Production of Seabird and Marine Mammal Distribution Models for the East of Scotland

2. List of figures and tables

Figure 1. The graph shows flown transects for each surveyed month between by month February 2020 to March 2021

Figure 2. The graph showing variation in depths (in metres) across the survey region. The study area is deepest in the central and northern part.

Figure 3. Total area covered (in km²) by (left) surveyed year (2020 and 2021) and (right) by calendar month (numbered, Jan. = 1 etc.), including months when survey did not occur. March had the largest cover but January the lowest but note that March was included in 2020 and 2021 survey. N.B no data for May, August or December were recorded.

Figure 4. A graph showing estimated numbers of northern fulmars over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to July) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of fulmars peaked in September and had lowest values in winter months (January to April).

Figure 5. A graph showing point estimates of northern fulmar densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of birds with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 6. A graph showing lower confidence bound estimates (2.5%) of northern fulmar densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 7. A graph showing upper confidence bound estimates (97.5%) of northern fulmar densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 8. A graph showing coefficients of variation (CV, in %) in estimated densities of northern fulmars for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at the peripheries of the study area, especially in the south.

Figure 9. A graph showing mean fulmar density surfaces for breeding (April – August) and non-breeding (September – March) seasons.

Figure 10. Graphs showing effect of (left) monthly sea surface temperature and (right) mean monthly salinity range on northern fulmar observed density assuming the middle of the survey area during the breeding season.

Figure 11. Graphs showing effect of (upper left) monthly sea surface temperature, (upper right) mean monthly salinity and (lower left) mean monthly salinity range on northern fulmar observed density assuming in the middle of survey area outside of the breeding season.

Figure 12. A graph showing estimated numbers of northern gannets over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to October) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of gannets peaked in July and September and had lowest values in winter months (January to April).

Figure 13. A graph showing point estimates of northern gannet densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of birds with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 14. A graph showing lower confidence bound estimates (2.5%) of northern gannet densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 15. A graph showing upper confidence bound estimates (97.5%) of northern gannet densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 16. A graph showing northern gannet coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at the eastern and western part of the study area.

Figure 17. A graph showing mean northern gannet density (birds/km²) surfaces for breeding (April – October) and non-breeding (November – March) seasons.

Figure 18. A graph showing effectof mean monthly sea surface temperature on northern gannet observed density assuming the middle of the survey area during the breeding season.

Figure 19. A graph showing estimated numbers of great skuas over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to July) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of great skuas peaked in June and had lowest values in late autumn and winter months (November to March). The results of this model should be treated with caution as not all model assumptions were met.

Figure 20. A graph showing point estimates of great skua densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of birds with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals. The results of this model should be treated with caution as not all model assumptions were met.

Figure 21. A graph showing lower confidence bound estimates (2.5%) of great skua densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. The results of this model should be treated with caution as not all model assumptions were met.

Figure 22. A graph showing upper confidence bound estimates (97.5%) of great skua densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. The results of this model should be treated with caution as not all model assumptions were met.

Figure 23. A graph showing great skua coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at the peripheries of he study area outside the breeding season and in the southern art during the breeding season. The results of this model should be treated with caution as not all model assumptions were met.

Figure 24. A graph showing mean great skua density (birds/km²) surfaces for breeding (April – July) and non-breeding (August – March) seasons. The results of this model should be treated with caution as not all model assumptions were met.

Figure 25. A graph showing effect of ( mean monthly sea surface temperature range in (red) and out (black) the breeding season on great skua observed density assuming the middle of the survey area during respective season.

Figure 26. A graph showing estimated numbers of common gulls over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to July) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of common gulls was lowest during the breeding season.

Figure 27. A graph showing point estimates of common gull densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of birds with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals. As the spatial pattern in density was consistent and uniform outside the breeding season, the graphs show mean estimates for non-breeding season for each surveyed month within this season.

Figure 28. A graph showing lower confidence bound estimates (2.5%) of common gull densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. As the spatial pattern in density was consistent and uniform outside the breeding season, the graphs show mean estimates for non-breeding season for each surveyed month within this season.

Figure 29. A graph showing upper confidence bound estimates (97.5%) of common gull densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. As the spatial pattern in density was consistent and uniform outside the breeding season, the graphs show mean estimates for non-breeding season for each surveyed month within this season.

Figure 30. A graph showing common gull coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. As the spatial pattern in density was consistent and uniform outside the breeding season, the graphs show mean estimates for non-breeding season for each surveyed month within this season.

Figure 31. A graph showing the effect of depth on common gull observed densities assuming the middle of survey area. The effect of depth is estimated breeding season only.

Figure 32. A graph showing estimated numbers of herring gull over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to July) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of herring gulls peaked in November.

Figure 33. A graph showing point estimates of herring gull densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of birds with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 34. A graph showing lower confidence bound estimates (2.5%) of herring gull densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 35. A graph showing upper confidence bound estimates (97.5%) of herring gull densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 36. A graph showing herring gull coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at the peripheries of the study area.

Figure 37. A graph showing mean herring gull density (birds/km²) surfaces for breeding (April – July) and non-breeding (August – March) seasons.

Figure 38. Graphs showing effect of day of year (upper left), depth (upper right) and mean monthly sea surface temperature range (lower left) on herring gull density assuming the middle of survey area outside of the breeding season. The effect of depth differs dependent on whether it is the breeding (red) or non-breeding period (black).

Figure 39. A graph showing estimated numbers of great black-backed gulls over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to July) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of gulls peaked in November.

Figure 40. A graph showing point estimates of great black-backed densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of birds with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 41. A graph showing lower confidence bound estimates (2.5%) of great black-backed gull densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 42. A graph showing upper confidence bound estimates (97.5%) of great black-backed gull densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 43. A graph showing great black-backed gull coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at southern part of the study area during non-breeding season and at the centre during the breeding season.

Figure 44, A graph showing mean great black-backed gull density (birds/km²) surfaces for breeding (April – July) and non-breeding (August – March) seasons.

Figure 45. Graphs showing effect of (left) monthly range of salinity and (right)seabed roughness on great black- backed gull observed density assuming the middle of survey area in the breeding season.

Figure 46. Graphs showing effect of (left) mean monthly sea surface temperature and (right) depth on great black-backed gull observed density assuming the middle of survey area outside of the breeding season.

Figure 47. A graph showing estimated numbers of kittiwakes over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to July) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of kittiwakes peaked in June.

Figure 48. A graph showing point estimates of black-legged kittiwake densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of birds with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 49. A graph showing lower confidence bound estimates (2.5%) of black-legged kittiwake densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 50. A graph showing upper confidence bound estimates (97.5%) of black-legged kittiwake densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 51. A graph showing kittiwake coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at the peripheries of the study area.

Figure 52. A graph showing mean black-legged kittiwake density (birds/km²) surfaces for breeding (April – August) and non-breeding (September – March) seasons.

Figure 53. Graph showing effect of monthly mean salinity on black-legged kittiwake observed density assuming the middle of survey area within (red) the breeding season and (black) outside the breeding season. .

Figure 54. A graph showing estimated numbers of common guillemots over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to July) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of guillemots peaked in June-September and had lowest values throughout non-breeding season.

Figure 55. A graph showing point estimates of common guillemot densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of birds with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 56. A graph showing lower confidence bound estimates (2.5%) of common guillemot densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 57. A graph showing common guillemot coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at the north eastern part of the study area.

Figure 58. A graph showing mean common guillemot density (birds/km²) surfaces for breeding (April – July) and non-breeding (August – March) seasons.

Figure 59. Graphs showing effect of (upper left) depth, (upper right) mean monthly sea surface temperature and (lower left) seabed roughness on common guillemot observed density.

Figure 60. A graph showing estimated numbers of razorbills over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to July) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of razorbills peaked in June and had lowest values in winter months (January to March).

Figure 61. A graph showing point estimates of razorbill densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of razorbill with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 62. A graph showing lower confidence bound estimates (2.5%) of razorbill densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 63. A graph showing upper confidence bound estimates (97.5%) of razorbill densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 64. A graph showing razorbill coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at the northern and eastern part the study area.

Figure 65. A graph showing mean razorbill density (birds/km²) surfaces for breeding (April – July) and non-breeding (August – March) seasons.

Figure 66. Graph showing effect of mean monthly sea surface temperature range on razorbill observed density assuming the middle of the survey area during the breeding season.

Figure 67. Graphs showing effect of (upper left) monthly sea surface temperature,(upper right) mean monthly salinity and (lower left) mean monthly salinity range on razorbill density assuming the middle of the survey area outside of the breeding season.

Figure 68. A graph showing estimated numbers of puffins over the duration of the study from February 2020 to March 2021. Red points indicate the breeding season (April to July) and the dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of puffins peaked in June. High uncertainty is generated in peripheral regions in the non-breeding season.

Figure 69. A graph showing point estimates of puffin densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of fulmars with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals

Figure 70. A graph showing lower confidence bound estimates (2.5%) of puffin densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 71. A graph showing upper confidence bound estimates (97.5%) of puffin densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 72. A graph showing Atlantic puffin coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at the northern and eastern of the study area.

Figure 73. A graph showing mean puffin density (birds/km²) surfaces for breeding (April – August) and non-breeding (September – March) seasons

Figure 74. Graphs showing effect of (upper left) monthly sea surface temperature, (upper right) mean monthly SST range and (lower left) mean monthly salinity on Atlantic puffin observed density assuming themiddle of survey area during the breeding season.

Figure 75. Graph showing effect of current on Atlantic puffin observed density assuming the middle of the survey area outside of the breeding season

Figure 76. A graph showing estimated numbers of minke whales over the duration of the study from February 2020 to March 2021. Dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of minke whales peaked in June.

Figure 77. A graph showing point estimates of minke whales densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of whales with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 78. A graph showing lower confidence bound estimates (2.5%) of minke whale densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 79. A graph showing upper confidence bound estimates (97.5%) of minke whale densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 80. A graph showing locations of sightings of common dolphins. No model was fitted to the data; hence the distribution model outputs are not presented. Black lines indicate sampling locations in that month. Red dots indicate observed numbers of common dolphin with areas proportional to number.

Figure 81. A graph showing estimated numbers of white-baked dolphins over the duration of the study from February 2020 to March 2021. Dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of dolphins peaked in July.

Figure 82. A graph showing point estimates of white-beaked dolphins densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of dolphins with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals

Figure 83. A graph showing lower confidence bound estimates (2.5%) of white-beaked dolphins densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 84. A graph showing upper confidence bound estimates (97.5%) of white-beaked dolphin densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 85. A graph showing white-beaked dolphin coefficients of variation (CV, in %) in estimated densities of dolphins for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month.

Figure 86. A graph showing estimated numbers of harbour porpoises over the duration of the study from February 2020 to March 2021. The dashed lines represent upper and lower bounds of the 95% confidence intervals. Numbers of porpoises peaked between April and June.

Figure 87. A graph showing point estimates of harbour porpoise densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month. Red dots indicate observed numbers of porpoises with size proportional to observed number. Note that scale is matching the following graphs depicting lower and upper confidence intervals.

Figure 88. A graph showing lower confidence bound estimates (2.5%) of harbour porpoise densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 89. A graph showing upper confidence bound estimates (97.5%) of harbour porpoise densities for each surveyed month from February 2020 to March 2021. Colours represent estimated densities per km². Black lines indicate sampling locations in that month.

Figure 90. A graph showing harbour porpoise coefficients of variation (CV, in %) in estimated densities of birds for each surveyed month from February 2020 to March 2021. Black lines indicate sampling locations in that month. The largest CVs are at the peripheries of the study area.

Figure 91. A graph showing effect of mean monthly salinity on surface porpoise observed density assuming the middle of the survey area.

Table 1. Species recorded during the eight digital aerial surveys, Feb 2020 – Mar 2021.

Table 2. Spatial predictors for consideration in modelling.

Table 3. Bird seasons used in this analysis (following Searle et al. 2022).

Table 4. Mean surface and dive times used for individuals of target species.

Table 5. Allocation of vaguely identified animals.

Table 6. Total area covered (in km²) by each survey.

Table 7. Initial models including list of environmental covariates for each species. All covariates were fitted as continuous variables (indicated by s()).

Table 8. Selected models for seabirds.

Table 9. Selected models for marine mammals (offsets not given).