Content uploaded by Gavin Snow
Author content
All content in this area was uploaded by Gavin Snow on Feb 02, 2017
Content may be subject to copyright.
CONTRIBUTIONS TO THE USE OF MICROALGAE IN
ESTUARINE FRESHWATER RESERVE DETERMINATIONS
Submitted in fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
(SCIENCE)
of
Nelson Mandela Metropolitan University
by
GAVIN CHARLES SNOW
December 2007
2
Estuaries are spiritual and physical meeting places between man and nature.
Their health reflects our health and they too have moods.
This picture of Knysna Estuary
1
was taken shortly after the August 2006 flood.
1
Photo taken by Ms Anusha Rajkaran (NMMU).
3
Promoter:
Prof. J.B. Adams Botany Department, PO Box 77000, Nelson Mandela Metropolitan
University, South Africa, 6031.
Co-promoter:
Prof. G.J.C. Underwood Department of Biological Sciences, University of Essex, Wivenhoe
Park, Colchester, Essex, CO4 3SQ, England.
The research described in this thesis was conducted at the Botany Department,
Nelson Mandela Metropolitan University, Port Elizabeth.
4
Acknowledgements
Many people and organizations have provided advice, finance and support
throughout the period of study and I am indebted to all of you. A big thank you to my
promoters Prof. Janine Adams and Prof. Graham Underwood, as well as to Prof. Guy
Bate, who have continually gone the extra mile for me; I have and will continue to
learn so much from you.
Research, bursary and travel funds were provided by UPE / NMMU, National
Research Foundation (NRF), Water Research Commission (WRC) and British
Council (Association of Commonwealth Universities).
Thanks to all of the staff and students in the Botany Department at UPE /
NMMU and the Biological Sciences Department at University of Essex who assisted
me in the laboratory and in the field; I would still be stuck in the mud without you. To
Ms Pat Smailes, thank you for the assistance and an endless source of enthusiasm
with regards to counting and identifying microalgal cells.
Thanks to Dr Angus Paterson (formerly of CES) and Dr Steve Mitchell (WRC)
for arranging finance, on such short notice, for the post-flood study of Knysna and
Swartvlei estuaries and to the South African National Parks for the assistance.
To my family and friends, thank you for all of the encouragement you have
given me that has helped me to reach this point.
To my wife Berny, I really appreciate the sacrifices made during my studies and
the continuous support you have provided, I could not have wished for a more
devoted partner. To Payton and Heather, my special daughters born during this
study, this is just the start of an exciting journey together.
5
Contents
Page
Acknowledgements 4
List of abbreviations 6
List of tables 7
List of figures 9
Chapter 1 General Introduction. 18
Chapter 2 Physico-chemical factors determining the distribution of benthic
microalgae and carbohydrates in permanently open South
African estuaries.
24
Chapter 3 Response of microalgae in the Kromme Estuary to managed
freshwater inputs.
57
Chapter 4 Responses of microalgae in the Berg Estuary to reduced
freshwater flow.
80
Chapter 5 Post-flood recovery in an estuarine embayment and an
estuarine lake, South Africa.
112
Chapter 6 Water quality in South African temporarily open/closed
estuaries: A conceptual model.
146
Chapter 7 Relating microalgal spatial patterns to flow, mouth and nutrient
status in the temporarily open/closed Mngazi Estuary, South
Africa.
180
Chapter 8 The value of microalgae in the determination of the freshwater
requirements of South African estuaries.
206
References 224
Appendix 246
6
Abbreviations
AFDW - Ash-Free Dry Weight
ANOVA - Analysis of Variance
BOD - Biological Oxygen Demand
Chl a - Chlorophyll a
CSIR - Council for Scientific and Industrial Research
DIN - Dissolved Inorganic Nitrogen (nitrate + nitrite + ammonium)
DIP - Dissolved Inorganic Phosphorus (SRP)
DO - Dissolved Oxygen
DSi - Dissolved Silicate
DWAF - Department of Water Affairs and Forestry
EC - Electrical Conductivity
EPS - Extracellular Polymeric Substances (exopolymers)
HPLC - High Performance Liquid Chromatography
ICOLLs - Intermittently Closed and Open Lakes and Lagoons
LMW - Low Molecular Weight Carbohydrates
MPB - Microphytobenthos (benthic microalgae)
MSL - Mean Sea Level
NTU - Nephelometric Turbidity Units
ORP - Redox Potential
OM - Organic Matter
PCA - Principal Components Analysis
POE - Permanently Open Estuaries
ppt - Parts per thousand (measure of salinity, often reported as ‰)
REI - River-Estuary Interface
RDM - Resource Directed Measures
SEM - Standard Error of the Mean (or SE)
SRP - Soluble Reactive Phosphorus
TDS - Total Dissolved Solids
TOCE - Temporarily Open/Closed Estuaries
TOxN - Total Oxidised Nitrogen (nitrate + nitrite)
TSS - Total Suspended Solids
7
List of tables
Table
Description Page
2.1. Sampling dates, estuary coordinates and number of sites sampled
in each estuary (ordered from west to east).
31
2.2. Pearson’s correlation coefficients relating sediment associated
variables and water chemistry to benthic chl a and carbohydrates in
six permanently open estuaries (N = 290). Coefficients that are
significantly correlated (P < 0.01) are in bold.
43
2.3. Pearson’s correlation coefficients (r) relating sediment associated
variables and water chemistry to benthic chl a and carbohydrates in
the Swartkops Estuary (N = 90). Coefficients that are significantly
correlated (P < 0.01) are in bold.
49
2.4. Microphytobenthic chl a ranges from intertidal sediments in
different estuaries.
52
3.1. Average salinity and phytoplankton chl a (± standard error) during
this study (italics) and previous studies (Snow et al. 2000a & b;
Snow 2000; Scharler et al. 1997) in the Kromme and nearby
Gamtoos Estuary.
73
3.2. Average intertidal chl a during this study (italics) and studies of
other South African estuaries (Snow, unpublished data).
74
3.3. Proposed future run-off scenarios indicating flow releases from the
Mpofu Dam and the annual volume of river inflow as a percentage
of the reference / natural condition.
76
4.1. The influence of flow rate on the concentration of chl a in the
Gamtoos Estuary. (after Snow 2000).
83
4.2. List of benthic diatom taxa collected in August 2005. The relative
abundances of dominant species (dominance) and total number of
species (No. taxa) at each site are included.
103
4.3 List of benthic diatom taxa collected in November 2005. The
relative abundances of dominant species (dominance) and total
number of species (No. taxa) at each site are included.
104
8
Table
Description Page
4.4. Phytoplankton chl a ranges published for South African estuaries
(modified from Adams et al. 1999). Western Cape estuaries are
marked with an asterisk.
106
5.1. Coordinates of sampling stations with distance (km) from the
mouths of the Knysna and Swartvlei estuaries. Station numbers are
shown in Fig. 5.1.
121
5.2. Average (± SEM) temperatures, pH’s and attenuation coefficients
(K
d
) in the Knysna and Swartvlei estuaries following the 2 August
2006 flood.
127
6.1. Brief descriptions of six estuaries that were compared to the
conceptual model.
173
7.1. Average temperature, dissolved oxygen (DO), secchi depth and
average phytoplankton chl a in the Mngazi Estuary from June 2002
to June 2005 (± standard error).
192
8.1. Data requirements on microalgae for the Intermediate
Determination of RDM in estuaries (DWAF 1999).
213
8.2. Classification, based on Shipunov 2007, and brief description of
phytoplankton groups used in the Estuarine Freshwater Reserve
method.
215
8.3. Classification scheme of median phytoplankton chl a obtained
using intermediate RDM methods.
218
8.4. Intertidal benthic microalgal biomass classification scheme based
on median chl a contents obtained using the intermediate RDM
method.
219
8.5. Intertidal benthic chl a, phytoplankton chl a and water-column
nutrient ranges in the Mhlanga and Mdloti estuaries and averages
(± SD) in the Colne Estuary.
220
9
List of figures
Figure
Description Page
2.1 Locality map of the Keurbooms, Gamtoos, Swartkops, Sundays,
Mngazana and Mngazi estuaries indicating the approximate
locations of sampling stations in the study.
32
2.2 Box-whisker plots of water-column nutrients, ash-free dry weight
(AFDW), sediment moisture content and benthic chl a in the
Keurbooms (Kb = 25/08/2002), Gamtoos (Gt1 = 21/02/2003; Gt2 =
07/08/2002), Swartkops (Sk1 = 30/10/2001; Sk2 = 12/02/2002; Sk3
= 05/12/2002; Sk4 = 15/08/2003), Sundays (Sd1 = 25/07/2002;
Sd2 = 19/02/2003), Mngazana (Ma1 = 25/01/2003; Ma2 =
22/06/2003) and Mngazi (Mi = 26/01/2003) estuaries in the study.
The line within the box, the boundaries of the box and the whiskers
represent the median, 25
th
and 75
th
percentiles and the 10
th
and
90
th
percentiles respectively.
37
2.3 Stacked columns comparing the relative contributions of sediment
particle size (> 125 µm and < 125 µm grain sizes) and organic
content in the intertidal sediments of six open South African
estuaries.
39
2.4 Scatter plot of average sample scores (n = 5) on principal
components 1 and 2, with variability scores, from a PCA of the
physical and chemistry variables of the sediment and overlying
water. Estuaries are colour coded and environmental vectors
included as arrows.
41
2.5 Benthic chl a content (µg g
-1
) in relation to chl a concentration (mg
m
-2
) in the Gamtoos, Sundays, Swartkops and Keurbooms
estuaries (N = 130). A linear regression with associated goodness
of fit (R
2
) included.
46
2.6 Benthic chl a content, expressed as µg Chl a g
-1
freeze-dried
sediment (sediment + organic content) and µg Chl a g
-1
organic
matter content (organic content only), in relation to distance from
the estuary mouths.
47
10
Figure
Description Page
2.7 Sediment chl a and colloidal carbohydrate content in the
Keurbooms, Gamtoos, Swartkops, Sundays, Mngazana and
Mngazi estuaries compared to the model (regression line with 95%
limits) of Underwood & Smith (1998b).
50
3.1 Map of the Kromme Estuary and Geelhoutboom Tributary, showing
distances of the sampling stations (km from the mouth) for the
2004 surveys (modified from Bickerton & Pierce 1988).
61
3.2 Physico-chemical variables measured in the Kromme Estuary (30
July 2004) in relation to distance from the mouth. Arrows indicate
measurements recorded in the Geelhoutboom Tributary. Vertical
bars represent ± SE mean.
65
3.3 Water-column chl a in the Kromme Estuary (30 July 2004). Arrows
indicate measurements recorded in the Geelhoutboom Tributary.
Vertical bars represent ± SE mean.
66
3.4 Intertidal and subtidal benthic chl a in the Kromme Estuary (30 July
2004). Arrows indicate measurements recorded in the
Geelhoutboom Tributary. Vertical bars represent ± SE mean.
67
3.5 Physico-chemical variables measured in the Kromme Estuary (24
November 2003) in relation to distance from the mouth. Arrows
indicate measurements recorded in the Geelhoutboom Tributary.
Vertical bars represent ± SE mean.
68
3.6 Water-column chl a in the Kromme Estuary (24 November 2003).
Arrows indicate measurements recorded in the Geelhoutboom
Tributary. Vertical bars represent ± SE mean.
69
3.7 Intertidal and subtidal benthic chl a in the Kromme Estuary (24
November 2003). Arrows indicate measurements recorded in the
Geelhoutboom Tributary. Vertical bars represent ± SE mean.
70
3.8 Relative abundances of Protista groups and large cyanobacteria in
surface and bottom water in the Kromme Estuary (1.8 and 4.6 km
from mouth) and Geelhoutboom tributary (9.9 and 11.4 km from the
mouth) on 24 November 2003.
71
11
Figure
Description Page
3.9 Cell densities of diatoms, flagellates, dinoflagellates and
chlorophytes in the Mpofu Dam prior to the 2 x 10
6
m
3
release of
water (16 November 1998) (Snow et al., 2000a).
72
4.1 Map of the Berg River Estuary indicating locations of the sampling
sites.
85
4.2 Vertically averaged salinity (ppt) and light attenuation (K) measured
along the longitudinal axis in the Berg Estuary, August 2005.
Vertical bars represent standard error of the means. Symbols
shaded grey represent samples collected in the blind arm, 0.8 km
from the mouth of the estuary.
90
4.3 Vertically averaged salinity (ppt) and light attenuation (K) measured
along the longitudinal axis in the Berg Estuary, November 2005.
Vertical bars represent standard error of the means. Symbols
shaded grey represent samples collected in the blind arm, 0.8 km
from the mouth of the estuary.
91
4.4 Vertically averaged temperature measured along the longitudinal
axis in the Berg Estuary, August and November 2005. Vertical bars
represent standard error of the means. The unshaded symbols
represent measurements in the blind arm, 0.8 km from the mouth
of the estuary.
92
4.5 Total oxidised nitrogen (TOxN), ammonium and soluble reactive
phosphorus (SRP) concentrations measured along the longitudinal
axis in the Berg Estuary, August 2005. Concentrations are the
average of 0.5 m and bottom samples. The symbols shaded light
grey represents measurements in the blind arm, 0.8 km from the
mouth of the estuary.
93
4.6 Total oxidised nitrogen (TOxN) and soluble reactive phosphorus
(SRP) concentrations in relation to salinity in the Berg Estuary,
November 2005 (Anchor Environmental Consultants, unpub. data).
94
12
Figure
Description Page
4.7 Vertically averaged water-column chl a measured along the
longitudinal axis in the Berg Estuary, August and November 2005.
Vertical bars represent standard error of the means. The unshaded
symbols represent measurements in the blind arm, 0.8 km from the
mouth of the estuary.
95
4.8 Filled contour diagrams of phytoplankton group cell density (cell ml
-
1
) in the Berg Estuary, August 2005. Samples were collected at 0,
0.5, 1, 2 and 3 m depths except at 2 m deep in the case of the
mouth site where 3m was not possible.
97
4.9 Vertically averaged total phytoplankton cell density (cell ml
-1
) and
chl a (µg l
-1
) measured in August 2005 along the longitudinal axis of
the Berg Estuary. Vertical bars represent standard error of the
means.
98
4.10 Surface phytoplankton cell density (cell ml
-1
) measured in
November 2005 along the longitudinal axis of the Berg Estuary.
99
4.11 Dominant phytoplankton cells collected in November 2005 (A =
large flagellate, B = small diatom and C = small flagellate).
100
4.12 Filled contour diagram of phytoplankton chl a (µg l
-1
) in the Berg
Estuary, November 2005.
101
4.13 Intertidal and subtidal benthic chl a along the longitudinal axis of
the Berg Estuary, August and November 2005.
102
5.1 Outline of the (A) Knysna and (B) Swartvlei estuaries. Positions of
sampling stations are indicated on the plots and the inset shows
the location of Knysna and Swartvlei on the south coast of South
Africa.
119
13
Figure
Description Page
5.2 Rainfall, flow and height above geodetic mean sea level (MSL) for
the Knysna and Swartvlei estuaries during the period August 2005
to October 2006: (A) Daily rainfall (mm) measured at George, a
town ~30 and 60 km west of Swartvlei and Knysna estuaries
respectively, (B) Knysna River flow measured in the Millwood
Forest Reserve (~33% of total flow), (C) Karatara River flow
(tributary into Swartvlei; ~6.5% of total flow), and (D) mean height
above sea level (m) measured in the Swartvlei Lake.
124
5.3 Longitudinal salinity distribution in the Knysna Estuary 8, 19 and 53
days after the 2 August 2006 flood. Measurements were taken from
3.8 km to 16.9 km from the mouth of the estuary.
126
5.4 Longitudinal DO distribution in the Knysna Estuary following the 2
August 2006 flood. The caps at the end of each box indicate the
extreme values (minimum and maximum), the box defines the
lower and upper quartiles and the line in the centre of the box is the
median.
128
5.5 Longitudinal nutrient concentrations (µM) in the Knysna Estuary
following the 2 August 2006 flood. Standard errors are represented
by vertical bars.
131
5.6 The ratios between dissolved inorganic nitrogen (DIN) and
dissolved inorganic phosphorus (DIP); DIN: DIP, following the 2
August 2006 flood in the Knysna estuary.
132
5.7 Longitudinal phytoplankton chl a in the Knysna Estuary following
the 2 August 2006 flood. Standard errors are represented by
vertical bars.
133
5.8 Longitudinal salinity distributions in the Swartvlei Lake and Estuary
following the 2 August 2006 flood. The x-axes are split to represent
salinity contours along the two mouth-tributary axes (Wolwe and
Hoëkraal Rivers).
134
14
Figure
Description Page
5.9 DO along the longitudinal axis of the Swartvlei Estuary and Lake
following the 2 August 2006 flood. The caps at the end of each box
indicate ± standard deviation, the box represents 25% and 75%
quartiles and the line in the box the median.
136
5.10 Longitudinal nutrient concentrations (µM) in the Swartvlei Lake (7.3
to 11.9 km from mouth) and Estuary (0.6 to 6.2 km from mouth)
following the 2 August 2006 flood. Standard error represented by
vertical bars. The points where sites in the Wolwe and Hoëkraal
River plots overlap is highlighted ().
138
5.11 The ratios between dissolved inorganic nitrogen (DIN) and
dissolved inorganic phosphorus (DIP), DIN: DIP, following the 2
August 2006 flood in the Swartvlei system. The points where sites
in the Wolwe and Hoëkraal River plots overlap is highlighted ().
138
5.12 Longitudinal water-column chlorophyll a (µg l
-1
) in the Swartvlei
Lake (7.3 to 11.9 km from mouth) and Estuary (0.6 to 6.2 km from
mouth) following the 2 August 2006 flood. Standard error
represented by vertical bars. The points where sites in the Wolwe
and Hoëkraal River plots overlap is highlighted ().
140
6.1 Diagrammatic representation of the three dominant states in which
TOCE mouths can exist; open, semi-closed and closed.
150
6.2 Normal salinity conditions in a TOCE when the mouth is open to
both seawater and riverwater exchange.
151
6.3 Vertically-stratified salinity conditions becoming mixed by wind to
become homogeneous brackish in the semi-closed mouth state.
152
6.4 The closed-mouth condition when salinity is nearly homogeneous
throughout the water-column. Some vertical and longitudinal
stratification may be evident immediately after closure. Overwash
events may introduce more saline water to deeper reaches of the
estuary.
153
15
Figure
Description Page
6.5 Schematic illustration of the progressive change expected in
dissolved oxygen characteristics in a TOCE under the semi-closed
state.
156
6.6 A simplified representation of the main potential sources of
inorganic nutrients to estuaries.
157
6.7 Hypothetical relationship between salinity and inorganic nutrient
concentrations during the open-mouth state.
159
6.8 A hypothetical relationship between water-column chl a and
inorganic nutrient concentrations as a function of time during the
semi-closed mouth state.
160
6.9 Hypothetical relationship between water-column chl a and
inorganic nutrient concentrations as a function of time in the
closed-mouth state.
161
6.10 Map of South Africa showing locations of estuaries that were
compared to the conceptual TOCE model.
162
7.1 Outline of the Mngazi Estuary with the distances of the sampling
stations indicated on the plot. The inset shows the location of the
estuary on the east coast of South Africa.
185
7.2 Average daily flow data for the Mngazi River (T7H001), 01 May
2002 to 30 June 2005 (courtesy of the South African Department of
Water Affairs and Forestry). The gauge monitors runoff from 315
km
2
of the 591 km
2
catchment. Horizontal bar at top of graph
indicates mouth state (? = unknown, black = closed, hatched =
semi-closed and white = open) and arrows indicate sampling dates.
187
7.3 Salinity profiles in the Mngazi Estuary during the period June 2002
to June 2005. State of the mouth and estimated river flow are
included in each plot.
190
7.4 Ammonium, SRP and TOxN relative to distance from the mouth of
the Mngazi Estuary mouth during semi-closed (June 2002 and
January 2003) and closed (June 2003 and November 2003) mouth
conditions.
193
16
Figure
Description Page
7.5 Dissolved inorganic nitrogen (DIN; TOxN + NH
4+
) and phosphorus
(DIP; SRP) ratios in relation to distance from the mouth, June 2002
to November 2003. The DIN: DIP of 16: 1 is represented by grey
lines.
195
7.6 Phytoplankton chl a in relation to distance from the mouth of the
estuary during the semi-closed (June 2002 and January 2003) and
closed (June 2003 and November 2003) mouth conditions.
Standard error represented by vertical bars.
196
7.7 Micro- (nitex; >20.0 µm), nano- (GF/D; 2.7 - 20 µm), pico- (GF/C;
1.2 - 2.7 µm) and total phytoplankton chl a (µg l
-1
) relative to
distance from the mouth of the Mngazi Estuary in March and June
2005 (open mouth phase).
198
7.8 Intertidal mud (<125 µm) and organic matter (AFDW) contents in
relation to distance from the Mngazi Estuary mouth (January 2003).
Standard error bars displayed.
199
7.9 Benthic chl a (µg g
-1
) in relation to distance from the mouth (km).
Intertidal and subtidal concentrations given for semi-closed and
open mouth conditions, and shallow (close to estuary bank) and
channel concentrations given for closed mouth conditions.
Standard error represented by vertical bars.
200
7.10 A conceptual model of the transition from open (flooded) to closed
mouth conditions in the Mngazi Estuary. River input influences the
mouth phase, stratification and the import of dissolved and
particulate organic matter, sediments and nutrients (TOxN and DSi
in particular). As river flow decreased, there was a trend towards
the water-column becoming well-mixed, nutrients being mineralised
from organic matter in the sediment and the microphytobenthos
(MPB) dominating primary production (modified from Eyre & Twigg,
1997).
205
8.1 Locations of South African estuaries mentioned in the text (Chapter
8).
208
17
Figure
Description Page
A.1 The Mpofu Dam, approximately 45% full, located approximately 4
km from the tidal head of the Kromme Estuary, Eastern Cape
Province.
246
A.2 The mouth region (open) of the temporarily open/closed Mngazi
Estuary, Eastern Cape Province. Mngazi River Bungalows Resort
is slightly to the right of centre.
246
A.3 The flood-tide delta of the permanently open Mngazana Estuary,
Eastern Cape Province. The mouth and entrance to creek 1 are to
the right of the image.
247
A.4 A well developed sand berm at the mouth region of the temporarily
open/closed Van Stadens Estuary, Eastern Cape Province.
247
A.5 Tannin-stained upper reaches of the Knysna Estuary, Western
Cape Province, approximately 2 km below the Chelmsford Weir (10
August 2006).
247
A.6 Flood debris against a house located in the upper reaches of the
Knysna Estuary, Western Cape Province. (10 August 2006).
248
A.7 Knysna Embayment and heads; picture taken from South African
National Parks offices, Thesen Island (10 August 2006).
248
A.8 Dense beds of Potamogeton pectinatus L., colonising the shallow
littoral zone of the Swartvlei Lake, Western Cape Province (9
August 2006).
249
A.9 Maximum water level recorded during August 2006 floods at Pine
Lake Marina, western shores of Swartvlei Lake (9 August 2006).
249
18
CHAPTER 1
General introduction
19
The ecologist Garrett Hardin (1968) introduced a useful concept called the tragedy of
the commons, which describes how ecological resources become threatened or lost.
The term “commons” is based on the commons of old English villages and is
symbolic of a resource that is shared by a group of people. If every person were to
use each resource in a sustainable fashion it would be available in perpetuity.
However, if people use more than their share they would only increase their personal
wealth to the detriment of others. In addition, an increase in the population would
mean that the size of each share would have to decrease to accommodate the larger
number of people. As a result, resources are threatened by personal greed and
uncontrolled population growth.
Freshwater is an example of a common resource that is under threat in South
Africa where the average annual rainfall is less than 60% of the global average
(Mukheibir & Sparks 2006). The increasing demands for freshwater as well as its
eutrophication are major concerns with regards to estuarine health, environmental
resource management and human health. The correct management of water is
necessary to ensure that it is utilised in a sustainable manner. The National Water
Act (No. 36 of 1998) has provided the rights to water for basic human needs and for
sustainable ecological function; the Basic Human Needs Reserve and Ecological
Reserve are both provided as a right in law.
The amount of water necessary for an estuary to retain an acceptable
ecological status, known as the Estuarine Ecological Reserve, is determined through
the implementation of procedures (rapid, intermediate or comprehensive) compiled
by the Department of Water Affairs and Forestry (1999) in its Resource Directed
Measures (RDM) for the Protection of Water Resources. The impact of restricted flow
on estuaries can be reduced by manipulating the water released from
impoundments, the regulation of water abstractions within the river catchment or both
(Hirji et al. 2002). The reserve assessment method is designed to evaluate
ecosystem requirements by employing groups of specialists from different disciplines.
In South Africa, this includes hydrologists, sedimentologists, water chemists and
biologists (including microalgae specialists). The use of microalgae in ecological
assessments has largely been based on research that was initiated at the Nelson
Mandela Metropolitan University (formerly University of Port Elizabeth) and
subsequently at Rhodes University (Grahamstown) and the University of KwaZulu-
20
Natal (Durban). The microalgal research can be divided into two main focus areas;
phytoplankton and benthic microalgae.
Phytoplankton
The spatial distribution of phytoplankton, which consists of microalgal cells adapted
to life spent suspended in the water-column, is almost totally dependent on water
motion. Local studies (Hilmer & Bate 1990; 1991; Allanson & Read 1995; Grange &
Allanson 1995) have found a strong relationship between phytoplankton biomass,
measured using chlorophyll a (chl a) as an index, and freshwater inflow. Factors such
as circulation patterns, nutrient concentration and turbidity were shown to have an
effect on the spatial distribution, growth rates and species composition of
phytoplankton communities in South African estuaries (Adams et al. 1999). In
addition, research by Hilmer and Bate (1991) found that vertical and longitudinal
salinity gradients, an indication of consistent freshwater input, resulted in
phytoplankton dominance in the Sundays Estuary but this did not extend to estuaries
further west along the south Cape coast, largely due to lower nutrient concentrations
in the latter systems. Drainage from the nutrient-poor Table Mountain quartzite in the
Cape fold mountains results in most southern Cape rivers and estuaries being
oligotrophic in contrast to the Sundays and Gamtoos estuaries, which are enriched
with fertilizer nutrients from agricultural return flow. The maximum phytoplankton
biomass in the Sundays and Gamtoos estuaries was dependent on the flow rate and
a retention time of 3 spring tidal cycles (~42 days) was optimal (Hilmer 1990; Snow et
al. 2000a). The maximum phytoplankton biomass generally occurred at a vertically
averaged salinity of less than 10 ppt and was termed the river-estuary interface zone
(REI). However, further research is required to determine whether an REI can exist in
permanently open estuaries starved of freshwater, in other types of estuaries or in
other biogeographical zones.
Estuaries have been classified into permanently open (POE), temporarily
open/closed (TOCE), estuarine lakes, estuarine bays and river mouths (Whitfield
1992). Although the Whitfield (1992) classification included river mouths, the
definition of an estuary as being one where either there is a tidal influence from the
sea or a measurable salinity gradient, precludes them from this discussion. Estuarine
lakes and bays are the least common estuary types with only 8 and 3 examples
respectively in the country. In addition, the climate along the South African coast
21
ranges from cold temperate on the west coast, warm temperate along the south and
south-east coasts and sub-tropical along the east coast, creating a complex diversity
of estuaries. The research presented in this thesis addresses some of the
relationships between microalgal biomass and abiotic factors in different types of
estuaries.
Benthic microalgae
A large proportion of research on estuarine benthic microalgae, or
microphytobenthos (MPB), has found a poor relationship between microphytobenthic
(MPB) biomass and the chemistry of the overlying water (Snow 2000a). As a result,
further research was necessary to determine whether spatial patterns of MPB could
be related to water chemistry and hydrology or whether the distribution was
determined by biogeochemical parameters in the sediment.
Recent research of European estuaries has described the stabilising effect that
benthic microalgae have on soft-sediment habitats in intertidal and shallow subtidal
marine ecosystems (Paterson & Black 1999; Underwood & Paterson 2003; Stal
2003). This is largely the result of extracellular carbohydrates produced by motile
benthic diatoms, creating a matrix of cells, sediment particles and extracellular
polymeric substances (EPS) (Underwood & Paterson 2003). Du Preez (1996) looked
at EPS production by Anaulus australis, a neritic diatom that is both planktonic and
benthic depending on conditions in the surf-zone. It is known that the nature of the
sediment and the nutrient loading of the overlying water does affect both MPB
biomass and biofilm properties (Underwood 2002; Underwood & Paterson 2003).
However, no work has previously been conducted in South African estuaries, hence
research investigating the spatial patterns of carbohydrate fractions and the potential
sediment stabilising effects in different South African systems was initiated.
Objectives
The basic questions underlying the research described in this thesis were:
1. What factors determine the spatial patterns of microalgal biomass
(phytoplankton and MPB) in South African estuaries?
22
2. How is MPB biomass related to sediment carbohydrates and sediment
stability?
3. How can microalgae be used to determine the freshwater inflow requirements
of estuaries?
Chapters 2, 3 and 4 focus on microalgae in permanently open estuaries. The
factors that determine the spatial patterns of benthic microalgae and their
extracellular carbohydrates in estuaries along the southern and Eastern Cape coasts
are dealt with in Chapter 2. Chapter 3 is a comprehensive freshwater reserve study
of microalgae in the freshwater starved Kromme Estuary and Chapter 4 describes
the spatial patterns of microalgae in the strongly seasonal Berg Estuary in the cool
temperate biogeographic zone.
A flood in August 2006 along the southern Cape coast provided an opportunity
to study the post flood response of phytoplankton in two rare types of estuaries;
Knysna estuarine bay and Swartvlei estuarine lake (Appendix Fig.s A.5-A.9). The
water chemistry and phytoplankton results of the studies that are described in
Chapter 5 contribute to the information used by researchers to determine the
freshwater inflow requirements (or freshwater reserves) for those two systems.
Allanson (2001) discussed the diversity of estuarine types and the individual
responses of these systems to physical and chemical determinants in this highly
dynamic environment. In addition, he highlighted the need for techniques with which
to integrate these features into conceptual and numerical models. Chapter 6 provides
a conceptual model describing the water quality and characteristics of TOCE’s in
South Africa. Thereafter, available literature on estuaries in the warm- and cool-
temperate biogeographic regions of South Africa is reviewed and assessed against
this model. An understanding of the processes was a necessary step to better
understand the spatial patterns of microalgae in TOCE’s, particularly in relation to
mouth condition. Research conducted by Perissinotto (University of KwaZulu-Natal),
Froneman (Rhodes University) and Gama (Nelson Mandela Metropolitan Univesity)
have described the spatial patterns of microalgae in TOCE’s along the KwaZulu-
Natal and Eastern Cape coasts respectively. A comprehensive study of microalgae in
the Mngazi Estuary, related to the mouth phases is presented in Chapter 7. The
Mngazi Estuary is located in the transition zone between warm temperate and sub-
tropical biogeographic zones and the anthropogenic influence on the quality and
23
quantity of the river water is unusually low for South Africa; potentially making the
Mngazi a suitable example of an estuary in its reference state (Appendix Fig. A.2).
24
CHAPTER 2
Physico-chemical factors determining the distribution of benthic microalgae
and carbohydrates in permanently open South African estuaries
25
Abstract
Flocculation is an important process by which suspended sediment, organic matter,
nutrients and microalgal cells are deposited from the water-column to the benthos in
estuaries. The rate and site of deposition are largely functions of the morphology and
flow rates within an estuary. Intertidal sediment from six South African estuaries was
sampled for benthic chlorophyll a (chl a) and carbohydrates with the aim to establish
which environmental variables determine the spatial distribution of benthic microalgal
biomass and associated carbohydrates. Chl a content, based on mass (µg g
-1
) (n =
320), and concentration, based on area (mg m
-2
) (n = 130), were significantly
correlated (r = 0.33, P < 0.01 and r = 0.16, P < 0.01 respectively) to total oxidised
nitrogen (TOxN) in the overlying water. However, the correlation coefficients were
relatively low. The strongest associations were with sediment related variables; ash-
free dry weight (AFDW) (r = 0.53, P < 0.001), % moisture (r = 0.62, P < 0.001), very
fine sand (63-125 µm) (r = 0.31, P < 0.001) and the silt-clay fraction (< 63 µm) (r =
0.23, P < 0.001). The lowest chl a content was below the detectable limit in the
coarse sand (oligotrophic) at the mouth region of the Keurbooms Estuary and the
highest, 104.8 µg g
-1
, was measured in the middle reaches of the Mngazana Estuary.
Chl a content was significantly correlated to chl a concentration in 130 samples that
had a broad range of organic and mud contents (r = 0.87, P < 0.001). The largest
deviation from the linear regression line of these two measures were the muddy sites
in the Gamtoos and Sundays estuaries, probably as a result of higher pore space
and lower bulk density of these sediments.
Exopolymers were strongly associated with nutrients in the Swartkops Estuary,
particularly soluble reactive phosphorus (SRP) (r = 0.357; P < 0.001), where a strong
gradient in SRP was present (ranging from 6.9 µM at the mouth to 30.4 µM near to
the head of the estuary). A PCA of environmental variables separated sampling sites
and estuaries along gradients in sediment type and nutrients (PC1) and sediment
type and organic content (PC2). Chl a and the carbohydrate fractions were most
strongly correlated with PC2. These results suggest that sites with a high organic
content (> 3%) and fine sediment content (< 125 µm sediment contributes more than
20% of the sediment) are most likely to support a high microalgal biomass. Colloidal
carbohydrates are also likely to be high at these sites providing a significant energy
source to bacteria and potentially increasing the stability of the sediment.
26
Introduction
There are five main types of estuarine systems in southern Africa: Permanently open,
temporarily open/closed (TOCE), river mouths, estuarine lakes and estuarine bays.
Permanently open estuaries, the focus of this study, generally have a steady
discharge of riverwater maintaining a longitudinal salinity gradient between the head
and mouth of the estuary (Whitfield 1992). Rivers that meander to the sea form vast
intertidal flats with soft or semi-soft sandy or muddy substrates. These intertidal flats
are ideal shallow water habitats, which become colonised by a wide variety of fauna
and flora. Sediments accumulate and the rich nutrient influx and shallow warm
waters results in a highly productive ecosystem (Lubke & de Moor 1998). Many
South African estuaries such as the Gamtoos Estuary (Snow 2000a) have channel-
like morphologies. As a result, the intertidal zones are steep and narrow restricting
the overall area available for benthic microalgal production and the transfer of
autochthonous energy to higher trophic levels.
The process of flocculation is important as this leads to the sedimentation of
organic matter, fine sediment, microalgal cells and nutrients. The process provides
an allochthonous supply of energy for resident micro- and macrobenthic fauna and
has been the focus of a number of recent studies; van der Lee (2000), Winterwerp
(2002), Jiang et al. (2004), Garcia (2005) and Manning et al. (2006). In general, the
microtidal (< 2 m tidal variance), wave-dominated estuaries along the southern coast
of South Africa have three distinct geomorphological zones (Cooper et al. 1999); a
sandy barrier in which a constricted tidal inlet is formed and landward of which is an
extensive flood-tidal delta. The flood-tidal delta is deposited as a result of the flood
dominance of tidal currents in constricted inlets. Landward of these barrier-
associated environments is a deeper water area typified by fine sediment deposition,
in many instances enhanced by the flocculation of suspended clays in the saline
estuarine water. At their upstream limit such estuaries typically exhibit a fluvial delta
where bottom sediments are coarse and water depths are shallow.
Day (1981) described the distribution of sediments within estuaries and
emphasised that sediments of both marine and terrestrial origin are present in most
estuaries. Normally in an estuary there is a preponderance of fine silt and clay of
fluvial origin at the head of the estuary, grading to medium or coarse sand of marine
origin at the mouth. Both Day (1981) and Dyer (1979) described the process of
flocculation, which occurs in estuaries at a salinity of 1-4 ppt for very fine sediments
27
(< 2 µm) and over the full salinity range for other fine sediment types (e.g.
montmorillonite), and is the cause of the turbidity maximum found in many estuaries.
The turbidity maximum is the zone where suspended particulate matter
concentrations is higher than further upstream or downstream in the estuary and is
usually located near to the point of the saline intrusion (Dyer 1994). Dyer (1979) did
mention that sand and gravel-type sediment moves downstream to the tip of the
saline intrusion in the river-dominated reaches of an estuary. When temperatures
exceed 25 ºC, as frequently occurs during summer in sub-tropical regions, the
phenomenon of flocculation is likely to accelerate resulting in the deposition of large
amounts of mud particles (Jiang et al. 2004).
Benthic microalgal biomass (chlorophyll a)
Microalgae, being primary producers, are important components of estuarine
ecosystems. They influence the exchange of nutrients between the water-column
and the sediment (Rysgaard et al. 1995; Nedwell et al. 1999), provide a food source
for deposit feeding fauna (Davis & Lee 1983; Hillebrand & Kahlert 2002), contribute
to sediment stability (Sullivan 1999; Underwood 2000) and both stimulate and
compete with various bacterial sediment processes (Haynes et al. 2007).
Chlorophytes, euglenoids, cyanobacteria and diatoms, collectively termed
microphytobenthos (MPB), inhabit the top few centimetres of intertidal sediment in
estuarine and coastal sediments worldwide (Cahoon et al. 1999; Underwood &
Kromkamp 1999). The photosynthetic pigment chlorophyll a (chl a) is frequently used
as an measure of MPB biomass and has been used extensively in a number of
studies describing the distribution of benthic microalgae in aquatic ecosystems
(Thornton et al. 2002; Cartaxana et al. 2006; Jesus et al. 2006; Snow & Adams
2006).
Large standard deviations in chl a measurements sampled over a small
distance are common phenomena in intertidal sediments. MacIntyre et al. (1996)
reported on small-scale variation in the distribution of MPB biomass and
recommended that at least five cores were needed to reduce the coefficient of
variance to 45%. Recent studies of the Gamtoos (Snow et al. 2000b) and Kromme
estuaries (Snow et al. 2000a) in South Africa described a relationship between
average benthic chl a and river flow. However, no apparent relationship was evident
between benthic chl a and distance from the mouth of the estuaries, despite the
28
presence of a strong TOxN gradient along the length of the Gamtoos Estuary. Similar
results have been found in other estuaries and coastal waters (Underwood et al.
1998; Perissinotto et al. 2002; Welker et al. 2002; Mundree et al. 2003), suggesting
that the quality of overlying water at the time of measurement was not significantly
related to benthic microalgal chl a. Instead, there was a much stronger correlation
between the microphytobenthos and the sediment, a possible source of nutrients.
However, similar studies have found a closer association between water-column
nutrients and benthic chl a (Sündback 1996) suggesting that the relationship is not
simple and hence further research is required to improve our understanding of this
relationship.
Recent publications (Perkins et al. 2003; Tolhurst et al. 2005) have highlighted a
problem encountered with regards to presenting chl a data as content (mass per unit
mass; e.g. µg Chl a g
-1
) or concentration (mass per unit volume; e.g. mg Chl a m
-3
).
Natural soft marine sediments consist of six components; (1) non-cohesive mineral
grains (sand particles), (2) cohesive particles (fine silt and clay), (3) water, (4) gas,
(5) biota and (6) other matter (detritus, extracellular polymeric substances, heavy
metals and salt) (Tolhurst et al. 2005). The first five are interdependent but the
densities can vary (e.g. sand has a higher density and usually has a lower water
content than fine sediment), which means that a change in one will affect at least one
of the other four. By reporting chl a as a content, chl a is being expressed as a
fraction relative to the remaining components. In contrast, chl a expressed as a
concentration is expressing chl a as an exact amount of chl a in a fixed volume of
sediment. As such, trends in the distribution of chl a can be incorrectly interpreted,
particularly when comparing muddy and sandy sites. Muddy sites tend to have higher
chl a contents than sandy sediments (Tolhurst et al. 2005).
Carbohydrates
The MPB inhabiting fine intertidal sediment is usually dominated by epipelic diatoms,
which move through the sediment by excreting extracellular polymeric substances
(EPS) from the raphe slit present in each of the silica cell walls (valves) that make up
the cell (Stal 2003; Underwood & Paterson 2003). In coarser sediments, the
episammic microphytobenthos is generally dominated by diatoms that either attach
themselves to sand particles by a pad or short stalk of EPS or are capable of
movement by excreting EPS. Cyanobacteria commonly found in fine sandy
29
sediments are also capable of excreting a polysaccharide sheath, which allows for
gliding mobility. Through the remineralisation of organic matter they enrich the
sediment with nutrients, helping to create a suitable environment for diatoms (Stal
2003).
The extracellular carbohydrate exudates of microphytobenthos, or exopolymers,
can be broadly grouped into low molecular weight (LMW), small sugar units and
glycollates, and larger, increasingly polymeric molecules (EPS). Collectively, this
material is termed “colloidal carbohydrate”, and high contents of colloidal
carbohydrates can be present on mudflats (typically between 50 and 5000 µg g
-1
).
The production of colloidal carbohydrates is usually closely related to the rate of
photosynthesis and can also increase significantly when nutrients are limiting,
referred to as the overflow hypothesis (Stal 2003; Underwood & Paterson 2003). The
amounts of colloidal carbohydrate and EPS generally increase over a tidal emersion
period (Hanlon et al. 2006). Although other organisms, such as bacteria, produce
extracellular carbohydrates, colloidal carbohydrates present in microphytobenthic
biofilms are closely related to microalgal biomass and photosynthetic activity (Smith
& Underwood 2000). In diatom-rich biofilms, between 20% and 40% of the
extracellular carbohydrate is polymeric (i.e. EPS). The remaining LMW carbohydrates
are an important food source for bacteria and may play an important role in the
ecology of fine sediments. Bacterial production is usually closely associated with
LMW compounds, resulting in rapid decreases in LMW material in the absence of
photosynthesis, particularly in sandy sediments (van Duyl et al. 1999; Underwood
2002). Microalgal-bacterial coupling is generally strongest in sandy sediments with
low organic content compared to fine mud with a high organic content (Köster et al.
2005). Carbohydrate content is more closely associated to epipelic diatom biomass
than episammic biomass.
Underwood & Smith (1998b) investigated the chl a: colloidal carbohydrate
distribution for a range of European mudflats and published a model describing the
relationship; Log [colloidal carbohydrate (µg g
-1
) + 1] = 1.40 + [log chl a (µg g
-1
) + 1].
This relationship was only found to fit for sediments dominated by fine clay particles
and at sites where epipelic diatoms dominated (> 50%) the microphytobenthic
assemblage. Assemblages dominated by green algae and cyanobacteria did not
show a significant colloidal carbohydrate: chl a relationship. Similar trends have been
30
found in more recent studies (Blanchard et al. 2000; Thornton et al. 2002; Bellinger et
al. 2005).
To date, no published reports of carbohydrates or of their stabilising effects in
the intertidal sediment of South African estuaries have been produced. Du Preez
(1996), however, has presented a considerable amount of information on the EPS of
the diatom Anaulus australis in the surf-zone. This diatom also spends time on/in the
sediment behind the breaker-line in calm weather.
Research aims
The aims of the study were:
1. To determine the relative importance of the water-column chemistry and
sediment characteristics in determining the spatial distribution of
intertidal benthic microalgae, using chl a as an index of biomass, in six
permanently open estuaries along the south-eastern coast of South
Africa.
2. To describe the spatial pattern of carbohydrate fractions (colloidal
carbohydrates in particular) in six South African estuaries and relate
these to physical and chemical variables in the estuary and benthic
microalgal biomass (chl a).
Study location
Six estuaries (Table 2.1) were selected for the study and all except for the
Keurbooms (Western Province) occur in the Eastern Cape Province of South Africa
(Fig. 2.1). All six estuaries are examples of drowned river valley estuaries, being
typically confined within bedrock valleys and influenced, to varying degrees, by
aeolian sediment deposition as many dunes along the coast are unvegetated
(Cooper et al. 1999). There is a shift from a warm temperate climate in the west to
subtropical in the east. Intertidal benthic samples were collected with the result that
only permanently open estuaries could be considered in this study except for the
Mngazi, which closes temporarily and was open for more than a month prior to
sampling. The Swartkops was sampled four times, the Sundays, Mngazana and
31
Gamtoos twice and the Keurbooms and Mngazi estuaries only once because of
floods during March 2003 and mouth closure in June 2003 respectively.
Table 2.1
Sampling dates, estuary coordinates and number of sites sampled in each estuary
(ordered from west to east).
Estuary name Coordinates Date sampled Number of sites
Keurbooms 34
o
02’S; 23
o
23’E 25/08/2002 5
Gamtoos 35
o
58’S; 25
o
01’E 21/02/2003 5
07/08/2002 5
Swartkops 33
o
52’S; 25
o
38’E 12/02/2002 6
05/12/2002 6
15/08/2003 6
30/10/2001 6
Sundays 33
o
43’S; 25
o
51’E 25/07/2002 5
19/02/2003 5
Mngazana 31
o
42’S; 29
o
25’E 22/06/2003 5
25/01/2003 5
Mngazi 31
o
41’S; 29
o
27E 26/01/2003 5
Thirty-one sites were sampled ensuring that a broad range of benthic chl a and
associated environmental variables were collected. Two of the six estuaries had
distinct point-source discharges of high nutrient water. An agricultural return flow pipe
enters the Gamtoos near to the head of the estuary and a polluted stormwater canal
enters from the nearby Motherwell Township into the mid to lower reaches of the
Swartkops Estuary. Water samples were collected from these sources during
sampling trips and analysed for quality. There were six sampling sites in the
Swartkops and five sites in each of the remaining estuaries. Sites were fairly evenly
spaced along the length of the estuaries to capture the salinity gradient in the water-
column and changes in sediment type along their longitudinal axes.
32
Figure 2.1. Locality map of the Keurbooms, Gamtoos, Swartkops, Sundays,
Mngazana and Mngazi estuaries indicating the approximate locations of sampling
stations in the study.
33
Materials and Methods
Microalgal biomass
Five replicate samples of benthic chl a were collected from just above the water line
within a square metre of each other from each site by scraping a known area of
surface sediment (< 2 mm depth) during spring low tide. Scrapes of known areas of
intertidal sediment were necessary to convert chl a content (µg g
-1
) to chl a
concentration (mg m
-2
). The samples were freeze-dried, approximately 0.1 g was
added to 4 ml of 95% ethanol and then stored for 24 hrs at 0
o
C. Once the chl a was
extracted the samples were whirli-mixed then filtered through glass-fibre filters
(Whatman GF/C). The extract was analysed on a high performance liquid
chromatograph (HPLC) attached to a Waters Lambda-Max 481 LC
spectrophotometer and Waters LM-45 solvent delivery system. A 30% methanol /
70% acetone mixture was used as a carrier. The system was calibrated using the chl
a of red seaweed (Plocamium corallorhiza) because it contains no chlorophyll b
which interferes with the chl a reading at 665 nm (du Preez pers. comm.). Chl a was
expressed on a mass per unit mass basis (micrograms of chl a per gram of freeze-
dried sediment (µg g
-1
) and 130 samples were also measured on a mass per unit
area basis (mg m
-2
) to determine the difference between the two.
Fractionation of carbohydrates
Sediments remaining from the freeze-dried replicate samples were analysed for
carbohydrate. The method used was based on the phenol-sulfuric acid assay
(Dubois et al. 1956). This assay is used to measure total carbohydrate in the
sediment (i.e. intracellular, extracellular and particle-bound) by hydrolyzing more
complex carbohydrates into simple carbohydrates using acid and heat. Phenol is
then added to the sample, which binds to the sugars to form fluorogen. The yellow
product is measured using a spectrophotometer at an absorbance of 485 nm.
Colloidal carbohydrates were extracted from the sediment using saline water
(Underwood et al. 1995) and include low molecular weight (LMW) carbohydrates and
extracellular polymeric substances (EPS). The EPS is extracted from the colloidal
carbohydrate solution using ethanol (70% final concentration) followed by
centrifuging (Underwood et al. 1995). EPS are long-chain molecules produced and
excreted by microbial metabolism. The absorbance of a standard series of anhydrous
34
glucose solutions was used to calibrate the analysis. Carbohydrate content (m/m)
was expressed in micrograms of glucose equivalents per gram of freeze-dried
sediment (µg g
-1
).
Overlying water nutrients and salinity
Salinity was recorded at each site using a YSI 30 CTD with a 10 ft cable and probe.
Four replicate water samples for nutrient analysis were collected from just below the
water surface near each intertidal site, filtered through Whatman (GF/C) filter paper
and preserved using HgCl
2
(20 µl of 5% solution added to 30 ml sample) and frozen
at –20
o
C. Total oxidised nitrogen (TOxN) was determined using the reduced copper-
cadmium method described by Bate and Heelas (1975). Dissolved inorganic
phosphorus (DIP) otherwise referred to as soluble reactive phosphorus (SRP),
ammonium and dissolved silicate (DSi) were determined manually using standard
colorimetric methods (Parsons et al. 1984).
Sediment particle size
A scrape of sediment to a depth of approximately 1 cm was collected from directly
below the area where each chl a replicate was collected on all occasions (excl.
Swartkops Estuary, October 2001). In the laboratory each sediment sample was
dried at 105 ºC to constant mass. Samples were then disaggregated using a mortar
and pestle before being shaken through a series of steel mesh sieves (mesh
apertures of 500, 250, 125, and 63 µm). The dry sieve method could underestimate
the proportion of cohesive sediments (< 63 µm) but it was decided that the error
would be small because South African estuary sediment is generally sandier than the
U.K. and European counterparts (personal observation). Each fractionated sediment
size-class was weighed and the dry mass of each fraction was expressed as a
percentage of the total. Sediments were allocated to five different classes; sand (mud
content < 10%), muddy sand (mud content 10-25%), sandy mud (mud content 25-
50%), mud (mud content 50-85%), and clay-mud (mud content > 85%) based on the
sediment’s mud content (proportion of < 125 µm sediment), according to Figge et al.
(1980) cited in Riethmüller et al. (2000).
35
Ash-free dry weight (AFDW) and moisture content
A subsample of sediment collected for sediment particle size analysis was placed
into predried and weighed crucibles. Once the wet mass was recorded, the samples
were dried in an oven at 105 ºC for 24 hours, reweighed (dry mass) and then placed
into an ashing-oven at 550 ºC for 1 hr. The cool ashed sediments were weighed and
the data analysed.
Data analysis
Averages of the four replicate water sample nutrient concentrations and five benthos-
related variables were used when analyzing the data. Before testing for significant
differences, the Kolmogorov-Smirnov Test was used to check if the data were
parametric. If the data were parametric, then a Students t-test was used to compare
two sets of data for significant differences and the Tukey’s test was used if there
were more than two sets. However, if the data were non-parametric, then a Mann-
Whitney Rank Sum test was used to compare two sets of data and the Kruskal-Wallis
Anova on Ranks test was used if more than two sets were compared. Pearsons
Product Moment Correlation was used to test the strength of association between
variables. All tests were performed using the statistical software package Statistica
(Version 7). Principal Component Analysis (PCA) using Minitab (Version 13) was
used to examine the entire physical and chemistry data sets for variability.
Environmental variables and microalgal biomass (chl a) are presented as box-
whisker plots using Grapher 6.1.21 (Golden Software, Inc.) and sediment
composition is presented using Excel 2000 (Microsoft
®
). Mean values are expressed
as mean ± standard error of the mean.
Results
Overlying water and sediment characteristics
A high variance of salinity and nutrient concentrations was measured in the estuarine
water during the study (Fig. 2.2). Low salinity (< 5 ppt) was recorded in the upper
reaches of the Keurbooms, Gamtoos, Mngazana and Mngazi estuaries, and high
salinity (> 30 ppt) in all estuaries other than the Keurbooms and Gamtoos. Nutrient
concentrations measured in the overlying water of the Swartkops Estuary were
36
relatively high compared to the other five estuaries (TOxN
max
= 32.0 µM; NH
4+max
=
13.4 µM; DIP
max
= 53.9 µM; DSi
max
= 128.3 µM). By contrast, nutrients in the
overlying water, excluding silicate, were low in the Mngazana and Mngazi estuaries
(TOxN
max
= 2.2 µM; NH
4+max
= 4.2 µM; DIP
max
= 0.3 µM). There was a general
increase in average DSi from westerly estuaries to those in easterly areas. Intertidal
sediments in the Mngazi and Mngazana estuaries had the highest organic contents;
up to 9.9% and 7.5% respectively, and the highest moisture contents (> 35%).
Salinity in the Keurbooms Estuary in August 2002 was relatively low throughout,
ranging from less than 1 ppt to 19.5 ppt. A relatively high concentration of TOxN was
measured at the head of the estuary (23.2 µM). Ammonium was below detectable
limits throughout the estuary.
An agricultural drainage pipe emptying into the upper reaches of the Gamtoos
Estuary (Fig. 2.1) was an important source of nutrients. TOxN concentration
measured in water flowing out of the pipe was 489 µM in February 2003 and 267 µM
in August 2002. In February 2003, TOxN in the estuary decreased from 21.5 µM at
the head of the estuary to below detectable limits just 5 km downstream (site 3; Fig.
2.1). The TOxN concentration in the overlying water was significantly higher in
August 2002 compared to February 2003 (n = 20; t = 14.5; P < 0.001). The higher
average TOxN concentration could have been the result of increased river flow,
which was evident from the decrease in salinity 8.1 km from the mouth from 18 ppt in
February to 11 ppt in August. Sediment was coarsest at site 3 where the proportion
of > 125 µm sediment was 93% and was finest at site 2 where < 125 µm sediment
was > 52% (Fig. 2.3). The organic content of the sediment in the upper estuary was
generally low (< 1%) and only exceeded 1% at sites 1 and 2.
37
Figure 2.2. Box-whisker plots of water-column nutrients, ash-free dry weight
(AFDW), sediment moisture content and benthic chl a in the Keurbooms (Kb =
25/08/2002), Gamtoos (Gt1 = 21/02/2003; Gt2 = 07/08/2002), Swartkops (Sk1 =
30/10/2001; Sk2 = 12/02/2002; Sk3 = 05/12/2002; Sk4 = 15/08/2003), Sundays (Sd1
= 25/07/2002; Sd2 = 19/02/2003), Mngazana (Ma1 = 25/01/2003; Ma2 = 22/06/2003)
and Mngazi (Mi = 26/01/2003) estuaries in the study. The line within the box, the
boundaries of the box and the whiskers represent the median, 25
th
and 75
th
percentiles and the 10
th
and 90
th
percentiles respectively.
38
The Swartkops Estuary had two distinct sources of nutrients, the Swartkops
River, which flows through industrialised areas, and a stormwater canal draining the
large informal Motherwell Township. Swartkops riverwater was consistently high in
SRP and the Motherwell canal was high in DIN and DSi (maximum concentrations of
239 µM and 1107 µM respectively). However, in August 2003, the DIN concentration
was higher (31.4 µM) at the site nearest to the head of the estuary than water near
the Motherwell canal. As a result, DIP was generally highest at the head of the
estuary and TOxN highest in the lower reaches. This was not the case in August
2002 when both DIN and DIP, 39.4 µM and 30.4 µM respectively, were highest in the
upper reaches of the estuary. In October 2001, DIN followed a similar pattern to that
of February 2002, decreasing upstream and being below detectable limits in the
upper reaches.
The seepage of fertilizers from citrus agriculture in the catchment of the
Sundays River is a major source of nutrients into the Sundays Estuary. In July 2002
the DIN: DIP ratio was > 44, which indicates possible P-limited microalgal growth
(assuming a threshold ratio of 16:1). The sediment at site 3, 5.1 km from the mouth
was sandy mud, being a mix of sediment dominated by 125-250 µm sediment
particles and derived from fringing coastal dunes and fluvial sediment. The intertidal
zone at this site was sloped gently in a broad stretch of the estuary, favouring the
settling of suspended sediments and organic matter. Highest water-column nutrient
concentrations and the finest sediment were located nearest to the head of the
estuary but the intertidal zone in this region was steep, the sediment compacted and
the estuary channel narrow. This favours the resuspension of settled particles during
pulses in flow, which does not favour the accumulation of microalgal biomass.
The organic contents in the Mngazana and Mngazi estuaries had maxima of
9.9% (site 4; 3.4 km from the mouth) and 7.3% (site 3; 1.7 km from the mouth)
respectively. The sediment at site 4 in the Mngazana Estuary was mud and at site 3
in Mngazi was clay-mud, having much higher proportions of fine sediment particles
than other sites (Fig. 2.3). The nutrient concentration in the overlying water was low
by comparison with the other estuaries in the study, with TOxN generally being < 2.0
µM in January and June, and DIP generally being < 0.2 µM in January. Ammonium
concentration was elevated throughout both estuaries, ranging from 1.1 to 4.2 µM.
In January 2003, TOxN and DIP were low in the overlying water of the
Mngazana Estuary (< 2.2 µM and 0.3 µM respectively). Ammonium was 4.2 µM at
39
the head of the estuary where surface salinity was 5.7 ppt and herds of cattle
frequent. A second maximum of 1.6 µM was measured at site 4 where the benthic
sediment was finest and organic content highest. Site 3 in the Mngazi Estuary
followed a similar pattern to that in the Mngazana. The highest contents of organic
matter, fine sediment (< 125 µm) and moisture occurred at this site in the middle
reaches of the estuary (Fig. 2.3). TOxN and DIP were less than 2 µM and 0.2 µM
respectively throughout the estuary and ammonium was highest in the middle and
lower reaches of the estuary, ranging from 1.5 µM to 2.7 µM.
Figure 2.3. Stacked columns comparing the relative contributions of sediment
particle size (> 125 µm and < 125 µm grain sizes) and organic content in the intertidal
sediments of six open South African estuaries.
40
Variability of environmental variables
In the PCA, 45.7% of the total variability that was in the physical and chemistry data
was described by the first two axes, PC’s 1 and 2 (Fig. 2.4), and a further 26.5% was
described by PC’s 3 and 4. The PC loadings of the first axis showed that the
sampling stations in the Gamtoos, Swartkops and Keurbooms estuaries were largely
separated by the type of sediment and nutrient concentrations in the overlying water.
The sediment in the Keurbooms Estuary was predominantly medium grained sand
(250-500 µm) and the water was low in nutrients. In contrast, nutrient concentrations
were much higher in the Gamtoos and Swartkops estuaries and there was a full
gradient in sediment particle sizes; ranging from sites high in fines (< 125 µm) to sites
with a high coarse sand content (> 125 µm).
Sampling sites along PC2 were separated by gradients in the organic matter,
moisture and fine particulate sediment contents. This was most evident in the
Mngazana, Mngazi and Sundays estuaries. Sites located near to the mouth of the
Sundays Estuary or near to the head of the Mngazana and Mngazi estuaries were
dominated by sand and gravel, whereas the sediment in the middle reaches of the
estuaries had a high content of organic matter and fines.
Correlations between benthic microalgal biomass and environmental variables
The PCA highlighted high variation in the physical variables and chemistry of the six
estuaries. This variation was used to determine the strength of association, through
the use of correlation analyses, between benthic microalgal biomass and any single
or group, using axes scores, of environmental variables (Table 2.2).
41
Figure 2.4. Scatter plot of average sample scores (n = 5) on principal components 1
and 2, with variability scores, from a PCA of the physical and chemistry variables of
the sediment and overlying water. Estuaries are colour coded and environmental
vectors included as arrows.
42
Benthic chl a was positively correlated to TOxN, AFDW, moisture content and to
sediment particle sizes of less than 125 µm. The strongest of these correlations were
with AFDW and moisture contents. DSi was the only variable that chl a was
negatively correlated to.
It is unlikely that any single environmental variable was responsible for the
distribution of microalgal biomass in the estuaries and this was confirmed by the
significant correlations between the PC axis scores and chl a (Table 2.2). Sites with
high PC1 scores; sites at the head of the Mngazi and Mngazana estuaries and site 2
in the Keurbooms Estuary, had low concentrations of nutrients in the overlying water
and the sediment had a high content of coarse particles (> 250 µm). In contrast, sites
with low PC1 scores; sites 3 in the Swartkops and Sundays estuaries, had relatively
high concentrations of DIN (> 9 µM), DIP (> 1 µM) and DSi (> 20 µm) and had a high
proportion of fine sediment (< 250 µm). PC1 was significantly correlated to all
environmental variables excluding AFDW.
PC2 was significantly correlated to all environmental variables except for water-
column salinity and TOxN. There was a slightly stronger correlation between the PC2
scores and chl a (r = 0.379), the result of strong associations with AFDW (r = 0.534)
and the moisture content of the sediment (r = 0.622). The highest chl a
concentrations of the study (38.9 to 104.8 µg g
-1
) were measured in the middle
reaches of the Mngazana Estuary (site 4) where the intertidal sediment had a high
proportion of fine sediment (> 30%) (Fig. 2.3), AFDW (approximately 8%) and
moisture content (approximately 40%). In contrast, sites closest to the mouths of the
Mngazi (appendix figure A.2), Mngazana (appendix figure A.3) and Sundays
estuaries were dominated by coarse marine sediment that contained very little
organic matter and was easily drained.
43
Table 2.2 (continued to following page)
Pearson’s correlation coefficients relating sediment associated variables and water chemistry to benthic chl a and carbohydrates in
six permanently open estuaries (N = 290). Coefficients that are significantly correlated (p < 0.01) are in bold.
Water chemistry variables Sediment associated variables
Salinity TOxN NH4 SRP DSi AFDW Moisture > 500 250-500 125-250 63-125 < 63
TOxN -0.37
NH4 0.35 0.20
SRP 0.27 -0.03 0.26
DSi -0.48 -0.13 -0.12 -0.20
AFDW 0.32 -0.29 0.01 -0.23 0.14
Moisture 0.16 -0.12 -0.12 -0.11 -0.16 0.73
> 500 -0.23 -0.28 -0.12 -0.25 0.60 0.37 -0.04
250-500 -0.12 -0.11 -0.08 -0.10 -0.10 -0.32 -0.23 -0.06
125-250 0.29 0.05 0.18 0.16 -0.46 -0.25 -0.03 -0.63 -0.38
63-125 0.08 0.32 0.02 0.17 -0.06 0.25 0.36 -0.30 -0.66 0.08
< 63 0.03 0.39 0.03 0.21 0.05 0.15 0.21 -0.27 -0.46 -0.08 0.70
PC1 -0.46 -0.32 -0.31 -0.41 0.53 -0.09 -0.32 0.67 0.58 -0.57 -0.73 -0.61
PC2 0.18 0.13 0.22 0.27 -0.47 -0.81 -0.62 -0.53 0.48 0.47 -0.47 -0.39
PC3 0.00 -0.72 0.12 0.03 -0.36 0.48 0.37 0.05 0.13 0.19 -0.31 -0.46
PC4 0.00 -0.11 0.63 0.53 0.36 -0.01 -0.40 0.36 -0.20 -0.16 -0.02 0.09
Chl a 0.16 0.15 -0.03 0.00 -0.27 0.53 0.62 -0.09 -0.14 -0.05 0.31 0.23
Total 0.00 -0.18 -0.06 -0.11 -0.08 0.86 0.81 0.13 -0.25 -0.14 0.31 0.19
Colloidal 0.00 -0.02 -0.01 0.01 -0.25 0.64 0.67 0.02 -0.20 -0.06 0.28 0.17
EPS 0.16 0.08 0.06 0.24 -0.30 0.40 0.51 -0.13 -0.27 0.08 0.35 0.28
44
Table 2.2 continued
Physico-chemical PCA scores Biological variables
PC1 PC2 PC3 PC4 Chl a Total Colloidal
TOxN
NH4
SRP
DSi
AFDW
Moisture
> 500
250-500
125-250
63-125
< 63
PC1
PC2 0.00
PC3 0.00 0.00
PC4 0.00 0.00 0.00
Chl a -0.32 -0.38 0.15 -0.28
Total -0.24 -0.67 0.43 -0.20 0.77
Colloidal -0.29 -0.45 0.32 -0.19 0.89 0.84
EPS -0.43 -0.28 0.15 -0.09 0.81 0.65 0.85
45
Comparison of chl a content and concentration
A total of 130 chl a samples, measured on a m/m (content) or m/a (concentration)
basis were collected from the Gamtoos (08/2002), Sundays (06/2002 and 02/2003),
Swartkops (12/2002) and Keurbooms (08/2002) estuaries. The linear regression of
chl a content (x) compared to concentration (y) had a good fit (y = 1.67x; R
2
= 0.69),
and the two variables were significantly correlated (r = 0.87; P < 0.001) (Fig. 2.5).
Samples with chl a contents or concentrations that were off the regression line
included site 3 in the Sundays Estuary (July ’02) and sites 1 and 2 in the Gamtoos
Estuary (August ’02). Chl a content at sites 1 and 2 in the Gamtoos Estuary averaged
40.3 ± 5.3 µg g
-1
and the sediment was muddy (proportion of sediment <125 µm;
55.5%). In the Sundays Estuary, chl a was 20.6 ± 0.9 µg g
-1
at site 3, and the
sediment had a high sand content (proportion of sediment <125 µm; 33.6%). The
larger amount of pore space in the Gamtoos mud, which has a lower bulk density,
probably contained a greater number of microalgal cells than the same mass of
Sundays muddy sand resulting in the uncoupling of these sites from the regression
line. It is likely that other muddy sites, with a high chl a content, would have a lower
chl a m/a than the average.
Chl a content in organic matter
Organic matter can include living (e.g. microalgal cells) or non-living material (e.g.
detritus) whereas mineral particles only consist of inorganic material. By reporting chl
a as a proportion of organic matter (µg Chl a g
-1
AFDW), referred to here as OM chl
a, the effect of sediment particle size was removed. The ratio, chl a: AFDW, is the
inverse of the Autotrophic Index (Collins & Weber 1978), which was developed to
distinguish the effects of inorganic nutrients and organic enrichment. The OM chl a
followed similar trends to chl a content in freeze-dried sediment (µg Chl a g
-1
sediment) in the Sundays, Mngazana and Mngazi estuaries but there were distinct
differences in the other three estuaries (Fig. 2.6). This indicates areas of an estuary
that have been recently recolonised (pioneer stage) or areas that have reached a
more mature stage, i.e. if sediment chl a was high at a site with high organic content,
then it is likely that the area has not been eroded recently and there has been a build
up of detritus with time resulting in a low OM chl a (mature stage).
46
Figure 2.5. Benthic chl a content (µg g
-1
) in relation to chl a concentration (mg m
-2
) in
the Gamtoos, Sundays, Swartkops and Keurbooms estuaries (N = 130). A linear
regression with associated goodness of fit (R
2
) included.
OM chl a was low in relation to sediment chl a in the Bitou tributary of the
Keurbooms, middle reaches of the Gamtoos and mid-lower reaches of the Swartkops
estuaries suggesting that these habitats were stable and there had been an
accumulation of organic matter with elevated microalgal biomass at these sites. By
contrast, sites near to the head of the Keurbooms and Swartkops, as well as in the
mouth region of the Mngazana estuaries had high OM chl a relative to sediment chl
a. This could indicate areas of estuaries that have been recently eroded, or the
accumulation of organic matter was very slow, and where the microalgae have
successfully recolonised. It is likely that areas with an accumulation of detritus, which
have low OM chl a’s, are areas with a high biomass of heterotrophs and an elevated
biological oxygen demand.
47
Figure 2.6. Benthic chl a content, expressed as µg Chl a g
-1
freeze-dried sediment (sediment + organic content) and µg Chl a g
-1
organic
matter content (organic content only), in relation to distance from the estuary mouths.
48
Spatial patterns of carbohydrate
There were a number of weak, yet significant (P < 0.01), correlations between the
carbohydrate fractions and water chemistry. Total carbohydrate was negatively
correlated with TOxN as a result of the high carbohydrate content, more than 10000
µg g
-1
, in the middle reaches of the Mngazana and Mngazi estuaries where water-
column TOxN was low (< 1.8 µM). Total carbohydrate at all other sites, including
sites with high TOxN concentrations, was much lower (< 5000 µg g
-1
). In addition,
EPS was positively correlated with salinity and negatively correlated with DSi. The
highest EPS contents (> 400 µg g
-1
) were measured in the middle reaches of the
Mngazi, Mngazana and Swartkops estuaries at sites where water-column salinity and
DSi were greater than 22 ppt and 500 µM respectively. Site 1 in the Gamtoos Estuary
was the one exception where salinity was 11 ppt.
There was a significant correlation (r = 0.36; P < 0.01) between EPS and SRP
(Table 2.3). SRP concentrations in all estuaries, excluding the Swartkops, ranged
from 0 to 2.6 µM and was negatively correlated with EPS (r = -0.204; P = 0.004). This
association was also the result of the close association between chl a and
carbohydrate. In contrast, there was a strong correlation between EPS and SRP in
the Swartkops Estuary (r = 0.357; P < 0.001) where SRP ranged from 6.9 µM at the
mouth to 30.4 µM at the site nearest to the head of the estuary. There was a stronger
association between benthic chl a, colloidal carbohydrate and EPS in this estuary
with water chemistry than sediment associated variables, i.e. they were significantly
correlated to salinity, TOxN, SRP and DSi and there was only a weak correlation with
silt content (< 63 µm sediment).
Total carbohydrate is a measure of the organic matter present in sediment that is
digestable using the Dubois assay and hence it is not unexpected that AFDW and
total carbohydrate were so strongly correlated (r = 0.86). The water-soluble
carbohydrate fractions (colloidal and EPS) were also strongly correlated with organic
(r = 0.64 and 0.40 respectively) and moisture contents (r = 0.67 and 0.51
respectively).
49
Table 2.3
Pearson’s correlation coefficients (r) relating sediment associated variables and water chemistry to benthic chl a and carbohydrates in the
Swartkops Estuary (N = 90). Coefficients that are significantly correlated (P < 0.01) are in bold.
Water chemistry Sediment associated variables Biological variables
Salinity TOxN NH4 SRP DSi AFDW Moisture > 500 250-500 125-250 63-125 < 63 Chl a Total Colloidal
TOxN -0.56
NH4 -0.09 0.42
SRP -0.91 0.50 -0.02
DSi 0.09 -0.75 -0.65 0.13
AFDW 0.75 -0.17 0.24 -0.71 -0.24
Moisture 0.12 0.07 0.44 -0.27 -0.33 0.58
> 500 0.11 0.06 0.02 -0.08 -0.10 -0.17 -0.57
250-500 -0.24 -0.15 -0.22 0.30 0.35 -0.47 -0.40 0.31
125-250 0.38 0.01 0.15 -0.42 -0.29 0.27 0.11 -0.14 -0.77
63-125 -0.24 0.14 0.25 0.14 -0.11 0.21 0.61 -0.53 -0.57 0.03
< 63 0.08 0.18 -0.12 0.00 -0.07 0.48 0.29 -0.32 -0.34 -0.17 0.38
Chl a -0.45 0.71 0.09 0.47 -0.39 -0.16 -0.01 -0.01 -0.06 -0.04 -0.02 0.32
Total 0.01 0.27 0.00 -0.02 -0.20 0.42 0.47 -0.41 -0.36 -0.05 0.46 0.71 0.51
Colloidal -0.49 0.77 0.20 0.39 -0.59 -0.13 0.13 -0.07 -0.141 0.01 0.09 0.30 0.91 0.52
EPS -0.49 0.74 0.21 0.36 -0.57 -0.13 0.16 -0.07 -0.16 0.00 0.13 0.30 0.90 0.54 0.97
50
Underwood and Smith (1998b) model
Benthic chl a and colloidal carbohydrate were significantly correlated in this study
and the results were compared to the model published by Underwood and Smith
(1998b), which described the relationship using data from European estuaries (Fig.
2.7). The β regression parameter of chl a and colloidal carbohydrates obtained in this
study (β = 1.066) was similar to that (β = 1.108) reported for the model and more
than 95% of the samples analysed fell within the 95% prediction limits. A β
regression parameter of > 1 implies that there was a gradual gain of colloidal
carbohydrates, relative to benthic microalgal biomass, in the estuaries studied.
Figure 2.7. Sediment chl a and colloidal carbohydrate content in the Keurbooms,
Gamtoos, Swartkops, Sundays, Mngazana and Mngazi estuaries compared to the
model (regression line with 95% limits) of Underwood & Smith (1998b).
51
Discussion
Benthic microalgae are important contributors to primary production in shallow
aquatic ecosystems, particularly those with large intertidal areas. This study
described the spatial patterns of benthic microalgal biomass and associated benthic
carbohydrates, in relation to physical and chemical variables, in the intertidal areas of
six permanently open estuaries along the warm temperate and sub-tropical coast of
South Africa. Nutrient concentrations in the Keurbooms, Mngazi and Mngazana
estuaries were relatively low, with DIN and DIP concentrations less than 2.0 µM and
0.2 µM in the Mngazana and Mngazi estuaries. In the Gamtoos, Swartkops and
Sundays estuaries the concentrations of nutrients in the water-column were highest
during the winter months (July and August) when rainfall is generally highest. TOxN
concentrations at the sites nearest to the heads of the Gamtoos, Swartkops and
Sundays estuaries were 61.9, 31.4 and 78.4 µM respectively. SRP was generally
less than 2.5 µM in all of the estuaries except for Swartkops Estuary where
concentrations ranged from 17.1 to 30.4 µM in the upper reaches during the study.
The catchment areas of these estuaries are either heavily industrialised (Swartkops)
or are impacted by fertilizers in agricultural return flow (Gamtoos and Sundays
estuaries).
All of the estuaries have the three geomorphological zones that are typical of
micro-tidal, wave dominated estuaries along the southern and south-eastern coast of
South Africa (Cooper et al. 1999); sand-dominated flood-tidal deltas near to the
mouths, deeper and mud-dominated middle reaches and sand-dominated fluvial
deltas near to the heads of the estuaries. The highest proportion of fine particle
sediment (< 125 µm) in the intertidal zones occurred where salinity in the overlying
water ranged from 8.4 ppt in the Sundays to 32.2 ppt in the Mngazana estuary. Tidal
flow and the high variability of river flow frequently resuspend sediment in the lower
and upper reaches of estuaries, not favouring the accumulation of fine particulate
matter or the colonisation of microalgae (Adams et al. 1999).
Fine particle sediment (< 125 µm), organic matter and sediment water content
were closely associated in the estuaries studied. This was especially noticeable in
the Mngazana and Mngazi estuaries. The high content of organic-rich muds in the
middle reaches of these two estuaries was probably the result of the geology of the
catchment areas (allochthonous supply) and an accumulation of organic matter
generated in the estuary itself (autochthonous supply). The catchment areas are
52
dominated by shale, which easily erode to form clay (Blyth & de Freitas 1984), and
the water temperature frequently exceeded 25 ºC which has been found to
accelerate the rate of flocculation and sedimentation of fine particulate materials
(Jiang et al. 2003). A strong covariance between fine particle sediments, organic
matter and the water content in intertidal estuarine sediment has been found in
previous studies (Lucas et al. 2003; Perkins et al. 2003).
Microphytobenthic biomass
The average chl a content during the study was 13.9 ± 0.9 SE µg g
-1
, ranging from
below detectable levels found in marine sand at the mouth of the Keurbooms Estuary
to 104.8 µg g
-1
in organic-rich sediment in the middle reaches of the Mngazana
Estuary. These chl a contents, measured in the surface two millimetres of sediment,
were generally lower but comparable to contents measured in English, Scottish and
Portuguese estuaries (Table 2.4).
Table 2.4.
Microphytobenthic chl a ranges from intertidal sediments in different estuaries.
Estuary Sample depth
(mm)
Chl a range
(µg g
-1
) References
Southampton Water Estuary, England < 1 21-168 Friend et al. (2003)
Eden Estuary, Scotland 2 30-238 Perkins et al. (2003)
Arlesford Creek, England 2 < 0.1-460 Perkins et al. (2003)
Blackwater Estuary, England 2 88-210 Snow (unpublished data)
Colne Estuary, England 2 6.7-272 Snow (unpublished data)
Stour Estuary, England 2 21-52 Snow (unpublished data)
Colne Estuary, England 2 22-42 Hanlon et al. (2005)
Tagus Estuary, Portugal 2 21
s
-77
m
Cartaxana et al. (2006)
Ria Formosa, Portugal 2 1-202 Friend et al. (2003)
s
= average in sand and
m
= average in mud
53
Chl a measured in this study was significantly correlated to TOxN in the water-
column and moisture, organic matter and the fine particled sediment contents in the
intertidal sediments. This close association between chl a content and sediment
associated variables has been found in a number of other studies (Friend et al. 2003;
Cartaxana et al. 2006; Jesus et al. 2006). However, reports by Perkins et al. (2003)
and Tolhurst et al. (2005) have described the interdependence of chl a content, a
mass ratio, with four other biogeochemical parameters in sediments. As the content
of one changes, this affects at least one other parameter, e.g. an increase in fine
cohesive sediment results in an increase in pore space, moisture content, chl a
content and a reduction in total density. This study found that chl a content (µg g
-1
)
was significantly correlated to chl a concentration but variance did occur at muddy
sites with an elevated organic content. In addition, there was a significant, positive
correlation between chl a concentration and DIN in the water-column. However, the
correlation coefficients of chl a content (0.16) and concentration (0.33) were relatively
low so this association should be regarded with caution. The strength of association
between microalgal biomass and nutrients was much stronger in the Swartkops
Estuary, in relation to the other estuaries, and may be the result of the consistently
high nutrient concentrations entering the system from industry or stormwater runoff.
When the effects of sediment content were removed, by adjusting chl a content
to the organic content, then sites that had high chl a in relation to organic content
could be easily identified. This could be a useful tool to identify sites that have
recently been eroded and have been successfully recolonised by benthic microalgae.
In contrast, sites with high organic matter and chl a content were more likely to be in
a mature state. Factors that may affect the accumulation of fine sediment and
organic matter at a site include the intervals between floods, water temperatures in
excess of 25 ºC (Jiang et al. 2003), geohydrology, accelerated erosion in the
estuary’s catchment area and the growth rates of microalgae, macroalgae and
macrophytes in rivers that flow into the estuary as well as in the estuary itself.
Eutrophication can accelerate the growth rates of flora, accelerating the accumulation
of decaying plant material.
54
Carbohydrates
All carbohydrate fractions measured during this study were significantly correlated to
chl a, the proportion of fine sediment (< 125 µm) and the organic and water contents.
Similar associations have been found in a number of other estuaries (de Brouwer et
al. 2003; Friend et al. 2003; Lucas 2003) and colloidal carbohydrate content has
been related to sediment stability (Underwood 2000; Tolhurst et al. 2003). This
indicates a high carbohydrate content, and more stable sediment, in the middle
reaches of permanently open estuaries in the Eastern Cape, South Africa, based on
the tripartite geomorphology described by Cooper et al. (1999). In general, open
estuaries along the coast are micro-tidal (tidal variation is less than two metres),
wave-dominated and have three distinct geomorphological zones; a sandy barrier in
the mouth region with an associated flood-tidal delta. Sandy sediments generally
form wide intertidal flats. Further upstream of the flood-tidal delta there is a deeper
area where the deposition of fine particulate matter (clays, silt and organic matter) is
favoured, usually enhanced by flocculation. At the upstream limit, the water is
shallower and a fluvial delta dominated by coarse sediment is present. Results from
this study found the same general pattern in geomorphology and the carbohydrate
and microalgal contents were highest in the middle reaches of the estuaries. It is
expected that a reduction in river flow, either as a result of water abstraction or the
construction of impoundments in the catchment area, and the subsequent reduction
in the frequency and intensity of floods will allow the flood-tidal delta to penetrate
further upstream. This is likely to result in a larger area of the lower reaches of the
affected estuary becoming shallower, broader and the sediment less stable.
De Brouwer et al. (2003) attributed the correlation between median grain size
and carbohydrate contents (all fractions) to direct and indirect influences of sediment
grain size on carbohydrate content. A direct relationship exists between the amount
of organic matter adsorbed to sediment grains and sediment grain size. This is
because finer sediment particles have a larger surface area to volume ratio than
coarse sediment particles providing more adsorption sites. Sediment grain size has
an indirect effect on extracellular carbohydrate contents because diatoms maintain
higher growth rates on silt-dominated sediments. In turn, the particles at the sediment
surface become bound together through the secretion of extracellular
mucopolysaccharides, a process termed “biostabilisation”. Over an extended period
55
of stable conditions, the benthic environment becomes more mature, leading to
increased levels of extracellular carbohydrates.
Yallop et al. (2000) also found multicollinearity between biological and
biochemical variables in the Severn Estuary (U.K.). Once collinear variables were
eliminated, only three variables were used in a model to determine sediment stability;
water content, EPS and chl a. The model provided evidence that the relationship
between these variables could change from “pioneer” to “mature” biofilms. Benthic
chl a generally increased as the water content increased but sediment immersion or
an erosion event could reset the sediment to an early colonisation phase. An
example of an early pioneer phase during the current study would be in the sediment
in the mouth area of the Keurbooms Estuary. Tidal resuspension keeps this sediment
in a continual pioneer state resulting in low chl a, colloidal carbohydrate and organic
contents. The finer sediments in the middle reaches of the Mngazi, Mngazana and
Gamtoos estuaries occur in more stable environments allowing microalgal biofilms to
reach maturity and an accumulation of organic matter in the cohesive sediment to
occur. Yallop et al. (2000) described a high bacterial cell density during this phase,
which is also capable of contributing to the colloidal carbohydrate content in the
sediment.
The foregoing suggests that biogeochemical parameters play an important role in
determining the carbohydrate content in intertidal mudflats, supporting a simple
model developed by Underwood & Smith (1998b). This model describes a strong
relationship between colloidal carbohydrate and chl a and is valid for intertidal
mudflats where epipelic diatoms constitute > 50% of the microphytobenthic
assemblage. Microalgal biomass and colloidal carbohydrate were strongly correlated
in the study reported here and almost all samples fell within the 95% limits of the
Underwood and Smith (1998b) model. Other studies of mudflats in Western Europe
(de Brouwer et al. 2003; Lucas et al. 2003) and England (Bellinger et al. 2005) have
shown similar relationships.
Conclusions
Conditions that favour the accumulation of fine sediment and organic material in
estuarine sediments are likely to support a high biomass of benthic microalgae,
which will increase the stability of the sediment through the production of colloidal
carbohydrates. These conditions include reduced river flow through increased
56
abstraction and the impoundment of rivers, which decrease the frequency and
intensity of floods. Loads of suspended particulate material such as organic matter,
fine sediment particles, microalgal cells and absorbed nutrients (DIP in particular) are
then more likely to flocculate and settle out of the water-column as a result of these
more stable conditions. The process of eutrophication, in which there is a gradual
accumulation of organic matter as a result of elevated nutrient concentrations, can
also provide a suitable environment for high benthic microalgal biomass. Extremely
high benthic chl a concentrations (in excess of 300 mg m
-2
) were measured in the
eutrophic Mdloti and Mhlanga estuaries on the KwaZulu-Natal coast (Perissinotto et
al. 2004). These estuaries are temporarily open-closed estuaries (TOCE), which
makes them more susceptible to the effects of elevated nutrients.
Benthic microalgal biomass in the intertidal sediment of permanently open
estuaries was associated with TOxN in the water-column but the association was
much stronger with biogeochemical parameters in the sediment (organic and
moisture contents and the proportion of mud). This could be the result of, in part at
least, measuring chl a as a content and not as a concentration. However, chl a
content was significantly correlated to the concentration, and a similar association
between microalgal biomass and sediment type was found in the Swartkops Estuary
in a previous study using chl a concentration (Rodriguez 1993). The highest chl a
contents measured during this study were generally in the intertidal sediments
bordering the deeper middle reaches of the estuaries, where deposition typically
exceeds the resuspension of sediment. Sites in these reaches of the estuaries had
the highest proportion of fine sediment and organic matter, and it is likely that the
remineralisation of nutrients from the sediment, NH
4+
in particular, was an important
source of nutrients during the warmer, drier summer months when nutrient
concentration was lowest. However, there was a high level of variation in the
microalgal biomass and nutrient concentration in the water-column between sampling
sites, estuaries and sampling dates. From this it does not seem possible to generate
a generic model of factors that determine microalgal biomass in all six estuaries.
Instead, this study emphasises the need to conduct more extensive investigations of
each estuary to get a better understanding of processes affecting the distribution of
the microphytobenthos.
57
CHAPTER 3
Response of microalgae in the Kromme Estuary to managed freshwater inputs
Snow, G.C. & Adams, J.B. 2006. Response of micro-algae in the Kromme Estuary to
managed freshwater inputs. Water SA. 32 (1), 71-80.
58
Abstract
The Kromme is a permanently open estuary that receives little freshwater input
because the capacity of the dams is equivalent to the mean annual run-off of the
catchment. The estuary is marine dominated and phytoplankton chlorophyll a (chl a)
is low because of reduced freshwater pulses that introduce nutrient-rich freshwater.
Water released (2 x 10
6
m
3
) from the Mpofu Dam in 1998 showed little biological
effect on the estuary. This study addresses other run-off scenarios to determine
which would be beneficial in stimulating microalgal production. Recent surveys
together with past research were used to describe the present state and reference
condition of the estuary. Average intertidal chl a was 12.9 ± 2.5 µg g
-1
and 4.9 ± 0.4
µg g
-1
during November 2003 and July 2004. These concentrations are low but
comparable to those found in intertidal sediments in other South African estuaries
which could indicate that intertidal microalgal biomass is not severely limited by low
freshwater inputs. Average water-column chl a concentrations have ranged from 0.6
± 0.1 µg l
-1
to 5.6 ± 0.3 µg l
-1
.
Present state conditions can thus be described as that where water-column chl
a seldom exceeds 5 µg l
-1
and small flagellates dominate the phytoplankton. The
diatoms introduced via freshwater have been lost. Under reference conditions
baseflow would have been greater than 1 m
3
s
-1
for approximately 8 months of the
year. The flocculation of fine particles associated with the mixing of fresh and saline
water would have resulted in phytoplankton peaks (chl a >10 µg l
-1
) in the middle
reaches of the estuary. A more suitable habitat would also have been present for the
epipelic (motile microalgae that inhabit muddy sediments) microalgae. An
assessment of the possible future run-off scenarios indicated that the most beneficial
for the microalgae would be a flow release from the Mpofu Dam of 5 x 10
6
m
3
in
October and January. This would lead to a 25-33% increase in phytoplankton chl a
and a 50% increase in intertidal benthic chl a for approximately 2 months after the
releases.
59
Introduction
The Kromme Estuary is located in the Eastern Cape Province, 80 km west of Port
Elizabeth. The estuary is relatively narrow with a mean width of approximately 80 m,
and extends for 14 km from a permanently open mouth to a rocky sill that forms the
tidal head of the estuary. Its major tributary is the Geelhoutboom River which enters
8 km from the mouth. The catchment size and length of the Kromme River are
approximately 1000 km
2
and 100 km respectively. Rainfall shows a bimodal pattern
with maxima in autumn and spring. January and February are the driest seasons
(Bickerton & Pierce 1988).
There are two dams upstream of the estuary and their combined holding
capacity (Mpofu Dam 107 x 10
6
m
3
, appendix figure A.1, and the Churchill Dam 33.3
x 10
6
m
3
) exceeds the mean annual runoff (MAR) of the Kromme River (estimated to
be in excess of 105 x 10
6
m
3
) (Reddering 1988). This has resulted in increased
salinity towards the head of the estuary. In addition to reduced river flow, the dams
have also affected the frequency and magnitude of flood events, and the availability
of riverine material replenishing the estuarine nutrient pool (Scharler et al. 1997). A
release policy, which provides 2 x 10
6
m
3
per annum was proposed to account for the
evaporative loss of the estuary (Jezewski & Roberts 1986). However, flow records
and personal communication with the Mpofu Dam managers indicate that very few or
no releases have been executed for 2002-2005. The construction of farm dams and
a weir, together with abstraction of water for agriculture, have significantly reduced
the flow of water input from the Geelhoutboom River.
A freshwater release study in 1998 (Bate & Adams 2000; Snow et al. 2000a)
indicated that a release of 2 x 10
6
m
3
had little beneficial effect on the estuary and
that a consistent baseflow was probably necessary to maintain water-column
production. The results reported in this study were a component of a Department of
Water Affairs and Forestry comprehensive ecological reserve (freshwater
requirement) study on the Kromme Estuary. The response of the microalgae to
different run-off scenarios was investigated to determine optimal conditions for
microalgal production.
The present state (distribution and biomass) of the microalgae was documented
based on two recent surveys and past research results (Scharler et al. 1997; Bate &
Adams 2000; Snow et al. 2000a). The reference condition (before anthropogenic
60
influences) was predicted as well as the response of the microalgae to different
freshwater inflow scenarios. These scenarios were:
i) 5 x 10
6
m
3
release from the Mpofu Dam during November,
ii) 5 x 10
6
m
3
release during November and another in January,
iii) Maintain present flow in the Kromme River but increase flow in the
Geelhoutboom tributary
iv) 5 x 10
6
m
3
release during November and increased flow from the
Geelhoutboom tributary
v) 7.5 x 10
6
m
3
release over a two-month period (October-November).
Materials & methods
Study site
The Kromme Estuary lies in a relatively undisturbed and pristine area. It meanders
through rural land and there is little urban, industrial and agricultural development
(Baird & Heymans 1996). Recently, there has been a lot of clearing along the banks
of the estuary associated with recreational / residential developments. A road bridge
has been constructed across the estuary, approximately 3 km from the mouth, and a
marina development consisting of a number of canals with waterfront housing and
mooring facilities, is located on the west bank near to the mouth (Emmerson et al.
1982) (Fig. 3.1). The substrate at the mouth area varies between medium and
coarse sand with some silt. In the lower and middle reaches, mud and silt increase
whereas the upper reaches consist of coarse silty sand, stones and rock.
The estuary was sampled for microalgae in November 2003 and in July 2004.
These surveys focussed on the Geelhoutboom tributary where few data are
available. During November 2003, two sampling sites were located below the
confluence of the Kromme Estuary and the Geelhoutboom tributary, 1.7 and 8.1 km
from the mouth, and two sites were located within the Geelhoutboom, 9.9 and 11.1
km from the mouth. An additional site was located near to the head of the estuary but
only physico-chemical measurements were taken there. On 30 July 2004 there were
10 sites in total, all matching the locations of Scharler et al. (1997). Sites at 1.7, 5.1,
6.6, 8.1, 10.2, 11.8 and 13.2 km from the mouth closely match the sites sampled
during the 1998 freshwater release study (Snow et al. 2000a).
61
Figure 3.1. Map of the Kromme Estuary and Geelhoutboom Tributary, showing
distances of the sampling stations (km from the mouth) for the 2004 surveys
(modified from Bickerton & Pierce 1988).
Physico-chemical factors
Water-column conductivity, salinity, dissolved oxygen (DO) content, total dissolved
solids and pH were measured using a YSI multiprobe. Measurements were taken
every 50 cm from the surface at each site. Water transparency was measured with a
Secchi disc.
Phytoplankton biomass
Water-column samples were collected from the surface and bottom and then
gravity filtered through plastic Millipore towers using Whatman (GF/C) glass fibre
filters. The samples were collected using a 500 ml weighted pop-bottle. Chl a was
extracted by placing the filters into glass vials containing 10 ml of 95% ethanol
(Merck 4111). The samples were extracted overnight at 1-2 ºC and filtered.
Absorbances were read at 665 nm, before and after acidification, using a UV/Vis
spectrophotometer. Chl a concentration was calculated according to Hilmer (1990).
62
Microphytobenthos biomass
Microphytobenthos biomass was estimated using benthic chl a (µg Chl a g
-1
freeze-
dried sediment; expressed as µg g
-1
). Four 1 cm deep intertidal and subtidal
sediment cores were collected from each site, frozen and kept in the dark before
being freeze-dried in the Secfroid Lausanne Suisse freeze-drier. The process of
freeze-drying removes interstitial water that improves chlorophyll extraction. Once
freeze-dried, 4 ml of 95% ethanol were added to approximately 100 mg of sample
and pigments were extracted in a fridge for 24 hours.
After extraction the samples were well mixed using a whirlmixer (WM/250/SC/P)
and the extract injected into a high performance liquid chromatograph (HPLC)
attached to Waters-Lambda-Max 481 LC spectrophotometer and Waters LM-45
solvent delivery system for chl a analysis. A 30% methanol and 70% acetone mixture
was used as a carrier. The system was calibrated using the chl a of red seaweed
(Plocamium collorhiza) because it contains no chlorophyll b that might interfere with
the chl a reading. Chl a absorbance was measured at 665 nm and the concentration
determined using the modified equation of Nusch (1980) (Snow et al. 2000a).
Phytoplankton identification
Surface and bottom water (500 ml) were collected from each site and preserved with
1 ml of 25% Glutaraldehyde solution for phytoplankton identification. The filtrate from
phytoplankton chl a was preserved using mercuric chloride and frozen for nutrient
analysis.
Two drops of Rose Bengal were added to 60 ml of the preserved water samples
and poured into a 26.5 mm internal diameter settling chamber and allowed to stand
for 24 hours before identification with a Zeiss IM 35 inverted microscope at 630X
magnification. A minimum of 200 cells was counted in each sample and the cells
were classified as small flagellate