Ecology of Shallow Lakes
Ecology of Shallow Lakes
Ecology of Shallow Lakes
16. Regulation and Stabilization Paradigms in Popula- 23. Dynamics of Coral Communities
tion Ecology Ronald H. Karlson
P.J. den Boer, J. Reddingius Hardbound, ISBN 0-412-79550-7, August 1999
Hardbound, ISBN 0-412-57540-X, September 1996 Paperback, ISBN 1-4020-1046-X, October 2002
17. Biology of Rarity
William E. Kunin, Kevin J. Gaston 24. Invasive Species and Biodiversity Management
Hardbound, ISBN 0-412-63380-9, December 1996 Odd Terje Sandlund, Peter Johan Schei, Åslaug Viken
Hardbound, ISBN 0-412-84080-4, April 1999
18. Structured-Population Models in Marine, Terres- Paperback, ISBN 0-7923-6876-2, June 2001
trial, and Freshwater Systems
Shripad Tuljapurkar, Hal Caswell 25. Structure and Dynamics of Fungal Populations
Paperback, ISBN 0-412-07271-8, January 1997 James J. Worrall
Hardbound, ISBN 0-412-80430-1, March 1999
19. Resource Competition
James P. Grover 26. Competition
Hardbound, ISBN 0-412-74930-0, July 1997 Paul A. Keddy
Hardbound, ISBN 0-7923-6064-8, February 2001
20. Individual Behavior and Community Dynamics Paperback, ISBN 1-4020-0229-7, November 2001
John Fryxell, Per Lundberg
Paperback, ISBN 0-412-99411-9, October 1997
27. Chaos in Real Data
21. Spatiotemporal Models of Population and Com- Joe N. Perry, Robert H. Smith, Ian P. Woiwod, David
munity Dynamics R. Morse
Tamás Czárán Hardbound, ISBN 0-412-79690-2, May 2000
Hardbound, ISBN 0-412-57550-7, November 1997
Marten Scheffer
Department of Environmental Sciences,
Wageningen University,
The Netherlands
First edition Hardbound 1998; Reprinted Hardbound 2001; Reprinted paperback with
corrections 2004
Preface ix
Symbols xii
Introduction xiv
What is a shallow lake? xiv
Management problems xiv
Theory versus nature xvii
Set-up of the book xix
3 Phytoplankton 76
3.1 The regulation of algal biomass 76
Empirical relationships between chlorophyll and nutrients 76
The logistic equation as a model of algal growth 79
Losses due to sinking and flushing 81
Lake depth and light limitation 85
The combined effects of nutrients and light 89
Phytoplankton control by grazers 96
Allelopathic effects 97
3.2 Competition between algae and cyanobacteria 99
Empirical relationships with nutrients and turbidity 101
Hysteresis as an implication 105
Competition as an explanation 107
Other mechanisms involved 114
3.3 Multi-species competition and succession 116
Causes of non-equilibrium dynamics 117
Reasons to expect chaotic dynamics 118
S Vegetation 210
5.1 Implications of vegetation for the animal community 213
Invertebrates 214
Zooplankton 216
Fish 222
Birds 224
5.2 Effect of vegetation on turbidity 225
Correlations and causality 225
Seasonal dynamics 229
Case studies revealing mechanisms 233
Are there general patterns? 237
Brackish lakes as an exception 241
5.3 The regulation of vegetation abundance 245
Turbidity and water depth 245
Periphyton 250
Temperature 253
Substrate and nutrients 253
Wave action 254
Birds 256
viii Contents
Fish 257
5.4 Vegetation and phytoplankton dominance
as alternative equilibria 258
Stabilizing mechanisms 258
A simple graphical model 260
The vegetation-turbidity interaction elaborated 263
Hysteresis as an implication 266
A minimal mathematical model 269
Predictions from a mechanistic vegetation model 273
Implications of seasonality 279
Other mechanisms influencing the hysteresis 281
Hallmarks of hysteresis 284
A review of evidence from the field 286
References 314
Lake index 344
Subject index 346
Preface
Looking into the water of a shallow lake or pond one might have a crystal
clear view of gently waving submerged plants, fishes darting off and small
animals moving busily around. More likely, however, the water is murky,
troubled by blooming algae and suspended sediment particles hiding what-
ever is going on below. Remarkably, intermediate situations between these
two extremes seem relatively rare. This impression may result from our
persistent tendency to dichotomize in an attempt to simplify the world, but
research suggests that in this case there is some truth in the dichotomy.
The two situations represent strongly contrasting community states, both
of which have stabilizing feedback mechanisms. In the turbid state,
the development of submerged vegetation is prevented by low underwater
light levels. The unprotected sediment is frequently resuspended by wave
action and by fish searching for food causing a further decrease of transpar-
ency. Since there are no plants that could serve as refuges, zooplankton is
grazed down by fish to densities insufficient to control algal blooms. In
contrast, the clear state in eutrophic shallow lakes is dominated by aquatic
macrophytes. The weed beds prevent sediment resuspension, take up nutri-
ents from the water, and provide a refuge for zooplankton against fish
predation.
In view of these feedback mechanisms it is not surprising that shallow
lakes refuse to obey simple rules such as the classical relationships between
algal biomass and nutrient loading. The response of shallow lakes to
eutrophication is often catastrophic rather than smooth, and several lakes
are known to have switched back and forth between a clear and a turbid
state repeatedly without obvious external forcing. In view of the relative
neglect of shallow lakes in limnologicalliterature, this disparate behaviour
has apparently discouraged rather than tempted researchers for a long time.
Even in countries where almost all the lakes are shallow, like Denmark and
The Netherlands, limnological research has traditionally been focused en-
tirely on the few deep lakes that were available. Over the last decade,
however, this situation has changed.
Many of the shallow lakes have become turbid during this century be-
cause of eutrophication, and efforts to restore the clear state by means of
reduction of the nutrient loading are often unsuccessful. This has invoked
experiments with additional methods such as temporary reduction of the
fish stock. The potential impact of such disturbances on shallow Jake ecosys-
tems appears to be huge, and several lakes have switched to a stable clear
state in response. The careful monitoring of these changes has catalysed the
x Preface
development of insights into the mechanisms that govern the dynamics of
shallow lake communities.
Writing this book has been an (unfinished) journey towards the goal of
constructing a coherent picture of 'how shallow lakes work'. Unravelling the
origin of main ideas is probably as difficult as unravelling the functioning of
lake ecosystems, but some things can be said about the intellectual roots of
the views presented in this book. The feeling that there was something like
alternative stable states in our lakes, and ideas about the roles of fish and
vegetation have been fermenting for at least a decade in a group of English,
Danish and Dutch Iimnologists. Surely, Brian Moss and Erik Jeppesen were
very important inspiring thinkers in this group. At our institute Andre
Breukelaar, Harry Hosper and Marie-Louise Meijer were my partners in
discussing these exciting views.
Erik Jeppesen was the one who after reading the first draft encouraged
me, explaining that it was going to be great, but pointing out kindly that I
needed to do quite a bit of extra homework. The book ripened during the
subsequent phase when I tried to understand the often conflicting results
from countless descriptive and experimental studies, a process that defi-
nitely helped the theoretician in me to become more modest. Sergio Rinaldi
was my guide in the fascinating and dazzling world of dynamical systems
theory with its attractors, bifurcations and other abstract structures. To-
gether we looked for parallels to the even more dazzling world of aquatic
food webs. Rob de Boer and Yuri Kuznetsov developed wonderful software
for analysing models and guided me in using it.
Don DeAngelis, who suggested me to write this book, invited me over at
the Oak Ridge National Laboratory to work, and gave many useful advices
and comments. Eric Marteijn and the rest of the staff at RIZA have been
very supportive and flexible, allowing excursions into research topics that go
beyond a diJ;.ectly applied context, and giving me the time to write this book.
It was a great pleasure to write a large portion of it in the beautiful
Tennessean country house of Bob and Dorothy Jolley.
With Adriaan Achterberg I shared an unforgettable time doing our first
work in aquatic ecology trying to figure out what happens in ditches. He
made a number of pen drawings especially for this book (Figs. 4.3, 4.42, 1.4,
4.49, 3.13, 5.1, 2.15, 4.46). The cover illustrations are made by Ad Swier. Bert
Jansen prepared most of the technical drawings and kindly tolerated my
numerous changes of mind. Rita van Leeuwen did an invaluable job local-
izing relevant literature. Egbert van Nes has been a crucial help all the way,
solving problems with software packages, writing better ones, helping or-
ganizing and finding literature, discussing countless important and unimpor-
tant ideas during our lunch-walks and quietly pointing out crucial problems
in seemingly brilliant reasonings.
The complete draft ofthe book has been read by Irmgand Blindow, Steve
Carpenter, Hugo Coops, Milena Holmgren, Harry Hosper, Mark Hoyer,
Bas Ibelings, Erik Jeppesen, Eddy Lammens, Marie-Louise Meijer, Stuart
Preface xi
Mitchell, Brian Moss, Egbert van Nes, Ruurd Noordhuis, Marcel van der
Berg and Diederik van der Molen, while portions were reviewed by Herman
Gons and Henddk Buiteveld. Without their extensive comments the book
would have been much harder to read.
Symbols
The main symbols used in the book. Figures are produced using the listed
default values unless indicated otherwise. Symbols that used only once, are
explained in the text and can not be found in this list.
MANAGEMENT PROBLEMS
The pristine state of the majority of shallow lakes is probably one of clear
water and a rich aquatic vegetation. Nutrient loading has changed this
situation in many cases. The lakes have shifted from clear to turbid, and with
the increase in turbidity, submerged plants have largely disappeared. The
sequence of changes during eutrophication is rarely documented well, but
some elements are agreed upon by most workers in the field (Moss, 1988).
Introduction xv
Shallow lakes with a low nutrient content usually have a vegetation domi-
nated by relatively small plants. With increased nutrient loading the biomass
of aquatic macrophytes increases and plants that fill the entire water column
or concentrate much of their biomass in the upper water layer become
dominant. Such dense weedbeds are often experienced as a nuisance by the
fishing and boating public. When weed control programmes eradicate the
vegetation, turbidity in shallow lakes tends to increase strongly due to algal
blooms and wind resuspension of the sediment. Also, when vegetation is not
controlled explicitly, further eutrophication of vegetated lakes can lead to a
gradual increase of phytoplankton biomass and of the periphyton layer that
covers the plants. Shading by these organisms ultimately leads to a collapse
of the vegetation due to light limitation.
Restoration of non-vegetated turbid shallow lakes to the clear vegetated
state is notoriously difficult. Reduction of the nutrient loading may have
little effect, as during the period of eutrophication a large amount of phos-
phorus has often been adsorbed by the sediment. When the loading is
reduced and its concentration in the water drops, phosphorus release from
the sediment becomes an important nutrient source for phytoplankton.
Thus a reduction of the external loading is often compensated by 'internal
loading', delaying the response of the lake water concentration to the reduc-
tion of external loading.
However, internal loading is not the only reason why restoration of
turbid shallow lakes is difficult. With the disappearance of aquatic veg-
etation the structure of the shallow lake community changes dramatically
(Fig. 1).
Invertebrates that are associated with vegetation disappear and with
these animals the birds and fishes that feed on them or on the plants.
Also, vegetation provides an important refuge against predation for
many animals, and hence its disappearance causes crucial shifts in many
predator-prey relationships. Large zooplankton use vegetation as a day-
time refuge against fish predation. In vegetated lakes they can contribute
significantly to the control of phytoplankton biomass. In the absence of
vegetation their numbers are strongly reduced. This, and the increased
nutrient availability, allows phytoplankton biomass to be higher in the
absence of vegetation. In addition, wave resuspension of the unprotected
sediment can' cause a considerable additional turbidity once the vegetation
has vanished. The fish community of unvegetated lakes becomes dominated
by species that forage on benthic invertebrates. Their activity promotes
the nutrient flux from the sediment into the water and causes an extra
resuspension of sediment particles, contributing to the already high
turbidity.
Return of submerged plants in this situation is unlikely, in part because
their absence has allowed a further increase in turbidity, but also because
the frequent disturbance of sediment by wind and benthivorous fish ham-
pers resettlement. Ecological feedback mechanisms are thus an important
xvi Introduction
reason why restoration of the vegetated clear water state is difficult. In many
cases, nutrient reduction alone may be insufficient to restore the clear state
in shallow lakes. Additional measures, however, such as removal of part of
the fish stock and changes in the water level, have been successfully used as
a way to break the feedback that keeps such lakes turbid.
This chapter tells the history of a number of shallow lakes in which conspicu-
ous changes in community structure have been observed (Table 1.1).
The lakes range in area from 1 to 18200 hectares and in geographic
position from New Zealand to Sweden. Although some of the lakes have
been studied in detail for several years, the information remains fragmen-
tary in most cases. Nonetheless, when viewed in combination, clear patterns
emerge. Together, these histories give an idea of the large and often rapid
shifts that can occur in shallow lake ecosystems, but also of the internal
mechanisms and external forces involved in causing such changes. While the
next chapters zoom in on specific regulatory mechanisms, the series of short
stories in this chapter sets the stage by depicting the main phenomena
that the work described in the rest of the book is ultimately aimed at
understanding.
Chlorophyll-a
75
25
60
year
Fig. 1.2 Changes in the mean summer concentrations of chlorophyll-a and total-P
and in the biomass of submerged plants in Alderfen Broad, UK, following discon-
nection of the lake from a nutrient rich inflow stream. Redrawn from Perrow et al.
(1994).
1950s, while the decline with eutrophication in the late 1960s is the mirror
image of these changes (Fig. 1.1).
(Brouwer and Tinbergen, 1939; Leentvaar, 1961; Leentvaar, 1966;
Hosper, 1985; Scheffer et al., 1992; Scheffer et al., 1994a; Scheffer et al.,
1994b; Noordhuis, 1997; Van den Berget al., 1997)
Alderfen Broad
Alderfen Broad is one of the fifty or so small lakes in the peaty Norfolk
Broadland (UK). These so-called Broads are connected by rivers that also
serve to transport the sewage of about 400000 people to the nearby North
Storm effects 5
Sea. This and the draining of farmland has caused a strong eutrophication
of the Broads. Once low in both phosphorus and nitrogen compounds,
the lakes now have 10 times their former concentration of these nutrients.
The water that according to the locals used to be 'gin-clear' is now turbid
and the once famous fields of submerged plants are largely gone.
Alderfen Broad is surrounded by a wetland dominated by alder (Alnus
glutinosa). It had clear water and a dense submerged vegetation (mainly
Ceratophyllum demersum) un~l at least the late 1960s, but in the next
decade it lost its aquatic plants completely and became turbid due to high
algal biomass. This was linked with discharge of sewage effluent to the
inflow stream. In 1979, the inflow stream was diverted into existing ditches
from where the water rejoined the outflow stream below the lake. This
isolation led to a sequence of conspicuous changes in the lake over the next
12 years (Fig. 1.2).
The total phosphorus concentration as well as the release of phosphorus
from the sediments declined considerably over the first three years, and this
led to a substantial decrease of algal biomass. During 1982 and 1983 a dense
vegetation of Ceratophyllum was found and the phytoplankton populations
remained low. In 1984, however, vegetation was less abundant and the
phosphorus level in the lake water had become high again. Very few plants
were found over the next three years. Some blooms of blue-green algae
(Anabaena spiroides) developed in 1985 and 1986, but on average
phytoplankton biomass remained relatively low during these years, prob-
ably because of nitrogen limitation. From 1987 onwards vegetation ex-
panded again and algal biomass dropped to lower levels. With the increase
in vegetation, phosphorus concentrations also rose again to reach a peak
value in 1991 when vegetation biomass was once again declining.
The observed pattern (Fig. 1.2) suggests a cycle with a period of about 8
years. It is not clear what drives the decline of the vegetation. It seems
unlikely that turbidity prevented plant growth in the 1985-1987 period.
Conspicuous absence of plants has been observed in several other clear
Broads, suggesting that another factor is at work.
(Moss et al., 1986; Moss et al., 1990; Perrow et al., 1994.)
Lake Apopka
The large and shallow Lake Apopka (128km2; mean depth 1.65m) is situ-
ated in Florida. The lake was well known as an outstanding sport fishing lake
with exceptionally clear water and dense beds of aquatic plants until the
autumn of 1947 when a hurricane wiped out the vegetation. Soon after this
catastrophic event the first phytoplankton bloom was observed. Over the
following 4 years the total fish stock was estimated to have increased 10-
fold, and game fish populations had greatly expanded. By 1957, however,
6 The story of some shallow lakes
the omnivorous shad had become the dominant fish species in the lake
and the abundance of game fish had strongly decreased. Rotenone applica-
tions killed an estimated total of 9 million kg of shad during the late 1950s
in an unsuccessful attempt to restore the game fish dominance. Today the
lake is still turbid and vegetation has not recovered since the hurricane of
1947. A thick layer (1.5m) of unstable sediment is involved in frequent
resuspension, and turbidity of the water increases strongly during windy
periods.
(Schelske and Brezonik, 1992; Schelske et al., 1995; Schelske et al., 1997.)
Lake EUesmere
The story of the New Zealand Lake Ellesmere is comparable to that of Lake
Apopka. During the 1950s and 1960s the lake was known to harbour a
population of up to 80000 black swans (Chenopsis atrata} living on the
abundant aquatic vegetation. In 1968, however, a violent storm killed 5000
swans. Probably more importantly, the storm annihilated the extensive
weedbeds in this huge unprotected lake. Vegetation has not recovered since
then, and waves stir up the sediments most of the time. Together with high
concentrations of phytoplankton this suspended sediment causes the lake to
be very turbid. By 1986 only 4000 swans resided in the lake, a mere 5% of
the original population.
(Mitchell et al., 1988; McKinnon and Mitchell, 1994; Hamilton and
Mitchell, 1996.)
Tiimnaren
The large (35km2} and shallow (now 1m mean depth} Lake Tiimnaren is
situated 120 km northwest of Stockholm in a flat part of Sweden dominated
by forests and farmlands. It used to be common in these areas to reduce the
lake levels to minimize flooding of farm areas. Farming also increased the
nutrient loading of the originally oligotrophic lakes, and caused the sparse
plant stands to be replaced by a lush aquatic vegetation in many cases.
The water level of Lake Tamnaren was intentionally lowered twice. The
first time, in 1870, by 1m, and the second time, in 1950-54, by another 0.5 m.
After the last draw-down, aquatic macrophytes expanded strongly, and the
lake became famous for its waterfowl. In the spring and autumn many
migratory birds flocked in great numbers in the lake and in summer about
500 mute swans (Cygnus olor) used to live there. The water was clear
enough to see the bottom anywhere in the lake. The lush vegetation,
mapped in 1973, was dominated by Elodea canadensis and by floating leafed
species (Nuphar lutea and Potamogeton natans).
VVaterlevel changes 7
In the spring of 1977 the water level rose 0.3 m leading to dramatic
changes in the community. The swans as well as many other birds disap-
peared from the lake and the submerged and floating leafed vegetation as
well as a large part of the emergent vegetation vanished. Aerial photographs
taken in 1973 and 1983 revealed that the total vegetation coverage, includ-
ing reed, was reduced from 80% to 14% of the lake area. The water had
become turbid with a Secchi-depth of only 0.6m. This turbidity was partly
due to an increased phytoplanKton density (up to 70 !lgr'), but resuspended
sediments contributed strongly as well.
(Wallsten and Forsgren, 1989; Bengtsson and Hellstrom, 1992.)
Rice Lake
Rice Lake is a medium-sized shallow lake in Wisconsin (USA), named for
the extensive beds of wild rice ( Zizania) that used to grow along the shores
together with cattails (Typha) and bulrushes (Scirpus). Pollen analysis
shows that quillwort (Isoetes) used to grow in the lake, a species that re-
quires very clear water. Aerial photographs since 1938 and testimony from
local residents suggest that the lake stayed clear until the 1970s when high
water flooded the wild rice. The high water, promoted by a newly con-
structed beaver dam in the outlet creek allowed summer winds to sweep
across the lake unabated by the flooded vegetation. Waves eroded the
bottom, tearing away chunks of marsh that subsequently disintegrated,
producing a soft sediment The erosion of the marsh vegetation was further
promoted by a peculiar interplay of freezing and water level dynamics.
Water levels rose in March when the lake was still frozen. This lifted the
ice cover within the marsh border, pulling up mats of rhizomes and roots.
Wind and wave action in early spring pulled out the frozen mats to form
floating islands. Many islands drifted downstream and grounded in the
outlet creek. This caused the water levels to rise even higher, making the
situation worse.
A decade later the water levels fell, but the wild rice never recovered and
the lake has stayed turbid. Secchi-depth in summer is now only 0.3m on
average. This is only partly caused by phytoplankton, reaching a maximum
of only 40 !lg Chi r' during blooms of green algae. Probably, resuspension of
sediment by waves and fish is a more important source of turbidity. The
sediment of Rice Lake is now so loose and flocculent that a steel pipe shoved
into the lake bottom hits firm bottom only after penetrating 6m of this soft
sediment. The role of wave resuspension is illustrated by the large differ-
ence in suspended solid concentration during ice cover (<2mgr') and dur-
ing ice free periods (59mgr'). The present submerged vegetation is very
sparse, 90% of which is sago pondweed (Potamogeton pectinatus). Cattail
dominates the shore vegetation, and some stands of water lilies and floating
leaf pondweed are found along the shores.
(Engel and Nichols, 1994.)
8 The story of some shallow lakes
1.4 FISH STOCK MANAGEMENT
Zwemlust
Zwemlust is a very small shallow lake in the centre of The Netherlands. In
summer it is used as a swimming pool. The lake receives water through
seepage from the polluted River Vecht running closely along the lake. As a
result the nutrient loading of the lake is high. The chlorophyll-a concentra-
tion in the lake was as high as 250 mg t' and blooms of blue-green algae
(Microcystis) frequently turned the lake bright green. After several unsuc-
cessful attempts to improve the water quality it was decided to manipulate
the fish stock. This measure had spectacular results (Fig. 1.3).
In March 1987 the water was pumped out of the lake to facilitate com-
plete removal of the fish. It appeared that lOOOkgha·' of fish had been
present, 75% of which was bream. Seepage refilled the lake in three days,
and a small fish stock of pike and rudd (Scardinius erythrophtalmus) was
introduced together with water fleas and some charophytes and yellow
waterlilies. Shortly after the refill there was an algal bloom, but soon large
water fleas became abundant and grazed down algal biomass to a mere 2%
of the pre-manipulation values and the water became crystal clear. Some
filamentous green algae (Hydrodictyon; Enteromorpha) developed in the
shallower parts, but only a small part of the lake bottom became covered by
macrophytes in the first summer.
200,--------------------------------------------.
<>--<> Chlorophyll-a (!lg J·1)
•·-·---• Macrophytes (g DW m·2) 240
r r
250
150
100
50
Biomanipulation
Linford lakes
The Great Linford sand and gravel-pit complex covers an area of about
300ha in the flood plain of a river near Newport Pagnell (UK). The site
10 The story of some shallow lakes
Fig. 1.4 Coot (Fulica atra) are largely herbivorous. In the breeding season birds are
territorial and population densities are usually not very higb. However, in the
autumn and winter large numbers of animals may concentrate in lakes and reduce
the biomass of aquatic vegetation.
encompasses 14 lakes which have been excavated over the past four dec-
ades. Two distinct ways of gravel extraction were used in this area: wet-
digging and dry-digging. When a new quarry is opened, groundwater enters.
The wet-digging method proceeds by sucking the gravel from the bottom
Fish stock management 11
with a suction dredger and depositing it in a floating barge. The gravel
rapidly settles out and the silt laden water flows directly back into the newly
forming lake. There the fine silt forms a thick layer of loose sediment.
During the process of dry-digging the entering groundwater is continuously
pumped out of the quarry. The excavated material is transported to a
washing and grading plant where the washings are run off into special silt
settlement lagoons. When pumping ceases the dry-dug lakes fill up with
water.
These different digging methods have resulted in two different lake
types in the area. Wet-dug lakes have typically remained turbid, even
though most of them have been left undisturbed for over 20 years. They
have very little submerged vegetation and the sediment is flocculent
and easily resuspended by waves. During storms the lakes tum chocolate
brown, with suspended solid concentrations of up to 0.2gt1• The dry-dug
lakes, on the other hand, are clear and densely vegetated. They also have
many more birds than the wet-dug lakes. This led workers of the Game
Conservancy research station to try to determine why the turbid lakes are a
poor habitat for ducks and other waterfowl, and find out ways of changing
this.
One of the wet-dug lakes that has been studied in some detail is Main
Lake. The lake was very turbid during a survey in 1982 and less than 1% of
its area was found to be vegetated. Sago pondweed was the only recorded
species. In the winter of 1987-88 the lake was pumped down and almost the
entire fish stock was removed with seine nets. It appeared that about
356 kgha-' of fish, mainly bream and roach (Rutilis rutilis ), had been present.
The following years vegetation expanded spectacularly (Fig. 1.5) to reach a
coverage of 93% in 1989 when Elodea canadensis had become the dominant
species.
At the same time, the density of midge larvae and snails increased mark-
edly (Fig. 1.5). These changes were followed by a sharp increase in the
number of overwintering coots, ducks and swans. Also, the nesting success
of ducks increased, as early survival of ducklings improved, presumably
because of a better food situation. Transparency of the lake water increased
strongly after the fish removal. This was in part due to an increase in
large water fleas (Daphnia spp.) that graze on phytoplankton, but after the
development of vegetation the resuspension of sediment by waves has also
decreased.
In 1990 a bay was isolated from the rest of the lake by nets and stocked
with the original fish community. This led to a reversal of the changes
observed after the fish removal. Vegetation development was strongly sup-
pressed again in the bay and the densities of midge larvae and snails
dropped to their original values (Fig. 1.5). In the main lake, however, the
clear and vegetated state has persisted (Traill-Stevenson pers. comm. ).
(Giles, 1987; Hill et al., 1987; Wright and Shapiro, 1990; Giles, 1992.)
12 The story of some shallow lakes
fish
normal
200
Aquatic plants
~
i
. 150
~ 100
~
E
50
~
12
.,E
..
i
10
2:- 6
'3
~ 4
E
i 2
20
~
i
. 15
~ 10
'3
~
E 5
.Q
"'
1986 1987 1988 1989 1990 1991
Fig. 1.5 Effect of a removal of fish and subsequent restocking on the biomass of
aquatic vegetation, midge larvae and snails in Main Lake, Great Linford, UK.
Redrawn from Giles (1992).
Miscellaneous cases 13
1.5 MISCELLANEOUS CASES
Lakes THkem and Krankesjiin
The southern Swedish lakes Takern and Krankesjtin have changed repeat-
edly from a clear state with abundant vegetation to a turbid state with few
submerged plants and vice versa over the last century (Fig. 1.6).
Early data on the vegetation are sparse, but waterfowl that are associated
with plants have been continuously recorded, and these data were used as an
indirect indicator of vegetation abundance. In neither of the lakes are
changes in the external nutrient loading thought to have occurred. Instead,
variations in the water level seem to be involved in causing the switches
from one state to the other.
Lake Takern was largely covered by a dense vegetation dominated by
charophytes at the beginning of the century. In 1914 after a dry period, the
vegetation disappeared from the lake, but rapidly recovered again. In the
early 1930s, large parts of the lake dried out. This desiccation and subse-
quent freezing of the bottom in winter are thought to have caused the
complete disappearance of submerged plants during these years. Within a
few years, however, a dense vegetation had returned. In the early 1950s sub-
merged plants disappeared completely again and the water became turbid.
This time there was no obvious cause of the changes. At the beginning of the
1960s the vegetation recovered starting with stands of angiosperms and
some charophytes. In 1969 the charophytes had expanded into dense mats
covering large parts of the lake again. The lake has stayed clear and veg-
etated until1995 when summer transparency decreased and the condition of
the vegetation deteriorated. The next year this downward trend continued,
leaving the lake in a turbid and poorly vegetated state once again.
Like Lake Takern, Lake Krankesjtin was covered by a charophyte domi-
nated vegetation at the beginning of the century. In the 1940s, vegetation
disappeared for some years, presumably due to low water levels in winter
that allowed the bottom to freeze, eliminating the plants. However, vegeta-
tion recovered soon, and the lake stayed clear and densely vegetated till the
early 1970s. Inspection in 1975 revealed that submerged macrophytes had
disappeared entirely. Exceptionally high water levels during the growing
-
--1••-••----------------
~----------------~
1900
~~Lake Krankesjon
2000
lake Ukern
• Turbid
Macrophyte dominated
Fig. 1.6 Repeated shifts between a turbid state (thick line) and a clear vegetation
dominated state (thin line) in the Swedish lakes Krankesjon and Tilkern. The win-
dow marks the period represented in Fig. 1.7. Modified from Blindow et al (1993).
14 The story of some shallow lakes
40 60
50
30
5..... 40
~
!;
;::> :;
f 20 30 u
"
-e .Q
-:;;
"
..... 20
10 ~
10
0
1983 1984 1985 1986 1987 1988 1989 1990
Fig. 1.7 Shift from a turbid state to a vegetation dominated clear situation in Lake
Krankesj6n. Redrawn from Hargeby eta/. (1994).
season may have shaded out the vegetation, although the absence of ice
cover during mild winters resulting in strong wave erosion is another pos-
sible explanation.
From 1983 onward the lake has been monitored more closely revealing
the scenario of a subsequent vegetation recovery episode in some detail. It
is unclear what initiated the recovery, but low summer water level allowing
more light to penetrate to the bottom, and a disease causing a reduction of
the bream population are mentioned as possible causes. Despite the fact
that the sediment contained very high densities of oospores of charophytes
the first macrophyte to expand was sago pondweed (Potamogeton
pectinatus) (Fig. 1.7).
Only after some years, the relatively sparse stands of pondweed were
replaced by exponentially increasing dense mats of charophytes, leaving
angiosperms to play a minor role. Turbidity decreased only slightly during
the expansion of pondweed, but subsequently dropped almost an order of
magnitude during the expansion of charophytes. A decrease was also ob-
served in the chlorophyll-a and total phosphorus levels and in the density of
large herbivorous zooplankton. Conspicuous changes occurred in the bird
community. During the turbid years, only some fish-eating birds resided in
the lake. With the increase of vegetation biomass, however, the number of
coots and swans (Fig. 1.8) and the populations of dabbling ducks increased
steeply.
(Blindow, 1992b; Blindow et al., 1993; Hargeby et al., 1994.)
30 300
c::
"'
;:
"'
2
:s
20 200 8
u
::E
10 100
0 0
1985 1986 1987 1988 1989 1990
1964 a dam was constructed turning the swamp into a shallow (2.5m mean
depth) reservoir, intended to create a storage of irrigation water, and a place
for trout fishery. The majority of the original macrophytes died as a result of
the flooding, but other submersed vegetation colonized the lake bottom.
Over the following two decades, trout fishery flourished, but due to its small
storage capacity the lake was found to be an unreliable source of irrigation
water. In 1984, a canal was constructed to divert water into the storage from
a nearby creek. Subsequently, high water levels were maintained at the
lagoon for several years. In the summer of 1987-88 the water quality de-
clined and so did the condition of the trout. The problems which persisted
during the next year were thought to be due to a suppression of macrophyte
growth caused by the increased water levels and elevated nutrient loading.
These factors caused a switch from a diverse macrophyte dominated com-
munity, to a simpler one dominated by phytoplankton, and apparently
unfavourable for the production of trout. It was therefore decided to switch
to a management regime of lower water levels. This has led to a recovery of
the macrophytes, water quality and trout fisheries.
(Sanger, 1992; Sanger, 1994.)
Lake Christina
Lake Christina is a large and shallow prairie lake in Minnesota (USA). The
lake freezes in winter, and ice is out usually by the beginning of April. In the
autumn, it is heavily used by migrating waterbirds. The lake was one of
the most important feeding and staging areas for migrating diving ducks in
the so-called Mississippi flyway during the first half of the century. The lake
supported lush beds of aquatic plants in this period, and the water was clear.
By 1959, however, transparency had rather suddenly declined to <25cm,
16 The story of some shallow lakes
100.------------------------------------.
I:- I
~ ~ ~
Caov.,back
~ ~ ~ IE LesserScaup
c:::J Other ducks and coot
80
§ 60
3
.l!l
"
8"
~ 40
....
"'
&
20
~~n~~oo~~~M~~~oo~~
year
Fig. 1.9 Trends in peak autumn waterfowl counts in Lake Christina, Minnesota,
USA. The period of low waterfowl numbers from 1978 till 1987 corresponds to a
period with little vegetation and turbid water. Redrawn from Hanson and Butler
(1994b).
vegetation had become very sparse and bird numbers had dropped about 2
orders of magnitude. These changes were probably associated with higher
water levels and much larger fish populations. The latter may result in part
from the high water, as winter-kills of fish during periods of ice-cover are
Jess likely in deeper water. In 1965, state biologists eliminated part ofthe fish
using toxaphene, and water clarity, macrophytes and duck numbers in-
creased. In the mid 1970s, however, water transparency and vegetation
abundance declined again. Peak autumn counts of waterfowl dropped spec-
tacularly from more than 130000 in 1977 to Jess than 5000 in the following
years (Fig. 1.9).
In the autumn of 1987 all fish in the lake were killed by spraying 627001
of rotenone from the air. Most of the original species reinvaded the lake
relatively quickly, but predatory fish (largemouth bass and walleye) was
stocked regularly in an attempt to slow down the recovery of the fish
biomass. During May and June of the following year transparency of the
water increased to about 90cm in contrast with the 30-40cm in the years
before the fish manipulations (Fig. 1.10).
Miscellaneous cases 17
100
80
~ 60
8c
§ 40
!j
0
20
1990
Fig. 1.11 Recovery of the aquatic vegetation between 1980 and 1990 in Lake
Christina, Minnesota, USA. Redrawn from Hanson and Butler (1994a).
This spring clear water phase was caused by an increase of large bodied
herbivorous water fleas (Daphnia) that were estimated to have a filtration
potential of 100-200% of the lake volume per day, as opposed to <10%
during the years before the fish kill. In summer Daphnia populations de-
creased, filtration rates dropped and turbidity increased. Nonetheless, veg-
etation expanded that year and became more diverse (Fig. 1.11), including
the originally common sago pondweed and Ruppia maritima, but also large
stands of charophytes, Myriophyllum exalbescens and Najas fiexis.
The same pattern of plankton dynamics occurred in the second year but,
unlike before, the water became clear again in September. Remarkably, it
18 The story of some shallow lakes
-----~~----
450 120
400 macrophytes
100
phytoplankton
~
1\
350
.\
~E
~
.
rt
u 300 80
./,
.!'9
J:
250 -r
\
e 60
-6'
... ,.: \
te
] 200
t
Q.
40
150 ~
:;:
f 100
20
50
stayed clear throughout the next year. With the return of the vegetation,
waterfowl numbers recovered as well (Fig. 1.9). The following years, the
lake stayed vegetated and rich in waterfowl. Since 1995 transparency has
gradually declined again although vegetation has remained quite abundant.
(Hanson and Butler, 1990; Hanson et a/.,1990; Hanson and Butler, 1994a;
Hanson and Butler, 1994b, Hanson pers.comm.).
Tomahawk Lagoon
Tomahawk Lagoon is a small and very shallow lake on South Island, New
Zealand. The lake is fed by a stream draining mainly pastures. Due to the
oceanic climate the lake rarely freezes for more than a few hours, and
summer temperatures are usually below 20°C. During 1963 when studies
started and the following year, the lake was turbid with high phytoplankton
densities and few macrophytes. In 1965, however, it became clear with
abundant macrophyte growth, until 1970 when it became turbid again for
two years, switching back to a clear and vegetated state in 1972. During
periods when vegetation dominated the lake, phytoplankton density was as
much as two orders of magnitude lower than in the years when vegetation
was sparse. In years with abundant vegetation large numbers of black swans
foraged on the plants in the lake. This herbivory contributed to the decline
Miscellaneous cases 19
of vegetation in some periods, but the more dramatic vegetation declines
must have been due largely to other factors, storms being a possible
candidate.
From 1969 to 1973 the seasonal dynamics of plankton and macrophytes
were studied more closely (Fig. 1.12).
These studies revealed the details of a switch from a turbid to a vegeta-
tion dominated state. In the spring, both macrophyte biomass and
phytoplankton productivity usually increase. During the first two years of
the study the macrophyte development was short lived and a dense
phytoplankton bloom developed over the summer. In tbe third year, how-
ever, the spring bloom of algae collapsed and vegetation became abundant
for tbe rest of the summer. In the autumn of that year charophytes started
to become dominant. Vegetation biomass remained high over the winter
and expanded further in the subsequent year, when no phytoplankton peak
occurred. Vegetation was thought to suppress the phytoplankton develop-
ment. No single clear mechanism could be revealed, but during some peri-
ods nitrogen-limitation of the algae could be demonstrated, while at other
times tbe grazing pressure by zooplankton living between the plants was
shown to be important.
It is not quite clear what causes the switches from phytoplankton domi-
nance and vice versa. Herbivory by black swans and storms are thought to
be involved in vegetation decline, while collapse of phytoplankton blooms
may be related to zooplankton grazing and to pulses of highly turbid inflow
during rainy periods.
(Mitchell et at., 1988; Mitchell, 1989; McKinnon and Mitchell, 1994.)
(Fig. 2.1).
Light under water 21
lrradiance (%)
10 30 50 100
g -------------------------------------------------- Secchi·depth
t
Cl
2 red (630 nm)
Fig. 2.1 Attenuation of irradiance with depth in each of three spectral blocks in
Crose Mere, UK. Irradiance in each spectral block is expressed as percentage of the
irradiance in the corresponding spectral block just beneath the water surface. Re-
drawn from Reynolds (1984).
22 The abiotic environment
called the Photosynthetic Active Radiation (PAR). The vertical attenuation
coefficient is often denoted as Kd. However, since K is reserved in this book
for the carrying capacity in logistic growth equation, E is used throughout
the text to prevent confusion. A problem with the use of E for characterizing
light attenuation is that its value is different for light of different colours
(Fig. 2.1). Green light usually penetrates the water column deeper than
other colours that can be used for photosynthesis. As a result of this differ-
entiallight attenuation, the attenuation coefficient measured over the PAR
spectrum as a whole is not constant over depth. The colours that are ab-
sorbed the most attenuate first, and, consequently, the remaining light pen-
etrates the water better. As a result, E diminishes with depth. Fortunately,
this effect appears to be relatively small in turbid waters (Kirk, 1994).
Therefore, the vertical attenuation coefficient (E) of PAR is, as Kirk states
(Kirk, 1986): 'the best single parameter in terms of which to compare the
light-attenuating properties of one water-body with another'.lt follows that
E can be roughly characterized from a simultaneous measurement of the
irradiance at depth z and just under the water surface:
In b._
E=____!_z_ (2)
z
The attenuation of radiation with depth depends on scattering as well as
absorption, but is not simply the sum of the two (called the beam attenua-
tion, c). Scattering does not remove light as absorption does; it merely
changes its direction. Because of this, scattering increases the average path-
length travelled by an incoming photon to reach a given depth, and there-
fore the chance of being absorbed. In addition, a small proportion is
scattered in a backward direction, and the fraction of this back-scattered
light that is not reflected downward again by the surface leaves the water.
For monochromatic light there is a simple empirical relationship between
E and the coefficients of absorption (a) and scattering (b):
(3)
where J1<J is the cosine of the angle of the underwater light to the vertical
(Kirk, 1994). To obtain an impression of what this implies for an average
condition in the temperate climate zone we can substitute the value of 0.8
for Jl<J, giving:
Secchi-depth
The simplest way to characterize the optical properties of Jake water is by
means of a so-called Secchi-disc. This approach, systematically studied more
than a century ago by the Italian physicist Angelo Secchi, is still widely used.
A black-and-white disc is lowered into the water, and the depth at which it
just disappears from view is noted as the Secchi-depth. A problem of the use
of Secchi-depth is that in clear shallow lakes, the bottom can be visible
throughout the Jake. In that case Secchi-depth is not measurable. However,
the method is simple and relatively robust, and, not surprisingly, data on the
Secchi-depth of waters are abundant.
Unfortunately Secchi-depth (Sd) is not a very good indicator of light
penetration into the water. On the basis of measurements in marine waters,
Poole and Atkins (1929) noted an inverse relationship between E and S":
(5)
However, the Poole Atkins coefficient, c., appears to vary strongly (roughly
around 2) from case to case. Later workers showed that the relationship
between Secchi-depth and light attenuation can be more accurately de-
scribed if the beam attenuation coefficient (c = a + b) is taken into account
(Tyler, 1968):
s __9_ (6)
"-c+E
In view of this empirical relationship, the simple inverse proportionality
proposed by Poole and Atkins would hold only if c varies proportionally to
E in the data set, which is of course unlikely (see Eq. 3). Basically, the
reason why the simple Pool-Atkins relationship does not work well, is that
scattering has a stronger effect on (inverse) Secchi-depth than on the verti-
callight attenuation (Fig. 2.2).
Therefore, two waters can have the same Secchi-depth, but differ mark-
edly in light attenuation if the relative importance of scattering differs. For
instance, a lake in which turbidity is mainly caused by suspended clay
particles (which scatter rather than absorb), will have a lower light at-
tenuation than a lake with the same Secchi-depth in which the turbidity is
mainly due to phytoplankton.
24 The abiotic environment
Fig. 2.2 Vertical attenuation coefficient (E) and the inverse Secchi-depth (liS,),
both plotted as a function of the scattering coefficient (b) and the absorption coeffi-
cient (a) of lake water. Note that Secchi-depth is strongly affected by scattering,
whereas the effect of scattering on vertical light attenuation is minor.
14
X
• Veluwemeer
"
Wolderwijd
12
..
0 Markenneer
"
IJsselmeer
E X Drontermeer
~ 10 /}
D Reeuwijk
0
•• "'
1111>< Haringvliet
~ X
•
- --
Volkerakmeer
~ /} Veluwemeer
------- Wolderwijd
~ ()
--Markerrneer
Ei - • - • Drontermeer
'E 4 --Reeuwijk
~ - ••- Haringvliet
-
-- Volkerakmeer
- Usselmeer
10
inverse Secchi-depth (m-1)
Fig. 2.3 Poor relationship between inverse Secchi-depth (liS,) and the vertical light
attenuation (E) illustrated by data sets from eight Dutch lakes. The deviating posi-
tion of Markermeer is explained by the high concentration of suspended clay parti-
cles in this lake. These particles contribute relatively much to scattering of the light
which affects Secchi-depth stronger than it affects vertical light attenuation.
Nephelometric turbidity
In this book the word turbidity is used in a loose sense to indicate the lack
of clarity of lake water. There are, however, also laboratory devices that
measure so-called nephelometric turbidity. In such turbidimeters a light
beam is sent through a cylindrical glass container that holds the lake water
to be studied. A light cell on the side of this container measures the light that
is scattered out of the beam in an angle perpendicular to the beam. The
stronger this scattering, the higher the turbidity which is expressed in
nephelometric turbidity units (NTU). This is an essentially arbitrary unit
determined relative to that of artificial standard suspensions made up in a
prescribed manner. In practice, nephelometric turbidity corresponds closely
to the scattering coefficient, b (Kirk, 1994). Since Secchi-depth and vertical
light attenuation also depend on absorption, it follows that these apparent
optical properties can not simply be derived from nephelometric turbidity
measurements.
Euphotic depth
Sometimes the light climate of a lake is characterized in terms of its
'euphotic depth'. This is the depth beyond which the light level falls below
1% of the surface irradiation which is considered too low for algae to
maintain a positive net photosynthesis. Obviously, this is a rough approxi-
mation, as the absolute light at that level depends on the surface irradiation,
and different algal species will require different amounts of light.
Substituting the ratio 100:1 of / 0 to I, in Eq. 2, it can be seen that there is
a fixed inverse proportionality between euphotic depth (z,.) and the vertical
attenuation coefficient:
z ,4.6 (7)
eu E
However, the most penetrating colours that can be used for photosynthesis
have an attenuation coefficient (Em 1.) of about 75% of that measured over
26 The abiotic environment
the whole PAR spectrum. Therefore, Emin is often used instead of E, and in
that case the relationship becomes:
(8)
6
5
., 4
E 3
1
0
5
contribution of seston fractions
4 to inverse Secchi-depth
'e
0
Zeeltje Galgje Wolderwijd Usselmeer Markenneer
1987 1987 1981 1990 1990
Fig. 2.4 The average summer concentrations of different seston fractions in five
Dutch lakes and the contribution of these fractions to the vertical light attenuation
(E) and to the inverse Secchi-depth (liS) in the lakes estimated using the regression
equations presented in Tables 2.1 and 2.2.
• Volkerakmeer
Veluwemeer
- - Wolderwijd
X
--Markermeer
- • - • Drontermeer
- - Reeuwijk
Haringvliet
-· -- -· -- · Volkerakmeer
- • Usselmeer
0~'---~---.-----.----~----------~---4
0 4 10 12 14
Predicted light attenuation (m- 1)
dS =.!:.-~s (13)
dt D D
Note that the depth of the water column occurs in both terms: in the gain
term because the suspended material becomes more diluted if the water is
deeper, in the loss term because a sinking particle reaches the bottom
sooner in shallower water. As a result of this inverse proportionality to
depth, the rates of these processes can become very high in shallow water.
The sinking rate of a particle depends on its specific weight but also on its
size and shape. Light particles with irregular shapes sink relatively slowly.
Since seston usually consists of a large variety of particles, some fractions
will settle much slower than others. Typically, however, sinking velocities of
suspended solids are more than a few decimetres per day. As a result the
water column of many shallow lakes could potentially clear out in a few days
if all resuspension through wave action and fish activity was excluded (r "'
Sedimentation and resuspension 33
0). Indeed, when shallow lakes freeze, the quiet water under the ice often
becomes very clear. The rapid settling of material is also apparent when a
bottle of turbid lake water is left to rest. In a day, the water usually becomes
clear, and a layer of settled particles becomes visible at the bottom.
The equilibrium concentration of suspended solids in the water column
(S*) is reached when sedimentation equals resuspension. From the above
equation it follows that:
S* =!:. (14)
s
If the effect of factors like wind velocity or fish activity on resuspension (r)
is known, these simple equations allow a translation into effects on sus-
pended matter concentrations in a lake.
·~
~ 0.16
."
~
~
_g·~ 0.08
wind velodty (m s-1)
E
"
E
0 3
Depth (m)
Fig. 2.6 Relationship between the maximum horizontal water velocity and water
depth for a fixed fetch of 1000 m and wind velocities of 2.5, 5.0 and lOms-'respec-
tively. Redrawn from Aalderink et aL (1985).
variation in the relation between shear velocity and suspended solids (Ham-
ilton and Mitchell, 1996).
Another, more pragmatic, approach is to use relatively simple empirical
formulae that give wavelength as a function of wind velocity and fetch, and
subsequently apply the rule of thumb that resuspension occurs if the waves
'touch the bottom' which is considered the case if the wavelength exceeds
twice the water depth. Carper and Bachmann (1984) show that this simple
approach, developed originally by engineers working with problems of
beach erosion, actually works well to describe resuspension in the shallow
prairie lake that they studied. Because the approach is relatively transpar-
ent, the formulations are used here to explore the effect of lake depth and
size on susceptibility to resuspension.
As long as waves do not touch the bottom they are called 'deep water
waves'. The size of such waves increases in a predictable way with wind
velocity, W (ms-1) and with the fetch, F (km) which is the distance to the
shore measured in the direction from which the wind comes, i.e. the distance
over which the waves have been allowed to build up. A relatively simple
Sedimentation and resuspension 35
empirical formula gives wavelength (L.) as a function of fetch and wind
velocity:
(15)
Wavelength increases almost linearly with wind velocity, while the increase
with fetch is clearly non-linear (Fig. 2.7).
The latter can be observed in any pond if there is some breeze. At the
sheltered shore the water is quiet, but the size of waves rapidly increases
with the distance from the shore. Further from the sheltered shore waves
keep growing but this increase with fetch is less steep than that observed
over the initial few metres. The formula can be used to generate a map of
wavelengths in a lake, given its contours and the wind speed. The area
where sediment is predicted to become resuspended at given wind speeds
can subsequently be found by overlaying the map of water depths, and
applying the rule of thumb that resuspension occurs if the waves 'touch the
ground', that is if the wavelength exceeds twice the depth (L. > 2D). Obvi-
ously, the resuspended area increases with wind speed (Fig. 2.8), and shel-
tered areas are affected only at the highest wind velocities.
-4
fetch (km)
Wind velocity (m s- 1)
0
Fig. 2. 7 Increase of wavelength (L.) with fetch (F) and wind velocity (W) in deep
water as described by Eq. 15.
36 The abiotic environment
Fig. 2.8 Contours showing the wind velocities (kmh-1) necessary for resuspension to
occur at a southeasterly wind in the shallow Little Wall Lake. Wave resuspension is
predicted to occur in the area northwest of the contours. From Carper and
Bachmann (1984).
Fig. 2.9The proportion (a) of a hypothetical square lake where wave resuspension
occurs depends on the critical fetch (F,",) at which the wavelength exceeds twice the
depth (D), relative to the total length (I) of the lake measured in the direction of the
wind.
(16)
Substituting this and the critical condition for resuspension (at D = 2 Lw) in
Eq. 15, we obtain a single formula that relates the resuspended fraction (a)
to the maximum fetch (Fmax) and the depth (Dm) of the lake and to the wind
speed (Wms-1):
{17)
For a given {hypothetical) lake of fixed depth and size, this formula can be
used to plot the increase of the resuspended area with wind speed (Fig.
2.10).
At low wind speeds no resuspension occurs as the critical fetch is larger
than the maximum fetch in the lake. Above a critical wind speed, the
resuspended fraction of the lake rises asymptotically to 1 with increasing
wind.
To see better how the effects of fetch and depth interact, we change the
viewpoint and ask the question which combination of lake depth and size
leads to a 50% resuspension (a = 0.5) at a given wind speed (Fig. 2.11 ).
38 The abiotic environment
"'~
_,.""'
..!ll
0 E
:e" :Ei"
0
0.5
~
al
.,g.
""0 li
""6l- .5
0 5 10
Wind speed (m s·1)
Fig. 2.10 Relationship between wind velocity (W) and the fraction (a) of the surface
area of hypothetical lakes of 1m 1 km (Fig. 2.9) where resuspension occurs as
predicted for three different lake depths.
The resulting iso-resuspension lines are not straight lines, implying that it
is not simply a fetch/depth ratio that counts. Therefore, scale models of
lakes can not be used for studying resuspension. The sediment of a pond
with a maximum fetch of lOOm and a depth of 0.5m is more easily
resuspended than a lake with a maximum fetch of 1 km and a depth of 5 m.
A plot with the logarithm of the lake area shows more precisely how
resuspension susceptibility depends on size and depth of lakes (Fig. 2.11b}.
All other things being equal, lakes that are on the same iso-resuspension
line in this plot, should have a comparable sensitivity to wind resuspension
according to our simple model. Thus, it can be seen, for instance, that a pond
with a size of one hectare and a depth of 0.5 m is comparable to a lake of
lOOha and a depth of 1.3m. Because of the sharp decrease of resuspension
with water depth, a change in water level can affect wind resuspension in a
lake rather strongly. This is illustrated, for instance, by the case of Lake
Chapala in Mexico (Lind et al., 1994). A drop in water level caused a
considerable increase in clay resuspension and turbidity in this lake.
The amount of material that is brought into suspension when the waves
touch the sediment depends on the situation. In general, the ongoing pro-
cess of sedimentation and resuspension leads to a sorting of material in
lakes. The sediment in exposed shallow areas with frequent resuspension is
coarse, because lateral transport causes the fine material to concentrate in
0
0
Lake length (km) (a)
6
§
.r;
fr3
"0
~
2
(b)
Fig. 2.11 Iso-resuspension lines indicating at which conditions 50% of the bottom of
hypothetical lakes (Fig. 2.9) is subject to wave resuspension at different wind veloc-
ities (W). Lakes that are on the same iso-resuspension line are comparable in their
susceptibility to wind resuspension. Note that susceptibility to resuspension de-
creases rapidly with lake depth. Size of the lake can be expressed as maximum fetch,
i.e. the length of the lake measured in the direction of the wind (a), or as lake surface
area (b).
40 The abiotic environment
deeper sheltered parts where resuspension occurs rarely (Evans, 1994). As
pointed out by Carper and Bachmann (1984) this implies that the exposed
'erosion areas' are often not an important source of suspended solids. In
such lakes, resuspension only becomes important in periods when winds are
strong enough to affect areas in which resuspension occurs rarely, as easily
resuspendable material is restricted to these areas. By definition,
res'uspension is thus relatively unimportant most of the time in lakes in
which the exposed areas contain just coarse sediment due to horizontal
sorting.
On the other hand, there are many lakes in which there is hardly any
horizontal sorting because there are no deep parts where soft sediment can
accumulate, and resuspension frequently occurs over most of the area. Such
lakes often have a more or less discrete top layer of sediment consisting of
fine material that is frequently resuspended (Luettich, Jr. et al., 1990;
Bengtsson and Hellstrom, 1992). Due to the frequent resuspension there is
little consolidation of this layer and the material is resuspended easily.
Obviously, if there is such a relatively discrete resuspendable layer, the
amount of suspended sediment should simply increase linearly with the area
over which resuspension occurs. This is indeed found, for instance, by
Bengtsson and Hellstrom (1992) in their studies of Lake Tiimnaren (Fig.
2.12).
In principle, a linear increase of suspended solids with the resuspended
100
0
-bll
80
5
t
:1:'
"'E 60 0 0
"0
"0
~
" 0
"5l- 40
;:;
0
b
~
b 20
£
Fig. 2.12 Measured concentration of suspended matter versus the computed fraction
of the lake area where wave resuspension occurred at the time of measurements in
the Swedish Lake Tamnaren. From Bengtsson and Hellstrom (1992).
Sedimentation and resuspension 41
,-, .............
/,---- --
'I
,,,
t::l/
I
/,'
"'I
"'I
&/,
I
I
I
So ===:::. _____________ _: I
Wind speed
Fig. 2.13 Predicted increase in suspended solid concentration (S) with wind velocity
for a large shallow versus a smaller or deeper hypothetical lake (Fig. 2.9) (see text).
+~
w•·
S*-S
- b s (19)
The background concentration (S.), sinking rates and the parameters a, and
b, are simply tuned in such a way that the model results fit to time-series of
wind speeds and suspended solid concentrations in the lake. This approach
has been used to describe the situation in Lakes Balaton (Somlyody, 1982;
Somlyody and Stanbury, 1986), Arres¢ (Kristensen et at., 1992) and·
Veluwemeer (Aalderink et al., 1985). The increase of resuspension in the
fitted model can be either concave (b,; 0.4 in Veluwemeer) or convex (b,;
1.45 in Arres!ll) (Fig. 2.14).
Although this difference is surprising at first sight, a look at the theoreti-
cally derived relationship (Fig. 2.13) may explain the discrepancy. As ar-
100
80
-
5""
,~ 60
,~
,"' 40
"lit-"'
:::1
V> 20
0
0 2 4 6 8 10
Wind speed (m s·1)
Fig. 2.14 Increase of suspended solid concentration (S) with wind velocity (W)
according to Somlyody's empirical model fitted to data from two different lakes (see
text).
Sedimentation and resuspension 43
Fig. 2.15 Benthivorous fishes such as bream (Abramis brama) usually dominate the
fish community of turbid shallow lakes. These animals can increase turbidity by
whirling up the sediment in search for benthic food. They also stimulate algal blooms
by enhancing the nutrient flow from the sediment to the water column and by
consuming waterfieas that would otherwise graze on algae.
50.---------------------------------------.
• - - - - • Galgje (treated area)
- Zeeltje (control area)
~40
.§.
:2
~ 30
13
'0
~
~ 20 ,
u
-·-
-~
.
!'!'
~ 10
'\
,'•
,'
................ -.... . . . . . .........
... ......
0+-.----.---,----.----.----.----.--~
May June July Aug. Sep. Oct. Dec.
1987
Fig. 2.16 Inorganic solids concentration in Lake Bleiswijkse Zoom following fish
removal in April1987. From Meijer et aL (1989).
Sedimentation and resuspension 45
in transparency which appeared to be largely due to a drop in the concentra-
tion of inorganic suspended solids (Fig. 2.16).
Theoretically, the effect of benthivores on suspended solids can be un-
derstood from the basic equations that describe the balance of resuspension
and sedimentation in a similar way as wind resuspension (Eqs. 18 and 19). If
the daily amount of material that is resuspended is assumed to be propor-
tional to the biomass of benthivorous fish Bb the model can be written as:
dS = q Bb _.!...(s-s) (20)
dt D D o
(21)
where q is the amount of sediment stirred up per unit of fish biomass each
day.
This simple model is surprisingly well in line with observations. For
instance, the concentration of inorganic suspended solids in several Dutch
ponds and small lakes where wind resuspension is unimportant shows a
linear increase with the biomass of benthivorous fish (Fig. 2.17).
40-
..
~
s~
30
:g
Si 20
5}-
0!
.!.!
""'e!'
0
E
Fig. 2.17 Relationship between inorganic suspended solid concentrations and the
biomass of benthivorous fish in several Dutch ponds. From Meijer et al. (1989).
46 The abiotic environment
"~ 300
"tJ
~
.!:9
..,
<ii
200
~
E
"tJ
~
"
E 100
'0
&I
1:bl}
·a:;
;;:
0
<='
Cl 0 200 400 600
Benthivorous fish (kg ha-1)
80
- 60
.s
bl}
32
~ 40
"tJ
"
"tJ
""~ 20
"
V>
0
0 200 400 600
Benthivorous fish (kg ha-1)
Fig. 2.18 Increase of the flux of sedimentating material (upper panel) and the
concentration of suspended solids with the biomass of benthivorous bream in a
series of experimental ponds. Modified from Breukelaar eta!. (1994).
The effect of sediment resuspension by fish has also been studied experi-
mentally in a series of ponds that were stocked with carp (Cyprinus carpio)
and bream (Abramis brama) of different size classes and in different den-
sities (Breukelaar et al., 1994). As expected, both the sedimentation rate,
measured with sediment traps, and the concentration of suspended solids in
the water column increased approximately linearly with fish density (Fig.
2.18).
The impact of resuspension on turbidity in these ponds is considerable.
Using regression models to separate the effect of changes in phytoplankton,
Breukelaar et al. (1994) estimated that, roughly speaking, resuspension by a
Sedimentation and resuspension 47
moderate benthivorous stock of 30kgha-1 suffices to reduce the Secchi-
depth from crystal clear water to less than 1 m. This impact is not surprising
in view of the activity of the fish. An average bream was computed to
suspend five times its own body weight per day (q = 5gg-1 day-1) in these
ponds.
Obviously, the obtained results can not be simply extrapolated to other
situations. The activity of fish may vary from case to case, and the type of
sediment will affect resuspension as well as the settling rate of particles. The
sediment in the experimental ponds consisted mainly of clay. On sandy soils
the effect will probably be less, while, on the other hand, certain soft organic
sediments can have much lower settling rates, increasing the potential im-
pact of benthivores on the concentration of suspended material.
As mentioned earlier, the sensitivity of sediment to resuspension
by waves depends strongly on the state of the sediment surface layer. If
the sediment is left undisturbed, the critical shear needed for resuspension
increases over time due to consolidation. In view of this mechanism,
benthivorous fish may be expected to increase the sensitivity of shallow
lakes to wind. Their -activity keeps the sediment from consolidating in
periods with low wind. As a consequence, a smaller shear and therefore
less wind is needed for resuspension. Obviously, this mechanism will not
be important if the fetch to depth ratio of a lake is such that wind
resuspension itself is very frequent, preventing consolidation altogether, as
in Lake Arres(ll mentioned in Chapter 2. Nor will it be of any significance
in cases where wind resuspension is very rare, as in the experimental
ponds discussed above. In intermediate situations, however, this indirect
effect of benthivorous fish on resuspension may be expected to contribute to
turbidity.
700
Spring (no vegetation present)
Summer (vegetation present)
600
j 500
E
~ 400
1
-'=' 300
:e
'C
:l
f- 200
100
0 0 0
---o---o---~-------·
6 10 14 18 22 26 30 34
Wind speed (miles per hour)
Fig. 2.19 Turbidities of Lake Chautauqua, Illinois, USA, occurring at various wind
velocities in the spring when no vegetation is present and in the summer when
vegetation has developed. Redrawn from Jackson and Starret (1959).
F M A M J A 5 0 N D
Fig. 2.20 Difference between the seasonal changes in total-P concentrations in the
epilimnion of stratified lakes (upper panel) as opposed to total-P dynamics in shal-
low (well mixed) lakes (lower panel) in Denmark. Data are from a eutrophic set of
lakes (0.2 < total-P < 0.5mgll). Redrawn from Jeppesen (1996).
release of nutrients from the sediment (Jeppesen eta/., 1996). As a result the
nutrient concentration in shallow lakes tends to follow the opposite seasonal
pattern of what is generally observed in stratified lakes (Fig. 2.20).
Riley and Prepas (1985) found that, on average, the total phosphorus
concentration in mixed lakes increased by 57% from the spring to summer
while the summer values in the epilimnion of stratified lakes in their data set
were 13% below the spring concentrations on average.
Phosphorus has probably received more attention than any other nutri-
ent in limnology. In shallow lakes, the intense sediment-water contact gives
an extra dimension to the eutrophication problem. Much of the phosphorus
that has been absorbed by the sediment during eutrophication can be re-
leased to the water column later. This 'internal loading' can cause a delay of
many years in the response of lake water concentrations to a reduction of
the external loading. For nitrogen the sediment-buffer effect is less relevant
(Jensen et al., 1991). Instead, it has been shown that substantial amounts of
nitrogen can disappear from shallow lakes as a result of denitrification.
Although nitrogen limitation occurs frequently, its dynamics have been
studied less extensively.
Particulate P
Total P
Soluble P <~·---·
Soluble Unreactive P
Fig. 2.21 The total pool of phosphorus in lake water is split up in fractions that can
be distinguished by simple techniques. The particulate fraction is separated by
filtration from the total soluble phosphorus. The latter is further divided by chemical
methods into soluble reactive phosphorus (SRP) and soluble unreactive phosphorus
(SUP).
Unfortunately, this subdivision does not give much insight into what is
actually available for algal growth. First of all, it is important to distinguish
between immediate availability, and long-term availability. Immediately
available is what can be taken up by phosphorus starved test algae in the
laboratory within a few hours (Bostrom et al., 1988b). It was long assumed
that SRP was a good estimate of the immediately available fraction. The
idea was that SRP was largely equivalent to orthophosphate (HPO/-,
H 2P04-), and that this orthophosphate was the sole form of phosphorus
utilized by algae. It is now clear that neither of these assumptions is really
correct and that there is not even a fixed proportionality between immedi-
ately bioavailable phosphorus and chemically assessed SRP (Bostrom et al.,
1988b). Nonetheless, laboratory experiments to estimate the immediately
available pool are tedious, and SRP still gives the best estimate in practice.
With respect to understanding why some lakes have a higher algal
biomass than others, immediately available phosphorus is not really the
most relevant statistic. Many of the forms of phosphorus that are not di-
rectly available to algae can be transformed into available forms relatively
quickly. Desorption and dissolution can make part of the inorganic
particulate phosphorus available, and also the turnover of phosphorus that
is present in algae can be extremely rapid (e.g. Rigler, 1956). Obviously, it
would be useful to have an indication of the total amount of phosphorus that
is available to algae. In practically all eutrophication studies the total con-
centration in the water column ('total-P') is used as such. This pragmatic
solution has two problems. First of all, part of the phosphorus fractions in
the Jake water can not be converted into available phosphorus. Secondly,
and in our context more importantly, in shallow lakes there is an intensive
exchange between phosphorus in the water column and phosphorus in the
52 The abiotic environment
sediments. Thus, a substantial part of the relevant available phosphorus
pool in shallow lakes is present in the sediment rather than the water
column. Therefore, the traditionally used 'total-P' is in fact a far from
perfect indicator for the nutrient status of shallow lakes. Release of SRP
from the sediment into the water depends on the composition of the sedi-
ment and the SRP concentration in the lake water (S!Zindergaard et a/.,
1992), but also varies strongly depending on the conditions at the sediment-
water interface. Understanding this sediment-water interaction is therefore
crucial for understanding the phosphorus dynamics of shallow lakes.
(22)
Later an exponent was added to the function to obtain a slightly better fit
(Vollenweider and Kerekes, 1982), but the above formula has become well
known as the Vollenweider model.
Since the empirical Vollenweider equation describes a generic relation-
ship between input and the equilibrium concentration in the lake, it can in
principle also be used to predict the effect of a reduction in nutrient loading
on the concentration in the lake water. After a transient period the nutrient
concentration in the lake should settle to a new equilibrium value depending
on the new input concentration and hydraulic retention time in the way
described by Vollenweider's equation. If the hydraulic situation (and hence
t,) is unaltered by the restoration measures, a short-cut is sometimes made
by estimating that the mean P concentration in the lake should decrease
roughly in proportion to the change in P input (Sonzogni et al., 1976).
Jeppesen and co-workers (1991) used this approximation to study the set-
Nutrient dynamics 53
n.a.
0
15
>. 10
1
!:!:
"0 5
"'-!l"
""2
~
n.a.
0 0
.<:
~
.<:
"-
-5
-10
78 79 80 81 82 83 84 85 86 87 88 89 90
Year
Fig. 2.22 External phosphorus loading (upper panel) and phosphorus retention
(lower panel) in the Danish Lake S!i!bygaard from 1978 to 1990. Before 1983 the
loading was high and the lake accumulated phosphorus. After a strong reduction in
the external loading at the end of 1982 there is a net loss of phosphorus from the lake
that continues for years, indicating that the sediment keeps releasing phosphorus to
the lake water. From S!i!ndergaard eta/. (1993).
tling to the predicted new state for 27 Danish lakes that had received a
substantial reduction in nutrient loading. It appeared that even 4-16 years
after the reduction in loading, the decrease in concentration in most lakes
was still far less than expected from the reduction in the input concentration.
54 The abiotic environment
Part of this delay can be explained from the time it takes to dilute the
nutrient rich lake water with the cleaner inflowing water. Assuming a homo-
geneous well mixed system with no exchange between sediment and water,
it can be derived that it takes about three times the hydraulic retention time
to reduce the surplus pool of phosphorus in the lake water by 95% (Sas,
1989). In some cases the delay can indeed be largely explained by this
dilution effect, but usually the response to a reduction of the inflow concen-
tration takes much longer. The main reason for this is that the sediment
starts acting as a source rather than a net sink of phosphorus (Marsden,
1989; Sas, 1989; Jeppesen et al., 1991). This is illustrated, for instance, by the
response of Lake S!llbygard (Fig. 2.22) to a reduction of the external nutrient
loading.
In the period prior to the restoration efforts the lake showed a net
retention of phosphorus (as expected from the Vollenweider relation). Af-
ter a strong reduction of the inflow concentration, however, the lake started
showing a negative retention (that is a net release) of phosphorus. Although
this 'internal phosphorus loading' decreased gradually over the eight years
studied, the authors suggest that the sediment contains enough phosphorus
to support a net release for another 10 years or so.
In view of these buffer effects it is not surprising that in the years follow-
ing a reduction of the external loading, phosphorus concentrations in shal-
low lakes are correlated with release from the sediment rather than with
inflow concentrations (Vander Molen and Boers,1994). From a restoration
point of view it is therefore useful to be able to predict the effect of internal
loading on the lake water after an intended reduction of the inflow concen-
tration. A reasonable guess would be that the phosphorus content of the
sediment is indicative of the potential internal loading. Correlative studies,
however, show that the concentration of phosphorus in the lake water is not
(Jensen et al.,1992) or only weakly (Vander Molen and Boers, 1994) related
to the phosphorus concentration in the sediment. Instead, the concentra-
tions in the water tend to correlate well with the ratio between phosphorus
and iron concentrations (P: Fe) in the sediment (Jeppesen et al., 1991;
Jensen et al., 1992; Vander Molen and Boers, 1994). This is presumably
because iron is the most important agent binding phosphorus in the aerobic
upper layer of the sediment in most lakes. Interestingly, for the subset of
lakes where the P/Fe ratio (gig) in the sediment is lower than 1110, the
correlation with lake water concentrations becomes weak (Jensen et al.,
1992; Vander Molen and Boers, 1994). This suggests the simple rule that
iron in the sediments of these shallow lakes is able to bind more or less
permanently an amount of phosphorus equivalent to about 10% of its own
weight, and that it is basically only the surplus phosphorus that constitutes
the pool from which there can be a release to the lake water.
This empirical 10% rule is also reflected in the vertical concentration
gradient of iron and phosphorus in the sediment of Lake S!llbygard
(S!Ilndergaard et al., 1993). As mentioned earlier, eight years after reduction
Nutrient dynamics 55
of the external loading, the sediment was still releasing phosphorus (Fig.
2.22). The depth profile of the Fe:P ratio (Fig. 2.23), however, shows
that the Fe: P ratio in the upper sediment layers has already stabilized at a
value of approximately 10, suggesting that the main source of released
phosphorus is now the sediment at a depth of around 20cm where the ratio
is lower.
Besides being able to predict the effect of internal loading after a reduc-
tion of the external loading, it is also of practical interest to predict for how
many years this phenomenon is likely to delay the recovery. Since the
phosphorus released by the sediment has to be washed out somehow by the
water flowing through the lake, it seems reasonable to expect that lakes with
a higher throughflow ('flushing rate') recover faster. Danish data, however,
indicate that this is not the case (Jeppesen et al., 1991). Lakes with a high
flushing rate are just as slow to approach the predicted new equilibrium as
the rest. A possible explanation for this is that a high inflow generally also
implies a high overall nutrient loading in the past, allowing a large accumu-
lation of phosphorus in the sediment during the period before restoration.
Indeed hydraulic retention time in the Danish lakes is strongly correlated
with the yearly P-load and the P-pool in the upper 20cm of the sediment.
Thus, although lakes with high flushing rates may have a better potential to
wash out their phosphorus, this advantage seems to be counterbalanced by
the fact that they also tend to have accumulated more phosphorus in the
past.
"E 12
.!l.
~ 16
t3
20
24
28
0 5 10 15 20 25 30
Fe /Totai-P
Fig. 2.23 Variation of the Fe:P ratio with depth in the sediment of Lake Sl1!bygaard.
In the top layer the ratio has decreased to about 10, suggesting that the remaining
phosphorus in this layer can now be entirely immobilized by iron under aerobic
conditions. Redrawn from S!llndergaard eta/. (1993).
56 The abiotic environment
The mechanisms that govern sediment phosphorus release
Although some empirical relationships between lake concentrations, exter-
nalloading and sediment characteristics have been found, a large part of the
variation remains unexplained. Several studies have shown that differences
in, for instance, turbulence, animal activity and plant growth can cause large
variations in the release of phosphorus from the sediment. To understand
these effects, it is necessary to zoom in more closely on the mechanisms that
govern phosphorus dynamics at the sediment surface.
ITuhulence I
;'':
.,/' ~
"
Aerobic layer
t
' 0
'-···-------------------------------·
0 •
Fig. 2.24 Schematic representation of the main processes involved in the internal
phosphorus cycle of a shallow lake. Turbulence promotes diffusion of phosphorus
out of the sediment, but also helps maintaining the anaerobic surface layer where
phosphorus is immobilized by iron. The decomposition process supplies mineral
phosphorus, but also uses oxygen, thereby reducing the size of the aerobic layer.
58 The abiotic environment
at the sediment surface. The maintenance of an aerobic layer depends
critically on the balance between microbial consumption of oxygen in the
sediment and oxygen supply from the aerobic water layer. If oxygen supply
from the water column is insufficient to counterbalance the microbial up-
take, the sediment surface becomes anoxic and phosphorus can no longer be
immobilized by iron.
The importance of turbulence for oxygenating the sediment surface is
well illustrated by an incident that occurred during experiments in an in situ
chamber in Gullmarsfjorden in Sweden (Sundby et al., 1986). When stirring
device in the chamber failed, black anaerobic patches soon developed on
the sediment surface and the SRP concentration within the chamber in-
creased fivefold, even though the water column had remained aerobic.
Although animal activity can contribute significantly to mixing at the sedi-
ment surface, wind-induced turbulence may explain most of the differences
between lakes. Since resuspension and oxygenation of the sediment are
both driven by the turbulence at the sediment surface it is not surprising that
iron controlled phosphorus release is in practice related to lake size and
depth in much the same way as resuspension (Fig. 2.25).
Lakes that are on the borderline in this plot may stratify only during short
periods of warm weather in summer. Such periods may suffice to cause
anoxic conditions in the ephemeral hypolimnion that induce a shot of
anaerobic phosphorus release from the sediment. Subsequent mixing of the
lake may then bring about a significant increase of phosphorus concentra-
tion in the whole water column (Riley and Prepas, 1984; Kallio, 1994). Even
1000,-------------------------------------------.
g •
f1oo
"'
~ 0
E
"
E 50 0
~
0~-----.-----.------.-----,------,----__,
0,1 10 100 1000 10,000 100,000
Surface area (km 2)
Fig. 2.25 The distribution of tbe reported form of sediment phosphorus release
relative to the lake surface area and maximum depth. Closed circles represent
anaerobic release, open circles aerobic release and partly open circles are lakes with
both forms of release. Redrawn from Marsden (1989).
Nutrient dynamics 59
in the absence of distinct stratification, however, a reduced turbulence can
lead to anoxic conditions at the sediment surface as shown by the experi-
mental mixing problem in Gullmarsfjorden.
The overall relationship between turbulence and sediment phosphorus
release is complicated by the fact that turbulence has two opposite effects
(Fig. 2.24 ). It prevents excessive anaerobic phosphorus release by oxidizing
the sediment surface, but it also promotes diffusion of phosphorus from the
aerobic top sediment into the water. The subtle balance between these two
effects is illustrated by a time-series analysis of phosphorus dynamics and
wind velocity for the shallow (2.7m) Dutch Lake Westeinder (De Groot,
1981 ). In this lake the phosphorus release from the sediment peaks on windy
days probably due to enhanced diffusion from the sediment. However,
when longer periods (months) are considered the correlation between aver-
age wind and phosphorus release is reversed; The flux of phosphorus from
the sediment into the water column being higher during calm periods than
in periods with windy weather. Supposedly, diffusion of oxygen into the top
sediment is insufficient in such periods of reduced turbulence to prevent
anaerobic conditions at the sediment, resulting in an elevated phosphorus
release.
Resuspension
At the most turbulent side of the range of mixing conditions, sediment is
resuspended. Depending on the situation resuspended particles can either
adsorb phosphorus from the surrounding water or release it (Serruya, 1977;
Yousef et al., 1980; Gunatilaka, 1982; Lennox, 1984), but in the P-loaded
sediments of many eutrophic lakes release is likely to dominate in summer.
The potential importance of resuspension in enhancing sediment P release
is illustrated well by the work of S!l)ndergaard and co-workers (1992). They
measured how phosphorus release from sediment cores collected in Lake
Arresp increased during experimental resuspension. Internal P-loading in-
duced by resuspension appeared to be 20-30 times greater than release from
the undisturbed sediment. This could not be explained simply from entrance
of nutrient rich pore water into the water column. Also, the SRP release was
not related to the amount of sediment that was brought into suspension.
This suggests that SRP release during resuspension depends largely on the
absorption-desorption kinetics. When the suspended particles are over-
saturated, they release phosphorus. The net release becomes nil when the
SRP concentration in the water has increased enough to be in equilibrium
with the particle bound P-fraction. For sediment sampled in summer, an
extra portion of SRP could be released when the sediment was resuspended
again one day later, but such an extra release could not be obtained from
sediment sampled in the spring. This indicates that it is largely the minerali-
zation of fresh organic material present in the summer sediment that feeds
the pool of exchangeable phosphorus in the top layer of Lake Arres!l).
60 The abiotic environment
Phosphorus-release
~mfdFm~··~~------------,---------------.-------------,
,0
,.
turbulence at sediment surface
Fig. 2.26 Schematic representation of the effect of turbulence at the sediment sur-
face on phosphorus release from the sediment (see text).
Nitrogen dynamics
Nitrogen is less often reported as a limiting nutrient in lakes than phospho-
rus. Also it does not show the strong resilience to reduction efforts described
for phosphorus. Both factors probably contributed to the fact that its dy-
namics have received less attention than those of phosphorus. Although
there are some similarities, the processes that govern nitrogen cycling differ
widely from those implied in the phosphorus dynamics. Three major fea-.
tures set nitrogen aside from phosphorus: it does not accumulate in the
sediment that strongly; it can disappear as gas into the atmosphere under
certain conditions; and some cyanobacteria can use atmospheric nitrogen as
a nutrient (Fig. 2.27).
Decomposition of organic material normally leads to the release of nitro-
gen as ammonium (NH/). The ammonium produced by this so-called
'ammonification' of detritus can diffuse into the water column where it
can be readily used as a nitrogen source by algae and macrophytes. In
the aerobic top layer of the sediment, ammonium can be transformed
microbially to nitrate (NO,-). This process is called nitrification. Although
it can happen under the aerobic conditions in the water column as well, it
-<=:; ...-
Fig. 2.27 Schematic representation of the main processes involved in the internal
nitrogen cycle of a shallow lake (see text).
62 The abiotic environment
is reported to occur predominantly in the sediment where ammonium
concentrations are normally high (Lijklema, 1994). Unlike phosphate,
ammonium and nitrate are hardly adsorbed by sediment particles and do
not normally precipitate to insoluble forms in the sediment either. There-
fore, the strong accumulation in the sediment that is found for phosphorus
in eutrophied lakes does not occur for nitrogen (Jensen eta/., 1991 ). This is
probably the reason why lake water concentrations of nitrogen tend to
respond more promptly to reduction of the external loading than phospho-
rus concentrations.
When nitrate ends up in an anaerobic situation, it can be microbially
transformed into N, which can not be used as a nutrient by most algae, and
largely disappears as gas to the atmosphere. This process, called
denitrification, can constitute the major loss of nitrogen from lakes (Jensen
eta/., 1991; Lijklema, 1994; Windolf et a/., 1996). Because denitrification
itself requires anaerobic conditions, but its substrate (nitrate) is produced
under aerobic conditions, denitrification works mainly where both condi-
tions co-occur or alternate in time. Therefore, the sediment surface is a very
important site for denitrification. Denitrification rates are difficult to meas-
ure in the field. An indirect way to estimate denitrification is through the
mass balance. What goes into the lake and does not leave it is normally
called retention. In the case of nitrogen this is obviously not a very adequate
term, as part of it may be retained in the sediment, but another part is
released into the atmosphere. Jensen and co-workers (1991) showed for a
set of 69 shallow Danish lakes that burial in the sediment could only explain
23% of the total nitrogen loss. The remaining 77% is likely to disappear
largely through denitrification, as the contribution of other loss processes
are thought to be minor.
Empirical studies show that nitrogen loss is highest in shallow systems
(Vollenweider and Kerekes, 1982; Lijklema eta/., 1989; Jensen eta/., 1991).
This fits well with the idea that denitrification is the dominant process
involved and that the sediment surface is the most important site for
denitrification. As with phosphorus, the retention of nitrogen in a lake is
larger when the water passes through the lake more slowly. In shallow lakes
with a hydraulic retention time of more than a month it is not unusual to find
that more than 50% of the nitrogen is 'retained' (Fig. 2.28), most of which
has probably disappeared into the air. Because of this, shallow lakes can
play an important role in reducing nitrogen concentrations in the passing
water.
Obviously, the effect of nitrogen loss on the lake water concentration is
also highest in lakes with a high hydraulic retention time, as the longer water
stays in the lake, the stronger nitrogen concentrations can be reduced by
denitrification and burial in the sediment. Windolf and co-workers (1996)
analysed this effect statistically and found that 83% of the variation in
annual mean lake water concentrations (N) in a set of well studied Danish
Nutrient dynamics 63
100,---------------------------,
Fig. 2.28 Annual mean percentage retention of nitrogen (Nm) versus hydraulic
retention time ( r,) during four subsequent years in 16 Danish lakes. From Windolf et
a/. (1996).
shallow lakes from the inflow concentration (N;) and hydraulic retention
time ( ;,) by the following equation:
(23)
"Totai-P"
"Totai-P"
Fig. 2.29 Effect of algal biomass on the total-P concentration in the water column
(see text).
66 The abiotic environment
Macrophytes
'0 20
2 3 4 5yr
Totai·N
3
-z 2
·,-....~...... .. ......... ...
-----···"..-::-·-· ... ·""' .......·-·-·-.....
""E .............. _______________ _
0
Control 2 3 4 5yr
.
';' 1.00
0.
E 0.50
0.00 +---.-----r---"-::::.:;c==r-------1
Control 2 3 4 5yr
Fig. 2.30 Changes in total-N and total-P concentrations in the water of lakes where
macrophytes have developed in response to a reduction of the fish stock. From
Meijer et at. (1994a).
Bleiswijk
Noorddiep
- -o- - Zwemlust
----+- Vreng
-b/)
0
.5
z
]
.8
c
~ -1 ·············•·····•··················.....
c
-5"'
-2
-30 10 50 90
change in macrophytes (% of lake area)
Fig. 2.31 Relationship between the year-to-year changes in the percentage of lake
area covered by vegetation and change in the total-N concentrations for the
biomanipulated lakes represented in Fig. 3.30.
0~--L_~--~--~~L-----~
2 1987
0
1988
0~------.-----_.~--------~
2 1989
1990
Fig. 2.32 The amount of N and Pin water, macrophytes and phytoplankton before
(1986) and development of aquatic vegetation in response to a reduction of the fish
stock (N.B. water-N is NH4-N and N0 3-N; water-Pis SRP). Modified after van Donk
et al. (1993).
Benthic algae
The sediment surface is usually a place of high microbial activity. As
decribed in this chapter the continuous supply of settling organic material
and the aerobic conditions ensure optimal conditions for decomposing bac-
teria. However, if resuspension does not occur too frequently and some light
70 The abiotic environment
penetrates to the bottom, the surface can also be quickly colonized by
benthic algae. The resulting microbial community of algae and bacteria can
form a soft crust that further reduces the probability of resuspension
(Delgado et at., 1991) and forms a barrier to diffusion between sediment and
water. Obviously, the growing benthic algae benefit from the high nutrient
concentrations at the sediment surface and may take up nutrients that would
otherwise have been released to the water column. Also, they oxygenate the
upper sediment layer, facilitating immobilization of phosphorus by iron. In
addition, the increase of the aerobic surface layer tends to enhance the
coupled denitrification-nitrification process as explained in the previous
section. A set of laboratory experiments with undisturbed sediment cores
from Lake Wolderwijd demonstrates the impact of benthic algae (mostly
diatoms) on nutrient dynamics at the sediment surface (Van Luijn et al.,
1995). The overlying water of the cores was replaced by nutrient free water
in a continuous flow set-up, and the temperature was kept at 20"C. Half of
the cores were incubated in the dark, while the others were continuously
illuminated, allowing development of benthic diatoms. After 10 days the
upper 2cm of the illuminated cores had a markedly higher oxygen content
than the dark cores. The release of mineral nitrogen and silicon from the
sediment into the water column was reduced by a factor 6 in the illuminated
cores relative to the dark ones where the benthic algae were unable to grow.
Also, the flux of ortho-phosphorus was three times lower in the illuminated
cores. In the field, temperature and light conditions are less favourable and
algal productivity was estimated to be only 10--30% of what was realized in
the lab. Nonetheless, the results indicate that benthic diatoms may markedly
reduce nutrient release from the sediment.
In addition to the relatively modest development of benthic diatoms, a
thick mat of filamentous algae can sometimes develop at the bottom of clear
shallow water. Often these mats will be pulled up by trapped gas bells
forming the so-called flab (from 'FLoating Algal Beds'). Strong increases in
SRP can occur in water dominated by filamentous algae, presumably due to
anaerobic phosphorus release from the sediment under the beds (Meijer et
at., 1994a).
Benthic invertebrates
In the case of benthic invertebrates the effect on nutrient cycling is more
complex (Andersson et at., 1988). Not only do they excrete mineral nutri-
ents, they also enhance mixing at the sediment surface. As argued above, the
latter can work in two ways (Figs. 2.24 and 2.26), as it promotes diffusion of
phosphorus across the sediment surface, but also reduces the chance of
anaerobic phosphorus release because of enhanced aeration of the
top sediment. The effect. of tubificids and chironomid larvae on oxygen
penetration in the sediment has been amply documented (e.g. Davis, f974;
Graneli, 1979). Nonetheless, the overall phosphorus release from the
sediment tends to increase with the density of benthic invertebrates, as
shown from comparisons between lakes (Wisniewski and Planter, 1985).
Obviously, such correlative studies do not allow the conclusion that the
benthos is actually the cause of the elevated phosphorus release. Various
laboratory experiments, however, have clearly demonstrated a causal link.
Gallepp (1979), for instance, showed a linear increase of phosphorus release
from the sediment with increasing densities of Chironomus tentans larvae, in
a throughflow system (Fig. 2.33).
Benthivorous fish
Benthivorous fish eat benthic invertebrates. By reducing invertebrate
biomass they can thus potentially reduce the positive effect of these animals
on sediment phosphorus release. On the other hand, most benthivorous fish
whirl up a lot of sediment while foraging. Also, they excrete much of the
nutrients originating from the consumption of benthos, directly into the
water column, thereby acting as a nutrient pump from the sediment to
the water. The relative importance of these direct and indirect effects is hard
to separate in practice, but many experiments have shown that, overall,
72 The abiotic environment
/.
10
.
·,._
8
..,.""0"' /
/.
E
btl 6
.5
2
~
~
4
/.
~
0..
2
/.
0 •
0 5 10 15 20 25
C. tentans density
Fig. 2.33 The release rate of total-Pas a function of the density of chironomid larval
density (Chironomus tentans) per test tube (37.5cm2 ). Redrawn from Gallepp
(1979).
15
'o
~
btl
.5 10
btl
'C"
..9
~
"' 5
2
0
.<:
0
Sl-
.<:
0.. 0
0 25 50 75 100 125
Fish density (g m·2 )
Fig. 2.34 Effect of carp density on phosphorus release into the water column 7-14
days after introduction of carp in enclosures. Redrawn from Lamarra (1974).
100
50
10 15 20 25 30 10 15 20 25 30
day day
Fig. 2.35 Total-P and chlorophyll concentrations in lake water and in enclosures set
up in a shallow isolated bay of Lake Erie (Ohio, USA). In enclosures where fish are
prevented from reaching the sediment by installing a net just above the bottom fish
effects on total-P and chlorophyll are reduced almost to no-fish levels. Modified from
Havens (1993).
While these results indicate that the effect of fish on phytoplankton in this
case was mainly due to the benthivorous behaviour, the exact mechanism is
not clear. The most obvious causal link is through nutrient enrichment.
Indeed all published enclosure experiments show an elevated total-P
concentration in response to the benthivorous behaviour of fish. Most of the
total phosphorus content of the water column, however, is usually contained
in algae and other suspended material rather than being present as dissolved
ortho-phosphate. As argued this implies a chicken-and-egg problem. Other
mechanisms may contribute to the elevated phytoplankton biomass, and a
high algal biomass can be the cause rather than the effect of the high total-
p content of the water, as explained earlier.
Therefore, to check whether the phosphorus pump mechanism can be
responsible for the effect of benthivorous fish on phytoplankton biomass,
ortho-phosphorus concentrations are more informative than the total-P
content. The results of the experimental pond study discussed earlier in the
section on resuspension by fish (Breukelaar et al., 1994) show that at least in
this case, enhanced phosphorus input into the water phase can not be the
mechanism behind the increase of total-P and chlorophyll concentrations
with fish biomass (Fig. 2.36).
Indeed, the chlorophyll-a concentrations are lowest in the ponds with low
benthivorous fish densities, and this pattern is also recognizable in the total-
p concentrations. However, the ortho-phosphate available for algal growth
shows no consistent relation to benthivorous fish biomass, indicating that
phosphorus-limitation is an unlikely explanation for the effect of fish on
algae.
0
30
c
25
-
""
3- 20
~
""
~ 15
-a
0
"" 10
"0
0
June July Aug. Sept. Oct.
0.6-.-------..,.,------------,
- 0.4
.••..•. .
§
'l' •
~
~
...•
0 0.2
:E
u
Fig. 3.1 Relationship between the summer average totai-P concentration (P mgl- 1)
and the summer average chlorophyll-a (Chla mgl-1) concentration in a set of 88
Dutch shallow (2.lm mean depth) lakes. The line (Chla = 0.9P) is fitted by eye to
represent the upper limit of the cloud of 406 data points.
78 Phytoplankton
a sigmoidal increase is probably the general pattern (McCauley et ill., 1989;
Prairie et al., 1989; Watson et al., 1992).
Although the correlations are usually significant, the unexplained
variation in relationships between algal biomass and the phosphorus
level is typically huge in shallow lakes. It has been noted that only the upper
limit of the cloud of points in scatter plots is relatively well delineated
(Fig. 3.1).
This suggests that it might be more appropriate to interpret the effect of
nutrients as posing an upper limit to the summer chlorophyll concentration,
than to fit a regression model through the points (Hosper, 1980). All points
above the limiting line in this figure represent lakes dominated by
filamentous cyanobacteria. This phytoplankton group is able to reach a
higher biomass with the same amount of phosphorus than other algae. As
explained later, this has important implications for the potential of
cyanobacteria to dominate the plankton in turbid shallow lakes.
Since relatively few lakes have chlorophyll concentrations close to the
upper limit, other factors apparently have a pronounced effect on algal
biomass in most cases. One obvious possibility is nitrogen limitation.
Phytoplankton cells contain approximately 10 times more nitrogen than
phosphorus. Therefore, nitrogen is more likely to be limiting than phospho-
rus when the N:P ratio in the water column falls below a critical value of
about 10 (see Smith 1982). In macrophyte dominated lakes in particular
available nitrogen concentrations can be strongly reduced as explained in
the previous chapter, and nitrogen is often found to be limiting algal growth
in such cases (Van Donk et al., 1993). Indeed, a plot of chlorophyll con-
centrations against total-N shows a sharp upper limit (Fig. 3.2) indicating
that nitrogen availability limits algal biomass in many cases. Note that
intersection of the limiting line with the horizontal axis suggests that at least
about 0.7mgt' of the total-N content of the water is unavailable for algal
growth.
The approach of using such upper limits of scatter plots to predict the
maximum algal biomass at a given nutrient concentration has been followed
for many years in The Netherlands. The majority of the data points, how-
ever, are neither close to the P-limitation nor to the N-limitation line,
suggesting that other factors are often limiting. Light is an important limit-
ing factor in many turbid lakes. Also, on sheltered sites and between aquatic
plants losses due to sinking can be large while in fast flushed lakes algal
concentrations can be low because the inflowing water that is initially almost
free of algae stays in the lake far too brief a time to allow the algae to reach
the maximum biomass. Especially spectacular and important are the effects
that zooplankton grazing can have on phytoplankton. To obtain an insight
into how factors like lake depth, flushing, resuspension and grazing interact
with the effect of nutrients in determining algal biomass it is useful to go
beyond the empirical regression approach and analyse algal growth in a
more mechanistic way. In the next sections some simple models of algal
The regulation of algal biomass 79
0.6
0.5
-biJ
.s 0.4
'l'
'5. 0.3
•
.<:
c..
e0 •
:E 0.2
u
0.1
0
0 2 4 6 8 10
Total-N (mg 1"1)
growth are presented for this purpose. The insights from these models serve
as a framework to discuss the patterns observed in the ·field.
dA
dt
=rA(l-~)
K
(1)
K
Population density (A)
~
b ~
e
lJ
c:
0
~
:;
"-
0
Q.
K
Population density (A)
K
$
a
·o;
c:
c "c:
"0
0
~
:;
"-
0
Q.
time
Fig. 3.3 Some properties of the logistic growth equation: (a) The relative ('per
capita') growth rate (dA!dvA) decreases linearly with the population density. (b)
The growth rate (dA!dt) of the total population has a maximum at half the carrying
capacity. (c) Population density of a growing population evolves in a sigmoidal way
over time.
The regulation of algal biomass 81
zero due to competition. Note that the logistic equation predicts that opti-
mal harvest can be obtained by maintaining the population at a density of
half its carrying capacity. Plotted against time the density of a logistically
growing population follows a sigmoidal curve, with the steepest growth
occurring at half the carrying capacity (Fig. 3.3c).
Its simplicity makes the logistic equation very attractive for use in simple
models. Obviously, the model is a crude simplification. A strictly linear
decrease of per capita growth with density seems unlikely to be found in any
real-life situation, and it may seem even more unlikely that the dynamics of
the total biomass of a complex algal community with many species can be
described by such a naive model. Nonetheless, the development in time of
natural algal communities in lakes appears to fit reasonably well to the
logistic equation in practice (Heyman and Lundgren, 1988).
In a simple model like the logistic equation, many underlying regulatory
mechanisms are not explicitly included. Clearly, factors that affect growth
such as the nutrient status and the depth of a lake should affect the param-
eters, but a priori it is not obvious how. The maximum growth rate r is
probably rather species specific. It varies with physical conditions such as
incident light and temperature, but does not change too much with the
nutrient level (Heyman and Lundgren, 1988), although a shift in species
composition with enrichment may of course alter the r at the community
level. The effects of different factors on K are somewhat easier to estimate
from field data, as data on biomass are far more abundant than information
on maximum growth rates. The empirical evidence shows that the maximum
algal density in lakes increases with the nutrient content, and, indeed,
increasing K is a usual way to mimic nutrient enrichment in the logistic
model.
To explore in a more mechanistic way how the combined effects of lake
depth and nutrients can be explained, it is necessary to elaborate the de-
scription of algal growth further as explained later. The simple logistic
equation, however, serves as a useful basis to analyse the effects of losses
due to sedimentation, lake flushing and grazing by herbivorous zooplankton
and other filter feeders.
k=_!_ (2)
' D
In principle algal cells are prone to the same processes as sediment particles.
Some species are able to swim actively or regulate their buoyancy. Most
species, however, lack these abilities and may therefore suffer large losses
due to sedimentation when there is little or no resuspension. Such a situa-
tion is most likely to arise on sheltered or vegetated sites and on days when
wind speed is low and warm weather enhances the chance of thermal strati-
fication. An important feature that sets algae aside from dead sediment
particles is their ability to reproduce. Even for species that are unable to
swim or float, reproduction can at least partially balance the settling loss. A
basic idea of how this potentially affects the phytoplankton dynamics can be
obtained using the logistic equation for algal growth, adding an extra term
to represent the settling loss:
dA
dt
=rA(l-~J-_!_A
K D
(3)
The algal density in equilibrium A*, when growth is zero by definition, can
be found by solving the equation for dA/dt = 0:
(4)
s
r>- (5)
D
Phrased in biological terms this result is intuitively straightforward. At the
border of extinction algal density is very low. In that situation competition
is practically nil and the remaining algae reproduce at the maximum rate r.
However, this high reproduction will only save the population if r exceeds
the sinking loss s/D. This implies that settling should select for algal species
with a low sinking velocity and a high growth rate. Both properties tend to
go with small cell size. Even for small algae, however, s will be no less than
about 0.25md-' and r no more than about 0.5d-' in general, implying a
required depth of at least 0.5 m to prevent extinction in completely still
water.
The regulation of algal biomass 83
1.0
Q
"
0
1
:2"
0.5 15
"
1i
8"
E
"
'i:
&
'5
g-
0
Fig. 3.4 Equilibrium concentration of a logistically growing phytoplankton popula-
tion with sinking losses depending on the sinking rate of cells (s md- 1} and the depth
of the water column (D m)
Obviously, species that have high sinking rates and are unable to swim or
float upwards thus depend on turbulence for survival in the water column of
shallow lakes. As discussed in the previous chapter, the resuspension fre-
quency depends mainly on the depth and size of a lake. Large exposed lakes
should therefore be expected to support species with potentially high sink-
ing losses better than smaller sheltered ones. Also windy periods should
favour such species, and therefore variation in wind could influence the
seasonal dynamics of algae.
The importance of resuspension for phytoplankton in shallow lakes is
well illustrated by the close correlation between wind speed and the biomass
and species composition of the algal community in the Florida Lake Apopka
(Schelske et al., 1995). In this large and shallow lake (128km2 , 1.7m mean
depth) as much as 53% of the variation in chlorophyll contents of the water
column is explained by wind (Fig. 3.5). During calm periods relatively large
diatoms (20-200 ,urn) settle to the bottom and the surface water is domi-
nated by small pico and nannoplankton (<20,um).
Systematic seasonal variations in wind may also play a role in the sea-
sonal succession of phytoplankton. Gons et al. (1991), for instance, showed
that the seasonal pattern of diatom abundance in the shallow Loosdrecht
lakes suggests a dependence on turbulence. He computed an estimation of
sinking loss assuming that sinking loss in parts of the lake where no
84 Phytoplankton
250-,--------------------------- ----------,
•
200
1150
~
.• .• •
~ •
~
0 100
:;:
u
•
50 •
0 10 15 20 25
Wind speed (km h" 1)
Fig. 3.5 Relationship between average daily wind speed and the chlorophyll-a
concentration in surface water of Lake Apopka. Redrawn from Schelske et al.
(1995).
resuspension occurs equals the ratio of settling velocity to lake depth (Eq.
2), while sinking loss was assumed to be nil in parts where waves stir up the
sediment. The fraction of the lake area where resuspension occurred on any
given day was estimated from wind speed data using a simple model (Gons
et al., 1986). Indeed, the spring peak of diatoms coincides with relatively
windy periods, whereas the calm weather in summer leads to estimated
sinking losses of 50% of the population per day, which is unlikely to be
compensated by growth even under favourable conditions.
As explained later, the effect of settling on the algal community also
explains part of the pronounced differences that often exist between the
open water and dense weedbeds where turbulence is low. In the vegetation,
algal biomass is usually lower and the community is dominated by small
species that have high growth rates and low settling velocities and by flagel-
late species that can actively swim.
Resuspension will affect algal dynamics not only by bringing settled algae
back into the water column but also indirectly through its impact on the light
climate and the nutrient dynamics. These light and nutrient effects are
treated in the next sections.
Settling causes a so-called density independent mortality. Unlike density
dependent mechanisms such as predation, density independent processes
The regulation of algal biomass 85
kill a fixed proportion of the population independently of the population
density. Flushing of a lake with water that is free of algae also causes a
density independent loss, and in some aspects the effects of flushing and
sinking are therefore comparable. When 10% of the lake water with algae is
washed out every day, this implies a loss rate of 0.1 day-' for the population.
Such a flushing loss (f) can be added to the logistic equation in the same way
as the sinking loss (s/D), and by analogy the equilibrium biomass drops
linearly with f, while the requirement for survival is r > f In Danish lakes
algae disappear when f exceeds 0.3 day-' (Jeppesen, pers. comm. ). Assuming
flushing to be the main loss factor in these situatio.ns, this can thus be
directly interpreted as an indication of the maximum algal growth rate r in
these lakes. In general, density independent losses affect slow growing
species more, offering a competitive advantage to fast growing species. As
explained later this explains why flushing may cause a switch from domi-
nance by large, slowly growing cyanobacteria to other algae.
__h_,_ (6)
h,+ED
E = e,A + E, (7)
If light is the only limiting factor, algal growth can now be described as the result of
light dependent productivity and a fixed loss rate (() resulting from respiration and
mortality:
dA =rA h, lA (8)
dt h,+D(e,A+Eb)
It can be seen from this formula that algal density in equilibrium A* is predicted to
decrease with lake depth D (Fig. 3.6a).
This fits with the observations (Fig. 3.10) and with the intuitively straightforward
explanation that shallow lakes can become more turbid than deep ones, simply
because in a shallow water column higher turbidity is needed to cause enough shade
The regulation of algal biomass 87
120
100
~
!:9 80
c
0
"'
i
c
8
60
40
],
<
20
0
0
Lake depth (m)
"'120
'E 8
!:9
~ 100-
i~ 80-
........................
b 60-
=2
------------------
---
~ .............. Eb= 1
8. 40-
~
E 20-
~ ---------------------
~ 0
I I
0 6 8 10
Lake depth (m)
Fig. 3.6 Theoretically expected relationship between (mixed) lake depth and
phytoplankton biomass expressed per unit of volume (a) and expressed per unit of
lake surface area (b) with and without background turbidity.
to stop algal growth. The prediction can be made more specific if we consider algal
biomass per unit area A·,~ (g m-2). This can be done simply by multiplying the above
formula for equilibrium algal concentration A* (g m-3) with the depth D (m):
(10)
88 Phytoplankton
This formulation shows that in the absence of background turbidity (E, = 0) the
equilibrium algal biomass per square metre is predicted to be independent of the
depth of the water. If, on the other hand, there is some background turbidity, the
algal biomass per unit area declines linearly with the (mixed) depth of the lake (Fig.
3.6b ). Similar results have been obtained from elaborate models that use realistic
photosynthetic responses to light and integrations over depth and time (Straskraba,
1980) indicating that these patterns are not an artifact arising from our very simple
formulation of light limitation.
Another interesting suggestion can be derived with respect to the light on the
bottom of the lake when algae are in equilibrium. Since light at any point under
water is an exponential function of the product of depth and turbidity (Chapter 2,
Eq. 1), light at the bottom when algae are in equilibrium is therefore a function of:
(11)
Both background turbidity E, and lake depth D have disappeared. Thus, the
model suggests that light on the bottom of a lake with light-limited phytoplankton
depends neither on the depth of the lake nor on the background turbidity of the
water.
:s
..
c:
0 8
"'"'c:
...
::J
I
.
:g" 6
·~
.. ....'·
1:
Jf ~··
.:-~'
;;; 4
u
...•,L• .tJ
'E
~
0
0 2
"' 4 6 8 10 12 14 16 18
mean depth, D (m)
Fig. 3.8 Relationship between the vertical light attenuation coefficient (E) and the
mean depth (D) of lakes in the data set used for Fig. 3.7. The line represents a shade
level (ED) of 16.
_!L (13)
hp+Pr
When P1 equals h. this Monod function takes the value of 112, and will thus
reduce the growth by 50%. Free nutrients are depleted in the course of
population growth, and supplied again by processes like desorption from
particle bound forms. To describe this properly, a separate set of differential
equations for nutrient dynamics would be needed. A way to avoid that is to
assume that free phosphorus (P1) at any time is simply the total pool minus
the phosphorus present in algae:
(14)
where p, is the phosphorus content of algae. Substituting this into the
Monod formulation (Eq. 13), we obtain a function that describes nutrient
limitation directly in terms of total nutrients and algal biomass:
(15)
The regulation of algal biomass 91
Note that this short-cut necessarily implies an overestimation of the phos-
phorus concentration directly available to the algae, as it neglects the fact
that part of the total available phosphorus that is not contained in the algae
is bound to sediment particles. Therefore, this approach will tend to overes-
timate the immediately available concentration and thus the growth rate at
which the equilibrium is approached. However, since algal growth becomes
phosphorus-limited only at very low SRP concentrations (hp is small) this
problem is probably minor compared with the inaccuracies caused by other
simplifications.
The easiest way to 'combine' light and nutrient limitation is to apply the
so-called Liebig law of the minimum, stating that growth is only affected by
one factor at a time, namely tbe factor that poses the strongest limitation. A
more neutral assumption is that both nutrients and light affect productivity
at any time. This means that we should simply multiply the growth rate by
botb limiting functions:
(16)
0
Totai·P (mg 1·1)
Fig. 3.9 Theoretically expected effect of total-P concentration (P) and (mixed) lake
depth (D) on phytoplankton concentration (A). In deeper lakes light limitation
occurs already at low total-P concentrations, whereas in shallow lakes
phytoplankton concentration can reach very high levels.
92 Phytoplankton
0.5 0 1•o mean depth > 3 m
mean dePth < 3 m
0.4
.s"" 0.3
'l'
'5.
l0
0.2
:;:
u 0 0 0 0 0
.
o~• cPQ oeo
0.1 0 0 Q 0 ~
o 0o o 0 •
OO.o~oo co
~• . , 00 • •'lf' 0
0 0
0.0 0
0.5
I•Totai-P < 0.2 mg 1·1 l
o Totai-P ~ 0.2 mil: 1·1 I
0.4
l 0.3
'l'
~
o,_ 0.2 0 0
1:1
.
0
0 0 ••
:;:
u
. •i..
0 0 <E
0.1 o eo ~
8 8> 0 • ~ 0
I•••3
.~ 01 0
00 0
I o!
0.0
0 3 4 5 6
(b) mean depth (m)
Fig. 3.10 (a) Relationship between the summer (July-August) average concentra-
tion oftotal-P and chlorophyll-a in 142 Dutch lakes. In shallow lakes (open circlesD
<3m) chlorophyll concentrations can become higher than in deeper lakes (closed
circles D > 3m). (b) Relationship between the mean depth (D) and the summer
(July-August) average concentration of chlorophyll-a in the same data set. The
maximum chlorophyll concentration decreases with depth, but in lakes with a low
total-P concentration (open circles P < 0.2mg r') chlorophyll concentrations remains
limited even in shallow lakes.
The regulation of algal biomass 93
Solving this equation to find the equilibrium biomass can be done, but leads
to a very lengthy formula.
Figure 3.9 shows the predicted effect of nutrients and lake depth on algal
biomass graphically.
The model suggests that algal biomass should initially increase with the
total-P level, until a light-limited maximum value is reached that depends on
the depth of the lake. At low nutrient levels lake depth does not affect algal
biomass. At high nutrient levels, however, the concentration of algae rises
with decreasing lake depth until nutrient limitation poses a limit. Note that,
although the minimum law is not applied, regions can be distinguished
where either light or nutrients seem to be 'the' limiting factor. The transition
between these regions is smooth rather than abrupt, but in large areas one
limiting factor dominates completely. At high nutrient concentrations and
in 'deep' water, for instance, algal biomass hardly increases with nutrients.
Here, light is the dominant factor limiting algal growth.
As shown in previous sections, field data confirm that nutrients (Figs. 3.1
and 3.2) but also light (Fig. 3.8) can impose an upper limit to algal concen-
trations. A simple way to check the combined effects of nutrients and depth
suggested by the model is to classify the points in such graphs (Fig. 3.10).
This confirms that the maximum chlorophyll level increases with phos-
phorus to very high values in the shallowest lakes, while in the deeper ones
chlorophyll concentrations soon level off and become largely independent
of the phosphorus concentration (Fig. 3.10a). Similarly, the maximum chlo-
rophyll concentration increases with decreasing depth, but only when the
phosphorus concentration is high enough (Fig. 3.10b). Thus, as predicted by
the simple model, at low levels nutrients impose an upper limit on algal
biomass that does not depend upon depth while in very eutrophic lakes,
depth set an upper limit to algal biomass that does not vary with the
phosphorus concentration. Another way to visualize the combined effects of
phosphorus and lake depth on algal biomass is to fit a three-dimensional
response surface through the data points using a local interpolation tech-
nique (Fig. 3.11).
Note that this bumpy surface shows how the average chlorophyll concen-
tration changes with nutrients and lake depth and does not indicate an
upper boundary as shown in the other figures. Nonetheless, it suggests
roughly the same patterns. Note that the patterns in field data correspond
quite well with the predictions from the model (Fig. 3.9), suggesting that the
explored simple formulations and reasonings may still capture much of the
essence of the complex mechanisms involved.
As shown in the previous chapter, the contribution of resuspended sedi-
ment particles to turbidity can be very high in shallow lakes. The expected
effect of such background turbidity on the response of algal biomass to
enrichment can also be explored using the model (Fig. 3.12).
Not surprisingly, algal biomass is predicted to be reduced at high back-
ground turbidity. An analysis of data from 96 reservoirs in the Midwest of
94 Phytoplankton
~0
2.5
1
96 Phytoplankton
reduction of algal productivity will be followed relatively soon by a decrease
in detritus related background turbidity.
Fig. 3.13 The zebra mussel, Dreissena polymorpha cau massively colonize poles,
rocks, boats and other hard substratum. Filtration by these animals has cleared up
the water remarkably in some lakes in the USA where the species is a recent invader
that spreads rapidly.
usually restricted to safe sites such as plants, rocks, boats and poles (Fig.
3.13).
Allelopathic effects
Aquatic macrophytes have long been suspected of suppressing
phytoplankton through the excretion of chemical substances that inhibit
phytoplankton growth (Hutchinson, 1975). Such chemical suppression,
called allelopathy, is known to play an important role in some vegetations
98 Phytoplankton
(Rice, 1974), and seems an obvious candidate to explain the apparent con-
spicuous reduction of phytoplankton growth in densely vegetated lakes as
well. Especially Chara has been subject of many investigations. Stands of
this macrophyte are often surrounded by remarkably clear water. Also the
plants have a pungent smell, suggesting that they might excrete something
suspect.
Indeed, chemical compounds isolated from Chara and from some other
plants have been shown to inhibit the photosynthesis of natural
phytoplankton assemblages and of an epiphytic diatom in the laboratory.
(Anthoni et al., 1980; Wiurn-Andersen et al., 1982; Wium-Andersen et al.,
1987). Whether or not these effects occur in the field has been a topic of
much debate. Forsberg et al. (1990) found, for instance, that phytoplankton
biomass in a set of Chara dominated lakes was not lower than expected from
their total-P concentrations, and concluded that allelopathy apparently did
not reduce algal growth in these lakes. Clearly, to have an effect on
phytoplankton the allelopathic substances have to be released naturally in
the surrounding water and be stable enough to stay there for some time.
This appears less easy to demonstrate in practice.
More convincing than experiments with plant extracts are experiments
with intact plants or water from natural weed beds. To exclude the possibil-
ity that depletion of nutrients by the macrophytes reduces algal growth, the
water is artificially enriched in such experiments. Mostly, the observed
effects on phytoplankton growth are not spectacular. Water from Chara
dominated ponds, for instance, appeared to reduce the biomass of a culture
of green algae (Scenedesmus) by only 10-15% compared with water from
ponds with little or no vegetation (Hootsmans and Breukelaar, 1990), while
experiments with intact Ceratophyllum plants showed little effects at all on
total biomass of a natural phytoplankton assemblage (Jasser, 1995). Re-
markably, bio-assays have even demonstrated positive effects on algal
growth of (non-nutrient) exudates of early growth stages of Myriophyllum
spicatum (Godmaire and Planas, 1983).
Although overall results of aquatic allelopathy research are rather
equivocal, several studies indicate that cyanobacteria (or 'blue-green algae')
are much more sensitive to allelopathic exudates from macrophytes than
other algae. Turkmenistan workers, for instance, studying the effect of
Ceratophyllum and Myriophyllum plants on cultures of the blue-green algae
Anabaena and Anabaenopsis recorded drops in the algal density of almost
90% in the presence of the macrophytes (Kogan and Chinnova, 1972). Also,
German work shows that the submerged plant Myriophyllum spicatum
releases polyphenols into the surrounding water that can strongly suppress
the growth of cyanobacteria, but have much smaller effects on green algae
and diatoms (Gross and Siitfeld, 1994). This difference in sensitivity to
allelopathic substances suggests that exudates from macrophytes may be
important in tipping the competitive balance between cyanobacteria and
other algae. Indeed, recent Polish experiments show this (Jasser, 1995).
Competition between algae and cyanobacteria 99
Natural phytoplankton assemblages in which cyanobacteria were abundant
were incubated in the field in bags together with intact plants of
Ceratophyllum demersum. Although total algal biomass was not signifi-
cantly affected, cyanobacterial abundance was greatly reduced and green
algae became dominant. Extracts of four other plant species had largely the
same effects.
Thus, although release of allelopathic substances by aquatic macrophytes
has not been demonstrated to reduce total phytoplankton biomass mark-
edly in natural situations, it may well be an important factor preventing
cyanobacterial dominance in vegetated lakes. The mechanisms involved in
the competition between cyanobacteria and other algae are treated further
in the next section.
Allelopathic substances are not only produced by macrophytes.
Cyanobacteria of the genus Anabaena, for instance, have been shown to
release heat labile substances that repress the growth of diatoms (Keating,
1977; Keating, 1978). Also, filtrates from Anabaena cultures led to a reduc-
tion of oxygen production by the submersed plant Zanichellia peltata (Van
Vierssen and Prins, 1985).
100
..
....·.
•
80
~
..
.I
~
60
.. .
40
~
11'-
20 .. ......
0
0 2 4 6 8 10 12 14
shade (E Zl
Q
30
25 -----
10
§
~ 8
c:
""'"
::l
!:!
u..
6
0
0 50 100
% Oscillatoriaceae
('
80 80
-~ "'
1 60 )g 60
I
~ !l,;
(; 40 u 40
~ ~
20 20
0 0 ···········
0 0.1 0.2 0.3 0.4 0.5 0.6 0 0.3 0.6 0.9 1.2
Total P (mg 1'1) Zeu/Zm
------- Veluwemeer
- - Schlachtensee
•
20~----------------------------------~
16
0
• •
••
0,4-----~------~-----r----~r---~
0.0 0.2 0.4 0.6 0.8 1.0
totai-P (mg 1"1)
20
•
16
• •
N' 12 • •• o•
••
~
0
<IJ
0 0
1l
-iii 8
4
~0 00 0
0 oo
0
0.0 0.1 0.2 0.3
totai-P (mg 1"1)
Fig. 3.17 Under water shade (EZ) as a function of the total phosphorus concentra-
tion. Solid dots refer to lakes where Oscillatoriaceae comprise more than 50% of the
algal biovolume open dots are for lakes with less than 50% Oscillatoriaceae. The
upper panel represents the entire data set described in the legend to Fig. 3.14. The
lower panel represents the subset of lakes where the total phosphorus concentration
is less than 0.3mgl"1• The regression lines are computed for the subset of data
represented in the lower panel. The upper regression line is for the Osci/latoria
dominated lakes (solid dots), the lower line for the other lakes. From Scheffer et a/.
(1997a).
Competition between algae and cyanobacteria 105
Usselmeer Eemmeer
-
150
A
... ;:
79
200
B •••
.,•
oo•
.
88
J1oo 0 8 aoo~'l, o
92 90 0
~c. 0
0 oo o ss
n 87 75
84 0 81
100 .,o
93 0
e
.2
50 91
.<!
u
0 0
0 0.1 0.2 0.3 0.2 0.4 0.6 0.8
Totai-P (mg 1"1) Totai-P (mg 1"1)
Fig. 3.18 Average summer chlorophyll-a concentrations plotted against the total-P
concentration in different years for two eutrophic shallow lakes. Blue-green algae
dominated years are marked as solid dots. chlorophyll is the main light attenuating
seston component in these lakes, and shade will be higher in the years with higher
chlorophyll levels. From Scheffer et al. (1997a).
chicken and egg problem and causality cannot be inferred from correlations
alone. Other factors that vary between lakes may influence shade and
cyanobacteria in such a way that the same pattern is produced. Therefore, it
is informative to see that individual lakes that alternate between
cyanobacterial dominance and another algal community tend to show the
same pattern (Fig. 3.18).
In IJsselmeer, for instance, filamentous cyanobacteria have been rare
over the past 20 years. However, during the summers of 1976 and 1989 the
summer algal community was dominated by Planktothrix agardhii. In both
summers chlorophyll-a was exceptionally high. Another Dutch lake,
Eemmeer, is usually dominated by Planktothrix. In 1991, however, the
cyanobacterial density was low for most of the summer. This coincided with
a drop in chlorophyll-a.
Hysteresis as an implication
The patterns in the field data suggest that shade promotes Oscillatoria
dominance (Fig. 3.14), but that Oscillatoriaceae also promote shady condi-
tions (Fig. 3.17 and 3.18). This would imply an positive feedback in the
development of blooms of Oscillatoriaceae. The consequences can be seen
more clearly by setting up a simple graphical model (Fig. 3.19), based on the
empirically derived patterns.
As argued before, shade experienced by the algae in a well mixed lake
depends on the depth and the vertical light attenuation coefficient. The
model, however, describes a given hypothetical lake in which depth is fixed
106 Phytoplankton
Totai-P
Fig. 3.19 Generalized diagram of the equilibrium states of the algal community of
shallow lakes inferred from the patterns observed in the field (see text for explana-
tion). From Scheffer eta/. (1997a).
and shade therefore depends only on turbidity (E). Turbidity will increase
with the phosphorus level, starting at a background value (E.) and leveling
off at high P concentrations when light becomes limiting (Fig. 3.19 lower
curve). When cyanobacteria dominate, however, turbidity will be higher at
the same nutrient concentration (cf. Fig. 3.17). Thus a separate turbidity-
nutrient relationship should apply to Oscillatoria dominated situations (Fig.
3.19 upper curve).
The field patterns further suggest that the probability that
Oscillatoriaceae will dominate the community depends strongly on the
shade level. Since many other factors may have an effect on the probability
of cyanobacterial dominance, it would be surprising if the response to shade
would be identical in all situations. However, for a given well mixed lake,
the simplest assumption is that there is a single critical shade level, which
(because of the fixed depth) translates into a critical turbidity, E"'" above
which cyanobacteria will become dominant (Fig. 3.19 horizontal line).
Above this critical shade level Oscillatoriaceae will become dominant, at
lower turbidities other algae will dominate. This implies that below the
horizontal line the cyanobacterial turbidity-nutrient relation is irrelevant,
whereas above the critical level the turbidity-nutrient relation for other
algae is irrelevant. Neglecting these irrelevant (dashed) parts, the two
curves combine with the middle segment of the horizontal line to an S-
shaped curve of steady states that is typical of so-called catastrophic
systems. The figure suggests that at low total-P levels only the non-
cyanobacterial state is possible, whereas at very high total-P levels only the
cyanobacteria dominated state exists. However, over a range of intermedi-
Competition between algae and cyanobacteria 107
ate nutrient levels (P 1 < P < P2) both states are possible. Here, the commu-
nity will tend to settle in either of the two states depending on whether the
turbidity in the initial state is above or below the critical value (£,"',).
This graphical model implies that the response to changes in the nutrient
level should be discontinuous ('catastrophic') rather than smooth. When
starting from a low totai-P level, the nutrient loading of the lake is slowly
increased, turbidity will gradually increase too. This smooth response ends
when the critical total-P value (P2) is reached, since above this level only the
blue-green algae dominated state exists. When this 'breakpoint' is passed
the system will jump to a higher turbidity at the Oscillatoria dominated
upper branch of the diagram. If from this point the total-P concentration is
gradually reduced, the algal community will stay on the cyanobacteria domi-
nated branch until the lower critical nutrient concentration (P1) is reached,
and then jump back to the lower branch. It can also be inferred that
Oscillatoria will not easily disappear from lakes that have a high background
turbidity (£,).
The tendency of systems with alternative stable states to stay in the same
state despite changes in external conditions is called hysteresis, and that
term is also used in a broader sense to indicate that a system has alternative
stable states over a range of conditions.
Competition as an explanation
The hysteresis inferred from the field patterns can be understood from the
distinct differences in physiology between cyanobacteria and other algae.
This can be shown by extending the model for algal growth presented in the
previous section (Eq. 16). To make it a competition model we write an
equation for each of the algal types, fitting one with parameter values
measured in the laboratory for green algae (G) and the other one with
values representing the physiology of Planktothrix agardhii as a typical
representative of filamentous blue-green algae (B):
(18)
where turbidity (E) depends on the biomass of both groups and their specific light
attenuation coefficients (e,and e,):
(19)
and free available phosphorus (P1) depends on the total available pool (P) and the
concentrations of phosphorus in the two groups (p, and p,):
108 Phytoplankton
(20)
In addition to the specific loss rates due to respiration, settling losses and other
mortality causes of both groups (!, and 1,), this model has a loss rate f due to flushing
of the lake, which is the same for both species. Note that background turbidity is not
considered explicitly in this model. Its potential implications, however, can be easily
inferred from the previous analyses as pointed out in the next section.
Table 3.1 Parameter dimensions and the values used to produce Figs. 3.20-3.23 from
the model of competition between filamentous cyanobacteria and other algae (Eqs.
17 and 18, p. 107). Derivation and sources of the values can be found in Scheffer et
al. (1997a). (Note that in that paper algal abundance is expressed in grams of
phosphorus rather than biomass, and sensitivity to turbidity, q, in a 3m water column
is used rather than shade tolerance h,)
,
__~-
highest isocline wins and the system ends up in the trivial equilibrium at
which its competitor is absent. If the isoclines intersect, the properties of the
system depend on the stability of the equilibrium at the intersection. If it is
stable, and there is no other intersection, the trivial equilibria are both
unstable. Therefore, the intersection is the only stable equilibrium. Any
simulation starting with both species will end up in this state of stable co-
existence. In the case of our model (Fig. 3.20), the intersection is always an
unstable equilibrium. This can be formally proven from the inequalities, as
stated above. This specific type of unstable equilibrium is called a 'saddle'
because of the pattern produced by the paths of trajectories in its vicinity.
The saddle repels trajectories in the direction of the trivial equilibria but
attracts them from the perpendicular directions. The saddle point lays on
the 'separatrix', a line that starts from the origin and separates the attraction
basins of the two trivial equilibria which are in this case stable. Simulations
starting from an initial state above the separatrix always end up in a
monoculture of cyanobacteria (B*) whereas trajectories starting below the
separatrix lead to the other trivial equilibrium ( G*).
Obviously, the positions of the isoclines and equilibria depend on the
value of the control parameters. By playing with the nutrient level (P) and
the flushing rate (f), either of the isoclines can be brought entirely above the
other. As explained, these are situations in which only one stable (trivial)
equilibrium exists. Changing the value of one of the control variables gradu-
ally makes the saddle move in the direction of either of these equilibria.
Since the separatrix moves in the same direction the region of attraction of
that equilibrium progressively declines, until the saddle collides with the
110 Phytoplankton
i
- . Green algae (G)
Fig. 3.21 Effect of the total-P concentration of the lake on the isoclines used to
analyse the competition between cyanobacteria (c) and other phytoplankton (a)
(Fig. 3.20) (see text). Two alternative stable states exist at phosphorus concentra-
tions between P, and P,.
equilibrium and makes it unstable. In fact, the saddle moves out of the
positive quadrant to negative concentrations of cyanobacteria where it
does not make biological sense any more. The collision of the saddle
with the stable equilibrium is an example of a so-called bifurcation. Bifurca-
tions happen when, due to change of a control variable, equilibria meet
and change their nature. If, as in this case, one of the equilibria goes through
the axis (driving a species extinct) the bifurcation is called transcritical.
Bifurcations always mark a change in the qualitative properties of the
system. Therefore, tracing the parameter values for which they occur can
be a very useful technique for analysing the system's behaviour, as shown
later.
The effect of a control variable on the isoclines and equilibria can be
illustrated by adding the control variable as an extra dimension to the
isocline picture (Fig. 3.21 ). Over the range of nutrient values for which the
saddle exists (P1 < P < Pz) the two trivial equilibria (G* and B*) occur as
alternative stable states.
It is easier to grasp a two-dimensional plot in which both algal groups are
combined in one variable, and the equilibria of the model are shown as a
function of the control variable. In our case, an interesting state variable
that combines the density of both algal groups is the total turbidity (E).
Figure 3.22a shows the turbidity in equilibrium predicted by the competition
model as a function of the total nutrient concentration (P). The two
Competition between algae and cyanobacteria 111
Turbidity (E) Turbidity (E)
2.5,---------------, 2,---------------,
p =0.3
1.5
1.5
0.5
t,
0 0.05 0.1 0.15 0.2 0.25 0.15 0.16 0.17 0.18 0.19
Total phosphorus (P) Flush rate(f)
Fig. 3.22 Hysteresis shown as the response of the turbidity (E) with respect to the
control parameters total-P (P) and flush rate (f). Modified from Scheffer et a/.
(1997a).
~ 0.6
~
cyano bacteria (8*)
~
-a 0.4
~
! 0.2
o+---------.-------_,
0.1 0.15 0.2
flush rate (f)
Fig. 3.23 Bifurcation diagram of the model showing for which combinations of flush
rate (f) and total-P concentration (P) cyanobacteria [c) or other algae [a] will
dominate, and for which combinations those states are alternative equilibria. From
Scheffer et at. (1997a).
the control parameter, f, the situation now becomes less favourable for
cyanobacteria. Interestingly, the catastrophic transitions between the two
branches (at f, and / 2) do not occur at the same turbidity any more as was the
case in the hysteresis with respect to P (Fig. 3.22a). This illustrates that the
effect of flushing can not simply be explained as affecting the competitive
balance through a reduction of turbidity. Instead, the clue to understanding
the effect of flushing is that cyanobacteria have low rates of productivity (r.
< '•) but usually compensate this by having low loss rates (1. < t.). Conse-
quently, the relative impact of an extra loss due to flushing is much higher
for the slow growing blue-green algae than for the other algae. In fact,
the effect of flushing on the algal community is analogous to the much
discussed effect of 'disturbance' on terrestrial vegetations. It keeps the
slow-growing K-select species from outcompeting the fast growing r-select
species.
Depicting the combined effects of flushing (Fig. 3.22a) and nutrients (Fig.
3.22b) on the turbidity would require a three-dimensional graph. It is easier
to produce and read a projection of such a graph in the parameter plane
(Fig. 3.23).
The two lines in this graph represent the edges of the hysteresis where the
catastrophes occur ([f;,P,] and [f,P2]). Because, as explained earlier, the
corresponding connections of the stable (trivial) equilibria with the unstable
saddle equilibrium are so-called 'bifurcations', this type of representation is
known as a bifurcation graph. The cyanobacterial monoculture (B*) is
Competition between algae and cyanobacteria 113
stable above the lower bifurcation line, whereas the equilibrium without
cyanobacteria ( G*) is stable below the upper line. The area between the
lines represents the conditions under which the two states are present
as alternative equilibria. Note that Figs. 3.22 a and b correspond to a
vertical and a horizontal transect through this bifurcation diagram re-
spectively, and can be used to understand how turbidity varies with the
two parameters in this graph. The bifurcation graph adds a lot of informa-
tion to the separate hysteresis curves. It shows, for instance, that the hyster-
esis with respect to P is largest if there is some flushing, whereas the
hysteresis with respect to flushing shrinks to almost nothing at high phos-
phorus levels.
Restoration strategies, manipulating nutrients and flushing rate can be
interpreted in terms of this diagram. Obviously, they are not independent,
as reduction of the nutrient content normally requires reduction of the
concentration in the inflow. A dilemma occurs when water flows can be
diverted to flush the lake but the nutrient concentration of the available
water is high. Theory predicts that flushing with high nutrient water can
still work (Fig. 3.23) if the flushing rate is sufficiently high. This view is
supported by information from Danish lakes (Jeppesen, pers. comm. ). Blue-
green algal dominance is never observed in lakes where the hydraulic
retention time is shorter than five days, even though such lakes can have
very high nutrient concentrations. Importantly, flushing effects will be
stronger in winter when algal growth rates are low. Indeed winter flushing
with water from the nearby polder areas has probably helped to break
cyanobacterial dominance in Veluwemeer (The Netherlands) (Hosper,
1985; Hosper and Meijer, 1986). Unfortunately, causality is complicated in
this case because the flushing also caused a strong reduction in phosphorus
concentration.
The above treatment of hysteresis is phrased in terms of dynamical
systems theory. However, studying competition has a long tradition in ecol-
ogy. A short note on the connection to existing theory and terminology is
therefore appropriate. Importantly, our alternative equilibria or 'unstable
coexistence' case is also one of the qualitative possibilities in standard
Volterra competition models based on the logistic growth equation. In such
models the mechanisms of competition are not specified. Instead, the inten-
sity of inter- and intra-specific competition is defined directly. For alterna-
tive equilibria to occur, inter-specific competition needs to be stronger than
intra-specific competition; loosely phrased; it should be better to have
conspecifics around than individuals of the other species. Obviously, this
is in line with the mechanism in our more specific model. Cyanobacteria
have a competitive advantage in the turbid situation caused by their own
dominance.
Our general result is also analogous to that obtained from the graphical
models of competition for two resources elaborated by Tilman (Taylor and
Williams, 1975; Tilman, 1977; Tilman, 1982; Tilman, 1985). Tilman's re-
114 Phytoplankton
source-ratio theory predicts that the coexistence between two competing
species is unstable if each species consumes relatively more of the resource
of which it also tolerates the lowest levels. In our case competition is for
light and nutrients. Blue-green algae cause a higher turbidity at the same
nutrient level. In resource-ratio terms, this means that they use relatively
more light. Since they are also the most shade tolerant group, this fits with
the resource-ratio requirement for unstable coexistence.
a b
c d
Fig. 3.24 Some interaction structures for which simple models have been demon-
strated to produce chaotic behaviour: (a) a consumer exploiting two competing
preys, (b) two consumers exploiting prey that compete, (c) a carnivore on top of a
simple consumer-food system, (d) a network of at least four competing species.
From Scheffer (1991b).
0 500 1000
days------>-
carnivores
herbivores
autotrophs
Fig. 4.1 Top-down control in food chains of different lengths as suggested by the
HSS hypothesis, stating that top-down control of primary producers occurs only in
food chains with an odd number of links (see text). From Scheffer (1997).
well known, in two categories. One category admits the usefulness of birds
since these destroy insects which damage crops, and believes that by promot-
ing increase of bird numbers the number of insects and the extent of the
damage they do could be reduced. The other category, on the other hand,
believes that the effect of birds is of little importance concerning the destruc-
tion of insects harmful to crops, and that the development of birds would not
prevent the development of insects.
Naturalists belonging to the first category reason this way: the number of
insects which cause damage to crops increases; that of birds, on the other
hand, decreases. Now, birds feed to a great extent on insects; so if we increase
the numbers of birds, the number of insects will decrease. The second cat-
egory of naturalists think differently: the number of birds is high particularly
in those places where insects are very abundant When the number of insects
decreases, so does the number of birds. Regions with low insect abundance
also have few birds. The amount of insects in a region depends essentially on
the amount of food found in it[. . .]. Hence, they conclude: birds play only a
small effect in destroying insects which may damage crops. Well-known
naturalists have argued in favor of either one of the theories mentioned.
However, the number of naturalists who support the first theory is decreasing
every day, while those in favor of the second one increase
Camerano also brings up several familiar sounding explanations for the fact
that no clear insights on this topic had been obtained yet, such as the fact
that applied science has been too sloppy ('an inclination to hasten to appli-
cations while disregarding data from pure science') and neglecting impor-
124 Trophic cascades
tant feedbacks in the food web ('without taking into account the many and
very important relations among different groups of animals'). He proceeds,
presenting a theoretical framework for understanding food-chain dynamics
that contains many of the key concepts of later ecological theory, for in-
stance, the idea that consumer and food populations are in dynamic equilib-
rium: 'It is a well accepted fact by all that animals and plants develop in
direct proportion to the available food. From this it follows that no species,
be it carnivore or herbivore, can develop beyond a certain limit which, if
surpassed, would destroy the source of its own nourishment. Equilibrium
broken by the excessive growth of either kind of animal, would again be
reestablished.' Camerano explains in detail how the effect of disturbances
on one trophic level will cascade through the food chain, the same idea that
would provoke so much debate almost a century later (Hairston et al., 1960),
and describes how equilibrium will tend to be re-established through
damped oscillations.
Camerano's work has been rediscovered only recently (by Jacobi and
Cohen, see Camerano's reference) and apparently his systems ecology avant
Ia lettre has not been appealing enough to the scientists of his days to create
a school that kept the ideas alive. Much more influential were the simple
mathematical models of species interactions presented about half a century
later by Alfred Latka (1925) and by Vito Volterra (1926). Ever since their
contribution, these and later, more realistic minimal models have catalysed
the understanding of the dynamics resulting from trophic interactions. In-
deed, the dynamic results of 'eating and being eaten' are often rather intri-
cate and in many cases simple models have provided the little push needed
to grasp the full implications of consumer - food interactions intuitively. In
this chapter such minimal models are used to explain some basic features of
the interactions between phytoplankton, zooplankton and fish. The result-
ing insights serve as a basis from which more complex aspects of trophic
relations such as size selective predation and predator avoidance strategies
are discussed.
140
~ 120
c.
~
~
100
"":<:
~ 80
60
40
20
Fig. 4.2 Changes in Secchi-depth over a 1-year period in experimental ponds with
(treated) and without (reference) zebra mussels. From Reeders eta!. (1992)
less likely to affect transparency throughout the water column. The increase
in Secchi-depth associated with massive Dreissena populations is about a
factor 2 in the cited cases. Although zooplankton grazing can cause more
spectacular increases in transparency, the effect of zebra mussel grazing is
less ephemeral than that of the often short-lived outbreaks of Daphnia.
Since, in addition, zebra mussels can filter out particles of a much larger size
range than Daphnia, research has been done to find ways of stimulating
zebra mussel populations as a possible aid in reducing turbidity of shallow
lakes (Reeders and Bij de Vaate, 1990). In practice, it appears that lack of
suitable hard substrate prevents the species from becoming abundant in
many of the European turbid shallow lakes. The continuous resuspension
and sedimentation in unvegetated shallow lakes buries the mussels except
on safe sites like stones, poles, trunks or boats. When lakes clear up due
to mussel grazing, aquatic vegetation can expand (Griffiths, 1992). Since
macrophytes are a suitable substratum for settlement of juvenile mussels
and help to prevent wave resuspension of sediments, this may further im-
prove expansion possibilities for Dreissena. The resulting positive feedback
may boost the changes observed after Dreissena invasions in shallow parts
of American lakes (Macisaac, 1996).
Top-down control of phytoplankton 127
Several molluscivorous fish species can feed on zebra mussels (French,
1993) and when the water is not too deep, diving waterfowl are known to
consume considerable quantities of mussels in some cases (Wormington and
Leach, 1992; Hamilton et al., 1994). Nonetheless, strong top-down control of
Dreissena populations has not been reported. Lack of suitable substratum
for settlement is probably a main restriction of Dreissena expansion in many
shallow lakes.
Fig. 4.3 Waterfleas (Daphnia) are small planktonic crustaceans that filter the water
to obtain their algal food. They can reach densities of hundreds of animals per litre
and reduce phytoplankton to low levels. However, waterfleas are also very vulner-
able to fish predation and this explains their absence in many situations.
However, probably the most important weak point in the success formula
of Daphnia is that they are very profitable food for planktivorous fish. This
leads to their absence in many situations as explained further in the next
section. Many studies illustrate the strong impact of Daphnia on algal
biomass when they are released from fish predation. Sometimes long ice-
cover in winter leads to massive fish-kills due to lack of oxygen. Such natural
winter kills are typically followed by the occurrence of dense Daphnia
populations that graze down algal biomass (Schindler and Comita, 1972;
Haertel and Jongsma, 1982; Samelle, 1993). The same is observed after
artificial strong reductions of the fish stock by means of fishing or rotenone
Top-down control of phytoplankton 129
treatments (Shapiro and Wright, 1984; Van Donk et al., 1990; Meijer et al.,
1994a).
The potential of Daphnia to graze down algal biomass to very low levels
is also illustrated by the spring clear-water phase that occurs in many lakes
(Lampert et al., 1986; Luecke et al., 1990; Carpenter et al., 1993; Rudstam et
al., 1993; Samelle, 1993; Hanson and Butler, 1994a; Townsend et al., 1994;
Jurgens and Stolpe, 1995). The details of this phenomenon are treated more
thoroughly later, but the general scenario is easy to understand. The spring
bloom of algae provides a wealth of food for Daphnia, allowing a high
individual growth rate and reproduction. The population expands in a
couple of weeks to a density at which its grazing capacity exceeds the
algal production. As a result, the algal community collapses to a low level.
During this phase the water can be spectacularly clear. This state of over-
exploitation of algae does not last long. The condition of the Daphnia
individuals deteriorates due to food shortage. The number of eggs per
female decreases and reproduction practically stops. Finally the Daphnia
population collapses and the algal community recovers. Obviously, this is a
classic predator - prey cycle scenario, and indeed in laboratory populations
such cycles tend to produce a steady oscillation (Pratt, 1943). Also, in the
field a regular sequence of cycles is sometimes observed in the course of the
summer, but often, young-of-the-year planktivorous fish will prohibit a next
Daphnia outbreak after the spring peak.
In the following, a simple classic model of the Daphnia - algae re-
lationship is presented and it is shown how its dynamics are affected by
factors like the nutrient level and spatial heterogeneity. This minimal model
will serve as a basis to explore the effects of fish and seasonality in later
parts.
dAdt =rA(1-i!.)-g
K
z~
' A+h.
(1)
130 Trophic cascades
dZ A
-=egZ---mZ (2)
dt ' ' A+h. '
The basic growth of algae (A) is logistic, but an extra term is added to
account for the consumption by Daphnia. This consumption depends on the
amount of Daphnia (Z) and its maximum consumption rate (g,) and on the
phytoplankton density. The latter dependence is formulated as a Monod
function representing a simple saturating functional response with a fixed
half-saturation value (h,). The zooplankton population converts the in-
gested food into growth with a certain efficiency (e,) and suffers losses due
to respiration and mortality at a fixed rate (m,).
This model or a similar one can be found in most ecological textbooks.
Such simple predator-prey models have a long history of analysis and
consequently much is known about their behaviour (Rosenzweig and
MacArthur, 1963; Rosenzweig, 1971; DeAngelis et al., 1975; Murdoch and
Oaten, 1975; Scheffer, 1991a). The traditional way to explore the properties
of these minimal models is through analysis of the zero-isoclines. The for-
mula of these isoclines is obtained simply by solving the above equations for
zero growth (dA/dt = 0 and dZ!dt = 0).
The resulting formula for the algal isocline is:
Z* = rA(1-~) A+ h. (4)
K g,A
The first part is really the productivity curve of the logistically growing algae
while the second part is the inverse of the functional response of
zooplankton. The height of the algal isocline (dA/dt = 0 in Fig. 4.4) at any
point can be interpreted as the amount of zooplankton needed to consume
exactly the production of the algae at that density, thus balancing their
population growth to zero.
The logistic growth causes the isocline to be humped: at low algal
densities the total productivity of the population is low and little grazing
is needed to balance it, whereas at high algal densities productivity
drops again because of competition and again little grazing is needed to
keep it from growing further. The (inverse) functional response causes
the isocline to be asymmetrical; the left-hand side is higher than the right-
hand side. This is because more zooplankton can be tolerated at low algal
densities when zooplankton can not gather food as effectively as at high
algal densities.
The isocline of zooplankton (dZ!dt = 0 in Fig. 4.4) is simply vertical, as Z
is eliminated from the equation:
A*=--h_._ (5)
e,g, -1
m,
Top-down control of phytoplankton 131
A* K
Algae(A)
In biological terms, the reason for the vertical isocline is that there is no
negative feedback of high population densities other than via the food in the
model. Thus there is simply one fixed food density (A*) at which the popu-
lation gains just outbalance the losses.
The assumption that the consumers affect each other only through
depleting food is probably quite realistic in the case of Daphnia (Slobodkin,
1954), but certainly not for all consumers. Consumer interference
(mostly termed 'predator interference') can either be incorporated by
modelling direct interactions between the animals (Rosenzweig, 1971;
Gilpin, 1972) or by using a predator-dependent functional response
(Reddington, 1975; DeAngelis eta/., 1975; Ruxton et al., 1992; Abrams
and Roth, 1994). A special case of predator dependence is the ratio
dependent functional response (Arditi and Ginzburg, 1989). Although
this formulation captures the essence of predator dependence in a simple
way, its use has some serious theoretical problems (Abrams and Roth,
1994).
132 Trophic cascades
As explained in the previous chapter the isoclines of zero-growth divide
the 'phase plane' (Fig. 4.4) into regions with positive and regions with
negative growth of the populations. In this case algal growth is negative
above the algal isocline and positive below it. Similarly, zooplankton growth
is negative to the left of the vertical isocline and positive on the right-hand
side. Intersections of isoclines are equilibria as growth of both populations
is zero. Since on the axis the density of either of the populations is zero and
thus its growth rate is zero, the axes are also (trivial) isoclines. Conse-
quently, the origin and part of the intersections of isoclines with the axes are
(trivial) equilibria.
In this case the intersection of the algal isocline with the x-axis is such a
trivial equilibrium. Since zooplankton is absent, the logistic equation alone
determines algal biomass and the algae equilibrate at carrying capacity (K).
This trivial equilibrium is unstable. The system will not return to it if we add
a little bit of zooplankton. Note that unlike in the competition model dis-
Algae (A)
A* Nutrients (K)
Algae,C(AA)l-----------~'::_',;
Fig. 4. 7
Effect of enrichment (increasing K) on the equilibrium and limit cycle of
the zooplankton-algae model. In the Hopf bifurcation point (H) the equilibrium
point becomes unstable and the limit cycle evolves around it (see text for full
explanation).
biomass
I
,,----------- ...... ... , ,
I '
: Algae
I
I
I
I
I
I
I
I
I
I
I
I
Fig. 4.8On the limit cycle of the classical minimal model extreme oscillations of
zooplankton and algal populations occur.
Top-down control of phytoplankton 137
cycles pass very close along both axes of the phase plane (Fig. 4.5), cor-
responding to periods during which zooplankton or phytoplankton reach
densities that are close to zero. Although, natural populations and labora-
tory populations of Daphnia do tend to oscillate, their cycles typically have
a much smaller amplitude. Also, the period of the model oscillations
produced by the model is unrealistically large, almost half a year as opposed
to 20-45 days in real populations (McCauley and Murdoch, 1987). The
problems of the large cycle and the low frequency are in fact closely related.
The episodes in which either of the population densities becomes almost
zero stretch the cycle period (Fig. 4.8) because recovery from the near
extinctions is very slow. In the following sections it is shown how the pres-
ence of alternative food, inedible algae, spatial heterogeneity and fish
modify the above patterns to produce a more realistic view of Daphnia
dynamics.
Stabilizing mechanisms
In the ecological literature about the mechanisms that can stabilize predator
- prey relations, spatial heterogeneity and prey switching are probably the
two most abundant topics. In deep lakes Daphnia does not really have
the option of switching to different food, as it has to rely solely on the
seston in the epilimnion. In vegetated shallow lakes, however, there are
indications that Daphnia feeds on detritus accumulated at the bottom
when phytoplankton densities are low (Jeppesen et al., 1996). Although
this is probably a relatively low-quality food source it may keep the
populations from collapsing completely, thus stabilizing the system.
Unfortunately, little information is available on the diet selection of
Daphnia in the field, but the idea that the presence of detritus as an
alternative food source should stabilize Daphnia populations seems
reasonable.
Another potentially stabilizing factor is the presence of inedible algae
such as large cyanobacterial colonies. Obviously, unravelling the subtleties
of the interaction of competing algal groups with grazers is rather compli-
cated. However, put simply, there are two reasons why the presence of
inedible algae may be stabilizing: (1) they compete for resources with the
edible algae, thus lowering the 'effective carrying capacity', and (2) inedible
species reduce the efficiency with which zooplankton can consume the
edible algae (Giiwicz and Lampert, 1990). In terms of isoclines, as explained
earlier (1) moves the top of the humped algal isocline to the left, while (2)
moves the vertical zooplankton isocline to the right. Since the oscillations
disappear when the top in the algal isocline is on the left side of the
zooplankton isocline (1) and (2) should thus be expected to stabilize
Daphnia dynamics. The potentially stabilizing effect of inedible algae has
been demonstrated with simple models (Kretzschmar et al., 1993; Gragnani,
1997), but has not been shown experimentally yet.
138 Trophic cascades
One of the most frequently mentioned topics in the literature about
stabilization of predator - prey dynamics is spatial heterogeneity. Using
models many authors have shown a stabilizing effect of partial isolation
of habitat patches (Gurney and Nisbet, 1978; Nisbet et al., 1989). Other
models have been formulated to show that predator - prey oscillations
are stabilized when the predators aggregate in patches with high prey
density (Hassel and May, 1974). Stabilization can also be achieved by limit-
ing the speed of movement of individuals in individual-based predator -
prey models (De Roos et al., 1991). All of these mechanisms are in fact
closely related. The space outside the patches where the predator is con-
centrated can be considered a 'partial refuge' where part of the prey popu-
lation can escape predation. As explained in the next section, spatial
heterogeneity is likely to be an important reason why the extreme model
oscillations with near extinctions of both algae and Daphnia are not found
in the field.
Fig. 4.9 Simple spatial structure assumed in the spatial version of the zooplankton
algae model (Eqs. 5.3, 5.4 and 5.5). Zooplankton (Z) is confined to one part of the
space. Their algal food grows both inside (A 1) and outside (A 2) the zooplankton
compartment, and diffuses (d) between both parts of space. From Scheffer and De
Boer (1995).
dAdt 2 = rA 2 (1- A
K
2 )- -..!!:__(A
1-q
2- A,) (7)
dZ =e g Z~-m Z (8)
dt ' ' A 1 +h. '
The combined effect of the mixing rate (d) and the fraction of the lake
occupied by Daphnia (q) on the dynamics of the model can be summarized
in a bifurcation diagram (Fig. 4.10).
Since this does not show what the oscillations look like there is a separate
plot displaying the corresponding dynamical patterns (Fig. 4.11).
still oscillates, but the amplitude of the cycles is reduced, and the period
falls nicely in the range reported from the field. The algae in the ungrazed
part now show a mild oscillation too, driven by an exchange with the grazed
part.
Top-down control of phytoplankton 141
Algae outside (A 2)
a biomass (mg 1" 1) 10 ·····:;-*"---------::·::·.:.·::·:···········································?·--
' ' I
1 \Algae 1nside (A 1)
'
''
'
' ..
.
'
I'' ''
I'
''
'
50 100 150 200 Algae outside <A:2l
time {days) Algae inside (A1 )
Zooplankton (Z)
/
' '1 Algae ins1de (A1)
I Zooplankton (Z)
Fig. 4.11 Time plots showing the behaviour of the spatial plankton model for the
parameter settings indicated in Fig. 4.10. Parameters: q ~ 0.5 and d ~ 0.0, 0.02, and 0.4
for panels a, band c, respectively. From Scheffer and De Boer (1995).
142 Trophic cascades
ingly alike, and the density of Daphnia in its refuge becomes very high due
to the inflow of food from the rest of the lake.
Biologists usually describe the dynamics of populations by sampling at
various points at a lake, and averaging the samples. In the model we can do
the same by averaging the densities of algae (Ar) and zooplankton (Zr) over
the total volume:
(9)
5 0.5
+-------.-----~,------,-------+0
50 100 150 200
time (days)
Fig. 4.12 Time plot showing the dynamics of the concentrations of zooplankton and
algae averaged over the total volume (Zrand Ar) generated by the spatial model for
a small value of the exchange rate and the grazed fraction, i.e. d = 0.001 and q = 0.1.
Zooplankton oscillates while their algal food stays practically constant. From
Scheffer and De Boer (1995).
Top-down control of phytoplankton 143
Total algae CA,-l
a 10.--------------------------.
--
_;:.0.:~--~-=-~-:-~:-~-~::.7.::.~......
0 v·· f.,0.6 ·············•·•··•············•·
0 5 10
nutrients (K)
Total zooplankton (Zy)
b 10.-------------------------,
5 10
nutrients (K)
dA
dt
=rA(1-~)-g
K
Z~+i(K-A)
' A+h.
(10)
Note that to simplify further, the single parameter i now replaces the more
explicit ratio d/q to represent the rate at which ungrazed algae enter the
studied volume. Since the two-compartment model is a rather abstract
representation of the situation in any real lake the value of the parameters
d and q will be hard to characterize in practice. Thus, although these param-
eters were useful to clarify some general principles, their preservation is not
of much importance for further analyses.
If the mixing rate is low the above formulation yields a reasonable ap-
proximation of the dynamics in the zooplankton occupied compartment of
the spatial model. (Note that, because the model produces only these local
Top-down control of phytoplankton 145
dynamics, the average over a Jake, including the parts without Daphnia,
should be computed as explained earlier, assuming algae to be at carrying
capacity in the remaining space.) The model can be simplified further by
replacing the entire last term, i(K-A), by a single constant inflow of algae
(mg r' d- 1). This works quite well in most situations (Scheffer and De Boer,
1995). However, since in the following sections K is varied with enrichment
and change of seasons, it is more appropriate to keep the full term, i(K-A),
to make it vary in concert with carrying capacity.
The isocline picture of the new model differs from the classical one in that
the algal isocline rises sharply at low algal densities (Fig. 4.14).
A small inflow term (i) suffices to get rid of the unnaturally large limit
cycles with near extinction periods. Increase of i causes the range over which
the algal isocline has a positive slope to decrease until the complete isocline
has a negative slope (Fig. 4.14). Since oscillations occur only when the algal
isocline has a positive slope at the intersection, it can be seen from change
in the algal isocline that increasing i tends to stabilize the model.
A very similar isocline picture arises if a sigmoidal functional response, or
Holling type Ill, instead of a simple saturating one is used. Indeed the use of
Algae (A)
Fig. 4.14 Zero growth isoclines of zooplankton (vertical) and algae (curved) of a
simplified version of the spatial model for increasing values of the parameter i which
represents import of algae from an ungrazed part of space (Eq. 10, p. 144). Open
circles denote unstable equilibria~ closed circles represent stable equilibrium points.
146 Trophic cascades
a sigmoidal functional response is an effective way to stabilize predator -
prey models. A problem with the use of sigmoidal functional responses
for Daphnia is that they may be inferred from the presence of an alternative
food source like bottom detritus, but are never really measured. Even if
there is prey switching, though, there is a conceptual problem of modelling
it through the use of a sigmoidal response, as in the above model. From
the point of view of the prey it describes the situation well, but for the
predator does not. Put simply, Daphnia stops eating the algae when they
are rare, but does not get the alternative food to which it is supposed to
switch.
!
,!J!
0 ...-"
l ,.____ without fish
~~ ~with fish
-
Chlorophyll-a
threshold
• •
•
•
••
•
•
age-a yellow perch
Fig. 4.16 Relationship between the density of Daphnia pulex and the number of age-
D yellow perch on the first of August in Oneida Lake over a period of 15 years.
Redrawn from Mills et al. (1987)
I II Ill IV
I->
I
~ .
I
consumption
~I
''
\production
''
/ i
K
food biomass
Fig. 4.17 Graphical analysis of the stability of an exploited food population. At the
intersections of tbe two curves consumption equals production and the food popula-
tion is in equilibrium. However the equilibrium on the intersection marked by an
open circle is unstable {for further explanation see text).
150 Trophic cascades
consumer density
'•,
''
''
'' \
''
'
''
'
'
food biomass
Fig. 4.18 Consumption increases with consumer density and this affects the position
of the stable and unstable equilibrium points at the intersections of the production
and the consumption curves (see also Fig. 4.17).
For high food densities it saturates as the maximum consumption rate of the
consumer individuals is approached. Assuming that the food population can
never be eaten completely because there is always a part unreachable for
the consumers, the consumption starts at a food level slightly higher than
zero.
If the functional response saturates at sufficiently low food densities the
curves can intersect in three points as in the illustrated case. Obviously, the
food population will increase if production is higher than the consumption
losses (sections I and III) and decrease if consumption exceeds production
(sections II and IV). All three intersections are equilibria, as consumption
balances growth. The middle one, however, is unstable. This is because after
a small disturbance, the system will move further away from it rather than
returning as in the case of the other two intersection points. The unstable
point represents the breakpoint food density below which the system col-
lapses into an over-exploited state with low food biomass where production
is very low.
Since the overall consumption increases with the amount of consumers
present, the height of the consumption curve will increase in proportion to
the consumer density (Fig. 4.18).
If one tracks the shift in the equilibria with changing consumer density, it
can be seen that this configuration implies a hysteresis much like the one in
the competition between cyanobacteria and other algae, but in this case due
to a completely different mechanism. Starting from the lowest consumer
density there is just one equilibrium. Increasing consumer density, this
equilibrium moves to the left. Food density decreases but productivity in-
creases until the consumption curve becomes too high to intersect with the
The effect of planktivorous fish 151
equilibriumr-----------------------,
food
biomass
,':'
o'"'···· l
r
1 over exploited state
consumer density
Fig. 4.19 Food density in equilibrium plotted as a function of consumer density. The
dashed middle section of the curve corresponds to the unstable breakpoints of the
system (open circles in Figs. 4.17 and 4.18). In the range of consumer densities over
which this unstable equilibrium exists the system tends to either of the two alterna-
tive stable equilibria depending on the initial density of the food population relative
to the breakpoint (see text for further discussion).
production curve. At that point the equilibrium hits the unstable breakpoint
and disappears. As a result, the system collapses into the overexploited
state. If after this collapse the consumer density is reduced in order to
restore the productive state, the system shows hysteresis. It stays in the over-
exploited equilibrium with low food densities until the consumption curve
has become low enough to let the intersections at the left side disappear.
Again this happens when the breakpoint collides with the stable state.
Plotting the position of the three equilibria against consumer density (Fig.
4.19) one obtains a hysteresis curve that is analogous in interpretation to the
ones obtained from the cyanobacteria model (Figs. 3.19 and 3.22) although
the processes involved are very different.
Noy-Meir used this model to explain the effect of overgrazing by cattle
that is often observed in arid vegetations. There is, however, in principle no
reason why it should not apply to the effect of increasing fish densities on
Daphnia as well. Indeed, the response of Daphnia to fish seems to be non-
linear (e.g. Fig. 4.16) and, as explained later, a collapse into an overexploited
state is likely to be the underlying reason. However, the situation is obvi-
ously more complex here because Daphnia dynamics depend also on their
interaction with phytoplankton. In fact, the Daphnia- phytoplankton cycles
themselves are oscillations between over-exploitation and under-exploita-
tion of the algae. To understand better how the effect of fish cascades down
to phytoplankton, we need to consider the dynamics of the planktonic
system explicitly.
152 Trophic cascades
A minimal model of planktivory
To explore the potential impact of fish on the dynamical interaction of
Daphnia and algae we slightly modify the zooplankton- algae model devel-
oped in the previous section. The algal equation remains the same, as
fish generally do not eat phytoplankton. To account for the effect of preda-
tion by fish on Daphnia we simply add an extra loss term to their growth
equation:
dA
dt
=rA(l-~)-g
K
Z~+i(K-A)
' A+h.
The new loss term in the zooplankton equation represents the impact of the
fish community as a whole. In reality, different groups of fish forage on
Daphnia with different functional responses. Therefore, this term is really
just a pragmatic solution to mimic the effect of many different animals
switching to forage on Daphnia at different moments with different
efficiencies. Since most of the larger individuals usually switch to Daphnia
only when it is not too scarce (Lammens, 1985; Lammens et al., 1985), the
predation pressure is likely to increase more than linearly with Daphnia
density over part of the range. Therefore, the overall functional response is
made sigmoidal by adding a square to the formulation. The maximum
consumption rate ( G1) is set directly, rather than as the product of the fish
biomass and their weight specific maximum intake. The latter is not easily
defined for a whole community, since large animals consume less per gram
of body weight than small ones.
Note that fish is not modelled dynamically. The effect of varying fish
predation pressure over the year will be shown later, but even then,
fish growth will not be modelled as a function of Daphnia consumption.
This is reasonable, since, as argued, Daphnia represents only a small
part of the diet of most fish. Therefore, overall fish density depends
on the productivity of the lake, but does not react directly to Daphnia
density.
Algae (A)
Fig. 4.20 Effect of increasing fish predation pressure (Gt) on the zero-growth iso-
clines of the plankton model. The zooplankton isocline bends out with increasing fish
density while the algal isocline remains unaffected. Dots denote stable equilibria,
open circles denote unstable ones. The half open circles and corresponding isoclines
are associated with fold bifurcations. From Scheffer et at. (1997b ).
This causes the intersection (that represents the unstable focus of the
limit cycle) to move to the right. When it is pushed far enough to the right
it can become stable, as discussed in the paradox of enrichment section. In
this case the corresponding Hopf bifurcation is close to the top of the hump,
but not exactly on it, as that conjunction is only valid for vertical predator
isoclines. In addition to the shift in the existing equilibrium, new intersec-
tions, and thus new equilibria, with high algal biomass and very little
zooplankton arise at the lower right-hand part of the graph for sufficiently
high fish densities.
A more complete view is obtained by adding fish as an extra dimension to
the isocline picture (Fig. 4.21 ). This shows how the position of the intersec-
tions changes smoothly with fish predation to form one continuous equilib-
rium curve. Projected in the fish - algae plane or the fish - zooplankton
plane this curve shows an S-shape (Fig. 4.22) similar to the one indicating
hysteresis in the simple Noy-Meir model (Fig. 4.19) and in the competition
model of algae and cyanobacteria (Fig. 3.22).
At low fish predation, zooplankton density is high and algal density
low, whereas at high fish predation, a single equilibrium with high algal
biomass and very little zooplankton occurs. These contrasting states
co-exist as alternative equilibria over a range of intermediate fish densities
(F1 < G1 < F2 ).
154 Trophic cascades
Zooplankton (Z)
/
iLo
dt
Fish (G~
Fig. 4.21 Three dimensional representation of the effect of fish on the isoclines of
the plankton model (see Fig. 4.20). The isoclines have now become zero-growth
surfaces of zooplankton and algae. Note that the intersection representing the sys-
tems equilibria has a sigmoidal shape. From Scheffer eta/. (1997b ).
----~
----------- i' ,~
:-
"'
" ""
" ""
,I " " "
~
F,
Algae (A)
::~ -- -
............
........
:'' ''
\
I )
l ~,''[ l
------r--- ii
Fish (Gp
lmp6cations of osciUations
Since isoclines do not tell much about the limit cycle, we look directly at the effect
of fish on the equilibria to obtain an overview of the implications of the oscillatory
equilibrium on the hysteresis (Fig. 4.23), using snapshots at fixed fish densities to
clarify the link to the change in isoclines (Fig. 4.24). For zero fish, we simply have the
original zooplankton- algae model and the limit cycle is the only stable equilibrium.
Increasing fish, we first arrive at the lower left inflection point of the hysteresis curve.
Because it occurs at the folding point of the hysteresis curve, this is called a fold
bifurcation (F1 ). If we look at the isoclines, we can see that two intersections arise
after the curved zooplankton isocline hits the algal isocline (Fig. 4.24a). Because one
is a stable 'node' equilibrium and the other one an unstable 'saddle', this bifurcation
is sometimes called saddle-node bifurcation. The stable point is an equilibrium with
almost no zooplankton and a high algal density, close to carrying capacity. Note that
Zooplankton (Z)
Fish (G1)
Fig. 4.23 Schematic representation of the effect of fish predation pressure on the
equilibria and limit-cycles of the system. The limit cycle does not exist between
the two homoclinic bifurcations ( 0 1 and 0 2). In these bifurcations the stability ofthe
limit cycle is destroyed due to the collision with the saddle that marks the border of
the basin of attraction of the stable algal dominated equilibrium. From Scheffer et aL
(1997b).
The effect of planktivorous fish 157
IHomoclinic o,j
(D Zooplankton (Z)
Fig. 4.24 Trajectories and isocline configurations corresponding to the first (a) and
second (d) fold bifurcations, the first homoclinic (b) and a fish density in the range
between the two homoclinics (c). From Scheffer et al. (1997b).
Zooplankton (Z)
Algae (A)
Fig. 4.25 Six qualitatively different diagrams showing the equilibria and cycles of the
system that can be obtained by manipulating the carrying capacity for algae (K) and
the stabilizing diffusive inflow of algae (i). From Scheffer et al. (1997b ).
sufficiently small to avoid touching the saddle. In that case homoclinic bifurcations
do not occur (Fig. 4.25d). The most simple situations arise if there are no oscillations
(Fig. 4.25e) and eventually no folds (Fig. 4.25f). In the latter case there are no
bifurcations at all, and the system responds smoothly to an increase in fish, shifting
160 Trophic cascades
@ Nutrients (K)
15~-----..--------------------.
F1 01
---------- ---------------- ---------------- -- -Fig. 4.25 b
12 BT
Fig. 4.23
9
Fo 2 02
6 -------- -------- ---------------- ----------- -Fig. 4.25 c
Fo1
------- - - K-------------------------------- -Fig. 4.25 d
H 2
3
F*
0+-----.-----.-----.-----.---~
0 0.07 0.14 0.21 0.28 0.35
Fish (G 1)
@ Nutrients (K)
15~-.--~--------.-------,----.
12
9
---------------- ----------- - Fig. 4.25 a
Fig. 4.25 e
3
F*
Fig. 4.25 f
0+-----~-------.------,-----~
0 0.2 0.4 0.6 0.8
Fish (G1)
The effect of planktivorous fish 161
from a stationary regime with predominantly Daphnia to a regime with mainly algae.
These simple cases with no or small oscillations are obviously associated with rela-
tively stable basic Daphnia algae systems. As argued in the previous chapter, such
stability may be promoted by aggregated spatial distributions of Daphnia, low nutri-
ents, the availability of detritus as an alternative food source, and the presence of
inedible algae.
Although this inventory gives an overview of what might happen if the conditions
change, the discussion with respect to the parameters remains ad hoc. A more
systematic image of the effect of various parameters on the system's mode of behav-
iour can be obtained by drawing the lines at which the bifurcations occur in a two-
parameter plane. We have already used this bifurcation analysis technique to map
the behaviour of the cyanobacteria competition model (Fig. 3.23) and the stabilizing
effect of spatial aggregation of Daphnia (Fig. 4.10). Here we apply it to explore how
change in the nutrient level (expressed through K) influences the effect of fish on the
system (Fig. 4.26).
At zero fish predation the system reduces to the original Daphnia -
phytoplankton model, which has only a Hopf bifurcation (H). This bifurcation stays
present if fish is increased until it meets the second fold (F2 ) and homoclinic ( 0 2 )
bifurcations in a so-called Bogdanov - Takens point (BT). Beyond this point the
bifurcations are no longer relevant as they are not associated with attractors. Such
points in parameter space where different bifurcations meet are called codimension-
two points. Note that those represent a higher hierarchical level than the simple
bifurcations of equilibria. Our model has three other codimension-two points: a cusp
point for low values of K and G1 at which the hysteresis is born and two fold
bifurcations emerge (F*), and two points at which the fold and homoclinic curves
merge (F01 and F02 ). As indicated in the figure, the scenarios of change in equilibria
and cycles with fish density (Fig. 4.25), represent horizontal sections through the
bifurcation graph.
Obviously, the configuration of the bifurcation lines in the two parameter plane
(Fig. 4.26a) depends on the values of the other parameters. A parameter that is really
related to the environment rather than to the physiology of the species is i, used to
Fig. 4.26 Bifurcation diagrams showing when the Hopf (H), fold (F) and homoclinic
( 0) bifurcations occur depending on the fish predation ( G1) and the carrying capac-
ity for algae (K). The horizontal sections that are indicated correspond to the
diagrams in Figs. 4.23 and 4.25 as indicated. a. The Hopf, homoclinic and fold are
very close together before merging in the Bogdanov-Takens point, indicating that
the part between 0 2 and F 2 in Figs. 4.23 and 4.25 is very small in reality. Dots denote
codimension-two bifurcation points. b. Diagram for a higher value of the stabilizing
diffusive inflow parameter i. The Hopf bifurcation is no longer rooted in the vertical
axis, implying that in the absence of fish, there are no plankton oscillations. From
Scheffer et al. (1997b).
162 Trophic cascades
mimic the stabilizing effect of a non-homogeneous spatial distribution of Daphnia. If
i is increased a bit, the homoclinic no longer merges with the first fold (not shown).
A further increase of the stabilizing factor i causes the Hopf bifurcation to be no
longer rooted in the K-axis of the bifurcation graph (Fig. 4.26b). In this case the
system does not oscillate in the absence of fish. tncreasing fish can be destabilizing in
this case, as it can bring the system through the Hopf. The horizontal cross-sections
indicated in this bifurcation graph (Fig. 4.26b) correspond to the three left panels of
Fig. 4.25.
Roughly speaking, all lines in the bifurcation graphs have a positive slope,
indicating that in systems which are richer in nutrients (higher algal carrying
capacity (K)) all bifurcations appear at higher fish levels. A biological interpreta-
tion is that more fish is needed to let Daphnia collapse if the system is more
eutrophic. There is, however, an important complication on top of this. If K is low
·enough the system does not go through homoclinic bifurcations with increasing fish
predation, and Daphnia does not collapse until the second fold is reached (F2 ),
whereas if K is larger DaphnitJ collapses much earlier due to the homoclinic bifurca-
tion (0 1).
Put in biological terms, the oscillation in the Daphnia population makes it vulner-
able to fish. This is because it periodically brings the density down to a level that is
low enough to let a relatively small amount of fish prevent recovery of the popula-
tion. If there is no such oscillation, a much higher fish predation is needed to let the
Daphnia population collapse into the over-exploited state. Note that in general any
factor that stabilizes Daphnia oscillations will help to prevent the homoclinic bifur-
cation that triggers the collapse of DaphnitJ.
These theoretical diagrams and the abstract terminology of dynamic systems
theory may give the impression that we fiy high in the sky far away from any
down-to-earth biology. However, the next sections will show that this mathematical
world does indeed reflect crucial mechanisms that govern plankton dynamics in the
field.
Daphnia pulex
(mgm·'>
400 0 + perch (kg ha" 1 )
1976
300
200 Daphnia
20
100
10
0 0
400 1975
300
30
200
20
100
10
0 0
Apr May Jun Jul Aug Sept Oct Nov
Fig. 4.27 Comparison of biomass dynamics of age-0 yellow perch and Daphnia pulex
in Oneida Lake for two contrasting years. Daphnia is able to recover after the spring
collapse only when the density of young yellow perch is low. Redrawn from Mills and
Forney (1983).
164 Trophic cascades
food in summer is poor due to the increase in inedible colonies of blue-green
algae (Threlkeld, 1985; Lampert et al., 1986). With the rise in interest in the
effect of planktivorous fish, however, evidence accumulated in favour of an
alternative explanation. The early summer increase predation pressure due
to the young-of-the-year fish development could be the main responsible
mechanism of suppression of Daphnia in summer.
The long-term studies of the food-web dynamics in Lake Oneida show
especially suggestive patterns in this respect (Mills and Forney, 1981; Mills
and Forney, 1983). Young yellow perch are the most important planktivores
in this lake. The biomass of a cohort of young-of-the-year perch increases
sharply in early summer due to individual growth and this increase coincides
with the collapse of Daphnia (Fig. 4.27).
In years with few perch, however, Daphnia usually recovers after
the spring collapse, while in years with many perch Daphnia stays
practically absent for the rest of the season. As mentioned earlier, analysis
of data from many years suggests that there is a well defined critical perch
density above which summer Daphnia populations are suppressed (Fig.
4.16).
The coincidence of the biomass increase of young-of-the-year fish with
the collapse of Daphnia suggests that fish predation may actually be an
important cause of this collapse. On the other hand, food shortage is likely
to be a major problem for the zooplankters during the clear-water phase
when algal biomass can be extremely low. A good example of how causality
in these things can be unravelled is the analysis of the 1987 Daphnia galeata
dynamics in Lake Mendota, Wisconsin, USA (Luecke et al., 1990). The
study shows that the number of eggs per adult female decreased dramati-
cally in the course of the spring peak (Fig. 4.28a).
This number of eggs is an indication of the nutritional status, as starving
individuals can not support the production of eggs. Thus the low egg num-
bers indicate a severe food shortage in the course of the clear-water phase.
In July, the Daphnia density stays very low, but the numbers of eggs carried
by the remaining animals are high again, suggesting that their nutritional
status is not the problem. Using the egg numbers, one can estimate the
potential reproductive rate of the population over the spring and summer.
Comparison of this with the realized rate of increase gives an estimate of
Daphnia mortality. Subsequently, the Daphnia consumption by the fish
community was reconstructed over the same period, using fish censuses,
stomach analysis and bioenergetic models. The results show that fish
consumption could completely account for Daphnia mortality in summer,
while contributing only 2 o/o to the mortality at the collapse of the spring
peak (Fig. 4.28b ). Thus the emerging picture is that the spring peak of
Daphnia collapses due to food shortage, while the populations are subse-
quently kept low by fish predation. A comparable analysis with similar
outcome has been published for the Dutch lake Tjeukemeer (Boersma et at.,
1996).
Seasonal dynamics of plankton and fish 165
A
150 6
.c:
~ 100
ci 4
6 "*E
~
~
~ "'"
1il
"'
""C
.!!!
<: ~
50
!
2 :J!]
0 0
Apr May Jun Jul Aug Sept
B
20.9
2
/total mortality
~
·~
~
ci
6
~
<a
-e0
E
-!!! fish consumption
!Cl -(
Fig. 4.28 (a) During the spring peak of Daphnia galeata in Lake Mendota, egg
numbers per adult female drop dramatically, indicating food shortage. After the
collapse egg numbers recover, suggesting that food shortage is not the reason that
the population does not recover. (b) Estimated mortality of Daphnia during the
spring collapse is much higher than can be explained from consumption by fish. In
the summer, however, fish consumption can account for all Daphnia mortality.
Redrawn from Luecke eta/. (1990).
166 Trophic cascades
Other seasonal scenarios
Although the spring clear-water phase has received most attention, a single
clear-water phase in the spring is just one of the possible scenarios of
seasonal plankton dynamics. In particular, a repetition of the spring pattern
in the autumn appears to be very common in lakes. Indeed, an international
assemblage of lake plankton specialists (the Plankton Ecology Group,
PEG) described the pattern with a spring and an autumn peak as the typical
scenario for eutrophic lakes (Sommer et al., 1986) (Fig. 4.29).
There are also several case studies that demonstrate that recurrent
Daphnia peaks can cause several clear-water phases during the summer.
Examples are the French Lake Aydat where three Daphnia peaks and
corresponding clear-water phases were found in one year (Lair and Ayadi,
1989), and the German Lake Grosser Binnensee where four Daphnia peaks
occurred in one season, three of which led to a conspicuous clear-water
phase (Lampert and Rothhaupt, 1991).
It is likely that such repeated outbreaks of large Daphnia are only possi-
ble in lakes with few planktivorous fish in summer. An example that sup-
ports this idea is the development of the community dynamics in Bough
Beech Reservoir, a newly created water reservoir in south-east England
from which all coarse fish was removed in the first year (Munro and Bailey,
1980; Harper and Ferguson, 1982). Fish populations established slowly and
Fig. 4.29 A seasonal cycle with a spring and an autumn peak of large Daphnia and
corresponding dips in algal biomass, considered the typical pattern for moderately
eutrophic lakes. Redrawn from Sommer et al. {1986).
Seasonal dynamics of plankton and fish 167
ind.l-1
Daphnia
80
60
40
20
F M A M J A 5 0 N D
Fig. 4.30 A sequence of peaks in Daphnia numbers observed in the newly created
Bough Beech Reservoir, UK, before the development of a significant fish stock.
Redrawn from data of Harper and Ferguson (1982).
400,----------------.----------------.----------------.
300
200
100
300
200
200
100
100
Fig. 4.32 Conspicuous clear-water phases in Lake Tjeukemeer some times occur in
the summer (left panel) or autumn (right panel) rather than in the spring. From
Scheffer et al. (1997c).
Seasonal dynamics of plankton and fish 169
no. of cases
30 _,........
,........
24
r--,.......
18
- r--
-
12 r--
r--
6
0 h---11
0.00 0.14 0.28 0.42 0.56 0.70 0.84
minimum Chi-a
average Chi-a
Fig. 4.33 Frequency distribution of the relative depth of the deepest dip (minimum/
average of recorded values) in 257 time-series of chlorophyll-a from 71 Dutch lakes.
Each time-series covers the period from 1st April to 1st November of a given year.
From Scheffer eta/. (1997c).
0.9
0.8
o,
0 O'
0.7 oooooo'
0 '
0.6 '
,, ' .Oo
0 0 0000:
0
% Oo '0 '
0
0.5
0.4 o
oo:ooo oooo
o oo
o:oog
o
o o!tlo
o
'llloo
cP o
o,
~
J
0
0 g,_o•:&o o 8oo oo , 0 o
1) o o 0 : o 0 o o
0.3 t:P..
-- ,}-.# --.-;;!-Q.iP---.·-- ~---------------- -------------------
0 O O 1 0 0
0.2
.\ ,....::.....' ;.. .. ... .. ..": .
, ,• • • • • • • . , :
...
.:).· ·. .
0.1
0.0
~o-----~,------1oo,_----1~+-----2·oo~--~2~~--~3~oo~--~35~o--__J
Fig. 4.34 Relative depth of the deepest dip in chlorophyll-a plotted against the
average chlorophyll concentration in each of the analysed time-series (see Fig. 3.14).
Drops in chlorophyll to less 25% of the average value of that time-series (heavy
dots) are practically absent in lakes where the average level exceeds 150mgl·'. From
Scheffer et al. (1997c).
0.0001). This result is in line with the earlier analyses by Gulati (1983)
showing that strong peaks in zooplankton grazing and associated clear-
water phases are rare in highly eutrophic situations.
To analyse the timing of clear-water phases, the date of occurrence of the
dips deeper than 0.25 was checked. Deep dips in algal biomass appear to
occur at any time of the studied period of the year, although the majority of
the clear-water phases are found around May, with another peak of occur-
rences in the autumn (Fig. 4.3Sa). Obviously this fits well with reported
bestuary of seasonal patterns described in more detailed case studies.
The analysed chlorophyll data set does not allow a check whether grazing
is really the cause of the deep dips in algal biomass, as zooplankton densities
are not available. However, Danish workers have systematically analysed
time-series of zooplankton abundance from many lakes (Jeppesen et al.,
1996). To estimate the potential grazing pressure of zooplankton on the
phytoplankton, they used the rule of thumb that cladocerans can consume a
daily amount of algae equal to their own body weight, while copepods can
consume only half their own weight per day. The results indicate that in
Seasonal dynamics of plankton and fish 171
no. of cases
20
clear-water phases
15
10
0 ----,-----,-----+--+--~-~---1,_-4~-1--+--+
M A M A 0 N D (a)
80
zooplankton grazing
60
40
20
0 ----,-----~-~-~-~~~~---.---,-...,.
J F M A M A 0 N D (b)
Fig. 4.35 (a) Frequency distribution of the moment of occurrence of drops in chlo-
rophyll concentration below 25% of the average of a time-series (see Fig. 3.14)
(heavy dots in Fig. 4.34). From Scheffer et al. (1997c). (b) Seasonal variation in the
zooplankton grazing pressure on phytoplankton (% of the phytoplankton biomass
ingested per day) for moderately eutrophic Danish lakes (TP 0.05-0.10mgl" 1). The
curve indicates the median, the bars represent the 25-75% percentiles. From
Jeppesen et al. (1996).
172 Trophic cascades
moderately eutrophic lakes (TP 0.05-0.lmg 1"1) the peaks in potential graz-
ing pressure of zooplankton on phytoplankton occur precisely in the same
periods as the majority of the clear-water phases observed in the Dutch data
set (Fig. 4.35b ), suggesting that the Dutch clear-water phases are indeed the
result of top-down control. The estimated potential grazing pressures reach
values close to lOOo/o of the algal biomass per day in lakes with moderate
nutrient contents. Obviously, this should be sufficient to graze down even
fast growing algal populations.
Interestingly, the Danish analysis also shows that grazing pressure varies
in a systematic way with the nutrient status of the lake. In hypertrophic lakes
grazing pressure is low during the entire year. This is in line with the Dutch
results, showing that clear-water phases are usually absent in lakes with high
chlorophyll levels (Fig. 4.34).
Fig. 4.36 Seasonal variation in biomass of planktivorous fish (right-hand panel) can
cause the planktonic system (left-hand panel) to switch between cycles with periodic
Daphnia peaks (cylinder in left-hand panel) and a stable algal dominated state
(vertical equilibrium line in left-hand panel), depending on whether the two main
bifurcations in the system are crossed. If planktivory is high (I) no Daphnia peaks
will occur at all. When planktivorous fish stays low (III) Daphnia can remain abun-
dant throughout the summer. However, intermediate scenarios cause Daphnia to
collapse through a homoclinic bifurcation in the spring, leaving algae to dominate
throughout the summer. From Scheffer et al. (1997c).
corresponding (technical) text sections. Here, the figure is tilted so that fish
predation is on the vertical axis. At low fish predation there are zooplankton
- algae oscillations (cylinder) whereas at higher fish predation there is a
stable algal dominated equilibrium where zooplankton is overexploited by
fish (curve rooted in F 1). These two modes of behaviour co-exist as alterna-
tive 'equilibria' for a small range of fish predation. Two 'bifurcation points'
mark the transitions from the oscillatory regime to the stationary one (0 1,
the 'homoclinic bifurcation') and vice versa (F1, the 'fold bifurcation') .The
exact development of planktivory over the year is hard to know, but if we
assume that it can be mimicked in the model by a cycle in the consumptive
capacity for plankton (G1) with a maximum in summer (right-hand panel),
three qualitatively different scenarios can be distinguished:
I Planktivory remains entirely above the threshold at which Daphnia can
recover from over-exploitation (the fold bifurcation, F 1).
II The annual minimum in planktivory is below this threshold but the
maximum is above the threshold at which the oscillating Daphnia state
collapses due to the homoclinic bifurcation ( 0 1).
III Planktivory stays entirely below the homoclinic threshold ( 0 1) at which
Daphnia would collapse.
174 Trophic cascades
The first scenario corresponds to lakes where large zooplankters are absent
throughout the year and there is no clear-water phase. The second situation
corresponds to the classical clear-water phase scenario, with a Daphnia peak
in the spring but low Daphnia numbers and high algal biomass in summer.
The third scenario represents a situation where Daphnia remains oscillating
throughout the summer. Indeed, as mentioned earlier, all of these scenarios
are observed in lakes, and model results suggest that differences in their fish
communities could be a reasonable explanation.
As argued, scenario I is especially common in hypertrophic shallow lakes
(Fig. 4.34), where, as explained in section 4.5, the fish stock can be very high
due to the availability of benthic food. Since these benthivorous fish switch
to feeding on zooplankton when it is available, the potential grazing pres-
sure on Daphnia is high even in the absence of young-of-the-year fish
(Jeppesen et al., 1996).
The model suggests that the collapse of Daphnia at the end of the clear-
water phase and their subsequent absence in summer (scenario II) corre-
sponds to a 'homoclinic bifurcation'. The essence of this bifurcation in
biological terms is that Daphnia collapses due to food shortage (the limit
cycle), and that this brings the population down to a level that is low enough
to let a relatively small amount of fish prevent recovery (the over-exploited
state). Indeed, this corresponds precisely to the mechanism revealed by the
analysis of the factors that drive the plankton dynamics in Lake Mendota
(Fig. 4.28).
Note that it is the starvation collapse of the Daphnia population that
makes it vulnerable to fish. If there is no such oscillation in Daphnia, a much
higher fish predation is needed to let the Daphnia population arrive at the
over-exploited state, as explained more thoroughly in the corresponding
technical section (Fig. 4.26). This implies that, factors that stabilize Daphnia
oscillation will help to prevent the collapse of Daphnia. As argued above,
vegetation may help to stabilize Daphnia dynamics in shallow lakes by
supplying refuges in which the animals aggregate and detritus that can be
used as an alternative food source.
It is interesting to note that the mechanism that causes the famous snow-
shoe hare - lynx cycles is thought to be closely related to the emerging
picture of the clear-water phase. The decline in the hare population that sets
the cycle is initiated by a food limitation of the hares that have overgrazed
the vegetation. When the hare population has collapsed to low numbers, the
impact of the predator populations, whose numbers stay roughly constant,
becomes increasingly important, keeping the hare population low for an
extended period (Keith, 1983; Keith, 1990).
1-ecos( 2 m)
365
(14)
where t = 0 stands for the first of January. In this formulation the minimum
value of each parameter (i.e. the value in the middle of the winter) is equal
to its maximum (in summer) multiplied by (1 -e)/(1 =e). Thus, the summer
maximum of a parameter corresponds to the default value and e determines
the amplitude of seasonal change.
In addition to the temperature induced variation, fish predation pressure
( G1) should show a seasonal variation due to the reproductive cycle. Assum-
ing this cycle to be sinusoidal and in phase with the variation in temperature
and light we can include the effect simply by multiplying G1 with an extra
seasonal impact ( u). Thus the complete seasonal model becomes:
dA
dt
=uwA(l-~J-zu(,w~+d(u(,lK-A)
O'(t)K A+hA
(15)
dZ A Z2
- = eu(r)gZ--- O'(r)mZ- O'(r)O'(t)G/ - 2 - - 2 (16)
dt A+hA z +h,
A simple way to look at the consequences of such a seasonal variation is to
compute the cycles and equilibria of the system for each day of the year with
176 Trophic cascades
Algae (A) 1
-
time (days)
Fig. 4.37 The attractors (equilibria and cycles) of the seasonal plankton model
constructed by computing the cycles and/or stable equilibria separately for each
day t of the year using tbe appropriate value of cr,. The dotted line represents the
true asymptotic behaviour of the seasonally forced model. From Scheffer et al.
{1997c).
the appropriate seasonal values of the parameters and assemble them into a
picture that shows the entire year (Fig. 4.37).
This representation shows that plankton has the tendency to oscillate in
the spring and autumn whereas it is stationary with high algal biomass in
summer and stationary with low algal biomass in winter. The discontinuities
show that the transitions to and from the turbid summer equilibrium are
catastrophic (corresponding to the homoclinic and fold bifurcations). How-
ever, the dynamic behaviour of a periodically forced system can not really
be inferred from such an assemblage of frozen asymptotic behaviours. This
becomes apparent if real simulated dynamics are plotted in the same figure
(dashed line in Fig. 4.37). The depicted simulated path shows the asymptotic
behaviour to which the seasonal model always converges after many simu-
lation years. Therefore, this seasonal cycle is a real attractor, comparable to
the stable points and limit cycles of the non-seasonal model. This seasonal
attractor follows the artificially constructed 'frozen attractor set' only ap-
proximately. The discrepancy is understandable. The populations never
have time to reach the asymptotic behaviour corresponding to the condi-
tions at a certain day of the year because the conditions change continu-
Seasonal dynamics of plankton and fish 177
ously. Therefore, the behaviour of the seasonal model can only be studied
properly by analysing the real seasonal attractors.
A(o)
A(tl
A(o)
A(t)
0.12
0.10
o, ' \\ I
0.08 I
F, I I
I I
0.06 I I
I I
3(')
I I
I I
0.04 I I
I I
I I
0.02 I I
I I
I I
I I
0+-----.----.-----.----.---~
0 0.2 0.4 0.6 0.8 1.0
seasonal forcing (E)
Fig. 4.39 Bifurcation diagram showing how seasonal forcing and planktivory affect
the behaviour of the model. The bifurcation boundaries merging from points 0 1 and
F 1 divide the parameter space into three distinct regions where the asymptotic
behaviour of the model is different (see text). In region (1), clear-water phases are
absent, while in region (3) clear-water phases occur. In region (2) the asymptotic
behavioural regimes of region (1) and (3) co-exist. On the left side of the dashed
curves the dynamics in region (3(1 1) are not locked with the rhythm of the seasons
(see text), on the right side of these curves (region 3(1)) the system behaves periodi-
cally with a period of one year. The dynamic behaviour at the parameter settings
marked as a, b, c and dis illustrated in Fig. 4.40. From Scheffer et al. (1997c).
band is crossed from below, the torus disappears through homoclinic contacts. The
boundary rooted at F 1 is a tangent bifurcation curve. As this curve is crossed from
above, a stable cycle and a saddle cycle collide and disappear. Thus, region 1 contains
a unique attractor, namely a stable seasonal cycle with high algal biomass (Fig. 4.40a)
corresponding to the 'turbid equilibrium' of the constant parameter case. In the
small region 2, this turbid regime coexists with a quasiperiodic regime characterized
by clear-water episodes. Region 3 is characterized by the existence of a torus,
although in many subregions (3(il) the regime is locked to be purely periodic with a
period of i years (see below). In our case all of these subregions are small with the
exception of region 3( 0 where the behaviour on the torus is locked to a period of one
year. The bifurcation line demarcating the border of region 3(1l corresponds to a
tangent bifurcation of cycles on torus. Although numerical experiments have shown
in which zone of the parameter space it occurs (3l'l) the exact position could not be
Seasonal dynamics of plankton and fish 181
Biomass (mg 1" 1) Biomass(mgl-1)
12r----------------------, 12r------------------------,
10 10
Algae (A)
#:
,,
/ ''~
____ ............''' '~
------------ ... __________________ _
Zooplankton (Z)
Zooplankton (Z) --------
JFMAMJJASOND JFMAMJJASONO
10
Fig. 4.40 Seasonal dynamics of the plankton model for parameter setting corre-
sponding to points a, b, c and din Fig. 4.39 (s = 0.7; Gr = 0.12, 0.09, O.Q75 and 0.025
respectively). From Scheffer et aL (1997c).
detected due to numerical problems. In this parameter zone (3(' 1) several small
islands occur where cycles with a period of more than one year exist. At a seasonal
forcing of e = 0.7 which is probably reasonable for simulating temperate lakes
(Scheffer et al., 1997c) the frequency of oscillations in the plankton tends to be
locked with that of the seasons, in such a way that the same pattern of plankton
dynamics repeats each year (Fig. 4.40).
~ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
years
Algal Biomass
b
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
years
Algal Biomass
c
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
years
Fig. 4.41 Dynamic patterns occurring in the narrow band of bifurcations that sepa-
rates regions 1 and 3<'1in Fig. 4.39. (a) A clear-water phase occurring two out ofthree
years (G1 = 0.1008); (b) a clear-water phase occurring two out of four years (G1 =
0.1011); (c) a segment of a complex cycle where the occurrence of clear-water phases
appears erratic (G1 = 0.1013). From Scheffer et al. (1997c).
Daphnia peaks and clear-water phases in a year being predicted only for a
small parameter ninge at low fish densities. For a large range of parameter
settings the model predicts the clear-water phase to be either entirely
absent (Fig. 4.40a ), or to happen in the spring (Fig. 4.40b) with a possible
repetition in the autumn (Fig. 4.40c). Indeed, in the field these three sce-
narios seem to be the most common ones also (Figs. 4.29, 4.31 and 4.35). The
timing of the simulated patterns corresponds remarkably well to the field
observations as well. Without any special tuning the spring clear-water
phase occurs around May (Figs. 4.40b,c,d) as it does in lakes (Fig. 4.35). The
model suggests that the timing of the spring clear-water phase should de-
pend on the fish density, coming later if fish density is higher (Figs.
4.40b,c,d). Suitable data to check this prediction are rare, but in Lake
184 Trophic cascades
Mendota where the relationship between fish and plankton dynamics has
been analysed for many years, it has indeed been noted that the spring clear-
water phase can come earlier when the density of planktivorous fish is low
(Temte et al., 1988; Vanni et al., 1990; Rudstam et a/., 1993). Even the
peculiar prediction that incidental isolated clear-phases can occur in the
summer or autumn in lakes that normally show no clear-water phase (Fig.
4.41) is supported by data (Fig. 4.32).
Importantly, the results are quite independent of the precise model for-
mulation. The use of more realistic time lags between the light and tempera-
ture forcing functions and variation in their amplitudes hardly changes the
patterns produced. Also, the dynamic patterns produced by a comparable
but much more elaborate model (Doveri eta/., 1993) are very similar to the
repertoire of the minimal model shown here. In that model, nutrient recy-
cling and the dynamics of young-of-the-year fish are modelled dynamically,
the role of adult fish that prey on zooplankton facultatively is modelled
explicitly, and realistic temperature and light scenarios at different latitudes
are analysed.
Inedible algae
A potentially important aspect that is not considered in the model is the
presence of inedible cyanobacteria. The analysis suggests that the absence
of clear-water phases can be explained as the effect of a high density of
planktivorous fish. The time-series analysis of Dutch and Danish data indi-
cates that the absence of clear-water phases is especially frequent in hyper-
trophic lakes with high algal biomass (Fig. 4.34, Fig. 4.35b). Indeed, fish
biomass is typically very high in such lakes, and this may well explain the
absence of clear-water phases. However, hypertrophic shallow lakes are
also frequently dominated by cyanobacteria as explained in the previous
chapter. Several studies show that these algae are usually not well edible
(Arnold, 1971; Schindler, 1971) and that Daphnia growth can be severely
reduced in their presence (Gliwicz, 1990; Gliwicz and Lampert, 1990). Also,
toxic substances released by cyanobacteria have been shown to reduce
filtering rates of daphnids by 50% or more (Haney eta/., 1994). Obviously,
such adverse effects of cyanobacteria may well contribute to the absence of
clear-water phases.
At first sight the size structure of the zooplankton community of hyper-
trophic lakes seems to indicate that fish is the dominant factor involved. The
community is typically dominated by small zooplankters. As explained ear-
lier, this is thought to be characteristic of situations with a high predation
pressure from planktivorous fish. Selective predation of fish removes the
larger individuals (Brooks and Dodson, 1965; Shapiro and Wright, 1984;
Hambright, 1994; Seda and Duncan, 1994) and Daphnia also tends to
change its life history strategy, becoming smaller in response to chemical
186 Trophic cascades
cues released by fish (Weider and Pijanowska, 1993; Engelmayer, 1995).
However, it has also been shown that large Daphnia species are less able
than small ones to forage and grow in the presence of filamentous blue-
green algae (Hawkins and Lampert, 1989; Gliwicz, 1990; Gliwicz and
Lampert, 1990), thus the absence of large Daphnia from hypertrophic lakes
may at least in part be caused by the poor food situation.
Causality is difficult to unravel in this matter, but an extensive analysis of
data from many Danish lakes indicates that top-down control by fish is
probably the dominant mechanism preventing large zooplankters from
peaking and grazing down algal biomass in most cases (Jeppesen et al.,
1996). Many of the hypertrophic Danish lakes that entirely lack episodes of
high zooplankton grazing pressure are dominated by easily edible green
algae. Also the mid-summer decline of Daphnia in lakes of lower nutrient
contents occurs every year in many Danish lakes, irrespective of whether or
not cyanobacteria are present.
An interesting study showing the complexity of the interaction of ined-
ible cyanobacteria and Daphnia in a eutrophic lake has been published by
Samelle (1993). The natural seasonal pattern in the lake was that the species
composition of algae switched from diatoms and green algae during the
spring bloom to small flagellate algae during the clear-water phase and
filamentous cyanobacteria after the clear-water phase. At first sight this
seems to support the view that grazing favours dominance of the algal
community by the 'grazing resistant' cyanobacteria as it eliminates edible
competitors. However, after a fish kill, Daphnia grazing was able to retard
succession to 'inedible' cyanobacteria in the summer. Instead the
phytoplankton community remained dominated by edible small flagellates.
The idea that in the absence of fish predation Daphnia could suppress
filamentous blue-green algae was confirmed by enclosure studies in the lake.
Thus cyanobacteria may affect Daphnia development negatively but
Daphnia may also suppress cyanobacteria.
The observations of Samelle (1993) suggest that the effect of Daphnia
grazing on algal composition may depend on grazing intensity. Strong
Daphnia grazing seems to lead to a dominance by small flagellate algae,
while mild grazing may favour 'inedible' cyanobacteria. This view is sup-
ported by observations of Danish workers (Jeppesen, pers. comm.). The
mechanisms through which these two very different groups can survive
grazing pressure are entirely different. Selection of grazers against the
large colonies is an intuitively straightforward mechanism that may favour
blue-green algae. On the other hand larger algae have lower growth rates
in general (Reynolds, 1988). Therefore, if the population suffers losses
due to grazing (or sinking or flushing) these are not easily compensated
for by growth. The small flagellate algae that are typical for heavily
grazed situations, on the other hand, have very high growth rates. Thus
even severe losses can be compensated for by the fast reproduction. As
explained in the previous chapter this is also the reason why fast growing
Seasonal dynamics of plankton and fish 187
algae tend to dominate in situations with high flush-rates or large sinking
losses.
There is some evidence that the outcome of the mutually negative inter-
action between cyanobacteria and Daphnia is dependent not only on preda-
tion pressure from planktivorous fish but also on the initial density and the
condition of cyanobacteria. Laboratory studies show that especially high
densities of cyanobacterial filaments have a strong negative effect on
Daphnia growth but that the animals are better able to handle the blue-
green algae if the filaments are short or deteriorating (Gliwicz, 1990; Gliwicz
and Lampert, 1990). At lower filament densities, Daphnia is less affected
and some studies indicate that Daphnia grows even better on a mix of
cyanobacteria and green algae than on a culture of green algae alone (Gulati
pers. comm.).
Invertebrate predators
Another factor that may complicate the straightforward trophic cascade
from fish to algae is the presence of large densities of invertebrate predators
which forage on zooplankton but are themselves prey to fish. Especially
notorious are the predacious Leptodora and some mysid shrimp species
(Fig. 4.42).
Obviously, it is difficult to predict a priori what will be the effect of a
reduction of the fish density in situations where such planktonic predators
are present, since fish suppresses Daphnia but also its invertebrate predators
(Fig. 4.43).
Reduction of the fish stock will reduce its direct impact on herbivorous
zooplankton but can also allow an increase of invertebrate predators. Pre-
dicting the overall effect on Daphnia requires detailed specific information
on the prey selection and population dynamics of the species involved.
An attempt to combine such information in an integrated model has been
made for the food web of Lake Mendota (Luecke et at., 1996) where
Leptodora may be in part responsible for the midsummer decline in
Daphnia (De Stasio, Jr et al., 1995). The model predicted that in general
Daphnia biomass should benefit from a decrease in the density of the two
dominant planktivorous fish species, except at very low fish densities where
Daphnia would be suppressed by an increasing population of Leptodora
(Fig. 4.44).
Indeed, since the size of most invertebrate predators makes them very
attractive food for fish, it seems reasonable to assume that planktonic inver-
tebrate predators can only play a significant role if the density of
planktivorous fish is very low. Even in fishless situations, however, inverte-
brate impact on zooplankton may be minor in practice (Paterson, 1994).
Since small invertebrate predators such as Chaoborus can only handle rela-
tively small-bodied zooplankton (Pastorok, 1980; Luecke and O'Brien,
1983), their presence may often help to cause the shift to dominance by large
188 Trophic cascades
Fig. 4.42 Mysid shrimps (Neomysis sp.) occur massively in some brackish lakes.
They are omnivorous animals that consume detritus, but can also catch zooplankton.
Daphnia that is usually observed in fishless lakes (Vanni, 1988) rather than
controlling the abundance of large herbivores.
Field evidence for the role of invertebrates, in reducing the density
of larger Daphnia species is rare but some suggestive cases have been
described. In Bautzen Reservoir (Germany), stocking of piscivorous
fish has led to a decline in planktivorous fish (Benndorf et al., 1988).
Daphnia abundance increased first, but decreased in later years despite
a further reduction in planktivorous fish. Possibly, the simultaneous
increase in Chaoborus larvae and Leptodora is in part responsible
for this decrease in Daphnia, but this link has not been clearly
demonstrated.
In Lake Wolderwijd (Meijer et al., 1994a) a 75% reduction of the fish
stock in winter resulted in a spring peak of Daphnia and a clear-water
phase during which the lake was clearer than ever before. In summer,
· Seasonal dynamics of plankton and fish 189
Fig. 4A3 When fish consumes Daphnia but eats also a potential invertebrate preda-
tor, the net effect of fish on Daphnia is not clear a priori.
225
;; 200
E
~ 175
"
·;:
~ 150
"ill, 125
j
Fig. 4A4 The simulated net effect of two planktivorous fish species, Cisco and Perch
on the biomass of large daphnids in Lake Mendota. At very low fish biomass the
model predicts that daphnids are suppressed by an increase in the invertebrate
predator Leptodora. Since this invertebrate is a preferred and vulnerable food for
fish, it survives only at the lowest fish densities. Redrawn from Luecke et al. (1996).
o+----.---.---,----r-~
c:
.a 5
""-'!!c:
c.
0
~ 0
0 3 6 9 12 15
Mean lake depth (m)
Fig. 4.45 Regression lines showing that zoobenthos biomass is higher in shallow
lakes than in deeper ones, while the ratio of zooplankton to zoobenthos biomass
increases systematically with lake depth. From Jeppesen et al. (1996).
Fig. 4.46 Tench (Tinea tinea) can be a common fish in vegetated lakes. It eats all kind
of macro-invertebrate prey but has a preference for snails.
The benthic connection 195
12
10
NE
8 control
~
a ~ 6
E
~
"0 4
~
.5i 2
NE 1200
b;,
.: 1000
~
E 800
1::'
b "0
QJ 600
-9
.<: tench
:g 400
c:
~ 200
.<: control
0.
-~
0.. 0
June July August
140
120 control
NE
100
~
~ 80
c E tench
1::' 60
"0
:3 40
"'w8
20
0
June July August
Fig. 4.47 Enclosure experiments in a shallow Swedish pond that in the presence of
the molluscivorous fish tench, (a) snail biomass decreases (b) periphyton biomass
increases and (c) the growth of the submerged plant Elodea canadensis is reduced.
Error bars denote ±1 Standard Error. From Bronmark (1994).
196 Trophic cascades
sure by snails. Also the composition of the periphyton changed markedly
with grazing. In the presence of fish, when snails were rare, large
'overstorey' species, such as stalked diatoms and filamentous algae domi-
nated the community. When snail grazing was severe, small tightly adherent
species became more abundant. This shift in species composition is com-
monly observed in studies of the effect of snail grazing (Bronmark, 1989),
and is quite comparable to the effect of large herbivores on the structure of
terrestrial vegetations.
Shading by a layer of periphyton can limit the growth of submerged
plants (Sand-Jensen and Borum, 1984). It may therefore be expected that
snails that graze down periphyton biomass should thereby indirectly en-
hance plant growth (Carpenter and Lodge, 1986; Thomas, 1987). Such a
positive effect of snails has been shown in experiments with fresh-water
macrophytes (Bronmark, 1985; Underwood, 1991; Daldorph and Thomas,
1995), and also the growth of seagrass (Zostera marina) has been shown to
benefit from the activity of periphyton grazers (Hootsmans and Vermaat,
1985; Howard and Short, 1986). In view of this link between snails and
plant growth, molluscivorous fish should potentially have an indirect nega-
tive effect on submerged macrophytes. Indeed, this effect has also been
demonstrated. In a set of enclosure experiments in a shallow Swedish pond,
tench was used as a molluscivore (Bronmark, 1994). Snail biomass de-
creased in the presence of this fish (Fig. 4.47a), leading to a increase of
periphyton biomass (Fig. 4.47b) and a significant decrease in the biomass of
Elodea canadensis, the dominant macrophyte {Fig. 4.48c). Comparable re-
sults were obtained by Martin et al. (1992) who found that the biomass of
submerged plants increased when molluscivorous redear sunfish were
excluded.
Fig. 4.48 Size distribution of crucian carp (Carassius carassius) in southern Swedish
ponds with and without piscivores (data pooled from a survey of 46 ponds). From
Briinmark eta/. {1995)
Piscivores 197
4.5 PISCIVORES
The effects of piscivory on shallow lake food webs have been studied less
intensively than the effects of planktivory and benthivory. Nonetheless it
has become clear that predation is an important structuring force in fish
communities. In the presence of piscivorous fish, potential prey fish often
change their behaviour in order to reduce predation risk, and this can lead
to crowding and increased food competition in safe vegetated areas
(Werner et al., 1983; Persson et al., 1993). Although such predation
avoidance strategies help reduce predation losses, marked impacts of
piscivores on the abundance and size structure of fish populations have been
found.
Fig. 4.49 Northern pike (Esox lucius) is a voracious predator that hunts from an
ambush.lt can swallow fish of up to half it own length. Cannibalism is a major source
of mortality for juvenile pike and their survival depends largely on the presence of
vegetation as a refuge.
Piscivores 199
instance, in a three-year replicated pond study, found that the presence of
piscivores caused a shift from smaller to larger size classes in pumpkinseed.
A whole lake example is the German Bautzen reservoir, where the stock of
pikeperch (Stizostedion lucioperca) was enhanced by restriction of the catch
by sports fisherman, and repeated stocking with pond-raised juveniles. This
manipulation led to a reduction in the total biomass of the main prey fish,
perch, and to an increase in perch average individual size (Benndorf et al.,
1988).
Several studies show that survival of juvenile fish can be severely re-
duced, especially when a high density of small piscivores occurs (He and
Wright, 1992; Prejs et al., 1994; Berget al., 1997). In the Polish Lake Wirbel,
for instance, juvenile pike (Esox lucius) (Fig. 4.49) were stocked in high
densities (up to 3000ha-1} during three subsequent springs (Prejs et al.,
1994). In the fourth year a rotenone treatment killing all fish revealed that
in the years with pike-stocking, recruitment of cyprinids had been strongly
repressed. Populations of roach and white bream (Blicca bjoerkna) con-
sisted almost entirely of individuals older than 3 years. Stocking of juvenile
pike in the autumn instead of the spring had little impact as few survived
until the following spring.
A Danish study confirms the potentially strong impact of young-of-the-
year pike on cyprinid (mainly roach} recruitment (Berget al., 1997). Differ-
ent densities of pond raised juvenile pike were stocked in the spring of four
subsequent years in Lake Lyng. Annual censuses in the autumn indicated
that the number of small fish (<lOcm) in the vegetated littoral zone where
pike hunt decreased markedly with increasing numbers of stocked pike (Fig.
4.50).
250~ 89
E
u
0 200
v • 92
~ 150
"""E
.0
100 • 94
.91
.90
5"
w
50
•
::> 93
0..
u
0
0 1000 2000 3000 4000
Stocking density of pike (no./ha)
Fig. 4.50 Effect of juvenile stocked pike on the density of small prey fish (<10cm) in
the littoral zone of the Danish lake Lyng. From Berget al. (1996).
200 Trophic cascades
The large impact of juvenile pike on their prey density can be understood
from a back-calculation of how much food is needed to allow an individual
pike to grow up to 18cm in the first year as they usually do. Depending on
the contribution of fish to the total diet such animals are estimated to
consume up to about 600 juvenile cyprinids per capita during their first
season (Grimm, 1989).
'2
15
fishery,'\
:
'
.
~ 10
''
''
~
.Q
no fishery
"' 5
0
15 25 35 45 55 65 75 85 95 105
length (em)
12
Planktivorous fish
...
'2 8
.c
~
~
E
.2 4
"'
5 7 9 11 13 15
length (em)
Fig. 4.51 An intensive gill-net fishery removing pike-perch >60cm from a set of
shallow lakes in The Netherlands has produced a shift to dominance by smaller size-
classes (upper panel), while the density of small cyprinid prey-fish in the lakes
decreased (lower panel) presumably due to an increase in predation pressure from
the changed pike-perch population. After data from Lammens et at. (1997).
1\
piscivores
fish
f-1----'/\--11
/\
zooplankton
phytoplankton
Fig. 4.52 The trophic cascade involves several critical size matches between preda-
tors and prey. Most prey fishes are too large to be handled by most piscivores, most
zooplankton is too small to be a profitable food for most fish, and much of the
phytoplankton size classes can only be handled by the largest herbivorous
zooplankters. From Scheffer (1997)
'"il
c: 9
• •••
...:/".
•
•• • • • •
~ 6
'Iii
u::
•
0 ~ . .
.s:;: •• •
B
<1£
!S
80
•
E!
-~c: 60
l'!
;,'! 40
"•••'-1
•
0
,· .....J,. c
...·- ..
0
~ 0.6
0. •
~ ••
if 0.4
• ••
0.
0
0 0.2
tC ... :
•• ••
N
0
0 0.2 0.4 0.6 0.8 1.0
Total phosphorus (mg P 1·1)
Fig. 4.53 Relationship between the total phosphorus concentration and fish biomass,
the share of piscivores and the ratio of zooplankton to phytoplankton biomass in
Danish lakes. From Jeppesen eta/. (1996)
mately half the value found in the otherwise comparable IJsselmeer, sup-
posedly due to light limitation. A very similar difference exists between the
biomass of fish of the two lakes (Lammens et at., 1996).
Fish biomass is often closely correlated to the biomass of benthic inver-
tebrates (Hanson and Leggett, 1982), suggesting that benthos is an impor-
tant food source that largely determines the total biomass of many fish
communities. This is especially so in shallow lakes where benthos is rela-
tively more important than in deep lakes. As argued earlier, the increase
of (omnivorous) fish with eutrophication is thought to result in a
disproportionally high predation pressure on zooplankton in such lakes
(Jeppesen et at., 1996).
208 Trophic cascades
100
"0 D daylight
E
0
80 - r= II darkness
0 §
~ ·= 60 -
_o..C:
n
E u
40 -
~]
"- 20 -
::l
0
Perch Bream Roach
Fig. 4.54 Efficiency of perch, bream and roach in capturing chironomid larvae
during experiments at daylight and darkness. From Diehl (1988).
More surprising than the increase of fish biomass with nutrients are the
changes in trophic structure. Above a total-P level of 0.1 mg r' the share of
piscivores in the fish community drops sharply and so does the ratio of
zooplankton to phytoplankton biomass (Fig. 4.53b,c). This suggests a sys-
tematic shift in the role of top-down control with nutrients. At low nutrient
levels piscivory is relatively important and planktivory is low allowing
zooplankton to graze down algae. At high nutrient levels the strong top-
down links have shifted one trophic level, and strong top-down control of
phytoplankton by zooplankton is rare.
To understand the shift in the fish community it is necessary to consider
the difference in species composition between lakes of different productiv-
ity in more detail. The general pattern is that in oligotrophic lakes
Salmoniformes are the most important fish group, while in moderately
productive lakes percids dominate and in eutrophic and hypertrophic lakes
dominance shifts to cyprinids (Kitchell et a/., 1977; Leach et al., 1977;
Persson eta!., 1991). On the American continent there is another group, the
centrarchids, that increases with eutrophication more or less like the
cyprinids do (Oglesby et al., 1987).
The shift from percids to cyprinids corresponds to the strong decrease in
the percentage of piscivores with increasing nutrient status reported, for
instance, in the Danish study (Fig. 4.53b ). In European lakes the large
proportion of piscivores in mesotrophic lakes is almost entirely due to
dominance by perch, while roach and bream are usually the most important
cyprinids (Persson eta/., 1991). The reason for the shift has been explained
from the differences in competitive ability of these species depending on
turbidity of the water and the presence of aquatic vegetation. As discussed
further in the next chapter perch is a superior competitor for food in dense
vegetation, which is predominantly found in the less eutrophic lakes. In
turbid water, characteristic of hypertrophic lakes, roach and bream are the
better competitors (Diehl, 1988; Persson et al., 1991). Juvenile animals of
these cyprinids have been found to feed at maximum rates even at very low
General patterns 209
light intensities characteristic of turbid lakes (Bohl, 1980; Diehl, 1988) (Fig.
4.54), and adult bream can switch to a filter-feeding mode for catching
zooplankton (Lammens et al., 1985) that does not require spotting the prey
visually. Bream and roach also keep foraging efficiently on chironomid
larvae in unvegetated sediments until the density of this prey is quite low
(Diehl, 1988). As a result zooplankton as well as zoobenthos are reduced by
cyprinids to a density that is too low for young perch to forage efficiently in
these turbid conditions. The poor food conditions caused by the cyprinids
can prevent young perch from growing enough to become piscivorous.
Thus, competition by bream and roach prevents juvenile perch from
becoming a predator in eutrophic turbid lakes. In clear water with aquatic
vegetation, on the other hand, juvenile perch are good competitors and
many individuals can become large enough to be piscivorous (Persson, 1986;
Persson, 1987a; Persson, 1987b; Persson and Greenberg, 1990a; Persson and
Greenberg, 1990b).
In practice, much of the changes in food-web structure with the nutrient
status of lakes are probably related to the disappearance of submerged
vegetation with eutrophication rather than to productivity or turbidity per
se. The direct and indirect impacts of vegetation of shallow lake communi-
ties are treated further in the next chapter.
5 Vegetation
Fig. 5.1 The rounded leaves of the yellow waterlily (Nuphar lutea) and the finely
dissected leaves of Ceratophyllum are extremes in the wide spectrum of plant forms
that are found in lakes. The growth form of plants largely determines their effect on
wave resuspension and their effectiveness as a refuge for small animals against
predation by bigger ones.
212 Vegetation
develop rapidly at the sediment surface in the spring entangling and shading
other submersed plants. Later such algal masses tend to become buoyant
and form dense floating algae beds ('flab') that deteriorate gradually
in summer. All of these structural groups have their specific impacts on
the functioning of a lake, and in systems where vegetations with differ-
ent growth forms co-occur, each one typically has its own associated in-
vertebrate fauna (Dvorak and Best, 1982; Scheffer et al., 1984; Dvoi'lik,
1987).
With respect to the functioning of most temperate shallow lakes, rooted ,
submerged plants are probably the most important group. They can develop
massively over the entire Jake bed, and have a tremendous impact on the
system. Different types of submerged plants develop under different condi-
tions and these types can have quite different impacts on the system. At
least two extremes in growth form need to be distinguished: plants that can
grow long and concentrate most biomass in a canopy just under the water
surface in shallow situations, and plants that stay short and produce low but
sometimes very dense vegetations. Vegetation in turbid shallow lakes is
typically dominated by angiosperms that develop growth forms of the first
type, such as sago pondweed (Potamogeton pectinatus) and Hydrilla. An
example of the seasonal development of this vegetation structure has been
described for an Elodea dominated weedbed in a Czechoslovakian fish pond
(Pokornyet a[., 1984). Early in the growing-season plant biomass is distrib-
uted relatively evenly over the water column, but soon overall biomass
increases and most plant matter becomes concentrated just below the water
surface (Fig. 5.2).
{!l'J 0.1
iil
'
1i 0.2
~ f
E 0.3
.g
8
•
0.4
~
i5 0.5
0 100 200 0 100 200 0 100 200
Plant biomass (g DWT m·2)
Fig. 5.2 Change in the depth distribution of plant biomass in an Elodea vegetation
from spring to summer. By the end of July most of the biomass is concentrated close
to the water surface. From Pokorny et at. (1984).
Implications of vegetation 213
Isoetides are an extreme example of the low growing structural type.
These rosette forming plants are typical of very clear water. Another
very important group with a low growth form are the charophytes
(Charophyceae). Despite the similarity in growth form to higher plants,
charophytes are in fact algae. Consequently they differ strongly in physiol-
ogy from other aquatic plants. Charophyceae can be characterized as typical
r-select pioneers that have a high growth rate, and produce many small but
highly persistent diaspores. They are often among the first species to colo-
nize new water bodies, and although they usually represent a transitional
vegetation stage, they can also form permanent vegetations in some situa-
tions (Wood, 1950).
Invertebrates
Vegetation stands usually have a much richer invertebrate community
than unvegetated sites, both in terms of species numbers and total bio-
mass (Gilinsky, 1984; Diehl, 1988; Engel, 1988; Hargeby et al., 1994). The
differences with open water fauna are especially pronounced if vegetation
stands are dense. As an example consider the local differences in macro-
invertebrate fauna found in Lake Krankesjon (Hargeby et al., 1994). This
lake has unvegetated sites, stands of sago pondweed and Chara fields. The
heavy Chara mats have a 12 times higher vegetation biomass than
pondweed stands. Macroinvertebrate biomass in the plant stands is higher
than in the unvegetated parts, but the invertebrate biomass is much more
elevated in the dense Chara fields than in the relatively sparse pondweed
stands (Fig. 5.3).
Also diversity is higher in the vegetated areas. In the unvegetated areas
Chironomidae and Oligochaeta consitute as much as 74-100% of the total
biomass, whereas in the vegetation stands a wide range of invertebrate taxa
is found. Again pondweed stands are intermediate in invertebrate fauna
diversity between unvegetated parts and Chara beds.
~
~
~ 10
E
~
GJ
"E!
..c
-I!
~ 5
-~
:;!
~
No Potamogeton Chara
Fig. 5.3 Marcroinvertebrate biomass (g dry mass m-2) in a Chara tomentosa vegeta-
tion, a Potamogeton pectinatus vegetation and in an unvegetated area in Lake
Krankesjon. From Hargeby et al. (1994).
Implications of vegetation 215
It seems reasonable to assume that the sheltering effect of plants against
predation by fish is an important reason why invertebrate communities are
richer in dense plant stands, but other factors play a role too. This can be
seen from the fact that even when fish are excluded, the macro-invertebrate
density and species richness are usually found to be higher in experimental
treatments with submerged plants than in treatments with few or no plants
(Gilinsky, 1984; Rabe and Gibson, 1984; Gregg and Rose, 1985). For many
invertebrates, food availability is an obvious reason to be in the vegeta-
tion. Few invertebrates seem to consume significant amounts of intact
macrophytes, but decomposing plants provide a relatively high quality detri-
tus that is eaten by animals such as isopods, snails and insect larvae
(Komij6w et a/., 1995). However, the periphyton layer that covers the
macrophytes is probably more important as a food source for invertebrates
in vegetated areas. This is illustrated by a study of the carbon flow through
the food webs of a vegetated and an unvegetated Florida lake based on a
comparison of stable carbon isotope ratios in different biota (Hoyer et a/.,
1997). Phytoplankton was the most important carbon source in the
unvegetated lake, while periphyton was the main source of carbon in the
vegetated lake. The abundant macrophytes contributed little to the food
web in the vegetation dominated lake. The idea that periphyton is the main
food source in vegetation dominated lakes also fits with the observation that
macrofauna! distribution depends strongly on colonizable plant surface
(Dvorak and Best, 1982; Dvorak, 1987).
Although most of the invertebrates in aquatic vegetations feed on
periphyton or decomposing plants (Engel, 1988), or are predators on such
herbivores and detritivores (Dvofak and Best, 1982; Scheffer et a/., 1984),
there are also species that do not really depend on vegetation for food.
Chironomid larvae, for instance, are usually an important group in the top
layer of lake sediments both in vegetation and in unvegetated areas. None-
theless, their density is usually much higher in the vegetation than outside
(e.g. Fig. 5.4 ).
This is at least in part the result of active habitat choice of the animals.
Habitat choice has been experimentally tested in predator free aquaria in
which one half of the bottom area was planted with artificial Chara vegeta-
tion and the other half was left unvegetated (Diehl, 1988). Equal numbers of
Chironomus anthracinus larvae were introduced in both parts, but after 24
hours already twice as many animals were found in the sediment of the
vegetated part than in the unvegetated sediment. Predator avoidance is
likely to be an important ultimate reason for the fact that vegetation is the
preferred habitat even for such sediment inhabiting invertebrates. The work
of Diehl (1988) shows that the feeding efficiency of different fish species on
chironomids decreases markedly with the presence of vegetation (Fig. 5.4 ),
although the apparent sheltering effect of Chara is much greater than that of
sago pondweed.
Summarizing, submerged plant stands have higher densities of most in-
216 Vegetation
10000
8000
NE
~
..!l!
6000
"E
0
1:
e
~
b 4000
:;;
_Q
~
z
2000
vertebrate groups than unvegetated areas, due mainly to the high availabil-
ity of suitable food, and the relatively low predation pressure from fish.
Zooplankton
As mentioned earlier, vegetation stands can also be an important refuge for
Daphnia and other pelagic copepods against fish predation. The first study
indicating the importance of vegetation as a refuge for herbivorous
zooplankton against fish addressed a striking difference in transparency
between two small connected lakes in the Norfolk Broads area in eastern
England (Timms and Moss, 1984). The smallest one, Hudsons Bay, had a
large stand of water lilies, and the adjacent open water was clear. In con-
trast, the larger unvegetated Hoveton Great Broad had chlorophyll concen-
trations that were mostly an order of magnitude higher. Both lakes received
their water from the same nutrient rich river, and bio-assays showed that the
water in the clear Hudsons Bay could actually support a high algal growth.
Indeed, phytoplankton densities in Hudsons Bay were high in the spring and
autumn, but dropped strongly in the summer when the lilies were present.
Zooplanktivorous fish was present in both basins, and in Hoveton Great
Implications of vegetation 217
100
D no vegetation
IIIII Potamogeton
• Chara
-l'l 80
·e: ,.:I
g
~
.,~ 60
~
i
0
40
1i
~
0
Perch iream Roach
Fig. 5.5 Efficiency of perch, bream and roach in capturing chironomid larvae in an
unvegetated situation as opposed to the capture efficiency in artificially constructed
vegetations of Potamogeton pectinatus and Chara tomentosa. From Diehl (1988).
120
open water vegetation
100
-
ci
5 80
"'c::
"'"'E 60
·c"' 40
-"'
c.
0"' 20
-4 -3 -2 -1 0 4
Distance from edge (m)
Fig. 5.6 Distribution of Daphnia magna in along a spatial gradient across the edge of
a Potamogeton pectinatus field in the daytime (open symbols) and at night (closed
symbols). From Lauridsen and Buenk (1996).
This suggests that large homogeneous fields of aquatic vegetation are not
very effective as a daytime refuge for pelagic zooplankton as they have little
edge relative to the surface area. This is confirmed by a study in Lake
Stigsholm (DK) where diurnal fluctuations of zooplankton numbers in es-
tablished vegetation stands with diameters of 2, 10 and 25m were compared
with open water dynamics (Lauridsen et al., 1996). In the 2m stands strik-
ing diurnal patterns occurred for Ceriodaphnia spp., Bosmina spp. and
Diaphanosoma brachyurum, the decrease in density at night being mirrored
by an increase in the open water (Fig. 5.7).
In the larger vegetation stands, Ceriodaphnia and Bosmina densities
were much lower and diurnal fluctuations were found hardly at all. In
contrast, Sida crystal/ina, Eurycercus lame/latus and Simocephalus velutus
were abundant in the large vegetation stands but rare in the 2m ones and
absent in the open water. These cladocerans are known to_ be macrophyte
associated (Quade, 1969; Paterson, 1994), and are suspected to have a
considerable filtration capacity explaining at least in part the water transpar-
ency in some plant stands (Irvine et al., 1990; Jeppesen eta!., 1996). Several
species live largely attached to the plants. Therefore their abundance is
easily underestimated with standard plankton sampling techniques.
Obviously, the high concentration of filter feeding zooplankton in plant
stands leads to a high grazing pressure on phytoplankton, which already
has a low productivity between the macrophytes due to shading, low nutri-
ent availability allelopathic exudates and high sinking losses. As a result
algal density is very low and in practice mainly small fast growing algae
and bacterioplankton survive (Hasler and Jones, 1949; Schriver et al., 1995;
Implications of vegetation 219
3000-.---------------------------------.
Bosmina spp.
t;;: 2000
"*:?!"
"0
] 1000
8 11 14 17 20 23 2 5
Time (hour)
Lauridsen et al., 19%). This may explain why among the pelagic
zooplankton, species such as Ceriodaphnia spp. and Diaphanosoma
brachyurum which can filter small particles do relatively well in larger plant
stands (Lauridsen et al., 1996). In general, however, submerged weed beds
with their low phytoplankton concentrations should be an unfavourable
foraging habitat for pelagic zooplankton, and reduction of predation risk
seems the obvious explanation why the animals nevertheless concentrate in
the vegetation in the daytime. The predation risk theory is supported by the
fact that larger and thus more vulnerable species seem to show a more
pronounced migration. For instance, diurnal fluctuations of Daphnia magna
were much stronger than for the smaller D. hyalinalgaleata in Lake Ring
(Lauridsen and Buenk, 1996).
It has also been shown experimentally that it is fish that drives
cladocerans into the vegetation. Long ago Pennak demonstrated experi-
mentally that macrophytes as well as water that has been into contact with
macrophytes has a repellent effect on Daphnia (Pennak, 1973). Recent
work confirms the repellent effect of plants but shows that it can be overrid-
den by fish related stimuli (Lauridsen and Lodge, 1996). In aquaria that
were partly planted with Myriophyllum a majority of the animals stayed in
the open water, however when fish was added most of the animals preferred
the vegetated part. Plastic plants had the same effect, although not as
strongly as the real ones, indicating that not only plant odour but also the
structure itself repels the animals. On the other hand, adding fish odour
works as strongly as adding a caged fish, suggesting that it is largely
the chemical cue that drives Daphnia into the refuge. This is well in line with
the much better studied induction of die! vertical migration into the
220 Vegetation
hypolimnion of deep lakes (Dodson, 1988; Leibold, 1990; Loose and
Dawidowicz, 1994), where chemical cues from fish induce a phototaxic
response in Daphnia that drives them to the dark deep-water refuge at day
(Ringelberg et al., 1991 ).
It is not yet very clear how well submerged plants protect zooplankton
against fish predation, but several experiments indicate that the refuge
effectiveness depends among other things on the fish species and on
the density of vegetation. Laboratory experiments with artificial reed and
waterlily stands show that zooplankton consumption by juvenile roach
(Rutilis rutilis) decreases with vegetation density, whereas capture rates of
rudd (Scardinius erythrophtalmus) and perch can be enhanced by vegetation
when it is not too dense (Winfield, 1987). The positive effect of sparse
vegetation on prey capture of juvenile rudd and perch results in part from a
higher activity of the animals, which, as Winfield suggests, may be due to a
lower percepted risk of predation by larger piscivores in the vegetation.
The idea that the refuge effect depends critically on plant density is
supported by experiments in Lake Stigsholm showing that daytime aggrega-
tion of pelagic zooplankton occurs in dense but not in sparse plant stands
(Fig. 5.8).
A series of enclosure experiments (Schriver et al., 1995) with different
densities of Potamogeton plants and planktivorous fish confirms that sparse
2000
dense vegetation
1500
1000
500
-0
.s 0
"'
.!;
2000
&
.g
1500 sparse vegetation
1000
500~
0,_--r--r--.--r--r--r--.-~
8 11 14 17 20 23 2 5 8
Time (hour)
Fig. 5.8 Diurnal change in the density of Bosmina in a dense (PVI = 70%) and a
sparse (PVI = 23%) vegetation stand in the Danish Lake Stigsholm. Only the dense
stand is apparently used as a daytime refuge. From Jeppesen et al. (1996).
Implications of vegetation 221
./~~ 0
I -
~ -----,;/----~--~--
~1800
,.,
J
/
1200
---------~----~/«a_________~/~----------
.; 600
.
i
'
/' ,''
0 / //
20
40
Macrophytes (PVI)
stands are hardly protective, and that even dense vegetation cannot prevent
Daphnia and Bosmina populations from collapsing when the fish density is
too high (Fig. 5.9).
Exclosure experiments in a macrophyte bed in a Finnish lake confirm
that fish can enter even dense macrophyte stands and suppress Daphnia and
Bosmina populations there (Kairesalo et al., 1997).
In view of these results it seems surprising that even the sparse under-
water structure of lily stands could act as an effective refuge as suggested by
Timms and Moss (1984). In another study, juvenile fish have actually been
found to be more abundant in Nuphar lutea vegetation than in the open
water (Venugopal and Winfield, 1993). In practice the use of plant stands by
fish and the resulting predation pressure on zooplankton apparently varies
strongly from case to case. Importantly, vegetation is also a refuge against
predation for juvenile fish, and crowding of small fish in the vegetation
refuge may lead to high predation on zooplankton there. In fact fish preda-
tion avoidance patterns are quite similar to those observed in zooplankton.
The most vulnerable individuals (in this case the smallest ones) use the
vegetation more than the less vulnerable ones (Engel, 1988), and the use of
222 Vegetation
vegetation increases when predators are present even though the food
situation in this refuge is often inferior to that in the open water (Eklov and
Persson, 1995; Persson and Eklov, 1995; Eklov and Persson, 1996). Al-
though the discussed experimental results show that outcome of this cascade
of hide and seek is quite variable, the high abundance of juvenile fish as well
as zooplankton in vegetation stands suggests that overall survival of both
groups is usually enhanced by the availability of vegetation structure.
Summarizing, pelagic zooplankton tends to leave the open water during
the daytime and concentrate in the edge zone of plant stands to avoid
predation by fish. Juvenile planktivorous fish, however, also gather in veg-
etation to avoid predation by bigger fish. Sparse plant beds allow efficient
foraging of some juvenile planktivores and are therefore a poor protection
for pelagic zooplankton. However, other cladocerans which are closely
associated with macrophytes appear to escape predation and can be abun-
dant in plant beds even if fish density is high.
Fish
The presence of vegetation is one of the main factors structuring the fish
community of eutrophic shallow lakes (Lammens, 1989). In Europe, turbid
lakes that lack vegetation are usually dominated by cyprinids such as bream
(Abramis brama), roach and carp (Cyprinus carpio), and many of these
cyprinid dominated communities also have a relatively high density of
pikeperch (Stizostedion lucioperca) (Lammens, 1989; Persson et al., 1991).
In vegetated systems, on the other hand, perch and tench are more abun-
dant and dense populations of relatively small pike can be found (Grimm,
1983; Kipling, 1983). Because of relatively high perch and pike densities,
potential piscivory can be high in vegetated situations. But, as explained in
the previous section, vegetation also provides a refuge for juvenile fish
against predation.
The importance of vegetation in structuring the fish community is illus-
trated by a review of the response to biomanipulation in some small lakes
(Meijer et al., 1995). A drastic reduction of the fish stock ('biomanipulation')
has led to the recovery of aquatic vegetation and clear water in all the
studied cases. After the biomanipulation, fish biomass usually recovered
largely in a few years, but in the new clear and vegetated state other species
became dominant than in the previous unvegetated state. The new vegeta-
tion associated community was usually more diverse. Bream and carp be-
came less abundant and were partly replaced by roach and perch. Also the
importance of pikeperch, the main predator in the unvegetated state, was
reduced but an increase in pike and perch ensured an enhanced share of
piscivores in the total fish stock on all studied sites. The overall recruitment
of young-of-the-year tended to be higher in the vegetated situation, but the
survival of these animals to older year classes differed strongly between
species.
Implications of vegetation 223
The effect of vegetation on food availability and predation risk are
thought to be the main explanations for the markedly different fish commu-
nity in vegetated systems. The high density of invertebrates among sub-
merged macrophytes represents a potentially rich food source for most
fishes, but not all species can explore this resource equally well. For a
specialist feeder like tench that preys on various invertebrates in plant
stands and has a strong preference for snails, weed beds are obviously a
good foraging habitat. For other species the situation is less clear, but
vegetation may just tip the competitive balance in favour of the slightly
better adapted ones. Feeding of bream and roach on chironomids, for
instance, is strongly hampered by vegetation while the feeding efficiency of
perch is much less affected (Fig. 5.5). Perch is also more efficient at captur-
ing Daphnia in vegetation than are bream and roach (Winfield, 1987). Thus,
the fact that perch is especially abundant in vegetated situations whereas
bream and roach dominate in turbid lakes without vegetation is at least in
part explained by the fact that perch is better adapted to forage in vegetated
situations than are the cyprinids. Although the open water is the preferred
habitat for juvenile perch, they are better able to outcompete the cyprinids
in vegetated situations (Persson et al., 1993).
Also, the ability to use vegetation as a refuge against predators differs
with the species (Eklov and Persson, 1995; Persson and Eklov, 1995; Eklov
and Persson, 1996). The crucial importance of vegetation as a refuge has
been noted especially for young-of-the-year pike (Grimm, 1983; Grimm and
Backx, 1990; Wright and Shapiro, 1990). Cannibalism is a main source of
mortality for small individuals of this species and, in the absence of vegeta-
tion as a shelter, few survive the first year. This is in part the explanation for
the striking difference in size structure between the pike populations of
vegetated and unvegetated lakes. In lakes with little or no vegetation, pike
populations often consist of relatively large individuals, while vegetated
lakes typically have a high density of relatively small individuals. An exam-
ple of such contrasting populations is found in the adjacent St Peters Lake
and Main Lake at the Great Linford gravel-pit area (Giles, 1992). Pike are
large on average in the unprotected open water of Main Lake whereas the
vegetation beds of St Peters harbour many small individuals. The dense
vegetation stands of St Peters Lake that promote survival are apparently not
the best habitat for larger pike, and on average individual growth is much
better in Main Lake. Also the poor egg survival due to silting of the eggs
after wind resuspension of the sediment contributes to the low juvenile
numbers in the unvegetated Main Lake. Indeed, beside offering different
food and providing a refuge against predation vegetation is important as a
substrate for egg deposit for various species.
In summary, the fish community in vegetated lakes differs widely from
that of non-vegetated situations. This is largely due to shifts in the competi-
tive balance and to differences between species in the use of vegetation as a
substrate for egg deposition and as a refuge against predation.
224 Vegetation
Birds
Shallow lakes that shift from a vegetated to an unvegetated state or vice
versa show large shifts in the bird communities (Wallsten and Forsgren,
1989; Hanson and Butler, 1994b; Hargeby et al., 1994). Some spectacular
examples mentioned in the opening chapter are the lakes Veluwemeer,
Ellesmere, Titkem, Krankesjiin, Linford, Tiimnaren and Christina. The spe-
cies that are involved are different from case to case but the general trend is
usually the same. In vegetation-rich lakes large numbers of migrating swans,
coots and ducks come to forage on vegetation and its associated inverte-
brates, while piscivores such as grebes forage on the fish. If the vegetation
disappears only piscivorous birds remain abundant (Fig. 5.10).
The large numbers of migrating birds visiting vegetated lakes in the
autumn and winter can be especially spectacular (Wallsten and Forsgren,
1989; Hanson and Butler, 1994b) and switches between macrophyte domi-
nance and phytoplankton dominance in shallow lakes are often noted by the
conspicuous change in bird abundance first by visitors. The developments in
Lake Christina are a good example of the effect of macrophytes on lake use
by migrating waterfowl (Fig. 1.9). Vegetation in the lake has been depressed
from the late 1970s till the late 1980s and this is reflected by a tremendous
drop in the autumn counts of ducks and coots that can be as high as 160 birds
ha-1 in years when vegetation is abundant.
Summer breeding populations never reach such high densities. Typically
only a few coots or ducks per hectare are present during the summer
1975-1984 1988-1991
Herbivorous Diving Fish feeding
waterflow! ducks waterflow!
Fig. 5.10 Schematic description of the trophic web before (1975-1984) and after
(1988-1991) the shift from a turbid to a clear state in the Swedish Lake Krankesjon.
Boxes represent biomass, arrows represent energy flow. The estimation of
invertebrate-feeding fish is uncertain. From Hargeby et al. (1994).
Effect of vegetation on turbidity 225
(Hargeby et al., 1994; Perrow eta/., 1996; S!ilndergaard eta/., 1997). Shifts are
usually observed in these breeding bird populations when vegetation be-
comes dominant (Fig. 5.11).
Not only plants but also invertebrates are used as a food source by many
species. Invertebrate food is especially important to ensure a proper protein
level in the diet of young ducks (Street, 1977), and duckling survival has
been shown to increase with invertebrate abundance (Hunter eta/., 1986;
Hill et al., 1987). Thus the high invertebrate densities in vegetated lakes may
be a major factor determining their suitability as a reproduction habitat for
waterfowl.
Abundance and diversity of the resident bird community does not neces-
sarily increase with vegetation coverage. Censuses in 46 Florida lakes re-
vealed a shift in species composition with macrophyte abundance from
mainly piscivorous to vegetation associated birds, but total bird numbers or
species diversity did not change systematically with vegetation abundance
(Hoyer and Canfield, 1994). However, many examples show that the au-
tumn use by migrating waterfowl can differ by as much as two orders of
magnitude between vegetated and non-vegetated situations.
40 400 100
30 300 80
~
~ 20 200 g
u 160
z
::;;
10 100 40
Fig. 5.11 Change in the summer populations of mute swan (Cygnus olor}, coot
(Fulica atra) and diving and dabbling ducks in the Swedish Lake Krankesjon during
the switch from the turbid to the clear state. From Hargeby et al. (1994).
226 Vegetation
macrophytes (see Chapter 1). Analyses of the relationship between the
transparency and the macrophyte abundance in large sets of lakes confirm
that there is a systematic correlation. Danish work (Jeppesen et al., 1990a),
for instance, shows that in a large set of shallow lakes the ones with a high
cover of submerged macrophytes have a higher transparency than compara-
ble lakes that lack dense vegetation (Fig. 5.12).
The Danish result is based on a presence-absence qualification of (exten-
sive) vegetation. A more detailed view is suggested by the analysis of a
group of 84 Dutch shallow lakes where a quantitative estimate of vegetation
coverage was available (Fig. 5.13).
An interpolated surface through these data shows a gradual decline in
chlorophyll with the amount of submerged vegetation. In the absence of
vegetation, the classic increase of chlorophyll with total-P concentrations is
found, but in lakes with a high coverage of submerged plants chlorophyll
hardly increases with the phosphorus concentration.
A very similar pattern is obtained by regression analysis of data from 32
Florida lakes (Fig. 5.14) (Canfield et al., 1984).
The authors showed that the existing regression model for predicting the
chlorophyll concentration from total-P and total-N could be improved sig-
nificantly by adding the percentage of the lake volume occupied by aquatic
macrophytes as an explanatory variable. The resulting model was tested
using a series of data from Lake Pearl, where vegetation was eliminated by
a two-year period of repeated treatments with herbicides and grass carp.
Indeed, with the exception of some algal peaks, the resulting increase in
algal biomass could well be explained by the regression model (Fig. 5.15).
5-r--------------------------- --------------------.
"' "'
"' "'.ti'
:0.~
', 6.
"'
A
1-
0
'':.~>~~;~}:-i~:-~t··,,-.~--~-.--~-·.: __~-:-•:----·--·;·-·-----.·
I I I I
0 100 200 300 400 500 600 700 800 900 1000
Total phosphorus (~g P 1- 1)
Fig. 5.12 Summer mean transparency (Secchi-depth) in relation to the total phos-
phorus concentration of the lake water in shallow Danish lakes with (triangles) and
without (dots) substantial aquatic vegetation. From Jeppesen eta/. (1990a).
Effect of vegetation on turbidity 227
200
150
~
1
t 100
~
.9
.<::
u
50
0
0
'~
Fig. 5.13 The concentration of chlorophyll-a plotted against the total-P concentra-
tion and submerged macrophyte coverage in 84 Dutch shallow lakes. The surface is
interpolated through the data points. In lakes with a high coverage of submerged
plants chlorophyll hardly increases with the phosphorus concentration.
Fig. 5.16 Three sets of causal relations that could explain the negative correlation
between vegetation and turbidity abundance in the field (see text).
the spring when overwintering structures such as tubers, turions, and seeds
resprout or germinate. In the autumn growth ceases, plants become senes-
cent and eventually decompose.
Several studies show a conspicuous inverse correlation of this vegetation
cycle with the seasonal dynamics of phytoplankton. Goulder (1969), for
instance, describes the contrasting patterns observed in two English gravel-
pits (Goulder, 1969). In one pond, chlorophyll-a concentrations dropped
dramatically during the summer when a dense vegetation of Ceratophyllum
demersum developed, while upon disappearance of the macrophytes in
early winter phytoplankton peaked again (Fig. 5.17).
In the other pond where macrophytes were scarce, no such depression in
summer phytoplankton was observed. Similarly, Norwegian workers analys-
ing data from 24 small Norwegian lakes found that lakes with low
macrophyte cover have high phytoplankton biomass throughout the sum-
mer, while in lakes with a high cover of macrophytes, the diatom biomass in
the spring declines sharply in early summer when macrophyte biomass
increases (Mjelde and Faafeng, 1997).
A caveat of comparisons between lakes is that they usually differ in many
aspects, so that the link to vegetation as a cause of clear water in the summer
is not entirely convincing. Therefore, it is interesting to see that similar
differences in seasonal dynamics are observed between vegetated and open
areas within lakes. A good example is the situation in Veluwemeer (see also
Section 1.1). Part of this large shallow lake (3300ha, 1.4m mean depth) is
covered with dense Chara fields, and in summer water in these vegetation
stands is crystal clear, whereas in the unvegetated part of the lake Secchi-
depth is only about O.Sm (Scheffer eta/., 1994b; Van den Berget al., 1997).
Seasonal dynamics of turbidity inside and outside the vegetation show con-
trasting patterns (Fig. 5.18).
In both parts of the lake a clear-water phase occurs in the spring but in
the unvegetated area turbidity becomes high again in the summer whereas
inside the Chara stands the water remains clear throughout the vegetated
Effect of vegetation on turbidity 231
100
Chlorophyll-a
80 100
I ' -
~ Vegetation
80
1
b), 60
s * c
0
1
:;, 60
~bll
_c
"- ~
E! 40 0
0
:;: [;
u >
40 8
20
j
20
F M A M A
period. Only upon senescence of the plants in the autumn, turbidity in-
creases again in the vegetated parts.
A similar pattern has been described for a Czechoslovakian fish culture
pond (Pokornyet at., 1984). This old (1513 AD) shallow (1.5 m mean depth)
pond is small (7ha) compared with Veluwemeer. Nonetheless in summer a
pronounced difference between the turbidity of the water exists between a
dense vegetation stand (mainly Elodea canadensis) and the non-vegetated
area. Chlorophyll concentrations in the open water are around 120 pgr 1
whereas in the vegetation chlorophyll contents drop sharply in the spring
and stay around 5 11g r 1 for most of the vegetated period of the year. On an
even smaller scale I have observed an intense local phytoplankton bloom in
an unvegetated spot with a diameter of only 2m in an otherwise clear and
vegetated pond with a diameter of just Sm.
Seasonality of submerged vegetations depends strongly on the climate. In
the mild winters of Florida, for instance, submerged plants stay present and
productive. Seasonality also varies with the species. As a result, a shift in
5.50 100
4.50 80
g
"'~
c
0
3.50 60
~
~ 5
~ 2.50
~w
40
~
>
1.50 20
0.50 0
M A M A 0 N
5.50
Tubidity
4.50 unvegetated site
g 3.50
c
0
~
1!
~ 2.50 i
~
~w
l
> 1.50
0.50
M A M
'~ ~ A 0 N
Fig. 5.18 Seasonal dynamics of the vertical light attenuation coefficient and the
estimated contribution of different seston fractions to it in the water column inside
a Chara field (left-hand panel) and in the open water (right-hand panel) of Lake
Veluwemeer. The line in the left-hand panel indicates the seasonal development of
vegetation cover in the field. From van den Berget al. (1997).
Effect of vegetation on turbidity 233
dominant species may affect the duration of vegetation presence and asso-
ciated clear water conditions as observed in the Dutch Lake Zwemlust (Van
Donk and Gulati, 1995). In the first years after fish stock reduction, winter-
green Elodea nuttallii vegetation dominated the lake and phytoplankton
biomass remained low during all the seasons. In the following years
dominance shifted to Ceratophyllum demersum and later Potamogeton
berchtoldii. The shorter growing season of these species was associated with
the occurrence of spring and autumn phytoplankton blooms (Fig. 1.13).
- limited by light
all not limited
c:::::J limited by zooplankton
c:::::J limited by nitrogen
t biomanipulation
Fig. 5.19 Seasonal changes in factors limiting the phytoplankton growth in the
Dutch Lake Zwemlust before (1986) and after (1987-1990) reduction of the fish
stock. Vegetation became abundant from 1988 onward. Redrawn from van Dank et
a!. (1993).
234 Vegetation
pattern in the vegetation dominated state fits well with those reported in
the studies described in the previous section. The sequence of mechanisms
that causes the pattern in this case is thought to be as follows: unlimited
phytoplankton growth occurs at the end of winter when light is no longer
limiting, and this spring bloom is followed by a period of severe zooplankton
grazing. So far, this is the classical spring clear-water phase scenario. How-
ever, when vegetation appears, nitrogen availability drops sharply, and bio-
assays reveal that this is the main limiting factor for phytoplankton in the
summer. In the autumn when macrophytes become senescent, release of
nutrients leads to a short period of unlimited phytoplankton growth again,
followed by a second grazing limited episode. The spring clear-water phase
(Meijer et al., 1994a) and the autumn release of nutrients causing a
phytoplankton peak at the end of the growing season (Landers, 1982) are
probably quite common in vegetated lakes. The mix of mechanisms respon-
sible for suppressing phytoplankton in the summer may differ from case to
case, as demonstrated by other case studies.
The causes of the striking contrast in Veluwemeer between clear Chara
fields and turbid open water have also been analysed (Van den Berget al.,
1997). Regulation of algal growth has not been tested with bio-assays in this
case, but the seston composition and sedimentation rates have been re-
corded in some detail, allowing inferences about the regulating processes. In
this large exposed lake, turbidity is caused not only by phytoplankton, but
also to a large extent by resuspended inorganic sediment particles and
detritus (Fig. 5.18). The concentration of all of these seston fractions is lower
inside the Chara fields than in the open water, but differences are especially
pronounced for large particles. Since larger particles sink faster in general,
this suggests that sedimentation in the absence of wave resuspension is an
important reason why the water is clear in the Chara fields. This is confirmed
by the sedimentation records. In the open water sedimentation is as much as
100 g DWT m-2 d- 1, whereas in the centre of the Chara field practically no
deposition is measured in sediment traps despite the fact that some seston
remains present here. Apparently this remaining seston is hardly prone to
sedimentation. The same picture arises from the shift in phytoplankton
composition over a gradient from the open water into the vegetation (Fig.
5.20). In the centre of the Chara bed motile flagellates that can swim up to
prevent settlement dominate.
Cladoceran density is lower in the Chara fields than in the open water in
the summer but to have an idea of the potential importance of zooplankton
for phytoplankton regulation, it is necessary to relate the consumption by
the animals to the available amount of algae. A simple way to estimate
the potential grazing pressure is to assume that cladocerans can consume a
daily amount of phytoplankton equal to their own body weight (Schriver
et al., 1995; Jeppesen eta/., 1996). It then appears that despite relatively low
numbers of Daphnia and Bosmina, their potential daily grazing capacity is
estimated to about 10 times the total phytoplankton biomass in the veg-
Effect of vegetation on turbidity 235
18 July 1995
~
100
•
Ill
Others
Flagellates
c:
0
"'·~ 80
1;!1 Cyanobacteria
0..
E D Diatoms
8 60
p
c: D Greenalgae
~c:
cy
"'
c..
.8
40
£0..
>
~
" 20
,
i
~ I
100 100 50
Chara cover(%)
eta ted area. Despite the fact that the animals will probably consume alterna-
tive food as well, it seems likely that the realized grazing pressure on
phytoplankton is high enough to suppress populations of motile flagellates
and other algae that are able to overcome sinking losses in the Chara fields.
The border between clear and turbid water usually coincides precisely
with the border of the Chara fields, indicating that mixing between the open
water and the vegetation stands is slow relative to the clearing processes in
the vegetation. Nonetheless, heavy winds can sometimes cause turbid water
to enter the vegetation. Interestingly the water in the fields clears up after
such a mixing event in less than a day (Scheffer et al., 1994b), showing that
the processes responsible for clearing the water are fast. Again this fits well
with the idea that sedimentation is a dominant process (see Chapter 2).
Water depth in the Chara stands is only 30-80cm, and the mixed layer of
water above the dense canopy is a few decimetres only. Since the average
sinking velocity of a particle is something in the order of 1m per day, much
of the seston may be expected to settle in a matter of hours. As argued
zooplankton grazing may deal with the remaining non-settled algae.
How fast the water column in vegetation stands can be cleared is also
illustrated by a study in the freshwater tidal Potomac river (1990). Here,
chlorophyll concentrations are found to be up to seven times lower in dense
Hydrilla stands than in the open water during low tides. At high tides
differences are much less pronounced, indicating that semi-diurnal mixing
236 Vegetation
with changing tides is considerable, and that apparently algal loss processes
in the vegetation are fast enough to clear up the water considerably in a
matter of hours. There was no difference between photosynthetic rates of
phytoplankton in the water from inside or outside the weed beds, indicating
that severe nutrient limitation or allelopathic effects were not the cause
of the observed differences in chlorophyll concentrations. This leaves
zooplankton grazing, sedimentation and shading by the plants as possible
explanations.
A Danish enclosure study designed to study the refuge effect of vegeta-
tion for zooplankton (Fig. 5.9} also provides some insight into the possible
role of zooplankton grazing in reducing phytoplankton biomass in weed
beds (Schriver et at., 1995). In this case the main plant species are
Potamogeton pectinatus, P. pusillus and Callitriche hermaphroditica.
Phytoplankton biovolume in the enclosures decreases with vegetation abun-
dance (measured as plant volume infested), but this effect is less pro-
nounced when fish predation has eliminated Daphnia and Bosmina
populations (Fig. 5.21).
Estimates of the potential grazing pressure of these cladocerans indicate
Fig. 5.21 Phytoplankton biovolume in relation to the vegetation density (PVI) and
the biomass of planktonic cladocerans (Daphnia and Bosmina). From Schriver et al.
(1995).
Effect of vegetation on turbidity 237
that their maximum daily consumption can be as much as 36 times the
available amount of phytoplankton in the vegetated enclosures. Although
other food sources such as detritus and periphyton apparently sustain
cladoceran density at such high levels, grazing pressure on the preferred
food, phytoplankton, is obviously severe under these conditions. Nonethe-
less, the fact that phytoplankton biovolume also decreases with vegetation
abundance when these cladocerans are absent indicates that grazing
by Daphnia and Bosmina is certainly not the only process involved (Fig.
5.21).
The importance of zooplankton grazing is also supported by the study by
Timms and Moss (1984) who found clear water in the waterlily dominated
Hudsons Bay, but not in the unvegetated Hoveton Great Broad receiving
water from the same nutrient rich source. Bio-assays ruled out the possibil-
ity of nutrient limitation, and grazing by high numbers of large plant associ-
ated and pelagic cladocerans was probably the reason why chlorophyll
concentrations in Hudsons Bay were less than 10% of those measured in the
adjacent Hoveton Great Broad.
Shading
Although the role of shading by macrophytes in reducing phytoplankton
productivity in plant beds is rarely mentioned, it may explain reduced algal
abundance in vegetated areas at least in part as remarked already by Wetzel
in his textbook on limnology (Wetzel, 1975). Light attenuation in the veg-
etation is among other things a function of plant biomass. Ikusima (1970)
measured the specific attenuation coefficient of various plant types, and
reports values of about 0.001m2g-'. For submerged vegetation with a
biomass of say 500 g m_, this implies that even in crystal clear water less than
1% of the light that enters the water will reach the sediment (Ir/10 = e-0·00 '"500
= 0.0067). How much shading will be experienced by phytoplankton de-
pends on the vertical distribution of plant biomass, but clearly it is not an a
priori negligible factor. Measurements of the verJical light gradient in a
dense Elodea bed, for instance, show that irradiance can be reduced by
more than 95% within the upper 20cm of the water column (Pokornyet al.,
1984).
238 Vegetation
Nutrient limitation
Phosphorus availability in the water column may be reduced due to uptake
by macrophytes (Kufel and Ozimek, 1994) but the majority of the studies
show unaltered or even increased ortho-phosphorus levels (Moss et al.,
1990; Van Donk et al., 1993; Perrow et al., 1994; Van den Berget al., 1997).
In contrast, very low inorganic nitrogen concentrations in the water column
of vegetation stands are frequently found (Goulder, 1969; Van Donk et al.,
1993) and the importance of nitrogen as a limiting factor for algal growth
has been confirmed in bio-assay experiments (Van Donk et al., 1993).
Low nutrient levels in vegetation stands may be due to uptake by plants but
also to uptake by periphyton and in case of nitrogen by denitrification.
Nonetheless, nutrient limitation has also been excluded as a possible expla-
nation in various studies were reduced algal densities are observed among
macrophytes (Pokorny et al., 1984; Timms and Moss, 1984; Jones, 1990;
Schriver et al., 1995).
Allelopathy
Few studies indicate more than marginal allelopathic suppression of
phytoplankton by macrophytes in natural situations, but the results so far
indicate that cyanobacteria have a relatively high sensitivity to allelopathic
exudates (see Chapter 3). Indeed the relative share of cyanobacteria in the
phytoplankton community of weedbeds is often low (Hasler and Jones,
1949; Timms and Moss, 1984; Schriver et al., 1995; Van den Berget al., 1997),
but at least one study indicates that this is due to zooplankton grazing rather
than plant exudates (Schriver et al., 1995).
Resuspension prevention
Although several studies address the alteration of the sedimentation
resuspension cycle of seston by macrophyte beds, only the Veluwemeer
study (Van den Berget al., 1997) indicates the importance of this mechanism
for the clearing effect of vegetation stands explicitly. Indeed in dense Chara
stands in shallow water only motile or buoyant algae are likely to survive,
as estimated sinking losses of more than 100% d-1 are unlikely to be com-
pensated by growth. Although potential sinking losses in these Chara beds
may be extreme, many studies demonstrate reduced resuspension in the
presence of vegetation structure (Jackson and Starrett, 1959; Dieter, 1990;
James and Barko, 1990; Petticrew and Kalff, 1992). In view of the rapid
settling losses in shallow water, the presence of non-living suspended parti-
cles and most phytoplankton groups depends critically on resuspension,
as explained earlier. Whether or not motile algae that are not affected
by settling losses in dense vegetations can build up high biomasses depends
on other factors. Note that one such factor is the residence time of water in
Effect of vegetation on turbidity 239
the vegetation field. In some situations, residence time may be long
enough to allow settling of particles, but too short to allow an alterna-
tive phytoplankton community to build up. The clearing of water in vegeta-
tion stands in the tidal Potomac river could potentially represent such a
situation.
Zooplankton grazing
Several studies indicate that the amount of zooplankton in weed beds
should be able to control the sparse phytoplankton populations present
(Timms and Moss, 1984; Schriver et al., 1995; Van den Berget al., 1997).
Even in relatively dense vegetation, however, planktonic cladocerans can be
driven to extinction by fish predation (Schriver et al., 1995; Kairesalo et al.,
1997), and phytoplankton density can be reduced in the vegetation despite
the absence of noticeable Daphnia and Bosmina populations (Schriver eta[.,
1995). Plant associated cladocerans that hook up to the macrophytes may
play a role in reducing algal biomass. These animals seem to be less vulner-
able to fish predation (Kairesalo et al., 1997) and their density is probably
underestimated in standard sampling procedures. Nonetheless it is clear
that grazing is just part of the story, and other factors may be more impor-
tant in controlling phytoplankton at least in some cases (Van Donk et al.,
1993; Moore et al., 1994).
dA
-=rA
dt K
A)
( 1-- -g Z--+!
A
' A+h,
·(K-A ) (1)
: zooplankton
!
Fig. 5.22 Effect of zooplankton biomass and algal carrying capacity (K} on the
equilibrium biomass of phytoplankton as predicted by a minimal model (Eq. 1).
Since algal carrying capacity (K) decreases with vegetation biomass the lower left-
hand axis may be interpreted as representing the effect of vegetation. The projection
on the horizontal plane (bottom) shows that at high vegetation density less
zooplankton is needed to cause over-exploitation of phytoplankton leading to clear
water.
Brackish lakes
4
g3
Fig. 5.23 Relationship between the total-P concentration of the lake water and
Secchi-depth in a brackish Danish lakes with (open symbols) and without (dots)
substantial aquatic vegetation. Unlike in freshwater lakes (Fig. 5.13) vegetation does
not make a difference fer water clarity in brackish situations. From Jeppesen et aL
(1997). .
1994; Jeppesen et al., 1997). A Danish survey confirms that the connection
between water clarity and submerged vegetation characteristic for fresh
shallow lakes (Fig. 5.12), is not found in brackish systems (Fig. 5.23).
Changes observed in the Dutch Lake Volkerak-Zoommeer (5000ha av-
erage depth Sm) support the idea that the turbidity of brackish lakes is
causally related to their salinity. This lake changed from brackish to fresh in
about a year due to flushing with river water following isolation from the
marine tidal Oosterschelde. Despite the high nutrient loading, transparency
increased spectacularly when the lake became fresh, and submerged
macrophytes expanded (Schutten et al., 1994). Grazing by Daphnia was the
main factor controlling algal biomass in the lake. Later, colonization by
freshwater fish species changed the situation to one more typical for
eutrophic lakes (Ligtvoet and Grimm, 1992; Ligtvoet and de Jong, 1995).
Perch was the first species to expand, and individuals grew extremely rapid
on a diet of mysid shrimps and Daphnia (Houthuijzen et al., 1993). Only in
the fifth year when the first generation of roach that invaded as juveniles
became adult did cyprinid fry become abundant. In that year D. pulex was
replaced by the smaller D. galeata and transparency decreased from three to
one metre. In view of the high nutrient load the clear state is probably
transient in this lake, but the fact that Daphnia became abundant and
cleared the lake after freshening suggests that the turbidity of brackish state
may have been due to absence of such phytoplankton grazers.
A four-year study of a Danish lake that shifts between a brackish (1-3
p.p.thousand) and a slightly brackish state (0.5-1 p.p.thousand) supports
this view (Jeppesen et al., 1997). In years with higher salinity, chlorophyll-a
Effect of vegetation on turbidity 243
- 350.,-------------, D spring
'i.. 300
.1t 250 • summer
!! 200 • autumn
t 150
~ 100
ii 50+-L----,...C:
3,5
g
t
2,5
1,5
1
0,5
0
..E
~
500
400
i
300
200
100
}j
.8
3,5
i 2,5
f 1,5
L1
0,5
0
1986 1993 1994 1995
Fig. 5.24 In years with a higher salinity (lower panel) the concentrations of chloro-
phyll-a and total phosphorus are elevated and Secchi-depth is low in a shallow
coastal Danish lake. From Jeppesen et al. (1997).
levels are much higher and Secchi-depths lower than in fresher years (Fig.
5.24).
Zooplankton composition in earlier years is not well documented, but the
shift from turbid to clear in 1995 appears to be related to a strong increase
in Daphnia numbers, as was observed in the Dutch Volkerak-Zoommeer.
Interestingly, total-P levels are also elevated in the turbid years (Fig. 5.24)
despite a relatively constant external loading. This is in line with the view
presented in Chapter 3 that total-P concentration can in part be caused by
high phytoplankton biomass, although causality behind their correlation
is usually interpreted the other way round. Sulphide formation in the
sediments of brackish water (which contains more sulphur than fresh water)
244 Vegetation
may be another explanation of elevated phosphorus concentrations as it can
cause part of the iron to become unavailable for phosphorus immobilization
because Fe(II) is removed from the pore water due to precipitation with
sulphide as insoluble PeS (see Chapter 2).
Comparison of the zooplankton communities of many brackish and fresh
lakes confirms that substantial Daphnia populations are usually found
hardly at all at salinities higher than 2-4 pp thousand (Jeppesen et al., 1994;
Moss, 1994). Instead, zooplankton in these situations is dominated by less
efficient filter-feeders such as the copepods Eurytemora spp. and Acartia
spp. and rotifers (Jeppesen et al., 1994). Daphnids are known to be rather
intolerant to salinity, and this may well be the main reason for their absence
in brackish situations. An additional explanation may be that predation
rates on zooplankton can be high in brackish lakes (Jeppesen et al., 1997).
The fish community is usually dominated by sticklebacks (Gasterosteus
aculeatus and Pungitius pungitius). These animals may spawn several times
during the summer leading to continuously high numbers of planktivorous
juveniles. Since sticklebacks enter the vegetation, the refuge function of
submerged plants for zooplankton in freshwater lakes may not work well
against stickleback predation.
Mysid shrimps are another important component of the food web in most
brackish lakes. This may be due to an adaptation to salinity per se, but the
outbreak of Neomysis after strong reduction of the fish stock in Lake
Wolderwijd (Meijer et al., 1994a) shows that they can do well in fresh water.
Possibly, the relative scarcity of larger planktivorous fish that prey on such
relatively large invertebrates is an important factor explaining the abun-
dance of mysid shrimps in brackish lakes. Mysids are omnivores that feed on
detritus and periphytic and benthic algae but they are also known to con-
sume zooplankton up to relatively large sizes (Chigbu and Sibley, 1994),
which may be another stress factor to Daphnia populations in brackish
lakes. Also, total-P levels may be elevated in the presence of Neomysis
integer because excretion by these largely benthic feeders represents a net
nutrient flow to the water column (Jeppesen et al., 1997) a mechanism
studied more extensively for benthivorous fish (Section 2.3).
Thus, predation pressure on Daphnia and nutrient regeneration by
Neomysis may in part explain the turbidity of brackish lakes. However, the
fact that daphnids respond more rapidly to decreasing salinity than the rest
of the food web suggests that salinity per se is the main explanation for their
absence in brackish water and the corresponding lack of top-down control
of phytoplankton.
Note that the observations in brackish lakes also seem to suggest that
cladocerans are crucial for causing water to clear up in aquatic vegetation. A
caveat in this reasoning is that vegetation in brackish lakes is usually domi-
nated by sago pondweed (Potamogeton pectinatus), a species that co-exists
with turbid conditions in freshwater lakes too. In the lakes Krankesjon
and Veluwemeer, for instance, the water becomes clear only after sago
The regulation of vegetation abundance 245
pondweed vegetation is replaced by Chara fields. The reason why sago
pondweed seems to have so little impact on turbidity is unclear. Possibly it
just provides a poor refuge against predation. but other factors such as its
canopy forming structure in combination with its typically low biomass may
also imply relatively little effect on nutrients. light and resuspension com-
pared with other plants.
(3)
c
sd ~ ~ (4)
where c" is the so-called Poole Atkins coefficient. Although this relationship
is far from accurate (see Chapter 2) it suggests a linear relationship between
z""" and transparency:
(5)
13
12
-•
~ 11
11,., 10
9
•
..<:
c.
e:;!
8
E 7
'0 6 •••
]
~
....~
4
3
2
'.• • • •
•
0.2 0.4 0.6 0.8 1.0 1.2 1.4
Km;n<m-1)
Fig. 5.25 Relationship between the maximum depth at which macrophytes occur
(zm~l and the vertical attenuation coefficient of the most penetrating colours that
can be used for photosynthesis (Em.,) in a set of 15 UK lakes. Redrawn from Spence
(1982).
The regulation of vegetation abundance 247
where c, is a constant depending on the Poole Atkins coefficient, the incom-
ing radiation (10) and the critical light level needed by the vegetation (!"").
Although this linearity is indeed confirmed remarkably well by some data
sets, regression lines tend to intersect the y-axis (Fig. 5.26), suggesting that
even under extremely turbid conditions, submerged plants can grow at
shallow sites.
This can be represented by the formula:
(6)
where z, is the maximum depth that can be colonized by submerged plants
even under the most turbid conditions.
A simple explanation for this phenomenon is that it is not the light
reaching the bottom that counts but the light reaching the canopy of the
vegetation. This would imply that the effect of turbidity on plant growth in
shallow lakes should critically depend on the growth form. Plants that stay
low, such as many charophytes, depend more on a clear water column than
plants that grow tall and concentrate much of their leaves just under the
water surface. Indeed, canopy forming species such as Potamogeton
pectinatus and Hydrilla verticillata usually dominate the vegetation in turbid
shallow water, and there is a systematic increase of canopy forming species
with the nutrient level (Chambers, 1987; Moss, 1988). Also, some plants
respond to low light levels by extensive shoot elongation (Barko and Smart,
1981; Tanner et al., 1993). It might be argued that the light that reaches the
bottom remains crucial, as plants usually start their growth from the sedi-
ment surface in the spring. Many species in turbid lakes, however,
overwinter in the form of vegetative underground structures. Such tubers
]: 6
t
.r::
c.
5
el;! 4
E
b 3
~ 2
:;;
~
-'
0
0 2 3 4 5 6 7 8 9
Secchi-depth (m)
Fig. 5.26 Linear increase of the maximum depth inhabited by macrophytes with
Secchi-depth in a set of 27 Finnish lakes. Redrawn from Hutchinson (1975).
248 Vegetation
and rhizomes contain enough energy storage to support early growth in the
spring even under low irradiance levels (Hodgson, 1966). Once the upper
few decimetres of the water column are reached, the plants have escaped the
effects of turbidity. Shoots spread horizontally, and photosynthesis is largely
restricted to the upper layer of the water.
The idea that it is this morphological trait that allows persistence in highly
turbid shallow water also fits with the shifts in dominant growth form that
are observed over the past decades in Veluwemeer (see Chapter 1). In the
early years when the lake was clear, low growing Chara meadows covered
the bottom over large areas. With increasing turbidity, however, Chara
disappeared and the canopy forming Potamogeton pectinatus became the
dominant species. Recently, with increasing transparency, Chara has again
replaced Potamogeton pectinatus in large part. Similar patterns are reported
for Swedish lakes (Blindow, 1992b).
Surprisingly, charophytes are known to be quite shade tolerant, and
indeed in deep lakes they usually grow to greater depths than angiosperms
(Hutchinson, 1975; Spence, 1982). Their absence in turbid lakes has long
been ascribed to a toxic effect of high phosphorus levels (Forsberg, 1964;
Forsberg, 1965). However, regressions of the maximum inhabited depth
against transparency support the view that it is light limitation due to their
low growth-form that limits their occurrence in turbid shallow lakes (Fig.
5.27). The high intersection with the vertical axis explained by the canopy
formation of angiosperms is absent for charophytes.
In practice, the simple Zm,x regression models are not very useful for
describing the response of vegetation to changes in transparency in turbid
shallow lakes. A useful data set to show this is the series of vegetation maps
gathered over a period of 20 years for the Randmeren. Annual routine
mapping started in the early 1970s because the abundant vegetation in
the eutrophic lakes had become a nuisance. Soon after, the ongoing
eutrophication reduced vegetation to a few sparse stands of pondweed
Secchi·depth (m)
Fig. 5.27 Schematic relationship between Secchi-depth and the maximum depth
inhabited by charophytes and angiosperms. From Blindow (1992a).
The regulation of vegetation abundance 249
(Potamogeton pectinatus). Mapping, however, has been continued ever
since and also covers the gradual recovery of vegetation with improving
water quality that started in the late 1980s. Since maps of lake depth,
sediment characteristics and data on water quality and the abundance of fish
and birds over that period are also available, a statistical analysis of the
possible impact of these factors on the vegetation could be made (Scheffer
eta/., 1992). To do this, 5% of the map surfaces were sampled at random in
50 x 50m blocks. The presence of plants in these blocks was scored and
related to the environmental factors.
As expected, the analysis suggests that turbidity and depth are the most
important factors explaining vegetation abundance. However, although
most of the plant beds are found in water shallower than 1m, the probability
of finding vegetation decreases with depth almost asymptotically, and no
real limit is observed (Fig. 5.28).
Healthy plant stands are found at depths down to 3m, even though
Secchi-depth is less than 0.5 m. Model simulations (Scheffer et a/., 1993a)
suggest that this can be explained by dispersal of diaspores from productive
plant stands at shallower sites (Fig. 5.29).
The potentially large effects of propagule dispersal on plant distribution
have also been reported in terrestrial plant ecology. A detailed study of the
population regulation of the dune annual Cakile edentula, for instance,
revealed that landward migration of seeds sustains substantial populations
in areas where the species would otherwise not be able to persist
(Watkinson, 1985; Watkinson and Davy, 1985). Obviously, variation in
100
l/1-
1lc:
~
:s
tl0
"E 50
-[
0
~
~
e"'a.
.0
0
0 0.5 1.0 1.5 2.0
depth (m)
isolated sites
"seed" input= 5 g DW m-2 year-1
400
~
:;;:
Cl
!!9 300
J
~ 200
~ 100
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
depth (m}
Fig. 5.29 Simulated effect of an input of seeds or other propagules on the equilib-
rium biomass of submerged vegetation over a range of water depths. From Scheffer
et al. (1993a).
Periphyton
The practically floating canopies of Potamogeton pectinatus seem an almost
perfect escape from the shading effect of turbid water. Indeed, the
The regulation of vegetation abundance 251
1.o
o.a
~
-~
\:l
~ 0.6
....Cl.:
C)
~
~ 0.4
!:;:
8
C)
0 ..2
•..;>
Randmeren study shows that even under very turbid conditions plants stay
present down to about 1m depth. However, despite the growth form, abun-
dance of this species does decrease with turbidity even in the shallow areas
(Fig. 5.30).
Obviously, this decrease does not necessarily imply a direct causallinl{.
The correlation may well have more indirect reasons (cf. Fig. 5.16). Turbid-
ity increases with the nutrient loading, but so do other things that may affect
the performance of submerged plants such as fish density and the growth of
the periphyton covering the plants. To check if light was limiting the growth
in the pondweed vegetation of the Randmeren, shade cloth of different
252 Vegetation
t:tJ\1/\_r- 1··- 1
~ A 1 M 1 J 1 J 1 A 1 s 1 o 1 A 1 M 1 J 1 J 1 A 1 s 1 o 1 ~~s 1 o 1
0
Month
Wave action
Wave action can be an important factor for plants in shallow wind-exposed
lakes even if the sediment is solid enough to prevent uprooting. The sea-
sonal cycle of submerged plants, for instance, may differ with the degree of
exposure. In the large, wind-exposed Randmeren the entire aboveground
part of the vegetation is swept away with the first autumn storms. Van Wijk
(1988) who compared the life cycle of sago pondweed vegetation in the
Randmeren and several contrasting lakes, found that the growing season is
The regulation of vegetation abundance 255
shorter in the exposed Randmeren than in any other studied case. In small
protected water bodies the species can even be perennial. Interestingly, the
moment at which the vegetation is wiped out in the autumn has been shown
to be related to the degree of eutrophication in the Randmeren. The grow-
ing season was longer in early years when the conditions were better. When
eutrophication proceeded, lakes that were less affected kept vegetation
longer at the end of the summer (Leentvaar, 1966). Also in the in situ shade
experiments, heavily shaded plots lost aboveground biomass earlier at the
end of the growing season (Van Dijk and Van Vierssen, 1991). Wave action
thus seems to shorten the seasonal period over which submerged vegetation
is present on exposed sites, and more so if growth conditions are bad due to
light limitation.
Although vegetation abundance and species richness is often found to be
better on sheltered sites (Spence, 1982), this is not necessarily the case in
shallow lakes. In the Randmeren, even though the length of the vegetation
period is limited by wave action, presence of submerged plants in summer is
correlated positively with exposure (Fig. 5.32).
Apparently, positive effects of factors related with exposure overrule the
effect of wave damage. One possible positive effect is that wave action
partly removes the periphyton layer from the plants, whereas at sheltered
sites, deposition of suspended solids on the vegetation only adds to the
periphyton complex. Indeed, in Lake Krankesjon where vegetation abun-
dance is also found to be reduced at sheltered sites, periphyton biomass is
inversely related to wave exposure (Weisner et al., 1997). Also, water move-
ment caused by waves may enhance photosynthesis through an improved
carbon exchange between water and plant. An even more indirect effect of
I
Fig. 5.32 Abundance of submerged vegetation in relation to water depth and wind
exposure (fetch relative to the prevailing direction of the wind) in Lake Veluwe for
a good year (1969) and a bad year (1975). The surfaces are obtained by interpolation
through presence-absence data from 0.25ha plots. From Scheffer eta/. (1994a).
256 Vegetation
waves on vegetation is that birds may prefer sheltered sites where they
vandalize the vegetation (Lauridsen et al., 1994; Weisner et al., 1997). Obvi-
ously, waves are important for vegetation in shallow lakes, but several
different mechanisms are involved and the net effect can not be predicted a
priori.
Birds
It has long been assumed that herbivory on submerged plants is negligible,
but this is certainly not always true. Macrophytes probably contribute little
to the total food web even in vegetation dominated lakes (Hoyer et al.,
1997), but various birds, fish and invertebrates include aquatic macrophytes
in their diet and several studies report reductions in vegetation of over 50%
due to herbivory (Lodge, 1991). Few invertebrates consume significant
amounts of intact macrophytes (Kornij6w et al., 1995), and although there
are indications that herbivory in aquatic systems is not systematically lower
than in terrestrial plants (Jacobsen and Sand-Jensen, 1992), the potential
impact of vertebrates on submerged plants is probably much larger.
Bird grazing on aquatic vegetation has received particular attention.
Coots and swans are the most notorious plant eaters in lakes during the
summer season. Exclosure experiments in Scottish (Jupp and Spence, 1977),
Danish (Lauridsen et al., 1993) and Dutch (Van Wijk, 1988) lakes where
birds were observed to forage on plants give remarkably similar estimates of
maximum grazing impact. Aboveground plant biomass in gauze protected
plant stands was found to be S-7 times larger than on unprotected sites in all
cases. A possible explanation for this convergence is that birds tend to clip
plants, keeping shoots short throughout the summer (Jupp and Spence,
1977; Sfl)ndergaard et al., 1997). Indeed, vegetation dominated by species
that usually grow tall can appear like neatly cut meadows when grazing
coots are abundant. In less firm soils individual plants may be uprooted
entirely by birds (Lauridsen et al., 1993; Van Dank et al., 1994b). This
'vandalism' can cause the damage to exceed the actual consumption. How-
ever, a large number of studies indicates that bird effects are usually smaller
than the reductions of about 80% found in the quoted enclosure experi-
ments (Ki!llrboe, 1980; Mitchell, 1989; Van Dank et al., 1994b; Woolhead,
1994; Perrow et al., 1996; Sfl)ndergaard et al., 1997).
There is a strong seasonal variation in bird grazing. In the spring
macrophyte consumption is low (S!Ilndergaard et al., 1997), due to the fact
that territorial behaviour keeps bird density low and filamentous macro
algae and invertebrates rather than plants are the preferred food at that
time of the year (Perrow et al., 1996). In the summer a diet shift to
macrophytes and an increase in bird density leads to an increase in grazing
pressure on aquatic vegetation (Perrow et al., 1996) and effects on plants
become apparent (S!Ilndergaard et al., 1997). In the autumn bird numbers
rise spectacularly in many lakes due to the arrival of migratory animals.
The regulation of vegetation abundance 257
Densities of over 100 individuals ha- 1 are not unusual (Hanson and Butler,
1994b; Van Donk et al., 1994b) as opposed to typically less than lOha-1 in
summer (Van Donk et al., 1994b; Perrow eta/., 1996; Si1Sndergaard eta/.,
1997). Consumption by such high autumn and winter populations of the
remaining shoots and dormant structures may conceivably have an impact
on next year's vegetation development. Exclosure experiments show that in
the Randmeren overwintering Bewick's swans (Cygnus columbianus) can
reduce the tuber bank of sago pondweed vegetation to less than 5% of its
initial biomass in a few weeks (M. van Eerden and M. Scheffer, unpublished
results). Since other experiments have demonstrated that summer biomass
in these vegetation stands is closely related to initial tuber density in the
spring (Van Dijk et al., 1992), winter consumption of tubers may well
contribute to the sparsity of the pondweed stands (about 50g m-2) in the
Randmeren.
Fish
Some fish can consume significant amounts of vegetation. Rudd (Scardinus
erythrophtalmus) and roach (Rutilus rutilus) are omnivores that may include
macrophytes in their diet. These animals have not been reported to reduce
total plant biomass significantly. However, they may affect vegetation com-
position by selectively grazing on some macrophyte species. In Lake
Zwemlust, for instance, these fish are thought to have caused a shift from an
Elodea dominated vegetation to dominance by the more calcareous
Ceratophyllum (Van Donk eta/., 1994b). A much more vigorous grazer is
the grass carp (Ctenopharyngodon idella). This exotic species is often intro-
duced in European and American lakes to reduce vegetation abundance in
cases where submerged plants are experienced as a nuisance (Shireman et
al., 1985; Santha et al., 1991). These animals can wipe out submersed plants
entirely from lakes where they are introduced, leading to increased turbidity
due to algal blooms and increased sediment resuspension (Small, Jr. eta/.,
1985). Only if the stock is managed very carefully by sequential harvesting
and restocking grass carp may balance the amount of vegetation to a desired
level (Shireman eta/., 1985).
Even when plants are not consumed, disturbance of the sediment by
benthivorous foraging fish such as bream, roach and carp may well be an
important factor inhibiting the colonization by plants. As mentioned earlier,
the stock of these animals in shallow lakes can be extremely high and the
sediment surface is typically covered by little craters left by their foraging
activities. Several observations are suggestive of an effect of bream on
vegetation development. For instance, plants soon developed in wire cages
placed on barren sandy sediment in Lake Maarseveen to study the impact of
bream on the benthic fauna (Ten Winkel and Meulemans, 1984). Unfortu-
nately enclosures usually exclude birds and fish. Therefore it is hard to say
which of the two groups is more important from such studies.
258 Vegetation
In the case of carp, the detrimental effect has been unequivocally demon-
strated in many experimental studies (Threinen and Helm, 1954; Tryon,
1954; King and Hunt, 1967; Crivelli, 1983). Indeed the introduction of carp
for sport fishery purposes is likely to be responsible for the disappearance of
vegetation from many lakes in western Europe. In the United States, where
carp was introduced more than a century ago, the devastating effect of this
species on vegetation and its associated waterfowl and game-fish has been
noted long since (Cahn, 1929). Several biomanipulation experiments 'avant
Ia lettre' removing carp and other 'coarse fish' were conducted, resulting in
vegetation recovery and enhanced transparency of the water (Rose and
Moen, 1952; Cahoon, 1953; Threinen and Helm, 1954). In view of their
destructive effects carp are even referred to as the swine of lakes (Threinen
and Helm, 1954). King and Hunt (1967) found that while Chara plants were
consumed by carp, pondweeds were damaged more by uprooting. Sediment
resuspension may be another important way in which carp and other
benthivorous fish affect vegetation performance. Not only does this cause a
high turbidity, settling of the suspended sediment may also cover plants. The
magnitude of this effect is illustrated by an observation of Threinen and
Helm (1954). They found that wire exclosures helped little in stimulating
vegetation growth, probably due to a deposition in the exclosures of as much
as 20cm of sediment within two months. As wave action was minor in the
sheltered area where the experiments were conducted, carp was held re-
sponsible for these sediment dynamics.
Stabilizing mechanisms
In the previous sections it has been shown that vegetation tends to enhance
water clarity, but also that light limitation is one of the main problems for
submerged plants in eutrophic lakes. This implies a positive feedback in the
development of submerged vegetation: once they grow, the water clears up
and they grow even better. Figure 5.33 summarizes the main mechanisms
involved. A simple way of evaluating the overall effect of the depicted
Vegetation and phytoplankton dominance 259
Resuspended
~~~~
Sed. ~
0 <B_/'f'
<B ~~p
Vegetation
0
0
~
~
~
1 0
~
:::=----....AIIelopathic
Subs.
~ 0 Zooplankton
Fig. 5.33 Feedbacks that may cause a vegetation dominated state and a turbid state
to be alternative equilibria. The qualitative effect of each route in the diagram can be
computed by multiplying the signs along the way. This shows that both the vegetated
and the turbid state are self-reinforcing. From Scheffer et al. (1993b).
interactions, is to multiply the signs along the way of a path through the
scheme. This exercise shows that through all depicted routes turbidity en-
hances turbidity, and vegetation enhances vegetation.
Although the vegetation-turbidity feedback is probably a major mecha-
nism causing hysteresis in shallow lakes, other factors may well be important
too. As explained later, vegetationless lakes tend to stay vegetationless not
only because they are turbid, but also because sediment disturbance by
waves and benthivorous fish prevents plant settlement, and herbivory may
help to prevent vegetation recovery. Vegetated systems, on the other hand,
tend to stay vegetated because the resulting water clarity promotes plant
growth, but also because the sediment is stable, the fish community is more
shifted towards piscivores, and the overall vegetation productivity is high
enough to sustain a large population of herbivores without collapsing.
The existence of stabilizing mechanisms that tend to keep the system in
either a vegetation dominated or a phytoplankton dominated state suggests
the potential for alternative stable states. However, in mathematical models
alternative equilibria usually occur for limited ranges of parameter settings
only, and, likewise, real systems will normally have these properties only for
a limited set of conditions. Indeed, the hypothesized stabilization of the
vegetated state seems unlikely in deep lakes where the narrow littoral zone
260 Vegetation
that can be vegetated has a less dramatic impact on turbidity than in shallow
lakes that can be entirely vegetated. Also, in shallow lakes the existence of
alternative stable states will be limited to an intermediate range of nutrient
levels, as oligotrophic lakes are rarely turbid and very high nutrient loading
usually excludes vegetation dominance. Therefore, the demonstration of
stabilizing mechanisms per se is not sufficient to conclude that a lake has
alternative stable states. The following sections discuss the effect of factors
such as nutrient loading and lake depth on the potential for alternative
equilibria, and show how the presence of hysteresis may be inferred from
the dynamics of Jakes in practice.
i
•• • • • • ·• critical
·-·-=---- ----------
turbidity
In some cases this may be very difficult; for instance when wave
resuspension causes a high background turbidity that is ·not due to
phytoplankton, or when sediments contain a large amount of buffered phos-
phorus. In such situations a disturbance, such as biomanipulation, that tem-
porarily reduces the turbidity to a value below the critical level needed for
macrophyte colonization may cause a permanent shift to the alternative
stable state of clear water and vegetation dominance. This is discussed
further in section 6.1.
Note that, at the extremes of the range of nutrient levels over which
alternative stable states exist, either of the equilibrium lines approaches the
critical turbidity that represents the breakpoint of the system. This corre-
sponds to a decrease of stability. Near the edges, a small perturbation is
enough to bring the system over the critical line and to cause a switch to the
other equilibrium.
Water level in the lake is another important control variable with respect
to aquatic macrophyte dominance. Since vegetation can resist a higher
turbidity if the lake is shallower, the horizontal breakpoint line in the
diagram will be at a higher critical turbidity in shallower lakes. It can be seen
from the graphical model that a small shift in critical turbidity resulting from
a change in water level can bring about a switch from one state to the other
in lakes that are close to the breakpoint already. This is in line with obser-
vations in several lakes (Wallsten and Forsgren, 1989; Blindow et al., 1993;
Sanger, 1994).
262 Vegetation
Nutrients
Fig. 5.35 Combining the results from the cyanobacterial competition analysis (Fig.
3.19) and the graphical vegetation-turbidity model (Fig. 5.34), it follows that under
certain conditions shallow lakes may posses three alternative equilibria: one domi-
nated by aquatic macrophytes, one dominated by a diverse phytoplankton assem-
blage and one dominated by cyanobacteria (see text).
maximum vegetated
:w :rn
Vegetation cover Vegetation cover
I2J :CSJ
Vegetation cover Vegetation cover
Fig. 5.36 Schematic representation of the response of the percentage of the lake
area covered by aquatic vegetation to changes in Secchi disk transparency or turbid-
ity (vertical light attenuation) in hypothetical flat bottomed lakes versus lakes with
a v-shaped depth profile (the two left-hand drawings) as predicted from the empiri-
cally derived effects of transparency and turbidity on the maximum depth inhabited
by plants (zm~) represented in the top two diagrams (see text).
50%
Turbidity (m- 1)
accounting for more realistic lake shapes, the overall response of vegetation
abundance to turbidity should typically look sigmoidal for a shallow lake
(Fig. 5.37).
Over a range of very high turbidities submerged plants will be virtually
absent. Over a range of very low turbidities, the whole lake will be occupied
by plants. Between these extremes there is a range of turbidities over which
vegetation responds relatively strongly to turbidity. The response over this
range is steeper if the bottom of the lake is flatter. Obviously, the critical
turbidity where vegetation reacts relatively steeply to changes is higher if
the lake is shallower.
0 100%
Vegetation coverage
L
Turbidity
b Turbidity
Fig. 5.40 Possible effects of nutrient loading on the equilibrium vegetation abun-
dance and turbidity as derived from the graphical isocline model (Fig. 5.39). Alter-
native stable states can (a) but need not (b) arise.
(Fig. 5.39) can intersect in more than one point, and this can only happen if
the middle part of the vegetation isocline is steep enough relative to the
slope of the turbidity isocline. As argued, the slope of the vegetation re-
sponse to turbidity depends on how flat the bottom of the lake is, whereas
the slope of the turbidity isocline depends on the impact of vegetation on
turbidity. In lakes where vegetation declines more gradually with increasing
turbidity and where the impact of vegetation on turbidity is smaller, multi-
ple intersections are less likely to occur, and hence hysteresis is unlikely.
Note that even when there is no hysteresis, however, the response of the
system will still tend to be rather discontinuous (Fig. 5.40b ). Minor changes
in control variables such as nutrient loading or water depth may therefore
have large effects when the lake is close to the critical range of conditions.
(8)
where the power (p) determines the slope of the response of vegeta-
tion abundance to turbidity, and the half-saturation constant (hE) represents
the turbidity allowing 50% of the lake to be covered by vegetation. Obvi-
ously, these formulae are a rather sterile representation of real relation-
ships, and by freezing things into this simple form we sacrifice detail.
Nonetheless, the minimal model is a useful tool to highlight some aspects of
the hysteresis.
The model has only four parameters. Since it describes the system on a
high integration level, the parameter values can not be derived from the
physiology of the organisms as in some other models. Nonetheless, each of
the parameters has a clear qualitative interpretation. In summary:
The first two parameters affect the turbidity isocline, while the last two
determine the shape of the vegetation isocline (Fig. 5.41).
Although the effect of the parameters on the equilibria can be seen by
making overlays of these graphs and tracking the intersections, an easier
way to explore the effect of parameters on the systems equilibria is to look
directly at the equilibrium turbidity or vegetation abundance as a function
of one or more parameters. To obtain these functions we substitute V in the
first equation or, alternatively E in the second one:
Vegetation and phytoplankton dominance 271
Vegetation Vegetation
1
Turbidity isoclines
Turbidity Turbidity
Vegetatlon
Vegetation isoclines
10
Turbidity
Fig. 5.41 Different shapes of the isoclines of turbidity (dE/dt = 0 upper panels) and
vegetation (dV/dt = 0 lower panels) resulting from changes in the parameters £ 0 , hv,
hE and p in the minimal mathematical model of the vegetation-turbidity interaction
(Eqs. 69 and 70) (see text for interpretation of the parameters).
V*= h~ (10)
( ~)"
h.+V*
+h"
E
The formulae look awkward, but we can now plot them with any appro-
priate software to see how the equilibria depend on the parameters.
272 Vegetation
Vegetation (V) Turbidity (E)
1 ~~~--------~ 10,-------------------,
f2
0.5
11
11 f2
0 o+-==~~~------~
0
If we fix all parameters to their default values and plot turbidity and
vegetation coverage as a function of E 0 to explore the response of the system
to the nutrient level a folded line arises, indicating the hysteresis (Fig. 5.42).
The interpretation is similar to that of the hysteresis derived earlier. The
upper and lower branch represent stable equilibria. The middle part is the
unstable saddle equilibrium that marks the border of the basins of attraction
of the stable states. The two inflection points (f1 and / 2) are fold bifurcations
where the saddle collides with either of the stable equilibria. For intermedi-
ate nutrient values the system has two alternative equilibria, a clear one with
abundant vegetation and a turbid one with little vegetation.
As explained already for the simpler graphical rriodel (Fig. 5.34) the
resulting hysteresis with respect to changes in the nutrient level can be seen
by slowly moving from the oligotrophic left end to the hypertrophic right
end of the diagram. On the left where the basic turbidity (E0) is low there is
only one equilibrium, the clear vegetation dominated state. With increasing
nutrient loading the lake tends to stay clear until / 2 is reached. Further
eutrophication will cause a catastrophic transition to the turbid state. Subse-
quent reduction of turbidity does not have much effect, as the system stays
in the turbid state. Only when nutrients have been reduced enough to reach
the left-hand fold bifurcation (M, another catastrophic transition will return
the lake to the clear water state.
A simple way to explore how this hysteresis depends on the other param-
eters of the model is to add these parameters as an extra dimension to the
hysteresis plots (Fig. 5.43). The resulting three-dimensional plots show that
the range over which the turbid and the clear state exist as alternative
equilibria increases with hE and p. Recalling the interpretation of these
Vegetation and phytoplankton dominance 273
parameters, the model thus confirms the earlier conclusion that hysteresis
should be most pronounced in shallow lakes with a flat depth profile. The
hysteresis is also enhanced by a high vegetation impact (low h,). This rela-
tive impact of macrophytes is likely to be higher in shallow water where
wind resuspension can cause strongly increase turbidity in the absence of
vegetation. All of these theoretical results lead to the same general conclu-
sion: hysteresis due to the vegetation-turbidity feedback is more likely to
occur in shallow water.
In fact, deep lakes are not even considered in the model. The maximum
vegetation coverage in the model is 100%, occurring at zero turbidity. Even
in very clear water, however, aquatic macrophytes do not completely colo-
nize deep lakes. Obviously, the vegetation-turbidity feedback is unlikely to
cause hysteresis in deep lakes where only a minor part of the total surface
and an even smaller part of the overall volume can be colonized by
macrophytes.
Vegetation
Vegetation
Fig. 5.44 Seasonal cycle represented in the simulation model MEGAPLANT with
the main environmental factors (outside circle) and plant characteristics (inside
circle) that are taken into account. After Scheffer eta/. (1993a).
model at a fixed day in the spring. From that moment onwards, each
overwintering structure starts transforming a daily percentage of its remain-
ing biomass into sprout growth. At the end of the growing season, a propor-
tion of the vegetation biomass is transformed into overwintering structures,
and the remaining shoots disappear, although optionally part of the vege-
tation can be allowed to remain wintergreen. Growth depends on photo-
synthesis and respiration, both of which are temperature dependent.
Photosynthesis on a given site on the plant also depends on in situ light and
the distance from the tissue to the top of the plant. The latter is due to the
decrease in activity with tissue ageing. Daily irradiance and temperature in
the model follow a sine wave over the year. Light also follows a daily cycle
and is attenuated in the water column according to the Lambert-Beer law
(Section 2.1 ). In addition, in situ light is affected by self-shading. Plants grow
until they reach a maximmn length. After this a proportional increase in
vegetation biomass over the whole length axis occurs. When plants hit the
water surface, shoots spread just under the water surface, forming a canopy.
Throughout the growing year mortality of plants occurs due to factors such
as wave action or crowding effects.
As a first step in the analysis of the vegetation-turbidity feedback the
impact of vegetation on turbidity is left out, and the effect of turbidity on the
276 Vegetation
summer standing crop at a water depth of 1m is computed (Fig. 5.45 upper
panel).
The numerical procedure to produce this picture is to increase the turbid-
ity (E) of the water in small steps from 0 tillS. Vegetation development is
simulated at each of these intervals for several years until the summer
biomass has stabilized. Then the equilibrium biomass is plotted and used as
a starting point for simulations at the next turbidity step. This analysis gives
the same result when it is run the other way around, decreasing turbidity
stepwise from 8 till 0. The result is in line with the idea that vegetation
declines relatively sharply at a critical turbidity. Note that these simulations
give biomass at a lake depth of 1m. As argued, the depth profile of a lake
will affect the response of the total vegetated area to turbidity (used in the
minimal models). The latter may be smoother than a local biomass response
depicted here.
To check if the vegetation turbidity feedback may cause alternative
equilibria, the model is expanded to include the effect of vegetation on
turbidity. Turbidity in the vegetation (E) is assumed to be proportional to
the turbidity of unvegetated water (E0) and decline with vegetation biomass
(Bg AFDW m-2 ) in a way that is analogous to the effect of vegetation
coverage assumed in the previous model:
(11)
The half saturation constanth 8 is set to 150gDWm-2 which fits well to the
data from Charophyte vegetation in Veluwemeer (Van den Berg et al.,
1997). It implies a turbidity reduction of 77% in a dense vegetation of
500gm-2• As explained earlier, as much as 90% reduction of turbidity is
measured in various dense vegetations, and hence the used setting may be
considered a conservative estimate of the vegetation effect.
When the equilibrium analysis is repeated for this version of the model,
it appears that the vegetation collapse with increasing turbidity occurs at a
much higher threshold (I in Fig. 5.45 lower panel). Obviously, this is the
result of the implemented effect of vegetation on turbidity. By keeping
the water within the vegetation clear it can survive even when turbidity in
the absence of vegetation (E0 displayed on the horizontal axis) would be
high. Reversing the scenario, however, the vegetation biomass follows a
different path (II). With decreasing turbidity vegetation does not recover
until the critical turbidity where it also recovered in the simulation in which
plants were not allowed to influence turbidity (Fig. 5.45 upper panel). The
explanation is simple: in the colonizing phase plant biomass is insufficient to
cause a significant reduction in turbidity. Therefore, the ability of
macrophytes to keep the water clear helps vegetation to persist, but not to
colonize a lake. As a result there is a range of conditions over which the
vegetated and the unvegetated state are both stable. As in the previous
hysteresis models, the dashed line (III) represents the threshold biomass for
Vegetation and phytoplankton dominance 277
450,--------------------------------------------,
no vegetation effect on turbidity
~
E!
J
0
"
c
0
150
0
Turbidity, Eo (m-1)
450,--------------------------------------------,
1: 300
~c
t0
150
:g
0 5
Turbidity of unvegetated water, E0 (m-1)
,.
. :
.,................ , ...
240
. I
'
':
''
'
I
''
+
'' '' ''
'1'
E ! A
~
~
E
160
~' t'
~
'' '''
6 ''' ''
i
~
100
''''
''
'''
'''
'' '''
0
: ~
Fig. 5.46 Response of charophyte biomass to water level computed from the model
MEGAPLANT as in the lower panel of Fig. 4.45 showing a hysteresis similar to the
one with respect to turbidity.
Vegetation and phytoplankton dominance 279
simplifications in the minimal model, or of particular details of the complex
model. To use the phrase from Levins (1966) once more, we are more likely
to accept something as the truth when it comes out 'as the intersection of
independent lies'.
Implications of seasonality
It is surprising at first sight that temperate aquatic vegetations are able to
sustain an alternative clear state despite the fact that in large shallow lakes
the aboveground biomass often disappears entirely in winter. Indeed, tur-
bidity often recovers to high values during the absence of plants (e.g. Fig.
5.18). Since over the range where alternative equilibria exist, the vegetation
dominated state can only be reached if the initial biomass is high enough, it
can be inferred that the amount of biomass invested in overwintering struc-
tures such as seeds, spores, rhizomes or tubers must be essential to allow a
successful return of the vegetations that are not wintergreen. When the
biomass of overwintering structures falls below a certain threshold value,
the spring vegetation will be too sparse to clear up the water sufficiently to
prevent shading from driving the vegetation extinct.
The model CHARISMA includes seasonality and this offers the possibil-
ity to explore the theoretical implications of seasons for the stability of the
vegetation dominated state somewhat further. Chara aspera, the species
mimicked by the model, invests as much as 25% of its ashfree dry weight in
spores and tuber-like structures called 'bulbils'. The importance of this
investment for vegetation survival can be checked in the model by simply
reducing it. Indeed, a fourfold reduction of the percentage invested in
overwintering structures already diminishes the potential for the vegetation
to maintain dominance markedly (Fig. 5.47).
Obviously, the importance of investing in underground overwintering
structures depends upon the winter survival chances of aboveground
biomass. Some species are more often wintergreen than others, but in
practice overwintering of shoots also varies strongly with shelter and cli-
matic conditions. In (sub)tropical areas such as Florida where winters are
very mild, plants usually remain present throughout the year as they can
simply keep growing. In temperate regions wintergreen vegetation occurs
especially on sheltered sites or under ice cover where wave action is minor.
One would expect plants to invest less in overwintering structures in situa-
tions where wintergreen survival is more likely. This is confirmed nicely by
a study of the variation in life cycles of Potamogeton pectinatus in the Baltic
(Kautsky, 1987). Populations on exposed sites invest about four times as
much of their biomass in tubers as populations on sheltered sites where
more shoots survived the winter.
Some species, such as Elodea canadensis can not produce special
overwintering organs and depend entirely on the survival of shoots or
fractions of shoots during the winter. This may represent an Achilles heel
280 Vegetation
450,-----------------------~~~~~~-----.
less investment in tubers+ spores
.... +e:.................. ~ ... ....
~
~ '?'
t!. 'it
0 4
Turbidity of unvegetated water, Eo (m- 1)
Fig. 5.47 Response of charophyte biomass to turbidity computed from the model
MEGAPLANT as in the lower panel of Fig. 5.45 but reducing investment in
overwintering structures such as tubers and spores to only 25% of the normal value.
This reduces the extent of the stable upper branch of the hysteresis curve suggesting
that large investments in survival structures are crucial for persistence of the veg-
etated state in lakes where aboveground biomass disappears seasonally.
©~:
® erosion mortality
I~ : ~ ii
0.._[_~9-
'
''
:-
''
'
_vegetation --+ Vegetation ----.Vegetation
biomass (v) biomass(v) biomass(v)
Fig. 5.48 (a) The relative growth rate (dV!dVV) declines linearly with vegetation
biomass (V) due to competition in a logistically growing vegetation. Biomass stabi-
lizes at carrying capacity (K) where the net growth is zero. (b) Relative erosion
mortality (-dV/dt!V) of plants caused by uprooting by waves or animals from unsta-
ble sediment approaches zero with increasing vegetation biomass due to consolida-
tion of the sediment, dampening of wave action and exclusion of benthivorous fish.
(c) Growth and mortality balance each other in two intersection points. The left
hand point (open symbol) represents an unstable breakpoint: below this threshold
biomass, vegetation decreases to zero because mortality exceeds growth, whereas
above this threshold vegetation biomass will increase until the stable intersection
point (filled dot) is reached.
this mechanism could lead to alternative stable states even if the vegetation-
turbidity feedback did not exist. A graphical model may help to see this (Fig.
5.48).
We assume the vegetation to grow logistically, implying that the relative
growth rate declines linearly with vegetation density due to increased com-
petition (Fig. 5.48a). Now we want to study mortality due to sediment
instability as an extra regulating mechanism. Such erosion mortality will also
decrease with vegetation density because vegetation stabilizes the sediment.
However, the decline in erosion mortality will be concave, as it can never
become negative even at very high vegetation densities (Fig. 5.48b). If the
growth and mortality are plotted in the same graph, intersections represent
equilibria of the vegetation as both processes balance each other exactly
(Fig. 5.48c). In the depicted case the intersection on the left-hand side
represents an unstable breakpoint. Only when vegetation density is higher
than this threshold can growth exceed erosion mortality. Thus there are two
alternative stable states: one in which vegetation is absent and one in which
vegetation has a biomass close to the carrying capacity (the other intersec-
tion point). This situation with alternative stable states arises only if two
conditions are fulfilled: erosion mortality at very low vegetation densities
must be high enough to exceed the maximum growth rate, and the decrease
of mortality with increasing vegetation density must be steeper than the
decrease in growth rate due to competition.
In practice the effect of vegetation on turbidity will also play a role and
Vegetation and phytoplankton dominance 283
®~dV/dlfvi
dV /dtfv
@~:
g"'wth
i
-l-
:- growttr :
: -j-
0 0 : '
Ki '' ''
' '
-+Vegetation ---+Vegetation ___..,Vegetation
biomass (v) biomass{v) biomass(V)
Fig. 5.49 (a) Relative growth of vegetation (dV!dVV) as a function of biomass (V) in
a situation with alternative stable states due to the fact that vegetation reduces
turbidity. Below a critical threshold (open symbol) vegetation biomass is too low to
make the water clear enough for growth and the vegetation become extinct. Above
this threshold vegetation will stabilize at carrying capacity (K). (b) If turbidity in the
unvegetated state is low enough to allow growth, the clearing effect will not lead to
alternative stable states. However, unlike in the case of simple logistic growth (Fig.
5.48) relative growth will still tend to be a humped function of biomass. (c) In this
situation (b) even a relatively low erosion mortality (cf. Fig. 5.48) may lead to
alternative stable states.
the combination of these mechanisms will increase the chance that alterna-
tive stable states occur. To see this consider the way in which inclusion of the
vegetation-turbidity feedback should be expected to modify the logistic
growth. Instead of a linear decrease of growth rate with vegetation biomass,
the relationship of growth to biomass will become humped (Fig. 5.49a).
At very low vegetation densities turbidity can be so high that growth is
negative. Only if the biomass exceeds a critical threshold value can turbidity
then be reduced sufficiently to allow a positive growth bringing the vegeta-
tion to carrying capacity. This represents a situation with alternative stable
states. Even if turbidity in the absence of plants does not prevent growth and
cause an alternative unvegetated equilibrium, however, the relative growth
curve will remain convex as long as vegetation promotes water clarity (Fig.
5.49b). The combination of erosion mortality with the vegetation-turbidity
feedback may now still cause alternative equilibria (Fig. 5.49c) even if each
of these mechanisms separately would not be sufficient to have that effect:
a moderate erosion mortality may be enough to prevent colonization if
seedlings grow slowly due to high turbidity.
In conclusion, the vegetation-turbidity feedback tends to cause alterna-
tive stable states in shallow lakes but other mechanisms may well contribute
to the hysteresis. Vegetationless lakes tend to stay unvegetated not only
because they are turbid, but also because sediment disturbance by waves
and benthivorous fish prevents plant settlement, and herbivory may help to
prevent vegetation recovery, while vegetated systems, on the other hand,
tend to stay vegetated because they are clear but also because the sediment
284 Vegetation
is more stable, the fish community is more shifted towards piscivores, and
the overall vegetation productivity is high enough to sustain a large popula-
tion of herbivores.
Hallmarks of hysteresis
Although there is abundant evidence for the existence of mechanisms that
tend to cause a positive feedback in the development of aquatic vegetation,
our quantitative insight into the functioning of the system is still relatively
poor. Obviously, we are still far from able to produce a mechanistic model
that incorporates the discussed mechanisms in a way that allows us to
predict whether a given lake will possess alternative clear and turbid
equilibria and for which set of conditions.
Therefore, the best way to find out if hysteresis is important in real lakes
is to check if specific predicted patterns can be found in the field. Obvious
hallmarks of a system with alternative equilibria are the hysteresis in its
response to a control variable and the catastrophic transitions at the fold
bifurcations. In theory, experiments in which a control variable is gradually
increased and subsequently decreased are obviously the best way to check
this. Besides playing with the control variable, one may perturb the system
state. Following small perturbations the system will return to its original
state. A sufficiently large perturbation, however, should bring the system
into an alternative stable state if the values of the control variables are in the
range for which multiple equilibria exist.
The possibility for such controlled experiments with real lakes is limited,
but we may also take advantage of natural experiments. The response of a
hysteretic system to a variable environment leads to the expectation that
states sampled over a long time-series should fall in two contrasting clusters
as the contrasting stable states should occur more often than the unstable
transients. Likewise, sets of comparable lakes should have bimodal state
distributions at any given instant of time.
Summarizing there are four types of indicative observations (Fig. 5.50):
a. The response to a slow increase of a control factor (C) such as nutrient
loading or water level should be discontinuous. The system switches to a
contrasting state when a critical value of C is exceeded.
b. A subsequent decrease of the control variable should lead to a switch
back, occurring at a lower threshold value of C than the forward switch.
c. It should be possible to bring the system from one stable state to another
one by means of a perturbation, if the control variables are in the range
that allows alternative equilibria.
d. Distributions of system states should be bimodal.
It is important to realize that most of these observations are not sufficient
for a positive diagnosis of true hysteresis. Therefore we scrutinize them
somewhat further:
Vegetation and phytoplankton dominance 285
state state
A B
state state
c ... ------- 0 ·a\.r~·~-· ....-------
·-e·.. .
<-F..... -· . •t!
.. ........ !-"e!.e,.-:• ••
!1.~
Fig. 5.50 Four types of observations indicating that a system may have alternative
stable states: (a) The response to a slow increase of a control factor (C) such as
nutrient loading or water level is discontinuous. The system switches to a contrasting
state when a critical value of Cis exceeded. (b) A subsequent decrease of the control
variable leads to a switch back, occurring at a lower threshold value of C than the
forward switch. (c) It is possible to bring the system from one stable state to another
one by means of a perturbation, provided that the control variables are in the range
that allows alternative equilibria. (d) Distributions of system states are bimodal.
Much of the work presented in this book has been invoked by the need to
find ways of restoring shallow lakes, but the emphasis in the text has been
on unravelling the mechanisms rather than on the direct application to
lake management. This chapter reviews the main points from an applied
point of view. It is not meant as a practical guide for lake restoration.
Several such guides including information on material, costs and legisla-
tion are available now. A good general treatise on reservoir and lake
management is the work by Cooke and co-authors (Cooke eta/., 1993).
Restoration guides that are more directed towards shallow lake problems
are published in The Netherlands (Hosper et al., 1992) and England (Moss
eta/., 1996).
The following sections briefly summarize the implications of alternative
stable states for management and the practical measures that can be taken
to change the state of a lake.
Stability properties
Although the mechanisms involved in causing alternative equilibria and the
details of the resulting patterns can be quite intricate the overall stability
properties can be summarized in a simple and intuitively straightforward
way by means of a 'stability landscape' or 'marble-in-a-cup diagram' (Fig.
6.1).
290 Managing the ecosystem
turbidity
The system, like a ball, tends to move downhill and settle in the deepest
point which is an equilibrium. The slope of the surface determines the
direction and speed of movement. Such a stability landscape can be com-
puted from a mathematical model of the system, for instance, by using the
derivative of a state variable (such as turbidity) as the slope of the hills in the
stability landscapes. As indicated in the figure they correspond to transverse
sections through the familiar sigmoidal hysteresis curves presented in
Chapter 5. On the hill-tops and in the deepest point of the valleys the
slope is zero, corresponding to a derivative of zero and thus to an equilib-
rium. However, only the valleys of the stability landscapes represent
stable equilibria. The hill-tops are unstable equilibria and represent the
breakpoints that mark the limits of the basins of attraction of the stable
equilibria.
With respect to the response of the lake to management it is important to
distinguish between disturbances and measures that affect the stability
properties. In terms of stability landscapes (Fig. 6.1 ), disturbances are
displacements of the ball but do not alter the pattern of hills and valleys.
Fish-kills, herbicide treatments and heavy storms are examples. If there is
Implications of alternative stable states 291
only one stable state (valley), the effect of a disturbance will be temporal as
the system will always settle to this same state again. However, if two
alternative stable states (valleys) are present, the system may settle to the
alternative stable state if the disturbance has been strong enough to move it
past the breakpoint (hill-top). The permanent loss of vegetation caused by
a single heavy storm event as observed in Lakes Apopka and Ellesmere
(Section 1.2), and the reverse switch to a vegetation dominated state in
response to a single drastic reduction in fish stock as observed in Lakes
Zwemlust and Linford (Section 1.4,5) illustrate this possibility. On the con-
trary, changes in external conditioning factors such as the nutrient loading
(e.g. Veluwemeer, Chap. 1.1) or the average water level of the lake (e.g.
Lake Tiimnaren, Section 1.3) will change the stability properties (the land-
scape, 6.1) which may also cause a shift but has distinctly different implica-
tions with respect to management than disturbances.
Fig. 6.2 Deep lakes are less likely to show hysteresis than shallow ones, but there is
a continuum of possibilities between the extremes of a full-blown hysteresis and a
smoothly responding deep lake.
Implications of alternative stable states 293
Thus, even in lakes that are, for instance, too deep to have alternative
stable states, reduction of the nutrient loading may have little effect until a
critical concentration range is reached where the response becomes much
stronger.
Note that in the discussion of hysteresis with respect to nutrients it is
important to distinguish between nutrient loading and nutrient concentra-
tions in the Jake water. The change of the stability properties is a function of
external nutrient loading which is a conditioning factor, whereas inlake
nutrient concentration is a systems property that is strongly affected by the
biological structure of the system. Plant dominance, for instance, often
causes nitrogen concentrations to drop strongly even when external nitro-
gen loading remains the same.
6.3 BIOMANIPULATION
Biomanipulation has proved to be one of the most successful measures to
let a turbid shallow lake switch to an alternative clear state. Carp removal
was used as a way to restore lake communities in the early 1950s (Rose
and Moen, 1952; Cahoon, 1953; Threinen and Helm, 1954) but the broad
practice of biomanipulation is of the last decade only. Fish manipulation is
likely to affect plankton abundance in most lakes, but the approach is
especially powerful in shallow lakes because, there, a single fish reduction
can produce long lasting results if an alternative stable state is reached.
Although there are many striking examples of successful biomanipulation,
there are also cases in which little effect was observed or the effects were
Biomanipulation 297
not long lasting. In the next sections I briefly review the way in which
successful biomanipulation works, and the main reasons why it may fail.
The literature on this topic grows exponentially, but as a starting point
readers interested in shallow lake work may refer to the proceedings
of a 1989 conference on the topic (Gulati et at., 1990) for early case studies,
and to a study of long-term responses of some lakes (Meijer et at., 1994a)
and a preliminary 'test' to assess the chances of success (Hosper and Meijer,
1993).
Draw-down
Complete draw-down is an extreme form of water level management.
It is applied frequently to fish-ponds and reservoirs but there is less ex-
perience with applying this approach to natural lakes. In vegetated lakes,
prolonged draw-down is used as a way to control aquatic plants as it
usually results in the loss of most submerged species (Cooke et al., 1993).
302 Managing the ecosystem
On the other hand, in unvegetated turbid lakes where sediment
resuspension is a major problem, it seems reasonable to expect that draw-
down could promote a shift to a clear vegetation dominated state. No
case studies are available so far, but the potential scenario seems
straightforward. When the sediment is left to dry out, consolidate and be
colonized by terrestrial vegetation, resuspension is unlikely when the lake is
allowed to fill up with water again. Helophytes that have developed will
survive in the shallowest parts and subsequent colonization of the rest of the
lake bottom by submerged vegetation will help stabilize the clear state in the
lake.
Draw-down also facilitates other modifications that may help improving
lake conditions. Sediment, for instance, can often be removed relatively
easily with bulldozers and scrapers from a dry lake. Importantly, partial
draw-down makes it much easier to remove the fish. As explained earlier,
the chances of shifting a lake from a turbid to a stable clear state by means
of biomanipulation increase strongly with the percentage of the fish stock
that is removed. In Lake Zwemlust (Chapter 1) partial draw-down was used
to be able to remove fish more effectively, and the fish reduction that was
realized in this way has resulted in a spectacular shift to a plant-dominated
state.
Ephemeral ponds that periodically experience a natural draw-down
when they dry out are usually fishless. When they fill up they may have a
rapidly growing vegetation of charophytes or other plants that have ways to
survive the dry period. Such systems may have dense populations of
Daphnids that filter the water.
Flushing
As mentioned in the section on nutrient managem.~nt, flushing a lake with
relatively clean water can reduce its nutrient level, but may also help to get
rid of colonial cyanobacteria in a more direct way. This is because the
growth rate of these algae is relatively small, causing the relative effect
of an extra population loss due to wash-out larger. Put simply, an algal
group can be eliminated from a lake if the loss rate due to flushing exceeds
the growth rate (Section 3.1). Extreme flushing rates that replace more than
about one third of the lake volume per day are likely to eliminate all
phytoplankton, but slow growing species may be washed out at much lower
flushing rates. Since growth rates are very low in winter, flushing in this
season may be particularly effective. Indeed, winter flushing has probably
been an important reason for the decrease of cyanobacterial density in
Veluwemeer (Chapter 1). In practice even small increases in hydraulic
flushing rates may lead to disappearance of blue-green algal in situations
where the competitive balance with other algae is already close to shifting,
due, for instance, to a reduction of the nutrient level in the lake (Section
3.2).
Other measures 303
6.5 OTHER MEASURES
Barley-straw
The addition of barley-straw to ponds can lead to a remarkable reduction in
phytoplankton biomass. The phenomenon is well known, and packages of
straw are even sold in gardening shops for this purpose. There has been
relatively little research on this straw effect, but the available work confirms
the strong overall effect and gives some indications as to the mechanisms
(Gibson et al., 1990; Welch et al., 1990; Everall and Lees, 1996). A recent
case-study (Everall and Lees, 1996), for instance, shows that the addition of
barley-straw (50gm-3) to a small English reservoir resulted in a reduction of
summer chlorophyll concentrations from about 100mgl"1 to about 20mgl"1•
The cyanobacterial blooms that occurred in the previous years were absent
in the summer after the straw addition. These changes did not occur in an
adjacent control.
It has been suggested that uptake of nutrients from the water by bacteria
developing on the decomposing straw causes phytoplankton biomass to
decrease as a result of nutrient limitation (Wingfield et al., 1985). Another
explanation is that the algal control results from release of phytotoxic com-
pounds by the decomposing straw (Gibson et al., 1990; Pillinger et al., 1994).
The mentioned reservoir study (Everall and Lees, 1996) shows no significant
drops in available nutrients while showing that the total 'cocktail' of
algicidal, phytotoxic, unidentified or toxicologically unknown organic com-
pounds reaches concentrations of 0.48-4.31mgl"1 near the straw. Although
toxic substances are thus likely to be the dominant cause of algal decline,
rotifer density was also enhanced after straw addition, and grazing by these
animals together with a moderate spring peak of Daphnia may have helped
to reduce algal biomass.
Although the huge amounts of straw that would be needed make this
approach unlikely to be useful for manipulating large lakes, it might help in
clearing up ponds. Straw can also stimulate macro-invertebrate populations
(Everall and Lees, 1996). Since invertebrates are an essential part of the diet
of small ducklings (Section 4.5), straw addition has been promoted as a way
to make unvegetated gravel-pits more suitable for duck reproduction
(Street, 1978).
Dredging
Resuspension of a thick layer of unconsolidated sediment and phosphorus
release from sediment are conspicuous causes of turbidity in many shallow
lakes, and several ways to solve the sediment problem have been proposed.
Removal of the accumulated sediment is the most straightforward ap-
proach, and it has been applied to many lakes over the years (1993). Dredg-
ing is the most common procedure for sediment removal, although
bulldozer excavation can also work after lake draw-down.
304 Managing the ecosystem
Many different dredge types and dredging approaches have been devel-
oped. A problem of some dredging techniques is that they cause a consider-
able resuspension, with the associated problems of turbidity and release of
nutrients and sometimes toxic substances. Another difficulty is the loose
structure of the sediment layer in many shallow lakes. Often the substance
behaves almost as a fluid. As a result local removal results in spreading out
of the remaining material. The intense horizontal redistribution of loose
sediment in shallow lakes can also be used to our advantage. Sediment tends
to accumulate on deeper sites where wave action does not cause
resuspension (Section 2.2). When a deep site is excavated in an otherwise
shallow lake, sediment is trapped there and may be dredged out relatively
efficiently.
The best studied case of sediment removal from a shallow lake is prob-
ably that of the Swedish Lake Trummen (Andersson, 1988). Dredging in-
creased the mean depth of the lake from 1.1 to 1.75 m and resulted in a
strong reduction of sediment phosphorus release into the water column.
Since then numerous sediment removal cases have been published (Cooke
et al., 1993). ·
Although the recent intensive work on shallow lakes has led to an impres-
sive expansion of our knowledge of these systems, the array of poorly
understood problems remains equally impressive. To name just a few: Un-
der which conditions do alternative stable states exist? How can switches
of states be induced? How does vegetation as a refuge affect various
predator-prey interactions? What are the implications of chemical signal-
ling in trophic interactions for community dynamics? Could we use such
signals to manipulate the food web? What is the role of inedibility and
toxicity of cyanobacteria in their success? Why are brackish lakes turbid
even when they have vegetation? How do predator-prey relations evolve
over the seasons?
The best way to resolve these and other questions will differ from case to
case, but as a rule a combination of approaches is often the most powerful
strategy. Controlled experiments, whole lake manipulations, minimal mod-
els and elaborate simulation models all have their specific strong and weak
points. Addressing the same question simultaneously with different ap-
proaches is the best way to identify artifacts of each of them. Clearly, well
documented whole lake experiments are very valuable at this point. Not
only do they reveal what works and what does not in restoring certain lakes,
they can also help to identify the main regulatory mechanisms. The recent
interest in biomanipulation, for instance, has catalysed the insight in the
major forces that govern the dynamics of shallow lakes. On the other hand
the conditions in whole lake experiments are difficult to control and there
are usually no replicates. As a result it is often difficult to understand what
happened for what reason in hindsight. Therefore, small-scale, well control-
led and replicated experiments remain indispensable further to enhance our
understanding, even though such experiments are necessarily conducted
under rather unnatural conditions, and do not reveal how the specific
mechanism addressed interacts with other mechanisms in the field. Elabo-
rate individual based models and other simulation models can help in
putting different processes into perspective but are difficult to study due to
their mere complexity. Very simple 'minimal' models are easier to under-
stand but do not help to reveal the relative importance of the addressed
mechanism in the field.
Although the merits of combining modelling, laboratory experiments
and fieldwork are broadly recognized the approaches remain quite segre-
gated in practice. Whole lake manipulations are often disguised as rather
crude and uninterpretable by experimentalists, and the lack of integration
between modellers and 'real' biologists is notorious. Even within the mod-
elling world there is a distinct separation between theoreticians working
with abstract minimal models and groups working on more applied quanti-
tative simulation models. Combining approaches requires an investment of
time and energy to explain results and assumptions to relative outsiders, and
The necessity of mechanistic insight 313
to try to understand the basic ins and outs of other disciplines. Also, the
integration may be confronting as it inevitably brings shortcomings in the
separate approaches to the surface.
Nonetheless, a combined approach seems the most effective way to ad-
dress the many unsolved questions about the functioning of shallow lakes.
Questions that are worth resolving, not only because they are exciting from
a scientific point of view but also because of the applied urgency. Many
shallow lakes have degraded badly as a result of human activities from an
attractive clear water state with a high diversity to a monotonous murky
pool. Finding out how to restore this damage will help to make this world a
prettier place, for us and many other species.
References
Aalderink, R.H., Lijklema, L., Breukelman, J., Van Raaphorst, W. and Brinkman,
A.G. (1985) Quantification of wind induced resuspension in a shallow lake. Water
Science and Technology, 17, 903-914.
Abrams, P.A. and Roth, J.D. (1994) The effects of enrichment of three-species food
chains with nonlinear functional responses. Ecology, 75, 1118-1130.
Adams, S.M. and DeAngelis, D.L. (1987) Indirect effects of early bass-shad interac-
tions on predator population structure and food web dynamics, in W.C. Kerfoot
and A. Sih (eds.) Predation Direct and Indirect Impacts on Aquatic Communities,
University Press of New England, Hannover, pp. 103--117.
Ahlgren, I., Frisk, T. and Kamp-Nielsen, L. (1988) Empirical and theoretical models
of phosphorus loading retention and concentration vs. lake trophic state.
Hydrobiologia, 170, 285-304.
Andersen, J.M. (1982) Effect of nitrate concentration in lake water on phosphate
release from the sediment. Water Research, 16, 1119-1126.
Anderson, M.R. and Kalff, J. (1987) Regulation of submerged aquatic plant distribu-
tion in a uniform area of a weedbed. Journal of Ecology, 74, 953--962.
Anderson, M.R. and Kalff, J. (1988) Submerged aquatic macrophyte biomass in
relation to sediment characteristics in ten temperate lakes. Freshwater Biology,
19, 115-121.
Andersson, G. (1988) Restoration of Lake Trummen, Sweden: effects of sediment
removal and fish manipulation, in G. Balvay (eds.) Eutrophication and Lake
Restoration. Water Quality and Biological Impacts, Thonon-les-Bains, pp. 205-
214.
Andersson, G., Graneli, W. and Stenson, J. (1988) The influence of animals on
phosphorus cycling in lake ecosystems. Hydrobiologia, 170, 267-284.
Anthoni, U., Christophersen, C., Madsen, J.O., Wium-Andersen, S. and Jacobsen,
N. (1980) Biologically active sulfur compounds from the green alga Chara
globularis. Phytochemistry, 19, 1228-1229.
Arditi, R. and Ginzburg, L.R. (1989) Coupling in predator-prey dynamics: Ratio-
dependence. Journal of Theoretical Biology, 139, 311-326.
Arneodo, A., Coullet, P., Peyraud, J. and Tresser, C. (1982) Strange attractors in
Volterra equations for species in competition. Journal of Mathematical Biology,
14, 153-157.
Arnold, D.E. (1971) Ingestion, assimilation, survival, and reproduction of Daphnia
pulex fed seven species of blue-green algae. Limnology and Oceanography, 16,
906--920.
Arruda, J.A., Marzolf, G.R. and Faulk, R.T. (1983) The role of suspended sediments
in the nutrition of zooplankton in turbid reservoirs. Ecology, 64, 1225-1235.
Bales, M., Moss, B., Phillips, G., Irvine, K. and Stansfield, J. (1993) The changing
ecosystem of a shallow brackish lake Hickling Broad Norfolk UK II. Long-term
trends in water chemistry and ecology and their implications for restoration of the
lake. Freshwater Biology, 29, 141-165.
References 315
Barko, J.W., Hardin, D.G. and Matthews, M.S. (1982) Growth and morphology of
submersed fresh water macrophytes in relation to light and temperature. Cana-
dian Journal of Botany, 60, 877-Sirl.
Barko, J.W. and Smart, R.M. (1980) Mobilization of sediment phosphorus by sub-
mersed freshwater macrophytes. Freshwater Biology, 10, 229-238.
Barko, J.W. and Smart, R.M. (1981) Comparative influences of light and tempera-
ture on the growth and metabolism of selected submersed fresh water
macrophytes. Ecological Monographs, 51, 219-236.
Beddington, J.R. (1975) Mutual interference between parasites or predators and its
effects on searching efficiency. Journal of Animal Ecology, 51, 597-<i24.
Bengtsson, L. and Hellstrom, T. (1992) Wind-induced resuspension in a small shal-
low lake. Hydrobiologia, 241, 163-172.
Benndorf, J., Schultz, H., Benndorf, A., Unger, R., Penz, E., Kneschke, H., Kossatz,
K., Dumke, R., Hornig, U., Kruspe, R. and Reichel, S. (1988) Food web manipu-
lation by enhancement of piscivorous fish stocks long-term effects in the
hypertrophic Bautzen Reservoir East Germany. Limnologica, 19, 97-110.
Berg, S., Jeppesen, E. and Sfl!ndergaard, M. (1997) Pike (Esox lucius L.) stocking as
a biomanipulation tooll. Effects on the fish population in Lake Lyng, Denmark.
Hydrobiologia, 342,311-318.
Berger, C. (1975) Occurrence of Oscillatoriacea agardhii Gom. in some shallow
eutrophic lakes. Verhandlungen /nternationale Vereinigung Theoretisch
Angewandte Limnologie, 19,2689-2697.
Best, E.P.H. (1982) A preliminary model for growth of Ceratophyllum demersum.
Verhandlungen Internationale Vereinigung Theoretisch Angewandte Limnologie,
21, 1484-1491.
Best, E.P.H. and Dassen, J.H.A. (1987) Biomass stand area primary production.
characteristics and oxygen regime of the Ceratophyllum demersum L. population
in Lake Vechten, the Netherlands. Archiv fUr Hydrobiologie, 76, 347-368.
Blindow, I. (1992a) Decline of charophytes during eutrophication: Comparison with
angiosperms. Freshwater Biology, 28, 9-14.
Blindow, I. (1992b) Long- and short-term dynamics of submerged macrophytes in
two shallow eutrophic lakes. Freshwater Biology, 28, 15-27.
Blindow, I., Hargeby, A. and Andersson, G. (1996) Alternative stable states in
shallow lakes - what causes a shift? in prep.
Blindow, I., Andersson, G., Hargeby, A. and Johansson, S. (1993) Long-term pattern
of alternative stable states in two shallow eutrophic lakes. Freshwater Biology, 30,
159-167.
BJorn, G., Van Duin, E.H.S. and Lijklema, L. (1994) Sediment resuspension and
light conditions in some shallow Dutch lakes. Water Science and Technology, 30,
243-252.
Bloomfield, J.A., Park, R.A., Scavia, D. and Zahorcak, C.S. (1974) Aquatic model-
ling in the eastern deciduous forest biome, U.S.-International biological program,
in E. Middlebrooks, D.H. Falkenberg and T.E. Maloney (eds.) Modeling the
Eutrophication Process, Ann Arbor Science, Ann Arbor, Ml, pp.139-158.
Boers, P., Van der Does, J., Quaak, M., Van der Vlucht, J. and Walker, P. (1993)
Fixation of phosphorus in lake sediments using iron(III)chloride: experiences,
expectations. Hydrobiologia, 233, 211-212.
Boersma, M. (1995) Competition in natural populations of Daphnia. Oecologia, 103,
309-318.
316 References
Boersma, M., Van Tongeren, O.F.R. and Mooij, W.M. (1996) Seasonal patterns in
the mortality of Daphnia species in a shallow lake. Canadian Journal of Fisheries
and Aquatic Sciences, 53, 1&-28.
Bob!, E. (1980) Die! pattern of pelagic distribution and feeding in planktivorous fish.
Oecologia, 44, 36&-375.
Bostrl:lm, B., Andersen, J.M., Fleischer, S. and Jansson, M. (1988a) Exchange
of phosphorus across the sediment-water interface. Hydrobiologia, 170, 229-
244.
Bostrl:lm, B., Persson, G. and Broberg, B. (1988b) Bioavailability of different phos-
phorus forms in freshwater systems. Hydrobiologia, 170, 133-156.
Brancelj, A. and Blejec, A. (1994) Diurnal vertical migration of Daphnia hyalina
Leydig, 1860 (Crustacea, Cladocera) in Lake Bled (Slovenia) in relation totem-
perature and predation. Hydrobiologia, 284, 125-136.
Breukelaar, A.W., Lammens, E.H.R.R. and Breteler, J.G.P.K. (1994) Effects of
benthivorous bream (Abramis brama) and carp ( Cyprinus carpio) on sediment
resuspension and concentrations of nutrients and chlorophyll-a. Freshwater Biol-
ogy, 32, 113-121.
Brl:lnmark, C. (1985) Interactions between macrophytes epiphytes and herbivores an
experimental approach. Oikos, 45, 26-30.
Bronmark, C. (1988) Effects of vertebrate predation on freshwater gastropods an
exclosure experiment. Hydrobiologia, 169, 363-370.
Brl:lnmark, C. (1989) Interactions between epiphytes macrophytes and freshwater
snails, a review. Journal of Molluscan Studies, 55, 299-311.
Bronmark, C., Pettersson, L.B. and Nilsson, P.A. (1993) Induced morphological
defense in crucian carp cues costs and benefits. Bulletin of the Ecological Society
of America, 74, 176
Bronmark, C. (1994) Effects of tench and perch on interactions in a freshwater,
benthic food chain. Ecology, 75, 1818--1828.
Bronmark, C., Paszkowski, C.A., Tonn, W.M. and Hargeby, A. (1995) Predation as
a determinant of size structure in populations of crucian carp (Carassius
carassius) and tench (Tinea tinea). Ecology of Freshwater Fish, 4, 85-92.
Bronmark, C. and Weisner, S.E.B. (1992) Indirect effects of fish community struc-
ture on submerged vegetation in shallow eutrophic lakes an alternative mecha-
nism. Hydrobiologia, 243--244,
Brooks, J.L. and Dodson, S.I. (1965) Predation, body size, and composition of
plankton. Science, 150, 2&-35.
Brouwer, G.A. and Tinbergen, L. (1939) De verspreiding der kleine zwanen, Cygnus
bewickii Yarr., in de Zuiderzee, v66r en na de verzoeting. Limosa, 12, 1-18.
Buiteveld, H. (1995) A model for calculation of diffuse light attenuation (PAR) and
Secchi depth. Netherlands Journal of Aquatic Ecology, 29, 55-65.
Cahn, A.R. (1929) The effect of carp on a small lake: the carp as dominant. Ecology,
10, 371-374.
Cahoon, W.G. (1953) Commercial carp removal at Lake Mattamuskeet, North
Carolina. Journal of Wildlife Management, 17,312-316.
Camerano, L. (1880) Dell'equilibrio dei viventi merce la reciproca distruzione,
Accademia delle Scienze di Torino 15: 393-414. (translated in the cited source by
C.M.Jacobi and J.E.Cohen, 1994, into: On the equilibrium of living beings by
means of reciprocal destruction), in S.A. Levin (eds.) Frontiers in Mathematical
Biology, pp. 360-380.
References 317
Canfield, D.E.J., Shireman, J.V., Colle, D.E. and Haller, (1984) Prediction of chloro-
phyll a concentrations in Florida lakes importance of aquatic macrophytes. Cana-
dian Journal of Fisheries and Aquatic Sciences, 41, 497-501.
Canfield, D.E.J., Langeland, K.A., Linda, S.B. and Haller, W.T. (1985) Relations
between water transparency and maximum depth of macrophyte colonization in
Jakes. Journal of Aquatic Plant Management, 23, 25-28.
Carpenter, S.R. (1980) Enrichment of Lake Wingra, Wisconsin USA, by submersed
macrophyte decay. Ecology, 61, 1145-1155.
Carpenter, S.R. (1981) Submersed vegetation an internal factor in lake ecosystem
succession. American Naturalist, 118, 372-383.
Carpenter, S.R., Kitchell, J.F. and Hodgson, J.R. (1985) Cascading trophic interac-
tions and lake productivity. Bioscience, 35, 634~39.
Carpenter, S.R., Kitchell, J.F., Hodgson, J.R., Cochran, P.A., Elser, J.J., Elser, M.M.,
Lodge, D.M., Kretchmer, D., He, X. and Ende, C.N. (1987) Regulation of lake
primary productivity by food web structure. Ecology, 68, 1863-1876.
Carpenter, S.R., Cottingham, K.L. and Schindler, D.E. (1992) Biotic feedbacks in
Jake phosphorus cycles. Trends in Ecology and Evolution, 7, 332-336.
Carpenter, S.R., Lathrop, R.C. and Munoz-Del-Rio, A. (1993) Comparison of dy-
namic models for edible phytoplankton. Canadian Journal of Fisheries and
Aquatic Sciences, 50, 1757-1767.
Carpenter, S.R. and Lodge, D.M. (1986) Effects of submersed macrophytes on
ecosystem processes. Aquatic Botany, 26, 341-370.
Carpenter, S.R. and Pace, M.L. (1997) Dystrohy and eutrophy in lake ecosystems:
implications of fluctuating inputs. Oikos, 78, 3-14.
Carper, G.L. and Bachmann, R.W. (1984) Wind resuspension of sediments in a
prairie lake. Canadian Journal of Fisheries and Aquatic Sciences, 41, 1763-1767.
Carr, N.G. and Whitton, B.A. (1982) The Biology of Cyanobacteria. Blackwell
Scientific, Oxford
Carvalho, L. (1994) Top-down control of phytoplankton in a shallow hypertrophic
lake Little Mere (England). Hydrobiologia, 276, 53~3.
Chambers, P.A. (1987) Light and nutrients in the control of aquatic plant community
structure: II. In situ observations. Journal of Ecology, 75, 621~28.
Chambers, P.A. and Kalff, J. (1985a) The influence of sediment composition and
irradiance on the growth and morphology of Myriophyllum spicatum. Aquatic
Botany, 22, 253-264.
Chambers, P.A. and Kalff, J. (1985b) Depth distribution and biomass of submersed
aquatic macrophyte communities in relation to Secchi depth. Canadian Journal of
Fisheries and Aquatic Sciences, 42, 701-709.
Chen, R.L. and Barko, J.W. (1988) Effects of freshwater macrophytes on sediment
chemistry. Journal of Freshwater Ecology, 4, 279-290.
Chigbu, P. and Sibley, T.H. (1994) Relationship between abundance, growth, egg
size and fecundity in a landlocked population of Iongtin smelt, Spirinchus
thaleichthys.Journal of Fish Biology, 45, 1-15.
Colebrook, J.M. (1960) Some observations of zooplankton swarms in Windermere.
Journal of Animal Ecology, 29,241-242.
Connell, J.H. and Sousa, W.P. (1983) On the evidence needed to judge ecological
stability or persistence. American Naturalist, 121, 789-824.
Connor, E.F. and Simberloff, D. (1979) The assembly of species communities:
chance or competition? Ecology, 60, 1132-1140.
318 References
Cooke, G.D., Welch, E.B., Peterson, S.A. and Newroth, P.R. (1993) Restoration and
Management of Lakes and Reservoirs. Lewis Publishers, Boca Raton
Coops, H. (1996) Helophyte Zonation, Impact of Water Depth and Wave Exposure.
RIZA, Lelystad, The Netherlands, Thesis, 150.
Cottingham, K.L. {1996) Phytoplankton responses to whole-lake manipulations of
nutrients and food webs. University of Wisconsin, Madison, USA
Crawford, S.A. (1979) Farm pond restoration using Chara vulgaris vegetation.
Hydrobiologia, 62, 17-32.
Crivelli, AJ. (1983) The destruction of aquatic vegetation by carp Cyprinus carpio a
comparison between southern France and the USA. Hydrobiologia, 106, 37-42.
Cuddington, K.M. and McCauley, E. (1994) Food-dependent aggregation and mo-
bility of the water fleas Ceriodaphnia dubia and Daphnia pulex. Canadian Journal
of Zoology, 72, 1217-1226.
Daldorph, P.W.G. and Thomas, J.D. (1995) Factors influencing the stability of
nutrient-enriched freshwater macrophyte communities: The role of stickle-
backs Pungitius pungitius and freshwater snails. Freshwater Biology, 33, 271-
289.
Davies, J. (1985) Evidence for a diurnal horizontal migration in Daphnia hyalina
lacustris. Hydrobiologia, 120, 103-106.
Davis, R.B. (1974) Tubificids alter profiles of redox potential and pH in profunda!
lake sediment. Limnology and Oceanography, 19, 342-346.
De Groot, W.T. (1981) Phosphate and wind in a shallow lake. Archiv for
Hydrobiologie, 91, 475-489.
DeRoos, A.M., McCauley, E. and Wilson, W.G. (1991) Mobility versus density-
limited predator-prey dynamics on different spatial scales. Proceedings of the
Royal Society of London Series B - Biological Sciences, 246, 117-122.
De Roos, A.M., Diekmann, 0. and Metz, J.A.J. (1992) Studying the dynamics of
structured population models a versatile technique and its application to
Daphnia. American Naturalist, 139, 123-147.
De Stasio, B.T., Jr, (1990) The role of dormancy and emergence patterns in the
dynamics of a freshwater zooplankton community. Limnology and Oceanogra-
phy, 1079-1090.
De Stasio, B.T., Jr, Rudstam, L.G., Haning, A., Soranno, P. and Allen, Y.C. (1995)
An in situ test of the effects of food quality on Daphnia population growth.
Hydrobiologia, 307,221-230.
DeAngelis, D.L., Goldstein, R.A. and O'Neill, R.V. (1975) A model for trophic
interaction. Ecology, 56, 881-892.
DeAngelis, D.L., Cox, D.C. and Coutant, C.C. (1979) Cannibalism and size dispersal
in young-of-the-year largemouth bass: experiments and model. Ecological Mod-
elling, 8, 133-148.
DeAngelis, D.L. and Gross, L.J. (1992) Individual-Based Models and Approaches in
Ecology. Chapman and Hall, New York
DeAngelis, D.L. and Rose, K.A. (1992) Which individual-based approach is most
appropriate for a given problem? in D.L. DeAngelis and L.J. Gross (eds.) Indi-
vidual-based Models and Approaches in Ecology, Chapman and Hall, New York,
pp. 509-520.
Delgado, M., DeJonge, V.N. and Peletier, H. (1991) Experiments on resuspension
of natural microphytobenthos populations. Marine Biology, 108, 321-328.
Demeester, L., Weider, L.J. and Tollrian, R. (1995) Alternative antipredator de-
References 319
fences and genetic polymorphism in a pelagic predator-prey system. Nature, 378,
483-485.
Demelo, R., France, R. and McQueen, D.J. (1992) Biomanipulation: Hit or myth?
Limnology and Oceanography, 192-207.
Demott, W.R. (1982) Feeding selectivities and relative ingestion rates of Daphnia
and Bosmina. Limnology and Oceanography, 27, 518-527.
Diehl, S. (1988) Foraging efliciency of three freshwater lishes effects of structural
complexity and light. Oikos, 53, 207-214.
Dieter, C.D. (1990) The importance of emergent vegetation in reducing sediment
resuspension in wetlands. Journal of Freshwater Ecology, 5, 467-474.
Dodson, S. (1988) The ecological role of chemical stimuli for the zooplankton
predator-avoidance behavior in Daphnia. Limnology and Oceanography, 33,
1431-1439.
Doveri, F., Scheffer, M., Rinaldi, S., Muratori, S. and Kuznetsov, Y.A. (1993)
Seasonality and chaos in a plankton-fish model. Theoretical Population Biology,
43, 159-183.
Dvorak, J. (1987) Production-ecological relationships between aquatic vascular
plants and invertebrates in shallow waters and wetlands, a review. Ergebnisse der
Limnologie, 181-184.
Dvorak, J. and Best, E.P.H. (1982) Macro invertebrate communities associated with
the macrophytes of Lake Vechten the Netherlands structural and functional
relationships. Hydrobiologia, 95, 115-126.
Edmondson, W.T. (1991) The Uses of Ecology: Lake Washington and Beyond.
University of Washington Press, Seattle
Einsele, W. (1936) Ober die Beziehungen des Eisenkreislaufs zum
Phosphatkreislauf im eutrophen See. Archiv fUr Hydrobiologie, 29, 664-686.
Einsele, W. (1938) Ober chemische und kolloidchemische Vorgiinge in Eisen-
Phosphat-Systemen unter limnochemischen und limnogeologischen Gesi-
chtspunkten. Archiv fUr Hydrobiologie, 33, 361-387.
Eklov, P. and Persson, L. (1995) Species-specific antipredator capacities and prey
refuges: Interactions between piscivorous perch (Perea fiuviatilis) and juvenile
perch and roach (Rutilus rutilus). Behavioral Ecology and Sociobiology, 37, 169-
178.
Eklov, P. and Persson, L. (1996) The response of prey to the risk of predation:
Proximate cues for refuging juvenile fish. Animal Behaviour, 51, 105-115.
Engel, S. (1988) The role and interactions of submersed macrophytes in a shallow
Wisconsin Lake USA. Journal of Freshwater Ecology, 4, 329-342.
Engel, S. and Nichols, S.A. (1994) Aquatic macrophyte growth in a turbid windswept
lake. Journal of Freshwater Ecology, 9, 97-109.
Engelmayer, A. (1995) Effects of predator-released chemicals on some life history
parameters of Daphnia pulex. Hydrobiologia, 307, 203-206.
Evans, R.D. (1994) Empirical evidence of the importance of sediment resuspension
in Jakes. Hydrobiologia, 284, 5-12.
Everall, N.C. and Lees, D.R. (1996) The use of barley-straw to control general and
blue-green algal growth in a Derbyshire reservoir. Water Research, 30, 269-276.
Fogg, G.E., Steward, W.D.P., Fay, P. and Walsby, A.E. (1973) The Blue-Green
Algae. Academic Press, London
Forsberg, C. (1964) Phosphorus, a maximum factor in the growth of Characeae.
Nature, 201, 517-518.
320 References
Forsberg, C. (1965) Nutritional studies Chara in axenic cultures. Physiologia
Plantarum, 18, 275-290.
Forsberg, C., Kleiven, S. and Willen, T. (1990) Absence of allelopathic effects of
Chara on phytoplankton in situ. Aquatic Botany, 38, 289-294.
French, J.R.P. (1993) How well can fishes prey on zebra mussels in eastern north
America? Fisheries (Bethesda), 18, 13-19.
Gallepp, G.W. (1979) Chironomid influence on phosphorus release in sediment-
water microcosms. Ecology, 60, 547-556.
Gibson, M.T., Welch, I.M., Barrett, P.R.F. and Ridge, I. (1990) Barley straw as an
inhibitor of algal growth II: laboratory studies. Journal of Applied Phycology, 2,
241-248.
Giles, N. (1987) Differences in the ecology of wet-dug and dry-dug gravel pit lakes.
The Game Conservancy Annual Review, 18, 130--133.
Giles, N. (1992) Wildlife After Gravel; Twenty Years of Practical Research by the
Game Conservancy and ARC. Game Conservancy Limited, Fordingbridge,
Hampshire, UK
Gilinsky, E. (1984) The role of fish predation and spatial heterogeneity in determin-
ing benthic community structure. Ecology, 65, 455-468.
Gilpin, M.E. (1972) Enriched predator-prey systems: Theoretical stability. Science,
177' 902-904.
Gliwicz, Z.M. (1986) Predation and the evolution of vertical migration in
zooplankton. Nature, 320, 746--748.
Gliwicz, Z.M. (1990) Why do cladocerans fail to control algal blooms?
Hydrobiologia, 200/201, 83-98.
Gliwicz, Z.M. and Lampert, W. (1990) Food thresholds in Daphnia species in the
absence and presence of blue-green filaments. Ecology, 71, 691-702.
Godmaire, H. and Planas, D. (1983) Potential effect of Myriophyllum spicatum on
the primary production of phytoplankton, in Anonymous(eds.) Periphyton of
Freshwater Ecosystems, Dr. W. Junk Publishers, The Hague, pp.227-232.
Gons, H.J., Gulati, R.D. and VanLiere, L. (1986) The eutrophic Loosdrecht lakes:
current ecological research and restoration perspectives. Hydrobiological bulle-
tin, 20, 61-75.
Gons, H.J., Otten, J.H. and Rijkeboer, M. (1991) The significance of wind
resuspension for the predominance of filamentous cyanobacteria in a shallow,
eutrophic lake. Memorie dell'Istituto Italiano di Idrobiologia, 233-249.
Goulder, R. (1969) Interactions between the rates of production of a freshwater
macrophyte and phytoplankton in a pond. Oikos, 20, 300--309.
Gragnani, A., Scheffer, M., Rinaldi, S. (1997) Top-down control of cyanobacteria a
theoretical analysis. Submitted.
Graneli, W. (1979) The influence of Chironomus plumosus larvae on the oxygen
uptake of sediment. Archiv fUr Hydrobiologie, 87, 385-403.
Gregg, W.W. and Rose, F.L. (1985) Influences of aquatic macrophytes on inverte-
brate community structure guild structure and microdistribution in streams.
Hydrobiologia, 128, 45-56.
Griffiths, R.W. (1992) Effects of zebra mussels (Dreissena polymorpha) on the
benthic fauna of Lake St. Clair, in T.F. Nalepa and D.W. Schloesser (eds.) Zebra
Mussels - Biology, Impacts and Control, Lewis Publishers, London, pp. 415-
437.
References 321
Grimm, M.P. (1981a) Intraspecific predation as a principal factor controlling the
biomass of northern pike Esox lucius. Fish Management, 12, 77-80.
Grimm, M.P. (1981b) The composition of northern pike Esox lucius populations
in four shallow waters in the Netherlands with special reference to factors in-
fluencing pike in their first growing season biomass. Fish Management, 12, 61-
76.
Grimm, M.P. (1983) Regulation of biomasses of small (<41 em) northern pike (Esox
lucius L.), with special reference to the contribution of individuals stocked as
fingerlings (4--6cm). Fish Management, 14, 115-133.
Grimm, M.P. (1989) Northern pike (Esox lucius L.) and aquatic vegetation, tools in
the management of fisheries and water quality in shallow waters. Hydrobiological
bulletin, 23, 59-65.
Grimm, M.P. and Backx, J.J .G.M. (1990) The restoration of shallow eutrophic lakes
and the role of northern pike, aquatic vegetation and nutrient concentration.
Hydrobiologia, 200/201, 557-566.
Gross, E.M. and Siitfeld, R. (1994) Polyphenols with algicidal activity in the sub-
merged macrophyte Myriophyllum spicatum L. Acta Horticultura, 381, 710--716.
Gulati, R.D. (1983) Zooplankton and its grazing as indicators of trophic status in
Dutch lakes. Environmental Monitoring and Assessment, 343-354.
Gulati, R.D., Lammens, E.H.R.R., Meijer, M.-L. and Van Donk, E. (1990)
Biomanipulation Tool for Water Management. Proceedings of an International
Converence held in Amsterdam, The Netherlands, 8-11 August 1989. Kluwer
Academic Publishers, Dordrecht, Boston, London
Gunatilaka, A. (1982) Phosphate adsorption kinetics of resuspended sediments in a
shallow lake Neusiedlersee, Austria. Hydrobiologia, 91-92, 293-298.
Gurney, W.S.C. and Nisbet, R.M. (1978) Predator prey fluctuations in patchy envi-
ronments. Journal of Animal Ecology, 47, 85-102.
Guy, M., Taylor, W.D. and Carter, J.C.H. (1994) Decline in total phosphorus in the
surface waters of lakes during summer stratification, and its relationship to size
distribution of particles and sedimentation. Canadian Journal of Fisheries and
Aquatic Sciences, 51, 1330--1337.
Haertel, L. and Jongsma, D. (1982) Effect of winterkill on the water quality of
prairie lakes. Proc. S. D. Acad. Sci. 61, 134-151.
Hairston, N., Smith, F.E. and Slobodkin, D. (1960) Community structure, population
control, and competition. American Naturalist, 94, 421-425.
Hairston, N.G.J. (1987) Diapause as a predator-avoidance adapation, in W.C.
Kerfoot and A. Sih (eds.) Predation, Direct and Indirect Impacts on Aquatic
Communities, University Press of New England, Hannover, pp. 281-290.
Hakkari, L. and Bagge, P. (1984) On the fry densities of pike (Esox lucius L.) in Lake
Saimaa, Finland. V erhandlungen Internationale V ereinigung Theoretisch
Angewandte Limnologie, 22, 2560--2565.
Hambright, K.D. (1994) Can zooplanktivorous fish really affect lake thermal dynam-
ics? Archiv for Hydrobiologie, 130, 429-438.
Hamilton, D.J., Ankney, C.D. and Bailey, R.C. (1994) Predation of zebra mussels by
diving ducks: an exclosure study. Ecology, 75, 521-531.
Hamilton, D.P. and Mitchell, S.F. (1996) An empirical model for sediment
resuspension in shallow lakes. Hydrobiologia, 317, 209-220.
Haney, J.F., Forsyth, D.J. and James, M.R. (1994) Inhibition of zooplankton filtering
322 References
rates by dissolved inhibitors produced by naturally occurring cyanobacteria.
Archiv fUr Hydrobiologie, 132, 1-13.
Hanson, J.M. and Leggett, W.C. (1982) Empirical prediction of fish biomass and
yield. Canadian Journal of Fisheries and Aquatic Sciences, 39, 257-263.
Hanson, M.A., Butler, M.G., Richardson, J.L. and Arndt, J.L. (1990) Indirect effects
of fish predation on calcite supersaturation precipitation and turbidity in a shal-
low prairie lake. Freshwater Biology, 24, 547-556.
Hanson, M.A. and Butler, M.G. (1990) Early responses of plankton and turbidity to
biomanipulation in a shallow prairie lake. Hydrobiologia, 20()....201, 317-328.
Hanson, M.A. and Butler, M.G. (1994a) Responses of plankton, turbidity, and
macrophytes to biomanipulation in a shallow prairie lake. Canadian Journal of
Fisheries and Aquatic Sciences, 51, 118()....1188.
Hanson, M.A. and Butler, M.G. (1994b) Responses to food web manipulation in a
shallow waterfowl lake. Hydrobiologia, 280, 457-466.
Hargeby, A., Andersson, G., Blindow, I. and Johansson, S. (1994) Trophic web
structure in a shallow eutrophic lake during a dominance shift from
phytoplankton to submerged macrophytes. Hydrobiologia, 280, 83-90.
Harper, D.M. and Ferguson, A.J.D. (1982) Zooplankton and their relationships with
water quality and fisheries. Hydrobiologia, 88, 135-145.
Harris, G.P. (1986) Phytoplankton Ecology: Structure, Function and Fluctuation.
Chapman and Hall, London, New York
Hasler, A.D. and Jones, E. (1949) Demonstration of the antagonistic action of large
aquatic plants on algae and rotifers. Ecology, 30, 346--359.
Hassel, M.P. and May, R.M. (1974) Aggregation of predators and insect parasites
and its effect on stability. Journal of Animal Ecology, 43, 567-594.
Havel, J.E. (1987) Predator-induced defenses: a review, in W.C. Kerfoot and A. Sib
(eds.) Predation, Direct and Indirect Impacts on Aquatic Communities, University
Press of New England, Hannover, pp. 263-278.
Havens, K.E. (1991) Fish-induced sediment resuspension: Effects on phytoplankton
biomass and community structure in a shallow hypereutrophic lake. Journal of
Plankton Research, 13, 1163-1176.
Havens, K.E. (1993) Responses to experimental fish manipulations in a shallow,
hypereutrophic lake: The relative importance of benthic nutrient recycling and
trophic cascade. Hydrobiologia, 254, 73-!!0.
Hawkins, P. and Lampert, W. (1989) The effect of Daphnia body size on filtering
rate inhibition in the presence of a filamentous cyanobacterium. Limnology and
Oceanography, 34, 1084--1089.
He, X. and Wright, R.A. (1992) An experimental study of piscivore-planktivore
interactions population and community responses to predation. Canadian Jour-
nal of Fisheries and Aquatic Sciences, 49, 1176--1183.
Heller, R. and Milinski, M. (1979) Optimal foraging of sticklebacks Gasterosteus
aculeatus on swarming prey. Animal Behaviour, 27, 1127-1141.
HellstrOm, T. (1991) The effect of resuspension on algal production in a shallow
lake. Hydrobiologia, 213, 183-190.
Hessen, D.O. (1990) Carbon, nitrogen and phosphorus status in Daphnia at varying
food conditions. Journal of Plankton Research, 12, 1239-1249.
Hessen, D.O. and Van Donk, E. (1993) Morphological changes in Scenedesmus
induced by substances released by Daphnia. Archiv fUr Hydrobiologie, 127, 129-
140.
Heyman, U. and Lundgren, A. (1988) Phytoplankton biomass and production in
relation to phosphorus – Some conclusions from field studies. Hydrobiologia,
170, 211–227.
Hill, D., Wright, R. and Street, M. (1987) Survival of mallard ducklings Anas
platyrhynchos and competition with fish for invertebrates on a flooded gravel
quarry in England UK. Ibis, 129, 159–167.
Hodgson, R.H. (1966) Growth and carbohydrate status of Sago pondweed. Weeds,
14, 263–268.
Hogeweg, P. and Hesper, B. (1990) Individual-oriented modelling in ecology.
Mathematical and Computer Modelling, 13, 83–90.
Holland, R.E. (1993) Changes in planktonic diatoms and water transparency in
Hatchery Bay, Bass Island Area, western Lake Erie since the establishment of
the zebra mussel. Journal of Great Lakes Research, 19, 617–624.
Holt, R.D. (1977) Prediction apparent competition and the structure of prey
communities. Theoretical Population Biology, 12, 197–229.
Hootsmans, M.J.M. and Vermaat, J.E. (1991) Macrophytes, a key to understanding
changes caused by eutrophication in shallow freshwater ecosystems. Thesis,
IHE Delft 1–412.
Hootsmans, M.J.M. and Breukelaar, A.W. (1990) De invloed van waterplanten op
de groei van algen. H2O, 23, 264–266.
Hootsmans, M.J.M. and Vermaat, J.E. (1985) The effect of periphyton grazing by
three epifaunal species on the growth of Zostera marina under experimental
conditions. Aquatic Botany, 22, 83–88.
Hosper, S.H. (1980) Development and practical application of limiting values for
the phosphate concentration in surface waters in the Netherlands. Hydrobiolog-
ical bulletin, 14, 64–72.
Hosper, S.H. (1985) Restoration of Lake Veluwe the Netherlands by reduction of
phosphorus loading and flushing. Water Science and Technology, 17, 757–768.
Hosper, S.H., Meijer, M.-L. and Walker, P.A. (1992) Handleiding Actief Biologisch
Beheer – Beoordeling van de mogelijkheden van visstandbeheer bij het herstel
van meren en plassen. RIZA, OVB, Lelystad, The Netherlands.
Hosper, S.H. and Meyer, M.-L. (1986) Control of phosphorus loading and flush-
ing as restoration methods for Lake Veluwe, the Netherlands. Hydrobiological
bulletin, 20, 183–194.
Hosper, S.H. and Meyer, M.-L. (1993) Biomanipulation, will it work for your lake?
A simple test for the assessment of chances for clear water, following drastic
fishstock reduction in shallow, eutrophic lakes. Ecological Engineering, 63–72.
Houthuijzen, R.P., Backx, J.J.G.M. and Buijse, A.D. (1993) Exceptionally rapid
growth and early maturation of perch in a freshwater lake recently converted
from an estuary. Journal of Fish Biology, 43, 320–324.
Howard, R.K. and Short, F.T. (1986) Seagrass Halodule wrightii growth and
survivorship under the influence of epiphyte grazers. Aquatic Botany, 24,
287–302.
Howarth, R.W., Marino, R., Lane, J. and Cole, J. (1988) Nitrogen fixation in fresh-
water estuarine and marine ecosystems 1. rates and importance. Limnology and
Oceanography, 33, 669–687.
Hoyer, M.V., Gu, B. and Schelske, C.L. (1997) Sources of organic carbon in the
foodwebs of two Florida lakes indicated by stable isotopes, in E. Jeppesen, M.
324 References
Ss;ndergaard, M. Ss;ndergaard and K. Christoffersen (eds.) The structuring role of
submerged macrophytes in lakes, Springer-Verlag, New York, in press.
Hoyer, M.V. and Canfield, D.E. (1994) Bird abundance and species richness on
F1orida lakes - Influence of trophic status, lake morphology, and aquatic
macrophytes. Hydrobiologia, 280, 107-119.
Hoyer, M.V. and Jones, J.R. (1983) Factors affecting the relation between phospho-
rus and chlorophyll a in USA midwestern reservoirs. Canadian Journal of Fisher-
ies and Aquatic Sciences, 40, 192-199.
Hrba~ek, J., Dvol'akova, M., Korinek, V. and Prochazkova, L. (1961) Demonstra-
tion of the effect of the fish stock on the species composition of zooplankton and
the intensity of metabolism of the whole plankton association. Verhandlungen
Internationale Vereinigung Theoretisch Angewandte Limnologie, 14, 192-195.
Huisman, J. and Weissing, F.J. (1994) Light-limited growth and competition for light
in well-mixed aquatic environments- An elementary model. Ecology, 75, 507-
520.
Hunter, M.L.J., Jones, J.J., Gibbs, K.E. and Moring, J.R. (1986) Duckling responses
to lake acidification do black ducks Anas rubripes and fish compete. Oikos, 47,
26-32.
Huston, M.A., DeAngelis, D.L. and Post, W. (1988) New computer models unify
ecological theory. Bioscience, 38, 682-691.
Hutchinson, G.E. (1961) The paradox of the plankton. American Naturalist, 95, 137-
145.
Hutchinson, G.E. (1967) A Treatise on Limnology. Volume II, Introduction to Lake
Biology and the Limnoplankton. John Wiley and Sons, New York
Hutchinson, G.E. (1975) A Treatise on Limnology. Volume III, Limnological
Botany. John Wiley and Sons, New York
Ibelings, B.W., Mur, L.R. and Walsby, A.E. (1991) Diurnal changes in buoyancy and
vertical distribution in populations of Microcystis in two shallow lakes. Journal of
Plankton Research, 13, 419-436.
Ibelings, B.W. (1992) Cyanobacterial waterblooms: the role of buoyancy in
watercolumns of varying stability. Thesis University of Amsterdam, 172 pp
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 photo-
synthetic photon flux densities, simulation light regimes induced by mixing in
lakes. New Phytologist, 128, 407-424.
Ikusima, I. (1970) Ecological studies on the productivity of aquatic plant communi·
ties IV. Light condition and community photosynthetic production. Botanical
Magazine Tokyo, 83, 330-341.
Irvine, K., Moss, B. and Stansfield, J. (1990) The potential of artificial refugia for
maintaining a community of large-bodied Cladocera against fish predation in a
shallow eutrophic lake. Hydrobiologia, 200-201, 379-390.
Jackson, H.O. and Starrett, W.C. (1959) Turbidity and sedimentation at Lake
Chautauqua, Illinois. Journal of Wildlife Management, 23,157-168.
Jacobsen, D. and Sand-Jensen, K. (1992) Herbivory of invertebrates on submerged
macrophytes from Danish freshwaters. Freshwater Biology, 28, 301-308.
Jakobsen, P.J. and Johnsen, G.H. (1987) The influence of predation on horizontal
distribution of zooplankton species. Freshwater Biology, 17,501-508.
James, W.F. and Barko, J.W. (1990) Macrophyte influences on the zonation of
References 325
sediment accretion and composition in a north-temperate reservoir. Archiv far
Hydrobiologie, 120, 129-142.
Jasser, I. (1995) The influence of macrophytes on a phytoplankton community in
experimental conditions. Hydrobiologia, 306, 21-32.
Jensen, J.P., Kristensen, P. and Jeppesen, E. (1991) Relationships between N loading
and in-lake N concentrations in shallow Danish lakes. Verhandlungen
Internationale Vereinigung Theoretisch Angewandte Limnologie, 24,201-204.
Jensen, J.P., 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.
Jensen, J.P., Jeppesen, E., Olrik, K. and Kristensen, P. (1994) Impact of nutrients
and physical factors on the shift from cyanobacterial to chlorophyte dominance in
shallow Danish lakes. Canadian Journal of Fisheries and Aquatic Sciences, 51,
1692-1699.
Jeppesen, E., Jensen, J.P., Kristensen, P., S!l)ndergaard, M., Mortensen, E., Sortkjrer,
0. and Olrik, K. (1990a) Fish manipulation as a lake restoration tool in shallow,
eutrophic, temperate lakes 2: threshold levels, long-term stability and conclu-
sions. Hydrobiologia, 200/201, 219-228.
Jeppesen, E., S!l)ndergaard, M., Sortkjrer, 0., Mortensen, E. and Kristensen, P.
(1990b) Interactions between phytoplankton zooplankton and fish in a shallow
hypertrophic lake a study of phytoplankton collapses in Lake Sobygaard Den-
mark. Hydrobiologia, 191, 149-164.
Jeppesen, E., Kristensen, P., Jensen, J.P., S!l)ndergaard, M., Mortensen, E. and
Lauridsen, T.L. (1991) Recovery resilience following a reduction in external
phosphorus loading of shallow eutrophic Danish lakes: duration, regulating fac-
tors and methods for overcoming resilience. Memorie dell'Istituto Italiano di
ldrobiologia, 48, 127-148:
Jeppesen, E., S!l)ndergaard, M., Kanstrup, E. and Petersen, B. (1994) Does the
impact of nutrients on the biological structure and function of brackish and
freshwater lakes differ. Hydrobiologia, 276, 15-30.
Jeppesen, E., Jensen, J.P., S!1lndergaard, M., Lauridsen, T.L., Pedersen, L.J. and
Jensen, L. (1996) Top-down control in freshwater lakes with special emphasis on
the role of fish, submerged macrophytes and water depth. Hydrobiologia,-in
press.
Jeppesen, E., S!l)ndergaard, M., Jensen, J.P., Kanstrup, E. and Pedersen, B. (1997)
Macrophytes and turbidity in brackish lakes, with special emphasis on the role of
top-down control, in E. Jeppesen, M. S!1lndergaard, M. S!l)ndergaard and K.
Kristoffersen (eds.) The structuring role of submerged macrophytes in lakes,
Springer Verlag, New York, in press.
Johnson, D.S. and Chua, T.E. (1973) Remarkable schooling behavior of a water-flea
Moina sp., Cladocera. Crustaceana, 24, 332-333.
Jones, R.C. (1990) The effect of submersed aquatic vegetation on phytoplankton
and water quality in the tidal freshwater Potomac River USA. Journal of Fresh-
water Ecology, 5, 279-288.
Jupp, B.P. and Spence, D.H.N. (1977) Limitations of macrophytes in a eutrophic
lake Loch Leven, Scotland. Part 2. Wave action sediments and waterfowl grazing.
Journal of Ecology, 65, 431-446.
Jurgens, K. and Stolpe, G. (1995) Seasonal dynamics of crustacean zooplankton,
326 References
heterotrophic nanoflagellates and bacteria in a shallow, eutrophic lake. Freshwa-
ter Biology, 33, 27-38.
Kairesalo, T., Komij6w, R. and Luokkanen, E. (1997) Trophic cascade structuring a
plankton community in a strongly vegetated lake littoral, in E. Jeppesen, M.
S!21ndergaard, M. S12indergaard and K. Christoffersen (eds.) The structuring role of
submerged macrophytes in lakes, Springer-Verlag, New York, in press.
Kallio, K. (1994) Effect of summer weather on internal loading and chlorophyll a in
a shallow lake: a modeling approach. Hydrobiologia, 275-276, 371-378.
Kautsky, L. (1987) Life-cycles of three populations of Potamogeton pectinatus L. at
different degrees of wave exposure in the Asko Area, Northern Baltic Proper.
Aquatic Botany, 27, 177-186.
Keating, K.J. (1977) Allelopathic influence on blue-green bloom sequence in a
eutrophic lake. Science, 196, 885-887.
Keating, K.J. (1978) Blue-green algal inhibition of diatom growth: Transition ftom
mesotrophic to eutrophic community structure. Science, 199, 971-973.
Keith, L.B. (1983) Role of food in hare population cycles. Oikos, 385-395.
Keith, L.B. (1990) Dynamics of snowshoe hare populations. Current Mammalogy, 2,
119-195.
Kersting, K. (1985) Development and use of an aquatic micro-ecosystem as a test
system for toxic substances properties of an aquatic micro-ecosystem 4.
Internationale Revue der Gesamten Hydrobiologie, 69, 567-607.
King, D.R. and Hunt, G.S. (1967) Effect of carp on vegetation in Lake Erie Marsh.
Journal of Wildlife Management, 31, 181-188.
Ki!21rboe, T. (1980) Distribution and production of submerged macrophytes in Tip-
per Grund, Ringkobing Fjord, Denmark, and the impact of waterfowl grazing.
Journal of Applied Ecology,17, 675-688.
Kipling, C. (1983) Changes in the population of pike Esox lucius in Windermere
England UK 1944--1981. Journal of Animal Ecology, 52, 989-1000.
Kirk, J.T.O. (1986) Optical limnology -A manifesto, in P. De Deckker and W.D.
Williams (eds.) Limnology in Australia, Dr W. Junk Publishers, Dordrecht,
pp.33-62.
Kirk, J.T.O. (1994) Light and photosynthesis in aquatic ecosystems. Cambridge Uni-
versity Press, Cambridge
Kirk, K.L. (1991) Inorganic particles alter competition in grazing plankton the role
of selective feeding. Ecology, 72, 915-923.
Kirk, K.L. and Gilbert, J.J. (1990) Suspended clay and the population dynamics of
planktonic rotifers and cladocerans. Ecology, 71, 1741-1755.
Kitchell, J.F., Koonce, J.F. and Tennis, P.S. (1975) Phosphorus flux through fishes.
Verhandlungen Internationale Vereinigung Theoretisch Angewandte Limnologie,
19, 2478-2484.
Kitchell, J.F., Johnson, M.G., Minns, C.K., Loftus, K.H., Greig, L. and Olver, C.H.
(1977) Percid habitat the river analogy. J Fish Res Board Can, 34, 1936-1940.
Klemetsen, A. (1970) Plankton swarms in Lake Gjokvatn East Finnmark. Astarte. 3,
83-85.
Kogan, S.I. and Chinnova, G.A. (1972) Relations between Ceratophyllum demersum
and some blue-green algae. Hydrobiol. J. (Engl. Trans/. Gidrobiol. Zh). 8, 14--19.
Kooijman, S.A.L.M. (1986) Population dynamics on the basis of budgets, in J.A.J.
Metz and 0. Diekmann (eds.) The dynamics of physiologically structured
populations, Springer Verlag, Berlin, pp. 266-297.
References 327
Kornij6w, R., Gulati, R.D. and Ozimek, T. (1995) Food preference of freshwater
invertebrates: Comparing fresh and decomposed angiosperm and a filamentous
alga. Freshwater Biology, 33, 205-212.
Kretzschmar, M., Nisbet, R.M. and McCauley, E. (1993) A predator-prey model for
zooplankton grazing on competing algal populations. Theoretical Population Bi-
ology, 44, 32~6.
Kristensen, P., S!l)ndergaard, M. and Jeppesen, E. (1992) Resuspension in a shallow
eutrophic lake. Hydrobiologia, 228, 101-109.
Kuenne, C. (1925) Uber Schwarmbildung bei Bosmina longirostris o.f.m. Archiv fti.r
Hydrobiologie, 16, 508
Kufel, L. and Ozimek, T. (1994) Can Chara control phosphorus cycling in Lake
Luknajno (Poland). Hydrobiologia, 276, 277-283.
Kuznetsov, Y.A. (1995) Elements of Applied Bifurcation Theory. Springer-Verlag,
New York
Lair, N. and Ayadi, H. (1989) The seasonal succession of planktonic events in Lake
Aydat, France: a comparison with the PEG model. Archiv fti.r Hydrobiologie,
115,58~02.
Lamarra, V.A. (1974) Digestive activities of carp as a major contributor to the
nutrient loading of lakes, in K.E. Marshall (eds.) Xix Congress International
Association of Limnology, Winnipeg, Canada, Aug. 22-29, 1974. 238P, Freshwa-
ter Institute, Department of Environment
Lammens, E.H.R.R. (1985) A test of a model for planktivorous filter feeding by
bream Abramis brama. Environmental Biology of Fishes, 13, 289-296.
Lammens, E.H.R.R., DeNie, H.W., Vijverberg, J. and Densens, W.L.T. (1985) Re-
source partitioning and niche shifts of bream (Abramis brama) and eel (Anguilla
anguilla) mediated by p.redation of smelt (Osmerus eperlanui) on Daphnia
hyalina. Canadian Journal of Fisheries and Aquatic Sciences, 42, 1342-135i.
Lammens, E.H.R.R. (1989) Causes and consequences of the success of bream in
Dutch eutrophic lakes. Hydrobiological bulletin, 23, 11-18.
Lammens, E.H.R.R. (1991) Diets and feeding behaviour, in I.J. Winfield and J.S.
Nelson (eds.) Cyprinid Fishes: Systematics, Biology and Exploitation, Chapman
and Hall, London, pp.353--376.
Lammens, E.H.R.R., Buijse, T., Butijn, G., Blaakman, E., Buiteveld, H., Dekker,
W., Hosper, S.H., Vander Hut, R. and Slingerland, T. (1996) Relation of chloro-
phyll-a and suspended matter to transparency and fish biomass in IJsselmeer and
Markermeer in 1970-1992. Limiting factors for the carrying capacity of both
ecosystems. Verhandlungen Intemationale Vereinigung Theoretisch Angewandte
Limnologie, 26,
Lammens, E.H.R.R., Klein Breteler, J.G.P., Mooij, W.M. and Claassen, T. (1997)
The direct and indirect effects of a seine and gill-net fishery on fish stock and
water quality in the Frisian Lakes, The Netherlands. Canadian Journal of Fisher-
ies and Aquatic Sciences, submitted,
Lampert, W., Fleckner, W., Rai, H. and Taylor, B.E. (1986) Phytoplankton control
by grazing zooplankton: A study on the spring clear-water phase. Limnology and
Oceanography, 31, 478-490.
Lampert, W. (1987a) Vertical migration of freshwater zooplankton indirect effects
of vertebrate predators on algal communities, in W.C. Kerfoot and A. Sih (eds.)
Predation, Direct and Indirect Impacts on Aquatic Communities, University Press
of New England, Hannover, pp. 291-299.
328 References
Lampert, W. (1987b) Laboratory studies on zooplankton-cyanobacteria inter-
actions. New Zealand Journal of Marine and Freshwater Research, 21, 483-
490.
Lampert, W. (1992) Bottom-up control of freshwater plankton communities applica-
tions of threshold growth and functional responses. Bulletin of the Ecological
Society of America, 73,
Lampert, W. and Rothhaupt, K.O. (1991) Alternating dynamics of rotifers and
Daphnia magna in a shallow lake. Archiv fiir Hydrobiologie, 120, 447-456.
Landers, D.H. (1982) Effects of naturally senescing aquatic macrophytes on nutrient
chemistry and chlorophyll a of surrounding waters. Limnology and Oceanogra-
phy, 27, 428-439.
Lauridsen, T.L., Jeppesen, E. and Andersen, F.O. (1993) Colonization of submerged
macrophytes in shallow fish manipulated Lake Vaeng impact of sediment compo-
sition and waterfowl grazing. Aquatic Botany, 46, 1-15.
Lauridsen, T.L., Jeppesen, E. and S!ilndergaard, M. (1994) Colonization and succes-
sion of submerged macrophytes in shallow Lake Vaeng during the first five years
following fish manipulation. Hydrobiologia, 276, 233-242.
Lauridsen, T.L., Pedersen, L.J., Jeppesen, E. and Sjijndergaard, M. (1996) The im-
portance of macrophyte bed size for cladoceran composition and horizontal
migration in a shallow lake. in prep.
Lauridsen, T.L. and Buenk, I. (1996) Die! changes in the horizontal distribution of
zooplankton in the littoral zone of two shallow eutrophic lakes. in prep.
Lauridsen, T.L. and Lodge, D.M. (1996) Avoidance by Daphnia magna of fish and
macrophytes: chemical cues and predator-mediated use of macrophyte habitat.
Limnology and Oceanography, 41,794-798.
Leach, J.H., Johnson, M.G., Kelso, J.R.M. and Hartmann, J. (1977) Responses of
percid fishes and their habitats to eutrophication. J Fish Res Board Can, 34, 1964-
1971.
Leentvaar, P. (1961) Hydrobiologische waarnemingen in het Veluwemeer. De
Levende Natuur, 64, 273-279.
Leentvaar, P. (1966) Plant en dier in het Veluwemeer. De Waterkampioen, 1164,18-
20.
Lehman, J.T. (1980) Release and cycling of nutrients between planktonic algae and
herbivores. Limnology and Oceanography, 25, 620-632.
Leibold, M.A. (1990) Resources and predators can affect the vertical distributions of
zooplankton. Limnology and Oceanography, 35, 938-944.
Lennox, L.J. (1984) Lough Ennell, Ireland, laboratory studies on sediment phospho-
rus release under varying mixing aerobic and anaerobic conditions. Freshwater
Biology, 14, 183-188.
Levins, R. (1966) The strategy of model building in population biology. American
Scientist, 54, 421-431.
Levitan, C., Kerfoot, W.C. and Demott, W.R. (1984) Ability of Daphnia to buffer
trout lakes against periodic nutrient inputs. Verhandlungen Internationale
Vereinigung Theoretisch Angewandte Limnologie, 22, 3076-3082.
Ligtvoet, W. and DeJong, S.A. (1995) Ecosystem development in Lake Volkerak-
Zoom: concept and strategy. Water Science and Technology, 31, 239-243.
Ligtvoet, W. and Grimm, M.P. (1992) Fish in clear water- Fish-stock development
and management in Lake Volkerak/Zoom. Proceedings and Information CHO-
TNO, 46, 69-84.
References 329
Lijklema, L. (1977) The role of iron in the exchange of phosphate between water and
sediments, in H,L, Golterman (eds.) Interactions between sediments and freshwa-
ter, Junk Pub!. The Hague, pp. 313-317.
Lijklema, L., Jansen, J.H. and Roijackers, R.M.M. (1989) Eutrophication in The
Netherlands. Water Science and Technology, 21, 1899-1891.
Lijklema, L. (1994) Nutrient dynamics in shallow lakes: effects of changes in loading
and role of sediment-water interactions. Hydrobiologia, 275-276, 335-348.
Lind, O.T., Davaloslind, L.O., Chrzanowski, T.H. and Limon, J.G. (1994) Inorganic
turbidity and the failure of fishery models. lnternationale Revue der Gesamten
Hydrobiologie, 79, 7-16.
Lindegaard, C. (1994) The role of zoobenthos in energy flow in two shallow lakes.
Hydrobiologia, 275-276, 313-322.
Lodge, D.M. (1991) Herbivory on freshwater macrophytes. Aquatic Botany, 41,195-
224.
Loose, C.J. and Dawidowicz, P. (1994) Trade-offs in die! vertical migration by
zooplankton: The costs of predator avoidance. Ecology, 75, 2255-2263.
Lotka, A.J. (1925) Elements of Physical Biology (Reprinted in 1956). Dover, New
York
Luecke, C., Vanni, M.J., Magnuson, J.J. and Kitchell, J.F. (1990) Seasonal regulation
of Daphnia populations by planktivorous fish implications for the spring clear-
water phase. Limnology and Oceanography, 35, 1718-1733.
Luecke, C., Lunte, C.C., Wright, R.A., Robertson, D. and McLain, A.S. (1996)
Impacts of variation in planktivorous fish on abundance of daphnids: a simulation
model of the Lake Mendota food web, in J.F. Kitchell (eds.) Food Web Manage-
ment- A Case Study of Lake Mendota, Springer-Verlag, New York, pp. 405-
423.
Luecke, C. and O'Brien, W.J. (1983) The effect of Heterocope predation on
zooplankton communities in arctic ponds. Limnology and Oceanography, 28,
367-377.
Luettich, R.A., Jr., Harleman, D.R.F. and Somlyody, L. (1990) Dynamic behavior of
suspended sediment concentrations in a shallow lake perturbed by episodic wind
events. Limnology and Oceanography, 35, 1050-1067.
Macisaac, H.J., Sprules, W.G. and Leach, J.H. (1991) Ingestion of small-bodied
zooplankton by zebra mussels Dreissena polymorpha can cannibalism on larvae
influence population dynamics? Canadian Journal of Fisheries and Aquatic Sci-
ences, 48, 2051-2060.
Macisaac, H.J. {1996) Potential biotic and abiotic impacts of zebra mussels on the
inland waters of North America. American Zoologist, 36, 287-299.
Malone, B.J. and McQueen, D.J. (1983) Horizontal patchiness in zooplankton
populations in two Ontario kettle lakes. Hydrobiologia, 99, 101-124.
Marsden, M.W. (1989) Lake restoration by reducing external phosphorus loading
the influence of sediment phosphorus release. Freshwater Biology, 21, 139-162.
Marshall, B.E. and Junor, F.J.R. (1983) The decline of Salvinia molesta on Lake
Kariba. Hydrobiologia, 83, 477-484.
Martin, T.H., Crowder, LB., Dumas, C.F. and Burkholder, J.M. (1992) Indirect
effects of fish on macrophytes in bays mountain lake evidence for a littoral
trophic cascade. Oecologia, 89, 476-481.
McAllister, C.D., Brasseur, R.J., Parsons, T.R. and Rosenzweig, M.L. (1972) Stabil-
ity of enriched aquatic ecosystems. Science, 175, 562-565.
330 References
McCauley, E., Downing, J.A. and Watson, S. (1989) Sigmoidal relationships be-
tween nutrients and chlorophyll among lakes. Canadian Journal of Fisheries and
Aquatic Sciences, 46, 1171-1175.
McCauley, E. and Murdoch, W.W. (1987) Cyclic and stable populations: Plankton as
paradigm. American Naturalist, 129, 97-121.
McKinnon, S.L. and Mitchell, S.F. (1994) Eutrophication and black swan (Cygnus
atratus Latham) populations: tests of two simple relationships. Hydrobiologia,
279-280, 163-170.
McQueen, D.J., Johannes, M.R.S., Post, J.R., Stewart, T.J. and Lean, D.R.S. (1989)
Bottom-up and top-down impacts on freshwater pelagic community structure.
Ecological Monographs, 59,289-310.
McQueen, D.J. and Post, J.R. (1988) Cascading trophic interactions uncoupling at
the zooplankton-phytoplankton link. Hydrobiologia, 159, 277-296.
Meijer, M.-L., Raat, A.J.P. and Doef, R.W. (1989) Restoration by biomanipulation
of Lake Bleiswijkse Zoom the Netherlands first results. Hydrobiological bulletin,
23,49-58.
Meijer, M.-L., DeHaan, M.W., Breukelaar, A.W. and Buiteveld, H. (1990) Is reduc-
tion of the benthivorous fish an important cause of high transparency following
biomanipulation in shallow lakes? Hydrobiologia, 200-201, 303-316.
Meijer, M.-L., Jeppesen, E., Van Donk, E. and Moss, B. (1994a) Long-term re-
sponses to fish-stock reduction in small shallow lakes- Interpretation of five-year
results of four biomanipulation cases in the Netherlands and Denmark.
Hydrobiologia, 276, 457-466.
Meijer, M.-L., Van Nes, E.H., Lammens, E.H.R.R. and Gulati, R.D. (1994b) The
consequences of a drastic fish stock reduction in the large and shallow Lake
Wolderwijd, The Netherlands - Can we understand what happened?
Hydrobiologia, 276, 31-42.
Meijer, M.-L., Lammens, E.H.R.R., Raat, A.J.P., Klein Breteler, J.G.P. and Grimm,
M.P. (1995) Developments of fish communities in lakes after biomanipulation.
Netherlands Journal of Aquatic Ecology, 29, 91-102.
Melack, J.M. (1980) Temporal variability of phytoplankton in tropical lakes.
Oecologia, 44, 1-7.
Mills, E.L., Forney, J.L. and Wagner, K.J. (1987) Fish predation and its cascading
effect on the Oneida Lake food chain, in W.C. Kerfoot and A. Sih (eds.) Preda-
tion: Direct and Indirect Impacts on Aquatic Communities, University Press of
New England, pp. 118-131.
Mills, E.L. and Forney, J.L. (1981) Energetics food consumption and growth of
young yellow perch Perea flavescens in Oneida Lake, New York. Transactions of
the American Fisheries Society, 110, 479-488.
Mills, E.L. and Forney, J.L. (1983) Impact on Daphnia pulex of predation by young
yellow perch Perea fiavescens in Oneida Lake, New York. Transactions of the
American Fisheries Society, 112, 154-161.
Mitchell, S.F., Hamilton, D.P., MacGibbon, W.S., Nayar, K. and Reynolds, R.N.
(1988) Interrelations between phytoplankton, submerged macrophytes, black
swans (Cygnus atratus) and zooplankton in a shallow New Zealand lake.
Internationale Revue der Gesamten Hydrobiologie, 73, 145-170.
Mitchell, S.F. (1989) Primary production in a shallow eutrophic lake dominated
alternately by phytoplankton and by submerged macrophytes. Aquatic Botany,
33, 101-110.
Fteferences 331
Mittelbach, G.G., Turner, A.M., Hall, D.J., Rettig, J.E. and Osenberg, C.W. (1995)
Perturbation and resilience: A long-term, whole-lake study of predator extinction
and reintroduction. Ecology, 76, 2347-2360.
Mjelde, M. and Faafeng, B.A. (1997) Ceratophyllum demersum hampers
phytoplankton development in some small Norwegian lakes over a wide range of
phosphorus level and geographic latitude, in E. Jeppesen, M. S!1lndergaard, M.
S!1lndergaard and K. Christoffersen (eds.) The structuring role of submerged
macrophytes in lakes, Springer Verlag, New York, in press.
Mooij, W.M., Van Densen, W.L.T. and Lammens, E.H.R.R. (19%) Formation of
year-class strength in the bream population in the shallow eutrophic Lake
Tjeukemeer. Journal of Fish Biology, 48, 30-39.
Moore, B.C., Funk, W.H. and Anderson, E. (1994) Water quality, fishery, and
biologic characteristics in a shallow, eutrophic lake with dense macrophyte
populations. Lake Reservoir Management, 175-188.
Moriarty, D.J.W. (1973) The physiology of digestion of blue-green algae in the
cichlid fish Tilapia nilotica. J Zoo/ Proc Zool Soc Lond, 171, 25-39.
Mortimer, C.H. (1941) The exchange of dissolved substances between mud and
water in lakes. I. Journal of Ecology, 29, 280-329.
Mortimer, C.H. (1942) The exchange of dissolved substances between mud and
water in lakes. II. Journal of Ecology, 30, 147-201.
Moss, B., Balls, K., Irvine, K. and Stansfield, J. (1986) Restoration of two lowland
lakes by isolation from nutrient- rich water sources with and without removal of
sediment. Journal of Applied Ecology, 23, 391-414.
Moss, B. (1988) Ecology of Fresh Waters 2nd Ed. Man and Medium. Blackwell
Scientific, Oxford
Moss, B., Stansfield, J. and Irvine, K. (1990) Problems in the restoration of a
hypertrophic lake by diversion of a nutrient-rich inflow. Verhandlungen
Internationale Vereinigung Theoretisch Angewandte Limnologie, 24, 568-572.
Moss, B., Stansfield, J. and Irvine, K. (1991) Development of daphnid communities
in diatom-dominated and cyanophyte-dominated lakes and their relevance to
lake restoration by biomanipulation. Journal of Applied Ecology, 28, 586-602.
Moss, B. (1994) Brackish and freshwater shallow lakes- Different systems or vari-
ations on the same theme. Hydrobiologia, 276, 1-14.
Moss, B. (1995) The microwaterscape - A four-dimensional view of interactions
among water chemistry, phytoplankton, periphyton, macrophytes, animals and
ourselves. Water Science and Technology, 32, 105-116.
Moss, B., Stansfield, J., Irvine, K., Perrow, M. and Phillips, G. (1996) Progressive
restoration of a shallow lake: a 12-year experiment in isolation, sediment removal
and biomanipulation. Journal of Applied Ecology, 33,71-86.
Moss, B., Madgewick, J. and Phillips, G. (19%a) A Guide to the Restoration of
Nutrient-Enriched Shallow Lakes. Broads Authority/Environment Agency
Mulholland, P.J., Steinman, A.D., Palumbo, A.V., Elwood, J.W. and Kirschtel, D.B.
(1991) Role of nutrient cycling and herbivory in regulating periphyton communi-
ties in laboratory streams. Ecology, 72, 966-982.
Munro, I.G. and Bailey, R.G. (1980) Early composition and biomass of the crusta-
cean zooplankton in Bough Beech Reservoir, Southeast England, UK. Freshwa-
ter Biology, 10,85-%.
Mur, L.R., Schreurs, H. and Visser, P. (1993) How to control undesirable
cyanobacterial dominance, in G. Giussani and C. Callieri (eds.) Proceedings of
332 References
the 5th international conference on the conservation and management of lakes,
Stresa, Italy, pp. 565-569
Murdoch, W.W. and Oaten, A. (1975) Predation and population stability. Advances
in Ecological Research, 9, 1-131
Neary, J., Cash, K. and McCauley, E. (1994) Behavioural aggregation of Daphnia
pulex in response to food gradients. Functional Ecology, 8, 377-383.
Nicholls, K.H. and Hopkins, G.J. (1993) Recent changes in Lake Erie (north shore)
phytoplankton: cumulative impacts of phosphorus loading reductions and the
zebra mussel introduction. Journal of Great Lakes Research, 19, 637-fJ47.
Nilsson, P.A., Briinmark, C. and Pettersson, L.B. (1995) Benefits of a predator-
induced morphology in crucian carp. Oecologia, 104, 291-296.
Nisbet, R., Gurney, W.S.C., Murdoch, W.W. and McCauley, E. (1989) Structured
population models: a tool for linking effects at individual and population level.
Biological Journal of the Linnean Society, 37, 79-99.
Noordhuis, R. (1997) Biologische monitoring zoete rijkswateren, Watersysteem
raportage randmeren. 95003,
Noy-Meir, I. (1975) Stability of grazing systems an application of predator prey
graphs. Journal of Ecology, 63, 459-482.
Ogilvie, S.C. and Mitchell, S.F. (1995) A model of mussel filtration in a shallow New
Zealand lake, with reference to eutrophication control. Archiv fiir Hydrobiologie,
133, 471-482.
Oglesby, R.T., Leach, J.H. and Forney, J. (1987) Potential Stizostedon yield as a
function of chlorophyll concentration with special reference to Lake Erie. Cana-
dian Journal of Fisheries and Aquatic Sciences, 44,166-170.
Oksanen, L., Fretwell, S.D., Arruda, J. and Niemela, P. (1981) Exploitation ecosys-
tems in gradients of primary productivity. American Naturalist, 118, 240-261.
Oreskes, N., Shraderfrechette, K. and Belitz, K. (1994) Verification, validation, and
confirmation of numerical models in the earth sciences. Science, 263, 641-fJ46.
Pace, M.L. (1984) Zooplankton community structure, but not biomass, influences
the phosphorus-chlorophyll a relationship. Canadian Journal of Fisheries and
Aquatic Sciences, 41, 1089-1096.
Padisak, J., Reynolds, C.S. and Sommer, U. (1993) Intermediate Disturbance Hy-
pothesis in Phytoplankton Ecology. Developments in Hydrobiology 81. Kluwer
Academic Publishers, Dordrecht.
Pastorok, J. (1980) Selection of prey by Chaoborus larvae: A review and new
evidence for behavioral flexibility. Am. Soc. Limnol. Oceanogr. Spec. Symp. 3,
538-555.
Paterson, M.J. (1993) The distribution of microcrustacea in the littoral zone of a
freshwater lake. Hydrobiologia, 263, 173-183.
Paterson, M.J. (1994) Invertebrate predation and the seasonal dynamics of
microcrustacea in the littoral zone of a fishless lake. Archiv fiir Hydrobiologie, 1-
36.
Pennak, R.W. (1973) Some evidence for aquatic macrophytes as repellents for a
limnetic species of Daphnia. Internationale Revue der Gesamten Hydrobiologie,
58, 569-576.
Perrow, M.R., Moss, B. and Stansfield, J. (1994) Trophic interactions in a shallow
lake following a reduction in nutrient loading - A long-term study.
Hydrobiologia, 276, 43-52.
Perrow, M.R., Schutten, J., Howes, J.R., Holzer, T. and Jowitt, A.J.D. (1997) Inter-
342, 241–255.
334 References
Prejs, A., Martyniak, A., Borori, S., Hliwa, P. and Koperski, P. (1994) Food web
manipulation in a small, eutrophic Lake Wirbel, Poland - Effect of stocking with
juveniie pike on planktivorous fish. Hydrobiologia, 276, 65-70.
Prentki, R.T., Adams, M.S., Carpenter, S.R., Gasith, A., Smith, C.S. and Weiler, P.
(1979) The role of submersed weedbeds in internal loading and interception
of allochthonous materials in Lake Wingra, Wisconsin USA. Archiv fUr
Hydrobiologie, 57,221-250.
Prieur, L. and Sathyendranath, S. (1981) An optical classification of coastal and
oceanic waters based on the specific spectral absorption curves of phytoplankton
pigments, dissolved organic-matter, and other particulate materials. Limnology
and Oceanography, 26, 671--{i89.
Quade, H.W. (1969) Cladoceran faunas associated with aquatic macrophytes in
' some lakes in northwestern Minnesota. Ecology, 50,170-179.
Quinn, J.F. and Dunham, A.E. (1983) On hypothesis testing in ecology and evolu-
tion. American Naturalist, 122, 602--{i17.
Rabe, F.W. and Gibson, F. (1984) The effect of macrophyte removal on the distribu-
tion of selected invertebrates in a littoral environment. J oumal of Freshwater
Ecology, 2, 359-372.
Ramberg, L., Bjork-Ramberg, S., Kautsky, N. and Machena, C. (1987) Development
and biological status of Lake Kariba, a man-made tropical lake. Ambia, 16, 314-
321.
Reeders, H.H., Bij de Vaate, A. and Noordhuis, R. (1992) Potential of tbe zebra
mussel (Dreissena polymorpha) for water quality management, in T.F. Nalepa
and D.W. Schloesser (eds.) Zebra Mussels- Biology, Impacts, and Control, Lewis
Publishers, London, pp. 439-451. •
Reeders, H.H. and Bij de Vaate, A. (1990) Zebra mussels Dreissena polymorpha a
new perspective for water quality management. Hydrobiologia, 200/201, 437-450.
Reynolds, C.S. (1975) Interrelations of photosynthetic behavior and buoyancy regu-
lation in a natural population of a blue-green alga. Freshwater Biology, 5, 323-
338.
Reynolds, C.S. (1984) The Ecology of Freshwater Phytoplankton. Cambridge Uni-
versity Press, Camebridge
Reynolds, C.S. (1988) Functional morphology and the adaptive strategies of fresh-
water phytoplankton, in C.D. Sandgren (eds.) Growth and Survival Strategies of
Freshwater Phytoplankton, Cambridge University Press, New York, pp. 388-433.
Reynolds, C.S. (1993) Scales of disturbance and their role in plankton ecology.
Hydrobiologia, 249, 157-171.
Reynolds, C.S., Padisak, J. and Sommer, U. (1993) Intermediate disturbance in the
ecology of phytoplankton and the maintenance of species diversity a synthesis.
Hydrobiologia, 249, 183-188.
Rice, E.L. (1974) Allelopathy. Academic press, New York
Rigler, F.H. (1956) A tracer study of the phosphorus cycle in lake water. Ecology, 37,
550-562.
Riley, E.T. and Prepas, E.E. (1984) Role of internal phosphorus loading in two
shallow productive lakes in Alberta, Canada. Canadian Journal of Fisheries and
Aquatic Sciences, 41, 845-855.
Riley, E.T. and Prepas, E.E. (1985) Comparison of the phosphorus-chlorophyll
relationships in mixed and stratified lakes. Canadian Journal of Fisheries and
Aquatic Sciences, 42, 831-835.
Fteferences 335
Page numbers in bold refer to figures. Hudsons Bay 216, 237, 288
Huron 125
Albufera 116
Alderfen Broad !Jsselmeer 1, 24, 27, 28, 29, 31, 207
overview 2,4-5,4
phosphorus in relation to vegetation Kariba 301
66 Krankesjon
Apopka biomass of invertebrates in Chara
overview 2, 5-6 214, 214, 216
perturbation by storm 287, 291 birds 224, 224, 225
resuspension 47, 83, 84 foodweb 224
Arresl<! 42, 47, 59 overview 2, 13-14, 13, 14, 15
Aydat 166 periphyton 255
water level 286
Balaton 42, 57
Bautzen Reservoir 190, 200 Lac Hertel 192
Bleiswijkse Zoom 29, 44, 44, 45, 67, Lagoon of the Islands 2, 14-15
68 Lake, see name of lake
Border-lakes, see Randmeren linford Lakes
Bough beech Reservoir 166 birds 224
Breukeleveen 254 fish stock management 291
Broads, see Norfork Broads overview 2, 9--12
Little Mere 295
Chapala 38 Little Wall Lake 34, 36
Chatauqua 47,48 Loosdrecht 83, 114
Christina see also Lakes, Breukeleveen
birds 224 Luknajno 66
overview 2, 15-18, 16, 17 Lyng 199, 200
Crose Mere 21
Maarseveen 257
Drontermeer 1, 24, 31 Main Lake 12, 223
Markermeer 24, 27, 28, 29, 31,
Ellesmere 207
birds 224 Mendota 164, 165, 184
overview 2, 6 Michigan 200
perturbation by storm 287, 291
Erie 72, 73, 125 Noorddiep 45, 67, 68
Norfolk Broads 4-5, 216
Galgje, see Bleiswijkse Zoom
Great linford 193, 223, 288 Oneida 145, 148, 163, 164
Grosser Binnen see 166
Gullmarsfjorden 58 Paul 200
Pearl 226, 229
Haringvliet 24, 31 Peter 200
Hoveton Great Broad 216, 237, 288 Potomac river 235, 239
Lake index 345
Randmeren 1, 248, 249, 251, 254-257 Vechten 253
see also Veluwemeer Veluwemeer
Reeuwijk 24, 31 birds 224
Rice lake Chara fields 66, 230, 232, 235, 288,
overview 2, 7 294
water level change 286, 287 cyanobacterial dominance 103, 103
Ring 217, 219 flushing effects 113
light climate 24, 27, 28, 29, 31
Schlachtensee 103, 103 nutrient effect 291
S¢bygaard 53, 55 overview 1-4, 2, 2
St. Clair 125 periphyton 252
St. Peters Lake 223 resuspension 42, 49, 238
Stigsholm 218, 219, 220 Vogelenzang 296
Volkerak 24, 27, 28, 31, 242
Takern
birds 224 Washington 190
overview 2, 13-14 Westeinder 59
switch between states 288 Wirbel 199
water level change 286, 291 Wolderwijd
Tamnaren benthivorous fish effect 45
birds 224 light climate 24, 27, 28, 29, 31
effect of inorganic suspended solids role of Neomysis 190
on algae 94
overview 2, 6-7 Zeeltje, see Bleiswijkse Zoom
resuspension 40, 47 Zuiderzee 1
water level change 286, 291 see also !Jsselmeer
Tjeukemeer 164, 167, 168 Zwemlust
Tomahawk Lagoon fetch·vegetation 255
overview 2, 18-19, 18 fish stock management 291
switch between states 288 nutrient dynamics 67, 68, 68
Trummen 304 overview 2, 8-9
Tuakitoto 124 seasonal dynamics 8, 233, 233
Tuesday 200 shading 252
Vreng 217,254,281
Subject index
Winter kill of fish, see Fish kill in Zero isoclines, see Isoclines
winter Zero-plus fish, see Young-of-the-year
fish
Yellow perch (Percaflavescens) 148, Zizania, see Wild rice
163, 164 Zooplankton
Yellow water-lily, see Nuphar lutea affected by fish 146--190, 297
Young-of-the-year fish 148, 164, 172, affected by vegetation 216--222, 239
221-223, 298 grazing algae in vegetation 236--237,
239-241
Zander, see Pikeperch see also Sida crystallina,
Zanichellia peltata 99 Simocephalus vertilus
Zebra mussel, see Dreissena see also Daphnia, Planktivorous fish
polymorpha Zostera marina (eelgrass) 196