This document is given in support of seal density maps provided to Scottish Government

Prepared by: M. I. D. Carter & D. J. F. Russell,
Sea Mammal Research Unit, Scottish Oceans Institute,
University of St Andrews, KY16 8LB, UK
March 2024

midc@st-andrews.ac.uk
dr60@st-andrews.ac.uk

At time of publishing, a manuscript (Carter et al) is in preparation that will feature this work. Please check the Marine Scotland data portal and search “Updated habitat-based distribution maps for harbour and grey seals in Scotland “. A link to the publication will be available on the dataset landing page once published. Please preferentially cite the paper where possible.

If no paper is listed, please cite this resource as follows:
Carter, M. I. D, Bivins, M., Duck, C. D., Hastie, G. D., Morris, C. D., Moss, S. E. W., Thompson, D., Thompson, P. M., Vincent, C., Russell, D. J. F. (2024)
Updated habitat-based distribution maps for harbour and grey seals in Scotland. Report to Scottish Government by Sea Mammal Research Unit, University of St Andrews.

For information specific to the absolute density estimates, please see:
SCOS (2021) Scientific advice on matters related to the management of seal populations. Main advice report.

#

FILES

This document relates to interpretation and use of the following files:

"Harbour_seal_absolute_density_SCOTLAND_202403.tiff"
"Harbour_seal_relative_density_SCOTLAND_202403.tiff"
"Grey_seal_absolute_density_SCOTLAND_202403.tiff"
"Grey_seal_relative_density_SCOTLAND_202403.tiff"

File projection: Universal Transverse Mercator 30°N World Geodetic System 1984 projection; UTM30NWGS84 (EPSG: 32630).

Each .tiff file contains three rasters with estimates of mean and 95% lower and upper confidence intervals for
seal distribution on a 5 km x 5 km grid covering the maximum foraging range of seals from haulouts in Scotland.

IMPORTANT NOTE

These estimates represent predicted at-sea density of seals from haulouts in Scotland only (with the exception of the St Kilda archipelago) and do not account for
density of seals from haulouts elsewhere.

#

#

INTERPRETATION

Relative density refers to the percentage of the at-sea population expected to be in a particular cell during the main foraging season.
Thus, values for the mean relative density surface sum to 100.
Absolute density reflects an estimate of the number of animals expected to be in a particular cell.

In each file, three seal density surfaces are provided: mean and associated lower and upper 95% confidence limits.

For each cell, the confidence intervals represent the range of values within which, based on
the underlying model, the true seal density is likely encompassed. The mean represents a
measure of the centre of this range.

For the mean estimates, cell values can be summed across an area. For example, to estimate the total abundance of seals in an area of interest,
it is possible to sum all cells in the mean distribution surface that fall within that area.

However, confidence interval estimates are provided on a cell-wise basis. Therefore, the sum of all cells in the lower confidence
interval surface will be less than 100% of the at-sea population total, and the sum of all cells in the upper confidence interval
surface will be greater than 100% of the at-sea population total. Cells in the confidence limit surfaces therefore CANNOT be
summed across an area, and must be treated on a cell-by-cell basis. Summing these values across an area would give inflated
confidence intervals. For area-based confidence intervals, please contact the authors on a case-by-case basis.

ABSOLUTE DENSITY

The number of animals per cell is calculated using two scalars estimated from telemetry data:
(1) the proportion of total population hauled-out during low tide in August and thus available to count,
and (2) the mean proportion of time seals spend at-sea during the main foraging season.

To generate these absolute values, the total population size in Scotland for each species was first estimated by summing
counts for each haulout (most recent available counts were used per haulout), then scaling up to 100% using the first scalar
(mean estimated proportion of the population hauled out during low tide in August).
Total population size was then converted to at-sea population size by multiplying by the second scalar
(estimated mean proportion of time spent at-sea during the main foraging season).
Relative estimates of density (% at-sea population per cell) were then converted to absolute density using the
estimated at-sea population size.

Scalars used in this process are as follows:

Scalar 1 (proportion of time spent hauled out during low tide in August):

Grey seals: 0.2515 (Russell & Carter 2021, SCOS Briefing Paper 21/02 in SCOS Report 2021)
Harbour Seals: 0.72 (Lonergan et al. 2013, DOI:10.1002/aqc.2277)

Scalar 2 (proportion of time spent at-sea during marin foraging season):

Grey seals: 0.8616 (derived from analysis in Russell et al. 2015, DOI:10.1111/oik.01810)
Harbour seals: 0.8236 (derived from analysis in Russell et al. 2015, DOI:10.1111/oik.01810)

Cells in the mean distribution surface sum to the total current at-sea population estimate for Scotland (grey seals = 86,797, harbour seals = 30,425).

#

TEMPORAL RELEVANCE

These estimates of seal density are generated using telemetry data from the main foraging season
for each species, thus the distribution maps represent seal density during the main foraging
season (grey seals: summer, harbour seals: autumn-winter-spring).

Density estimates were spatially weighted using the most recent population count available per haulout.

Density estimates were predicted using static environmental covariates. Thus, the density estimates
represent absolute seal density during the main foraging season in a "typical" year.

#

CONSIDERATIONS AND EXCLUSIONS

In contrast to previous iterations of seal density maps (usage maps), confidence intervals given here for absolute density do not
account for uncertainty relating to the population scalars.

Scalar 2 used here relating to the proportion of time spent at-sea is averaged across days and years, and thus does not account
for the variability in proportion of time spent at-sea with season and state of tide.

Confidence intervals also do not account for uncertainty relating to individual variation in habitat preference
or temporal changes in abundance (e.g. throughout the day). See Carter et al. (2020) Section 5.4 for details.
https://assets.publishing.service.gov.uk/media/6022c0bf8fa8f51478436975/...

Estimates for the east coast of Scotland should be treated with caution as they may contain a high degree of
un-modelled uncertainty due to lack of recent telemetry data.