A Functional Analysis of the Benthic Macrofauna of the São Sebastião Channel (Southeastern Brazil)

Problem

The increase of environmental pollution in urban coastal zones during the last decades is one of the main problems facing developing countries. Knowledge on the dynamics of these coastal areas through the study of marine communities allows the ecological and environmental conditions to be assessed. Benthic macrofauna is used to describe changes in the environment because the organisms are relatively sedentary and have comparatively long life spans (Thouzeau et al., 1991). In addition, the macrofauna consists of different species that exhibit different tolerances to stress (Dauer, 1993) and permit monitoring of environmental quality. Macrobenthos can also be employed to understand the dominance of certain ecological factors responsible for the structure and productivity of the communities (Saiz-Salinas, 1997).

Biological variables are important components in water quality assessment because they may uncover problems undetected in measurements of different physicochemical parameters or underestimated by other methods (Dauer, 1993). For these reasons, biotic indices were developed recently, especially in North America and Europe (Grall & Glémarec, 1997; Van Dolah et al., 1999). All these studies emphasised the importance of biological indicators to measure the ecological quality in the marine environment. The marine Biotic Coefficient (BC) recently proposed by Borja et al. (2000) applies benthic macrofauna as bio-indicators and explores the response of soft-bottom communities to natural and man-induced changes in water quality, integrating long-term environmental conditions. Recent papers showed good results using it and suggested its application in a more wide range of environments (Borja et al., 2003, 2004).

The present paper discusses the distribution patterns and structure of the benthic macrofauna in relation to environmental conditions along the São Sebastião Channel, an area influenced by anthropogenic activities. This region has an important commercial port (São Sebastião) and oil terminal (Dutos e Terminais Centro Sul – DTCS) which processes up to 55% of the Brazil's oil (Zanardi et al., 1999a). Also, the area is influenced by two sewage discharges, one located on the continent at Ponta do Araçá, near the harbour, and the other at Ilha Bela.

In the channel area, structural characteristics of the macrobenthos were examined by Flynn et al. (1999), who studied the species composition and spatial pattern of benthic fauna in the middle section of the channel. Other studies conducted by Muniz & Pires (1999, 2000) analysed the trophic structure and distribution patterns of polychaetes throughout this area. The goal of the present paper is to study the spatial and temporal patterns of benthic macrofauna in view of the anthropic activities developed in the area, using the species classification and trophic group approach, combining univariate biological and multivariate analyses.

Material and Methods

1. Study area

The São Sebastião Channel is located on the northern coast of São Paulo State (23°41′ to 23°53.5′ S; 45°19′ to 45°30′ W) (Fig. 1). The channel separates the continent from the São Sebastião Island, the second largest island in Brazil, forming a well-sheltered area ideal for shipping operations. The extension is about 25 km long with two relatively large entrances (6–7 km wide). The middle portion is the narrowest (2 km in width) and the deepest (about 40 m). This narrowness increases the current speed and consequently, no sediments are deposited in this area (Furtado, 1995). Furtado (1995) and Furtado et al. (1998) intensively studied the channel topography. Briefly, the northern region of São Sebastião Channel has a counter-clockwise vortex, which transports fine grains to the south, depositing fine sediment at the continental margin, a region with low hydrodynamic energy conditions.

The marine currents here are mainly NE driven, except in summer when there is a two-layer water flow, the superficial one towards the SW and the deep one directed to the NE (Fontes, 1995). This bottom current is not influenced by wind but may be associated with the intrusion of the South Atlantic Central Water (SACW), which is characterised by low temperature (<18 °C) and high salinity (>35.4) (Miranda, 1985). Coastal Water (CW) dominates in the other seasons and is associated with high temperatures (>24 °C) and low salinity due to the combined effects of small and medium-sized estuaries (Castro & Miranda, 1998).

The São Sebastião Channel contains one of the largest oil terminals in Brazil (Dutos e Terminais Centro Sul, DTCS); it also has a harbour and is an important tourist site. The Araçá sewer accommodates 27% of São Sebastião city (CETESB, 2000) and is located in an area with low current speeds, making the dispersion of the effluent difficult.

2. Data collection

Four surveys were conducted with 15 stations distributed systematically on five transects perpendicular to the axis of the channel (Fig. 1). The surveys were seasonal from spring to winter – November/93 to August/94 – with the R/V ‘Veliger II’ and ‘Albacora’, both from the Oceanographic Institute, University of São Paulo. Stations were located with Global Positioning System (GPS). Water for hydrographic data was collected with Nansen bottles at three intervals (surface, mid-depth and bottom) along water column. In the laboratory, salinity measurements were determined with an inductive salinometer. Dissolved oxygen concentration was obtained by the Winkler titration method (Strickland & Parsons, 1968) and conversion tables were employed to calculate saturation levels (UNESCO, 1973).

Sediment and biological data were sampled with a 0.1 m² van Veen grab at the 15 stations, from 10 to 45 m depth (Table 1). About 200 g of the grab sediment was used for grain-size analysis as described in Suguio (1973). Parameters described by Folk & Ward (1957) were obtained for sedimentological data. The organic carbon content was calculated as described in Gaudette et al. (1974) and the nitrogen was estimated according to Kabat & Mayer (1948).

For biological analysis, the sediment was washed in situ through a 0.5 mm mesh sieve and the material retained was fixed with formalin (4%) and then preserved in 70% ethanol. In the laboratory, the sediment was washed again, applying the elutriation technique (Santos et al., 1996) before sorting and identification under a stereoscopic microscope. Most benthic organisms were classified to genus or species, except Nematoda, Nemertea, Echiura, Ascidiacea, Pycnogonida and Priapulida.

3. Data analysis

The frequency of occurrence (F) of the species was calculated to discriminate the most representative species. The F used was described by Guille (1970): F = p_a/P × 100 (p_a is the number of stations where species a occurred; P is the total number of stations). Only the species classified as common (F ≥ 15%), plus some species that were abundant in a few stations, were selected to perform the subsequent analyses. This group of species was also classified into trophic groups based on Fauchald & Jumars (1979), Rios (1994) and Lastra et al. (1991). Five trophic groups were considered: carnivores (C), surface deposit-feeders (S), subsurface deposit-feeders (B), omnivores (O) and suspension-feeders (F).

The marine Biotic Coefficient (BC) used (Borja et al., 2000) allows the classification of the benthic environment according to the environmental quality. The BC is based on five groups of soft-bottom macrofauna, according to their sensitivity to an increasing stress gradient. Initially, these groups were determined by Grall & Glémarec (1997), and an updated list with more than 2000 taxa is available at AZTI's web page (http://www.azti.es). The BC is based upon the percentages of abundances of each of the five ecological groups and has the advantage of being a continuous value, from 0 to 6, reaching 7 when the sediment is azoic.

Pearson's correlation coefficient (Zar, 1984) was used to evaluate the degree of relation between the BC and the environmental variables.

A PCA ordination was performed for each survey to ordinate the samples based on their physico-chemical and sedimentological characteristics, using four ‘environmental variable × sampling sites’ matrices (standardised values, 12 × 15). In order to test statistically if these groups of stations were different in relation to species and trophic groups’ abundance data, a one-way analysis of similarity (ANOSIM) (Clarke, 1993) was applied using the Bray-Curtis similarity index and fourth root transformed abundance values. The statistical package Plymouth Routines in Multivariate Ecological Research (PRIMER) was used.

Results

1. The benthic environment

Analysis of temperature and salinity data showed the dominance of coastal water (CW) in the São Sebastião Channel over the study period, except in spring when South Atlantic central water (SACW) was present in the deepest part (45 m, Table 1). The concentration of dissolved oxygen in the bottom water varied, with lowest values in spring and highest in winter. Despite the high heterogeneity in the bottom sediments, a general trend is evident. There is deposition of fine and very fine sands in the south mouth of the channel and of coarse sediments in the northern portion, whereas the continental margin is dominated by fine fractions. In the central zone, pelitic fractions (silt and clay) prevailed and the organic carbon values were highest (Table 1).

The PCAs (Fig. 2) showed that in general the area can be divided in three regions based on the environmental variables studied. The first two axes explained 79% of the data variation in the spring survey, 73% in summer, 74% in autumn and 72% in winter. Stations 2, 7 and 10 always occurred together in one of the extremes of the diagram: this group was characterised by the highest values of organic carbon, nitrogen and mean sediment diameter and by a low contribution of medium and coarse sand (Fig. 2), suggesting an environment with low hydrodynamic energy. Depending on the sampling period, one or two additional sampling stations completed this group (Fig. 2). Stations 3, 9, 11, 14 and sometimes stations 12, 13 and 15 formed the other group, which was characterised by high contributions of medium and coarse sand and the lowest organic load. The third group in the PCAs was characterised by mixed sediments and intermediate values of organic carbon and nitrogen (Fig. 2).

2. The fauna

About 19,000 individuals (81%) were identified to genus or species, and 392 species were determined: 129 Polychaeta, 127 Mollusca, 98 Crustacea, 28 Echinodermata and 10 from other phyla. Only 62 species were common and constant to the area, and two gastropods, Caecum striatum and C. pulchellum, were abundant in specific regions of the channel. Arasaki (1997) analysed the density and diversity patterns of these communities, presenting a detailed inventory of the macrofaunal species. A species list is available on request.

3. The Biotic Coefficient

The BC values varied between 0.7 and 2.6, both values occurring in the spring survey (Table 2). According to the classification proposed by Borja et al. (2000), most of the stations were unpolluted and slightly polluted, corresponding to impoverished and unbalanced benthic community health. Stations 7, 8, 10, 12 and 13 in spring; stations 7, 8 and 13 in summer; stations 5, 8, 10, 12 in autumn and stations 1, 2, 5 and 7 in winter are highlighted for their higher BC values (2.0 and 2.6) (Table 2). BC was positively correlated (P < 0.05) with the organic carbon content, organic nitrogen content and mean grain size (φ) of the sediment, and negatively correlated with the percentage of coarse sand. Thus, higher BC values tend to be associated with a greater organic load.

4. Multivariate analysis

The analysis of similarities (ANOSIM) of the species and trophic group abundances between the groups of samples generated with the PCAs showed significant differences between almost all the possible pairs of groups (P < 0.05) in the four surveys. The only exception was the pair I and III in the spring survey (Table 3). Table 4 presents the abundance of the most important species for the groups defined. Note that characteristic species of group III were absent in the other groups or were more abundant in group III than in the others.

Group I always showed lower abundances and characteristic deposit-feeder polychaetes such as nereids, spionids and paranoids (Table 4). This group had the highest abundance of opportunistic polychaete species in addition to the ophiuroid Amphiodia atra and the sipunculid Aspidosiphon gosnoldi. A predominance of stations with fine and very fine sand and intermediate values of organic carbon characterised group II. Here, magelonids, lumbrinerids and the suspension and surface deposit-feeder Owenia fusiformis dominated (Table 4). Group III had a high abundance of small molluscs like the gastropods Caecum striatum and C. pulchellum and the bivalve Corbula caribaea, besides the polychaetes Exogone arenosa, Parandalia americana, Aricidea (Acmira) taylori and the anthozoan Edwardsia sp. (Table 4). It corresponded to the channel region dominated by medium and coarse sand and having the least organic carbon. The deepest station (8) was isolated from the three groups by having a high abundance of Cirrophorus americanus and the exclusive presence of the large suspension feeder Spirographis spallanzani.

Although surface deposit-feeders were the most abundant trophic group throughout the channel (Table 5), carnivores were particularly numerous in the stations of group III and omnivores and subsurface deposit-feeders in group I.

The BC varied within a narrow range, but the BC group mean values revealed a trend: generally BC decreased from group I to III (Table 5).

5. Temporal variation

Mean values of BC were almost constant (spring = 1.72 ± 0.53, summer = 1.65 ±0.46, autumn = 1.64 ± 0.45, winter = 1.64 ± 0.46) (Table 2) and, based on these results, the whole area may be classified as slightly polluted during the study period.

A seasonal analysis of the station groupings showed that almost the same group structure was maintained during the study period, except by stations 6, 13 and 15, which change their location from one survey to the other. Although the faunal composition was similar throughout the study (Table 4), some species were quite abundant in a given season. For example, Caecum species were abundant in spring and Exogone arenosa reached peak abundance in winter, both species being absent in autumn. Apseudes paulensis was only present and abundant in winter. In general, density was higher in winter and lower in autumn.

Discussion

The southeastern Brazilian inner shelf is occupied mainly by Coastal Water (CW), which tends to be vertically homogeneous due to mixing processes caused by wind stress and tidal shear. However, the shelf waters show a two-layered structure near mid depths in summer, with the South Atlantic Central Water (SACW) filling the bottom layer. In this season, prevailing winds usually blow up from the northeast quadrant, favouring an upwelling type of cross-shelf circulation. This situation brings SACW near the coast (Castro & Miranda, 1998). The detection of the SACW in the spring survey (station 8, 45 m) indicated an early intrusion of this water mass in the channel in 1993.

The homogeneity of temperature and salinity detected in the water column in winter probably contributed for the highest density of macrofauna observed, even though the abundance of individuals may change daily or weekly in a same period (Morrisey et al., 1992). Although the temporal scale adopted in this study does not adequately represent seasonal variation, we recorded a clear winter increase in density at all sites, including low-density stations 2, 7 and 10. The results point to higher values of organic matter in the depositional areas of stations 2, 7 and 10, as reported by Furtado (1995). Nonetheless, these values remain below those predicted for coastal areas (up to 2.0%; Rashid, 1985), which can be explained by the incipient fluvial drainage in the region, restricting organic carbon input (Furtado et al., 1998).

The three main regions in the channel were represented by characteristic species. One area had organisms typical of medium sand and high current speed such as Caecum species, Branchiostoma platae, Apanthura sp., Ophiactis lymani, Aspidosiphon gosnoldi, Exogone arenosa, Aricidea (Acmira) taylori, Parandalia americana and Chone insularis (group III). Parker (1975) observed the suspension-feeder gastropod Caecum pulchellum living in areas influenced by strong currents. He also observed that motile carnivores and suspension-feeder polychaetes, like those recorded in this group, were well developed in areas of medium sand, as also reported by Muniz & Pires (1999). Within group III, all these species were more abundant at station 9, located at the insular margin, during spring and winter. This site is located in a narrow and deep trench along the channel, and the ophiuroid Ophiactis lymani, a typical species of coarse sand, was found here in abundance. Although current velocities were not measured in the present study, hydrodynamics are the principal factor controlling sediment characteristics here (Furtado et al., 1998) and an important indirect cause of community patterns (Turner et al., 1995).

The spatial variation of sediment types in the south and central region of São Sebastião Channel can be related to the input of grains eroded from continental and insular margins. Other sediment sources are the suspended matter from the Juqueriquerê River (located in Caraguatatuba Bay, northern part of the channel) and the reworking of bottom sediments from the inner shelf (Furtado, 1995). The present study shows that drastic changes in hydrodynamics modify the size and shape of the sediment mosaics through movement and/or transport, a phenomenon also described by Gray (1974). These mosaics are ideal areas for higher infauna diversity, inducing the distribution of organisms in patches. Furthermore, these patches may be induced by the accumulation of organic matter, mainly in the muddy areas. The capability of sediment to adsorb organic matter is related to its clay content (Sharp, 1973), which in turn influences the feeding habitats of the fauna. In the São Sebastião Channel, the characteristic species found at the muddy continental margin were detritus-feeders such as Hemipholis elongata, Microspio pigmentata, Cirrophorus americanus and Neanthes species and carnivores such as Pinnixa sayana and Aspidosiphon albus.

The distributional pattern of the benthic macrofauna can also be indirectly related to anthropogenic influences. Stations 7 and 10, near the oil terminal, and station 2 always presented an impoverished fauna despite the high organic matter of the sediments. Aromatic and aliphatic petroleum hydrocarbons have been consistently measured in water and sediments in the channel and adjacent areas (Zanardi et al., 1999a, 1999b). According to these authors, hydrocarbon concentrations are generally low, but the amount of oil that the area receives during each event is significant. One event described by these authors occurred in May 1994, during the period of our sampling, spilling 2700 m³ of crude oil into the São Sebastião Channel and adjacent areas. As a result the bottom sediments were visibly impregnated with oil. The water rapidly recovered to normal concentrations, but the sediments still contained typical petroleum hydrocarbon 7 months later (Zanardi et al., 1999b). Migotto et al. (1993) argued that the continuous input of contaminants reduced the number of species and the abundance of molluscs in the São Sebastião littoral. The results obtained in stations 7 and 10 (situated near the DTCS terminal) confirmed this trend of lower density values and reduced species number. In station 2, located in the south mouth of the channel, the low density values might be due to clandestine tank washing by petroleum ships there (Zanardi et al., 1999b).

Although this study was not designed to detect anthropogenic and/or natural disturbance, the marine Biotic Coefficient (BC) indicated human stress in some areas of the channel, as on stations 2, 7 and 10. Heavy pollution is relatively easy to detect due to the presence of opportunistic species (Grall & Glémarec, 1997), but when incipient or slight pollution is present, as in the case of the São Sebastião Channel ecosystem, the diagnosis is difficult. Due to the complexity of this problem, further investigation of the area is urged, specifically where stress or anthropogenic disturbances have been detected.

No consistent seasonal pattern was observed for the overall data in the present study. Pires (1992) and Pires-Vanin (2001) found clear evidence of seasonality in the megabenthic fauna of Ubatuba and São Sebastião Channel. In both areas, the number of species and individuals peaked in summer, when SACW reached the inner shelf and coastal regions. Perhaps the dominance of CW throughout the sampling period can explain the lack of variation in the present study. The slight decrease of macrofaunal abundance in autumn may be linked to the high density of the penaeid shrimp Xiphopenaeus kroyeri at the channel area (Pires-Vanin, 2001). This species feeds on benthic macrofauna, especially the infauna (Pires, 1992), and polychaetes are generally a major food item for several penaeid species in marine shrimp culture systems, as pointed out by Nunes & Jay Parsons (2000).

Summary

The present results showed that the channel has a heterogeneous pattern of bottom sediment, which is reflected in the structure of the benthic macrofauna assemblages. Sediment type seems to be the main variable determining the distribution of macrobenthic species, and the marine Biotic Coefficient (BC) allowed detection at those regions of the channel subject to human stress. The use of trophic groups and species abundance data showed the same ordination of the sampling stations. Thus, for rapid evaluations, trophic groups may be useful and recommendable, especially when working with diverse and little-known tropical communities such as those of the São Sebastião Channel.