Factors influencing spatial patterns of the
ichthyofauna of low energy estuarine beaches
in southern Brazil
carlos werner hackradt1, fabiana ce’zar fe’lix-hackradt2, helen audrey pichler2,3,
henry louis spach1,4 and lilyane de oliveira e santos4
1Programa de Po´s-Graduac¸a˜o em Ecologia e Conservac¸a˜o, Setor de Cieˆncias Biolo´gicas, UFPR, 2Instituto Nautilus de Pesquisa e
Conservac¸a˜o da Biodiversidade, Avenida Senador Souza Naves, 655, cj 252, Cristo Rei, Curitiba, Parana´, CEP: 88.050-040, 3Programa
de Po´s-Graduac¸a˜o em Zoologia, Departamento de Zoologia, UFPR, 4Centro de Estudos do Mar, UFPR, Avenida Beira Mar s/n,
CP: 50002, CEP: 83.255-000; Pontal do Sul, Pontal do Parana´, Parana´, Brasil
Estuarine beaches are low energy environments found along the coast in several protected places around the world, such as
estuaries, bays or areas protected by submerged bars or coral reefs. Although common, these places have been poorly studied.
Using a seine net, six beaches were sampled along an environmental gradient for 12 months. During this period, an increase of
the dominant species following the increase in salinity and energy values was observed, whilst the opposite pattern occurred for
species richness and abundance. Univariate and multivariate analyses showed spatial and temporal variations between
beaches and months. Depth and marginal habitats appear to be more important factors than wave height and period, temp-
erature, salinity and pH, which are normally used to describe fish habitat utilization patterns in beaches and estuaries.
Nevertheless, specific sampling designs should be proposed to evaluate properly these questions.
Keywords: fish, estuarine beaches, dominance, spatial variation
Submitted 26 August 2008; accepted 12 August 2010; first published online 2 November 2010
I NTRODUCT ION
Sandy ocean beaches cover only a small portion of Earth’s
total surface; and in Brazil, these areas extend along almost
all the 9200 km of coastline (Hoefel, 1998). However, low
energy sandy beaches are limited to gulfs, bays, barrier
protected lagoons, islands protected by reefs or submerged
bars and principally estuaries (Nordstrom, 1992; Jackson et al.,
2002; Goodfellow & Stephenson, 2005), which are common
in southern and south-eastern Brazil (Borzone et al., 2003).
Nordstrom (1992) defines estuarine beaches as sand, gravel
or shell beaches located at partially closed places connected to
an ocean or sea. In these areas, dominant sediment reworking
processes are driven by local waves smaller than 0.25 m
(Jackson et al., 2002) with the wave formation centre not
further than 50 km (Nordstrom, 1992; Jackson et al., 2002).
In addition, beach face widths have to be narrow, measuring
less than 20 m in micro-wave regimes, and morphological fea-
tures must include those inherited by highly energetic events
(Jackson et al., 2002). Borzone et al. (2003) suggested that
these places give rise to a new morphodynamic environment
type which is characterized as a transition between wave-
dominated sandy beaches and tide-dominated flats.
All morphological features generated by beach peculiarities
have influenced the biotic communities of beach environ-
ments (Brow & McLachlan, 1990; Romer, 1990), and many
studies have been dedicated to investigate their role in the
fish life cycle. These investigations found a numerical preva-
lence of both few species and juvenile individuals (Lasiak,
1984, 1986; Santos & Nash, 1995; Gibson et al., 1993; Clark
et al., 1996; Clark, 1997; Strydom, 2003). Most fish remained
during short periods in this environment (Gibson et al., 1993)
and only a reduced number of species showed annual resi-
dence (Brown & McLachlan, 1990). Another important
factor in beach environment is the high food availability due
to continuous wave action, which makes nutrients available
in the water column. This process favours phytoplankton
enrichment, and consequently, the planktophagic organisms
(McLachlan, 1980).
Previous studies have found greatest fish abundance during
warmer months, decreasing with temperature reduction
(Modde & Ross, 1981; Ross et al., 1987; Gibson et al., 1993;
Santos & Nash, 1995; Clark et al., 1996). These differences
are attributed to the effect of the environmental set, such as
wind, wave and water temperature (Lamberth et al., 1995).
However, some authors have found greatest abundances
during spring rather than in the summer, as it was expected
(Godefroid et al., 1997; Fe´lix et al., 2007a).
Many authors have studied fish communities at Brazilian
beaches. The first studies were focused on understanding
spatial and temporal community patterns, characterizing
species composition and comparing sites (Cunha, 1981;
Paiva-Filho et al., 1987; Monteiro-Neto et al., 1990; Grac¸a
Lopes et al., 1993; Monteiro-Neto & Musick, 1994; Giannini
& Paiva-Filho, 1995; Saul & Cunningham, 1995; Teixeira &
Almeida, 1998; Lopes et al., 1999; Gomes et al., 2003). Most
Corresponding author:
C.W. Hackradt
Email: hackradtcw@gmail.com
1345
Journal of the Marine Biological Association of the United Kingdom, 2011, 91(6), 1345–1357. # Marine Biological Association of the United Kingdom, 2010
doi:10.1017/S0025315410001682

recently, authors have investigated the daily ichthyofaunal
variation and the influence of morphodynamic gradients
(Gaelzer & Zalmon, 2003; Pessanha & Arau´jo, 2003).
With respect to the coast of Parana´, beach ichthyofauna is
poorly studied, with few dispersed investigations, mainly on
local comparisons between sandy beaches (Pinheiro, 1999),
temporal variations (Godefroid et al., 1997, 2004), ichthyo-
plankton (Godefroid et al., 1999) and the influence of mor-
phodynamism on fish community (Fe´lix et al., 2007b).
Estuarine beaches were only evaluated based on species com-
position (Hackradt et al., 2009) and temporal variation
(Godefroid et al., 1997; Fe´lix et al., 2006).
Despite the range of advances in studies of different beach
environments and the great number of biological assessments
at several estuarine habitats, such as mangroves, seagrass beds,
tidal flats and water columns, estuarine beaches remain largely
unstudied (Nordstrom, 1992; Hoefel, 1998). They are unique
environments that differ from sandy beaches by presenting
a stable substrate that allows fauna and flora attachment
(Nordstrom, 1992). In this context, the aim of the present
study is to understand the ichthyofaunal structuring at six
estuarine beaches along a salinity–energy gradient inside
the largest estuary in southern Brazil.
MATER IALS AND METHODS
Study area
The Bay of Paranagua´ is described as a type B, partially-mixed
estuary, with lateral heterogeneity (Knoppers et al., 1987). The
estuary penetrates 50 km into the continent, with a mean
width of 10 km and an average depth of 5.4 m (Noernberg
et al., 2004). The occurrence of a salinity and energy gradient
along the east–west axis divides the bay into three zones: (1)
an external high energy region with average salinity of about
30, called the euhaline region, which includes the following
beaches: Encantadas (EN: 25833′49.1′′S 48819′05.1′′W),
Brası´lia (BR: 25831′36.4′′S 48820′35.7′′W), Coroinha (CO:
25830′40.9′′S 48822′38.8′′W) and Cobras’ Island (IC:
25829′03.1′′S 48825′50.6′′W); (2) an intermediary, polyhaline
region where Piac¸aguera beach is located (PI: 25829′03.1′′S
48829′40.0′′W); and (3) an innermost low energy and salinity
region, called the oligo-mesohaline region, with salinity values
between zero and 15, where Europinha beach is located (EU:
25827′39.2′′S 48836′41.1′′W) (Figure 1). Usually, waves are
originated by south-eastern winds at the estuary mouth
region, displaying on average half a metre height and three
to seven seconds period duration. In stormy conditions,
waves can reach a maximum of 3 m height (Lana et al., 2001).
Data collection
Fish assemblages at the 6 locations were sampled during day-
light hours, between 6.00 and 13.00 h from June 2005 to May
2006, using a beach seine net, 15 m long and 2.6 m height with
a stretched mesh size of 5 mm. Three 20-m hauls were made
at each site, separated 5 m apart to minimize the influence on
the following haul. All sampling campaigns began at neap low
tide, following the same beach visiting sequence. Hauls were
pulled simultaneously and parallel to the beach face by two
persons, one at each end of the net. All fish collected were
identified to species level following Figueiredo & Menezes
(1978, 1980, 2000) Fischer (1978), Menezes & Figueiredo
(1980, 1985) and Barletta & Correˆa (1992). These were then
weighed (g) and measured to the nearest 1 mm (total length
and standard length), except when samples were very large.
On such occasions, measurements were restricted to a sub-
sample of 30 individuals per species. The excess was
weighed, counted and incorporated as weight and number
counts.
Environmental data were measured concomitantly with
beach hauls: surface water salinity (using a refractometer),
surface water temperature (through a mercury thermometer),
pH (through a portable pH meter), wave height and wave dur-
ation. Wave height was taken with a 2-m ruler and obtained
from the metric difference between crest and sea level of the
largest waves breaking on the surf zone. Wave period was
measured from the duration (in seconds) of 11 successive
breaking waves divided by 10 to obtain the period of a single
wave. This procedure was applied twice to produce an average.
Seasons of the year were defined as follows: summer
(December, January and February), autumn (March, April
and May), winter (June, July and August) and spring
(September, October and November).
Data analysis
To determine whether species can be classified as dominant,
the following criteria have been used: frequency of occurrence
in the samples exceeding 10%; abundance exceeding 1%; and,
constant occurrence, i.e. present in at least eight collection
months.
To test whether the abundance (N), number of species (S),
catch weight (P), Margalef’s richness (d), Pielou’s evenness (J′)
and Shannon–Wiener diversity (H′) were spatio-temporally
different, two-way analysis of variance (ANOVA) (Pielou,
1969; Ludwig & Reynolds, 1988) was applied. Before conduct-
ing the test, biotic data were tested for homoscedasticity and
normality by the Bartlett and Kolmogorov–Smirnov tests,
respectively (Sokal & Rohlf, 1995). To fulfil ANOVA assump-
tions, abundance (N), catch weight (P), Pielou’s evenness (J′)
and Shannon–Wiener diversity index (H′) data were trans-
formed by Log(x + 1). When differences were significant
(P , 0.05), the a posteriori Student–Newman–Keuls test
was used to identify which averages were different.
Data on species abundance (log transformed) were con-
verted into similarity matrices using the Bray–Curtis simi-
larity index, with all field points separated by seasons.
Following, ANOVA results were displayed on a dendrogram
using group average linking (cluster), and an ordination
plot, generated by a non-metric multidimensional scaling
(MDS) procedure (Clarke & Warwick, 1994). To evaluate
the statistical importance of group formation, a similarity
analysis (ANOSIM) was performed and, to reveal species con-
tribution to group formation, a similarity of percentages
(SIMPER) procedure was conducted subsequently. To evalu-
ate the correlation level between environmental data that
best explained fish community patterns, the BIOENV
routine was applied.
RESULTS
Environmental results showed beach singularities and marked
temporal differences. Temperature reached highest values
1346 carlos werner hackradt et al.