Temporal Variation in the Diversity and Cover of Sessile Species in

PII: S0025-326X(00)00201-0
Marine Pollution Bulletin Vol. 42, No. 7, pp. 554±568, 2001
Ó 2001 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
0025-326X/01 $ - see front matter
Temporal Variation in the Diversity and
Cover of Sessile Species in Rocky
Intertidal Communities A€ected by
Copper Mine Tailings in Northern Chile
Departamento de Ecologõa, Facultad de Ciencias Biol
ogicas, Ponti®cia Universidad Cat
olica de Chile, Alameda 340,
C.P. 6513677, Santiago, Chile
Several coastal rocky shores in the northern Chile have
been a€ected by the discharges of copper mine tailings. In
spite of this, the temporal and spatial variation on the
diversity and composition of their intertidal benthic communities has scarcely been studied. The objectives of the
present study were to analyse and to compare quantitatively the temporal variation on the diversity, cover and
composition of sessile species in rocky intertidal benthic
communities of the northern Chilean coast, in relation to
the presence of copper mine tailings. The results show that
the drastic reduction on the sessile species diversity and
the monopolization of the substrate exerted by the green
algae Enteromorpha compressa, are common and permanent features of the intertidal rocky shores a€ected by
copper mine tailings. Such spatial (between sites) and
temporal (seasonal) variation of these changes has been
associated with the relative concentrations of trace metals
and inorganic particles of the mining wastes. Our results
suggest that the mechanical e€ects of resuspended and
settling tailings are a more likely cause. Ó 2001 Elsevier
Science Ltd. All rights reserved.
Keywords: copper; Chile; Enteromorpha compressa;
rocky intertidal; ®ne ground wastes; diversity.
Copper mining has historically been the most important
economic activity in the northern Chilean region. The
major copper mines and deposits in this region are found
on the western slopes of the Andes Mountains, but minor mines and some processing plants are located on the
coast (Castilla, 1983). As a consequence, for a long time,
several sites of the coast had been a€ected by the dis-
*Corresponding author. Fax: +56-2-6862610.
E-mail address: [email protected] (J.M. Fari~
charges of copper mine tailings (Miethke et al., 1992;
Vasquez et al., 1999). In spite of this, the analysis of the
e€ects of copper wastes on marine intertidal communities has been done mainly at one area a€ected by the El
Salvador Copper Mine (Fig. 1) discharges (Castilla and
Nealler, 1978; Castilla, 1983; Castilla, 1995; Correa et al.,
1999, 2000; Pasko€ and Petiot, 1990).
The occurrence of copper mine tailings in the seawater
drastically change the composition and spatial distributions of the rocky intertidal communities, particularly
on sessile species. One of the most dramatic changes
reported has been a decrease in the number of sessile
species and the monopolization of the rocky substrate
by the green ephemeral alga Enteromorpha compressa
(Castilla, 1995). Another result has been that in addition
to the trace metals, copper mine tailings have a physical
component (ground wastes) which modify drastically
the coastal landscape (Castilla, 1983; Ellis, 1987; Pasko€
and Petiot, 1990). Nevertheless, intertidal communities'
modi®cations have been explained solely with reference
to the high concentration of trace metals (Correa et al.,
1999, 2000), neglecting the possible role of the ground
wastes as abrasive agents.
So far, there have been no attempts to evaluate the
temporal (i.e., seasonal) and/or spatial (i.e., between
sites) variability in the changes exerted by copper mine
tailings on the intertidal diversity and species composition. The aims of this work are to analyse the patterns of
diversity on rocky intertidal communities of several
northern Chilean coastal sites, with regard to the presence of copper mine wastes (trace metals and ®ne
ground wastes), and to evaluate quantitatively their
temporal and spatial variability.
Materials and Methods
Study sites
The study was carried out on the rocky intertidal
shores of four sites located on the northern Chilean
Volume 42/Number 7/July 2001
Fig. 1 Location of the discharge and reference sampled (*) and
previously described (Castilla, 1995) sites.
coast: Enami, Las Conchas, Bandurrias and Santo
Domingo (Fig. 1). Since the 1960s both Enami and
Santo Domingo have been a€ected by the discharges of
copper mine tailings derived from processing plants located close to the shore. At Enami, the e‚uents are
discharged southwards and have formed a tailing bed of
approximately 2 km in length, along the southern side of
the sampling site. At Santo Domingo, the waste discharges move northwards and have formed a tailing bed
of approximately 10 km in length on the northern side of
the sampling site. Las Conchas and Bandurrias, located
about 15 km north of Enami and 19 km south of Santo
Domingo, respectively, were considered to be non-contaminated sites, and used as reference sites.
Seawater analysis
Based on the results of previous studies, the main
trace metals, Cu, Zn, and Cd, reported as abundant on
Chilean copper mine orebodies and tailings e‚uents
were analysed (Castilla and Nealler, 1978; Veermer and
Castilla, 1991). To evaluate the levels of these trace
metals, six 1-l samples of seawater were taken from the
shore of each site in July 1997, January 1998 and June of
1998. In the case of Enami and Santo Domingo, the
samples were taken 1 km from the waste discharge
points. Water samples were treated and analysed as indicated by Correa et al. (1996). At CESMEC Inc. lab-
oratories, the samples were ®ltered through 0.45 lm
Whatman GF/F ®lters and ®xed with nitric acid (Merck,
supra pur). Dissolved trace metals in the samples were
quanti®ed by potentiometric stripping analysis in stationary solution, using a computarized Radiometer ISS
820 analyser.
To analyse the concentration of ®ne ground wastes,
we used standard methods designed to measure the
seawater concentrations of particulate inorganic (PIM)
and organic matter (POM) (Bustamante, 1994; Rodriguez, 1999). During each sampling date, four samples of
0.5 l were ®ltered using Whatman GF/F glass ®ber ®lters. The ®lters, with a pore size of 0.45 lm, retain organic and inorganic particles of more than 0.7 lm in
size. Concentrations of PIM and POM were determined
by the di€erences in weight between dried versus burned
(at 500°C) ®lters. Prior to the drying and burning process the organic and inorganic particles caught in each
®lter were visually inspected using a stereomicroscope
(Nikon SMZ-2T). This analysis allowed us to identify
the possible sources for inorganic (i.e., sand or minederived sediments) and organic (i.e., small pieces of intertidal algae or planktonic organisms) particles. In the
case of inorganic particles the di€erent colours and
textures of the sand (white in colour and smooth) and
mine-derived (brown or black in colour and rough)
particles, were used for their recognition.
Marine Pollution Bulletin
Sampling design
The sampling of the rocky intertidal at each site was
carried out, every two months, from July 1997 to May
1998 (six sampling dates). A series of 20 plots of
0:25 0:25 m were randomly placed parallel to the
shoreline, between the high and low intertidal levels.
Plot randomization was developed using an equally 0.25
m spaced grid, which covered all the intertidal study
area of each site. Each point of the grid was numbered
and the random position of the plots in the grid was
chosen taking the sequence numbers from a random
numbers table and ranking them from the smallest to
the largest (Kuehl, 2000). The levels were determined a
priori using the high and low water marks predicted by
tide tables, during days of similar tidal and sea conditions. Each plot was divided into a grid of 100 equally
spaced intersection points. The sessile algae or ®lter
feeding species occurring underneath each point were
identi®ed to the lowest possible taxonomic level. The
total cover of each species in the plot was obtained directly by the sum of their intersection points (Castilla,
Data analysis
The results of the seawater analysis (trace metals,
PIM and POM concentrations) were contrasted using
two-factor analysis of variance (ANOVA). The variations in the mean numbers and the mean covers of
species were also contrasted using two-factor ANOVA
(Winer, 1971). In both cases, sites and dates were considered as random factors because no a priori hypothesis
was speci®ed about them, and because their levels represent one of the several potential combinations of sites
and dates over which the study could be done (Underwood, 1997). After each analysis Tukey (HSD) multiple
comparisons test was used (Day and Quinn, 1989). The
normality of the data was tested using graphical procedures, and when appropriate (i.e., for cover), standard
procedures of data transformations (i.e., Arc-sin) were
used. Homogeneity of variance and independence of the
data were veri®ed using Levene's and Durbing±Watson
tests, respectively (Wilkinson et al., 1996).
Spatial dominance and temporal incidence of the species
To identify the most representative species occurring
at each locality, we analysed the relationship between
their spatial dominance and temporal incidence. We
plotted the mean percentage of cover vs. the mean
number of plots, in which each species was registered
during the entire period of study, choosing in each case
the species with signi®cant high cover and/or spatial
incidence. Signi®cance (at a < 0:05) was calculated from
the frequency distribution of the means generated using
a bootstrap procedure, in which the original matrix
containing both variables was re-sampled 1000 times
(Manly, 1991). The patterns of spatial dominance of the
sessile species at each date and sites were compared
using rank/abundance plots (Magurran, 1988), consid556
ering the logged mean percentage of cover as an estimate
of dominance (Clarke, 1990).
Similarity on the species composition
The Jaccard coecient of similarity (Jaccard, 1901;
Sneath, 1957) was used to analyse the similarity and
variability in the species composition observed at each
site. This coecient, based on the presence ± absence of
the species, has been widely used in ecology, and highly
recommended for the analysis of species composition
similarities (Hubalek, 1982). The patterns of similarity
on the species composition were contrasted using one
cluster diagram of the Jaccard coecients calculated for
each combination of site and date. The cluster was
constructed using un-weighted average linkage methods
(Wilkinson et al., 1996) and signi®cant level of similarity
was estimated using the distribution of the Jaccard coecients, calculated after bootstraping 1000 times the
species composition matrix (Clarke, 1993).
To analyse the temporal variability of species composition at each site, non-metric multidimentional scaling analysis (MDS) on the calculated matrix of Jaccard
coecients was used. The number of dimensions to scale
(2) was chosen considering the lowest value of the
Kruskal MDS stress statistic calculated for 1±5 dimensions. The number of iterations (50), was selected considering the lowest value of the minimized Kruskal
MDS stress statistics …ˆ 0:068†. The result of the MDS
analysis was corroborated by the Sheppard diagram,
which showed a signi®cant monotonic decreasing relation between the observed and estimated similarities
(Clarke, 1993; Wilkinson et al., 1996).
Using a cluster diagram of the Jaccard coecients, we
compared the patterns of similarity on the species
composition of the sites used in this study with those
used in previous studies (Castilla, 1995). The comparison was done considering the species compositions reported on the rocky intertidal areas of four sites located
near the city of Cha~
naral: Zenteno, La Lancha, Palito
and Pan de Az
ucar. La Lancha and Palito have been
impacted by tailings waste from the El Salvador copper
mine since the mid 1970s, while Zenteno and Pan de
ucar have been considered non-contaminated
(Castilla, 1995).
Seawater analysis
Di€erent concentrations of copper (Fig. 2(a)) were
observed in the seawater at each of the sites
(F0:05;3 ˆ 98:13, p < 0:0001). The sequence of signi®cant
di€erences was Santo Domingo …25:78 1:98 lg l 1 † >
Enami …12:23 1:90 lg l 1 † > Bandurrias …3:08 1:04
lg l 1 † and Las Conchas …1:53 0:36 lg l 1 †. Bandurrias and Las Conchas were not signi®cantly di€erent,
and the pattern was persistent over time (F0:05;2 ˆ 0:14,
p ˆ 0:87).
Volume 42/Number 7/July 2001
Fig. 2 Temporal variation of trace metal concentrations
…mean 1 EE† in the seawater of the study sites. (a) copper,
(b) zinc and (c) cadmium.
In spite of the low concentrations of zinc (Fig. 2(b)),
there were signi®cant di€erences between sites
(F0:05;3 ˆ 29:08, p ˆ 0:0006†. The sequence of di€erences
was Enami …1:98 0:28 lg l 1 † > Las Conchas …0:45 0:18 lg l 1 † > Bandurrias (less than 0:10 0:0 lg l 1 )
and Santo Domingo (less than 0:10 0:0 lg=l). Santo
Domingo and Bandurrias were not signi®cantly di€erent
and the pattern was persistent over time (F0:05;2 ˆ 0:55,
p ˆ 0:61).
For cadmium (Fig. 2(c)) there were signi®cant di€erences between sites (F0:05;3 ˆ 432:73, p < 0:0001). The
concentration in Enami …3:76 0:62 lg l 1 † was higher
than that observed at Santo Domingo …1:94 0:07 lg l 1 †, Las Conchas …1:56 0:19 lg l 1 † or Bandurrias …1:02 0:08 lg l 1 †. Santo Domingo, Las Con-
chas and Bandurrias did not show signi®cant
di€erences, and the pattern was persistent with time
(F0:05;2 ˆ 1:18, p ˆ 0:368).
High levels of Particulate Inorganic Matter in the
seawater (Fig. 3(a)) were observed at Enami during
September±October of 1997 and March±April of 1998.
This site …221:91 212:12 mg l 1 † showed signi®cant
(F0:05;3 ˆ 3:78, p ˆ 0:033) di€erences with Santo Domingo …47:79 10:99 mg l 1 †, Las Conchas …32:16 7:68 lg l 1 † and Bandurrias …38:50 9:49 lg l 1 †. The
last three sites did not show signi®cant di€erences
between them.
The concentration of POM in the seawater (Fig. 3(b))
did not show di€erences between sites …F0:05;3 ˆ 2:13,
p ˆ 0:14) and time (F0:05;5 ˆ 1:38, p ˆ 0:28). However,
Marine Pollution Bulletin
Fig. 3 Temporal variation …mean 1 EE† of the (a) PIM and (b)
POM concentrations in the seawater of the sites.
due to the high concentrations observed at Enami during September±October of 1997, the interaction between
the factors of site and time was signi®cant (F0:05;15 ˆ
30:08, p < 0:0001).
Species composition and diversity patterns
A total of 31 taxa were identi®ed in the 480 sampled
plots (Table 1). Bandurrias and Las Conchas showed the
higher number of taxa (26 and 29, respectively). Enami
and Santo Domingo showed 10 and 8, respectively. The
di€erences in the number of taxa observed were due to
the number of algae and to the number of ®lter feeding
The mean number of species (Fig. 4(a)) showed differences between sites and time (Table 2(a)). Between
sites, Enami …2:70 2:06 spp† and Santo Domingo
…3:22 1:70 spp† had the lowest number of species and
did not show signi®cant di€erences between them. The
highest mean numbers of species observed in Bandurrias
…13:24 2:77 spp† was signi®cantly higher than the one
observed in Las Conchas …10:85 2:08 spp†. Over time,
the number of species observed during the periods of
July±August …8:85 5:54 spp† and September±October
of 1997 …9:21 5:95 spp† were signi®cantly higher than
the observed in other dates (which ¯uctuated between
6.62±7.23 spp and did not show signi®cant di€erences).
This was due to the high mean number of species observed
in Las Conchas and Bandurrias during these two periods.
The mean number of algal species (Fig. 4(b)) followed
the variations of the mean number of total species (see
Fig. 4(a)). There were signi®cant di€erences
(F0:05;3 ˆ 129:55, p < 0:0001) between the contaminated
sites, Enami and Santo Domingo, and non-contaminated ones, Las Conchas and Bandurrias. Enami and
Santo Domingo had the lower mean number of species
(2:42 1:66 and 3:06 1:31, respectively) and there
were no di€erences between them. Las Conchas and
Bandurrias showed the highest mean numbers
(9:18 2:77 spp and 10:25 2:07 spp, respectively) and
there were no signi®cant di€erences between them. Over
time, there were signi®cant di€erences (F0:05;5 ˆ 4:14,
p ˆ 0:014). For Las Conchas, the period September±
October of 1997 was signi®cantly higher in algal species
…12:05 0:95† than the other periods.
Volume 42/Number 7/July 2001
List of the sessile species observed (X) at each site during the entire
period of study.
Enteromorpha compressa
Enteromorpha prolifera
Ulva lactuca
Chaetomorpha linum
Ectocarpus confervoides
Ralfsia expansa
Colpomenia phaeodactyla
Colpomenia sinuosa
Petalonia fascia
Scytosiphon lomentaria
Glosophora kunthii
Polysiphonia paniculata
Porphyra columbina
Gelidium chilense
Gelidium lingulatum
Corallina ocinalis
Ceramium rubrum
Hildenbrandtia lecannellieri
Lithothamnion sp
Filter feeding species
Anthothoe chilensis
Phymactis clematis
Notochthamalus scabrosus
Jehlius cirratus
Balanus laevis
Austromegabalanus psittacus
Semimytilus algosus
Perumytilus purpuratus
Pyura chilensis
Bandurrias Santo
The pattern of variations in the mean number of ®lter
feeding species (Fig. 4(c)), showed some di€erences between sites …F0:05;3 ˆ 48:88, p < 0:0001†. The sequence of
di€erence was: Bandurrias …2:99 1:07† > Las Conchas
…1:67 0:86† > Enami …0:28 0:58† and Santo Domingo …0:16 0:48 spp†. Enami and Santo Domingo
did not show signi®cant di€erences.
Percentage of cover of the primary space
The mean cover of species (Fig. 5(a)) did not show
signi®cant di€erences between sites or over time (Table
2(b)). However, due to the wide variability observed in
Enami and Santo Domingo the interaction between sites
and date was signi®cant (Table 2(b)). The mean cover of
species observed in Enami during November±December
of 1997 …56:13 9:21%† and in Santo Domingo during
May±June of 1998 …50:85 8:12%†, were signi®cantly
lower than all the other values registered (which ¯uctuated between 90% and 70%).
For the mean cover of algal species (Fig. 5(b)), no
signi®cant di€erences were found between sites
(F0:05;3 ˆ 1:516, p ˆ 0:251) and over time (F0:05;5 ˆ 0:608,
p ˆ 0:695). Nevertheless, due to the important varia-
tions observed within each of the four sites, the interaction between sites and date was signi®cant
(F0:05;15 ˆ 20:845, p < 0:001). In Enami, the mean cover
observed during July±August …77:78 15:01%† decreased through to November±December of 1997
…55:90 9:08%† and then it was followed by a recovery.
For Las Conchas, from July±August of 1997 to January±February of 1998, there were no signi®cant di€erences (with values near 77%). However, an important
increase was observed from January±February of 1998
until May±June of 1998 …90:57 5:42%†. For Bandurrias a decreasing trend was observed from September±
October of 1997 …75:08 6:59%† to March±April of
1998 …56:98 5:73%†. The trend was followed by a recovery. For Santo Domingo, from July±August of 1997
to March±April of 1998, there were no signi®cant differences (with values near 77%). However, a signi®cant
decrease was observed during May±June of 1998
…50:82 8:09%†. As a result of all these changes the
mean values observed at Enami during November±December of 1997, at Bandurrias from January to April of
1998, as well as at Santo Domingo and Las Conchas
during May±June of 1998 were signi®cantly lower than
the other values (see Fig. 5(b)).
The mean cover of the ®lter feeding species
(Fig. 5(c)) showed signi®cant di€erences between sites
(F0:05;3 ˆ 27:57, p < 0:0001) but not over time
(F0:05;5 ˆ 0:31, p ˆ 0:899). Enami and Santo Domingo
showed the lowest mean cover (0:15 0:44% and
0:02 0:06%, respectively) but there was no signi®cant
di€erence between them; while the highest covers of
Bandurrias …16:92 8:66%† and Las Conchas
…5:69 5:64%† were signi®cantly di€erent. Over time
within each site, Enami and Santo Domingo showed
no signi®cant di€erences. Important variations were
observed at Bandurrias and Las Conchas. Due to that,
the interaction between site and time was signi®cant
(F0:05;15 ˆ 13:90, p < 0:0001). For Las Conchas, from
July±December of 1997 the mean cover increased,
from 4:82 2:88% to 12:52 8:51%, and then decreased until May±June of 1998 …2:02 2:14%†. For
Bandurrias, the most important variation was an increase in the mean cover during January to April of
1998, from 10:90 3:51 to 22:33 7:39%. Due to this,
the mean cover of the ®lter feeding species during the
period of January to April of 1998 were signi®cantly
higher than the values observed in the other sites and
Spatial dominance and temporal incidence of the species
The spatial dominance and incidence of the species at
each site showed important di€erences (Fig. 6). In both
Enami and Santo Domingo the green algae E. compressa
and `Bare Rock' were `items' with the highest dominance and incidence. The main di€erence between these
two sites was that at Santo Domingo the red encrusting
alga Hildenbrandtia lecannellieri was important in terms
of incidence. For Las Conchas, the green alga Ulva
Marine Pollution Bulletin
Fig. 4 Temporal variation …mean 1 EE† on (a) the number of
sessile species, (b) the number of algae species and (c) the
number of ®lter feeding species, at the study sites.
Results of the 2-way ANOVA on: (a) the mean number and (b) the mean cover of sessile species, observed during the study period at each site.
Site date
Site date
Volume 42/Number 7/July 2001
Fig. 5 Temporal variation …mean 1 EE† on (a) the covers of sessile
species, (b) the covers of algae species and (c) the covers of
®lter feeding species at the study sites.
lactuca and `Bare Rock' were important in terms of
incidence and dominance, while the brown algae Ectocarpus confervoides and H. lecannellieri were important
in terms of incidence. For Bandurrias, the `Bare Rock'
`item' showed high levels of dominance and incidence;
U. lactuca and H. lecannellieri showed high levels of
incidence and the barnacle Jehlius cirratus, the red algae
Gelidium chilense and E. confervoides were important in
terms of dominance.
The dominance curve showed important di€erences
between sites. At Enami and Santo Domingo the
rocky substrate was dominated by one or two `items'
(E. compressa and/or `Bare Rock'). In Las Conchas
and Bandurrias, there were no dominant species and
the substrate was shared by several species. These
di€erences were consistent during the study period
(Fig. 7).
Similarity on the species composition
Two main groups (Fig. 8), with a high and signi®cant
level of similarity of their species composition, were
separated in the cluster analysis: Enami ± Santo Domingo and Las Conchas ± Bandurrias. The subgroups
separated within each main group showed a high level of
similarity in the species compositions within each site.
The diagram generated by the MDS analysis (Fig. 9),
separated the group of Enami-Santo Domingo from the
group of Las Conchas±Bandurrias. Enami was the site
with the highest temporal variations. In the case of Las
Conchas, Bandurrias and Santo Domingo, the temporal
variations were restricted to a speci®c region of the diagram and followed a closed trend with cyclical tendencies.
The species compositions of the contaminated and
non-contaminated sites analysed in our study were in
Marine Pollution Bulletin
Fig. 6 Spatial representation (% of cover, mean 1 EE) vs. Temporal incidence (mean no. of plots 1 EE) of sessile species at
the study sites. The grey lines represent the limits at the 95%
of the means.
agreement with the results of Castilla (1995). The cluster
analysis (Fig. 10) showed a signi®cant level of similarity
between the non-contaminated sites Las Conchas,
Bandurrias, Zenteno and Pan de Az
ucar, and between
the contaminated, Enami, Santo Domingo, La Lancha
and Palito. In spite of their dissimilarity with the noncontaminated sites, the species composition patterns of
the contaminated sites analysed by Castilla (1995) were
di€erent to the patterns observed in Santo Domingo and
Seawater analysis
The concentrations of trace metals and PIM in the
seawater (Figs. 2 and 3) were in agreement with the
subjective status of contaminated sites for Enami and
Santo Domingo, and non-contaminated sites for Bandurrias and Las Conchas.
Copper was the most important trace metal and their
concentrations in both Enami and Santo Domingo
were in the ranges reported in other copper contaminated coastal areas of Chile (Correa et al., 1999) and of
the world (Bryan and Langston, 1992). The concentrations observed in Las Conchas and Bandurrias were
in the range of those observed in other non-contaminated sites of Chile (Castilla, 1983; Vasquez and Guerra, 1996) and of the world (Lewis, 1995). For zinc and
cadmium, the concentrations observed in Enami and at
Las Conchas, Bandurrias and Santo Domingo were in
the range of other contaminated and non-contaminated
sites respectively (Correa, 1999; Knauer and Martin,
Both Enami and Santo Domingo presented tailing
beds on their shores (Ingenierõa y Geotecnica, 1990).
However, high concentrations of PIM were only detected in Enami. The di€erence could be due to the
particle size of the wastes occurring at each site. Based
on stereomicroscope visual inspection, we found that
sampled PIM at Santo Domingo was ®ner (less than 60
lm in average) than at Enami (more than 500 lm in
average). Considering that the ®lters of both sites were
saturated with PIM, it is probable that the method utilized (based on the weight di€erences) does not re¯ect
this pattern. The di€erences in size of the particles
sampled at Enami and Santo Domingo could be related
with the mining extraction process used at each site. At
Enami copper is puri®ed through lixiviation and at
Santo Domingo by ¯otation (Cortez et al., 1996). Flotation produces ®ner particles (ranging between 50±100
lm) than lixiviation (ranging between 150±1000 lm). In
Volume 42/Number 7/July 2001
Fig. 7 Dominance curves based on the mean cover of the sessile
species occurring at each site, during the study period.
¯otation the mined rock is ground at least three times
before the extraction, while in lixiviation the rock is
ground only once (Biswas and Davenport, 1980).
The high temporal variability of PIM concentration
observed at Enami may be related to the occurrence of
storms. As pointed out by Ellis (1987),in mining contaminated coastal waters, during windstorm a resuspension of the particles from tailing beds occurs. Hence
PIM concentrations are highly dependent on the local
oceanographic and atmospheric conditions. In the case
of Enami, and for the northern Chilean coast, during
September of 1997 and April of 1998, at least two important storms associated with the event of El Ni~
Southern Oscillation (ENSO) of 1997 occurred (Glynn,
1988; Camus and Andrade, 1999).
For POM, the high concentrations found at Enami
during September±October of 1997 may also be related
to the resuspension of ground wastes. Visual inspection
showed that POM was composed of small pieces of
green alga which perhaps were detached from the rocky
intertidal zone by the abrasive e€ect of the resuspended
PIM. In fact, after this period a signi®cant decrease on
the mean cover of algae was observed at Enami (Fig.
5(b)). The high concentrations of PIM occurring in
Enami during the period of March±April of 1998 were
not associated with a high concentration of POM.
During this period no algae particles were observed in
the collected POM and moreover, no decreases of the
mean cover of intertidal algae occurred at this site.
Another aspect related with POM, is the di€erence in
the copper concentration between Enami and Santo
Domingo. If the high concentration of POM observed at
Enami provided more ligands for the complexation of
copper (Sunda and Lewis, 1978), then it is possible that
this locality showed a lower concentration of the dissolved metal in the water than Santo Domingo. The
relationship between copper and POM must be analysed
in future ecotoxicological studies.
Marine Pollution Bulletin
Fig. 8 Cluster analysis of the species composition similarity of each
site during the study period. The key for the sites is:
E ˆ Enami, SD ˆ Santo Domingo, LC ˆ Las Conchas and,
B ˆ Bandurrias. The dashed line represents the limit of similarity at a < 0:05.
Species composition
At Enami and Santo Domingo, we observed lower
numbers and di€erent compositions of sessile species
than at Las Conchas and Bandurrias. The most important di€erences in the species composition were related to the absence at the contaminated sites of
®lamentous Porphyra columbiana, and G. chilense and
encrusting red algae C. ocinalis, as well as the absence
of ®lter feeding species such as Jehlius cirratus, Notochthamalus scabrosus and Semimytilus algosus.
Several factors may explain these di€erences. Correa
et al. (2000) proposed three hypotheses for the decreased diversity of algal species at one site of the
northern Chilean coast a€ected by copper mine
1. High copper concentrations in the seawater prevent
the growth of red and brown algae. Nevertheless,
the result of their work showed that, under labora564
tory conditions, several species of red and brown
Chilean algae are able to grow at high concentrations
of copper (between 10 and 150 lg l 1 ).
2. The mixture of trace metals (Cu, Zn, and Cd) occurring on the seawater of contaminated sites prevented
the growth of algae. Nevertheless, the result of their
work showed that under controlled conditions, several species of red and brown Chilean algae were able
to grow in the water of the contaminated sites (which
contained the trace metal mixture).
3. The unusual high density of herbivore gastropods occurring on the contaminated site eliminated the normal algal diversity. To test this hypothesis, at the
contaminated site, they conducted experimental exclusions of the grazers. This resulted in the fast colonization of red and brown algae on the treatment
plots. It was concluded that the main explanation
for the decreased algal diversity at the contaminated
site was the high grazing pressure exerted by these
The last explanation (No. 3) has two problems. At
their polluted site the gastropods had been observed for
the last 6 years; however the pattern of low diversity of
red and brown algal species has been observed for more
than 20 years (J.C. Castilla pers. comm.). Likewise,
during the entire period of our study, gastropods were
absent (in Enami) or occurred in low densities (at Santo
Domingo). However the pattern of low diversity in algal
species and cover domination by E. compressa was observed at both sites. A complementary explanation for
these patterns may be that the abrasive e€ects of tailings
prevent the colonization of the algae, and moreover
facilitate the growth of the ephemeral green algae. There
are no studies directly related to the e€ect of ground
mining-derived wastes on the diversity of rocky shore
marine communities. However, in several studies the
abrasive e€ects of naturally derived particles (i.e., sand)
on the composition and diversity of marine organisms
have been addressed (Airoldi, 1998; Daly and Mathieson, 1977; Menge et al., 1994; Gibbons, 1988; Graf,
1987; Moore, 1972; Robles and Cubit, 1981; Robles,
1982; Stewart, 1983). Daly and Mathieson (1977)
working on the east coast USA demonstrated that
highly sand-abraded rock surfaces, in the lower intertidal zone, were dominated by opportunistic annuals
like Enteromorpha spp and showed a lower diversity of
®lter feeding species, like Mytilus edulis and Balanus
balanoides. Robles (1982) showed that after the occurrence of storms, which drives sand on to the intertidal
rocky shores, a decrease in the diversity of sessile species
occurred. This decrease was followed by a temporary
colonization and dominance by green algae, mainly
Enteromorpha spp. Menge et al. (1994) reported that on
intertidal rocky shores, the interactions between echinoderm predators and their ®lter feeding prey could be
modi®ed by the burying e€ect of sand.
The relationship between the occurrence of ®ne
ground particles and the patterns of species composition
Volume 42/Number 7/July 2001
Fig. 9 MDS plot of the species composition at each site during the
study period. The key for the dates is 1 ˆ July±August 1997,
2 ˆ September±October 1997, 3 ˆ November±December 1997,
4 ˆ January±February 1998, 5 ˆ March±April 1998, 6 ˆ May±
June 1998.
Fig. 10 Cluster analysis of the similarity on the species composition
between the sites analysed in this study (Enami, Las Conchas,
Bandurrias and Santo Domingo) and the sites analysed in
previous studies (La Lancha, Palito, Zenteno and Pan de
ucar) (see text for details). The dashed line represents the
limit of similarity at a < 0:05.
at the sites a€ected by copper mine tailings, must be
addressed in future studies.
Temporal variations on the species mean numbers and
In contrast to non-contaminated, the contaminated
sites showed a low temporal variability in their number
of species (Fig. 4). These patterns were observed for
both algal and ®lter feeding species. There are no studies
on the range and the temporal variation in the number
of species at other contaminated sites of the northern
Chilean coast. However, both the mean and the variation in the total number of species observed in Las
Conchas and Bandurrias were similar to that observed
at three Northern-Chile non-contaminated sites analysed during 1996±1997 (Camus and Andrade, 1999).
During July±October of 1997 both Bandurrias and
Las Conchas showed a high mean number of species
but, due to the di€erences in the mean number of ®lter
feeding species (Fig. 4(c)) the values of Bandurrias were
higher than Las Conchas (Fig. 4(a)). These di€erences
have also been observed between Chilean non-contaminated sites, and may be related with the local abundance of predators (i.e, sea stars) or with the local
occurrence of upwelling (Camus and Andrade, 1999).
Percentage of cover
The mean percentage of cover ranged between 70±
90% at contaminated and non-contaminated sites.
However, at non-contaminated sites the rock surface
was occupied by algae and ®lter feeding species, while at
contaminated sites it was almost solely occupied by algal
species (Fig. 5(a)±(c)).
Non-contaminated sites showed a lower temporal
variability in the mean cover of species than contaminated sites. This di€erence is related to the fact that at
both Bandurrias and Las Conchas there was an inverse
Marine Pollution Bulletin
temporal relation between the cover of algae and ®lter
feeder species. This relation was not observed at Enami
and Santo Domingo (compare Fig. 5(b) and (c)). At
non-contaminated sites when a patch of bare rock is
created, algae or ®lter feeder animals colonize it, while at
contaminated sites it would either be colonized solely by
algae or remain uncolonized (i.e., rock). Considering
that primary substrate has been widely recognized as a
limiting resource for intertidal communities (Paine,
1994), the results of our work indicate that at contaminated sites, apart from the changes in diversity, some
important changes in the use and colonization of the
substrate may occur.
Spatial dominance and temporal incidence of the species
In all sites, the `item' `Bare Rock' was important. The
importance could be related with desiccation stress
(Bertness and Leonard, 1997), especially important in
the high intertidal zones of rocky shores located in desert areas (like the northern Chilean coast, Rojas et al.,
2000). Apart from `Bare Rock', contaminated and noncontaminated sites showed important di€erences for the
rest of the analysed items. At Enami and Santo Domingo E. compressa dominated the substrate and the
main di€erence between both sites was the incidence of
the red encrusting alga H. lecannellieri. This species was
absent at Enami while at Santo Domingo it showed the
same temporal and spatial incidence as at the non-contaminated sites. Considering that in terms of contamination the main di€erence between Enami and Santo
Domingo was related to the concentration, and more
speci®cally to the size of PIM, this alga could be a good
candidate to analyse the e€ects of ®ne ground wastes on
intertidal organisms.
In addition to the high incidence of H. lecannellieri, at
Bandurrias and Las Conchas U. lactuca and E. confervoides were important. The main di€erence between
these sites was that the red alga G. chilense and the
barnacle J. cirratus showed high dominance at Bandurrias and were not important at Las Conchas.
At the contaminated sites, there was a spatial dominance of E. compressaand `Bare Rock'. In the case of
non-contaminated sites, the spatial and temporal representation of `Bare Rock', U. lactuca, J. cirratus, E.
confervoides, and H. lecannellieri have been recognized
in other studies conducted in the northern Chilean coast
(Camus and Lagos, 1996).
During the period of study the contaminated sites
showed a pattern of species dominance completely different from the one observed at non-contaminated sites.
The main di€erence was that at the contaminated sites a
reduced number of species occurred and that only one of
these species dominated the primary substrate. This
pattern of dominance has been referred to as an indicator of highly polluted sites (Clarke, 1990; Warwick,
1986). The contribution of our study is the demonstration that the pattern of species dominance, at the sites
a€ected by copper mine tailings, occurs permanently
over time, in spite of seasonal variation in the cover of
other species.
Similarity of the species composition
During the study, the patterns of species composition
observed at contaminated sites were signi®cantly different from the ones observed at the non-contaminated
sites. Due to the major number of shared species, the
non-contaminated sites showed a level of similarity in
their species composition higher than that observed at
the contaminated sites.
The MDS diagram showed that within Bandurrias,
Las Conchas and Santo Domingo, the pattern of species
composition had restricted temporal variation, while in
the case of Enami a high temporal variability was observed. These di€erences were due to the fact that at
Enami, in spite of the high incidence of `Bare Rock' and
E. compressa, several species, like H. lecannellieri and E.
confervoides, were observed for no more than one sampling date. This pattern could be related with the major
temporal variability of some of the contaminants (like
®ne ground wastes).
The comparative analysis of the species composition
of our sites with the patterns of other sites located near
the city of Cha~
naral, showed interesting results. In spite
of the fact that the sites are separated by more than 200
km, the non-contaminated ones showed a high and
signi®cant level of similarity. The contaminated sites of
our study had signi®cant di€erences with the contaminated sites of the Cha~
naral area. The high level of similarity at the non-contaminated sites may be related to
the geographical characteristics of the intertidal communities on the northern Chilean coast. Based on a wide
geographical analysis, Camus (1998) proposed that, due
to the high levels of similarity in the species compositions of intertidal rocky shores communities, from 18°S
to 33°S the northern Chilean coast should be recognized
as a bio-geographic province.
The pattern of dominance of E. compressa and `Bare
Rock' has been observed in the four contaminated sites
compared. The main di€erence refers to the presence of
rare species. Due to the occurrence of E. confervoides, R.
expansa, and P. purpuratus (see Table 1), our contaminated sites showed a higher number of species (8±10)
than those a€ected (4±5) by El Salvador copper mine
discharges (Castilla, 1995). These di€erences could be
related to the concentrations of contaminants in the
seawater. The sites located near the city of Taltal (our
study) and the city of Cha~
naral (previous studies) have
similar levels of dissolved trace metals (Correa et al.,
1999) but they have important di€erences in the incidence of ®ne ground wastes. There is no analysis published on the concentration of PIM in the seawater of
the sites a€ected by El Salvador copper mine discharges,
although the in¯uence of ®ne ground wastes in these
sites has been widely recognized. Castilla and Correa
(1997), pointed out that at least, 130 million tons of solid
wastes have been discharged in the coast of Cha~
Volume 42/Number 7/July 2001
Pasko€ and Petiot (1990) reported that these discharges
formed arti®cial beaches more than 7 km long and so
they had a huge e€ect on the coastal landscape.
Finally, our study demonstrates that both the decreased diversity and the substrate monopolization exerted by the green algae E. compressa, are common and
permanent features of intertidal rocky shores of the
northern Chilean a€ected by copper mine tailings. The
spatial (between sites) and temporal (seasonal) variation
of this e€ect may be associated with the relative concentrations of trace metals and inorganic particles, on
the mining wastes occurring at each locality. The explanations proposed in previous studies for these changes (Correa et al., 2000), based on the e€ects of
herbivorous gastropods over the sessile species compositions, are not supported by our results.
Maximino Villaroel, Jorge Eraso, Jose Miguel Rojas, Marco Ramirez,
Carolina Oliva, Ignacia Toro, Carmen Espoz and Andrea Angel are
thanked for their advice during the di€erent stages of this work. Randy
Finke, Matthew Lee and one anonymous referee provided improvement on the manuscript. This work was funded by FONDECYT
2970075 and 490021 to JMF and partially with the 1997 JCC PEW
Fellowship on Marine Conservation and the Mellon Foundation
Grant to S. Navarrete and J.C. Castilla.
Airoldi, L. (1998) Roles of disturbance, sediment stress, and substratum retention on spatial dominance in algal turf. Ecology 79, 2759±
Bertness, M. D. and Leonard, G. H. (1997) The role of positive
interactions in communities: lessons from intertidal habitats.
Ecology 78, 1976±1989.
Biswas, A. K. and Davenport, W. G. (1980) Extractive Metallurgy of
Copper. Oxford Pergamon Press, Oxford, 488 pp.
Bryan, G. W. and Langston, W. J. (1992) Bioavailability, accumulation and e€ects of heavy metals in the sediments with special
reference to UK estuaries: a review. Environmental Pollution 76, 89±
Bustamante, R. H. (1994) Patterns and causes of intertidal community
structure around the coast of south Africa. Dissertation, University
of Capre Town, South Africa.
Camus, P. A. and Lagos, N. (1996) Variacion espacio-temporal del
reclutamiento en ensambles intermareales sesiles del norte de Chile.
Revista Chilena de Historia Natural 69, 193±204.
Camus, P. A. (1998) Estructura espacial de la diversidad en ensambles
sesiles del intermareal rocoso de Chile Centro-Norte: la diversidad
local como un resultado de determinantes de multiescala. Ph.D.
Thesis Dissertation. Ponti®cia Universidad Cat
olica de Chile,
Santiago, Chile.
Camus, P. A. and Andrade, Y. (1999) Diversidad de comunidades
intermareales rocosas del norte de Chile y el efecto potencial de la
surgencia costera. Revista Chilena de Historia Natural 72, 389±410.
Castilla, J. C. (1983) Environmental impact in sandy beaches of cooper
mine tailings at Ch~
naral, Chile. Marine Pollution Bulletin 14, 459±
Castilla, J. C. (1988) Earthquake-caused coastal uplift and its e€ects on
rocky intertidal kelp communities. Science 242, 440±443.
Castilla, J. C. (1995) Copper mine tailing disposal in northern Chile
rocky shores: Enteromorpha compressa (Chlorophyta) as a sentinel
species. Environmental Monitoring and Assessment 40, 171±184.
Castilla, J. C. and Correa, J. (1997) Copper Tailing impacts in coastal
ecosystems of northern Chile: from species to community responses.
National Health Forum Monographs, Metal Series No. 3, pp. 81±
Castilla, J. C. and Nealler, E. (1978) Marine environmental impact due
to mining activities of El Salvador Copper Mine, Chile. Marine
Pollution Bulletin 9, 67±70.
Clarke, K. R. (1990) Comparisons of dominance curves. Journal of
Experimental Marine Biology and Ecology 138, 143±157.
Clarke, K. R. (1993) Non-parametric multivariate analyses of changes
in community structure. Australian Journal of Ecology 18, 117±143.
Correa, J. A., Gonzalez, P. and Sanchez, P. (1996) Copper-algae
interactions: inheritance or adaptation? Environmental Monitoring
and Assesment 40, 41±54.
Correa, J., Castilla, J. C., Ramirez, M., Varas, M., Lagos, N., Vergara,
S., Moene, A. and Brown, M. (1999) Copper, copper mining tailings
and their e€ects on marine algae in northern Chile: from molecular
to community approach. Journal of Applied Phycology 11, 57±67.
Correa, J., Ramirez, M., de la Harpe, J. P., Roman, D. and Rivera, L.
(2000) Copper, copper mining e‚uents and grazing as potential
determinants of algal abundance and diversity in the northern
Chile. Environmental Monitoring and Assessment (in press).
Cortez, R., Solis, R. and Weisser, M. (1996) Compendio de la Minerõa
del Cobre. Editorial Antartica, Santiago. 661 pp.
Daly, M. A. and Mathieson, A. C. (1977) The e€ects of sand
movement on intertidal seaweds and selected invertebrates at
Bound Rock, New Hampshire, USA. Marine Biology 43, 45±55.
Day, R. and Quinn, G. P. (1989) Comparisons of treatments after an
analysis of variance in ecology. Ecological Monographs 59, 433±463.
Ellis, D. V. (1987) A decade of environmental impact assessment at
marine and coastal mines. Marine Mining 6, 385±417.
Gibbons, M. J. (1988) The impact of sediment accumulations, relative
habitat complexity and elevation on rocky shore meiofauna.
Journal of Experimental Marine Biology and Ecology 122, 225±241.
Glynn, P. W. (1988) EL Ni~
no Southern Oscillation 1982±1983:
nearshore population, community, and ecosystem responses. Annual Review of Ecology and Systematics 19, 309±345.
Graf, G. (1987) Benthic energy ¯ow during a simulated autumn bloom
sedimentation. Marine Ecology Progress Series 39, 23±29.
Hubalek, Z. (1982) Coecients of association and similarity based on
binary (presence- absence) data: an evaluation. Biological Review
57, 669±689.
Ingenierõa y Geotecnica (1990) Levantamiento catastral de los tranques de relave en Chile, Technical Report. Servicio Nacional de
Geologõa y Minerõa, Santiago, 1200 pp.
Jaccard, P. (1901) Distribution de la ¯ore alpine dans le basin des
dranses et dans quelquesregions voisines. Bulletin de la Societe
Vaudoise de Sciences Naturelles 37, 241±272.
Knauer, G. A. and Martin, J. H. (1973) Seasonal variations of
cadmium, copper, manganese, lead, and zinc in water and phytoplankton in monterrey bay, California. Limnology and Oceanography 18, 597±604.
Kuehl, R. O. (2000) Design of Experiments: Statistical Principles of
Research Design and Analysis. Duxbury Press, Paci®c Grove, 666
Lewis, A. G. (1995) Copper in Water and Aquatic Environments.
International Copper Association, LTD, New York, 65 pp.
Magurran, A. E. (1988) Ecological Diversity and its Measurement.
Princeton University Press, Princeton, NJ, 179 pp.
Manly, B. F. J. (1991) Randomization and Monte Carlo Methods in
Biology. Chapman & Hall, London, 281 pp.
Menge, B. A., Berlow, E. L., Blanchette, C. A., Navarrete, S. and
Yamada, S. B. (1994) The keystone species concept: variation in
interaction strength in a rocky intertidal habitat. Ecological
Monographs 64, 249±286.
Miethke, S., Castilla, J. C., Espoz, M. C. and Oliva, D. (1992) Impacto
ambiental por relaves de cobre en comunidades intermareales
rocosas de la III regi
on. XII Jornadas de Ciencias del Mar, Stgo
Chile 76.
Moore, P. G. (1972) Particulate matter in the sublittoral zone of an
exposed coast and its ecological signi®cance with special reference
to faunal inhabiting kelp holdfast. Journal of Experimental Marine
Ecology and Biology 10, 59±80.
Paine, R. T. (1994) Marine Rocky Shores and Community Ecology: An
Experimentalist's Perspective. Ecology Institute, Oldendorf/Luhe,
152 pp.
Pasko€, R. and Petiot, R. (1990) Coastal progradation as a by-product
of human activity: an example from Cha~
naral Bay, Atacama
Desert, Chile. Journal of Coastal Research 6, 91±102.
Robles, C. (1982) Disturbance and predation in an assemblage of
herbivorous diptera and algae on rocky shores. Oecologia 54, 23±
Robles, C. and Cubit, J. (1981) In¯uence of biotic factors in an upper
intertidal community: dipteran larvae grazing on algae. Ecology 62,
Rodriguez, S. R. R (1999) Subsidios tr
o®cos en ambientes marinos: la
importancia de las macroalgas pardas a la deriva como fuente
ogena de recursos alimentarios para el erizo Tetrapygus niger
Marine Pollution Bulletin
(Echinodermata: echonoidea) en el intermareal rocoso de Chile
central. Dissertation. P.U. Cat
olica de Chile, 172 pp.
Rojas, J. M., Fari~
na, J. M., Soto, R. E. and Bozinovic, F. (2000)
Variabilidad geogr
a®ca en la tolerancia termica y economõa hõdrica
del gastr
opodo intermareal Nodilittorina peruviana (Gastropoda:
Littorinidae, Lamarck, 1822). Revista Chilena de Historia Natural
(in press).
Sneath, P. H. (1957) Some thoughts on bacterial classi®cation. Journal
of General Microbiology 17, 184±200.
Stewart, J. G. (1983) Fluctuations in the quantity of sediments trapped
among algal thalli on intertidal rock platforms in southern
California. Journal of Experimental Marine Ecology 73, 205±211.
Sunda, W. G. and Lewis, J. A. M. (1978) E€ect of complexation byb
natural organic ligands on the toxicity of copper to an unicelular
alga, Monochrysis lutheri. Limnology and Oceanography 23, 870±876.
Underwood, A. J. (1997) Experiments in Ecology. Cambridge University Press, Melbourne, 504 pp.
Vasquez, J. A. and Guerra, N. (1996) The use of seaweds as
bioindicators of natural and anthropogenic contaminants in northern Chile. Hydrobiologia 326/327, 327±333.
Vasquez, J. A., Vega, J. M. A., Matsuhiro, B. and Urzua, C. (1999)
The ecological e€ects of mining discharges on subtidal habitats
dominated by macroalgae in northern chile: population and
community level. Hydrobiologia 398/399, 217±223.
Veermer, K. and Castilla, J. C. (1991) High Cadmiun residues
observed during a pilot study in shore birds and their prey
downstream from the El Salvador Copper Mine, Chile. Bulletin of
the Environmental Contamination and Toxicology 46, 242±248.
Warwick, R. M. (1986) A new method for detecting pollution e€ects
on marine macrobenthic communities. Marine Biology 92, 557±562.
Wilkinson, L., Blank, G. and Gruber, C. (1996) Desktop Data Analysis
with SYSTAT. Prentice-Hall, Englewood Cli€s, NJ, 798 pp.
Winer, B. J. (1971) Statistical Principles in Experimental Design.
McGraw-Kogakusha, Tokyo, 760 pp.