Proefschrift Immers
Click here to load reader
-
Upload
nicole-nijhuis -
Category
Documents
-
view
268 -
download
6
description
Transcript of Proefschrift Immers
N E T H E R L A N D S I N S T I T U T E O F E C O L O G Y
Preven
ting or p
redictin
g cyanob
acterial bloom
sN
IOO
Th
esis 115A
nn
e K. Im
mers
Invitation to attend the public defence of my thesis:
Preventing or predicting cyanobacterial blooms
Iron addition as a whole lake restoration tool
Monday December 22nd
at 14.30
SenaatskamerUtrecht University
Domplein 29Utrecht
Anne K. [email protected]
Paranymphs:
Tânia Vasconcelos [email protected]
Dennis [email protected]
Reception to follow
Preventing or predicting cyanobacterial blooms
Iron addition as a whole lake restoration tool
Anne K. Immers
Preventing or predicting cyanobacterial blooms
Iron addition as a whole lake restoration tool
Thesis committee: Prof.dr. W. Admiraal
Dr. L.M. Dionisio Pires
Prof.dr. L.M.P. Lamers
Prof.dr. J.C.M. Smeekens
Prof.dr. J.T.A. Verhoeven
Layout and printed by: Gildeprint
Cover artwork: Niek & Rosa Immers
This thesis should be cited as: Immers AK (2014) Preventing or predicting cyanobacterial
blooms - Iron addition as a whole lake restoration tool. PhD thesis. Utrecht University,
Utrecht, The Netherlands.
ISBN 978-94-610-87959
Preventing or predicting cyanobacterial blooms
Iron addition as a whole lake restoration tool
Voorspellen of voorkomen van blauwalg-drijflagen
IJzeradditie als maatregel om een meer te herstellen
(met een samenvatting in het Nederlands)
Proefschrift
ter verkrijging van de graad van doctor aan de Universiteit Utrecht
op gezag van de rector magnificus, prof.dr. G.J. van der Zwaan,
ingevolge het besluit van het college voor promoties
in het openbaar te verdedigen
op maandag 22 december 2014 des middags te 2.30 uur
door
Anne Katherine Immers
geboren op 24 september 1985 te Delft
Promotoren: Prof.dr. E. van Donk
Prof.dr. B.W. Ibelings
Copromotor: Dr. E.S. Bakker
This thesis was accomplished with financial support from the Water Framework Directive
Innovation Fund from Agentschap NL from the Dutch Ministry of Economic Affairs, Agriculture
and Innovation, STOWA and the Netherlands Institute of Ecology (NIOO-KNAW).
Voor mijn ouders, Ben & Tiny
CONTENTS
Chapter 1 General Introduction 9
Chapter 2 Lake restoration by in-lake iron addition: a review 21
of iron impact on aquatic organisms and lake ecosystems
Chapter 3 Iron addition as a measure to restore water quality: 39
implications for macrophyte growth
Chapter 4 Iron addition as a shallow lake restoration measure: 61
impacts on charophyte growth
Chapter 5 Invasive crayfish threaten the development of submerged 77
macrophytes in lake restoration
Chapter 6 Iron addition and biomanipulation as complementary 99
measures for the restoration of a shallow peaty lake
Chapter 7 Gone with the wind - Stability of cyanobacterial scums 123
under turbulent conditions
Chapter 8 Synthesis 151
Summary 161
Samenvatting 167
Dankwoord 173
References 177
Curriculum Vitae 195
List of publications 197
CHAPTER 1
General Introduction
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 1
10
EUTROPHICATION – A GLOBAL PROBLEM
Numerous freshwater lakes throughout the world have been suffering from eutrophication due to
increased nutrient loading during the second half of the 20th century. This excess input of nutrients
into the system, mainly nitrogen (N) and phosphorus (P), has in many cases led to an increase in
phytoplankton productivity, shifting the lake from a macrophyte dominated system towards an
algal dominated system (Scheffer et al., 1993; Smith and Schindler, 2009). Whereas macrophytes
play a key role in enhancing water quality by acting as a nutrient sink, preventing resuspension of
the sediment and providing a habitat for a variety of zooplankton and macrofauna (Jeppesen et al.,
1998; Van Donk and Van de Bund, 2002; Bakker et al., 2010), algal dominated systems are often
characterised by low water transparency and a decreased aquatic biodiversity (Moss, 1990; Pearl
and Huisman, 2008). Moreover, as many cyanobacteria are able to produce toxins and can form
dense blooms at the surface of lakes (scums; Figure 1.1), they pose a risk to the aquatic biota, but
also to humans who come into contact with this water via consumption or recreational activities
(Guo, 2007; Pearl and Huisman, 2008). In order to tackle this deterioration of our freshwater
systems, the European Union has set up a directive (e.g. Water Framework Directive; WFD),
which requires European lakes to meet the standards of a good ecological state by 2015 (European
Commission, 2000). One of the criteria in the WFD for a good ecological state is a decrease in
lake phosphorus concentrations, after which lakes are expected to switch back to a self-stabilising
macrophyte dominated system (Moss et al., 2003; Smith and Schindler, 2009).
LAKE RESTORATION
In the past, lakes were often fed by nutrient-rich river inlet water, and this high external loading
of P has been the main cause of the high phosphorus concentrations in lakes. Moreover, additional
nutrients frequently reached the system via run-off from nearby agricultural fields and improper
connections to sewage systems. However, due to the modernization of sewage systems and removal
of phosphorous in river inlet water, a large part of this external P loading has now been reduced
in many European and North American lakes (Klapwijk et al., 1982; Jeppesen et al., 1991; Van
Liere and Janse, 1992). Nonetheless, external nutrient loading remains a serious issue on a global
scale, especially in developing countries (Jeppesen et al., 2012).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
General Introduction
11
1
Figure 1.1 – Scum of the genus Microcystis accumulating in a harbour at Lake Westeinderplassen, The Netherlands.
Whereas in some cases the reduction of external nutrient loading resulted in an improvement
of the water quality (Marsden,1989), in many cases the recovery was delayed by biological or
chemical in-lake mechanisms, such as unbalanced foodweb interactions or high sediment P
concentrations (Hansson et al., 1998; Gulati and Van Donk, 2002; Søndergaard et al., 2007).
Foodweb interactions, such as high densities of plankti- and benthivorous fish, can reduce the
standing stock of grazing zooplankton and decrease water transparency and macrophyte vegetation
by resuspension of seston and inorganic matter from the lake sediment (Moss, 1990; Gulati
and Van Donk, 2002). Biomanipulation, e.g. the removal of benthi- and planktivorous fish or
stocking of piscivorous fish, could therefore be an effective method to stabilize these trophic
interactions and increase the density of herbivorous zooplankton (mainly Daphnia) and submerged
macrophytes. Biomanipulation has been highly successful in shifting turbid lakes to the clear
water state (Van Donk et al., 1990; Meijer et al., 1994; Søndergaard et al., 2007; Jeppesen et
al., 2012). The longevity of biomanipulation success can, however, be impeded by a multitude
of factors. In order for the biomanipulation to have long-term success, yearly biomanipulation
should continue during the following years in order to decrease the number of young-of-the-year
recruits, which are enhanced by reduced intra- and interspecific competition (Hansson et al.,
1998; Gulati et al., 2008). Moreover, macrophyte return can be impeded by a missing seedbank,
grazing by waterfowl, or unfavourable abiotic conditions, such as upwelling of the sediment by
wind induced waves (Bakker et al., 2013). But most importantly, biomanipulation in highly
eutrophied water bodies can only be effective on a longer term when phosphorus concentrations
are reduced, not only from external but also internal sources (Meijer et al., 1994; Hansson et al.,
1998; Søndergaard et al., 2007; Gulati et al., 2008).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 1
12
High build-up of excess nutrients over the years has in many cases led to high P concentrations
in the sediment, which can hamper or delay the recovery of lakes due to the slow release of P from
the sediment into the surface water, so called internal loading (Jeppesen et al., 1991; Søndergaard
et al., 2003; Smolders et al., 2006). It was calculated that this internal P loading can, depending
on loading history, persist up to 10-15 years after the reduction of external P loading of lakes
(Jeppesen et al., 2005). Therefore, in order to tackle the high phosphorus concentrations in lakes,
restoration measures are nowadays frequently focused on reducing the internal P loading from
the sediment.
RESTORATION MEASURES TO REDUCE INTERNAL P LOADING
Multiple restoration methods have been proposed to tackle internal P loading in lakes, either
through physical or chemical methods.
Physical restoration measures
Physical restoration methods include dredging, flushing, and the addition of passive capping
agents, during which the nutrient-rich top layer of the sediment is removed from the lake (Van
der Does et al., 1992), the lake is flushed with nutrient-poor water (Jagtman et al., 1992), or the
sediment is covered with a layer of sand, gravel or clay (Hickey and Gibbs, 2009; Bakker et al.,
2011), respectively. By completely removing the P rich top layer of the sediment, a new and less
reactive layer is exposed which will reduce the amount of P that is released to the overlying water
column (Cooke et al., 1993b). On numerous occasions, however, a new nutrient-rich layer was
uncovered after dredging, which not only made the measure redundant, but also deteriorated lake
nutrient conditions (Hosper, 1998; Annadotter et al., 1999). Additionally, dredging can harm
the macrofaunal community due to physical damage and burial by dredge trailings (Krueger et
al., 2007). Completely removing the top layer of the sediment is, in addition, a costly and time-
consuming process and often problems arise with finding appropriate areas to store the (polluted
or toxic) sediment (Gulati and Van Donk, 2002). Flushing a lake with nutrient-poor water can
decrease TP concentrations in a lake, but experiments by Jagtman et al. (1992) showed that
after an initial success of decreased chlorophyll and nutrient concentrations, water transparency
quickly decreased due to resuspension of both detritus and inorganic suspended matter. Capping
the sediment with sand, gravel, or clay lowers the diffusion rate of nutrients into the overlaying
water column (Bakker et al., 2011). Thick layers (> 5 cm) of fine material are most effective
(Hickey and Gibbs, 2009), but adding such a large quantity of material to the sediment limits
this method to small lakes and ponds. Therefore, lake managers usually try to find more cost-
effective restoration measures and only suggest these physical restoration techniques when other
measures have failed in reducing internal P loading.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
General Introduction
13
1Chemical restoration measures
Chemical restoration measures are based on adding chemical substances to a lake, such as
aluminium (Al), calcium (Ca), iron (Fe), or lanthanum-enriched benthonite clay (Phoslock®),
which naturally bind to P and enhance sediment P binding capacity (Cooke et al., 1993a; Burley
et al., 2001; Smolders et al., 2006; Kleeberg et al., 2013; Lürling and Van Oosterhout, 2013). If
added on a regular basis, these chemicals will not only precipitate with the available phosphate
(PO4) in the water column and sediment, but can also provide long-term control of internal
P loading from the sediment (Boers et al., 1994; Cooke et al., 1993a; Smolders et al., 2006;
Kleeberg et al., 2013). Which and how much of each capping agent to use highly depends on lake
properties, such as lake size, depth, flushing rate, wind fetch, type of sediment, and water quality
(alkalinity, pH, and organic content).
Aluminium is widely used as a flocking and P capping agent, because it has the advantage
that it forms irreversible bonds with P and it works also under anoxic conditions (Cooke et al.,
1993a; Reitzel et al., 2005; Hickey and Gibbs, 2009). Treatment of lakes with aluminium indeed
resulted in decreased TP and chlorophyll concentrations that lasted 2-20 years after application
(Welch and Cooke, 1999; Reitzel et al., 2005). The use and dose of aluminium is however
restricted by the pH and alkalinity of the water, as aluminium is toxic to fish and other organisms
when pH decreases below 6.5 (Gensemer and Playle, 1999; Hickey and Gibbs, 2009).
Calcium addition (in the form of calcite or lime) is most efficient in binding P during periods
of high photosynthetic activity, when pH values exceed 9 (Cooke et al., 1993a). When water pH
drops, for instance during respiration, calcite becomes soluble and P is released back into the
system (Andersen, 1975). Addition of calcite and lime in two hardwater lakes proved, however, to
be very effective in decreasing lake TP concentrations (Prepas et al., 2001). Whereas the positive
effects were visible for over 7 years after application and water transparency increased, macrophyte
biomass slowly declined, which was probably caused by the high pH of the lake water (Prepas et
al., 2001).
The binding capacity of iron is regulated by the redox state of the water (Lijklema, 1977;
Burley et al., 2001; Smolders et al., 2006). Under oxic conditions in the top layer of the sediment,
oxidized ferric iron can freely precipitate with P, but under anoxic conditions, reduced ferrous
iron is formed and iron loses this binding capacity and consequently P will be released (Mortimer,
1941; Lijklema, 1977; Cooke et al., 1993a; Golterman, 2001). To avoid P release from the
sediment during anoxia, the treatment has in some cases been complemented with hypolimnetic
oxygenation (Cooke et al., 1993a; Jaeger, 1994). Recent experiments by Kleeberg et al. (2013)
did, however, show that the success of iron addition is not hindered by the redox sensitivity
of iron and P can be efficiently precipitated independent of the nature of the oxygen supply.
That is, when iron is added to reach a sediment molar Fe:P ratio of 7. These conditions will
assure continuous P elimination independent of oxygen supply, as both will be released from the
sediment in a ratio close to 1 and consequently coprecipitate due to natural oxygenation processes
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 1
14
(Kleeberg et al., 2013). Indeed, iron addition without additional aeration in soft-water lakes,
reservoirs and deep dimictic lakes yielded low TP and chlorophyll concentrations for up to eight
years (Daldorph and Price, 1994; Jaeger, 1994; Kleeberg et al., 2012).
Addition of Phoslock® (developed by CSIRO Australia) is a relatively new method to combat
internal P loading, during which the lanthanum-enriched clay particles bind to soluble P in
the water column and the accumulated thin layer of clay on the sediment prevents P release
(Robb et al., 2003; Douglas et al., 2004; Lürling and Van Oosterhout, 2013). Phoslock® addition
successfully decreased TP under a range of environmental conditions (pH 5-10) and was unaffected
by anoxia (Robb et al., 2003; Meis et al., 2013). The use of lanthanum is, however, rather
expensive compared to the other chemical P binding agents (Spears et al., 2013) as lanthanum is
a rare earth metal and is used to manufacture computer hard drives, mobile phones, and electric
car batteries (Thomas et al., 2014).
Although the effects of these chemical restoration measures on lake water quality prove to
be positive, adding large quantities of chemical substances to a lake can have serious negative
consequences for the aquatic community. Various experiments have shown that the addition of
these substances or their precipitates can negatively affect littoral and benthic communities,
either directly due to toxic effects (Cooke et al., 1993a) or indirectly by affecting life history traits
(Lürling and Tolman, 2010), as a side effect of changing environmental conditions such as pH
(Prepas et al., 2001; Hickey and Gibbs, 2009), by covering organisms, food sources or habitats
(Gerhardt and Westermann, 1995; Linton et al., 2007; Hickey and Gibbs, 2009), or by changing
community composition due to differences in tolerance for these chemicals (Vuori, 1995).
ACTIVELY CONTROLLING CYANOBACTERIAL DOMINANCE
Removal or suppression of cyanobacteria
Nutrient reduction (both external and internal) is on a long term the most successful method to
shift lakes from cyanobacterial to macrophyte dominance (Smith and Schindler, 2009; Søndergaard
et al., 2013), but other restoration measures can on a shorter term induce the required changes
by either killing cyanobacteria or by suppressing cyanobacterial growth or abundance (Visser et
al., 2005). Cyanobacteria can be killed by adding algicides, such as copper sulphate, to a lake.
Not only is this method non-selective and can negatively affect the whole aquatic community,
experiments by Kenefick et al. (1993) also showed that the addition of algicides resulted in release
of cyano-toxins to the water, thereby possibly deteriorating the problem. A more selective method
is injecting hydrogen peroxide (H2O
2) in the water, which selectively kills cyanobacteria and is
relatively harmless to eukaryotic algae (Matthijs et al., 2012). Whereas during the application
negative effects on eukaryotic phytoplankton, zooplankton, and macrophytes appeared mild, the
long-term effects of H2O
2 addition on the aquatic community are still unknown (Matthijs et al.,
2012).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
General Introduction
15
1Other techniques to suppress cyanobacterial growth are designed to reduce or remove the
advantage cyanobacteria have over other phytoplankton, through their ability to regulate their
buoyancy (Reynolds et al., 1987; Ibelings et al., 1991; Brookes et al., 1999). Cyanobacteria have
the ability to track the illuminated surface mixed layer by altering the density of their cells and
can hence outcompete other phytoplankton in the competition for light (Walsby, 1992; Visser
et al., 2005; Jöhnk et al., 2008). This ability is provided by intracellular gas-vesicles, hollow
structures filled with air, which decrease the density of the cell and can make it buoyant. At
low irradiance, these gas-vesicles provide the cell with buoyancy and consequently move the
cell to the epilimnion. During the day, when increased irradiance causes the cell to build up
carbohydrates as a by-product of photosynthesis, cell density increases again and causes the cell
to sink to nutrient rich deeper waters. This diel rhythm of buoyancy controlled migration is,
however, only possible at low turbulence. At higher mixing rates cyanobacteria will simply be
entrained in the turbulent flows. At low wind speed and strong insolation, as during hot summer
days, fast floating cyanobacterial colonies are able to dis-entrain from the weakening turbulence,
which results in tracking the near surface mixed layer or in the complete absence of mixing, the
formation of surface scums (Reynolds et al., 1987; Brookes et al., 1999).
Methods to prevent cyanobacterial dominance and scum formation are therefore focussed
on either collapsing these gas vesicles or on removing the advantage of buoyancy by entraining
cyanobacterial cells to deeper waters with the help of artificial mixing systems (Walsby, 1992; Visser
et al., 2005). When gas vesicles are collapsed, the advantage of large size for big cyanobacterial
colonies becomes a disadvantage due to high sinking velocities and the colonies are quickly lost
to deeper layers in the water column. Successful use of pressure to collapse gas vesicles has been
reported using devices such as ultrasonic transducers, explosives, or by transporting colonies to
deep water via pipes (see Walsby, 1992 and Visser et al., 2005 for detailed descriptions). The
pressure that is needed to collapse sufficient gas vesicles to remove the advantage of cell buoyancy
depends on both the species and lake depth (Walsby, 1992). Species that are adapted to deep
lakes have much stronger, pressure resistant gas vesicles, whereas gas vesicles of species in shallow
lakes are more easily collapsed (Walsby, 1994). Ideally, the maximum needed pressure should be
applied to lakes in order to collapse all gas vesicles, whereas in practice a consideration should be
made between efficacy, costs, and possible drawbacks for the aquatic environment.
Entraining cyanobacterial colonies to deeper layers in the water column with the help of
artificial mixing systems decreases cyanobacterial dominance in a lake via a number of ways.
By transporting cyanobacterial colonies to deeper layers in the water column, cyanobacterial
colonies will receive insufficient irradiance for net growth (Walsby, 1992). Moreover, circulating
the planktonic community of a lake will reduce sedimentation losses for other non-buoyant
phytoplankton, such as diatoms and green algae (Visser et al., 1996). Increased abundance
of these eukaryotic algae will increase the competition for light and nutrients. The resulting
changed conditions of increased nutrient availability, decreased pH, and a dynamic light regime
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 1
16
will favour fast growing eukaryotic algae over slow growing cyanobacteria (Ibelings et al., 1994;
Visser et al., 2005). Artificial mixing, either via aeration or mechanical pumps or propellers, has
proven to be successful in shifting the phytoplankton dominance towards eukaryotic species,
but the applicability of the method is reserved to (high-nutrient) deep lakes (Visser et al., 1996,
2005; Hickey and Gibbs, 2009).
Prediction of the occurrence of cyanobacterial scums
Whereas most restoration techniques are focused on removing nutrients from the system,
increasing the biomass of phytoplankton grazers, or decreasing cyanobacterial biomass, hence
pushing the system towards a macrophyte dominated state, these measures often prove to be
expensive, time consuming, and only effective on a very long term. Moreover, climate change is
predicted to not only affect water temperature, but also water column stability, nutrient loading
(Carey et al., 2012; Elliot, 2012), and lake residence time (Visser et al., 2005), which will likely
result in increased abundance and duration of cyanobacterial blooms. Therefore prediction of
bloom and scum formation, e.g. to timely warn the public against the risks of cyanobacteria or
to inform drinking water companies, remains a necessity, if only to bridge the years before water
quality is fully restored. Scum formation is probably the most pressing risk posed by cyanobacteria,
as during scum formation biomass increases manifold over a short time interval. Since the main
cyanobacterial toxins (microcystins) occur intracellular, toxin concentrations in scums can quickly
increase to alarmingly high levels (Ibelings et al., 2012). Protocols for risk assessment and risk
management in most countries take this enhanced risk level through scum formation into account
and presence of scums typically results in the highest alert level (Ibelings et al., 2014).
Scum formation depends on the presence of a buoyant cyanobacterial population and a
stable water column (Ibelings et al., 2003). Wind induced turbulence can in turn decrease water
column stability and break up cyanobacterial scums. By defining these different parameters in
advance, detailed model predictions can be made on the time and location of scum formation and
disappearance. Model work by Ibelings et al. (2003) showed that it was possible, by combining
these parameters with meteorological forecasts, to not only correctly predict scum occurrence in
the open water of the large Lake IJssel (The Netherlands), but also predict scum formation several
days in advance. Using models to predict scums in more sheltered places, such as harbours, ponds,
or ship locks where contact with people is most extensive and scums may persist longer, still
remained problematic, however.
In 2007, an elaborate study was set up to develop a model which would predict the time and
location of scum formation and disappearance in five shallow lakes in The Netherlands (Burger
et al., 2009). Whereas in two of the five lakes 80% of the model predictions correctly matched
field observations, the predictions for the other lakes only matched 50% of the time (Burger et
al., 2009). This low number of correct predictions was mainly due to a high number of scum
predictions by the model, when no scum was observed in the field (false positives). The number of
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
General Introduction
17
1false positives could be reduced by changing the parameter settings of the model, but this resulted
in more scums being missed (false negatives). One explanation for these mismatches could be
that only one value for buoyancy was used for all cyanobacterial species, which was derived from
the genus Microcystis. Field measurements showed that Microcystis sp. did occur in the five lakes,
but that most of the time the cyanobacterial biomass was dominated by Anabaena sp. (Burger
et al., 2009). A variety of surface bloom forming cyanobacterial species can be dominating in a
scum, which differ in shape, size, favourable growth-, and scum forming conditions, but also in
flotation velocity. One solution for improvement of these model predictions could therefore be
the incorporation of flotation velocities and scum formation characteristics of the cyanobacterial
species that most commonly occur in the particular location.
LAKE RESTORATION IN THE NETHERLANDS
Restoration of shallow freshwater lakes remains a hot topic in The Netherlands, as well as
elsewhere, amongst others since many of the lakes suffer from internal P loading. Remediation
of this problem is often proposed by adding chemical P binding agents to the surface water of
sediment of a lake. Of the different chemical P binding agents, iron is a compound that was
naturally present in high quantities in lake sediments in The Netherlands, but due to changes
in water regimes such as damming and excess use of groundwater for agriculture and drinking
water, the input of this iron-rich groundwater (seepage) has decreased and consequently lake
sediments have gradually become iron depleted (Lamers et al., 2002; Van der Welle et al., 2007b).
Therefore, addition of chemicals that naturally occur in high quantities in lakes might be more
favourable and sustainable than adding substances which are not commonly found in lakes, such
as lanthanum-enriched benthonite clay.
The successful iron dose in order to regulate P release can be calculated by using the molar
Fe:P ratio in the pore water or the sediment of a lake. Various ratios are suggested in literature,
ranging from a pore water molar ratio of 1-3.5 (Smolders et al., 2001; Zak et al., 2004; Geurts
et al., 2008) or a ratio of 15 Fe:P by weight (Jensen et al., 1992) to a molar ratio of 8-10 for the
sediment (Hansen et al., 2003; Geurts et al., 2008), which would need to be reached or exceeded
to enable P retention in the (oxidised) sediment. High sulphate (SO4) and dissolved organic
carbon (DOC) concentrations can, however, facilitate internal eutrophication by competing with
PO4 for Fe anion adsorption sites, which ultimately results in mobilization PO
4 to the water
column (Zak et al., 2004; Smolders et al., 2006; Van der Welle et al., 2007b). To compensate
for this co-precipitation of Fe with other compounds, depending on lake characteristics, field
experiments therefore successfully used high iron doses of 100-200 g Fe m-2 (Walker et al., 1989;
Boers et al., 1994; Kleeberg et al., 2012). Although iron addition in these field experiments
significantly reduced lake P concentrations (Walker et al., 1989; Boers et al., 1994; Kleeberg et
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 1
18
al., 2012), possible side effects of this addition on the aquatic community remained unstudied,
whereas toxicity studies have shown that high amounts of iron can be toxic to various aquatic
organisms (Gerhardt and Westermann, 1995; Linton et al., 2007).
AIMS OF THIS THESIS
Iron addition
The first part of this thesis covers the aim to gain new insight in the efficacy of iron addition on
P reduction and possible negative side-effects of this addition on aquatic organisms, particularly
on macrophytes. Whereas a reduction of lake P concentrations ideally would shift a lake from
turbid algal dominated system to a clear macrophyte dominated system, iron addition might
hamper the recovery of macrophytes due to possible toxic side-effects. The effects of iron addition
on macrophyte growth and germination were therefore experimentally studied for a variety of
macrophyte species using a combination of small scale laboratory tests, in-lake mesocosms and a
whole lake iron addition experiment in the peaty Lake Terra Nova, The Netherlands (Figure 1.2).
In addition to the effects of iron on macrophytes, both water quality and community composition
of phyto- and zooplankton were evaluated during and up to two years after the whole lake iron
addition experiment.
Figure 1.2 – Addition of iron(III)chloride in Lake Terra Nova using a mobile wind-driven pump. (Photo by Gerard ter Heerdt)
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
General Introduction
19
1Scum prediction models
The second part of this thesis covers the aim to gain a better understanding of cyanobacterial
scum formation under wind induced turbulence, which could benefit scum prediction models.
The combined effort of water boards (Waternet) and research institutes (Deltares and NIOO)
in The Netherlands resulted in a scum prediction model EWACS (Early Warning Against
Cyano Scums), which due to the high number of false positive predictions, needed additional
information of species specific scum behaviour, in particular for persistent scums in sheltered
places (Burger et al., 2009). For this reason, the formation and disappearance of scums of two
different cyanobacterial species were studied in specially designed mesocosms with oscillating
grids which generate turbulence. The conclusions of this research were based on a combination of
experimental results and civil engineering theories.
Figure 1.3 – Schematic overview of the possible interactions (both direct and indirect) of iron addition on the aquatic foodweb covered in the first part of this thesis. Grey and orange arrows represent simplified consumptive foodweb interactions and possible iron addition effects, respectively. Letters indicate tested relationships, with (A) iron effects on P availability (Chapters 2 and 6), (B) iron effects on macrophyte growth and survival (Chapters 3, 4, 5 and 6), (C) herbivory by fish and crayfish suppressing macrophyte recovery (Chapters 5 and 6), and (D) iron effects on the aquatic community composition (Chapter 2 and 6). Figure adapted from http://www.pkgills.com/wp-content/uploads/2010/03/thefoodweb.png.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 1
20
OUTLINE OF THIS THESIS
The first and main part of this thesis focuses on iron addition as a restoration measure to combat
internal P loading and its effect on the aquatic environment (Figure 1.3). The second part of this
thesis describes the mesocosm experiments on the formation and disappearance of cyanobacterial
scums under the influence of turbulence.
Iron is a compound that naturally occurs in lakes, and is well known to be toxic to aquatic
organisms when concentrations exceed certain limits. Chapter 2 therefore covers a literature
study on direct and indirect effects of iron on the aquatic community (Figure 1.3; A, D). In
this study toxicity experiments are compared to iron restoration experiments in the field to get
a complete overview of possible side-effects of iron addition on both primary and secondary
producers.
Laboratory studies on the effect of iron addition on individual macrophytes are described
in Chapter 3 and 4, during which the eutrophic macrophytes Elodea nuttallii and Potamogeton
pectinatus (Chapter 3) and oligotrophic charophytes Chara globularis and Chara virgata (Chapter
4) were subjected to a gradual iron dosing of 20 and 40 g Fe m-2 (Figure 1.3; B). The effects
of the different iron concentrations and different types of iron dosing (e.g. in the water or a
combination of water and sediment) on macrophyte growth and propagule germination were
followed for a duration of 12 (Chapter 3) or 5 (Chapter 4) weeks. The question whether high
iron concentrations or invasive crayfish hamper the recovery of macrophytes in two closed-off
ponds in Terra Nova is addressed in Chapter 5 (Figure 1.3; B, C). By means of a full factorial
design, herbivory effects of the exotic crayfish Procambarus clarkii and other herbivores were
tested on macrophyte transplants in an iron rich and iron poor enclosure. To conclude the iron
research, using long-term monitoring data spanning a period of 27 years, Chapter 6 evaluates
the sequential effects of biomanipulation and a large scale iron addition on the water quality and
biotic communities (phytoplankton, zooplankton, and macrophytes) of the shallow eutrophic
peaty Lake Terra Nova (Figure 1.3; A-D).
The second part of this thesis is covered in Chapter 7, where cyanobacterial scum formation
under decreasing turbulence and scum disappearance under increasing turbulence was
investigated. In this chapter, scums of the cyanobacterial species Aphanizomenon flos-aquae and
Woronichinia naegeliana were subjected to changing oscillation speeds in specially designed 920 L
tanks (Limnotrons), after which depth measurements and civil engineering theories described the
distribution and scum behaviour of the two species.
Finally, in Chapter 8 the conclusions and implications of the previous chapters are discussed.
CHAPTER 2
Lake restoration by in-lake iron addition:
a review of iron impact on aquatic organisms and
lake ecosystems
Anne K. Immers, Ellen van Donk, and Elisabeth S. Bakker
Submitted to Freshwater Biology
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 2
22
ABSTRACT
Internal phosphorus loading has become a major problem in many freshwater lakes due to the
build-up of nutrient stocks in the sediment over the past decades. Iron is a natural capping agent
which naturally binds to P and can enhance sediment P binding capacity. Various restoration
experiments using iron addition have been a success by reducing P availability and shifting a
lake from an algal dominated state to a macrophyte dominated state. Adding iron to a lake could
however also negatively affect lake ecosystems, as iron could impose toxic effects on the biota. We
review therefore iron toxicity studies and lake restoration experiments using iron addition, and
combine this knowledge to formulate guidelines for lake restoration using iron addition without
posing extra risks to the environment. Iron toxicity studies reveal that even though iron is an
essential nutrient for growth, when added in excess, it can negatively affect aquatic organisms,
either directly due to toxic effects or indirectly due to precipitation of ironhydroxides. These
precipitations could alter food quality, food availability, habitat structure, and could attach to
vital parts of the aquatic organisms, resulting in stress and tissue damage. A review of restoration
studies using iron addition shows that several have successfully shifted eutrophic ecosystems
to macrophyte dominated oligotrophic ecosystems with higher biodiversity. However, in other
studies, local environmental constraints masked the effect of iron addition, resulting in moderate
or no effects of iron addition. Whereas high iron concentrations can have toxic effects on both
primary and secondary producers, these effects remained absent during field studies, as dilution
and chemical interactions quickly reduced the high amount of dissolved iron in the system. We
conclude that differences in species response to iron addition might lead to shifts in aquatic
communities, favouring the more iron-tolerant species. Furthermore, iron addition is effective
in lowering lake P concentrations, which could eventually have the most important effect on
the aquatic community composition. Long term effects of iron on the community composition,
however, have barely been tested and still remain largely unknown.
Guidelines and perspectives. The reviewed studies show that the following factors should
be taken into account when applying iron addition as a measure for lake restoration. In order to
regulate P release from the sediment, the amount of iron should be added to reach a sediment molar
Fe:P ratio of 7-10. To prevent a quick drop in pH and direct effects of high iron concentrations
during the iron addition period on aquatic organisms, slow addition of iron over a longer term
(months to a year) is necessary. Lastly, iron addition in lakes with high concentrations of organic
matter or other chemicals with high affinity for Fe does not increase P retention until these
concentrations have sufficiently decreased. These constraining environmental factors should be
addressed to improve the success rate of iron addition.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Lake restoration by in-lake iron addition: a review of iron impact on aquatic organisms and lake ecosystems
23
2
INTRODUCTION
The water quality of many freshwater lakes has been declining since the second half of the
20th century due to high input of nutrients, mainly phosphorus (P) and nitrogen (N), often
resulting in a shift from a clear macrophyte dominated system to a turbid algal dominated system
(Søndergaard et al., 2003; Smith and Schindler, 2009). Various restoration measures have been
proposed by both scientists and water managers to combat these changes and return these lakes
to their ‘natural’ situation which occurred prior to these eutrophication events. Great efforts
have been made ever since, largely by reducing external input of nutrients by either closing off
nutrient rich input sources or by pre-treating the nutrient rich water before it enters the lakes
(Klapwijk et al., 1982; Jeppesen et al., 1991; Van Liere and Janse, 1992). Yet a full recovery has
not been reached in many cases, as restoration measures are often hindered by internal loading
from nutrients that have been building up in the lake sediment (Cooke et al., 1993a; Søndergaard
et al., 2003; Smolders et al., 2006).
One way to combat internal loading is by adding chemical substances to a lake, such as
aluminium, calcium, or iron, which naturally bind to P (Cooke et al., 1993a; Burley et al.,
2001; Smolders et al., 2006; Kleeberg et al., 2013). Of these compounds, iron is a compound
that can be naturally found in high quantities in lake sediments, but due to changes in water
regimes such as damming and excess use of groundwater for agriculture, the input of iron-rich
groundwater has decreased and consequently lake sediments have become iron depleted (Lamers
et al., 2002; Van der Welle et al., 2007b). The addition of iron has frequently been used in the
past for pre-treatment of P-rich inlet water (Klapwijk et al., 1982; Bootsma et al., 1999), but it
has also successfully been used in both mesocosm experiments and the field to combat internal
P loading by either adding the iron to lake sediment (Quaak et al., 1993; Boers et al., 1994;
Smolders et al., 2001) or to the water column of a lake (Jaeger, 1994; Burley et al., 2001; Deppe
and Benndorf, 2002; Hansen et al., 2003; Kleeberg et al., 2012). Although the effects of this
restoration measure on biogeochemistry are well documented, the effects on different parts of the
foodweb are often not taken into account.
By adding iron to a lake to bind to the excess phosphate in the system, the lake is expected
to shift towards a clear water macrophyte dominated state which is generally considered positive
for biodiversity (Smith and Schindler, 2009). However iron, when added in excess, could also
negatively affect organisms as iron in high doses can be toxic. Different toxicity experiments
have been carried out in the lab, testing the effect (EC50
) or lethal (LC50
) doses of iron on various
animals and plants. These investigations into the impact of toxic metals have tended to rely on
only single species toxicity tests, whereas ecological effects of iron addition, such as competition,
plant-herbivore interactions, and predator-prey relationships eventually determine the ecosystem
impact. In this review we will aim to combine both lines of iron research (e.g. toxicity and
restoration studies) to indicate the potential effects of iron as a restoration measure on different
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 2
24
levels in the foodweb, from individual species to a whole lake ecosystem. First we will explore
the direct and indirect effects of iron on both primary and secondary producers and assess the
possible effects of iron addition on the aquatic community composition. Next we will evaluate
lake restoration studies using iron and determine guidelines/lessons for successful restoration,
both chemically and biologically.
IRON AND ITS BIOTIC ENVIRONMENT
Primary producers
The addition of iron can have several different effects on growth and reproduction of primary
producers, both direct and indirect (Wheeler et al., 1985; Snowden and Wheeler, 1993; Lucassen
et al., 2000). The element iron can form covalent bonds with many nutrients. The formation
of these bonds with essential nutrients, such as P, Mn, K, Ca, Mg, and Zn, can lead to nutrient
limitation and consequently to nutrient deficiencies within plants (Ponnamperuna, 1972;
Wheeler et al., 1985; Sahrawat, 2004). On the other hand, by forming covalent bonds with
excess P or highly insoluble metal-sulphides with sulphur (FeS, FeS2, or pyrite), iron can improve
water quality for plants and act as a detoxification mechanism by reducing the availability of
phytotoxins to plants (Smolders et al., 2001). Iron itself is also an essential nutrient for primary
producers, where it is involved in photosynthesis, chlorophyll synthesis, respiration, and
nitrogen assimilation (Lucaç and Aegerter, 1993). The essentiality however is limited to a certain
concentration, after which iron becomes toxic, the so called ‘window of essentiality’ (Walker et
al., 2012). At low concentrations, iron increases primary producers’ productivity, but at elevated
concentrations, iron can induce oxidative stress on a cellular level and disrupt cell membranes,
proteins, pigments, and even damage DNA, eventually leading to death of the organism (Linton
et al., 2007; Sinha et al., 2009; Keller et al., 2012). Moreover, high metal concentrations within
plants and algae can cause metal binding to the cell wall, which could reduce growth by inhibiting
nutrient uptake or efflux pumping of metals at the plasma membrane (Spijkerman et al., 2007).
Iron toxicity can also directly influence productivity and reproduction of plants by reducing
leaf size or causing leaf and shoot dieback, by forming necrotic spots on leaves, by inducing root
flaccidity, and by reducing root branching (Lucassen et al., 2000; Van der Welle et al., 2007a).
Until now these direct effects of iron toxicity have only been observed for terrestrial or emergent
wetland plant species growing on sediment with high pore water or sediment iron concentrations
of 50-68 mg L-1 and 109-438 mg g-1, respectively (Jones and Etherington, 1970; Wheeler et al.,
1985; Macfie and Crowder, 1987; Lucassen et al., 2000; Van der Welle et al., 2007a; Siqueira-
Silva et al., 2012), whereas some wetland species were already showing signs of iron toxicity at
pore water iron concentrations of 1 mg Fe L-1 (Batty and Younger, 2003). Adding iron (25-100 g
Fe m-2) to the water column did, however, not directly affect growth and physical appearance of
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Lake restoration by in-lake iron addition: a review of iron impact on aquatic organisms and lake ecosystems
25
2
the fully submerged aquatic species Elodea nuttallii, Potamogeton pectinatus (Immers et al., 2014),
P. acutifolius, Stratiotes aloides (Van der Welle et al., 2006, 2007b), Myriophyllum aquaticum (Kamal
et al., 2004), and the charophyte species Chara virgata and C. globularis (Immers et al., 2013).
Experiments with macrophytes growing in industrial metal-rich areas (8.6 mg Fe L-1) also showed
that aquatic macrophytes were able to grow well without showing any external abnormalities
(Nayek et al., 2010). While these macrophytes did not show any visible symptoms of iron stress,
the tolerance of aquatic plants to higher iron concentrations (10-100 mg Fe L-1) has been found to
be species specific and could be negatively related to growth rate (Snowden and Wheeler, 1993;
Nayek et al., 2010).
Effects of iron addition on phytoplankton have been intensely investigated for oceans, where
phytoplankton growth in certain ‘high nitrate, low chlorophyll areas’ is highly limited by iron
(Martin et al., 1991). Ocean iron addition on various occasions consequently resulted in an increase
of phytoplankton growth and abundance (Martin et al., 1991; Boyd et al., 2007). Freshwater
systems, however, differ greatly in nutrient composition and iron availability and iron addition
does therefore not necessarily yield the same response in phytoplankton growth. Micro- and
macronutrient addition experiments by Downs et al. (2008) showed that most phytoplankton
in freshwater lakes was limited by phosphate, although growth of certain heterocystous
cyanobacterial species was promoted by iron addition (1.6 mg Fe L-1 in a eutrophic lake) due
to the high Fe demands of these species for nitrogen assimilation. In contrast, iron addition
experiments with the freshwater green algae Pseudokirchneriella subcapita showed that additions
of 10 mg Fe2+ L-1 and 25 mg Fe3+ L-1 yielded lower growth rates compared to control conditions
without iron (Keller et al., 2012). Also toxin production in cyanobacteria can be affected by iron,
which decreases with higher iron concentrations (Lucaç and Aegerter, 1993), but this response
was not consistent for all tested cyanobacterial species (Utkilen and Gjolme, 1995).
Whereas iron addition eventually could alleviate light limitation by returning the ecosystem
to a macrophyte dominated state with high water transparency, it can simultaneously precipitate
as iron-hydroxides on plants and lake sediments, which in turn could induce light limitation and
inhibit growth of both plants and periphyton (Gerhardt and Westermann, 1995). Not only at
the surface of the plants, but also in the oxygenated sediment near the roots iron hydroxides are
formed, which can be visible as red plaques coating the surface of roots. When iron concentrations
in the water column or sediment are high, excess uptake of iron within plants may lead to the
formation of toxic Reactive Oxygen Species (ROS) within cells (Sinha et al., 2009). In order to
avoid cellular damage, oxygen can be excreted at the tips of roots, which in turn reacts with iron
to form iron oxyhydroxides. The plaques could serve as iron storage in case of iron shortage, serve
as a protective barrier against uptake of (other) toxic metals, but could also inhibit the uptake
of essential nutrients by the roots (Macfie and Crowder, 1987; Otte et al., 1989; St-Cyr and
Campbell, 1996). The effectiveness of the formation of plaques as a protection against hyper-
accumulation of iron within cells is however debated (Siqueira-Silva et al., 2012).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 2
26
Tab
le 2
.1 –
Eff
ect
(EC
50) a
nd le
thal
dos
e (L
C50
) tes
ts o
f iro
n on
zoo
plan
kton
, mac
roin
vert
ebra
tes,
am
phib
ians
, and
fish
as
repo
rted
in li
tera
ture
.
Spec
ies
Ord
erE
C50
(mg
Fe L
-1)
LC50
(mg
Fe L
-1)
Ref
eren
ce
48 h
96 h
48 h
96 h
Ann
elid
a
Bra
nchi
ura
sow
erby
iO
ligo
chae
te58
0M
ukho
padh
yay
and
Kon
ar, 1
984
Nai
s eli
ngui
sO
ligo
chae
te0.
12Sh
uhai
mi-
Oth
man
et
al.,
2012
a
Tubi
fex
tubi
fex
Oli
goch
aete
101.
8410
1.84
Kha
ngar
ot, 1
991
Mol
lusc
a
Mel
anoi
des t
uber
cula
taG
astr
opod
a21
.78
8.49
Shuh
aim
i-O
thm
an e
t al
., 20
12b
Lym
naea
acu
min
ata
Gas
trop
oda
Kha
ngar
ot a
nd R
ay, 1
989
Phy
sell
a gy
rina
Gas
trop
oda
12.0
9B
irge
et
al.,
1985
; Shu
haim
i-O
thm
an e
t al
., 20
12b
Pla
norb
ariu
s sp.
Gas
trop
oda
7.32
Furm
ansk
a, 1
979
Sem
isul
cosp
ira
libe
rtin
aG
astr
opod
a76
.0N
ishi
uchi
and
Yos
hida
, 197
2
Cru
stac
ea
Ase
llus
aqu
atic
usIs
opod
a81
.112
4.0
Furm
ansk
a, 1
979;
Ger
hard
t, 1
994
Cra
ngon
yx p
seud
ogra
cili
sA
mph
ipod
a12
0.0
Mar
tin
and
Hol
dich
, 198
6
Che
rax
dest
ruct
orD
ecap
oda
50.0
Kha
n an
d N
ugeg
oda,
200
7
Cyc
lops
vir
idis
Cop
epod
a35
.2M
ukho
padh
yay
and
Kon
ar, 1
984
Dap
hnia
long
ispi
naC
lado
cera
11.4
8R
anda
ll e
t al
., 19
99
Dap
hnia
mag
naC
lado
cera
7.2
5.9
Kha
ngar
ot a
nd R
ay, 1
989;
Bie
sing
er a
nd C
hris
tens
en, 1
972
Mac
robr
achi
um la
nche
ster
iD
ecap
oda
3.72
Shuh
aim
i-O
thm
an e
t al
., 20
12a
Sten
ocyp
ris m
ajor
Ost
raco
da0.
28Sh
uhai
mi-
Oth
man
et
al.,
2012
a
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Lake restoration by in-lake iron addition: a review of iron impact on aquatic organisms and lake ecosystems
27
2
Inse
cta
Chi
rono
mus
java
nus
Dip
tera
0.62
Shuh
aim
i-O
thm
an e
t al
., 20
12a
Lep
toph
lebi
a m
argi
nata
Eph
emer
opte
ra70
.010
6.3
Ger
hard
t, 1
994
Cho
rdat
a
Buf
o ja
poni
cus
Anu
ra4.
2H
ashi
mot
o an
d N
ishi
uchi
, 198
1
Dut
taph
rynu
s mel
anos
tict
usA
nura
0.6
0.4
Nis
hiuc
hi a
nd Y
oshi
da, 1
972;
Shu
haim
i-O
thm
an e
t al
., 20
12a
Poe
cili
a re
ticu
lata
Cyp
rino
dont
ifor
mes
1.46
Shuh
aim
i-O
thm
an e
t al
., 20
12a
Ran
a he
xada
ctyl
aA
nura
17.6
Kha
ngar
ot a
nd R
ay, 1
989
Ran
a li
mno
char
isA
nura
79.7
Pan
and
Lia
ng, 1
993
Ras
bora
sum
atra
naC
ypri
nifo
rmes
1.71
Shuh
aim
i-O
thm
an e
t al
., 20
12a
Salm
o tr
utta
Salm
onid
ae47
.0D
alze
ll a
nd M
acfa
rlan
e, 1
999
Til
apia
mos
ambi
caP
erci
form
es11
9.6
Muk
hopa
dhya
y an
d K
onar
, 198
4
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 2
28
Secondary producers
Even though animals require iron for haemoglobin in blood cells and various enzymes (such as
cytochromes which are involved in ATP production) and use iron as a detoxification mechanism
against heavy metals (Vuori, 1995), iron can cause serious damage to the animal when
concentrations reach beyond the window of essentiality. High concentrations of iron can on a
cellular level disrupt cell membranes, damage DNA, and enhance lipid peroxidative damage
through the formation of ROS (Gerhardt and Westermann, 1995). Moreover, iron may also affect
behaviour or life cycle strategies as high iron concentrations can cause a decrease in number of
offspring (Dave, 1985; Myllynen et al., 1997), reduce the viability of offspring (Myllynen et al.,
1997; Van Anholt et al., 2002; Sotero-Santos et al., 2005), increase susceptibility to bacterial
pathogens (Sealey et al., 1997), and interfere with digestion and consequently reduce the uptake
of nutrients (Gerhardt, 1992; Van Anholt et al., 2002). The severity of these effects is strongly
coupled to the concentration of iron encountered by the animal and differs greatly among species.
Direct toxicity tests
Direct toxicity experiments have been carried out on many occasions to test the effect (EC50
) and
lethal dose (LC50
) of iron on both benthic and pelagic animals (Table 2.1). These tests often used
high concentrations of iron to represent lakes or rivers which had been acidified or polluted with
heavy metals due to mining or other industrial activities (Wepener et al., 1992; Van Anholt et
al., 2002; Verberk et al., 2012). The results clearly show a big difference in the response of the
tested animals to iron concentrations, even among species of the same order (Table 2.1). The
high variation could partly be explained as dissolved and particulate iron, iron speciation, water
hardness, possible effects of iron addition on pH, and concentrations of other toxic metals were
not always carefully separated. In the case of Daphnia, for example, Biesinger and Christensen
(1972) showed that relatively low additions of iron(III)chloride impaired survival of both adult
and young. Yet follow-up experiments showed that Daphnia magna and D. longisperma seemed
unaffected by higher dissolved iron concentrations and that the particulate nature of the added
iron sulphate and the decrease in pH caused the mortalities and reduced number of broods, not
the toxicity of the metal itself (Randall et al., 1999; Van Anholt et al., 2002). Acute toxicity
experiments with FeCl3 yielded low LC
50 values for other pelagic animals, such as for the warm
water fish Rasbora sumatrana and Poecilia reticulata and the amphibian species Duttaphrynus
melanostictus (Shuhaimi-Othman et al., 2012a). Even though iron addition showed physical
damage within tissues of these animals (Shuhaimi-Othman et al., 2012a), the animals were tested
in water with low water hardness, whereas low water hardness has been known to increase toxicity
of metals to organisms (Khangarot, 1991). Moreover, according to Randall et al. (1999), acute
iron toxicity rarely occurs in fish, but chronic toxicity might occur after prolonged exposure.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Lake restoration by in-lake iron addition: a review of iron impact on aquatic organisms and lake ecosystems
29
2
Various benthic and pelagic macroinvertebrates show a high tolerance for iron, such as the
mollusc Melanoides tuberculata which could withstand high concentrations of iron by closing
its tightly sealing operculum (Gerhardt, 1992; Shuhaimi-Othman et al., 2012a). The mayfly
Leptophlebia marginata stopped feeding during the time of high iron concentrations up to 50 mg
Fe L-1 and showed 95% survival after two weeks (Gerhardt, 1992). However, after prolonged
exposure of high iron concentrations, the mayflies started to die due to starvation and constipation
(Gerhardt, 1992). A big difference was found between the different oligochaete species and their
tolerance for iron. Whereas Tubifex tubifex and Branchiura sowerbyi could withstand extremely
high iron concentrations (Mukhopadhyay and Konar, 1984; Khangarot, 1991), Nais elinguis was
only able to survive very low concentrations (Shuhaimi-Othman et al., 2012a). Nonetheless, it
is not clear whether other confounding factors such a low pH were carefully separated during
these iron toxicity tests. According to Chapman et al. (1982), the tolerance for low pH was for
both T. tubifex and B. sowerbyi relatively low (3.6 and 3.7 respectively), which could therefore
indicate that a drop in pH after iron addition in the N. elinguis toxicity tests had interfered
with the results. Aquatic oligochaete species are often used as environmental indicators for water
quality due to the fact that some species can withstand highly polluted areas whereas others are
only found in unpolluted areas. Therefore pollution tolerance, or in this case iron tolerance, is
for oligochaetes species specific, even for species within the same genus (Chapman et al., 1982).
Iron toxicity under natural conditions
Where in iron toxicity studies iron and pH effects need to be carefully separated, during field
experiments these effects may occur together and could increase toxicity, not to mention co-
precipitation of other toxic metals. Moreover, iron could precipitate as iron hydroxides, which
could alter food quality, food availability, habitat structure, and could attach to vital parts of the
animal, resulting in stress and tissue damage (Gerhardt and Westermann, 1995; Vuori, 1995;
Linton et al., 2007; Siqueira-Silva et al., 2012). These indirect effects of iron precipitates on
animals, plants, lake sediment, and other surfaces have shown to be eventually more detrimental
to animal growth than possible toxic effects of iron within cell tissues (Gerhardt and Westermann,
1995; Vuori, 1995; Linton et al., 2007).
Iron hydroxide precipitations both above- and belowground (iron plaque layers) can decrease
periphyton and plant growth, which could lead to a decrease in food quality and availability
for herbivores (Gerhardt and Westermann, 1995). Moreover, when ingested ironhydroxide
precipitates can attach to gill and gut membranes disturbing animal metabolism and mobility,
thereby restricting foraging behaviour (Rasmussen and Lindegaard, 1988; Gerhardt and
Westermann, 1995; Siqueira-Silva et al., 2012). Iron hydroxide layers on the sediment could
alter the structure and quality of benthic habitats and destroy spawning grounds for fish
(Rasmussen and Lindegaard, 1988; Gerhardt and Westermann, 1995; Linton et al., 2007).
Direct accumulation of iron precipitates on fish and macroinvertebrate gills has led to restricted
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 2
30
respiration in various animals (Gerhardt and Westermann, 1995; Vuori, 1995; Linton et al.,
2007). Moreover, precipitated iron deposits on eggs showed a decrease in hatching success as
the iron clogged the egg pores, resulting in suffocation of the offspring (Vuori, 1995; Linton et
al., 2007). Nonetheless, these negative effects of iron precipitates on zooplankton and fish were
not observed during the iron addition restoration experiment of Jaeger (1994), even though the
sediment was covered with an ironhydroxide layer and surface water iron concentrations reached
4 g Fe m-3.
EFFECTS OF IRON ON COMMUNITY SHIFTS
As shown in the previous paragraphs, iron can have several negative and positive effects on species,
both primary and secondary producers. Therefore, iron addition in the field is expected to induce
changes on a community level due to the differences in iron tolerance between species or groups
of species. The formation of iron precipitates on plants has for example been observed to restrict
the distribution of various plant and periphyton species in streams (Vuori, 1995). Therefore,
differences in plant responses to iron addition, both direct and indirect, might lead to a shift
in community composition, favouring growth of the more iron-tolerant species. Nonetheless,
Geurts et al. (2008) showed that the occurrence of endangered plant species such as charophytes
was related to high Fe:PO4 ratios in the sediment pore water of peat lakes, indicating that iron
addition could therefore lead to a higher abundance and diversity of endangered macrophyte
species. Additionally, the germination of several charophyte species from peat sediment was
not hindered by iron additions up to 40 g Fe m-2 (Immers et al., 2014). Therefore, the shift
in community composition after iron addition would not necessarily lead to dominance of fast
growing macrophyte (or algal) species.
Differences in iron tolerance between macroinvertebrate species has also been shown to affect
community composition. High iron concentrations in a Danish lowland river led to a decrease
in macroinvertebrate taxa, with only the taxa Tubificidae, Chironomidae, and Tipulidae present,
whereas the pollution sensitive taxa Ephemeroptera and Plecoptera were confined to areas with
low iron concentrations (Rasmussen and Lindegaard, 1988). Diversity of macroinvertebrates was
shown to decrease at iron concentrations above 1.2 mg Fe L-1, but even at low concentrations of
0.2 – 0.3 mg Fe L-1 the number of macroinvertebrate taxa decreased from 67 to 53 (Rasmussen
and Lindegaard, 1988; Gerhardt and Westermann, 1995). Moreover, precipitations of iron on
plants, periphyton, and sediments have shown to eliminate macroinvertebrate grazers which
feed on biofilm and periphyton (Rasmussen and Lindegaard, 1988). According to Gerhardt and
Westermann (1995), iron tolerance in macroinvertebrates is related to high nutrient tolerance
and the same species that are found in eutrophicated areas are also found in areas with high
iron concentrations. In contrast, Chapman et al. (1982) showed that oligochaetes adapted to
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Lake restoration by in-lake iron addition: a review of iron impact on aquatic organisms and lake ecosystems
31
2
oligotrophic conditions were more tolerant to high metal concentrations (mercury and cadmium)
than species adapted to eutrophic conditions. Nonetheless, metal tolerance in macroinvertebrates
changed with varying environmental conditions, such as shifts in pH and temperature (Chapman
et al., 1982).
Higher iron requirements of certain species can also induce community changes such as
in phytoplankton communities, where iron additions have caused a shift towards N-fixing
cyanobacterial species over green algae (Morton and Lee, 1974; Downs et al., 2008; Molot et
al., 2010). In this case iron availability changed and the species with higher iron requirements,
such as the heterocystous cyanobacteria, could grow faster, resulting in a shift in phytoplankton
dominance (Downs et al., 2008; Molot et al., 2010).
Lastly, iron additions can change communities due to behavioural avoidance, as was shown by
Verberk et al. (2012) for two stickleback species. Verberk et al. (2012) concluded that the three-
spined stickleback showed behavioural avoidance to areas with high iron concentrations, whereas
the nine-spined stickleback preferred these areas. Nonetheless after iron concentrations were
reduced, the three-spined stickleback returned to the formerly iron contaminated areas (Verberk
et al., 2012). This non-lethal effect of high iron concentrations on community composition was
also shown for other fish and benthic invertebrates (Rasmussen and Lindegaard, 1988; Gerhardt
and Westermann, 1995; Vuori, 1995; Randall et al., 1999).
While all previous mentioned consequences of high iron concentrations could result in
considerable changes in the community composition of the aquatic ecosystem, high iron
concentrations in the water could bind to excess P in the system, thereby shifting eutrophic
ecosystems to macrophyte dominated oligotrophic ecosystems with higher biodiversity, which
could eventually have the most important effect on community composition (Jeppesen et al., 2012).
Lower P concentrations in the water favour macrophyte over phytoplankton growth, resulting
in an increase in water transparency. Moreover, excess iron could bind to phytotoxins, such as
sulphate, thereby decreasing their availability to freshwater organisms. As high concentrations of
iron in the water can potentially be toxic to the aquatic community, the concentration of excess
(non-bound) iron in the water will therefore define whether these indirect positive effects of iron
(low P and phytotoxin concentrations) will prevail over the potential negative effects of iron
addition.
IRON AS RESTORATION MEASURE
Addition of iron: chemical interactions
The goal of adding iron to the sediment or surface water of a lake is to bind the available
phosphate in the water and form a ‘phosphate-trap’ on the sediment-water interface. The binding
capacity of Fe is however regulated by the redox state in the top layer of the sediment (Lijklema,
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 2
32
1977; Burley et al., 2001; Smolders et al., 2006). Under oxic conditions, oxidized ferric iron
(Fe3+) can freely precipitate with PO4, but under anoxic or reduced conditions, ferrous iron (Fe2+)
is formed and Fe loses this binding capacity and consequently PO4 will be released from the
sediment (Mortimer, 1941; Lijklema, 1977; Cooke et al., 1993a; Golterman, 2001). Moreover,
high sulphate (SO4) concentrations can facilitate internal eutrophication by competing with PO
4
for Fe anion adsorption sites, which ultimately results in mobilization of previously bound PO4
to the water column (Smolders et al., 2006; Van der Welle et al., 2007b). Additionally, high
SO4 reduction rates lead, under anaerobic conditions, to the formation of toxic sulphide (S2-),
which reduces the formed iron-phosphates to form FeSx (Smolders et al., 2006). Therefore, Fe
addition to reduce internal P loading can only be successful when the top layer of the sediment is
oxidized and when SO4 concentrations are low or when sufficient Fe is added to cope with these
SO4 interactions.
The success of iron addition in order to regulate P release can be calculated by using the Fe:P
ratio in the pore water of the sediment. Various ratios are suggested in literature, ranging from
a molar pore water ratio of 1-3.5 (Smolders et al., 2001; Zak et al., 2004; Geurts et al., 2008)
or a ratio of 15 Fe:P by weight (Jensen et al., 1992) to a molar ratio of 8-10 for the sediment
(Hansen et al., 2003; Geurts et al., 2008), which would need to be reached or exceeded to enable
P retention in the (oxidised) sediment. In order to increase pore water Fe:P ratios in lakes, field
studies have added different iron salts as a restoration measure, which included FeCl3, FeCl
2,
FeSO4, and Fe
2O
3, with or without extra aeration with oxygen in the lake (Quaak et al., 1993;
Boers et al., 1994; Jaeger, 1994; Smolders et al., 2001; Hansen et al., 2003). Additional aeration
of iron treatments with O2 did not yield better results regarding P retention, provided enough
oxidised Fe was present in the upper layer of the sediment (Jensen et al., 1992; Hansen et al.,
2003). This is in accordance with Kleeberg et al. (2013), who showed that no additional aeration
is needed to oxidise Fe. According to Kleeberg et al. (2013), the success of iron addition is not
hindered by the redox sensitivity of iron as P can be efficiently precipitated independent of the
nature of the oxygen supply. That is, when iron is added to reach a sediment molar Fe:P ratio of
7. These conditions will assure continuous P elimination independent of oxygen supply, as both
will be released from the sediment in a ratio close to 1 and will co-precipitate due to natural
oxygenation processes (Kleeberg et al., 2013).
Additionally, humic compounds can form stable humic-iron complexes with iron, which
could inhibit the formation of ironphosphates and ironoxides (Myllynen et al., 1997; Zak et
al., 2004; Spijkerman et al., 2007). High concentrations of organic matter and other chemicals
with high affiliation to Fe (such as sulphate) therefore need to be measured before application in
order to calculate the appropriate iron dose. Iron addition in organic-rich lakes does not increase
P retention until DOC concentrations have sufficiently decreased (Zak et al., 2004). Part of
the reactive Fe will bind with humic compounds, thereby lowering the effective iron dose to
immobilise sediment P. For instance, in lake Groß-Glienicke (Germany) a high dose of 200 g
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Lake restoration by in-lake iron addition: a review of iron impact on aquatic organisms and lake ecosystems
33
2
Fe m-2 was needed to bind all sediment P and compensate for this co-precipitation of iron with
organic matter (Kleeberg et al., 2012).
Lastly, due to the low pH of iron(III)chloride, adding large quantities of iron may lead to a
drop in pH, which in turn leads to increased solubility of other metals in water. To prevent a
quick drop in pH and direct effects of high iron concentrations during the iron addition period
on aquatic organisms, slow addition of iron over a longer term (months to a year) is necessary.
Lake restoration by iron addition - lessons learned
The success of iron addition as a restoration measure, by lowering P concentrations without
imposing negative effects on the aquatic community, depends on the dose and environmental
conditions. Several iron addition experiments have been performed in the past, which results
could be used to explore guidelines for successful restoration, both chemically and biologically.
We have compiled the results of these iron addition experiments in Table 2.2. Of these
experiments, 3 were performed by adding iron compounds to sediment cores in the lab and 10
by adding iron in the field (lake or pond), either to the sediment (2 occasions) and/or to the water
column (9 occasions). P retention increased in all experiments using iron salts (FeCl2, FeCl
3, and
FeSO4), whereas it was barely affected after addition of Fe
2O
3 (Smolders et al., 2001; Table 2.2).
While the decrease in P concentrations in these experiments resulted in decreased
concentrations of chlorophyll, the longevity of these restorations was in some cases cut short due
to a variety of factors influencing both P concentrations and macrophyte success. The short term
success was in these cases due to either high external P loading (Boers et al., 1994), short water
retention time (Boers et al., 1994), heavy wind effects or seasonal turnover (Quaak et al., 1993;
Walker et al., 1989), a high population of plankti- and benthivorous fish (Van Donk et al., 1994),
or invasive crayfish inhibiting the development of submerged macrophytes (Van der Wal et al.,
2013; Table 2.2). Therefore, the success of iron addition as a restoration measure is affected by
location specific confounding factors, which may obscure the effects of iron addition itself.
The longevity of the success of iron addition also appears to depend greatly on the type of lake.
Addition in soft-water lakes, reservoirs and deep dimictic lakes yielded positive results for up to
eight years (Daldorph and Price, 1994; Jaeger, 1994; Kleeberg et al., 2012; Table 2.2), whereas iron
addition in alkaline lakes proved to be only a temporary solution due to elevated concentrations
of phosphate and sulphate with high affiliations for Fe (Geurts, 2010; Table 2.2). Therefore,
addition of iron in these lakes might need to be repeated to ensure positive effects on water
quality. On the other hand, iron addition might not be the best suitable measure for restoration
of lakes with high consumption rates of iron due to high P, SO4, and OM concentrations, which
are generally lower in lakes with sandy or clay sediments. For that reason, the choice of capping
agent depends on site specific conditions, and the use of other additions, such as aluminium or
lime, could in that case also be considered (Cooke et al., 1993a; Burley et al., 2001).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 2
34
Tab
le 2
.2 –
An
over
view
of p
erfo
rmed
res
tora
tion
exp
erim
ents
usi
ng ir
on a
ddit
ion
and
thei
r ef
fect
on
both
P r
eten
tion
and
aqu
atic
bio
ta.
Res
tora
tion
ex
peri
men
tsFi
eld
/ Lab
Loca
tion
/ O
rigi
n se
dim
ent
Add
itio
nA
mou
ntLo
cati
on o
f ad
diti
onE
ffec
t on
P
rete
ntio
nR
epor
ted
effe
cts
on o
rgan
ism
s an
d/or
the
lake
eco
syst
em
Bur
ley
et a
l., 2
001
Lab,
sed
imen
t co
res
Cro
oked
Lak
e,A
mis
k La
ke a
nd
Bap
tist
e La
ke, C
anad
a
FeC
l 3, Fe
Cl 3 +
O2
100
g Fe
m-2
Wat
er
colu
mn
Pos
itiv
eN
ot a
vail
able
Dal
dorp
h an
d P
rice
, 19
94Fi
eld
Foxc
ote
Res
ervo
ir,
Eng
land
FeSO
43.
5 m
g Fe
L-1
Wat
er
colu
mn
Pos
itiv
eR
eser
voir
shi
fted
from
ph
ytop
lank
ton
dom
inat
ed t
o m
acro
phyt
e do
min
ated
sys
tem
th
ree
year
s af
ter
dosi
ng
Dep
pe a
nd
Ben
ndor
f, 20
02Fi
eld
Bau
tzen
Res
ervo
ir,
Ger
man
yFe
Cl 3,
FeC
l 2, Fe
ClS
O4
21.3
and
18.
7 g
Fe m
-2W
ater
co
lum
nP
osit
ive
Not
ava
ilab
le
Geu
rts,
201
0Fi
eld,
m
esoc
osm
sLa
ke U
ddel
mee
r, T
he N
ethe
rlan
dsFe
Cl 3 /
FeC
l 250
and
100
g F
e2+
m-2/ 5
and
10
g Fe
3+ m
-2
Sedi
men
t / W
ater
co
lum
n
Pos
itiv
eC
hlor
ophy
ll a
nd s
uspe
nded
m
atte
r de
crea
sed.
Mac
roph
ytes
re
mai
ned
abse
nt d
ue t
o th
e an
aero
bic
sedi
men
t w
hich
co
unte
ract
ed g
erm
inat
ion
Han
sen
et a
l., 2
003
Lab,
sed
imen
t co
res
Lake
Ved
sted
, D
enm
ark
FeC
l 3 (±
O2)
3.7
mm
ol (F
e:P
O4
= 1
0)W
ater
co
lum
nP
osit
ive
Not
ava
ilab
le
Jaeg
er, 1
994
Fiel
dLa
ke K
rupu
nder
, G
erm
any
FeC
lSO
4 + O
2 (‘F
erri
Flo
c’)
5 g
Fe m
-3W
ater
co
lum
nP
osit
ive
No
fish
kill
s or
adv
erse
eff
ects
by
iron
-hyd
roxi
de-fl
akes
on
the
zoo
plan
kton
wer
e ob
serv
ed d
urin
g or
aft
er ir
on
prec
ipit
atio
n
Kle
eber
g et
al.,
20
12Fi
eld
Lake
Gro
ß-G
lien
icke
, Ger
man
yFe
(OH
) 3, Fe
Cl 2
250
g Fe
m-2
Wat
er
colu
mn
Pos
itiv
eC
hlor
ophy
ll d
ecre
ased
si
gnifi
cant
ly
Qua
ak e
t al
., 19
93;
Boe
rs e
t al
., 19
94Fi
eld
Gro
ot V
ogel
enza
ng,
The
Net
herl
ands
FeC
l 310
0 g
Fe m
-2Se
dim
ent
Pos
itiv
eD
urab
ilit
y of
pos
itiv
e ef
fect
s w
as o
nly
3 m
onth
s du
e to
sh
ort
wat
er r
esid
ence
tim
e of
lake
(35
days
) and
hig
h ex
tern
al lo
adin
g
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Lake restoration by in-lake iron addition: a review of iron impact on aquatic organisms and lake ecosystems
35
2
Smol
ders
et
al.,
2001
Lab,
sed
imen
t co
res
De
Bru
uk,
The
Net
herl
ands
FeC
l 3, Fe
Cl 2,
FeSO
4
150,
500
and
150
0 m
g Fe
kg-1
Sedi
men
tP
osit
ive
Not
ava
ilab
le
Smol
ders
et
al.,
2001
Lab,
sed
imen
t co
res
De
Bru
uk,
The
Net
herl
ands
Fe2O
315
0, 5
00 a
nd 1
500
mg
Fe k
g-1Se
dim
ent
Neg
ativ
eN
ot a
vail
able
Ter
Hee
rdt
et a
l.,
2012
Fiel
dTe
rra
Nov
a,
The
Net
herl
ands
FeC
l 333
g F
e m
-2W
ater
co
lum
nP
osit
ive
No
visu
al e
ffec
ts o
n fis
h.
Dec
reas
e in
chl
orop
hyll
, su
spen
ded
mat
ter,
and
phyt
opla
nkto
n bi
ovol
ume
Van
der
Wal
et
al.,
2013
Fiel
d,
mes
ocos
ms
Terr
a N
ova,
T
he N
ethe
rlan
dsFe
Cl 3
85 g
Fe
m-2
Wat
er
colu
mn
Pos
itiv
eD
ecre
ase
in c
hlor
ophy
ll
and
susp
ende
d m
atte
r. M
acro
phyt
es r
emai
ned
abse
nt
due
to h
igh
abun
danc
e of
ex
otic
cra
yfish
Van
Don
k et
al.,
19
94Fi
eld,
m
esoc
osm
sLa
ke B
reuk
elev
een,
T
he N
ethe
rlan
dsFe
Cl 3
28.9
mg
Fe L
-1W
ater
co
lum
nN
oIr
on a
ddit
ion
did
not
have
an
y ef
fect
on
chlo
roph
yll o
r su
spen
ded
mat
ter
Wal
ker
et a
l., 1
989
Fiel
dV
adna
is L
ake,
USA
FeC
l 3 + O
210
0 kg
Fe
day-
1W
ater
co
lum
nP
osit
ive
Dur
abil
ity
of p
osit
ive
effe
cts
was
cut
sho
rt d
ue t
o au
tum
n tu
rnov
er o
f the
lake
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 2
36
Whereas the previous paragraphs have shown that high iron concentrations can have toxic
effects on both primary and secondary producers, these effects remained absent during the
restoration experiments which monitored biological effects (Table 2.2). One explanation for this
could be that the high iron concentrations used in the iron toxicity studies in Table 2.1 are rarely
reached during restoration experiments with iron addition, as dilution and chemical interactions
quickly reduce the amount of dissolved iron in the system. For example, addition of 40 and
250 g Fe L-1 to the water column by Immers et al. (2014) and Kleeberg et al. (2012) resulted in
dissolved iron concentrations in the water column of only 0.12 and 0.2 mg L-1, respectively. On
the other hand, Kleeberg et al. (2012) noted that sediment iron concentrations after iron addition
reached high values of 533 g Fe L-1. When comparing these sediment iron concentrations to
EC50
and LC50
values of benthic organisms in Table 2.1, these concentrations would have a severe
impact on the aquatic life. Nonetheless, the bioavailability of the iron will be much lower as most
iron will be bound to either phosphates or sulphates in the aerobic top layer of the sediment.
CONCLUSIONS
Differences in species response to iron addition might lead to shifts in aquatic communities,
favouring the more iron-tolerant species. Nevertheless, various experiments and lake restoration
measures have shown that iron addition is effective in lowering lake P concentrations, shifting
the lake towards a clear macrophyte dominated system without hampering the germination and
development of various valuable macrophyte species (Daldorph and Price, 1994; Jaeger, 1994;
Ter Heerdt et al., 2012; Table 2.2). Due to interactions between iron and its environment, it still
remains difficult to predict the effects of iron addition on aquatic life. Precipitation of iron can
negatively affect the benthic macroinvertebrate community on the short term, but due to wind
induced mixing and bioactivity in the sediment surface, these iron-hydroxides would on the long
term gradually be mixed into the sediment. Short-term effects of iron addition in lake restoration
are neutral to positive, but the long term effects of iron addition on the aquatic life still remain
largely unknown.
We conclude that iron addition as a lake restoration measure can yield positive results
by lowering P availability and improving both water transparency and the development of
the macrophyte vegetation. However, environmental constraints should be addressed before
considering the use of iron addition. Iron addition seems most successful when external P loading
and concentrations of organic matter and other chemicals with high affiliation to Fe are reduced.
Moreover, submerged macrophytes can only develop when herbivory and sediment upwelling by
benthivorous fish and invasive crayfish are low.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Lake restoration by in-lake iron addition: a review of iron impact on aquatic organisms and lake ecosystems
37
2
ACKNOWLEDGEMENTS
This study was funded by the Water Framework Directive Innovation Fund from Agentschap NL
from the Dutch Ministry of Economic Affairs, Agriculture and Innovation.
CHAPTER 3
Iron addition as a measure to restore water quality:
implications for macrophyte growth
Anne K. Immers, Kirsten Vendrig, Bas W. Ibelings, Ellen van Donk,
Gerard N. J. ter Heerdt, Jeroen J. M. Geurts, and Elisabeth S. Bakker
Aquatic Botany (2014) 116, 44-52.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
40
ABSTRACT
Eutrophication of shallow lakes in North-West Europe has resulted in cyanobacterial blooms,
turbid water, and a decline in submerged macrophytes. Even though external inputs of phosphorus
(P) are declining, internal loading of P from the sediment may delay the recovery of these aquatic
ecosystems. Iron can be a useful chemical binding agent to combat internal P loading in shallow
lakes, but may potentially be harmful for macrophyte growth. We tested whether iron addition
as a restoration measure harms the growth of submerged macrophytes. We hypothesized that
this depends on the iron dosage and the rooting strategy of the macrophytes. We experimentally
tested the effects of Fe (FeCl3) on the submerged macrophytes Potamogeton pectinatus L. and Elodea
nuttallii (Planch.) H. St. John. Iron was dosed at a concentration of 20 g Fe m-2 and 40 g Fe m-2
to the surface water or to both the surface water and sediment. Elodea nuttallii growth was not
affected by iron addition, whereas P. pectinatus growth significantly decreased with increasing
iron concentrations. Nonetheless, biomass of both species increased in all treatments relative to
starting conditions. During the experiment, propagules sprouted from a propagule bank in the
sediment including species with a high conservation value and this spontaneous emergence was
not influenced by increasing iron concentrations. We conclude that adding iron(III)chloride in
dosages of 20-40 g m-2 may reduce growth of some macrophyte species, but does not prevent
overall macrophyte recovery. It may however affect macrophyte community composition due to
differential responses of macrophyte species to iron addition.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
41
3
INTRODUCTION
High nutrient loading from agricultural runoff and wastewater discharge during the second half
of the 20th century has led to eutrophication of many shallow lakes in north-western Europe. The
excess input of phosphorus (P) and nitrogen (N) has resulted in (toxic) cyanobacterial blooms and
subsequently turbid water, biodiversity loss, and a decline in submerged macrophytes (Tilman et
al., 2001; Hilt et al., 2006; Hickey and Gibbs, 2009). Submerged macrophytes play a key role in
the functioning of shallow water bodies by acting as a nutrient sink, providing a habitat for fauna,
and preventing resuspension of lake sediment. Through these actions macrophytes stabilize the
clear water state of shallow lake ecosystems (Scheffer et al., 1993; Jeppesen et al., 1998; Bakker
et al., 2010). After eutrophication, a strong reduction in P loading of a lake is required to restore
a lake to this self-stabilizing clear water state (Cooke et al., 1993a; Jaeger, 1994). However,
internal loading of P from the sediment, particularly from nutrient rich organic lake sediment
(Lamers et al., 2002), may significantly delay the recovery of aquatic ecosystems, even after the
external loading has been reduced (Cooke et al., 1993a; Jeppesen et al., 1998; Søndergaard et al.,
2003, 2013).
Before the intensification of agriculture, many peaty lakes would not suffer from high internal
P loading, as iron in upwelling groundwater naturally binds to phosphorus (in the form of
phosphate, PO4; Lamers et al., 2002). However, the upwelling of this iron-rich groundwater
has declined due to changes in hydrological regimes and desiccation through extraction of
groundwater for agricultural purposes, which consequently has led to a reduction in the amount
of iron reaching the top layer of the sediment (Van der Welle et al., 2007b). Hence, one way to
cope with internal P loading is by improving the P binding capacity of the lake sediment by
adding iron (Fe) or other chemical P binding agents such as aluminium (Al), calcium (Ca), or
lanthanum-enriched benthonite clay (Phoslock®) to the sediment (Cooke et al., 1993a; Burley et
al., 2001; Smolders et al., 2006; Hickey and Gibbs, 2009; Van Oosterhout and Lürling, 2011).
These chemical binding agents, if added on a regular basis, will not only precipitate with the
available PO4 in the sediment, but can potentially provide long-term control of internal P loading
from the sediment (Boers et al., 1992, 1994; Cooke et al., 1993a; Smolders et al., 2006; Kleeberg
et al., 2013).
Various mesocosm and field experiments have shown that the addition of Fe to the sediment
indeed results in lower total phosphorus (TP) concentration in the water column, which is why
iron is often used to decrease P concentrations of lake inlet water before the water enters the lake
(Klapwijk et al., 1982; Boers et al., 1992; Van Donk et al., 1994; Smolders et al., 1995; Kleeberg
et al., 2013). High Fe concentrations in the sediment, however, can have deleterious effects on
macrophytes (Kamal et al., 2004). Recent experiments have shown that growth of plants can
be directly inhibited by high iron concentrations in the sediment for instance by the formation
of necrotic leaf spots and iron plaques on roots (Lucassen et al., 2000; Van der Welle et al.,
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
42
2007a). Evidently, iron can also have indirect effects on macrophytes by lowering the phosphorus
concentration in the sediment, thereby decreasing the available nutrients for growth and by
lowering the pH of the water (Boers et al., 1994). Moreover, the addition of iron to the sediment
may be possible in mesocosms, but is a real challenge for a whole lake. Alternatively, iron could be
added to the surface water. However, the effects of adding iron to the surface water on submerged
macrophytes are not yet known, whereas they are directly exposed to the added iron when this
is added in the surface water in contrast to addition in the sediment. The place of addition may
affect macrophyte species differently, as macrophytes differ in rooting strategies (Jeppesen et al.,
1998). Macrophytes depending for their growth on nutrients from the water column might be
more affected by iron in the water column than rooting macrophytes, which generally take up
nutrients from the sediment. Over time, rooting species may become affected as well, when the
iron added in the water column precipitates and mixes with the sediment through macrofaunal
activity or wind-driven sediment movement (Søndergaard et al., 2003).
The objective of this study was to test whether iron addition as a restoration measure affects
the growth of submerged macrophytes. We hypothesized that this depends on the iron dosage,
the application mode (surface water or sediment plus water), and the rooting strategy of the
macrophytes (uptake of nutrients from sediment or water column). We experimentally tested
potential negative effects of iron (Fe) on the growth of two submerged macrophytes, the facultative
rooting species Elodea nuttallii (Planch.) H. St. John and the rooting species Potamogeton pectinatus L.
as well as on the sprouting of propagules present in the sediment propagule bank. Furthermore, to
simulate a condition in which wind driven sediment resuspension and subsequent sedimentation
would lead to an accumulation of iron in the sediment, we added a treatment in which we, prior
to the start of the experiment, mixed half of the total dosage of iron in the sediment. To study
the effect of iron addition we focused on changes in macrophyte growth and appearance, biomass
allocation, nutrient composition, and sprouting of propagules from the sediment. The experiment
is based upon planned restoration measures in peat Lake Terra Nova, The Netherlands (Van de
Haterd and Ter Heerdt, 2007), where water managers are proposing to add iron to the surface
water.
MATERIAL AND METHODS
Study species and study site
The study species Elodea nuttallii and Potamogeton species are often observed to be the first dominant
species after lake restoration measures have been taken (Van Donk et al., 1994; Van Donk and
Otte, 1996; Perrow et al., 1997; Irfanullah and Moss, 2004; Hilt et al., 2006). This was also
the case in lake Terra Nova, where Elodea nuttallii became dominant and multiple Potamogeton
species occurred (including P. pectinatus) after sediment disturbing fish were removed through
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
43
3
biomanipulation in a restoration attempt in the past (Van de Haterd and Ter Heerdt, 2007;
Bakker et al., 2013). Therefore, these are the first species that are expected to return when iron
supplementation is a successful restoration measure and thus the first to be exposed to potential
negative effects of iron addition. Lake Terra Nova (52º 12’ 55.87” N, 5º 2’ 23.00” E) is an 85
ha shallow peat lake with a mean depth of 1.4 m. The bottom is covered with a 0.9 m organic
sediment layer, with an organic matter concentration of 62.8% and a moisture content of 95.2%.
According to Brouwer and Smolders (2006), the internal loading of P from the sediment is
estimated at 0.10 g m-2 year-1. The restoration measure of iron supplementation was proposed to
bind this extra influx of P into the lake surface water.
Experimental set-up
In February 2010, 90 polyethylene tanks (w × l × h = 0.19 × 0.19 × 0.29 m3) were set up at the
NIOO-KNAW in Nieuwersluis. The tanks were placed in a temperature and light controlled
culture room with a constant temperature of 18 °C and a light intensity of 100 ± 5 µmol photons
m-2 s-1 at the water surface in the tanks and a 14:10 h light:dark cycle. Each tank contained 2 L
peat sediment, collected from Lake Terra Nova. Before tanks were filled combinations of three
different variables, (i) levels of iron addition (0, 20, and 40 g Fe m-2), (ii) mode of iron application
(in the water or in the sediment plus water), and (iii) two different macrophyte species plus
control (E. nuttallii, P. pectinatus, no macrophyte), were randomly allocated to the tanks in a full
factorial design, each with 5 replicates.
The total iron concentration that is planned to be dosed in Lake Terra Nova is 100 g Fe m-2
over a period of 1.5 years. This dosage was also used in various experimental and field studies
(Boers et al., 1994; Burley et al., 2001). To mimic the effects of (gradual) iron addition in Terra
Nova, we recalculated this dose according to the volume of our experimental units (containing
only 7.3 L water) and the short duration of our iron addition, which resulted in a total addition
of 100 mg Fe per tank and this corresponds to an iron treatment of 20 g Fe m-2 in Terra Nova.
A high iron addition was calculated which would receive a total addition of 200 mg Fe, which
corresponds to an iron treatment of 40 g Fe m-2 in Terra Nova. The iron added to both treatments
was in the form of FeCl3. Two control treatments were designed; a zero Fe addition treatment
and a treatment without macrophytes, as a control for the effects of macrophytes on pore and
surface water composition. The no iron control treatment received NaCl in equal molar amounts
of chloride as in the high iron treatments. The goal of the experiment was to test whether iron
addition would affect macrophyte growth. We hypothesized that a no iron control treatment
without a dose of iron would show differences in growth compared to macrophytes with iron
dosing, as iron naturally binds to phosphate in the water, resulting in different growth rates due
to differences in phosphate availability. Therefore we added once a low dose of 0.73 mg FeCl3 on
day 1 to the surface water of the no iron control treatments to bind the available 0.1 µmol L-1 P in
the water column and sediment in order to equalize these differences.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
44
The sediment of tanks in which iron was added to both the water column and sediment (i.e.
mixed treatments) was pre-mixed with half of the total dosage of FeCl3 and NaCl. Subsequently,
7.3 L of filtered (ME 24, Whatman, Brentford, UK) Terra Nova water (Fe = 0.06 ± 0.02 µmol
L-1, PO4 = 0.01 ± 0.01 µmol L-1, NO
3 = 0.15 ± 0.05 µmol L-1) was poured very carefully onto the
sediment. To enable pore water sampling, Rhizon soil moisture samplers (Eijkelkamp Agrisearch
Equipment, Giesbeek, The Netherlands) attached to 50 mL vacuum syringes were inserted into
the upper layer of the sediment. Elodea nuttallii shoots and P. pectinatus tubers were collected at
ponds close to Lake Terra Nova. Potamogeton pectinatus tubers and E. nuttallii shoots were pregrown
for 2 weeks under the experimental conditions to let both macrophytes get acclimatized and to
let the P. pectinatus tubers sprout. Three E. nuttallii shoots (mean total dryweight per tank 0.07
± 0.01 g; n=30), and three P. pectinatus shoots (mean total dryweight per tank 0.07 ± 0.01 g;
n=30) were each planted in the sediment of 60 tanks, and 30 tanks were kept empty (macrophyte
control treatment). Elodea nuttallii was planted as shoots of about 8 cm, without any belowground
material; P. pectinatus shoots were about 6 cm long and still contained tubers. Water loss due
to evaporation and sampling was replaced with filtrated (ME 24, Whatman, Brentford, UK)
Terra Nova water. During the experiments, plants were checked for several visual observable
characteristics of iron toxicity on plants, such as the formation of black spots or discolorations
of leaves. Moreover, macrophytes that sprouted from the sediment propagule bank during the
experiment were counted, removed, and identified to the species level.
Iron was not added to the surface water all at once, but slowly during 36 addition days,
namely 3 times a week over a period of 12 weeks. The additions correspond for the low and high
iron addition treatments to 2.9 and 5.7 mg Fe per tank per addition day respectively. The mixed
treatments, in which half of the total FeCl3 and NaCl dose was already mixed in the sediment,
received only half of the aforementioned dose in the surface water per tank per addition day. The
slow addition of iron over 12 weeks enabled addition of high dosages of iron to the surface water
as a concentrated addition of iron would result in a quick drop in pH.
Sampling and sample analysis
Every other week, at days 1, 13, 27, 41, 55, 69, and 83 of the experiment, 105 mL of surface
water and sediment pore water samples were taken from each tank for chemical analyses. Directly
after the pore water had been collected, 50 mL was fixed in polyethylene bottles with 1 mL
nitric acid (2 M) for Fe, Al, Ca, and SO4 analysis. Another 20 mL of pore water was stored in
polyethylene bottles for Cl- analysis. Surface water samples of 20 mL were filtered over a 0.45 μm
membrane filter (ME 25, Whatman, Brentford, UK) before storage in polyethylene bottles and
fixation with nitric acid. Membrane filters that were used for the filtration of 20 mL surface water
were dried for 24 hours at 60 °C and afterwards stored in 50 mL centrifuge tubes before analysis
of precipitated Fe. Subsamples of 10 mL were taken from both surface and pore water and filtrated
over (1.2 µm) Whatman GF/C filters. All samples were stored at -20 °C before analyses.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
45
3
A 25 mL subsample from both surface and pore water was used to measure conductivity with
a portable parameter instrument (WTW Multi 350i, Weilheim, Germany) and pH and alkalinity
with a TIM840 titration manager (Radiometer Analytical, Copenhagen, Denmark). Alkalinity
was determined by titrating with 0.01 M HCl down to pH 4.2. The 10 mL subsamples were
used to colourimetrically determine PO4, NH
4, NO
3, and NO
2 with a QuAAtro CFA flow
analyser (Seal Analytical, Beun de Ronde, Abcoude, The Netherlands). Dissolved Fe, Al, Ca,
and S were measured using an inductively coupled plasma emission spectrophotometer (ICP;
Liberty 2, Varian, Bergen op Zoom, The Netherlands) according to Dutch NEN-EN-ISO
17294. Total S concentrations provided a good estimate of SO4 concentrations, because only a
small percentage of S was in organic form, as was verified by capillary ion analysis (Geurts et
al., 2008). Precipitated Fe on the collected 0.45 μm membrane filters was also measured using
ICP, however filters were treated with 8 mL nitric acid (2 M) prior to analysis. Chloride was
measured spectrophotometrically (Aquakem 250, Thermo Fisher Scientific, Waltham, MA,
USA) with extinction at 480 nm. Acquired sediment pore water Fe and PO4 concentrations were
subsequently used to calculate molar Fe:PO4 ratios, which, when the ratio reaches values below
10 mol mol-1, indicate P release from the sediment (Geurts et al., 2008). Pore water Fe:PO4 ratios
from the last week of the experiment (week 12) could not be calculated with pore water Fe and
PO4 concentrations, as PO
4 concentrations by then decreased below the detection limit, therefore
we used data from week 10.
At the end of the experiment, all macrophytes were harvested and separated in shoots and
roots. A small branch of about 3 cm was separated in each tank from the macrophyte shoot
material for epiphytic macroalgae determination. Each branch was placed in a closed cup with
20 ml of deionised water and subsequently shaken for exactly 60 seconds. Epiphytic material was
determined according to the method described in Zimba and Hopson (1997) and epiphyton per
unit surface area (SLA) was calculated using SLA values from literature of 1309 cm2 g-1 dryweight
for E. nuttallii (James et al., 2006) and 900 cm2 g-1 dryweight for P. pectinatus (Pilon and Santamaria.,
2002). Afterwards, the branches used for determination and all other macrophyte material were
dried for 24 hours at 60 °C and subsequently weighed to determine the total dryweight. Total
dryweight at the start of the experiment was calculated with a conversion factor, which was
acquired from the fresh and dryweight of several subsamples. For E. nuttallii dryweight = 8.7 %
of fresh weight (fresh weight at start of the experiment 0.77 ± 0.16 g), for P. pectinatus dryweight
= 15.9 % of fresh weight (fresh weight at start of the experiment 0.44 ± 0.08 g). Relative growth
rate (RGR) for both macrophyte species was calculated using dryweight according to Barrat-
Segretain (2004): (ln(w2)-ln(w
1))/d, in which w
1 and w
2 are the dryweight at the start and at the
end of the experiment, respectively and d is the duration of the experiment in days (84 days).
To determine both C and N concentrations in macrophytes, macrophytes were grounded and a
homogenized portion of dry macrophyte material was analysed with a FLASH 2000 Organic
Elemental Analyser (Interscience, Breda, The Netherlands). Macrophyte P concentrations were
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
46
determined by incinerating homogenized dry material for 30 minutes at 500 °C, followed by
digestion in H2O
2 (Murphy and Riley, 1962) before analysis with a QuAAtro CFA flow analyser.
Collected nutrient concentrations in macrophytes were subsequently used to calculate N:P, C:N,
and P:N ratios to get an insight into the nutrient availability at the end of the experiment.
Statistical analysis
Statistical analyses were carried out with SPSS 18.0 (SPSS, Chicago, IL, USA). Differences between
treatments for plant biomass and plant nutrient composition were tested for each macrophyte
species with two-way ANOVA’s with iron treatment and application mode (water vs. sediment
plus water) as fixed factors followed by a Tukey’s post-hoc test. Differences in chemical variables
and number of spouting species were tested with three-way ANOVA’s with iron treatment,
application mode and macrophyte species (consisting of the levels Elodea, Potamogeton or control)
as fixed factors followed by a Tukey’s post-hoc test. Prior to analysis, all data were tested for
normality and homogeneity of variance, and if necessary, data were log 10 transformed. For data
that had no normal distribution, even after transformation, a nonparametric Kruskal-Wallis test
was used with Statistica 9.1 (StatSoft Inc., Tulsa, OK, USA) to analyse variances. Values are
presented as means (± SEM) and P ≤ 0.05 was accepted for statistical significance.
RESULTS
Macrophyte response
Adding iron to the water column or to both the water column and the sediment did not
differentially affect macrophyte growth (Table 3.1), therefore we pooled these data for the analysis
of macrophyte growth. Total macrophyte biomass (roots plus shoots) increased over time in all
treatments, but iron addition induced a different response in the two macrophyte species (Table
3.1, Figure 3.1). Elodea nuttallii biomass did not differ significantly between the three iron
treatments with an average RGR of 0.034 ± 0.001 g dryweight day-1 (Figure 3.1a). In contrast,
iron concentrations had a significant negative effect on the growth of Potamogeton pectinatus, which
grew less with increasing iron addition (Figure 3.1b, Table 3.1) resulting in a mean RGR of 0.041
± 0.001 g dryweight day-1 for the no iron treatment, a mean RGR of 0.039 ± 0.001 g dryweight
day-1 for the 20 g Fe m-2 treatment, and a mean RGR of 0.036 ± 0.001 g dryweight day-1 for the
40 g Fe m-2 treatment (Figure 3.2a). Biomass allocation was not affected by either iron addition
or application mode (Figure 3.2b, Table 3.1). Epiphyton measurements from macrophyte shoots
did not show any significant differences between any of the treatments (Figure 3.2c). No changes
or plant abnormalities were detected on all macrophytes during observations on the direct toxic
effects of iron on macrophytes (i.e. necrotic leaf spots).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
47
3
Table 3.1 – Results of analysis of the effects of iron addition on biomass, growth, shoot-root ratio, and nutrient composition of E. nuttallii and P. pectinatus. Data were analysed with a two-way ANOVA with the amount of iron (0, 20, or 40 g m-2) and the application mode (in water or sediment plus water) as fixed factors, n=5. Bold values indicate P ≤ 0.05.
Effect
Iron amount Application mode Iron × Mode
Df = 2, 24 Df = 1, 24 Df = 2, 24
F P F P F P
E. nuttallii
Biomass roots 0.81 0.46 0.21 0.65 4.55 0.02
Biomass shoots 2.11 0.14 0.01 0.91 2.05 0.15
Total biomass 1.83 0.18 0.04 0.84 2.74 0.08
Total biomass increase 1.79 0.19 0.06 0.81 2.82 0.08
Shoot-root ratio 0.34 0.72 0.18 0.67 3.72 0.04
RGR 1.31 0.29 0.07 0.78 3.99 0.03
N concentration shoots 1.50 0.24 0.01 0.95 1.69 0.21
N concentration roots 3.62 0.04 0.03 0.87 1.12 0.34
P concentration shoots 0.95 0.40 0.06 0.80 0.54 0.59
P concentration roots 0.17 0.84 0.28 0.60 0.01 0.99
N:P ratio shoots 0.09 0.91 0.00 0.98 0.85 0.44
N:P ratio roots 2.32 0.12 0.17 0.69 0.06 0.95
Epiphyton 3.26 0.06 0.81 0.38 1.50 0.24
P. pectinatus
Biomass roots 4.55 0.02 0.24 0.63 0.63 0.54
Biomass shoots 3.60 0.04 0.16 0.70 0.66 0.53
Total biomass 4.74 0.02 0.01 0.92 0.65 0.53
Total biomass increase 4.91 0.02 0.01 0.93 0.71 0.50
Shoot-root ratio 1.00 0.38 1.00 0.33 0.95 0.40
RGR 5.87 0.01 0.07 0.79 2.18 0.13
N concentration shoots 0.80 0.46 0.07 0.80 1.65 0.21
N concentration roots 0.35 0.71 2.34 0.14 0.68 0.52
P concentration shoots 3.21 0.06 0.88 0.36 0.53 0.60
P concentration roots 1.12 0.34 0.91 0.35 1.10 0.35
N:P ratio shoots 1.79 0.19 0.14 0.71 1.11 0.35
N:P ratio roots 0.40 0.68 0.21 0.66 0.35 0.71
Epiphyton 1.60 0.22 0.78 0.39 0.05 0.95
During the experiment several macrophyte species sprouted from the sediment, including
Nitella mucronata, Chara virgata, C. globularis, and Nuphar lutea. There were no significant effects
of iron on differences in abundance, however seedlings sprouted more often in no macrophyte
control tanks compared to tanks with E. nuttallii and P. pectinatus (ANOVA: F2, 72
= 5.45, P <
0.01, data not shown).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
48
Figure 3.1 – Root, shoot, and total biomass in g dryweight at the end of the experiment in response to iron addition for (a) E. nuttallii and (b) P. pectinatus. Open circles, closed triangles, and open squares represent respectively root, shoot, and total biomass. Significant differences between iron treatments are indicated by different letters (Analysis of variance, Tukey test, P ≤ 0.05).
Tissue nutrient concentrations
The final P concentrations of both E. nuttallii and P. pectinatus shoots (0.96 ± 0.05 and 1.35 ± 0.08
mg g dryweight-1) decreased steeply compared to starting concentrations (6.29 ± 0.32 and 6.17 ±
0.57 mg g dryweight-1) respectively. N concentrations showed this decrease as well with low mean
final concentrations (7.51 ± 0.26 and 8.31 ± 0.24 mg g dryweight-1) respectively in macrophyte
shoots compared to starting concentrations (45.79 ± 0.58 and 34.99 ± 1.87 mg g dryweight-1) for
E. nuttallii and P. pectinatus, respectively. No statistical differences in macrophyte tissue nutrient
concentrations were found among iron treatments (Figure 3.3a, b; Table 3.1).
Mean shoot N:P ratios increased from 16.10 ± 0.63 and 12.54 ± 1.88 mol mol-1 at the start of
the experiment to 17.38 ± 1.42 and 17.66 ± 1.56 mol mol-1 (or 8.24 ± 0.33 and 6.61 ± 0.31 g g-1)
at the end of the experiment for E. nuttallii and P. pectinatus respectively, indicating a relative higher
decrease in mean shoot P concentrations over time compared to shoot N concentrations for both
macrophyte species, but differences were not significant among treatments (Figure 3.3c, Table 3.1).
Tissue nutrient concentrations in macrophyte roots showed a similar reaction to the different iron
treatments as nutrient concentrations in macrophyte shoots (Table 3.1).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
49
3
Figure 3.2 – (a) Relative growth rate (RGR) in g dryweight day-1, (b) macrophyte shoot:root ratio in g g-1, and (c) epiphyton per unit plant area in mg cm-2 (average ± SEM) in response to the different iron additions after 12 weeks. Open triangles and closed circles represent respectively E. nuttallii and P. pectinatus. Significant differences between treatments are indicated for each species separately by different letters (Analysis of variance, Tukey test, P ≤ 0.05).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
50
Surface and pore water nutrient concentrations
During the experiment, Fe concentrations in both surface and pore water of the high iron treatments
increased significantly (Supplementary Table 3.1). Precipitated Fe reached in the high iron
treatment the highest mean concentration (Supplementary Table 3.1). In addition, precipitated
Fe was significantly higher in the treatments in which iron was only added to the surface water
compared to mix treatments in which iron was partly added to the sediment (Supplementary Table
3.1). PO4 concentrations in both pore and surface water of all treatments decreased to values below
the detection limit (< 0.03 μmol L-1; Supplementary Table 3.1). As a result of these low PO4 values,
pore water Fe:PO4 after 10 weeks ratios reached high mean values (Supplementary Table 3.1), which
did not differ between iron and no iron control treatments (Table 3.2). Tanks with macrophytes had
significant lower pore water PO4 concentrations and consequently higher pore water Fe:PO
4 ratios
compared to pore water Fe:PO4 ratios in control tanks (Table 3.2).
Surface water pH and alkalinity decreased significantly during the experiment due to iron
additions. At the end of the experiment the pH was significantly lower in the high iron treatments
(Table 3.2). Alkalinity only differed significantly between no iron control and iron treatments, with
a higher alkalinity in the no iron treatments compared to lower values in the low and high iron
treatments (Table 3.2, Supplementary Table 3.1). Conductivity significantly increased over time
for the iron treatments and at the end of the experiment values in the high iron (40 g Fe m-2) and
no iron control treatments were significantly higher than in the low iron treatments (20 g Fe m-2;
Table 3.2, Supplementary Table 3.1). The presence of macrophytes resulted in a lower conductivity
compared to tanks without macrophytes (Table 3.2, Supplementary Table 3.1).
DISCUSSION
Macrophyte growth under iron addition
Iron addition in the sediment of various lakes in The Netherlands resulted in an improvement
of the water quality by a decrease of the surface water PO4, chlorophyll-a, and suspended solids
concentrations (Boers et al., 1992; Smolders et al., 1995; Van der Welle et al., 2007a). By
improving water transparency, iron addition can stimulate macrophyte growth, which is often light
limited (Bornette and Puijalon, 2011). Epiphyton measurements from macrophyte shoots in our
experimental units did not show any statistical differences between the iron and no iron treatments,
thus excluding differences in epiphyton- induced light limitation among iron treatments.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
51
3
Figure 3.3 – Macrophyte shoot (a) N concentration and (b) P concentration in mg g dryweight-1 and (c) N:P ratio in g g-1 (average ± SEM) in response to the different iron additions after 12 weeks. Open triangles and closed circles represent respectively E. nuttallii and P. pectinatus. There were no significant differences between iron treatments.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
52
Tab
le 3
.2 –
Res
ults
of a
naly
sis o
f the
eff
ects
of i
ron
addi
tion
on
surf
ace
and
pore
wat
er n
utri
ent c
ompo
siti
on. D
ata
wer
e an
alys
ed w
ith
a th
ree-
way
AN
OV
A (F
) or
non
-par
amet
ric
Kru
skal
-Wal
lis (
H) w
ith
the
amou
nt o
f iro
n (0
, 20,
or 4
0 g
m-2),
the
appl
icat
ion
mod
e (i
n w
ater
or s
edim
ent p
lus w
ater
) and
the
mac
roph
yte
spec
ies
(E. n
utta
llii
, P. p
ecti
natu
s, or
con
trol
tre
atm
ents
) as
fixed
fact
ors,
n=
5. B
old
valu
es in
dica
te P
≤ 0
.05.
Eff
ect
Iron
M
ode
Pla
nt
Iron
× M
ode
Iron
× P
lant
Mod
e ×
Pla
ntIr
on ×
Mod
e ×
Pla
nt
Df =
2,7
2D
f = 1
,72
Df =
2,7
2D
f = 2
,72
Df =
4,7
2D
f = 2
,72
Df =
4,7
2
F/H
PF/
HP
F/H
PF/
HP
F/H
PF/
HP
F/H
P
Surf
ace w
ater
Fea
28.4
6<
0.00
15.
070.
022.
300.
3240
.62
<0.
001
35.7
0<
0.00
17.
650.
1857
.57
<0.
001
Fe17
.04
<0.
001
10.1
7<
0.00
10.
340.
710.
410.
671.
360.
260.
240.
791.
600.
19
(pre
cipi
tate
d)
PO
4a1.
010.
602.
020.
161.
010.
604.
050.
547.
080.
534.
050.
5416
.18
0.51
Cl
185.
18<
0.00
122
.50
<0.
001
0.39
0.68
54.2
6<
0.00
10.
430.
790.
810.
450.
810.
52
Al
68.6
3<
0.00
11.
870.
184.
020.
023.
440.
041.
010.
410.
710.
500.
940.
45
Ca
23.5
5<
0.00
11.
300.
2614
.71
<0.
001
2.01
0.14
0.95
0.44
0.01
0.99
0.31
0.87
SO4
10.5
0<
0.00
10.
140.
713.
260.
043.
330.
040.
910.
460.
510.
600.
640.
63
NH
4a 1.
160.
560.
180.
671.
310.
524.
570.
473.
030.
932.
260.
817.
700.
97
NO
2 0.
560.
577.
700.
011.
470.
242.
190.
120.
850.
500.
430.
651.
980.
11
NO
3a0.
700.
700.
240.
620.
870.
391.
320.
935.
050.
754.
720.
459.
240.
93
pH66
.30
<0.
001
3.53
0.06
22.0
6<
0.00
11.
380.
261.
350.
260.
710.
500.
390.
81
Alk
alin
ity
16.5
9<
0.00
110
.45
<0.
001
32.2
0<
0.00
11.
380.
261.
100.
371.
020.
370.
480.
75
Con
duct
ivit
y12
4.13
<0.
001
30.8
4<
0.00
126
.06
<0.
001
9.07
<0.
001
1.20
0.32
1.07
0.35
0.79
0.54
Por
e wat
er
Fe1.
420.
250.
360.
555.
90<
0.00
10.
460.
530.
240.
910.
570.
570.
680.
61
PO
4a0.
190.
910.
000.
957.
290.
021.
970.
8511
.87
0.12
11.4
50.
0420
.60
0.24
Cl
147.
38<
0.00
112
.44
<0.
001
0.17
0.84
35.0
7<
0.00
10.
210.
930.
930.
400.
740.
57
Ala
36.1
9<
0.00
10.
000.
975.
110.
0837
.82
<0.
001
42.2
5<
0.00
15.
320.
3845
.24
<0.
001
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
53
3
Ca
21.5
5<
0.00
11.
240.
273.
980.
022.
310.
110.
420.
790.
050.
950.
700.
59
SO4
4.48
0.02
0.00
1.00
4.25
0.02
1.66
0.20
0.44
0.78
0.95
0.39
0.47
0.76
NH
4a 2.
980.
230.
100.
750.
050.
983.
940.
5613
.79
0.09
0.84
0.97
16.6
40.
48
NO
2 0.
030.
970.
000.
980.
380.
691.
380.
260.
710.
591.
440.
250.
660.
62
NO
3a 5.
260.
071.
190.
270.
800.
677.
200.
219.
550.
308.
130.
1518
.99
0.33
pH13
.27
<0.
001
0.42
0.52
14.3
4<
0.00
12.
570.
081.
010.
410.
070.
931.
070.
38
Alk
alin
ity
5.29
0.01
1.44
0.23
5.59
0.01
1.56
0.22
0.97
0.43
0.16
0.85
0.66
0.62
a Non
-par
amet
ric
Kru
skal
-Wal
lis
test
(H) p
erfo
rmed
inst
ead
of A
NO
VA
(F)
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
54
The addition of iron can furthermore have several direct and indirect effects on macrophyte
growth (Wheeler et al., 1985; Snowden and Wheeler, 1995; Lucassen et al., 2000). Whereas
iron addition can alleviate light limitation, it can simultaneously induce nutrient limitation
for macrophyte growth by limitation of P availability due to the precipitation of PO4 with Fe
(Wheeler et al., 1985). In our experiment the P concentrations measured in the surface and pore
water dropped below the detection limit at the end of the experiment, suggesting potential
P limitation for macrophyte growth. Macrophyte P concentrations followed this change as
concentrations decreased during the experiment as well. Nevertheless, no statistical difference
in P availability between iron and no iron control treatments was found in the surface or pore
water, suggesting that all treatments may have been P-limited. The low P availability in the
no iron control treatment could be the result of our initial action of adding a small dosage of
iron to the no iron control treatment to level out P limitation, but the Terra Nova sediment
itself also contained iron (83.60 ± 11.38 µmol Fe L-1; Van der Wal et al., 2013), which may
have caused these low P concentrations. Nutrient limitation for plant growth can be deduced
from the nutrient concentrations measured in plant tissue. According to Krombholz and Gerloff
(1966) both species were limited by P, but also by N. Macrophyte N:P ratios may also be used
as an indicator to determine which of the two elements is most likely to be limiting. However,
different threshold values for N- and P-limitation have been suggested for phytoplankton and
terrestrial plants, respectively P limitation at N:P > 16 mol mol-1 or 7 g g-1 (Redfield ratio;
Redfield, 1958) and N:P > 16 g g-1 (Koerselman and Meuleman, 1996), whereas Duarte (1992)
suggests a threshold of N:P > 12 g g-1 for aquatic primary producers in general. These ratios
would indicate that E. nuttallii and P. pectinatus were most likely N limited across all treatments.
The surface and pore water nutrient data from our experiments indeed show low N values, which
is also in accordance to the low N concentrations found in Lake Terra Nova itself (see Van der
Wal et al., 2013).
Nutrient limitation may have affected the RGR of both macrophytes, which were slightly
lower than relative growth rates of these species growing under optimal conditions (Pilon et al.,
2002, Barrat-Segretain, 2004). Nevertheless, microcosm experiments with P. pectinatus report
growth rates of 0.015 (Van Dijk and Van Vierssen, 1994), 0.028 (Spencer and Rejmánek, 2010),
and 0.039 g dryweight day-1 (Pilon et al., 2002), which are in the same range as the RGR
of this species in our experiment. Relative growth rates of E. nuttallii can be higher than the
RGR measured in our experiment (0.066 g dryweight day-1; Barrat-Segretain, 2004), but the
high biomass of epiphyton relative to the aboveground biomass of E. nuttallii could have been
involved in reducing the RGR of E. nuttallii. High biomass of epiphyton on E. nuttallii shoots
was also reported by Irfannulah and Moss (2004), who monitored E. nuttallii growth in mesocosm
experiments over a period of 40 days, which resulted in a markedly lower RGR of 0.03-0.04 g
dryweight day-1.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
55
3
As there was no significant effect of iron treatment on availability of PO4 in surface water or
pore water, nor in macrophyte P concentration, differences in growth response between the two
macrophytes seemed not dependent on P availability and are therefore likely to be the outcome
of direct effects of iron (under P limitation). Direct effects of iron can be seen in the plants’
physical structure. It can act on the leaves by reducing their size or by the formation of black
necrotic spots or complete discoloration of leaves. Iron also acts on roots which can blacken, stop
growing, or lack branching (Wheeler et al., 1985; Snowden and Wheeler, 1995; Van der Welle
et al., 2007a). These physical symptoms, indicating direct iron toxicity could not be detected in
our experiment with E. nuttallii and P. pectinatus. Moreover, effects of iron addition, regardless of
application mode, were not detected on biomass allocation between roots and shoots. This could
imply that high Fe concentrations around macrophyte roots did not induce root die-off, which
would be expressed in higher shoot:root ratios. Until now these direct effects of iron toxicity
were only observed in experiments with both terrestrial and emergent wetland species (Jones and
Etherington, 1970; Wheeler et al., 1985; Lucassen et al., 2000; Van der Welle et al., 2007a),
but not in experiments using fully aquatic plants or charophytes (Van der Welle et al., 2007b;
Immers et al., 2013). Even though the direct effects of toxicity are not shown, it could be that the
(energetic) costs of iron tolerance in P. pectinatus are merely expressed by a decrease in biomass,
as found for floating macrophytes and non-aquatic plants (Snowden and Wheeler, 1995; Van der
Welle et al., 2007a). The fact that out of the two macrophytes tested, only P. pectinatus showed
this response to iron addition may be explained by the obligate sediment rooting of P. pectinatus,
whereas E. nuttallii relies less on rooting in the sediment and has many roots in the water layer.
This implies that through its flexible rooting strategy E. nuttallii enables itself better access to
nutrients, both in the water and in the sediment. Possibly, the ability to alter rooting strategies
might also allow macrophytes to be more resistant to changes in their environment, but this
should be further investigated, as we only tested one species per rooting strategy and any of the
other differences between these species may cause this different response to iron addition.
Phosphate inactivation through iron addition
The goal of adding Fe to Lake Terra Nova was to lower surface water P and to control internal
P release. According to Geurts et al. (2008), sediments with pore water Fe:PO4 ratios < 10
(mol mol-1) would indicate an enhanced potential for P release from the sediment. The required
Fe:PO4 ratios > 10 (mol mol-1) to prevent P release were reached in all tanks at the end of
our experiment and consequently surface water P concentrations remained low. Nevertheless,
the P reduction in the no iron control tanks might be only temporary as Fe can be depleted
quickly through interactions with SO4 (Smolders et al., 2006; Van der Welle et al., 2007b). In
contrast, high iron concentrations in the iron treatments, which were detected in the form of
iron-phosphates and iron-oxides, will provide long term control of P release from the sediment
(Boers et al., 1994). The higher Fe:PO4 ratios in tanks with macrophytes compared to control
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
56
tanks was due to the fact that macrophytes take up PO4 via their roots, resulting in lower PO
4
concentrations in pore water. This could implicate that reduction of internal P loading is most
effective when macrophytes are already present. Alternatively, the presence of macrophytes can
function as a nutrient pump, where macrophytes take up P from the sediment and release it in
the water column through their leaves (Carpenter, 1981). However, in our experiment there was
no significant effect of macrophyte presence on PO4 concentrations in the surface water, most
probably as the release of P would only take place during decay after death of the macrophyte for
which the duration of the experiment was too short.
Iron addition and lake restoration
According to Cooke et al. (1993a), lakes with high internal loading are only able to improve if
P is inactivated by addition of chemical binding agents. We conclude from our experiments that
adding up to 40 g Fe m-2 in the surface water can, depending on the species, negatively affect
macrophyte growth, but is not lethal for macrophytes and their propagules in the sediment
bank. Furthermore, by increasing light availability through inducing nutrient limitation for
phytoplankton, iron addition can have net positive effects on macrophyte growth. The different
response of both macrophyte species to iron addition, however, indicates that iron addition can
result in a shift in species composition (Kamal et al., 2004; Van der Welle et al., 2007b). According
to Geurts et al. (2008), the occurrence of endangered species in peat lakes is correlated with high
Fe:PO4 ratios in the sediment and iron addition may thus benefit these species. Moreover, the
amount of species that sprouted from the sediment was equal for all treatments, which means that
adding iron did not seem to hinder this process. The species that sprouted from the propagule
bank in the Lake Terra Nova sediment (Nitella mucronata, Chara virgata, and Chara globularis) are
also species of high conservation value which are typically found in meso- to oligotrophic water
bodies (Simons and Nat, 1996).
The addition of iron also resulted in a decrease in pH in the tanks receiving high iron
additions, but values stayed well above 7 and the pH varied only ± 0.5 between treatments. The
slow addition of iron over 12 weeks thus enabled addition of high cumulative dosages of iron to
the surface waters, whereas previous studies, where iron was added at once (in the sediment), were
restricted to lower dosages (Jaeger, 1994; Van der Welle et al., 2007a).
The combined addition of iron to the surface water and sediment showed the ‘long-term’
effect of iron addition, when repeated cycles of wind driven sediment resuspension and subsequent
sedimentation lead to burial of iron in the sediment. The only difference that was found between
this ‘long-term’ treatment and the treatment in which iron was added only to surface water was
the high concentration of precipitated iron in the latter treatment. The precipitated layer of
iron (iron-oxides and iron-hydroxides) can potentially be a nuisance for macrophytes or other
organisms as the layer can block incoming light or form a physical barrier for macrophyte
emergence, although a difference in macrophyte biomass between addition in surface water or
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
57
3
both sediment and surface water was not found in our study. Formation of iron hydroxides (ochre)
will probably not play a big part in decreasing water transparency as well, as ochre formation
usually only occurs in waters with low pH and low concentrations of dissolved oxygen (Macintosh
and Griffiths, 2013). Moreover, ochre has a much higher density compared to other sediment
particles and will most likely be deposited on the top layer of the sediment.
We conclude that adding iron(III)chloride in the dosages used in our experiment (20 – 40
g Fe m-2 to the surface water) does not prevent macrophyte recovery but may affect macrophyte
community composition due to differential responses of macrophyte species. However, despite
these positive indications, the application of iron addition in lake restoration is still in an
experimental phase as long term effects on the biota are currently unknown.
ACKNOWLEDGEMENTS
We are grateful to Leon Lamers for his valuable theoretical insights and useful discussions. We
would also like to thank Naomi Huig, Thijs de Boer, and Koos Swart for their practical assistance
in the field and Hans Kaper, Nico Helmsing, and Harry Korthals for performing multiple
chemical analyses. This study was funded by the Water Framework Directive Innovation Fund
from Agentschap NL from the Dutch Ministry of Economic Affairs, Agriculture and Innovation.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 3
58
Sup
ple
men
tary
Tab
le 3
.1 –
Mea
n (±
sta
ndar
d er
ror
of m
ean)
nut
rien
t co
ncen
trat
ions
at
the
end
of t
he e
xper
imen
t m
easu
red
in b
oth
surf
ace
and
pore
wat
er fo
r ea
ch t
reat
men
t (t
he a
mou
nt o
f iro
n, t
he a
ppli
cati
on m
ode,
and
the
mac
roph
yte
spec
ies)
.M
acro
phyt
e sp
ecie
sE
lode
a nu
ttal
lii
Pot
amog
eton
pec
tina
tus
Pla
ce o
f add
itio
nN
on-
Mix
edM
ixed
Non
- M
ixed
Iron
add
itio
n0
g Fe
m-2
20 g
Fe
m-2
40 g
Fe
m-2
0 g
Fe m
-220
g F
e m
-240
g F
e m
-20
g Fe
m-2
20 g
Fe
m-2
40 g
Fe
m-2
Surf
ace w
ater
Fe (µ
mol
L-1)
1.18
± 0
.08
1.42
± 0
.12
1.67
± 0
.14
1.35
± 0
.19
1.45
± 0
.07
1.76
± 0
.26
0.71
± 0
.08
0.92
± 0
.10
1.29
± 0
.18
Fe
(pre
cipi
tate
d; µ
mol
L-1)
1.33
± 0
.72
1.41
± 1
.22
4.43
± 3
.84
0.59
± 0
.47
0.62
± 0
.41
0.59
± 0
.31
4.06
± 0
.91
18.5
2 ±
11.
576.
64 ±
1.0
9
PO
4 (µ
mol
L-1)
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
Cl (
mm
ol L
-1)
3.17
± 0
.06
3.09
± 0
.03
3.27
± 0
.08
3.98
± 0
.06
4.10
± 0
.18
3.90
± 0
.16
2.62
± 0
.05
2.65
± 0
.05
2.76
± 0
.06
Al (
µmol
L-1)
1.47
± 0
.18
1.33
± 0
.12
1.65
± 0
.32
1.06
± 0
.22
1.53
± 0
.17
1.32
± 0
.16
0.54
± 0
.10
0.98
± 0
.20
0.66
± 0
.23
Ca
(mm
ol L
-1)
1.10
± 0
.05
1.27
± 0
.09
1.40
± 0
.14
1.08
± 0
.13
1.23
± 0
.08
1.27
± 0
.21
1.11
± 0
.05
1.22
± 0
.06
1.50
± 0
.08
SO4 (m
mol
L-1)
0.62
± 0
.04
0.66
± 0
.04
0.63
± 0
.04
0.61
± 0
.02
0.61
± 0
.06
0.53
± 0
.02
0.65
± 0
.06
0.65
± 0
.05
0.68
± 0
.02
NH
4 (µ
mol
L-1)
24.9
9 ±
1.0
424
.92
± 1
.76
24.6
3 ±
0.7
326
.68
± 1
.34
25.6
0 ±
1.4
626
.09
± 1
.53
25.9
8 ±
1.3
726
.10
± 0
.98
26.5
0 ±
1.2
8
NO
2 (µ
mol
L-1)
0.24
± 0
.01
0.24
± 0
.01
0.25
± 0
.01
0.23
± 0
.01
0.25
± 0
.01
0.25
± 0
.00
0.23
± 0
.01
0.23
± 0
.00
0.24
± 0
.01
NO
3 (µ
mol
L-1)
0.02
± 0
.01
0.12
± 0
.12
0.35
± 0
.30
0.04
± 0
.04
0.07
± 0
.03
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.17
± 0
.12
pH8.
51 ±
0.1
88.
92 ±
0.1
58.
42 ±
0.1
38.
48 ±
0.1
78.
95 ±
0.0
98.
29 ±
0.1
87.
60 ±
0.2
38.
32 ±
0.2
07.
49 ±
0.0
9
Alk
alin
ity
(mE
q L-1
)1.
18 ±
0.0
81.
42 ±
0.1
21.
67 ±
0.1
41.
35 ±
0.1
91.
45 ±
0.0
71.
76 ±
0.2
60.
71 ±
0.0
80.
92 ±
0.1
01.
29 ±
0.1
8
Con
duct
ivit
y (1
0 w
eeks
; µS
cm-1)
598
± 1
5.66
470.
6 ±
6.9
054
7 ±
11.
4871
0 ±
20.
7248
2.4
± 1
4.69
559.
2 ±
14.
9259
6 ±
11.
2648
4 ±
11.
0855
4.4
± 1
1.84
Por
e wat
er
Fe (µ
mol
L-1)
1.45
± 0
.19
1.45
± 0
.22
1.72
± 0
.16
1.43
± 0
.27
1.29
± 0
.15
1.64
± 0
.25
1.04
± 0
.16
1.2
± 0
.08
1.38
± 0
.12
PO
4 (µ
mol
L-1)
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.01
± 0
.01
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
Fe:P
O4 r
atio
(1
0 w
eeks
; mol
mol
-1)
234.
63 ±
87.
2420
8.26
± 6
1.60
268.
88 ±
81.
0516
7.56
± 5
2.50
314.
68 ±
44.
4216
4.05
± 5
6.14
157.
60 ±
25.
8421
8.09
± 7
5.27
224.
17 ±
113
.31
Cl (
mm
ol L
-1)
3.29
± 0
.07
3.12
± 0
.04
3.37
± 0
.09
4.17
± 0
.08
4.29
± 0
.17
4.06
± 0
.18
2.65
± 0
.03
2.69
± 0
.06
2.75
± 0
.06
Al (
µmol
L-1)
0.53
± 0
.15
0.90
± 0
.13
0.67
± 0
.09
0.53
± 0
.13
0.51
± 0
.06
0.54
± 0
.19
0.22
± 0
.03
0.28
± 0
.05
0.22
± 0
.03
Ca
(mm
ol L
-1)
1.30
± 0
.08
1.41
± 0
.08
1.46
± 0
.10
1.32
± 0
.09
1.30
± 0
.08
1.38
± 0
.17
1.35
± 0
.03
1.42
± 0
.11
1.60
± 0
.05
SO4 (
mm
ol L
-1)
0.63
± 0
.04
0.73
± 0
.05
0.63
± 0
.02
0.64
± 0
.02
0.69
± 0
.08
0.58
± 0
.03
0.66
± 0
.09
0.69
± 0
.08
0.71
± 0
.02
NH
4 (µ
mol
L-1)
21.8
0 ±
2.4
427
.02
± 1
.28
24.7
7 ±
3.6
323
.03
± 1
.80
22.2
5 ±
2.6
222
.35
± 1
.67
22.6
3 ±
1.7
422
.29
± 1
.50
24.1
5 ±
2.1
8
NO
2 (µ
mol
L-1)
0.23
± 0
.02
0.26
± 0
.01
0.25
± 0
.02
0.24
± 0
.01
0.23
± 0
.01
0.24
± 0
.01
0.24
± 0
.01
0.24
± 0
.01
0.24
± 0
.02
NO
3 (µ
mol
L-1)
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.03
± 0
.03
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
pH7.
46 ±
0.0
77.
50 ±
0.1
07.
73 ±
0.0
87.
50 ±
0.1
47.
29 ±
0.1
57.
59 ±
0.1
27.
24 ±
0.1
07.
25 ±
0.0
67.
46 ±
0.0
9
Alk
alin
ity
(mE
q L-1
)1.
45 ±
0.1
91.
45 ±
0.2
21.
72 ±
0.1
61.
43 ±
0.2
71.
29 ±
0.1
51.
64 ±
0.2
51.
04 ±
0.1
61.
20 ±
0.0
81.
38 ±
0.1
2
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a measure to restore water quality: implications for macrophyte growth
59
3
Sup
ple
men
tary
Tab
le 3
.1 –
Con
tinu
edM
acro
phyt
e sp
ecie
sP
otam
oget
on p
ecti
natu
sE
mpt
y
Pla
ce o
f add
itio
nM
ixed
Non
- M
ixed
Mix
ed
Iron
add
itio
n0
g Fe
m-2
20 g
Fe
m-2
40 g
Fe
m-2
0 g
Fe m
-220
g F
e m
-240
g F
e m
-20
g Fe
m-2
20 g
Fe
m-2
40 g
Fe
m-2
Surf
ace w
ater
Fe (µ
mol
L-1)
0.87
± 0
.09
0.95
± 0
.17
1.64
± 0
.18
0.75
± 0
.12
0.71
± 0
.07
1.24
± 0
.07
0.91
± 0
.06
1.04
± 0
.12
1.83
± 0
.19
Fe (p
reci
pita
ted;
µm
ol L
-1)
2.60
± 1
.18
1.81
± 0
.77
2.00
± 0
.79
12.1
7 ±
4.4
210
.89
± 1
.40
3.88
± 2
.24
4.98
± 1
.43
2.95
± 1
.08
4.91
± 1
.96
PO
4 (µ
mol
L-1)
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
Cl (
mm
ol L
-1)
2.52
± 0
.17
2.46
± 0
.06
2.59
± 0
.06
3.20
± 0
.05
3.20
± 0
.07
3.22
± 0
.06
3.20
± 0
.10
3.11
± 0
.17
3.13
± 0
.03
Al (
µmol
L-1)
1.17
± 0
.48
1.20
± 0
.12
0.79
± 0
.12
0.32
± 0
.07
0.38
± 0
.04
0.24
± 0
.02
0.34
± 0
.05
0.61
± 0
.10
0.30
± 0
.03
Ca
(mm
ol L
-1)
1.21
± 0
.04
1.35
± 0
.09
1.64
± 0
.11
1.45
± 0
.05
1.50
± 0
.03
1.64
± 0
.04
1.54
± 0
.05
1.54
± 0
.09
1.87
± 0
.11
SO4 (m
mol
L-1)
0.73
± 0
.04
0.73
± 0
.05
0.68
± 0
.02
0.75
± 0
.02
0.74
± 0
.04
0.63
± 0
.02
0.71
± 0
.03
0.72
± 0
.03
0.64
± 0
.02
NH
4 (µ
mol
L-1)
26.3
5 ±
1.4
326
.74
± 2
.36
24.9
7 ±
1.3
625
.65
± 0
.86
26.1
3 ±
1.2
825
.76
± 1
.51
27.0
9 ±
1.3
525
.46
± 0
.89
26.0
1 ±
1.5
3
NO
2 (µ
mol
L-1)
0.24
± 0
.01
0.26
± 0
.01
0.24
± 0
.00
0.23
± 0
.01
0.23
± 0
.00
0.23
± 0
.00
0.25
± 0
.01
0.23
± 0
.01
0.25
± 0
.01
NO
3 (µ
mol
L-1)
0.08
± 0
.00
0.04
± 0
.04
0.04
± 0
.03
0.12
± 0
.12
0.15
± 0
.15
0.51
± 0
.31
0.02
± 0
.02
0.03
± 0
.01
0.65
± 0
.46
pH7.
84 ±
0.3
78.
61 ±
0.1
77.
73 ±
0.1
67.
12 ±
0.2
37.
43 ±
0.1
27.
25 ±
0.1
27.
11 ±
0.0
98.
01 ±
0.1
97.
47 ±
0.1
8
Alk
alin
ity
(mE
q L-1
)0.
87 ±
0.0
90.
95 ±
0.1
71.
64 ±
0.1
80.
75 ±
0.1
20.
71 ±
0.0
71.
24 ±
0.0
70.
91 ±
0.0
61.
04 ±
0.1
21.
83 ±
0.1
9
Con
duct
ivit
y (1
0 w
eeks
; µS
cm-1)
756
± 2
8.18
508
± 6
.06
582.
8 ±
29.
8467
6.4
± 1
0.33
555.
8 ±
14.
6960
4.6
± 1
2.63
752.
2 ±
46.
4457
9 ±
22.
2063
2.2
± 2
2.35
Por
e wat
erFe
(µm
ol L
-1)
1.10
± 0
.14
1.76
± 0
.34
1.68
± 0
.23
0.96
± 0
.06
1.08
± 0
.20
1.23
± 0
.04
1.06
± 0
.16
0.98
± 0
.13
1.54
± 0
.25
PO
4 (µ
mol
L-1)
0.00
± 0
.00
0.00
± 0
.00
0.02
± 0
.02
0.01
± 0
.01
0.01
± 0
.01
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
Fe:P
O4 r
atio
(1
0 w
eeks
; mol
mol
-1)
157.
60 ±
25.
8419
8.74
± 8
2.92
272.
42 ±
139
.19
37.5
6 ±
12.
6156
.57
± 1
1.52
64.6
7 ±
13.
2162
.68
± 1
9.38
87.3
4 ±
22.
0660
.76
± 1
6.73
Cl (
mm
ol L
-1)
2.54
± 0
.16
2.53
± 0
.09
2.47
± 0
.20
3.27
± 0
.05
3.30
± 0
.07
3.16
± 0
.11
3.23
± 0
.10
3.27
± 0
.20
3.18
± 0
.07
Al (
µmol
L-1)
0.24
± 0
.01
0.47
± 0
.16
0.29
± 0
.08
0.22
± 0
.03
0.26
± 0
.03
0.22
± 0
.05
0.20
± 0
.02
0.31
± 0
.06
0.22
± 0
.02
Ca
(mm
ol L
-1)
1.46
± 0
.05
1.75
± 0
.16
1.70
± 0
.14
1.67
± 0
.05
1.68
± 0
.10
1.74
± 0
.07
1.67
± 0
.05
1.68
± 0
.10
1.89
± 0
.11
SO4 (
mm
ol L
-1)
0.75
± 0
.04
0.79
± 0
.07
0.68
± 0
.03
0.77
± 0
.03
0.76
± 0
.02
0.70
± 0
.04
0.71
± 0
.04
0.78
± 0
.04
0.64
± 0
.03
NH
4 (µ
mol
L-1)
22.6
7 ±
2.9
821
.01
± 2
.02
25.5
1 ±
0.9
625
.21
± 2
.32
23.5
6 ±
1.9
821
.50
± 2
.22
26.6
8 ±
3.3
223
.72
± 2
.03
24.0
2 ±
2.1
4
NO
2 (µ
mol
L-1)
0.23
± 0
.02
0.22
± 0
.01
0.26
± 0
.01
0.24
± 0
.01
0.23
± 0
.01
0.22
± 0
.01
0.25
± 0
.02
0.24
± 0
.02
0.25
± 0
.02
NO
3 (µ
mol
L-1)
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.00
± 0
.00
0.16
± 0
.16
0.00
± 0
.00
0.02
± 0
.02
0.36
± 0
.28
pH7.
11 ±
0.0
37.
42 ±
0.1
07.
53 ±
0.0
87.
00 ±
0.0
77.
06 ±
0.1
27.
32 ±
0.0
87.
15 ±
0.0
67.
16 ±
0.1
57.
56 ±
0.1
1
Alk
alin
ity
(mE
q L-1
)1.
10 ±
0.1
41.
76 ±
0.3
41.
68 ±
0.2
30.
96 ±
0.0
61.
08 ±
0.2
01.
23 ±
0.0
41.
06 ±
0.1
60.
98 ±
0.1
31.
54 ±
0.2
5
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
CHAPTER 4
Iron addition as a shallow lake restoration measure:
impacts on charophyte growth
Anne K. Immers, Masha T. van der Sande, Rene M. van der Zande,
Jeroen J. M. Geurts, Ellen van Donk, and Elisabeth S. Bakker
Hydrobiologia (2013) 710, 241-251.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 4
62
ABSTRACT
Eutrophication has caused a decline of charophyte species in many shallow lakes in Europe. Even
though external inputs of phosphorus are declining, internal loading of P from the sediment
seems to delay the recovery of these systems. Iron is a useful chemical binding agent to combat
internal phosphorus loading. However, the effects of iron addition on charophytes are not yet
known. In this study we experimentally tested the potential toxicity of iron(III)chloride (FeCl3)
on two different charophytes, Chara virgata Kützing and Chara globularis Thuiller added at the
concentration of 20 g Fe m-2 and 40 g Fe m-2 to the surface water. Chara virgata growth was
not significantly affected, whereas C. globularis growth significantly decreased with increasing
iron concentrations. Nonetheless, biomass of both species increased in all treatments relative
to starting conditions. The decrease of C. globularis biomass with high iron additions may have
been caused by a drop in pH and alkalinity in combination with iron induced light limitation.
Iron addition over a longer time scale, however, will not cause this rapid drop in pH. Therefore
we conclude, that adding iron(III)chloride in these amounts to the surface water of a lake can
potentially be a useful restoration method.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a shallow lake restoration measure: impacts on charophyte growth
63
4
INTRODUCTION
Submerged macrophytes play a crucial role in the maintenance of water transparency and aquatic
biodiversity in shallow water bodies (Timms and Moss, 1984; Scheffer et al., 1993). However,
macrophyte species seem to differ in the success at which they perform this role (Engelhardt and
Ritchie, 2001). Particularly the group of charophytes (Characeae) has been documented to be
more successful in maintaining water clarity than for example Potamogeton species (Hargeby et
al., 2007, Ibelings et al., 2007, Bakker et al., 2010). Charophytes are green macroalgae, the closest
ancestors of land plants (Karol et al., 2001), which are known as species of high conservation
value (Lamers et al., 2006) and are commonly found in clear, hard, and nutrient poor water
bodies of relatively high alkalinity (Simons and Nat, 1996; Van den Berg et al., 1998b; Kufel
and Kufel, 2002). Under these conditions, charophytes can improve their own light climate by
forming dense beds on the sediment surface (Kufel and Kufel, 2002; Van Donk and Van de Bund,
2002), which have a high nutrient uptake, enhance sedimentation, and counteract fish or wind
induced sediment resuspension (Scheffer et al., 1993; Van den Berg et al., 1998a; Van den Berg
et al., 1999; Kufel and Kufel, 2002). Charophytes may also directly reduce phytoplankton and
periphyton growth by releasing allelopathic substances (Mulderij et al., 2003).
High nutrient loading and a subsequent increase in water turbidity due to phytoplankton
surface blooms have led to a decrease of charophytes in many shallow lakes in Europe (Van den
Berg et al., 1998a; Van den Berg et al., 1998b; Klosowski et al., 2006; Lambert and Davy,
2010). Recent restoration measures, where external phosphorus (P) input and water turbidity
were experimentally reduced, have led to the return of dense charophyte beds (Van den Berg et
al., 1998a; Meijer et al., 1999; Ibelings et al., 2007). These restoration measures, however, were
performed in sandy lakes, whereas peaty lakes are suffering from high internal loading of P from
the sediment and are more prone to sediment resuspension (Cooke et al., 1993a; Jeppesen et al.,
1998; Søndergaard et al., 2003). Under natural conditions, peaty lakes in The Netherlands would
not suffer from internal P loading, as upwelling iron rich groundwater binds to phosphorus (in
the form of phosphate, PO4) in the sediment. This seepage, however, has disappeared over the
years due to high regional and local use of groundwater (Smolders and Roelofs, 1996; Van der
Welle et al., 2007b). Water managers have tried to resolve this problem by adding iron (Fe), in
the form of iron(III)chloride, to the lake sediment as a natural P binding agent (Cooke et al.,
1993a; Boers et al., 1994; Burley et al., 2001). In this way, the iron would not only precipitate
with the available P in the sediment, but would also form a barrier on the top layer of the
sediment, preventing internal P loading of the lake in the future. However, lake restoration by
adding iron in the lake sediment is a costly and time consuming process, therefore adding iron to
the surface water may be more feasible in case of restoration of a whole lake. The effect of this iron
addition, and the consequential potential drop in pH, on various organisms in the aquatic food
web is not yet well studied, whereas it is very important to know whether iron addition may be
harmful for the target species that are aimed to return to the restored lake.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 4
64
Charophytes are desirable species for water managers to grow in a lake as they are indicators
of good water quality (Lambert and Davy, 2010) and have been shown to return in peat lakes after
restoration measures had been taken including external nutrient reduction (Rip et al., 1992) and
biomanipulation (Ter Heerdt and Hootsmans, 2007). As charophytes primarily utilize nutrients
from the water column instead of the sediment (Kufel and Kufel, 2002; Hidding et al., 2010),
possible effects of iron on charophytes would be more pronounced when adding iron in the water
column.
The aim of this study was to test whether iron affects the growth, biomass allocation and
nutrient concentration of two different charophyte species. The experiment was based upon the
situation of Lake Terra Nova, The Netherlands, in which this method of FeCl3 addition to the
surface water is now being applied.
METHODS
Experimental set-up
Mesocosm experiments were performed in May 2010 in 45 Perspex cylinders (d × h = 10 cm
× 50 cm) which were placed in a temperature controlled culture room at the NIOO-KNAW in
Nieuwersluis. Temperature was kept constant at 19 °C and light regime was set at 12 hours light
and 12 hours darkness with a light intensity at the water surface of 100 ± 5 µmol photons m-2
s-1. Each cylinder was filled up with 0.50 L peat sediment, collected on April 2010 in Lake Terra
Nova (52º12’N, 5º02’E, The Netherlands), and subsequently very carefully 3.25 L of filtrated
(0.2 µm, ME 24, Whatman, Brentford, UK) Terra Nova water was poured on the sediment. To
enable pore water sampling, Rhizon soil moisture samplers (Eijkelkamp Agrisearch Equipment,
Giesbeek, The Netherlands) attached to 50 mL vacuum syringes were inserted into the upper
layer of the sediment.
During the experiment we manipulated 2 factors: namely the iron addition and the plants on
which the effects of iron addition were tested. The iron and plant treatments consisted each of
three levels. The effects of iron addition were tested during 5 weeks, with three different levels
of iron which would correspond to additions in Lake Terra Nova of 20 g Fe m-2 (low) and 40 g Fe
m-2 (high) in the form of FeCl3 and a control addition (0 g Fe m-2) was designed which received
NaCl in equal molar amounts of chloride in the high iron additions. The plant treatment levels
consisted of cylinders filled with Chara virgata Kützing, Chara globularis Thuiller, and empty
cylinders. All nine combinations of levels were experimentally tested with 5 replicates, which
were randomized in blocks.
Chara virgata was collected from experimental ponds in Loenderveen (52°12’N, 5°02’E, The
Netherlands) on 29 April 2010. Chara globularis was prior to the experiment grown in aquaria
from propagules in Terra Nova sediment. A bundle composed of 3 C. virgata shoots was planted
in the sediment of 15 cylinders (total FW per cylinder 0.16 ± 0.04 g), a bundle of 3 C. globularis
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a shallow lake restoration measure: impacts on charophyte growth
65
4
shoots in 15 other cylinders (total FW per cylinder 0.89 ± 0.38 g), and the last 15 cylinders were
not planted with macroalgae as controls.
To distinguish between the effects of iron toxicity and P limitation we reduced P in control
iron additions at the onset of the experiment with a low dose of 0.33 mg FeCl3 per cylinder. During
the experiment, iron was added two times every week on 8 addition days, which corresponds to
the low and high iron addition of 28.75 and 57.50 mg FeCl3 per addition day, respectively.
Sampling and analysis
Once every week during the experiment 35 mL samples of surface water were taken from each
cylinder for chemical analyses. A subsample of 10 mL from each cylinder was filtrated over
Whatman GF/C (1.2 µm) filters and subsequently stored at -20 °C before nutrient analysis. The
remaining 25 mL subsample was used to measure pH and alkalinity with a TIM840 titration
manager (Radiometer Analytical, Copenhagen, Denmark). Alkalinity was determined by titrating
with 0.01 M HCl down to pH 4.2. The stored 10 mL subsamples were used to colorimetrically
determine PO4, NH
4, and NO
3 with a QuAAtro CFA flow analyzer (Seal Analytical, Norderstedt,
Germany).
During the last sample day, in addition to prior analyses, 50 mL of sediment pore water
samples were collected from each cylinder using Rhizon soil moisture samplers. Samples were
stored in 50 mL centrifuge tubes at -20 °C directly after the pore water had been collected. The
same volume of surface water was, prior to storage in 50 mL centrifuge tubes at -20 °C, filtrated
over a 0.45 μm membrane filter (ME 25, Whatman, Brentford, UK). Membrane filters that were
used were afterwards dried for 24 hours at 60 °C and later stored in 50 mL centrifuge tubes at
-20 °C. Analyses of stored samples were performed using an inductively coupled plasma emission
spectrophotometer (ICP; Liberty 2, Varian, Bergen op Zoom, The Netherlands) according to the
Dutch NEN-EN-ISO 17294 to estimate dissolved Fe, Al, Ca, and S in surface and pore water.
The same method was used to measure precipitated Fe in the surface water, which was prior to
analysis collected by filtration of surface water on 0.45 μm membrane filters (ME 25, Whatman,
Brentford, UK), that were subsequently treated with 8 mL nitric acid (2 M).
At the end of the experiment, ± 3 cm of shoot material from each cylinder was placed in a
plastic cup with 20 mL of demineralized water for periphyton determination following Zimba
and Hopson (1997). Each cup was shaken gently for 1 minute and subsequently shoot material
was taken out, dried for 24 hours at 60 °C and weighed. Demineralized water with periphyton
was filtered over a Whatman GF/C (1.2 µm) filter, and afterwards filters were dried for 24 hours
at 60 °C and weighed. Subsequently all charophytes were harvested and separated in above-
and belowground material. All material was dried for 24 hours at 60 °C, dried shoots from
periphyton determination were added and subsequently all material was weighed to determine
the total above- and belowground dryweight. Total dryweight at the start of the experiment was
calculated with a conversion factor, which was acquired from the fresh and dryweight of several
subsamples (for C. virgata dry weight = 30% of fresh weight, for C. globularis dryweight = 18%
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 4
66
of fresh weight). A homogenized portion of dry charophyte material was used to determine both
C and N concentrations with a FLASH 2000 Organic Elemental Analyzer (Interscience, Breda,
The Netherlands). Charophyte P concentrations were acquired by incinerating homogenized dry
material for 30 minutes at 500 °C, followed by digestion in H2O
2 (Murphy and Riley 1962)
before analysis with a QuAAtro CFA flow analyzer.
Statistical analysis
Statistical analyses were carried out with SPSS 18.0 (SPSS, Chicago, IL, USA). Differences
between treatments for plant biomass, shoot:rhizoid ratio, and plant nutrient composition
were tested with one-way ANOVA’s with iron treatment as a fixed factor followed by a Tukey’s
post-hoc test. Differences in chemical variables and periphyton growth were tested with two-
way ANOVA’s with iron treatment and plant treatment (consisting of the levels C. virgata,
C. globularis, or empty cylinders) as fixed factors followed by a Tukey’s post-hoc test. Prior to
analysis, all data were tested for normality and homogeneity of variance, and if necessary, data
were log 10 transformed. For data that had no normal distribution, even after transformation, a
nonparametric Kruskal-Wallis test was used with Statistica 9.1 (StatSoft Inc., Tulsa, OK, USA)
to analyze variances. Results were expressed as mean ± standard error of mean and P ≤ 0.05 was
accepted for statistical significance.
RESULTS
Charophyte response
Both charophyte species biomass increased notably over the 5 weeks that the experiment
ran. Chara virgata experienced on average a 4-fold increase, from 0.05 ± 0.00 to 0.20 ± 0.02
g dryweight, whereas Chara globularis, which started with a higher mean biomass of 0.15 ±
0.02 g dryweight, increased on average 3-fold to 0.51 ± 0.04 g dryweight. Iron additions had
different effects on the two species (Figure 4.1). Chara virgata above ground and below ground
biomass were not significantly affected by iron additions (Table 4.1), although at the highest
level of iron addition C. virgata biomass tended to be somewhat lower (Figure 4.1). The growth
of C. globularis, however, was negatively affected by iron additions (Figure 4.1). Chara globularis
below ground material, which only on average made up 6 % of total biomass, did not differ
between iron additions, but above ground material was considerably lower in cylinders which
received iron compared to cylinders in which no iron was added (Table 4.1). Total biomass, which
was on average composed of 94 % above ground material thus decreased with increasing iron
concentrations (Table 4.1). Biomass allocation of both C. virgata and C. globularis was not affected
by iron addition, as charophyte shoot:rhizoid ratio did not differ between iron additions
(Table 4.1).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a shallow lake restoration measure: impacts on charophyte growth
67
4
Figure 4.1 – Biomass increase (average ± sem) in reaction to iron addition after 5 weeks for Chara virgata and Chara globularis. White, grey, and black bars represent respectively additions of 0, 20, and 40 g Fe m-2. Significant differences between iron additions are indicated for each species separately by different letters (Analysis of variance, Tukey test, P ≤ 0.05).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 4
68
Table 4.1 – Mean (± SEM) end results of charophyte biomass, growth, shoot:rhizoid ratio, and nutrient composition of C. virgata and C. globularis at different iron additions. Data were analysed with a one-way ANOVA with the levels of iron treatment (0, 20, or 40 g m-2) as a fixed factor, n=5. Significant differences between iron additions are indicated for each species separately by different letters (Analysis of variance, Tukey test, P ≤ 0.05). Bold values indicate P ≤ 0.05.
Mean ± sem Effect iron amountDf=2, 14
0 g Fe m-2 20 g Fe m-2 40 g Fe m-2 F P
C. virgata
Biomass below ground (g) 0.03 ± 0.01 0.03 ± 0.01 0.02 ± 0.00 1.49 0.26
Biomass above ground (g) 0.19 ± 0.04 0.20 ± 0.04 0.13 ± 0.02 1.03 0.39
Total biomass (g) 0.22 ± 0.05 0.23 ± 0.04 0.15 ± 0.02 1.13 0.36
Total biomass increase (g) 0.17 ± 0.05 0.18 ± 0.04 0.10 ± 0.02 1.14 0.35
Shoot:rhizoid ratio (g g-1) 0.87 ± 0.03 0.89 ± 0.01 0.90 ± 0.02 0.66 0.54
C (mg g dryweight-1) 273.90 ± 14.25 272.51 ± 2.79 291.96 ± 10.85 1.07 0.37
N (mg g dryweight-1) 20.47 ± 2.51 21.38 ± 0.08 24.95 ± 0.92 2.35 0.14
P (mg g dryweight-1) 1.81 ± 0.13 1.66 ± 0.13 1.82 ± 0.06 0.62 0.56
C:N ratio (mol mol-1) 16.14 ± 1.25 14.87 ± 0.13 13.68 ± 0.41 2.60 0.12
N:P ratio (mol mol-1) 25.43 ± 3.06 29.18 ± 2.32 30.50 ± 1.64 1.19 0.34
Periphyton (g g dryweight-1) 0.38 ± 0.06ab 0.21 ± 0.04a 0.44 ± 0.07b 3.39 0.04
C. globularis
Biomass below ground (g) 0.03 ± 0.01 0.02 ± 0.00 0.02 ± 0.00 3.07 0.08
Biomass above ground (g) 0.65 ± 0.05a 0.44 ± 0.02b 0.34 ± 0.03b 22.03 < 0.001
Total biomass (g) 0.69 ± 0.05a 0.46 ± 0.02b 0.39 ± 0.02b 21.85 < 0.001
Total biomass increase (g) 0.51 ± 0.02a 0.34 ± 0.01b 0.23 ± 0.01c 66.66 < 0.001
Shoot:rhizoid ratio (g g-1) 0.96 ± 0.01 0.96 ± 0.01 0.95 ± 0.01 0.91 0.43
C (mg g dryweight-1) 258.12 ± 5.66 267.11 ± 4.70 270.17 ± 15.76 0.42 0.67
N (mg g dryweight-1) 14.86 ± 0.82a 20.91 ± 1.20b 23.12 ± 1.70b 10.93 0.002
P (mg g dryweight-1) 1.10 ± 0.02a 1.21 ± 0.01b 1.46 ± 0.03c 67.47 < 0.001
C:N ratio (mol mol-1) 20.52 ± 1.22a 15.08 ± 0.83b 13.70 ± 0.48b 14.39 0.001
N:P ratio (mol mol-1) 29.89 ± 1.68a 38.19 ± 1.90b 35.17 ± 1.80ab 5.45 0.02
Periphyton (g g dryweight-1) 0.17 ± 0.05a 0.50 ± 0.08ab 0.81 ± 0.18b 7.63 0.01
Tissue nutrient concentrations for C. virgata increased significantly during the experiment
for N and P respectively from 12.58 ± 0.35 to mean end concentrations of 22.27 ± 1.14 mg N g
dryweight-1 and from 1.05 ± 0.01 to mean end concentrations of 1.76 ± 0.06 mg P g dryweight-1.
Different iron additions, however, did not induce any differences in N or P concentrations and
their relative ratios in this charophyte (Table 4.1). This relationship was not seen in the tissue of
C. globularis, where the control iron addition (0 g Fe m-2) remained similar to the start conditions
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a shallow lake restoration measure: impacts on charophyte growth
69
4
(1.18 ± 0.01 mg P g dryweight-1 and 12.67 ± 0.52 mg N g dryweight-1) and only the iron
additions of 20 and 40 g Fe m-2 induced a significant increase in N and P concentrations and their
relative ratios (Table 4.1).
The amount of periphyton, the reddish colored material growing on the charophyte shoots
(Figure 4.2), was clearly affected by iron additions. For cylinders containing C. virgata, the high
iron addition (40 g Fe m-2) yielded significantly more periphyton than the low iron addition (20 g
Fe m-2). Cylinders containing C. globularis, on the other hand, showed no difference in periphyton
biomass between the iron additions, but the high iron addition had considerably more periphyton
biomass than the control iron addition (0 g Fe m-2; Table 4.1).
Moreover, during the experiment a large number of charophyte propagules sprouted from the
sediment, which did not seem to be affected by the different iron additions.
Changes in water properties
Surface water pH decreased significantly due to iron additions and at the end of the experiment
surface water pH reached mean values of 6.95 ± 0.17 in the high iron additions, 7.81 ± 0.13
in the low iron additions and mean values of 8.35 ± 0.22 in the control additions (Table 4.2;
Figure 4.3a). Alkalinity showed the same relationship with low mean values of 0.62 ± 0.04 mEq
L-1 in the high iron additions, 0.95 ± 0.08 mEq L-1 for the low iron additions and the highest
mean values of 1.55 ± 0.20 mEq L-1 in the control additions. Moreover, alkalinity also differed
between the charophyte species, with a significant lower alkalinity of 0.62 ± 0.03 mEq L-1 in the
C. globularis cylinders compared to the empty cylinders or cylinders with C. virgata (1.25 ± 0.16
and 1.24 ± 0.18 mEq L-1; Table 4.2; Figure 4.3b).
Iron and aluminum concentrations in the surface water decreased with higher iron additions,
however, concentrations in the surface water were very low with mean iron concentrations ranging
between 0.37 ± 0.05 and 0.14 ± 0.04 µmol Fe L-1 and mean aluminum concentrations ranging
between 1.93 ± 0.15 and 0.21 ± 0.05 µmol Al L-1. This difference was possibly due to the
precipitation of iron with phosphate, however phosphate concentrations did not differ between
iron and control additions, as P was reduced in the control additions (0 g Fe m-2) at the onset of
the experiment.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 4
70
Figure 4.2 – (a) Periphyton material on shoots in g g dryweight-1 (average ± sem) in reaction to different iron additions. Periphyton may include other material such as precipitated iron. White, grey, and black bars represent respectively additions of 0, 20, and 40 g Fe m-2. Significant differences between iron additions are indicated for each species separately by different letters (Analysis of variance, Tukey test, P ≤ 0.05). Pictures taken at the end of the experiment of Chara globularis receiving (b) 0 g Fe m-2 and (c) 40 g Fe m-2.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a shallow lake restoration measure: impacts on charophyte growth
71
4
Table 4.2 – Results of analysis of the effects of iron addition on surface and pore water nutrient composition. Data were analysed with a two-way ANOVA (F) or non-parametric Kruskal-Wallis (H) with the levels of iron treatment (0, 20, or 40 g m-2) and the levels of plant treatment (Chara virgata, Chara globularis, or empty cylinders) as fixed factors, n=5. Bold values indicate P ≤ 0.05
Effect Iron amountDf=2, 36
Macrophyte species Df=2, 36
Iron × MacrophyteDf=4, 36
F / H P F / H P F / H P
Surface water
pH* 18.31 < 0.001 1.73 0.42 22.81 0.004
Alkalinity* 14.45 < 0.001 14.66 < 0.001 33.96 < 0.001
Fe* 16.64 < 0.001 1.67 0.43 19.15 0.01
Fe (precipitated)* 1.29 0.52 6.05 0.05 11.77 0.16
Al* 31.22 < 0.001 0.28 0.87 33.10 < 0.001
PO4
2.86 0.07 2.80 0.07 1.63 0.19
NO3* 5.71 0.06 18.48 < 0.001 28.50 < 0.001
NH4* 3.27 0.20 33.31 < 0.001 37.50 < 0.001
Ca* 5.57 0.06 2.39 0.30 13.28 0.10
S 0.21 0.81 2.18 0.13 0.28 0.89
Pore water
Fe* 1.59 0.45 0.31 0.86 4.52 0.81
Al* 21.55 < 0.001 0.36 0.83 25.69 0.001
PO4* 0.05 0.98 10.50 0.01 12.44 0.13
Fe:PO4* 2.20 0.33 5.34 0.07 9.98 0.27
NO3* 9.90 0.01 14.80 < 0.001 25.80 0.001
NH4* 0.37 0.83 2.10 0.35 3.96 0.86
Ca 3.16 0.04 2.65 0.08 0.90 0.47
S 0.04 0.96 0.26 0.77 0.95 0.49
* Non-parametric Kruskal-Wallis test (H) performed instead of ANOVA (F)
Iron and phosphate concentrations in the pore water showed the same ratio with the different
iron additions. As a result Fe:PO4 ratios in sediment, which are often used as a tool to determine
internal phosphorus loading, reached mean values of 16.98 ± 4.21 mol mol-1, but did not differ
significantly between the iron additions. Phosphate also seemed to be lower in the surface water
of the cylinders containing C. globularis where P decreased to mean values of 0.05 ± 0.00 µmol
L-1 compared to cylinders with C. virgata (0.08 ± 0.01 µmol L-1) and empty cylinders (0.08 ±
0.01 µmol L-1), however this difference was not significant (Table 4.2; Figure 4.3c). Precipitated
iron, which was measured in the surface water, reached highest values in the cylinders which
contained no charophytes (Figure 4.3d). No difference was found for precipitated iron between
iron additions.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 4
72
Figure 4.3 – Surface water (a) pH, (b) Alkalinity, (c) PO4, (d) precipitated Fe, (e) NO
3, and (f) NH
4
concentrations in mEq L-1 and µmol L-1 (average ± sem) after 5 weeks for the different plant treatment levels under different iron additions. White, grey, and black bars represent respectively cylinders receiving iron additions of 0, 20, and 40 g Fe m-2. Significant differences between iron additions are indicated for each species separately by different letters (Kruskal-Wallis, P ≤ 0.05).
Nitrogen, in the form of NO3 and NH
4, decreased significantly during the experiment in
the surface water of all cylinders. Nitrate showed a clear significant relationship for the type of
charophyte presence in cylinders, with constantly lower values (approaching 0) in cylinders with
C. globularis compared to higher values in empty cylinders and cylinders with C. virgata (Table
4.2; Figure 4.3e). Ammonium reached highest mean concentrations in cylinders containing
C. virgata (107.93 ± 0.42 µmol L-1), which differed significantly from cylinders containing
C. globularis (105.48 ± 0.20 µmol L-1) and empty cylinders (104.33 ± 0.19 µmol L-1; Figure
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a shallow lake restoration measure: impacts on charophyte growth
73
4
4.3f). No significant differences were found between treatments for calcium and sulphur and
concentrations remained constant at 962.54 ± 48.42 and 331.48 ± 12.32 μmol L-1 for Ca and
S, respectively. Over all, pore water nutrient concentrations seemed to be less affected by the
presence/ absence of charophyte species (Table 4.2).
DISCUSSION
The decrease of C. globularis biomass with increasing iron concentrations might be related to iron
toxicity. Negative effects of iron addition on the growth of macrophytes are usually distinguished
in two different kinds, namely direct and indirect (Wheeler et al., 1985). According to Van
der Welle et al. (2007b), direct effects of iron toxicity can be seen in the physical structure of
plants. It can act on the leaves by reducing the size or by the formation of black necrotic spots or
complete discoloration of leaves and even die-back of old leaves, or in roots which can blacken,
stop growing or lack branching (Van der Welle et al., 2006). Other described unfavorable effects
were the formation of iron plaques on roots, which could prevent plant nutrient uptake (Van
der Welle et al., 2007b). These physical symptoms, indicating direct iron toxicity could not be
detected in our experiment with C. virgata and C. globularis. Charophytes differ greatly from
vascular macrophytes in having only a rhizoid system, on which they do not rely on for nutrient
uptake (Kufel and Kufel, 2002). These processes of direct iron toxicity as found in vascular
macrophytes therefore may not apply for charophytes.
For most higher plant species, iron can have an indirect negative effect on growth by mainly
limiting the macronutrient P due to the precipitation of phosphate with iron (Wheeler et al.,
1985). According to Koerselman and Meuleman (1996), macrophytes are P limited at N:P ratios
measured in plant biomass above 16 and N limited at N:P ratios below 14. Charophytes, however,
are usually only found in lakes with low inorganic P concentrations (Bloemendaal and Roelofs,
1988; Simons and Nat, 1996), and are known to give way to higher plants with increasing
phosphorus concentrations (Kufel and Kufel, 2002; Lambert and Davy, 2011). Moreover, for
charophyte species, the measured concentrations of the macronutrients N and P in plant material
not only varies greatly between species, it also differs within species, and usually only gives an
indication of the environment in which the charophytes are growing (Kufel and Kufel, 2002). In
our experiment N and P concentrations in C. globularis increased with increasing iron addition
whereas this did not happen in C. virgata, at least not significantly. For both species the N:P
ratio was above 16, suggesting P-limitation if this threshold can be used for Characeae. However,
if considering actual concentrations for both N and P, both species were always above limiting
levels of 13 and 1.3 mg g dryweight-1 (Gerloff and Krombholz, 1966) for N and P, respectively,
indicating that these plants were not limited by these nutrients. Measurements of water nutrients
did not show evidence of increasing P limitation as well, but indicated a strong reduction of
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 4
74
nitrate by C. globularis relative to C. virgata and the cylinders with no plants, whereas there were
no differences in phosphate. Most research on charophyte growth limitation has focused on the
effects of P, but recently Lambert and Davy (2011) showed that N, particularly nitrate, may
strongly affect growth and abundance of Characeae. The accumulation of N and P in the tissue
of C. globularis in our study may be explained by the reduced growth of this species at higher
iron addition, which would simultaneously explain the lack of significant changes in N and P
concentration in C. virgata tissue as this species did not experience a significant growth reduction
with iron addition. As less biomass is formed, nutrients may accumulate in plant tissue. Reduced
growth can in this case be the result of toxic effects of iron, or the fact that other factors have
become limiting.
In addition to nutrients, light can be a limiting factor of plant growth. Another indirect
negative effect of iron addition could be the formation of iron precipitates and their shading effect
on shoots. No differences were found between iron additions for the presence of precipitated iron,
however, precipitated iron was only measured in surface water and not on charophyte shoots,
cylinders or on the sediment surface. Most of the iron could have accumulated on these surfaces as
iron-phosphates or iron oxides. The amount of measured periphyton material on shoots did show
a relation with iron concentrations, as highest periphyton biomass for both species in the high
iron additions. Whereas the method of shaking plant shoots is commonly applied to quantify
periphyton biomass on the plants, other material on the leaves, such as the iron precipitates is
included in this measurement. When looking at the color of the periphyton and the difference
between periphyton in the high iron and in the control additions, the reddish colored periphyton
in iron additions does most probably contain iron precipitates. For charophytes, light is a crucial
factor for growth (Kufel and Kufel, 2002; Rip et al., 2007). Consequently, dense growth of
periphyton and iron precipitation could have limited charophyte growth in high iron additions.
The addition of iron also resulted in a decrease in pH and alkalinity in the cylinders receiving
high iron additions. Even though the pH stayed well within the optimal range of 5-7 for
maximal iron phosphate binding capacity (Cooke et al., 1993a), the lower pH and alkalinity
were suboptimal for the charophytes, as they require a high pH and high alkalinity of the surface
water (Van den Berg et al., 1998b; Klosowski et al., 2006; Lambert and Davy, 2011). Not only
was there a significant difference in alkalinity between the different iron additions, there was also
a difference between charophyte species. Cylinders containing C. globularis proved to have a lesser
buffer capacity than empty cylinders and cylinders containing C. virgata. This difference might
well explain the difference in iron sensitivity, where C. globularis was considerably more affected
by iron additions than C. virgata. According to Van den Berg et al. (2002), growth of charophytes
is strongly correlated to the bicarbonate (HCO3
-) concentrations in the water. The inability of C.
globularis to maintain the buffer capacity in combination with light limitation could therefore
have resulted in decreasing photosynthesis rates and a steady drop in pH in cylinders of the iron
additions due to the quick addition of iron.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition as a shallow lake restoration measure: impacts on charophyte growth
75
4
Iron as a measure to control eutrophication
The goal of adding Fe to the surface water of lakes is to lower surface water P and to control
internal P release. The binding capacity of Fe, however, is regulated by the redox state of the agent
(Burley et al., 2001). Under oxic conditions, oxidized ferric iron (Fe3+) can freely precipitate with
PO4, but under anoxic conditions, reduced ferrous iron (Fe2+) is formed and Fe loses this binding
capacity and consequently PO4 will be released (Cooke et al., 1993a). Charophytes are able to
oxidize the sediment, thereby preventing this redox-reaction to occur (Kufel and Kufel, 2002).
Moreover, the possibility for charophytes to use bicarbonate as a carbon source for photosynthesis
leads to the formation of carbonate, which in turn can precipitate with calcium to form calcite
(Otsuki and Wetzel, 1972). Calcite can subsequently co-precipitate with phosphate, which is a
redox-insensitive reaction (Otsuki and Wetzel, 1972). Charophytes can thus enhance the binding
capacity of iron.
The negative effects of the addition of 40 g Fe m-2 on C. globularis biomass may have partly
been due to the fact that iron was added over a short period of 5 weeks. When using iron addition
as a lake restoration measure, the choice can be made for addition distributed over a longer time
period. Moreover, a drop in pH and alkalinity as observed in this experiment will probably not
occur in a lake such as Terra Nova with the same amount of iron, as the water column above the
sediment is much larger and therefore negative consequences of iron addition such as a drop in
pH and alkalinity would be much less dramatic (Boers et al., 1994).
From the fact that both species reacted differently on iron addition it might follow that
after iron addition, lakes would become dominated by more iron tolerant species, which could
possibly cause a shift in community composition. However, the fact that the addition of iron
to a fresh water ecosystem will reduce the phosphate concentration in the water and sediment
by forming a Fe-trapping barrier on the sediment-water interface will be favorable to push the
equilibrium towards a clear, charophyte dominated ecosystem. And as charophyte establishment
was not hampered by the iron layer on the sediment, dense charophyte beds can provide a positive
feedback loop resulting in a resilient, clear water state.
ACKNOWLEDGEMENTS
The authors would like to thank Thijs de Boer, Koos Swart, and Martijn Dorenbosch for
their practical assistance in the field and Nico Helmsing and Harry Korthals for performing
multiple chemical analyses in the lab. This study was funded by the Water Framework Directive
Innovation Fund from Agentschap NL from the Dutch Ministry of Economic Affairs, Agriculture
and Innovation.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
CHAPTER 5
Invasive crayfish threaten the development of
submerged macrophytes in lake restoration
Jessica E. M. van der Wal, Martijn Dorenbosch, Anne K. Immers,
Constanza Vidal Forteza, Jeroen J. M. Geurts, Edwin T. H. M. Peeters,
Bram Koese, and Elisabeth S. Bakker
PLoS One (2013) 8, e78579.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
78
ABSTRACT
Submerged macrophytes enhance water transparency and aquatic biodiversity in shallow water
ecosystems. Therefore, the return of submerged macrophytes is the target of many lake restoration
projects. However, at present, North-Western European aquatic ecosystems are increasingly
invaded by omnivorous exotic crayfish. We hypothesize that invasive crayfish pose a novel
constraint on the regeneration of submerged macrophytes in restored lakes and may jeopardize
restoration efforts. We experimentally investigated whether the invasive crayfish (Procambarus
clarkii Girard) affects submerged macrophyte development in a Dutch peat lake where these
crayfish are expanding rapidly. Seemingly favourable abiotic conditions for macrophyte growth
existed in two 0.5 ha lake enclosures, which provided shelter and reduced turbidity, and in one lake
enclosure iron was added to reduce internal nutrient loading, but macrophytes did not emerge. We
transplanted three submerged macrophyte species in a full factorial exclosure experiment, where
we separated the effect of crayfish from large vertebrates using different mesh sizes combined with
a caging treatment stocked with crayfish only. The three transplanted macrophytes grew rapidly
when protected from grazing in both lake enclosures, demonstrating that abiotic conditions for
growth were suitable. Crayfish strongly reduced biomass and survival of all three macrophyte
species while waterfowl and fish had no additive effects. Gut contents showed that crayfish were
mostly carnivorous, but also consumed macrophytes. We show that P. clarkii strongly inhibit
macrophyte development once favourable abiotic conditions for macrophyte growth are restored.
Therefore, expansion of invasive crayfish poses a novel threat to the restoration of shallow water
bodies in North-Western Europe. Prevention of introduction and spread of crayfish is urgent, as
management of invasive crayfish populations is very difficult.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
79
5
INTRODUCTION
Submerged macrophytes play a key role in shallow freshwater ecosystems by increasing nutrient
retention, stabilizing sediment, and providing food and habitat for macro-invertebrates, fish, and
birds (Carpenter and Lodge, 1986). A high abundance of submerged macrophytes is therefore
considered to be an important variable in maintaining the clear water state in shallow lakes
(Scheffer, 2001). However, increased nutrient loading of shallow water systems during the last
decades resulted in turbid waters and a strong decline of macrophyte abundance (Jeppesen et al.,
1998; Gulati and Van Donk, 2002). To restore water transparency and macrophyte vegetation,
external nutrient loading has been reduced and additional measures like the removal of
benthivorous fish have been taken (Moss, 1989; Jeppesen et al., 2005; Søndergaard et al., 2007;
Gulati et al., 2008). These measures have only been temporarily successful (Søndergaard et al.,
2007). Especially in lakes that are rich in organic sediments, internal phosphorus (P) loading still
leads to high nutrient levels (Phillips et al., 1994; Søndergaard et al., 2003).
To minimize P release from lake sediments into the water column, several chemical
phosphorus-binding agents have been applied, like calcium, aluminium, and iron (Boers et al.,
1994; Burley et al., 2001; Hickey and Gibbs, 2009), leading to reduced internal P loading
and increased water transparency in several studies (Boers et al., 1994; Smolders and Roelofs,
1995). However, increased water transparency does not always result in the return of submerged
macrophytes (Lauridsen et al., 2003; Jeppesen et al., 2005). This can be due to other unsuitable
abiotic conditions for macrophyte development or to limiting biotic factors such as grazing by
herbivores (Bakker et al., 2013). Waterfowl and fish can strongly reduce biomass of planted
macrophytes in restored lakes (Lauridsen et al., 1993; Søndergaard et al., 1996; Lauridsen et al.,
2003; Irfanullah and Moss, 2004; Van de Haterd and Ter Heerdt, 2007) as well as spontaneous
development of macrophyte communities (Van Donk and Otte, 1996; Hilt, 2006), even though
the latter is not found in all restoration projects (Perrow et al., 1997; Strand and Weisner, 2001;
Marklund et al., 2002). However, large fish and waterfowl are no longer the only potential grazers
as European shallow lakes are increasingly colonised by invasive crayfish such as the red swamp
crayfish (Procambarus clarkii; Geiger et al., 2005; Gherardi, 2006; Souty-Grosset et al., 2006).
In The Netherlands currently six species of exotic crayfish have established, whereas the
native crayfish Astacus astacus is almost extinct due to the crayfish plague (Koese and Soes, 2011).
Crayfish may reduce the standing stock of macrophytes by direct consumption (Lodge and
Lorman, 1987; Gherardi et al., 2011), increase water turbidity through sediment resuspension
(Rodríguez-Villafañe et al., 2003), and destroy macrophyte biomass by non-consumptive plant
shredding (Cronin et al., 2002), leading to a severe reduction of macrophyte abundance in lakes
where they have been introduced (Lodge and Lorman, 1987; Chambers et al., 1990; Nyström and
Strand, 1996; Gherardi and Acquistapace, 2007). Additionally, invasive crayfish may prevent the
recruitment of macrophytes as shown in rice fields and mesocosm studies (Matsuzaki et al., 2009).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
80
Therefore, invasive crayfish may potentially inhibit or prevent the return of macrophytes
when abiotic conditions for macrophyte growth have been restored, but their impact in lake
restoration projects remains untested. In The Netherlands, P. clarkii was first observed in 1985
(Koese and Soes, 2011) and has rapidly spread throughout the peat district in the west of the
country in the last decade (Figure 5.1). Many restoration projects have been executed to restore
the water transparency and promote the return of macrophytes in the shallow water bodies of
this peat district (Gulati and Van Donk, 2002; Lamers et al., 2002; Ter Heerdt and Hootsmans,
2007).
Figure 5.1 – Map of records of the exotic crayfish Procambarus clarkii in The Netherlands. The data are a combination of (muskrat) trapping surveys, netting surveys, and sightings of specimens migrating overland, n=1534 records. The study site is located at the lower tip of the black line.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
81
5
We hypothesize that invasive crayfish pose a novel constraint on the regeneration of submerged
macrophytes in lake restoration projects and may jeopardize restoration efforts.
The innovation of our study is that (1) we study the impact of crayfish in the field in an
additive design, using different mesh size exclosures to study the role of crayfish versus other
potential herbivores, and that (2) we study whether crayfish inhibit the return of macrophytes,
when abiotic conditions for growth seem favourable. There has been documentation that water
birds and large fish may jeopardize restoration efforts (Lauridsen et al., 1993; Søndergaard et
al., 1996; Van Donk and Otte, 1996; Lauridsen et al., 2003; Irfanullah and Moss, 2004; Hilt,
2006; Van de Haterd and Ter Heerdt, 2007), but we are the first, to our knowledge, to show that
invasive crayfish may also threaten successful lake restoration, e.g. the return of macrophytes. We
show that invasive crayfish P. clarkii strongly inhibit macrophyte development once favourable
abiotic conditions for macrophyte growth are restored. We conclude that invasive crayfish
may compromise restoration measures and that the continuing expansion of invasive crayfish
populations throughout North-Western Europe poses a new threat to successful restoration of
clear water with abundant submerged vegetation.
MATERIAL AND METHODS
Ethics statement
The study was conducted on the terrain of Waternet. Waternet gave permission to work on their
property as well as to conduct this study. No further permits were required for the described
study, which complied with all relevant regulations. The study did not involve endangered or
protected species.
Study design
We experimentally tested the effect of the invasive crayfish P. clarkii on the development of
submerged macrophytes within a restored shallow peat lake in The Netherlands. We used two
enclosed lake sections, hereafter called ponds, where seemingly favourable abiotic conditions for
macrophyte growth were found. In situ enclosures and exclosures in both ponds allowed us to
investigate separate and combined effects of crayfish and native herbivores (fish and waterfowl)
on the growth of three introduced plants. We analysed diet composition of P. clarkii using gut
content analysis to determine whether they consumed the plants.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
82
Figure 5.2 – Overview of Lake Terra Nova and design of the cage-experiment. (a) Lake Terra Nova with ponds indicated in the black box. (b) Enlarged overview of the study ponds with the grazing treatments arranged in blocks within the iron pond (iron suppletion) and non-iron pond. (c) Legend ovf the grazing treatments applied. In the partial exclosure, mesh size was 5 cm height and 10 cm width to allow undisturbed access for large crayfish.
Study area
The experiment was conducted in the western part of Lake Terra Nova (52º13’N, 5º02’E), The
Netherlands (Figure 5.2). Lake Terra Nova is an 85 ha shallow peat lake in which different
restoration measures were taken in the past. The lake has a mean depth of 1.4 m and the bottom
is covered with a 0.9 m organic sediment layer. Until the early 1970’s, a highly developed
macrophyte community consisting of various Characeae and Potamogeton sp., covered the lake
bottom (Van de Haterd and Ter Heerdt, 2007). An increase in P loading was observed after 1977
and as a consequence the lake shifted from a clear macrophyte-dominated system to a turbid
algae-dominated system in which only floating and sparse submerged macrophytes remained
(Van de Haterd and Ter Heerdt, 2007). In 2003, biomanipulation was applied in which the
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
83
5
benthivorous sediment disturbing fish assemblage was reduced from 180 kg ha-1 to less than 25
kg ha-1 cyrpinid fish biomass, which resulted in clear water and the return of many macrophyte
species (Van de Haterd and Ter Heerdt, 2007). However, despite continued fishing keeping the
cyprinid fish at low biomass, the macrophyte revival was only brief and in 2010 most of the
lake contained bare sediment with scattered floating plant vegetation and turbid water through
summer algal blooms. Red swamp crayfish were first reported in 2006 in the lake area (Figure
5.1) and may have been present since the early 2000’s, but numbers have not been documented.
To test whether restoration measures would prevent algal blooms and stimulate the return
of submerged macrophytes, two ponds of approximately 0.5 ha each were constructed in the
western part of Lake Terra Nova in 2003 (Figure 5.2). In one pond FeCl3 was applied in 2009 to
reduce internal P loading (gradual addition over a period of 102 days to a total of 85 g Fe m-2).
However, in both ponds clear water conditions existed, whereas no submerged macrophytes were
observed in either pond in 2009 or 2010 prior to this study and only floating leaved species
(Nuphar lutea L. and Nymphaea alba L.) were present and Phragmites australis (Cav.) Trin. ex Steud.
was the dominant species along the shores. We counted and sampled the potential herbivores,
respectively water birds, fish, and crayfish in and around the ponds (see Table 5.1 and 5.2 for
methods, densities, and species of waterbirds and fish). Crayfish abundance was determined by
surveying both ponds simultaneously with 12 cylindrical crayfish traps (75 cm long, diameter 30
cm, 1.2 × 1.2 cm mesh) baited with cat food, which were checked every three days for five weeks
prior to the experiment. Crayfish were individually marked. Only two crayfish were recaptured;
numbers are therefore minimum number of crayfish present.
At the start and the end of the experiment we sampled environmental variables from the water
column and sediment in both ponds; see Appendix 5.1 for the methodology and Supplementary
Table 5.1 for the results.
Experimental set-up
To analyse the effect of different herbivores on the development of macrophytes we performed
an experiment in both ponds with four different grazing treatments: a full exclosure in which all
studied herbivores were excluded, a partial exclosure providing access to crayfish and small fish,
an enclosure, stocked with only crayfish, and a control where all herbivores had access to (Figure
5.2). Exclosures and enclosures consisted of cages of 1 m3 and were closed on all six sides, control
plots were 1 m2. The corners of each cage were fixed with bamboo poles in the sediment and the
control plots were marked with a pole. In each pond, each treatment was replicated seven times
following a randomized block design (Figure 5.2); plots within a block were 2 m apart from each
other. Each block of four treatments was placed randomly in the pond, but at least 15 m from the
nearest other treatment block at the start of the growing season in 2011 (April 18th 2011). Water
depth in the cages ranged between 0.7-0.9 m; none of the cages was completely submerged and
thus no algae were growing on the top, allowing maximum light availability inside the cage.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
84
Since no submerged macrophytes were present in the ponds, three species of submerged
macrophytes known to have occurred in Lake Terra Nova (Van de Haterd and Ter Heerdt, 2007)
were collected from nearby ponds and introduced. Two shoots of Chara virgata Kützing (mean
DW 0.54 g ± 0.02 SE), Elodea nuttallii (Planch.) St. John (0.12 ± 0.01 g), and Myriophyllum
spicatum L. (0.14 ± 0.02 g) were planted in separate square plastic pots (11 × 11 × 12 cm3; one
pot per species and two shoots per pot) filled with sediment originating from the pond where
they were subsequently planted. Two replicate pots of the three species were randomly mounted
on metal frames (50 × 50 cm2). These frames, thus containing a total of 6 pots each (2 replicates
× 3 species), were subsequently placed in each grazing treatment.
For the enclosure treatment, crayfish were caught with crayfish traps in Lake Terra Nova
at about 500 m distance from the ponds. Crayfish were placed in the enclosures on the day of
capture. At the start of the experiment, four adult crayfish were introduced in each enclosure (mean
biomass per crayfish 37.4 g ± 2.0 SE, Ntot
= 56, female:male ratio 1:1.7). The crayfish density in
the enclosures (150 g m-2 wet weight) approached the higher densities estimated for Lake Terra
Nova (up till 191 g m-2 wet weight; Van Giels, 2011). Crayfish densities vary widely in the field
and are reported to range from 0.8-13 individuals m-2 in the meta-analysis of Matsuzaki et al.
(2009), who use 140 g m-2 as a high density in their own experiments. Gherardi and Acquistapace
(2007) report 4 and 8 individuals m-2 as natural densities in Italy, whereas Rodriguez-Villafañe et
al. (2003) estimate a density of approximately 1 individual m-2 for a Spanish lake, although they
indicate that this is probably an underestimation of the real density.
Harvest
Six weeks later (May 31st 2011), when the canopy-forming species M. spicatum and E. nuttallii
had reached the water surface in a majority of the full exclosure plots, the plants were harvested.
Macrophytes from all treatments were harvested and transported to the lab, rinsed with running
fresh water, dried for 48 h at 60°C and weighed. Crayfish were collected from the enclosure cages
and frozen at -20°C for gut analysis.
Table 5.1 – Overview of densities of waterfowl around the ponds.
Waterfowla Individuals ha-1 Tufted duck (Aythya fuligula L.) 4.3Eurasian coot (Fulica atra L.) 2.9Common pochard (Aythya farina L.) 1.4Greylag goose (Anser anser L.) 1.4Gadwall (Anas strepera L.) 1.4Egyptian goose (Alopochen aegyptiacus L.) 0.7Mallard (Anas platyrhynchos L.) 0.7Mute swan (Cygnus olor Gmelin) 0.6
a Water birds present in the water in and around the ponds (an area encompassing 0.07 km2) were counted weekly in April and May 2011 using binoculars. Data are means of the weekly counts.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
85
5
Crayfish diet
Crayfish gut content analysis was performed on 41 individuals in total from the enclosures from
both ponds (22 from the iron pond and 19 from the non-iron pond) and 20 from the natural
population in the ponds (10 per pond) caught outside the treatment blocks at the end of the
experiment with the same traps used to estimate crayfish numbers (see Table 5.2). The crayfish
were dissected and the stomach was removed from each individual and subsequently washed out
to dilute the gut contents (Gilling et al., 2009). Food items (recorded as either present or absent
in each specimen) were identified to the nearest recognizable taxonomic level with a dissecting
microscope.
Table 5.2 – Numbers of fish caught in the study ponds.
Fish CPUEa Electro fishing (Individuals ha-1) Gill nets (Individuals m-1 net)
Non-iron
pond
Iron
pond
Fish length
range (cm)
Non-iron
pond
Iron
pond
Fish length
range (cm)
Rudd (Scardinius erythrophthalmus L.) 35 69 3-7 0 0.008 14
Perch (Perca fluviatilis L.) 2482 414 7-15 0.16 0.24 8-22
Ruffe (Gymnocephalus cernuus L.) 0 0 0.03 0.008 7-13
Pike (Esox Lucius L.) 69 0 30-74 0 0
Tench (Tinca tinca L.) 35 0 3 0.016 0 43-47
Roach (Rutilus rutilus Rafinesque) 2 0 4-6 0 0
a Fish catch per unit effort. Fish abundance in each of the ponds was determined on 25 and 26 October 2011. Shoreline abundance was determined by electrofishing (200 volt, 5 amp, 290 m shore line length sampled per pond, 1 m transect width). Open water fish abundance was determined by overnight placement of multi-mesh gill nets (10-110 mm; total length 75 m) and an additional gillnet (140 mm; length 50 m) and additionally for 2 hours during the day on 25 October.
Presence of plant propagules
To investigate whether the sediment of the ponds contained viable plant propagules, in total 25 L
of the upper 5 cm of the sediment from three random locations in each pond was collected during
the harvest of the transplants (on May 31st 2011). The pooled sediment sample of each pond was
taken to the lab and distributed over three 60 L aquaria, resulting in a ca. 3 cm sediment layer in
each aquarium. Aquaria were subsequently filled with tap water (15 cm depth), and placed in a
greenhouse at 20 °C under natural light conditions. Plants were allowed to emerge during 18 weeks
after which all plants that had emerged were counted and identified to species level.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
86
Data analysis
Survival and biomass data of the plants were analysed using R version 2.15.0 (R Development
Core Team, 2011). Since survival of transplants followed a binomial distribution, effects of
grazing treatment and pond on survival were analysed by fitting generalized linear mixed effect
models with pond, grazing treatment and their interaction as fixed factors and treatment block
and plant duplicate as random factors. The biomass of the plants (logarithmically transformed)
was analysed by fitting general linear mixed effect models. Models were fitted with the lmer
function in the lme4 package (Bates et al., 2011). To determine effects of fixed factors a likelihood
ratio test was used to compare models with and without the variable of interest (Crawley, 2007).
Post-hoc comparisons of means were made based on Tukey contrasts available in the multcomp
package. Assumptions of normality for general linear mixed models were checked by plotting
residuals and performing a Shapiro test on residuals.
RESULTS
Herbivore presence
The herbivores and omnivores present in and around the ponds were water birds, fish and crayfish
(Table 5.1 and 5.2). With respect to crayfish, only Procambarus clarkii was caught in the ponds. In
total 178 crayfish were caught in the non-iron pond and 66 in the iron pond, corresponding to
respectively 0.42 and 0.16 CPUE (individuals per trapnight, based on 12 traps and 35 nights in
each pond). Both ponds were characterized by low numbers of fish, predominantly existing of smaller
sized perch, although the non-iron pond also harboured some larger individuals of pike and tench
(Table 5.2). The biomass of benthivorous fish (rudd, ruffe, tench, and roach) amounts to 0.2 kg ha-1
averaged over both ponds (based on CPUE of electrofishing, weight data not shown).
Effect of herbivores on macrophyte development
Macrophyte growth and survival was significantly affected by grazing treatment (Figure 5.3; Table
5.3). Free herbivore access strongly reduced survival and growth of all three macrophytes, which
produced most biomass when fully protected from grazing (Figure 5.3; Table 5.3). Biomass of E.
nuttallii and C. virgata was strongly reduced in all three treatments with herbivores. Similarly,
biomass of M. spicatum was reduced in all treatments with herbivores in the iron pond, whereas in
the non-iron pond, biomass in the partial exclosure was intermediate and not significantly different
from the full exclosure or full enclosure and control (Figure 5.3; Table 5.3). The effect of grazing was
stronger in the iron pond compared to the non-iron pond for E. nuttallii and M. spicatum. Biomass of
M. spicatum was significantly higher in the full exclosures in the iron pond compared to the non-iron
pond, whereas there was a similar trend, but no statistical differences, for E. nuttallii and C. virgata
(Figure 5.3; Table 5.3). Survival of the macrophytes was similar in both ponds (Figure 5.3; Table 5.3).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
87
5
Figure 5.3 – Biomass and survival of transplanted macrophytes under different grazing treatments. Mean biomass (left panels) and survival (right panels) of C. virgata (a, b), E. nuttallii (c, d), and M. spicatum (e, f) transplants at the end of the experiment for the non-iron and iron pond. Different letters or numbers in biomass panels indicate significant differences between treatments for the iron pond and non-iron pond respectively (Tukey post hoc comparisons, P < 0.050). Significant differences in transplant biomass between ponds within a single treatment were found for Elodea biomass in the partial exclosure and for Myriophyllum biomass in the full exclosure and are indicated by asterisks (Tukey post-hoc comparisons, * P < 0.050; ** P < 0.01; ***P < 0.001). For the survival panels, different letters indicate significant differences between treatments only. See Table 5.3 for results of the statistical analyses.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
88
There was some mortality of crayfish in the enclosures, which had reduced stocked crayfish
biomass in the enclosures whereas the surviving crayfish were growing, resulting in a final mean
biomass per enclosure of 151.1 g ± 17.7 SE in the iron pond, and 132.4 g ± 9.4 in the non-iron
pond which was not significantly different (t-test, df = 12, t = 0.932, P = 0.370).
Table 5.3 – Effect of pond (iron and non-iron) and grazing treatment on biomass and survival of three transplanted macrophyte species.
Pond Grazing treatment Pond × Grazing
Parameter Species Χ2 df P Χ2 df P Χ2 df P
Biomass E. nuttallii 11.56 4 0.021 95.77 6 <0.001 11.21 3 0.011
M. spicatum 33.60 4 <0.001 91.84 6 <0.001 25.29 3 <0.001
C. virgata 6.77 4 0.149 51.40 5 <0.001 5.55 3 0.136
Survival E. nuttallii 7.17 4 0.127 27.74 6 <0.001 5.24 3 0.155
M. spicatum 55.78 4 0.233 58.88 6 <0.001 5.32 3 0.150
C. virgata 16.61 4 0.002 34.17 6 <0.001 5.42 3 0.143
Results (likelihood ratio tests) of general linear mixed effects models (biomass) and generalized linear mixed effect models (survival) per macrophyte species (see also Figure 5.3). Df – degrees of freedom.
Germination of propagules
Each plot was checked for naturally emerging macrophytes in the field, but none were found
on 31 May, after 6 weeks of exclosure treatments. Germination in the greenhouse showed that
the sediment of both ponds contained viable propagules of macrophytes. Forty-eight individual
macrophytes germinated from the sediment of both ponds combined, representing 8 species. In
the sediment from the non-iron pond we found Chara globularis (3 individuals), Myriophyllum
spicatum (4), and Tolypella prolifera (Ziz ex A.Braun) Leonhardi (1), in the iron pond Potamogeton
pusillus L. (1) as submerged species. Nuphar lutea (L.) Sm. was the only floating species and was
found in both ponds (5 individuals in total). The emergent species were more abundant: Typha
angustifolia L. (19), Juncus articulatus L. (4), and Lythrum salicaria L. (7), all species found in both
ponds.
Crayfish diet
Gut content analysis of the crayfish in the enclosures showed that the percentage of crayfish with
animal remains in their stomach was considerably larger than the percentage of crayfish with
vegetal remains in their stomach, whereas the majority of the free-living crayfish in the ponds had
both animal as well as vegetal remains in their stomach (Table 5.4).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
89
5
Table 5.4 – Occurrence of food items (presence – absence) in crayfish guts from individuals collected from the full enclosures (n=41) and the natural population in the field (n=20), at the end of the experiment.
Identified food item: Crayfish in enclosures Free-living crayfish
Detritus 44 70
Remains of higher plants 17 75
Remains of filamentous algae 20 50
Diptera larvae 51 10
Crustacea 51 30
Gastropoda 5 10
Hydrachnidia 10 20
Protozoa - Amoeba 7 45
Unknown animal remains 22 55
Unknown remains 44 0
Subtotals
Animal remains 66 80
Vegetal remains 34 85
Data show the percentage of crayfish (in relation to the total number of dissected individuals) for which the given food item was present in the stomach.
Environmental conditions
The abiotic conditions were very similar in both ponds (see Supplementary Table 5.1). The iron
pond had a higher attenuation of light, despite lower chlorophyll-a concentration, but in both
ponds there was on average more than 15% of ambient light available at the bottom. The iron
pond had a significantly higher Fe concentration in the surface water and sediment and a higher
sediment P concentration. P and PO4 in the water column were higher at the start but lower at
the end of the experiment, whereas NO3 was lower at the start and higher at the end in the iron
pond compared to the non-iron pond respectively (Supplementary Table 5.1).
DISCUSSION
Invasive crayfish P. clarkii can inhibit the development (growth and survival) of submerged
macrophytes while abiotic conditions for macrophyte growth were favourable as demonstrated in
our experiment. Survival and biomass of the three submerged macrophytes was significantly lower
when crayfish were present, whereas the plant species grew well in both study ponds when they
were protected from crayfish and other herbivores. When protected from grazing, Myriophyllum
grew better in the iron pond, but there was no significant difference for the other species.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
90
The establishment of the ponds as lake enclosures may have provided enough shelter from the
wind to prevent sediment resuspension and allow clear water conditions (Van de Haterd and Ter
Heerdt, 2007) regardless of iron addition, whereas differences among the ponds may have been
present before the iron addition as well. We conclude that in both ponds, the light availability
was with more than 15% of ambient light on the lake bottom (and often much more) above the
minimum light requirements for growth of caulescent submerged angiosperms and charophytes
(Middleboe and Markager, 1997) and therefore abiotic conditions were suitable for macrophyte
growth in both ponds. During our experiment, we did not observe naturally emerging vegetation,
which may perhaps be due to the short term (6 weeks) or early season (April-May) in which we
performed the experiment. The presence of viable propagules of several submerged species in the
sediment suggests that the absence of submerged vegetation in the entire ponds is not due to
a lack of propagules per se. We therefore further focus on the role of invasive crayfish and their
potential to inhibit macrophyte growth and development once favourable abiotic conditions for
growth have been created.
Whereas invasive crayfish are known to reduce macrophyte abundance in Southern and
Northern Europe (Nyström et al., 1999; Rodríguez-Villafañe et al., 2003; Gherardi and
Acquistapace, 2007) and inhibit propagule establishment in mesocosms (Matsuzaki et al., 2009),
their impact on macrophyte establishment in field restoration projects has not yet been tested
to our knowledge. We show that invasive crayfish may present a new bottleneck for macrophyte
development in North-Western European waters when abiotic conditions for macrophyte growth
are restored. In North-Western Europe, many lake restoration projects have been executed and are
still being implemented, aimed at improving water transparency and development of abundant
macrophyte vegetation (Moss, 1989; Jeppesen et al., 2005; Hilt et al., 2006; Søndergaard et al.,
2007; Gulati et al., 2008). Our results suggest that these projects may face a new constraint with
the increasing spread of invasive crayfish, particularly P. clarkii.
Effects of crayfish versus other potential herbivores
The enclosure treatments with only crayfish present showed that crayfish strongly reduced
survival and growth of submerged macrophytes. Furthermore, the very small differences between
the enclosure treatment (access for crayfish only) and the partial exclosure (access for crayfish and
small fish) and the control treatment (access for all herbivores) indicate strong effects of crayfish
and no significant additive effects of waterfowl and larger fish. Smaller fish that could enter the
partial exclosures were present in the study ponds. Technically, very small fish could even have
entered the full exclosure or crayfish enclosure with the mesh size of 1 × 1 cm and reduce plant
growth. However, this would have led to reduced growth of the macrophytes in the full exclosure,
whereas we observed a much higher plant growth in the full exclosure compared to the treatments
where larger herbivores had access. Therefore, if very small fish did enter the full exclosure, we
estimate their impact on plant growth to be very small. Small fish may have entered the partial
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
91
5
exclosure, in which the mesh was oriented such that it was 10 cm wide and 5 cm in height (to
allow optimal access for large crayfish, which are wider than tall due to their claws). However,
the density of fish in the study ponds was generally very low and most fish were not herbivorous.
Of the fish that include macrophytes in their diet, e.g. rudd and tench, the smaller size classes
are mostly carnivorous (Lake et al., 2002; Nurminen et al., 2003) and even the large fish of these
species preferentially feed on macrofauna under temperate conditions, as demonstrated for rudd
(Dorenbosch and Bakker, 2011, 2012). When feeding on invertebrates, fish may inadvertently
ingest the macrophyte leaves which have macrofauna on them. Smaller roach (of 7 cm and larger)
for instance have been observed to pluck macrophyte leaves when consuming macro-invertebrates
on the leaves, although they mostly do so when zooplankton and other food sources are scarce
(Körner and Dugdale, 2003). This is in line with observations in a Finnish lake, where in spring,
when zooplankton is abundant, small (<10 cm) rudd does not ingest plant material and only
larger rudd consumed plants (Nurminen et al., 2003). Furthermore, significant effects of plant
plucking on macrophyte growth were observed in Lake Müggelsee at a fish biomass of >150 kg
ha-1 of which 70-80% consisted of bream and roach (Körner and Dugdale, 2003; Hilt, 2006). In
contrast, fish density in our study ponds was much lower with 0.2 kg ha-1 for benthivorous fish,
estimated from the electrofishing CPUE. Previous removal of benthivorous fish in our study lake
showed that a reduction from 180 to <25 kg ha-1 biomass of cyprinid fish, resulted in strong
growth of submerged macrophytes (Van de Haterd and Ter Heerdt, 2007). Therefore, whereas
we cannot entirely exclude that small fish may have had an additional impact on macrophyte
growth in our study, a large part of the difference in plant growth among the partial exclosure and
crayfish enclosure versus the full exclosure is likely caused by crayfish considering the low density
and diet preferences of small fish and the high crayfish density. Whereas it was known that
grazing by water birds or fish can be a limiting factor in the appearance of submerged vegetation
(Lauridsen et al., 1993; Søndergaard et al., 1996; Van Donk and Otte, 1996; Lauridsen et al.,
2003; Irfanullah and Moss, 2004; Hilt, 2006; Van de Haterd and Ter Heerdt, 2007), we now show
that the presence of crayfish can inhibit the establishment of submerged macrophytes in a lake
restoration project. The absence of an additional effect of water birds and large fish demonstrates
that crayfish alone are potentially able to prevent restoration of submerged vegetation.
Crayfish grazing versus bioturbation
It is often unclear whether observed crayfish impact on macrophytes is caused by herbivory or
bioturbation (Matsuzaki et al., 2009). In our study, gut content analysis showed that P. clarkii
had an omnivorous diet, with animal and plant material and detritus found equally often in
free living crayfish. The gut of the crayfish in the enclosures contained more frequently animal
material. This may be due to the fact that most plant material had already been consumed at
the end of the experiment and thus was no longer available. These results agree with previous
studies that showed crayfish to be omnivorous (Nyström et al., 1999; Correia, 2002; Körner and
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
92
Dugdale, 2003; Gherardi, 2006). In our study the crayfish did consume macrophytes and thus at
least part of their impact on macrophytes was due to herbivory. However, we cannot exclude that
part of the observed effects of crayfish may also be due to bioturbation, particularly destruction or
uprooting of the planted macrophytes (Anastácio et al., 2005).
Effect of crayfish during lake restoration
Invasive crayfish may reduce macrophyte abundance and induce a shift to a turbid, algae
dominated, state of the ecosystem (Rodríguez-Villafañe et al., 2003; Geiger et al., 2005). The
goal of many restoration projects is to reverse a turbid state into a clear water state dominated
by submerged macrophytes (Scheffer, 2001). Once appropriate measures have been taken
macrophytes may return, when propagules are available (Lauridsen et al., 2003; Hilt et al., 2006;
Bakker et al., 2013). The question is to what extent invasive crayfish may inhibit the return of
submerged macrophytes and therefore compromise restoration efforts. The impact of crayfish on
the establishment and development of submerged macrophytes is potentially large as they live on
the sediment, which is where macrophytes emerge from propagules. Crayfish have been shown
to strongly suppress macrophyte establishment from a propagule bank in mesocosm studies
(Matsuzaki et al., 2009). Contrary to herbivorous waterfowl, which are frequently mentioned
as consumers of establishing macrophytes (Søndergaard et al., 1996; Lauridsen et al., 2003;
Irfanullah and Moss, 2004), crayfish stay in a lake year round and are able to feed on alternative
sources like detritus (Momot, 1995) on which they can sustain themselves when macrophytes are
absent (Grey and Jackson, 2012). As a result, crayfish density will not be strongly coupled to the
availability of macrophytes in lakes with organic sediments, such as our study lake. Therefore,
grazing pressure on macrophytes is potentially high, particularly when predation on the crayfish
is low, for instance when fish densities are low due to biomanipulation, as is the case in our study
lake (Ter Heerdt and Hootsmans, 2007).
Species invasions in general occur more often in disturbed situations (Hobbs and Huenneke,
1992) where exotic species can opportunistically invade (temporarily) empty niches (Jackson et
al., 2012). Procambarus clarkii is an opportunistic species due to its omnivorous feeding habits and
semi-amphibious life style (Gheradi, 2006; Grey and Jackson, 2012). Possibly lake restoration
projects are more prone to colonization by invasive crayfish, but to our knowledge, this has not
been investigated.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
93
5
CONCLUSIONS
We conclude that P. clarkii strongly reduced the biomass development and survival of establishing
macrophytes. Invasive crayfish may form a new constraint on the development of submerged
aquatic vegetation when abiotic conditions for macrophyte growth are improved. Invasive crayfish
may compromise restoration measures and pose a new threat to successful restoration of clear water
with abundant submerged vegetation. The continuing expansion of invasive crayfish populations
throughout North-Western Europe is worrying. Strong emphasis should be put on prevention of
introduction and where possible spread of the crayfish, since removal or management of invasive
crayfish populations is very difficult (Gherardi et al., 2011).
ACKNOWLEDGMENTS
Gerard ter Heerdt from Waternet provided access to the study site, water quality data and a boat.
Ivan Mettrop collected data in the iron suppletion project. ATKB provided fish data. Thijs de
Boer and Koos Swart helped collecting sediment, building cages, and sampling.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
94
APPENDIX 5.1
Methods Chapter 5. Collection of environmental and chemical variables
Environmental and chemical variables were measured in both ponds at the start and the end
of the experiment (Supplementary Table 5.1). Surface water samples from 10 cm below the
water surface were collected in 500 mL polyethylene bottles. Sediment pore water was collected
anaerobically using 60 mL vacuum syringes connected to ceramic soil moisture samplers
(Eijkelkamp Agrisearch Equipment, Giesbeek, The Netherlands), which were installed in the
upper 10 cm of the sediment. The first 10 mL was discarded to enable anaerobic sampling. The
pH of the water samples was measured using a combined pH electrode with an Ag/AgCl internal
reference (Orion Research, Beverly, CA, USA), and a TIM800 pH meter. Subsequently, surface
water samples were filtered through glass microfiber filters (type GF/C, Whatman, Brentford,
UK). The samples were stored in polyethylene bottles at -20 °C until further analyses.
Additionally to nutrients, water transparency (by absorption at 750 nm in a Helios delta
photospectrometer, Unicam, Cambridge, UK) and chlorophyll-a concentrations (in a PhytoPAM
phytoplankton Analyser, Heinz Walz GmbH, Effeltrich, Germany) of the surface water were
measured in 100 ml water samples (replicated three times), whereas light extinction in the water
column was measured at a depth of 60 cm (by a LI-CORLI-250 quantum photometer, LI-COR
Biosciences, Lincoln, NE, USA, replicated seven times).
Nutrients (Fe, S, P, organic P, and Olsen-P) in the pond sediments were only measured prior to
the transplant experiment. Samples of the upper sediment layer were taken with a multisampler
(Eijkelkamp Agrisearch Equipment, Giesbeek, The Netherlands), transported in airtight bags
and kept in the dark at 4°C until further analyses.
Homogenized portions of 5 g wet sediment were used to determine organic P concentrations
using a P-fractionation analysis according to Golterman (1996). The rest of the sediment was
dried for 48 hours at 70° C. Homogenized portions of 3 g dry sediment were used to determine
Olsen-P concentrations by extraction according to Olsen et al. (1954). Homogenized portions
of 200 mg dry sediment were digested with 4 mL HNO3 (65%) and 1 mL H
2O
2 (30%), using
an Ethos D microwave labstation (Milestone srl, Sorisole, Italy). Digestates were diluted and
concentrations of Fe, S, and P were determined by ICP.
The concentrations of PO4, NO
3, and NH
4 in surface water and sediment pore water were
measured colorimetrically with an Auto Analyser 3 system (Bran+Luebbe, Norderstedt, Germany)
according to Geurts et al. (2008). The concentrations of Fe, S, P, organic P, and Olsen-P were
measured using an ICP Spectrometer (IRIS Intrepid II, Thermo Electron Corporation, Franklin,
USA).
Differences in abiotic characteristics were analysed using general linear mixed effect models
(see methods in Chapter 5). Pond, sampling time and their interaction were fixed factors. The
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
95
5
factor sampling time was defined as a random slope and nested in the random factor sampling
location to account for repeated measurement correlations.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
96
Sup
ple
men
tary
Tab
le 5
.1 –
Abi
otic
cha
ract
eris
tics
(mea
n ±
SE
) of s
urfa
ce w
ater
, por
e w
ater
, and
the
sed
imen
t of
the
tw
o ex
peri
men
tal p
onds
(iro
n an
d no
n-ir
on).
Pla
ce o
f sam
plin
gV
aria
ble
nP
ond
At
the
star
tA
t th
e en
d (6
wee
ks la
ter)
Mod
el r
esul
ts
Surf
ace
wat
erM
ean
SEM
ean
SEP
ond
Tim
eP
ond
×
Tim
e
pH4
Iron
8.25
0.49
7.19
0.02
NS
***
NS
4
Non
-iro
n8.
460.
027.
200.
02
Fe (µ
mol
L-1)
4Ir
on18
.71
3.93
28.7
53.
44**
**
NS
4
Non
-iro
n2.
420.
044.
300.
37
P (µ
mol
L-1)
4Ir
on1.
810.
091.
240.
09**
****
*
4
Non
-iro
n1.
220.
041.
500.
16
S (µ
mol
L-1)
4Ir
on94
.13
2.74
49.6
88.
35N
S**
*N
S
4
Non
-iro
n98
.14
0.29
48.0
01.
15
PO
4 (µm
ol L
-1)
4Ir
on0.
540.
130.
230.
05*
***
4
Non
-iro
n0.
260.
020.
450.
10
NO
3 (µm
ol L
-1)
4Ir
on3.
853.
287.
074.
15**
***
***
4
Non
-iro
n16
.52
0.68
0.00
0.00
NH
4 (µm
ol L
-1)
4Ir
on4.
852.
4414
.91
13.1
1N
SN
SN
S
4
Non
-iro
n5.
571.
2428
.08
15.6
1
Chl
orop
hyll
-a (µ
g L-1
)3
Iron
0.74
0.07
0.80
0.13
**N
SN
S
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Invasive crayfish threaten the development of submerged macrophytes in lake restoration
97
5
3
Non
-iro
n1.
330.
175.
400.
14
Ligh
t ex
tinc
tion
(%)
7Ir
on61
.52
2.43
84.8
20.
73**
***
***
*
7
Non
-iro
n52
.47
2.20
55.6
32.
17
Wat
er t
rans
para
ncy
(750
nm
)3
Iron
0.02
0.00
0.02
0.00
NS
NS
NS
3
Non
-iro
n0.
030.
000.
050.
01
Por
e w
ater
pH6
Iron
6.60
0.06
6.57
0.03
*N
SN
S
6
Non
-iro
n6.
700.
026.
770.
07
Fe (µ
mol
L-1)
6Ir
on92
.43
18.7
510
2.20
17.0
1N
SN
SN
S
6
Non
-iro
n83
.60
11.3
871
.15
13.9
4
P (µ
mol
L-1)
6Ir
on39
.37
12.3
544
.95
12.8
1N
SN
SN
S
6
Non
-iro
n28
.09
4.84
26.4
36.
13
S (µ
mol
L-1)
6Ir
on25
.78
7.03
18.4
90.
64*
NS
NS
6
Non
-iro
n14
.34
1.01
14.4
61.
05
PO
4 (µm
ol L
-1)
6Ir
on10
.46
4.47
6.38
1.70
NS
NS
NS
6
Non
-iro
n5.
652.
164.
321.
36
NO
3 (µm
ol L
-1)
6Ir
on1.
500.
442.
661.
17N
SN
SN
S
6
Non
-iro
n1.
180.
130.
440.
15
NH
4 (µm
ol L
-1)
6Ir
on10
11.8
037
6.79
1081
.04
354.
92N
SN
SN
S
6
Non
-iro
n87
8.16
146.
7981
0.22
170.
75
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 5
98
Sedi
men
t
Fe (µ
mol
L-1)
6Ir
on41
5.59
36.5
2
**
6
Non
-iro
n27
1.67
13.0
3
P (µ
mol
L-1)
6Ir
on37
.59
2.33
*
6
Non
-iro
n31
.13
1.32
S (µ
mol
L-1)
6Ir
on51
2.19
28.8
2
N
S
6
Non
-iro
n48
5.80
20.9
0
Org
anic
-P (µ
mol
L-1)
6Ir
on16
.75
1.95
NS
6
Non
-iro
n17
.79
2.13
P-O
lsen
(µm
ol L
-1)
6Ir
on9.
611.
06
**
*
6
Non
-iro
n3.
520.
98
For
each
var
iabl
e, r
esul
ts (l
ikel
ihoo
d ra
tio
test
s) o
f gen
eral
line
ar m
ixed
mod
els
are
repo
rted
in t
he c
olum
ns ‘p
ond’
, ‘ti
me’
, and
‘pon
d ×
tim
e’. S
edim
ent
vari
able
s w
ere
only
col
lect
ed o
nce.
NS:
P >
0.0
50; *
P <
0.0
50; *
* P
< 0
.010
; ***
P <
0.0
01, n
= n
umbe
r of
sam
ples
tak
en.
CHAPTER 6
Iron addition and biomanipulation as
complementary measures for the restoration
of a shallow peaty lake
Anne K. Immers, Elisabeth S. Bakker, Ellen van Donk,
Gerard N. J. ter Heerdt, and Steven A. J. Declerck
Submitted to Ecological Engineering
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
100
ABSTRACT
Abatement of external phosphorus (P) loading and biomanipulation are measures that are
often applied with the aim to restore a macrophyte dominated clearwater state in turbid,
anthropogenically eutrophied lakes. The recovery of such lakes, however, is often hampered by
‘internal eutrophication’, as a result of the release of historically accumulated P from the sediment
into the water column. One way to combat this internal P loading is by adding iron (Fe) into
the lake, which naturally binds to phosphate. Although studied in the laboratory or mesocosms,
the effects of iron addition on a whole-lake scale are largely unknown. In this study we therefore
compiled long-term lake monitoring data to evaluate the effect of a gradual dose of 33 g Fe m-2
on the water quality and biotic communities (phytoplankton, zooplankton, and macrophytes) of
Lake Terra Nova. Lake Terra Nova is a eutrophied, shallow peaty lake that has been subjected
to biomanipulation measures for nearly 10 years. Despite an initial success of biomanipulation,
continued fish removal efforts did not reduce the high phosphate concentrations in the lake.
As a consequence, yearly summer blooms of cyanobacteria re-occurred soon after the initiation
of biomanipulation. The combination of biomanipulation with large scale addition of iron,
however, resulted in a substantial reduction of lake concentrations of TP, suspended matter (SM)
and cyanobacterial biomass. The decrease in cyanobacterial biomass was coincided by an increase
in macrophyte coverage, which remained abundant until the end of the study period. However,
two years after the onset of iron addition, lake TP concentrations slowly started to increase again.
This increase might indicate that the reservoir of surplus iron, which ideally should form a buffer
against sediment P release, is increasingly bonding with dissolved organic carbon (DOC) in this
highly organic lake. Addition of iron might therefore be even more effective when applied to DOC
poor lakes. Our results show that the combination of both restoration measures, biomanipulation
and iron addition, can be an effective tool to restore lakes which suffer from internal P loading.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
101
6
INTRODUCTION
Eutrophication in fresh water lakes has been a major environmental problem all over the world
due to high anthropogenic input of phosphorus (P) and nitrogen (N) from agriculture, industry,
and waste water during the last century (Davis et al., 2010). This high input of nutrients impacted
aquatic ecosystems by enhancing lake productivity, leading to a change in community structure
and a decrease in light availability and biodiversity of these ecosystems (Davis et al., 2010).
During the last few decades, however, measures are taken in order to restore aquatic ecosystems.
Great efforts have been made ever since, largely by reducing external input of nutrients by either
closing off nutrient rich input sources or by pre-treating nutrient rich water before it entered the
lakes (Jeppesen et al., 2007). Whereas this has led to considerable improvements of water quality
(Jeppesen et al., 2007), a full recovery has not yet been reached in many cases, through internal
loading from nutrients that have been building up in the lake sediment (Cooke et al., 1993a;
Smolders et al., 2006).
Various restoration measures have been applied to lakes in order to tackle this delay in recovery,
including managing foodweb dynamics through removal of bioturbating and zooplanktivorous
fish (e.g. biomanipulation; Meijer et al., 1994; Søndergaard et al., 2007) and by adding one, or
various, chemical P binding agents in the lake (Cooke et al., 1993a; Burley et al., 2001; Smolders
et al., 2006). Biomanipulation has been highly successful in shifting turbid lakes to the clear water
state (Meijer et al., 1994; Jeppesen et al., 2012). However, the effect is often short-lived and lakes
return to the turbid state when the external or internal nutrient loading is not simultaneously
reduced by additional measures (Gulati and Van Donk, 2002; Søndergaard et al., 2007; Jeppesen
et al., 2012). Indeed, low P concentrations are a necessity for long-term biomanipulation success
(Meijer et al., 1994; Hansson et al., 1998; Jeppesen et al., 2012). By adding chemical P binding
agents to reduce internal P loading, P is chemically precipitated from the water column and
P sorption of the lake sediment is enhanced. This is often done through an addition of either
iron (Fe) or other chemical P binding agents such as aluminium (Al), calcium (Ca) or lime, or
lanthanum-enriched benthonite clay (Phoslock®) to the water column or sediment (Burley et
al., 2001; Smolders et al., 2006; Lürling and Van Oosterhout, 2013). The long-term effects and
potential consequences of these chemical additions have been described for case studies with Al
and lime (Cooke et al., 1993a; Burley et al., 2001). The restoration success of iron addition has
long been debated (see Cooke et al., 1993a), because iron is redox sensitive and the formed bond
with phosphate is expected to become unstable under anoxic conditions. However, Kleeberg et al.
(2013) showed that the success of iron addition is not hindered by this redox sensitivity of iron
and P can be efficiently precipitated independent of the nature of the oxygen supply.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
102
Adding Fe to a lake is not as artificial as it seems, as under natural conditions, Fe can seep in
the sediment of lakes via upwelling. However, due to regional changes in hydrological regimes
and desiccation through excessive extraction of groundwater this iron-rich seepage has decreased
(Van der Welle et al., 2007b). Therefore, re-adding a natural component in a lake could be a
more favourable method than adding substances which are not commonly found in lakes, such as
lanthanum-enriched benthonite clay. Various mesocosm experiments have shown that addition
of Fe to lake sediments may indeed lower total phosphorus (TP) concentrations in the water
column (Boers et al., 1994; Burley et al., 2001; Van der Welle et al., 2007b). Moreover, possible
undesirable effects of iron dosing on macrophytes, such as iron toxicity or a decrease in pH
remained absent when tested under experimental conditions (Immers et al., 2013, 2014). A field
experiment in which ironchloride (FeCl3) was added to Lake Vogelenzang indeed resulted in a
decrease in TP while no undesirable ecological side effects to the lake ecosystem were observed
(Boers et al., 1994). The longevity of the restoration measure was however cut short to a mere
three months due to both the low water retention time of this particular lake and high external
P loading from nearby rivers (Boers et al., 1994). Therefore, knowledge on the effects of iron
addition on aquatic ecosystems at the whole lake scale remains largely unknown.
In this study we evaluated the effect of iron addition on the aquatic community composition
as a complementary restoration measure to biomanipulation. Terra Nova is a eutrophied peaty
lake in The Netherlands with pronounced cyanobacterial blooms during summer months, even
after reductions in external nutrient loading (Hofstra and Van Liere, 1992). To shift the lake
from a turbid to clear water state from 2003 onwards, biomanipulation has been performed on
a yearly basis, involving the removal of both benthi- and planktivorous fish (Ter Heerdt and
Hootsmans, 2007). Whereas the first year of biomanipulation was characterized by increased
water transparency and macrophyte coverage, the system quickly deteriorated to pre-restoration
conditions, likely as a result of high P concentrations in the sediment. In order to tackle this
high internal loading of P, the biomanipulation efforts were complemented with the addition of
P binding iron. We expected that the addition of iron would lower the available P in the system,
resulting in increased water transparency and expanding macrophyte coverage.
MATERIAL AND METHODS
Study area
Lake Terra Nova (52°13’N, 5°02’E) is a shallow peaty lake located in the centre of The Netherlands,
with a lake surface area of 85 ha, a mean depth of 1.4 m and a bottom covered with a 0.9 m
organic sediment layer. As the lake is shallow, the water is well-mixed throughout the seasons
and stratification only occurs during ice cover. The lake originated due to peat excavation from
the 17th to the 18th century, and over time a highly diverse macrophyte community developed
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
103
6
(Hofstra and Van Liere, 1992). During the second half of the 20th century the trophic status of
the lake changed from mesotrophic to eutrophic due to the input of nutrient rich river water
and runoff from nearby agricultural fields. Consequently, macrophyte abundance decreased
considerably and phytoplankton blooms started to occur frequently during summer periods (Ter
Heerdt and Hootsmans, 2007).
Restoration Measures
Biomanipulation
In an attempt to restore aquatic macrophyte vegetation and shift the lake from a turbid to clear
water state, biomanipulation was started in 2003 with the removal of a large part of the benthi-
and planktivorous fish (157.2 kg fish ha-1 in the first two years, mostly bream and roach). This
resulted in a pronounced increase of lake water transparency and during the following year
(2004), macrophytes reappeared and algal blooms remained absent during the summer months
(Ter Heerdt and Hootsmans, 2007). Despite sustained, yearly efforts to remove fish since then
(see Supplementary Table 6.1 for an overview of yearly removed fish biomass), this initial success
proved short term and during the following years macrophytes disappeared and phytoplankton
blooms dominated by toxic blue-green algae re-occurred again. A study by Brouwers and
Smolders (2006) suggested that these cyanobacterial blooms were caused mainly by high P
concentrations. By that time, phosphate concentrations originating from external sources were
low due to a modernization of the sewage system and dephosphatation of river inlet water. High
P concentrations in the lake can therefore be largely attributed to internal P loading. Indeed,
Brouwer and Smolders (2006) estimated the yearly P loading from the sediment to be as high as
0.1 g P m-2 y-1 (45.5% of total P loading; Supplementary Table 6.2).
Iron addition
Addition of iron(III)chloride (FeCl3) has often been suggested as a way to reduce internal P
loading in shallow lakes (Burley et al., 2001; Smolders et al., 2006; Kleeberg et al., 2013). In
a study of the lake, Brouwer and Smolders (2006) suggested that addition of iron would allow
reduction of P-concentrations in the water column. They expected also that the surplus of iron
would form a protective P barrier on the surface of the sediment, which would be gradually mixed
into the deeper layers of the sediment by aerobic bottom dwellers. From the start of May 2010
till the end of August 2011, a mobile pump (Figure 6.1a) gradually supplied 203 tonnes of FeCl3
(33 g Fe m-2) to the lake over the course of a 1.5 year period. As this pump was driven by a wind
mill, daily quantities of iron that were added depended on prevailing wind speeds. By doing so,
local build-up of high FeCl3 concentrations on low-wind days could be avoided, preventing acute
exposure of biota to high levels of the added chemicals. At the end of the addition period, molar
iron:phosphate ratios in the sediments of most areas reached 10, which is shown to be sufficient
to reduce P mobilisation from the sediment (Geurts et al., 2008).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
104
Sampling procedures
We compiled already existing data from the Dutch water board Waternet in order to evaluate
changes in water chemistry and plankton species composition in response to biomanipulation
and Fe-addition. From 1986 onwards, Lake Terra Nova water chemistry has been monitored on 9
different locations on a biweekly or monthly basis. From 1996 onwards, biweekly phytoplankton
and zooplankton samples have been collected on these locations. The sampling point at the centre
of the lake (Figure 6.1b) consisted of the longest and most complete data set. A comparison of
this location with short term monitoring data from other locations in the lake (phytoplankton,
zooplankton, chemistry), showed that the location was representative for the whole lake, and data
from this location was therefore used for the different data analyses.
Figure 6.1 – (a) Iron addition (wind powered pump) and (b) Lake Terra Nova with the location of the central sampling site (red ×).
During each sampling day, water measurements were performed, including oxygen
concentration (HQ30D flexi with a LDO101 probe, Hach, Tiel, The Netherlands), pH (HQ30D
flexi with a WTW SenTix 41 probe, Hach, Tiel, The Netherlands), and temperature (TLC 1598,
Ebro, Ingolstadt, Germany). Depth-integrated samples (a total of 20 L) for both chemical and
biological analysis were collected with a polyethylene tube of 1 m length and a volume of 2 L.
Zooplankton samples were divided in two size fractions by subsequently filtering the lake water
over 50 and 30 µm filters. Subsequently, phytoplankton and zooplankton samples were preserved
with Lugol solution and afterwards samples were counted with an inverted microscope (DMI
4000B, Leica Microsystems b.v., Münster, Germany) and a stereomiscroscope (MZ 16, Leica
Microsystems b.v., Münster, Germany), respectively. For each genus, 25 body size measurements
were performed, which were used to calculate population biovolumes. In the period after iron
addition (2012-2013), only large zooplankton (copepoda and cladocera) were counted in 2013.
Macrofauna was collected from both the water column and bottom substrates on five different
locations in Lake Terra Nova during the spring (May) and summer (August) of the years 2008,
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
105
6
2009, and 2011. Water column samples were collected by sweeping a handnet (30 × 20 cm, 0.5
mm mesh size) over a range of 5 m along the lake shore. Sediment samples were collected using
a bottom grab (2 L; Van Veen, Eijkelkamp, Giesbeek, The Netherlands). During the sediment
sampling, plants, rocks, and other substrates were also collected and carefully checked for
macrofauna. Samples were subsequently pooled and organisms were identified to their respective
genus.
Water samples were filtered over a 1.2 µm Whatman GF/C filter (Whatman, Brentford,
UK) and the filters and filtrate were stored at -20 ºC until further analysis. Concentrations
of PO4, NO
3, NH
4, and NO
2 were colorimetrically determined with a QuAAtro CFA flow
analyser (Seal Analytical, Beun de Ronde, Abcoude, The Netherlands). Chloride was measured
spectrophotometrically (Aquakem 250, Thermo Fisher Scientific, Waltham, MA, USA) with
extinction at 480 nm. Analyses of Fe and SO4 were performed using an inductively coupled
plasma emission spectrophotometer (ICP; Liberty 2, Varian, Bergen op Zoom, The Netherlands)
according to the Dutch NEN-EN-ISO 17294 to estimate dissolved Fe and S in the water column.
To determine the organic N fraction, filtered field samples were analysed with a FLASH 2000
Organic Elemental Analyser (Interscience, Breda, The Netherlands). TP concentrations were
determined by incinerating filtered field samples for 30 minutes at 500 °C, followed by digestion
in H2O
2, before analysis with a QuAAtro CFA flow analyser.
Phytoplankton chlorophyll (total) from stored filters was extracted in 80% ethanol (according
to the Dutch NEN 6520 protocol) and was measured spectrophotometrically on a UV/Vis Thermo
spectophotometer (UV3, Unicam Instruments, Cambridge, England) at 665 nm with a turbidity
correction conducted at 750 nm. The concentration of chlorophyll was determined using the
calibration equation from Lorenzen (1967). Suspended matter (SM) dry weight was measured by
filtering 1 L of field sample over a prewashed and preweighed Whatman GF/A filter (Whatman,
Brentford, UK). Subsequently, filters were dried for 24 hours at 60 °C and afterwards weighed to
determine the total dry weight.
From 2004 onwards (except for the years 2010 and 2012) the submerged macrophyte
vegetation was monitored each summer (July/August) on at least 43 locations representing most
of the surface area of the lake. At each of these sites, dominant species were assessed on a visual
basis using a hydroscope. Additional rake sampling was done at turbid sites to complement
visual recordings of coverage and species identification. Macrophyte coverage was estimated as a
percentage of sampling sites with macrophytes.
Data analyses
We distinguished 5 different time periods: I, before biomanipulation (1986 – 2003); II, the first
year of biomanipulation (2004); III, the years between the onset of biomanipulation and the start
of iron addition (2005 – 2009); IV, during iron addition (2010 – 2011); and V, after iron addition
(2012 – 2013).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
106
To visualize the long term trends in chemical and biological variables as well as their responses
to the start of biomanipulation and Fe-addition measures, we performed principal components
analyses (PCA) using data from 2001 onwards. One analysis was done using spring data (April,
May, and June), whereas the second analysis was performed to characterise summer conditions
(July, August, and September). Input data for these analyses were, respectively, spring and
summer means of the variables Fe, TP, TN, SM, pH, and total phytoplankton chlorophyll as well
as biovolumes of cyanophytes, chlorophytes, diatoms, copepods, Daphnia, and the remaining
cladoceran zooplankton. The PCA was carried out on standardized and log-transformed data
(both species and environmental data) using Canoco v. 4.5 (microcomputer Power, Ithaca, USA),
with the different time periods and Fe as supplementary environmental variables.
RESULTS
Overall effects of biomanipulation and iron addition on water quality and plankton dynamics
Average spring water quality and plankton biovolume changed markedly between the subsequent
time periods (Figure 6.2a). The first PCA axis, which represents 44% of the total variation, shows
strong negative associations with suspended matter, chlorophyll, and the biovolume of each of the
major phytoplankton functional groups. The different restoration periods show a shift from the
left to the right side of the first axis, indicating a shift from high to low levels of phytoplankton
biovolume, suspended matter, and TP in response to the onset of biomanipulation. This trend
was reinforced by the start of the Fe-addition treatment. The second axis, which accounts for 17%
of the variation, mainly tends to represent the first-year response of the lake to biomanipulation,
with an increase in cladoceran biomass (both Daphnia and non-Daphnia cladocerans) and nutrients
(TP and TN) during spring.
The PCA-biplot on summer averages revealed patterns that were very similar to those observed
for the spring (Figure 6.2b). With the exception of the first year, however, biomanipulation
did not result in a persistent shift along the first axis. The second axis, however, differentiated
between the pre-biomanipulation period (I) and the periods after biomanipulation but before Fe-
addition (Periods II and III), the latter showing strong affinity with higher TP-levels and biomass
of cladoceran zooplankton, especially Daphnia.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
107
6Figure 6.2 – PCA ordination of (a) mean spring (April – June) and (b) mean summer (July – September) water chemistry and plankton variables measured from 2001 to 2013, with the different restoration periods and Fe as supplementary variables. The symbols next to the years correspond to the different restoration periods with yellow, blue, green, red, and purple representing restoration periods before restoration measures (2001-2003), the onset of biomanipulation (2004), after biomanipulation but before iron addition (2005-2009), during (2010-2011), and after iron addition (2012-2013), respectively.
Effects of restoration measures on water column chemistry
Biomanipulation
The year following the onset of the biomanipulation measures (2004) was characterised by
lower total chlorophyll and suspended matter concentrations compared to the period before
biomanipulation (Figure 6.3a, b). During the following years (2005-2009) however, summer
chlorophyll and SM values slowly increased again and reached even higher values than those
measured before biomanipulation (Figure 6.3a, b; Supplementary Figure 6.1a). Conversely,
chlorophyll and SM concentrations in spring and winter experienced a more than two-fold
decrease (Figure 6.3a, b). During the summer months, total phosphorus seemed to show no
response to biomanipulation (Figure 6.3c; Supplementary Figure 6.1b).
Iron addition
Iron addition (2010-2011) resulted in substantial reductions of chlorophyll, SM, and TP
concentrations during the summer months, whereas concentrations in spring and winter
remained low. The decrease was especially pronounced for chlorophyll and SM during the two
years after iron addition (2012-2013), of which concentrations dropped to respectively ca. 30
and 20% compared to the average of the 5 years preceding Fe-addition, a decrease that was even
more pronounced in the 2 years after application (Figure 6.3a, b; Supplementary Figure 6.1a).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
108
Iron addition decreased total phosphorus concentrations to an average of 0.02 mg L-1 throughout
the seasons. However, as soon as iron addition had stopped, summer TP concentrations raised
quickly to pre-restoration concentrations again. This increase was however not noticeable for
orthophosphate, which remained low throughout the summer months after iron addition had
stopped (Supplementary Figure 6.2d). In reaction to iron addition, iron concentrations in the water
column showed a steep increase in the years 2010 and 2011, especially in the period between the
months of August and February (Figure 6.3d). The highest concentration of iron was measured
during the summer of the second year of addition, when water column iron concentrations
reached 0.74 mg L-1. When iron addition had stopped (in the years 2012 and 2013), this seasonal
peak disappeared, but overall iron concentrations in summer tended to remain higher than in
the period preceding the start of iron addition. Throughout the iron addition process, surface
water pH remained well above 7 (Supplementary Figure 6.2a). Moreover, we observed a strong
reduction in dissolved organic carbon (DOC) during iron addition, which, in combination with
other additional nutrient measurements, can be found in Supplementary Figure 6.2.
Figure 6.3 – Responses of phytoplankton chlorophyll (a), suspended matter (b), total phosphorus (c), and iron (d) for each month of the year during each of five time periods. Boxplots represent among-year variability for the period before restoration measures (1986-2003), blue symbols represent the onset of biomanipulation (2004), and green, red, and purple symbols represent medians for the periods after biomanipulation but before iron addition (2005-2009), during (2010-2011), and after iron addition (2012-2013), respectively.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
109
6
Community changes in phytoplankton, zooplankton, and macrofauna
Biomanipulation
Phytoplankton composition before biomanipulation (1996-2003) shows an increase in summer
biomass for all groups, with increasing dominance of cyanobacteria (Figure 6.4b; Supplementary
Figure 6.3a). The start of biomanipulation (2004), however, resulted in substantial reductions of
summer biovolume for all phytoplankton groups, whereas winter concentrations of cyanobacteria
and diatoms noticeably increased (Figure 6.4b; Supplementary Figure 6.3a, c). Phytoplankton
biovolume slowly increased during the spring and summer of the next years of biomanipulation,
with cyanobacteria reoccurring as the dominant group again, similar to the years preceding
biomanipulation (Figure 6.4a, b).
Figure 6.4 – Spring (April – June) and summer (July – September) mean (a, b) phytoplankton and (c, d) zooplankton biovolume in µm3 ml-1 from 1996 to 2013. Black, dark grey, white, and light grey bars represent in (a, b) nitrogen-fixing cyanobacteria, other cyanobacteria, green algae, and diatoms, respectively and in (c, d) Daphnia, other cladocera, copepoda, and rotifera, respectively. Dashed arrows indicate the start of biomanipulation and solid and dashed-dotted arrows indicate the start and stop of iron addition, respectively. a No measurements performed, b no rotifera counted, c values are means of 2 months (May-June or July-August).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
110
Whereas the cyanobacterial community before biomanipulation was dominated by nitrogen-
fixing Aphanizomenon, N-fixing Anabaena became the most abundant taxon between 2005 and
2007, and stayed co-dominant with Microcystis from 2008 onwards (Figure 6.4a, b).
Of the different zooplankton groups, mainly Daphnia and other cladocerans appear to be
influenced by biomanipulation (Figure 6.4c, d). At the onset of biomanipulation (2004) and the
succeeding years of removal of benthi- and planktivorous fish (2005-2009), relative abundance
of Daphnia increased substantially (Figure 6.4c, d). The first year of biomanipulation was
characterised by high spring overall cladoceran biomass (Figure 6.4c). During the following
years, these spring concentrations decreased again, especially for other cladocerans. The overall
measured concentrations remained however higher than the average measured concentrations
during the years before biomanipulation, especially during early spring (April; Supplementary
Figure 6.4a). Furthermore, biomanipulation also resulted in higher summer cladoceran biomass,
which was dominated by Daphnia (Figure 6.4d), although there was a tendency for a decline
towards 2009.
Iron addition
Although the relative abundance of cyanobacteria during the summer of the first year of iron
addition (2010) remained relatively high, contributing with ca. 50% to the total phytoplankton
biomass, phytoplankton biomass was strongly reduced a year after iron addition had started
(2011; Figure 6.4b). After this reduction in 2011, absolute cyanobacterial biovolume remained
relatively low until the end of the study period (2013; Figure 6.4a, b; Supplementary Figure
6.3a).
During and after Fe-addition we observed markedly lower cladoceran biomass during the
summer months August and September, although spring concentrations remained relatively
constant (Figure 6.4c, d; Supplementary Figure 6.4a). Whereas Daphnia dominated the
cladoceran biomass in spring, this group was only found in low numbers during the summer
months. Copepods also appeared to be negatively affected by iron addition as summer biovolume
in the period of iron addition (2010-2011) showed a 30% decline (Supplementary Figure 6.4b).
Nonetheless, copepod summer biovolume recovered quickly upon termination of iron addition
(Supplementary Figure 6.4b). In contrast, cladoceran spring and summer biovolume remained at
the same level as during iron addition (Supplementary Figure 6.4a).
The total number of collected macrofauna taxa in the period after the onset of biomanipulation
remained constant at 124 and 126 genera in 2008 and 2009, respectively. In the second year of
iron addition (2011) however, this number had increased to 157 genera, of which Chironomus
was the most abundant genus.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
111
6
Macrophyte coverage
During the summer of the first year of biomanipulation (2004), submerged macrophytes
emerged in 86% of the sampling points, dominated by meso- and eutrophic macrophyte species,
particularly Ceratophyllum demersum (Figure 6.5). During the succeeding years of biomanipulation
(2005-2009), the abundance of the submerged macrophytes declined from 56% in 2005 over
23%-35% during 2006-2008 to complete disappearance in 2009 (Figure 6.5). During the second
year of iron addition (2011), submerged macrophytes re-appeared and were found in 63% of
the sampling points with Elodea nuttallii as the most dominant species (Figure 6.5). Two years
after iron addition, in the summer of 2013, submerged macrophytes were found in 51% of the
sampling points.
DISCUSSION
Biomanipulation resulted during the first year in reduced sediment resuspension, reduced
levels of phytoplankton biomass and suspended matter, increased biomass of large cladoceran
zooplankton and the development of an extensive submerged macrophyte vegetation (Ter Heerdt
and Hootsmans, 2007). During the succeeding years, at least during the summer, the lake
reverted to pre-restoration conditions, despite continued fish removal. This was probably due to
the high phosphorus concentrations in the lake, which had not changed during biomanipulation.
It is widely recognised that biomanipulation in highly eutrophied water bodies can only be
effective on a longer term when phosphorus concentrations are reduced, both from external and
internal sources (Meijer et al., 1994; Hansson et al., 1998; Søndergaard et al., 2007). Thus, even
though the removal of benthi- and planktivorous fish resulted in an increase of large bodied
cladocerans, the high P concentrations in the lake facilitated algal growth to a point where grazers
were unable to suppress it. High blue-green algal biomass is known to counteract the success of
biomanipulation, as many zooplankton are unable to eat large (toxic) cyanobacterial colonies
(Hansson et al., 1998). Our results show high cyanobacterial biomass before biomanipulation,
which increased even more after the restoration measure had started. Whereas total nitrogen
concentrations in Lake Terra Nova remained low throughout the restoration process, high P
concentrations sustained growth of nitrogen-fixing cyanobacteria, such as Anabaena. Consequently,
as a result of this increase in cyanobacterial biomass in 2005, the light climate deteriorated and
macrophytes started to disappear.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
112
Figure 6.5 – Estimation of submerged macrophyte coverage on 43 locations in Lake Terra Nova during the summer of the years 2004-2013. Coloured symbols represent areas with a coverage ≥ 1%, with light green, orange, and red symbols representing Ceratophyllum demersum, Elodea nuttallii, and either alone or a combination of Potamogeton sp., Zannichellia palustris, and Najas marina, respectively. White symbols represent areas with no submerged macrophytes. When macrophytes were present with less than 1% coverage the vegetation was categorized as sparse vegetation (blue symbols).
Additionally, besides high P concentrations, other factors may have reduced biomanipulation
success. There are various examples of biomanipulated lakes with similar short-term successes,
where biomass of large zooplankton gradually decreased again after the initial improvements
during the first year(s) (e.g. Van Donk et al., 1990; Meijer et al., 1994). Removal of fish results
in reduced intra- and interspecific competition and may therefore enhance the recruitment and
survival of planktivorous young-of-the-year fish (Hansson et al., 1998). However, the observed
increase of zooplankton biomass after biomanipulation, especially of large and efficient grazers
like Daphnia, makes such scenario unlikely in the case of Lake Terra Nova.
Iron addition had strong positive effects on water quality, at least within the time frame
of our monitoring. Both spring and summer TP and SM concentrations and phytoplankton
biomass decreased considerably in the two years after addition. The effectiveness of iron addition
on the long term, however, depends on various chemical properties of the lake. High DOC and
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
113
6
sulphate concentrations in lakes may, for example, reduce the phosphorous binding capacity of Fe
(Smolders et al., 2006). Iron should therefore be added in sufficient concentrations to avoid the
treatment being ineffective. In Lake Terra Nova, DOC concentrations showed a steep decrease
during iron addition, which could imply that a large part of the added iron precipitated with
DOC to form humic-iron complexes. Formation of such stable complexes may have considerably
reduced the amount of free iron to form a P barrier on the water-sediment interface. This may
explain why water column TP concentrations slowly increased to pre-restoration concentrations
during the two years after the termination of the iron addition efforts (2012-2013).
High quantities of iron may also have negative side-effects on lake biota. When added in
excess, high iron concentrations may lower pH and form ironhydroxide (ochre) precipitates. These
iron precipitates can attach to plant surfaces, fish gills, or form a layer on the sediment which
could alter the structure and quality of benthic habitats (Gerhardt and Westermann, 1995).
Although in Lake Terra Nova iron precipitates were observed to form at the site of addition, local
accumulation of high precipitate concentrations were prevented because dosing was controlled by
the strength of wind. The amount of iron precipitates that was found on the sediment surface did
not seem to affect macrofaunal taxon richness. According to Gerhardt and Westermann (1995),
precipitation of iron on fish gills can result in physical stress and tissue damage. Discoloured gills
or any other signs of iron precipitates on fish were however not encountered during the yearly
biomanipulation fish removal (pers. obs. G. ter Heerdt).
Iron in the water column can also impose direct toxic effects on aquatic organisms. Iron
toxicity experiments have shown that iron can be lethal to various fish when concentrations
exceed 47 mg Fe L-1 (Mukhopadhyay and Konar, 1984). The macrofaunal community shows a
wider range of sensitivities, ranging from 0.28 mg Fe L-1 for mayflies (Shuhaimi-Othman et al.,
2012a) to 580 mg Fe L-1 for various oligochaete species (Mukhopadhyay and Konar, 1984). Even
though the total amount of iron added to the water column was high, average iron concentrations
in the water column throughout the addition process (2010-2011) only reached 0.23 ± 0.15 mg
Fe L-1. The highest concentration of 0.75 mg Fe L-1 was measured in the summer of 2011, but
this maximum concentration was only reached on one occasion, which makes it unlikely that
aquatic communities have suffered from high iron concentrations. Other possible distresses of
iron on aquatic organisms, such as a low pH, were avoided by slowly dosing the iron over a long
period of time.
During iron addition, the zooplankton community, especially large bodied cladocera and
copepods, showed a strong decline. This decline seemed, however, to have already started before
iron addition. After the termination of iron addition, copepods slowly increased again whereas
Daphnia and other cladoceran biomass remained low. Iron toxicity studies have shown that
cladocerans and copepods are tolerant to surface water iron concentrations up to 5.9 and 35.2
mg Fe L-1, respectively (Biesinger and Christensen, 1972; Mukhopadhyay and Konar, 1984). The
decline in cladocerans during the summer months of iron addition was therefore probably not
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
114
caused by direct iron toxicity, but more likely by the concomitant decline in P-rich algal biomass,
which may have resulted in food limitation for these filter feeders (Meijer et al., 1994).
The macrophyte community showed a positive response to iron addition, as they returned
in more than half of the sampling points, compared to their absence in 2009. The dominant
macrophyte species that emerged during iron addition (Elodea nuttallii) is known to be a typical
eutrophic species, but various mesotrophic species also slowly reappeared at several locations,
including Potamogeton obtusifolius and Najas marina. Experiments have shown that the lake
sediment does contain seeds and propagules of eutrophic but also mesotrophic species, including
several charophyte species (Van der Wal et al., 2013; Immers et al., 2014). Charophytes were
not (yet) encountered during the summer surveys in 2011 and 2013. The absence of charophytes
after iron addition could possibly be related to lower iron tolerance of these species, as high iron
concentrations in rivers have been observed to restrict the distribution of several iron-intolerant
macrophytes (Vuori, 1995). Various terrestrial plants and helophytes cannot cope with high
iron concentrations, which lead to decreased growth rates, leaf die-off, and even death of the
plants (Van der Welle et al., 2007b). Nonetheless, aquatic plants such as Elodea nuttallii, various
Potamogeton species, and charophytes are known to be relatively tolerant to effects of iron toxicity
(Van der Welle et al., 2007b; Immers et al., 2013, 2014). Moreover, recent experiments have
shown that the germination of charophyte propagules from the sediment is not hindered by iron
addition up to 40 g Fe m-2 (Immers et al., 2014).
The absence of a diverse mesotrophic vegetation could also be related to the presence of invasive
crayfish, which both consume macrophytes as well as disturb the sediment by bioturbation. In
Lake Terra Nova, the amount of invasive crayfish removed during the biomanipulation efforts
steadily increased from 2008 onwards, which most likely reflects an increase in crayfish population
size in the lake. Invasive crayfish, which are increasingly becoming a nuisance in European lakes,
are well-known for their ability to alter aquatic ecosystems by decreasing water transparency
and destroying macrophyte biomass, particularly Procambarus clarkii, the dominant species in
Lake Terra Nova (Bakker et al., 2013; Van der Wal et al., 2013). An experiment showed that
transplanted Chara virgata grew well in Lake Terra Nova, but survival and growth was reduced in
the presence of crayfish (Van der Wal et al., 2013). Even though yearly biomanipulation removed
a large number of crayfish, a considerable amount of the population could have still remained in
the lake.
CONCLUSIONS
The addition of iron contributed substantially to the restoration of the peaty Lake Terra Nova as
it resulted in a general improvement of water quality, a reduction of cyanobacterial biomass, and
a recovery of macrophyte vegetation. TP concentrations, nevertheless, increased shortly after iron
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
115
6
addition had stopped. In order for iron addition to be effective on the long-term, the sediment-
water interface needs to be covered with a surplus of iron to prevent P-leakage from the sediment
into the water column. In our study case, Fe-addition resulted in strong reductions of DOC,
suggesting that DOC bonded to Fe. The fixation of Fe by DOC may to a large extent have reduced
the freely available Fe-reservoir and as such have depleted the long term Fe-buffer available for the
binding of P. Therefore, careful consideration of both the dose and type of the capping agent are
a necessity when planning to restore an organic-rich lake. Nonetheless, the conspicuous success
of Fe-addition despite high DOC suggests that iron addition has the potential to be even more
effective when applied to DOC poor lakes.
The success of Fe-addition in Lake Terra Nova was probably also facilitated by the ongoing
fish removal. Even though sustained biomanipulation alone had not resulted in any important
long-term improvements during the summer months, it may still have enhanced the recovery of
macrophyte cover and diversity through the removal of fish and crayfish.
We conclude that the success of biomanipulation in lakes that suffer from internal P loading
may thus be strongly enhanced by the addition of iron. Ideally, however, nutrient concentrations
are lowered first. In eutrophied lakes where the external input of nutrients has been reduced to
sufficiently low levels, Fe-addition has strong potential to also reduce internal eutrophication.
Once this is achieved, the long term success of biomanipulation is best guaranteed.
ACKNOWLEDGEMENTS
The authors would like to thank Het Waterlaboratorium, Waterproef, Rob van de Haterd from
Bureau Waardenburg, and Kuiper & Burger for providing additional data on fish, macrofauna,
and macrophyte development in Lake Terra Nova. We are also grateful to Erik Reichman for
analysing the zooplankton samples. This study was funded by the Water Framework Directive
Innovation Fund from Agentschap NL from the Dutch Ministry of Economic Affairs, Agriculture
and Innovation.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
116
Supplementary Table 6.1 – Overview of the yearly removed fish and crayfish biomass in Terra Nova.
Removed fish
(kg) per size fraction
Removed crayfish
(kg)
Total removed
(kg ha-1)
Total biomass present
(kg ha-1)*
0 – 15 cm ≥ 15 cm
Winter 2003-2004 6564 9643 190.7 47.8
Winter 2004-2005 4585 1072 66.6
Winter 2005-2006 3563 926 52.8
Winter 2006-2007 1198 290 18.5
Autumn 2007 738 24 9.0
Autumn 2008 733 626 26 16.3 50.4
Autumn 2009 1078 0 119 14.1
Autumn 2010 798 0 188 11.6
Autumn 2011 625 0 108 8.6
Autumn 2012 1002 0 453 17.1
Autumn 2013 911 0 273 13.9 58.7
Total removed (kg) 21795 12581 1167
* Excluding pike (28.0, 11.7, and 8.8 kg ha-1 for 2003, 2008, and 2013 respectively)
Supplementary Table 6.2 – Calculated yearly phosphate input in Terra Nova from both external and internal sources.
Origin of phosphate contribution in Lake Terra Nova Phosphate (g m-2 y-1)
Precipitation 0.02
Seepage 0.04
External input from Lake Loenderveen 0.02
External input (other) 0.01
Agricultural (adjacent fields) 0.01
Birds 0.01
Sediment 0.10
Total 0.22
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
117
6
Supplementary Figure 6.1 – Summer (July – September) mean chlorophyll, suspended matter (a), total phosphorus, and iron (b) from 1986 to 2013. White bars represent in (a) suspended matter and in (b) total phosphorus, grey bars represent in (a) total chlorophyll and in (b) iron. No data available from 1991 to1994. Additionally, suspended matter and Fe were not measured from 1986 to1990, and during the years 1997, 1998, 1999 (and 2000 for Fe). Dashed arrows indicate the start of biomanipulation and solid and dashed-dotted arrows indicate the start and stop of iron addition, respectively.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
118
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
119
6
« Previous page Supplementary Figure 6.2 – Responses of pH (a), chloride (b), sulphate (c), ortho-phosphate (d), total nitrogen (e), anorganic nitrogen (f), oxygen (g), and dissolved organic carbon (h) for each month of the year during each of five time periods. Boxplots represent among-year variability for the period before restoration measures (1986-2003), blue symbols represent the onset of biomanipulation (2004), and green, red, and purple symbols represent medians for the periods after biomanipulation but before iron addition (2005-2009), during (2010-2011), and after iron addition (2012-2013), respectively. The start of biomanipulation caused an increase in pH, TN, and DOC in the water column, which decreased again after iron addition (a, e, h). Due to iron(III)chloride addition, chloride concentrations increased in the water column (b), whereas sulphate concentrations showed a small decrease (c). Moreover, during and after iron addition, the water column remained well oxygenated (g).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 6
120
Supplementary Figure 6.3 – Responses of cyanobacteria (a), green algae (b), and diatoms (c) for each month of the year during each of five time periods. Boxplots represent among-year variability for the period before restoration measures (1986-2003), blue symbols represent the onset of biomanipulation (2004), and green, red, and purple symbols represent medians for the periods after biomanipulation but before iron addition (2005-2009), during (2010-2011), and after iron addition (2012-2013), respectively.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Iron addition and biomanipulation as complementary measures for the restoration of a shallow peaty lake
121
6
Supplementary Figure 6.4 – Responses of cladocera (a), copepoda (b), and rotifera (c) for each month of the year during each of five time periods. Boxplots represent among-year variability for the period before restoration measures (1986-2003), blue symbols represent the onset of biomanipulation (2004), and green, red, and purple symbols represent medians for the periods after biomanipulation but before iron addition (2005-2009), during (2010-2011), and after iron addition (2012-2013), respectively.
CHAPTER 7
Gone with the wind - Stability of cyanobacterial
scums under turbulent conditions
Anne K. Immers, Rob E. Uittenbogaard, and Bas W. Ibelings
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
124
ABSTRACT
High nutrient loading has in many freshwater lakes led to increased cyanobacterial abundance.
Cyanobacteria have the ability to track the illuminated surface mixed layer by altering the
density of their cells and can outcompete other phytoplankton in the competition for light.
During periods of calm weather buoyant cyanobacteria rapidly float to the lake surface where they
may form dense scums. Since cyanobacterial toxins are intracellular, when cells accumulate in a
scum, toxin concentrations increase manifold, posing a threat to lake users, Researchers and lake
managers have tried to predict the timing and location of scum formation well in advance enabling
lake managers to timely warn the public. These models take into account the three essential
preconditions for scum formation: (i) cyanobacterial biomass, (ii) buoyancy state, and (iii) stability
of the water column. Whereas these scum prediction models have been successful in predicting
scums in open water of large lakes, the ability to predict scums in more sheltered places, such as
harbours, ship locks, or urban ponds still remains unreliable. One of the limitations of existing
early warning models is that they use information of only one cyanobacterial species, whereas
a variety of surface bloom forming cyanobacterial species can be present in a lake, which differ
in shape, flotation velocity, and favourable growth conditions. For this reason, we investigated
the formation and disappearance of scums of two different cyanobacterial species, Aphanizomenon
flos-aquae and Woronichinia naegeliana under a range of artificially induced turbulence intensities.
Both our experiments and theoretical computations show that increasing the turbulence to
the highest level completely mixed the distribution of Woronichinia, while Aphanizomenon cell
density remained highest in the upper millimetres of the water column. After the turbulence was
decreased, Woronichinia slowly re-formed a scum at the surface. Aphanizomenon, however, appeared
to be less tolerant to mixing as it lost its ability to float back to the surface. We conclude that
buoyant cyanobacterial species may differ in their response to turbulence. Prediction models
can therefore be improved by identifying the dominant cyanobacteria in the target lake and
incorporating their scum forming characteristics and resistance to turbulence, which may lead to
a more reliable early warning of the public.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
125
7
INTRODUCTION
Eutrophication is still threatening biodiversity and ecosystem functioning of freshwater ecosystems.
One of the most noticeable consequences of eutrophication is the formation of nuisance blooms of
harmful cyanobacteria (Schindler, 1978; Pearl and Huisman, 2008; Pearl et al., 2011). Even though
water quality is increasing in many lakes after measures were taken to control eutrophication during
the past decades, bloom forming cyanobacteria still pose a problem in many freshwater lakes (Carey
et al., 2012; Rigosi et al., 2014). Various cyanobacteria have the ability to produce harmful toxins,
which in majority, like the dominant toxins microcystins, are intracellular. Consequently, when
cells accumulate in scums, toxin concentrations increase manifold. The risks of high concentrations
of toxins in scums for water contact sports and recreation is recognised by water managers around
the world, since in protocols for cyanobacterial risk assessment and management scum formation
typically results in the highest alert-level (Ibelings et al., 2014).
Bloom-forming cyanobacterial taxa typically possess gas-vesicles, hollow structures filled with
air, which decrease the density of the cells and can make them positively buoyant. Using their
variation in cell density, genera like Microcystis are able to track the (illuminated) near surface mixed
layer of lakes (Humphries and Lynne, 1988; Ibelings et al., 1991). During periods of calm weather,
as on hot summer days, when irradiation by the sun and low wind speed reduce turbulent mixing
and enhance the stability of the water column, large, buoyant cyanobacteria may rapidly float to
the lake surface where they form dense scums (Ibelings et al., 2003; Jöhnk et al., 2008; Carey et al.,
2012). Hence there are three pre-conditions for scum formation: (i) presence of cyanobacteria in the
phytoplankton (biomass), (ii) buoyancy, and (iii) a stable water column.
There is increasing evidence that climate warming may further promote blooms (Pearl and
Huisman, 2008) and to some extent this may undo efforts to restore eutrophic systems (Rigosi
et al., 2014). Cyanobacterial blooms are here to stay, and lake managers need to be given the best
possible tools to manage the problem and reduce the risks for lake users (drinking water production,
recreation, fisheries - see Ibelings and Chorus, 2007). As argued above, given the potentially extreme
concentrations of microcystins, scums typically form the highest risk factor (up to tens of thousands
of μg per litre – Ibelings et al., 2012). Up to date information on the timing and location of scums
gives managers time to take appropriate actions, update the status of warning protocols, dissuade, or
even ban swimming (Ibelings et al., 2014). Traditional sampling methods are, however, insufficient
to capture the dynamics in scum formation, both because of a low frequency (perhaps once or twice
per month, whereas scum formation varies on a diel time scale) and an inadequate spatial resolution
(typically a single location is sampled whereas scum formation is highly patchy; Ahn et al., 2008).
Therefore accurate prediction on scum formation to timely warn the public, remains a necessity.
Ibelings et al. (2003) combined modelling of biomass, buoyancy and wind induced turbulence
into an early warning system. This model was able to correctly predict scum formation on basis of
the medium term weather forecast, in the large open water of the IJsselmeer in The Netherlands.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
126
Yet the ability to predict scums in more sheltered water bodies, such as harbours, ship locks, or
urban ponds where human contact with the water tends to be intense, and scums may persist longer,
remained problematic. Further improvements resulted in correct prediction scores up till 50%, when
based on actual measured meteorological data (Burger et al., 2009). The “low” number of correct
predictions was mainly due to the model predicting more scums then were observed in the field
(false positives). This mismatch of model and field measurements was caused by validation issues,
but also the fact that one standard value for buoyancy and flotation velocity (based upon Microcystis)
was used, whereas the lakes were mainly dominated by other cyanobacterial genera (e.g. Anabaena;
Burger et al., 2009). Therefore, the forecasting model would benefit from a better understanding of
scum dynamics of different cyanobacterial species under turbulent conditions (Burger et al., 2009).
We used an experimental approach studying cyanobacterial scum formation under controlled
conditions supported by technical engineering models to study the effect of turbulence on species-
specific scum formation. A common method to mimic wind-generated turbulence in lakes is
using a vertically oscillating grid in mesocosms (DeSilva and Fernando, 1994; Bache and Rasool,
1996; O’Brien et al., 2004; Regel et al., 2004). We hypothesize that (i) larger colony forming and
buoyant cyanobacteria are able to produce scums at higher grid frequencies (i.e. more elevated wind
speeds) and that (ii) the resulting scums are more stable compared to taxa of smaller size. The latter
will remain entrained by turbulent flows even at lower grid frequencies. Moreover, (iii) we expect
a difference between grid frequencies that allow scum formation and those needed to break up
existing scums.
MATERIALS AND METHODS
We experimentally tested the effect of increasing and decreasing levels of turbulence on scum
formation (and breakdown) of the filamentous scum forming cyanobacteria Aphanizomenon flos-aquae
Ralfs ex Bornet & Flahault and colonial Woronichinia naegeliana (Unger) Elenkin.
Collection of material
A quantity of 60 litres of two different scum forming cyanobacteria was harvested in the summer
(August) of 2012. We collected Aphanizomenon flos-aquae in a shallow pond (± 1 m) in the Wilhelmina
park in Rijswijk (Zuid-Holland, The Netherlands) on the 31st of July and two weeks later we
harvested Woronichinia naegeliana in a shallow (± 1 m) city pond in Someren (Noord-Brabant, The
Netherlands) on the 14th of August. The collected cyanobacteria were both the dominant species
in their respective samples. The two sampling locations were high in nutrients and were notorious
locations for cyanobacterial scums. During the harvest, only floating cyanobacteria in scums
were collected from the water surface, while carefully avoiding old and dead material. Collected
material was afterwards carefully transported to the lab of the NIOO-KNAW in Wageningen (The
Netherlands), taking care to avoid collapse of gas-vesicles through sudden pressure shocks.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
127
7
Experimental setup
To measure the distribution of the cyanobacteria in the water column under the influence of
turbulence levels, 3 specially designed 920 litres mesocosms, called Limnotrons (Verschoor et al.,
2003) were filled with 900 litres of groundwater. The Limnotrons are stainless steel vessels (with
a Perspex lid) with a diameter of 0.97 m and a depth of 1.32 m (side) - 1.37 (centre) m. Each
Limnotron was equipped with a frequency controlled oscillating grid below the water surface (depth
of the grid Dgrid
= 110 mm), generating turbulence at four different grid-oscillation frequencies
(ƒ; Table 7.1). The turbulence generated by these oscillating grids represented equivalent physical
conditions to wind-generated turbulence in lakes. The grids were constructed of 1 cm diameter
Perspex bars assembled into a horizontal rectangular grid with a mesh size (M) of 6 cm. The double
amplitude of the grid (or stroke length, S) was 2.8 cm. Temperature and light were left at ambient
conditions (dim light and 21º C).
In order to estimate the turbulence properties used in our experiments and their correspondence
with equivalent wind speeds, we used calculations from Fernando and DeSilva (1993) and O’Brien
et al. (2004) to acquire the energy dissipation (), Kolmogorov length scale (ηk) and the rms of
vertical velocity
137
Aphanizomenon flos-aquae Ralfs ex Bornet & Flahault and colonial Woronichinia naegeliana
(Unger) Elenkin.
Collection of material
A quantity of 60 litres of two different scum forming cyanobacteria was harvested in
the summer (August) of 2012. We collected Aphanizomenon flos-aquae in a shallow pond (±
1 m) in the Wilhelmina park in Rijswijk (Zuid-Holland, The Netherlands) on the 31st of July
and two weeks later we harvested Woronichinia naegeliana in a shallow (± 1 m) city pond in
Someren (Noord-Brabant, The Netherlands) on the 14th of August. The collected
cyanobacteria were both the dominant species in their respective samples. The two sampling
locations were high in nutrients and were notorious locations for cyanobacterial scums.
During the harvest, only floating cyanobacteria in scums were collected from the water
surface, while carefully avoiding old and dead material. Collected material was afterwards
carefully transported to the lab of the NIOO-KNAW in Wageningen (The Netherlands),
taking care to avoid collapse of gas-vesicles through sudden pressure shocks.
Experimental setup
To measure the distribution of the cyanobacteria in the water column under the
influence of turbulence levels, 3 specially designed 920 litres mesocosms, called Limnotrons
(Verschoor et al., 2003) were filled with 900 litres of groundwater. The Limnotrons are
stainless steel vessels (with a Perspex lid) with a diameter of 0.97 m and a depth of 1.32 m
(side) - 1.37 (centre) m. Each Limnotron was equipped with a frequency controlled
oscillating grid below the water surface (depth of the grid Dgrid = 110 mm), generating
turbulence at four different grid-oscillation frequencies (ƒ; Table 7.1). The turbulence
generated by these oscillating grids represented equivalent physical conditions to wind-
generated turbulence in lakes. The grids were constructed of 1 cm diameter Perspex bars
assembled into a horizontal rectangular grid with a mesh size (M) of 6 cm. The double
amplitude of the grid (or stroke length, S) was 2.8 cm. Temperature and light were left at
ambient conditions (dim light and 21º C).
In order to estimate the turbulence properties used in our experiments and their
correspondence with equivalent wind speeds, we used calculations from Fernando and
DeSilva (1993) and O’Brien et al. (2004) to acquire the energy dissipation (𝜖𝜖), Kolmogorov
length scale (ηk) and the rms of vertical velocity |𝑤𝑤′(𝑧𝑧)| for each grid frequency (see for each grid frequency (see Appendix 7.1). An overview of the resulting
values and equivalent wind speeds derived from these equations are presented in Table 7.1.
Table 7.1 – Experimental turbulence conditions during the two experiments on day 1 and day 2 with respectively increasing and decreasing turbulence intensities.
Time (h)
Pump speed (V)
Grid frequency (Hz)
Turbulent energy dissipation
138
Appendix 7.1). An overview of the resulting values and equivalent wind speeds derived from
these equations are presented in Table 7.1.
Table 7.1 – Experimental turbulence conditions during the two experiments on day 1 and day 2 with
respectively increasing and decreasing turbulence intensities.
In each of the Limnotrons, 20 litres of the collected Aphanizomenon was on the day of
the harvest carefully poured in each of the 3 Limnotrons. Hence the experiment was carried
out in triplicate. After transfer to the Limnotrons, the cyanobacteria were left undisturbed for
12 hours in order to allow surface bloom formation. Two weeks later, after cleaning, the
experiment was repeated using the newly collected Woronichinia material. Each of the
species was subjected to both increasing and decreasing levels of turbulence in the Limnotron
(Figure 7.1).
Increasing turbulence levels
When a scum had appeared at the surface of the Limnotrons on the day following the
harvest, the experiment started by sampling depth specific samples of 3 mL using a 5 ml
syringe (internal diameter of tube 1 mm) at depths of 2, 5, 10, and 15 mm below the water
surface. Sampling closer to the surface (< 2 mm) resulted in clogging of the syringe. At each
reference depth, samples were collected at 6 random locations in order to take spatial
variation into account. After sampling the Limnotrons, the oscillating grid was set at the
lowest frequency of 0.53 Hz. One hour after oscillating at this frequency, the whole sampling
Time (h)
Pump speed (V)
Grid frequency (Hz)
Turbulent energy dissipation 𝜖𝜖10% (m2 s-3)
Kolmogorov scale ηk (mm)
Wind speed (m s-1)
|𝑤𝑤′(𝑧𝑧10%)| (mm s-1)
Increasing turbulence (day 1) 0 0 0 - - - - 1 3 0.527 1.710 × 10-7 1.771 0.68 1.22 2 6 1.093 1.510 × 10-6 1.031 1.40 2.56 3 9 1.691 5.610 × 10-6 0.742 2.17 4.0 4 12 2.409 1.610 × 10-5 0.569 3.09 5.6 Decreasing turbulence (day 2) 1 12 2.409 1.610 × 10-5 0.569 3.09 5.6 2 9 1.691 5.610 × 10-6 0.742 2.17 4.0 3 6 1.093 1.510 × 10-6 1.031 1.40 2.56 4 3 0.527 1.710 × 10-7 1.771 0.68 1.22 5 0 0 - - - -
(m2 s-3)
Kolmogor-ov scale η
k
(mm)
Wind speed(m s-1)
138
Appendix 7.1). An overview of the resulting values and equivalent wind speeds derived from
these equations are presented in Table 7.1.
Table 7.1 – Experimental turbulence conditions during the two experiments on day 1 and day 2 with
respectively increasing and decreasing turbulence intensities.
In each of the Limnotrons, 20 litres of the collected Aphanizomenon was on the day of
the harvest carefully poured in each of the 3 Limnotrons. Hence the experiment was carried
out in triplicate. After transfer to the Limnotrons, the cyanobacteria were left undisturbed for
12 hours in order to allow surface bloom formation. Two weeks later, after cleaning, the
experiment was repeated using the newly collected Woronichinia material. Each of the
species was subjected to both increasing and decreasing levels of turbulence in the Limnotron
(Figure 7.1).
Increasing turbulence levels
When a scum had appeared at the surface of the Limnotrons on the day following the
harvest, the experiment started by sampling depth specific samples of 3 mL using a 5 ml
syringe (internal diameter of tube 1 mm) at depths of 2, 5, 10, and 15 mm below the water
surface. Sampling closer to the surface (< 2 mm) resulted in clogging of the syringe. At each
reference depth, samples were collected at 6 random locations in order to take spatial
variation into account. After sampling the Limnotrons, the oscillating grid was set at the
lowest frequency of 0.53 Hz. One hour after oscillating at this frequency, the whole sampling
Time (h)
Pump speed (V)
Grid frequency (Hz)
Turbulent energy dissipation 𝜖𝜖10% (m2 s-3)
Kolmogorov scale ηk (mm)
Wind speed (m s-1)
|𝑤𝑤′(𝑧𝑧10%)| (mm s-1)
Increasing turbulence (day 1) 0 0 0 - - - - 1 3 0.527 1.710 × 10-7 1.771 0.68 1.22 2 6 1.093 1.510 × 10-6 1.031 1.40 2.56 3 9 1.691 5.610 × 10-6 0.742 2.17 4.0 4 12 2.409 1.610 × 10-5 0.569 3.09 5.6 Decreasing turbulence (day 2) 1 12 2.409 1.610 × 10-5 0.569 3.09 5.6 2 9 1.691 5.610 × 10-6 0.742 2.17 4.0 3 6 1.093 1.510 × 10-6 1.031 1.40 2.56 4 3 0.527 1.710 × 10-7 1.771 0.68 1.22 5 0 0 - - - -
(mm s-1)
Increasing turbulence (day 1)
0 0 0 - - - -
1 3 0.527 1.710 × 10-7 1.771 0.68 1.22
2 6 1.093 1.510 × 10-6 1.031 1.40 2.56
3 9 1.691 5.610 × 10-6 0.742 2.17 4.0
4 12 2.409 1.610 × 10-5 0.569 3.09 5.6
Decreasing turbulence (day 2)
1 12 2.409 1.610 × 10-5 0.569 3.09 5.6
2 9 1.691 5.610 × 10-6 0.742 2.17 4.0
3 6 1.093 1.510 × 10-6 1.031 1.40 2.56
4 3 0.527 1.710 × 10-7 1.771 0.68 1.22
5 0 0 - - - -
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
128
In each of the Limnotrons, 20 litres of the collected Aphanizomenon was on the day of the
harvest carefully poured in each of the 3 Limnotrons. Hence the experiment was carried out in
triplicate. After transfer to the Limnotrons, the cyanobacteria were left undisturbed for 12 hours
in order to allow surface bloom formation. Two weeks later, after cleaning, the experiment was
repeated using the newly collected Woronichinia material. Each of the species was subjected to
both increasing and decreasing levels of turbulence in the Limnotron (Figure 7.1).
Increasing turbulence levels
When a scum had appeared at the surface of the Limnotrons on the day following the harvest,
the experiment started by sampling depth specific samples of 3 mL using a 5 ml syringe (internal
diameter of tube 1 mm) at depths of 2, 5, 10, and 15 mm below the water surface. Sampling
closer to the surface (< 2 mm) resulted in clogging of the syringe. At each reference depth,
samples were collected at 6 random locations in order to take spatial variation into account. After
sampling the Limnotrons, the oscillating grid was set at the lowest frequency of 0.53 Hz. One
hour after oscillating at this frequency, the whole sampling process was repeated, followed by an
increase in grid frequency (Figure 7.1). Sampling was carried out at hourly intervals with grid
frequencies of 1.09, 1.69, and 2.41Hz
Figure 7.1 – Schematic overview of grid-generated turbulence over time during the experiments with increasing and decreasing grid frequencies.
Decreasing turbulence levels
During the following day, after 16 hours of stagnant (non-mixing) conditions overnight, the
procedure was reversed with the same cyanobacterial species to test the process of scum formation
under decreasing turbulence levels. This time the measurements started one hour after the grids had
been oscillating at the highest frequency of 2.41 Hz, with the intention of evenly distributing the
cyanobacteria throughout the depth of the Limnotron at the start of the experiment. The oscillation
speed was decreased stepwise after each sampling routine using the same oscillation frequency as
mentioned above (Figure 7.1).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
129
7
Sampling and analyses
For the purpose of investigating the influence of different turbulence levels, the distribution of
Aphanizomenon and Woronichinia over depth was measured as chlorophyll-a and cell concentrations
mL-1. Photographs were taken of the water surface for each of the two species at each grid frequency
for visual comparison. Five of the collected subsamples for each time and depth point of each
Limnotron were stored at -20 ºC for chlorophyll analysis. The sixth subsample was fixed with Lugol
for microscopic analysis. Chlorophyll-a from samples was extracted according to the Dutch NEN
6520 protocol and was measured using a quartz microplate (Hellma, Müllheim, Germany) on a
microplate reader (Biotek Synergy HT, Beun de Ronde, Abcoude, The Netherlands). Chlorophyll-a
concentrations (µg L-1) were calculated using the calibration equation (1) from Lorenzen (1967):
140
Beun de Ronde, Abcoude, The Netherlands). Chlorophyll-a concentrations (µg L-1) were
calculated using the calibration equation (1) from Lorenzen (1967):
𝐶𝐶ℎ𝑙𝑙𝑎𝑎 = 𝐴𝐴𝐶𝐶ℎ𝑙𝑙𝑙𝑙∗𝐾𝐾∗𝑉𝑉𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑙𝑙𝑒𝑒𝑒𝑒∗{(𝐸𝐸𝜆𝜆6650 −𝐸𝐸𝜆𝜆750
0 )−(𝐸𝐸𝜆𝜆665𝑙𝑙 −𝐸𝐸𝜆𝜆750
𝑙𝑙 )}𝑉𝑉𝑠𝑠𝑙𝑙𝑠𝑠𝑠𝑠𝑙𝑙𝑒𝑒 ∗𝑙𝑙𝑠𝑠𝑙𝑙𝑙𝑙𝑒𝑒𝑒𝑒
(1)
where the absorption coefficient of chlorophyll-a for cyanobacteria AChla equals 11.90 as
determined by Ritchie (2006), K is the factor to equate the reduction in absorbance to initial
chlorophyll-a concentrations which equals 2.43 (Lorenzen, 1967), E0 and Ea are the
absorbance at given wavelength before and after acidification respectively, Vextract and Vsample
the volume of ethanol and sample used in ml and litres respectively, and lplate the length of the
light path in the microplate in cm.
Microscope measurements on cell number, cells per colony or filament and colony
size (l w) were performed on an inverted microscope (DMI 4000B, Leica Microsystems
CMS GmbH, Münster, Germany) and the image analysis program Cell-D (Olympus Soft
Imaging System GmbH, Münster, Germany). Moreover, presence of cells specialized in
nitrogen fixation (heterocysts), which could under turbulent conditions decrease the strength
of the filaments, were noted for Aphanizomenon.
Statistical analyses
Statistical analyses were carried out using SPSS 19 (SPSS, Chicago, IL, USA).
Differences between cell counts, colony and filament formation, colony size, and
chlorophyll-a concentrations were tested with a two-way ANOVA using grid frequency and
sampling depth as fixed factors, followed by a Tukey’s post-hoc test. Spatial variation within
the Limnotrons was taken into account by sampling per depth at 6 different horizontal
locations, of which 5 samples were used for the chlorophyll-a measurements. The spatial
variation was not accounted for in the sixth sample, which was used for the microscope
measurements only.
Prior to analysis, all data were tested for normality and homogeneity of variance, and
if necessary, data were log 10 transformed. For data that had no normal distribution, even
after transformation, a nonparametric Kruskal-Wallis test was used with Statistica12 (StatSoft
Inc., Tulsa, OK, USA) to analyze variances. Results were expressed as mean ± standard error
of mean (sem) and P ≤ 0.05 was accepted for statistical significance.
(1)
where the absorption coefficient of chlorophyll-a for cyanobacteria AChla
equals 11.90 as determined
by Ritchie (2006), K is the factor to equate the reduction in absorbance to initial chlorophyll-a
concentrations which equals 2.43 (Lorenzen, 1967), E0 and Ea are the absorbance at given wavelength
before and after acidification respectively, Vextract
and Vsample
the volume of ethanol and sample used in
ml and litres respectively, and lplate
the length of the light path in the microplate in cm.
Microscope measurements on cell number, cells per colony or filament and colony size
(l × w) were performed on an inverted microscope (DMI 4000B, Leica Microsystems CMS
GmbH, Münster, Germany) and the image analysis program Cell-D (Olympus Soft Imaging
System GmbH, Münster, Germany). Moreover, presence of cells specialized in nitrogen fixation
(heterocysts), which could under turbulent conditions decrease the strength of the filaments, were
noted for Aphanizomenon.
Statistical analyses
Statistical analyses were carried out using SPSS 19 (SPSS, Chicago, IL, USA). Differences between
cell counts, colony and filament formation, colony size, and chlorophyll-a concentrations were
tested with a two-way ANOVA using grid frequency and sampling depth as fixed factors,
followed by a Tukey’s post-hoc test. Spatial variation within the Limnotrons was taken into
account by sampling per depth at 6 different horizontal locations, of which 5 samples were used
for the chlorophyll-a measurements. The spatial variation was not accounted for in the sixth
sample, which was used for the microscope measurements only.
Prior to analysis, all data were tested for normality and homogeneity of variance, and if
necessary, data were log 10 transformed. For data that had no normal distribution, even after
transformation, a nonparametric Kruskal-Wallis test was used with Statistica12 (StatSoft Inc.,
Tulsa, OK, USA) to analyze variances. Results were expressed as mean ± standard error of mean
(sem) and P ≤ 0.05 was accepted for statistical significance.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
130
RESULTS
Effect of increasing turbulence levels
Aphanizomenon
In the absence of turbulent mixing, Aphanizomenon formed a dense scum over 12 hours that
contrary to the expectation did not really change in appearance with increasing grid frequencies.
The top view of the cyanobacterial surface distribution one hour after each increase in grid
frequency is shown in Figure 7.2a. Just a few open areas started to form at larger grid frequencies
of 1.70 and 2.41 Hz, although the overall surface area remained dark green (Figure 7.2a, upper
panel).
Overall Aphanizomenon cell and filament number mL-1 in the top 2 mm were higher at all
grid frequencies (0.53 - 2.41 Hz) than under stagnant conditions (Figure 7.3a, c; P < 0.001).
Depth distribution for both cell and filament number did not change with increasing turbulence
(P = 0.200 and P = 0.219) and the highest average number of cells and filaments mL-1 was always
found at 2 mm and the lowest number at depths of 10 and 15 mm. Overall, filament size, i.e. the
number of cells per filament, did not change with increasing turbulence but the biggest filaments
were always found in the top 2 mm of the Limnotrons, while the smaller filaments were found at
a depth of 15 mm (Figure 7.3e; P < 0.001). The distribution of heterocysts per total cell number
over depth did not change with turbulence, but the highest number of heterocysts per total cell
number (0.028 ± 0.002) was always found near the surface at a depth of 5 mm, and the lowest
(0.021 ± 0.002) at a depth of 10 mm (Supplementary Figure 7.1a; P = 0.012).
Chlorophyll-a concentrations measured 2 and 5 mm below the water surface of the Limnotron
increased with increasing grid frequencies, particularly in the top 2 mm, with higher concentrations
measured when the grid was oscillating compared to stagnant conditions (Supplementary Figure
7.2a, P = 0.016). The dense scum which had formed was measured at a depth of 2 mm and to a
lesser degree at 5 mm, and did not disappear with an increase in grid frequency (Supplementary
Figure 7.2a). Compared to the stagnant conditions, concentrations at 10 and 15 mm did neither
significantly increase nor decrease during the stepwise increase in grid frequency (P = 0.579 and
P = 0.241, respectively).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
131
7
Figure 7.2 – Pictures taken from the surface of Limnotron 1 for Aphanizomenon (a) and Woronichinia (b) 1 hour after setting each mixing regime. The upper panels show surface views of day 1 with increasing turbulence levels and the bottom panels show surface views of day 2 with decreasing turbulence levels. Grid frequencies are depicted in the upper left corner of each picture.
Woronichinia
After 12 hours of stagnant conditions overnight, Woronichinia had formed a scum at the surface.
However, in contrast to Aphanizomenon, an increase in turbulence changed the appearance of
the Woronichinia scum and holes and gaps started to appear already at the lowest grid frequency
of 0.53 Hz, corresponding to 0.68 m s-1 wind speed (Figure 7.2b, upper panel). At 1.70 Hz
grid frequency, just small surface areas remained covered with (floating) scums and the scum
completely disappeared at a turbulence frequency of 2.41 Hz, corresponding to 3.10 m s-1 wind
speed.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
132
Figure 3a-f – Cell number mL-1 (a, b), filament number mL-1 (c, d), and cells per filament (e, f) for Aphanizomenon measured at four different depths with (a, c, e) increasing and (b, d, f) decreasing turbulence levels.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
133
7
Figure 3g-l – Cell number mL-1 (g, h) colony number mL-1 (i, j), and cells per colony (k, l) for Woronichinia measured at four different depths with (g, i, k) increasing and (h, j, l) decreasing turbulence levels.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
134
Woronichinia was found both as single cells and as colonies, which numbers showed a
comparable depth distribution (Figure 7.3g, i). Highest concentrations of cells and colonies
were found in the top 2 mm at stagnant conditions and at the two lowest grid frequencies of
0.53 and 1.09 Hz, but at grid frequencies exceeding 1.09 Hz, concentrations in the top 2 mm
decreased. At the highest grid frequencies of 1.7 and 2.41 Hz, the depth distribution changed and
measured cell and colony concentrations became uniformly distributed over the four measured
depths. The number of cells per colony did not show any significant differences between depth
and turbulence speed (P = 0.123 and P = 0.228; Figure 7.3k), whereas colony size did show a
significant difference between the different turbulence speeds (P = 0.002), with bigger colonies
measured at the grid frequencies of 0.53 and 2.41 Hz compared to situations without mixing
(Supplementary Figure 7.1c).
Woronichinia chlorophyll-a concentrations at the depth of 2 mm showed a large variation
between grid frequencies and Limnotrons, with high concentrations remaining at this depth,
right up to the highest oscillation frequency of 2.41 Hz, when chlorophyll-a became more evenly
distributed over the different measured depths (Supplementary Figure 7.2c). This change in depth
distribution between the different mixing regimes was however, not significant (P = 0.237).
Effect of decreasing turbulence levels
Aphanizomenon
Surface pictures taken during the stepwise decrease in turbulence clearly showed a decrease in
abundance of Aphanizomenon at the surface compared to the preceding phase of the experiment –
described above – in which turbulence was increased. Big gaps of clear water already appeared at
the surface after 16 hours of stagnant conditions overnight and no change was noted one hour after
mixing at the highest grid frequency and when grid frequencies were stepwise reduced (Figure
7.2a, bottom panel). The pictures did, however, show that the scum still had not completely
disappeared one hour after mixing at the highest grid frequency and during the stepwise decrease
in grid frequency.
Cell and filament number mL-1 for Aphanizomenon under decreasing turbulence started
with slightly higher concentrations in the top 2 mm compared to the other depths, but this
difference disappeared with decreasing turbulence speeds (Figure 7.3b, d). Once more, overall
cell and filament concentrations per grid frequency were considerably lower compared to the
measurements from the experiment with increasing grid frequencies. This difference was also
visible for both the number of cells per filament and heterocysts mL-1 (Figure 7.3f; Supplementary
Figure 7.1b), which did not show any significant differences between the measured depths and
grid frequencies (P = 0.437 and 0.765 for cells per filament, P = 0.052 and 0.379 for heterocysts
mL-1).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
135
7
Average chlorophyll-a concentrations measured for Aphanizomenon when grid frequency
decreased over time, were ten-fold lower than concentrations measured at equal grid frequencies
in the previous series of increasing grid frequencies during the first phase of the experiment and
showed at the depths of 2 and 5 mm a great variance between the different grid frequencies
(Supplementary Figure 7.2b). Nonetheless, a scum at the top layer (2 mm) in the Limnotrons
remained present, even at the start of the experiment when mixing was at its highest setting (P
= 0.003), but a pattern with decreasing turbulence was not found (Supplementary Figure 7.2b).
Woronichinia
Woronichinia appeared uniformly mixed over depth after 1 hour of mixing at the highest grid
frequency (Figure 7.2b, bottom panel). Decreasing the grid frequency did not change this well-
mixed appearance, except for when mixing was completely stopped, at which time a very thin
layer of cyanobacteria re-appeared at the surface.
Woronichinia cell and colony concentrations, cells per colony and colony size at the start of
the experiment with decreasing turbulence were similar to the concentrations measured at the
highest turbulence during the previous day and remained stable throughout the experiment with
no differences between the measured depths (P = 0.095, 0.429, 0.192, and 0.190 for cell and
colony number, cells per colony, and colony size, respectively; Figure 7.3h, j, l; Supplementary
Figure 7.1d).
Decreasing the grid frequency from 2.41 to 0.53 Hz did not significantly affect the depth
distribution of chlorophyll-a concentrations for Woronichinia and depth distribution remained
similar to the distribution which was measured at the highest grid frequency of 2.41 Hz during
the previous day (Supplementary Figure 7.2d). The chlorophyll-a distribution changed, however,
when mixing was completely stopped, with a higher concentration measured at 2 mm compared
to the other measured depths (P < 0.001).
DISCUSSION
In this experiment we followed scum formation and scum disappearance of two different buoyant
cyanobacterial species under increasing and decreasing turbulence levels. Cell and chlorophyll-a
distribution over the four measured depths in response to the applied grid frequencies were
different for the two species, with a more stable scum for Aphanizomenon compared to a more
easily disturbed scum of Woronichinia. Some of the results, perhaps in particular for Aphanizomenon
were contrary to the expectations on basis of field observations (e.g. Ibelings et al., 1991, 2003).
Differences in turbulence resistance or scum stability may partly be explained by the
differences in flotation velocity of each of the cyanobacterial species (Walsby, 1991). Although we
lack direct flotation velocity measurements, Stokes Law shows a quadratic dependency of flotation
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
136
velocity on particle (cell or colony) radius. Known size distributions for these species show that
Aphanizomenon can form much bigger string-like aggregates of up to 30 mm long (McLachlan
et al., 1963), compared to the much smaller colonies of Woronichinia with an average diameter
of 180 µm (Wilk-Wozniak et al., 2003). Larger cells or aggregates therefore are expected to
more easily dis-entrain from weakening turbulence than smaller ones, leading to the expectation
that Aphanizomenon scums will form more readily and remain in position even under increasing
turbulence. However, in our experiments it was unclear whether an increase in turbulence
(O’Brien et al., 2004) or the sampling process itself (mainly the small internal diameter of the
syringe) had reduced the colonial aggregates to single filaments (Aphanizomenon) and single cells
(Woronichinia).
The effect of turbulence on (buoyant) particles in a fluid has been intensely investigated
in the field of civil and chemical engineering (Brumley and Jirka, 1987; Bache and Rasool,
1996; Penning et al., 2013). In order to assess and understand scum behaviour of both species
with increasing and decreasing turbulence, we therefore turn to technical engineering methods
based on turbulence models tested for near-bed mixing of sand and mud particles up to high
concentrations – See Box 1 for a full explanation.
Comparison of the turbulence model simulations to the experiments
Both our experiments (Figures 7.1 – 7.3 in the Results section) and simulations based on
turbulence models for sediment particles (Figures 7.4 – 7.5 in Box 1) show that increasing
the turbulence to the highest level completely mixed the distribution of Woronichinia over the
four measured depths, while Aphanizomenon cell density remained highest in the upper layer (2
mm) of the water column. The turbulence dissipation rates in our experiment might thus not
have reached the critical level required to negatively influence Aphanizomenon scum formation.
When comparing the turbulence dissipation rates of our experiments (
147
compared to a more easily disturbed scum of Woronichinia. Some of the results, perhaps in
particular for Aphanizomenon were contrary to the expectations on basis of field observations
(e.g. Ibelings et al., 1991, 2003).
Differences in turbulence resistance or scum stability may partly be explained by the
differences in flotation velocity of each of the cyanobacterial species (Walsby, 1991).
Although we lack direct flotation velocity measurements, Stokes Law shows a quadratic
dependency of flotation velocity on particle (cell or colony) radius. Known size distributions
for these species show that Aphanizomenon can form much bigger string-like aggregates of
up to 30 mm long (McLachlan et al., 1963), compared to the much smaller colonies of
Woronichinia with an average diameter of 180 µm (Wilk-Wozniak et al., 2003). Larger cells
or aggregates therefore are expected to more easily dis-entrain from weakening turbulence
than smaller ones, leading to the expectation that Aphanizomenon scums will form more
readily and remain in position even under increasing turbulence. However, in our
experiments it was unclear whether an increase in turbulence (O’Brien et al., 2004) or the
sampling process itself (mainly the small internal diameter of the syringe) had reduced the
colonial aggregates to single filaments (Aphanizomenon) and single cells (Woronichinia).
The effect of turbulence on (buoyant) particles in a fluid has been intensely
investigated in the field of civil and chemical engineering (Brumley and Jirka, 1987; Bache
and Rasool, 1996; Penning et al., 2013). In order to assess and understand scum behaviour of
both species with increasing and decreasing turbulence, we therefore turn to technical
engineering methods based on turbulence models tested for near-bed mixing of sand and mud
particles up to high concentrations – See Box 1 for a full explanation.
Comparison of the turbulence model simulations to the experiments
Both our experiments (Figures 7.1 – 7.3 in the Results section) and simulations based
on turbulence models for sediment particles (Figures 7.4 – 7.5 in Box 1) show that increasing
the turbulence to the highest level completely mixed the distribution of Woronichinia over
the four measured depths, while Aphanizomenon cell density remained highest in the upper
layer (2 mm) of the water column. The turbulence dissipation rates in our experiment might
thus not have reached the critical level required to negatively influence Aphanizomenon scum
formation. When comparing the turbulence dissipation rates of our experiments (𝜖𝜖10% = 1.7 ×
10-7 to 1.6 × 10-5 m2 s-3) with dissipation rates often found in (shallow) lakes (10-11 – 10-5 m2
s-3, Zülicke et al., 1998; Wüest and Lorke, 2003), our experiments can be compared to field
conditions with low to moderate levels of turbulence. This was also shown by the translation
= 1.7 × 10-7 to 1.6
× 10-5 m2 s-3) with dissipation rates often found in (shallow) lakes (10-11 – 10-5 m2 s-3, Zülicke et
al., 1998; Wüest and Lorke, 2003), our experiments can be compared to field conditions with
low to moderate levels of turbulence. This was also shown by the translation of the turbulence
frequencies exerted by our grids to wind speeds, which varied from 0.68 m s-1 at the lowest grid
frequency to 3.10 m s-1 at the highest grid frequency. According to Webster and Hutchinson
(1994) and Wallace and Hamilton (2000), wind speeds higher than 2-3 m s-1 were needed to re-
entrain Microcystis aeruginosa scums. In this context, Aphanizomenon scum persistence is similar to
that of this well known scum forming species.
Whereas Figure 7.5a and 7.5b, using flotation velocities of 0.5 m h-1 and 5 m h-1, correctly
follow the scum behaviour of Woronichinia and Aphanizomenon with increasing turbulence as
measured during our experiments, Figure 7.5a shows an increase in scum density at the surface
after the stepwise decrease in turbulence intensity, which was not visible in the experiments
with Aphanizomenon. The ability to form scums after a period of high turbulence appears to differ
between cyanobacterial species. Whereas a low level of turbulence (1-2 Hz) has shown to increase
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
137
7
metabolic activity in M. aeruginosa strains without affecting cell viability (Regel et al., 2004) and
high intensity (i.e. deep) mixing actually enhanced buoyancy in M. aeruginosa strains due to a
decrease in rate of carbohydrate accumulation (Wallace and Hamilton, 2000), Aphanizomenon lost
its ability to float back to the surface after a decrease in turbulence. Heterocystous cyanobacteria
such as Aphanizomenon, however, are known to be less shear tolerant than other cyanobacteria
(Moisander et al., 2002), and the genus Aphanizomenon is usually only found under stable or
low turbulence conditions (Berman and Shteinman, 1998). Experiments by Moisander et al.
(2002) showed that filament length significantly decreased with increasing shear, resulting in a
distribution of these species over the entire mixed layer. This is in accordance with our results,
where after a whole day of increased mixing, filament size was significantly reduced at the start of
the experiment with decreasing mixing speeds. As a result, cell and chlorophyll-a concentrations
and the number of heterocysts per total cells in the upper 15 mm of the water column, were a
hundred-fold lower compared to day 1, even in the absence of turbulence, which indicates that
most of the cells had sunk to a lower depth.
Implications for scum prediction models
Our experiments give two important insights in species specific scum dynamics that can be used
to improve scum prediction models.
First of all, scum prediction models would benefit from using species specific information
of the dominant cyanobacteria in their target lakes. The scum prediction model of Burger et al.
(2009) used flotation characteristics of the fast floating genus Microcystis, whereas the target lakes
were mainly dominated by N-fixing genera, such as Anabaena and Aphanizomenon. Aphanizomenon,
like Microcystis, can produce stable scums that persist at the surface with wind speeds up to 3 m s-1,
but this genus has shown to be less shear tolerant (Moisander et al., 2002), with high turbulence
breaking up the aggregates, hence reducing flotation velocity. Incorporating this information
into a model would translate into a time lag for the formation of heterocystous cyanobacterial
scums after a period of high wind. It could partially explain why Burger et al. (2009) produced
many false positives in their predictions.
Secondly, in our Limnotron experiments we also noted a difference in turbulence intensity
between a situation where an increase in turbulence intensity breaks up an existing scum and
a situation where a decrease in turbulence intensity enabled scum formation. This difference in
turbulence shows that the turbulence which still allows scum formation, allowing particles to
float up to the surface, is lower than the turbulence needed to break up an already formed scum
(e.g. erode the layer of particles at the surface). This phenomenon, which can also be applied
to other forms of surface layers such as oil spills, shows that the formation of a layer at the
surface dampens the effect of turbulence, causing a self-stabilising state. Our results emphasize
the importance of this difference, which was already incorporated in the original early warning
models of Ibelings et al. (2003).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
138
By combining the knowledge of civil engineering and biological processes, we improved our
understanding of scum formation processes under turbulent conditions, which can be used to
benefit shallow lake scum predictions models. We believe that the number of correctly predicted
scums will increase when scum prediction models include species specific information, such as
flotation velocity and shear resistance, which will ultimately lead to better protection of lake
users. We also note however, that a further series of Limnotron experiments would be helpful,
incorporating yet more species and ideally adjusting the oscillating grids so that higher levels of
turbulence can be generated.
Box 1 – Understanding cyanobacterial scum formation – learning from engineering
studies
Understanding turbulence effects on cyanobacterial particles
There is some correspondence between the breakdown of oil slicks and the stability of
cyanobacterial scums under turbulent conditions in water, as investigated here. However,
due to the larger buoyancy of oil droplets and the large surface tension of oil, the destruction
of oil slicks occurs mainly under breaking waves (Delvigne and Sweeny, 1988; Tkalich and
Chan, 2002), which generate notably larger turbulence levels and occur at larger wind speeds
than the wind speeds which are relevant for the breakdown of scums in lakes. In terms
of turbulence levels, a better correspondence between the stability of cyanobacteria scums
created by buoyant cyanobacteria is the erosion of fluffy mud layers created by sinking mud
flocs by wind-wave generated turbulence in shallow lakes (Penning et al., 2013). Fluffy mud
layers flow and mix as a heavy fluid and can be treated as such using turbulence-mixing
models (e.g. Winterwerp et al., 2001). To understand the vertical profiles of cell and filament
concentrations in our experiments, we therefore treat the scum as a buoyant fluffy layer of
cyanobacteria, possible weakly adhered by mucus.
Whereas mud beds are eroded by water flowing over the bed, thereby creating turbulence
above the bed, cyanobacterial scums are eroded by wind blowing over the lake surface,
generating turbulence in the water column, which we attempted to recreate in our experiments
by using an oscillating grid below the scum layer. Detailed observations of Brumley and
Jirka (1987) describe the vertical mixing of particles with and without this grid-generated
turbulence near the water surface, which we have summarised in Supplementary Figure 7.3.
Vertical mixing is determined by the product of the mixing-length scale proportional to
depth (z) below the water surface and the rms of the vertical velocity (w), written as
150
Box 1 – Understanding cyanobacterial scum formation – learning from engineering studies
Understanding turbulence effects on cyanobacterial particles
There is some correspondence between the breakdown of oil slicks and the stability of
cyanobacterial scums under turbulent conditions in water, as investigated here. However, due
to the larger buoyancy of oil droplets and the large surface tension of oil, the destruction of
oil slicks occurs mainly under breaking waves (Delvigne and Sweeny, 1988; Tkalich and
Chan, 2002), which generate notably larger turbulence levels and occur at larger wind speeds
than the wind speeds which are relevant for the breakdown of scums in lakes. In terms of
turbulence levels, a better correspondence between the stability of cyanobacteria scums
created by buoyant cyanobacteria is the erosion of fluffy mud layers created by sinking mud
flocs by wind-wave generated turbulence in shallow lakes (Penning et al., 2013). Fluffy mud
layers flow and mix as a heavy fluid and can be treated as such using turbulence-mixing
models (e.g. Winterwerp et al., 2001). To understand the vertical profiles of cell and filament
concentrations in our experiments, we therefore treat the scum as a buoyant fluffy layer of
cyanobacteria, possible weakly adhered by mucus.
Whereas mud beds are eroded by water flowing over the bed, thereby creating
turbulence above the bed, cyanobacterial scums are eroded by wind blowing over the lake
surface, generating turbulence in the water column, which we attempted to recreate in our
experiments by using an oscillating grid below the scum layer. Detailed observations of
Brumley and Jirka (1987) describe the vertical mixing of particles with and without this grid-
generated turbulence near the water surface, which we have summarised in Supplementary
Figure 7.3. Vertical mixing is determined by the product of the mixing-length scale
proportional to depth (z) below the water surface and the rms of the vertical velocity (w),
written as |𝑤𝑤′(𝑧𝑧)|. For sufficiently large ratios of the cyanobacteria flotation velocity (wr) to
the vertical velocity variations |𝑤𝑤′(𝑧𝑧)|, cyanobacteria can concentrate near the water surface,
whereas at a low ratio they remain well mixed. We are thus interested in |𝑤𝑤′(𝑧𝑧)| for our
experiments and lacking direct observations, we derive these estimates from Brumley and
Jirka (1987).
Over a depth interval of about 10% of the mean grid depth (Dgrid = 11 cm) below the
water surface, the vertical turbulence velocity increases proportional to 13z , but beyond this
depth the turbulence properties agree with those of grid-generated turbulence without the
.
For sufficiently large ratios of the cyanobacteria flotation velocity (wr) to the vertical velocity
variations
150
Box 1 – Understanding cyanobacterial scum formation – learning from engineering studies
Understanding turbulence effects on cyanobacterial particles
There is some correspondence between the breakdown of oil slicks and the stability of
cyanobacterial scums under turbulent conditions in water, as investigated here. However, due
to the larger buoyancy of oil droplets and the large surface tension of oil, the destruction of
oil slicks occurs mainly under breaking waves (Delvigne and Sweeny, 1988; Tkalich and
Chan, 2002), which generate notably larger turbulence levels and occur at larger wind speeds
than the wind speeds which are relevant for the breakdown of scums in lakes. In terms of
turbulence levels, a better correspondence between the stability of cyanobacteria scums
created by buoyant cyanobacteria is the erosion of fluffy mud layers created by sinking mud
flocs by wind-wave generated turbulence in shallow lakes (Penning et al., 2013). Fluffy mud
layers flow and mix as a heavy fluid and can be treated as such using turbulence-mixing
models (e.g. Winterwerp et al., 2001). To understand the vertical profiles of cell and filament
concentrations in our experiments, we therefore treat the scum as a buoyant fluffy layer of
cyanobacteria, possible weakly adhered by mucus.
Whereas mud beds are eroded by water flowing over the bed, thereby creating
turbulence above the bed, cyanobacterial scums are eroded by wind blowing over the lake
surface, generating turbulence in the water column, which we attempted to recreate in our
experiments by using an oscillating grid below the scum layer. Detailed observations of
Brumley and Jirka (1987) describe the vertical mixing of particles with and without this grid-
generated turbulence near the water surface, which we have summarised in Supplementary
Figure 7.3. Vertical mixing is determined by the product of the mixing-length scale
proportional to depth (z) below the water surface and the rms of the vertical velocity (w),
written as |𝑤𝑤′(𝑧𝑧)|. For sufficiently large ratios of the cyanobacteria flotation velocity (wr) to
the vertical velocity variations |𝑤𝑤′(𝑧𝑧)|, cyanobacteria can concentrate near the water surface,
whereas at a low ratio they remain well mixed. We are thus interested in |𝑤𝑤′(𝑧𝑧)| for our
experiments and lacking direct observations, we derive these estimates from Brumley and
Jirka (1987).
Over a depth interval of about 10% of the mean grid depth (Dgrid = 11 cm) below the
water surface, the vertical turbulence velocity increases proportional to 13z , but beyond this
depth the turbulence properties agree with those of grid-generated turbulence without the
, cyanobacteria can concentrate near the water surface, whereas at a low
ratio they remain well mixed. We are thus interested in
150
Box 1 – Understanding cyanobacterial scum formation – learning from engineering studies
Understanding turbulence effects on cyanobacterial particles
There is some correspondence between the breakdown of oil slicks and the stability of
cyanobacterial scums under turbulent conditions in water, as investigated here. However, due
to the larger buoyancy of oil droplets and the large surface tension of oil, the destruction of
oil slicks occurs mainly under breaking waves (Delvigne and Sweeny, 1988; Tkalich and
Chan, 2002), which generate notably larger turbulence levels and occur at larger wind speeds
than the wind speeds which are relevant for the breakdown of scums in lakes. In terms of
turbulence levels, a better correspondence between the stability of cyanobacteria scums
created by buoyant cyanobacteria is the erosion of fluffy mud layers created by sinking mud
flocs by wind-wave generated turbulence in shallow lakes (Penning et al., 2013). Fluffy mud
layers flow and mix as a heavy fluid and can be treated as such using turbulence-mixing
models (e.g. Winterwerp et al., 2001). To understand the vertical profiles of cell and filament
concentrations in our experiments, we therefore treat the scum as a buoyant fluffy layer of
cyanobacteria, possible weakly adhered by mucus.
Whereas mud beds are eroded by water flowing over the bed, thereby creating
turbulence above the bed, cyanobacterial scums are eroded by wind blowing over the lake
surface, generating turbulence in the water column, which we attempted to recreate in our
experiments by using an oscillating grid below the scum layer. Detailed observations of
Brumley and Jirka (1987) describe the vertical mixing of particles with and without this grid-
generated turbulence near the water surface, which we have summarised in Supplementary
Figure 7.3. Vertical mixing is determined by the product of the mixing-length scale
proportional to depth (z) below the water surface and the rms of the vertical velocity (w),
written as |𝑤𝑤′(𝑧𝑧)|. For sufficiently large ratios of the cyanobacteria flotation velocity (wr) to
the vertical velocity variations |𝑤𝑤′(𝑧𝑧)|, cyanobacteria can concentrate near the water surface,
whereas at a low ratio they remain well mixed. We are thus interested in |𝑤𝑤′(𝑧𝑧)| for our
experiments and lacking direct observations, we derive these estimates from Brumley and
Jirka (1987).
Over a depth interval of about 10% of the mean grid depth (Dgrid = 11 cm) below the
water surface, the vertical turbulence velocity increases proportional to 13z , but beyond this
depth the turbulence properties agree with those of grid-generated turbulence without the
for our experiments and
lacking direct observations, we derive these estimates from Brumley and Jirka (1987).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
139
7
Over a depth interval of about 10% of the mean grid depth (Dgrid
= 11 cm) below the
water surface, the vertical turbulence velocity increases proportional to
150
Box 1 – Understanding cyanobacterial scum formation – learning from engineering studies
Understanding turbulence effects on cyanobacterial particles
There is some correspondence between the breakdown of oil slicks and the stability of
cyanobacterial scums under turbulent conditions in water, as investigated here. However, due
to the larger buoyancy of oil droplets and the large surface tension of oil, the destruction of
oil slicks occurs mainly under breaking waves (Delvigne and Sweeny, 1988; Tkalich and
Chan, 2002), which generate notably larger turbulence levels and occur at larger wind speeds
than the wind speeds which are relevant for the breakdown of scums in lakes. In terms of
turbulence levels, a better correspondence between the stability of cyanobacteria scums
created by buoyant cyanobacteria is the erosion of fluffy mud layers created by sinking mud
flocs by wind-wave generated turbulence in shallow lakes (Penning et al., 2013). Fluffy mud
layers flow and mix as a heavy fluid and can be treated as such using turbulence-mixing
models (e.g. Winterwerp et al., 2001). To understand the vertical profiles of cell and filament
concentrations in our experiments, we therefore treat the scum as a buoyant fluffy layer of
cyanobacteria, possible weakly adhered by mucus.
Whereas mud beds are eroded by water flowing over the bed, thereby creating
turbulence above the bed, cyanobacterial scums are eroded by wind blowing over the lake
surface, generating turbulence in the water column, which we attempted to recreate in our
experiments by using an oscillating grid below the scum layer. Detailed observations of
Brumley and Jirka (1987) describe the vertical mixing of particles with and without this grid-
generated turbulence near the water surface, which we have summarised in Supplementary
Figure 7.3. Vertical mixing is determined by the product of the mixing-length scale
proportional to depth (z) below the water surface and the rms of the vertical velocity (w),
written as |𝑤𝑤′(𝑧𝑧)|. For sufficiently large ratios of the cyanobacteria flotation velocity (wr) to
the vertical velocity variations |𝑤𝑤′(𝑧𝑧)|, cyanobacteria can concentrate near the water surface,
whereas at a low ratio they remain well mixed. We are thus interested in |𝑤𝑤′(𝑧𝑧)| for our
experiments and lacking direct observations, we derive these estimates from Brumley and
Jirka (1987).
Over a depth interval of about 10% of the mean grid depth (Dgrid = 11 cm) below the
water surface, the vertical turbulence velocity increases proportional to 13z , but beyond this
depth the turbulence properties agree with those of grid-generated turbulence without the
, but beyond this
depth the turbulence properties agree with those of grid-generated turbulence without the
damping effect of the water surface. From the profiles observed in Brumley and Jirka (1987),
the rms vertical velocity
151
damping effect of the water surface. From the profiles observed in Brumley and Jirka (1987),
the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| can be approximated by:
𝑧𝑧 ≤ 0.1 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ |𝑤𝑤′(𝑧𝑧)| = (𝜖𝜖10%𝑧𝑧)13 (B1)
where the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| is a function of the rate of turbulence-energy
dissipation at 10% of the mean grid depth below the water surface (𝜖𝜖10%).
For our oscillating grid experiments, 𝜖𝜖10% is estimated using equation (A1; Appendix
7.1) and is presented in Table 7.1, which results in a rms vertical velocity at 10% of the grid
depth of 1.2 mm s-1 (4 m h-1) for the lowest grid frequency and 5.6 mm s-1 (20 m h-1) for the
highest grid frequency (Table 7.1). For the cyanobacteria species in our experiments we
expect their flotation velocity to be notably less than 20 m h-1, as the highest flotation
velocity mentioned in literature (11.88 m h-1; Reynolds et al., 1987) was measured using
Microcystis aeruginosa with a radius of 200 µm, whereas the cyanobacteria used in our
experiments are considerably smaller (Supplementary Figure 7.1c, d).
Based on conditions for sediment remaining in suspension while flowing over a bed,
we would expect cyanobacteria with a flotation velocity of an order of magnitude less than 20
m h-1 to remain well mixed in the Limnotron up until the highest grid frequency (and should
never accumulate very close to the water surface). The previous analogy, however, is not
strict since in our experiments there is no turbulence production between the most upper grid
position and the water surface. Additionally, just below the water surface, molecular viscosity
damps the eddies advected upward from the grid towards the water surface. Hence, close to
the water surface there is an unsteady laminar zone. The thickness of this zone is expressed
by an equivalent Reynolds number z+ defined by 𝑧𝑧+ = |𝑤𝑤′(𝑧𝑧)| 𝑧𝑧/𝑣𝑣. We estimate the minimal
thickness of the laminar zone below the water surface of about 2 mm based on the estimate
𝑧𝑧+ < 10 for a laminar wall-boundary layer with [𝑤𝑤′(𝑧𝑧)] = 1.2 [𝑚𝑚𝑚𝑚 𝑠𝑠−1] for the lowest grid
frequency (see Table 7.1), z = 2 mm and 𝑣𝑣 = 1.10−6 [𝑚𝑚2𝑠𝑠−1] ≡ 1 [𝑚𝑚𝑚𝑚2𝑠𝑠−1]. In this
unsteady but laminar surface boundary layer the floating cyanobacteria are but weakly mixed.
Similar to the settling of sediment (as explained by Richardson and Zaki, 1954), the
accumulation of cyanobacteria near or at the water surface may be hindered by two
phenomena. Firstly, while approaching the water surface, the volumetric concentration of
cyanobacteria increases. The accumulation expels water causing a return flow that reduces
the net flotation velocity relative to a fixed reference frame. Richardson and Zaki (1954)
can be approximated by:
151
damping effect of the water surface. From the profiles observed in Brumley and Jirka (1987),
the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| can be approximated by:
𝑧𝑧 ≤ 0.1 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ |𝑤𝑤′(𝑧𝑧)| = (𝜖𝜖10%𝑧𝑧)13 (B1)
where the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| is a function of the rate of turbulence-energy
dissipation at 10% of the mean grid depth below the water surface (𝜖𝜖10%).
For our oscillating grid experiments, 𝜖𝜖10% is estimated using equation (A1; Appendix
7.1) and is presented in Table 7.1, which results in a rms vertical velocity at 10% of the grid
depth of 1.2 mm s-1 (4 m h-1) for the lowest grid frequency and 5.6 mm s-1 (20 m h-1) for the
highest grid frequency (Table 7.1). For the cyanobacteria species in our experiments we
expect their flotation velocity to be notably less than 20 m h-1, as the highest flotation
velocity mentioned in literature (11.88 m h-1; Reynolds et al., 1987) was measured using
Microcystis aeruginosa with a radius of 200 µm, whereas the cyanobacteria used in our
experiments are considerably smaller (Supplementary Figure 7.1c, d).
Based on conditions for sediment remaining in suspension while flowing over a bed,
we would expect cyanobacteria with a flotation velocity of an order of magnitude less than 20
m h-1 to remain well mixed in the Limnotron up until the highest grid frequency (and should
never accumulate very close to the water surface). The previous analogy, however, is not
strict since in our experiments there is no turbulence production between the most upper grid
position and the water surface. Additionally, just below the water surface, molecular viscosity
damps the eddies advected upward from the grid towards the water surface. Hence, close to
the water surface there is an unsteady laminar zone. The thickness of this zone is expressed
by an equivalent Reynolds number z+ defined by 𝑧𝑧+ = |𝑤𝑤′(𝑧𝑧)| 𝑧𝑧/𝑣𝑣. We estimate the minimal
thickness of the laminar zone below the water surface of about 2 mm based on the estimate
𝑧𝑧+ < 10 for a laminar wall-boundary layer with [𝑤𝑤′(𝑧𝑧)] = 1.2 [𝑚𝑚𝑚𝑚 𝑠𝑠−1] for the lowest grid
frequency (see Table 7.1), z = 2 mm and 𝑣𝑣 = 1.10−6 [𝑚𝑚2𝑠𝑠−1] ≡ 1 [𝑚𝑚𝑚𝑚2𝑠𝑠−1]. In this
unsteady but laminar surface boundary layer the floating cyanobacteria are but weakly mixed.
Similar to the settling of sediment (as explained by Richardson and Zaki, 1954), the
accumulation of cyanobacteria near or at the water surface may be hindered by two
phenomena. Firstly, while approaching the water surface, the volumetric concentration of
cyanobacteria increases. The accumulation expels water causing a return flow that reduces
the net flotation velocity relative to a fixed reference frame. Richardson and Zaki (1954)
(B1)
where the rms vertical velocity
151
damping effect of the water surface. From the profiles observed in Brumley and Jirka (1987),
the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| can be approximated by:
𝑧𝑧 ≤ 0.1 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ |𝑤𝑤′(𝑧𝑧)| = (𝜖𝜖10%𝑧𝑧)13 (B1)
where the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| is a function of the rate of turbulence-energy
dissipation at 10% of the mean grid depth below the water surface (𝜖𝜖10%).
For our oscillating grid experiments, 𝜖𝜖10% is estimated using equation (A1; Appendix
7.1) and is presented in Table 7.1, which results in a rms vertical velocity at 10% of the grid
depth of 1.2 mm s-1 (4 m h-1) for the lowest grid frequency and 5.6 mm s-1 (20 m h-1) for the
highest grid frequency (Table 7.1). For the cyanobacteria species in our experiments we
expect their flotation velocity to be notably less than 20 m h-1, as the highest flotation
velocity mentioned in literature (11.88 m h-1; Reynolds et al., 1987) was measured using
Microcystis aeruginosa with a radius of 200 µm, whereas the cyanobacteria used in our
experiments are considerably smaller (Supplementary Figure 7.1c, d).
Based on conditions for sediment remaining in suspension while flowing over a bed,
we would expect cyanobacteria with a flotation velocity of an order of magnitude less than 20
m h-1 to remain well mixed in the Limnotron up until the highest grid frequency (and should
never accumulate very close to the water surface). The previous analogy, however, is not
strict since in our experiments there is no turbulence production between the most upper grid
position and the water surface. Additionally, just below the water surface, molecular viscosity
damps the eddies advected upward from the grid towards the water surface. Hence, close to
the water surface there is an unsteady laminar zone. The thickness of this zone is expressed
by an equivalent Reynolds number z+ defined by 𝑧𝑧+ = |𝑤𝑤′(𝑧𝑧)| 𝑧𝑧/𝑣𝑣. We estimate the minimal
thickness of the laminar zone below the water surface of about 2 mm based on the estimate
𝑧𝑧+ < 10 for a laminar wall-boundary layer with [𝑤𝑤′(𝑧𝑧)] = 1.2 [𝑚𝑚𝑚𝑚 𝑠𝑠−1] for the lowest grid
frequency (see Table 7.1), z = 2 mm and 𝑣𝑣 = 1.10−6 [𝑚𝑚2𝑠𝑠−1] ≡ 1 [𝑚𝑚𝑚𝑚2𝑠𝑠−1]. In this
unsteady but laminar surface boundary layer the floating cyanobacteria are but weakly mixed.
Similar to the settling of sediment (as explained by Richardson and Zaki, 1954), the
accumulation of cyanobacteria near or at the water surface may be hindered by two
phenomena. Firstly, while approaching the water surface, the volumetric concentration of
cyanobacteria increases. The accumulation expels water causing a return flow that reduces
the net flotation velocity relative to a fixed reference frame. Richardson and Zaki (1954)
is a function of the rate of turbulence-energy
dissipation at 10% of the mean grid depth below the water surface (
151
damping effect of the water surface. From the profiles observed in Brumley and Jirka (1987),
the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| can be approximated by:
𝑧𝑧 ≤ 0.1 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ |𝑤𝑤′(𝑧𝑧)| = (𝜖𝜖10%𝑧𝑧)13 (B1)
where the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| is a function of the rate of turbulence-energy
dissipation at 10% of the mean grid depth below the water surface (𝜖𝜖10%).
For our oscillating grid experiments, 𝜖𝜖10% is estimated using equation (A1; Appendix
7.1) and is presented in Table 7.1, which results in a rms vertical velocity at 10% of the grid
depth of 1.2 mm s-1 (4 m h-1) for the lowest grid frequency and 5.6 mm s-1 (20 m h-1) for the
highest grid frequency (Table 7.1). For the cyanobacteria species in our experiments we
expect their flotation velocity to be notably less than 20 m h-1, as the highest flotation
velocity mentioned in literature (11.88 m h-1; Reynolds et al., 1987) was measured using
Microcystis aeruginosa with a radius of 200 µm, whereas the cyanobacteria used in our
experiments are considerably smaller (Supplementary Figure 7.1c, d).
Based on conditions for sediment remaining in suspension while flowing over a bed,
we would expect cyanobacteria with a flotation velocity of an order of magnitude less than 20
m h-1 to remain well mixed in the Limnotron up until the highest grid frequency (and should
never accumulate very close to the water surface). The previous analogy, however, is not
strict since in our experiments there is no turbulence production between the most upper grid
position and the water surface. Additionally, just below the water surface, molecular viscosity
damps the eddies advected upward from the grid towards the water surface. Hence, close to
the water surface there is an unsteady laminar zone. The thickness of this zone is expressed
by an equivalent Reynolds number z+ defined by 𝑧𝑧+ = |𝑤𝑤′(𝑧𝑧)| 𝑧𝑧/𝑣𝑣. We estimate the minimal
thickness of the laminar zone below the water surface of about 2 mm based on the estimate
𝑧𝑧+ < 10 for a laminar wall-boundary layer with [𝑤𝑤′(𝑧𝑧)] = 1.2 [𝑚𝑚𝑚𝑚 𝑠𝑠−1] for the lowest grid
frequency (see Table 7.1), z = 2 mm and 𝑣𝑣 = 1.10−6 [𝑚𝑚2𝑠𝑠−1] ≡ 1 [𝑚𝑚𝑚𝑚2𝑠𝑠−1]. In this
unsteady but laminar surface boundary layer the floating cyanobacteria are but weakly mixed.
Similar to the settling of sediment (as explained by Richardson and Zaki, 1954), the
accumulation of cyanobacteria near or at the water surface may be hindered by two
phenomena. Firstly, while approaching the water surface, the volumetric concentration of
cyanobacteria increases. The accumulation expels water causing a return flow that reduces
the net flotation velocity relative to a fixed reference frame. Richardson and Zaki (1954)
).
For our oscillating grid experiments,
151
damping effect of the water surface. From the profiles observed in Brumley and Jirka (1987),
the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| can be approximated by:
𝑧𝑧 ≤ 0.1 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ |𝑤𝑤′(𝑧𝑧)| = (𝜖𝜖10%𝑧𝑧)13 (B1)
where the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| is a function of the rate of turbulence-energy
dissipation at 10% of the mean grid depth below the water surface (𝜖𝜖10%).
For our oscillating grid experiments, 𝜖𝜖10% is estimated using equation (A1; Appendix
7.1) and is presented in Table 7.1, which results in a rms vertical velocity at 10% of the grid
depth of 1.2 mm s-1 (4 m h-1) for the lowest grid frequency and 5.6 mm s-1 (20 m h-1) for the
highest grid frequency (Table 7.1). For the cyanobacteria species in our experiments we
expect their flotation velocity to be notably less than 20 m h-1, as the highest flotation
velocity mentioned in literature (11.88 m h-1; Reynolds et al., 1987) was measured using
Microcystis aeruginosa with a radius of 200 µm, whereas the cyanobacteria used in our
experiments are considerably smaller (Supplementary Figure 7.1c, d).
Based on conditions for sediment remaining in suspension while flowing over a bed,
we would expect cyanobacteria with a flotation velocity of an order of magnitude less than 20
m h-1 to remain well mixed in the Limnotron up until the highest grid frequency (and should
never accumulate very close to the water surface). The previous analogy, however, is not
strict since in our experiments there is no turbulence production between the most upper grid
position and the water surface. Additionally, just below the water surface, molecular viscosity
damps the eddies advected upward from the grid towards the water surface. Hence, close to
the water surface there is an unsteady laminar zone. The thickness of this zone is expressed
by an equivalent Reynolds number z+ defined by 𝑧𝑧+ = |𝑤𝑤′(𝑧𝑧)| 𝑧𝑧/𝑣𝑣. We estimate the minimal
thickness of the laminar zone below the water surface of about 2 mm based on the estimate
𝑧𝑧+ < 10 for a laminar wall-boundary layer with [𝑤𝑤′(𝑧𝑧)] = 1.2 [𝑚𝑚𝑚𝑚 𝑠𝑠−1] for the lowest grid
frequency (see Table 7.1), z = 2 mm and 𝑣𝑣 = 1.10−6 [𝑚𝑚2𝑠𝑠−1] ≡ 1 [𝑚𝑚𝑚𝑚2𝑠𝑠−1]. In this
unsteady but laminar surface boundary layer the floating cyanobacteria are but weakly mixed.
Similar to the settling of sediment (as explained by Richardson and Zaki, 1954), the
accumulation of cyanobacteria near or at the water surface may be hindered by two
phenomena. Firstly, while approaching the water surface, the volumetric concentration of
cyanobacteria increases. The accumulation expels water causing a return flow that reduces
the net flotation velocity relative to a fixed reference frame. Richardson and Zaki (1954)
is estimated using equation (A1; Appendix
7.1) and is presented in Table 7.1, which results in a rms vertical velocity at 10% of the grid
depth of 1.2 mm s-1 (4 m h-1) for the lowest grid frequency and 5.6 mm s-1 (20 m h-1) for
the highest grid frequency (Table 7.1). For the cyanobacteria species in our experiments we
expect their flotation velocity to be notably less than 20 m h-1, as the highest flotation velocity
mentioned in literature (11.88 m h-1; Reynolds et al., 1987) was measured using Microcystis
aeruginosa with a radius of 200 µm, whereas the cyanobacteria used in our experiments are
considerably smaller (Supplementary Figure 7.1c, d).
Based on conditions for sediment remaining in suspension while flowing over a bed,
we would expect cyanobacteria with a flotation velocity of an order of magnitude less than
20 m h-1 to remain well mixed in the Limnotron up until the highest grid frequency (and
should never accumulate very close to the water surface). The previous analogy, however, is
not strict since in our experiments there is no turbulence production between the most upper
grid position and the water surface. Additionally, just below the water surface, molecular
viscosity damps the eddies advected upward from the grid towards the water surface. Hence,
close to the water surface there is an unsteady laminar zone. The thickness of this zone is
expressed by an equivalent Reynolds number z+ defined by
151
damping effect of the water surface. From the profiles observed in Brumley and Jirka (1987),
the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| can be approximated by:
𝑧𝑧 ≤ 0.1 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ |𝑤𝑤′(𝑧𝑧)| = (𝜖𝜖10%𝑧𝑧)13 (B1)
where the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| is a function of the rate of turbulence-energy
dissipation at 10% of the mean grid depth below the water surface (𝜖𝜖10%).
For our oscillating grid experiments, 𝜖𝜖10% is estimated using equation (A1; Appendix
7.1) and is presented in Table 7.1, which results in a rms vertical velocity at 10% of the grid
depth of 1.2 mm s-1 (4 m h-1) for the lowest grid frequency and 5.6 mm s-1 (20 m h-1) for the
highest grid frequency (Table 7.1). For the cyanobacteria species in our experiments we
expect their flotation velocity to be notably less than 20 m h-1, as the highest flotation
velocity mentioned in literature (11.88 m h-1; Reynolds et al., 1987) was measured using
Microcystis aeruginosa with a radius of 200 µm, whereas the cyanobacteria used in our
experiments are considerably smaller (Supplementary Figure 7.1c, d).
Based on conditions for sediment remaining in suspension while flowing over a bed,
we would expect cyanobacteria with a flotation velocity of an order of magnitude less than 20
m h-1 to remain well mixed in the Limnotron up until the highest grid frequency (and should
never accumulate very close to the water surface). The previous analogy, however, is not
strict since in our experiments there is no turbulence production between the most upper grid
position and the water surface. Additionally, just below the water surface, molecular viscosity
damps the eddies advected upward from the grid towards the water surface. Hence, close to
the water surface there is an unsteady laminar zone. The thickness of this zone is expressed
by an equivalent Reynolds number z+ defined by 𝑧𝑧+ = |𝑤𝑤′(𝑧𝑧)| 𝑧𝑧/𝑣𝑣. We estimate the minimal
thickness of the laminar zone below the water surface of about 2 mm based on the estimate
𝑧𝑧+ < 10 for a laminar wall-boundary layer with [𝑤𝑤′(𝑧𝑧)] = 1.2 [𝑚𝑚𝑚𝑚 𝑠𝑠−1] for the lowest grid
frequency (see Table 7.1), z = 2 mm and 𝑣𝑣 = 1.10−6 [𝑚𝑚2𝑠𝑠−1] ≡ 1 [𝑚𝑚𝑚𝑚2𝑠𝑠−1]. In this
unsteady but laminar surface boundary layer the floating cyanobacteria are but weakly mixed.
Similar to the settling of sediment (as explained by Richardson and Zaki, 1954), the
accumulation of cyanobacteria near or at the water surface may be hindered by two
phenomena. Firstly, while approaching the water surface, the volumetric concentration of
cyanobacteria increases. The accumulation expels water causing a return flow that reduces
the net flotation velocity relative to a fixed reference frame. Richardson and Zaki (1954)
. We estimate
the minimal thickness of the laminar zone below the water surface of about 2 mm based on
the estimate z+< 10 for a laminar wall-boundary layer with
151
damping effect of the water surface. From the profiles observed in Brumley and Jirka (1987),
the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| can be approximated by:
𝑧𝑧 ≤ 0.1 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ |𝑤𝑤′(𝑧𝑧)| = (𝜖𝜖10%𝑧𝑧)13 (B1)
where the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| is a function of the rate of turbulence-energy
dissipation at 10% of the mean grid depth below the water surface (𝜖𝜖10%).
For our oscillating grid experiments, 𝜖𝜖10% is estimated using equation (A1; Appendix
7.1) and is presented in Table 7.1, which results in a rms vertical velocity at 10% of the grid
depth of 1.2 mm s-1 (4 m h-1) for the lowest grid frequency and 5.6 mm s-1 (20 m h-1) for the
highest grid frequency (Table 7.1). For the cyanobacteria species in our experiments we
expect their flotation velocity to be notably less than 20 m h-1, as the highest flotation
velocity mentioned in literature (11.88 m h-1; Reynolds et al., 1987) was measured using
Microcystis aeruginosa with a radius of 200 µm, whereas the cyanobacteria used in our
experiments are considerably smaller (Supplementary Figure 7.1c, d).
Based on conditions for sediment remaining in suspension while flowing over a bed,
we would expect cyanobacteria with a flotation velocity of an order of magnitude less than 20
m h-1 to remain well mixed in the Limnotron up until the highest grid frequency (and should
never accumulate very close to the water surface). The previous analogy, however, is not
strict since in our experiments there is no turbulence production between the most upper grid
position and the water surface. Additionally, just below the water surface, molecular viscosity
damps the eddies advected upward from the grid towards the water surface. Hence, close to
the water surface there is an unsteady laminar zone. The thickness of this zone is expressed
by an equivalent Reynolds number z+ defined by 𝑧𝑧+ = |𝑤𝑤′(𝑧𝑧)| 𝑧𝑧/𝑣𝑣. We estimate the minimal
thickness of the laminar zone below the water surface of about 2 mm based on the estimate
𝑧𝑧+ < 10 for a laminar wall-boundary layer with [𝑤𝑤′(𝑧𝑧)] = 1.2 [𝑚𝑚𝑚𝑚 𝑠𝑠−1] for the lowest grid
frequency (see Table 7.1), z = 2 mm and 𝑣𝑣 = 1.10−6 [𝑚𝑚2𝑠𝑠−1] ≡ 1 [𝑚𝑚𝑚𝑚2𝑠𝑠−1]. In this
unsteady but laminar surface boundary layer the floating cyanobacteria are but weakly mixed.
Similar to the settling of sediment (as explained by Richardson and Zaki, 1954), the
accumulation of cyanobacteria near or at the water surface may be hindered by two
phenomena. Firstly, while approaching the water surface, the volumetric concentration of
cyanobacteria increases. The accumulation expels water causing a return flow that reduces
the net flotation velocity relative to a fixed reference frame. Richardson and Zaki (1954)
for the
lowest grid frequency (see Table 7.1), z = 2 mm and
151
damping effect of the water surface. From the profiles observed in Brumley and Jirka (1987),
the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| can be approximated by:
𝑧𝑧 ≤ 0.1 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ |𝑤𝑤′(𝑧𝑧)| = (𝜖𝜖10%𝑧𝑧)13 (B1)
where the rms vertical velocity |𝑤𝑤′(𝑧𝑧)| is a function of the rate of turbulence-energy
dissipation at 10% of the mean grid depth below the water surface (𝜖𝜖10%).
For our oscillating grid experiments, 𝜖𝜖10% is estimated using equation (A1; Appendix
7.1) and is presented in Table 7.1, which results in a rms vertical velocity at 10% of the grid
depth of 1.2 mm s-1 (4 m h-1) for the lowest grid frequency and 5.6 mm s-1 (20 m h-1) for the
highest grid frequency (Table 7.1). For the cyanobacteria species in our experiments we
expect their flotation velocity to be notably less than 20 m h-1, as the highest flotation
velocity mentioned in literature (11.88 m h-1; Reynolds et al., 1987) was measured using
Microcystis aeruginosa with a radius of 200 µm, whereas the cyanobacteria used in our
experiments are considerably smaller (Supplementary Figure 7.1c, d).
Based on conditions for sediment remaining in suspension while flowing over a bed,
we would expect cyanobacteria with a flotation velocity of an order of magnitude less than 20
m h-1 to remain well mixed in the Limnotron up until the highest grid frequency (and should
never accumulate very close to the water surface). The previous analogy, however, is not
strict since in our experiments there is no turbulence production between the most upper grid
position and the water surface. Additionally, just below the water surface, molecular viscosity
damps the eddies advected upward from the grid towards the water surface. Hence, close to
the water surface there is an unsteady laminar zone. The thickness of this zone is expressed
by an equivalent Reynolds number z+ defined by 𝑧𝑧+ = |𝑤𝑤′(𝑧𝑧)| 𝑧𝑧/𝑣𝑣. We estimate the minimal
thickness of the laminar zone below the water surface of about 2 mm based on the estimate
𝑧𝑧+ < 10 for a laminar wall-boundary layer with [𝑤𝑤′(𝑧𝑧)] = 1.2 [𝑚𝑚𝑚𝑚 𝑠𝑠−1] for the lowest grid
frequency (see Table 7.1), z = 2 mm and 𝑣𝑣 = 1.10−6 [𝑚𝑚2𝑠𝑠−1] ≡ 1 [𝑚𝑚𝑚𝑚2𝑠𝑠−1]. In this
unsteady but laminar surface boundary layer the floating cyanobacteria are but weakly mixed.
Similar to the settling of sediment (as explained by Richardson and Zaki, 1954), the
accumulation of cyanobacteria near or at the water surface may be hindered by two
phenomena. Firstly, while approaching the water surface, the volumetric concentration of
cyanobacteria increases. The accumulation expels water causing a return flow that reduces
the net flotation velocity relative to a fixed reference frame. Richardson and Zaki (1954)
.
In this unsteady but laminar surface boundary layer the floating cyanobacteria are but weakly
mixed.
Similar to the settling of sediment (as explained by Richardson and Zaki, 1954),
the accumulation of cyanobacteria near or at the water surface may be hindered by two
phenomena. Firstly, while approaching the water surface, the volumetric concentration of
cyanobacteria increases. The accumulation expels water causing a return flow that reduces
the net flotation velocity relative to a fixed reference frame. Richardson and Zaki (1954)
explored the latter theoretically and experimentally for settling solid particles, called
hindered settling. For rising cyanobacteria the analogy is obvious and we introduce the term
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
140
hindered rising. In addition, the accumulation increases the probability of mutual collisions
and contacts between cyanobacteria colonies or filaments, possibly imbedded in mucilage,
finally causing their arrest into a scum layer. After a night of accumulation, the observed
concentrations at the start of our experiments are indicated at 0 Hz in Figure 7.3a and 7.3g.
With a typical size of 0.02 mm of the filamentous colonies of A. flos-aquae (Ploug et al., 2010)
and 5.106 filaments mL-1 2 mm below the water surface (Figure 7.3c) their volume fraction
is about 2%, thus very dilute. With a typical size of W. naegeliana colonies of 30 - 180 µm
(Wilk-Wozniak et al., 2003, Supplementary Figure 7.1c, d) and 104 colonies mL-1 sampled
2 mm below the water surface (Figure 7.3i), their volume fraction would yield just about
4%. Note that the 1 mm diameter syringe clogged when sampling closer to the water surface
(< 2 mm), indicating a further increase in volume fraction towards the water surface. Some
substantiation of the existence of low volume fractions can be derived from Alldredge and
Gotschalk (1989) and Smayda (1971), presenting a proportionality between the settling
velocity of aggregates and their size while Stokes Law would predict a quadratic dependency
on size. From the latter follows that the volumetric cell fraction decreases with the inverse of
the aggregate size and to the dilution levels estimated above.
Numerical simulation using turbulence models
For the purpose of demonstration we present numerical simulations based on the previous
analyses, for details of the numerical methodology see (Aparicio Medrano et al., 2013). We
combine the hindered rising phenomenon and the clogging by contacts between filaments
or colonies up to the accumulation of a scum at concentration nscum
into the single formula
152
explored the latter theoretically and experimentally for settling solid particles, called hindered
settling. For rising cyanobacteria the analogy is obvious and we introduce the term hindered
rising. In addition, the accumulation increases the probability of mutual collisions and
contacts between cyanobacteria colonies or filaments, possibly imbedded in mucilage, finally
causing their arrest into a scum layer. After a night of accumulation, the observed
concentrations at the start of our experiments are indicated at 0 Hz in Figure 7.3a and 7.3g.
With a typical size of 0.02 mm of the filamentous colonies of A. flos-aquae (Ploug et al.,
2010) and 5.106 filaments mL-1 2 mm below the water surface (Figure 7.3c) their volume
fraction is about 2%, thus very dilute. With a typical size of W. naegeliana colonies of 30 -
180 µm (Wilk-Wozniak et al., 2003, Supplementary Figure 7.1c, d) and 104 colonies mL-1
sampled 2 mm below the water surface (Figure 7.3i), their volume fraction would yield just
about 4%. Note that the 1 mm diameter syringe clogged when sampling closer to the water
surface (< 2 mm), indicating a further increase in volume fraction towards the water surface.
Some substantiation of the existence of low volume fractions can be derived from Alldredge
and Gotschalk (1989) and Smayda (1971), presenting a proportionality between the settling
velocity of aggregates and their size while Stokes Law would predict a quadratic dependency
on size. From the latter follows that the volumetric cell fraction decreases with the inverse of
the aggregate size and to the dilution levels estimated above.
Numerical simulation using turbulence models
For the purpose of demonstration we present numerical simulations based on the
previous analyses, for details of the numerical methodology see (Aparicio Medrano et al.,
2013). We combine the hindered rising phenomenon and the clogging by contacts between
filaments or colonies up to the accumulation of a scum at concentration nscum into the single
formula
𝑤𝑤𝑟𝑟(𝑛𝑛) = 𝑤𝑤𝑟𝑟,𝑜𝑜(1 − 𝑛𝑛/𝑛𝑛𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠)4.65 (B2)
where wr,0 is the flotation velocity of solitary filaments or colonies and n their volumetric
concentration experiencing the reduced flotation velocity ѡr(n) and the power 4.65 taken
from Richardson and Zaki (1954).
For correspondence between the simulation and the oscillating grid experiments, the
vertical eddy-diffusivity at 10% of the grid depth is taken as reference. The vertical
(B2)
where wr,0
is the flotation velocity of solitary filaments or colonies and n their volumetric
concentration experiencing the reduced flotation velocity wr(n) and the power 4.65 taken
from Richardson and Zaki (1954).
For correspondence between the simulation and the oscillating grid experiments,
the vertical eddy-diffusivity at 10% of the grid depth is taken as reference. The vertical
turbulence mixing coefficient
153
turbulence mixing coefficient Γ𝑇𝑇(𝑧𝑧) readily follows from equation (B1) by multiplication of
the mixing-length scale (𝜅𝜅𝑧𝑧) yielding:
𝑧𝑧 ≤ 0.1𝑧𝑧𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ Γ𝑇𝑇(𝑧𝑧) = 𝜅𝜅𝑧𝑧|𝑤𝑤 ′(𝑧𝑧)| = 𝜅𝜅𝑧𝑧43𝜖𝜖10%
13 (B3)
with Von Kármán coefficient 𝜅𝜅 = 0.40. At 10% of the grid depth the vertical eddy-
diffusivity (B3) is about 5.10-6 m2 s-1 for the lowest grid frequency and about 25.10-6 m2 s-1
for the highest grid frequency.
The numerical simulation covers a depth of 1 meter using 2,000 non-equidistant
levels of which 136 levels over the top 11 mm, starting with 50 µm thickness at the water
surface. The bed floor is an impermeable wall and the water surface shear free. Turbulence is
generated by imposing a current, yielding the desired eddy-diffusivity profile from equation
(B3) as by grid-generated turbulence below the water surface (Figure 7.4a).
A scum layer is initiated in the top 1 mm with a reference concentration of 1000 (in
arbitrary units) so that the well-mixed concentration would be 1. The simulation runs the first
12 hours without turbulence generation (similar to the overnight stagnant case) and
subsequently the turbulence mixing is stepwise increased in agreement with the oscillating
grid frequency and eddy-diffusivity, using equation (B3) as reference. The first increment,
equivalent to an increase of turbulence frequency of 0.53 Hz in the Limnotron experiments,
lasts 2 hours allowing for the spin-up of turbulence over the depth but the subsequent
frequency increments last a single hour, as in the Limnotron experiments; the numerical time
step is 10 seconds.
Figure 7.4b shows the vertical profiles of concentration for fictional species rising in
dilute suspension with wr,0 = 0.5 m h-1 (black lines) or wr,0 = 5 m h-1 (red lines) at the two
highest grid frequencies (ƒ= 1.7 and 2.4 Hz). Note that for the lower grid frequency (dashed
lines) the concentration at 2 mm or deeper is larger than at higher frequency (solid lines).
The explanation is that turbulence erodes the contents of the scum layer downward thereby
reducing the scum layer concentration proper but increasing the concentration below the
scum layer by mass conservation. The slower rising species (black lines in right panel) are
well mixed at the highest grid frequencies and approach the depth-averaged magnitude of 1.
At the highest grid frequency the simulations indicate that the surface (scum) concentration
can be reduced by more than one order of magnitude depending on the unhindered flotation
velocity wr,0 of 5 m h-1 or 0.5 m h-1. For 0.5 m h-1 unhindered flotation velocity, the
readily follows from equation (B1) by multiplication of
the mixing-length scale (
153
turbulence mixing coefficient Γ𝑇𝑇(𝑧𝑧) readily follows from equation (B1) by multiplication of
the mixing-length scale (𝜅𝜅𝑧𝑧) yielding:
𝑧𝑧 ≤ 0.1𝑧𝑧𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ Γ𝑇𝑇(𝑧𝑧) = 𝜅𝜅𝑧𝑧|𝑤𝑤 ′(𝑧𝑧)| = 𝜅𝜅𝑧𝑧43𝜖𝜖10%
13 (B3)
with Von Kármán coefficient 𝜅𝜅 = 0.40. At 10% of the grid depth the vertical eddy-
diffusivity (B3) is about 5.10-6 m2 s-1 for the lowest grid frequency and about 25.10-6 m2 s-1
for the highest grid frequency.
The numerical simulation covers a depth of 1 meter using 2,000 non-equidistant
levels of which 136 levels over the top 11 mm, starting with 50 µm thickness at the water
surface. The bed floor is an impermeable wall and the water surface shear free. Turbulence is
generated by imposing a current, yielding the desired eddy-diffusivity profile from equation
(B3) as by grid-generated turbulence below the water surface (Figure 7.4a).
A scum layer is initiated in the top 1 mm with a reference concentration of 1000 (in
arbitrary units) so that the well-mixed concentration would be 1. The simulation runs the first
12 hours without turbulence generation (similar to the overnight stagnant case) and
subsequently the turbulence mixing is stepwise increased in agreement with the oscillating
grid frequency and eddy-diffusivity, using equation (B3) as reference. The first increment,
equivalent to an increase of turbulence frequency of 0.53 Hz in the Limnotron experiments,
lasts 2 hours allowing for the spin-up of turbulence over the depth but the subsequent
frequency increments last a single hour, as in the Limnotron experiments; the numerical time
step is 10 seconds.
Figure 7.4b shows the vertical profiles of concentration for fictional species rising in
dilute suspension with wr,0 = 0.5 m h-1 (black lines) or wr,0 = 5 m h-1 (red lines) at the two
highest grid frequencies (ƒ= 1.7 and 2.4 Hz). Note that for the lower grid frequency (dashed
lines) the concentration at 2 mm or deeper is larger than at higher frequency (solid lines).
The explanation is that turbulence erodes the contents of the scum layer downward thereby
reducing the scum layer concentration proper but increasing the concentration below the
scum layer by mass conservation. The slower rising species (black lines in right panel) are
well mixed at the highest grid frequencies and approach the depth-averaged magnitude of 1.
At the highest grid frequency the simulations indicate that the surface (scum) concentration
can be reduced by more than one order of magnitude depending on the unhindered flotation
velocity wr,0 of 5 m h-1 or 0.5 m h-1. For 0.5 m h-1 unhindered flotation velocity, the
) yielding:
153
turbulence mixing coefficient Γ𝑇𝑇(𝑧𝑧) readily follows from equation (B1) by multiplication of
the mixing-length scale (𝜅𝜅𝑧𝑧) yielding:
𝑧𝑧 ≤ 0.1𝑧𝑧𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ Γ𝑇𝑇(𝑧𝑧) = 𝜅𝜅𝑧𝑧|𝑤𝑤 ′(𝑧𝑧)| = 𝜅𝜅𝑧𝑧43𝜖𝜖10%
13 (B3)
with Von Kármán coefficient 𝜅𝜅 = 0.40. At 10% of the grid depth the vertical eddy-
diffusivity (B3) is about 5.10-6 m2 s-1 for the lowest grid frequency and about 25.10-6 m2 s-1
for the highest grid frequency.
The numerical simulation covers a depth of 1 meter using 2,000 non-equidistant
levels of which 136 levels over the top 11 mm, starting with 50 µm thickness at the water
surface. The bed floor is an impermeable wall and the water surface shear free. Turbulence is
generated by imposing a current, yielding the desired eddy-diffusivity profile from equation
(B3) as by grid-generated turbulence below the water surface (Figure 7.4a).
A scum layer is initiated in the top 1 mm with a reference concentration of 1000 (in
arbitrary units) so that the well-mixed concentration would be 1. The simulation runs the first
12 hours without turbulence generation (similar to the overnight stagnant case) and
subsequently the turbulence mixing is stepwise increased in agreement with the oscillating
grid frequency and eddy-diffusivity, using equation (B3) as reference. The first increment,
equivalent to an increase of turbulence frequency of 0.53 Hz in the Limnotron experiments,
lasts 2 hours allowing for the spin-up of turbulence over the depth but the subsequent
frequency increments last a single hour, as in the Limnotron experiments; the numerical time
step is 10 seconds.
Figure 7.4b shows the vertical profiles of concentration for fictional species rising in
dilute suspension with wr,0 = 0.5 m h-1 (black lines) or wr,0 = 5 m h-1 (red lines) at the two
highest grid frequencies (ƒ= 1.7 and 2.4 Hz). Note that for the lower grid frequency (dashed
lines) the concentration at 2 mm or deeper is larger than at higher frequency (solid lines).
The explanation is that turbulence erodes the contents of the scum layer downward thereby
reducing the scum layer concentration proper but increasing the concentration below the
scum layer by mass conservation. The slower rising species (black lines in right panel) are
well mixed at the highest grid frequencies and approach the depth-averaged magnitude of 1.
At the highest grid frequency the simulations indicate that the surface (scum) concentration
can be reduced by more than one order of magnitude depending on the unhindered flotation
velocity wr,0 of 5 m h-1 or 0.5 m h-1. For 0.5 m h-1 unhindered flotation velocity, the
(B3)
with Von Kármán coefficient
153
turbulence mixing coefficient Γ𝑇𝑇(𝑧𝑧) readily follows from equation (B1) by multiplication of
the mixing-length scale (𝜅𝜅𝑧𝑧) yielding:
𝑧𝑧 ≤ 0.1𝑧𝑧𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 ∶ Γ𝑇𝑇(𝑧𝑧) = 𝜅𝜅𝑧𝑧|𝑤𝑤 ′(𝑧𝑧)| = 𝜅𝜅𝑧𝑧43𝜖𝜖10%
13 (B3)
with Von Kármán coefficient 𝜅𝜅 = 0.40. At 10% of the grid depth the vertical eddy-
diffusivity (B3) is about 5.10-6 m2 s-1 for the lowest grid frequency and about 25.10-6 m2 s-1
for the highest grid frequency.
The numerical simulation covers a depth of 1 meter using 2,000 non-equidistant
levels of which 136 levels over the top 11 mm, starting with 50 µm thickness at the water
surface. The bed floor is an impermeable wall and the water surface shear free. Turbulence is
generated by imposing a current, yielding the desired eddy-diffusivity profile from equation
(B3) as by grid-generated turbulence below the water surface (Figure 7.4a).
A scum layer is initiated in the top 1 mm with a reference concentration of 1000 (in
arbitrary units) so that the well-mixed concentration would be 1. The simulation runs the first
12 hours without turbulence generation (similar to the overnight stagnant case) and
subsequently the turbulence mixing is stepwise increased in agreement with the oscillating
grid frequency and eddy-diffusivity, using equation (B3) as reference. The first increment,
equivalent to an increase of turbulence frequency of 0.53 Hz in the Limnotron experiments,
lasts 2 hours allowing for the spin-up of turbulence over the depth but the subsequent
frequency increments last a single hour, as in the Limnotron experiments; the numerical time
step is 10 seconds.
Figure 7.4b shows the vertical profiles of concentration for fictional species rising in
dilute suspension with wr,0 = 0.5 m h-1 (black lines) or wr,0 = 5 m h-1 (red lines) at the two
highest grid frequencies (ƒ= 1.7 and 2.4 Hz). Note that for the lower grid frequency (dashed
lines) the concentration at 2 mm or deeper is larger than at higher frequency (solid lines).
The explanation is that turbulence erodes the contents of the scum layer downward thereby
reducing the scum layer concentration proper but increasing the concentration below the
scum layer by mass conservation. The slower rising species (black lines in right panel) are
well mixed at the highest grid frequencies and approach the depth-averaged magnitude of 1.
At the highest grid frequency the simulations indicate that the surface (scum) concentration
can be reduced by more than one order of magnitude depending on the unhindered flotation
velocity wr,0 of 5 m h-1 or 0.5 m h-1. For 0.5 m h-1 unhindered flotation velocity, the
. At 10% of the grid depth the vertical eddy-
diffusivity (B3) is about 5.10-6 m2 s-1 for the lowest grid frequency and about 25.10-6 m2 s-1
for the highest grid frequency.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
141
7
The numerical simulation covers a depth of 1 meter using 2,000 non-equidistant levels
of which 136 levels over the top 11 mm, starting with 50 µm thickness at the water surface.
The bed floor is an impermeable wall and the water surface shear free. Turbulence is generated
by imposing a current, yielding the desired eddy-diffusivity profile from equation (B3) as by
grid-generated turbulence below the water surface (Figure 7.4a).
A scum layer is initiated in the top 1 mm with a reference concentration of 1000 (in
arbitrary units) so that the well-mixed concentration would be 1. The simulation runs the
first 12 hours without turbulence generation (similar to the overnight stagnant case) and
subsequently the turbulence mixing is stepwise increased in agreement with the oscillating
grid frequency and eddy-diffusivity, using equation (B3) as reference. The first increment,
equivalent to an increase of turbulence frequency of 0.53 Hz in the Limnotron experiments,
lasts 2 hours allowing for the spin-up of turbulence over the depth but the subsequent
frequency increments last a single hour, as in the Limnotron experiments; the numerical time
step is 10 seconds.
Figure 7.4b shows the vertical profiles of concentration for fictional species rising in
dilute suspension with wr,0
= 0.5 m h-1 (black lines) or wr,0
= 5 m h-1 (red lines) at the two
highest grid frequencies (ƒ= 1.7 and 2.4 Hz). Note that for the lower grid frequency (dashed
lines) the concentration at 2 mm or deeper is larger than at higher frequency (solid lines).
The explanation is that turbulence erodes the contents of the scum layer downward thereby
reducing the scum layer concentration proper but increasing the concentration below the
scum layer by mass conservation. The slower rising species (black lines in right panel) are
well mixed at the highest grid frequencies and approach the depth-averaged magnitude of 1.
At the highest grid frequency the simulations indicate that the surface (scum) concentration
can be reduced by more than one order of magnitude depending on the unhindered
flotation velocity wr,0
of 5 m h-1 or 0.5 m h-1. For 0.5 m h-1 unhindered flotation velocity,
the concentrations observed at 2, 5, 10 and 15 mm are nearly equal but for 5 m h-1 still
strongly stratified. Comparing these vertical profiles with our experiments, the slow (0.5 m
h-1) and fast (5 m h-1) floating predictions are similar to the observations of Aphanizomenon
and Woronichinia, respectively (see the photographs, Figure 7.2), provided Aphanizomenon
rises faster than Woronichinia.
Figure 7.5a presents the evolution of the vertical concentration profile of species with 5
m h-1 unhindered flotation velocity yielding a scum layer much thinner than 1 mm at the
highest grid frequency (ref. 15-16 h). With decreasing grid frequency the concentration
profile re-establishes in close symmetry to its earlier breakdown at increasing grid frequency.
We assume in these calculations, however, that filament or colony size is not affected by the
increase in grid frequencies. This temporal pattern changes notably when using 0.5 m h-1
unhindered settling velocity. Figure 7.5b shows a long temporal delay time for the scum to
return both due to the slowly rising material, spread over the 1 m depth, and because the
grid turbulence decays after the final experiments (ref 18-19 h). Note the simulated temporal
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
142
evolution of the concentration 2 mm below the water surface, i.e. the blue lines in upper
the panels of Figure 7.5a and 7.5b. Initially, the concentration at 2 mm increases by the
downward mixing of the scum layer at lower grid frequency. Subsequently, the concentration
at 2 mm decreases with higher grid frequencies as the species are more homogenously mixed
over the depth of the Limnotron. These complex concentration patterns in time and depth
correspond to some extent with the observations.
Figure 7.4 – The simulated eddy-diffusivity profile (a) with blue lines for ƒ = 2.4 Hz, dashed blue for ƒ = 1.7 Hz, and the theoretical profile in red for ƒ = 2.4 Hz, see equation (B3). The vertical concentration profiles of theoretical species (b) with 0.5 m h-1 (black lines) and 5 m h-1 (red lines) flotation velocity at zero concentration, full lines for ƒ = 2.4 Hz, dashed for ƒ = 1.7 Hz.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
143
7Figure 7.5a – Erosion of the 2 mm thick scum layer by a stepwise increasing grid frequency or eddy-diffusivity (central panel) at 5 m h-1 unhindered rising velocity (w
r,0). At the highest grid frequency (ref.
15-16 h) most of the scum layer is eroded from below and distributed over the simulation depth of 1 m.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
144
Figure 7.5b – Erosion of the 2 mm thick scum layer by a stepwise increasing grid frequency or eddy-diffusivity (central panel) at 0.5 m h-1 unhindered rising velocity (w
r,0). Soon after starting the grid
oscillations (ref. 13 h) the scum layer has been eroded entirely and distributed over the simulation depth of 1 m. Note the slow re-establishment of the scum layer after arresting grid oscillations.
ACKNOWLEDGEMENTS
We are grateful to Miguel Dionisio Pires and Hans Los for their valuable theoretical insights and
useful discussions. We would also like to thank Dennis Waasdorp and Nico Helmsing for their
practical assistance with the Limnotrons and Erik Reichman, Nico Helmsing, and Dilara Deniz
for performing multiple chemical analyses. This study was funded by STOWA and supported by
a grant from Deltares for data processing.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
145
7
APPENDIX 7.1
Methods Chapter 7. Estimation of turbulence properties and equivalent wind speeds
In this section we exploit the turbulence velocity observations of Fernando and DaSilva (1993)
and their empirical relations to grid size and frequency, derived in similar grid experiments but
without an air-water interface. Reversely, given the grid size and grid frequency in our Limnotron
experiments we estimate turbulence properties for vertical mixing, using the Kolmogorov length
scale of maximum fine-scale shearing (ηk) and the rms of vertical velocity
157
Appendix 7.1
Methods Chapter 7. Estimation of turbulence properties and equivalent wind speeds
In this section we exploit the turbulence velocity observations of Fernando and DaSilva
(1993) and their empirical relations to grid size and frequency, derived in similar grid
experiments but without an air-water interface. Reversely, given the grid size and grid
frequency in our Limnotron experiments we estimate turbulence properties for vertical
mixing, using the Kolmogorov length scale of maximum fine-scale shearing (ηk) and the rms
of vertical velocity |𝑤𝑤′(𝑧𝑧)|. In addition, we relate the turbulence level at 10 mm below the
water surface to the corresponding wind speed over a lake.
In general, we follow the approach of O’Brien et al. (2004) by first estimating 𝜖𝜖, the
rate of viscous dissipation of turbulent kinetic energy, by:
𝜖𝜖 = 1𝛽𝛽 (2𝐶𝐶12+𝐶𝐶22
3 )32 𝑀𝑀
32 𝑆𝑆
92 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 = 𝛼𝛼 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 for z > 10 mm (A1)
here M is the mesh size of the grid (m), S the stroke length (m), (Dgrid - z) the distance from
the grid (m), and ƒ the grid oscillating frequency (Hz). The constant value β equals 0.1 and
the values C1 and C2 are derived from the geometry of the grid and are 0.18 and 0.22,
respectively (Fernando and DeSilva, 1993; O’Brien et al., 2004). With a mesh size of 0.06 m
and a stroke length of 0.028 m, α equals 1.11 × 10-10 m6. Brumley and Jirka (1987) analyzed
that Equation (A1) is applicable to 10% of the grid depth below the water surface, thus to z =
10 mm. Closer to the water surface the length scales and the vertical velocity variance are
damped by the air-water interface.
Subsequently the acquired energy dissipation from Equation (A1) was used to
calculate the Kolmogorov length scale (ηk) for each grid frequency in Equation (A2)
(O’Brien et al., 2004):
𝜂𝜂𝐾𝐾 = (𝑣𝑣3
𝜖𝜖 )14 = (𝑣𝑣3
𝛼𝛼 )14 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 − 𝑧𝑧
ƒ34
for z > 10 mm (A2)
. In addition, we
relate the turbulence level at 10 mm below the water surface to the corresponding wind speed over
a lake.
In general, we follow the approach of O’Brien et al. (2004) by first estimating
157
Appendix 7.1
Methods Chapter 7. Estimation of turbulence properties and equivalent wind speeds
In this section we exploit the turbulence velocity observations of Fernando and DaSilva
(1993) and their empirical relations to grid size and frequency, derived in similar grid
experiments but without an air-water interface. Reversely, given the grid size and grid
frequency in our Limnotron experiments we estimate turbulence properties for vertical
mixing, using the Kolmogorov length scale of maximum fine-scale shearing (ηk) and the rms
of vertical velocity |𝑤𝑤′(𝑧𝑧)|. In addition, we relate the turbulence level at 10 mm below the
water surface to the corresponding wind speed over a lake.
In general, we follow the approach of O’Brien et al. (2004) by first estimating 𝜖𝜖, the
rate of viscous dissipation of turbulent kinetic energy, by:
𝜖𝜖 = 1𝛽𝛽 (2𝐶𝐶12+𝐶𝐶22
3 )32 𝑀𝑀
32 𝑆𝑆
92 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 = 𝛼𝛼 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 for z > 10 mm (A1)
here M is the mesh size of the grid (m), S the stroke length (m), (Dgrid - z) the distance from
the grid (m), and ƒ the grid oscillating frequency (Hz). The constant value β equals 0.1 and
the values C1 and C2 are derived from the geometry of the grid and are 0.18 and 0.22,
respectively (Fernando and DeSilva, 1993; O’Brien et al., 2004). With a mesh size of 0.06 m
and a stroke length of 0.028 m, α equals 1.11 × 10-10 m6. Brumley and Jirka (1987) analyzed
that Equation (A1) is applicable to 10% of the grid depth below the water surface, thus to z =
10 mm. Closer to the water surface the length scales and the vertical velocity variance are
damped by the air-water interface.
Subsequently the acquired energy dissipation from Equation (A1) was used to
calculate the Kolmogorov length scale (ηk) for each grid frequency in Equation (A2)
(O’Brien et al., 2004):
𝜂𝜂𝐾𝐾 = (𝑣𝑣3
𝜖𝜖 )14 = (𝑣𝑣3
𝛼𝛼 )14 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 − 𝑧𝑧
ƒ34
for z > 10 mm (A2)
, the rate of
viscous dissipation of turbulent kinetic energy, by:
157
Appendix 7.1
Methods Chapter 7. Estimation of turbulence properties and equivalent wind speeds
In this section we exploit the turbulence velocity observations of Fernando and DaSilva
(1993) and their empirical relations to grid size and frequency, derived in similar grid
experiments but without an air-water interface. Reversely, given the grid size and grid
frequency in our Limnotron experiments we estimate turbulence properties for vertical
mixing, using the Kolmogorov length scale of maximum fine-scale shearing (ηk) and the rms
of vertical velocity |𝑤𝑤′(𝑧𝑧)|. In addition, we relate the turbulence level at 10 mm below the
water surface to the corresponding wind speed over a lake.
In general, we follow the approach of O’Brien et al. (2004) by first estimating 𝜖𝜖, the
rate of viscous dissipation of turbulent kinetic energy, by:
𝜖𝜖 = 1𝛽𝛽 (2𝐶𝐶12+𝐶𝐶22
3 )32 𝑀𝑀
32 𝑆𝑆
92 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 = 𝛼𝛼 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 for z > 10 mm (A1)
here M is the mesh size of the grid (m), S the stroke length (m), (Dgrid - z) the distance from
the grid (m), and ƒ the grid oscillating frequency (Hz). The constant value β equals 0.1 and
the values C1 and C2 are derived from the geometry of the grid and are 0.18 and 0.22,
respectively (Fernando and DeSilva, 1993; O’Brien et al., 2004). With a mesh size of 0.06 m
and a stroke length of 0.028 m, α equals 1.11 × 10-10 m6. Brumley and Jirka (1987) analyzed
that Equation (A1) is applicable to 10% of the grid depth below the water surface, thus to z =
10 mm. Closer to the water surface the length scales and the vertical velocity variance are
damped by the air-water interface.
Subsequently the acquired energy dissipation from Equation (A1) was used to
calculate the Kolmogorov length scale (ηk) for each grid frequency in Equation (A2)
(O’Brien et al., 2004):
𝜂𝜂𝐾𝐾 = (𝑣𝑣3
𝜖𝜖 )14 = (𝑣𝑣3
𝛼𝛼 )14 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 − 𝑧𝑧
ƒ34
for z > 10 mm (A2)
(A1)
here M is the mesh size of the grid (m), S the stroke length (m), (Dgrid
- z) the distance from the grid
(m), and ƒ the grid oscillating frequency (Hz). The constant value β equals 0.1 and the values C1
and C2 are derived from the geometry of the grid and are 0.18 and 0.22, respectively (Fernando and
DeSilva, 1993; O’Brien et al., 2004). With a mesh size of 0.06 m and a stroke length of 0.028 m, α
equals 1.11 × 10-10 m6. Brumley and Jirka (1987) analyzed that Equation (A1) is applicable to 10%
of the grid depth below the water surface, thus to z = 10 mm. Closer to the water surface the length
scales and the vertical velocity variance are damped by the air-water interface.
Subsequently the acquired energy dissipation from Equation (A1) was used to calculate the
Kolmogorov length scale (ηk) for each grid frequency in Equation (A2) (O’Brien et al., 2004):
157
Appendix 7.1
Methods Chapter 7. Estimation of turbulence properties and equivalent wind speeds
In this section we exploit the turbulence velocity observations of Fernando and DaSilva
(1993) and their empirical relations to grid size and frequency, derived in similar grid
experiments but without an air-water interface. Reversely, given the grid size and grid
frequency in our Limnotron experiments we estimate turbulence properties for vertical
mixing, using the Kolmogorov length scale of maximum fine-scale shearing (ηk) and the rms
of vertical velocity |𝑤𝑤′(𝑧𝑧)|. In addition, we relate the turbulence level at 10 mm below the
water surface to the corresponding wind speed over a lake.
In general, we follow the approach of O’Brien et al. (2004) by first estimating 𝜖𝜖, the
rate of viscous dissipation of turbulent kinetic energy, by:
𝜖𝜖 = 1𝛽𝛽 (2𝐶𝐶12+𝐶𝐶22
3 )32 𝑀𝑀
32 𝑆𝑆
92 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 = 𝛼𝛼 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 for z > 10 mm (A1)
here M is the mesh size of the grid (m), S the stroke length (m), (Dgrid - z) the distance from
the grid (m), and ƒ the grid oscillating frequency (Hz). The constant value β equals 0.1 and
the values C1 and C2 are derived from the geometry of the grid and are 0.18 and 0.22,
respectively (Fernando and DeSilva, 1993; O’Brien et al., 2004). With a mesh size of 0.06 m
and a stroke length of 0.028 m, α equals 1.11 × 10-10 m6. Brumley and Jirka (1987) analyzed
that Equation (A1) is applicable to 10% of the grid depth below the water surface, thus to z =
10 mm. Closer to the water surface the length scales and the vertical velocity variance are
damped by the air-water interface.
Subsequently the acquired energy dissipation from Equation (A1) was used to
calculate the Kolmogorov length scale (ηk) for each grid frequency in Equation (A2)
(O’Brien et al., 2004):
𝜂𝜂𝐾𝐾 = (𝑣𝑣3
𝜖𝜖 )14 = (𝑣𝑣3
𝛼𝛼 )14 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 − 𝑧𝑧
ƒ34
for z > 10 mm (A2)
(A2)
where ν is the kinematic viscosity of the fluid (water = 1.004 × 10-6), (Dgrid
- z) the distance from the
grid (m), ƒ the grid oscillating frequency (Hz),
157
Appendix 7.1
Methods Chapter 7. Estimation of turbulence properties and equivalent wind speeds
In this section we exploit the turbulence velocity observations of Fernando and DaSilva
(1993) and their empirical relations to grid size and frequency, derived in similar grid
experiments but without an air-water interface. Reversely, given the grid size and grid
frequency in our Limnotron experiments we estimate turbulence properties for vertical
mixing, using the Kolmogorov length scale of maximum fine-scale shearing (ηk) and the rms
of vertical velocity |𝑤𝑤′(𝑧𝑧)|. In addition, we relate the turbulence level at 10 mm below the
water surface to the corresponding wind speed over a lake.
In general, we follow the approach of O’Brien et al. (2004) by first estimating 𝜖𝜖, the
rate of viscous dissipation of turbulent kinetic energy, by:
𝜖𝜖 = 1𝛽𝛽 (2𝐶𝐶12+𝐶𝐶22
3 )32 𝑀𝑀
32 𝑆𝑆
92 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 = 𝛼𝛼 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 for z > 10 mm (A1)
here M is the mesh size of the grid (m), S the stroke length (m), (Dgrid - z) the distance from
the grid (m), and ƒ the grid oscillating frequency (Hz). The constant value β equals 0.1 and
the values C1 and C2 are derived from the geometry of the grid and are 0.18 and 0.22,
respectively (Fernando and DeSilva, 1993; O’Brien et al., 2004). With a mesh size of 0.06 m
and a stroke length of 0.028 m, α equals 1.11 × 10-10 m6. Brumley and Jirka (1987) analyzed
that Equation (A1) is applicable to 10% of the grid depth below the water surface, thus to z =
10 mm. Closer to the water surface the length scales and the vertical velocity variance are
damped by the air-water interface.
Subsequently the acquired energy dissipation from Equation (A1) was used to
calculate the Kolmogorov length scale (ηk) for each grid frequency in Equation (A2)
(O’Brien et al., 2004):
𝜂𝜂𝐾𝐾 = (𝑣𝑣3
𝜖𝜖 )14 = (𝑣𝑣3
𝛼𝛼 )14 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 − 𝑧𝑧
ƒ34
for z > 10 mm (A2)
the energy dissipation (m2 s-3), and α is derived
using Equation (A1).
Translating turbulence to wind speeds
In order to couple our turbulence statistics to wind speeds encountered in the field, we used the
following equations. In limnology and oceanography, the wind shear stress τwind
(in Pa) exerted on
the water surface is related to the wind velocity U10
(in m s-1) observed 10 meters above the water
surface:
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
146
158
where ν is the kinematic viscosity of the fluid (water = 1.004 × 10-6), (Dgrid - z) the distance
from the grid (m), ƒ the grid oscillating frequency (Hz), 𝜖𝜖 the energy dissipation (m2 s-3), and
α is derived using Equation (A1).
Translating turbulence to wind speeds
In order to couple our turbulence statistics to wind speeds encountered in the field, we
used the following equations. In limnology and oceanography, the wind shear stress τwind (in
Pa) exerted on the water surface is related to the wind velocity U10 (in m s-1) observed 10
meters above the water surface:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑎𝑎𝐶𝐶𝐷𝐷|𝑈𝑈10|𝑈𝑈10 (A3)
where ρa is the density of air (1.2 kg m-3) and CD is the dimensionless wind-drag coefficient,
which is a function of the wind velocity (U10). We apply the approximation of Smith and
Banke (1975):
𝐶𝐶𝐷𝐷 = 10−3(0.63 + 0.066|𝑈𝑈10|) (A4)
Below a wind-sheared water surface, the vertical distribution of the turbulence energy
dissipation (𝜖𝜖) reads:
𝜖𝜖 = 𝑢𝑢∗3
𝐾𝐾 (𝑧𝑧+𝑧𝑧0) [𝑚𝑚2𝑠𝑠−3] (A5)
with z (in m) downward from the water surface and z0 (in m) the so-called surface roughness
height. For low wind speeds, z0 is less than 1 mm. With 𝐾𝐾 the Von Kármán constant which
equals 0.4 and u* the wind-shear stress velocity, by definition related to the wind-shear stress
(𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤) through:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑤𝑤𝑢𝑢∗2 (A6)
with ρw the density of water (1000 kg m-3).
Setting all equalities we can relate the equivalent wind speed (U10) to 𝜖𝜖 through:
(A3)
where ρa is the density of air (1.2 kg m-3) and C
D is the dimensionless wind-drag coefficient, which
is a function of the wind velocity (U10
). We apply the approximation of Smith and Banke (1975):
158
where ν is the kinematic viscosity of the fluid (water = 1.004 × 10-6), (Dgrid - z) the distance
from the grid (m), ƒ the grid oscillating frequency (Hz), 𝜖𝜖 the energy dissipation (m2 s-3), and
α is derived using Equation (A1).
Translating turbulence to wind speeds
In order to couple our turbulence statistics to wind speeds encountered in the field, we
used the following equations. In limnology and oceanography, the wind shear stress τwind (in
Pa) exerted on the water surface is related to the wind velocity U10 (in m s-1) observed 10
meters above the water surface:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑎𝑎𝐶𝐶𝐷𝐷|𝑈𝑈10|𝑈𝑈10 (A3)
where ρa is the density of air (1.2 kg m-3) and CD is the dimensionless wind-drag coefficient,
which is a function of the wind velocity (U10). We apply the approximation of Smith and
Banke (1975):
𝐶𝐶𝐷𝐷 = 10−3(0.63 + 0.066|𝑈𝑈10|) (A4)
Below a wind-sheared water surface, the vertical distribution of the turbulence energy
dissipation (𝜖𝜖) reads:
𝜖𝜖 = 𝑢𝑢∗3
𝐾𝐾 (𝑧𝑧+𝑧𝑧0) [𝑚𝑚2𝑠𝑠−3] (A5)
with z (in m) downward from the water surface and z0 (in m) the so-called surface roughness
height. For low wind speeds, z0 is less than 1 mm. With 𝐾𝐾 the Von Kármán constant which
equals 0.4 and u* the wind-shear stress velocity, by definition related to the wind-shear stress
(𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤) through:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑤𝑤𝑢𝑢∗2 (A6)
with ρw the density of water (1000 kg m-3).
Setting all equalities we can relate the equivalent wind speed (U10) to 𝜖𝜖 through:
(A4)
Below a wind-sheared water surface, the vertical distribution of the turbulence energy dissipation
(
157
Appendix 7.1
Methods Chapter 7. Estimation of turbulence properties and equivalent wind speeds
In this section we exploit the turbulence velocity observations of Fernando and DaSilva
(1993) and their empirical relations to grid size and frequency, derived in similar grid
experiments but without an air-water interface. Reversely, given the grid size and grid
frequency in our Limnotron experiments we estimate turbulence properties for vertical
mixing, using the Kolmogorov length scale of maximum fine-scale shearing (ηk) and the rms
of vertical velocity |𝑤𝑤′(𝑧𝑧)|. In addition, we relate the turbulence level at 10 mm below the
water surface to the corresponding wind speed over a lake.
In general, we follow the approach of O’Brien et al. (2004) by first estimating 𝜖𝜖, the
rate of viscous dissipation of turbulent kinetic energy, by:
𝜖𝜖 = 1𝛽𝛽 (2𝐶𝐶1
2+𝐶𝐶22
3 )32 𝑀𝑀
32 𝑆𝑆
92 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 = 𝛼𝛼 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 for z > 10 mm (A1)
here M is the mesh size of the grid (m), S the stroke length (m), (Dgrid - z) the distance from
the grid (m), and ƒ the grid oscillating frequency (Hz). The constant value β equals 0.1 and
the values C1 and C2 are derived from the geometry of the grid and are 0.18 and 0.22,
respectively (Fernando and DeSilva, 1993; O’Brien et al., 2004). With a mesh size of 0.06 m
and a stroke length of 0.028 m, α equals 1.11 × 10-10 m6. Brumley and Jirka (1987) analyzed
that Equation (A1) is applicable to 10% of the grid depth below the water surface, thus to z =
10 mm. Closer to the water surface the length scales and the vertical velocity variance are
damped by the air-water interface.
Subsequently the acquired energy dissipation from Equation (A1) was used to
calculate the Kolmogorov length scale (ηk) for each grid frequency in Equation (A2)
(O’Brien et al., 2004):
𝜂𝜂𝐾𝐾 = (𝑣𝑣3
𝜖𝜖 )14 = (𝑣𝑣3
𝛼𝛼 )14 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 − 𝑧𝑧
ƒ34
for z > 10 mm (A2)
) reads:
158
where ν is the kinematic viscosity of the fluid (water = 1.004 × 10-6), (Dgrid - z) the distance
from the grid (m), ƒ the grid oscillating frequency (Hz), 𝜖𝜖 the energy dissipation (m2 s-3), and
α is derived using Equation (A1).
Translating turbulence to wind speeds
In order to couple our turbulence statistics to wind speeds encountered in the field, we
used the following equations. In limnology and oceanography, the wind shear stress τwind (in
Pa) exerted on the water surface is related to the wind velocity U10 (in m s-1) observed 10
meters above the water surface:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑎𝑎𝐶𝐶𝐷𝐷|𝑈𝑈10|𝑈𝑈10 (A3)
where ρa is the density of air (1.2 kg m-3) and CD is the dimensionless wind-drag coefficient,
which is a function of the wind velocity (U10). We apply the approximation of Smith and
Banke (1975):
𝐶𝐶𝐷𝐷 = 10−3(0.63 + 0.066|𝑈𝑈10|) (A4)
Below a wind-sheared water surface, the vertical distribution of the turbulence energy
dissipation (𝜖𝜖) reads:
𝜖𝜖 = 𝑢𝑢∗3
𝐾𝐾 (𝑧𝑧+𝑧𝑧0) [𝑚𝑚2𝑠𝑠−3] (A5)
with z (in m) downward from the water surface and z0 (in m) the so-called surface roughness
height. For low wind speeds, z0 is less than 1 mm. With 𝐾𝐾 the Von Kármán constant which
equals 0.4 and u* the wind-shear stress velocity, by definition related to the wind-shear stress
(𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤) through:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑤𝑤𝑢𝑢∗2 (A6)
with ρw the density of water (1000 kg m-3).
Setting all equalities we can relate the equivalent wind speed (U10) to 𝜖𝜖 through:
(A5)
with z (in m) downward from the water surface and z0 (in m) the so-called surface roughness height.
For low wind speeds, z0 is less than 1 mm. With the
158
where ν is the kinematic viscosity of the fluid (water = 1.004 × 10-6), (Dgrid - z) the distance
from the grid (m), ƒ the grid oscillating frequency (Hz), 𝜖𝜖 the energy dissipation (m2 s-3), and
α is derived using Equation (A1).
Translating turbulence to wind speeds
In order to couple our turbulence statistics to wind speeds encountered in the field, we
used the following equations. In limnology and oceanography, the wind shear stress τwind (in
Pa) exerted on the water surface is related to the wind velocity U10 (in m s-1) observed 10
meters above the water surface:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑎𝑎𝐶𝐶𝐷𝐷|𝑈𝑈10|𝑈𝑈10 (A3)
where ρa is the density of air (1.2 kg m-3) and CD is the dimensionless wind-drag coefficient,
which is a function of the wind velocity (U10). We apply the approximation of Smith and
Banke (1975):
𝐶𝐶𝐷𝐷 = 10−3(0.63 + 0.066|𝑈𝑈10|) (A4)
Below a wind-sheared water surface, the vertical distribution of the turbulence energy
dissipation (𝜖𝜖) reads:
𝜖𝜖 = 𝑢𝑢∗3
𝐾𝐾 (𝑧𝑧+𝑧𝑧0) [𝑚𝑚2𝑠𝑠−3] (A5)
with z (in m) downward from the water surface and z0 (in m) the so-called surface roughness
height. For low wind speeds, z0 is less than 1 mm. With 𝐾𝐾 the Von Kármán constant which
equals 0.4 and u* the wind-shear stress velocity, by definition related to the wind-shear stress
(𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤) through:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑤𝑤𝑢𝑢∗2 (A6)
with ρw the density of water (1000 kg m-3).
Setting all equalities we can relate the equivalent wind speed (U10) to 𝜖𝜖 through:
Von Kármán constant which equals 0.4 and
u* the wind-shear stress velocity, by definition related to the wind-shear stress (
158
where ν is the kinematic viscosity of the fluid (water = 1.004 × 10-6), (Dgrid - z) the distance
from the grid (m), ƒ the grid oscillating frequency (Hz), 𝜖𝜖 the energy dissipation (m2 s-3), and
α is derived using Equation (A1).
Translating turbulence to wind speeds
In order to couple our turbulence statistics to wind speeds encountered in the field, we
used the following equations. In limnology and oceanography, the wind shear stress τwind (in
Pa) exerted on the water surface is related to the wind velocity U10 (in m s-1) observed 10
meters above the water surface:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑎𝑎𝐶𝐶𝐷𝐷|𝑈𝑈10|𝑈𝑈10 (A3)
where ρa is the density of air (1.2 kg m-3) and CD is the dimensionless wind-drag coefficient,
which is a function of the wind velocity (U10). We apply the approximation of Smith and
Banke (1975):
𝐶𝐶𝐷𝐷 = 10−3(0.63 + 0.066|𝑈𝑈10|) (A4)
Below a wind-sheared water surface, the vertical distribution of the turbulence energy
dissipation (𝜖𝜖) reads:
𝜖𝜖 = 𝑢𝑢∗3
𝐾𝐾 (𝑧𝑧+𝑧𝑧0) [𝑚𝑚2𝑠𝑠−3] (A5)
with z (in m) downward from the water surface and z0 (in m) the so-called surface roughness
height. For low wind speeds, z0 is less than 1 mm. With 𝐾𝐾 the Von Kármán constant which
equals 0.4 and u* the wind-shear stress velocity, by definition related to the wind-shear stress
(𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤) through:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑤𝑤𝑢𝑢∗2 (A6)
with ρw the density of water (1000 kg m-3).
Setting all equalities we can relate the equivalent wind speed (U10) to 𝜖𝜖 through:
) through:
158
where ν is the kinematic viscosity of the fluid (water = 1.004 × 10-6), (Dgrid - z) the distance
from the grid (m), ƒ the grid oscillating frequency (Hz), 𝜖𝜖 the energy dissipation (m2 s-3), and
α is derived using Equation (A1).
Translating turbulence to wind speeds
In order to couple our turbulence statistics to wind speeds encountered in the field, we
used the following equations. In limnology and oceanography, the wind shear stress τwind (in
Pa) exerted on the water surface is related to the wind velocity U10 (in m s-1) observed 10
meters above the water surface:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑎𝑎𝐶𝐶𝐷𝐷|𝑈𝑈10|𝑈𝑈10 (A3)
where ρa is the density of air (1.2 kg m-3) and CD is the dimensionless wind-drag coefficient,
which is a function of the wind velocity (U10). We apply the approximation of Smith and
Banke (1975):
𝐶𝐶𝐷𝐷 = 10−3(0.63 + 0.066|𝑈𝑈10|) (A4)
Below a wind-sheared water surface, the vertical distribution of the turbulence energy
dissipation (𝜖𝜖) reads:
𝜖𝜖 = 𝑢𝑢∗3
𝐾𝐾 (𝑧𝑧+𝑧𝑧0) [𝑚𝑚2𝑠𝑠−3] (A5)
with z (in m) downward from the water surface and z0 (in m) the so-called surface roughness
height. For low wind speeds, z0 is less than 1 mm. With 𝐾𝐾 the Von Kármán constant which
equals 0.4 and u* the wind-shear stress velocity, by definition related to the wind-shear stress
(𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤) through:
𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 = 𝜌𝜌𝑤𝑤𝑢𝑢∗2 (A6)
with ρw the density of water (1000 kg m-3).
Setting all equalities we can relate the equivalent wind speed (U10) to 𝜖𝜖 through:
(A6)
with ρw the density of water (1000 kg m-3).
Setting all equalities we can relate the equivalent wind speed (U10
) to
157
Appendix 7.1
Methods Chapter 7. Estimation of turbulence properties and equivalent wind speeds
In this section we exploit the turbulence velocity observations of Fernando and DaSilva
(1993) and their empirical relations to grid size and frequency, derived in similar grid
experiments but without an air-water interface. Reversely, given the grid size and grid
frequency in our Limnotron experiments we estimate turbulence properties for vertical
mixing, using the Kolmogorov length scale of maximum fine-scale shearing (ηk) and the rms
of vertical velocity |𝑤𝑤′(𝑧𝑧)|. In addition, we relate the turbulence level at 10 mm below the
water surface to the corresponding wind speed over a lake.
In general, we follow the approach of O’Brien et al. (2004) by first estimating 𝜖𝜖, the
rate of viscous dissipation of turbulent kinetic energy, by:
𝜖𝜖 = 1𝛽𝛽 (2𝐶𝐶12+𝐶𝐶22
3 )32 𝑀𝑀
32 𝑆𝑆
92 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 = 𝛼𝛼 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 for z > 10 mm (A1)
here M is the mesh size of the grid (m), S the stroke length (m), (Dgrid - z) the distance from
the grid (m), and ƒ the grid oscillating frequency (Hz). The constant value β equals 0.1 and
the values C1 and C2 are derived from the geometry of the grid and are 0.18 and 0.22,
respectively (Fernando and DeSilva, 1993; O’Brien et al., 2004). With a mesh size of 0.06 m
and a stroke length of 0.028 m, α equals 1.11 × 10-10 m6. Brumley and Jirka (1987) analyzed
that Equation (A1) is applicable to 10% of the grid depth below the water surface, thus to z =
10 mm. Closer to the water surface the length scales and the vertical velocity variance are
damped by the air-water interface.
Subsequently the acquired energy dissipation from Equation (A1) was used to
calculate the Kolmogorov length scale (ηk) for each grid frequency in Equation (A2)
(O’Brien et al., 2004):
𝜂𝜂𝐾𝐾 = (𝑣𝑣3
𝜖𝜖 )14 = (𝑣𝑣3
𝛼𝛼 )14 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 − 𝑧𝑧
ƒ34
for z > 10 mm (A2)
through:
159
𝑈𝑈10 = { 𝜌𝜌𝑤𝑤𝜌𝜌𝑎𝑎𝐶𝐶𝐷𝐷
}12 {𝐾𝐾(𝑧𝑧 + 𝑧𝑧0)𝜖𝜖(𝑧𝑧)}
13 (A7)
Equation (A4) shows that the drag coefficient CD depends on the wind speed U10, yielding in
equation (A7) an implicit relation for U10. For simplicity, we set CD = 10-3.
The final step is to relate at a certain depth the turbulence energy dissipation (A1) created by
the oscillating grid to the turbulence energy dissipation (A5) created by wind over a lake.
From this equivalence follows the corresponding wind speed U10. As reference depth we
select z = 10 mm below the water surface where equation (A1) is still applicable while
neglecting z0 for low wind speeds, yielding:
𝑈𝑈10~150 𝜖𝜖13 𝜖𝜖 at z = 10 mm below the water surface (A8)
Solving equation (A8) with energy dissipation rates (𝜖𝜖) from Table 7.1 yields low to
moderate wind speeds of 0.7 - 3 m s-1. Further, from equation (A8) and (A1) follows that the
corresponding grid frequency (ƒ) is proportional to U10, hence doubling the highest grid
frequency would correspond to 6 m s-1 wind velocity. The motor of the grid oscillation
system did, however, not allow higher frequencies.
(A7)
Equation (A4) shows that the drag coefficient CD depends on the wind speed U
10, yielding in
equation (A7) an implicit relation for U10
. For simplicity, we set CD = 10-3.
The final step is to relate at a certain depth the turbulence energy dissipation (A1) created by the
oscillating grid to the turbulence energy dissipation (A5) created by wind over a lake. From this
equivalence follows the corresponding wind speed U10
. As reference depth we select z = 10 mm
below the water surface where equation (A1) is still applicable while neglecting z0 for low wind
speeds, yielding:
159
𝑈𝑈10 = { 𝜌𝜌𝑤𝑤𝜌𝜌𝑎𝑎𝐶𝐶𝐷𝐷
}12 {𝐾𝐾(𝑧𝑧 + 𝑧𝑧0)𝜖𝜖(𝑧𝑧)}
13 (A7)
Equation (A4) shows that the drag coefficient CD depends on the wind speed U10, yielding in
equation (A7) an implicit relation for U10. For simplicity, we set CD = 10-3.
The final step is to relate at a certain depth the turbulence energy dissipation (A1) created by
the oscillating grid to the turbulence energy dissipation (A5) created by wind over a lake.
From this equivalence follows the corresponding wind speed U10. As reference depth we
select z = 10 mm below the water surface where equation (A1) is still applicable while
neglecting z0 for low wind speeds, yielding:
𝑈𝑈10~150 𝜖𝜖13 𝜖𝜖 at z = 10 mm below the water surface (A8)
Solving equation (A8) with energy dissipation rates (𝜖𝜖) from Table 7.1 yields low to
moderate wind speeds of 0.7 - 3 m s-1. Further, from equation (A8) and (A1) follows that the
corresponding grid frequency (ƒ) is proportional to U10, hence doubling the highest grid
frequency would correspond to 6 m s-1 wind velocity. The motor of the grid oscillation
system did, however, not allow higher frequencies.
(A8)
Solving equation (A8) with energy dissipation rates
157
Appendix 7.1
Methods Chapter 7. Estimation of turbulence properties and equivalent wind speeds
In this section we exploit the turbulence velocity observations of Fernando and DaSilva
(1993) and their empirical relations to grid size and frequency, derived in similar grid
experiments but without an air-water interface. Reversely, given the grid size and grid
frequency in our Limnotron experiments we estimate turbulence properties for vertical
mixing, using the Kolmogorov length scale of maximum fine-scale shearing (ηk) and the rms
of vertical velocity |𝑤𝑤′(𝑧𝑧)|. In addition, we relate the turbulence level at 10 mm below the
water surface to the corresponding wind speed over a lake.
In general, we follow the approach of O’Brien et al. (2004) by first estimating 𝜖𝜖, the
rate of viscous dissipation of turbulent kinetic energy, by:
𝜖𝜖 = 1𝛽𝛽 (2𝐶𝐶12+𝐶𝐶22
3 )32 𝑀𝑀
32 𝑆𝑆
92 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 = 𝛼𝛼 ƒ3
(𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔−𝑧𝑧)4 for z > 10 mm (A1)
here M is the mesh size of the grid (m), S the stroke length (m), (Dgrid - z) the distance from
the grid (m), and ƒ the grid oscillating frequency (Hz). The constant value β equals 0.1 and
the values C1 and C2 are derived from the geometry of the grid and are 0.18 and 0.22,
respectively (Fernando and DeSilva, 1993; O’Brien et al., 2004). With a mesh size of 0.06 m
and a stroke length of 0.028 m, α equals 1.11 × 10-10 m6. Brumley and Jirka (1987) analyzed
that Equation (A1) is applicable to 10% of the grid depth below the water surface, thus to z =
10 mm. Closer to the water surface the length scales and the vertical velocity variance are
damped by the air-water interface.
Subsequently the acquired energy dissipation from Equation (A1) was used to
calculate the Kolmogorov length scale (ηk) for each grid frequency in Equation (A2)
(O’Brien et al., 2004):
𝜂𝜂𝐾𝐾 = (𝑣𝑣3
𝜖𝜖 )14 = (𝑣𝑣3
𝛼𝛼 )14 𝐷𝐷𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 − 𝑧𝑧
ƒ34
for z > 10 mm (A2)
from Table 7.1 yields low to moderate wind
speeds of 0.7 - 3 m s-1. Further, from equation (A8) and (A1) follows that the corresponding grid
frequency (ƒ) is proportional to U10
, hence doubling the highest grid frequency would correspond
to 6 m s-1 wind velocity. The motor of the grid oscillation system did, however, not allow higher
frequencies.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
147
7
Supplementary Figure 7.1 – Number of heterocysts per total cell number for Aphanizomenon (a, b) and Woronichinia colony size in µm2 measured at four different depths (c, d) with increasing (a, c) and decreasing (b, d) turbulence levels.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 7
148
Supplementary Figure 7.2 – Chlorophyll-a concentrations in µg L-1 for Aphanizomenon (a, b) and Woronichinia (c, d) measured at four different depths with increasing (a, c) and decreasing (b, d) turbulence levels.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Gone with the wind - Stability of cyanobacterial scums under turbulent conditions
149
7
Supplementary Figure 7.3 – Schematic overview of grid-generated turbulence properties below a water surface as observed in (Brumley and Jirka, 1987) with
162
Supplementary Figure 7.3 – Schematic overview of grid-generated turbulence properties below a
water surface as observed in (Brumley and Jirka, 1987) with |𝑤𝑤′(𝑧𝑧)| the rms of the vertical velocity
and |𝑢𝑢′(𝑧𝑧)| of the horizontal velocity. Through this turbulent regime a colony rises with velocity wr.
the rms of the vertical velocity and
162
Supplementary Figure 7.3 – Schematic overview of grid-generated turbulence properties below a
water surface as observed in (Brumley and Jirka, 1987) with |𝑤𝑤′(𝑧𝑧)| the rms of the vertical velocity
and |𝑢𝑢′(𝑧𝑧)| of the horizontal velocity. Through this turbulent regime a colony rises with velocity wr.
of the horizontal velocity. Through this turbulent regime a colony rises with velocity w
r.
CHAPTER 8
Synthesis
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 8
152
INTERNAL P LOADING – IS IRON THE SOLUTION?
Increased nutrient loading in shallow freshwater lakes has led to enhanced lake productivity,
resulting in a changed community structure, turbid water, a decrease in biodiversity, reduced
water quality, and formation of (toxic) cyanobacterial blooms (Davis et al., 2010). Whereas a large
part of this external phosphate (PO4) loading has now been reduced in many European and North
American lakes, the recovery of these lakes is often hampered or delayed by internal loading of
phosphate from the sediment (Jeppesen et al., 1991; Søndergaard et al., 2003; Smolders et al.,
2006). One way to reduce this internal loading is by adding chemical P-binding agents, such
as iron (Fe), to the water column or sediment. The use of iron as a restoration measure to reduce
internal loading has been tested and verified on many occasions, both using lab (Burley et al.,
2001; Smolders et al., 2001; Hansen et al., 2003) and field studies (Boers et al., 1994; Daldorph
and Price, 1994; Kleeberg et al., 2012). By adding iron to a lake to bind to the excess phosphate
in the system, the lake is expected to shift towards a clear water macrophyte dominated state
which is generally considered positive for biodiversity (Smith and Schindler, 2009). High iron
concentrations in the water column and sediment can, however, have serious negative effects on
aquatic organisms (Gerhardt and Westermann, 1995; Kamal et al., 2004).
The first part of this thesis therefore described and tested these possible negative effects
of iron addition and presented guidelines and constraints for the use of iron as a restoration
tool. I started by reviewing the known effects of high iron concentrations on aquatic organisms
from literature. This was followed by lab experiments, testing potential toxic effects of iron
on macrophytes, both common species and species with high conservation value. Subsequently
I compared growth of transplanted macrophyte species, with and without herbivory effects of
invasive crayfish, in an iron-rich and iron-poor pond in the field. And lastly, to conclude the iron
research, I combined long-term monitoring data from Lake Terra Nova with an evaluation of a
whole-lake iron addition experiment.
IRON ADDITION AS A RESTORATION TOOL
Iron toxicity
While iron is an essential nutrient for both primary and secondary producers, when in excess,
it can negatively affect growth, behaviour, reproduction, or even cause death of the organism
(Wheeler et al., 1985; Vuori, 1995). Moreover, iron can precipitate as iron hydroxides, which can
alter food quality, food availability, habitat structure, and attach to vital parts of the organism,
resulting in stress and tissue damage (Gerhardt and Westermann, 1995; Vuori, 1995; Linton et
al., 2007). Toxicity studies have shown a big difference in the response of organisms to high iron
concentrations (Chapter 2; Khangarot and Ray, 1989; Shuhaimi-Othman et al., 2012a). Due to
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Synthesis
153
8
these differences in species’ sensitivities to iron, adding iron to a lake can result in a change in
community composition, favouring the more iron resistant species (Chapter 2).
Despite these possible toxic side effects of iron addition, visual signs of iron toxicity remained
absent in four different macrophyte species during iron addition experiments with gradual doses
of 20 and 40 g Fe m-2 (Chapters 3 and 4; Figure 8.1; Figure 8.2). Adding up to 40 g Fe m-2 was
not lethal to any of the tested macrophytes, but a difference was found in growth rate, with
iron addition resulting in slower growth of the species Potamogeton pectinatus and Chara globularis
compared to unaffected growth of the species Elodea nuttallii and Chara virgata (Chapters 3 and
4). Thus, even though the direct effects of toxicity were not visible in these aquatic plants, it
could be that the (energetic) costs of iron tolerance in P. pectinatus and C. globularis were merely
expressed by a decrease in growth rate, as was found for floating macrophytes and non-aquatic
plants (Snowden and Wheeler, 1995; Van der Welle et al., 2007a).
Addition of iron in the water column, as opposed to partly mixing it in the sediment, also
resulted in higher concentrations of precipitated iron, both in the water column (Chapter 3) and
on the surface of the macrophytes and experimental tanks (Figure 8.3; Chapter 4), which could
have induced light limitation or form a physical barrier for macrophyte emergence from the
sediment. Light is a crucial factor for growth of macrophytes, especially that of charophytes (Kufel
and Kufel, 2002; Rip et al., 2007). These high concentrations of precipitated iron could therefore
have limited charophyte growth in the high iron treatments. The sprouting of propagules from
the sediment, however, was not hindered by iron addition or high concentrations of precipitated
iron and during the addition experiments a range of charophyte species (Nitella mucronata, Chara
virgata, and Chara globularis) sprouted from the sediment (Chapter 3).
The negative effects of the addition of 40 g Fe m-2 on P. pectinatus and C. globularis biomass
may have partly been due to the fact that iron was added in the water column over a short period
of only 12 and 5 weeks, respectively. When using iron addition as a lake restoration measure, the
choice can be made for water column iron addition distributed over a longer time period or iron
injections in the sediment. Moreover, pH and alkalinity will be more stable in lakes compared to
small experimental units and therefore potential negative consequences of iron addition such as
iron hydroxide formation and a drop in pH and alkalinity would be reduced (Chapters 2 and 6).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 8
154
Figure 8.1 – Schematic overview of the possible interactions (both direct and indirect) of iron addition on the aquatic foodweb covered in the first part of this thesis. Grey and orange arrows represent simplified consumptive foodweb interactions and possible iron addition effects, respectively. Letters indicate tested relationships, with (A) negative effects of iron addition on P availability (Chapters 2 and 6), (B) slightly negative (lab) to positive (experimental ponds and lake) effects of iron on macrophyte growth and survival (Chapters 3, 4, 5, and 6), (C) negative effects of fish and crayfish on macrophyte recovery (Chapters 5 and 6), and (D) no recorded effects of iron addition on zooplankton and fish, but negative effects on phytoplankton abundance (Chapter 2 and 6). Figure adapted from http://www.pkgills.com/wp-content/uploads/2010/03/thefoodweb.png.
Indeed, transplant experiments involving both fast growing macrophyte species E. nuttallii
and Myriophyllum spicatum and the species of higher conservation interest C. virgata, showed that
the tested macrophytes were not negatively affected by high iron concentrations when grown in
a pond that had previously received 85 g Fe m-2 (Chapter 5; Figure 8.1). In fact, at the end of the
experiment, M. spicatum biomass was even higher in this iron-rich pond compared to a pond that
was not dosed with iron, but there was no significant difference for the other species (Chapter
5). This is in accordance with restoration experiments in lakes, either by iron addition in the
sediment or slow addition in the water, which monitored biological effects and did not notice any
negative effects of iron on aquatic organisms (Chapters 2 and 6; Figure 8.1; Daldorph and Price,
1994; Jaeger, 1994). Additionally, although iron restoration studies use high quantities of iron,
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Synthesis
155
8
dilution and chemical interactions of iron with sulphate (SO4), PO
4, and dissolved organic matter
(DOC) will quickly reduce the bioavailability of iron to the aquatic community (Chapter 2). Due
to these chemical interactions, the lake is expected to shift from a turbid algal dominated state
towards a clear macrophyte dominated state, which could eventually have the most important
effect on the aquatic community (Jeppesen et al., 2012). Nonetheless, it still remains difficult to
predict the long-term effects of iron addition on aquatic life.
Figure 8.2 – Experimental tanks containing Chara globularis, C. virgate, and an empty control two weeks into gradually receiving 20 g Fe m-2 (Chapter 4).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 8
156
Figure 8.3 – Elodea nuttallii growing in experimental tanks (Chapter 3) after 2 months of iron dosing receiving (a) 0, (b) 20, and (c) 40 g Fe m-2 in the water column. Notice the iron precipitates on the shoots and experimental tanks in the 40 g Fe m-2 treatment (c).
Sustainable Fe application
Application of iron in Lake Terra Nova, dosing 33 g Fe m-2 over a period of 1.5 years, drastically
improved water quality by lowering TP, suspended matter (SM), and chlorophyll concentrations,
giving way to a diverse macrophyte community (Chapter 6; Figure 8.1). This was in accordance
with earlier iron addition experiments, where iron addition resulted in increased sediment P
retention and decreased chlorophyll concentrations (Boers et al., 1994; Daldorph and Price, 1994;
Jaeger, 1994; Kleeberg et al., 2012). Several restoration studies using iron, however, reported
only short term success (1 - 3 months) due to location specific confounding factors, which either
influenced lake P concentrations or inhibited macrophyte success (Chapter 5; Walker et al., 1989;
Boers et al., 1994; Van Donk et al., 1994). The limited longevity of the positive effects was in
these cases due to high external P loading (Boers et al., 1994), short water retention time (Boers
et al., 1994), heavy wind effects or seasonal turnover (Quaak et al., 1993; Walker et al., 1989), a
high population of plankti- and benthivorous fish (Chapter 6; Van Donk et al., 1994), or invasive
crayfish inhibiting the development of submerged macrophytes (Chapter 5). Long term success
of iron dosing on water quality, without negatively affecting the aquatic community depends,
therefore, on both chemical and biological lake characteristics.
Confounding factors for long term success
Firstly, iron dosing should be done carefully over a longer time period in order to prevent a quick
drop in pH which could directly affect pelagic and bottom dwelling aquatic organisms (Chapter
2). Iron dosing in the macrophyte experiments was stretched over a period of 5 to 12 weeks
(Chapters 3 and 4), which, due to the buffer capacity of the Terra Nova sediment, only slightly
decreased water column pH but stayed well above 7. Dosing in Lake Terra Nova was applied over
an even longer time period of 1.5 years with the help of a wind-driven mill at the centre of the
lake (Chapter 6). Due to this slow dosing, local build-up of high iron concentrations was avoided,
resulting in a stable surface water pH and acute exposure of biota to high levels of the added
chemicals was prevented. Slow addition of iron over a longer time (months to a year) thus enables
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Synthesis
157
8
addition of cumulative high total concentrations of iron to the water column, whereas dosing at
once restricts the method to lower dosages (Jaeger, 1994; Van der Welle et al., 2007a).
Due to chemical interactions of iron with DOC or other compounds with high affiliation to
iron, the P binding capacity of the added iron is reduced. High concentrations of DOC, such as
found in the organic-rich peaty Lake Terra Nova, can precipitate with iron to form humic-iron
complexes. Iron addition in Lake Terra Nova resulted in a steep decrease in DOC concentrations,
which could imply that a large part of the added iron precipitated with DOC (Chapter 6).
Formation of such stable complexes may have considerably reduced the amount of free iron to
form a P barrier on the water-sediment interface, which was visible by a slow increase in lake
TP concentrations in the two years after iron addition had stopped (Chapter 6). Therefore, iron
dosing in organic rich lakes, or lakes with elevated concentrations of other compounds with high
affiliations for iron (such as SO4), should be repeated or carried out with a surplus of Fe to avoid
the treatment being ineffective. Conversely, in the case of these lakes the choice can be made for
a different P capping agent that does not react with these compounds, such as aluminium (Cooke
et al., 1993).
Aluminium addition has in some cases been proposed over iron addition due to the redox
sensitivity of iron, as iron loses its P-binding capacity during anoxic conditions (Lijklema, 1977;
Cooke et al., 1993). However, Kleeberg et al. (2013) showed that the P-binding capacity of iron
is assured even under anoxic conditions, save enough iron is added to reach a sediment molar
Fe:P ratio ≥ 7. Addition of high amounts of iron to reach high sediment or pore water Fe:P ratios
are often suggested in literature in order for iron to guarantee long-term P regulation (Jensen
et al., 1992; Smolders et al., 2001; Zak et al., 2004; Geurts et al., 2008). Whether adding high
quantities of iron to reach high Fe:P ratios without extra oxygenation can on the long-term
prevent P release, even during thermal stratification and high O2 consumption events, is however
still debated (Cooke et al., 1993; Kleeberg et al., 2013).
Besides chemical characteristics of a lake, also the biological community (e.g. birds, fish, and
crayfish) can hamper the success of iron addition by inhibiting the return of submerged macrophytes
due to grazing, sediment upwelling or non-consumptive plant shredding (Chapter 5; Bakker et al.,
2013). Iron addition by Van Donk et al. (1994) in large mesocosms in Lake Breukeleveen did not
result in decreased chlorophyll concentrations, but removal of sediment upwelling benthivorous fish
considerably increased water transparency and macrophyte biomass. In contrast, biomanipulation in
Lake Terra Nova alone did not result in decreased chlorophyll and suspended matter concentrations,
but the combination of continuous fish removal and iron addition did (Chapter 6). The crayfish
enclosure and exclosure experiments with macrophyte transplants in the iron-rich and iron-poor ponds
also showed that the invasive crayfish species Procambarus clarkii can strongly reduce the survival and
growth of macrophytes, even in absence of any other herbivores (Chapter 5; Figure 8.1). Crayfish impact
on macrophyte establishment and development is potentially even larger than other herbivores as they
are able to feed on alternative sources like detritus and live on the sediment (Momot, 1995), which is
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 8
158
where macrophytes emerge from propagules. Improvement of water quality alone might therefore not
result in a return of macrophyte vegetation due to continuous suppression of propagule germination
by crayfish. The success of Fe-addition in Lake Terra Nova was for that reason probably facilitated by
the ongoing fish and crayfish removal, as it may have enhanced the recovery of macrophyte cover and
diversity through reduced herbivory and sediment disturbance (Chapter 6).
PREDICTION OF CYANOBACTERIAL BLOOMS
While iron addition can improve water quality and shift a lake towards a macrophyte dominated state,
adding iron to a whole lake is a costly and time consuming process. During the process of restoration
a cyanobacterial bloom can still form, as occurred in Lake Terra Nova where during the iron addition
period in the summer of 2010 a cyanobacterial bloom flourished (Chapter 6; Ter Heerdt et al., 2012).
Moreover, whereas external P loading in Europe and North-America is decreasing, eutrophication still
remains an issue in other parts of the world, which are increasingly experiencing nuisance cyanobacterial
blooms (Guo, 2007; Jeppesen et al., 2012). Additionally, climate change might undo lake restoration
efforts via increasing temperatures, nutrient loading and water column stability, which can result in
increased abundance and duration of cyanobacterial blooms and scums (Pearl and Huisman, 2008;
Carey et al., 2012; Rigosi et al., 2014). Prediction on place and time of cyanobacterial scum occurrence
could therefore be a solution to protect the public from contact with these toxic scums and bridge the
years before water quality is fully restored.
A model designed by Ibelings et al. (2003) successfully predicted the occurrence of
cyanobacterial scums in the open water of Lake IJssel in The Netherlands, but scum predictions
for more sheltered areas such as lake shores, where scums may accumulate and persist longer, still
remains problematic. A team of researchers and water managers in The Netherlands have tried
to accomplish this issue by designing a scum prediction model (EWACS) for four shallow lakes,
designed to predict scums both in the open water and on the lake shores (Burger et al., 2009).
Whereas the model correctly predicted most of the occurring scums, it also predicted scums that
were not observed in the field (false positives). It was proposed that the prediction model would
benefit from species specific information on flotation velocities and scum formation characteristics
under turbulence, instead of using only one value derived from the species Microcystis. A variety
of cyanobacterial species can dominate in a scum, which differ in their shape, size, and flotation
velocity. Species specific information could therefore increase the accuracy of the models, especially
information on scum appearance and disappearance with decreasing and increasing turbulence.
The last chapter and second part of this thesis therefore focussed on gaining more insight
in the scum behaviour of two different cyanobacterial species under increasing and decreasing
turbulence. To do so, I used a combination of experiments and technical engineering models
to follow and evaluate scum behaviour of the species Aphanizomenon flos-aquae and Woronichinia
naegeliana with increasing and decreasing turbulences.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Synthesis
159
8
Stability of cyanobacterial scums under turbulent conditions
In our experiments I set out to test the effect of turbulence on scum formation (and breakdown)
by floating cyanobacteria (Chapter 7). For this I used specially designed 920 L mesocosms
(Limnotrons; Verschoor et al., 2003) which were fitted with a motor-controlled oscillating grid
below the water surface, which generated turbulence at four different frequencies. In order to
predict scums, you not only need to predict the timing and intensity of scum formation but also
the timing of their breakdown. Therefore I performed two consecutive experiments, one where
I slowly increased the oscillation frequency to the highest speed and one with slowly decreasing
oscillation frequencies, which was initiated by an hour of mixing at the highest intensity.
The distribution of cell and chlorophyll-a concentrations over depth showed a different reaction
for each species to the increasing turbulence speeds, with a more stable scum for Aphanizomenon
(Figure 8.4a) compared to a more easily disturbed scum of Woronichinia (Figure 8.4b; Chapter
7). Decreasing the turbulence intensity after a day of mixing showed, however, that Woronichinia
formed a surface scum once mixing had stopped, while after a whole day of intense mixing
Aphanizomenon cell concentration at the surface had significantly decreased (Chapter 7). These
differences in scum behaviour highlight the importance of identifying the dominant species in
a lake in order to make accurate model predictions. The scum prediction model from Burger et
al. (2009) incorporated species specific information from the genus Microcystis, whereas the lakes
were mainly dominated by heterocystous cyanobacteria, such as Anabaena and Aphanizomenon
(Burger et al., 2009).
Figure 8.4 – Surface view of scums of Aphanizomenon (a) and Woronichinia (b) during the increased mixing experiment at the highest oscillation frequency (2.41 Hz).
The experiments showed that Aphanizomenon formed stable scums under turbulent conditions,
which are comparable to those of Microcystis (Wallace and Hamilton, 2000), yet the colonial
aggregates appear less resistant to shear (Chapter 7; Moisander et al., 2002). Scum formation
of Aphanizomenon after high wind events will therefore most likely take much longer, which to
some extent explains why Burger et al. (2009) produced many false positives in their predictions
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Chapter 8
160
(Chapter 7). Conversely, scums produced by Woronichinia will be quickly mixed to deeper layers
of the water column at low wind speeds, but these will slowly reappear at the surface once the
water column is stable (Chapter 7).
Additionally, cyanobacteria appeared to be able to maintain a position at the surface at higher
turbulence frequencies once a scum had formed, as opposed to a situation where particles were
originally mixed and floated upward to the surface. Scum formation appeared to dampen the
effects of turbulence, which shows that higher wind speeds need to be allocated in prediction
models for scum disappearance compared to scum appearance. This is in accordance with the
prediction model of Ibelings et al. (2003), where they assigned wind functions which were shifted
to higher wind speeds needed for disappearance than to inhibit appearance of scums.
The newly gained information on scum behaviour of the notorious scum forming species
Aphanizomenon flos-aquae and the lesser known, but recently more commonly sighted, Woronichinia
naegeliana (Wilk-Wozniak et al., 2003; Oberholster et al., 2006; Mooney et al., 2011) shows that
the stability of scums differs between cyanobacterial species (Chapter 7). Applying species specific
information of the dominant cyanobacterial species in the target lake in prediction models will
therefore most probably result in more accurate predictions, which will ultimately lead to better
warning of lake users against encounters with toxic scums.
GENERAL CONCLUSIONS
Lake restoration remains a hot topic on a global scale, not only due to the history of eutrophication
in the past, but also as future predictions forecast increased cyanobacterial dominance due
to increased temperatures, stratification, and nutrient loading. This thesis focussed on the
applicability of iron addition as a tool to restore lakes that have been suffering from high internal
loading. Iron toxicity, as tested on macrophytes and plankton community composition does not
occur with the doses that are used for restoration, mostly because the bioavailability of iron will
be much lower due to dilution and chemical interactions in the lake. Moreover, iron addition
would shift a lake from a turbid algal dominated state to a clear macrophyte dominated state,
which could eventually have the most important effect on the aquatic community. Interference of
chemical and biological interactions which might limit long term success can be prevented when
following the presented guidelines for iron addition in this thesis.
Whereas restoration of freshwater systems can in some cases take a long time, cyanobacterial
scum prediction models can in the meantime prevent encounters with these toxic species. The
combined information presented in this thesis can benefit both ecologists who study the effects of
restoration measures on aquatic ecosystems and lake managers who are searching for an applicable
method to restore lakes that suffer from these historical P-loads or help predict cyanobacterial
scums to protect the public before lakes are fully restored.
SUMMARY
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Summary
162
IRON ADDITION AS A RESTORATION TOOL
Eutrophication of shallow lakes has globally resulted in a decline in water quality, causing a
shift from clear macrophyte dominated systems to turbid cyanobacterial dominated systems.
Even though in Europe and North America external inputs of phosphorus (P) are declining,
internal P loading has become a major problem in many freshwater lakes due to the build-up
of nutrient stocks in the sediment over the past decades. Various restoration experiments using
chemical P-capping agents, such as aluminium, lime, and iron, have been a success by reducing
P availability and increasing sediment P retention, resulting in a shift back to a macrophyte
dominated state.
Of these different P-capping agents, iron is a capping agent that naturally reached lake
sediments in high quantities via seepage, but due to changes in water regimes, this input of iron-
rich groundwater has decreased. Adding a chemical that naturally occurs in high quantities in
lakes might therefore be more favourable than adding substances which are not commonly found
in lakes. Whereas restoration experiments using iron effectively lowered lake P concentrations,
adding high quantities of iron to a lake could negatively affect lake ecosystems, as iron could
impose toxic effects on the biota. Even though iron is an essential nutrient for growth, when
added in excess, it can negatively affect aquatic organisms, either directly due to toxic effects or
indirectly due to precipitation of iron hydroxides. These precipitations could alter food quality,
food availability, habitat structure, and could attach to vital parts of the aquatic organisms,
resulting in stress and tissue damage. Toxicity studies have shown a big difference in the response
of organisms to high iron concentrations which could lead to a change in community composition,
favouring the more iron resistant species (Chapter 2).
Therefore, the aim of this study was to test whether iron addition, in the dosages used for lake
restoration, is toxic to macrophytes, the target species that are aimed to return after the use of this
restoration measure. In this thesis I therefore experimentally tested the effects of iron using both
lab experiments, where we tested the effects of iron on the growth of four different macrophyte
species, and field experiments, which were performed in both closed-off ponds and on a whole
lake scale (Terra Nova, The Netherlands).
In the lab experiments I tested the effects of iron addition, with doses of 20 and 40 g Fe
m-2, on the growth, survival, nutrient allocation, and propagule germination of four different
macrophyte species, both fast growing species (Chapter 3) and species with high conservation
value (Chapter 4). Growth of Elodea nuttallii and Chara virgata was not affected by iron addition,
whereas Potamogeton pectinatus and C. globularis growth significantly decreased with increasing
iron concentrations. Nonetheless, biomass of all species increased in all experiments relative to
starting conditions and during the experiments several charophyte species sprouted from the
sediment, which was not hindered by iron addition. The decrease of P. pectinatus and C. globularis
biomass with high iron additions may have been caused by iron induced light limitation, as
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Summary
163
S
concentrations of precipitated iron in the water and on the surface of plants significantly increased
in the high iron treatments. While iron in these lab experiments was added over a short time
period of only 5 and 12 weeks in small experimental units, lake restoration studies can dose
over a longer time period and have a much larger body of water above the sediment. Negative
consequences of iron addition such as this iron hydroxide formation, but also a drop in pH and
alkalinity, would therefore be much less dramatic. Indeed, growth and survival of the transplanted
macrophyte species E. nuttallii, C. virgate, and Myriophyllum spicatum in closed-off ponds were not
affected by the higher iron concentrations in iron-rich ponds (which had received an iron dose of
85 g Fe m-2 prior to the experiment) when compared to growth and survival in iron-poor ponds
(Chapter 5).
The whole-lake experiment was performed by dosing 33 g Fe m-2 over 1.5 years in the water
column of the shallow peaty Lake Terra Nova, during which, and up to two years after the
dosing had stopped, I followed macrophyte development, phyto- and zooplankton community
composition, and lake nutrient concentrations. The restoration experiment resulted in a positive
change in water quality, where after 1.5 years of dosing water column P, suspended matter
(SM), and chlorophyll concentrations considerably decreased, without negatively affecting the
biota (Chapter 6). The increase in water transparency coincided with the return of a diverse
macrophyte community, a process that continued during two years after addition had stopped.
Nonetheless, interactions of dissolved organic carbon (DOC) with Fe had considerably reduced
the amount of free iron to form a P barrier on the water-sediment interface and consequently lake
P concentrations slowly rose to pre-restoration conditions once iron addition had stopped.
Iron dosing in organic-rich lakes with high affiliation for Fe (or with high concentrations of
other compounds that react with iron, such a sulphate) should therefore be repeated or carried
out with a surplus of Fe to avoid the treatment being ineffective. In order to guarantee long term
success of iron addition a surplus of iron should be added to reach a molar sediment Fe:P ratio ≥
7, as in these ratios the sediment P binding capacity is maintained. Addition should, however,
be performed over a longer time period of a few months to a year to prevent the aforementioned
accumulation of iron hydroxides. Alternatively, the choice can be made for other chemical
P-binding agents, such as aluminium, which forms an irreversible bond with P.
Other confounding factors for long-term success could be external P loading, which should
be tackled before iron application, and a high abundance of benthi- and planktivorous fish. The
addition of iron Lake Terra Nova was complemented with ongoing biomanipulation measures,
which considerably reduced the amount of sediment upwelling benthivorous fish. Additionally,
invasive crayfish could inhibit the return of macrophytes, as the invasive crayfish Procambarus clarkii
strongly reduced the biomass and survival of transplanted macrophytes in the experimental ponds
(Chapter 5). These crayfish not only inhibit the return of macrophytes due to direct consumption,
but can also increase water turbidity through sediment resuspension, destroy macrophyte biomass
by non-consumptive shredding and alter macrophyte community composition due to selective
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Summary
164
feeding (Chapter 5). Improvement of water quality alone might therefore not result in a return of
macrophyte vegetation due to continuous suppression by benthivorous fish or crayfish.
To conclude, differences in species response to iron addition might lead to a change in
community composition, favouring the more iron-tolerant species. Long term effects of addition
on the biotic community are, however, relatively unknown. Nonetheless, iron addition can
indirectly induce a shift from a eutrophic to a mesotrophic system that might eventually have
the most important effect on the diversity of the aquatic community. To guarantee long term
success of iron addition, the constraining chemical and biological factors that might reduce long
term sediment P retention or inhibit macrophyte return should be addressed before or during
iron application.
PREDICTION OF CYANOBACTERIAL SCUM FORMATION
Iron addition as a restoration measure can shift a lake from a cyanobacterial dominated state
to a macrophyte dominated state, but in some cases these effects are only visible on a longer
term. Increasing evidence also shows that climate warming may lead to a rise of cyanobacterial
abundance in lakes, which to some extend may undo the efforts to restore eutrophic systems.
Moreover, whereas external loading in Europe and North America is declining, eutrophication
remains an issue on a global scale. Therefore, in order to bridge the years before water quality
is fully restored, prediction of cyanobacterial scum occurrence could be a solution to protect
the public from contact with these toxic scums. Scum prediction models for open water have
successfully predicted both time and place of scum formation, but prediction of scums in more
sheltered areas still remains difficult, whereas many recreational areas typically are sheltered
locations. One reason for the mismatch with these prediction models for sheltered areas could be
that these models mostly use scum characteristics of only one cyanobacterial species (Microcystis
sp.), whereas a variety of cyanobacterial species can be dominating in a scum. A way to benefit
these prediction models is therefore to improve our knowledge of turbulence effects on different
cyanobacterial species and their scum formation as scums of cyanobacterial species could differ in
their response to turbulence.
In the last part of this thesis I therefore experimentally investigated the effect of turbulence
(induced by an oscillating grid) on scum formation and disappearance of the notorious scum
forming species Aphanizomenon flos-aquae and the lesser known, but recently more commonly
sighted Woronichinia naegeliana (Chapter 7). A combination of depth measurements and
turbulence model simulations showed that the two species differed in their response to increasing
grid frequencies in large mesocosms (Limnotrons), with a more stable scum of Aphanizomenon
compared to the scum of Woronichinia.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Summary
165
S
Decreasing grid frequencies from the highest intensity to 0 Hz resulted in a scum of
Woronichinia after mixing had completely stopped. Aphanizomenon, however, did not fully recover
after the intense mixing treatment and cell distribution over depth remained similar to the
distribution at the start of the decreasing turbulence experiment. These differences in scum
behaviour, e.g. turbulence resistance and shear resistance, highlight the importance of identifying
the dominant cyanobacterial species in a lake. Hence shallow lake prediction models can be
improved by incorporating species specific information of the dominant cyanobacteria in the
model target lake, which will ultimately result in more accurate model predictions.
Increased nutrient loading, climate change and imbalanced foodweb interactions are some of the
many factors that have resulted, and in the near future will result, in increased cyanobacterial
dominance in freshwater lakes. This thesis has shown that we, as scientists, policy makers, water
boards, and other water users are able to prevent further degradation of our freshwater systems.
Whereas some of these intended changes could take a while to take place, prediction models of
cyanobacterial scum occurrence can protect lake users against unforeseen encounters with toxic
scums before lakes are fully restored.
SAMENVATTING
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Samenvatting
168
IJZERADDITIE ALS HERSTELMAATREGEL
De toename van voedingsstoffen in ondiepe meren heeft wereldwijd geleid tot een afname van de
waterkwaliteit, waardoor heldere, door planten gedomineerde plassen zijn veranderd in troebele,
door cyanobacteriën gedomineerde plassen. Ook al is de toestroom van voedingsstoffen (fosfor, P)
vanuit externe bronnen in Europa en Noord-Amerika al sterk afgenomen, toch wordt het herstel van
deze meren vaak nog belemmerd door interne fosfaatbelasting vanuit de bodem, die in de afgelopen
decennia is ontstaan door een ophoping van voedingsstoffen in het sediment. Verschillende herstel-
experimenten waarbij men chemische fosfaatbindende stoffen, zoals aluminium, kalk of ijzer, aan een
meer heeft toegevoegd, hebben met succes de fosfaatbeschikbaarheid in het water kunnen verlagen en
de fosfaatbinding van de bodem kunnen verhogen. Dit heeft geresulteerd in een verschuiving in het
systeem van fytoplankton dominantie naar planten (macrofyten) dominantie.
Van de verschillende fosfaatbindende stoffen is ijzer (Fe) de stof die vaak in hoge concentraties
te vinden was in meren en wel door aanvoer via ijzerrijk kwel. Veranderingen in grondwaterstanden
hebben er echter voor gezorgd dat de aanvoer van ijzerrijk kwelwater is afgenomen. Het toevoegen
van een stof die van nature in grote hoeveelheden in meren voorkomt heeft de voorkeur boven het
toevoegen van stoffen die niet of in mindere mate worden aangetroffen in deze meren. Ondanks het
feit dat Fe als fosfaatbindende stof weliswaar positieve effecten kan hebben voor de diversiteit van
het aquatische ecosysteem, kan het toevoegen van grote hoeveelheden ook negatieve effecten hebben,
aangezien hoge concentraties ijzer giftig kunnen zijn. Hoewel ijzer voor veel organismen een essentiële
voedingsstof is voor de groei, kan een overdaad schadelijk zijn, hetzij direct als gevolg van toxische
effecten, hetzij indirect als gevolg van neerslag van ijzerhydroxiden. Deze neerslag kan de kwaliteit
en beschikbaarheid van voedsel verslechteren, kan de structuur van het habitat veranderen en kan
zich hechten aan vitale delen van aquatische organismen, wat kan zorgen voor stress en weefselschade.
Studies naar de toxiciteit van ijzer laten grote verschillen zien in resistentie van verschillende aquatische
organismen. Dit kan leiden tot een verschuiving in de soortensamenstelling naar meer ijzer resistente
soorten (Hoofdstuk 2).
Het doel van deze studie is het testen of ijzer, in de doseringen die worden gebruikt tijdens
herstelmaatregelen, schadelijk is voor waterplanten, aangezien deze organismen verwacht worden terug
te komen na het toevoegen van ijzer en de daarmee samenhangende verbetering van de waterkwaliteit.
In dit proefschrift heb ik dit onderzocht aan de hand van laboratorium experimenten waarbij ik heb
gekeken naar de effecten van het toevoegen van ijzer op de groei van vier verschillende waterplanten.
Daarnaast heb ik deze effecten ook getest aan de hand van veldexperimenten, die op kleine schaal
werden uitgevoerd in afgesloten proefvijvers en op grote schaal in de ondiepe veenplas Terra Nova.
De effecten van ijzeradditie werden in het laboratorium onderzocht op basis van doses van
20 en 40 g Fe m-2. Daarbij werd gekeken naar het effect op groei, overleving, concentratie en
verdeling van de voedingsstoffen en kiemkracht van vier verschillende waterplanten, zowel
snelgroeiende soorten als soorten met een hogere conserveringswaarde (Hoofdstuk 3 en 4).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Samenvatting
169
S
De groei van Elodea nuttallii en Chara virgata werd niet beïnvloed door ijzeradditie, terwijl de
groei van Potamogeton pectinatus en C. globularis afnam bij toenemende concentraties toegevoegd ijzer.
Toch nam de biomassa van alle soorten toe vergeleken met de startcondities. Daarnaast kiemden
tijdens de experimenten verschillende kranswieren uit het sediment, een proces dat niet werd
gehinderd door de toevoeging van ijzer aan het water of het sediment. De afname in groei van P.
pectinatus en C. globularis bij de hoge ijzeraddities kan zijn veroorzaakt door licht limitatie, aangezien
de concentraties neergeslagen ijzer in het water en op het oppervlak van de waterplanten bij de hoge
ijzeraddities significant waren toegenomen. In de laboratorium experimenten is het ijzer gedurende
korte periodes van slechts 5 en 12 weken in kleine proefopzetten toegevoegd, maar in het veld kan
de dosering worden uitgespreid over veel langere periodes en zal de dosering verder worden verdund
door de grotere kolom water boven de waterbodem. Negatieve effecten van ijzeradditie, zoals de
formatie van ijzerhydroxides, maar ook een afname in pH en buffervermogen, zullen daardoor
minder dramatisch zijn. Veldexperimenten met de getransplanteerde waterplanten E. nuttallii, C.
virgata en Myriophyllum spicatum in afgesloten proefvijvers lieten inderdaad zien dat de groei en
overleving niet werden beïnvloed door de hogere concentraties ijzer in de ijzerrijke vijver (welke
vooraf was behandeld met 85 g Fe m-2) indien vergeleken met groei en overleving in de ijzerarme
vijver (Hoofdstuk 5).
Tijdens het veld experiment in de veenplas Terra Nova werd 33 g Fe m-2 over een periode
van 1,5 jaar langzaam in de waterkolom gedoseerd. Gedurende de periode van ijzeradditie en een
periode van twee jaar daarna heb ik de ontwikkeling van de onderwater vegetatie (waterplanten), de
samenstelling van de zoöplankton en fytoplankton gemeenschappen en de nutriënten concentraties in
het meer gevolgd. Het veldexperiment resulteerde in een verbetering van de waterkwaliteit, waarbij
na een periode van 1,5 jaar ijzer toevoegen de hoeveelheid fosfor (P), het aantal opgeloste deeltjes en
de hoeveelheid chlorofyl in de waterkolom aanzienlijk waren afgenomen, zonder negatieve effecten
te hebben op de aquatische flora en fauna (Hoofdstuk 6). De verbetering van de waterkwaliteit en
de toename van het doorzicht in het water vielen samen met de terugkeer van waterplanten in de
veenplas, een proces dat onveranderd bleef tijdens de twee jaar na het stoppen van de ijzeradditie.
Toch zorgde de reactie van opgelost organisch koolstof met ijzer voor een daling van de beschikbare
hoeveelheid ijzer noodzakelijk voor het vormen van een fosfaat-barrière op het grensvlak van het
water en de bodem van het meer. Het gevolg was dat de P concentraties in het water, nadat de
ijzeradditie was gestopt, langzaam stegen naar de waarden van voor het veldexperiment.
IJzeradditie in vergelijkbare organische meren met een hoge consumptie van ijzer (of hoge
concentraties van andere stoffen die reageren met ijzer, zoals sulfaat) moet daarom worden herhaald
of worden uitgevoerd met een overschot aan ijzer, dit om te voorkomen dat de behandeling geen
effect heeft (op P). Om ook op de lange termijn het succes van ijzeradditie te garanderen moet
een overschot aan ijzer worden toegevoegd om op deze wijze een molaire Fe:P verhouding ≥ 7
te bereiken, een verhouding waarbij de P-bindingscapaciteit van de waterbodem kan worden
verzekerd. Daarnaast moet de dosering worden uitgevoerd over een langere periode van enkele
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Samenvatting
170
maanden tot jaren om de eerder genoemde vorming en opeenstapeling van ijzerhydroxiden te
voorkomen. Als alternatief kan ook gekozen worden voor andere chemische fosfaatbindende stoffen
zoals aluminium, dat een onomkeerbare verbinding vormt met P.
Andere factoren die het lange termijn succes van ijzeradditie kunnen verstoren zijn een hoge
externe fosfaatbelasting, die moet worden aangepakt voordat er kan worden gestart met ijzeradditie,
en een overvloed aan bodem-omwoelende en plankton etende vis. IJzeradditie in Terra Nova ging
vergezeld van biomanipulatie maatregelen, die de hoeveelheid bodem-omwoelende vis aanzienlijk
hebben verminderd. Daarnaast kunnen ook invasieve zoetwaterkreeften de terugkeer van
waterplanten belemmeren, zoals de invasieve zoetwaterkreeft Procambarus clarkii die de biomassa en
overleving van de getransplanteerde waterplanten in de proefvijvers sterk verminderde (Hoofdstuk
5). Deze kreeften verhinderen de terugkeer van waterplanten niet alleen door directe consumptie,
maar ook door het vertroebelen van de waterkolom door sediment resuspensie, door het vernietigen
van de waterplant biomassa via niet-consumptie gerichte versnippering en het door veranderen van
de waterplant gemeenschap als gevolg van selectieve consumptie (Hoofdstuk 5). Verbetering van
de waterkwaliteit alleen zal dus niet altijd leiden tot een terugkeer van waterplanten als gevolg van
continue onderdrukking van de groei door bodem-omwoelende vis en invasieve zoetwaterkreeften.
Verschillen in reacties van uiteenlopende organismen op ijzeradditie kunnen leiden tot een
verandering in de samenstelling van het aquatisch milieu, waarbij de meer ijzer-tolerante soorten
een voordeel zullen hebben. De lange termijn effecten van ijzeradditie op het aquatisch milieu zijn
echter relatief onbekend. In ieder geval zal ijzeradditie indirect zorgen voor een verschuiving van de
waterkwaliteit van eutroof naar mesotroof, wat uiteindelijk de belangrijkste invloed zal hebben op
de diversiteit van het aquatisch milieu. Om het succes van ijzeradditie ook voor de lange termijn te
kunnen garanderen moeten de chemische en biologische factoren die de P-bindingscapaciteit van
de waterbodem kunnen verminderen of de terugkeer van waterplanten kunnen belemmeren, voor of
tijdens de ijzeradditie worden aangepakt.
VOORSPELLEN VAN CYANOBACTERIE DRIJFLAAGVORMING
Herstelmaatregelen zoals ijzeradditie kunnen de toestand van een meer doen verschuiven van
een door cyanobacteriën gedomineerde staat naar een door waterplanten gedomineerde staat. In
sommige gevallen zijn deze effecten echter alleen zichtbaar op de langere termijn. Daarnaast is er
steeds meer bewijs dat de opwarming van de aarde kan leiden tot een toename van cyanobacteriën,
die tot op zekere hoogte de inspanningen van herstelmaatregelen ongedaan kunnen maken.
Bovendien is vermesting op globale schaal nog steeds een groot probleem, dit terwijl de externe
fosfaatbelasting in Europa en Noord-Amerika geleidelijk daalt. Om de bevolking te beschermen
tegen ongewenst contact met deze giftige drijflagen en de periode te overbruggen die nodig is om
de waterkwaliteit volledig te herstellen, is het voorspellen van drijflaagvorming van cyanobacteriën
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Samenvatting
171
S
mogelijk een oplossing. Voorspellingsmodellen hebben al met succes drijflaagvorming in open water
voorspeld, waarbij zowel de tijd als de plaats van vorming correct werd voorspeld. Voorspelling van
drijflagen in meer beschutte gebieden blijft echter nog een probleem, en recreatiegebieden zijn vaak
beschutte locaties. Een mogelijke verklaring voor deze mismatch tussen de modelvoorspellingen
en de daadwerkelijke drijflaagvorming in beschutte locaties kan zijn dat de voorspellingsmodellen
vaak gebruik maken van drijflaag eigenschappen en kenmerken van maar één soort (Microcystis sp.),
terwijl verschillende soorten cyanobacteriën drijflagen kunnen vormen. Aangezien drijflagen van
cyanobacteriën kunnen verschillen in hun reactie op turbulentie zouden de voorspellingsmodellen
kunnen worden verbeterd door onze kennis van turbulentie effecten op verschillende soorten
cyanobacteriën en hun drijflagen uit te breiden.
In het laatste deel van dit proefschrift heb ik daarom het effect van turbulentie (gecreëerd door een
oscillerend grid) op het verdwijnen en vormen van drijflagen van de beruchte drijflaagvormende soort
Aphanizomenon flos-aquae en de minder bekende, maar wel steeds vaker voorkomende Woronichinia
naegeliana experimenteel onderzocht (Hoofdstuk 7). Een combinatie van dieptemetingen in grote
920 L tanks (Limnotrons) en turbulentie modelvoorspellingen liet zien dat de drijflagen van de
twee geteste soorten verschilden in hun reactie op toenemende grid oscillatie frequenties, waarbij
Aphanizomenon een stabielere drijflaag vormde dan Woronichinia.
Bij afnemende grid oscillatie frequenties, vanaf de hoogste intensiteit tot 0 Hz, vormde
Woronichinia een drijflaag nadat het mixen compleet was gestopt. Aphanizomenon bleek echter niet
volledig hersteld na de intense menging en de verdeling van cellen over de diepte was aan het eind
van het experiment gelijk aan de celverdeling die was gemeten aan het begin van het experiment.
Deze verschillen het gedrag van drijflagen, namelijk de weerstand tegen turbulentie en wrijving,
benadrukken het belang van het identificeren van de dominante soorten cyanobacteriën in een meer.
Voorspellingsmodellen voor ondiepe meren kunnen daarom worden verbeterd door het gebruiken
van soort specifieke informatie van de dominante soort in het betreffende meer, hetgeen uiteindelijk
zal zorgen voor betere, nauwkeurigere modelvoorspellingen.
Vermesting, klimaatverandering en onevenwichtige voedselwebben zijn enkele van de vele
factoren die hebben geleid, en in de nabije toekomst zullen leiden, tot toenemende dominantie van
cyanobacteriën in ondiepe meren. Dit proefschrift heeft aangetoond dat wij als wetenschappers,
beleidsmakers, waterschappen en andere watergebruikers in staat zijn om verdere degradatie van
onze meren te voorkomen. Aangezien de realisatie van een aantal van deze veranderingen veel tijd
vraagt, kunnen in de tussentijd drijflaag voorspellingsmodellen worden ingezet om de recreanten te
beschermen tegen onvoorziene en ongewenste ontmoetingen met giftige drijflagen.
DANKWOORD
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Dankwoord
174
ACKNOWLEDGEMENTS – DANKWOORD
Ook al sta ik als enige op de voorkant van dit proefschrift, het was mede dankzij de hulp van vele
anderen dat dit boek in deze vorm er nu is. Ik ben niet alleen dankbaar voor alle kennis en hulp
die er is gegeven vanuit de aquatische hoek, maar ik ben ook zeer dankbaar voor de steun van
vrienden en familie waardoor ik uiteindelijk zo ver ben gekomen.
Allereerst wil ik graag mijn promotoren Ellen en Bas bedanken voor het delen van hun kennis,
het geven van hulp en het vertrouwen dat ze in mij hadden. Daarnaast heb ik erg veel steun gehad
van mijn co-promotor Liesbeth, ze was zeer betrokken bij mijn onderzoek en de deur stond altijd
open. Vanaf het moment dat Ellen bij je aanklopte met het verzoek in het project mee te draaien,
was je er. Ik denk met veel plezier terug aan onze samenwerking.
Het werken op het NIOO deed ik nooit alleen. Tijdens mijn experimenten heb ik intensief
samengewerkt met veel mensen. Ik wil graag Dennis (en je zangkunsten!), Nico, Erik, Thijs,
Harry en Koos super bedanken voor al hun hulp en de vele leuke avonturen die we samen hebben
beleefd. Tânia, thanks for all the hugs, laughs and serious conversations, I never knew engineers
could be so much fun. Ook mijn andere kamergenoten Susanne, Sven, Suzanne, Dirk, Mandy
en Thijs zorgden er voor dat ik mij nooit hoefde te vervelen, stonden altijd klaar voor vragen en
zorgden voor de nodige snacks tijdens het werk. De reis naar Japan was niet hetzelfde geweest
zonder Alena en Susanne. Arigato voor dit prachtige avontuur. Steven heeft heel wat avonduurtjes
voor mij opgeofferd in een voor mij leerzame samenwerking. Ontzettend bedankt hiervoor. Ook
heb ik bij de verschillende projecten veel hulp gehad van mijn studenten Kirsten, Rene, Masha en
Mandy. Additionally, I want to thank Ying and Rémi for their help and the energy they brought
from their own countries China and France. Vele andere collega’s op het NIOO waren maar kort
of niet bij mijn project betrokken, maar indirect hebben ze wel bijgedragen aan een prettige
werkomgeving. Anne, Anne, Annette, Bart, Casper, Daan, Dedmer, Dick, Dilara, Edith, Hennie,
Jan, Jessica, Judith, Kostas, Lisette, Luuk, Marion, Martijn, Mayra, Michaela, Michiel, Naomi,
Paul, Ramesh, Sasha en Spiros, bedankt! Verder wil ik de prettige samenwerking vermelden met
Jeroen en Leon van de Radboud Universiteit Nijmegen, met Gerard van Waternet en met Rob,
Evelyn en Miguel van Deltares.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Dankwoord
175
D
En al ben ik er pas kort, toch wil ik mijn nieuwe collega’s bij Vitens bedanken voor de
ruimte die zij mij hebben gegeven om dit proefschrift fatsoenlijk af te ronden en voor de vele
complimenten die ik al heb mogen ontvangen. Daar zal ik niet snel aan kunnen wennen!
Een hele warme knuffel voor mijn vrienden en vriendinnen. Jullie lieten me vaak weten dat
er meer in de wereld is dan blauwalgen en ijzer! Bedankt voor de afleiding en dat jullie er waren
toen het zo hard nodig was.
Mijn broer en zus hebben mij altijd gesteund en geholpen ‘out-of-the-box’ te denken. Beide
wil ik ook bedanken voor het maken van de prachtige cover van dit boek.
Mijn ouders, zo ontzettend betrokken, maar ook bezig met hun eigen gevecht. Niets maakt
een familie zo hecht als een ziekteproces. Ondanks alle drukte in jullie eigen leven bleven jullie
me steunen en motiveren. Lieve Ben, bedankt voor al het nakijkwerk, maar vooral voor de
inspiratie die je me al mijn hele leven hebt gegeven. Tiny, bedankt voor alle knuffels, de uren die
wij aan de telefoon hebben doorgebracht en je betrokkenheid. Ik weet dat alles goed gaat komen.
Als laatste wil ik Michiel bedanken, mijn rots in de branding. Bedankt dat je er al zo lang voor
me bent en dat je ook de laatste 4.5 jaar bij me bent gebleven! Het laatste jaar was voor jou soms
even frustrerend als voor mij. Ik beloof je dat ik vanaf nu alle weekenden met jou zal doorbrengen,
en dat zijn er heel veel…!
REFERENCES
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
178
Ahn, C.-J., Joung, S.-H., Park, C.-S., Kim, H.-S., Yoon, B.-D. and Oh, H.-N., 2008. Comparison of sampling and analytical methods for monitoring of cyanobacteria-dominated surface waters. Hydrobiologia 596, 413-421.
Alldredge A.L. and Gotschalk, C.C., 1989. Direct observation of the mass flocculation of diatom blooms: characteristics, settling velocities and formation of diatom aggregates. Deep-Sea Research 36, 159-171.
Anastácio, P.M., Correia, A.M. and Menino, J.P., 2005. Processes and patterns of plant destruction by crayfish: effects of crayfish size and developmental stages of rice. Archiv für Hydrobiologie 162, 37-51.
Andersen, J.M., 1975. Influence of pH on release of phosphorus from lake sediments. Archiv für Hydrobiologie 76, 411-419.
Annadotter, H., Cronberg, G., Aagren, R., Lundstedt, B., nilsson, P.-A. and Ströbeck, S., 1999. Multiple techniques for lake restoration. Hydrobiologia 395-396, 77-85.
Aparicio Medrano, E., Uittenbogaard, R.E., Dionisio Pires, L.M., Van de Wiel, B.J.H. and Clercx, H.J.H., 2013. Coupling hydrodynamics and buoyancy regulation in Microcystis aeruginosa for its vertical distribution in lakes. Ecological Modelling 248, 41-56.
Bache, D.H. and Rasool, E., 1996. Measurement of the rate of energy dissipation around an oscillating grid by an energy balance approach. The Chemical Engineering Journal 63, 105-115.
Bakker, D., Osté, L., Roskam, G., De Weert, J. and Hemelraad, J., 2011. De Bodem Bedekt – et onderzoeken en aanbrengen van een fosfaatbindende afdeklaag in de Bergse Voorplas. 119 pp. Project Report, Deltares. (In Dutch)
Bakker, E.S., Sarneel, J.M., Gulati, R.D., Liu, Z. and Van Donk, E., 2013. Restoring macrophyte diversity in shallow temperate lakes: biotic versus abiotic constraints. Hydrobiologia 710, 23-37.
Bakker, E.S., Van Donk, E., Declerck, S.A.J., Helmsing, N.R., Hidding, B., Nolet, B.A., 2010. Effect of macrophyte community composition and nutrients enrichment on plant biomass and algal blooms. Basic and Applied Ecology 11, 432-439.
Barrat-Segretain, M-H., 2004. Growth of Elodea canadensis and Elodea nuttallii in monocultures and mixture under different light and nutrient conditions. Archiv für Hydrobiologie 161, 133-144.
Bates, D., Maechler, M. and Bolker, B., 2011. lme4: linear mixed-effects models using S4 classes. R package version 0.999375-39.
Batty, L.C. and Younger, P.L., 2003. Effects of external iron concentration upon seedling growth and uptake of Fe and phosphate by the common reed, Phragmites australis (Cav.) Trin ex. Steudel. Annals of Botany 92, 801-806.
Berman, T. and Shteinman, B., 1998. Phytoplankton development and turbulent mixing in Lake Kinneret (1992-1996). Journal of Plankton Research 20, 709-726.
Biesinger, K. and Christensen, G.M., 1972. Effects of various metals on survival, growth, reproduction, and metabolism of Daphnia magna. Journal of the Fisheries Research Board of Canada 29, 1691-1700.
Birge, W.J., Black, J.A. and Westerman, A.G., Short, T.M., Taylor, S.B., Bruser, D.M. and Wallingford, E.D.,1985. Recommendations on numerical values for regulating 26 iron and chloride concentrations for the purpose of protecting warmwater species of aquatic life in the commonwealth of Kentucky, University of Kentucky, Lexington, USA.
Bloemendaal, F.H.J.L. and Roelofs, J.G.M., 1988. Waterplanten en waterkwaliteit. Koninklijke Nederlandse NatuurhistorischeVereniging, Utrecht, The Netherlands. (In Dutch)
Boers, P., Van der Does, J., Quaak, M. and Van der Vlucht, J., 1994. Phosphorus fixation with iron(III)chloride: A new method to combat phosphorus loading in shallow lakes? Archiv für Hydrobiologie 129, 339-351.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
179
R
Boers, P., Van der Does, J., Quaak, M., Van der Vlucht, J., Walker, P., 1992. Fixation of phosphorus in lake sediments using iron(III)chloride: experiences, expectations. Hydrobiologia 233, 211-212.
Boers, P., Van der Does, J., Quaak, M. and Van der Vlucht, J., 1994. Phosphorus fixation with iron(III)chloride: A new method to combat phosphorus loading in shallow lakes? Archiv für Hydrobiologie 129, 339-351.
Bootsma, M.C., Barendregt, A. and Van Alphen, J.C.A., 1999. Effectiveness of reducing external nutrient load entering a eutrophicated shallow lake ecosystem in the Naardermeer nature reserve, The Netherlands. Biological Conservation 90, 193-201.
Bornette, G. and Puijalon, S., 2011. Response of aquatic plants to abiotic factors: a review. Aquatic Sciences 73, 1-14.
Boyd, P.W., Jickells, T., Law, C.S., Blain, S., Boyle, E.A., Buesseler, K.O., Coale, K.H., Cullen, J.J., de Baar, H.J., Follows, M., Harvey, M., Lancelot, C., Levasseur, M., Owens, N.P., Pollard, R., Rivkin, R.B., Sarmiento, J., Schoemann, V., Smetacek, V., Takeda, S., Tsuda, A., Turner, S. and Watson, A.J., 2007. Mesoscale iron enrichment experiments 1993-2005: synthesis and future directions. Science 315, 612-617.
Brookes, J.D., Ganf, G.G., Green, D. and Whittington, J., 1999. The influence of light and nutrients on buoyancy, filament aggregation and flotation of Anabaena circinalis, Journal of Plankton Research 21, 327-341.
Brouwer, E. and Smolders, A. P. J., 2006. Nutrientenhuishouding in de veenplas Terra Nova en mogelijkheden tot herstel. Rapport nummer 2006.01, B-Ware, Nijmegen. (In Dutch)
Brumley B.H. and Jirka, G.H., 1987. Near-surface turbulence in a grid-stirred tank. Journal of Fluid Mechanics 183, 235-263.
Burger, D., Los, H., Groot, S., Hulsbergen, R. and Ibelings, B., 2009. EWACS (Early Warning Against Cyano Scums) Phase 2: Model performance and validation. 92 pp. Project Report, Deltares.
Burley, K.L., Prepas, E.E. and Chambers, P.A., 2001. Phosphorus release from sediments in hardwater eutrophic lakes: the effects of redox-sensitive and -insensitive chemical treatments. Freshwater Biology 46, 1061-1074.
Carey, C.C., Ibelings, B.W., Hoffmann, E.P., Hamilton, D.P. and Brookes, J.D., 2012. Eco-physiological adaptations that favour freshwater cyanobacteria in a changing climate. Water Research 46, 1394-1407.
Carpenter, S.R., 1981. Submersed vegetation: an internal factor in lake ecosystem succession. American Naturalist 118, 372-383.
Carpenter, S.R. and Lodge, D.M., 1986. Effects of submersed macrophytes on ecosystem processes. Aquatic Botany 26, 341-370.
Chambers, P.A., Hanson, J.M., Burke, J.M. and Prepas, E.E., 1990. The impact of the crayfish Orconectes virilis on aquatic macrophytes. Freshwater Biology 24, 81-91.
Chapman, P.M., Farrell, M.A. and Brinkhurst, R.O., 1982. Relative tolerances of selected aquatic oligochaetes to individual pollutants and environmental factors. Aquatic Toxicology 2, 47-61.
Cooke, G.D., Welch, E.B., Martin, A.B., Fulmer, D.G., Hyde, J.B. and Schrieve, G.D., 1993a. Effectiveness of Al, Ca, and Fe salts for control of internal phosphorus loading in shallow and deep lakes. Hydrobiologia 253, 323-335.
Cooke, G.D., Welch, E.B., Peterson, S.A. and Newroth, P.R., 1993b. Restoration and management of lakes and reservoirs. Lewis Publishers, London.
Correia, A.M., 2002. Niche breath and trophic diversity: feeding behaviour of the red swamp crayfish (Procambarus clarkii) towards aquatic macroinvertebrates in a rice field (Portugal). Acta Oecologica 23, 421-429.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
180
Crawley, M.J., 2007. The R Book. John Wiley & Sons, Ltd., Chichester, UK.
Cronin, G., Lodge, D.M., Hay, M.E., Miller, M., Hill, A.M., Horvath, T.M., Bolser, R.C., Lindquist, N. and Wahl, M., 2002. Crayfish feeding preferences for freshwater macrophytes: the influence of plant structure and chemistry. Journal of Crustacean Biology 22, 708-718.
Daldorph, P. and Price, R., 1994. Long-term phosphorus control at 3 eutrophic reservoirs in south-eastern England. Advances in Limnology 40, 231-243.
Dalzell, D.J.B. and Macfarlane, N.A.A., 1999. The toxicity of iron to brown trout and effects on the gills: a comparison of two grades of iron sulphate. Journal of Fish Biology 55, 301-315.
Dave, G., 1985. The influence of pH on the toxicity of aluminum, cadmium, and iron to eggs and larvae of the zebrafish, Brachydanio rerio. Ecotoxicology and Environmental Safety 10, 253-267.
Delvigne G.A.L. and Sweeney, C.E., 1988. Natural dispersion of oil. Oil & Chemical Pollution 4, 281-310.
Davis, J., Sim, L. and Chambers, J., 2010. Multiple stressors and regime shifts in shallow aquatic ecosystems in antipodean landscapes. Freshwater Biology 55, 5-18.
Deppe, T. and Benndorf, J., 2002. Phosphorus reduction in a shallow hypereutrophic reservoir by in-lake dosage of ferrous iron. Water Research 36, 4525-4534.
DeSilva, I.P.D. and Fernando, H.J.S., 1994. Oscillating grids as a source of nearly isotropic turbulence. Physics of Fluids 6, 2455-2464.
Dorenbosch, M. and Bakker, E.S., 2011. Herbivory in omnivorous fishes: effect of presence of plant secondary metabolites and prey stoichiometry. Freshwater Biology 56, 1783-1797.
Dorenbosch, M. and Bakker, E.S., 2012. Effects of contrasting omnivorous fish on submerged macrophyte biomass in temperate lakes: a mesocosm experiment. Freshwater Biology 57, 1360-1372.
Douglas, G.B., Robb, M.S., Coad, D.N. and Ford, P.W., 2004. A review of solid phase adsorbents for the removal of phosphorus from natural and wastewaters. In: Valsami-Jones, E. (Ed.), Phosphorus in Environmental Technology: Principles and Applications. p. 291-320. WA, IWA Publishing, London.
Downs, T.M., Schallenberg, M. and Burns, C.W., 2008. Responses of lake phytoplankton to micronutrient enrichment: a study in two New Zealand lakes and an analysis of published data. Aquatic Sciences 70, 347-360.
Duarte, C. M., 1992. Nutrient concentration of aquatic plants: patterns across species. Limnology and Oceanography 37, 882-889.
Elliot, J.A., 2012. Is the future blue-green? A review of the current model predictions of how climate change could affect pelagic freshwater cyanobacteria. Water Research 46, 1364-1371.
Engelhardt, K.A.M. and Ritchie, M.E., 2001. Effects of macrophyte species richness on wetland ecosystem functioning and services. Nature 411, 687-689.
European Union, 2000. Directive 2000/60/EG of the European Parliament and of the Council Establishing a Framework for the Community Action in the Field of Water Policy of 23 October. PB L 327 of 22 December 2000.
Fernando, H.J.S. and DeSilva, I.P.D., 1993. Note on secondary flows in oscillating-grid, mixing-box experiments. Physics of Fluids 5, 1849-1851.
Furmanska, M., 1979. Studies of the effect of copper, zinc, and iron on the biotic components of aquatic ecosystems. Polskie Archiwum Hydrobiologii 26, 213-220.
Geiger, W., Alcorlo, P., Baltanás, A. and Montes, C., 2005. Impact of an introduced Crustacean on the trophic webs of Mediterranean wetlands. Biological Invasions 7, 49-73.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
181
R
Gensemer, R.W. and Playle, R.C., 1999. The bioavailability and toxicity of aluminum in aquatic environments. Critical Reviews in Environment Science and Technology 29, 315-450.
Gerhardt, A., 1992. Effects of subacute doses of iron (Fe) on Leptophlebia marginata (Insecta: Ephemeroptera). Freshwater Biology 27, 79-84.
Gerhardt, A., 1994. Short term toxicity of iron (Fe) and lead (Pb) to the mayfly Leptophlebia marginata (L.) (Insecta) in relation to freshwater acidification. Hydrobiologia 284, 157-168.
Gerhardt, A. and Westermann, F., 1995. Effects of precipitations of iron hydroxides on Leptophlebia marginata (L.) (Instecta: Ephemeroptera) in the field. Archiv für Hydrobiologie 133, 81-93.
Gerloff, G.C. and Krombholz, P.H., 1966. Tissue analysis as a measure of nutrient availability for the growth of angiosperm aquatic plants. Limnology and Oceanography 11, 529-537.
Geurts, J.J.M., 2010. Restoration of fens and peat lakes: a biogeochemical approach. Ph.D. Thesis. Radboud Universiteit, Nijmegen, The Netherlands.
Geurts, J.J.M., Smolders, A.J.P., Verhoeven, J.T.A., Roelofs, J.G.M. and Lamers, L.P.M., 2008. Sediment Fe:PO
4 ratio as a diagnostic and prognostic tool for the restoration of macrophyte biodiversity in fen waters.
Freshwater Biology 53, 2101-2116.
Gheradi, F., 2006. Crayfish invading Europe: the case study of Procambarus clarkii. Marine and Freshwater Behaviour and Physiology 39, 175-191.
Gherardi, F. and Acquistapace, P. 2007. Invasive crayfish in Europe: the impact of Procambarus clarkii on the litteral community of a Mediterranean lake. Freshwater Biology 52, 1249-1259.
Gherardi, F., Aquiloni, L., Diéguez-Uribeondo, J. and Tricarico, E., 2011. Managing invasive crayfish: is there a hope? Aquatic Sciences 73, 185-200.
Gilling, D., Reich, P. and Thompson, R., 2009. Loss of riparian vegetation alters the ecosystem role of a freshwater crayfish (Cherax destructor) in an Australian intermittent lowland stream. Journal of the North American Benthological Society 28, 626-637.
Golterman, H.L., 1996. Fractionation of sediment phosphate with chelating compounds: a simplification, and comparison with other methods. Hydrobiologia 335, 87-95.
Golterman, H.L., 2001. Phosphate release from anoxic sediments or `What did Mortimer really write?’ Hydrobiologia 450, 99-106.
Gulati, R.D., Pires, L.M.D. and Van Donk, E., 2008. Lake restoration studies: Failures, bottlenecks and prospects of new ecotechnological measures. Limnologica 38, 233-247.
Gulati, R.D. and Van Donk, E., 2002. Lakes in the Netherlands, their origin, eutrophication and restoration: state-of-the-art review. Hydrobiologia 478, 73-106.
Grey, J. and Jackson, M.C., 2012. Leaves and eats shoots’: direct terrestrial feeding can supplement invasive red swamp crayfish in times of need. PLoS One 7, e42575.
Guo, L., 2007. Doing battle with the green monster of Taihu Lake. Science 317, 1166.
Hansen, J., Reitzel, K., Jensen, H.S. and Andersen, F.Ø., 2003. Effects of aluminum, iron, oxygen and nitrate additions on phosphorus release from the sediment of a Danish softwater lake. Hydrobiologia 492, 139-149.
Hansson, L.-A., Annadotter, H. Bergman, E., Hamrin, S.F., Jeppesen, E., Kairesalo, T., Luokkanen, E., Nilsson, P.-Å., Søndergaard, M. and Strand, J., 1998. Biomanipulation as an application of food-chain theory: constraints, synthesis, and recommendations for temperate lakes. Ecosystems 1, 558-574.
Hargeby, A., Blindow, I. and Andersson, G., 2007. Long-term patterns of shifts between clear and turbid states in Lake Krankesjön and Lake Tåkern. Ecosystems 10, 28-35.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
182
Hashimoto, Y. and Nishiuchi, Y., 1981. Establishment of bioessay methods for the evaluation of acute toxicity of pesticides to aquatic organisms. Journal of Pesticide Science 6, 257-264.
Hickey, C.W. and Gibbs, M.M., 2009. Lake sediment phosphorus release management – Decision support and risk assessment framework. New Zealand Journal of Marine and Freshwater Research 43, 819-856.
Hidding, B., Brederveld, R.J. and Nolet, B.A., 2010. How a bottom-dweller beats the canopy: inhibition of an aquatic weed (Potamogeton pectinatus) by macroalgae (Chara spp.). Freshwater Biology 55, 1758-1768.
Hilt, S., 2006. Recovery of Potamogeton pectinatus L. stands in a shallow eutrophic lake under extreme grazing pressure. Hydrobiologia 570, 95-99.
Hilt, S., Gross, E.M., Hupfer, M., Morscheid, H., Mahlmann, J., Melzer, A., Poltz, J., Sandrock, S., Scharf, E.M., Schneider, S. and Van de Weyer, K., 2006. Restoration of submerged vegetation in shallow eutrophic lakes - a guideline and state of the art in Germany. Limnologica 36, 155-171.
Hobbs, R.J. and Huenneke, L.F., 1992. Disturbance, diversity, and invasion: implications for conservation. Conservation Biology 6, 324-337.
Hofstra, J.J. and Van Liere, L., 1992. The state of the environment of the Loosdrecht lakes. Hydrobiologia 233, 11-20.
Hosper, S.H., 1998. Stable states, buffers and switches: An ecosystem approach to the restoration and management of shallow lakes in The Netherlands. Water Science and Technology 37, 151-164.
Humphries, S.E. and Lyne, V.D., 1988. Cyanophyte blooms: the role of cell buoyancy. Limnology and Oceanography 33, 79-91.
Ibelings, B. W., Backer, L. C., Kardinaal, E. and Chorus, I., 2014. Current approaches to cyanotoxin risk assessment and risk management around the globe. Harmful Algae, in press.
Ibelings, B.W. and Chorus, I., 2007. Accumulation of cyanobacterial toxins in freshwater “seafood” and its consequences for public health: a review. Environmental Pollution 150, 177-192.
Ibelings, B.W., Kroon, B.M.A. and Mur, L.R., 1994. Acclimation of photosystem II in a cyanobacterium and a eukaryotic green alga to high and fluctuating photosynthetic photon flux densities, simulating light regimes induced by mixing in lakes. New Phytologist 128, 407-424.
Ibelings, B.W., Mur, L.R. and Walsby, A.E., 1991. Diurnal changes in buoyancy and vertical distribution in populations of Microcystis in 2 shallow lakes. Journal of Plankton Research 13, 419-436.
Ibelings, B.W., Portielje, R., Lammens, E.H.R.R., Noordhuis, R., Van den Berg, M.S., Joosse, W. And Meijer, M.-L., 2007. Resilience of alternative stable states during recovery of shallow lakes from eutrophication: Lake Veluwe as a case study. Ecosystems 10, 4-16.
Ibelings, B.W., Stroom, J.M., Lürling, M.F.L.L.W. and Kardinaal, E.A., 2012. Risks of toxic cyanobacterial blooms in recreational waters and guidelines. In: I. Chorus (Ed.), Current approaches to cyanotoxin risk assessment, risk management and regulations in different countries. p. 82-96. Federal Environment Agency (Umweltbundesamt), Dessau-Roβlau.
Ibelings, B.W., Vonk, M., Los, H.F.J., Van der Molen, D.T. and Mooij, W., 2003. Fuzzy modeling of cyanobacterial surface waterblooms: Validation with NOAA-AVHRR satellite images. Ecological Applications 13, 1456-1472.
Immers, A.K., Van der Sande, M.T., Van der Zande, R.M., Geurts, J.J.M., Van Donk, E. and Bakker, E.S., 2013. Iron addition as a shallow lake restoration measure: impacts on charophyte growth. Hydrobiologia 710, 241-251.
Immers, A.K., Vendrig, K., Ibelings, B.W., Van Donk, E., Ter Heerdt, G.N.J., Geurts, J.J.M. and Bakker, E.S., 2014. Iron addition as a measure to restore water quality: implications for macrophyte growth. Aquatic Botany 116, 44-52.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
183
R
Irfanullah, H. M. and Moss, B., 2004. Factors influencing the return of submerged plants to a clear-water, shallow temperate lake. Aquatic Botany 80, 177-191.
Jackson, M.C., Donohue, I., Jackson, A.L., Britton, J.R., Harper, D.M. and Grey, J., 2012. Population-level metrics of trophic structure based on stable isotopes and their application to invasion ecology. PLoS ONE 7, e31757.
Jaeger, D., 1994. Effects of hypolimnetic water aeration and iron-phosphate precipitation on the trophic level of Lake Krupunder. Hydrobiologia 275/276, 433-444.
Jagtman, E., Van der Molen, D.T. and Vermij, S., 1992. The influence of flushing on nutrient dynamics, composition and densities of algae and transparency in Veluwemeer, the Netherlands. Hydrobiologia 233, 187-196.
James, C.S., Eaton, J.W. and Hardwick, K., 2006. Responses of three invasive aquatic macrophytes to nutrient enrichment do not explain their observed field displacements. Aquatic Botany 84, 347-353.
Jensen, H.S., Kristensen, P., Jeppesen, E. and Skytthe, A., 1992. Iron:phosphorus ratio in surface sediment as an indicator of phosphate release from aerobic sediments in shallow lakes. Hydrobiologia 235-236, 731-743.
Jeppesen, E., Kristensen, P., Jensen, J.P., Søndergaard, M., Mortensen, E. and Lauridsen, T., 1991. Recovery resilience following a reduction in external phosphorus loading of shallow, eutrophic Danish lakes: Duration, regulating factors and methods for overcoming resilience. Memorie dell’ Istituto Italiano di Idrobiologia 48, 127-148.
Jeppesen, E., Lauridsen, T.L., Kairesalo, T. and Perrow, M.R., 1998. Impact of submerged macrophytes on fish-zooplankton interactions in lakes. In: Jeppesen, E., Søndergaard, M., Søndergaard, M., Christoffersen, K. (Eds.), The structuring role of submerged macrophytes in lakes. Ecological Studies, Vol. 131, p. 91-114. Springer-Verlag, Berlin.
Jeppesen, E., Søndergaard, M., Jensen, J.P., Havens, K.E., Anneville, O., Carvalho, L., Coveney, M.F., Deneke, R., Dokulil, M.T., Foy, B., Gerdeaux, D., Hampton, S.E., Hilt, S., Kangur, K., Kohler, J., Lammens, E.H.H.R., Lauridsen, T.L., Manca, M., Miracle, M.R., Moss, B., Noges, P., Persson, G., Phillips, G., Portielje, R., Romo, S., Schelske, C.L., Straile, D., Tatrai, I., Willen, E. and Winder, M., 2005. Lake responses to reduced nutrient loading – an analysis of contemporary long-term data from 35 case studies. Freshwater Biology 50, 1747-1771.
Jeppesen, E., Søndergaard, M., Jensen, J.P., Mortensen, E., Hansen, A.M. and Jørgensen, T., 1998. Cascading trophic interactions from fish to bacteria and nutrients after reduced sewage loading: An 18-year study of a shallow hypertrophic lake. Ecosystems 1, 250-267.
Jeppesen, E., Søndergaard, M., Lauridsen, T.L., Davidson, T.A., Liu, Z., Mazzeo, N., Trochine, C., Özkan, K., Jensen, H.S., Trolle, D., Starling, F., Lazzaro, X., Johansson, L.S., Bjerring, R., Liboriussen, L., Larsen, S.E., Landkildehus, F., Egemose, S. and Meerhoff, M., 2012. Biomanipulation as a restoration tool to combat eutrophication: recent advances and future challenges. In: Woodward, G., Jacob, U., O’Gorman, E.J. (Eds.), Advances in Ecological Research, Vol. 47, p. 411-488. Elsevier, London.
Jeppesen, E., Søndergaard, M., Meerhoff, M., Lauridsen, T.L. and Jensen, J.P., 2007. Shallow lake restoration by nutrient loading reduction - some recent findings and challenges ahead. Hydrobiologia 584, 239-252.
Jöhnk, K.D., Huisman, J., Sharples, J., Sommeijer, B., Visser, P.M. and Stroom, J.M., 2008. Summer heatwaves promote blooms of harmful cyanobacteria. Global Change Biology 14, 495–512.
Jones, H.E. and Etherington, J.R., 1970. Comparative studies of plant growth and distribution in relation to waterlogging: I. The survival of Erica cinerea L. and E. tetralix L. and its apparent relationship to iron and manganese uptake in waterlogged soil. Journal of Ecology 58, 487-496.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
184
Kamal, M., 2004. Phytoaccumulation of heavy metals by aquatic plants. Environment International 29, 1029-1039.
Karol, K.G., McCourt, R.M., Cimino, M.T. and Delwiche, C.F., 2001.The closest living relatives of land plants. Science 294, 2351-2353.
Keller, A.A., Garner, K., Miller, R.J. and Lenihan, H.S., 2012. Toxicity of nano-zero valent iron to freshwater and marine organisms. PLoS One 7, e43983.
Kenefick, S.L., Hrudey, S.E., Peterson, H.G. and Prepas, E.E., 1993. Toxin release from Microcystis aeruginosa after chemical treatment. Water Science & Technology 27, 433-440.
Khan, S. and Nugegoda, D., 2007. Sensitivity of juvenile freshwater crayfish Cherax destructor (Decapoda: Parastacidae) to trace metals. Ecotoxicology and Environmental Safety 68, 463-469.
Khangarot, B.S., 1991. Toxicity of metals to a freshwater tubificid worm, Tubifex tubifex (Muller). Bulletin of Environmental Contamination and Toxicology 46, 906-912.
Khangarot, B.S. and Ray, P.K., 1989. Sensitivity of midge larvae of Chironomus tentans Fabricius (Diptera Chironomidae) to heavy metals. Bulletin of Environmental Contamination and Toxicology 42, 325-330.
Klapwijk, S.P., Kroon, J.M.W. and Meijer, M.-L., 1982. Available phosphorus in lake sediments in The Netherlands. Hydrobiologia 92, 491-500.
Kleeberg, A., Herzog, C. and Hupfer, M., 2013. Redox sensitivity of iron in phosphorus binding does not impede lake restoration. Water Research 47, 1491-1502.
Kleeberg, A., Köhler, A. and Hupfer, M., 2012. How effectively does a single or continuous iron supply affect the phosphorus budget of aerated lakes? Journal of Soils and Sediments 12, 1593-1603.
Klosowski, S., Tomaszewicz, G.H. and Tomaszewicz, H., 2006.The expansion and decline of charophyte communities in lakes within the Sejny Lake District (north-eastern Poland) and changes in water chemistry. Limnologica 36, 234-230.
Koerselman, W. and Meuleman, A.F.M., 1996. The vegetation N:P ratio: A new tool to detect the nature of nutrient limitation. Journal of Applied Ecology 33, 1441-1450.
Koese, B. and Soes, D.M., 2011. De Nederlandse rivierkreeften (Astacoidea & Parastacoidea). Entomologische Tabellen 6. Supplement bij Nederlandse Faunistische Mededelingen. EIS-Nederland, Leiden. (In Dutch)
Körner, S. and Dugdale, T., 2003. Is roach herbivory preventing re-colonization of submerged macrophytes in a shallow lake? Hydrobiologia 506-509, 497–501.
Krueger, K., Chapman, P., Hallock, M. and Quinn, T., 2007. Some effects of suction dredge placer mining on the short-term survival of freshwater mussels in Washington. Northwest Science 81, 323-332.
Kufel, L. and Kufel, I., 2002. Chara beds acting as nutrient sinks in shallow lakes – a review. Aquatic Botany 72, 249-260.
Laan, P., Smolders, A.J.P. and Blom, C.W.P.M., 1991. The relative importance of anaerobiosis and high iron levels in the flood tolerance of Rumex species. Plant and Soil 136, 153-161.
Lake, M.D., Hicks, B.J., Wells, R.D.S. and Dugdale, T.M., 2002. Consumption of submerged aquatic macrophytes by rudd (Scardinius erythrophthalmus L.) in New Zealand. Hydrobiologia 470, 13-22.
Lambert, S.J., and Davy, A.J., 2010. Water quality as a threat to aquatic plants: discriminating between the effects of nitrate, phosphate, boron and heavy metals on charophytes. New Phytologist 189, 1051-1059.
Lamers, L.P.M., Geurts, J.J.M., Bontes, B., Sarneel, J.M., Pijnappel, H.W., Boonstra, H., Schouwenaars, J.M., Klinge, M., Verhoeven, J.T.A., Ibelings, B.W., Verberk, W.C.E.P., Kuijper, B., Esselink, H. and Roelofs, J.G.M., 2006. Onderzoek ten behoeve van het herstel en beheer van Nederlandse laagveenwateren. Eindrapportage 2003–2006. 286 pp. Ede: Netherlands Ministry of Agriculture, Nature and Food Quality. (In Dutch).
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
185
R
Lamers, L.P.M., Smolders, A.J.P. and Roelofs, J.G.M., 2002. The restoration of fens in The Netherlands. Hydrobiologia 478, 107-130.
Lauridsen, T.L., Jensen, J.P., Jeppesen, E. and Søndergaard, M., 2003. Response of submerged macrophytes in Danish lakes to nutrient loading reductions and biomanipulation. Hydrobiologia 506/509, 641-649.
Lauridsen, T., Jeppesen, E. and Østergaard Andersen, F., 1993. Colonization of submerged macrophytes in shallow fish manipulated Lake Væng: impact of sediment composition and waterfowl grazing. Aquatic Botany 46, 1-15.
Lauridsen, T.L., Sandsten, H. and Hald Møller, P., 2003. Restoration of a shallow lake by introducing Potamogeton spp.: Impact of waterfowl grazing. Lakes & Reservoirs: Research & Management 8, 177-187.
Lijklema, L., 1977. Interactions between sediments and freshwater. Golterman, H.L. (Ed.), p. 313-317, Dr W. Junk B. V. Publ., The Hague.
Linton, T.K., Pacheco, M.A.W., McIntyre, D.O., Clement, W.H. and Goodrich-Mahoney, J., 2007. Development of bioassessment-based benchmarks for iron. Environmental Toxicology and Chemistry 26, 1291-1298.
Lodge, D.M. and Lorman, J.G., 1987. Reductions in submerged macrophyte biomass and species richness by the crayfish Oronectes rusticus. Canadian Journal of Fisheries and Aquatic Sciences 44, 591-597.
Lorenzen, C., 1967. Determination of chlorophyll and pheo-pigments: spectrophotometric equations. Limnology and Oceanography 12, 343-346.
Lucaç, M. and Aegerter, R., 1993. Influence of trace metals on growth and toxin production of Microcystis aeruginosa. Toxicon 31, 293-305.
Lucassen, E.C.H.E.T., Smolders, A.J.P. and Roelofs, J.G.M., 2000. Increased groundwater levels cause iron toxicity in Glyceria fluitans (L.). Aquatic Botany 66, 321-327.
Lürling, M. and Tolman, Y., 2010. Effects of lanthanum and lanthanum-modified clay on growth, survival and reproduction of Daphnia magna. Water Research 44, 309-319.
Lürling, M. and Van Oosterhout, F., 2013. Controlling eutrophication by combined bloom precipitation and sediment phosphorus inactivation. Water Research 47, 6527-6537.
Macfie, S.M. and Crowder, A.A., 1987. Soil factors influencing ferric hydroxide plaque formation on roots of Typha latifolia L. Plant and Soil 102, 177-184.
Macintosch, K. A. and Griffiths, D., 2013. Catchment and in-stream influences on metal concentration and ochre deposit density in upland streams, Northern Ireland. Environmental Earth Sciences 70, 3023-3030.
Marklund, O., Sandsten, H., Hansson, L.A. and Blindow, I., 2002. Effects of waterfowl and fish on submerged vegetation and macroinvertebrates. Freshwater Biology 47, 2049-2059.
Marsden, M.W., 1989. Lake restoration by reducing external phosphorus loading; the influence of sediment phosphorus release. Freshwater Biology 21, 139-162.
Martin, J.H., Gordon, M. and Fitzwater, S.E., 1991. The case for iron. Limnology and Oceanography 36, 1793-1802.
Martin, T.R. and Holdich, D.M., 1986. The acute lethal toxicity of heavy metals to peracarid crustaceans (with particular reference to freshwater asellids and gammarids). Water Reseach 20, 1137-1147.
Matsuzaki, S., Usio, N., Takamura, N. and Washitani, I., 2009. Contrasting impacts of invasive engineers on freshwater ecosystems: an experiment and meta-analysis. Oecologia 158, 673-686.
Matthijs, H.C.P., Visser, P.M., Reeze, B., Meeuse, J., Slot, P.C., Wijn, G., Talens, R. and Huisman, J., 2012. Selective suppression of harmful cyanobacteria in an entire lake with hydrogen peroxide. Water Research 46, 1460-1472.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
186
McLachlan, J., Hammer, U.T. and Gorham, P.R., 1963. Observations on the growth and colony habits of ten strains of Aphanizomenon flos-aquae. Phycologia 2, 157-168.
Meijer, M.-L., De Boois, I., Scheffer, M., Portielje, R. and Hosper, H., 1999. Biomanipulation in shallow lakes in The Netherlands: an evaluation of 18 case studies. Hydrobiologia 408/409, 13-30.
Meijer, M.-L., Jeppesen, E., Van Donk, E., Moss, B., Scheffer, M., Lammens, E., Van Nes, E., Van Berkum, J.A., De Jong, G.J., Faafeng, B.A. and Jensen, J.P., 1994. Long-term responses to fish-stock reduction in small shallow lakes: interpretation of five-year results of four biomanipulation cases in The Netherlands and Denmark. Hydrobiologia 275/276, 457-466.
Meis, S., Spears, B.M., Maberly, S.C. and Perkins, R.G., 2013. Assessing the mode of action of Phoslock® in the control of phosphorus release from the bed sediments in a shallow lake (Loch Flemington, UK). Water Research 47, 4460-4473.
Middleboe, A.L. and Markager, S., 1997. Depth limits and minimum light requirements of freshwater macrophytes. Freshwater Biology 37, 553-568.
Moisander, P.H., Hench, J.L., Kononen, K. and Pearl, H.W., 2002. Small-scale shear effects on heterocystous cyanobacteria. Limnology and Oceanography 47, 108-119.
Molot, L.A., Li, G., Findlay, D.L. and Watson, S.B., 2010. Iron-mediated suppression of bloom-forming cyanobacteria by oxine in a eutrophic lake. Freshwater Biology 55, 1102-1117.
Momot, W.T., 1995. Redefining the role of crayfish in aquatic ecosystems. Reviews in Fisheries Science 3, 33-63.
Mooney, K.M., Hamilton, J.T.G., Floyd, S.D., Foy, R.H. and Elliot, C.T., 2011. Initial studies on the occurrence of cyanobacteria and microcystins in Irish lakes. Environmental Toxicology 26, 566-570.
Mortimer, C.H., 1941. The exchange of dissolved substances between mud and water in lakes. Journal of Ecology 29, 280-329.
Morton, S.D. and Lee, T.H., 1974. Algal blooms - Possible effects of iron. Environmental Science & Technology 8, 673-674.
Moss, B., 1989. Water pollution and the management of ecosystems: a case study of science and scientist. In: Grubb, P.J., Whittaker, J.B. (Eds.), Towards a more exact ecology. p. 401-422. Blackwell Scientific Publications, Oxford.
Moss, B., 1990. Engineering and biological approaches to the restoration from eutrophication of shallow lakes in which aquatic plant communities are important components. Hydrobiologia 200, 367-377.
Moss, B., Stephen, D., Alvarez, C., Becares, E., Van De Bund, W., Collings, S.E., Van Donk, E., De Eyto, E., Feldmann, T., Fernández-Aláez, C., Fernández-Aláez, M., Franken, R.J.M., García-Criado, F., Gross, E.M., Gyllström, M., Hansson, L.A., Irvine, K., Järvalt, A., Jensen, J.P., Jeppesen, E., Kairesalo, T., Kornijów, R., Krause, T., Künnap, H., Laas, A., Lill, E., Lorens, B., Luup, H., Miracle, M.R., Nõges, P., Nõges, T., Nykänen, M., Ott, I., Peczula, W., Peeters, E.T.H.M., Phillips, G., Romo, S., Russell, V., Salujõe, J., Scheffer, M., Siewertsen, K., Smal, H., Tesch, C., Timm, H., Tuvikene, L., Tonno, I., Virro, T., Vicente, E. and Wilson, D., 2003. The determination of ecological status in shallow lakes – a tested system (ECOFRAME) for implementation of the European water framework directive. Aquatic Conservation: Marine and Freshwater Ecosystems 13, 507-549.
Mukhopadhyay, M.K. and Konar, S.K., 1984. Toxicity of copper, zinc, and iron to fish, plankton and worm. Geobios 11, 204-207.
Mulderij, G. E., Van Donk, E. and Roelofs, J.G.M., 2003. Differential sensitivity of green algae to allelopathic substances from Chara. Hydrobiologia 491, 261-271.
Murphy, J. and Riley, J.P., 1962. A modified single solution method for determination of phosphate in natural waters. Analytica Chimica Acta 26, 31-36.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
187
R
Myllynen, K., Ojutkangas, E. and Nikinmaa, M., 1997. River water with high iron concentration and low pH causes mortality of Lamprey Roe and newly hatched larvae. Ecotoxicology and Environmental Safety 36, 43-48.
Nayek, S., Gupta, S. and Saha, R., 2010. Effects of metal stress on biochemical response of some aquatic macrophytes growing along an industrial waste discharge channel. Journal of Plant Interactions 5, 91-99.
Nishiuchi, Y. and Yoshida, K., 1972. Toxicities of pesticides to some fresh water snails. Bulletin of the Agricultural Chemicals Inspection Station 12, 86-92.
Nurminen, L., Horppila, J., Lappalainen, J. and Malinen, T., 2003. Implications of rudd (Scardinius erythrophthalmus) herbivory on submerged macrophytes in a shallow eutrophic lake. Hydrobiologia 506-509, 511-518.
Nyström, P., Brönmark, C. and Granéli, W., 1999. Influence of an exotic and native crayfish species on a littoral benthic community. Oikos 85, 545-53.
Nyström, P. and Strand, J., 1996. Grazing by a native and an exotic crayfish on aquatic macrophytes. Freshwater Biology 36, 673-682.
Oberholster, P.J., Botha, A.-M. and Cloete, T.E., 2006. Toxic cyanobacterial blooms in a shallow, artificially mixed urban lake in Colorado, USA. Lakes & Reservoirs: Research and Management 11, 111-123.
O’Brien, K.R., Meyer, D.L., Waite, A.M., Ivey, G.N. and Hamilton, D.P., 2004. Disaggregation of Microcystis aeruginosa colonies under turbulent mixing: laboratory experiments in a grid-stirred tank. Hydrobiologia 519, 143-152.
Olsen, S.R., Cole, C.V., Watanabe, F.S. and Dean, L.A., 1954. Estimation of available phosphorus in soils by extraction with sodium bicarbonate. Vol. Cir. No 939, U.S. p. 1-19. Dept. of Agriculture, Washington DC.
Otsuki, A. and Wetzel, G.R., 1972. Coprecipitation of phosphate with carbonates in a marl-lake. Limnology and Oceanography 17, 763-767.
Otte, M.L., Zozema, J., Koster, L., Haarsma, M.S. and Broekman, R.A., 1989. Iron plaque on roots of Aster tripolium L.: Interaction with zinc uptake. New Phytologist 111, 309-317.
Pearl, H.W., Hall, N.S. and Calandrino, E.S., 2011. Controlling harmful cyanobacterial blooms in a world experiencing anthropogenic and climatic-induced change. Science of the Total Environment 409, 1739-1745.
Pearl, H.W. and Huisman, J., 2008. Blooms like it hot. Science 320, 57-58.
Pan, D.Y. and Liang, X.M., 1993. Safety study of pesticides on bog frog, a predatory natural enemy of pest in paddy field. Journal of Human Agricultural College 19, 47-54.
Penning, W.E., Genseberger, M., Uittenbogaard, R.E. and Cornelisse, J.C., 2013. Quantifying measures to limit wind-driven resuspension of sediments for improvement of the ecological quality of some shallow Dutch lakes. Hydrobiologia 710, 279-295.
Perrow, M.R, Schutten, J.H., Howes, J.R., Holzer, T., Madgewick, F.J. and Jowitt, A.J.D., 1997. Interactions between coot (Fulica atra) and submerged macrophytes: the role of birds in the restoration process. Hydrobiologia 342/343, 241-255.
Phillips, G., Jackson, R., Bennet, C. and Chilvers, A., 1994. The importance of sediment phosphorus release in the restoration of very shallow lakes (The Norfolk Broads, England) and implications for biomanipulation. Hydrobiologia 275/276, 445-456.
Pilon, J. and Santamaria, L., 2002. Clonal variation in morphological and physiological responses to irradiance and photoperiod for the aquatic angiosperm Potamogeton pectinatus. Journal of Ecology 90, 859-870.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
188
Pilon, J., Santamaría, L., Hootsmans, M. and Van Vierssen, W., 2002. Latitudinal variation in life-cycle characteristics of Potamogeton pectinatus L.: vegetative growth and asexual reproduction. Plant Ecology 165, 247-262.
Ploug, H., Musat, N., Adam, B., Moraru, C.L., Lavik, G., Vagner, T., Bergman, B. and Kuypers, M.M.M., 2010. Carbon and nitrogen fluxes associated with the cyanobacterium Aphanizomenon sp. in the Baltic Sea. The ISME Journal 4, 1215-1223.
Ponnamperuna, F.N., 1972. Advances in Agronomy. Brady, N.C. (Ed.), p. 29-96, Academic Press Inc., London.
Prepas, E.E., Babin, J., Murphy, T.P., Chambers, P.A., Sandland, G.J., Ghadouani, A. and Serediak, M., 2001. Long-term effects of successive Ca(OH)
2 and CaCO
3 treatments on the water quality of two eutrophic
hardwater lakes. Freshwater Biology 46, 1089-1103.
Quaak, M., Van der Does, J., Boers, P. and Van der Vlugt, J., 1993. A new technique to reduce internal phosphorus loading by in-lake phosphate fixation in shallow lakes. Hydrobiologia 253, 337-344.
Randall, S., Harper, D. and Brierley, B., 1999. Ecological and ecophysiological impacts of ferric dosing in reservoirs. Hydrobiologia 395/396, 355-364.
Rasmussen, K. and Lindegaard, C., 1988. Effects of iron compounds on macroinvertebrate communities in a Danish lowland river system. Water Research 22, 1101-1108.
Redfield, A.C., 1958. The biological control of chemical factors in the environment. American Scientist 64, 205-221.
Regel, R.H., Brookes, J.D., Ganf, G.G. and Griffiths, R.W., 2004. The influence of experimentally generated turbulence on the Mash01 unicellular Microcystis aeruginosa strain. Hydrobiologia 517, 107-120.
Reitzel, K., Hansen, J., Andersen, F.Ø., Hansen, K.S. and Jensen, H.S., 2005. Lake restoration by dosing aluminum relative to mobile phosphorus in the sediment. Environmental Science & Technology 39, 4134-4140.
Reynolds, C.S., Oliver, R.L. and Walsby, A.E., 1987. Cyanobacterial dominance: The role of buoyancy regulation in dynamic lake environments. New Zealand Journal of Marine and Freshwater Research 21, 379-390.
R Development Core Team, 2011. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0.
Richardson, J.F. and Zaki, W.N., 1954. Sedimentation and fluidization: part I. Transactions of the Institution of Chemical Engineers 32, 82-100.
Rigosi, A., Carey, C.C., Ibelings, B.W. and Brookes, J.D., 2014. The interaction between climate warming and eutrophication to promote cyanobacteria is dependent on trophic state and varies among taxa. Limnology and Oceanography 59, 99-114.
Rip, W. J., Everards, K. and Houwers, A., 1992. Restoration of Botshol (The Netherlands) by reduction of external nutrient load: The effects on physico-chemical conditions, plankton and sessile diatoms. Aquatic Ecology 25, 275-286.
Ritchie, R.J., 2006. Consistent sets of spectrophotometric chlorophyll equations for acetone, methanol and ethanol solvents. Photosynthesis Research 89, 27-41.
Robb, M., Greenop, B., Goss, Z., Douglas, G. and Adaney, J., 2003. Application of PhoslockTM, an innovative phosphorus binding clay, to two Western Australian waterways: preliminary findings. Hydrobiologia 494, 237-243.
Rodríguez-Villafañe, C.L., Becares, E. and Fernández-Aláez, M., 2003. 2003. Shift from clear to turbid phase in Lake Chozas (NW Spain) due to the introduction of American red swamp crayfish (Procambarus clarkii). Hydrobiologia 506-509, 421-426.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
189
R
Sahrawat, K.L., 2004. Iron toxicity in wetland rice and the role of other nutrients. Journal of Plant Nutrition 27, 1471-1504.
Scheffer, M., Hosper, S.H., Meijer, M.-L., Moss, B. and Jeppesen, E., 1993. Alternative equilibria in shallow lakes. Trends in Ecology and Evolution 8, 275-279.
Scheffer, M., 2001. Ecology of shallow lakes. Kluwer Academic Publishers, Dordrecht, Boston, London.
Schindler, D.W., 1978. Factors regulating phytoplankton production and standing crop in the world’s freshwaters. Limnology and Oceanography 23, 478-486.
Sealey, W.M., Lim, C. and Klesius, P.H., 1997. Influence of the dietary level of iron from iron methionine and iron sulfate on immune response and resistance of channel catfish to Edwardsiella ictaluri. Journal of the World Aquaculture Society 28, 142-149.
Shuhaimi-Othman, M., Nadzifah, Y., Nur-Amalina, R. and Umirah, N.S., 2012a. Deriving freshwater quality criteria for iron, lead, nickel, and zinc for protection of aquatic life in Malaysia. The Scientific World Journal 2012, 1-7.
Shuhaimi-Othman, M., Nur-Amalina, R. and Nadzifah, Y., 2012b. Toxicity of metals to a freshwater snail, Melanoides tuberculata. The Scientific World Journal 2012, 1-10.
Simons, J. and Nat, E., 1996. Past and present distribution of stoneworts (Characeae) in The Netherlands. Hydrobiologia 340, 127-135.
Sinha, S., Basant, A., Malik, A. and Singh, K.P., 2009. Iron-induced oxidative stress in a macrophyte: a chemometric approach. Ecotoxicology and Environmental Safety 72, 585-595.
Siqueira-Silva, A.I., da Silva, L.C., Azevedo, A.A. and Oliva, M.A., 2012. Iron plaque formation and morphoanatomy of roots from species of restinga subjected to excess iron. Ecotoxicology and Environmental Safety 78, 265-275.
Smayda, T.J., 1971. Normal and accelerated sinking of phytoplankton in the sea. Marine Geology 11, 105-122.
Smith, S.D. and Banke, E.G., 1975. Variation of sea-surface drag coefficient with wind speed. Quarterly Journal of the Royal Meteorological Society 101, 665-673.
Smith, V.H. and Schindler, D.W., 2009. Eutrophication science: where do we go from here? Trends in Ecology and Evolution 24, 201-207.
Smolders, A.J.P., Lamers, L.P.M., Lucassen, E.C.H.E.T., Van der Velde, G. and Roelofs, J.G.M., 2006. Internal eutrophication: How it works and what to do about it – A review. Chemistry and Ecology 22, 93-111.
Smolders, A.J.P., Lamers, L.P.M., Moonen, M., Zwaga, K. and Roelofs, J.G.M., 2001. Controlling phosphate release from phosphate-enriched sediments by adding various iron compounds. Biogeochemistry 54, 219-228.
Smolders, A.J.P., Nijboer, R.C., and Roelofs, J.G.M., 1995. Prevention of sulphide accumulation and phosphate mobilisation by the addition of iron(III)chloride to a reduced sediment: An enclosure experiment. Freshwater Biology 34, 559-568.
Smolders, A.J.P. and Roelofs, J.G.M., 1995. Internal eutrophication, iron limitation and sulphide accumulation due to the inlet of river Rhine water in peaty shallow waters in the Netherlands. Archiv für Hydrobiologie 133, 349-365.
Smolders, A.J.P. and Roelofs, J.G.M., 1996. The roles of internal iron hydroxide precipitation, sulphide toxicity and oxidizing ability in the survival of Stratiotes aloides roots at different iron concentrations in sediment pore water. New Phytologist 133, 253-260.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
190
Snowden, R.E.D. and Wheeler, B.D., 1993. Iron toxicity to fen plant species. Journal of Ecology 81, 35-46.
Snowden, R.E.D. and Wheeler, B.D., 1995. Chemical changes in selected wetland plant species with increasing Fe supply, with specific reference to root precipitates and Fe tolerance. New Phytologist 131, 503-520.
Søndergaard, M., Bjerring, R. and Jeppesen, E., 2013. Persistent internal phosphorus loading during summer in shallow eutrophic lakes. Hydrobiologia 710, 95-107.
Søndergaard, M., Bruun, L., Lauridsen, T.L., Jeppesen, E. and Vindbæk Madsen, T., 1996. The impact of grazing waterfowl on submerged macrophytes. In situ experiments in a shallow eutrophic lake. Aquatic Botany 53, 73-84.
Søndergaard, M., Jensen, J.P. and Jeppesen, E., 2003. Role of sediment and internal loading of phosphorus in shallow lakes. Hydrobiologia 506/509, 135-145.
Søndergaard, M., Jeppesen, E., Lauridsen, T.L., Skov, C., Van Nes, E.H., Roijackers, R., Lammens, E. and Portielje, R., 2007. Lake restoration: Successes, failures and long-term effects. Journal of Applied Ecology 44, 1095-105.
Sotero-Santos, R.B., Rocha, O. and Povinelli, J., 2005. Evaluation of water treatment sludges toxicity using the Daphnia bioassay. Water Research 39, 3909-3917.
Souty-Grosset, C., Holdich, D.M., Noël, P.Y., Reynolds, J.D. and Haffner, P. (Eds.), 2006. Atlas of crayfish in Europe. Muséum national d’Histore naturelle, Paris.
Spears, B.M., Dudley, B., Reitzel, K. and Rydin, E., 2013. Geoengineering in lakes: a call for consensus. Environmental Science & Technology 47, 3953-3954.
Spencer, D. F. and Rejmánek, M, 2010. Competition between two submersed aquatic macrophytes, Potamogeton pectinatus and Potamogeton gramineus, across a light gradient. Aquatic Botany 92, 239-244.
Spijkerman, E., Barua, D., Gerloff-Elias, A., Kern, J., Gaedke, U. and Heckathorn, S.A., 2007. Stress responses and metal tolerance of Chlamydomonas acidophila in metal-enriched lake water and artificial medium. Extremophiles 11, 551-562.
St-Cyr, L. and Campbell, P.G.C., 1996. Metals (Fe, Mn, Zn) in the root plaque of submerged aquatic plants collected in situ: Relations with metal concentrations in the adjacent sediments and in the root tissue. Biogeochemistry 33, 45-76.
Strand, J.A. and Weisner, S.E.B., 2001. Dynamics of submerged macrophyte populations in response to biomanipulation. Freshwater Biology 46, 1397-1408.
Ter Heerdt, G.N.J., Geurts, J.J.M., Immers, A.K., Colin, M., Olijhoek, P., Yedema, E., Baars, E. and Voort, J.W., 2012. Iron suppletion in peat leakes, STOWA, Amersfoort. (In Dutch)
Ter Heerdt, G. and Hootsmans, H., 2007. Why biomanipulation can be effective in peaty lakes. Hydrobiologia 584, 305-316.
Thomas, P.J., Carpenter, D., Boutin, C. and Allison, J.E., 2014. Rare earth elements (REEs): Effects on germination and growth of selected crop and native plant species. Chemosphere 96, 57-66.
Tilman, D., Fargione, J., Wolff, B., D’Antonio, C., Dobson, A., Howarth, R., Schindler, D., Schlesinger, W.H., Simberloff, D. and Swackhamer, D., 2001. Forecasting agriculturally driven global environmental change. Science 292, 281-284.
Timms, R.M. and Moss, B., 1984. Prevention of growth of potentially dense phytoplankton populations by zooplankton grazing in the presence of zooplanktivorous fish in a shallow wetland ecosystem. Limnology and Oceanography 29, 472-486.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
191
R
Tkalich, P. and Chan, E.S., 2002. Vertical mixing of oil droplets by breaking waves. Marine Pollution Bulletin 44, 1219-1229.
Utkilen, H. and Gjolme, N., 1995. Iron-Stimulated Toxin Production in Microcystis aeruginosa. Applied and Environmental Microbiology 61, 797–800.
Van Anholt, R.D., Spanings, F.A.T., Knol, A.H., Van der Velden, J.A. and Wendelaar Bonga, S.E., 2002. Effects of iron sulfate dosage on the water flea (Daphnia magna Straus) and early development of carp (Cyprinus carpio L.). Archives of Environmental Contamination and Toxicology 42, 182-192.
Van de Haterd, R.J.W. and Ter Heerdt, G.N.J., 2007. Potential for the development of submerged macrophytes in eutrophicated shallow peaty lakes after restoration measures. Hydrobiologia 584, 277-290.
Van den Berg, M.S., Coops, H., Meijer, M.-L., Scheffer, M. and Simons, J. 1998a. Clear water associated with a dense Chara vegetation in the shallow and turbid Lake Veluwemeer, The Netherlands. In: Jeppesen, E., Søndergaard, M. Søndergaard, M. and Christoffersen, K. (Eds.), The structuring role of submerged macrophytes in lakes. Ecological Studies 131, p. 339-352. Springer Verlag, Berlin.
Van den Berg, M.S., Scheffer, M., Coops, H. and Simons, J., 1998b. The role of Characean algae in the management of eutrophic shallow lakes. Journal of Phycology 34, 750-756.
Van den Berg, M.S., Scheffer, M., Van Nes, E. and Coops, H., 1999. Dynamics and stability of Chara sp. and Potamogeton pectinatus in a shallow lake changing in eutrophication level. Hydrobiologia 408/409, 335-342.
Van den Berg, M.S., Coops, H., Simons, J. and Pilon, J., 2002. A comparative study of the use of inorganic carbon resources by Chara aspera and Potamogeton pectinatus. Aquatic Botany 72, 219-233.
Van der Does, J., Verstraelen, P., Boers, P., Van Roestel, J., Roijackers, R. and Moser, G., 1992. Lake restoration with and without dredging of phosphorus-enriched upper sediment layers. Hydrobiologia 233, 197-210.
Van der Wal, J.E.M., Dorenbosch, M., Immers, A.K., Vidal Forteza, C., Geurts, J.J.M., Peeters, E.T.H.M., Koese, B. and Bakker, E.S., 2013. Invasive crayfish threaten the development of submerged macrophytes in lake restoration. PLoS One 8, e78579.
Van der Welle, M.E.W., Cuppens, M., Lamers, L.P.M. and Roelofs, J.G.M., 2006. Detoxifying toxicants: Interactions between sulfide and iron toxicity in freshwater wetlands. Environmental Toxicology and Chemistry 25, 1592-1597.
Van der Welle, M.E., Niggebrugge, K., Lamers, L.P. and Roelofs, J.G., 2007a. Differential responses of the freshwater wetland species Juncus effusus L. and Caltha palustris L. to iron supply in sulfidic environments. Environmental Pollutution 147, 222-230.
Van der Welle, M.E.W., Smolders, A.J.P., Op den Camp, H.J.M., Roelofs, J.G.M. and Lamers, L.P.M., 2007b. Biogeochemical interactions between iron and sulphate in freshwater wetlands and their implications for interspecific competition between aquatic macrophytes. Freshwater Biology 52, 434-447.
Van Dijk, G.M. and Van Vierssen, W., 1994. Survival of a Potamogeton pectinatus L. population under various light conditions in a shallow eutrophic lake (Lake Veluwe) in The Netherlands. Aquatic Botany 39, 121-129.
Van Donk, E., Grimm, M.P., Gulati, R.D. and Klein Breteler, J.P.G., 1990. Whole-lake food-web manipulation as a means to study community interactions in a small ecosystem. Hydrobiologia 191, 275-289.
Van Donk, E., Grimm, M.P., Heuts, P.G.M., Blom, G., Everards, K. and Van Tongeren, O.F.R., 1994. Use of mesocosms in a shallow eutrophic lake to study the effects of different restoration measures. Archiv für Hydrobiologie 40, 283-294.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
192
Van Donk, E. and Otte, A., 1996. Effects of grazing by fish and waterfowl on the biomass and species composition of submerged macrophytes. Hydrobiologia 340, 285-290.
Van Donk, E. and Van de Bund, W.K., 2002. Impact of submerged macrophytes including charophytes on phyto- and zooplankton communities: allelopathy versus other mechanisms. Aquatic Botany 72, 261-274
Van Giels, J., 2011. Verspreidingsonderzoek kreeften Terra Nova en Loenderveense Plas. ATKB rapport 20090236/02, Geldermalsen, Netherlands. (In Dutch)
Van Oosterhout, F. and Lürling, M., 2011. Effects of the novel ‘Flock & Lock’ lake restoration technique on Daphnia in Lake Rauwbraken (The Netherlands). Journal of Plankton Research 33, 255-263.
Van Liere, L. and Janse, J.H., 1992. Restoration and resilience to recovery of the Lake Loosdrecht ecosystem in relation to its phosphorus flow. Hydrobiologia 233, 95-104.
Verberk, W.C.E.P., Munckhof, P.J.J. and Pollux, B.J.A., 2012. Niche segregation in two closely related species of stickleback along a physiological axis: explaining multidecadal changes in fish distribution from iron-induced respiratory impairment. Aquatic Ecology 46, 241-248.
Verschoor, A.M., Takken, J., Massieux, B. and Vijverberg, J. The Limnotrons: a facility for experimental community and food web research. Hydrobiologia 491, 357-377.
Visser, P.M., Ibelings, B.W., Mur, L.R. and Walsby, A.E., 2005. The ecophysiology of the harmful cyanobacterium Microcystis: features explaining its success and measures for its control. In: Huisman, J., Matthijs, H.C.P., Visser, P.M. (Eds.), Harmful cyanobacteria. p. 109-142.Springer, Dordrecht.
Visser, P.M., Massaut, L., Huisman, J. and Mur, L.R., 1996. Sedimentation losses of Scenedesmus in relation to mixing depth. Archiv für Hydrobiologie 136, 289-308.
Vuori, K.-M., 1995. Direct and indirect effects of iron on river ecosystems. Annales Zoologici Fennici 32, 317-329.
Walker, C.H., Sibly, R.M., Hopkin, S.P. and Peakall, D.B., 2012. Principles of ecotoxicology, Fourth Edition. 385 pp. Taylor & Francis Group, London.
Walker, W.W.J., Westerberg, C.E., Schuler, D.J. and Bode, J.A., 1989. Design and evaluation of eutrophication control measures for the St. Paul water supply. Lake and Reservoir Management 5, 71-83.
Wallace, B.B. and Hamilton, D.P.,1999. The effect of variations in irradiance on buoyancy regulation in Microcystis aeruginosa. Limnology and Oceanography 44, 273-381.
Walsby, A.E., 1992. The control of gas-vacuolate cyanobacteria. In: Sutcliffe, D.W., Gwynfryn, J. (Eds.), Eutrophication: Research and application to water supply. p. 150-162. Freshwater Biological Association, Ambleside.
Walsby, A.E., Kinsman, R., Ibelings, B.W. and Reynolds, C.S., 1991. Highly buoyant colonies of the cyanobacterium Anabaena lemmermanii form persistent surface waterblooms. Archiv für Hydrobiologie 121, 261-280.
Walsby, A.E., 1994. Gas vesicles. Microbiological Reviews 58, 92-144.Webster, I.T. and Hutchinson, P.A., 1994. Effect of wind on the distribution of phytoplankton cells in lakes revisited. Limnology and Oceanography 39, 365-373
Welch, E.B. and Cooke, G.D., 1999. Effectiveness and longevity of phosphorus inactivation with alum. Lake and Reservoir Management 15, 5-27.
Wepener, V., Van Vuren, J.H.J. and Du Preez, H.H., 1992. Effect of manganese and iron at a neutral and acidic pH on the hematology of the Banded Tilapia (Tilapia sparrmanii). Bulletin of Environmental Contamination and Toxicology 49, 613-619.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
References
193
R
Wheeler, B.D., Al-Farraj, M.M. and Cook, R.E.D., 1985. Iron toxicity to plants in base-rich wetlands: Comparative effects on the distribution and growth of Epilobium hirsutum L. and Juncus subnodulosus Schrank. New Phytologist 100, 653-669.
Wilk-Wozniak, E., Bucka, H. and Mrozinska, T., 2003. Contribution to a broadening of taxonomical and ecological knowledge on Woronichinia naegeliana (Unger) Elenkin. Algological Studies 109, 609-615.
Winterwerp, J.C., Uittenbogaard, R.E. and De Kok, J.M., 2001. Rapid siltation from saturated mud suspensions. In: McAnally, W.H., Mehta, A.J. (Eds.), Coastal and Estuarine Fine Sediment Processes. Proceedings in Marine Science, Vol. 3, p. 125-146. Elsevier Science, Amsterdam.
Wüest, A. and Lorke, A., 2003. Small-scale hydrodynamics in lakes. Annual Review of Fluid Mechanics 35, 373-412.
Zak, D., Gelbrecht, J. and Steinberg, C.E.W., 2004. Phosphorus retention at the redox interface of peatlands adjacent to surface waters in northeast Germany. Biogeochemistry 70, 357-368.
Zimba, P. V., Hopson, M. S., 1997. Quantification of epiphyte removal efficiency from submersed aquatic plants. Aquatic Botany 58, 173-179.
Zülicke, C., Hagen, E. and Stips, A., 1998. Dissipation and mixing in a coastal jet: A Baltic Sea case study. Aquatic Sciences 60, 220-235.
CURRICULUM VITAE
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
Curriculum Vitae
196
CURRICULUM VITAE
Anne was born on the 24th of September 1985 in Delft, The Netherlands. After finishing her
secondary school in Delft in 2004, she moved to Leiden to start studying Biology at Leiden
University. Her first internship (working with the fascinating ‘Marimo’) quickly spiked her
interest for the aquatic world which, after obtaining her Bachelor diploma, caused her to
switch universities in order to start the Master Limnology & Oceanography at the University
of Amsterdam. During the first year of her Master study she focused on competition studies
between toxic and non-toxic cyanobacterial species at the Institute for Biodiversity and Ecosystem
Dynamics (IBED) in Amsterdam. During the succeeding year toxicity studies were continued,
but this time at the Cawthron institute in Nelson, New Zealand, where she focused on toxic
dinoflagellates and their effects on fish, mussels and oysters. After graduation in 2009 she started
her PhD-project at The Netherlands Institute of Ecology (NIOO). Her research focused on
preventing and predicting cyanobacterial blooms, and resulted in this thesis. As of August 2014,
Anne works as an ecologist at the Dutch drinkingwater company Vitens.
LIST OF PUBLICATIONS
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
List of publications
198
LIST OF PUBLICATIONS
Published
Boedeker, C., Eggert, A. Immers, A. and Smets, E., 2010. Global decline of and threats to
Aegagropila linnaei, with special reference to the lake ball habit. Bioscience 60, 187-198.
Boedeker, C., Eggert, A., Immers, A. and Wakana, I., 2010. Biogeography of Aegagropila linnaei
(Cladophorophyceae, Chlorophyta): a widespread freshwater alga with low effective dispersal
potential shows a glacial imprint in its distribution. Journal of Biogeography 37, 1491-1503.
Boedeker, C. and Immers, A., 2009. No more lake balls (Aegagropila linnaei Kützing,
Cladophorophyceae, Chlorophyta) in The Netherlands? Aquatic Ecology 43, 891-902.
Immers, A.K., Van der Sande, M.T., Van der Zande, R.M., Geurts, J.J.M., Van Donk, E. and
Bakker, E.S., 2013. Iron addition as a shallow lake restoration measure: impacts on charophyte
growth. Hydrobiologia 710, 241-251.
Immers, A.K., Vendrig, K., Ibelings, B.W., Van Donk, E., Ter Heerdt, G.N.J., Geurts, J.J.M.
and Bakker, E.S., 2014. Iron addition as a measure to restore water quality: implications for
macrophyte growth. Aquatic Botany 116, 44-52.
Shi, F., McNabb, P., Rhodes, L., Holland, P., Webb, S., Adamson, J., Immers, A., Gooneratne,
R. and Holland, J., 2012. The toxic effect of three dinoflagellate species from the genus Karenia
on invertebrate larvae and finfish. New Zealand Journal of Marine and Freshwater Research 46,
149-165.
Ter Heerdt, G, Geurts, J., Immers, A., Colin, M., Olijhoek, P., Yedema, E., Baars, E., Voort,
J.W., 2012. IJzersuppletie in laagveenplassen: De resultaten. 2012-43, STOWA. (In Dutch)
Van de Waal, D.B., Verspagen, J.M.H., Finke, J.F., Vournazou, V., Immers, A.K., Kardinaal,
W.E.A., Tonk, L., Becker, S., Van Donk, E., Visser, P.M, Huisman, J., 2011. Reversal in
competitive dominance of a toxic versus non-toxic cyanobacterium in response to rising CO2.
The ISME Journal 5, 1438–1450.
Van der Wal, J.E.M., Dorenbosch, M., Immers, A.K., Vidal Forteza, C., Geurts, J.J.M., Peeters,
E.T.H.M., Koese, B. and Bakker, E.S., 2013. Invasive crayfish threaten the development of
submerged macrophytes in lake restoration. PLoS One 8, e78579.
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
R26
R27
R28
R29
R30
R31
R32
R33
R34
R35
R36
R37
R38
R39
List of publications
199
L
Submitted
Immers, A.K., Bakker, E.S., Van Donk, E., Ter Heerdt, G.N.J., Geurts, J.J.M. and Declerck,
S.A.J., 2014. Fighting internal phosphorus loading: An evaluation of the large scale application
of gradual Fe-addition to a shallow peat lake.
Immers, A.K., Van Donk, E. and Bakker, E.S., 2014. Lake restoration by in-lake iron addition:
A review of iron impact on aquatic organisms and lake ecosystems.
N E T H E R L A N D S I N S T I T U T E O F E C O L O G Y
Preven
ting or p
redictin
g cyanob
acterial bloom
sN
IOO
Th
esis 115A
nn
e K. Im
mers
Invitation to attend the public defence of my thesis:
Preventing or predicting cyanobacterial blooms
Iron addition as a whole lake restoration tool
Monday December 22nd
at 14.30
SenaatskamerUtrecht University
Domplein 29Utrecht
Anne K. [email protected]
Paranymphs:
Tânia Vasconcelos [email protected]
Dennis [email protected]
Reception to follow
Preventing or predicting cyanobacterial blooms
Iron addition as a whole lake restoration tool
Anne K. Immers