1 23

HydrobiologiaThe International Journal of AquaticSciences ISSN 0018-8158 HydrobiologiaDOI 10.1007/s10750-012-1273-z

Natural variability and referenceconditions: setting type-specificclassification boundaries for lagoonmacroinvertebrates in the Mediterraneanand Black SeasAlberto Basset, Enrico Barbone, AngelBorja, Michael Elliott, Giovanna Jona-Lasinio, João Carlos Marques, KrysiaMazik, et al.

1 23

Your article is protected by copyright and

all rights are held exclusively by Springer

Science+Business Media B.V.. This e-offprint

is for personal use only and shall not be self-

archived in electronic repositories. If you

wish to self-archive your work, please use the

accepted author’s version for posting to your

own website or your institution’s repository.

You may further deposit the accepted author’s

version on a funder’s repository at a funder’s

request, provided it is not made publicly

available until 12 months after publication.

WATER BODIES IN EUROPE

Natural variability and reference conditions: settingtype-specific classification boundaries for lagoonmacroinvertebrates in the Mediterranean and Black Seas

Alberto Basset • Enrico Barbone • Angel Borja • Michael Elliott • Giovanna Jona-Lasinio •

Joao Carlos Marques • Krysia Mazik • Inigo Muxika • Joao Magalhaes Neto •

Sofia Reizopoulou • Ilaria Rosati • Heliana Teixeira

Received: 23 February 2012 / Accepted: 4 August 2012

� Springer Science+Business Media B.V. 2012

Abstract The ecological status classification of

aquatic ecosystems using biological indices requires a

number of steps, including the description and stan-

dardisation of the indices’ natural variability. Here, we

address this point with reference to selected Mediterra-

nean and Black Sea lagoons, using benthic macroinver-

tebrates in order to: (i) explore the drivers and extent of

the indices’ natural variability; (ii) evaluate lagoon

type-specific reference conditions and related classifi-

cation boundaries; (iii) test the classification strength of

the derived boundaries; and, (iv) propose recommen-

dations for optimising ecological status classification.

The considered indices showed large variation between

and within the reference lagoons on both spatial and

temporal scales. Among the tested descriptors of the

proposed lagoon typologies, surface area, confinement

and water salinity were found to be significant sources of

index variability. Type-specific reference conditions

and classification boundaries were then defined,

improving the accuracy of ecological status assessment.

At the lagoon level, classification strength increased up

to 100 % in reference (least disturbed) lagoons and up to

83 % in an independent validation set of highly

disturbed sites. Nevertheless, a certain degree of

uncertainty was still found to affect classification at

the study site level. Recommendations concerning the

application of the various approaches to type-specific

reference conditions and classification boundaries are

given.

Guest editors: C. K. Feld, A. Borja, L. Carvalho &

D. Hering / Water bodies in Europe: integrative systems to

assess ecological status and recovery

A. Basset (&) � I. Rosati

Department of Biological and Environmental Sciences

and Technologies, University of Salento,

73100 Lecce, Italy

e-mail: alberto.basset@unisalento.it

E. Barbone

ARPA Puglia – Puglia Regional Environmental Protection

Agency, Corso Trieste 27, 70100 Bari, Italy

A. Borja � I. Muxika

Marine Research Division, AZTI-Tecnalia,

Herrera Kaia, Portualdea s/n, 20110 Pasaia, Spain

M. Elliott � K. Mazik

Institute of Estuarine & Coastal Studies, University of

Hull, Cottingham Road, Hull HU6 7RX, UK

G. Jona-Lasinio

DSS University of Rome La Sapienza, P.le Aldo Moro 5,

00185 Rome, Italy

J. C. Marques � J. M. Neto � H. Teixeira

IMAR—Institute of Marine Research, Marine and

Environmental Research Centre, Department of Life

Sciences, University of Coimbra, Largo Marques de

Pombal, 3004-517 Coimbra, Portugal

S. Reizopoulou

Institute of Oceanography, HCMR,

47km Athinon-Souniou, 1903 Anavyssos, Greece

123

Hydrobiologia

DOI 10.1007/s10750-012-1273-z

Author's personal copy

Keywords Reference conditions � Typology �Natural variability � Mediterranean and Black Sea

lagoons � Multimetric indices � Benthic

macroinvertebrates

Introduction

Benthic macroinvertebrates represent one of the most

extensively used biological quality elements in the

assessment of ecological status in European water

bodies under the Water Framework Directive (WFD,

2000/60/EC) (Borja et al., 2009a, b). Using benthic

macroinvertebrates to assess ecological status requires

the development and validation of: (i) tools (i.e.

indices) to assess the relative quality of the considered

habitats based on human pressures, (ii) thresholds of

natural variability along the driver gradients of each

assessment tool, and (iii) benchmark values or ‘refer-

ence conditions’ as the basis for developing a classi-

fication system of ecosystem ecological status. A large

number of benthic indices have been developed and

successfully validated over the last decade (see Pinto

et al., 2009 for a review; and Borja et al., 2009a, b;

Birk et al., 2012) but relatively little attention has been

paid to the methods used for setting adequate reference

conditions, although this step is clearly crucial for the

assessment of ecological status (Borja et al., 2012a,

but see Borja et al., 2012b).

Reference conditions are a description of the

biological quality elements with no or only very

minor disturbance from human activities (Borja et al.,

2004). They are equivalent to High Ecological Status

and define a benchmark for assessment in line with the

WFD. By comparison of a test site’s biology

(observed) with the water-body type-specific refer-

ence conditions (expected), the Ecological Status is

derived and expressed as ecological quality ratio

(EQR), ranging from 0 (Bad status) to 1 (High status).

The WFD identifies four options for deriving

reference conditions: (i) using the conditions at

existing undisturbed sites or sites with only very

minor disturbance, (ii) using historical data and

information on formerly undisturbed sites, (iii) apply-

ing modelling techniques to predict reference condi-

tions, or (iv) using expert judgement (Vincent et al.,

2002). Advantages and disadvantages of the four

options were recently highlighted by Borja et al.

(2012a). In some European regions and/or aquatic

ecosystem categories, one of the main problems arises

from the absence of unimpacted areas. Moreover,

setting reference conditions is particularly challenging

for transitional waters (i.e. estuaries and lagoons), due

to the difficulty of differentiating natural from anthro-

pogenic variability, the so-called estuarine quality

paradox (Elliott & Quintino, 2007). In the case of

lagoons, which are naturally eutrophic ecosystems

(Basset et al., 2012a) with strong spatial and temporal

variability, this is even more difficult.

In this study, we focus on Eastern Mediterranean

and Black Sea lagoons, which form a geo-morpho-

logically defined type of transitional waters according

to the WFD. The heterogeneity of environmental

conditions in lagoons has two main components: (i) an

inter-lagoon component, which is accounted for by

lagoon typology (sensu WFD, 2000) depending on the

most significant natural sources of variation, and

(ii) an intra-lagoon component, which arises from

habitat patchiness and seasonality and represents a

major source of uncertainty in the assessment of type-

specific reference conditions (Comin et al., 2004;

Arocena, 2007).

The first attempt to develop a general framework

for categorising the spatial heterogeneity of lagoons

into lagoon types dates back to the Venice System

(Battaglia, 1959) and confinement theory (Guelorget

& Perthuisot, 1983). More recent and specific attempts

to identify the most relevant sources of benthic

macroinvertebrate variation in Mediterranean and

Black Sea lagoons have highlighted the importance

of lagoon surface area (Basset et al., 2006, 2007), tidal

range (Basset et al., 2006, 2007; Barbone & Basset,

2010), water salinity (Boix et al., 2005; Basset et al.,

2007; Lucena-Moya et al., 2009; Barbone et al.,

2012a) and the presence of aquatic flora (Barbone

et al., 2012a).

Although threshold values for surface area, degree

of confinement (related to openness), tidal range, and

salinity gradients have been specified in Mediterra-

nean and Black Sea basins, the national transitional

water typologies of Member States differ, for instance,

in weighting these parameters. Surface area, tidal

range and salinity are considered in the national

lagoon typology of Italy (ISPRA, 2007), where

salinity thresholds are derived from the Venice System

(Battaglia, 1959) and tidal range and lagoon surface

area thresholds are set with reference to the distribu-

tion of values in the full set of Italian lagoons (Basset

Hydrobiologia

123

Author's personal copy

et al., 2006). Overall, 20 lagoon types are specified by

Italy (based on 2 levels of tidal range, 2 of surface area

and 5 of water salinity; ISPRA, 2007), grouped into 3

major macrotypes (ISPRA, 2010). On the other hand,

the Greek national transitional water typology con-

siders a single lagoon type, ‘Coastal Lagoons’

(HCMR, 2010; internal document). Italy has also

recently reorganized its national typology in its water

monitoring legislation (DM 56/2008, DM 131/2008,

DM 260/2010), considering a single lagoon type,

‘Coastal Lagoons’, for non-tidal lagoons (sensu Basset

et al., 2006). The Mediterranean lagoon typology

drawn up by the Mediterranean Geographical Inter-

calibration Group (MEDGIG) also includes the con-

cept of confinement, classifying lagoons into leaky,

restricted and choked, combined with water salinity

in a two-factor classification system (Reizopoulou,

personal communication).

According to the main sources of variation consid-

ered in the different typologies, every lagoon can be

classified into a specific type; long-term oscillations of

climate events (Comin et al., 2004) and changes in

hydro-dynamism or other environmental niche shifts,

affecting the abiotic descriptors of lagoon types, can

determine shifts of specific lagoons from one type to

another.

Even when a detailed typology with many different

types is adopted, significant spatial heterogeneity in

terms of hydro-morphology, habitat structure and

physico-chemical characteristics still occurs to some

extent within each lagoon, typically resulting in a

mosaic of different habitats (Escavarage et al., 2004).

Factors determining submerged landscape patchiness

in Mediterranean and Black Sea lagoons, such as

sediment grain size and organic carbon (Teske &

Wooldridge, 2003; Reizopoulou & Nicolaidou, 2004),

the presence and abundance of benthic macrophytes

(Kafanov & Plekhov, 2001; Arocena, 2007; Galuppo

et al., 2007), macrophyte type (seagrasses vs. sea-

weeds; Galuppo et al., 2007) and water depth (De

Casabianca & Posada, 1998) are known to affect

macroinvertebrate distribution, abundance and body

size (Basset et al., 2007, 2008).

The present study has the following aims:

1. to evaluate the classification strength of both

taxonomically and non-taxonomically based mul-

timetric indices with regard to current Mediterra-

nean and Black Sea lagoon typologies;

2. to investigate the most important environmental

factors controlling the variability of assessment

indices for Mediterranean and Black Sea lagoons;

3. to propose a methodology to set type-specific

classification boundaries for assessment indices;

4. to test the classification strength of new type-

specific boundaries for the classification of Med-

iterranean and Black Sea lagoon reference sites

using an independent validation set of lagoons/

lagoon areas.

Materials and methods

Study sites

Data on benthic macroinvertebrates were collected

from 14 lagoons located in 4 different countries: (from

west to east) Italy, Albania, Greece and Romania. Data

were collected during the TWReferenceNET Project

(Basset et al., 2008), as part of the INTERREG IIIB

CADSES (Central European, Adriatic, Danubian,

South-Eastern, European Space) Program (Fig. 1).

Reference lagoons were selected according to

expert judgement (based on pressure variables) and

the protective status of an ecosystem. The latter was

based on the assumption that the location within a

protected area is a guarantee that actual pressures are

at least controlled and in most cases minimised and

that a management plan for the conservation of good

environmental status for the whole lagoon landscape

has been developed. This includes a few relatively

pristine areas and/or areas that are highly valuable for

conservation purposes, such as Ramsar areas, Sites of

Community Interest, Special Protected Areas and

regional/national Parks. The lagoons were also pre-

classified according to existing pressures as described

and evaluated in Barbone et al. (2012a, b).

Disturbed lagoons and areas within lagoons (rang-

ing moderate/poor/bad status), respectively, were used

as an independent dataset for validation of type-

specific boundaries derived from the reference

lagoons, in order to account for potential overestima-

tion of the ecological status with the definition of new

type-specific boundaries.

The selected lagoons varied from microtidal to non-

tidal, large to small, hyperhaline to oligohaline and

choked to restricted ecosystems (Kjerfve & Magill,

Hydrobiologia

123

Author's personal copy

1989; Basset et al., 2006) (Table 1). Within each

ecosystem, spatio-temporal sources of variability that

were known to affect the quality and distribution of the

benthic fauna were taken into account, such as habitat

patchiness and seasonality. In order to account for

habitat patchiness, the sites were selected from two or

three dominant habitat types defined by a factorial

classification of sediment granulometry and vegeta-

tion cover/type as in Basset et al. (2008). Detailed

information on these lagoons including benthic com-

munity composition and distribution patterns and their

relationships to the typology drivers considered are

reported in Basset et al. (2008) and Barbone et al.

(2012a) with regard patterns of variation of simple

metric and in Barbone et al. (2012b) with regard

simple metric shifts with salinity changes in a saltern

ecosystem.

Overall, 101 reference sites were sampled within

the 14 lagoons, with 2/3 habitat types per reference

lagoon, 2 sites per habitat type and 5 replicates per site.

The same experimental design was applied to an

additional 57 study sites, considered here as disturbed

sites and used for validation of the new type-specific

boundaries derived from the reference sites.

Sampling procedures and analysis of samples

Sampling was carried out in autumn 2004 and spring

2005 (temporal component), which may represent a

critical period of discontinuity in Eastern Mediterra-

nean and Black Sea lagoons. Sampling sites in each

ecosystem were chosen so as to include a variety of

habitat types (mud/sand, with/without submerged and

Fig. 1 Map of study sites. 1 Grado Marano Lagoon, 2 Grado

‘Valle Cavanata’, 3 Grado fish farm, 4 Margherita di Savoia Salt

pans, 5 Torre Guaceto brackish Wetland, 6 Le Cesine brackish

Wetland, 7 Lake Alimini, 8 Patok Lagoon, 9 Karavasta Lagoon,

10 Narta Lagoon, 11 Logarou Lagoon, 12 Agiasma Lagoon, 13Sinoe Lagoon, 14 Leahova Lagoon

Hydrobiologia

123

Author's personal copy

Ta

ble

1L

ist

of

stu

die

dtr

ansi

tio

nal

wat

erec

osy

stem

ssh

ow

ing

geo

gra

ph

ical

po

siti

on

and

po

ten

tial

sou

rces

of

inte

r-an

din

trae

cosy

stem

het

ero

gen

eity

Lag

oo

nC

ou

ntr

yP

rote

ctio

nC

oo

rdin

ates

Su

rfac

e(k

m2)

Sal

init

yC

on

fin

emen

tV

eget

atio

nS

edim

ent

Dep

th(c

m)

Ox

yg

en(m

g/l

)

Lat

.L

on

g.

Mea

nS

EM

ean

SE

Gra

do

Mar

ano

Ital

yc

45

.74

2�

13

.22

1�

14

2.0

02

;3

22

11

17

.52

8.4

8.3

0.0

5

Gra

do

Val

leC

avan

ata

Ital

ya,

b4

5.7

12

�1

3.4

70

�2

.10

32

11

90

.05

.87

.60

.1

Gra

do

Val

lid

aP

esca

Ital

yc

45

.72

2�

13

.36

2�

0.3

03

;4

22

;3

16

1.3

8.1

8.3

0.5

Mar

gh

erit

ad

iS

avo

iaIt

aly

a,b

,c

41

.42

9�

15

.98

8�

12

.00

41

2;

31

38

.82

.38

.10

.5

To

rre

Gu

acet

oIt

aly

a4

0.7

11

�1

7.7

95

�1

.60

21

1;

31

45

.02

.05

.40

.8

Ces

ine

Ital

ya,

b4

0.3

58

�1

8.3

35

�0

.70

1;

21

1;

31

,2

10

8.3

12

.08

.70

.3

Ali

min

iIt

aly

b,

c4

0.2

02

�1

8.4

46

�1

.40

3;

41

1;

21

,2

10

6.7

14

.87

.90

.2

Pat

ok

Alb

ania

b,

c4

1.6

35

�1

9.5

90

�7

.10

3;

42

2;

31

50

.06

.09

.40

.5

Kar

avas

taA

lban

iaa

40

.92

0�

19

.47

2�

45

.00

41

2;

31

12

0.0

10

.09

.60

.6

Nar

taA

lban

iac

40

.52

9�

18

.42

6�

29

.90

41

21

47

.57

.59

.71

.5

Lo

gar

ou

Gre

ece

b3

9.0

62

�2

0.9

00

�2

4.2

03

;4

12

11

15

.02

6.0

6.9

1.0

Ag

iasm

aG

reec

ec

40

.88

0�

24

.62

0�

3.2

03

;4

13

17

0.0

2.7

7.1

0.6

Sin

oe

Ro

man

iaa

44

.62

0�

28

.88

8�

12

9.6

01

11

;2

1,

25

4.2

9.5

8.8

0.4

Lea

ho

va

Ro

man

iaa

44

.72

7�

29

.02

8�

22

.90

11

1;

21

84

.01

6.6

7.7

1.0

Sal

init

y:

1o

lig

oh

alin

e(\

0.5

);2

mes

oh

alin

e(5

–1

8);

3p

oly

hal

ine

(18

–3

0);

4eu

hal

ine

(30

–4

0).

Co

nfi

nem

ent:

1ch

ok

ed;

2re

stri

cted

;b

ased

on

lag

oo

no

pen

nes

s.V

eget

atio

n:

1em

erg

ed;

2ab

sen

ce;

3su

bm

erg

ed.

Sed

imen

t:1

mu

d;

2sa

nd

.P

rote

ctio

n:

aR

amsa

rsi

te;

bN

atu

ra2

00

0si

te;

clo

cal

pro

tect

ion

pla

n

Hydrobiologia

123

Author's personal copy

emergent vegetation), levels of lagoon confinement

(choked/restricted) and different depths (Table 1).

For each sampling site five replicates were col-

lected by manual Reineck box-corer (0.03 m2). Each

sample was sieved through a 0.5-mm mesh and the

remaining material was preserved in toto in jars

containing 4 % buffered formaldehyde in seawater.

Benthic samples were sorted under a stereomicroscope

in the laboratory, identified to the lowest possible

taxonomic level, counted, measured individually

(total length for most taxa) to the nearest 0.01 mm

using an image analysis device (Leica Qwin) and

weighed to the nearest 1 lg after drying for 72 h at

60�C. Individual weight was determined for all

specimens that were not damaged or broken while

handling the samples. Broken individuals (mostly

annelids) were handled only in the presence of at least

two-thirds of the individual body, including the head

capsule. Ash content was determined by combustion

in a muffle furnace for 24 h at 500�C. Large specimens

were combusted individually, whereas smaller

animals were combusted in groups of the same size

class. Individual body size was then transformed to

biomass using individual ash-free dry mass (AFDM).

For \5 % of the total sample, length–body mass

relationship was computed at the population level, as

direct measures of body size and AFDM were

impossible. In total, 31,569 individuals were identi-

fied, 99.1 % of which were subjected to body size

estimations. Thus, \1 % of the total sampled data

remained unconsidered (Basset et al., 2012b).

Physico-chemical water parameters (water salinity,

dissolved oxygen and temperature) were monitored

close to the bottom at each station during sampling

activities using a hand-held multi-probe (YSI 556).

Approaches to the definition of lagoon types

In order to define type-specific reference conditions

for Mediterranean and Black Sea lagoons, we fol-

lowed two different approaches. First, we applied

three existing typologies (Type A—‘Coastal Lagoons’;

Type B1—MEDGIG proposal; Type B2—Italian

extended typology proposal) and derived type-specific

reference conditions by using a sub-set of lagoons

for each type and calculating assessment indices

(Table 2). This approach, which is fully compliant

with the WFD, considers typology as metric a-specific,

implicitly assuming that all multimetric indices used

as assessment tools weigh potential sources of natural

variability in lagoons in exactly the same way.

An alternative approach (a posteriori approach;

Table 2), which is also compliant with the WFD,

implies a more complex definition of typology (i.e.

Type C—metric-specific proposal). It was developed

by deriving the optimal lagoon typology from analysis

of those environmental factors (i.e. environmental

lagoon niche dimensions) that significantly reduce the

uncertainty of ecological status assessment in refer-

ence lagoons. Hence, type-specific reference condi-

tions might also be metric specific, as uncertainty in

the assessment of ecological status by means of

multimetric indices can, in principle, be minimised by

accounting for a number of environmental factors. The

environmental factors included in the proposals for a

Mediterranean and Black Sea lagoon typology (see

factors listed in Table 2), as well as those determining

most of the spatial and temporal landscape heteroge-

neity within lagoon ecosystems, should be taken into

Table 2 Approaches to the definition of Mediterranean and

Black Sea lagoon types (Type A—Greek proposal; Type B1—

MEDGIG proposal; Type B2—Italian extended typology pro-

posal; Type C—metric-specific proposal)

National typologies (a priori approach)

A

Type a-specific

Metric a-specific

One type ‘Lagoons’

B

Type specific

Metric a-specific

Many types factorially combining drivers

B1

Confinement ? salinity

B2

Salinity ? surface area ? tidal range

Ottimized typology (a postetiori approach)

C

Type specific

Metric specific

Many types resulting from mixed model analysis of drivers

Potentially different types for different metrics

Hydrobiologia

123

Author's personal copy

consideration as potential sources of multimetric

index variation. Mixed-model procedures have

already been used to quantify the most significant

sources of metric variation in reference lagoons

(Barbone et al., 2012a).

Data analysis

Altogether, 4 multimetric indices were calculated

based on benthic invertebrate samples: Benthic

Assessment Tool (BAT; Marques et al., 2009; Teixeira

et al., 2009), Benthic Index based on Taxonomic

Sufficiency (BITS; Mistri & Munari, 2008), M-AMBI

(Borja et al., 2004; Muxika et al., 2007) and ISS

(Basset et al., 2012b).

While oxygen and depth were treated as continuous

variables, salinity and surface area were transformed

into categorical variables in accordance with the

Venice System (oligohaline = 0.5–5, mesohaline =

5.1–18, polyhaline = 18.1–30, euhaline = 30–40,

hyperhaline [ 40) and in accordance with the Med-

iterranean and Black Sea lagoon typology, respec-

tively, set out in Basset et al. (2006) (small\2.5 km2;

large [2.5 km2). Hyperhaline sites were underrepre-

sented (four sites only at Margherita di Savoia; mean

salinity: 58.21) and thus included in the euhaline

group. Typological descriptors were transformed into

categorical variables: lagoon confinement (two levels:

choked and restricted) based on lagoon openness

(Barbone & Basset, 2010), sediment grain size (two

levels: mud and sand), vegetation (three levels: no

vegetation, submerged, emergent) and season (two

levels: spring and fall).

Normality of variable distribution was checked for

continuous variables (oxygen, depth, M-AMBI, ISS,

BAT, BITS) using Q–Q plots. Outliers were identified

using visual examination of Cleveland dot-plots, box-

plots and scatter-plots (Zuur et al., 2009). Accord-

ingly, depth was log-transformed in order to approach

normality.

A general mixed-model approach (LME, Pinheiro

& Bates, 2000) was used to analyse the relationship

between environmental predictors and multimetric

indices (Barbone et al., 2012a). Mixed modelling takes

account of correlation and variance heterogeneity

within groups of observations and hence is often used

to estimate regression coefficients when covariation

exists among samples, for example, when repeated

measurements are taken at the same site (Zuur et al.,

2009). Here, the covariance is considered an obstacle

that needs to be removed to correctly estimate

regression coefficients.

As a first step, four general linear regression models

were fitted using all explanatory variables in a full

model:

Y ¼ salinity � surface þ log depthð Þ � sediment

� vegetationþ seasonþ oxygenþ confinement,

where an asterisk indicates interaction terms. We

considered the interaction among (log(depth) * sedi-

ment * vegetation) as a proxy of habitat variability,

and the interaction between salinity and surface as a

proxy of Mediterranean lagoon typology. The valida-

tion procedures for the four models (Q–Q plots;

residuals vs. fitted values) showed evidence of residual

heterogeneity among transitional water bodies and

unequal variance among the levels of some explana-

tory variables. Consequently, the four models were

independently developed.

The importance of independent variables was

evaluated by a likelihood ratio test, consisting of a

comparison of the ‘beyond optimal’ model with

models in which explanatory variables and interaction

terms were omitted (significance level: 5 %). The

coefficients of the final optimal LME models (M1–M4)

were calculated using restricted maximum likelihood

(REML). Mixed-effects modelling was conducted

using the ‘nlme’ package (Pinheiro et al., 2006) within

the ‘R’ statistical and programming environment

(R Development Core Team, 2006). The results of

the mixed-effect models were also used to define

typology a posteriori, at the metric-specific level.

One-Way ANOVA was used in order to assess and

compare the variability explained by factors consid-

ered in the a priori and a posteriori approaches to

lagoon typology classification.

Classification of ecological status was performed

by two different procedures: (i) using the standard

published boundaries for each multimetric assessment

tool and (ii) defining new boundaries at the lagoon

type level for each multimetric assessment tool, with

reference to the distribution of values observed in the

sample of reference sites available in the dataset for

each type (type-specific boundaries).

The ecological classification of 101 reference sites

with the first of these procedures was carried out in

accordance with the standard boundaries available in

Hydrobiologia

123

Author's personal copy

the literature for M-AMBI (Borja et al., 2004), BAT

(Teixeira et al., 2009), BITS (Mistri & Munari, 2008)

and ISS (Basset et al., 2012b).

Regarding the second procedure, the new type-

specific boundaries were refined for each type in the

same reference sites by setting the boundary between

high (H) and good (G) at the 50th percentile of the

values of each assessment tool, scaling all other

boundaries in accordance with the scaling procedures

originally used by the tool’s authors (Muxika et al.,

2007; Mistri & Munari, 2008; Teixeira et al., 2009;

Basset et al., 2012b). As an example, the new

boundary between good and moderate (M) classes is

equal to the new high-good boundary multiplied by the

ratio between standard good-moderate and high-good

boundaries [new G–M = new H–G * (standard G–M/

H–G)].

The proposed Mediterranean and Black Sea lagoon

typologies were slightly refined by grouping water

salinity into just two levels (i.e. B30; [30), in

accordance with the results of the mixed-model

analysis, to account for the small number of sites in

the oligohaline and mesohaline types. Therefore, new

type-specific boundaries were set for 8 types based on

the Italian typology (2 salinity levels, two surface area

levels, two tidal range levels) and for 4 types based on

the MEDGIG typology (2 salinity levels and 2

‘confinement’ levels). In the former, even after this

simplification, the lack of small lagoons with a tidal

range above 0.5 m in the dataset prevented the

definition of new boundaries for 2 out of the 8 types.

New type- and metric-specific classification bound-

aries were also defined by considering significant

variables in the mixed-effect models. The sites were

thus divided into 8 types when classification was

performed with BAT and M-AMBI (2 salinity levels, 2

surface area levels, 2 seasonal levels), into 12 types

when classification was performed with BITS (2

salinity levels, 3 vegetation type levels, 2 confinement

levels) and into 4 types, with a normalization of values

along the continuous variable oxygen in accordance

with the model equation, when the classification was

performed with ISS (2 salinity levels, 2 seasonal

levels).

The accuracy of the two classification procedures,

given as the distance of the assessed ecological status

from the high/good status expected for reference

ecosystems, was then tested within and among lagoon

typologies by comparing the distribution of sites in the

five ecological status quality groups (contingency

analysis; v2 test).

The same data analyses described above were

applied to an independent set (taken from the TWRe-

ferenceNet database) of disturbed ecosystems, or

disturbed sites within ecosystems, where strong

pressure gradients were identified (Basset et al.,

2008, 2012b).

Validation was performed both at the study site

level and at the lagoon/lagoon area level. The accuracy

of the classification of reference and disturbed lagoon

conditions was compared with the various methodo-

logical approaches so far proposed to define Mediter-

ranean and Black Sea lagoon typology (contingency

analysis; v2 test).

Results

Sources of variation

Overall, the multimetric indices showed significant

variability both within and among the studied refer-

ence ecosystems (Table 3). They varied significantly

among lagoon ecosystems (One-way ANOVA,

P \ 0.01, for all metrics) and were also significantly

affected by the environmental factors considered

(Table 4). As a general trend, multimetric index

values increased with salinity from oligohaline to

euhaline lagoons and with lagoon surface area.

Moreover, values were higher in spring than in fall

and, within lagoons, in habitats with submerged

vegetation or no vegetation than in habitats with

emergent vegetation (Table 3).

Variability was higher for BITS than for the other

indices. At the ecosystem level, the coefficients of

variation ranged 30–35 % for BAT, M-AMBI and

ISS, but 66 % for BITS. At the study site level, BAT,

M-AMBI and ISS were significantly less variable than

BITS (F-ratio, F99,99 3.83, F99,99 4.71, F99,99 3.61,

respectively).

The relevance of the various potential sources of

natural variation of the multimetric indices included in

Table 3 in terms of both typology and intra-lagoon

habitat patchiness was further analysed with an a pos-

teriori approach, using mixed models, at the metric-

specific level. Quantitatively, each multimetric index

showed co-variation patterns with various environ-

mental factors (mixed-effect models, Table 4). Among

Hydrobiologia

123

Author's personal copy

the factors considered in the proposed typologies, as

single or in interactions (Table 4): (i) water salinity and

seasonality were found to be significant sources of

variation for all multimetric indices, (ii) surface area

significantly affected BAT and M-AMBI, (iii) con-

finement and internal sources of variation, such as

sediment type or vegetation, were found to be a source

of variation for BITS and BAT (only vegetation for the

latter) and (iv) tidal range was not a significant source

of variation for any of the multimetric indices.

The four multimetric indices were affected to

differing degrees by the environmental factors. BAT

was significantly affected by season, being higher in

spring than in fall (Table 4; Fig. 2a), and by the

interactions between salinity and surface area

(Table 4; Fig. 2b) and between sediment and vegeta-

tion (Table 4; Fig. 2c).

BITS varied significantly as a function of lagoon

salinity, lagoon confinement, habitat patchiness and

season, being higher in polyhaline and euhaline than

oligohaline and mesohaline waters (Table 4; Fig. 3a),

in choked than in restricted lagoons (Table 4; Fig. 3b),

in the presence of submerged/no vegetation than in

patches with emergent vegetation (Table 4; Fig. 3c)

and in spring than in fall (Table 4; Fig. 3d).

M-AMBI was also significantly affected by lagoon

salinity and season, being higher in euhaline lagoons

than in polyhaline, mesohaline and oligohaline con-

ditions (Table 4; Fig. 4a) and in spring than in fall

(Table 4; Fig. 4b). It was also influenced by the

interaction between salinity and surface area (Table 4;

Fig. 4c).

ISS was significantly influenced by salinity levels

(Table 4; Fig. 5a), being higher in euhaline than in

polyhaline, mesohaline and oligohaline conditions,

and by season (Table 4; Fig. 5b), being higher in

spring than in fall. ISS increased with oxygen

concentration (Table 4; Fig. 5c).

Globally, the factors considered in the a priori and

a posteriori typologies significantly reduced the

unexplained residual variance of the overall dataset

(i.e. Type A—‘Coastal Lagoons’ typology) (One-way

ANOVA, P \ 0.01 for all typology schemes and

assessment tools). On average, the more complex

Table 3 Mean BAT, BITS, M-AMBI and ISS values measured at each level of factors for reference sites

Factors BAT BITS M-AMBI ISS

Salinity

Oligohaline 0.44 (0.04) 0.57 (0.11) 0.44 (0.04) 2.66 (0.15)

Mesohaline 0.54 (0.06) 1.08 (0.21) 0.55 (0.05) 2.65 (0.20)

Polyhaline 0.55 (0.03) 1.73 (0.20) 0.61 (0.04) 2.71 (0.16)

Euhaline 0.62 (0.03) 2.11 (0.13) 0.75 (0.03) 3.37 (0.14)

Surface area

Large 0.56 (0.03) 1.63 (0.15) 0.67 (0.03) 3.01 (0.12)

Small 0.54 (0.03) 1.40 (0.17) 0.58 (0.03) 2.86 (0.12)

Confinement

Choked 0.53 (0.02) 1.59 (0.12) 0.59 (0.02) 2.88 (0.11)

Restricted 0.63 (0.03) 1.34 (0.17) 0.73 (0.04) 3.15 (0.15)

Sediment granulometry

Mud 0.55 (0.02) 1.77 (0.15) 0.64 (0.02) 2.92 (0.10)

Sand 0.55 (0.04) 2.06 (0.18) 0.54 (0.04) 3.06 (0.23)

Vegetation

No vegetation 0.55 (0.02) 1.77 (0.15) 0.63 (0.04) 2.97 (0.13)

Submerged vegetation 0.58 (0.03) 1.52 (0.17) 0.67 (0.03) 3.00 (0.17)

Emerged vegetation 0.53 (0.03) 1.04 (0.18) 0.57 (0.04) 2.79 (0.14)

Season

Fall 0.53 (0.03) 1.48 (0.14) 0.60 (0.03) 2.90 (0.14)

Spring 0.55 (0.02) 1.53 (0.15) 0.63 (0.03) 2.98 (0.11)

Standard errors are shown in brackets

Hydrobiologia

123

Author's personal copy

typology schemes explained 21 % of the assessment

tools’ variability (measured as sums of squares)

among sites; the MEDGIG typology explained 19 %,

the mixed-model-based typology 20 % and the more

complex extended Italian typology 23 %. At the

assessment tool level, the average explained variabil-

ity was highest for M-AMBI (30 %) and lowest for

BAT (15 %).

Ecological classification

Type a-specific classification boundaries for the

assessment of lagoon ecological status based on the

selected multimetric indices are shown in Tables 5

and 6 as ‘standard boundaries’. The same tables also

show the type-specific boundaries (‘new boundaries’)

between ecological status classes computed for lagoon

typologies listed in Table 2 (i.e. Type A—‘Coastal

Lagoons’; Type B1—MEDGIG proposal; Type B2—

Italian typology proposal; Type C—a posteriori met-

ric-specific proposal).

When a single ‘Coastal Lagoons’ type was consid-

ered (Table 2; Type A), the classification using the

standard published boundaries strongly underesti-

mated the ecological status of the reference sites.

ISS and M-AMBI classified 56 % of the ‘reference’

sites as having high or good status, BITS 48 % and

BAT 46 %. The accuracy of the ecological status

classification was significantly improved by using the

new boundaries derived in this study (contingency

Table 4 Linear mixed-model results for environmental factors’ effects on multimetric indices of macroinvertebrate communities in

14 Mediterranean and Black Sea Lagoons

Model Response Model terms Factors df L ratio P

M1 BAT Salinity f 3 4.55 0.2075

Season f 1 5.29 0.0213

Surface f 1 0.42 0.5142

Salinity * surface f 3 19.88 \0.0001

Sediment f 1 0.52 0.4691

Vegetation f 2 2.10 0.3493

Sediment * vegetation f 2 7.10 0.0287

Lagoon r 1 12.51 \0.0001

Surface * salinity v–c 7 9.71 0.2052

M2 BITS Salinity f 3 23.11 \0.0001

Confinement f 1 11.40 \0.0001

Vegetation f 2 21.53 \0.0001

Season f 1 20.73 \0.0001

Lagoon r 1 3.10 0.0391

Surface * salinity v–c 7 31.43 \0.0001

M3 M-AMBI Salinity f 3 9.76 0.0207

Surface f 1 1.37 0.2414

Surface * salinity f 3 12.79 0.0051

Season f 1 9.31 0.0023

Lagoon r 1 10.36 0.0013

Surface * salinity v–c 7 7 0.0350

M4 ISS Salinity f 3 9.03 0.0289

Season f 1 7.76 0.0053

Oxygen f 1 4.61 0.0317

Lagoon r 1 5.99 0.0144

Surface * salinity v–c 7 15.58 0.0291

Only terms that are significant for optimal model are shown. Non-significant factors that were part of significant interactions were not

removed. Factors (f fixed; r random; v–c variance–covariance)

Hydrobiologia

123

Author's personal copy

analysis: BAT, v2 = 37.44, df = 4, P \ 0.01; BITS,

v2 = 5.15, df = 4, P = ns; M-AMBI, v2 = 24.37,

df = 4, P \ 0.01; ISS, v2 = 40.34, df = 4, P \ 0.01),

with 74 % of sites classified as good or high on

average, ranging from 62 % (BITS) to 83 % (ISS).

Similar results were observed for the other pro-

posed typologies as well as for the typology derived

a posteriori using the mixed-model approach. With all

typologies, the new type- and metric-specific bound-

aries (Tables 5, 6) significantly improved the accuracy

of the ecological status assessment for the studied

lagoons (contingency analysis: all cases P \ 0.05).

According to BAT, the proportion of reference sites

classified as good and high increased from 46 to 74 %

for Type B1, 73 % for Type B2 and 80 % for Type C;

with M-AMBI, the proportion increased from 56 to 84,

78 and 77 %; with ISS, from 56 to 81, 84 and 85 %;

and with BITS, from 48 to 66, 63 and 63 %. However,

on average, ecological status was still underestimated

at more than 25 % of the reference sites.

The average accuracy of the ecological status

assessment was not affected by the typologies tested

in this article (contingency analysis; P [ 0.05 for all

multimetric indices).

Validation procedures

The accuracy of the new type-specific reference

condition boundaries was validated using an indepen-

dent set of data on six lagoons, or lagoon areas, which

were known to be affected by strong anthropogenic

pressures (Basset et al., 2008, 2012b), using ISS as an

assessment tool. The data referred to a total of 57 new

study sites. At the study site level, the accuracy of the

ecological status classification of disturbed sites was

higher with the standard boundaries than with the

type-specific boundaries developed in this study. Only

19 % of the disturbed sites were classified as having

good or high status using the standard boundaries,

while on average 53 % of sites were classified as good

b a

c

Fall SpringSeason

0.0

0.2

0.4

0.6

0.8

1.0

BA

T

Oligo.LMeso.L

Poly.LEu.L

Oligo.SMeso.S

Poly.SEu.S

Salinity*Surface

0.0

0.2

0.4

0.6

0.8

1.0

Mud.EmeMud.No

Mud.SubSand.Eme

Sand.NoSand.Sub

Sediment*Vegetation

0.0

0.2

0.4

0.6

0.8

1.0

BA

T

Fig. 2 Comparisons of mean BAT values under environmental

conditions found to be significant sources of variation with

mixed-modelling statistical approach (MLE). a Sampling sea-

son; b interaction between salinity and surface area (L large;

S small); c interaction between sediment and vegetation (Eme

emergent vegetation; Sub submerged vegetation; no no vege-

tation). Central lines Median value; boxes range between lower

and upper quartiles (i.e. 25th–75th percentile); whiskers full

range of data, excluding outliers

Hydrobiologia

123

Author's personal copy

or high status using the new type-specific boundaries

(Table 7).

Averaging the study site values at the lagoon level,

the accuracy of ecological status classification was

higher than at the study site level. When the new

boundaries were set, all reference lagoons or lagoon

areas were classified as good or high with the Type B2

and C approaches and 93 % with the Type A approach

(Table 8). However, while the Type C approach

classified 50 % of disturbed lagoons or lagoon areas

as having good ecological status, the A, B1 and B2

approaches classified 83 % of them as moderate

(Table 8).

Discussion

The challenges involved in classifying ecological

status in Mediterranean and Black Sea lagoons are

well known (Basset, 2010). On the one hand, lagoons

are ecotone ecosystems (Basset et al., 2012a) naturally

enriched and characterised by strong natural gradients;

on the other hand, lagoon colonisers’ adaptations to

these conditions include both resistance and resilience

to anthropogenic external pressures. Hence, distin-

guishing between natural and anthropogenic pressures

is particular challenging and a prerequisite for the

quantification of biological responses to anthropo-

genic stress and the setting of class boundaries for

assessment (Birk & Hering, 2009; Dauvin & Ruellet,

2009), which need to be adjusted when compared to

ecosystem categories with lower natural disturbance

(Prato et al., 2009). This investigation highlights a

number of methodological issues that are key to

addressing and overcoming these challenges.

The results achieved in this study do not seem to

depend on the criteria used to define reference

conditions (Vincent et al., 2002; Borja et al., 2012a).

The criteria are consistent with the options proposed

by the WFD and with the available literature on

freshwater and marine ecosystems (Stoddard et al.,

2006; Borja et al., 2009a), as well as with a recent

a

d

b

c

Oligo Meso Poly Eu

Salinity

0

1

2

3

4

BIT

S

choked restricted

Lagoon Confinement

0

1

2

3

4

Emerged No Submerged

Vegetation

0

1

2

3

4

BIT

S

Fall Spring

Season

0

1

2

3

4

Fig. 3 Comparisons of mean BITS values under environmental

conditions found to be significant sources of variation with

mixed-modelling statistical approach (MLE). a Salinity levels

(Oligo oligohaline; Meso mesohaline; Poly polyhaline;

Eu euhaline); b lagoon confinement; c vegetation type

(emergent vegetation, no vegetation and submerged vegetation);

d sampling season. Central lines median value; boxes range

between lower and upper quartiles (i.e. 25th–75th percentile);

whiskers full range of data, excluding outliers

Hydrobiologia

123

Author's personal copy

study performed on Mediterranean and Black Sea

lagoon ecosystems (Barbone et al., 2012a). However,

as the criteria used are based on minimal disturbance

and expert view assessments, which incorporate a

certain degree of uncertainty, quantitative definitions

of the disturbance level beyond which a site can no

longer be considered as exemplifying ‘reference

conditions’ are still lacking. This has already been

acknowledged in other studies (Lucena-Moya et al.,

2009; Neto et al., 2010; Barbone et al., 2012a).

In this investigation, we considered all lagoon

typologies so far proposed by Mediterranean countries

(i.e. Type A—‘Coastal Lagoons’; Type B1—MEDGIG

typology; Type B2—Italian typology) as well as a more

complex definition (Type C—a posteriori metric-spe-

cific typology) in order to derive type-specific reference

conditions that significantly reduce the uncertainty of

assessment in reference condition ecosystems. In this

investigation, we also defined new boundaries for each

proposed typology of Mediterranean and Black Sea

lagoons. A key issue is the definition of the boundary

between ‘Good’ and ‘Moderate’ status, namely what is

‘acceptable’ (undegraded) or ‘not acceptable’ (degraded)

status or when it is necessary to invest resources in the

restoration of an ecosystem (Blanchet et al., 2008). The

applicability of multimetric indices to ecological

classification boundaries was recently analysed in a

study of transitional water ecosystems (Dauvin, 2007;

Ruellet & Dauvin, 2007), which pointed out that

arbitrary boundaries between categories sometimes

remain (Blanchet et al., 2008).

Our approach of defining type-specific reference

conditions and their boundaries in Mediterranean and

Black Sea lagoons is consistent with previous studies in

freshwater ecosystems, for instance, with probability

tools such as River Invertebrate Prediction and Clas-

sification System (RIVPACS; Wright et al., 1993) and

Benthic Assessment of SedimenT (BEAST; Reynold-

son et al., 1995). Such approaches use potential

reference sites for biological status evaluation and

ba

c

Oligo Meso Poly Eu

Salinity

0.0

0.2

0.4

0.6

0.8

1.0M

AM

BI

Fall Spring

Season

0.0

0.2

0.4

0.6

0.8

1.0

Oligo.LMeso.L

Poly.LEu.L

Oligo.SMeso.S

Poly.SEu.S

Salinity*Surface

0.0

0.2

0.4

0.6

0.8

1.0

MA

MB

I

Fig. 4 Comparisons of mean M-AMBI values under environ-

mental conditions found to be significant sources of variation

with mixed-modelling statistical approach (MLE). a Salinity

levels (Oligo oligohaline; Meso mesohaline; Poly polyhaline;

Eu euhaline); b season; c interaction between salinity and

surface area (L large; S small). Central lines median value;

boxes range between lower and upper quartiles (i.e. 25th–75th

percentile); whiskers full range of data, excluding outliers

Hydrobiologia

123

Author's personal copy

assess the level of disturbance on the basis of the

probability that a test site falls within the range of

variation of reference sites. Setting the boundary

between high and good status at the 50th percentile

of metric values in reference ecosystems is also

consistent with previous approaches based on type-

specific reference conditions data, for example, for

Mediterranean and Black Sea lagoons (Barbone et al.,

2012a). A consistent criterion was applied to freshwa-

ter ecosystems by the RIVPACS approach, which

generally considered sites falling within the central

80 % of reference spectra as reference sites. More

restrictive criteria, with the H–G boundary set at or

higher than the 90th percentile, have been used for

transitional and coastal water ecosystems when the

original data were from ecosystems covering a range of

disturbance conditions (Borja & Tunberg, 2011).

In this study, we showed that the lagoon typologies

adopted or proposed at the national or MEDGIG level

incorporate the key sources of variability to which

macroinvertebrate assessment tools are susceptible.

Indeed, all multimetric indices showed patterns of

variations with water salinity (Boix et al., 2005;

Lucena-Moya et al., 2009; Barbone et al., 2012a),

which is a key component of the Venice system and of

confinement theory (Guelorget & Perthuisot, 1983).

Moreover, BAT and M-AMBI showed patterns of

variation with surface area (Basset et al., 2006) and

BITS was affected by confinement. The rationale for

these relationships is discussed in the following

paragraphs.

Regarding lagoon salinity, various hypotheses can

be advanced to explain the patterns of variation shown

by all multimetric indices, which divide euhaline sites

from poly- to oligohaline ones. As nutrient input to

lagoon ecosystems is mainly through freshwater

inflows, these differences on the discrete gradient of

salinity, with lower values of all assessment tools

recorded at poly- to oligohaline sites than at euhaline

ones, might be attributable to processes of natural

eutrophication in low salinity lagoons or lagoon areas.

Indeed, inverse relationships among eutrophication

ba

c

Oligo Meso Poly Eu

Salinity

0

1

2

3

4

5

6IS

S

Fall Spring

Season

0

1

2

3

4

5

6

2 4 6 8 10 12 14

Oxygen (mg/l)

0

1

2

3

4

5

6

ISS

Fig. 5 Comparison of mean ISS values under environmental

conditions found to be significant sources of variation with

mixed-modelling statistical approach (MLE). a Salinity levels

(Oligo oligohaline; Meso mesohaline; Poly polyhaline; Eu

euhaline); b sampling season; c oxygen concentration. Regres-

sion equation ISS = 0.143 * oxy ? 1.79. Central lines median

value; boxes range between lower and upper quartiles (i.e. 25th–

75th percentile); whiskers full range of data, excluding outliers

Hydrobiologia

123

Author's personal copy

Ta

ble

5S

tan

dar

dan

dn

ewb

ou

nd

arie

so

fp

rop

ose

dty

po

log

ies

AB

1B

2

B3

0B

30

[3

0[

30

B3

0[

30

Lar

ge

Sm

all

Lar

ge

Sm

all

Sta

nd

ard

bo

un

dar

ies

New

bo

un

dar

ies

Ch

ok

edR

estr

icte

dC

ho

ked

Res

tric

ted

\0

.5[

0.5

\0

.5[

0.5

\0

.5[

0.5

\0

.5[

0.5

BA

T

Hig

h/g

oo

d0

.80

.60

.50

.60

.60

.70

.50

.70

.50

.60

.70

.7

Mo

der

ate/

go

od

0.6

0.4

0.4

0.4

0.5

0.5

0.4

0.5

0.4

0.5

0.5

0.5

Po

or/

mo

der

ate

0.4

0.3

0.3

0.3

0.3

0.3

0.2

0.3

0.3

0.3

0.3

0.3

Bad

/po

or

0.3

0.2

0.2

0.2

0.2

0.3

0.2

0.2

0.2

0.2

0.3

0.3

BIT

S

Hig

h/g

oo

d1

.81

.41

.10

.82

.21

.50

.61

.21

.22

.12

.11

.9

Mo

der

ate/

go

od

1.4

1.1

0.8

0.6

1.7

1.2

0.4

0.9

0.9

1.7

1.6

1.5

Po

or/

mo

der

ate

0.9

0.7

0.5

0.4

1.1

0.8

0.3

0.6

0.6

1.1

1.0

0.9

Bad

/po

or

0.5

0.4

0.3

0.2

0.6

0.4

0.2

0.3

0.3

0.6

0.6

0.5

M-A

MB

I

Hig

h/g

oo

d0

.80

.60

.50

.60

.70

.80

.50

.80

.50

.80

.80

.9

Mo

der

ate/

go

od

0.6

0.5

0.4

0.4

0.5

0.6

0.3

0.6

0.4

0.6

0.6

0.7

Po

or/

mo

der

ate

0.4

0.3

0.3

0.3

0.3

0.4

0.2

0.4

0.3

0.4

0.4

0.4

Bad

/po

or

0.2

0.2

0.1

0.1

0.2

0.2

0.1

0.2

0.1

0.2

0.2

0.2

ISS Hig

h/g

oo

d4

.02

.92

.82

.73

.73

.42

.72

.22

.93

.62

.53

.0

Mo

der

ate/

go

od

2.8

2.0

1.9

1.9

2.6

2.4

1.9

1.6

2.0

2.5

1.8

2.1

Po

or/

mo

der

ate

2.2

1.6

1.5

1.5

2.1

1.9

1.5

1.2

1.6

2.0

1.4

1.7

Bad

/po

or

1.2

0.9

0.8

0.8

1.1

1.0

0.8

0.7

0.9

1.1

0.8

0.9

A—

Sin

gle

typ

e‘C

oas

tal

Lag

oo

ns’

;B

1—

ME

DG

IGp

rop

osa

l;B

2—

Ital

ian

exte

nd

edty

po

log

y.

Sta

nd

ard

bo

un

dar

ies

app

lyto

all

typ

olo

gie

san

dfo

rty

pes

B1

and

B2

on

lyn

ew

bo

un

dar

ies

are

rep

ort

ed

Hydrobiologia

123

Author's personal copy

Ta

ble

6S

tan

dar

db

ou

nd

arie

san

dty

pe/

met

ric

spec

ific

bo

un

dar

ies

for

fou

rm

ult

imet

ric

ind

ices

con

sid

ered

BA

TS

tan

dar

db

ou

nd

arie

sB

30

B3

0[

30

[3

0

Lar

ge

Sm

all

Lar

ge

Sm

all

Fal

lS

pri

ng

Fal

lS

pri

ng

Fal

lS

pri

ng

Fal

lS

pri

ng

Hig

h/g

oo

d0

.80

.50

.60

.40

.50

.60

.60

.70

.7

Mo

der

ate/

go

od

0.6

0.4

0.4

0.3

0.4

0.5

0.5

0.5

0.6

Po

or/

mo

der

ate

0.4

0.3

0.3

0.2

0.3

0.3

0.3

0.4

0.4

Bad

/po

or

0.3

0.2

0.2

0.1

0.2

0.2

0.2

0.2

0.3

BIT

SS

tan

dar

db

ou

nd

arie

sC

ho

ked

Ch

ok

edR

estr

icte

dR

estr

icte

d

Fal

lS

pri

ng

Fal

lS

pri

ng

Em

erg

edN

oS

ub

mer

ged

Em

erg

edN

oS

ub

mer

ged

Em

erg

edN

oS

ub

mer

ged

Em

erg

edN

oS

ub

mer

ged

Hig

h/g

oo

d1

.81

.12

.21

.20

.82

.51

.20

.11

.00

.90

.71

.82

.3

Mo

der

ate/

go

od

1.4

0.9

1.7

0.9

0.6

2.0

0.9

0.1

0.8

0.7

0.5

1.4

1.8

Po

or/

mo

der

ate

0.9

0.6

1.1

0.6

0.4

1.3

0.6

0.1

0.5

0.5

0.3

0.9

1.1

Bad

/po

or

0.5

0.3

0.6

0.3

0.2

0.7

0.3

0.0

0.3

0.3

0.2

0.5

0.6

M-A

MB

IS

tan

dar

db

ou

nd

arie

sB

30

B3

0[

30

[3

0

Lar

ge

Sm

all

Lar

ge

Sm

all

Fal

lS

pri

ng

Fal

lS

pri

ng

Fal

lS

pri

ng

Fal

lS

pri

ng

Hig

h/g

oo

d0

.80

.50

.70

.50

.50

.70

.80

.70

.9

Mo

der

ate/

go

od

0.6

0.4

0.5

0.4

0.4

0.5

0.6

0.5

0.7

Po

or/

mo

der

ate

0.4

0.3

0.4

0.2

0.3

0.4

0.4

0.3

0.5

Bad

/po

or

0.2

0.1

0.2

0.1

0.1

0.2

0.2

0.2

0.2

ISS

Sta

nd

ard

bo

un

dar

ies

B3

0[

30

Fal

lS

pri

ng

Fal

lS

pri

ng

Hig

h/g

oo

d4

.02

.62

.83

.43

.5

Mo

der

ate/

go

od

2.8

1.8

2.0

2.4

2.4

Po

or/

mo

der

ate

2.2

1.4

1.5

1.9

1.9

Bad

/po

or

1.2

0.8

0.9

1.0

1.0

Hydrobiologia

123

Author's personal copy

and BAT (Neto et al., 2010 or Teixeira et al., 2009),

BITS (Munari et al., 2009), M-AMBI (Borja et al.,

2009a, b) and ISS (Basset et al., 2012b) values have

already been shown. Clearly, streams and rivers can

also transport organic and inorganic pollutants into

lagoon ecosystems, but this seems to be an unlikely

explanation of the patterns observed in the reference

lagoons, as they were affected by low-pressure

intensity. On the other hand, as all indices are, to

some extent, affected by species richness and as most

lagoon species are marine in origin, the higher values

of all assessment tools at euhaline sites might be

attributable to a greater influence of the sea and to a

higher colonisation rate by marine species (Whitfield

et al., 2012).

The variation of BAT and M-AMBI with lagoon

surface area is also consistent with the literature. Both

have a taxonomic-richness component and well-

defined species–area relationships have already been

observed for Mediterranean lagoons (Sabetta et al.,

2007; Guilhaumon et al., 2012). This means that the

definition of clear reference conditions for these indices

must account for species–area relationships, empha-

sised by some authors (Borja et al., 2004, 2012a).

On the other hand, the fact that in terms of direct

single-factor effects, the multimetric indices were not

affected by internal lagoon heterogeneity, with the

exception of BITS, suggests that the integration of

simple metrics increases the robustness of the assess-

ment tools to the spatial patchiness of lagoon

Table 7 Percentage of disturbed sites with ‘bad’, ‘poor’,

‘moderate’, ‘good’ and ‘high’ ecological quality status for each

proposed typology, considering ISS as assessment tool

Type A Type B1 Type B2 Type C

Standard New

High 0.00 14.04 10.53 7.02 8.77

Good 19.30 42.11 42.11 42.11 43.86

Moderate 22.81 17.54 14.04 12.28 14.04

Poor 49.12 17.54 24.56 29.82 24.56

Bad 8.77 8.77 8.77 8.77 8.77

Table 8 Ecological status

of reference lagoons/lagoon

areas and disturbed lagoons/

lagoon areas for each

proposed typology,

considering ISS as

assessment tool

Type A Type B1 Type B2 Type C

Standard New

Reference lagoons/lagoon area

Agi Moderate Moderate Moderate Good Good

Ali Good Good Good Good Good

Ces Good High High High High

GM Moderate Good Good High High

GVC Moderate Good Good Good Good

GVP Moderate Good Good Good Good

Kar High High High High High

Lea Moderate Good Good High High

Log Moderate Good Moderate Good Good

MdS Good High Good Good Good

Narta Good High Good Good Good

Patok High High High High High

Sinoe Moderate Good Good Good Good

TG Moderate Good Good Good Good

Disturbed

GM Poor Moderate Good Good Good

GVP Moderate Moderate Moderate Moderate Good

Log Poor Moderate Moderate Moderate Moderate

MdS Moderate Good Moderate Moderate Moderate

Narta Moderate Moderate Moderate Moderate Moderate

Varna Poor Moderate Moderate Moderate Good

Hydrobiologia

123

Author's personal copy

ecosystems. Indeed, simple metrics have been

observed to be sensitive to patchiness in vegetation,

sediment types or both (Barbone et al., 2012a). A

small degree of sensitivity of BITS to vegetation

habitat type has been shown for Mediterranean

lagoons (Munari et al. 2009), and a case of BAT

sensitivity to vegetation and sediment habitat types

has also been shown in Ria Formosa (Gamito et al.,

2012). However, our results show that a certain degree

of heterogeneity in habitat types sampled at different

spatial scales (lagoon, regional area, eco-regional

area) should not introduce significant biases in the

evaluation process of lagoon ecological status.

This investigation also highlights the important

influence of seasonal patterns on all assessment tools

tested here. It is supported by the fact that all

assessment tools showed significantly higher values

in spring than in fall. These patterns are likely to be

related both to abiotic patterns of temperature and

dissolved oxygen in Mediterranean and Black Sea

lagoons and to species responses in terms of life cycle

adaptation and space use behaviour. Summer crises

are common in Mediterranean and Black Sea lagoons

not only as a result of external inputs of nutrients from

anthropogenic activities (Zaldivar et al., 2008) but also

due to reduced hydrodynamics and wind-driven

mixing (Vignes et al., 2009). This temporal variability

must be taken into account, in order both to standard-

ise time schedules in Mediterranean/Black Sea lagoon

ecological status monitoring programs and to derive

season-specific reference conditions and classification

boundaries, as in this study.

In this investigation, we classified the sites inside

the lagoons, using both standard published boundaries

and new boundaries defined in this article as type-

specific boundaries for each multimetric assessment

tool. This approach is similar to that of the European

intercalibration exercise (Borja et al., 2007, 2009b)

and considers the proposed Mediterranean and Black

Sea lagoon typologies and the more complex typology

obtained with the results of mixed-effect models.

Classification with standard boundaries underesti-

mated the ecological status of the reference sites

considered, 50 % of which on average were classified

as moderate, poor or bad. The level of misclassifi-

cation was independent of the multimetric indices

selected. Consequently, all the indices tested appear to

be potentially appropriate for classifying the ecolog-

ical status of reference lagoons while their boundaries

between ecological quality classes were clearly not

adequate.

The fact that misclassification was not completely

removed by setting the new type- and metric-specific

boundaries, which did, however, significantly improve

the accuracy of all multimetric indices, is attributable

to various factors. Hidden anthropogenic pressures

might be affecting some lagoon areas by lateral or

aerial diffusion. Macroinvertebrate response may also

be the result of historical anthropogenic pressures

stored and slowly released from the sediments (Bar-

bone et al., 2012a). Moreover, natural pressures are

likely to have different intensities at different sites

within lagoons due to patchy sedimentation rates,

spatially explicit gradients of hydrodynamic forces, or

differential influence of fresh and marine waters,

causing locally biased estimates of the real ecological

status of sites within lagoon ecosystems or of lagoons

within lagoon complexes. Misclassification was sig-

nificantly less frequent when new rather than standard

boundaries were used, but still significant. Using type-

and metric-specific reference conditions for the

assessment tools considered, the status of some sites

(25 %) was underestimated, showing the persistence

of potential bias in the classification of reference

lagoon ecosystems in our data set. However, it can also

be argued that the low values of macroinvertebrate

assessment metrics are not necessarily an index of

underestimation of the actual ecological status of

lagoon ecosystems.

The potential misclassification, which was also

observed for the validation data set, where only ISS

was used as an assessment tool, suggested an alterna-

tive explanation based on the adaptation of benthic

macroinvertebrates to very harsh and variable condi-

tions. Macroinvertebrates are likely to be highly

resistant and resilient, as well as pre-adapted, to

disturbed conditions. The use of ISS for the validation

test is justified by the fact that multimetric indices

considered appear to be potentially appropriate for

classifying the ecological status of reference lagoons

and ISS have already been shown to be effective in

detecting anthropogenic pressures (Basset et al.,

2012b).

Both explanations acknowledge a degree of uncer-

tainty in the assessment of lagoon ecological status

with benthic invertebrates, necessitating optimisation

of all procedures in order to avoid misclassification.

Underestimating status may lead to unnecessarily high

Hydrobiologia

123

Author's personal copy

recovery costs where extensive recovery is not

required. In contrast, overestimation may lead to a

delay in remedial action.

The risk of bias in the classification of ecological

status is intrinsically related to the particular condi-

tions of Mediterranean and Black Sea lagoons, which

are naturally eutrophic and highly variable ecosys-

tems. In order to minimise misclassification, and

consequent delays in recovery or misallocation of

funds, Mediterranean and Black Sea lagoon monitor-

ing programmes need to include a sufficient number of

replicate sites within lagoons (which implies greater

effort), seasonal replicates or season-specific refer-

ence conditions and classification boundaries, and

type- (and metric-) specific reference conditions and

classification boundaries at the water-body level. As

assessments compliant with the WFD are made at the

water-body level, the results achieved in this study at

the lagoon level show that a correct sampling design

decreases the risk of misclassification. However,

moderate to bad sites within lagoons with otherwise

good ecological status need surveillance or investiga-

tive monitoring schemes until a clear pressure–

response relationship is quantified at the site level.

Conclusions

The proposed lagoon typologies successfully captured

the major sources of natural variability. Increasing

complexity in typology definition, from the simple

‘Coastal Lagoons’ type to the extended Italian typol-

ogy, increased the accuracy of assessment tools.

Environmental sources of natural variability con-

tributing to lagoon typology and seasonality have a

greater influence on assessment tool variability than

factors determining internal habitat patchiness.

Type- and metric-specific classification boundaries,

setting the H/G threshold at the 50th percentile, were

defined and successfully validated.

Use of the new type and metric-specific boundaries

represents a significant contribution to the accuracy of

ecological status assessment in both reference and

disturbed lagoons, when computed at the ecosystem

level, in disturbed lagoons. The risks of misclassifi-

cation are considered and discussed.

Acknowledgments We thank Joxe Mikel Garmendia, Jose

German Rodrıguez (both from AZTI-Tecnalia, Spain) and

Antoaneta Trayanova (from IO-BAS, Bulgaria) for providing

data and supporting computation of multimetric indices, and

ARPA Puglia for their permission to use regional monitoring

data to support WISER activities. The original data were

collected thanks to the TWReferenceNet project and Regional

Monitoring of transitional waters in Puglia. This article was

supported by the following projects: INTERREG IIIB CADSES

TWReferenceNet, MIUR PRIN08 and WISER (Water bodies in

Europe: Integrative Systems to assess Ecological status and

Recovery), funded by the European Union under the 7th

Framework Programme, Theme 6 (Environment including

Climate Change) (Contract No. 226273), www.wiser.eu.

References

Arocena, R., 2007. Effects of submerged aquatic vegetation on

macrozoobenthos in a coastal lagoon of the Southwestern

Atlantic. International Review of Hydrobiology 92: 33–47.

Barbone, E. & A. Basset, 2010. Hydrological constraints to

macrobenthic fauna biodiversity in transitional waters

ecosystems. Rendiconti Lincei 21: 301–314.

Barbone, E., I. Rosati, S. Reizopoulou & A. Basset, 2012a.

Linking classification boundaries to sources of natural

variability in transitional waters: a case study of benthic

macroinvertebrates. Ecological Indicators 12: 105–122.

Barbone, E., I. Rosati, A. Caldararu, A. Basset, L. P. Georgescu

& C. Iticescu, 2012b. Abiotic filtering of benthic inverte-

brates guilds in the Margherita di Savoia salt pan, Italy.

Environmental Engineering and Management Journal 11:

717–722.

Basset, A., 2010. Aquatic science and the water frame work

directive: a still open challenge towards ecogovernance of

aquatic ecosystems. Aquatic Conservation: Marine and

Freshwater Ecosystems 20: 245–249.

Basset, A., L. Sabetta, A. Fonnesu, D. Mouillot, T. Do Chi, P.

Viaroli, G. Giordani, S. Reizopoulou, M. Abbiati & G.

C. Carrada, 2006. Typology in mediterranean transitional

waters: new challenges and perspectives. Aquatic Conser-

vation: Marine and Freshwater Ecosystems 16: 441–455.

Basset, A., N. Galuppo & L. Sabetta, 2007. Environmental

heterogeneity and benthic macroinvertebrate guilds in

Italian lagoons. Transitional Water Bulletin 1: 48–63.

Basset, A., L. Sabetta, F. Sangiorgio, M. Pinna, D. Migoni, F.

Fanizzi, E. Barbone, N. Galuppo, S. Fonda Umani, S.

Reizopoulou, A. Nicolaidou, C. Arvanitidis, S. Moncheva,

A. Trajanova, L. Georgescu & S. Beqiraj, 2008. Biodi-

versity conservation in Mediterranean and Black Sea

lagoons: a trait-oriented approach to benthic invertebrate

guilds. Aquatic Conservation: Marine and Freshwater

Ecosystems 18: 4–15.

Basset, A., E. Barbone, M. Elliott, B. Li, S.E. Jorgensen, P.

Lucena-Moya, I. Pardo & D. Mouillot, 2012a. A unifying

approach to understanding transitional waters: fundamental

properties emerging from ecotone ecosystems. Estuarine,

Coastal and Shelf Science. http://dx.doi.org/10.1016/j.ecss.

2012.04.012.

Basset, A., E. Barbone, A. Borja, S. Brucet, M. Pinna, X.

D. Quintana, S. Reizopoulou, I. Rosati & N. Simboura,

2012b. A benthic macroinvertebrate size spectra index for

Hydrobiologia

123

Author's personal copy

implementing the water framework directive in coastal

lagoons in Mediterranean and Black Sea. Ecological

Indicators 12: 72–83.

Battaglia, B., 1959. Final resolution of the symposium on the

classification of brackish waters. Archo Oceanography

Limnology 11: 243–248.

Birk, S. & D. Hering, 2009. A new procedure for comparing

class boundaries of biological assessment methods: a case

study from the Danube Basin. Ecological Indicators 9:

528–539.

Birk, S., W. Bonne, A. Borja, A. Courrat, S. Poikane, A. Soli-

mini, W. van de Bund, N. Zampoukas & D. Hering, 2012.

Three hundred ways to assess Europe’s surface waters: an

almost complete overview of biological methods to

implement the Water Framework Directive. Ecological

Indicators 18: 31–41.

Blanchet, H., N. Lavesque, T. Ruellet, J. C. Dauvin, P. G. Sau-

riau, N. Desroy, C. Desclaux, M. Leconte, G. Bachelet, A.

L. Janson, C. Bessineton, S. Duhamel, J. Jourde, S. Mayot,

S. Simon & X. de Montaudouin, 2008. Use of biotic indices

in semi-enclosed coastal ecosystems and transitional

waters habitats—implications for the implementation of

the European Water Framework Directive. Ecological

Indicators 8: 360–372.

Boix, D., S. Gascon, J. Sala, M. Martinoy, J. Gifre & X.

D. Quintana, 2005. A new index of water quality assess-

ment in Mediterranean wetlands based on crustacean and

insect assemblages: the case of Catalunya (NE Iberian

peninsula). Aquatic Conservation: Marine and Freshwater

Ecosystems 15: 635–651.

Borja, A. & B. G. Tunberg, 2011. Assessing benthic health in

stressed subtropical estuaries, eastern Florida, USA using

AMBI and M-AMBI. Ecological Indicators 11: 295–303.

Borja, A., J. Franco, V. Valencia, J. Bald, I. Muxika, M.

J. Belzunce & O. Solaun, 2004. Implementation of the

European water framework directive from the Basque

Country (northern Spain): a methodological approach.

Marine Pollution Bulletin 48: 209–218.

Borja, A., A. B. Josefson, A. Miles, I. Muxika, F. Olsgard, G.

Phillips, J. G. Rodrıguez & B. Rygg, 2007. An approach to

the intercalibration of benthic ecological status assessment

in the North Atlantic ecoregion, according to the European

Water Framework Directive. Marine Pollution Bulletin 55:

42–52.

Borja, A., I. Muxika & J. G. Rodrıguez, 2009a. Paradigmatic

responses of marine benthic communities to different

anthropogenic pressures, using M-AMBI, within the

European Water Framework Directive. Marine Ecology

30: 214–227.

Borja, A., A. Miles, A. Occhipinti-Ambrogi & T. Berg, 2009b.

Current status of macroinvertebrate methods used for

assessing the quality of European marine waters: imple-

menting the Water Framework Directive. Hydrobiologia

633: 181–196.

Borja, A., D. M. Dauer & A. Gremare, 2012a. The importance of

setting targets and reference conditions in assessing marine

ecosystem quality. Ecological Indicators 12: 1–7.

Borja, A., M. Elliott, P. Henriksen & N. Marba, 2012b. Tran-

sitional and coastal waters ecological status assessment:

Advances and challenges resulting from implementing the

European Water Framework Directive. Hydrobiologia.

doi:10.1007/s10750-012-1276-9.

Comin, F. A., M. Menendez & J. A. Herrera, 2004. Spatial and

temporal scales for monitoring coastal aquatic ecosystem.

Aquatic Conservation: Marine and Freshwater Ecosystems

14: 5–17.

Dauvin, J. C., 2007. Paradox of estuarine quality: benthic

indicators and indices, consensus or debate for the future.

Marine Pollution Bulletin 55: 271–281.

Dauvin, J. C. & T. Ruellet, 2009. The estuarine quality paradox:

is it possible to define an ecological quality status for

specific modified and naturally stressed estuarine ecosys-

tems? Marine Pollution Bulletin 59: 38–47.

De Casabianca, M. L. & F. Posada, 1998. Effect of environ-

mental parameters on the growth of Ulva rigida (Thau

Lagoon, France). Botanica Marina 41: 157–165.

Elliott, M. & V. Quintino, 2007. The estuarine quality paradox,

environmental homeostasis and the difficulty of detecting

anthropogenic stress in naturally stressed areas. Marine

Pollution Bulletin 54: 640–645.

Escavarage, V., T. Ysebaert & P. Herman, 2004. Description of

the maximal and good ecological potentials (MEP/GEP)

for the benthic macrofauna for the European Water

Framework Directive (WFD), the Westerschelde. Report

of the Koninklijke Nederlandse Academie van Wetensc-

happen Nederland Instituut voor Ecologie, Yerseke, The

Netherlands.

Galuppo, N., S. Maci, M. Pinna & A. Basset, 2007. Habitat types

and distribution of benthic macroinvertebrates in a transi-

tional water ecosistema: alimini Grande (Puglia, Italy).

Transitional Waters Bulletin 1: 9–19.

Gamito, S., J. Patrıcio, J. M. Neto, J. C. Marques & H. Teixeira,

2012. The importance of habitat-type for defining the ref-

erence conditions and the ecological quality status based on

benthic invertebrates: the Ria Formosa coastal lagoon

(Southern Portugal) case study. Ecological Indicators 19:

61–72.

Guelorget, O. & J. P. Perthuisot, 1983. Le domaine paralique.

Travaux du Laboratoire de Geologie. Presses de l’Ecole

Normale Superieure, Paris.

Guilhaumon, F., A. Basset, E. Barbone & D. Mouillot, 2012.

Species–area relationships as a tool for the conservation of

benthic invertebrates in Italian coastal lagoons. Estuarine,

Coastal and Shelf Science. doi:http://dx.doi.org/10.1016/

j.ecss.2011.12.001.

ISPRA, 2007. Guida alla tipizzazione dei corpi idrici di tran-

sizione ed alla definizione delle condizioni di riferimento ai

sensi della direttiva 2000/60/CE.

ISPRA, 2010. Classificazione dello stato ecologico dei corpi

idrici delle acque marino costiere e di transizione. Imple-

mentazione della Direttiva 2000/60/CE.

Kafanov, A. I. & S. P. Plekhov, 2001. Bottom communities of

Semyachik Lagoon (Kronotskii biosphere reserve, Eastern

Kamchatka). Russian Journal of Marine Biology 27:

98–104.

Kjerfve, B. & K. E. Magill, 1989. Geographic and hydrody-

namic characteristics of shallow coastal lagoons. Marine

Geology 88: 187–199.

Lucena-Moya, P., I. Pardo & M. Alvarez, 2009. Development of

a typology for transitional waters in the Mediterranean

Hydrobiologia

123

Author's personal copy

ecoregion: the case of the islands. Estuarine, Coastal and

Shelf Science 82: 61–72.

Marques, J. C., F. Salas, J. Patrıcio, H. Teixeira & J. M. Neto,

2009. Ecological indicators for coastal and estuarine

environmental assessment. A user guide. WIT Press, UK.

ISBN: 978-1-84564-209-9.

Mistri, M. & C. Munari, 2008. BITS: a SMART indicator for

soft-bottom, non-tidal lagoons. Marine Pollution Bulletin

56: 587–599.

Munari, C., E. Manini, A. Pusceddu, R. Danovaro & M. Mistri,

2009. Response of BITS (a benthic index based on taxo-

nomic sufficiency) to water and sedimentary variables and

comparison with other indices in three Adriatic lagoons.

Marine Ecology 30: 255–268.

Muxika, I., A. Borja & J. Bald, 2007. Using historical data,

expert judgement and multivariate analysis in assessing

reference conditions and benthic ecological status,

according to the European Water Framework Directive.

Marine Pollution Bulletin 55: 16–29.

Neto, J. M., H. Teixeira, J. Patrıcio, A. Baeta, H. Verıssimo, R.

Pinto & J. C. Marques, 2010. The response of estuarine

macrobenthic communities to natural- and human-induced

changes: dynamics and ecological quality. Estuaries and

Coasts 33: 1327–1339.

Pinheiro, J. C. & D. M. Bates, 2000. Mixed effects models in S

and S-PLUS. Springer, New York.

Pinheiro, J., D. M. Bates, S. DebRoy, D. Sarkar, and the R Core

team, 2006. nlme: An R package for fitting and comparing

Gaussian linear and nonlinear mixed-effects models.

Available at http://www.stats.bris.ac.uk/R/.

Pinto, R., J. Patricio, A. Baeta, B. D. Fath, J. M. Neto & J.

C. Marques, 2009. Review and evaluation of estuarine

biotic indices to assess benthic condition. Ecological

Indicators 9: 1–25.

Prato, S., J. G. Morgana, P. La Valle, M. G. Finoia, L. Lattanzi,

L. Nicoletti, G. D. Ardizzone & G. Izzo, 2009. Application

of biotic and taxonomic distinctness indices in assessing

the ecological quality status of two coastal lakes: Caprolace

and Fogliano lakes (Central Italy). Ecological Indicators 9:

568–583.

R Development Core Team, 2006. R: A language and envi-

ronment for statistical computing; R foundation for statis-

tical computing: Vienna, Austria. ISBN 3-900051-07-0.

http://www.R-project.org.

Reizopoulou, S. & A. Nicolaidou, 2004. Benthic diversity of

coastal brackish-water lagoons in western Greece. Aquatic

Conservation: Marine and Freshwater Ecosystems 14:

93–102.

Reynoldson, T. B., R. C. Bailey, K. E. Day & R. H. Norris, 1995.

Biological guidelines for freshwater sediment based on

benthic assessment of sediment (BEAST) using a multi-

variate approach for predicting biological state. Australian

Journal of Ecology 20: 198–219.

Ruellet, T. & J. C. Dauvin, 2007. Benthic indicators: analysis of

the threshold values of ecological quality classifications for

transitional waters. Marine Pollution Bulletin 54:

1707–1714.

Sabetta, L., E. Barbone, A. Giardino, N. Galuppo & A. Basset,

2007. Species–area patterns of benthic macroinvertebrates

in Italian lagoons. Hydrobiologia 577: 127–139.

Stoddard, J. L., D. P. Larsen, C. P. Hawkins, R. K. Johnson & R.

H. Norris, 2006. Setting expectations for the ecological

condition of streams: the concept of reference condition.

Ecological Application 16: 1267–1276.

Teixeira, H., J. M. Neto, J. Patrıcio, H. Verıssimo, R. Pinto, F.

Salas & J. C. Marques, 2009. Quality assessment of benthic

macroinvertebrates under the scope of WFD using BAT,

the benthic assessment tool. Marine Pollution Bulletin 58:

1477–1486.

Teske, P. R. & T. H. Wooldridge, 2003. What limits the dis-

tribution of subtidal macrobenthos in permanently open

and temporarily open/closed South African estuaries?

Salinity vs. sediment particle size. Estuarine, Coastal and

Shelf Science 57: 225–238.

Vignes, F., E. Barbone, P. Breber, R. D’Adamo, L. Roselli, N.

Ungaro, S. Focardi, M. Renzi & A. Basset, 2009. Spatial

and temporal description of the dystrophic crisis in Lesina

lagoon during summer 2008. Transitional Water Bulletin 3:

114–119.

Vincent, C., H. Heinrich, A. Edwards, K. Nygaard & J. Hay-

thornthwaite, 2002. Guidance on Typology, Reference

Conditions and Classification Systems for Transitional and

Coastal Waters Produced by: CIS Working Group 2.4

(Coast), Common Implementation Strategy of the Water

Framework Directive, European Commission.

Whitfield, A. K., M. Elliott, A. Basset, S. J. M. Blaber & R.

J. West, 2012. Paradigms in estuarine ecology—the Re-

mane diagram revisited. Estuarine, Coastal and Shelf Sci-

ence 97: 78–90.

Wright, J. F., M. T. Furse & P. D. Armitage, 1993. RIVPACS: a

technique for evaluating the biological quality of rivers in

the UK. European Water Pollution Control 3: 15–25.

Zaldivar, J. M., A. C. Cardoso, P. Viaroli, A. Newton, R. De Wit,

C. Ibanez, S. Reizopoulou, F. Somma, A. Razinkovas, A.

Basset, M. Holmer & N. Murray, 2008. Eutrophication in

transitional waters: an overview. Transitional Water

Monograph 1: 1–78.

Zuur, A. F., E. N. Ieno, N. J. Walker, A. A. Saveliev & G.

M. Smith, 2009. Mixed effects models and extensions in

ecology with R. Springer, New York.

Hydrobiologia

123

Author's personal copy