Do multivariate analyses incorporating changes in pattern across taxonomic levels reveal anthropogenic stress in Mediterranean lagoons?

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright

Author's personal copy Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

Contents lists available at ScienceDirect

Journal of Experimental Marine Biology and Ecology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / j e m b e

Do multivariate analyses incorporating changes in pattern across taxonomic levels reveal anthropogenic stress in Mediterranean lagoons? Christos Arvanitidis a,⁎, Paul John Somerfield b, Georgios Chatzigeorgiou c, Sofia Reizopoulou d, Theodoros Kevrekidis e, Anastasios Eleftheriou a a

Institute of Marine Biology of Crete, Hellenic Centre for Marine Research, Heraklion, 71003, Crete, Greece Plymouth Marine Laboratory, Prospect Place, The Hoe, Plymouth PL1 3DH, United Kingdom Biology Department, University of Crete, Vasilika Vouton, 71409 Heraklio, Crete, Greece d Institute of Oceanography, Hellenic Centre for Marine Research, 19013, Anavyssos Attikis, Greece e Laboratory of Environmental Research and Education, Democritus University of Thrace, 68100, Alexandroupolis, Greece b c

a r t i c l e

i n f o

Article history: Received 7 July 2006 Received in revised form 29 October 2008 Accepted 31 October 2008 Keywords: Biodiversity Biodiversity MDS Delta MDS Lambda MDS Mediterranean, lagoons Number of taxa MDS Taxonomic distinctness

a b s t r a c t It is accepted that observed patterns in community structure change as analyses are carried out at higher taxonomic levels. Univariate analyses which incorporate higher taxonomic structure within assemblages have been shown to be informative. In this paper we suggest ways in which changes in multivariate relationships at higher taxonomic levels and associated with higher taxonomic/phylogenetic structure of the community may be incorporated into multivariate analyses, an aspect never occurred before in this type of analysis. Four approaches, namely: biodiversity MDS (bdMDS), number of taxa MDS (ntMDS), delta MDS (δMDS) and lambda MDS (λMDS), are proposed, and applied to theoretical data as well as to data collected from the literature on the Mediterranean lagoonal environment. Results show that these approaches have the capacity to distinguish severely impacted lagoons from naturally disturbed ones, although in practice the simplest method (ntMDS) was the most successful. Analyses based on the most abundant groups (polychaetes, molluscs, crustaceans) did not always match analyses based on the entire macrofauna, mirroring the performance of taxonomic distinctness indices in the Mediterranean lagoons. The important characteristics of the approaches introduced, as well as potential criticisms are provided. Application of these techniques on smaller scales and to other habitats, is suggested prior to their wider use in the region. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Major global environmental concerns, such as the biodiversity crisis (genomes-species-habitats extinction) and global change, have brought the need to assess and conserve biodiversity to the fore in many international fora (e.g. Gray, 1997). Governmental commitments under a number of European and International treaties and conventions, limited resources, and the rapidity with which environmental managers and decision-makers need biodiversity information, have necessitated the development and application of various Rapid Assessment Techniques (RATs) in the field of marine science (e.g. Féral et al., 2003). Many techniques for measuring changes in biodiversity have been developed over the years (see for example Clark, 1997; Gray, 2000; Clarke and Warwick, 2001; Magurran, 2004). Though these techniques are generally applied to species information, several studies have shown that information about the distribution of higher taxa may be useful in

⁎ Corresponding author. Tel.: +30 2810 337748; Fax: +30 2810 337822. E-mail address: [email protected] (C. Arvanitidis). 0022-0981/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jembe.2008.10.032

studies of marine biodiversity and environmental issues (Warwick, 1988; Warwick et al., 1990; Ferraro and Cole, 1995; Somerfield and Clarke, 1995; Olsgard et al., 1997, 1998; Warwick and Clarke, 2001; Warwick and Light, 2001; Féral et al., 2003). Two approaches employing higher taxonomic information are commonly used in marine benthic ecology: (1) analyses performed at taxonomic levels higher than species (Warwick, 1988; Warwick et al., 1990; Ferraro and Cole,1995; Somerfield and Clarke, 1995; Olsgard et al., 1997, 1998); (2) analyses using biodiversity indices which include information about the phylogenetic/taxonomic relationships between taxa, such as average taxonomic distinctness (Δ+) and variation in taxonomic distinctness (Λ+) (Clarke and Warwick, 1998a,b; Izsak and Price, 2001; Warwick and Clarke, 2001; Clarke et al., 2006). It is often postulated that the taxonomic level required to meet the objectives of an environmental monitoring study with the least possible cost and time spends is higher than the species level (Warwick, 1988; Kingston and Riddle, 1989; Ferraro and Cole, 1995). This is because identification of organisms to the species level may not always be necessary to describe spatial patterns, especially when patterns are strong such as along established pollution gradients (Pearson and Rosenberg, 1978; Boesch and Rosenberg, 1981; Warwick,

Author's personal copy C. Arvanitidis et al. / Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

1988; Ferraro and Cole, 1990; Olsgard et al., 1997, 1998). Multivariate analyses deriving from higher phylogenetic/taxonomic categories may reflect gradients of contamination more closely than those based on species abundances, which are considered to be more affected by natural variability (the “hierarchic-response-to-stress” hypothesis) and therefore the taxonomic level most highly correlated with environmental variables indicative of anthropogenic impacts will increase with increasing levels of stress (Olsgard et al., 1997, 1998; Olsgard and Somerfield, 2000). Comparisons of patterns at the species level and those from the higher taxonomic categories show that patterns are more closely related in polluted areas than in pristine ones (Olsgard et al., 1998; Olsgard and Somerfield, 2000). In anthropogenically impacted areas species abundances may be discriminated into a smaller number of lower taxonomic categories. If pollution, in the extreme example, reduces the number of species in each family to a single species, then patterns deriving from the genus and family level will be identical. Recently, the efficiency and robustness of taxonomic distinctness indices have been tested using data derived from Mediterranean lagoons at various geographic scales (Arvanitidis et al., 2005a,b). The main difficulty faced in the course of biodiversity/environmental health assessments in such lagoons is that natural disturbance, such as increased nutrient concentrations (fresh water resources) and high temperatures, as well as anthropogenic activities (e.g. intensive resource exploitation, sewage discharges) often result in dystrophic episodes (uncontrolled microbial activity and oxygen depletion), with identical community modifications: low species diversity with almost complete dominance of a few tolerant species; intercrossed ABC curves, associated with near zero W-statistic values (e.g. Cognetti, 1992; Reizopoulou et al., 1996). Analyses using taxonomic distinctness indices, incorporating phylogenetic/ taxonomic relationships, could differentiate severely impacted lagoons from naturally disturbed ones but standard multivariate analyses making use of information at the species level only, could not (Arvanitidis et al., 2005a,b). In this study we explore multivariate techniques, which incorporate higher taxonomic structure into the analysis, to see if such analyses match the capacity of taxonomic distinctness indices to distinguish between anthropogenic impact and natural disturbance in Mediterranean coastal lagoons. The performance of these techniques is tested by using both theoretical datasets as well as literature data from the Mediterranean. 2. Material and Methods 2.1. Data An artificial dataset was constructed (http://elnet-net.hcmr.gr/ data_sets/data_sets.zip), representing abundances of 48 species in samples from 4 coastal lagoons (A, B, C, D), together with taxonomic relationships between species (Table 1). Five replicate samples from each of 3 sites along an environmental gradient in each of 4 seasons for each lagoon are simulated, giving a total of 240 samples. Species presences, abundances and taxonomic relationships are used to represent different conditions. Seasonal shifting was taken into account based on the general trend that species abundances are known to produce a spatially homogenized multivariate pattern during summer and autumn and a well-defined gradient in winter and spring. This trend emerges from the Mediterranean lagoons literature and is well documented in the papers cited below. Accordingly, the two extremes were that the abundances values in the artificial dataset were within the same order of magnitude along the sea-land axe in summer sampling period and differed several orders of magnitude during spring period, with higher abundance values towards the land side. This theoretical dataset closely follows that proposed by Pearson and Rosenberg (1978) in that it provides

101

Table 1 Examples of the theoretical datasets. a: example of species presence/absence data; b: example of the aggregation matrix used for the classification of the species a Sp1 Sp2 Sp3 Sp4 Sp5 Sp6 Sp7 Sp8 Sp9 Sp10 Sp11 Sp12 Sp13 Sp14 Sp15 Sp16 Sp17 Sp18 Sp19 Sp20 Sp21 Sp22 Sp23 Sp24 Sp25

LagoonA

LagoonB

LagoonC

LagoonD

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0

1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1

1 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0

1 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 1

b SPECIES

GENUS

FAMILY

ORDER

CLASS

PHYLUM

Sp1 Sp2 Sp3 Sp4 Sp5 Sp6 Sp7 Sp8 Sp9 Sp10 Sp11 Sp12 Sp13 Sp14 Sp15 Sp16 Sp17 Sp18 Sp19 Sp20 Sp21 Sp22 Sp23 Sp24 Sp25

genus 1 genus 2 genus 51 genus 52 genus 10 genus 10 genus 21 genus 22 genus 23 genus 24 genus 25 genus 18 genus 18 genus 18 genus 26 genus 27 genus 28 genus 29 genus 12 genus 12 genus 12 genus 12 genus 30 genus 31 genus 16

family 1 family 1 family 99 family 100 family 5 family 5 family 7 family 7 family 8 family 9 family 10 family 6 family 6 family 6 family 11 family 12 family 13 family 14 family 2 family 2 family 2 family 2 family 15 family 16 family 2

order 1 order 1 order 3 order 3 order 4 order 4 order 6 order 6 order 7 order 7 order 7 order 5 order 5 order 5 order 7 order 7 order 8 order 9 order 2 order 2 order 2 order 2 order 10 order 11 order 2

class 1 class 1 class 3 class 3 class 4 class 4 class 6 class 6 class 7 class 7 class 7 class 5 class 5 class 5 class 7 class 7 class 8 class8 class 2 class 2 class 2 class 2 class 9 class 10 class 2

phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1 phylum 1

less species with higher abundances towards the land side and more species in lower abundance values towards the sea side of the lagoons included. Lagoons A and D represent one biogeographic sector, B and C another. Lagoons C and D represent severely impacted conditions, while A and B do not. Species abundances in the simulated dataset were aggregated to higher taxonomic groupings, namely genera, families, orders and classes, and averaged within seasons for each lagoon, standardised (converted to percentages) and fourthroot transformed. Eighteen Mediterranean lagoons, sampled at least on a seasonal basis, are considered (Fig. 1, Table 2). Macrobenthic datasets were collected from the following Hellenic lagoons: Tsopeli, Vivari (Reizopoulou et al., 1996), Gialova (Koutsoubas et al., 2000), Papas (Reizopoulou and Nicolaidou, 2004), Monolimni, Laki, and Drana

Author's personal copy 102

C. Arvanitidis et al. / Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

Fig. 1. Map indicating locations of the Mediterranean lagoonal systems taken into account in the present study. Numbers on the top of columns indicate number of species belonging to the most abundant macrobenthic groups (polychaetes, molluscs, and crustaceans. Abbreviations: Pr: Prevost; Fo: Fogliano; Mo: Monaci; Ca: Caprolace; Lu: Lungo; Fn: Fondi; Gh: Ghar el Melh; Gp: Sacca di Goro polluted; Gd: Sacca di Goro dredged; Ts: Tsopeli; Pa: Papas; Gl: Gialova; Vv: Vivari; La: Laki; Dr: Drana; Mn: Monolimni; Bu: Burollus; Ba: Bardawil.

(Kevrekidis et al., 2000; Kevrekidis, 2004a). Datasets on lagoonal macrobenthos elsewhere in the Mediterranean were collected from: Prevost (Guelorget and Michel, 1979a,b), Fogliano, Monaci, Caprolace, Fondi, Lungo (Gravina et al., 1989), Ghar el Melh (Romdhane and Chakroun, 1986), Sacca di Goro (both polluted and dredged areas; Reizopoulou et al., 1996), Burollus (Samaan et al., 1989) and Bardawil (Aboul-Ezz, 1988). The majority are euryhaline systems experiencing regular dystrophic crises during summer and early autumn associated with the rapid decomposition of organic material accumulated over the year. These crises frequently result in hypoxic or anoxic conditions in the water column and in sediments, with hydrogen sulphide releases into the water column. Although subjected to periodic aquaculture stress, most lagoons are naturally stressed ecosystems that do not experience severe anthropogenic impacts (e.g. Guelorget and Michel, 1979a,b; Gravina et al., 1989; Reizopoulou et al., 1996; Koutsoubas et al., 2000). Drana, Papas, Sacca di Goro and Burollus are considered to be anthropogenically stressed. Drana, Sacca di Goro, and Burollus are severely impacted. Drana lagoon is impacted by anthropogenic activities associated with the occlusion of its sea connection, which led to irregular reclamation of much of the lagoon. Papas lagoon, opening to the Patraikos Gulf and the Ionian Sea, is strongly affected by

eutrophic phenomena, characterized by prolonged seasonal hypoxia and frequent anoxic events (Reizopoulou and Nicolaidou, 2004). Sacca di Goro at the south of the Po River Delta is considered a severely impacted ecosystem (Reizopoulou et al., 1996; Viaroli et al., 2006; Marchini et al., 2004). Burollus is one of the most impacted lagoons in the Nile delta, subjected to inputs from agriculture, urbanization and industrial pollution. Dramatic alterations in fish and macrofaunal species diversity have been observed in recent years (Prof. M.R. Fishar, pers. comm.). On the other hand, for those lagoons reported as not receiving severe anthropogenic impact, sufficient information on runoffs from agriculture, industry, sewage or from any other kind of urban waste, is often not provided by the relevant publications. Average abundance values, over seasonal samplings, data for the most abundant macrofaunal species (polychaetes, molluscs and crustaceans combined) from the Hellenic lagoons were aggregated to genus, family, order, class and phylum. These taxa have been selected because the respective taxonomic knowledge and expertise is considered to be evenly distributed in the Mediterranean. Abundances of each macrofaunal group were also aggregated separately. Taxonomic classification followed the European Register of Marine Species (http:// www.marbef.org/data/erms.php) and ITIS (Integrated Taxonomic

Table 2 Diversity values, geomorphological and physico-chemical characteristics of the Mediterranean lagoons considered (modified from Dounas et al., 1998) Lagoons

Mean macrobenthic density1 H'2

J'2

Surface (Km2) Mean Depth (m) Salinity range Temperature range Reference

Bardawil Burollus Carpolace Drana Fogliano Fondi Ghar El Melh

3,850 440 1,239 9,051 2,489 10,482 —

— — 4.23 1.15 2.6 1.55 —

— — — 0.43 — — —

1,440 500 2.26 2.2 4.04 38 30

2.5 1 1.3 1.2 0.9 9.1 —

33.1-93.56 — — 0.9-8.7 18-45 13-33 —

15.6-32.7 17-29.3 8-30 2-34 7.5-28.2 — 9-29.5

Gialova Laki Lungo Monaci Monolimni Papas Prevost Sacca di Goro Sacca di Goro (polluted) Sacca di Goro (dredged) Tsopeli Vivari

5,792 9,063 4,516 1,687 27,538 8,633 5,485 — 5,866 13,921 1,648 3,716

2.5 1.16 1.36 2.99 1.5 2.33 0.51 — 1.7 1.7 2.35 1.42

0.57 0.43 — — 0.57 0.61 — — 0.53 0.52 0.59 0.41

2.4 1 47 0.95 1.12 — 38 25 — — 1 0.5

0.7 0.2 4.5 0.8 0.65 — 1 1.5 — — — —

13-60 0.1-35.1 13-33 — 0.3-5.7 20-42.5 17-40 18.4-35.6 18.4-34.7 19.9-35.6 21-35 31.5-40

14-24 6.7-25.9 — 8.9-29.2 1.8-28.5 10-32 24.9-12.2 4.9-26.9 — — 8-29 12-34

1: Yearly averaged values; 2: Index calculated from yearly averaged abundance values.

Aboul-Ezz (1988); Siliem (1989) Samaan et al. (1989) Gravina et al. (1989) Malea et al. (2004) Gravina et al. (1989) Gravina et al. (1989) Romdhane and Chakroun (1986); Ben Abdallah and Maamouri (2006) Dounas et al. (1998); present study Mogias and Kevrekidis (2005) Gravina et al., (1989) Gravina et al., (1989) Kevrekidis (2004a,b) Reizopoulou and Nicolaidou (2004) Guelorget and Michel (1979a,b) Reizopoulou et al. (1996) Reizopoulou et al. (1996) Reizopoulou et al. (1996) Reizopoulou et al. (1996) Reizopoulou et al. (1996)

Author's personal copy C. Arvanitidis et al. / Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

103

Fig. 2. Schematic representation of the non-metric multivariate scaling (nMDS) technique, applied on the theoretical dataset.

Information System; http://www.itis.usda.gov). Quantitative data were standardised (converted to percentages) and fourth-root transformed. Seasonal abundances were not available for many of the Mediterranean lagoons. For all lagoons data matrices were constructed with the numbers of taxa recorded in each taxonomic level from species to phylum, average (Δ+) and variation in taxonomic distinctness (Λ+) values as calculated for taxonomic levels from species to class, both for polychaetes, molluscs and crustaceans combined, and for each group separately. The original datasets are available from http://elnet-net. hcmr.gr/data_sets/data_sets.zip.

2.2. Analyses Non-metric multidimensional scaling (MDS) was performed on resemblance matrices calculated using Steinhaus' similarity coefficient (Legendre and Legendre, 1998), more commonly called the BrayCurtis coefficient (Clarke et al., 2006), on both the quantitative data (standardised and fourth-root transformed) and qualitative data (number of taxa, Δ+ and Λ+ values, per taxonomic category and lagoon). These analyses were applied to the artificial dataset as well as to those from the Hellenic and the Mediterranean lagoons. Stress of

Fig. 3. Schematic representation of the third-stage MDS, biodiversity MDS (bdMDS), applied on the theoretical dataset.

Author's personal copy 104

C. Arvanitidis et al. / Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

Fig. 5. MDS plot resulting from the average macrofaunal density values at the sectoral scale (Hellenic lagoons).

Fig. 4. Ordination plots resulting from the application of: (a) the MDS technique on the species density values from the seasonal sampling data, in included in the model file; (b) the bdMDS technique on the same dataset.

2-dimensional plots was measured by Kruskal's stress formula I (Clarke and Green, 1988). Changes in pattern deriving from different taxonomic levels in different lagoons were assessed by means of second-stage MDS (Somerfield and Clarke, 1995). This kind of analysis requires, however, quantitative data and was, therefore, applied only to the artificial dataset and to those from the Hellenic lagoons considered. Lagoons in the same biogeographical sector will have more species in common than lagoons in different sectors. In a standard multivariate analysis of species abundance data from the 4 lagoons in the simulated dataset lagoons in the same sector should cluster together. Lagoons A and D should be well-distinguished from lagoons B and C (Fig. 2). To strip out the effect of shared species, and to focus on changes in pattern across taxonomic levels, the key is to construct second-stage resemblance matrices (Fig. 3). First, matrices of resemblances between seasonal abundances of taxa at each taxonomic level within each lagoon are constructed. These resemblance matrices are produced by taxa-by-season (averaged abundance values) matrices, at each of the taxonomic levels. These matrices are then correlated using a rank correlation between corresponding elements. Thus a second-stage resemblance matrix is constructed for each lagoon (Fig. 3), which could be used to display interrelationships between patterns derived from different taxonomic levels, as in Somerfield and Clarke (1995). To display interrelationships between lagoons, in terms of how similar they are with respect how patterns

change across taxonomic levels, a second second-stage resemblance matrix (here termed a third-stage resemblance matrix) is constructed using rank correlations between corresponding elements in the set of second-stage matrices (Fig. 3). This third-stage matrix is further processed by MDS (a third-stage MDS) in which lagoons showing similar changes in pattern as species abundances are aggregated to higher taxonomic levels, regardless of whether they have any species in common, will group together. The anticipated pattern in the third-stage MDS plot, henceforward termed biodiversity MDS (bdMDS), is that impacted lagoons (in which interrelations of the patterns deriving from the different taxonomic levels are expected to be similar, according to the hierarchic-response-tostress hypothesis) should cluster separately from non-impacted ones, irrespective of the species present. In such an analysis of the simulated dataset, therefore, lagoons C and D should cluster separately from A and B. The bdMDS has been explored using only the simulated dataset and those from the Hellenic sector which have been collected by using hand-operating grabs (0.025-0.05 m2 sampling surface). At the panMediterranean scale, however, there is insufficient data to assess changes in the full multivariate pattern across taxonomic levels. Heterogeneity in sampling gears used was another reason for not taking into account the quantitative information but only information on the presence of species. Non-metric MDS was performed on resemblance matrices summarising similarity between lagoons in the way that the number of taxa, Δ+, and Λ+ values change across taxonomic levels, which from now on are referred to as ntMDS, δMDS and λMDS, respectively. 3. Results In the standard MDS ordination plot derived from Steinhaus similarities calculated from yearly averaged abundance values in the simulated dataset (Fig. 4a), as expected, lagoons A and D representing impacted and unimpacted conditions in the same biogeographic sector, are clearly separated from lagoons B and C representing impacted and unimpacted conditions in another sector. Using presence-absence data, or seasonally averaged abundance values,

Fig. 6. Left column: ordination bdMDS plots of ranked intermatrix correlations resulted from the third-stage MDS; matrices involved, include information on changes of patterns from the species and higher taxonomic levels. Data include both the total of the most abundant macrofaunal groups (a), as well as each of these groups: polychaetes (b), molluscs (c) and crustaceans (d). Right column: ordination ntMDS plots; matrices involved, include numbers of taxa identified for each lagoon. Data include both the total of the most abundant macrofaunal groups (e), as well as each of these groups: polychaetes (f), molluscs (g) and crustaceans (h).

Author's personal copy C. Arvanitidis et al. / Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

105

Author's personal copy 106

C. Arvanitidis et al. / Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

resulted in the same lagoonal differentiation (not shown). The bdMDS (Fig. 4b) shows, as intended, a completely different grouping of the lagoons: the lagoons designated as impacted (C, D) cluster separately from those designated as non-impacted (A, B). Application of the standard multivariate analysis to the 7 lagoons in the Hellenic sector (Fig. 5) clearly separates lagoons biogeographically, with lagoons in the vicinity of the Evros delta (Monolimni, Laki, Drana) clustering separately from lagoons on the coasts of the Peloponnese and the Ionian Sea. Again, and as expected, analyses based on presence-absence data and averaged abundance values of the seasonal sampling data resulted in similar MDS ordination plots (not shown). Applying bdMDS to the data from these lagoons, however, generally shows different patterns (Fig. 6a-d). This is expected, as lagoons are now ordinated on the basis of their taxonomic structure, not on the basis of the species composition and abundance. The exception, however, is the plot derived from polychaetes (Fig. 6b) where the biogeographic pattern is evident. In none of the plots (Fig. 6a-d) is there a clear separation of impacted lagoons (Drana, and to a lesser extent Papas) from non-impacted lagoons, although there is a suggestion in Fig. 6b that Drana is separated from other lagoons in the Evros delta. Similar results, for lagoons in the Hellenic sector, were also derived from macrobenthos and polychaetes with ntMDS, crustaceans with δMDS and polychaetes, molluscs and crustaceans with λMDS (not shown). The advantage of these techniques, however, is that they can be applied to all the Mediterranean lagoons (Figs. 6e-h and 7) and are very much easier to calculate. For macrobenthos in all lagoons it is immediately apparent that ntMDS clearly discriminates the severely impacted lagoons (Drana, Borullus and Sacca di Goro) from the rest (Fig. 6e). The division of the severely impacted lagoons from the remainder is also clear for the plot derived from molluscs (Fig. 6g), but less so for plots derived from polychaetes (Fig. 6f) or crustaceans (Fig. 6h). Additionally, there is no evidence, from these plots, that Papas is impacted in comparison to other lagoons in the Mediterranean. Plots from δMDS and λMDS for the total macrobenthic fauna in all Mediterranean lagoons and for the most abundant groups (Fig. 7) indicate that δMDS (Fig. 7a-d) does not tend to discriminate severely impacted lagoons as a group consistently, although it does separate out some impacted lagoons, such as Sacca di Goro, Borullus or Drana, some of the time. However, the λMDS plot derived from polychaetes (Fig. 7f) clearly distinguishes severely impacted lagoons from the remainder. The plot derived from mollusc inventories (Fig. 7g) discriminates impacted lagoons less convincingly, and the remaining λMDS plots, like the δMDS plots, tend to be inconsistent in the way they discriminate severely impacted and not severely impacted lagoons. 4. Discussion Several techniques are described and tested, which aspire to focus on taxonomic structure to highlight changes associated with impacts of anthropogenic stress in Mediterranean lagoons. The aim is to determine multivariate methods, which are not confounded by biogeographic differences, to be used to discriminate the effects of anthropogenic impacts in an environment which experiences high levels of natural stress. The results of bdMDS analysis of the simulated dataset demonstrate that the technique has, indeed, the potential to detect changes

in assemblages associated with differences in higher taxonomic structure, independent of changes in species composition or abundance. Application of the technique to real data from a number of lagoons indicated that, in general, changes in taxonomic structure did not correlate with changes in species composition, as expected, but failed to discriminate clearly between lagoons considered to be anthropogenically impacted and those considered to be naturally disturbed. The possibility of measuring the degree of agreement between second-stage plots using the Spearman correlation of their underlying second-stage resemblance matrices (of Spearman correlations), leading to a third-stage MDS and further third-stage tests, was discussed by Clarke et al. (2006), albeit in a different context. They concluded that third-stage MDS could answer questions about whether some repeats of an experimental protocol at certain times or places tend to give different conclusions, but cautioned that a practically useful application of third-stage MDS is unlikely, since each successive stage involves a further step away from the data of species counts or area cover. While there may well be contexts in which third-stage MDS proves to be a useful technique, the discrimination of severely impacted lagoons in the Mediterranean does not appear to be one of them, at least using the datasets explored here. The simplest way of examining taxonomic structure at increasing levels in a taxonomic hierarchy is to look at the number of taxa at each successive level. While this information may be displayed as a table, a simple graphical alternative is to plot the number of taxa against taxonomic level to produce a profile for each sample. Such approaches are useful where only a limited number of samples are to be compared, but tend to become unwieldy if more than a few samples are involved. As described by Clarke (1990) for dominance curves, a logical alternative is to use a measure of distance between profiles in a subsequent multivariate approach. In essence this is what is achieved by ntMDS. This technique does discriminate the lagoons in the Mediterranean considered to be under severe anthropogenic (in addition to natural) stress, namely Sacca di Goro, Drana and Borullus, from all of the others. The method is simple, relies only on species lists with their associated taxonomic hierarchy, is not influenced by differences in the identities of species (biogeographic differences), and is not difficult to apply. As such it seems a sensible approach to employ to detect anthropogenic impact in Mediterranean lagoons. The discrimination of impacted lagoons using ntMDS is clear using data on all macrofauna, but also using only data on the number of mollusc taxa present. Analyses of polychaete and crustacean data do discriminate some of the severely impacted lagoons, but not all of them consistently. This is at odds with the findings of Olsgard and Somerfield (2000) who demonstrated that, among the major groups of macrofauna present, polychaetes tended to match patterns in the total macrofauna in impacted areas of the North Sea most closely. They attributed this to the diversity of life-history strategies, ecological roles and sensitivity to impacts within the group. The taxonomic and ecological diversity of molluscs in lagoons in the Mediterranean, and their relative sensitivity, merits further research. Do they, for example, tend to fill roles normally associated with polychaetes in more diverse sublittoral assemblages? While there are hints in the resulting plots that δMDS is detecting some difference between lagoons that are severely impacted and those that are not, the technique does not provide clear or consistent results. The related method, λMDS, performs better. For polychaetes

Fig. 7. Left column: ordination plots resulted from the δMDS; matrices involved, include information on the Δ+ values, calculated at each taxonomic level. Data include Δ+ values calculated both for the total of the most abundant macrofaunal groups (a), as well as for each of these groups: polychaetes (b), molluscs (c) and crustaceans (d). Right column: ordination plots resulted from the λMDS; matrices involved, include information on the Λ+ values, calculated at each taxonomic level. Data include Λ + values calculated both for both the total of the most abundant macrofaunal groups (e), as well as for each of these groups: polychaetes (f), molluscs (g) and crustaceans (h).

Author's personal copy C. Arvanitidis et al. / Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

107

Author's personal copy 108

C. Arvanitidis et al. / Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

λMDS clearly discriminates all of the anthropogenically impacted lagoons from the naturally stressed ones. It is perhaps surprising that λMDS and ntMDS discriminate impacted lagoons most effectively using data from different taxa, but it should be remembered that each technique relies on very different source of information. Variation in taxonomic distinctness (Λ+) is not mechanistically correlated with the number of taxa involved, instead reflecting how evenly taxa are distributed across the underlying taxonomic hierarchy. The λMDS using mollusc data does discriminate Sacca di Goro and Drana from the others, but the relationships between Borullus and the impacted lagoons, and Vivari and the moderately disturbed lagoons, are not as expected. A potential problem with all of these techniques is that the phylogenetic/taxonomic resolution is more stable in some groups than in others: a family within the polychaetes does not necessarily correspond to a molluscan or a crustacean family. Such difficulties are unlikely ever to be resolved, at least until phylogenetic analysis reveals the entire tree-of-life (e.g. Pleijel and Rouse, 2003). However, the original observations that prompted Warwick and Clarke (1995) to develop indices based on the relatedness of organisms, such as that in severely impacted situations large numbers of individuals from closely related species, such as sibling species of Capitella and Tisbe, are often found, and also that along pollution gradients different phyla tend to show differing sensitivities to pollution, hold true. This can be reflected in the polychaetes, in the case of λMDS for instance, because it is either the dominant group in terms of abundance or biomass values, or an important contributor to the entire macrobenthic communities in the coastal lagoons (e.g. Arvanitidis et al., 2005a,b). Consequently, severely impacted lagoons imposing a high degree of niche filtering through less and more homogenised microhabitats available, host a limited number of closely related taxa. Although this may be overlooked at the level of the species it can, nevertheless, be apparent in their higher phylogenetic structure. The latter is not far from an interpretation which could derive from other approaches such as niche filtering and functional traits analysis: almost all of the feeding methods occur in this taxon (e.g. Fauchald and Jumars, 1979; Rouse and Pleijel, 2001). Feeding methods is one of the main functional traits, closely linked to ecosystem processes and which make the group more reliable than other macrobenthic components in the natural representation of the communities. Therefore, species sharing feeding methods are able to survive under severe anthropogenic impact because they most probably have similar feeding structures and feeding mechanisms. If this is true, then these species are more likely to have many characters shared in common and thus common higher classification (e.g. congeners, species classified under the same family). The latter may explain, up to a certain degree, why polychaetes did not perform with ntMDS but they did perform with λMDS: the former method is an approximation through raw numbers of taxa at each taxonomic level while the latter is a measure of their actual taxonomic relatedness. Many studies have looked at taxonomic structure in relation to pollution and other stressors, using a range of multivariate, graphical and univariate techniques. As discussed by Ellingsen et al. (2005), there tends to be a trade-off between the confidence one has in a taxonomic hierarchy and the number of species one is considering, but as long as one taxonomic hierarchy is consistently applied within a study there is no reason to discard the Linnean classification in its entireity because it is not perfect. There is valid, albeit imperfect, information to be utilized in the taxonomic structure of assemblages, and we suggest some approaches in this paper which use it and may prove useful. Methods based on the taxonomic relatedness are not the only ones proposed for the detection of anthropogenic impacts to macro- and meiobenthic communities (Warwick and Clarke, 2001). Moulliot et al. (2006) have summarized an array of alternative methods, such as

those based on: (i) body size or size spectra (e.g. Holling, 1992; Robson et al., 2005), (ii) biomass distribution among functional groups (e.g. Pearson, 2001; Gerino et al., 2003), (iii) functional diversity (e.g. Diaz and Cabido, 2001; Petchey and Gaston, 2002; Bremner et al., 2003), and (iv) biomass or productivity measures (e.g. Pearson and Rosenberg, 1978). Some of the methods have also been applied to coastal Mediterranean lagoons with a varying degree of success (e.g. Basset et al., 2006; Reizopoulou and Nicolaidou, 2007). Recently, Moulliot et al. (2007) have developed methodology for the testing of the niche filtering hypothesis, with promising results in two Mediterranean lagoons. Bremner (2008) has reviewed the biological traits analysis in benthic communities and their potential use in conservation and management. The fundamental principle behind the methodologies hitherto developed and tested is a presumed mechanistic link between the shape of the assemblages and how their members (species, taxa) are related to each other as competitors or members of a web of interactions, facing (dis)similar environmental constrains (Moulliot et al., 2007). All of the above methods, however, require additional information (e.g. biomass measures, functional traits), the collection and systematization of which may be highly demanding in energy and time. Certainly, these data are not readily available from the literature for the total of the Mediterranean lagoons considered here. The pros and cons of each of the previouslymentioned methodologies, including those based on taxonomic relatedness can only be tested through inter-calibration exercises on datasets produced under similar conditions and for the same purposes. Finally, the incorporation of the special attributes of each of the methods into a broader theory, such as the metabolic theory of ecology (Brown et al., 2004) should also be attempted in order to answer questions relevant with space, time and biological organization scales. The novel feature in the multivariate techniques presented above is one that allows comparisons of any habitat throughout the Mediterranean in terms of changes in multivariate patterns derived from data at each taxonomic level (from species to phylum). Consequently, these methods make use of information that has never been utilized by the multivariate techniques: they do not simply use the “frontpage” information (species occurrence and abundance) but also their taxonomic/phylogenetic relationships. The efficiency of these methods to distinguish severely impacted lagoons from the remainder in the Mediterranean, when applied to specific macrobenthic components, indicates that their capability to compare changes in pattern as the information is aggregated into higher than species taxonomic categories, as an intrinsic character, has a strong potential in their wider application as rapid assessment techniques. This potential is further enhanced by the general acceptance of the multivariate techniques as the more sensitive ones among those used for environmental assessment, and also as those techniques which give the most reliable results compared to other standard methods such as calculation of “classical” diversity indices and the utilization of graphical methods (e.g. Clarke and Warwick, 1994). Coastal lagoons, especially those in the Mediterranean, are severely fluctuating environments. The populations within them show dramatic seasonal, annual and inter-annual variations, ranging from disappearance to complete dominance during the periods of dystrophic crises. Although the techniques described here seem to have a degree of success in detecting anthropogenic impact in Mediterranean coastal lagoons, their application in studies of lagoonal macrofaunal biodiversity on smaller scales, or in other Mediterranean habitats, should be tested before they are proposed for wider use in the region. All of these analyses should be seen as merely providing evidence for the existence of particular forms of, or changes in, assemblage structuring. The contributions from individual species in establishing observed patterns are fundamental, and understanding the ecological mechanisms underpinning those contributions should continue to be a primary goal of any scientific investigation.

Author's personal copy C. Arvanitidis et al. / Journal of Experimental Marine Biology and Ecology 369 (2009) 100–109

Acknowledgements The authors are much indebted to Mrs. M. Eleftheriou for the critical reading of the manuscript. The authors acknowledge support by the MARBEF EU funded Network of Excellence (6th EU FP) and TWReferenceNET (INTERREG IIIB). Support was also received from the UK NERC through PML's CSRP, and Defra, Project ME3109. The three anonymous reviewers are acknowledged for their fruitful comments and suggestions. [RH] References Aboul-Ezz, S.M., 1988. Periodicity and distribution of bottom fauna of the hyper-saline Bardawil lagoon (Egypt). Bull. Natl. Inst. Oceanogr. Fish ARE 14, 159–174. Arvanitidis, C., Chatzigeorgiou, G., Koutsoubas, D., Dounas, C., Eleftheriou, A., Koulouri, P., 2005a. Mediterranean lagoons revisited: weakness and efficiency of the rapid biodiversity assessment techniques in a severely fluctuating environment. Biodivers. Conserv. 14, 2347–2359. Arvanitidis, C., Chatzigeorgiou, G., Koutsoubas, D., Kevrekidis, T., Dounas, C., Eleftheriou, A., Koulouri, P., Mogias, A., 2005b. Estimating lagoonal biodiversity in Greece: comparison of rapid assessment techniques. Helgol. Mar. Res. 59, 177–186. Basset, A., Galuppo, N., Sabetta, L., 2006. Environmental heterogeneity and benthic macroinvertebrate guilds in italian lagoons. Transit. Waters Bull. 1, 48–63. Ben Abdallah, L.G., Maamouri, F., 2006. Spatial-temporal dynamics of the nematode Anguillicolla crassus in Northest Tunisian lagoons. C.R. Biologies 329, 785–789. Boesch, D.F., Rosenberg, R., 1981. Response to stress in marine benthic communities. In: Barrett, G.W., Rosenberg, R. (Eds.), Stress effects on natural ecosystems. Wiley, New York, pp. 179–200. Bremner, J., 2008. Species' traits and ecological functioning in marine conservation and management. J. Exp. Mar. Biol. Ecol. 366, 37–47. Bremner, J., Rogers, S.I., Frid, C.L.J., 2003. Assessing functional diversity in marine ecosystems: a comparison of approaches. Mar. Ecol. Prog. Ser. 254, 11–25. Brown, J.H., Gilloogy, J.F., Allen, A.P., Van Savage, M., West, G.B., 2004. Toward a metabolic theory of ecology. Ecology 85, 1771–1789. Clark, R.B., 1997. Marine Pollution. Clarendon Press, Oxford. Clarke, K.R., 1990. Comparisons of dominance curves. J. Exp. Mar. Biol. Ecol. 138, 143–157. Clarke, K.R., Green, R.H., 1988. Statistical design and analysis for a “biological effects” study. Mar. Ecol. Prog. Ser. 46, 213–226. Clarke, K.R., Warwick, R.M., 1994. Change in marine communities: an approach to statistical analysis and interpretation. Natural Environmental Research Council, Plymouth Marine Laboratory, Plymouth. Clarke, K.R., Warwick, R.M., 1998a. Quantifying structural redundancy in ecological communities. Oecologia 113, 278–289. Clarke, K.R., Warwick, R.M., 1998b. A taxonomic distinctness index and its statistical properties. J. Appl. Ecol. 35, 523–531. Clarke, K.R., Warwick, R.M., 2001. A further biodiversity index applicable to species lists: variation in taxonomic distinctness. Mar. Ecol. Prog. Ser. 216, 265–278. Clarke, K.R., Somerfield, P.J., Chapmen, M.G., 2006. On resemblance measures for ecological studies, including taxonomic dissimilarities and a zero-adjusted BrayCurtis coefficient for denuded assemblages. J. Exp. Mar. Biol. Ecol. 330, 55–80. Cognetti, G.,1992. Colonization of stressed coastal environments. Mar. Pollut. Bull. 24,12–14. Diaz, S., Cabido, M., 2001. Vive la difference: plant functional diversity matters to ecosystem processes. Trends Ecol. Evol. 16, 646–655. Dounas, C., Koutsoubas, D., Arvanitidis, C., Petihakis, G., Drummond, L., Eleftheriou, A., 1998. Biodiversity and the impact of anthropogenic activities in Mediterranean lagoons: the case of Gialova lagoon, SW Greece. Oebalia 24, 77–91. Ellingsen, K.E., Clarke, K.R., Somerfield, P.J., Warwick, R.M., 2005. Taxonomic distinctness as a measure of diversity applied over a large scale: the benthos of the Norwegian continental shelf. J. Anim. Ecol. 74, 1069–1079. Fauchald, K., Jumars, P., 1979. The diet of worms: A study of polychaete feeding guilds. Oceanogr. Mar. Biol. Annu. Rev. 17, 193–284. Féral, J.-P., Fourt, M., Perez, T., Warwick, R.M., Emblow, C., Heip, C., van Avesaath, P., Hummel, H., 2003. European Marine Biodiversity Indicators. NIOO-CEME, Yerseke. Ferraro, S.P., Cole, F.A., 1990. Taxonomic level and sample size sufficient for assessing pollution impacts on the Southern California Bight macrobenthos. Mar. Ecol. Prog. Ser. 67, 251–262. Ferraro, S.P., Cole, F.A., 1995. Taxonomic level sufficient for assessing pollution impacts on the southern California bight macrobenthos – revisited. Environ. Toxicol. Chem.1,1031–1040. Gerino, M., Stora, G., Francois-Carcaillet, F., Gilbert, F., Poggiale, J.C., Mermillod-Blondin, F., Desrosiers, G., Vervier, P., 2003. Macro-invertebrate functional groups in freshwater and marine sediments: a common mechanistic classification. Vie Milieu 53, 221. Gravina, M.F., Ardizzone, G.D., Scaletta, F., Chimenz, C., 1989. Descriptive analysis and classification of benthic communities in some Mediterranean coastal lagoons (central Italy). PSZN Mar. Ecol. 10, 141–166. Gray, J.S., 1997. Marine biodiversity: patterns, threats and conservation needs. Biodivers. Conserv. 6, 153–175. Gray, J.S., 2000. The measurement of marine species diversity with an application to the benthic fauna of the Norwegian continental shelf. J. Exp. Mar. Biol. Ecol. 250, 23–49. Guelorget, O., Michel, P., 1979a. Les peuplements benthiques d'un étang littoral Languedocien, l'étang du Prevost (Herault). Téthys 9, 49–64. Guelorget, O., Michel, P., 1979b. Les peuplements benthiques d'un étang littoral Languedocien, l'étang du Prevost (Herault). 2. Ètude quantitative de la macrofaune des sables. Téthys 9, 65–77.

109

Holling, C.S., 1992. Cross-scale morphology, geometry, and dynamics of ecosystsems. Ecol. Monogr. 62, 447–502. Izsak, C., Price, A.R.G., 2001. Measuring β-diversity using a taxonomic similarity index, and its relation to spatial scale. Mar. Ecol. Prog. Ser. 215, 69–77. Kevrekidis, T., 2004a. Seasonal variation of the macrozoobenthic community structure at low salinities in a Mediterranean lagoon (Monolimni Lagoon, Northern Aegean) Int. Rev. Hydrobiol. 89, 407–425. Kevrekidis, T., 2004b. Population dynamics, growth and reproduction of Corophium insidiosum (Crustacea: Amphipoda) at low salinities in Monolimni Lagoon (Evros delta, North Aegean Sea). Hydrobiologia 522, 117–132. Kevrekidis, T., Mogias, A., Gouvis, N., 2000. Interannual changes in the composition of the macrobenthic fauna of Drana lagoon (Evros Delta, N. Aegean Sea): preliminary note. Belg. J. Zool. 130, 105–111. Kingston, P.F., Riddle, M.J., 1989. Cost effectiveness of benthic faunal monitoring. Mar. Pollut. Bull. 20, 490–496. Koutsoubas, D., Dounas, C., Arvanitidis, C., Kornilios, S., Petihakis, G., Triantafyllou, G., Eleftheriou, A., 2000. Macrobenthic community structure and disturbance assessment in Gialova lagoon, Ionian Sea. ICES J. Mar. Sci. 57, 1472–1480. Legendre, P., Legendre, L., 1998. Numerical ecology. Developments in environmental modelling, vol. 20. Elsevier Scientific Publishers, Amsterdam. Magurran, A.E., 2004. Measuring Biological Diversity. Blackwell Publishing, Madlen, Oxford, Carleton. Malea, P., Kevrekidis, T., Mogias, A., 2004. Annual versus perennial growth cycle in Ruppia maritime L.: temporal variation in population characteristics in Mediterranean lagoons (Monolimni and Drana lagoons, Northern Aegean Sea). Bot. Mar. 47, 357–366. Marchini, A., Gauzer, K., Occhipinti-Ambrogi, A., 2004. Spatial and temporal variability of hard-bottom macrofauna in a disturbed coastal lagoon (Sacca di Goro, Po River Delta northwestern Adriatic Sea). Mar. Pollut. Bull. 48, 1084–1095. Mogias, A., Kevrekidis, T., 2005. Macrobenthic community structure in a poikilohaline Mediterranean lagoon (Laki lagoon, northern Aegean). Helgol. Mar. Res. 59, 167–176. Moulliot, D., Spatharis, S., Reizopoulou, S., Laugier, T., Sabetta, L., Basset, A., Do Chi, T., 2006. Alternatives to taxonomic-based approaches to assess changes in transitional water communities. Aquat. Conserv.: Mar. Freshw. Ecosyst. 16, 469–482. Moulliot, D., Dumay, O., Tomasini, J.A., 2007. Limiting similarity, niche filtering and functional diversity in coastal lagoon fish communities. Estuar. Coast. Shelf Sci. 71, 443–456. Olsgard, F., Somerfield, P.J., 2000. Surrogates in marine benthic investigations – which taxonomic unit to target? J. Aquat. Ecosyst. Stress Recovery 7, 25–42. Olsgard, F., Somerfield, P.J., Carr, M.R., 1997. Relationships between taxonomic resolution and data transformations in analyses of a macrobenthic community along an established pollution gradient. Mar. Ecol. Prog. Ser. 149, 173–181. Olsgard, F., Somerfield, P.J., Carr, M.R., 1998. Relationships between taxonomic resolution, macrobenthic community patterns and disturbance. Mar. Ecol. Prog. Ser. 172, 25–36. Pearson, T.H., 2001. Functional group ecology in soft-sediment marine benthos: the role of bioturbation. Oceanogr. Mar. Biol. 39, 233. Pearson, T.H., Rosenberg, R., 1978. Macrobenthic succession in relation to organic enrichment and pollution of the marine environment. Oceanogr. Mar. Biol. Annu. Rev. 16, 229–311. Petchey, O.L., Gaston, K.J., 2002. Functional Diversity (FD), species richness and community composition. Ecol. Lett. 5, 402–411. Pleijel, F., Rouse, G.W., 2003. Ceci n'est pas une pipe: names, clades and phylogenetic nomenclature. J. Zool. Syst. Evol. Res. 41, 162–174. Reizopoulou, S., Nicolaidou, A., 2004. Benthic diversity of coastal brackish-water lagoons in Western Greece. Aquat. Conserv. 14, 93–102. Reizopoulou, S., Nicolaidou, A., 2007. Index of Size Distribution (ISD): a method of quality assessment for coastal lagoons. Hydrobiologia 577, 141–149. Reizopoulou, S., Thessalou-Legaki, M., Nicolaidou, A., 1996. Assessment of disturbance in Mediterranean lagoons: an evaluation of methods. Mar. Biol. 12, 189–197. Robson, B.J., Barmuta, L.A., Fairweather, P.G., 2005. Methodological and conceptual issues in the search for a relationship between animal body-size distributions and benthic habitat architecture. Mar. Freshw. Res. 56, 1–11. Romdhane, M.S., Chakroun, F.L., 1986. Les peuplements benthiques de la lagune de Ghar el Mehl. Bull. Inst. Natl. Sci. Tech. Océanogr. Pêche Salammbô 13, 95–108. Rouse, G.W., Pleijel, F., 2001. Polychaetes. Oxford University Press, Oxford. Samaan, A.A., Ghobashy, A.F.A., Aboul-Ezz, S.M., 1989. The benthic fauna of lake Burollus. 1. Community composition and distribution of the total fauna. Bull. Natl. Inst. Oceanogr. Fish ARE 15, 217–224. Siliem, T.A.E., 1989. Chemical conditions in Bardawil lagoon. III. – Some limnological studies. Bull. Natl. Inst. Oceanogr. Fish ARE 15, 21–33. Somerfield, P.J., Clarke, K.R., 1995. Taxonomic levels, in marine community studies, revisited. Mar. Ecol. Prog. Ser. 127, 113–119. Viaroli, P., Giordani, G., Bartoli, M., Naldi, M., Azzoni, R., Nizzoli, D., Ferrari, I., Zaldívar, J.M., Bencivelli, S., Castaldelli, G., Fano, E.A., 2006. The Sacca di Goro and an arm of the Po river. In: Wangersky, P.J. (Ed.), The Handbook of Environmental Chemistry, pp. 197–232. Estuaries 5/H. Warwick, R.M., 1988. The level of taxonomic discrimination required to detect pollution effects on marine benthic communities. Mar. Pollut. Bull. 19, 259–268. Warwick, R.M., Clarke, K.R., 1995. New “biodiversity” measures reveal a decrease in taxonomic distinctness with increasing stress. Mar. Ecol. Prog. Ser. 129, 301–305. Warwick, R.M., Clarke, K.R., 2001. Practical measures of marine biodiversity based on relatedness of species. Oceanogr. Mar. Biol. Annu. Rev. 39, 207–231. Warwick, R.M., Light, J., 2001. Death assemblages of molluscs on St. Martin's flats, Isles of Scilly: a surrogate for Mediterranean biodiversity? Biodivers. Conserv. 11, 99–112. Warwick, R.M., Platt, H.M., Clarke, K.R., Agard, J., Gobin, J., 1990. Analysis of macrobenthic and meiobenthic community structure in relation to pollution and disturbance in Hamilton harbour, Bermuda. J. Exp. Mar. Biol. Ecol. 138, 119–142.