Arch. Hydrobiol.
162
3
349–362
Stuttgart, March 2005
Interactions between sediment propagule banks
and sediment nutrient fluxes explain floating plant
dominance in stagnant shallow waters
Ger Boedeltje, Alfons J. P. Smolders, Leon P. M. Lamers and
Jan G. M. Roelofs1
Radboud University Nijmegen, Department of Ecology, Research group of
Aquatic Ecology and Environmental Biology
With 1 figure and 2 tables
Abstract: Mats of floating plants are known to have detrimental effects on aquatic life
as they reduce the penetration of light so that submerged species are unable to develop.
These mats also prevent gaseous exchange between water and atmosphere resulting in
anoxic conditions of the water layer. In shallow waters, nutrient fluxes from the sediment to the water layer are expected to play a key role in the shift to floating plant
dominance. This study investigated the potential recruitment and development of
duckweed mats from propagule banks in field enclosures containing either muddy or
sandy sediment. It was observed that free-floating fronds of Lemna minor were derived
from both sediment types, but that dense duckweed mats developed exclusively in
enclosures with muddy sediment. The results can be explained by differences in nutrient release from the sediments to the overlying water during summer, when high
fluxes, especially of phosphorus and nitrogen, were observed almost exclusively in enclosures with muddy sediment involving anoxia of the water layer. This study discusses the results in relation to free-floating plant dominance and the management of shallow waters.
Key words: anoxia, duckweed, eutrophication, Lemna minor, nitrogen flux, phosphorus flux, seed bank, stable state.
Introduction
Mats of floating plants including duckweeds (Lemnaceae) are known to have
detrimental effects on aquatic life. The penetration of light is strongly reduced
1
Authors’ address: Radboud University Nijmegen, Department of Ecology, Research Group of Aquatic Ecology and Environmental Biology, Toernooiveld 1, 6525
ED Nijmegen, The Netherlands; E-mail: [email protected]
DOI: 10.1127/0003-9136/2005/0162-0349
0003-9136/05/0162-0349 $ 3.50
 2005 E. Schweizerbart’sche Verlagsbuchhandlung, D-70176 Stuttgart
350
G. Boedeltje et al.
so that submerged species are unable to develop (Sculthorpe 1967) and gaseous exchange between water and atmosphere is precluded resulting in anoxic
conditions of the water layer (Morris & Barker 1977, Janse & Van Puijenbroek 1998). Therefore, restoration efforts focussed on reducing freefloating plant dominance in favour of recovering submerged vegetation will
highly enhance aquatic biodiversity.
Dominance of Lemnaceae is related to high nutrient (P and N) concentrations of the water column (De Groot et al. 1987). In experimental ditches receiving different levels of external nutrient input, it was found that the submerged Elodea nuttallii (Planch.) St. John was replaced by the free-floating
Lemna minor L. at high nutrient loading (Portielje & Roijackers 1995). In
competition experiments including E. nuttallii and Lemna gibba L., Scheffer
et al. (2003) demonstrated as well that L. gibba became dominant exclusively
at high nutrient (N) concentrations of the water, E. nuttallii at low concentrations, whereas at moderate nutrient concentrations the experiments ended in
either floating or submerged plant dominance depending on the initial biomass
of the species. Reducing the external and internal nutrient load of shallow waters is therefore required for the shift from free-floating to submerged vegetation.
After reducing the external nutrient input and mechanical removal of duckweed layers, sediments may play the major role in (renewed) development and
maintenance of duckweed layers. First, sediment propagule banks can be a
source of duckweed fronds from which free-floating vegetation may derive.
Second, nutrient-rich sediment can significantly impact the nutrient loading of
shallow waters (Van Luijn et al. 1999, Wetzel 2001) and therefore the development of duckweed mats from single propagules would not be nutrient-limited. Whether, at high nutrient concentrations, the presence of duckweed propagules results ultimately in dense, floating layers, additionally depends on
wind and wave action (Scheffer et al. 2003).
This study investigated the potential development of duckweed mats from
Lemna-rich propagule banks in field enclosures with two contrasting sediments. It was conducted in a relatively exposed, eutrophic backwater along a
navigation canal where thick layers of nutrient-rich, muddy sediment were deposited (Boedeltje et al. 2001). It was shown that similar densities of Lemnapropagules occurred in both superficial (muddy) and deeper (sandy) sediments
(Boedeltje et al. 2003), but this system has never been faced with duckweeddominance. By contrast, the water layer of these sites is dominated by planktonic algae and cyanobacteria in summer (Van Beek 2000). Its Lemna-free
state is likely the result of boat-induced water currents as in more isolated
parts of these backwaters, where such currents are absent, duckweed dominance does occur (Boedeltje et al. 2001). To obtain two contrasting sediments, in four out of eight enclosures the muddy layer was removed and here
Nutrient fluxes and floating plant dominance
351
the initially sandy sediment surfaced, in the others the muddy layer was not removed. During one year, the effect of the presence or absence of muddy sediment was investigated on: (i) temporal variation of nutrient concentrations in
pore water and the water column, (ii) vegetation development and (iii) nutrient
fluxes from sediments to the water column.
It was hypothesized that the Lemna-free state would turn to a Lemna-dominated state in enclosures with muddy sediments. In contrast, such shift was not
expected to occur when sandy sediments with similar densities of Lemnapropagules would be present, because nutrient-fluxes from sandy sediments
are small compared to muddy sediments (Van Luijn et al. 1999). The results
are discussed in relation to nature restoration and development strategies to
avoid floating plant dominance and concomitant biodiversity loss.
Material and methods
Study site
From December 2000 until December 2001 inclusive, the study was conducted in a
wave-protected water zone (backwater) along the Twentekanaal, a navigated canal in
the Netherlands (52˚ 11′ N, 6˚ 28′ W). This 200 m long, 6 m wide and 0.5 m deep backwater is separated from the main water body by steel sheet piles (Boedeltje et al.
2001). Three 0.5 m wide gaps in these piles facilitate water exchange between backwater and canal. Apart from the surroundings of the gaps, deposition of resuspended
sediment derived from the navigated canal and of organic matter takes place across the
backwater resulting in the accumulation of muddy sediment. In 2000, the initially
sandy bottom or sediment was covered by 13.6 ± 3.1cm (mean ± SE) muddy sediment.
Percentage loss on ignition of this muddy sediment was significantly higher than that
of the uppermost part (0 – 5 cm) of the sandy sediment; the same holds for total P, N, Fe
and Mn (Table 1).
The mean number m – 2 of Lemna minor turions (212 ± 82) found in the surface (0 –
5 cm) layer of the muddy sediment did not significantly differ from that of the uppermost (0 – 5 cm) part of the sandy sediment (262 ± 216) (Boedeltje et al. 2003). The
water margins of the backwater were characterised by stands of Phragmites australis
Table 1. Concentrations (means + SE; n = 4) of parameters measured in the sandy and
muddy sediment (0 – 5 cm) in January 2002. d = P ≤ 0.0001 (one-way ANOVA).
Sandy sediment
Loss on ignition (%)
N (µmol g – 1 DW)
P (µmol g – 1 DW)
Fe (µmol g – 1 DW)
Mn (µmol g – 1 DW)
Muddy sediment
Mean
SE
Mean
SE
1.3
28.6
8.2
101
1.5
(0.4)
(14.3)
(2.3)
(20)
(0.4)
9.2
193
84.9
657
16.4
(0.7)
(21)
(6.5)
(49)
(1.2)
Sign
d
d
d
d
d
352
G. Boedeltje et al.
(Cav.) Steud. Potamogeton pectinatus L. and L. minor occurred occasionally in the
open water.
Experimental set-up and sample collection
Eight enclosures consisting of transparent polycarbonate cylinders (Ø 0.5 m, height
1.3 m, open on both sides) were pushed into the sediment of the backwater to a depth
of 50 cm. In four replicates the muddy sediment layer was present; in four others this
layer had been conscientiously removed by hand. The enclosures were alternately
placed in a line, 3 m from the sheet piles.
To obtain pore water, two ceramic soil moisture samplers (Eijkelkamp Agrisearch
Equipment) were placed into the sediment of each enclosure at depths of 0–10 cm.
Each sampler was connected to a polyethylene tubing protruding above water surface.
Samples were obtained with syringes connected to the free end of the tubing. The first
50 ml were discarded to exclude dead volume water. At the same date, surface water
samples were collected in iodated polyethylene bottles at 20 cm below water surface.
Sampling occurred monthly between 8 : 00 and 10 : 00 a. m. In May, June and July, additional water samples were taken for chlorophyll-a determinations.
Monthly, cover percentages of L. minor and floating algae were estimated. In addition, potentially emerging submerged plants were counted. In September, samples of
L. minor were obtained by pressing an edge cylinder (Ø 11 cm; open at both sides)
through the duckweed mat and collecting the plants inside this cylinder. To test
whether and how many propagules were present in the propagule banks after the experiment had ended, five sediment cores (surface area 100 cm2, 5 cm depth) per enclosure were collected for germination experiments. Per enclosure the cores were pooled
into one sample and treated according to the seedling emergence technique of Boedeltje et al. (2002). For chemical analyses, three additional sediment samples (0 –
5 cm) were taken from each enclosure using a Plexiglas coring tube (Ø 60 mm) and
pooled.
Sample handling and analyses
Temperature and O2 concentration of the water layer were recorded in situ at 20 cm below water surface with a WTW CellOx 325 sensor, pH with a WTW SenTix 41-3 electrode. Redox potential (the average of 3 replicates per enclosure) was determined in
the upper 5 – 8 cm of the sediment with a WTW Pt Ag/AgCl electrode and converted to
the normal hydrogen potential (Eh). Water column chlorophyll-a concentrations were
determined according to Roijackers (1981). Surface water was filtered through a
Whatman GF/C filter. Citric acid was added to a final concentration of 0.6 mmol l –1 to
prevent metal precipitation. Because of high Fe-concentrations, additional pore water
sub-samples were acidified with 65 % HNO3 (200 µl HNO3 per 20 ml pore water).
Duckweed and sediment samples were dried at 70 ˚C and 105 ˚C, respectively.
Duckweed-biomass of the enclosures was estimated by multiplying the measured dry
weight of each sample by the surface area-ratio of enclosure and sample cylinder. Of
each dried sample, 200 mg were digested in a mixture of 1ml 30 % H2O2 and 4 ml 65 %
HNO3 in a Milestone microwave type mls 1200 Mega and finally diluted to 100 ml
Nutrient fluxes and floating plant dominance
353
with bidistilled water. Herein total concentrations of dissolved Fe, Mn and P were determined by inductively-coupled plasma emission spectrophotometry (Spectroflame
ICP). Ortho-phosphate (PO43 – ) concentrations were measured colorimetrically with a
Technicon AA II system, using ammonium molybdate (Henriksen 1965). Nitrate
(NO3 – -N) and ammonium (NH4 + -N) were determined colorimetrically with a Traacs
800 + auto-analyser, using hydrazine sulphate (Technicon 1969) and salicylate (Grasshoff & Johannsen 1977), respectively. Organic matter was determined as loss on ignition (5 h at 550 ˚C). N and C concentrations of L. minor were measured in dried samples using a CNS analyzer (Carlo Erba Instruments NA1500).
Data analyses
The net flux of N, P, Fe and Mn from the sediment to the water column in each enclosure was estimated for two periods of 150 days each (December – May and May – October) using the equation:
F=
C150 – C0 V CLemna 150 × BLemna 150 – CLemna 0 × BLemna 0 1
× +
×
t150 – t0 A
t150 – t0
A
F = net flux (mmol m – 2 d – 1)
C0, C150 = concentrations (mmol l – 1) of the water column at time t0 and t150
t150 – t0 = time (in days)
V = volume of the cylinder (l)
A = area of the cylinder (m2)
CLemna 0, CLemna 150 = Concentration (mmol g – 1 DW) in L. minor tissue at time t0 and t150
BLemna 0, BLemna 150 = Biomass (g DW) of L. minor at time t0 and t150
SPSS version 11.0 (SPSS Inc., Chicago, USA) was used to perform statistical analyses.
Percentages were arcsine square-root transformed and other data were log (x + 1) transformed to increase homogeneity of variances and normality (Zar 1999). To examine
the response of (pore)water parameters and L. minor to the treatment during the research period, the GLM (General Linear Models) procedure for repeated measures,
model one-way ANOVA, was used. A conservative P-value (Greenhouse-Geisser) was
calculated to examine the significance of time and interaction effects. Differences at a
given time were tested with a one-way ANOVA (α = 0.05). Differences between the
mean number of propagules per sediment type were also tested using a one-way
ANOVA. Homogeneity of variance was tested and established with a Levene’s test.
Results
Temporal variation in (pore)water parameters
At the start, redox potential (Eh) of muddy sediments was significantly (F =
71.2, P < 0.0001) lower than that of sandy sediments and this difference continued to exist until the end of the experiment (Fig. 1). During spring and summer, the average Eh of muddy sediments dropped from c. 45 mV in March to
354
G. Boedeltje et al.
.
.
.
.
.
.
.
.
.
.
Fig. 1. Course of sediment-redox potential (Eh), Lemna minor-cover and some (pore)water parameters from December 2000 – December 2001. Means ( +SE) are given (n = 4).
Nutrient fluxes and floating plant dominance
355
values between – 10 and 0 mV in the period July – September, whereas Eh of
sandy sediments, by contrast, decreased from c. 200 mV to 80 mV in the similar period (Fig. 1). From October onwards, Eh increased again for both treatments.
From January onwards, a thin layer of filamentous algae became attached
to the inner walls and sediments of all enclosures. As a result, the O2 concentration of the water column of all enclosures increased between January and
April (Fig. 1). From April – June, the O2 concentration decreased in both treatments but was still elevated compared to that of the start (F = 8.7, P < 0.01;
Fig. 1). Between July and November, the O2 concentration of the enclosures
with muddy sediment was significantly (F = 93.4, P < 0.0001) lower than that
of the enclosures with sandy sediment (Fig. 1). In September, the water column of the enclosures with muddy sediment was nearly anoxic with O2 concentrations ranging from 6 to 16 µmol l – 1. From December – May, pH of the
water layer increased from c. 7.4 to 9.7, but it decreased to c. 8.0 for both treatments in June – July (Fig. 1). Between July and December, the pH of the enclosures with muddy sediment was significantly (F = 37.6; P < 0.001) lower than
that of the enclosures with sandy sediment (Fig. 1).
For both treatments, the average NO3 – concentrations of the surface water
dropped rapidly from c. 400 µmol l – 1 at the start to < 1 µmol l – 1 in May (Fig. 1).
Pore water NO3 – concentrations of the sandy sediment showed a similar
course; those of the muddy sediment were below 1 µmol l – 1 throughout the experiment (Fig. 1). Pore water NH4 + concentrations were significantly higher in
the muddy sediment compared to those of the sandy sediment (Appendix 1;
Fig. 1). During summer, NH4 + levels in the pore water of muddy sediments increased to values as high as 1200 µmol l – 1 (Fig. 1). As to the NH4 + concentrations of the surface water, there was a significant sediment effect (Appendix
1), which was notably visible in July. Then, the NH4 + concentrations in enclosures with muddy sediment were on average 13 µmol l – 1, significantly (F =
27.2, P = 0.002) higher than those in the enclosures with sandy sediment
(3 µmol l – 1).
Throughout the experiment, pore water Fe concentrations were higher in
muddy sediment than in sandy sediment (Appendix 1; Fig. 1). The same held
for Mn and P (Appendix 1; graphs not shown). From July – September, surface
water Fe and PO43 – concentrations were significantly higher in the enclosures
with muddy sediment than in the enclosures with sandy sediment (F = 65.4, P
< 0.0001 and F = 47.3, P < 0.0001, respectively; Fig. 1). For Mn and P concentrations, similar peaks were measured in the surface water of enclosures with
muddy sediment only (graphs not shown).
356
G. Boedeltje et al.
Vegetation development and propagule bank
From June onwards, c. 30 ± 15 % (SE) of the water surface of the enclosures
with muddy sediment was covered by filamentous algae. In contrast, no algal
growth was observed at the water surface of the enclosures with sandy sediment. On average, chlorophyll-a concentration did not change significantly (F
= 1.12, P = 0.33) over the period it was determined (May – July) and no sediment effect could be demonstrated (F = 0.43, P = 0.53). Over this period, its
concentration was on average 9.9 ± 2.4 µg/l (n = 24). Apart from Zannichellia
palustris L. which emerged in May from the muddy sediment in only one enclosure, no other submerged and emergent aquatic species were observed.
In April, the first fronds of L. minor became visible in all enclosures, but
during spring and early summer no further growth was observed. In August,
however, all enclosures with muddy sediment were covered by a duckweed
mat, whereas on the water surface of enclosures with sandy sediment only few
duckweed plants were present (Fig. 1). In September, the average L. minor biomass m – 2 was estimated at 537.6 ± 61.1 g DW for muddy sediments, versus 1.0
± 1.0 g DW for sandy sediments. In this month, Z. palustris had disappeared
from the water column.
Germination experiments of sediment samples showed that L. minor propagules were present in all enclosures but that the average densities were significantly (F = 14.4; P < 0.01) higher in muddy sediment (11985 ± 7373 m – 2) than
in sandy sediment (265 ± 219 m – 2). Moreover, propagule density of L. minor
in muddy sediments was significantly (F = 15.5, P < 0.01) higher than the density estimated before the experiment started (214 ± 82 m – 2). Over the same period, L. minor propagule density in the sandy sediments did not change (F =
0.2, P = 0.9). From muddy sediments only, Z. palustris emerged at low densities (25 ± 19 m – 2). Additionally, propagules of several semi-aquatic and terrestrial species were found (results not shown).
Nutrient fluxes from sandy and muddy sediments
From December to May, there was a net loss of N from the water column of c.
1.60 mmol m – 2 d – 1 for both treatments (Table 2). Net P-losses were not observed and net losses of Fe and Mn from the water layer were very low
(0.02 mmol m – 2 d – 1) for both treatments (Table 2).
From May to October, there were relatively high net fluxes of N, P, Fe and
Mn from the muddy sediment to the water column (Table 2). Net N and P
fluxes to the water column were estimated at 4.35 and 0.46 mmol m – 2 d – 1, respectively (Table 2). In enclosures with sandy sediment, by contrast, a negative net N flux (of 0.05 mmol m – 2 d – 1) to the sediment was still found,
whereas for P, Fe and Mn no or very low net fluxes to the water column were
found (Table 2).
357
Nutrient fluxes and floating plant dominance
Table 2. Net fluxes of N, P, Fe and Mn from sediments to the water column for December – May (= period 1) and for May – October (= period 2). Data represent average
amounts ( + SE) dissolved in the water column and fixed in L. minor biomass. For biomass in the enclosures with sandy sediment: n = 2, otherwise: n = 4; × = Lemna biomass absent.
Parameter
Sediment
Period
N
Sandy
1
P
Muddy
2
1
Sandy
2
1
Fe
Muddy
2
1
Sandy
2
1
Mn
Muddy
2
1
Sandy
2
1
Dissolved – 1.56 – 0.06 – 1.64 0.02 0.00 0.01 0.00 0.01 – 0.02 0.01 – 0.02 0.08 – 0.02
SE
0.04 0.04 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.00
In biomass ×
0.01
×
4.33 × < 0.01 × 0.45
×
< 0.01
×
0.75
×
SE
0.00
0.39
0.14
0.16
Total
– 1.56 – 0.05 – 1.64 4.35 0.00 0.01 0.00 0.46 – 0.02 0.01 – 0.02 0.85 – 0.02
Muddy
2
1
2
0.00 – 0.02 0.18
0.00 0.00 0.06
0.00
×
0.26
0.00
0.09
0.00 – 0.02 0.44
Discussion
The hypothesis that there would be a shift from a Lemna-free to a Lemnadominated state in enclosures with muddy sediment, was supported by the results. Because at the start of the experiment no duckweed fronds were observed, the plants found at the water surface from April onwards, must have
originated from dispersal or from propagule banks. Given the presence of
wave-breaking sheet piles between backwater and canal and the height of the
cylinders above the water level, it is unlikely that L. minor fronds were dispersed into the enclosures from the main water body. Although waterbirdmediated diaspore transport of duckweeds is also known (Ridley 1930), it is
most likely that L. minor fronds were derived from propagule banks. According to Boedeltje et al. (2003), the muddy as well as the sandy sediments (0 –
5 cm) contained more than 200 L. minor propagules m – 2. The results of the
germination experiments of sediment samples confirm the presence of L.
minor propagules in the enclosures with sandy sediment.
The significant increase of the surface water PO43 – and NH4 + concentrations of the enclosures with muddy sediment in July preceded the mass development of L. minor which is in line with our second hypothesis that nutrient
fluxes from the sediment to the water column would drive the growth and
multiplication of duckweeds. From April to July, growth of filamentous algae
at the enclosure walls and moderate phytoplankton growth in the water column may have inhibited the growth of duckweeds by their removal of N, P
and Fe from the water layer as well as by their photosynthetic increase of pH
(Szabó et al. 1999, Roijackers et al. 2004). In this period, surface water
PO43 – and N (NH4 + + NO3 – ) concentrations in the enclosures of both sediment types were below or equal to the K50 values of 0.2 and 2.9 µmol l – 1 respectively, the concentrations at which the maximum growth rate of L. minor
358
G. Boedeltje et al.
is reduced by 50 % (Lüönd 1983). In July, however, the PO43 – and N concentrations of the enclosures with muddy sediment were 6 – 10 times higher than
the K50, whereas those of the sandy enclosures did not exceed the K50 for P
and N.
The amount of N, P, Fe and Mn accumulated in L. minor tissue in September, reflects the magnitude of sediment nutrient releases from July onwards
since L. minor plants take up these nutrients exclusively from the water body.
Even if these nutrients would have originated from the mineralization of algae
(cf. Kleeberg & Schlungbaum 1993), they were previously derived from
the water column and/or sediments. Therefore, these data additionally give
evidence for the important role of sediment nutrient fluxes in the development
of duckweed mats.
The estimated N and P release from muddy sediments is within the range
reported for several eutrophic lakes (Marsden 1989, Dudel & Kohl 1992,
Van Luijn et al. 1999). Despite this similarity, eutrophic shallow lakes are
phytoplankton-dominated (Scheffer et al. 2003) whereas stagnant backwaters with muddy sediment, on the other hand, are duckweed-dominated (Boedeltje et al. 2001). This difference may be attributed to wind and wave action
by which the small duckweeds are washed ashore in exposed waters (Scheffer et al. 2003).
The higher proportion of organic matter of the muddy sediment may be
considered an important factor determining differences in (pore)water parameters between the treatments (Sinke et al. 1990, Van Luijn et al. 1999). In the
muddy sediment, microbial O2 consumption was much higher than that in the
sandy sediment as can be deduced from differences in O2 concentration of the
water column during summer. Eh values and dissolved Fe concentrations show
that inside the muddy sediment Fe-reduction took place, making PO43 – more
mobile than in the sandy sediment. Because the release of PO43 – and Fe from
muddy sediments was strong particularly during periods of anoxia and negative Eh, it is likely that anoxic mobilisation occurred (e. g. Boström et al.
1982). The similar course of the exchange curves of surface water Fe and
PO43 – additionally supports this hypothesis. As the entire water column became more or less anoxic for a considerable period of time, both PO43 – and Fe
could accumulate in the water layer. The increase of pore water NH4 + concentrations during summer may have been the result of temperature-dependent
degradation of N-bearing organic matter (Van Luijn et al. 1999). During periods of anoxia the nitrification and the coupled nitrification-denitrification
were likely suppressed (Van Luijn et al.1999) implying that NH4 + was released to the overlying water where it was consumed by L. minor.
According to the hypothesis of Scheffer et al. (2003), rooted submerged
species were expected to grow from the sandy sediment and to dominate in the
nutrient-poor overlying water. It was shown, however, that submerged species
Nutrient fluxes and floating plant dominance
359
were absent from the propagule banks of sandy sediments which corresponds
to previous observations of Boedeltje et al. (2003). Although the observed
replacement of the submerged Z. palustris by the free-floating L. minor at high
nutrient concentrations supports the theory of Scheffer et al. (2003), the low
initial density of propagules and plants do not allow general conclusions.
As to the question whether phytoplankton or L. minor would dominate at
high nutrient concentrations (Scheffer et al. 2003), our results clearly show
the ultimate dominance of the duckweed, despite the relatively high cover of
floating algae in July. This observation supports the hypothesis of Roijackers
et al. (2004) that in the absence of mortality, Lemna will outcompete algae at
high nutrient concentrations. In order to prevent disturbance of the system, information on changes in the biomass of floating algae could not be obtained.
However, the decreased pH and O2 concentration of the water layer suggest
the decline of phytoplankton beneath the duckweed mat.
Wave breaking structures have been constructed in front of canal banks and
lake shorelines to protect the established bank vegetation from erosion and to
enhance the establishment and growth of submerged and emergent macrophytes in shallow backwaters. Although positive effects of such structures on
the establishment and growth of emergent macrophytes in backwaters have
been shown (Foote & Kadlec 1988), negative effects appeared as well, including the deposition of muddy sediments with relatively high contents of organic matter (Rolletschek 1999, Boedeltje et al. 2001). Accumulation
zones with a low degree of water exchange between backwater and bordering
lake or canal are faced with increased O2 consumption rates during summer
resulting in anoxia of the water layer and in enhanced concentrations of soluble sulphides and H2S (Rolletschek 1999). Anoxia and increased nutrient releases from the sediment ultimately may lead to duckweed dominance, as was
shown in this study. Wave damping structures which allow sufficient water exchange between backwater and open water should therefore be preferred in
ecological engineering (cf. Rolletschek 1999). Alternatively, backwaters
may be completely isolated from the main canal after sediment removal. However, due to the lack of connectivity between main canal and backwater, this
may have other disadvantages.
Once established, duckweed mats may influence nutrient fluxes from sediments by maintaining anoxic conditions of the water layer which results in increased nutrient fluxes from the sediment and, consequently, in increased
plant growth. Furthermore, the density of Lemna propagules in the sediment
increases after the development of a Lemna mat. By such positive feedbacks,
the well-known stability of the floating plant state may be maintained for
years. To reach a shift from a duckweed-dominated state to a state dominated
by submerged plants, additional removal of floating plant layers and accumulated muddy sediment is required (Scheffer et al. 2003). This study has
360
G. Boedeltje et al.
clearly shown that sediment removal (dredging) does enhance the habitat quality of shallow (back)waters. The subsequent (re)establishment of rooted submerged species from propagule banks, however, is not likely and depends
largely on dispersal (Boedeltje et al. 2004).
Acknowledgements
The assistance of Germa Verheggen, Jelle Eygensteyn and Liesbeth Pierson in
the laboratory is highly appreciated. Thanks to Jikkie Boedeltje-Elzinga for her
help with data processing. We thank Dr. S. Szabó, Dr. H. Rolletschek and an anonymous referee for comments that improved the manuscript. This project was supported by the Ministry of Transport, Public Works and Water Management, Direction
Oost-Nederland, Department Twenthekanalen & IJsseldelta.
Supplementary material
The Appendix 1 (Repeated measures ANOVA of treatment and time effects and their
interaction on the water parameters measured) is available from
http//www.eco.sci.kun.nl/mibiol/Boedeltje.htm
References
Boedeltje, G., Bakker, J. P. & Ter Heerdt, G. N. J. (2003): Potential role of propagule banks in the development of aquatic vegetation in backwaters along navigation canals. – Aquat. Bot. 77: 53 – 69.
Boedeltje, G., Bakker, J. P., Ten Brinke, A., Van Groenendael, J. M. & Soesbergen, M. (2004): Dispersal phenology of hydrochorous plants in relation to discharge, seed release time and buoyancy of seeds: the flood pulse concept supported. – J. Ecol. 92: 786 – 796.
Boedeltje, G., Smolders, A. J. P., Roelofs, J. G. M. & Van Groenendael, J. M.
(2001): Constructed shallow zones along navigation canals: vegetation establishment and change in relation to habitat characteristics. – Aquat. Conserv. Mar.
Freshwat. Ecosyst. 11: 453 – 471.
Boedeltje, G., Ter Heerdt, G. N. J. & Bakker, J. P. (2002): Applying the seedlingemergence method under waterlogged conditions to detect the seed bank of
aquatic plants in submerged sediments. – Aquat. Bot. 72: 121 – 128.
Boström, B., Jansson, M. & Forsberg, C. (1982): Phosphorus release from lake sediments. – Arch. Hydrobiol. Beih. Ergebn. Limnol. 18: 5 – 59.
De Groot, W. T., De Jong, F. M. W. & Van den Berg, M. M. H. E. (1987): Population dynamics of duckweed cover in polder ditches. – Arch. Hydrobiol. 109: 601 –
618.
Dudel, G. & Kohl, J.-G. (1992): The nitrogen budget of a shallow lake (Grosser
Müggelsee, Berlin). – Internat. Rev. ges. Hydrobiol. 77: 43 – 72.
Foote, A. L. & Kadlec, J. A. (1988): Effects of wave energy on plant establishment
in shallow lacustrine wetlands. – J. Freshwat. Ecol. 4: 523 – 532.
Grasshoff, K. & Johannsen, H. (1977): A new sensitive method for the determination of ammonia in sea water. – Water Res. 2: 516.
Nutrient fluxes and floating plant dominance
361
Henriksen, A. (1965): An automated method for determining low-level concentrations
of phosphate in fresh and saline waters. – Analyst 90: 29 – 34.
Janse, J. H. & Van Puijenbroek, P. J. T. M. (1998): Effects of eutrophication in
drainage ditches. – Environ. Pol. 102: 547 – 552.
Kleeberg, A. & Schlungbaum, G. (1993): In situ phosphorus release experiments in
the Warnow River (Mecklenburg, northern Germany). – Hydrobiologia 253: 263 –
274.
Lüönd, A. (1983): Das Wachstum von Wasserlinsen (Lemnaceae) in Abhängigkeit des
Nährstoffangebots, insbesondere Phosphor und Stickstoff. – Veröff. Geobot. Inst.
ETH, Stift. Rübel, Zürich 80: 1 – 116.
Marsden, M. W. (1989): Lake restoration by reducing external phosphorus loading:
the influence of sediment phosphorus release. – Freshwat. Biol. 21: 139 – 162.
Morris, P. F. & Barker, W. G. (1977): Oxygen transport rates through mats of Lemna
minor and Wolffia spp. and oxygen tension within and below the mat. – Can. J.
Bot. 55: 1926 – 1932.
Portielje, R. & Roijackers, R. M. M. (1995): Primary succession of aquatic macrophytes in experimental ditches in relation to nutrient input. – Aquat. Bot. 50: 127 –
140.
Ridley, H. N. (1930): The dispersal of plants throughout the world. – L. Reeve & Co.,
Ashford, pp. 542 – 543.
Roijackers, R. (1981): A comparison between two methods of extracting chlorophylla from different phytoplankton samples. – Hydrobiol. Bull. 15: 179 – 183.
Roijackers, R., Szabó, S. & Scheffer, M. (2004): Experimental analysis of the
competition between algae and duckweed. – Arch. Hydrobiol. 160: 401 – 412.
Rolletschek, H. (1999): The impact of reed-protecting structures on littoral zones. –
Limnologica 29: 86 – 92.
Scheffer, M., Szabó, S., Gragnani, A., Van Nes, E. H., Rinaldi, S., Kautsky, N.,
Norberg, J., Roijackers, R. M. M. & Franken, R. J. M. (2003): Floating plant
dominance as a stable state. – Proc. Nat. Acad. Sci. USA 100: 4040 – 4045.
Sculthorpe, C. D. (1967): The biology of aquatic vascular plants. – Edward Arnold,
London, pp.456 – 483.
Sinke, A. J. C., Cornelese, A. A., Keizer, P., Van Tongeren, O. F. R. & Cappenberg, T. E. (1990): Mineralization, pore water chemistry and phosphorus release
from peaty sediments in the eutrophic Loosdrecht lakes, The Netherlands. – Freshwat. Biol. 23: 587 – 599.
Szabó, S., Braun, M. & Borics, G. (1999): Elemental flux between algae and duckweeds (Lemna gibba) during competition. – Arch. Hydrobiol. 146: 355 – 367.
Technicon (1969): Industrial Method 33 – 69 W, nitrate + nitrite in water. Technicon
Autoanalyser Methodology. – Technicon Corporation, Karrytown, New York, pp.
1 – 2.
Van Beek, G. (2000). Plankton. – In: Bak, A., Kaper, A., Reeze, A. J. G. & Van
Splunder, I. (eds): Biologische monitoring zoete rijkswateren: Watersysteemrapportage Noordzeekanaal, Amsterdam-Rijnkanaal, Kanaal Gent-Terneuzen, Twentekanalen 1997. – RIZA-rapport, Lelystad, pp. 93 – 101.
Van Luijn, F., Boers, P. C. M., Lijklema, L. & Sweerts, J.-P. R. A. (1999): Nitrogen
fluxes and processes in sandy and muddy sediments from a shallow eutrophic
lake. – Water Res. 33: 33 – 42.
362
G. Boedeltje et al.
Wetzel, R. G. (2001): Limnology, Lake and River Ecosystems. – Academic Press,
London, pp. 1 – 1006.
Zar, J. H. (1999): Biostatistical Analysis. – Prentice Hall, Upper Saddle River, pp.
1 – 663.
Submitted: 20 April 2004; accepted: 9 November 2004.