Fish assemblages on fringe coral reefs of the northern coast of Cuba near Havana Harbor

Transcription

1 Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] Fish assemblages on fringe coral reefs of the northern coast of Cuba near Havana Harbor Consuelo Aguilar, a Gaspar Gonza lez-sanso n, a Kelly R. Munkittrick, b and Deborah L. MacLatchy b, a Centro de Investigaciones Marinas, Universidad de La Habana, Havana, Cuba b Department of Biology and Canadian Rivers Institute, University of New Brunswick, Saint John, New Brunswick, Canada E2L 4L5 Received 4 October 2002; received in revised form 10 June 2003; accepted 19 June 2003 Abstract It is difficult to separate effects on fish community assemblages due to anthropogenic stressors from natural factors. We examined small-bodied-fish communities along the northern coast of Cuba near Havana Harbor during the dry (February/March of 2000) and wet (June 2000) seasons. Over 35,000 individual fish were visually counted at 15 sites along the coast in three areas located 0 2.4, , and km from the entrance to Havana Harbor. Fish communities in four substrate biotopes did not vary significantly between wet and dry seasons, but did vary with water depth. Proximity to Havana Harbor was the second most important factor affecting fish assemblages, and sites closest to the harbor had reduced populations of the bluehead wrasse (Thalassoma bifasciatum) and an increased abundance of slippery dick (Halichoeres bivittatus). More studies are required at the population and individual levels to link stressors (e.g., contaminants, siltation) directly to observed effects. r 2003 Elsevier Inc. All rights reserved. Keywords: Fish assemblages; Coral reefs; Contamination; Cuba 1. Introduction Coastal fish assemblages are under increasing pressure from human activities around the world due to a variety of stressors, including overfishing, land-based pollution, and juvenile habitat destruction (FAO, 1993). Little information is available on the impacts of human activities on coastal fish populations of Cuba, especially around large population centers. A narrow shelf with rocky and sandy bottoms allows the growth of fringing coral reefs along the northern coast of Cuba near Havana Harbor, and the fish communities of these fringing reefs have been poorly studied. The main invertebrate groups in fringing reef areas are sponges (Niphates sp., Holopsama welggi, Callispongia vaginalis), gorgonians (Plexaura flexuosa, Eunicea spp.), and scleractinian corals (Siderastrea radians, Porites astreoides), and these communities show a decreased abundance and gradient of responses near Havana Harbor (Guardia and Gonza lez-sanso n, 2000). Corresponding author address: maclatch@unbsj.ca (D.L. MacLatchy). There are a variety of pressures on fish communities near Havana, and selective overfishing occurs for medium- and large-bodied fish of the families Lutjanidae (snappers), Serranidae (groupers), Carangidae (jacks), and Sphyraenidae (barracudas). This selective overfishing alters species composition (Aguilar and Gonza lez-sanso n, 2000) and increases the dominance of smaller species (Baisre, 2000), similar to changes seen on Jamaican coral reefs (Hughes, 1994) and on coral reefs of the Indo-Pacific Region (Harmelin-Vivien, 1992). The resulting fish assemblages have a high level of diversity and are dominated by small-bodied species of the families Pomacentridae (damselfish), Labridae (wrasses), and Acanthuridae (surgeonfish). Havana Harbor may be the most polluted coastal marine habitat in Cuba, with untreated discharges of crude oil and related substances, sewage, industrial effluents, and heavy metals (Areces and Toledo, 1985; Gonza lez, 1991; Beltra n et al., 1998). The Almendares River is the main waterway flowing through Havana and is also heavily contaminated with wastes of domestic and industrial origin (e.g., from a brewery, a /$ - see front matter r 2003 Elsevier Inc. All rights reserved. doi: /s (03)

2 2 ARTICLE IN PRESS C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] Fig. 1. Location of the 15 sites in areas 1, 2 and 3 along the north coast of Cuba used in this study. Arrow thickness is a qualitative representation of the relative strength of the currents. For identification of the site locations, see Table 1. paper mill, juice plants, food processing including chicken processing plants, and cement plants). The domestic waste from Havana (population 2.2 million) is discharged into the harbor or east of the Harbor entrance without treatment. Previous studies have shown a decreased density of fishes at the mouth of Havana harbor and at the mouth of the Almendares River (Aguilar and Gonza lez-sanso n, 2000; Gonza lez-sanso n and Aguilar, 2000), but it is not known how far from the harbor these impacts extend. This study was designed to confirm earlier indications of impacts on fish near Havana Harbor. As a basic assumption, it was recognized that fishing activities have removed a groupof large-bodied species from the original fish assemblages. The remaining small-bodied species form modified assemblages that are expected to be homogeneous along the coast. There are changes in abundance and species composition of assemblages along a depth gradient for invertebrates (Guardia and Gonza lez-sanso n, 2000) and previous studies have suggested that similar gradients exist for fish communities (Fausch et al., 1990; Zapata and Morales, 1996), including near Havana Harbor (Aguilar and Gonza lez- Sanso n, 2000; Gonza lez-sanso n and Aguilar, 2000). The fish assemblages can also change seasonally (Luckhurst and Luckhurst, 1978; Aliaume et al., 1990; McGehee, 1994). This is the first detailed study along the northern coast of Cuba that focuses on three key factors affecting fish community structure: season, habitat zone, and proximity to potential contaminant input (Havana Harbor). 2. Materials and methods 2.1. Study area and sampling The study area was a section of the coastal waters off Havana City (Fig. 1). In this region the upper subtidal zone is a submarine terrace with a gentle slope until an approximately 8-m depth. At that point, the slope increases significantly until m, where a sandy plain is present. The terrace is about 300 m in width and can be divided into four distinct biotopes, which exist as bands running parallel to the coastline (Fig. 2). These biotopes are: * Echinometra (Ec): The most shoreward and shallowest biotope. Wave strength is very high. The main feature of this zone is the huge population of the sea urchin, Echinometra lucunter, a species that excavates the rocky bottom for shelter. This zone has a relatively low structural complexity. * Rocky plain (Pl): The main feature of this biotope is the smooth rocky bottom that is covered by seaweed seasonally. It presents the lowest structural complexity. Wave strength is high. In some places, however, small, elongated trenches (1 m deepand 5 m long) perpendicular to the shoreline can be found. In these trenches, high concentrations of juvenile fishes can be occasionally observed. * Terrace edge (Te): This is the portion just before the point where the bottom slope becomes steeper. A mixed community of scleractinian corals, gorgonians, and sponges of variable abundance dominates the

3 C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] 3 Fig. 2. Typical bottom profile in the study areas showing the habitat zones Echinometra (Ec), Rocky plain (Pl), Terrace edge (Te) and Terrace base (Tb). bottom. In some places this community is almost completely absent and substituted by seaweed growth. This biotope has the highest structural complexity. * Terrace base (Tb): This is the deepest biotope where the rocky wall of the terrace meets the sandy plain. A mixed community of scleractinean corals, gorgoneans, and sponges of variable abundance also dominates. In some places this community is almost completely absent and substituted by seaweed growth. It has a high structural complexity. The main influences from the Almendares River and Havana Harbor are on the eastern portion of the study area as the main oceanic and in-shore currents are easterly (Fig. 1). The western portion, therefore, was considered in the present study to be relatively clean and an acceptable reference region. The coastal waters were divided into three areas according to proximity to Havana Harbor (area 1, within km; area 2, within km; and area 3, within km) (Fig. 1). In each area, a set of representative sites was defined. Fifteen sites were selected along the coast (Fig. 1). The site positions were referenced using easily identifiable points on the shoreline as described in Table 1. The fish assemblages at each site were visually counted during 2000 in two different periods, from (a) the 15th of February to the 21st of March (middle of the dry season) and from (b) the 6th to the 30th of June (beginning of the rainy season, after the massive recruitment peak). Four stationary censuses (based on Bohnsack and Bannerot, 1986) were made at each biotope at all sites for a total of 16 censuses per site during each sampling period Community composition A total of 480 censuses were made during this study. No attempt was made to estimate the size of the fishes and only those which were very small (less than 3 cm body length for Scaridae, Acanthurus, Haemulon, and Clepticus parrae; less than 8 cm body length for Caranx ruber) and/or had a distinct juvenile color phase were recorded as juveniles. The number of individuals of each species was recorded. This value was used to calculate the percentage numeric representation of each species included in the analyses. Early juveniles of the genera Acanthurus and Haemulon, as well as some species of the family Scaridae, were common in some sites. Although most fish were classified to species level, it was not possible to classify these fish to species level in situ, due to either similarities in appearance or time constraints. In order to make a comparative analysis of fish assemblage composition, the 10 most abundant species for each biotope in each sampling period were determined for the three areas defined according to proximity to Havana Harbor. In the analyses of these data sets there are three main guiding questions: 1. Did the dominant species change between seasons? 2. Did habitat preferences change? 3. Was there evidence of a gradient according to distance from Havana Harbor? 2.3. Diversity indices Sites were considered as replicates inside each of the three areas and data on numbers of individuals by species obtained at the sites were pooled for each combination of biotope, sampling period, and level of pollution.

4 4 ARTICLE IN PRESS C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] Table 1 Location of each of the sites on the northern coast of Cuba selected for this study Area Site (number) Location and Global Positioning System coordinates Area 1 (0 2.4 km from Havana Harbor, San La zaro s cove) BR (1) Red buoy at the entrance of the harbor N W HD (2) Deauville Hotel N W DM (3) A point halfway between HD and PM N W PM (4) Maceo s Park N W 25 (5) 25th Street (Vedado) N W HN (6) National Hotel N W Area 2 ( km from Havana Harbor) Mr (7) Marti s Park N W Pa (8) Paseo Street N W 12 (9) 12th Street (Vedado) N W Pu (10) La Puntilla N W Area 3 ( km from Havana Harbor) 16 (11) 16th Street (Miramar) N W 30 (12) 30th Street (Miramar) N W 42 (13) 42nd Street (Miramar) N W 70 (14) 70th Street (Miramar) N W 84 (15) 84th Street (Miramar) N W See Fig. 1 for locations of sites along coast. Total diversity and its components were estimated using the following indices: H 0 ¼ SP i ln P i ; Shannon index, for total diversity (P i being the proportional representation of individuals of each species). S is the total number of species, for species richness. J 0 ¼ H 0 =ln S; Pielou index, for evenness Cluster analysis and multidimensional scaling The numerical classification of the pooled samples was obtained using agglomerative cluster analysis employing the unweighted pair-group method using arithmetic average (UPGMA) clustering algorithm (Sneath and Sokal, 1973; Boesch, 1977). The ordination of the same data was made using the nonmetric multidimensional scaling (MDS) method (Kruskal, 1964; Clarke and Warwick, 1994). The dissimilarity matrix was prepared using the coefficient of percentage dissimilarity, CPD jk ¼ 1 minðp ij ; P ik Þ; where CPD jk is the coefficient of percentage dissimilarity between sites j and k; and P ik and P ij are proportions (by number of individuals) of species i in samples j and k; respectively. Cluster and MDS analyses were performed by means of the computer package Statistica 5.0 (Statsoft, Inc., Tulsa, OK, USA). The dissimilarity matrix was calculated using an original program written in Statistica Basic. Diversity indices were calculated by using the MVSP shareware computer package (Multivariate Statistical Package, MVSP Shareware 2.0). 3. Results The surveys enumerated 15,634 individuals and 102 species in February March (dry season) and 19,441 individuals and 100 species in June (wet season). Of the species surveyed, only 4 species accounted for 450% of the fishes surveyed (Table 2). Fish species composition was consistent within each of the three areas and within each biotope, so data were pooled within biotope and within area for replicate sites. The same four species

5 C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] 5 Table 2 Species included among the 10 most abundant for at least one combination of habitat, sampling period, and distance from Havana Harbor (see Tables 3 5) Scientific name Family Common name N %N F %F Abudefduf saxatilis (Linnaeus) Pomacentridae Sargeant major Acanthurus bahianus Castelnau Acanthuridae Ocean surgeonfish Acanthurus coeruleus Bloch and Schneider Acanthuridae Blue tang Acanthurus chirurgus (Bloch) Acanthuridae Doctorfish Cantherhines pullus (Ranzani) Balistidae Filefish Caranx crysos (Mitchill) Carangidae Blue runner Caranx ruber (Bloch) Carangidae Bar jack Clepticus parrae (Bloch and Schneider) Labridae Creole wrasse Chaetodon capistratus (Linnaeus) Chaetodontidae Foureye butterflyfish Chaetodon ocellatus (Bloch) Chaetodontidae Spotfin butterflyfish Chaetodon striatus (Linnaeus) Chaetodontidae Banded butterflyfish Chromis cyanea (Poey) Pomacentridae Blue chromis Chromis multilineata (Guichenot) Pomacentridae Brown chromis Emblemaria pandionis (Evermann and Marsh) Clinidae Sailfin blenny Haemulon flavolineatum (Desmarest) Haemulidae French grunt Haemulon plumieri (Lacépède) Haemulidae White grunt Halichoeres bivittatus (Bloch) Labridae Slippery dick Halichoeres garnoti (Valenciennes) Labridae Yellowhead wrasse Halichoeres maculipinna (Mu ller and Troschel) Labridae Clown wrasse Holocentrus ascencionis (Osbeck) Holocentridae Squirrelfish Lutjanus synagris (Linnaeus) Lutjanidae Lane snapper Malacanthus plumieri (Bloch) Malacanthidae Sand tilefish Malacoctenus triangulatus Springer Clinidae Saddled blenny Mulloidichthys martinicus (Cuvier) Mullidae Yellow goatfish Myripristis jacobus (Cuvier) Holocentridae Black bar soldierfish Ocyurus chrysurus (Bloch) Lutjanidae Yellowtail snapper Ophioblennius atlanticus (Valenciennes) Blenniidae Redlipblenny Sparisoma aurofrenatum (Valenciennes) Scaridae Redband parrotfish Stegastes leucostictus (Mu ller and Troschel) Pomacentridae Beaugregory Stegastes partitus (Poey) Pomacentridae Bicolor damselfish Thalassoma bifasciatum (Bloch) Labridae Bluehead wrasse N; total number of individuals counted in the study. %N; percentage representation of the species, calculated over a total of 35,075 individuals counted for both sampling periods. F; number of censuses in which the species was present. %F; frequency of occurrence of the species, calculated over a total of 480 stationary censuses. were the most abundant for both seasons and accounted in most cases for more than 40% of individuals. These fishes were the bluehead wrasse (Thalassoma bifasciatum), the bicolor damselfish (Stegastes partitus), the ocean surgeonfish (Acanthurus bahianus), and the slippery dick (Halichoeres bivittatus). The abundance of these species per sampling site, however, varied notably along the coast. Two contrasting patterns were found for S. partitus (Fig. 3) and H. bivittatus (Fig. 4). T. bifasciatum showed a pattern very similar to that of S. partitus, whereas A. bahianus was much more evenly distributed along the coast. The habitat preferences of the four most abundant fishes varied according to species. At the area farthest from Havana Harbor (area 3, km), the bluehead wrasse was the most abundant species at all biotopes during the dry season. In deeper water (Tb, Te biotopes), damselfish and surgeonfish were the next most abundant species (Table 3). In shallow water (Pl, Ec biotopes) surgeonfish were the second most common species, followed by slippery dick. During the wet season, damselfish were more common at all depths than during the dry season because of the peak of breeding activity and the presence of juvenile fish. Bluehead wrasse were still the most abundant species, except at the deepest biotope, and surgeonfish were again common at all depths. The brown chromis (Chromis multilineata) was the fourth most common fish in the wet season and was usually found at shallower depths. Slippery dick comprised less than 5% of the fish encountered in both seasons. Area 2 ( km) was similar to the reference site in the dry season, with the bluehead wrasse being the most abundant species at all biotopes except the deepest (Tb), where bicolor damselfish dominated (Table 4). Surgeonfish represented the next most abundant species at all depths in both seasons, followed by the French grunt (Haemulon flavolineatum). Slippery dick comprised o5% of the fish observed at this site and, as in area 3, were less commonly observed in the wet season. At the site closest to Havana Harbor (area 1, km), there was a marked decrease in the

6 6 ARTICLE IN PRESS C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] Fig. 3. Total number of individuals of the species H. bivittatus (slippery dick) visually censused at each sampling site and biotope for both seasons. Site numbers correspond to those in Table 1. Biotopes: Echinometra ; Rocky plain ; Terrace edge &; Terrace base. Fig. 4. Total number of individuals of the species S. partitus (bicolor damselfish) visually censused at each sampling site and biotope for both seasons. Site numbers correspond to those in Table 1. Biotopes: Echinometra ; Rocky plain ; Terrace edge &; Terrace base.

7 C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] 7 Table 3 Percentage representation of the total number of dominant individuals counted for each biotope in area 3 ( km from Havana Harbor) sites Species Percentage representation February March June Tb Te Pl Ec Tb Te Pl Ec Thalassoma bifasciatum Stegastes partitus Acanthurus bahianus Halichoeres bivittatus Acanthurus chirurgus Sparisoma aurofrenatum Holocentrus ascencionis Acanthurus coeruleus Chromis cyanea Acanthurus (juv.) Ocyurus chrysurus Haemulon plumieri 1.9 Chromis multilineata Abudefduf saxatilis Mulloidichthys martinicus 3.4 Chaetodon capistratus Ophioblennius atlanticus 3.0 Chaetodon ocellatus 1.3 Clepticus parrae (juv.) 5.5 Malacanthus plumieri 2.5 Halichoeres garnoti 1.9 Caranx ruber 5.6 Myripristis jacobus 1.9 Scaridae (juv.) 1.5 Emblemaria pandionis 4.6 Halichoeres maculipinna 2.3 Lutjanus synagris 2.2 Subtotal 10 species The 10 most abundant species in each biotope for both sampling periods are included. Biotopes are Tb, terrace base; Te, terrace edge; Pl, rocky plain; Ec, Echinometra. Species in Tb in the sampling period February March are ordered from most to least abundant. abundance of the bicolor damselfish (Stegastes), especially in shallow waters and in June (wet season). The most striking change along the coast is the greater percentage representation of the slippery dick and the grunts in the area closest to Havana Harbor (area 1) compared to areas farther away (Table 5). Grunts are represented by two closely related species, the white grunt, Hae. plumieri, and the French grunt, Hae. flavolineatum. The first was more abundant in February March and the second dominated in June. The bluehead wrasse, T. bifasciatum, had the same trend as the bicolor damselfish, although the decrease was less dramatic. The creole wrasse, Cl. parrae, exhibited a very high percentage representation at the area 1 Te biotope in February March and was the most abundant species closest to Havana Harbor. All the individuals counted were adults forming small groups feeding in the water column. In areas 2 and 3, however, this species had a lower rank and was present in June only as early juvenile individuals near the bottom. Other, less abundant species showed clear habitat preferences. Among the species which represent a significant proportion of the community in deep waters are Halichoeres garnoti, Malacanthus plumieri, Chromis cyanea, Myripristis jacobus, and Cl. parrae. These fishes clearly preferred the Tb and Te biotopes. Included in the 10 most abundant species in the shallower habitats of Pl and Ec were Chaetodon capistratus, Chaetodon ocellatus, and most juvenile forms. Other species did not show clear patterns of habitat preference. Evidence of a change in habitat preference associated with the sampling period (season) was not found for any of the species included in this analysis. However, several species were represented only by early juvenile individuals classified among the 10 dominant species in June. These species were the bar jack, Ca. ruber, several parrotfish species of the family Scaridae, the surgeonfishes, Acanthurus spp., and several species of grunts of the genus Haemulon Diversity indices The total number of species, Shannon s index of diversity (H 0 ), and Pielou s index of evenness (J 0 ) were

8 8 ARTICLE IN PRESS C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] Table 4 Percentage representation of the total number of dominant individuals counted for each biotope in area 2 ( km from Havana Harbor) sites Species Percentage representation February March June Tb Te Pl Ec Tb Te Pl Ec Stegastes partitus Acanthurus bahianus Haemulon flavolineatum Thalassoma bifasciatum Halichoeres bivittatus Halichoeres garnoti Acanthurus coeruleus Haemulon plumieri Lutjanus synagris 2.7 Sparisoma aurofrenatum Abudefduf saxatilis 2.2 Chaetodon capistratus Ocyurus chrysurus Holocentrus ascencionis 1.5 Chaetodon striatus Emblemaria pandionis 6.1 Stegastes leucostictus 0.3 Chromis cyanea Mulloidichthys martinicus Myripristis jacobus 4.0 Chromis multilineata Clepticus parrae (juv.) 3.0 Acanthurus (juv.) Caranx ruber (juv.) Haemulon spp. (juv.) Scaridae (juv.) Subtotal 10 species The 10 most abundant species in each biotope for both sampling periods are included. Biotopes are Tb, terrace base; Te, terrace edge; Pl, rocky plain; Ec, Echinometra. Species in Tb in the sampling period February March are ordered from most to least abundant. calculated for the pooled data in each combination of biotope, period of sampling, and area (Table 6). A strong dependence of H 0 and S (number of species) on the total number of individuals counted was found for values of N (number of individuals) less than 1500 (Fig. 5). Therefore, a comparative analysis among biotopes was not reliable, as the number of individuals counted was always less than 1500 in the shallower biotopes (Pl and Ec). The values of the diversity measures were clearly independent of the number of individuals for those biotopes with high values of N; i.e., Tb and Te (Table 6). Therefore, a comparative analysis of the values among periods of sampling and proximity to Havana Harbor can be made for these biotopes. In all cases, area 1 (near Havana Harbor) sites showed the highest values of H 0 and for area 3 (sites farthest from Havana Harbor) the values of this index were the lowest. This change is strongly related to the evenness index rather than to number of species. The lower percentage representations of the bluehead wrasse and the bicolor damselfish in area 1 result in a more even representation of the dominant species, thus increasing the diversity index Cluster analysis and multidimensional scaling The cluster analysis of the data pooled for each combination of biotope, sampling period, and area yielded well-differentiated clusters (Fig. 6). The first one (A) includes all combinations related to shallower biotopes (Pl and Ec) plus those related to the biotope Te in area 1. The other main cluster, B, contains the rest of the combinations related to deeper biotopes. Inside each cluster, smaller ones are formed that include mainly pooled samples from the same biotope (A2, B1, B2), although they also include some samples from other biotopes. This indicates that the main axis of change in the fish assemblages are among biotopes and two main subassemblages can be identified: one living on shallow and less complex biotopes and the other living in deeper, more complex, biotopes. There is no evidence of any clustering due to seasonal differences and in many cases the two seasons for the same biotope and area are joined at a high level of similarity. The two-dimensional MDS ordination based on the same dissimilarity matrix resulted in a final configuration with a stress value of (Fig. 7), indicating that

9 C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] 9 Table 5 Percentage representation of the total number of dominant individuals counted for each biotope in area 1 (San Lazaro s Cove; km from Havana Harbor) sites Species Percentage representation February March June Tb Te Pl Ec Tb Te Pl Ec Halichoeres bivittatus Stegastes partitus Thalassoma bifasciatum Acanthurus bahianus Haemulon plumieri Haemulon flavolineatum Abudefduf saxatilis Sparisoma aurofrenatum Halichoeres garnoti Lutjanus synagris 2.6 Chromis multilineata 9.2 Clepticus parrae 22.7 Emblemaria pandionis Chaetodon capistratus Acanthurus coeruleus Caranx ruber Acanthurus chirurgus Chaetodon ocellatus 1.7 Ophioblennius atlanticus 3.2 Malacoctenus triangulatus Acanthurus (juv.) Caranx crysos 3.3 Scaridae (juv.) Cantherhines pullus 1.9 Subtotal 10 species The 10 most abundant species in each biotope for both sampling periods are included. Biotopes are Tb, terrace base; Te, terrace edge; Pl, rocky plain; Ec, Echinometra. Species in Tb in the sampling period February March are ordered from most to least abundant. Table 6 Diversity measures calculated for each combination of biotope, period of sampling, and area Biotope Area Diversity values February March June N S H 0 J 0 N S H 0 J 0 Terrace base Terrace edge Rocky plain Echinometra N; number of individuals counted. S; number of species. H 0 ; Shannon s index of total diversity. J 0 ; Pielou s index of evenness. Area 1, San Lazaro s Cove; km from Havana Harbor; area 2, km from Havana Harbor; area 3, km from Havana Harbor.

10 10 ARTICLE IN PRESS C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] two dimensions are enough for a good ordination. The combination of clustering and ordination analyses is an effective way of checking the adequacy and mutual consistency of both representations (Clarke and Warwick, 1994). In this case there is a great coincidence between the results obtained from the two methods. All (a) 4.0 Shannon's index (H') Tb Pl Te Ec Total number of individuals (N) (b) 80 Number of species (S) Tb Pl Te Ec Total number of individuals (N) Fig. 5. (A) Shannon s index (H 0 ) and (B) Number of species (S) against the total number of individuals (N) counted in pooled samples for each combination of biotope, sampling period and area. Samples are classified according to biotope to emphasize the effect of the number of individuals counted on the value of the indices. E Echinometra (Ec); Rocky plain (Pl); m Terrace edge (Te); K Terrace base. the pooled samples pertaining to shallow biotopes had positive values on the first ordination axis, whereas those pertaining to deep biotopes had negative values. For a clear representation of the similarity of the MDS ordination (Fig. 7) and the previous dendrogram (Fig. 6), circles in Fig. 7 enclose the shallow (GroupA in Fig. 6) and deep(groupb in Fig. 6) biotopes. A dotted line (Fig. 7) encircles the members of the A1 cluster from Fig. 6, which are all area 1 sites. 4. Discussion Three possible factors influencing the composition of fish assemblages along the northern coast of Cuba have been considered in the present study: season, habitat, and proximity to Havana Harbor. The changes between sampling periods due to season can be considered to be minor along the northern coast of Cuba. Overall, this study has demonstrated that the main source of variation of the composition of fish assemblages is the biotope. The secondary factor is proximity to Havana Harbor. Fausch et al. (1990) pointed out several reasons to consider fish as useful organisms for measuring environmental degradation, including that fish communities act as integrators of direct and indirect stress on aquatic ecosystems. In this study, analyses of changes in marine fish communities have been conducted using different approaches and quantitative methods. It is generally agreed that there is not a best method and that it is useful to use different methods on the same data set (Fausch et al., 1990; Clarke and Warwick, 1994). The use of various techniques allows the testing of different models of response at the individual and assemblage levels and coincident results give more strength to the data interpretation. Fig. 6. Dendrogram showing the numerical classification of pooled samples for each combination of biotope, sampling period, and area. Symbols in the labels are F, February March; J, June; 1, area 1; 2, area 2; 3, area 3. EC, Echinometra; PL, Rocky plain; TE, Terrace edge; TB, Terrace base. See text for explanation of groups formed.

11 C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] 11 Fig. 7. Multidimensional scaling of the pooled samples for each combination of biotope, sampling period, and distance from Havana Harbor. Symbols are B/E for Echinometra, &/ for Rocky plain, n/m for Terrace edge, J/K for Terrace base. Open symbols: February March. Solid symbols: June. Sampling areas: area 1 (AR1), area 2 (AR2), and area 3 (AR3). Continuous lines encircle the members of clusters A and B from Fig. 6. A dotted line encloses members of cluster A1 from Fig. 6. Three main natural factors can be identified as influencing the fish distribution among biotopes (McGehee, 1994). These are water motion (Hilomen and Gomez, 1988) and type and complexity of substrate of the bottom from coral growth and other sessile organisms forming the reef (i.e., sponges, octocorals) (Bell and Galzin, 1984; Grigg, 1994; Chabanet et al., 1997; Lirman, 1999). All these factors are strongly correlated with depth, which is reflected in the zonation of bottom communities as bands running parallel to the shore (Nun ez-lara and Arias-Gonza lez, 1998). It is not possible to make a direct comparison of our findings with results obtained by authors working at other Caribbean reefs because those describe biotopes which are not the same as the ones considered in the present study. Previous works do not give explicit information on the abundance of the species included (McGehee, 1994; Nun ez-lara and Arias-Gonza lez, 1998) or include only selected species in the analysis (Alevizon et al., 1985). The main agreement of the present study with others in the Caribbean is the observed change in the composition of fish assemblages related to biotope, e.g., McGehee (1994) reported that species of the genus Acanthurus preferred a shallow biotope, whereas S. partitus was more associated with deepbiotopes. The sampling periods were selected to determine if there were any changes in composition of the ichthyofauna before and after the massive juvenile settlement period, which takes place in late spring (Aguilar and Gonza lez-sanso n, 1998). This can be considered the major factor influencing temporal change in the fish assemblages in the study area. The most dominant species (T. bifasciatum, S. partitus, A. bahianus, and H. bivittatus), however, were the same in both periods, suggesting that changes in the structure of fish assemblages are not very significant on the whole, although juvenile stages of some species (Acanthurus spp., Scaridae, Ca. ruber, Haemulon spp., and Cl. parrae) showed differences in representation between seasons (higher representation in June). Research carried out previously in the same study area has yielded very similar results in terms of dominant fish species (Aguilar and Gonza lez-sanso n, 2000). Therefore, the composition of the ichthyofauna can be considered to be stable during the dry and wet seasons. The low representation of the two most abundant species (S. partitus and T. bifasciatum) in area 1 (sites closest to Havana Harbor in San Lazaro s cove) is evidence of stressor effects. Area 1 biota are exposed directly to Havana Harbor inputs while area 2 contaminants would be predominately influenced by the Almendares River. S. partitus is a bottom-dwelling, highly territorial species, which feeds mainly on benthic algae and small invertebrates (Emery, 1973; Booth and Hixon, 1999), while T. bifasciatum is a benthic, territorial fish with a natural diet composed of small benthic invertebrates (Randall, 1967; Clifton and Motta, 1998). S. partitus is a bottom-nesting fish (Knapp and Warner, 1991; Cole and Sadovy, 1995; Knapp et al., 1995) which needs clean rocky substrate for the construction of its nests. Research made as part of a regional project funded by the Global Environmental Facility (GEF) and the United Nations Development Program (UNDP) found high concentrations of nonfilterable solids ( mg L 1 ) in the water column of San Lazaro s cove (area 1), which could account for the lack of suitable habitat (and therefore much lower abundance) of this species (GEF/UNDP, 1998). While data were being collected for this study, the texture of sand in some sampling sites near the bay s entrance and low visibility in the water column were noted. The presence of high concentrations of hydrocarbons (936 mg of total hydrocarbons per gram of dry sediment matter) in the bottom of San La zaro s cove has also been documented (GEF/UNDP, 1998). The petroleum contamination and/or siltation could potentially be negatively affecting populations of benthic-dwelling fish species, and this warrants further study. The increased representation of H. bivittatus (the fourth most abundant species in the study) in area 1, deserves some attention. This wrasse has behavioral characteristics and feeding habits which are very similar to those of T. bifasciatum (Randall, 1967; Valdés- Mun oz and Mochek, 1994; Clifton and Motta, 1998). Its representation was greater in the sites within San La zaro s cove in comparison to S. partitus and T. bifasciatum. It may be the case that H. bivittatus has a higher tolerance for stressor effects or that relaxed competition pressure on this species by T. bifasciatum and S. partitus in area 1 occurred. H. bivittatus appears to be more tolerant to siltation than other species of

12 12 ARTICLE IN PRESS C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] wrasses (Aliaume et al., 1990; Baelde, 1990; Bouchon- Navaro et al., 1992; Thomas and Logan, 1992; Aguilar et al., 1997; Gonza lez-sanso n et al., 1997), which could account for its presence in the area. Further work is required to confirm a greater tolerance of H. bivittatus for stressors present around Havana Harbor compared to other wrasses. It is expected that due to dilution and currents, the contaminants and siltation coming from the harbor and the river do not reach very high levels and are not lethal for any fish species. However, indirect effects on habitat and possibly chronic effects on fish health (e.g., growth and reproduction) may occur. This is supported by the variations in total diversity and evenness values observed for the deeper biotopes (Tb and Te) closer to Havana Harbor. These habitats appear to be experiencing intermediate disturbances as defined by Conell (1978) and Jones and Syms (1998). In area 3, stressors on the fish assemblages are very low and competition probably results in the dominance of very few species, lowering the value of the Pielou index for evenness (J 0 ) and consequently the value of Shannon s index (H 0 ). In areas 1 and 2, the stressors are higher in the form of contaminants such as heavy metals and crude oil (Gonza lez, 1991; Beltra n et al., 1998) and dominant fish species perhaps find a less favorable environment, permitting other species to increase their own numbers and relative importance. This causes an increase in evenness and, therefore, total diversity. Further research is required to determine the extent of stressor effects and whether stressors are acting directly (e.g., contaminant effects) or indirectly (e.g., changes in habitat) on the resident fish populations. The allocation of two pooled samples pertaining to a deep biotope in area 1 for both sampling periods (i.e., Te) in the cluster containing the samples of shallow biotopes can also be considered an indication of an effect on the composition of fish assemblages due to proximity to Havana Harbor. Further evidence of this is the fact that the two samples from the deep biotope are included in the same cluster (A1) with the pooled samples pertaining to shallow biotopes in area 1. This is the only case in which area designation dominates over biotope in the clustering. This result indicates that proximity to Havana Harbor can be considered a second axis of variation in composition of the fish assemblages. In summary, this is the first research at the community level to show effects on fringing coral reef fish assemblages in relation to proximity to Havana Harbor. These effects are independent of the natural factors (habitat and season) studied. Other authors have also found an effect of pollution, eutrophication, and increased siltation on fish community composition (Marszalek, 1987; Fausch et al., 1990; Dawson- Shepherd et al., 1992). Research is currently proceeding to determine if there are effects in area 1 due to contaminant exposure at the population and individual levels. Baseline data are required for monitoring purposes, as stress on the ecosystem continues to increase, e.g., through industrial growth and tourism development. Present studies focus on S. partitus, T. bifasciatum, and H. bivittatus. These studies include caging studies with naïve fish (K.E. Costain and D.L. MacLatchy) and analyses of morphological and organosomatic indices and reproductive endpoints of wild fish at selected sites in areas 1, 2, and 3 (C. Aguilar, G. Gonza lez-sanso n, D.L. MacLatchy, and K.R. Munkittrick). Acknowledgments The authors thank Luis Sanchez and Armando Perez for their assistance in the field and at the Centro de Investigaciones Marinas, Universidad de La Habana. Funding for this work was provided by a Natural Sciences and Engineering Research Grant to D.L.M and a Canadian International Development Agency Tier 2 Grant to the Centro de Investigaciones Marinas and the University of New Brunswick Saint John. References Aguilar, C., González-Sansón, G., Variacio n estacional de la abundancia de juveniles de peces en una zona del sublitoral rocoso de la Habana, Cuba. Rev. Invest. Mar. 19, Aguilar, C., González-Sansón, G., Influencia de la contaminacio n de la bahía de la Habana sobre las asociaciones de peces costeros: I. Abundancia y diversidad. Rev. Invest. Mar. 21, Aguilar, C., González-Sansón, G., Angulo, J., González, C., Variacio n espacial y estacional de la ictiofauna en un arrecife de coral costero de la regio n noroccidental de Cuba I. Abundancia total. Rev. Invest. Mar. 18, Alevizon, W., Richardson, R., Pitts, P., Serviss, G., Coral zonation and patterns of community structure in Bahamian reef fishes. Bull. Mar. Sci. 36, Aliaume, C., Laserre, G., Louis, M., Organisation spatiale des peuplements ichtyologiques des herbiers à Thalassia du Grand Culde Sac marin en Guadeloupe. Rev. Hydrobiol. Trop. 23, Areces, A., Toledo, L., Características tro ficas de la Bahía de La Habana durante el período de seca. Rep. Invest. Inst. Oceanol. 40, 32p. Baelde, P., Differences in the structures of fish assemblages in Thalassia testudinum beds in Guadeloupe, French West Indies, and their ecological significance. Mar. Biol. 105, Baisre, J., Chronicle of Cuban marine fisheries ( ): trend analysis and fisheries potential. FAO Fish. Tech. Pap Bell, J.D., Galzin, R., Influence of live coral cover on coral-reef fish communities. Mar. Ecol. Prog. Ser. 15, Beltrán, J., Ruiz, F., Vega, L., Contaminacio n por hidrocarburos del petro leo en la Bahía de La Habana, Cuba. Rev. Cie. Te c. IIT: Transporte Desarrollo y Medio Ambiente, Cuba 18, Boesch, D.F., Application of numerical classification in ecological investigations of water pollution. Va. Inst. Mar. Sci. Spec. Sci. Rep. 77, ix+113.

13 C. Aguilar et al. / Ecotoxicology and Environmental Safety ] (]]]]) ]]] ]]] 13 Bohnsack J.A., Bannerot, S.P., A stationary visual census technique for quantitatively assessing community structure of coral reef fishes. NOAA Tech. Rep. NMFS 41, iii+15. Booth, D.J., Hixon, M.A., Food ration and condition affect early survival of the coral reef damselfish, Stegastes partitus. Oecologia 121, Bouchon-Navaro, Y., Bouchon, C., Louis, M., L ichthyofaune des herbiers de phanerogames marines de la baie de Fort-de-France (Martinique, Antilles Fran@aises). Cybium 16, Chabanet, P., Ralambondrainy, H., Amanieu, M., Faure, G., Galzin, R., Relationships between coral reef substrata and fish. Coral Reefs 16, Clarke, K.R., Warwick, R.M., Change in marine communities: an approach to statistical analysis and interpretation. Natural Environment Research Council, UK. Clifton, K.B., Motta, P.J., Feeding morphology, diet, and ecomorphological relationships among five Caribbean labrids (Teleostei: Labridae). Copeia 4, Cole, K.S., Sadovy, Y., Evaluating the use of spawning success to estimate reproductive success in a Caribbean reef fish. J. Fish Biol. 47, , doi: /jfbi Conell, J.H., Diversity in tropical rainforests and coral reefs. Science 199, Dawson-Shepherd, A.R., Warwick, R.M., Clarke, K.R., Brown, B.E., An analysis of fish community responses to coral mining in the Maldives. Environ. Fish. Biol. 33, Emery, A.R., Comparative ecology and functional osteology of fourteen species of damselfish (Pisces: Pomacentridae) at Alligator Reef, Florida Keys. Bull. Mar. Sci. 23, FAO, Marine fisheries and the law of the sea: A decade of change. FAO Fish. Circular 853, v+65. Fausch, K.D., Lyons, J., Karr, J.R., Argeimeier, P.L., Fish communities as indicators of environmental degradation. Am. Soc. Symp. 8, GEF/UNDP, Planificacio n y manejo ambiental de bahías y zonas costeras fuertemente contaminadas del Gran Caribe: Estudio de caso Bahía de La Habana, Cuba. Resultado 1.1. Estudios que identifican las condiciones actuales de la bahía y de las áreas costeras contiguas, incluyendo el grado de impacto ambiental sobre sus componentes biolo gicos. Final Report for the Project RLA/93/ G41. Havana. González, H., Heavy metal surveys in sediments of five important Cuban bays. Biochemistry 14, González-Sansón, G., Aguilar, C., Influencia de la contaminación de la Bahía de la Habana (Cuba) sobre las asociaciones de peces costeros. II: Análisis multidimensional. Rev. Invest. Mar. 21, González-Sansón, G., Aguilar, C., Angulo, J., González, C., Variación espacial y estacional de la ictiofauna en un arrecife de coral costero de la región noroccidental de Cuba II: diversidad. Rev. Invest. Mar. 18, Grigg, R.W., Effects of sewage discharge, fishing pressure and habitat complexity on coral ecosystems and reef fishes in Hawaii. Mar. Ecol. Prog. Ser. 103, Guardia, E., González-Sansón, G., Asociaciones de corales, gorgonias y esponjas del sublitoral habanero al oeste de la bahía de La Habana. I. Gradiente ambiental. Rev. Invest. Mar. 21, 1 8. Harmelin-Vivien, M., Impact des activités humaines sur le peuplements ichthyologiques des re cifes coralliens de polynésie Fran@aise. Cybium 16, Hilomen, V.V., Gomez, E.D., Distribution of fish communities in some Philippine reefs. Proceedings 6th International Coral Reef Symposium, Vol. 3. Townsville, Australia, pp Hughes, T.P., Catastrophes, phase shifts, and large scale degradation of a Caribbean coral reef. Science 265, Jones, G.P., Syms, C., Disturbance, habitat structure and the ecology of fishes on coral reefs. Aust. J. Ecol. 23, Knapp, R.A., Warner, R.R., Male parental care and female choice in the bicolour damselfish, Stegastes partitus: bigger is not always better. Anim. Behav. 41, Knapp, R.A., Sikkel, P.C., Vredenburg, V.T., Age of clutches in nests and the within-nest spawning-site preferences of three damselfish species (Pomacentridae). Copeia 1, Kruskal, J.B., Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis. Psychometrika 29, Lirman, D., Reef fish communities associated with Acropora palmata: relationships to benthic attributes. Bull. Mar. Sci. 65, Luckhurst, B.E., Luckhurst, K., Analysis of the influence of substrate variables on coral reef fish communities. Mar. Biol. 49, Marszalek, D.S., Sewage and eutrophication. In: Salvat, B. (Ed.), Human Impacts on Coral Reefs: Facts and Recommendations. Antenne Museum EPHE, French Polynesia, pp McGehee, M.A., Correspondence between assemblages of coral reef fishes and gradients of water motion, depth, and substrate size off Puerto Rico. Mar. Ecol. Prog. Ser. 105, Nun ez-lara, E., Arias-González, E., The relationshipbetween reef fish community structure and environmental variables in the southern Mexican Caribbean. J. Fish Biol. 53, , doi: /jfbi Randall, J.E., Food habits of reef fishes of the West Indies. Stud. Trop. Oceanogr. 5, Sneath, P.H.A., Sokal, R.R., Numerical Taxonomy. Freeman, San Francisco, pp. xv+573. Thomas, M.L.H., Logan, A., A guide to the ecology of shore line and shallow-water marine communities of Bermuda. Bermuda Biol. St. Res. Spec. Pub. 30. Valdés-Muñoz, E., Mochek, A.D., Estructura etolo gica de las comunidades de peces. In: Claro, R. (Ed.), Ecología de los Peces Marinos de Cuba. CIQRO, Quintana Roo, México, pp Zapata, F.A., Morales, Y.A., Spatial and temporal patterns of fish diversity in a coral reef at Gorgona Island, Colombia. Proceedings 8th International Coral Reef Symposium. Panama. Vol. 1,