Hydrobiologia 515: 137–146, 2004. 137 © 2004 Kluwer Academic Publishers. Printed in the .

Changes in marsh nekton communities along the salinity gradient of the Schelde river, and The Netherlands

Henrietta Hampel1, Andre Cattrijsse2 & Jan Mees2 1Ghent University, Marine Biology Section, Krijgslaan 285, S-8, 9000 Gent, Belgium Tel: +00-32-9-264-85-24. E-mail: [email protected]; [email protected]. 2Flanders Marine Institute, Vismijn Pakhuizen 45-52, 8400 Oostende, Belgium

Received 4 February 2003; in revised form 18 August 2003; accepted 1 September 2003

Key words: marsh habitat types, salinity gradient, nekton, Schelde

Abstract Nekton was sampled in five marshes along the salinity gradient of the Schelde River. The utilisation of three different habitats (large and small creek, marsh pond) by fish and macrocrustacean species was compared among the five sampling sites. In the larger channels fyke nets were deployed to capture fish and macrocrustaceans leaving the marsh at ebb while block nets were set in smaller intertidal creeks. Fish traps passively sampled fish and shrimp in the marsh ponds. The tidal freshwater marsh had a species poor fauna and only a low number of fish was caught. Besides some freshwater species ( bipunctatus, Carassius carassius) the European eel, Anguilla an- guilla was still present. The four other marshes had a similar community structure although Platichthys flesus was absent from the euhaline area. Among fish species, dominance of Dicentrarchus labrax, Platichthys flesus and Pomatoschistus microps was observed. Carcinus maenas and Palaemonetes varians were the most abundant macrocrustacean species in every marsh. Between the large and small intertidal creeks there was no difference in nekton species composition. The main species used both habitats. Marsh ponds were utilized intensively only by two species, Pomatoschistus microps and Palaemonetes varians in every marsh.

Introduction ences their physicochemical and biological properties. These properties differ between stream order (Rozas Shallow marshes are among the most productive zones & Odum, 1987; Desmond et al., 2000). In Europe, the of estuaries for supporting high standing crop of fish importance of the position of a marsh within an estuary and shellfish (Weinstein et al., 1980). Fish belong to was reported by Cattrijsse et al. (1994). ‘nekton’ which refers to that live in water Studies dealing with the nekton fauna of European and are capable of sustained self-propulsion through tidal marshes have so far not addressed the topic of that medium in a horizontal direction. Estuarine nek- spatial differences in the aquatic marsh fauna within ton include a variety of fishes, crustaceans and even one region or estuary (Drake & Arias, 1991; Laffaille cetaceans (Kneib, 1997). et al., 2000; Lefeuvre et al., 2000; Costa et al., 2001). In the North American marshes, the occurrence of Mathieson et al. (2000) compared the fauna of several nekton species in marsh creeks is influenced by the European tidal marshes but did not include this topic. hydrology and channel morphology (McIvor & Rozas, Even global level studies describing faunistic changes 1996), the presence or absence of vegetation (Rozas & in marsh nekton along salinity changes are few (Rako- Zimmerman, 2000) and the type of the marsh whether cinski et al., 1992). Cattrijsse et al. (1994) supplied a natural vs. constructed (Zedler et al., 1997; Willi- detailed description of the aquatic fauna of the brack- ams & Zedler, 1999). The stream order of the creek ish marsh of Saeftinghe (Belgium) but information is also important since tidal creeks of the same stream lacking on the species occurrence in marshes situated order have similar physical attributes which influ- in other salinity ranges within the Schelde estuary. 138

period when tidal amplitude ranged between 4.5– 5.5 m NAP (Dutch ground level) in the estuary. Zwin (Zw) is situated at the mouth of the Westerschelde estuary, in the euhaline zone, with the total surface area of 125 ha. The Zuidgors (Z) marsh belongs to the polyhaline zone of the estuary and Waarde (W) to the mesohaline part. The total surface of the two marshes are 50 ha and 107 ha, respectively. ‘Het Verdronken Land van Saeftinghe’ (S) also lies in the mesohaline part of the Westerschelde estuary but with lower salin- ity as measured in Waarde. Saeftinghe has the largest surface area (2078 ha). In the tidal freshwater part of the Schelde reed and willow marshes are still present. Figure 1. Location of the five marshes sampled along the Wester- One of the largest tidal freshwater marshes (8.77 ha) is schelde. Grembergen (G), Saeftinghe (S), Waarde (W), Zuidgors (Z), Zwin (Zw). the ‘Groot Schoor’ or the marsh of Grembergen (G).

The first objective of this study was to examine Material and methods how the fish and macrocrustacean fauna of tidal marsh creeks within one estuary differ between salinity re- Intertidal channels gions. In the Schelde, freshwater tidal marshes are still present and marshes occur in all salinity ranges. Large intertidal channels were sampled with fyke nets In the present study, large creeks, smaller high to capture larger fish species that were only occasion- marsh creeks and marsh pools were sampled to further ally caught by Cattrijsse et al. (1994). The dimensions evaluate the habitat value of salt marshes for nekton of the larger creeks varied between 10–20 m wide and species. Cattrijsse et al. (1994) used a small mesh size 2–3 m deep. These fyke nets had an opening of 1 m (1 mm) to estimate the nursery and feeding ground and a mesh size of 15 × 15 mm. The nets were 5 m value of large intertidal marsh creeks. Large individu- long and contained 6 parlours. At the mouth 3 m long als and some species would have been inefficiently wings enlarged the sampling area of each net. Two caught with their method. By using other techniques fyke nets were set close to the mouth of each creek and we tried to broaden our knowledge on marsh usage by two deeper into the marsh. The fyke nets were set be- fish and macrocrustaceans. In North America, marsh fore water entered the system and faced the outgoing habitats different than large creek are equally used by ebb currents to sample fish and macrocrustaceans leav- nekton species (Rozas & Reed, 1993; Kneib & Wag- ing the creek with ebb. Collections were made after the ner, 1994) but information on the use of other marsh water receded. habitat types is lacking in Europe (Frid, 1988; Frid & James, 1989). Therefore it is of value to determine Intertidal creeks differences of habitat utilization by fish, shrimps or crabs, within the marshes. In each marsh one small creek was chosen and sampled with one block net, which was set at the mo- ment of the high water near the mouth of the creek. Study area The small creek directly opened from the middle of the large channel. These smaller creeks were generally 2– In this study five marshes were chosen to investigate 3 m wide and 1–1.8 m deep. This net was constructed possible changes in the fish and macrocrustaceans as- of 5 mm gauze and had two lateral wings of 2 m and semblages along the salinity gradient of the Schelde a central bag. The whole net was 2 m height and the river (Fig. 1). topline was provided with float. The ground rope was The river is divided by the saline part of the West- fitted with a heavy chain. The chain in the ground rope erschelde and the freshwater part of Zeeschelde. All was manually placed on the bottom of the creek and marshes sampled are subjected semi-diurnal tidal re- sticks kept the net and ground rope in place. After all gime. Marshes were sampled during the spring tide water had left the creek the net was emptied. 139

Tidal pools Fish trap collections were expressed as number of individuals per hour per fish trap. Tidal pools inside the marsh were also selected as po- Group Average Sorting cluster analysis based on tential nekton habitat types. In Waarde and Zwin one the Bray–Curtis similarities (Bray & Curtis, 1957) was larger and one small pond were chosen situated 3m used to describe the spatial structure of nekton based from each other while in Zuidgors one large pond was on density per station and per month. selected. Fish bottle traps were constructed to pass- ively catch small fish or macrocrustaceans inhabiting these shallow environments. The number of traps was Results and discussion adjusted to the size of the pools. Six fish traps were placed in the large and 4 in the small pond in Waarde Environmental parameters and Zwin while 10 traps were situated in the large pool of Zuidgors at each sampling day. The traps were Table 1 shows the change of environmental paramet- placed at the edge of the pools before the tide started ers in the five marshes during the sampling period. and stayed there for 5 h. Fish traps were made of trans- Samples were collected along the salinity gradient parent plastic bottles. The neck of the bottle was cut between 0.2–30 psu. Salinity ranged from 30 psu in off and inserted backwards into the bottle. The bottom Zwin, 25 psu in Zuidgors, over 15 psu in Waarde was also opened to insert another neck. This created a and 10 psu in Saeftinghe to fresh water in Grember- trap with two openings of 25 mm, a diameter of 75 mm gen. Temperature showed similar temporal variation ◦ and a total length of 250 mm. In Grembergen such in all marshes. Generally 12–15 C were measured at pools were not present while in Saeftinghe we found spring and autumn while in the summer months the ◦ no pool that could easily be accessed. temperature reached the 18–22 C. Dissolved oxygen concentration showed the lowest value in Saeftinghe General material and methods and Grembergen around 4–6.8 mg l−1, except in Oc- tober when 7.7–7.2 mg l−1 oxygen were measured Samples were collected every six weeks between April respectively. Waarde and Zuidgors had similar dis- and October 2000. The five marshes were sampled on solved oxygen concentrations ranging between 5.6– consecutive days during spring tide periods. Waarde 8.1mgl−1. Approximately 9 mg l−1 oxygen con- could not be sampled during April while in the ‘Groot centrations were measured in Zwin except in April Schoor’ the block net could only be set during that when the concentration stayed only at 4.2. The de- month. Between May and October too dense ve- crease in oxygen concentration in the upstream part getation prevented the placement of the net in this of the estuary is a well-known phenomenon. Large marsh. Traps were put in the pools from July onwards. areas of the tidal freshwater parts of the Schelde river Previous observations had shown that early juvenile are anoxic throughout most of the year (Maes et al., Pomatoschistus microps inhabits the creeks from June 1997). In the marsh ponds, salinity showed gener- onwards (Cattrijsse et al., 1994). In the ‘Groot Schoor’ ally higher values than in the channels. Temperature freshwater marsh pools were absent. changes reflected the seasonal trend with the highest At each location dissolved oxygen, salinity and values in July (22.4–25.5 ◦C) and lowest in October temperature were measured at the moment of high wa- (12–14.3 ◦C). In every marsh, the dissolved oxygen ter. Similar environmental parameters were measured was lower in the marsh ponds than in the creeks. in the small and large creeks therefore values from the large creeks are presented in this study. Temporal variation of total number of nekton during Individuals per fyke net per hour were calculated the sampling period for the large creek in every month in each marsh. These values were summed up to obtain the total Total number of individuals caught for each sampling number of nekton in each month in this habitat type. technique over the complete sampling campaign is Similar calculation was made for the other two habitat presented in Fig. 2. Fyke nets (Fig. 2Ia–b) and the types such as small creek and marsh ponds. block net (Fig. 2IIa–b) catches increased over the 5 To standardise the catches in the different marshes months reaching maxima (27.7 ind. hr−1 net−1 and the total number of individuals caught per hour per 191 ind. ebb period−1) in August and October. The fyke net was calculated. Block net samples were ex- maximum catch of fish traps happened in August pressed as number of individuals leaving the creek. (18.5 ind. hr−1 trap−1) (Fig. 2IIIa–b). Cattrijsse et al. 140

Table 1. Predicted water height (cm) in the creek and the measured water height in the marsh pond, salinity (psu), − ◦ dissolved oxygen (mg l 1) and temperature ( C) measured in the water column in the large creek and the marsh pond.

Location April May July August October Creek Pond Creek Pond Creek Pond Creek Pond Creek Pond Waterheight

Grembergen 524 506 532 524 541 Saeftinghe 424 517 513 515 523 Waarde 517 521 22 552 30 513 25 Zuidgors 575 555 518 30 562 42 477 28 Zwin 388 423 423 25 451 40 402 30

Salinity

Grembergen 0.3 0.4 0.2 0.2 0.3 Saeftinghe 11.4 8.4 8.0 10.8 9.0 Waarde 16.5 14.8 22.5 16.7 17.3 15.9 16.8 Zuidgors 22.6 22.8 25.3 25.8 28.4 25.3 25.0 24.0 Zwin 25.5 27.5 30.8 31.2 28.5 30.1 25.5 25.7

Temperature

Grembergen 13.3 17.0 17.2 20.3 16.0 Saeftinghe 13.3 17.5 18.8 21.9 15.2 Waarde 14.0 16.7 25.1 21.2 19.1 12.6 13.0 Zuidgors 12.2 15.1 20.5 22.4 16.9 18.4 13.6 14.3 Zwin 12.7 21.4 21.9 25.5 15.8 16.2 11.6 12.0

Oxygen

Grembergen 4.1 5.2 5.0 4.8 7.2 Saeftinghe 4.6 6.8 5.6 4.6 7.7 Waarde 8.0 7.3 5.2 6.2 5.8 6.6 4.2 Zuidgors 5.6 8.1 7.5 6.3 7.3 5.9 6.8 4.2 Zwin 4.2 9.2 8.9 5.6 9.0 7.4 9.2 9.1

(1994) have shown that during this period of the year a flounder appear in the Ythan estuary in spring (April, maximum number of species use the marsh creeks as May) and all year classes were present until winter. nursery and feeding ground. Investigations mainly focused on the larval or postlarval stages of D. labrax. Drake & Arias (1991) Intertidal channels found high densities of postlarval seabass from March In the large creeks, numbers of shore crab Car- till July in a salt marsh in Spain. Aprahamian & Barr cinus maenas and seabass Dicentrarchus labrax in- (1985) observed postlarvae in the Severn estuary, UK creased towards autumn reaching a total 9.9 and during August and September. In the Westerschelde 16.5 ind. hr−1 net−1, respectively (Fig. 2Ia-b,). The estuary, high density of postlarvae of seabass migrated abundance of flounder, Platichthys flesus decreased in the salt marsh creek from August till November during the sampling campaign. Highest occurrence (Cattrijsse et al., 1994) and the juvenile of this spe- was detected in July (5.7 ind. hr−1 net−1). Hampel cies reached the highest abundances from December (pers.obs.) found high number of shore crab in autumn till March in the estuary subtidal (Hostens, 2003). in a mesohaline marsh creek, Sieperda, The Nether- lands. Raffaelli et al. (1990) reported large number of 141

− − Table 2a. Number of nekton (ind. hr 1 net 1) captured by fyke net in the five different marshes during the sampling campaign. Marshes are indicated as following: Grembergen (G), Saeftinghe (S), Waarde (W), Zuidgors (Z), Zwin (Zw)

Month April May July August October Location GS ZZwGS WZZwGS WZZwGS WZZwGS WZZw Alburnoides bipunctatus 1.7 Anguilla anguilla 0.3 0.1 0.1 0.3 1.7 1.7 0.2 0.3 0.1 0.3 0.1 Carassius carassius 0.1 0.3 Clupeidae spp. 0.4 0.8 0.3 Dicentrarchus labrax 0.2 0.7 2.3 1.0 0.7 0.6 5.0 1.9 6.8 7.8 8.7 5.2 0.2 0.3 Gasterosteus aculeatus 0.1 Liza spp. 0.5 Platichthys flesus 2.9 0.8 1.1 2.2 1.0 1.1 3.3 1.3 0.2 0.8 1.3 1.0 0.1 Pleuronectes platessa 0.1 Pomatoschistus microps 0.1 Syngnathus rostellatus 1.7

Carcinus maenas 0.2 0.6 0.3 0.3 6.0 0.1 0.4 0.3 0.7 4.4 1.7 2.7 1.9 3.3 7.9 0.2 1.3 0.6 Eriocheir sinensis 0.1 0.3 Palaemonetes varians 1.7

Figure 2. Total number of individuals caught by fyke net (I), block net (II) and fish trap (III) in the five marshes during the five sampled months. Total number of fish (a) and crustaceans (b) are separated in the figure. 142

Table 2b. Number of individuals caught by blocknet during one ebb period. Marshes are indicated as following: Grembergen (G), Saeftinghe (S), Waarde (W), Zuidgors (Z), Zwin (Zw)

Month April May July August October Location GSZZw SWZZwSWZZw SWZZwSWZZw Alburnoides bipunctatus 2 Atherina presbiter 1 Clupeidae spp. 3 1 Dicentrarchus labrax 4 11 13 236 10 Platichthys flesus 11

Pomatoschistus microps 3112 4 9 52269 Carcinus maenas 3 12 24 3446 25 8 Crangon crangon 7 Palaemonetes varians 16 2 24444810 1 121121

− − Table 2c. Number of individuals (ind. hr 1 trap 1) in the marsh ponds captured by fish traps. Marshes are indicated as following: Waarde (W), Zuidgors (Z), Zwin (Zw)

Month July August October Location WZ Zw WZ Zw WZZw Gasterosteus aculeatus 0.03 Pomatoschistus microps 0.1 0.06 0.05 1.0 1.73

Carcinus maenas 0.03 Crangon crangon 0.03 Palaemonetes varians 2.03 0.48 1.15 17.3 0.15 4.13 1.0 0.2

Intertidal creeks highly abundant in this marsh habitat type (Fig. 2IIIa– The block net (Fig. 2IIa–b,) catches showed that b). Frid (1988) also observed high number of Po- fish and shrimp visit the small intertidal creeks in matoschistus and Palaemon species in marsh ponds large numbers only during late summer and early during summer and autumn in Eastern England (geo- autumn. In spring and early summer the block net graphically comparable to the Westerschelde). Juven- catches were low. Pomatoschistus microps appeared ile common gobies make maximum use of the high mainly in August and October (25 and 96 ind. ebb marsh pools as nurseries between August and October. period−1, respectively), which supports the results of These maximum catches in the marsh pools coincided Cattrijsse et al. (1994). In the subtidal of the Wester- with maximum catches of these species in the block schelde estuary, the adult common goby reached the net. highest density during the winter period from Novem- ber till March (Hostens, 2003). The seasonal change Spatial variation of abundance of nekton species in in abundance of Carcinus maenas was also clear from different marsh habitat types these samples. Increasing number of shore crab util- ize the small creeks towards autumn reaching a total Intertidal channels of 33 individuals. Dicentrarchus labrax was present Densities of nekton in the large creek are given in in the small creeks with low numbers (0–15 ind. Table 2a. Nekton communities in the large channels ebb period−1) during the sampling period. Platichthys of the fresh water and euhaline marsh separated from flesus appeared in this habitat type only in May. the other marshes by using Bray–Curtis similarities (Fig. 3a). The nekton assemblages in the other three Tidal pond marshes with intermediate salinity level did not dif- Fish traps collected maximum numbers (18.5 ind. hr−1 fer from each other. This result differed that Hostens trap−1) in August when Palaemonetes varians was (2003) found in the Westerschelde estuary. In the 143

ror the poor fish fauna of the tidal freshwaters of the Schelde river. The catches in the other marshes were dominated by gobies Pomatoschistus spp., flounder Platichthys flesus, seabass Dicentrarchus labrax and shore crab Carcinus maenas. Only the mentioned species seem to be able to use the marsh habitat as nursery or as foraging ground. These fish species together with mullets dominate the European fish fauna of marshes (Frid & James, 1989; Drake & Arias, 1991; Cattrijsse et al., 1994; Costa & Cabral, 1999; Laffaille et al., 1998, 2000). Mathieson et al. (2000) reported P. fl e su s and Anguilla anguilla to be the two most widely re- corded species in European estuarine and lower river marsh systems. Mullets were not caught during this study but the juveniles do enter marsh creeks. Cat- trijsse et al. (1994) observed their occurrence but since the juveniles swim close to the surface they could hardly be caught with the method used. In Saeft- inghe and Waarde, Platichthys flesus migrated into the creeks until August. Maximum abundance occurred in April in Saeftinghe (2.9 ind. hr−1 net−1)andin July in Waarde (3.3 ind. hr−1 net−1). In August and October high numbers of Dicentrarchus labrax and Figure 3. Dendogram for nekton densities in the intertidal channel Carcinus maenas were caught. Also in the Zuidgors (a), intertidal creek (b) and tidal pool (c). Marshes are labelled as marsh, P. fl e su s mainly appeared during the first three indicated in Figure 1. Sampled months were April (4), May (5), July (7) August (8) and October (10) indicated next to the marsh months but here D. labrax was more numerous in labels. every month except for October when only C. maenas was captured. In the euhaline marsh, Zwin, the only species estuary, the mesohaline and the polyhaline areas were caught in high numbers was C. maenas. This species is characterized by two different epibentic communities. known to avoid low salinities during most of the year Fyke net catches (Fig. 4Ia–b,) in the freshwater (De Veen et al., 1979). The low turbidity of the water marsh were always low and were composed of com- most likely maximised net avoidance by fish. The ab- pletely different species than in the other marshes. sence of flounder in this marsh might derive from the The eel, Anguilla anguilla was the dominant species low catching efficiency of the net. Although along the in each sample irrespective of the sampling period. salinity gradient of the Westerschelde estuary, Hos- Crucian carp, Carassius carassius and schneider, Al- tens (2003) reported P. fl e su s as common species and burnoides bipunctatus were the other species caught. observed that the density of flatfish stayed relatively The low catches in the Groot Schoor can partly be at- low in the marine part and increased in the brackish tributed to the very high amount of organic material area. The salinity preference of this species might ex- trapped in the fyke net but the fish community of the plain better the lack of flounder in the euhaline marsh. tidal freshwater part of the estuary is very poor in dens- Young Dicentrarchus labrax were regularly observed ity and diversity (Maes et al., 1997). Based on fyke swimming in the larger creek but caught only in low samples, these authors recorded 7 fish species and one numbers. In the small channels seabass was captured shrimp species on the mudflats of the tidal freshwater with the block net in July and August. Besides the part of the estuary. Their catches in the most upstream possibility that seabass avoids the net, D. labrax was sampling location consisted of only two species: A. reported mainly from the brackish part of the estu- anguilla and Rutilus rutilus, the later one was recor- ary (Hostens, 2003). In contrast to the marshes, the ded only once. Grembergen was even further upstream seabass was classified as species occurring only oc- than this location. Our low catches seem thus to mir- 144

Figure 4. Number of individuals caught in the large creek by fyke net (I), in the small creek by block net (II) and in the marsh pond by fish trap (III). Letters are used to label the marshes (Grembergen – G, Saeftinghe – S, Zuidgors – Z, Waarde – W, Zwin – Zw) and the numbers represent the month of the sampling (April 4, May 5, July 7, August 8, October 10). Number of fish (a) and crustaceans (b) are separated in the figure. casionally in the subtidal of the estuary (Hostens, to visit this habitat type. Rozas & Odum (1987) ob- 2003). served in a North American freshwater marsh that fish and macroinvertebrates concentrate in the upper Intertidal creeks reaches of tidal creeks where the creeks become more Number of nekton captured in the small creek is given shallow and narrow. In addition, our data also showed in Table 2b. Using Bray–Curtis similarities the fresh- that not only juveniles venture into this habitat type. water catch in April was separated from the other Adult flounder was regularly caught in these small marshes. There was no spatial difference detected in channels to feed upon the infauna. Desmond et al. species composition between the other four marshes (2000) argued that in a salt marsh, situated at the US– (Fig. 3b). Only the common goby Pomatoschistus Mexico border, the low oxygen environment and the microps, the shore crab Carcinus maenas, the brack- higher temperatures in the smallest intertidal creeks ish water shrimp Palaemonetes varians and young restricted their use by fish. Excessive high temperat- Dicentrarchus labrax were the dominant species in the ures were however never recorded during our study in small creeks (Figure 4IIa–b). The environmental para- these small creeks. meters were similar in the small creeks and the differ- ence of salinity seemed not to restrict these few species 145

Tidal ponds In North America, researchers argued about the Several authors from North America reported high primary factor of large-scale environmental gradients habitat use of marsh ponds by nekton, which offer (salinity or hydrology and current patterns) influen- refugia from predators but may compromise growth cing community structure of estuarine fishes (Wein- (Werner et al., 1983; Power, 1984; Holbrook & stein et al., 1980; Ross & Epperly, 1985) On the east Schmitt, 1988). In Europe, Frid (1988) and Frid & coast of the US, distributions of estuarine fishes and James (1989) reported utilization of tidal pools by macroinvertebrates have been shown to directly follow fish and macrocrustaceans like juvenile mullet, small primary salinity gradients in a salt marsh (Weinstein plaice and flounder, gobiid fish and Palaemon spe- et al., 1980). In another salt marsh along the east cies. In the marshes of the east coast of US, Fundulus coast of the US, Rakocinski et al. (1992) also found heteroclitus is the typical species use intensively the that salinity is the primary abiotic factor correlate marsh ponds during the whole year (Smith & Able, with fish community structure. Similar researches are 1994; Halpin, 2000; Layman et al., 2000). This spe- lacking in European intertidal areas. In this study cies has a similar ecological role in North American the nekton fauna of the four saline marshes did not marshes as Pomatoschistus microps in the European differ. Similar species utilized the eu-, poly- and meso- tidal areas. F. heteroclitus found both growth advant- haline intertidal areas. Among fish species, dominance ages and predator refuge in marsh ponds. of Dicentrarchus labrax, Platichthys flesus and Po- Densities of nekton in the ponds are given in matoschistus microps was observed. Carcinus maenas Table 2c. During the sampling campaign the fish traps and Palaemonetes varians were the most abundant captured Pomatoschistus microps and Palaemonetes macrocrustacean species in each marsh. Only the varians and only in Waarde large numbers of fish freshwater marsh had a distinct fish fauna with some and shrimp were caught (Fig. 4IIIa–b). The nekton typically freshwater species. Additionally, no differ- communities in the three sampled marshes did not ence in species composition was observed between differ from each other using Bray–Curtis similarities the nekton of large or small intertidal creeks. The (Fig. 3c). P. varians was caught in the highest num- main species used tidal marsh creeks as habitats en- bers in the three marshes indicating that this species tering both the larger and the smaller channels. Marsh is more tolerable for more the extreme abiotic condi- ponds were only utilized by two species, Pomatoschis- tions. In Waarde the pools were close to small creeks tus microps and Palaemonetes varians. These pans so that tides might have regularly well refreshed the are shallow and not regularly flooded which might water. In Zwin and Zuidgors, the ponds were situ- represent too stressful environment for other species. ated farther from the creek and the water entered very The description of the community structure and the slowly to the ponds with the tide and this less intense abundance of nekton species give an overview about the water supply might explain the lower presence the utilization of the marsh habitat types but future of gobies and shrimp. The physical environment of research is suggested to focus on the feeding habits marsh ponds can be very variable (Bulger, 1984; Co- of the nekton species on large- and small spatial scale. chran & Burnett, 1996) and only a few species have Investigations about the feeding habits of marsh visit- the tolerance to withstand these conditions. The spe- ing species would contribute further to assess the value cies, which can adapt to the environmental conditions of marsh habitat types as feeding ground. of the marsh ponds, can find a good refuge and for- aging area (Talbot & Able, 1984; Smith & Able, 1994; References Kneib, 1997). Aprahamian, M. W. & D. C. Barr, 1985. The growth, abundance and diet of O-group sea bass, Dicentrarchus labrax, from the Severn Conclusions Estuary. Journal of Marine Biological Association of the United Kingdom 65: 169–180. As hydrography is the most important factor in the de- Bray, J. R. & J. T. Curtis, 1957. An ordination of the upland forest velopment and functioning of wetlands (Reed, 1993), communities of southern Wisconsin. Ecological Monographs 27: 325–349. salinity is considered the major influence on the dis- Bulger, A. J., 1984. A daily rhythm in heat tolerance in the tribution and movements of estuarine nekton (Gunter, salt marsh fish Fundulus heteroclitus. Journal of Experimental 1961, Marshall & Elliott, 1998). Fish and invertebrate Zoology 230: 11–16. assemblages of estuarine salt marshes fluctuate both Cattrijsse, A., E. S. Makwaia, H. R. Dankwa, O. Hamerlynck & M. A. Hemminga, 1994. Nekton communities of an intertidal temporary and spatially (Rakocinski et al., 1992). 146

creek of a European estuarine brackish marsh. Marine Ecological Maes, J., A. Taillieu, P. Van Damme & F. Ollevier, 1997. The com- Progress Series 109: 195–220. position of the fish and crustacean community of the Zeeschelde Cochran, R. E. & L. E. Burnett, 1996. Respiratory responses of Estuary (Belgium). Belgian Journal of Zoology 127 (1): 47–55. the salt marsh animals, Fundulus heteroclitus, Leiostomus xan- Marshall, S. & M. Elliott, 1998. Environmental influences on fish thurus,andPalaemonetes pugio to environmental hypoxia and assemblages of the Humber estuary, UK. Estuarine, Coastal and hypercapnia and to the organophosphate pesticide, azinphos- Shelf Science 46: 175–184. methil. Journal of Experimental Marine Biology and Ecology Mathieson, S., A. Cattrijsse, M. J. Costa, P. Drake, M. Elli- 195: 125–144. ott, J. Gardner & J. Marchand, 2000. Fish assemblages of Costa, M. J. & H. N. Cabral, 1999. Changes in the Tagus nurs- European tidal marshes: a comparison based on species, families ery function for commercial fish species: some perspectives for and functional guilds. Marine Ecological Progress Series 204: management. Aquatic Ecology 33: 287–292. 225–242 Costa, M., F. Catarino & A. Bettencourt, 2001. The role of salt McIvor, C. C. & L. P. Rozas, 1996. Direct nekton use of intertidal marshes in the Mira estuary (Portugal). Wetlands Ecology and salt marsh habitat and linkages with adjacent habitats: a review Management 9: 121–134. from the southeastern United States. In Nordstorm, K. F. & C. Desmond, J. S., J. B. Zedler & G. D. Williams, 2000. Fish use T. Roman (eds), Estuarine Shores: Evolution, Environments, and of tidal creek habitats in two southern California salt marshes. Human Alteration. John Wiley and Sons, West Sussex: 311–334. Ecological Engineering 14: 233–252. Power, M. E., 1984. Depth distribution of armored catfish: De Veen, J. F., R. Boddeke & K. H. Postuma, 1979. Tien jaar predatior-induced resource avoidance? Ecology 65: 523–528. kinder-kamer-opnames in Nederland: I. Het Zeeuws estuarium Raffaelli, D., H. Richner, R. Summer & S. Northcott, 1990. Tidal (in Dutch). Visserij 32: 3–32. migrations in the flounder. Marine Behavioral Physiology 16: Drake, P. & A. M. Arias, 1991. Composition and seasonal fluc- 249–260. tuations of the ichthyoplankton community in a shallow tidal Rakocinski, C. F., D. M. Baltz & J. W. Fleeger, 1992. Corres- channel of Cadiz Bay (S.W. Spain). Journal of Fish Biology 39: pondence between environmental gradients and the community 245–263. structure of marsh-edge fishes in a Louisiana estuary. Marine Frid, C. L. J., 1988. The marine fauna of the north Norfolk salt Ecological Progress Series 80: 135–148 marshes and their ecology. Transactions of the Norfolk and Reed, D. J., 1993. Hydrology of temperate wetlands. Progress in Norwich Naturalists Society 28: 46–50. Physical Geography 17: 20–31. Frid, C. L. J. & James R., 1989. The marine invertebrate fauna of a Ross, S. W. & S. P. Epperly, 1985. Utilization of shallow estuarine British coastal salt marsh. Holarctic Ecology 12: 9–15. nursery areas by fishes in Pamlico Sound and adjacent tributaries Gunter, G., 1961. Some relations of estuarine organisms to salinity. North Carolina. In Yanez-Arancibia, A. (eds), Fish Community Limnology and Oceanography 6: 182–190. Ecology in Estuaries and Coastal Lagoons: Towards an Eco- Halpin, P. M., 2000. Habitat use by an intertidal salt-marsh fish: system Integration. Universidad National Autonoma de Mexico, trade-offs between predation and growth. Marine Ecological Mexico City: 207–232. Progress Series 198: 203–214. Rozas, L. P. & W. E. Odum, 1987. Use of tidal freshwater marshes Holbrook, S. J. & R. J. Schmitt, 1988. The combined effects of by fish and macrofaunal crustaceans along a marsh creek-order predation risk and food reward on patch selection. Ecology 69: gradient. Estuaries 10: 36-43. 123–134. Rozas, L. P. & D. J. Reed, 1993. Nekton use of marsh-surface Hostens, K., 2003. The demersal fish and macro-invertebrate as- habitats in Louisiana (U.S.A.) deltaic salt marshes undergoing semblages of the Westerschelde and Oosterschelde estuaries submergence. Marine Ecological Progress Series 96: 147–157. (Southern Bight of the North Sea). University of Ghent, Gent, Rozas, L. P. & R. J. Zimmerman, 2000. Small-scale patterns of nek- 164 pp. ton use among marsh and adjacent shallow nonvegetated areas of Kneib, R. T., 1997. The role of tidal marshes in the ecology the Galveston Bay Esturay, Texas (U.S.A.). Marine Ecological of estuarine nekton. Oceanography and Marine Biology 35: Progress Series 193: 217–239. 163–220. Smith, K. J. & K. W. Able, 1994. Salt-marsh tide pools as winter Kneib, R. T. & S. L. Wagner, 1994. Nekton use of vegetated marsh refuges for the mummichog, Fundulus heteroclitus,inNew habitats at different stages of tidal inundation. Marine Ecological Jersey. Estuaries 18: 226–234. Progress Series 106: 227–238. Talbot, C. W. & K. W. Able, 1984. Composition and distribution of Laffaille, P., S. Brosse, E. Feunteun, A. Baisez & J. C. Lefeuvre, larval fishes in New Jersey high marshes. Estuaries 7: 434–443. 1998. Role of fish communities in particulate organic matter Weinstein, M. P., S. L. Weiss & M. F. Walters, 1980. Multiple fluxes between salt marshes and coastal marine waters in the determinants of community structure in shallow marsh habit- Mont Saint-Michel Bay. Hydrobiologia 373/374: 121–133. ats, Cape Fear River estuary, North Carolina. Marine Ecological Laffaille, P., E. Feunteun & J.C. Lefeuvre, 2000. Composition of Progress Series 58: 227–243. fish communities in a European macrotidal salt marsh (the Mont Werner, E. E., J. F. Gilliam, D. J. Hall & G. G. Mittelbach, 1983. Saint-Michel Bay, ). Estuarine, Coastal and Shelf Science An experiment test of the effects of predation risk on habitat use 51: 429–438. in fish. Ecology 64: 1540–1548. Layman, C. A., D. E. Smith & J. D. Herold, 2000. Seasonally vary- Williams, G. D. & J. B. Zedler, 1999. Fish assemblage composition ing importance of abiotic and biotic factors in marsh-pond fish in constructed and natural tidal marshes of San Diego Bay: Re- communities. Marine Ecological Progress Series 207: 155–169. lative influence of channel morphology and restoration history. Lefeuvre, J. C., V. Bouchard, E. Feunteun, S. Grare, P. Laffaille & Estuaries 22: 702–716. A. Radureau, 2000. European salt marshes diversity and func- Zedler, J. B., G. D. Williams & J. S. Desmond, 1997. Wetland mit- tioning: The case study of the Mont Saint-Michel Bay, France. igation: can fishes distinguish between natural and constructed Wetlands, Ecology and Management 8: 147–161. wetlands? Fisheries 22: 26–28.