Do environmental factors influence the movement of estuarine fish?

A case study using acoustic

telemetry

A.-R. Childs*, P.D. Cowley, T.F. Næsje, A.J. Booth, W.M. Potts, E.B. Thorstad and F. Økland

*Department of Ichthyology and Fisheries Science, Rhodes University, PO Box 94, Grahamstown,
6140, South Africa

Abstract

Telemetry methods were used to investigate the influence of selected environmental variables on
the position and movement of an estuarine-dependent haemulid, the spotted grunter Pomadasys
commersonnii (Lacepède 1801), in the Great Fish Estuary, South Africa. Forty individuals (263–
698 mm TL) were surgically implanted with acoustic coded transmitters and manually tracked during
two periods (7 February to 24 March 2003; n = 20 and 29 September to 15 November 2003;
n = 20). Real-time data revealed that spotted grunter are euryhaline (0–37) and are able to tolerate
large variations in turbidity (4–356 FTU) and temperature (16–30 °C). However, the fish altered their
position in response to large fluctuations in salinity, temperature and turbidity, which are
characteristic of tidal estuarine environments. Furthermore, tidal phase had a strong influence on
the position of spotted grunter in the estuary.

1. Introduction

Estuarine-dependent fish species are defined as those who would be adversely affected by the loss
of estuarine habitats (Whitfield, 1994c). Estuaries, like many other types of wetland worldwide, are
under long-term threat of damage and destruction. Human degradation, coupled with an increase in
fishing pressure on estuarine systems ([Wallace et al., 1984], [Houde and Rutherford, 1993],
[Cattrijse et al., 2002], [Hartill et al., 2003] and [Lamberth and Turpie, 2003]), places estuarine fish
species under threat. As a result, the importance of estuarine systems to fish species is widely
acknowledged ([Lenanton and Potter, 1987], [Potter et al., 1990] and [Hartill et al., 2003]). Although,
studies on the relationships between estuarine fish and environmental factors, such as salinity
([Whitfield, 1994a] and [Shervette et al., 2007]), temperature (Marshall and Elliot, 1998), turbidity
(Cyrus and Blaber, 1987) and tidal currents (Szedlmayer and Able, 1993) have been examined
world-wide, the relationship between fish movement and environmental factors in estuaries is not
well documented. Studies of this nature require knowledge on the real-time movement of estuarine
fishes in relation to the fluctuating environmental factors characteristic of estuarine environments.
Such high resolution studies can only be conducted using telemetry techniques. Telemetry enables
the monitoring of real-time movements of individual fish, and as a result determine the exact abiotic
environment in which fish are found. Previous studies have only examined the effect of
environmental factors on fish distribution, abundance and/or assemblages, and only a limited
number of studies, namely (Almeida, 1996), (Szedlmayer and Able, 1993) and (Taverny et al.,
2002) and Kelly et al. (2007), have tested the influence of these factors on fish movements in
estuarine environments. Knowledge on the relationship between environmental factors and
estuarine fish movements within an estuary is essential to improve our biological understanding of
estuarine-dependent species.

The spotted grunter Pomadasys commersonnii, is an estuarine-dependent species found in inshore

coastal regions and estuaries of the Western Indian Ocean (Smith and Heemstra, 2003). It spawns
at sea and juveniles enter estuaries where they remain for a period of up to three years before
returning to the marine environment ([Wallace, 1975a] and [Wallace, 1975b]). Consequently, early
juveniles are considered to be entirely dependent on estuaries, while adult fish frequent estuaries,
presumably to feed ([Wallace, 1975b] and [Whitfield, 1994c]). The survival of spotted grunter and
other estuarine-dependent species in South African waters is determined by the existence of
numerous estuarine systems along the coast ([Wallace et al., 1984] and [Whitfield, 1994c]). These
estuaries are dynamic, unpredictable environments characterized by large environmental variation.
Frequent, abrupt changes in salinity, temperature and turbidity place considerable physiological
demands on fishes that occupy these systems (Harrison and Whitfield, 2006). Although estuarine
fishes have evolved to tolerate large environmental fluctuations, they can also use movements to
decrease environmental stress by finding areas that better suit their physiological needs ([Whitfield,
1994b] and [Whitfield, 1994c]). The aim of this study was to determine whether an estuarine-
dependent haemulid, the spotted grunter, tolerates large fluctuations in salinity, temperature and
turbidity, or if it uses behavioural responses such as tidal transport to move in response to
fluctuating environmental conditions.
2. Materials and methods

2.1. Study site

The Great Fish River enters the Indian Ocean approximately halfway between the coastal cities of
Port Elizabeth and East London at 33° 29′ 28″S, 27° 13′ 06″E (Fig. 1). It receives large volumes of
freshwater from an interbasin transfer scheme, which ensures a continuous input of nutrients and
suspended particles, resulting in a highly productive, but turbid estuary. The estuary is riverine in
appearance, with an increase in turbidity towards the upper reaches. The perennial river flow
together with the tidal exchange ensures a permanently open connection to the sea (Grange et al.,
2000). The Great Fish Estuary is approximately 12 km in length and has a tidal prism of
1.6 × 106 m3 (Whitfield et al., 1994). The spring tidal range is between 1 and 1.5 m in the lower
reaches and decreases towards the head (Whitfield et al., 1994). The tidal prism volume exceeds
the river water volume by six times during an average tidal cycle. The lower reaches are mainly
marine-dominated, the middle reaches represent the mixing zone between river and sea, and the
upper reaches are freshwater dominated (Grange et al., 2000). The estuary channel is narrow (30–
100 m wide) and its depth (0.5–3.5 m) is dependent on flooding events (Whitfield et al., 1994). The
bathymetry of the Great Fish Estuary is uniform and is mostly shallow, ranging between 1 and 2 m
(mean 1.4 m), except for an area in the lower (±3–4 m), middle (3 m) and two areas in the upper
reaches (±5–6 m). The shallow nature of the estuary is a result of the large fluvial sediment load
from the catchment (Grange et al., 2000). The mouth region is restricted by the presence of
extensive sand banks. The estuary is characterised by strong longitudinal and vertical physico-
chemical gradients. The abiotic characteristics of the Great Fish Estuary were monitored at eight
fixed stations during both study periods (the same locations as the ALS sites in the second period,
see Fig. 1). Salinity (Practical Salinity Scale, Atago handheld refractometer), temperature (°C,
digital/electronic thermometer), turbidity (FTU, Hanna 93703 turbidity meter), depth (m, graduated
weighted rope) and current speed (m s−1) were recorded. Current speed was calculated from the
time that it took a neutrally buoyant object to move 2 m on the water surface. Water samples were
taken approximately 30 cm above the bottom using a Van Doorn water sampler. River discharge (m
s−1) and conductivity (m s m−1) were supplied by the Department of Water Affairs and Forestry,
South Africa.
Fig. 1. Map of South Africa showing the location of the Great Fish River and a detailed map showing
the position of the automated listening stations (ALSs) (black dots) in the Great Fish Estuary during
the second period (29 Sep 2003 to 12 Feb 2004).

2.2. Research approach

Acoustic telemetry methods were used to track the movements of spotted grunter in the Great Fish
Estuary during two periods. In the first period (7 February to 24 March 2003), 20 individuals
between 263 and 387 mm TL (mean 336 mm TL) (Table 1) were tagged with VEMCO V8SC-2L-
R256 (VEMCO Ltd, Halifax, Canada) coded transmitters (69 kHz). These tags had an expected
battery life of 112 days, were 8.5 mm in diameter, 28 mm in length and weighed approximately 3.1 g
in water being on average 0.75% of the fish body mass (min 0.5%, max 1.6%). In the second period
(29 September to 15 November 2003), 20 fish between 362 and 698 mm TL (mean 478 mm TL)
(Table 1) were tagged with VEMCO V13SC-1L-R256 coded transmitters (69 kHz). These tags had
an expected battery life of 130 days, were 13 mm in diameter, 36 mm in length and weighed
approximately 6 g in water being on average 0.5% of the fish's body mass (min 0.2%, max 1.2%).
Each transmitter emitted a unique acoustic pulse train randomly every 5–15 s. Prior to this study, a
pilot laboratory test was conducted on spotted grunter. Fish were tagged with exact replicates (size
and weight) of the VEMCO transmitters using the tagging techniques (capture and retention of fish
and the appropriate surgical procedure) described by Næsje et al. (2007). No post-tagging infection,
haemorrhaging or abnormal post-tagging behaviour was observed during the 90-day trial period. In
addition, field trials were conducted by Kerwath et al. (2005), which revealed that the adopted
techniques were appropriate for an acoustic telemetry study on this species. According to Wallace's
(1975b) estimated size at sexual maturity (300–400 mm TL), the fish tagged in the first study period
were mainly juveniles, while both adults and juveniles were tagged in the second period. Manual
tracking of spotted grunter commenced 6 and 4 days after the last fish was released in the first and
second periods, respectively, to allow for acclimation. During the acclimation period, fish were
tracked intermittently to check for any possible tagging effects. None of the fish showed any
noticeable abnormal post-tagging behaviour.

Table 1.

Details of the 40 acoustically tagged spotted grunter manually tracked in the Great Fish Estuary
during the first (7 Feb to 24 Mar 2003) and second (29 Sep to 15 Nov 2003) periods, and
acoustically monitored during the extended 137-day second period (29 Sep 2003 to 12 Feb 2004).
Asterisks (*) indicate the end of the study period

Total No.
Study Fish Date last recorded: Date last recorded: ALS
length positional
period code manual tracking monitoring (Period 2 only)
(mm) fixes

1 20 317 25 16 March 2003

1 21 334 34 24 March 2003*

1 22 297 20 7 March 2003

1 23 380 36 24 March 2003*

1 24 330 18 1 March 2003

1 25 313 19 3 March 2003

1 26 314 35 24 March 2003*

1 27 328 12 19 February 2003

1 28 382 34 24 March 2003*

1 29 377 36 24 March 2003*

1 30 308 36 24 March 2003*

1 31 357 12 10 March 2003

1 32 318 36 24 March 2003*

1 33 329 13 9 March 2003

1 34 263 8 17 February 2003

 Total No.
Study Fish Date last recorded: Date last recorded: ALS
length positional
period code manual tracking monitoring (Period 2 only)
(mm) fixes

1 35 357 8 18 February 2003

1 36 387 15 25 February 2003

1 37 363 23 23 March 2003

1 38 319 12 23 February 2003

1 39 355 36 24 March 2003*

2 50Aa 449 12 11 October 2003 11 October 2003

2 50Bb 515 15 15 November 2004* 28 January 2004

2 51 469 32 15 November 2004* 12 February 2004*

2 52 385 42 15 November 2004* 12 February 2004*

2 53 428 37 15 November 2004* 12 February 2004*

2 54 620 22 15 November 2004* 12 February 2004*

2 55 432 29 10 November 2003 10 November 2003

2 56 440 42 15 November 2004* 12 February 2004*

2 57 364 42 15 November 2004* 12 February 2004*

2 58 625 21 15 November 2004* 12 February 2004*

2 59 472 28 15 November 2004* 12 February 2004*

2 60 527 29 15 November 2004* 12 February 2004*

2 61 489 30 15 November 2004* 28 January 2004

2 62 504 29 15 November 2004* 12 February 2004*

2 63 534 20 15 November 2004* 25 December 2003

2 64 387 35 15 November 2004* 12 February 2004*

2 65 698 20 15 November 2004* 26 January 2004

2 66 403 42 15 November 2004* 12 February 2004*

2 67 428 35 15 November 2004* 12 February 2004*

2 68 538 32 15 November 2004* 12 February 2004*

2 69 362 41 15 November 2004* 12 February 2004*

a
Fish that was caught during study.
b
Fish that was caught and tagged during the study.
Tagged fish were manually tracked for 36 days in the first, and 42 days in the second period. Each
period included two 16-consecutive day sampling sessions and an interim period of 14 days, during
which fish were tracked every third day in the first and every second day in the second period. Each
16-consecutive day session was standardised according to the lunar phase and tracking was
conducted over two semi-lunar cycles. Each session began 2 days prior to the first quarter (waxing)
moon, and the last day of each session was the last quarter (waning) moon.

Manual tracking was conducted using a VEMCO VR60 receiver and VH10 hydrophone from a
motorised boat. The position of each fish was recorded once a day. Manual tracking sessions
began at the river mouth at approximately 08:00 h. When a fish was located, the boat was
anchored, the coordinates were recorded using a GPS (Garmin 12) and water chemistry variables
(salinity, temperature, turbidity, depth and current speed) were measured, as described above,
when obtaining a positional fix for each fish. If all the fish were not recorded in the estuary, the
sampling was extended into the riverine environment between 13 and 14 km upstream. If all the fish
were still not recorded, the procedure was repeated on the return trip to the estuary mouth, where
the session ended at approximately 18:00 h. The number of positional fixes per fish varied during
both sampling periods, as some fish undertook short-term sea trips or emigrated and did not return
during the respective study periods (Table 1). Tests were conducted to determine the precision of
tracking, using transmitters that were hidden at random locations within the estuary. On each
occasion, the position was recorded within 1 m of the transmitter location.

During the second period, eight VEMCO VR2 acoustic listening stations (ALSs) were deployed
along the length of the estuary from 29 September 2003 to 12 February 2004 (Fig. 1). The up and
downstream movements of spotted grunter were therefore only monitored during the second period.
These movements were monitored over the manual tracking period (29 September 2003 to 15
November 2003) and for an additional 95 days.

A reward system and an awareness campaign were implemented to ensure that fishers returned the
transmitter of a tagged fish they had caught. One fish (Fish 50A) was caught and returned on 10
October 2003 and was replaced by another (Fish 50B) later in the study (Table 1). Subsequently
these fish were excluded from all data analyses.

2.3. Data analyses

Data were analysed using GIS (ArcView® GIS 3.2) and the VEMCO software package. The position
of the fish was expressed as ‘distance from the estuary mouth’. Given the longitudinal nature of the
estuary, the estuary was divided into 500 m sections and the number of positional fixes recorded for
each fish within each 500 m section was calculated as a proportion of the total number of fixes. The
proportion of positional fixes calculated for each fish within each stretch was then averaged and
presented as a frequency histogram. This ensured that the assumption of independence was not
contravened (Grafen and Hails, 2002) and that the contribution of each individual fish was equally
weighted.

2.3.1. Environmental variables

The frequency histograms for each of the environmental variables salinity, temperature, turbidity,
depth and current speed were calculated as described above. Each environmental variable was
binned into specific categories (e.g., 16–17 °C). The proportion of positional fixes for each fish
within each category was averaged and the mean proportion of positional fixes within each category
was presented as a frequency histogram. A Pearson product–moment correlation was used to
assess the relationships between salinity, temperature and turbidity during the first and second
periods.

The chi-square test of independence was used to test the hypothesis that the direction of fish
movement was dependent on the tidal phase. A binomial test, corrected with the Bonferroni
adjustment, was used to determine the probability of movement with and against the tide. An
upstream or downstream movement was only considered if an individual fish passed more than one
ALS in the same direction within 6 h. However, in the cases where the ALSs were situated more
than 1 km away from each other, the assumption of an individual passing more than one ALS to
constitute a movement did not apply, but instead the movement to each of these receivers within a
6 h period was considered as an upward or downward movement. All upstream movements were
assigned a 1 and downstream movements a 2. The corresponding tidal state was then assigned to
each of the upstream or downstream movements. For each fish, the number of movements
upstream and downstream, as well as with and/or against the tide was calculated.

The relationship between fish position and tidal phase was analysed using a Rayleigh test of
randomness (Batschelet, 1981). For this calculation, the estuary was divided into the lower (0–2 km
from the estuary mouth), middle (2–6 km from the estuary mouth), upper (6–11 km from the estuary
mouth) and riverine (>11 km from the estuary mouth) regions, based on the mean bottom salinities
recorded at the eight fixed positions during the study. The mean time after low tide (h) that the
positional fixes were recorded within each region of the estuary was expressed as theta (θ), the
mean direction of the resultant vector (measured in radians). Theta for each fish was then used to
calculate the mean time after low tide that all fish were recorded in each region. The number of
observations in each stretch of the estuary varied and the significance of these observations could
only be tested when more than five observations were recorded. The tidal phase was recorded
when each fish was located. The time delay of the tide from sea to 10.3 km upstream was
calculated using the daily depth values measured at each fixed station and was considered when
assigning the tidal phase at each fish position.

2.3.2. Effect of environmental variables on fish movement

A generalized linear model was used to model the relationship between the relative change in a
fish's position from time t to t + 1 and the relative change in salinity, temperature, turbidity and tide
from time t to t + 1, fish size and season. Given the inherent autocorrelation generally found within
telemetry data ([Dunn and Gipson, 1977], [Swihart and Slade, 1985] and [De Solla et al., 1999]),
and hence the autocorrelation found in the present data, the relative change between all variables,
dependent and independent, removed the inherent autocorrelation in the data and did not violate
the assumptions of the linear regression approach. The ‘Wald’ statistic (W) and its p-level were
used to test the significance of each regression coefficient. The response variable was the relative
change in position (distance from the mouth) from time t to t + 1, and the continuous independent
variables were the relative change in salinity, temperature, turbidity, and tide and the categorical
variables were season (summer = period 1 and spring = period 2) and size of fish (small ≤400 mm
TL and large >400 mm TL). Since the data of the response variable was continuous, the identity link
function was used for variables with a normal distribution (McCullagh and Nelder, 1995). The
following parameters were included in the generalized linear model (GLM):

∆Fish
position=β0+β1(∆Salinity)+β2(∆Temperature)+β3(∆Turbidity)+β4(∆Tide)+β5(Season)+β6(Size)+
where ∆ is the relative change from time t to t + 1, β* are the regression coefficients, and is the
model residual where N(0,σ2).

Movement was assumed to be upstream if the change was positive and downstream (towards the
mouth) if the change was negative. Only data where consecutive days were sampled were included
in the model.

Partial correlations (Yule, 1907) were used to test if any of the environmental variables (salinity,
temperature, turbidity and tide) had an underlying influence on the relationship between these
variables.
3. Results

Bottom temperatures recorded at the eight fixed water chemistry stations ranged from 15.2 to
29.5 °C in the first and from 15.7 to 25.8 °C in the second period. Bottom turbidities recorded at the
fixed stations ranged from 12.4 to 762 FTU in the first and from 3.67 to 358 FTU in the second
period, and bottom salinities ranged from 0 to 38 in both periods. Using the Venice System
(Whitfield, 1998), the bottom salinity profile revealed euhaline conditions (30.0–39.9) from the mouth
to 1 km upstream, polyhaline conditions (18.0–29.9) from 1 to 3.5 km upstream, mesohaline
conditions (5.0–17.9) from 3.5 to 6 km upstream and oliogohaline conditions (0.5–4.9) from 6 km to
the head of the estuary (12 km). Surface current speeds recorded at the fixed stations ranged from
0 to 0.76 m s−1 in the first and from 0 to 0.69 m s−1 in the second period. The depth profile of the
estuary during the first study was uniform, ranging between 1 and 2 m, except for a few deep areas
in the lower and upper reaches of the estuary. However, the bathymetry of the estuary changed
dramatically after a flash flood in May 2003 (149 mm rainfall overnight), creating large scours and
holes in the middle (±4.5 km from the estuary mouth) and upper (±7 km from the estuary mouth)
reaches of the estuary. The most affected area was in the upper reaches of the estuary between 6
and 8 km from the mouth.

River discharge was relatively constant during both the first and second periods, with a mean
discharge of 6.28 ± 1.86 m s−1 (range 1.48–9.76 m s−1) and 5.13 ± 1.19 m s−1 (range 3.21–7.90 m
s−1), respectively. Mean conductivity was slightly higher in the second period (mean 176.17 m s m−1;
range 159–190 m s m−1) when compared to the first period (mean 137.88 m s m−1; range 121.3–
149 m s m−1).

The number of positional fixes per fish varied (range 8–42, mean 28 ± 10 SD) during both sampling
periods, as some fish went to sea (Table 1).

In the first period, the position of spotted grunter in the estuary extended from the mouth to 12.1 km
upstream. The mean proportion of observations was highest (90%) within 6.0 km of the estuary
mouth, of which 68% were within the first 3.0 km and half (50%) between 1.0 and 1.5 km from the
mouth (Fig. 2a).
Fig. 2. Position of tagged spotted grunter in the Great Fish Estuary, based on the mean proportion
of observations per 500 m zone recorded while manual tracking, during (a) the first (7 Feb 2003 to
24 Mar 2003) and (b) second (29 Sep 2003 to 15 Nov 2003) period. Numerical values above the
bars indicate the number of individuals recorded in each 500 m zone.

In the second period, tagged individuals were positioned from the mouth to 13.4 km upstream. The
distribution of fish along the estuary was bimodal, with 43% of the mean proportion of positional
fixes recorded between the mouth and 3.0 km upstream and one-third (33%) in the upper reaches
(6–8 km upstream). Only 19% were found in the middle reaches (3–6 km) and 5% in the uppermost
region (8–13.5 km upstream) (Fig. 2b).

3.1. Environmental variables

3.1.1. Salinity

Spotted grunter were found in salinities ranging between 0 and 36 during both periods. The mean
salinity in which tagged individuals were recorded was 22.1 and 15.5 in the first and second periods,
respectively. Observations were not uniformly distributed (Fig. 3a). During the first period, the mean
proportion of observations was highest (35%) in the euhaline range, while during the second period
the mean proportion of observations was highest (29%) in the oligohaline range (Fig. 3a).
Fig. 3. The mean proportion of observations describing (a) salinity (Oligo, oligohaline region; Meso,
mesohaline region; Poly, polyhaline region; Eu, euhaline region, according to the Venice System),
(b) temperature, (c) turbidity, (d) surface current speed and (e) depth at which tagged spotted
grunter were recorded in the Great Fish Estuary during the first (7 Feb 2003 to 24 Mar 2003)
(Period 1) and second (29 Sep 2003 to 15 Nov 2003) periods (Period 2).

3.1.2. Temperature

Mean water temperature at which spotted grunter were located was 23.0 °C (range 17.3–30.5 °C) in
the first and 20.2 °C (range 16.3–25.3 °C) in the second period. The distribution of observations
within each temperature range was not uniform (Fig. 3b). In the first period, the mean proportion of
observations was highest (63%) between temperatures of 22 and 25 °C, while in the second period,
65% were in temperatures ranging between 18 and 21 °C (Fig. 3b). Although in the first period, the
mean proportion of observations was only 10% in water temperatures higher than 25 °C, no
observations were recorded in this range during the second period (Fig. 3b).

3.1.3. Turbidity

Spotted grunter were located in water ranging from 6.0 to 567.0 FTU (mean 111.5 ± 83.5 (SD) FTU)
in the first and from 4.1 to 358.0 FTU (mean 92.7 FTU, SD 63.4) in the second period. In both the
first (49%) and second (60%) periods, the mean proportion of positional fixes was greatest in water
of 20–100 FTU, while in water less than 20 FTU it was low in both the first (1.8%) and second (4%)
periods. The mean proportion of observations was also high in turbid water (exceeding 100 FTU) in
both the first (49%) and second (36%) periods (Fig. 3c).

3.1.4. Depth

The mean water depth at which spotted grunter were located was 1.6 m (range 0.1–3.6 m) in the
first and 1.7 m (range 0.4–5.9 m) in the second period. The mean proportion of positional fixes was
highest at depths between 1 and 2 m in both the first (74%) and the second (65%) periods, but
much lower in depths less than 1 m, in both the first (7.7%) and second (11%) periods. The mean
proportion of observations was only 19% in water deeper than 2 m in the first, and 25% in the
second period (Fig. 3d).

3.1.5. Current speed

Spotted grunter were located at an average surface current speed of 0.32 m s−1 (range 0–0.83 m
s−1) in the first, and 0.27 m s−1 (range 0–0.93 m s−1) in the second period. The maximum surface
current speed (0.93 m s−1) in which spotted grunter were recorded was in the mouth region of the
estuary. The observations were not uniformly distributed. The mean proportion of observations was
highest at current speeds ranging between 0 and 0.39 m s−1 in both the first (64%) and second
(79%) periods (Fig. 3e).

3.1.6. Tidal phase

The position of spotted grunter in the estuary was influenced by the tidal phase. Spotted grunter
were found in the lower reaches of the estuary during low tide and in the upper reaches during the
incoming and high tides. In the first period, spotted grunter were located in the lower reaches of the
estuary around low tide (mean time after low tide, θ = 11:20 ± 01:59; p < 0.005; r = 0.50; n = 20), in
the middle reaches on the incoming tide (mean time after low tide, θ = 01:31 ± 02:15; p > 0.01;
r = 0.36; n = 18), and in the upper region of the estuary on the high tide (mean time after low tide,
θ = 05:26 ± 00:30; p < 0.001; r = 0.97; n = 8). While only one fish was recorded in the riverine
environment, this fish was recorded on average during the incoming and high tide (mean time after
low tide, θ = 04:53 ± 01:49). Similarly, in the second period, spotted grunter were found in the lower
reaches of the estuary during low tide (mean time after low tide, θ = 00:52 ± 01:48; p < 0.001;
r = 0.59; n = 17), in the middle reaches of the estuary on the incoming tide (mean time after low tide,
θ = 02:22 ± 01:43; p < 0.001; r = 0.63; n = 18) and in the upper reaches on the incoming tide (mean
time after low tide, θ = 03:38±01:16; p < 0.001; r = 0.79; n = 19). Although the number of fish
located in the riverine environment was too low to test the significance of the tidal phase on fish
position, the three fish recorded in this region were located on average during the high tide (mean
time after low tide, θ = 06:06 ± 00:28).

The probability of movements with the tide for each tagged fish was significant (Binomial test:
p < 0.01). The mean percentage of movements made by the tagged spotted grunter with the tide
was 93%, compared with movements made against the tide (5%), during high (1%) and low tide
(1%). The direction of fish movement up and downstream was dependent on the tidal cycle
(χ2 = 5462.3; p < 0.001). The mean percentage of upstream movements made with the incoming
tide was 95%, while the mean percentage of downstream movements made with the outgoing tide
was also 95%.
3.2. Effect of environmental variables on fish movement

All the environmental variables (salinity, temperature, and turbidity) measured at each spotted
grunter location were significantly correlated to each other. In the first period, the strongest
correlation was between salinity and temperature (p < 0.001; r = −0.61; r2 = 0.38), followed by
salinity and turbidity (p < 0.001; r = −0.56; r2 = 0.31), and temperature and turbidity (p < 0.001;
r = 0.50; r2 = 0.25). In the second period, the strongest correlation was also between salinity and
temperature (p < 0.001; r = −0.81; r2 = 0.66), followed by salinity and turbidity (p < 0.001; r = −0.73;
r2 = 0.54), and temperature and turbidity (p < 0.001; r = 0.61; r2 = 0.37).

The results from the linear model showed that the change in salinity (p < 0.001; W(1) = 48.76),
temperature (p < 0.01; W = 10.58), turbidity (p < 0.01; W = 9.58), and tide (p < 0.05; W = 4.74) from
time t to t + 1 had a significant effect on the change in fish position from time t to t + 1. There was no
significant difference between season (summer and spring) (p = 0.78; W = 0.08) and fish size
(p = 0.92, W = 0.01). Therefore, the relative change (∆) in spotted grunter position in the estuary
from time t to t + 1 was determined by the positive relative change in temperature and turbidity, and
negative change in salinity and tide, from time t to t + 1. This is described by the equation:

∆Fish position=0.01−0.06∆Salinity+0.16∆Temperature+0.003∆Turbidity−0.16∆Tide

Partial correlation values revealed that after controlling for salinity, turbidity, temperature and tide,
the relationship between these variables remained constant.

4. Discussion

The conservation of biodiversity and management of aquatic environments in particular has become
a major concern in recent years. The linear nature of most estuaries, and their high degree of
linkage with freshwater and marine ecosystems, makes estuarine habitats highly vulnerable to
external perturbations ([Whitfield, 1998] and [Cattrijse et al., 2002]). It was noted more than a
decade ago that anthropogenic activities could lead to the periodic or permanent elimination of
estuarine-dependent fish species from individual estuarine systems ([Cyrus, 1991], [Peterson et al.,
2000] and [Kennish, 2002]). Such activities include poor farming practices which result in siltation
(extremely high turbidity levels) and construction of impoundments in the catchment which result in
freshwater abstraction (salinity extremes and hypersaline conditions). Since the adults of many
estuarine-dependent species are exploited commercially, the preservation of estuarine habitats is
critical for the maintenance of many marine fisheries (Lenanton and Potter, 1987). Therefore,
knowledge of the response of estuarine fishes to changes in environmental conditions will not only
enhance our biological understanding of estuarine fish, but will contribute to our understanding of
the potential affects of anthropogenic impacts on estuarine fish species.

Salinity has been viewed as one of the most important variables influencing the utilisation of
organisms in estuaries (Marshall and Elliot, 1998). Spotted grunter are euryhaline (Whitfield, 1980)
and have been found to tolerate salinities from 0 to 90 (Whitfield et al., 1981). In this study, spotted
grunter were located in a wide range of salinities (ranging from 0 to 36). The variation in the mean
salinity between the first and second periods can be ascribed to the large proportion of fish located
in the freshwater upper reaches of the estuary during the initial stages of the second period. Ter
Morshuizen et al. (1996) and Bate et al. (2002) suggested that the high conductivity levels of the
Great Fish Estuary, in comparison to the upper reaches of other Eastern Cape estuaries, promotes
the utilisation of the upper and head regions of the estuary by euryhaline fish species, such as the
spotted grunter. The higher mean conductivity in the second period, when compared to the first,
could have facilitated the utilization of the upper reaches of the Great Fish Estuary by spotted
grunter in the second period.

Temperature has been identified as the primary abiotic factor controlling key physiological,
biochemical and life-history processes of fish (Beitinger and Fitzpatrick, 1979), and has been found
to influence the utilisation of estuaries by fishes worldwide ([Morin et al., 1992], [Thiel et al., 1995],
[Baldwin et al., 2002] and [Harrison and Whitfield, 2006]). Generally, fish have a thermal preference
that optimizes physiological processes. Spotted grunter were however located in a wide range of
temperatures during both periods. The large variation in water temperature, between and within
both periods, was due to the large tidal fluctuation in the estuary, with cold incoming seawater and
warm outgoing freshwater. Although the mean water temperature at which spotted grunter were
located was 23 °C in the first and 20 °C in the second period, the thermal preference of 0+ juveniles
under culture conditions was found to be between 24 and 25 °C (Deacon and Hecht, 1995). Lower
temperatures are likely to reduce metabolism and growth. The very low sea temperatures (<16 °C)
recorded at the beginning of the second period (unpublished data) may account for the noticeable
peak in spotted grunter position observed in the upper reaches of the estuary during this period.
The upper reaches may have provided a thermal refuge and caused the fish to move to this region
where they maintained position for an extended period (10–14 days). This suggests that spotted
grunter may use movement in response to temperature variations within the Great Fish Estuary.

Turbidity has also been found to influence the utilization of estuaries by fishes ([Blaber and Blaber,
1980], [Blaber, 1981], [Cyrus, 1992], [Akin et al., 2005] and [Bennett et al., 2005]). Spotted grunter
were found in both exceptionally clear and turbid waters. Field sampling and laboratory experiments
have also shown that spotted grunter are indifferent to turbidity ([Cyrus and Blaber, 1987], [Hecht
and van der Lingen, 1992] and [Whitfield et al., 1994]), probably because they are macrobenthic
carnivores and rely primarily on tactile stimuli when foraging (Whitfield, 1998). The real-time data
collected in this study suggest that spotted grunter are physically adapted to tolerate large variations
in turbidity.

The large tidal prism in the Great Fish Estuary creates wide fluctuations in environmental conditions
during a single tidal phase. During low tide, conditions in the lower reaches of the estuary were
characterized by reduced salinity, increased turbidity and warmer freshwater (particularly in
summer). During high tide, conditions in the lower reaches are characterized by high salinity, low
turbidity and cooler sea water. Spotted grunter were found in the lower reaches during the outgoing
and low tides, and in the middle and upper reaches during the incoming and high tides. Continuous
24-h data collected in the second period by the eight acoustic listening stations corroborate these
findings of the manual tracking data by showing that spotted grunter moved upstream during the
incoming tide and downstream during the outgoing tide. Although it was found that spotted grunter
have a broad physico-chemical tolerance, this information suggests that spotted grunter may follow
a particular suite of environmental variables while making use of the strong currents of the Great
Fish Estuary. Such behaviour may alleviate the physiological demands (e.g. thermoregulatory
stress) that are placed on fishes that occupy estuarine systems (Harrison and Whitfield, 2006) and
consequently minimize energy expenditure. The use of tidal currents for movement thereby
minimizing energy expenditure, has been suggested by other authors studying the movements of
adult thin-lipped grey mullet Liza ramado (Risso 1810) (Almeida, 1996), American eel Anguilla
rostrata (Lesueur 1817) (Helfman et al., 1983), and salmon smolts Oncorhynchus kisutch (Walbaum
1792) (Miller and Sadro, 2003). Although fish are likely to occupy positions in an estuary that
optimises their physiological needs, Matthew (1990) suggests that strong selection also exists for
animals to occupy areas of optimal resource availability. For example, the major food source of
spotted grunter, the mud prawn Upogebia africana (Ortmann 1894) are most abundant and
concentrated in the muddy intertidal lower reaches of the Great Fish Estuary (Hecht and van der
Lingen, 1992). This mirrors with the noticeable peaks observed in the position of spotted grunter in
the lower reaches during both periods. Furthermore, while the influence of the tidal cycle on the
feeding intensity of spotted grunter is not known, optimal foraging theory (McArthur and Pianka,
1966) suggests that these fish would feed when prey is most readily available to them. Hill (1981)
showed that at low tide, mud prawns move to the air–water interface of their burrows. It is therefore
possible that these prey items are more vulnerable at low tide and that spotted grunter would
concentrate their feeding effort on the submerged mud banks at low tide. Since spotted grunter
were mostly found in the lower reaches of the estuary during low tide, this study provides some
circumstantial evidence for this hypothesis.

Although this study has found that spotted grunter are tolerant to large fluctuations in salinity and
turbidity, and to a fairly wide range in temperature during the daily tidal phases, the results also
indicate that a change in tide and the subsequent changes in salinity, temperature and turbidity do
cause a change in position of spotted grunter in the estuary. Season had no effect on the relative
change in position of the fish, despite the decrease in temperature and turbidity observed during the
second period. However, low temperatures (≤16 °C), typical of the spring season, may have
induced a behavioural response in spotted grunter to evade low temperatures by moving upstream.
The size of fish had no effect on the relative change in position of the fish and even the smaller
individuals (whose distribution was confined to the lower reaches of the estuary in the first period)
altered their position with changes in environmental conditions. Spotted grunter may well be
physiologically adapted to survive in this wide range of environmental conditions, but it appears that
they use movement and selective use of tidal currents to minimise energy expenditure and remain
in optimal environmental conditions.

5. Conclusion

The results of this study suggest that the most important environmental factor governing the
movement of the estuarine-dependent spotted grunter in the Great Fish Estuary is tidal phase and
the associated changes in salinity, temperature and turbidity. However, low sea temperatures may
supersede these factors and determine the position of spotted grunter in the estuary. These results
suggest that anthropogenic impacts (construction of dams, water abstraction, and inter-basin
transfer schemes) that are able to cause large environmental fluctuations, particularly in salinity and
turbidity, could influence the movement of spotted grunter in estuarine environments.

Acknowledgements

We would like to thank Rupert Harvey most sincerely for his dedicated support during the fieldwork.
Tia and Hendrik Swart are thanked for providing access to the caravan park and its facilities during
the study. This research was funded by the SA/Norway program on research co-operation (2003–
2005). The collaborating institutions, SAIAB and NINA, are thanked for additional financial and
infrastructure support.
References

Akin, S., Buhan, E., Winemiller, K.O., Yilmaz, H. Fish assemblage structure of Koycegiz Lagoon-
Estuary, Turkey: Spatial and temporal distribution patterns in relation to environmental variation
(2005) Estuarine, Coastal and Shelf Science, 64 (4), pp. 671-684.

Almeida, P.R. Estuarine movement patterns of adult thin-lipped grey mullet, Liza ramada (Risso)
(Pisces, Mugilidae), observed by ultrasonic tracking (1996) Journal of Experimental Marine Biology
and Ecology, 202 (2), pp. 137-150.

Batschelet, E. (1981) Circular Statistics in Biology. Academic Press, London 371 pp.

Baldwin, C.M., Beauchamp, D.A. Seasonal and diel distribution and movement of cutthroat trout
from ultrasonic telemetry (2002) Transactions of the American Fisheries Society, 131 (1), pp. 143-
158.

Bate, G.C., Whitfield, A.K., Adams, J.B., Huizinga, P., Wooldridge, T.H. The importance of the
river-estuary interface (REI) zone in estuaries (2002) Water SA, 28 (3), pp. 271-279.

Beitinger, T.L., Fitzpatrick, L.C. Physiological and ecological correlates of preferred temperature in
fish (1979) American Zoology, 19 (1), pp. 319-329.

Bennett, W.R., Edmondson, G., Lane, E.D., Morgan, J. Juvenile white sturgeon (Acipenser
transmontanus) habitat and distribution in the Lower Fraser River, downstream of Hope, BC,
Canada (2005) Journal of Applied Ichthyology, 21 (5), pp. 375-380.

Blaber, S.J.M. The zoogeographical affinities of estuarine fishes in south-east Africa. (1981) South
African Journal of Science, 77 (7), pp. 305-307.

Blaber, S.J.M., Blaber, T.G. Factors affecting the distribution of juvenile estuarine and inshore fish
(1980) Journal of Fish Biology, 17, pp. 143-162.

Cattrijse, A., Codling, I., Conides, A., Duhamel, S., Gibson, R.N., Hostens, K., Mathieson, S., (...),
McLusky, D.S. Estuarine development/habitat restoration and re-creation and their role in estuarine
management for the benefit of aquatic resources (2002) Fishes in Estuaries, pp. 266-321. Cited 6
times. Elliot M., and Hemingway K.L. (Eds), Blackwell Science Ltd, UK

Cyrus, D.P. Fish conservation in South African estuaries: pressures, problems and prospects (1991)
South African Journal of Aquatic Science, 17, pp. 19-27.

Cyrus, D.P. Turbidity gradients in two indo-pacific estuaries and their influence on fish distribution
(1992) South African Journal of Aquatic Science, 18 (1-2), pp. 51-63.

Cyrus, D.P., Blaber, S.J.M. The influence of turbidity on juvenile marine fish in the estuaries of
Natal, South Africa. (1987) Continental Shelf Research, 7 (11-12), pp. 1411-1416.

Deacon, N., Hecht, T. Observations on the thermoregulatory behaviour of juvenile spotted grunter,
Pomadasys commersonnii (Haemulidae: Pisces) (1995) Journal of Applied Ichthyology, 11 (1-2),
pp. 100-110.
De Solla, S.R., Bonduriansky, R., Brooks, R.J. Eliminating autocorrelation reduces biological
relevance of home range estimates (1999) Journal of Animal Ecology, 68 (2), pp. 221-234.

Dunn, J.E., Gipson, P.S. Analysis of radio telemetry data in studies of home range (1977)
Biometrics, 33 (1), pp. 85-101.

Grafen, A., Hails, R. (2002) Modern Statistics for the Life Sciences. Oxford University Press Inc.,
New York

Grange, N., Whitfield, A.K., De Villiers, C.J., Allanson, B.R. The response of two South African east
coast estuaries to altered river flow regimes (2000) Aquatic Conservation: Marine and Freshwater
Ecosystems, 10 (3), pp. 155-177.

Harrison, T.D., Whitfield, A.K. Temperature and salinity as primary determinants influencing the
biogeography of fishes in South African estuaries (2006) Estuarine, Coastal and Shelf Science, 66
(1-2), pp. 335-345.

Hartill, B.W., Morrison, M.A., Smith, M.D., Boube?e, J., Parsons, D.M. Diurnal and tidal movements
of snapper (Pagrus auratus, Sparidae) in an estuarine environment (2003) Marine and Freshwater
Research, 54 (8), pp. 931-940.

Helfman, G.S., Stoneburner, D.L., Bozeman, E.L., Christian, P.A., Whalen, R. Ultrasonic telemetry
of American eel movements in a tidal creek (1983) Transactions of the American Fisheries Society,
112, pp. 105-110.

Hecht, T., van der Lingen, C.D. Turbidity-induced changes in feeding strategies of fish in estuaries
(1992) South African Journal of Zoology, 27 (3), pp. 9-107.

Hill, B. Respiratory adaptations of three species of Upogebia (Thalassinidea, Crustacea) with

special reference to low tide periods (1981) Biological Bulletin, 160 (2), pp. 272-279.

Houde, E.D., Rutherford, E.S. Recent trends in estuarine fisheries: predictions of fish production
and yield (1993) Estuaries, 16 (2), pp. 161-176.

Kelly, J.T., Klimley, A.P., Crocker, C.E. Movements of green sturgeon, Acipenser medirostris, in the
San Francisco Bay estuary, California (2007) Environmental Biology of Fishes, 79 (3-4), pp. 281-
295.

Kennish, M.J. Environmental threats and environmental future of estuaries (2002) Environmental
Conservation, 29 (1), pp. 78-107.

Kerwath, S.E., Go?tz, A., Cowley, P.D., Sauer, W.H.H., Attwood, C. A telemetry experiment on
spotted grunter Pomadasys commersonnii in an African estuary (2005) African Journal of Marine
Science, 27 (2), pp. 389-394.

Lamberth, S.J., Turpie, J.K. The role of estuaries in South African fisheries: Economic importance
and management implications (2003) African Journal of Marine Science, 25, pp. 131-157.

Lenanton, R.C.J., Potter, I.C. Contribution of estuaries to commercial fisheries in temperate

Western Australia and the concept of estuarine dependence. (1987) Estuaries, 10 (1), pp. 28-35.
Marshall, S., Elliott, M. Environmental influences on the fish assemblage of the Humber estuary,
U.K. (1998) Estuarine, Coastal and Shelf Science, 46 (2), pp. 175-184.

Matthews, K.R. An experimental study of the habitat preferences and movement patterns of copper,
quillback, and brown rockfishes (Sebastes spp.) (1990) Environmental Biology of Fishes, 29 (3), pp.
161-178.

McArthur, R.H., Pianka, E.R. On optimal use of a patchy environment (1966) The American
Naturalist, 100 (916), pp. 603-609.

McCullagh, P., Nelder, J.A. Log-linear models (1995) Generalized Linear Models McCullagh P., and
Nelder J.A. (Eds), Chapman and Hall, London

Miller, B.A., Sadro, S. Residence time and seasonal movements of juvenile coho salmon in the
ecotone and lower estuary of Winchester Creek, South Slough, Oregon (2003) Transactions of the
American Fisheries Society, 132 (3), pp. 546-559.

Morin, B., Hudon, C., Whoriskey, F.G. Environmental influences on seasonal distribution of coastal
and estuarine fish assemblages at Wemindji, eastern James Bay (1992) Environmental Biology of
Fishes, 35 (3), pp. 219-229.

Næsje, T.F., Childs, A.-R., Cowley, P.D., Potts, W.M., Thorstad, E.B., Økland, F. Movements of
undersized spotted grunter (Pomadasys commersonnii) in the Great Fish Estuary, South Africa:
Implications for fisheries management (2007) Hydrobiologia, 582 (1), pp. 25-34.

Peterson, C.H., Summerson, H.C., Thomson, E., Lenihan, H.S., Grabowski, J., Manning, L., Micheli,
F., Johnson, G. Synthesis of linkages between benthic and fish communities as a key to protecting
essential fish habitat (2000) Bulletin of Marine Science, 66 (3), pp. 759-774.

Potter, I.C., Beckley, L.E., Whitfield, A.K., Lenanton, R.C.J. Comparisons between the roles played
by estuaries in the life cycles of fishes in temperate Western Australia and Southern Africa (1990)
Environmental Biology of Fishes, 28 (1-4), pp. 143-178.

Shervette, V.R., Ibarra, N., Gelwick, F. Influences of salinity on growth and survival of juvenile
pinfish Lagodon rhomboides (Linnaeus) (2007) Environmental Biology of Fishes, 78 (2), pp. 125-
134.

Smith, M.M., Heemstra, P.C. (2003) Smiths' Sea Fishes Struik Publishers, Cape Town, South Africa
1047 pp.

Swihart, R.K., Slade, N.A. Testing for independence of observations in animal movements. (1985)
Ecology, 66 (4), pp. 1176-1184.

Szedlmayer, S.T., Able, K.W. Ultrasonic telemetry of age-0 summer flounder, Paralichthys dentatus,
movements in a southern New Jersey estuary (1993) Copeia, 1993 (3), pp. 728-736.

Taverny, C., Lepage, M., Piefort, S., Dumont, P., Rochard, E. Habitat selection by juvenile
European sturgeon Acipenser sturio in the Gironde estuary (France) (2002) Journal of Applied
Ichthyology, 18 (4-6), pp. 536-541.
Ter Morshuizen, L.D., Whitfield, A.K., Paterson, A.W. Distribution patterns of fishes in an Eastern
Cape estuary and river with particular emphasis on the ebb and flow region (1996) Transactions of
the Royal Society of South Africa, 51, pp. 257-280.

Thiel, R., Sepulveda, A., Kafemann, R., Nellen, W. Environmental factors as forces structuring the
fish community of the Elbe Estuary (1995) Journal of Fish Biology, 46 (1), pp. 47-69.

Wallace, J.H. The estuarine fishes of the east coast of South Africa. I. Species composition and
length distribution in the estuarine and marine environments. II. Seasonal abundance and
migrations (1975) Investigational Report of the Oceanographic Research Institute, 40, pp. 1-72.

Wallace, J.H. The estuarine fishes of the east coast of South Africa. III. Reproduction (1975)
Investigational Report of the Oceanographic Research Institute, 41, pp. 1-51.

Wallace, J.H., Kok, H.M., Beckley, L.E., Bennett, B., Blaber, S.J.M., Whitfield, A.K. South African
estuaries and their importance to fishes. (1984) South African Journal of Science, 80 (5), pp. 203-
207.

Whitfield, A.K. A quantitative study of the trophic relationships within the fish community of the
Mhlanga Estuary, South Africa (1980) Estuarine and Coastal Marine Science, 10, pp. 417-435.

Whitfield, A.K. Abundance of larval and O+ juvenile marine fishes in the lower reaches of 3 southern
African estuaries with differing freshwater inputs (1994) Marine Ecology Progress Series, 105 (3),
pp. 257-268.

Whitfield, A.K. Fish species diversity in southern African estuarine systems: An evolutionary
perspective (1994) Environmental Biology of Fishes, 40 (1), pp. 37-48.

Whitfield, A.K. An estuary-association classification for the fishes of southern Africa (1994) South
African Journal of Science, 90 (7), pp. 411-417.

Whitfield, A.K. Biology and ecology of fishes in southern African Estuaries. Ichthyological
Monographs of the J.L.B (1998) Smith Institute of Ichthyology, 2, pp. 1-223.

Whitfield, A.K., Blaber, S.J.M., Cyrus, D.P. Salinity ranges of some southern African fish species
occurring in estuaries (1981) South African Journal of Zoology, 16, pp. 151-155.

Whitfield, A.K., Paterson, A.W., Bok, A.H., Kok, H.M. A comparison of the ichthyofaunas in two
permanent open eastern Cape estuaries (1994) South African Journal of Zoology, 29 (2), pp. 175-
185.

Yule, G.U. On the theory of correlation for any number of variables treated by a new system of
notation (1907) Proceedings of the Royal Society. Series A, 79, pp. 182-193.