Natural Heritage Zones Bird Population Estimates SWBSG Commissioned Report Number: 1504

Transcription

1 Natural Heritage Zones Bird Population Estimates SWBSG Commissioned Report Number: 1504 i

2 Natural Heritage Zones Bird Population Estimates Commissioned Report No.: SWBSG_1504 Project no: 1403 Contractor: British Trust for Ornithology Scotland Year of publication: 2015 For further information on this report please contact SWBSG Data and Research Coordinator C/O RSPB Scotland, Ground Floor, 2 Lochside View, Edinburgh Park, EH12 9DH This report should be cited as: Wilson, M. W., Austin, G. E., Gillings S. and Wernham, C. V. (2015). Natural Heritage Zone Bird Population Estimates. SWBSG Commissioned report number SWBSG_1504. pp72. Available from: This report, or any part of it, should not be reproduced without the permission of the Scottish Windfarm Bird Steering Group. This permission will not be withheld unreasonably. The views expressed by the author(s) of this report should not be taken as the views and policies of the Scottish Windfarm Bird Steering Group. SWBSG 2015 i

3 Contents ACRONYMS... iii SUMMARY INTRODUCTION DATA SOURCES AND METHODS General methods General approach to deriving population estimates Use of data from Bird Atlas Deriving confidence limits General limitations of estimates Caveats for breeding species Caveats for wintering species Methods for particular groups of species Wintering wildfowl Skuas and gulls Other species SPECIES POPULATION ESTIMATES Whooper swan (Cygnus cygnus) - wintering Pink-footed goose (Anser brachyrhynchus) - wintering Greenland white-fronted goose (Anser albifrons flavirostris) - wintering Barnacle goose (Branta leucopsis) - wintering Black grouse (Tetrao tetrix) - breeding Red-throated diver (Gavia stellata) - breeding Black-throated diver (Gavia arctica) - breeding Red kite (Milvus milvus) - breeding White-tailed eagle (Haliaeetus albicilla) - breeding Hen harrier (Circus cyaneus) - breeding Goshawk (Accipiter gentilis) - breeding Golden eagle (Aquila chrysaetos) - breeding Osprey (Pandion haliaetus) - breeding Kestrel (Falco tinnunculus) - breeding Merlin (Falco columbarius) - breeding Peregrine (Falco peregrinus) - breeding Golden plover (Pluvialis apricaria) - breeding ii

4 3.18 Dunlin (Calidris alpina) - breeding Snipe (Gallinago gallinago) - breeding Whimbrel (Numenius phaeopus) - breeding Curlew (Numenius arquata) - breeding Greenshank (Tringa nebularia) - breeding Short-eared owl (Asio flammeus) - breeding Arctic skua (Stercorarius parasiticus) - breeding Great skua (Stercorarius skua) - breeding Lesser black-backed gull (Larus fuscus) - breeding Herring gull (Larus argentatus) - breeding Great black-backed gull (Larus marinus) - breeding References Acknowledgements Appendices Appendix A: Monthly WeBS counts by species and NHZ Appendix B: Peak NHZ wildfowl counts from all surveys ACRONYMS Acronym Explanation AOT Apparently occupied territory APEP Avian Population Estimates Panel BBS Breeding Bird Survey C.I. Confidence Interval GB Great Britain GSMP Goose and Swan Monitoring Programme GWfS Greenland white-fronted goose Study IGBGC International Greenland Barnacle Goose Census IGC Icelandic-breeding Goose Census ISC International Swan Census NHZ Natural Heritage Zone S2K Seabird 2000 SNH Scottish Natural Heritage SRMS Scottish Raptor Monitoring Scheme SRSG Scottish Raptor Study Group SWBSG Scottish Windfarm Bird Steering Group WeBS Wetland Birds Survey iii

5 SUMMARY This report provides Natural Heritage Zone (NHZ) population estimates for 24 species of breeding birds and a further four species of wintering wildfowl, considered by the Scottish Windfarm Bird Steering Group to occur frequently in environmental impact assessments. NHZs are usually considered appropriate biogeographical spatial units against which regional impacts of development proposals can be assessed. For each species, NHZ-scale population estimates have been derived using the most appropriate methods and data sources available currently. NHZ-specific population estimates generated by previous work on a species were used if the quality of the estimates was equivalent to, or higher than, that of any estimates that could be derived via alternative means. The quality of estimates was assessed according to: age (and likelihood of major changes in populations since the time of the estimate); precision; spatial explicitness (feasibility of apportioning estimates to specific NHZs); and appropriateness of methods (any likelihood of bias due to lack of stratification, or poor coverage of some areas or habitats). Where spatially explicit survey datasets were available and more up-to-date than alternative information, NHZ-specific estimates were derived by assigning birds or breeding units to the relevant NHZs. In the absence of spatially explicit survey datasets, NHZ estimates were derived as rigorously as possible from estimates from larger spatial units (larger regions, whole of Scotland or whole of UK) by apportioning the population across NHZs using relative abundance data from the most recent Britain and Ireland Bird Atlas project. Where available, more recent trend information was used to update population estimates for periods subsequent to the point at which the most up-to-date population estimate for a species was derived. Winter population estimates include individual birds of all ages. Birds are generally more mobile during the winter than when breeding, and consequently the number of birds in an NHZ can vary markedly outside the breeding season, over a range of temporal scales. Birds can move between feeding and roosting sites on a daily basis, while variations between days may occur as food resources are depleted, and variation over the whole season may result from through-flow of migrating birds. For the four wintering wildfowl species, the largest count or estimate available for each NHZ is provided as an indication of the maximum number of individuals that might be exposed to impact within an NHZ. Two important caveats are attached to the use of peak population estimates: - the true peak in numbers within an NHZ might not be reflected in the available survey data, because it occurs at a time when no survey is taking place or when survey effort is not at its highest (no current surveys are specifically designed to provide estimates of regional population peaks); and 1

6 - quantitative information on the extent of turnover of individuals within regions through the winter is lacking. This means that apparently stable regional populations could be much more mobile than survey data suggest, such that regional impacts potentially affect many more individuals than would be apparent under an assumption of no turnover. Breeding bird population estimates most often comprise breeding units (such as breeding or territorial pairs) rather than individual birds. They do not include nonbreeding birds or floaters (birds that do not hold territories). These are an important component of many populations, particularly in long-lived species where individuals take several years to reach maturity. Sedentary species typically remain present (and thus susceptible to impact) all year round. Breeding population estimates do not take account of movements of individuals away from their breeding grounds outside of the breeding season. In Scotland, where winter conditions can be relatively severe, such movements may be common. Each species account contains estimates of population size in each NHZ, a brief description of the methods and datasets used to derive the estimates, any specific caveats associated with the estimates, and a map showing the proportion of the overall Scottish population within each NHZ. The scale at which it is appropriate to interpret the information in both maps and tables is strictly that of the NHZ. Values in tables and shading on maps do not illustrate variation within NHZs - areas of an NHZ that are occupied, or where a species is extremely abundant, are not distinguished from other parts of the same NHZ where the species is rare or absent. Population estimates in the species accounts should be interpreted in the context of the information on derivation and assumptions provided in Data Sources and Methods (Section 3). 2

7 1. INTRODUCTION Development proposals are often assessed in relation to designated nature conservation sites. However, where work is proposed which is not specific to these sites (a wider countryside proposal), Natural Heritage Zones (NHZs, Figure 1) have been deemed appropriate biogeographical spatial units against which regional impacts of development proposals on many bird species can be assessed. Further information on this process may be obtained from the SNH (2006). This report provides NHZ population estimates for 24 species of breeding birds and a further 4 species of wintering wildfowl, all of which are frequently considered in environmental impact assessments. For each species, NHZ-scale population estimates have been derived using the most appropriate methods and data sources available to us. This report describes both the methods by which the estimates are derived and the sources of data on which the estimates are based. Where existing data are considered inadequate to generate estimates at the appropriate scale, where this information is dated, insufficient to support reliable estimates, or derived from poorly designed surveys this has been noted in the text, and recommendations are made for further data collection. 1. Shetland 2. Orkney and North Caithness 3. Coll, Tiree and the Western Isles 4. North West Seaboard 5. Peatlands of Caithness and Sutherland 6. Western Seaboard 7. Northern Highlands 8. Western Highlands 9. North East Coastal Plain 10. Central Highlands 11. Cairngorm Massif 12. North East Glens 13. East Lochaber 14. Argyll West and Islands 15. Loch Lomond, Trossachs & Breadalbane 16. Eastern Lowlands 17. West Central Belt 18. Wigtown Machars & Outer Solway 19. Western Southern Uplands & Inner Solway 20. Border Hills 21. Moray Firth Figure 1. Location and names of Scottish Natural Heritage Zones (NHZs). 3

8 2. DATA SOURCES AND METHODS 2.1 General methods General approach to deriving population estimates The following section outlines the approach taken to selecting data sources and deriving NHZ population estimates. More detailed information on selection and analysis of the data available for each species can be found in the text accompanying the NHZ estimates for that species. Where NHZspecific population estimates had been generated by previous work on a species, these were used, provided that the quality of the estimates was deemed equivalent to, or higher than, that of any estimates that could have been derived via alternative means. Quality of estimates was assessed according to: age (taking into account, wherever available, the opinion of experts as to the likelihood of major changes in populations since the time of the estimate in question); precision (with wide confidence limits indicating a lack of confidence in the central estimate); spatial explicitness (this being directly related to the feasibility of apportioning estimates to different NHZs); and appropriateness of methods (taking into account likely bias due to lack of stratification, or poor coverage of some areas or habitats). Where NHZ-specific estimates were not available from previous work, but could be derived directly from spatially explicit survey datasets, this was achieved by assigning every bird or breeding unit recorded in a survey to the relevant NHZ. For many species, individual population estimates were not available (and could not be derived directly from spatial datasets) for all NHZs. In such cases, NHZ estimates were derived from the best estimates available (in terms of their quality, as described above), whether for regions of Scotland, the whole of Scotland, or the whole of Britain. Two useful sources of population estimates for all species, applying to Scotland and Britain respectively, are the Birds of Scotland publication, edited by Forrester et al. (2007), and the formal estimates from the Avian Population Estimates Panel (APEP) by Musgrove et al. (2013). These were often used to help identify the most suitable population estimates for species lacking an obvious source of NHZ-specific information. Where it was possible to estimate population trends for periods subsequent to the point at which the most up-to-date population estimate for a species was derived, these trends were used to update population estimates. For relatively abundant species, high quality trend data for Scottish populations was available from the Breeding Bird Survey (BBS; e.g. Harris et al. 2014). For some other species, alternative sources of trend information, including changes in relative abundance derived from bird atlas data (Balmer et al. 2013; Gibbons et al. 1993), as well as multiple species or taxon-specific surveys, were used to estimate national or regional trends (e.g. Sim et al. 2005) Use of data from Bird Atlas Where appropriate, and where NHZ-specific estimates were not available (see above), national and regional base population estimates were apportioned between NHZs according to relative abundance metrics derived from Bird Atlas (henceforth referred to as the Atlas; Balmer et al. 2013). Relative abundance estimates, along with confidence limits describing the uncertainty around these, were calculated by randomisation of the observed timed tetrad visit counts, as follows: 4

9 1. Each tetrad in Great Britain (GB) was assigned to an individual NHZ, or to the non-scotland portion of GB, according to which of these was overlapped by the largest proportion of that tetrad. 2. For each NHZ, abundance estimates were generated by randomly sampling tetrad counts from that NHZ with replacement, for each of the 10km squares overlapping the NHZ. Sampling with replacement means that each of the tetrads being sampled can be selected multiple times. Tetrads from each 10km square were sampled until the square was represented by a full complement of tetrad counts (up to 25, for 10km squares not partially overlapping sea, England or another NHZ). 3. These counts were summed across all 10km squares overlapping an NHZ, to derive an abundance figure for the NHZ. This process was repeated for all NHZs, plus the rest of Great Britain, and abundance figures were then converted to proportions of the total abundance across all zones. 4. This exercise was repeated 100 times to produce 100 estimates of the proportion of the GB abundance found in a particular NHZ. Finally, the 100 estimates for each NHZ were ordered, and the following values calculated: 5. Average of 50 th and 51 st values: estimated median proportion of GB abundance found in the NHZ. 6. Average of the 2 nd and 3 rd values: the lower 95% confidence limit for the estimate. 7. Average of the 97 th and 98 th values: the upper 95% confidence limit for the estimate. Population estimates for NHZs were then derived from the base estimates using the abundance proportions. In the most straightforward case (see species account for snipe for an example), an overall population estimate for the whole of Scotland was apportioned between NHZs according to the proportional abundance values associated with each NHZ. The approach was slightly more involved where sub-national, regional population estimates were available for a species but the regions for the base estimates did not match NHZ boundaries. In this latter case, the following methods were used: 1. For each NHZ overlapping the area associated with a base estimate (henceforth, the base area), the Atlas-derived relative abundance value was divided by the area of the NHZ to yield a relative population density estimate for that NHZ. 2. Each density estimate was multiplied by the area of overlap between NHZ and the base area. 3. This product was summed across all NHZs overlapping with the base area. 4. The product for each overlapping NHZ (output of step 2) was divided by the sum of these products (output of step 3), to give the proportional contribution of each NHZ to the base estimate. 5. For each NHZ, the output of step 4 was multiplied by the base estimate, to yield a population estimate for the area overlapping each NHZ. For NHZs overlapping multiple base areas, the relevant contribution of each base area was summed to derive the population size for the whole NHZ. For a few species for which neither recent estimates nor BBS trends to update these were available, changes in tetrad occupancy between the last two Bird Atlases were used to estimate changes in abundance. The relationship between occupancy and abundance is exponential, such that, except at extremely low or high rates of tetrad occupancy, an increase in occupancy of 0.2 is approximately equivalent to a doubling in number of individuals. 5

10 Deriving confidence limits When dividing up real population estimates between NHZs using Atlas-derived estimates of relative abundance (as described in 2.1.2), overall confidence limits for the newly-derived NHZ population estimates had to be generated. Typically, the base population estimates (2.1.1) and Atlas-derived relative abundances (2.1.2), both had their own confidence intervals (or, in the case of many population estimates, maximum and minimum values). These were converted to variance values, first by converting the distance between central estimate and confidence limit to an approximate standard deviation (by dividing it by 1.96) and then converting this to an approximation of variance by squaring it. Combined variance values of multiplied estimates were derived according to the delta method (Powell 2007). This method can also be used to combine variance measures of summed estimates, provided these are independent of one another, or their covariance is known. Once combined variance values are calculated, they can be converted back to confidence limits (by taking their square root, multiplying this by two, and adding to or subtracting from the central estimate, as appropriate. Where variance of summed estimates was likely to be non-independent, overall confidence intervals were derived by summing confidence intervals of each estimate. For combined estimates with asymmetrical confidence intervals, combination of variance was carried out separately for upper and lower confidence intervals. Values for estimates and confidence limits are rounded to the nearest whole number. 2.2 General limitations of estimates Of the 28 species dealt with here, population estimates are for breeding populations of 24 species, and wintering (non-breeding) populations of the remaining four species. Each approach has particular limitations that users of these estimates should consider carefully Caveats for breeding species Breeding bird population estimates are typically of breeding units rather than individual birds. Most often, the relevant unit is the breeding (or, at least, territorial) pair, though for lekking species like black grouse (Tetrao tetrix), breeding populations are typically assessed in numbers of displaying males. In all cases, it is important to remember that the number of breeding units is lower than the number of breeding individuals. Strictly monogamous populations contain twice as many breeding individuals as breeding pairs, but the relationship between breeding units and number of breeders is less straightforward for polygamous birds (like black grouse and some regional populations of hen harrier Circus cyaneus). Whatever the relationship between number of breeding individuals and the size of the breeding population, a greater difficulty in using breeding population estimates to assess the importance of NHZs for a species is that the estimates do not include non-breeding birds. Referred to as floaters in many cases (because they tend not to hold territories), these are an important component of many populations, particularly in long-lived species where individuals take several years to reach maturity. In strictly sedentary species, the individuals comprising the breeding population in an NHZ will typically remain present (and thus susceptible to impact there) all year round. However, breeding population estimates do not take account of movements of individuals away from their breeding grounds outside of the breeding season. In areas like Scotland, where winter conditions can be relatively severe, many bird populations redistribute themselves substantially between summer and 6

11 winter, with birds moving to areas and habitats where their winter survival prospects can be maximised. Such movements need to be considered when assessing the numbers of birds that developments could impact on within an NHZ outside of the breeding season Caveats for wintering species Winter population estimates apply to numbers of individual birds of all ages, and so are more demographically representative of overall numbers than breeding estimates. However, nonbreeding individuals are generally more mobile than breeding birds. As a result, the number of birds supported by a site (or within an NHZ) can vary markedly over the course of the non-breeding season. This has clear implications for the feasibility of evaluating impacts of wind energy development based on single, winter population estimates. Numbers of wintering wildfowl in an area may vary considerably at different temporal scales. Variation within a day occurs as birds move between feeding and roosting sites, variation between days may reflect changes in the value of feeding sites as food resources are depleted, and variation over the whole season may result from migrating birds stopping at sites on their way to or from their mid-winter quarters. The graphs in Appendix A, which are based on Wetland Bird Survey (WeBS) counts, provide an indication of intraseasonal changes in NHZ abundances for each of the four wildfowl species covered by this report. For each of the four wintering wildfowl species, the largest count or estimate available for each NHZ has been identified (see section 2.3.1), as this provides some indication of the maximum number of each species that might be exposed to impact within an NHZ at any one time. However, there are two important caveats to the use of peak population estimates in this way: 1. The true peak in numbers supported in an NHZ might not be captured by any survey. This will be the case if peak abundance occurs at a time when no survey is taking place or when survey effort is not at its highest, or if the surveys that coincide with the peak do not provide complete coverage of the population s distribution within the NHZ. 2. There is a lack of quantitative information on the extent of turnover of individuals within sites or regions through the winter. In at least some cases, regional populations that appear stable and static from the survey data available could be much more mobile, with many individuals moving between regions either within or between winter periods. Where this occurs, regional impacts on a population could potentially affect many more individuals than would be apparent under an assumption of no turnover. 2.3 Methods for particular groups of species Wintering wildfowl As mentioned in section 2.2.2, the mobility of individual birds over the winter means that any one survey or series of surveys cannot be guaranteed to record the peak number of a species wintering within a region, particularly as none of the available surveys has been designed specifically with the generation of regional estimates as a core aim. The approach taken here has been to consider the information available at appropriate regional scale from each survey, in order to illustrate the variability of available estimates, as well as maximising the likelihood that one or more of them will give a reasonable estimate of peak numbers. For this reason, counts are reported from three to four surveys for each of the species considered. Where the highest counts from two or more of these surveys are similar for an NHZ, this provides greater certainty that the numbers reported are close to the true winter peak. However, particularly where there is no clear consensus between counts from 7

12 different surveys, there is usually no way of knowing whether the highest count is close to the true peak. Census programmes aiming to cover the whole UK population of each of the four species of wildfowl covered in this report are carried out either annually (for pink-footed goose Anser brachyrhynchus and Greenland white-fronted goose Anser albifrons flavirostris) or every five years (for Greenland barnacle goose Branta leucopsis and whooper swan Cygnus cygnus). More detail on each of these surveys is provided in the relevant species account. These are coordinated surveys aimed primarily at generating whole-population estimates, typically during times when numbers at the main wintering sites for each species are relatively stable. The Goose and Swan Monitoring Programme (GSMP) is an umbrella programme of surveys that includes those described above, as well as others aimed at counting different wintering populations of large wildfowl (see Although these surveys are generally aimed at collecting data on one or two species, flocks of non-target species of swans and geese are also counted. For each species, the GSMP therefore provides data that are supplementary to those collected during dedicated counts. It can be a useful source of information about numbers and distributions of species at times of the year when they may be more mobile than when they are counted by dedicated surveys. The Wetland Bird Survey (WeBS) comprises monthly counts of wetland birds, including geese and swans, at many of the sites known to be important for them (Austin et al. 2014). Because the focus of the survey is on roosting shorebirds, many geese and swans (especially flocks feeding inland) may be missed, and coverage of these species is certainly less comprehensive than in the dedicated surveys described above. However, because WeBS counts are monthly, they can capture information from months when numbers were substantially higher than during the period targeted by the main surveys carried out for these species. WeBS counts may also pick up small flocks of geese and swans outside of the main areas targeted by dedicated surveys. Atlas-derived (in this case winter) relative abundance data for each NHZ can also be applied to population estimates for the whole of Scotland derived from coordinated census counts. For each combination of NHZ and species, the most reliable estimate of the maximum number of wintering birds was selected from the surveys described above. Although the sources of data on which these counts are based derive from slightly different time periods, they should be comparable as there is broad overlap between these periods. For each NHZ the abundance value selected was based principally on size larger counts should be closer to the true peak in numbers being estimated. However, because all estimates apart from those derived from the Atlas were based directly on counts of real birds, Atlas estimates were only preferred if their 95% lower confidence limits were not less than the counts from any of the other surveys. Where either WeBS or Atlas estimates were identified as being more reliable than those from the other surveys, peak abundance is probably an under-estimate, due to the likelihood that coverage of non-dedicated surveys was incomplete. The tables in Appendix B present all abundance estimates, for every combination of species and NHZ, from the surveys described above. For wildfowl and WeBS counts, the peak count of birds for each NHZ during the period considered is presented, along with five-year means of the annual peak count. For Atlas data, estimated abundances with 95% confidence interval are shown. 8

13 It is important to understand that the derivation of national (Scottish) totals by the summing of peak counts across all 21 NHZs, as is generally feasible for breeding birds, is not appropriate for wintering waterfowl. This is because, within one winter, the same birds can contribute to the peak totals for several NHZs because of their mobility. By way of example, a substantial proportion of birds present in NHZs when numbers peak (i.e. passing through Scotland) spend a substantial proportion of the winter further south. The total published APEP estimate for GB is presented for each species, in order to provide a context for the peak numbers reported in each NHZ Skuas and gulls All five skua and gull species were the subject of whole-population surveys, which aimed to comprehensively census all breeding units, as part of the Seabird 2000 (S2K) programme (Mitchell et al. 2004). The S2K dataset provided by the Joint Nature Conservation Committee (JNCC) comprises counts of breeding pairs or active nests for all known colonies of these species at the time of the survey. The location of each colony was either given as a single point or, in the case of large, coastal colonies, as start and end coordinates. For each species, population estimates for each NHZ were derived by summing the relevant counts for all colonies in the NHZ. Counts from a few large coastal colonies that crossed the border of two NHZs were apportioned between NHZs according to the approximate length of coastline between the start and end points of the colony that lay in each NHZ. Population estimates for skuas and gulls from S2K were updated by estimating population trends from a subset of colonies where more recent population data had been gathered as part of the Seabird Monitoring Programme (see Trends at these colonies were used to estimate trends for individual NHZs, allowing for possible variation in population trends between different regions. In order to allow for stochastic inter-annual variability in colony sizes, trends were derived from the difference between average colony size during the five-year periods and When combining proportional trends from multiple colonies, the contribution of each colony was weighted according to its size, and also according to the total number of years contributing data to the estimated trend for that colony. For NHZs with fewer than five colonies for which trends could be calculated, NHZ-level trends were derived by combining data from the focal NHZ with data from colonies in all surrounding NHZs. Inland colonies of gulls are likely to have been systematically under-represented in S2K (this particularly applies to lesser black-backed gull Larus fuscus and herring gull Larus argentatus here) because(i) rooftop-nesting gulls surveyed using the recommended methodologies for these species (vantage point surveys) are likely to be under-counted; and (ii)many small clusters of inland-nesting gulls are likely to have been missed by the survey, as potentially suitable inland sites were not comprehensively surveyed (Mitchell et al. 2004, John Coulson pers. comm., John Calladine pers. comm.). In addition, in parts of the UK where recent surveys allow an assessment of trends since S2K surveys, new inland colonies of lesser black-backed gull and herring gull have been found, and many existing inland colonies of both species have increased in size (Sellers & Shackleton 2011). It is therefore likely that even accurate counts of gulls made in the original survey could now underrepresent many inland gull populations. The evidence to date (Mitchell et al. 2004, Sellers & Shackleton 2011) suggests that recent declines in coastal gull colonies are likely to outweigh any gains in inland colonies. However, given that only one of the 149 herring gull and 79 lesser blackbacked gull colonies that were surveyed repeatedly (allowing trend information to be calculated) 9

14 was inland, it is fair to say that, for regions holding significant numbers of inland colonies, regional declines since S2K (based on repeated counts at coastal colonies) may be over-estimated Other species Specific methods and datasets applicable to other species are dealt with in the relevant speciesspecific sections. 10

15 3. SPECIES POPULATION ESTIMATES For each of 28 species, a table is presented providing estimates of population size in each NHZ, along with a brief description of the methods and datasets used to derive the estimates. A map is shown for each species in which NHZs are labelled with their number, and shaded according to relative population size (between 0 and 100%, according to the proportion of the Scottish population held by the NHZ). The scale at which it is appropriate to interpret the information in both maps and tables is strictly that of the NHZ. Neither values in tables nor shading on maps distinguish between areas of an NHZ that are occupied, or where a species is extremely abundant, and other parts of the same NHZ where that species is rare or absent. 3.1 Whooper swan (Cygnus cygnus) - wintering All estimates for wintering wildfowl are of peak numbers in each NHZ, so they may add up to substantially more than the total population supported in Scotland. However, the numbers reported for any one survey may seriously underestimate peak counts for some NHZs. For this reason, the estimates reported here are drawn from several surveys (see sections and for further detail). The best overall population estimate for whooper swan is provided by counts from the International Swan Census (ISC), which is carried out once every five years (Hall et al. 2012). Data from this census are presented together with those from wider GSMP surveys, from which an annual peak count is available for each year ( Counts from WeBS and the Bird Atlas have been considered separately. It is important to bear in mind that these counts are minimum estimates of the peak numbers of this species that occur in each NHZ during the non-breeding season. Particularly where figures from WeBS or the Atlas are substantially higher than the more comprehensive ISC and GSMP counts, it is Figure 2. Relative wintering abundance of whooper swan among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. likely that the latter surveys do not coincide with peak numbers. In such cases, although the timing of WeBS and Atlas counts may be better, their spatial coverage for migrating wildfowl is likely to be incomplete, resulting in underestimation of peak abundance. 11

16 Intra-seasonal variation in numbers recorded did not follow exactly the same pattern across all NHZs (Appendix A, Figure A1). The two NHZs (16 and 19) where dedicated surveys identified highest peak numbers had peaks in early winter. In NHZ 16 there was a sharp decline to a lower stable level in November-December, followed by departure of the remaining birds by April. In NHZ 19, there was a more gradual decline from December to April. Early winter peaks (of between 100 and 1500 birds) are also seen in some other NHZs (notably 1, 9, 14 and 21), but others returned reasonably stable counts (of between birds) for most of the winter (particularly 2, 3 and 18). In NHZs where numbers peaked in late winter (e.g. 4, 7, 8, 12 and 15), the mean monthly peak was 100 or less. 12

17 Table 1: Estimated peak abundance of wintering whooper swan individuals in each of 21 NHZs, derived from ISC, GSMP, WeBS and Bird Atlas data. Counts from the ISC and GSMP are taken from the period from the winter of 2008/2009 to the winter of 2012/2013. WeBS counts and Atlas estimates are from the five year period from the winter of 2009/2010 to the winter of 2013/2014. The latest APEP estimate for the total wintering population in Great Britain is given at the bottom of the table. See Table B1 for details of counts and estimates from all surveys for each NHZ. NHZ Estimated peak abundance Source 1. Shetland 287 ISC/GSMP 2. Orkney and North Caithness 706 ISC/GSMP 3. Coll, Tiree and the Western Isles 813 (587-1,069) Atlas North West Seaboard 94 (66-138) Atlas Peatlands of Caithness and Sutherland 190 WeBS 6. Western Seaboard 50 (29-73) Atlas Northern Highlands 98 (67-135) Atlas Western Highlands 43 WeBS 9. North East Coastal Plain 1,260 WeBS 10. Central Highlands 28 (4-62) Atlas Cairngorm Massif 145 (57-227) Atlas North East Glens 229 ISC/GSMP 13. East Lochaber 2 WeBS 14. Argyll West and Islands 135 ISC/GSMP 15. L. Lomond, Trossachs & Breadalbane 146 ISC/GSMP 16. Eastern Lowlands 1,524 ISC/GSMP 17. West Central Belt 424 ISC/GSMP 18. Wigtown Machars & Outer Solway 288 ISC/GSMP 19. W. Southern Uplands & Inner Solway 1,188 ISC/GSMP 20. Border Hills 33 (18-52) Atlas Moray Firth 1,600 WeBS APEP GB estimate (2004/ /10): 11,000 (Musgrove et al. 2013) 13

18 3.2 Pink-footed goose (Anser brachyrhynchus) - wintering All estimates for wintering wildfowl are of peak numbers in each NHZ, so they may add up to substantially more than the total population supported in Scotland. However, the numbers reported for any one survey may seriously underestimate peak counts for some NHZs. For this reason, numbers are reported from several surveys (see sections and for further detail). UK populations are surveyed, along with Greylag (Anser anser) populations, by the annual Icelandic-breeding Goose Census (IGC; Mitchell et al. 2014), with 1-2 coordinated counts each year. Additional, independent information on peak numbers in each NHZ is available from the wider GSMP survey, and also the more generic WeBS and Bird Atlas surveys. It is important to bear in mind that these counts are minimum estimates of the peak numbers of this species that occur in each NHZ during the non-breeding season. Particularly where figures from WeBS or the Atlas are substantially higher than the more comprehensive IGC and GSMP counts, it is likely that the latter surveys do not coincide with peak numbers. In such cases, although Figure 3. Relative wintering abundance of pink-footed goose among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. the timing of WeBS and Atlas counts may be better, their spatial coverage for migrating wildfowl is likely to be incomplete, resulting in underestimation of peak abundance. There is a high seasonal turnover of birds in Scotland, as many birds spend the winter in England, only passing through Scotland in transit. Scotland therefore holds more pink-footed geese in the autumn than it does for most of the winter. Intra-seasonal variation in numbers recorded during WeBS varied between NHZs (Appendix A, Figure A2). Dedicated surveys (IGC and GSMP) identified nine NHZs that held peak numbers of at least 10,000 birds. Counts from all surveys were highest in NHZs 9 and 16, indicating that these are the most important for this species. Numbers followed a similar seasonal pattern in both of these NHZs, peaking in October, declining to substantially lower levels by January, and remaining relatively stable until departure in March-May. A similar October peak was recorded in NHZs 12, 20 and 21, but declining to negligible numbers by January except in NHZ 21, where post-january numbers were variable. Pink-footed goose numbers over the winter were maintained at relatively stable levels in NHZs 17, 18 and 19, while in NHZ 2 there appears to be a spring peak in numbers counted during WeBS. 14

19 Table 2: Peak counts of wintering pink-footed goose individuals in each of 21 NHZs, derived from IGS, GSMP and WeBS data. All survey counts and Atlas estimates are from the five year period from the winter of 2009/2010 to the winter of 2013/2014. The latest APEP estimate for the total wintering population in Great Britain is given at the bottom of the table. See Table B2 for details of counts and estimates from all surveys for each NHZ. NHZ Estimated peak abundance Source 1. Shetland 148 GSMP 2. Orkney and North Caithness 20,746 GSMP 3. Coll, Tiree and the Western Isles 65 WeBS 4. North West Seaboard 6 WeBS 5. Peatlands of Caithness and Sutherland 2,070 WeBS 6. Western Seaboard 1 WeBS 7. Northern Highlands 4 WeBS 8. Western Highlands 0-9. North East Coastal Plain 77,859 IGC 10. Central Highlands 7 (0-17) Atlas Cairngorm Massif 36 (0-87) Atlas North East Glens 12,000 GSMP 13. East Lochaber Argyll West and Islands 407 IGC 15. L. Lomond, Trossachs & Breadalbane 6,200 GSMP 16. Eastern Lowlands 162,039 IGC 17. West Central Belt 16,237 IGC 18. Wigtown Machars & Outer Solway 47,491 IGC 19. W. Southern Uplands & Inner Solway 34,621 GSMP 20. Border Hills 47,407 GSMP 21. Moray Firth 35,370 IGC APEP GB estimate (2004/ /10): 360,000 (Musgrove et al. 2013) 15

20 3.3 Greenland white-fronted goose (Anser albifrons flavirostris) - wintering All estimates for wintering wildfowl are of peak numbers in each NHZ, so they may add up to substantially more than the total population supported in Scotland. However, the numbers reported for any one survey may seriously underestimate peak counts for some NHZs. For this reason, numbers are reported from several surveys (see sections and for further detail). The Irish and UK population of Greenland white-fronted geese is surveyed through bi-annual, coordinated counts by the Greenland white-fronted goose Study (GWfS; Fox et al. 2014). Additional, independent information on peak numbers in each NHZ is available from the wider GSMP survey, and also the more generic WeBS and Bird Atlas surveys. It is important to bear in mind that these counts are minimum estimates of the peak numbers of this species that occur in each NHZ during the non-breeding season. Particularly where WeBS figures are substantially higher than the more comprehensive GWfS and GSMP counts, it is likely that these surveys do not coincide with peak numbers. In such cases, although the timing of WeBS counts may be better, the spatial coverage of this survey for migrating wildfowl is likely to be incomplete, resulting in underestimation of peak abundance. Figure 4. Relative wintering abundance of Greenland white-fronted goose among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. Table 3 clearly shows the hierarchical importance of different NHZs for this species, with NHZ 14 consistently holding a maximum of around 10,000 birds, NHZ 3 holding a maximum of just over 1000 birds, and NHZs 2, 6, 17, 18 and 19 all typically holding peak numbers of a few hundred birds. Within NHZs 3 and 14, some localised sites are of disproportionately high importance ( Figure A3 in Appendix A shows that, withinseason, WeBS counts of Greenland white-fronted goose peaked around November-December in NHZs 6, 14 and 18. In NHZs 2, 3, 17 and 19 peak counts were not clearly associated with early or late winter though, in NHZ 3, there was a suggestion of increased numbers in March. However, the most striking feature of monthly WeBS counts of this species is that, for almost all combinations of NHZ and month, the minimum count over the five-year period considered was zero. This indicates that wintering numbers of this species were not reliably captured by monthly WeBS counts, so the variation in Figure A3 should be interpreted with caution. 16

21 Table 3: Peak counts of wintering Greenland white-fronted geese individuals in each of 21 NHZs, derived from GWfS, GSMP and WeBS data. All survey counts and Atlas estimates are from the five year period from the winter of 2009/2010 to the winter of 2013/2014. The latest APEP estimate for the total wintering population in Great Britain is given at the bottom of the table. See Table B3 for details of counts and estimates from all surveys for each NHZ. NHZ Estimated peak abundance Source 1. Shetland peak abundance 2 WeBS 2. Orkney and North Caithness 492 GSMP 3. Coll, Tiree and the Western Isles 2,516 GWfS 4. North West Seaboard 2 WeBS 5. Peatlands of Caithness and 95 WeBS 6. Western Seaboard 360 WeBS 7. Northern Highlands 0-8. Western Highlands 73 WeBS 9. North East Coastal Plain 39 GSMP 10. Central Highlands Cairngorm Massif 36 WeBS 12. North East Glens 25 GSMP 13. East Lochaber 217 WeBS 14. Argyll West and Islands 15,733 WeBS 15. L. Lomond, Trossachs & 220 GSMP 16. Eastern Lowlands 140 GSMP 17. West Central Belt 237 WeBS 18. Wigtown Machars & Outer Solway 470 GWfS 19. W. Southern Uplands & Inner 550 WeBS 20. Border Hills Moray Firth 156 WeBS APEP GB estimate (all races, 2004/ /10): 16,000 (Musgrove et al. 2013) 17

22 3.4 Barnacle goose (Branta leucopsis) - wintering All estimates for wintering wildfowl are of peak numbers in each NHZ, so they may add up to substantially more than the total population supported in Scotland. However, the numbers reported for any one survey may seriously underestimate peak counts for some NHZs. For this reason, numbers are reported from several surveys (see sections and for further detail). Barnacle geese wintering in Britain derive from two distinct breeding populations; the majority wintering in southern Scotland come from Svalbard, whereas those concentrated in Islay derive from Greenland. However, individual geese from these populations cannot be reliably distinguished in the field, so the estimates presented here are totals of birds from both flyways. A census of the wintering population of Greenland barnacle geese, which is based entirely in Britain and Ireland, is carried out every 5 years (International Greenland Barnacle Goose Census - IGBGC; Mitchell et al. 2013). Additional, independent information on peak numbers in each NHZ is available from the wider GSMP survey, and also the more generic WeBS and Bird Atlas surveys. Because IGBGC counts are carried out Figure 6. Relative wintering abundance of barnacle geese among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. in spring, whereas the other counts are carried out in autumn and winter as well, variation between these provides a useful indication of how numbers of this species may vary over the season within some NHZs. It is important to bear in mind that these counts are minimum estimates of the peak numbers of this species that occur in each NHZ during the non-breeding season. Particularly where WeBS figures are substantially higher than the more comprehensive IGBGC and GSMP counts, it is likely that these surveys do not coincide with peak numbers. In such cases, although the timing of WeBS counts may be more suitable, the spatial coverage of this survey for migrating wildfowl is likely to be incomplete, resulting in underestimation of peak abundance. Atlas-derived relative abundances suggest that NHZs 14 and 19 together account for 79% of the Scottish winter population of barnacle goose. However, as shown in Table 4, a far greater proportion of the birds present are picked up by WeBS counts in NHZ 19 than in NHZ 14. The graphs in Figure A4, in Appendix A, indicate that numbers of barnacle geese are relatively stable over the winter in NHZ 19, but indicate an early winter peak in NHZ 14. Whether the late-winter decline in WeBS counts of barnacle goose is due to a genuine drop in numbers held by this NHZ, or to more local movements away from WeBS sites, is unclear. The three NHZs holding the next largest proportions 18

23 of the wintering population (together accounting for 19.6% of the population, according to the Atlasderived relative abundance data) are NHZs 2, 3 and 4. Intra-seasonal variation in WeBS counts of barnacle goose within these NHZs suggest a gradual build of numbers from early to late winter in NHZ 2, multiple peaks in early winter, late winter and spring in NHZ 3, and a single spring peak in NHZ 4. However, as in NHZ 14, the maximum number of birds recorded during WeBS counts in each of these NHZs is substantially smaller than the number of birds recorded during dedicated goose counts. With such a clear indication of incomplete coverage, the intra-seasonal variation indicated by WeBS data may not be reliable. Table 4: Peak counts of wintering barnacle geese individuals in each of 21 NHZs, derived from IGBGC, GSMP and WeBS data. IGBGC counts are from the two censuses carried out in the winters of 2009/10 and 2013/14, while GSMP and WeBS counts are from the five year period from the winter of 2009/2010 to the winter of 2013/2014. The latest APEP estimate for the total wintering population in Great Britain is given at the bottom of the table. See Table B4 for details of counts and estimates from all surveys for each NHZ. NHZ Estimated peak abundance Source 1. Shetland 32 WeBS 2. Orkney and North Caithness 6,514 GSMP 3. Coll, Tiree and the Western Isles 12,604 IGBGC 4. North West Seaboard 1,148 IGBGC 5. Peatlands of Caithness and Sutherland 600 WeBS 6. Western Seaboard 983 IGBGC 7. Northern Highlands 0-8. Western Highlands 4 WeBS 9. North East Coastal Plain 361 WeBS 10. Central Highlands Cairngorm Massif North East Glens 7 WeBS 13. East Lochaber Argyll West and Islands 48,815 IGBGC 15. L. Lomond, Trossachs & Breadalbane 2 WeBS 16. Eastern Lowlands 1,384 GSMP 17. West Central Belt 152 WeBS 18. Wigtown Machars & Outer Solway 5,116 WeBS 19. W. Southern Uplands & Inner Solway 47,000 GSMP 20. Border Hills 328 GSMP 21. Moray Firth 800 GSMP APEP GB estimate (all flyways, 2004/ /10): 94,000 (Musgrove et al. 2013) 19

24 3.5 Black grouse (Tetrao tetrix) - breeding This species was surveyed across its British range in 2005, the Scottish population estimate being 3344 (95% CI: ) displaying males (Sim et al. 2008). The survey also generated population estimates for four Scottish regions: North Scotland 770 (95% C.I ), Northeast Scotland 1499 (95% C.I ), Southeast Scotland 257 (95% C.I ) and Southwest Scotland 807 (95% C.I ). The national estimate and its 95% confidence intervals were apportioned between overlapping NHZs, using Atlas-derived relative abundance. In order to check that Atlas-derived abundance apportioned the national estimate correctly, each NHZ estimate was apportioned among the four survey regions, according to the area of overlap between NHZ and survey region. The resulting, NHZbased regional estimates compared favourably with the regional estimates produced by the survey (North Scotland 804, Northeast Scotland 1728, Southeast Scotland 137 and Southwest Scotland 644). It should be borne in mind that trends Figure 6. Relative breeding abundance of black grouse among 21 Scottish NHZs. Shading indicates breeding abundance as a proportion of the whole breeding population. based on counts of displaying males, especially in a polygynous species like black grouse, could be misleading if populations are subject to gender-specific survival or recruitment. Annual counts of lekking males, carried out by regional black grouse study groups, are available for many parts of Scotland. However, these are variable in quality, and the (often poorly defined) boundaries of counting regions do not align exactly with those of NHZs, most of which are incompletely covered by these regions. Considering the relative stability of most Scottish black grouse populations over the past 10 years (borne out by recent trends from most black grouse study groups), it is likely that apportioning the national survey estimate between NHZs using Atlas-derived relative abundance, as described above, is the most efficient and reliable approach (Phil Warren pers. comm.). 20

25 Table 5: Estimated number (with 95% confidence limits) of displaying black grouse males in each of 21 NHZs in 2005, based on single population estimate from a national survey, apportioned between NHZs according to Atlas-derived relative abundance. NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif 1, , North East Glens East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane , Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth TOTAL 3,344 2,580 4,171 21

26 3.6 Red-throated diver (Gavia stellata) - breeding A national survey was carried out in 2006, which generated population estimates for Shetland, Orkney, Outer Hebrides, Inner Hebrides and mainland Scotland (Dillon et al. 2009). Each of the estimates for Shetland, Orkney and Outer Hebrides were assigned to the relevant NHZ, while the estimates for Inner Hebrides and mainland Scotland were apportioned between multiple overlapping NHZs according to Atlas-derived proportional abundance estimates (section 2.1.2). The estimates for NHZs 2 and 3 are taken entirely from the abundance estimated by Dillon et al. (2009) for Orkney and the Outer Hebrides, respectively. No breeding divers were found in the mainland part of NHZ 2, and any bias resulting from the exclusion of Coll & Tiree pairs from the estimate for NHZ 3 is likely to be negligible (Andrew Stevenson pers. comm.). All values were rounded to the nearest whole number. Figure 7. Relative breeding abundance of red-throated diver among 21 Scottish NHZs. Shading indicates breeding abundance as a proportion of the whole breeding population. 22

27 Table 6: Number of breeding pairs of red-throated divers in each of 21 NHZs in These numbers are based on the regional densities estimated by Dillon et al (2009), which are apportioned between the NHZs overlapping these regions according to relative abundance values derived from Balmer et al. (2013). NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands L. Lomond, Trossachs & Breadalbane Eastern Lowlands West Central Belt Wigtown Machars & Outer Solway W. Southern Uplands & Inner Solway Border Hills Moray Firth TOTAL 1, ,722 23

28 3.7 Black-throated diver (Gavia arctica) - breeding RSPB hold comprehensive survey records on the breeding population of this species up to 2006, after which monitoring effort was greatly reduced. The number of occupied territories found in this year was used to derive the population estimate for each NHZ. The estimate includes all confirmed and probable territories plus half of the possible territories identified in The minimum estimate is taken as the number of confirmed territories only, while the maximum is taken as the total number of confirmed, probable and possible territories plus all unchecked territories known to have held birds previously. Nevertheless, as some pairs could have been missed, the numbers presented in Table 7 may be slightly conservative. Figure 8. Relative breeding abundance of blackthroated diver among 21 Scottish NHZs. Shading indicates breeding abundance as a proportion of the whole breeding population. 24

29 Table 7: Numbers of breeding pairs of black-throated divers in each of 21 NHZs in These numbers are based directly on the numbers of confirmed, probable and possible pairs found by the RSPB in NHZ Estimate Minimum Maximum 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands L. Lomond, Trossachs & Breadalbane Eastern Lowlands West Central Belt Wigtown Machars & Outer Solway W. Southern Uplands & Inner Solway Border Hills Moray Firth TOTAL

30 3.8 Red kite (Milvus milvus) - breeding This species was monitored comprehensively by RSPB until 2012, with regional breakdown of breeding numbers reported in the Scottish Raptor Monitoring Scheme (SRMS) annual report (Challis et al. 2014). Since 2013, the Scottish Raptor Study Group (SRSG) has taken on responsibility for coordinating monitoring of red kite and providing data to the SRMS. The number of pairs recorded as breeding in 2013 was 244. This is the highest number of red kite pairs recorded in Scotland since the species reintroduction in 1989 (Challis et al. 2014), and it is likely that less than 5% of pairs were missed (Duncan Orr-Ewing, pers. comm.). The data available through the SRMS for 2013 are therefore likely to be the best available. Furthermore, because Scottish populations of red kites are still centred around the four reintroduction sites (in North Scotland, Aberdeenshire, Central Scotland and Dumfries Galloway), and most of the SRMS data for this species are associated with breeding locations, Figure 9. Relative breeding abundance of red kite among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. pairs can be assigned to individual NHZs with a high degree of certainty. However, it should be borne in mind that the Scottish population of red kite is likely to continue its recent expansion in range and abundance, such that many of the estimates given in Table 8 below may quickly become outdated. & 26

31 Table 8: Number of pairs of red kites reported breeding in 2013 in each of 21 NHZs. Based on data from the Scottish Raptor Monitoring Scheme, supplemented by data from SRSG. NHZ Estimate 1. Shetland 0 2. Orkney and North Caithness 0 3. Coll, Tiree and the Western Isles 0 4. North West Seaboard 0 5. The Peatlands of Caithness and Sutherland 0 6. Western Seaboard 0 7. Northern Highlands 9 8. Western Highlands 0 9. North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth 50 TOTAL

32 3.9 White-tailed eagle (Haliaeetus albicilla) - breeding Comprehensive monitoring of this species was coordinated by RSPB until 2013, with regional breakdown of breeding numbers reported in the SRMS annual report. The number of breeding pairs monitored and reported to the SRMS in 2013 (82) was the highest since the species reintroduction in The numbers for 2013 can be regarded as being close to comprehensive. The majority of records in the SRMS are spatially explicit and can be assigned to individual NHZs. Remaining records have been divided between possible NHZs on the basis of probability (determined by overlap between the recording area and the relevant NHZs, weighted by the Atlas-derived proportional abundance of whitetailed eagles within each of these NHZs). As the Scottish population of white-tailed eagle is currently expanding, with several pairs in the newly established east coast population now breeding, many of the estimates given in Table 9 may quickly become outdated. Figure 10. Relative breeding abundance of white-tailed eagle among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. 28

33 Table 9: Estimated number of territorial pairs of white-tailed eagle in 21 NHZs, from 2013 Scottish Raptor Monitoring Scheme data. NHZ Estimate 1. Shetland 0 2. Orkney and North Caithness 1 3. Coll, Tiree and the Western Isles North West Seaboard 5 5. The Peatlands of Caithness and Sutherland 1 6. Western Seaboard Northern Highlands 1 8. Western Highlands 6 9. North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane 0 a 16. Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth 0 TOTAL 82 a. A pair was recorded breeding in NHZ 15 in 2014, for the first time since white-tailed eagle was reintroduced. 29

34 3.10 Hen harrier (Circus cyaneus) - breeding This species was the subject of a national survey in 2010 (Hayhow et al. 2013). The estimates below were largely produced by SNH based on the results of this survey, supplemented by additional information from the SRMS, the Rare Breeding Birds Panel and the SRSG (Andrew Stevenson pers. comm.). In some parts of its range, breeding aggregations of hen harrier are semi-colonial and, particularly in years when prey is abundant, polygyny can be common, making it difficult to assess the breeding population in terms of the number of breeding pairs. As a ground-nesting bird, this species can be more vulnerable to nest predation by ground predators than most other Scottish raptor species, and there is evidence that populations in some areas are limited by illegal predator control (Fielding et al. 2011). However, regional populations of this species can quickly increase in abundance during conditions Figure 11. Relative breeding abundance of hen harrier among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. favouring both breeding productivity and survival (Redpath & Thirgood 2007, Langholm Moor Demonstration Project 2014). The NHZ abundance figures in Table 10 may, therefore, be subject to substantial changes over relatively short timescales, as indicated by the regional increases in southern Scotland and Orkney since 2010 (Table 10). There will be another UK hen harrier survey in 2016, which will provide a timely opportunity to update the figures below more comprehensively. 30

35 Table 10: Upper and lower estimates for number of breeding hen harrier pairs in each of 21 NHZs in Based on data from the 2010 hen harrier Survey, supplemented by data from SRSG. NHZ Central Lower Upper estimate estimate estimate 1. Shetland Orkney and North Caithness 105 a Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands Loch Lomond, Trossachs and Breadalbane Eastern Lowlands 3 b West Central Belt 8 b Wigtown Machars and Outer Solway Coast 0 b Western Southern Uplands and Inner Solway 18 b Border Hills 13 b Moray Firth TOTAL a. The estimate for NHZ 2 is taken from the 2014 SRMS report, which reports a minimum of 105 occupied territories (23 more than estimated in 2010) out of 216 checked, for Orkney alone. b. The total number of breeding hen harriers estimated here for southern Scotland (NHZs 16, 17, 18, 19 and 20), based on 2010 survey data, is 42 pairs. This is likely to be an under-estimate, as the minimum number of breeding pairs recorded in the 2014 SRMS report in three SRSG regions within this area (South Strathclyde, Lothian and Borders, and Dumfries and Galloway) was

36 3.11 Goshawk (Accipiter gentilis) - breeding Forrester et al. (2007) base their estimate of the Scottish goshawk population on the fact that around 100 of 162 known goshawk home ranges were occupied during the period (Forrester et al. 2007). They inferred from this observation that the breeding population might have been somewhere between these two figures, tentatively placing the estimate at a minimum of 130 breeding pairs. However, the main source of uncertainty in this estimate lies in an unknown quantity - the number of goshawk home ranges in Scotland that have not yet been found. Rather than proportional abundances being calibrated using an arbitrarily derived national total, the reference estimate is taken from areas where more is known about goshawks than elsewhere in Scotland. This estimate (of 103 pairs) is the total number of breeding goshawk records submitted to the SRMS in 2013, in three parts of Scotland where the species is intensively studied. The Figure 12. Relative breeding abundance of goshawk among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. proportional abundance values (derived from the Bird Atlas data) were then summed for the six NHZs that broadly overlap this area (0.81). Under a scenario where most of the pairs breeding within these intensively studied areas were found, the national total could be estimated by dividing pairs found (103) by proportional abundance for these NHZs (0.81) to give 127. The NHZ population estimates in Table 11 were derived by redistributing these 127 pairs among NHZs, according to proportional abundance. However, the assumption that most pairs in NHZs overlapping intensive studies were found is almost certainly unjustified (Mick Marquiss pers. comm.). The 2013 Scottish Raptor Monitoring report (Challis et al. 2014) states that 174 home ranges were checked (compared to 132 in 2004), and 124 breeding pairs found (compared to 86 in 2004). Moreover, the fact that the Atlas reports no breeding birds in NHZs 14 and 17, despite the fact that small numbers of birds are known to breed in these regions, highlights the likelihood that goshawks probably occur in other NHZs with estimates of zero pairs. Thus, an estimate of 127 pairs is probably highly conservative, and the true number of goshawk pairs in Scotland almost certainly higher. 32

37 Table 11: Number of goshawk breeding pairs in 2013 estimated for 21 NHZs. Based on data from the Scottish Raptor Monitoring Scheme, adjusted using Atlas-derived relative abundance data. No overall confidence limits are available for the whole population, because NHZ confidence limits are based largely on relative abundance data. NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands <5 a Loch Lomond, The Trossachs and Breadalbane Eastern Lowlands West Central Belt <5 a Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth TOTAL a. Minimum population estimate provided by Andrew Stevenson (pers. comm.). 33

38 3.12 Golden eagle (Aquila chrysaetos) - breeding Estimates of the number of occupied territories in each NHZ population, based on data from the 2003 national golden eagle survey (Eaton et al. 2007), were included in the golden eagle Framework (Whitfield et al. 2008). Although somewhat dated now, these estimates are of high quality, as national surveys of this species typically achieve near complete coverage. A national survey of this species, organised by the RSPB, is taking place in 2015, and when it has been completed a comprehensive update of the estimates In Table 12 should be possible. Figure 13. Relative breeding abundance of golden eagle among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. 34

39 Table 12: Number of golden eagle occupied breeding territories in 2003 estimated for 21 NHZs. Based on data from the 2003 golden eagle Survey. NHZ Estimate 1. Shetland 0 2. Orkney and North Caithness 0 3. Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth 0 TOTAL

40 3.13 Osprey (Pandion haliaetus) - breeding Forrester et al. (2007) estimated that, in 2004, when 182 pairs of breeding ospreys were reported in the SRMS annual report, the actual breeding population was close to 200 pairs, due to the increasing probability that some pairs were missed as the population expanded. The number of known breeding sites checked at this time was 230, implying an occupancy rate of 79% (Challis et al. 2014). In 2013, 197 pairs of osprey were reported in Scotland, and 290 known sites were checked, implying either a lower occupancy rate of previous breeding sites, or more comprehensive coverage of these sites (Challis et al. 2014). However, it is likely that some pairs remain unrecorded, particularly in areas where there is limited searching for new breeding sites. Range expansion and population increase currently appear to be strongest in NHZs 14, 19 and 20 (Andrew Stevenson pers. comm.). Figure 14. Relative breeding abundance of osprey among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. From the 2013 SRMS data, 85 breeding attempt locations recorded at a spatial resolution of 1km or better were assigned to the NHZs in which they were known to be located. A further 26 breeding sites recorded at the 10km square level were assigned to the NHZ overlapping with the largest portion of the relevant 10km square, and contributed values of 0 and 1, respectively, to the lower and upper confidence limits of all overlapping NHZs. The remaining 86 breeding records, for which no grid reference was recorded, were assigned to NHZs according to SRMS sub-regions (Challis et al. 2014). One thousand randomised sets of osprey locations were generated, with the chance of an NHZ holding each nest being proportional to the overlap between that nest s sub-region and the NHZ. For each NHZ, the estimated number of nests and the lower and upper confidence limits were taken as the means of the 500 th and 501 st, the 25 th and 26 th, and 975 th and 976 th values, respectively. All values were rounded to the nearest whole number. 36

41 Table 13: Number of osprey breeding pairs in 2013 estimated for 21 NHZs. Based on 2013 breeding records submitted to the Scottish Raptor Monitoring Scheme. No overall confidence limits are available for the whole population, because NHZ confidence limits are based on uncertainties in assigning known records between NHZs, rather than on variance of total estimate. NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland Orkney and North Caithness 1 a Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth TOTAL a. Information on 1-2 pairs in NHZ 2 received from SNH staff (Andrew Stevenson, pers. comm.). 37

42 3.14 Kestrel (Falco tinnunculus) - breeding As noted in Birds of Scotland (Forrester et al. 2007), this is one of the few raptor species that is not the subject of dedicated national surveys, and there is a lack of reliable information on the absolute size of the Scottish population. All recent published estimates for Scotland can be traced back to an estimate in Greenwood et al. (2003) of 11,000 breeding pairs in the period This is based on extrapolation from densities found in intensive study areas which, at least in Scotland, represent only a small proportion of the species range. What can be stated with much greater certainty, based on smoothed annual trends calculated from BBS data, is that the Scottish population of kestrel has undergone a decline of 65% in the period (Harris et al. 2014). Applying this estimate of change to the abundance estimate of Greenwood et al. (2003), gives an estimated Scottish population of 3,850 (95% C.I. 2,750-5,500). This figure was divided between NHZs Figure 15. Relative breeding abundance of kestrel among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. according to Atlas-derived estimates of relative abundance, which were rounded to the nearest whole number. The two main caveats to place around these estimates are that: (i) the baseline population estimate from which they are derived is itself derived from a small and potentially unrepresentative sample; and (ii) the population may still be declining, and so could be smaller than suggested by the 2012 index value. 38

43 Table 14: Estimated number (with 95% confidence limits) of breeding pairs of kestrel in each of 21 NHZs in 2013, based on 1990 extrapolation from densities in intensively studied areas, and adjusted for population trend in Scotland from 1990 to NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth TOTAL 3,850 2,750 5,500 39

44 3.15 Merlin (Falco columbarius) - breeding This status of merlins in Scotland was assessed in 2008 (Ewing et al. 2011), with a national survey comprising a mix of comprehensive coverage in intensive study areas, and stratified coverage elsewhere. It found that there were 733 (95% C.I ) pairs of breeding merlins in Scotland but the conclusion was subsequently questioned by others in the ornithological community who thought that this estimate was too high (Haworth & Fielding 2011). Disagreement on interpretation of the survey results lay principally in the extrapolation from the stratified element of the survey, the numbers of merlins in intensively studied areas being relatively undisputed. Rather than taking any national estimate as a starting point for estimating NHZ population sizes, merlin numbers were extracted from some NHZs where they were intensively studied, where it is likely that a high proportion of merlin pairs were found. These NHZs were 2, 3, 11 and 12. In NHZ 2, coverage Figure 16. Relative breeding abundance of merlin among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. by the SRSG during the merlin survey was supplemented by a separate study of the Lewis Peatlands (Rae 2010). Between them, these NHZs held a total of 133 pairs of merlin, with a further 17 territories showing signs of occupation. There is a remarkably tight relationship between the numbers of merlin pairs found in these NHZs and the Atlas-derived proportional abundance scores (r 2 > 0.999). According to the relative abundance scores from the Bird Atlas, these NHZs held 33% of the Scottish population. Dividing the numbers of pairs and territories found by 0.33 gives 403 pairs and 455 territories, with a central estimate of 434 pairs. These are the numbers predicted for the whole of Scotland if the relationship between relative and real abundance was the same across all NHZs (at the time of the survey). These were apportioned between all 21 NHZs according to Atlasderived proportional abundance scores and their confidence intervals. It should be noted, however, that if the NHZs with intensive studies held appreciable numbers of breeding merlin that were not detected during the survey, the numbers presented here are likely to be underestimates. 40

45 Table 15: Estimated number (with 95% confidence limits) of breeding pairs of merlin in each of 21 NHZs in 2008, based on calibration of Atlas-derived proportional abundances against densities in intensively studied areas. NHZ Estimate Lower Upper 95% CL 95% CL 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth TOTAL

46 3.16 Peregrine (Falco peregrinus) - breeding This species was the subject of national surveys in 2002 (Banks et al. 2010) and 2014 (Wilson et al. 2015). Data for some NHZs are from complete (or near-complete) checks of all known territories, and so are likely to closely reflect the number of pairs actually breeding in the NHZ. An estimate based on records of observed pairs and singletons is the basis for the minimum figure given in Table 16. In most NHZs, a proportion of known territories was not checked. For these areas, the population is estimated by extrapolating from numbers found at checked sites to estimate the numbers present at all known sites. The maximum estimate assumes that unchecked sites were occupied at the same rate as checked sites. However, likelihood of a site being checked by raptor workers with prior knowledge of an area is positively related to likelihood of occupation. The best estimate is therefore based on a more realistic rate of occupancy, derived from a subsample of sites visited by non- SRSG members. This approach assumes that few Figure 17. Relative breeding abundance of peregrine among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. unknown sites were occupied, which is likely to be reasonable for the majority of NHZs where populations are stable or declining. However, in a few lowland NHZs where peregrine populations appear to be increasing (i.e. 9, 16, 17 and 18), some newly occupied sites may not have been found yet, so numbers in Table 16 could be slight underestimates. Note also that at the time of writing the analysis of the 2014 survey results was still ongoing, and final results may vary from those in Table 16 by small numbers of pairs. 42

47 Table 16: Estimated number (with minimum and maximum estimates) of breeding pairs of peregrine in each of 21 NHZs in 2014, based directly on 2014 national survey dataset. NHZ Estimate Minimum Maximum 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth TOTAL

48 3.17 Golden plover (Pluvialis apricaria) - breeding NHZ-specific population estimates were generated by Massimino et al. (2011) using predictive models based on the relationship between BBSderived densities in 2009, and a suite of habitat and other environmental explanatory variables. The sum of these predicted values gives a national estimate of 37,475 pairs. This is in line with the 95% C.I. of 27,468 42,647, generated by applying the Atlas-derived Scottish proportion of the GB population to the GB estimate for golden plover given by Musgrove et al. (2013). Also, a recent estimate of breeding golden plover numbers in NHZ 5, assembled as part of a study on wader populations in and around protected areas and reserves (Bellamy & Eaton 2009, Paul Bellamy pers. comm.), is 3,628 breeding pairs, which is broadly similar to the estimate of 3,125 pairs generated by the model (Massimino et al. 2011). Figure 18. Relative breeding abundance of golden plover among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. 44

49 Table 17: Estimated number (with 95% confidence limits) of golden plover breeding pairs in each of 21 NHZs in 2005, based on predictions from a habitat model using bird distribution data from NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland 5,665 5,195 6, Orkney and North Caithness 1,474 1,365 1, Coll, Tiree and the Western Isles 4,194 3,876 4, North West Seaboard 3,569 3,297 3, The Peatlands of Caithness and Sutherland 3,125 2,907 3, Western Seaboard 1,606 1,467 1, Northern Highlands 3,009 2,762 3, Western Highlands North East Coastal Plain Central Highlands 2,702 2,476 2, Cairngorm Massif 3,616 3,309 3, North East Glens East Lochaber Argyll West and Islands 1,429 1,309 1, Loch Lomond, The Trossachs and Breadalbane 1, , Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills 1, , Moray Firth TOTAL 37,480 34,472 40,477 45

50 3.18 Dunlin (Calidris alpina) - breeding Musgrove et al. (2013) take their estimate of 8,000-10,000 pairs of dunlin from Forrester et al. (2007) who, in turn, derive their estimate from a number of regional studies. The geographical boundaries of some of these studies are welldefined enough that the breeding pairs of dunlin they report can be assigned to particular NHZs. These include a 1987 estimate of 1,700 pairs for Shetland (Pennington et al. 2004), and estimates from between 1995 and early 2000s, deriving from Jackson (2001) and a variety of unpublished sources, for NHZ 3 (5,996, in total). Abundance change data can be used to update the Shetland estimate; since the 1990 Atlas, there has been an estimated 21% increase in the Shetland population, bringing the estimate for NHZ 2 to 2,054 pairs. In addition, a recent estimate of 2,627 pairs for NHZ 5 was based on intensive stratified sampling (Bellamy & Eaton 2009, Paul Bellamy pers. comm.). Together, these Figure 19. Relative breeding abundance of dunlin among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. estimates add up to 10,677 pairs, and account for all of the birds in 3 NHZs holding 77% of the Scottish population (according to Atlas-derived relative abundance scores). Dividing 10,677 by 0.77 gives a total Scottish abundance estimate of 13,871, with 3,194 pairs to be apportioned among the remaining NHZs in accordance with Atlas-derived relative abundance scores. This estimate is in line with the stated expectation of Forrester et al. (2007) that their estimate of 8,000 10,000 was too low. In fact, given the fact that many multi-species surveys that aren t focussed on finding nests will underestimate numbers of breeding dunlin (Rae & Watson 1998), our estimate may also be too low. However, breeding dunlin populations have declined recently in several regions of Scotland and Northern England (Sim et al. 2005) and machair-breeding dunlins in the Uists (NHZ 3) have declined by approximately 50% (from 1004 to 513) (Fuller et al. 2007, Calladine et al. 2014). In conclusion, this species is difficult to survey, and our poor understanding of dunlin abundance is underscored by the very large confidence intervals around most of the estimates in Table 18. Our knowledge about dunlin breeding numbers would benefit considerably from further, dedicated survey work. 46

51 Table 18: Estimated numbers of breeding pairs of dunlin in 21 NHZs, based on surveys carried out between 1980 and 2000, summarised in Forrester et al. (2007). NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland 2,054 1,767 2, Orkney and North Caithness 1, , Coll, Tiree and the Western Isles 5,996 2,634 13, North West Seaboard , The Peatlands of Caithness and Sutherland 2, , Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands , Loch Lomond, The Trossachs and Breadalbane Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth TOTAL 13,313 5,904 28,939 47

52 3.19 Snipe (Gallinago gallinago) - breeding Snipe trends, in common with those of other wader species, can vary widely between regions (e.g. Sim et al. 2005). While some regional estimates for this species exist (e.g. Murray et al. 1998), updating national population estimates was considered to be a more reliable way to derive current regional population sizes (O Brien et al. 2002, O Brien & White 2004) using BBS trends. The estimate for the farmland population of snipe in Scotland was 32,977 pairs, and dates from This was updated using the BBS trend applying to the period (a decrease of 26%). The estimate for the upland population of snipe in Scotland was 9,099 pairs, and dates from This was updated using the BBS trend applying to the period (an increase of 12%). The total updated estimate for both upland and lowland breeding snipe was then apportioned between NHZs according to Atlasderived proportional abundance estimates. It should be borne in mind that this approach assumes that these trends apply uniformly to all Figure 20. Relative breeding abundance of snipe among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. habitats, which may not be the case. However, in the absence of habitat-specific trends (or population estimates based on more recent surveys) it is the best approach that can be taken. 48

53 Table 19: Estimated number of breeding pairs of snipe in 21 NHZs. The data on which these are based were taken from surveys from for birds breeding in farmland, and from for birds breeding in upland areas. These were updated according to BBS trends, and apportioned between NHZs using Atlas-derived relative abundances. NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland 6,728 5,672 8, Orkney and North Caithness 3,326 2,604 4, Coll, Tiree and the Western Isles 6,780 5,622 8, North West Seaboard 1, , Peatlands of Caithness and Sutherland 2,673 2,214 3, Western Seaboard 2,025 1,669 2, Northern Highlands 1,309 1,024 1, Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif 1,614 1,218 2, North East Glens 1, , East Lochaber Argyll West and Islands 1,289 1,026 1, L. Lomond, Trossachs & Breadalbane , Eastern Lowlands West Central Belt Wigtown Machars & Outer Solway W. Southern Uplands & Inner Solway 1, , Border Hills , Moray Firth TOTAL 34,594 29,780 43,553 49

54 3.20 Whimbrel (Numenius phaeopus) - breeding Musgrove et al. (2013) estimate the national population of whimbrel to be pairs. This estimate is taken from Forrester et al. (2007), and is based on a handful of primary sources. These are widely acknowledged to be out of date (Mark Eaton pers. comm., Pete Ellis pers. comm.). An unpublished study by Natural Research in 2009, based on a mixture of comprehensive surveys and stratified sampling, placed the Shetland population at 290 pairs. The Atlas-derived proportional abundance scores place the Shetland population at 95% of the national population. In the absence of comparably thorough estimates for the rest of Scotland, it was considered appropriate to estimate the (small) numbers of birds breeding in other NHZs by scaling the proportional abundance scores so that they are in line with the Shetland estimate of 290 pairs. More recent, intensive work, associated with Figure 21. Relative breeding abundance of whimbrel among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. assessment for a proposed wind farm, suggests that the main concentrations of whimbrel on mainland Shetland are relatively stable (Digger Jackson pers. comm.). The same study suggests that numbers in outlying parts of mainland Shetland, as well as some of the smaller islands, may have been slightly under-represented in the 2009 estimate, such that the Shetland population may be slightly larger than previously thought (Digger Jackson pers. comm.). However, more analysis is needed before the estimate can be formally revised. 50

55 Table 20: Estimated number of breeding pairs of whimbrel in 21 NHZs. These are based on an estimate of the Shetland population from 2009, which was used to assign populations to other NHZs based on Atlas-derived relative proportional abundances. NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard Peatlands of Caithness and Sutherland 1 a Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands L. Lomond, Trossachs & Breadalbane Eastern Lowlands West Central Belt Wigtown Machars & Outer Solway W. Southern Uplands & Inner Solway Border Hills Moray Firth TOTAL a. At least one probable record of breeding Whimbrel reported in NHZ 5 by Balmer et al. (2013) 51

56 3.21 Curlew (Numenius arquata) - breeding The NHZ-specific population estimates presented in Table 21 were taken from the work of Massimino et al. (2011), who used predictive models based on the relationship between BBS-derived densities in 2009, and a suite of habitat and other environmental explanatory variables. The number of breeding pairs in Scotland estimated by Massimino et al. (2011) across all NHZs was 30,194. The last studies to generate Scottish population estimates for curlew were O Brien et al. (2002) and O Brien & White (2004). These estimated, respectively, 45,627 pairs of curlew annually breeding on Scottish farmland between 1997 and 2000; and 13,057 pairs of curlew breeding in upland habitats in Scotland between 1985 and Each of these estimates can be updated using BBS trends for changes in the breeding curlew population (approximately 35% decrease for farmland estimate and 55% reduction for upland component), and then Figure 22. Relative breeding abundance of curlew among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. added together to give a contemporary estimate of 35,533. This is sufficiently similar to the estimate of Massimino et al. (2011) to provide further confidence in that estimate. Further confidence is inspired by the close relationship between Atlas derived proportional abundance and modelled population estimates, though NHZ 2 does not conform as closely to this relationship as the other NHZs. The cause of this discrepancy is not clear, and doubt has been cast on both estimates for NHZ 2, that of Massimino et al. probably being too low, at least in relation to the Shetland estimate (David Douglas pers. comm.) and the estimate based on Atlas-derived abundance probably being too high (Alan Leitch pers. comm.). However, given recent declines of curlew across Scotland, and the numbers reported for Orkney by O Brien et al. (2002) and O Brien & White (2004), it seems likely that the habitat-based estimate of Massimino et al. (2011) is more realistic. 52

57 Table 21: Estimated number (with 95% confidence limits) of curlew breeding pairs in each of 21 NHZs in 2005, based on predictions from a habitat model using bird distribution data from NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland 4,227 a 3,643 4, Orkney and North Caithness 3,233 a 2,915 3, Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland 1,737 1,555 1, Western Seaboard Northern Highlands Western Highlands North East Coastal Plain 1,037 9,29 1, Central Highlands Cairngorm Massif 1,322 1,187 1, North East Glens 2,815 2,556 3, East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane 1,434 1,287 1, Eastern Lowlands 3,253 2,963 3, West Central Belt 2,303 2,085 2, Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway 4,284 3,851 4, Border Hills 1,400 1,261 1, Moray Firth TOTAL 30,194 26,002 33,457 a. RSPB staff on Shetland and Orkney have expressed doubt that the former should support more breeding Curlew than the latter, and apparent discrepancies for NHZ 2 between the estimate presented in this table and relative abundance from the Atlas (see Appendix C for details), suggest that the estimates for NHZs 1 and 2 should be treated with caution. 53

58 3.22 Greenshank (Tringa nebularia) - breeding The last national survey of greenshank in 1995 estimated 1,080 breeding pairs in Scotland (Hancock et al. 1997). Since that time, there has been relatively little systematic work on this species, such that the estimate provided by Musgrove et al. (2013) for GB (which for this species is equivalent to the estimate for Scotland) was taken directly from that survey. Between the last two breeding bird atlases, occupancy rates of greenshank within its range increased by 5.2% (95% confidence interval 2.8% - 7.7%). This is equivalent to a numerical increase of approximately 20% (10% - 30%), which is in line with increases in NHZ 3 (Andrew Stevenson pers. comm.). It is also in keeping with the finding of Sim et al. (2005) that, between 1980 and 2000, greenshank populations increased by 79% in part of NHZ 5, but did not change appreciably in other parts of this NHZ. So, before apportioning the national estimate Figure 23. Relative breeding abundance of greenshank among 21 Scottish NHZs. Shading indicates peak count as a percentage of all peak counts. from 1995 between NHZs according to Atlas-derived relative abundance, the estimate was updated to take account of the increase in number of greenshanks since that time. Greenshank was included in a near-comprehensive survey of NHZ 5 in 2009, which estimated that the breeding population was 1052 (95% C.I ,752) pairs (Bellamy & Eaton 2009, Paul Bellamy pers. comm.). While the 95% confidence interval of this estimate encompasses the estimate for this area in Table 22, the estimate itself is more than twice as large. However, without more detailed study, it is not possible to say whether this is because the 1995 survey underestimated the national greenshank population, whether the assessment of Bellamy & Eaton (2009) is an overestimate, or whether there has been an increase between the two periods that was not captured in the Atlas data. If the newer estimate is accurate, it is possible that other NHZs also hold more pairs than estimated here, and the total in Table 22 may be a significant underestimate of the national breeding population. 54

59 Table 22: Estimated numbers of breeding pairs of greenshank in 21 NHZs. These are based on the findings of a survey in 1995, with abundance updated according to inter-atlas changes in tetrad-occupancy. NHZ Estimate Lower 95% CL Upper 95% CL 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard The Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands Loch Lomond, The Trossachs and Breadalbane Eastern Lowlands West Central Belt Wigtown Machars and Outer Solway Coast Western Southern Uplands and Inner Solway Border Hills Moray Firth TOTAL 1, ,792 55

60 3.23 Short-eared owl (Asio flammeus) - breeding The dearth of knowledge on shorteared owl and, in particular, on the size of its Scottish population, is reflected in the entry for this species in Forrester et al. (2007), which places the Scottish breeding population at 125-1,250 pairs. This estimate is loosely based on extrapolation from regional estimates and reporting, and part of the variation it encompasses derives from the difference in the number of pairs thought to breed during poor and good vole years. The equivalent estimate made by Musgrove et al. (2013) is 483 1,693 pairs (610 2,140 in Great Britain, minus the 21% of the population estimated from Balmer et al. (2013) to occur outside of Scotland). This is taken from Gibbons et al. (1993), who extrapolated from the densities found in some intensively studied areas to the rest of the species range in Great Britain. These estimates are relatively close to one another, given the range of values they span. According to changes in occupancy between the two Figure 24. Relative breeding abundance of short-eared owl among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. most recent Bird Atlases, the Scottish short-eared owl population appears to have undergone a decrease of around 8% since the estimate of Gibbons et al. (1993). However, the influence this would have on NHZ estimates is small in the context of the wide confidence intervals around these. The NHZ estimates in Table 23 are therefore calculated by apportioning the 1993 estimate between the NHZs, on the basis of Atlas-derived proportional abundance scores. Given both the incompleteness of our understanding about this species population and ecology, and also the propensity for breeding numbers in many areas to be highly variable, these figures should be treated only as loosely indicative. 56

61 Table 23: Estimated number of breeding pairs of short-eared owls in 21 NHZs. These are based on the GB population estimate reported in Musgrove et al. (2013) (which, in turn, is based on estimates made in 1990) and Atlas-derived relative proportional abundances. NHZ Central estimate Lower estimate Upper estimate 1. Shetland Orkney and North Caithness Coll, Tiree and the Western Isles North West Seaboard Peatlands of Caithness and Sutherland Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands L. Lomond, Trossachs & Breadalbane Eastern Lowlands West Central Belt Wigtown Machars & Outer Solway W. Southern Uplands & Inner Solway Border Hills Moray Firth TOTAL 1, ,852 57

62 3.24 Arctic skua (Stercorarius parasiticus) - breeding Assessments of breeding skua numbers in Seabird 2000 (S2K) were based on counts of apparently occupied territories (AOTs). This was primarily to accommodate evidence of single birds apparently on territory in suitable habitat, as well other types of evidence more directly linked to breeding such as nest defence, eggs and chicks, pair behaviour etc. For each NHZ, an 8-16 year population trend applying to changes in numbers between the periods and was calculated from Seabird Monitoring Programme data. Trends could be estimated directly for NHZs 1 and 2, which respectively had 6 and 27 colonies with surveys in both time periods. NHZs 3, 4, 5 and 14 all had between 0 and 2 colonies that received repeat surveys, so a combined trend was estimated for all of these colonies. Combined trends were based on data from NHZs that were as close a biogeographic match as possible. Nevertheless, the risk that these Figure 25. Relative breeding abundance of Arctic skua among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. trends will not be representative of changes in the NHZ populations to which they were applied is relatively high. Differences between the original S2K estimates and the current estimates for these NHZs should therefore be treated with caution.. 58

63 Table 24: Estimated number breeding pairs of Arctic skua in 21 NHZs. These are based on estimates from a nearcomprehensive national survey around 2000 (Seabird 2000 or S2K), which has been updated using NHZ-specific trend information derived from Seabird Monitoring Programme data (see section for more details). The number of colonies in each NHZ that could be used to calculate trends is reported in the Trend Colonies column, along with average number of breeding pairs in these colonies in in brackets. NHZ Current S2K Trend Trend Colonies Estimate Estimate (%) 1. Shetland 516 1,120 6( 214) Orkney and North Caithness (869) Coll, Tiree and the Western Isles (1) -57 a 4. North West Seaboard (39) -57 a 5. Peatlands of Caithness and Sutherland (8) -57 a 6. Western Seaboard Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands a 15. L. Lomond, Trossachs & Breadalbane Eastern Lowlands West Central Belt Wigtown Machars & Outer Solway W. Southern Uplands & Inner Solway Border Hills Moray Firth 0 0 TOTAL 1,027 2, (1,131) a. Data for colonies from NHZs 3, 4, and 5 contributed to the trend for NHZs 3, 4, 5 and

64 3.25 Great skua (Stercorarius skua) - breeding Assessments of breeding skua numbers in Seabird 2000 (S2K) were based on counts of apparently occupied territories (AOTs). This was primarily to accommodate evidence of single birds apparently on territory in suitable habitat, as well other types of evidence more directly linked to breeding such as nest defence, eggs and chicks, pair behaviour etc. For each NHZ, an 8-16 year population trend applying to changes in numbers between the periods and was calculated from Seabird Monitoring Programme data. Trends could be estimated directly for NHZs 1, 2 and 3, which respectively had 4, 24 and 5 colonies with surveys in both time periods. However, trend estimates for NHZs that are based on a relatively small number of colonies may be biased if these are not representative of the entire NHZ population. For example, NHZ 3 is dominated by colonies in St Kilda, and is Figure 26. Relative breeding abundance of great skua among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. therefore unlikely to be representative of population change in the entire NHZ (Andrew Stevenson pers. comm.). NHZs 4, 5 and 6 all had between 0 and 3 colonies that received repeat surveys, so combined trends were estimated for these NHZs (see footnotes below Table 25 for details). Combined trends were based on data from NHZs that were as close a biogeographic match as possible. Nevertheless, the risk that these trends will not be representative of changes in the NHZ populations they were applied to is relatively high. Differences between the original S2K estimates and the current estimates for these NHZs should therefore be treated with caution. 60

65 Table 25: Estimated number of breeding pairs of great skua in 21 NHZs. These are based on estimates from a nearcomprehensive national survey around 2000 (Seabird 2000 or S2K), which has been updated using NHZ-specific trend information derived from Seabird Monitoring Programme data (see section for more details). The number of colonies in each NHZ that could be used to calculate trends is reported in the Trend Colonies column, along with average number of breeding pairs in these colonies in in brackets. NHZ Current S2K Trend Trend Estimate Estimate Colonies (%) 1. Shetland 10,377 6,846 4 (1,319) Orkney and North Caithness 1,868 2, (2,258) Coll, Tiree and the Western Isles (283) North West Seaboard (198) +1 a 5. Peatlands of Caithness and Sutherland b 6. Western Seaboard (5) +1 a 7. Northern Highlands Western Highlands North East Coastal Plain Central Highlands Cairngorm Massif North East Glens East Lochaber Argyll West and Islands 6 c L. Lomond, Trossachs & Breadalbane Eastern Lowlands West Central Belt Wigtown Machars & Outer Solway W. Southern Uplands & Inner Solway Border Hills Moray Firth 0 0 TOTAL 12,832 9,634 a. Data for colonies from NHZs 3, 4, and 6 contributed to the trend for NHZs 4 and 6. b. Data for colonies from NHZs 2 and 4 contributed to the trend for NHZ 5. c. Approximate estimate taken from Rheniallt et al. (2007), who suggest between a small number of pairs breeding in Argyll since Seabird

66 3.26 Lesser black-backed gull (Larus fuscus) - breeding Assessments of breeding gull numbers in Seabird 2000 (S2K) were based either on counts of nests or incubating adults or on spacing of individuals or pairs viewed from a vantage point. Coverage of coastal breeding areas in S2K was much better than that of inland colonies. Particularly for NHZs with large urban areas, this is likely to have resulted in underestimates of gull populations (see section for more detail). For each NHZ with a non-zero population estimate in S2K, an 8-16 year population trend applying to changes in numbers between the periods and was calculated from Seabird Monitoring Programme data. Trends were estimated directly for NHZs 2, 6, 14 and 16, which NHZs had between 6 and 16 colonies with surveys in both time periods. A trend was also estimated directly for NHZ 1, despite its having only 3 colonies with repeat visits, because the unique Figure 27. Relative breeding abundance of lesser blackbacked gull among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. biogeographical situation of Shetland renders it poorly suited to deriving trend information from other NHZs. NHZs 3, 4, 5, 7, 8, 9, 10, 11, 12, 13, 15, 17, 18, 19, 20 and 21 all had between 1 and 3 colonies that received repeat surveys (see the footnotes below Table 26 for details of how combined trends for these NHZs were derived). Combined trends were based on data from NHZs that were as close a biogeographic match as possible. Nevertheless, the risk that these trends will not be representative of changes in the NHZ populations they were applied to is relatively high. Differences between the original S2K estimates and the current estimates for these NHZs should therefore be treated with caution. 62

67 Table 26: Estimated number of breeding pairs of lesser black-backed gull in 21 NHZs. These are based on estimates from a near-comprehensive national survey around 2000 (Seabird 2000 or S2K), which has been updated using NHZ-specific trend information derived from Seabird Monitoring Programme data (see section for more details). The number of colonies in each NHZ that could be used to calculate trends is reported in the Trend Colonies column, along with average number of breeding pairs in these colonies in in brackets. NHZ Current S2K Trend Trend Estimate Estimate Colonies (%) 1. Shetland (48) Orkney and North Caithness 1,158 1,044 6 (483) Coll, Tiree and the Western Isles 547 1,191 2 (12) -54 a 4. North West Seaboard (4) -56 b 5. Peatlands of Caithness and Sutherland c 6. Western Seaboard (200) Northern Highlands d 8. Western Highlands (6) e 9. North East Coastal Plain (2) -83 f 10. Central Highlands d 11. Cairngorm Massif d 12. North East Glens d 13. East Lochaber e 14. Argyll West and Islands 1,577 3, (1,327) L. Lomond, Trossachs & Breadalbane (33) 15 g 16. Eastern Lowlands 11,662 8, (7,895) West Central Belt 7,681 8,258 2 (3,677) -7 h 18. Wigtown Machars & Outer Solway i 19. W. Southern Uplands & Inner Solway 1,048 1, i 20. Border Hills i 21. Moray Firth f TOTAL 24,457 25, (13,689) a. Data for colonies from NHZs 3 and 6 contributed to the trend for NHZ 3. b. Data for colonies from NHZs 4, 6 and 8 contributed to the trend for NHZ 4. c. Data for colonies from NHZs 2, 4 and 21 contributed to the trend for NHZ 5. d. Data for colonies from NHZs 8, 13, 15 and 21 contributed to the trend for NHZs 7, 10, 11 and 12. e. Populations for NHZs 8 and 13 were taken as the average breeding pair number recorded by the SMP in these NHZs in f. Data for colonies from NHZs 9, 16 and 21 contributed to the trend for NHZs 9 and 21. g. Data for colonies from NHZs 13, 14, 15 and 16 contributed to the trend for NHZ 15. h. Data for colonies from NHZs 14, 16 and 17 contributed to the trend for NHZ 17 i. Data for colonies from NHZs 16 and 17 contributed to the trend for NHZs 18, 19 and 20 63

68 3.27 Herring gull (Larus argentatus) - breeding Assessments of breeding gull numbers in Seabird 2000 (S2K) were based either on counts of nests or incubating adults or, where this wasn t feasible, on spacing of individuals or pairs viewed from a vantage point. Coverage of coastal breeding areas in S2K was much better than that of inland colonies. Particularly for NHZs with large urban areas, this is likely to have resulted in underestimates of gull populations (see section for more detail). For each NHZ with a non-zero population estimate in S2K, an 8-16 year population trend applying to changes in numbers between the periods and was calculated from Seabird Monitoring Programme data. Trends were estimated directly for NHZs 2, 6, 8, 14 and 16, which had between 11 and 46 colonies with surveys in both time periods. A trend was also estimated directly for NHZ 1, despite its having only 3 colonies with repeat visits, because the unique biogeographical situation of Figure 28. Relative breeding abundance of herring gull among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. Shetland makes it poorly suited to deriving trend information from other NHZs. NHZs 3, 4, 5, 7, 9, 10, 11, 12, 13, 15, 17, 18, 19, and 21 all had 0-4 colonies that received repeat surveys. See footnotes below Table 27 for details of how combined trends for these NHZs were derived. Combined trends were based on data from NHZs that were as close a biogeographic match as possible. Nevertheless, the risk that these trends will not be representative of changes in the NHZ populations they were applied to is relatively high. Differences between the original S2K estimates and the current estimates for these NHZs should therefore be treated with caution. 64

69 Table 27: Estimated number of breeding pairs of herring gull in 21 NHZs. These are based on estimates from a nearcomprehensive national survey around 2000 (Seabird 2000 or S2K), which has been updated using NHZ-specific trend information derived from Seabird Monitoring Programme data (see section for more details). The number of colonies in each NHZ that could be used to calculate trends is reported in the Trend Colonies column, along with average number of breeding pairs in these colonies in in brackets. NHZ Current S2K Trend Trend Estimate Estimate Colonies (%) 1. Shetland 2,526 4,016 3 (98) Orkney and North Caithness 3,455 2, (306) Coll, Tiree and the Western Isles 1,251 3,975 2 (28) -69 a 4. North West Seaboard 458 1,058 4 (97) -57 b 5. Peatlands of Caithness and Sutherland 2,962 2, c 6. Western Seaboard 1,278 4, (1,946) Northern Highlands d 8. Western Highlands (593) North East Coastal Plain 7,136 11,669 3 (450) -39 e 10. Central Highlands d 11. Cairngorm Massif d 12. North East Glens d 13. East Lochaber (14) -37 f 14. Argyll West and Islands 9,372 15, (8,977) L. Lomond, Trossachs & Breadalbane (42) -23 g 16. Eastern Lowlands 16,782 17, (9,775) West Central Belt 2,928 3,886 1 (1,500) -25 h 18. Wigtown Machars & Outer Solway i 19. W. Southern Uplands & Inner Solway 1,130 1, i 20. Border Hills Moray Firth 1,490 2,437 2 (142) -39 e TOTAL 52,089 72, (23,968) a. Data for colonies from NHZs 3 and 6 contributed to the trend for NHZ 3. b. Data for colonies from NHZs 4, 6 and 8 contributed to the trend for NHZ 4. c. Data for colonies from NHZs 2, 4 and 21 contributed to the trend for NHZ 5. d. Data for colonies from NHZs 8, 13, 15 and 21 contributed to the trend for NHZs 7, 10, 11 and 12. e. Data for colonies from NHZs 9 and 21 contributed to the trend for NHZs 9 and 21. f. Data for colonies from NHZs 8, 13 and 14 contributed to the trend for NHZ 8. g. Data for colonies from NHZs 13, 14, 15 and 16 contributed to the trend for NHZ 15. h. Data for colonies from NHZs 14, 16 and 17 contributed to the trend for NHZ 10. i. Data for colonies from NHZs 16 and 17 contributed to the trend for NHZs 18 and

70 3.28 Great black-backed gull (Larus marinus) - breeding Assessments of breeding gull numbers in Seabird 2000 (S2K) were based either on counts of nests or incubating adults or, where this wasn t feasible, on spacing of individuals or pairs viewed from a vantage point. Coverage of coastal breeding areas in S2K was much better than that of inland colonies. Particularly for NHZs with large urban areas, this is likely to have resulted in underestimates of gull populations (see Skuas and gulls for more detail). For each NHZ with a non-zero population estimate in S2K, an 8-16 year population trend applying to changes in numbers between the periods and was calculated from Seabird Monitoring Programme data. Trends could be estimated directly for NHZs 1, 2, 3, 5, 6, 8, 14 and 16, which NHZs had between 5 and 52 colonies with surveys in both time periods. NHZs 4, 7, 9, 12, 13, 17, 18, 19 and 21 all had between 0 and 4 colonies Figure 29. Relative breeding abundance of great blackbacked gull among 21 Scottish NHZs. Shading indicates estimated breeding abundance as a proportion of the whole breeding population. that received repeat surveys. See footnotes below Table 28 for details of how combined trends for these NHZs were derived. Combined trends were based on data from NHZs that were as close a biogeographic match as possible. Nevertheless, the risk that these trends will not be representative of changes in the NHZ populations they were applied to is relatively high. Differences between the original S2K estimates and the current estimates for these NHZs should therefore be treated with caution. 66

71 Table 28: Estimated number of breeding pairs of great black-backed gull in 21 NHZs. These are based on estimates from a near-comprehensive national survey around 2000 (Seabird 2000 or S2K), which has been updated using NHZ-specific trend information derived from Seabird Monitoring Programme data (see section for more details). The number of colonies in each NHZ that could be used to calculate trends is reported in the Trend Colonies column, along with average number of breeding pairs in these colonies in in brackets. NHZ Current S2K Trend Trend Estimate Estimate Colonies (%) 1. Shetland 2,106 2,856 4 (101) Orkney and North Caithness 1,712 5, (1,847) Coll, Tiree and the Western Isles 734 2,351 5 (373) -69 a 4. North West Seaboard (447) -63 b 5. Peatlands of Caithness and Sutherland (100) Western Seaboard (308) Northern Highlands c 8. Western Highlands (52) North East Coastal Plain (6) 87d 10. Central Highlands Cairngorm Massif North East Glens East Lochaber e 14. Argyll West and Islands 679 1, (692) L. Lomond, Trossachs & Breadalbane Eastern Lowlands (38) West Central Belt (246) -29 f 18. Wigtown Machars & Outer Solway g 19. W. Southern Uplands & Inner Solway g 20. Border Hills Moray Firth h TOTAL 6,820 14, (4,210) a. Data for colonies from NHZs 3 and 6 contributed to the trend for NHZ 3. b. Data for colonies from NHZs 4, 6 and 8 contributed to the trend for NHZ 4. c. Data for colonies from NHZs 8 contributed to the trend for NHZ 7. d. Data for colonies from NHZs 9 and 16 contributed to the trend for NHZs 9. e. Data for colonies from NHZs 14 contributed to the trend for NHZs 13. f. Data for colonies from NHZs 14, 16 and 17 contributed to the trend for NHZs 17. g. Data for colonies from NHZs 16 and 17 contributed to the trend for NHZs 18 and 19. h. Data for colonies from NHZs 5 and 9 contributed to the trend for NHZ

72 4. References 1. Austin, G.E., Calbrade, N.A., Mellan, H.J., Musgrove, A.J., Hearn, R.D., Stroud, D.A., Wotton, S.R., Holt, C.A. (2014). Waterbirds in the UK 2012/13: The Wetland Bird Survey. BTO/RSPB/JNCC. Thetford Balmer, D. E., Gillings, S., Caffrey, B. J., Swann, R. L., Downie, I. S., Fuller, R. J. (2013). Bird Atlas : the breeding and wintering birds of Britain and Ireland. BTO Books, Thetford. 3. Banks, A. N., H. Q. P. Crick, R. Coombes, S. Benn, D. A. Ratcliffe, E. M. Humphreys The breeding status of Peregrine Falcons Falco peregrinus in the UK and Isle of Man in Bird Study Bellamy, P. E., Eaton, M. A. (2009). CSM bird monitoring of Caithness and Sutherland Peatlands SPA. Report to SNH by RSPB, Sandy. 5. Calladine, J., Pakeman, R. J., Humphreys, E., Huband, S., Fuller, R. J. (2014). Changes in breeding wader assemblages, vegetation and land use within machair environments over three decades. Bird Study Challis, A., Holling, M., Stevenson, A., Roos, S., Stirling-Aird, P., Wilson, M. (2014). Scottish Raptor Monitoring Scheme Report BTO Scotland/SRMS, Stirling. 7. Dillon, I. A., Smith T. D., Williams S. J., Haysom S., Eaton M. A. (2009). Status of Red-throated Divers Gavia stellata in Britain in Bird Study Eaton, M.A., Dillon, I.A., Stirling-Aird, P., Whitfield, D.P. (2007). The status of the Golden Eagle Aquila chrysaetos in Britain. Bird Study Ewing, S.R., Rebecca, G.W., Heavisides, A., Court, I.R., Lindley, P., Ruddock, M., Cohen, S., Eaton, M.A. (2011). Breeding status of Merlins Falco columbarius in the UK in Bird Study Fielding, A., Haworth, P., Whitfield, P., McLeod, D., Riley, H. (2011). A conservation framework for Hen harriers in the United Kingdom. JNCC report 441. JNCC, Peterborough. 11. Forrester, R. W., Andrews, I. J., McInerny, C. J., Murray, R. D., McGowan, R. Y., Zonfrillo, B., Betts, M. W., Jardine, D. C., Grundy, D. S. (2007). The Birds of Scotland. Scottish Ornithologists' Club. Aberlady. 12. Fox, T., Francis, I., Norriss, D., Walsh, A. (2014). Report of the 2013/14 International Census of Greenland White-fronted Geese. Greenland White-fronted Goose Study, Aarhus University, and National Parks and Wildlife Service, Wexford. 13. Fuller, R. J., Humphreys, E. M., Wilson, J. D., Hoccom, D. G., Calladine, J. (2010). Changes in the breeding wader populations of the machair of the Western Isles, Scotland, between 2000 and Bird Study Gibbons, D. W., Reid, J. B., Chapman, R. A. (1993). The New Atlas of Breeding Birds in Britain and Ireland: T & AD Poyser, London, for the British Trust for Ornithology, Scottish Ornithologists' Club, Irish Wildbird Conservancy. 15. Greenwood, J. J. D., Crick, H. Q. P., Bainbridge, I. P. (2003). Numbers and international importance of raptors and owls in Britain and Ireland, in: Thompson, D. B. A., Redpath, S. M., Fielding, A. H., Marquiss, M., Galbraith, C. A. (Eds.), Birds of Prey in a Changing Environment. The Stationary Office, Edinburgh. 68

73 16. Hall, C., J.R. Glanville, H. Boland, Ó. Einarsson, G. McElwaine, C.A. Holt, C.J. Spray, E.C. Rees. (2012). Population size and breeding success of Icelandic Whooper Swans Cygnus cygnus: results of the 2010 international census. Wildfowl Hancock, M.H., Gibbons, D.W., Thompson, P.S. (1997). The status of breeding Greenshank Tringa nebularia in the United Kingdom in Bird Study Harris, S. J., Risely, K., Massimino, D., Newson, S. E., Eaton, M. A., Musgrove, A. J., Noble, D. G., Procter, D., Baillie, S. R. (2014). The Breeding Bird Survey BTO Research Report 658. British Trust for Ornithology, Thetford. 19. Haworth, P., Fielding, A. (2011). The Identification of Important Localities for Breeding Merlins in Scotland. Report to SNH by Haworth Conservation, Mull. 20. Hayhow, D. B., Eaton, M. A., Bladwell, S., Etheridge, B., Ewing, S. R., Ruddock, M., Saunders, R., Sharpe, C., Sim, I. M. W., Stevenson, A. (2013). The status of the Hen Harrier, Circus cyaneus, in the UK and Isle of Man in Bird Study Jackson, D.B. (2001). Wader surveys: provisional estimates and recent trends, Outer Hebrides. Western Isles Bird Report Langholm Moor Demonstration Project. (2014). The Langholm Moor Demonstration Project: seven year review December Langholm Moor Demonstration Project Ltd., Langholm. 22. Massimino, D., Johnston, A., Pearce-Higgins, J.W. (2011). Producing Regional Population Estimates for Upland Wader. BTO Research Report 58. BTO, Thetford. 23. Mitchell, C., C. Hall (2013). Greenland Barnacle Geese Branta leucopsis in Britain and Ireland: results of the international census, spring Wildfowl & Wetlands Trust, Slimbridge. 24. Mitchell, C. (2014). Status and distribution of Icelandic-breeding geese: results of the 2013 international census. Wildfowl & Wetlands Trust Report, Slimbridge. 25. Mitchell, P.I., Newton, S.F., Ratcliffe, N., Dunn, T.E. (2004). Seabird Populations of Britain and Ireland. T & AD Poyser, London. 26. Murray, R.D., Holling, M., Dott, H.E.M., van Dome, P. (1998). The Breeding Birds of South- East Scotland. Scottish Ornithologists Club, Edinburgh. 27. Musgrove, A., Aebischer, N., Eaton, M., Hearn, R., Newson, S., Noble, D., Parsons, M., Risely, K., Stroud, D. (2013). Population estimates of birds in Great Britain and the United Kingdom. British Birds O Brien, M., Tharme, A., Jackson, D. (2002). Changes in breeding wader numbers on Scottish farmed land during the 1990s. Scottish Birds O Brien, M., White, C. S. (2003). Estimating the breeding wader populations of Scottish uplands and salt marshes. Scottish Birds Pennington, M.G., Osborn, K., Harvey, P.V., Riddington, R., Okill, J.D., Ellis, P.M., Heubeck, M. (2004). The Birds of Shetland. Christopher Helm, London. 31. Powell, L.A. (2007). Approximating variance of demographic parameters using the delta method: a reference for avian biologists. The Condor Rae, S. (2010). Density and productivity of ground-nesting Merlins on an island with no indigenous terrestrial predators. Scottish Birds Redpath, S. M., Thirgood, S. J. (2007). Birds of prey and red grouse. Centre for Ecology and Hydrology, Banchory. 69

74 34. Sellers, R. M., Shackleton, D. (2011). Numbers, distribution and population trends of large gulls breeding in Cumbria, northwest England. Seabird Sim, I. M. W., Gregory, R. D., Hancock, M. H., Brown, A. F. (2005). Recent changes in the abundance of British upland breeding birds. Bird Study Sim, I. M. W., Eaton, M. A., Setchfield, R. P., Warren, P. K., Lindley, P. (2008). Abundance of male Black Grouse Tetrao tetrix in Britain in 2005, and change since Bird Study SNH (2006). Assessing significance of impacts from onshore windfarms on birds outwith designated areas. SNH Guidance Document. 20pp. 38. Whitfield, D. P., Fielding, A. H., McLeod, D. R. A., Haworth, P. F., Watson, J. (2006). A conservation framework for the golden eagle in Scotland: refining condition targets and assessment of constraint influence. Biological Conservation Wilson, M. W., Wernham, C. V., Balmer, D. E., Noble, D. (2015). Peregrine Survey Draft Report to SNH. BTO Scotland, Stirling. 5. Acknowledgements Many thanks to the following for provision of datasets and accompanying advice: Roddy Mavor at JNCC, David Wood and Andrew Stevenson at SNH, Richard Evans and Tessa Cole at RSPB, the Scottish Raptor Monitoring Scheme and members of the Scottish Raptor Monitoring Group, and the Scottish Raptor Study Group. Ali Johnston and Mark Miller at BTO provided invaluable and patiently delivered guidance and R code relating to confidence intervals. John Calladine, Liz Humphreys, Andy Musgrove and James Pearce-Higgins at BTO, Phil Warren at Game and Wildlife Conservancy Trust, Digger Jackson at Natural Research, Paul Bellamy, David Douglas, Mark Eaton, Alan Leitch and Staffan Roos at RSPB, and Amy Challis at SRMS all contributed significantly through discussion and comments. The report also benefitted from the helpful reviewing of the Scottish Windfarm Bird Steering Group Research Group representatives. 70

75 6. Appendices 71

76 Appendix A: Monthly WeBS counts by species and NHZ Figure A1. Five-year ( ) means of monthly total WeBS counts of whooper swan in each NHZ. Pale green shaded areas around the mean delineate the range (minimum to maximum) of count values during this period. 72

77 Figure A2. Five-year ( ) means of monthly total WeBS counts of Pink-footed goose in each NHZ. Pale green shaded areas around the mean delineate the range (minimum to maximum) of count values during this period. 73