5). Seals significantly preferred sand, marine sediment over gravel, sandy, marine, gravel marine sediment and mud, marine, sandy sediment. This difference in sediment type preference may reflect a change in prey preferences during the summer period. Seals preferred mid-water depths and shallow slopes. They also preferred farther distances from the haulouts (compared to the distribution of control points), although not as great (Figure 9). Table 5: Results of GEE for seal habitat preference in relation to square root of water depth, square root of seabed slope, logarithm (to the base 10) of distance to nearest haulout and seabed sediment type (reference level: sand, marine sediment). Term: Estimate Standard Wald Error Statistic P-value Intercept < 0.001* Depth * Depth < 0.001* Slope * Distance to nearest haulout < 0.001* Sediment Sand, marine, gravelly Gravel, sandy marine < 0.001* Gravel, marine < 0.001* Sand, marine, muddy Mud, marine, sandy < 0.001* * denotes statistical significance at 5% level 96

183 a) b) Figure 9: a) Map of seal SSM summer (April to July) daily positions and control points, b) GEE predicted values of seal habitat preference (white cells indicate no data). 97

184 3.6 Trends in abundance at haul-out sites Counts made during the breeding season indicate that there has been a steady decline in the number of seals occupying the Dornoch Firth SAC since the mid- 1990s. Over this same period, numbers in Loch Fleet have gradually increased, and the area has also become established as an important breeding site used by over 70 individually recognisable adult females (Thompson & Wheeler; Cordes et al. 2011; Cordes Unpublished Data) Figure 10. Trends in the mean pupping season count of harbour seals (not including pups) at haul-out sites within the Dornoch Firth (closed triangles) and Loch Fleet (open circles). (Taken from Cordes et al to be updated)) 3.7 Abundance of seals at sea Based upon the highest levels of abundance seen over the last two decades, the results of the presence-absence GAM indicate that seals from the Moray Firth population may be dispersed widely across the Moray Firth, particularly over offshore sandbanks (Figure 11). These data suggest that there is variability in the importance 98

185 of different parts of the BOWL and MORL sites, but that some grid squares in this region might be expected to hold up to 8 seals, representing a density approaching 0.5 individuals per km 2. Figure 11. Predicted numbers of harbour seals from Moray Firth haul-out sites in different 4 x 4 km grid cells across the Moray Firth. 4. Acknowledgements We would like to think all the funding bodies who supported the collection of data used in this study, and the colleagues who carried out the original capture, tagging and tracking of the seals. A particular thanks to Phillip Hammond, Ruth Sharples and Line Cordes for access to data and for useful discussion about the work. 99

186 5. References Aarts G, MacKenzie M, McConnell B, Fedak M, Matthiopoulos J (2008) Estimating space-use and habitat preference from wildlife telemetry data. Ecography 31: Bailey H, Shillinger G, Palacios D, Bograd S, Spotila J, Paladino F, Block B (2008) Identifying and comparing phases of movement by leatherback turtles using statespace models. Journal of Experimental Marine Biology and Ecology 356: Butler, J.R.A., Middlemas, S.J., McKelvey, S.A., McMyn, I., Leyshon, B., Walker, I., Thompson, P.M., Boyd, I.L., Duck, C., Armstrong, J.D., Graham, I.M., Baxter, J.M., (2008). The Moray Firth Seal Management Plan: an adaptive framework for balancing the conservation of seals, salmon, fisheries and wildlife tourism in the UK. Aquatic Conservation: Marine and Freshwater Ecosystems, 18; Cordes, L.S., Duck, C.D., Mackey, B.L., Hall, A.J., & Thompson, P.M. (2011) Longterm patterns in harbour seal site-use and the consequences for managing protected areas. Animal Conservation. DOI: /j x Costa DP, Robinson PW, Arnould JPY, Harrison AL, Simmons SE, Hassrick JL, Hoskins AJ, Kirkman SP, Oosthuizen H, Villegas-Amtmann S, Crocker DE (2010) Accuracy of ARGOS locations of pinnipeds at-sea estimated using Fastloc GPS. PLoS ONE 5:e8677 Craiu RV, Duchesne T, Fortin D (2008) Inference methods for the conditional logistic regression model with longitudinal data. Biometrical Journal 50: Fieberg J, Rieger RH, Zicus MC, Schildcrout JS (2009) Regression modelling of correlated data in ecology: subject-specific and population averaged response patterns. Journal of Applied Ecology 46: Fieberg J, Matthiopoulos J, Hebblewhite M, Boyce MS, Frair JL (2010) Correlation and studies of habitat selection: problem, red herring or opportunity? Philosophical Transactions of the Royal Society B: Biological Sciences 365: Gardiner, K.J., Boyd, I.L., Racey, P.A., Reijnders, P.J.H & Thompson, P.M. (1996). Plasma progesterone concentrations measured using an enzyme-linked immunosorbent assay useful for diagnosing pregnancy in harbor seals (Phoca vitulina). Marine Mammal Science, 12: Hardin JW, Hilbe JM (2003) Generalized estimating equations, Vol. Chapman and Hall, Florida Härkönen, T., Dietz, R., Reijnders, P., Teilmann, J., Harding, K., Hall, A., Brasseur, S., Siebert, U., Goodman, S., Jepson, P., Dau Rasmussen, T. & Thompson, P.M. (2006) A review of the 1988 and 2002 Phocine Distemper Virus seal epidemics in European Harbour Seals. Diseases in Aquatic Organisms, 68;

187 Hazel J (2009) Evaluation of fast-acquisition GPS in stationary tests and fine-scale tracking of green turtles. Journal of Experimental Marine Biology and Ecology 374:58-68 Jonsen ID, Fleming JM, Myers RA (2005) Robust state-space modeling of animal movement data. Ecology 86: Jonsen ID, Myers RA, James MC (2007) Identifying leatherback turtle foraging behaviour from satellite-telemetry using a switching state-space model. Marine Ecology Progress Series 337: Koper N, Manseau M (2009) Generalized estimating equations and generalized linear mixed-effects models for modelling resource selection. Journal of Applied Ecology 46: Lonergan, M., Duck, C.D., Thompson, D., Mackey, B.L., Cunningham, L. & Boyd, I.L. (2007). Using sparse survey data to investigate the declining abundance of British harbour seals. J. Zool. (Lond) 271, Matthiopoulos J, McConnell B, Duck C, Fedak M (2004) Using satellite telemetry and aerial counts to estimate space use by grey seals around the British Isles. Journal of Applied Ecology 41: McConnell B, Beaton R, Bryant E, Hunter C, Lovell P, Hall A (2004) Phoning home - A new GSM mobile phone telemetry system to collect mark-recapture data. Marine Mammal Science 20: McConnell BJ, Fedak MA, Lovell P, Hammond PS (1999) Movements and foraging areas of grey seals in the North Sea. Journal of Applied Ecology 36: Middlemas, S., Barton, T., Armstrong, J. & Thompson, P.M. (2005) Functional and aggregative responses of harbour seal predators to changes in salmonid abundance. Proceedings of the Royal Society, B. 273, Patterson TA, McConnell BJ, Fedak MA, Bravington MV, Hindell MA (2010) Using GPS data to evaluate the accuracy of state-space methods for correction of Argos satellite telemetry error. Ecology 91: Sharples, R. J., Matthiopoulos, J. and Hammond, P. S. (2008). Distribution and movements of harbour seals around the coast of Britain: Outer Hebrides, Shetland, Orkney, the Moray Firth, St Andrews Bay, The Wash and the Thames, Sea Mammal Research Unit, University of St Andrews: 65 pp. Springer JT (1979) Some sources of bias and sampling error in radio triangulation. Journal of Wildlife Management 43: Thompson, P.M., Corpe, H.M. & Reid, R.J. (1998). Prevalence and intensity of the parasite Echinophthirius horridus on harbour seals (Phoca vitulina); effects of host age and inter-annual variability in host food availability. Parasitology, 117:

188 Thompson PM, Mackay A, Tollit DJ, Enderby S, Hammond PS (1998) The influence of body size and sex on the characteristics of harbour seal foraging trips. Canadian Journal of Zoology 76: Thompson, P.M., Mackey, B.L., Barton, T.R., Duck, C. & Butler, J.R.A. (2007) Assessing the potential impact of salmon fisheries management on the conservation status of harbour seals in NE Scotland. Animal Conservation, 10; Thompson PM, McConnell BJ, Tollit DJ, Mackay A, Hunter C, Racey PA (1996) Comparative distribution, movements and diet of harbour and grey seals from the Moray Firth, N.E. Scotland. Journal of Applied Ecology 33: Thompson PM, Miller D (1990) Summer foraging activity and movements of radiotagged common seals (Phoca vitulina L.) in the Moray Firth, Scotland. Journal of Applied Ecology 27: Thompson, P.M., Miller, D., Cooper, R. & Hammond, P.S. (1994). Changes in the distribution and activity of female harbour seals during the breeding season; implications for their lactation strategy and mating patterns. Journal of Animal Ecology, 63: Thompson, P.M., Thompson, H. & Hall, A..J. (2002). Prevalence of morbillivirus antibodies in Scottish harbour seals. Veterinary Record, 151, Thompson PM, Tollit DJ, Wood D, Corpe HM, Hammond PS, Mackay A (1997) Estimating harbour seal abundance and status in an estuarine habitat in north-east Scotland. Journal of Applied Ecology 34:43-52 Thompson, P.M. & Wheeler, H. (2008). Photo-ID based estimates of reproductive patterns in female harbor seals. Marine Mammal Science, 24; Tollit, D.J. & Thompson, P.M. (1996). Seasonal and between-year variations in the diet of harbour seals in the Moray Firth, NE Scotland. Canadian Journal of Zoology, 74: Tollit DJ, Black AD, Thompson PM, Mackay A, Corpe HM, Wilson B, Van Parijs SM, Grellier K, Parlane S (1998) Variations in harbour seal Phoca vitulina diet and divedepths in relation to foraging habitat. Journal of Zoology 244: Van Parijs, S.M., Hastie, G.D. & Thompson, P.M. (1999). Geographic variation in temporal and spatial vocalisation patterns of male harbour seals, Phoca vitulina, during the mating season. Animal Behaviour, 58: Van Parijs, S.M., Thompson, P.M., Tollit, D.J. & Mackay, A. (1997). Distribution and activity of male harbour seals during the mating season. Animal Behaviour, 54: Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009) Mixed Effects Models and Extensions in Ecology with R, Vol. Springer, New York 102

189 Grey seal usage maps for MORL/BOWL developments Phase 2 delivery Project Name: BOWL EIA Work Marine Mammals Reference: RPS Project Manager: Dr Kate Grellier Drafted by: Esther Jones and Jason Matthiopoulos QA by: Dr Kate Grellier Date: Wednesday, 07 December 2011 VAT reg. No. GB SMRU LIMITED is a limited company registered in Scotland, Registered Number: Registered Office: 5 Atholl Crescent, Edinburgh EH3 8EJ

190 TABLE OF CONTENTS Table of contents... 2 Summary... 3 Methods... 4 Usage maps References Appendix Data waterfall

191 GREY SEAL USAGE MAPS FOR MORL/BOWL DEVELOPMENTS PHASE 2 DELIVERY AUTHORS: ESTHER JONES 1 & JASON MATTHIOPOULOS 1,2 1. Sea Mammal Research Unit, Scottish Oceans Institute, University of St Andrews, St Andrews, KY16 8LB 2. Centre for Ecological & Environmental Modelling, The Observatory, Buchanan Gardens, University of St Andrews, St Andrews, KY16 9LZ SUMMARY Grey seal (Halichoerus grypus) telemetry data from were combined with aerial survey data from to produce maps of estimated total, at sea, and hauled out usage in a study area surrounding the proposed Moray Offshore Renewables Ltd (MORL) and Beatrice Offshore Wind Farm Ltd (BOWL) wind farm developments. Temporal and demographic (age class and sex) information on seals were aggregated to produce the spatial maps. The usage maps in this (Phase 2) report utilise developments in the software used to produce the maps. As such, there have been several key changes: 1. The Phase 1 report showed tracks that had zero and non zero aerial survey counts associated with them. To improve transparency between the number of tracks shown in Table 1 and Figures 1 and 2, and the usage maps (Figure 6, 7 and 8), only tracks that were matched to non zero aerial survey counts were used in the final analysis. This explains why 7 tracks shown in Table 1 and Figures 1 and 2 of the Phase 1 report do not feature in the corresponding Table 1 and Figures 1 and 2 of this (Phase 2) report. 2. Additional data have become available since the Phase 1 report was submitted. This has resulted in 11 additional tracks being incorporated into the final analysis (Table 1; Figures 1 and 2). 3

192 3. The second driver in the differences between total and at sea usage maps (Figures 6 and 7) is the development of a new null usage model (see Null (Accessibility) Model). This is a bespoke model using only telemetry tracks from the analysis, and therefore represents a more accurate picture of null usage at the study site. In the Phase 1 report, a generalised null usage map was used. 4. Because the total and at sea usage appear graphically similar, a hauled out usage map was included to give a clearer understanding of where the differences in total and at sea usage occurs. METHODS AVAILABLE DATA AERIAL SURVEY Aerial surveys are conducted each year by the Sea Mammal Research Unit (SMRU) and are funded by Scottish National Heritage (SNH) and the National Environmental Research Council (NERC). They take place throughout August and both grey and harbour seals (Phoca vitulina) are counted. At that time, harbour seals are moulting and are in aggregated groups. Grey seals are in dispersed haul outs along the coast. Over a number of consecutive years the entire Scottish coastline is surveyed and counts are marked using OS Landranger maps (1:50,000) to an accuracy of 50m. Data from surveys were used in the analysis. Fixed wing aerial surveys for grey and harbour seals are also carried out during the August moult each year in the Moray Firth, Tay Estuary and The Wash in East Anglia. These surveys are funded by NERC and Natural England. Counts from were used in the analysis. TELEMETRY Telemetry data from individual grey seals have been collected by SMRU since These comprise two sources: Satellite Relay Data Logger (SRDL) tags developed by SMRU use the Argos satellite system and were deployed between 1988 and GPS phone tags that use the GSM mobile phone network with a hybrid Fastloc protocol (McConnell et al., 2004) have been deployed since

193 Telemetry data were selected from the SMRU database by species and processed through a set of data cleansing protocols to remove null and missing values, duplicated records and ineligible data. Tracks were then selected based on the criteria that if any part of a track passed within a 100km buffer zone of the proposed MORL/BOWL development sites, regardless of where tagging had taken place, that track was included. Forty four tracks were used in the final analysis (Table 1), from seals that were tagged between 1992 and 2008 (mostly using Argos tags). Thirty seven of the tagged animals were adults, four were juveniles and three were moulted pups. The male to female ratio was 26:18. Year Tag type Number of tags Sex ratio (m:f) Mean tag lifespan (days) Mean number of location fixes (per day) 1992 Argos 4 2: Argos 2 1: Argos 9 5: Argos 2 1: Argos 16 12: Argos 1 0: Argos 2 2: Argos 1 1: Argos 1 0: Argos/GPS 6 (2 Arg, 4 GPS) 2: TOTAL 44 MEAN Table 1. Summary of telemetry tracks used in the final analysis. Figure 1 shows the geographical locations of tracks used in the analysis, split by tag type. GPS tags have a smaller spatial extent, concentrated to the south of the Moray Firth. Figure 2 shows the tracks split by year from

194 Figure 1. Map showing telemetry track locations by tag type. Figure 2. Map showing telemetry track locations by year. 6