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

1. Executive Summary

The Scottish Government has an interest in monitoring the status of mobile marine megafauna especially with regard to the Marine Protected Area (MPA) network and offshore wind farm development. Knowledge of the spatial and temporal distribution and abundance of marine species is important in order for the Government to make evidence-based decisions regarding the status of these species and management of marine activities. This report describes temporal and spatial patterns of density for seabird and marine mammal species in the eastern waters of Scotland from digital aerial surveys undertaken by APEM Ltd. for Marine Scotland between February 2020 and March 2021.

Eleven bird species recorded sufficiently regularly within the surveys for modelling and abundance estimation included northern fulmar (Fulmarus glacialis), northern gannet (Morus bassanus), great skua (Stercorarius skua), common gull (Larus canus), herring gull (Larus argentatus), great black-backed gull (Larus marinus), black-legged kittiwake (Rissa tridactyla), common guillemot (Uria aalge), razorbill (Alca torda), and Atlantic puffin (Fratercula arctica). Design-based estimates were obtained for abundance only for lesser black-backed gull (Larus fuscus) where a model could not be fitted.

There are four cetacean species of particular interest in this study due to their regular occurrence: minke whale (Balaenoptera acutorostrata), common dolphin (Delphinus delphis), white beaked dolphin (Lagenorhynchus albirostris) and harbour porpoise (Phocoena phocoena). One species, common dolphin (Delphinus delphis) could not be modelled but designed based estimated were made. Other species such as killer whale (Orcinus orca), long-finned pilot whale (Globicephala melas), Atlantic white-sided dolphin (Lagenorhynchus acutus), Risso’s dolphin (Grampus griseus), and humpback whale (Megaptera novaeangliae) do occur in the region on a regular basis but were identified too few times during the digital aerial surveys for their density distributions to be modelled. The number of individuals recorded within each species or species group per survey is given in Table 1.

Species/species group identifications were checked from a sample totalling c. 300 digital images, with emphasis on species with uncertain identification, species outside their normal range, and ones of similar appearance (see Appendix 1). Since many images could not be identified to species level, whilst some others were ascribed to a particular species when image resolution in our opinion was too poor for confident identification, in order to allocate species groups to individual species, and to improve the precision of estimates for some species with low sample sizes, an updated version of the MERP (Marine Ecosystems Research Programme) database was utilised in certain cases.

General additive mixed models were fitted separately to each species of interest. The response variable for these models was the number of animals detected on the surveys. The candidate explanatory variables associated with digital photos were location and environmental features (e.g. depth, seabed roughness, sea surface temperature (both mean and range), salinity (both mean and range), and stratification index).

For those seabird species with reasonable sample sizes, density and abundance estimates generally aligned well with estimates from previous published findings of earlier survey collations, taking account of recent status changes, where applicable. Estimates of abundance and density for cetacean species, on the other hand, tend to be much larger than from the SCANS survey. The reasons for the differences are not clear but may be partly due to corrections made for availability bias in this study.

The selected models were used to estimate abundance throughout the region of interest and to derive maps of estimated densities (numbers of individuals per km2) from each survey. Overall abundance estimates, and spatial and seasonal patterns of densities were compared with previously published results from analyses of larger survey datasets. For breeding seabirds, comparisons were made with colony counts, taking account of status changes recorded since the last published census (Seabird 2000) conducted between 1998-2002.

The limitations of using only digital aerial survey data are discussed. Recommendations for future monitoring include use of visual survey data, greater survey effort east of the Northern Isles, particularly Shetland, and during some critical periods, for example April to June when circumstances prevented coverage. The analysis is also based on data collected largely in a single calendar year (2020) and so may not be applicable to the longer term. With respect to seabirds, two significant mortality events have occurred after the survey work was completed. The first of these was a large seabird wreck (affecting primarily Auk spp.) in late summer of 2021, and the second, an outbreak of highly pathogenic avian influenza during the seabird breeding season of 2022. These could lead to significant changes in the at-sea densities of some species.