MARINE ECOLOGY PROGRESS SERIES Mar Ecol INTRODUCTION

Full text

MARINE ECOLOGY PROGRESS SERIES
Mar Ecol Prog Ser
Vol. 222: 51–62, 2001 Published November 5
INTRODUCTION
Potential impacts of large sewage outfalls on inter-
tidal rocky shore assemblages have been described in
the literature (Borowitzka 1972, Littler & Murray 1975,
© Inter-Research 2001
*Present address: Regional Oceans and Environment Branch,
Fisheries and Oceans Canada, Institut Maurice-Lamon-
tagne, 850, Route de la Mer, CP 1000, Mont-Joli, Québec
G5H 3Z4, Canada. E-mail: archambaultp@dfo-mpo.gc.ca
Temporal variation in the structure of intertidal
assemblages following the removal of sewage
P. Archambault1,*, K. Banwell2, A. J. Underwood1
1Centre for Research on Ecological Impacts of Coastal Cities, Marine Ecology Laboratories A11, University of Sydney,
Sydney, New South Wales 2006, Australia
2New South Wales Department of Environmental Health, North Sydney, New South Wales 2059, Australia
ABSTRACT: Following the closure of 2 outfalls, changes in the number of species and abundance of
intertidal organisms at 2 decommissioned outfalls were compared with control (sewage outfalls that
remained in operation) and reference (natural) areas. Two intertidal levels (mid- and low-shore) were
sampled 5 times over 2 yr following the closure of these outfalls. It was proposed that the number of
species would increase and the abundance of green algae would decrease through time at the
decommissioned outfalls, while no noticeable changes in the number of species and the abundance
of green algae were predicted at the reference and control locations. The 4 outfalls (2 decommis-
sioned and 2 controls) were analysed separately with an asymmetrical ANOVA to identify differences
between the outfall and the average of their 2 respective reference locations. Non-metric multi-
dimensional scaling ordination using Bray-Curtis similarity was used to identify differences between
outfalls and reference locations in the structure of the assemblages at the 2 heights on the shore. Dur-
ing the first (6 mo after closure) and second sampling period, fewer species and greater abundance of
green algae were observed at every outfall than at their respective reference locations in low-shore
areas. In the subsequent sampling periods, the number of species at the decommissioned outfalls
increased through time while there were consistently fewer species at the control outfalls. The oppo-
site pattern was observed for the abundance of green algae (i.e., decreases through time at the
decommissioned outfalls). Assemblages at control outfalls never clustered with their reference loca-
tions. At midshore levels, no pattern was observed at any sampling date. Furthermore, at the first
time of sampling, the number of species was not smaller at the outfalls than at reference locations.
These results suggest that sewage outfalls have little impact on the number of midshore species. The
results of univariate analysis were similar to those from multivariate analyses. Assemblages of spe-
cies in low-shore areas at the decommissioned outfalls were different from those at reference loca-
tions at the first sampling date, but became more similar by the last sampling date. Some exceptions
were observed at the site closest to the point of discharge. Again no differences in abundances of
organisms were observed at any time for any outfall. The results showed a recovery of the benthic
assemblages in the low-intertidal zone after the closure of 2 sewage outfalls. Furthermore, this study
supports the importance of using more than 1 reference and control areas to measure recovery of a
site without ambiguity.
KEY WORDS: Sewage outfalls · Recovery · Human disturbance · Species richness · Green ephemeral
algae · Benthic assemblage · Experimental design
Resale or republication not permitted without written consent of the publisher

Mar Ecol Prog Ser 222: 51– 62, 2001
Axelrad et al. 1981, Smith et al. 1981, May 1985, Fair-
weather 1990, López Gappa et al. 1990, 1993, Bell-
grove et al. 1997). Outfalls discharging waste water of
varying quality and volume have been shown to alter
the structure of nearby assemblages (Borowitzka 1972,
Fairweather 1990). Particular types of algae such as
ephemeral green algae and animals such as Serpulor-
bis squamigerus (Littler & Murray 1975, Brown et al.
1990, Fairweather 1990, Bellgrove et al. 1997) have
been described as occurring near outfalls.
Most studies on ecological impacts of sewage on
marine assemblages have been comparisons between
reference and affected locations (Borowitzka 1972, Lit-
tler & Murray 1975, Brown et al. 1990, Fairweather
1990, Chapman et al. 1995, Bellgrove et al. 1997) or
before and after (Before After Control Impact—BACI
design, Underwood 1991, 1992, 1994) the start of dis-
charge (Roberts et al. 1998, Ajani et al. 1999, Smith et
al. 1999). Relatively few studies have documented the
recovery of benthic assemblages following the cessa-
tion of a long-standing discharge of sewage (Pearson &
Rosenberg 1978, Smith et al. 1981). In the 32 mo after
the cessation of the discharge of sewage into Kaneohe
Bay, Hawaii, Smith et al. (1981) observed a large de-
cline in the biomass of benthic animals, especially in
the first year. They concluded that the response of those
benthic assemblages to the diversion was very rapid.
Furthermore, the authors assumed that the major role
of the sewage was nutritional rather than toxic. This
differs from many other studies, which have concluded
that sewage has only negative effects (Fairweather
1990, López Gappa et al. 1990, Roberts et al. 1998).
Few studies have shown the full scale of these
changes, with data gathered before the pollution
ceased and through sufficient time until no observable
differences between the previously polluted sites and
reference sites were distinguishable. Further, despite
improvements in sampling designs and analytical tech-
niques to measure outcomes of disturbances (Eberhert
& Thomas 1991, Underwood 1992, 1994, 1996), many
recent studies (Calcagno et al. 1997, Calcagno et al.
1998, Taylor et al. 1998, Lopez-Rodriguez et al. 1999)
remain poorly replicated and proper controls are com-
monly lacking. Recovery following cessation of a dis-
turbance has to be measured as an interaction in time
and space with an appropriate experimental design
(Green 1979).
In 1991, 2 large metropolitan shoreline outfalls in
the most heavily populated region of Australia
(Sydney-Illawarra region) were closed (Otway 1995a).
This provided an ideal opportunity for a spatially
replicated investigation of the effects of the removal
of sewage from rocky intertidal assemblages. This
closure had largely been driven by public concerns
about human health and the environment. The effects
of sewage on intertidal assemblages had been studied
at these outfalls before they were decommissioned
(Borowitzka 1972, Fairweather 1990). Borowitzka
(1972) observed a smaller number of algal species,
especially Phaeophyceae and the Rhodophyceae,
near an outfall than in reference locations. Green
ephemeral algae (Ulvoid), which dominated rocky
shores adjacent to outfalls, were studied by Fair-
weather (1990). The ulvoid mat reached 100% cover
in many places near outfalls and abundances of ani-
mals were smaller than at reference locations. Gener-
ally, assemblages close to outfalls were dominated by
early-stage colonising species.
From these studies and a study of the same locations
by Banwell (1996, unpubl. data), it appears that
sewage was the dominant factor determining the
structure of intertidal assemblages near outfalls. De-
commissioning of several local outfalls provided an
excellent opportunity to examine this model for differ-
ent assemblages near outfalls. We had reference loca-
tions where there were no outfalls and therefore no
sewage. Assemblages in these locations represent
background, ‘undisturbed’ shores , i.e. unaffected by
the particular disturbance due to discharge of sewage.
Outfalls to be decommissioned can be considered as
replicates of the experimental treatment, but were
analysed separately and the results compared. Outfalls
that continued to discharge sewage provided repli-
cated controls. This allowed tests of specific hypothe-
ses about recovery of biota (to match the current situa-
tion in reference areas). The design also unconfounds
changes due to decommissioning from any coincident
background or natural changes, which would also
affect controls. The rationale and logic of this experi-
mental design have been discussed in detail by Chap-
man (1999) in the context of recovery of disturbed
habitats.
METHODS
Study sites. Four shore-line outfalls in the Sydney
metropolitan region of New South Wales, Australia
(Fig. 1) were studied. Two outfalls, North Head and
Malabar, were to cease operation with the opening of
new deepwater ocean outfalls (Otway 1995a). Two
other outfalls at Potter Point and Bellambi, which were
to continue to operate and discharge effluent near the
shore, were selected as controls. The average (during
dry season) daily flows of sewage through North Head,
Malabar, Potter Point and Bellambi in 1991 were
295, 475, 46 and 20 Ml d–1, respectively. The 4 out-
falls received primary treatment with the exception
of Bellembi during the dry season, where the effluent
received an additional treatment with chemically
52

Archambault et al.: Temporal variation in intertidal assemblages
assisted sedimentation. The depth of the discharge at
North Head and Malabar was approximately 7 m be-
low mean water level and just below the mean low tide
level at Potter Point. The discharge pipe at Bellambi
was exposed during low tide; effluent was discharged
at the sea surface. More detail on sewage at each out-
fall is available in the Camp, Dresser & McKee Inter-
national Inc. (CDM 1989) review and in Otway (1995a).
For each region, 2 reference locations were selected to
be as similar as possible to the outfall with respect to
aspect, wave exposure, slope of the shore and geologi-
cal features. Wherever possible, they were chosen to
be either side of the outfall. The reference locations for
North Head and Potter Point were both located north
of the outfall (Fig. 1) because of restrictions of access
and because there were no suitable locations to the
south of the outfalls. Reference locations were approx-
imately 600 and 1000 m from North Head and 400 and
5000 m from Potter Point. At Malabar, the reference
locations were north and south of the outfall at 1300
and 1000 m, respectively. The north and south refer-
ences at Bellambi were 300 and 500 m from the out-
fall. At 2 tidal heights (low- and midshore), 3 sites
(4 ×4 m) were randomly selected within each location.
Sites were marked to ensure that the same areas were
sub-sampled independently at later times. These sites
were between 25 and 50 m from the point of discharge
on each shore. Sampling was done 5 times between
1991 and 1993 (July/August 1991, 1992 and 1993,
October/November 1991 and April/May 1992). Sea-
sonal differences could not be analysed because there
was only 1 sample in each season (Underwood 1993).
The first samples were taken 6 mo after the sewage
outfall ceased to be discharged. It was not possible to
sample earlier because the construction of the new
deep-water outfall was completed 10 mo ahead of
schedule. During these 6 mo, however, effluent was
released from the decommissioned outfalls when re-
pairs and improvements were made. Sampling in
July/August 1991 was therefore deemed to be ‘before’
decommissioning. There was no sampling in July/
August 1991 at Bellambi and no data were collected at
the midshore level at this outfall at any time because
there was no midshore habitat close to the outfall.
Sampling. Five randomly placed quadrats (50 ×
50 cm) were sampled for intertidal assemblages at
each site at each time. The percentage covers of algae
and sessile animals in each quadrat were estimated
from 100 regularly spaced points. Organisms under
each point were recorded, and organisms seen but
not recorded under a point were noted as present.
Canopy-forming species were sampled first, then were
moved to 1 side to sample the understorey. Organisms
were identified to the greatest taxonomic resolution
possible, but many groups, particularly red algae,
were hard to identify because individual species are
not well known or documented in Australia (Clayton &
King 1990). These groups were given common names
to identify the particular morphotaxa. Specimens were
collected to confirm field identifications. Motile ani-
mals found within the quadrats were all identified to
species level and the numbers of individuals were
recorded. Where numbers were too large to count, 5
small sub-quadrats (4.5 ×4.5 cm) were counted and the
numbers multiplied up to the area of the quadrat. No
attempt was made to count small fast-moving animals
such as amphipods.
Statistical analysis. Univariate analyses: Differen-
ces in the mean number of species in the low (Table 1)
and midshore (Table 2) habitats and the mean percent-
age cover of green algae (Table 3) were analysed using
an asymmetrical design to identify differences be-
tween the outfall and the average of the 2 reference
locations (Underwood 1993, 1994, Glasby 1997).
Recovery was measured as an interaction in time and
space between the 2 decommissioned outfalls and
their respective reference locations. Data for the 4
regions were analysed separately. Variances were
homogeneous in all analyses (determined using
53
Fig. 1. Map of the coast of Sydney, New South Wales (NSW),
Australia, showing the 4 regions and the positions of the loca-
tions sampled. *Reference locations; arrows: outfalls

Mar Ecol Prog Ser 222: 51– 62, 2001
Cochran’s C-test) except in 2 cases where the total
number of species in the low- and midshore for North
Head were ln(x+ 1) transformed to achieve homo-
geneity. In addition, Student-Newman-Keuls (SNK)
tests were done on appropriate means, so that differ-
ences could be identified in relation to the specified
hypotheses.
Multivariate analyses: Differences in the structure
of whole assemblages among the 3 locations in each
region were identified by non-metric multi-dimen-
sional scaling (nMDS) ordination on untransformed
data, using the Bray-Curtis similarity measure (Clarke
1993). The centroid of the 5 quadrats in each site was
used to evaluate the dissimilarity. Following the uni-
variate analyses, it was decided to use only July/
August 1991 and July/August 1993 in the analyses on
total number of species in the low- and midshore (see
‘Results’ section. Analysis of similarities (ANOSIM)
testes for differences in the composition of assem-
blages and is a procedure that cannot handle experi-
mental designs with more than 2 factors (Clarke 1993)
and therefore was not used here. To test predictions
54
Source of variation Decommissioned outfalls Operational outfalls (controls)
North Head Malabar Potter Point Bellambi
df MS FpMSFpMSFpdf MS Fp
Location (L) 2 2
Outfall vs reference 1 24.94 10.7 ns 2954.24 19.96 ns 8862.24 167.82 * 1 6084.44 1267.59 *
Between references 1 2.33 20.44 * 148.01 2.09 ns 52.81 3.12 ns 1 4.8 0.07 ns
Time (T) 4 1.19 2.11 ns 115.57 1.06 ns 102.47 5.09 * 3 14.16 2.07 ns
T ×L8 6
T ×outfall vs reference
40.84 2.85 ns 149.72 2.21 ns 39.62 2.21 ns 3 3.24 0.31 ns
T ×between references
40.29 5.2 ** 67.84 1.87 ns 0.64 0.04 ns 3 10.4 0.86 ns
Sites (location) 6 6
Sites (outfall) 2 9.14 176.84 *** 341.85 56.08 *** 95.52 14.62 *** 2 120.98 69.05 ***
Sites (references) 4 0.11 2.2 ns 70.81 11.62 *** 16.91 2.59 * 4 219.11 22.66 ***
T ×sites (L) 24 18
T ×sites (outfall) 8 0.61 11.75 *** 21.75 3.57 *** 8.25 1.26 ns 6 0.84 0.26 ns
T ×sites (references) 16 0.06 1.09 ns 36.25 5.95 *** 17.27 2.64 *** 12 12.06 3.8 ***
Residual 180 0.05 6.1 6.53 144 3.17
Total 224 179
Table 1. Asymmetrical ANOVA comparing the total number of species in low-shore habitats at different times at 3 sites at 1 out-
fall and 2 reference locations. The data from North Head were ln(x+ 1) transformed. ns: p > 0.05; *p < 0.05; **p ≤0.01;
***p ≤0.001
Source of variation Decommissioned outfalls Operational outfalls (controls)
North Head Malabar Potter Point
df MS FpMSFpMSFp
Location (L) 2
Outfall vs reference 1 2.81 10.43 ns 3.38 0.08 ns 0.27 0.02 ns
Between references 1 0.27 1.21 ns 41.61 5.97 ns 12.33 2.42 ns
Time (T) 4 1.98 0.91 ns 33.62 0.65 ns 9.47 0.64 ns
T ×L8
T ×outfall vs reference 4 1.04 0.31 ns 93.26 9.88 * 14.49 0.98 ns
T ×between references 4 3.3 23.96 *** 9.44 1.3 ns 14.74 2.3 ns
Sites (location) 6
Sites (outfall) 2 8.03 134.98 *** 31.21 18.1 *** 239.77 120.69 ***
Sites (References) 4 0.22 3.72 ** 6.97 4.04 ** 5.09 2.56 *
T ×sites (L) 24
T ×sites (outfall) 8 1.52 25.4 *** 5.21 3.02 ** 24.64 12.4 ***
T ×sites (references) 16 0.14 2.3 ** 7.24 4.2 *** 6.41 3.23 ***
Residual 180 0.06 1.72 1.99
Total 224
Table 2. Asymmetrical ANOVA comparing the total number of species in the midshore habitats at 5 different times, at 3 sites at
1 outfall and 2 reference locations for each outfall sampled. The data from North Head were ln(x+ 1) transformed. ns: p > 0.05;
*p < 0.05; **p ≤0.01; ***p ≤0.001

Archambault et al.: Temporal variation in intertidal assemblages
about the assemblages and interactions through time
and space, percentage dissimilarities were calculated
between locations for the 2 times. Percentage dissimi-
larities were evaluated only for the nMDS ordinations
showing a specific pattern. If the 2 decommissioned out-
fall locations were recovering through time, there
should be a decrease in percentage dissimilarities
between the outfall and the reference locations
through time. This would indicate that assemblages at
outfall locations were becoming more similar to those
at the reference locations. It was predicted that the
non-decommissioned outfall locations were going to
stay largely dissimilar from their reference locations
through time. One site at each of North Head and Mal-
abar outfalls (the site closest to the point of discharge
on each shore) had to be excluded from the evaluation
of the percentage dissimilarity because the site was
considerably disturbed by recreational fishermen
(Banwell 1996) throughout the study. Banwell (1996)
observed that fishermen kept these 2 outfall sites clear
55
Source of variation Decommissioned outfalls Operational outfalls (controls)
North Head Malabar Potter Point Bellambi
df MS FpMSFpMSFpdf MS Fp
Location (L) 2 2
Outfall vs reference 1 18 611.6 458.87 * 260 257 7.79 ns 1004 039 92.5 ns 1 532 110 277.7 *
Between references 1 40.56 0.01 ns 33 391 101.73 *** 10855 3.92 ns 1 1916 1.52 ns
Time (T) 4 7644.6 1.73 ns 22 862 2.65 ns 13206 3.97 * 3 5047 1.79 ns
T ×L8 6
T ×outfall vs reference
46698.5 3.17 ns 12 100 2.35 ns 5079 3.24 ns 3 5337 18 *
T ×between references
42115.8 0.69 ns 5139 8.17 *** 1570 2.48 ns 3 297 0.93 ns
Sites (location) 6 6
Sites (outfall) 2 156226.7 194.61 *** 16437 31.7 *** 7362 17.83 *** 2 109997 1393 ***
Sites (references) 4 3967.4 4.94 *** 328 0.63 ns 2767 6.7 *** 4 1257 15.92 ***
T ×sites (L) 24 18
T ×sites (outfall) 8 4903.1 6.11 *** 8155 15.73 *** 1501 3.64 *** 6 7540 95.51 ***
T ×sites (references) 16 3045.9 3.79 *** 629 1.21 ns 632 1.53 ns 12 318 4.02 ***
Residual 180 802.8 518 413 144 79
Total 224 179
Table 3. Asymmetrical ANOVA comparing the total percentage cover of green algae in low-shore habitats at different times at
3 sites at 1 outfall and 2 reference locations for each outfall sampled. ns: p > 0.05; *p < 0.05; ***p ≤0.001
Fig. 2. Mean (+SE) num-
ber of species through
time in low-shore habi-
tats for each site at the
outfall (black column),
and north (white column)
and south (hashed col-
umn) reference locations.
North Head and Malabar
are the 2 experimental
regions, i.e. where the out-
fall was decommissioned.
No data could be col-
lected at Bellambi in July/
August 1991. The site
closest to the point of dis-
charge at the outfalls is
the left black column for
each sampling time

Mar Ecol Prog Ser 222: 51– 62, 2001
of algae to increase their safety while standing there.
This disturbance has been shown experimentally (Ban-
well 1996) to explain why there was no recovery in the
number of species at these 2 sites. The great abun-
dance of green ephemeral algae at these sites sug-
gested that these sites are frequently disturbed.
Human trampling (Brosnan & Crumrine 1994, Banwell
1996) and baits harvested by fishermen have been
shown to modify assemblages on rocky shores, includ-
ing in New South Wales (Underwood & Kennelly 1990,
Kingsford et al. 1991). This disturbance was major and
these sites were clearly not appropriate to include in
the design.
RESULTS
Forty-six species of algae and 39 species of inverte-
brates were identified during this study (the most com-
mon species are listed in Table 4). Seven and 11 spe-
cies decreased in their abundance at the decom-
missioned outfalls of North Head and Malabar, respec-
tively. Decreases were mainly of ephemeral and op-
portunistic species (e.g. Chlorophycae). At least 7 spe-
cies were never recorded at the outfalls throughout the
experiment. More species increased than decreased in
abundance at the decommissioned outfalls. Species
that increased were mostly Phaeophyta and Rhody-
phyta algae. Many, particularly Phaeophyta, were not
present in July/August 1991, but recruited and in-
creased in number of species and in abundance during
the experiment. Twenty-one (North Head) and 15
(Malabar) species increased their abundance at these 2
outfalls (Table 4).
Univariate analyses
Low-shore areas
For the total number of species, there were signifi-
cant interactions of time ×site within reference loca-
tions in the 2 control regions with the operational out-
falls (Table 1, Fig. 2, Potter Point and Bellambi), but not
at the outfall location. Thus, there was greater vari-
ability through time between reference sites than at
outfall sites. Significant interactions of time ×site
within the outfall locations were also identified in the
2 regions with decommissioned outfalls (Fig. 2). In the
first and second samples, there were fewer species at
all outfall locations (decommissioned and control) than
at their respective reference locations. The total num-
ber of species remained smaller at the outfall locations
than the reference locations in the control sets of data
but increased through time at Malabar and North
Head (decommissioned outfalls; Fig. 2). There were
always significantly fewer species at the outfall site
closer to the point of discharge than at the 2 others sites
for North Head, Malabar and Potter Point (SNK tests
on data in Fig. 2). These differences persisted for 2 yr.
Because of the significance of interactions at the spa-
tial scale of sites, there is no appropriate hypothesis for
56
Fig. 3. Mean (+SE) percentage
cover of green algae through
time in low-shore habitats for
each site at the outfall (black
column), and north (white
column) and south (hashed
column) reference locations.
North Head and Malabar are
the 2 experimental regions, i.e.,
where the outfall was de-
commissioned. No data could
be collected at Bellambi in
July/August 1991. The site
closest to the point of discharge
at the outfalls is the left black
column for each sampling
time. The data illustrated here
include canopy and under-
storey algae

Archambault et al.: Temporal variation in intertidal assemblages
examining interactions at the larger spatial scale of
locations (Underwood 1992).
Asymmetrical analyses of mean total percentage
cover of green algae showed a significant interaction
of time ×site within outfalls in the 4 regions (Table 3).
The total percentage cover of green algae was highly
variable among outfall sites in 3 regions (North Head,
Malabar and Bellambi). At North Head, the outfall site
closest to the point of discharge had a large mean
cover at all times. The other 2 sites at this outfall had a
smaller cover of green algae than at most of the refer-
ence sites (Fig. 3). At Malabar, there was a greater per-
centage cover of green algae at the first sampling date
at the 3 sites around the outfall than at any reference
sites. Percentage cover decreased through time, but
remained, generally greater at the outfall than at the
reference sites, apart from the last date of sampling
(July/August 1993; Fig. 3). Percentage cover at the out-
fall sites in the Potter Point region were always greater
than at the respective reference sites. The same pat-
tern was observed at the other control outfall (except
for 1 site in July/August 1992 and 1993; Fig. 3). The
abundance of green algae at the outfall site closest to
the point of discharge at North Head was significantly
greater than any outfall sites or reference sites for 4
times (except in July/August 1992; SNK tests).
Midshore areas
Sewage had little effect on the number of midshore
species. The total number of midshore species varied
significantly through time and site within outfall and
reference locations for the 3 regions (Table 2), but
there were no patterns through time. There were more
species at the first sampling date for the 2 decommis-
sioned outfalls (Fig. 4). At Potter Point, there was no
pattern of difference among sites at the outfall, nor
among sites at the 2 reference locations. The number
of species at the outfall site closest to the point of dis-
charge was significantly smaller (SNK tests; p < 0.05,
see data in Fig. 4) than at the 2 other outfall sites at
Malabar and Potter Point.
Multivariate analysis
Low-shore areas
The assemblages in the outfall sites closest to the
point of discharge at North Head and Malabar differed
from those in the 2 other sites at each of these outfalls
in July/August 1991. In July/August 1993, assem-
blages at outfall sites tended to become more similar to
57
Species North Head Malabar Species North Head Malabar
↓=↓=↑=↑↓=×
Decreasing or equal Mostly increasing
Cyanobacteria ••Cladophora spp. ••
Chaetomorpha aurea ••Colpomenia sinuosa ••
Enteromorpha spp. ••Ectocarpus spp. ••
Ulva lactuca ••Petalonia fascia ••
Corallina officinalis ••Ralfsia verrucosa ••
Gelidium pusillum ••Sargassum spp. ••
Littorina unifasciata ••Ceramium spp. ••
Siphonaria virgulata •• Dictyothamnion sp. ••
Tesseropora rosea •• Gigartina sp. ••
Patelloida latistrigata ••Gracilaria sp. ••
Hildenbrandia prototypus ••
Laurencia botyroides ••
Laurencia pinnosa ••
Encrusting Corallines ••
Polysiphonia spp. ••
Pterocladia capillacea ••
Cellana tramoserica ••
Class Polyplacophora ••
Galeolaria caespitosa ••
Kerguelenella sp. ••
Montfortula rugosa ••
Siphonaria denticulata ••
Table 4. Common species of macroalgae and animals that changed in abundance at the decommissioned outfalls, Malabar and
North Head. Species that increased in abundance are labelled with an upward arrow (↑), species that decreased in abundance
are labelled with a downward arrow (↓), and those that were not altered in abundance are represented by ‘=’. Species not
present at every time are labelled ‘×’. Biota found in mid- and low-shore habitats are combined in the table

Mar Ecol Prog Ser 222: 51– 62, 2001
those at reference sites at North Head, but not at Mal-
abar. At North Head, the assemblage in the outfall site
closest to the pipe remained different, while the other
2 sites became similar to reference sites and clearly
moved to be within the cluster of reference sites
(Fig. 5a). The patterns in the 2 regions with decom-
missioned outfalls were supported by comparisons of
the average percentage dissimilarities
among the 3 locations (Table 5). Per-
centage dissimilarities between the
outfall location at North Head and the
2reference locations were greater in
July/August 1991 than the percentage
dissimilarity between the 2 reference
locations. The outfall differed from
other places more than due to natural
spatial variation. After 2 yr (July/August
1993), the outfall sites at North Head
(the closest site to the point of dis-
charge was removed from the analysis,
see ‘Methods’) became less dissimilar
from the northern (49 %) and southern
(61%) reference locations. Even after
2yr of removal of sewage, the outfall
location at Malabar stayed dissimilar
from the reference locations (Table 5).
This is also visible in the nMDS ordina-
tion (Fig. 5b)
There was a distinct separation of the
outfall sites and all reference sites for
the 2 times at the operational outfalls, Potter Point and
Bellambi (Fig. 5c,d). No changes through time were
observed. Moreover, the percentage dissimilarities
between outfall and reference locations tended to stay
large and were always larger than the percentage dis-
similarities between the relevant reference locations
(Table 5).
58
Outfall- Outfall- Northern reference
Southern Northern –
reference reference Southern reference
Decommissioned outfalls
North Head
Time 1 69 54 43
Time 5 61 49 60
Malabar
Time 1 69 79 68
Time 5 67 81 58
Operational outfalls (control)
Potter Point
Time 1 89 92 38
Time 5 86 96 46
Bellambi
Time 2 92 97 64
Time 5 87 83 60
Table 5. Percentage dissimilarity between locations in low-shore habitats at
2 different times for the 2 decommissioned outfalls (North Head, Malabar) and
the 2 control outfalls (operational; Potter Point, Bellambi). The 2 times are the
first sampling date after the closure of the outfalls and 2 yr after the closure, respec-
tively. The first date at Bellambi was different from that at the other outfalls
Fig. 4. Mean (+SE) number of species through
time in midshore habitats for each site at the
outfall (black column), and north (white col-
umn) and south (hashed column) reference lo-
cations. North Head and Malabar are the 2
experimental regions, i.e. with a decommis-
sioned outfall. The site closest to the point of
discharge at the outfalls is the left black col-
umn for each sampling time. No data were
collected at Bellambi in midshore habitats

Archambault et al.: Temporal variation in intertidal assemblages
Midshore areas
The nMDS ordinations comparing assem-
blages in the midshore at outfall and refer-
ence sites at North Head, Malabar and Potter
Point showed no evidence of difference at
any time (Fig. 5e to g). Symbols for all sites
were interspersed in July/August 1991, sug-
gesting no important impact of the effluent
on the composition of the midshore assem-
blages. The sites at the outfall and reference
locations were also interspersed in July/
August 1993 at North Head and Potter Point.
It does, however, seem that the assemblages
in reference locations in the 2 experimen-
tal regions where outfalls were decommis-
sioned changed through time (Fig. 5f,g). Fur-
thermore, at Malabar, the sites at the outfall
location diverged from the sites at the refer-
ence locations and became less dissimilar
from each other by July/August 1993.
DISCUSSION
Recovery of benthic assemblages at 2 out-
falls following decommissioning has been
examined and compared with natural (refer-
ence) locations and with 2 operational (con-
trol) outfalls over a period of 24 mo. Numbers
of species and percentage cover of green
ephemeral algae changed through time at
the decommissioned outfalls, but not at the
reference nor at the control locations.
Low-shore areas
At the beginning of the study, assemblages
at all 4 outfalls were different from those at
the respective reference locations, despite
there being great variability from site to site.
These results are similar to those reported
in other studies (Borowitzka 1972, Littler &
Murray 1975, May 1985, Brown et al. 1990,
Fairweather 1990, López Gappa et al. 1993).
In July/August 1993, 2 yr after the closure of
2 outfalls, a clear increase in the number of
species and a decrease in the percentage
cover of green algae were observed at the 2
experimental outfall locations. Furthermore,
the whole assemblage had changed. In con-
trast, assemblages, number of species and
percentage cover of green algae at the con-
trol outfalls did not change throughout the
59
Fig. 5. Non-metric multi-dimensional scaling (nMDS) ordinations of
assemblages of fauna and flora in quadrats (n = 5) for North Head (a,e),
Malabar (b,f), Potter Point (c,g) and Bellambi (d) comparing composition
of species at 3 low-shore sites (left graphs) and 3 midshore sites (right
graphs) at the outfall (3 circles) and at 2 reference locations (6 diamonds).
The empty symbols represent samples at July/August 1991 and the filled
symbols at July/August 1993. Quadrats at each site have not been dis-
tinguished; thus, n = 5 per site. Note that the sample at Bellambi in
October/November 1991 was used in the nMDS ordination because there
was no sample at July/August 1991 and no data for mid-intertidal habitats

Mar Ecol Prog Ser 222: 51– 62, 2001
study. The results from this study support the conclu-
sions of the studies mentioned above and provided fur-
ther evidence that sewage alters the structure of
assemblages in low intertidal areas.
One or only a few species of algae occupied large
areas of low-shore habitat near the outfall. Such exten-
sive mats of macroalgae (mainly green algae) associ-
ated with discharges of sewage have been described
(Borowitzka 1972, Littler & Murray 1975, May 1985,
Soulsby et al. 1985, Fairweather 1990, Bellgrove et al.
1997). In this study, the percentage cover of green
algae decreased dramatically at the 2 decommissioned
outfalls but stayed very large at the 2 control outfalls
and was small at the respective reference locations. In
contrast, Soulsby et al. (1985) found that Ulva and
Enteromorpha did not decrease in abundance follow-
ing the closure of a sewage discharge. They suggested
that the ambient nutrient regimen of the study area
(Langstone Harbour, UK) naturally exceeded the re-
quirements for growth of these macroalgae. This sug-
gestion was supported by Montgomery et al. (1985),
who showed that the nutrients in seawater entering
Langstone Harbour could support the observed stand-
ing crop of macroalgae without the presence of the
sewage outfall. In a long-term study, Smith et al. (1981)
showed that the biomass of Ulva decreased and be-
came less abundant after sewage was diverted at
Kaneohe Bay, Hawaii. These and our results suggest
that sewage outfalls provided the necessary amount of
nutrients to enable these green algae to dominate the
substratum. Many other models (e.g. the addition of
freshwater, the abundance of grazers, rates of grazing
by individual animals, etc.) must, however, be investi-
gated before claiming that nutrients are the only factor
controlling the abundance of green algae at sewage
outfalls.
Midshore areas
A different pattern was apparent at midshore habi-
tats compared with low-shore habitats. No single
effect was common to all outfalls. Similar numbers of
species occurred at reference and outfall locations at
all times (before and up to 2 yr after decommission-
ing) suggesting that the outfalls had little impact on
organisms in midshore areas. Furthermore, the whole
assemblages in outfall locations and their respective
reference locations did not differ before the closure
of the outfalls. There was no evidence that the mid-
shore assemblages and number of species at the out-
fall locations differed significantly from those at ref-
erence locations. Spatial and temporal patterns were
variable more among sites within any location than
between locations (reference versus outfall or refer-
ence versus reference). This finding is in contrast
with most of the studies (except Roberts & Scanes
2000) on the effects of sewage on subtidal and inter-
tidal assemblages in New South Wales, most of
which has shown a localised effect (May 1985, Fair-
weather 1990, Smith 1996). Underwood & Chapman
(1996) observed a similar pattern to the results de-
scribed here in subtidal habitats close to an outfall.
Their study and the present study began between
4and 6 mo after the closure of outfalls. Underwood
& Chapman (1996) suggested that organisms were
not stressed because wave action, tides and currents
continuously removed effluents. In contrast, Roberts
et al. (1998) described rapid changes (within 3 mo)
following the discharge of sewage effluent in a simi-
lar subtidal habitat.Perhaps midshore organisms
were not exposed as often or for as long as were
organisms lower on the shore to effluents because all
discharges were below low-tide levels. During high
tide, the sewage would be diluted. Data before the
outfalls were decommissioned indicate no persistent
influence of sewage on midshore assemblages (un-
publ. data)
Recovery
The present study defined recovery as conver-
gence of the various variables in affected locations to
the values in reference locations. From this study, the
overall conclusion is that recovery of assemblages
occurred in most areas of the low-shore at North
Head and Malabar outfalls following their decommis-
sioning. At the 2 control outfalls, no changes were
observed. In general, the detection of whether recov-
ery has occurred can be problematic (Fairweather
1993, McDonald & Erickson 1994, Chapman 1999,
Underwood & Chapman 1999, Underwood 2000)
with regard to the time-course involved and the
selection of the appropriate reference and control
locations. Furthermore, the level of acceptable recov-
ery has generally not been well defined. McDonald
& Erickson (1994) highlighted the advantages of
using bioequivalence to test whether recovery has
occurred between treated areas and reference areas,
but this procedure is not common in biological
sciences. Moreover, studies designed to detect the
influence of human activities must include replication
at relevant spatial and temporal scales. This can be
achieved by hierarchical sampling, with nested
designs such as that used in this study (Underwood
1991, 1992, 1993, 2000). Data need not be collected
before an impact, although it is better to be able to
do so (Green 1979, Underwood 1994). When it is
impossible to collect prior data, subsequent samples
60

Archambault et al.: Temporal variation in intertidal assemblages
must be compared with data collected at appropriate
reference locations (Bernstein & Zalinski 1983,
Underwood 1994). When using asymmetrical designs
as advocated by various authors (Underwood 1994,
Otway 1995b, Roberts et al. 1998), the number of
reference locations sampled must be considered. The
present study confirms the importance of making
comparisons between a contaminated location and
more than 1 reference area. Chapman (1999) sug-
gested that restoration (i.e., after closure of the out-
fall) cannot be properly assessed without multiple
control locations (here, outfalls that continued to
operate) and reference locations (ideally, natural
undisturbed areas). The use of 2 control outfalls in
this study increased the reliability of our finding that
the closure of 2 effluents resulted in recovery of
intertidal assemblages (Chapman 1999). This study
was designed to be able to test the hypothesis that
the number of species or abundances of algae would
recover after the closure of the outfalls and that such
changes would not be due to natural variability or
some other larger-scale influences.
It had been predicted that decommissioning 2 out-
falls would lead to recovery of the intertidal benthic
assemblages at these outfalls. The decommissioning
of the outfall led to a recovery (increase in number of
species similar to the reference locations and a
decrease in abundance of green algae), except at sites
where fishermen removed the algae. The primary aim
of the closure of the outfalls was to reduce the effluent
into beaches for human health issues, but it was also
important to test the stated predictions about the
recovery of intertidal organisms. If the aim of man-
agerial decisions is recovery, further action is needed
to manage fishermen. This requires careful designs to
show what happens following changed management
of an area. The design to follow the recovery in our
study was replicated at 2 decommissioned outfalls
and control areas (operational outfalls), each with
replicated reference areas and assessments at a hier-
archy of spatial scales to allow valid conclusions to be
made. In general, spatial and temporal variations
must be estimated properly to discern the environ-
mental signal that would identify the real response to
a change in management. Otherwise confusion will
continue to reign.
Acknowledgements. We thank V. Mathews for help with the
graphics, W. Green for help with analyses, Drs C. W. McKind-
sey, M. G. Hoskin and M. G. Chapman, and 4 anonymous
referees for constructive advice on earlier drafts of the manu-
script. P.A. was supported by an NSERC (Natural Sciences
and Engineering Research Council of Canada) Postdoctoral
Fellowship. The preparation of the paper was funded by the
Australian Research Council through the Centre for Research
on Ecological Impacts of Coastal Cities.
LITERATURE CITED
Ajani PA, Roberts DE, Smith AK, Krogh M (1999) The effect of
sewage on two bioindicators at Port Stephens, New South
Wales, Australia. Ecotoxicology 8:253–267
Axelrad DM, Poore GCB, Arnott GH, Bauld J, Brown V,
Edwards RRC, Hickman NJ (1981) The effects of treated
sewage discharge on the biota of Port Phillip Bay, Victoria,
Australia. In: Neilson BJ, Cronin LE (eds) Estuaries and
nutrients. Humana Press, Clifton, p 279–306
Banwell K (1996) Structure of intertidal rocky shore assem-
blages following the removal of sewage. MSc thesis, Uni-
versity of Sydney
Bellgrove A, Clayton MN, Quinn GP (1997) Effects of secon-
darily treated sewage effluent on intertidal macroalgae
recruitment processes. Mar Freshw Res 48:137–146
Bernstein BB, Zalinski J (1983) An optimal sampling design
and power tests for environmental biologist. J Environ
Manage 65:35– 43
Borowitzka MA (1972) Intertidal algal species diversity and
the effect of pollution. Aust J Mar Freshw Res 23:73–84
Brosnan DM, Crumrine LL (1994) Effects of human trampling
on marine rocky shore communities. J Exp Mar Biol Ecol
177:79– 97
Brown VB, Davies SA, Synnot RN (1990) Long-term monitor-
ing of the effects of treated sewage effluent on the inter-
tidal macroalgal community near Cape Schanck, Victoria,
Australia. Bot Mar 33:85– 98
Calcagno JA, Gappa JL, Tablado A (1997) Growth and pro-
duction of the barnacle Balanus amphitrite in an intertidal
area affected by sewage pollution. J Crustac Biol 17:
417–423
Calcagno JA, Gappa JL, Tablado A (1998) Population dynam-
ics of the barnacle Balanus amphitrite in an intertidal area
affected by sewage pollution. J Crustac Biol 18:128–137
CDM (1989) Review of Sydney’s beach protection programme.
Camp, Dresser & McKee International Inc, Cambridge, MA
Chapman MG (1999) Improving sampling designs for mea-
suring restoration in aquatic habitats. J Aquat Ecosyst
Stress Recov 6:235–251
Chapman MG, Underwood AJ, Skilleter GA (1995) Vari-
ability at different spatial scales between a subtidal as-
semblage exposed to the discharge of sewage and two
control assemblages. J Exp Mar Biol Ecol 189:103–122
Clarke KR (1993) Non-parametric multivariate analyses of
changes in community structure. Aust J Ecol 18:117–143
Clayton MN, King RJ (1990) Biology of marine plants. Long-
man Cheshire, Melbourne
Eberhert LL, Thomas JM (1991) Designing environmental
field studies. Ecol Monogr 61:53– 73
Fairweather PG (1990) Sewage and the biota on seashores:
assessment of impact in relation to natural variability.
Environ Monit Assess 14:197–210
Fairweather PG (1993) Links between ecology and eco-
philosophy, ethics and the requirements of environmental
management. Aust J Ecol 18:3–19
Glasby TM (1997) Analysing data from post-impact studies
using asymmetrical analyses of variance: a case study of
epibiota on marinas. Aust J Ecol 22:448– 459
Green RH (1979) Sampling design and statistical methods for
environmental biologists. Wiley, Chichester
Kingsford MJ, Underwood AJ, Kennelly SJ (1991) Humans as
predators on rocky reefs in New South Wales, Australia.
Mar Ecol Prog Ser 72:1–14
Littler MM, Murray SN (1975) Impact of sewage on the distri-
bution, abundance and community structure of rocky
intertidal macro-organisms. Mar Biol 30:277–291
61

Mar Ecol Prog Ser 222: 51– 62, 2001
López Gappa JJ, Tablado A, Magaldi NH (1990) Influence of
sewage pollution on a rocky intertidal community domi-
nated by the mytilid Brachidontes rodriguezi. Mar Ecol
Prog Ser 63:163–175
López Gappa JJ, Tablado A, Magaldi NH (1993) Seasonal
changes in an intertidal community affected by sewage
pollution. Environ Pollut 82:157–165
López-Rodriguez MC, Barbara I, Perez-Cirera JL (1999)
Effects of pollution on Fucus vesiculosus communities on
the Northwest Iberian Atlantic coast. Ophelia 51:129–141
May V (1985) Observations on algal floras close to two sewer-
age outlets. Cunninghamia 1:385– 394
McDonald LL, Erickson WP (1994) Testing for bioequivalence
in field studies: has a disturbed sites been adequately
reclaimed? In: Fletcher DJ, Manly BFJ (eds) Statistics in
ecology and environmental monitoring. University of
Otago Press, Dunedin, p 183–197
Montgomery HAC, Soulsby PG, Hart IC, Wright SL (1985)
Investigation of a eutrophic tidal basin II— Nutrients and
environmental aspects. Mar Environ Res 15 285–302
Otway NM (1995a) Assessing impacts of deepwater sewage
disposal: a case study from New South Wales, Australia.
Mar Pollut Bull 31:347–354
Otway NM (1995b) Environmental problems due to the dis-
posal of waste. 2. Disposal of waste. In: Underwood AJ,
Chapman MG (eds) Coastal marine ecology of temperate
Australia. New South Wale University Press, Sydney,
p 304– 310
Pearson TH, Rosenberg R (1978) Macrobenthic succession in
relation to organic enrichment and pollution of the marine
environment. Annu Rev Oceanogr Mar Biol 16:229–311
Roberts DE, Scanes PR (2000) Spatial patterns in the macro-
benthic assemblages inhabiting kelp (Ecklonia radiata)
forests exposed to sewage effluent. Aust J Ecotoxicol 5:
89–102
Roberts DE, Smith A, Ajani P, Davis AR (1998) Rapid changes
in encrusting marine assemblages exposed to anthropo-
genic point-source pollution: a ‘Beyond BACI’ approach.
Mar Ecol Prog Ser 163:213–224
Smith AK, Ajani PA, Roberts DE (1999) Spatial and temporal
variation in fish assemblages exposed to sewage and
implications for management. Mar Environ Res 47:
241–260
Smith SDA (1996) The effects of domestic sewage effluent on
marine communities at Coffs Harbour, New South Wales,
Australia. Mar Pollut Bull 33:7–12
Smith SV, Kimmerer WJ, Laws EA, Brock RE, Walsh TW
(1981) Kaneohe Bay sewage diversion experiment: per-
spectives on ecosystem responses to nutritional perturba-
tion. Pac Sci 35:279– 395
Soulsby PG, Lowthion D, Houston MA, Montgomery HAC
(1985) The role of sewage effluent in the accumulation of
macroalgal mats on intertidal mudflats in two basins in
Southern England. Neth J Sea Res 19:257–263
Taylor LA, Chapman PM, Miller RA, Pym RV (1998) The
effects of untreated municipal sewage discharge to
the marine environment off Victoria, British Columbia,
Canada. Water Sci Technol 38:285–292
Underwood AJ (1991) Beyond BACI: experimental designs
for detecting human environmental impacts on temporal
variations in natural populations. Aust J Mar Freshw Res
42:569– 587
Underwood AJ (1992) Beyond BACI: the detection of environ-
mental impacts on populations in the real, but variable,
world. J Exp Mar Biol Ecol 161:145–178
Underwood AJ (1993) The mechanics of spatially replicated
sampling programmes to detect environmental impacts in
a variable world. Aust J Ecol 18:99–116
Underwood AJ (1994) On beyond BACI: sampling designs
that might reliably detect environmental disturbances.
Ecol Appl 4:3–15
Underwood AJ (1996) Detection, interpretation, prediction
and management of environmental disturbances: some
roles for experimental marine ecology. J Exp Mar Biol Ecol
200:1–27
Underwood AJ (2000) Trying to detect impacts in marine
habitats: comparisons with suitable reference areas. In:
Sparks T (ed) Statistics in ecotoxicology. John Wiley &
Sons, Toronto, p 279– 308
Underwood AJ, Chapman MG (1996) Subtidal assemblages
on rocky reefs at a cliff-face sewage outfall (North Head,
Sydney, Australia): what happened when the outfall was
turned off? Mar Pollut Bull 33:7–12
Underwood AJ, Chapman MG (1999) The role of ecology in
coastal zone management: perspectives from South-East
Australia. In: Salomons W, Turner RK, de Lacerda LD,
Ramachandran S (eds) Perspectives on integrated coastal
zone management. Springer, Berlin, p 99–128
Underwood AJ, Kennelly SJ (1990) Pilot studies for the
designs of surveys of human disurbance of intertidal habi-
tats in New South Wales. Aust J Mar Freshw Res 41:
165–173
62
Editorial responsibility: Otto Kinne (Editor),
Oldendorf/Luhe, Germany
Submitted: September 25, 2000; Accepted: March 22, 2001
Proofs received from author(s): October 18, 2001