Estuarine, Coastal and Shelf Science 67 (2006) 461e474www.elsevier.com/locate/ecss

Seasonal, tidal and diurnal changes in fish assemblages in theRia Formosa lagoon (Portugal)

Joaquim Ribeiro*, Luis Bentes, Rui Coelho, Jorge M.S. Goncalves,Pedro G. Lino, Pedro Monteiro, Karim Erzini

Centro de Ciencias do Mar, Faculdade de Ciencias do Mar e do Ambiente, Universidade do Algarve,

Campus de Gambelas, 8005-139 Faro, Portugal

Received 17 June 2005; accepted 28 November 2005

Available online 24 February 2006

Abstract

Fish fauna were collected in two different subtidal shallow sites, seagrass and sand, using a small beam trawl in the Ria Formosa lagoon(South Portugal). Samples were taken at low and high tides, during day and night, and in each season. Fish assemblages associated witheach site were significantly different, with seagrass site supporting greater fish abundance and higher number of species than sand. Thesesite-related differences in fish assemblages were stronger than any other factor studied. Both sites showed seasonal variations in their fish as-semblages, mainly because of recruitment of marine juvenile migrants during spring and summer. No significant tidal or diel changes were ob-served in the fish assemblages of either site, but there was a significant siteetide interaction, with higher fish abundance in seagrass at low tide.In sand, tide effect was evident only for certain species, with resident species more abundant at high tide and marine species more abundant atlow tide. Within the Ria Formosa coastal lagoon, ichthyofaunal composition and structure is mainly influenced by site followed by season.� 2005 Elsevier Ltd. All rights reserved.

Keywords: fish fauna; coastal lagoon; seagrass; sand; temporal variations; Ria Formosa; South Portugal

1. Introduction

Shallow areas of estuaries and coastal lagoons containsome of the most productive coastal habitats, such as tidalflats, seagrass beds, subtidal channels and salt marshes (Pihland Rosenberg, 1982; Weinstein, 1982; Kneib, 1997). Manyof these shallow habitats support diverse and abundant fishassemblages, and sustain significant populations of juvenilesof many commercially important fish species (Bell and Pol-lard, 1989; Potter et al., 1990; Edgar and Shaw, 1995a; Lazzariand Tupper, 2002). However, the ichthyofauna is heteroge-neously distributed among the different types of habitats,according to sediment type, vegetation preference and tempo-ral scale (tidal, diel, lunar, seasonal). Several studies have

* Corresponding author.

E-mail address: jribeiro@ualg.pt (J. Ribeiro).

0272-7714/$ - see front matter � 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.ecss.2005.11.036

compared seagrass beds with non-vegetated substrates, inparticular sand, and have generally reported that the seagrasssupports different and more abundant and diverse fish assem-blages than non-vegetated habitats (e.g. Heck et al., 1989;Sogard and Able, 1991; Connolly, 1994a; Gray et al., 1996;Jenkins et al., 1997; Guidetti, 2000; Travers and Potter,2002). Besides the differences in fish assemblages associatedwith the various habitat types, ichthyofaunal compositionsand structure can also undergo consistent cyclical and tempo-ral changes within habitats. Seasonal shifts in fish communi-ties are common, as a result of sequential immigration andemigration of certain fish species (Hyndes et al., 1999; Thieland Potter, 2001). Tidal and diel shifts in fish assemblageshave also been reported by several authors (e.g. Sogardet al., 1989; Rountree and Able, 1993; Gray et al., 1998;Griffiths, 2001; Methven et al., 2001; Morrison et al., 2002;Guest et al., 2003). The aim of this study was to examine

462 J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

the effect of season, tide and diel variations in fish assem-blages associated with two subtidal sites, seagrass and sand,present in the Ria Formosa coastal lagoon (Algarve, SouthPortugal).

2. Materials and methods

2.1. Study area

This study was carried out in the western part of the RiaFormosa (Ria Faro-Olh~ao). The Ria Formosa is a large tidallagoon extending for about 55 km along the south coast ofPortugal, with a maximal width of 6 km (Fig. 1). A seawardbelt of dunes protects a system of salt marshes, subtidal chan-nels and tidal flats, with a total surface area of approximately170 km2. The tidal elevations are 1.30 and 2.80 m at meanneap tide and spring tide, respectively, and the minimum andmaximum areas covered by water during spring tides are14.1 and 63.1 km2 (�Aguas, 1986). A strongly branched systemof creeks and channels is connected with the ocean by six out-lets. The average depth is less than 3 m, with 14% of the la-goon surface permanently submersed (subtidal channels) andthe intertidal area covers approximately 1/3 of the total areaof the lagoon. The system has semi-diurnal tides, with 50e75% of the water volume exchanged during each tide (�Aguas,1986). The lagoon does not receive any permanent freshwaterinput: salinity is 35.5e36.9 all year, except for surface watersfor brief periods after winter heavy rainfalls (Falc~ao et al.,1992). Water temperature varies from 12 to 28 �C (Sprung,1994).

2.2. Sampling design

Fish fauna were collected from two different subtidal shal-low sites, separated by 1.5 km (Fig. 1). One site representedthe seagrass habitat and was located in a medium size subtidalchannel (37 �000N, 07 �570W), with around 50 m width and anaverage depth of 1.5 m at low tide. The tidal current velocityis relatively low compared to that observed in the vicinity of

the inlets (Lima and Vale, 1977), and the bottom consists ofmuddy sandy sediments (sand e 35.7%; silt e 31.7%; clay e32.6%), covered with several extensive patches of seagrass,Cymodocea nodosa. The second site represented the sand hab-itat and was located in a subtidal channel that gives access to aninlet, artificially opened in 1997 (36 �590N, 07 �570W). The tidalcurrent velocity is much stronger than in the seagrass site, andthe sediments were dominated by well calibrated medium-coarse sand (sand e 99.4%; silt e 0.3%; clay e 0.3%). Thischannel was 30 m wide with an average depth of 1.9 m at lowtide and no vegetation was present during the sampling period.

Both sites were sampled on a seasonal basis over a one-yearperiod, with samples taken in May 2001 (spring), August 2001(summer), November 2001 (autumn) and February 2002 (win-ter). In each season the sampling took place in one single day,and the two sites were visited at the beginning of the ebb tide(high tide sampling) and at the beginning of the flood tide (lowtide sampling), both during the day and night. At each visit,three replicate tows of 200 m each were made, using a beamtrawl (2.6 m wide and 0.45 m high at the mouth) witha stretched mesh size of 9 mm in the cod end. Since this sam-pling gear is not effective for catching pelagic species (e.g.Atherina spp., Liza spp. and Sardina pilchardus), this study fo-cuses on benthic and epibenthic fish communities, present ineach sampled site. All samples were taken during neap tide,near the last quarter moon. In the laboratory, all fish were iden-tified to species and counted.

2.3. Data analysis

The independent variables were site (seagrass and sand),season (spring, summer, autumn and winter), tide stage(high/ebb tide and low/flood tide) and diel period (day andnight). A four-way factorial analysis of variance (ANOVA)was used to test for differences in the number of species andindividuals, with all factors considered fixed (SAS Institute,1988). Number of species was square-root transformed andnumber of individuals was ln transformed. When significantmain effects were detected in the multifactor ANOVAs,

Fig. 1. Western part of the Ria Formosa lagoon (Ria Faro-Olh~ao) showing the location of the two sampling sites: (1) seagrass and (2) sand.

463J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

Tukey’s Honestly Significant Difference (HSD) test was usedto find which means differed. Analysis was performed on thetransformed data but the level of significance set to P� 0.01 tominimize the chances of Type I errors occurring (Underwood,1981). Significant interactions between factors were examinedgraphically.

To test for differences in number of individuals betweenseason, tide and diel period, three-way factorial analysis ofvariance (ANOVA) were performed for the 10 most abundantspecies in each site. Data from seagrass site were square-roottransformed and Tukey’s HSD test was used when significantdifferences were detected (P� 0.05).

Spatial and temporal variations in the structure of the fishassemblages were assessed by non-metric multi-dimensionalscaling (nMDS) and analysis of similarities (ANOSIM) (Fieldet al., 1982; Clarke, 1993) using the statistical softwarePRIMER 5.0 (Clarke and Warwick, 2001). Both presence/absence and abundance data were used. For the abundancedata square-root transformed values were used so that eachspecies contributed fairly evenly to each analysis (Clarkeand Green, 1988), except in the seasonal analysis for thesand site because species abundances were very low. Samplesimilarity matrices based on the BrayeCurtis similarity coef-ficient were generated, after which non-metric multi-dimensionalscaling (nMDS) was used to create two-dimensional ordinationplots, and one-way analysis of similarities (ANOSIM) wasused to determine which assemblages differed (Clarke,1993). Pairwise ANOSIM comparisons were made betweenall groups, using 10 000 simulations in each case.

To compare the fish assemblage functional structure in re-lation to each independent variable, species were classified ac-cording to their ecological guild adapted from Elliott andDewailly (1995): (D) diadromus migrant species, (MA) ma-rine adventitious visitors, (MJ) marine juvenile migrants and(R) resident species. The vertical distribution guild adaptedfrom the same authors was used to compare the fish assem-blages present in the two sites: (B) benthic and (EB) epi-benthic species.

3. Results

A total of 17 232 fish were caught, representing 63 speciesand 21 families. Seagrass supplied 15 796 fish (44 species) andsand 1435 fish (44 species) (Table 1). Twenty-six species(41.3%) were common to both sites, 19 species (30.2%) oc-curred only in the seagrass, and 18 species (28.6%) occurredonly in sand (Table 1). In sand, seven species were presentonly during the low tide, 10 were caught exclusively at hightide, and 30 were present during both tidal stages. In seagrass,12 species were found only during the low tide, seven werecaught exclusively at high tide and 29 were common to bothtidal stages. Species present in both tidal stages representedmore than 97% of the total catch in each tidal stage in bothsites. Again in sand, eight species were present only duringthe day, 10 were caught exclusively at night and 29 were pres-ent during both day and night. In seagrass, nine species were

found only during the day, 10 were caught exclusively at nightand 29 were common to both periods. In each site, the speciespresent during both day and night, represented more than 99%of the total catch in each period, indicating that species thatoccur only at night or daytime were poorly associated withthe site assemblages.

3.1. Number of species and individuals

Overall number of species was similar for both sites (Table 1),but mean number of species was significantly higher in sea-grass (Table 2). Season effect was also responsible for signif-icant differences in number of species, but siteeseasoninteraction was also significant and therefore trends withinsite should be investigated (Table 2). In seagrass, the TukeyHSD test revealed that mean number of species differed sig-nificantly between all seasons except for spring and summer,which had significantly higher numbers than autumn andwinter (Fig. 2). In sand, mean number of species did not dif-fer between spring and summer, and between autumn andwinter, with the first two seasons having significantly highernumbers than the later. This interaction also indicated thatnumber of species was significantly higher in seagrass thansand, across all seasons. All other main effects and interac-tions were not significant.

The overall fish abundance was 10 times higher in seagrassthan sand, with a significant difference in mean number of fishbetween sites (Tables 1 and 2). There was also a significant sea-sonal variation in mean number of fish. Site� season interactionwas significant, with seagrass presenting significantly highermean numbers of fish in spring and summer compared to au-tumn, which had significantly higher abundance than winter(Fig. 3). In sand, only spring differed significantly from otherseasons, with higher mean fish abundance. There was no signif-icant effect of tide and diel period on mean fish abundance, butthere was significant siteetide, seasonediel and siteeseasonetideediel interactions. Siteetide interaction indicated thatmean number of fish was significantly higher at low tide in theseagrass site, while in sand there was no significant differencebetween tidal stages (Fig. 4). For seagrass, mean fish abundancewas never significantly higher during high tide, while it was sig-nificantly higher during low tide for at least one sampling visit ineach season (Fig. 5). With respect to seasonediel and siteesea-sonetideediel interactions, although significant differences inmean fish abundance were identified, no clear pattern of meanfish abundance was identified, with F-values lower than othersignificant effects (Table 2).

3.2. Species composition

The species common to both sites Gobius niger, Diplodusvulgaris and Symphodus bailloni had a similar rank order inabundance in each site, but were much more abundant in the sea-grass (Table 1). Gobius paganellus, Spondyliosoma cantharusand Symphodus cinereus showed a much greater abundanceand a higher rank order of abundance in this site (Table 1).The syngnathidae species, Hippocampus guttulatus, Syngnathus

464 J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

Table 1

List of fish families and species caught over seagrass and over sand, total abundance (n), relative abundance (n%), and ecological and vertical distribution

classification

Family Species Sand Seagrass Guilds

n n(%) n n(%) Ecological Vertical

Anguillidae Anguilla anguilla 1 0.07 64 0.41 D B

Batrachoididae Halobatrachus didactylus 0 e 621 3.93 MJ B

Blenniidae Salaria pavo 0 e 3 0.02 R B

Parablennius gattorugine 4 0.28 15 0.09 R B

Parablennius pilicornis 6 0.42 0 e R B

Bothidae Arnoglossus laterna 1 0.07 0 e MA B

Arnoglossus thori 19 1.32 0 e R B

Bothus podas 30 2.09 0 e MA B

Callionymidae Callionymus maculatus 37 2.58 0 e MA B

Callionymus risso 1 0.07 0 e MA B

Congridae Conger conger 1 0.07 5 0.03 MA B

Gobiidae Gobius couchi 2 0.14 284 1.80 R B

Gobius cruentatus 1 0.07 3 0.02 MA B

Gobius niger 145 10.10 5080 32.16 R B

Gobius paganellus 25 1.74 2851 18.05 R B

Pomatoschistus microps 53 3.69 28 0.18 R B

Pomatoschistus minutus 798 55.57 0 e R B

Pomatoschistus pictus 16 1.11 2 0.01 MA B

Gobiesocidae Diplecogaster bimaculata 0 e 22 0.14 R B

Labridae Ctenolabrus rupestris 0 e 1 0.01 MA EB

Labrus bergylta 0 e 1 0.01 MA EB

Labrus merula 0 e 1 0.01 MA EB

Symphodus bailloni 44 3.06 302 1.91 R EB

Symphodus cinereus 14 0.97 3151 19.95 R EB

Symphodus melops 1 0.07 0 e MA EB

Symphodus ocellatus 0 e 2 0.01 MA EB

Symphodus roissali 1 0.07 4 0.03 MA EB

Symphodus rostratus 0 e 1 0.01 MA EB

Mullidae Mullus surmuletus 12 0.84 0 e MJ B

Rajidae Raja undulata 7 0.49 0 e MA B

Scophthalmidae Scophthalmus maximus 1 0.07 0 e MA B

Scophthalmus rhombus 4 0.28 0 e MA B

Scorpaenidae Scorpaena notata 9 0.63 1 0.01 MA B

Scorpaena porcus 9 0.63 57 0.36 MJ B

Serranidae Serranus cabrilla 0 e 1 0.01 MA EB

Serranus hepatus 1 0.07 2 0.01 MA EB

Soleidae Microchirus azevia 3 0.21 4 0.03 MJ B

Microchirus boscanion 0 e 3 0.02 MA B

Monochirus hispidus 4 0.28 0 e R B

Solea lascaris 22 1.53 0 e MJ B

Solea senegalensis 0 e 1 0.01 MJ B

Synaptura lusitanica 2 0.14 0 e MA B

Sparidae Dentex marrocanus 0 e 1 0.01 MA EB

Diplodus annularis 0 e 650 4.11 MJ EB

Diplodus bellottii 0 e 105 0.66 MJ EB

Diplodus cervinus 1 0.07 0 e MA EB

Diplodus puntazzo 0 e 7 0.04 MJ EB

Diplodus sargus 2 0.14 48 0.30 MJ EB

Diplodus vulgaris 47 3.27 873 5.53 MJ EB

Lithognathus mormyrus 1 0.07 0 e MJ EB

Sarpa salpa 3 0.21 50 0.32 MJ EB

Spondyliosoma cantharus 15 1.04 1057 6.69 MJ EB

465J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

Table 1 (continued)

Family Species Sand Seagrass Guilds

n n(%) n n(%) Ecological Vertical

Syngnathidae Hippocampus guttulatus 15 1.04 223 1.41 R EB

Hippocamus hippocampus 6 0.42 10 0.06 R EB

Nerophis lumbriciformes 0 e 1 0.01 MA EB

Nerophis ophidion 0 e 4 0.03 R EB

Syngnathus abaster 1 0.07 12 0.08 R EB

Syngnathus acus 5 0.35 133 0.84 R EB

Syngnathus typhle 2 0.14 110 0.70 R EB

Torpedinidae Torpedo marmorata 0 e 1 0.01 MA B

Trachinidae Echiichthys vipera 59 4.11 0 e MA B

Trachinus draco 5 0.35 0 e MA B

Triglidae Trigla lucerna 0 e 1 0.01 MA B

Total no. of individuals 1436 e 15 796 e

Total no. of species 44 e 44 e

Resident 1143 79.6 12 228 77.4

Marine juvenile migrants 114 8 3473 22

Marine adventitous 178 12.4 31 0.2

acus and Syngnathus typhle, although less abundant than the pre-vious three species, were much more abundant in the seagrassthan in sand. On the other hand, although the overall abundanceof Pomatoschistus microps, Pomatoschistus pictus and Scor-paena porcus was low, these species were more abundant inthe sand site. Of the species present only in the seagrass, Diplo-dus annularis and Halobatrachus didactylus were numericallythe most important. Of the species present only in sand, themost important were Arnoglossus thori, Bothus podas, Calliony-mus maculatus, Echiichthys vipera, Mullus surmuletus, Solealascaris and, in particular, Pomatoschistus minutus (55.6% ofthe total fish abundance in sand).

The 10 most numerically important species in the seagrassall revealed significant differences in abundance with season(Table 3). The resident species Gobius niger, Gobius

Table 2

F-values and significance levels for four-way ANOVAs of mean number of

species and fish, testing for differences between sites (H), season (S), tide

(T) and diel (D)

Effects df No. of species No. of fish

F P F P

Site (H) 1 202.71 <0.001 1117.38 <0.001

Season (S) 3 30.51 <0.001 29.21 <0.001

Tide (T) 1 1.39 0.234 3.69 0.059

Diel (D) 1 0.55 0.462 3.76 0.057

H� S 3 4.98 0.004 12.19 <0.001

H� T 1 0.04 0.848 13.03 <0.001

H�D 1 3.52 0.065 1.81 0.183

S� T 3 0.26 0.856 0.47 0.704

S�D 3 1.59 0.201 6.32 <0.001

T�D 1 0.14 0.711 0.46 0.498

H� S� T 3 0.42 0.740 3.76 0.015

H� S�D 3 0.40 0.757 0.51 0.677

H� T�D 1 0.23 0.634 1.86 0.178

S� T�D 3 0.50 0.682 0.21 0.887

H� S� T�D 3 2.59 0.060 8.30 <0.001

paganellus and Symphodus cinereus clearly dominated thefish assemblages in the seagrass throughout the year, withsome seasonal variation. Gobius niger abundance was highestin the spring, with lowest values in the summer (Table 4). Incontrast, the abundance of G. paganellus was lowest in thespring and highest in the summer. Symphodus cinereus abun-dance was maximal in autumn and minimal in the spring.The juvenile migrant species Diplodus vulgaris, Diplodus an-nularis, Halobatrachus didactylus and Spondyliosoma cantha-rus were absent or almost absent in the autumn and winter, butwere both abundant and frequent in the spring and/or summer.In seagrass, the three-way ANOVA also detected a significanttide effect on number of fish for all species, except forHippocampus guttulatus (Table 3). Tukey HSD test revealedthat those species were significantly more abundant at lowtide. Diel effect on fish abundance was present in only one spe-cies, H. didactylus, which was significantly more abundant atnight. Interaction terms were significant in very few cases, andno three factor interactions were found. F-values of interac-tions were almost always less than main effects, in particular

0

4

8

12

16

20Sand Seagrass

A

B B

C

D

EA

Sprin SprinSummer SummerAutum AutumWinte Winter

C

Mea

n nu

mbe

r of

spe

cies

Fig. 2. Mean number of species over seagrass and sand across all seasons. Er-

ror bars represent þ1 SE and unlike letters denote seasons that differed statis-

tically from each other in Tukey HSD test.

466 J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

for season and tide, indicating that these factors made a greatercontribution to the source of variance.

In sand, a significant seasonal effect on number of fish wasfound for all 10 most important species, except for Calliony-mus maculatus for which no factors or interactions were sig-nificant (Table 5). Pomatoschistus minutus clearly dominatedthe fish assemblage in sand throughout the year, with greatestabundance in spring and lowest in summer (Table 6). In thesummer Bothus podas, Gobius paganellus and Symphodusbailloni were relatively abundant, while they were absent oralmost absent in the rest of year. Diplodus vulgaris and Solealascaris were relatively abundant in spring and summer, butthey were absent or almost absent in the rest of year. Gobiusniger and Pomatoschistus microps were abundant only inspring. Tide effect was responsible for differences in fish abun-dance in half of the species tested for sand (Table 5). The Tu-key HSD test revealed that the abundance of marineadventitious species, B. podas and Echiichthys vipera, and ma-rine juvenile migrant species, D. vulgaris, were significantlyhigher at high tide, while those of resident species G. nigerand P. microps, were significantly higher at low tide. Diel pe-riod has a significant effect on the abundance of three species,E. vipera, P. microps and S. bailloni.

3.3. Functional guilds

In terms of ecological guilds, both sites were clearly dom-inated by resident species, but marine juvenile migrant specieswere more important in seagrass, and marine adventitious vis-itors were more important in sand (Table 1). In both sites, therelative importance of each ecological guild follows a clear

0

100

200

300

400

500

AB B B

Spring Summer Autum Winter Spring Summer Autum Winter

Sand Seagrass

C

D

C

D

Mea

n nu

mbe

r of

fis

h

Fig. 3. Mean number of fish over seagrass and sand across all seasons. Error

bars represent þ1 SE and unlike letters denote seasons that differed statisti-

cally from each other in Tukey HSD test.

0

100

200

300

400

500

Mea

n nu

bmer

of

fish

A

B

Low tide High tide

Sand Seagrass

Low tide High tide

A

C

Fig. 4. Mean number of fish over seagrass and sand across at high and low tide.

Error bars represent þ1 SE and unlike letters denote seasons that differed sta-

tistically from each other in Tukey HSD test.

seasonal pattern, with marine juvenile migrant species beingmore abundant in spring and summer, and absent or almost ab-sent in the autumn and winter, while the resident species dom-inated the assemblage year round (Tables 4 and 6). In seagrass,the relative importance of each ecological guild was similarfor both tide stages, while in sand, marine adventitious and mi-grant species (Bothus podas, Callionymus maculatus, Echiich-thys vipera and Diplodus vulgaris) were more important at lowtide, and the resident species (e.g. Pomatoschistus minutus,Pomatoschistus microps, Gobius niger and Gobius paganellus)were more important at high tide (Table 6). The relative im-portance of the ecological groups was similar during dayand night in both sites.

Relative fish abundance in the vertical distribution guildsfor each site shows that sand was clearly dominated by benthicspecies, while both benthic species and epibenthic speciespopulated seagrass (Fig. 6a). In terms of ecological guildsboth habitats were clearly dominated by resident species,while the marine juvenile migrant species were more impor-tant in the seagrass than in the bare sand and marine adventi-tious visitors were more important in the bare sand than in theseagrass (Fig. 6b).

3.4. Fish assemblage structure

The nMDS ordination plot based on presence/absence dataclearly shows a separation of samples from sand and seagrass(Fig. 7a). The seagrass samples are tightly grouped, indicatinghigh similarities, while the sand samples constitute a more dis-persed group, with a lower similarity between samples. A simi-lar pattern was evident using abundance data (Fig. 7b),indicating that it was the presence/absence of species, ratherthan their abundance, that separated these two groups. ANOSIMindicated that the fish assemblages associated with the two sitesdiffered significantly (R-statistic¼ 0.985, P< 0.001).

The nMDS ordination showed a clear separation by seasonof seagrass samples, both for presence/absence and abundancedata (Fig. 8a,c), while sand samples showed some seasonaloverlap, especially for presence/absence data (Fig. 8b,d).ANOSIM revealed an overall significant seasonal effect onthe fish assemblage structure for both sites, with significantdifferences between all pairwise seasonal comparisons, exceptfor autumn and winter in the sand (Table 7). ANOSIM re-vealed no significant differences in fish assemblage structureassociated with tidal effect for seagrass (R-statistic¼ 0.053,P¼ 0.069) or sand (R-statistic¼ 0.005, P¼ 0.337), and nosignificant differences in fish assemblage structure associatedwith diel effect for seagrass (R-statistic¼ 0.023, P¼ 0.184)or sand (R-statistic¼ 0.029, P¼ 0.112).

4. Discussion

4.1. Site differences

The fish fauna of Ria Formosa was more diverse and con-siderably more abundant over seagrass than over sand, asgenerally reported in many studies elsewhere (e.g. Heck

467J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

0

100

200

300

400

500

600

700

800

900

(a)

Spring Summer AutumnWinter

Low tide Low tide Low tide Low tideHigh tide High tide High tide High tide

0

10

20

30

40

50

60

70

80

90

Mea

n nu

mbe

r of

fis

hM

ean

num

ber

of f

ish

(b)Spring Summer AutumnWinter

Low tide Low tide Low tide Low tideHigh tide High tide High tide High tide

Fig. 5. Mean number of fish over seagrass (a) and sand (b) across season, tide stage and diel period. Error bars represent þ1 SE. White bars e day samples and

dark bars e night samples.

et al., 1989; Sogard and Able, 1991; Connolly, 1994a; Grayet al., 1996; Guidetti, 2000; Travers and Potter, 2002). Previ-ously, Monteiro et al. (1990) found that the fish assemblagesassociated with seagrass were more diverse and abundantthan other habitat types inside the Ria Formosa. These differ-ences in fish diversity and abundance are likely due to higherstructural complexity and higher productivity of the seagrassbeds compared to unvegetated areas, providing protection andshelter from predators (Stoner, 1983; Orth et al., 1984; Pol-lard, 1984; Bell and Pollard, 1989; Connolly, 1994a) and of-fering higher food availability and diversity (Burchmoreet al., 1984; Connolly, 1994b). In the Ria Formosa, Calvario(unpublished data) found that invertebrate communities of

subtidal seagrass habitats were more diverse and abundantthan those of subtidal sand, in part due to an important pres-ence of detritivorous and herbivorous fauna associated withstructural elements of the macrophytes (e.g. leaves andrhizomes).

In the Ria Formosa, fish assemblage structures associatedwith seagrass also differed from those found over sand, likein other parts of the world (e.g. Gray et al., 1996; Jenkinset al., 1997; Gray et al., 1998). The multivariate analysesshowed that site-related differences were greater than any sea-sonal, tidal or diel variation, with several species occurringconsistently in only one site, or predominantly caught in onesite, indicating that different types of fish species were

Table 3

F-values and significance levels for three-way ANOVAs of mean number of fish of 10 most numerically abundant species in seagrass, testing for differences be-

tween season (S), tide (T) and diel (D). *P< 0.05; **P< 0.01; ***P< 0.001; (ns) not significant

Species Season (S) Tide (T) Diel (D) S� T S�D T�D S� T�D

D. annularis 19.905*** 4.186* 1.744 (ns) 1.150 (ns) 6.662*** 7.053* 3.561 (ns)

D. vulgaris 465.828*** 10.557** 0.645 (ns) 8.065*** 3.851* 0.594 (ns) 2.924 (ns)

G. couchi 36.492*** 8.736** 0.785 (ns) 3.103* 0.382 (ns) 1.046 (ns) 2.466 (ns)

G. niger 9.257*** 31.126*** 0.311 (ns) 0.792 (ns) 1.946 (ns) 1.470 (ns) 3.361 (ns)

G. paganellus 20.581*** 7.547** 1.042 (ns) 1.900 (ns) 2.335 (ns) 0.189 (ns) 3.247 (ns)

H. didactylus 45.788*** 17.902*** 6.404** 1.986 (ns) 4.720** 0.057 (ns) 2.648 (ns)

H. guttulatus 8.970*** 0.045 (ns) 1.463 (ns) 0.783 (ns) 1.115 (ns) 2.107 (ns) 0.553 (ns)

S. cantharus 11.151*** 4.602* 2.003 (ns) 1.558 (ns) 1.738 (ns) 0.012 (ns) 0.303 (ns)

S. bailloni 61.273*** 5.102* 0.689 (ns) 1.961 (ns) 2.559 (ns) 1.551 (ns) 1.090 (ns)

S. cinereus 8.308*** 5.062* 2.737 (ns) 0.518 (ns) 0.900 (ns) 0.003 (ns) 0.332 (ns)

Table 4

nce, and frequency of occurrence ( f %)

Diel

Low Day Night

n n(%) n n(%) n n(%)

26 0.27 32 0.39 32 0.42

14 0.14 20 0.25 2 0.03

422 4.31 327 4.02 323 4.21

61 0.62 60 0.74 45 0.59

36 0.37 26 0.32 22 0.29

546 5.58 385 4.73 488 6.37

191 1.95 139 1.71 145 1.89

3441 35.15 2443 30.04 2637 34.41

1766 18.04 1368 16.82 1483 19.35

419 4.28 292 3.59 329 4.29

118 1.21 123 1.51 100 1.30

10 0.10 22 0.27 6 0.08

28 0.29 8 0.10 42 0.55

26 0.27 24 0.30 33 0.43

597 6.10 639 7.86 418 5.45

177 1.81 181 2.23 121 1.58

1744 17.82 1861 22.88 1290 16.83

69 0.70 82 1.01 51 0.67

55 0.56 58 0.71 52 0.68

9561 e 8132 e 7664 e

37 e 37 e 35 e

7384 77.2 6319 77.7 5909 77.1

2133 22.3 1762 21.7 1711 22.3

18 0.2 19 0.2 12 0.2

46

8J.

Ribeiro

etal.

/E

stuarine,C

oastaland

ShelfScience

67(2006)

461e474

List of the most abundant species caught in the seagrass site, by season, tidal stage and diel period, presenting their absolute (n) and relative (n%) abunda

Species Season Tide

Spring Summer Autumn Winter High

n n(%) f(%) n n(%) f(%) n n(%) f(%) n n(%) f(%) n n(%)

Anguilla anguilla 13 0.30 50.0 51 1.03 91.7 0 e 0.0 0 e 0.0 38 0.63

Diplecogaster

bimaculata

16 0.36 50.0 1 0.02 8.3 5 0.15 25.0 0 e 0.0 8 0.13

Diplodusannularis

188 4.27 91.7 391 7.87 91.7 51 1.51 75.0 20 0.66 58.3 228 3.80

Diplodus

bellottii

0 e 0.0 98 1.97 91.7 7 0.21 41.7 0 e 0.0 44 0.73

Diplodus sargus 11 0.25 66.7 15 0.30 50.0 20 0.59 66.7 2 0.07 16.7 12 0.20

Diplodus vulgaris 822 18.66 100.0 50 1.01 91.7 1 0.03 8.3 0 e 0.0 327 5.44

Gobius couchi 220 5.00 91.7 10 0.20 58.3 11 0.33 33.3 43 1.41 91.7 93 1.55

Gobius niger 1847 41.94 100.0 652 13.12 100.0 1161 34.4 100.0 1420 46.6 100.0 1639 27.28

Gobius

paganellus

218 4.95 100.0 1503 30.24 100.0 731 21.66 100.0 399 13.09 100.0 1085 18.06

Halobatrachus

didactylus

50 1.14 83.3 405 8.15 100.0 126 3.73 100.0 40 1.31 83.3 202 3.36

Hippocampus

guttulatus

86 1.95 100.0 44 0.89 75.0 19 0.56 83.3 74 2.43 100.0 105 1.75

Pomatoschistus

microps

26 0.59 58.3 2 0.04 8.3 0 e 0.0 0 e 0.0 18 0.30

Sarpa salpa 41 0.93 66.7 0 e 0.0 0 e 0.0 9 0.30 16.7 22 0.37

Scorpaena

porcus

9 0.20 33.3 23 0.46 58.3 24 0.71 83.3 1 0.03 8.3 31 0.52

Spondyliosoma

cantharus

473 10.74 100.0 504 10.14 100.0 74 2.19 100.0 6 0.20 25.0 460 7.66

Symphodus

bailloni

19 0.43 66.7 242 4.87 100.0 38 1.13 100.0 3 0.01 16.7 125 2.08

Symphodus

cinereus

312 7.08 100 880 17.71 100.0 1017 30.13 100.0 942 30.92 100.0 1407 23.42

Syngnathus

acus

15 0.34 75.0 39 0.78 100.0 36 1.07 83.3 43 1.41 100.0 64 1.07

Syngnathus

typhle

15 0.34 75.0 21 0.42 50.0 34 1.01 91.7 40 1.31 91.7 55 0.92

Total no.

of individuals

4404 e e 4970 e e 3375 e e 3047 e e 6235 e

Total no.

of species

29 e e 32 e e 23 e e 18 e e 35 e

Resident 2786 63.3 e 3413 68.7 e 3065 90.8 e 2964 97.3 e 4844 77.7

Marine

juvenile migrants

1600 36.3 e 1491 30.0 e 303 8.0 e 79 2.6 e 1340 21.5

Marine

adventitious

5 0.1 e 15 0.3 e 7 0.2 e 4 0.1 e 13 0.2

469J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

Table 5

F-values and significance levels for three-way ANOVAs of mean number of fish of 10 most numerically abundant species in sand, testing for differences between

season (S), tide (T) and diel (D). *P< 0.05; **P< 0.01; ***P< 0.001; (ns) not significant

Species Season (S) Tide (T) Diel (D) S� T S�D T�D S� T�D

B. podas 11.971*** 4.261* 0.087 (ns) 2.232 (ns) 0.841 (ns) 2.174 (ns) 1.768 (ns)

C. maculatus 0.447 (ns) 2.750 (ns) 2.750 (ns) 2.750 (ns) 0.568 (ns) 3.841 (ns) 1.538 (ns)

D. vulgaris 8.782*** 7.750** 2.331 (ns) 6.503*** 0.825 (ns) 2.331 (ns) 1.212 (ns)

E. vipera 16.87*** 23.439*** 37.098*** 11.732*** 19.667*** 12.902*** 11.211***

G. niger 11.095*** 4.538* 2.579 (ns) 3.424* 3.659* 0.671 (ns) 1.195 (ns)

G. papganellus 4.231* 0.338 (ns) 0.043 (ns) 0.396 (ns) 0.303 (ns) 0.051 (ns) 0.300 (ns)

P. microps 29.344*** 15.044*** 3.967 (ns) 13.989*** 4.553* 10.560** 11.498***

P. minutus 96.590*** 3.490 (ns) 10.990** 5.880** 22.600*** 0.130 (ns) 2.720 (ns)

S. lascaris 3.455* 0.727 (ns) 2.909 (ns) 0.121 (ns) 1.576 (ns) 0.182 (ns) 0.303 (ns)

S. bailloni 3.237* 1.306 (ns) 5.226* 2.758 (ns) 3.366* 1.306 (ns) 2.242 (ns)

associated with these two sites. The dominance of benthic spe-cies like Arnoglossus thori, Bothus podas, Callionymus macu-latus, Echiichthys vipera, Mullus surmuletus, Solea lascarisand Pomatoschistus minutus, observed only in sand showa clear adaptation of the fish assemblage to exposed hydrody-namic conditions, associated with this site. These conditionsfavour the settlement of benthic species that avoid strong tidalcurrents by close association with the sandy bottom. In con-trast, the seagrass site located over muddy sediment is associ-ated with weaker current velocities, allowing the settlement ofepibenthic fish along with benthic fish species. These condi-tions explain the important presence of species with poor mo-bility, like syngnathids and young-of-the-year (YOY) ofmarine migrant species. The higher structural complexityalso enhances settlement capabilities, working like a trap forYOY (Bell et al., 1987) and allowing the attachment of syn-gnathids to the leaves and rhizomes (Howard and Koehn,1985). The benthic fish assemblage also differs from sand,with larger gobids like Gobius niger and Gobius paganellus,dominating this part of the assemblage.

The high abundance of YOY in the seagrass, almost 20times higher than in the sand, is clear evidence that seagrassis an important habitat for juvenile fish, in particularcommercially important species like sparids. Structural com-plexity may aid the settlement of YOY by providing protec-tion against predation, food and enhancing early growth andsurvival. The sand site, in spite of the low abundance ofYOY, is nevertheless an important habitat for juvenile fishfrom a different suite of species including Callionymus mac-ulatus, Echiichthys vipera, Mullus surmuletus and Solealascaris.

4.2. Seasonal changes

The seasonal patterns in fish assemblage structure, with in-creased number of species and fish in the spring and summer,were consistent with the findings of other studies in the RiaFormosa (Monteiro et al., 1987), and elsewhere in estuariesand other temperate shallow habitats (e.g. Nash and Gibson,1982; Potter et al., 1986; Guidetti and Bessotti, 2000). The re-cruitment of marine juvenile migrant species during the springand summer is consistent with the late winter and springspawning season of most of these species (Goncalves and

Erzini, 2000a,b). The decline in numbers of these fish speciesin autumn and winter suggests they may migrate to the adja-cent coastal areas within their first year. The seasonal fluctua-tions in the abundance of the dominant resident species couldalso be attributed to their life cycle, since the majority of thesespecies breed during the warmer seasons precisely whenchanges in their abundance occurs.

Seasonal changes in fish assemblage structure in both sitesis a consequence of the seasonal shifts in species abundanceand species composition, reflecting both spawning eventswithin the lagoon and seasonal movements to and from the ad-jacent marine environment (Livingston et al., 1976; Yoklavichet al., 1991; Potter et al., 1997; Jackson and Jones, 1999).

4.3. Tidal changes

In the seagrass site, the tides affect the different species ina similar way, with the most important species having higherabundances at low tide than at high tide. Tidal changes infish abundance could be due to the movement of fish from sub-tidal channels throughout the inundated intertidal areas withthe flooding tide. This pattern in fish abundance along the tidalcycle has been reported for tidal flats and marshes located inestuarine areas, coastal lagoons and other shallow coastalareas (Kneib and Wagner, 1994; Rozas, 1995; Morrisonet al., 2002; Greenwood and Hill, 2003). Other studies also re-ported that intertidal habitats are temporarily used by fish andother nekton during the period of submersion (e.g. Rey et al.,1990; Minello and Zimmerman, 1992; Rountree and Able,1992; Rozas and Reed, 1993; Thomas and Connolly, 2001).The transversal movements to the intertidal habitats duringthe flood tide are generally explained by the fact that they al-low fish to access the productive intertidal foraging grounds,only available during the periods of tidal inundation (Weisbergand Lotrich, 1982; Boesch and Turner, 1984; Rozas andOdum, 1988; Cattrijsse et al., 1994; Edgar and Shaw,1995b). Several studies show that some fish species forageduring their stay in intertidal areas (Kelley, 1988; Raffaelliet al., 1990; Laffaille et al., 1999), supporting the hypothesisthat intertidal areas function as foraging grounds for fish. Cal-vario (unpublished data) found that number of species, densityand biomass of benthic macroinvertebrate in the intertidalareas of the Ria Formosa were higher than in subtidal areas,

Table 6

List of the abundance, and frequency of occurrence ( f %)

Species Diel

Low Day Night

n(%) n n(%) n n(%) n n(%)

Arnoglossu 1.69 6 0.90 16 1.97 3 0.48

Bothus pod 1.04 22 3.30 16 1.97 14 2.24

Callionymu

maculatu

1.69 24 3.60 24 2.96 13 2.08

Diplodus vu 1.04 39 5.86 15 1.85 32 5.13

Echiichthys 1.82 45 6.76 49 6.04 10 1.60

Gobius nige 13.65 40 6.01 48 5.92 97 15.54

Gobius pag 2.60 5 0.75 17 2.10 8 1.28

Hippocamp

guttulatu

0.91 8 1.20 7 0.86 8 1.28

Mullus surm 0.26 9 1.35 2 0.12 10 1.60

Pomatoschi

microps

5.85 8 1.20 17 2.10 36 5.77

Pomatoschi

minutus

55.4 372 55.86 477 58.82 321 51.44

Pomatoschi

pictus

1.04 8 1.20 7 0.86 9 1.44

Solea lasca 1.69 9 1.35 15 1.85 7 1.12

Spondylioso

cantharu

0.78 9 1.35 6 0.74 9 1.44

Symphodus

bailloni

4.03 13 1.95 40 4.93 4 0.64

Symphodus

cinereus

1.17 5 0.75 10 1.23 4 0.64

Total no. of

individua

e 666 e 812 e 623 e

Total no. of

species

e 38 e 35 e 35 e

Resident 87.5 470 70.6 652 80.3 491 78.8

Marine juve

migrants

4.7 78 11.7 49 6.0 65 10.4

Marine

adventiti

7.8 118 17.7 111 13.7 67 10.8

47

0J.

Ribeiro

etal.

/E

stuarine,C

oastaland

ShelfScience

67(2006)

461e474

most abundant species caught in the sand site, by season, tidal stage and diel period, presenting their absolute (n) and relative (n%)

Season Tide

Spring Summer Autumn Winter High

n n(%) f(%) n n(%) f(%) n n(%) f(%) n n(%) f(%) n

s thori 11 1.51 58.3 6 2.11 25.0 2 1.05 8.3 0 e 0.0 13

as 2 0.28 8.3 25 8.80 66.7 3 1.57 25.0 0 e 0.0 8

s

s

10 1.38 50.0 12 4.23 50.0 8 4.19 41.7 7 3.00 33.3 13

lgaris 35 4.81 58.3 12 4.23 50.0 0 e 0.0 0 e 0 8

vipera 14 1.93 58.3 10 3.52 41.7 2 1.05 16.7 33 14.16 58.3 14

r 112 15.41 100.0 15 5.28 50.0 3 1.57 16.7 15 6.44 83.3 105

anellus 2 0.28 8.3 22 7.75 58.3 1 0.52 8.3 0 e 0.0 20

us

s

4 0.55 16.7 8 2.82 50.0 1 0.52 8.3 2 0.86 16.7 7

uletus 0 e 0.0 11 3.87 41.7 0 e 0.0 0 e 0.0 2

stus 52 7.15 75.0 0 e 0.0 1 0.52 8.3 0 e 0.0 45

stus 440 60.52 100.0 67 23.59 100.0 143 74.87 100.0 148 63.52 100.0 426

stus 5 0.69 16.7 0 e 0.0 0 e 0.0 11 4.72 41.7 8

ris 9 1.24 58.3 9 3.17 58.3 4 2.09 25.0 0 e 0.0 13

ma

s

1 0.14 8.3 11 3.87 41.7 2 1.05 16.7 1 0.43 8.3 6

3 0.41 25.0 32 11.27 50.0 7 3.66 33.3 2 0.86 16.7 31

4 0.55 25.0 3 1.06 16.7 5 2.62 25.0 2 0.86 16.7 9

ls

727 e e 285 e e 190 e e 233 e e 769

28 e e 31 e e 20 e e 18 e e 33

634 87.2 e 172 60.4 e 164 86.3 e 173 74.3 e 673

nile 46 6.3 e 56 19.6 e 9 4.7 e 3 1.2 e 36

ous

47 6.5 e 57 20.0 e 17 9.0 e 57 24.5 e 60

471J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

suggesting that fish feeding activities may play an importantrole in recycling the intertidal productivity within the coastallagoon. These foraging activities allow energy transfer be-tween intertidal benthic invertebrate assemblages and fish con-sumers (Kneib, 1997). Protection from predation is also anargument used to explain these tidal-related movements(Minello and Zimmerman, 1983; Boesch and Turner, 1984;Kneib, 1987; Rozas and Odum, 1988). Although the immersedarea of the Ria Formosa coastal lagoon increases by up to 5times during the high tide (�Aguas, 1986), the fish abundancein the subtidal channel only declines by 1.5e2 times, suggest-ing that most of the fish species probably only use the adjacentintertidal areas close to the edge of the subtidal channels(Kneib and Wagner, 1994).

In contrast, on sand there was no evidence of tidal changes infish assemblage structure, number of species or in general fishabundance that supports the possibility of dispersal movements

Sand Seagrass

Per

cent

age

of f

ish

num

ber

Per

cent

age

of f

ish

num

ber

EB

B

Verticaldistribution

guilds

(a)

Sand Seagrass

R

MA

MJM

Ecologicalguilds

(b)

100

80

60

40

20

0

100

80

60

40

20

0

Fig. 6. Percentage of fish number in each guild, over sand and seagrass sites:

(a) vertical distribution guilds (B e benthic; EB e epibenthic); and (b) ecolog-

ical guilds (R e resident; MA e marine adventitious; MJM e marine juvenile

migrant).

to the immersed intertidal areas. The habitat and benthic macro-invertebrate homogeneity of sand subtidal habitat and the sur-rounding intertidal areas (Calvario, unpublished data), alongwith the low fish density, seems not to offer any additional advan-tage in terms of dispersal foraging and predation avoidance. How-ever, the fish functional organization suffers some tidal changes,with a higher proportion of marine species (marine adventitiousand marine juvenile migrants) observed at low tide, and a higherproportion of resident species at high tide. Since the high tidesampling took place at the beginning of ebb tide, and the lowtide sampling was at the beginning of the flood tide, it seems plau-sible that the direction of a relatively intense tidal current, due tothe close presence of an inlet, was the main factor explaining theobserved tidal changes in sand fish community organization. Thelongitudinal movement of the water from the coastal areas into thelagoon during the flood tide facilitates the immigration of fishfrom the surrounding coastal areas, followed by their emigrationon the ebb tide (Greenwood and Hill, 2003). On the other hand,the flood tide tends to push the resident species towards the inte-rior of the lagoon, while the ebb tends to drag them to more pe-ripherical areas.

4.4. Diel changes

The number of species and fish abundance, as well as thefish assemblage structure associated with both sites did not dif-fer substantially between day and night. These results are con-sistent with those found by several authors in estuarinevegetated habitats in Australia (Gray et al., 1996; Grayet al., 1998). However, others studies reported diel changesin certain species abundance, and/or greater overall fish abun-dances and number of species at night in seagrass (e.g. Adams,1976; Horn, 1979; Robblee and Zieman, 1984; Mattila et al.,1999; Guest et al., 2003), and over sand in various shallow ma-rine and estuarine environments (Gray et al., 1998; Methvenet al., 2001; Travers and Potter, 2002; Guest et al., 2003; Pes-sanha and Araujo, 2003; Pessanha et al., 2003).

The absence of significant diel changes in the fish assem-blage and the observation of fish movements associated withtidal rhythm, suggest that the tide pulse is the major effect re-sponsible for daily fluctuations in the fish community of theRia Formosa coastal lagoon. This is an expected result sincethis ecosystem is affected by a semi-diurnal mesotidal regime,where approximately 70e80% of its surface represents a very

Fig. 7. Non-metric multi-dimensional scaling ordination plots of the presence/absence data (a) and abundance data (b). Stress values are given in the top right

corner of each plot. (:) e Sand samples and (P) e seagrass samples.

472 J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

Fig. 8. Non-metric multi-dimensional scaling ordination plots of the presence/absence data for seagrass (a) and sand (b); and of the abundance data for seagrass

(c) and sand samples (d), both classified by season. Stress values are given in the top right corner of each plot. (-) spring; ( ) summer; (P) autumn; (:) winter.

productive intertidal area that functions as an important feed-ing ground. The imposition of a diel rhythm over a tidal onewould mean that many foraging opportunities would be lost.In the summer, when the days are longer, this would be partic-ularly critical for species that only used the intertidal areasduring the night.

5. Conclusions

The fish assemblages associated with seagrass and sandsites were substantially different, both in terms of speciescomposition and their relative abundance. Structurally morecomplex, the seagrass supports a much more abundant and di-verse fish community, and was an important juvenile fish hab-itat for more species than sand. These site-related differenceswere greater than any seasonal, tidal or diel variations in fishassemblages, suggesting that the habitat type (e.g. degree ofstructural complexity, sediment type, topography) mighthave had an important role in the definition of the fish

Table 7

Results of ANOSIM tests, with overall and the pairwise for differences

amongst the seasonal combinations in both sites

Seasons

comparisons

Sand Seagrass

R-statistic P-value R-statistic P-value

Overall 0.531 P< 0.001 0.831 P< 0.001

Winteresummer 0.637 P< 0.001 0.953 P< 0.001

Winterespring 0.756 P< 0.001 0.923 P< 0.001

Wintereautumn 0.209 P¼ 0.090 0.440 P< 0.001

Autumnesummer 0.492 P< 0.001 0.615 P< 0.001

Autumnespring 0.747 P< 0.001 0.940 P< 0.001

Summerespring 0.953 P< 0.001 0.957 P< 0.001

assemblage structure in the Ria Formosa coastal lagoon. Inspite of these differences, both sites showed similar seasonalchanges in the fish assemblage structure. These seasonal pat-terns in the fish assemblages were related to ontogenic shiftsin the presence and abundance of marine juvenile migrant spe-cies, mainly due to reproduction timing and large-scale move-ments, to and from the adjacent marine environment. On theother hand, small-scale temporal changes in fish assemblagesappear related to tidal and diel cycles, which may be relatedto small-scale movements associated with feeding, shelter us-age and predator avoidance. In the Ria Formosa these small-scale movements are mainly associated with tidal rhythmrather than diel rhythm.

Acknowledgements

This research was funded in part by the European Commis-sion (EC-DG XIV/C1; Study project no. 99/061) and was alsosponsored by the Fundac~ao para Ciencia e Tecnologia througha PhD grant awarded to the first author (grant reference:SFRH/BD/6812/2001, Portuguese Ministry of Science andTechnology). The authors would like to thank to all the col-leagues and volunteers who assisted with the field samplingand laboratory work and also to the fishing boat skipper Isi-doro Costa.

References

Adams, S.M., 1976. The ecology of eelgrass, Zostera marina (L.), fish com-

munities, I. Structural analysis. Journal of Experimental Marine Biology

and Ecology 22, 269e291.�Aguas, M.P.N., 1986. Simulac~ao da circulac~ao hidrodinanica na Ria Formosa

(Hydrodynamic circulation of Ria Formosa). In: Gomes Gerreiro, M. (Ed.),

473J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

Proceedings of the Os Sistemas Lagunares do Algarve. Universidade do

Algarve, Faro, Portugal, pp. 78e90.

Bell, J.D., Pollard, D.A., 1989. Ecology of fish assemblages and fisheries as-

sociated with seagrasses. In: Larkum, A.W.D., McComb, A.J.,

Shepard, S.A. (Eds.), Biology of Seagrasses: A Treatise on the Biology

of Seagrasses with Special Reference to the Australian Region. Elsevier,

Amsterdam, pp. 565e609.

Bell, J.D., Westoby, M., Steffe, A.S., 1987. Fish larvae settling in seagrass: do

they discriminate between beds of different leaf density? Journal of Exper-

imental Marine Biology and Ecology 111, 133e144.

Boesch, D.F., Turner, R.E., 1984. Dependence of fishery species on salt

marshes: the role of food and refuge. Estuaries 7, 460e468.

Burchmore, J.J., Pollard, D.A., Bell, J.D., 1984. Community structure and

trophic relationships of the fish fauna of an estuarine Posidonia australis

seagrass habitat in Port Hacking, New South Wales. Aquatic Botany 18,

71e87.

Cattrijsse, A., Makwaia, E.S., Dankwa, H.R., Hamerlynck, O.,

Hemminga, M.A., 1994. Nekton communities of an intertidal creek of

a European estuarine brackish marsh. Marine Ecology Progress Series

109, 195e208.

Clarke, K.R., 1993. Non-parametric multivariate analyses of changes in com-

munity structure. Australian Journal of Ecology 18, 117e143.

Clarke, K.R., Green, R.H., 1988. Statistical design and analysis for a ‘‘biolog-

ical effects’’ study. Marine Ecology Progress Series 46, 213e226.

Clarke, K.R., Warwick, R.M., 2001. Change in Marine Communities e An

Approach to Statistical Analysis and Interpretation. Primer-E Ltd.,

Plymouth.

Connolly, R.M., 1994a. A comparison of fish assemblages from seagrass and

unvegetated areas of a southern Australian estuary. Australian Journal of

Marine and Freshwater Research 45, 1033e1044.

Connolly, R.M., 1994b. The role of seagrass as preferred habitat for juvenile

Sillaginodes punctata (Cuv. & Val.) (Sillaginidae, Pisces): habitat selection

or feeding? Journal of Experimental Marine Biology and Ecology 180,

39e47.

Edgar, G.J., Shaw, C., 1995a. The production and trophic ecology of shallow-

water fish assemblages in Southern Australia. I. Species richness, size

structure and production of fishes in Western Port, Victoria. Journal of

Experimental Marine Biology and Ecology 194, 53e81.

Edgar, G.J., Shaw, C., 1995b. The production and trophic ecology of shallow-

water fish assemblages in Southern Australia. II. Diets of fishes and trophic

relationships between fishes and benthos at Western Port, Victoria. Journal

of Experimental Marine Biology and Ecology 194, 83e106.

Elliott, M., Dewailly, F., 1995. The structure and components of European es-

tuaries assemblages. Netherlands Journal of Aquatic Ecology 29,

397e417.

Falc~ao, M., Pissarra, J., Cavaco, M., 1992. Caracteristicas quımico biologicas

da Ria Formosa: Analise de um ciclo anual. Boletim do Instituto Nacional

de Investigac~ao das Pescas 16, 1e14.

Field, J.G., Clarke, K.R., Warwick, R.M., 1982. A practical strategy for ana-

lysing multispecies distribution patterns. Marine Ecology Progress Series

8, 37e52.

Goncalves, J.M.S., Erzini, K., 2000a. The reproductive biology of Spondylio-

soma cantharus (L.) from the SW coast of Portugal. Scientia Marina 64,

403e411.

Goncalves, J.M.S., Erzini, K., 2000b. The reproductive biology of the two-

banded sea bream (Diplodus vulgaris) from the southwest coast of Portu-

gal. Journal of Applied Ichthyology 16, 110e116.

Gray, C.A., Chick, R.C., McElligott, D.J., 1996. Intra- and inter-estuary differ-

ences in assemblages of fishes associated with shallow seagrass and bare

sand. Marine and Freshwater Research 47, 723e735.

Gray, C.A., McElligott, D.J., Chick, R.C., 1998. Diel changes in assemblages

of fishes associated with shallow seagrass and bare sand. Estuarine,

Coastal and Shelf Science 46, 849e859.

Greenwood, M.F.D., Hill, A.S., 2003. Temporal, spatial and tidal influences on

benthic and demersal fish abundance in the Forth estuary. Estuarine,

Coastal and Shelf Science 58, 211e225.

Griffiths, S.P., 2001. Diel variation in the seagrass ichthyofaunas of three in-

termittently open estuaries in south-eastern Australia: implications to

improving fish diversity assessments. Fisheries Management and Ecology

8, 123e140.

Guest, M.A., Connolly, R.M., Loneragan, N.R., 2003. Seine nets and beam

trawls compared by day and night for sampling fish and crustaceans in

shallow seagrass habitat. Fisheries Research 64, 185e196.

Guidetti, P., 2000. Differences among fish assemblages associated with near-

shore Posidonia oceanica seagrass beds, rocky-algal reefs and unvegetated

sand habitats in the Adriatic Sea. Estuarine, Coastal and Shelf Science 50,

515e529.

Guidetti, P., Bessotti, S., 2000. Fish fauna of a mixed meadow composed by

the seagrasses Cymodocea nodosa and Zostera noltii in the western Med-

iterranean. Oceanologica Acta 23, 759e770.

Heck Jr., K.L., Able, K.W., Fahay, M.P., Roman, C.T., 1989. Fishes and

decapod crustaceans of Cape Cod eelgrass meadows: species composition

seasonal abundance and comparison with unvegetated substrates. Estuaries

12, 59e65.

Horn, M.H., 1979. Diel and seasonal variation in abundance and diversity of

shallow-water fish populations in Morro Bay, California. Fishery Bulletin

78, 759e770.

Howard, R.K., Koehn, J.D., 1985. Population dynamics and feeding ecology of

pipefish (Syngnathidae) associated with eelgrass beds of Western Port,

Victoria. Australian Journal of Marine and Freshwater Research 36,

361e370.

Hyndes, G.A., Platell, M.E., Potter, I.C., Lenanton, R.C.J., 1999. Does the

composition of demersal fish assemblages in temperate coastal waters

change with depth and undergo consistent seasonal changes? Marine Biol-

ogy 134, 335e352.

Jackson, G., Jones, G.K., 1999. Spatial and temporal variation in nearshore

fish and macroinvertebrate assemblage from a temperate Australian

estuary over a decade. Marine Ecology Progress Series 182, 253e268.

Jenkins, G.P., May, H.M.A., Wheatley, M.J., Holloway, M.G., 1997. Comparison

of fish assemblages associated with seagrass and adjacent unvegetated hab-

itats of Port Philip Bay and Corner Inlet, Victoria, Australia, with emphasis

on commercial species. Estuarine, Coastal and Shelf Science 44, 569e588.

Kelley, D.F., 1988. The importance of estuaries to sea bass (Dicentrarchus lab-

rax). Journal of Fish Biology 33 (Suppl. A), 25e33.

Kneib, R.T., 1987. Predation risk and use of intertidal habitats by young fishes

and shrimp. Ecology 68, 379e386.

Kneib, R.T., 1997. The role of tidal marshes in the ecology of estuarine

nekton. Oceanography and Marine Biology: An Annual Review 35,

163e220.

Kneib, R.T., Wagner, S.L., 1994. Nekton use of vegetated habitats at different

stages of tidal inundation. Marine Ecology Progress Series 106, 227e238.

Laffaille, P., Feunteun, E., Lefeuvre, J.-C., 1999. Competition alimentaire en-

tre deux especes de gobie, Pomatoschistus lozanoi (de Buen) et P. minutus(Pallas), dans un marais sale macrotidal. Comptes Rendus de l’Academie

des Sciences Paris, Life Sciences 322, 897e906.

Lazzari, M.A., Tupper, B., 2002. Importance of shallow water habitats for de-

mersal fishes and decapod crustaceans in Penobscot Bay, Maine. Environ-

mental Biology of Fishes 63, 57e66.

Lima, C., Vale, C., 1977. Ria de Faro-Olh~ao: aspectos fısicos, quımicos e bac-

teriologicos. Secretaria de Estados das Pescas, Direcc~ao Geral de Inves-

tigac~ao e Protecc~ao dos Recursos Vivos e Ambiente Aquatico, Lisboa,

21 pp.

Livingston, R.J., Kobylinsky, G.J., Lewis III, F.G., Sheridan, P.F., 1976. Long-

term fluctuations of epibenthic fish and invertebrate populations in Apala-

chicola Bay, Florida. Fishery Bulletin 74, 311e321.

Mattila, J., Chaplin, G., Eilers, M.R., Heck Jr., K.L., O’Neal, J.P.,

Valentine, J.F., 1999. Spatial and diurnal distribution of invertebrate and

fish fauna of a Zostera marina bed nearby unvegetated sediments in Dam-

ariscotta River Maine (USA). Journal of Sea Research 41, 321e332.

Methven, D.A., Haedrich, R.L., Rose, G.A., 2001. The fish assemblage of

a Newfoundland estuary: diel monthly and annual variation. Estuarine,

Coastal and Shelf Science 52, 669e687.

Minello, T.J., Zimmerman, R.J., 1983. Fish predation on juvenile brown

shrimp, Penaeus aztecus (Ives): the effect of simulated Spartina structure

on predation rates. Journal of Experimental Marine Biology and Ecology

72, 211e231.

474 J. Ribeiro et al. / Estuarine, Coastal and Shelf Science 67 (2006) 461e474

Minello, T.J., Zimmerman, R.J., 1992. Utilization of natural and transplanted

Texas salt marshes by fish and decapod crustaceans. Marine Ecology Prog-

ress Series 90, 273e285.

Monteiro, C., Lam Hoai, T., Lasserre, G., 1987. Chronological assemblages of

fishes in 2 stations of the Ria Formosa lagoon (Portugal). Oceanologica

Acta 10, 359e371.

Monteiro, C., Lasserre, G., Lam Hoai, T., 1990. Spatial organization of the ich-

thyological community in the Ria Formosa lagoon (Portugal). Oceanolog-

ica Acta 13, 79e96.

Morrison, M.A., Francis, M.P., Hartill, B.W., Parkinson, D.M., 2002. Diurnal

and tidal variation in the abundance of fish fauna of a temperate tidal mud-

flat. Estuarine, Coastal and Shelf Science 54, 793e807.

Nash, R.D.M., Gibson, R.N., 1982. Seasonal fluctuations and compositions of

two populations of small demersal fishes on the west coast of Scotland. Es-

tuarine, Coastal and Shelf Science 15, 485e495.

Orth, R.J., Heck Jr., K.L., van Montfrans, J., 1984. Faunal communities in sea-

grass beds: a review of the influence of plant structure and prey character-

istics on predatoreprey relationships. Estuaries 7, 339e350.

Pessanha, A.L.M., Araujo, F.G., 2003. Spatial, temporal and diel variations of

fish assemblages at two sandy beaches in the Sepetiba Bay, Rio de Janeiro,

Brazil. Estuarine, Coastal and Shelf Science 57, 817e828.

Pessanha, A.L.M., Araujo, F.G., Azevedo, M.C.C., Gomes, I.D., 2003. Diel

and seasonal changes in the distribution of fish on a southeast Brazil sandy

beach. Marine Biology 143, 1047e1055.

Pihl, L., Rosenberg, R., 1982. Production, abundance and biomass of mobile

epibenthic marine fauna in shallow waters, western Sweden. Journal of Ex-

perimental Marine Biology and Ecology 72, 211e231.

Pollard, D.A., 1984. A reviewed of ecological studies on seagrass-fish commu-

nities with particular reference to recent studies in Australia. Aquatic Bot-

any 18, 3e42.

Potter, I.C., Beckley, L.E., Whitfield, A.K., Lenanton, R.C.J., 1990. Compar-

isons between the roles played by estuaries in the life cycles of fishes in

temperate Western Australia and South Africa. Environmental Biology

of Fishes 28, 143e178.

Potter, I.C., Claridge, P.N., Hyndes, G.A., Clarke, K.R., 1997. Seasonal, an-

nual and regional variations in ichthyofaunal composition in the Inner Sev-

ern estuary and Inner Bristol Channel. Journal of the Marine Biology

Association, UK 77, 507e526.

Potter, I.C., Claridge, P.N., Warwick, R.M., 1986. Consistency of seasonal

changes in an estuarine fish assemblage. Marine Ecology Progress Series

32, 217e228.

Raffaelli, D., Richner, H., Summers, R., Northcott, S., 1990. Tidal migrations

in the flounder (Platichthys flesus). Marine Behaviour and Physiology 16,

249e260.

Rey, J.R., Shaffer, J., Tremain, D., Crossman, R.A., Kain, T., 1990. Effects of

re-establishing tidal connections in two impounded subtropical marshes on

fishes and physical conditions. Wetlands 10, 27e46.

Robblee, M.B., Zieman, J.C., 1984. Diel variation in the fish fauna of a tropical

seagrass feeding ground. Bulletin of Marine Science 37, 335e345.

Rountree, R.A., Able, K.W., 1992. Foraging habits, growth and temporal

patterns of salt marsh creek habitat use by juvenile summer flounder in

New Jersey. Transactions of the American Fisheries Society 121, 765e

776.

Rountree, R.A., Able, K.W., 1993. Diel variation in decapod crustacean and

fish assemblages in New Jersey polyhaline marsh creeks. Estuarine,

Coastal and Shelf Science 37, 181e201.

Rozas, L.P., 1995. Hydroperiod and its influence on nekton use of salt marsh:

a pulsing ecosystem. Estuaries 18, 579e590.

Rozas, L.P., Odum, W.E., 1988. Occupation of submerged aquatic vegeta-

tion by fishes: testing the roles of food and refuge. Oecologia 77,

101e106.

Rozas, L.P., Reed, D.J., 1993. Nekton use of marsh-surface habitats in Loui-

siana (USA) deltaic salt marsh undergoing submergence. Marine Ecology

Progress Series 96, 147e157.

SAS Institute, 1988. SAS/STAT Users Guide, Release 6.03 Edition. SAS Insti-

tute Inc., Cary, NC, 1028 pp.

Sogard, S.M., Able, K.W., 1991. A comparison of eelgrass, sea lettuce macro-

algae, and marsh creeks as habitats for epibenthic fishes and decapods. Es-

tuarine, Coastal and Shelf Science 33, 501e519.

Sogard, S.M., Powell, G.V.N., Holmquist, J.G., 1989. Utilization by fishes of

shallow, seagrass-covered banks in Florida Bay: 2. Diel and tidal patterns.

Environmental Biology of Fishes 24, 53e65.

Sprung, M., 1994. Macrobenthic secondary production in the intertidal zone of

the Ria Formosa e a lagoon in southern Portugal. Estuarine, Coastal and

Shelf Science 38, 539e558.

Stoner, A.W., 1983. Distribution of fishes in seagrass meadows: role of

macrophyte biomass and species composition. Fishery Bulletin 81, 837e

846.

Thiel, R., Potter, I.C., 2001. The ichthyofaunal composition of the Elbe estu-

ary: an analysis in space and time. Marine Biology 138, 603e616.

Thomas, B.E., Connolly, R.M., 2001. Fish use of subtropical saltmarshes in

Queensland, Australia: relationships with vegetation, water depth and

distance onto the marsh. Marine Ecology Progress Series 209, 275e288.

Travers, M.J., Potter, I.C., 2002. Factors influencing the characteristics of fish

assemblages in a large subtropical marine embayment. Journal of Fish Bi-

ology 61, 764e784.

Underwood, A.J., 1981. Techniques of analysis of variance in experimental

marine biology and ecology. Oceanography and Marine Biology: an An-

nual Review 19, 513e605.

Weinstein, M.P., 1982. A need for experimental work in estuarine fisheries

ecology. Northeast Gulf Science 5, 59e64.

Weisberg, S.B., Lotrich, V.A., 1982. The importance of an infrequently

flooded intertidal marsh surface as an energy source for mummichog

Fundulus heteroclitus: an experimental approach. Marine Biology 66,

307e310.

Yoklavich, M.M., Cailliet, G.M., Barry, G.P., Ambrose, D.A., Antrim, B.S.,

1991. Temporal and spatial patterns in abundance and diversity of fish

assemblages in Elkhorn Slough, California. Estuaries 14, 465e480.