9. Appendix 1. Review of sample images used by APEM for marine mammal and seabird identification from digital aerial surveys off East Scotland
The aim of this exercise was to review a selection of digital images to check species identifications. The motivation for this was as a validation procedure for species occurring out of their normal range / habitat, and to better understand the reasons why there was large variation in the proportions of sightings ascribed to species (e.g. harbour porpoise) vs species group (porpoise/dolphin). Taking that example, the first two aerial surveys had no porpoises recorded but 51 and 13 porpoise/dolphin species groups respectively. Although there were several instances where we would recommend a change in ID, we did no updates for the analysis. This would have necessitated a complete new evaluation involving going through many thousands of images.
Species out of normal habitat Three cases were examined: 1) a black guillemot recorded far offshore in the northern North Sea on 6 Mar 2021 (@12:21). This species is normally very coastal; 2) a red-throated diver also recorded far offshore on 17 Sep 2020 and 3) a red-throated diver recorded far offshore on 29 Oct 2020. This species tends to occupy shallow waters, estuaries and bays. In all three cases, the species identification could be correct although we would have been more cautious for the latter two, and assigned them as diver sp. we checked also black guillemot images from more coastal areas, and agreed that they were of this species.
Species out of normal range Images identified as common dolphin were examined for 4 Mar 2020, 14 Apr 2020, 2 Jun 2020 (multiple sightings), 8 Jun 2020, and 24 Jun 2020. This species is rare in the North Sea but occasionally enters the northern sector (presumably from the Atlantic – it occurs regularly in the Hebrides). Images from each of those surveys were examined. Very few images were unambiguously common dolphin, but most were probably so. Of 22 images, we agreed with ten, disagreed with three, and were uncertain for nine.
Another species that is rare in the North Sea (south of Caithness) is the Risso’s dolphin. There was one record of five animals some distance offshore due east of the Firth of Forth on 2 June 2020. On checking, we confirmed these as Risso’s dolphins.
White-beaked Dolphin vs Atlantic White-sided Dolphin The main dolphin species inhabiting the North Sea, particularly offshore, is the white-beaked dolphin. However, in northern Scotland (Caithness to the Northern Isles), in recent years white-beaked dolphins have become scarce and the main dolphin species recorded there is Atlantic white-sided dolphin. In the APEM data set, no Atlantic white-sided dolphins are recorded. These two species can be confused although they should be easier from the air due to the conspicuous white area over the back behind the dorsal fin in the white-beaked dolphin. Atlantic white-sided dolphins tend to occur in large groups (numbering 25 to a few hundred) whereas white-beaked dolphin average group size is <10. Checking the larger groups recorded as white-beaked dolphins, all images were confirmed as that species, and no definite white-sided dolphins were identified.
Harbour Porpoise vs Porpoise/Dolphin There are thousands of images in one or other of these two categories. We therefore concentrated upon one of the first two surveys, on 2 June 2020. A total of 242 images assigned to Porpoise/Dolphin were reviewed. Of those, two images were identified as white-beaked dolphins, 122 were classified as harbour porpoise, and 118 were of indeterminate species. Condition of light and sea state were no worse on those two first surveys compared with subsequent ones, suggesting that perhaps the person going through the images had not yet got their eye in to differentiating porpoises (NB they have a characteristic body shape: blunt head, rotund body tapering to a narrow tail stock but relatively wide tail fluke). When animals are below a certain depth in the water or visibility underwater is poor, it is not possible to be confident in determining body shape because it becomes distorted. This will always account for a proportion of animals. It is quite likely that the great majority of the 118 images unassigned to species, were in fact porpoises but we cannot be sure.
Other species Some random checks were made on other species / species groups (e.g. minke whale, whale sp., marine mammal sp., small gull sp., razorbill/guillemot, razorbill, guillemot). There were no cases where there was disagreement on the identification but there were some which would be better downgraded to a species group. For other species, there were none from within a species group that could confidently be assigned to a species. Some of the whale sp. looked as if they were minke whales and one appeared to be a possible fin whale but without knowing the sizes, it was impossible to confirm.
There was general agreement over the species identification assigned to the images reviewed. In some cases, however, more caution would have been applied leading to downgrading of species ID. On the other hand, for porpoise/dolphin, one could more confidently separate porpoises, at least for the early surveys and probably to a lesser extent for all surveys. Clearly, this project could not go through everything again in that latter category, re-assign those, and re-calculate numbers and densities. One option might be to consider all porpoises/dolphins as porpoises and use that as the upper bound for estimates since the great majority of encounters will be of that species.
There is a problem
Thanks for your feedback