101
New marine IBAs for the Mediterranean Shag Phalacrocorax 
aristotelis desmarestii in Slovenia
Nova morska območja IBA za sredozemskega vranjeka Phalacrocorax aristotelis 
desmarestii v Sloveniji
Urška Koce, DOPPS – BirdLife Slovenia, Tržaška cesta 2, 1000 Ljubljana, e-mail: ursa.koce@dopps.si
The IBA network is being regularly updated, based on new data and their 
better quality. There have been three previous stages of the marine IBA 
identification in Slovenia and the Mediterranean Shag has been included as 
the qualifying species only in the most recent stage in 2011. However, the 
sites were limited to inshore coastal roost-sites and thus insufficient to cover 
the foraging areas of the species. To fill this gap in the Slovenian territorial 
sea, new marine IBAs have been identified for the Mediterranean Shag 
within the scope of the SIMARINE-NATURA (LIFE10NAT/SI/141) 
project in the 2011–2015 period. The new sites were identified following 
standardized methodology for the identification of marine IBAs from 
BirdLife International. The data on the Mediterranean Shag distribution 
and population size were collected using four field methods: (1) monthly 
monitoring at in-shore communal roost-sites, (2) monthly monitoring at 
sea following the standardized ESAS method, (3) GPS telemetry, and (4) 
unsystematic census of in-shore floating groups. Based on these data, one 
new site, the IBA Osrednji Tržaški zaliv, and one extension to the existing 
IBA Debeli rtič were identified, covering 8,218 ha and 155 ha, respectively. 
The new sites cover 39.2% of the Slovenian territorial sea.
Acrocephalus 39 (178/179): 101–128, 2018
10.1515/acro-2018-0009
1. Introduction
1.1. Marine IBAs in Slovenia
Conservation of marine biodiversity is widely 
implemented through IBA programme – Important 
Bird and Biodiversity Areas, established by BirdLife 
International. The main objectives of the programme 
are identification, protection and management of 
worldwide network of sites of crucial importance for 
the long-term conservation of wild bird populations 
(BirdLife International 2010c). In contrast to 
the birds of the terrestrial environment, the seabird 
populations were poorly assessed until recently due to 
difficult access to the ornithological data in the marine 
environment, and this was reflected in a low number 
of identified marine IBAs, especially in off-shore areas; 
however, the number of BirdLife partners (corre-
sponding to number of countries) involved in marine 
IBA identification and protection increased to over 40 
by 2010 (BirdLife International 2010a, c). 
The IBA network is being updated regularly based 
on new data and their better quality. There have been 
three previous stages of IBA identification in Slovenia 
involving seabirds, however, the resulting marine 
IBA and Natura 2000 network has been assessed as 
insufficient by BirdLife International (BirdLife 
International 2014). Common characteristic 
of previous marine IBA identification is that it was 
part of the pan-national IBA identification and / or 
revisions, focused on terrestrial sites (Božič 2003, 
Denac et al. 2011, Polak 2000). Within these 
stages, four marine IBAs covering six species of 
seabirds were identified in the 2000–2011 period 
(Ibid.) The fourth (and most recent) stage of marine 
IBA identification in Slovenia, presented in this work, 
102
Shags in the Gulf of Trieste coincides also with the 
swing of shellfish farming on filamentous floating 
farms that enable the shags’ undisturbed roosting 
on otherwise highly populated coast of the Gulf. 
There are three such communal roost-sites along 
the Slovenian coast, near Debeli rtič, Strunjan and 
Sečoveljske soline (Bordjan et al. 2013, Koce & 
Lipej 2016). All sites have already been identified 
as marine IBAs and included in the Natura 2000 
network (Denac et al. 2011, Koce & Lipej 2016).
The Mediterranean Shags forage predominantly 
on fish (Bazin & Imbert 2012, Cosolo et al. 
2011, Lipej et al. 2016). They are foot propelled 
pursuit-divers, usually preying near sea-bed, but 
occasionally also in pelagic waters, especially in 
shallow coastal belt (Bazin & Imbert 2012). 
Their foraging dives are regularly 30 m deep, often 
up to 60 m and in extreme cases even up to 80 m 
(Bazin & Imbert 2012) and can last over one 
minute (Sponza et al. 2010). They usually forage 
solitarily or in small groups independently from 
one another, but they also socialize in large groups 
of up to few hundred individuals and communally 
prey on schools of small pelagic fish in coastal 
shoals (Nelson 2005), sometimes in the company 
of other seabird species, e.g. the Black-headed Gulls 
Chroicocephalus ridibundus (pers. observation). 
This is a common phenomenon in the marine 
environment known as multi species foraging 
groups (Camphuysen & Garthe 2004).
The Mediterranean Shag is a dietary opportunist 
preying on the commonest and most easily accessible 
prey in the benthic zone. The majority of its prey in 
the Adriatic sea represent economically unimportant 
fish species: around the Croatian breeding sites 
(Oruda Island), the diet consisted predominantly of 
big-scale sand smelt Atherina boyeri, brown comber 
Seranus hepatus and peacock wrasse Crenilabrus 
tinca, however, in the Gulf of Trieste their main food 
is black goby Gobius niger (Cosolo et al. 2011, Lipej 
et al. 2016). 
As a species, the European Shag (Phalacrocorax 
aristotelis) is not endangered by the IUCN criteria and 
has a favourable conservation status at the European 
level (non-SPEC) (BirdLife International 
2004). The status assessment after SPEC classification 
for the Mediterranean subspecies was not made. 
Although historical population estimates for the 
Mediterranean Shag are rather poor, the experts 
was implemented in the years 2011–2015 as part of 
the project SIMARINE-NATURA (LIFE10NAT/
SI/141) (www.simarine-natura.ptice.si). Contrary to 
previous stages when seabird censuses were limited 
to coastal and inshore areas, the monitoring of target 
species, i.e. the Mediterranean Shag Phalacrocorax 
aristotelis desmarestii, was conducted systematically 
across the entire range of the Slovenian territorial sea.
1.2. Target species
The Mediterranean Shag belongs to the 
cormorant family (Phalacrocoracidae). It is 
distributed solely in the Mediterranean and the 
Black Sea. Their entire breeding population is 
estimated at 8,700–11,130 pairs, breeding in over 
400 colonies (Bazin & Imbert 2012). The largest 
national breeding numbers are found in Croatia, 
Italy, Greece and France, ranging from 1,000 to 
2,000 pairs (Bazin & Imbert 2012). The total non-
breeding population of the Mediterranean Shags 
is estimated at 30,000 individuals (Wetlands 
International 2004). In the post-breeding 
season, part of the population spreads out of their 
breeding area, to the non-breeding areas up to 
several hundred kilometres from their breeding 
colonies (Škornik et al. 2011). The majority of the 
Adriatic population, almost entirely breeding in 
Croatian waters, migrates to the Gulf of Venezia 
(Sponza et al. 2013). In summer and autumn, 
only the Gulf of Trieste hosts around 6,000 
and exceptionally even up to 10,000 individuals 
(Škornik et al. 2011), representing 20–33% of 
the total non-breeding population (Wetlands 
International 2004). The roosting population 
along the Slovenian coast was estimated at 2,000–
3,000 individuals in the 2006–2011 period (Denac 
et al. 2011). Regular seasonal migration from the 
Croatian breeding areas to the Gulf of Trieste 
was established fairly recently, in the 1980s, and 
became massive at the end of the 1990s (Škornik 
et al. 2011). The onset of post-breeding migration 
to the north coincides with the period of extensive 
overfishing in the Croatian waters and could be 
triggered by the lack of food around the breeding 
colonies and facilitated by high foraging efficiency 
in the shallow Gulf of Trieste, abundant with easy 
accessible fish prey (Sponza et al. 2010). Moreover, 
the increase in the numbers of the Mediterranean 
U.  Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
103
agree that its population has decreased (Aguilar 
& Fernández 2002). Due to the limited range and 
small population, the Mediterranean Shag is listed 
on the Annex I of the Bird Directive, Annex II of the 
Bern Convention and on the list of priority species of 
the Barcelona Convention.
2. Methods
2.1. Target geographic area
Within the scope of the SIMARINE-NATURA 
project (2011–2016), the target geographic area 
for the identification of marine IBAs was defined 
as part of the then-claimed Slovenian territorial 
sea within the Gulf of Trieste (hereafter referred 
to as ‘project area’). The project area was selected 
based on previous data on importance of the Gulf 
of Trieste for the post-breeding population of the 
Mediterranean Shags (Vrezec 2006b). However, 
the Slovenian territorial sea was redefined in 2017 
as a result of the arbitration between Slovenia and 
Croatia. As the present Slovenian territorial sea 
overlaps very well with the project area it represents 
the target geographic area in this work (Figure 1). 
Surface area of the target area, i.e. the Slovenian 
territorial sea, is 214  km2. Surface area of the 
project area is 211 km2.
2.2. Marine IBA identification methodology
The identification of marine IBAs in the Slovenian 
territorial sea was conducted following the protocol 
proposed by BirdLife International (2010b). 
The protocol recommends nine analytical steps: (1) 
identification of target species; (2) seabird and envi-
ronmental data gathering from all possible sources, 
including systematic surveys, existing databases, 
reports and published work; (3) spatial data analysis 
resulting in GIS layers based on different data 
sources, species by species and organization of data 
for comparison between different months/seasons/
years; (4) classification of data layers based on data 
quality to primary and supplementary; (5) identifi-
cation of candidate sites for each seabird species; (6) 
application of standard IBA criteria and confirma-
tion that the IBA candidates sites comply with the 
criteria; (7) delineation of final boundaries. Steps 
5 and 7 were merged in our study because only one 
target species was defined. The IBA proposals with 
the descriptions are entered into the “World Birds 
and Biodiversity Database”  (WBDB) managed 
by BirdLife International (https://www.global-
conservation.info/). As the final step, the BirdLife 
Secretariat reviews the proposals and decides about 
the confirmation of proposed IBAs. 
2.2.1. Identification of target species
According to BirdLife International 
(2010b), about 340 extant species are categorized 
as seabirds and theoretically all of them are suitable 
for the analysis in the process of the marine IBA 
identification. However, for practical reasons it is 
recommended that one or some priority species 
are defined based on their conservation status, 
data availability, and expected distribution at 
sea, etc. (BirdLife International 2010b). In 
Slovenia, 42 seabirds have been recorded in coastal 
and marine environment since 1950 (Koce & 
Lipej 2016), six of which met the criteria for 
either identification of new IBAs or inclusion in 
the existing IBAs in previous stages of the IBA 
identification (Božič 2003, Denac et al. 2011, 
Koce & Lipej 2016, Polak 2000). In opposition 
to previous stages when the identification of 
marine IBAs was limited to coastal and inshore 
areas and was focused on breeding colonies and 
communal roost-sites, the focus in the present 
stage was an identification of offshore marine 
IBAs with emphasis on species’ foraging areas. 
Considering these recommendations from 
BirdLife International (2010b), the 
Mediterranean Shag was selected as the only target 
species, based on the following characteristics: (1) 
it is listed in the Annex I of the Birds Directive, 
Annex II of the Bern Convention and among 
priority species of the Barcelona Convention, 
(2) it has a negative population trend in a 
geographically limited range (Bazin & Imbert 
2012), (3) it has a relatively large population in the 
target area (Vrezec 2006a), (4) it has expected 
concentrations rather than extremely dispersed 
distribution at offshore sea (Vrezec 2006b), (5) 
it qualifies as an umbrella species in the target 
marine environment (Koce & Lipej 2016). 
Other seabird species occurring at Slovenian sea 
at present failed to meet one or more of these 
Acrocephalus 39 (178/179): 101–128, 2018
104
criteria and were thus not selected as priority 
species for the identification of marine IBAs in 
this stage. Moreover, the Mediterranean Shag was 
a priority species for financing by LIFE, the EU 
financial instrument supporting the conservation 
projects for Natura 2000 qualifying species that 
enhanced the marine IBA identification process 
in the 2007–2013 programming period, including 
through the SIMARINE-NATURA project. 
2.2.2. Data gathering
The data on the Mediterranean Shag distribution 
and population size were collected in the years 
2011–2014, using four field methods: (1) monthly 
monitoring at in-shore communal roost-sites, (2) 
monthly monitoring at sea following standardized 
ESAS method, (3) GPS telemetry, and (4) unsys-
tematic census of in-shore foraging groups.
2.2.2.1. Monitoring at communal roost-sites
Monitoring of roosting individuals was conducted 
to assess the size and temporal dynamics of the 
roosting population in the Slovenian sea. The 
monitoring was conducted between November 
2011 and October 2013 at three inshore locations 
previously known as their main communal 
roost-sites along the Slovenian coast. They are 
situated on floating infrastructure (buoys) of 
shellfish farms, located in the Bay of St. Bar-
tholomeus (Zaliv Sv. Jerneja), Strunjan Bay 
(Strunjanski zaliv) and Piran Bay (Piranski zaliv). 
The sites had been previously identified as marine 
IBAs Debeli rtič, Strunjan and Sečoveljske soline, 
respectively (Denac et al. 2011). The censuses 
of shags at the roost-sites were conducted once a 
month, synchronously at all three locations. Counts 
on one census occasion were done every half an 
Figure 1. Target geographic area (Slovenian territorial sea) for the identification of marine IBAs for the Mediterranean 
Shag (Phalacrocorax aristotelis desmarestii). Black dashed line on the bottom map denotes the non-overlapping part 
of the project area border as defined in the SIMARINE-NATURA (LIFE10NAT/SI/141) project.
U.  Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
105
hour, starting two hours before sunset and ending 
at the time when light conditions were too poor to 
continue counting. The shags were classified by age 
in two classes: adults and non-adults (juveniles and 
sub-adults). All shags within the area of a roost-site 
were counted, including individuals swimming 
in the sea among buoys, because most of them 
stayed within the roost-site and occupied buoys 
within short period of time. The roost-site in the 
Bay of St. Bartholomeus is divided by national 
border with Italy in two parts but the shags were 
counted on both sides of the border as the location 
is functionally one roost-site. Based on the results 
of the roost-site monitoring, the peak non-breeding 
season was defined as the period when the total 
number of individuals was higher than mid-range, 
calculated as 0.5 x (nmax + nmin.) (i.e. 763 individuals).
2.2.2.2. Monitoring at sea following ESAS 
method
Monitoring at sea was done every month between 
July 2012 and August 2013 following the standardi-
zed ESAS method (‘European Seabirds At Sea’), 
whereby the birds were counted along line-transect, 
using boat (Camphuysen & Garthe 2004). 
The transect line in this survey was placed on the 
project area in a pattern that ensured representativeness 
for the area (Figure 2). It covered an entire gradient 
of distances from the coast to the outer border of 
the territorial sea as well as longitude gradient. The 
total length of the transect line corresponded to the 
distance travelled by the survey boat in 4–5  hours 
(81.6  km), corresponding to the time available for 
an optimal at-sea census of the Mediterranean Shags 
in one day. The censuses were conducted during the 
diurnal period when most shags were expected to be 
active outside their roost-sites.
The boat was travelling at constant speed of 
10  knots. The birds were recorded within 5  minute 
intervals. Given constant speed of 10  knots, this 
corresponds to segments of length 1,540  m. In case 
the boat changed its direction or speed, an ongoing 
census interval was aborted and a new interval started 
at the location of the change. The birds were recorded 
on both sides of the transect line in two bands: 300 m 
inner band and an unlimited outer band. The instant 
counting area was limited to a distance of 300 m in 
front of the boat. The distance was estimated by the 
rule of thumb by each observer upon prior calibration 
at a measured distance 300  m from the coast. The 
birds which were in contact with water were counted 
continuously, whereas the flying birds were counted 
in snapshots in one minute intervals. The birds were 
detected with a naked eye; binoculars were used only 
for the identification of individual’s age class, whereby 
two classes were defined based on their plumage: non-
adults and adults.
The data were entered in a standardized ESAS 
monitoring form. The following types of data were 
recorded: (1) census metadata (date, boat name 
and type, names of observers, type of census, group 
of recorded birds), (2) data about census interval 
(geographic coordinates of the starting point, 
interval start and end time, sea state on the Beaufort 
scale, visibility, floating objects and matter), and 
(3) interval specific bird data (age class, transect 
band, number of individuals, movement direction, 
association with objects or non-avian species, 
behaviour, prey, multi-species group membership).
2.2.2.3. GPS telemetry
The shags were being trapped in autumn 2012 and 
in summers 2013–2014. Self-made clap-traps with 
remote triggering system were used. The traps were 
adapted for mounting on three different floating 
objects: vertical cylindrical buoys (r = 1 m), cubic 
rafts (a = 1  m) and horizontal cylindrical buoys 
(r = 0.4 m). The trap was triggered remotely by the 
user when one or more shags sat on the buoy or raft.
The traps were set at three sites along the Slovenian 
coast: (1) in Viližan Bay on the east side of the town 
of Izola and within two communal roost-sites of the 
Mediterranean Shags near Strunjan (2) and Sečovlje 
(3). At the Izola trapping site, the traps were mounted 
on four existing vertical cylindrical buoys, two at a 
time. At the Strunjan and Sečovlje trapping sites, two 
cubic self-made rafts were tied among the existing 
buoys of the shellfish farms where the shags usually 
rest, one at each site. In addition, three traps were set 
on the existing surrounding horizontal cylindrical 
buoys at the Strunjan trapping site. 
We tagged the shags with GPS-GSM loggers 
of Polish producer ECOTONE. Three units of 
DUCK-3  model and 26 units of SAKER model 
were used. 16 units of the SAKER model were 
filled with polyurethane and 10 units with resin. 
Acrocephalus 39 (178/179): 101–128, 2018
106
The dimensions of the DUCK-3  model were 
45 × 25 × 25 mm, weight = 40 g, and those of the 
SAKER model were 45 × 25 × 18 mm, weight = 25 
(series with polyurethane filling) or 40 g (series with 
resin filling). The devices were mounted on birds 
permanently, as backpacks with teflon ribbon straps.
The devices were supplied with energy through 
solar panels, which provided enough energy in good 
light conditions to fix a GPS position every half an 
hour. The data were transferred to server by SMSs 
after every 4 GPS fixes. They were accessible to the 
user by the Internet through a password protected 
web panel. Frequency and accuracy of GPS fixes, and 
operating hours of the loggers could be set remotely 
via the web panel. We tended to set the highest 
frequency of fixes during the day but sometimes it 
needed to be lowered due to low natural light levels 
and consequent need for energy savings. The loggers 
were switched off during the mid-night hours. 
2.2.2.4. Land-based census of large floating 
groups
In the years 2011–2014, the data on large floating 
groups of the Mediterranean Shags (≥ 10 individuals) 
were collected unsystematically based on participa-
tion of experienced volunteer observers. The data 
were reported by observers through online web 
form (http://simarine-natura.ptice.si/sodeluj/). To 
enhance the census effort, the public was also called 
to participate in data collection through different 
means of communication: DOPPS  – BirdLife 
Facebook page, SIMARINE-NATURA project 
website, leaflets and newspaper articles as well as 
personal communication.
2.2.3. Production of basic spatial data layers 
and ecological models
2.2.3.1. ESAS dataset
Filtering of the ESAS data. The aim of the analysis 
was to identify the areas important for foraging 
shags, hence we omitted the following data from 
the original ESAS dataset: flying individuals, as the 
shags never predate from the air, and the individuals 
resting along the shore or floating objects at sea. The 
individuals which were not observed during their 
foraging activity but were swimming in the sea were 
retained in the dataset, assuming that they have 
most likely been foraging at or near the locations 
where observed, based on the information from the 
literature and results of the GPS telemetry analysis 
in this study. According to Nelson (2005), the 
breeding shags fly directly to their foraging area and 
return to the colony in the same manner. According 
to the results of the GPS telemetry survey in this 
study this holds also in case of non-breeding shags 
that take foraging trips from communal roost-sites. 
We refer to the resulting dataset as ‘ESAS data 
subset 1’. Furthermore, we omitted the individuals 
which were associated with large (> 12  m) fishing 
boats, assuming that they were momentarily foraging 
on the discards and were thus not associated with 
their usual foraging habitat, i.e. marine benthos. We 
refer to the resulting dataset as ‘ESAS data subset 2’.
Calculation of shag densities. The densities of 
shags were calculated at the level of transect and at 
the level of transect segments (intervals) based on 
the number of individuals registered in the inner 
300  m band. They were calculated separately for 
each census occasion (denoted by date) and from 
each of the two ESAS data subsets.
Furthermore, a raster layer of mean densities 
in the peak non-breeding season at the transect 
was produced in ArcGIS 10.3 (ESRI 2014) with 
'Polyline to Raster' tool, based on each of the two 
ESAS data subsets. Linear vector layers consisting of 
transect segments with shag density as an attribute 
were used as input data. Rasterization was conducted 
separately for each census conducted in peak non-
breeding season. The final raster layer with mean 
densities was made using 'Cell Statistics' tool, using 
census-specific raster layers as input data.
Production of ecological models of the 
Mediterranean Shag distribution: To explain the 
distribution of the foraging Mediterranean Shags 
in the Slovenian territorial sea in peak non-breeding 
season, the number of shags in transect segments 
was modelled as a function of several ecological 
variables (Table 1). Two sets of models were fit 
each based on one of the ESAS data subsets (1, 2), 
whereby only the censuses from peak non-breeding 
season were included. The modelling was done in 
R (R Development Core Team 2018) using 
packages ‘MASS’ and ‘mgcv’ (Venables & Ripley 
2002, Wood 2004). GAMM models with negative 
binomial distribution of the response variable (number 
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
107
of shags in transect segment) were fit, following Zuur 
et al. 2009. Logarithmically transformed surface area 
of a transect segment was included in the models as an 
offset variable due to differing length of the transect 
segments. The potential effects of pseudo-replication 
due to repeated censuses at the same transect were 
taken into account with the inclusion of ‘census date’ 
random variable. The ecological data represented by 
a set of fixed variables were obtained from GIS layers 
at the intersections with central points of the transect 
segments, or recorded at sea during census (Table 1). 
The variable 'sea depth' was eliminated from further 
analysis because it was in strong correlation with the 
variable 'distance from the coast' (Pearson’s correlation 
coefficient = 0.7). The decision to eliminate the 
former instead the latter was based on the facts that 
(1) the bathymetry of the Slovenian sea is more or less 
uniform and most of the sea bottom belongs to the 
same depth class (20‒30 m), (2) the depth of the sea 
in the area is not a limiting factor for the shags. The 
variables describing the presence of fishing boats were 
included only in the models which were fit based on 
ESAS data subset 1. The best model in each set was 
selected based on AIC (Akaike information criterion) 
(Burnham & Anderson 2002)
2.2.3.2. GPS telemetry dataset
Filtering of the GPS data. The original dataset was 
cleared before it was used for the analysis. In this 
step the following GPS locations were omitted from 
the dataset: locations which were fixed (1) before 
the shag was released, (2) after the shag died, was 
suspected to be dead or severely injured, (3) when 
the shag was apparently on migration or at breeding 
site, and (4) locations with an obvious error (i.e. 
more than 100 m inland). The outcome data subset 
in this step is referred to as ‘clear dataset’. Moreover, 
the GPS locations in the ‘clear dataset’ were grouped 
by a date-individual grouping variable (i.e. GPS 
locations from one individual fixed in the same day 
were assigned unique and equal group number). 
Furthermore, the locations were classified into three 
daily periods according to the fixing time: (1) day, 
(2) twilight and (3) night, based on astronomical 
data about time of sunrise, sunset and beginning 
or end of nautical twilight in Ljubljana, Slovenia 
(http://www.timeanddate.com/sun/slovenia/
ljubljana). Furthermore, GPS locations of five shags 
which were tracked for less than three days were 
excluded from the analysis at this point. 
Identification of roost-sites along the 
Slovenian coast. Besides the three traditionally 
known main communal roost-sites on shellfish 
farms, the shags were suspected to regularly roost 
at several other locations along the Slovenian coast. 
These regularly used roost-sites were identified 
based on clusters of roosting locations of the 
tracked individuals. A roosting location of an 
individual shag was defined as the first fixed GPS 
location in a day (usually early in the morning 
before the shags leave their roost-sites). Roost-sites 
were then defined as clusters of at least five roosting 
locations not further than 500 m from one another 
Table 1. Independent variables used in the modelling of the Mediterranean Shag Phalacrocorax aristotelis desmarestii 
distribution in the Slovenian territorial sea 
Independent variable Variable type Level Data source
census date date census ESAS dataset
presence of small fishing boats (up to 12 m) binary transect segment ESAS dataset
presence of large fishing boats (above 12 m) 
<0,5 km away binary transect segment ESAS dataset
presence of large fishing boats (above 12 m) 
0,5–2 km away binary transect segment ESAS dataset
sea state (Beaufort scale) factorial transect segment ESAS dataset
sediment type factorial transect segment GIS layer (Geodetski inštitut Slovenije)
distance from the coast continuous transect segment calculated in ArcGIS
Acrocephalus 39 (178/179): 101–128, 2018
108
and delineated as minimal convex polygons. Only 
roost-sites along the Slovenian coast were taken 
into account and roosting events in the Italian and 
Croatian territorial sea were excluded from further 
analysis. A centre of each roost-site was defined as 
the mean mid-point of the cluster, which was used 
in the following steps of the analysis (referred to as 
a roost-site centre from hereon). 
Further filtering of the data. In this step, 
only those locations were retained in the dataset 
that belonged to individual shags roosting at the 
three main communal roost-sites and roost-sites 
identified during the previous step. Furthermore, 
in order to focus on the areas where the shags have 
been most likely foraging and not only resting, more 
locations were omitted: (1) diurnal locations within 
roost-sites, (2) GPS locations at other daytime 
resting places, or (3) artificial feeding locations 
(i.e. locations next to the fish market where at least 
two tagged shags have been known to be fed by fish 
sellers), (4) night and twilight locations outside 
roost-sites (probably erroneous fixes). We refer to 
the resulting data subset as ‘end dataset’. It was used 
during the following steps as an input dataset for 
the delineation of the marine IBA candidates.
2.2.3.3. Large floating groups dataset
The data were digitalized in ArcGIS 10.3 (ESRI 
2014) and the groups distinguished according to 
their foraging activity at the time of observation to 
foraging and non-foraging. 
2.2.4. Classification of spatial data layers
The spatial data layers were classified as primary or 
supplementary based on their quality, following the 
recommendations by BirdLife International 
(2010b). 
2.2.5. Identification of marine IBA candidates 
by spatial data layers
2.2.5.1. ESAS dataset
Selection of 5% transect segments with the 
highest densities. Identification of the marine 
IBA candidate was based on the densities of shags 
at transect segments, calculated from both ESAS 
subsets. Only the peak-nonbreeding season of the 
Mediterranean Shags was considered, as revealed 
by the results of the roost-sites monitoring (Table 
2). Following the recommendations by BirdLife 
International (2010b), 5% of transect segments 
with highest densities of shags were chosen for each 
census occasion (referred to as the best segments 
from hereon), separately for each ESAS subset (1, 2).
Selection of grid cells intersected by best 
transect segments. The target area was then 
overlaid with a grid of cells with side size 1,540 m 
(corresponding to the length of one standard 
transect segment, rounded to 10 m). The cells which 
were intersected by the segments identified during 
the previous step were selected from the grid. The 
selected cells were used as a basic framework for the 
delineation of the marine IBA candidate.
2.2.5.2. GPS telemetry dataset
The marine IBA candidates were identified 
following BirdLife International’s protocol for 
the identification of marine IBAs using seabird 
tracking data (BirdLife International 2013). 
The protocol assumes central-place foraging, 
whereby an individual’s roost-site was defined as 
the central place. The analytical method aims at the 
identification of core use areas (kernels) of tracked 
individuals on their round trips, taken from the 
roost-site to foraging areas and back to the roost-site 
(hereafter referred to as ‘round foraging trips’). A 
core use area identified for a round foraging trip 
represents a foraging site of the individual shag 
making that trip.
The protocol consists of eight analytical steps. 
In step 1, the GPS data from the ‘end dataset’ 
were grouped by roost-sites. We thus obtained 10 
subsets of ‘end dataset’, one for each roost-site, and 
from hereon we describe the method for analysing 
individual roost-site subsets. In step 2, each subset 
was further split to round foraging trips. A set 
of locations was considered to represent a trip 
when the shag moved at least a defined distance 
from the roost-site centre, stayed on its journey 
for at least 1 hour, and returned back to the same 
roost-site i.e. closer than a defined distance from 
the roost-site centre. The distances (referred to 
as inner and outer buffer) for each roost-site were 
defined according to the roost-site characteristics 
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
109
(Table 9). Furthermore, trips with less than six 
locations were automatically omitted from the 
analysis as they were below the numerical threshold 
for identification of core use areas. In step 3 the 
scale of interaction with the environment was 
calculated and further used in step 4 for fitting the 
kernels, i.e. identification of core use areas. In step 
5, a variance test was applied on the resulting set 
of kernels to check for significant site fidelity of 
any individual. In case site fidelity of one or more 
individuals was significant, only one randomly 
chosen core use area per individual was chosen 
for further analysis, to avoid bias due to pseudo 
replication. In step 6, bootstrapping was used to 
assess the representativeness of the dataset and to 
calculate the percentage of the roosting population 
utilizing the marine IBA candidate. If the dataset 
was not representative (bootstrap outcome < 70%), 
no IBA candidates could be identified. In step 7, the 
areas more intensively used by several individuals 
(i.e. marine IBA candidates) were identified, and in 
step 8 threshold values were applied to these areas, 
estimated as  % of roosting population that visits 
each area. Steps 2–8 were applied separately to each 
data group as defined in step 1.
2.2.5.3. Large floating groups dataset
The marine IBA candidates were delineated based 
on a 1,000 m buffer around locations of the groups 
that were recorded at the time of active foraging. 
Additional criterion was used to include only the 
groups that were larger than the median group size, 
i.e. the threshold was set at minimum 130 individuals. 
2.2.6. Delineation of final marine IBA 
proposals
The delineation of final marine IBA proposals 
was done based on integration of IBA candidates 
identified from different source data layers. The 
strength of each IBA proposal was evaluated 
against the rules outlined in BirdLife Interna-
tional (2010b). According to these rules, the most 
defendable cases of marine IBAs are those identified 
based on two primary data layers, followed by the 
cases identified based on one primary and one 
supplementary data layer, and the cases identified 
based on one primary data layer. The cases 
identified based on two supplementary data layers 
are defendable only exceptionally, depending on 
the data involved, and the cases identified based on 
one supplementary data layer cannot be confirmed.
2.2.7. Application of IBA criteria
2.2.7.1. Relevant IBA criteria
BirdLife International lists 20 criteria for the iden-
tification of IBAs (http://datazone.birdlife.org/
site/ibacriteria), nine of which apply to the IBAs in 
the marine environment. The marine IBA proposals 
for the Mediterranean Shag in the Slovenian sea 
have been evaluated against three relevant criteria:
B1ii. Congregations: The site is known or 
thought to hold ≥ 1% of a distinct population of a 
seabird species.
C2. Concentrations of a species threatened at 
the European Union level: The site is known to 
regularly hold at least 1% of a flyway population 
or of the EU population of a species threatened at 
the EU level (listed on Annex I and referred to in 
Article 4.1 of the EC Birds Directive).
C6. Species threatened at the European Union 
level: The site is one of the five most important in 
the European region (NUTS region) in question 
for a species or subspecies considered threatened in 
the European Union (i.e. listed in Annex I of the 
EC Birds Directive).
2.2.7.2. Numerical thresholds
Two of the relevant IBA criteria, B1ii and C2, 
prescribe a numerical threshold that needs to be 
reached in the proposed IBA, i.e. the site has to 
hold at least 1% of migratory or biogeographic 
population (http://datazone.birdlife.org/site/
ibacriteria). Based on the population estimates 
by Wetlands International (2004) for the 
non-breeding population of the Mediterranean 
Shag, this threshold is at 300 individuals. To 
check whether the numerical threshold has been 
reached in the marine IBA proposals, the number 
of individuals using the sites has been assessed for 
each proposal, using the data from different sources 
that were used for their delineation. 
IBA Osrednji Tržaški zaliv. The number of 
individuals using the site was estimated based on 
Acrocephalus 39 (178/179): 101–128, 2018
110
(1) the extrapolation from transect densities and 
(2) model predictions. 
In case of the extrapolation, the ESAS dataset 1 
was used as input data, i.e. the number of all floating 
shags in the area was estimated. The number of 
individuals was calculated within the target area 
and within the marine IBA proposal, for each 
census occasion separately, by multiplying density of 
shags at the transect with the surface area of either 
site. In case of the IBA estimates, only the part of 
the transect overlapping the IBA was considered 
for the calculation of shags’ density. To check 
whether the number of shags reached the numerical 
threshold required in IBA criteria C2 and B1ii over 
two different years in at least one season, we tested 
whether the mean population estimates in the IBA 
proposal in meteorological summer (June to August) 
of both census years exceeded the threshold of 300 
individuals, using one tailed Student t-test. The 
assumption of the test that the sample population 
size estimates need to be normally distributed was 
checked and confirmed with Shapiro-Wilcoxon test 
of normality (W = 0.98, p = 0.93).
The model predictions were done based on 
a new dataset, represented by a grid of cells that 
was used for the delineation of the marine IBA 
candidate. The grid was clipped to exactly match 
the target area or the area of the IBA proposal, 
resulting in several partial grid cells at the edges. 
To account for the different cell surface areas, 
logarithmically transformed surface area of cells 
was calculated (i.e. representing the offset variable 
as used in the models). The values of the variable 
‘distance from the coast’ were obtained in ArcGIS 
10.3 (ESRI 2014) as the distance of the cell central 
point from the coastal line. To simulate the effect of 
fishing boats variable ‘presence of large fishing boats 
<0.5  km away’ was set to “yes” in one randomly 
chosen cell, and variable ‘presence of large fishing 
boats 0.5–2  km away’ in two randomly chosen 
cells. The sea state was set to level zero to account 
for the optimal monitoring conditions. Any non-
significant estimates were interpreted as not 
different from zero (i.e. the predictor has no effect). 
The mean predicted values and standard errors (SE) 
were calculated using ‘predict.gam’ function in R 
package ‘mgcv’ (Wood 2004), and 90% confidence 
intervals were calculated by the equation: mean ± 
1,645 x SE. All values were calculated on the scale 
of the linear predictor and then transformed to the 
scale of the response variable, i.e. the number of 
individuals present in a grid cell. The total number 
of individuals in the IBA proposal was calculated as 
a sum of individuals across the cells which intersect 
the area of the site.
Extension to IBA Debeli rtič. The number of 
birds visiting the area was assessed based on the census 
of large foraging groups and on the estimated number 
of individuals using the IBA candidate 1A, identified 
in the analysis of GPS telemetry data (Table 10). 
2.2.7.3. Regularity of use of the area
Besides the fact that the numerical thresholds have 
been reached it needs to be demonstrated that the 
site is being regularly used by the target species. 
According to the BirdLife International protocol for 
the identification of marine IBAs this means that the 
birds are visiting the area in different periods (seasons 
or years) or that the area is visited by the birds from 
different sites (i.e. breeding colonies or roost-sites) 
(BirdLife International 2010b). 
3. Results
3.1. Monitoring at communal roost-sites
Altogether, 24 monthly censuses were conducted at 
three communal roost-sites of the Mediterranean 
Shags between November 2011 and the end of 
October 2012 (Table 2). Their number peaked in 
August 2013 when 1,494 individuals were counted. 
The minimum was observed in March 2012 when 
only 32 individuals were counted.
3.2. Monitoring at sea following the ESAS method
Altogether, 16 censuses were conducted between 
July 2012 and August 2013 (Table 3). The length 
of a standard transect was 81.58 km and the width 
was 600 m. The mean census time was 4:31. Two 
censuses (22.6.2012 and 21.5.2013) were not 
conducted entirely due to bad weather conditions 
and hence omitted from further analysis.
Densities of the Mediterranean Shags at the 
transect. The density of shags at the transect varied 
greatly between censuses (Table 14). The highest 
density at the transect was observed in July 2013 
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
111
when it reached 4.4 ind./km2 (all floating shags) or 
3.4 ind./km2 (floating shags without those associated 
with fishing boats) (Table 14). The lowest densities 
were observed in January, February and March 2013 
when less than 10 floating individuals were observed 
at the transect (Table 14). The mean density of shags 
in peak non-breeding season (data pooled for years 
2012 and 2013) was generally higher off-shore than 
in-shore whether the shags associated with fishing 
boats were considered or not (Figure 2). 
Ecological models of the Mediterranean 
Shag distribution. Altogether, seven (set 1) and five 
(set 2) models were fit to describe the distribution 
of the Mediterranean Shags in the Slovenian sea 
(Table 4). The selected, most parsimonious model 
in set 1 (M1.5) contained five fixed variables. Four 
of these variables had significant coefficients: the 
‘distance from the coast’, ‘presence of large fishing 
boats <0.5 km away’, ‘presence of large fishing boats 
0.5–2 km away’ and the ‘sea state’. All but the latter 
had positive effects on the number of shags in the 
transect segments (Table 5). The selected, most 
parsimonious model in set 2 (M2.5) contained two 
fixed variables, the ‘distance from the coast’ and 
Table 2. The number of roosting Mediterranean Shags Phalacrocorax aristotelis desmarestii at three main communal 
roost-sites on shellfish farms along the Slovenian coast in 2011–2013; the difference between the total number of 
individuals (Total ind.) and the sum of adults (Ad.) and non-adults (Non-ad.) equals the number of individuals of 
unknown age. The shadowed censuses define the peak non-breeding season.
Date
All roost-
sites Debeli rtič Strunjan Sečovlje
Total ind. Ad. 
Non-
ad.
Total 
ind. Ad. 
Non-
ad. 
Total 
ind. Ad.
Non-
ad.
Total 
ind.
22.11.2011 725 344 53 328
21.12.2011 485 141 32 173 10 34 44 268
14.1.2012 172 3 27 32 18 122
16.2.2012 90 6 14 22 0 46 46 7 15 22
15.3.2012 32 3 7 10 12 0 10 10
18.4.2012 89 33 2 12 14 1 41 42
16.5.2012 170 75 3 39 42 2 51 53
20.6.2012 1047 373 191 483
16.7.2012 1485 330 174 169 380 775
13.8.2012 1406 371 309 726
17.9.2012 1241 531 276 434
13.10.2012 1261 532 209 520
15.11.2012 691 357 108 226
19.12.2012 148 68 20 88 31 29
13.1.2013 56 29 8 19
19.2.2013 39 2 14 20 0 2 2 0 17 17
22.3.2013 114 0 17 17 0 18 18 79
17.4.2013 82 28 0 6 6 48
21.5.2013 395 128 37 165 83 147
17.6.2013 905 337 18 355 195 355
17.7.2013 1357 450 255 652
23.8.2013 1494 577 33 610 269 615
17.9.2013 1424 529 267 628
24.10.2013 1215 538 38 576 120 519
Acrocephalus 39 (178/179): 101–128, 2018
112
‘sea state’, both having significant coefficients with 
similar effects as in case of M1.5.
Identification of marine IBA candidates. 
The density of shags in top 5% transect segments 
(i.e. two segments per census occasion) spanned 
between 4.3 and 72.2 ind./km2 in case of all 
floating shags were considered (ESAS data 
subset 1), and between 4.3 and 28.7 ind./km2 in 
case of floating shags without those associated 
with fishing boats (ESAS data subset 2) (Table 6). 
The spatial distribution of the selected segments 
was very similar for both ESAS data subsets, 
resulting in equal distribution of the intersecting 
grid cells that represent the spatial framework 
for the delineation of the marine IBA candidate 
(Figure 3).
3.3. GPS telemetry
A total of 29 shags were tagged with GPS loggers 
between 3.10.2012 and 27.8.2014: 27 at the Izola 
trapping site and two at the Strunjan trapping site. 
There was no trapping success at the Sečovlje site. 
Two, 15, and 12 Shags were tagged in 2012, 2013 and 
2014, respectively. In 2013, several devices stopped 
working soon after they were mounted on birds due 
to technical faultiness. The data from the devices 
that stopped working less than three days after they 
were mounted on birds (five devices) were excluded 
from further analysis. Data from 24 birds were thus 
analysed for the purpose of marine IBA identification. 
The tracking data used in the analysis were 
obtained between 3.10.2012 and 30.9.2014. The 
number of tracking days in ‘clear dataset’ differed 
highly among shags (n = 24): from minimum of 
five to maximum of 445 days (mean: 78) and so did 
the number of fixed GPS locations: from minimum 
of 168 to maximum of 8,380 (mean: 1,653). The 
number of tracking days in ‘end dataset’ ranged 
from minimum of 5 to maximum of 437 days 
(mean: 70) and so did the number of fixed GPS 
locations: from minimum of 67 to maximum of 
3,449 (mean: 608).
The core use areas (kernels) were fit for 21 of 24 
individuals. The number of GPS locations in each 
round foraging trip of three individuals (Andro, 
Table 3. Summary of ESAS censuses conducted in the Slovenian sea in 2012 and 2013. Censuses marked with 
asterisk were conducted within the peak non-breeding season. Duration of census: total census duration including 
short pauses.
Census date
Length of 
censused 
transect[km]
% of 
censused 
transect
Census 
start time 
(h:mm)
Census 
end time 
(h:mm)
Duration 
of census 
(h:mm)
No. of 
transect 
segments
No. of 
transect 
segments < 
1000 m
22.6.2012* 74.73 91.6 7:03 11:30 4:27 54 6
17.7.2012* 81.58 100 7:15 11:41 4:26 62 12
7.8.2012* 81.58 100 7:53 12:29 4:36 60 11
23.8.2012* 81.58 100 8:47 13:43 4:56 60 12
7.9.2012* 81.58 100 9:00 13:54 4:54 62 12
18.10.2012* 81.58 100 11:09 15:56 4:47 59 11
16.11.2012 81.58 100 10:30 14:52 4:22 61 12
19.12.2012 81.58 100 10:18 15:05 4:47 60 12
23.1.2013 81.58 100 9:58 14:06 4:08 58 10
19.2.2013 81.58 100 11:05 15:36 4:31 62 13
22.3.2013 81.58 100 11:00 15:32 4:32 60 12
19.4.2013 81.58 100 11:10 15:31 4:21 60 12
21.5.2013 41.55 50.9 11:30 14:22 2:52 33 9
19.6.2013* 81.58 100 11:14 15:53 4:39 57 9
26.7.2013* 81.58 100 10:55 16:35 5:40 59 11
23.8.2013* 81.58 100 10:47 15:08 4:21 60 12
Total 1.258,4 72:19 927 176
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
113
Roko and Dino) was below the required threshold 
(i.e. six locations) to fit the kernels. The number of fit 
core use areas varied greatly between the individuals 
(1–112) (Table 7), as well as between roost-sites 
(1–160) (Table 8). Mean surface area of core use 
areas was 383.79 ± 222.56 ha. Mean distance from 
the coast, calculated as aerial distance of core use 
area central point from the corresponding roost-
Figure 2. Mean densities of the Mediterranean Shags Phalacrocorax aristotelis desmarestii (ind./km2) at the ESAS 
transect in peak non-breeding season; the data were pooled for the years 2012 and 2013. Left: density of all 
floating shags; right: density of floating shags without those associated with fishing boats.
Table 4. Models of the Mediterranean Shag Phalacrocorax aristotelis desmarestii distribution in the Slovenian territorial 
sea in peak non-breeding season; the data were pooled for the years 2012 and 2013. Set 1: the response variable 
was defined as the number of all floating individuals in a transect segment (ESAS data subset 1). Set 2: the response 
variable was defined as the number of floating individuals not associated with fishing boats in a transect segment (ESAS 
data subset 2). The selected models in each set are printed in bold. Fixed variables included in a given model are 
denoted with 'Y'. Distance: distance from the coast; Sfishb: presence of small (<12 m) fishing boats; Lfishb <0.5 km: 
presence of large (>12 m) fishing boats <0.5 km away; Lfishb 0.5–2 km: presence of large fishing boats 0.5–2 km 
away; Sediment: sediment type; Sea state: sea state on the Beaufort scale; *GAM was fit. **Model did not converge.
Set of 
models Model
Degrees of 
freedom AIC Δ AIC
Random 
variable
Date
Fixed variables
D
at
e
D
is
ta
nc
e
Sfi
sh
b
Lfi
sh
b 
<0
.5
 k
m
Lfi
sh
b 
0.
5–
2 
km
Se
di
m
en
t
Se
a s
ta
te
Set 1
M1.5 11 1439.4 0 census date Y Y Y Y Y
M1.6 8 1449.6 10.2 census date Y Y Y Y
M1.2 13 1451.6 12.2 census date Y Y Y Y Y Y
M1.1 17 1467.8 28.4 none* Y Y Y Y Y Y Y
M1.7 8 1536.9 97.5 census date Y Y
M1.3 5 1539.6 100.2 census date Y Y
M1.4** 7 1540.9 101.5 census date Y Y
Set 2
M2.5 8 1437.54 0 census date Y Y
M2.2 10 1449.35 11.81 census date Y Y Y
M2.3 5 1452.98 15.44 census date Y
M2.1 14 1457.27 19.73 none* Y Y Y Y
M2.4 7 1468.32 30.78 census date Y Y
Acrocephalus 39 (178/179): 101–128, 2018
114
site was 3.61 ± 2.3 km. Nine of 21 individuals were 
using more than one roost-site (Table 7).
Using the GPS telemetry data, eight marine 
IBA candidates were identified in total, based on 
the individuals making round foraging trips from 
roost-sites 1, 3 and 5 (Figure 6). No sites were 
identified based on foraging trips from other roost 
sites (Table 9). It was estimated that the maximum 
numbers of shags visiting individual areas ranged 
between 7 and 220 individuals (Table 10).
3.4. Census of large floating groups
A total of 40 cases of large floating groups (≥10 
individuals) of the Mediterranean Shags were 
reported between October 2011 and November 
2014 (Table 11, Figure 7). All reports coincide 
with the autumn season, the earliest being from 
21th September and the latest from 6th December. 
13 (32.5%) groups consisted of 130 or more 
individuals, the largest two groups holding 300 and 
310 individuals. 69.2 % of groups with 130 or more 
individuals and 40.7 % of groups with less than 130 
individuals were observed during active foraging, 
in several cases in the company of Black-headed 
Gulls (Chroicocephalus ridibundus). Most groups 
were foraging in the intertidal zone preying on 
schools of small pelagic fish.
3.5. Data layer classification
The spatial data layers produced based on the ESAS 
dataset were classified as primary and the spatial 
data layers produced based on the GPS telemetry 
data and census of large floating groups as supple-
mentary (Table 12). An additional environmental 
layer, i.e. the bathymetry layer, was used for a 
detailed delineation of final marine IBA borders. 
Table 5. Regression coefficients of the selected GAMM models describing distribution of the Mediterranean Shags 
Phalacrocorax aristotelis desmarestii at the Slovenian sea in peak non-breeding season; M1.5: the response variable was 
defined as the number of all floating individuals in a transect segment (ESAS data subset 1). M2.5: the response variable 
was defined as the number of floating individuals not associated with fishing boats in a transect segment (ESAS data 
subset 2). te: denoting the smoothing function used in the additive modelling. Sea state referential level: Beaufort level 0. 
Model Predictor Description ß p
M1.5
Intercept Model intercept -12.9 <0.001
te(Distance) Distance from the coast 1.22 <0.001
Sfishb Presence of small (<12 m) fishing boats 0.20 non-sig.
Lfishb <0.5 km Presence of large (>12 m) fishing boats 0.5 km away 2.74 <0.001
Lfishb 0.5‒2 km Presence of large (>12 m) fishing boats 0.5‒2 km away 0.87 <0.05
Sea state 
level = 1
Sea state on the Beaufort scale
-0.53 <0.05
level = 2 -0.89 <0.01
level = 3 -1.33 <0.01
level = 5 -0.77 non-sig.
M2.5
Intercept Model intercept -12.8 <0.001
te(Distance) Distance from the coast 1.33 <0.001
Sea state 
level = 1
Sea state on Beaufort scale
-0.59 <0.01
level = 2 -1.06 <0.001
level = 3 -1.50 <0.01
level = 5 -0.84 non-sig.
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
115
3.6. New marine IBA proposals and their 
compliance with the IBA criteria
3.6.1. Marine IBA proposals
The final marine IBA proposals for the Mediter-
ranean Shags were delineated based on the IBA 
candidates that were identified as part of the 
analysis of the three data sources: ESAS dataset 
(Figure 3), GPS telemetry dataset (Figure 6), 
and land-based observations of large foraging 
groups (Figure 7). One new marine IBA and 
one extension to an existing marine IBA were 
delineated (Figure 8). The IBA Osrednji Tržaški 
zaliv was delineated based on one primary data 
layer (Table 13). The extension to IBA Debeli rtič 
was delineated based on two overlapping supple-
mentary data layers and the bathymetry layer as an 
additional environmental data layer to limit the 
area to 0–10  m depth zone, as social foraging of 
shags on schools of small pelagic fish occurred most 
intensively in the shallow in-shore waters and the 
intertidal zone (Table 13). The new proposed areas 
and the previously identified IBAs represent four 
distinct IBAs for the Mediterranean Shag in the 
Slovenian territorial sea (Figure 8). Note that the 
marine IBA Škocjanski zatok does not include the 
Mediterranean Shag as a qualifying species.
Figure 3. Marine IBA candidate as identified based on the locations of the top 5% transect segments with the 
highest densities of the Mediterranean Shags Phalacrocorax aristotelis desmarestii in peak non-breeding season in 
2012 and 2013; light grey segments: year 2012; dark grey segments: year 2013; black segments: both years. 
Left: density of all floating shags; right: density of floating shags without those associated with fishing boats. 
The segments are overlaid with intersecting grid cells (a = 1,540 m), representing the spatial framework for the 
delineation of the marine IBA candidate.
Table 6. Densities of the Mediterranean Shags 
Phalacrocorax aristotelis desmarestii in the top 5% 
transect segments in peak non-breeding season by 
census occasion. The densities were calculated based 
on the number of all floating shags (ESAS data subset 
1) or the number of floating shags without those 
associated with fishing boats (ESAS data subset 2). 
*Transect segments differ between data subsets.
Census date
Mediterranean Shag density 
[ind./km2]
ESAS data 
subset 1 
ESAS data 
subset 2
17.7.2012
21.6 21.6
19.3 19.3
7.8.2012
56.3* 4.3*
13.0 13.0
23.8.2012
72.2* 7.9*
10.8 10.8
7.9.2012
5.4 5.4
4.3 4.3
18.10.2012
10.0 10.0
5.4 5.4
19.6.2013
26.0 24.9
15.2 15.2
26.7.2013
28.7 28.7
21.6 21.6
23.8.2013
19.5 19.5
11.9 11.9
Acrocephalus 39 (178/179): 101–128, 2018
116
Figure 4. GPS locations of 24 Mediterranean Shags Phalacrocorax aristotelis desmarestii within the Gulf of Trieste, 
tagged along the Slovenian coast and tracked between 3.10.2012–30.9.2014; grey dots: clear dataset (n=39,670); 
black dots: end dataset (n=14,595)
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
117
Figure 5. Core use areas of 21 of 24 tracked Mediterranean Shags Phalacrocorax aristotelis desmarestii making 
round foraging trips from roost-sites 1–10. Note that no core use areas were identified for three individuals (Andro, 
Dino and Roko). Numbers 1, 8 and 9 represent the main communal roost-sites along the Slovenian coast, Debeli 
rtič, Strunjan and Sečovlje, respectively.
Acrocephalus 39 (178/179): 101–128, 2018
118
Table 7. Core use area (CUA) statistics by individuals for 21 Mediterranean Shags Phalacrocorax aristotelis 
desmarestii. Note that no core use areas were identified for three individuals, Andro, Dino and Roko.
Individual No. of CUA
Mean surface 
area (ha) SD (ha)
Mean coast 
distance (km) SD (km)
No. of roost-
sites
Ante 1 147.24 NA 1.90 NA 1
Ari 3 457.79 77.48 1.34 1.36 1
Bepo 29 630.59 203.39 5.50 5.68 1
Dado 6 220.65 27.12 0.71 0.69 2
Ilija 4 871.15 392.35 11.62 8.65 1
Ivek 8 238.30 33.89 1.55 1.50 1
Jakomo 112 240.37 101.77 2.89 3.61 4
Karlo 44 595.18 164.78 6.59 6.74 2
Mihi 69 395.54 144.60 3.78 3.94 4
Momo 4 389.72 71.13 2.47 1.47 1
Nace 38 254.11 83.20 2.59 3.21 1
Nikola 63 587.95 209.13 4.50 4.38 3
Ogi 10 678.16 235.74 5.79 5.81 1
Oto 2 861.29 94.64 3.33 3.33 1
Pino 15 246.68 93.74 2.77 2.00 3
Srečko 26 332.43 77.60 1.75 1.56 4
Šime 5 329.61 169.54 2.67 2.87 1
Štelio 10 459.52 312.37 3.38 3.15 1
Tartini 10 527.98 179.13 2.93 2.62 1
Tonin 3 264.21 17.46 1.23 1.41 2
Ugo 55 183.21 68.87 2.58 1.69 3
Table 8. Core use area (CUA) statistics by roost-sites for 21 Mediterranean Shags Phalacrocorax aristotelis 
desmarestii. Note that no core use areas were identified for three individuals, Andro, Dino and Roko. 
Roost-site No. of CUA
Mean area 
(ha) SD (ha)
Mean coast 
distance (km) SD(km)
No. of 
individuals
1 119 426.76 223.89 4.70 2.38 9
2 1 169.12 NA 2.76 NA 1
3 160 420.45 220.79 3.82 1.81 7
4 8 246.32 36.12 1.31 0.45 2
5 93 243.58 122.23 1.33 1.21 7
6 8 335.92 30.25 1.69 0.58 1
7 10 385.34 79.38 2.13 0.54 1
8 103 412.11 246.93 4.10 1.51 8
9 8 584.18 406.85 9.20 5.04 2
10 7 272.95 128.14 4.29 1.78 1
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
119
Figure 6. Marine IBA candidates as identified in the analysis of the GPS telemetry data of 21 Mediterranean Shags 
Phalacrocorax aristotelis desmarestii tagged along the Slovenian coast and tracked between 3.10.2012–30.9.2014. 
Note that no candidates were identified for shags roosting at roost-sites 2, 4, 6–7 and 9–10. Letters A to D denote 
individual IBA candidates.
Table 9. Parameters and outcomes of the identification of marine IBA candidates based on the GPS telemetry data of 
21 Mediterranean Shags Phalacrocorax aristotelis desmarestii tagged along the Slovenian coast and tracked between 
3.10.2012–30.9.2014.
Roost-site
Inner 
buffer [km]
Outer 
buffer [km] Ars scale Site fidelity
Sample 
representa-
tiveness
Resulting marine 
IBA candidates
1 (Debeli rtič) 0.7 0.7 0.732 non-significant 97.42 1A–B
2 0.3 0.3 / sample too small
3 0.35 0.35 0.675 significant 79.8 3A–D
4 0.4 0.4 0.7 / sample too small
5 0.3 0.5 0.545 non-significant 99.31 5A–B
6 0.2 0.2 0.600 / sample too small
7 0.2 0.2 0.629 / sample too small
8 (Strunjan) 0.4 0.4 0.829 significant 56.07 sample not representative
9 (Sečovlje) 0.6 0.6 1.471 / sample too small
10 0.3 0.3 1.067 / sample too small
Acrocephalus 39 (178/179): 101–128, 2018
120
3.6.2. Estimated numbers of the 
Mediterranean Shags in proposed IBAs
Estimates based on extrapolation: Estimated 
number of shags in the target area was the highest in 
July 2013 and the lowest in March 2013 (Table 14). 
The estimated number of shags in the proposed 
IBA Osrednji Tržaški zaliv (OTZ) exceeded 300 
individuals (1% of biogeographic population) in 
5 out of 8 censuses within the peak non-breeding 
season, i.e. in July 2012, twice in August 2012, 
in June 2013 and July 2013 (Table 14). The mean 
number in summer months, pooled for both 
monitoring years, was 374 individuals (321–427, 
90% CI). The probability that the mean was lower 
than 300 individuals was below 5% (Student 
t = 2.8, df = 5, p = 0.02). 
Estimates based on the models: The 
estimated mean number of all f loating shags 
within the Slovenian territorial sea was 866 
(471–1,705, 90% CI), and 495 (265–988, 
90  % CI) within the IBA Osrednji Tržaški 
zaliv (predictions from the model M1.5). The 
estimated mean number of f loating shags without 
individuals associated with fishing boats was 
745 (472–1,187, 90% CI) within the Slovenian 
territorial sea and 415 (263–658, 90% CI) within 
the IBA Osrednji Tržaški zaliv (predictions 
from the model M2.5). Note that the predictions 
of the models account for the number of shags 
averaged over the entire peak non-breeding 
season and are thus not directly comparable with 
the estimates based on extrapolation (expressed 
as mean number of shags in summer months).
Table 10. Marine IBA candidates identified based on the GPS telemetry data of the Mediterranean Shags 
Phalacrocorax aristotelis desmarestii in the Slovenian sea, and estimates of the number of individuals utilizing each 
area. Roost-sites represent the central place of round foraging trips of the shags.
Roost-site Area ID Surface area [ha]
Roosting 
population
max. % of roosting 
population
max. mIBA site 
population
1
A 573
530—610 1
24.37 129‒149
B 888 36.13 191‒220
3
A 168
~100 2
14.29 14
B 1407 57.14 57
C 147 14.29 14
D 239 14.29 14
5
A 14
~50 2
12.9 7
B 291 78.49 39
1 Roost-site monitoring yearly maxima in 2012 and 2013; 
2 numbers assessed based on land observations (best expert opinion).
Table 11. The number of groups of the Mediterranean 
Shags Phalacrocorax aristotelis desmarestii along the 
Slovenian coast from October 2011 to December 2014 
by months and years; the data were reported by random 
observers; Total foraging: number of groups observed 
during active foraging. The number in brackets denotes 
the number of groups with ≥ 300 individuals. 
Month < 130 ind. ≥ 130 ind.
September 1 4
October 11 7(2)
November 14 2
December 1 0
Year < 130 ind. ≥ 130 ind.
2011 0 2(1)
2012 9 5
2013 10 3
2014 8 3(1)
Total 27 13(2)
Total foraging 11 9(2)
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
121
Figure 7. Large groups (≥ 10 individuals) of the Mediterranean Shags Phalacrocorax aristotelis desmarestii randomly 
registered between October 2011 and November 2014 along the Slovenian coast; black circles: the group was 
observed during active foraging; grey circles: the group was not actively foraging; the groups selected for delineation 
of marine IBA candidates are marked with asterisk. 
Table 12. Classification of the spatial data layers used for delineation of marine IBA candidates for the 
Mediterranean Shag Phalacrocorax aristotelis desmarestii in the Slovenian sea.
Data layer Data source Data layer class
1
local densities of shags in the peak non-
breeding season, i.e. top 5% transect segments 
with highest densities (the best segments) ESAS monitoring primary
2 grid cells (1,540 m) intersecting with the best transect segments
3 core use areas of individual tracked shags GPS telemetry supplementary
4 large foraging groups of shags unsystematic land-based census supplementary
5 bathymetry layer Geodetic Administration of the Republic of Slovenia (GURS) environmental
Acrocephalus 39 (178/179): 101–128, 2018
122
4. Discussion
The Slovenian territorial sea has been previously 
recognized as an area of international importance 
for the Mediterranean Shag in the non-breed-
ing season (Denac et al. 2011, Koce & Lipej 
2016), however, this work delivers important new 
knowledge about the spatio-temporal distribution 
of this seabird within the area. Two new marine 
IBAs for the Mediterranean Shag were identified 
based on the results of this study, which essentially 
complement the previously existing marine IBAs 
covering three main communal roost-sites on 
the shellfish farms located along the Slovenian 
coast (Denac et al. 2011). The main added value 
of the new sites to the conservation of the global 
non-breeding population of the Mediterranean 
Shag is that they encompass two areas regularly 
used for foraging by at least 1% of the population. 
The two new areas, IBA Osrednji Tržaški zaliv 
and the extension to the existing IBA Debeli rtič, 
differ in their ecological characteristics reflected in 
different ways the shags are using them. The IBA 
Osrednji Tržaški zaliv covers an off-shore area, 
spanning between approximately 3.5 and 10  km 
from the coast and lying within the 20–25 m depth 
zone, dominated by muddy and sandy benthic 
substrate and corresponding benthic community 
that is poorly investigated (Peterlin et al. 2013). 
We suggest that its importance for the foraging 
Figure 8. Marine IBAs in Slovenia; dashed line: new marine IBAs for the Mediterranean Shag Phalacrocorax 
aristotelis desmarestii identified within the scope of the SIMARINE-NATURA project (this work); dark grey polygons: 
marine IBAs identified during previous stages in 2000–2011 (denac et. al 2011). 1: IBA Osrednji Tržaški zaliv; 2: 
IBA Debeli rtič; 3: extension to IBA Debeli rtič; 4: IBA Strunjan; 5: IBA Sečoveljske soline; 6: IBA Škocjanski zatok. 
IBAs 2, 4 and 5 cover the three main roost-sites of the Mediterranean Shags along the Slovenian coast. The IBA 
Sečoveljske soline crosses the national border since it was identified before the marine border between Slovenia and 
Croatia was defined in the arbitration. Note that the IBA Škocjanski zatok does not include the Mediterranean Shag 
as a qualifying species.
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
123
Table 13. Marine IBAs proposed based on the data on the Mediterranean Shag Phalacrocorax aristotelis desmarestii 
distribution and numbers, gathered within the project SIMARINE-NATURA (LIFE10NAT/SI/141) in 2011‒2014.
Marine IBA 
proposal
Underlying 
data sources
Surface 
area (ha)
Estimated  
population size Regularity of use
Complies with 
IBA criteria
IBA Osrednji 
Tržaški zaliv – ESAS dataset 8.218 ha
374 (321‒427, 90% CI) 1 
495 (265‒988, 90% CI) 2
1% population 
threshold exceeded 
in two years (2012 
and 2013)
B1ii, C2, C6
Extension to 
IBA Debeli rtič
– GPS 
telemetry 
dataset
– large floating 
groups 
dataset
– bathymetry 
layer
155 ha up to 310 
3
129‒149 4
1% population 
threshold 
exceeded in two 
years (2011 and 
2014)
B1ii, C2, C6
1 Mean number of shags using the area in summer months over two years, estimated based on the extrapolation; 
2 mean number of shags using the area in peak non-breeding season, estimated based on the model M1.5; 
3 maximum number of shags observed within the same foraging group; 
4 estimated number of shags using the marine IBA candidate 1A, identified based on the GPS telemetry data.
Table 14. Estimated numbers of the Mediterranean Shags Phalacrocorax aristotelis desmarestii in the Slovenian 
territorial sea and in the IBA Osrednji Tržaški zaliv (OTZ), based on the extrapolation from the ESAS dataset. 
(1): Number of all floating shags, (2): number of floating shags without individuals associated with fishing boats. 
Results of the censuses from 22.6.2012 and 17.5.2013 were omitted due to unrepresentativeness of the sample. 
Censuses marked with asterisk were conducted within the peak non-breeding season.
M
et
eo
ro
lo
gi
ca
l 
se
as
on
C
en
su
s d
at
e
N
o.
 o
f S
ha
gs
at
 
th
e t
ra
ns
ec
t (
1)
N
o.
 o
f S
ha
gs
 at
 
th
e t
ra
ns
ec
t (
2)
D
en
si
ty
 o
f S
ha
gs
 
at
 th
e t
ra
ns
ec
t (
1)
 
[n
o.
/k
m
2 ]
D
en
si
ty
 o
f S
ha
gs
 
at
 th
e t
ra
ns
ec
t (
2)
[n
o.
/k
m
2 ]
Es
ti
m
at
ed
 n
o.
 o
f 
Sh
ag
s a
t S
LO
 (1
)
Es
ti
m
at
ed
 n
o.
 o
f 
Sh
ag
s i
n 
O
T
Z
 (1
)
Es
ti
m
at
ed
 n
o.
 o
f 
Sh
ag
s a
t S
LO
 (2
)
Es
ti
m
at
ed
 n
o.
 o
f 
Sh
ag
s i
n 
O
T
Z
 (2
)
Summer
17.7.2012* 131 131 2.7 2.7 572 349 572 349
7.8.2012* 112 60 2.3 1.2 489 392 262 184
23.8.2012* 145 95 3.0 1.9 633 425 415 225
Autumn
7.9.2012* 50 50 1.0 1.0 218 110 218 110
18.10.2012* 36 36 0.7 0.7 157 124 157 124
16.11.2012 21 13 0.4 0.3 92 26 57 26
Winter
19.12.2012 11 9 0.2 0.2 48 20 39 20
23.1.2013 7 7 0.1 0.1 31 4 31 4
19.2.2013 2 2 0.0 0.0 9 0 9 0
Spring
22.3.2013 1 1 0.0 0.0 4 0 4 0
19.4.2013 19 19 0.4 0.4 83 44 83 44
Summer
19.6.2013* 142 127 2.9 2.6 620 344 555 337
26.7.2013* 215 167 4.4 3.4 939 458 729 458
23.8.2013* 101 101 2.1 2.1 441 277 441 277
All year 71 58 1.4 1.2 310 184 255 154
Summer 141 114 2.9 2.3 616 374 496 305
Peak non-breeding season 117 96 2.4 2.0 509 310 419 258
Acrocephalus 39 (178/179): 101–128, 2018
124
shags is established on high abundance and easy 
access of suitable prey species, especially the black 
goby Gobius niger. The species has been identified as 
the dominant prey of the Mediterranean Shags in 
the Gulf of Trieste in two recent studies (Cosolo 
et al. 2011, Lipej et al. 2016). The black goby is a 
common benthic fish species throughout the Gulf 
of Trieste (Lipej et al. 2016), however, its spatio-
temporal dynamics within the Slovenian sea is 
not well known. Based on the existing data, it is 
not possible to conclude that the peak densities of 
the Mediterranean Shags, consistently found in 
off-shore areas within the scope of the systematic 
ESAS monitoring, reflect higher densities of prey in 
these areas, although this is a plausible hypothesis. 
Moreover, the distribution of shags at sea could also 
be affected by the level of disturbance but the effect 
of this factor was not considered in our study.
The IBA Osrednji Tržaški zaliv overlaps with 
the national B fishing zone (Klanjšček et al.), 
regularly exploited by bottom and midwater otter 
trawlers, purse seiners and pelagic pair trawlers 
(MKGP 2011). All these fishing techniques 
use vessels defined in this study as large fishing 
boats. Groups of shags have been occasionally 
observed to follow the boats and forage on the 
discards (own unpublished data and B. Marčeta, 
personal communication). This suggests that higher 
abundances of shags in the off-shore areas could 
be the consequence of fishing activities in these 
areas. In fact, the presence of large fishing boats in a 
transect segment or in its close vicinity (i.e. < 500 m 
from the instant observation point) had a strong 
positive effect on the number of shags, as revealed 
by one of the selected models fit for the ESAS 
dataset in this study. Nevertheless, a comparison 
of both selected models revealed that the number 
of shags within the transect segment increases with 
increasing distance from the coast independently 
of the presence of fishing boats. Moreover, the 
occasional discards probably do not represent an 
essential source of food for the Mediterranean Shag, 
as they are relatively scarce in this area (MKGP 
2011). We thus suggest that fishing activity is not an 
ultimate factor influencing the distribution of shags 
in the Slovenian sea, but it can cause temporary 
aggregations of the individuals that primarily visit 
the overlapping area to prey on the bottom dwelling 
prey species, predominantly the black goby. 
The IBA Osrednji Tržaški zaliv was identified 
based on one primary data source, the spatial 
distribution of shags, as revealed by the ESAS 
monitoring. However, the results of GPS telemetry 
do not comply with these findings as they suggest 
that the shags’ most important foraging areas 
are located along the coast. According to the 
recommendations by the marine IBA identification 
protocol (BirdLife International 2010b), the 
GPS telemetry data, retrieved from more than 10 
individuals, classifies as the primary data source, 
but we classified it as supplementary despite 
adequate number of tracked individuals. The 
reason for this lies in the fact that the sample of 
tracked shags was biased due to non-randomized 
catching. The majority of individuals were caught 
at one near-coast location, characterized as a resting 
spot used by the shags during their diurnal activities 
(i.e. between foraging episodes). As revealed by the 
results of this telemetry study, many shags tend to 
be faithful to their foraging areas, meaning that our 
sample was at least to some degree biased towards 
the locally-foraging population and thus not 
representative for the entire roosting population. 
To ensure random sampling, the majority of 
shags should be caught at their communal roost-
sites. However, these individuals were extremely 
hard to catch as they tended to avoid the floating 
rafts and buoys with mounted traps, placed in 
the environment with abundant alternative 
traditional roosting structures, i.e. the shellfish 
farm buoys. Another possible source of bias could 
be the difference in total diving time when shags 
are foraging in shallow coastal waters or foraging 
in deep off-shore waters, which would affect the 
frequency of successfully fixed GPS locations. 
However, the total diving time of individuals was 
not assessed in this study. 
Nevertherless, the GPS telemetry and land-
based census of large floating groups importantly 
supplement the ESAS censuses that tend to overlook 
the shags closely associated with the sea-shore. As 
part of the maritime traffic rules, the ESAS census 
boat was not allowed to travel with required speed 
closer than 400 m from the shoreline, missing out 
at least the innermost 100  m coastal belt where 
most communal foraging occurs. Although the 
supplementary datasets do not provide adequate 
data to evaluate the entire Slovenian coastal belt, 
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
125
they do provide adequate information to identify 
an important foraging area extending from the 
existing marine IBA Debeli rtič along several 
kilometres of the coast. The importance of the site 
for the socially foraging shags is supported also 
with the findings of their diet analysis that revealed 
disproportionally high frequencies of sand-smelts 
in the diet of individuals roosting at the Debeli 
rtič roost-site (Lipej & Mavrič 2013, Lipej et al. 
2016). However, the rest of the innermost coastal 
belt should be further evaluated in terms of its 
importance for foraging shags through systematic 
land-based monitoring from several spots at the 
coast and through systematic monitoring of 
innermost coastal belt by slow-speed travelling boat.
The standard IBA criteria used for site 
justification require two conditions to be met: (1) 
regular use, and (2) threshold number of individuals 
utilizing the site. For the purpose of marine IBA 
justification, the regularly used areas were defined 
as areas visited by birds from more than one site 
or during different periods (i.e. seasons or years) 
(BirdLife International 2010b). 
Regular use of the IBA Osrednji Tržaški zaliv 
was demonstrated by repeated systematic censuses 
over the period of 13 months and two non-breeding 
seasons in 2012 and 2013. It was demonstrated that 
in both years the number of shags using the area 
exceeded 300 individuals in several months and that 
the average summer number over both years was 
above 300 individuals. It is of utmost importance 
to stress that these calculations do not take into 
account the turnover, i.e. the phenomenon that 
more birds are actually using the site over a period 
of time than the number of individuals recorded at 
any given moment during the census. The turnover 
rates for the foraging Mediterranean Shags in this 
study were not assessed and thus the true number 
of individuals using the IBA within one season 
cannot be estimated, but it could be considerably 
higher than the estimated numbers. This is very 
important to consider in the attempts to compare 
these estimated numbers against the size of the 
roosting population, acquired by total counts of 
static individuals within the roost-site monitoring. 
Moreover, the first few censuses implemented in 
2012 (from July to September) were not conducted 
in the optimal time of the day, as their starting 
time was too early. Based on the data from a few 
all-day monitorings of the communal roots-sites, 
conducted after the onset of ESAS monitoring 
(Bordjan et al. 2013, Pavletič 2013, Rozman 
2012), the optimal time for at-sea census in the peak 
non-breeding season is between 10 a.m. and 4 p.m. 
It was estimated that 48–70% shags were staying at 
the roost-sites during the first hour of the censuses 
that started around 8 a.m., 31–43% in the first hour 
of censuses starting around 9 a.m. and only 13–28% 
in the first hour of censuses started between 10–11 
a.m. (own unpublished data). Based on these data 
we conclude that the estimated number of shags 
censuses between July and September 2012 does 
not reflect the peak diurnal numbers of shags in the 
IBA Osrednji Tržaški zaliv. 
The estimated number of shags in the IBA 
Osrednji Tržaški zaliv and even in the entire target 
area was well below 300 individuals in two of the 
eight censuses during peak non-breeding season 
(September and October 2012), although the 
number of roosting population was at its highest 
level. These low estimated numbers of shags at sea 
coincide with the autumn period when the shags 
frequently formed large groups and foraged on 
schools of small pelagic fish in the shallow coastal 
waters. These groups were often formed of 100 
or more individuals, and nine such cases were 
reported in September and October 2012 (max. 
230 individuals in one group) only in random 
observations. Since the large groups foraging in the 
innermost coastal belt might have been overlooked 
from the ESAS census boat, this seasonal 
phenomenon could explain the lower numbers 
of shags in the autumn peak non-breeding season 
months, estimated based on the ESAS dataset.
The IBA Debeli rtič has been estimated to hold 
about 800 roosting individuals in the 2006–2011 
period (Denac et al. 2011). In this study, however, 
a maximum of 532 and 610 individuals were 
registered during the peak non-breeding seasons 
in 2012 and 2013, respectively. The roost-site itself 
meets IBA criteria, but the extension can also be 
justified as an important foraging area. The shags 
tend to form large foraging groups around this 
roost-site, diving after schools of small pelagic fish 
along the coast outside the original IBA borders. 
Two groups of 300 and 310 individuals were 
observed in the area in two different years, proving 
that the numerical threshold has been met without 
Acrocephalus 39 (178/179): 101–128, 2018
126
even considering the turnover rates. Regular use of 
the extension to the existing IBA was demonstrated 
by both types of data used for its delineation. The 
large groups of 130 or more individuals were 
observed in the area in four different years and the 
area was visited by shags from two different roost-
sites as revealed by the GPS telemetry data. 
Least but not last, the results of this study 
should be viewed on the spatial scale of the entire 
Gulf of Trieste. The non-breeding population of 
the Mediterranean Shags within this region cannot 
be partitioned to national fragments. Although the 
identification of marine IBAs for the Mediterranean 
Shag was limited to the Slovenian territorial sea 
within the scope of the SIMARINE-NATURA 
(LIFE10NAT/SI/141) project, the data collected 
in this study as well as previous knowledge about 
the distribution of Mediterranean Shags in the 
region call for the identification of complementary 
sites within the Italian waters or transboundary 
sites spanning across the political borders. 
Povzetek
Omrežje IBA redno posodabljamo na osnovi 
novih podatkov in njihove boljše kakovosti. 
V Sloveniji smo morska območja IBA v preteklosti 
opredelili v treh fazah, vendar je bil sredozemski 
vranjek vključen kot kvalifikacijska vrsta le v zadnji 
fazi leta 2011. Toda ker so bila območja omejena 
na njegova obalna prenočišča, niso zadostovala za 
pokrivanje prehranjevalnih območij te vrste. Da 
bi zapolnili to vrzel v slovenskem teritorialnem 
morju, smo v okviru projekta SIMARINE-
NATURA (LIFE10NAT/SI/141) v obdobju 
2011–2015 opredelili nova morska območja IBA za 
sredozemskega vranjeka, in sicer po standardizirani 
metodologiji naravovarstvene organizacije BirdLife 
International za opredelitev morskih območij 
IBA. Podatki o razširjenosti in velikosti populacij 
sredozemskega vranjeka so bili zbrani z naslednjimi 
terenskimi metodami: (1) mesečnim monitoringom 
na skupnih obalnih prenočiščih, (2) mesečnim 
monitoringom na morju po standardizirani metodi 
ESAS, (3) z GPS-telemetrijo, in (4) nesistematičnim 
štetjem obalnih skupin plavajočih osebkov. Na 
osnovi teh podatkov smo opredelili eno novo 
območje, IBA Osrednji Tržaški zaliv, in razširitev 
že obstoječega IBA Debeli rtič, ki pokrivata 8218 
oziroma 155 ha. Novi območji tako pokrivata 39,2 % 
slovenskega teritorialnega morja.
Acknowledgements
This research was done as part of the project 
SIMARINE-NATURA (LIFE10NAT/SI/141) 
project co-financed by LIFE, the EU’s financial 
instrument for the Environment. Other co-financ-
ers were: Ministry of the Environment and Spatial 
Planning, City Municipality of Koper, Municipali-
ty of Izola and Municipality of Piran. 
I am especially grateful to many who took 
part or otherwise helped in this study either as 
professionals or volunteers. This research could not 
be accomplished without their participation. These 
were (in alphabetical order): Andrej Bibič, Dr Dejan 
Bordjan, Igor Brajnik, Bruna Campos, Dr Damijan 
Denac, Katarina Denac, Dr Maria Dias, Ana 
Dolenc, Dare Fekonja, Jernej Figelj, Dr Irena Fonda, 
Jakob Fric, Tilen Genov, MSc, Pedro Geraldes, Dr 
Jonathan Green, Dr Mateja Grego, Eva Horvat, 
Lech Iliszko, Kaja Jensterle, Martina Kačičnik 
Jančar, MSc, Luka Kastelic, Mirko Kastelic, Dr 
Primož Kmecl, Brina Knez, Neža Kocjan, Dr 
Ben Lascelles, Julijana Lebez-Lozej, MSc, Bojana 
Lipej, Dr Lovrenc Lipej, Metka Lotrič, Bojan 
Marčeta, Dr Borut Mavrič, Andrej Medved, MSc, 
Ana Meirinho, Nace Mihelič, Tomaž Mihelič, Dr 
Roberto Odorico, Polona Pagon, Miša Pavletič, Dr 
Polona Pengal, Nevenka Pfajfar, Patrik Pucer, Anja 
Pitamic, Dejan Putrle, Bia Rakar, Rok Rozman, 
Borut Rubinić, Dr Stefano Sponza, Domen 
Stanič, Dr Boštjan Surina, Nataša Šalaja, MSc, 
Bojan Škerjanc, Iztok Škornik, Tanja Šumrada, 
Marguerite Tarzia, Dr Davorin Tome, Robert Turk, 
MSc, Paolo Utmar, Duša Vadnjal, Barbara Vidmar, 
Dr Al Vrezec, Eva Vukelič. A big thank you also to 
anonymous referees and the editor for their valuable 
comments to the manuscript.
Last but not least, I am sincerely grateful also to 
many unnamed for their kind support during the 
course of the project. 
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia
127
5. References 
Bazin N., Imbert M. (2012): Mediterranean Shag 
Phalacrocorax aristotelis desmarestii. Updated 
state of knowledge and conservation of the nesting 
populations of the Mediterranean Small Islands.  - 
Initiative PIM. 
Aguilar, J. S., Fernández, G. (2002): Species Action 
Plan for the Mediterranean Shag (Phalacrocorax 
aristotelis desmarestii). Final draft. – BirdLife 
International, Strasbourg.
BirdLife International (2004): Birds in Europe: 
population estimates, trends and conservation status. 
BirdLife Conservation Series No. 12. – BirdLife 
International, Cambridge.
BirdLife International (2010a): Marine IBAs in the 
European Union. Version 1.1. – BirdLife International, 
Brussels.
BirdLife International (2010b): Marine Important 
Bird Areas toolkit: standardised techniques for 
identifying priority sites for the conservation of seabirds 
at sea. – BirdLife International, Cambridge.
BirdLife International (2010c): Marine Important 
Bird Areas: priority sites for the conservation of 
biodiversity. – BirdLife International, Cambridge.
BirdLife International (2013): Methods for 
identifying marine IBAs using seabird tracking data. 
Preliminary version – 23rd Dec 2013. 
BirdLife International (2014): Marine Natura 
2000 Progress Assessment. – BirdLife International, 
Cambridge.
Bordjan D., Gamser M., Kozina A., Novak J., 
Denac M. (2013): Roost-site characteristics of 
the Mediterranean Shag Phalacrocorax aristotelis 
desmarestii along the Slovenian coast. – Acrocephalus 
34 (156/157): 5–11.
Božič L. (2003): Mednarodno pomembna območja 
za ptice v Sloveniji 2. Predlogi posebnih zaščitenih 
območij (SPA) v Sloveniji. Monografija DOPPS št. 
2. – DOPPS, Ljubljana.
Burnham K. P., Anderson D. R. (2002): Model 
Selection and Multimodel Inference. A Practical 
Information-Theoretic Approach.  – Springer, New 
York.
Camphuysen K. C. J., Garthe S. (2004): Recording 
foraging seabirds at sea. Standardised recording 
and coding of foraging behaviour and multi-species 
foraging associations. – Atlantic Seabirds 6 (1): 1–32.
Cosolo M., Privileggi N., Cimador B., Sponza S. 
(2011): Dietary changes of Mediterranean Shags 
Phalacrocorax aristotelis desmarestii between the 
breeding and post-breeding seasons in the upper 
Adriatic Sea. – Bird Study 58 (4): 461–472.
Denac K., Mihelič T., Božič L., Kmecl P., Jančar 
T., Figelj J., Rubinić B. (2011): Strokovni predlog 
za revizijo posebnih območij varstva (SPA) z uporabo 
najnovejših kriterijev za določitev mednarodno 
pomembnih območij za ptice. Končno poročilo 
(dopolnjena verzija). – DOPPS, Ljubljana.
Klanjšček, M., Zdešar Kadunc, Š., Karničnik, I., 
Radovan, D. (2015): Ažuriranje kart rab na morju. 
Kartografska in geoinformacijska podpora za prikaz 
stanja prostora za območje slovenskega dela Jadranskega 
morja. Končno poročilo - IZVLEČEK. - Geodetski 
inštitut Slovenije, Ljubljana.
Koce U., Lipej B. (2016): Varstvo sredozemskega 
vranjeka in drugih morskih ptic v slovenskem morju. 
Priročnik za uporabnike in upravljavce morskega 
prostora.– DOPPS, Ljubljana.
Lipej L., Mavrič B. (2013): Prehranjevalna ekologija 
sredozemskega vranjeka (Phalacrocorax aristotelis 
desmarestii) v slovenskem delu Jadranskega morja. 
Poročilo za projekt: SIMARINE-NATURA 
(LIFE10NAT/SI/141) – Preparatory inventory 
and activities for the designation of marine IBAs 
and SPAs for Phalacrocorax aristotelis desmarestii in 
Slovenia. – Shoreline, Trieste.
Lipej L., Mavrič B., Odorico R., Koce U. (2016): 
The diet of the Mediterranean Shag Phalacrocorax 
aristotelis desmarestii roosting along the Slovenian 
coast. – Acrocephalus 37: 151–158.
MKGP (2011): Program upravljanja z morskim 
ribištvom v vodah pod suverenostjo ali jurisdikcijo 
Republike Slovenije. 
Nelson J. B. (2005): Pelicans, Cormorants and their 
relatives, The Pelecaniformes. Bird Families of the 
World. – Oxford University Press, New York.
Pavletič M. (2013): Monitoring sredozemskih vranjekov 
(Phalacrocorax aristotelis desmarestii). Poročilo 
terenskega dela pri predmetu Ornitologija. – Univerza 
v Ljubljani, Biotehniška fakulteta, Ljubljana.
Peterlin M., Gabrijelčič E., Orlando Bonaca M., 
Lipej L., Malej A., Francé J., Čermelj B., Bajt O., 
Kovač N., Mavrič B., Turk V., Mozetič P., Ramšak 
A., Kogovšek T., Šiško M., Flander Putrle V., 
Grego M., Tinta T., Petelin B., Vodopivec M., 
Jeromel M., Martinčič U., Malačič V., Marčeta 
B., Pengal P., Strojan, I. (2013): Začetna presoja 
morskih voda v pristojnosti Republike Slovenije. 
Bistvene lastnosti in značilnosti morskih voda.  – 
Inštitut za vode Republike Slovenije, Ljubljana.
Polak S. (2000): Mednarodno pomembna območja 
za ptice v Sloveniji. Monografija DOPPS št. 1.  – 
DOPPS, Ljubljana.
R Development Core Team (2018): R: A language 
and environment for statistical computing.  – R 
Foundation for Statistical Computing, Wien.
Rozman R. (2012): Dnevna dinamika številčnosti 
sredozemskih vranjekov (Phalacrocorax aristotelis 
desmarestii) na prenočiščih Debeli rtič in Strunjan. 
Individualna naloga pri predmetu Terensko delo 
Acrocephalus 39 (178/179): 101–128, 2018
128
iz botanike in zoologije.  – Univerza v Ljubljani, 
Oddelek za biologijo BF, Ljubljana.
Sponza S., Cimador B., Cosolo M., Ferrero E. 
(2010): Diving costs and benefits during post-
breeding movements of the Mediterranean shag in 
the North Adriatic Sea.  – Marine Biology 157 (6): 
1203–1213.
Sponza S., Cosolo M., Kralj, J. (2013): Migration 
patterns of the Mediterranean Shag Phalacrocorax 
aristotelis desmarestii (Aves: Pelecaniformes) within 
the northern Adriatic Sea.  – Italian Journal of 
Zoology 2013: 1–12.
Škornik I., Utmar P., Kravos K., Candotto S., 
Crnković, R. (2011): Important post-breeding 
roosting area of Mediterranean Shag Phalacorax 
aristotelis desmarestii in Gulf of Trieste (N 
Adriatic). – Medmaravis, Alghero.
Venables W. N., Ripley B. D. (2002): Modern Applied 
Statistics with S. – Springer, New York.
Vrezec A. (2006a): Marine and coastal birds of 
Slovenia: status, population size and conservation 
of Mediterranean action plan species.  – RAC/SPA, 
Tunis Vilanova i la Geltrú.
Vrezec A. (2006b): Tržaški zaliv  – mednarodno 
morsko območje IBA / SPA?  – Acrocephalus 27 
(130/131): 117–129.
Wetlands International (2004): Waterbird 
Population Estimates, 4th edition.  – Wetlands 
International, Wageningen.
Wood S. N. (2004): Stable and efficient multiple 
smoothing parameter estimation for generalized 
additive models. – Journal of the American Statistical 
Association 99: 673–686.
Zuur A. F., Ieno E. N., Walker N. J., Saveliev A. A., 
Smith G. M. (2009): Mixed Effects Models and 
Extensions in Ecology with R. Statistics for Biology 
and Health. – Springer, New York.
Prispelo / Arrived: 12. 12. 2018
Sprejeto / Accepted: 27. 2. 2019
U. Koce:  New marine IBAs for the Mediterranean Shag Phalacrocorax aristotelis desmarestii in Slovenia