taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary,...

8
Taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary, Guanabara Bay, Southeast Brazil Carla Lima Torres Mendes a, * , Marcos Tavares b , Abı ´lio Soares-Gomes a a Departamento de Biologia Marinha, Universidade Federal Fluminense, Nitero ´ i, Rio de Janeiro 24010-970, Brazil b Museu Paulista de Zoologia, Universidade de Sa ˜o Paulo, Sa ˜o Paulo, SP 04263-000, Brazil Abstract Guanabara Bay (GB) is considered to be one of the most polluted environments of the southern Brazilian coastline. This typical estu- arine system is impacted by the heavy discharge of both industrial and domestic waste from the Rio de Janeiro metropolitan area. The mollusc community structure and distribution was investigated between 2000 and 2001, using a three month sampling design of 38 sta- tions, according to austral seasons. Species abundance was aggregated into progressively higher taxa matrices (genus, family, order) and were analysed using multivariate techniques. Mollusc distribution in GB varied significantly in space and time and was probably ruled by the organic enrichment effects of hypoxia and altered redox conditions coupled with prevailing patterns of circulation. Within the sectors of GB an increasing gradient in mollusc diversity and occurrence was observed, ranging from the azoic and impoverished stations in the inner sector to a well-structured community in terms of species composition and abundance inhabiting the outer sector. The non-metric multidimensional scaling (nMDS) and cluster analysis showed similar results when species were aggregated into genera and families, while greater difference occurred at coarser taxonomic identification (order). The literature about taxonomic sufficiency has demon- strated that faunal patterns at different taxonomic levels tend to become similar with increased pollution. In Guanabara Bay, an analysis carried out solely at family level is perfectly adequate to describe the ecophysiological stress. Further aggregation to order level changed the perceived patterns of differences. However, a different taxonomic resolution can be chosen depending on the type of ecological pat- terns investigated. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Taxonomic sufficiency; Macrobenthos; Pollution monitoring; Tropical estuary 1. Introduction Benthic communities are key indicators of estuarine environmental status, responding predictably to many kinds of natural and human-induced disturbances (Thomp- son et al., 2003; Venturini et al., 2004). However, as benthic community monitoring has been criticised as being time- consuming and expensive, both in terms of sampling effort in the field and laboratory analysis, there has been interest in examining ways of improving the cost-effectiveness of these studies. Therefore, a special focus has been given on the use of coarser levels of taxonomic resolution, if the loss of information is acceptable (Warwick, 1988; Ferr- aro and Cole, 1990; Gray et al., 1990; Thompson et al., 2003; De Biasi et al., 2003). Taxonomic identification to species or to the lowest pos- sible level is a common but labour intensive practice in monitoring programs (Ferraro and Cole, 1990). The sort- ing, identification and quantification stages, necessary to compile species abundances arrays, require a considerable degree of taxonomic expertise and familiarity with the local fauna. In addition to these time-consuming tasks, when several species within the genus are similar, the identifica- tion can be more error prone due to inaccuracy and impre- cision (Warwick, 1988). 0025-326X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpolbul.2006.08.026 * Corresponding author. Tel.: +55 21 3325 5986; fax: +55 21 3325 9341. E-mail addresses: [email protected] (C.L.T. Mendes), [email protected] (M. Tavares), [email protected]ff.br (A. Soares- Gomes). www.elsevier.com/locate/marpolbul Marine Pollution Bulletin 54 (2007) 377–384

Upload: carla-lima-torres-mendes

Post on 09-Sep-2016

216 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: Taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary, Guanabara Bay, Southeast Brazil

www.elsevier.com/locate/marpolbul

Marine Pollution Bulletin 54 (2007) 377–384

Taxonomic sufficiency for soft-bottom sublittoral mollusksassemblages in a tropical estuary, Guanabara Bay, Southeast Brazil

Carla Lima Torres Mendes a,*, Marcos Tavares b, Abılio Soares-Gomes a

a Departamento de Biologia Marinha, Universidade Federal Fluminense, Niteroi, Rio de Janeiro 24010-970, Brazilb Museu Paulista de Zoologia, Universidade de Sao Paulo, Sao Paulo, SP 04263-000, Brazil

Abstract

Guanabara Bay (GB) is considered to be one of the most polluted environments of the southern Brazilian coastline. This typical estu-arine system is impacted by the heavy discharge of both industrial and domestic waste from the Rio de Janeiro metropolitan area. Themollusc community structure and distribution was investigated between 2000 and 2001, using a three month sampling design of 38 sta-tions, according to austral seasons. Species abundance was aggregated into progressively higher taxa matrices (genus, family, order) andwere analysed using multivariate techniques. Mollusc distribution in GB varied significantly in space and time and was probably ruled bythe organic enrichment effects of hypoxia and altered redox conditions coupled with prevailing patterns of circulation. Within the sectorsof GB an increasing gradient in mollusc diversity and occurrence was observed, ranging from the azoic and impoverished stations in theinner sector to a well-structured community in terms of species composition and abundance inhabiting the outer sector. The non-metricmultidimensional scaling (nMDS) and cluster analysis showed similar results when species were aggregated into genera and families,while greater difference occurred at coarser taxonomic identification (order). The literature about taxonomic sufficiency has demon-strated that faunal patterns at different taxonomic levels tend to become similar with increased pollution. In Guanabara Bay, an analysiscarried out solely at family level is perfectly adequate to describe the ecophysiological stress. Further aggregation to order level changedthe perceived patterns of differences. However, a different taxonomic resolution can be chosen depending on the type of ecological pat-terns investigated.� 2006 Elsevier Ltd. All rights reserved.

Keywords: Taxonomic sufficiency; Macrobenthos; Pollution monitoring; Tropical estuary

1. Introduction

Benthic communities are key indicators of estuarineenvironmental status, responding predictably to manykinds of natural and human-induced disturbances (Thomp-son et al., 2003; Venturini et al., 2004). However, as benthiccommunity monitoring has been criticised as being time-consuming and expensive, both in terms of sampling effortin the field and laboratory analysis, there has been interestin examining ways of improving the cost-effectiveness of

0025-326X/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.marpolbul.2006.08.026

* Corresponding author. Tel.: +55 21 3325 5986; fax: +55 21 3325 9341.E-mail addresses: [email protected] (C.L.T. Mendes),

[email protected] (M. Tavares), [email protected] (A. Soares-Gomes).

these studies. Therefore, a special focus has been givenon the use of coarser levels of taxonomic resolution, ifthe loss of information is acceptable (Warwick, 1988; Ferr-aro and Cole, 1990; Gray et al., 1990; Thompson et al.,2003; De Biasi et al., 2003).

Taxonomic identification to species or to the lowest pos-sible level is a common but labour intensive practice inmonitoring programs (Ferraro and Cole, 1990). The sort-ing, identification and quantification stages, necessary tocompile species abundances arrays, require a considerabledegree of taxonomic expertise and familiarity with the localfauna. In addition to these time-consuming tasks, whenseveral species within the genus are similar, the identifica-tion can be more error prone due to inaccuracy and impre-cision (Warwick, 1988).

Page 2: Taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary, Guanabara Bay, Southeast Brazil

378 C.L.T. Mendes et al. / Marine Pollution Bulletin 54 (2007) 377–384

Indeed, Gray et al. (1988) pointed out that there is aconsiderable degree of redundancy in species data and con-cluded that aggregation into higher taxonomic levels wouldreduce this problem, preventing natural variables frommasking the effects of pollution. Accordingly, Warwick(1988) suggested that pollution effects are detectable atcoarser taxonomic groups as natural variability affectscommunity structure mainly by species replacement. Manyother studies provided the best evidence to date for the tax-onomic resolution of higher taxa in pollution studies, spe-cially regarding the family-level identification (Somerfieldand Clarke, 1995; Olsgard et al., 1997; Olsgard et al.,1998; Ferraro and Cole, 1990; Thompson et al., 2003;Gesteira et al., 2003; Dauvin et al., 2003). These studiesdemonstrated the equivalency of pollution assessments atthe species to family level for several macrobenthic datasets, suggesting that identification to the level of familymay be satisfactory in many routine monitoring surveys.

Therefore, to increase the cost-effectiveness of a moni-toring survey all components must be optimised whilemaintaining the ability to reliably detect an impact.Consideration of taxonomic resolution and also the sieve-mesh-size should be made in conjunction with other sam-pling design variables such as number of replicates, samplersize, and sample unit size (Thompson et al., 2003) whenseeking a statistically rigorous, cost-effective study designsufficient to meet pollution assessment objectives.

The aim of this study was to investigate the taxonomicresolution sufficiency for the sublittoral mollusk assem-blage of Guanabara Bay, Rio de Janeiro State, southernBrazil. The present research is a contribution to what con-cerns the taxonomic sufficiency approach in tropical envi-ronments, once previous studies have been conductedmostly in temperate areas worldwide.

2. Study area

Guanabara Bay (GB), located at 23� 50 0 S, 43� 08 0 W, isa 384 Km2 eutrophic coastal bay in southeast Brazil(Kjerfve et al., 1997). It is considered to be one of the mostpolluted environments of the southern Brazilian coastline(Carreira et al., 2002). Due to the rapid degradation of thisecosystem, the bay represents an important focus of envi-ronmental interest but little information is available aboutthe local benthos. The ongoing socio-economic pressureupon GB requires updated knowledge of its ecosystemsstructure and function.

This typical estuarine system is impacted by the heavydischarge of both industrial and domestic waste from theRio de Janeiro metropolitan area. Its drainage basinreceives polluted effluents from 24 sub basins with about6000 industries, two airports, two harbours and 15 oil ter-minals located in its vicinities (Kjerfve et al., 1997). Apartfrom the residues of this large industrial park, the sewageof an increasing population of 10 million inhabitants is alsoreleased into the bay’s water through about 45 rivers, 6 ofwhich are responsible for 85% of the total runoff (Kjerfve

et al., 2001). As a result, 75% of the organic wastes origi-nate from urban untreated sewage and 25% from indus-tries. Further, high inputs of heavy metals, petroleumhydrocarbons, pesticides (Xavier et al., 2002) and othertoxic chemical compounds (Ventura et al., 2002) enter thebay daily, especially in the inner portion, accumulating inthe bottom sediments. However, these pollutants seemsto be buried in anoxic sediments and are therefore, unavail-able for biological uptake (Carvalho and Lacerda, 1992).

The bay experiences a large spatial and temporal vari-ability of its water quality, mainly caused by circulationpatterns and pollution foci (Kjerfve et al., 1997). Circula-tion is controlled by tides and winds, allowing water inflowfrom the ocean through the bottom layers. As described byKjerfve et al. (2001), BG has a complex bathymetry with a400 m wide central channel, which stretches from themouth (1.6 m wide) more than 5 km into the bay, and isdefined by the 30 m isobath. The channel rapidly becomesshallower further into the bay, reaching to a maximum 1 mdepth in the inner portion. Because of the rapid mud sedi-mentation, the mean depth measures only 5.7 m.

According to Kjerfve et al. (2001), the worst water qual-ity is indicated by average faecal coliform counts higherthan 1000 ml�1 and by the average chlorophyll concentra-tion exceeding 130 lg l�1 in the inner bay, the most criticalzone, in response to high nutrient loading. Sediments arenot evenly distributed at the bottom, predominating mudat the inner areas and fine sand near the mouth (JICA,1994). As a result of increase pollution levels and poor cir-culation towards the inner region, the sediments in thisarea presents reducing conditions, being considered anoxicduring the wet season.

3. Material and methods

3.1. Sampling design

The sublittoral soft-bottom molluscs of GB were inves-tigated through four oceanographic surveys conductedbetween the years of 2000 and 2001. It was performed atthree month interval sampling design of 38 stations,according to austral seasons. The sampling stations weregrouped into subsets and named as: inner sector (stations1–14), intermediary sector (stations 15–26), and outer sec-tor (stations 27–38) in relation to the sampling design ofprevious works in the area, natural hydrodynamic charac-teristics and pollution intensity (Fig. 1).

At each station a 0.1 m2 van-Veen grab sample wastaken in triplicate. Sediment samples were analysed forgrain size distribution, organic matter, carbonate contentand oxi-redox potential. The methods of mechanical drysieving and decantation described by Suguio (1973) wereused to determine the grain size fractions. The percentageof total organic matter was determined by loss of masson ignition and sediments samples were oven-dried at105 �C for 12 h and subsequently ashed at 500 �C for 2 h.Biodetritic carbonate (CaCO3) was obtained by HCl 10%

Page 3: Taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary, Guanabara Bay, Southeast Brazil

Fig. 1. Geographical location of Guanabara Bay in Rio de Janeiro State, southern Brazil, and sampling stations within the studied area.

C.L.T. Mendes et al. / Marine Pollution Bulletin 54 (2007) 377–384 379

attack. The oxi-redox potential was measured through anAnalyser platinum electrode model 6 A05/AG.

Biological samples were sieved out through a 1.0 mmmesh size and the mollusks sorted under stereomicroscopefor identification to the finest possible level. The faunawas characterized by its abundance, diversity (Shannon–Wiener), evenness (Pielou) and richness.

3.2. Data analysis

Species abundance was aggregated into progressivelyhigher taxa matrices (genus, family, order) after calculatingthe mean abundance values for all four surveys carried out.For multivariate analyses, non-metric multidimensionalscaling (MDS), a technique widely employed in taxonomicsufficiency studies (Warwick, 1988), and a cluster analysis,using WPGMA method, were carried out separatelyfor each taxonomic level. Azoic stations and stationsrepresented by only one species were not included in theanalyses. Various data transformations (untransformedabundance data, square root, forth root, log (x + 1) andpresence/absence) have been applied and the Bray–Curtissimilarity index was used throughout (Olsgard et al.,1998). Although most of the ordination plots and stressvalues yielded similar results, the square root transforma-tion was chosen by its best representation of the abundancedata. K-dominance curves for abundance data were alsoplot for each taxonomic level according to established sec-tors in GB.

Formal significance tests for difference between theouter, intermediary and inner sampled stations were per-

formed using a one-way analysis of similarity (ANOSIM)permutation test (Clarke, 1993).

Correlations between similarity matrices from speciesabundances and matrices derived from abundances ofhigher taxonomic levels were calculated using the Spear-man rank correlation and the significance of the correla-tions was determined by a permutation procedure, usingthe PRIMER software routine RELATE (Thompsonet al., 2003). In addition, a cost/benefit ratio (CBL) was cal-culated as to identify the taxonomic level with the least lossof information and the least taxonomic effort. The CBL

index, originally introduced by Karakassis and Hatziyanni(2000) and successfully applied by Muniz (2002), is exp-ressed by:

CBL ¼1� qL

ðS � tLÞ=S

qL Spearman rank correlation between the level taxo-nomic L and the specific.

tL number of taxa present in the level L.

S number of species.

4. Results

The mean abundance of mollusks per sector of GB ran-ged from 483 individuals in the inner sector (52% bivalves),1.038 individuals in the intermediary sector (68% gastro-pods) to 1.540 individuals in the outer sector (50% each).According to Table 1, diversity values (H 0) were eitherinfluenced by evenness (J 0) or richness (S), not revealing

Page 4: Taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary, Guanabara Bay, Southeast Brazil

Table 1Values of Shannon–Wiener Diversity (H 0), Eveness (J 0) and richness (S) for each sampling station (St.) at different seasons

Very light gray represents the inner stations; Light gray represents the intermediary stations; and dark gray represents the outer ones. Azoic stations orthose represented by one species only were omitted from the table.

380 C.L.T. Mendes et al. / Marine Pollution Bulletin 54 (2007) 377–384

a clear pattern of distribution of these biological descrip-tors within the sectors. High diversity was observed inthe outer sector, especially in the dry season.

The K-dominance plots (Fig. 2) revealed the same pat-tern of mollusk diversity among the bay’s sectors at species,genera and family levels. The inner sector is the leastdiverse while a gradual increasing diversity is observedfrom the intermediary to the outer sector.

Inner

Intermediary

Outer

Cum

ulat

ive

Dom

inan

ce %

Cum

ulat

ive

Dom

inan

ce %

Order rank

0

20

40

60

80

1 10

Inner

Intermediary

Outer

Genus rank

0

20

40

60

80

100

100

1 10 100

Fig. 2. K-dominance curves for mollusks

The species abundance matrix contained 32 mollusksspecies grouped into 24 genus, 18 families and 7 orders-the largest changes in the number of taxonomic units beingbetween the levels of family and order. MDS ordination ofmatrices, derived from square root transformed abun-dances of various taxonomic levels distributed into 25 sta-tions, appear similar at the levels of species, genus andfamily (Fig. 3). The separation between the outer and the

Cum

ulat

ive

Dom

inan

ce %

Inner

Intermediary

Outer

Family rank

0

20

40

60

80

100

1 10 100

Species rank

0

20

40

60

80

100

1 10 100

Cum

ulat

ive

Dom

inan

ce %

Inner

Intermediary

Outer

abundance at each taxonomic level.

Page 5: Taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary, Guanabara Bay, Southeast Brazil

37 38 35 27 36 29 339

26 34 6 10 19 23 12 7 31 14 32 28 25 21 22 20 24

100

80

60

40

20

0

Sim

ilarit

y

37 38 6 23 25 21 22 19 20 24 7 10 12 31 14 32 29 35 9 33 26 34 28 27 36

100

80

60

40

20

0

Sim

ilarit

y

37 38 9 33 26 34 35 27 36 28 29 6 10 19 23 25 21 22 20 24 12 7 31 14 32

100

80

60

40

20

0

Sim

ilarit

y

6 37 23 38 10 7 14 32 12 31 26 34 9 33 29 35 19 25 21 22 27 36 28 20 24

100

80

60

40

Sim

ilarit

y

6

7

9

10

12

14

19

20

2122

23

242526

27

28

29

3132

33

34

35

36

37

38

Stress: 0.13

6

7

910

12

14

19

20

2122

23

2425

26

27

2829

31

32

3334

3536

37 38

Stress: 0.14

6

7

910

12

14

19

20

2122

23

2425

26

27

2829

31

32

3334

3536

37 38

Stress: 0.14

6

7 9

1012

14

19

20

2122

23

2425 2627

28

29

31

32

3334

35 36

37

38

Stress: 0.14

Fig. 3. MDS ordinations and cluster dendograms of square root transformed mollusks abundances at different taxonomic levels of (a) species, (b) genus,(c) family and (d) order.

C.L.T. Mendes et al. / Marine Pollution Bulletin 54 (2007) 377–384 381

intermediary stations is slightly more evident in the species-level plot than at family level, suggesting that the overallpatterns of community structure are retained as abun-dances are aggregated to higher taxonomic levels. Stressvalues maintained relatively low (0.13–0.14) and almostconstant along the different taxonomic levels.

Hierarchical cluster analysis (Fig. 3) also recognized,although roughly, the separation of stations into an outerand an intermediary in all level plots. The similarity values

Table 2ANOSIM test results at different taxonomic levels (TL)

(GR) Global-R; (SL) Significance level of global-R; (SG) Sampling groups; (odark gray are highlighted the statistic levels (%) of R statistic values (in light

in station clustering at species level are very similar to thoseat family level (17–20). When taxa were identified to order,the separation of stations was less apparent and not asrobust as identification to family or species.

The Spearman rank correlations between similaritymatrices for abundance data were consistently higher than0.9 between species-genus (0.97) and species-family (0.96),decreasing to 0.67 at species-order. The CBL indexesreflected the same pattern, corroborating with the findings

ut) outer stations; (inter) intermediary stations; (inner) inner stations. Ingray).

Page 6: Taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary, Guanabara Bay, Southeast Brazil

382 C.L.T. Mendes et al. / Marine Pollution Bulletin 54 (2007) 377–384

previously presented. There is little loss of informationfrom species to family levels (0.08) while much informationis lost at the coarser level of taxonomic resolution (0.417).

Results from the ANOSIM test, comparing sectors atvarious taxonomic levels, had similar global R values,increasing slightly from species to family levels (Table 2).Groupings became less clear when taxa were aggregatedinto order. In agreement with the MDS ordination andcluster results, the outer and the intermediary sectors wereconsistently dissimilar in all four taxonomic levels. As tax-onomic resolution decreased, the inner and the intermedi-ary sectors became even more similar.

5. Discussion

Previous studies in taxonomic sufficiency (TS) have beenconducted in relatively shallow marine environments andmost of them were undertaken in temperate areas. Thesehave found that using results of higher taxonomic levelsallowed the determination of the response of the benthiccommunity to environmental or pollution gradients (e.g.:Gesteira et al., 2003; Narayanaswamy et al., 2003; Thomp-son et al., 2003). In the tropics, TS surveys of soft bottombenthic communities are scarce (e.g.: Olsgard et al., 1998;Guzman and Carrasco, 2005) and were expected not toshow the same patterns and effectiveness as for temperateimpacted areas. As far as diversity is concerned, increasingspecies follows a latitudinal gradient from high latitudes tolow latitudes and this has been described for terrestrial,freshwater, and marine ecosystems, both in the northernand southern hemisphere (Clarke and Crame, 1998; Gray,1998; Willig et al., 2003; Hillebrand, 2004). However, dueto differential species’ responses to variable conditionsand perturbations, some debate continues to surround con-troversies over this broad latitudinal gradients speciesrichness.

High diversity communities harbor a great number ofredundant species that, theoretically, are able to expandtheir roles in the ecosystem to compensate for those specieswhich are becoming extinct due to natural or man-madeperturbation. Guzman and Carrasco (2005) found such evi-dence when investigating a tropical macroinfaunal assem-blage of the Southern Caribbean, which exhibited a highdegree of structural redundancy. The results show that,on analyzed temporal and spatial scale, the pattern wasmaintained as data were aggregated to families. The sub-sets of genera acted as ‘‘response units’’, comprising thesame or different taxonomic groups and, therefore, appear-ing to be functioning equivalent to another ‘‘responseunit’’. The authors concluded that these units respondedin a similar way to environmental changes and, therefore,higher taxonomic levels (TL) would be adequate for suffi-ciency studies in the region.

In the present study, the non-metric multidimensionalscaling (NMDS) and hierarchical cluster analysis showedsimilar results when species were aggregated into generaand families, while a certain loss of information occurred

at coarser taxonomic identification (order). Low stress val-ues indicated satisfactory reliability of the results (Gesteiraet al., 2003). Furthermore, strong correlations and low CBL

indexes were observed between the similarity matrices atthe species, genera and family levels. Aggregation abovethe family level reduced the ability to reliably distinguishbetween groups of stations and much loss of informationwas observed at the order level according to the CBL indexcalculated. A consistent explanation can be drawn from thefact that most mollusc families were represented by a singlespecies and thus, high correlation values and very similarrepresentation plots from both taxonomic levels are to beexpected. Small numbers of species per genera and per fam-ily have likewise been observed in several previous studiesof this kind (e.g.: Warwick, 1988; Ferraro and Cole,1990; Gesteira et al., 2003; Guzman and Carrasco, 2005)and therefore found that multivariate analysis at genusand family level data is no less effective for identifying sam-ple groups than ordination of the species level data. In fact,one of the main requirements for taxonomic sufficiencyapproaches to be effective is the low species/genera and spe-cies/family ratios (Somerfield and Clarke, 1995). However,the taxonomic resolution also seems to work out success-fully when dealing with higher species/genera and species/family ratios, as revealed by most studies in the area(e.g.: Fiori and Soares-Gomes, 2001; Dauvin et al., 2003;De Biasi et al., 2003; Olsgard et al., 2003). The data fromthe Norwegian coast obtained by Olsgard et al. (2003)comprised 517 polychaete species grouped into 278 genera,58 families and 25 orders. The results therein add weigh tothe usefulness of the taxonomic resolution approach.

The ANOSIM test was not able to significantly discrim-inate stations as abundance data was aggregated to orderlevel. However, the outer and intermediary sectors of GBwere clearly dissimilar at lower TL. For the levels of spe-cies, genera and family, a general decrease in the signifi-cance level of global R was observed, indicating a moreconsistent result at family level. These results corroboratewith those of other multivariate analysis previouslydescribed (e.g.: Olsgard et al., 1997).

The literature about taxonomic sufficiency has demon-strated that faunal patterns at different taxonomic levelstend to become similar due to increased pollution- the hier-archic-response-to-stress theory (Ferraro and Cole, 1990).According to Warwick (1988), anthropogenic effects mod-ify community composition at a higher taxonomic levelthan natural environmental variables, which influence thefauna more by species replacement than by changes inthe proportions of major taxa present. As stress increases,the adaptability of first the individual, then the species,genera, and family is exceeded (Ferraro and Cole, 1990).Consequently, impacts resulting from increasing stress aremanifested at higher and higher levels of biological organi-zation and, therefore, taxonomic sufficiency studies arepotentially useful tools in assessing this differentialresponse, separating pollution effects from those of naturalvariables. Olsgard et al. (1998) and Guzman and Carrasco

Page 7: Taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary, Guanabara Bay, Southeast Brazil

C.L.T. Mendes et al. / Marine Pollution Bulletin 54 (2007) 377–384 383

(2005) suggested that if a gradient in the environment orpollution reduces the number of species in each family toa single species, then analysis of these samples could beundertaken at family level, as the outcome would be iden-tical to that of species level analysis.

The results presented herein supports evidence to statethat BG is going through the same processes and conse-quences described for highly polluted environments intemperate areas. Mollusc distribution in GB varied signifi-cantly in space and time and was probably controlled bythe organic enrichment effects of hypoxia and altered redoxconditions coupled with prevailing patterns of circulation.GB has been a focus of heavy pollution discharge for overdecades (Carreira et al., 2002), specially in the inner sector.Along this period many environmental and pollution gra-dients have been established. The strength of such gradi-ents has undoubtly influenced the local taxonomicalsufficiency in that redundant species were probably elimi-nated, and therefore, investigations based on higher TLare likely to be of use in describing the main pattern offaunal composition. Fiori and Soares-Gomes (2001) havealso found a use of TS in a tropical continental shelf, wheregradients were considerable though not as well establishedas for GB.

In GB, an analysis carried out solely at family level isperfectly adequate to describe the ecophysiological stressgradient (De Biasi et al., 2003) and no extra informationwhatsoever would have been gained by analyzing to spe-cies. This conclusion is in line with that of most previouslypublished papers on TS approaches in marine macrobentosstudies (e.g.: Warwick, 1988; Ferraro and Cole, 1990; Grayet al., 1990; Karakassis and Hatziyanni, 2000; Fiori andSoares-Gomes, 2001; Dauvin et al., 2003; De Biasi et al.,2003; Gesteira et al., 2003; Thompson et al., 2003). Work-ing at the family level would allow sampling of additionalreplicates, control locations or times of sampling (Thomp-son et al., 2003) while obviating nearly all of the time-con-suming taxonomic constraints on the sorting of samples.Thereby, increasing the statistical power to detect animpact. On the other hand, levels higher than family havegiven useful results in multivariate ordinations in somestudies (Somerfield and Clarke, 1995; Olsgard et al.,1997, 1998; Venturini et al., 2004). Our findings have gen-eral significance in that they support accumulative evidenceindicating that taxonomic sufficiency is a useful concept inpollution ecology.

The decision about the resolution to which organismsmust be identified depends on the precise aims of the study.Therefore, when the above findings are taken into consid-eration together with the paucity of previous work, we con-clude that for the purpose of detecting change in generaldistribution pattern and pollution effects of molluscs, iden-tification of molluscs to family level is sufficient. However,a different taxonomic resolution can be chosen dependingon the type of ecological patterns investigated (e.g. compi-lation of species lists, temporal drifts in communityidentity).

This study has not only contributed to the knowledge ofdiversity, especially molluscs, in the region (GB) but hasalso ecological relevance for soft bottom macrobenthosstudies from similar areas, i.e. semi-enclosed polluted bayswith impacts of hypoxia and sedimentation linked to ariver discharge in its innermost part. However, it has beenrecommended that additional investigations comparing dif-ferent taxonomic levels and other taxonomically difficultgroups (e.g.: polychaete, crustaceans) should be under-taken, not only in GB but also in the tropical ecosystemsas a whole, so that all components of an effective pollutionmonitoring survey can be optimized for these areas.

Acknowledgements

Financial support for this study was granted by CAPESand FAPERJ, for which we are very grateful. The authorsare greatly indebted to PhD. Ricardo S. Absalao, FranklinNoel dos Santos, Alexandre Dias Pimenta and Carlos Hen-rique Soares Caetano from Universidade Federal do Riode Janeiro, for providing assistance in the mollusks identi-fication. We also would like to thank Gabriela Benkendor-fer for helping in the gastropods identification.

References

Carreira, R.S., Wagener, A.L.R., Readman, J.W., Fileman, T.W., Macko,S.A., Veiga, A., 2002. Changes in the sedimentary organic carbon poolfor a fertilized tropical estuary, Guanabara Bay, Brazil: an elemental,isotropic and molecular marker approach. Marine Chemistry 79, 207–227.

Carvalho, C.E.V., Lacerda, L.D., 1992. Heavy metals in the GuanabaraBay biota: Why such low concentrations? Journal of the BrazilianAssociation for the Advancement of Science 44 (2/3), 184–186.

Clarke, K.R., 1993. Non-parametric multivariate analyses of changes incommunity structure. Australian Journal of Ecology 18, 117–143.

Clarke, A., Crame, J.A., 1998. Diversity, latitude and time: Patterns in theshallow sea. In: Ormond, Gage, Angel (Eds.), Marine Biodiversity.Cambridge University Press, pp. 122–147, Chapter 6.

Dauvin, J.C., Gesteira, J.L.G., Fraga, M.S., 2003. Taxonomic sufficiency:an overview of its use in the monitoring of sublittoral benthiccommunities after oil spills. Marine Pollution Bulletin 46, 552–555.

De Biasi, A.M., Bianchi, C.N., Morri, C., 2003. Analysis of macrobenthiccommunities at different taxonomic levels: an example from anestuarine environment in the Ligurian Sea (NW Mediterranean).Estuarine, Coastal and Shelf Science 58, 1–8.

Ferraro, S.P., Cole, F.A., 1990. Taxonomic level and sample size sufficientfor assessing pollution impacts on the Southern California Bightmacrobenthos. Marine Ecology Progress Series 67, 251–262.

Fiori, C.S., Soares-Gomes, A., 2001. Taxonomic sufficiency for amonitoring program in a tropical continental shelf, Rio de Janeiro.Contaminated Soil Sediment and Water, 34–39.

Gesteira, J.L.G., Dauvin, J.C., Fraga, M.S., 2003. Taxonomic level forassessing oil spill effects on soft-bottom sublittoral benthic communi-ties. Marine Pollution Bulletin 46, 562–572.

Gray, J.S., Aschan, M., Carr, M.R., Clarke, K.R., Green, R.H., Pearson,T.H., Rosenberg, R., Warwick, R.M., 1988. Analysis of communityattributes of the benthic macrofauna of Frierfjord/Langensundfjordand in a mesocosm experiment. Marine Ecology Progress Series 46,151–165.

Gray, J.S., Clarke, K.R., Warwick, R.M., Hobbs, G., 1990. Detection ofinitial effects of pollution on marine benthos: an example from the

Page 8: Taxonomic sufficiency for soft-bottom sublittoral mollusks assemblages in a tropical estuary, Guanabara Bay, Southeast Brazil

384 C.L.T. Mendes et al. / Marine Pollution Bulletin 54 (2007) 377–384

Ekofisk and Elfisk oilfields, North Sea. Marine Ecology Progress Series66, 285–299.

Gray, J.S., 1998. Gradients in marine biodiversity. In: Ormond, Gage,Angel (Eds.), Marine biodiversity. Cambridge University press, pp.18–34, Chapter 2.

Guzman, A.I., Carrasco, F., 2005. Taxonomic aggregation and redun-dancy in a tropical macroinfaunal assemblage of the southernCaribbean in the dection of temporal patterns. Scientia Marina 69(1), 133–141.

Hillebrand, H., 2004. Strength, slope and variability of marine latitudinalgradients. Marine Ecology Progress Series 273, 251–267.

Jica, 1994. In: The study on Recuperation of the Guanabara Bayecosystem, 8 vols. Japan International Cooperation Agency (JICA).Kokusai Kogyo Co., Ltd., Tokyo.

Karakassis, I., Hatziyanni, E., 2000. Benthic disturbance due to fishfarming analysed under different levels of taxonomic resolution.Marine Ecology Progress Series 203, 247–253.

Kjerfve, B., Ribeiro, C.H.A., Dias, G.T.M., Filipo, A.M., Quaresma,V.S., 1997. Oceanographic characteristics of an impacted coastal bay:baıa de Guanabara, Rio de Janeiro, Brazil. Continental Shelf Research17 (13), 1609–1643.

Kjerfve, B., Lacerda, L.D., Dias, G.T.M., 2001. Baıa de Guanabara, Riode Janeiro, Brazil. In: Seeliger, Kjerfve (Eds.), Coastal marineecosystems of Latin America. Springer-Verlag, Berlin, pp. 107–117.

Narayanaswamy, B.E., Nickell, T.D., Gage, J.D., 2003. Appropriate levelsof taxonomic discrimination in deep-sea studies: species vs family.Marine Ecology Progress Series 257, 59–68.

Olsgard, F., Somerfield, P.J., Carr, M.R., 1997. Relationships betweentaxonomic resolution and data transformations in analysis of amacrobenthic community along an established pollution gradient.Marine Ecology Progress Series 149, 173–181.

Olsgard, F., Somerfield, P.J., Carr, M.R., 1998. Relationships betweentaxonomic resolution, macrobenthic community patterns and distur-bance. Marine Ecology Progress Series 172, 25–36.

Olsgard, F., Brattegard, T., Holthe, T., 2003. Polychaetes as surrogates formarine biodiversity: lower taxonomic resolution and indicator groups.Biodiversity and Conservation 12, 1033–1049.

Somerfield, P.J., Clarke, K.R., 1995. Taxonomic level, in marinecommunity studies, revisited. Marine Pollution Bulletin 46, 562–572.

Suguio, K., 1973. Introducao a sedimentologia. Sao Paulo: EdgardBlucher/EDUSP. pp. 317.

Thompson, B.W., Riddle, M.J., Stark, J.S., 2003. Cost-efficient methodsfor marine pollution monitoring at Casey Station, east Antarctica: thechoice of sieve mesh-size and taxonomic resolution. Marine PollutionBulletin 46, 232–243.

Ventura, E.C., Gaelzer, L.R., Zanette, J., Marques, M.R.F., Bainy,A.C.D., 2002. Biochemical indicators of contaminant exposure inspotted pigfish (Orthopristis ruber) caught at three bays at Rio deJaneiro coast. Marine Environmental Research 54, 775–779.

Venturini, N., Muniz, P., Rodrıgues, M., 2004. Macrobenthic subtidalcommunities in relation to sediment pollution: the phylum-level meta-analysis approach in a south-eastern coastal region of South America.Marine Biology 144, 119–126.

Warwick, R.M., 1988. The level of taxonomic discrimination required todetect pollution effects on marine benthic communities. MarinePollution Bulletin 19, 259–268.

Willig, M.R., Kaufman, D.M., Stevens, R.D., 2003. Latitudinal gradientsof biodiversity: patterns, process, scale, and synthesis. Annual Reviewof Ecology, Evolution and Systematics 34, 273–309.

Xavier De Brito, A.P., De Andrade Bruning, I.M.R., Moreira, I., 2002.Chlorinated pesticides in mussels from Guanabara Bay, Rio deJaneiro, Brazil. Marine Pollution Bulletin 44, 71–81.