Ecological Indicators 5 (2005) 19–31

This article is also available online at:

www.elsevier.com/locate/ecolind

The suitability of the marine biotic index (AMBI) to

new impact sources along European coasts
I. Muxikaa, Á. Borjaa,*, W. Bonnea,b
a
AZTI Foundation, Marine Research Division, Herrera Kaia, Portualdea s/n, 20110 Pasaia, Spain
b
Marine Biology Section, Biology Department, Ghent University, Krijgslaan 281/S8, 9000 Ghent, Belgium
Accepted 19 August 2004

Abstract

In recent years, several benthic biotic indices have been proposed to be used as ecological indicators in estuarine and coastal
waters. One such indicator, the AZTI Marine Biotic Index (AMBI), was designed to establish the ecological quality of European
coasts. The index examined the response of soft-bottom benthic communities to natural and man-induced disturbances in coastal
and estuarine environments. It has been successfully applied to different geographical areas and under different impact sources,
with increasing user numbers in European marine waters (Baltic, North Sea, Atlantic and Mediterranean). The AMBI has been
used also for the determination of the ecological quality status (EcoQ) within the context of the European Water Framework
Directive (WFD).
In this contribution, 38 different applications including six new case studies (hypoxia processes, sand extraction, oil platform
impacts, engineering works, dredging and fish aquaculture) are presented. The results show the response of the benthic
communities to different disturbance sources in a simple way. Those communities act as ecological indicators of the ‘health’ of
the system, indicating clearly the gradient associated with the disturbance.
# 2004 Elsevier Ltd. All rights reserved.

Keywords: AMBI; Biotic indices; Environmental quality assessment; Coastal waters; Estuaries; Macrozoobenthos; Soft-bottom; Anoxia; Sand
extraction; Hydrocarbon pollution; Civil works; Aquaculture

1. Introduction coastal waters (Hily, 1984; Washington, 1984; Rygg,

1985; Majeed, 1987; Codling and Ashley, 1992;
Recently, several benthic biotic indices have been Dauer, 1993; Engle et al., 1994; Grall and Glémarec,
proposed as ecological indicators in estuarine and 1997; Weisberg et al., 1997; Roberts et al., 1998; Van
dolah et al., 1999; Smith and Rule, 2001; and Eaton,
2001), to determine natural and man-induced impacts.
* Corresponding author. Tel.: +34 943 004800;
fax: +34 943 004801.
One such indicator, the AZTI Marine Biotic Index
E-mail addresses: imuxika@pas.azti.es (I. Muxika), (AMBI), which was developed by Borja et al. (2000)
aborja@pas.azti.es (Á. Borja), wbonne@pas.azti.es (W. Bonne). has been applied successfully to different geographi-

1470-160X/$ – see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ecolind.2004.08.004
20 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31

cal areas and under different impact sources (Borja site, representing the benthic community ‘health’
et al., 2003a,b, 2004a), with increasing user numbers (sensu Grall and Glémarec, 1997). Secondarily, it has
within Europe (Table 1). been used for the determination of the ecological
The AMBI was designed primarily to establish the quality status (EcoQ) within the context of the
ecological quality of European coastal and estuarine European Water Framework Directive (WFD) (Borja
waters by examining the response of soft-bottom et al., 2003b, 2004a,b). The ultimate aim of the WFD
benthic communities to natural and man-induced is to achieve, by 2015, a good EcoQ within all the
disturbances in the environment. Hence, the AMBI European waters, by the elimination of priority
offers a ‘disturbance or pollution classification’ of a hazardous substances, and contribute to achieving

Table 1
Different impact sources and geographical areas for which AMBI has been applied, in recent years
Impact sources Locations (countries) Seas Author
Various sources along UK (United Kingdom) Atlantic A. Miles, A. Prior (p.c., 2003)
Outfall and harbour Brittany (France) Atlantic Borja et al. (2003a)
Engineering works (dyke) Basque Country (Spain) Atlantic Borja et al. (2000, 2003a)
Sewerage works Basque Country (Spain) Atlantic Borja et al. (2000, 2003a)
Harbour construction Basque Country (Spain) Atlantic This contribution
Submarine outfall Basque Country (Spain) Atlantic Borja et al. (2000, 2003b)
Harbour and river inputs Basque Country (Spain) Atlantic Muxika et al. (2003)
Various sources Tejo estuary (Portugal) Atlantic M.J. Gaudencio (p.c., 2003)
Eutrophy Mondego estuary (Portugal) Atlantic Salas et al. (2004)
River inputs Guadalquivir (Spain) Atlantic AZTI (unpublished data)
Heavy metals Huelva (Spain) Atlantic Borja et al. (2003a)
Estuarine inputs Cádiz (Spain) Atlantic A. Rodrı́guez-Martı́n (p.c., 2003)
Various sources (Morocco) Atlantic H. Bazairi (p.c., 2003)
Various sources Latvia Baltic V. Jermakovs (p.c., 2004)
Anoxia-hypoxia Sweden Baltic This contribution
Dredging mud disposal Sweden Baltic S. Smith (p.c., 2003)
Various sources along Sweden Sweden Baltic M. Blomqvist (p.c., 2003)
Various sources in a lagoon Smir (Morocco) Mediterranean A. Chaouti (p.c., 2003)
Dredging in harbour Ceuta (Spain) Mediterranean This contribution
Diffuse pollution (mines, agriculture, . . .) Almerı́a and Murcia (Spain) Mediterranean Borja et al. (2003a)
Aquaculture cages Murcia, Valencia (Spain) Mediterranean AZTI (unpublished data)
Mining debris Mar Menor (Spain) Mediterranean L. Marı́n (p.c., 2004)
Submarine outfall Catalonia (Spain) Mediterranean M.J. Cardell (p.c., 2003)
Marina Catalonia (Spain) Mediterranean S. Pinedo (p.c., 2003)
Wastewater discharge in a lagoon (France) Mediterranean G. Reimonenq (p.c., 2003)
Inputs to a coastal lagoon Adriatic Sea (Italy) Mediterranean Casselli et al. (2003)
Various sources Adriatic Sea (Italy) Mediterranean Forni and Occhipinti Ambrogi (2003)
Stagnation and industrial and urban pollution Port of Trieste (Italy) Mediterranean Solı́s-Weiss et al. (2004)
Submarine outfall Gulf of Trieste (Italy) Mediterranean Solı́s-Weiss (p.c., 2004)
Various sources Adriatic Sea (Italy) Mediterranean R. Simonini (p.c., 2004)
Submarine outfall Saronikos Gulf (Greece) Mediterranean Borja et al. (2003a)
Aquaculture cages Three locations (Greece) Mediterranean This contribution
River inputs Thames (United Kingdom) North Sea M. Davison (p.c., 2002)
Oil-based drilling muds (oil platforms) Eleven locations (United Kingdom) North Sea This contribution
Impacts on sandy shores (Netherlands) North Sea S. Mulder (p.c., 2003)
Ester-based drilling muds (oil platforms) North Sea (Netherlands) North Sea Borja et al. (2003a)
Re-opening of a brackish lake to sea influence Veerse Meer (Netherlands) North Sea V. Escaravage (p.c., 2004)
Sand extraction Belgium North Sea Bonne et al. (2003); this contribution
Key: p.c., personal communication.
 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31 21

Table 2
Summary of the AMBI values and their equivalences (modified from Borja et al., 2000)
Biotic coefficient Dominating ecological group Benthic community health Site disturbance classification Ecological status
0.0 < AMBI  0.2 I Normal Undisturbed High status
0.2 < AMBI  1.2 Impoverished
1.2 < AMBI  3.3 III Unbalanced Slightly disturbed Good status
3.3 < AMBI  4.3 Transitional to pollution Moderately disturbed Moderate status
4.3 < AMBI  5.0 IV–V Polluted Poor status
5.0 < AMBI  5.5 Transitional to heavy pollution Heavily disturbed
5.5 < AMBI  6.0 V Heavy polluted Bad status
6.0 < AMBI  7.0 Azoic Azoic Extremely disturbed
The last column shows the proposed equivalent ecological status for the application of the WFD (Borja et al., 2003b).

concentrations in the marine environment near back- the various ecological groups (see Fig. 2, in Borja
ground values for naturally occurring substances. et al., 2000). These thresholds (Table 2) are coincident
EcoQ is established on the basis of physico-chemical with the benthic community health proposed by Grall
and biological variables (Borja et al., 2004a). In and Glémarec (1997) (based upon Reish, 1959,
coastal and estuarine waters, benthic community Bellan, 1967 and Pearson and Rosenberg, 1976).
measures are especially important, as they integrate The AMBI is expected to be used to calculate the
impacts over a wide period of time. In the benthic EcoQ, but it would be only a part of a set of measures
EcoQ determination, three parameters are proposed by and indices in the WFD, such as diversity, richness,
the WFD: diversity, species abundance and the etc. The thresholds used for site pollution classifica-
presence/absence of indicator and stress-sensitive tion are not the same as the thresholds proposed for the
species. The latter is represented by the AMBI (Borja EcoQ, to accomplish WFD specifications (Borja et al.,
et al., 2003b, 2004a,b). 2003b, 2004b) (Table 2).
The AMBI is based upon ecological models, such The increasing use of this tool requires integration
as those of Glémarec and Hily (1981) and Hily (1984) of different applications and results obtained by
(for the development of the model, see Borja et al., various authors (see, also, Table 1). Hence, the main
2000; for additional details, see Borja et al., 2004b). objectives of this contribution are to explore: (i) the
The most novel contribution of AMBI has been a suitability of the AMBI to its use in the Atlantic,
formula (1) to allow the derivation of a series of Baltic, North Sea and Mediterranean European coasts;
continuous values (firstly called the ‘Biotic Coeffi- and (ii) its usefulness in relation to different new
cient’, in Borja et al. (2000), and then AMBI, as impact sources, not explored previously, i.e. sand
outlined above), based upon the proportions of five extraction, hypoxia processes, oil platform impacts,
ecological groups (EG) to which the benthic species dredging and fish aquaculture.
are allocated:
AMBI ¼ ½ð0  %EG IÞ þ ð1:5  %EG IIÞ 2. Methods
þ ð3  %EG IIIÞ þ ð4:5  %EG IVÞ
þ ð6  %EG VÞ=100 (1) Six different ‘case studies’ along the European
coast are compared. These case studies are based on
with EG I being the disturbance-sensitive species, EG data from the authors and other published sources
II the disturbance-indifferent species, EG III the dis- representing a variety of pollution types, environ-
turbance-tolerant species, EG IV the second-order mental problems and geographic settings.
opportunistic species and EG V the first-order oppor- Case study 1 is located in the Gullmarsfjord
tunistic species (see Borja et al., 2000). (Swedish west coast; see Fig. 1). This fjord has a
Several thresholds have been established over the maximum depth of 118 m. The bottom water is
scale of the AMBI, based upon proportions amongst usually renewed with oxygen-rich water, each spring.
22 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31

installed. Biological and physico-chemical data were

obtained from the Marine Environmental Surveys
Database on the UKCS-UK Benthos, provided by the
UK Offshore Operators Association (UKOOA). From
this database, the 11 piles sampled in 1988 were
selected for this contribution; five of these are situated
in the northern North Sea (Beryl A, Beryl B, Buchan,
Miller and Thistle), whilst six are located in the
southern/central North Sea (Audrey, Barque, Cleeton,
Cilpper, Ravenspurn and Sole). Moreover, comple-
mentary biological and chemical information was
obtained from Davies et al. (1984), Shimmield et al.
(2000), and Breuer et al. (1999, 2004).
Case study 4 is situated on the Spanish Basque
coast (Fig. 1). For the construction of a new dyke in
Fig. 1. Location of the six case studies, within the context of an
Bilbao harbour, a large area of the seabed was dred-
European framework. Note that there are two sites for Case 3
(northern and central North Sea) and two sites for Case 6 (western ged. These works commenced in 1993. An intensive
and eastern Greece). dredging period extended from 1995 to 1997,
changing the bathymetry of the area from 20 to
In spring 1997, this water renewal did not occur and 30 m, in some places. As a result, all of the benthic
the fauna were: more or less unaffected at 60 and 75 m fauna disappeared (V. Valencia, AZTI, personal
water depth (oxygen saturation > 15%); significantly communication, 2003). The works finished in 1999.
reduced at 85 and 95 m (saturation < 10%); and The zone was monitored before the construction (from
eliminated below about 100 m water depth (Rosen- 1989 to 1993) and three surveys were carried out (in
berg et al., 2002). In spring 1998, the fjord was re- 2000, 2001 and 2003) to monitor the impact of the
oxygenated and the succession of benthic fauna was new structure.
studied at five sampling stations and over a 2-year The immediate recolonisation following a harbour
period by Rosenberg et al. (2002). dredging in 1999 (Case study 5) was studied in the
The Kwintebank (Case study 2) is an intensively Spanish Mediterranean (Ceuta, North Africa; see
exploited sandbank, located in the southern North Sea Fig. 1). The impact on the benthic communities was
(Belgian coast; see Fig. 1). Three stations (1, 6 and 9) studied using a before–after control impacted (BACI)
were sampled for macrobenthos during three periods approach, by Guerra-Garcı́a et al. (2003). These
(in the late 1970s, late 1990s and 2001, respectively), authors undertook six samplings, at dredged and
within the framework of different projects (Bonne control locations: before dredging and after 3, 15, 30,
et al., 2003). These data were used to assess the impact 90 and 180 days. Community structure data and MDS
of sand extraction on subtidal sandbanks. In the late ordination were used to analyse the results.
1970s about 310,000 m3 yr 1 sand was extracted on The impact of cage culturing of fish on benthic
the Kwintebank, whereas the extracted volume amo- communities was investigated at three commercial
unted to about 1,360,000 m3 yr 1 in the late 1990s, fish farms with different types of sediment (ranging
and 1,700,000 m3 in 2001 (Fund for Sand Extraction, from silt to coarse sand) at 20–30 m water depth in
FPS Economy, S.M.E.s, Self-employed and Energy, Cephalonia, Ithaki and Sounion (Greece; Fig. 1), by
Brussels, Belgium). The sand extraction intensity, in Karakassis et al. (2000) (Case study 6). A transect of
the late 1990s, was very high at Stations 1 and 6 stations and a control station were sampled near each
(56,000 and 92,000 m3 0.5 km 2 yr 1, respectively); farm, for macrofauna and geochemical variables in
and low at Station 9 (4,000 m3 0.5 km 2 yr 1) (Bonne July and October 1995 and April 1996.
and Vincx, 2003). Most of the selected studies have, in common, the
Case study 3 is located in the central and northern aim to explain the effect of different impact sources on
North Sea (Fig. 1), where numerous oilfields are soft-bottom communities; this is based upon the study
 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31 23

of structural parameters, such as abundance, biomass,

richness, diversity, evenness, abundance-biomass
comparison (ABC) curves (Warwick, 1986; Clarke,
1990), or multivariate methods. The latter include:
clustering (Sokal and Sneath, 1963); multi-dimen-
sional scaling (MDS) (Kruskal and Wish, 1977;
Schiffman et al., 1981; Field et al., 1982); principal
component analysis (PCA) (Kendall, 1975; Jolliffe,
1986); or correspondence analysis (CA) (Hill, 1974;
Fielding, 1992). Based upon the abundance of
individuals, as provided by the authors in the Fig. 2. AMBI values in 1998, 1999 and 2000, for each sampling
above-mentioned papers, the corresponding Biotic station in Gullmarsfjiord (note that the label for each sampling
Coefficient (AMBI), sensu Borja et al. (2000), was station coincides with the water depth). Key: UD = undisturbed;
SD = slightly disturbed; MD = moderately disturbed; HD = heavily
calculated using a freeware program available on disturbed; and ED = extremely disturbed.
www.azti.es, which includes the EG of more than
2,700 taxa, updated continuously. Whenever the
species composition per replicate was available, the the AMBI values calculated with those data (based
AMBI was calculated for each of the replicates, then throughout upon more than 80% of the total
averaged for the entire station, as recommended by abundance, for each station) show a clear increasing
Borja et al. (2004b). These values have been used to gradient with water depth, during 1998 (Fig. 2). The
illustrate, in a simple format: (i) spatial disturbance or sampling stations at 75 and 85 m water depths were
pollution gradients; (ii) the evolution of the effect of classified as slightly disturbed (AMBI values < 3.3)
disturbance or pollution on the communities; (iii) and and dominated by the EG III (47–91% of total
the sensitivity of AMBI to different impact sources. abundance), which represents the tolerant species. The
The assessment was undertaken according to the station at 95 m water depth was classified as
classification listed in Table 2. moderately disturbed (AMBI values near 4.5), with
In this contribution, the term ‘disturbed’ has been the EG III and V (first-order opportunistic) species
used with the same meaning as ‘polluted’ (which was being co-dominant, at 54 and 46% of the total
used in the original contribution of Borja et al., 2000). abundance, respectively. The 105 and 118 m water
Hence, ‘unpolluted’, ‘slightly polluted’, etc., are depth stations were classified as heavily disturbed
presented here as ‘undisturbed’, ‘slightly disturbed’, (AMBI values near 6), with the EG V species
etc. The use of ‘disturbed’ is recommended when the representing more than 95% of the community
impact source is natural (e.g. the inner part of an abundance.
estuary, with high levels of natural stress and unlikely In 1999 and 2000, a clear recovery was detected by
to be classified as undisturbed), mechanical (e.g. high the AMBI at the 95, 105 and 118 m water depth
exposure or dynamics, with changing characteristics stations; these were classified as slightly disturbed or
in the substrata) or physical (e.g. dredging or undisturbed. The EG III became dominant and the EG
engineering works). For comparison, the use of V disappeared within these communities. At the 75
‘polluted’ is recommended when the impact source and 85 m water depth stations a recovery was also
is chemical. detected, but was not so important (Fig. 2). The EG I
(sensitive) species became more important as a
consequence of a decrease in the percentage of EG
3. Results IV (second-order opportunistic) species.

3.1. Case study 1 3.2. Case study 2

Although the mean abundance was only available The three stations are classified as undisturbed or
for the dominant species in Rosenberg et al. (2002), slightly disturbed, during all the sampling periods
24 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31

differences were found among the stations (d.f. = 2;

F = 3.31; p = 0.050), whereas the AMBI was sig-
nificantly lower during the periods of high sand
extraction intensity (d.f. = 1; F = 6.58; p = 0.013).

3.3. Case study 3

AMBI values show a clear decreasing gradient, as

Fig. 3. AMBI values, with standard error, for Stations 1, 6 and 9 on one moves away from the stations located near the
the Kwintebank during different campaigns (yy/mm on x-axis). platform wells in all the studied cases (Fig. 4) and in
the prevailing current direction: (i) from 0 to 100 m,
AMBI values lie between 5 and 6 (heavily polluted)
(Fig. 3). A two-way ANOVA was carried out to and the benthic community is dominated by first-order
compare the AMBI values between periods of low opportunistic species; (ii) from 100 to 500 m, the
(late 1970s and 1980s) and high sand extraction AMBI values lie between 3.3 and 5 (moderately
intensity (late 1990s and 2001s) and between stations polluted), except at Audrey (Fig. 4a), with increasing
(with Stations 1 and 6 being heavily exploited and dominance of EG IV and III, and the presence of I and
Station 9 only sporadically exploited). No significant II; and (iii) from 500 to 1,000 m, the AMBI values are

Fig. 4. AMBI values for each sampling station at the Audrey (a), Beryl A (b), Beryl B (c and d) and Thistle (e and f) oil platforms. The stations
are labelled with the distance along the transect from the platform, which is located by an arrow. The transect orientations are: (a) and (e) 1358,
(b) 1688, (c) 1808, (d) 908 and (f) 458, relative to North. Key: UD = undisturbed; SD = slightly disturbed; MD = moderately disturbed;
HD = heavily disturbed; and ED = extremely disturbed.
 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31 25

normally <1.2 (unpolluted), with EG I and II

dominating. This gradient depends upon the regional
prevailing current direction. Hence, the regression
between the distance (from 0 to 1,200 m) and AMBI
relating to stations in the prevailing current direction
is: AMBI = 0.004  distance + 5.354, with the
correlation being strong and highly significant
(F = 168.31; p = 0.000; r = 0.928). At greater dis-
tances, there was only a weak correlation (F = 0.389;
p = 0.000; r = 0.150).
Conversely, a strong and highly significant correla-
tion was found between the total hydrocarbons in the
sediment and the AMBI values, when data were
available (Beryl A and Beryl B), following a
logarithmic model (F = 157.02; p = 0.000;
r = 0.914) (Fig. 5a). Hence, at the furthest stations,
sensitive species are dominant in all cases. Approach-
ing the oil platforms, they are progressively sub-
stituted by indifferent, tolerant and second- and first-
order opportunistic species. These changes are related
to the high hydrocarbon values in the sediment.
Likewise, correlations between grain size and AMBI
values, together with those between organic matter
and AMBI, were only moderate (p = 0.000; r < 0.50). Fig. 5. Regressions between (a) AMBI values and total hydrocar-
bons (THC) measured in mg kg 1, (b) percentage of organic matter
content in sediment and (c) mean grain diameter, in F units.
3.4. Case study 4

Before the engineering works commenced (1989– (90%). Only after 180 days had the proportion
1993) in Bilbao harbour, the AMBI values were very returned to the situation before dredging (Fig. 7b).
similar (except in 1990, with an AMBI = 2.1), with the Likewise, in the control location there was a decrease
area being classified as undisturbed or slightly in the sensitive species abundance (EG I), after 3 days
disturbed and the EG I being dominant (Fig. 6). (Fig. 7a). The most important change occurred after 15
When intensive dredging finished in 1997, the area days, when the second-order opportunistic (EG IV)
became totally azoic (AMBI = 7) as a result of the species increased in their proportion to 75%,
elimination of a surficial sediment layer of about 10 m. decreasing subsequently (Fig. 7a).
Over recent years, the area has improved in terms of its
classification, but with AMBI values higher than
previously, i.e. nearer to 3 (except in 2001,
AMBI = 2.3); these represent a moderately or slightly
disturbed situation.

3.5. Case study 5

At the dredged location, the proportion of first-

order opportunistic species (EG V) increased within 3 Fig. 6. Relative abundance of each ecological group for each
days after dredging (from 30 to 60%; Fig. 7b). replicate and average AMBI values, with standard error bars for
Between 15 and 90 days after dredging, the second- each of the sampling occasions: 89, 90, 91, . . . represent the
order opportunistic species (EG IV) largely dominated sampling year.
26 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31

Fig. 9. Regression lines between AMBI values and the distance

from the aquaculture cages, for each of the sites. Key: (—) Cepha-
lonia; (- - -) Ithaki; ( ) Sounion.

highly significant for the model (F = 10.38; d.f. = 5;

p = 0.002). No significant differences were detected
between the Cephalonia and Ithaki regression lines,
but significant differences were detected between
Fig. 7. Relative abundances of the five ecological groups (EG I–EG
V), during each sampling (key: B = before dredging, A3 = 3 days
Cephalonia and Sounion and between Ithaki and
after dredging; A15 = 15 days after dredging; etc.; C = control Sounion, both in terms of the intercept and the slope.
station; and D = dredged station) at the control station (a) and the Hence, a clear gradient is detected by the AMBI, in
dredged station (b). Cephalonia and Ithaki. The AMBI values decrease, as
the distance from the cages increases. In Sounion, the
AMBI is a little bit higher below the cages than at the
This pattern is reflected in the AMBI values control site, but the gradient is not very clear. Under
(Fig. 8): hence, the dredged location increased from the Cephalonia and Ithaki cages, benthic communities
4.05 to 5.07 (moderately disturbed to heavily dis- should be considered as heavily disturbed: being: (a)
turbed) after 3 days, decreasing to 4.54 after 15 days, moderately disturbed at a distance of between 5 and
then fluctuated around 4.3 (always moderately 10 m; (b) slightly disturbed at a distance of 25 m; and
disturbed). Likewise, the control location improved (c) undisturbed at 50 m.
within 3 days after dredging; it worsened after 15 days
(reaching 4.09 values), then improved.
4. Discussion
3.6. Case study 6
Even though the impact sources studied in this
A comparison of the derived regression lines was contribution are different, the effects can be grouped
carried out between the sites (Fig. 9). The AMBI was into three classes: (i) oxygen demand (Case 1:
used as dependent variable, with the distance from the dissolved oxygen depletion); (ii) physical disturbance
cages as independent variable. The regression was (Case 2: sand extraction; Case 4: engineering works;
and Case 5: dredging activities); and (iii) increasing
organic matter and associated pollutants (Case 3: drill
mud dumping; and Case 6: aquaculture).
In Case study 1, 1 year after the anoxic episode, the
deepest communities were still dominated by oppor-
tunistic species and were classified as heavily disturbed,
the station at 95 m water depth was moderately
disturbed and the shallowest stations were slightly
Fig. 8. Evolution of AMBI values throughout the BACI study, at the
disturbed. Two years after the anoxic episode, the
control station and the dredged station. Note: 0 day corresponds to benthic community had already recovered, with the
the sampling undertaken before dredging commenced. area being classified as slightly disturbed or undis-
 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31 27

turbed. Such a rapid improvement was probably Case study 4 is related also with physical
possible because the only impact was a punctual disturbance. Before the engineering works com-
anoxic episode, rather than a chronic impact. This menced, the area was classified as unpolluted or
conclusion agrees with that of Rosenberg et al. (2002), slightly polluted. The pollutants carried by the
i.e. following multivariate analysis, that the benthic Nervión river did not cause any detectable effect,
communities at all depths more or less returned to the due to dilution (Gorostiaga et al., 2004). Following the
same faunal composition as during pre-disturbed dredging period, a recovery was expected in the
conditions, with such a return being slowest at the AMBI, from 7 to values close to those before the
deepest stations. These researchers stated that the works. No data were available from 1997 to 1999, but
benthic fauna succession in this fjord followed the from 2000 to 2003 AMBI values lie between 2 and 3.5;
Pearson–Rosenberg successional model, on which the this indicates partial recovery in the macrozoobenthic
AMBI is based (Borja et al., 2000, 2004b). community. The monitoring should be continued, to
There are no major changes in the AMBI values in confirm this trend and ensure that the impact of the
Case study 2, where all the stations are classified as works, on the benthic community, was only transient.
undisturbed or slightly disturbed. Increasing sand However, it is not expected to reach those values
extraction intensity, from the late 1970s onwards, did before the works due to the enclosure, which: (i) slows
not result in an increase in the AMBI. Rather, it was water renewal; (ii) causes important pollutant reten-
lower during the period of high sand extraction tion; and (iii) increases organic matter levels, as
intensity, than in the period of low extraction. The observed in other harbours (Muxika et al., 2003).
lower AMBI results from a decrease in the EG III. The pattern of recovery following dredging in
Moreover, the AMBI did not differ between stations Ceuta, as detected by the AMBI, is more evident than
characterised by different sand extraction intensities. through the use of other tools (total abundance,
Therefore, in this case study, the AMBI is not a good diversity, evenness and Margalef’s index), as shown
indicator for detecting the impacts of sand extraction. by Guerra-Garcı́a et al. (2003). By means of AMBI, it
This conclusion is in agreement with that of Bonne is possible to deduce two different impact effects: (i)
et al. (2003), who did not detect an increased in the dredged area, there is an immediate effect due to
abundance of opportunistic species as a result of the the physical disturbance (detected within 3 days), with
extractions. a posterior gradual recolonisation (occuring in the
Elsewhere, it has been shown that the usefulness of same period as mentioned by Dernie et al. (2003a));
the AMBI is not only limited to organic pollution and (ii) an effect on the ‘near-control’ area (due
assessment, but also reflects disturbances, for exam- probably to the deposition of suspended materials,
ple, by hydrocarbons, engineering works and harbour after several days), as detected by the AMBI after 15
dredging (see Table 1). Conversely, the reworking of days. The general pattern is similar to that shown by:
‘organic-poor’ sediments of subtidal sandbanks (a) Sánchez-Moyano et al. (2004), in dredging works
(Vanosmael et al., 1979, 1982) does not appear to in the south of Spain; and (b) Case study 4, where the
favour typical opportunistic macrobenthic species. recovery is nearly total after 3 months.
The competitive ability of the species classified as Oil platforms can produce several environmental
opportunistic species in the AMBI is probably not impacts, in response to the platform itself and to the
advantageous in organically poor and naturally– discharge of drilling muds and cuttings (Frascari et al.,
physically stressed environments, such as offshore 1992): (i) physical impacts, such as the generation of
subtidal sandbanks or the inner part of estuaries (as turbulence, erosion, changes in grain size, etc.; and (ii)
outlined by Borja et al., 2004b). Pearson and biological, such as community changes and pollutant
Rosenberg (1978) pointed out that the use of any incorporation. For example, the amount of diesel oil
indicator schemes must be accompanied by a detailed discharged (associated with drill cuttings used in
knowledge of both the abundance and range of species drilling operations), in 1981, into UK continental shelf
in the area concerned, together with complementary waters, was estimated as 7,000 t (Davies et al., 1984).
community structure information (as outlined by Drilling chemicals discharged in the same area up to
Borja et al. (2004b)). 1989 were 39,902 t yr 1 (Breuer et al., 2004).
28 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31

In Case study 3, the highest AMBI values example, Capitella cf. capitata dominated the
(therefore, the highest disturbance) are reached near macrofauna over distances of up to 10 m at
the oil platforms, in all situations. Clear gradients are Cephalonia and Ithaki, whereas the dominant species
shown in all directions, but preferential currents are at Sounion was Protodorvillea kefersteini. This parti-
shown by the smoothest gradients (Fig. 4). The impact cular study concluded that ‘impacts of fish farming
of the oil platforms reached up to 500–1,000 m, as on benthos in the Mediterranean vary considerably
detected by AMBI (in the northern North Sea, some of depending on site characteristics’ (Karakassis et al.,
the stations located at 10,000 m from the platform are 2000). These conclusions, which are similar to those
still slightly disturbed). The same pattern has been using the AMBI, were obtained using ABC curves and
detected by Davies et al. (1984), for the same oilfields, MDS ordination plot approaches (Karakassis et al.,
using community structure parameters: likewise, by 2000).
Borja et al. (2003a,b) using AMBI on ester-based Benthic communities, affected by different impact
muds in the Dutch area of the North Sea. Davies et al. sources, react similarly; essentially, they change from
(1984) detected oil concentrations between 1,000 and sensitive groups (I and II) to lower successional stages
10,000 times the background levels, within 250 m of (the opportunistic IVand V groups). The percentage of
the platforms; this explains the high correlation each EG in the community depends upon: (i) the
obtained, in this contribution, between the AMBI intensity of the impact; (ii) the duration of the impact;
values and the total hydrocarbons. The pattern of and (iii) the distance from the source (as shown in this
distribution of the pollutants (together with its impact contribution). In the case of sand extraction, the
on the benthic community, detected by means of the community behaves in a different way, not detected by
AMBI) coincides with the axis of the most persistent the AMBI as disturbance in this contribution. As
current, often producing an ellipsoidal distribution mentioned by Rosenberg et al. (2002), benthic
(Davies et al., 1984). Further, Shimmield et al. (2000) communities show great resilience and elasticity.
found high disturbances in cores obtained from Hence, the resilience of these communities following
sediments retrieved at a distance of 65 m from the the cessation of the impact source is dependent upon
drilling cutting piles. Higher depletion of interstitial the species composition, their different life-cycles,
dissolved oxygen concentrations were found in these reproduction periods and larval dispersal patterns
cores, in comparison with those obtained at 165 m and (Rosenberg et al., 2002). However, the AMBI appears
300 m; likewise, higher heavy metal concentrations in to be independent of these controlling factors in most
the superficial layer and higher Ba concentrations. cases, because the case studies shown in this
The same distributional impact pattern is shown in contribution, together with others presented pre-
Case study 6, in which AMBI values decrease as the viously (Borja et al., 2000, 2003a; Gorostiaga et al.,
distance from the cages increases. All the stations are 2004; Salas et al., 2004), include different settings,
moderately or heavily polluted, up to 25 m from the species and time-scales. The proportion between the
edge of the cages (the same pattern has been detected different EGs appears to control the final result.
in fish-farms elsewhere (Mazzola et al., 2000). At Conversely, elasticity (rapid recovery) is promoted
Sounion, this gradient is not detected; this can be by the presence of undisturbed communities in the
explained by the high current velocities at this site, vicinity of a particular site, as demonstrated when local
which are 6.3 cm s 1 (compared with only 3.5 and impacts are produced (such as local hypoxia, dredging
2.8 cm s 1 at Cephalonia and Ithaki, respectively). works, dumping, etc.). On this basis, Dernie et al.
Hence, high current velocities could spread the (2003b) suggest that physical and biological recovery
pollutants (organic carbon and nitrogen content of rates are mediated by a combination of physical,
the sediment), avoiding a localised impact on the bed chemical and biological factors; these, in turn, differ in
near the farm. This interpretation agrees with the their relative importance in different habitats. Hence,
conclusions of Karakassis et al. (2000), except that the AMBI could integrate, simply and usefully, these
these authors detect also an impact on the macrofaunal different factors into an unique number; this incorpo-
community at Sounion. However, the faunal composi- rates an equilibrium between the five EGs, connecting
tion of Sounion differed from that at the other sites: for with the classical ecological theories on disturbance
 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31 29

models and recovery of impacted, or stressed, com- e.g. high hydrodynamic energy areas, subtidal
munities (Bellan, 1967; McArthur and Wilson, 1967; sandbanks, and the inner parts of the estuaries, etc.
Pianka, 1970; Pearson and Rosenberg, 1978; Gray,
1979). This feature explains how the AMBI is able to
respond successfully to very different environmental Acknowledgements
impact sources, including: drill cutting discharges;
submarine outfalls; harbour and dyke construction; One of the authors (I. Muxika) was supported by a
heavy metal inputs; eutrophication processes; diffuse grant from the Technological Centres Foundation of
pollutant inputs; recovery in polluted systems, under the the Basque Country. The macrobenthic data of the
impact of sewerage schemes; dredging processes; mud Kwintebank were gathered during different projects
disposal; and oil spills (see Table 1, together with the by the Marine Biology Section of the University of
associated references). That is why, combined with Gent; these were used for the AMBI within the Ph.D.
other metrics, it can also be a successful tool in thesis of Wendy Bonne, at the same laboratory. We
implementing the WFD (Borja et al., 2004a,b). would like to thank Prof. Magda Vincx and Dr. Steven
Degraer for providing the data for Case study 2 and
Annamaria Rekecki for identifying the additional
5. Conclusions samples (from 2001) for the same case study. The
benthic data, prior to the construction of the harbour in
This contribution has demonstrated the usefulness Case study 4, were provided by the ‘Consorcio de
of the AMBI in detecting different and new impact Aguas Bilbao-Bizkaia’; those following dredging by
sources and disturbance gradients. The results the Bilbao Port Authority. We wish to thank Professor
obtained with the AMBI are comparable with those Michael Collins (School of Ocean and Earth Science,
obtained using other methods and parameters (includ- University of Southampton, UK) and also Victoriano
ing univariate and multivariate statistical analyses); as Valencia (AZTI Foundation) for kindly advising us on
such, they are appropriate to what physico-chemical some details of this paper.
data show. The AMBI values provide a single and
clear way (useful in terms of environmental advice)
to establish the ecological quality of soft-bottom References
benthos, which is complementary to the above-
mentioned methods. Bellan, G., 1967. Pollution et peuplements benthiques sur substrat
meuble dans la région de Marseille. 1 Partie. Le secteur de
The AMBI is appropriate to use for all European Cortiu. Rev. Int. Océanogr. Med. VI–VII, 53–87.
coastal environments, as it is independent of longitude Bonne, W., Vincx, M., 2003. The impact of marine sand extraction
and latitude, because it is based upon general on benthic copepod communities of a subtidal sandbank. In:
ecological principles and paradigms. In this contribu- Benthic Copepod Communities in Relation to Natural and
Anthropogenic Influences in the North Sea, Chapter III. Ph.D.
tion, it has been used for the Atlantic Ocean region
Thesis, Ghent University, Biology Department, Marine Biology
(Baltic Sea, North Sea and Bay of Biscay), as well as Section, Belgium, pp. 165–202.
in the Mediterranean Sea (Gibraltar Strait and Bonne, W., Rekecki, A., Vincx, M., 2003. Impact assessment of sand
Greece). Results from 38 case studies, from all the extraction on subtidal sandbanks using macrobenthos. In:
European seas, have been collated. The AMBI reacts Benthic copepod communities in relation to natural and antho-
in the same way to different disturbance sources, e.g. pogenic influences in the North Sea, Chapter IV. Ph.D. Thesis,
Ghent University, Biology Department, Marine Biology Section,
anoxic episodes, hydrocarbon pollution, engineering Belgium, pp. 207–226.
works, dredging or fish farming cages. Borja, Á., Franco, J., Pérez, V., 2000. A marine biotic index to
The AMBI is useful to compare the ecological establish the ecological quality of soft-bottom benthos within
quality of the soft-bottom benthos in BACI experi- European estuarine and coastal environments. Mar. Poll. Bull.
ments, environmental impact studies and in determin- 40, 1100–1114.
Borja, Á., Muxika, I., Franco, J., 2003a. The application of a Marine
ing the ecological status (as in WFD). Biotic Index to different impact sources affecting soft-bottom
However, the AMBI has not been shown to be benthic communities along European coasts. Mar. Poll. Bull. 46,
useful in naturally-stressed and poor communities, 835–845.
30 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31

Borja, Á., Franco, J., Muxika, I., 2003b. Classification tools for Fielding, A., 1992. Axiomatic approaches to scoring ordered clas-
marine ecological quality assessment: the usefulness of macro- sifications. Department of Economics, University of Birming-
benthic communities in an area affected by a submarine outfall. ham. Discussion Paper 92-06, 30 pp.
ICES CM2003/Session J-02, Tallin, Estonia, 24–28 September Forni, G., Occhipinti Ambrogi, A., 2003. Applicazione del coeffi-
2003. ciente biotico (Borja et al., 2000) alla comunità macrobentonica
Borja, Á., Franco, J., Valencia, V., Bald, J., Muxika, I., Belzunce, del Nord Adriático. Meeting of the Italian Society of Marine
M.J., Solaun, O., 2004a. Implementation of the European Water Biology, Tunisia, June 2003.
Framework Directive from the Basque country (northern Spain): Frascari, F., Rosso, G., Bortoluzzi, G., Barbanti, A., Bonvicini,
a methodological approach (viewpoint). Mar. Poll. Bull. 48, A.M., Pagliai, R., Crema, A., Castelli, M., Mauri, R., Zunarelli,
209–218. E., Orlando, D., Prevedelli, L., Ceffa, Ratti, S., 1992. Environ-
Borja, Á., Franco, J., Muxika, I., 2004b. The biotic indices and the mental impact of water based drilling muds and cuttings in a
Water Framework Directive: the required consensus in the new northern Adriatic Sea site. Bull. Inst. Océanogr., Monaco, no
benthic monitoring tools (correspondence). Mar. Poll. Bull. 48, special 11, 305–324.
405–408. Glémarec, M., Hily, C., 1981. Perturbations apportées à la macro-
Breuer, E., Howe, J.A., Shimmield, G.B., Cummings, D., Carroll, J., faune benthique de la baie de Concarneau par les effluents
1999. Contaminant Leaching from Drill Cuttings Piles of the urbains et portuaires. Acta Oecol. 2, 139–150.
Northern and Central North Sea: A review. Centre for Coastal & Gorostiaga, J.M., Borja, Á., Dı́ez, I., Francés, G., Pagola-Carte, S.,
Marine Sciences and Scottish Association for Marine Science, Sáiz-Salinas, J.I., 2004. Recovery of benthic communities in
Dunstaffnage Marine Laboratory, Scotland, UK, 49 pp. polluted systems. In: Borja, Á., Collins, M. (Eds.). Oceanogra-
Breuer, E., Stevenson, A.G., Howe, J.A., Carroll, J., Shimmield, phy and Marine Environment of the Basque Country, Elsevier,
G.B., 2004. Drill cutting accumulations in the Northern and Amsterdam. Elsevier Oceanography Series no. 70, pp. 549–578.
Central North Sea: a review of environmental interactions and Grall, J., Glémarec, M., 1997. Using biotic indices to Estimate
chemical fate. Mar. Poll. Bull. 48, 12–25. Macrobenthic Community Perturbations in the Bay of Brest. Est.
Casselli, C., Ponti, M., Abbiati, M., 2003. Valutazione della qualità Coast. Shelf Sci. 44, 43–53.
ambientale della laguna costiera Pialassa Baiona attraverso lo Gray, J.S., 1979. Pollution-induced changes in populations. Phil.
studio dei suoi popolamenti bentonici. In: XIII Congresso Trans. Royal Soc., London Ser. B 286, 545–561.
Societá Italiana de Ecologı́a, Como, Villa Olmo, 8–10 Settembre Guerra-Garcı́a, J.M., Corzo, J., Garcı́a-Gómez, J.C., 2003. Short-
2003, poster. term benthic recolonization after dredging in the harbour of
Clarke, K.R., 1990. Comparisons of dominance curves. J. Exp. Mar. Ceuta, North Africa. P.S.Z.N.: Mar. Ecol. 24, 217–229.
Biol. Ecol. 138, 143–157. Hill, M.O., 1974. Correspondence analysis: a neglected multivariate
Codling, I.D., Ashley, S.J., 1992. Development of a biotic index for method. Appl. Stat. 23, 340–354.
the assessment of pollution status of marine benthic commu- Hily, C., 1984. Variabilité de la macrofaune benthique dans les
nities. Final Report to SNIFFER and NRA, NR 3102/1. milieux hypertrophiques de la Rade de Brest. Thèse de Doctorat
Dauer, D.M., 1993. Biological criteria, environmental health and d’Etat, Univ. Bretagne Occidentale. vol. 1, 359 pp.; vol. 2, 337
estuarine macrobenthic community structure. Mar. Poll. Bull. pp.
26, 249–257. Jolliffe, I.T., 1986. Principal component analysis. Springer–Verlag,
Davies, J.M., Addy, J.M., Blackman, R.A., Blanchard, J.R., Fer- New York, 271 pp.
brache, J.E., Moore, D.C., Somerville, H.J., Whitehead, A., Karakassis, I., Tsapakis, M., Hatziyanni, E., Papadopoulou, K.N.,
Wilkinson, T., 1984. Environmental effects of the use of oil- Plaiti, W., 2000. Impact of cage farming of fish on the seabed in
based drilling muds in the North Sea. Mar. Poll. Bull. 15, 363– three Mediterranean coastal areas. ICES J. Mar. Sci. 57, 1462–
371. 1471.
Dernie, K.M., Kaiser, M.J., Richardson, E.A., Warwick, R.M., Kendall, M.G., 1975. Multivariate analysis. Charles Griffin &
2003a. Recovery of soft sediment communities and habitats Company Ltd, London, 210 pp.
following physical disturbance. J. Exp. Mar. Biol. Ecol. 285– Kruskal, J.B., Wish, M., 1977. Multidimensional Scaling. Sage
286, 415–434. Publications, Beverly Hills, California, 96 pp.
Dernie, K.M., Kaiser, M.J., Warwick, R.M., 2003b. Recovery dates McArthur, R.H., Wilson, E.O., 1967. The theory of island biogeo-
of benthic communities following physical disturbance. J. Anim. graphy. Princeton University Press, Princeton, USA, 224 pp.
Ecol. 72, 1043–1056. Majeed, S.A., 1987. Organic Matter and Biotic Indices on the
Eaton, L., 2001. Development and validation of biocriteria using Beaches of North Brittany. Mar. Poll. Bull. 18, 490–495.
benthic macroinvertebrates for North Carolina estuarine waters. Mazzola, A., Mirto, S., la Rosa, T., Fabiano, M., Danovaro, R., 2000.
Mar. Poll. Bull. 42, 23–30. Fish-farming effects on benthic community structure in coastal
Engle, V.D., Summers, J.K., Gaston, G.R., 1994. A benthic index of sediments: analysis of meiofaunal recovery. ICES J. Mar. Sci.
environmental condition of Gulf of Mexico estuaries. Estuaries 57, 1454–1461.
17, 372–384. Muxika, I., Borja, Á., Franco, J., 2003. The use of a biotic index
Field, J.G., Clarke, K.R., Warwick, R.M., 1982. A practical strategy (AMBI) to identify spatial and temporal impact gradients on
for analysing multispecies distribution patterns. Mar. Ecol. Prog. benthic communities in an estuarine area. ICES CM2003/Ses-
Ser. 8, 37–52. sion J-01.
 I. Muxika et al. / Ecological Indicators 5 (2005) 19–31 31

Pearson, T., Rosenberg, R., 1976. A comparative study of the effects of the northern and central North Sea: field results from the Beryl
on the marine environment of wastes from cellulose industries in ‘‘A’’ cuttings pile. Scottish Association for Marine Science,
Scotland and Sweden. Am. Biol. 5, 77–79. 28 pp.
Pearson, T., Rosenberg, R., 1978. Macrobenthic succession in Smith, S.D.A., Rule, M.J., 2001. The effects of dredge-spoil dump-
relation to organic enrichment and pollution of the marine ing on a shallow water soft-sediment community in the Solitary
environment. Oceanogr. Mar. Biol. Ann. Rev. 16, 229–311. Islands Marine Park, NSW, Australia. Mar. Poll. Bull. 42, 1040–
Pianka, E.R., 1970. On r- and K-selection. Am. Nat. 104 (940), 592– 1048.
597. Sokal, R.R., Sneath, P.H.A., 1963. Principals of Numerical Taxon-
Reish, D.J., 1959. An Ecological Study of Pollution in Los omy. Freeman and Co, San Francisco, CA, USA, 339 pp.
Angeles—Long Beach Harbors, California. Allan Hancock Solı́s-Weiss, V., Aleffi, F., Bettoso, N., Rossin, P., Orel, G., Fonda-
Foundation Publications. Occasional Paper 22, 117 pp. Umani, S., 2004. Effects of industrial and urban pollution on the
Roberts, R.D., Gregory, M.R., Foster, B.A., 1998. Developing an benthic macrofauna in the Bay of Muggia (industrial port of
efficient macrofauna monitoring index from an impact study—a Trieste, Italy). Sci. Tot. Env. 328, 247–263.
dredge spoil example. Mar. Poll. Bull. 36, 231–235. Van dolah, R.F., Hyland, J.L., Holland, A.F., Rosen, J.S., Snoots,
Rosenberg, R., Agrenius, S., Hellman, B., Nilsson, H.C., Norling, T.R., 1999. A benthic index of biological integrity for assessing
K., 2002. Recovery of marine benthic habitats and fauna in a habitat quality in estuaries of the southeastern U.S.A. Mar.
Swedish fjiord following improved oxygen conditions. Mar. Environ. Res. 48, 269–283.
Ecol. Prog. Ser. 234, 43–53. Vanosmael, C., Heip, C., Vincx, M., Claeys, D., Rappé, G., Braeck-
Rygg, B., 1985. Distribution of species along pollution-induced man, A., Van Gansbeke, D., 1979. De invloed van zandwinning
diversity gradients in benthic communities in Norwegian fjords. op de bodemfauna voor de Belgische kust. Ministerie van
Mar. Poll. Bull. 16 (12), 469–474. volksgezondheid en leefmilieu, Brussels, 50 pp.
Salas, F., Neto, J.M., Borja, A., Marques, J.C., 2004. Evaluation of Vanosmael, C., Willems, K.A., Claeys, D., Vincx, M., Heip, C.,
the applicability of a marine biotic index to characterise the 1982. Macrobenthos of a sublittoral sandbank in the Southern
status of estuarine ecosystems: the case of Mondego estuary Bight of the north Sea. J. Mar. Biol. Assoc. UK 62, 521–
(Portugal). Ecol. Indic. 4, 215–225. 534.
Sánchez-Moyano, J.E., Estacio, F.J., Garcı́a-Adiego, E.M., Garcı́a- Warwick, R.M., 1986. A new method for detecting pollution
Gómez, J.C., 2004. Dredging impact on the benthic community effects on marine macrobenthic communities. Mar. Biol. 92,
of an unaltered inlet in southern Spain. Helgol. Mar. Res. 58, 32– 557–562.
39. Washington, H.G., 1984. Diversity, biotic and similarity indices: a
Schiffman, S.S., Reynolds, M.L., Young, F.W., 1981. Introduction to review with special relevance to aquatic ecosystems. Water Res.
Multidimensional Scaling: Theory, Methods and Applications. 18, 653–694.
Academic Press, New York, 440 pp. Weisberg, S.B., Ranasinghe, J.A., Schaffner, L.C., Dı́az, R.J., Dauer,
Shimmield, G.B., Breuer, E., Cummings, D.G., Peppe, O., Shim- D.M., Frithsen, J.B., 1997. An estuarine benthic index of biotic
mield, T., 2000. Contaminant leaching from drill cuttings piles integrity (B-IBI) for Chesapeake Bay. Estuaries 20, 149–158.