This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

Author's personal copy

Habitat change by the formation of alien Crassostrea-reefs in theWadden Sea and its role as feeding sites for waterbirds

Alexandra Markert a,b,*, Wiebke Esser a,c, Dietrich Frank d, Achim Wehrmann a,b,Klaus-Michael Exo c

a Senckenberg am Meer, Marine Research Department, Südstrand 40, 26382 Wilhelmshaven, GermanybBiodiversity and Climate Research Centre, Senckenberganlage 25, 60325 Frankfurt, Germanyc Institute of Avian Research, “Vogelwarte Helgoland”, An der Vogelwarte 21, 26386 Wilhelmshaven, GermanydWadden Sea National Park Administration of Lower Saxony, Virchowstraße 1, 26382 Wilhelmshaven, Germany

a r t i c l e i n f o

Article history:Received 18 June 2012Accepted 5 August 2013Available online 15 August 2013

Keywords:Pacific oysterecosystem engineeringforaging behaviourwadersoystercatcherNorth Sea

a b s t r a c t

Non-indigenous Pacific oysters (Crassostrea gigas) have been invading the central Wadden Sea since 1998,predominantly settling on intertidal blue mussel (Mytilus edulis) beds which are increasingly trans-formed into Crassostrea-reefs. Pacific oysters are strong ecosystem engineers and the habitat change isconsidered to be a threat for waterbirds losing important feeding sites in the intertidal of the WaddenSea. This study has increased our understanding of the use of foraging habitats by birds according tochanging food resources.

During the spring and autumn migration period in 2007, we recorded bird densities at two reef typesvarying in Pacific oyster density and at the adjacent sand flat as a reference site. We also recorded feedingbehaviour, choice of prey and assessed peck and intake rate of three target species: Eurasian oyster-catcher Haematopus ostralegus, Eurasian curlew Numenius arquata and European herring gull Larusargentatus. To evaluate the use of the Crassostrea-reef in the central Wadden Sea, we compared birddensities of the target species at different intertidal feeding habitats in various regions and compared thebiomass intake of Eurasian oystercatcher feeding on different prey species. We show that Eurasianoystercatcher and Eurasian curlew have adapted to the new situation and learned to exploit the foodsupply offered by Crassostrea-reefs. While foraging mainly on Pacific oysters, Eurasian oystercatchersattained sustainable intake rates even though food resource at dense reef during autumn was very poordue to a lack in harvestable oysters. Consolidation of reefs limits the accessibility of prey for Eurasianoystercatchers whereas a successful recruitment of Pacific oysters enhances the suitability of the habitat.Eurasian curlew was promoted by the engineering effects of the Pacific oyster while feeding extensivelyon shore crabs at the reefs. In contrast, European herring gulls appear hampered in foraging during lowtide and hereby negatively affected by the habitat change from Mytilus-beds to Crassostrea-reefs.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

TheWadden Sea is themost important staging site formigratorybirds on the East Atlantic flyway and the most important breedingsite for waders in central Europe. Annually, 10e12 million migra-tory waterbirds use the habitat diversity of the Wadden Sea forroosting, breeding and feeding (Meltofte et al., 1994; Laursen et al.,2010). Due to their considerably high biomass and productivity,intertidal Mytilus-beds are important foraging sites for waders and

waterbirds (Nehls et al., 1997). In addition to common eiders(Somateria mollissima), Eurasian oystercatchers (Haematopusostralegus) are the main avian predators of blue mussels in theWadden Sea. Other bird species such as the European herring gull(Larus argentatus), Eurasian curlew (Numenius arquata), commonredshank (Tringa totanus), black-headed gull (Chroicocephalus rid-ibundus) and dunlin (Calidris alpina) mainly feed on associatedmacrofauna, such as polychaetes and crustaceans living betweenthe blue mussels (Nehls et al., 1997; Van de Kam et al., 2004).Decreasing blue mussel stocks and deteriorating body conditionindices of blue mussels in most of the regions of the Wadden Seamay have decreased abundances of especially shellfish eating birds,but causes for significantly declining trends of population sizes ofEurasian oystercatcher and European herring gull during the last

* Corresponding author. Senckenberg am Meer, Marine Research Department,Südstrand 40, 26382 Wilhelmshaven, Germany.

E-mail address: amarkert@senckenberg.de (A. Markert).

Contents lists available at ScienceDirect

Estuarine, Coastal and Shelf Science

journal homepage: www.elsevier .com/locate/ecss

0272-7714/$ e see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.ecss.2013.08.003

Estuarine, Coastal and Shelf Science 131 (2013) 41e51

Author's personal copy

two decades in the Wadden Sea remain unclear (Ens et al., 2009;Laursen et al., 2010).

Meanwhile, the habitat type “Mytilus-bed” experiences novelchanges throughout the entire Wadden Sea. Non-indigenous Pa-cific oysters (Crassostrea gigas) are transforming intertidal Mytilus-beds into Crassostrea-reefs. Oysters were imported for aquacultureto the Dutch Oosterschelde estuary in 1964 (Drinkwaard, 1999)(Fig. 1) and were deliberately introduced to the western WaddenSea (The Netherlands) where wild oysters have been propagatingsince 1983 (Fey et al., 2010), settling predominantly on intertidalMytilus-beds. In the central Wadden Sea, the first occurrence ofPacific oysters was observed in 1998, invading the area by larvaldrift via residual currents (Wehrmann et al., 2000). After an initiallag phase of the bioinvasion, the regional abundance of Pacificoysters increased exponentially. Since 2006, all intertidal Mytilus-beds in the central Wadden Sea have been colonized (Nehls et al.,2009). Juvenile oysters may grow 7 cm in their first year and reachshell lengths of up to 25 cmwithin only a few years (Schmidt et al.,2008). Oyster larvae prefer settling on conspecifics and the adultindividual stays cemented to its substratum of settlement. Hence,shells of dead individuals contribute to the physical structure ofthe habitat. Increasing densities of this strong physical ecosystemengineer (Jones et al., 1994) lead to the formation of dense reefs.Both rapid growth and reef formation protects the oysters againstpredators and generally support invasion success (Colautti et al.,2004). In contrast to the flexible meshworks and dynamicoccurrence of Mytilus-beds (Nehls and Thiel, 1993), Pacific oysterreefs constitute new rigid and persisting bioconstructions (Smaalet al., 2005).

The invasion and spread of the Pacific oyster overlap temporallywith a long-term decline of blue mussels in the area. In the centralWadden Sea, blue mussel density is now increasing in the shelter ofCrassostrea-reefs while habitat characteristics are mainly engi-neered by Pacific oysters (Nehls et al., 2009). Dense reefs may beconstructed by about 2000 live and dead oysters per m2 and mayinhabit even more live blue mussels than live oysters. Blue musselpopulations are built up by all size classes, especially individuals

between 30 and 40 mm now contribute to the population structurein Crassostrea-reefs (Markert et al., 2010). Markert et al. (2010)detected an increase in biodiversity, and high abundance andbiomass of associated macrofauna at areas dominated by Pacificoysters. They found prey species for birds, such as shore crabs orpolychaetes, in high abundances in the reefs. However, the pres-ence of prey does not imply accessibility at the same time, as su-perficial structure and habitat matrix differ essentially betweenCrassostrea-reefs and Mytilus-beds. Pacific oysters create complexthree-dimensional rigid biogenic structures which may influencefood detection as well as food accessibility for birds (Zwarts andWanink, 1993; Zwarts et al., 1996a; Ruesink et al., 2005).Scheiffarth et al. (2007) suggested that no difference in habitatquality has to be expected for birds feeding on associated fauna butalso emphasised the lack of data, especially for the Wadden Sea asthemost important stop-over site for waders using the East AtlanticFlyway.

The present study aimed to increase knowledge of the intensityand quality of the use of Crassostrea-reefs by waterbirds. Wefocused on three target species (i) Eurasian oystercatcher, (ii)Eurasian curlew and (iii) European herring gull (hereafter: oyster-catcher, curlew, herring gull) to answer the following questions:

� Are the birds using Crassostrea-reefs as feeding sites and is theecological function of Mytilus-beds replaced?

� Which of the target bird species benefits or suffers from thehabitat change?

� What kind of long-term effect can be expected from the habitatchange?

The data for bird density are compared with different feedinghabitats in various intertidal regions andwith regional pre-invasiondata using a BACI (Before-After/Control-Impact) approach. It is thefirst study on foraging behaviour and energy intake rates ofoystercatcher at Crassostrea-reefs. The investigation increases ourunderstanding of the impact of a non-indigenous ecosystem engi-neering species affecting ecosystem functions.

Fig. 1. Location of the Pacific oyster reef, study plots at the habitat types dense reef, reef carpet and sand flat (photos for details) and position of the observation tower at theDornumer Nacken in the central Wadden Sea, North Sea (tidal flat area in light grey, Wadden Sea area indicated by dashed lines).

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e5142

Author's personal copy

2. Methods

2.1. Study area and habitat characteristics

The study area (Dornumer Nacken) is located in the backbarriertidal flats of the island Baltrum, central Wadden Sea (North Sea(Fig. 1). Three habitat types with different characteristics in Pacificoyster density were distinguished: (i) dense reef, (ii) reef carpet and(iii) sand flat (Fig. 1). Dense reef and reef carpet are parts of a formerMytilus-bed with an extension of 8.3 ha. The Mytilus-bed trans-formed to a Crassostrea-reef between 2003 and 2005. Analogous tothe structure of Mytilus-beds (Nehls et al., 2009), the Crassostrea-reef consists of bivalve covered patches e in the followingexpressed as “reef density” in kg total shell mass of live and deadPacific oysters per m2 patch e and irregularly interspersed openspaces where biodeposits accumulate to mounts or which aredeepened to tidal pools. The heterogeneous reef structure includessmall gullies which mainly open into the creek bordering thesoutheast of the oyster reef. Mean coverage of reef patches was 40%in the dense reef and 18% in reef carpet. In 2007, the total area of theCrassostrea-reef was characterized by varying reef densities be-tween 10 and 90 kg shell mass m�2 patch.

The habitat type dense reef was situated close to the creek andparts were located in the lower intertidal which, especially duringautumn, stayed submerged for a longer time than reef carpet. Meanabundance (see photos Fig. 1) within 1 m2 patch in dense reef was884 � 153 live oysters larger 25 mm with a wet mass of53.15�17.67 kg, equivalent to 287.8� 103.0 g AFDM, and 915� 422left valves of dead oysters (>25 mm). Reef density of 1 m2 patch indense reef corresponded to a shell mass of 61.9 � 4.5 kg (55.7% live).The same m2 inhabited 697 � 299 live blue mussels larger 10 mmwith a wet mass of 4.7 � 1.9 kg, equivalent to 71.0 � 29.3 g AFDM.Abundance of the associated macrofauna (1 mm mesh size) was10,650 � 1746 individuals m�2 which corresponded to a biomass of194.4 � 212.3 g AFDMm�2 (data March 2007; Appendix Table A.1).

The habitat type reef carpet was located northwest of dense reef.Mean abundance within 1 m2 patch in reef carpetwas 680� 69 liveoysters (>25 mm) with a wet mass of 31.46 � 0.71 kg, equivalent to167.0 � 2.9 g AFDM, and 180 � 35 left valves of dead oysters(>25 mm). Reef density of a 1 m�2 patch in reef carpet corre-sponded to a shell mass of 23.3 � 0.7 kg (88.6% live). The same m2

inhabited 549 � 332 live blue mussels larger than 10 mm with awet mass of 4.0 � 2.5 kg, equivalent to 59.1 � 37.4 g AFDM. Theabundance of the associated macrofauna was 10,050 �2666 individuals m�2 which corresponded to a biomass of328.3 � 131.3 g AFDM m�2 (data March 2007; Appendix Table A.1).

The oyster reef is surrounded by a homogeneous structuredsand flat which is the dominant habitat type at the DornumerNacken and which has no epibenthic bivalve coverage. The habitattype sand flat south of the reef was colonized by 5450 � 3247macrofaunal individuals m�2 which corresponded to a biomass of121.0 � 55.3 g AFDM m�2. Whereas its abundance was dominatedmainly by smaller polychaetes (68%) and oligochaetes (18%), thehighest share in biomass was observed for the edible cockle Cera-stoderma edule (41%), followed by the sand mason Lanice conchilega(30%) and the lugworm Arenicola marina (16%) (data March 2007;Appendix Table A.1).

2.2. Bird data

To analyze differential habitat use of birds, low tide counts(spotting scope: Optolyth TBS 80, 20e60 � 60) were carried outfrom a 5.5m high observation tower duringmainmigration periodsfrom March through May during spring and from August throughOctober during autumn 2007. Three plots (Fig. 1) of 30 m � 60 m

(0.18 ha) per habitat type were marked by wooden sticks. For eachplot, bird species and their abundance was recorded by low tidecounts during 5 days per month and counts at intervals of 20 minfor the whole exposure period during daytime (mainly 2 h beforethrough to 2 h after low tide). At the same time, we recordedwhether or not the registered bird was showing foraging behaviour(“scan sampling” after Martin and Bateson, 1986). During autumn,we also registered the small scale distribution of birds at coveredpatches and interspersed open spaces of the habitat type dense reefand reef carpet.

Three target species were selected: oystercatcher and herringgull are common bird species in the Wadden Sea and showdeclining population trends, curlew was selected due to its wide-spread appearance in the Wadden Sea. For these species werecorded detailed foraging behaviour, e.g. diet composition, pecksand successful pecks per minute and success rate, during 4e5additional days per month and habitat type (“focal animal sam-pling” after Martin and Bateson, 1986). Additionally, biomass intakein gram ash free dry mass per minute [g AFDM/min] was deter-mined based on the size of ingested prey (for methods see Petersenand Exo, 1999). For oystercatchers, we additionally distinguishedbetween pecks and successful pecks of adult and immature in-dividuals and assessed mean handling-time for feeding successwhen preying on varying Pacific oyster sizes. We observed the shellopening technique of the birds, collected remaining right valves ofconsumed Pacific oysters, measured their shell length and analysedthe contact area by allocating typical predation marks according to4 sectors.

2.3. Data processing e statistical analyses

Bird density was calculated by averaging total bird numbersover all observation units per exposure period during daytime perday and habitat type to daily means and afterwards to mean birds/ha � standard deviation per exposure period during daytime ac-cording to migration period (spring and autumn).

Depending on the approach and database, we applied Mann-Whitney-U test or for analyses of multiple factors KruskaleWallistest after Bonferroni-correction (p/3) followed by post hoc testing(Sachs, 1997). Significance is displayed by standardised levels to* ¼ p < 0.05, ** ¼ p < 0.01 and *** ¼ p < 0.001.

3. Results

3.1. Bird density

We registered a total of 19 bird species during both migrationperiods of which 13 were present at dense reef, 10 at reef carpet and12 at sand flat (Appendix Table A.2). Oystercatcher, curlew, herringgull, black-headed gull, common gull and common redshank wereobserved at all habitat types during both migration periods. Duringspring migration, all habitat types were used by similar total birdnumbers, varying between 9.71 � 4.12 birds/ha at sand flat and13.47 � 7.33 at reef carpet (Fig. 2). Overall bird density was higherduring the autumnmigration than during spring which was mainlycaused by higher bird numbers of the consistently present andmost numerous species, e.g. the target species of the investigatione oystercatcher, curlew and herring gull. The total abundance of thethree target species represented 62e87% of all birds counted perhabitat type and season, excluding observations at sand flat duringautumn. 46.85 � 33.64 birds/ha at sand flat during autumnwas thehighest bird number of all habitat types during both seasons, beingsignificantly higher than during spring and significantly higherthan at reef carpet or dense reef during autumn. These differenceswere mainly due to the significantly higher bird number of

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e51 43

Author's personal copy

common gull (0.03e18.33 birds/ha; p < 0.001) and due to theexclusive and high abundance of dunlin at sand flat (10.27 birds/ha;Appendix Table A.2).

Curlew was the most abundant bird species at both reef typesduring spring (Fig. 2) and used reef carpet with significantly highernumbers during autumn than during spring. Moreover, density atreef carpet during autumn was significantly higher than at densereef and at sand flat.

During spring, oystercatchers were observed in higher densitiesat both reef types than at sand flat (Fig. 2) and compared to densereef, there was a significantly lower density at sand flat. During theautumn migration period, oystercatchers had significantly highernumbers at sand flat than during spring.

Excluding common gull and dunlin during autumn, herring gullwas the most common bird species at sand flat during bothmigration periods (Fig. 2). Significantly more herring gulls wereencountered at dense reef during autumn than during spring.Densities at dense reef and at sand flat during autumn were signif-icantly higher than at reef carpet where bird numbers stayed at alow level during both migration periods.

3.2. Foraging behaviour

Oystercatcher and curlew used all habitat types mainly forforaging (Table 1). The relative number of foraging birds at allhabitat types was lower during autumn, in particular for oyster-catcher and especially at reef carpet. Herring gulls showed lessforaging behaviour than the other two target species at all habitattypes during spring migration, in particular at the reef types. Anespecially low foraging behaviour of herring gulls was observed atreef carpet during autumn, whereas foraging behaviour at dense reefincreased.

During autumn, bird species were observed by their small scaledistribution at dense reef and reef carpet. Oystercatchers were pre-dominantly observed at exposed oyster covered patches of bothreef types (>95%) while curlews only used the patches of reef carpetless frequently (dense reef 94%, reef carpet 72%), respectively morebirds were registered at interspersed open spaces. Herring gullscomprised a small proportion at the exposed covered patches ofboth reef types (dense reef 24%, reef carpet 34%) but their occurrencewas high above dense reef when the lower parts stayed submergeddue to the high water level in autumn (64%).

3.3. Diet composition

The diet composition of oystercatcher was dominated by bi-valves at all habitat types whereas curlew and especially herring

Table 1Foraging behaviour of the target species during 15 exposure periods in daytime at dense reef, reef carpet and sand flat during spring and autumn migration 2007, given arepercentages of foraging birds (N ¼ total number of observed birds). Significance level superscripted for foraging vs. other behaviour within season and significance level listedfor foraging behaviour between spring and autumn (Mann-Whitney-U Test).

dense reef reef carpet sand flat

Observedbirds (N)

Foragingrate [%]

Foraging behaviourspring/autumn

Observedbirds (N)

Foragingrate [%]

Foragingbehaviourspring/autumn

Observedbirds (N)

Foragingrate [%]

Foraging behaviourspring/autumn

SpringEurasian oystercatcher 380 96*** 282 97*** 66 95***Eurasian curlew 467 98*** 502 97*** 307 98***European herring gull 108 68 64 60 286 86**

AutumnEurasian oystercatcher 178 91** * 547 78* *** 368 86***Eurasian curlew 271 96*** 677 89*** ** 206 93***European herring gull 333 88*** 84 52 450 88***

Fig. 2. Mean bird density in numbers per hectare [ha] þ standard deviation of allobserved bird species and of the target species per exposure period during daytime atdense reef, reef carpet and sand flat during spring and autumn migration 2007. Sig-nificance level between habitat types within season below bars and of a single habitattype between seasons above bars (KruskaleWallis test after Bonferroni correction).N ¼ 15 observation days per season.

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e5144

Author's personal copy

gull mainly fed on crustaceans (Fig. 3). At both reef types, oyster-catcher primarily fed on Pacific oysters. At dense reef, besides someblue mussels during spring, oystercatcher exclusively preyed onPacific oysters. At reef carpet, diet composition was more diversebut also dominated by Pacific oysters. During autumn, proportionsof other prey organisms such as cockles (Cerastoderma edule) andblue mussels increased and oystercatchers also fed on shore crabs(e.g. Carcinus maenas) and polychaetes. At sand flat, oystercatchersconsumed more cockles than polychaetes during spring. Duringautumn, the proportion of polychaetes increased while fewercockles and additionally some shore crabs were consumed.

The diet composition of curlew was comparable at both reeftypes during both migration periods. It was dominated by shorecrabs and moderate proportions were observed for brown shrimp(Crangon crangon), especially at reef carpet during autumn. Poly-chaetes and starfish (Asterias rubens) were additional prey speciesduring spring while, during autumn, sea anemones supplementedthe diet. Also at sand flat, shore crabs were the preferred preyspecies but proportions were lower during autumn when morepolychaetes and more brown shrimp were consumed.

Herring gulls seemed to be restricted to prey on shore crabs atall habitat types even though moderate proportions of brownshrimp at reef carpet and at sand flat as well as some starfish at bothreef types were added to its diet during autumn. Interspecifickleptoparasitism was observed for herring gulls which stole oysterflesh from oystercatchers before the bird was able to swallowits prey.

3.4. Pecks, successful pecks and biomass intake

The number of pecks per minute of oystercatcher was compa-rable at all habitat types during both seasons (Table 2). Duringspring migration, oystercatcher had significantly higher feedingsuccess at sand flat than at both reef types. During autumn, suc-cessful pecks showed no differences between habitat types. Foroystercatcher, we registered pecks/min and success rate of adult(N ¼ 3 during spring, 15/autumn) and immature (11/spring, 6/autumn) individuals at dense reef during spring and autumn. Pecks/min of adults or immature birds did not differ significantly duringor between seasons (Mann-Whitney-U test) whereas adultsreached a significantly (p < 0.05) higher success rate (22%) thanimmature birds (7%) during spring. Success rate was lower duringautumn and no difference in feeding success of adults (6%) andimmature birds (3%) was observed.

For curlew and herring gull, pecks/min at sand flat duringboth seasons was significantly higher than at dense reef(Table 2). With reference to the spring migration period, pecks/min of both species at dense reef and at sand flat almostremained unchanged in autumn but increased at reef carpet,being significantly higher than at dense reef. Feeding success ofcurlew at all habitat types and during both seasons remained ata moderate level. While during spring no differences occurred,feeding success of curlew was significantly lower at dense reefthan at reef carpet or sand flat during autumn. For herring gull,feeding success at dense reef and at sand flat was comparable

Fig. 3. Diet composition in relative number of prey individuals [%] consumed by the target species at the habitat types dense reef, reef carpet and sand flat during spring and autumnmigration 2007. Crust. ¼ Crustacea. N ¼ Total number of prey individuals.

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e51 45

Author's personal copy

during spring, no successful pecks were observed at reef carpet,but numbers of observation were low. During autumn, feedingsuccess of herring gull was significantly higher at dense reefthan at sand flat.

Biomass intake of oystercatcher at all habitat types duringboth seasons showed no significant differences but mean valueswere conspicuously lower at sand flat in autumn (Fig. 4).Compared to oystercatcher, biomass intake of curlew was higherduring both seasons and at all habitat types. Curlew reached asignificantly higher biomass intake at sand flat during spring thanat reef carpet whereas no differences among habitat typesoccurred in autumn. In contrast to fewest pecks/min at dense reef,considering all species and all habitat types, herring gull had thehighest biomass intake at dense reef during both seasons beingsignificantly higher than at sand flat in autumn (Fig. 4).

3.5. Pacific oyster predation by oystercatchers

At both reef types, oystercatcher primarily fed on Pacific oysters(Fig. 3). The birds were mainly following the falling and risingwaterline to look for gaping individuals. Mean shell length ofPacific oysters consumed by oystercatchers during spring was30.5 � 7.7 mm (range 17e53 mm) and during autumn46.6 � 11.3 mm (range 22e68 mm) (Fig. 5). The difference betweenseasons was significant (p < 0.001, Mann-Whitney-U Test). Oys-tercatchers preyed on the most abundant shell length class of thelive Pacific oyster stock during spring. Due to proliferation, Pacificoysters up to 100 mm are abundant during autumn but were notpreyed on by oystercatchers. Oystercatcher needed on average19.25�14.32 s (range 11e54 s; N¼ 8) to open up Pacific oysters outof different shell length classes.

Oystercatchers inserted their bill in between the valves ofgaping individuals (stabbing technique) or they pecked at aparticular side of the thin shell margin of closed individuals tolaterally insert their bill. In both cases, the valves were then prisedopen while the shell margin cracked by leaving typical marks(Fig. 5). 60% of the predation marks at consumed and collectedPacific oyster shells were at the area in opposite to the hinge while26% were located in opposite to the adductor muscle and only 14%were counted directly at the side of the adductor.

4. Discussion

Food provision, food quality and intake rate are decisive vari-ables for birds to choose a feeding site but also population size andsubsequently time of the year can strongly affect the intensity of ahabitat use (Goss-Custard et al., 1982). The invasion and massive

Table 2Pecks and successful pecks per minute (min)� standard deviation (SD) and success rate [%] of the target species at dense reef, reef carpet and sand flat during spring and autumnmigration 2007. Significance between habitat types: dense reef vs. reef carpet (a), dense reef vs. sand flat (b) and reef carpet vs. sand flat (c), significance level superscripted(KruskaleWallis test after Bonferroni correction). N ¼ Number of observations.

dense reef reef carpet sand flat Significance level

a, b a, c b, c

N Pecks/min (SD) N Pecks/min (SD) N Pecks/min (SD)

Successfulpecks/min (SD)

Successrate [%]

Successfulpecks/min (SD)

Successrate [%]

Successfulpecks/min (SD)

Successrate [%]

SpringEurasian oystercatcher 44 12.15 (3.62) 10 12 13.31 (2.49) 6 28 14.73 (4.73) 15 b*, c*

1.21 (0.74) 0.84 (0.58) 2.19 (1.57)Eurasian curlew 10 16.54 (5.52) 6 19 14.85 (2.94) 4 28 24.53 (7.77) 5 b**, c**

1.06 (0.68) 0.60 (0.73) 1.12 (0.78)European herring gull 3 4.82 (3.06) 10 1 3.75 (0) w 40 12.23 (4.31) 7 b***

0.49 (0.44) 0 (0) 0.90 (0.87)AutumnEurasian oystercatcher 21 11.11 (3.78) 4 20 11.76 (3.58) 6 20 15.06 (8.96) 7

0.45 (0.42) 0.68 (0.74) 1.02 (0.81)Eurasian curlew 25 15.49 (5.55) 4 20 22.01 (5.72) 5 22 24.09 (6.77) 6 a***, b***

0.58 (0.48) 1.06 (0.65) 1.55 (1.19) a*, b**European herring gull 20 4.34 (2.75) 15 20 8.52 (2.85) 6 20 9.83 (3.51) 2 a***, b***

0.66 (0.52) 0.52 (0.71) 0.22 (0.43) b*

Fig. 4. Biomass intake in gram ash free dry mass per minute [g AFDM/min] þ standarddeviation of the target species at dense reef, reef carpet and sand flat during spring andautumn migration 2007. Significance level between habitat types within season belowbars (KruskaleWallis test after Bonferroni correction). Note different scales of y-axis.Numbers of observations see (Table 2).

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e5146

Author's personal copy

spread of the non-indigenous strong physical ecosystem engineer,the Pacific oyster, and the subsequent increasing transformation ofintertidal native Mytilus-beds into Crassostrea-reefs has led to theconcern that birds may lose feeding habitats since Mytilus-bedsserve as important food source for resident and migratory birdpopulations. In accordance with the general seasonal pattern(Petersen and Exo,1999; Laursen et al., 2010) and differences in birdnumbers observed at Mytilus-beds between seasons (Goss-Custardet al., 1982), densities of all observed birds at the studied reef typeswere higher during autumn than during spring migration. Withinthis study and before oyster invasion at the same site in the 1990s(Frank et al., 1997), disproportional higher bird densities duringautumn than during spring were observed at the sand flat (spring1994e96/2.42 to 2007/9.71 birds/ha, autumn 1993e95/29.35 to2007/46.85). In both studies, this can mainly be traced back to thestrong presence of common gulls and dunlins. These species areknown to highly frequent the Wadden Sea especially duringautumn migration. Although population sizes of common gull anddunlin have had stable populations since the 1990s (Laursen et al.,2010), the sand flat was used more intensively within the presentstudy than in the 1990s. Spatially-extended ecosystem engineeringeffects of Crassostrea-reefs impacting on adjacent areas of oysterreefs (Ruesink et al., 2005; Van der Zee et al., 2012) may haveincreased food availability at the sand flat in 2007.

4.1. Habitat use of the target species

Irrespective of regional contrasting population sizes (Laursenet al., 2010), numbers of herring gull at Mytilus or cockle beds(Umland, 2000; Van de Kam et al., 2004; Scheiffarth et al., 2007)

vary strongly and exceptionally high numbers were only observedwhen high amounts of blue mussel spat were available (Hilgerlohet al., 1997). Recorded densities were often higher than at ourstudy site (Table 3), irrespective of the year of investigation.Compared to bird numbers at our study site before oystersestablished (Frank et al., 1997), fewer herring gulls visited bothreef types during spring or reef carpet during autumn than theMytilus-bed in the 1990s. In contrast, bird number at dense reefduring autumn 2007 was considerably higher and comparable toother studies. Longer submergence time of lower parts at densereef due to tidal conditions during autumn may have impacted onthe differential use which we observed as a proportional shift inbird numbers of all target species at the two reef types betweenseasons. The preferred foraging area of oystercatcher and curlew,respectively the area of exposed oyster patches, was reduced andnumbers were lower than in spring. In contrast, the submergedpatches were used by high numbers of herring gulls. Herring gullsforage intensively in front of Crassostrea-reefs during falling andrising tide, especially at gullies, but these areas were not investi-gated within the present study. A similar observation was made byBaptist (2005) when herring gulls were the first to arrive above aCrassostrea-reef in the Oosterschelde but also the density of thebirds before reef exposure was not assessed. Herring gulls may besensitive to the sharp surface structure of exposed dense patchesor the habitat complexity hinders detection and accessibility ofprey such as shore crabs while blue mussel spat may completelystay out of reach. However, once detected prey was pecked, mainlycrabs, success rate and biomass intake was high (Table 2; Fig. 4).Despite lower feeding success and lower biomass intake than atboth reef types, high numbers of herring gulls used the reef

Fig. 5. Predation of Pacific oysters by Eurasian oystercatcher during spring and autumn 2007. (A) Shell length frequency distribution of consumed Pacific oysters as relative number[%] of collected shells, significance level of mean shell lengths between seasons, and the relevant part of the Pacific oyster stock up to 100 mm shell length as live individuals (Ind.)per m2 according to shell length class. Stock data for spring from June 2007 and for autumn from September 2006 as prospective population structure after a cohort shift fromspring to autumn 2007 (Markert, unpublished). Arrows point to the most consumed shell length class in spring (grey) and autumn (white). (B) View of the exterior of a Pacific oystervalve with a predation mark at the lower right and schematic view in the same orientation with 4 sectors, proportion of predation marks of collected shells according to sector.N ¼ Number of shells collected.

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e51 47

Author's personal copy

adjacent sand flat and numbers mostly exceeded observations atMytilus or cockle beds in other studies. We assume that foragingon high abundances of crabs which leave and return late in thetidal cycle by gullies and creeks of the reefs may make the birdsstay in the area.

The curlew is a common wader in the Wadden Sea with largestpopulations in the western and smallest populations in thenorthern Wadden Sea. Contrasting regional trends in populationsizes are observed in the recent decades as increasing numbers areregistered in the western, stable populations in the central anddecreasing trends are observed in the northern Wadden Sea(Laursen et al., 2010). Curlew is known to mainly feed on crusta-ceans and polychaetes (Petersen and Exo, 1999) on muddy tidalflats and cockle beds are also visited frequently (Van de Kam et al.,2004). Bird numbers of curlew visiting Mytilus-beds as foraging

habitat are low in all parts of theWadden Sea with lowest densitiesin the northern Wadden Sea (Table 3). We observed more curlewusing dense reef or reef carpet than sand flat and densities at thereefs mainly exceed numbers observed at other habitat types inprevious studies. Birds find high densities of crabs at the reefs(Appendix Table A.1; Markert et al., 2010) which may be easy toaccess by their long beak, reaching deep between rigid oysterclusters. Reef carpet during autumn of this study had the highestdensity registered at different habitat types in the entire WaddenSea as shown by decades of investigation. A high number of curlewwas also attracted to use the oyster reef Zandkreek in the Ooster-schelde estuary, The Netherlands (Baptist, 2005).

The oystercatcher is one of the most common waders in theWadden Sea. The transformation of Mytilus-beds into Crassostrea-reefs was assumed to have an impact especially on blue mussel

Table 3Density in bird numbers per hectare [ha] per exposure period during daytime of the target species at different habitats and in various study areas.

Bird density [numbers/ha] per exposure periodduring daytime

Period Area (country code) Reference

Eurasianoyster catcher

Eurasiancurlew

Europeanherringgull

HabitatCrassostrea-reef (dense) 4.2 4.5 1.2 S 2007 Dornumer Nacken, c WS (D) This studyCrassostrea-reef (carpet) 3.5 5.0 0.7 S 2007 Dornumer Nacken, c WS (D) This studyCrassostrea-reef (dense) 2.3 3.2 3.9 A 2007 Dornumer Nacken, c WS (D) This studyCrassostrea-reef (carpet) 7.3 8.3 1.0 A 2007 Dornumer Nacken, c WS (D) This studyCrassostrea-reef 2.6 0.2 2.8 S 2005 Uthoerner Aussenwatt, n

WS (D)Schieck 2010

Crassostrea-reef 1.3 3.7 0.01 23.08. 2005 Zandkreek, Oosterscheldeestuary (NL)

Baptist 2005

Crassostrea-reef 1.2 0.9 2.5 07.09. 2005 Vondelingsplaat, Oosterscheldeestuary (NL)

Baptist 2005

Crassostrea-reef 2* n.s. 1** Summer 2002 Isla del Jabalí, NorthernPatagonia (AR)

Escapa et al., 2004

Mytilus-bed 0.7 0.7 1.5 S 1994e96 Dornumer Nacken, c WS (D) Frank et al., 1997Mytilus-bed 1.2 0.9 1.2 A 1993e95 Dornumer Nacken, c WS (D) Frank et al., 1997Mytilus-bed 7.3 1.0 9.1 A 1993 Uthoerner Aussenwatt, n

WS (D)Scheiffarth et al., 2007

Mytilus-bed (new) 32 n.s. 70 A 1994 Neuharlingersieler Nacken,c WS (D)

Hilgerloh et al., 1997

Mytilus-bed (old) 1.7 0.5 2.3 A 1999 Mellum island, c WS (D) Umland 2000Mytilus-bed 18 n.s. n.s. Winter 1976/77 Texel, w WS (NL) Zwarts and Drent,

1981/Van de Kamet al., 2004

Mytilus-bed (old) 30 5 4 S-A 1971e73 Schiermonnikoog, w WS (NL)

Mytilus-bed (oldwith spat)

4.0 1.3 0.7 Summer 1987 Schiermonnikoog, w WS (NL) Ens and Alting 1996

Mytilus-bed (oldwithout spat)

0.9 0.8 0.5 Summer 1987 Schiermonnikoog, w WS (NL) Ens and Alting 1996

Mytilus-bed (withcockles)

78 n.s. n.s. 17.10. 1984 Slikken van Vianen,Oosterscheldeestuary (NL)

Meire 1996

Cockle bed 0-50 n.s. n.s. All year 1977e86 Frisian coast, w WS (NL) Zwarts et al., 1996bCockle bed (old) 10 5 0.2 Summer-A 1980/81 Frisian coast, w WS (NL) Van de Kam et al., 2004Cockle bed (cockle/

Mytilus spat low)8.3 0.7 0.5 A 2001 Ameland, w WS (NL) Ens et al., 2004

Cockle bed (cockle/Mytilus spat high)

11.5 1.3 7.9 A 2001 Ameland, w WS (NL) Ens et al., 2004

Lanice flat 13.6 5.2 1.3 A 1994 Neuharlingersieler Nacken,c WS (D)

Petersen and Exo 1999

n.s. ¼ not specified.* American oystercatcher Haematopus palliatus.** Kelp gull Larus dominicanus.S ¼ spring (MarcheMay).A ¼ autumn (AugeOct).w ¼ western.c ¼ central.n ¼ northern.WS ¼ Wadden Sea.D ¼ Germany.NL ¼ The Netherlands.AR ¼ Argentina.

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e5148

Author's personal copy

feeding birds such as the oystercatcher. The intensity of the use offoraging habitats is highly variable and numbers at Mytilus-bedsrange between 0 and 78 birds/ha (Table 3). Higher numbers ofoystercatchers at Mytilus- or cockle beds were observed in olderstudies, especially decades ago, and highest densities are docu-mented for thewesternWadden Seawhere largest population sizesare registered (Laursen et al., 2010). Densities of oystercatchers atboth reef types were higher than at other Crassostrea-reefs in theWadden Sea (Baptist, 2005; Schieck, 2010) or in Argentina (Escapaet al., 2004). Densities exceed those observed in 1999 at a Mytilus-bed in the central Wadden Sea (Umland, 2000) and in particularnumbers observed between 1993 and 1996 at the same site beforeoysters established (Frank et al., 1997). In contrast, higher numbersthan in 2007 were observed at another Mytilus-bed in 1994 wherethe birds were attracted by the availability of blue mussel spat(Hilgerloh et al., 1997). During the past three decades, populationsizes of oystercatchers have continuously decreased in all parts ofthe Wadden Sea and, especially in the last decade, a significantdecrease has been observed in the central and northern WaddenSea during autumn (Laursen et al., 2010). The decline was related todecreasing stocks of blue mussels due to failures in spatfall anddeteriorating body indices (Laursen et al., 2010) which may be

responsible for low numbers at the study site in the 1990s. Theintense use of our study site in 2007may be related to the observedshift in main prey species, as oystercatchers favoured predatingPacific oysters at both reef types. Alternate prey, such as medium-sized blue mussels or cockles, was available in high abundances(Appendix Table A.1) but contributed to a larger proportion of thedietary spectrum only at reef carpet. Especially during autumn,when fewer harvestable Pacific oysters were available, most oys-tercatchers were observed at reef carpet where the accessibility ofmussels, cockles, polychaetes or crabs was presumably easier thanat dense reef.

4.2. Pacific oyster predation by oystercatchers

There is little evidence worldwide that oystercatchers prey onoysters. Dewar (1922) reported Eurasian oystercatchers feeding onEuropean oysters Ostrea edulis, Tomkins (1947) and also Nol andHumphrey (1994) found American oystercatchers Haematopuspalliatus which opened Eastern oysters Crassostrea virginica andCampbell (1966) and Butler and Kirbyson (1979) found Black oys-tercatchers H. bachmani preying on Pacific oysters. The oyster-catcher was observed to feed on cultured Pacific oysters at the

Table 4Biomass intake in gram ash free dry mass per minute [g AFDM/min.] of main prey organisms of Eurasian oystercatcher at different habitats and in various study areas.

Main prey organism/s (habitat) Biomass intake [g AFDM/min] Period Area (country code) Reference

Eurasian oystercatcher

Mean Range

Pacific oyster (Crassostrea-dense reef) S 2007 Dornumer Nacken, c WS (D) This studyPacific oyster (Crassostrea-reef carpet) 0.07 S 2007 Dornumer Nacken, c WS (D) This studyPacific oyster (Crassostrea-dense reef) 0.05 A 2007 Dornumer Nacken, c WS (D) This studyPacific oyster (Crassostrea-reef carpet) 0.07 A 2007 Dornumer Nacken, c WS (D) This studyBlue mussel (Mytilus-bed) (S) 0.14 0.05e0.21 S 1974e92 Northern Europe, estuaries/WS

(UK, NL, DK)Cal. after 7 ref. inZwarts et al., (1996c)

Blue mussel (Mytilus-bed) (H) 0.07 0.03e0.12 S 1956e84 Northern Europe, estuaries/WS(UK, NL)

Cal. after 3 ref. inZwarts et al. (1996c)

Blue mussel (Mytilus-bed) (S) 0.11 0.06e0.24 A 1973e92 Northern Europe, estuaries/WS(UK, NL, DK)

Cal. after 9 ref. inZwarts et al. (1996c)

Blue mussel (Mytilus-bed) (H) 0.14 0.07e0.26 A 1966e86 Northern Europe, estuaries/WS(UK, NL)

Cal. after 7 ref. inZwarts et al., (1996c)

Blue mussel (Mytilus-bed) 0.26 A 1993 Königshafen, n WS (D) Hertzler 1995Cockle, blue mussel (Mytilus-bed) 0.15 A 1999 Mellum island, c WS (D) Umland 2000Cockle (cockle bed) 0.15 0.04e0.23 S 1984 Northern Europe, estuaries/

WS (NL)Cal. after 1 ref. inZwarts et al., (1996c)

Cockle (cockle bed) 0.14 0.03e0.27 A 1983, 86 Northern Europe, estuaries/WS (NL)

Cal. after 5 ref. inZwarts et al., (1996c)

Cockle (mixed flat) 0.22 S 2007 Dornumer Nacken, c WS (D) This studyCockle, polychaetes (mixed flat) 0.04 A 2007 Dornumer Nacken, c WS (D) This studyCockle, polychaetes (mixed flat) 0.10 A 1999 Mellum island, c WS (D) Leyrer 2001Blue mussel spat, bivalves (mixed flat) (S) 0.14 S 1995 Neuharlingersieler Nacken,

c WS (D)Bradter 1996

Blue mussel spat, bivalves (mixed flat) (S) 0.10 A 1994 Neuharlingersieler Nacken,c WS (D)

Bradter 1996

Macoma (mixed flat) 0.12 S 1998 Minsener Oog, c WS (D) Wolff 2000Macoma (mixed flat) 0.09 A 1998 Minsener Oog, c WS (D) Wolff 2000Polychaetes, bivalves (mixed flat) 0.09 S 1999 Mellum island, c WS (D) Leyrer 2001

(S) ¼ stabber.(H) ¼ hammerer.S ¼ spring (March–May).A ¼ autumn (Aug–Oct).cal. ¼ calculated.ref. ¼ reference.w ¼ western.c ¼ central.n ¼ northern.WS ¼ Wadden Sea.D ¼ Germany.UK ¼ United Kingdom.NL ¼ The Netherlands.DK ¼ Denmark.

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e51 49

Author's personal copy

French coast (Lunais, 1975), which Baptist (2005) reported from anoyster reef in the Dutch Oosterschelde estuary.

Two opening techniques are described. Those using a stabbertechnique force their bill between gaping valves and pry them apartwhile those using a hammerer technique hammer a small holethrough the valve, insert their bill and sever or paralyze theadductor muscle. Oystercatchers are known to use both techniquesto open blue mussels (Goss-Custard et al., 1993) and single in-dividuals may be specialized in one of the techniques. Analogous tothe Black oystercatcher, who needed 8 years to learn to feed onoysters (Butler and Kirbyson, 1979), 9 years after the first record ofPacific oysters in the central Wadden Sea (Wehrmann et al., 2000),we observed oystercatchers preying intensively on Pacific oystersclustered in a reef. They followed the waterline and applied thestabbing technique on gaping individuals. Mainly around low tide,a combination of hammering and stabbing was used on closedindividuals. Peck ratewas unchanged between seasons but foragingbehaviour, success rate and biomass intake during autumn waslower at dense reef. Due to the failure of a pronounced spatfall inlate summer 2007 and the proliferation of the standing Pacificoyster stock between spring and autumn, individuals between 70and 100 mm were presumably more abundant at both reef typesthan smaller individuals (Fig. 5, prospective cohort shift by autumn2007). As the most consumed shell lengths during autumn werebetween 40 and 50 mm and the largest individual was 68 mm, weassume a limitation in opening Pacific oysters larger than 70 mmalthough further studies are needed.

The flexibility of oystercatchers in food choice (Zwarts andDrent, 1981; Boates and Goss-Custard, 1992; Zwarts et al., 1996b;Tuckwell and Nol, 1997) was reflected by high bird densities at reefcarpet during autumn where half of the diet served as alternateprey to Pacific oysters. The latter might be responsible for the stablesuccess rate and the same biomass intake at reef carpet during bothmigration periods. In contrast, at dense reef only a very low pro-portion of blue mussels contributed to the diet during spring. Bluemussels prefer to hide from predation by living on the bottom ofthe reef (Eschweiler and Christensen, 2011) and it remains open ifoystercatchers can pick up other prey than oysters within denselypacked reefs as their beak lengthmight be limited to reach betweenclusters of large oysters. At reef carpet, a habitat type with a morescattered oyster distribution, food accessibility of alternate prey isapparently easier.

Mean biomass intake of oystercatchers feeding on blue musselsor cockles varied between 0.07 and 0.15 g AFDM/min during springand between 0.11 and 0.14 during autumn (Table 4). Monthly singlerates can be lower (minimum rate 0.03) or higher (maximum rate0.27) when feeding on smaller or larger specimens, respectively.Similar mean values were found within studies on intake rates atdifferent tidal flat types calculated as well for other prey species orprey species mixtures. The biomass intake of Pacific oyster feedingoystercatchers during spring was at a low (0.07) to medium (0.11)level and within mean values assessed for blue mussels and cocklesin other Northern European habitats. In contrast, biomass intake atdense reef during autumnwas at the minimum of 0.05 g AFDM/min(Zwarts et al.,1996d)which requires the birds to feed over thewholelow tide period to maintain body condition. Oystercatcherscontinuously make feeding decisions which enhance intake rate,but the higher the biomass intake exceeds the digestive processingrate of 0.03 g AFDM/min, the longer birds are forced to interruptforaging for digestive pauses (Kersten and Visser,1996; Zwarts et al.,1996d). On the other hand, high biomass intake reduces the time thebird has to spend foraging. During autumn, when the energeticdemand of the bird increases or highwater levels shortens time andarea for feeding, dense Crassostrea-reefs may be less profitable,especially at times of failed or weak Pacific oyster spatfalls.

5. Conclusions

Decreasing blue mussel stocks in all regions of the Wadden Sea,which is mainly attributed to be shellfish fishery induced (Ens et al.,2009; Laursen et al., 2010) but also is related to a reduced repro-ductive success by the absence of cold winters (Beukema et al.,2010), may have led to decreasing abundances of especiallyshellfish eating birds during recent decades. The spread of the Pa-cific oyster overlaps with the decline of blue mussels (Nehls et al.,2009) but the transformation of Mytilus-beds into Crassostrea-reefs may not aggravate the situation for most of the birds in theWadden Sea. Crassostrea-reefs rather replenish the ecologicalfunction as feeding sites, respectively serving as a new prey foroystercatchers, enhancing the provision of associated macrofaunalprey organisms for curlew and increasing the attractiveness ofadjacent tidal flats for various bird species (Van der Zee et al., 2012).

Aswe observed distinct use and behaviour of the target specieseoystercatcher, curlew and herring gull e at the two reef types, wepostulate that reef structure impacts differentially on the suitabilityof the reefs as foraging sites for birds. Since the beginning of theinvasion in the central Wadden Sea, we observe an increasinglystable distribution in reef density (Markert, unpublished). Long-termstable heterogeneity in reef density within and between sites facil-itates habitat choice by oystercatchers such as visiting lesser engi-neered areas to feed on different prey than oysters. In some areas ofthe western and northern Wadden Sea the Pacific oyster plays aminor role as ecosystem engineer. Here,Mytilus- or cockle beds stilldominate and provide alternate feeding habitats (Nehls et al., 2009).Hence, bird behaviour may be adapted to different conditions indifferent regions and our findings may not be appropriate for allregions of the Wadden Sea. Thus, further studies are needed.

A repetitive failure in Pacific oyster spatfall over a couple ofyears may reduce the overall food supply for oystercatchers assmaller oyster size classes would not be available for consumptionand presumably alternate prey stay out of reach at consolidatedreefs. The provision of small oysters by a good year to year repro-duction with successful late summer spatfalls accompanied by lowwintermortalitiesmay be a new determinant in the futurewhich inturn depends on climate conditions and is positively correlatedwith climate change.

Acknowledgements

We thank theWadden SeaNational Park Administration of LowerSaxony (Wilhelmshaven) for appropriate authorisations, the Waterand Shipping Authority (Aurich) for the permission to rear theobservation tower. Several assistants are thanked for rearing andremoving the tower, especially Gerold Gembler for essential logis-tical support. We thank the University Münster for providing hous-ing facility at the “MeeresbiologischeWattstation Carolinensiel”. Weare indebted to Sarah Narkus for sorting macrofauna samples. Spe-cial thanks are given to Gregor Scheiffarth for constructive supportand comments on early versions of the manuscript. We thank Ger-hard Cadée and an anonymous referee for their constructive com-ments. The present study was carried out within the framework ofthe project “Management der Bioinvasion der Pazifischen Auster”which was financially supported by the SGN Senckenberg Gesell-schaft für Naturforschung and the Niedersächsische WattenmeerStiftung. The study is a contribution to the scientific concept of theBiodiversity and Climate Research Centre BiK-F.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.ecss.2013.08.003.

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e5150

Author's personal copy

References

Baptist, H., 2005. Habitattoets proef weghalen oesterbanken in de Oosterschelde.Report 2005/19. Ecologisch Adviesbureau Henk Baptist, Kruisland, TheNetherlands (unpublished).

Beukema, J.J., Dekker, R., Philippart, C.J.M., 2010. Long-term variability inbivalve recruitment, mortality, and growth and their contribution to fluctua-tions in food stocks of shellfish-eating birds. Marine Ecology Progress Series414, 117e130.

Boates, J.S., Goss-Custard, J.D., 1992. Foraging behaviour of oystercatchers Haema-topus ostralegus specializing on different species of prey. Canadian Journal ofZoology 7012, 2398e2404.

Bradter, U., 1996. Raum-Zeit-Muster und Nahrungskonsumption des Aus-ternfischers (Haematopus ostralegus) während der Zugzeit im Rückseitenwattvon Spiekeroog. Diploma thesis. Carl von Ossietzky Universität Oldenburg,Germany (unpublished).

Butler, R.W., Kirbyson, J.W., 1979. Oyster predation by the black oystercatcher inBritish Columbia. Condor 81, 433e435.

Campbell, R.W., 1966. On black oystercatchers “catching” oysters. Victoria Naturalist23, 26.

Colautti, R.I., Ricciardi, A., Grigorovich, I.A., Macisaac, H.J., 2004. Is invasion successexplained by the enemy release hypothesis? Ecology Letters 7, 721e733.

Dewar, J.M., 1922. Ability of the oystercatcher to open oysters and its bearing uponthe history of the species. British Birds 16, 118e125.

Drinkwaard, A.C., 1999. Introductions and developments of oysters in the North Seaarea: a review. Helgoländer Meeresuntersuchungen 52, 301e308.

Ens, B.J., Alting, D., 1996. The effect of an experimentally created mussel bed on birddensities and food intake of the oystercatcher Haematopus ostralegus. Ardea84A, 493e507.

Ens, B.J., Blew, J., Van Roomen, M.W.J., Van Turnhout, C.A.M., 2009. ExploringContrasting Trends of Migratory Waterbirds in the Wadden Sea. Wadden SeaEcosystem No. 27. Trilateral Monitoring and Assessment Group, Joint Moni-toring Group of Migratory Birds in the Wadden Sea, Common Wadden SeaSecretariat, Wilhelmshaven, p. 47.

Escapa, M., Isacch, J.P., Daleo, P., Alberti, J., Iribarne, O., Borges, M., Dos Santos, E.P.,Gagliardini, D.A., Lasta, M., 2004. The distribution and ecological effects of theintroduced Pacific oyster Crassostrea gigas (Thunberg, 1793) in Northern Pata-gonia. Journal of Shellfish Research 23, 765e772.

Eschweiler, N., Christensen, H.T., 2011. Trade-off between increased survival andreduced growth for blue mussels living on Pacific oyster reefs. Journal ofExperimental Marine Biology and Ecology 403, 90e95.

Fey, F., Dankers, N., Steenbergen, L., Goudswaard, K., 2010. Development and dis-tribution of the non-indigenous Pacific oyster (Crassostrea gigas) in the DutchWadden Sea. Aquaculture International 18, 45e59.

Frank, D., Exo, K.M., Wahls, S., Grünkorn, T., 1997. Ökologische Beglei-tuntersuchungen zu dem Projekt ’Europipe’ - Teilprojekt Avifauna. Absch-lußbericht Teil II. Final Report. Nationalparkverwaltung NiedersächsischesWattenmeer, Wilhelmshaven, Germany (unpublished).

Goss-Custard, J.D., Le V. Dit Durell, S.E.A., McGrorty, S., Reading, C.J., 1982. Use ofmussel Mytilus edulis beds by Oystercatchers Haematopus ostralegus accordingto age and population size. Journal of Animal Ecology 51, 543e554.

Goss-Custard, J.D., West, A.D., Le V. Dit Durell, S.E.A., 1993. The availability andquality of the mussel prey Mytilus edulis of Oystercatchers Haematopus ostra-legus. Netherlands Journal of Sea Research 31, 419e439.

Hilgerloh, G., Herlyn, M., Michaelis, H., 1997. The influence of predation by herringgulls Larus argentatus and oystercatchers Haematopus ostralegus on a newlyestablished mussel Mytilus edulis bed in autumn and winter. HelgoländerMeeresuntersuchungen 51, 173e189.

Jones, C.G., Lawton, J.H., Shachak, M., 1994. Organisms as ecosystem engineers.Oikos 69, 373e386.

Kersten, M., Visser, W., 1996. The rate of food processing in the oystercatcher: foodintake and energy expenditure constrained by a digestive bottleneck. Func-tional Ecology 10, 440e448.

Laursen, K., Blew, J., Eskildsen, K., Günther, K., Hälterlein, B., Kleefstra, R.,Lüerßen, G., Potel, P., Schrader, S., 2010. Migratory Waterbirds in the WaddenSea 1987e2008. Wadden Sea Ecosystem No. 30. Joint Monitoring Group ofMigratory Birds in the Wadden Sea, Common Wadden Sea Secretariat, Wil-helmshaven, p. 165.

Leyrer, J., 2001. Die nahrungsökologische Bedeutung von Arenicola-Sandwatten fürdurchziehende Wasser- und Watvögel. Diploma thesis. Carl von OssietzkyUniversität Oldenburg, Germany (unpublished).

Lunais, B., 1975. Charactéristique et signification du comportement prédateur del’Huîtrier-pie Haematopus ostralegus sur l’Huître de culture Crassostrea gigas.Memoires DEA Ecologie de Université de Tour. In: Goss-Custard, J.D. (Ed.),The Oystercatcher. From Individuals to Populations. Oxford University Press,Oxford.

Markert, A., Wehrmann, A., Kröncke, I., 2010. Recently established Crassostrea-reefsversus native Mytilus-beds: differences in ecosystem engineering affects themacrofaunal communities Wadden Sea of Lower Saxony, southern GermanBight. Biological Invasions 12, 15e32.

Martin, P., Bateson, P., 1986. Measuring Behaviour. An Introductory Guide. Cam-bridge University Press, New York, p. 200.

Meire, P.M., 1996. Distribution of Oystercatchers Haematopus ostralegus over a tidalflat in relation to their main prey species, Cockels Cerastoderma edule and

MusselsMytilus edulis: did it change after a substantial habitat loss? Ardea 84 A,525e538.

Meltofte, H., Blew, J., Frikke, J., Rösner, H.-U., Smit, C.J., 1994. Numbers and Distri-bution of Waterbirds in the Wadden Sea. Results and Evaluation of 36 Simul-taneous Counts in the DutcheGerman-Danish Wadden Sea 1980e1991. IWRBPubl. 34/Wader Study Group Bull. 49, Special Issue, p. 192.

Nehls, G., Thiel, M., 1993. Large-scale distribution patterns of the Mussel Mytilusedulis in the Wadden Sea of SchleswigeHolstein: do storms structure theecosystem? Netherlands Journal of Sea Research 31, 181e187.

Nehls, G., Hertzler, I., Scheiffarth, G., 1997. Stable mussel Mytilus edulis beds inthe Wadden Sea e they’re just for the birds. Helgoland Marine Research 51,361e372.

Nehls, G., Witte, S., Büttger, H., Dankers, N., Jansen, J., Millat, G., Herlyn, M.,Markert, A., Kristensen, P.S., Ruth, M., Buschbaum, C., Wehrmann, A., 2009. Bedsof blue mussels and Pacific oysters. Thematic Report No. 11. In: Marencic, H., DeVlas, J. (Eds.), Quality Status Report 2009. Wadden Sea Ecosystem No. 25.Trilateral Monitoring and Assessment Group, Common Wadden Sea Secretariat,Wilhelmshaven, p. 29.

Nol, E., Humphrey, R.C., 1994. American oystercatcher (Haematopus palliatus). In:Poole, A. (Ed.), The Birds of North America Online. Cornell Lab of Ornithology,Ithaca. http://dx.doi.org/10.2173/bna.82. http://bna.birds.cornell.edu/bna/species/082.

Petersen, B., Exo, K.-M., 1999. Predation of waders and gulls on Lanice conchilegatidal flats in the Wadden Sea. Marine Ecology Progress Series 178, 220e240.

Ruesink, J.L., Lenihan, H.S., Trimble, A.C., Heiman, K.W., Micheli, F., Byers, J.E.,Kay, M.C., 2005. Introduction of non-native oysters: ecosystem effects andrestoration implications. Annual review of Ecology, Evolution and Systematics36, 643e689.

Sachs, L., 1997. Angewandte Statistik. Springer-Verlag, Berlin, p. 881.Scheiffarth, G., Ens, B., Schmidt, A., 2007. What will happen to birds when Pacific

Oysters take over the mussel beds in the Wadden Sea? Wadden Sea Newsletter1, 10e14.

Schieck, P., 2010. Nahrungsökologische Bedeutung des Neozoen Pazifische Auster(Crassostrea gigas) für Wat- und Wasservögel im Nordfriesischen Wattenmeer.Diploma thesis. Universität Bremen, Germany (unpublished).

Schmidt, A., Wehrmann, A., Dittmann, S., 2008. Population dynamics of the invasivePacific oyster Crassostrea gigas during the early stages of an outbreak in theWadden Sea. Helgoland Marine Research 62, 367e376.

Smaal, A., Van Stralen, M., Craeymeersch, J., 2005. Does the introduction of thePacific oyster Crassostrea gigas lead to species shifts in the Wadden Sea? In:Dame, R.F., Olenin, S. (Eds.), The Comparative Roles of Suspension-feeders inEcosystems, Nato Science Series IV. Earth and Environmental Science, vol. 47,pp. 277e289.

Tomkins, I.R., 1947. The oyster-catcher of the Atlantic coast of North America and itsrelation to oysters. Wilson Bulletin 59, 204e208.

Tuckwell, J., Nol, E., 1997. Foraging behaviour of American oystercatchers inresponse to declining prey densities. Canadian Journal of Zoology 75, 170e181.

Umland, J., 2000. Die nahrungsökologische Bedeutung einer Wat-tlebensgemeinschaft für Wat- und Wasservögel im ostfriesischen Wattenmeer.Diploma thesis. Universität Hamburg, Germany (unpublished).

Van de Kam, J., Ens, B., Piersma, T., Zwarts, L., 2004. Shorebirds. An IllustratedBehavioural Ecology. KNNV, Utrecht, p. 368.

Van der Zee, E., van der Heide, T., Donadi, S., Eklöf, J., Eriksson, B.K., Olff, H., van derVeer, H., Piersma, T., 2012. Spatially extended habitat modification by intertidalreef-building bivalves has implications for consumer-resource interactions.Ecosystems 15, 664e673.

Wehrmann, A., Herlyn, M., Bungenstock, F., Hertweck, G., Millat, G., 2000. Thedistribution gap is closed e first record of naturally settled Pacific oystersCrassostrea gigas in the East Frisian Wadden Sea, North Sea. Senckenbergianamaritima 30, 153e160.

Wolff, S., 2000. Nahrungsökologische Bedeutung von Mischwatten im ost-friesischen Rückseitenwatt für Larolimikolen. Diploma thesis. Carl vonOssietzky Universität Oldenburg, Germany (unpublished).

Zwarts, L., Drent, R.H., 1981. Prey depletion and the regulation of predatordensity: oystercatchers Haematopus ostralegus feeding on mussels Mytilusedulis. In: Jones, N.V., Wolff, W.J. (Eds.), Feeding and Survival Strategiesof Estuarine Organisms, Marine Science, vol. 15. Plenum Press, NY,pp. 193e216.

Zwarts, L., Wanink, J., 1993. How the food supply harvestable by waders in theWadden Sea depends on the variation in energy, density, body weight, biomass,burying depth and behaviour of tidal flat invertebrates. Netherlands Journal ofSea Research 31, 441e476.

Zwarts, L., Cayford, J.T., Hulscher, J.B., Kersten, M., Meire, P.M., Triplet, P., 1996a.Prey size selection and intake rate. In: Goss-Custard, J.D. (Ed.), The Oyster-catcher: from Individuals to Populations, Oxford Ornithological Series 7,pp. 30e55.

Zwarts, L., Wanink, J.H., Ens, B.J., 1996b. Predicting seasonal and annual fluctuationsin the local exploitation of different prey by oystercatchers Haematopus ostra-legus: a ten-year study in the Wadden Sea. Ardea 84A, 401e440.

Zwarts, L., Ens, B.J., Goss-Custard, J.D., Hulscher, J.B., Le, V., dit Durell, S.E.A., 1996c.Causes of variation in prey profitability and its consequences for the intake rateof the Oystercatcher Haematopus ostralegus. Ardea 84A, 229e268.

Zwarts, L., Ens, B.J., Goss-Custard, J.D., Hulscher, J.B., Kersten, M., 1996d. Why Oys-tercatchers Haematopus ostralegus cannot meet their daily energy requirementsin a single low water period. Ardea 84A, 269e290.

A. Markert et al. / Estuarine, Coastal and Shelf Science 131 (2013) 41e51 51