Content uploaded by Ian O'Connor
Author content
All content in this area was uploaded by Ian O'Connor on Feb 19, 2015
Content may be subject to copyright.
Content uploaded by Simon Berrow
Author content
All content in this area was uploaded by Simon Berrow on Sep 20, 2014
Content may be subject to copyright.
DENSITY ESTIMATES OF HARBOUR
PORPOISES PHOCOENA PHOCOENA AT
EIGHT COASTAL SITES IN IRELAND
Simon Berrow, Ronan Hickey, Ian O’Connor and David McGrath
ABSTRACT
Single platform line-transect surveys with distance sampling were carried out at eight sites between
July and September 2008 to derive density and abundance estimates of harbour porpoises. Over 37
days at sea, a total of 475 track-lines were surveyed for a total distance of 20,623km. From the 332
sightings, a total of 618 individual harbour porpoise were recorded. Overall density estimates ranged
from 0.53 to 2.03 porpoises km
2
(without correction for g(0)). Mean group size varied from 1.41
to 2.67. These data provide baseline information to help identify important habitats for harbour
porpoise and reference values for monitoring future changes in harbour porpoise distribution and
densities in Ireland.
INTRODUCTION
The harbour porpoise (Phocoena phocoena L.) is the
most widespread and abundant cetacean species in
Irish waters (Rogan and Berrow 1996). It has been
recorded off all coasts and over the continental shelf
but is thought to be most abundant off the south-
west coast (Reid et al. 2003). Harbour porpoise are
also consistently the most frequently recorded
species stranded in Ireland (Berrow and Rogan
1997; O’Connell and Berrow 2010).
Despite this, relatively little is known about
their abundance or population trends. Leopold et al.
(1992) carried out five line-transects from Galway
to Cork in 1989 on a platform of opportunity and
derived an abundance estimate of 19,120 harbour
porpoises, equating to an overall density of 0.779
0.26 harbour porpoises per km
2
. On a broader
scale, a density of 0.18 harbour porpoises per km
2
was calculated for the Celtic Sea as part of an
international harbour porpoise survey called
SCANS (Small Cetacean Abundance in the North
Sea) which was carried out in July 1994 (Hammond
et al. 2002). This survey was repeated in July 2005
(SCANS-II) but encompassed all Irish waters
including the Irish Sea (SCANS II 2008). Harbour
porpoise densities were calculated for three areas;
Celtic Sea (0.41 porpoise km
2
) Irish Sea (0.34)
and Atlantic coastal Ireland (0.28). The offshore
Ireland survey area included Scotland and a density
of 0.07 porpoise km
2
was generated for both areas
combined. These results suggested the harbour
porpoise density in the Celtic Sea had doubled
between SCANS I and SCANS II, representing an
increase of 11% per annum between 1994 and
2005. The only small-scale survey of harbour
porpoise densities carried out in Irish coastal waters
was that of Berrow et al. (2009), who calculated a
density of 1.33 porpoise km
2
during a survey in
2007 of the Blasket Islands candidate Special Area
of Conservation (cSAC).
As a coastal species, harbour porpoise are
vulnerable to many anthropogenic influences such
as pollution, fisheries interactions and disturbance,
including acoustic disturbance (Jenkins et al. 2009;
Tougaard et al. 2009). In order to assess these
impacts it is essential to understand population
distribution and abundance and to establish baseline
estimates for long-term monitoring (Evans and
Hammond 2004). Increasingly, cetacean sightings
surveys are being carried out as part of environ-
mental impact assessments of inshore sites. How-
ever, there are no data available with which to
compare these site surveys in order to assess their
relative importance for this species.
The harbour porpoise is also a species of high
conservation value throughout Europe. Harbour
porpoise is listed on Annex II of the European
Union Habitats Directive which requires Member
States to designate Special Areas of Conservation
(SAC) for their protection. Guidance on what
constitutes an important site for harbour porpoises
in EU waters is limited, but factors such as good
population density (in relation to neighbouring
areas) and high ratio of young to adults during
certain periods of the year are considered important
(EC 2001). Some EU Member States have used a
number of methods to try to identify important sites
for harbour porpoises. In the UK, Embling et al.
(2010) used spatial modelling to identify sites with
elevated densities off southwest Scotland, while
static acoustic monitoring and aerial surveys were
Received 12 December
2012. Accepted 15
January 2014.
Published 28 March
2014.
Simon Berrow
(corresponding
author; email:
simon.berrow@
iwdg.ie),
Irish Whale and
Dolphin Group,
Merchants Quay,
Kilrush, County Clare
and Marine and
Freshwater Research
Centre, Galway-
Mayo Institute of
Technology, Dublin
Road, Galway;
Ronan Hickey, Irish
Whale and Dolphin
Group, Merchants
Quay, Kilrush,
County Clare; Ian
O’Connor and David
McGrath, Marine
and Freshwater
Research Centre,
Galway-Mayo
Institute of
Technology, Dublin
Road, Galway.
Cite as follows:
Berrow, S., Hickey,
R., O’Connor, I. and
McGrath, D. 2014
Density estimates of
harbour porpoises
Phocoena phocoena
at eight coastal sites
in Ireland.
Biology
and Environment:
Proceedings of the
Royal Irish Academy
2014. DOI: 10.3318/
BIOE.2014.03
DOI: 10.3318/BIOE.2014.03
B
IOLOGY AND
E
NVIRONMENT:
P
ROCEEDINGS OF THE
R
OYAL
I
RISH
A
CADEMY,
V
OL.
114, N
O.
1, 116 (2014).
#
R
OYAL
I
RISH
A
CADEMY
1
used in Germany to identify sites of importance for
harbour porpoise (Verfußet al. 2007).
Here we present the results of dedicated line-
transect surveys for harbour porpoise, using small
vessels at eight coastal sites around Ireland,
including the Blasket Islands and Roaringwater
Bay candidate SACs (cSAC). We present density
estimates and discuss variability in the dataset.
This is used to explore the limitations and
usefulness of using small vessels to record harbour
porpoise in relatively small coastal sites. These
data could provide reference values to identify
sites with elevated densities of harbour porpoise
and to compare with similar surveys carried out
elsewhere.
METHODS
SITE SELECTION
Eight sites were surveyed, two on the east coast,
one in the southeast, three in the southwest, one in
the west and one in the northwest (Fig. 1). Sites
were selected by the National Parks and Wildlife
Service (NPWS) because sighting schemes (Berrow
et al. 2002; 2010) have suggested they have
potential for designation as SACs, together with
those sites (Blasket Islands and Roaringwater Bay)
already designated as cSAC for harbour porpoise.
At three sites (North County Dublin, Roaring-
water Bay, Galway Bay) surveys were carried out
Fig. 1
*
Map of Ireland showing location of sites surveyed for harbour porpoise during 2008.
2
B
IOLOGY AND
E
NVIRONMENT
on six days, two each in July, August and September
2008. At two sites (Dublin Bay, Cork Coast) five
surveys were carried out and at three sites (Carnsore
Point, Blasket Islands and Donegal Bay) three surveys
were carried out, one in each in July, August and
September 2008. Each survey at each site was to be
completed in one day as more than one consecutive
day of suitable sea conditions to complete a survey is
rare in Ireland, especially on the west coast.
SURVEY TRAINING
A team of 19 surveyors were used in this survey,
eleven experienced primary observers and eight
recording survey effort, all organised into three
teams. This made it possible to take advantage of
suitable weather conditions at short notice, with
occasionally all three teams surveying different sites
on the same day. However, the use of multiple
primary observers can increase variability within a
dataset through differences in observer performance.
Trials were organized to train observers in the survey
methodology and identify sources of variability (see
O’Brien 2009). Two teams of six observers were
sent to two sites adjacent to the Shannon Estuary on
two occasions. All observers were visually excluded
from each other and asked to record the time of any
sightings, therefore allowing for the assessment of
variability between observers in time taken to record
first sighting, estimation of distance to the observed
animals and the number of groups.
To assist in estimating distance, two trials were
carried out where observers on land were asked to
estimate distances to a Rigid Inflatable Boat on the
estuary. The distance was then determined using a
Leica Rangemaster 1200. This range finder reports
an accuracy to within 92m over 800m or 90.5%
over 600m. In the first trial observers were given
ten distances to estimate between 50 and 1000m
with no feedback provided between estimations as
to the actual distance. In the second trial, ten more
distances were estimated but observers were told
the actual distance between each estimate.
SURVEY METHODOLOGY
Twelve different vessels were chartered over the
duration of the survey period. Vessels were chosen
that could navigate close to the shore, had a shallow
draft and a flying bridge and were relatively cheap to
charter. They ranged in length from 11m to 18m and
platform height above sea-level, from 2.3m to 3.2m.
Every vessel had active high-frequency navigation
sonar on during each survey due to health and
safety requirements: coastal sites navigated have a
complex bathymetry and often only poor navigation
charts are available. Conventional single platform
line-transect surveys were carried out within or in
close proximity to the boundaries of survey sites along
pre-determined routes as described in detail by
Berrow et al. (2009). Transect lines were chosen to
cross depth gradients and provide as close to equal
coverage probability as possible following the recom-
mendations of Dawson et al. (2008) who suggested
systematic line spacing resulted in better precision
than randomised line spacing. At most sites, lines were
changed for each survey to try to get full coverage of
the site over the study period to ensure no potentially
important porpoise concentrations were overlooked.
In the Blasket Islands cSAC, track-lines were repeated
during each survey day to explore the variability
on the same track-lines between visits, as Berrow
et al. (2009) had already carried out a survey of the site
in 2007 using randomly placed track-lines.
Each survey vessel travelled at a speed of
1216km hr
1
, which was 23 times the typical
average speed of the target animal (harbour por-
poise) as recommended by Dawson et al. (2008).
Two primary observers were positioned on the
flying bridge, which provided an eye-height above
sea-level of between 4m and 6m, depending on the
height of the platform and each individual observer.
Primary observers watched with naked eye from
dead ahead to 908to port or starboard depending on
which side of the vessel they were stationed. All
sightings were recorded but sightings over 200m
(300m if sea-state 0 predominated) perpendicular
distance from the track-line were not used in the
distance model, following the recommendations of
Buckland et al. (2001). These extreme values do not
contribute much to the density estimate and they
make it difficult to fit the detection function. Calves
were defined as animals less than half the length of
the accompanying animal (adult) and juveniles half to
two-thirds of the adult length.
During each transect, the position of the survey
vessel was tracked continuously through a GPS
receiver connected directly to a laptop, while
survey effort, including environmental conditions
(sea-state, wind strength and direction, glare etc.)
were recorded directly onto LOGGER software
(#IFAW) every 15 minutes. When a sighting was
made, the position of the vessel was recorded
immediately and the angle of the sighting from
the track of the vessel and the perpendicular
distance of the sighting from the vessel recorded.
These data were communicated to the recorder in
the wheelhouse via VHF radio. The angle was
recorded to the nearest degree via an angle board
attached to the vessel immediately in front of each
observer. Accurate distance estimation is essential
for distance sampling. At some sites during each
survey an orange buoy 225mm in diameter was
towed 200m astern of the observers’position on the
survey vessel. This provided a reference point
against which to estimate distance.
3
D
ENSITY ESTIMATES OF HARBOUR PORPOISES AT
I
RISH COASTAL SITES
ABUNDANCE ESTIMATE
The software programme DISTANCE (Version 5,
University of St Andrews, Scotland) was used for
calculating the detection function, which is the
probability of detecting an object a certain distance
from the track-line. The detection function is used
to calculate the density of animals on the track-line
of the vessel. In this survey we assumed that all
animals on the track-line were observed, i.e. that
g(0) 1. This was clearly not the case but without
a double-platform survey technique the true value
of g(0) cannot be determined. This survey was
intended to be conducted from small vessels
operating inshore, and the inability to conduct a
double-platform methodology was one of the
inherent constraints. The DISTANCE software
allows the user to select a number of models in
order to identify the most appropriate for the data.
It also permits the truncation of outliers when
estimating variance in group size and testing for
evasive movement prior to detection.
The track-line was used as the sample with
sightings as observations following Berrow et al.
(2009). Estimates of abundance were calculated for
each survey day, providing there were sufficient
sightings to generate an estimate. The overall abun-
dance estimate was derived from all track-lines in sea-
state 2 or less from all days combined. This was
necessary to obtain sufficient sightings (minimum
required is 4060, Buckland et al. 2001) for a robust
estimate. For the model, we have assumed that there
were no major changes in distribution within each site
between sample days or any immigration or emigra-
tion into or out of the site over the survey period.
We fitted the data to a number of models. We
found that a half-normal model with hermite
polynomial series adjustments best fitted the data
according to Akaike’s information criterion. The
recorded data were grouped into equal distance
intervals of 020m, 2040m and up to 180200m
for surveys where sea-state 1 and 030, 3060
and up to 300m for surveys when sea-state 51.
The detection functions were calculated from all
survey data combined, as the number of sightings
recorded each day were too few to generate robust
density estimates. This would be compounded
further if data were stratified by sea-state, vessel,
etc., thus the detection function was determined
from a pooled dataset in each site. Cluster size was
analysed using size-bias regression method with
log(n) of cluster size against estimated g(x). The
variance was estimated empirically.
A Chi-squared test is associated with each
detection function. If significant then this indicated
that the detection function was a good fit and the
estimate is robust. Some fits were not significant
and thus the results from these detection functions
should be treated with extreme caution. We present
them to show the effect of small samples sizes
(number of survey days, sightings) on the model
and thus the limitations of the methodology. The
proportion of the variability accounted for by the
encounter rate, detection probability and group size
is presented with each detection function. Varia-
bility associated with the encounter rate reflects the
number of sightings on each track-line, which will
vary from zero to up to seven or eight. The
detection probability reflects how far the sightings
were from the track-line and group size the range
of group sizes recorded at each site.
Maps were created using Irish Grid (TM65
Irish Grid) with ArcView 3.2 while maps of the
survey areas were obtained from the National Parks
and Wildlife Service
RESULTS
SURVEY TRAINING
The results from the distance trials show very
accurate distance estimation up to 200m with a
very small under-estimate at short distances (Fig. 2a).
At greater distances accuracy was less, with a
tendency to under-estimate the distance. In the
second trial, observers were told the actual distance
between each estimate, which enabled them to
improve slightly on estimation of large distances
following this feedback (Fig. 2b).
SITE SURVEYS
During 37 survey days, a total of 475 track-lines
were surveyed for a total distance of 20,623km in
1
0
200
400
600
800
1000
(a)
(b)
Distance (m)
2345
Test number (re-ordered in ascending order)
678910
Actual
Estimate
1
0
200
400
600
800
1000
Actual distance
Estimate
Distance (m)
2345
Test number (re-ordered in ascendin
g
order)
678910
Fig. 2
*
Mean distance estimates during blind trials a) with
no feedback and b) with feedback.
4
B
IOLOGY AND
E
NVIRONMENT
sea-state 52. From the 332 sightings, a total of
618 individual harbour porpoise was recorded
(Table 1). The proportion of the variability ac-
counted for by the encounter rate ranged from 80%
to 49% with between 15% and 34% attributed to
detection probability and 29% to 22% due to group
size. This shows that generally it is the number of
sightings on each track-line that shows the greatest
variability, which is to be expected as many track-
lines will have no sightings, while others will have
many. However significant variability at some sites
was attributed to detection probability (e.g. Galway
Bay) and group size (e.g. Carnsore Point). The
detection functions from each site indicated some
minor evasive movement with a peak in sightings
generally between 20m and 60m from the track-
line. A summary of results from each site is
presented.
North County Dublin
The track-lines and sightings for the six surveys
carried out in North County Dublin are shown in
Fig. 3. The distribution of effort in sea-state 0 and 1
(black lines) was good, with effort in all areas of the
study site. There was considerable variability in the
number of sightings per survey day, ranging from
Fig. 3
*
Map showing location of all track lines surveyed and harbour porpoise observed in North County Dublin.
Table 1
*
Date, sea-state and number of sightings of harbour porpoises at each site during 2008.
Site No. of
survey
days
No. of
track
lines
Total distance
in sea-state
52 (km)
Sea-state (% of total survey
time)
Number
of sightings
Total
Animals
0123
North County Dublin 6 69 293.75 15.4 47.6 19.5 14.7 82 111
Dublin Bay 5 75 289.38 15.8 34.0 34.0 14.1 56 69
Carnsore Point 3 33 183.59 2.7 34.5 43.0 19.8 13 23
Cork Coast 5 58 435.97 3.2 34.3 48.3 14.2 28 72
Roaringwater Bay 6 70 330.63 10.9 53.7 19.7 15.8 47 110
Blasket Islands 3 54 208.26 19.3 25.2 47.8 7.8 31 57
Galway Bay 6 82 627.10 23.2 33.6 25.8 17.5 62 134
Donegal Bay 3 34 230.95 51.1 26.9 18.9 2.8 19 42
5
D
ENSITY ESTIMATES OF HARBOUR PORPOISES AT
I
RISH COASTAL SITES
none to 48; even in excellent sea-state variability
was large. On 12 July, despite sea-state 0 and 1
occurring for 94% of the survey only 8 sightings of
a total of 9 individuals were recorded, however on
29 August a total of 48 sightings of a total of 67
individuals were recorded in sea-state 0 and 1 for
88% of survey effort. However, only two weeks
later in sea-state 0 or 1 for 96% of survey effort only
15 sightings of 21 individuals were recorded.
Density estimates for North County Dublin ranged
from 0.54km
2
on 12 July to 6.93km
2
on 29
August. Mean group size was consistent at between
1.14 and 1.41 per sighting. The overall density
estimate was 2.03km
2
, which gave an abundance
estimate of 211947 (CI and CV values for all site
estimates are given in Table 3).
Dublin Bay
Track-lines and the position of each sighting
during five survey days in Dublin Bay are shown
in Fig. 4. Effort in the inner bay was restricted by
water depth (B5m). Effort in sea-state 0 and 1 was
distributed throughout the site though there was
more effort in sea-state 2 (red lines). There were
concentrations of harbour porpoises to the north of
the site but porpoises were distributed throughout.
For the DISTANCE analysis, data from the first
two days (13 and 28 July) were omitted as the sea-
state was high and the number of sightings low (two
and three on each day). Thus a total of 54 track-
lines and 50 sightings were used in the analysis.
Density estimates ranged from 0.48 to 2.05 km
2
.
The mean group size was quite consistent ranging
from 1.08 to 1.50. The overall density estimate was
1.19 km
2
, which gave an abundance of 138933.
Carnsore Point
Three surveys were carried out off Carnsore Point
(Table 2). On 21 July, sea-state 2 or less was
recorded during 77% of the survey, and we
recorded seven sightings. On 22 August although
sea-state 2 or less was recorded on 95% of the
survey, nearly one-half was in sea-state 2 and only
two sightings were recorded. On the last day (13
September) sea-state was mainly 2 but sea-state 3
accounted for 32% of the survey time; only four
sightings were made. An important constraint at
this site was the strong tides. With a mean survey
duration of 5-6 hours there was always a period
of strong tides during the survey causing high
sea-states especially on the shallow sand banks
to the northwest of the study site. Track-lines and
sightings are shown in Fig. 5. The distribution of
effort in sea-state 0 and 1 (black lines) is concen-
trated to the southern half of the study site. Harbour
porpoises were distributed throughout the study
area with concentrations to the south-east. The
detection function at this site was a poor fit
Fig. 4
*
Map showing location of all track lines surveyed and harbour porpoise observed in Dublin Bay.
6
B
IOLOGY AND
E
NVIRONMENT
Table 2
*
The goodness of fit of the detection function using DISTANCE analysis and the
proportion of variability attributed to each main variable at each site surveyed.
Site Detection Function (X2) Proportion of variability (%)
Encounter Rate Detection Probability Group Size
North County Dublin 13.7, p0.09 80.3 15.3 4.4
Dublin Bay 22.4, p0.01 78.0 19.0 2.9
Carnsore Point 6.10, p0.64 49.3 28.3 22.4
Cork Coast 6.86, p0.55 55.0 26.8 18.2
Roaringwater Bay 6.43, p0.60 61.2 25.1 13.7
Blasket Islands 7.33, p0.50 68.0 23.7 8.4
Galway Bay 3.88, p0.09 50.2 34.3 15.3
Donegal Bay 5.92, p0.66 59.6 23.8 16.7
Fig. 5
*
Map showing location of all track lines surveyed and harbour porpoise observed off Carnsore Point.
Table 3
*
Mean overall density and abundance of harbour porpoise at eight sites
Site N (95% CI) SE CV Density (km
2
) Group size mean (95% CI)
North County Dublin 211 (137327) 47.1 0.22 2.03 1.41 (1.261.56)
Dublin Bay 138 (86221) 33.2 0.24 1.19 1.22 (1.111.34)
Carnsore Point 87 (39196) 36.3 0.42 0.58 1.91 (1.252.92)
Cork Coast 173 (92326) 56.6 0.33 0.53 2.67 (1.963.64)
Roaringwater Bay 159 (95689) 42.4 0.27 1.24 2.21 (1.852.64)
Blasket Islands 372 (216647) 105.3 0.28 1.65 1.76 (1.502.07)
Galway Bay 402 (267605) 84.1 0.21 0.73 2.15 (1.842.51)
Donegal Bay 249 (106586) 111.5 0.45 0.88 2.40 (1.633.53)
7
D
ENSITY ESTIMATES OF HARBOUR PORPOISES AT
I
RISH COASTAL SITES
(P0.64) reflecting the very low number of
sightings. The proportion of the variability ac-
counted for by the encounter rate was 49.3%, with
28.3% attributed to detection probability and 22.4%
due to group size. These figures again reflect the
small number of sightings and this density estimate
should be treated with extreme caution. The
overall density estimate for Carnsore Point was
0.58 harbour porpoises km
2
. Mean group size was
ca. two animals. This resulted in an abundance
estimate of 87936.3.
Cork Coast
Five survey days were carried out in the Cork coast
site (Table 8). Track-lines and sightings for the
Cork coast are shown in Fig. 6. Sea-state 51 and
sea-state 2 were distributed throughout the site but
most sightings were off the Old Head of Kinsale
and to a lesser extent Seven Heads and Galley
Head (Fig. 6). The detection function was con-
sidered a poor fit (P 0.55) and thus the results
should be treated with caution. The proportion of
the variability accounted for by the encounter rate
was 55.0%, with 26.8% attributed to detection
probability and 18.2% due to group size. This is
somewhat different to other sites where the varia-
tion due to encounter rate was higher with lower
variation due to group size. However the dataset for
this site is small and the low number of sightings
(albeit of relatively large group sizes (up to eight
individuals)) will have had a large negative influ-
ence on the significance of the detection function.
The overall density estimate from the Cork coast
was 0.53 harbour porpoise km
2
. This estimate is
higher than might have been expected from the
low sighting rate which is due to the high group
size estimates, including two observations of eight
harbour porpoises and one of six.
Roaringwater Bay cSAC
Six survey days were carried out in Roaringwater
Bay cSAC with good sea conditions recorded on
two days, which returned 13 and 23 sightings.
Overall there were 47 sightings of a total of 110
individuals. Track-lines and sightings for Roaring-
water Bay are shown in Fig. 7. Track-lines surveyed
in sea-state 52 were distributed throughout the
site. Most sightings were around Gascanane Sound
between Sherkin and Clear Islands and off the
western tip of Cape Clear (Fig. 7). The detection
function was not considered a good fit (P 0.60).
There was some evidence of evasive behaviour
by the porpoises, with a peak on the track-line
but also at 2040m and 60100m from the
track-line. The proportion of the variability ac-
counted for by the encounter rate was 61.2%, with
25.1% attributed to detection probability and 13.7%
due to group size. This was similar to the Cork
coast and may be due to the high variability in the
number of sightings per survey day. Mean group
Fig. 6
*
Map showing location of all track lines surveyed and harbour porpoise observed in the Cork coast site.
8
B
IOLOGY AND
E
NVIRONMENT
size was consistent at around two animals. The
overall density estimate was 1.24 km
2
with a CV
of 0.27. This gave an abundance estimate of 1599
42.
Blasket Islands cSAC
Three surveys were carried out within the Blasket
Islands. The track-lines and position of each
sighting are shown in Fig. 8. Effort in sea-state 0
and 1 was distributed throughout the site though
generally there was more effort in sea-state 2 (red
lines). There were concentrations of harbour
porpoises on the south side of Great Blasket and
to a lesser extent in Blasket Sound (Fig. 8). The
number of sightings varied between five and 19
sightings and a total of seven and 37 individuals
between the three survey days along the same
track-lines. The distribution of sightings was con-
sistent between survey days with most sightings to
the south of Great Blasket, north of Inistooskert and
in Blasket Sound (Fig. 8). A total of 54 track-lines
and 31 sightings were used in the distance analysis.
The detection function was not considered a good
fit (P0.50) and thus estimates should be treated
with caution. The mean group size also ranged
greatly from 1.00 to 2.29. The overall density
estimate was 1.65 km
2
, which gave an abundance
of 3729105.
Galway Bay
Six surveys were carried out in Galway Bay
(Table 1). A total of 84 track-lines were surveyed
covering 627km in sea-state 52 which resulted in
62 sightings of a total of 134 individuals. The track
lines surveyed in Galway Bay are shown in Fig. 9.
Effort in sea-state 0 and 1 and sea-state 2 is
distributed throughout the survey area. Harbour
porpoises were distributed throughout the survey
area with concentrations off Black Head, Co. Clare,
and towards the middle of the bay. The proportion
of the variability accounted for by the encounter
rate was 50.2%, with 34.3% attributed to detection
probability and 15.3% due to group size. This is
somewhat different to other sites where the varia-
tion due to encounter rate was higher due to
smaller group sizes. This indicates there was more
variability in the number of sightings recorded per
track-line and a greater range in group size. This
may reflect the larger area of this site. The overall
density estimate was 0.73 km
2
giving an abun-
dance 9SE of 402984. Galway Bay is 547km
2
in
area and even though the density estimate was low
the overall abundance was high.
Donegal Bay
Three surveys were carried out in Donegal Bay
(Table 1). The track-lines and sightings are shown in
Fig. 10. Sea-state 51 was distributed throughout
Fig. 7
*
Map showing location of all track lines surveyed and harbour porpoise observed in Roaringwater Bay cSAC.
9
D
ENSITY ESTIMATES OF HARBOUR PORPOISES AT
I
RISH COASTAL SITES
the site but most harbour porpoise sightings were
concentrated in the centre of the bay. The detection
function was a poor fit (P 0.66). The proportion of
the variability accounted for by the encounter rate
was 59.6%, with 23.8% attributed to detection
probability and 16.7% due to group size. The
Fig. 9
*
Map showing location of all track lines surveyed and harbour porpoise observed in Galway Bay.
Fig. 8
*
Map showing location of all track lines surveyed and small cetaceans observed in the Blasket Islands cSAC.
10
B
IOLOGY AND
E
NVIRONMENT
variability attributed to the encounter rate was
relatively low with a higher proportion attributed
to group sizes, which reflects the large range in group
size recorded. The overall estimate was 0.88 km
2
giving an abundance of 2499111.5.
PROPORTION OF ADULT TO YOUNG
We calculated the proportion of adults to young
(combining records of those animals described as
juveniles or calves) for each site. No calves or
juveniles were observed off the Cork coast.
The proportions of juveniles and calves in sites,
excluding Cork Coast, were consistent with a
percentage of 6%8% (Table 4).
DISCUSSION
Statistical inference using distance sampling rests on
the validity of several assumptions (Buckland et al.
2001). These include that objects are spatially
distributed according to some stochastic process. If
transect lines were randomly placed within the
study area one can safely assume that objects are
uniformly distributed with respect to the perpendi-
cular distance from the line in any given direction.
Another assumption is that objects on the track-line
are always detected (i.e. g(0) 1) and are detected
at their initial location prior to any movement in
response to the observer. If objects on or near to the
track-line are missed then the density estimate will
be biased low. To minimise the effect of movement
Fig. 10
*
Map showing location of all track lines surveyed and all small cetaceans observed in Donegal Bay.
Table 4
*
Proportion of adult to young for all sites surveyed in 2008.
Site No. of sight-
ings
No. of
animals
No. of
Adults
No. of
Juveniles
No. of
calves
Proportion of
young (%)
North County Dublin 82 111 102 1 8 8
Dublin Bay 56 69 65 1 3 6
Carnsore Point 12 22 19 3 0 14
Cork Coast 28 72 72 0 0 0
Roaringwater Bay 47 110 102 8 0 7
Galway Bay 62 134 124 2 8 7
Blasket Islands 30 55 45 0 10 18
Donegal Bay 18 40 37 0 3 8
11
D
ENSITY ESTIMATES OF HARBOUR PORPOISES AT
I
RISH COASTAL SITES
it is recommended that the speed of the observer
is at least twice the speed of the object (Buckland
et al. 2001).
Typically for surveys of harbour porpoise
g(0)0.4 or 0.5, i.e. only one-half of the animals
on the track-line are detected (Buckland et al.
2001). If this was the case with the present survey,
then we should double the density estimates.
Without a double-platform methodology it is not
possible to accurately determine the numbers
missed on the track-line. This methodology would
require an increase in the number of surveyors and
a larger vessel, which may restrict how close to
shore the vessel could survey as well as considerably
increasing costs. This is a major constraint of single-
platform methods and this can only be used for
between site comparison if the same method and
analysis is used between surveys and sites. This may
realistically only provide a relative rather than
absolute abundance estimate. The detection func-
tions for most sites suggested there was some
evasive movement from the boat. The model can
be adjusted for this but the effect of this evasive
behaviour was not consistent between days and sites
so no account was taken of this reaction to boats
for consistency. This factor will also cause the
density estimates to be under-estimated. We have
attempted to keep the survey methodology and
treatment of the data consistent at all sites, through-
out the present survey. Survey methodology and
data analysis were also consistent with that used
by Berrow et al. (2009), which facilitates compar-
ison of the data between, and within, surveys. It is
interesting to note that the results of the survey of
the Blasket Islands cSAC in 2007 and 2008 were
quite similar giving some confidence in the meth-
odology. Only additional surveys will see if this
similarity in estimates is consistent.
The ability to detect harbour porpoise visually
at sea and thus the accuracy of density and
abundance estimates is extremely dependent on
sea-state. During the present study, transects were
carried out, whenever possible, in sea-state 2 or
less as the ability to detect harbour porpoise
decreases significantly in sea-state ]3 (Teilmann
2003). Berrow et al. (2009) recommended that all
harbour porpoise surveys should be carried out in
sea-state 0 or 1 to ensure all animals are detected
and g(0) is as close to 1 as possible. This is rarely
achievable and given the poor weather throughout
the summer in 2008 we were fortunate to be able
to carry out as many surveys as we did in relatively
good sea-state (sea-state 52). The data could be
stratified by sea-state within each site if necessary if
subsequent surveys record any significant changes
in densities but the number of sightings available
for the model will decrease.
SURVEY TRAINING
The use of a team of surveyors contributed to the
successful completion of the survey as this enabled
full advantage to be taken of favourable weather
conditions. Having a large pool of observers meant
that a team could be put together at short notice
and more than one site could be surveyed on the
same day. However, having such a large number of
observers, with different experience and abilities,
could lead to large variability in the accuracy of
parameters recorded such as distance to sightings
and in estimations of group size. O’Brien (2009)
explored these variables in more detail and showed
that experienced observers were good at recording
sightings and estimating distance consistently.
Variability in estimation of group size could
potentially be significant during surveys of small
cetaceans (O’Brien 2009), but harbour porpoise in
the present survey generally only occurred in small
groups so the effect should be limited. We have
shown that if using a large team of experienced
observers, heterogenity in the data caused by
different observers could be minimised especially
if training and observer trials are carried out at the
start of the survey.
COMPARISON OF HARBOUR PORPOISE
DENSITY ESTIMATES
There are a number of variables in the data
collected during a survey which may contribute
to large standard errors and wide confidence
intervals that should be considered when presenting
results and interpreting the data. The mean group
sizes of harbour porpoise varied considerably
between sites with larger groups recorded on the
south and west coasts compared to the east coast.
The number of sightings, and thus density esti-
mates, during each visit to the same site can vary
considerably over relatively short time periods. For
example, 48 sightings of a total of 67 harbour
porpoises were recorded on 29 August in North
County Dublin followed by only 15 sightings and
21 individuals 14 days later on the 12 September,
despite sea-state 51 for 88% and 96% of the survey
effort. The high number of sightings recorded on
29 August in North County Dublin suggests
immigration into the site may have occurred.
This violates one of the assumptions made for this
survey methodology, namely that there were no
changes in distribution within each site during the
survey period. North County Dublin and Roaring-
water Bay cSAC sites are small in area, which makes
them sensitive to even small local movements of
harbour porpoises in the adjacent area and this
should be considered when determining the size of
a site to be monitored. Strong short-term or
seasonal patterns in abundance of harbour porpoise
have been shown elsewhere (e.g. Marubini et al.
12
B
IOLOGY AND
E
NVIRONMENT
2009;; Gilles et al. 2011) which suggests that by
moving the date of a survey by only a few weeks
can produce large differences in density estimates.
Ideally survey day should be randomised but as
suitable weather conditions dictate when a survey
can be carried out this is not possible. However if
effort is consistent with a defined period, such as six
survey days within the same three month period,
then this variable may not be too influential.
The coefficients of variation (CV) at five of the
eight sites surveyed were between 0.22 and 0.28.
However, Englund et al. (2007) recommended that
the CV of abundance estimates, if used for
monitoring the population status of bottlenose
dolphins in the Shannon Estuary, should be as
low as 0.12 if changes in abundance were to be
detected within reasonable time-frames. To achieve
lower CVs for harbour porpoise density estimates
more surveys will be required at each site to
increase the number of sightings. The data on
detection rates from the present survey could be
used in a power analysis to predict the number of
surveys required to reduce these CVs. This could
inform managers on the effort and cost required to
monitor harbour porpoise densities. Large inter-
annual fluctuations in the densities of harbour
porpoises within a site may occur and data should
be collected over a number of years before sites are
assessed for site designation. For example, porpoise
densities in the Cork Coast site were lower than
might have been expected. This site was chosen as
concentrations of harbour porpoise sightings have
been reported off the Old Head of Kinsale and
Galley Head (Berrow et al. 2010). This suggests
2008 was a poor year for harbour porpoise at this
site and the results from the present survey may not
reflect the importance of this site for this species.
The presence of other cetacean species within a site
may influence the distribution and densities of
harbour porpoise. Bottlenose dolphins are regularly
recorded in Donegal Bay (Ingram et al. 2003;
Berrow et al. 2008) and the density of harbour
porpoises in the bay might vary depending on their
presence as this species is known to attack and kill
harbour porpoises (Ross and Wilson 1996). How-
ever as the dolphins in Donegal Bay are known to
be transient (O’Brien et al. 2009), harbour porpoise
densities may increase when dolphins are absent.
Habituation to ships may also be a factor with
porpoises less likely to show evasive behaviour in
areas exposed to heavy boat traffic. This would
negatively bias estimates at sites with les traffic as
porpoises may react more to the survey vessel.
All these variables should be taken into account
if short-duration surveys are to be conducted in a
single season and year. Results from a similar survey
of the Blasket Islands cSAC in 2007 were used to
compare density estimates from 2008. It was
encouraging that the estimate in 2008 (1.65) was
similar to 2007 (1.33) with similar CVs (0.28
compared to 0.25). As this site was already
designated for harbour porpoise it was considered
to provide good habitats for harbour porpoise and
thus the density recorded there (1.33) may repre-
sent an upper range for coastal sites in Ireland.
During the present survey only North County
Dublin (2.03) had a higher overall density than
the Blasket Islands cSAC. The density estimates
in Roaringwater Bay cSAC (1.24), which is also
designated for harbour porpoises, was very similar
to that recorded in the Blasket Islands cSAC during
2008. Dublin Bay, which is around 20km south of
the North County Dublin site at 1.19 harbour
porpoises per km
2
was only slightly lower again.
Densities at the other four sites were less than 1.0
harbour porpoises per km
2
with densities at some
sites (Carnsore Point and the Cork Coast) were
around 0.500.60, nearly one-third of the densities
recorded in the two cSACs and may reflect
relatively poor habitat for this species.
Comparing the results from the present survey
to harbour porpoises surveys elsewhere is difficult as
different methods were used (single v double-plat-
form, dedicated v opportunistic surveys) and over
very different spatial scales. Most density estimates
reported in the present study were higher than that
recorded during SCANS-II (2008), where the
highest reported density was 0.81 harbour porpoises
km
2
off the east Danish coast (Table 5). This is to
be expected as the survey areas in SCANS II were
very large and included offshore waters. If densities
of harbour porpoises are greater closer to the coast
(say within 12 nmls) then this will result in higher
overall estimates in the small coastal sites surveyed
during the present study. Thus it is not reasonable to
compare the present study with results from
SCANS-II: line transect surveys, which cover the
whole range of a species, are less likely to be strongly
influenced by this than small-scale surveys.
PROPORTION OF ADULTS TO YOUNG
The proportion of adults to young at the four sites
reported with young, was remarkably consistent, at
between 6% and 8%. Given the large differences in
the number of sightings and numbers of individual
porpoises recorded between sites, it was surprising
that the proportion of calves was so consistent. This
proportion was consistent for both small (North
County Dublin, 104km
2
8%) and relatively
large sites (Galway Bay, 547km
2
7%). Sonntag
et al. (1999) summarised data on the proportion of
calves from 13 aerial surveys and 10 ship-based
surveys throughout the North Sea and Kattegat
area, including data from SCANS (Hammond et al.
2002). The proportion of calves ranged from 5.1%
(Inner Danish waters) to 17.9% (Isle of Sylt) from
aerial surveys and 2.2%6.7% from ship-based
13
D
ENSITY ESTIMATES OF HARBOUR PORPOISES AT
I
RISH COASTAL SITES
surveys. Irish Sea records were 5.1% calves and only
3.3% in British coastal waters. The 6%8% pre-
sented in this survey (Table 4) is likely to be higher
than the data from the SCANS surveys as the sites
are all small and coastal compared to much larger
areas surveyed during the SCANS surveys. Sonntag
et al. (1999) suggested the high proportion of calves
of the Isle of Sylt in Germany (9.6%17.9%)
indicated that it was a preferred calving ground
for harbour porpoise in the southern North Sea.
Our data do not suggest such elevated levels but
proportions are probably typical of Irish coastal
waters (Table 5).
USEFULNESS OF METHODOLOGY FOR
MONITORING HARBOUR PORPOPOISES
Estimating densities and abundance of harbour
porpoises in Irish coastal waters presents many
difficulties. These are mainly associated with diffi-
culties in observing this species in high sea-states
and swell. We have attempted to use conventional
single platform line-transect sighting surveys to
record densities of harbour porpoises for identifying
important habitats and monitoring abundance.
Small vessels can be chartered at short notice to
take advantage of suitable weather. We also
designed the survey methodology to obtain full
coverage of the site in one day again to take full
advantage of suitable weather windows. We have
demonstrated that this method can be used to
estimate densities providing suitable vessels, obser-
ver training and appropriate weather windows are
fully considered. A number of surveys of the entire
area over a number of days are preferable to
incomplete surveys of the study site during a single
day.
Sources of variability are many and are difficult
to stratify for unless there are a large number of
sightings. Given that each survey was carried out on
a single day it is unlikely that the number of
sightings per survey will be large enough to satisfy
the requirements of the model (4060 sightings) for
robust estimates. The low number of sightings may
result in a poor fit of the model to the data. It may
be best to ignore these results or ideally use them to
carry out a power analysis to determine how many
track-lines or days are required to provide enough
sightings for a good fit and a low CV of the density
estimates. However, providing surveyors only use
data collected in sea-state 52, use experienced
observers and make the same assumptions when
analysing or modelling the data, then similar surveys
may be used to quantify porpoise densities and
make comparison between sites, providing a similar
methodology is adopted. Changes in prey type
(benthic or pelagic fish species), their distribution
and abundance may occur, and thus the use of
Table 5
*
Density estimates of harbour porpoise during dedicated sighting surveys in Ireland and
elsewhere in the EU.
Location Year Area
(km
2
)
Density
(km
-2
)
Abundance
9
SE
(95% Confidence Intervals)
CV Reference
North County
Dublin
2008 104 2.03 211947.1 (137327) 0.23 This study
Dublin Bay 2008 116 1.19 138933.2 (86221) 0.24 This study
Carnsore Point 2008 151 0.58 87936.3 (39196) 0.42 This study
Cork Coast 2008 326 0.53 173956.6 (92326) 0.33 This study
Roaringwater
Bay
2008 128 1.24 159942 (95689) 0.27 This study
Blasket Islands 2008 227 1.65 3729105.3 (216647) 0.28 This study
Galway Bay 2008 547 0.73 402984.1 (267605) 0.21 This study
Donegal Bay 2008 281 0.88 2499111.5 (106586) 0.45 This study
Blasket Islands 2007 227 1.33 303976 (186494) 0.25 Berrow et al. (2009)
Northern North
Sea
1994 118,985 0.78 - 0.25 Hammond et al. (2002)
Orkney and
Shetland
1994 31,059 0.78 - 0.34 Hammond et al. (2002)
East Danish
Coast
1994 7,278 0.81 - 0.27 Hammond et al. (2002)
South Central
North Sea
2005 156,972 0.56 - 0.23 SCANS-II (2008)
Coastal NW
Denmark
2005 20,844 0.56 - 0.43 SCANS-II (2008)
14
B
IOLOGY AND
E
NVIRONMENT
habitats by porpoises may change considerably. This
can cause significant changes between seasons. We
have attempted to survey sites in a three month
period in the summer but a greater time series of
data will be required to explore the scale of this
variability between seasons and years within sites.
The density estimates presented here could be
used as a reference with which to compare results of
similar surveys elsewhere, in order to assess the
importance of these sites for harbour porpoise. In
addition to these baselines, more surveys during the
same period are required to determine if estimates
are consistent or show large inter-annual variability.
This includes identification of sites with elevated
densities for designation as protected areas. This
methodology could also be used in impact studies in
Irish coastal waters, such as those associated with
offshore wind and wave-farms and other major
developments, where an assessment of the impor-
tance of the site for harbour porpoises and other
species of small cetacean is required. In time, with
more surveys this method may be suitable, espe-
cially when combined with acoustic monitoring
techniques, to monitor population dynamics and
status.
ACKNOWLEDGEMENTS
We would like to thank all those surveyors who
made this survey possible: John Brophy, Eileen
Diskin, Bojana Ferjan, Brian Glanville, Sophie
Hansen, Eugene McKeown, Dana Miller, Fiacc
O’Brolchain, Joanne O’Brien, Mick O’Connell,
Tracey O’Shea, Debbie Pedreschi, Conor Ryan,
Dave Wall, Faith Wilson, Peter Wilson and Pa
´draig
Whooley. We would like to thank all the boat
operators: Colin Barnes, Joe Brady, Jeff Brownlee,
William van Dyke, Declan Kilgannon, Se
´an
McDonogh, Ruairi O’Brien, Edmund O’Byrne,
Micha
´el O’Chonneile and Fachtna O’Driscoll,
Mick Sheeran, and Richard Timony, who have
been extremely tolerant of late changes to plans and
some early starts. This survey was funded by the
Department of Environment, Heritage and Local
Governments’National Parks and Wildlife Service,
and we thank Dr David Lyons for his support
during this project.
REFERENCES
Berrow, S.D. and Rogan, E. 1997 Cetaceans stranded
on the Irish coast, 19011995. Mammal Review
27 (1), 5176.
Berrow, S.D., Whooley, P. and Ferriss, S. 2002 Irish
Whale and Dolphin Group cetacean sighting review
(19912001). Irish Whale and Dolphin Group.
34pp. ISBN 0-9540552-1-7.
Berrow, S., Whooley, P., O’Connell and M., Wall, D.
2010 Irish cetacean review (20002009). Irish Whale
and Dolphin Group, 60pp. ISBN 0-9540552-4-1.
Berrow, S.D., O’Brien, J., O’Connor, I. and McGrath,
D. 2008 Abundance estimate and acoustic mon-
itoring of harbour porpoise Phocoena phocoena in the
Blasket Islands candidate Special Area of Conserva-
tion. Biology and Environment: Proceedings of the Royal
Irish Academy 109B (1), 3546.
Berrow, S.D., Hickey, R., O’Connor, I. and McGrath,
D. 2008 Small cetacean site investigations survey
2008. Unpublished Report to the National Parks
and Wildlife Service. Irish Whale and Dolphin
Group.
Buckland, S.T., Anderson, D.R., Burnham, K.P., Laake,
J.L., Borchers, D.L. and Thomas, L. 2001 An
introduction to distance sampling: estimating abun-
dance of biological populations. Oxford. Oxford
University Press.
Dawson, S., Wade, P., Slooten, E. and Barlow,
J. 2008 Design and field methods for sighting
surveys of cetaceans in coastal and riverine habitats.
Mammal Review 38 (10), 1949.
Embling, C.B., Gillibrand, P.A., Gordon, J., Shrimpton,
J., Stevick, P.T. and Hammond, P.S. 2010 Us-
ing habitat models to identify suitable sites for
marine protected areas for harbour porpoises (Pho-
coena phocoena). Biological Conservation 143, 26779.
Englund, A., Ingram, S. and Rogan, E. 2007 Popula-
tion status report for bottlenose dolphins using the
Lower River Shannon SAC, 20062007. Unpub-
lished Final Report to the National Parks and
Wildlife Service.
European Community 2001 Interpretation note: Se-
lection of sites for the species of Annex II Phocoena
phocoena. Habitats Committee Hab. 1/05.
Evans, P.G.H. and Hammond, P.S. 2004 Monitoring
cetaceans in European waters. Mammal Review 34
(1), 13156.
Gilles, A. Adler, S., Kaschner, K., Scheidat and M.,
Siebert, U. 2011 Modelling harbour porpoise
seasonal density as a function of the German Bight
environment: implications for management. Endan-
gered Species Research 14,15769.
Hammond, P. S., Benke, H., Berggren, P., Borchers,
D.L., Buckland, S.T., Collet, A., Heide-Jorgensen,
M.P., Heimlich-Boran, S., Hiby, A.R., Leopold,
M.F. and Oien, N. 2002 Abundance of harbour
porpoise and other cetaceans in the North Sea and
adjacent waters. Journal of Applied Ecology 39, 361
76.
Jenkins, R.E., Brown, R.D.H. and Phillips, M.R.
2009 Harbour porpoise (Phocoena phocoena) con-
servation management: a dimensional approach.
Marine Policy 33, 7449.
Leopold, M. F., Wolf, P.A. and Van de Meer, J.
1992 The elusive harbour porpoise exposed: strip
transect counts off southwestern, Ireland. Netherlands
Journal of Sea Research 29 (4), 395402.
Marubini, F., Gimona, A., Evans, P.G.H., Wright, P.J.
and Pierce, G.J. 2009 Habitat preferences and
interannual variability in occurrence of the harbor
porpoise Phocoena phocoena off northwest Scotland.
Marine Ecology Progress Series 381, 297310.
15
D
ENSITY ESTIMATES OF HARBOUR PORPOISES AT
I
RISH COASTAL SITES
O’Brien, J. 2009 The inshore distribution and abun-
dance of small cetaceans on the west coast of Ireland:
site assessment for SAC designation and an evalua-
tion of monitoring techniques. Unpublished PhD
thesis, GalwayMayo Institute of Technology.
O’Brien, J.M., Berrow, S.D., Ryan, C, McGrath, D.,
O’Connor, I., Pesante, P., Burrows, G., Massett,
N., Klo
¨tzer, V. and Whooley, P. 2009 A note
on long-distance matches of bottlenose dolphins
(Tursiops truncatus) around the Irish coast using
photo-identification. Journal of Cetacean Research
and Management 11 (1), 716.
O’Connell, M. and Berrow, S. 2010 Records from
the Irish Whale and Dolphin Group. Irish Natural-
ists’ Journal 31 (1), 5055.
Reid, J. B., Evans, P.G.H. and Northridge, S.P.
2003 Atlas of cetacean distribution in north-west
European waters. Peterborough. Joint Nature Con-
servation Committee 175.
Rogan, E. and Berrow, S.D. 1996 Review of har-
bour porpoises Phocoena phocoena L. in Irish waters.
Report of the International Whaling Commission 46,
595605.
Ross, H. and Wilson, B. 1996 Violent interactions
between bottlenose dolphins and harbour porpoises.
Proceedings of the Royal Society (London) 263, 2836.
SCANS-II 2008 Small Cetaceans in the European
Atlantic and North Sea. Final Report to the
European Commission under Project LIFE04-
NAT/GB/000245. Available from SMRU, Gatty
Marine Laboratory, University of St Andrews,
St Andrews, Fife, KY16 8LB, UK.
Sonntag, R.P., Benke, H., Hiby, A.R., Lick, R. and
Adelung, D. 1999 Identification of the first
harbour porpoise (Phocoena phocoena) calving ground
in the North Sea. Journal of Sea Research 41, 22532.
Teilmann, J. 2003 Influence of sea state on density
estimates of harbour porpoises (Phocoena phocoena).
Journal of Cetacean Research and Management 5(1),
8592.
Tougaard, J., Carstensen, J., Teilmann, J., Skov, H. and
Rasmussen, P. 2009 Pile driving zone of respon-
siveness extends beyond 20km from harbour por-
poise (Phocoena phocoena (L.)). Journal of the Acoustical
Society of America 126 (1), 1114.
VerfußU.K., Meding A., Honnef C.G., Da
¨hne, M. and
Benke H. 2007 Geographical and seasonal var-
iation of harbour porpoise (Phocoena phocoena) pre-
sence in the German Baltic Sea revealed by passive
acoustic monitoring. Journal of the Marine Biological
Association of the United Kingdom 87, 16576.
16
B
IOLOGY AND
E
NVIRONMENT
