Ptt

JRa

b

a

ARRAH

KLLCB

1

mReiLicsteM2

tmm

h0

Fisheries Research 164 (2015) 26–34

Contents lists available at ScienceDirect

Fisheries Research

j ourna l ho me pa ge: www.elsev ier .com/ locate / f i shres

erch (Perca fluviatilis) and pikeperch (Sander lucioperca) in the diet ofhe great cormorant (Phalacrocorax carbo) and effects on catches inhe Archipelago Sea, Southwest coast of Finland

uhani A. Salmia,∗, Heikki Auvinena, Jari Raitaniemia, Mika Kurkilahti a, Juha Liljab,iikka Maikolaa

Finnish Game and Fisheries Research Institute, Itäinen Pitkäkatu 3A, FI-20520 Turku, FinlandFinnish Game and Fisheries Research Institute, Survontie 9, FI-40500 Jyväskylä, Finland

r t i c l e i n f o

rticle history:eceived 20 May 2014eceived in revised form 8 October 2014ccepted 12 October 2014andling Editor Prof. George A. Rose

a b s t r a c t

We investigated cormorant diet in the Archipelago Sea, Southwest coast of Finland. Samples of cormorantpellets, regurgitates and stomachs were collected in 2010–2011. During these years, pikeperch and perchwere found to make up 0.04–10% and 21–43% of the diet respectively, by mass. The length distribution ofperch eaten by cormorants was more similar to the length distribution in purse seine samples rather than

eywords:ength distributionength-weight relationshipoastal fisheryaltic sea

to that in commercial catches. The size structure of pikeperch in the cormorants’ diet was intermediatebetween those in purse seine samples and commercial catches. Combined commercial and recreationalcatches have varied between 500–2000 t for perch and 225–525 t for pikeperch in the 2000s. Potentialyield losses to fisheries caused by cormorant feeding in 2010 were 340–420 t for perch and 110–140 tfor pikeperch, thus cormorant predation may affect catches and the profitability of fishing.

© 2014 Elsevier B.V. All rights reserved.

. Introduction

Some studies in Europe and North America indicate that cor-orants can have an impact on fish stocks and fisheries (e.g.

udstam et al., 2004; Fielder, 2010; Jepsen et al., 2010; Vetemaat al., 2010; Östman et al., 2012; Mustamäki et al., 2013) whilen other studies no impact has been found (e.g. Engström, 2001;iordos and Goutner, 2007; Lehikoinen et al., 2011). The differencesn the results depend on differences in the fish species studied,ormorant predation pressure, fishing pressure, fish communitytructure and research methods. Especially percids (Percidae) seemo be particularly vulnerable to cormorant predation (Rudstamt al., 2004; Fielder, 2010; Vetemaa et al., 2010; Östman et al., 2012;ustamäki et al., 2013; Skov et al., 2014, but see Lehikoinen et al.,

011).The great cormorant (Phalacrocorax carbo) started to breed on

he southern coast of Finland in 1996 (Rusanen et al., 1998). Cor-

orants had colonized the Archipelago Sea by the turn of theillennium; the number of nesting pairs was over 4000 in 2011.

∗ Corresponding author.E-mail address: juhani.a.salmi@rktl.fi (J.A. Salmi).

ttp://dx.doi.org/10.1016/j.fishres.2014.10.011165-7836/© 2014 Elsevier B.V. All rights reserved.

Temperature during the summer is the main factor affecting theyear-class strength of percids (Böhling et al., 1991; Pekcan-Hekimet al., 2011). The mean water temperature in the Archipelago Seaduring July-August has been favorable to pikeperch (Sander luci-operca) reproduction (Pekcan-Hekim et al., 2011) in several yearsduring the first decade of the millennium (Heikinheimo et al., 2014).The eutrophication and turbidity levels have been fairly stable dur-ing the past decade (Anonymous, 2011a). Despite these favorableconditions for percids, commercial catches of pikeperch and perch(Perca fluviatilis) have diminished during the first decade of themillennium (Fig. 1). Recreational catches of perch and pikeperchin the Archipelago Sea reached their lowest level since 1986 in2010 (Anonymous, 2004, 2011b). Commercial fishermen argue thatcormorants are contributing to the reduced catches in traditionalfishing areas (Salmi et al., 2010). So far, research on the spatial vari-ation in the diet of cormorants and their effects on fisheries havebeen scarce in the Baltic Sea region.

In this paper we studied the diet of cormorants in the FinnishArchipelago Sea. We examined both spatial variation and inter-annual variation between 2010 and 2011 in the proportions of

perch and pikeperch in the diet. We compared length distribu-tions of perch and pikeperch from cormorant diet samples withdistributions from test fishing and commercial fishing samples. Wecollected these data in order to estimate the amount of perch and

J.A. Salmi et al. / Fisheries Rese

Fi

pm

2

mcTETtiAaÄtacKKbK(

tsspstsSoat

afiFr

nfw

ig. 1. Commercial yield of perch and pikeperch and the number of cormorant nestsn the Archipelago Sea from 1998–2011.

ikeperch in the cormorants’ diet and cormorants’ effect on com-ercial and recreational fishery catches in the Archipelago Sea.

. Materials and methods

Cormorant diet samples were collected at three of the eight cor-orant breeding colonies in the study area in 2010–2011. Study

olonies were Kluppi, Kustavi (60.6◦ N, 21.1◦ E); Kalmanhohde,aivassalo (60.5◦ N, 21.7◦ E) and Äggskär, Parainen (60.3◦ N, 21.8◦

) in the Archipelago Sea on the southwest coast of Finland (Fig. 2).hese colonies were selected because they are situated in habitatshat differ from each other in e.g. openness, depth, and turbid-ty, and are representative of the different habitat types of therchipelago Sea (Kangas et al., 2003). Kalmanhohde is located inn inner bay with closed shoreline, and turbid and shallow water.ggskär is situated in the central part of the Archipelago Sea, where

he water area is greater than the land area, and waters are deepernd more transparent. Kluppi is in the outer Archipelago Sea andharacterized by open shores with deep and cool water. Coloniesluppi and Äggskär were populated by cormorants in 2003 andalmanhohde in 2009 (P. Rusanen, pers. comm.). The numbers ofreeding pairs in 2010 and 2011 were, by area: 717 and 417 inluppi, 1436 and 2000 in Äggskär, and 325 and 241 in Kalmanhohde

P. Rusanen, pers. comm.).We consider the presence of cormorants during two periods of

he year; the breeding season (April–July) and the non-breedingeason (August–November). The first cormorants start to migrateouthwards after the breeding season. Breeding season diet sam-les were collected from 1 May to 31 July and non-breeding seasonamples from 1 August to 15 November. Samples collected duringhe breeding season primarily represent the diet of subspecies P. c.inensis, which inhabits the study colony area during that period.amples collected outside the breeding season represent the dietf migrating birds from both subspecies, sinensis and carbo, whichre present in the study area for a period of variable length duringheir migration.

We collected cormorant pellets (n = 647), regurgitates (n = 101)nd culled birds (n = 44). The samples contained remains from 8494sh. Samples were collected with the permission of Southwestinland Centre for Economic Development, Transport and the Envi-onment (Southwest Finland ELY Centre).

During the breeding season, diet samples were collected in theesting colonies. After the breeding season, samples were collected

rom resting places located up to nine km from the colonies. Pelletsere collected at intervals of three weeks from May to October in

arch 164 (2015) 26–34 27

2010 and from May to November in 2011. 20 individual sampleswere taken from each colony per visit; only fresh and completepellets were collected. Samples of regurgitates of nestlings werealso collected. In 2010, bulk samples from each study colony werecollected over three weeks in June and July. In 2011, we collectedsamples of regurgitated fish of 20 nestlings from each colony pervisit. Samples were collected at intervals of three weeks in June andJuly. Stomach contents of culled birds were gathered from huntersfrom the 20th August to the 31st October in 2010 and 2011.

All samples were frozen for at least 24 h. Pellets were washedbefore analysis. Fish from the defrosted sample were identifiedto species level (or if that was not possible, to the genus level).Total length (TL) was measured (whole fish) or calculated (bones)to the nearest mm and fish were weighed or calculated to thenearest gram. The lengths of partly digested fish were estimatedand masses were calculated using the length–mass formulae.Remnants of cyprinid species in pellets and stomachs were iden-tified using pharyngeal teeth and chewing pads, while other fishspecies were identified using otoliths and vomer bones (Wheeler,1978; Härkönen, 1986; Veldkamp, 1995; Radu, 2005; Cech, 2006;Svetocheva et al., 2007). Fish lengths were calculated from otolithswith Härkönen’s (1986) formula, from pharyngeal teeth with Cech’s(2006) formula and from chewing pads with Veldkamp’s (1995)formula. Only non-eroded bones were used for length–mass cal-culations. Fish mass was calculated by using length–mass relationformula (Le Cren, 1951):

W = aLb (1)

where W is the wet mass in g, L is the total length in cm, a is acoefficient and b an exponent (Table 1). The values used in thelength–mass relation formulae were based on measurements fromfish caught in test fishing, surveys and commercial fishing (EUData Collection Framework on fisheries) of the Finnish coast of theBaltic Sea. For fourhorned sculpin we used values from Swedishtest fishing (Institute of Coastal Research, Department of AquaticResources, Swedish University of Agricultural sciences). We calcu-lated mass for black goby and small sandeel (Ammodytes tobianus)using Härkönen’s (1986) formula. For three-spined stickleback weused a mean mass of two grams and for common goby (Pomatoschi-tus microps) and sand goby (Pomatoschitus minutus) a mean massof one gram.

We used pellets as primary samples and supplemented thepellet data with regurgitates and samples from culled birds. Wecalculated proportions of different prey species in the diet by com-bining individual samples from the time frame and area of interest.Samples of different types were treated equally when calculatingproportions of different prey species in the diet.

Purse seine (cod end stretched mesh size 6 mm) test fishing sam-ples, cormorant diet samples and commercial catch samples forperch and pikeperch were collected in the Mynälahti Bay area insummer 2011 (Fig. 2). The size structures of perch and pikeperch inthe samples were compared. The size distributions of gill net (GN)and fyke net (FN) commercial catches were combined and the ratio(GN/FN) in the total catch was used as a weighting factor.

The total fish mass consumed by cormorants was calculated sep-arately for the eight colonies in the study area. The number of nestsin each colony in 2010–2011 was provided by Finnish EnvironmentInstitute (P. Rusanen, pers. comm.). For colonies from which welacked diet data, we used values from the nearest colony situatedin similar habitat.

The total fish mass consumed by cormorants in each colony wascalculated using two methods, which differ from each other in the

estimation of the total number of birds and in the daily consump-tion assumed for one bird. Method 1 was adopted from Vetemaaet al. (2010). They obtained the total number of birds in a colonyof increasing size by multiplying the number of nests by a factor of

28 J.A. Salmi et al. / Fisheries Research 164 (2015) 26–34

F arkedM

ftRcson

TEs

ig. 2. Cormorant colonies studied in 2010 and 2011. Colonies with diet data are mynälahti bay is marked with a rectangular.

our (Engström, 2001) and used a conservative figure of 300 g forhe mean daily food intake. Method 2 was based on the review byidgway (2010) who suggested a per capita consumption by adult

ormorants of 542 and 436 g per day in the nesting and non-nestingeasons, respectively. Estimates of the number of birds during thepen water season were based on a method developed by P. Rusa-en (pers. comm.): the number of adult birds is estimated as the

able 1stimated parameters of the equation (W = aLb) for the calculation of fish mass from fish

tandard deviation (SD), median length (cm) and range (cm) from measurements of fish f

Species Mean (cm) Median (cm)

White bream (Blicca bjoerkna) 14.6 14.0

Roach (Rutilus rutilus) 19.4 19.7

Bream (Abramis brama) 27.0 25.8

Bleak (Alburnus alburnus) 10.8 11.0

Rudd (Scardinius erythrophthalmus) 18.5 18.0

Tench (Tinca tinca) 33.1 35.3

Ide (Leuciscus idus) 30.7 30.0

Smelt (Osmerus eperlanus) 13.9 14.0

Perch (Perca fluviatilis) 24.2 25.6

Ruffe (Gymnocephalus cernua) 11.6 11.0

Pikeperch (Sander lucioperca) 38.2 39.2

Baltic herring (Clupea harengus membras) 14.4 14.5

Sprat (Sprattus sprattus) 11.3 11.5

Pike (Esox lucius) 47.2 46.0

Burbot (Lota lota) 44.1 43.0

Eelpout (Zoarces viviparous) 16.4 16.0

Whitefish (Coregonus lavaretus) 33.7 34.4

Flounder (Platichtys flesus) 22.2 23.0

with black squares. Colonies without diet data in 2010 are marked with black dots.

number of nests in the colony times by two, while the number ofsub-adults is estimated as 33% of the adult population in April and66% of the adult population in May–July (Lehikoinen, 2003). We

assumed that the number of chicks per nest is four during the firstweek after hatching, three in the second week, and two in the thirdand the fourth week. The mean number of fledglings is around twoin Finnish colonies (Lehikoinen, 2006). In the first four weeks of

length (W = wet mass (g) and L = total length (cm)) and values of mean length (cm),rom the Finnish coast. r2 is the coefficient of determination for the equation.

Range (cm) SD a b r2 n

6–27 4.7 0.0092 3.0759 0.9767 3458–32 5.2 0.0046 3.2789 0.9852 8617–49 7.2 0.0052 3.2214 0.9865 9315–14 2.3 0.0064 3.0516 0.9247 2069–31 5.1 0.0094 3.1581 0.9844 1357–48 11.3 0.0118 3.1165 0.9905 30

14–49 8.6 0.0081 3.1372 0.9866 905–24 3.4 0.0052 3.0063 0.9608 2675–37 5.1 0.0069 3.1704 0.9823 23866–19 2.9 0.0158 2.8722 0.9431 3147–80 6.7 0.0053 3.1153 0.9889 44195–33 3.7 0.0050 3.0558 0.9805 60067–16 1.9 0.0049 3.0631 0.9253 2116

14–88 13.2 0.0062 3.0190 0.9707 14819–80 11.1 0.0024 3.2801 0.9670 397

5–30 3.8 0.0024 3.1357 0.9098 26012–53 5.5 0.0026 3.3349 0.9646 3449

8–34 5.1 0.0174 2.8848 0.9711 235

J.A. Salmi et al. / Fisheries Research 164 (2015) 26–34 29

Table 2Annual mortality values (M = instantaneous natural mortality rate, F = instantaneous fishing mortality rate), the mean masses of landed fish in fyke net and gill net catches,and the percentage of undersized discarded pikeperch in gill net fishery, used in the model.

Age Perch Pikeperch Discarded (%)

M F Fyke net mean (g) Gill net mean (g) M F Fyke net mean (g) Gill net mean (g)

1 0.6 0.00 10 – 0.4 0.00 – – 100.02 0.4 0.05 18 76 0.3 0.05 – – 100.03 0.3 0.20 48 183 0.2 0.10 398 467 93.44 0.2 0.30 108 224 0.2 0.30 419 507 25.15 0.2 0.40 130 248 0.2 0.60 425 535 15.66 0.2 0.50 143 263 0.1 1.00 449 537 13.57 0.2 0.50 157 272 0.1 1.00 477 543 9.6

0.1

0.10.1

labttF

wfiwtcsd

nacmhimwHof

c1

N

wa

TT

8 0.2 0.50 154 283

9 0.2 0.50 163 282

10 0.2 0.50 159 298

ife, the daily food intake of chicks was assumed to be 82, 202, 332nd 433 g respectively, and thereafter 436 g (Ridgway, 2010). Inoth methods, the total fish consumption by cormorants duringhe non-breeding period was estimated to be 25% of the consump-ion during the breeding period (April to July) as suggested in theinnish cormorant management plan (Anonymous, 2005).

The relative consumption of perch and pikeperch in each colonyas calculated separately using the mean percentage of the totalsh mass of these species in the food samples from the whole open-ater season in 2010. For the breeding colony Måsgrund we used

he results of Korhonen (2010). The number of perch and pikeperchonsumed by cormorants was calculated as the total mass con-umed divided by the individual mean mass of the species in theiet samples in 2010.

We estimated cormorant predation on perch and pikeperch inumbers by age group in 2010. The age distributions of perchnd pikeperch were based on a length-age key derived from theommercial fyke net samples in 2010. Then, we calculated howuch perch and pikeperch predated by cormorants in 2010, would

ave yielded to fisheries in the following years in the hypothet-cal absence of cormorants. We used annual fishing and natural

ortality values (F and M) from assessments in years (1978–1997)hen cormorants were absent (updated values of Setälä et al., 2003,eikinheimo et al., 2006). For the first year (2010) we ran the modelnly for a half year, because cormorant predation takes place mainlyrom May to October.

The number of fish surviving from one age group to the next wasalculated as (e.g. Beverton and Holt, 1957; Hilborn and Walters,992):

1 = N0e−(Z) (2)

here N0 is the initial number of fish and N1 is the number of fishfter one unit of time (year) and Z is the total mortality rate which

able 3he proportions of perch and pikeperch in the diet of cormorants in 2010 and 2011 by m

Species Year Period Kalmanhohde

% of mass % of number n

Perch 2010 Total 21 19 239

Breeding 20 15 138

Non-breeding 22 28 101

2011 Total 26 23 241

Breeding 13 12 83

Non-breeding 37 42 158

Pikeperch 2010 Total 10 4 46

Breeding 13 4 33

Non-breeding 6 4 13

2011 Total 10 3 36

Breeding 6 1 8

Non-breeding 14 7 28

1.00 505 539 6.91.00 658 577 6.11.00 658 639 5.0

can be obtained as the regression coefficient from log-transformed(ln + 1) annual age frequencies of fish in samples. The total mortalityrate (Z) was obtained as the sum of fishing mortality (F) and naturalmortality (M) rates (Z = F + M) (e.g. Beverton and Holt, 1957; Hilbornand Walters, 1992; Cadima, 2003). The number of fish that died infisheries (DF) in a year was calculated by the formula (e.g. Hilbornand Walters, 1992):

DF = F

ZN(1 − e−(Z)) (3)

where N is the number of fish in the beginning of the year. The yieldY (i.e. kg fish caught in the fishery) was calculated as the product ofnumbers and mass.

We evaluated the effect of different natural mortality values forthe youngest age groups on catches. In the pikeperch model weused natural mortality rates of 0.8, 0.5 and 0.4 for age groups oneto three, respectively instead of 0.4, 0.3 and 0.2. In the perch modelwe used natural mortality rates of 1.0, 0.8 and 0.5 instead of 0.6, 0.4and 0.3 for age groups one to three (Table 2). Because there is a min-imum landing size limit for pikeperch (37 cm), the undersized fishwere removed from the potential catches in the gill nets accord-ing to the distributions of age specific samples to landed/discardedfish. When caught by other gear it is possible to release the under-sized pikeperch alive. However, we assumed that five percent ofpikeperch caught in other gear than gill nets in each age group willbe discarded and die. As most of the pikeperch and perch in thediet of the cormorant were too small to be caught as catch in thefisheries in 2010, we made the following assumptions based onthe length data of the cormorants’ diet: pikeperch that were below

35 cm in length were not supposed to be able to reach the mini-mum landing size in 2010 and were thus omitted from the catches,and perch that were below 20 cm in length were not targets for gillnet fishing in 2010.

ass and number.

Äggskär Kluppi

% of mass % of number n % of mass % of number n

24 21 382 29 15 34126 21 341 27 13 25018 20 41 33 27 9137 30 343 43 23 16940 30 326 34 17 10416 22 17 57 64 65

8 2 39 1 0.3 87 2 32 2 0.2 89 3 7 0 0 02 1 10 0.04 0.1 12 1 10 0.1 0.2 10 0 0 0 0 0

30 J.A. Salmi et al. / Fisheries Research 164 (2015) 26–34

Table 4Observed mean mass (g), standard deviation (SD), range (g), and number of speci-mens (n) of perch and pikeperch in the cormorant diet per colony in 2010–2011.

Species Colony Year Mean (g) SD Range (g) n

Perch Total 2010 42 31 1–206 9622011 52 41 2–305 753

Kalmanhohde 2010 44 34 1–181 2392011 52 42 3–255 241

Kluppi 2010 44 22 3–155 3412011 55 46 3–305 169

Äggskär 2010 38 32 1–206 3822011 50 38 2–267 343

Pikeperch Total 2010 110 86 13–439 932011 123 83 6–439 47

Kalmanhohde 2010 110 92 17–439 462011 134 79 36–439 36

Kluppi 2010 86 57 24–190 82011 10 – – 1

g(iaoathc(g

3

3

ipibd2cb5pnmtTn

aimmw41

Amst

Fig. 3. Size frequency distribution of perch of different length groups in Mynälahtibay from purse seine samples, cormorant diet samples and commercial gill-net andfyke-net catches in 2011.

Äggskär 2010 114 84 13–351 392011 97 91 6–276 10

The distribution of potential catches according to fishingear was based on commercial and recreational catch statisticsAnonymous, 2011b, 2012). We used samples of landed pikeperchn 2010–2013 to estimate the age and gear specific mean masses,nd random gill net samples of the same years to estimate the sharef discarded undersized fish in the catch. For perch we did not haveged samples after 2011, and we used the data from 2005 to 2011o estimate the age and gear specific mean masses. As we did notave samples from non-commercial catches, we used the age spe-ific mean mass of commercial catch from gears other than gill netsmainly fyke nets) to represent recreational fisher’s catches withears other than gill nets.

. Results

.1. Diet and size distribution

The diet of cormorants consisted of 27 fish species and 12 fam-lies in 2010–2011. The most common diet items in biomass wereerch (28%) and roach (Rutilus rutilus; 17%). Numerically, the most

mportant prey species were perch (19%) and three-spined stickle-ack (Gasterrosteus aculeatus; 19%). The share of pikeperch in theiet was 6% (range 0.04–10% per year and colony) by mass and% (range 0.1–4% per year and colony) by number in the studyolonies (Table 3). Perch comprised 21–43% by biomass and 15–30%y number (Table 3). The length range of consumed perch was–29 cm (mean 15 cm, median 15 cm, SD 3.8 and n = 1715), while forikeperch it was 10–38 cm (mean 23 cm, median 24 cm, SD 5.6 and

= 140). The length range of all prey fish was 3–43 cm. The meanass of perch was 42 g in 2010 and 52 g in 2011; for pikeperch,

he mean mass was 110 g in 2010 and 123 g in 2011 (Table 4).he share of perch and pikeperch fluctuated between breeding andon-breeding season (Table 3).

In the breeding colony of Kalmanhohde in the inner part of therchipelago, the diet consisted of 20 species. The most common diettems by mass were perch (24%) and roach (18%). Numerically, the

ost important prey species were Baltic herring (Clupea harengusembras; 21%), perch (20%) and roach (18%). The share of pikeperchas 10% by mass and 3% by number. The mean mass of perch was

4 g in 2010 and 52 g in 2011; for pikeperch, the mean mass was10 g in 2010 and 134 g in 2011 (Table 4).

In the breeding colony of Äggskär in the central part of the

rchipelago Sea, the diet was composed of 21 fish species. Theost common food items by biomass were perch (30%), fourhorned

culpin (Triglopsis quadricornis; 15%) and roach (13%). Numerically,he most important prey species were perch (24%) and ruffe (19%).

Fig. 4. Size frequency distribution of pikeperch of different length groups inMynälahti bay from purse seine samples, cormorant diet samples and commercialgill-net and fyke-net catches in 2011.

The share of pikeperch was 7% by biomass and 2% by number.The mean mass of perch was 38 g in 2010 and 50 g in 2011; forpikeperch, the mean mass was 114 g in 2010 and 97 g in 2011(Table 4).

In the breeding colony of Kluppi in the outer archipelago, thediet of cormorant consisted of 20 fish species. The common preyitems by mass were perch (34%) and eelpout (Zoarces viviparous;24%). Numerically, the most important prey species were eelpout(30%), three-spined stickleback (18%) and perch (17%). The share ofpikeperch was 1% by biomass and 0.3% by number. The mean massof perch was 44 g in 2010 and 55 g in 2011; for pikeperch, the meanmass was 86 g in 2010 and 10 g in 2011 (Table 4).

In Mynälahti bay, 95% of perch in purse seine samples werebetween 7 and 18 cm in length, perch consumed by cormorants var-ied between 7 and 22 cm in length, and perch in commercial catchesvaried between 13 and 30 cm in length (Fig. 3). For pikeperch,the corresponding figures were 9–28 cm, 17–30 cm and 25–44 cm,respectively (Fig. 4).

3.2. Consumption of perch and pikeperch and the potential loss ofyield

The estimated total mass of fish consumed by cormorants inthe Archipelago Sea in 2010 was between 679–835 t (Table 5). In2010 cormorants consumed 168–208 t of perch and 50–63 t of

J.A. Salmi et al. / Fisheries Rese

Table 5Number of cormorant nests (P. Rusanen, pers. comm.) and estimates of annual fishconsumption by cormorants (t) in the Archipelago Sea in 2010 using two methods.

Colony No. of nests Method 1 Method 2Consumption (t) Consumption (t)

Kalmanhohde 325 75 93Kluppi 717 165 201Mäntykaria 24 6 6Måsgrund 425 96 120Smörgrunda 10 3 3Urpoinenb 25 6 6Äggskär 1436 327 405Äijäkaric 5 1 1Total 2967 679 835

Diet from aÄggskär, bKluppi, cKalmanhohde.

Table 6Estimated numbers of perch and pikepearch eaten by cormorants in the ArchipelagoSea in 2010.

Age (years) Perch Pikeperch

Method 1 Method 2 Pikeperch Method 1 Method 2

1 161 000 199 000 155 000 191 0002 613 000 757 000 77 000 95 0003 1 613 000 1 991 000 100 000 124 0004 958 000 1 183 000 70 000 87 0005 494 000 609 000 50000 62 0006 94 000 116 000 3000 40007 20 000 25 000 4000 50008 49 000 61 000 2000 20009 6000 7000 0 010 2000 2000 0 0

pa

btohWpp

Ft2

Total 4 010 000 4 950 000 461 000 570 000

ikeperch. In numbers, this corresponds to 4.0–4.9 million perchnd 460 000–570 000 pikeperch (Table 6).

The majority of perch and pikeperch eaten by cormorants wereetween one and five years of age, but four percent of perch andwo percent of pikeperch were older (Table 6). The potential lossf yield to fisheries caused by cormorant feeding in 2010 wouldave been 110–140 t for pikeperch and 340–420 t for perch (Fig. 5).hen using higher natural mortality values for young age groups

otential yield losses were lower by 14% for pikeperch and 7% forerch as compared with the original mortality values.

Pote

ntia

l loss

of y

ield

, t

0

25

50

75

100

125

150

Yea r1 2 3 4 5 6 7 8 9 10

hcreP hcrepekiP

ig. 5. The potential loss of yield by year caused by cormorant predation in 2010 inhe Archipelago Sea. The shaded area shows the difference between methods 1 and.

arch 164 (2015) 26–34 31

4. Discussion

4.1. Diet and size distribution

The number of fish species in the diet of cormorants was approx-imately the same in all sampled colonies. Previous studies showthat the diet of cormorants usually consists of 10–59 different preyspecies, depending on the location and the duration of the study(Engström, 2001; Lunneryd and Alexandersson, 2005). The 27 dif-ferent fish species found in this study represent a typical numberof species seen in cormorant diet studies.

The cormorant is not a species-specific predator (Gremillet et al.,1996), but the main part of its diet consists of a few prey speciesthat are common in the cormorant’s habitat (Liordos and Goutner,2007). Perch comprised the largest share of the diet by mass in allcolonies and was also among the three most abundant species bynumber. Perch has also been found to be an important prey item inother studies in the northern Baltic Sea (Lehikoinen, 2005; Boströmet al., 2012a). In our study area, perch is a common species, and itssize is suitable for the cormorant to swallow whole, which explainsits abundance in the diet. Pikeperch, on the other hand, was nota common prey item in our study area. Its share of the diet washighest in the inner parts of the archipelago. Pikeperch spawningareas are clustered around the shallow bays in the inner part of theArchipelago Sea, close to the continent (Veneranta et al., 2011). Forthe birds in colonies located in these areas or near them, pikeperchis an important prey species. Lehikoinen (2005) and Boström et al.(2012b) found few pikeperch in their study in the northern BalticSea, but Eschbaum et al. (2003) found high quantities of pikeperchconsumed by cormorants in the shallow areas (mean depth < 10 m)off the west coast of Estonia.

The share of perch in the diet increased from 2010 to 2011 inall colonies. At the same time, the share of pikeperch stayed atthe same level or decreased. Proportions of perch and pikeperchalso differed between the breeding and non-breeding seasons.Temporal changes in the diet are related to the life-cycles ofprey fish species, water temperature changes, food require-ments of cormorant nestlings and the density of prey speciesin the sea. Lehikoinen (2005) and Boström et al. (2012a) founddiet changes during the breeding season, while Keller (1995)and Cech et al. (2008) found diet changes in the non-breedingperiod.

The diet of cormorants fluctuates spatially even within a rela-tively small area, depending on the habitat types near the coloniesand resting places. The proportion of perch and pikeperch in thediet fluctuated between colonies in certain years. The spatial vari-ation is explained by different foraging grounds, which differ inspecies composition and the density of each species (Lilliendahland Solmundsson, 2006; Liordos and Goutner, 2007; Boström et al.,2012a). The study colonies were located in geographically differentparts of the Archipelago Sea at least 30 km from each other. Thedistances between the colonies are thus great enough that differ-ences in foraging habitat and species composition explain spatialvariation in diet composition.

Cormorants were found to display preference for a certainsize range of perch and pikeperch. The size ranges of perch andpikeperch eaten by cormorants agree with the results of previousstudies where cormorants were found to prefer slender, 5–36 cmprey (Lunneryd and Alexandersson, 2005; Liordos and Goutner,2007; Cech et al., 2008; Boström et al., 2012b). Looking at the wholeprey size spectrum, the majority of consumed perch fall on the lefthand side of the median, while most pikeperch fall to the right hand

side of the median. The similar increase of mean mass of perch eatenby cormorant from 2010 to 2011 in all colonies suggests an increaseon mean mass in the perch population, as well. The total mean massof pikeperch eaten by cormorants was higher in 2011 than in 2010,

3 s Rese

bc

fswtuemct1tctgs

hmtbcietfcltttwwau2

4y

mtrcesdtet2woclttp

mmip

2 J.A. Salmi et al. / Fisherie

ut a similar trend of increased mean mass in all colonies as in thease of perch was not observed.

The size distributions of both perch and pikeperch were dif-erent among purse seine, commercial catch and cormorant dietamples. The majority of perch and pikeperch eaten by cormorantsere smaller than those in commercial catches, but larger than

hose in test fishing samples. Pikeperch eaten by cormorants aresually below 37 cm in length, the minimum legal size for fish-ries. The length distribution of perch eaten by cormorants wasore similar to that in purse seine samples than in commer-

ial catches. There is overlap in the size spectrum of perch inhe cormorant’s diet and commercial catches (perch in the range9–22 cm), but commercial fishermen generally catch larger perchhan cormorants. Eschbaum et al. (2003) found that on the westoast of Estonia, the length distribution of perch was almost iden-ical in the diet of the cormorant and in multi-mesh experimentalillnetting samples, but in the case of pikeperch, smaller juvenilepecimens prevailed in the diet than in commercial catches.

Of the available methods in diet studies, examination of pelletsas been the most popular method to investigate the diet of cor-orants (Carss et al., 1997). Seefelt and Gillingham (2006) observed

hat different methods in diet studies (pellets, regurgitate, or culledirds) can lead to differing estimations of prey abundance in theormorant’s diet. These methods can give slightly different resultsn the numbers of prey items, but presence or absence of differ-nt prey types shows similarity. A diet study should cover at leasthe whole breeding season, if its purpose is to estimate cormorantood consumption at the species level and its effect on commercialatches. From these three traditional methods, usually only pel-ets can provide enough samples to estimate temporal changes inhe diet from breeding to the autumn migration. Best practice iso supplement pellet data with samples from nestlings’ regurgi-ates. We have tried to compensate these uncertainties associatedith different methods by using all three methods together, bute have kept pellet analysis as the main method, since several

dvantages and good knowledge of its weaknesses support itsse (Suter and Morel, 1996; Carss et al., 1997; Johnson et al.,010).

.2. Consumption of perch and pikeperch and the potential loss ofield

Our results indicate that both perch and pikeperch populationsay be affected by cormorant predation resulting in changes in

he catch levels of fishers. These results are in line with severalecently published results indicating that percids are vulnerable toormorant predation (Rudstam et al., 2004; Fielder, 2010; Vetemaat al., 2010; Skov et al., 2013; Östman et al., 2013). Especially inituations similar to the Archipelago Sea, where cormorant pre-ation has occurred together with an intensive fishing, effects onhe catch level have been found (Mustamäki et al., 2013; Östmant al., 2013). In some occasions, the affected fish stock has startedo strengthen after the onset of cormorant control (Rudstam et al.,004; Fielder, 2010; Dorr et al., 2012; Ridgway and Fielder, 2013),hich supports these results. The fact that in some cases no effects

n percid populations by cormorant predation have been found,an depend on many factors, e.g. the intensity of predation can beow because of the distance between the colonies and the place ofest fishings, other more attractive food may be readily available,he fishing pressure can be low, or the methods used may not beowerful enough to detect changes.

The calculation of the potential loss of yield in the fishery was

ade in order to evaluate the consequences of the presence of cor-orants for catches. The results thus represent potential catches

n the hypothetical absence of cormorants, from the number oferch and pikeperch eaten by cormorants in one year. The model

arch 164 (2015) 26–34

estimated catch losses of 340–420 t and 110–140 t, for perch andpikeperch respectively, caused by the predation of cormorants inone year. Cormorant predation shows annual variation, both asto the total volume and to the species selected. Therefore it isnot possible to estimate the long term annual catch losses basedonly on one year’s data. However, provided that the year 2010was representative of the mean intensity of cormorant predation,the losses estimated are considerable, when compared with theactual catch statistics of the area. The annual commercial catch ofperch has varied between 170 and 420 t, and that of pikeperchbetween 150 and 370 t in 1998–2010. The recreational catcheshave varied between 270 and 1550 t of perch and 70–280 t ofpikeperch. Catch estimates for the recreational fishery have wideconfidence limits (Anonymous, 2011b) and thus those figures mustbe treated with caution. The total annual catch of perch, from bothcommercial and recreational fisheries in the Archipelago Sea hasvaried between 500 and 2000 t since 1998, and that of pikeperchbetween 225 and 525 t (Anonymous, 2004, 2011b; Fig. 1). Theannual potential loss of yield would correspond to 20–80% and20–60% of the mean annual total perch and pikeperch catch in theArchipelago Sea in 1998–2010, respectively. These values are onthe same level as the estimated indirect effects of cormorant pre-dation on perch in the Swedish coast (Östman et al., 2013). Thecatch declines occurred simultaneously with the increasing num-ber of cormorants (Fig. 1). In addition to affecting population sizeand size structure of fish, the presence of cormorants may result inchanges in fish behavior and distribution (Skov et al., 2013), whichcan affect the profitability of fisheries. The results of the model-ing can only show the order of magnitude of the changes in thecatch level caused by cormorants. It seems however evident thatthe cormorant predation has caused at least a part of the changesin the perch and pikeperch catches that have been documented inthe previous years.

The success of cormorants in the Baltic region may depend oneutrophication, a warming climate (Pekcan-Hekim et al., 2011), andincreased density of small fish. These factors have provided cor-morants a suitable niche in the Archipelago Sea. The amount ofpikeperch and perch consumed annually varies according to thedensity of suitably-sized individuals, the number of cormorantsand the locations of their colonies. Cormorant colonies and foraginggrounds are not equally distributed in the Archipelago Sea (Fig. 2),which means that predation pressure caused by cormorants is notequally distributed to the area (Table 5). Although our consump-tion calculations are made for the whole Archipelago Sea, the mainpart of consumption takes place in the foraging grounds aroundbig colonies. This means that actual effects for the local fish popu-lations might be relatively higher than those of the whole area. Onthe other hand, both cormorants and fishers may concentrate onthe most profitable fishing sites.

Many parameters in the calculation of then potential catch lossare yet poorly known, e.g. the lack of diet data from all colonies,the lack of bird counts outside the breeding season, the naturalmortality rate of young fish, the effect of possible compensatoryeffects, the consequences of possible differences in the growth rateswithin a cohort etc. Thus at present state of knowledge, the modelgives only rough estimates of the real situation.

Since cormorant predation shows annual variation, the cor-morants’ diet should be monitored for several years in order tomake reliable conclusions about the long-term effects of predationon catches. In order to gain better estimates on cormorants’ effectson the mortality of different fish species, size groups and cohortsin the Archipelago Sea, tagging of fish (Skov et al., 2014) and GPS-tagging of birds should be considered. This would enable the useof more sophisticated models to estimate the direct and indirect

effects of cormorant predation on the development of fish stocksand catches.

s Rese

A

S2mGePfWoFes

R

A

A

A

A

A

B

B

B

B

C

C

C

C

D

E

E

F

G

H

H

J.A. Salmi et al. / Fisherie

cknowledgments

We thank Katja Korhonen from Turku University of Appliedciences for collection and preparation of regurgitate samples in010, Timo Asanti and Pekka Rusanen and the Finnish Environ-ental Institute for providing us the cormorant population data,unilla Sandberg from Swedish University of Agricultural sci-nces for letting us to use their length–mass relation formulae,ekka Jounela, Outi Heikinheimo and two anonymous reviewersor providing valuable comments on the manuscript and Rebecca

hitlock for revising the English. University of Turku provided lab-ratory facilities during the study. The Ministry of Agriculture andorestry, Southwest Finland ELY Centre, Finnish Game and Fish-ries Research Institute and the European Fisheries Fund financiallyupported the study.

eferences

nonymous, 2004. Recreational Fishing 2002. SVT Maa-, metsä- ja kalatalous nro51. Official Statistics of Finland – Agriculture Forestry and Fishery.

nonymous, 2005. Merimetson kannanhoitosuunnitelma. Ympäristöministeriönmoniste 161 (in Finnish).

nonymous, 2011a. From clear to turbid. State of the Archipelago Sea and sea areasaround the Åland Islands and nutrient loading to them. In: Publications of Cen-tre for Economic development, Transport and the Environment of SouthwestFinland 6/2011.

nonymous, 2011b. Recreational Fishing 2010. Riista- ja kalatalous – Tilastoja7/2011. Official Statistics of Finland – Agriculture Forestry and Fishery.

nonymous, 2012. Commercial Marine Fishery 2011. Riista-ja kalatalous – Tilastoja2/2012. Official Statistics of Finland – Agriculture Forestry and Fishery.

everton, R.J.H., Holt, S.J., 1957. On the Dynamics of Exploited Fish Populations.Fishery Investigations Series 2, 19. Chapman & Hall, London.

öhling, P., Hudd, R., Lehtonen, H., Karås, P., Neuman, E., Thoresson, G., 1991. Vari-ations in year-class strength of different perch (Perca fluviatilis) populations inthe Baltic sea with special reference to temperature and pollution. Can. J. Fish.Aquat. Sci. 48, 1181–1187.

oström, M.K., Lunneryd, S.-G., Sthålberg, H., Karlsson, L., Ragnarsson, B., 2012a. Dietof the Great Cormorant (Phalacrocorax carbo sinensis) at two areas at Lövstabuk-ten, South Bothnian Sea, Sweden, based on otolith size-correction factors. OrnisFennica 89, 157–169.

oström, M.K., Östman, Ö., Bergenius, M.A.J., Lunneryd, S.-G., 2012b. Cormorant dietin relation to temporal changes in fish communities. ICES J. Mar. Sci. 69 (2),175–183.

adima, E.L., 2003. Fish Stock Assessment Manual. FAO Fisheries Technical Paper.No. 393.

arss, D.N., Bevan, R.M., Bonetti, A., Cherubini, G., Davies, J., Doherty, D., El Hili, A.,Feltham, M.J., Grade, N., Granadeiro, J.P., Grémillet, D., Gromadzka, J., Harari,Y.N.R.A., Holden, T., Keller, T., Lariccia, G., Mantovani, R., McCarthy, T.K., Mellin,M., Menke, T., Mirowska-Ibron, I., Muller, W., Musil, P., Nazirides, T., Suter, W.,Trauttmansdorff, J.F.G., Volponi, S., Wilson, B., 1997. Techniques for assessingCormorant diet and food intake: towards a consensus view. Suppl. Ric. Biol.Selvag. XXVI, 197–230.

ˇech, M., 2006. Keys of Fish Head Identification Bones. Biology Centre AS CR, CeskéBudejovice.

ˇech, M., Cech, P., Kubecka, J., Prachalová, M., Drastík, V., 2008. Size selectivity insummer and winter diets of great cormorant (Phalacrocorax carbo): does it reflectseason-dependent difference in foraging efficiency? Waterbirds 31, 438–447.

orr, B.S., Hanisch, S.L., Buttcho, P.H., Fielder, D.G., 2012. Management of double-crested cormorants to improve sport fisheries in Michigan: three case studies.Hum. Wildl Interact. 6, 155–168.

ngström, H., 2001. Effects of great cormorant predation on fish populations andfishery. Comprehensive Summaries of Uppsala Dissertations from the Faculty ofScience and Technology, vol. 670., pp. 1–39.

schbaum, R., Veber, T., Vetemaa, M., Saat, T., 2003. Do Cormorants and fishermencompete for fish resources in the Väinameri (Eastern Baltic) area? In: Cowx, G.,Eschbaum, R., Veber, T., Vetemaa, M., Saat, T. (Eds.), Interactions Between Fishand Birds: Implications for Management. Blackwell Publishing Ltd., Oxford, pp.72–83.

ielder, D.G., 2010. Response of yellow perch in Les Cheneaux Islands, Lake Huron todeclining numbers of double-crested cormorants stemming from control activ-ities. J. Great Lakes Res. 36, 207–214.

remillet, D., Argetin, G., Schulte, B., Culik, B.M., 1996. Flexible foraging techniquesin breeding cormorants Phalacrocorax carbo and shags Phalacrocorax aristotelis:benthic or pelagic feeding? Ibis 140, 113–119.

eikinheimo, O., Setälä, J., Saarni, K., Raitaniemi, J., 2006. Impacts of mesh-size regu-

lation of gillnets on the pikeperch fisheries in the Archipelago Sea, Finland. Fish.Res. 77, 192–199.

eikinheimo, O., Pekcan–Hekim, Z., Raitaniemi, J., 2014. Spawning stock-recruitment relationship in pikeperch Sander lucioperca (L.) in the Baltic Sea,with temperature as an environmental effect. Fish. Res. 155, 1–9.

arch 164 (2015) 26–34 33

Hilborn, R., Walters, C.J., 1992. Quantitative Fisheries Stock Assessment. Choice,Dynamics and Uncertainty. Chapman Hall, New York.

Härkönen, T., 1986. Guide to the Otoliths of the Bony Fishes of the Northeast Atlantic.Danbiu Aps, Hellerup.

Jepsen, N., Klenke, R., Sonnesen, P., Bregnballe, T., 2010. The use of coded wire tagsto estimate cormorant predation on fish stocks in estuary. Mar. Freshw. Res. 61,320–329.

Johnson, J.H., Ross, R.M., McCullough, R.D., Mathers, A., 2010. A comparative analysisof double-crested cormorant diets from stomachs and pellets from two LakeOntario colonies. J. Freshw. Ecol. 25, 669–672.

Kangas, P., Bäck, S., Kauppila, P. (Eds.), 2003. Ehdotuksia Euroopan yhteisön vesipoli-tiikan puitedirektiivin (2000/60/EY) mukaiseksi rannikkovesien tyypittelyksiSuomessa. Suomen ympäristökeskuksen moniste 284. (in Finnish).

Keller, T., 1995. Food of Cormorants Phalacrocorax carbo sinensis wintering in Bavaria,Southern Germany. Ardea 83, 185–192.

Korhonen, K., (Bachelor’s thesis) 2010. Merimetson (Phalacrocorax carbo sinensis)poikasajan ravinnonkäyttö Saaristomerellä kesinä 2009-2010. Turku Universityof Applied Sciences, pp. 71 (in Finnish).

Le Cren, E.D., 1951. The length-weight relationship and seasonal cycle in gonadweight and condition in the perch (Perca fluviatilis). J. Anim. Ecol. 20, 201–219.

Lehikoinen, A., (M.Sc. thesis) 2003. Merimetson kannankehitys Suomessa–pesintämenestys, ravinto ja vaikutus muuhun pesimälinnustoon. University ofHelsinki, 39 pp. (in Finnish).

Lehikoinen, A., 2005. Prey-switching and diet of the great cormorant during thebreeding season in the Gulf of Finland. Waterbirds 28, 511–515.

Lehikoinen, A., 2006. Cormorants in the Finnish archipelago. Ornis Fennica 83, 34–46.Lehikoinen, A., Heikinheimo, O., Lappalainen, A., 2011. Temporal changes in the

diet of great cormorant (Phalalacrocorax carbo sinensis) on the southern coastof Finland – comparison with available fish data. Boreal Env. Res. 16 (Suppl. B),61–70.

Lilliendahl, K., Solmundsson, J., 2006. Feeding ecology of sympatric European shagsPhalacrocorax aristotelis and great cormorants Phalacrocorax carbo in Iceland.Mar. Biol. 149, 979–990.

Liordos, V., Goutner, V., 2007. Spatial patters of winter diet of the great cormorantin coastal wetlands of Greece. Waterbirds 30, 103–111.

Lunneryd, S.-G., Alexandersson, K., 2005. Födoanalyser av storskarv Phalacrocoraxcarbo i Kattegatt – Skagerrak. Finfo 2005 (11), 1–21 (in Swedish).

Mustamäki, N., Bergström, U., Ådjers, K., Sevastik, A., Mattila, J., 2013. Pikeperch(Sander lucioperca (L.)) in decline: high mortality of three populations in theNorthern Baltic Sea. AMBIO 42, http://dx.doi.org/10.1007/s13280-013-0429-z.

Östman, Ö., Bergenius, M., Boström, M.K., Lunneryd, S.-V., 2012. Do cormorantcolonies affect local fish communities in the Baltic Sea? Can. J. Fish. Aquat. Sci.69, 1047–1055.

Östman, Ö., Boström, M.K., Bergström, U., Anderson, J., Lunneryd, S.-V.,2013. Estimating competition between wildlife and human-A caseof cormorants and coastal fisheries in the Baltic Sea. PLOS ONE 8,http://dx.doi.org/10.137/journal.pone.0083763.

Pekcan-Hekim, Z., Urho, L., Auvinen, H., Heikinheimo, O., Lappalainen, J., Raitaniemi,J., Söderkultalahti, P., 2011. Climate warming and pikeperch year-class catchesin the Baltic Sea. AMBIO 40, 447–456.

Radu, V., 2005. Atlas for the Identification of Bony Fish Bones from ArchaeologicalSites. Contrast, Bucuresti.

Ridgway, M.S., 2010. A review of estimates of daily energy expenditure and foodintake in Cormorants (Phalacrocorax spp.). J. Great Lakes Res. 36, 93–99.

Ridgway, M.S., Fielder, D.G., 2013. Double-crested cormorants in the LaurentianGreat Lakes: issues and ecosystems. In: Taylor, W.W., Lynch, A.J., Leonard, N.J.(Eds.), Great Lakes Fisheries Policy and Management, A Binational Perspective.Michigan State University Press, East Lansing, Michigan, pp. 733–766.

Rudstam, L.G., vanDeValk, A.J., Adams, C.M., Coleman, J.T.H., Forney, J.L., Richmond,M.E., 2004. Cormorant predation and the population dynamics of walleye andyellow perch in Oneida Lake. Ecol. Appl. 14, 149–163.

Rusanen, P., Mikkola-Roos, M., Asanti, T., 1998. Merimetso Phalacrocorax carbo– Musta viikinki, Merimetson kannan kehitys ja siihen vaikuttavat tekijätItämeren piirissä ja Suomen ympäristö, luonto ja luonnonvarat sarja 182 (inFinnish).

Salmi, J., Salmi, P., Moilanen, P., 2010. Ammattikalastus ja merimetso: merestäelantoaan hankkivien näkemyksiä. Riista- ja kalatalous – Selvityksiä 1/2010 (inFinnish).

Seefelt, N.E., Gillingham, J.C., 2006. A comparison of three methods to investi-gate the diet of breeding double-crested cormorants (Phalacrocorax auritus)in the Beaver Archipelago, northern Lake Michigan. Hydrobiologia 567,57–67.

Setälä, J., Heikinheimo, O., Saarni, K., Raitaniemi, J., 2003. Verkon solmuvälinsuurentamisen vaikutus Saaristomeren ammattikalastuksen kuha- ja ahven-saaliin arvoon. Kala- ja riistaraportteja 297 (in Finnish).

Skov, C., Chapman, B.B., Baktoft, H., Brodersen, J., Brönmark, C., Hansson, L.-A., Hulthen, K., Nilsson, P.A., 2013. Migration confers survival benefitsagainst avian predators for partially migratory freshwater fish. Biol. Lett. 9,http://dx.doi.org/10.1098/rsbl.20121178.

Skov, C., Jepsen, N., Baktoft, H., Jansen, T., Pedersen, S., Koed, A., 2014. Cormorantpredation on PIT-tagged lake fish. J. Limnol. 73, 177–186.

Suter, W., Morel, P., 1996. Pellet analysis in the assessment of great cormorant Pha-lacrocorax carbo diet: reducing biases from otolith wear when reconstructingfish length. Colon. Waterbird. 19, 280–284.

Svetocheva, O., Stasenkova, N., Fooks, G., 2007. Guide to the Bony Fishes Otoliths ofthe White Sea. PINRO, Murmansk.

3 s Rese

V

V

4 J.A. Salmi et al. / Fisherie

eldkamp, R., 1995. The use of chewing pads for estimating the consumption ofcyprinidis by cormorants Phalacrocorax carbo. Ardea 83, 135–138.

eneranta, L., Urho, L., Lappalainen, A., Kallasvuo, M., 2011. Turbidity characterizesthe reproductions areas of pikeperch (Sander lucioperca (L.)) in the northernBaltic Sea. Estuar. Coast. Shelf Sci. 95, 199–206.

arch 164 (2015) 26–34

Vetemaa, M., Eschbaum, R., Albert, A., Saks, L., Verliin, A., Jürgens, K., Kesler, M.,Hubel, K., Hannesson, R., Saat, T., 2010. Changes in fish stocks in an Estonianestuary: overfishing by cormorants? ICES J. Mar. Sci. 67, 1–8.

Wheeler, A., 1978. Key to the Fishes of Northern Europe. Frederick Warne Ltd.,London.