Restoration and Recovery of Shallow Eutrophic Lake

Ecosystems in The Netherlands

Developments in Hydrobiology 74

Series editor
H. J . D u m o nt
Restoratio n an d Recover y o f Shallo w
Eutrophi c Lak e Ecosystem s
in Th e Netherland s
Proceedings of a conference held in A m s t e r d a m ,
The Netherlands, 1 8 - 1 9 April 1 9 9 1

Edited by
Louis Van Liere & Ramesh D. Gulati

Reprinted from Hydrobiologia, vol. 233 (1992)

Springer-Science+Business Media, B.V.

 Library of Congress Cataloging-in-Publication Data

R e s t c a t i o n and recovery of shallow eutrophic lake ecosystems in the

Netherlands : proceedingisj of a con erence
r
held in Amsterdam, The
Netherlands, 1 6 - 1 9A p r i l 1991 / edited by L o u i s van Liere & Ramesh
D. Gu1 a t l .
p. cm. — (Developments in hydrotnology ; v. 74)
"Reprinted from hydrobiologia, v o l . 233 (1992)."
I S B N 978-94-010-5073-9 I S B N 978-94-011-2432-4 (eBook)
DOI 10.1007/978-94-011-2432-4
1. Lake eco1ogy--Nether1 anas—Loosarecnt—Congresses. 2. Water
q u a l i t y — N e t h e r 1ands--Loosdrecht—Congresses. 3. Lake renewal¬
-Netherlands—Congresses. 4. Eutrophication--Control—Nether 1ands-
-Congresses. I. Liere, Louis v a n . I I . G u l a t i , R. D.
III. Hydrobiologia. IV. Series: Developments in hydrobiology ; 74.
QH159.R47 1992
333.91 63153'Q9492--dc2C 92-12305

ISBN 978-94-010-5073-9

The image on the front cover is a L A N D SAT Thematic Mapper water quality concentration map
on the basis of the 5 July 1987 band 2 [green] image (left) and the red band 3 image (right). For
further explanation see the article of Dekker et ah on page 137.

Printed on acid-free paper

A ll Rights Reserved
© 1992 Springer-Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 1992
Softcover reprint of the hardcover 1st edition 1992
No part of the material protected by this copyright notice may be reproduced or utilized in any
form or by any means, electronic or mechanical, including photocopying, recording, or by any
information storage and retrieval system, without written permission from the copyright owners.
 Dedicated to Sikko Parma
on the occasion of his retirement
April 1, 1992
 vii

Contents

Preface . IX

List of referees . Xl
Sikko Parma, limnologist, environmentalist and scientific manager
by L. Van Liere, R. D. Gulati, B. Z. Salome, J. Ringelberg, C. Davids & K. Bakker xiii

Part one: Aspects of water quality research in Loosdrecht lakes

Working group Water Quality Research Loosdrecht Lakes: its history, structure, research pro-
gramme, and some results
by L. Van Liere, S. Parma & R. D. Gulati .
The state of the environment of the Loosdrecht lakes
by J. J. Hofstra & L. Van Liere . 11
The hydrology of the Loosdrecht lakes area
by G. B. Engelsen, B. F. M. Kal, J. J. Buyse & F. G. M. Van Pruissen . 21
Phosphorus in the sediment of the Loosdrecht lakes and its implications for lake restoration
perspectives
by P. Keizer & A. J. C. Sinke . 39
Coupling of phytoplankton and detritus in a shallow, eutrophic lake (Lake Loosdrecht, The
Netherlands)
by H. J. Gons, T. Burger-Wiersma, J. H. Otten & M. Rijkeboer . 51
Dynamics of phytoplankton detritus in a shallow, eutrophic lake (Lake Loosdrecht, The Nether-
lands)
by J. H. Otten, H. J. Gons & M. Rijkeboer . 61
The dynamics and role oflimnetic zooplankton in Loosdrecht lakes (The Netherlands)
by R. D. Gulati, A. L. Ooms-Wilms, O. F. R. Van Tongeren, G. Postema & K. Siewertsen
69
P-Ioad, phytoplankton, zooplankton and fish stock in Loosdrecht Lake and Tjeukemeer: con-
founding effects of predation and food availability
by E. H. R. R. Lammens, P. J. Boesewinkel-de Bruyn, H. Hoogveld & E. Van Donk .... 87
Restoration and resilience to recovery of the Lake Loosdrecht ecosystem in relation to its phos-
phorus flow
by L. Van Liere & J. H. Janse . 95
Multivariate analysis of the plankton communities in the Loosdrecht lakes: relationship with the
chemical and physical environment
by O. F. R. Van Tongeren, L. Van Liere, R. D. Gulati, G. Postema & P. J. Boesewinkel-
de Bruyn . 105
A mathematical model of the phosphorus cycle in Lake Loosdrecht and simulation of additional
measures
by J. H. Janse, T. Aldenberg & P. R. G. Kramer . 119
Remote sensing as a tool for assessing water quality in Loosdrecht lakes
by A. G. Dekker, T. J. Malthus, M. M. Wijnen & E. Seyhan . 137
VIII

Water quality research in the Loosdrecht lakes: proposals for the follow-up restoration measures
by M. A. De Ruiter 161
The significance of the Loosdrecht lakes research project for eutrophication policy in the Nether-
lands
by H. W. Kroes 165
Water quality research in Loosdrecht lakes: the salient features
by R. D. Gulati & L. Van Liere 171
Publications of the workgroup WQL (Water Quality Research Loosdrecht Lakes) 1983 - present
by L. Van Liere 179

Part two: Eutrophication research in The Netherlands, with emphasis on additional measures
The influence offlushing on nutrient dynamics, composition and densities of algae and transparency
in Veluwemeer, The Netherlands
by E. Jagtman, D. T. Van der Molen & S. Vermij 187
Lake restoration with and without dredging of phosphorus-enriched upper sediment layers
by J. Van Der Does, P. J. T. Verstraelen, P. C. M. Boers, J. Van Roestel, R. Roijackers &
G. Moser 197
Fixation of phosphorus in lake sediments using iron(III)chloride: experiences, expectations
(extended abstract)
by P. C. M. Boers, J. Van Der Does, M. Quaak, J. C. Van Der Vlugt & P. A. Walker. .. 211
Fisheries management as an additional lake restoration measure: biomanipulation scaling-up
problems
by J. C. Van Der Vlugt, P. A. Walker, J. Van Der Does & A. J. P. Raat 213
The effects of dredging and fish stocking on the trophic status of shallow, peaty ditches
by I. R. M. Hovenkamp-Obbema & W. Fieggen 225
Eutrophication control strategies for three shallow Vecht lakes in the province of North Holland
by P. J. T. Verstraelen, J. Wisserhof, Lj. Rodic & R. Eijsink . . . . . . . . . . . . . . . . . . . . . . . .. 235
Restoration of fen ecosystems in the Vecht River plain: cost-benefit analysis of hydrological
alternatives
by A. Barendregt, S. M. E. Starn & M. J. Wassen 247
Phosphorus eutrophication in the SW Frisian lake district. 1. Monitoring and assessment of a
dynamic mass balance model
by H. J. W. J. Van Huet 259
Phosphorus eutrophication in the SW Frisian lake district. 2. Phosphorus balances and simulation
of reduction scenarios
by H. J. W. J. Van Huet 271

Part Three: Europhication control in The Netherlands

Restoration and recovery of shallow eutropic lake ecosystems in The Netherlands: epilogue
by L. Van Liere & R. D. Gulati 283
Hydrobiologia 233, 1992.
L. Van Liere & R. D. Gulati (eds). Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. IX

Preface

The Symposium on 'Restoration and Recovery of Shallow Eutrophic Ecosystems in The Netherlands'
was held on April 18 and 19, 1991 in Amsterdam at the residence of the Royal Netherlands Academy
of Arts and Sciences. This book presents the results of a ten-year study carried out by the multidisciplinary
working group WQL (Water Quality research Loosdrecht lakes), and discusses these findings with
colleagues who have also been engaged in water management measures to combat eutrophication oflakes
other than the Loosdrecht lakes.
It was the second WQL symposium, the first having taken place in Amsterdam in 1985, when WQL
presented its 'state of the art', and discussed future research with invited specialists, both Dutch and
foreign. In addition there were special sessions devoted to WQL's research at the SIL congres in Lyon
(1983) and Munich (1989). Also at the Conference 'Ecosystem Research in Freshwater Environment
Recovery', organized by the Commission of the European Communities (Pallanza, 1990), WQL
participants contributed richly.
The participants at this final WQL Symposium in Amsterdam were scientists working at institutes or
Dutch ministries, involved in environmental affairs, and from regional water authorities and scientifically-
oriented water quality managers. The Symposium deliberations had a more national character and so
the papers were presented in Dutch. Because the working group WQL was funded largely by the Dutch
ministries and regional authorities (although the contribution of the Commission of the European
Communities was considerable), the Symposium was a good opportunity for the WQL to demonstrate
how judicially it utilized its resources, especially the funds. About 170 people attended the Symposium,
so the target number of the Symposium Organizing Committee was achieved.
The papers presented were divided into two sessions: the first represented the WQL participants (14
oral presentations) and the second (4 review papers and 7 posters presentations) represented the
non-WQL participants. It is almost impossible to summarize all the aspects investigated by WQL
between 1979 and 1990. This would require reviewing 135 scientific publications to date (the list is
presented in this volume). Thus, the authors were requested to present in their contributions the
quantitative phosphorus flux in the compartments studied as a central theme, paying special attention
to supplementary measures to combat eutrophication.
It has resulted in papers that portray the state of the different aquatic environments, especially
pertaining to ecological assessment of their water quality, complex hydrology, and interactions at the
sediment-water interface, as well as between phytoplankton, detritus, zooplankton and fish. An overview
of the food chain in Loosdrecht lakes is given, and light shed on the relationships between the physical-
chemical environment on the one hand, and between phytoplankton and zooplankton on the other, using
multivariate analysis. The vast amount of data collected and processes studied by the WQL, made it
worthwhile to attempt mathematical modelling. The logistics employed and the infrastructure developed
by the WQL facilitated groundtruth verification of the remote sensing studies. Two authors from outside
the WQL were invited to give critical analysis of the WQL research: Dr H. W. Kroes (the Ministry of
Housing, Physical Planning and Environment), who dealt with the significance of WQL for national
eutrophication policy in The Netherlands, and Ir. M. A. De Ruiter (the Province of Utrecht, the Water
Quality Authority in the Loosdrecht lakes area) who emphasized especially what the Province has gained
from the WQL research in decision-making for further activities on the restoration of Loosdrecht lakes.
Moreover, the papers and posters presented by the non-WQL colleagues in the second session had
x

a critical look at eutrophication research, more from a practical point of view. Most of these studies were
directed to additional measures, since they were carried out at institutes dealing with regional water
problems, and at restoration measures such as flushing, dredging, chemomanipulation and biomanipula-
tion, either in isolation or in combination.
The Symposium, which ended with a vigourous but stimulating discussion on the contributions, was
chaired by Dr H. M. De Boois (the Chairwoman of the Water Pollution Control Authority of the Amstel
and Gooiland) with as members of the panel: Ir. M. A. De Ruiter; Dr H. W. Kroes; Dr S. Parma
(Director of the Limnological Institute and Secretary of the WQL Board); Dr L. Van Liere (the WQL's
Research-Coordinator, and one of the undersigned) and Drs P. J. T. Verstraelen (the Water Pollution
Control Authority of the Amstel and Gooiland). The deliberations reaffirmed the complex nature of the
shallow lake ecosystem, especially those driven to a state of hypertrophy by decades of uninterrupted
eutrophication. That the eutrophication should be reduced primarily by decreasing the inputs of external
nutrients, and that secondary measures can hasten the recovery process and reduce the ecosystem
resilience, was a general consensus.
The Symposium Organization Committee consisted of Jeannine Ebert, Cecilia C. C. Janssen-Kroon,
Herman J. Gons and Henk De Haan, all from the Limnological Institute and was chaired by Louis Van
Liere from the National Institute of Public Health and Environmental Protection. The Symposium was
financially supported by the Dutch Ministry of Housing, Physical Planning and Environment, the
Limnological Institute and the National Institute of Public Health and Environmental Protection.
We, as editors, are highly obliged to the referees for their critical reviews and suggestions. Each
manuscript was sent to at least two referees, one of them a native of an English-speaking country for
linguistic improvements; the second was generally a Dutch colleague, acquainted with the geographical
situation and regional problems. They all responded very promptly, sparing us the need to send extra
reminders, or look for alternative referees. The authors too were generally cooperative, with a few
exceptions.
We are very grateful to Cecilia C. C. Janssen-Kroon. Not only was she one of the organizers of the
Symposium, she also skilfully handled the editorial correspondence.
On the front cover of this volume there is an image of the LANDSAT Thematic Mapper. We thank
Arnold Dekker, who prepared the image for the Loosdrecht lakes area; The International Journal of
Remote Sensing is acknowledged for the permission to use the image.
From the inception of the WQL Dr S. Parma (Director of the Limnological Institute) has been one
of the driving forces; as a Secretary of the WQL Board he exhibited great interest in WQL and contributed
significantly to its successful outcome. We dedicate this volume to Sikko Parma on the occasion of his
retirement on April 1, 1992.

January 6, 1992 LOUIS VAN LIERE, Bilthoven

RAMESH D. GULATI, Nieuwersluis
Hydrobiologia 233, 1992.
L. Van Liere & R. D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands.
List of referees
Auer, M. T., Michigan Technol. University,
Dept. of Civil Engineering, Michigan, U.S.A.
Bailey-Watts, A. E., Institute of Terrestrial Ecol-
ogy, Edinburgh Research Station, Scotland,
U.K.
Blom, G., Agricultural University at Wageningen,
Dept. Of Nature Conservation, The Nether-
lands.
Boers, P. C. M., National Institute of Integral
Freshwater Management, The Netherlands.
Buyse, J. J., IWACO Consultants, The Nether-
lands.
Cooke, G. D., Kent State University, Dept. of
Biological Sciences, Ohio, U.S.A.
Curran, P. J., University of Sheffield, Dept. of
Geography, U.K.
De Haan, H., Limnological Institute, The Nether-
lands.
DePinto, J. V., SUNY at Buffalo, New York,
U.S.A.
Donze, M., Technical University at Delft,
Limnological Laboratory, The Netherlands.
Gibson, C. E., Fresh Water Biological Unit.
Dept. of Agriculture, Northern Ireland, U.K.
Gons, H. J., Limnological Institute, The Nether-
lands.
Gulati, R. D., Limnological Institute, The Nether-
lands.
Higler, L. W. G., National Institute of Nature
Conservation, The Netherlands.
Hosper, S. H., National Institute for Integral
Freshwater Management, The Netherlands.
Jewson, D., The University of Ulster, Limnology
Laboratory, Northern Ireland, U.K.
Keizer, P., Limnological Institute, The Nether-
lands.
Klapwijk, S. P., Water Authority of Rijnland, The
Netherlands.
Latour, J., National Institute of Public Health
and Environmental Protection, The Nether-
lands.
Xl
Loogman, J. G., European Parliament, Luxem-
bourg.
Malthus, T. J., Univ. of Nottingham, Dept. of
Physiol. and Environmental Sciences, U.K.
Moss, B., Univ. of Liverpool, Dept. of Environ-
mental and Evolutionary Biology, U.K.
Nawalany, N., University of Warsaw, Poland.
Rhee, G.-Y., School of Public Health, SUNY at
Albany, NY State Dept. of Health, Albany,
U.S.A.
Parma, S., Limnological Institute, The Nether-
lands.
Roijackers, R. M. M., Agricultural University at
Wageningen, Dept. of Nature Conservation,
The Netherlands.
Schot, P. P., University of Utrecht, Dept. of
Environmental Sciences, The Netherlands.
Ter Braak, K., National Institute of nature Con-
servation, The Netherlands.
Van Donk, E., Agricultural University at Wage-
ningen, Dept. of Nature Conservation, The
Netherlands.
Van Huet, H., University of Nijmegen, The
Netherlands.
Van Liere, L., National Institute of Public Health
and Environmental Protection, The Nether-
lands.
Van Straaten, G., Agricultural University at
Wageningen, Dept. of Agricultural Techniques
and Agricultural Physics, The Netherlands.
Van Tongeren, O. F. R., Limnological Institute,
The Netherlands.
Van Vierssen, W., Institute for Hydraulic and
Environmental Engineering, The Netherlands.
Verstraelen, P. J. T., Water Authority of Amstel
and Gooiland, The Netherlands.
Vijverberg, J., Limnological Institute, The Nether-
lands.
Walsby, A. E., University of Bristol, Dept. of
Botany, U.K.
Winfield, I., Institute of Freshwater Ecology, The
Windermere Laboratory, U.K.
Young, T. c., Clarkson University, Dept. of Civil
and Environmental Engineering, New York,
U.S.A.
Hydrobiologia 233: xiii-xviii, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. Xl11
© 1992 Kluwer Academic Publishers.

Sikko Parma: Iimnologist, environmentalist and scientific manager

Louis Van Liere 1,2, Ramesh D. Gulati I, Bert Z. Salome \ Joop Ringelberg 3 , Cees Davids 4 &
Kees Bakker 5
1 Limnological Institute, Rijksstraatweg 6, 3631 AC Nieuwersluis, The Netherlands; 2 National Institute of

Public Health and Environmental Protection, P. O. Box 1, 3720 BA Bitthoven, The Netherlands (address for
correspondence); 3 Dept. of Pure and Applied Ecology, University of Amsterdam, Kruislaan 320,
1098 8M Amsterdam, The Netherlands; 4 Dept. of Aquatic Ecotoxicology, University of Amsterdam,
Kruislaan 320, 1098 8M Amsterdam, The Netherlands; 5 Delta Institute for Hydrobiological Research,
Vierstraat 28.4401 EA Yerseke. The Netherlands

It's work, the most important thing is work,

Work, the most important thing is work.
Lou Reed: 'Songs for Drella'

We dedicate the Proceedings of the final Water Quality research Loosdrecht lakes Symposium 'Restoration and
recovery of shallow eutrophic lake ecosystems in The Netherlands' to Dr. Sikko Parma who retired on April I, 1992 as
director of the Limnological Institute of the Royal Netherlands Academy of Arts and Sciences. Here we attempt to
give a thumbnail sketch of the person of Sikko Parma, including his early career, as well as his accomplishments as a
limnologist, as a director of the Limnological1nstitute, and as an environmentalist: in all Sikko is 'mission-oriented'.Sikko
was an active member of the WQL right from its inception and secretary of the WQL-board.His job as a director of
the Limnological Institute from 1978 to March 31, 1992 was none-too-easy, in view of the financial, scientific and the
reorganisation problems, the government-run research laboratories were faced with. To lead a group of limnologists, one
must be a generalist. Sikko Parma, though specialist by his research training, is a generalist with main interests,
besides, in general limnology, in problems of water pollution and environmental deterioration. Being a dedicated
environmentalist, Sikko Parma was member of various advisory committees of the national government, but also of
action groups to arouse public awareness and warn the national government on environmental problems.

Sikko Parma, student of biology
Sikko Parma was born in November of 1931 at
IJmuiden in the Province of North Holland. He
passed the State High School at the 'Paarden-
markt', Alkmaar in 1950. He was a very enthu-
siastic scout but also member of a Youth Club for
Nature Study, so much that he failed to pass his
third class at school on his first attempt. After
finishing high school, Sikko joined the undergrad-
uate course in biology (then with only 20 other
students) at the University of Amsterdam. He
successfully completed the requirements for a
'candidate in biology' (1950-1955, 'cum laude').
He majored in animal physiology (Prof. Dr.
A. Punt), with a minor in hydrobiology (Dr.
A. G. Vorstman). Besides hydrobiology, his spe-
cial interests were bats, midges and Prof. Punt's
technician Netteke. Netteke and Sikko were mar-
ried in 1956. During his study period Sikko was
an assistant during several field courses and ex-
cursions abroad (Ambleteuse, Pont du Gard,
Bangor, Dubrovnik) and in The Netherlands
(Limburg and the Zoological Station at Den
Helder). He passed his doctoral examination in
1959 'cum laude', and was appointed, without
having the time for a holiday, as a scientist at the
Laboratory of Animal Physiology (Prof. Dr.
A. Punt). During that time Sikko was active in
the scientific 'camps' of the Netherlands Hydro-
biological Society in the Biesbosch, and ecologi-
cally unique tidal area in the Rhine and Meuse
estuaries.
Sikko Parma, Limnologist
In 1961 he was appointed as a zoologist at the
Hydrobiological Institute of the Royal Nether-
xiv
lands Academy of Arts and Sciences at Nieuw-
ersluis. The choice of research object fell on
Chaoborus flavicans, the midge larvae, which were
abundant in the Lake Vechten, the main object of
research at the institute at that time. On Decem-
ber 10, 1971, Sikko Parma successfully defended
his doctorate thesis entitled 'Chaoborus flavicans
(Meigen) (Diptera, Chaoboridae): an autecologi-
cal study' at the State University of Groningen.
From 1971 to 1978 he worked at the Delta
Institute for Hydrobiological Research of the
Royal Netherlands Academy of Arts and Sci-
ences at Yerseke. Later, he became leader of the
working group 'Structure and dynamics of insta-
ble aquatic ecosystems'. Because of his continu-
ing interests in the life cycle of dipteran larvae he
studied the distribution of chironomid larvae in
brackish waters. In addition he got involved in
actions against the complete damming (damn-
Fig. 1. Sikko Parma at one of the Eastern ScheIdt excursions.
ing?) of the Eastern Scheidt estuary, which would
have led to its transformation into a freshwater
lake with loss of estuarine flora and fauna. Due
to the campaign of 'mission-oriented scientists'
like Sikko Parma and other environmentalists the
choice fell on a unique tidal surge barrier-dam
allowing exchange with the North Sea, so that
essential features of the former estuary are pre-
served.
'Sikko is a biologist in his heart ofhearts. and no
water is safe to him. On holidays he invariably car-
ries at least one net and a series of vials with him.
In the course of time he also has become an enthu-
siastic bird watcher, and now he never travels with-
out his binoculars' (Netteke Parma, pers. comm.).
Sikko Parma, a scientific manager
Sikko Parma has the inherent qualities of a leader,
a manager. He is a staunch devotee of fundamen-
tal sciences, more so if it is socially relevant and
morally acceptable. He is a good listener; and
could often bring order to situations in which
working-group leaders and co-workers experi-
enced confusion and chaos. He took time to lis-
ten to staff members and their problems, irrespec-
tive of their age or experience. The door of his
room was 'always open' and people could gener-
ally walk in without a prior appointment.
When Sikko Parma rejoined the Limnological
Institute in 1978 the research organization was
somewhat in disarray, lacking in a clear integral
approach to studies. Soon on his arrival Sikko
streamlined the institute's organization by setting
up different working-groups. He undertook many
actions to set up new projects, which were often
financed by sources outside the institute. This
paid rich dividends so far as attracting young
scientists working for their Ph.D. degrees. The
Limnological Institute has now become a 'Centre
for Limnology', both nationally and internation-
ally. Sikko has the ability not only to inspire the
scientists but also to exploit their talent for the
best interest of science. More than ten years ago
Sikko initiated a monthly bulletin for the insti-
 xv

tute's personnel to provide in a nutshell first hand

information on the different developments and
activities, scientific and others, including memo-
randa of working group meetings, summaries of
scientific lectures, etc.
As a director of the Limnological Institute
Sikko Parma fulfilled multifarious functions out-
side the institute. He is actively involved in socio-
political aspects of environmental problems.
The Limnological Institute had two establish-
ments; the Vijverhof Laboratory at Nieuwersluis
and the Tjeukemeer Laboratory at Oosterzee.
With his office at Nieuwersluis it was not easy for
Sikko to pay equal or optimal attention to both
the laboratories. He, however, succeeded in pre-
venting the two from drifting apart in their re-
search programmes. He developed plans to merge
the two departments, and expand and reorganize
the research facilities at Nieuwersluis. Thanks to
Sikko Parma, the new facilities at Nieuwersluis
have now been realized and personnel from
'Tjeukemeer' institute has shifted since 1 January
1992 to Nieuwersluis. Also, he was involved in
the founding of a new national institute, the Neth-
erlands Institute of Ecology (NIE) in January Fig. 2. Sikko Parma in his office at the Limnological Institute.
1992 (incorporating the three former ecological
institutes of the Royal Netherlands Academy of
Arts and Sciences: the Institute of Terrestrial the institute. During his directorship scientists
Ecology, the Limnological Institute and the Delta from several countries (Soviet Union, Poland,
Institute for Hydrobiological Research). Sikko Czechoslovakia, Hungary, Spain, U.K., New
showed a lot of patience, perseverance and stead- Zealand, U.S.A., Canada, China, India, Sri
fast attitude in defending the interests of limnol- Lanka, Sweden, Japan, Israel, Libya, Germany,
ogists in the world of ecology at large. Belgium and Finland) participated in the Insti-
The Limnological Institute with its new name tute's research programme. Besides, he encour-
the Centre for Limnology, has since January 1, aged contacts with limnologists from the Peoples
1992, become an integral part of the NIE. Republic of China, India and Sri Lanka through
exchange of visits by the scientist from the Insti-
tute to these countries and vice versa. He also is
International contacts a stimulating international congress organizer. In
1990 he put up the facilities of the Limnological
Sikko Parma initiated many international con- Institute at the disposal of the GAP (Group on
tacts. To name just a few examples, in the re- Aquatic Productivity) workshop of the SIL (So-
search project 'Water Quality research Loos- cietas Internationalis Limnologiae). A group of
drecht lakes' (WQL), he encouraged foreign about 20 international scientists performed re-
colleagues to come to Nieuwersluis, providing search on photosynthetic activity and metabolism
them research facilities to participate in the of phytoplankton, especially in relation to the di-
WQL-research programme, and guest facilities at urnal light-dark cycle. In 1990 he chaired the
XVI
organizing committee of the international confer-
ence: 'Biomanipulation: Tool for Water Manage-
ment' (Hydrobiologia Vol. 200/201). Sikko Parma
has been also a national representative at SIL,
and an active one at that.
Sikko played an important role in initiating and
stimulating scientific cooperation in the European
Community, starting with the national-based re-
search programme 'Water Quality research Loos-
drecht lakes; studying and modelling the impact
of water management measures on the internal
nutrient cycle'. This led the Limnological Institute
to playa coordinating role in the EC group 'Eco-
system Research in Freshwater Environment Re-
covery' with the Italian Institute of Hydrobiology,
the Institute of Agricultural Entomology of the
University of Milano, the School of Environmen-
tal Sciences of the University of East Anglia, the
University of Liverpool, the Institute of Inland
Water Management and Waste Water Treatment,
and the Limnological Institute. Sikko was also
involved in setting up scientific cooperation with
the Peoples Republic of China, and in formally
organizing cooperation between the Royal Neth-
erlands Academy of Arts and Sciences and with
its counter parts in Poland and the former Soviet
Union. Also for the other main projects of the
Limnological Institute (Lake Vechten, Tjeuke-
meer, etc.) he initiated and stimulated interna-
tional cooperation.
All and sundries
The foregoing memoir on Sikko, rather than his
biography, does not have the contention of com-
pleteness, nor of perfection. It contains some
reminiscences of a part of Sikko's life devoted to
science and professional pursuits.
From the 'coffee break gossip' at the Institute
it was clear that Sikko has a great sense of hu-
mour, and capacity to mimic and imitate, hum
and sing and crack jokes. He has a tremendous
memory especially in remembering names and
keeping an updated diary of addresses of profes-
sional colleagues and friends. He is apparently a
fanatic television fan with special interest in the
current affairs, satirical programmes and football.
He has a large collection of Dutch stamps,
which he collects and catalogues with the same
enthusiasm as the snaps taken at the SIL con-
gresses and other scientific meetings.
He reads a great deal, his interest varying from
detectives and thrillers to literature and poetry.
He recites poems of Adriaan Roland Holst, but
also those of John O'Mili during coffee-breaks.
Despite all these very social and sociable char-
acteristics, Sikko is modest, compromising if the
occasion demands, and resisting to be involved
into controversies.
Before returning to the Limnological Institute
as a director he played badminton and volleyball.
The latter was quite surprising since he never
joined the lunchtime volleyball-game for which
the Limnological Institute was locally well known.
While lunching he often strolled in the garden. He
took great interest if it came to preserving the
flora in the garden of the old mansion 'Vijverhof,
the institute's main building.
Membership of committees and advisory boards
Sikko Parma is an active member of the councils
engaged in scientific and socio-political aspects of
environmental issues, and intends to continue in
some of these after his retirement. Because of his
broad knowledge, both scientific as well as so-
ciopolitical, he was invited to join many such
groups. He considered some of these as an im-
portant platform for interaction with the policy
makers and planners. His subtle humour served
as a great help when scientific arguments did not
quite succeed, and there was a stalemate. Thus,
he was also invited many a time to chair discus-
sion groups at conferences and symposia. A se-
lection of commissions and committees in which
Sikko took active part are listed alphabetically.
• Advisory Council for the environment (section
on Water)
• Biological Research in The Netherlands
(BION) (Member of the Board and Chairman
subgroup Aquatic Ecology)

• Biological Council
• Central Committee of Fisheries Research
(NRLO)
• Commission for Ecological Standards in
Water Management (Health Council: Chair-
man)
• Commission Biological Problems Cooling
Water (Chairman)
• Commission on Hydrological Research
(CHOjTNO)
• Committee of Nature- and Landscape Con-
servation
• Committee on the Inventarization of the River
Vecht and the Eastern and Western Areas of
the River Vecht
• Coordination Committee of Nature-, Land-
scape- and Environmental Conservation in
Zealand (Chairman)
• Coordination Committee for the Execution of
the Surface water Pollution Law (CUWVO)
• Advisory council for Nature- and Environmen-
tal Research (RMNO)
• Foundation for Nature and Environment
(chairman)
• Health Council
• Hydrobiological Society
• International Water Tribunal, Rotterdam, The
Netherlands
• Naples-Roscoff Committee (KNAW)
• Netherlands Zoological Society (Secretary)
• Planning Committee for Soil Research
• Royal Zealand Academy of Sciences (Secre-
tary)
• Society of the Conservation of Nature Mon-
uments in the Netherlands
• Study Group Environmental Management
(Ecological Society)
• Water Quality research Loosdrecht lakes
(Secretary)
Publications
As an active scientist, Sikko Parma, besides pub-
lishing fundamental studies on Chaoborus, also
published several papers relating to environmen-
tal issues, which revealed in an early stage his
xvii
'mission orientation'. As a director of the Lim-
nological Institute he was time-limited both to
carry out active research or to publish his earlier
research studies. He gave his consent to be a
co-author of a joint paper only if he had actively
participated in the preparation of the manuscript,
and not per se because he was director. He greatly
stimulated his staff and students to publish, and
kept an updated inventory of the papers 'in press'.
Scientific papers of Sikko Parma, and those in
which he was involved as a co-author are listed
in the references below in a chronological order.
Since we did not check with Sikko some omis-
sions and errors are not ruled out.
Acknowledgements
The most important thing is work, but it is not
only work. However, the citation at the beginning
refers to the creativity of Andy Warhol, and his
ability to inspire the artists at The Factory to
reach for a higher level. And that needs work.
Sikko Parma can work, and he likes it. Sikko
gave his consent to chair a national symposium
on water management on April 8, i.e. just one
week after his retirement.
We acknowledge on behalf of very many col-
leagues all the work that Sikko Parma did: for
science, for limnology, for the Limnological In-
stitute, and for the environment and its 'Sustain-
able Future'.
Sikko, we wish you a very happy retired but
active life, which undoubtedly will be devoted
partly to science and the environment. As a good
start we offer you this book.
References
Punt, A. & S. Parma, 1964. On the hibernation of bats in a
marl cave. PubI. Natuurhist. Genootsch. Limburg. Reeks
13: 45-59.
Parma, S., 1967. Oorzaken van de verspreiding van
Chaoborus-larven in de bodem van meren. Meded. Hydro-
bioI. Ver. I: 28-30.
Parma, S., 1967. Het onderscheiden van watertypen. VakbI.
BioI. 47: 141-150.
XVlll
Parma, S. & K. F. Vaas, 1968. Verslag van het hydrobiolo-
gisch onderzoek in de Biesbosch, voorwoord. Meded. Hy-
drobiol. Ver. 2: 87-91
Parma, S., 1968. Hydrografie van de Biesbosch. Meded. Hy-
drobiol. Ver. 2: 92-145.
Parma, S., 1969. Notes on the larval taxonomy, ecology, and
distribution of the Dutch Chaoborus species (Diptera, Cha-
oboridae). Beaufortia 17: 21-50.
Parma, S., 1969. The life cycle of Chaoborus crystallinus (De
Geer) in a Dutch pond. Verh. int. Ver. Limnol. 17: 888-
894.
Roth, J. C. & S. Parma, 1970. A Chaoborus bibliography. Bull.
Entomol. Soc. Am. 16: 100-110.
Parma, S., 1971. The morphology of the larval instars of
Chaoborusflavicans (Meigen, 1818) (Diptera, Chaoboridae).
Beaufortia 18: 173-182.
Parma, S., 1971. Chaoborusflavicans (Meigen) (Diptera, Cha-
oboridae): an autecological study. Thesis, University of
Groningen, 128 pp.
Parma, S., 1972. Muggenplagen op en om het Zeeuwse Meer.
Vakbl. Biologen 52: 134-140.
Parma, S. & A. J. Cave, 1972. De consequenties van de men-
ing van den Hartog en de Jonge (VB 11, 1972): de afsluiting
van de Oosterschelde. Vakbl. BioI. 52: 344-346.
SwUste, H. F. J., R. Cremer & S. Parma, 1973. Selective pre-
dation by larvae of Chaoborus flavicans (Diptera, Cha-
oboridae). Verh. int. Ver. Limnol. 18: 1559-1563.
Parma, S., 1973. De afsluiting van de Oosterschelde opnieuw
in discussie. Natuur en Landschap 27: 29-48.
Parma, S., 1974. De Oosterschelde, open of dicht. Maritiem
Journaal 1974: 189-194.
Parma, S., 1975. De Oosterschelde, open of dicht. Maritiem
Journaal 1975: 153-159.
Parma, S. & B. P. M. Krebs, 1977. The distribution of chi-
ronomid larvae in relation to chloride concentration in a
brackish water region of The Netherlands. Hydrobiologia
52: 117-126.
Boissevain, E. C. & S. Parma, 1977. Hoe kan de Ooster-
schelde gered worden? Nat. Milieu 1: 12-15.
Parma, S., 1978. Political aspects of the closure of the East-
ern Scheidt estuary. Hydrobiol. Bull. 12: 163-175.
Parma, S., 1978. Kreken en wielen in Zeeland. Zeeuws Nieuws
over Natuur, Landschap en Milieu 3: 5-10.
Parma, S., 1980. The history of the eutrophication concept
and the eutrophication in The Netherlands. Hydrobiol Bull.
14: 5-11.
Parma, S., 1982. The twenty-fifth anniversary of the Limno-
logical Institute, The Netherlands (1957-1982). Hydrobio-
logia 95: 1-9.
Gulati, R. D. & S. Parma (eds), 1982. Studies on Lake
Vechten and Tjeukemeer, The Netherlands. Hydrobiologia
95: 1-383. (Also published as: Developments in Hydrobi-
ology 11: 383 pp. 25 th Anniversary of the Limnological In-
stitute of the Royal Netherlands Academy of Arts and Sci-
ences).
Parma, S., H. M. Van Embden & J. Castelein (eds), 1983.
Oecologie van meren en plassen. Biologische Raad Reeks.
PUDOC, Wageningen. pp. 157.
Parma, S., 1983. Aquatisch-oecosyteemonderzoek in weten-
schappelijk en maatschappelijk kader. In: S. Parma, H. M.
Van Embden & J. Castelein (eds) Oecologie van meren en
plassen. Biologische Raad Reeks pp. 7-18. PUDOC, Wa-
geningen.
Parma, S., 1983. Wensen voor Natuur- en Milieuonderzoek.
Nat. Milieu 7: 12-15
Van Liere, L., S. Parma, L. R. Mur, P. Leentvaar & G. B.
Engelen, 1984. Loosdrecht lakes restoration project: an in-
troduction. Verh. int. Ver. Limnol. 22: 829-834.
Weekenstroo, J. E. & S. Parma, 1984. Aspects of the role of
Chaoborus flavicans (Meigen) in Lake Vechten (The Neth-
erlands). Verh. int. Ver. Lirnnol. 22: 904-908.
Parma, S., 1987. Ontwikkelingsmogelijkheden voor aquatis-
che ecosystemen. In: Produktiebeperkingen in de land-
bouw; nieuwe kansen voor natuur en landschap! Verslag
van het WLO-Iustrumcongres, Wageningen. Werkgemeen-
schap Landschapsoecologisch Onderzoek. pp. 29-43.
Parma, S., 1988. Management of water quality in surface
waters of The Netherlands. Verh. int. Ver. Limnol. 23:
1465-1470.
Parma, S., 1989. Strategie van onderzoek voor de milieube-
weging. In: J. G. Waardenburg, W. Dijkrnan & N. G. Van
Vennedaal-Koevoets (eds), Samenleving en onderzoek.
Verslag van een studiedag op 31 October 1986, Zoetermeer,
Comrnissie van Overleg Sectorraden, p. 37-43.
Parma, S. & A. J. Murk, 1990. Inleiding: het advies van de
gezondheidsraad over Ecologische Normen Waterbeheer.
pp.9-17. In: A. J. Murk, A. A. A. Van der Schaaf,
H. A. M. De Kruijf et al. (eds), Strategieen voor ecologis-
che normstelling waterbeheer: het spel en de knikkers.
SDU, 's Gravenhage.
Van Liere, L., S. Parma & R. D. Gulati, 1992. Working group
Water Quality Research Loosdrecht Lakes: its history,
structure, research programme, and some results. Hydro-
biologia 233: 1-9.
Wolff, W. J., S. Parma, 1974. Oosterscheide pleidooi voor
veilig voorbestaan Stichting Zeeuws Coordinatieorgaan
voor Natuurlandschap en milieubescherming Middelburg,
95 p.
Editor of Progress Reports of the Limnological Institute since
1978.
Hydrobiologia 233: 1-9, 1992.
L. Van Liere & R.D. Gulati (eds). Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands.
© 1992 Kluwer Academic Publishers.

Working group Water Quality Research Loosdrecht Lakes: its history,

structure, research programme, and some results

Louis Van Liere,I,2* Sikko Parma 1 & Ramesh D. Gulati 1

I Limnological Institute, Rijksstraatweg 6,3631 AC Nieuwersluis, The Netherlands; 2 National Institute of

Public Health and Environmental Protection, P. O. Box 1, 3720 BA Bilthoven, The Netherlands; (* author for
correspondence)

Key words: interdisciplinary research, lake restoration

Abstract

In 1984 the external phosphorus load to the Loosdrecht lakes ecosystem was decreased substantially.
A working group, Water Quality research Loosdrecht lakes (WQL), was formed to study the conse-
quences for recovery of the lakes, and to evaluate the water management measures taken. Its history
from the start in 1979 is described. The working group had an interdisciplinary character; its organi-
zation structure (project management) is depicted. From March 1983 a coordinate multidisciplinary
research programme was started, carried out uninterrupted up to December 1990. The programme paid
special attention to the flow of phosphorus through the lake ecosystems. The results were presented at
several international conferences. The WQL was also involved in the international working group
ERiFER (Ecosystem Research in Freshwater Environment Recovery). Important financial support was
given by Dutch ministries, the Commission of European Communities, regional water authorities, and
private funds.
The measures taken to counteract eutrophication were not successful. It has been the task of the WQL
to explain the ecosystem resilience.

Introduction of the water quality in Loosdrecht lakes. The

Loosdrecht lakes have been subjected to severe
eutrophication since the beginning of this century
(Gulati et al., 1991; Hofstra & Van Liere, 1992).
The water quality lakes had already deteriorated
in the early sixties. One of the lakes of the area
served as source of raw material for the drinking
water production of the city of Amsterdam.
Around 1970 the demand for drinking water of
the city had increased and increasingly more
water from the lake area had to be used. This
meant a higher demand for extraneous (phospho-
rus-rich water) to maintain the water level. This
led to fears about rapid rise in the deterioration
Stichtse Milieufederatie (The Sticht Environmen-
tal Federation) warned about the consequences,
upon which the Province of Utrecht formed a
working group to assess the water-quality prob-
lems of Loosdrecht lakes. This working group
consisted of representatives of the Province of
Utrecht, Amsterdam Municipal Waterworks, the
Research Institute of Nature Conservation, the
National Institute of Drinking Water Supply and
the Limnological Institute. Soon thereafter the
group reported its assessment about the state of
the environment (Province of Utrecht, 1973). A
second working group was instituted to explore
the various alternatives for external water supply
2
in order to safeguard the water quality of the lakes
in the future (Province of Utrecht, 1980). The
model that was chosen envisaged on its applica-
tion on a drastic reduction in external phospho-
rus loading. Some members ofthe working groups
opted to carry out detailed ecological studies of
the lakes, and founded the working group WQL
(Water Quality research Loosdrecht lakes). The
history, structure, research programme of the
working group, and some results are subject of
this paper.
WQL, founding and first years (1979-1983)
The first meetings to found WQL were called by
Ing. Wolter Kats and Dr. Ir. Anton Graveland
(Amsterdam Municipal Waterworks) on March
7, 1979 followed by a second meeting on May 3,
1979. No notes were made. The first 'official'
meeting of the working group was held on June
15, 1979. The goals of the working group were 'to
qualify and quantify the effects of a considerable
preduction on the functional and structural as-
pects of a shallow lake ecosystem, to translate
these effects in terms of water quality and to con-
tribute, if necessary, to water management mea-
sures with a view to attaining and maintaining the
desired water quality'. The objective was to per-
form a baseline study prior to measures relating
the phosphorus reduction came into effect, and to
perform a follow-up study thereafter to evaluate
the effects of the measures taken.
Already in July 1979 a fourweekly monitoring
programme was started by the Amsterdam Mu-
nicipal Wat~rworks. About twenty physical and
chemical param~ters were monitored at ten sam-
pling stations spread over the whole Loosdrecht
lakes system. This programme was carried out
uninterrupted up to December 1990. From 1981
phytoplankton was counted routinely, in samples
from a selected number of stations. Zooplankton
research was started in summer 1981. In 1982
preliminary projects started to the possibilities of
hydrological research and exchange at the sedi-
ment water interface.
The structure of the working group to start with
was rather loose, which is not surprising, since
the group owed its foundation to several personal
initiatives, lacking finances to have a steering
function. A board was formed which convened
two to three times a year meetings of the scien-
tists involved to discuss the results and future
plans of work. The participants of the working
group came from various disciplines. University
departments and institutes belonging to the Royal
Netherlands Academy of Arts and Sciences,
which were assigned to perform fundamental re-
search. Laboratories of water authorities and
water boards, and consultant firms which are
confined to applied aspects, i.e. to problem solv-
ing research.
Since the scientist is generally 'curiosity-
oriented', the funds to perform such research be-
comes a limiting factor in terms of economics.
The Dutch policy makers demand rightly, that
research institutes pay attention to social relevant
problems. This does not necessarily mean that
the level of scientific research is lowered. On the
contrary, such 'mission-oriented' research may be
equally valuable as 'curiosity-oriented' research.
In the Dutch scientific jargon it is often called
'strategic fundamental research'. The primary
motivation, at least for the Limnological Institute,
was to perform fundamental research, to under-
stand the structure and functioning of the lake
ecosystems.
Such studies are by definition multidisciplinary
in character, with their roots in hydrology, phys-
ics, chemistry, biology and mathematics. How-
ever, from experience we learned that it is rather
difficult for scientists to cooperate, especially if
the involved disciplines are remote. The integral
level of ecosystem research is such that it is dif-
ficult to put forth a joint question, and this is even
true for the practical research.
Multidisciplinary ecosystem research has an
excellent chance if the objective has the carrying
capacity of the environmental concern. The eu-
trophication problem was the basis of executing
fundamental research in WQL, although some
participants were inclined to carry out purely
problem-solving research. However, the com-
puted drastic reduction in external phosphorus
 3

load to the Loosdrecht lakes was an experiment

conducted at the ecosystem level. It has been a
chance to study effects, but also to get a better
understanding of the causal processes. And, if
water management measures were ineffective
WQL would not only explain this, but suggest
additional measures.
A main advantage of 'mission oriented re-
Subgroup
search' is that it brings together scientists, policy Mathematical
Remote
Sensing
Mode1l1ng
makers and water quality managers. It would not
be honest to say that also financial motives did
not playa role. Eutrophication research, which is
socially relevant, attracts financial support.
A few years after the start of the studies it was '"
-g.~
~i i
felt that separate projects, as they were carried
i" ;~ "E
out, would not meet the objectives. Ecosystem ~j "~
~z~ '"
research is only possible if logistics and scientific a: ::l
::l!

demands are coordinated, let alone of integration

of the various disciplines. The Ministry of Edu-
-"
"
~

-<l
0",
$..,
c
u
!
~
z
c
:> c 2
~
cation and Science funded the appointment of a
research-coordinator (first author of this paper)
~:
=<
ct:

'"-
u 0
i
£
~

'"
iii
C
0.>- ~ o

1
.2 E !!
to WQL from 1983 to 1987, in order to stream- c..,
o "
3:'"
line the functioning of WQL. The Ministry of E '"
:J:l "
a: l1.

Housing, Physical Planning and Environment en-

abled the coordinator's work to be extended. Fig. I. Organization scheme of Water Quality research Loos-
drecht lakes.

WQL, organization structure

For optimizing exchange of knowledge and set-
ting up new programmes, an organization struc-
ture was developed to enable the participating
scientists to meet and discuss results. Represen-
tatives of the eight participating institutes were
arranged in working groups according to four
themes (Fig. 1). In these group study plans,
progress of research and evaluations were dis-
cussed. Several members participated in more
than one group in order to optimize exchange of
knowledge. All WQL-participants were kept in-
formed of the progress of the four groups through
notes and internal reports. The research-coordi-
nator chaired the working group meetings, each
group meeting four to five times per year. Only the
Remote Sensing group met ad hoc. The research-
coordinator was added to the Board, which con-
sisted of the directors/working-group leaders of
the main WQL institutes. Twice a year all par-
ticipants of WQL, as well as policy makers and
water managers, met in a plenary session, in which
papers of various WQL-subjects were presented
and discussed. WQL consisted of about 60 mem-
bers. Depending on the number of projects in
operation about 30 scientists were actually work-
ing in the field or in the laboratories, or both.
WQL, integrated research (1983-1990)
In 1983 the coordinated research programme was
set up. It consisted of fourweekly measurements
of physical, chemical and biological parameters
and processes. Because of the available man-
power and apparatus it was possible to run the
routine programme only at four sampling stations,
two in the Lake Loosdrecht, the other two were
4

situated in the Lake Breukelveen and the Lake
Vuntus. Apart from that, the number of projects
in various disciplines was extended in the course
of time. In Fig. 2 the hierarchical levels of the
research are depicted in a hierarchial scheme.
The integrating theme of the research pro-
gramme was to quantify the flow of phosphorus
through the ecosystem, with special attention to
the food chain.
The basis of lake ecosystem research is a reli-
able water- and matter balance. We chose for a
phosphorus balance because phosphorus has
been indicated as the factor limiting algal growth
in the past (Van Heusden, 1942), and because
phosphorus was likely to increasingly become
limiting for algal growth in the future due to the
planned reduction in external load of phospho-
rus. For practical reasons it was not possible to
include nitrogen. Also we knew that, except for
some short periods in late summer, the concen-
tration of bound nitrogen was so high that it was
not expected to limit algal growth. The physical
and chemical monitoring programmes gave sound
support to the various projects, although they are
not indicated in Fig. 2.
Not all projects started in 1983, mostly because
of financial restrictions. Mesocosm research
could start only in 1985 through a financial injec-
tion of the European Community to build and
develop them. After developing and testing the
system the Dutch Ministry of Housing, Physical
Planning and Environmental Protection funded
further research. The investigation of the physi-
ology of Prochlorothrix hollandica could only start
after its discovery and isolation (Burger-Wiersma
et aI., 1986). Water plants were studied mainly in
1983 (Best et al., 1984) and 1985 (Malthus et al.,
1990); since submerged water plants were scarce;
and because water plants in general were negli-
gible on the phosphorus balance it was decided
not to prolong the study. Studies on rotifer feed-
ing awaited the coming of Dr. Jolanta Ejsmont-

CATCHMENT
AREA

LOOSDRECHT
LAKES

Lake Lake Lake

Vuntus Loosdrecht Breukeleveen
--- --

. ..
i3 ...;:
'" ; .-i" m
~

~ ii
~
~
2. ~
•~
j. il
1
Fig. 2. Hierarchical scheme of WQL's research programme from 1985-1990.
 5

Karabin from Poland (Ejsmont-Karabin et a!.,
1989; Gulati eta!., 1992), and fishery research
also started later (Van Densen et at., 1986). It
took in fact three years to optimize the pro-
gramme.
Performing experiments with complete ecosys-
tems in The Netherlands is for various reasons
impossible, if not absurd. Therefore, individual
scientists have to incline to lower hierarchical
level, instead of being subject to an integrated
cooperative ecosystem approach. From the var-
ious integration levels WQL had to extrapolate to
the ecosystem level.
The results and spin-off ofWQL are presented
in Gulati & Van Liere (1992). One main conclu-
sion is that the measures to reduce the external
phosphorus load to Loosdrecht lakes have not
yet resulted in the clear macrophyte-dominated
system, like the one that existed before eutroph-
ication. From the concentrations of total-phos-
phorus and chlorophyll a (Figs. 3 and 4). Van
Liere et at. (1990) calculated, after correction for
autocorrelation, that the concentration of total-
phosphorus rose from 1979 to 1983 at a rate of
9 ± 1.5,ug 1- 1 y- \ and decreased with 12
± 2.0,ug 1- 1 y- I from 1984 to 1988. Chlorophyll
a rose very rapidly in the years before the resto-
ration measures: 24 ± 3 ,ug 1- I Y- I, but did not
react at all after 1984 (0.1 ± 0.25 ,ug 1- 1 Y- I). It
has been the task of WQL to explain this resil-
ience.
WQL, highlights
WQL presented its research studies at an inter-
national symposium: 'Restoration of shallow eu-
trophic ecosystems, with special emphasis on

LOOSDRECHT

250....---------------------------------.,

200
~I

.........
Ol
::J

(f) 150
::J
L
0
L-
a.
(f)
0
L- 100
a.
r-l
co
->J
0
I-
50

79 80 81 82 83 84 85 86 87 88 89 90
Fig. 3. Concentration of total-phosphorus in Lake Loosdrecht from 1979-1990. Data before 1984 measured by Amsterdam
Municipal Waterworks, after 1984 the measurement were done at the Limnological Institute, after intercalibration in 1983 and
1984.
6

LOOSDRECHT

350

300

...... 250
"-
OJ
:J

200
co
......
......
>-
£ 150
Q
0
C-
o
......
£ 100
U

50

0
79 80 81 82 83 84 85 86 87 88 89 90
Fig. 4. Concentration of chlorophyll a in Lake Loosdrecht from 1979-1990. Data before 1984 measured by Amsterdam Municipal
Waterworks, after 1984 the measurement were done at the Limnologica1 Institute, after intercalibration in 1983.

Loosdrecht lakes', Amsterdam, September 1985. (School of Environmental Sciences, University of

The proceedings were published as a special vol-
ume of the Hydrobiological Bulletin (Loogman &
Van Liere, 1986). Also at the xxn th (Lyon, 1983)
and XXIV th (Munich, 1989) SIL Symposia spe-
cial sessions were devoted to WQL's research. At
the conference 'Ecosystem Research in Freshwa-
ter Environment Recovery' (Pallanza, 1990) or-
ganized by the Commission of the European
Communities, WQL had an important contribu-
tion.
In 1984 WQL entered the international money
funding, scientific scene by a grant from the Com-
mission of the European Communities, which in
1987 led to the founding of ERiFER (Ecosystem
Research in Freshwater Environment Recovery),
a research group consisting of two Italian insti-
tutes (Italian Institute of Hydrobiology, Pallanza)
and the (Institute of Agricultural Entomology,
Milan), one group from the United Kingdom
East-Anglia) and two Dutch institutes (Institute
for Inland Water management and Waste Water
Treatment) and the coordinating Limnological
Institute). The common framework was intended
to increase the understanding of aquatic ecosys-
tem dynamics i.e. the physical-chemical environ-
ment and its interactions with the biological com-
ponents of the system. In practice we have studied
the reversibility of eutrophication by biomanipu-
lation; here defined as scientifically oriented water
management of the internal structure and func-
tioning of the system ('active biological manage-
ment'). Field studies, laboratory experiments, me-
socosm studies, and a whole lake experiment were
included. ERiFER held its third International
Workshop at Pallanza, Italy in September 1990
(Giussani et al., 1991). The proceedings of the
first two workshops were reported only to the
ERiFER and the Commission of the European

Communities. An extension of the group both in
institutes and research subjects is in preparation.
Financial support
When all projects ofthe research programme were
in operation, as depicted in Fig. 2, in 1985 the
annual budget was 3.5 106 Dutch guilders (ca. 1.5
106 ECU). About half of it came from the regu-
lar annual budgets of the participating institutes
themselves, the other half coming from projects
submitted to various institutions, and from inter-
national, national, regional as well as private
funds. The main ones in order ofimportance were:
the Netherlands Ministry of Housing, Physical
Planning and Environment, the Commission of
the European Communities, the Dutch Ministry
of Education and Science and the Province of
Utrecht. The other funding institutions in alpha-
betical order are: the Amsterdam Municipal Wa-
terworks, the Batavian Commission of Experi-
mentally Supported Philosophy, the Beijerinck-
Popping Fund, the Foundation for Technical
Research, the National Institute of Nature Con-
servation, the National Institute of Public Health
and Environmental Protection, the Netherlands
Integrated Soil Research Programme, the Neth-
erlands Foundation for Biological Research, the
Netherlands Remote Sensing Board, the Royal
Netherlands Academy of Arts and Sciences, the
Netherlands Organization for the Advancement
of Pure Research.
Acknowledgements
A coordinators personal point of view
It needs both enthusiastic, and motivated women
and men to do ecosystem research. Thus, many
are to be acknowledged, but it is virtually impos-
sible to name, and thank all of them. Most of
them had just finished their degree and entered
the world of science, with short term contracts,
varying from one year to four years. Their zeal,
enthusiasm and ideas were extremely stimulating,
7
and have guided WQL through periods when top-
down control nearly failed. The largest WQL
group was at the Limnological Institute, an im-
portant reason why I chose may office to be there,
because research coordination has its optimum
during coffee and luncheon break. I do not feel
the need to apologize for the probably nasty re-
marks about the delay in writing scientific publi-
cations and reports to be submitted to the spon-
sors. A look at the publication list of the working
group (Van Liere, 1992) reveals that it has been
worth it.
Very important in ecosystem research is the
role of sampling crew. Almost throughout the field
study Heb Roon, the skipper of the sampling ves-
sel the Limnos was there, and demonstrated many
scientists how to sample and helped with great
enthusiasm. In the first years I was glad to be able
to join him. When the management task became
more time-consuming, I really missed those early
morning sampling tours. My sons Bram and Joost
always looked forward to their annual summer-
holiday when they could join Heb and me on
sampling tour on the Loosdrecht lakes.
Writing projects-proposals and submitting
them to raise funds for research, and keeping
watch on the financial part needs an excellent
administration. In this respect Jan Landstra and
Sietze Landmeter were indispensible. Cecilia
Janssen-Kroon greatly improved the layout of
many manuscripts and proposals. During the
course of the projects they were a tremendous
support. When I had to organize workshops or
symposia Cecilia was my co-organizer. In the in-
stitutions to which we submitted the above men-
tioned proposals I would like to acknowledge
some representatives: Henk W. Kroes and Hans
R. Van Otterloo (Ministry of Housing, Physical
Planning and Environment), Hartmut Barth
(Commission of the European Communities) and
Marien A. de Ruiter (Province of Utrecht) for
their keen interest and help.
In the working group 'Primary and Secondary
Production' with Ramesh D. Gulati as working-
group leader many new ideas and projects were
initiated. I have cooperated in various projects,
including WQL, and even in different institutes
8
since 1983 with Herman J. Gons. We have ex-
perienced lots of difficulties, but also lots of fun.
The workshop of the Limnological Institute pro-
vided many research tools, and even complete
mesocosms.
My assisting coordinators (Leo van Ballegoo-
ijen, Loes Breebaart and Jeannine Ebert), with
whom I shared the extremely small office at the
Limnological Institute, deserve a special acknowl-
edgement. Their presence and support smoothed
the cooperation and integration of WQL's re-
search.
Concerning WQL so far I have restricted ac-
knowledging colleagues at the Limnological In-
stitute; this is not surprising since more than
three-fourths of the scientists engaged in WQL
are stationed there. However, there were many
good contacts with other institutes, especially with
the hydrologists at the Free University at Am-
sterdam. J anJ aap Buyse always promptly pro-
vided me with the data I needed for my lecture
presentations and publications, and we spent to-
gether some suspenseful time in controlling the
leaking sluices. The work on groundtruth of the
remote sensing programme with Arnold G. De-
kker was an exciting experience. We succeeded in
accomplishing what experts considered was im-
possible. (And a research-coordinator needs to
measure chlorophyll a personally every now and
then).
The international contacts, especially in ERi-
FER (Ecosystem Research in Freshwater Envi-
ronment Recovery) were very stimulating to
WQL. I acknowledge the relaxed cooperation
with Gianluigi Giussani (Istituto Italiano di Idro-
biologia), Brian Moss (Dept. of Environmental
and Evolutionary Biology, University of Liver-
pool), Marco Vighi (Istituto di Entomologia
Agraria) and Marie-Louise Meijer (Rijksinstituut
voor Integraal Zoetwaterbeheer and Afvalwater-
behandeling). I am not coordinating the ERiFER
group any more but I would strongly advise the
Commission of the European Communities to ex-
tend the programme of this group in the future.
Finally, I would like to thank Sikko Parma,
who offered me the coordinator's job, and stim-
ulated me in becoming and being research-
coordinator. And Hans A. M. de Kruijf who not
only helped me to get a permanent position at the
National Institute of Public Health and Environ-
mental Protection, but also was a help in finish-
ing the WQL tasks yet unfinished. After seven-
teen years on various contract positions it was
very restful, but fortunately not uneventful. I am
obliged to the National Institute of Public Health
and Environmental Protection to let me finish the
WQL, in the form of this Symposium Proceed-
ings.
References
Best, P. H., D. De Vries & A. Reins, 1984. The macro-
phytes in the Loosdrecht lakes: a story of their decline in
the course of eutrophication. Verh. int. Ver. Limno!. 22:
868-875.
Burger-Wiersma, T., M. Veenhuis, H. J. Korthals, C. C. M.
Van De Wiei & L. R. Mur, 1986. A new prokaryote con-
taining chlorophylls a and b. Nature 320: 262-264.
Ejsmont-Karabin, J., R. D. Gulati & J. Rooth, 1989. Is food
availability the main factor controlling the abundance of
Euchlanis dilatata lucksiana Haver in a shallow eutrophic
lake? Hydrobiologia 186(187: 29-34.
Giussani, G. L. Van Liere & B. Moss, 1991. Ecosystem Re-
search in Freshwater Environment Recovery. Mem. 1st.
ita!' Idrobio!. 48: 1-362.
Gulati, R. D., L. Van Liere & K. Siewertsen, 1991. The Loos-
drecht lake system: Man's role in its creation, perturbation
and rehabilitation. In: O. Ravera (ed.), Terrestrial and
aquatic ecosystems: perturbation and recovery pp. 593-
606. Ellis Horwood Limited, Chicester.
Gulati, R. D. & L. Van Liere, 1992. Waterquality research in
Loosdrecht lakes: the salient features. Hydrobiologia 233:
171-177.
Gulati, R. D., A.L. Ooms-Wilms, O. F. R. Tongeren, G.
Postema & K. Siewertsen, 1992. The dynamics and role of
limnetic zooplankton in Loosdrecht lakes (The Nether-
lands) Hydrobiologia 233: 69-86.
Hofstra, J. J. & L. Van Liere, 1992. The state of the environ-
ment of the Loosdrecht lakes. Hydrobiologia 233: 11-20.
Loogman, J. G. & L. Van Liere (eds), 1986. Restoration of
shallow lake ecosystems, with emphasis on Loosdrecht
lakes. Hydrobio!. Bull. 20: 1-259.
Malthus, T. J., E. P. H. & A. G. Dekker, 1990. An assess-
ment of the importance of emergent and floating-leaved
macrophytes to trophic status in the Loosdrecht lakes (The
Netherlands). Hydrobiologia 191: 257-264.
Province of Utrecht, 1980. Over de keuze van een suppletie-
model ter beveiliging van de waterkwaliteit van de Loos-
drechtse Plassen (The choice of a supply model to safe-

guard Loosdrecht lakes water quality). Report of the
Province of Utrecht, Utrecht. 90 pp.
Province of Utrecht, 1973. Over de veiligstelling van de wa-
terkwaliteit van de Loosdrechtse Plassen (Safeguarding the
Loosdrecht lakes water quality). Report of the Province of
Utrecht, Utrecht. 63 pp.
Van Densen, W. L. T., C. Dijkers & R. Veerman, 1986. Fish
community of the Loosdrecht lakes and the perspective for
biomanipulation. Hydrobiol. Bull. 20: 147-163.
Van Heusden, G. P. H., 1942. Planktongroei in en fosfaat- en
nitraatgehalte van de Loosdrechtse Plassen (in Dutch -
Growth of plankton and concentrations of phosphate and
nitrate in the Loosdrecht lakes) Amsterdam Municipal wa-
terworks. Report 11. 11 pp.
9
Van Liere, L., O. F. R. Van Tongeren, L. Breebaart & W.
Kats, 1990. Trends in chlorophyll and total-phosphorus in
Loosdrecht lakes, The Netherlands. Verh. int. Ver. Limnol.
24: 707-710.
Van Liere, L., 1992. Publications of the workgroup WQL
(Water Quality Research Loosdrecht Lakes) 1983-present.
Hydrobiologia 233: 179-186.
Van Tongeren, O. F. R., L. Van Liere, R. D. Gulati, G.
Postema & P. J. Boesewinkel-De Bruyn, 1992. Multivariate
analysis of the plankton communities in the Loosdrecht
lakes: relationships with the physical and chemical envi-
ronment. Hydrobiologia 233: 105-117.
Hydrobiologia 233: 11-20, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 11
© 1992 Kluwer Academic Publishers.

The state of the environment of the Loosdrecht lakes

Jaap J. Hofstra 1,3 & Louis Van Liere 1,2,*

1 National Institute of Public Health and Environmental Protection, P. O. Box 1, 3720 BA Bitthoven, The
Netherlands; 2 Limnological Institute Nieuwersluis, Vijverhof Laboratorium, Rijksstraatweg 6, 3631 AC
Nieuwersluis, The Netherlands; 3 present address: P.O. Box 80300, 3508 TH Utrecht, The Netherlands
(* author for correspondence)

Key words: eutrophication, Loosdrecht lakes, 'AMOEBE-approach', water quality, ecological

assessment

Abstract

The Loosdrecht lakes are a system of shallow, interconnected, peat lakes in the centre of The Nether-
lands. The main environmental functions of the Loosdrecht lakes are nature and recreation. From the
point of view of the Dutch policy, a Specific Environmental Quality ('Bijzondere Milieukwaliteit') should
be set for these lakes.
The most serious environmental problem of the area is eutrophication. The Loosdrecht lakes have,
by increasing external phosphorus loading, changed, from clear lakes with few macrophytes, followed
by a period of abundant characean growth, to turbid lakes dominated by cyanobacteria and detrital
matter. Eutrophication was counteracted by use of sewerage systems and dephosporization of the supply
water. The resultant decrease in external phosphorus loading did not result in a decrease of turbidity
by suspended particles.
The eutrophication of the lake ecosystems was described as a series of phases. One of those phases,
the status around 1940, has been used as an ecological reference system.
By means of a graphical presentation technique, the so-called 'AMOEBE-approach', the state of the
environment of the Loosdrecht lakes has been visualized. Thirty-two ecological parameters, including
both biotic and abiotic factors, have been selected and quantified. Concrete target values for these pa-
rameters have been derived from historical reports and from Lake Western Loenderveen, located close
to the Loosdrecht lakes, but less eutrophic.
The general conclusion is that the state of the environment of the Loosdrecht lakes is far from what
is required with respect to a Specific Environmental Quality, as many of the selected parameters, like
water transparency, total phosphorus, mineral nitrogen, cyanobacteria, bream, pike, macrophytes, birds
and otter, deviate by over an order of magnitude from their desired levels.

Introduction collected for about a century. Historical data have

been reported by The Amsterdam Municipal Wa-
The state of the environment of the Loosdrecht terworks (Heymann, 1922; Van Heusden, 1942
lakes is well-documented. This is especially true and Geelen, 1955). From 1964 until present the
for the water quality. Data on this topic have been water quality of the Loosdrecht lakes has been
12

monitored by the Water Authorities of the Pro-

vince of Utrecht (Province of Utrecht 1975, 1982
and 1989). Papers describing the former flora and
fauna of the area are: Van Heusden (1942), Geelen
(1955), Leentvaar & Morzer Bruins (1962) and
Prud'homme Van Reine & Van Der Meulen
(1961 ).
In the last two decades the Loosdrecht lakes
have become important topics of ecological re-
search. The project 'Water Quality Research
Loosdrecht lakes' (WQL), which has been under-
taken from 1979 to 1991, was a research pro-
gramme involving eight institutes (Loogman &
Van Liere, 1986).
In this paper we describe the lakes' morphol-
ogy and present information on their creation,
hydrology and history of eutrophication. The
present environmental state is compared to a
chosen, former, mesotrophic phase. We make use
of the so-called 'AMOEBE-approach', a graphi-
cal presentation technique which has been devel-
oped by Ten Brink & Hosper (1989) for The Fig. 1. A map of the Loosdrecht lakes area. Arrows indicate
Dutch Water Management Plant (Ministry of the sluices used to supply water from the R. Vecht to the lakes.
Traffic and Public Works, 1989). AMOEBE L. Eastern Loenderveen (LEL), L. Eastern Loenderveen
means 'General method for ecological assess- (LEL) and The Reservoir are hydrologically separated from
the Loosdrecht lakes' system.
ment' (Algemene Methode voor Oecologische
BEoordeling) and was devised for policy makers
and environmental management authorities.
44 km 2, including Lake Western Loenderveen
(LWL), the Reservoir and Lake Eastern Loender-
Description of the lakes
veen. The hydraulic retention time of the Loos-
General drecht lakes is about six months. More detailed
information on the morphometry and topography
The Loosdrecht lakes (Fig. 1), a system of shal- of the Loosdrecht lakes has been reported by
low, interconnected lakes, are typical of Dutch Gulati et al. (1991) and Van Liere et al. (1991).
peat lakes. They lie in the centre of the Nether-
lands in the Province of Utrecht in an area of
Sphagnum peatland. The Loosdrecht lakes (Lake Creation of the lakes
Loosdrecht, Lake Breukeleveen, Lake Vuntus
and Kievits Area) are bordered by the River Vecht The Loosdrecht lakes originated from human ac-
to the west, Lake Eastern Loenderveen (LEL) to tivities. In medieval times the land was drained
the north, the Pleistocene, ice-pushed sand ridge for agricultural purposes. In 1633, so-called 'in-
't Gooi to the east and Polder Bethune and Tien- dustrial peatmining' was started. The peat was
hoven lakes to the south. The total water area of dredged from underneath the watersurface and
the lakes is 14.5 km 2 , including the Kievits Area. left to dry on adjacent banks. In this wayan area
The mean depth is 1.85 m and the total water with cuts ('trekgaten') and banks ('legakkers'),
volume is 20.10 6 m 3 . The catchment area is was formed. As the peatmining continued, the

banks got smaller and subsequent wind and wave
action eroded the banks. A system of shallow,
interconnected lakes originated. The pattern of
cuts and banks has been preserved in certain parts
of the lakes, especially in the Kievits Area, which
was mined at the end of the 19th century. At this
time the Loosdrecht lakes had already attained
their present morphometry (Van Liere et aI.,
1991). More details on the creation of the lakes
have been reported by Gulati et al. (1991).
Hydrology
For about 150 years the hydrology of the Loos-
drecht lakes has been largely controlled by tech-
nical operations (Engelen, 1986). Halfway
through the 19th century several polders, includ-
ing Polder Bethune, were reclaimed. Conse-
quently groundwater, which initially percolated
into the lakes, seeped for the greater part to the
lower lying polders. To maintain water levels dur-
ing dry spells in summer, water from the River
Vecht was pumped into the lakes.
In 1932 the Amsterdam Municipal Waterworks
started to use water from the Polder Bethune for
drinking water supply. Water from the Polder Be-
thune was pumped into Lake Loosdrecht and
then to L. Loenderveen, and from there through
a pipeline to Amsterdam. Since 1959 most of the
water from Polder Bethune has flowed through a
separate canal directly into a deepened part of
Lake Loenderveen, called the 'Reservoir' (Fig. 1).
The water let into the Reservoir is dephospho-
rized. The remaining water from Polder Bethune
has flowed untreated into the Loosdrecht lakes.
The exploitation of water from Polder Bethune,
and the withdrawal of groundwater of 'Het Gooi',
resulted in the requirement for the continued inlet
of allochthonous water to the Loosdrecht lakes.
Until 1984 the River Vecht, which was highly
polluted with organic waste and nutrients from
the city of Utrecht, remained the main source of
supplemental supply. In 1984 the inlet of water
from the River Vecht was replaced by inlet of
dephosphorized water from the Amsterdam-
Rhine Canal.
13
The environmental problems of the area
The main environmental problems of the Loos-
drecht lakes are eutrophication as a result of the
high external phosphorus loading in past decades,
accumulation of heavy metals, especially cad-
mium, in the sediments (Kuhnel, 1985), the de-
struction of banks by erosion in the 18th and 19th
centuries and the disturbance of the fauna by rec-
reation and habitation over the second half of the
20th century. Eutrophication research has been
the major focus of the Water Quality Research
Loosdrecht lakes programme (WQL) and is well-
documented (Van Liere, 1992). This paper mainly
focusses on the eutrophication of the lakes but
comments on some aspects of other environmen-
tal problems are also made.
Eutrophication
Due to a prolonged external nutrient loading in
this century the Loosdrecht lakes have become
eutrophic. Important sources of external nutri-
ents, especially phosphorus, have been the supply
of water from the R. Vecht, the discharge of un-
treated sewage from surrounding villages, run-off
from the eastern hinterland and recreation.
U ntil1944 the external phosphorus loading was
estimated to be 0.6gPm- 2 y-l (Gulati etal.,
1991). Between 1944 and 1984 the mean external
phosphorus loading was about 1.0 g P m - 2 Y- 1.
In 1976, an extremely dry year, the external phos-
phorus loading was as high as 4.0 g P m - 2 Y- 1.
Due to measures taken by The Amsterdam Mu-
nicipal Waterworks and The Water Authorities of
the Province of Utrecht the external phosphorus
load has been reduced to 0.35 g P m - 2 y- 1 since
1984. These measures included the sewering of
surrounding villages between 1970 and 1986
(sewage is now transported to purification plants
outside the lake area) and replacement of the inlet
of highly polluted water from the R. Vecht by the
inlet of dephosphorized water from the Amster-
dam-Rhine Canal in 1984.
The high and prolonged external nutrient load-
ing has had dramatic consequences for the envi-
ronment of the Loosdrecht lakes. In about 50
14
years the Loosdrecht lakes have changed from
oligotrophic ecosystems, with a clearly visible
bottom, into highly eutrophic ones, with low
transparency (Secchi-disc depth < 0.4 m). The
course of eutrophication of the Loosdrecht lakes
is described below as a series of phases. A sim-
ilar approach has been used by Moss (1989) to
describe the eutrophication of the Norfolk Broads
(UK). Although the exact distinction ofthe phases
is arbitrary, Moss' concept (1989) is supported by
other studies (Van Liere et al., 1991; lanse & Al-
denberg, 1990; Scheffer, 1989 and Gulati et al.,
1991).
Phase 1
Time span: 1800-1920
Characterization: strongly nutrient limited (olig-
otrophic)
Clear water (Heymann, 1922); a virtual absence
of submerged waterplants (Best et al., 1984); a
estimated fish population biomass of about
50 kg ha - 1 (Lammens, pers. comm.);
Phase 2
Time span: 1930-1955
Characterization: nutrient limited; domination by
submerged macrophytes (mesotrophic)
Clear water (Van Heusden, 1942); well-developed
submerged vegetation of Characeae (open water)
and Potamogetanaceae (periphery) (Van Heusden,
1942); a relatively large amount of zooplankton
compared to phytoplankton (Geelen, 1955); a
well-developed population of zebra mussels
(Dreissena polymorpha) (Leentvaar & Morzer
Bruijns, 1962); a large population of hibernating
waterbirds, especially pochards (Anthya jerina),
tufted ducks (Anthya juligula) and coots (Fulica
atra) (Leentvaar & Morzer Bruijns, 1962); a res-
ident population of otters (Lutra lutra)
(Creutzberg et al., 1969); an estimated fish biom-
ass of about 250 kg ha - I, the dominant species
being pike (Esox lucius, 45 kg ha - I) and bream
(Abramis brama, 50 kg ha - I) (Lammens, pers.
comm.); annual average concentrations of
chlorophyll a about 25 J.lg 1- 1 (Geelen, 1955;
CUWVO, 1988; Hofstra, 1990) and total phos-
phorus about 5.5 J.lg P 1- 1 (Van Liereet al., 1991);
a mineral nitrogen concentration of about
17 J.lmoll- 1 (summer average) (Van Liere et al.,
1991);
In Lake Breukeleveen Phase 2 was followed by a
period of abundant growth of tall, submerged,
floating and emergent macrophytes, especially
Potamogetonaceae (Prud'homme Van Reine &
Van Der Meulen, 1961). This period lasted from
1955 to 1965.
Phase 3
Time span: 1960-1980
Characterization: nutrient limited; cyanobacteria
domination (eutrophic)
No waterplants submerged (Best et al., 1984);
dense cyanobacterial growth, limited in its growth
by light-energy, the dominant species being
Oscillatoria redekei, Oscillatoria limnetica and pos-
sibly the recently discovered prochlorophyte
Prochlorothrix hollandica (Van Liere et al., 1991);
a large population of small zooplankters (Gulati,
1984); no fresh water mussels; no otters, few hi-
bernating waterbirds (Morel, 1989; Hofstra,
1990); a fish biomass of about 300 kg ha - I, dom-
inated by bream (180 kg ha - 1) and pike-perch
(Stizostedion lucioperca, 40 kg ha - I) (Lammens,
pers. comm.); annual average concentrations of
chlorophyll a about 100 J.lg 1- 1 and total phos-
phorus about 100 J.lg P 1- I; a mineral nitrogen
concentration of about 60 J.lmoll- 1 (summer av-
erage) (Van Liere et al., 1991);
Phase 4
Time span: 1980-present
Characterization: high internal phosphorus load-
ing rate; detritus domination (eutrophic)
Like Phase 3 but seston « 150 J.lm) is now dom-
inated (50-75 per cent) by detritus (Van Liere
et al., 1991).
Phase 4 is at present a stable state of the lakes.
Measures that have been taken to reduce the ex-

ternal phosphorus loading rate to the Loosdrecht
lakes have resulted in only a small decline of total
phosphorus concentration since 1984. The eco-
system is apparently well-buffered against
changes in external nutrient supply, due to high
internal nutrient loading rates (Van Liere & J anse,
1992).
Accumulation of heavy metals in the Loosdrecht
Lakes
The main sources of heavy metals to the Loos-
drecht lakes are the inlet of allochthonous water
(carrying polluted sediment), atmospheric depo-
sition, agricultural effluents and local emissions
(Kuhnel, 1985). The inlet of water from the R.
Vecht caused peak concentrations of heavy met-
als in the sediments in the Kievits Area and in the
southern part of Lake Loosdrecht (Table 1). The
average concentration of heavy metals in the sedi-
ment of the Loosdrecht lakes as a whole is lower.
Between 1979 and 1984 the cadmium concentra-
tion in the sediment of the Loosdrecht lakes has
increased sharply (Kuhnel, 1985).
Destruction of banks
Topographical maps show that most residual
banks, the strips of land that remained after the
industrial peatmining, have disappeared. The de-
Tahle 1. Concentrations (mg kg - I) of cadmium, zinc and
lead in the sediments of the Loosdrecht lakes (spatial average
of si.\ stations) and Kievits Area (spatial average of six sta-
tions) (KOhne!. 1985). estimated background concentrations
(Hofstra. 1990). and General Environmental Quality (GEQ)
target 'alues for fresh "·ater systems from the Dutch Water
t-Ianagement Plan (t-linistry of Traffic and Public Works,
1989il
Cadmium Zinc Lead
Loosdrecht lakes 6.3 17 90
Kievits area 17.5 928 168
Background (reference value) 0.3 100 30
Objective (GEQ)
2 480 530
15
struction of the banks as well as eutrophication
has contributed to the present amount of detritus
in the Loosdrecht lakes. On windy days signifi-
cant amounts of detritus are resuspended lower-
ing the transparency of the water. The finer frac-
tion ofdetritus, which consists merely of remnants
of algae, may remain suspended for days or even
weeks (GOllS & Van Keulen, 1989).
Disturbance of the fauna by recreation and habita-
tion
Since World War II the Loosdrecht lakes area
has become increasingly populated. Recreation in
the area is still on the increase. In recent decades
yachting harbours, camping sites and holiday cot-
tages have been constructed.
A decline in the breeding populations of Sedge
Warbler (Acrocephalus schoenobaenus), Bittern
(Botaurus stellaris), Little Bittern (Ixobrychus
minutus), Night Heron (Nycticorax nycticorax) and
Purple Heron (Ardea purpurea) and the extinction
of the otter (Lutra lutra) may be partially attrib-
uted to disturbance (Morel, 1989; Creutzberg
et al., 1969). However, a study by Reijnen (1988)
on the environmental aspects of water recreation
demonstrated that quantified dose-effect relation-
ships are not yet available.
Policy aims and ecological target values for fresh
waters in The Netherlands
A 'General Environmental Quality' (GEQ; 'AI-
gemene Milieukwaliteit') is the policy objective
for all ecosystems in The Netherlands, both ter-
restrial and aquatic. G EQ guarantees 'health and
maintenance of plants, animals, goods and forms
of use' (Ministry of Housing, Physical Planning &
Environmental Protection, 1985). In some areas
a 'Specific Environmental Quality' (SEQ; 'Bij-
zondere Milieukwaliteit') is required, correspond-
ing to specific functions, such as nature, drinking
water supply, recreation or fisheries The Water
Authorities of the Province of Utrecht have as-
signed the following functions to the Loosdrecht
16
lakes: nature (Loosdrecht lakes), drinking water
supply (Polder Bethune and Lake Loenderveen)
and recreation (Loosdrecht lakes). Therefore a
SEQ seems appropriate for the area.
Ecological target values for the Loosdrecht lakes
Ecological target values for various freshwater
types, including shallow lakes, matching to SEQ
target values, have been drafted by a workgroup
'CUWVO' (1988). The CUWVO target values,
however, should merely be regarded as general
guidelines and not as operational tools for envi-
ronmental quality assessment (Kroes, 1987). So
far, concrete and verifiable ecological target val-
ues, for both biotic and abiotic parameters, have
been set only for a few Dutch aquatic ecosystems.
The Loosdrecht lakes, unfortunately, do not be-
long to them. To assess the state of the environ-
ment of the Loosdrecht lakes operational ecolog-
ical target values have first to be set.
Lakes originating as a result of peatmining and
receiving seepage water could have clear, nutrient-
poor water with corresponding plant and animal
communities (CUWVO, 1988). Therefore
Phase 2 above, corresponding to the situation
around 1940 and being a well-documented period
in the ecological history of the Loosdrecht lakes,
has been chosen as a 'frame of reference'. Target
values for 32 ecological parameters that prevailed
during this period have been chosen (Hofstra,
1990; Hofstra et al., 1991) and are summarized in
Table 2 in which also the present, Phase 4, values
of the different parameters are summarized.
Criteria for the selection of parameters
A first choice of the ecological parameters was
made according to expert judgement. A final se-
lection of the parameters was made according to
the method of Klijn et al. (1990). The criteria of
Klijn et al. (1990) were: firstly, parameters should
correspond to specific environmental functions of
the area in study, especially nature, drinking
water, recreation, agriculture, forestry and fish-
ery. Secondly, parameters should reflect the en-
vironmental problems of the area. Thirdly, data
on the parameters should be available and
fourthly, parameters should appeal to the interest
of policy makers.
A graphical presentation of the state of the lake
environment
In the Dutch Water Management Plan (Ministry
of Traffic and Public Works, 1989) a graphical
presentation technique, called the 'AMOEBE ap-
proach' was introduced. AMOEBE is an abbre-
viation for 'Algemene Methode voor OEcologis-
che BEoordeling', which means 'General Method
for Ecological Assessment'. The method was de-
signed for policy makers and environmental au-
thorities and has been applied to the North Sea,
the River Rhine and some terrestrial ecosystems
(Ten Brink & Hosper, 1989; Klijn et al., 1990;
Latour et al., 1990). In Fig. 2 an AMOEBE, con-
taining the parameters from Table 2, is presented.
The sequence of the parameters corresponds to
the method of Klijn et al. (1990).
Evaluation of the state of the environment
The AMOEBE (Fig. 2) clearly demonstrates that
the state of the environment of the Loosdrecht
lakes does not match with the frame of reference.
Some parameters, total phosphorus, cadmium,
cyanobacteria, diatoms and seston have increased
since 1940. Others, transparency, green algae,
zebra mussel larvae, Characeae, pike, sedge war-
bler and otter, are below their target value. Only
a few parameters, oxygen, calcium-ions and sa-
linity, approach their aims.
The AMOEBE-approach forces the choice of
concrete target values for ecological parameters.
A complete and verifiable set of target values can
only be obtained if the study area is well-
documented. The Loosdrecht lakes are a good
example of such an area. But even if historical
data are lacking or incomplete the setting of eco-
logical aims is an indispensable step in environ-
mental quality assessment (Public Health Coun-
cil, 1989). In fact, it should be the first step in
 17

Table 2. A survey of the present concentrations/values (Phase 4) and target values (Phase 2) for 32 ecological parameters (after
Hofstra, 1990).

Parameter Present concentration/ Desired concentration/

value value

Abiotic parameters
Transparency (yearly average, Secchi-disc depth, m) 0.4 *a 1.9 *a
Sestion < 150 flm (yearly average, mg C 1- 1) 8.0 *b 0.2 *cd
Total phosphorus (yearly average, mg P 1- 1) 0.1 *a 0.0054 *ef
Soluble Reactive Phosphorus (yearly avg., mg P 1- 1) 0.002 *b 0.0015 *g
Mineral nitrogen (*1, summer average, mg N 1- ') 0.840 *a 0.238 *a
Salinity (*2, summer average, mg 1- ') 305 *a 231 *a
Calcium-ions (summer average, mg 1- ') 49.0 *a 40.8 *a
Sulphate (summer average, mg 1- ') 19.2 *a 33.6 *a
Bicarbonate (summer average, mg 1- 1) 153 *a 92 *a
pH (summer average, pH-units) 8.9 *a 7.8 *a
Oxygen (summer average, mg O 2 1- ') 10.4 *a 11.0 *a
Cadmium (in sediment, mg kg-I) 6.3 *h 0.3 *f

Plankton (yearly averages)

Zebra mussel larvae (fresh weight, mg 1- ') o *c 0.0038 *e
Total zooplankton (carbon, mg C 1- 1) 0.575 *d 0.075 *de
Total cyanobacteria (fresh weight, mg 1- 1) 30.6 *b 1.25 *e
Total diatoms (fresh weight, mg 1- ') 0.81 *b 0.06 *e
Total green algae (fresh weight, mg 1- 1) o *b 0.48 *e
Chlorophyll a (mg 1- 1) 0.1 *a 0.025 *fi

Vegetation-groups
unit: Total covering according to Tansley (1946) (sum of Tansley abundance terms)
Characeae (4 species) 0 *fjk 15 *fjk
Potamogeton-group (10 species) 1 *fjk 11 *fjk
Menyanthus-group (7 species) 2 *fjk 14 *fjk
Thelypteris-group (15 species) 6 *fjk 25 *fjk
Nymphaea-group (7 species) 14 *fjk 18 *fjk
Butomus-group (6 species) 11 *fjk 15 *fjk
Caltha-group (10 species) 14 *fjk 24 *fjk

Fauna
Pike (fresh weight, kg ha - I) 1 *1 45 *1
Bream (fresh weight, kg ha) 180 *1 50 *1
Otter (*3, nr. individuals) 0 *f 15 *m
Pochard (*4, nr. individuals) 750 *f 7500 *0
Sedge Warbler (*5, breeding pairs) 5 *n 50 *f
Little Bittern (*5, breeding pairs) I *n 5 *f
Great Crested Grebe (*5, breeding pairs) 70 *n 20 *f

Remarks
*1 Sum of the ions N0 2- , N0 3- and NH;
*2 Sum of ions SO~-, HC0 3-, Cl-, Mg2 +, Na + and K +
*3 Number of individuals in Southern Vecht lakes area (Crutzberg et al., 1969)
*4 Number of individuals hibernating in the Loosdrecht lakes ares (Leentvaar & Morzer Bruijns, 1962)
*5 Number of breeding pairs in Soutern Kievits Area (Morel, 1989)

References supporting this table

*a Van Liere etal., 1991; *b Breebaart et al., 1989; *c Gulati 1990, pers. comm.; *d Gulati, 1984; *e Geelen, 1955; *fHofstra,
1990; *g Amsterdam Municipal Waterworks, 1940; *h KOhnel, 1985; *i CUWVO, 1988; *j Best et al., 1984; *k Runhaar et al.,
1987; *1 Lammens, pers. comm.; *m Creutzberg et al., 1969; *n Morel, 1989; *0 Leentvaar & Morzer-Bruijns, 1962.
18
." .~~:1~~"""" '"
,..... .... ..... ..... ....
......
Ik~~<i'~"
'k>
Ipresent situation
_ _ _ reference circle (situation around 1940)
logarithmic scale
• pH has been plotted linearly
Fig. 2. 'AMOEBE' presentation of the state of the environ-
ment of the Loosdrecht lakes. The circle (thicker line) corre-
sponds to a state of the ecosystem (around 1940; see 'Phase 2'
under the text), used as an ecological frame of reference. The
shaded parts represent the present state of the environment
(,Phase 4'). The magnitude of deviations from the circle indi-
cates the extent of unfavourable environmental conditions,
which are due to a number of courses, described in the text,
such as eutrophication, disturbance of the fauna and destruc-
tion of banks. Note that the graph has been plotted on a
logarithmic scale except for pH.
translatiqg general policy alms to directives for
managemt<nt·
One m~st ~ear in mind, however, that the tar-
get values that have been set in this study are
based on a variety of sources, including historical
reports and data from Lake Western Loender-
veen (Hofstra, 1990). No attempts have been
made to describe seasonal fluctuations or spatial
variation in each parameter. Therefore, the target
values should be considered as a rough draft of
Specific Environmental Quality for the Loos-
drecht lakes. Exact upper and lower limits for
each parameter have not been established, nor
could have been at this stage. In the future a more
precise definition of ecological aims for the area
might be desirable, for instance if, by proper man-
agement, the lakes approach their Phase 2 situa-
tion again.
Although we are unable presently to formulate
exact ecological aims, a general conclusion can be
generated from this study: the state of the envi-
ronment of the Loosdrecht lakes is poor with
respect to the Specific Environmental Quality that
is required, especially with respect to its main
functions, nature and recreation. Measures that
combat eutrophication are strongly recom-
mended.
Conclusions
1. The Loosdrecht lakes have changed in the past
century from a oligotrophic ecosystem with very
clear water into a highly eutrophic one with very
low transparency.
2. The state of the environment ofthe Loosdrecht
lakes is poor with respect to the Specific Envi-
ronmental Quality target values that are required
for the area.
3. The situation around 1940, described in the
text as 'Phase 2', is recommended as a concrete
and verifiable ecological frame of reference or tar-
get, corresponding to the policy aim of Specific
Environmental Quality.
4. The AMOEBE presentation technique, de-
vised for policy makers, is a valuable tool in en-
vironmental quality assessment as it visualizes in
one glimpse the state of the environment of an
area as well as a frame of reference, i.e. level
desired.
References
Amsterdam Municipal Waterworks, 1940. Rapport 1940. In-
zake de watervoorziening van Amsterdam - Report 1940.
The watersupply of the city Amsterdam. Amsterdam.
575 pp.
Best, E. P. H., D. De Vries & A. Reins, 1984. The macro-
phytes in the Loosdrecht lakes: a story of their decline in

the course of eutrophication. Verh. Int. Ver. Limnol. 22:
829-834.
Breebaart, L., J. Ebert & L. Van Liere, 1989. Waterkwaliteit-
sonderzoek Loosdrechtse Plassen. Basisgegevens 1983-
1988 - Water Quality research Loosdrecht lakes. Data from
the period 1983-1988. WQL-report 1989-5. Limnological
Institute, Nieuwersluis. 153 pp.
Creutzberg, F., P. Leentvaar, R. Rense, C. T. B. Rikkert de
Koe, H. A. de Vries & K. W. R. Zwart, 1969. De Zuidelijke
Vechtplassen. Flora en fauna - The Southern Vecht lakes.
Flora and fauna. P. Leentvaar (ed) RIVON-verhandeling 7.
205 pp.
CUWVO (Coordinatiecommissie Uitvoering Wet Verontrein-
iging Oppervlaktewateren. Werkgroep V-I, 1988. Ecologis-
che normdoelstellingen voor Nederlandse oppervlaktewa-
teren - Ecological objectives for Dutch surface waters.
212 pp.
Engelen, G. B., 1986. Interaction of hydrological systems and
eutrophication of the Loosdrecht lakes. Hydrobiol. Bull.
20: 17-25.
Geelen, J. F. M., 1955. Het plankton in de plassen in 1954-
The plankton of the Loosdrecht lakes in 1954. Amsterdam
Municipal Waterworks. Report 38. 74 pp.
Gons, H. J. & R. van Keulen, 1989. De relatie tussen doorz-
icht en slib in de Loosdrechtse Plassen - The relation be-
tween transparency and sludge in the Loosdrecht lakes.
WQL-Report 1989-4. Limnological Institute, Nieuwer-
sluis. 35 pp.
Gulati, R. D., 1984. The zooplankton and its grazing as mea-
sures of trophy in the Loosdrecht lakes. Verh. int. Ver.
Limnol. 22: 863-867.
Gulati, R. D., L. Van Liere & K. Siewertsen, 1991. The Loos-
drecht lake system: Man's role in its creation, perturbation
and rehabilitation. In: O. Ravera (ed.), Terrestrial and
aquatic ecosystems. Perturbation and recovery. Ellis Hor-
\\ood. New York, London. pp: 593-606.
Heymann. J. A .. 1922. Onderzoek van de Loosdrechtsche
Plassen op maandag 12 juni 1922 - Research in the Loos-
drecht lakes on monday. June 12th, 1922. Amsterdam Mu-
nicipal \\·atemorks. 7 pp.
Hofstra. J. J .. 1990. De milieutoestand van de Loosdrechtse
Plassen \\eergege\en door mid del van een 'AMOEBE' -
State of the en\ironment of the Loosdrecht lakes reflected
by means of an ·A~IOEBE'. RIVM-report No. 751901004.
72 pp.
Hofstra. J. J .. J. H. Janse & L. Van Liere, 1991. De milieu-
toestand \an de Loosdrechtse Plassen. Toepassing van de
presentatietechniek AMOEBE en het eutrofieringsmodel
PCLOOS - The state of the environment of the Loosdrecht
lakes. Application of the presentation technique
'.-\MOEBE' and the eutrophication model 'PCLOOS'. H 2 0
24: 260-265.
Janse, J. H. & T. Aldenberg, 1990. PCLOOS: a eutrophica-
tion model of the Loosdrecht lakes. WQL-report No. 1990-
1. RIVM-report No. 714502001. 91 pp.
Klijn, F., J. B. Latour, M.1. Nip, H. A. Udo De Haes,
19
M. M. H. E. Van Den Berg & J. J. Hofstra, 1990. Milieu-
kwaliteit van ecodistricten. Deel 2: Methode en aanzet tot
uitwerking - Environmental quality of'ecodistricts'. Part 2.
Method and initial application results. RIVM-report
No. 75190lO03/CML-report No. 63. 1lOpp.
Kroes, H. W., 1987. Van biologische waterbeoordeling naar
ecologische normdoelstelling: de verbinding tussen weten-
schap en beleid. In: Biologische waterbeoordeling. Instru-
ment voor waterbeheer? - From biological water quality
assessment toward ecological target values: the link be-
tween science and policy. P. F. M. Verdonschot en
L. W. G. Higler (red.). RIN, Leersum. pp: 51-62.
Kuhnel, V., 1985. Zware metalen in het Loosdrechtse
Plassen-sediment - Heavy metals in sediment of the Loos-
drecht lakes. WQL-report No. 1985-5. 36 pp.
Latour, J. B., J. J. Hofstra & M. 1. Nip, 1990. Toepasbaar-
heid van de AMOEBE-benadering op terrestrische ecosys-
temen - Application of the 'AMOEBE-approach' on ter-
res tic ecosystems. RIVM-report No. 751901001. 32 pp.
Leentvaar, P. & M. F. Morzer Bruijns, 1962. De verontrein-
iging van de Loosdrechtse Plassen en haar gevolgen - Pol-
lution of the Loosdrecht lakes and its consequences. De
Levende Natuur 65: 42-67.
Loogman, J. G. & L. Van Liere (eds.), 1986. Proceedings of
the WQL-symposium: Restoration of shallow lake ecosys-
tems, with emphasis on Loosdrecht lakes. Hydrobiol. Bull.
20: 1-259.
Ministry of Housing, Physical Planning & Environmental
Protection, 1985. Indicatief Meerjarenprogramma Mi-
lieubeheer 1986-1990 - Indicative Programme of Environ-
mental Management 1986-1990. The Hague. 141 pp.
Ministry of Traffic & Public Works, 1989. Water voor nu en
later. Derde Nota Waterhuishouding - Dutch Water Man-
agement Plan. The Haque. 297 pp.
Morel, G. A., 1989. De Zuidelijke Kievitsbuurt: Moerasvo-
gels in een natuurgebied - Southern Kievits Area: marsh-
birds in a nature area. De Kruisbek 5: 166-173.
Moss, 1989. Water pollution and the management of ecosys-
tems: a case study of science and scientist. Reprinted from:
Toward a more exact ecology. The 30th symposium of the
British Ecological Society. pp: 401-422. Blackwell Scien-
tific Publications. London.
Province of Utrecht, 1975. Kwaliteit oppervlaktewater
Utrecht. April 1964 tot en met maart 1975 - Quality of
surface water in the Province of Utrecht. April 1964 until
March 1975. 105 pp.
Province of Utrecht, 1982. Kwaliteit oppervlaktewater pro-
vincie Utrecht 1975-1980. Quality of surface water in the
Province of Utrecht 1975-1980. 154 pp.
Province of Utrecht, 1989. Waterkwaliteitsverslag 1980 tim
1986. Het waterkwaliteitsonderzoek in het oppervlaktewa-
ter van de Provincie Utrecht. Deel 1: tekst, deel 2: tabellen.
Quality of surface water in the Province of Utrecht 1980-
1986. Volume 1: text, volume 2: tables. 66 pp. 106 pp.
Prud'homme Van Reine, W. F. & P. E. Van Del' Meulen,
1961. Verslag van het onderzoekje naar het voorkomen van
20
Chara spec. in het plassengebied van Loosdrecht - Charac-
eae in the Loosdrecht lakes. RIN-report. 2 pp.
Public Health Council, 1989. Ecologische normen waterbe-
heer. Deeladvies III: beschrijving van de parameters - Eco-
logical target values for water management. Part 3. De-
scription of the parameters. Report 89(21. The Hague.
408 pp.
Reijnen, 1989. Invloed van watersport op de natuur - Influ-
ence of waterrecreation on nature. RMNO-report No. 37.
107 pp.
Runhaar, J., C. L. G. Groen, R. Van Der Meijden & R. A. M.
Stevers, 1987. Een nieuwe indeling in ecologische groepen
binnen de Nederlandse flora - A new way of dividing the
Dutch flora into ecological plant communities. Gorteria 13:
277-359.
Scheffer, M., 1989. Alternative stable states in eutrophic shal-
low freshwater systems: a minimal model. Hydrobiol. Bull.
23: 73-83.
Tansley, A. G., 1946. Introduction to plant ecology. Allen &
Unwin, London.
Ten Brink, B. J. E. & S. H. Hosper, 1989. Naar toetsbare
ecologische doelstellingen voor het waterbeheer: de
AMOEBE benadering - Towards verifiable ecological tar-
get values for water management: the AMOEBE-approach.
H 2 0 22: 612-617.
Van Heusden, G. P. H., 1942. Planktongroei in en fosfaat- en
nitraatgehalte van de Loosdrechtse Plassen - Growth of
plankton and concentration of phosphate and nitrate in de
Loosdrecht lakes. Amsterdam Municipal Waterworks. Re-
port 11. 11 pp.
Van Liere, L., 1992. Publications of the workgroup WQL
(Water Quality Research Loosdrecht Lakes) 1983 - present.
Hydrobiologia 233: 179-186.
Van Liere, L., J. Ebert, W. Kats & J. J. Buyse, 1991. The
water quality of Loosdrecht lakes, a review. Mem. 1st. ital.
Idrobiol. 48: (in press).
Van Liere, L. & J. H. Janse, 1992. Restoration and resilience
to recovery of the Lake Loosdrecht ecosystem in relation
to its phosphorus flow. Hydrobiologia 233: 95-104.
Hydrobiologia 233: 21-38, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 21
© 1992 Kluwer Academic Publishers.

The hydrology of the Loosdrecht lakes area

G. B. Engelen, B. F. M. Kal, J. J. Buyse & F. G. M. Van Pruissen

Free University, Fac. of Earth Sciences, Dept. of Hydrology, De Boelelaan 1085, 1081 HV Amsterdam,
The Netherlands

Key words: lake restoration, hydrological systems, nutrient balance, groundwater modelling, water
management

Abstract

The hydrological research of the Loosdrecht Lakes and their surroundings over the period 1982-1990
was based on the integrated systems approach to surface- and ground water, with respect to their quantity
and quality.
The description of the water systems includes an evaluation of the structure and interactions of water
systems in and around the lakes since the beginning of the Holocene until recent times, focussing pro-
gressively from the regional to the local systems and from the lake complex to the individual lakes.
Quantitative research concerned: monthly and annual water balances for the complex of lakes and
for individual lakes; numerical two- and semi-three-dimensional groundwater flow models; field exper-
iments on turbulence and diffusion; and experimental laboratory studies on wind-driven currents.
Qualitative research concerned: behaviour of environmental isotopes in groundwater; heavy metals in
lake bottom sediments; phosphorus balances of the complex of lakes and of individual lakes for the
existing and the simulated conditions; and remote sensing studies for the water quality pattern recog-
nition.
The interaction between the various water systems in the area has been influenced strongly in historic
and recent times by human interference on both the local and regional scales of the water resources, e.g.
through the reclamation of lakes, the extraction of water for public water supply, the regional transfer
of water, the agricultural and urban activities etc. Two major remedial measures have been taken thus
far: construction of sewer systems, and the supply with dephosphorized water from the Amsterdam-
Rhine canal. These measures resulted in a decrease of the external P-load but did not result in a sig-
nificant decrease in eutrophication as yet.
Scenario's to improve this situation either pertain to tackling the fundamental causes of the eutroph-
ication or to redressing their effects. The first category includes: drastic reductions in groundwater ab-
straction for public supply in the Gooi catchment area of the lakes and (partial?) flooding of adjacent
deep polder areas to reduce the leakage losses from the lakes (both measures aim at restoration of the
water balance). The second category involves less drastic water management alternatives for lake res-
toration: dredging P-enriched bottom sediments; reinforced flushing during summer with dephosphorized
Amsterdam-Rhine Canal water; interception and rerouting of polluted inflows from the watershed; or
additional P-removal from yet untreated inflows.
22
Introduction
The hydrological studies in the interdisciplinary
workgroup WQL (Water Quality research Loos-
drecht lakes) on restoration of the Loosdrecht
Lakes have been carried from 1982-1988 (Kal
et al., 1984; Engelen & Kal, 1985; Kal, 1986). The
aim of the study was to provide an integrated
description of the abiotic part of the water sys-
tems and its contribution to the functioning of the
aquatic ecosystems in the Loosdrecht Lakes (En-
gelen, 1986).
The results had to serve the following purposes:
General framework. The integrated hydrological
studies had to provide a general frame of refer-
ence for the other project participants represent-
ing other disciplines as well as for the national,
regional and provincial authorities which are in
charge of planning and management of the lakes
area.
Input data for biological and management models.
The data on water quantity and quality were to be
used as: part of the input of a eutrophication
simulation model for the Loosdrecht lakes
(PCLOOS, Janse & Aldenberg, 1990); input for
the simulation models for scenario's of surface
water management (Buyse, 1989); input for the
simulation models for scenario's of groundwater
management (Van Pruissen, unpublished); input
for the development of plans for engineering
works in the surface water network, related to
different management scenario's.
The Loosdrecht Lakes are connected to their sur-
roundings by natural, semi-natural and artificial
flows of surface- and groundwaters, nutrients,
etc., which act as external boundary conditions
for the development of the ecosystems in the lakes
(Engelen & Schot, 1990).
The complexity of the interactions between
groundwater and surface water in the area - am-
plified by intensive human interferences through
the water management activities in the area -
required a methodological approach in which the
interactions of surface waters and groundwater
are considered quantitatively and qualitatively
and within a wide range of spatial and time scales.
The present hydrological situation in the area
is the result of a complex sequence of events of
natural and man-made origin. The present distri-
bution of nutrients in the water systems of the
area can be understood only when the hydrolog-
ical history is traced back. One has to start at the
beginning of the Holocene period and to follow
up closely the increasing interference of man.
Over the last centuries and decades man has been
changing the flow systems around the lakes by
digging peat, draining lakes, extracting ground-
water, maintaining fixed surface water levels in
polders and transferring water from adjacent riv-
ers and canals to supply the lakes to maintain
their water levels.
This paper presents a description of the Loos-
drecht lakes hydrology, its history and interac-
tions with other systems. A water- and matter-
balance is set up for the whole lake area, as well
for its compartments. Since the measures taken
did not led to the result expected, additional mea-
sures, from a hydrological point of view, will be
presented.
The study revealed that the main forcing func-
tions for the response of the ecosystems and the
size and distribution of the eutrophication are
external factors like hydrological structure of the
area and its modifications by water management
practices and superposed pollution. The hydro-
logical studies had both purely scientific and ap-
plied aspects. Moreover a large number of pub-
lic agencies at various levels of administration
have been involved: national agencies dealing with
the environment, the provincial government of
Utrecht, the water management board of the
lakes, the public water supply company of the city
of Amsterdam, etc. The water systems approach
provided the logical background for the integra-
tion of all different aspects.
Methodology
The complicated interactions between natural and
human, biotic and abiotic factors in the area re-

quire an interdisciplinary, holistic systems ap-
proach. The hydrological integration has been
achieved by applying the regional hydrological
systems concept (Engelen & Jones, 1986). In this
concept the hierarchy of flow systems of the sur-
face waters and groundwater and their coupling
are emphasized. Moreover the quantity- and
quality aspects of both are intimately related and
studied as such. The integration of hydrology and
ecology takes advantage of the close correlation
between hydrological and ecological zoning.
The large number of sub-themes of this hydro-
logical research can be summarised under the fol-
lowing themes.
Genesis, types and distribution of the water
systems. Descriptive and semi-quantitative stud-
ies of the water systems in the area concerning the
development of the major surface water and
groundwater flow systems in the area with time
after the end of the last glaciation in the Pleis-
tocene period (Engelen & Schot, 1989).
Research on the flow quantities and the flow pro-
cesses in individual water systems. The surface
water flows into, between and within the lakes
have been studied by the study of records of
pumping stations, the recording of water levels,
areal surveys of electrical conductivity and water
chemistry of the water bodies, the monitoring of
changes in electrical conductivity and chemistry
in connecting canals and passages, the applica-
tion of dyes and other tracers (Teunissen, 1985;
Fisser, 1986; Van Der Spek, 1986). The effects of
different water management practices as well as
of wind on currents and turbulent mixing in the
lakes were investigated in situ (Buyse, 1986).
Physical scale models of the lakes were used in
the laboratory to get an idea about the currents
generated by wind and by water management
practices and to compare their outcome with field
data (Wasser, 1988). The remote sensing study
contributed as well to this type of information
(Dekker et al., 1991). Water balances for monthly,
seasonal and annual periods have been calculated
for the lake area as a whole and for individual
lakes within the complex of lakes. The reader is
23
referred to Engelen et al. (1988) and Buyse (1989)
for full details on the calculation methods for the
balances of water, chloride and phosphorus and
for their results. Chloride has been used as a con-
servative tracer to check and correct the water
balance calculations. On the basis of those water
balances and the concentration data from field
measurement the phosphorus balances have been
calculated. Balances have been calculated for the
system of lakes as a whole and for individual
lakes within it for the following periods (Engelen
et al., 1988; Buyse, 1989):
- 1978-1983: a period with supply of strongly
polluted water from the River Vecht;
- 1983-1987: a period with supply of less pol-
luted water from the Amsterdam-Rhine Canal
in combination with phosphate removal.
Hettling (1985a) surveyed the isotope situation in
the Loosdrecht lakes area to study the interaction
between the surface water- and groundwater sys-
tems and to determine groundwater ages. After
the qualitative identification of the hierarchy of
groundwater flow systems in the project zone the
systems have been modelled. This was done first
roughly with the two-dimensional, finite differ-
ence model FLOWNET (Engelen & Jones, 1986;
Hettling, 1985b; Kuhnel, 1985a) and subse-
quently with the multi-layer, finite element model
MICROFEM (Van Pruissen, 1989). The models
were calibrated by means of surface-water-level-
and hydraulic head data, water balances of polder
units, isotopes and hydrochemical information.
The following situations have been modelled: a
wet (winter-) situation; a dry (summer-) situation;
a situation after dredging the top of the lake bot-
tom sediments; and a situation with a flooded
Polder Bethune.
Research on remedial water management measures
to combat the eutrophication of the lakes. The ef-
fects have been quantified and evaluated of:
(1) the installation of sewer systems in the area;
(2) the substitution in 1986 of water supply from
the polluted River Vecht by water supply from the
Amsterdam-Rhine canal and of phosphate re-
moval on the nutrient balances of the lakes; (3) a
24

future drastic reduction in groundwater abstrac-
tion in the catchment area of the lakes to restore
the original groundwater flow patterns and areas
of upward seepage as existed at the beginning of
the century; (4) (partial) flooding of adjacent deep
polder areas, restoring thus partly the hydrologi-
cal situation with the Lake Rorstermeer and es-
pecially the Bethune area, before their reclama-
tion into deep polders; (5) the phosphate removal
from the lake system by dredging of nutrient-rich
lake sediments. This may result in an increased
leakage from the lakes and a shift in the water-
and nutrient balances of the lakes. The possible
effect of such a measure has been evaluated (Ri-
gler et al., 1987); (6) the reinforced flushing of the
lakes during summer with dephosphorized water
from the Amsterdam-Rhine Canal to progres-
sively deplete the stock of P in the lake complex;
(7) the diversion of the nutrient-rich surface water
sources at that time were the sluices with Vecht
water intake.
Description, functioning and interactions of the
water systems
Development phases of the water systems in the
Loosdrecht lakes area
The four major phases in the development of the
landscape and of the water systems in and along
the western border of the ice-pushed hills of 't
Gooi/Utrecht since the Pleistocene period are
given in Fig. 1 (Engelen & Schot, 1989). In the
first phase, at the end of the Pleistocene epoch
WEST SANDRIDGE UTRECHT EAST

flows into the lake complex from the foothills of

the Gooi recharge area directly into the River PHASE 1
BND OF PLEISTOCENE
Vecht. It resulted in recommendations for
changes in the water management and adaptation
of the surface water network (see below); and
(8) the additional P-removal from yet untreated
inflows would decrease the P-load on the lakes PHASE 2
END OF HOLOCENE
further.
Apart from these main themes the hydrological
studies formed a framework for other research,
such as investigations to patterns of water qual-
ity in the surface and groundwater systems. These PHA$E3
were established for different seasons and water END 19TH CENTURY

management conditions (Engelen et al., 1988;

Buyse, 1989). The quality aspects were focussed
on the major inorganic components and on the
nutrients phosphorus and nitrogen. Some work
has been done on heavy metals in the lake sedi-
ments (Kuhnel, 1985b). The distribution of the
heavy metals Fe, Zn, Cd, Pb, Cu and Zn was
measured in 1984 in 17 core-samples of river- and
lake sediments. The mineralogy of three samples
was studied by differential thermal analysis and
X-ray diffraction methods. In that period the ar-
Fig. 1. Four phases in the development of the groundwater
tificial summer supply of water to the lakes was
flow systems in the Gooi/Utrecht ice-pushed hills and the
still from the polluted River Vecht via three intake Loosdrecht Lakes area: 1. End Pleistocene 2. End Holocene
sluices. Many different diffuse sources of heavy 3. End 19th century 4. Recent, before the construction of the
metals occur in the lake area, the main point Flevo polders to the north. (After Engelen & Schot, 1989).

with low sea level conditions, the infiltration in
the pervious hills of ice-pushed older fluviatile
sands and gravels generated a large groundwater
flow system with flow branches to the adjacent
valleys. In the second phase, at the end of the
Holocene period and before the historic Middle
Ages, a semi-permeable cover of Holocene sands,
clays and peat had filled in the low areas on both
sides of the ice-pushed hills. The outlet zones of
the GooijUtrecht Hills groundwater flow system
had retracted to the base of the slope at the tran-
sition between the Pleistocene and Holocene
landscapes. Groundwater emerged in upward
seepage zones and was discharged by small riv-
ers across the surrounding peat belt to larger me-
andering distributary branches of the River Rhine
such as the River Vecht. In the third phase
(Fig. 2), at the end of the 19th century, lakes and
marshes had been formed by peat extraction for
fuel since the Middle Ages and some (Horster-
meer and Bethune area) had been reclaimed as
deep polders. The GooijUtrecht Hills groundwa-
IJSSEL-
MEER
Fig. 2. Situation of the Loosdrecht Lakes with respect to the
most important hydrological sybsystems in the region (After
Engelen, 1986).
25
ter flow system extended again with additional
flow branches towards the deep polders. More-
over the lakes had become the recharge areas for
new groundwater flow systems by leakage to-
wards the deep polders.
In the fourth phase, especially since the middle
of the 20 th century, the natural outflow of the
GooijUtrecht Hills groundwater flow system has
decreased considerably because of groundwater
extraction and has been polluted by urbanisation
and agriculture (Witmer, 1989; Schot, 1991). In
the summer the intake of strongly polluted water
from the river Vecht was necessary to maintain
the water level of the lakes. Thus the losses due
to evapotranspiration decreased groundwater in-
flow and the heavy leakage losses to the nearby
deep polders were compensated. The eutrophica-
tion of the lakes was the result. Since 1984 the
polluted supply from the Vecht has been replaced
by supply from the Amsterdam-Rhine canal with
P-removal before entering the Loosdrecht Lakes
complex. Figs. 3-6 show in flow charts the struc-
ture of the interaction of the various surface- and
groundwater flow systems of the Loosdrecht
Lakes area from the 15 th century to the present.
Surface water flow systems
The relative location, form, depth and size of the
individual lakes in the lake complex and their
hydraulic connections playa very important role
in the physico-chemical and biological behaviour
of the lakes (Fig. 7). All lakes have rather steady
leakage losses via a 'cascade' of coupled surface
water and groundwater systems to the adjacent
lower polders throughout the year (Fig. 8).
Roughly during the winter half year, however,
there is a water surplus due to decreased evapo-
transpiration and increased seepage from the
GooijUtrecht Hills catchment. The excess water
is pumped out to the River Vecht. The simulated
flow pattern during this period is given in Fig. 9.
In the summer half year the lake complex has
a more or less negative water balance depending
upon the meteorological conditions, requiring
supply from outside to maintain the lake levels.
26
Recharge area Gool
~ I
oligotrophic
Looadrecht Lakea
16 th century - mid 19th century
~
, ~
{7
Fig. 3. Water flow in the Loosdrecht Lakes area: 15 th century - mid 19th century.
The simulated flow pattern under those condi-
tions is given in Fig. 10. The lakes Vuntus and
Breukeleveen have a net inflow from the Loos-
drecht Lakes in summer and are at the end of the
surface water part of the cascade in the leakage
route (Fig. 8). Thus biomass and detritus gener-
ated in the Loosdrecht Lakes is swept by lateral
transport into those lakes and is trapped there.
This may explain in part the consistently higher
concentration of sestonic matter in these com-
partments of the lake complex.
Four mechanisms are responsible for the dy-
namics of the surface water in the Loosdrecht
Lakes:
- the leakage losses through the lake bottoms to
the adjacent polder areas. The areal pattern of
the losses differs strongly within the lakes and
varies from zero in the central western part of
the lake complex to locally more than 15 mm
day - I in a narrow strip in the southern part of
the Loosdrecht Lakes and Lake Breukeleveen,
adjacent to the Polder Bethune;
- the water level control of the lake complex by
...
Vecht ,
unpolluted
aurface water flow . .
groundwater flow S>
pumping of excess water or intake of Amster-
dam-Rhine canal water during dry and warm
periods;
- the influence of wind in setting up seiches and
currents (Fig. 11);
- the influence of differences in evapotranspira-
tion between compartments of the lake com-
plex.
The interaction of those four factors differs per
season and with the meteorological conditions.
Despite the hydrochemical differences in the
various inputs the rather homogeneous water
quality in the Lake Loosdrecht - often differing
from the qualities in the more secluded parts Lake
Breukeleveen, Lake Vuntus and Kievitsbuurt area
- is remarkable. This homogenisation is the result
of wind-induced flow and turbulence; and water
management-induced flow and turbulence.
The average residence time of the water in the
lakes varies from 255 days for the Lake Loos-
drecht through 220 days for Lake Vuntus to 110
days for Lake Breukeleveen. The residence time
of the water in the lakes differs depending on their

water
Recharge area Gool
Horster- oligotrophic
meer Loosdrecht Lakes
polder
situation till 1920
1883
Mljdrecht
polder
1880
Fig. 4. Water flow in the Loosdrecht Lakes area: mid 19 th century - 1920.
position in the network ofinteracting surface- and
groundwater systems. According to the 180_2H
ratio Lake Vuntus has the highest evaporative
influence and consequently the longest residence
time (Hettling, 1985a). This agrees very well with
a net inflow through one narrow inlet channel
from the Lake Loosdrecht (Fisser, 1986) to com-
pensate for the leakage losses towards the distant
Polder Horstermeer according to the groundwater
modelling. The Lake Breukeleveen (Teunissen,
1985) has the second-highest residence time due
to better exchange with the Lake Loosdrecht
through several inlet channels and a higher
throughflow due to the very strong leakage losses
to the adjacent Polder Bethune. The good mixing
in the Lake Loosdrecht, also evident from its
rather uniform hydrochemistry, is equally found
in the isotopic composition. Resuspension of bot-
tom sediments often follows a daily cycle along
with a similar cycle in wind velocities. Resuspen-
sion becomes particularly important from hourly
average wind velocities above 5 m s - I (Gons &
Van Keulen, 1989). An increase of the depth of
27
Bethune
polder
1888
Vecht
unpolluted
surface water flow ....
groundwater flow =;:>
the NE parts of the lake complex to about 3.5 m
by dredging the Pleistocene sands in the lake bot-
tom is mentioned by those authors as a possible
means to reduce the resuspension and improve
the transparency of the water.
Groundwater flow systems
The groundwater flow systems in the area are the
result of the natural topography and geological
structure and the subsequent interference by man
(Engelen& Schot, 1989; Witmer, 1989). Figure 12
gives a schematic picture of the groundwater flow
systems in the study area and of the pattern of
their recharge and discharge areas. The major
regional flow system (UH-system) has its re-
charge area in the ice-pushed hills of the Utrecht/
Gooi area, east of the lakes and has upward seep-
ing flow branches in the project area in the eastern
part of the Loosdrecht area and in the Polder
Bethune. Superimposed are a small man-made
local flow system of the pumping station east of
28

Recharge area Gool

Horster- eutrophication
Bethune
meer Looadrecht Lakes
1960 -1984 polder
polder

Water tr sport canal

1930

water supply Mljdrecht

surface water flow ...
Amsterdam polder
groundwater flow c::::;>

Fig. 5. Water flow in the Loosdrecht Lakes area: 1950-1984.

the lake (not indicated in Fig. 12) and the local
Nieuw Loosdrecht (NL-) groundwater flow sys-
tem, due to the topographic configuration, east of
the lakes. The Loosdrecht Lakes have their own
groundwater flow system by leakage to surround-
ing lower polder areas, having separate perma-
nent flow branches to the deep Polder Horster-
meer to the north, the deep Polder Bethune to the
south and to the polders to the west. In summer
a temporary eastward flow branch develops when
groundwater levels drop along the eastern border
zone of the lakes. The decrease in influx of high-
quality deep groundwater to the lakes, by ground-
water extraction in the adjacent ice-pushed hills,
played an important role in the deterioration of
the water quality of the lakes. This was reinforced
by the pollution of the inflowing shallow ground-
water over the last decades.
However, the fundamental cause for the gene-
sis of the eutrophication problem in the lakes is
the need to supply water, that contained high con-
centrations of nutrients, during dry spells in the
summer season to maintain their levels because of
the heavy leakage losses via groundwater flows to
the adjacent deep polders. Table 1 shows the re-
sults of the modelling with MICROFEM.
Isotopes
Specific combinations of 18 0, deuterium and tri-
tium enable a distinction of three types of re-
charge water for the groundwater flow systems:
local precipitation water in the recharge area of
ice-pushed hills "t Gooi', waters from the Rhine
system in the Amsterdam-Rhine canal and the
Vecht river; and lake waters in the compartments
of the Loosdrecht lakes, subjected in a varying
degree to evaporation.
The distribution patterns of these isotopes
agree fairly well with other types of information
like the results of the groundwater modelling, the
hydrochemical patterns in the groundwater (Het-
ding, 1985b) and the degree of flushing of the
lake compartments. Near the inlets of Vecht
water, e.g. in the Kievitsbuurt Area, the deviating
 29

water
Recharge area Gool
supply
Gool
treatment

Horster- Bethune
Loosdrecht Lakes
meer
polder
polder 1984 - present

P-rem Water trans ort canal

1984
Vecht

Amsterdam-Rhine canal 1938

water supply
Amsterdam
L; lJdrecht
polder
------
surface water flow ...
groundwater flow =>
Fig. 6. Water flow in the Loosdrecht Lakes area: 1984-present.

Table 1. MICROFEM modelling results. Flows in mm d - I The groundwater flow systems in the lake area
for the major water balance units in terms of upward seepage
were conceived initially on the basis of a hydro-
( + ) and seepage losses ( - ) in mm d - 1 for the dry situation
(1) and for the wet situation (2) under three different scenar- geological and hydrochemical survey (Hettling,
io's: present situation, situation with Bethune Polder flooded, 1985b) and a primitive version of the groundwa-
and situation after dredging. ter flow model FLOSA (Kuhnel, 1985a). The
leakage from the lakes in the unpolluted situation
Lakes Present Lake Dredging
around the middle of the century could not be
situation Bethune scenario
scenario traced so well by hydrochemical means because
quality differences between the various sources of
2 2 2 water were not so great at that time. It appeared
that the penetration depth of the grondwater flow
dry wet dry wet dry wet
systems and their flow branches had been under-
Loosdrecht -0.9 -0.4 +0.1 0.0 -1.3 - 0.5 estimated once the isotope data became avail-
Breukeleveen -l.l -0.5 +0.3 + 0.1 -1.7 -0.3 able. The subsequent pollution of the lakes and
Tienhoven - 5.8 -0.2 + 0.4 + 0.2 - 5.8 - 0.2 their surroundings introduced much clearer hy-
Kievits Area - 3.6 -0.8 - 0.2 -0.3 - 3.6 -0.8 drochemical markers on the flow lines of the
Vuntus +0.2 -0.2 +0.2 -0.2 +0.4 -0.2
groundwater systems.
Loenderveen -0.2 -0.3 -0.1 -0.3 -0.1 -0.3
Polder
Bethune + 12.4 + 5.0 - 1.0 - 0.5 + 13.1 + 5.0
Water balances

i80-values indicate the yet insufficiently mixed The following percentages may give an indication
influx of supply water. of the relative orders of magnitude of the differ-
30
/I
Loenderveen
Vuntus
Fig. 7. Layout of the Loosdrecht Lakes complex (After Engelen et al., 1988).
ent items of the water balance in the period 1978-
1987. The external supply of water from the
Amsterdam-Rhine Canal varies on average be-
tween 20 % of the total input on an annual basis
to 90% during the summer season. The annual
input to the Loosdrecht Lakes from the catch-
ment area in the adjacent hills amounts to 40%.
The output of the system oflakes consists of 40%
through forced discharge by pumping in winter,
25 % through infiltration losses to the groundwa-
ter flow systems and 20% is return flow to the
upstream part of the catchment during the dry
summer periods.
Phosphorus balances
The results from the balance calculations for
total phosphorus indicate that the external
P-Ioad on the whole lake system amounts to
0.8 mg m - 2 d - 1 under the present (1984-1987)
conditions. The contributions from various
central
Loosdrecht
sources calculated for 1984-1987 are: 30% from
the catchment in the hills, 30 % from the intake of
water from Polder Bethune, 10% supplied by
Amsterdam-Rhine Canal water; 30% from leak-
age inflows through sluices, precipitation, water-
based tourism and unsewered housing.
The annual phosphorus balance, though inac-
curate, indicates that a netto flux of P of about
50% of the total P-input disappears through
downward seepage, or is adsorbed by the lake
sediments. However, monthly balances indicate
that there is phosphorus release from this stock
into the supernatant water during the summer
months in the order of 0.2 to 0.4 mg P m - 2 d - I.
Fig. 13 shows the contribution to the P-Ioad by
the supply water before and after the start of P-
removal. In 1983 the summer input of P from the
supply of Vechtwater is still dominant in the P-
balance. In 1986 the supply component has been
reduced greatly by the change to Amsterdam-
Rhine-Canal water. The inputs to the Loosdrecht
lakes are notably high in the winter half year due

Loo8drecht lake8
IL. Vuntu8!
eer
~
IPolder Groot Mijdrecht
'L
Fig. 8. Cascades of surface water- and groundwater flow systems in the Loosdrecht Lakes area due to water supply to compensate
for leakage losses.
to polluted surplus flow from the GooijUtrecht
Hills catchment. During the summer half-year the
upwelling P-enriched groundwater in the Bethune
Polder is largely used for public water supply by
Amsterdam Municipal Waterworks. In the win-
ter half-year part is discharged into the Lake
Loosdrecht.
Heavy metals in the lake sediments
The occurrence of heavy metals in the lake sed-
iments is of two types: (a) comparatively high but
localized concentrations of heavy metals bound
to rapid settlement of suspended sediments of the
River Vecht, dragged into the lakes by the previ-
ous water intake from the River Vecht during dry
spells in summer. The concentrations show fairly
steep decreases in the channels downstream of
the intake sluices (the point sources) over dis-
tances of 1 to 2 kilometers; (b) more evenly dis-
tributed lower concentrations over the entire lake
31
Gool/Utrecht hill8
recharge area
8urface water flow
groundwater flow
area due to diffuse sources and from the Vecht-
intake points.
The Vecht-water heavy metals are absorbed
chiefly on ferri-hydroxydes and expandable clays.
Therefore the heavy metals entering the lakes
through intake points did not spread farther than
2 km downstream of the intake points due to rapid
settling of the suspended river sediments and an
increase in pH due to the mixing of Vecht water
and lake water. The latter factor was limiting the
mobility of most metals.
Dispersed, suspended Fe reached the lakes in
considerable quantities but accumulated rather
rapidly in the sediment due to adsorption of phos-
phates depending upon the turbulence conditions.
In non-turbulent environments like the Kievits-
buurt area of the lake the highest Fe- and P-
concentrations were measured in the sediments.
The behaviour of Cd differed slightly from that
of the other heavy metals. A significant fraction
of the Cd in the Vechtwater remained in solution
at first as uncomplexed bivalent Cd and precipi-
32
Fig. 9. Simulated flow pattern in the Loosdrecht Lakes com-
plex with inflow from the eastern catchment area, leakage
losses and pumped discharge for the winter situation without
wind, in a physical model at the scale I: 10000. (After Wasser,
1988).
tated farther on in the lakes, mostly as carbonate.
The maximum value of about 70 ppm Cd (in dry
weight minus loss-on-ignition sample) in the
Kievits Area at the SW intake point is probably
due to the cumulative effects of intake at that
point of both Vecht water and of water from the
agriculturally-used Polder Bethune. In the period
1979-1984 Cd has been the metal whose concen-
tration in the sediment of the 'open' compart-
ments of the Lake Loosdrecht has most strongly
increased, viz. from 1.6 to 6.3 ppm. Fe follows in
second place in this respect. The concentrations
of the other heavy metals did not change signifi-
cantly over that period.
No data are available on the situation of heavy
metals in the sediment after the shift from Vecht-
water supply to dephosphatized Amsterdam-
Rhine Canal water. The point sources ofthe Vecht
Fig. 10. Simulated flow pattern in the Loosdrecht Lakes com-
plex with water supply from the Amsterdam-Rhine canal,
leakage losses and reversed flow to the eastern catchment area
for the summer situation without wind, in a physical model at
the scale I: 10000 . (After Wasser, 1988).
sluices have strongly decreased in strength since
only leakage of the sluices and water exchange
due to ships passing the locks remain here as
pollutant fluxes.
It may be assumed that the bulk of heavy metal
load in the water supplied from the Amsterdam-
Rhine canal will be trapped nowadays in the pro-
cess of phosphate removal by flocculation, copre-
cipitation and settling upon addition of iron
chloride. Consequently the total input of heavy
metals to the lakes has probably decreased
strongly.
Scenario's for water quality restoration
Remedial measures until 1984
The construction of sewer systems in the catch-
ment area of the lakes and discharge of the effiu-

Fig. II. Simulated flow pattern in the Loosdrecht Lakes com-
plex, induced by a southwestern wind in a physical model at
the scale 1:10000 in a wind tunnel. (After Wasser, 1988).
ent outside the catchment. The sewage load from
the permanent population in the villages around
the lakes has been removed largely by the instal-
lation of sewer systems, sewage treatment and
discharge of the effluent outside the catchment of
the lakes in the period 1970-1986. The sewage
load from heavy recreational use of the lakes dur-
ing summer has been reduced by the installation
of sanitary facilities. As a result of these measures
the P-input from sewage decreased from
2360 kg P y- 1 in 1980 to 613 kg P y- 1 in 1987.
Currently the P-loading due to remaining unsew-
ered discharges is about 13 /~ of the total exter-
nal P-load in summer and 6 % in winter.
Since 1984 additional water supply to the lakes
has changed from the intake through sluices of
highly polluted Vecht water to supply by pipeline
from the Amsterdam-Rhine canal via a P-removal
facility of Municipal Waterworks of the city of
33
Amsterdam. The annual external P-loading of the
lakes has decreased due to these two measures
considerably: from 1.2 to 0.4 g P m - 2 Y- 1. How-
ever, those measures have not resulted as yet in
a visible improvement of the water quality of the
lakes.
Optional restoration measures, removing the original
cause of eutrophication
The results of the balance studies indicated which
additional measures might be taken for a further
reduction of the P-load. The first category of rad-
ical solutions aims at the removal of the original
cause of the eutrophication problem: the negative
water balance of the lakes in the summer half year
which requires an additional supply to maintain
the lake levels. This could be achieved by: (1) re-
storing the original groundwater influx; and
(2) decreasing/terminating the leakage losses
from the lakes by flooding anew (parts of) the
empoldered former Lakes Horstermeer and es-
pecially the nearby Polder Bethune.
Drastic reductions in groundwater abstraction
in the Gooi catchment area of the lakes (Engelen
et al., 1989; Engelen & Schot, 1989; Witmer,
1989; Wassen, 1990; Schot, 1991). Reductions
will lead to increased seepage input of good qual-
ity and thus less input of water from Amsterdam-
Rhine Canal. The provincial government of North
Holland has initiated recently a policy to reduce
the groundwater abstraction in the northern part
of the Gooi/Utrecht Hills area. Such a policy is
still lacking in the province of Utrecht.
(Partial) flooding of deep polders (former lakes
or marshy areas) The possible hydrological ef-
fects of restoring the original hydrological situa-
tion by flooding Polder Bethune were considered
as an option in the groundwater modelling. The
upward seeping flow branch of the regional
groundwater flow system along the eastern bor-
der and within the central part of the Lake Loos-
drecht would be restored to a large degree. The
need for water supply in the summer half year to
compensate for the leakage losses would be re-
duced completely, because under wet conditions
 34

subregional upward
,--_r....~_loTca_1s..:..ys_teTm_La_k_eTL_oo_sd_'.,.eC~ht~,---se--:ep~a.::,ge_W_in_dO,W,--_,--lo_ca_1 sr-ys_te_m,-_,-_'-negiOnal system

=:::::::t:t::::::::::::::::I~:) 50 m

100 m

regional flow branch

UH - system

I:EI<=========================E8nk<rmn::================================~) I 150 m

Fig. 12. SW-NE cross section (from the city of Hilversum to the Polder Bethune) of the groundwater flow systems in the
Loosdrecht Lakes area with the two-dimensional model FLOWNET. (Van Pruissen, unpublished). Note the vertical distortion
by the difference in horizontal and vertical scales.

(Table I). The results of the groundwater model-

Loosdrecht Lakes ling seem odd, because the infiltration (or 'leak-
P-Ioad by water supply age') losses in Table I are generally lower under
P-Ioad, kg/day
140 -'-,_--'--.C.---'- ---.--------------
wet than under dry conditions. This can be ex-
I plained as follows: the fluxes are related to the
120r
differences in hydraulic head between the ground-
100~"
water in the recharge area in the hills, the water
80-1 .. ·.. ·· .. ·...... ···· .. ·· •
level of the lake complex and the water levels in
80+···· .. ·················•
the adjacent polders. Under the dry (summer)
40+ .. ·.. ·· .. ·.. ······ .. ···· •
conditions lake levels are kept more stable by
2: :~~~~-~;~~_~~f-"I-'l~· . . f -"I-r-,-..,.-r-1--\"-' ~1'11"-'t~'---r---,-.,-,--,r' supply, whereas the phreatic groundwater levels
in the polder areas decrease more. Thus the avail-
1 2 3 4 6 8 7 8 9 10 ft n 1 2 3 4 6 8 7 8 9 10 ft n
able head difference to drive the groundwater flow
_ 1983. Vachl aupply I22a 1988. ARC aupply is larger under dry conditions.
f,o.. Enee'e. et a'. teee

Fig. 13. Comparison of the P-load by water supply for the

Loosdrecht lakes for the years 1983 (supply from polluted Optional restoration measures, attacking the inter-
Vecht river) and 1986 (supply from Amsterdam- Rhine Canal). nal and external P-Ioads

The second, less drastic, category of solutions

there would be a zero net flow for the groundwa- aims at mitigating/removing the secondary main
ter component and under dry conditions there factor causing the eutrophication: the external
would be even a small upward seepage excess and internal P-Ioads on the lake complex. Besides

the already executed measures, further measures
along this line could be: reducing the internal
P-Ioading by dredging of the P-enriched bottom
sediments; reinforced flushing of the lakes during
summer with dephosphorized water to deplete
the stock of P in the lake complex: intercepting
and rerouting polluted inflows from the water-
shed; and construction of additional P-removal
facilities to treat yet untreated inflows.
Dredging of the lake system
Part of the phosphorus stock of the lake system
is stored in the loose bottom sediments and partly
released to the water in summer. Removal of this
amount by dredging the loosely packed upper part
of the bottom sediments with their high organic
content has been considered. The hydrological
consequence would be increased leakage from the
lakes towards the adjacent lowlying polders, es-
pecially from the southern edges of the lake com-
plex to the deep nearby Polder Bethune. Two-
dimensional groundwater modelling (Van
Pruisen, unpublished) showed that actual leakage
losses along the southern edge of Lake Breu-
keleveen amount to 5 mm d - I and in the south-
ern border zone of the Lake Loosdrecht up to
15 mm d - I. Groundwater modelling for the
dredging scenario even shows an increase in av-
erage upward seepage for the Polder Bethune
from 5 to 13,1 mm d - 1 (Table 1). So dredging
would have most effect on the water balance there
and only minor effects on the leakage elsewhere
in the lake complex due to the topographic con-
figuration of the lakes in relation to the deep pold-
ers in the surroundings. The additional leakage
losses due to dredging would have to be compen-
sated by additional intake of water from the
Amsterdam-Rhine canal. Notwithstanding the
phospate removal from this source of supply it
still contains some P and consequently the exter-
nal P-load on the lakes would increase slightly
during supply in summer periods, counteracting
slightly the beneficial effects of dredging. The re-
sults from semi-3D groundwater modelling with
MICROFEM (Van Pruisen, unpublished) are
35
that the average leakage from the Loosdrecht
lakes, when dredging the entire lake bottom,
would increase under dry conditions from 0.9 to
1.3 mm d - 1 (Table 1). If the southern border
would be left undregded this increase would be
considerably less and the same would hold for the
required additional water supply.The increased
permeability of a dredged lake bottom would
probably diminish again with time due to con-
tinuing sedimentation of detritus. However, no
data are available about the intensity of such a
process.
Flushing in summer
Another management option to clean up the
Loosdrecht Lakes could be a scheme of reinforced
flushing of the lakes by supplying de-phospho-
rised water in considerable excess of the amounts
needed to maintain the water levels of the lake.
The flushed algal biomass would be, however, an
additional burden to the receiving water system
outside the Loosdrecht area (the River Vecht
flowing into Lake Gooi) and thus problems might
just be shifted. The flushing studies indicate that
there are bottlenecks by the present lay-out of the
technical infrastructure and capacity of intake-
and discharge structures, treatment plant and the
narrow limits of allowable surface water levels.
A flushing flux of 10 x 106 m - 3 month - I during
the summer period would be the order of magni-
tude required to expect a notable reduction in
P-concentration in the water.
Interception and rerouting ofpolluted inflows
The P-influx from the catchment in the hills (es-
pecially during the winter period) amounts to 30 %
of the present load. Interception and deviation of
this polluted surface water inflow in winter is
technically possible by adaptation of the physical
water management infra-structure. The provin-
cial government of Utrecht commissioned a tech-
nical study to elaborate this option resulting from
the water balance studies.
36
Construction of additional P-removal facilities
Instead of directing the polluted influx from the
hills of the catchment in winter directly to the
River Vecht, this polluted water could be col-
lected and processed for P-removal and then dis-
charged into the Loosdrecht Lakes. In addition,
the input of 30% of the external P-Ioad from the
intake of water from the Polder Bethune could
also be largely eliminated if an additional treat-
ment plant for phosphorus removal were built or
if these waters were routed through the existing
treatment plant before discharge into the lake
complex. It would require additional technical in-
frastructure works and adaptation of the water
extraction schemes of the Amsterdam Municipal
Waterworks, which uses part of the excess seep-
age waters of the Polder Bethune.
The last major balance item of 30 % of the
external P-load consists of a combination ofloads
from various sources of a distributed nature.
These are very difficult to tackle by systematic
collection- or treatment schemes. The 10% load
from the supply water could be reduced further by
technical improvements in removing phosphorus
from the supply water in the treatment plant.
However, the very large quantity of needed sup-
ply water produces - even at the current, already
fairly high level of phosphate removal in the treat-
ment plant - still about 10% of the total present
load.
Discussion
The scenario's have different social and political
feasibilities, financial and technical boundary
conditions, time spans to become effective, and
net effects on the solution of the eutrophica-
tion problem. The proposed rating is rather
approximate because systematic and complete
feasibility studies of the scenario's which take
into account net effect on P load, technical/
financial constraints and social/political feasi-
bility have not been made as yet: (1) addi-
tional P-removal scenario; (2) deviation scenario;
(3) flushing scenario; (4) dredging scenario;
(5) Bethune Lake scenario; (6) Combination sce-
nario (Fig. 14).
The full additional P-removal scenario would
remove a considerable portion (60 %) of the re-
maining external P-Ioad and constitute no addi-
tional P-burden on the receiving waters as the
deviation scenario would do. The flushing sce-
nario requires major adaptations in the water
management infra-structure (pumping capacity
and relocation of inlets and outlets). It would
tackle the internal P-Ioad of the lakes by deplet-
ing the stock of P in the bottom sediments and
flushing out biomass. During a transitional phase
it might be the necessary trigger to overcome a for
the biological systems critical boundary value of
e.g. light penetration, necessary for the restora-
tion of meso- to oligotrophic conditions and sub-
merged vegetations. It would loose its significance
once the lake water quality should be restored.
The partial dredging scenario would affect the
internal P-Ioad by removal of part of the
P-enriched bottom sediments and improve visi-
bility by a decrease in suspended sediment con-
centrations. Costs and disposal of the dredged
sediments form major obstacles. Dredging of the
lakes would be expensive and technically difficult
by stirring the mud due to the very low cohesion
of the upper layer of the bottom sediments. Proper
disposal or use of the huge quantities of dredged
material would constitute another practical and
environmental problem to be solved. However
from a hydrological point of view no major ob-
jections against dredging should be raised. The
Bethune Lake scenario would be a very funda-
mental and thorough solution, but difficult to re-
alize. It would require substantial physical plan-
ning efforts, evacuation and relocation of the
sparse agricultural population of the polder and
the cooperation of the Amsterdam Municipal
Waterworks to relinquish its rights to the excess
seepage water of the present Polder Bethune. The
advantages could be great however in terms of a
fundamental solution to the eutrophication prob-
lem, improved water supply to the nature protec-
tion areas along the Gooi/Utrecht Hills and the
extension of part of the crowded recreational fa-
cilities to the new lake.
 37

sewage~

Recharge area Gool ~nt

~tlon

Horster- Loosdrecht Lakes

meer Combination
polder

P-rem Flushing

Vecht

water supply MIJdrecht surface water flow ...

Amsterdam polder
groundwater flow c::;>

Fig. 14. Water flow and remedial measures in the Loosdrecht Lakes area in a comprehensive combination scenario.

Finally the combination scenario contains all
scenario's as an overview and could be the end
result of successive steps in the process of resto-
ration of the Loosdrecht lakes area by measures
affecting the P-load. P-removal as a single ap-
proach still remains a far cry from a return to a
complete restoration of the original ecological sit-
uation because of the considerable remaining
input of other components to the lake complex in
the supply water, e.g. the continuing overloading
of the lake complex with other nutrients like ni-
trogen, dissolved organic compounds, etc.
References
Loosdrecht lakes, in Dutch) WQL-report 1988-7. Free
University, Institute of Earth Sciences, Amsterdam, 71 pp.
Dekker, A. G., T. J. Malthus, S. W. M. Peters & E. Seyhan,
1991. The remote sensing Loosdrecht Lakes project. BeRS
Report no. 90-29. Free University, Institute of Earth Sci-
ences, Amsterdam, 105 pp.
Engelen, G. B., 1986. Interaction of hydrological systems and
eutrophication of the Loosdrecht Lakes. Hydrobio!. Bull.
20: 17-25.
Engelen, G. B. & G. P. Jones (eds), 1986. Developments in
the analysis of groundwater flow systems. IAHS Pub!.
no. 163,356 pp.
Engelen, G. B. & B. F. M. Kal, 1985. Hydrologisch onder-
zoek in WOL kader van het Loosdrechtse Plassengebied in
verband met beleid en beheer voor waterkwaliteitsherstel
(Hydrological research in the Loosdrecht lake area in re-
lation to policy and watermanagement for lake restoration,
in Dutch). WQL-report 1985-1. Free University, Institute
of Earth Sciences, Amsterdam, 37 pp.
Engelen,G. B. B. F. M. KaI,LJ .. Bllyse,A. Wakjira & S. O.
T

Buyse, J. J., 1986. Turbulente diffusie in de Loosdrechtse
Plassen (Turbulent diffusion in Loosdrecht lakes, in Dutch).
WQL-report 1986-12. Free University, Institute of Earth
Sciences, Amsterdam, 42 pp.
Buyse, J. J., 1989. Water- en stotbalansmodellen voor de
Loosdrechtse Plassen (Water- and matter balances in
Los, 1988. Waterhuishoudkundige beleids- en beheers-
mogelijkheden in verband met waterkwaliteitsherstel van
het Loosdrechtse Plassen-system (Water management and
policy in relation to Loosdrecht lakes' restoration, in
Dutch). WQL-report 1988-1. Free University, Institute of
Earth Sciences, Amsterdam, 43 pp.
38
Enge!en, G. B. & P. P. Schot, 1989. Analyse van de water-
systemen in het GooisjUtrechts stuwwallengebied, de
Vechtstreek en het Eemdal: probleemsignalering (Analysis
of watersystems in the GooijUtrecht sandridge area; the
Vecht area and the Valley of the Eem, defining the problem,
in Dutch). CHOjTNO Rapporten en Nota's 22: 21-31.
Fissel', C. H. B., 1986. Hydrogeografisch vooronderzoek van
de Vuntus (Preliminary research of Lake Vuntus. WQL-
report 1986-5. Free University, Institute of Earth Sciences,
Amsterdam, 89 pp.
Gons, H. J. & R. Van Keulen, 1989. De relatie tussen door-
zicht en slib in the Loosdrechtse Plassen in verband met de
zwemwaternorm: afsluitend onderzoek (Relation between
water cl-arity and suspended matter in relation to swim-
mingwater standards: final report, in Dutch) WQL-report
1989-4. Limnological Institute, Nieuwersluis. 35 pp.
Hettling, H. K., 1985a. Isotopic groundwater study of the
Loosdrecht Lakes area (the Netherlands). WQL-report
1985-4. Free University, Institute of Earth Sciences, Am-
sterdam, 55 pp.
Hettling, H. K., 1985b. Een hydrogeologische inventarisatie
van het Loosdrechtse Plassengebied (A hydrogeological in-
ventarisation of the Loosdrecht lakes area, in Dutch).
WQL-report 1985-2, Free University, Institute of Earth
Sciences, Amsterdam, 116 pp.
Higler, L. W. G., S. Parma & L. Van Liere (eds), 1987. Bag-
geren in de Loosdrechtse Plassen (Dredging in Loosdrecht
lakes, in Dutch). WQL-report 1987-5, National Inst. of
Nature Protection, Leersum. 18 pp.
Janse, J. H. & T. Aldenberg, 1990. Modelling phosphorus
fluxes in the hypertrophic Loosdrecht lakes. Hydrobiol.
Bull. 24: 69-89.
Kal, B. F. M., G. B. Engelen & Th. E. Cappenberg, 1984. The
Loosdrecht lakes restoration project: hydrology and phys-
ico-chemical characteristics of the lakes. Verh. int. Ver.
Limnol. 22: 835-841.
Kal, B. F. M., 1986. Monthly mass balances for compart-
ments of the Loosdrecht Lakes system: approach and pre-
liminary results. Hydrobiol. Bull. 20: 27-39.
KUhnel, V., 1985a. Toepassing van het grondwaterstromings-
model FLOSA binnen het Loosdrechtse Plassen restau-
ratie-project (The application of the groundwaterflowmodel
FLOSA in the Loosdrecht lakes restoration project, in
Dutch). WQL-report 1985-3. Free University, Institute of
Earth Sciences, Amsterdam, 66 pp.
KUhne!, V., 1985b. Zware metalen in het Loosdrechtse
Plassen-sediment (Heavy metals in Loosdrecht lakes' sed-
iments, in Dutch). WQL-report 1985-6. Free University,
Institute of Earth Sciences, Amsterdam, 36 pp.
Schot, P., 1991. Solute transport by groundwater flow to wet-
land ecosystems. Ph.D. thesis. State University at Utrecht,
134 p.
Van Pruissen, F. G. M., 1989. Modelleren met 'MICRO-
FEM'. H 2 0 22: 618-623.
Van Del' Spek, T. R., 1986. Onderzoek naar de mogelijkheid
van het simuleren van oppervlaktewater-stromingen in het
Loosdrechtse Plassen-gebied met behulp van een Hele-
Shaw model (Research to the possibilities of the simulation
of surface-currents in the Loosdrecht lakes area by means
of a He!e-Shaw model, in Dutch). WQL-report 1986-4.
Free University, Institute of Earth Sciences, Amsterdam,
61 pp.
Teunissen, A. c., 1985. De Breukeleveense Plas. Een hydro-
geografisch vooronderzoek (Lake Breukeleveen. A hydro-
geographic study, in Dutch). WQL-report 1985-6. Free
University, Institute of Earth Sciences, Amsterdam, 103 pp.
Wassen, M. J., 1990. Water flow as a major landscape eco-
logical factor in fen development. Ph.D. thesis, State Uni-
versity at Utrecht, 199 p.
Wasser, H. J., 1988. Fysisch-model onderzoek naar door
wind aangedreven stroming in het Loosdrechtse Plassen
gebied (Physical model research to wind-driven currents in
Loosdrecht lakes, in Dutch). WQL-report 1986-6. Free
University, Institute of Earth Sciences, Amsterdam, 27 pp.
Witmer, M. C. H., 1989. Integral water management at re-
gional level. An environmental study of the Gooi and the
Vechtstreek. Ph.D. thesis, State University at Utrecht,
172 pp.
Hydrobiologia 233: 39-50, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 39
© 1992 Kluwer Academic Publishers.

Phosphorus in the sediment of the Loosdrecht lakes and its implications

for lake restoration perspectives

Peer Keizer & Anja J, C. Sinke

Limnological Institute, Rijksstraatweg 6,3631 AC Nieuwersluis, The Netherlands

Key words: phosphate, iron hydroxide, dredging, downward seepage, mineralization, eutrophication

Abstract

In 1984 the external phosphorus load of the shallow eutrophic Loosdrecht lakes was reduced from 3,3
to 1.0 mg m - 2 d - 1, The effect of phosphorus release from the sediment on lake restoration was inves-
tigated. Diffusive release under aerobic conditions (20 0 C) decreased from 1 mg m - 2 d - 1 in 1984 to
0.3 mg m - 2 d - 1 in 1990, The generation of inorganic phosphorus due to mineralization during summer
equals 3 mg m - 2 d - 1, which is much higher than the measured rate of diffusive release, Despite that,
the phosphorus release is hardly stimulated by anaerobic conditions, which indicates that only a small
amount of phosphorus is adsorbed by ferric iron in the top sediment layer. This apparent discrepancy
is probably caused by the uptake of inorganic phosphorus uptake during resuspension and the loss of
inorganic phosphorus with downward seepage.
The estimated removal of phosphorus due to downward seepage of 0,8 mg m - 2 d - 1 agrees well with
the average phosphorus retention in the lake. This indicates that sediment burial and diagenesis are
unimportant mechanisms for withdrawing phosphorus from the nutrient cycle.
Between 1982 and 1991 the total phosphorus content of the upper 2 cm of the sediment decreased
from 0.94 to 0.60 g kg- 1 DW. At present, about 20% of total phosphorus in this layer is potentially
bioavailable, but largely incorporated in easily degradable organic matter. This pool is much smaller in
deeper layers, Based on the estimated and measured rates and pool sizes, the annual average phosphorus
cycle in the lakes was modelled to evaluate the effects of various restoration measures, The main pre-
dictions of the model are: I) further reduction of the external load may cause a gradual decrease of the
total phosphorus concentration in the lake water; 2) dredging and iron addition, without reduction of
the external load, may give a rapid improvement followed by a slow return to the present situation; and
3) reduction of the external load, combined with a cut off of downward seepage will not improve the water
quality.

Introduction management measures to improve water quality

Most surface waters in the Netherlands are eu-
trophic or even hypereutrophic, having total
phosphorus concentrations exceeding 0.15 g m - 3
(Lijklema et al., 1988). During the last decade,
have been focussed mainly on a reduction of the
external phosphorus load. The response of a lake
to a reduction of the phosphorus load depends on
the recycling of phosphorus from the sediment
(Marsden, 1989; Sas, 1989), The phosphorus re-
40

lease from the sediment, i.e. internal loading, is past centuries (Van Liere, 1986). Because of con-
identified as a major drawback for lake recovery tinuous wind-induced turbulence (Gons et al.,
(Ryding, 1985; Cullen & Forsberg, 1988). 1986), the surface sediment remains aerobic
The diffusive release of phosphorus from the throughout the year. General lake characteristics
sediment, measured in columns under standard- are presented in Table 1.
ized conditions, has been thoroughly investigated The sediment has an extremely high organic
(Holdren & Armstrong, 1980; Boers & van Hese, matter content, exceeding 50 % on a dry weight
1988). Release rates are stimulated by high tem- basis (Boers et al., 1984). The lakes receive water
perature (Kamp-Nielsen, 1974; Sinke & Cappen- with a relatively high alkalinity from the surround-
berg, 1988) and a low redox potential (Mortimer, ing area and the sediment contains 3-6~~ CaC0 3
1941; Tessenow, 1972; Redshaw et al., 1990). In despite its peaty origin. The pH of the lake water
shallow lakes resuspension of the upper layer of varies around 8.4.
the sediment can be especially important for in- The hydrological setting of the lakes induces an
ternal loading (Ryding & Forsberg, 1977; Sem- average downward seepage of 1.5 mm d - I (En-
dergaard et al., in press). To predict the release gelen et al., 1992). Therefore, approximately
rates of phosphorus, several workers used the 6.10 6 m 3 of supplemental water is needed during
total phosphorus content of the sediment (NOrn- summer to maintain the water table at a fixed
berg, 1988; Sas, 1989). However, it is likely that level. The high phosphorus concentration of this
the remobilization of phosphorus from the sedi- water has largely contributed to a serious deteri-
ment is determined by its speciation. Especially oration in the water quality during the past de-
the loosely adsorbed fraction (S0ndergaard, 1989) cades (Van Liere et al., 1989). Since 1984 the
and the iron bound fraction (Hosomi et al., 1982; supplemental water has been dephosphorized.
NOrnberg, 1988; Shaw & Prepas, 1990) are sub- This measure reduced the external phosphorus
ject to mobilization. Bostrom et al. (1985) showed
that organic bound phosphorus was the most dy-
Table I. General characteristics of the Loosdrecht lakes and
namic fraction in the sediment of shallow eutro-
data on the composition of the water column and the sediment.
phic lake Vallentunasjon.
In this paper we evaluate the role of the sedi- Lake area 14.5 km 2
ment in the phosphorus cycle of a peaty eutrophic Average depth 1.8 m
Water inflow and precipitation 2.3 m 3 m- 2 y-l
system. Extraction procedures are introduced to
Water outflow (surface) 1.3 m 3 m- 2 y-l
estimate the potentially bioavailable pools of or- 0.65m 3 m- 2 y-1
Water outflow (seepage)
ganic and inorganic phosphorus in the sediment. Average external P load 1.0 mg m - 2 d - I
Results are used in a simple descriptive model for Average P discharge 0.4 mgm- 2 d- '
evaluating the impact of additional restoration
measures. The different scenarios presented con- LAKE WATER (1990) (N= 13)
Total phosphorus 75 (±22) mg m- 3
cern: a) further reduction of the external phos- 2 (±3) mg m- 3
Soluble reactive phosphorus
phorus load; b) sediment removal; c) Fe(III) ad- Chlorophyll a 93 (± 18) mg m - 3
dition to the sediment; and d) reduction of pH 8.4 (± 0.3)
downward seepage by changing the hydrology of
the lake area. LAKE SEDIMENT 0-10 cm (1989) (N= 8)
Dry weight 9.0 (± 1.0) %
Organic carbon content 53.6 (± 3.0) %
Iron (total) 20.3 (± 2.0) g kg - I
Lake description Iron (acid extractable) 11.3 (± 1. 9) g kg - I
Phosphorus total 0.67 (± 0.08) g kg- I

The Loosdrecht lakes are shallow, interconnected Phosphorus (acid extractable) 0.15 (± 0.02) g kg- I

Cacium (total) 40.2 (± 3.4) g kg- I

lakes which are partly surrounded by polders.
CaC0 3 46.1 (± 13.8) g kg- I
They were created by peat excavation during the

load from 3.3 to 1 mg m - Z d - 1. However, the
rate of decrease in the phosphorus concentration
of the lake of ca. 12 mg m - 3 y- I is less pro-
nounced and amounts to < 5% per year (Van
Liere et al., 1990). The lake is still highly eutro-
phic and in summer the chlorophyll a level often
exceeds 100 mg m - 3. Because the average Sechhi
depth of the Loosdrecht lakes is still limited to
0.3 m, application of additional restoration mea-
sures is considered desirable.
Materials and methods
Phosphorus release and uptake experiments
Sediment cores were sampled at the centre of the
Loosdrecht lakes with a stainless steel sampler
having a perspex inner tube of 0.40 m length and
a diameter of 0.05 m. After sampling, intact sedi-
ment cores were transferred immediately to the
laboratory. The overlying lake water was si-
phoned off and replaced with artificial lake water
(ALW= 1.0mM Caz +, 2mM Na+, 0.2mM
Mg z +, 0.2 mM SO~-, 2.4 mM HC0 3- and
1.6 mM CI -; pH 8.3). The water level was made
up to 15 cm and flushed with air for the aerobic
incubations, or with a mixture of 99.7% N z and
0.3 % CO z for the anaerobic incubations. The
CO z was added to maintain the pH in the over-
lying water at 8.0 ± 0.3.
Aerobic phosphorus release was determined
monthly in cores sampled in the period 1983-
1991 according to Boers & van Hese (1988). The
overlying water was replaced continuously with
ALWand filtered over 0.45 J-lm membrane filters
and acidified to pH = 1 before analyses. Anaero-
bic release of phosphorus was measured in 1990.
To prevent contamination with oxygen, sampling
was performed through needles in the rubber
stoppers of the columns, using a peristaltic pump.
The samples were acidified before exposure to the
air in order to prevent oxidation of dissolved fer-
rous iron and subsequent precipitation of iron
hydroxyphosphate.
The uptake of phosphorus by the sediment was
measured under aerobic and anaerobic condi-
41
tions. 10 ml of 1 mM NaH zP0 4 was added to the
overlying water (final volume 290 ml) of the aer-
obic cores at time zero and once again seven days
later. The anaerobic cores received only one ad-
dition after 20 days incubation. One week later,
the remaining overlying water was removed as
much as possible and the water level made up to
a height of 15 cm with ALW. Thereafter, the water
was sampled by replacing half of the overlying
water with fresh ALW once or twice a week
(Keizer et al., 1991).
The effect of iron addition on the phosphorus
uptake was measured in anaerobic cores to which
10 ml of 0.01 M freshly-prepared Fe(OHh
(2.8 g Fe m - Z) was added prior to the phospho-
rus addition. After settling, the precipitate was
gently mixed with the upper 5 mm of the sedi-
ment. The Fe(OHh suspension was prepared by
titrating a known volume of 0.2 M FeCI 3 in 2 h to
pH 7 with 0.6 M NaOH, and diluted 10 times
with ALW just before use.
Sediment analysis and phosphorus fractionation
Total phosphorus and iron contents of the top
layer (0-2 cm) of the lake sediment were deter-
mined in cores sampled at four different locations
in 1982, 1985 and 1991. After determining dry
weight (24 h at 105 ° C) and loss on ignition (4 h
at 550 ° C), the remaining ash was digested for
24 h with a 3: 1 mixture of concentrated HCI and
HN03 at 150°C. After filtration (0.45 jtm}phos-
phorus and iron in the samples were analyzed.
Total inorganic phosphorus in the sediment
was determined by a 24 h extraction with 0.24 N
H zS0 4 (1 % wjw) of slices of the upper 10 cm of
the sediment. After that, the slurry was filtered
(0.45 J-lm) and reactive phosphorus was measured
in the filtrate. In mixtures of sediment and iron
hydroxyphosphate, this extraction method gave a
> 95 % recovery of the added iron and phospho-
rus.
To distinguish between adsorbed phosphorus
and phosphorus included in minerals, 0.15 %
w jw sediment suspension in ALW were incubated
for two months in an anaerobic glove box (Coy
42
laboratories, Ann Arbor, Michigan) filled with 2 %
H 2 , 0.3% CO 2 and 97.7% N 2 . During this incu-
bation the Fe(OH)3 in the sediment is reduced
completely and all adsorbed phosphorus is des-
orbed.
Phosphorus in degradable organic matter was
determined by aerobic incubation of the sediment
from 0-2 cm and 8-10 cm depth, suspended in
ALW (0.5% wjw). To ensure the presence of
aerobic microorganisms the sediment from the
deeper layer was inoculated with material from
0-2 cm (0.01 % wjw). The suspensions were
shaken continuously in the dark at 22 ± 4 C in
0
polyethylene bottles. The amount of oxygen in the
headspace was sufficient to maintain aerobic con-
ditions between the sampling periods. Homoge-
neous subsamples were taken after various incu-
bation times, diluted 1:1 with 0.48 N H 2 S0 4 and
filtered over 0.45 Jim 24 hours later. Phosphorus
and iron were determined on a SKALAR au-
toanalyser according to Murphy & Riley (1973)
and Fries & Getrost (1977) respectively.
Results
After the reduction of the external phosphorus
load in the lakes in 1984, the average diffusive
phosphorus release rate, measured in aerobic
cores for four weeks, decreased from 0.5-
1.5mgm- 2 d- 1 to 0-0.5mgm- 2 d- 1 (Fig. 1).
In 1990 the average aerobic release measured
during a three months period was less than
0.2 mg m - 2 d - 1 (Fig. 2A). After addition of
phosphorus to the overlying water, an uptake of
phosphorus up to 200 mg m - 2 occurred. How-
ever, this enrichment of the cores with phospho-
rus did not increase the aerobic release rate. After
a shift to anaerobic conditions however, the phos-
phorus was returned to the overlying water al-
most completely in about two months. Under
anaerobic conditions, the average release rate of
phosphorus during the incubation was
0.6 mg m - 2 d - 1 (Fig. 2B). The anaerobic incu-
bation of the sediment before the phosphorus ad-
dition largely reduced the uptake capacity of the
sediment. This provides indirect evidence that in-
5
I"
u
N
I
4
E
CJ)
3
E
Q)
~ 2
Q)
Ul
0
Q)
Q)
'-
(L
1983 1984 1985 1986 1987 1988 1989 1990
Fig. 1. Diffusive phosphorus release rate (four week average)
from intact sediment cores of the Loosdrecht lakes sampled
between 1983-1991.
cubation under anaerobic conditions for three
weeks is sufficient to reduce the Fe(III) in the top
layer of the sediment almost completely. Addition
of 2.8 g m - 2 Fe as Fe(OHh to the top layer of
anaerobic sediment cores resulted in a fast initial
phosphorus uptake of 150 mg m - 2. Nearly all the
adsorbed phosphorus was released over the next
two months under anaerobic conditions. From
this experiment the reduction of Fe(OHh was
estimated at 0.05 g m - 2 d - I.
The composition of the upper 10 cm of the
sediment was determined in cores sampled at the
centre of the lake in 1989 (Table 1). Dry weight
increased from 7 to 10% with depth. For the
other variables no significant trends with depth
were observed. Total inorganic phosphorus, ex-
tracted with 0.24 N H 2 S04 , was 0.15 ± 0.02 g
kg- lOW. The difference between total phospho-
rus and the acid extractable fraction is considered
to be incorporated in organic matter and bio-
mass. Acid extractable iron equals 11 ± 2 g kg - I,
almost 60% of the total iron content. The non-
extractable iron is included in relatively stable
minerals and probably of minor importance for
the chemical equilibria in the sediment.
From the deeper sediment only a small amount
of phosphorus was mobilized under anaerobic
conditions (Fig. 3). Upon anaerobic incubation
of diluted slurries (1.5 g 1- lOW) of mixed sedi-
ment from 1-10 cm depth, the phosphorus con-
 43

100 40 .06
A V1

-
;u

N
, -"
.05 (")
I
E a CJ)
.>:
30 0

-
Z
CJ) 0 .04 (")

-
0>
E E
r'1
z
--<
w 20 .03 ;u
0-
V1 :to-
<! o :::!

--
w w 0
..J
w
-100 m .02 z
'" '"0 10
3
-
0- V1
w .01 <D
0
-
-200 L~0-~-:=:~;=:::=:=::l 0 0
~

o 20 40 60 80 100 0 10 20 30 40 50 60
TIME (days) TliAE (days)

Fig. 3. Desorption of phosphorus and phosphorus concen-

100 - . - - - - - - - - - - - - - - - - - - - ,
B tration in solution in a diluted slurry (1.5 g 1- lOW) of mixed
sediment from 1-10 cm depth, under anaerobic conditions.
N
,
E
CJ)
0fa=*"'='~;::::-----___:::,.L=--__1
E
w
250

-
V1 '0>
<! .:<
w
cj -100 0>
E 200
'" 0-
0-
w 150
..J

o 20 40 60 80 100
m
<!
f-
••
u 100
<!
TIME (days)
'"
f-
X o- 2 em _
Fig. 2. Phosphorus release and uptake (indicated as a nega- w 50
8 - 10 em •
tive release) in intact sediment cores under various experi- 0
U
mental conditions: a) aerobic release, b) aerobic release after <! 0
0 20 40 60 80 100
aerobic uptake, c) anaerobic release after aerobic uptake,
d) anaerobic release, e) anaerobic uptake and f) anaerobic up- TIME (days)
take after Fe(OHh addition. Fig. 4. Acid-extractable phosphorus in sediment samples
from different depths during aerobic incubation of diluted
sediment slurries.

centration in solution slowly increased to a max-

imum of about 0.05 mg 1- I due to a phosphorus
The total phosphorus content of the sediment
desorption of 30 mg kg- I DW. The 1-10 cm
top layer decreased from 0.9 (± 0.3) g kg- 1 in
layer has an average dry weight content of 9% or
1982 to 0.6 (± 0.1) g kg- 1 in 1991. Loss on igni-
8.1 kg m - 2. On an areal basis, the observed
tion and total iron were more or less constant
phosphorus desorption thus corresponds to
(Table 2). Simple regression analysis showed a
250mgm- 2 •
significant positive correlation between phospho-
The increase of the inorganic phosphorus pool
during aerobic incubation of the sediment slurries
reflects the phosphorus mobilization due to min- Table 2. Composition (SO) of surface sediment (0-2 cm)
eralization. A gradual increase of the acid extract- sampled between 1982-1991.
able phosphorus pool with time from 110 to
230 mg kg - t was observed in sediment from 0- 1982 (11= 16) 1985 (11=21) 1991 (n= 19)

2 cm (Fig. 4). No such increase was observed for Total phosphorus (g kg - ') 0.9 (0.3) 0.7 (0.2) 0.6 (0.1)
the 8-10 cm sediment layer, indicating that the Total iron (g kg - ') 21.9(4.7) 22.3 (3.5) 22.2 (2.9)
easily degradable organic pool in this material is Loss on ignition (%) 55.3 (13.5) 56.9 (7.5) 55.4 (5.1)
P/Fe (mmol mol- ') 75.4 (22.0) 54.8 (8.1) 48.1 (9.6)
negligible.
44
rus and iron content. A negative correlation be-
tween phosphorus content and year of sampling
was obtained (Table 3). Multiple regression anal-
ysis with total phosphorus as independent varia-
ble, and iron content and year as predictive
variables, indicated that between 1982 and
1991 the total phosphorus content in this layer
had decreased with an average rate of 34
(± 10) mg kg- 1 y- I (p<O.OOI).
Discussion
The contribution of phosphorus in the sediment
to the internal nutrient cycle of a lake is deter-
mined by the rate at which phosphorus is regen-
erated by mineralization and transported into the
water column by diffusion or resuspension. Dur-
ing summer, the sediment oxygen demand of the
peaty sediment of the Loosdrecht lakes is approx-
imately 30 mmol m - 2 d - I (Sweerts & Cappen-
berg, 1988). With the assumption that the CjP
ratio of the organic matter that is mineralized in
the sediment equals the CjP ratio of the seston
< 150 ~m, which is about 300 (Gulati et al., 1991),
the regeneration of inorganic phosphorus due to
mineralization during summer is estimated at
about 3 mg P m - 2 d - 1. The diffusive release of
phosphorus under aerobic conditions as mea-
sured with intact sediment cores is an order of
magnitude lower. Obviously, the adsorption of
phosphorus in the aerobic top layer is quite rapid,
so that inorganic phosphorus regenerated by min-
eralization is immobilized almost completely. Our
study confirms that under aerobic conditions the
sediment may adsorb considerable amounts of
phosphorus, without stimulating the release sig-
Table 3. Correlation matrix (r values) of sediment 0-2 cm
composition sampled in 1982-1991 (N = 56). *** p < 0.00 1.
P Fe P/Fe Year
Loss on ignition 0.16 0.29 -0.10 -0.01
Total phosphorus 0.48*** 0.80*** - 0.46***
Total iron 0.11 0.02
P/Fe ratio - 0.56***
nificantly (Fig. 2A). The measured uptake capac-
ity of the aerobic sediment has the same order of
magnitude as the annual phosphorus retention of
0.2 g m- 2 •
These results appear to agree with the classical
studies (Einsele, 1936; Mortimer, 1941, 1942) ac-
cording to which the phosphorus adsorption by
iron under oxidized conditions prevents diffusive
release to the overlying water. Both the rapid de-
crease of the uptake capacity and the relative fast
release of adsorbed phosphorus during the anaer-
obic incubation (Fig. 2B) confirm that the exper-
imental conditions were adequate to mobilize
most of the phosphorus adsorbed by ferric iron.
Nevertheless, the untreated sediment cores re-
leased < 60 mg m - 2 phosphorus during the in-
cubation period. Therefore, the amount of phos-
phorus adsorbed by ferric iron under field
conditions must be extremely small.
Compared to the mineralization rate, diffusive
release of phosphorus from the sediment cores is
extremely low throughout the year. Therefore,
processes other than diffusion must operate to
explain the low level of inorganic phosphorus in
the sediment. Downward seepage in the Loos-
drecht lakes equals 1.5 mm d - 1 and is likely to
contribute considerably to removal of phospho-
rus from the top sediment layer (Keizer et al.,
1991). Also the recurrent sediment resuspension
in the Loosdrecht lakes (Gons et at., 1986) can
promote a reintroduction of adsorbed phospho-
rus into the nutrient cycle.
The potentially available phosphorus fraction
in the sediment acts as a buffer in the nutrient
cycling of the lake (Bostrom et al., 1982; Mars-
den, 1989). A desorption of phosphorus from the
solid phase occurs if concentrations in the inter-
stitial water decrease (e.g. Froelich, 1988). The
anaerobic release from the sediment cores indi-
cates that desorption from the solid phase in the
top layer is less than 60 mg m - 2. The results in
Fig. 3 indicate that a phosphorus desorption from
the 1-10 cm layer of 250 mg m - 2 is possible if the
phosphorus concentration in solution is reduced
to 0.05 mg 1- 1. This value is about 6 times smaller
than the actual average phosphorus concentra-
tion in the interstitial water (Sinke et al., 1990).

Thus the total amount of inorganic phosphorus in
the uppermost 10 cm of the sediment that will be
mobilized after a severe reduction of the dissolved
phosphorus concentration in the sediment is
about 300 mg m - 2. An overestimate is likely be-
cause part of the dissolved phosphorus in both
experiments may have been produced by anaer-
obic mineralization processes as well.
In addition, there is also a large organic pool
that serves as a phosphorus source. The size of
this pool is derived from the mineralization ex-
periments. In the 0-2 cm layer of the sediment a
gradual increase of the inorganic phosphorus pool
with time, from 110 mg kg- I to a maximum of
240 mg kg - I, is observed upon aerobic incuba-
tion. This increase gives the amount of phospho-
rus that may become readily available for the in-
ternal cycle upon mineralization of easily
degradable organic matter. In the 8-10 cm layer
of the sediment such a pool is virtually absent. In
the same sediment Sinke et al. (1991) reported an
exponential decrease of chlorophyll a and alka-
line phosphatase activity with depth. Assuming
that a similar pattern is likely for the content of
degradable organic matter, we estimated the total
pool size in the upper 10 cm at 500 mg m - 2. This
is only about 10 % of the total organic phospho-
rus fraction in this layer. Obviously the major part
of the organic matter in this peaty sediment is
highly resistent to microbial decomposition.
Modelling
The phosphorus cycle in the sediment of the
Loosdrecht lakes is described most simply by a
model with three state variables (Fig. 5). In the
sediment there are two phosphorus pools: phos-
phorus in easily degradable organic matter
(PORG) and dissolved and adsorbed phospho-
rus (PADS). Phosphorus in the watercolumn is
confined to one pool (PWAT) that consists of
both algae and detritus. Soluble reactive phos-
phorus in the water column is ignored because its
contribution to the total phosphorus concentra-
tion in the Loosdrecht lakes is negligible.
The phosphorus content in the lake water is the
resultant of the external phosphorus load (Pext ),
45
~@
1.0 PWAT
150
Water
se
P ,:
1.5 ~Pint
0.9
----_ _-_ _---_ _--- _-----._ _-----------
Sediment
r=\ ~
~
Pmin
1.5
Fig. 5. Schematic representation of the phosphorus dynamics
at the sediment-water interface and in the sediment of the
Loosdrecht lakes. Pool sizes in mg m - 2 Rates in
mg m - 2 d - I. PWAT = total phosphorus in the water column,
PADS = adsorbed and dissolved inorganic phosphorus,
PORG = phosphorus in easily degradable organic matter.
Pex, = external P-load, Pout = P-discharge, P sed = P-
sedimentation, P min = P-mineralization, Pint = total gross in-
ternal P-load, P seep = P-removal with downward seepage.
the discharge (Pout) and the net flux to the sedi-
ment. The latter is the resultant of sedimentation
of easily degradable organic matter (P sed ) and the
inorganic phosphorus release from the sediment
(Pint). Sedimentation of refractory organic matter
and inorganic particulate phosphorus is not con-
sidered in the model. The contribution of these
fractions to P ext and Pout is believed to be rela-
tively small and will hardly affect the numeral
values presented in Fig. 5.
Pint includes both diffusive release and uptake
of adsorbed inorganic phosphorus by algae dur-
ing resuspension. PORG has its origin in sedi-
menting algae (P sed ) and loses phosphorus due to
mineralization (Pmin)' PADS is fed by mineral-
ization (P min), and loses phosphorus due to
downward seepage (Pseep ), and transport into the
water column (PinJ.
PWAT is calculated from Van Liere et al.
(1990) and the average depth of the lakes. PORG
and PADS were determined experimentally as
described above. P ext and Pout are based on an-
nual phosphorus balances reported by Engelen
et at. (1992). The average mineralization rate
(P min) was calculated from the measured water
46
temperatures and temperature dependent oxygen
consumption rates reported by Sweerts (1990).
With these input variables the missing 'rates for
Pseep , Psed and Pint can be estimated under the
assumption of steady state.
According to Van Liere et al. (1990) total phos-
phorus in the water column decreased between
1984 and 1989 at an average rate of 12 mg
m - 3 Y- 1, which corresponds to 0.054 mg
m - 2 d - 1 on an areal basis. The total phosphorus
content of the 0-2 cm sediment has decreased
since 1982 with an average rate of 34 mg
kg- 1 y- I, which corresponds to 0.15 mg
m - 2 d - 1. It seems reasonable to assume that the
present decrease rate is smaller than immediately
after the reduction of the external phosphorus
load in 1984. So the average value over the past
years presumably overestimates the present de-
crease. Although it cannot be ruled out that a
decrease still proceeds at a slow rate, it is obvi-
ous that the present average net input of phos-
phorus to the sediment approximates the differ-
ence between Pext and Pout> viz. 0.6 mg m - 2 d - I.
In addition, the input of phosphorus on top of the
sediment must be almost offset by downward
seepage. Based on an average seepage of
1.5 mm d - I and an annual average phosphorus
concentration of 0.5 mg l- 1 at 10 cm depth (A.
Sinke, pers. comm.), downward transport of
phosphorus due to seepage is estimated at
0.8 mg m - 2 day - 1. Thus removal by downward
seepage may indeed balance the major part of the
phosphorus input onto the top the sediment.
Moreover, phosphorus retention by sediment
burial or diagenesis must be very small. However,
because the difference between annual phospho-
rus retention and the estimated phosphorus re-
moval by seepage is only a rough estimate, it can-
not be said whether the phosphorus content of
the upper 10 cm is still decreasing or not. Al-
though spatial variability of seepage rates and
phosphorus concentrations in pore water have to
be examined in more detail, the results emphasize
the importance of downward seepage as a limit-
ing process for the accumulation of potentially
available phosphorus near the sediment-water in-
terface.
The estimated rate for the phosphorus trans-
port from the sediment to the overlying water
(Pint) is much higher than the diffusive release of
0.3 mg m - 2 d - 1 measured with the intact sedi-
ment cores. The limited validity of laboratory re-
lease experiments to determine the total flux of
phosphorus from the sediment to the overlying
water in shallow lakes, was earlier suggested by
Ryding & F orsberg (1977) and confirmed recently
by Sinke et al. (1990).
The presented sedimentation of phosphorus of
1.5 mg m - 2 d - 1 corresponds to a settling velo-
city of 2 cm d - 1, which is a factor two smaller
than measured by Gons & van Keulen (1987) in
a laboratory set up. This discrepancy is partly a
result of wind-induced turbulence in the field and
partly of using annual averages rather than sea-
sonal values. Actual summer values for Pmin' Psed
and Pint may be approximately twice as high as
indicated in Fig. 5, but in winter they may be close
to zero.
In order to predict future trends after applica-
tion of various additional management measures,
relations between rates and pool sizes were de-
fined according to Equations 1-3.
dPWATjdt = Pext - (K sed + K aut ) PWAT
+ KintPADS (1)
dPORGjdt = KsedPWAT - KminJpORG (2)
dPADSjdt = KminJpORG - KintPADS
(3)
It is assumed that the total flux from the sediment
to the overlying water (Pint) and the removal by
downward seepage (P seep ) are linearly propor-
tional with factors K int and K seep to PADS. Sed-
imentation (P sed ) and the discharge of phospho-
rus (Pout) are assumed to be linearly proportional
with factors K sed and K aut to the total amount of
phosphorus in the overlying water (PWAT). P min
is assumed to be proportional to the square root
of the degradable organic pool size, similar to the
relation between the sediment oxygen demand
and the volumetric oxygen demand reported by
Bouldin (1968). The values of the proportionality

constants were calculated from the steady state
values presented in Fig. 5.
With Equations 1-3 the trends are calculated
after the following treatments: A) 50% reduction
of the external phosphorus load; B) 50% reduc-
tion of the external phosphorus load and com-
plete reduction of downward seepage by inunda-
tion of surrounding polders; C) sediment dredging
of the sediment top layer which removes 80% of
the easily degradable organic matter (PORG);
and D) combination of dredging and a 50% re-
duction of the external phosphorus load.
Due to the boundary condition, that at present
the system is in steady state, no changes occur
without any additional management measure-
ments. According to the model, a reduction of the
external phosphorus load will result in a gradual
reduction of PWAT (Fig. 6). The new equilibrium
will be reached much faster if this measure is
accompanied by sediment dredging. Dredging
without a reduction of the external phosphorus
load will initially improve the water quality, but
eventually the system will return to its former
equilibrium. These predictions reflect the role of
the mobile sediment fractions (PORG and
PADS) in the internal nutrient cycle. The rates at
which these pools decrease determine largely the
changes in PWAT. A reduction of the external
load is followed by a gradual decrease of these
100
D
<;>
80
E reported by Ryding (1982) for lake Trehorningen.
C1l
E 60
CL
40
<i. c simultaneous 50% reduction of the present ex-
-0 20
0 a rate of 0.05 g m - 2 d - 1 (Fig. 2), i.e. after six
o 4 6 8 10
TIME (years)
Fig. 6. Simulated effect on total phosphorus concentrations
in the water column after various additional management
measures. Dotted line: no additional measurements; A) 50%
reduction of the external phosphorus load; B) sediment re-
moval; C) sediment removal and 50% reduction of external
phosphorus load and D) 50% reduction of external phospho-
rus load without downward seepage.
47
pools as long as the total phosphorus input Pexl
is lower than the total output (PaUL + P seep )' Sedi-
ment dredging of the top layer results in an in-
stantaneous reduction of the mobile phosphorus
pool. Therefore, even without a reduction of the
external load, dredging results in an initial reduc-
tion ofPWAT. However, after dredging the total
output of phosphorus from the system is lower
than the total input. Under these conditions the
mobile fractions of phosphorus in the sediment
are likely to increase in size till the original steady
state conditions are reached again. If a 50% re-
duction of the external phosphorus load is com-
bined with a complete elimination of the down-
ward seepage, the changes in PWAT are relatively
small, because the difference between input and
output hardly differs from the present situation.
The trends revealed by the model calculations
demonstrate that the difference between phospho-
rus input and output, rather than the external
phosphorus load alone, determines the final con-
centration of phosphorus in the water. In lakes in
which the phosphorus output by either discharge
or seepage is less favourable, restoration perspec-
tives are worse. A reduction of the external phos-
phorus load has a limited effect in case of a large
potentially available pool of phosphorus and a
small discharge of phosphorus after the measure-
ment. In such cases dredging may contribute sig-
nificantly to the improvement of the water qual-
ity. Such a behaviour was, indeed, observed for
Lake Trummen (Bengtsson et al., 1975). Limited
success of sediment dredging has been, however,
The effects of application of 100 gram m - 2 iron
as Fe(OHh were calculated with and without a
ternal phosphorus load. Four additional assump-
tions were made: 1) the added iron is reduced at
years all the added iron is converted into reduced
precipitates; 2) initially, the addition of iron stops
the removal of phosphorus by downward seepage
completely (Kseep is set at zero); 3) to simulate the
effects of improved coagulation in the sediment
top layer and suppressed resuspension of sedi-
ment material, the internal phosphorus load de-
48
creases (Kint is reduced by a factor five) immedi-
ately after the iron addition; 4) K seep and K int
return to their original values after complete re-
duction of the added iron.
After the addition of iron an initial decrease of
PWAT occurs, because of the improved adsorp-
tion of phosphorus in the sediment (Fig. 7). How-
ever, the adsorption capacity of the iron decreases
because of chemical and microbial reduction pro-
cesses. At the same time PADS increases as min-
eralization proceeds at its original rate, while
downward seepage and transport to the water are
suppressed because of the strong adsorption to
the added iron. The iron treatment initially im-
proved the retention of phosphorus in the sedi-
ment, but this results eventually in a relatively
large internal load as the reduction of the iron
proceeds. Finally, the gradual increase of PADS
and K int results in an internal load (Pint) at a
somewhat higher level than in the absence of the
iron treatment.
The time scale on the abscissae in Fig. 6 and
Fig. 7 must be considered as a rough indication
because of the limited accuracy of the model. The
scaling is in particular sensitive to the difference
between input and output relative to the pool sizes
in the sediment. Because of its simplicity, further
adaptations will be needed to predict the behav-
ior of the lake if the structure of the nutrient cycle
100 A
f0'
I 80
t
0;
t 60
CL
--'
40
<!
c-
o
c- 20
o be reduced by flushing through the water column.
0 2 4 6 8 10
TIME (years)
Fig. 7. Simulated effects of iron additions on total phospho-
rus concentrations in the Lake water. Dotted line: no treat-
ment; A) iron addition, no reduction of external phosphorus
load; B) 50% reduction of external load and C) combination
of iron addition and 50% reduction of external load.
changes e.g. after the introduction of submersed
marcophytes or after biomanipulation. Part of
these processes are included in more sophisti-
cated models like PCLOOS (Janse et at., 1992)
which contains a description of sediment phos-
phorus dynamics that is largely comparable to the
approach presented here.
Conclusions
The diffusive release of phosphorus from the sedi-
ment of the Loosdrecht lakes is small, both under
aerobic and anaerobic conditions. Inorganic
phosphorus generated by mineralization pro-
cesses is removed from the sediment by down-
ward seepage and uptake in the nutrient cycle
during resuspension. The phosphorus cycle in the
lake, including the upper 10 cm of the sediment is
close to steady state. In the Loosdrecht lakes
phosphorus removal by downward seepage is re-
sponsible for the withdrawal of phosphorus from
the nutrient cycle. It seems that diagenesis and
sediment burial hardly act as a permanent sink
for phosphorus. The total phosphorus content of
the sediment top layer has decreased significantly
since 1982. The amount of potentially bioavail-
able phosphorus in the upper 10 cm equals
< 15 % of the total phosphorus content of the
sediment, but is still three times larger than the
annual external phosphorus load.
The simulations indicate that in the Loosdrecht
lakes a further reduction of the external load will
reduce the total phosphorus content of the lake
water since the historical phosphorus load in the
sediment will be reduced relatively fast by down-
ward seepage. In comparable lakes without
downward seepage, restoration perspectives are
less favourable, because the bioavailable phos-
phorus load accumulated in the sediment can only
Sediment dredging and iron addition without re-
duction of the external load may be successful on
a short term, but not on a longer time scale. Elim-
ination of downward seepage by hydrological
measures will counteract the effect of a reduction
of the external load.
 49

Acknowledgements wind on a shallow lake ecosystem: resuspension of

particles in the Loosdrecht Lakes. Hydrobiol. Bull. 20:
109-120.
This study is part of the 'Water Quality Research Gulati, R. D., L. Van Liere & K. Siewertsen, 1991. The Loos-
Loosdrecht Lakes'. It was financially supported drecht lake system: Man's role in its creation, perturbation
by the Provincial Waterboard of Utrecht and the and rehabilitation. In O. Ravera (ed.), Terrestrial and
Netherlands Integrated Soil Research Pro- aquatic ecosystems: perturbation and recovery. Ellis Hor-
wood Ltd: 593-606.
gramme. The authors with to thank M. Buysman,
Holdren Jr, G. C. & D. E. Armstrong, 1980. Factors affect-
P. de Joode, F. Cottaar and S. Visser for tech- ing phosphorus release from intact lake sediment cores.
nical assistance. Envir. Sci. Technol. 14: 79-87.
Hosomi, M., M. Okada & R. Sudo, 1982. Release of phos-
phorus from lake sediments. Envir. Int. 7: 93-98.
References Janse, J. H., T. Aldenberg & R. P. G. Kramer, 1992. A math-
ematical model of the phosphorus cycle in Lake Loosdrecht
and simulation of additional measures. Hydrobiologia 233:
Bengtsson, L., S. Fleisher, G. Lindmark & W. Ripl, 1975.
119-136.
Lake Trummen restoration project I. Water and sediment
Kamp-Nielsen, L., 1974. Mud-water exchange of phosphate
chemistry. Verh. int. Ver. Limnol. 19: 1080-1087.
and other ions in undisturbed sediment cores and factors
Boers, P. C. M. & O. van Hese, 1988. Phosphorus release
affecting the exchange rates. Arch. Hydrobiol. 73: 218-237.
from the peaty sediments of the Loosdrecht Lakes (The
Keizer, P., M. C. N. P. Buysman & Th. E. Cappenberg, 1991.
Netherlands). Wat. Res. 22: 355-363.
Sorption and release of phosphorus in a peaty sediment.
Boers, P. C. M., J. w. Th. Bongers, A. G. Wisselo & Th. E.
Verh. int. Ver. Limnol. 24: 722-725.
Cappenberg, 1984. Loosdrecht Lakes Restoration Project:
Lijklema, L., J. H. Janse, R. M. M. Royackers & M.-L. Me-
Sediment phosphorus distribution and release from the sed-
ijer, 1988. Eutrofiering in Nederland. H 2 0 21: 462-467.
iments. Verh. int. Ver. Limnol. 22: 842-847.
Marsden, M. W., 1989. Lake restoration by reducing the ex-
BostrOm, B., M. Jansson & C. Forsberg, 1982. Phosphorus
ternal phosphorus loading: the influence of sediment phos-
release from lake sediments. Arch. Hydrobiol. Beih. 18:
phorus release. Freshwat. BioI. 21: 139-162.
5-59.
Mortimer, C. H., 1941. The exchange of dissolved substances
BostrOm, B., I. Ahlgren & R. Bell, 1985. Internal nutrient
between mud and water in lakes I. J. Ecol. 29: 280-329.
loading in a eutrophic lake, reflected in seasonal variations
Mortimer, C. H., 1942. The exchange of dissolved substances
some sediment parameters. Verh. int. Ver. Lirnnol. 22:
between mud and water in lakes II. J. Ecol. 30: 147-201.
3335-3339.
Murphy, J. & J. P. Riley, 1962. A modified single solution
Bouldin, D. R., 1968. Models for describing the diffusion of
method for the determination of phosphate in natural wa-
oxygen and other mobile constituents across the mud-water
ters. Analyt. Chim. Acta 27: 31-36.
interface. J. Ecol. 56: 77-87.
Nilrnberg, G. K., 1988. Prediction of phosphorus release rates
Cooke, G. D., E. B. Welch, S. A. Peterson & P. R. Newroth,
from total and reductant soluble phosphorus in anoxic lake
1986. Lake and reservoir restoration. Butterworth Publ.
sediments. Can. J. Fish. aquat. Sci. 45: 453-462.
Boston. pp. 392.
Redshaw, C. J., C. F. Mason, C. R. Hayes & R. D. Roberts,
Cullen, P. & C. Forsberg, 1988. Experiences with reducing
1990. Factors influencing phosphate exchange across the
point sources. Hydrobiologia 170: 321-336.
sediment-water interface ofeutrophic reservoirs. Hydrobio-
Einsele, W., 1936. Uber die Beziehungen des Eisenkreislaufs
logia 192: 233-245.
zum Phosphatkreislaufim eutrophen See. Arch. Hydrobiol.
Ryding, S. 0., 1982. Lake Trehorningen restoration project.
29: 664-686.
Changes in water quality after sediment dredging. Hydro-
Engelen, G. B., B. F. M. Kal, J. J. Buyse & F. G. M. Van
biologia 92: 549-558.
Pruissen. The hydrology of the Loosdrecht lakes area. Hy-
Ryding, S. 0., 1985. Chemical and microbiological processes
drobiologia 233: 21-38.
as regulators of exchange of substances between sediments
Fries, J. & H. Getrost, 1977. Organische Reagenzien fur die
and water in shallow eutrophic lakes. Int. Revue. ges. Hy-
Spurenanalyse. E. Merck, Darmstadt.
drobiol. 70: 657-702.
Froelich, P. N., 1988. Kinetic control of dissolved phosphate
Ryding, S. O. & C. Forsberg, 1977. Sediments as a nutrient
in natural rivers and estuaries: A primer on the phosphate
source in shallow polluted lakes. In H. L. Golterman (ed.),
buffer mechanism. Lirnnol. Oceanogr. 33: 649-668.
Interactions between sediments and freshwater. Dr. W.
Gons, H. J. & R. Van Keulen, 1987. The relatie tussen door-
Junk, The Hague: 227-234.
zicht en slib in the Loosdrechtse plassen in verband met de
Sas, H. (ed.), 1989. Lake restoration by reduction of nutrient
zwemwaternorm. WQL-rapport 1989-4. Limnological In-
loading: expectations, experiences, extrapolations. Aca-
stitute Nieuwersluis. 35 pp.
Gons, H. J., R. Veeningen & R. Van Keulen, 1986. Effects of demia Verlag Richarz, Sankt Augustin, 497 pp.
50
Shaw, J. H. F. & E. E. Prepas, 1990. Exchange of phospho-
rus from shallow sediments at nine Alberta lakes. J. envir.
Qual. 19: 249-256.
Sinke, A. J. C. & Th. E. Cappenberg, 1988. Influence ofbac-
teria1 processes on the phosphorus release in the eutrophic
Loosdrecht lakes, The Netherlands. Arch. Hydrobiol. Beih.
30: 5-13.
Sinke, A. J. c., A. A. Corne1ese & Th. E. Cappenberg, 1991.
Phosphatase activity in sediments of the Loosdrecht lakes.
Verh. int. Ver. Limnol. 24: 719-721.
Sinke, A. J. C., A. A. Cornelese, P. Keizer, O. F. R. Van
Tongeren & Th. E. Cappenberg, 1990. Mineralization,
porewater chemistry and phosphorus release from peaty
sediments in the eutrophic Loosdrecht lakes. Freshwat.
BioI. 23: 587-599.
Sondergaard, M., 1989. Phosphorus release from a hypertro-
phic lake sediment: Experiments with intact sediment cores
in a continuous flow system. Arch. Hydrobiol. 116:
45-59.
Sondergaard, M., P. Kristensen & E. Jeppesen, 1992. Phos-
phorus release from resuspended sediment in the shallow
and wind-exposed Lake Arreso, Denmark. Hydrobiologia
228: 91-99.
Sweerts, J-P. R. A., 1990. Oxygen consumption, mineraliza-
tion and nitrogen cycling at the sediment-water interface of
north temperate lakes. Ph.D. Thesis University of Gron-
ingen.
Sweerts, J-P. R. A. & Th. E. Cappenberg, 1988. Use of
micro-electrodes for measuring oxygen profiles in lake sed-
iments. Arch. Hydrobiol. Beih. 31: 365-371.
Tessenow, U., 1972. Losungs- Diffusions- und Sorptionsproz-
esse in der Oberschicht von Seesedimenten. 1. Ein
Langzeit-experiment unter aeroben und anaeroben Bedi-
nungen im Fliessgleichgewicht. Arch. Hydrobiol. 4, Suppl.
38: 353-398.
Van Liere, L., 1986. Loosdrecht lakes, origin, eutrophication,
restoration and research programme. Hydrobiol. Bull. 20:
9-15.
Van Liere, L., L. Breebaart, W. Kats & J. J. Buyse, 1989. De
waterkwaliteit in het Loosdrechtse plassengebied. CHO/
TNO rapport 22: 265-279.
Van Liere, L., O. F. R. Van Tongeren, L. Breebaart & W.
Kats, 1990. Trends in chlorophyll and total phosphorus in
the Loosdrecht lakes, The Netherlands. Verh. int. Ver. Lim-
nol. 24: 707-710.
Hydrobiologia 233: 51-59, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 51
© 1992 Kluwer Academic Publishers.

Coupling of phytoplankton and detritus in a shallow, eutrophic lake

(Lake Loosdrecht, The Netherlands)

Herman J. Gons,l Tineke Burger-Wiersma,2 Jerko H. Otten I & Machteld Rijkeboer 1

1 Limnological Institute, 'Vijverhof Laboratory, Rijksstraatweg 6, 3631 AC Nieuwersluis, The Netherlands;

2 Laboratory for Microbiology, University of Amsterdam, Nieuwe Achtergracht 127, 1018 WS

Amsterdam, The Netherlands

Key words: phytoplankton, Oscillatoria limnetica, Prochlorothrix hollandica, detritus, shallow lakes, lake
restoration

Abstract

An oscillating steady state is described of phytoplankton, dominated by Prochlorothrix hollandica and

Oscillatoria limnetica, and sestonic detritus in shallow, eutrophic Lake Loosdrecht (The Netherlands).
A steady-state model for the coupling of the phytoplankton and detritus is discussed in relation to field
and experimental data on phytoplankton growth and decomposition. According to model predictions,
the phytoplankton to detritus ratio decreases hyperbolically at increasing phytoplankton growth rate and
is independent of a lake's trophic state. The seston in L. Loosdrecht contains more detritus than
phytoplankton as will apply to many other lakes. The model provides a basis for estimating the loss rate
of the detritus, including decomposition, sedimentation and hydraulic loss. In a shallow lake like L.
Loosdrecht detritus will continue to influence the water quality for years.

Introduction The low transparency is due to phytoplankton

Reductions of the external phosphorus loading
did not significantly improve the water quality of
the shallow and highly eutrophic Lake Loosdrecht
(Loogman & Van Liere, 1986; Van Liere et al.,
1990). The lake is still characterized by predom-
inance of filamentous prokaryotic species, includ-
ing Oscillatoria spp. and Prochlorothrix hollandica
(De Kloet et al., 1990; Van Liere et al., 1989),
absence of submerged vegetation (Malthus et al.,
1990), high stocks of bream (Abramis brama), lack
oflarge-bodied grazing zooplankton (Gulati et al.,
1992), very low Secchi-disc transparency (almost
always < 0.5 m), and high annual mean concen-
trations of chlorophyll a and total phosphorus
(both ca. 100 mg m - 3).
and particulate detritus. The detritus consists of
peat fragments and fine-sized « 15 J.!m) parti-
cles. The peat fragments appear in the water col-
umn after resuspension in periods of high wind-
speed, but rapidly resettle in calm conditions. The
fine-sized detritus is likely to derive mainly from
the phytoplankton, has higher dry weight concen-
trations and a greater partial attenuation coeffi-
cient for downwelling light than the phytoplank-
ton (Gons, 1987; Gons et al., 1991; Otten et al.,
1992; Rijkeboer & Gons, 1990).
Each of the phytoplankton and detritus com-
ponents alone would prevent the Secchi-depth
from reaching 1 m. Because the lake shows fre-
quent resuspension of sediments (GollS et al.,
1986b; 1991) and the uniformity of depth almost
52
entirely precludes the permanent deposition of
particles, it might be inferred that even a drastic
reduction in phosphorus loading would not
achieve this water-quality goal. However, the
history of L. Loosdrecht (see further) and data
on other lakes of various trophic states (Saun-
ders, 1972; Wetzel, 1983; Wetzel et al., 1972),
demonstrate that the detritus concentration in
lakewater is coupled to phytoplankton produc-
tion. Furthermore, in having a higher concentra-
tion of particulate detritus than of phytoplankton,
L. Loosdrecht is no exception but rather repre-
sents the rule (Wetzel, 1983). It can thus be sur-
mised that a decrease in phytoplankton produc-
tion will inevitably involve decrease of the detritus
as well. The question remains how long the re-
fractory, fine-sized detritus that accumulated in
the past may prevent a clear state from develop-
mg.
Lake Loosdrecht exhibits only a weak season-
ality in the concentrations of chlorophyll and
seston dry weight, and their dynamics during
summer can be characterized as an oscillating
steady state (Gons et al., 1986a; Gons & Rijke-
boer, 1990). The validity of applying this concept
is illustrated on a basis of a time series of cWo-
rophyll a and seston dry weight concentrations in
summer 1988. The question of how change in
detritus concentration is related to the phyto-
plankton growth in this very shallow lake is ad-
dressed with a simple steady-state model. Model
predictions are discussed in relation to concen-
trations of chlorophyll a and seston dry weight,
and experimental data on phytoplankton growth
and mortality.
Theory
We presume that the phytoplankton growth in the
lake in summer is very nearly at steady state. The
specific growth rate of the phytoplankton (J1) var-
ies slightly around a mean value set by the spe-
cific rate of loss (D) due to grazing, sinking, and
other causes, hence J1 = D (time - 1). The phyto-
plankton biomass depends on the availability of
light and nutrients. Steady-state phytoplankton
growth may also involve constant concentrations
of detritus and seston.
Besides the steady-state condition, three fur-
ther assumptions are made:
(1) The seston dry weight concentration
(S g m - 3) consists exclusively of phytoplankton
(PH g m - 3) and fine-sized detritus (FD g m - 3):
[S] = [PH] + [FD] (Eq. 1)
For defined phytoplankton this partitioning can
be solved using chlorophyll (CHL g m - 3) and
seston dry weight measurements as in Eq. (2),
using f3 (g PH g- 1 CHL) as a symbol for a con-
stant of conversion of chlorophyll into phyto-
plankton dry weight:
[S] = f3[CHL] + [FD] (Eq.2)
(2) Regardless of the cause of phytoplankton
mortality, the only product is a fixed refractory
fraction (r) of the biomass, which is instanta-
neously released into the pool of fine-sized detri-
tus.
(3) Decrease of the fine-sized detritus has a
constant over-all specific loss rate (J (time - 1),
which includes decomposition, permanent depo-
sition and hydraulic wash-out.
Change in the concentration of fine-sized de-
tritus can now be given as:
d [FD]/dt = D [PH]r - (J [FD] (Eq.3)
We presume [FD] to be also at steady state:
d[FD]/dt = 0, hence:
[FD] = D/(J [PH] r (Eq.4)
Consequently, the concentration of fine-sized de-
tritus associated with a given concentration of
phytoplankton with defined r-value depends on
the ratio of the loss rates of the two components.
We note that also [S] will be constant (Eq. 1).
After re-arrangement ofEq. (4) the phytoplank-
ton to detritus ratio R p / d is given by:
R p / d = [PH]/[FD] = (J/Dl/r (Eq.5)
and the value of (J.
(J = Rp/dDr (Eq. 6)
Equation (5) implies that, in principle, the shares

of phytoplankton and detritus in seston are inde-
pendent of a lake's trophic state.
The value of (J indicates the time needed for
restoration of L. Loosdrecht's transparency once
the phytoplankton growth has been minimized.
For phytoplankton in L. Loosdrecht values of D
( = Ji) have been estimated using field and exper-
imental studies. Values of f3 and r have been ob-
tained for O. limnetica and P. hol/andica, the dom-
inating species in summer, at defined conditions
of light and phosphorus limitation.
The lake
Lake Loosdrecht (situated ca. 20 km SE of
Amsterdam, The Netherlands; A = 9.8 km 2 ;
z = 1. 9 m) has its origin in peat dredging and sub-
sequent erosion by wind and waves. Until the
nineteen-fifties the lake was clear and had exten-
sive benthic growth of characeans. Only storms
might turn the water turbid for one or two days.
Green algae and diatoms dominated the phyto-
plankton. Phytoplankton dry weight concentra-
tion, estimated from the biovolume, was ca.
0.2 g m - 3. Studies in the early sixties showed dis-
appearance of the characeans and a drastic de-
crease in water transparency and changes in
plankton composition symptomatic of eutrophi-
cation.
Lake Loosdrecht is a seepage lake needing sub-
stantial supplementary water supply to maintain
its water table in summer. It is thought that phos-
phorus loading due to inlet of water from the
highly polluted River Vecht has been the main
cause of the eutrophication. Since 1984 onwards
water from the Amsterdam-Rhine Canal has re-
placed the supply of the River Vecht water. Ad-
ditional measures included sewerage diversion
and phosphate removal of the inlet water by co-
agulation with FeCI 3 .
Throughout the nineteen-eighties the phyto-
plankton showed perennial predomination by fil-
amentous prokaryotic species, mainly O. limnet-
ica (including incompletely identified filaments),
O. redekei, and Prochlorothrix hol/andica (Burger-
Wiersma et aI., 1986, 1989; De Kloet et al., 1990;
53
Van Tongeren et al., 1992). Unlike in other shal-
low, eutrophic Dutch lakes (Berger, 1987; Hosper
& Meijer, 1986; Moed & Hoogveld, 1982)
Oscil/atoria agardhii is not predominant. Many
eukaryotic species co-exist, among which
nanoflagellates may have high cell densities, but
only the diatoms Diatoma elongatum and Melosira
italica reach a considerable biomass in early
spring and autumn (Gons, 1991).
For details on the lake's morphometry, and
hydrological and biological history, see Engelen
et al. (1992), Gulati et al. (1991), Hofstra &
Van Liere (1992) and Loogman & Van Liere
(1986).
Materials and methods
In summer 1988 turbidity and the concentrations
of seston dry weight and chlorophyll were mon-
itored at a small island in the northern part of L.
Loosdrecht (Gons et al., 1991). From July on-
ward, in the morning and evening, water was col-
lected and stored using a refrigerating sampler
(Ibuk Abwasser-technik, Konigsbronn, Ger-
many). The storage time varied from two hours to
longer than three days, but all samples were pro-
cessed within four days after collection. The in-
fluence of storage time on the analyses appeared
to be insignificant.
Lakewater was prefiltered through 144 Jim
mesh screen. Seston dry weight was determined
by filtering through weighed Whatman G F IF fil-
ters. Chlorophyll a and phaeopigment were mea-
sured after extraction in hot ethanol (Moed &
Hallegraeff, 1978). Phaeopigment generally con-
stituted about 20% and always less than 30% of
the uncorrected chlorophyll values.
Growth of the natural phytoplankton was
studied using bags of dialysis membrane in situ
(Burger-Wiersma & Mur, 1986) and laboratory
scale enclosures (LSE; Rijkeboer et al., 1990b).
The enclosure in dialysis bags aimed at isolat-
ing the seston from large-bodied zooplankton and
exchange of particles, while preserving the phys-
ical and chemical conditions in situ. After filling
with prefiltered (144 Jim) lakewater, the dialysis
54

bags (length 0.6 m; pore size 2.4 nm) were sus-
pended vertically in the 0.1-0.7 m depth layer.
Change in protein was directly converted into
specific growth rate of the phytoplankton. Growth
rate estimates from carbohydrate decrease in the
dark were made according to results with cultures
of P. hollandica (Burger-Wiersma & Mur, 1986;
Burger-Wiersma & Baard, 1987).
The LSE (Rijkeboer et al., 1990ab) is a con-
tinuous flow system simulating the natural water
column under light and temperature conditions
in situ. Growth rates of phytoplankton in prefil-
tered (144 /lm) lakewater were estimated by com-
parison of the chlorophyll-specific column-inte-
grated oxygen production with the values for
cultures of P. hollandica and O. limnetica in LSE
(Rijkeboer et al., 1990b).
Light- and phosphorus-limited growth, and
chemical characteristics of P. hollandica were
studied using turbidostat cultures (Burger-
Wiersma & Baard, 1987), and those of O. limnet-
ica and P. hollandica using chemostat cultures in
LSE (Rijkeboer et al., 1990b).
Fractions of refractory particulate matter of
O. limnetica were determined after three weeks of
incubation in the dark in aerobic conditions
(Otten eta/., 1992).
Results and discussion
averages of CHL and S in these months remained
close to those of the whole period, i.e. 160 mg m - 3
and 34 g m - 3, respectively.
Stability of the phytoplankton and detritus re-
lationship is reflected by the chlorophyll content
of seston (Fig. 1). From May until September the
values fluctuated around 5 mg CHL g - 1, within
the range of 3 to 8 mg g - 1. This range is quite
narrow considering the variability of pigment
content in phytoplankton cultures alone (see fur-
ther). Meteorological variability appears to be the
main cause of these fluctuations (Gons & Rijke-
boer, 1990; Gons et al., 1991). For example, the
lowest record was due to resuspension of sedi-
ment during a severe storm in September. There-
after more resuspended phytoplankton relative to
other particles remained in suspension, hence the
values increased to 7 mg g - I, but later decreased
to 5 mg g- I again.
In 1988 prolonged stability of the phytoplank-
ton growth was aided by evenness of insolation
and temperature. A warm, sunny spring was fol-
lowed by a remarkably cloudy summer. Through-
out May-September the water temperature was
invariably between 15 and 20°C.
However, it needs to be emphasized that evi-
dently there is feed-back control of the chloro-
phyll content of the seston. In the theory section
it was explained that the seston concentration
may be constant at given steady-state phytoplank-

Seston dynamics ~ 10
en
.~

QJ

In a system at steady state the concentrations of ~

8
>-
all components are constant, a situation made '-
"0

-vV ~
feasible by finely tuned feed-back control. Envi- en
6
'-
ronmental perturbations may temporarily cause QJ
a.
great oscillations, but the feed-back mechanisms ~

>-
J
force the concentrations to return to the values ~
a.
imposed by the steady-state relationships. 0
'-
0
2
From May until September in 1988 the con- ~

~
u
centrations of chlorophyll a (CHL, including en
E
0
phaeopigment) and seston dry weight (S) ranged May Jun Jul Aug Sep Oct
from 103 to 202 mg m - 3 and from 16 to 67 g m - 3, 1988
respectively. Though the short-term fluctuations Fig. 1. Time series of chlorophyll a content (including phae-
were marked, seasonal trends were absent. The opigment) of seston dry weight in Lake Loosdrecht.
 55

ton production, but not how such a steady state
is controlled. The answer must be sought in the
functioning of the detritus derived from the phy-
toplankton itself. It has been recognized for a
long time (Hutchinson, 1957) that detritus plays
an important role in the metabolism of lakes.
Saunders (1972) explicitly showed how detritus
may function as a buffer as regards flows of en-
ergy and nutrients. With respect to phytoplank-
ton production in L. Loosdrecht, this capacity of
a function of J.l ( = D) for hypothetical values of (J
(Eq.6).
The relationship (Fig. 2) has J.l = 0 as an as-
ymptote, because there is no detritus formation
when there is neither growth nor loss of the phy-
toplankton population. As J.l increases, R p / d de-
creases hyperbolically and only for values of (J
> 0.02 d - 1 there can be more phytoplankton than
detritus over a wide range of growth rates. For
(J= 0.01 and 0.005 d - \ the concentration ofphy-

detritus has been found to control the biomass toplankton will be lower than of detritus at J.l
under both light and phosphorus limitation (Rij- > 0.1 and > 0.05 d - 1, respectively. For
keboer & Gons, 1990; Rijkeboer et al., 1990b; in (J= 0.001 d - 1 there will be much more detritus

press). than phytoplankton, except at extremely low

Furthermore, there is evidence that in the case
of L. Loosdrecht the detritus precludes the de-
velopment of the large-bodied zooplankton able
to graze the filamentous prokaryotic species
(Gulati etal., 1992). In this way the range of the
phytoplankton loss rate will be restricted to low
values. The deposition of these organisms may be
due to entrainment in flocs of detritus resettling
after resuspension rather than sinking of the in-
dividual trichomes (Gons et al., 1991). This
mechanism might set the lower limit to the loss
rate.
We conclude that the conditions for coupling
of phytoplankton and detritus concentrations
(Eq. 4) do exist. The seston dynamics in L. Loos-
drecht in summer 1988 possesses the features of
an oscillating steady state. The described theory
is applicable to the mean concentrations of CHL
and S measured in this period.
growth rates.
The value of (J, for fine-sized particles discussed
here, will be lower than 0.02 d - 1 in most lakes.
First, reported decomposition rates of refractory
detritus are invariably one to two orders of mag-
nitude lower (Otten et al., 1992; Saunders et al.,
1980). Secondly, although the sedimentation
rates, i.e. intrinsic settling velocities over depth in
mixed layers (Smith, 1982), of the fine-sized de-
tritus in a shallow lake like Loosdrecht can be
about 0.02 d - \ the net loss rate will be much
lower due to resuspension (GOllS, 1991; Gons
etal., 1986b; 1991). In epilimnia of all but very
2.0
~
~
1.6
If)
:::>
.....
.~

1.2
.....
L
Ql
"0

Phytoplankton to detritus ratios C

.....a 0.8
:.:
c
The maximum value of the specific growth rate .....'"
CJ. 0.4
a
(J.lmax) of cyanobacteria at about 20 °C is approx- .....
>-
imately 1.0 d - 1 (Van Liere & Mur, 1980; Rey-
o.o~;::~~~~~~~~
J::.
CJ.

nolds, 1987). P. hollandica has a J.lmax values of 0.0 0.2 0.4 0.6 0.8 1.0
0.52 d - 1 at 20°C (Burger-Wiersma et al., 1989). speci fie growth rate (~ D) (d- 1 )

Refractory fractions, estimated as the particulate
matter remaining after 21 days of aerobic decom-
position of O. limnetica, averaged 0.18 (Otten
et al., 1992). This information suffices for explor-
ing the phytoplankton to detritus ratio (R p / d ) as
Fig. 2. Dependence of the phytoplankton to detritus ratio on
the steady state growth rate of phytoplankton, for values of
the loss rate of detritus (d - I) as indicated. Loss rate com-
puted for fraction of refractory particulate matter = 0.18, see
text. Asterisk represents position of Lake Loosdrecht in sum-
mer 1988.
56
small deep lakes the sedimentation rates will be
much lower than 0.02 d - 1, and also in this case
the loss will be reduced by resuspension of litoral
deposits. Thirdly, water retention times are gen-
erally too long for hydraulic loss rates as high as
0.02 d - 1. In a seepage lake like Loosdrecht, this
loss will probably be lower than indicated by the
value of 0.006 d - 1 corresponding to the water
retention time of ca. half a year, because the par-
ticles may be retained at the sediment/water in-
terface.
Even though refractory fractions may greatly
depend on phytoplankton species and growth
conditions (Jewell & McCarty, 1971; Fallon &
Brock, 1979; Otten et al., 1992) it can be con-
cluded that in lakes in general and L. Loosdrecht
in particular the mean R p / d value will be less than
unity over almost the whole range of phytoplank-
ton growth rates. The proportions ofphytoplank-
ton and detritus may greatly vary seasonally in
non-steady-state situations. In Frains Lake
(Michigan) R p / d ranged from 0.06 to 0.77 (Saun-
ders, 1972).
The specific loss rate (J for Lake Loosdrecht
The parameter f3 is needed for partitioning the
seston dry weight into phytoplankton and detri-
tus based on measured values of CHL and S
(Eq. 2). For known specific growth rate of the
phytoplankton the value of (J is fixed (Eq. 4). The
values of f3 and J.1 are not independent of each
other.
The chlorophyll content of phytoplankton is
highly dependent on both energy and nutrient sta-
tus of the cells. Depending on mean irradiance
and growth rate, chlorophyll a contents of O. ag-
ardhii in phosphorus- and light-limited chemo-
stats varied from 3 to 8 and 5 to 17 mg g - 1 dry
weight, respectively (Zevenboom et aI., 1982).
Thus, for this organism, f3 may range 60 to 330.
The high value would imply that phytoplankton
constituted > 100% of the seston in L. Loos-
drecht. In light-limited turbidostat cultures of
P. hollandica, f3-values were 28 and 142 at mean
irradiances of 10 and 240 J.1E m - 2 S - 1 PAR
( = photosynthetically available radiation), re-
spectively (Burger-Wiersma & Post, 1989).
Should the high f3-value of these cultures apply to
the phytoplankton in L. Loosdrecht in summer
1988, two-thirds of the seston was phytoplank-
ton.
However, vertical light attenuation coefficients
in the lake in summer are so high (~ 5 m - 1), that
the euphotic depth is much less than 1 m, and the
depth- and time-averaged irradiance is less than
50 J.1E m - 2 S - 1 of PAR. For steady-state growth
in the 'natural' light conditions in the laboratory
scale enclosures (LSE), receiving 550 J.1E m - 2
S - 1 as averaged incident daylight (PAR) in sum-
mer, the range of f3 (including phaeopigment) was
relatively narrow. The minimum was 42 - for
P. hollandica under both light and phosphorus
limitation - and the maximum was 77 - for
O. limnetica under phosphorus limitation (Rijke-
boer et al., 1990b). The growth rate ranged 0.04
to 0.17 d - 1. The highest f3-value from these re-
sults would imply that the phytoplankton consti-
tuted one-third of the seston dry weight.
In summer 1988, the phytoplankton growth rate
estimates were made using only the LSE method
(Rijkeboer et al., 1990b). For both light- and
phosphorus-limited growth the derived values of
J.1 were 0.09-0.12 d - 1. These J.1-values were about
half of those obtained for 1987 and 1989 using the
same method. The reasons for annual variability
in J.1 remain to be investigated. In the summers of
1984-1986, both the protein change over 24 h
and the carbohydrate decrease in the dark gave
about 0.1 d - 1 as mean value of J.1 (Burger-
Wiersma & Baard, 1987).
In summary, the mean value of J.1 ( = D) in sum-
mer 1988 probably was close to 0.10 d - 1. At this
growth rate, the variation in f3 for light- and
phosphorus-limited steady state cultures of P. hol-
landica and O. limnetica was 42 to 50 (Rijkeboer
et aI., 1990b). A mean value of 46 implies that the
mean contribution of phytoplankton to the seston
dry weight in summer 1988 was 22 %. Assuming
the validity of Eq. (1), the remaining 78 % of the
seston dry weight was refractory fine-sized detri-
tus. Thus, the value of R p / d would be 0.28.
The validity of Eq. (1) is questionable because

part of the seston may be constituted by peat
fragments and mineral particles not associated
with phytoplankton decomposition. However, re-
membering the former clear state of the lake, with
limitation of the Secchi-depth by the bottom only,
it is reasonably safe to conclude that these ma-
terials cannot represent more than 10% of the
present mean of seston dry weight. Even a dry
weight concentration as low as 5 g m - 3 would
allow a Secchi-depth of only 1 m (GOllS, 1987).
No data are available to substantiate that the
former submerged vegetation dampened the
wind-resuspension of sediment particles.
The share of the refractory detritus could also
be over-estimated because the labile particulate
detritus has been ignored in Eq. (1). For steady-
state conditions in L. Loosdrecht the labile de-
tritus can be neglected, because the decomposi-
tion rate of this fraction often will be higher than
the phytoplankton growth rate (Otten et al., 1992)
and, consequently, its concentration must be very
low.
Thus, a phytoplankton to detritus ratio of 0.3
appears reasonable, and, from the value of J.1, the
position of L. Loosdrecht in summer 1988 can be
located (Fig. 2). The value of (J can be precised
using Eq. (6). For D (= J.1) = 0.1 d - " r = 0.18 and
R r d = 0.3, the yalue of (Jis 0.0054 d - '. This value
equals the decomposition rate of the refractory
matter remaining after three weeks of incubation
of O. lilllllerica in the dark (Otten er al., 1992). It
is too early. ho\yever, to conclude that other
causes of decline of the detritus ill situ are insig-
nificant.
The fare of rhe re(i'acroIT derrirus
Besides D and (J, the refractory fraction r controls
the coupling of phytoplankton and detritus
(Eq.4). Both the formation of the detritus
( = D[PH ]r) and the value of a(Eq. 6) are linearly
equated to r. Consequently, values of r greatly
influence predictions about formation and fate of
the fine-sized detritus. The value of r is likely to
reflect metabolic adaptation to environmental
conditions, and seems to be related to the growth
57
rate in particular (Otten et al., 1992). However,
the data needed to predict values of r in relation
to the growth conditions of the organisms are too
fragmentary to allow discussion.
For D (=J.1)=0.10d- 1 and r=0.18, the dry
weight of detritus produced (= D[PH]r) during
May-September 1988 was 22 g m - 3. Extrapola-
tion of (J= 0.0054 d - 1 to a whole year involves
that 3 g m - 3 would remain in the water column
one year later, which does not seem to be a bad
perspective for the lake's restoration. However,
this annual decrease of 'old' detritus, will be over-
estimated for two reasons. First, the decomposi-
tion component of (J is likely to gradually decline
with ageing of the particles, and also will decrease
with temperature in winter (Saunders et al., 1980).
Secondly, the deposition component of (Jwill vary
seasonally. Sinking rates of the particles decrease
with temperature due to change in viscosity, and
wind-resuspension occurs more frequently and in
larger areas of the lake from September through
April than in summer (GOllS et al., 1986b; GOllS,
1991). Therefore, the deposition rate in summer
can be considerably higher than the annual mean,
which has the effect of substantially lowering the
value of (J on an annual basis.
A general conclusion is that refractory detritus
may continue to significantly influence the water
transparency in L. Loosdrecht for years after its
origin, even after the dominance of the cyanobac-
teria and P. hollandica has ceased. Apparently,
the situation is very similar to the shallow L. Ve-
luwe (Hosper & Meijer, 1986). It remains to be
seen which mechanisms may accelerate the elim-
ination of the phytoplankton remnants in shal-
low, wind-exposed lakes.
Acknowledgements
We wish to thank Louis van Liere for the pleas-
ant co-operation. Acknowledgements are due to
Ronny van Keulen for technical assistance. This
research has been financially supported by the
Netherlands Ministry of Housing, Physical Plan-
ning and Environment, and the Commission of
the European Communities (EVj4V-0137-NL).
58
References
Berger, c., 1987. Habitat and ecology of Oscillatoria agardhii
Gomont. Ph.D. thesis, Univ. Groningen (in Dutch with
summary in English). 233 pp.
Burger-Wiersma, T. & L. R. Mur, 1986. Carbohydrate dy-
namics and growth rate of natural phytoplankton popula-
tions. Hydrobiol. Bull. 20: 87-92.
Burger-Wiersma, T. & R. Baard, 1987. Een veldecologisch en
experimentee1 onderzoek naar het gedrag van fytoplankton
in de Loosdrechtse Plassen. Internal report WOL 1987-6,
Laboratory for Microbiology, Univ. Amsterdam (in Dutch),
36pp.
Burger-Wiersma, T. & A. F. Post, 1989. Functional analysis
of the photosynthetic apparatus of Prochlorothrix hollandica
(Prochlorales), a chlorophyll b containing procaryote. Plant
Physiol. 91: 770-774.
Burger-Wiersma, T., L. J. Stal & L. R. Mur, 1989.
Prochlorothrix hollandica gen. nov., sp. nov., a filamentous
oxygenic photoautotrophic procaryote containing chloro-
phylls a and b: assignment to Prochlorotrichaceae fam. nov.
and order Prochlorales Florenzano, Balloni, and Materassi
1986, with emendation of the ordinal description. Int. J.
Syst. Bact. 39: 250-257.
Burger-Wiersma, T., M. Veenhuis, H. J. Korthals et al., 1986.
A new prokaryote containing chlorophylls a and b. Nature
(London) 320: 262-264.
De Kloet, W. A., P. J. Boesewinkel-de Bruyn & L. Breebaart,
1990. Phytoplankton and its production rates in the Loos-
drecht lakes before and after P-load reduction. Verh. int.
Ver. Limnol. 24: 715-718.
Engelen, G. B., B. F. M. Kal, J. J. Buyse & F. G. M. Van
Pruissen, (1992). The hydrology of the Loosdrecht lakes
area. Hydrobiologia 233: 21-38.
Fallon, R. D. & T. D. Brock, 1979. Decomposition of blue-
green algal (Cyanobacterial) blooms in Lake Mendota,
Wisconsin. Appl. env. Microbiol. 37: 820-830.
Gons, H. J., 1987. De relatie tussen doorzicht en slib in the
Loosdrechtse Plassen in verband met de zwemwaternorm.
Internal report WOL 1987-3, Limnoiogicaiinstitute, Nieu-
wersluis (in Dutch with abstract in English), 75 pp.
Gons, H. J., 1991. Diatom sedimentation loss reduction by
resuspension in shallow lakes. Verh. int. Ver. Limnol. 24:
729-733.
Gons, H. J., R. D. Gulati & L. Van Liere, 1986a. The eutro-
phic Loosdrecht lakes: current ecological research and res-
toration perspectives. Hydrobiol. Bull. 20: 61-75.
Gons, H. J., R. Van Keulen & R. Veeningen, 1986b. Effects
of wind on a shallow lake ecosystem: redistribution of par-
ticles in the Loosdrecht lakes. Hydrobiol. Bull. 20: 109-
120.
Gons, H. J., J. H. Otten & M. Rijkeboer, 1991. The signifi-
cance of wind resuspension for the predominance of fila-
mentous cyanobacteria in a shallow, eutrophic lake. Mem.
1st. ital. Idrobiol. 48: (in press).
Gons, H. J. & M. Rijkeboer, 1990. Algal growth and loss
rates in Lake Loosdrecht: first evaluation of the roles of
light and wind on a basis of steady state kinetics. Hydro-
biologia 191: 129-138.
Gulati, R. D., L. Van Liere & K. Siewertsen, 1991. The Loos-
drecht lake system: man's role in its creation, perturbation
and rehabilitation. In O. Ravera (ed) Terrestrial and aquatic
ecosystems: perturbation and recovery: 593-606. Ellis
Horwood, New York.
Gulati, R. D., A. L. Ooms-Wilms, O. F. R. Van Tongeren, G.
Postema & K. Siewertsen, 1992. The dynamics and role of
Iimnetic zooplankton in Loosdrecht lakes (The Nether-
lands). Hydrobiologia 233: 69-86.
Hofstra, J. J. & L. Van Liere, 1992. The state of the environ-
ment of the Loosdrecht lakes. Hydrobiologia 233: 11-20.
Hosper, S. H. & M. L. Meijer, 1986. Control of phosphorus
loading and flushing as restoration methods for Lake Ve-
luwe, The Netherlands. Hydrobiol. Bull. 20: 183-194.
Hutchinson, G. E., 1957. A treatise on limnology. Vol. 1.
Wiley, New York, 1015 pp.
Jewell, W. J. & P. L. McCarty, 1971. Aerobic decomposition
of algae. Env. Sci. Technol. 5: 1023-1031.
Loogman, J. G. & L. Van Liere (eds), 1986. Restoration of
shallow lake ecosystems with emphasis on Loosdrecht
lakes. Hydrobiol. Bull. 20: 5-259.
Malthus, T. J., E. P. H. Best & A. G. Dekker, 1990. An as-
sessment of the importance of emergent and floating-leaved
macrophytes to trophic status in the Loosdrecht lakes (The
Netherlands). Hydrobiologia 191: 257-263.
Moed, J. R. & G. M. Hallegraeff, 1978. Some problems in the
estimation of chlorophyll-a and phaeopigments from pre-
and post-acidification spectrophotometric measurements.
Int. Revue ges. Hydrobiol. 63: 787-800.
Moed, J. R. & H. L. Hoogveld, 1982. The algal periodicity in
Tjeukemeer during 1968-1978. Hydrobiologia 95: 223-234.
Otten, J. H., H. J. Gons & M. Rijkeboer, 1992. Dynamics of
phytoplankton detritus in a shallow, eutrophic lake (Lake
Loosdrecht, The Netherlands). Hydrobiologia 233: 61-68.
Reynolds, C. S., 1987. Cyanobacterial water-blooms. In J. A.
Callow (ed.), Advances in botanical research 13: 67-143.
Academic Press, London.
Rijkeboer, M. & H. J. Gons, 1990. Light-limited algal growth
in Lake Loosdrecht: steady state studies in laboratory scale
enclosures. Hydrobiologia 191: 241-248.
Rijkeboer, M., F. De Bles & H. J. Gons, 1990a. Laboratory
scale enclosure: concept, construction and operation. J.
Plankton Res. 12: 231-244.
Rijkeboer, M., F. De Bles & H. J. Gons, 1990b. Fosfor, licht
and sestondynamiek. Internal report WOL 1990-18. Lim-
noiogicaiinstitute, Nieuwersluis (in Dutch), 46 pp.
Rijkeboer, M., F. De Bles & H. J. Gons, 1991. The role of
sestonic detritus as a phosphorus buffer. Mem. 1st. ital.
Idrobiol 48: (in press).
Saunders, G. W., 1972. The transformation of artificial detri-
tus in lake water. Mem. 1st. ital. Idrobiol. 29 (Suppl.):
261-284.
Saunders, G. W., K. C. Cummins, D. Z. Gak etal., 1980.

Organic matter and decomposers. In E. D. Le Cren & R. H.
Lowe-McConnell (eds) The functioning of freshwater eco-
systems. Cambridge University Press, Cambridge: 341-
392.
Smith, I. R., 1982. A simple theory of algal deposition. Fresh-
wat. BioI. 12: 445-449.
Van Liere, L. & J. H. Janse, 1992. Restoration and resilience
to recovery of the Lake Loosdrecht ecosystem in relation
to its phosphorus flow. Hydrobiologia 233: 95-104.
Van Liere, L. & L. R. Mur, 1980. Occurrence of Osci/latoria
agardhii and some related species; a survey. In J. Barica &
L. R. Mur (eds), Hypertrophic ecosystems, Dev. Hydro-
bioI. 2: 67-78. Dr. W. Junk Publ., The Hague.
Van Liere, L., L. Breebaart & Y. J. Dullemont, 1989. Deter-
mining the relative number of prochlorophytes in lake phy-
toplankton using epifluorescence microscopy. Br. phycol. J.
24: 391-394.
Van Liere, L., O. F. R. Van Tongeren, L. Breebaart & W.
59
Kats, 1990. Trends in chlorophyll and total phosphorus in
Loosdrecht lakes, The Netherlands. Verh. int. Ver. Limnol.
24: 707-710.
Van Tongeren, O. F. R., L. Van Liere, R. D. Gulati, G.
Postema & N. Boesewinkel-de Bruyn, 1992. Multivariate
analysis of the plankton communities in the Loosdrecht
lakes: relationship with the chemical and physical environ-
ment. Hydrobiologia 233: 105-117.
Wetzel, R. G., 1983. Limnology. Saunders, Philadelphia (2nd
edition), 766 pp.
Wetzel, R. G., P. H. Rich, M. C. Miller & H. L. Allen, 1972.
Metabolism of dissolved and particulate detrital carbon in
a temperate hard-water lake. Mem. 1st. ital. Idrobiol. 29
(Suppl.): 185-241.
Zevenboom, W., A. Bij de Vaate & L. R. Mur, 1982. Assess-
ment of factors limiting growth rate of Osci/latoria agardhii
in hypertrophic Lake Wolderwijd. Limnol. Oceanogr. 27:
39-52.
Hydrobiologia 233: 61-67, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 61
© 1992 Kluwer Academic Publishers.

Dynamics of phytoplankton detritus in a shallow, eutrophic lake (Lake

Loosdrecht, The Netherlands)

Jerko H, Otten, Herman J. Gons & Machte1d Rijkeboer

Limnological Institute, 'Vijverhof Laboratory, Rijksstraatweg 6,3631 AC, Nieuwersluis, The Netherlands

Key words: phytoplankton, epipelon, Oscillatoria limnetica, detritus, shallow lakes, lake restoration

Abstract

A study was made of the mortality and aerobic decomposition of light- and phosphorus-limited cultures
of Oscillatoria limnetica, a dominant phytoplankton species in shallow, eutrophic Lake Loosdrecht (The
Netherlands), When placed in the dark at 20 °e, most cells died and lysed within twelve days. The labile
organic matter was completely decomposed within three weeks. Absorbance spectra indicated that blue
green algae may contributed significantly to the refractory dissolved substances in the lake. Refractory
particulate matter constituted from 7 to 24 % of the biomass of O. limnetica, depending on the growth
rate before incubation in the dark. The decomposition rate of this fraction was 0.005 d - I. On a basis
of a steady-state model of the dynamics of phytoplankton detritus, the areal organic dry weight con-
centration of the detritus in the lake is ca. 60 g m - 2, This means the quantities of detritus in the seston
and epipelon are about equal.

Introduction eutrophication (Loogman & Van Liere, 1986; Van

The role of detritus in the structure and function-
ing of aquatic ecosystems has been widely recog-
nized (Wetzel, 1983; Round, 1981; Saunders
et at., 1980). Both particulate and dissolved frac-
tions are important, and the total amount of de-
tritus is considerably higher than that of the biota.
In many lakes, the particulate detritus occurs in
aggregates that also include microorganisms and
inorganic matter. The detritus in these aggregates
is a complex mixture of remains of organisms in
various stages of decomposition.
This article describes how the detritus origi-
nates from the phytoplankton in Lake Loos-
drecht (The Netherlands). It is a shallow lake
(Area = 9.8 km 2 ; mean depth = 1,85 m) with a
very low Secchi-disk depth (ca. 0.3 m), caused by
massive phytoplankton growth following cultural
Liere et al., 1992). Phytoplankton species are
mainly blue-green algae, including Oscillatoria
limnetica and Prochlorothrix hollandica (De Kloet
et al., 1990; Van Liere et al., 1989; Van Tongeren
et al., 1992). Detritus rather than phytoplankton
contributes to the high seston concentration (dry
weight ca. 35 g m - 3), which reduces light pene-
tration into the water column. The underwater
light climate also is influenced by dissolved de-
tritus, i.e. blue light absorbing humic substances
(Gons etal., in press; Dekker etal., 1991;
Malthus & Dekker, 1990). Submerged vegetation
is lacking, but floating-leaved plants occur along
the· shores (Malthus et at., 1990),
The sources of detritus are diverse. The more
recently formed detritus is associated mainly with
phytoplankton dynamics, and occurs as fine sized
particles in the water (Gons et al., 1991; 1992).
62
The majority of the detritus occurs in particle
aggregates in the sediment/water interface, the
epipelon. It includes fragments of eroded peat,
settled algae and fine-sized detritus. The detritus
is resuspended frequently over large areas of the
lake (Gons et al., 1991; 1986b). The peat frag-
ments settle again in a few hours but the fine-
sized detritus remains in the water column for
several weeks after resuspension.
Algal mortality is caused by different factors.
Besides the zooplankton grazing in the water layer
(Gulati et aI., 1992), sedimented algae on the lake
bottom die after a few days (Otten & Gons, 1991a)
due to lack of light. Microfauna, bacteria and
viruses may also contribute to algal mortality in
the epipelon (Daft & Steward, 1973; Granhall &
Berg, 1972).
Little is known about the inputs of dissolved
detritus in L. Loosdrecht. The main sources are
likely to be decomposing phytoplankton and
aquatic plants. Allochthonous inputs include
humic substances leaching from the lake's peaty
substratum and those formed during decomposi-
tion of terrestrial plants (Christman & Gjessing,
1983; Wetzel, 1983; De Haan, 1982), entering the
lake via run-off.
Several stages can be distinguished in the de-
composition of phytoplankton (Fig. 1; Otten &
Gons, 1991 a, b). After the death of cells, part of
the biomass is released into the DOM pool. The
labile
organic
matter
phytoplankton f---+--"'-- DOM
'--re-;f~r a-,c"7""to--:r-y----,
organic
matter
Fig. 1. Model of dynamics of particulate detritus (Otten &
Gons, 1991b). Mortality = phytoplankton death rate coeffi-
cient (time - 1); L = fraction of biomass released as dissolved
substances at cell lysis; II = labile fraction of biomass; 12 =
refractory fraction of phytoplankton biomass; k I = decom-
position rate coefficient of labile detritus (time - I); k 2 = de-
composition rate coefficient of refractory detritus (time - 1);
DOM = dissolved organic matter.
larger fraction of the remaining particulate matter
is labile (i.e. readily utilized) whereas the rest is
refractory (i.e. hardly utilized by heterotrophic
microorganisms). Not included in Fig. 1 is the
origin of refractory dissolved compounds in the
decomposition of the particulate phases.
Refractory material appears to constitute the
major fraction of detritus in the lake, and the
accumulation of refractory detritus in the epipelon
has been experimentally confirmed (Otten &
Gons, 1991a). Due to its slow decomposition rate,
this detritus might retard the lake's restoration
(Gons et aI., 1992). Therefore, it is of special in-
terest to obtain data on the refractory fraction of
phytoplankton, its decomposition rate and its dis-
tribution in the lake. For this study, on the origin
and fate of phytoplankton detritus, an experimen-
tal approach was chosen, using cultures of O. lim-
netica.
Materials and methods
Growth of Oscillatoria limnetica
Oscillatoria limnetica (isolate Limnological Insti-
tute) was grown in synthetic medium in a labo-
ratory scale enclosure (LSE) operated as a
chemostat (Rijkeboer et al., 1990a). The condi-
tions simulated averaged in situ summer condi-
tions with respect to incident light (120 W m - 2
PAR), pH (8-8.5) and water temperature
(20 0e). The cultures were non-axenic. Phospho-
rus was supplied to the algal cultures at a rate of
105 J.lg P 1- I d - 1 (light-limited algal growth in
steady state), or at a rate of 7.5 J.lg P 1- 1 d - I
(phosphorus limited algal growth in steady state).
Refractory fractions
Refractory fractions of algal biomass were esti-
mated in batch experiments after transferring
phosphorus- or light-limited O. limnetica suspen-
sions to 5 I vessels. The suspensions were kept in
the dark for three weeks at 20 ° C. Each experi-
ment was performed in duplicate. Conditions
 63

were kept aerobic by aeration with water-
saturated air. The pH remained close to 8 during
these experiments. After three weeks, the partic-
ulate detritus appeared to be decomposed at a
very low rate. The remaining particulate material
was defined as the refractory part of the particu-
late matter initially present. Refractory fractions
of seston of L. Loosdrecht were estimated after
incubating lake water in duplicate as described
above.
Dissolved detritus
A nutrient-saturated batch culture of O. limnetica
was kept in the dark at 20 0 C during three weeks.
Absorbance spectra of pre-filtered samples
(0.4511m Mil1ipore HA) were measured at 2 nm
intervals using a Philips PU 8600 UVjVIS spec-
trophotometer. Dry weight concentrations and
absorbance spectra of non-filtered suspensions
were also measured. In the latter case particle
scattering caused part of the absorbance.
Decomposition of refractory detritus
A 12 I continuous flow system, designated
epipelon trap (ET), was fed with effluent from an
O. limnetica culture at about 400 ml h - I (Otten &
Gons, 1991a, b). The ET was kept in the dark at
20°C and aerated to maintain the dissolved ox-
ygen content of the water close to saturation. Res-
piration was measured daily by monitoring the
dissolved oxygen change during interruption of
the aeration and the dilution of the system (Otten
& Gons, 1991b), and the layer of detritus was
sampled weekly. After six weeks the inflow of
culture effluent was stopped and the ET diluted
ometry for complete oxidation ofC6 H 12 0 6 (Otten
& Gons, 1991b).
Results and discussion
Refractory fractions ofphytoplankton and seston
After incubating a suspension of steady-state
light-limited O. limnetica (specific growth rate
11 = 0.083 d - I) in the dark at 20 0 C, almost all
algae died within twelve days. Absorbance spec-
tra and dry weight concentrations of the suspen-
sion were measured to establish the refractory
fraction of the particulate matter and to study the
transformation of materials from the particulate
into the dissolved phase.
Absorbance of the particles (Fig. 2) was ob-
tained by subtracting the absorbance of 0.45 11m
filtered water from the total absorbance. The ab-
sorbance above 700 nm is mainly due to particle
scattering. In the spectra, until day 9, not only the
absorption peaks of chlorophyll a (around 440
and 680 nm) but also that of phycoerythrin
(around 570 nm) are pronounced. Absorption by
C-phycocyanin (around 630 nm) was rather low.
The sudden decrease of absorbance between
day 9 and 12 was caused by mortality and lysis
of most O. limnetica cells. Except around 680 nm,
the presence of photosynthetic pigments was no
longer indicated. From day 15 onwards, the ab-
150
125
;:;--
I
E 100
CD
u
c 75
D
co

with synthetic medium instead. The labile detri- L

0
If)
50
tus formed during these first six weeks decom- D
co
25
posed within two to three weeks, thus only re-
fractory detritus remained. From then on, the 0
measured oxygen consumption was due to respi- 400 500 600 700
blue yellow red nm
ration by microorganisms decomposing refrac- Fig. 2. Absorbance spectra of particulate matter during decay
tory detritus. Decomposition rate coefficients (k]> of Oscillatoria limnetica. Time course indicated by arrow and
see Fig. 1) were calculated assuming the stoichi- day numbers.
 64

sorbance decreased only slightly: mainly refrac-
tory detritus was left and it decomposed at a
much lower rate.
The decomposition process shows a sigmoid
growth rate, but is quite insensitive to the type of
growth limitation.
Seston from L. Loosdrecht was incubated in
the dark at 20 C under aerobic conditions and
0

course (Fig. 3; upright triangles), considering the
dry weight concentration vs. time data. The data
points until day 9 represent the stage character-
ized by respiration and fragmentation of O. lim-
netica preceeding death (Otten & GOllS, 1991b).
Cell lysis and rapid decomposition of the labile
fraction occurred mainly around day 11. After
another ten days of aerobic decomposition the
refractory fraction left was 0.24 of the initial bi-
omass. The experiment was repeated with phos-
phorus-limited algae at the same growth rate
(fJ. = 0.083 d - I), the refractory fraction being
again one-quarter of the initial biomass (Fig. 3;
upside-down triangles).
Refractory fractions of freshwater algae vary
from 0 to 0.86 (Fallon & Brock, 1979; Jewell &
McCarthy, 1971), depending on algal species
(Foree & McCarty, 1970; Jewell & McCarty,
1971; Tezuka, 1989), temperature and the pre-
vailing communities of bacteria and zooplankton
(Jewell & McCarty, 1971). Earlier, we found the
low value of 0.07 for light-limited O. limnetica at
fJ. = 0.17 d - I (Otten & Gons, 1991a). Our study
indicates that the proportion of the refractory
fraction of phytoplankton also depends on the
100
. ment with light-limited O. limnetica, described
"-
~
-0
90
I .
80
E t
-0"
70
the dry weight concentration was measured
over three weeks. This experiment was repeated
five times in 1990 and 1991 (Table 1). The par-
ticulate refractory fraction of the sestonic matter
ranged from 0.50 to 0.63: no seasonality has been
found so far. The biomass of non-blue-green algal
species in the lake was negligible (P. J.
Boesewinkel-de Bruyn, pers. comm.). Thus the
phytoplankton biomass constituted maximally 37
to 50 % of the total seston dry weight during the
study period 1990/1991. These percentages indi-
cate that the relative share of phytoplankton was
higher than that estimated for summer 1988
(GOllS et al., 1992).
Dissolved detritus
Organic matter mainly enters the dissolved phase
during lysis of the phytoplankton cells (Fig. 1) but
there is also some contribution during further de-
composition of the particles. Part of the dissolved
organic matter is available to free-living and at-
tached bacteria, whereas the rest can be consid-
ered to be refractory humic compounds (Geller,
1986).
Absorbance spectra of filtrates in the experi-
above, showed a marked shift towards the blue

:J 60
Table I. Refractory detritus fractions of O. limnetica and
~0 50
seston from L. Loosdrecht. Steady-state specific growth rate
a.

~c
40
... • and the growth-limiting factor are given in parenthesis.

:,
30

20
U .
Source Fraction
(;
<!R 10
O. limnetica (0.083 d -1; phosphorus) 0.24
0 O. limnetica (0.083 d - 1; light) 0.24
0 4 8 12 16 20 24 28
O. limnetica (0.17 d- 1 ; light) 0.07
time (days) Seston (Aug. 1990) 0.50
Seston (Oct. 1990) 0.55
Fig. 3. Time course of dry weight concentration during decay Seston (Nov. 1990) 0.53
of light-limited (upright triangles) and phosphorus-limited Seston (July 1991) 0.63
(upside-down triangles) Osci//atoria limnetica. Dotted line in- Seston (Aug. 1991) 0.50
dicates the refractory matter as % of the initial biomass.

(Fig. 4) in comparison with the spectra of the
particles (Fig. 2). The chlorophyll a peak in the
red was still present but not that in the blue wave-
length band. The absorption peak of phyco-
erythrin was also absent. Differences with ab-
sorption spectra offiltered lake water (Gons et al.,
1992; Malthus & Dekker, 1990) are the absorp-
tion peak of chlorophyll a in the red band and the
flattening near 400 nm in the spectra in Fig. 4.
These characteristics indicate that the filtrates in
the experiment contained very small particles be-
sides true dissolved compounds.
During the experiment the mainly in blue-light
absorbing dissolved substances increased
(Fig. 4). This increase was greatest in the period
of lysis of the organisms but continued until the
end of the experiment. The absorption coefficient
of filtered lake water at 440 nm is 1-1.5 10 - I. In
the experiment an absorbance at 440 nm of 8 was
reached, which corresponds to an absorbtion co-
efficient of about 3. The initial dry weight con-
centration of the O. limnetica suspension was
45 g 10 - 3, which is equivalent to the phytoplank-
ton production during 15 to 30 days in situ (Ri-
jkeboer et al., 1990b).
Further kinetic studies, including fluxes of al-
lochthonous humic substances, are needed to es-
tablish whether a causal relationship underlies the
positive correlation between the growth of O. lim-
nelica and the concentration of dissolved organic
12 5 - r - - - - - - - - - - - - - - - - - - ,
400 500 600 700
blue yellow red nm
Fig. 4. Absorbance spectra of dissolved matter during decay
of Oscillatoria limnetica. Time course indicated by arrow and
day numbers.
65
carbon in situ (Van Tongeren et al., 1992). Also
needed is a qualitative comparison of the dis-
solved detritus derived in the laboratory with that
in the lake (De Haan et al., 1990; De Haan, 1983).
Decomposition rate of particulate detritus
The oxygen uptake of refractory detritus derived
from O. limnetica was measured using the
epipelon trap (see Methods). From these results,
the decomposition rate coefficient (k2 ; see Fig. 1)
was calculated. After three weeks of incubation of
O. limnetica in the dark the remaining particulate
matter was decomposed very slowly. The decom-
position rate of this refractory detritus was
0.005 d - 1 at 20 C, implying that 16 % could still
0
be present after one year. Because the decompo-
sition is likely to slow down later (Saunders et al.,
1980) and because decomposition is slower at
lower temperatures, these particles may influence
the water quality for several years after their or-
igm.
As has been shown above, the major part of the
biomass after mortality flows into the pool of la-
bile detritus, which has a quite high decomposi-
tion rate (k j ; see Fig. 1): varying from 0.06 to
0.19 d - I, depending on algal growth conditions
(Otten & Gons, 1991a, b). Because these
decomposition rates are more than an order of
magnitude higher than those of the refractory de-
tritus, the labile detritus in general will be a small
part of the total of phytoplankton detritus in L.
Loosdrecht, like in many other lakes (Wetzel,
1983).
Dynamics ofparticulate detritus in L. Loosdrecht
To evaluate the impact of the detritus on the un-
derwater light climate and nutrient cycling, it is
important to estimate the areal concentration of
the detritus in both seston and epipelon
([ DA] gm - 2). According to the model for dynam-
ics of detritus (Fig. 1, and Otten & Gons, 1991b),
the steady-state value of D A in L. Loosdrecht is
given approximately by the quotient of formation
66
of refractory detritus (g m - 2 d - I) and the de-
composition rate k 2 (d - I):
[PH] z Dr
[D A] = -'----"---- (1)
k2
where
[PH] = phytoplankton dry weight concentration
(gm- 3 ),
z = mean water depth (m),
D = phytoplankton loss rate (d - I), and
r = refractory fraction of the phytoplankton
dry weight.
In L. Loosdrecht an oscillating steady state of
phytoplankton and detritus may exist from May
to September (Gons et al., 1992). At steady-state
the combined loss rates (D) equal the growth rate
(fJ.) of the phytoplankton. Using laboratory scale
enclosures (LSE), Rijkeboer et al. (1990b) esti-
mated phytoplankton growth rates ranging from
0.09 to 0.24 d - I during the summers 1987-1989.
Also the growth of O. limnetica and P. hollandica
in LSE were studied at incident irradiances sim-
ilar to those in situ in summer. For fJ. = 0.09-
0.17 d - I at both light- and phosphorus-limited
growth, the content of chlorophyll a (including
phaeopigment) of O. limnetica and P. hollandica
were 19 ± 1.5 and 24 ± 0.5 mg g- I dry weight, re-
spectively. For equal biomass of these organisms
(Van Liere et al., 1989) the mean dry weight con-
centration of phytoplankton in L. Loosdrecht in
the recent summers was about 9 g m - 3.
Using Eq. (1) the areal concentration ofrefrac-
tory detritus in the lake derived from phytoplank-
ton can now be estimated. We obtained a value
of 0.005 d - I for k 2 . By substituting for D ( = fJ.)
and r minimum values of 0.09 d - I and 0.07, and
maximum values of 0.24 d - 1 and 0.24, [DA] is
calculated as 21 and 192 g m - 2, respectively.
The extremes of this range are unlikely for the
following reasons. The areal concentration of
sestonic dry weight in L. Loosdrecht in summer
is about 50 g m - 2. Consequently, the minimum
estimate of [D A] (21 g m - 2) might be compatible
with the concentration and estimated refractory
fractions of seston (Table 1) only when the seston
consists largely of other detritus than that origi-
nating in phytoplankton. Quantities of particulate
detritus from other sources cannot be very large
(Gons et aI., 1992). On the other hand, the max-
imum estimate of 192 g m - 2 is contradictory to
the observations that show that fine-sized detri-
tus is a small part of the epipelic matter (Gons
et al., 1990). The estimate can be improved as the
value of r appeared to depend greatly on the phy-
toplankton growth rate (Table 1). By substituting
in Eq. (1) D = 0.09 d - I together with r = 0.24, and
D = 0.24 d - I with r = 0.07, the range of [D A] is
reduced to 72 to 56 g m - 2. These concentrations
imply that the phytoplankton detritus in L. Loos-
drecht is about equally distributed between seston
and epipelon. Such a distribution is consistent
with the small increase in seston concentration
after wind resuspension (Gons et al., 1986b;
1991) and might prove to be a characteristic of
other 'typical Dutch' lakes. Although sedimenta-
tion rates of the fine-sized particles are much
higher than their decomposition rate, the frequent
resuspension prevents their accumulation on the
lake bottom. Paradoxically, the greatest share of
the phytoplankton detritus may originate in mor-
tality in the epipelon.
Acknowledgements
This research has been financially supported by
the Commission of the European Communities
(EVj4V-0137-NL) and the Netherlands Ministry
of Housing, Physical Planning and Environment.
References
Christman, R. F. & E. T. Gjessing (eds), 1983. Aquatic and
terrestrial humic materials. Michigan, Ann Arbor Science.
Daft, M. J. & w. D. P. Stewart, 1973. Light and electron
microscope observations on algal lysis by bacterium CP 1.
New Phytol. 72: 799-808.
De Haan, H., 1982. Physico-chemical environment in Tjeu-
kemeer with special reference to speciation of algal nutri-
ents. Hydrobiologia 95: 205-221.
De Haan, H., 1983. Use of ultraviolet spectroscopy, gel fil-
tration, pyrolysis/mass spectrometry, and numbers of
benzoate-metabolizing bacteria in the study ofhumification
and degradation of aquatic organic matter. In: R. F. Christ-

man and E. T. Gjessing (eds), Aquatic and terrestrial humic
materials. Michigan, Ann Arbor Science, pp. 165-182.
De Haan, H., T. De Boer, J. Voerman, H. A. Kramer &
O. F. R. van Tongeren, 1990. Size class distribution of
dissolved ( < 200 nm) nutrients and essential metals in shal-
low eutrophic humic lakes. Verh. int. Ver. Limno!. 24: 298-
301.
Dekker, A. G., E. Seyhan & T. J. Malthus, 1991. Quantita-
tive modelling of inland water quality for high resolution
MSS-systems. Trans. Geosc. Remote Sensing 29: 89-95.
De Kloet, W. A., P. J. Boesewinkel-de Bruyn & L. Breebaart,
1990. Phytoplankton and its production rates in the Loos-
drecht lakes before and after P-load reduction. Verh. int.
Ver. Limno!. 24: 715-718.
Fallon, R. D. & T. D. Brock, 1979. Decomposition of blue-
green algal (Cyanobacterial) blooms in lake Mendota, Wis-
consin. App!. Envir. Microbio!. 37: 820-830.
Foree, E. G. & P. L. McCarthy, 1970. The rate and extent of
algal decomposition in anaerobic waters. Proc. 24th Ind.
Waste Conf. 1969, Purdue University, Indiana.
Geller, A., 1986. Comparison of mechanisms enhancing bio-
degradability of refractory lakewater constituents. Limno!.
Oceanogr. 31: 755-764.
Gons, H. J., R. D. Gulati & L. Van Liere, 1986a. The eutro-
phic Loosdrecht Lakes: current ecological research and
restoration perspectives. Hydrobio!. Bull. 20: 61-75.
Gons, H. J., R. Veeningen & R. Van Keulen, 1986b. Effects
of wind on a shallow lake ecosystem: redistribution of par-
ticles in the Loosdrecht Lakes. Hydrobio!. Bull. 20: 109-
120.
Gons, H. J., J. H. Otten & M. Rijkeboer, 1991. The signifi-
cance of wind resuspension for the predominance of fila-
mentous cyanobacteria in a shallow, eutrophic lake. Mem.
1st. Ita!. Idrobio!. 48:
Gons, H. J., T. Burger-Wiersma, J. H. Otten & M. Rijkeboer,
1992. Coupling of phytoplankton and detritus in a shallow,
eutrophic lake (Lake Loosdrecht, The Netherlands). Hy-
drobiologia 233: 51-59.
Granhall, U. & B. Berg, 1972. Antimicrobial effects of Cell-
vibrio on blue-green algae. Arch. Hydrobio!. 84: 234-242.
Gulati, R. D., A. L. Ooms-Wilms, O. F. R. Van Tongeren, G.
Postema & K. Siewertsen, 1992. The dynamics and role of
limnetic zooplankton in Loosdrecht lakes (The Nether-
lands). Hydrobiologia 233: 69-86.
Jewell, W. J. & P. L. McCarty, 1971. Aerobic decomposition
of algae. Envir. Sci. Techno!. 5: 1023-1031.
Loogman, J. G. & L. Van Liere (eds), 1986. Restoration of
67
shallow lake ecosystems with emphasis on Loosdrecht
lakes. Hydrobio!. Bull. 20: 1-269.
Malthus, T. J., E. P. H. Best & A. G. Dekker, 1990. An as-
sessment of the importance of emergent and floating-leaved
macrophytes to trophic status in the Loosdrecht lakes (The
Netherlands). Hydrobiologia 191: 257-263.
Malthus, T. J. & A. G. Dekker, 1990. Spectral light attenu-
ation in a hypertrophic lake system (Loosdrecht Lakes, The
Netherlands). Verh. int. Ver. Limno!. 24: 711-714.
Otten, J. H. & H. J. Gons, 1991a. Aerobic decomposition of
settled algae in a laboratory system: the impact of resus-
pension on microbial activity. Verh. int. Ver. Limno!. 24:
734-737.
Otten, J. H. & H. J. Gons, 1991b. Detritus formation from
blue-green algae. Mem. 1st. Idrobiol. Ital. 48:
Rijkeboer, M., F. De Bles & H. J. Gons, 1990a. Laboratory
Scale Enclosure: concept, construction, and operation. J.
Plankton Res. 12: 231-244.
Rijkeboer, M., F. De Bles & H. J. Gons, 1990b. Fosfor, licht
en sestondynamiek. WQ L-report 1990-18, Limnological
Institute, The Netherlands. 46 pp. (in Dutch).
Round, F. E., 1981. The ecology of algae. Cambridge Uni-
versity Press, 653 pp.
Saunders, G. W., K. W. Cummins, D. Z. Gak, E. Pieczyn-
ska, V. Straskrabova & R. G. Wetzel, 1980. Organic matter
and decomposers. In: E. D. LeCren & R. H. Lowe-
McDonnel (eds), The functioning of freshwater ecosystems.
Cambridge University Press, pp. 341-392.
Tezuka, Y., 1989. The C:N:P ratio of Microcystis and
Anabaena (Blue-green Algae) and its importance for nutri-
ent regeneration by aerobic decomposition. Jap. J. Limnol.
50: 149-155.
Van Liere, L., L. Breebaart & Y. J. Dullemont, 1989. Deter-
mining the relative number of prochlorophytes in lake phy-
toplankton using epifluorescence microscopy. Br. phycol. J.
24: 391-394.
Van Liere, L., S. Parma & R. D. Gulati, 1992. Working group
Water Quality Research Loosdrecht Lakes: its history,
structure, research programme and some results. Hydro-
biologia 233: 1-9.
Van Tongeren, O. F. R., L. Van Liere, R. D. Gulati, G.
Postema, N. Boesewinkel-de Bruyn, 1992. Multivariate
analysis of the plankton communities in Loosdrecht lakes:
relationship with the chemical and physical environment.
Hydrobiologia (this issue)
Wetzel, R. G., 1983. Limnology. Philadelphia, Saunders (2nd
edn) 766 pp.
Hydrobi%gia 233: 69-86, 1992.
L. Vall Liere & R.D. Gu/ati (eds), Restoratioll alld Recovery of Shallow Eutrophic Lake Ecosystems ill The Nether/allds. 69
© 1992 K/uwer Academic Publishers.

The dynamics and role of Iimnetic zooplankton in Loosdrecht lakes (The

Netherlands)

R. D. Gulati, A. L. Ooms-Wilms, O. F. R. Van Tongeren, G. Postema & K. Siewertsen

Limnological Institute, Vijverhof Laboratory, Rijksstraatweg 6,3631 AC Nieuwersluis, The Netherlands

Key words: eutrophication, lake restoration, zooplankton grazing, rotifers, crustaceans, P-Ioading,
P-excretion

Abstract

The paper summarizes the results of a ten-year (1981-1991) zooplankton research on the Lake Loos-
drecht, a highly eutrophic lake. The main cause of the lake's eutrophication and deteriorating water
quality was supply up to mid 1984 of water from the River Vecht. This supply was replaced by dephos-
phorized water from the Amsterdam-Rhine Canal in 1984. The effects of this and other restoration
measures on the lake's ecosystem were studied. Despite a reduction in the external P-Ioad from ca. 1.0 g
P m - 2 Y- 1 to ca. 0.35 g m - 2 Y- 1 now, the filamentous prokaryotes, including cyanobacteria and
Prochlorothrix, continue to dominate the phytoplankton.
Among the crustacean plankton Bosmina spp, Chydorus sp. and three species of cyclopoid copepods
and their nauplii are quite common. Though there was no major change in the composition of abun-
dant species, Daphnia cucullata, which is the only daphnid in these lakes, became virtually extinct since
1989. Among about 20 genera and 40 species of rotifers the important ones are: Anuraeopsis fissa,
Keratella cochlearis, Filinia longiseta and Polyarthra. The rotifers usually peak in mid-summer following
the crustacean peak in spring. The mean annual densities of crustaceans decreased during 1988-1991.
Whereas seston « 150 J.1m) mean mass in the lake increased since 1983 by 20-60%, zooplankton
(> 150 J.1m) mass decreased by 15-35%.
The grazing by crustacean community, which was attributable mainly to Bosmina, had mean rates
between 10 and 25% d- 1 • Between 42 and 47% of the food ingested was assimilated. In spring and early
summer when both rotifers and crustaceans have their maximal densities the clearance rates of the
rotifers were much higher. Based on C/P ratios, the zooplankton (> 150 J.1m) mass contained 2.5 times
more phosphorus than seston « 150 J.1m) mass so that the zooplankton comprised 12.5% of the total-P
in total particulate matter in the open water, compared with only 4.5 % of the total particulate C. The
mean excretion rates of P by zooplankton varied narrowly between 1.5 and 1.8 J.1g P 1- 1 d - \ which
equalled between 14 and 28 % d - 1 of the P needed for phytoplankton production.
The lack of response to restoration measures cannot be ascribed to one single factor. Apparently, the
external P-Ioading is still not low enough and internal P-loading, though low, may be still high enough
to sustain high seston levels. Intensive predation by bream is perhaps more important than food qual-
ity (high concentrations of filamentous cyanobacteria) in depressing the development of large-bodied
zooplankton grazers, e.g. Daphnia. This may also contribute to resistance of the lake's ecosystem to
respond to rehabilitation measures.
70

Introduction

There are only a few long-term studies in Europe

dealing with the response of zooplankton to res-
toration measures (e.g. see in Faafeng & Nilssen,
1981; Gulati, 1990; Ewald, 1991). The present
studies on zooplankton in Loosdrecht lakes were
started in July 1981 in anticipation of the planned
restoration of the lakes (Van Liere et al., 1992). A
five-year (1982-1986) study on zooplankton in
three shallow and highly eutrophic lakes in the
Loosdrecht area did not reveal any significant
changes following the restoration measures,
which have resulted in considerable reduction in
external phosphorus loading since mid-1984
(Gulati, 1990). The research work
on the lakes was continued uninterrupted until
1991 when the results were presented at the
concluding 'Water Quality research Loosdrecht
lakes' (WQL) symposium (Van Liere & Gulati,
1992).
The major aim of the zooplankton research in Fig. 1. Location and surroundings of the Lake Loosdrecht.
Loosdrecht lakes has been: 1) to provide base-
line data needed for understanding the structure
and functioning of these lake ecosystems, and The lake
2) to examine changes in zooplankton dynamics,
especially those in response to the restoration The zooplankton study in the Lake Loosdrecht,
measures. Several aspects of zooplankton in the the largest among lakes of the Loosdrecht area
lake have been investigated and published. These (Fig. 1) was started in July 1981. Important mor-
included changes in zooplankton densities and phometric and limnological features of the Lake
grazing, lake's trophic status and food quality Loosdrecht are summarized in Table 1. For a de-
(Gulati, 1983, 1984; Gulati et al., 1985); food se- tailed account of the creation, history of forma-
lection among cladocerans (Irvine, 1986; Gulati tion, morphology and hydrology of Loosdrecht
etal., 1991a); feeding in rotifers (Gulati etal., lakes the reader is referred to papers in Loogman
1987, 1992; Ooms-Wilms, 1991; Ooms-Wilms & Van Liere (1986) and in Van Liere & Gulati
et al., 1992; Wilms et al., 1991); carbon and phos-
phorus contents and relationships with food, and
phosphorus regeneration by zooplankton (Den Table 1. Morphometry and general limnological characteris-
Oude & Gulati, 1988; Gulati et al., 1989, 1991c). tics of the Lake Loosdrecht (data from various sources).
The main objective of this paper is to present
Area 979 ha
an integrated picture of all aspects of zooplank-
Mean depth 1.85 m
ton studied in the Lake Loosdrecht between 1981 Temperature 11.3 ± 6.1 °C
and 1991, with special remarks on changes. Water residence time 0.6 year
External P-Ioading 0.35 g P m- 2 y-l
Total-P 89 ± 22 mg m - 3
Chlorophyll 93 ± 18 mg m - 3
Secchi disc 0.3-0.5 m

(1992). Recently, Gulati et al. (1991b) have pro-
vided a concise account of man's role in the cre-
ation and perturbation and the recent attempts to
rehabilitate these lakes. \ techniques used are given in Gulati (1990). The
The lakes are situated about 20 km southeast
of Amsterdam (Lat. 52° 20' N; Long. 5° 5' E) in
the province of Utrecht (Fig. 1). They are stripped
Sphagnum bogs. The first indications of the ex-
istence of the Loosdrecht lakes date back to early
seventeenth century when dredging of peat under
the water surface was formalized in legislation.
Precipitation and water pumped from the Polder
Bethune on the south and from the River Vecht
(up to 1983) on the west formed up to two-thirds
of the lake's water input. Evaporation and infil-
tration (seepage), mainly to the Polder Bethune,
are the main sources of water loss from the Loos-
drecht lakes.
The water, generally supplied during short pe-
riods in summer, from the River Vecht, was the
main cause of the lakes' eutrophication and de-
teriorating water quality. This supply was stopped
in 1984 and replaced by water from the Amster-
dam-Rhine Canal. This water was first stripped
of its phosphorus content by treating with FeCI 3 .
6H 2 0 in a treatment basin (Fig. 1). This main
restoration measure has led to a considerable re-
duction of the external P-loading of the Lake
Loosdrecht since 1984, viz. from about 1.1 g P
m- 2 y-l to about 0.35g P m- 2 y-l now (Van
Liere & J anse, 1992).
The Lake Loosdrecht is dominated by fila-
mentous prokaryotes, cyanobacteria and Pro-
chlorothrix hollandica; the former persist through-
out the year reaching concentrations of 200 x 10 3
filaments 1- 1 or more in summer (De Kloet et al.,
1991). The mean seston ( < 150 11m) mass varies
between 8 and 12 mg C 1- 1 (Gulati, 1990), about
two-thirds being detritus (Gons & Rijkeboer,
1990). The seasonal means of chlorophyll con-
centration ranged from 100 to 155 I1g 1- \ and the
mean primary production rates, based on 14C_
assimilation, from 1.7 to 2.5 mg C m - 2 d - 1 (De
Kloet et al.,1991).
71
Methods
The details of sampling procedures and counting
lakes were sampled routinely every four weeks
but fortnightly during 1983, 1984 and 1985. A
30-L composite sample was made up of five-litre
samples taken from a drifting boat at six stations.
The changes in the zooplankton densities and
environmental variables were investigated apply-
ing Principal Component Analysis (PCA) to the
log-transformed zooplankton species counts,
computing linear correlations of the environmen-
tal (independent) variables, with PCA axes (for
details see in Van Tongeren et ai., 1992). The
zooplankton densities in 1985 were compared
with those measured by Geelen (1955) in 1955 in
the Lake Loosdrecht.
The zooplankton (> 150 11m) mass measure-
ments refer to mainly crustaceans. The 'contam-
ination' of this mass by rotifers, coarser algae and
detritus was insignificant, except during the oc-
casional stormy periods when resuspension of the
peaty sediments did not allow such community
biomass measurements. Both the zooplankton
mass and seston « 150 11m) mass were deter-
mined as COD and converted to carbon as out-
lined in Gulati (1990). Setting the mean values of
the annual biomasses in 1983 at 100% the data
of 1981-1991 were compared. The biomass data
of 1985 were compared with those calculated from
Geelen (1955) for 1955.
Carbon and phosphorus in the zooplankton
and seston fractions were measured routinely
starting from 1983 (Gulati et ai., 1991c). Phos-
phorus regeneration by zooplankton was mea-
sured in the laboratory (Den Oude & Gulati,
1988) to assess the role of zooplankton in the
internal P-loading in the lake; this was done for
the period 1987-1990. From the weight-specific
P-excretion rates (0.17 I1g P mg - 1 DW h - I) cor-
rected for temperature, P-excretion rates (Den
Oude & Gulati, 1988) for the entire study period
(1981-1991) were computed.
Grazing rates were determined using the 14C_
technique (Gulati et al., 1982), both for the zoop-
lankton (> 150 11m) community in 1981-1984
72

(Gulati, 1984), and separately for the dominant
filter-feeders (Irvine, 1986; Gulati et al., 1991a).
Also, rotifer grazing measurements have been at-
tempted for the dominant species, viz. Anuraeop-
sisfissa (Wilms etal., 1991; Ooms-Wi1ms, 1991)
and Euchlanis dilatata lucksiana (Gulati et al.,
1987; Gulati et al., in press). The clearance rates
of the dominant species of both c1adocerans and
rotifers were measured simultaneously using latex
spheres (Sanders et al., 1989) in the diameter
range 0.5-6.5 /lm as tracers (Ooms-Wilms et al.,
in press). This was done to assess and compare
the grazing pressure of the two communities on
food particles which will generally include most of
the nanoplanktonic algae and flagellates, as well
as bacterioplankton.
Results
Composition and densities
The zooplankton species compositIOn did not
change much during the study years. Virtual dis-
appearance of Daphnia cucullata since 1981 and
the absence of large-bodied Daphnia sp. (D.
galeata) are quite conspicuous of the Lake Loos-
drecht. Among the crustaceans, Bosmina spp.
(B. longirostris and B. coregoni), Chydorus
sphaericus (Fig. 2) and cyclopoid copepods
(Diacyclops bicuspidatus, Mesocyclops leuckarti,
Acanthocyclops robustus, and Cyclops vicinus) and
their nauplli were important (for details see in
Gulati, 1990). The annual means of crustacean
densities varied more than two-fold, between
402 ± 395 ind 1- I in 1990 and 943 ± 883 ind 1- 1
in 1985 (Table 2). From about one-half to two-
thirds (45-69%) of the crustaceans on the aver-
age were cyclopoid copepods, including nauplii.
Virtually in all the study years the annual maxi-
mum for the crustaceans fell within a seven-week
period from 23 April to 10 June (Fig. 3).
Among about 20 genera and 40 species of ro-
tifers that were recorded in the Loosdrecht lakes
(Gulati, 1990), the four most abundant species
were Anuraeopsis fissa, Keratella cochlearis, Filinia
longiseta and Polyarthra (see in Fig. 2). The size
among the rotifers varied greatly: from about

Table 2. Mean annual densities (ind 1- I ± SD) of crustaceans and relative proportions of copepods and cladocerans in the total
and Bosl11ina spp. (B. longirostris and B. coregoni) in the Lake Loosdrecht.

Year Crustacean Copepods C1adocerans Bosl11ina spp.

% %
Total B. long. B. cor.
ind I - I % %
1981 447 ± 284 55 45 81 ± 33
1982 636 ± 612 45 55 257 ± 430
1983 559 ± 381 61 39 187 ± 213
1984 714 ± 527 62 38 215 ± 294
1985 943 ± 883 54 46 204 ±459
1986 675 ±474 62 38 190 19 81
1987 775 ± 647 55 45 192 67 33
1988 465 ± 281 65 35 113 1 99
1989 651 ± 452 69 31 157 27 73
1990 405 ±418 56 46 119 14 86
1991 402± 395 62 38 129 ± 269 11 89

Fig. 2. Photo-micrographs of some important species of cladocerans and rotifers in the Loosdrecht lake; scale for each species
is given by vertical bars (100 Ilm). The names of the species are: la, Bosl11ina longirostris; Ib, B. coregoni; 2, Daphnia cucul!ata;
3, Chydorus sphaericus; 4, Anuraeopsis fissa; 5, Keratella cochlearis; 6, Filinia 1000giseta; 7, Polyarthra sp.; and 8, Euchlanis dilatata
lucksiana.
73
74

Lake Loosdrecht
4000

Tot.Crustacea

3000

III
L
.....
...
.....
...... 2000
"0
C
H

1000

81 82 83 84 85 86 87 88 89 90 91

Lake Loosdrecht
18000
Tot.Rotifers

15000

12000

III
L
.....
.....
... 9000
......
"0
C
H

6000

81 82 83 84 85 86 87 88 89 90 91
Fig. 3. Densities of pelagial Crustacea and Rotifera in the Lake Loosdrecht.

60 11m in A.fissa to ca. 350 11m in Euchlanis di- Anuraeopsis contributed mainly to the seasonal
latata lucksiana. Their annual rotifer maximum trends of total rotifers throughout the summer
generally occurred in mid summer, about 3-6 period. Keratella cochlearis, the second most im-
weeks after the main crustacean peak (Fig. 3). portant rotifer in the lakes, generally peaked in
 75

Lake Loosdrecht
350
Daphnia spp.

2BO

210
cu
'-
.....
....
""'
"-
u
c:
H 140

70

'l i
Lake Loosdrecht

""'
';;; 2 -
'-
cu
D
E
:J
Z

''"-
0
u
0
.....
~ 1 -
-'

o -+-B~14-I"--B""2""""Ir-'BI-3""'I-4B~4""""I~B""5~-r-
\
I~BC:-6"'-r-
r~
I~BB.,.......IB~9:+--~19:-:0++-r--:19~1~1
I+B7=-'-'- ....

Fig. 4. Daphnia sp. and LeplOdora kindeii in the Lake Loosdrecht.

spring concomitantly with the maximum of cla-
doceran filter-feeders (Bosmina spp.). But for an
increase in the annual maxima of rotifers observed
in the years 1984-1986, there was no other note-
worthy trend in the rotifer dynamics. The ratio of
mean annual densities of rotifers to those of the
crustaceans, which was between 3.4 and 5 in the
years from 1981 to 1987, increased to between 6
and 7.7 in the following four years.
Bosmina longirostris was generally the first cla-
doceran to start the spring increase; and it coex-
isted for a short period in late spring and early
76
summer with B. coregoni which generally peaked
in mid summer. The mean annual densities of
Bosmina spp. during 1988-1991 (130 ind I-I)
were only about one-third lower than the mean
densities in the four-years preceding 1988 (Ta-
ble 2). The copepods exhibited a rather similar
trend in this regard.
Daphnia cucullata was characterized by mod-
erate densities and transient peaks up to 1987,
became very sparse in 1989 and was nearly ab-
sent in 1990 and 1991 (Fig. 4, upper panel). Its
maximal densities in the last three years ranged
between 5 and 15 ind 1- I and its occurrence was
generally restricted to a period during late spring
- early summer.
Leptodora kindtii was the only invertebrate
predator in the lakes. In the Lake Loosdrecht it
AXIS 2
SPRING 1 EARLY SUMMER
Keratella quadrata
•
Conoenilus unieornis
Synenaeta •
r-----r--NO~~~~~;::::2===~·=;;
-'1
AXIS 1
• Anuraeopsis
SI02
Secchi depth
WINTER -1
Fig. 5. PCA correlation biplot of limnetic zooplankton and environmental variables; species vectors are not drawn. Correlations
between species, between environmental factors, and between species and environmental factors can be derived by projecting the
vectors onto each other.
was encountered in the limnetic zooplankton
samples (> 150 Jim) from June to September in
densities that usually varied between 0.5 and 1.5
ind 1- I, but occasionally reached between 2 and
3 ind I i I as in 1985 (Fig. 4, lower panel). The
high n?mber~ in 1985 coinc~ded with those of
zoopl<fnkton In .gene~al, but. with those of Daphnia
sp. and the rotIfers In particular.
The PCA revealed that the within-year varia-
tion in the composition of zooplankton commu-
nity is larger than that between the years. A trend
over the years was lacking. The order in which the
species exhibited their maxima can be read from
the plot by viewing them clockwise, starting at the
lower end of axis I (left end of the abscissa in
Fig. 5). The changes in densities of the species
located in the centre of the plot are relatively
• Keratella eoenl .
• Cnydorus
• Pompnolyx
• • Daphnia
Bosmina
• Cye1 . eopepods
• Nauplii
Filinia
• Tr i enoeerea
fissa
• Temperature
Leeane DOC • Polyartnra
• Euehlanis
• Col1otneea
AUTUMN
 77

5000 Rotifers 160 - Lake Loosdrecht

1983:100 %
160 -
4000
140 -
.,
~

"- 3000 ..
c
120 -
.
u
'"
<-
III
c..
.c a.
E 100 -
z
::> 2000

~ l l
60 -
1000
60
nn n ...,
61 62 83 64 85 66 67 66 69 90 91

0 Year
1955 1985 1955 1985
Fig. 7. Changes in the annual means of seston « 150 j.lm)
and zooplankton masses in the Loosdrecht lake expressed as
Seston <150 p.,m. mg!l C per cent of the respective biomasses in 1983, which were set
7.5 at 100%; hatched bars, seston; unshade bars, zooplankton.

~ 5.0
"-
u
Ol Temporal changes in densities and mass
E 2.5

0.0
1955 1985
Fig. 6. A comparison of mean annual zooplankton densities
and seston mass in 1985 with those in 1955 in the Loosdrecht
lake. The data of 1955 are derived from Geelen (1955).
poorly described, and those towards the periph-
ery are better described. The general pattern of
the seasonal succession and deviation from it can
be derived from the PCA diagram and from data
for separate species. The succession starts in
spring with the rotifer species Keratella cochlearis.
The correlation of the PCA axis with environ-
mental variables suggests that temperature is the
main controlling factor (Fig. 5). This is demon-
strated, e.g., by the very high positive correlation
between A. fissa and temperature on one hand
and the high negative correlation between tem-
perature and K. cochlearis, a spring form, on the
other. The contrast between the late-spring-early-
summer period and autumn period lies in the con-
centrations of Total-P which are higher in the
former period than in the latter. The other envi-
ronmental variables depicted in the figure were
rather weakly correlated with the species abun-
dances.
The long-term comparison indicates 13-fold and
5-fold increases in the crustacean and rotifer den-
sities between 1955 (Geelen, 1955) and 1985, re-
spectively (Fig. 6). The most pronounced increase
was, however, noted for seston « 150 Jim) which
has increased by about forty times since 1955.
The trends in mass of seston and zooplankton
during 1981-1991 appeared to be roughly oppo-
site (Fig. 7). The seston mass has increased since
1983 and was between 20 and 60% higher in the
years 1984-1991. The zooplankton mass in-
creased during 1984 and 1985 only. Thereafter, it
slumped, in the years 1986-1991, to between 65
and 84% of the mean value in 1983.
Grazing
The grazing by the crustacean community had
generally a spring maximum caused mainly by
Bosmina spp. and Daphnia sp. (Fig. 8). In 1983,
however, both the spring and late autumnal peaks
were attributable to Bosmina alone. Most other
grazing values were less than 25 % d - 1. The
means for daily grazing rates in 1982, 1983 and
1984 were 25%, 10.6% and 11.7%, respectively,
and the mean net assimilation rates measured
78

125

Lake Loosdrecht
100

>-
ro
"C
"-
~

en 75
.~
c
N
ro
<-
en
c
ro
OJ
u 50
....,ro
'<-"
OJ
u

25

Fig. 8. Daily grazing (%) by zooplankton community (> 150 11m) in the Lake Loosdrecht (1981-1984), measured using I~C­
technique (see in Gulati, 1984).

simultaneously in these years were 12%, 4.7%
and 4.8 % d - t, respectively.
Based on the measured clearance rates of
Bosmina species in the Loosdrecht lakes during
1983 and 1984 (0.4-0.6 ml ind - t d - I: Gulati,
1984; Gulati et al., 1991a) and Bosmina densities,
the calculated consumption of Bosmina popula-
tion as % of the community grazing, was found
to be ca. 61 % in 1983 and 72 % in 1984. Thus,
bosminids, which accounted for about two-thirds
of the community grazing, were the most impor-
tant filter-feeding crustaceans in the Lake Loos-
drecht.
However, these community grazing rates do
not represent total zooplankton grazing since we
measured grazing for the zooplankton fraction
> 150 Jim, excluding a major part of rotifers. A
comparison of the clearance rates (CR) of the
most abundant rotifer grazers and important cla-
docerans shows marked differences in CR be-
tween the two taxonomic groups. The mean CR
values for rotifers except for Brachionus decreased
to nearly zero for tracer particles of 6.3 Jim; but
those for cladocerans increased by one to two
orders of magnitude as particle size increased
from 0.5 Jim to 2.4 Jim and 6.3 Jim (Ooms-Wilms,
unpub!. data). The specific clearance rates (SCR)
derived from the respective CR values and
weights were inversely related (P < 0.05) with the
filter-feeders' weight for all the three tracer types
(Fig. 9). In the range of weights of our animals,
viz. decreasing from Daphnia (2.7 Jig C ind - 1) to
Anuraeopsis (0.009 Jig C ind - I), whereas CR de-
creased by about 30 times, the SCR increased
between 80 and 100 times, depending on the size
of tracer particles (Fig. 9).
Based on the CR, using latex spheres and lit-
erature values, for the cladocerans and rotifers
and their respective densities on four dates in
May-June 1990, we observed the following daily
grazing rates for the two communities: crusta-
ceans, 29 %, 8% 18 % and 5% ; and rotifers, 66 %,
32%, 26% and 60%, respectively. This clearly
shows that in spring-early-summer, when both
communities have their maximal densities, the CR
of rotifers may far exceed those of crustaceans.
However, this observation is tentative and
restricted to food particles in the range 0.5-6 Jim.
 79

10000
f-
FI • 0.5 um
+- 2.4 um
+ <> 6.3 um
<>
<>

•At
-
I 1000
"0

I
<>
u

+
<>
100 +
cc
u
r.n •
Ke

• •
Ba Cs De
10
Cu • •
Be
•
0.001 0.01 0.1
Weight lugC)
Rotifers Cladocerans

Fig. 9. A comparison of the regression relationships between specific clearance rates (SCR) and weights of some important species
of crustaceans and rotifers based on three sizes (0.5 /lm; 2.4 /lm and 6.0 /lm) of microspheres. The SCR was derived from un-
published data of A. Ooms-Wilms. The abbreviations for filterfeeders species are given along the SCR for only the 0.5 /lm par-
ticles, but unless mentioned refer also to the SCR for the other two particle sizes (see along the upper two regression lines): Af,
Anuraeopsisjissa; Kc, Keratella cochlearis; Cu, Conochilus unicornis; FI, Filinia longiseta; Ba, Brachionus angularis; Cs, C. sphaericus;
Bc, B. coregoni; and Dc, D. cucullata.

Carbon and phosphorus in seston and zooplank-
ton
Carbon and phosphorus relationships of zoop-
lankton and its seston ( < 150 J-lm) food in Loos-
drecht lakes revealed pronounced seasonal dif-
ferences in the CjP ratios of seston, except in the
Lake Loosdrecht (Gulati et al., 1991c). In this
lake the ratios were generally concentrated
around 115 during 1983-1985 but increased in
the subsequent years and had a mean annual level
of ca. 150 in 1990. Also seasonal trends, which
were lacking earlier, have become evident
(Fig. 10, upper panel), because of increased P-
limitation in summers. In contrast to the CjP
ratios of seston, the CjP ratios of zooplankton
fluctuated more narrowly and were constant
around 45. For the six-year period (1983-1988),
in which both C and P were measured simulta-
neously, the zooplankton comprised ca. 4.5% C
but 12.7% P of the total contents of these ele-
ments, respectively, in the particulate suspended
matter. Thus, changes in the C and P contents of
zooplankton were strongly correlated (P<O.OOI)
such that the CjP ratio remained unaffected.
One main effect of relatively more P in zoop-
lankton than in their sestonic food is that zoop-
lankton contributes more to the total particulate-
P (seston-P + zooplankton-P) in the lake, than to
the total particulate-C (Fig. 10, lower panel).
During the peak densities and biomass of zoop-
lankton as in spring periods, the share of
zooplankton-P in the total particulate-P increased
markedly and reached up to 20% or more for
short periods.
Phosphorus excretion by zooplankton
The computed rates of P excretion by zooplank-
ton (Fig. 11) show that but for two high values in
1982, all others are less than 5 J-lg P0 4 -P 1- I d - I.
80

250
Lake Loosdrecht

Zooplankton
200 Seston < 150 urn

(J)
:::J
~ 150
J:::
Cl.
(J)
o
J:::
Cl.
"-
C
o 100
.D
L

'"
U

50

Lake Loosdrecht
0.30

I
\
as Carbon
"
0.25 "
as Phosphorus I",
I,
II

"
'1
" ,
6 o. 20
....,
I
I
(J) I
I
OJ
(J) I
I
I
I I
~ O. 15 I
, I I

....,0 I
I,
I I

, I I
"- I

,
I "
'II
I
I
I
I
Cl.
0 ,,
I /ll'
Ill' I
I
r
I
I

~ o. 10 I
11,1

:) "
,.

:~ I~ :
I I

:
I
, I ,II, I I
I \ I" 1\ I I

I : I I \: ~ : :
~/\: \/\/ v \ I I

0.05 I ~ / \,': II
,,' \
,
\'
\
\~ \.

o. 00 -+---,----,----,-----,-----,---,---,-----,----,------,
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990

Fig. 10. Carbon-phosphorus ratios of zooplankton (> 150 11m) and seston ( < 150 11m) in the Lake Loosdrecht (upper panel); and
relative proportions sestonic carbon and phosphorus in zooplankton (lower panel).

Generally, the primary annual peak which oc-
curred in spring was caused mainly by the in-
crease in crustacean biomass, especially Bosmina
spp. The second peak, if recorded, was in mid
summer when zooplankton was generally present
in modest densities, but the temperature was at its
annual maximum (18-20 cC). Mean excretion
rates in May-September varied narrowly between
1.5 and 1.8 Jig P 1- I d - I,
The measured excretion rates in Euchlanis di-
 81

20
Lake Loosdrecht

Calculated P-excretion of zooplankton

(from 1985 data)

15

1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991

Fig. 11. Daily excrelion of phosphorus by the zooplankton (> 150 pm) community in the Lake Loosdrecht. The data are based
on routine measurements on the in situ community during 1985-1986 in the ambient lake water, and the regression model
(Y = 0.436eo.145X) for relationship between water temperature (X) and specific P excretion (Y): pg P mg- 1C zoop d - 1(Den Oude
& Gulati, 1988).

lalala lucksiana (isolated from and cultured in the
lake water) of 0.06-0.18,ug P mg- 1 OW h- 1
(Gulati el al., 1989), were similar to the mean
weight-specific rates for crustacean community.
However, these rates were not representative for
the main rotifer species in the Lake Loosdrecht,
because these species (A. fissa and K. cochlearis)
were much smaller, weighing about an order-and-
one-half less than Euchlanis. Moreover, this large
euchlanid is relatively sparse and restricted to late
summer and early autumn periods.
Table 3. Phosphorus recycling by crustacean zooplankton in
the Lake Loosdrecht, based on means of May-September
during 1987-1990. The ingestion (I), assimilation (A), and
egestion of sestonic P by zooplankton were measured as car-
bon but converted to P using C/P ratios of seston (Gulati
et al., 199Ic); the zooplankton P-excretion (E) is based on
measurements by Den Oude & Gulati (1988). The zooplank-
ton P losses other than via excretion (E) i.e. due to predation
and natural mortality, are derived as the difference between A
and E. Uptake of P during phytoplankton production (item 8)
was assumed to be in the same ratio as C/P ratios measured
in seston, and item 6 divided by item 8 gives item 9, i.e., the
phytoplankton daily P requirements covered by daily zoop-
lankton excretion. All the biomass data are expressed as
pg P 1- 1 and rates as ~g PI- 1d - I

Role of zooplankton in P-recycling Parameter 1987 1988 1989 1990

The calculated P-recycling in the Lake Loos- 1. Seston mass 86 84 74 59

2. Zooplankton mass 8.7 8.8 6.5 8.3
drecht, for the years 1987-1990, is based on
3. Ingestion (I) 5.5 5.9 3.4 3.5
phosphorus in seston and zooplankton and its 4. Assimilation (A) 2.2 2.6 1.5 1.5
recycling via zooplankton grazing and excretion 5. Egestion (I-A) 3.3 2.3 1.9 2.0
and phytoplankton production rates (Table 3). 6. Excretion (E) 1.4 1.5 1.8 1.6
Zooplankton, via its grazing, removed daily be- 7. Zooplankton loss (A-E) 0.8 2.1 -0.3 - 0.1
8. Primary produktion (PP) 10.2 9.1 7.1
tween 4 and 6 % of the P stock in seston (item 3, 9. E/PP x 100 (% d - I)
5.8
13.7 16.1 25.3 27.6
Table 3) but returned more than half of this gross
82
P-uptake as egestion. The fate ofP in the egestion
products was difficult to quantify since these
products will partly lose some P0 4 -P due to leach-
ing but most of it in particulate form will re-enter
the seston as detrital-Po Ifwe consider leaching of
P from the egestion products as insignificant, the
daily P-input to seston (item 1) pool via zoop-
lankton egestion (item 5), will be roughly 25-33 %
(item 5/item 8) of the P input via the phytoplank-
ton production. Simultaneously, zooplankton ex-
cretion equalled between 14 and 28 % of the phy-
toplankton production demand ofP (see item 9 in
Table 3).
The importance of zooplankton P-excretion, as
a potential P-source for phytoplankton produc-
tion, increased markedly in 1989 and 1990 com-
pared with the preceding two years. This is due
to a decrease in primary production in 1989-
1990, rather than to increased zooplankton ex-
cretion rates.
The zooplankton recycling of its body-P via
excretion (item 6/item 2, in Table 3) was about
0.20 d - 1, the annual means ranging between 0.19
and 0.28 d - I.
Discussion
Here we attempt an integrated discussion of the
results of 10-year monitoring of zooplankton den-
sities and biomass, and studies on zooplankton
feeding and excretion rates, including the more
recent work on rotifer feeding.
There is practically no change since 1986 in
composition of abundant species of either rotifers
or crustaceans (Gulati, 1990). Moreover, because
of the normal year-to-year oscillations, the am-
plitude of changes in annual mean densities do
not warrant clear-cut conclusions on the likely
trends, either natural or in response to restoration
measures. This has been also confirmed by anal-
ysis of zooplankton data in the light of physico-
chemical variables, using Canonical Ordination
Techniques (Van Tongeren et al., 1992). The
question arises: 'have the basic conditions that
should bring about a change in zooplankton com-
position or abundance been brought about or met
by the rehabilitation measures'? A cursory an-
swer to this is: no! In this regard the important
aspects to be examined critically are the zoop-
lankton food and predation. We have noted that,
contrary to the expected decreases, the seston
mass (this study) and the filamentous cyanobac-
teria have, in fact, increased (see e.g. in Van
Tongeren et al., 1992). Moreover, this latter has
occurred despite the increasing evidence of P-
limitation derived both from laboratory data (Ri-
jkeboer et al., 1991 ) and increase in CjP ratios of
seston (Gulati etal., 1991c). The reduction in P-
load to ca. 0.3 g P m - 2 Y- 1 is apparently less
than expected (Van Liere & Janse, 1992) or
needed. Therefore, filamentous cyanobacteria still
continue to dominate this lake system. There are
no recent data on predation by planktivorous fish,
nor do we know if the standing crop of bream
(Abramis brama) has changed. It may be men-
tioned that a major fish kill involving bream', be-
sides management and P-reduction measures in
the Schlachtensee, resulted within three years in
important shifts in zooplankton structure (Ewald,
1991). The changes relate to the appearance of
large-bodied Daphnia (D. galeata) and reduction
in the numbers of Bosmina sp.
The Loosdrecht lakes differ markedly from
other eutrophic lakes in The Netherlands, e.g.
Lake Tjeukemeer which has a net P-retention of
about 10% of the external-P load of 25-50 mg P
m -2 d - I (Van Huet, 1991). This is more than an
order-of-magnitude higher than the net daily re-
tention calculated for the Lake Loosdrecht by
lanse et al. (1992). Moreover, the total-P concen-
tration in the Tjeukemeer is also much higher
(Lammens et al., 1992). However, despite these
differences in phosphorus concentrations, the
mean chlorophyll contents during spring and
summer in the two lakes (130-150 Jlg 1- 1) are
similar. Nevertheless, in the Lake Loosdrecht
throughout the year more than 90 % of the chlo-
rophyll comes from filamentous cyanobacteria
and the prochlorophyte Prochlorothrix hollandica.
On the other hand, in Tjeukemeer green algae and
diatoms contribute nearly as much to the chloro-
phyll as the filamentous prokaryotes, at least in
spring period. Therefore, the food situation in the

Lake Loosdrecht during early spring is less opti-
mal for cladocerans than in Tjeukemeer. Even
though the total fish stock in the two lakes is
comparable, namely ca. 300 kg ha - I, and both
have bream as the most dominant planktivorous
fish, essential differences can be observed in the
size composition of fish (Lammens et al., 1992).
Whereas in the Tjeukemeer standing stock of
bream < 25 em is only about 50 kg ha - I, in the
Lake Loosdrecht it is three times as high -150 kg
ha - I. On the other hand, whereas the Tjeuke-
meer contains about 150 kg ha - I bream> 25 em,
the Loosdrecht has about 80 kg ha - I of this.
These are, however, very rough estimates based
on annual means of several years since differ-
ences in individual years may be pronounced but
will be offset by averaging over the years. Un-
doubtedly, the effects of the differences between
the lakes in their nutrient status, producers' com-
position and relative abundance and of fish size
composition and standing crop are likely to be
reflected in the producer-consumer interphase,
the zooplankton.
We do not know to what extent the absence in
the Lake Loosdrecht of Daphnia galeata (which is
quite common in many other eutrophic lakes in
the Netherlands) can be attributed to food qual-
ity or fish predation. From the lakes undergoing
biomanipulation (see e.g. Lyche et al., 1990; and
other references in Gulati et aI., 1990) it is clear
that with a decrease in the standing crop of plank-
tivorous fish, the share of larger grazers in zoop-
lankton generally rises and remains high. An in-
triguing factor also in the Lake Loosdrecht is the
virtual extinction since 1989 of D. cucullata, which
is the smallest among the Daphnia species in the
Dutch lakes, and which generally co-occurs with
D. galeata, irrespective of the trophic status. At
present one can only conjecture that predation by
bream has become even more intense and has
markedly reduced the Daphnia population in
Loosdrecht lakes.
In the absence of large-bodied Daphnia species
and because of the disappearance of D. cucullata,
one would expect the planktivores, particularly
bream < 25 em, to exert a high pressure on cy-
clopoid copepods and bosminids. But because of
83
relatively smaller size, more so for copepods with
their much smaller juvenile stages, can escape fish
predation. The bosminids are apparently well
adapted to the food dominated by cyanobacteria
(Gulati et al., 1991a), which are abundant even in
early spring. The bosminids can generally quickly
build-up their populations, in the absence of
daphnid competitors, which may cause quick and
severe depression of food resources (see e.g. in
Romanovsky, 1985).
The grazer-induced net mortality (assimilation)
ofcyanobacteria is essentially low (4.7-12 % d - 1)
as observed from the grazing studies, and due
mainly to B. coregoni. In view of the assimilation
efficiencies of around 45 %, the amount egested
will be about 55 % of that ingested. The role of
bosminids in transforming cyanobacteria into de-
tritus may not be unimportant since in the Lake
Loosdrecht constantly a high proportion of seston
« 150 ,urn) exists as detritus (GOllS et al., 1990).
Also, model studies on phosphorus flow in the
lake's ecosystem (J anse et al., 1992: Fig. 2) reveal
that, based on carbon, the proportion of detritus
consumed will be even higher because of the ex-
pected higher CjP ratios of detritus than cyano-
bacteria and other planktonic food-items. Be-
sides, lack of competition among cladocerans to
feed on cyanobacteria may be related to differen-
tial use of food particles by Bosmina and Daphnia
species (Irvine, 1986). Moreover, the bosminids
exert a much stronger effect due to a more marked
effect of body length on clearance rates, than
Daphnia, feeding on particles both in the size range
0-7 ,urn and 7-15 ,urn. In laboratory studies using
the Lake Loosdrecht water and large-bodied
Daphnia (D. magna) Dawidowicz et al. (1988)
demonstrated that the fecundity of the reared
daphnids markedly decreased as the cyanobac-
terial concentration was increased to 80000 fila-
ments 1- I. Though the measured ingestion rates
were higher than the animals reared on 'diluted'
lake water, the advantage of higher consumption
rates at the high food levels appears to be more
than off set by the expected very high respiratory
cost of food gathering/rejection. Thus, besides
the intense predation by bream, cyanobacterial
dominance in the lake is not conducive to the
84
establishment of large-bodied forms. Leptodora
kindtii, despite its large size, can escape predation
by fish, both because of its transparent body and
tendency to migrate to the bottom layers in the
day-time (J. Vijverberg, pers. comm.). It has per-
sisted, but we do know its role as a predator on
microzooplankton on which it is most likely to
feed.
Though the planktivore predators, and perhaps
in some situations food, adversely affect the op-
timal development of crustacean zooplankton,
especially the large-bodied filter-feeding forms,
the rotifers, in contrast, are abundant and have
high diversity indicating favourable conditions.
Two important features of rotifers need to be em-
phasized: 1) they appear to be an important fac-
tor in the mortality of nanoplankton ( < 10 Jim),
including the unidentified flagellates, diatoms and
bacterioplankton; and 2) they have much higher
specific clearance rates on the small food types
than cladocerans, and because of their much
higher densities appear to be an important link in
the microbial food chain. Besides, the small size
confers upon the rotifers an added advantage in
escaping predation better than the crustaceans.
These characteristics make rotifers superior to
crustaceans under the existing conditions, as well
as have probably allowed them to virtually out-
compete the latter during the process of eutroph-
ication of the Loosdrecht lakes. Their success
appears to have gone hand in hand with the in-
creased success of bream in this lake ecosystem.
The same factors which adversely the affect the
crustaceans in Loosdrecht lakes perhaps also
prevent the large-bodied rotifers, namely the bra-
chionids (Euchlanis sp. and Brachionus sp.), from
establishing. Among these, Euchlanis dilatata
lucksiana represents both a unique and extreme
organism: being among the largest of the rotifer
taxa, it is able to consume prokaryotic filamen-
tous plankton, cyanobacteria and Prochlorothrix
hollandica, of varying lengths (Gulati et al., in
press). However, the failure of this rotifer to es-
tablish and build up a sizeable population, in mid
summer, can be explained by its vulnerability to
predation (Ejsmont-Karabin et al., 1988). Also,
Brachionus species which are known to attain
densities as high as 17000 ind 1- I in eutrophic
lakes, as in Lake S0bygard (J eppesen et al., 1990),
were never abundant in the Lake Loosdrecht.
Their numbers are generally high if green algae
(e.g. Scenedesmus spp.) on which they feed (Gil-
bert & Starkweather, 1978) are available in high
concentrations.
The role of zooplankton in P-recycling in the
lake will be grossly underestimated if the excre-
tion rates of rotifers are not included. This point
has been earlier emphasized by Gulati (1990).
The computed excretion rate of rotifers for the
summer of 1985, using the rotifer densities and
the relationships of excretion rate with size and
temperature given by Ejsmont-Karabin (1983)
was 33 ± 19.7 J.!g P 1- I d - I which is thirteen times
higher than that measured for the crustaceans by
Den Oude & Gulati (1988), viz 2.4 Jig P 1- 1 d - 1.
Though these rates are very high, as well as their
extrapolation to lakes is questionable, it does,
nevertheless, indicate that zooplankton excretion,
especially that of rotifers, can be an extremely
important factor in the lake's internal P-loading.
Lastly, the zooplankton research as well as
other studies on Loosdrecht lakes have gone a
long way in increasing our insight into the func-
tioning as well as the resilience of a hypertrophic
lake ecosystem. The expected changes, or, taking
a retrospective view, the lack thereof, have been
the driving force of the long-term, integrated stud-
ies on the Lake Loosdrecht.
Acknowledgements
The first author is highly indebted to Dr. Sikko
Parma, the Director of the Limnological Institute,
who encouraged zooplankton research in the
Loosdrecht lakes from its start in 1981; he pro-
vided the modest funds at his disposal to support
the stay at the Institute of several foreign scien-
tists to cooperate in the zoological research. Our
colleague Dr. Koos Vijverberg made several use-
ful suggestions to improve the manuscript. The
European Community funded the investigations
on zooplankton excretion and rotifer grazing. We
are grateful to Dr. Louis van Liere, the WQL co-

ordinator, who contributed to the success of
zooplankton studies silently but surely. We also
thank Mrs. Cecilia Janssen-Kroon for taking care
of the lay-out of this typescript and of publica-
tions and reports on zooplankton that have ap-
peared since 1984.
References
Dawidowicz, P., Z. M. Gliwicz & R. D. Gulati, 1988. Can
Daphnia prevent a blue-green algal bloom in hypertrophic
lakes? A laboratory test. Limnologica 19: 21-26.
De Kloet, W. A., P. J. Boesewinkel-de Bruyn & L. Breebaart,
1991. Phytoplankton and its production rates in the Loos-
drecht lakes before and after P-load reduction. Verh. int.
Ver. Limnol. 24: 715-718.
Den Oude, P. J. & R. D. Gulati, 1988. Phosphorus and ni-
trogen excretion rates of zooplankton from the eutrophic
Loosdrecht lakes, with notes on other P sources for phy-
toplankton requirements. Hydrobiologia 169: 379-390.
Ejsmont-Karabin, J., 1983. Ammonia nitrogen and phospho-
rus excretion by planktonic rotifers. Hydrobiologia 104:
231-236.
Ejsmont-Karabin, J., R. D. Gulati & J. Rooth, 1989. Is food
availability the main factor controlling the abundance of
Euchlanis dilatata lucksiana Hauer in a shallow, hypertro-
phic lake? Hydrobiologia 186/187: 29-34.
Ewald, S., 1991. Long-term changes of crustacean plankton
during successful restoration of Lake Schlachtensee (Berlin-
West). Verh. int. Ver. Limnol. 24: 866-872.
Faafeng, B. A. & J ens Petter Nilssen, 1981. A twenty-year
study of eutrophication in a deep, soft-water lake. Verh. int.
Ver. Limnol. 21: 412-424.
Gee/en, J. F. M., 1955. The plankton of the lakes. Unpub-
lished report no. 38. Amsterdam Municipal Waterworks (in
Dutch).
Gilbert, J. J. & P. L. Starkweather, 1978. Feeding in the ro-
tifer Brachiollus calyciflorus III. Direct observations on the
effect of food type, food density, change in food type and
starvation on the incidence of pseudotrochal screening.
Verh. int. Ver. Limnol. 20: 2382-2388.
Gons, H. J. & M. Rijkeboer, 1990. Algal growth and loss
rates in Lake Loosdrecht: first evaluation of the roles of
light and wind on a basis of steady state kinetics. Hydro-
biologia 191: 129-138.
Gulati, R. D., 1983. Zooplankton and its grazing as indica-
tors of trophic status in Dutch lakes. Envir. Monit. Ass. 3:
343-354.
Gulati, R. D., 1984. Zooplankton and its grazing as measures
of trophy in the Loosdrecht lakes. Verh. int. Ver. Limnol.
22: 863-867.
Gulati, R. D., 1990. Zooplankton structure in the Loosdrecht
lakes in relation to trophic status and recent restoration
measures. Hydrobiologia 191: 173-188.
85
Gulati, R. D., K. Siewertsen & G. Postema, 1982. The zoop-
lankton: its community structure, food and feeding and role
in the ecosystem of Lake Vechten. Hydrobiologia 95: 127-
163.
Gulati, R. D., K. Siewertsen & G. Postema, 1985. Zooplank-
ton structure and grazing activities in relation to food qual-
ity and concentration in Dutch lakes. Arch. Hydrobiol.
Beih. 21: 91-102.
Gulati, R. D., J. Rooth & J. Ejsmont-Karabin, 1987. A lab-
oratory study of feeding and assimilation in Euchlanis di-
latata lucksiana. Hydrobiologia 147: 289-296.
Gulati, R. D., J. Ejsmont-Karabin, J. Roath & K. Siewertsen,
1989. A laboratory study of phosphorus and nitrogen ex-
cretion of Euchlanis dilatata lucksiana. Hydrobiologia 186/
187: 347-354.
Gulati, R. D., E. H. R. R. Lammens, M.-L. Meijer & E. Van
Donk (eds), 1990. Biomanipulation, tool for water man-
agement (First International Conference 8-11 August, 1989,
Amsterdam). Developments in Hydrobiology 61: Kluwer
Academic Publishers, Dordrecht, 628 pp.
Gulati, R. D., C. Vuik, K. Siewertsen & G. Postema, 1991a.
Clearance rates of Bosmina species in response to changes
in trophy and food concentration. Verh. int. Ver. Limnol.
24: 745-750.
Gulati, R. D., L. Van Liere & K. Siewertsen, 1991b. The
Loosdrecht lake system: Man's role in its creation, pertur-
bation and rehabilitation. In: Terrestrial and Aquatic Eco-
systems: Perturbation and Recovery; ed. O. Ravera. Chich-
ester, Ellis Horwood. pp. 593-606.
Gulati, R. D., K. Siewertsen & L. Van Liere, 1991c. Carbon
and phosphorus relationships of zooplankton and its seston
food in Loosdrecht Lakes. Mem. 1st. ital. Idrobiol. 48:
279-298.
Gulati, R. D., J. Ejsmont-Karabin & G. Postema, in press.
Feeding in Euchlanis dilatata lucksiana Hauer on filamen-
tous cyanobacteria and a prochlorophyte. Hydrobiologia.
Irvine, K., 1986. Differential feeding behaviour of the domi-
nant Cladocera as an explanation of zooplankton commu-
nity structure in the Loosdrecht Lakes. Hydrobiol. Bull. 20:
121-134.
Janse, J. H., T. Aldenberg & P. R. G. Kramer, 1992. A math-
ematical model of the phosphorus cycle in Lake Loosdrecht
and simulation of additional measures. Hydrobiologia 233:
119-136.
Jeppesen, E., M. S0ndergaard, O. Sortkjaer, E. Mortensen &
P. Kristensen, 1990. Interactions between phytoplankton,
zooplankton and fish in a shallow, hypertrophic lake: a
study of phytoplankton collapses in Lake S0bygard, Den-
mark. Hydrobiologia 191: 149-164.
Lammens, E. H. R. R., P. J. Boesewinkel-de Bruyn, H.
Hoogveld & E. Van Donk, 1992. Phosphorus-load, phyto-
plankton, zooplankton and fish stock in Loosdrecht Lake
and Tjeukemeer: confounding effects of predation and food
availability. Hydrobiologia 233: 87-94.
Loogman, J. G. & L. van Liere (eds), 1986. Proceedings of the
Waterquality Research Loosdrecht Lakes Symposium on:
86
Restoration of shallow lake ecosystems with special em-
phasis on Loosdrecht Lakes. Hydrobio!. Bull. 20: 259 pp.
Lyche, A., B. A. Faafeng & A. Braband, 1990. Predictability
and possible mechanisms of plankton response to reduc-
tion of planktivorous fish. Hydrobiologia 200/20 I: 251-
261.
Ooms-Wilms, A. L., 1991. Ingestion of fluorescently-Iabelled
bacteria by rotifers and cladocerans in Lake Loosdrecht as
measure of bacterivory: preliminary results. Mem. 1st. ital.
Idrobio!. 48: 269-278.
Ooms-Wilms, A. L., G. Postema & R. D. Gulati, in press.
Clearance rates of bacteria by the rotifer Filinia longiseta
(Ehrb.) measured using three tracers. Hydrobiologia.
Rijkeboer, M., F. de Bles & H. J. Gons, 1991. Role of sestonic
detritus as a P-buffer. Mem. 1st. ita!. Idrobiol. 48: 251-260.
Romanovsky, Y. E., 1985. Food limitation and life history
strategies in cladoceran crustaceans. Arch. Hydrobiol.
Beih. 21: 363-372.
Sanders, R. W., K. G. Porter, S. J. Bennett & A. E. DeBiase,
1989. Seasonal pattern ofbacterivory by flagellates, ciliates,
rotifers and cladocerans in a freshwater planktonic com-
munity. Limno!. Oceanogr. 34: 673-687.
Van Huet, H., 1991. Modelling water transport and phospho-
rus eutrophication in an interconnected lake system. A sce-
nario study. Ph. D. Thesis, Agricultural University, Wa-
geningen, 129 pp.
Van Liere, L. & R. D. Gulati (eds), 1992. Restoration and
recovery of shallow eutrophic lake ecosystems in The Neth-
erlands: epilogue. Hydrobiologia 233: 283-287.
Van Liere, L. & J. H. Janse, 1992. Restoration and resilience
to recovery of the Lake Loosdrecht ecosystem in relation
to its phosphorus flow. Hydrobiologia 233: 95-104.
Van Liere, L., S. Parma & R. D. Gulati, 1992. Working group
Water Quality Research Loosdrecht Lakes: its history,
structure, research programme, and some results. Hydro-
biologia 233: 1-9.
Van Tongeren, O. F. R., L. Van Liere, R. D. Gulati, G.
Postema & P. J .. Boesewinkel-de Bruyn, 1992. Multivariate
analysis of the plankton communities in the Loosdrecht
lakes: relationship with the chemical and physical environ-
ment. Hydrobiologia 233: 105-117.
Wilms, A. L., R. D. Gulati & G. Postema, 1991. First attempt
to measure the clearance rate of Anuraeopsisjissa. Verh. int.
Ver. Limnol. 24: 742-744.
Hydrobiologia 233: 87-94, 1992.
L. Van Liere & R.D. Gulati (eds). Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 87
© 1992 Kluwer Academic Publishers.

P-Ioad, phytoplankton, zooplankton and fish stock in Loosdrecht Lake

and Tjeukemeer: confounding effects of predation and food availability

Eddy H. R. R. Lammens,l * Nel Boesewinkel-De Bruyn,1 Hans Hoogveld I & Ellen Van Donk 2
I Limnological Institute, Vijverhof, Rijksstraatweg 6,3631 AC, Nieuwersluis, The Netherlands; 2 Agricultural
University Wageningen, Department of Nature Conservation, Laboratory of Aquatic Ecology. P.O. Box
8080, 6700 DD Wageningen; * Present address: Institute of Inland Water Management and Waste Water
Treatment, P.O. Box 17, 8200 AA Lelystad, The Netherlands

Key words: bream, Daphnia, cladocerans, copepods, cyanobacteria, green algae, trophic interactions,
feeding conditions

Abstract

The fish community in the Loosdrecht lakes is dominated by bream, pikeperch and smelt and is char-
acteristic of shallow eutrophic lakes in The Netherlands. The biomasses of the respective fish species
amount to ca. 250, 25 and 10 kg ha - I and correspond to those in Tjeukemeer, another lake in The
Netherlands. The average size of bream, however, is much smaller in the Loosdrecht lakes as a con-
sequence of poorer feeding conditions. The zooplankton community in the Loosdrecht lakes is predom-
inantly composed of relatively small species such as Daphnia cucullata, Bosmina coregoni and cyclopoid
copepods, whereas in Tjeukemeer, Daphnia hyalina is permanently present in relatively high densities and
the other species show a larger mean length. In the Loosdrecht lakes, the absence of D. hyalina and the
smaller sizes of the other zooplankton species could be the consequence of a higher predation pressure,
in combination with unfavourable feeding conditions for the zooplankton including the low density of
green algae and the high density of filamentous cyanobacteria. A biomanipulation experiment in Lake
Breukeleveen, one of the Loosdrecht lakes, indicated that feeding conditions were too unfavourable for
large zooplankton to develop in spring, when the reduced fish biomass was not yet supplemented by
natural recruitment and immigration.

Introduction last ten years emphasize the role of fish in the

Most lakes in The Netherlands are very turbid
because of algal blooms, particularly filamentous
cyanobacteria, and are devoid of macrophytes. In
these lakes, bream Abramis brama, roach Rutilus
rntilus, smelt Osmerns eperlanus, ruffe
Gymnocephalus cernua and pikeperch Stizostedion
lucioperca are the dominant fish species, feeding
mainly on zooplankton and chironomids (Lam-
mens et al., 1991). Biomanipulation studies of the
expression of eutrophication (Gulati et al., 1990)
particularly in stabilizing this condition (Moss,
1990; Scheffer, 1990). The cascading trophic in-
teractions are now well known (Carpenter, 1985;
Persson et al., 1988), although there are many
exceptions to the rule (McQueen et al., 1986,
1989; Lammens et aI., 1990) and the removal or
reduction of fish usually leads to a dramatic in-
crease in transparency and a consequent growth
of macrophytes (Hosper & J agtman, 1990). One
88

important exception to the rule was the applica-

tion ofbiomanipulation in lake Breukeleveen, one
of the Loosdrecht lakes, where the reduction in
fish biomass had hardly any effect (Van Donk
et al., 1990). The high density of filamentous cy-
anobacteria and the large impact of turbulence
causing wind seemed to be the most important
counteracting effects. Enclosure experiments in
Tjeukemeer did not result in clear water either,
even though no fish were present (Richter et al.,
1987).
Because of the difficulties experienced in im-
proving the transparency in Lake Breukeleveen
RYINE
and in Tjeukemeer enclosures by biomanipula-
tion, we were interested in comparing the food WEST-GERMANY

chains in these lakes. We used data concerning

external P load, total-P concentration, concentra- ,
,- ,
BELGIUM
tions of the main phytoplankton groups (Cyano-
bacteria, Chlorophyceae and Diatomeae), the
density and average size of main zooplankton I

species or groups and fish species density and ---1

composition for the period 1984-1990. We were
mainly interested in the difference between the
lakes rather than the seasonal or yearly trends
(see Van Tongeren et al., this volume) and there-
fore we compared only the averages of the May-
October period.
Study area
Tjeukemeer is a shallow (average depth 1.5 m),
eutrophic and turbid (Secchi-disc 25-35 cm)
freshwater lake with a surface area of 21.5 km 2 .
Vegetation is rare, and is represented by only a
limited amount of reed. As the lake is shallow and
exposed to wind, summer stratification does not
occur and in consequence the phytoplankton and
zooplankton are almost homogeneously distrib-
uted (De Nie et al., 1978). The lake is part of an
interconnected system of water bodies, which re-
ceives water from the nearby lake IJsselmeer dur-
ing summer and from the surrounding polders in
winter, when precipitation exceeds evaporation
(Fig. 1). The management of the hydrological re-
gime is designed to control the water level which
is required for agricultural purposes. The Loos-
Fig. 1. Location of the Loosdrecht lakes (LoJ and "ljeuke-
meer (Tj) in the Netherlands.
drecht lake (980 ha) and Lake Breukeleveen
(180 ha and a part of the Loosdrecht system) are
similar to Tjeukemeer regarding depth and tur-
bidity, although the external P-Ioad is different.
Phosphorus reduction measures have reduced the
yearly external P-Ioad in Loosdrecht lakes to ca.
0.37 g m - 2 (Van Liere et al., 1990a; Van Liere &
J anse, this volume), whereas the yearly external
P-Ioad in Tjeukemeer is 25-50 g m - 2, mainly be-
cause the lake receives drainage from the agricul-
tural area (Van Huet, 1991). However, P-
retention in the latter lake is only 10%, while in
the Loosdrecht lakes it is about 60 %.
Methods and materials
Fish were collected in the period May-October in
the years 1984-1990. In Tjeukemeer the fish were
caught once a month, while in Loosdrecht Lake
they were taken three times per year. In both
lakes a small-mesh trawl (5.5 mm cod end) was
used at five stations and standard hauls were
 89

made at a speed of 1 m s - 1 for 10 min. Length- 320 I~~~~~~---;:::====~I

frequency distributions (fork length) and total
numbers of each species were determined. 240

Catches of young-of-the-year were only used for ~

i!:
the months September-October, because before "' 160
that period the fish were still too small to be re- ~
iii
tained by the trawl or confined to the inshore 80
area.
Zooplankton was sampled at the same time
o
and place of trawling. Using a 5-L Friedinger LOOSDRECHT T..ELKErvEER

sampler with a cylinder of 60 cm height, two sam- _ SA _ RO~ AU E'm PE ~ PI [==:J SM

ples were taken at each of the five stations, one

just below the water surface and the other just Fig. 2. Biomass per ha of bream (BR), roach (RO), ruffe (RU),
perch (PE), pikeperch (PI) and smelt (SM) in the Loosdrecht
above the bottom. The samples were concentrated lakes and Tjeukemeer as an average for the periods May-
by filtering through a 120 J1m mesh and preserved September in 1984-1990.
in 4 % formalin. The combined sample was sub-
sampled with a whirling vessel after Kott (1953),
Pikeperch
and organisms were identified and their lengths
measured. The length of the cladocerans was 10,--------------------,
measured from the top of the head or, if a helmet
was present, from the base of the helmet to the
base of the tail spine.
In Tjeukemeer phytoplankton was sampled
fortnightly, while in the Loosdrecht lakes it was 5

sampled every four weeks. Further details of the

sampling procedure and the method used for
counting are given in Moed & Hoogveld (1982),
Gulati (1990) and Van Tongeren et al. (1992).
For sampling and determination of chlorophyll-a 10 20 30 40 50 60 70 80
and total P concentration see Moed (1973), De Length (em)

Haan & Moed (1984) and Van Liere et al. _ Tje<J<emeer [==:J Loosdreeht

(1990b). Fig. 3. Size distribution of pikeperch in Loosdrecht Lake and

Results
The biomass and species composition of the fish
communities in Loosdrecht Lakes and Tjeuke-
meer are similar. The total biomass in both lakes
is ca. 300 kg ha - I and is dominated by bream
(68% in Tjeukemeer, 77% in Loosdrecht Lake)
(Fig. 2). Pikeperch is the most abundant predator
(9% in Tjeukemeer, 8 % in the Loosdrecht Lake)
and in both lakes its population is dominated by
large specimens (45-85 cm forklength) (Fig. 3).
The other fish species present are smelt, perch
(Perea fluviatilis), roach, ruffe and white bream
Tjeukemeer expressed in biomass per 10 cm length class av-
eraged for the periods May-September in 1984-1990.
(Blicca bjorkna) varying from 7.6% to 0.5% of the
total biomass (Fig. 2). These species are the main
prey fish of pikeperch and selected in the order as
indicated (Lammens, unpublished results). Smelt,
white bream and roach are more abundant in
Tjeukemeer and ruffe and perch occur in similar
densities in both lakes.
The fish communities in Loosdrecht Lake and
Tjeukemeer differ mainly in the size composition
of the bream population. In Loosdrecht Lake the
largest part of the fish biomass is comprised of
bream <25 cm (ca. 65%), whereas in Tjeuke-
90

70 8000
meer bream> 25 cm make up the largest part (ca. r--------~----------,

75%, Fig. 4). Van Densen et al. (1986) and Lam- 60

mens (1982) determined the growth rate of bream 50
6000
in Loosdrecht Lake and in Tjeukemeer respec-
40
tively. Bream up to 25 cm grew very slowly in 4000
Loosdrecht Lake, but very fast in Tjeukemeer. 30

During the study period the recruitment in 0 + 20

2000
bream did not significantly differ, but we did find
10
a significant (p < 0.005), t-test) difference in size
of the 0 + bream at the end of its first growing o
TJ LO TJ LO
season (5.8 cm in Tjeukemeer, 5.2 cm in Loos- _ Length (mm) Ei£] N/ha
drecht; Fig. 5). Therefore the difference in length-
frequency distribution is caused not by different Fig. 5. Size and recruitment of 0 + bream in Loosdrecht Lake
(LO) and Tjeukemeer (TJ) as determined in September av-
recruitment but by different growth rates which
eraged for 1984-1990. Bars indicate SE.
result in year-classes accumulating in the length
ranges up to 25 cm in Loosdrecht Lake.
abundant in Loosdrecht Lake (Fig. 6a). Only the
smallest species, B. longirostris and C. sphaericus
Zooplankton did not differ significantly (p> 0,35 and p> 0.39
respectively, t-test) in size composition, but all the
The lakes mainly differ in the absence of Daphnia other species are significantly larger (p < 0.0 1 or
hyalina in the Loosdrecht Lakes in most years smaller, t-test) in Tjeukemeer (Fig. 6b).
(Fig. 6a), whereas it is an important daphnid in
Tjeukemeer. The abundances of Bosmina
coregoni, B. longirostris, and cyclopoid copepods Phytoplankton, chlorophyll-a, Secchi-depth and to-
is not significantly different (p>0.10, p>0.35, tal-P
and p>0.30 respectively, t-test). Only Daphnia
cucullata was significantly (p<O.OI, t-test) more Filamentous cyanobacteria constituted the bulk
of the phytoplankton in both lakes. They were
significantly (p< 0.0001, t-test) more abundant in
Bream the Loosdrecht Lakes, in which the average num-
ber of cells amounted to ca. 150000 ml- 1. In
80 contrast the average density in Tjeukemeer was
ca. 30000 ml- 1 in Tjeukemeer (Fig. 7), but the
60 filaments were much thicker and longer (H.
Hoogveld, pers. comm.) than in Loosdrecht Lake
~
;,; and therefore the chlorophyll-a concentration
(J) 40
(J)
hardly differed (Fig. 8). In both lakes the green
§ algae had relatively low densities, but their num-
Iii
20 bers were in Tje\lkemeer much higher than in
Loosdrecht Lake (ca. 500 vs. 100 ml- 1). Dia-
o toms formed the third relative abundant phyto-
5 10 15 20 25 30 35 40 45 50
Length (em)
plankton group and were on average more abun-
_ TJeukemeer ~ Loosdrecht
dant (p < 0,01, t-test) in Loosdrecht Lake (2000
vs. 1000 ml- 1) (Fig. 7). Average chlorophyll-a
Fig. 4. Size distribution of bream in Loosdrecht Lake and
Tjeukemeer expressed in biomass per 5 cm length class aver- concentrations in both lakes were around
aged for the periods May-September in 1984-1990. 150 J.tg 1- 1 and the Secchi-depth (ca. 35 cm) did
 91

300 , - - - - - - - - - - - - - - - - - - - - - - - , 200 8000

200 E~
~ &iX 4000
'"~
100

.. -
~
'iii =r-
~ 100
U i

0 o
I CI Irn ~ o
DH BC CC DC BL CS SA LO TJ SA LO TJ SA LO TJ

_ LOOSDRECHT ~ T -ELKEMEER _ cyan. ~ chlor. ~ diat.

Fig. 7. Average density of the three main phytoplankton

120
(cyan = Cyanobacteria, left scale; chlor. = chlorophyta, right
scale; dial. = diatomeae, right scale) groups for the periods
8 May-September in 1984-1990 in two Loosdrecht lakes
f* 80 (BR = Lake Breukeleveen, LO = Loosdrecht Lake) and in

II
.c Tjeukemeer (TJ). Bars indicate SE.
~
. I
I
0 40
r-----------------------,
~'"
>
OAO

«
0.30
o
DH BC CC DC BL CS

_ LOOSDRECHT c::::LJ T-ELKEtvEER

Fig. 6a. Average density of the most abundant zooplankton
species (DH = Daphnia hyalina, BC = Bosmina coregoni,
CC = cyclopoid copepods, DC = Daphnia cucculata,
BL = Bosmina longirostris, CS = Chydorus sphaericus) for the
periods May-September in 1984-1990. Bars indicate SE.
6b. Average length of most abundant zooplankton species
(see 6a for abbreviations) for the periods May-September in
1984-1990. Bars indicate SE.
not differ either (Fig. 8). Surprisingly the total-P
concentration was 2-3 times higher (p<O.OOI,
t-test) in Tjeukemeer (ca. 250 j.Lg 1- 1), but the
daily external P-Ioad is even 50 times higher in
Tjeukemeer (50-100j.LgI-l) (Van Huet, 1991).
Discussion
Considering the potential predation pressure on
zooplankton the main difference between the fish
communities in Loosdrecht Lake and Tjeukemeer
is the relatively small individual size of bream in
Loosdrecht Lake. Because the bream population
makes up almost 70-80% of the total fish biom-
0.20
0.10
0.00
Looscrecht T jeU<emeer
_ Secchi (m) ~ Chl-a (mglU ~ Tot-P (mg/LI
Fig. 8. Secchi-depth (m), chlorophyll-a (mg 1- 1) and total-P
(mg 1- 1) as an average for the periods May-September in
1984-1990. Bars indicate SE.
ass the predation effect on zooplankton will con-
sequently be different in the two lakes. In Tjeu-
kemeer the bream population can feed only on
large zooplankton, i.e. Daphnia hyalina > 1 mm,
because of the size composition of the population
(Lammens, 1985). When the Daphnia density is
lower than 20 1- 1 large bream will switch to
benthivorous feeding (Lammens & Hoogen-
boezem, 1991; Hoogenboezem et aI., 1991) and
therefore unable to eradicate large zooplankton.
This is probably the main reason why D. hyalina
never disappears from Tjeukemeer. In Loos-
92
drecht lake, however, the bream population is
composed of relatively smaller individuals and
can therefore also feed on the smaller zooplank-
ton and will not switch readily to benthivorous
feeding because its benthivorous feeding effi-
ciency is much lower than that of large bream
(Lammens et aI., 1987). The smaller bream can
thus continue feeding on the smaller Bosmina
coregoni when D. hyalina has disappeared and will
therefore prevent the latter species from recover-
ing. The lower average sizes of B. coregoni and
cyclopoid copepods in Loosdrecht Lake support
the conclusion that there is a much higher preda-
tion pressure on the zooplankton in Loosdrecht.
However, the above does not explain the high
biomass of small bream in Loosdrecht. Over the
last seven years there was no significant differ-
ence in recruitment of 0 + bream in the two lakes
and therefore the density feedback mechanism
cannot be responsible for slow growth. Already in
the first year growth of bream is slower in Loos-
drecht Lake and up to a length of 25 cm growth
remains slower than in Tjeukemeer (Van Densen
et al., 1986), probably because the production of
zooplankton is higher in Tjeukemeer than in
Loosdrecht. So as the growth of the zooplankton
is relatively slow in Loosdrecht, the larger spec-
imens selected by bream are likely to be more
slowly replaced by growth and thus the size dis-
tribution will be more easily affected. The high
density of filamentous cyanobacteria and low
density of Chlorophyceae in lake Loosdrecht sus-
tain the hypothesis that the feeding conditions for
large daphnids are poorer in Loosdrecht (Daw-
idowicz, 1990; Gliwicz, 1990) and even a decrease
in the predation pressure on the zooplankton
would hardly change this picture.
In 1989 a large scale experiment was performed
in Lake Breukeleveen, one of the Loosdrecht
lakes. The fish stock was reduced to 57 kg ha - 1,
but D. hyalina did not develop, although the lake
was stocked with D. hyalina and D. magna
(12 ind m - 3), and the turbidity and chlorophyll-
a concentration remained high (Van Donk et al.,
1990). It is, however, debatable whether the re-
duction was successful especially regarding the
very successful recruitment of 0 + fish in 1989,
and the fact that older fish could not be prevented
from immigrating. Despite this set-back it was
clear that a period of a few months (up to July)
when the fish biomass was still low, was not suf-
ficient for D. hyalina to develop and that the high
densities of filamentous cyanobacteria and the
high sensitivity of the lake for resuspension by
wind were the most important factors preventing
the increase in transparency (see also Gons et al.,
1986). However, in Tjeukemeer enclosure exper-
iments in which fish was excluded also failed to
show an increase in transparency, even in the
presence of high Daphnia hyalina densities (Rich-
ter et al., 1987). A major factor is probably the
density of filamentous cyanobacteria at the mo-
ment the experiment starts. It seems, however,
almost impossible to have a high D. hyalina den-
sity before the filamentous cyanobacteria develop.
Going down the food chain trying to explain
the difference between the two lakes, the factor of
hydrology determining the external nutrient load
and inoculation with phytoplankton and zoop-
lankton seems crucial. Tjeukemeer is located in a
drainage area of agricultural land and receives
25-50 mg P m - 2 day - 1 (Van Huet, 1991), which
is 75-150 times more than the Loosdrecht lakes
(Van Liere et aI., 1990b). Although only less than
10% of this P is retained in Tjeukemeer com-
pared with 60 % in Loosdrecht, it is still 30 times
more. Average chlorophyll-a and turbidity are
similar in both lakes, but total-P concentration is
2-3 times higher in Tjeukemeer. The largest part
ofthe external load comes from the drainage water
from the agricultural area. However, during sum-
mer when evaporation exceeds precipitation
water from the less eutrophic Lake IJ sselmeer is
let in and replaces the 2000 ha Tjeukemeer water
several times inoculating the lake with phyto-
plankton, zooplankton and even fish larvae,
particularly smelt (Van Densen & Vijverberg,
1982). So probably a large part of the difference
in phytoplankton composition in the two lakes
can be explained by the hydrological relation to
IJ sselmeer. It is clear, however, that Tjeukemeer
receives and contains more nutrients than Loos-
drecht Lake and has probably a higher phyto-
plankton production, leading to a faster growth of

particularly large zooplankton and also more
rapid fish growth. But despite the relative high
D. hyalina density it seems also very difficult to
increase the transparency in Tjeukemeer, which is
mainly affected by filamentous cyanobacteria.
References
Carpenter, S. R., J. F. Kitchell & J. R. Hodgson, 1985. Cas-
cading trophic interactions and lake productivity. Bio-
Science 35: 634-639.
Dawidowicz, P., 1990. Effectiveness of phytoplankton control
by large-bodied and small bodied zooplankton. Hydrobio-
logia 200/201: 43-49.
De Haan, H. & J. R. Moed, 1984. Phosphorus, nitrogen and
chlorophyll-a concentrations in a typical Dutch polder lake,
Tjeukemeer in relation to its water regime between 1968
and 1982. Wat. Sci. Technol. 17: 733-743.
De Nie, H. W., H. J. Bromley & J. Vijverberg, 1978. Distri-
bution patterns of zooplankton in Tjeukemeer, The Neth-
erlands. J. Plankton. Res. 2: 317-334.
Gons, H. J., R. Van Keulen & R. Veeningen, 1986. Effects of
wind on shallow lake ecosystem. Redistribution of particles
in the Loosdrecht lakes. Hydrobiol. Bull. 20: 109-120.
Gliwicz, Z. M., 1990. Why do c1adocerans fail to control algal
blooms? Hydrobiologia 201: 83-97.
Gulati, R. D., E. H. R. R. Lammens, M-L. Meyer & E. Van
Donk (eds), 1990. Biomanipulation, tool for water man-
agement. Kluwer. pp 628.
Gulati, R. D., 1990. Structural and grazing responses of
zooplankton community to biomanipulation of some Dutch
water bodies. Hydrobiologia 200/201: 99-119.
Hoogenboezem, W., E. H. R. R. Lammens, Y. Van Vugt &
J. W. M. Osse, 1991. A model for switching between
particulate-feeding and filter-feeding in the common bream
Abramis brama (L.). Envir. BioI. Fishes, 33: 13-21.
Hosper, S. H. & E. Jagtman, 1990. Biomanipulation addi-
tional to nutrient control for restoration of shallow lakes in
The Netherlands. Hydrobiologia 200/201: 523-534.
Kott, P.. 1953. A modified whirling apparatus for subsampling
of plankton. Aust. J. mar. Freshwat. Res. 4: 387-393.
Lammens, E. H. R. R.. 1982. Growth, condition and gonad
development of bream (Abramis brama) in relation to its
feeding conditions in Tjeukemeer. Hydrobiologia 95: 311-
321.
Lammens, E. H. R. R., 1985. A test of a model for filter feed-
ing in planktivorous bream Abramis brama. Envir. BioI.
Fish. 13: 288-296.
Lammens, E. H. R. R.,J. Geursen & P. J. MacGillavry, 1987.
Diet shifts, feeding efficiency and coexistence of bream
(Abramis brama), roach (Rutilus rutilus) and white bream
(Blicca bjorkna) in eutrophic lakes. In: Proceedings Fifth
Congress of European Ichthyologists p. 153-162.
93
Lammens, E. H. R. R., A. Frank-Landman, P. J. Mac
Gillavry & B. Vlink, 1991. The role of predation and com-
petition in determining the distribution of common bream,
roach and white bream in Dutch eutrophic lakes. Envir.
BioI. Fishes, 33: 195-205.
Lammens, E. H. R. R., R. D. Gulati, M-L. Meijer & E. Van
Donk, 1990. The Biomanipulation Conference in Amster-
dam: A synthesis. Hydrobiologia 200/201: 619-627.
Lammens, E. H. R. R. & W. Hoogenboezem, 1991. Diets and
feeding behaviour. In: 'Cyprinid fishes: Systematics, Biol-
ogy and Exploitation' (ed. Winfield & Nelson). Chapman
& Hall. London: 353-376.
McQueen, D. J., J. R. Post & E. L. Mills, 1986. Trophic re-
lationships in freshwater pelagic ecosystems. Can. J. Fish.
aquat. Sci. 43: 1571-1581.
McQueen, D. J., M. R. S. Johannes & J. R. Post, 1989.
Bottom-up and top-down impacts on freshwater pelagic
community structure. Ecol. Monogr. 59: 289-309.
Moed, J. R., 1973. Effect of combined action of light and
silicon depletion on Asterionella formosa Hass. Verh. int.
Ver. Limnol. 18: 1367-1374.
Moed, J. R. & H. L. Hoogveld, 1982. The algal periodicity in
Tjeukemeer during 1968-1978. Hydrobiologia 95: 205-223.
Moss, B., 1990. Engineering and biological approaches to the
restoration from eutrophication of shallow lakes in which
aquatic plant communities are important components. Hy-
drobiologia 200/201: 367-379.
Persson, L., G. Andersson, S. F. Hamrin & L. Johansson,
1988. Predator regulation and primary production along the
productivity gradient of temperate lake ecosystems. In:
S. R. Carpenter (ed.), Complex interactions in lake com-
munities. Springer-Verlag, New York: 45-68.
Richter, A. F., G. J. Schrotenboer, J. Vijverberg & S. Visser,
1987. Enclosure studies (in Dutch). Limnologisch Instituut
Oosterzee/Nieuwersluis. 161 pp.
Scheffer, M., 1990. Multiplicity of stable states in fresh water
systems. Hydrobiologia 200/201: 475-487.
Van Densen, W. L. T. & J. Vijverberg, 1982. The relations
between 0 + fish density, zooplankton size and vulnerabil-
ity of pike-perch, Stizostedion lucioperca, to angling in the
Frisian lake. Hydrobiologia 95: 321-336.
Van Densen, W. L. T., C. Dijkers & R. Veerman, 1986. The
fish community of the Loosdrecht lakes and perspectives
for biomanipulation. Hydrobiol. Bull. 20: 147-163.
Van Donk, E., M. P. Grimm, R. D. Gulati, P. G. M. Heuts,
W. A. De Kloet & E. Van Liere, 1990. First attempt to
apply whole-lake food-web manipulation on a large scale in
The Netherlands. Hydrobiologia 201: 291-301.
Van Huet, H., 1991. Modelling water transport and phospho-
rus eutrophication in an interconnected lake system. A sce-
nario study. Ph.D. Thesis. Agricultural University Wa-
geningen. 129 pp.
Van Liere, L. Van, R. D. Gulati, E. H. R. R. Lammens &
R. G. Wortelboer, 1990a. Phosphorus dynamics following
restoration measures in Loosdrecht lakes. Hydrobiologia
191: 89-96.
94
Van Liere, L., O. F. R. Van Tongeren, L. Breebaart & W.
Kats, 1990b. Trends in chlorophyll and total phosphorus in
Loosdrecht lakes. The Netherlands. Verh. int. Ver. Limnol.
24: 707-710.
Van Liere, L. & J. H. Janse 1997 Restoration and resilience
to recovery of the Lake Loosdrecht ecosystem in relation
to its phosphorus flow. Hydrobiologia 223: 95-104.
Van Tongeren, O. F. R., L. Van Liere, R. D. Gulati, G.
Postema & N. Boesewinkel-de Bruyn 1992. Multivariate
analysis of the plankton communities in the Loosdrecht
lakes: relationship with the chemical and physical environ-
ment. Hydrobiologia 233/105-117.
Hydrobiologia 233: 95-104, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 95
© 1992 Kluwer Academic Publishers.

Restoration and resilience to recovery of the Lake Loosdrecht ecosystem

in relation to its phosphorus flow

Louis Van Liere 1,2,* & Jan H. Janse 2

1 Limnological Institute, Rijkstraatweg 6, NL-3631 AC Nieuwersluis, The Netherlands; 2 National Institute
of Public Health and Environmental Protection, P.O. Box 1, NL-3720 Bilthoven, The Netherlands
(* author for correspondence)

Key words: water management, food-web, phosphorus flow, lake restoration, recovery, resilience

Abstract

A reduction in external phosphorus loading since 1984 to Loosdrecht lakes system by the dephospho-
rization of the inlet water, yielded only minor effects in Lake Loosdrecht. This reduction measure turned
out to have decreased the loading only by a factor of two. A conceptual model was constructed based
on laboratory measurements to describe phosphorus flow in the lake ecosystem for the summer of 1987.
The role of zooplankton and fish was more important in phosphorus recycling than diffusion at the
sediment-water interface. The input and output of phosphorus of the lake were at equilibrium and
therefore, further reduction in external loading was needed for recovery. The results of the conceptual
model agreed well with the output of the mathematical model PCLOOS. Additional measures such as
dredging, flushing, chemomanipulation, or biomanipulation would be ineffective at the present level of
external loading. Only a significant further reduction in external input will restore Lake Loosdrecht's
water quality over a long period of time.

Introduction Van Liere, 1992). To restore the water quality, a

The Loosdrecht lakes system (Fig. 1) consists of
the three shallow (average depth about 2 m) in-
terconnected lakes, Loosdrecht, Vuntus, and
Breukeleveen (for detailed description, see Gulati
et al. 1991a; Hofstra & Van Liere, 1992). The
banks and ditches of the Kievits Area also form
a part of the system influencing its water and
phosphorus balances. The eutrophication of these
lakes started from the beginning of the 20th cen-
tury from various sources (Engelen et al., 1992).
The trophic status of these lakes, passing through
various intermediate stages, has become eutro-
phic, with cyanobacteria and detritus dominating
the particulates in the water column (Hofstra &
series of measures have been taken to reduce
phosphorus input. The most important step was
the replacement of inflow to the lake in 1984; the
inflow from River Vecht, with its phosphorus level
at 1-2 mg PI-I, was switched to Amsterdam-
Rhine Canal by means of a pipeline (Fig. 1). The
Canal water was dephosphorized before entering
the lakes by coagulation with iron (III)-chioride
followed by sedimentation of the resulting floc
(Van der Veen et al., 1987). L. Breukeleveen and
L. Vuntus have shown little effect by the diver-
sion, probably because they are situated far from
the inlet. Thus, this paper will focus on L. Loos-
drecht near the inlet.
To understand the persistence of trophic status
96

conceptual model of phosphorus cycling through

various ecosystem compartments was made as
presented below. Using various parameters mea-
sured in water and various compartments of the
ecosystem, a mathematical model was developed
(lanse & Aldenberg, 1990). This model output
was compared with empirical data, in order to
test its possibilities to simulate further water man-
agement measures in L. Loosdrecht (J anse et at.,
1992). The needs for further (and additional)
measures to restore Loosdrecht lakes are indi-
cated.

Reduction of external phosphorus

Earlier calculations (Province of Utrecht, 1973,

1980) predicted that the replacement of inflow
....... :.;~:: ... from R. Vecht with dephosphorized water from
. ::: :~.:.::.':.

.~.=/:?~.::/~ ;X:·://}:::~~ the Amsterdam-Rhine Canal would reduce the

Fig. 1. The Loosdrecht lakes area. Arrows indicate the sluices
used before 1984 to supply the lakes with water from the River
Vecht. The Reservoir, L. Eastern Loenderveen (LEL) and L.
Western Loenderveen (LWL) are hydrologically separated
from the L. Loosdrecht and L. Vuntus. Asterix: Limnological
Institute.
in the lake, an accurate phosphorus budget for the
lake was first estimated (Engelen et al., 1992). A
phosphorus load of the lake system (L. Loos-
drecht, L. Vuntus and L. Breukeleveen) from 1.35
to 0.10 g P m - 2 Y- I. Thus, the lake system was
expected to react shortly to the replacement.
However, no significant changes have been ob-
served (Table 1).
In L. Vuntus and L. Breukeleveen, the external
phosphorus load did not change. (Engelen et al.,
1992). For the whole lake area, the external phos-
phorus load after 1984 was 0.36-0.46 g P

Table 1. External phosphorus load (g Pm - 2 Y- 1) of the whole Loosdrecht lakes' system and its separate compartments, excluding
Kievits Area, before and after the restoration measures. Data according to Van Liere et al., 1991.

Lake system Before the After measures Increased load: Present estimated
measures (average for leakage of sluices load (if no extra
(average for 1984 and 1985) (average for measures were to be
1982 and 1983) 1986 and 1987) taken)

Loosdrecht 0.69 0.31 0.44 0.35

(excluding Kievits area)

Breukeleveen 0.57 0.59 0.59 0.60

Vuntus 0.24 0.26 0.25 0.25

Whole lake system 1.05 0.36 0.46 0.37

(including Kievits area)

m - 2 Y- 1(Table 1), which is much higher than the
predicted 0.1 g P m - 2 Y- 1 by the earlier reports
(Province of Utrecht, 1973, 1980), probably be-
cause they did not account for the load from the
agricultural hinterland. The load to L. Loosdrecht
was even higher in 1986 and 1987 because of a
severe leakage of the Southernmost sluice. This
leakage, however, did not have any effect on L.
Vuntus and L. Breukeleveen.
Besides, the earlier calculation of external load-
ing appeared to be an overestimate. Before 1984,
water from R. Vecht was supplied mainly through
two sluices (Fig. 1), passing through the Kievits
Area with its banks and ditches where wind in-
fluence, and concomitant mixing, were negligible.
The water lost a part of its phosphorus passing
this area, resulting in a lower load for L. Loos-
drecht. Phosphorus concentration in Kievits Area
sediments was high when compared to other lakes
(Boers et al., 1984).
The phosphorus balance (Engelen et al., 1992)
revealed that the external load to the whole lake
system had been reduced from 1.05 to
0.37 g P m - 2 Y- I. The reduction of external load
to L. Loosdrecht resulted in decrease only by a
factor of about two (0.69-0.35 g P m - 2 Y- I, Ta-
ble 1).
However, certain parts of the system showed
clear responses. Phosphorus release rate has de-
creased from 1 g P m - 2 d - 1 in 1983 to
0.3 g P m - 2 d - 1 in 1990 (Keizer & Sinke, 1992).
The concentration of total phosphorus in L.
Loosdrecht corrected for autocorrelation de-
creased from 1984 to 1988 by 12 ± 2 /lg 1- 1P Y- I,
but the concentration of chlorophyll a did not
react (Van Liere et al., 1990a; Van Tongeren et al.,
1992). Light energy was important as a growth-
limiting factor before 1985, but later phosphorus
has become more important (Rijkeboer et al.,
1991; Gons et al., 1992). This shift in growth-
limiting factor was also reflected in the C/P ratio
of the seston ( < 150 /lm) which increased from 90
to 145 between 1985 and 1989 (Gulati et at.,
1991 b). Other parts of the food chain showed
little reaction (Gulati et at., 1992; Lammens et at.,
1992).
97
The flow of phosphorus in Lake Loosdrecht
To gain more insight in the factors contributing to
the persistence of the ecosystem's trophic status,
a conceptual model for phosphorus flow was con-
structed. The model was based largely on the data
from the scientific projects by the work group
Water Quality Research Loosdrecht Lakes
(Loogman & Van Liere, 1986; Van Liere & Gulati,
1992). The input and output of the system were
derived from hydrological data (Buyse, 1988).
Data on phosphorus-dynamics at the sediment-
water interface were taken from Keizer & Sinke
(1992) and Gons & Van Keulen (1989). Primary
production was measured and calculated accord-
ing to Van Liere et al. (1986a). A striking simi-
larity was noted between the model that described
primary productivity and the results derived from
the mass balances obtained in Laboratory Sys-
tem Enclosures (Rijkeboer et al., 1991).
Zooplankton grazing, consumption, assimila-
tion, and egestion were determined according to
Gulati (1984) and Gulati et al. (1992). Phospho-
rus flows within seston and zooplankton were
calculated on the basis of the C/P ratios, of zoop-
lankton and their seston-food, which were mea-
sured by the 14C-tracer method in the laboratory.
Zooplankton (> 150 /lm) excretion was com-
puted from zooplankton biomass and water tem-
perature using the relationship between tempera-
ture and specific excretion rate (Den Oude &
Gulati, 1989). Mineralization, which was mea-
sured only three times in 1986, was not included
in the diagramme. The phosphorus mineraliza-
tion rate of sestonic matter was in the same order
of magnitude as phosphorus consumption rate
(Van Liere et al., 1986b). This suggests that the
mineralization of seston occurs mainly in the
water column and the top layer of sediments
(epipelon) in contact with oxygenated water (see
also Otten et al., 1992).
Data of phosphorus flow to and from water
plants were from calculations by Malthus et al.
(1990).
Phosphorus recycling in the fish compartments
was derived from Lammens et at. (1992), rates of
consumption of zooplankton and benthic fauna
98

were derived from observations in the fish intes-
tine and fish growth rate (Lammens, personal
communication).
The phosphorus flow diagramme shown in
Fig. 2 represents the time period between April
and September, 1987. It was shown earlier that
phosphorus dynamics did not change appreciably
(Van Liere et al., 1990b); this was not surprising,
considering the relatively small changes in exter-
nal phosphorus loading (Table 1). However, there
are some striking features in the flow diagramme
which might be of importance for future develop-
ments of the Lake Loosdrecht ecosystem.
- The concentration of soluble phosphorus was
very low at 12mgPm- 2 (6f.1.g1-1). Of this,
about 30-40 % was soluble reactive phospho-
rus (SRP). This means that whatever its source,
available phosphorus would be taken up very
fast by either phytoplankton, detritus, or both.
- Sediments contained a vast amount of phos-
phorus compared to other compartments. Ap-
parently, only a small part of this was released
to overlying water by diffusion (Keizer & Sinke,
1992). Diffusing phosphorus was reported to
be largely from the mineralization of organic
particles in the sediments (Sinke & Cappen-
berg, 1988). Adsorption was considered to be
small because the aerobic layer in which it takes
place was only a few-millimeters deep.
- Desorption of phosphorus after resuspension
might be an important source of phosphorus
for the phytoplankton. However, it was diffi-
cult to separately quantify adsorption/
desorption processes between living and dead
material within the complex resuspension-
sedimentation compartment. Rijkeboer et al.
(1991) showed that the exchange of phospho-
rus between detritus and phytoplankton com-
ponents in the seston might be of paramount
importance in phosphorus exchange.
- Phosphorus concentrations in the seston and
fish were equal. One half of phosphorus in the
water column was in fish and most fish phos-
phorus (70-80 %) was associated with bream

AsslmllatlonB~~
Eg~atlon 5 Predation ,~1 160
/ 5. 5 / Egestlon ~
9.8 / Exc2r.2etlon
~1.4 Eyetton 1- 7./' ~~,..;----.o
y
Consumption
\ ,
.
Losses . j).1MOrl'jfalit •
__ \ _
160
O~- . A
Inllow

Primary production
L(/"
\
'F-e-e-di-n9-

i
Output ~~

3.0 .~ -0.15~
I'
r.
I Mineralization ¢:::;. 005
;edimentation 0 c::. --::::: ~
0.3 r?~ Decay -rT\ ~
2. "'----
ResuspenSion ~Diffusion
I I
Adsorption
/) ,
~J-Y
~~
- ~ ...

~ 850

s'V:;'i/ .6
Fig. 2. Flowdiagramme of phosphorus (mg P m - 2) and its flux between compartments of Lake Loosdrecht (mg P m - 2 d - 1) for
the period April-September 1987.

(Abramis brama). If bream was of any com-
mercial value the problem of eutrophication
might have been partly solved because then it
would be harvested.
- Although zooplankton may seem unimportant
quantitatively, they are highly important for the
fast recycling of phosphorus to the pool of
available phosphorus for phytoplankton
through excretion, perhaps even more if rotifer
phosphorus excretion is also included (Gulati
et al., 1990).
- In recycling phosphorus to the available phos-
phorus pool, fish is important next to zoop-
lankton. However, a halfof phosphorus in dead
fish is in the form of calcium phosphate, an
insoluble form. Therefore, the death of fish is
a phosphorus loss factor for the ecosystem.
- Egestion products by fish and the death of phy-
toplankton, zooplankton, and fish is a source
of detritus. Detritus is partly refractive (Otten
et al., 1992) and therefore, eutrophication
builds recalcitrant debris in sediments. How-
ever, their density allows an easy resuspension
in the water column by wind action (Gons
et al., 1986), and they may play the role of sor-
bate in sorbing and desorbing the available
phosphorus for phytoplankton (Rijkeboer
et al., 1991).
- The largest flow was from the soluble phos-
phorus pool to the seston compartment at
18 mg P m - 2 d - I. This daily flow is the most
important process in the internal phosphorus
cycle of the lake ecosystem.
- Submerged plants were unimportant in phos-
phorus recycling. They virtually disappeared
from the system (Hofstra & Van Liere, 1992).
Only emergent plants or those with floating
leaves played a minor role (Malthus et at.,
1992). The balance between phosphorus uti-
lized by the plants and phosphorus release
upon their death is probably negative, meaning
a small net loss to the system. This loss was
probably less than shown in the diagram, be-
cause a part of phosphorus in plants entering
the seston compartment would become avail-
able by mineralization. Thus, the loss factor by
plants was considered negligible.
99
- Very little quantitative research has been per-
formed on benthos, although the presented es-
timation of the flow of phosphorus from
benthos to fish, which was derived from fish
intestine investigations, revealed it to be im-
portant.
- Output from the ecosystem was indicated by
the arrow 'output' (0.4 mg P m - 2 d - I) in the
seston compartment. This may not be always
true, since the discharge or outflow to the hin-
terland also contains zooplankton which would
carry P. Another main loss was downward
seepage to the Polder Bethune. This loss was
calculated to be 0.6 mg P m - 2 d - 1 (Keizer &
Sinke, 1992).
The most important point in the conceptual
model is that input and output are nearly in equi-
librium. This implies that if there is no further
reduction of external P, the system will remain in
its present state. However, the flow diagramme
was based on the summer of 1987, the time pe-
riod during which the leakage of the Southern
sluice occurred (Table 1). During other years, the
summer input was in the order of
0.75mgPm- 2 d- l . Then, the phosphorus loss
factor due to downward seepage the summer
value of 1987 was somewhat larger.
The mathematical model PCLOOS was devel-
oped to evaluate the input, output, and recycling
over various periods with a better precision (Janse
& Aldenberg, 1990; Janse et al., 1992). The model
was calibrated with hydrological and material-
balance data (Engelen et al., 1992; Buyse, 1988)
as well as with the measurements made by the
WQL work group. Following the calibration, it
was used to simulate the results of a further re-
duction in external input and additional measures
(dredging, flushing, biomanipulation; Janse et al.,
1992). The flow rates shown in the flow dia-
gramme were also calculated for the same time
period (Fig. 3). The model output of these param-
eters was strikingly similar to those in the diagram,
including those of the sediment compartment and
its exchange with the water compartment.
In open water the results were variable, but the
difference was no more than by a factor of two.
The reason for this difference was that the math-
100

150 e estion + death

0.2

outflow
0.4

ads.P

sediment

seepage

Phosphorus fluxes [mgP m-2 d-1]:

>8 0.5 - 1.25
3-8 0.2 - 0.5
1.25 - 3 -- - - - < 0.2
Fig. 3. Phosphorus content (mg P m - 2) and fluxes (mg P m - 2 day - I) and phosphorus amounts in various compartments of Lake
Loosdrecht for the time period April-September 1987 as calculated by the mathematical model PCLOOS.

ematical model separated seston into two frac-
tions, phytoplankton and detritus, whereas the
flow diagramme used them as a single compart-
ment. Calculated growth rates were different from
the estimated values based on primary produc-
tion measurements (Van Liere et al., 1986). The
calculated phosphorus flow from zooplankton
deviated only a little from the measurements.
Overall, the model gave a good description of
phosphorus flow through various compartments
of the ecosystem. It predicted with a reasonable
accuracy the outcome of further reduction of ex-
ternal phosphorus load over a short term, point-
ing out at the same time additional measurements
to be taken to reduce primary production (Janse
et al., 1992).

Additional measures
If an ecosystem does not react instantaneously on
the external phosphorus load additional measures
are often called for. Such measures are reviewed
here, especially in regard to their applicability or
otherwise in the Loosdrecht lakes.
Further reduction of the external phosphorus load
For the Loosdrecht lakes, any further reduction
of external phosphorus load is technically diffi-
cult, but possible. As mentioned earlier, a reduc-
tion of the external load by a factor of two was
not sufficient to induce significant responses from
the ecosystem. Engelen et al. (1992) have pointed
out several possible hydrological measures which
could decrease the load an additional 50 %, in-
cluding the diversion (or dephosphorization, De
Ruiter, 1992) of the Eastern hinterland water with
its high phosphorus concentrations.
Flushing
Hosper & Meijer (1986) and J agtman et al. (1992)
have shown positive results with flushing in the
recovery of L. Veluwe and L. Wolderwijd. How-
ever, the chemical composition of the flushing
water with high concentration of calcium and ni-
trate facilitated chemomanipulation (J agtman
et al., 1992). In L. Loosdrecht the situation is
more problematic. The inlet point of the dephos-
phorized Amsterdam-Rhine Canal water is close
to the discharging pumping station. A mixing time
of at least a week is needed and thus flushing has
to be repeated several times. It also does not deal
with the problem of organic deposits on the lake
bottom. Therefore, it is not recommended.
Dredging
Keizer & Sinke (1992) have shown that a major
phosphorus exchange at the sediment-water in-
terface involves the diffusion of mineralized phos-
phorus to the overlaying water. Therefore, it is
101
unlikely that phosphorus release will be reduced
by removing the top layer of sediments. In fact
Van Liere et al., (1986) reported only 30 % reduc-
tion in phosphorus release rate from Lake Loos-
drecht sediments following the removal of its top
layer. With the present high external load of phos-
phorus, organic matter in the top layer will be
replenished rapidly. However, removing more
than just the top layer will reduce resuspension
(Gons & Van Keulen, 1989), which is an impor-
tant factor in phosphorus exchange between
seston and sediments (Rijkeboer et al., 1990).
Dredging may also increase downward seepage
(Engelen et al., 1992), which can readily be com-
pensated for by an increased supply of dephos-
phorized water. Although dredging would in-
crease external phosphorus load slightly per unit
area, its load per unit volume would decrease
relatively. The cost of dredging L. Loosdrecht is
estimated to be 35 ± 10 106 ECU. Technical
problems of dredging a large amount of organic
deposits with its specific density close to that of
water have not yet been solved (Van der Does
et al., 1992), not to speak of the disposal of the
dredged material. Considering the present high
external load and the equilibrium situation be-
tween its input and output, it is questionable
whether dredging will significantly help the recov-
ery of L. Loosdrecht.
Biomanipulation
As can be derived from Dawidowicz et al. (1988),
the removal of a large amount of bream (Abramis
brama) failed to improve L. Breukeleveen (Van
Donk et aI., 1990), because this top-down control
is effective only when the food quality for large
Daphnia can sustain their growth. Apparently, this
was not the case for L. Breukeleveen, one of the
lakes in the Loosdrecht lakes' area. Furthermore,
with the present high amount of organic deposits
in the lake where wind action plays such a dom-
inant role, any increase in water transparency re-
sulting from biomanipulation may not be enough
to sufficiently enhance the growth of submerged
plants to remove any significant amount of phos-
102
phorus. Thus, this measure does not appear to be
suitable for L. Loosdrecht (see also Van der Vlugt
etal., 1992).
However, since so many phosphorus is stored
in the fish compartment (Fig. 2) annual reduction
of fish stock may be a tool for removing phos-
phorus, thus also diminishing flows to and from
other compartments.
Chemomanipulation
In laboratory chemomanipulation experiments,
the addition ofiron(III)-chloride to L. Loosdrecht
sediments was unsuccessful (Keizer & Sinke,
1992), because the aerobic layer of the sediments
was extremely thin. With significant downward
seepage in the lake, the chemical would also move
out of the system in weeks or months when the
aerobic layer is mixed with the anaerobic part by
wind. In a whole lake experiment in a compara-
ble lake, the addition of iron(III)-chloride also
failed to improve water quality (Boers et al.,
1992).
External nitrogen reduction
Based on a comparative study of nitrogen-fixing
and fixed-nitrogen utilizing strains of cyanobac-
teria, Zevenboom & Mur (1980) concluded that
in turbid lakes, the former could not compete with
the latter because of unfavourable energy balance.
This implies that if the input of external N is re-
duced, it would increase light transmission, thus
favouring nitrogen-fixing species. For example,
Oscillatoria spp. and Prochlorothrix hollandica will
be replaced by species as Aphanizomenon fios-
aquae. However, Burger-Wiersma (1991) re-
ported that nitrogen fixation was already an im-
portant component in the nitrogen budget of
Dutch eutrophic lakes. External N reduction thus
may effect the composition of biota, but not bio-
mass production.
Return of the upwelling water from the ice-pushed
sand-ridge of Het Gooi
At the beginning of this century, the Loosdrecht
lake system used to be oligotrophic because of the
low phosphorus level in the upwelling water from
the Eastern Sandridge (Hofstra & Van Liere,
1992). It was thus suggested that an optimum
corrective measure would be the return of this
water (Engelen & Schot, 1989). However, the
quality of ground water in the Eastern hinterland
has degraded by the inflow from the Loosdrecht
lakes and agricultural activities. Furthermore, the
water table in that area has been lowered, result-
ing in the oxydation of peat in the region. If the
flow of the upwelling water is restored, its present
phosphorus level would be more than 1 mg P I - I
(Hettling, 1985). Such water should not enter the
lakes; rather it should be diverted as suggested by
Engelen et al. (1992) in the case of waters from
the hinterland.
The return of the former upwelling water can be
achieved by inundating the Polder Bethune.
However, this is not practical, because Polder
Bethune water is the source of the processed
drinking water for the city of Amsterdam with an
important socio-economic function. Further-
more, the inundation would prevent downward
seepage from the Loosdrecht lake system, thus
removing a phosphorus sink of the system. Be-
sides, it will also increase upwelling water in the
Eastern hinterland, which will in turn force a large
volume of water with a high phosphorus level to
flow into the lakes (Hettling, 1985). Therefore, the
inundation of the Polder Bethune will mean an
increased net external phosphorus load to L.
Loosdrecht than now.
Conclusions
A reduction of external phosphorus load to L.
Loosdrecht did not diminish the in-lake phospho-
rus sufficiently enough to restore the clear water
in which macrophytes dominate primary produc-
tion. However, with the reduction it has become
a P-limited system in a relatively short time. The

input and output of phosphorus appear to be in
equilibrium. Therefore, any additional measures
not involving the curtailment of the present load-
ing will fail. A further reduction in external phos-
phorus load will decrease phytoplankton produc-
tion with concomitant reduction in organic
accumulation in sediments. It will also facilitate
phosphorus loss from the system by downward
seepage and discharge and reduce the amount of
phosphorus recycling in the food web. These re-
sults would eventually bring the ecosystem to a
new equilibrium with a lower external input, which
can maintain an acceptable water quality through
the future.
However, the eutrophication of the system
which has been going on for decades is not ex-
pected to be reversed or solved within a few more
years. The timespan needed for recovery may be
as long as the time needed to eutrophicate the
system.
References
Boers, P. C. M., J. W. Th. Bongers, A. G. Wisselo & Th. E.
Cappenberg, 1984. Loosdrecht lakes restoration project:
sediment phosphorus distribution and release from the sed-
iments. Verh. int. Ver. Limnol. 22: 842-847.
Boers, P. C. M., J. Van der Does, M. P. Quaak, J. C. Van Der
Vlugt & P. Walker, 1992. Fixation of phosphorus in lake
sediments using iron(III)chloride: experiences, expecta-
tions. Hydrobiologia 233: 211-212.
Burger-Wiersma, T., 1991. Inventarisatie van de stikstoffix-
atiecapaciteit in twee ondiepe Nederlandse meren ge-
durende het groeiseizoen van 1990 (Inventarization of ni-
trogen fixing capacity in two shallow Dutch lakes during the
growth season of 1990). Technical Report University of
Amsterdam 65 pp.
Buyse, J. J., 1988. Water- en stofbalansmodellen voor de
Loosdrechtse Plassen (Water- and matterbalances of the
Loosdrecht lake system). WQL report 1988-7. 87 pp.
Dawidowicz, P., Z. M. Glicwicz & R. D. Gulati, 1988. Can
Daphnia prevent a blue-green algal bloom in hypertrophic
lakes? A laboratory test. Limnologica 19: 21-26.
Den Oude, P. J. & R. D. Gulati, 1989. Phosphorus and ni-
trogen excretion rates of zooplankton from the eutrophic
Loosdrecht lakes, with notes on other P-sources for phy-
toplankton requirements. Hydrobiologia 169: 379-390.
De Ruiter, M., 1992. Water quality research in the Loosdrecht
lakes: proposals for the follow-up restoration measures.
Hydrobiologia 233: 161-163.
Engelen, G. B. & P. P. Schot, 1989. Analyse van de water-
103
systemen in het GooisfUtrechts stuwwallengebied, de
Vechtstreek en het Eemdal, probleemsignalering. Analysis
of watersystems in the GooifUtrecht Sandridge-area, the
lakes bordering River Vecht, and the valley of River Eem.
CHO/TNO Reports 22: 21-32.
Engelen, G. B., B. F. M. KaI, J. J. Buyse & F. Van Pruissen,
1992. The hydrology of the Loosdrecht lakes area. Hydro-
biologia 233: 21-38.
Gons, H. J., R. D. Gulati & L. Van Liere, 1986. The eutro-
phic Loosdrecht lakes: current ecological research and res-
toration perspectives. Hydrobiol. Bull. 20: 61-75.
Gons, H. J. & R. Van Keulen, 1989. De relatie tussen doorz-
icht en slib in de Loosdrechtse Plassen (Relation between
Secchi depth and suspended matter in Lake Loosdrecht).
WQL-report 1989-4. 35 pp.
Gons, H. J., T. Burger-Wiersma, J. H. Otten & M. Rijkeboer,
1992. Coupling of phytoplankton and detritus in a shallow,
eutropic lake (Lake Loosdrecht, The Netherlands). Hydro-
biologia 233: 51-59.
Gulati, R. D., K. Siewertsen & G. Postema, 1982. The zoop-
lankton: its community structure, food and feeding and its
role in the ecosystem of Lake Vechten. Hydrobiologia 95:
127-163.
Gulati, R. D., 1984. Zooplankton and its grazing as measures
of trophy in the Loosdrecht lakes. Verh. int. Ver. Limnol.
22: 848-852.
Gulati, R. D., J. Ejsmont-Karabin, J. Rooth & K. Siewertsen,
1990. A laboratory study of phosphorus and nitrogen ex-
cretion of Euchlanis dilatata lucksiana. Hydrobiologia 186/
187: 347-354.
Gulati, R. D., L. Van Liere & K. Siewertsen, 1991a. The
Loosdrecht lake system: Man's role in its creation, pertur-
bation and rehabilitation. In: O. Ravera (ed.), Terrestrial
and aquatic ecosystems. Ellis Horwood Limited, London.
pp. 593-606.
Gulati, R. D., K. Siewertsen & L. Van Liere, 1991b. Carbon
and phosphorus relationships of zooplankton and its seston
food in Loosdrecht lakes. Mem. 1st. ital. Idrobiol. 48: (in
press).
Gulati, R. D., A; L. Ooms-Wilms, O. F. R. Van Tongeren, G.
Postema & K. Siewertsen, 1992. The dynamics and the role
oflimnetic zooplankton in the Loosdrecht lakes (The Neth-
erlands). Hydrobiologia 233: 69-86.
Hettling, H., 1985. Een hydrogeologische inventarisatie van
het Loosdrechtse Plassengebied (A hydrogeological inven-
tarization of the Loosdrecht lakes area). WQL report
1985-2. 116 pp.
Hofstra, J. J. & L. Van Liere, 1992. The state of the environ-
ment of the Loosdrecht lakes. Hydrobiologia 233: 11-20.
Hosper, S. H. & M-L. Meijer, 1986. Control of phosphorus
loading and flushing as restoration methods for Lake Ve-
luwe, The Netherlands. Hydrobiol. Bull. 20: 183-194.
Jagtman, E., D. T. Van Der Molen & s. Vermij, 1992. The
influence of flushing on nutrient dynamics, composition and
densities of algae and transparency in Veluwemeer, The
Netherlands. Hydrobiologia 233: 187-196.
104
J anse, J. H. & T. Aldenberg, 1990. A mathematical model of
the P cycle in the Loosdrecht lakes. Hydrobio\. Bull. 24:
69-89.
Janse, J. H., T. Aldenberg & P. R. G. Kramer, 1992. A math-
ematical model of the phosphorus cycle in the Lake Loos-
drecht, and simulation of additional measures. Hydrobio-
logia 233: 119-136.
Keizer, P. & A. J. C. Sinke, 1992. Phosphorus in the sediment
of the Loosdrecht lakes and its implications for lake res-
toration persectives. Hydrobiologia 233: 39-50.
Lammens, E. H. R. R., N. Boesewinkel-De Bruyn, H.
Hoogveld & E. Van Donk, 1992. Phosphorus load, phyto-
plankton, zooplankton and fish stock in Loosdrecht Lake
and Tjeukemeer: confounding effects of predation and food
availability. Hydrobiologia 233: 87-94.
Loogman, J. G. & L. Van Liere, 1986a. Restoration of shal-
low lake ecosystems with emphasis on Loosdrecht lakes.
Proceedings of the Water Quality research Loosdrecht lakes
Symposium. Hydrobio\. Bull. 20: 5-259.
Malthus, T. J., E. P. H. Best & A. G. Dekker, 1990. An as-
sessment of the importance of emergent and floating-leaved
macrophytes to trophic status in the Loosdrecht lakes (The
Netherlands). Hydrobiologia 191: 257-263.
Otten, J. H., H. J. Gons & M. Rijkeboer, 1992. Dynamics of
of phytoplankton detritus in a shallow, eutropic lake (Lake
Loosdrecht, The Netherlands). Hydrobiologia 233: 61-68.
Province of Utrecht, 1973. Over de veiligstelling van de wa-
terkwaliteit van de Loosdrechtse Plassen (Securing the
Loosdrecht lakes water quality). Report of the Province of
Utrecht, Utrecht. 63 pp.
Province of Utrecht, 1980. Over de keuze van een suppletie-
model ter beveiliging van de waterkwaliteit van de Loos-
drechtse Plassen (The choice of a supply model to secure
Loosdrecht lakes water quality. Report of the Province of
Utrecht, Utrecht. 90 pp.
Rijkeboer, M., F. De Bles & H. J. Gons, 1991. The role of
sestonic detritus as a P-buffer. Mem. 1st. ita\. Idrobio\. 48:
(in press).
Sinke, A. J. C. & Th. E. Cappenberg, 1988. Influence ofbac-
terial processes on the phosphorus release from sediments
in the eutrophic Loosdrecht lakes, The Netherlands. Arch.
Hydrobio\. Beih. Ergebn. Limno\. 30: 5-13.
Van der Does, J., P. Verstraelen, P. Boers, J. Van Roestel, R.
Roijackers & G. Moser, 1992. Lake restoration with and
without dredging of phosphorus-enriched upper sediment
layers. Hydrobiologia 233: 197-210.
Van Der Veen, C., A. Graveland & W. Kats, 1987. Coagu-
lation of two different kinds of surface water before inlet
into lakes to improve the self-purification process. Wat. Sci.
Techno\. 19: 803-812.
Van DerVlugt,J. C., P. A. Walker,J. Van Der Does &A. J. P.
Raat, 1992. Fisheries management as an additional lake
restoration measure: biomanipulation scaling-up problems.
Hydrobiologia 233: 213-224.
Van Donk, E., M. P. Grimm, R. D. Gulati, P. G. M. Heuts,
W. A. De K10et & L. Van Liere, 1990. First attempt to
apply whole-lake food-web manipulation on a large scale in
The Netherlands. Hydrobiologia 200/201: 291-301.
Van Liere, L., L. Van Ballegooijen, W. A. De K1oet, K.
Siewertsen, P. Kouwenhoven & T. A1denberg, 1986a. Pri-
mary production in the various parts of Loosdrecht lakes.
Hydrobio\. Bull. 20: 77-85.
Van Liere, L., P. C. M. Boers, P. J. Den Oude, H. J. Gons,
R. D. Gulati, M. Rijkeboer, A. J. C. Sinke, J. P. R. A.
Sweerts, J. Bril & L. Postman, 1986b. Water Quality Re-
search Loosdrecht lakes, studying and modelling the im-
pact of water management measures on the internal nutri-
ent cycle. Final report to the European Community ENV
839-NL(N). 155 pp.
Van Liere, L., O. F. R. Van Tongeren, L. Breebaart & W.
Kats, 1990a. Trends in chlorophyll and total phosphorus in
Loosdrecht lakes, The Netherlands. Verh. int. Ver. Limno\.
24: 707-712.
Van Liere, L., R. D. Gulati, F. G. Wortelboer & E. H. R. R.
Lammens, 1990b. Phosphorus dynamics following restora-
tion measures in the Loosdrecht lakes (The Netherlands).
Hydrobiologia 191: 87-95.
Van Liere, L., 1. Ebert, W. Kats & J. J. Buyse, 1991. The
waterquality of Loosdrecht Lakes reviewed. Mem. 1st. ita\.
Idrobio\. 48: (in press).
Van Liere, L. & R. D. Gulati, 1992. Restoration and recov-
ery of shallow eutrophic lake ecosystems in The Nether-
lands: epilogue, Hydrobiologia 233: 283-287.
Van Tongeren, O. F. R., L. Van Liere, R. D. Gulati, G.
Postema & P. J. Boesewinkel-De Bruijn, 1992. Multivari-
ate analysis of the plankton communities in the Loosdrecht
lakes: relationship with the chemical and physical environ-
ment. Hydrobiologia 233: 105-117.
Zevenboom, W. & L. R. Mur, 1980. N 2 -fixing cyanobacteria:
why they do not become dominant in Dutch hypertrophic
lakes. In: J. Barica & L. R. Mur (eds), Hypertrophic eco-
systems pp. 123-130. Junk Publishers, The Hague.
Hydrobiologia 233: 105-117, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 105
© 1992 Kluwer Academic Publishers.

Multivariate analysis of the plankton communities in the Loosdrecht

lakes: relationship with the chemical and physical environment

O. F. R. Van Tongeren, L. Van Liere* R. D. Gulati, G. Postema & P. J. Boesewinkel-De Bruyn

Limnological Institute, Rijksstraatweg 6,3631 AC Nieuwersluis, The Netherlands; *present address: National
Institute of Public Health and Environmental Protection, P.O. Box 1,3720 BA Bilthoven,
The Netherlands

Key words: canonical ordination, lake restoration, phosphate, phytoplankton, redundancy analysis,
succession, trends, zooplankton

Abstract

Data on the water temperature, Secchi depth, water chemistry and phytoplankton and zooplankton of
the Loosdrecht lakes, collected during an eight-years monitoring study (1983-1990), were examined using
Principal Components Analysis (PCA) and Redundancy Analysis (RDA). The Loosdrecht lakes are
highly eutrophic lakes, in which several restoration measures were taken from 1970 to 1986. These
measures included the diversion of sewage by the construction of sewerage and additional reduction of
external phosphorus-loading by supplying from 1984 onwards water from the Amsterdam Rhine Canal,
which had phosphorus removed from it, and which replaced the inlet of highly eutrophic water from the
River Vecht. The descriptive statistical models based on this study show that, although the total P in
the lake was decreasing, probably because of the reduced external loading, the sum of the densities of
cyanobacteria and Prochlorothrix hollandica was increasing. The dominant species of the zooplankton
community (rotifers and crustaceans) did not show changes related to the restoration measures. It is
concluded that the available data are insufficient to predict future changes, but that the changes in the
past can be described well by a simple linear ordination model.

Introduction changes in such eutrophic systems. During this

Eutrophication of lakes is a worldwide problem,
resulting in systems dominated by a few species
of algae in which the diversity of the fish and
zooplankton communities is low and where mac-
rophytes are virtually absent. Nowadays, the res-
toration of such systems is a major issue in sci-
entific research, but only in a few occasions
detailed information of the communities seems to
be available. The Water Quality Loosdrecht
(WQL) project was one of the intensive research
projects ofthe last decade, planned to obtain more
insight into the mechanisms responsible for the
project detailed information was collected in the
Loosdrecht Lakes. Information on overall pro-
cess rates, e.g. primary production, grazing rates
and phosphorus release from the sediments, as
well as detailed information on the composition
of the planktonic communities and physical and
chemical environment were collected in a system-
atic way over a period of eight years. It was
planned to use the available data for the construc-
tion of models describing the major changes in
the system and predicting future changes.
The Loosdrecht lakes system consists of three
shallow, mainly hypertrophic lakes situated in the
106
central part ofThe Netherlands: Lake Loosdrecht
(surface area 979 ha, mean depth 1.85 m), Lake
Breukeleveen (surface area 179 ha, mean depth
1.45 m) and Lake Vuntus (surface area 88 ha,
mean depth 1.45 m. Chlorophyll concentrations
of more than 150 mg m - 3 are frequently reached
during the growing season. For a detailed de-
scription of the lake system see Loogman & Van
Liere (1986), Van Liere et al. (1990b), Gulati et al.
(1991) and Hofstra & van Liere (1992). The lakes
developed from a oligo/mesotrophic system in the
1930's to a eutrophic system in the 1950's (Hof-
stra & Van Liere, 1992). From 1976 onwards
several measures were taken in an effort to restore
the lakes, including the construction of a sewer-
age system, in the period 1976-1986, for the vil-
lages which earlier discharged their sewage into
the lake. Also, the inlet of highly eutrophic water
from the River Vecht was replaced by supply of
phosphorus depleted water from the Amsterdam
Rhine Canal in 1984 (Van Liere et al., 1984, 1986).
These measures were expected to lead to a re-
duction in the available phosphorus in the lakes,
thus leading to a decrease in the abundance of
cyanobacteria.
Because the lakes are shallow, wind-induced
resuspension of epipelon and peat fragments fre-
quently occurs (GOllS et al., 1986; Gons & Otten,
1991; Otten et al., 1992). The light climate is neg-
atively affected by the resuspended matter, which
favours the filamentous cyanobacteria (GOllS &
Otten, 1991). The phytoplankton community is
dominated by filamentous procaryotes (GOllS
et al., 1992) i.e. cyanobacteria and Prochiorothrix
hollandica, discovered in 1985 (Burger-Wiersma
et al., 1986, 1989). The zooplankton biomass is
dominated in spring by small cladocerans, mainly
Bosmina, the zooplankton numbers by rotifers,
mainly Keratella. Copepods dominate the biom-
ass and rotifers the numbers in summer (Gulati,
1990; Gulati et at., 1985, 1992). The fish commu-
nity is dominated by bream (Abramis brama), the
main planktivore, which is also supposed to con-
tribute to the resuspension of benthic sediments
by its foraging behaviour; pikeperch (Stizostedion
Iucioperca) is the main top predator (Lammens
et at., 1992). The structure of the fish community
resembles the structure in most eutrophic lakes in
The Netherlands.
The aim of this study is to detect trends in the
densities of phytoplankton and zooplankton and
in the chemical and physical variables frequently
monitored in the Loosdrecht lakes, and to exam-
ine the possible relationships between the com-
position of the plankton communities and the res-
toration measures as well as the natural variation
in the abiotic environment. For this purpose sev-
eral essentially different methods can be used,
including regression analysis, ordination tech-
niques and dynamic modelling. For dynamic
models a huge number of parameters have to be
estimated, even in the case that only a few state
variables are modelled. Most of the parameters
can not be reliably estimated from field data, and
therefore calibration techniques are often used to
obtain a better fit. Within the WQL project such
a dynamic model, the PC-LOOS model (lanse &
Aldenberg, 1989, 1990; lanse et al., 1992), was
developed for the same lake system, using data of
the same period of time. Canonical ordination
combines the simplicity of regression models with
the power of ordination models to detect the in-
ternal structure in a data set including a relatively
large number of interrelated variables. We there-
fore chose to use canonical ordination techniques
to study the trends and the relationships between
environmental factors and the planktonic com-
munities.
Material and methods
Three of the several sampling stations chosen in
the lake area at the start of the WQL project (Van
Liere et al., 1984) were sampled at least once in
four weeks throughout the study period (1983-
1990). These are P6 in Lake Breukeleveen, P7 in
Lake Vuntus and P9 in Lake Loosdrecht. The
other sampling stations are omitted from this
study. The lake water was sampled using a 5 I
Friedinger sampler at four-weekly intervals (in
some years at fortnightly intervals). A 30 I water
sample was collected at each of the three sam-
pling stations. Subsamples were used to measure

chemical characteristics (for methods see Van
Liere et al., 1989b) and to count phytoplankton
and zooplankton. Chemical variables included
several species of phosphorus (total P, soluble
total P and reactive phosphate), several species of
nitrogen (nitrate, ammonium, total N), dissolved
organic carbon (DOC) and total Si0 2 . The phy-
toplankton sample (0.5 I) was fixed using 5 ml
Lugol and counted using an inverted microscope.
Zooplankton was separated from the other
sestonic particles by sieving over alSO /lm filter
and preserved in 4 % formaldehyde. Prior to
counting zooplankton was subsampled using the
whirling vessel of Kott (1953). These subsamples
were transferred into counting chambers and
counted using an inverted microscope.
Separate data files (Lotus spreadsheets) on
zooplankton, phytoplankton and abiotic factors
of each sampling station for each year were com-
bined into one large file using the program CED IT
(Van Tongeren, 1991). This program was also
used to replace missing values by estimates using
interpolation techniques and to select the samples
for the analysis. Interpolation was not applied if
- more than 50 % of the chemical and physical
data were missing for one sampling date,
- the same variable was missing in two samples
following each other directly and
- a chemical or physical variable highly fluctu-
ated.
Samples were selected applying the following
three criteria to the data:
- all chemical and physical variables had been
determined or reliably estimated by interpola-
tion,
- both phytoplankton and zooplankton counts
were available, and
- samples for chemical and physical data and for
phytoplankton and zooplankton counts had
been collected within 30 hours. If this criterion
was not met, the phytoplankton and zooplank-
ton counts were included in the analyses as
'passive' samples.
Inspection of the data and a first analysis revealed
several taxonomic problems in the phytoplankton
counts. It appeared to be necessary to combine
several species into groups in order to eliminate
107
personal errors in the identification, to reduce the
effects of different taxonomic levels (species or
genus) used in different years and to reduce the
effects of species with very low frequencies in the
combined data set. The most important assem-
blage, dominating the phytoplankton for most of
the year, was that of the so-called Oscillatoria
limnetica group, including O. limnetica and other
filamentous cyanobacteria closely resembling
O. limnetica, as well as Prochlorothrix hollandica.
The combination of chemical and physical data
with zooplankton and phytoplankton counts re-
sulted in a total of 430, partly incomplete, sam-
pling units. After subsequent combination of data
obtained from samples, collected separately but
within a period of 30 h, and deletion of the sam-
pling units without zooplankton counts or phy-
toplankton counts, 295 sampling units remained.
Environmental variables being measured in less
than 150 of the remaining sampling units or not
being measured over longer periods of time were
deleted. From the 295 sampling units 173 con-
tained sufficient information on the remaining en-
vironmental variables. About 2 % of the values
were missing and replaced by interpolation esti-
mates.
Some more variables were added to the set of
environmental variables. The sampling stations
can be represented by a nominal variable with
values P6, P7 and P9. However the program does
not recognize nominal variables. Therefore, the
sampling stations were coded as three dummy
(value 0 or 1) variables (P6, P7 and P9 respec-
tively), as a replacement for all the differences
between lakes not covered by the chemical and
physical variables. Since exact data on the exter-
nal phosphorus load are missing, we assumed a
linear relationship with time, which is mimicked
by the variable year, having the discrete values
from 83 to 90.
Data were analyzed using Principal Compo-
nents Analysis (PCA, see e.g. Pielou, 1977, 1984;
Gauch, 1982; Ter Braak, 1987a) for the indepen-
dent (i.e. environmental) variables: the chemical
and physical characteristics measured and the
extra variables described in the previous para-
graph. Redundancy Analysis (RDA: Rao, 1964,
108
1973; Ter Braak, 1987a) and a hybrid of RDA
and PCA were applied to the dependent variables
(phytoplankton and zooplankton counts) and the
independent variables. For all analyses program
CANOCO (Ter Braak, 1987b, 1989, 1991) was
used.
Environmental variables were analyzed using
PCA after centering and standardization of the
variables, i.e. PCA on the correlation matrix of
the environmental variables. The correlation
structure within the set of environmental varia-
bles was visualized by plotting the correlations of
the variables with the extracted PCA axes. Be-
cause year and sampling station are not chemical
or physical characteristics, these variables were
analyzed passively: i.e. their correlations with the
ordination axes were computed after the correla-
tion matrix between the other variables was used
to obtain the first four principal components
(axes).
Plankton counts were log transformed
(log(y + 1)) and centered prior to the analysis. The
results of the thus obtained RDA and PCA on the
respective covariance matrices are presented as
so-called correlation biplots (e.g. Ter Braak,
1987a) in which the correlations with the ordina-
tion axes for both the dependent and the inde-
pendent variables are plotted. Besides, centroids
(means) of the sample scores were computed for
classes of samples and also plotted.
Forward selection of environmental variables
was used in the analysis of plankton counts to
obtain a small set of environmental variables
maximally related to the species counts. An ar-
bitrary threshold value of 3% of the variance in
the species data 'explained' by the addition of the
next environmental variable was used to decide
whether an environmental variable should be in-
cluded in the model as independent variable. This
arbitrary percentage approximates the mean pro-
portion of the total variance for each species, i.e.
100% divided by 38 for phytoplankton taxa or by
32 in case of zooplankton taxa. Variables like
total phosphorus, chlorophyll and transparency
(Secchi-depth) were considered not to be real in-
dependent variables, since their values and the
densities and composition of zooplankton as well
as phytoplankton are mutually dependent. There-
fore, such variables were not entered into the
model. The environmental variables not included
in the model were analyzed passively, i.e. their
correlations with the ordination axes were com-
puted after the computations for the ordination.
Phytoplankton and zooplankton data were ana-
lyzed separately. Additionally, the zooplankton
counts were analyzed passively together with the
phytoplankton counts, i.e. the correlations of
zooplankton counts with the four ordination axes
were computed after extraction of the axes. The
CANOCO results were post-processed, mainly to
compute centroids, using program CEDIT (Van
Tongeren, 1991).
Results
The PCA analysis of the environmental variables
describes 71 % of the variance in the chemical
and physical data in 4 axes. Passive analysis of
the relationship with sampling station and year
indicates that 54 % of the relationship between
chemical and physical data and sampling station
and year is represented by this PCA solution.
Temperature and year, which accounted for more
than the arbitrary 3% of the variance in the trans-
formed species counts, were selected to enter into
the RDA for the phytoplankton. However, for
zooplankton only temperature accounted for
more than 3% of the total variance in the trans-
formed counts. The hybrid (RDA and PCA)
analyses ofthe dependent variables accounted for
34% and 64% of the variance in the phytoplank-
ton and zooplankton counts, respectively, in 4
axes. The temperature constrained (RDA with
temperature) axis of the zooplankton counts ac-
counted for 31 % of the variance. The tempera-
ture-and-year constrained axes accounted for
only 11 % of the variance in the phytoplankton
counts. The unconstrained (PCA) axes describe
33% (64% minus 31%) and 23% (34% minus
11 %) of the variance in the zooplankton (3 axes)
and phytoplankton data (2 axes), respectively.
 109

Environmental variables Table l. Environmental variables with their abbreviations,

units, and means with standard deviations.

The four peA axes of the environmental varia- Variable Variable Unit Mean s.d.
bles account for 28.3%, 17.5%, 14.7% and name
10.3% of the variance, respectively. The yearly
Year minus 1900 Year Year 87.9 1.6
temperature cycle was clearly related to the cycles Temperature Temp. °C 11.3 6.1
of some of the environmental variables. Fig. 1 Nitrate NO) mg N I-I 0.06 0.13
Free ammonium NH 4 -free mg N I-I 0.08 0.12
summarizes the results for the environmental var- mg N I-I
Total ammonium NH 4 ·tot. 0.36 0.18
iables as a plot of PCA axis 1 (abscissa) against Secchi depth Secchi m 0.25 0.08
axis 3 (ordinate) displaying most of the informa- Dissolved organic carbon DOC mgC I-I 12.4 3.0
Si0 2 Total solicate mg Si0 2 l 2.0 1.5
tion on temperature. Temperature and concen-
Soluble reactive phospate SRP Ilg P [' 3.2 3.0
trations of NH 4+ , N0 3- and SiO z were negatively Soluble total phosphate STP Jlg P 1- I 9.5 4.2
correlated (arrows pointing more or less into the Total phospate TP Jlg P 1- I 89.0 22.0
opposite direction of the temperature arrow). The
negative correlation of soluble total phosphate
(STP) with temperature was weak, which is in-
dicated by the short projection of the arrow on TP
the temperature arrow. Rough estimates of the 0.5

concentrations at different temperatures can be

S102
made from the means and standard deviations of
the environmental variables (Table 1). In Fig. 2 -t
the year arrow roughly points into the direction of STP
axis 2 (ordinate), including that trends, displayed SRP
approximately along axis 2 (Fig. 2, ordinate), are -0.5

very weak. Differences between stations are best

displayed by axis 3 (contrast between P9 and P7:
Fig. 1) combined with axis 4 (contrast between -t

P6 and the other stations: Fig. 2). The arrow for Fig. 2. PCA diagram of environmental variables, axis 2 (ab-
Lake Breukeleveen (P6) roughly points into the scissa, 17.5% of the variance) against axis 4 (ordinate, 10.3 %
ofthe variance. Year, P6, P7 and P9 are passive variables. For
abbreviations see Table I.

0.5
same direction as that for dissolved organic car-
N03 bon (DOC), whereas the arrow for Secchi depth
S102 points roughly into the opposite direction. There-
rr.-.,.--,--,-.--r--;;r;;~~~IsTP:;=:;::;::;..:N~H~4 - f r ee fore, from Fig. 2 we derive that DOC was rela-
-t tively highest and Secchi depth relatively lowest
NH4-tot. in Lake Breukeleveen (P6), and that STP and
Temp. total phosphorus (TP) were decreasing in the
Secchi
course of years (arrows pointing into the opposite
DOC direction of the Year arrow). These observations
-t
are confirmed by inspection of the full correlation
matrix among environmental variables. The cor-
Fig. l. PCA diagram of environmental variables, axis I (ab-
scissa, 28.3% of the variance) against axis 3 (ordinate, 14.7% relations for soluble reactive phosphate (SRP)
of the variance. Year, P6 (not visible), P7 and P9 are passive with the other environmental variables were low,
variables. For abbreviations see Table I. probably because of the low accuracy of the SRP
110

measurements, the concentrations being near the
detection limit.
Phytoplankton
Temperature (Temp.) and the course of years
(Year), accounted for 6% and 5% of the variance
in the phytoplankton counts, respectively. This
seemingly poor fit is caused mainly by the chosen
transformation (log(y + 1)), which leads to a large
weight (loading) for species with a high ratio be-
tween the highest and lowest numbers counted.
These species are not necessarily the most abun-
dant ones. However, the cumulative percentages
fit for some important species indicate a fairly
good description of these species by the model
consisting of Temperature and Year (Table 2);
however, only for the Oscillatoria limnetica group
more than 50 % of the variance was related to the
independent variables. The two constrained axes
(axes in which the relationship with the indepen-
dent variables is expressed) account for 59.3 and
40.7% of the species environment relationship.
The biplot of species and environmental variables
(Fig. 3a) reveals several, fairly strong relation-
ships. The length of the species vectors (not
drawn, but being the arrows connecting the ori-
gin with the species points) reflects the proportion
of the variance in the log-transformed counts re-
lated to the independent variables (Temp. and
Year).
Some phytoplankton species were correlated
negatively with temperature, indicating their
peaks at lower temperatures and their relatively
low numbers at high temperatures. Important
among these are Oscillatoria redekei (max. 94000
cells ml- 1 in June 1983), Asterionella formosa
(max. 1040 cells ml- 1 in April 1988, max. 160
cells ml - I in August and September over the
whole period) and Dinobryon (max. 1700
cells ml- 1 in April 1984, max. in August and Sep-
tember over all years 203 cells ml- 1). Species
correlated positively with temperature are:
Oscillatoria limnetica groups (max. 388000
cells ml- 1 in August 1983 and 350000 ml- 1 in
September 1990, max. spring value ever reached
203000 cells ml - 1), Aphanizomenon jim-aquae
(16000 cellsml- 1 in August 1983, 13600
cells ml- 1 in the extremely warm month of June
1986, max. spring value over all other years 9500
cells ml - 1) and Tetraedron (200-400 cells ml- 1

Table 2. Cumulative fit per species for the first four ordination axes as fraction of the total variance and total variance per spe-
cies (after log-transformation) for some important phytoplankton species. Fit is related to environmental data (temperature and
year, axes I and 2) and latent variables (axes 3 and 4). Species (groups) not mentioned, but apearing in the figure 3: Actinast -
Actinastrum; AnkifMon - AnkistrodesmusfMonoraphidium; Chodatel - Chodatella; Cosmariu - Cosmarium; Cryptomo - Crypto-
monas; Dinobryo - Dinobryon; Gymnodin - Gymnodinium; Lynglimn - Lyngbya limnetica; Oocystis; Peridini - Peridinium; Phacus;
Restbitr - unidentified big trichomes; RestCent - unidentified centric diatoms; RestCyan - unidentified Cyanophyceae; Rhizosol-
Rhizosolenia; and Tetraedr - Tetraedron. Note that Prochlorothrix hollandica is included in the Oscil/atoria limnetica group.

Name Abbreviation Fraction of the variance 'explained'

Constrained axes Unconstrained axes Variance

VAR (y)
Axis I Axis 2 Axis 3 Axis 4

Oscillatoria limnetica group Osciligr 0.516 0.547 0.610 0.663 0.22

Aphanizomenon tlos-aquae Aphatl-aq 0.274 0.28 0.353 0.354 0.49
Oscillatoria redekei Oscirede 0.242 0.242 0.356 0.370 1.36
Remaining Chlorophyceae RestChlo 0.171 0.257 0.258 0.521 l.l5
Pediastrum Pediastr 0.040 0.139 0.148 0.396 1.36
Scenedesmus Scenedes 0.019 0.059 0.067 0.428 0.72
Asterionella formose Asteform 0.087 0.127 0.334 0.336 1.86
Melosira Melosira Om5 0.036 0.594 0.764 3.08
Diatoma elongatum Diatelon 0.004 0.026 0.650 0.776 3.53

O.S
6. StCY6
-Q.S
1I11i DSDJ
Year -1
84Jn
Fig. 3a. RDA correlation biplot of phytoplankton species and
environmental variables, constrained by temperature and year:
axes I (abscissa, 6.6~o of the variance in the species counts)
and 2 -ordinate, 4.5% of the variance in the species counts).
Species vectors are represented as symbols at the position of
the arrow-heads, vectors of environmental variables are
drawn. For abbreviations see Tables I and 2.
b. Succession in the phytoplankton community represented
by lines connecting the centroids (average scores) for all sam-
ples taken in each month. The beginning of each year is in-
dicated by an arrow-head, for clarity some other points are
labelled with year and month.
in summer, max. 150 cells ml- 1 in spring) indi-
cating summer peaks for these species.
An increase of numbers with time (Year) is
inferred for the Oscillatoria limnetica group (min-
imum 770 cells ml- 1 in 1983, gradually increas-
ing to 51 500 cells ml - 1 in 1990, maxima of the
111
same order of magnitude in 1983 and 1990, being
ca 400000 cells ml- 1, gradually increasing from
1985, ca 200000 cells ml- 1 towards 1990), Rhizo-
solenia (reaching peaks of 1000-4000 cells ml- 1
in 1987, 1989 and 1990, max. in the period 1983-
1986 270 cells ml- 1) and Ankistrodesmus/
Monoraphidium (max. 400 cells ml- I), but Pedi-
astrum, Actinastrum, Phacus and Scenedesmus
(species of minor importance, which never
reached counts of more than 400 cells ml- 1) de-
crease. For Melosira the decrease observed from
the diagram is due to one extreme peak of 19700
cells ml- 1 in april 1984. Omitting 1984, no trend
can be detected, maxima being invariably around
4000 cells ml- I.
The plot of centroids (mean scores) of the sam-
pie scores for the months within years (Fig. 3b)
also shows the seasonal variation and the trend
over years. The phytoplankton exhibited excep-
tional behaviour in 1985: the community in spring
of 1985 was like that in the springs of the pre-
ceding years 1983, 1984 and 1986, but the com-
munity composition in summer 1985 closely re-
sembled that during the summers of the following
years 1987, 1988 and 1989. The extremely mild
winter of 1989-1990 is noted from the triplot ob-
tained by overlaying Fig. 3a on Fig. 3b. The
abundance of species correlated positively with
temperature remained high. None of the stations
seems to have differed markedly from the other
stations (first four axes).
Zooplankton
Only the water temperature was well related to
the zooplankton numbers, accounting for 31 % of
the variance in the species counts. The second,
unconstrained, axis accounts for another 18 % of
the variance, and it mainly describes the differ-
ences between spring and autumn. The biplot of
species with environmental variables in Fig. 4a
clearly shows the recurrent seasonal succession
in the zooplankton community. The length of a
species vector reflects the explained variance of
the species (after log-transformation and center-
ing), i.e. the square root of the proportion of the
 112

variance in the zooplankton counts related to sea-

sonal variation (see Table 3 for the most impor-
ACflyl1spns
tant species). The general pattern in the cyclic
TP succession thus derived is: in winter almost all
0.5
~8Ds.pp ABr.cspp
species were nearly absent; in early spring rela-
APrJ_sulc
ASyncopp
tively low peaks of Asplanchna (10-16 ind I- 1 )
AsplS~A
STPA and Synchaeta (500-600 ind 1- 1) occurred, fol-
N03 ::>ec
lowed by higher peaks of Keratella quadrata
J1~JsJ em p.
_rl ....,NrH'4-_Tf-rTe-e'jil-r.:IIiI~-.It:::Oilll:r£AP'-r--r--r-r-,.......,Ar-F...
(200-300 ind 1- 1) and K. cochlearis (2000-3000
ind 1- I); late spring and early summer Bosmina
species (B. coregoni and B. longirostris (ca.
-0.5 600 ind 1- I) were followed by Brachionus (ca.
ACDl1spp 100 ind 1- 1) and Pompholyx sulcata (ca.
250 ind 1- 1) and later nauplii (ca. 500 ind 1- I),
cyclopoid copepods (ca. 200 ind 1- 1) and Daph-
-I
nia species (ca. 40 ind 1- 1); the carnivorous cla-
doceran Leptodora kindtii was most abundant (ca.
86May 3 ind 1- 1) at the same time, and later in summer
r"" the zooplankton community was dominated by
/ '"" rotifers, mainly Filinia spp. (ca. 500-800 ind 1- 1),
/ Anuraeopsis ftssa (ca. 5000 ind 1- 1), Trichocerca
:'.,l~
I ........
(ca. 300-600 ind 1- 1), and Polyarthra (ca.
700 ind 1- 1), the arrow of this latter species being
shorter, because it was present during the whole
year and only varied ca. tenfold in numbers). Col-
lotheca spp. (ca. 500 ind 1- I) and Euchlanis
90Jn
dilatata (ca. 20 ind 1- 1) were most abundant at
the end of the season. Most species are well rep-
resented by the model: of the 32 species there are
10 for which more than 30% of the variance is
described by the first two axes (Table 3). Species
not recorded in Table 3 do not fit well in the four
axes of the ordination, but are also of minor im-

Fig. 40. RDA correlation biplot of zooplankton species and

--P6
environmental variables axes 1 (abscissa, constrained by tem-

,,,<,,;f<·'·~~ .:::>>.".,.,
----- P7
.......... pg perature, 30.9% of the variance in the species counts) and 2
H"" (ordinate, unconstrained, 17.5% of the variance in the species
counts), samples are not shown. Vectors are indicated as in
I .~
Fig. 3a. The direction of the species vectors reflects the tim-
:: '''l, June ing of the maxima: turning clockwise around the origin, the
January/ I successive peaks in the species densities can be read. For
Feb r uar v-.l!::-r'''::''~.-.-.--.--+---r--r--r--r-~i--.+\-':'r--'--'
,--.. abbreviations see Tables 1 and 3.
b. Succession in the zooplankton community. For details see
Fig.3b.
c. Recurrent seasonal succession for the three stations rep·
resented by lines connecting the centroids (average scores)
over 8 years for all zooplankton.
 113

Table 3. Cumulative fit per species as in Table 2, but for important zooplankton species. Fit is related to temperature (axis 1) and
latent variables (axes 2, 3 and 4). Species not mentioned, but appearing with label in figures 4a and 5; Alonspp - Alona; Asplspp
- Asplanchna; Cephspp - Cephalodella; Ceripulc - Ceriodaphnia pulchella; Conounic - Conochilus unicomis; Kellong - Kellikottia
longispina; Lecaspp - Lecane; Leptkind - Leptodora kindtii; and Pompsulc - Pompholyx sulcata. The remaining species are of minor
importance and indicated by symbols only.

Name Abbreviation fraction of the variance 'explained'

Constrained Unconstrained axes VAR (y)

Axis 1 Axis 2 Axis 3 Axis 4

Daphnia Daphspp 0.238 0.247 0.395 0.517 0.89

Bosmina Bosmspp 0.009 0.260 0.262 0.482 l.l
Chydorus sphaericus Chydspha 0.030 0.474 0.479 0.483 1.65
Cyclopoid copepods CyclCope 0.402 0.439 0.42 0.486 0.50
Nauplii Nauplii 0.386 0.474 0.476 0.556 0.38
Synchaeta Syncspp 0.157 0.270 0.323 0.664 1.87
Polyarthra Polyspp 0.240 0.288 0.291 0.338 0.59
Keratella quadrata Keraquad 0.198 0.778 0.795 0.833 1.83
Keratella cochlearis Keracoch 0.019 0.688 0.747 0.748 1.23
Filinia Filispp 0.661 0.665 0.850 0.850 3.03
Brachionus Bracspp 0.156 0.374 0.416 0.535 1.45
Anuraeopsis fissA Anurfiss 0.654 0.714 0.896 0.896 4.49
Trichocerca Tricspp 0.543 0.618 0.890 0.891 2.80
Collotheca Colspp 0.344 0.725 0.726 0.801 2.82
Ascomorpha Ascospp 0.196 0.222 0.239 0.302 1.38
Euchlanis diJatata Euchdila 0.082 0.232 0.277 0.278 0.82

portance considering their lower densities or low

variances after log-transformation.
The plot of monthly centroids of the zooplank-
ton counts (Fig. 4b) and the correlations of the
zooplankton counts with axes in the ordination
space of the phytoplankton species (Fig. 5) con- 0.5
Temp .. Tricspp
firm the absence of a clear trend over years in the • FiJ IjPp N8upJi i
tl>CYCJCOP8PO."SUJC CeripuJc
zooplankton community. However, from the third Anurfiss .. 8r8Ctpp Le 8S'16 /f.er8cocfl "Syncspp
axis of the ordination (not shown) and Fig. 5 we "EUCfldl,,1 KeJJJong

derive that Daphnia sp. and Keratella cochlearis -1 .. ..CoJuspp .. Ker8qu8d

ASCOSpp6
were slightly decreasing. From the third zoop- sl!studin

lankton ordination axis alone we infer that also .. AspJspp

Filinia spp. Trichocerca spp. and Anuraeopsisfissa

were decreasing over years. Some other species
that seem to have decreased or increased were
too low in numbers to infer a trend from this
analysis. The year 1985 was conspicuous by hav-
-1
ing a secondary peak in Daphnia sp. in autumn.
In contrast to the phytoplankton densities Fig. 5. Passive ordination (axis 1, abscissa: and axis 2, ordi-
nate) of the zooplankton species in the same ordination space
which did not differ among the lakes (not pre-
of temperature and year as in Fig. 3. Points are correlations
sented in a figure) the zooplankton densities did of species with axes. The correlations of temperature and year
slightly differ. The Keratella peaks were higher in are plotted as arrows.
114
Breukeleveen (P6) than at the other two stations
during April and May and those of the rotifers
were higher in lake Vuntus (P7) and the central
lake Loosdrecht (P9) than in Lake Breukeleveen
(P6) during July and August (Fig. 4c).
Discussion
The proportion of the total variance described by
the hybrid analysis (2 constrained, 2 uncon-
strained axes) of the phytoplankton is low
(33.7%) compared with the proportion of vari-
ance explained in other analyses either with the
same material (ca. 50%), with data from other
lakes (not published, max. ca. 60%) or with the
zooplankton data analyzed in this paper (64 %).
This is due to several reasons. First, we chose to
use only a few independent variables, thereby re-
ducing the proportion of variance in the species
data that could maximally be explained by the
independent variables. Secondly, at the start of
the WQL project less abundant species were con-
sidered to be of minor importance and, therefore,
were lumped. Thus, adequate numbers have been
counted only for the dominant species, which are
a few in number. Densities for the other species
have been derived from very low numbers of cells
actually counted, their errors, therefore, are large.
We, nevertheless, chose this analysis (using log-
transformed data and the covariance matrix) be-
cause relatively more weight is given to the less
abundant species, and less weight to species with
less variation in abundance. Using untransformed
counts generally results in information on a few
important species only and in a high proportion
of explained variance, because the variance in
untransformed data of the other species is very
low compared with the variance in the dominant
species. However, in the present analysis this is
less important, because few independent varia-
bles have been used and, therefore, the full cor-
relation structure with the independent variables
is displayed on the first axis for the zooplankton
analysis and on the first two axes for the phyto-
plankton analysis. Retaining as much informa-
tion as possible on the relative changes in abun-
dance of all species in a low-dimensional
ordination space, thus implicitly giving the abun-
dant species a relatively lower weight, was con-
sidered to be more important than the statistical
reasons for transformation, which would have led
to the same transformation anyhow. If more in-
dependent variables were included in the model,
the fit for species now of minor importance would
have been better, although not significantly.
In the zooplankton community the abundances
of species differ less than in the phytoplankton
community. Therefore, more species have been
counted with sufficient precision. Thus, the over-
all accuracy is much higher than that for most
phytoplankton species. The proportion of the
total variance that can be ascribed to temperature
alone (the constrained first axis) is 31 %; the sec-
ond, unconstrained, axis (mainly describing the
difference between spring and autumn) accounts
for an additional 18 % of the total variance in the
species data. The next two, unconstrained, axes
account for only 15 % of the variance. The pos-
itive correlation (not shown) of the third axis with
time (Year, r = 0.249) is confirmed by the results
of the phytoplankton analysis (Fig. 5), but mainly
concerns species having low numbers, which have
been counted less accurately. Nevertheless, using
supplementary information on the species that
seem to have increased or decreased could pro-
vide more circumstantial evidence for changes in
abundance.
Because, as expected, we observed consider-
able autocorrelation in space and time, a sound
statistical test of the inferred relationships is
hardly possible. Since the sampling units are not
equally spaced in time, there is no simple permu-
tation scheme for a Monte Carlo permutation test
as provided by CANOCO. However, although it
is possible to feed the program with a permuta-
tion scheme, we did not consider this necessary.
Since most of the chemical and physical charac-
teristics are related to each other in some way,
deducing causal relationships would be incorrect.
The reduction in the external phosphorus load
since 1984 most probably resulted in the decrease
in total phosphorus observed in this study and by
Van Liere et al. (I 990b). However, the expected

decrease in the abundance of filamentous cyano-
bacteria (Van Liere, 1986) did not occur. Several
explanations can be given for this. Gons et al.
(1986, 1992) report the effect of wind on this shal-
low lake system as an important cause for the
sub-optimal light conditions for both green algae
and macrophytes. However, this does not explain
the increase in the filamentous cyanobacteria
since Rijkeboer et al. (1990) indicate that from
1987 onwards the dominant algae in the Loos-
drecht Lakes were phosphorus-limited rather
than light-limited. The CjP ratios in the system
have tended to increase since 1987 (Gulati et al.,
1992), also indicating that, with almost stable
seston mass, phosphorus limitation is likely at
this moment. Neither the resilience of the system
(Van Liere et al., 1990a), nor the existence ofmul-
tiple stable states (Scheffer, 1990) are good ex-
planations for the increase in filamentous cyano-
bacteria, since one can hardly expect an increase
in numbers and biomass of phosphorus-limited
cyanobacteria when total phosphorus is decreas-
ing.
Although the reduction in external phosphorus
load does appear to have resulted in a decrease
in total P concentration of the lakes, the latter
seem to have developed into an even more unde-
sirable direction. Partly, however, variation in the
climate may have caused the increase in filamen-
tous cyanobacteria. All winters between 1987 and
1990 were mild and this may reduce the algal
mortality rate, and, consequently, a high starting
standing crop of algae in spring may result in
extreme dominance of the filaments in summer.
The raw data, however, do not clearly corrobo-
rate effects of climate on the observed trend. The
observations of the more severe winter of 1990-
1991 and thereafter may confirm the extent to
which the climate was responsible for the ob-
served trends in the Loosdrecht lakes. An alter-
native, even more speculative, explanation that
probably cannot be checked is that during the
study years Prochlorothrix hollandica, discovered
in 1985 (Burger-Wiersmaetal., 1986, 1989) grad-
ually superseded Oscillatoria limnetica or even
some other Oscillatoria species. The two genera
are difficult to distinguish in preserved samples;
115
in fresh material, however, they can be distin-
guished using epi-fluorescence (Burger-Wiersma
et al., 1989; Van Liere et al., 1989a), but this has
been done only since summer 1988.
Strikingly, even though the phytoplankton
community changed markedly, due mainly to a
two-fold increase of filamentous cyanobacteria
(including Prochlorothrix hollandica), the zoop-
lankton community did not change much. Most
species slightly decreased with time; only some
species, e.g. Asplanchna, occurring in low num-
bers, seem to have increased. But, because their
densities were low and, consequently, the error
variances in the numbers high, the variance ex-
plained is low and their increase is, therefore, not
significant. Apparently, most zooplankton spe-
cies were influenced either by the stochastic
changes in the climate, which is, however, not
expected to result in a decrease, or by changes in
the quantity and quality of their food. For exam-
ple, green algae (e.g. Scenedesmus and Ankis-
trodesmus), considered to be good food for most
zooplankters, decreased and filamentous cyano-
bacteria, which possibly hinder the uptake offood,
especially at higher concentrations, as demon-
strated by Dawidowicz (1990; Dawidowicz et al.,
1988), increased.
The ordination models presented here describe
the main changes in the planktonic communities
in Loosdrecht lakes using only a few independent
variables. The absence of clear relationships with
changes in nutrient concentrations of the system
indicates that the range of nutrient concentrations
encountered during the study years is too narrow,
compared with the seasonal fluctuations, to draw
conclusions. This may also be due to the consid-
erable error variance in the observations due to
sampling and counting methods, but, since the
PCA of environmental data alone did not reveal
strong trends in time, the former reason, a narrow
range in nutrient concentrations, seems to be more
important. Nevertheless, both the phytoplankton
and the zooplankton communities indicate that
the Loosdrecht lakes are hypertrophic, and we
may expect that further reductions in phosphorus
load finally will lead to restoration of the lakes.
Without data from other comparable systems or,
116
preferrably, experimental studies on the responses
to nutrients of the now important species and the
species expected to take over if the system be-
comes less eutrophic, we will not be able to pre-
dict the course of this restoration.
These ordination models describe the behav-
iour of small species groups or separate species in
relation to temperature and in the course of time.
The relationship with time had to be introduced
because causal relationships with nutrients could
not be deduced. Moreover, further lumping of the
species can lead to a reduction of the error var-
iance, but also to undesired central tendency ef-
fects: the decrease of one species and concurrent
increase of an other within the same group will
offset each other. Despite the apparent constancy
in time or apparent lack of response to changes
in a nutrient, the community may change consid-
erably.
Besides the drawbacks of lumping based on
rough taxonomic criteria, both descriptive models
like the one in this study, as well as dynamic
ecosystem models like PC-LaOS (Janse & AI-
denberg, 1989, 1990; Janse et al., 1992), should
be based on a few, relatively homogeneous groups
with respect to response of the species to relevant
variables. If not, any model is likely to fail to
describe and predict the changes in the ecosys-
tem. Thus, it is unlikely that a dynamic model,
calibrated on lumped data from this same system
in the same period of time, will describe the
changes that took place in the Loosdrecht lakes
better than a descriptive model fitted on as many
separate species as possible. However, the value
of dynamic ecosystem models as a tool for un-
derstanding the mechanisms leading to changes
cannot be denied.
Summarizing, no model based on the available
data for the Loosdrecht Lakes only can be used
for prediction purposes: collinearity and non-
linear relationships between temperature and nu-
trients do not allow reliable estimates of param-
eters for the species now present. It is in any case
impossible to estimate parameters for species,
now absent, that are likely to take over if the
restoration measures will be successful. The
present study indicates that the Loosdrecht lakes
system is changing, but the relation between this
change and the restoration measures is not yet
clear.
Acknowledgements
The first author is especially thankful to Cajo ter
Braak for stimulating discussions and Koos Vij-
verberg for his useful remarks on an earlier ver-
sion of this typescript.
References
Burger-Wiersma, T., L. J. Stal & L. R. Mur, 1989.
Prochlorothrix hollandica gen. nov., sp. nov.: a filamentous
oxygenic photoautotrophic procaryote containing chIoro-
phylls a and b: assignment to Prochlorothrichaceae fam. nov.
and order Prochlorales, Florenzano, Balloni & Materassi
1986, with emendation of the original description. Int. J.
Syst. Bact. 39: 250-257.
Burger-Wiersma, T., M. Veenhuis, H. J. Korthals, C. C. M.
Van de Wiel & L. R. Mur, 1986. A new prokaryote con-
taining chlorophylls a and b. Nature 320: 262-264.
Dawidowicz, P., 1990. The effect of Daphnia on filament length
of blue-green algae. Hydrobiologia 191: 265-269.
Dawidowicz, P., Z. M. Gliwicz & R. D. Gulati, 1988. Can
Daphnia prevent a blue-green algal bloom? A laboratory
test. Limnologica 19: 21-26.
Gauch, H. G., 1982. Multivariate Analysis in Community
Ecology. Cambridge University Press, Cambridge, 298 pp.
Gons, H. J., J. Otten & M. Rijkeboer, 1991. The significance
of wind resuspension for the predominance of filamentous
cyanobacteria in a shallow, eutrophic lake. Mem. 1st. ital.
Idrobiol. 48: 233-249.
Gons, H. J., R. Van Keulen & R. Veeningen, 1986. Effects of
wind on a shallow lake ecosystem: redistribution of partic-
ulates in the Loosdrecht Lakes. Hydrobiological Bulletin
20: 109-120.
Gons, H. J., T. Burger-Wiersma, J. H. Otten & M. Rijkeboer,
1992. Coupling of phytoplankton and detritus in a shallow,
eutrophic lake (Lake Loosdrecht, The Netherlands). Hy-
drobiologia 233: 51-59.
Gulati, R. D., 1990. Zooplankton structure in the Loosdrecht
lakes in relation to trophic status and recent restoration
measures. Hydrobiologia 191: 173-188.
Gulati, R. D., K. Siewertsen & G. Postema, 1985. Zooplank-
ton structure and grazing activities in relation to food qual-
ity and concentrations in Dutch lakes. Arch. Hydrobiol.
Beih. Ergebn. Limnol. 21: 91-102.
Gulati, R. D., L. Van Liere & K. Siewertsen, 1991. The Loos-
drecht lake system: Man's role in its creation, perturbation

and rehabilitation. In: Terrestrial and Aquatic Ecosystems:
Perturbation and Recovered. O. Ravera, Chichester Ellis
Horwood. pp 593-596.
Gulati, R. D., A. L. Ooms-Wilms, O. F. R. Van Tongeren, G.
Postema & K. Siewertsen, 1992. The dynamics and role of
limnetic zooplankton in Loosdrecht lakes (The Nether-
lands) Hydrobiologia 233: 69-86.
Hofstra, J. J. & L. Van Liere, 1992. The state of the environ-
ment of the Loosdrecht Lakes. Hydrobiologia 233: 11-20.
Janse, J. H. & T. Aldenberg, 1989. Modelling the eutrophi-
cation of the shallow Loosdrecht Lakes. Verh. int. Ver.
Limnol. 24: 751-757.
Janse, J. H. & T. Aldenberg, 1990. PCLOOS: eutrophication
model of the Loosdrecht Lakes. WQL report 1990-1,
RIVM, Bilthoven.
Janse,J. H., T. Aldenberg& P. R. G. Kramer, 1992. A math-
ematical model of the phosphorus cycle in the Loosdrecht
Lakes and simulation of additional measures. Hydrobiolo-
gia 233: 119-136.
Kott, P., 1953. Modified whirling apparatus for the subsarn-
pIing of zooplankton. Aust. J. mar. Freshwat. Res. 4: 387-
393.
Lammens, E. H. R. R., P. J. Boesewinkel-de Bruyn, H.
Hoogveld & E. Van Donk, 1992. Phosphorus-load, phyto-
plankton, zooplankton and fish stock in Loosdrecht Lake
and Tjeukemeer: confounding effects of predation and food
availability. Hydrobiologia 233: 87-94.
Otten, J. H., H. J. Gons & M. Rijkeboer, 1992. Dynamics of
phytoplankton detritus in a shallow, eutrophic lake (Lake
Loosdrecht, The Netherlands). Hydrobiologia 233: 61-68.
Pielou, E. c., 1977. Mathematical Ecology. 2nd edn. Wiley,
New York, 385 pp.
Pielou, E. c., 1984. The interpretation of ecological data. A
primer on classification and ordination. Wiley, New York,
263 pp.
Rao, C. R., 1964. The use and interpretation of principal com-
ponent analysis in applied research. Sankhya A 26: 329-
358.
Rao, C. R., 1973. Linear statistical inference and its applica-
tions. 2nd edn. Wiley, New York, 625 pp.
Rijkeboer, M., F. de Bles & H. J. Gons, 1990. Phosphorus,
light and seston dynamics (in Dutch). Internal Report
1990-11, Limnological Institute, 85 pp.
117
Scheffer, M., 1990. Multiplicity of stable states in freshwater
systems. Hydrobiologia 200/201: 475-486.
Ter Braak, C. J. F., 1987a. Ordination. In: Jongman,
R. H. G., C. J. F. ter Braak & O. F. R. van Tongeren (eds),
1987. Data Analysis in Community and Landscape Ecol-
ogy. PUDOC, Wageningen, 299 pp.
Ter Braak, C. J. F., 1987b. CANOCO-A FORTRAN pro-
gram for canonical community ordination by [partial][de-
trended][ canonical] correspondence analysis, principal
component analysis and redundancy analysis (version 2.1).
TNO Institute of Applied Computer Science, Wageningen,
the Netherlands.
Ter Braak, C. J. F., 1989. CANOCO - an extension of DEC-
ORANA to analyze species-environment relationships.
Hydrobiologia 184: 169-170.
Ter Braak, C. J. F., 1991. CANOCO update notes. GLW-
DLO, Wageningen, the Netherlands (supplied with the pro-
gram).
Van Liere, L., 1986. Loosdrecht lakes, origin, eutrophication,
restoration and research programme. Hydrobiol. Bull. 20:
9-15.
Van Liere, L., S. Parma, L. R. Mur, P. Leentvaar & G. B.
Engelen, 1984. Loosdrecht lakes restoration project, an in-
troduction. Ver. int. Ver. Limnol. 22: 829-834.
Van Liere, L., L. Breebaart & Y. J. Dullemont, 1989a. De-
termining the relative number of prochlorophytes in lake
phytoplankton using epifiuorescence microscopy. Br. phy-
col. J. 24: 391-394.
Van Liere, L., L. Breebaart, W. Kats & J. J. Buyse, 1989b.
The water quality of the Loosdrecht Lakes area. Commissie
voor Hydrologisch Onderzoek TNO. Rapporten en Nota's
22: 265-279. (In Dutch)
Van Liere, L., R. D. Gulati, F. G. Wortelboer & E. H. R. R.
Lammens, 1990a. Phosphorus dynamics following restora-
tion measures in the Loosdrecht Lakes (The Netherlands).
Hydrobiologia 191: 87-95.
Van Liere, L., O. F. R. Van Tongeren, L. Breebaart & W.
Kats, 1990b. Trends in chlorophyll and total phosphorus in
the Loosdrecht lakes, The Netherlands. Ver. int. Ver. Lim-
nol. 24: 707-710.
Van Tongeren, 1991. CEDIT. Preliminary manual, Internal
Report 1991-9. Limnological Institute, Nieuwersluis.
123 pp.
Hydrobiologia 233: 119-136, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 119
© 1992 Kluwer Academic Publishers.

A mathematical model of the phosphorus cycle in Lake Loosdrecht and

simulation of additional measures

J. H. Janse, T. Aldenberg & P. R. G. Kramer

National Institute of Public Health and Environmental Protection (RIVM), P.O. Box 1, 3720 BA
Bilthoven, The Netherlands

Key words: model, simulation, eutrophication, phosphorus, PIC ratio, lake ecosystem

Abstract

The phosphorus cycle in the ecosystem of the shallow, hypertrophic Loosdrecht lakes (The Netherlands)
was simulated by means of the dynamic eutrophication model PCLOOS. The model comprises three
algal groups, zooplankton, fish, detritus, zoobenthos, sediment detritus and some inorganic phospho-
rus fractions. All organic compartments are modelled in two elements, carbon and phosphorus. Within
the model system, the phosphorus cycle is considered as completely closed. Carbon and phosphorus are
described independently, so that the dynamics of the PIC ratios can be modelled. The model has been
partly calibrated by a method based on Bayesian statistics combined with a Range Check procedure.
Simulations were carried out for Lake Loosdrecht for the periods before and after the restoration
measures in 1984, which reduced the external phosphorus loading to the lake from ca. 2 mgP m - 2 d - 1
to 1 mgP m - 2 d - I. The model outcome was largely comparable with the measured data. Total phos-
phorus has slowly decreased from an average 130 JlgP 1- I to ca. 80 JlgP 1- I, but chlorophyll-a (ca.
l50Jlgl-\ summer-averaged) and seston concentrations (8-l5mgCI- 1 ) hardly changed since the
restoration measures. About two-thirds of the seston consisted of detritus, while the phytoplankton
remained dominated by filamentous cyanobacteria. The PIC ratio of the seston decreased from ca. 1.0%
to 0.7%, while the PIC ratios of zooplankton, zoobenthos and fish have remained constant and are much
higher. The system showed a delayed response to the decreased phosphorus loading until a new equi-
librium was reached in ca. five years. Major reasons for the observed resilience of the lake in responding
to the load reduction are the high phosphorus assimilation efficiency of the cyanobacteria and the high
internal recycling of phosphorus. A further reduction of nutrient loading, perhaps in combination with
additional measures like biomanipulation, will be the most fruitful additional restoration measure.

Introduction 150-200 Jlg 1- I, high primary production and

Like many other shallow lakes in the Netherlands
the Loosdrecht lakes have become highly eutro-
phic due to heavy nutrient loadings for the last
forty years or so. This has led to almost perma-
nent blooms of filamentous cyanobacteria with
mean summer chlorophyll-a concentrations of
seston concentrations and loss of biological di-
versity (I;Iofstra & VanLiere, 1992). In order to
reduce the algal growth, several measures have
been taken to reduce the external phosphorus
loading. From 1984 onwards, the source of sum-
mer inlet water, the polluted River Vecht, has
been replaced by dephosphorized canal water.
120
This has resulted in a decrease in the average
external phosphorus loading to the main lake from
ca. 2mgPm- 2 d- 1 in 1983 to 1 mgPm- 2 d- 1
from 1984 onwards (Engelen et al., 1992; Buyse,
1988). This has, however, not led to a decrease of
the seston concentration (Van Liere et al., 1990).
The Water Quality research Loosdrecht lakes
project (WQL) was carried out to assess and
quantify the effects of reduction in phosphorus
load on the lake ecosystem and to recommend
additional water management measures (Van
Liere et al., 1984). A mathematical model, called
PCLOOS, was developed at the National Insti-
tute of Public Health and Environmental Protec-
tion (RIVM), one of the participants in the WQL
project. The aim of the model is to provide a
better insight into the functioning of the system by
integrating the results of the WQL project and to
predict the effects of different restoration mea-
sures.
Several phosphorus models have been reported
elsewhere, ranging from very simple to very com-
plex ones. The most simple models, the so-called
empirical or 'black box' models (OECD, 1982;
Reckhow & Chapra, 1983; Sas, 1989), relate the
in-lake total phosphorus concentration to the ex-
ternal loading and the water retention time and
describe statistically the relationship between
algal biomass and total P concentration. A sec-
ond group are the dynamic models which de-
scribe phytoplankton growth as function of nu-
trient availability (see overview in Straskraba &
Gnauck, 1985). Some of these models do, others
do not include nutrient cycling in the sediment.
The third group are detailed, multi-species mod-
els of the trophic web in aquatic ecosystems (e.g.
Patten et al., 1975).
The empirical models are, in fact, one-compart-
ment models which cover the processes of inflow,
dilution and net retention in the sediment. Such
models are not sophisticated enough for the
present study for two reasons. Firstly, in a shal-
low lake like Loosdrecht, with a very intense con-
tact between water and sediment due to contin-
uous mixing, the net retention is likely to decrease
following the restoration measures (Sas, 1989).
This implies that we need to include the upper
sediment layer explicitly in the model. Secondly,
the relationship between total phosphorus on one
hand and algal biomass and total seston on the
other hand is not unequivocal, so that a static
description of this relationship is not sufficient. In
the WQL project a dynamic model is needed,
with a distinction between algal biomass and algal
phosphorus and a variable ratio between them.
Furthermore, detritus has to be considered as a
separate compartment, also expressed in two
units. Some dynamic phytoplankton models (e.g.
Lehman et at., 1975; Bierman & Dolan, 1981)
use, in fact, a variable stoichiometry approach,
but they either lack a sediment module or the
description of the food web does not match the
needs of the present study.
Apart from the algal growth in response to the
phosphorus load we were also interested in the
phosphorus exchange with higher trophic levels,
especially zooplankton and fish. This required
some modifications in the modelling of the phos-
phorus cycle. A basic feature of the PCLOOS
model is that the phosphorus cycle is completely
closed, except for external loading to and losses
from the lake. This feature is important not only
for describing algal growth or the exchange be-
tween water and sediment, but also for consider-
ing the phosphorus flow through the food web. It
has been observed that at higher levels in the food
chain the weight-specific phosphorus content of
the organisms increases due to accumulation of
phosphorus. The PIC ratio of the seston can be
highly variable and is quite low in the Loosdrecht
lakes (less than 1% ; Gu1ati et al., 1991), whereas
this ratio is much higher for the zooplankton
(2.0-2.5%; Gulati et al., 1991) and zoobenthos
(2.0-2.5%; Dr. J. Vijverberg, Pers. Comm.) and
still higher for fish (ca. 5.5%; Dr. E. H. R. R.
Lammens, Pers. Comm.). These differences can
be maintained only if the animals have mecha-
nisms to regulate their P/C ratio irrespective of
the PIC ratio of their food. These mechanisms
can involve a more efficient assimilation of phos-
phorus from the food and/or a reduced phospho-
rus excretion rate. In this respect, our model dif-
fers from several existing ones (e.g. Di Toro et al.,
1975; Di Toro & Matystik, 1980; Jorgensen,

1980). Aldenberg & Peters (1990) have stated that
in a model in which the PIC ratios of the trophic
levels differ, the PIC ratios of one or more com-
partments should be variable in order to maintain
a closed P balance. In their model, only the de-
tritus has a variable PIC ratio while the other
compartments, phytoplankton and zooplankton,
have a fixed one.
The PCLOOS model, version 2.4 (J anse & AI-
denberg, 1990a, b, 1991) is based on the model by
Aldenberg & Peters (1990), but extended with a
sediment module, a more complete food web in-
cluding zoobenthos and fish and variable PIC
ratios of all the biotic and organic compartments.
The model presented in this paper is PCLOOS
version 2.5. It differs from the previous version in
the incorporation of chemical adsorption, an im-
proved modelling of some sediment processes and
a number of minor changes in the phytoplankton,
fish and zoobenthos growth equations. Moreover,
the model was extended with a calibration routine
based on Range-Check and Bayesian statistics.
Description of the lake
The model has been applied to the Loosdrecht
lakes, a system of interconnected shallow lakes
which originate from peat digging, located be-
tween Amsterdam and Utrecht, The Netherlands.
A detailed description of the lakes system can be
found in Gulati et al. (1990). The main lake, Lake
Loosdrecht, has a water area of 989.1 ha, a mean
depth of 1.91 m and an average water residence
time of 0.7 years. Downward seepage averages
1.5 mm d - I (Engelen et al., 1992). The bottom
consists of peat covered by a layer of loose or-
ganic matter of variable thickness. The lake is
very much exposed to the wind and can be re-
garded as completely mixed.
Model structure
General features
The model comprises both the water column and
the upper sediment layer (Fig. 1). The model is
121
'sheries
egestion + death
Fig. 1. Overview of the PCLOOS model. Compartments
modelled in two units are indicated by two boxes behind each
other. The algal compartment is composed of three functional
groups: cyanobacteria, diatoms and green algae. The arrows
in the figure denote both the carbon and the phosphorus fluxes.
Some minor arrows have been lumped or left out for the sake
of clarity.
zero-dimensional in space: the lake is assumed to
be completely mixed, a reasonable assumption
regarding the high correlation of the data from the
two sampling stations in Lake Loosdrecht, as can
be derived from Breebaart et al. (1989). Carbon
is used as the unit of biomass. The model is com-
posed of eight organic compartments, modelled
both in carbon and in phosphorus units, indi-
cated as two partly overlapping boxes, and four
inorganic phosphorus compartments, making
twenty state variables altogether. Inorganic car-
bon is not included in the model: decomposition
and respiration losses of CO 2 go to an 'outside'
pool of dissolved inorganic carbon (DIC), from
which CO 2 for algal growth is withdrawn. This
assumption seems reasonable, because the phy-
toplankton growth in the lakes is not carbon-
limited. For analogous reasons, other nutrients
than phosphorus are not considered.
Inorganic phosphorus is modelled both in the
surface water and in the interstitial water. In both
phases, a distinction is made between soluble re-
active phosphorus (SRP), which is assumed to be
biologically available, and adsorbed inorganic
phosphorus. The organic matter in the lake, which
122
is rich in iron, has a high adsorption capacity for
phosphorus (Otten et al., 1992).
The organic compartments of the model are:
three functional groups of phytoplankton, detri-
tus, zooplankton, sediment detritus, macrozoob-
enthos and fish. Of the three groups distinguished
in the phytoplankton, diatoms, green algae and
filamentous cyanobacteria, only the last-named is
important at present. The term 'zooplankton' re-
fers to the herbivorous filter feeders, mainly con-
sisting of Crustacea. The fish mainly consists of
bream and represents an important phosphorus
pool in the lake. The macrozoobenthos, mainly
Chironomidae larvae, is an important food source
for the fish besides zooplankton. The functional
groups of organisms are not split any further to
avoid further complexity of the model. The bac-
teria are modelled implicitly in the decomposition
processes.
For each state variable a differential equation
is formulated, made up of the processes (mass
fluxes) linking them (Fig. 1). Although the model
has a complex structure with many parameters, it
is not as complex as it might look at first sight.
Phosphorus fluxes related to natural mortality,
sedimentation or outflow, among others, simply
equal the corresponding C fluxes times the actual
PIC ratio of the compartment. These processes
do not affect the compartment's PIC ratio. In
other cases, however, corresponding fluxes are
independent: for instance, the PIC ratio of food
assimilated by animals may differ from the ratio
of food egested; similarly, the carbon and phos-
phorus uptake by algae may be in a different pro-
portion from the availability of these elements.
These processes determine the PIC ratio of the
species and are themselves controlled by it. The
main process formulations in the model will be
discussed briefly. A complete description of the
version 2.4 model can be found in J anse & AI-
denberg (1990a, b); a description of the version
2.5 model can be obtained from the authors.
Processes in the water
The external phosphorus loading to the lake oc-
curs mainly as detrital, adsorbed or dissolved
phosphorus; the inflow of P in biota is assumed
to be very small. The total phosphorus input is
divided over the different fractions according to
the estimated average values. For detritus and the
biota, the external carbon input is calculated by
dividing the phosphorus input by constant, esti-
mated PIC ratios. The detritus in the input flows
directly to the detritus pool in the lake and is thus
assumed to react in the same way. All compart-
ments in the water, except fish, are subject to
dilution due to water inflow into the lake, cor-
rected for evaporation. As we assume a constant
volume, the same amount of water leaves the lake
as flows into it, the greater part of it by surface
outflow. This water has the same composition of
dissolved and particulate substances as the lake
water. The same applies, except for the zooplank-
ton component, for the water that leaves the lake
by infiltration to the sediment.
The inorganic phosphorus can be present as
dissolved phosphorus (SRP) or can be chemically
adsorbed onto detrital particles. This is modelled
as a, fast, chemical equilibrium between the two
states, modifying the phosphorus flows which are
due to the, relatively slow, biological processes
(Di Toro, 1976). The adsorbed phosphorus is as-
sumed to be subject to sedimentation, infiltration
and grazing at the same rates as detritus, but not
to other biological processes.
Primary production of phytoplankton is de-
fined in the model as the integral carbon fixation
in a 24 h period. For simplicity, the carbon fixa-
tion is assumed to equal growth. Algal biomass is
calculated as a carbon concentration; the growth
rate depends on temperature, available light and
phosphorus. The light dependent growth of cy-
anobacteria and diatoms is described according
to Di Toro & Matystik (1980), using Steele's
equation integrated with respect to depth and time
(1 day). This equation implies growth inhibition
at high light intensities. For green algae, a simi-
lar equation is based on a Monod-type equation,
assuming no light inhibition. The available light is
determined by the light intensity at the water sur-
face and its extinction in the water column. The
extinction coefficient equals the sum of the back-
ground extinction of the water and the contribu-

tions of algae and detritus to it, thus accounting
for the self-shading effect. The temperature does
not affect the initial light affinity, but it does af-
fect the maximum growth rate that can be reached.
This dependence is described using an optimum
function, viz. a modified Gauss' curve, normal-
ized to 1.0 at 20°C.
The phosphorus dependence is modelled as a
two-step process. The first one is phosphorus up-
take, increasing the internal PIC ratio or 'cell quo-
ta' Q. The uptake rate increases with increasing
external SRP concentration up to a maximum.
The initial affinity, viz. the slope of the curve at
low external SRP concentrations, is taken as con-
stant, according to the results of cultures studies
by Riegman (1985). The maximum uptake rate,
however, is determined by the cell quota: the min-
imum cell quota gives the highest maximum rate
(lanse & Aldenberg, 1991). The second process
is growth, taken as increase in carbon, depending
on Q according to the Droop (1974) equation:
above the minimum cell quota, the growth rate
increases asymptotically with Q until the maxi-
mum growth rate is reached.
The actual growth rate is calculated by multi-
plying the maximum growth rate with the com-
bined reduction functions for light and tempera-
ture and for phosphorus (Table 1). Whether light
and nutrient functions operate independent of
one another in reality is discussable. It is gener-
ally agreed that the effects of multiple nutrients,
for instance phosphorus and nitrogen, are best
described by the minimum law of Liebig, viz. only
one of them is limiting at a time. The effect of
temperature is generally reported as independent
of nutrient availability. Concerning light, neither
of the two options seems entirely satisfactory.
We have followed the common assumption by
modellers that light and nutrients affect photo-
synthesis independently. There are some indica-
tions that these factors might influence each
other when both are low (Riegman, 1985), but
we did not incorporate this for sake of sim-
plicity. The chlorophyll-a content of the phyto-
plankton, a derived variable in the model, de-
creases with temperature and increases with the
degree of light limitation (Riegman, 1985). The
123
growth rate [d - I] multiplied by the algal concen-
tration gives the gross primary production
[mgC I - 1 d - 1 ] .
Counteracting algal growth are the loss pro-
cesses respiration, sedimentation and zooplank-
ton grazing. Respiration is the sum of mainte-
nance respiration, taken as a percentage of the
biomass [d - 1], and growth respiration, a rela-
tively small fraction of the daily growth. During
respiration the algae excrete also phosphorus, the
relative amount of which is controlled by their
actual PIC ratio. Sedimentation is defined as
gross sedimentation, the rate of which is slightly
depending on temperature. We have assumed that
settled algae die instantaneously on reaching the
bottom and contribute to the bottom detritus,
which can be resuspended into the water column.
Phytoplankton death is thus modelled indirectly.
The parameter values of the three algal groups
in the model differ. The cyanobacteria have a
higher affinity both for phosphorus and light as
well as a higher maximum uptake rate than the
other groups. On the other hand, they have a
much lower maximum growth rate and a stronger
sensitivity to temperature. The diatoms have a
lower temperature optimum, while the green algae
are not inhibited by high light intensities. Both
these groups have higher growth rates, but also
higher loss rates through settling and zooplank-
ton grazing.
The zooplankton feeds on both phytoplankton
and detritus. Grazing is described as a Monod~
like function of the seston concentration, the spe-
cific filtering rate decreasing hyperbolically with
increasing seston concentration (Gulati et at.,
1982; Gulati et at., 1985). A selectivity constant
is used for each food species to account for pref-
erence of the zooplankton for certain kinds of
food: 1.0 for green algae, 0.5 for diatoms, ca. 0.1
for filamentous cyanobacteria, which are not eas-
ily handled by most filter feeders (Gliwicz, 1980)
and 0.3 for detritus. The assimilation efficiency
for the consumed food is constant and quite low
(0.3) for carbon (Gulati et at., 1985), but variable
(depending on the PIC ratio of the food) and,
therefore, mostly higher for phosphorus. This is
one of the mechanisms by means of which the
124
differences in PIC ratio between the trophic lev-
els are maintained. The egested material becomes
detritus, but from this about half of the phospho-
rus 'leaks out' and enters the SRP pool. Losses
of zooplankton include, besides egestion of the
unutilised food, respiration, P excretion, natural
mortality and fish predation. The P excretion is
related to the respiration, but is reduced in times
of P shortage.
Fish feeds, besides on zooplankton, also on the
zoobenthos (midge larvae). Both predation pro-
cesses are modelled as a so-called 'type III' re-
sponse (Holling, 1965): the predation rate de-
pends on prey density and this relation is
described by a sigmoid curve. Selectivity between
the two food species has not been included. The
phosphorus assimilation efficiency, like for zoop-
lankton, depends on the PIC ratio of the food.
Also fish can reduce its P excretion rate with
respect to the respiration rate in order to save
phosphorus. Natural mortality of fish varies sea-
sonally, being the highest after spawning in May-
July (Dr. E. H. R. R. Lammens, Pers. Comm.).
The scales and bones of the dead fish are lost
from the system. The softer body parts become
detritus, besides a small part of the phosphorus
which immediately becomes available in dissolved
form. Only small amounts of fish are removed by
active fisheries.
Detritus is an important compartment in the
model. It has an extensive exchange with the sedi-
ment detritus, the organic matter in the upper
sediment, by means of sedimentation and resus-
pension. Sedimentation is, like the sedimentation
of algae, described as a first-order process, the
velocity being slightly influenced by temperature.
Detritus disappears due to decomposition and
mineralization, a first-order, strongly temperature
dependent process, and the dissolved P is re-
leased and enters the SRP pool.
Processes in the sediment
The sediment module is confined to the exchange-
able, degradable material assumed to be impor-
tant for the nutrient recycling in the lake ecosys-
tern, according to Keizer & Sinke (1992). These
authors found that only the recently sedimented
organic fraction and 10-20% of the inorganic
fraction in the upper 10 cm can be recycled. To-
gether they amount to less than 1.0 gP m - 2. Other
sediment fractions, like peat particles and the
phosphorus chemically bound to humic sub-
stances etc. are considered as unimportant for the
nutrient recycling (see e.g. Bostrom et al., 1982)
and are not included, although they can be in
considerable amount. The total P concentration
of the sediment is about 1.1 mgP g - 1 dry weight,
at a dry weight content of 90 g 1- 1 (Boers et at.,
1984), which equals nearly 10 gP m - 2 in the upper
10 cm. The sediment has a very high porosity of
0.91.
The most important compartment in the sedi-
ment is the sediment detritus, originating from
sedimentation (and infiltration) of seston from
the water column. This material can be recycled
back to the water column by resuspension, which
is estimated to affect 10% of the sediment surface
of the lake per day at average wind velocity
(Gons & Van Keulen, 1989). The average total
phosphorus flux by resuspension is estimated as
30-40mgPm- 2 d- 1, of which 10%, viz. 3-
4 mgP m - 2 d - 1, consists of 'true', fine-sized, de-
tritus (Gons et al., 1991). The other 90 % of the
resuspendable material, also called 'epipelon',
consist of coarse particles of peaty origin which
settle again within a few hours and which are not
modelled. The resuspension flux is, of course,
highly variable and greatly affected by wind speed,
ice cover and also - to a lesser extent - by boating
and by browsing activity of bream in the sedi-
ment. For simplicity, however, the flux is mod-
elled as a constant amount of carbon per day
(provided that the size of the sediment detritus
pool is sufficient). The amount of phosphorus re-
suspended this way depends on the actual PIC
ratio of the sediment detritus.
Mineralization of the sediment detritus is de-
scribed as a first-order process with a strong
temperature dependency. The dissolved phospho-
rus in the interstitial water is again assumed to
be in chemical equilibrium with the adsorbed
form. However, since the anoxic conditions in

most of the sediment except in the upper few
millimeters, the adsorption capacity of the sedi-
ment is less than that in the water (Keizer et al.,
1991). Dissolved phosphorus can diffuse from the
pore water to the surface water or can get lost by
downward seepage.
The macrozoobenthos is assumed to feed only
on the sediment detritus. Its feeding rate has been
modelled by a Monod-like function of the avail-
able food, a so-called 'Type II' functional re-
sponse (Holling, 1965). Loss processes are res-
piration, mortality (first-order) and fish predation.
Also the zoobenthos can regulate its PIC ratio to
a level higher than that of the food ingested by
means of a more efficient P assimilation or by a
reduced P excretion rate.
Parameter estimation and calibration
The main input parameters for the model were:
the measured water temperature, daily radiation,
water inflow, evaporation and total phosphorus
input. Radiation data from the meteorological
station at De Bilt, about 20 km from the lakes,
were obtained from the Royal Netherlands Me-
teorological Institute (KNMI) and were read in
as weekly values of a two-week moving average,
to smoothe the input and to reduce the number
of data. Hydrological data and phosphorus load-
ing were read from the monthly balance models
(1982-1987) made by Buyse (1988) and Engelen
et al. (1992), both for the Loosdrecht lakes to-
gether and for each lake separately. For 1988-
1990, the balance models for 1986 were used.
Other important input parameters were the water
depth, infiltration rate, sediment porosity, back-
ground extinction and a number of process pa-
rameters. The sources of these parameters were:
experimental data of WQL, literature data and
calibration. Experimental data on the Loosdrecht
lakes concerned the growth of cyanobacteria, the
growth of crustacean zooplankton, settling and
resuspension of seston and the phosphorus mo-
bilisation. Also, good estimates were available of
the phosphorus pools in the sediment and of
stocks and mortality offish. Data on zoobenthos,
125
food selectivity in grazing, fish predation and
mineralization in the water were scanty. Field
data for calibration were collected by the Munic-
ipal Water Works of Amsterdam and by the Lim-
nological Institute at Nieuwersluis. The sum of
phytoplankton and detritus was measured as
seston < 150 f.1m of which, among others, dry
weight and C, P and chlorophyll-a content were
measured. These measurements do not allow to
separate algae from detritus; this was done with
use of the chlorophyll-a data and estimated
C:chlorophyll-a ratios of the phytoplankton
(Gons & Van Keulen, 1989) and with use ofbio-
volume data which were collected in some peri-
ods. Zooplankton was measured as the seston
> 150 f.1m. In this paper, the term 'seston' refers
to the seston < 150 f.1m. Other measurements
used were SRP, total phosphorus and Secchi-
depth transparency.
For calibration of the model, we first made use
of so-called 'minimodels' of subsystems, e.g. the
sediment subsystem, and of steady state analyses
with summer averages. Secondly, a formal cali-
bration procedure was developed, based on Bay-
esian statistics (Box & Tiao, 1973), combined with
a Range-Check procedure. The aim of the method
of Bayesian parameter estimation was to explore
how the combined uncertainty in a large number
of parameters affects the uncertainty in the model
results, and to estimate the parameter values more
accurately. An important feature of the method is
that a number of parameters are observed to-
gether rather than separately. Because the method
is quite computer time-consuming, it was applied
to only a subset of 11 of the ca. 100 parameters
together. The other parameters were kept at their
nominal value. The choice of the 11 parameters
was based on their high uncertainty, their sensi-
tivity and because they appear in different mod-
ules of the model. For each of them, a range of
possible values was established. At the same time,
output ranges of all the state variables were es-
tablished, based on data and estimates on Lake
Loosdrecht. Next, a large number of model runs
was performed with randomly chosen parameter
combinations. In every run, the model results were
first compared with the output ranges, using the
126
values on August 1 of the first year for compari-
son. If the simulation fell beyond one or more of
these ranges, the parameter combination was
rejected, the run was stopped and a new combi-
nation was chosen. If the simulation was within
the ranges, the run was allowed to continue.
This Range-Check procedure saved a consider-
able amount of computing time spent other-
wise on unproductive runs. The successful runs
were subjected to Bayesian statistics. For every
measurement the difference between measured
and simulated values was calculated and then
used to compute the sum of squares for the var-
iable:
SumOfSquares var = L (obs var • t - model var .J 2 ,
t
with t = time. The eight obtained sum of squares
values, one for each measured variable, were used
to calculate the marginal posterior density (Box,
1971 ):
posterior density =
8 1
fl (S um Of Squares var ) - "2' nva,
var = 1
in which n = number of measurements, which can
be different for different variables. The parameter
combination with the maximum posterior density
gives the best fit of the model to the measure-
ments. This combination was used in the simu-
lations depicted.
The model c~lculations were performed on a
SUN Spare station or on a 80386 PC, by means
of the simulation program ACSL, level lOB, de-
veloped by Mitchell & Gauthier Associates
(1991). The integration routine used was Gear's
Stiff algorithm, with a variable order and variable
step size, which were adapted according to pre-
set error criteria. For our model, which includes
processes with highly different time constants, this
algorithm proved to be considerably faster than
others. A ten-years simulation took 5-10 minutes
per run. Initial values of the state variables were
estimated from the data. Prior to a simulation, the
model was allowed to equilibrate for a simulated
period of three years.
Results
Simulations
Simulations have been carried out for the period
1982-1990 for three lakes of the Loosdrecht sys-
tem, but here only the results for the largest lake,
Lake Loosdrecht, are described. In the years up
to 1984, the average total phosphorus concentra-
tion in the inflowing water (viz. the external P
loading divided by the water inflow) was very high
in the summer months, due to the inlet of highly
polluted supply water from the River Vecht
(Fig. 2). After restoration measures in 1984, these
peak loadings have disappeared, but the loading
from other sources remains high (Buyse, 1989;
Engelen et al., 1992). On annual basis, the exter-
nal loading has been reduced by a factor 2 (Van
Liere & lanse, 1992).
During the calibration of the model, about one
run out of every six was successful, i.e. the results
fell within the established ranges for the state var-
iables. 200-300 runs were sufficient to find the
maximum posterior density: the natural logarithm
,
of it was -503. Simulations, together with the
data, show that total phosphorus (Fig. 3) has
tended to decrease steadily following the restora-
tion measures, although the u'nexplained varia-
tion in the data is high. The solid line is the 'best
0.8 ~------------------,
%,0.6
oS
~
£
.£ 0.4
I/)
::>
o
.c
a.
I/)
,ga. 0.2
(ij
9
Fig. 2. Average total phosphorus concentration in the inflow-
ing water to Lake Loosdrecht, 1982-1990. Based on data of
Buyse (1988) and Engelen et al. (1992). The values for 1988-
1990 are based on those of 1986.
 127

0.20 25 ,--------------.<5',---------------,

"

20
:;:0.15
Q.

.s
Ol

"' 0. 10
0'"
~
0.
"'0
~
0.
0
00
0

Cii
§ 0.05 0 0
0
0
0
'I>
;,
;,
:''.
. ,
"
0.00 .J---1-98-2-'--1-98-3-'--1-98-4---'--19-85---r-19-8-6-r-1-98~7-'---:19-:-88-'---:-19~89:-r~19::::90:::- 0 '. -. .. "
1982 1983 1984 1985 1986 1987 1988 1989 1990

Fig, 3, Total phosphorus concentration in Lake Loosdrecht: Fig. 5. Simulation of total seston (solid line) and cyanobac-
simulation of nominal run (solid line), error bounds (dashed teria (dashed line) and measurements of total seston (sym-
lines) and measurements (symbols), 1982-1990, bols) in Lake Loosdrecht, in 1982-1990.

fit', the dashed lines indicate the ranges, due to consistent with estimates by Gons & Van Keulen
variation of the 11 calibration parameters, of (1989) of an average detrital fraction in the seston
runs giving a natural logarithm of the posterior mass of two-thirds. The zooplankton (Fig. 6),
density higher than -520. A corresponding de- zoobenthos and fish stock are not much affected
crease in chlorophyll-a was, however, not ob- by the reduced nutrient loading. The simulated
served (Fig. 4), nor in the total seston (Fig. 5) or densities of fish and zoobenthos are in the range
Secchi depth, which still varies between 0.3 and of field estimates: fish 275 kg ha - 1 (correspond-
0.4 m. The cyanobacteria were the dominant ing to about 2.8 gC m -2) (Lammens et al., 1992)
group during most of the year, but in spring, also and zoobenthos 1 gC m - 2 on the average (Dr. J.
the diatoms developed a significant biomass, ac- Vijverberg, Pers. Comm.). Strikingly, the seston
cording to both the simulations and the field ob- concentration if calculated as phosphorus gives a
servations (Boesewinkel-De Bruyn et al., 1988). stronger response to the decreased phosphorus
Major part of the seston mass (ca. 60% in sum- input than the corresponding carbon concentra-
mer, > 90% in winter) consists of detritus; this is tion. The simulated PIC ratio of the seston grad-

400,----------------------, 2.0 , - - - - - - - - - - - - - - - - - - - - - - ,

_300 1.5

'"Eo ()

.s
Ol

.s
Ol
f!
1200 1\ 5 1.0
>- I&'
Si
~
0. 1\ c
o
e \ '"o
a.
E o
() 100 NO.5

0.0 +---r---.----r-----,------.-----.-----r-~---I
1982 1983 1984 1985 1986 1987 1988 1989 1990 1982 1983 1984 1985 1986 1987 1988 1989 1990

Fig. 4. Chlorophyll-a concentration in Lake Loosdrecht: sim- Fig. 6. Zooplankton concentration in Lake Loosdrecht: sim-
ulation (line) and measurements (symbols), 1982-1990, ulation (line) and measurements (symbols), 1982-1990,
128
ually decreases, both in the simulations and in the
lake (Fig. 7), but those of the zooplankton, zoob-
enthos and fish do not change very much, because
of the regulation mechanisms of the animals. The
field zooplankton ratios also do not show a
change (Gulati et aI., 1991). The phosphorus pool
in fish, about 80 /l-g P 1- 1, roughly equals the Pin
all the other compartments in the water phase
together (Van Liere et al., 1990), as found in many
hypertrophic lakes. The SRP values (not shown)
are very low during most of the year, according
to both the model and the field data. As in the
water phase, the simulated P concentration in the
sediment (not shown) also decreases steadily be-
cause of the load reduction; although data are
sparse, observations in the field seem to confirm
this (Keizer & Sinke, 1992). The concentrations
in the pore water vary considerably during the
year, depending on the mineralization activity
which is highly temperature dependent.
The characteristic response time of the model
system to adapt itself to a decrease in nutrient
loading is about five years. A new equilibrium,
which has now been nearly reached, does not
differ much from the one before the restoration
measures, except for a decrease in phosphorus
concentration, not accompanied by a correspond-
ing decrease in chlorophyll-a and seston (Fig. 8).
The shift to the left towards the maximum line
based on a large number of Dutch lakes
(CUWVO, 1987) clearly shows that the chloro-
2.0 , - - - - - - - - - - - - - - - - - -
1.5
c:
o
1ii
~ 1.0
.2
"§
0.5
0.0 +-----;r-~-___,__-~-_,__-,..___.-___,__-_I
1982 1983 1984 1985 1986 1987 1988 1989 1990
Fig. 7. PIC ratio of seston in Lake Loosdrecht: simulation
(line) and measurements (symbols), 1982-1990.
c;)
E
Ol 200
.s
(/)
Q>
Ol
'"
ID 150
iU
0;
E
E
::J
(/)
100
1
>.
£. 50
eo
:cu
50 100 150 200
total phosphorus, summer averages [mg m-3]
Fig. 8. Eutrophication diagram (summer-averaged chloro-
phyll-a against total phosphorus) in Lake Loosdrecht: simu-
lation (circles) and measurements (asterisks), 1982-1990. The
dashed line is the maximum ratio found in a database of
Dutch lakes (Lijklema et al., 1988).
phyll-a: total P ratio is among the highest re-
corded.
The phosphorus turnover rates greatly differ
for the different model compartments, as can be
concluded from the simulated flow scheme, av-
eraged for the summer of 1987 (Fig. 9). For the
SRP compartment, which is very small in size,
this rate is by far the highest (ca. 10 d - 1): SRP
formed is taken up almost immediately by algae
or, when these are saturated (for instance in win-
ter), adsorbed onto iron-rich particles in the
water. This pool of adsorbed phosphorus also
has a high turnover rate (ca. 1 d - t, on the aver-
age), due to rapid chemical exchange with the
SRP pool. Of the organic compartments, the
zooplankton has the fastest turnover, about
0.2 d - 1, the algae, detritus and zoobenthos have
about 0.1 d - 1 and the turnover rate offish is only
0.02 d - 1. The compartment with the slowest
turnover rate is the organic phosphorus in the
sediment: about 0.01 d - 1. About 10 % ofthe daily
phosphorus needs of the phytoplankton is met,
directly or indirectly, by the external loading,
about 10% by desorption from the adsorbed P
pool and only 3 % by diffusion from the sediment.
The larger part, i.e. ca. 75%, becomes available
again by recycling via zooplankton, fish, detritus
or the algae themselves. The phosphorus flow
 129

The modelled diffusive flux approximately equals

the flux measured in sediment cores (Boers & Van
Hese, 1988; Keizer et al., 1991); it has decreased
a little since the restoration measures (Figs. 9 and
10).

Simulation of additional measures

Because the achieved reduction in phosphorus

loading of the lakes did not lead to a discernible
improvement in the water quality, additional
measures were considered by the regional water
sediment quality authorities (see also Van Liere & J anse,
1992). The possible effects of the following mea-

-
sures have been modelled, with the same param-
eter settings as established during the calibration:

-
Phosphorus /luxes {mgP m-2 d-l]:
>8
3-8
1.25 - 3
Fig. 9. Simulated phosphorus flow scheme of Lake Loos-
drecht, summer average 1987. Concentrations in mgP m -
flows in mgP m - 2 d - I.
from the sediment to the water phase occurs
mainly through resuspension; the flow due to fish
feeding and the diffusive flux are much smaller.
0.5 -1.25
0.2-0.5
< 0.2
(a) No further change in the present phosphorus
loading (1.0 mgP m - 2 d - 1).
(b) A further reduction in the phosphorus load-
2, ing to 0.35 mgP m - 2 d - 1 (= 0.1 gP m - 2
yr - 1), i.e. a reduction which is the maximum
technically possible.
(c) Extra flushing with 1'106 m 3 dephosphorised
water per month. This is the amount techni-
cally possible given the present hydrological

P-liSh 130
inflow inflow
230 l-""o~utl~IO'::'W-!-ll~ 170 l----:o~u~tlIo~W+_
1.8 P-waler 0.5 1.1 P-water 0.4

settling se"ling 62 57 resuspension

8.5 7.6
infiltration infiltration . . diffusion
fish feeding

-
1983 1987

Phosphorus /luxes {mgP m-2 d-l]:

>8 0.5 -1.25
3-8 0.2 -0.5
1.25 - 3 < 0.2

2
Fig. 10. Simplified phosphorus flow schemes of Lake Loosdrecht, yearly averages; concentrations in mgPm- , flows in
mgP m - 2 d - I (a) 1983. (b) 1987.
130

situation and the maximally allowed fluctua-
tions in water level (Buyse, 1988).
(d) Removal, by dredging, of 80 % of the organic
material in the sediment top layer, carried out
once. It is questionable whether this measure
is technically possible in the Loosdrecht lakes
(Van Liere & lanse, 1992).
(e) Reduction of the resuspension with 50% by
deepening of the lake (Gons & Van Keulen,
1989).
(0 Biomanipulation, involving removal of 80%
of the bream population, to be carried out
once.
(g) Phosphorus load reduction (measure a), com-
bined with biomanipulation, introduction of
predatory fish and 80% reduction of resus-
pension, under the assumption that sub-
merged vegetation will recolonize and sur-
vive.
The possible effects of the measures are presented
as time series of the total P and seston concen-
trations (Fig. 11) and by means of the so-called
AMOEBE method (Fig. 12). By this method, the
state of an ecosystem is visualised, in an inte-
grated way, with respect to a 'reference state'.
This is a goal state, based on the state of the lakes
in the period 1930-1950 (Hofstra & Van Liere,
1992). The results of the PCLOOS model for the
present situation are comparable with those of the
AMOEBE method. Fig. 12 shows the yearly av-
eraged results of the various measures, six years
after their application. The 'no change' scenario
results in no further effects, because the system is
now (nearly) in equilibrium with the prevailing
phosphorus loading. A further reduction of phos-
phorus loading (measure b) would certainly lead
to some permanent improvement of water qual-
ity, with total P concentration, seston and
chlorophyll-a about half the present values and
consequently, the transparency will increase.
However, according to the simulations, the phy-
toplankton will probably still be dominated by
cyanobacteria and the improvement will not be
sufficient to reach the reference state. The same
applies for resuspension reduction as a stand-
alone measure (measure e). Biomanipulation
(measure 0 or removal of the top layer (mea-
sure d), without a concomitant reduction in phos-
phorus loading, will be effective only temporarily.
Scenario (g), a combination of phosphorus load
reduction and additional measures, seems to be
promising. One must bear in mind, however, that
this scenario is rather speculative, because it as-
sumes the returning of aquatic macrophytes,
which are not an explicit part of the model but are
regarded as a boundary condition (see discussion
section).

20 ~--------------------,
0.20,-----------------------,

15
~015
0..
0>
oS '5,
0'"
:::J
0 . 10
oS 10
.r::: c
Q. B
~ '"
Ql
.r:::
Q. '"
B 0.05
(1j

0.00 .L-_,-----_~____,-~-__,__-~-___,_-__.________j
o 4 5 0 5 6

~~ ~~

Fig. I I. Simulations of additional measures. (a) total phosphorus, (b) seston. Solid lines: 'no change' scenario (measure a); dot-
ted lines: biomanipulation (measure f); dashed lines: P load reduction (measure b); dotted-dashed lines: P load reduction together
with biomanipulation (measure g). For other details see text.
 131

·
2ll 4x ax 4x ax

···
,
·
·· ·
·
.· ..··
··
measure (b) measure (c)

····
ex 2ll 4x ax

··
2ll 4x

····
",'
·.· .,,/ ,1"
···· .
",/

r~' #
br~# ,.'

measure (d)
-~I::... r:1+~L::::::"'/
cyanobaclena ·c'hTd'~r-a·· measure (e)

.,
4x ax 2ll 4x ax

··
.
·
·· .··
·
.· ...···
.
··
measure (g)

Fig. 12. Simulations of additional measures, depicted by means of the AMOEBE method (slightly adapted from Hofstra & Van
Liere, 1992). Measure b: P load reduction. Measure c: flushing. Measure d: removal of top layer. Measure e: reduction of resus-
pension. Measure f: biomanipulation. Measure g: P load reduction together with biomanipulation. White slices: simulation of 1987;
dark slices: simulations of additional measures (results after 6 years); solid circle: reference state. All values are based on yearly
averages.

Discussion net fluxes between water and sediment. The flux

The lake ecosystem
The overall behaviour of phosphorus in the sys-
tem can be analysed from the simulated flow
schemes (Fig. 10) based on the yearly averaged
from water to sediment is the sum of settling and
infiltration, the opposite flux is the sum of resus-
pension, diffusion and fish feeding. It is obvious
that the sediments act, on the average, as a net
sink of phosphorus and that they continue to do
so after load reduction. These schemes can be
132
compared with 'empirical' lake parameters like
the net retention coefficient, R, defined as the
fraction of the input which is retained in the sed-
iments
P - P
R = In out [ _ ] ,
Pin
with Pin denoting the phosphorus inflow and POUI
the surface outflow of phosphorus. The retention
coefficient has decreased from 0.7 to 0.6 after
the load reduction, a phenomenon observed fre-
quently in lakes under restoration and often at-
tributable to an increased internal loading in the
summer months (e.g. Ryding & Forsberg, 1977;
De Pinto, 1981; Sas, 1989). In Lake Loosdrecht,
where the direct phosphorus mobilisation due to
diffusion is very small (Keizer et al., 1991), the
cause of the decreased retention lies in the high
efficiency of phosphorus utilisation by the phyto-
plankton and in the high resuspension. The ob-
served retention coefficients comply with empir-
ical relations with lake depth and water retention
time (OECD, 1982; Lijklema et al., 1988; Sas,
1989). The decrease in phosphorus that does
occur, follows the load reduction with a certain
delay and a new equilibrium is reached after ca.
five years. This 'characteristic time' of the system
depends mainly on the processes which have the
lowest specific rates, i.e. mainly processes in the
sediment.
Apart from changes in the phosphorus flows,
another important feature is the decrease of the
PIC ratios of phytoplankton and seston, follow-
ing the load reduction, both in the simulations
and the field measurements (Fig. 7; see also
Gulati et al., 1991). The simulated summer aver-
ages decrease, between 1983 and 1990, from
1.0% (w/w) to 0.7% for total seston and from
2.1 % to 1.6 % for the cyanobacteria, while the
growth rate remains the same at 0.11 d - I. There
is a concomitant rise in the maximum initial
phosphate uptake rate from about 0.48 to
0.64 mgP mgC - I d - I, a trend also observed in
experimental tests (Baard & Burger-Wiersma,
1991). Both phenomena indicate that phosphorus
limitation of the phytoplankton is becoming more
severe (Healey, 1978; Riegman, 1985). The phy-
toplankton, especially the cyanobacteria, appar-
ently have increased their efficiency of utilising
phosphorus and thus can maintain a high biom-
ass despite a decrease in available phosphorus.
Parallel with a decrease in PIC ratio of the seston
the phosphorus assimilation efficiency of the
zooplankton increases, in order to maintain a high
PIC ratio when the food becomes poorer in phos-
phorus.
Although the general trends in the data are
reproduced correctly by the model, there are dif-
ferences between simulations and measurements,
due to assumptions which may not always be
true. The winter values of chlorophyll-a are often
underpredicted, possibly because the phytoplank-
ton can survive some days in the upper sediment
layer (Otten et at., 1992) and, if resuspended
within that period, will be measured in the water
again. The measured summer chlorophyll-a con-
centration in Lake Loosdrecht in 1982, when the
loading was still high, is lower than simulated by
the model and also lower than the concentration
in 1983. It is not clear whether this difference is
due to methodical aspects, as the measurements
in 1982 have been done by another laboratory, or
that some other factors might have reduced the
algal biomass. The species composition was not
significantly different from the one in later years
(Boesewinkel-De Bruyn et al., 1988).
The underestimation of the variations in the
seston concentration, especially the detritus com-
ponent, in the model, is partly due to the fact that
the resuspension flux is assumed to be constant,
i.e. its dynamics due to variation in wind speed
and other factors, is not modelled. Some of the
field data, however, might have been influenced
by storm events, which increase the contribution
of resuspended matter to the seston. This also
includes peat-derived particles, which most often
settle again within a few hours (GOllS & Van
Keulen, 1989). These particles have a much lower
phosphorus content than the detritus, but a com-
parable carbon content, leading to a temporary
increase in particulate carbon in the water, while
particulate phosphorus is less affected. Apart
from this variation, also the more or less regular

seasonal dynamics of the seston concentration is
underpredicted by the model. In fact, total seston
follows the changes in chlorophyll-a more closely.
The variation in the chlorophyll-a content of the
algae might have been overestimated, or miner-
alization in the water column might have been
more important than assumed.
The seasonal pattern of the zooplankton data
is not adequately reproduced by the model; for
that purpose, the model formulation is too much
a generalisation, while the zooplankton, due to its
position as intermediary trophic level, is quite
sensitive to parameter changes. The zooplankton
is especially sensitive to the food quality, deter-
mined by the ratio of well-edible to poorly-edible
algae; the algal types in turn depend on light con-
ditions and phosphorus availability. Unfortu-
nately, the literature data on the food type selec-
tivity in zooplankton grazing leave a lot of
uncertainty.
Why, then, is the system so resilient in respond-
ing to the load reduction? One reason is that the
reduction has been, on the average, only a factor
two, i.e. less than that expected (Van Liere &
J anse, 1992). But even so, the system seems to be
resistant to react to P decrease. Three factors
seem to be crucial for the continuation of the
prevailing cyanobacterial dominance. Firstly, the
large amount of sediment detritus, which can be
easily resuspended in a wind-exposed, turbulent
lake like Loosdrecht (Gons et al., 1991). Sec-
ondly, the high phosphorus assimilation efficiency
of the cyanobacteria, which go on forming new
organic matter despite a decreased phosphorus
input. For cyanobacteria, this input must be re-
duced further than for other algal types before
their growth becomes phosphorus-limited (Rieg-
man, 1985). Thirdly, the low grazing pressure on
the seston, which makes that most of the phos-
phorus in the lake is recycled within the seston
and that only a small amount is transferred to
higher trophic levels (Van Liere et al., 1990).
Grazing pressure is low because of the poor ed-
ibility of cyanobacteria and also because of the
dense population of bream predating selectively
on the large-bodied zooplankton (Gulati, 1990).
The bream population remains high because of
133
the near lack oflosses by harvesting or predators;
the low light conditions which favour bream are
unfavourable to predatory fish (Lammens et al.,
1992). The load reduction has not been sufficient
to break through this complex of interconnected
factors which tend to keep the system in the
present state.
Additional measures
An extra load reduction is necessary to make a
long-term recovery possible, because high nutri-
ent loadings are the ultimate cause of the hyper-
trophic state of the system. However, a load re-
duction to 0.35 mgP m - 2 d - 1 alone (measure b)
would probably not be sufficient to break the cy-
anobacterial dominance, because the turbidity
will remain high enough to favour this group, due
to the resuspension of sediment detritus. Mea-
sures like removal of the top layer (if technically
possible) or biomanipulation, carried out once
without a further reduction of the external input,
will have only a temporary effect or will have to
be repeated regularly, because the amount of or-
ganic material in the lake will be replenished rap-
idly, due to the high external nutrient load. Flush-
ing is not well possible in Lake Loosdrecht within
the given constraints.
Additional measures may thus only be effective
combined with load reduction, as is done, for
instance, in the scenario g, which involves phos-
phorus load reduction combined with biomanip-
ulation, introduction of predatory fish and 80 %
reduction of resuspension, under the assumption
that submerged vegetation will recolonize and
survive (Figs. 11 and 12). In this scenario, im-
provement is achieved in three ways. Firstly, the
zooplankton grazing pressure on the seston will
increase because the bream population would be
lowered by predatory fish; secondly, the seston
concentration will decrease also because of the
low external nutrient loading and thirdly, it will be
replenished less because of the low resuspension.
This result is still speculative, however, because it
assumes that predatory fish, e.g. pike, will be able
to survive. This implies the returning of aquatic
134
macrophytes, which will reduce resuspension by
stabilizing the sediment and which yield hiding
and nesting sites for pike (Van Liere et al., 1990).
They are, however, not an explicit part of the
model but are regarded as a boundary condition;
possible effects of macrophytes on the nutrient
cycling are not considered. Anyway, development
of macrophytes will be only possible if a suffi-
ciently long period of clear water can be achieved,
which implies reduction of the formation of or-
ganic material in the system. Several authors (e.g.
Moss, 1980) have shown that a lake ecosystem
can have several stable equilibrium states,
whereof two extrema are the hypertrophic phyto-
plankton-dominated state and the mesotrophic
macrophytes-dominated state. Transition be-
tween these states as a function of a changing
nutrient input is subject to hysteresis effects,
which obstruct restoration of hypertrophic lakes.
Scheffer (1989) has shown these effects in a two-
compartment bream-pike model, where the 'at-
tractability' of the clear, macrophytes-dominated,
'pike state' became less with increasing trophic
state. A system already being in the 'pike state'
could stand some eutrophication without being
drawn back to the turbid 'bream state'. In Schef-
fer's model the effect is due to the way the equa-
tions describe the interaction between the two fish
species. In PCLOOS, hysteresis effects occur as
a result of the differences in parameter values
between the phytoplankton species in the model,
which can lead to a rapid transition, in the sim-
ulations, to a state dominated by green algae as
soon as the system would pass a certain thresh-
old. These aspects need to be investigated further
when the model will be validated also on the data
of other shallow lakes.
Acknowledgements
This project would not have been possible with-
out cooperation with many other research work-
ers of the WQL project. We like to thank all those
people (who unfortunately can not be mentioned
all by name) for their valuable comments and
suggestions. We also thank Drs. J. J. Buyse, the
staff of the Limnological Institute and the Mu-
nicipal Waterworks of Amsterdam for putting
their data sets to our disposal. Mr. R. Lammers
is thanked for his technical support concerning
the computer network, Drs. J. J. Hofstra for the
AMOEBE pictures and Drs. F. G. Wortelboer
and two external referees for commenting the
draft of the manuscript. This project has been
made possible by financial support of the Dutch
Ministry of Housing, Physical Planning and En-
vironment.
References
Aldenberg, T. & J. S. Peters, 1990. On relating empirical water
quality diagrams and plankton-dynamical models: the
SAMPLE methodology applied to a drinking water storage
reservoir. Arch. Hydrobiol. Beih. Ergebn. Limnol. 33: 893-
911.
Baard, R. & T. Burger-Wiersma, 1991. Aspecten van fosfaat-
gelimiteerde groei van fytoplankton uit de Loosdrechtse
Plassen. WOL rapport 1991-1. Laboratorium voor Micro-
biologie, Universiteit van Amsterdam.
Bierman, V. J. & D. M. Dolan, 1981. Modeling of phyto-
plankton-nutrient dynamics in Saginaw Bay, Lake Huron.
J. Great Lakes Res. 7: 409-439.
Boers, P. C. M., J. W. Th. Bongers, A. G. Wisselo & Th. E.
Cappenberg, 1984. Loosdrecht lakes Restoration Project:
Sediment phosphorus distribution and release from the sed-
iments. Verh. int. Ver. Limnol. 22: 842-847.
Boers, P. C. M. & O. Van Hese, 1988. Phosphorus release
from the peaty sediments of the Loosdrecht lakes (The
Netherlands). Wat. Res. 22: 355-363.
Boesewinkel-De Bruyn, P. J., O. F. R. Van Tongeren & B. Z.
Salome, 1988. Kwantitatief fytoplanktononderzoek in het
Loosdrechtse P1assen-gebied 1985-1987; vergelijking met
de jaren 1981 tim 1984. WOL-rapport 1988-8. Limnolo-
gisch Instituut, Nieuwersluis.
Bostrom, B., M. Jansson & C. Forsberg (1982). Phosphorus
release from lake sediments. Arch. Hydrobiol. Beih. 18:
5-59.
Box, M. J., 1971. A parameter estimation criterion for mul-
tiresponse models applicable when some observations are
missing. Applied Statistics 20: 1-7.
Box, G. E. P. & G. C. Tiao, 1973. Bayesian inference in sta-
tistical analysis. Addison-Wesley Publ. Cy., Reading,
Mass., USA.
Breebaart, L., J. Ebert & L. Van Liere, 1989. Basisgegevens
WOL 1983-1988. WOL-rapport 1989-5. Limnologisch In-
stituut, Nieuwersluis.
Buyse, J. J., 1988. Water- en stofbalansmodellen voor de
Loosdrechtse Plassen. WOL rapport 1988-7. Instituut voor

Aardwetenschappen, Vrije Universiteit, Amsterdam, 87 pp.
De Pinto, J. V., 1981. Aquatic sediments, literature review. J.
Wat. Poll. Contr. Fed. 53: 999-1007.
Di Toro, D. M., 1976. Combining chemical equilibrium and
phytoplankton models - A general methodology. In: R. P.
Canale (ed.), Modeling biochemical processes in aquatic
ecosystems. Ann Arbor Science Pub\. Inc., Ann Arbor,
Mich.: 233-255.
Di Toro, D. M. & W. F. Matystik, 1980. Mathematical mod-
els of water quality in large lakes. Part 1: Lake Huron and
Saginaw Bay. EPA 600/3-80-56, Duluth, MI, USA.
Di Toro, D. M., D. J. O'Connor, R. V. Thomann & J. L.
Mancini, 1975. Phytoplankton-zooplankton-nutrient inter-
action model for Western Lake Erie. In: B. C. Patten (ed.),
Systems analysis and simulation in ecology, vo\. III. Aca-
demic Press: 423-474.
Droop, M. R., 1974. The nutrient status of algal cells in con-
tinuous culture. J. mar. bio\. Ass. UK 54: 825-855.
Engelen, G. B., B. F. M. Kal, J. J. Buyse & F. G. M. Van
Pruissen, 1992. The hydrology of the Loosdrecht lakes area.
Hydrobiologia 233: 21-38.
Gliwicz, Z. M., 1980. Filtering rates, food size selection and
feeding rates in cladocerans - another aspect of interspe-
cific competition in filter-feeding zooplankton. In: W. C.
Kerfoot (ed.), Evolution and ecology of zooplankton com-
munities. Am. Soc. Limno\. Oceanogr., Spec. Symp. Vo\.
3: 282-291.
Gons, H. J. & M. Rijkeboer, 1990. Algal growth and loss
rates in Lake Loosdrecht: first evaluation of the roles of
light and wind on a basis of steady state kinetics. Hydro-
biologia 191: 129-138.
Gons, H. J. & R. Van Keulen, 1989. De relatie tussen doorz-
icht en slib in de Loosdrechtse plassen in verband met de
zwemwaternorm. WOL-rapport 1989-4. Limnologisch In-
stituut, Nieuwersluis, 35 pp.
Gons, H. J., J. Otten & M. Rijkeboer, 1991. The significance
of wind resuspension for the predominance of filamentous
cyanobacteria in a shallow, eutrophic lake. Mem. 1st. ita\.
Idrobio\., in press.
Gulati, R. D., 1990. Zooplankton structure in the Loosdrecht
lakes in relation to trophic status and recent restoration
measures. Hydrobiologia 191: 173-188.
Gulati, R. D., K. Siewertsen & G. Postema, 1982. The zoop-
lankton: its community structure, food and feeding and role
in the ecosystem of Lake Vechten, Hydrobiologia 95: 127-
163.
Gulati, R. D., K. Siewertsen & G. Postema, 1985. Zooplank-
ton structure and grazing activities in relation to food qual-
ity and concentrations in Dutch lakes. Arch. Hydrobio\.
Beih. Ergebn. Limno\. 21: 91-102.
Gulati, R. D., L. Van Liere & K. Siewertsen, 1990. The Loos-
drecht lake system: Man's role in its creation, perturbation
and rehabilitation. In: O. Ravera (ed.), Terrestrial and
aquatic ecosystems, perturbation and recovery, pp. 593-
606. Ellis Harwood Ltd., Chichester, U.K.
Gulati, R. D., K. Siewertsen & L. Van Liere, 1991. Carbon
135
and phosphorus relationships of zooplankton and its ses-
ton food in Loosdrecht lakes. Mem. 1st. ita\. Idrobio\. (in
press).
Healey, F. P., 1978. Physiological indicators of nutrient defi-
ciency in algae. Mitt. int. verein. Limno\. 21: 34-41.
Hofstra, J. J. & L. Van Liere, 1992. The state of the environ-
ment of Loosdrecht lakes. Hydrobiologia 233: 11-20.
Holling, C. S., 1965. The functional response of predators to
prey density and its role in mimicry and population regu-
lation. Mem. Entomo\. Soc. Can. 45: 1-60.
Janse, J. H. & T. Aldenberg, 1990a. PCLOOS, a eutrophi-
cation model of the Loosdrecht lakes. WQL report
no. 1990-1. Report no. 714502001, Nat. Inst. of Public
Health and Env. Prot., Bilthoven, 92 pp.
Janse, J. H. & T. Aldenberg, 1990b. Modelling phosphorus
fluxes in the hypertrophic Loosdrecht lakes. Hydrobio\.
Bull. 24: 69-89.
Janse, J. H. & T. Aldenberg, 1991. Modelling the eutrophi-
cation of the shallow Loosdrecht Lakes. Verh. int. Ver.
Limno\. 24: 751-757.
Jorgensen, S. E., 1980. Lake management. (Water develop-
ment, supply and management, vo\. 14). Pergamon Press,
167 pp.
Keizer, P., M. N. C. P. Buysman & Th. E. Cappenberg, 1991.
Sorption and release of phosphorus in a peaty sediment.
Verh. int. Ver. Limno\. 24: 722-725.
Keizer, P. & A. J. C. Sinke, 1992. Phosphorus in the sediment
of the Loosdrecht lakes and its implications for lake res-
toration perspectives. Hydrobiologia 233: 39-50.
Kouwenhoven, P. & T. Aldenberg, 1986. A first step in mod-
elling plankton growth in the Loosdrecht lakes. Hydrobio\.
Bull. 20: 135-145.
Lammens, E. H. R. R., N. Boesewinkel-de Bruyn, H.
Hoogveld & E. Van Donk, 1992. P-load phytoplankton,
zooplankton and fish stock in Loosdrecht Lake and Tjeu-
kemeer: confounding effects of predation and food avail-
ability. Hydrobiologia 233: 87-94.
Lijklema, L., J. H. Janse, R. M. M. Roijackers & M.-L. Me-
ijer, 1988. Eutrofiering in Nederland. H 2 0 21: 462-467.
Mitchell & Gauthier Associates Inc., 1991. Advanced Con-
tinuous Simulation Language (ACSL), Reference Manual,
Edition 10.0 MGA, Concord, Mass., USA.
Moss, B., 1980. Ecology offresh waters. Blackwell Sci. Pub\.,
London, 332 pp.
OECD, 1982. Eutrophication of waters; monitoring, assess-
ment and contro\. OECD, Brusse\.
Otten, J. H., H. J. Gons & M. Rijkeboer, 1992. Dynamics of
phytoplankton detritus in a shallow, eutrophic lake (Lake
Loosdrecht, The Netherlands). Hydrobiologia 233: 61-68.
Patten, B. C. (ed.), 1975. Systems analysis and simulation in
ecology, vo\. III. Academic Press, New York.
Reckhow, K. H. & S. C. Chapra, 1983. Engineering ap-
proaches for lake management. Vo\. 1, Data analysis and
empirical modelling. Butterworth Publishers.
Riegman, R., 1985. Phosphate-phytoplankton interactions.
Ph.D. Thesis, University of Amsterdam.
136
Rijkeboer, M. & H. J. Gons, 1990. Light-limited algal growth
in Lake Loosdrecht: steady state studies in laboratory scale
enclosures. Hydrobiologia 191: 241-248.
Ryding, S. O. & C. Forsberg, 1977. Sediments as a nutrient
source in shallow polluted lakes. In: H. Golterman (ed.),
Interaction between sediments and fresh water. Dr W. Junk,
The Hague: 227-235.
Sas, H., 1989. -P!? Lake restoration by reduction of nutrient
loading: expectations, experiences, extrapolations. Aca-
demia-Verlag Richarz, St. Augustin, 497 pp.
Scheffer, M., 1989. Alternative stable states in eutrophic shal-
low freshwater systems: a minimal model. Hydrobiol. Bull.
23: 73-84.
Sinke, A. J. c., A. A. Cornelese, P. Keizer, O. F. R. Van
Tongeren & Th. E. Cappenberg, 1990. Mineralization, pore
water chemistry and phosphorus release from peaty sedi-
ments in the eutrophic Loosdrecht lakes, The Netherlands.
Freshwat. BioI. 23: 587-599.
Straskraba, M. & A. Gnauck, 1985. Freshwater ecosystems,
modelling and simulation. (Developments in environmental
modelling, 8.) Elsevier, Amsterdam/VEB Gustav Fischer
Verlag, Jena, 309 pp.
Van Liere, L. & J. H. Janse, 1992. Restoration and resilience
to recovery of the Lake Loosdrecht ecosystem in relation
to its phosphorus flow. Hydrobiologia 233: 95-104.
Van Liere, L., R. D. Gulati, F. G. Wortelboer & E. H. R. R.
Lammens, 1990. Phosphorus dynamics following restora-
tion measures in the Loosdrecht lakes (The Netherlands).
Hydrobiologia 191: 87-95.
Van Liere, L., S. Parma, L. R. Mur, P. Leentvaar & G. B.
Engelen, 1984. Loosdrecht lakes Restoration Project, an
introduction. Verh. int. Ver. Limnol. 22: 829-834.
Hydrobiologia 233: 137-159, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 137
© 1992 Kluwer Academic Publishers.

Remote sensing as a tool for assessing water quality in Loosdrecht lakes

A. G. Dekker,l T. J. Malthus,2 M. M. Wijnen 1 & E. Seyhan 1

1 Section Hydrology, Institute of Earth Sciences, Free University, De Boelelaan 1085, 1081 HV

Amsterdam, The Netherlands; 2 Dept Physiology & Environmental Sc., School of Agric., Univ. of
Nottingham, Sutton Bonington. Loughborough LE12 5RD, u.K.; current address: Sch. Appl. Sc.,
Wolverhampton Polytechnic, Wulfrunastr., Wolverhampton, WV1 1SB, u.K.

Key words: remote sensing, water quality, optical properties, spectral analysis, spectrometry

Abstract

The underwater light field in 7 lakes in the Loosdrecht lake area was measured in situ. Subsurface
upwelling irradiance and irradiance reflectance, together with estimations of scattering and laboratory
measurements of absorption by aquatic humus and particulate matter, enabled an analysis of the spectral
signature of these waters. Aircraft imaging spectrometer measurements of upwelling radiance at 1 km
altitude were used to simulate the PMI Chlorophyll # 1, the CAESAR Inland Water Mode spectral
bandsets and the Thematic Mapper bands 1 to 4. This made it possible to compare the effects of spectral
band width and selection on the estimation of water quality parameters. Correlations increased to
r> 0.94, at a significance level of 1% for the simulated C-IWM data with the 6 water quality parame-
ters. Images of the PMI Chlorophyll # 1 and of the TM were analysed and found to be in accordance
with the statistical modelling results.
A significant increase in correlation of remote sensing data with water quality parameters can be
achieved through the selective use of 10 to 20 nm wide bands in the spectral range of 500 to 720 nm
in these eutrophic waters. Sum of chlorophyll a and phaeopigments, seston dry weight, Secchi disc
transparency, and coefficients for vertical attenuation of light, absorption and scattering can be estimated
accurately. TM image data for water quality assessment is of limited use due to the relatively low spectral
and radiometric resolution. However, the revisit capability and relatively low price per area are positive
aspects of these satellite images.

Abbreviations:

CAESAR = CCD Airborne Experimental Scanner for Applications in Remote sensing

C-IWM = CAESAR Inland Water Mode
CCD = charge coupled device
EOS-A = Earth Observation System Platform A
PAR = photosynthetically active radiation from 400-700 nm.
PMI = Programmable Multispectral Imager
RSLL = Remote Sensing Loosdrecht Lakes Project
SPOT = Systeme Pour l'Observation de la Terre
138

SPOT-HRV = Sensor on board of the SPOT satellite

TM = Thematic Mapper instrument aboard the Landsat 5 satellite

Symbols
a = absorption coefficient for PAR range (m - I)
a(ah) = spectral absorption by aquatic humus (m - 1)
a(p) = spectral absorption by particulate matter (m - 1)
a(total) = spectral absorption of all components (m - I)
a(w) = spectral absorption by water (m - 1)
b = scattering coefficient for PAR range (m - I)
DN = digital number
DN-step = digital number step size
DN-step % = digital number step size as a percentage of the mean
DW = seston dry weight (mg 1- 1)
Eu = upwelling irradiance (W m - 2 nm - 1)
Eu s = subsurface upwelling irradiance (W m - 2 nm - 1)
Kd = vertical attenuation coefficient for Ed PAR (m - 1)
Ku = vertical attenuation coefficient for Eu PAR (m - 1)
Lair = airborne measured upwelling radiance (W m - 2 nm - 1 sr - 1 )
Pigm = sum of chlorophyll-a and phaeopigments (Jig 1- 1 )
Ptot = total phosphorus (Jig 1- 1 )
r = correlation coefficient
R(O) = subsurface irradiance reflectance
SD = Secchi depth or Secchi disc transparency (m)
(?)% = coefficient of determination
(sy. ) = standard error of estimate of the dependent variable
(sy.)% = standard error of estimate expressed as a percentage of the mean value of the dependent
variable
y = mean value of the dependent variable

Introduction water quality parameters, using remote sensing

This paper gives a summary of the results of a
study carried out by the Remote Sensing Loos-
drecht Lakes project (RSLL). The initial objec-
tives of the RSLL were to study temporal and
spatial variations in water quality (biological,
chemical and physical) of the Loosdrecht lakes
system through analysis of remote sensing images
(Dekker & Seyhan, 1988). In this context water
quality is defined as the changes in type or con-
centration of water constituents.
In the first year (1987) of the study it became
clear that it was necessary to increase knowledge
of analytical and statistical relationships between
aspects of the underwater light climate and the
data. The main theme of this study, as it pro-
gressed, developed into an analysis of the spectral
signature of the lake waters, through:
- theoretical and empirical analysis of the un-
derwater light climate;
- development of extraction methods for in-
formation derived from remotely sensed
spectral information;
- the study of the relationship between sub-
surface upwelling irradiance data and
airborne upwelling radiance data and their
subsequent relation with water quality pa-
rameters.
In this paper four aspects of remote sensing of
water quality of inland waters are treated:

(a) analysis of the composing factors of the spec-
tral signature of inland waters, (b) the manifesta-
tion of this spectral signature in the air and from
space, (c) the selective measurement of the spec-
tral signature through the width and positioning
of the spectral bands of remote sensing instru-
ments, and (d) the information extraction of water
quality information from these spectral bands.
The paper ends with a discussion on the use of,
and developments in, remote sensing for water
quality management and research.
The validity of the results presented here for
other inland waters is illustrated by the following
literature-based collection of similar spectra.
Similar subsurface irradiance spectra as encoun-
tered in the Loosdrecht lakes are reported by
Miyazaki et al. (1987) for the Lake Kasumigaura
in Japan and by Vos et al. (1976) for Frisian lakes
in The Netherlands. Similar subsurface reflec-
tance spectra for eutrophic water bodies are those
of Davies-Colley et al. (1988) for New Zealand
lakes and Seyhan et al. (1974) for ditch water in
The Netherlands. Normalised spectra with
shapes similar to those encountered in this study
are given by Gitelson et al. (1986) for Lake Bala-
ton in Hungary and by Gitelson & Nikanorov
(1987) for the Don and Seversky Donetz Rivers
in the Soviet Union. However, Miyazaki et al.
(1987), Vosetal. (1976), Seyhanetal. (1974) and
Gitelson et al. (1986) and Gitelson & Nikanorov
(1987) do not clarify the spectral absorption and
scattering features of these spectra.
The airborne systems were flown over the
Loosdrecht lakes (including The Lake Wijde Blik)
in 1987 and 1988 (the 1987 data was unsuitable
for further research as explained in Dekker et al.
(1990c)). The northern Vecht lakes were incor-
porated into the research in 1989, therefore only
satellite remote sensing data are available of the
northern Vecht lakes.
The lakes
General description
The Loosdrecht lakes and the northern Vecht
lakes lie in the central west part of The Nether-
139
lands (Fig. 1). The Loosdrecht lakes are a system
of lakes, separated by dikes and ranging in water
quality from a clear water water-storage basin
(Secchi depth up to 500 cm) water to very eutro-
phic, turbid lakes (Secchi depth 20-60 cm: Ta-
ble 1). Some lakes are interconnected through
narrow, shallow waterways. The lakes are impor-
tant as a nature reserve, for recreational purposes
and as reservoirs for drinking water purposes.
For a detailed description of the Loosdrecht
lakes and current restoration measures see Gulati
et al. (1991). The northern Vecht lakes are not all
interconnected or adjacent. Lakes Spiegel and
Wijde Blik are relatively deep lakes created by the
extraction of sand and gravel. Lake Wijde Gat is
a shallow, very eutrophic lake in open connection
with the Hilversums Canal which is in turn con-
nected to Lake Wijde Blik. These deep lakes dif-
fer in water quality from the other Vecht lakes in
algal composition and are not prone to wind-
induced resuspension of bottom material. Lakes
Hollands and Stichts Ankeveen are shallow lakes
created by peat excavation. Lake Hollands
Ankeveen is eutrophic and turbid with high algae
concentration. Water quality in Lake Stichts
Ankeveen is mesotrophic due to water replenish-
ment by groundwater. Lake Naardermeer is a
natural lake, consisting of several smaller water
bodies, lakes Grote Meer, Veertigmorgen and
Bovenste Blik. This system is one of the first na-
ture reserves in The Netherlands. In contrast to
the lakes Grote Meer and Veertigmorgen, Lake
Bovenste Blik is highly eutrophic.
The underwater light climate in the Loosdrecht
lakes
In a natural aquatic environment both direct solar
and diffuse skylight enter the water. Within the
water this light is modified by the absorption and
scattering processes of the water, algae, detritus
and aquatic humus. Subsurface upwelling radi-
ance is a function of spectral absorption and scat-
tering properties of these parameters. This sub-
surface upwelling radiance is the desired
information in a remotely sensed signal of a water
140

Fig. 1. The Loosdrecht lakes and northern Vecht lakes area. Based on a Thematic Mapper band 5 image.
 141

Table 1. Morphometry and average yearly primary water quality indicators of the Vecht lakes for 1988 (from north to south); SD
= Secchi Depth or Secchi disc transparency (m); DW = seston dry weight (mg 1- I); Pigm = sum of chlorophyll a and
phaeopigments (Jlg 1- I); (From Dekker et al., 1990c).

Lake Average Surface Volume Trophic SD DW Pigm

Depth Area Status
(m) (10 6m 2 ) (l06 m 3) (m) (mg/I) (Jlg/I)

NORTHERN VECHT LAKES

Naardermeer
-Groote Meer 1.00 1.16 1.2 000 0.70 12 45
-Bovenste Blik 0.65 0.29 0.2 00000 0.45 25 162
-Spookgat 1.00 0.07 0.1 00 0.70 11 54
Hollands Ankeveen 1.20 0.97 1.5 00000 0.45 19 114
Stichts Ankeveen 1.30 1.36 2.0 000 0.85 7 24
Spiegel 40.00 2.70 40.0 00 2.70 3 10
Wijde Gat 1.15 1.11 1.7 000000 0.35 28 177

LOOSDRECHT LAKES
Wijde Blik 20.00 2.90 20.0 0000 1.70 3 30
West Loenderveen* 1.50 0.60 0.9 00000 0.50 17 58
Water-storage* 8.00 1.23 9.8 a 2.25 3 6
East Loenderveen* 2.35 2.18 5.1 00000 0.45 23 59
Vuntus 1.45 0.88 1.3 000000 0.40 35 130
Nieuwe Polder** 2.10 0.30 0.5 000000 0.50 18 95
Loosdrecht 1.85 9.79 17.9 00000 0.35 35 125
Breukeleveen 1.45 1.79 2.6 000000 0.40 40 125

a oligotrophic
00 oligo-mesotrophic
000 mesotrophic
0000 mesotrophic-eutrophic
00000 eutrophic
000000 highly eutrophic
* = measurements in April 1988; ** = yearly average 1987.

body. Information concerning the constituents of
the water may be extracted from the remotely
sensed signal. This subsurface upwelling radiance
and its manifestation above the water surface are
often referred to as the spectral signature.
Ground-based spectral data collection in-
cluded spectrophotometric, spectroradiometric
and quantum irradiance measurements (over the
PAR range), which were used to determine the
apparent and the inherent optical properties. In-
herent optical properties are independent of the
ambient light field, i.e. changes in the angular dis-
tribution of radiant flux, such as the absorption
coefficient, a (m - I), the scattering coefficient b
(m - I), and the volume scattering function /3(0),
which describes the angular distribution of scat-
tered flux resulting from the primary scattering
process (Jerlov, 1976; Kirk, 1981a&b; Kirk,
1983). Direct measurement of these properties in
the field is not yet feasible. Apparent optical prop-
erties jointly depend on the inherent optical prop-
erties and on the ambient light field. Examples of
apparent optical properties are the attenuation
coefficients for upwelling (Ku) and downwelling
irradiance (Kd), the Secchi disc transparency and
the radiance reflectance. These apparent optical
properties are measurable in the field (See Dekker
et aI., 1990c and Malthus & Dekker (1990) for
further details).
142

Spectrophotometric measurements The spectral absorption properties of dissolved

The best values for absorption and scattering
spectra of pure water from 300 to 800 nm were
estimated by Buiteveld & Donze (pers. comm.)
based on data from: (1) Boivin (1986) for 400 to
436 nm, (2) Smith & Baker (1981) from 436 to
800 nm. Pure water absorbs light only very weakly
in the blue and green regions of the spectrum
(Fig. 2a). Absorption increases above 550 nm and
is significant in the red region. Scattering by water
is inversely proportional to wavelength, which
combined with low absorption, causes the blue
colour of pure water. In these lakes the molecu-
lar and Raman scattering effects are negligible
compared with the scattering caused by particu-
late matter.
aquatic humus (Fig. 2b) were determined using
the method described in Malthus and Dekker
(1990). Absorption of light by aquatic humus de-
clines exponentially with increasing wavelength
according to Davies et ai. (1987), although
Buiteveld et al. (1987) found significant deviations
from an exponential function. Data of Malthus
(unpub!.) indicate that aquatic humus absorption
in these lakes may be described by an exponen-
tial function provided a correction for residual
scattering has been made. An increase in absorp-
tion covaries with an increase in lake trophic sta-
tus. Aquatic humus removes blue light in the top
few centimetres of the water column, imparting a
yellow colour to the water. It is assumed that
aquatic humus does not scatter light.

Absorption Spectra of Aquatic Humus

Absorption Spectrum of Pure Water

""'oo
- Water-storage BasIn
+ WijdeBlik
3-
Eas! Loende<voen
• Loosdrecht

---- Yuntus
!-
2-

~
1-

o----===-=---~~~~~~-~~- _ a _ _ - - -
_ _ a _ _ ~ - - - ~
~ ~

wavelength (om) wavelength (om)

Absorption Spectra of Particulate Matter Absorption Spectra of all Components

12'1~~~~~~~~~~~~~~~~-,
I ""'oo
- Wal8l'-etorage Ba8in
10 ~
-+- WijdeBlik
East Loenderveen
I·· Loosdrec:hl
- - VunlUS
- - Bfeukeleveen
- • Pure Water·

o.~~~~~~~~~~~-~~
_ _ a _ _ _ _ ~ ~ ~
o_L~~:Ea:E~~~~:-::::==----~--------'
_ a _ _ ~_ _ ~ ~

wavelength (om) wavelength (om)

Fig. 2. Spectral absorption curves of (a) pure water (pers. comm. Buiteveld & Donze: see text); (b) aquatic humus; (c) suspended
particulate matter; (d) all constituents. (From Dekker et at., 1990c; note different scale Y-axis!)

The particles in the water consist of phyto-
plankton, dead organic matter and inorganic mat-
ter. Scattering by algae which are semitranspar-
ent and contain different materials within its cell
structure is complex. Cyanobacteria contain gas
vacuoles that have large breaking indices with the
surrounding cytoplasm causing anomalous scat-
tering at high angles (Donze et ai., 1987). Scat-
tering by algae is roughly inversely related to
wavelength. Scattering spectra of algae are given
in Prieur and Sathyenadrath (1981) and Donze
et ai. (1987).
Figure 2c shows typical particulate absorption
spectra for each lake from which several features
are apparent: high light absorption at short wave-
lengths due to absorption by detritus and by algal
pigments. A chlorophyll a absorption peak at
445 nm and an absorption shoulder at 485 nm,
probably a f3-carotene absorption feature, are ev-
ident. From 500 nm an exponential decrease in
absorption occurs. A local maximum in absorp-
tion occurs at 630 to 640 nm. This is probably due
to absorption by phycocyanin. At 675-680 nm
the in vivo chlorophyll a absorption band is clear.
From 680 nm onwards absorption by particulate
matter decreases.
The absorption coefficient of the medium as a
whole, at a given wavelength, is equal to the sum
of the individual absorption coefficients of the
Subsurface Upwelling Irradiance
Loosdrecht lakes April 1988
0.10----------------~
~
.~
- w_·_.. . ,:'
loenderveen ..:::.::::..
- West
E 0.06 - - - VUl'lllJI .:.' ·I~.
~
0.04 _ --:- WIjde Blik
--- - -
0.02 -
--
0.00 •. !iii~~===:::::::::::==::~~~~~~~
~ ~ ~ wavelength (nm)
Fig. 3. a: The subsurface upwelling irradiance (Eu s ) measurements of? lakes (see also Tables 1 and 2). (From Dekker et al., 1990c).
b: The R(O) (ratio of subsurface upwelling irradiance to subsurface downwelling irradiance) measurements of 7 lakes (see also
Tables 1 and 2). (From Dekkeretal., 1991a).
143
components present (Kirk, 1983). Assuming a
linear relationship between concentration and ab-
sorption, the absorption coefficient due to any
one component is proportional to the concentra-
tion of that component (Beer's Law). Therefore
a(total) = a(w) + a(ah) + a(p),
where a(w) is the absorption by pure water, a(ah)
by aquatic humus and a(p) by suspended partic-
ulate matter. By summing the spectral absorption
curves of water, aquatic humus and particulate
matter, the total spectral absorption curve is de-
rived (Fig. 2d). At short wavelengths absorption
by aquatic humus in the eutrophic lakes is a fac-
tor 2 to 4 larger than by the particulate matter.
The chlorophyll a absorption peak at 445 nm,
clearly visible in Fig. 2c is almost undistinguish-
able in Fig. 2d due to the aquatic humus absorp-
tion. From 710 nm onwards the water absorption
dominates the absorption spectrum. The pigment
absorption peaks at 630-640 nm and 675-
680 nm remain clearly visible in the graph.
Spectroradiometric measurements
The in situ downwelling and upwelling irradiance
above and under the water surface was measured
Subsurface Irradiance Reflectance R(O)
Loosdrecht lakes April 1988
0.10
0.08 -
..
~ 0.06
"e
~
0.04
:f
'"
. I
002 "
144

using an underwater spectroradiometer at 25 is supported by the local minimum in absorption,

sampling points in April 1988. These measure- visible at 700-705 nm in Fig. 2d.
ments reflect the combined spectral effects of ab-
sorption, scattering and solar-induced fluores-
cence from water, aquatic humus and the Quantum irradiance measurements
particulate matter. From these scans subsurface
spectral irradiance reflectance R(O) was calcu- Quantum irradiance measurements (integrated
lated: over 400 to 700 nm) were made in situ. The model
developed by Kirk (1983), based on radiative
R(O) = Eus/Eds' transfer theory and Monte Carlo simulations, was
used to estimate the inherent optical properties
where Eu s and Eds are subsurface spectral up- for the PAR wavelength range. The model uses
welling and downwelling irradiance. Representa- apparent optical properties such as net vertical
tive R(O) and Eus spectra are shown in Fig. 3. The irradiance attenuation coefficient and the under-
following features are apparent in the spectra of water irradiance distribution. The apparent opti-
the eutrophic lakes: cal properties were determined from in situ mea-
surements of upwelling and downwelling PAR
- Low R(O) and Eus in the wavelength region
(Malthus & Dekker, 1990; Dekker et al., 1990c).
of 400 to 500 nm due to absorption by
Table 2 shows typical values of absorption, a
aquatic humus and by algal pigments (e.g.
and scattering, b estimated from in situ PAR mea-
chlorophyll a, p-carotene, zeaxanthin).
surements in the Loosdrecht lake system. Scat-
- Increased R(O) and Eus up to 600 nm.
tering is the major controlling parameter of the
Greater R(O) and Eus in this wavelength re-
underwater light field in the more eutrophic lakes.
gion is the result of low absorption by pho-
Absorption was also highest in the eutrophic
tosynthetic pigments coupled with increased
lakes, resulting not only from higher pigment con-
backscattering with increased particulate
centrations but from the increased scattering
concentrations. which increased the pathlength of photons, in-
- A local minimum in R(O) and Eus occurs at creasing their chance of absorption. a and b show
630 to 640 nm. This is probably due to ab- strong covariance in these lakes (r = 0.85, N = 111,
sorption by phycocyanin. This pigment oc-
Malthus, unpub 1. data).
curs in cyanobacteria abundant in the eutro- The spectrophotometric and spectroradiomet-
phic lakes of this system. ric measurements demonstrate that the spectral
- At 675-680 nm the in vivo chlorophyll a ab-
sorption band is clear. Concentrations of Table 2. Some representative values of lake optical conditon
chlorophyll a may reach 200 J.1g 1- I in the parameters: vertical downwelling attenuation, Kd; vertical up-
most eutrophic of these lakes. welling attenuation, Ku; absorption coefficient, a and scatter-
- At 700-720 nm a peak in R(O) and Eus is ing coefficient, b calculated from PAR measurements in the
visible. Vos et al. (1986) studied the nature Loosdrecht lakes area
of this reflectance peak and gave several ex- Ku a b
Lakes Kd
planations for its presence: (m~ 1) (m -I) (m- ') (m -I)
(1) fluorescence of chlorophyll pigments;
(2) anomalous scattering caused by the ab- Water-storage 0.7 0.6 0.52 1.32
Wijde Blik 0.8 0.7 0.54 1.35
sorption peak at 675-680 nm;
East Loenderveen 3.4 2.9 2.02 12.20
(3) a minimum in the combined absorption 3.2 2.27 11.35
West Loenderveen 3.6
curves of algae and water. Loosdrecht 4.4 4.0 2.54 16.13
Vuntus 4.2 3.8 2.36 17.68
Vos et al. (1986) concluded that option (3) is 3.07 19.52
Breukeleveen 5.3 4.8
the most likely cause of this peak. This conclusion

area between 500 and 720 nm contains the rele-
vant information for the determination of water
quality parameters using remotely sensed data.
Below 500 nm aquatic humus, together with pig-
ments, absorbed almost all the light penetrating
the water in these water bodies. From 710 nm
onwards, absorption by water rises rapidly, thus
reducing the amount of reflected light.
Information extraction from remote sensing data
Ideally the remotely-sensed upwelling radiance
should be deconvoluted into each of the compos-
ing components of parameter specific absorption
and scattering features. The exact relation be-
tween the desired information and the radiance
component could then be established. Subse-
quently water quality feature extraction should
take place, preferably as maps depicting the de-
sired information. Such an ideal system is not yet
operational. Intermediate information extraction
methods are described in Dekker et al. (1990c).
Indices for information extraction based on com-
binations of two or more bands are preferred be-
cause they utilise the spectral information in a
reflectance signal from water analytically. Of these
techniques, the rationing of spectral bands is the
best known and most useful (Hilton, 1984). Ra-
tios have the advantage of yielding information
about the shape of the spectral signature curve
and extraneous factors, acting equally on all
bands, are annulled (Lillesand & Kiefer, 1987). A
drawback in using band ratios is the enhance-
ment of noise patterns that are uncorrelated in the
component images (Lillesand & Kiefer, 1987).
Therefore as few spectral bands as possible
should be used in calculations. Furthermore, there
is a normalisation problem, e.g. when two equally
high or two equally low signals are rationed, they
will present the same quotients.
To maximise the information content after
spectral band rationing and minimise noise am-
plification, ratios of bands maximally 100 nm
apart were used in this study. Philpot (1991) ad-
vises a maximum spectral distance of 50 nm when
applying derivative ratio algorithms to avoid at-
mospheric effects in remote sensing. Spectral
145
band ratios and water quality parameters were
examined for correlations.
Remote sensing systems
The remote sensing systems discussed vary from
high spectral resolution to low spectral resolution
systems; i.e. an airborne imaging spectrometer to
an airborne multispectral scanner to a space-
based multispectral system (Table 3). Location of
spectral bands of sensors used in the RSLL
project between 400 and 850 nm are depicted in
Fig. 4. Aerial photography is a spectrally low res-
olution analog imaging method. The ground-
based spectrophotometric and spectroradiomet-
ric measurements and the air-borne spectrometry
measurements were chosen as the reference data
set for analysing the information extraction po-
tential of the different remote sensing instruments.
The 17 spectra from the Programmable Multi-
spectral Imager, PMI, in spectral mode taken
from an altitude of 1000 m above lake level have
been used to simulate the spectral bands of the
PMI Chlorophyll # 1 spatial mode, The CAE-
SAR Inland Water Mode and the optical bands of
the Thematic Mapper, TM. This enabled a focus
on the comparison of relationships between op-
tical water quality parameters and spectral bands
or band combinations with improved spectral
band positioning and decreased spectral band
widths. Because the simulated spectral data orig-
inates from one data set, differences in measure-
ments caused by radiometry (irradiance, atmo-
spheric conditions, sensitivity), other sensor
related differences and flight conditions, are ex-
cluded. The results of this comparison are given
in Table 4.
The Programmable Multispectral Imager (PMI)
The PMI is an airborne imaging spectrometer. As
a pushbroom imager it images a line of 1925 el-
ements with five cameras. Both the modes of op-
eration of this system were used: a spectral mode
(288 spectral channels at 1.3 nm intervals, with
2.6 nm resolution from 430 to 805 nm) and a spa-
tial mode (computer programmable selection of
146

Table 3. Main characteristics of the remote sensing instruments used in the RSLL research.

Altitude Nominal pixel Nr. of bands Spectral resolution Dig. Number Swath
(km) size (m) vis. & NIR (nm) Range (max) (km)

Satellite
Sensors:
SPOT-HRV 832 20 1 3 80-110 256 60
TM 705 30 4 60-120 256 180

Airborne
Sensors 2 : - spatial mode at 6 km flight altitude -

PMI 0.5-10 4 8 2.6-300 + 4092/65536 3 8

C-SM 4 0.5-6 4 9 20-30 4000 2.6
C-IWM 5 0.5-6 4 9 10-30 4000 2.6

- spectral mode at 6 km flight altitude -

PMI 0.5-10 20 288 2.6 4092/65536 3

1 SPOT-HRV has a panchromatic resolution of 10 m

2 The geometric performances are dependent on flight speed and/or integration time and on flight altitude; the spectral range of
PMI is 430-805 nm; and CAESAR 400-1050 nm.
3 12 bit summation to 16 bit data
4 C-SM = CAESAR Sea Mode
5 C-IWM = CAESAR inland waler mode

up to 8 wider spectral bands for high resolution mISSIOn was flown over the Loosdrecht lakes
spatial imaging). These modes are switchable in using the PMI. Several tracks were flown in the
mid-flight. On 25 April 1988 a remote sensing spectral and in the spatial mode (Dekker et al.,
1991).

Table 4. Best regression equations and correlations between

the Lair simulated spectral bands (TM) and spectral band Results
ratios and water quality parameters. (TM = Thematic Map-
per; PMI Chi # 1 and C-IWM = CAESAR Inland Waler Only two out of five cameras produced data with-
Mode; Max. Sig. = maximum level of significance at which out radiometric distortions. A post-flight unifor-
correlation coefficient r> 0.80; (6/7) = ratio of spectral band 6 mity correction was applied by the company op-
to spectral band 7); see Fig. 4 for spectral band location.
erating the PMI, which had the effect of removing
TM PMI ChI #1 C-IWM changing atmospheric pathlength radiance with
Band Band Ratio Band Ratio changing angle of view across a swath. Despite
these technical problems the results were prom-
Max. Sign: 10% 5% 1% ising.
(N = 17) r r

SD (-) (6/7) 0.91 (2/5) 0.96 Spectral mode data

DW (3) 0.92 (6/7) - 0.94 (4/5) 0.95
Kd (3) 0.87 (6/7) -0.91 (4/5) 0.96 Figure 5 shows upwelling radiance spectra, from
Pigm (2) 0.86 (6/7) -0.96 (4/5) 0.98 the PMI from an altitude of 1000 m above lake
a (3) 0.83 (6/7) -0.89 (4/5) 0.95 level (For the same lakes as in Tables 1 and 2 and
b (-) (6/7) - 0.87 (3/5) 0.94
Figs 2 and 3). At wavelengths exceeding 500 nm
 147

~a1 Bands 400 - 850 nm

8ensOls Used in RSLL

350 400 500 550 600 650 700 750 800 850
wavelength (nm)

Fig. 4. Location of spectral bands of sensors used in the RSLL project between 400 and 850 nm. A subsurface reflectance spectrum
(dimensionless) of Eastern Lake Loenderveen is superposed as an illustration of the location of spectral features. (Codes to spectral
bandsets: SPOT = SPOT-HRV; TM = Landsat-TM; P-Chl # 1 = PMI Chlorophyll # I; C-SM = CAESAR Sea Mode; C-
IWM = CAESAR Inland Water Mode).

the same spectral characteristics were visible as in difference between these and the subsurface spec-
the subsurface upwelling irradiance spectra. tra is surface reflection and atmospheric compo-
Apart from differences in instrumentation, the nents in the airborne-measured spectra. The drop
in the measured radiance values below 450 nm in
the PMI data indicates a low radiometric accu-
Upwelling Radiance at 1000 m. A1t~ude
racy in this region (Borstad, 1985).
25 April 1988 - By PMI in Spectral Mode Because the spatial mode overflights covered
0.020 - - - - - - - - - - - - - - - - - - - - - , only 6 sample points (see further), 17 data points
in the different lakes were sampled from the spec-
0.015 -
tral mode PMI data. From these spectra inte-
-;':;.
.~
grated upwelling radiance values in the 8 spatial
"e mode bandsettings were calculated (the spatial
0.010 -
~ mode bands were simulated to obtain a signifi-
J cant data set). A regression analysis was carried
0.005 - out between ratios of these bands and the water

-- - --
quality parameters (only maximum correlation
o.()()()'---------~--~---~-
coefficients r are discussed here and given in Ta-
~ ~ ~
wavelength (nm)
~ ~ ble 4). Seston dry weight, DW, sum of chloro-
Fig. 5. Upwelling radiance measured at 1000 m above lake phyll a and phaeopigments, Pigm and the vertical
level (Lair) by the PMI in spectral mode. (From Dekker et al., attenuation coefficient, Kd, were highly correlated
1991) with the ratio of the Chlorophyll # 1 band 6 to
148

band 7. Secchi disc transparency, SD, was also
highly correlated, but the independent SD mea-
surements were distributed in two clusters: one
cluster of three measurements at 200 to 225 cm
and one at 30 to 50 cm transparency. No inter-
mediate values were present. These results sug-
gested that concentrations of Pigm, Kd and D W
could be determined with relatively high accuracy
in an image using the P(6j7) ratio.
Spatial mode image data
The Chlorophyll # 1 bandset wavelengths are
shown in Fig. 4. Colour Plate 1 is an image cre-
ated by rationing the spectral bands 6 and 7 of the
Chlorophyll # 1 spectral bandset. Spatial resolu-
tion was 2 m, but was reduced to 6 m for image
processing purposes. No atmospheric corrections
were made. The highest and lowest colour levels
were assigned to the highest and lowest measure-
ments of Pigm, DW, Kd, a and b. Intermediate
values were linearly extrapolated between these
colour levels, which appear to represent accurate
values of water quality parameters as predicted
by the band simulation modeling.
Due to deteriorating weather conditions simul-
taneously-measured water-based data were inad-
equate to validate the modelling statistically using
the spatial mode data. Two sampling points in the
northwestern part of Lake Loosdrecht, L4 and
L7, were measured during the flight (Fig. 6).
Points ELl, EL2 and EL3 were measured in the
Eastern Lake Loenderveen and WL2 in the
water-storage basin the following day. The mea-
sured values and the values obtained from the
ratio image are compared in Table 5; the pre-
dicted data are in agreement with observed
values.
Although Lake Wijde Blik is usually a clear
lake (SD of 200 cm or more), a plume of eutro-
phic water (green, yellow and red colours) can be
seen entering the lake from The Hilversums
Canal, to the north of the lake, and curving to the
southwest. The concentrations in the plume de-
duced from the image are: D W, 31.0-34.1 mg 1- I
and Pigm: 83-92 j.1g 1- 1. Wind direction during
the remote sensing flight was from the northeast.
It was assumed lake water flowed into this canal
instead of from the canal into the lake. Water
quality parameters of a sample taken in this canal
(2 km to the east) on the flight date are: DW,
25mgl- 1; SD, 40cm and Pigm, 146j.1g1-1.
These differences in estimation may be caused by
the distance between the sampling point and the
point of inflow, but indicate the influx of highly
eutrophic water into Lake Wijde Blik.
The determination of this plume is an example

Table 5. Simultaneous in situ and remote sensing measurements. See Fig. 6. for location (sample = parameter value measured in
situ; image = parameter value range deduced form the image;? = coefficient of determination between sample value and estimated
image values; Sy = standard error of estimate for image values; from Dekker et al., 199Ia).

DW(mgl- l
) Pigm (J.lgl - I) Kd (m- J) b (m - I) a (m - J)

sample image sample image sample image sample image sample image

L4 33.0 31.0-34.0 92 83-91 4.53 4.5-4.2 13.62 5.2-16.6 2.84 2.6-2.8

L7 37.5 24.1-37.0 98 92-100 4.43 4.6-4.8 14.37 13.8-15.1 2.66 2.8-3. I
ELI 22.7 24.7-27.7 68 66-73 3.41 3.4-37 13.26 12.3-13.7 2.19 2.1-2.4
EL2 23.5 24.7-27.7 66 66-73 3.51 3.4-3.7 12.47 12.3-13.7 2.07 2.1-2.4
EL3 22.8 21.5-24.6 59 57-65 3.43 3.1-3.3 12.20 10.8-12.2 2.02 1.9-2.1
WL2 3.7 2.6-5.6 6 4-12 0.93 0.8-1.1 1.68 2.1-3.4 0.65 0.5-0.7

? 0.97 0.99 0.98 0.95 0.96

Sy 2.1 3 0.21 1.17 0.17


!I2.20·
Fig. 6. Location of Colour Plate 1 and the simultaneously
measured sampling points.
of the advantages of remote sensing for water
quality research and monitoring. The synoptical
view provided by remote sensing makes it possi-
ble to determine the presence and extent of such
a phenomenon. More important, however, is that
the analysis of the high spectral resolution data
makes it possible to determine composition and
concentration of the optical water quality param-
eters. The presence of the plume also demon-
strates the disadvantages of point sampling data
as a measure of lake water quality. Samples for
routine water quality measurements are taken by
the water management authorities in Lake Wijde
Blik in the centre of the lake. Thus such sampling
confined to one point in the lake is not represent-
149
ative of the entire lake (under these circum-
stances) and may lead to wrong assessments
about water quality in the lake.
The other lakes that can be seen in the image,
are Eastern Lake Loenderveen, Western lake
Loenderveen, the water-storage basin and a part
of Lake Loosdrecht. These lakes exhibit a more
uniform (spatial) water quality. The water-storage
basin shows gradients, however, which may
be caused by the inlet of iron-rich water from
the dephosphorisation plant in the southwestern
corner of the lake (outside the limits of the
image).
CAESAR
CAESAR is a multispectral pushbroom scanner,
meant both for land and sea applications (Pou-
wels, 1987 and Looyen & Van Swol et al., 1989).
For water applications there are two modes,
each consisting of nine spectral bands: the Sea
Mode and the Inland Water Mode (Table 3 and
Fig. 4). The 'Sea Mode' is not suitable for detect-
ing quality of inland waters, due to the spectral
band selection, which is aimed at ocean type wa-
ters. In turbid, algae-rich and humus rich waters
different spectral bands are necessary to ade-
quately measure the desired spectral information
(Dekker et al., 1990a, b&c, 1991).
The CAESAR Inland Water Mode (C-IWM)
The Inland Water Mode was developed in the
course ofthe RSLL Project (Dekker et al., 1990b;
and Table 3 and Fig. 4). This spectral bandset
enables an accurate reconstruction of the spectral
signature between 500 to 800 nm in these lakes
(Fig. 7). A comparison of the spectral features in
the section on the underwater light climate and
the positioning of the C-IWM bandset illustrates
the rationale in the selection of the bands. A de-
tailed analysis of the choice of the C-IWM band-
set can be found in Dekker et al. (1990a&b).
The spectral bands of the C-IWM were simu-
lated using the PMI upwelling radiance spectra.
150

Plal~ I. Specual ratio image of (P 6/7). See legend for water quality parameter concentrations. See Fig. 6. for location and scale
(reprinted from Dekker lit oJ.)

--
'.

The main results (Table 4) confirm that in eutro-
phic systems such as the Loosdrecht lakes esti-
mation oCin-lake parameters can best be achieved
using band ratios based on wavelengths from
500 om to 800 om, with emphasis on the area
between 600 om and 720 om. An advantage of
using longer wavelengths is that atmospheric dis
tartian of the remotely-sensed signal decreases
with increasing wavelengths. The results show
that an increase in correlation and significance
features is strongly dependent on spectral band
positioning and width.
 PlaIt 1_ (Continued) 151
Pigm (pgl- I) OW (mgl- 1) kd(m- 1)

0- 3 0 - 2.5 0.+-0:;- 0.6-2.0 0.3-0.4

"
4- 2.6 - 5.6 0.8-1.1 2.1-3.4 0.5-0.6

13- 21 5.7 - 8.8 1.2-1.5 3.5-4.9 0.7-0.9

22- 29 8.9 - 11.9 1.6-1.8 5.0-6.3 1.0-1.1

30- 38 12.0 - 15.1 1.9-2.2 6A-7.8 1.2-1.3

39- 47 15.2-18.2 2.3-2.6 7.9-9.3 1.4-1.6

48- 56 Ih.3-21.4 2.7-3.0 9.4-10.7 1.7-1.8

57- 65 21.5 - 24.6 3.1-3.3 10.8-12.2 1.9-2.0

66-73 24.7 - 27.7 3.4-3.7 12.3-13.7 2.1-2.3

74- 82 27.8 - 30.9 3.8-4.1 2..1-2.5

83- 91 31.0 - 34.0 4.2-·;U 15.2-16.6 2.6-2.7

92-100 34.1 - 37.0 4.6-4.8 16.7-18.1 2.8-3.0

101-109 37.3 - 40.3 4.9-5.2 18.2-19.5 3.1-3.3

110< < 40.4< < 5.3« 19.6< < 3.4«

----....
-
CAESAR INLAND WATER MODE

... -
~.

SateUite sensor imagery

•
~

'.
--, r-
0.012 -

=
Although airborne remote sensing using very high ~

-- _...
spectral resolution data is most suited for quan- o- A
titative assessment of water quality, relatively low J !=~ , ..•
spectral and spatial resolution satellite imagery ~

may also be useful (Dekker & Peters, 1992). Due

-=---=--::_=-~-=---=-=-~-=---=-=-~-=---=--=­
_.-
to their sufficient spatial resolution, the Thematic
Mapper (TM) on board the Landsat satellite and
SPOT-HRV sensors on board the SPOT satellite
--
.-
-
---
are suited for imaging medium sized or smaller

---
---
inland waters, such as the Loosdrecht lakes. The

---
data are easily available at a moderate acquisition
cost (see Table 3 for main characteristics of these
systems). Furthermore, continuity in data acqui-
sition is guaranteed: Landsat 6 will be launched Fig. 7. La;. spectra of 7 lakes in the Loosdrecht lakes system
soon and Landsat 7 and SPOT 3 and 4 are in the as reconstructed from simulated CAESAR Inland Waler Modt
planning phase. Here only Landsat-TM data are Lab- spectral bandset values. (From Dekker tl 0/., 1990b).
152

THEMATIC MAPPER
discussed in detail. SPOT-HRV images of the Simulated spectral band values only

Loosdrecht lakes showed results similar to the 0.020 r - - - - - - - - - - - - - - - - - ,

TM images. Results of SPOT-HRV image anal- --=b:=an--=dC-'1------=b:=an--=d--=2--, ~ band 4 ->
yses are to be found in Dekker et al. (l990c). 1

The TM images discussed here are from the 0.015

I •••••••••••• j\
~§
Thematic Mapper instrument mounted on board i +-..\
on the Landsat 5 satellite. The primary use of this "e
~ 0.010
t•
:•:•=
• •'~~~'f-,
• •••• ....,--+,.-
.....-+ .......
system was for land observation, resulting in low -- \\J1 I
radiometric sensitivity at low radiance levels, i.e. J
0.005
\.~---......,.,~
a small range of digital numbers (DN) encom-
passes the water-leaving radiances. This small
0.000 , ' - - - - - - - " - - - " - - - - - - - - ' - - - - - ' - - ' - ' - - - ' - - - - - ' - - - '
DN range leads to a low signal to noise ratio. The 400 450 500 550 600 6SO 700 750 BOO 850
wavelength (nm)
overpass time over the lake area was 11 :06:49
M.E.T. The TM images the same area every 16
I Legend to Lakes
days; according to Van Der Laan (see Noordzij I- Br.uk.l....een·

et al., 1989), this results in an average of 1 to 3 i Vuntus

! . Loosdrecl'll
entirely cloud-free images of the entire Nether- I - East Loenderveen

lands per year over a 14-year period (for Landsat - - WestLoenderveen

- Wijde~k

MSS and TM combined). For a smaller area the ! -4- Water-storage

frequency of cloud-free images increases.

Fig. 8. Spectral signature reconstruction of TM spectral
bands simulated from Laic data.

Spectral signature analysis of Thematic Mapper

bands
Seven TM spectral bands are positioned to ob-
tain as much information as possible for a large
variety of targets (Table 3). To evaluate the spec-
tral information contained within each of the op-
tical TM spectral bands, signature reconstruction
is shown in Fig. 8.
Bands 1 to 4 of the TM are in the spectral range
where light transverses the water to a sufficient
extent to expect some information concerning the
water quality (see the discussion on the under-
water light climate). The main effects detected are
the attenuation of irradiance in band 1 (absorp-
tion by aquatic humus and by chlorophyll-a), and
the increase in reflected irradiance in band 2.
Band 3 information is difficult to interpret because
of the counter-acting processes of peaks in pig-
ment absorption with adjacent peaks in reflec-
tance. Band 4 reflectance is the product of rapidly
increasing water absorption and of reflection from
suspended matter in the near-infrared. Bands 5
and 7 are located in the middle-infrared, where
absorption by water is extremely high, based on
absorption data from Palmer & Williams (1974).
The thermal infrared band 6 makes temperature
mapping possible, but was not investigated due to
its low spatial resolution of 120 x 120 m pixels.
From the above it follows that the use of combi-
nations (sum, difference over sum, ratios etc.) of
the relatively wide spectral bands of TM to in-
crease correlations with water quality parameters
produces results which have no analytical mean-
ing. As an example, the information in each of the
bands 2 and 3 is not increased by adding or ra-
tioning these bands. Therefore, only a quantita-
tive analysis was performed on single bands of
the TM data.
TM band 1 to 4 were simulated using the PMI
spectral mode upwelling radiance spectra, mak-
ing the results comparable to those of PMI Chlo-
rophyll # 1 and CAESAR Inland Water Mode.
The highest correlations (Table 4) show that the
relatively large bandwidths of the TM lead to
relatively low correlations and significance levels
for spectral bands and optical water quality fea-
tures.

Image analysis
Images taken on 16 June 1986 and on 5 July 1987
were analysed. Results for the June 16th image
were similar to those of 5 July (Dekker et al.,
1990c and Dekker & Peters, 1992). Here data of
5 July is discussed.
The Thematic Mapper image of July 5, 1987 is
a pseudonatural colour composite (Plate 2a).
Contrast was enhanced to increase image bright-
ness. Qualitative differences are evident both
within and between the lakes. The water-storage
basin, Lake Wijde Blik and Lake Spiegel, appear
as dark blue, all of them being the deepest clearest
lakes in this area. The lakes with overall highest
reflectances are the eutrophic lakes Loosdrecht,
Nieuwe Polder, Vuntus and Breukeleveen.
An advantage of pseudonatural colour com-
posite images from space is their easy and
straightforward interpretability. However, it is
difficult to extract quantitative information. Quan-
titative information extraction involves prepara-
tory image analysis prior to statistical analysis
(Dekker & Peters, 1992).
Discussion of the satellite image data results
The results of bivariate linear regression analyses
for TM band DN data and water quality param-
eters having correlation coefficients larger than
0.80 are presented (Table 6a; after Dekker & Pe-
ters, 1992). The smallest interval measurable may
be expressed as the standard error of estimate
(s.,), and is expressed as a percentage of the mean
of the value, (sr' )~{, (Table 6a). The coefficient b
in the regression equation of Y = a + b*X is equal
Table 6a. Regression equations of the TM data and optical water quality parameters on July 5, 1987 (N = 10), (from Dekker el
at..1990c)
Regression equation r (?)~o
SD = 1199.93 - 55.90*TM3 0.81
DW - 207.02 + 8.96*TM2 0.96
DW - 165,84 + 9,43*TM3 0.92
Pigl11 = - 1137.52+48.93*TM2 0,93
Pigl11 = - 913,83 + 5l.S2*TM3 0.90
153
to the smallest interval of the characteristic to be
measured between two DNs. This interval is the
DN stepsize, and is give as a percentage of the
mean: DN-step %.
The (sy') % are all greater than 23 %. Secchi-
disc depth transparency does not have a correla-
tion r> 0.80 for the linear data. Seston dry weight
with a (Sy') of 4.9 mg 1- 1, as well as sum of chlor-
ophyll-a and phaeopigments with a (Sy') of
33.81/lg1-I, may best be determined using
band 2. Total phosphorous was also highly cor-
related with band 2 and band 3 DNs, a result prob-
ably due to the intercorrelation between organic
matter in the seston and total phosphorous. Cor-
relations for D W, Pigm and Ptot with band 2 and
band 3 are positive: an increase in DN (and thus
in reflectance) is proportional to an increase in the
concentration of the parameter. This is also true
for Pigm which supposedly is correlated to ab-
sorption. In this case backscattering, caused by
the algae and the suspended matter, variables
which also tend to covary strongly, dominates the
absorption by the pigments. Secchi-disc depth
could be estimated using band 3 on 5 July 1987
with a (Sy') of 65 cm.
Correlations higher than 0.80 for band 1 as in-
dependent variable show large (Sy')%' Atmo-
spheric path radiance contributions decrease the
correlations. The high aquatic humus concentra-
tions and chlorophyll absorption levels cause a
low absolute water leaving radiance. The TM has
large sensor calibration errors in band 1 as well.
Regression equations including the nearby infra-
red band 4 as independent variable did not reach
an r>0.80.
As can be seen from the table the intervals
measurable between two DNs is often larger than
-
(S),.)~o y DN-step
66 64.6 95.50 55.9
91 29,2 16.80 53.5
85 38.6 16,80 52.2
87 33,9 99.85 48.9
81 40.8 99.85 51.5
154

Table 6b. Regression equations using the log-transformed data
for July 5, 1987. (Regression formula: Y = a*X b ; coefficients
are determined using the formula: LN Y = LN a + b LN
X)(from Dekker et al., 1990c)
Regression equation r (?)%
Ln SD
+ 42.75 - 11.95*Ln TM3 -0.82
LnDW - 49.39 + 16.07*Ln TM2 0.85
LnDW - 40.87 + 14.56*Ln TM3 0.93
Ln Pigm = - 55.37 + 18.39*Ln TM2 0.87
Ln Pigm = - 43.54 + 15.97*Ln TM3 0.91
the (Sy')' Thus, in many cases it is the low radio-
metric sensitivity of the TM which is the limiting
factor in an accurate assessment of water quality
parameters (Dekker & Peters, 1992).
There was a drawback attached to using linear
regression analysis on untransformed data: when
applying the regression equations to the full range
of digital data present in TM images, negative
values were frequently calculated. An exponential
curve of the form Y = a *xt' fitted the data well.
There is physical basis for this: (1) negative val-
ues for the parameters do not occur; (2) most of
the correlation between radiance values and the
water quality parameters is based on the scatter-
ing properties of the water constituents. An in-
crease in scattering will also increase the multiple
scattering thus giving the increase an exponential
character. Thirdly, the data from highly eutrophic
lakes varies more in time than data from olig-
otrophic or mesotrophic lakes. Thus, a log-
transformation normalises the variance across the
range of water qualities encountered.
In contrast to the results from the non-
transformed data, the log-transformed data had a
higher correlation of SD, DW and Pigm with
band 3 than with band 2 (Table 6b). The log-
transformation caused calculated dimensions of
the dependent variables based on the bivariate
68
73
linear regression results to be more realistic, es-
86 pecially at low digital numbers for the DW and
77 Pigm concentrations and at high DNs for SD
83 transparencies. At the other end of the scale,
however, the values become unrealistically high
for D Wand Pigm. Therefore the lowest and high-
est DN values in Table 7 which lead to the cal-
culation of unrealistic values are marked with a
lower than ( <) or greater than (> ) sign instead
of an estimated parameter value. In Table 7 for
DN 28 & 29, therefore, only a minimum value at
DN 28 is given for D Wand Pigm.
Interpretation of water quality map based on a co-
lour coded TM band 2 image of 5 July 1987
The lakes Nieuwe Polder, Wijde Gat and Hol-
lands Ankeveen appear to be similar in water
quality (Plate 2b). Lake Vuntus differs (up to 2
DNs higher) from lakes Loosdrecht and Breu-
keleveen. Lake Vuntus had a concentration of
cyanobacteria of 210.000 ml- I and a Pigm of
189 Jig 1- I compared with 140.000 ml- I and a
Pigm of 143 Jig 1- I in Lake Loosdrecht and
140.000 per ml with a Pigm of 162 Jig 1- I in Lake
Breukeleveen. Lake Vuntus seems to be more yel-
lowish because of higher green and high red re-

Table 7. Concentration intervals of water quality parameters on 5 July 1987 calculated for TM band 2 DN values using the log-
transformed regression equations from Table 6b. (Ref. Colour Plate 2b.)

TM2 DW(mg/l) Pigm(jJ.g/l) SD(cm) colour code

DN22 - 1- 2 3 - 7 439 -255 blue

DN22 - 2 - 4 7 - 15 255 -151 light blue
DN24 - 4 - 7 15 - 32 151 - 92 dark green
DN25 - 7 -14 32 - 66 92 - 57 light green
DN26 - 14 -26 66 -133 57 - 36 green-yellow
DN27 - 26 -47 > 133 36 - 23 yellow
DN28 - >47 > 133 23 - 15 orange
DN29 - >47 > 133 < 15 salmon

Pla't 20. TM pseudonatural colour composite Thematic
Mapper image on July 5, 1987; bands J, 2, and 3 are repre-
sented in blue, green and red respectively. (From Dekker &
Peters, 1992)
ftectance, possibly due to the high amount of cy-
anobacteria. In band 3 Lake Vuntus is similar (0
these lakes. Other exceptions are Lake Naarder-
meer and the southeastern Lake Naardermeer-
Bovenste Blik, which differ by two DNs. The lakes
Stichts Ankeveen and Hollands Ankeveen are
155
P/o112b. Water quality map based on a colour coded TM
bond 2 image of the 5 July 1987. Concentration of water
quality parameters can be determined by comparing the DN
values in the image with the cakulated values in Table 7.
(From Dekker &. Peters., 1992.)
almost invisible because their reflectance values
approach those used to mask out the land reflec-
tance. This is probably due to emerse macrophyte
coverage. The results of June 16 1986 are similar.
The algorithms developed through statistical
156
analyses are based on the assumption that a 17
by 17 pixel area is a representative sample for the
average water quality measured there (Dekker
et al., 1990c). If each lake were homogeneous (i.e.
well mixed) the concentration maps of basis of
these algorithms would show no variation at all
within the lakes. In reality gradients and patterns
are observed (Plates 1 and 2a-b). There is enough
variability among the lakes for the calculated re-
gression line to be valid. So, the hypothesis that
the optimized spectral sample is representative
for the average measured parameters is not re-
jected.
Except for the (Sy') as a limiting factor for ac-
curacy of determination, the absolute range of
DN levels is also limiting. The amount of con-
centration levels maximally discernible expressed
as DNs, with a minimum threshold of at least 10
pixels per level, is 6 for band 2 and 9 for band 3
on both dates. If a threshold of 100 pixels per DN
level is applied the maximum range would de-
crease to 5 or 6 levels. The two images discussed
here were taken around maximum solar eleva-
tions (ca. 58 degrees) on cloud-free days and re-
present optimal images that can be achieved over
water at this geographic location. It is concluded
that Landsat TM images are of limited value for
the determination of surface water quality. Sub-
optimal images will not be applicable for these
purposes.
The use of remote sensing in water quality man-
agement and research
A simultaneous image of water quality over a
large area can be obtained in minutes (by satel-
lite sensor) or hours (airborne remote sensing).
U sing airborne systems a particular water body
can be imaged consecutively on one day or,
weather permitting, on consecutive days. Surface
streamflow dynamics may be visualised and cal-
culated. In lake restoration projects remote sens-
ing can help to establish changes during the course
of the restoration phase. The technique can be
applied to measure the dispersion of suspended
material during dredging operations of polluted
water bottoms. In general, spatial variability of
optical water quality parameters can be deter-
mined. This may lead to optimalisation of sam-
pling networks. Insight is obtained into how
representative point samples are for an entire
water body. Remote sensing enables non-intru-
sive monitoring (e.g. nature preservation areas).
Remote sensing images may have a demonstra-
tive function, since such information generally has
a higher impact factor in public awareness than
figures or tables. Remote sensing data may have
an archive function: unlike water samples digital
remote sensing data can be stored for long term
and will always be available for further analysis,
e.g. new information extraction methods can be
used on existing data. This latter was the case
with the Thematic Mapper images; only after ac-
quisition of the images, was the relevant water
quality data assembled.
Developments in remote sensing of inland water
quality: general conclusions
Any research involving remote sensing of water
quality in the optical and nearby infrared wave-
lengths involves analysis of the spectral signature
of the water under observation. The spectral sig-
nature is a result of the effects that spectral scat-
tering, spectral absorption and, possibly, fluores-
cence have on the water penetrating irradiance.
These are in turn the result of the effects of the
water itself and its constituents in dissolved, col-
loidal or suspended form on incoming light. The
upwelling radiance from the water column at the
surface is a convolution of the spectral effects of
these constituents on the incoming irradiance.
Bottom reflectance is assumed to be negligible in
the following discussion. Although it is an impor-
tant topic in remote sensing of water, the tech-
niques of removal of atmospheric influences are
not discussed.
A thorough research should incorporate the
following aspects to obtain reproduceable results.
- Measurement of surface and subsurface up-
welling (ir)radiance) and of optical proper-
ties, both inherent and apparent.
- Determination of the relationship between

the optical properties and the optically active
water constituents. This should preferably
be carried out analytically or else statistical
relationships may be used, provided an an-
alytical explanation of the results is possible.
- investigation of the covariance of chemical
substances with the optically active water
parameters. The statistical relationships be-
tween remote sensing data and chemical pa-
rameters is based upon a covariance with
one of the optically active parameters; most
such relationships will remain site specific
and dependent on time of day and season.
With this knowledge the performance of a re-
mote sensing instrument can be analysed. The
spectral bands should selectively measure spec-
tral features. Information extraction will inevita-
bly involve combining spectral band information.
As few spectral bands as possible must be used,
to minimise both the required amount of data and
the increase of the noise relative to the signal, and
to maximise the physical interpretability of the
data.
For the measurement of specific spectral fea-
tures, such as photosynthetic pigment absorption,
narrow bandwidths are required. Narrow bands
cause a decrease in radiance levels. To obtain
adequate signal to noise ratios the radiometric
resolution must be high at the typically low radi-
ance levels of water bodies. A trade-off must,
therefore, be made with regards to spectral band
width.
For spectral signature analysis imaging spec-
trometers are very useful. They are the only in-
struments that can measure the spatial variability
of the spectral signature of water bodies. Thus
they can also determine the relevance of surface-
based spectral measurements (which are point
measurements) to the entire water body. How-
ever, the amount of data generated by imaging
spectrometers is extremely high, is of an abstract
nature, and needs expert analysis methods to ex-
tract the necessary information.
Multispectral scanners for water quality assess-
ment should make use of specific spectral bands,
positioned to measure the expected spectral sig-
natures efficiently. They are relatively simple
157
compared with imaging spectrometers and are
presently at a higher operational level. The
amount of data will depend on the number of
spectral bands and the radiometric and spatial
resolution features, but will generally be a factor
10 to 20 lower than for imaging spectrometers. If
the spectral bands are positioned correctly, the
information extraction is relatively easy provided
the appropriate spectral signatures are known.
Satellite sensor systems, such as the Thematic
Mapper and SPOT-HRV, lack sufficient spectral
and radiometric resolution to accurately measure
water quality features. By virtue of their opera-
tional status and comparatively low costs per
square km imaged, they are feasible options for
the purposes of large-scale monitoring at low
water quality feature resolution. The near-infrared
channels of both satellite systems have been used
to monitor algal surface layers (not encountered
in this research) (Buiteveld et al., 1987).
The range and resolution in wavelengths on the
future Medium Resolution Imaging Spectrometer
(MERIS) planned on the European Space
Agency Polar Platform to be launched in 1997,
the Moderate Resolution Imaging Spectrometer
(MODIS) planned on the Earth Observation
System (EOS-A) platform to be launched in 1998
and the High Resolution Imaging Spectrometer
(HIRIS) planned on the second EOS-A platform
to be launched after 1998, instruments will make
it possible to monitor water quality accurately
from space in both a qualitative and quantitative
manner. The CAESAR Inland Water Mode illus-
trates that nine spectral bands are sufficient to
determine several water quality related optical pa-
rameters to a high degree of accuracy. This illus-
trates the possibility of data reduction when sys-
tems such as MERIS, MODIS and HIRIS
become operational. In this case a reduction of
the amount of spectral bands measured over the
range of 400 to 850 nm from 45 (in case of com-
plete coverage at 10 nm resolution) down to 9
spectral bands without losing relevant informa-
tion seems possible.
Until the next generation of earth observation
satellites is operational towards the end of the
nineties, airborne remote sensing is the main area
158
of research for water quality assessment of inland
waters. The airborne systems will pave the way to
optimise the use of these satellite based imaging
spectrometers as soon as they are operational.
The positive aspects of satellite remote sensing
will then be combined with the advantages of the
current airborne systems.
Specific recommendations to facilitate this
transition are:
(1) to perform spectral signature analysis of all
water bodies under consideration for sur-
veillance, monitoring, or analysis from the
air or from space;
(2) to develop reliable methods for information
extraction from the measured spectral sig-
natures;
(3) to continue research making use of imaging
spectrometers for thorough fundamental
analysis of the spectral signature as it is
measured by remote sensing;
(4) to develop, test and operationalise multi-
spectral scanners for water quality analy-
SIS;
(5) to make use of the results of airborne cam-
paigns to prepare the water quality com-
munity for this novel manner of studying,
monitoring and understanding water qual-
ity, especially in the spectral and spatial
domain.
Acknowledgements
Many institutions contributed to the RSLL
project. In alphabetical order they are: Labora-
tory for Microbiology at the University of Am-
sterdam; The Limnological Institute; Moniteq
Ltd. of Canada; Municipal Waterworks of Am-
sterdam (GWA); National Aerospace Laborato-
ries (NLR); Programme Bureau of the BCRS;
Free University of Amsterdam; Water Pollution
Control Authority Amstel- en Gooiland (ZAG).
The author is especially indebted to the follow-
ing persons for being allowed to disrupt their own
research or other activities, whenever it was
thought to be necessary: Loes Breebaart, drs.
Tineke Burger-Wiersma, drs. Jan Jaap Buysse,
dr. Herman Gons, drs. Jasper Griffioen, drs. Ben
Kal, Andre Lameris, dr. Louis van Liere, Ir. Wim
J. Looyen, dr. Frank Muller Karger, Ir. Hein
Noorbergen, Marchien A. van Oostende, drs.
Steef Peters and drs. Pierre Verstraelen.
Many more people contributed to this research,
especially the personnel of the Institute of Earth
Sciences at the Vrije Universiteit and of the Lim-
nological Institute who carried out the field-based
measurements and laboratory analyses.
References
Boivin, L. P., W. F. Davidson, R. S. Storey, D. Sinclair &
E. D. Earle, 1986. Determination of the attenuation coef-
ficient of visible and ultraviolet radiation in heavy water.
Appl. Optics 25: 877-882.
Borstad, G. A., H. R. Edel, J. F. R. Gower & A. B. Hollinger,
1985. Analysis of test and flight data from the Fluorescence
Line Imager. Can. Spec. Publ. J. Fish. aquat. Sci. 83: 38 pp.
Buiteveld, H., F. De Jong, R. Spanhoff & M. Donze, 1987.
Shape and variability of the absorption spectrum of aquatic
humus. Symp. on Rem. Sens. Resource Developm. & Env.
Managem., Enschede, the Netherlands, August 1986, IAHI
Netherl. Soc. for Rem. Sens.: 703-705.
Davies-Colley, R. J., W. N. Vant & R. J. Wilcock, 1988. Lake
water colour: comparison of direct observations with un-
derwater spectral irradiance. Wat. Res. Bull. 24: 11-18.
Dekker, A. G. & S. W. M. Peters, 1992. The use of the The-
matic Mapper for the analysis of eutrophic lakes: A case
study in The Netherlands. Int. J. Rem. Sens.
Dekker, A. G. & E. Seyhan, 1988. The Remote Sensing Loos-
drecht Lakes Project. Int. J. Rem. Sens. 10 & II: 1761-
1773.
Dekker, A. G., T. J. Malthus & E. Seyhan, 1990a. Improving
quantitative analysis of inland water quality using high
spectral resolution imaging and non-imaging data. Proc.
IGARSS'90 Symp., Washington, May 20-24, 1990: 117-
120.
Dekker, A. G., T. J. Malthus & E. Seyhan, 1990b. An inland
water quality bandset for the CAESAR system based on
spectral signature analysis. Proc. International Symposium
Remote Sensing and Water, Enschede, August 1990, IAHI
Neth. Soc. for Rem. Sens., The Netherlands: 597-606.
Dekker, A. G., T. J. Malthus & E. Seyhan, 1991. Quantita-
tive modelling of inland water quality for high resolution
MSS-systems. IEEE Trans. on Geosc. and Rem. Sens. 29:
89-95.
Dekker, A. G., T. J. Malthus, S. W. M. Peters & E. Seyhan,
1990c. The Remote Sensing Loosdrecht Lakes Project.
BCRS Report 90-28. Dutch Remote Sensing Board, Delft,
The Netherlands: pp. 99.

Donze, M., G. B. J. Dubelaar & J. W. M. Visser, 1987.
Anomalous behaviour of forward and perpendicular light-
scattering of a cyanobacteria due to intracellular gas vac-
uoles. BCRS report 87-08, Dutch Remote Sensing Board,
Delft, The Netherlands: pp. 43.
Gitel'son, A. A. & A. M. Nikanorov, 1988. Remote monitor-
ing of ecological condition of aquatic ecosystems. Proc.
USA-USSR Symposium, Athens, Georgia, Oct 19-21,
1987. on: Fate and Effects of Pollutants on Aquatic Or-
ganisms and Ecosystems: 166-182.
Gitel'son, A. A., A. M. Nikanorov, G. Y. Szabo & F. Sz-
ilagyi, 1986. Etude de la qualite des eaux de surface par
teledetection. Proc. Budapest Symp., July 1986, on: Mon-
itoring to Detect Changes in Water Quality Series: IAHS
Pub!. no. 157: 111-121.
Gulati, R. D., L. Van Liere & K. Siewertsen, 1991. The Loos-
drecht lake system: Man's role in its creation, perturbation
and rehabilitation. Terrestrial and Aquatic ecosystems:
Perturbation and recovery. Ellis Horwood Ltd.: 593-606.
Hilton, J., 1984. Airborne remote sensing for freshwater and
estuarine monitoring. Wat. Res. 18: 1195-1223.
JerlOV, N. G., 1976. Marine optics. Elsevier, Amsterdam, The
Netherlands.
Kirk, J. T. 0., 1981a. Monte Carlo Study of the nature of the
underwater light field in, and the relationships between op-
tical properties of, turbid yellow waters. Aust. J. Mar.
Freshwat. Res. 32: 517-532.
Kirk, J. T. 0., 1981b. Estimation of the scattering coefficient
of natural waters using underwater irradiance measure-
ments. Aust. J. Mar. Freshw. Res. 32: 533-539.
Kirk, J. T. 0., 1983. Light and photosynthesis in aquatic ec-
osystems. CSIRO, Canberra, Australia. Cambridge Uni-
versity Press: pp. 401.
Lillesand, T. M. & R. W. Kiefer, 1987. Remote sensing and
image interpretation. John Wiley & Sons, New York:
pp. 721.
Looyen, W. J., 1988. Geometric aspects of digital CAESAR
images. Neder!. Geod. Tijdschrift Geodesia, 30: 159-162.
Looyen, W. J. & R. W. Van Swol (Eds), et aI., 1989. CAE-
159
SAR: Performance and first evaluation of application pos-
sibilities. Report BCRS-89-06, Final Report Project TE-
1.3, June 1989. Netherlands Remote Sensing Board, Delft,
The Netherlands: pp. 101.
Malthus, T. J. & A. G. Dekker, 1988. Spectral light attenu-
ation in a hypertrophic lake system (Loosdrecht lakes, the
Netherlands). Verh. int. Ver. Limno!. 24: 711-714.
Miyazaki, T., H. Shimizu & Y. Yasuoka, 1987. High-speed
spectroradiometer for remote sensing. App!. Optics 26:
4761-4766.
Noordzij, W. J., F. B. van der Laan & W. Verhoef, 1989.
National Point of Contact, Products and Services, detailed
guide to imagery available. NPOC Annex to Dutch Remote
Sensing Newsletter 1989-no 38, (no page numbering).
Palmer, K. F. & D. Williams, 1974. Optical properties of
water in the near infrared. J. Opt. Soc. Am. 64: 1107-1110.
Philpot, W. D., 1991. The derivative ratio algorithm: avoiding
atmospheric effects in remote sensing. IEEE Trans. on
Geosc. and Rem. Sens. 29: 350-357.
Prieur, L. & S. Sathyendranath, 1981. An optical classifica-
tion of coastal and oceanic waters based on the spectral
absorption curves of phytoplankton pigments, dissolved
organic matter and other particulate materials. Limno!.
Oceanogr. 26: 671-689.
Seyhan, E., N. J. J. Bunnik, W. Verhoef & J. van Kuilenburg,
1974. Measurements of spectral signatures for water quality
monitoring. NIWARS Publication NO. 24, Delft, The
Netherlands. Paper presented at the First General Confer-
ence of the Remote Sensing Society, Birmingham, Great
Britain, Sept. 1974: pp. 29.
Smith, R. C. & K. Baker, 1981. Optical properties of the
clearest natural water (200-800 nm). App!. Optics 20: 177-
184.
Vos, W. L., M. Donze & H. Buiteveld, 1986. On the reflec-
tance spectrum of algae in water: the nature of the peak at
700 nm and its shift with varying concentration. Commun.
on San. Eng. and Water Management, nr. 7, ISSN-0169-
6246, TV Delft, The Netherlands 86-22: pp. 29.
Hydrobiologia 233: 161-163, 1992.
L. Van Liere & R.D. Gulati reds), Restoration and Recovery oj Shallow Eutrophic Lake Ecosystems in The Netherlands. 161
© 1992 KllIIver Academic Publishers.

Water quality research in the Loosdrecht lakes: proposals for the

follow-up restoration measures

Marien A. De Ruiter
Province of Utrecht, P.O. Box 80300, 3506 TH Utrecht, The Netherlands

Key words: eutrophication, departmental policy, Loosdrecht lakes, recovery, water management

Abstract

Reduction of the external phosphorus load in 1984 did not lead to recovery of Loosdrecht lakes.
Therefore the Provincial Water Authorities of the Province of Utrecht proposed, after consulting the
research group Water Quality Research Loosdrecht lakes (WQL), measures to reduce the present ex-
ternal phosphorus loading from 0.35 to 0.1 g Pm - 2 Y- I. Mathematical modelling suggested that such
a reduction would benefit the water quality of the lakes. The results of the WQL investigation did not
lead to additional measures (dredging, biomanipulation, chemomanipulation).

Introduction Breukeleveen, one of the lakes in the Loosdrecht

The provincial bylaw entrusts the Provincial
Water Authorities of Utrecht with the responsi-
bility of maintaining water quality within their
provincial jurisdiction. This has already resulted
in initiatives of the Province of Utrecht to restore
the water quality by providing sewerage systems
for the community inhabiting the Loosdrecht
lakes area, to divert domestical and industrial
loads, Besides, in close cooperation with Amster-
dam Municipal Water Works, supply of water to
the lakes from River Vecht has been replaced by
water from Amsterdam-Rhine Canal (Gulati
et al., 1991; Hofstra & Van Liere, 1992). Although
the concentration of total-phosphorus in the lakes
has showed a slight decrease, the concentration
of chlorophyll a has not decreased significantly
(Van Liere et aI" 1990). Also the required water
clarity for swimming water purposes, a Secchi-
disc transparency of 1 m was not achieved.
Therefore, Provincial Water Authorities directed
their attention to additional measures, for exam-
ple biomanipulation. This initial effort in Lake
lakes system, to stimulate large-bodied zooplank-
ton by reducing the planktivorous fish standing
crop did not produce the desired results (Van
Donk eta/., 1990).
This paper briefly discuss future proposals of
the Provincial Water Authorities for restoration
of the Loosdrecht lakes.
The proposed water management measures
The recommendations of the workgroup WQL
(Water Quality Research Loosdrecht lakes) have
shown clearly that further reduction of external
phosphorus loading to the lakes is still the crucial
measure in restoration of the lakes in this area
(see in Van Liere & Gulati, 1992), At present the
total external phosphorus sources are (Engelen
eta/.,1992):
1. The Polder Bethune water is supplied to the
lakes to prevent chloride concentration from
rising, because the chloride concentration in
the water from Amsterdam-Rhine Canal en-
162

tering the lakes is twice as high as in Loos-
drecht lakes' water. At present the water
from Polder Bethune enters the lakes via the
Southern Kievits Area. However, the concen-
tration of phosphorus in the polder-water is
not negligible, it is proposed to divert the
polder-water to the dephosphorization plant
near Loenen for phosphorus stripping before
entering the lakes (Fig. 1). This measure will
reduce the present external loading by about
30%.
2. The water that enters the lakes in wintertime
from the Eastern hinterland represents 30 % of
the external phosphorus load at present (En-
gelen et al., 1992). Collecting this water and
dephosphorizing it before it is left to flow into
the lakes will reduce the present phosphorus
load significantly.

LAKE

~ ~.,/{ '3.~;:~
·.>~:~X~·>'':i- ..·,;,:
.DEPHOSPH. PtAN'(
. ~Nieu';';:·.'·;
.. lo()sdrecht

.EASTfRN~" :,~,:~:~~~i;j?
HINTERLANQ;~;';!,:'·
..

LAKE :.:; . ,r,,:-:,~.' ,".. ,i •... ' r ~;i:',~ .

~. '~',;'~;/.~{'. ;;,.:<:,~,),:~~;~:;<~,~~,;~?~:r".i .; ~ "::
LAKE
LOOSDRECHT j;' BREUKELEVEEN

Fig, 1. The Loosdrecht lake area, with indication of the present and proposed phosphorus removal plants.

3. Expanding the existing dephosphorization fa-
cility near Loenen by adding a second unit
after the first one will reduce the external load
by another 10% (A. Graveland, pers. comm.).
4. Recreation lodges still not connected to the
sewerage system will be connected to it, or
supplied with storage tanks.
5. Toilets will be build on the recreational islands
and discharging or pumping of sewage to the
lakes will be prohibited.
6. Phosphorus entering the lakes via leakage of
sluices has been an important component of
the lakes' phosphorus balance during 1986-
1988 (Van Liere eta!., 1991). A proper main-
tenance of the three sluices still in use to trans-
port boats from River Vecht to the lakes will
minimize leakage, leading to about 10% re-
duction of the present load of phosphorus.
7. The Province of Utrecht has no means to pre-
vent of reduce atmospheric deposition, which
is estimated at 6 % of the total external phos-
phorus load to the lake area (Van Liere et a!.,
1991). To control this national and interna-
tional treaties have to be considered. After ef-
fectuation of the proposed reduction in exter-
nal phosphorus load the relative contribution
by atmospheric contribution may increase
threefold.
The total effect of all these measures will be a
reduction to one third in external phosphorus
load, i.e. from 0.35 to 0.1 g Pm - 2 Y- I. Mathe-
matical modelling has shown that already at a
reduction in phosphorus to one-half of the present
level, the chlorophyll a concentration will de-
crease (lanse et aI., 1992).
The total costs of the first three proposed mea-
sures is ca. 500000 ECU annually, which is 1.5%
fo the total budget of the Water Authority. The
costs of connecting recreational doges to the sew-
163
erage system or the construction of storage tanks
is roughly 9000 ECU. The costs of better main-
tenance of the sluices has not been estimated as
yet. However, the costs of pumping back the leak-
age water back into River Vecht are to be low.
The Province of Utrecht has currently no plans
for additional measures as dredging, biomanipu-
lation, chemomanipulation or reduction of exter-
nal nitrogen. Such measures are not likely to con-
tribute to the recovery of the lake system in the
present situation, as has been shown by the out-
come of the WQL-research programme (see e.g.
Van Liere & lanse, 1992).
References
Engelen, G. B., B. F. M. Kal, J. J. Buyse & F. G. M. Van
Pruissen, 1992. The hydrology of the Loosdrecht lakes area.
Hydrobiologia 233: 21-38.
Gulati, R. D., L. Van Liere & K. Siewertsen, 1991. The Loos-
drecht lake system: Man's role in its creation, perturbation
and rehabilitation. In: O. Ravera (ed.), Terrestrial and
aquatic ecosystems perturbation and recovery. Ellis Hor-
wood Limited, London.: 593-606.
Janse, J. H., T. Aldenberg & P. R. G. Kramer, 1992. A math-
ematical model of the phosphorus cycle in Lake Loosdrecht
and simulation of additional measures. Hydrobiologia 233:
119-136.
Van Donk, E., M. P. Grimm, R. D. Gulati, P. G. M. Heuts,
W. A. De Kloet & L. Van Liere, 1990. First attempt to
apply whole-lake food-web manipulation on a large scale in
The Netherlands. Hydrobiologia 200/201: 291-301.
Van Liere, L., O. F. R. Van Tongeren, L. Breebaart & W.
Kats, 1990. Trends in chlorophyll and total phosphorus in
Loosdrecht lakes, The Netherlands. Verh. int. Ver. Limnol.
24: 707-710.
Van Liere, L. & R. D. Gulati, 1992. Restoration and recov-
ery of shallow eutrophic lake ecosystems in The Nether-
lands: epilogue. Hydrobiologia 233: 283-287.
Van Liere, L. & J. H. Janse, 1992. Restoration and resilience
to recovery of the Lake Loosdrecht ecosystem in relation
to its phosphorus flow. Hydrobiologia 233: 95-104.
Hydrobiologia 233: 165-170, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 165
© 1992 Kluwer Academic Publishers.

The significance of the Loosdrecht lakes research project for

eutrophication policy in The Netherlands

H. W. Kroes
Ministry VROM (Housing, Physical Planning and Environment), Directorate of International Environmental
Affairs, P.O. Box 450, 2260 MB Leidschendam, The Netherlands

Key words: eutrophication, policy, environment

Abstract

The Project on the Water Quality Research in the Loosdrecht Lakes (WQL) has come up to the
expectations of the Netherlands Environment Ministry. The results reaffirm the main lines of national
eutrophication policy drawn up in 1979 (Policy Document on Phosphates) and further developed in the
eighties. Interesting new insights have been gained, for example into the role of sediment and seston as
well as into the relative importance of trophic levels. It is not possible, however, to definitely establish
the effect of WQL on eutrophication policy.
The Loosdrecht project is an example of genuine ecological research, incorporating several disciplines,
placing the object of research into its surroundings, emphasising the relation nature-culture and com-
mitting itself to certain value judgements and policy choices. As a consequence, there is a striking re-
semblance between the evolution of the project itself and that of ecosystems.
All in all the Loosdrecht project should be regarded as a paradigm for future ecosystem studies.

Introduction extrapolation to other surface waters in the Neth-

Some ten years ago two motives led to the con-
ception of the Loosdrecht Lakes research project:
a scientific one and a political one. From a sci-
entific viewpoint there was a need to gain insight
into the functioning of an aquatic ecosystem as a
whole, in relation to its environment. From the
viewpoint of policy and water management there
was a need to obtain information on the benefi-
cial effects of reducing the phosphate load of the
Loosdrecht lakes. The ministry of Housing, Phys-
ical Planning and Environment (VROM) made a
financial contribution to the project because of
the second motive. Of course, this decision was
influenced by the hope that enlarging the knowl-
edge on the Loosdrecht lakes would be useful for
erlands: VROM is a national policy-making body,
not a regional water board.
In this paper I will try to assess the success of
the project (indicated in this paper as WQL,
Water Quality research Loosdrecht lakes) by an-
swering the following four questions:
a. Has WQL come up to the expectations of
VROM in a general sense?
b. Has WQL affected policy on eutrophication
abatement or, in a broader sense, Dutch
government policy with respect to surface
water quality? Is it possible to indicate a pat-
tern of cause and effect and the extent of the
influence?
c. Has the government, especially VROM, ben-
efited from WQL in other respects?
166

d. Should WQL be an example for a similar
project in the future?
In trying to answer these questions I will also
touch on some general aspects of the relation
between research and environmental policy.
WQL and VROM's expectations
Each year the Directorate-General for Environ-
ment of the Ministry of VROM contracts with a
great number of research institutes, universities,
technical consultant bureaus etc. for carrying out
research. The objective is to update basic infor-
mation for environmental policy both for the short
and the long term, rather than to promote scien-
tific research as such.
An essential difference exists between studies
expected to provide specific information for the
short term and those intended to enlarge the gen-
eral basis of knowledge on environmental prob-
lems. VROM finances both categories as well as
the intermediate forms, but expectations as to the
outcome are not identical for the various types of
studies.
In case of a short term study immediate appli-
cability of the results in policy development is
expected. For example, after the accident in the
Swiss firm Sandoz in 1986 the Inspectorate for
Environmental Hygiene (a VROM agency) asked
a technical consultancy bureau to study the risks
of water polluting accidents in Dutch industrial
plants, and to draw up a list of such plants. The
results were used soon thereafter in the interna-
tional negotiations on the abatement of Rhine
pollution and for formulating a national policy to
prevent such accidents. Results of a strategic long
term study, on the other hand, at best may be
used for incorporation in long-term policy, but
sometimes only serve to develop general lines of
thought in policy-making departments.
The WQL research project is an example of a
strategic long-term study. Therefore, VROM's ex-
pectations were modest as to the direct applica-
bility of the results. The type of information col-
lected is important for the long term and can be
one of the many contributions to an intricate and
generally slow process of policy making (as will
be shown in the next section). In this sense, WQL
has come up to our expectations.
Eutrophication: research and policy
The second and most important question that has
to be dealt with, is whether WQL so far has had
an impact on Dutch national eutrophication pol-
icy. I cannot give an unambiguous answer to that
question, in spite of the fact that the project has
lasted for ten years. Answering that question in
my view is as complex a task as unraveling the

Table 1. Research on eutrophication 1960-1991. If only the year of start is given, the research is still continuing.

Period Name of surface water and research organisation

1960- Plas Vechten ('Lake Vechten'), Limnological Institute.

1967- Friese meren ('Frisian lakes'), Limnological Institute and Fosfri-project.
1967- IJsselmeer, Randmeren ('Lake IJssel' and 'Bordering lakes'), Limnological Institute, National Agency for
the IJsselmeerpolders, National Institute for Wastewater Treatment.
1972-1982 Eems-Dollard, Project Biological Research Ems-Dollart Estuary.
1975-1981 Grote Rug (Reservoir for drinking water supply, Dordrecht), National Institute for Drinking Water Supply.
1975- Reservoirs for drinking water supply (e.g. Biesbosch), Drinking water supply companies.
1975- Nieuwkoopse plassen, Reeuwijkse plassen etc. (lakes in the Rijnland area), Water Board 'Rijnland',
National Institute of Public Health and Environmental protection.
1979- Loosdrechtse plassen ('Loosdrecht lakes'), some 10 research institutes: WQL.
1986- Kortenhoefse(Ankeveense plassen (lakes north of the Loosdrecht area), Zuiveringsschap Amstel- en
Gooiland and Agricultural University Wageningen.
1987- Rhine, National Institute of Public Health and Environmental Protection.
 167

processes in the eutrophication of the Loosdrecht
Lakes.
To assess the possible impact of eutrophication
research in general and WQL in particular on
policy making, I have summarised (Tables 1 and
2) the major research projects since the sixties
and the highlights in policy making since the early
seventies. The information shows that, although
much has been done and is still in progress both
in research and in policy making, it is almost
impossible to establish a direct cause-and-effect
relationship between the two fields. Generally
speaking, the only certainty is that research
projects started much earlier than the policy mak-
ing, and that individual research projects last
longer than policy making projects.
Finding the causal relations would require an
extensive and detailed study. On the basis of my
own experience I fear, however, that the essential
relations would, in all probability, remain hidden.
The process of policy making is complicated; cru-
cial moments can only be pointed at by insiders,
and that too only in a subjective way, difficult to
verify by outsiders. The role of individual persons
can be important, but the role of collective deci-
sion making is important too, and, finally,
progress in policy making sometimes strongly de-
pends on accidents. The Sandoz accident for ex-
ample, resulting in the Rhine Action Programme
(ICPR, 1987), triggered a marked acceleration in
international and national eutrophication abate-
ment policy.
I can illustrate the above with an example taken
from policy developments in the seventies. The
decision to produce a policy statement on phos-
phates under the term of office of Mrs. Vorrink,
Minister of V&M (,Public Health and Environ-
mental Hygiene', the predecessor ofVROM), was
not taken after a linear process with consecutive
phases of research, policy preparation and policy
decisions, but more or less by coincidence. A
report of scientists (KNCV, 1976) had been pub-
lished, which in its turn was the result of a mix-
ture of scientific considerations and ideas of in-
dividual persons (Golterman & others) on a
desirable policy. There were action groups on the
subject of detergent phosphates and the Dutch
Parliament had questioned the Minister about

Table 2. Policy making on eutrophication 1972-1991.

Period Document/event

1972 'Urgentie nota milieuhyiene'. Government statement on urgent issues in environmental hygiene, including a rough
sketch of the eutrophication problem.
1978 'KNCV-rapport'. Royal Netherlands Chemical Society: recommendations on phosphate policy.
1979 'Fosfatennota'. Government policy statement on measures from 1980 to 1985 to reduce eutrophication:
phosphate removal in sewage treatment plants, banning phosphates from detergents, international actions,
regional actions.
1980 Agreement with detergent manufacturers in the Netherlands to reduce phosphate content of textile detergents
to below 7,5% (as P) in 1983.
1985 Objectives of Fosfatennota not realised.

1986 Sandoz accident (Rhine, Basel).

1987 Rhine Action Programme. One of the objectives: a 50% reduction of phosphate load of the Rhine between
1985 and 1995.
1987 Third International Conference on the Protection of the North Sea. One of the objectives: a 50% reduction
of phosphate load of the North Sea between 1985 and 1995.
1987 Decree on the application of animal manure (Soil Protection Act). Restricts the amounts of phosphate to be
applied per ha of grassland or arable land.
1989 'Nationaal Milieubeleidsplan'. National Environmental Policy Plan with objectives on the reduction of
eutrophication of water and soil.
1990 Decree on limit values for phosphate in effluents of sewage treatment plants (Surface Water Pollution Act).
1991 Draft Decree on limit values for nitrogen in effluents of sewage treatment plants (Surface Water Pollution Act).
168
phosphates in detergents. Within the ministry a
discussion had been going for some time on
whether or not a policy statement on phosphates
should be published; the minister's opinion had
not yet been asked.
Mrs. Vorrink was less patient than some of her
officials and, adding a remark to a draft answer
to parliamentary questions, asked when, at last,
the 'Fosfatennota' would be ready for publica-
tion. This tipped the balance in the departmental
offices in favour of those who wanted to publish.
In this way, the policy statement eventually ap-
peared. It was not Mrs. Vorrink, however, but her
successor, Dr. Ginjaar, who signed the 'Fosfaten-
nota' (Minister V&M & Minister V&W, 1979).
What would have happened without this coin-
cidence? Would the statement have been pub-
lished anyway? And would it have been published
if the research results up to that moment had not
been available?
Such questions are difficult to answer. It is
possible, however, to try and find a pattern from
a subjective viewpoint, i.e. from the impressions
of individual scientists and policy makers. I did
not conduct an enquiry, but on the basis of my
experience, I conclude that the impact of research
on policy:
- depends on the question whether a link be-
tween the two has been established in advance
in a particular case, and on the closeness of
that link,
- depends on the complexity of the problem,
- is often difficult to isolate from the multiplicity
of factors that may influence the decision mak-
ing by a minister and his/her advisers directly,
but also indirectly through effects on public
opinion, parliament, regional authorities, etc.
For the WQL project links between research
findings and specific policy decisions were not
stated at the start of the project. Before making
a financial commitment VROM did not state
whether or not policy would be adjusted accord-
ing to the outcome of the investigations. Conse-
quently, there was no reason to expect that on the
completion of the project important decisions
would be made on this particular basis. In gen-
eral the results of the investigations seem to sup-
port the policy developed in the past fifteen years
or so. Important among these are: (1) external
phosphate flows to surface waters have to be re-
duced drastically and (2) if the external load re-
mains high, the in-lake restoration measures, such
as dredging and biological management will be
vain. External phosphate loading can only be re-
duced sufficiently by tackling all sources, i.e. di-
rect discharges from sewage treatment plants and
(phosphate fertilizer) industry, pollution from ag-
ricultural areas and transboundary pollution, for
example via the River Rhine. In addition, as the
'Fosfatennota' indicates, sufficient attention
should be paid to regional and local circumstances
by drawing up regional management plans to com-
plement the national and international policy.
Without any doubt, the impact of WQL on
government policy would have been more spec-
tacular if the results had been more surprising, for
example if a hitherto neglected key factor in eu-
trophication had been discovered. Nevertheless,
in addition to this confirmation of the basic ele-
ments of eutrophication policy, WQL has pro-
vided policy makers with insight into many im-
portant aspects, for example the role of the
sediment, the seston and the higher trophic levels.
Many examples are given in the other papers in
this volume.
In this connection it is interesting to quote from
the opening address at a symposium on eutroph-
ication in 1979 by the then Environment Minis-
ter (Ginjaar, 1980):
'The role of the scientist nowadays con-
cerned with the problems of eutrophication
is not longer that of a pioneer, explorer or
prophet, as it was ten years ago, but that of
a patient filler-in and finisher, with an eye for
fine detail. This last is no less important than
the first; but it is different. The great discov-
ery has been made: it bears a capital P.'
For the management of the Loosdrecht Lakes
themselves WQL has evidently produced a wealth
of information. Details on this subject may be
found in the paper by De Ruiter (1992).
The eutrophication problem is 'not a simple

one'. Therefore, one cannot expect eutrophica-
tion research to produce simple recipes for policy
making. The relations between research and pol-
icy in a given field will therefore be diffuse and
iterative, and extend over a long period of time.
For the scientist as well as the civil servant and
the politician this can be frustrating. The scien-
tists may not recognise his results being put into
practice. On the other hand, politicians and civil
servants may be impatient and tempted to think
that such studies do not contribute to decisions
in policy making.
This takes us to the third factor affecting the
impact of research on policy making: the 'paths'
along which influence is exerted. The main paths
of influence in water management policy in the
Netherlands roughly run from scientists to policy-
making officials in the ministries. Rarely is there
a direct impact of research results on the opinion
of politicians. One should realise that there are
close contacts between policy making groups in
the ministries and aquatic research institutes. The
National Water Management Agency of the Min-
istry of Transport and Public Works ('Rijkswa-
terstaat') has policy making branches as well as
research divisions. The Environment Ministry
(VROM) has close connections with the National
Institute of Public Health and Environmental
Protection (RIVM) in the field of environmental
research. Provinces and Water Boards often em-
ploy their own research scientists. In these cases,
therefore, the relationship between research and
policy making is 'built-in'.
Some other institutes, however, do not have
such a close connection with policy making. This
is true for the major participants in the WQL
project, viz. the Limnological Institute, the Free
University, Amsterdam, and the University of
Amsterdam. They do have relations with policy
making authorities, albeit mostly on a temporary
basis, through incidental research contracts. I
have the impression that scientists in these insti-
tutes sometimes think their research does not get
due attention from the authorities. This may make
them refrain from attempts to get involved in pol-
icy making. In stead of trying to influence policy
more directly, for example by approaching po li-
169
ticians and officials in the ministries, they con-
centrate exclusively on research.
Other benefits for policy from WQL
Apart from the often elusive impact of research
on the material content of policy, there are vari-
ous effects on the way policy is made.
Policy makers may, for example, want to jus-
tify decisions already made in their own minds
with research projects still to be carried out. They
thus use the scientist as a notary, i.e. to provide
a document signed by an expert in the field. In
Germany this practice has even been institution-
alised in the so-called 'Gutachten' produced by
scientists both for companies and authorities.
Research projects also are used sometimes to
draw attention to specific environmental prob-
lems when it is impossible to reach a solution by
negotiations. Authorities as well as citizens use
this instrument. A good example is the study that
was started by the municipality of Rotterdam a
few years ago on the discharges of industries re-
sponsible for the pollution of Rhine sediment in
Rotterdam harbour (Rotterdam, 1986). The mu-
nicipality wanted to force a reduction of the dis-
charges of heavy metals and other toxic sub-
stances to a level not yet attained in the Rhine
negotiations between national governments. In
such a case research projects are being used as a
crowbar. A third motive for having research car-
ried out for other reasons than the enlargement of
knowledge can be a wish to postpone decisions.
Research projects may be used as an excuse for
not having to decide immediately.
I think that if WQL has acted as notary, crow-
bar or excuse at all, it has done so only to a
limited extent. In any case, the Environment Min-
istry, in deciding to finance parts of the pro-
gramme, mainly sought the enhancement of
knowledge on eutrophication in a general sense;
it did not attempt to justify its policies or to force
decisions and certainly did not try to find an ex-
cuse for refraining from taking decisions. What I
would like to emphasise is that scientists should
know when they are being used as a notary, as a
crowbar or as an excuse. Of course, scientists can
170
Table 3. Comparison of the evolution of ecosystems and the evolution of ecological research projects, exemplified by WQL.
Evolution of ecosystems
I. Individuals concentrating on survival and reproduction
2. Populations striving to establish and grow
3. Competition between individuals in a population
4. Competition between species
5. Cooperation of individuals (social species)
6. Coevolution of species (millions of years)
7. System characteristics: diversity, division of labour
8. Control: polycentric
9. Important product: man
Subsequently, left and right are exchanged and the whole process can 'restart' from the beginning. Ecosystem processes can then
be studied on the basis of the newly gained knowledge in new research projects. Therefore: WQL = paradigm for the future'
consciously accept such a role as part of the strat-
egy of a policy making authority.
Loosdrecht ecosystem research: a paradigm for
future studies
In my opinion WQL should have a successor, not
perhaps in the sense of a study of the same ob-
ject, but as to its approach. Ifwe assess WQL by
the criteria that according to Morin (1980) are
essential for ecology as a science, it has been a
programme of genuine ecological research:
the object itself and its environment were stud-
ied,
the approach was multidisciplinary,
the relationship between nature and culture
was emphasised,
policy choices and value judgments were dis-
cussed.
All these aspects required long term planning,
many investigators, adequate funding and excel-
lent coordination. Therefore it will not be easy to
initiate and carry out a similar programme of
studies in the future. I do see perspectives, how-
ever, for example in a Rhine ecosystem research
programme. In this regard the research in progress
in the Netherlands could be the core for a pro-
gramme that could gain an international dimen-
sion (de Wit et ai., 1989).
Lastly, I would like to reaffirm my plea for a
'do-it-again' with a brief comparison between the
evolution of two kinds of systems: ecosystems
and ecological research programmes, the latter
Evolution of WQL cooperation
I. Investigators intent on 'publish or perish'
2. Institutes striving for scientific recognition and growth
3. Competition between persons
4. Competition between institutes
5. Cooperation within working groups
6. Steady growth of cooperation between institutes (years)
7. System characteristics: unity in diversity
8. Control: polycentric (under supervision of coordinator')
9. Important products: knowledge and books
exemplified by the WQL project. This may seem
a strange comparison, because an example of the
first, the Loosdrecht lakes ecosystem, was the
object of investigation of the second (a research
programme). The special characteristic of ecolog-
ical research, however, is that in its optimal form
it shows a remarkable similarity with its object:
the ecosystem. See Table 3.
Acknowledgements. Thanks are due to Mr. H. R. Van Ot-
terloo for critically reading manuscripts for this paper and for
linguistic improvements.
References
De Ruiter, M., 1992. Water quality research in the Loosdrecht lakes:
proposals for the follow-up restoration measures. HYdrobiologia
233: 161-163.
De Wit, J. A. W., W. Admiraal, C. van der Guchte, W. C. Cazemier,
1989. Summary of results and conclusions from the first phase
(1988-1989) of the Netherlands research programme 'Ecological
Rehabilitation Rhine'. Publications and reports of the project 'Eco-
logical Rehabilitation of the River Rhine', nr. 10, 18 pp.
Ginjaar, L., 1980. Opening address. Proceedings of the Symposium
'Eutrophication in the Netherlands'. Hydrobiol. Bull. 14: 3-4.
ICPR, 1987. (International Commission for the Protection of the Rhine
against pollution). Rhine Action Programme. Technisch-wissen-
schaftliches Sekretariat, Koblenz, 24 pp.
KNCV, 1976. Fosfaten in het Nederlandse oppervlaktewater. Rapport
Stuurgroep Fosfaten van de Sectie Milieuhygiene, Stichting Vit-
geverij Sigma Chemie.
Minister V&M & Minister V&W, 1979. Maatregelen voor het ter-
ugdringen van de fosfaatbelasting van het Nederlandse oppervlak-
tewater (Fosfatennota). Tweede Kamer, zitting 1978-1979, 15640,
nrs. 1-2,30 pp.
Morin, E., 1980. La methode, 2: La vie de la vie. Editions du Seuil,
Paris, 470 pp.
Rotterdam, 1986. Havenbedrijf Rotterdam & Gemeentewerken Rot-
terdam. Project Onderzoek Rijn (naar schoon havenslib). Report,
48 pp. + 4 annexes.
Hydrobiologia 233: 171-177, 1992.
L. Van Liere & R.D. Gulati (eds). Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 171
© 1992 Kluwer Academic Publishers.

Water quality research in Loosdrecht lakes: the salient features

Ramesh D. Gulati 1 & Louis Van Liere 1,2

1 Limnological Institute, Rijksstraatweg 6, 3631 AC Nieuwersluis, The Netherlands; 2 National Institute of
Public Health and Environmental Protection, POB 1,3720 BA Bilthoven, The Netherlands

Key words: eutrophication, restoration, recovery, water quality

Abstract

Ecosystem research by the working group Water Quality Research Loosdrecht lakes (WQL) was car-
ried out from 1979 to 1990. A coordinated research programme, involving several research institutes and
laboratories in The Netherlands, was initiated in 1983, i.e. a year before the reduction of external
phosphorus loading by stripping, became effective. The paper summarizes the main results, with em-
phasis on insight they provide into the lake ecosystems.

Introduction Even though the WQL was formally dissolved

Routine work on the water chemistry of Loos-
drecht lakes was started in 1979 by the Amster-
dam Municipal Waterworks, initiated by the
working group 'Water Quality research Loos-
drecht lakes' (WQL). The restoration measures
involving external phosphorus-reduction came
into effect in 1984. From 1985 WQL was fully
operational in that most study aspects it had
planned to encompass were being investigated
concurrently, and on routine basis. The WQL
was unique in limnological research in The Neth-
erlands: it involved active cooperation between
several institutions (Van Liere et al., 1992). The
research carried out by WQL was fundamental,
and had great social relevance through its 'mission
orientation' (VanLiere et al., 1992; Kroes, 1992;
De Ruiter, 1992). The WQL was wound up at the
end of 1990, culminating in a two-day symposium
in the spring of 1991, in which a large part was
devoted to a review of the WQL followed, on the
second day, by studies on eutrophication and
water management measures in other Dutch lakes
(Van Liere & Gulati, 1992)
in April 1991, some essential elements of the re-
search have since then been incorporated into the
routine monitoring of the Loosdrecht lakes at the
Limnological Institute. A delay in response to
restoration measures (Van Liere & Janse, 1992)
has necessitated continuation of monitoring some
of the parameters like the underwater light cli-
mate, chlorophyll, total-phosphorus, and carbon
to phosphorus ratios of seston.
Scientific achievements
Van Liere (1992) provides a bibliography of more
than 132 papers that were published within a span
of ten years (1983-1992).
Here we briefly summarize the main WQL
achievements in the different disciplines, with
special emphasis on how they have added to our
existing knowledge about such shallow, eutrophic
lakes.
The hydrological research revealed that both
the upward and downward seepage and interac-
tion between ground- and surfacewater, both in
172
terms of quantities of the water involved and
qualitative differences between these, were impor-
tant in understanding the seasonal nutrient dy-
namics (Engelen et aI., 1992). Moreover, interac-
tions between the various systems in the area,
both in the past and recent, have had been strongly
influenced by human activities in the past as well
as at present. The main among these latter being
the recent increased need for public water supply
and the regional transfer of water. Unique fea-
tures of the hydrological research were: (1) the
development of two- and three dimensional mod-
els for ground water flow, based on monthly and
annual water balances: and (2) field experiments
on turbulence and diffusion using tracer tech-
niques and laboratory studies on wind driven cur-
rents. As regards surface water movements, the
remote sensing studies appear to simultaneously
provide the much-needed complementary infor-
mation. The hydrological studies strongly plead
for conserving the ground water in the catchment
area (the Gooi Region). Though the studies of
Engelen et ai. (1992) suggested flooding the
deeper polder areas to reduce seepage losses, the
repercussions of such a measure, if taken, need to
be seriously considered (see later in this paper).
The studies on phosphorus recycling in the sed-
iments (Keizer & Sinke, 1992) indicate that re-
generation via mineralization of phosphorus of
(ca. 3 mg P m - 2 d - 1) is an order of magnitude
higher than the diffusive release from sediments
(0.3 mg P m - 2 d - 1) under aerobic conditions at
20°C. Interestingly, this diffusive release de-
creased from about 1 mg P m - 2 d - 1 in 1984 to
0.3 mg P m - 2 d - 1 in 1990. Only a very small part
of the phosphorus appears to be adsorbed by
Fe(III) in the top sediment layer. This is attrib-
uted to a marked downward seepage loss of in-
organic phosphorus (0.8 mg P m - 2 d - 1 ), as well
as to uptake of this phosphorus during resuspen-
sion. Only the upper 2-cm sediment which con-
tains ca. 20 % potentially bioassimilable phospho-
rus is important. Thus, seepage, sediment burial,
and its diagenesis and mineralization are impor-
tant factors in the loss of phosphorus from the
cycle. The model predictions suggest that any fur-
ther reduction of external phosphorus load may
cause a gradual decrease in total phosphorus con-
centration (Keizer & Sinke, 1992).
The simulation of the phosphorus cycle using
the dynamic mathematical model PCLOOS
(l anse et ai., 1992) involved mass and rate pro-
cesses at all the trophic levels, as well as the in-
organic phosphorus fractions. The input of the
model were the hydrological water- and matter-
balance data (Engelen et ai., 1992). Also carbon
was included in modelling of the organic com-
partments, independent of phosphorus, so that
the PIC ratio's could be modelled too. However,
the separation of seston into phytoplankton and
detritus gave discrepancies in the flow of phos-
phorus in the open water, as well as for growth
rates compared with values based on primary
production (Van Liere & lanse, 1992), although
these discrepancies were not large. The outcome
for phosphorus was largely comparable with the
empirical data that indicate a decrease in total-
phosphorus concentration from ca. 130 jJ.g P 1- 1
to 80 jJ.g P 1- 1>, but no such decrease was pre-
dicted for concentrations of seston and chloro-
phyll a. A delayed response to the reduced phos-
phorus loading and a new equilibrium in about
five years was predicted. Both internal recycling
and high assimilation efficiency of cyanobacteria
are given as the causes for the delay in response
to rehabilitation measures. Additional restoration
measures such as flushing, chemomanipulation,
or biomanipulation are considered as ineffective
complementary measures at the present level of
phosphorus loading (Van Liere & lanse, 1992).
Therefore, research on phosphorus dynamics and
modelling presents a strong plea for further re-
duction in the present rate of external loading, viz.
0.35 g P m- 2 y-l.
The discovery of Prochiorothrix hollandica
(Burger-Wiersma et ai., 1986, 1989), the first free-
living oxychlorobacterium isolated from a fresh-
water lake, was an impressive scientific achieve-
ment. This filamentous form, which measures ca.
150 jJ.m in length and about 2 jJ.m in width, re-
sembles Oscillatoria iimnetica in its superficial ap-
pearance, but, unlike cyanobacteria, it contains
chlorophyll b, and lacks phycobilins, as well as
has other biochemical characteristics that distin-

guish it from cyanobacteria. In Loosdrecht lakes
P. hollandica forms an important proportion of
the total densities of filamentous prokaryotes, in-
cluding cyanobacteria. With actual growth rates
generally less than ca. 0.1 d - 1, its maximal den-
sities (ca. 120.000 ml- I) occur in mid summer
when the prevailing temperatures (18-22 0c) re-
sult in growth rates exceeding 0.1 d - 1. Prochlo-
rothrix, which has been considered as a missing
link between Cyanophyta and Chlorophyta, has
aroused renewed interest in molecular and evo-
lutionary biology of prokaryotes.
Studies on seston dynamics in shallow lakes
are rare and information on the effects of changes
in wind velocity and direction on the spatial and
temporal distribution is very scanty. The present
studies on Loosdrecht lakes (Gons et al., 1992)
have demonstrated that wind action is the main
driving force in the stirring up of sediments and
resuspension and translocation of seston. The ex-
tent to which seston is translocated depends
greatly on fetch and wind direction. The action of
wind in these lakes, thus, appears to be more
important than turbidity caused by bream brows-
ing in the sediments in search of food and the
'normal' water-based recreational activities of
man. Apparently frequent resuspension of seston,
coupled with the dominance of filamentous
prokaryotes and brown colour of the water, due
to dissolved humic substances, are the main fac-
tors for the poor light climate in these shallow,
peaty lakes (Gons & Otten, 1991).
The relationships between the dominant phy-
toplankton species and detritus were tested in the
laboratory in the specially developed Laboratory
Scale Enclosures (LSE's) (Rijkeboer et al., 1991).
These systems allowed studying the effect of dif-
ferent loads of phosphorus on growth rates of
Oscillatoria spp. and Prochlorothrix. The exten-
sion of studies using the LSE's led to the devel-
opment of the Epipelon Trap (ET) to collect the
overflow (among which the detritus produced),
and to follow its fate further. It may be mentioned
that detritus, which forms two-thirds of total
seston in the lakes, was found to be an important
phosphorus buffer apparently due to the affinity
of soluble phosphorus to adsorb onto this detri-
173
tus (Rijkeboer et al., 1991). On the basis of a
steady-state model of the dynamics of phyto-
plankton detritus, the areal organic dry weight
concentration of detritus in the lake is ca. 60 g
DW m - 2 distributed equally in the seston and
epipelon (Otten et al., 1992). The mineralization
rate of the refractory detrital material, based on
laboratory studies at 20°C is only 0.005 d - t, but
in the longer term will be even lower. Thus, re-
fractory detritus tends to accumulate, and may
influence the water quality for a long time to
come.
The zooplankton composition in the lakes is
typical of highly eutrophic lakes: 1) dominance in
the Crustacea of small-bodied cladocerans,
mainly Bosmina spp. and cyclopoid copepods;
and 2) a very high diversity (species composition)
of rotifers, with at least 20 taxa and 40 species
though only a few are dominant. (Gulati, 1990).
A complete absence of large-bodied daphnids
(Daphnia galeata) common in many other eutro-
phic lakes was quite striking; the cause of the
virtual disappearance in the more recent years of
the small-bodied daphnids (D. cucullata) is not
known. These changes in the daphnid popula-
tion, and lack of larger-sized species, appears to
be due to intensive predation by planktivorous
fish, mainly bream (Abramis brama).The food
quality and quantities are apparently influenced
by the very high concentrations of filamentous
prokaryotes (> 10 8 filaments 1- 1 ), with their likely
adverse effects on the grazing activities of crus-
taceans. In contrast to the Daphnia species, the
bosminids are well adapted to the abundant food
in the lakes and are less vulnerable to predation
by planktivores. About two-thirds of daily mean
grazing by crustaceans of 4-10 % was attribut-
able to the populations of Bosmina alone. On the
other hand, rotifers, especially the smallest among
them, appeared to be important grazers of food
particles < 10 j.1.m, dominated mainly by flagel-
lates, green algae and bacterioplankton (Wilms
et al., 1990; Ooms-Wilms, 1991).
The phosphorus excretion by crustaceans,
though low compared with values from the
literature, covered about one-third of the daily
phytoplankton phosphorus demand (Den Oude
174
& Gulati, 1988). In addition, the rotifers are as-
sumed to regenerate relatively much more phos-
phorus daily (Gulati et al., 1992). The PIC ratios
of the particulate matter exhibit an increase with
trophic level, with the more constant zooplankton
ratio of ca. 0.02 (Gulati et al., 1991), being inter-
mediate between those of seston (0.009) and fish
(ca. 0.03).
The role offish in phosphorus recycling through
its excretion and egestion as well as leaching from
dead fish could be crucial but remains unquanti-
fied. In view of the fish standing crop of ca. 320 kg
ha - 1 in these lakes the phosphorus in the fish was
estimated at about 160 mg P m - 2 (VanLiere &
lanse, 1992), i.e. about 50% of the total phos-
phorus in the water column. Even though fish
may be considered as a phosphorus sink, consid-
ering their phosphorus-stock, they form a contin-
uous source of phosphorus release into the sys-
tem through their excretion, egestion and
mortality.
Besides the achievements and lacunae of the
WQL research there are several areas in which
our knowledge of the Loosdrecht lakes' ecosys-
tems is very sketchy. More information is needed
on composition and structure of fish and their
food preferences. Bottom fauna, dominated by
Chironomus pulmosus, was found to be rich but its
importance has not been quantified. Besides, the
role of ciliates, which were found abundant in the
benthic collections (I.lllpublished data) in these
eutrophic systems has remained virtually un-
known. Both ciliates and flagellates may form
important pathways through which bacterial en-
ergy is channelled and grazer and decomposer
food chains are linked. In short, though a great
deal has been achieved by the WQL, much still
remains to be done to advance our knowledge
about these shallow, eutrophic and 'resilient' lake
ecosystems.
Spin-off and social relevance
A large coordinated and integrated long-term re-
search programme, as the one in Loosdrecht
lakes, offers an infrastructure which has a great
scope to attract other research subjects. Only
when a programme of multidisciplinary research
has been established can the ecosystem modellers
start their work. However, WQL chose to start
from the onset with the mathematical modelling
in order to have an integration between model-
ling, and field research and laboratory studies.
The logistics and infrastructure of WQL made it
possible to collect at an extremely short notice a
large number of samples over a vast area for
groundtruth verification of data collected by re-
mote sensing research. Together with the proper
remote sensing research this resulted in a water
quality model, which can be extended easily to
other lakes (Dekker et al., 1992).
Finally; the integration of various disciplines
and cooperation among scientists with different
interests and expertise, and working in different
institutes was among the WQL's major achieve-
ments. The contacts and exchange of scientific
information through discussons and joint publi-
cations led to greater understanding of the func-
tioning of the ecosystem, but also increased the
understanding of the scientific problems of one
another's discipline. It led also to a platform
for discussions with policy makers, water man-
agers and money givers, culminating in the bi-
annual plenary workshops, with all participants
involved.
The 'mission oriented' research programme of
WQL resulted in an efficient infrastructure for
consultancy to the water quality authorities. The
WQL scientists have been active participants in
various committees to advise the Water Quality
Authorities of the Province of Utrecht, but also in
committees dealing with eutrophication problems
in other Dutch lakes, but also European lakes in
general. This resulted in several reports on hy-
drology (see Engelen et al., 1992), flushing (Van
Liere, 1986), dredging (Boers et al., 1985; Higler
et al., 1987), and on the reactions of suspended
matter (Gons, 1987; Gons & Van Keulen, 1989).
A workshop on recovery of ecosystems from eu-
trophication was organized as part of a cooper-
ative international research programme of
the European Community (Giussani et al.,
1991).

However, the most important 'mission' in this
context was the advice on additional measures
(for a review see Van Liere & J anse, 1992).
Summarizing: all the WQL-authors of this vol-
ume agreed that the most important measure was
a further reduction in external phosphorus load.
This measure was preferred to measures to re-
duce the'internal' phosphorus load, and to eco-
system management by chemomanipulation or
biomanipulation.
To achieve improvement in the water quality
choices have to be made, and targets to be as-
sessed. Hofstra & Van Liere (1992) assessed the
target values for the ecosystem as those in the
period 1930-1940, when the lakes were me-
sotrophic, macrophyte dominated, and rich in flo-
ral and faunal species. Choice of such a 'refer-
ence' is arbitrary but has some realism. The
oligotrophic situation, as it existed at the begin-
ning of the century, is presumably impossible ever
to accomplish. And the mesotrophic situation
(1930-1950) was stable for many years at a higher
external phosphorus load than the present one
(Van Liere et ai. 1991).
Engelen et ai. (1992) suggest a return to the
'original' hydrological situation, the return of the
groundwater from the Pleistocene sandridge of
the region The Gooi. However, this choice is as
arbitrary as the one above. The groundwater
quality also has changed; it contains a fairly high
amount of soluble phosphorus (Hettling, 1985), it
is not clear what will be its fate on its way to the
lakes. Keizer & Sinke (1992) and Janse et ai.
(1992), as mentioned, calculated the present
downward seepage as an important loss factor
for phosphorus. Now the external load and the
loss factor are nearly in equilibrium (Van Liere &
J anse, 1992). If the upwelling ground water flow
were to be restored, by means of inundation of the
Polder Bethune, the downward seepage from the
Loosdrecht lakes will be reduced almost com-
pletely (Engelen et ai., 1992). Although the lakes
then need less supply of water, the phosphorus
entering the lakes will have a longer retention
time. In short, the incoming phosphorus will re-
main in the lake ecosystem, thus entering the
foodchain to be stored in lake sediment. If no
175
0.20 , - - - - - - - - - - - - - - - - - - - ,
-- no change
. - - - - - - inundation of polder Bethune
inundation of polder Bethune plus
~ 0.15 dephosphorization of hinterlands
0-
Cl
oS
VI
::J
0 0 . 10
.c
a.
VI
o
.c
a.
~ 0.05
I--
0.00 +--~--~--~-~--~----1
o 3
Time[YI
Fig. 1. Simulation of total-phosphorus concentration in the
Lake Loosdrecht using three scenarios. No change means
external phosphorus loading as at present (0.35 g P m - 2 Y- I);
see lanse et al., 1992). Hydrological situation as in 1986.
phosphorus were to be lost from the system by
seepage recovery would be counteracted.
Janse (pers. comm.) simulated the inundation
of the Polder Bethune using the hydrological
water- and matter-balance model (Engelen et al.,
1992), and its output as input for the ecosystem
model PCLOOS (Janse et ai., 1992) to evaluate
the influence of this inundation. The effect of the
formation of 'Lake Bethune' on the phosphorus
concentration in (Fig. 1) was compared, with the
'status quo' (present situation: 0.35 g P m - 2 y- 1).
Clearly the inundation of Polder Bethune has a
negative effect on Loosdrecht lakes' water qual-
ity. The water quality of the lakes will deteriorate,
and even an additional dephosphorization of
water from the eastern catchment area will not
counteract this negative effect on water quality
sufficiently.
On achieving reduction of phosphorus, with an
uninundated Polder Bethune, model calculations
revealed that, only after reduction to 50 % of the
present load, will the concentrations of total-
phosphorus and chlorophyll a decrease (Janse
et ai., 1992).
However, Loosdrecht lakes still will have its
detritus both suspended, sedimentated and often
resuspended (Gons, 1987). The rate of recovery
of these lakes will be related directly to the ratio
between loss of this detritus and input of organic
 176
matter via primary production (Gons et al., 1992;
Otten et al., 1992). Thus, a further reduction in
the external P-load will playa crucial role.
Acknowledgements
We gratefully acknowledge Jan Janse's model
calculations, and his permission to use them
which resulted in the presentation of Fig. 1. '
References
Boers, P. C. M., Th. E. Cappenberg & P. Van lersel, 1985.
Baggerexperiment in de Loosdrechtse Plassen, voorberei-
dend onderzoek (A dredging experiment in L. Loosdrecht,
spade-research, in Dutch). WQL-report 1985-10.
Burger-Wiersma, T., M. Veenhuis, H. J. Korthals, C. C. M.
Van Der Wiel & L. R. Mur, 1986. A new prokaryote con-
taining chlorophylls a and b. Nature 320: 262-264.
Burger-Wiersma, T., L. J. Stal & L. R. Mur, 1989. Prochlo-
rothrix hollandica gen. nov., sp. nov.: a filamentous oxygenic
photoautotrophic prokaryote containing chlorophylls a and
b: assignment to Prochlorotrichacae fam. nov. and order
Prochlorales Florenzano, Balloni and Materassi 1986 with
emendation of the ordial description. Int. J. Syst. Ba~t. 39:
250-257.
Dekker, A. G., T. J. Malthus, M. M. Wijnen & E. Seyhan,
1992. Remote sensing as a tool for assessing water quality
in Loosdrecht lakes. Hydrobiologia 233: 137-159.
Den Oude, P. J. & R. D. Gulati, 1988. Phosphorus and ni-
trogen excretion rates of zooplankton from the eutrophic
Loosdrecht lakes, with notes on other P sources for phy-
toplankton requirements. Hydrobiologia 169: 379-390.
De Ruiter, M. A., 1992. Water quality research in Loosdrecht
lakes: proposals for the follow-up restoration measures,
Hydrobiologia 233: 161-163.
Engelen, G. B., B. F. M. Kal, J. J. Buyse & F. G. M. Van
Pruissen, 1992. The hydrology of the Loosdrecht lakes area.
Hydrobiologia 233: 21-38.
Giussani, G., L. Van Liere & B. Moss, 1991. Ecosystem
Research In Freshwater Environment Recovery. Proceed-
ings of a Workshop, Pallanza, Italy, Sept. 26-29, 1990.
Mem. 1st. ita!. Idrobio!. 48: 1-362.
Gons, H. J., 1987. Relationship between transparency and
suspended matter in L. Loosdrecht, in regards to the stan-
dards of swimming-water (in Dutch). WQL-report 1987-3.
Gons, H. J. & R. Van Keulen, 1989. Relationship between
transparency and suspended matter in L. Loosdrecht, final
report (in Dutch). WQL-report 1989-4.
Gons, H. J., J. H. Otten & M. Rijkeboer, 1991. The signifi-
cance of wind resuspension for the predominance of fila-
mentous cyanobacteria in a shallow, eutrophic lake. Mem.
1st. ita!. Idrobio!. 48: 233-249.
Gons, H. J., T. Burger-Wiersma, J. H. Otten & M. Rijkeboer,
1992. Dynamics of phytoplankton and detritus in a shallow
eutrophic lake (Lake Loosdrecht, The Netherlands). Hy~
drobiologia 233: 51-59.
Gulati, R. D., 1990. Zooplankton structure in Loosdrecht
lakes in relation to the trophic status and the recent resto-
ration measures. Hydrobiologia 191: 173-188.
Gulati, R. D., K. Siewertsen & L. Van Liere, 1991. Carbon
and phosphorus relationships of zooplankton and its seston
food in Loosdrecht lakes. Mem. 1st. ita!. Idrobio!. 48: 279-
298.
Gulati, R. D., A. Ooms-Wilms, O. F. R. Van Tongeren, G.
Postema & K. Siewertsen, 1992. The dynamics and role of
limnetic zooplankton in Loosdrecht lakes (The Nether-
lands): Hydrobiologia 233: 69-86.
Hettling, H., 1985. A hydrological inventarization of the
Loosdrecht lakes area (in Dutch). WQL-report 1985-2.
Higler, L. W. G., S. Parma & L. Van Liere, 1987. Dredging
in Loosdrecht lakes, report of a workshop, (in Dutch).
WQL report 1987-5.
Hofstra J. J. & L. Van Liere, 1992. The state of the environ-
ment of the Loosdrecht lakes. Hydrobiologia 233: 11-20.
Janse,J. H., T. Aldenberg& P. R. G. Kramer, 1992. A math-
ematical model of the phosphorus cycle in Lake Loosdrecht
and simulation of additional measures. Hydrobiologia 233:
119-136.
Keizer, P. & A. J. C. Sinke, 1992. Phosphorus in the sediment
of the Loosdrecht lakes and its implications for lake res-
toration perspectives. Hydrobiologia 233: 39-50.
Kroes, H. W., 1992. The significance of the Loosdrecht lakes
research project for eutrophication policy in The Nether-
lands. Hydrobiologia 233: 165-170.
Ooms-Wilms, A. L., 1991. Ingestion of f1uorescently labelled
bacteria by rotifers and cladocerans in lake Loosdrecht as
measures of bacterivary: preliminary results. Mem. 1st. ita!.
Idrobio!. 48: 269-278.
Otten, J. H., H. J. Gons & M. Rijkeboer, 1992. Dynamics of
phytoplanktonic detritus in a shallow, eutrophic lake (Lake
Loosdrecht, The Netherlands). Hydrobiologia 233: 61-68.
Rijkeboer, M. & H. J. Gons, 1988. The relationship between
oxygen exchange and changes in seston in laboratory scale
enclosures. Verh. int. Ver. Limno!. 23: 756-761.
Rijkeboer, M., F. De Bles & H. J. Gons, 1991. Role ofsestonic
detritus as a P- buffer. Mem. 1st. ita!. Idrobio!. 48: 251-260.
Van Liere, L. (ed), 1986. Flushing as a tool to combat eu-
trophication in L. Vuntus (in Dutch). WQL Report 1986-
3.
Van Liere, L., 1992. Publications of the workgroup WQL
(Water Quality Research Loosdrecht Lakes) 1983-present.
Hydrobiologia 233: 179-186
Van Liere, L. & R. D. Gulati, 1992. Restoration and recov-
ery of shallow eutrophic lake ecosystems in The Nether-
lands: epilogue. Hydrobiologia 233: 283-287.
Van Liere, L. & J. H. Janse, 1992. Restoration and resilience

to recovery of the Lake Loosdrecht ecosystem in relation
to its phosphorus flow. Hydrobiologia 233: 95-104.
Van Liere, L., J. Ebert, W. Kats & J. J. Buyse, 1991. The
water quality of Loosdrecht lakes reviewed. Mem. 1st. ita!.
Idrobio!. 48: 219-232.
Van Liere, L., S. Parma & R. D. Gulati, 1992. Working group
177
Water Quality Research Loosdrecht Lakes: its history,
structure, research programme, and some results. Hydro-
biologia 233: 1-9.
Wilms, A. L., R. D. Gulati & G. Postema, 1990. First attempt
to measure the clearance rate of Anuraeopsis jissa. Verh. int.
Ver. Limno!. 24: 742-744.
Hydrobiologia 233: 179-186, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 179
© 1992 Kluwer Academic Publishers.

Publications of the workgroup WQL (Water Quality Research

Loosdrecht Lakes) 1983-present

Louis Van Liere

Limnological Institute, Rijksstraatweg 6, 3631 AC Nieuwersluis, The Netherlands & National Institute of
Public Health and Environmental Protection, P.O. Box 1, 3720 BA Bilthoven, The Netherlands
(correspondence address)

Abstract

The scientific publications of the workgroup Water Quality research Loosdrecht lakes (WQL) have been
arranged according to their research theme. Between 1983 and 1992 more than 130 publications found
their way to the scientific literature. Much more are submitted, or in preparation.

Introduction A graphical representation of the basic data is

From 1983 onwards there was a steady flow of
scientific publications from the workgroup WQL
(Water Quality research Loosdrecht lakes). It re-
sulted in 131 scientific research papers in the years
after 1983, many more are in preparation. The
number of publications followed the equation:
given in a report by J. Ebert & J. Van Liere:
Basisgegevens WOL 1979-1990 (Basic data
WQL 1979-1990). WQL-report 1991-1.
Publications
Organization and research programme

Number of publications = 31 e - 2.7/(year-1983) Van Liere, L., S. Parma, L. R. Mur, P. Leentvaar

The publications are arranged according to their
discipline below.
Besides these scientific articles, 65 WQL-
reports were written, almost all of them in Dutch.
The results of these reports have been published
at present in various scientific journals. However,
not all results have been submitted for publication
as yet. As for reports on hydrological research
subjects, which have not yet found their way to
the scientific literature, but contain information
that might be of general interest, and which have
English summaries, the reader is referred to the
article of Engelen et al., 1992.
All basic data of WQL have been stored on
MSDOS files. These are also available upon re-
quest to the author, if the purpose for which they
are needed is justifiable, and the 'owners' permit.
& G. B. Engelen, 1984. Loosdrecht lakes resto-
ration project, an introduction. Verh. Int. Ver.
Limnol. 22: 829-834.
Van Liere, L., 1986. Loosdrecht lakes, origin,
eutrophication, restoration and research pro-
gramme. Hydrobiol. Bull. 20: 9-15.
Van Liere, L., 1987. Watter iezze notte mai
responsabilitie, iezze mai collieg. Nuffic Visum
Nieuws (in Dutch), Maart 9-12.
Van Liere, L., S. Parma & R. D. Gulati, 1992.
Working Group Water Quality Research Loos-
drecht Lakes: its history, structure, research pro-
gramme, and some results. Hydrobiologia 233:
1-9.
Van Liere, L., 1992. Publications of the work-
group WQL (Water Quality Research Loosdrecht
Lakes) 1983-present. Hydrobiologia 233: 179-
186.
180
Hydrology
Engelen, G. B., 1986. Interaction of hydrological
systems and eutrophication of the Loosdrecht
lakes. Hydrobiol. Bull. 20: 17-25.
Engelen, G. B., B. F. M. Kal, J. J. Buyse &
F. G. M. Van Pruissen, 1992. The hydrology of
the Loosdrecht lakes area. Hydrobiologia 233:
21-38.
Kal, B. F. M., G. B. Engelen & Th. E. Cap-
penberg, 1984. Loosdrecht lakes restoration
project: Hydrology and physico-chemical charac-
teristics of the lakes. Verh. int. Ver. Limnol. 22:
835-841.
Kal, B. F. M., 1986. Monthly mass balances
for compartments of the Loosdrecht lakes sys-
tem: approach and preliminary results. Hydro-
bioI. Bull. 20: 27-39.
Exchange at the sediment-water interface
Boers, P. C. M., J. W. Th. Bongers, A. G.
Wisselo & Th. E. Cappenberg, 1984. Loosdrecht
lakes restoration project: Sediment phosphorus
distribution and release from the sediments. Verh.
int. Ver. Limnol. 22: 842-847.
Boers, P. C. M. & L. Van Liere, 1985. Het vr-
ijkomen van fosfor uit onderwaterbodems (Re-
lease of phosphorus from sediments, in Dutch).
In: Onderwaterbodem, rol en lot, pp. 92-100.
KNCV-Milieuchemie, Rotterdam.
Boers, P. C. M., O. Van Hese & Th. E. Cap-
penberg, 1985. The role of the sediment in the
phosphorus-cycling dynamics of the Loosdrecht
lakes (Abstract). Antonie Van Leeuwenhoek 51:
449-450.
Boers, P. C. M., O. Van Hese & Th. E. Cap-
penberg, 1985. The exchange of phosphorus be-
tween sediments and water in the Loosdrecht
lakes (Abstract). Antonie Van Leeuwenhoek 51:
575-576.
Boers, P. C. M., 1986. Studying the phospho-
rus release from the Loosdrecht lakes sediments,
using a continuous flow system. Hydrobiol. Bull.
20: 51-50.
Boers, P. C. M. & O. Van Hese, 1988. Phos-
phorus release from the peaty sediments of the
Loosdrecht lakes (The Netherlands), as measured
in a continuous flow system. Wat. Res. 22: 355-
363.
Boers, P. C. M. & J. J. Boon, 1988. Unmask-
ing particulate organic matter in an aquatic eco-
system: the origin and fate of POM in the shal-
low Loosdrecht lakes (The Netherlands). Arch.
Hydrobiol. Beih. Ergebn. Limnol. 31: 27-34.
Boers, P. C. M. & F. De Bles. Ion concentra-
tions in interstitial water: indicator for phospho-
rus release processes and reactions. Wat. Res. 25:
591-598.
Keizer, P., M. C. N. P. Buysman & Th. E.
Cappenberg, 1990. Sorption and release of phos-
phorus in a peaty sediment. Verh. Int. Ver. Lim-
nol. 24: 715-718.
Keizer, P. & A. J. C. Sinke, 1992. Phosphorus
in the sediment of the Loosdrecht lakes and its
implications for lake restoration perspectives.
Hydrobiologia 233: 39-50.
Meuleman, A. F. M. & A. J. C. Sinke. De rol
van sulfaatreduktie in de decompositie van orga-
nisch materiaal (The role of sulfate-reduction in
the decomposition of organic material, In Dutch).
The Utrecht Plant Ecology News Report 10: 23-
38 (1990).
Sinke, A. J. C. & Th. E. Cappenberg, 1988. In-
fluence of bacterial processes on the phosphorus
release from sediments in the eutrophic Loos-
drecht lakes, The Netherlands. Arch. Hydrobiol.
Beih. Ergebn. Limnol. 30: 5-13.
Sinke, A., F. De Bles & Th. E. Cappenberg,
1989. Phosphorus mineralization and release
from sediments in Loosdrecht Lakes (Abstract).
Antonie Van Leeuwenhoek 52: 361-362.
Sinke, A. J. C., A. A. Cornelese, Th. E. Cap-
penberg, 1990. Phosphatase activity in sediments
of the Loosdrecht lakes. Verh. int. Ver. Limnol.
24: 719-721.
Sinke, A. J. c., A. A. Cornelese, P. Keizer,
O. F. R. Van Tongeren & Th. E. Cappenberg,
1990. Mineralization, porewater chemistry and
phosphorus release from peaty sediments in the
eutrophic Loosdrecht lakes, The Netherlands.
Freshwat. BioI. 23: 587-599.
Sweerts, J-P. R. A. & Th. E. Cappenberg.

Methane oxidation at the sediment-water inter-
face of a shallow eutrophic lake (Lake Loos-
drecht) and a deep meso-eutrophic lake (Lake
Vechten). Arch. Hydrobiol. Beih. Ergebn. Lirn-
nol. (in press).
Phytoplankton
Burger-Wiersma, T., M. Veenhuis, H. J. Ko-
rthals, C. C. M. Van De Wiel & L. R. Mur, 1986.
A new prokaryote containing chlorophylls a and
b. Nature 320: 262-264.
Burger-Wiersma, T. & L. R. Mur, 1986. Car-
bohydrate dynamics and growth rate of natural
phytoplankton populations. Hydrobiol. Bull. 20:
87-92.
Burger-Wiersma, T., L. J. Stal & L. R. Mur,
1989. Prochlorothrix hollandica gen. nov., sp. nov.:
a filamentous oxygenic photoautotrophic pro-
caryote containing chlorophylls a and b: assign-
ment to Prochlorotrichaceae fam. nov. order
Prochlorales Florenzano, Balloni and Materassi
1986, with emendation of the ordinal description.
Int. J. Syst. Bacteriol. 39: 250-257.
De Kloet, W. A., P. J. Boesewinkel-De Bruyn
& B. Z. Salome, 1984. The phytoplankton and its
production rates in the Loosdrecht lakes. Verh.
int. Ver. Limnol. 22: 848-852.
De Kloet, W. A., P. J. Boesewinkel-De Bruyn
& L. Breebaart, 1990. Phytoplankton and its pro-
duction rates in the Loosdrecht lakes before and
after P-Ioad reduction. Verh. int. Ver. Limnol. 24:
715-718.
Leentvaar, P. & M. Van Der Meche-Jacobi,
1984. A method for the characterization of sur-
face waters as derived from plankton investiga-
tions in the Loosdrecht lakes. Verh. int. Ver. Lim-
nol. 22: 858-862.
Riegman, R. & L. R. Mur, 1986. Phosphate
uptake kinetics of the natural phytoplankton
population from the Loosdrecht lakes. Limnol.
Oceanogr. 31: 983-988.
Riegman, R., 1986. Een ecologische studie naar
de invloed van fosfaat op de groei van fytoplank-
ton (An ecological study to the influence of phos-
phate on the growth of phytoplankton, in Dutch).
Vakbl. BioI. 66: 121-126.
181
Schweizer, A. & R. E. Heusel. Picoplankton
photosynthesis and diurnal variations in photo-
synthesis-irradiance relationship in a eutrophic
and meso-oligotrophic lake. Hydrobiologia (sub-
mitted).
Sweerts, J-P. R. A., H. J. Gons & M. Rijke-
boer, 1986. Phosphate uptake capacity of sum-
mer phytoplankton of the Loosdrecht lakes in
relation to phosphorus loading and irradiance.
Hydrobiol. Bull. 20: 101-107.
Van Liere, L., L. Van Ballegooijen, W. A. De
Kloet, K. Siewertsen, P. Kouwenhoven & T. AI-
denberg, 1986. Primary production in the various
parts of Loosdrecht lakes. Hydrobiol. Bull. 20:
77-85.
Van Liere, L., L. Breebaart & Y. J. Dullemont,
1989. Determining the relative number of prochlo-
rophytes in lake phytoplankton using epifluores-
cence microscopy. Br. Phycol. J. 24: 391-394.
Seston-epipelon dynamics
Gons, H. J., R. D. Gulati & L. Van Liere, 1986.
The eutrophic Loosdrecht lakes: current ecolog-
ical research and restoration perspectives. Hy-
drobiol. Bull. 20: 61-75.
Gons, H. J., R. Van Keulen & R. Veeningen,
1986. Effects of wind on a shallow lake ecosys-
tem: redistribution of particulates in the Loos-
drecht lakes. Hydrobiol. Bull. 20: 109-120.
Gons, H. J., R. Van Keulen & R. Veeningen,
1986. Deposition and resuspension of seston in
the Loosdrecht lakes (Abstract). Br. Phycol. J.
21: 330.
GOllS, H. J. & M. Rijkeboer, 1990. Algal
growth and loss rates in Lake Loosdrecht: first
evaluation of the roles oflight and wind on a basis
of steady state kinetics. Hydrobiologia 191: 129-
138.
Gons, H. J., 1990. Diatom sedimentation loss
reduction by resuspension in shallow lakes. Verh.
int. Ver. Limnol. 24: 729-733.
Gons, H. J. & M. Rijkeboer, accepted. The
'true' growth efficiency of phytoplankton as influ-
enced by light attenuation and insolation: impli-
cations of the photosynthesis-irradiance relation-
ship. Hydrobiologia.
182
Gons, H. J., Otten,J. H. &M, Rijkeboer, 1991.
The significance of wind resuspension for the pre-
dominance of filamentous cyanobacteria in a
shallow, eutrophic lake. Mem. 1st. ital Idrobio!.
48: 233-249.
Gons, H. J., T. Burger-Wiersma, J. H. Otten &
M. Rijkeboer, 1992. Coupling of phytoplankton
and detritus in a shallow, eutrophic lake (Lake
Loosdrecht, The Netherlands). Hydrobiologia
233: 51-59.
Kromkamp, J. C. & H. J. Gons, 1984. Seston-
epipelon interrelationships in the shallow eutro-
phic Loosdrecht lakes (The Netherlands). Verh.
int. Ver. Limno!. 22: 853-857.
Otten, J. & H. J. Gons, 1990. Aerobic decom-
position of settled algae in a laboratory system:
the impact of resuspension on microbial activity.
Verh. int. Ver. Limno!. 24: 734-737.
Otten, J. H. & H. J. Gons, 1991. Detritus for-
mation from blue-green algae. Mem. 1st. ita!.
Idrobiol; 48: 261-268.
Otten, J. H., H. J. Gons & M. Rijkeboer, 1992.
Dynamics of phytoplankton detritus in a shallow,
eutrophic lake (Lake Loosdrecht, The Nether-
lands). Hydrobiologia 233: 61-68.
Rijkeboer, M., W. A. De Kloet & H. J. Gons,
1986. A comparison of primary production mea-
surements using two laboratory systems with dif-
ferences in light quality. Hydrobio!. Bull. 20:
93-99.
Rijkeboer, M. & H. J. Gons, 1988. The rela-
tionship between oxygen exchange and changes
in seston in laboratory scale enclosures. Verh. int.
Ver. Limno!. 23: 756-761.
Rijkeboer, M., F. De Bles & H. J. Gons, 1989.
Laboratory scale enclosures: concept, construc-
tion and operation. J. Plankton Res. 12: 231-244.
Rijkeboer, M. & H. J. Gons, 1990. Light-
limited algal growth in Lake Loosdrecht: steady
state studies in laboratory scale enclosures. Hy-
drobiologia 191: 241-248.
Rijkeboer, M. & H. J. Gons, 1990. Role of
detritus in the phosphorus balance of Lake Loos-
drecht water. Verh. int. Ver. Limno!. 24: 726-728.
Rijkeboer, M., W. A. De Kloet & H. J. Gons,
1992. Interspecific variation in pigmentation: im-
plications for production estimates for shallow
eutrophic lakes using an incubator. Hydrobiolo-
gla.
Rijkeboer, M., F. De Bles & H. J. Gons, 1991.
Role of sestonic detritus as a P-buffer. Mem. 1st.
ita!. Idrobiol. 48: 251-260.
Waterplants
Best, E. P. H., D. De Vries & A. Reins, 1984. The
macrophytes in the Loosdrecht lakes: a story of
their decline in the course of eutrophication. Verh.
int. Ver. Limnol. 22: 868-875.
Malthus, T. J., E. P. H. Best & A. G. Dekker,
1990. An assessment of the importance of emer-
gent and floating-leaved macrophytes to trophic
status in the Loosdrecht lakes (The Netherlands).
Hydrobiologia 191: 257-264.
Zooplankton dynamics
Dawidowicz, P., Z. M. Gliwicz & R. D. Gulati,
1988. Can Daphnia prevent a blue-green algal
bloom? A laboratory test. Limnologica 19: 21-
26.
Dawidowicz, P., 1990. The effect of Daphnia
on filament length of blue-green algae. Hydrobio-
logia 191: 265-269.
Den Oude, P. J. & R. D. Gulati, 1988. Phos-
phorus and nitrogen excretion rates of zooplank-
ton from the eutrophic Loosdrecht lakes, with
notes on other P sources for phytoplankton re-
quirements. Hydrobiologia 169: 379-390.
Ejsmont-Karabin, J., R. D. Gulati & J. Rooth,
1989. Is food availability the main factor control-
ling the abundance of Euchlanis dilatata lucksiana
Hauer in a shallow, hypertrophic lake? Hydro-
biologia 186/187: 29-34.
Ejsmont-Karabin, J., K. Siewertsen & R. D.
Gulati, accepted. Changes in size, biomass and
production of Euchlanis dilatata lucksiana Haver
during its life span. Hydrobiologia.
Gulati, R. D., 1983. Zooplankton and its graz-
ing indicators of trophic status in Dutch lakes.
Envir. Monit. Ass. 3: 343-354.
Gulati, R. D., 1984. Zooplankton and its graz-
ing as measures oftrophy in the Loosdrecht lakes.
Verh. int. Ver. Limno!. 2: 863-867.

Gulati, RD., K. Siewertsen & G. Postema,
1985. Zooplankton structure and grazing activi-
ties in relation to food quality and concentrations
in Dutch lakes. Arch. Hydrobio!. Beih. Ergebn.
Limno!. 21: 91-102.
Gulati, RD., J. Rooth & J. Ejsmont-Karabin,
1987. A laboratory study of feeding and assimi-
lation in Euchlanis dilatata lucksiana. Hydrobio-
logia 147: 289-296.
Gulati, RD., 1988. Concept of ecosystem
stress and its assessment. In: Ecological assess-
ment of environmental degradation, pollution and
recovery (0. Ravera, ed.) pp.55-80. Elseviers
Publishing Company, Amsterdam.
Gulati, RD., 1988. Concept of stress and re-
covery in aquatic ecosystems. In: Ecological as-
sessment of environmental degradation, pollution
and recovery (0. Ravera, ed.) pp. 81-119. Elsevi-
ers Publishing Company, Amsterdam.
Gulati, RD., C. W. M. Bodar, A. L. G.
Schuurmans,J. A. J. Faber & D. I. Zandee, 1988.
Effects of cadmium exposure on feeding of fresh-
water planktonic crustaceans. Compo Biochem.
Physio!. 90C: 335-340.
Gulati, RD., 1990. Zooplankton structure in
Loosdrecht lakes in relation to the trophic status
and the recent restoration measures. Hydrobio-
logia 191: 173-188.
Gulati, RD., J. Ejsmont-Karabin, J. Rooth &
K. Siewertsen, 1989. A laboratory study of phos-
phorus and nitrogen excretion of Euchlanis dila-
tata lucksiana. Hydrobiologia 186/187: 347-354.
Gulati, RD., C. Vuik, K. Siewertsen & G.
Postema, 1990. Clearance rates of Bosmina spe-
cies in response to changes in trophy and food
concentration. Verh. int. Ver. Limno!. 24: 722-
725.
Gulati, RD., L. Van Liere & K. Siewertsen,
1991. The Loosdrecht lakes system: Man's role in
its creation, perturbation and rehabilitation. In:
O. Ravera (ed.) Terrestrial and aquatic ecosys-
tems; perturbation and recovery. Ellis Horwood
Ltd., Chicester. pp. 593-606.
Gulati, RD., K. Siewertsen & L. Van Liere,
1991. Carbon and phosphorus relationships of
zooplankton and its seston food in Loosdrecht
lakes. Mem. 1st. ita!. Idrobio!. 48: 279-298.
183
Gulati, RD., J. Ejsmont-Karabin & G.
Postema, accepted. Feeding and assimilation in
Euchlanis on filamentous cyanobacteria and a
prochlorophyte. Hydrobiologia.
Gulati, R. D., Role of filter-feeding crustacean
zooplankton in Dutch lakes of varying depth and
trophy. Proc. of the 1990 Int. Crustacean Conf.,
Brisbane, Australia. Mem. Queensland Museum
31: 387.
Gulati, R D., A. L. Ooms-Wilms, O. F. R
Van Tongeren, G. Postema & K. Siewertsen,
1992. The dynamics and role of limnetic zoop-
lankton in Loosdrecht lakes (The Netherlands).
Hydrobiologia 233: 69-86.
Irvine, K., Differential feeding behaviour be-
tween the dominant cladocera as an explanation
of zooplankton community structure in the Loos-
drecht lakes. Hydrobio!. Bull. 20: 121-134.
Wilms, A. L., RD. Gulati & G. Postema,
1990. First attempt to measure the clearance rate
of Anuraeopsis fissa. Verh. int. Ver. Limno!. 24:
742-744.
Wilms, A. L., 1991. Ingestion of fluorescently
labelled bacteria by rotifers and cladocerans in
lake Loosdrecht as measures of bacterivory: pre-
liminary results. Mem. 1st. ita!. Idrobiol. 48: 269-
278.
Wilms, A. L., G. Postema & R D. Gulati, ac-
cepted . Clearance rate of bacteria by the rotifer
Filnia longisita (EHRB) measured by using three
tracer techniques. Hydrobiologia.
Fish
Lammens, E. H. R R, N. Boesewinkel-De
Bruyn, H. Hoogveld & E. Van Donk, 1992.
Phosphorus-load, phytoplankton, zooplankton
and fish stock in Loosdrecht Lake and Tjeuke-
meer: confounding effects of predation and food
availability. Hydrobiologia 233: 87-94.
Van Densen, W. L. T., C. Dijkers & R Veer-
man, 1986. Fish community of the Loosdrecht
lakes and the perspective for biomanipulation.
Hydrobiol. Bull. 20: 147-163.
184
Mathematical modelling
Aldenberg, T. & J. S. Peters, 1990. On relating
empirical waterquality diagrams and plankton-
dynamical models: the Sample methodology ap-
plied to a drinking water storage reservoir. Arch.
Hydrobiol. Beih. Ergebn. Limnol. 33: 893-911.
Auer, M. T. & M. Rijkeboer, 1990. Modelling
biomass and oxygen dynamics in Prochlorothrix
hollandica in Laboratory-Scale Enclosures. Verh.
int. Ver. Limnol. 24: 738-741.
Janse, J. H. & T. Aldenberg, 1990. Modelling
phosphorus fluxes in the hypertrophic Loosdrecht
lakes. Hydrobiol. Bull. 24: 69-89.
Janse, J. H., T. Aldenberg & P. R G. Kramer,
1992. A mathematical model of the phosphorus
cycle in Lake Loosdrecht and simulation of ad-
ditional measures. Hydrobiologia 233: 119-136.
Kouwenhoven, P. & T. Aldenberg, 1986. A
first step in modelling plankton growth in the
Loosdrecht lakes. Hydrobiol. Bull. 20: 135-145.
Van Tongeren, O. F. R, L. Van Liere, RD.
Gulati, G. Postema & P. J. Boesewinkel-De
Bruyn, 1992. Multivariate analysis of the plank-
ton communities in the Loosdrecht lakes: rela-
tionship with the chemical and physical environ-
ment. Hydrobiologia 233: 105-117.
Remote sensing
Dekker, A. G. & E. Seyhan, 1987. Operational
remote sensing analysis techniques applied to
spatial and temporal analysis of waterquality in a
eutrophied lake system in The Netherlands (Case
study: Loosdrecht lakes). In: 4th Australian Re-
mote Sensing Conf./Adelaide August 1987. Ed-
ited by D. Bruce, p. 690-712.
Dekker, A. G. & E. Seyhan, 1988. Application
of remote sensing techniques for detecting the
water quality of eutrophic shallow lake systems:
case study of the Loosdrecht lakes. Int. J. of Re-
mote Sensing 9: 1761-1773.
Dekker, A. G., T. J. Malthus & E. Seyhan,
1989. Quantitative modelling of inland water
quality for high resolution MSS-systems. IG-
ARSS'89, 12th Canadian Symposium on Remote
Sensing. RS: An economic tool for the nineties,
pp.2804-2807.
Dekker, A. G., L. Van Liere, E. Seyhan & T. J.
Malthus, 1990. Een helder beeld van troebel
water; remote sensing en waterkwaliteit (A clear
picture of turbid water; remote sensing and water
quality), GEA 123: 86-90.
Dekker, A. G., T. J. Malthus & E. Seyhan,
1990. Quantitative modelling of inland waterqual-
ity for high resolution MSS-systems. IEEE
Transactions on Geoscience and Remote Sens-
ing 29: 89-95.
Dekker, A. G., T. J. Malthus & E. Seyhan,
1990. An inland water quality bandset for the
CAESAR system based on spectral signature
analysis. Proc. Int. Symp. on Remote Sensing
and Water Resources, Enschede, pp. 597-606.
Dekker, A. G., T. J. Malthus, M. M. Wijnen &
E. Seyhan, 1992. Remote sensing as a tool for
assessing water quality in Loosdrecht lakes. Hy-
drobiologia 233: 137-159.
Malthus, T. J. & A. G. Dekker, 1990. Spectral
light attenuation in a hypertrophic lake system
(Loosdrecht lakes, The Netherlands). Verh. int.
Ver. Limnol. 24: 711-714.
Seyhan, E. & A. G. Dekker, 1985. Application
of remote sensing techniques for water quality
monitoring. Hydrobiol. Bull. 20: 41-50.
Water management
De Ruiter, M. A., L. Van Liere, B. F. M. Kal &
J. J. Buyse, 1988. Worden de Loosdrechtse Plas-
sen weer helder? (Will Loosdrecht lakes ever be-
come clear again?, in Dutch) H 2 0 21: 482-485.
De Ruiter, M. A., 1992. Water quality research
in the Loosdrecht lakes: proposals for the follow-
up restoration measures. Hydrobiologia 233:
161-163.
Leentvaar, P., 1984. Water transfer in a lake
complex. In: Technical Documents in Hydrology
(Ed. J. R Card), pp. 135-145. Hydro-Environ-
mental Indices: A review and evaluation of their
use in the assessment of the environmental im-
pact of water projects. Unesco Working Group
on IHP-II Project A 3.2, Paris.

Mur, L. R., L. Van Liere, M. A. De Ruiter,
1989. Scenario's voor het herstel van de Utrechtse
Vechtplassen (Scenario's to restore the lakes bor-
dering River Vecht in Utrecht). Comrnissie voor
Hydrologisch Onderzoek TNO. Rapporten en
Nota's 22: 59-68.
Van Der Veen, c., A. Graveland & W. Kats,
1987. Coagulation of two different kinds of sur-
face water before inlet into lakes to improve the
self-purification process. Wat. Sci. Technol. 19:
803-812.
Van Donk, E., M. P. Grimm, R. D. Gulati,
P. G. M. Heuts, W. A. De Kloet & L. Van Liere,
1990. First attempts to apply whole lake foodweb
manipulation on a large scale in The Netherlands.
Hydrobiologia 200/201: 291-301.
General limnology, integration
Giussani, G., L. Van Liere & B. Moss, 1991.
Ecosystem Research in Freshwater Environment
Recovery. Mem. 1st. Ital. Idrobiol. 48: (in press).
Gulati, R. D. & L. Van Liere, 1992. Water
quality research in Loosdrecht lakes: the salient
features. Hydrobiologia 233: 171-177.
Hofstra, J. J., J. H. Janse & L. Van Liere, 1991.
De milieutoestand van de Loosdrechtse Plassen.
Toepassing van de presentatietechniek
AMOEBE en het eutrofieringsmodel PCLOOS
(The status of the environment of Loosdrecht
lakes. Application of the AMOEBE technique
and the eutrophication model PCLOOS, in
Dutch). H 2 0 10: 260-265.
Hofstra, J. J. & L. Van Liere, 1992. The state
of the environment of the Loosdrecht lakes. Hy-
drobiologia 233: 11-20.
Kroes, H. W., 1992. The significance of the
Loosdrecht lakes research project for eutrophi-
cation policy in The Netherlands. Hydrobiologia
233: 165-170.
Loogman, J. G. & L. Van Liere (eds), 1986.
Proceedings of the WQL-symposium: Restora-
tion of shallow lake ecosystems, with emphasis
on Loosdrecht lakes. Hydrobiol. Bull. 20: 1-259.
The Loosdrecht lakes Restoration Project,
1983. In: Institutes of the Royal Netherlands
185
Academy of Arts and Sciences, Progress Report
1982. Verh. Kon. Ned. Akad. Wet. Afd. Natuurk.
Tweede reeks, vol 82: 36-52.
The Loosdrecht lakes Restoration Project,
1984. In: Institutes of the Royal Netherlands
Academy of Arts and Sciences, Progress Report
1983. Verh. Kon. Ned. Akad. Wet., Tweede
reeks, vol. 82: 25-47.
The Loosdrecht lakes Restoration Project,
1985. In: Institutes of the Royal Netherlands
Academy of Arts and Sciences, Progress Report
1984. pp. 23-52.
The Loosdrecht lakes Restoration Project,
1986. In: Institutes of the Royal Netherlands
Academy of Arts and Sciences. Progress Report
1985, pp. 9-33.
The Loosdrecht lakes Restoration Project,
1987. In: Limnological Institute, Progress Report
1986. Koninklijke Nederlandse Akademie van
Wetenschappen, Amsterdam, p. 17-37.
The Loosdrecht lakes Restoration Project,
1988. In: S. Parma, R. D. Gulati & B. Z. Salome
(eds.), Limnological Institute, Progress Report
1987, pp. 17-38.
The Loosdrecht lakes Restoration Project,
1989. In: S. Parma, R. D. Gulati & B. Z. Salome
(eds.). Limnological Institute, Progress Report
1988, pp. 38-54.
The Loosdrecht lakes Restoration Project,
1990. In: S. Parma & R. D. Gulati (eds.). Lim-
nological Institute, Progress Report 1989, pp. 42-
55.
The Loosdrecht lakes Restoration Project,
1991. In: S. Parma & R. D. Gulati (eds.). Lim-
nological Institute, Progress Report 1990, pp. 44-
56.
Van Liere, L., 1985. The eutrophication of
Loosdrecht lakes (Abstract). Antonie Van Leeu-
wenhoek 51: 450-451.
Van Liere, L., 1986. Concluding remarks (at
the First WQL Symposium, Amsterdam 1985),
with emphasis on remarks. Hydrobiol. Bull. 20:
257-259.
Van Liere, L., 1986. Water Quality Research
Loosdrecht lakes; studying and modelling the im
pact of the internal nutrient cycle. In: Environ-
ment and Quality of Life, Third Environmental
186
Research Programme, Commission of the Euro-
pean Communities.
Van Liere, L., R. D. Gulati & E. H. R. R.
Lammens, 1989. Van voedingsstof tot vis en van
vis tot voedingsstof (From food to fish and from
fish to food, in Dutch). Landschap 6: 33-46.
Van Liere, L., R. M. M. Roijackers & P. J. T.
Verstraelen (eds.), 1989. Integraal Waterbeheer in
het Goois/Utrechts stuwwallen- en plassengebied
(Integral water management in the lake area of
the sandridge of Gooi/Utrecht, in Dutch). Com-
missie voor Hydrologisch Onderzoek TNO. Rap-
porten en Nota's 22: 1-299.
Van Liere, L., L. Breebaart, W. Kats & J. J.
Buyse, 1989. De waterkwaliteit in het Loos-
drechtse Plassengebied (Water quality of the
Loosdrecht lakes area, in Dutch). Commissie
voor Hydrologisch Onderzoek TNO. Rapporten
en Nota's 22: 265-279.
Van Liere, L., R. D. Gulati, F. G. Wortelboer
& E. H. R. R. Lammens, 1990. Phosphorus dy-
namics following restoration measures in Loos-
drecht lakes. Hydrobiologia 191: 89-96.
Van Liere, L., O. F. R. Van Tongeren, L. Bree-
baart & W. Kats, 1990. Trends in chlorophyll and
total phosphorus in Loosdrecht lakes, The Neth-
erlands. Verh. int. Ver. Limno!. 24: 707-712.
Van Liere, L. & G. Giussani, 1991. Ecosystem
Research in Freshwater Environment Recovery,
an introduction. Mem. 1st. ita!. Idrobio!. 48:
3-8.
Van Liere, L., J. Ebert, W. Kats & J. J. Buyse,
1991. The waterquality of the Loosdrecht lakes
reviewed. Mem. 1st. ita!. Idrobio!. 48: 219-232.
Van Liere, L. & J. H. J anse, 1992. Restoration
and resilience to recovery of the Lake Loosdrecht
ecosystem in relation to its phosphorus flow. Hy-
drobiologia 233: 95-104.
Van Liere, L. & R. D. Gulati (eds), 1992. Res-
toration and recovery of shallow eutrophic lake
ecosystems in The Netherlands: epilogue. Hydro-
biologia 233: 283-287.
Hydrobiologia 233: 187-196, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 187
© 1992 Kluwer Academic Publishers.

The influence of flushing on nutrient dynamics, composition and densities

of algae and transparency in Veluwemeer, The Netherlands

E.Jagtman,1,3 D. T. Van der Molen 2 & S. Vermijl

1 Directorate Flevoland-R WS, P.O. Box 600, 8200 AP Lelystad, The Netherlands; 2 RIZA, P.O. Box 17,
8200 AA Lelystad, The Netherlands; 3 Current address R WS-Directorate Noord-Holland, P. O. Box
3119,2001 DC Haarlem, The Netherlands

Key words: eutrophication, lake restoration, flushing, Veluwemeer, algal species, transparency,
phosphorus, nitrogen

Abstract

Total phosphorus and chlorophyll decreased significantly after reduction of the external phosphorus
loading and the start of flushing Veluwemeer with polder water in 1979.
Flushing of Veluwemeer has had a large impact on nutrient dynamics. Especially in the first winter,
dilution was the main cause of changes in water quality. On a longer term the increase of the inactiva-
tion of phosphorus in sediments is important. Oscillatoria agardhii has been brought to the margins of
its habitat. Three successive cold winters were an additional causal factor in the disappearance of
Oscillatoria agardhii and the dominance of diatoms and green algae from 1985 onwards.
Due to higher detritus and inorganic suspended matter concentrations transparency increased less
than expected. Since 1985 chlorophyll only contributes for a small percentage to the transparency.
In the present situation further improvement of the water quality of Veluwemeer is questionable, as
the phosphorus concentration in the lake and the polder water is almost the same. Therefore it is
recommanded to shift flushing operations, at least in the winter period, from Veluwemeer towards
Wolderwijd.

Introduction (1986) also reported several cases in which flush-

Veluwemeer, the largest lake in a chain of shallow
lakes situated in the centre of The Netherlands,
is of great importance to recreation and nature. In
the past twenty years eutrophication has caused
severe water quality problems. Various measures
have been taken in the past to reverse this unde-
sirable situation (Hosper & Meijer, 1986). One of
these measures was flushing of Veluwemeer and
it proved to be extremely effective. Cooke et al.
ing achieved improvement in eutrophic lakes.
This paper examines the impact of flushing on
nutrient dynamics, the composition and densities
of algae species and transparency in Veluwemeer.
It describes and discusses the basic mechanism
on which the success of flushing is based, and
evaluates the effects of the present flushing regime
in relation to the water management of Veluwe-
meer and its neighbouring lakes in general.
188
Material and methods
Description of the area
The lakes Wolderwijd, Veluwemeer, Dronter-
meer, Vossemeer together form the Randmeren
(Fig. 1). The lakes have a typical hydrology due
to the special location between the 'old' land and
the new made polder, lying ca 5 m below mean
sea level. On one side of the lakes there is an
inflow of groundwater and a number of streams,
whereas on the other side there is a loss of water
by infiltration into the polders.
Veluwemeer is the largest lake of the Randme-
ren, with a surface area 32.8 km 2 , an average
depth of 1.35 m and a retention time of the water
of about 4 months. In the first years after its
creation in 1956, the water was clear and there
was a rich animal and plant life. In the second
half of the sixties the lake detoriated rapidly, be-
cause of a sharp increase of the phosphorus load-
ing. In the early seventies the lake was already
characterized by cyanobacteria. According to
Berger & Bij de Vaate (1983) the cyanobacterium
Oscillatoria agardhii Gomont was the dominating
species. The water became very turbid, Secchi
disc visibility was limited to 0.2 m. The other lakes
developed in a similar way.
The eutrophication ofthe Randmeren was con-
trolled by reducing the external phosphorus load-
ing. Phosphorus elimination was introduced at
the sewage treatment plants discharging into
Veluwemeer
Fig. 1. Sketch of Veluwemeer and bordering lakes.
Drontermeer in 1972 and Veluwemeer in 1979. In
the latter lake the measures lead to a decrease of
the external phosphorus loading from 3 to
1 gP m - 2 Y- 1 (PER, 1986). The remaining phos-
phorus loading is caused 30% each by the dis-
charge of effluents of the sewage treatment plant
and several small streams and by rainfall, seepage
and some minor discharges. Nutrient balance
studies in Wolderwijd revealed that sediment re-
lease of phosphorus occurred during the summer
season (PER, 1982), and such a release seemed
to be also important in Veluwemeer, in view of the
large increase in SRP during the first months of
the growing season (Fig. 2).
Monitoring of the water quality
Routine fortnightly measurements in the field were
carried out by the Institute for Inland Water
Management and Waste Water Treatment, Lelys-
tad, The Netherlands. The lake was sampled at
representative locations (PER, 1986). The field
measurements included water temperature, Sec-
chi-disc transparency and pH. Chlorophyll a was
determined spectrophotometrically after ethanol
extraction (Moed, 1973). The species composi-
tion of algae was measured microscopically, using
counting chambers, on samples fixed with iodine-
formaline. Inorganic suspended solids were de-
-&- SAP - - Total-P
1.0
0.9
0.8
M
0.7
~ 0.6
§ 0.5
~ 0.4
~ 0.3
0.2
0.1
M J J A SON DJFM A M J J A S
1978 1979
Fig. 2. Total- and soluble reactive phosphorus concentrations
in Veluwemeer before the start of the winter flushing.
 189

termined gravimetrically as loss on ignition at
550°C of the seston (Dankers & Laane, 1983),
after drying at 75°C. All nutrients were analyzed
on a HP 7673A auto analyzer. Soluble reactive
phosphorus (SRP) and total phosphorus, after
digestion with a sulphuric acid/persulphate mix-
ture, are analyzed according to the modified mo-
lybdenum blue method (Murphy & Riley, 1962).
Kjeldahl-nitrogen was measured as ammonium
after digestion with potassiumsulphate/sulphuric
acid and Wieninger selenium. Ammonium was
measured spectrophotometrically after complex-
ation with salicylate and chlorine. Nitrate was
reduced to nitrite, using cadmium as electron
donor and nitrite was complexed with sulfanil-
amide and iX-naphtylethleendiamine-dihy-
drochloride before measurement (modification of
the method described by Afghan & Ryan (1975».
Calcium was analyzed, also on the auto analyser,
after complexation with cresolphtaleine in a di-
ethylamine buffer. Particulate calcium is deter-
mined by the difference between total calcium
and calcium after filtration with 0.45,um Milli-
pore filters.
It was assumed that during the winter months a
reduction in the phosphorus concentration and a
satisfactory reduction in algae biomass would en-
able to prevent the recurrent blooms of cyano-
bacteria. In order to achieve this objective a trial
flushing of Veluwemeer was carried out. During
the winter of 1979/1980 pumping station Lovink
supplied an excess amount of high quality water
from the adjacent polder Flevoland into the lake.
The daily flushing rate was ca 0.01 d - '. This is
based on a monthly flushing discharge of 15
106 m 3 , an average winter situation for Veluwe-
meer. The polder water has relatively low con-
centrations of total-P (0.09 gP m - 3) and algae
(14 mg chlorophyll a m - 3), but is rich in calcium
(140 g Ca m - 3) and nitrate (3.0 gN m - 3 during
the winter). As these winter flushings were suc-
cessful, it was decided supplementary summer
flushings were carried out from 1985 on (Table 1).
This was aimed at preventing a 'concentration
effect' (increasing nutrient concentrations due to
infiltration and evapo-transpiration) and to dilute
the phosphorus released from sediments.

Results

Flushing Situation up to 1985

It is important to define the concept 'flushing' During the winter 1979/1980 89 106 m 3 flushing
carefully, because during the winter months, when water was supplied to the lake and the total-P
the excess water from the surrounding areas is
pumped into the lake, there may already be a Table I. Flushing discharges in Veluwemeer, corrected for
natural form of flushing. Our definition of 'flush- summer supplements of ca 6.0 106 m 3 per month (winter dis-
ing' is to deliberately supply more water than is charges refer to 1979/1980, 1980/1981 etc.).
necessary for maintaining the water level or a
Year Summer discharge Winter discharge
water reserve. The flushing discharge of Veluwe- (10 6 m 3 ) (10 6 m 3 )
meer is calculated as the difference between the
discharge supplied by the Lovink pumping sta- 1979 19.179 89.249
tion and the discharge to the bordering Wolder- 1980 14.596 79.075
1981 6.289 84.402
wijd (Fig. 1). The percentage oforiginal lake water
1982 19.425 82.797
can be calculated as: 1983 18.693 81.428
1984 69.258 116.117
percentage original lake water = 1OO(e - (Qd/QI» 1985 28.157 105.175
1986 101.270 68.824
1987 29.951 72.380
in which Qd is discharge volume in m 3 , and
1988 86.081 58.814
Ql is lake volume in m3 .
 190

concentration in the Veluwemeer decreased dra- tration on a higher level than the polder water
matically up to March 1980, when the concen- implies that flushing is counteracted by algal
tration approximated the concentration in the growth. Figure 5 shows growth curves for a num-
polder water (Fig. 3). Soluble reactive phospho- ber of algal groups derived from the algal growth
rus (SRP) decreased concomitantly to values model BLOOM (Los, 1991) and an average
around the detection limit. The chlorophyll a con- flushing rate. The curves are a function of the
centration also dropped dramatically, but already water temperature and have been calculated at
in January 1980 stabilization set in at a level of optimum light- and nutrient situation. At water
about 60 mg m - 3, which is above the average temperatures below 2 C, the growth rate of algae
0

concentration in the water used for flushing. The will not offset the effect of lake's flushing, so that
particulate-N concentration in Veluwemeer the algae will continue to be washed out. How-
dropped from 4.9 gN m - 3 when flushing was ever the net wash out effect will be 20 to 30% per
started to less than 2.0 gN m - 3 in May 1980 month. At water temperature above 2 C algae 0

(Fig. 4). can increase in numbers effectively, irrespective

The main cause of improvement of Veluwe-
meer water is the dilution with supply water. The - - - N02+ I'.H4 --_. N-Kjeld - N-tot

concentrations of total-P, Kjeldahl-N and chlo- N03

rophyll a in Veluwemeer were very much related 6,-----------------,

to the fluctuations in the quality of the supply 5
water (,-2> 0.95). If the dilution of the water is the 1'1
driving force for quality improvement, the nutri- ~ 4
,
ent concentration in Veluwemeer will start resem- 3 I'

bling the nutrient concentration in the supplied ~ " 1

.. J
\ 1\
\.
~ 2 /",.. __ J' .... _ /,', ... ""'_/
water more and more, as more polder water is I • ,_~

I \
supplied. Although this is the case for the total- ~I "
P concentration in Veluwemeer, the situation for o L...:."""-="""--.........-....l.,-LJ/'---".L..--'--L-'d..-"=_.....'''''-'-.-..I--l
chlorophyll a and nitrogen is more complex. JASONDJFMAMJJA
1979 1980
The stabilization of the chlorophyll a concen-
Fig. 4. Nitrogen concentrations in Veluwemeer, before and
after the winter flushing of 1979/1980.
- - Chlorophyll ---. Total-P

500 0.6

400
! Start flushing
0.5 0.7

("J
'I'0 0.6
I ("J

a
l; 0.4 I
(Jl 0.5
E 300 2
~ 0.4
0.3 a.
>.
~ 200 e
I
'"
l:
i' 0.3
0 0.2 0 '"
~

~~
f-
is 0.2
0
100 H.$LJPply.p' . ~ 0.1
H~LJPp'Iy'gt11 .
~ 0.0
oM A SON 0 F M A M
M
J J J J -0.1
1979 1980 0 1.5 2 3 4 5 6 7 8 9 10
Temperahre oC
Fig. 3. The effects of the first winter flushing (1979/1980) on
chlorophyll a and total-phosphorus concentrations (the hor- Fig. 5. Growth rates of the various algae species groups as
izontal dotted lines represent the concentrations in the sup- function of the water temperature (from: Los, 1991) and the
ply water). flushing rate in Veluwemeer.
 191

of flushing. Thus, it is reasonable to assume that
from early ~pring onwards (February/March)
algal growth could prevent the chlorophyll a con-
centration from decreasing to the level of the sup-
ply water.
Comparing the concentrations of nitrogen
compounds in Veluwemeer and the supply water,
it appears that at the start of flushing the nitrogen
in the lake water predominantly contains partic-
ulate-N and concentrations of nitrate and ammo-
nium are low. The supply water, however, is rich
in ammonium (1 gN m - 3) and nitrate
(3 gN m - 3). When the two water types are mixed,
the chlorophyll a concentration and the Kjeldahl-
N concentration decrease, but concentrations of
nitrate an ammonium rise rapidly to 2.5 gN m - 3
ably. Figure 7 shows an increase in total-P con-
centration again in March 1981, presumably be-
cause of sediment release, but from 1981 onwards
the phosphorus release from the sediment de-
creased and stabilized at a very low level from
1985 onwards (Fig. 8). The sediment release, cal-
culated by the model SED (Van der Molen, 1991)
and measured by continuous flow experiments
according to Boers & Van Hese (1988), decreased
proportionally more than the external load be-
cause of a number of processes, related to flush-
ing:
- The reduced organic matter production and
- Chlorophyll Total-P

and 0.5 gN m - 3 respectively, in February and 500 0.5

March (Fig. 4). In summer a decrease in nitrate
1
Start flushing
and ammonium sets in, probably because of den- 400 0.4
itrification at rising temperatures. Algal growth l'l
I l'l
exhausts the dissolved forms of nitrogen, with the E; I

result that in May 1980 all nitrogen in Veluwe- '"

E 300 0.3 E
dt

meer occurred again in particulate form (Kjeldahl- >. q.

£.0 200 0.2 iii
N). 15 0
f-
The dilution of Veluwemeer, started in 1979, (5
100
gives rise to a permanent improvement of the
water quality (Fig. 6). The flushing in winter 'H?LJPply Chi

1980/1981 had, therefore, less impact than the 0

M ... 0 N 0 F M
0.0

flushing in 1979/1980 (Fig. 7). The differences in 1980 1981

nutrient concentrations between supply water and Fig. 7. The effects of the second winter flushing (1980/1981)
lake water had already been reduced consider- on concentrations of chlorophyll a and total-P (the horizon-
tal dotted lines represent the concentrations in the supply
water).
- - Mean-surmer
total-P

0.75,..-----------------, 15,------------------.,

1 Start Flushing
-b
N
a. 0.50 10
&
I
II
'"
E

3ltil

!
0.25 5
Ql

~
Q.

0.00 L-..L--'----'-----'-~l_=-=:lIl==lIl:==lI>==lIl==-=:lIl==lIl:==lb
1975 1977 1979 1981 1983 1985 1987 1989

Fig. 6. Summer average total- and soluble reactive phospho- Fig. 8. Release of phosphorus in Veluwemeer based on model
rus during 1975-1989. SED (from: Van der Molen, 1991).
192

the considerable supply of nitrate in the polder ~ Cyano- IZ:ZI Green ~ Diatoms
bacteria algae
water have caused improvement of the sedi-

1
120000 .--------------------~
ment oxidation.
Moreover, the supply of calcium and to a lesser Start flushing

extent the supply of iron with the polder water 90000

(Brinkman & Van Raaphorst, 1986; Van det

Molen, 1990) have possibly led to an increased
60000
fixation of phosphorus in the bottom and a
decreased release rate.
30000

Further, flushing has caused a decrease in the

biovolume of the cyanobacteria (Fig. 9), while o1917 197. 1979 tgeo ,., '982 1983 1984 1985 1986 1987 1988 1989
the number of algae remained unchanged high
(Fig. 10). There was no distinct shift in the Fig. 10. Stacked presentation of the summer average of the
composition of the algae species up to 1985. densities of algae in Veluwemeer per species group over the
This means that flushing did not initially lead period 1979-1989.
to the realization of the objective to prevent the
cyanobacteria dominance. However, the de- - Total -- -- 1/Secchi
crease in biovolume resulted in an increased biovolLlTle depth

transparency (Fig. 11). I

M 250 7.5

~
200

Situation since 1985 ~

>
0 I
I
, --\
5.0
~
:c 150
I
I £;

The cyanobacteria disappeared from 1985 on- §

I
~
:E
..
100 u
wards, although the phosphorus concentrations "0
:!lto. 2.5 u

had already decreased considerable. In Fig. 12 !!2

II;
50
~
the number of the cyanobacteria, as a percentage
of the total number of algae, have been plotted, as j 0
1977 1979 1981 1983 1985 1987
0.0
1989

Fig. 11. Summer averages of the biovolume of the algae and

the reciprocal Secchi depth values in Veluwemeer for 1977-
~ Cyano- IZ:zl Green ~ Diatoms
bacteria algae 1989.
250 r------------------,
they occurred in the growing seasons (April-
October) of 1980 to 1989 in Veluwemeer and
Wolderwijd respectively. During the period 1985
to 1987 three successive cold winters occurred.
The number reduction was probably influenced
by the reduced productivity in winter (low tem-
peratures), poor light conditions under the ice
cover (Adams et al., 1984) as well as by increased
sedimentation under the ice cover. During a sim-
1978 1979 1980 1981 '982 1983 1984 1sa8S ,98e 1987 1988 1989 ilar cold winter in 1978/1979 the cyanobacteria
Fig. 9. Stacked presentation of the summer average of the
dominance was not broken, but only temporarily
biovolume of algae in Veluwemeer per species group over the reduced under prolonged ice cover (Fig. 13).
period 1977-1989. The following qualitative model can be devel-
 193

~ Wolderwijd IIIIIIIII Veluwemeer during ice coverage and as soon as the water
8 100
temperature exceeds the critical value of 2 C di-0

jij atoms and green algae will take up all the avail-
~ 80 able phosphorus. This causes a further deterio-
'0 ration in the competitive position of the
60 cyanobacteria. Table 2 shows this process after
the winter of 1984/1985. The direct contribution
40 of flushing to the reduction of cyanobacteria in
Veluwemeer in the winter of 1984/1985 is limited.
20
Although phosphorus levels still were relatively
low, the number of cyanobacteria started to in-
o
19801981 19821983198419851986198719881989 crease again after the mild winter of 1987/1988.
Low numbers of cyanobacteria in 1989 in Velu-
Fig. 12. Percentages of the Oscillatoria agardhii in the algae wemeer, compared to Wolderwijd, are caused by
population in Veluwemeer and Wolderwijd during the period low availability of phosphorus.
1980-1989.
During the summer another mechanism may
play a role in the disappearance of Oscillatoria
'" Wolderwijd + Veluwemeer
agardhii. The photosynthesis capacity of
E 100
~ + Oscillatoria agardhii gradually decreases and may
Q
90
even be reduced to zero at higher light intensities
§'" 80
'" '+" (Loogman, 1982). The increased transparency
11_ 70 '" +
'+"
t--8 60 ICE + during summer from 1985 onwards therefore con-
" 16
~ ~ 50
'"
+ '" +
tributes to unfavourable conditions for Oscilla-
"~
TIt: 40 '" +
toria agardhii.
'" 30 '"
0
'"
'0 20 '+"
!
10
+

0 '" '" Effects on lake-transparency

0 50 100 150 200 250 300 350 400

1979: dayrunber The transparency in Veluwemeer increased, but

Fig. 13. Influence of a period of prolonged ice coverage on the less than expected. Figure 11 showed a continu-
numbers of cyanobacteria in Veluwemeer and Wolderwijd ing decrease of the total biovolume since 1985,
during the winter of 1978/1979.
whilst stagnation occurs in the (reciprocal)
Secchi-depth values. The improvement of trans-
oped to explain the shift in algae composItIOn
during the period 1985-1987 in Veluwemeer and
Table 2. Number of algae per ml in Veluwemeer during the
Wolderwijd. Oscillatoria agardhii has a relatively
winter period 1984/1985
low growth rate, particularly at high light inten-
sities (Loogman, 1982), but it can maintain its Month Cyano- Green Diatoms
position because the loss rates are low also. This bacteria algae
species is hardly eaten by zooplankton and its
Oct 1984 52775 4238 1038
sedimentation rate is low (Reynolds & Walsby,
Nov. 1984 33038 2138 600
1975). Therefore, the number of cells at the be- Dec 1984 29700 3075 375
ginning of a growing season are important for the Jan 1985 10650 1050 0
dominance in the growing season. The combina- Feb 1985 54575 5138 75
tion of a severe winter and scarcity of phospho- Mar 1985 2425 2160 1250
Apr 1985 2871 16019 50699
rus has the following effect: the densities of
May 1985 4156 17265 32840
Oscillatoria agardhii will decrease dramatically
194
parency compared with the decrease in chloro-
phyll a is also much lower than expected. The
change in the algae species composition and the
role of other fraction in the water may have con-
tributed to this less than expected increase in
transparency.
Since 1985 the species composition has shifted
from cyanobacteria towards diatoms and green
algae. Per unit dry weight diatoms and green algae
contribute less to the extinction of light than cy-
anobacteria; they have a lower specific extinction
(Los, 1991). The biovolume/dry weight ratio is ca
the same for both groups, therefore further trans-
parency improvement is expected but not ob-
served. Apparently, suspended materials other
than live algae contribute more to the transpar-
ency.
Detritus has a delayed reaction on the decrease
in chlorophyll or biovolume, because a fraction of
dead algae decomposes very slowly. Lacking a
cellulose cell wall, detritus from cyanobacteria has
less resistance to mineralization than detritus
from other algae, such as diatoms and greens
(Gunnison & Alexander, 1975; Jewell & McCar-
thy, 1971). The turn-over rates of diatoms and
green algae are higher (Los, 1991). Consequently,
relatively more detritus will be formed by diatoms
and green algae than if cyanobacteria were dom-
inant.
The flushing of Veluwemeer has led to a four-
folds increase in the calcium load (Van der Molen,
1990) and the concentration increased from 60-
80 g m - 3 before 1980 to 80-140 g m - 3 after
1980. Calcium is present in predominantly solu-
ble form. Because of algal growth and high buffer
capacity, chemical balance shifts such that cal-
cium carbonate can be formed. Only a few per
cent of the calcium is present in particulate form.
In the summers since 1987 the concentration of
particulate calcium in Veluwemeer varied between
2 and 8 g m - 3, contributing ca 5 to 10 g m - 3 to
the concentration inorganic suspended matter in
the water. Compared with the situation before
1980 loss on ignition of the seston increased by
2.8 g m - 3 (Van der Molen et aI., 1989). Since the
start of the deliberate summer flushings in 1985
loss on ignition has been significantly higher than
in the first years after 1980. The low value of loss
on ignition in 1989 is possibly caused by a further
decrease in productivity and therefore in forma-
tion of calcium carbonate. Beside the increase in
inorganic suspended matter through calcium car-
bonate formation, the contribution of increasing
densities of diatoms is also significant from 1985
onwards.
Discussion
Cooke et at. (1986) reported that flushing can
achieve improvement in eutrophic lakes in two
ways:
1. reduction in the concentration of the limiting
nutrient (dilution) and
2. increasing water exchange rates (wash out).
We observed an additional effect, at least for Ve-
luwemeer, namely increase of inactivation of
phosphorus in the lake sediment.
Although the effect of reduced sediment release
is important, the first observed positive effects in
Veluwemeer depend mainly on dilution of the
total-P concentration. This is demonstrated by
the initial quick response of the lake towards
flushing, followed by a slow down in recovery as
nutrient concentrations tend to become equal to
supply water concentrations. Cooke et at. (1986)
found that, of the mechanisms induced by flush-
ing, dilution provoked the most pronounced ef-
fect in recovery of Green Lake and Moses Lake
(USA). In Moses Lake algal biomass declined as
a result of reduced nitrogen concentrations, as
supply water was poor in nitrates. In most part
of the lake dilution rates ranged between 0.01 and
0.10 d - 1 (Veluwemeer ca 0.01 d - 1). Only in
Parker Horn, an arm of the lake nearby the in-
flow, did the dilution rates exceed 0.20 d - 1 for
short periods, and under these circumstances
some wash out of algal cells was noted. As in
many cases the availability of dilution water will
be limited, wash out of algal cells will usually be
of limited significance. However, under special
conditions wash out may become significant. In
Veluwemeer, at low water temperatures (below

2 °C and preferably under ice cover) algal growth
rates fall to zero level or even less due to in-
creased mortality and sedimentation. This pro-
vides conditions for obtaining some additional
effect of flushing by wash out of algal cells.
The winter flushing started in 1979 appears to
have been very successful, resulting in transpar-
ency improvement due to a reduction in biovol-
ume. The projected decrease of the cyanobacteria
dominance, however, did not lead to transpar-
ency improvement, although this had been ex-
pected. Figure 14 schematically summarizes the
influence of flushing on the composition of the
algae species and transparency in Veluwemeer.
As long as the nutrient concentration of the sup-
ply water is below concentration in the lake, di-
lution will occur. High calcium and nitrate con-
centrations in the supply water had a positive
effect on the binding of phosphorus in the sedi-
ments, therefore decreasing the phosphorus avail-
ability for algal growth. The decrease of available
nutrients results in a reduction of the algae bio-
volume and consequently transparency improves.
A decrease in nutrient availability further influ-
ences competition between the various algae
groups. This is particularly important in spring,
FLUSHING
,
SUPPLY OF
CALCAREOUS
WATER
!
I tive. Chlorophyll a only contributes for a small
i percentage to the transparency. It is therefore not
="~-o
INCREASE
G
~ TRANSPARENCY ~ DECREASE IN
' I
IN LOI . BIOVOLUME I
Fig. 14. Scheme of the most important effects of the flushing
in Veluwemeer on transparency.
195
and especially after a long period of ice coverage.
After 1985 dominance of Oscillatoria Agardhii
is broken and diatoms and green algae become
the most important species. Because the specific
light-extinction per unit of diatoms and green
algae is lower than that of the cyanobacteria, a
transparency improvement was expected. How-
ever, the increase in detritus concentration be-
cause of the lower decomposition rates of dia-
toms and green algae, combined with a high
turnover rate, delayed further improvement of the
transparency. The left hand side of Fig. 14 shows
that flushing leads to the supply of calcareous
water. This had led to an evident increase in loss
on ignition and a decrease in transparency in Ve-
luwemeer.
Conclusions
We have to acknowledge that the flushing of Ve-
luwemeer has unmistakably led to water quality
improvement. Oscillatoria agardhii densities has
been brought to the margins of its habitat because
of the reduction of the phosphorus concentra-
tions in Veluwemeer. The severe winter condi-
tions is an additional causal factor for the disap-
pearance of Oscillatoria agardhii. In Veluwemeer,
the importance of flushing as a means to reduce
the spring concentration of cyanobacteria is very
limited when aimed for a direct wash out of algal
cells, whereas the effect of flushing on the total-
P concentrations is of great importance.
Higher detritus and inorganic suspended mat-
ter concentrations affect the transparency nega-
surprising that as from 1985 transparency has
hardly increased during the summer.
In the present situation improvement by dilu-
tion is minimal. On average over the past 4 years,
the summer average of the total-P concentration
in Veluwemeer is about the same or even lower
than in the supply water. Dilution by flushing is
not longer of great significance under such cir-
cumstances. During the summer period under av-
erage circumstances, a loss of water of about 36
196
106 m 3 may occur due to infiltration and evapo-
transpiration. These losses must be compensated,
for maintaining a certain water level. In order to
maintain the original total-P concentration in Ve-
luwemeer, ca 50 106 m 3 will be additionally re-
quired during the summer period to compensate
for the 'concentration effect'.
In view of the really acceptable (phosphorus)
quality of Veluwemeer in the present situation, it
seems that the efficiency of flushing is low at
present. The quality of water in the bordering lake
Wolderwijd offers more prospects for improve-
ment. Therefore, we recommend to shift the
flushing operations from Veluwemeer towards
Wolderwijd in the coming years.
References
Adams, W. A., W. P. Adams, P. A. Flavelle & N. T. Roulet,
1984. Variability of light beneath a modified portion of the
snow and ice cover of a lake. Verh. int. Ver. Limnol. 22:
65-71.
Afghan, B. K. & J. F. Ryan, 1975. A modified procedure for
the determination of nitrate in sediments and some natural
waters. Envir. Letters 9: 59-73.
Berger, C. & A. Bij de Vaate, 1983. Limnological studies on
the eutrophication of Lake Wolderwijd, a shallow hyper-
trophic Oscillatoria-dominated lake in the Netherlands.
Schweizerische Zeitschrift fUr Hydrologie 45: 458-479.
Boers, P. C. M. & O. Van Hese, 1988. Phosphorus release
from the peaty sediments of the Loosdrecht Lakes. Wat.
Res. 22: 355-363.
Brinkman, A. G. & W. Van Raaphorst, 1986. De fos-
faathuishouding in het Veluwemeer. Ph.D. thesis, Techni-
cal University Twente, The Netherlands, 481 pp.
Cooke, G. D., E. B. Welch, S. A. Peterson & P. R. Newroth
(eds), 1986. Lake and reservoir restoration. Ann Arbor
Science, Butterworth Publishers, 392 pp.
Dankers, N. & R. Laane, 1983. A comparison of wet oxida-
tion and loss on ignition of organic material in suspended
matter. Envir. Techn. Letters 4: 283-290.
Gunnison, D. & M. Alexander, 1975. Resistance and suscep-
tibility of algae to decomposition by natural microbioi com-
munities. Limnol. Oceanogr. 20: 64-70.
Hosper, S. H. & M-L. Meijer, 1986. Control of phosphorus
loading and flushing as restoration methods for Lake Ve-
luwe, the Netherlands. Hydrobiol. Bull. 20: 183-194.
Jewell, W. J. & P. L. McCarthy, 1971. Aerobic decomposi-
tion of algae. Envir. Sci. Technol. 5: 1023-1031.
Loogman, J. G., 1982. Influence of photoperiodicity on algal
growth kinetics. Ph.D. thesis, University of Amsterdam,
The Netherlands, 112 pp.
Los, F. J., 1991. Mathematical simulation of algae blooms by
the model BLOOM II. Delft Hydraulics, Delft, The Neth-
erlands, 113 pp.
Moed, J. R., 1973. Effects of combined action of light and
silicon depletion on Astrionellajormosa. Verh. int. Ver. Lim-
nol. 18: 1367-1374.
Murphy, J. & J. P. Riley, 1962. A modified single solution
method for the determination of phosphate in natural wa-
ters. Analyt. Chim. Acta 27: 31-36.
PER, 1982. Projektgroep Eutrofieringsbestrijding Randme-
reno Eutrofieringsonderzoek Wolderwijd-Nuldernauw.
RWS-nota, Lelystad, The Netherlands, 166 pp.
PER, 1986. Projektgroep Eutrofieringsbestrijding Randme-
reno Bestrijding van de eutrofiering van het Veluwemeer-
Drontermeer. RWS-nota, Lelystad, The Netherlands,
296 pp.
Reynolds, C. S. & A. E. Walsby, 1975. Water-Blooms. BioI.
Rev., 50: 437-481.
Van der Molen, D. T., P. C. M. Boers & L. Lijklema, 1989.
Gegevens-analyse nutrienten en zwevend materiaal Velu-
wemeer. H 2 0 22: 760-762.
Van der Molen, D. T., 1990. Inventarisatie gegevens m.b.t. de
waterhoeveelheden en makro-ionensamenstelling van de
beken, kwel, neerslag en de polderuitslag op het Veluwe-
meer. RIZA werkdokument 90.153X, Lelystad, The Neth-
erlands, 29 pp.
Van der Molen, D. T., 1991. Simple, dynamic model for the
simulation of the release of phosphorus from sediments in
shallow, eutrophic systems. Wat. Res. 25: 737-744.
Hydrobiologia 233: 197-210, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 197
© 1992 Kluwer Academic Publishers.

Lake restoration with and without dredging of phosphorus-enriched

upper sediment layers

Joop Van der Does,l Pierre Verstraelen,2 Paul Boers,3 Jan Van Roestel,4 Rudi Roijackers 5 &
Gerard Moser 1
1 Water Board of Rijnland, P.O. Box 156,2300 AD Leiden, The Netherlands (correspondence address);

2 Water Pollution Control Authority Amstel & Gooiland, P.O. Box 1061, 1200 BB Hilversum, The
Netherlands; 3 Institute for Inland Water Management and Waste Water Treatment, P. O. Box 17, 8200
AA Lelystad, The Netherlands; 4 Oranjewoud Ltd, P. O. Box 8590, 3009 AN Rotterdam, The Netherlands;
5 Agricultural University Wageningen Department of Nature Conservation, P.O. Box 8080, 6700 DD
Wageningen, The Netherlands

Key words: lake restoration, eutrophication, sediment, dredging, water quality management

Abstract

Human activity has been the cause of continuing decline of water quality in most Dutch lakes. Devel-
opment of lake restoration programmes must take into account the lake functions. Major reduction of
the nutrient and pollutant loading is the primary step in lake restoration. Still, the recovery of eutrophic
lakes is retarded frequently because of internal phosphorus loading by the lakes' sediments. Sediment
dredging, as an additional tool for water quality management to accelerate accomplishing the desired
water quality, is studied. In this paper we evaluate the preliminary results of eight lake restoration projects
in the Netherlands. The lakes are compared in order to estimate the magnitude of the internal phosphorus
loading. Dredging as an additional measure was carried out twice in the peatlake Geerplas. In the
Nieuwkoop Lakes only the external phosphorus loading was substantially reduced from 0.9 to 0.2 g
P m - 2 Y- 1. Provisional results of these two shallow peatlake restoration projects focussed on eutroph-
ication abatement with and without dredging, are presented. Both show a decrease in phosphorus
concentration in the lakes. The necessity to dredge the lakes is discussed.

Introduction of the programmes focused on reduction of the

Excessive loading with nutrients, particularly
phosphorus, often leads to increased algal growth
and poor water quality in most Dutch lakes. Con-
sequently many human and nature related func-
tions are hampered. To combat this process of
deterioration, restoration programmes that have
been drawn up relate to major reduction of load-
ing with nutrients in surface water and ground-
water. Despite restoration measures the success
external phosphorus loading is often limited (Van
der Does & Klapwijk, 1985, 1987; Van Liere et al.,
1990).
Apparently very large reductions of the phos-
phorus loading are needed to force a lake into
another trophic category (Cullen & Forsberg,
1988). There might be a resistance of the foodweb
to changes (Gulati et al., 1990; Van der Vlugt
et aI., 1992). In this respect foodweb manipula-
tion techniques have proven to be successful in
198
small lakes, but their use in large lakes is still a
matter of research. Another cause for a delayed
or only partial response on external loading re-
duction may be internal loading by the sediments.
Phosphorus release from enriched upper sediment
layers retards the decrease of phosphorus con-
centration in shallow lakes. The reduction of the
internal loading can be met by dredging the phos-
phorus rich upper sediment layers (Bjork, 1988;
Cooke et al., 1986) or by phosphorus inactivation
techniques (Cooke et al., 1986; Boers, 1991; Pre-
pas et aI., 1990). Such a selective dredging of
phosphorus is both costly and time consuming.
The latter method, being much more economic
and faster, still needs to be proven (Boers et al.,
1992). For these reasons, lake restoration pro-
grammes must consider the need for additional
restoration techniques that might be necessary
besides external loading reduction. The necessity
also depends on the lakes objectives, which are
related to assigned lake functions (e.g. recreation,
drinking water production, nature conservation)
and type of watersystem. An example is given in
Table 1.
This paper focusses on the response of lake
phosphorus concentrations to reduction of exter-
nal phosphorus loading by diminishing the phos-
phorus sources. The net internal phosphorus
loading of the lake is estimated in different ways.
Two lake restorations with and without dredging
of phosphorus enriched upper sediment layers are
Table I. Examples of objectives for phosphorous concentra-
tions, which differ according to the type of water system and
its function (Ministry of Agriculture, Nature Management and
Fisheries, 1990; Ministry for Transport and Public Work,
1989; Province of North Holland, 1990; Province of South
Holland, 1991):
Type of Function Objective for total
water system phosphorus concentration
Shallow lake AMK* Summer average <0.08 mg P \-1
Shallow lake Nature Summer average < 0.05 mg P 1- I
conservation
Deep lake AMK Summer average <0.15 mg P 1-1
Ditch AMK Summer trimester :S 0.30 mg P 1- 1
* AMK is the General Environmental Quality in The Neth-
erlands.
compared. Lake restoration measures that were
taken, are also related to water quality objectives
and hydrological circumstances.
Material and methods
Research area
The shallow peat lakes studied are situated in the
low-lying centre of The Netherlands. The area
borders on the dunes in the west, the North Sea
Canal, Lake IJ and Lake Gooi in the north, a
Pleistocene ice-pushed ridge in the east and the
river Oude Rijn in the south (Fig. 1). The lake
area is surrounded by cities i.e. Amsterdam, Lei-
den, The Hague, Utrecht and Hilversum and is
called 'the green heart'.
Originally the green heart was part of a peat-
land belt lying just above sea level and reaching
from the north-west of The Netherlands to the
south-west. It consisted of swamps and marshes,
slow flowing rivers and few shallow lakes. Water
economy was characterized by groundwater flow
from the adjacent sand ridges and dunes and an
upward groundwater seepage causing a constant
water surplus. Dense vegetation retained the
water and allowed only decelerated water dis-
charge into the rivers. Since the early Middle
Ages, agricultural development, peat mining, ca-
nalization and organized drainage caused a low-
ering of the groundwater table, peat mineraliza-
tion and subsidence, and, consequently, land
inversion. Finally, in the last three centuries, land
reclamation and drinking water extraction from
the sand ridges and dunes caused a further drop
of the groundwater table and increased ground-
water flow towards the deep polders. The year
around surplus changed into a winter surplus that
is pumped into the rivers and canals and summer
shortages that are neutralized by water supplies
from the heavily polluted rivers and canals. Now-
adays the green heart lies some 2 m below sea
level and comprises many shallow lakes (1-3 m
deep) created by peat mining, several deep lakes
(5-50 m deep) created by sand mining, slow flow-
ing rivers, canals, drained wetlands that are used
as pasture lands, and some undrained wetlands.

Amsterdam
NORTH SEA
Fig. 1. Research area with the situation of the investigated lakes.
Hydrological and morphological data of the
investigated lakes are presented in Table 3 and
discussed with the results.
Monitoring of water and sediment quality
Water samples were taken monthly and sediments
were sampled occasionally. Concentrations of
nutrients, chlorophyll a, Secchi disk depth, dry
weight, Fe, and LOI (Loss on ignition) were de-
termined according to Dutch standard methods.
The used symbols are explained in Table 2. The
availability of phosphorus in sediments and its
impact on the phosphorus concentration in lakes
is estimated as: P - release and bioavailable - P
from the sediment. Phosphorus release rates from
intact sediment cores were sampled and measured
according to the continuous flow technique (Boers
& Van Hese, 1988). Bio-available phosphorus of
the top 5 cm of the sediment was quantified as the
yield of algal biomass in batch cultures contain-
ing algae using sediment as the sole phosphorus
source (Bruning & Klapwijk, 1984; Klapwijk &
Bruning, 1986; Boers et al., 1984). Maximal algal
199
Utrecht
River Oude Rijn
biomass was expressed as summer average
chlorophyll-a and calculated from summer aver-
ages of total P and total N according to Lijklema
et al. (1988):
- Chl-a = 1428 *P-Iake,
- Chl-a = 89.6 + 89.6*N-Iake.
Net internal loading (NIL) expected after a re-
duction of the external loading was calculated
from TP, TP/2 ,urn, and Fe concentrations in sed-
iments according to Boers (1991):
- log (NIL + 1)= -0.11+0.82*log(TP)
- log (NIL + 1) = 1.87 + 1.48 *log (TP/Fe),
- log (NIL + 1) = - 0.77 + 1.08*
log (TP/ < 2 ,urn).
Results and discussion
Hydrological considerations of lake restoration
The lakes investigated have a surface area rang-
ing from 0.3 to 19.6 km 2 (Table 3). Most of the
water bodies in the lakes are surrounded by
marshlands and pastures with ditches. The lakes
are part of polder systems that need to be main-
200

Table 2. List of used symbols in the text.

Symbol Meaning Dimensions

Alg.av.-P Algal available phosphorus (/'0 TP)

Chl-a Summer chlorophyll-a concentration in the lake water (mg m- 3 )
DRWT Dry weight of the sediment (~o of wet weight)
Fe Total iron concentration in the sediments (mg Fe g- I)
h Average depth of the lake (m)
Lin External phosphorus loading (gPm- 2 y-l)
LOI Loss on ignition in sediments (~o DRWT)
NIL Net internal phosphorus loading (gPm-ly-l)
N-Kj Kjeldahl nitrogen in sediments (mgNg- l )
N-lake Summer lotal nitrogen concentration in the lakes' water (mgNI- 1)
P-Iake Summer total phosphorus concentration in the lakes' water (mgPI- I)
Prel Phosphorus release rate by the sediment (mg P m - 2 d - I)
Qevap Loss of water by evaporation (my-I)
Qin Hydraulic loading (m y-I)
rw Residence time (month)
TP Total phosphorus in sediments (mg P g- I)
TP/<2Jlm Total phosphorus in sediments normalized clay fraction (mg P g- I)

Table 3. Hydrological and morphological data of the investigated lakes. The lakes are represented by the following numbers:
(I) Lake Amstelveense Poel, (2) Lake Geerplas, (3) Lake Naarden-North, (4) Lake Naarden-South, (5) Lake Kortenhoef-West,
(6) Lake Hollands Ankeveen, (7) Lake Stichts Ankeveen, (8) Loosdrecht Lakes, 59) Nieuwkoop Lakes-Noord-einderplas,
(10) Nieuwkoop Lakes-Zuideinderplas.

Lake no. 2 3 4 5 6 7 8 9 10

Area (km 2 ) open water 2.3 0.3 1.4 0.5 1.1 0.1 1.4 19.6 1.5 1.0
Average h (m) 1.8 2.5 0.9 0.7 1.1 1.1 1.2 1.8 2.75 2.75
Qin(m y-I) 2.0 1.8 1.5 1.4 2.2 1.8 2.3 2.5 1.3
r w (month) 2 17 4 1.5 2 4 3-4 9 21 4-5
Wind fetch (SwiNe) (km) 0.8 0.5 0.4 0.7 0.5 0.5 0.4 4 2.3 1.2

tained at fixed water levels. The polder systems
with shallow lakes are surrounded by separate
polders which often have a much lower fixed water
table. The water losses from the lakes are due to
infiltration, evaporation and outlet. In summer
water is supplied to the lakes to compensate of
evaporation (average Qevap of the lakes =
0.6 m y - 1) and infiltration. In winter there is ex-
cess precipitation which is pumped out. The hy-
draulic load (Qin) of the open water areas also
includes water transit to and from surrounding
marshlands. Low Qin values (ca. 1.3-1.5 m y- 1)
are found only in Lake Naarden and in the
Nieuwkoop Lakes (Table 3). In four lakes, Am-
stelveense Poel, Lakes Kortenhoef-west, Lakes
Stichts Ankeveen and the Loosdrecht Lakes, Qin
exceeds 2.0 m y - 1, due to high infiltration rates or
transit flows. Water residence time in most lakes
is relatively short, some months. Because the
lakes are shallow and located in a 'flow through'
system for the surrounding marshlands and due
to infiltration to the lower polders. Two lakes,
lake Geerplas and lake Nieuwkoop-Noordeinder-
plas, have a residence time exceeding one year.
Infiltration rates are on the increase due to a
constant land subsidence caused by peat miner-
alisation and consequent lowering of water levels.
High hydraulic loads to the lakes with nutrient-
rich water, needed to maintain water levels, ham-
per the reduction of the external nutrient loading.
 201

Thus, for purposes of lake restoration nutrient
removal in the inlet water is needed and involves
high costs. Nature management may also require
low hydraulic loads to allow ecosystems to re-
cover and stabilize.
Possibilities to reduce the hydraulic load should
be considered to achieve a lake restoration pro-
gramme at minimal costs. This may involve ele-
vating water levels in the agricultural polders and
reducing groundwater extraction for drinking
water (Verstraelen et ai., 1992). However, hydro-
logical isolation or increasing the residence time
will increase the negative influence of phosphorus
release from the sediments, because the outflow
of phosphorus will decrease. On the other hand
short residence times combined with nutrient-
poor water supply may deplete the phosphorus
load in the sediments and accelerate phosphorus
removal.
This alternative for reduction of the internal
loading by dredging nutrient enriched upper sed-
iments layers is carried out in the Nieuwkoop
Lakes-Zuideinderplas.
Characteristics of peaty sediments
The sediments of the lakes studied consist of an
unconsolidated superficial fluid mud layer con-
taining fine-grained material. This layer varies in
thickness from several em up to 1 m (Table 4).
The upper peat layer is loose, between 0-40 em
thick, and is found between mud layers and solid
peat. The solid peat layer has at some locations
been removed in the course of the peat-mining.
Now it varies from 10 em to several meters. Sand
or clay layers below the solid peat mainly deter-
mine the infiltration rate. Total phosphorus in the
upper mud layer is relatively high (TP = 1.0-
1.9 mg P g - 1) but decreases with depth to a con-
stant concentration (0.3-0.4 mg P g-l) in the
solid peat layer.
Total phosphorus concentrations in the upper
sediment layer are high in Naarden-South, Am-
stelveense Poel and Geerplas (Table 4) and low
in Hollands Ankeveen, Stichts Ankeveen and the
Loosdrecht lakes. Water supply of the eastern
lakes (Fig. 1) is influenced by groundwater seep-
age from the bordering Pleistocene ice-pushed
ridge in the east (Verstraelen et ai., 1992). Seep-
age water is rich in iron. Total iron concentrations
in the sediment top layers of the investigated lakes
are presented in Table 4. Highest Fe-values are
indeed found in the eastern lakes sediment and
will decrease phosphorus release rates (Boers
et ai., 1992) and net internal loading (Fig. 3). The
0-5 em top layer of the fluid mud can be mixed
with lake water due to resuspension or stirring up
by for example bream or motorboats (Yousef,
1980). Algal available phosphorus of this layer is
expressed as percentage of TP (Table 4). Lake
Amstelveense Poel had the highest bioavailable
phosphorus (41 %) and the Loosdrecht Lakes the
lowest (5%). The TP content of the loose peat
and solid peat, is not only lower but has also a
lower bioavailability. In the Nieuwkoop - Noor-
deinderplas available phosphorus in the mud,
loose peat and peat was respectively 23 %, 10.5 %,
11%.
Highest phosphorus release rates from the

Table 4. Sediment data of the investigated lakes. For lake numbers see Table 3.

Lake nr 2 3 4 5 6 7 8 9 10

Mud (em) 15 40 55 40 90 105 115 30 2 10

DRWT C/o) 5.2 5.5 17 II 5.0 II 8.0 9.0 7.4 9.7
LOI (%) 51 63 27 31 65 59 54 58 73 54
N-Kj (mg N g- I) 23 5.0 6.5 13.6 28 11.1 15 20 5.3 5.0
TP (mg P g- I) 1.46 1.48 0.72 1.86 1.53 1.30 1.04 0.90 0.92 1.40
Fe (mg g- J) 11.6 5.4 21.1 19.2 24.5 18.3 18.8 22.0 11.6 14.2
Alg. av.-P (%) 42 32 18 16 34 28 26 5 23 26
P reI (g P m - 2 d - I) 5.1 11 0.5 7 6 2.4 2.5 0 0 16
202

sediment were measured in the Nieuwkoop Lakes
- Zuideinderplas (16 mg P m - 2 d - I), where mud
and loose peat are mixed with clay.
Water quality and loading
In 1989, the study year, P-lake ranged from
0.05 mg P 1- 1 in lake Stichts Ankeveen to 0.31 mg
P 1- 1 in lake Nieuwkoop-Zuideinderplas (Ta-
ble 5).
According to the nutrient-biomass relations
(Lijklema et al., 1988), in all lakes growth ofphy-
toplankton is limited by nitrogen (Table 5).
Chlorophyll-a concentrations are much lower
than those expected from the phosphorus con-
centrations except in Lakes Loosdrecht. Thus be-
sides phosphorus reduction lake restoration ef-
forts may also include nitrogen reduction as well.
Nitrogen reduction without phosphorus reduc-
tion, however, may be not feasible, because N 2 -
fixation by cyanobacteria was observed in three
of the investigated lakes having P-summer con-
centrations exceeding 0.08 mg P 1- 1 (Burger-
Wiersma, 1991).
Water quality objectives for all the lakes stud-
ied aim at measures that lead to change in the
composition of the algal community, from cyano-
bacteria dominance to that of diatoms and green
algae. Without bio-manipulation, i.e. without im-
Table 5. Water quality data (summer average: 1989) of the investigated lakes. For lake numbers see Table 3.
provement in light climate, this objective requires
a decrease of P-lake to ca. 0.04-0.08 mg P 1- 1
(Lijklema et al., 1988; Sas, 1989). The P concen-
tration in most of the lakes is still much higher,
causing excessive algal growth. Therefore resto-
ration programmes are based primarily on reduc-
tion in P loadings. Some programmes are in prep-
aration and others already carried out (Table 5).
This table also includes summer averages of
transparency (Secchi disk) and P-lake. In 1989
external phosphorus loading to the lakes ranged
from 0.06 to 5.1 g Pm -2 y- I.
The question how far the external phosphorus
loading should be reduced in order to meet the
desired trophic status cannot be answered accu-
rately with two recently evaluated models (Lijk-
lema, 1991). This indicates that the issue con-
cerning the relation between reduction in external
load and desired concentration, is not met satis-
factory yet. Besides, uncertainties in water and
nutrient balances and lack of data on internal
loading hamper prediction of required reduction
in external loading. Therefore accurate water and
nutrient balances and documentation of the net
internal loading is of most importance.
Reduction of external phosphorus loading does
not result in a proportional decrease of the total
phosphorus concentration in a lake (average P-
summer concentration) (Cullen & Forsberg,
1988). Sas (1989) selected shallow lakes success-
fully restored by reduction of the external phos-

Lakenr 2 3 4 5 6 7 8 9 10

P-Iake (mg 1- I) 0.26 0.29 0.06 0.25 0.23 0.13 0.05 0.08 0.12 0.31
N-Iake (mg 1- 1) 3.4 3.4 1.6 3.6 3.1 2.6 1.2 2.4 2.6 3.8
Secchi (m) 0.36 0.43 0.63 0.31 0.33 0.40 0.73 0.35 0.45 0.31
Chl-a 173 125 27 136 181 78 15 150 113 256
Chl-a max
according to
P-Iake 371 414 86 357 328 186 71 117 168 442
Chl-a max
according to
N-Iake 215 212 53 233 18 143 17 116 143 250
Lin 1989
(g P m- 2 y-I) 0.84 0.38 0.25 0.31 5.1 0.06 0.22 0.35 0.48 4.7
Lin before
(g P m- 2 r ') 0.84 0.9 ? ? 5.6 0.46 0.83 1.1 0.48 4.7
 203

phorus loading. Average decrease of the phos-
phorus concentration after restoration in these
lakes (expressed as percentage reduction)
amounted to 80% of the % reduction of the ex-
ternal phosphorus loading. We consider lake res-
toration by nutrient loading reduction less suc-
cessful, when the concentration of the nutrients in
the lake does not decrease more than 40 % of the
percentage of the reduction in nutrient loading.
A theoretical example of possible reactions of
P-concentrations on reduction of external phos-
phorus loading is presented in Fig. 2. The lowest,
dotted, line shows the 1: 1 response, which is usu-
ally not obtained. The middle, solid, line shows
an average development of a successful reduction
of the external loading (see above), whereas the
upper, broken, line represents a less successful
case. The last case illustrates that, following ex-
ternal phosphorus loading reduction, the sedi-
ments may become a major source for phospho-
rus (net internal loading). So data on the
magnitude of the net internal phosphorus loading
are needed to answer the question if dredging
should be applied. Calculations of net internal
loading (NIL) is based on three ways (Boers,
1991): a) total phosphorus (TP) in sediments;
(b) TP and Fe in sediments; c) TP in sediments
normalized to the clay fraction. The results are
shown in Fig. 3. High net internal loadings, e.g.,
are calculated for lake Geerplas and low net in-
ternalloadings are calculated for the Loosdrecht
Lakes and Lake Naarden-North. The differences
between the lakes correlated to, as earlier men-
tioned, the differences in concentrations of phos-
phorus and iron in the sediments. Net internal
loadings are > 0 in lakes Amstelveense Poel,
Geerplas, Nieuwkoop, Naarden-South and Ko-
rtenhoef.
A second method to estimate net internal load-
ing is based on calculations with measured phos-
phorus release rates from sediments. P-release is
supposed to occur in a period of 200 days.
NIL rei = P reI (mg P m - 2 d - 1) * 200 (d Y- 1).
However, one should take into account that the
NIL is the resultant of both release and sedimen-
tation. Also, NIL may be much smaller than the
measured release because of; the release experi-

260

240

220

200

--
"i
Cl
leo
160

21 140
W
:.::: 120
~
Q. 100

eo
60

= 1:1
40
; : 100%
20

0 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.J

0 0.1 0.2 0.3 0.4 O.! 0.6 0.7 0.8 0.9 1.0 1.1 1.2
EXTERNAL LOADING [9 P m-2y-l]
Fig. 2. Example of possible response of the decrease in total phosphorus concentration C/o) in a lake on the reduction of the
external loading. Explanation see text.
204
NET INTERNAL LOADING
A-POEl
GEERPLAS
NAARDERMEER-N
NAARDERMEER-S
KORTENHOEF
H. ANKEVEEN
S. ANKEVEEN
lOOSDRECHT
NIEUWKOOP-N
NIEUWKOOP-Z
-2 o 2
• BASED ON TP •
Fig. 3. Calculated net internal loading (NIL) of the investigated lakes; based on total phosphorus (TP) in sediments, based on
TP and Fe in sediments and based on TP in sediments normalized to the clay fraction, according to Boers, 1991.
ments overestimate the in situ sediment release if
SRP-concentrations in the continuous flow reac-
tor are much smaller than under field conditions;
differences in redox at the sediment/water inter-
face between incubated cores and in situ sediment
may affect the adsorption capacity of the sedi-
ment considerably. Moreover hydrological fac-
tors such as seepage, infiltration and resuspen-
sion, are excluded in laboratory experiments.
Algal available phosphorus of the sediment top
layer (0-5 cm) shows the potential available
phosphorus at a certain moment (Table 4), but
cannot be linked to a annual net internal phos-
phorus loading.
To assess possible retardation of lake restora-
tion by reduction of the external phosphorus
loading, water quality managers are especially in-
terested in NIL during summer. Net internal
4 6 8 10
[9 P m-ey-1]
BASED ON TP/<2J,lm 0 BASED ON TP/Fe
loading based on measured release rates (NIL rel
in the different lakes exceeds Lin in some lakes:
Amstelveense Poel, Geerplas, Naarden-South
and Ankeveen (Fig. 4). Since the net internal
loading is difficult to quantify the lakes are ranked
relatively (Table 6), according to the following
criteria, that may show the magnitude of the in-
ternalloading; calculated NIL (based on TP, TP/
Fe and TP/ < 2 fl m), phosphorus release rates,
% algal-availability P, ratio NILrel/Lin and P-
concentration (summer average). This results in a
sequence of the lakes based on an average score
shown in Table 6.
According to this enumeration internal phos-
phorus loading is relatively highest in lake Geer-
plas, followed by the lakes Nieuwkoop - Zuid-
einderplas, Naarden-South, Amstelveense Poel
and Kortenhoef. The relatively lowest internal
 205
PHOSPHORUS LOADING

I
A.POEL1• • • • • n
GEERPLAS •
1 •' -
i - -T - -T - -1
NAARDERMEER·N ~
~ -'- ----J'---_-,

NAARDERMEER-S I~~~I==~=-----,I
KORTENHOEF •
I •• ~• • •~.I• • • • • •
I
••••••
i
5.1

H·ANKEVEEN ~
f------t

$-ANKEVEEN ~
LOOSDRECHT _ .

NIEUWKOOP·N I~~~J,__._--'-
r =- __..-L-_~ I 32
NIEUWKOOP-S I 4"7
L._._.. __L_. J L.. __.. . . .1.... ....._. .:.... J •
o 0.5 1.0 1.5 2.0 2.5 3.0
[9 P m- 2y-1]
o INTERNAL LOADING • EXTERNAL LOADING

Fig. 4. External phosphorus loading (Lin) compared to the estimated net internal phosphorus loading based on phosphorus re-
lease rates (NIL'e1) of the different lakes.

Table 6. Lakes ranked on the average score of each lake according to the criteria that may express the magnitude of the net in-
ternalloading (NIL); calculated NIL by TP, TP/Fe and TP/ < 211m; P,el; % Algal available P; NIL'e1/Lin and P-Iake. A low score
indicates a high net internal loading.

Lake NIL NIL NIL P,el ALGAL NIL/ P-lake Average

TP TP/Fe TP/ <211m % Lin score

Geerplas 3 2 I 2 3 I 2 2.0
Nieuwkoop-Zuideind. 5 4 3 I 5 6 I 3.4
Naarden-South I 5 4 3 5 2 4 3.4
Amstelveense Poel 4 6 2 5 I 5 3 3.7
Kortenhoef 2 I 7 4 2 7 5 3.9
Hollands Ankeveen 6 3 6 7 4 3 6 4.9
Stichts Ankeveen 7 9 8 6 5 4 10 5.9
Nieuwkoop-Noordein. 8 7 5 10 8 10 7 7.6
N aarden-North 10 8 10 8 9 8 9 8.6
Loosdrecht 9 10 9 9 10 9 8 8.9

phosphorus loading will be in the Loosdrecht
Lakes followed by Naarden-North and Nieuw-
koop Noordeinderplas. Additional measures to
reduce internal loading seem urgently in the lakes
with average scores below 4.
Reduction of the internal loading is considered
if the external loading has been far reduced in
comparison with the internal loading for example
in lakes Geerplas and Naarden (Fig. 4). In the
lakes Nieuwkoop - Zuideinderplas, Kortenhoef
206
and Amstelveense Poel reduction of the external
loading is mostly needed. The response of the
lake after reduction of the external phosphorus
loading may indicate if the internal loading should
also be reduced (Fig. 2).
Dredging criteria for peaty lakes
Dredging of the nutrient-rich upper sediment lay-
ers, i.e. surface sediment skimming is applied to
reduce the internal phosphorus loading. While
preparing lake restoration projects the following
criteria were recommend - derived from the
ranked lakes in Table 6 - to select lakes for ad-
ditional dredging measures:
Characteristics Criteria
- external P-Ioading far reduced in relation with
the residence time
>0.08 mg PI-I
- P-lake
- TP-upper layer >0.08 mg P g-I
- bio-available P in the >25%
upper layer
1
- P-release rate in upper >2.5mgPm- 2 d-
layer
- TP lower layer <0.70 mg P g-l
- bio-available P in lower <20%
layer
In two lakes, with highest net internal loading
(Table 6), experimental restoration projects with
a different approach. were started: lakes Geerplas
and the Nieuwkoop lakes.
The external P-Ioading in both lakes is far re-
duced in 1989 and 1990. Lake Geerplas has a
long residence time (> 1 year), therefore the high
net internal loading is also reduced by dredging.
In the Nieuwkoop-Zuideinderplas the short res-
idence time in combination with a nutrient-poor
water supply is used to deplete the phosphorus
load in the sediment in order to decrease the net
internal loading without dredging.
Restoration of lakes Geerplas: dredging
Since lake Geerplas is characterized by a rela-
tively high internal loading and reduction of the
external loading alone may not lead to its recov-
ery, sediment dredging was considered as an ad-
ditional measure. Dredging was focussed on re-
moving the nutrient-rich fine-grained top mud
layer using a two sided suction head. The thick-
ness of the mud layer removed varied in thickness
from 1 to 100 cm, without causing high turbidity.
The first attempt to remove the lake's sediment
was hardly successful (period April-October
1989), as the fluid mud layer was transported
from the undredged area to the area from where
the sediment had been removed by dredging. The
upper sediment layer proved to be very mobile.
Within a month wave-induced changes in mud
thickness up to 30 cm were measured (Fig. 5).
The measurements of the changes in fluid mud
level in lake Geerplas are large. A theoretical ex-
planation is found in the behaviour of the fluid
mud layer in which three sub-layers can be dis-
tinguished: a lower consolidated mud layer at the
bottom, a middle layer of fluid mud suspension
and a upper layer of diluted mud. Fluidization of
the mud layer is initiated by wave induced pres-
sure variation of the bed which lead to increasing
water pressure in the pores and hence to a reduc-
tion of the internal soil shear strength (Van Rijn,
1989). Transport of the upper dilute mud suspen-
sion layer is caused by advection and diffusion of
the water flow. The wind induced circulation flow
is an important transport mechanism and related
to the water depth of a lake. The bathometric
features of lake Geerplas show that in the deeper
middle part of the lake, the wind induced hori-
zontal flow is contrary to the wind, while at less
deeper borders the water flow corresponds to the
wind direction, according to Sunderman (1979).
Transport of the middle fluid mud suspension
layer may take place by gravity forces, for in-
stance by sloping bottom. Wind causes a surface
slope in a lake. In that case the interface slope
between the water and the fluid mud layer is op-
posite to the surface slope and an order of mag-
nitude larger, due to the density differences be-
tween the water column and the fluid mud layer
(SchOnfeld & Kranenburg, 1981). The mobility of
mud can be minimised by increasing the lake's
depth, by reducing the transport of the fluid mud

Fig. 5. Bathometric features of Lake Geerplas with depth
profiles (solid lines) and circulation by prevailing winds (bro-
ken lines) and results of mud transport. For example + 15 is
the localy increased thickness of the mud layer between Oc-
tober 6th and October 27th, 1989.
layer with screens or by reducing the fetch length
of the wind. Reducing horizontal mud transport
by screens was tested in lake Geerplas.
The screens jutted about 1 m above the sedi-
ment and stretched up to about 1 m below the
water surface. They could withstand wave action
and stormy conditions. Differences in sediment
thickness of up to 30 cm were observed between
both sides of the screens during a stormy 4 week
period (January-February 1990). In this period
the sediment thickness increased about 2 cm
within the protected area of 5.000 m2 , whereas on
the outside of the screens sediment thickness in-
207
creased up to 30 cm. Thus the horizontal fluid
mud transport can be suppressed. Apparently
fluid mud transport with higher concentrations
fine grained material seems to occur in 1 m above
the original sediment upper layer.
The restoration programme in lake Geerplas
was completed in 1990 by further reducing the
external phosphorus load from 0.38 to 0.13 g
P m - 2 Y- I. Waste water discharges into the lake
were stopped and the lake was hydrologically iso-
lated. The inlet water was dephosphorized before
it entered the lake since 1989.
During a second attempt from October 1990
till May 1991 surface sediment was skimmed with
preventing horizontal mud transport. Screens
were used to separate the three compartments: 1)
dredged compartments; 2) compartments which
were dredged and 3) untreated compartments.
Phosphorus concentrations in the lake decreased
to 0.1 mg P 1- I at the end of the second dredg-
ing operation (Fig. 6).
Restoration of the Nieuwkoop Lakes: no dredging
The Nieuwkoop Lakes consist of two main ba-
sins; lake Zuideinderplas and lake Noordeinder-
plas (Fig. 1). The restoration programme was ex-
ecuted in 1990. Point sources of phosphorus
discharge (households and horticulture) were re-
duced. A watershed was created between areas
for nature management and agriculture. Transit
water flow and hydraulic loading decreased.
Phosphorus removal was done by treating the
inlet water with iron chloride. The external phos-
phorus loading in 1990 decreased from 0.92 to
0.22 g P m - 2 Y- I for the whole Nieuwkoop
Lakes area (Van der Does & de long, 1992). The
lakes Noordeinderplas and Zuideinderplas differ
greatly in residence time, because water supply
and outlet of excess of precipitation from the
Noordeinderplas passes through the Zuideinder-
plas. Therefore these lakes differ in their nutrient
loading, water quality and sediment loading.
Phosphorus release rate in the Noordeinderplas
was almost nil, but extremely high in the Zui-
deinderplas (Table 4).
Notwithstanding the high internal phosphorus
208
1.2,------------..,-------------------,
1
0.8
......
a..
en 0.6
oS
0.4
0.2
1986 1987
Fig. 6. Changes in the total phosphorus concentration in lake Geerplas from Jan. 1986-May 1991. The dredging periods: 1989
and 1990/1991.
loading in the Zuideinderplas, dredging is not yet
carried out pending an evaluation of the lake's
response to the reduction in external load. The
Zuideinderplas is chosen as pilot location to study
the effect of its low residence time (4-5 months)
in the removal of phosphorus released by the sed-
iments. Development of the total phosphorus
concentration in lake Zuideinderplas is presented
in Fig. 7. Reduction of the external phosphorus
load in the Zuideinderplas (about 90%) resulted
in a decrease of 41 % of the lake's total phospho-
rus concentration (summer average) in 1990. This
means a decrease of 46% in P-Iake of the per-
centage of reduction of the external phosphorus
loading in the first year after the measures were
implemented. In the Noordeinderplas no signifi-
cant changes were observed so far.
Sas (1989) indicated that phosphorus concen-
trations in treated lakes will stabilize after about
3-5 times the residence time. Therefore the ex-
dredging periods
AVERAGE I
-----·····--········1
I
1988 1989 1990 1991
periment in the Nieuwkoop lakes will continue for
several years.
Conclusions and recommendations
- Assignment of different lake functions which
lead to contradictory measures should be
avoided.
- Hydraulic loading of peaty lakes has to be con-
trolled and can be used as a tool for lake res-
toration by reducing the external phosphorus
loading or removal of phosphorus released by
the sediments.
- More accurate nutrient- and water balances
are still of most importance for evaluation and
prediction of effects of lake restoration mea-
sures.
- Decrease of phosphorus concentrations can be
achieved with and without dredging of nutrient
enriched upper sediment layers. The choice be-

0.5
0.4
.....1. 0.3
Q.
-
CJ)
E
0.2
0.1
o
Fig. 7. Concentration of the total phosphorus in lake Nieuwkoop Zuideinderplas in the period 1985-1990. External phosphorus
loading was reduced by ca. 90% in 1990.
tween dredging or not depends on specific hy-
drologic circumstances (e.g. depth and resi-
dence time) and sediment characteristics (e.g.
depth profile and net internal loading).
Nitrogen reduction without phosphorus reduc-
tion is not recommended for the investigated
shallow lakes.
Mobile sediments must be controlled if dredg-
ing is applied for reduction of internal nutrient
loading.
Horizontal sediment transport can be con-
trolled by screens. This new technique should
be used to minimize the area of sediment re-
moval and the costs of restoration of shallow
lakes.
References
Bjork, S., 1988. Redevelopment of lake ecosystems. A case-
study approach. Ambio 17: 90-98.
Boers, P. C. M., J. W. T. Bongers, A. G. Wisselo & Th. E.
Cappenberg, 1984. Loosdrecht Lakes restoration project:
209
AVERAGE
Sediment phosphorus distribution and release from the sed-
iments. Verh. int. Ver. Limnol. 22: 842-847.
Boers, P. C. M. & O. Van Hese, 1988. Phosphorus release
from the peaty sediments of the Loosdrecht Lakes (The
Netherlands). Wat. Res. 22: 355-363.
Boers, P. C. M. (1991). The release of dissolved phosphorus
from lake sediments. Ph.D. Thesis. Agricultural University
of Wageningen. 139 pp.
Boers, P. C. M., J. Van der Does, M. P. Quaak, J. C. Vander
Vlugt& P. A. Walker, 1992. Fixation of phosphorus in lake
sediments using iron (III) chloride: experiences, expecta-
tions. Hydrobiologia 233: 211-212.
Bruning, C. & S. P. Klapwijk, 1984. Application of derative
spectroscopy in bioassays estimating algal available phos-
phates in lake sediments. Verh. into Ver. Limnol. 22: 172-
178.
Burger-Wiersma, T., 1991. Orienterend onderzoek naar stik-
stof fixatie in Nederlandse meren. Inventarisatie van stik-
stof fixatie in Nederlandse meren gedurende het groe-
iseizoen van 1990. Report Nr. 1991-1. Laboratorium voor
Microbiologie, Universiteit van Amsterdam (in Dutch).
Cooke, G. D., E. B. Welch, S. A. Peterson & P. R. Newroth,
1986. Lake and reservoir restoration. Buttersworth Pub-
lishers, Boston.
Cullen, P. & C. Forsberg, 1988. Experiences with reducing
210
point sources of phosphorus to lakes. Hydrobiologia 170:
321-336.
Gulati, R. D., E. H. R. R. Lammens, M.-L. Meyer & E. Van
Donk (eds), 1990. Biomanipulation - Tool for Water Man-
agement. Developments in Hydrobiology 61. Kluwer,
Dordrecht. Reprinted from Hydrobiologia 200/201.
Klapwijk, S. P. & C. Bruning, 1986. Available phosphorus in
the sediments of eight lakes in the Netherlands. In: P. G.
Sly, (ed.), Sediment and Water Interactions, pp. 391-398.
Springer Verlag, New York.
Lijklema, L., J. H. Janse, R. M. M. Roijackers & M.-L.
Meijer, 1988. Eutrophication in the Netherlands. H 20 21:
462-467 (in Dutch with an English summary).
Lijklema, L., 1991. Response of lakes to the reduction of
phosphorus load. Hydrobiol. Bull. 24: 165-170.
Ministry of Agriculture, Nature Management and Fisheries,
1990. Nationaal Natuurbeleidsplan (In Dutch). Staatsuit-
geverij, 's-Gravenhage.
Ministry for Transport and Public Works, 1989. Derde nota
op de waterhuishouding. (In Dutch with an English sum-
mary). Staatsuitgeverij, 's-Gravenhage.
Prepas, E. E., T. P. Murphy, J. M. Crosby, D. T. Wally, J. T.
Lim, J. Babin & P. A. Chambers, 1990. Reduction of Phos-
phorus and chlorophyl-a concentrations following CaC0 3
and Ca(OHh additions to hypertrophic Figure Eight Lake,
Alberta. Envir. Sci. Technol. 24: 1252-1258.
Province of North Holland, 1990. Waterhuishoudingsplan.
Natuurlijk, Water. Ecologische aspecten en normdoelstell-
ingen (in Dutch). Haarlem.
Province of South Holland, 1991. Waterhuishoudingsplan.
Leven door Water. (in Dutch). 's-Gravenhage.
Sas, H., 1989. Lake restoration by reduction of nutrient load-
ing: Expectations. Experiences. Extrapolations. St. Augus-
tin I: Akademia VerI., Richarz. ISBN 3-88345379-X.
Schonfeld, J. C. & C. Kranenburg, 1981. Density currents
and internal waves. Department Civil Engineering. Univer-
sity of Technology Delft.
Sunderman, J., 1979. Numerical modelling of circulations in
lakes. Hydrodynamics of lakes. In: W. H. Graf & C. H.
Mortimer (eds), Developments in Water Science, Vol. II,
Elsevier Sci. Publ. Compo
Van der Does, J. & S. P. Klapwijk, 1985. Phosphorus re-
moval and effects on water quality in Rijnland. H 20 18:
381-387 (in Dutch with an English summary).
Van der Does, J. & S. P. Klapwijk, 1987. Effects of phos-
phorus removal on the maximal algal growth in bio-assay
experiments with water from four Dutch lakes. Int. Revue
ges. Hydrobiol. 72: 27-39.
Van der Does, J. & P. De Jong, 1992. Integrale eutrofierings-
bestrijding in the Nieuwkoopse Plassen verloopt succesvol
(in Dutch with and English summary) H 2 0 25(1): 2-10.
Van der Vlugt, J. c., P. A. Walker, J. Van der Does & A. J. P.
Raat, 1992. Fisheries management as an additional lake
restoration measure: biomanipulation scaling-up problems.
Hydrobiologia 233: 213-224.
Van Liere, L., R. D. Gulati, F. G. Wortelboer & E. H. R. R.
Lammers, 1990. Phosphorus dynamics following restora-
tion measures in the Loosdrecht Lakes (The Netherlands).
Hydrobiologia 191: 85-87.
Van Rijn, L. c., 1989. Handbook Sediment Transport by
Currents and Waves. Delft Hydraulics, Delft.
Verstraelen, P. J. T., J. Wisserhof, Lj. Rodic & R. Eijsink,
1992. Eutrophication control strategies for three shallow
Vecht lakes in the province of North Holland. Hydrobio-
logia 233: 235-245.
Yousef, A. Y., W. M. Me Lellon & H. H. Zebutch, 1980.
Changes in phosphorus concentrations due to mixing by
motorboats in shallow lakes. Wat. Res. 22: 841-852.
H.I'drobiologia 233: 211-212, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 211
© 1992 Kluwer Academic Publishers.

Fixation of phosphorus in lake sediments using iron(III)chloride:

experiences, expectations

Paul Boers,l J oop Van der Does, 2 Marinus Quaak, 3 J 0 Van der Vlugt 4 & Paddy Walker 5
1 Institute for Inland Water Management and Wastewater Treatment, P.O. Box 17,8200 AA Lelystad,

The Netherlands; 2 Waterboard ofRijnland, P.O. Box 156,2300 AD Leiden, The Netherlands; 3 Volker Stevin
Dredging, P. O. Box 2695, 3000 CR Rotterdam, The Netherlands; 4 National Institute of Public Health
and Environmental Protection, P.O. Box 1, 3720 BA Bilthoven, The Netherlands; 5 Organization for
Improvement of the Inland Fisheries, P.O. Box 433, 3430 AK Nieuwegein, The Netherlands

Key words: eutrophication, lake restoration, phosphorus, sediments, internal loading,

chemomanipulation

Abstract

After a reduction of the external phosphorus loading to a lake, an internal loading from the sediments
may delay the improvement of the water quality. The accepted method to combat internal loading is
careful dredging of the upper sediment layers (Cooke et al., 1986), but this method is costly and time
consuming. Addition of phosphorus binding agents to the sediments might offer an alternative. In the
Netherlands the use of aluminum compounds, the most common phosphorus binding agent, for water
quality improvement purposes is not favoured. Therefore a sediment treatment with a solution of
iron(III)chloride was tested. Iron was chosen because it is considered to be a natural binder of phos-
phate. 100 g m - 2 of Fe 3 + were added to the sediments of the shallow (1. 75 m average depth) and eu-
trophic Lake Groot Vogelenzang (The Netherlands) in October and November 1989. The iron(III)chlo-
ride solution was diluted 100 times with lake water and mixed with the surface sediments with a water-jet.
Following the addition the concentrations of total phosphorus (Fig. 1), chlorophyll-a and suspended
solids decreased. This improvement of the water quality lasted for three months. After this time the total
phosphorus concentration increased again, but remained at a lower level than in spring and summer of
1989. The phosphorus release rate from the sediments as measured from intact sediment cores decreased
from 4 to 1.2 mg P m - 2 d - 1 (n = 5), and the bioavailability of the sediment phosphorus, as measured
with bioassays, decreased from 34 to 23 % (n = 5) shortly after the treatment. One year after the treat-
ment the release rate was increased to 3 mg P m - 2 d - 1 (n = 5). Before treatment, the lake was thought
to have a residence time of over one year. However, the chloride added to the lake disappeared according
to a dilution rate of 0.03 d - 1 or a retention time of about 35 days. A high external loading due to rapid
flushing with phosphorus-rich water from surrounding lakes possibly prevented a more durable improve-
ment in water quality. Another possibility is that the iron addition has lost its phosphate binding capacity
due to reduction or binding with other anions like carbonate or sulphide. Therefore the suitability of the
method to reduce internal loading and especially the long term availability of added iron to bind
phosphorus needs additional proof.
The treatment of the 18 ha area of Lake Groot Vogelenzang took three weeks. The operational costs
were about US$ 125000. This is fast and cheap compared to dredging. Application of the technique is
limited to those cases where the sediments are not polluted with micro-pollutants and the water depth
need not be increased.
212

We could not detect an influence on the macrophytes, because they remained absent due to the poor
light climate.
Phosphorus inactivation should be applied only after the external phosphorus loading is reduced. It
can be used as stand-alone technique or in combination with other lake restoration techniques such as
dredging or fish stock manipulation.

0.3 ...,--~----..,...,..,r------------,
References

iE- 0.2 Cooke, G. D., E. B. Welch, S. A. Peterson & P. R. Newroth,

Cl
.§. 1986. Lake and reservoir restoration. Butterworths pub-
a.. lishers, Boston. 392 pp.
:i. 0.1
l5I-
o JFMAMJJASONDJFMAMJJASOND
1989 1990

Fig. 1. Concentrations in mg 1- I of total phosphorus in Lake

Groot Vogelenzang during 1989 and 1990. The shaded col-
umn represents the treatment period.
Hydrobiologia 233: 213-224, 1992.
L. Van Liere & R .D. Gulati (eds). Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 213
© 1992 Kluwer Academic Publishers.

Fisheries management as an additional lake restoration measure:

biomanipulation scaling-up problems

J. C. Van der Vlugt,' P. A. Walker,2 J. Van der Does 3 & A. J. P. Raat 2

'National Institute of Public Health and Environmental Protection, P.O. Box 1,3720 BA Bilthoven, The
Netherlands; 2 Organization for the Improvement of Inland Fisheries, P.O. Box 433, 3430 AK Nieuwegein,
The Netherlands; 3 Waterboard of Rijnland, P.O. Box 156,2300 AD Leiden, The Netherlands

Key words: lake-restoration, turbidity, biomanipulation, zooplankton, fish removal, internal loading,
eutrophic lakes, bream

Abstract

Restoration of the highly eutrophic Reeuwijk lakes (ca. 700 ha) started in 1986 by reducing the exter-
nal phosphorus loading. As an additional measure to improve the quality of the lake water, the struc-
ture of the fish population in Lake Klein Vogelenzang (18 ha) was altered in 1989 by the removal of ca.
100 kg ha -, bream from the lake in April and December. This constituted about 50% of the total bream
biomass in the lake.
The fish-stock reduction in April, 1989, was initially followed by high phosphorus concentrations,
probably the result of considerable phosphorus release from the sediments. The resulting heavy algal
blooms that occurred reduced the transparency to very low values. During the summer the zooplankton
population increased markedly in numbers coinciding with reductions in total suspended matter includ-
ing (blue-green) algae. A great improvement in Secchi-disc transparency was observed and by the end
of December, 1989, the bottom of the lake (1.5-2.0 m) was visible. After heavy storms in January and
February 1990, transparency dropped to < 1 m as a result of resuspension of high concentrations of
suspended matter from the bottom sediments. Although transparency over the rest of 1990 was higher
than in 1988, i.e. the year preceding the removal of fish (biomanipulation), it was lower than expected,
based on the results of 1989. The study shows that technical and biological factors can cause serious
management problems for the implementation of biomanipulation in larger water bodies.

Introduction
The eutrophication of the Reeuwijk lakes (The
Netherlands) is a typical example of the problems
brought about by increased phosphorus and ni-
trogen emissions in the lower Rhine catchment
area. After decades of increasing emissions of
these nutrients, remedial actions were taken and
the effects of reduced phosphorus loading studied
extensively, starting in 1973 with pilot studies
(Van der Vlugt, 1976; Bannink & Van der Vlugt,
1978a, b; Bannink et al., 1980; Van der Vlugt &
Aldenberg, 1982).
The restoration of the Reeuwijk lakes, which
started in April 1986, was planned in two phases.
The first phase involved the reduction of the ex-
ternal P-loading. This was achieved by the re-
moval of phosphorus from waste water discharges
at the sewage treatment plant of the village of
Reeuwijk. These reduction measures did not lead
214
to significant improvements in water quality (Van
der Vlugt & Klapwijk, 1988, 1990). In the second
phase (1989-1992) the reduction of the phospho-
rus load is supplemented by in-lake measures,
both biomanipulation and chemomanipulation.
The latter involves in-lake addition of iron com-
pounds and is discussed recently in detail by
Boers et al. (1992).
The biomanipulation measure, as carried out in
this study, involves the removal of large numbers
of planktivorous and benthivorous fish. This
measure has been used as a tool in the manage-
ment of water quality in eutrophic waters and has
been practised in several countries throughout the
world with varying degrees of success (Shapiro
etal., 1975; Andersson, 1988; Benndorf, 1990;
Van Donk & Gulati, 1989; Gulati et al. 1990). In
The Netherlands the history of biomanipulation
dates back to 1985 to the feasibility study of Rich-
ter (1985, 1986). This study was followed by var-
ious experiments in ponds as well as whole lake
experiments (Meijer et al., 1989; Van Donk et al.,
1990a, b). In small (0.1 ha) ponds it was shown
that 0 + cyprinids could have an effect on the
composition of the zooplankton and the trans-
parency of the water (Meijer et al., 1990b; Raat,
1989). In somewhat larger ponds (0.4 ha) it was
demonstrated that 0 + predatory fish (northern
pike, European catfish and pike-perch) realised a
considerable production on a diet composed of
zooplankton and young cyprinids (Raat, 1990).
This study showed that these 0 + predators have
a potential as a management tool in the regulation
of abundant 0 + cyprinids, but with the present
knowledge of their dynamics the predators could
not be used as the only tool in biomanipulation.
Large scale removal of planktivorous and
benthivorous fish has resulted in a considerable
improvement of the water transparency in small
lakes (Meijer et al., 1989; Van Donk et al., 1989b,
1990a). Experiments in deep lakes in Sweden and
North America have shown that changes in the
fish population in larger water bodies have been
successful, at least for a period of up to four years
(Shapiro, 1990; Gophen, 1990; Walker, 1989).
Experiments on a larger scale in shallow Dutch
lakes, however, have so far not been successful,
Table 1. Water bodies in The Netherlands in which bioma-
nipuiation was implemented.
Pond/Lake Area Authors
(ha)
Drainable ponds; 0.1 (Raat, 1989; Meijer et a/., 1990b)
Drainable ponds; 0.4 (Raat, 1990; Meijer eta/., 1990b)
Zwemlust; 1.5 (Van Donk etal., 1989b, 1990a)
Bleiswijkse Zoom; 3.1 (Meijer et al.. 1989. 1990a)
Noorddiep; 4.5 (Meijer et al.. 1990a)
Klein Vogelenzang; 18 (Van der Vlugt et al., 1990)
Breukeleveen; 180 (Van Donk et a/., I990b)
Wolderwijd; 2700 (Meijer, 1989; Hosper & Jagtman.
1990)
for example in lake Breukeleveen (180 ha) where
the water has remained turbid, despite the large
scale removal of bream (Van Donk et al., 1990b).
The reasons for the failure of the biomanipulation
measure are complex and depend on the interac-
tion of abiotic and biotic factors as well as man-
agement problems.
Because of the promising results in drainable
ponds and small lakes, biomanipulation was
carried out in one of the smaller lakes in the
Reeuwijk Lakes area. This study in Lake Klein
Vogelenzang forms part of a series in which bi-
omanipulation was implemented in increasingly
larger lakes in The Netherlands (Table 1).
In the winter 1990-1991 the restoration of Lake
Wolderwijd (2700 ha) has been started as
planned (Hosper & Jagtman, 1990). The first re-
sults in spring 1991 are promising. Transparency
has increased from 40 em to 100 em in the course
of the winter, and to 1.7 m at the start of summer.
This paper deals with a study in Lake Klein
Vogelenzang, one of the Reeuwijk lakes of 18 ha,
in which the fish-stock was manipulated. This
study shows that technical and biological factors
can cause serious management problems for the
implementation of biomanipulation in larger
water bodies.
Description of sites studied
The Reeuwijk lakes
The Reeuwijk lakes area consists of nine shallow
lakes situated at 52 0 2' Nand 4 0 45' E in the

Reeuwijk polder near the city of Gouda in the
western part of The Netherlands (Van der Vlugt
& Klapwijk, 1990). The lakes originated from peat
mining carried out in the 16th and 17th centuries;
they have a total surface area of 7 km 2 and an
average depth of about 2 m. They are all inter-
connected except for Lake Broekvelden/
Vettenbroek, which is completely isolated from
the other lakes.
The shallow and peaty Reeuwijk lakes have
become highly eutrophic as a result of decades of
discharge of household waste water and the inlet
of nutrient-rich water from the adjacent Breevaart
canal. The water in the canal originally comes
from the river Rhine, via the Hollandse IJ ssel and
the canals of the city of Gouda. Moreover, the
sewage treatment plant of the village of Reeuwijk
(with 6800 inhabitants) discharges its nutrient-
rich effluent into the Breevaart canal, through
which the level of the lakes is regulated and kept
constant at 2.2 m below mean sea level. Excess
water in the polder is pumped out near Gouda
and Bodegraven. Since April 1986, the effluent
from the sewage treatment plant has been de-
phosphorized and consequently the average
yearly total phosphorus concentration in the Bre-
evaart canal has decreased from about 0.8 mg 1- 1
to about 0.4 mg 1- I (Van der Vlugt & Klapwijk,
1988, 1990).
Lake Klein Vogelenzang
Lake Klein Vogelenzang is a highly eutrophic
small lake in the centre of the Reeuwijk lakes. It
is interconnected with the lakes Groot Vogelen-
zang and Ravensberg. Its morphological and hy-
drological characteristics, as well as the phospho-
rus loading, are summarized in Table 2.
Materials and methods
Sampling and water analyses
The lake water was sampled fortnightly for three
years (1988-1990) at three depths (0-0.5 m, 0.5-
215
Table 2. Hydrological information on Lake Klein Vogelen-
zang.
Water surface area: 0.2 km 2
Mean water depth: l.5m
Maximum water depth: 2.0m
Residence time: 2.0 years
Gross P-load: 0.5 gP . m - 2. Y- I
General sediment composition: peat
1.0 m, 1.0-1.5 m) with a 0.5 m plexiglass tube
sampler of 3 I. These samples were mixed to make
a composite sample which was used for chemical
and biological analyses. Samples were trans-
ported to the laboratory in dark polyethylene jer-
rycans, within three hours of sampling.
Nitrate + nitrite nitrogen (N0 2 + N0 3 -N), ammo-
nium nitrogen (NH 4 -N) and orthophosphate
(P0 4 -P) were measured with a Technicon AA II
autoanalyser. Kjeldahl-nitrogen (Kj-N) and total
phosphorus (total-P) were measured after diges-
tion with, respectively, sulphuric acid + Wieninger
selenium mixture and sulphuric acid + persul-
phate followed by analysis for NH 4 -N and P0 4 -P
as stated above. Total nitrogen was calculated as
the sum of Kj-N and N0 2 + N0 3 -N. Particulate
phosphorus (part-P) was calculated as the differ-
ence between total-P before and after filtration.
Particulate nitrogen (part-N) was calculated by
substraction of NH 4 -N from Kj-N.
Chlorophyll a (Chi a) was determined spectro-
photometrically, after extraction of glass-fibre fil-
tered algae with 80% ethanol (75 0q, according
to a standard method of the Netherlands Nor-
malisation Institute (NEN 6250). C-phycocya-
nine was determined spectrophotometrically at
615 nm after extracting samples in distilled water
by freezing and thawing it ten times. Water sam-
ples used for the determination of C-phycocya-
nine were concentrated by continuous flow cen-
trifugation at 15000 rpm (Van Liere, 1979).
Samples for seston weight and particulate organic
matter were collected on 1.2 lim membrane fil-
ters; these parameters were determined gravimet-
rically as dry weight (DW) after drying at 75°C
and ash free dry weight (AFDW), as loss of
weight, after ignition at 600 ° C. Transparency
216

was measured with a Secchi-disc. For the esti-
mation of the zooplankton fresh weight (FW) a
modified Utermohl microscopical method (Van
Heusden, 1972) was used.
Fisheries management in Lake Klein Vogelenzang
Three aspects of the fisheries management pro-
gramme in Lake Klein Vogelenzang are: (1) the
estimation of the fish-stock prior to the bioma-
nipulation measure; (2) the actual management
measure, i.e. the implementation of fish removal;
(3) the evaluation of the development of the fish-
stock after biomanipulation. These aspects will
be discussed below.
Fish-stock estimation
numbers and biomass of the most dominant fish
species are shown in Table 3. The biomass of the
total fish-stock was ca. 260 kg ha - 1.
The pike-perch population was composed of
two relatively strong year-classes, from 1985 and
1987. The young pike-perch from 1988 were small,
< 10 em, suggesting that having been unable to
switch to piscivory, they had remained planktiv-
orous. Few pike-perch individuals> 50 em were
caught, which indicates an intensive and efficient
fishery. The pike (Esox lucius) population in Klein
Vogelenzang was small and limited to the inshore
area. Most of the pike were caught in the vege-
tation on the west side of the lake.
The most dominant cyprinid was the bream
which, with a biomass of 230 kg ha - 1, comprised
nearly 90% of the biomass of the entire fish-stock
(Fig. 1). A major part of the biomass (ca.

Table 3. Estimations of the fish-stock composition and size in

In 1976 and 1980 the fish-stock was assessed by
Lake Klein Vogelenzang, January, 1989 (before biomanipu-
the Operational Group from the Ministry for Ag- lation). (YOY = young of the year)
riculture and Fisheries in the lakes Elfhoeven and
's Gravenbroek. In these years the fish-stock was Fish size Numbers Biomass
composed primarily of pike-perch (Stizostedion (SD) (kg)
lucioperca) and bream (Abramis brama) and the
Pike-perch:
growth was poor. Information from professional Stizostedion YOY 1547 (909) 6
fishermen bore this out and the fish-stock was lueioperea > 14cm 217(104) 196
described as one-sided in favour of planktivorous Pike:
species (Van Gemeren, 1987). Esox lucius <30em 53 (25) 4
>30em 105 (69) 91
Little is known about the fish-stock in Klein Perch:
Vogelenzang before January 1989, but it is as- Perea jluviatilus YOY few
sumed to be similar to stocks found in the lakes > 10 em 592 (339) 57
Elfhoeven and's Gravenbroek above. The Co- Bream:
operation of Professional Fishermen from the Abramis brama YOY few
10-26 em 3726 (668) 603
Reeuwijk lakes has stocked elvers (Anguilla >26em 4646 (358) 3416
anguilla) yearly in Klein Vogelenzang for the past White bream:
15 years. Several hundred kilogrammes of bream Blieea bjoerkna YOY few
were removed from the lake each December in 7-26 cm 81 5
1985, 1986 and 1987. The professional fishery is Roach:
Rutilus rutilus YOY few
based on eel (Anguilla anguilla) and pike-perch. 8-24 em 5430 (1467) 153
The first part of the fisheries management pro- Rudd:
gramme, the fish-stock assessment, was carried Seardinius YOY few
out in January 1989. Fifteen different species of erythrophthalmus 7-24 em 1387(511) 75
fish were caught, and the size of the fish popula- 369 48
Eel:
tion was estimated using the mark-and-recapture
Anguilla anguilla
technique (Ricker, 1975). Estimations of the
 217

FISH POPULATION STRUCTURE IN 1989

Results

Development of the fish-stock

The lake was sampled for young fish (yay) dur-

(J)
(J)
<l:
:; 100
o
m inant species of young fish were pike-perch, perch,
50
()-l<'.-........"I""'"'''-----....,.'''''''--........''I'''''"'''----/ Perch
January 1989 April 1989
Fig. 1. Fish population structure in Lake Klein Vogelenzang
in 1989.
145 kg ha - 1) was contributed by bream> 32 em.
The growth of the bream was poor, as was that
of the bream caught in the lakes 's-Gravenbroek
and Eltboeven of the Reeuwijk lakes area.
Fisheries management
It was decided to approach the management in
three ways. Firstly, to remove a significant por-
tion of the bream population, especially the large
bream. This should reduce the amount of sedi-
ment material being suspended by foraging activ-
ities and relieve predation pressure on the zoop-
lankton. Secondly, to prevent the recruitment of
large numbers of young bream from the adjoining
lakes, the lake was shut off by two moveable sluice
doors. Thirdly, young pike were stocked in the
inshore area of the lake, in the hope that the pike
ing the summer of 1989. The size of the popula-
tion of young fish, as well as their growth, was
followed (Kemper & Raat, 1989). The most dom-
Bream
bream and roach (Rutilus rutilus). The main peak
Rudd and Roach
Pikeperch in 0 + fish density (ca. 60000 indo ha - I) and bi-
Pike
omass (ca. 50 kg ha - 1) was found at the begin-
December 1989
ning of June (Fig. 2). By the end of August the
numbers had dropped to ca. 10000 indo ha - I and
the biomass to ca. 25 kg ha - 1. Bream and roach
0+ fish constituted 15 kg ha - 1 of this biomass.
Pike-perch was numerically the most dominant
fish species and the 0 + fish remained planktivo-
rous, which would suggest a lack of young fish as
prey. This is because numbers of 0 + bream and
roach were low and the fish grew fast, making
them unavailable as prey fish.
If the relatively low recruitment of the bream
and roach was caused by the biomanipulation
measure is difficult to assess. The pike-perch born
in 1989 will probably not form a strong year-
40.--------------,
DENSITY OF 0+ FISH IN 1989
~
~30
~
"'20
&l
a:
w
~10
~
Z

population would reproduce naturally as the o

25
amount of vegetation increased following bioma- BIOMASS OF 0+ FISH IN 1989
nipulation. Thus the recruitment of young bream 20
would be regulated by young pike.
The Cooperation of Professional Fishermen re-
moved 1200 kg of bream from the lake in April
1989, and another 800 kg in December 1989 (De

~. n
5
Laak et al., 1989). At the beginning of May, 1989,
1500 young of the year (yay) pike were released o LI
PIKEPERCH
I
PERCH BREAM
IT
ROACH REST

into the lake. The sluice doors were put in place -JUNE 7 EillJUNE27 DJULY2S DA.UGUST30

in June 1989, prior to which the lake had been Fig. 2. Density and biomass of 0 + fish in Lake Klein Voge-
shut off from the other lakes using nets. lenzang in 1989.
218

class. The eels caught by the professional fisher-
men in July and August were in poor condition,
probably as a result of anoxic conditions at the
sediment-water interface.
The fish-stock was assessed further during the
winter 1989-1990 and a number of large bream
were removed at the same time.
Effects on water quality
Phosphorus and nitrogen
Most of the phosphorus and nitrogen measured
was in particulate form, 84-90% for P and 78-
93 % for N, as yearly average values (Fig. 3A-B).
The particulate NIP ratio was 10-20. The total-
P concentrations in the three study years differed
0.6 , - - - - - - - - - , - - - " " - - , - - - - - -
TOTAL-P
0.4
0.2
'§, 0 -'- ---1 -1-
markedly (Table 4). In 1988 the seasonal fluctu-
ations of the total-P concentration ranged from
0.10 mg 1- 1 in winter to 0.35 mg 1- 1 in May.
In 1989 i.e., after biomanipulation measures
were carried out, the concentration of total-P in-
creased to a much higher level than in 1988. In
mid-summer 1989 the part-P level exceeded
0.60 mg 1- 1, but by mid-September there was a
sudden decrease of part-P from 0.55 mg 1- I to
0.15 mg 1- 1, with a simultaneous increase of P-
P0 4 from 0.04 mg 1- 1 to 0.30 mg 1- 1. At the end
of the year the concentration of P-P0 4 again de-
creased to 0.10 mg 1- 1. In 1990 the concentration
of total-P increased from 0.10 mg 1- 1 to a maxi-
mum of 0.30 mg 1- 1. The average summer con-
centration, however, was 20% less than in 1988
(Table 4).
The concentration of total nitrogen in the three
years ofinvestigation also differed greatly. In 1988
the total nitrogen concentration was 1. 5 mg 1- 1,
in the winter and in the summer it fluctuated
between 2.5 and 4.0 mg 1- 1. In the following
_
winter the N concentration decreased to about

I:~~I
2.0 mg 1- 1. There was a sudden increase in ni-
trogen in spring 1989, reaching its annual maxi-
mum, 5.5 mg 1- 1, in late summer. In spring 1990

8:~COCI 1988 1989 1990

Table 4. Summer averages (I April-l October) of some water
quality parameters in Lake Klein Vogelenzang before (1988)
and after biomanipulation (1989 and 1990), concentration
expressed as mg I - I.

6,------=--..,-:c:_:_:_~__:_::_:c:_:_:_-__=-___,_----_____, Parameter Year

B/OMAN/PU AT/ON TOTAL-N
4
1988 1989 1990
2
~ 0-'-- --' -'- ----' P0 4 -P 0.020 0.060 0.030
.sz Particulate-P 0.260 0.350 0.190

~~I
o Total-P 0.280 0.420 0.220
~
a: NH 4 -N 0.13 0.45 0.10
I-
z N0 2 +N0 3-N 0.02 0.08 0.04

~CIJi'"" )
W
()
z Particulate-N 2.85 3.88 2.37
8 Total-N 3.00 4.41 2.51
Chlorophyll a 0.122 0.147 0.060
C-phycocyanine ND 0.350 0.043
Dry weight 28 31 20
1988 1989 1990 AFDW 21 23 14
Zooplankton (FW) 4.3 16.2 19.2
Fig. 3. The concentrations of phosphorus and nitrogen in Secchi-depth (cm) 34 32 49
Lake Klein Vogelenzang in 1988-1990.
 219

the trend of total nitrogen was reversed: instead
of an increase in concentration, a decrease to
1.5 mg 1- I was observed, followed by an increase
to about 3.0 mg 1- I in summer. The curves of
particulate-N and total nitrogen are rather simi-
lar, except for the second half of 1989, when the
NH 4 -N concentration increased from 0.1 up to
2.0 mg I-I.
trations fluctuated around 50 p,g 1- I, increasing
to 300 p,g 1- I in summer but there was a sudden
drop to a low level after a few weeks. At the be-
ginning of 1990 the concentration was even less
than 10 p,g 1- 1, the average summer value being
only 50% of that in 1988 and 40% of that in 1989
(Table 4); values as high as 100 p,g 1- I were not
reached until the autumn.

Chlorophyll a C-phycocyanine
In the winter of 1988 the chlorophyll a concen- The C-phycocyanine, a measure of the content of
tration was 60 p,g 1- I, but increased in summer blue-green algae in the water, was not determined
when it ranged between 80 and 200 p,g 1- I in 1988 (Fig.4A). In 1989 the curve of the C-
(Fig.4A). During biomanipulation and directly phycocyanine concentration was similar to that
after it in spring, 1989, the chlorophyll a concen- of chlorophyll a. The peak in midsummer, with
concentrations ranging from 0.2 to 1.0 mg 1- \
0.3 ~-----,---~--,--------, was followed by very low values in late summer.
CHLOROPHYLL a
0.2
In 1990 the concentrations of C-phycocyanine
reached peaks of only 0.100 and 0.200 mg 1- I a
0.1
few times.

i:~~1~I)~U:'"~~'=I
~ 0 ~ _
Seston
Seasonal fluctuations in the suspended matter
differed from year to year during the investigation
(Fig.4A). In 1988 the winter values fluctuated
o 60 I around 10 mg 1- I and increased to 30-40 mg 1- I

O~~I in summer. In the first half of both 1988 and 1989

the concentrations fluctuated between 10 and
40 mg 1- I. In midsummer 1989, however, the
1988 1989 1990
concentration of suspended matter ranged be-
60 - r - - - - - - - - - - - - . - - - - - - - - - - , , - - - - - - - - - - , tween 40 and 60 mg 1- I, but decreased sharply
AFDW
~
.§. 40 within a few weeks. Values as low as 2.0 mg 1- 1
<.5
z 20
were measured at the turn of the year 1989-1990.
o The pattern of suspended matter was similar to
(,)
o -1.- -'-'-- -="-- -1
that of chlorophyll a, in both midsummer and at

i~LIJI:[1
the turn of the year. After the extremely low win-
ter concentrations the peaks in January and Feb-
ruary, 1990, were due to resuspension of particles

J::1:A;=1
caused by heavy storms, with wind speeds
I > 30 m sec - 1. Though the concentrations of sus-
I
(,) pended matter in 1990 varied more than in 1988,
(,)
W
If> the summer averages were 30 % lower than in
1988 1989 1990 1988 (Table 4).
AFDW (Fig. 4B), the particulate organic mat-
Fig. 4. The concentrations of chlorophyll a, C-phycocyanine,
seston dry weight, AFDW, the biomass of zooplankton and ter part of seston, showed almost the same con-
Secchi-disc transparency in Lake Klein Vogelenzang in 1988- centration pattern as suspended matter, the lev-
1990. els being, of course, somewhat lower.
220

Transparency Zooplankton
The transparency (Fig. 4B) before and after the Virtually no cladocerans (Figs. 4B and 6) were
biomanipulation reflected the fluctuations in the encountered in 1988, crustacean biomass being
pigment concentrations of chlorophyll a and C- composed mainly of cyclopoid copepods and
phycocyanine (Fig. 4A) and suspended matter their nauplii, with maximum values of 5 mg 1- 1
(Fig.4B). The annual variations, irrespective of fresh weight. After biomanipulation in April 1989,
the restoration measure were high: in 1988 the the density of cyclopoid copepods increased
maximum Secchi-depth in winter was 80 em, de- gradually until the end of summer. Cladocerans
creasing in the summer to 20 em; in the summer were not observed in appreciable densities in 1989
of 1989 i.e., after the biomanipulation measure, until midsummer (Figs.4B and 6). Bosmina
the Secchi-depth was as low as 20 em. There was coregoni was the main cladoceran to reach ex-
first a sudden increase in transparency in Sep- tremely high densities in mid September (Fig. 6),
tember, 1989, followed by a gradual increase until coinciding with the decreasing concentrations of
early 1990. In this period up to early January, suspended matter and chlorophyll a B. coregoni
1990, the bottom of the lake (150-200 em) could peaked only for a short time and then disappeared
be seen for a few weeks. After the heavy storms until midsummer 1990, when it again exhibited
in January and February, 1990, the transparency high densities for a few weeks. Daphnia galeata
decreased markedly to around 50 em, remaining and Daphnia cucullata were observed from the
at this level for the rest of the year. middle of October 1989 and developed a biomass
of 30 mg 1- 1 fresh weight in about six weeks after
Phytoplankton the BosminajCopepod peak. Daphnia persisted
In the first half of 1988 the phytoplankton con- during the winter at densities as high as 30 mg 1- 1
sisted mainly of blue-green algae (Fig. 5) and sub- up to mid January, 1990. After declining in Feb-
sequently by a mixed population of chlorococcal ruary and March, the daphnids peaked again in
green algae, flagellates, diatoms and filamentous spring 1990. A last peak of Daphnia galeata and
blue-greens. The massive algal bloom in summer Daphnia cucullata coincided with high densities of
1989 was caused mainly by Lyngbya limnetica, B. coregoni in the first half of July, 1990. During
Anabaena sp., Aphanizomenon fios aquae and the second half of 1990 low densities of copepods
Microcystis aeruginosa. These forms were replaced as well as cladocerans were observed. From July
thereafter by chlorococcales, flagellates and dia- 1990 onwards Daphnia was virtually absent.
toms.
70
........ BOSMINA
BIOMANIPU !ATION • - - - CYCLOPS
- - CYANOPHYCEAE

\
60 - - DAPHNIA
........ CHLOROPHYCEAE
_,.-
._.. _.. CHRYSOPHYCEAE '§, 50
._-_. OTHERS .s
(f)
(f)
« 40
:2
0
iIi
30

~,~
~\ 20
:

.r! ,,>. ........&\

,
10

. v.,
... \
....... ,/, ,.'
\~.
.' I
...... "'-'... ' \ -.. ,',jl \ 0
-~, .... ' --,~ ," I ~ ,
o
1988 1989 1990
Fig. 6. The biomass (fresh weight) of Bosmina spp., Cyclops
Fig. 5. The numbers of phytoplankton cells in Lake Klein spp. and Daphnia spp. in Lake Klein Vogelenzang in 1988-
Vogelenzang in 1988 and 1989. 1990.

Discussion
In April 1989 1200 kg bream (Fig. 1) were re-
moved from Lake Klein Vogelenzang, leaving ca.
150 kg ha - I. Despite the high fishing effort by
commercial fishermen this catch constituted only
part of the resident bream-stock. The removal did
not result in an immediate improvement in water
quality and in December 1989 a further 800 kg
were removed. In the late summer of 1989 and in
1990 there was a noticeable improvement in
transparency (Fig. 4B), a decrease in chlorophyll
a (Fig. 4A) and an increase in the zooplankton
species Bosmina coregoni, Daphnia galeata and
D. cucullata (Fig. 6), which had been absent prior
to the manipulation of the fish-stock. The sea-
sonal dynamics of the results will be discussed.
During the summer of 1989 the transparency of
the water was low and chlorophyll a and total-P
were high as a result of a bloom of blue-green
algae. This was caused by an excessive mobilisa-
tion of phosphorus and nitrogen from the sedi-
ments following high pH values (> 9) and oxygen
depletion of a water layer of 0.5 m above the sed-
iments. This bloom of blue-green algae could be
the reason that the zooplankton population did
not increase in numbers following the removal of
fish. Daphnia species are unable to graze effec-
tively on blue-green algae such as Lyngbya
limnetica, especially at the concentrations mea-
sured (> 100000 per ml) and could have been
inhibited in their growth (see also Gliwicz, 1990).
The algal bloom was followed by a peak in the
Bosmina coregoni population. There is evidence
that Bosmina coregoni can graze quite effectively
on blue-green algae, especially if there is little al-
ternative (Gulati et al., 1991). This could have
been the case in Lake Klein Vogelenzang in 1989
where the peak of blue-green algae occurred as a
result of the high phosphorus levels in that year.
Following the decline in cyanobacteria and other
phytoplankton species, Daphnia galeata and
D. cucullata appeared and remained present until
the following year.
It is probable that not only the presence of
blue-green algae, but also the removal of fish had
an effect on the dynamics of the zooplankton
221
population. The depletion of the fish-stock in
April 1989,just before spawning, meant that fewer
young fish would be recruited to the population
than in previous years. The low numbers of the
young bream and roach bear this out (Kemper &
Raat, 1989). These 0 + fish grew fast, in compari-
son to 0 + fish caught in other peat lakes sampled
in 1989, which indicates an adequate food supply.
Consequently there was probably a high preda-
tion pressure on the available zooplankton. The
large size probably of the 0 + fish late in the sum-
mer meant that they were unavailable to the 0 +
pike-perch as prey. The fact that the young pike-
perch remained planktivorous in 1989 supports
this. By September there were few young fish left
(Fig. 2) and the predation pressure on the zoop-
lankton would have been relieved. This could
have contributed to the appearance of Bosmina
and small Daphnia species in the late summer of
1989.
In December a further 800 kg of fish were re-
moved from Lake Klein Vogelenzang. Although
it showed several fluctuations the Daphnia popu-
lation did not collapse during the winter of 1989/
90, possibly as a result of this second fish re-
moval. The transparency remained high during
the same period. Severe storms in January/
February 1990 led to increased turbidity due to
the resuspension of bottom sediments. Secchi
depth decreased from 200 cm to 60 cm and the
zooplankton biomass dropped from 30 mg 1- 1 to
5 mg 1- 1. Although the transparency did not im-
prove again during the course of 1990 it was higher
than in 1988 and 1989. Moreover, the Daphnia
population increased, following by Bosmina, and
chlorophyll a remained low during the summer
months. A decrease in the numbers of Bosmina
and Daphnia in August 1990 coincided with a
doubling of the chlorophyll a concentration al-
though this did not correspond to a decrease in
transparency. The reason for this is unclear.
The water quality therefore, appears to have
improved in 1990, albeit slowly. However, there
are fluctuations apparent in the parameters mea-
sured, especially seston, AFDW and total-P. It is
possible that the bottom sediments were stirred
up to such an extent by the storms in January and
222
February that they remained unstable throughout
the summer of 1990. Wind action and foraging by
fish would lead to resuspension of the sediments.
Although Lake Klein Vogelenzang was shut off
from the adjoining lakes during and after the ma-
nipulation measure it is probable that fish have
migrated into the lake. The sluice doors, separat-
ing the lake from the surrounding lakes, were re-
peatedly and intentionally left open. The lakes are
used by sailors, surfers, fishermen and others for
recreational purposes and the sluice doors were
a hindrance to their freedom of movement. Ob-
servations by personnel from the OVB (Organi-
zation for the Improvement of Inland Fisheries)
whilst fishing suggest that well-fed and under-fed
bream can be distinguished in the lake. The un-
derfed fish have been observed in Lake Groot
Vogelenzang (De Laak, personal observation),
where the availability of benthic organisms is very
low following the treatment of the sediments with
iron-chloride in order to bind phosphorus (Boers
et al., 1992). It is possible that these fish migrated
into the adjacent Lake Klein Vogelenzang and are
responsible for the resuspension of bottom sedi-
ments during their search for food. This illus-
trates the need for effective measures for isolating
the fish-stock, which take into account the needs
of other users of the lake(s) in question.
The removal of large numbers of planktivorous
and benthivorous fish in the spring has resulted in
a number of lakes in an immediate improvement
in water clarity (e.g. Bleiswijkse Zoom (Meijer
et al., 1989; Noorddiep (Kemper et al., 1990).
These were shallow lakes with a clay/sand bot-
tom in which aquatic macrophytes developed
within several weeks offish removal (Meijer et al.,
1990a). This colonisation with macrophytes has
been defined as one of the stabilising factors in
maintaining clear water, either by forming a ref-
uge for zooplankton or by acting as a competitor
for nutrients with blue-green algae (Ozimek et al.,
1990; Scheffer, 1990). In Lake Klein Vogelenzang
the situation is quite different. The peat layer is
overlain with a (moveable) layer in fine mud,
forming a lake bottom which is easily resuspended
and in which it is difficult for macrophytes to take
root. In order for macrophytes to grow there has
to be sufficient light at the lake bottom or in the
inshore area. In Lake Klein Vogelenzang the bot-
tom of the lake was only visible during the win-
ter following the second removal of fish. This is
an unsuitable time for plant growth. This lack of
aquatic macrophytes could be one of the factors
contributing to the slow speed of improvement.
The results in Lake Wolderwijd (2700 ha) two
months following fish removal show that it is pos-
sible to increase water transparency in large lakes
within weeks of a biomanipulation measure. Col-
onisation by Potamogeton pectinatus has occurred
and it is hoped that the lake will remain clear.
The specific conditions in peat lakes with re-
spect to bottom substrate and wind action make
them 'risky' subjects for biomanipulation mea-
sures. The problems described in this paper are
not only scaling-up problems, especially consid-
ering Lake Wolderwijd, but rather the application
of a management measure without taking the par-
ticular requirements of the lake into consider-
ation. The extrapolation of results from clay/sand
ponds to peat lakes has led to incorrect expecta-
tions as to magnitude of results and the time re-
quired for acquiring them.
Summarizing, the problems described in this
study are two-fold:
- the isolation of interconnected lakes which
are used for recreation purposes requires
careful management;
- the use of biomanipulation as a management
measure in peat lakes requires more detailed
research.
Conclusions
The transparency of the water in Lake Klein Vo-
gelenzang did not improve immediately following
the removal of ca. 50% of the bream population
in April and December 1989. A bloom of blue-
green algae was responsible for the lack of an
immediate improvement in water transparency in
1989. The warm weather conditions in the sum-
mer of 1989 caused anoxia at the sediment-water
interface and the consequent release of phospho-
rus from the sediments. This phosphorus was im-

mediately fixed by the blue-green algae. In the late
summer of 1989 the water transparency did im-
prove and coincided with a peak in Bosmina and
Daphnia and a decrease in phytoplankton. The
heavy storms in January and February, 1990,
caused the bottom sediments of the lake to be
resuspended. The sediments remained loose and
easily resuspendable after the storms. Despite
this, an improvement in transparency was ob-
served in 1990. The migration offish into the lake,
as a consequence ofothers leaving the sluice doors
open intentionally, may mean that the bream
population has reached its former level.
Acknowledgements
The authors acknowledge the valuable assistence
of the fishermen from the Lake Reeuwijk area.
Frans Jacques and Gerard de Laak of the Orga-
nization for the Improvement of Inland Fisheries
were responsible for the fish monitoring in Lake
Klein Vogelenzang.
References
Andersson, G., 1988. Restoration of Lake Trummen Sweden.
Effects of sediment removal and fish manipulation. In G.
Balvay (ed.), Eutrophication and Lake Restoration. Water
Quality and Biological Impacts. French-Swedish Sympo-
sium, Thonon-les-Bains: 205-214.
Bannink, B. A. & J. C. Van der Vlugt, 1978a. Erfahrungen mit
dem Zusatz von Fallmitteln im Grote Rug Speicherbecken
und vergleichende Untersuchungen in 3 Modell-Reser-
voiren nach Lund. DVGW-Schriftenreihe Wasser 16: 216-
243.
Bannink, B. A. & J. C. Van der Vlugt, 1978b. Hydrobiological
and chemical response to the addition of iron and alumin-
ium salts, studied in three Lund-type butylrubber reser-
voirs. Verh. int. Ver. Limnol. 20: 1816-1821.
Bannink, B. A., J. H. M. van der Meulen, J. C. H. Peeters &
J. C. Van der Vlugt, 1980. Hydrobiological consequences of
the addition of phosphate precipitants to inlet water of
lakes. Hydrobiol. Bull. 14: 73-89.
Benndorf, J., 1990. Conditions for effective biomanipulation;
conclusions derived from whole-lake experiments in Eu-
rope. Hydrobiologia 200/201: 187-203.
Boers, P. C. M., J. Van der Does, M. P. Quaak, J. C. Van der
Vlugt & P. A. Walker, 1992. Fixation of phosphorus in lake
sediments using iron(III)chloride: experiences, expecta-
tions. Hydrobiologia 233: 211-212.
223
De Laak, G. A. J., A. J. P. Raat & J. H. Kemper, 1989. Uit-
dunningsvisserij in Klein Vogelenzang in april 1989. Nieu-
wegein, Organisatie ter Verbetering van de Binnenvisserij.
OVB-Onderzoeksrapport 1989-16, 10 pp.
Gliwicz, Z. M., 1990. Why do cladocerans fail to control algal
blooms? Hydrobiologia 200/20 I: 83-97.
Gophen, M., 1990. Biomanipulation: retrospective and future
development. Hydrobiologia 200/201: I-II.
Gulati, R. D., E. H. R. R. Lammens, M.-L. Meijer & E. Van
Donk (eds), 1990. Biomanipulation tool for water manage-
ment. Developments in Hydrobiology 61, 628 pp.
Hosper, S. H., 1989. Biomanipulation, a new perspective for
restoring shallow lakes in The Netherlands. Hydrobiol.
Bull. 23: 5-10.
Hosper, S. H. & E. Jagtman, 1990. Biomanipulation addi-
tional to nutrient control for restoration of shallow lakes in
The Netherlands. Hydrobiologia 200/201: 523-534.
Kemper, J. H. & A. J. P. Raat, 1989. Broedbemonstering in
Klein Vogelenzang zomer 1989. Nieuwegein. Organisatie
ter Verbetering van de Binnenvisserij. OVB-Onderzoek-
srapport 1989-17, 9 pp.
Meijer, M.-L., (ed.), 1989. Toepassing van Actief Biologisch
Beheer in het Wolderwijd-Nuldernauw. Rapportage van
projectgroep Actief Biologisch Beheer. DBW/RIZA nota
89-057, Lelystad.
Meijer, M.-L., A. J. P. Raat & R. Doef, 1989. Restoration by
biomanipulation of Lake Bleiswijkse Zoom (The Nether-
lands): first results. Hydrobiol. Bull. 23: 49-57.
Meijer, M.-L., M. W. De Haan, A. W. Breukelaar & H.
Buiteveld, 1990a. Is reduction of the benthivorous fish an
important cause of high transparency following bio-
manipulation in shallow lakes? Hydrobiologia 200/201:
303-315.
Meijer, M.-L., E. H. R. R. Lammens, A. J. P. Raat, M. P.
Grimm & S. H. Hosper, 1990b. Impact of cyprinids on
zooplankton and algae in ten drainable ponds: preliminary
results. Hydrobiologia 191: 275-284.
Raat, A. J. P., 1989. Growth and production of 0+ bream
(Abramis brama), 0 + roach (Rutilis ruti/is) and 0 + carp
(Cyprinus carpio) in 10 drainable 0.1 ha ponds. Hydrobiol.
Bull. 23: 67-72.
Raat, A. J. P., 1990. Production, consumption and prey avail-
ability of northern pike (Esox lucius), pikeperch (Stizostedion
lucioperca) and European catfish (Si/urus glanis): a bioen-
ergetics approach. Hydrobiologia 200/201: 497-509.
Richter, A. F., 1985. Mogelijkheden van biomanipulation ten
behoeve van het waterkwaliteitsbeheer in Nederland. Ver-
slag 1985-8. Limnologisch Instituut Oosterzee. 109 pp.
Richter, A. F., 1986. Biomanipulation and its feasibility for
water quality management in shallow eutrophic water bod-
ies in The Netherlands. Hydrobiol. Bull. 20: 165-172.
Ricker, W. E., 1975. Computation and interpretation of bio-
logical statistics of fish populations. Bull. Fish. Res. Bd
Can. 191, 382 pp.
Shapiro, J., 1990. Biomanipulation: the next phase - making
it stable. Hydrobiologia 200/201: 13-27.
224
Shapiro, J., V. Lamarra & M. Lynch, 1975. Biomanipulation:
an ecosystem approach to lake restoration. In: P. L. Br-
ezonik & J. L. Fox (eds) Water Quality Management
Through Biological Control, Report No. ENV-07-75-1,
University of Florida Gainsville, 85-96.
Scheffer, M., 1990. Multiplicity of stable states in freshwater
systems. Hydrobiologia 200/201: 475-486.
Van der Vlugt, J. c., 1976. Comparative limnological research
in the 'Grote Rug' and model reservoirs. Hydrobiol. Bull.
10: 136-144.
Van der Vlugt, J. C. & T. Aldenberg, 1982. Blooms of algae
and their control by phosphorus precipitation/inactivation.
(In Dutch with English summary). H 2 0, IS: 73-79.
Van der Vlugt, J. C. & S. P. Klapwijk, 1988. Dose-effect re-
lationships between phosphorus concentration and phyto-
plankton biomass in the Reeuwijk Lakes (The Nether-
lands). Verh. into Ver. Limnol. 23: 482-488.
Van der Vlugt, J. C. & S. P. Klapwijk, 1990. the Reeuwijk
Lakes: a five years water quality study in an eutrophic ec-
osystem. Hydrobiol. Bull. 24: 57-67.
Van der Vlugt, J. c., S. P. Klapwijk & A. J. P. Raat, 1990.
First results ofbiomanipulation in the Reeuwijk Lakes. (In
Dutch with English summary). H 2 0, 23: 118-121.
Van Donk, E. & R. D. Gulati (eds), 1989a. Biomanipulation
in the Netherlands: application in fresh-water ecosystems
and estuarine waters. Hydrobiol. Bull. 23, 99 pp.
Van Donk, E., R. D. Gulati & M. P. Grimm, 1989b. Food
web manipulation in Lake Zwemlust: positive and negative
effects during the first two years. Hydrobiol. Bull. 23: 19-
34.
Van Donk, E., M. P. Grimm, R. D. Gulati & J. P. G. Klein
Breteler, 1990a. Whole-lake food-web manipulation as a
means to study community interactions in a small ecosys-
tem. Hydrobiologia 200/201: 275-289.
Van Donk, E., M. P. Grimm, R. D. Gulati, P. G. M. Heuts.
W. A. De Kloet & L. Van Liere, 1990b. First attempt to
apply whole-lake food-web manipulation on a large scale in
The Netherlands. Hydrobiologia 200/201: 291-31.
Van Gemeren, J., 1987. Mens en natuur in het Reeuwijkse
Plassengebied: het leven van de mens met veen en water,
vogels en vissen in het heden en verleden. Reeuwijk, John
van Gemeren, 48 pp.
Van Heusden, G. P. H., 1972. Estimation of biomass of
plankton. Hydrobiologia, 39: 165-208.
Van Liere, L., 1979. On Oscil/aloria agardhii Gomont, exper-
imental ecology and physiology of a nuisance bloom-
forming cyanobacterium. Thesis University of Amster-
dam.
Walker, P. A., 1989. The feasibility oflake restoration through
biomanipulation, a literature study. DB W/RIZA nota nor.
89-033 M & W Aquasense & DBW/RIZA, Lelystad.
60 pp.
Hydrobiologia 233: 225-233, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 225
© 1992 Kluwer Academic Publishers.

The effects of dredging and fish stocking on the trophic status of

shallow, peaty ditches

I. R. M, Hovenkamp-Obbema & W. Fieggen

Waterboard of Uitwaterende Sluizen in Kennemerland and West-Friesland. P,O. Box 15, 1135 ZH
Edam, The Netherlands

Key words: ditches, fish stock, dredging, biomanipulation, eutrophication, phosphorus

Abstract

In the polder 'Wormer, lisp and Nek', which forms part of a grossly eutrophicated region in the pro-
vince North Holland, the effects of dredging and biomanipulation were studied in ditches. Four areas,
including 5.7 ha surface water in total, were dammed up and dredged. It was feasible to remove the loose
top-layer of sediments with a high content of phosphorus by modifications of the dredging machine. In
two areas the fish stock was managed.
The measures resulted in an improvement of the water quality: an increase of the transparency and
a decrease of suspended solids, phosphorus, nitrogen and algae. The chlorophyll-a concentration de-
creased nearly to 100 p,g 1- I (summer average). The relative densities of diatoms, green algae and
cyanobacteria did not change. The zooplankton community was dominated by small rotifers and crus-
taceans both before and after restoration measures. Cladocerans were scarcely present. In the second
year after the measures, submerged macrophytes developed.

Introduction English shallow lakes where the submerged plant

The polder 'Wormer, lisp and Nek' forms a part
of the peat-area of Holland. It derives its name
from the names of three villages within the polder.
It contains three shallow lakes. The meadows are
separated by a labyrinth of ditches. In the polder
the objectives for water quality are not met. Due
to algal bloom the water is coloured green. A
thick layer of sludge covers the sediment. The
origin of this sludge is unclear. Up to fifty years
ago the water was brackish. The intrusion of fresh
water since 1935 probably has intensified the
mineralization of the peat-bottom of the ditches.
A further factor could be the disappearance of
submerged macrophytes. This situation is similar
to that found in the Norfolk Broads, a group of
communities have disappeared completely since
the late sixties (Balls et al., 1989). It is unlikely
that nutrient addition alone is the cause of a switch
from dominance by submerged plant communi-
ties to dominance by phytoplankton.
The waterboard of Uitwaterende Sluizen in
Kennemerland and West-Friesland has the task
to improve the water quality. The main strategies
to control eutrophication in fresh waters have
been the reduction of external and internal load
of nutrients (e.g. Bjork, 1985). Its effects can be
enhanced by the control of the internal ecological
processes by fish stock management (McQueen
et al., 1986; Benndorf, 1988) i.e. the reduction of
the biomass of both planktivorous and benthiv-
orous fish. Particularly, bream has a negative ef-
226
fect on water quality, as it feeds on large zoo-
plankters and benthic fauna (chironomids). This
reduces grazing on algae by large zooplankters
and causes resuspension of the sediment, both of
which contribute to increase in turbidity and the
nutrient release from the bottom. Both the high
algal biomass and the resuspension of sediment
cause deterioration of the light climate in the
water. (For review see Lammens et al., 1990). In
the Norfolk Broads, artificial refugia for zoo-
plankton have been created in order to reduce fish
predation (Irvin et al., 1990).
According to a pilot study the main source of
phosphorus in the polder is sediment release
(Scholler et al., 1985). The contribution of exter-
nal sources is low. The villages in the polder are
connected to a sewage treatment plant discharg-
ing its effluent outside of the polder. The pilot
study included the simulation of restoration mea-
sures by use of a water quality model. These sim-
ulations indicated that reduction of phosphorus
release by dredging of the upper layer containing
a high phosphorus content would be the most
successful measure to improve the water quality
(Scholler et al., 1985). The highest amounts of
phosphorus are present in the top layer of the
sediment (Waterboard, 1988). The release of
phosphorus has been measured in laboratory ex-
periments with use of sediment columns with and
without removal of the top layer. Removal of this
layer decreased the release of phosphorus from
the sediment to the surface water considerably
(Smeitink et al., 1988; Fig. 1). This decrease could
be sufficient for a substantial reduction of algal
biomass.
Based on data about the density and compo-
sition of the fish stock in the polder it also could
be assurped that fish stock management would
have positiv~ effects on the water quality (Raat &
Laak, 1989).
The first step to improve the water quality in
the entire polder area was an experiment in a
small area of ditches. The construction of a num-
ber of dams in connection with of a reallocation
of land, offered the possibility to create without
excessive expenses four experimental sites to test
two restoration measures. These measures con-
[P] Jlg I"
600
500
Before dredging
400
300
200
100 After dredging
0
0 10 20 30 40
Time (day)
Fig. I. Simulation in the laboratory of the effects of dredging
on the release of phosphorus from the bottom to the surface
water with use of a start concentration of phosphorus of
o/lg 1- I.
stitute control of excessive algal growth by dredg-
ing, and controlling of the fish population, a tech-
nique introduced in the Netherlands by Hosper
(Hosper, 1989). This paper presents the first re-
sults of these experiments.
Material and methods
Site description
For the experiment four comparable experimen-
tal sites - total 5.7 ha surface water - were cho-
sen (Fig. 2). The sites were isolated from the rest
of the polder by damming at the end of 1989. An
overflow allowed maintenance of the water level.
Additionally three reference sites were chosen in
the polder. The average water depth of the ditches
is 50 to 100 cm. The bottom consists of a peaty
layer. On the red-brown peat we find a brownish-
black layer of peat mixed with sludge. On top of
it there is a solid layer of sludge of varying thick-
ness (0 to 50 cm) covered by a slushy layer. The
layer of sludge had a thickness of about one metre
or less. The water in the polder is eutrophic (Ta-
ble 1).
 227

fish remains mainly in the lakes within the polder.

The biomass of the fish stock was estimated to be
about 300 kg ha - '.
In the four experimental sites (I-IV) the fish
were chased away almost completely by the
dredging activities. In each experimental site the
last dam was closed in February 1990 at the end-
ing of dredging activities. The remaining fish was
further caught by seine- and electrofishing. Sites I
and II were supplied with fish representative for
the present fish composition and stock in this
polder, namely ca. 170 kg fish ha - ': pike-perch,
bream white bream and roach (Table 3). The pur-
Fig. 2. Location of the polder Wormer, lisp and Nek, the pose was to introduce 300 kg ha - '. In the exper-
experimental sites (I, II, III and IV) and the reference sites. imental sites III and IV fish introduced were: pike,
Table 1. Summer averages of two sampling points of Secchi-
white bream, roach, rudd and tench, ca. 75 kg
disc depth, concentrations of nutrients and chlorophyll-a in fish ha - '. This density should represent that in
the polder Wormer, lisp and Nek in the period clear peaty waters. In each experimental site an
April-September area of about 10 m2 was dredged to about 1 m
depth, in which the fish can survive in winter. At
Parameter Summer average
the overflow wire netting was placed to prevent
1984 1985 the fish from escaping.

Secchi-disc depth (cm) 19 18

Total phosphorus (mg 1- I) 0.49
Total nitrogen (mg I- I) 6.2
Chlorophyll- a (/lg 1- I) 357
Dredging activities
The four experimental sites were dredged in the
period November 1989 - February 1990. Circa
35 em of sludge was removed with a dredging
machine pumping up the sludge via a sucking
device. The sludge was collected in a funnel
shaped mouth with a horizontal plate at its bot-
tom side and pumped into a sludge lagoon out-
side the experimental site. Within the funnel two
jacks with worm drive had to be mounted in order
to move the sludge inside.
Biomanipulation
In the polder pike was absent, pike-perch being
the main piscivorous fish (Table 2). The plank-
tivorous fish prefers the ditches, the benthivorous
0.57
Research programme
6.1
254
Samples of the bottom were taken in November
1989 and March 1991. A special auger was used
consisting of a solid auger head, a half-cylindrical
sampling body with a cover fin and a shaft. When
the auger was pushed in the sediment, the sam-
pling body was kept closed by the cover fin. It was
filled and closed again at the desired sampling
depth by turning the auger half a circle. At each
sampling site composite samples were made for
each layer with use of five samples taken at dis-
tances of about 5 m from each other. In Novem-
ber 1989 composite samples of bottom sediment
were made for the layers at 0-20 em, 2035 em,
35-50 em, 50-65 em, 65-80 and 80-95 em
depths. In March 1991, after dredging of a layer
of35 em, composite samples were taken from the
layers at 0-15 em, 1530 em, 30-45 em and 45-
60 em depths. The samples of bottom sediment
were examined for phosphorus.
The surface water was analyzed monthly for
suspended solids, total-phosphorus, total-nitro-
228

Table 2. Estimated numbers, size range and biomass of the fish stock in the polder based on seine and electro fishing autumn
1989 (Raat & Laak 1989).

Species Number Size range (cm) Total weight

(kg)
Min. Max.

Bitterling (Rhodeus sericeus amarus) 24 3.2 4.8

Bullhead (Cottus gobio) 34 3.8 3.8
Bream (Abramis brama)/
White bream (Blicca bjoerkna) 0 + 1214 3.3 8.2 3.1
Bream (Abramis brama»O+ 2233 9.0 46.0 652.0
Carp (Cyprinus carpio) 29 7.9 57.0 15.9
Eel (Anguilla anguilla) 58 15.0 37.0 2.6
Gudgeon (Gobio gobio) 1 10.2
Perch (Perea fluviatilis) 0 + 1116 6.1 9.0 6.3
Perch (Perea fluviatilis) > 0 + 154 10.0 27.0 8.1
Pike-perch (Stizostedion lucioperca) 0 + 785 5.6 20.0 4.4
Pike-perch (Stizostedion lucioperca) > 0 + 135 21.0 85.0 74.8
Roach (Rutilis ruti/is) 0 + 866 4.6 7.1 1.0
Roach (Rutilis rutilis) > 0 + 104 9.0 19.0 4.5
Rudd (Scardinius erythrophthalmus) 0 + 64 3.2 5.8 0.1
Rudd (Scardinius erythrophtalmus) > 0 + 221 7.0 19.0 7.4
Ruff (Gymnocecephalus cernua) 86 8.0 14.0 1.0
Smelt (Osmerus eperlanus) 42 7.7 10.2
Stickleback (Gasterosteus aculeatus) 408 3.4 5.5
'Vetje' (Leuciscus delineatus) 2 2.5 3.9
White bream (Blicca bjoerkna) > 0 + 284 7.0 28.0 20.1

Total 7860 801.3

Table 3. Fish set out in the four experimental sites; March ods (NEN), which comply with the standards of
1990 (Walker etal., 1990) the International Organization for Standardiza-
Experimental sites I and II without biomanipulation (= the
fish composition in the polder); experimental sites with bio-
tion (ISO).
manipulation (= the fish composition in clear water with The plankton was sampled in May, July, Au-
aquatic plants in peaty areas); (See Table 2 for Latin fish gust and September. For the enumeration of the
names). phytoplankton one litre of surface water was con-
served in 4 % formalin, sedimented and counted.
Site Fish species Ratio Ratio
biomasses numbers
For the zooplankton counting 10 litre sample was
used. The zooplankton was concentrated after
Pike-perch / filtering the water over a plankton net of 50 /lm
Bream + White bream + Roach 1/8 1/16 mesh size. The concentrate was preserved in 5 %
II Pike-perch /
formalin. The phytoplankton and zooplankton
Bream + White bream + roach 1/9 1/18
III Pike /
were counted using an inverted microscope.
Roach + White bream + Rudd 1/12 1/78 Macrofauna was sampled in May and August
IV Pike / with a net, which was pulled alongside the littoral
Roach + White bream + Rudd 1/12 1/70 zone of the ditch over a distance of 5 m. The
concentrate in the net corresponded to the sam-
gen, and chlorophyll-a. The transparency of the pling of about 250 litre of water. This concentrate
water was also determined. The samples were was preserved in 5 % formalin and Neomysis in-
analysed according to the Dutch Standard Meth- teger was counted.
 229

Results tal and reference sites. The effects of biomanip-

Sediment phosphorus
A sludge layer of ca. 35 cm thick was removed
out of the experimental sites. Using a modified
dredging machine the resuspension of sludge was
prevented and the loose top layer could be re-
moved almost completely. The phosphorus con-
tent decreased with depth (Fig. 3). As a result of
dredging the phosphate content of the top layer
was lowered. The phosphate content ofthe deeper
layers differed from that before dredging due to
small local differences in phosphorus-content of
the bottom.
Waterquality
The waterquality data are summarized in Table 4.
This table gives average results for the experimen-
ulation in the biomanipulated waters (sites III and
IV) are compared with the controls (sites I and II)
in Table 5 for transparency, suspendid solids and
chlorophyll-a.
Transparency and suspended solids
The measures taken in the experimental sites im-
proved the transparency of the water. At the out-
set the average Sechhi depth was 15.6 cm and
after the measures 29.4 cm. These figures com-
pare to 15.9 and 18.4 cm for the reference sites,
respectively. The quantity of suspended solids de-
creased considerably, namely from an average of
61.4 mg 1- 1 (dry weight) to 24.3 mg 1- 1, com-
pared to 62.2 and 65.0 mg 1- 1 for the reference
sites. The data in table 5 indicate that the differ-
ences due to biomanipulation between sites I/II
and III/IV are not significant.

Before dredging After dredging

Water surface Water surface

Depth Bottom level before dredging

(em) 0- 0- .

Dredged part

20- 20-

Bottom level after dredging

35- 35-

50- 50-

65- 65-

80- 80-

95- 95-

a 0.5 1.0 1.5 a 0.5 1.0 1.5

mg tot· P g.-' d.W. mg tot· P g.-' d.w.

Fig. 3. Phosphorus profiles of the bottom of experimental sites before and after dredging, measured in November 1989 and March
1991.
230

Table 4. Average water quality data for summer periods 1988-1990 (April-September, for algae densities May-September).

Secchi-disc Suspendid solids t-P t-N Chlorophyll-a Algae density

depth (cm) (mg 1- I) (mg 1- I) (mgl- 1)
(Ilg l - I ) (ind. ml- 1 x 10 3 )

Experimental sites 1988

mean: 14.9 69.0 0.60 6.0 148 38.3
s.d.: 0.2 7.4 0.07 0.3 20 1.0

Reference sites 1988

mean: 14.7 70.0 0.67 5.9 147 36.7
s.d.: 0.5 2.7 0.12 0.3 18 1.5

Experimental sites 1989

mean: 16.3 53.8 0.52 5.6 163 49.3
s.d.: l.l 2.6 0.12 0.9 27 8.8

References sites 1989

mean: 17.1 54.3 0.52 5.7 183 48.3
s.d.: 0.6 9.8 0.08 0.6 19 13.4

Experimental sites 1990

mean: 29.4 24.3 0.3 3.7 106 63.3
s.d.: 4.1 8.7 0.09 0.9 36 20.3

Reference sites 1990

mean: 18.4 65.0 0.43 5.0 209 90.3
s.d.: 0.7 4.4 0.04 0.5 27 11.2

Phosphorus and nitrogen in 1990 the average chlorophyll-a was

In the experimental sites the phosphorus and the
nitrogen concentrations decreased. After the
measures the summer averages were 0.33 mg
total-phosphorus 1- 1 and 3.7 mg total nitrogen
1- 1. The external phosphorus load also had de-
creased in 1990, compared with the preceding
years due to diminished use of phosphorus in
detergents. This is indicated by the summer means
of the reference sites of 0.67 mg 1- 1 of total-
phosphorus in 1988 and 0.43 mg 1- 1 in 1990.
Algal biomass and composition
63.3 x 10 3 1- 1 compared to 90.3 x 103 1- 1 in the
reference sites. In the summer the phytoplankton
was dominated by green algae (Fig. 4). There was
no permanent dominance of cyanobacteria, as
found in other peat lakes in Holland, e.g. the
Nieuwkoopse and Reewijkse plassen (Hoven-
kamp et aI., 1987). For this experiment important
species are the filamentous forms like Lyngbya
limnetica, Oscillatoria agardhii and Oscillatoria
redekei. These species are not easily grazed by
cladocerans (Dawidowicz, 1990 & Gliwicz,
1990). They were the dominant cyanobacterial
species in 1988, 1989, but also in 1990, although
their levels were'lower.

The measures reduced both the algal biomass and

densities. In the experimental sites in 1990 the
average chlorophyll-a was 106 Ilg 1- 1 compared
to 209 Ilg 1- 1 in the reference sites. The data in
Table 5 indicate that the differences due to bio-
manipulation between sites 1/11 and III/IV are not
significant. Algae density in the experimental sites
Zooplankton and macrofauna
Cladocerans, the most important grazers of algae
in many biomanipulated Dutch lakes (Gulati,
1989; 1990), were almost absent. The highest
density measured in spring was 24 indo 1- 1.
 231

Table 5. Average transparency, suspendid solids and chloro- Small-sized species of cladocerans were also
phyll-a for sites 1/11 and III/IV (without resp. with
scarce: only empty exuviae of Bosmina spp. were
biomanipulation)
found. A few individuals were met of the follow-
1/11 III/IV Difference ing species: Bosmina longirostris, Chydorus globulo-
sus, C. sphaericus, C. sp., Daphnia hyalina, D. sp.,
Secchi-disc depth (cm) Eurycercus lamellatus, Pleuroxus sp., P. trigonel-
April 25.0 37.5 12.50 Ius, Pseudochydorus globosus, Scapholeberis mucro-
May 25.0 35.0 10.00 nata, Sida crystallina and Simocephalus expinosus.
June 35.0 32.5 -2.50 Copepods - less efficient grazers of algae -
July 25.0 27.5 2.50
August 20.0 25.0 5.00
were found frequently and were far more abun-
Sept. 32.5 32.5 0.00 dant than cladocerans. Their maximum density
Mean: 4.58 amounted to 800 indo 1- 1 (May 1990). Rotifers -
St. dev.: 5.79 eating mainly detritus and nanoplankton - were
also abundant; their maximum number amounted
Suspendid solids (mg 1- ')
to about 1000 indo 1- 1 in April and May 1990.
April 39.0 34.0 - 5.00 The density of the zooplankton (cladocerans,
May 31.5 18.0 - 13.50
copepods and rotifers) was not influenced by the
June 23.0 18.0 - 5.00
July 23.00 13.0 - 10.00 restoration measures. Neomysis integer was found
August 40.0 23.0 -17.00 in different numbers (Table 6).
Sept. 11.0 16.0 5.00
Mean: -7.58
St. dev.: 7.76 Macrophytes

Chlorophyll-a (Ilg 1- ') Submerged plants were absent during the years
up to 1990 in all experimental sites. In May/June
April 192.5 87.5 - 105.0
41.5 90.5 49.0
1991 submerged macrophytes and filamentous
May
June 58.0 12.0 -46.0 green algae developed. Species which were no-
July 71.0 89.0 18.0 ticed are Potamogeton crispus, P. pusillus,
August 222.0 173.0 -49.0 Zannichellia palustris, Callitriche spp., Chara
Sept. 99.5 85.0 - 14.5
globularis and Elodea nuttallii.
Mean: -24.6
St. dev.: 54.5
Discussion
The dredging of the bottom-layer, with a mini-
mum of resuspension of sludge turned out to be

Table 6. Estimated numbers of Neomysis integer (individuals per sample of 250 litre).

1988 1989 1990

May August May August May August

Site I 3 0 3 10 0 55
Site II 12 166 133 12 43 7
Site III 0 0 2 10 64 12
Site IV 0 2 2 41 88 0
Reference 1 17 19 I 39 1 44
Reference 2 17 48 54 52 54 23
232
Experimental sites
Distribution (%)
100
75
50
o 0
BLUE DIATOMS GREEN OTHER
WB1988
Fig. 4. Percentage average distribution of dominating algae groups in summer (period May-September).
possible after some modifications of the dredging-
machine. The method, however, was laborious
and expensive. The measures have improved the
water quality, especially with respect to transpar-
ency and density of algae. The results agree rea-
sonably well with the predictions based on the
pilot study (Scholler et al., 1985), especially re-
garding the phytoplankton biomass measured as
chlorophyll-a. As the summer of 1990 was rela-
tively dry the inlet of more water than normal was
necessary. This inlet water had a high phospho-
rus content (Table 4) and thus reduced the effect
of measures, which have been more effective in a
normal year. A further disturbing factor might
have been the irregular inlet of water in the year
1990, leading sometimes to the inlet oflarge quan-
tities of nutrient-rich water in a short period.
In the summer of 1990 no submerged plants
developed. An explanation could be insufficient
stabilization of the bottom sediment as dredging
ended not before February 1990. Secondly in the
summer light penetration in the water was not
optimal (Secchi-depth 30 cm) because of still too
high concentrations of algae and suspended
Reference sites
100 r - - - - - - - - - - - - - ,
75
~
BLUE DIATOMS GREEN OTHER
.1990
solids. The change-over from an algae-dominated
turbid water to a clear water should be followed
by the development of aquatic plants (Scheffer,
1990). Already in May/June 1991 six species of
submerged macrophytes developed without intro-
ducing them.
So far fish stock management has not shown
any positive results. This is possible because:
1. the fish stock in the experimental sites I and II
did not differ much from that in the control si-
tes III and IV; and 2. contrary to the expecta-
tions, large bodied cladocerans were scarce even
after dredging and fish stock management. This
could have been caused by the composition of
phytoplankton namely high concentrations of
filamentous cyanobacteria (Dawidowicz, 1990;
Gliwicz, 1990; Bernardi & Giussani, 1990). An
other causal factor might be predation by
Neomysis integer. In lake Washington predation
by Neomysis integer was found to be important in
reducing the density of zooplankton (Edmonson
& Abella, 1988). Neomysis integer is omnivorous,
but prefers animal food (Mauchline, 1971; Ha-
nazato, 1990). It feeds on detritus, algae, sus-

pended organic material and hunts zooplankton.
In the absence of aquatic plants the cladocerans
lack refuges from predators.
On the whole the measures taken have reduced
the levels of nutrients, the density of algae and has
increased the transparency. The recent return of
submerged plants indicates that the prospects for
further improvements in the ecosystem are
good.
Acknowledgement
We thank Dr R. D. Gulati and Dr E. Van Donk
for making helpful comments on an earlier ver-
sion of this manuscript.
References
Balls, H., B. Moss & K. Irvine, 1989. The loss of submerged
plants with eutrophication. I. Experimental design, water
chemistry, aquatic plant and phytoplankton biomass in ex-
periments carried out in ponds in the Norfolk Broadland.
Freshwat. BioI. 22: 71-87.
Benndorf, J., 1988. Objectives and unsolved problems in
ecotechnology and biomanipulation: A preface. Limnolog-
ica (Berlin) 19: 15-18.
Bjork, S., 1985. Lake restoration techniques. In: Proceedings
International Congress 'Lakes Pollution and Recovery',
Rome; p. 281-292.
Dawidowicz, P., 1990. Effectiveness of phytoplankton control
by large bodied and small bodied zooplankton. Hydrobio-
logia 200/201: 43-47.
De Bernardi, R. de & G. Giussani, 1990. Are blue-green alga
a suitable food for zooplankton? An overview. Hydrobio-
logia 200 201: 29-41.
Edmondson, W. T. & S. E. B. Abella, 1988. Unplanned Bi-
omanipulation in lake Washington. Limnologica 19: 73-79.
Gli\\"icz, Z. M., 1990. Why do cladocerans fail to control algal
blooms. Hydrobiologia 200/201: 83-97.
Gulati, R. D., 1989. Structure and feeding activities of zoop-
lankton in lake Zwemlust in the two years after biomanip-
ulation. Hydrobiol. Bull. 23: 5-10.
Gulati, R. D., 1990. Structural and grazing responses of
233
zooplankton community to biomanipulation of some Dutch
water bodies. Hydrobiologia 200/201: 99-118.
Hanazato, T., 1990. A comparison between predation effects
on zooplankton communities by Neomysis and Chaoborus.
Hydrobiologia 198: 33-40.
Hosper, S. H., 1990. Biomanipulation, new perspectives for
restoration of shallow eutrophic lakes in the Netherlands.
Hydrobiol. Bull. 23: 5-10.
Hovenkamp-Obbema, I. R. M., S. P. Klapwijk& Y. Scheffer-
Ligtermoet, 1987. Toepassing van het raamwerk van
Caspers en Karbe voor de waterbeoordeling in Noord- en
Zuid-Holland. In: Biologische Waterbeoordeling: Instru-
ment voor Waterbeheer; symposium 2 en 3 april 1987,
Rijksinstituut voor N atuurbeheer, Leersum.
Irvine, K., B. Moss & H. Balls, 1989. The loss of submerged
plans with eutrophication. II. Relationships between fish
and zooplankton in a set of experimental ponds, and con-
clusions. Freshwat. BioI. 22: 89-107.
Lammens, E. H. R. R., R. D. Gulati, M. L. Meijer & E. Van
Donk, 1990. The first biomanipulation conference: a syn-
thesis. Hydrobiologia 200/201. R. D. Gulati, E. H. R. R.
Lammens, M. L. Meijer & E. Van Donk (eds). Proceedings
of an international conference held in Amsterdam, The
Netherlands, 8-11 August, 1989; Kluwer 638 pp.
Mauchline, J., 1971. The biology of Neomysis integer (Crus-
tacea, Mysidacea). J. mar. bioI. Ass. U.K. 51: 347-357.
McQueen, D., J. R. Post & E. L. Mills, 1986. Trophic rela-
tionships in pelagic ecosystems. Can. J. Fish. aquat. Sci.:
1571-1581.
Raat, A. J. P. & G. A. J. Laak, 1989. De visstand in de polder
Wormer, Jisp en Nek najaar 1989. Hoogheemraadschap/
O.V.B. 32 pp.
Scheffer, M., 1990. Multiplicity of stable states in freshwater
systems. Hydrobiologia 200/201: 475-486.
Scholler, M., G. Van Straten & W. Fieggen, 1985. Eutrofi-
eringsmodel ten behoeve van het waterkwaliteitsbeheer van
de polder Wormer, Jisp en Neck. H 2 0 26: 548-554.
Smeitink, D. J. W., G. Van Straten & A. G. Brinkman, 1988.
Het effect van baggeren op de fosfaatafgifte van het sedi-
ment in de polder Wormer, Jisp en Nek. Universiteit van
Twente, 32 pp.
Walker, P. A., A. J. P. Raat & G. A. J. De Laak, 1990. De
afvissing en de bezetting van de proefvakken in de polder
Wormer, Jisp en Nek, voorjaar 1990; O.V.B. rapport, Nieu-
wegein.
Waterboard of Uitwaterende Sluizen in Kennemerland and
West Friesland, 1988. Onderzoek van de waterbodem in de
polder Wormer, Jisp en Nek. Rapport, Edam. 31 pp.
Hydrobiologia 233: 235-245, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 235
© 1992 Kluwer Academic Publishers.

Eutrophication control strategies for three shallow Vecht lakes in the

province of North Holland

P. J. T. Verstraelen,l J. Wisserhof,2 Lj. Rodit3 & R; Eijsink 4

1 Water Pollution Control Authority Amstel & Gooiland P.O. Box 1061, 1200 BB Hilversum, The
Netherlands; 2 Delft University of Technology, Faculty of Civil Engineering, P.O. Box 5048, 2600 GA
Delft, The Netherlands; 3 International Institute for Hydraulic and Environmental Engineering, Department
of Aquatic Ecology, P.O. Box 3015, 2601 DA Delft, The Netherlands; 4 University of Utrecht,
Department of Environmental Studies, P.O. Box 80115, 3508 TC Utrecht, The Netherlands

Key words: lake restoration, eutrophication, water balances, nutrient loading, sediment, alternative
supply options

Abstract

Like many shallow surface waters in the Netherlands the North Holland Vecht lakes, formerly known
for their rich variety of flora and fauna, now face a serious eutrophication problem. Nutrient enrichment
has been mainly in the form of (treated) wastewater discharges, and the continuing ingress of nutrient-
loaded water from the river Vecht. Yet, this water has to be supplied in order to compensate for water
shortages resulting from (i) changes in the groundwater flow pattern due to reclamation of the deep polder
Horstermeer, (ii) extensive groundwater extraction in the Gooi hills, and (iii) extensive drainage for ag-
ricultural purposes.
The present policy of eutrophication abatement and restoration of the Ankeveen and Kortenhoeflakes
ecosystems is focused on eliminating wastewater discharges and Vecht water supply. It also allows for
additional dredging measures. Because of the un-suitable major ion composition of the Vecht, the aim
is to compensate for this water supply by (i) partial restoration of the original groundwater flow from
the Gooi hills and (ii) periphere additional supply with fresh seepage water from the skirts of the
Horstermeer polder. However, uncertainty exists about the amounts of water needed.
Water balances and phosphorus budgets have been established to ascertain the water demands of the
lakes and to gain a detailed insight into the nutrient fluxes through the lakes. A groundwater flow model
is used to assess the beneficial effects of the proposed measures.
The results obtained, question the current unilateral restoration objectives. Calculations reveal that,
both in the present situation and after (total) reduction of groundwater extractions in the future, the
available quantity of fresh seepage water from the skirts of the polder Horstermeer is not sufficient to
replace the inlet from the river Vecht into the Kortenhoef lakes. Additional supply options are available
but the ones favoured from an ecological viewpoint are either the most expensive or less favoured from
a social point of view. Although the sediments of the lakes appear to be a major source of eutrophication,
the possibility of dredging the lakes will be considered only after reviewing results of a pilot-dredging
project in the Hollands Ankeveen lakes in 1991.
236
Introduction
Man-induced eutrophication of lakes and reser-
voirs is considered to be one of the most widely
spread water quality problems in the world. Eu-
trophication can result in (i) detrimental changes
in water quality and aquatic ecosystems, and (ii) it
can severely limit human use of water as a re-
source. The impacts of eutrophication very often
call for costly management measures. The most
effective strategy for controlling eutrophication is
considered to be reducing or eliminating the input
of nutrients in the waterbody (Sas & Ahlgren,
River
Vecht
- 0.40 m
Loenen
•
Fig. 1. The research area.
1989; Ryding & Rast, 1989). In the majority of
cases, phosphorus is the growth limiting nutrient
indicating that a control strategy should include
phosphorus control (OEeD, 1982).
The eutrophication problems of the shallow
Vecht lakes Ankeveen and Kortenhoef (Fig. 1)
became apparent in the late 50's and the early 60's
but find their origin in (i) the reclamation of lake
Horstermeer in 1883, (ii) the groundwater extrac-
tions in the Gooi hills that started in 1890 and
total ca. 16 million m 3 y - 1 in the 80's, and
(iii) ongoing lowering of the polder water levels to
keep up with the peat bottom subsidence (Ver-
's-Graveland
m
Kortenhoef
~"""'---Lakes
's-Graveland
Canal - 0.15

straelen et aI., 1989; Gast, 1989). Before these
actions the area used to be recharged by ground-
water flow originating from the Pleistocene
glacial-formed Gooi ridges in the east. This
groundwater seepage, with its characteristic
chemical composition, has proved to be essential
for the very specific ecosystems that developed
(Barendregt et al., 1989; Was sen, 1990). How-
ever, because of the interferences with the origi-
nal hydrology, seepage flow to the Vecht lakes
decreased substantially and extraneous (surface)
water supply became inevitable inflicting major
impacts on the nutrient load and chemical char-
acteristics of the lakes.
Thus, the originally oligo-mesotrophic lakes
were polluted by nutrient-loaded water from the
river Vecht. With time, the accumulation capac-
ity of the lakes was exceeded, and in the last three
decades, the Ankeveen and Kortenhoef lakes ex-
perienced serious eutrophication. The number of
plant species as well as the species characteristic
of an oligo-mesotrophic habitat, have declined in
their abundance (van den Berg & de Smidt, 1985;
Roijackers & Verstraelen, 1988). Submerged
water plants have completely vanished from some
parts of the area, while blue-green algae and float-
ing leaved macrophytes with rhizomes have be-
come far more common.
Several measures to solve the eutrophication
problem have been taken e.g., diverting municipal
effluent discharges and halting the extraneous
water supply to lake Hollands Ankeveen by pre-
serving local water within the system. Other mea-
sures are in preparation e.g., supplying the system
with fresh seepage water from the polder Horster-
meer, and decreasing the groundwater extractions
in the Gooi by 50% (Provincial Board of North-
Holland, 1986; Schouten et al., 1988; Roijackers
& Verstraelen, 1991).
Since the original conservation value of the
Vecht area is linked to good quality water supply,
it is essential to determine water quantity and
quality demands of the polders and, thus, enable
selection of appropriate restoration measures.
In the following sections, water balances and
phosphorus budgets are discussed, and the re-
sults of the groundwater flow model calculations
237
are presented. Based on the results several man-
agement options are considered.
Methods
Water balances
In general, a water balance can be described by
the following equation:
Qin - Qout = bV,
where
Qin is the input;
Qout is the output;
bV is the storage change.
The boundaries of the Ankeveen and Kortenhoef
polders (the western part of which form the
Ankeveen and Kortenhoef lakes respectively)
were chosen to be the boundaries of the balance
system, because the polders are reasonably well
defined hydrological units. Qin is comprised of
precipitation (P), upward seepage (Su) and inlet
of surface water (Is)' Qout is comprised of evapo-
transpiration (E), infiltration (Fd ) and outlet of
surface water (as), Finally bV consists of surface
water (bVs ) and groundwater (bVg ). Each of these
components is discussed below.
Two main criteria determined the choice of the
period 1985-1989 as a time span to be considered
throughout the work. Firstly the period is repre-
sentative for the situation analyzed; thus all the
relevant variables are true for the present state of
the system. Secondly the data necessary for the
calculations were available. The balance equa-
tions have been established for each month dur-
ing the five-year period.
Precipitation data representative of the area
were obtained from the Royal Dutch Meteoro-
logical Institute, gauging station Loenen, situated
about 5 km from the study area. The seepage term
(Su) in the balance equation was calculated ac-
cording to the following formula (based on Dar-
cy's law):
238
where
bh is the difference in metres between the piezo-
metric head of the groundwater in the first
(Pleistocene) aquifer and the surface water
level;
e is the vertical flow resistance in days of the
covering (Holocene) layer.
Monthly averages of daily records of the surface
water levels in the Ankeveen and Kortenhoef
polders, have been used for the calculations. A
map of contour lines of the groundwater piezo-
metric level in the first aquifer averaged over the
period 1974-1980 was obtained from the Depart-
ment of Environment and Water, Province of
North Holland. Use has also been made of the
findings of Witmer on the vertical flow resistance
of groundwater in the covering layer (Witmer,
1986 and 1989).
The inlet of surface water actually depends on
the water demand in the polders, and as such, is
crucial for the water economy of the polders.
Water enters the polders in several ways:
- deliberate inlet by operation of sluices and
pumping stations;
- leakage of sluice gates;
- ship passages through the sluices;
- percolation through the soil under and along
the sluices;
- drainage through the dikes along 's-Graveland
Canal (Fig. 1);
- deliberate inlet through connections along the
's-Graveland Canal.
Since accurate monitoring of most of the water
flows is not feasible precise amounts of water that
enter the polders are not known. Therefore, the
balance equations are arranged so that the
amount of surface water inlet is computed. Sim-
ulations have been carried out to assess the val-
ues for the different items of the inlet.
Data on evaporation were obtained as decanal
values from the monthly climatological bulletins
of the Royal Dutch Meteorological Institute at
De Bilt. The meteorological station at Schiphol
was taken to be the most representative of the
evaporation conditions in the polders concerned.
Evaporation rates of open water surface, built-up
areas, pasture and swamp forrest are distin-
guished. For the open water surface, Penman's
data are used. For surfaces covered by vegeta-
tion, the reference-crop evapo-transpiration ap-
proach according to Makkink is applied
(Hooghart, 1987; CHO-TNO, 1988). The formula
used for the calculation of the I d infiltration term
(as well as the input data) is analogous to the one
used for seepage, i.e.
Id = bhle, bh<O,
with the terms bh and e as already defined.
as takes place by means of pumps, except for
the polder Stichts Ankeveen. Pumping hours are
registered on a routine basis by the Water Quan-
tity Control Authority Amstel & Vecht.
bVs is calculated as the difference in surface
water level between sequential months, multiplied
by the area of the open water surface. Analogous
to this, bVg is calculated as the difference in
groundwater piezometric level between succes-
sive months multiplied by the area ofland surface
and the storage coefficient of peaty soils.
Values for some of the factors incorporated
into the water balance calculations are question-
able. Therefore, to check the results obtained
(mainly concerning groundwater flow), chloride
balances for the polders were established. This
approach was possible due to the conservative
properties of chloride ions in natural waters.
The change in storage of chloride consists of two
terms
bCI = V1 * [CI] - V ·[CI]· .
I }-I }-1'
where bCI is the change in chloride content per
month in g, Vi and Vi _ 1 are the volumes of water
in m 3 contained in the polder in the ith-month and
the previous month respectively, and [Cn and
[Cn _ I are the chloride concentrations (mg 1- l)
in the polder water in the i-month and the previ-
ous month respectively.
The three main chloride components are pre-
cipitation, surface water and groundwater. The
data on the chemical composition of precipitation
were taken from the nearest data collection point
to the polders concerned i.e., the meteorological

station at De Bilt. Chemical analyses of the sur-
face water samples in the lakes concerned are
carried out monthly by the Water Pollution Con-
trol Authority Amstel & Gooiland. The cWoride
balance and the phosphorus budget for the Ko-
rtenhoefpolder, enable to distinguish between the
deep lake Wijde Blik and the shallow lakes in the
rest of the polder. For the groundwater chloride
content of the seepage water coming from the first
aquifer a detailed outline is available from Schot
(1989); an average value was taken (P. P. Schot,
pers. comm.).
Phosphorus budgets
Although the phosphorus budget equations were
established analogous to the chloride balance
equations, there are some notable differences. The
balance systems are defined so as to include only
surface water bodies in the polders, since the nu-
trient content of surface water in the region is of
main concern. Compared with the water balance
calculations, there are two additional input terms:
- surface runoff and interflow;
- domestic waste water discharges and com-
bined sewer overflow.
The change in storage term is also different. It
consists of:
- a change in storage of surface water;
- an exchange with the bottom sediment.
The data on the phosphorus contents of each
particular term of the equations were obtained in
a similar way as those of the chloride balance.
For nutrient loading by surface runoff and inter-
flow estimates from literature were used (Steen-
voorden, 1988). Equations were arranged so that
the exchange at the water-sediment interface can
be derived from the calculations. It should be
noted that no biochemical transformations of
phosphorus have been taken into account.
Groundwater flow model
The regional hydrological system was modelled
by using the quasi-three-dimensional groundwa-
239
ter flow program Micro-Fern (Hemker & Elburg,
1988). The modelled area comprises parts of the
Gooi, the Vecht area and the Eem valley (Fig. 1).
The model schematisation relies to a considerable
extent on a similar study carried out by Witmer
(1989), although input data have been updated.
The hydrological system has two aquifers, divided
by an aquitard. For the water balances of the
Vecht lakes only the first (upper) aquifer and the
covering layer are of interest. Groundwater flow
in the aquifers is assumed to be horizontal. It is
described by the following equation:
where:
kD is transmissivity (m 2 d - I);
k is permeability (m d - 1 );
h is hydraulic head (m);
Q is groundwater recharge/discharge (m d - 1),
In Micro-Fern the groundwater flow equations
are resolved using the finite element method. In
the covering layer only vertical flow of ground-
water is assumed to take place (for the formula
see 'Water Balances').
For modelling, the area studied is divided into
elements that are considered to be homogeneous.
The applied network consists of2242 nodal points
and 4211 triangular elements. The model input
comprises the following parameters: vertical flow
resistances and transmissivities, fixed hydraulic
heads at the outer boundaries of the modelling
area, polder water levels, groundwater withdraw-
als and groundwater recharge by net-precipita-
tion. The major outputs of the model are hydrau-
lic heads (in the first aquifer) and water balances.
Results and discussion
Water balances
The water and chloride balance data (Table 1)
relate to the year 1986, an average year with re-
gard to the annual amount of precipitation (data
240

Table J. Water balances of Ankeveen and Kortenhoef pold- Potamogeton acutif6lius, Potamogeton obtusif6lius
ers in m 3 106 y - I (1986).
and Hottonia palustris.
Hollands Stichts Kortenhoef
Ankeveen Ankeveen Polder Stichts Ankeveen. The Stichts Ankeveen
polder is supplied with surface water both from
Input the 's-Graveland Canal and from the former sand
Precipitation 3.3 4.3 12.6
pit Lake Spiegel (Fig. I). The inlet from the 's-
Seepage 4.0 0.7 1.0
Inlet from Graveland Canal takes place only in dry periods
- Vecht 29.7 via connections which are managed by farmers.
- 's-Graveland Canal 0.2 0.1 2.4 Water from Lake Spiegel enters the Stichts
- lake Spiegel 2.4 Ankeveen polder via a pumping station. The data
of the pump capacity are not reliable because the
Output
Evapo(transpi)ration 2.4 3.1 10.1 pump capacity was measured years ago. The cal-
Infiltration 3.2 4.1 22.4 culated water and chloride balances indicate that
Outlet to the pump capacity is overestimated by 15-30%.
- River Vecht 7.3 Surplus surface water from the Stichts
- 's-Graveland Canal 2.0 5.5
Ankeveen polder is pumped to Lake Spiegel, but
- lake Spiegel 0.8
the volume discharged is not registered. The sim-
Annual balance -0.1 -0.5 0.4 ulations are made to correspondent to fluctua-
tions in the release from the polder to the lake and
to the 's-Graveland Canal. The infiltration in the
series 1910-1989; Royal Dutch Meteorological Stichts Ankeveen polder is much larger than the
Institute, 1985 + updates). one in the Hollands Ankeveen polder. This is
The discussion focuses mainly on the inlet and explained by the proximity on the south of the
outlet of surface water, the seepage and infiltra- Stichts Ankeveen polder ( -1.35 m below MSL)
tion being less certain terms. A sensitivity analy- to the lower Horstermeer polder ( -3.50 m below
sis reveals that the outcome is most sensitive to MSL; Fig. 1). The acceptable waterlevel fluctua-
the uncertain variables, namely the amount of tion in the lakes are insufficient to prevent inlet of
infiltration. surface water both into the Hollands and Stichts
Ankeveen polders (A. Barendregt, pers. comm.).
Polder Hollands Ankeveen. On an annual basis
the water balance between the 's-Graveland Canal Polder Kortenhoef The calculated surface water
and the Hollands Ankeveen polder appears to be inlet from the river Vecht into the Kortenhoef
negative (Table 1). This is confirmed by the data polder ranges between 25.1 and 32.3 106 m 3 y - t,
of the pumping station i.e., only sporadically sur- with an average of 29.7 106 m 3 y - 1. The resulting
face water is let into the Hollands Ankeveen inlet is determined almost solely by the value for
polder, which has a positive net-seepage. Al- infiltration (Table 1). But, as mentioned, the input
though the calculated amounts of seepage and data for the calculation of infiltration are not com-
infiltration bear uncertainties of the input data, pletely reliable. Vertical flow resistance in the cov-
the resulting net-seepage is confirmed by the bi- ering layer (c) is derived from modelling (Witmer,
ological data collected the Department of Envi- 1986). The contour lines of the piezometric head
ronment and Water, Province of North Holland. in the first aquifer are derived by linear interpo-
Unlike the Stichts Ankeveen and Kortenhoef lation between (scarce) measuring points (Depart-
polders, in the Hollands Ankeveen polder several ment of Environment and Water, Province of
plant species characteristic of seepage conditions North Holland). The pattern of the contour lines
are found e.g., Veronica beccabUnga, Equisetum based on the groundwater flow model differs sig-
fluviatile, Sparganium emersum, Carex rostrata, nificantly from the former. With these piezomet-
 241

ric head values, infiltration would amount to not sessed for the water balance of the polder as a
more than 8.2 10 6 m 3 y - I. Since inflows from the whole. This is due to the lakes being situated in
's-Graveland Canal through connections and by the west of the polder that is an infiltration area,
drainage through the dike along the canal, are not as opposed to the east of the polder that is a
monitored, an estimated 2.4 106 m3 y- 1 (Witmer, seepage area.
1989) is used. This value, however, could easily There appears to be P-release from the sedi-
be twice as high in an unusually dry year like ment; the average release rate is -1.7 mg
1989; the calculated inlet from the river Vecht will m - 2 d - 1 in summer time. The exchange with the
decrease by the same extent, due to the linear lakes sediment is the main input into the lakes
nature of balance equations. (Table 2). The external P-Ioad is only
0.06 g m - 2 Y- 1. Since the P-concentration in the
lakes (Fig. 2) is higher than the calculated limit-
Phosphorus budgets ing level of ca. 0.08 mg 1- J (Lijklema et al., 1989)
the phosphorus release from the sediment needs
The phosphorus budget calculations (Table 2) do to be reduced.
not include the input term domestic wastewater
discharge because the households were connected Lakes Stichts Ankeveen. Similar to the Hollands
to the sewerage system in 1985 and 1986. Data Ankeveen lakes, the Stichts Ankeveen lakes are
for the winter months are lacking. situated in an infiltration area. Although the inlet
of surface water from lake Spiegel is ca. 17 times
Hollands Ankeveen Lakes. The phosphorus bud- larger than from the 's-Graveland Canal (Ta-
get of the Hollands Ankeveen lakes has a net- ble 1), the P-inputs from the two are similar. This
seepage gain as against the net infiltration as- is caused by the lower P-concentration in Lake
Spiegel because of the long residence time of sur-
face water due to a much greater average depth
Table 2. Phosphorus budgets of Ankeveen and Kortenhoef
of the lake.
lakes in kg P Y- 1 in the period April-September (1986).
Contrary to the Hollands Ankeveen lakes P-
Hollands Stichts Kortenhoef concentration in the surface water of the Stichts
Ankeveen Ankeveen Ankeveen lakes (Fig. 2b) mostly complies with
the limiting P-concentrations of ca. 0.08 mg P 1- 1
Input
estimated by Lijklema et al. (1989). This may be
Precipitation 12.1 16.2 55
Seepage 37.9 5 because of the large infiltration which may sup-
Inlet from press recharge from the sediment. But, the phos-
- Vecht 20256 phorus concentrations in the southern parts of
- 's--Graveland Canal 24.4 39.0 1346 the lakes are much higher (up to 0.3 mg P 1- I)
- lake Spiegel 138.1
than in the northern parts; they are in open con-
Surface run-off and
interflow 10.5 28.8 510 nection with the ditches in the agricultural east-
ern part of the Stichts Ankeveen polder and re-
Output ceive their P-input mainly from surface run-off,
Infiltration 184.8 279.8 376 interflow and the 's-Graveland Canal.
Outlet to
- Vecht 364
- 's-Graveland Canal 144.0 1475 Kortenhoeflakes. P concentrations in the Korten-
- lake Spiegel 2.2 hoef lakes exceed the target concentration of ca.
0.08 mg total-P 1- 1 according to Lijklema et al.
Annual balance (1989). A reduction is, however, discernible since
- surface water 3.3 - 731.0 -1437
- exchange with bottom 247.2 -790.9 - 21394
the sewage treatment plant at Horstermeer started
operating and the sewerage system was completed
242

0,25,-----------------
Effect analysis based on the groundwaterflow model

0,20 ~ The groundwater flow model was applied to es-

timate the effects of reduction and abandonment
0,15
mg PI" of groundwater extraction in the Gooi area and
0,10
.................................................................................... 1.. ·I+·~L..· . ·· ··· · .. .-\.. 1
inundation of (parts of) the polder Horstermeer
(Table 3). Complete abandonment of the ground-
0,05
water extractions will lead to a net increase in
o,oo+---~--~---~--~----"
seepage of about 10* 106 m 3 y'" I in the Vecht
1985 1986 1987 1988 1989
area. This increase in seepage will lead to a de-
- Lakes Holl.Ankeveen P..limiting level
crease in net infiltration in the polders with im-
portant conservation values (Table 3). A 50-
0,25,-------------------,
100% reduction of the extraction will lead to a
0,20
1. 7-3.2 106 m 3 y - 1 decrease in infiltration. Still,
the 50 % reduction of groundwater extractions
0,15 planned by the North-Holland Provincial Board
mgPI .,
(1986) in the Gooi will have only a minor effect
0,10 .A.... "AJ. on the net-infiltration in the polders.

'"' iW; C;JV~V~ Raising the water level in the polder Horster-
meer can significantly reduce infiltration in the
o,oo+---~--~---~--~----" Vecht area. Interestingly a rise of ca. 0.5 m in
1985 1986 1987 1988 1989
water level in about three-quartes of the polder
- Lakes St.Ankeveen P-limiting level
Horstermeer will lead to a reduction in net-
0,50,-----------------, infiltration in the surrounding polders which will
be larger than that expected from total abandon-
0.40 ment ofgroundwater extractions in the Gooi (Ta-
ble 4). Flooding the entire polder Horstermeer
0,30
mgPI"
would, as expected, largely restore seepage con-
0,20
ditions that existed in the past. This will require
adjustments in the management of water levels in
0,10 most agricultural polders of the Vecht area.
o,oo+--_.. ~--~---~--~-----j
1985 1986 1987 1988 1989 Table 3, Net seepage ( + ) and infiltration (- ) in 106 m 3 y'" I
- Lakes Kortenhoef P..limiting levei in five polders of the River Vecht area following a 100%, a
50% and a 0% reduction of groundwater extraction.
Fig, 2. total-P concentrations in the study lakes;
(a) Hollands Ankeveen lakes; Polder Present Alternative
(b,) Stichts ankeveen lakes; and extraction extraction
(c) Kortenhoef lakes,
15 7,5 0
(l06 m 3 y ... l) (106 m3 y ... l)

Horstermeer polder 30.8 30.9 30.9

(Fig. 2c). The river Vecht is the major source of Kortenhoef polder - 8.3 -7.4 - 6.9
phosphorus in the Kortenhoef polder (Table 2); Hollands Ankeveen
most of this P-input enters in the lakes' sediment, polder -1.3 -1.2 -1.2
'but can be mobilized (e.g., by wind-wave action Stichts Ankeveen - 3,3 - 3.1 - 3.0
polder
or foraging fish) and, therefore, poses a serious
's-Graveland polder -2.4 -2.1 - 1.6
obstacle to eutrophication abatement of the lakes.
 243

Table 4. Net rates of seepage ( + ) and infiltration ( - ) in 106 management objectives are: 1) eutrophication
m 3 y - 1 in the polders of the River Vecht area following in-
abatement; and 2) restoration of the conserva-
undation of the Horstermeer polder. Alternatives:
o : present situation tional values. In view of the costs and hydrolog-
I-IV: Raising the water level in ca. 75% of the Horstermeer ical constraints, only one of the objectives - the
polder with steps of 0.5 m from Mean Sea Level eutrophication abatement - can be accomplished.
(MSL) - 3.5 to MSL - 1.5 m Extensively treated wastewater does not meet
V : raising the water level in the entire Horstermeer polder
the quality and quantity demands, let alone the
and the adjacent Meeruitderdijkse polder to a level of
MSL - 1.35 m. costs of purification. Although the present policy
of replacing Vecht water supply by fresh seepage
Alternative 0 I II III IV V water from the skirts of the polder Horstermeer
polder level - 3.5 - 3.0 - 2.5 - 2.0 -1.5 - 1.35
seems to meet the quality demands, this study
Horstermeer polder 30.8 25.7 21.0 16.2 10.9 -0.6 demonstrates that the available quantities are just
Kortenhoef polder - 8.3 - 5.6 -4.1 - 3.2 -1.1 + 2.0 not sufficient to supply all the lakes. A supple-
Hollands Ankeveen -1.3 -1.2 -l.! -1.0 -0.9 -0.2
polder
mentary measure, viz. complete abandonment of
Stichts Ankeveen - 3.3 - 2.8 - 2.5 -2.1 - 1.8 + 1.2 groundwater extractions, also does not make up
polder for the shortage. The proposals to use seepage
water from the Horstermeer polder for supple-
mental supply of the Vecht lakes already meet
The infiltration based on the presented water considerable public resistance. Inundation of the
balance of polder Kortenhoef and on the ground- polder Horstermeer, is therefore, an unrealistic
water flow model differ strikingly. This latter option; however, since policy on nature conser-
model predicts a net infiltration in the polder Ko- vancy seems to favour gradual land purchase in
rtenhoef of ca. 8 106 m3 y- 1 which is ca. 14 the polder Horstermeer, a step by step realization
106 m 3 y- 1 lower than based on the water bal- of the 75% inundation option might still be a
ance. The difference in net infiltrations lies in dif- future option. Also, economically this is a viable
ferences in the basic inputs for the calculations. option on the middle term. Despite this option
In the water balance and phosphorus balance the being realized additional extraneous supply will
input for the calculating infiltration consisted of remain necessary. Although supplying the lakes
resistance values of the top layer in combination with treated and desalinated discharge water from
with a measured average hydraulic head pattern the polder Horstermeer would meet both quantity
of the first aquifer and polder water levels for the and quality demands it is economically unrealis-
polder Kortenhoef. In the groundwater flow tic.
model calculations the measured hydraulic head Reformulation of the management objectives
pattern is replaced by the calculated one, the other aiming at adhering to the conservation objectives
calculation variables remaining the same. The un- for both Ankeveen lakes and confining to
certainty about the water demand of polder eutrophication abatement for the Kortenhoef
Kortenhoef can be eliminated by intensifying the lakes seems unavoidable. Choosing an additional
monitoring network in this polder. supply option for the polder Kortenhoef will be a
trade-off between ecological benefits and costs.

Management options
The potential alternative supply options including
available quantities, costs and expected ecologi-
cal benefits (based on response predictions for
macrophytes with the hydro-ecological model
ICHORS) are given in Table 5. The main water
Conclusions and recommendations
The allowable range of variation in water level of
the Hollands Ankeveen lakes, without adversely
affecting the lakes' ecosystems, is limited. How-
ever, the inlet of water from the 's-Graveland
244

Table 5. Alternativie Sypply options. Cost-estimates are based on an interest rate of 9%,75000 guilders per ha. for agricultural
land purchase and 15000 guilders per ha. for uncultivated land purchase, a write-off over 30 years for immoveables (covering ca.
70% of the building costs) and 15 years for mechanical parts (covering ca. 30% of the building costs), 0.05 guilder per cubic m
for dephosphorization of supply water, 0.01 guilder per cubic m dephosphorized supply water for processing the simultaneously
produced unpolluted sludge and 0.05 guilder per cubic m dephosphorized supply water for processing the simultaneously polluted
sludge.
* in accordance with Barendregt et al., in press.
** non-recurrent investment.

Alternative Available supply or Estimated costs Ecological benefit*

supply options infiltration reduction per potential of measure (expressed in 00)
(in 106 m 3 y-I) measure (in
guilders per m3 ) Polder Polder Stichts
Kortenhoef Ankeveen

Present River Vecht water supply >32 0.00 52.7-68.4 60.0-70.9

Dephosphorized Lake IJ water supply >32 0.10 49.4-53.4 45.8-46.2
Dephosporized Amsterdam-Rhine Canal >32 0.15 49.5-58.2 42.8-50.6
water supply
Dephosphorized River Vecht water supply >32 0.15 55.0-61.3
Dephosphorized + denitriphicated effluent 10 1.15 55.3-65.0
from the waste water treatment plant Horstermeer
Fresh seepage water from the Horstermeer polder 17 0.15 63.7-72.4 76.5-83.9
75% inundation of the Horstermeer polder 9 0.50** 61.8-69.6 73.0-73.0
100% inundation of the Horstermeer polder 16 1.45** 67.0-74.5 100.0-100.0
Dephosphorized + denitriphicated + desalination >32 1.95 99.7-100.0
discharge water from the Horstermeer polder

Canal is not any more the main source of the
lakes' eutrophication. Since 1986 when the waste
water treatment plant in the Horstermeer polder
became operational, the phosphorus dynamics
(sedimentation and release) at the sediment is the
most important item in the lakes' phosphorus bal-
ance. Therefore, dredging of the lakes' nutrient-
loaded bottom sediment will be the most effective
measures to reduce eutrophication (Van der Does
et al., 1992). A pilot-dredging project in this re-
gard is to commence shortly.
In case of the Stichts Ankeveen lakes the sur-
face water supply from lake Spiegel is much larger
than the inlet from the 's-Graveland Canal. Con-
siderable amount of water infiltrates from the
Stichts Ankeveen polder to the adjacent Horster-
meer polder, preventing recharge of the sediment
settled phosphorus. To prevent phosphorus re-
lease from the sediment and to reduce its eutroph-
ication as well as pollution of groundwater in the
Stichts Ankeveen polder, dredging the lakes'
sediment appears to be of utmost importance.
For the Kortenhoef lakes the situation is quite
similar to that for the Stichts Ankeveen lakes.
However, in the Kortenhoeflakes infiltration, and
thus settling of phosphorus in the lakes' sediment
is found to be much larger. An alternative water
supply to these lakes is more necessary than for
the Stichts Ankeveen lakes. However, because of
the uncertainties ofthe calculated infiltration rates
a range of 8-22 106 m 3 y- 1, rather than just one
value, should be considered.
Groundwater flow-model calculations reveal
that reduction of groundwater withdrawals in the
Gooi region would not completely eliminate water
shortages in the Vecht lakes. Raising the water
level in the polder Horstermeer would largely re-
store seepage conditions that existed in the past
and greatly eliminate the necessity of extraneous
water supply. However, from a social viewpoint
the inundation option does not seem feasible in
the near future.

References
Barendregt, A., De Smidt, J. T. & M. J. Wassen, 1985. The
impact of groundwater flow on wetland communities. Col-
loques Phytosociologiques 13, Vegetation et Geomorphol-
ogie, Bailleul.: 603-612.
Barendregt, A., Wassen, M. J. & J. T. De Smidt, 1989. Re-
generatie en behoud van ecosystemen door integraal water-
beheer. In: Van Liere, L., Roijackers, R. M. M. & P. J. T.
Verstraelen (eds.), 1989. Integraal waterbeheer in het
GooisjUtrechts stuwwallen- en plassengebied: 45-58.
Barendregt, A., S. M. E. Starn & M. J. Wassen, 1992. Res-
toration of Fen Ecosystems in the Vecht River Plain: Cost-
Benefit Analysis of Hydrological Alternatives. Hydrobiolo-
gia (this volume).
Commissie voor hydrologisch onderzoek-TNO, 1988. Van
Penman naar Makkink. Een nieuwe berekeningswijze voor
de klimatologische verdampingsgetallen. KNMIjCHO,
eindrapport van de project- en begeleidingsgroep verdamp-
ingsberekeningen. Commissie voor Hydrologisch onder-
zoek TNO, 's-Gravenhage, 1988. Rapporten en Nota's, 19,
74 pp. (in Dutch).
Gast, M. K. H., 1989. Integraal drinkwaterbeheer in het Gooi
en omgeving. In: Van Liere, L., R. M. M. Roijackers, &
P. J. T. Verstraelen (eds), 1989. Integraal waterbeheer in
het GooisjUtrechts stuwwallen- en plassengebied: 90-102.
Hemker, C J. & H. Van Elburg, 1988. Micro-Fern. Users
manual. MicroComputer MultiLayer Steady Finite Element
GroundWater Modelling: 152 pp.
Hooghart, J. C, 1987. Evaporation and weather. CHO-TNO
Committee on Hydrological Research. Proceedings and in-
formation 39: 106 pp.
Royal Dutch Meteorological Institute, 1985. Climatological
Data of Stations in The Netherlands. A climatology of the
main observatory at De Bilt for the period 1901-1984,25,
56 pp.
Lijklema, L., J. H. Janse & R. M. M. Roijackers, 1989. Eu-
trophication in The Netherlands. Wat. Sci. Tech. 21: 1899-
1902.
OECD, 1982. Eutrophication of waters. Monitoring, Assess-
ment and Control. OECD Paris, 154 pp.
Provincial Board of North-Holland, 1986. Provincial Ground-
water Management Plan North-Holland: 100 pp. (in
Dutch).
Roijackers, R. M. M. & P. J. T. Verstraelen, 1988. Ecological
investigations in three shallow lakes of different trophic
level in The Netherlands. Verh. int. Ver. Limnol. 23: 489-
495.
245
Roijackers, R. M. M. & P. J. T. Verstraelen, 1991. Eutroph-
ication abatement programme for three North Holland
Vecht lakes. Verh. int. Ver. Limnol. 24: 768-772.
Ryding, S-O. & W. Rast, 1989. The control of eutrophication
oflakes and reservoirs. Man and the Biosphere series. Pub-
lished by the UNESCO and the Parthenon Publishing
Group, 314 pp.
Sas, H., 1989. Lake restoration by reduction of nutrient load-
ing: expectations, experiences, extrapolations. Academia
Verlag Richarz GmbH. 497 pp.
Schot, P. P., 1989. Grondwatersystemen en grondwaterk-
waliteit in het Gooi en randgebieden. Rijksuniversiteit
Utrecht. Interfacultaire Vakgroep Milieukunde: 44 pp. (in
Dutch).
Schouten, P. H., Verstraelen, P. J. T. & J. J. Buntsma, 1988.
Restoration of the Vecht lakes ecosystems surrounding the
Horstermeer polder: a feasibility study. - Agricul. Wat.
Managmt. 14: 471-479.
Steenvoorden, J. H. A. M., 1988. How to reduce nitrogen
losses in intensive grasland management? Ecol. Bull. 39:
126-130.
Van den Berg, W. J. & J. T. de Smidt, 1985. De vegetatie van
het Oostelijk Vechtplassengebied 1935-1980. Stichting
Commissie voor de Vecht en het Oostelijk en Westelijk
Plassengebied. 155 pp. + annex.
Van der Does, J., P. J. T. Verstraelen, P. Boers, J. Van Roes-
tel, R. M. M. Rooijackers and G. Moser, 1992. Lake res-
toration with and without dredging of phosphorus-enriched
upper sediment layers. Hydrobiologia 233: 197-210.
Verstraelen, P. J. T., R. M. M. Roijackers & L. Van Liere,
1989. Geohydrologie van de GooisjUtrechtse stuwwal en
het Vechtplassengebied. In: Van Liere, L., Roijackers,
R. M. M. & P. J. T. Verstraelen (eds), 1989. Integraal
waterbeheer in het GooisjUtrechts stuwwallen- en plassen-
gebied: 151-165 (in Dutch).
Wassen, M., 1990. Water flow as a major landscape ecolog-
ical factor in fen development. Ph.D. Thesis, State Univer-
sity of Utrecht, 199 pp.
Witmer, M. C H., 1986. Management of groundwater and
surface water resources in the infiltration area 'Het Gooi'
and adjacent wetlands. Conjunctive Water Use. Proceed-
ings of the Budapest Symposium. IAHS Publ. 156: 534-
547.
Witmer, M. C H., 1989. Integral water management at re-
gional level. An environmental study of the Gooi and
Vechtstreek. Ph.D. Thesis, State University of Utrecht,
172 pp.
Hydrobiologia 233: 247-258, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 247
© 1992 Kluwer Academic Publishers.

Restoration of fen ecosystems in the Vecht River plain: cost-benefit

analysis of hydrological alternatives

A. Barendregt, S. M. E. Stam & M. J. Wassen

Department of Environmental Studies, University of Utrecht P.O. Box 80.115,3508 TC Utrecht,
The Netherlands

Key words: fen ecosystems, hydrology, deterioration, restoration, cost-benefit analysis

Abstract

In the Vecht river plain, the vegetation diversity has been diminished during recent decades, due to
hydrological changes and resulting eutrophication. Different options are discussed for hydrological
management aimed at abating the eutrophication and at stimulating the characteristic mesotraphent
succession phases of the fen vegetation. The ecological impact of the management plans are estimated
with the hydro-ecological model ICHORS; this model takes into account the relations between the
non-biotic environment and the presence of hydrophytes and phreatophytes. The impact of the hydro-
logical options is assessed, after rearrangement of the predicted values of the species to values of the
required fen vegetation. These results are incorporated in the management options with the water bal-
ances of the polders. Information on the costs of the options is used as a basis for a cost-benefit anal-
ysis. Technical hydrological solutions such as purification of the supplied water might contribute to the
eutrophication abatement, but they are on their own unlikely to restore the ecosystems. Contrastingly,
options involving the restoration of the original hydrology would probably lead to the characteristic
mesotraphent fen vegetation once more, and the financial implications of this are likely to be the more
favourable in the long-term.

Introduction Vecht River plain in The Netherlands, a wetland

Fen ecosystems are endangered, for their location
mostly near a river or an estuary can be eutroph-
icated easily. Moreover, after draining or reclaim-
ing, fen ecosystems become arable lands, desired
in agriculture (Williams, 1990). This hydrological
change influences the chemistry of water in the
ecosystem too, so that the ecological equilibrium
is affected both in nutrients and in major ions.
The restoration of polluted aquatic ecosystems
might be achieved in the hydrological manage-
ment. The question remains what type of man-
agement is profitable.
This problem is investigated in the Eastern
area situated north of the city of Utrecht and
southeast of Amsterdam (Fig. 1). Shallow fen
lakes (turfponds), marshes and ditched pasture
lands are characteristic of the Vecht area. The
original fen vegetation, so rich in species, is de-
teriorating as present. Species characteristic of
mesotrophic conditions have disappeared or have
become very rare, due to eutrophication of the
surface water and to structural changes in the
groundwater flow (Barendregt et al., 1989; Was-
sen, 1990; Roijackers & Verstraelen, 1991). Var-
ious water management options have been sug-
gested for restoring the vegetation.
This paper assesses the ecological implications
248
Fig. 1. Part of the Vecht River plain with Lakes Kortenhoef
and Lakes Stichts Ankeveen. Numbers correspond to the
locations of hydrological options described in Table 1.
as well as the financial consequences of possible
changes in water management. Past and present
geographical and hydrological conditions are out-
lined first, while the deterioration of fen vegeta-
tion due primarily to nutrient enrichment is dis-
cussed secondly. Next, various measures aimed
at reversing eutrophication trends - involving
management of the hydrological regime - are dis-
cussed. Ten options are considered, and here, not
only the likely ecological impacts (fen species re-
sponses) are taken into account, but also the cost
benefits.
Description of the lake areas
The Eastern Vecht area is bordered by the Pleis-
tocene ice-pushed hill ridge Ret Gooi to the east
and by the Vecht River to the west. It is a polder
area below mean sea level, where the level of
surface water is artificially controlled. The natu-
ral groundwater flow originates from the hill ridge
and is directed to the plain, where seepage water
discharges at the edge of the ridge. Before culti-
vation of the area, the high groundwater levels
induced the formation of peat as in the flood zone
next to the river, where eutrophic woodland peat
existed. Also, in the east of the river plain, under
the influence of nutrient-poor and calcium-rich
seepage water from Ret Gooi, mesotrophic peat
of Carex species was formed. In the intermediate
area, the peat rose above the groundwater level
and was only influenced by infiltrating rain water.
Under these nutrient-poor conditions, Sphagnum
peat developed.
Over the last few centuries, the natural hydro-
logical regime has changed significantly due to
human influence (Witmer, 1989). In the eleventh
century, ditches were dug in the wetland area to
drain the soil for agricultural purposes. The drain-
age caused mineralisation and setting of the peat.
The natural water flow to the River Vecht dimin-
ished. From the sixteenth century onwards, the
water was discharged artificially through ditches
and canals; peat excavation and dredging could
then be intensified. Peat was dredged up to the
Pleistocene sand deposits. The excavation re-
sulted in a landscape of elongated turfponds al-
ternating with small residual strips of land, that
were washed away. This resulted in the formation
of large, shallow peat lakes in the eighteenth cen-
tury.
The reclamation of Lake Rorstermeer in 1882
caused a structural change in the groundwater
flow system, such that the surface water level of
the new polder was significantly (ca. 2.5 m) lower
than that of surrounding areas. In order to main-
tain this low level, 30 x 106 m 3 water is discharged
annually from the Polder Rorstermeer into the
Vecht River. This low level also causes drainage
of water from the adjacent polders, which results
in water shortages in these polders in summer. At
present there is an annual deficit of 20 x 106 m 3 in
the polder Kortenhoef and of 3 x 106 m 3 in the
polder Stichts Ankeveen (Verstraelen et al.,
1992). A secondary effect is the discharge of
brackish water from the deeper aquifer to the sur-
face in the centre of the Polder Rorstermeer (Wit-
mer, 1989; Schot, 1991). A second cause of the
water shortages in the plain is the groundwater
abstraction by water supply companies since
1888. The population of Ret Gooi increased and,
subsequently, the demand for drinking water rose

(in 1988 15.6 x 106 m3 ; Schot, 1991). This signif-
icantly diminished the natural groundwater flow
into the wetlands.
To counteract water shortages, the inlet of sur-
face water, mainly from the Vecht River, was in-
evitable. Because of this partial substitution of
ridge water by surface water, the chemical com-
position of the water changed. Indeed, the qual-
ity of wetland water deteriorated because of the
poor quality of water entering from the Vecht
River. Since 1930, the amounts of nutrients and
major ions in the Vecht River have increased due
to pollution of the Rhine River and the discharge
of wastewater from the city of Utrecht (Baren-
dregt et al., 1989). At present the main source of
eutrophication in the Kortenhoef Polder is the
water from the Vecht River, which may supply up
to 20 tons phosphorus a year (Verstraelen et al.,
1992). The supply of nutrient-rich surface water
to the lakes over many years has increased the
nutrient content of the sediments (Verstraelen
et al., 1989). For example the concentration of
phosphorus in the mud of the Vecht lakes has
doubled since the beginning of this century. No-
tably, a third of the total amount of stored phos-
phorus is available to algal growth (Visser et al.,
1989).
The ecosystems: their deterioration, restoration
and water management options
After the creation of turfponds and larger lakes in
the eighteenth century, the successional phases
led to the development of a great variety of plant
communities. A diverse vegetation with Chara
species occurred in open water, and in the zone
between water and land a mixture of succession
phases was observed. When this vegetation
growed tightly together forming a floating inter-
tangled mat of plant roots, a species-rich quaking
fen developed. At other localities, there were
mashes with Phragmites australis and Carex spe-
cies. Shrubs and trees appear in succession when
plant growth extends above groundwater level.
This is how the last successional phase of Alnus
and Betula woodlands starts.
249
Since the first documented releves of the Vecht
wetland area in 1935, the vegetation has changed
(Best et al., 1984; Van den Berg & De Smidt,
1985; Roijackers & Verstraelen, 1988; Barendregt
et al., 1990). This is primarily due to changes in
hydrology. The decreasing abundance of species
that grow under nutrient-poor conditions is ob-
served and the characteristic mesotrophic wet-
land ecosystems are now restricted to seepage
areas that are unaffected by extraneous supply
water (Barendregt et al., 1989). Trends observed
over the last decades include those of succession,
eutrophication and acidification. Thus, there is a
general tendency for aquatic vegetation and early
succession phases to disappear, and this is fol-
lowed by an increase of later phases with acid
reedland and woodlands. Aquatic plants disap-
pear due to an explosive growth of algae and
increasing water turbidity. At the same time spe-
cies characteristic of mesotrophic fens disappear.
Eventually, this diminishes the biological diver-
sity in the wetlands. Nowadays the aquatic veg-
etation of the Lakes Stichts Ankeveen includes
no more than 20 vascular species and 4 Charas,
in Lakes Kortenhoef only 9 common aquatic plant
speCIes occur.
The ecological value of the Vecht River plain
and the key role of groundwater supply are rec-
ognised by national and provincial authorities
(Ministry of Agriculture, Nature conservation and
Fisheries, 1990; Province of North-Holland,
1990; Province of Utrecht, 1990). Until now,
however, this has not prevented further deterio-
ration of the wetlands. Restoration of the char-
acteristic mesotraphent fen ecosystem requires
decisive measures regarding the supply of ground-
water and the prevention of eutrophication. The
problem can be dealt with in several ways.
(1) The best route to a structural restoration is to
revert to the hydrological situation that ex-
isted before the reclamation of the Polder
Horstermeer. A large amount of water char-
acteristic of the system is drained from the
surrounding polders and the hill ridge into the
Polder Horstermeer from which it is pumped
into the Vecht River and is lost to the wetland
250

ecosystems. Loss ofthis water can be avoided
by inundation of the Polder Horstermeer
(Witmer, 1989).
(2) The discharge of groundwater from the ridge
into the river plain can be stimulated by a
structural restoration of the groundwater sup-
ply. A fifty percent reduction of groundwater
abstraction in Het Gooi within ten years has
been proposed in a Groundwater Manage-
ment Plan (Province of North-Holland,
1986).
(3) It is possible to recycle a large part of the
water. Fresh seepage water from the outskirts
of the Polder Horstermeer originates from the
adjacent polders, where it infiltrates. This
seepage water can be pumped back into these
polders instead of pumping it into the Vecht
River.
(4) Short-term improvements can be achieved by
reducing the nutrients in the artificially sup-
plied water. The supply of polluted water from
the Vecht River and Lake Spiegelplas can be
altered technically by introducing water of a
better quality. The quality of supply water de-
termines very much the conditions for the fen
vegetation. Therefore, choices concerning the
type of water to be used for supply should be
made with great care (Barendregt et al., 1986).
(5) Structural reduction of the nutrient loads in
the Vecht River can only be realised in the
long term. If the goal is a river with meso-
trophic water, then we will have to stop the
eutrophication and salinisation of the Rhine
River; waste water discharges will also have
to be eliminated in the whole region.
For Lakes Kortenhoef and Ankeveen, manage-
ment policies have been formulated with the aim
of reducing eutrophication (Roijackers & Ver-
straelen, 1988; Schouten et al., 1988). One such
measure relates to the waste water discharge plant
in the Horstermeer, in operation since 1986. This
plant brought to an end the discharge of untreated
waste water in the area. Removal of the lake sedi-
ment has also been suggested (Schouten et al.,
1988; Roij ackers & Verstraelen, 1991).
Ten options in water management are consid-
ered in this study (Table 1 and Fig. 1). All pos-
sible sources of supply water for Lakes Korten-
hoef and Lakes Stichts Ankeveen will be
compared to assess the ecological impact of hy-
drophytes and phreatophytes in turfponds and
ditches. The emerging will be ranked with the aim
to choose the one that can lead to optimal con-
ditions for a mesotraphent fen vegetation. A num-
ber of the available water types are very rich in
nutrients or are brackish, and can as such not be
used for supply purposes. Only after purification
(dephosphorization and denitriphication) or de-
salination can supply be considered (Starn &
Barendregt, 1990). The effect of P-stripping
(FeCI 3 treatment) on ionic composition is calcu-

Table 1. Hydrological options for Lakes Kortenhoef (1-10) and Lakes Stichts Ankeveen (1,2,5,7,9,10). Assumed is a 100% re-
placement of surface water by the option. Chemistry derives from average data in the period 1985-1990 (Starn & Barendregt, 1990).

1. Status quo, present surface water in the polders

2. Present supply water from the River Vecht (Kortenhoef) or Lake Spiegelplas (Stichts Ankeveen).
3. Supply with desalinated surface water derived from the Polder Horstermeer. The upward seepage in the centre of the Polder
is saline (up to ca. 4000 mg 1- 1 CI-) and can be used as supply water only after desalination.
4. Supply with (extra) purified effluent from the sewage treatment plant in the Horstermeer.
5. Supply with fresh seepage water from the border of the Polder Horstermeer (southern part, Kortenhoef; northern part,
Ankeveen).
6. Supply with dephosphorized water from the Vecht River. The ionic composition of the river water is based on analysis of
water from the river south of De Nes, where the water is not pollution by the effluent from the treatment plant.
7. Supply with dephosphorized water from Lake IJmeer.
8. Supply with dephosphorized water from the Amsterdam-Rijn Canal.
9. Supply with dephosphorized water from the Amsterdam-Rijn Canal and the Polder Bethune. The water types are to be mixed
in the ratio 5 to 1.
10. Supply with groundwater from the hill ridge in the east. This variant requires restoration of the groundwater flow system.

lated and the total effect estimated by comparing
available data on ion composition of water before
and after phosphorus removal. Aiming to restrict
maximum salt content to 100 mg 1- 1, 85 % of the
water from Polder Horstermeer needs to be
treated. For P0 4 and N0 3 , desired concentra-
tions of P < 0.05 mg 1- 1 and N < 2.0 mg 1- 1 are
taken into account.
Methods
Species responses
The ecological impact of changes in hydrology
was estimated with the hydro-ecological compu-
termodel ICHORS (Barendregt & Bootsma,
1991), that describes the relation between the
presence of aquatic and mire plant species and
the environmental conditions. For this study two
versions of the model ICHORS were used: ver-
sion 3.1 was used for turfponds and version 3.2
for ditches. For each plant species a regression
model with the essential environmental factors,
sorted out by stepwise multiple logistic regression
analysis, is incorporated in this computermodel.
With an input of a number of morphological, hy-
drological and chemical variables that describe a
certain environmental condition, ICHORS pre-
dicts the responses of hydrophytes and phreato-
phytes in the range between 0 and 1. Input of the
present hydrological condition and of some alter-
native conditions offers the required responses of
the plant species on hydrological changes. By
comparing this ecological impact of different hy-
drological options, the best option can be selected.
The ionic composition of the water types is an
important variable in the calculations. Other var-
iables, such as soil type and morphology of tur-
fpond or ditch, are not influenced by qualitative
changes in water supply. Examples of ICHORS
input (water chemistry) and ICHORS output for
four hydrological options are given in Table 1. It
is assumed that the new supply water completely
replaces present surface water. This enables the
comparison of the extreme impact of the different
supply options. In the actual situation, surface
251
water is a mixture of supplied surface water, rain
water and groundwater.
Ecological assessment
For the purpose of this study it is necessary to
translate the species responses as predicted by
ICHORS to conclusions on a higher ecological
level. Based on their habitats the plant species
were divided into different ecological groups as
described in Table 2. These groups partly match
the plant communities of polder water mentioned
in the Provincial Water Management Plan (Pro-
vince of North-Holland, 1990).
The calculation methods described below were
conducted separately with the results of both
ICHORS 3.1 and 3.2. Per hydrological option a
mean response for each group was obtained. Just
taking the mean of all responses of species has the
disadvantage that species with consistently low
responses (rare species) have no weight compared
to more common species with higher responses.
Therefore, for each species the responses were
calculated as a percentage, the highest occurring
response of a species (optimal hydrological vari-
ant) being 100 percent. Differences in absolute
responses were thus neglected. Mean group re-
sults in terms of percentage were obtained. Af-
terwards the mean results of some groups (Ta-
ble 3) were taken together.
Special attention in this study is paid to the
fresh water species (groups 2, 3 and 4 for both
turfponds and ditches) for they are characteristic
for fen vegetation. For these species a second,
more detailed method of assessment was used
that also accounts, besides the response, for their
rarity. For data on species rarity a code list was
used; the rarity codes (1-9) are based upon the
number of square kilometres in which a species
occurs (Province of North-Holland, 1987). Fen
species which especially need stimulation are the
ones that are endangered. For that reason an
estimated change in response of rare species
(which have a low rarity code) was considered of
more importance than a change in response of
more common plant species. Therefore a rarity
252

Table 2. Division of species from ICHORS 3.1 and ICHORS 3.2 in ecological groups. Number of species per group in brack-
ets. Group 2, 3 & 4: fresh water species; group 5 & 6: eutraphent species; group 7 & 8: brackish (tolerant) species.

Groups of species: ICHORS 3.1 (turfponds)

I Common fen species (22)

2 Succession species (10): more or less endangered, mostly floating species of clear water
3 Mesotraphent species (10): endangered species of unpolluted and undisturbed locations in the river plain
4 Gradient species (10): species characteristic for the zone between constant seepage of groundwater and infiltration
5 Eutraphent turfpound species (10): hydrophytes and phreatophytes indicating eutrophication and disturbance in fresh turfpounds
6 Bank species (12): bank species mostly living under eutrophic conditions

Groups of species: ICHORS 3.2 (ditches)

I Common fen species (28)

2 Hydrophytes of clean fresh water (13): mesotraphent species of nutrient-poor water
3 Species offresh calcium-rich water (13): endargered mesotraphent species occuring in ditches with water of the CaHCOrtype
4 Mesotrophic bank species (28): bank species often prefering nutrient-poor and fresh water
5 Eutraphent hydrophytes (10): mostly floating water plant species indicating eutrophic conditions
6 Eutraphent phreatophytes (16): phreatophytes tolerant to eutrophic conditions
7 Species of clear nOll-eutrophic water (14): species with some tolerance to brackisch water
8 Species of low brackisch dynamic conditions (13): species tolerating or prefering slightly brackisch and dynamic conditions

score was calculated: rarity score = (10 - rarity
code) x response. The rarer a species, the higher
the score. For each hydrological option the mean
rarity score for the fresh water species was cal-
culated, using absolute responses and responses
as percentage.
As a result for the fresh water species the means
were known of: (1) absolute responses, (2) re-
sponses as a percentage, (3) rarity scores based
on absolute responses and (4) rarity scores based
on responses as a percentage. To obtain one value
for fresh water species per hydrological option
instead of four, it is not possible to simply calcu-
late a combined average of these means. They
were obtained by four ways of transformation
and have different dimensions. This was over-
come by converting, by way of transformation,
the means into percentage, taking the highest
mean (optimal hydrological option) as 100 per-
cent. Finally one average value for fresh water
species per hydrological option was obtained.
In the calculation of the cost-benefits the final
results of freshwater species in the turfponds and
the ditches are integrated in one value of poten-
tial ecological improvement; as especially the tur-
fponds need ecological stimulation, they have a
double weight in this calculation. The potential
ecological improvement of a mixture of two dif-
ferent types ofchemistry (for instance: option with
less water supply) is a linear interpolation of the
ecological impact of these two types of chemistry.
It should be noted that the final value of poten-
tial ecological improvement is a weighted mean
which is based on ecological principles. This value
is expressed, however, only on a relative scale.
Ecological impact of the hydrological options
The results of the calculations are summarized in
Table 4 (the means of responses per species
group) and in Fig. 2 (second method, with the
fresh water species).
Lakes Kortenhoef
The supply of unpurified water from the Vecht
River is a very unfavourable option. The re-
sponses of the fresh water species (24-33%) are
the lowest and the responses of the eutraphent
species (67-69%) are high (Table 4). The present
supply, therefore, needs to be discontinued.
The most favourable option is supply of desal-
inated surface water from the Polder Horster-
 253

Table 3. An example of ICHORS input (water chemistry) and ICHORS output (response of species). Responses of species in
reference to four different water supply options of the Lakes Kortenhoef: Vecht = Vecht River water; AlB = dephosphorized water
from Amsterdam-Rijn Canal and Polder Bethune. Seep = fresh seepage water from Polder Horstermeer; Grw = (sub-) regional
groundwater.

ICHORS 3.2 input: Chemical composition of the supplied water in mg 1- I

Vecht AlB Seep Grw

pH 7.8 7.2 6.7 6.9

CI 222 139 100 25
HC0 3 172 188 250 170
S04 86.5 58.1 45.0 15.0
Na 127 86 60 15
K 9.3 6.0 2.0 1.9
Ca 85.6 80.9 80.0 45.0
Mg 13.3 10.8 5.0 5.5
Fe 0.08 0.50 10.0 2.5
Si 1.2 2.5 7.0 7.0
NH 4 0.5 0.9 2.5 0.8
N0 3 14.6 13.8 0.3 0.4
P0 4 0.98 <O.ol 0.40 0.36

ICHORS 3.2 output: Absolute responses of species (range 0-1).

Vecht AlB Seep Grw

Mesotraphent species
Achillea ptarmica 0.008 0.008 0.066 0.116
Cardamine flexuosa O.ol8 0.021 0.051 0.044
Elodea canadensis 0.030 0.022 0.022 0.031
Equisetum fluviatile 0.191 0.271 0.451 0.457
DD-Equisetum fluviatile 0.032 0.052 0.106 0.107
Hottonia palustris 0.000 0.001 0.003 0.066
Potamogeton obtusifolius 0.020 0.040 0.230 0.123
Potamogeton trichoides 0.034 0.033 0.043 0.071
Potentilla palustris 0.013 0.009 0.204 0.121
Ranunculus flammula 0.347 0.339 0.217 0.408
Veronica beccabunga 0.008 0.011 0.017 0.025
Chara globularis 0.000 0.001 0.000 0.003
Mean value (n = 12) 0.059 0.067 0.118 0.131

Eutraphent species
Hydrophytes (mean, n = 10) 0.345 0.158 0.245 0.202
Phreatophytes (mean, n = 16) 0.248 0.269 0.266 0.250
Mean value (n = 26) 0.297 0.214 0.256 0.226

Brackish species
Mean value (n = 13) 0.156 0.109 0.058 0.047

meer. Compared with the supply from the Vecht ter supply. That would substantially stimulate
River, this would considerably increase the fresh water species besides common and bank
chances for both common species and fresh water species (Fig. 2). Supply of fresh seepage water
species. Another favourable option is groundwa- derived from the Polder Horstermeer would have
254

Table 4. Summarized results of calculations with ICHORS for all hydrological options (numbers 1-10: see Table 1), for both
Lakes Kortenhoef and Lakes Stichts Ankeveen. Mean results (responses converted in % of maximum value) of fresh, common,
eutrophic, bank and brackish species for turfponds (t) and ditches (d). In brackets the number of species per group.

2 3 4 5 6 7 8 9 10

Lakes Pres Vecht HM RWZI Seep V IJ A AlB Grw

Kortenhoef

Fresh-water t (30) 31 24 53 31 41 30 26 29 29 53
species d (40) 45 33 64 33 49 36 33 34 34 59
Common t (22) 56 55 65 48 53 63 60 57 56 58
species d (28) 63 46 71 59 58 51 53 50 51 65
Eutraphent t (10) 38 70 46 53 43 72 60 60 58 59
species d (26) 50 67 45 54 54 54 41 50 48 51
Bank species t (12) 75 66 50 56 47 66 64 63 60 59
Brackish species d (27) 52 49 36 39 23 56 82 49 46 22

Lakes Pres Spie Seep IJ AlB Grw

Stichts Ankeveen

Fresh-water t (30) 48 25 61 28 32 73
d (40) 57 43 62 36 36 79
Common t (22) 75 64 56 66 64 71
species d (28) 66 62 72 56 56 71

Eutraphent t (10) 43 50 36 64 64 58
species d (26) 49 66 65 44 55 63
Bank species t (12) 79 82 52 68 66 79
Brackish species d (27) 50 59 25 84 49 25

a positive impact on fen vegetation. In addition,
those options would have a positive effect by sub-
stantially reducing the changes of brackish spe-
CIes.
The supply from the Vecht River near the sluice
Het Hemeltje (Fig. 1) could be replaced by sup-
ply with dephosphorized water from the Vecht
River south of De Nes (without effluent of the
treatment plant). This would have positive effects
on both common species and fresh water species.
A negative effect of dephosphorization using
FeCl 3 is a small increase in chloride concentra-
tion, increasing the chances for brackish species.
The effects on vegetation of supply with dephos-
phorized water from the Amsterdam-Rijn Canal,
with or without water from the Polder Bethune,
are similar. Compared with the present supply,
the chances for eutraphent species would dimin-
ish; for fresh water species of turfponds and com-
mon species the situation would improve some-
what. In general, inlet of this water would result
in a slight improvement of chances for fen vege-
tation.
Supply of dephosphorized water from Lake
IJ meer would affect the fresh water species sim-
ilar to that of Vecht River water now. Responses
of common species would increase somewhat.
Since the most important impact of such supply
would be a substantial increase of brackish spe-
cies, this alternative must be discouraged.
Lakes Stichts Ankeveen
For fresh water species, the supply water from
Lake Spiegelplas is the most unfavourable option

Kortenhoef Ankeveen
6,-"
-7 7 .. ~
8=
9 .. ~
9=
10 . 10 .~ •...~~~!!~
. ~~
30 40 50 60 70 80 90 100% 30 40 50 60
_ Turfponds
, , Ditches
Fig. 2. Ecological impact (expressed in % of the maximum
option; x-axis) in turfponds and ditches of ten hydrological
options (see Table 1; y-axis) in Lakes Kortenhoef and Lakes
Stichts Ankeveen on fresh water species. Vertical lines indi-
cate the impact of present supply (option 2).
(25-42 % in Table 4). Besides, the brackish spe-
cies score high with this option. For these reasons
the quality of this water does not permit the de-
velopment of a rich fen vegetation. Improvement
of conditions for fen vegetation can be achieved
by supply of fresh seepage water from the Polder
Horstermeer or by groundwater supply (option 5
& 10, Fig. 2). The best scoring option for fresh
water species in both turfponds and ditches is
groundwater supply. In addition to this positive
effect on fen vegetation, these options show the
lowest scores for brackish species.
Supply of dephosphorized water from the
Amsterdam-Rijn Canal and the Polder Bethune
will improve the responses of fresh water species
of turfponds to a minor extent. However, for
fresh water species of ditches an opposite effect
is predicted. The responses of brackish species
will be somewhat reduced. Summarizing, no
marked change in conditions for fen vegetation is
expected. The effects on fresh water species of the
supply of dephosphorized water from the Lake
IJ meer are comparable to those for dephospho-
rized water from Amsterdam-Rijn Canal and
Polder Bethune. However, this option would vig-
255
orously stimulate brackish species and should
therefore be firmly discouraged.
Cost-benefit analysis
Based on the likely impact of the options, we can
set priorities on the restoration measures. The
decision to implement an option will, however,
depend also on its financial consequences. In this
respect a cost-benefit analysis is necessary. This
analysis requires data on the hydrology (Ver-
straelen et at., 1992), the likely ecological impact
70 80 90 100%
of the suggested options (this paper) and the es-
timate costs (Verstraelen et al., 1992). The avail-
able amount of the different types of supply water
was taken into account. In practice some options
will have to be combined with other options, to
obtain the required amount of supply water for
the polders and lakes. The net infiltration in the
Polder Kortenhoef is 20 x 10 6 m 3 and in the
Polder Stichts Ankeveen it is 3 X 10 6 m 3 . Less
infiltration in a polder decreases water shortages
which consequently leads to a lower amount of
supply water needed and a relatively larger con-
tribution of discharged groundwater in the water
balance of the polder concerned. An inundation
of the Polder Horstermeer, either totally, or for
75%, would induce such changes in the water
balance of the surrounding polders. Thus ground-
water seepage becomes more important, with un-
polluted fresh water extending into the river plain.
The inundation of 75% of the Polder Horster-
meer, excluding the built-up area, requires an an-
nual investment of 5.9 million guilders for 30
years. Total inundation of Polder Horstermeer
requires 20.35 million guilders a year.
By plotting the potential ecological improve-
ment (stimulation of the characteristic ecosys-
tems) against the costs (Fig. 3), the ecological im-
pact of various management options can be
evaluated in terms of required investment. De-
phosphorized water from the Vecht River south
of the treatment plant (option c) is preferable to
the other supply options with dephosphorization
(a & b), for the costs are almost equal and the
ecological improvement is higher. However, with
256

• times as much and therefore is not a realistic

100
9

option. The inundation of 75% of the Polder

c
90
Horstermeer (f) yields an ecological benefit of
~ some percents higher than with (e) in Polder Ko-
•
e

""I•
~
C.
80
rtenhoef. Because of the minor hydrological im-
.~
m 1 plications the option (f) in Polder Ankeveen re-
.~ 70
a
<;
quires a comparatively larger sum of money with
~
m respect to the ecological benefits. In Polder
.~ 60
Ankeveen the total inundation (g) results in a
~
50 ~
Stichts Ankeveen ::: ...
positive water balance and in the optimal ecolog-
Kortenhoef = • ical impact; in Polder Kortenhoef option (g) of-
fers ecological improvement, although the hydro-
logical implications indicate only 50 % reduction
Stichts Ankeveen
I
4
i
5
of the water shortages. A positive water balance
,---,,-----,-----,--~~~~-~~~~~,'V' Kartenhoe! in Polder Kortenhoef is possible with option (h):
2 4 6 8 10 12 14 16 18 20 22 24 39

~ Costs (x 106 Dutch guilders year-') option (g) with an extra total supplementary in-
vestment of 160 million guilders. This creates
Fig. 3. Cost-benefit analysis of management options and
ecological development in fresh fen vegetation. Costs accord-
possibilities for supplementary activities, such as
ing to Verstraelen et al. (1992). Ecological benefit expressed in changing of levels of surface water in surround-
percentage of maximum option (see text). Codes: ing polders, so that an ecologically optimal chem-
ARK = Amsterdam-Rijn Canal + Polder Bethune; HM = istry (resembling groundwater) will be reached.
Polder Horstermeer). Option (i), that can only be applied in Polder
0= no change, status quo in water supply (from Vecht River
in Kortenhoef and from Lake Spiegelplas in Ankeveen); a, b
Kortenhoef, obtains the optimal ecological im-
and c = 100% dephosphorized water from IJmeer, ARK and provement and demands at the same time an ex-
Vecht, respectively; d = 10 X 10 6 m 3 dephosphorized effiuent tremely high investment.
and 10 x 10 6 m3 dephosphorized water ARK; e = in Korten-
hoef: 13 x 10 6 m3 fresh seepage water HM and 10 x 106 m 3
dephosphorized water ARK; in Ankeveen: 3 x 10 6 m3 fresh
Discussion
seepage water HM; f = in Kortenhoef: 7 x 10 6 m 3 less infiltra-
tion and 13 x 106 m3 dephosphorized water ARK; in
Ankeveen: 1.5 x 106 m 3 less infiltration and 1.5 x 10 6 m3 de- The essential problem of quality management and
phosphorized water ARK; g = in Kortenhoef: 10 x 106 m3 1ess nature conservation in the Vecht River plain is the
infiltration and 10 x 106 m 3 dephosphorized water ARK; in structural water shortage and its contaminant
Ankeveen: 4 x 106 m3 less infiltration; h = in Kortenhoef:
water supply with nutrient-loaded water from the
20 x 106 m 3 less infiltration (positive water balance); i = in Ko-
rtenhoef 20 x 106 m 3 desalinated water from HM. river. Solutions to this ecological problem can be
formulated in terms of technical options to pre-
vent eutrophication or more sustainable options
the same annual amount of money, a higher eco- to prevent a negative water balance.
logical improvement would be obtained by pump- By the technical options the water is purified,
ing fresh seepage water from the outskirts of the so that the input of nutrients is reduced to meet
Polder Horstermeer (e) to the nature reserves in the demands of eutrophication abatement. Op-
Polder Kortenhoef and Polder Stichts Ankeveen. tions included in this study are the purification of
The supply of extra purificated effluent from the different allochthonous water sources, further pu-
treatment plant (d), that according to the avail- rification of effluent from the treatment plant, and
able quantity has to be combined with other sup- desalinisation of water from the Polder Horster-
ply water, would result in the same ecological meer. These supply options can be realised on the
improvement as dephosphorized water from the short-term. However, the costs of desalinisation
Vecht River (about 60%), but costs about seven are extremely high and the ecological benefit of

the other options is relatively small. A supply
option that is realistic, both in terms of expendi-
ture and in ecological improvement, is the recy-
cling of fresh seepage water from the outskirts of
the Polder Horstermeer. This option utilizes orig-
inal water from the area. However, all these sup-
ply options include purification and the water has
to be pumped to the location. These are impor-
tant disadvantages of the technical solutions, for
they involve continuous expenditure and energy,
and are not sustainable.
If management objectives include both eu-
trophication abatement and restoration of the
original fen vegetation, restoration of the original
hydrology should be given serious consideration.
This necessitates structural investment to reduce
the negative water balance and the positive nu-
trient balance to negligible level. To realise these
goals, water deficits caused by extreme infiltration
have to be avoided. This can only be achieved by
inundation (partial or total) of the Polder
Horstermeer, together with the stimulation of the
discharge of groundwater from the hill ridge.
These options (f-h) will lead to a considerable
and sustainable improvement of the ecosystems
in the long-term. Because restoration of the orig-
inal hydrology does not completely meet the water
demand in summer, an additional supply option
has to be implemented, i.e. dephosphorized water
from the Vecht River according to this study.
Inundation of the Polder Horstermeer and
stimulation of the discharge of groundwater (f-h)
will meet the requirements of the ecosystems to a
much larger extent than the supply options (a-d)
do. The cost-benefit analysis indicates that by
inundation the financial investment will be high
the first 30 years. However, this problem is a
matter of time, for no further financial investment
will be needed after 30 years. In contrast, the
supply options would require continual funding
and therefore deserve lower priority than the sus-
tainable options. The essential choice in policy
can be reduced to an investment of little money
for hundreds of years to stimulate the ecosystems
for a minor part or an investment eight times
higher during thirty years to stimulate the ecosys-
tems sustainable to an optimal level.
257
References
Barendregt, A., J. T. De Smidt & M. J. Wassen, 1986. The
impact of groundwater flow on wetland communities. In:
Gehu, J. M. (ed.), Colloques Phytosociologiques XIII:
603-912. Kramer, Berlin.
Barendregt, A., M. J. Wassen & J. T. De Smidt, 1989. Re-
generation and conservation ofecosystems by integral water
management. In: L. Van Liere, R. M. M. Roijackers &
P. J. T. Verstraelen (eds), Integral water management in the
area with hill ridges and lakes in Het Gooi and Utrecht.
CHO-TNO 22: 214-228, 's Gravenhage (in Dutch).
Barendregt, A., M. J. Wassen & A. Van Leerdam, 1990.
Changes in hydrology and management adversely affect the
vegetation in the area of the river Vecht. Landschap 7:
17-32 (in Dutch).
Barendregt, A. & M. C. Bootsma, 1991. The hydro-ecological
model ICHORS (versions 3.1 en 3.2); the relations between
aquatic and fen plants and environmental factors in the
province of Utrecht. Department of Environmental Stud-
ies, University of Utrecht (in Dutch).
Best, E. P. H., D. De Vries & A. Reins, 1984. The macro-
phytes in the Loosdrecht Lakes: A story of their decline in
the course of eutrophication. Verh. int. Ver. Limnol. 22:
868-875.
Ministry of Agriculture, Nature Conservation and Fishery,
1990. Nature Management Plan. Den Haag (in Dutch).
Province of North-Holland, 1986. Provincial Grondwater
Management Plan; plan and explanation. Haarlem (in
Dutch).
Province of North-Holland, 1987. Wild plants in North-
Holland; an research by the Province of North-Holland.
Haarlem (in Dutch).
Province of North-Holland, 1990. Provincial Watermanage-
ment. Plan of North-Holland. Haarlem (in Dutch).
Province of Utrecht, 1990. Policy plan for Nature and Land-
scape in the province of Utrecht. Utrecht (in Dutch).
Roijackers, R. M. M. & P. J. T. Verstraelen, 1988. Ecological
investigations in three shallow lakes of different trophic
level in the Netherlands. Verh. int. Ver. Limnol. 23: 489-
495.
Roijackers, R. M. M. & P. J. T. Verstraelen, 1991. Eutroph-
ication abatement programme for three North Holland
Vecht Lakes. Verh. int. Ver. Limnol. 24: 768-772.
Schot, P. P., 1991. Solute transport by groundwater flow to
wetland ecosystems. Thesis, University of Utrecht.
Schouten, P. H., P. J. T. Verstraelen, J. J. Buntsma, 1988.
Restoration of the Vecht Lakes ecosystems surrounding the
Horstermeer polder: a feasibility study. Agric. Wat. Mgmt.
14: 471-479.
Starn, S. M. E. & A. Barendregt, 1990. Assessment of aquatic
and fen species by supply of water into the Lakes Stichts
Ankeveen and Lakes Kortenhoef. Department of Environ-
mental Studies, University of Utrecht (in Dutch).
Van den Berg, W. J.& J. T. De Smidt, 1985. The vegetation
of the Eastern Vecht area 1935-1980. Stichting Commissie
258
voor de Vecht en het Oostelijk en Westelijk Plassengebied.
Hilversum (in Dutch).
Verstraelen, P. J. T., R. M. M. Roijackers & A. Barendregt,
1989. Conditions for restoration in the area of the Lakes
Ankeveen and Kortenhoef. In: L. Van Liere, R. M. M. Ro-
ijackers, P. J. T. Verstraelen (eds), Integral water manage-
ment in the area with hill ridges and lakes in Het Gooi and
Utrecht. CHO-TNO 22: 214-228, 's Gravenhage (in
Dutch).
Verstraelen, P. J. T., J. Wisserhof, Lj. Rodil'; & R. Eijsink,
1992. Eutrophication control strategies for three shallow
Vecht lakes in the Province of North Holland. Hydrobio-
logia 233: 235-245.
Visser, C. M., P. J. T. Verstraelen & R. M. M. Rooijackers,
1989. Report of the lake sediments in the experimental
project of the Lakes Ankeveen and Kortenhoef: the dredg-
ing considered. Agricultural University, Wageningen (in
Dutch).
Wassen, M. J., 1990. Waterflow as a major landscape eco-
logical factor in fen development. Thesis, University of
Utrecht.
Williams, M., 1990. Wetlands, A Threatened Landscape. Basil
Blackwell, Oxford.
Witmer, M. C. H., 1989. Integral water management at re-
gional level. An environmental study of the Gooi and the
Vechtstreek. Thesis, University of Utrecht.
Hydrobiologia 233: 259-270, 1992.
L. Van Liere & R.D. Gulati (eds). Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 259
© 1992 Kluwer Academic Publishers.

Phosphorus eutrophication in the SW Frisian lake district

1. Monitoring and assessment of a dynamic mass balance model

Harry J. W. J. Van Huet*

Limnological Institute, Tjeukemeer Laboratory, De Akkers 47, 8536 VD Oosterzee, The Netherlands;
* Present address: University of Nijmegen, Department of Environmental Sciences, Toernooiveld 1, 6525
ED Nijmegen, The Netherlands

Key words: eutrophication, lakes, phosphorus, chloride, loading, dynamic mass balance model,
sensitivity analysis

Abstract

In 1984 a frequent monitoring programme was started in the hypertrophic S.W. Frisian lake district, with
emphasis on total phosphorus (TP) and chloride (Cl-). The main objectives of the project were: to
quantify the phosphorus flows, to gain insight in the process of eutrophication, and to simulate man-
agement scenarios. The seasonal variability in the lakes is due mainly to the man-made hydrology: re-
ception of humic-rich polder water in wet periods (winter) and inlet of chloride-rich IJsselmeer water
in dry periods (summer). The yearly means of TP concentrations in the lakes (Tjeukemeer, Groote
Brekken and Slotermeer) ranged from 0.23 to 0.29 mg 1- 1. However, much higher concentrations
(0.9 mg 1- 1) were found in periods with high inflow of polder water.
The simulations with a mass balance showed an acceptable similarity between measured and simu-
lated concentrations of TP as well as of Cl-. Chloride was modelled to verify the accuracy of a
hydrodynamic model. A sensitivity analysis of the apparent settling rate in the P model showed that
sensitivity was lowest in simulations of Groote Brekken and highest in simulations of Slotermeer, the
difference being attributable to the influence of the water residence time. The model was found to be
appropriate for simulating management scenarios.

Introduction these investigations will be published elsewhere

In 1984 a study was started to investigate the
effects of phosphorus (P)-related eutrophication
of S.W. Frisian lakes (Fig. 1). The main objective
of the study was to simulate scenarios for policy
and management, by means of a P model. In
order to support this model, the distribution of P
in the lake sediments and the P loads of the lakes
from the surrounding polders were investigated.
The water transport in the lake area was simu-
lated by use of a hydrodynamic model. Details of
(Van Huet & De Haan, in press; Van Huet
(1991b), and Van Huet (in press), The prelimi-
nary results of the project are reported in Van
Huet (1990).
This paper focuses on: a) a brief description of
the total phosphorus (TP) concentration and
some other water quality parameters of the lakes
during 1984-1987; b) the assessment of a simple
dynamic mass balance model; and c) the results
of this model, with emphasis on TP concentra-
tions.
260

A
WADDENZEE

ECHTENER VEENPOLDER

2.5

Fig. 1. A. The Frisian 'boezem'. B. Its position in The Netherlands. C. The S.W. lake district (e-:boundaries research area;
RWZI: waste water treatment plants; RW50: Motorway 50; from Van Huet, 1990).
 261

Area description and seasonal variation
The S.W. Frisian lakes and the interconnecting
canals (water surface area: 4000 ha) are part of
the 'boezem', the water network (14000 ha) in the
province of Friesland, used for water table regu-
lation in the surrounding polders and for flushing.
Table 1 shows the depth, surface area, catchment
area and water residence times of the shallow
lakes Tjeukemeer, Groote Brekken and Sloter-
meer, the lakes modelled here.
In summer (April-September), under typical
climatological conditions, a precipitation deficit
in the polders occurs and boezem water is let in.
To replace the loss, chloride-rich water from IJ s-
selmeer, originating from the river Rhine, is let in
at Lemmer, Tacozijl and Stavoren. In winter
(October-March), precipitation exceeds evapo-
ration and humic-rich water from the polders is
pumped into the boezem system, while, in turn,
boezem water is pumped out into IJ sselmeer at
Lemmer and Stavoren or is released into the
Waddenzee (Fig. 1). Thus, man strongly influ-
ences the hydrology of the lake system. In addi-
tion, the wind has considerable influence on
short-term water transport. For detailed informa-
tion about the boezem system and hydrology, see
Leenen (1982a) and Van Huet (1990).
The main differences between the three lakes
are given in Table 1. The influence of polder water,
reflected in the TP concentrations, in Tjeukemeer
in winter is high and relatively low in Slotermeer.
The influence of water from IJsselmeer in sum-
mer, reflected in the Cl- concentrations, is great
in Groote Brekken and less in Tjeukemeer and in
Slotermeer (see also Table 2). The mean water
residence time is short in Groote Brekken and
long in Slotermeer. Tjeukemeer is a typical polder
lake, while Slotermeer is a relatively isolated lake.
Water residence time and influence of water from
polders and IJ sselmeer on Koevorder Meer, the
fourth main lake in the area, are similar to those
of Groote Brekken. Despite a difference in bot-
tom composition of Tjeukemeer and Slotermeer,
sedimentary TP contents do not differ much (Van
Huet & De Haan, in press).
Eutrophication and use of models
Frisian lakes are highly eutrophicated. The con-
centrations of TP and TN (total nitrogen) gener-
ally exceed 0.2 and 2.0 mg 1- 1, respectively, while
summer chlorophyll-a contents often exceed
150 /lg 1- 1. The lake ecosystems are dominated
by the cyanobacteria, mainly Oscil/atoria agardhii
(Moed & Hoogveld, 1982; De Haan & Moed,
1984; Claassen, 1986). The main zooplankton
species are: Bosmina coregoni, Chydorus sphaericus
and Daphnia hyalina. The main fish species are:
bream (Abramis brama), and pike perch (Stizo-
stedion lucioperca), see Lammens (1986). Eutroph-
ication, chemistry, fish communities and trophic
relationships have been described by Beattie et al.
(1978), De Haan (1982), De Haan & Voerman
(1988), Vijverberg & Van Densen (1984), Lam-
mens (1990), Klink & Claassen (1988) and Moed
et al. (1988).

Table I. Characteristics of Tjeukemeer, Groote Brekken and Siotermeer (*: after Brinkman et al., 1989; **; Van Huet, in press).

Tjeukemeer Groote Brekken Slotermeer

Surface water area (ha)** 1958 300 1100

Catchment area (ha) 50000 5000 3800
Mean depth (m)** 1.75 2.0 1.65
Bottom structure (> 50%, < 50%) Peat, sand Sand, peat Sand, peat
Mean water residence time (days)** 20 5 54
Relative influence of polder water (winter) High Moderate Low
Relative influence of IJsseimeer water (summer) Moderate High Moderate
Number of lake chloride sampling stations 8 3 6
Number of lake TP sampling stations (mixed sample) 3 3 3
262

Table 2. Water quality parameters in Tjeukemeer (Tj), Groote Brekken (GB) and Siotermeer (SI) during 1984-1987 (*: sampling
started on May 14, 1984; **: mean value of mixed samples; ***: mean value of all lake stations).

TP (mg 1- ')** CI (mg 1- ')*** Secchi (cm) pH

Tj GB SI Tj GB Sl Tj GB Sl Tj GB SI

1984* 0.25 0.23 0.24 98 177 130 28 33 28 9.0 8.8 8.8

1985 0.29 0.24 0.24 56 131 101 37 35 27 8.5 8.6 8.6
1986 0.28 0.24 0.26 105 153 120 32 32 29 8.7 8.8 8.6
1987 0.29 0.24 0.23 59 113 96 30 30 25 8.6 8.7 8.7

Apr. 84-Sep. 84* 0.24 0.23 0.25 102 206 122 26 34 27 9.5 9.2 9.0
Apr. 85-Sep. 85 0.30 0.26 0.26 61 146 102 29 33 26 8.8 8.8 8.7
Apr. 86-Sep. 86 0.25 0.22 0.25 99 155 105 34 31 28 9.0 9.1 8.8
Apr. 87-Sep. 87 0.27 0.23 0.22 64 131 96 29 29 25 8.9 8.9 9.0

Oct. 84-Mar. 85 0.27 0.25 0.23 81 125 130 34 32 29 8.1 8.3 8.6
Oct. 85-Mar. 86 0.32 0.24 0.24 45 101 97 47 35 26 7.9 8.2 8.4
Oct. 86-Mar. 87 0.32 0.26 0.27 113 154 145 30 33 30 8.3 8.5 8.3

In Dutch lakes, including the S.W. Frisian
lakes, a significant direct correlation between TP
concentrations and chlorophyll-a contents has
been found (Lijklema et a/., 1988). There are in-
dications that undesirable cyanobacteria will no
longer dominate if TP concentrations can be re-
duced to < 0.07 mg 1- 1. Here, I assume that re-
ducing P levels in the Frisian lakes will eventually
lead to its limitation and, finally, to lake restora-
tion.
Golterman et al. (1980) used a mass balance
model, based on chloride, to quantify the water
balance of Tjeukemeer and the P load to this lake
during 1969-1975. They concluded that for reli-
able model results Cl - concentrations have to be
measured frequently (weekly or fortnightly) at
several stations on the lakes and in their inflow
and outflow canals. Frequent sampling is neces-
sary because of the spatial and temporal variabil-
ity of Cl- and TP concentrations, due to inflow
of polder water (lnd IJ sselmeer water, see also
Leenen (1982a), De Haan & Moed (1984).
Cl- was measured because time series of Cl-
concentrations can lead to understanding of water
transport, and the origin of water in the lakes (due
to concentration differences between polder and
IJ sselmeer water). The Cl- concentrations were
also used to verify the results of a hydrodynamic
model (see the section on assessment of the Cl-
model). TP concentrations were measured for the
calibration of the mass balance model and to pro-
vide an indication of trophic levels. For the lat-
ter purpose Secchi disc transparencies and pH
values were also measured.
A composite weekly sample of the lake stations
was analyzed for TP concentrations, while fort-
nightly samples of the inflow and outflow sites of
the lakes were analyzed for TP. Phosphorus was
analyzed colorimetrically with an autoanalyser
(molybdate reagent, Murphey & Riley, 1962), fol-
lowing hydrolysis in acedic persulphate. Cl- was
titrated with Ag + -ions. For a description of the
polder sampling programme and results, see Van
Huet, 1991 b.

Materials and methods Water quality during 1984-1987

From May 1984 to December 1987, 50 stations Information about the water quality parameters
in the S.W. lake district were sampled weekly in the three lakes during 1984-1987 is given in
with a plexiglass tube (length 1 m) from a boat. Table 2 (see also Van Huet, 1990). It can be con-

cluded that: (a) TP concentrations in polder water
were higher than in IJsselmeer water (e.g. the
mean values in 1985 were > 0.4 mg 1- 1 and
0.2 mg 1- 1, respectively); (b) during dry periods
Cl- concentrations increase, due to inlet of IJ s-
selmeer water, and in wet periods, high TP con-
centrations occur, due to the pumping of polder
water into the system; (c) Tjeukemeer is influ-
enced most by polder water, and Groote Brekken
and to a lesser extent Slotermeer by IJsselmeer
water; (d) concentrations in Tjeukemeer fluctu-
ated more than in Slotermeer, while concentra-
tions between Tjeukemeer west and Tjeukemeer
east differed; and (e) Secchi disc transparencies
were low and pH values were high in summer.
Modelling phosphorus dynamics
Model choice
Generally, two types of lake eutrophication mod-
els are used: empirical, steady state models and
theoretical, dynamic models. The empirical mod-
els are derived from statistical analysis of data
from large numbers of lakes. In general, these
models give good predictions for a group of lakes,
but might lead to large errors if applied to indi-
vidual lakes. Simple empirical models have ad-
vantages over complex theoretical models. Few
data and simple mathematics are needed and they
show generality (Kamp-Nielsen, 1985). However,
their limitations are the steady state assumptions
and the poor resolution in time and space. The
theoretical models are based on (detailed) math-
ematical descriptions of the nutrient and popula-
tion dynamics. Hence, these dynamic models can
differ in complexity, from simple input-output
models using a single net loss term to large eco-
logical n-layer, n-box, n-compartment models,
describing spatial variability of several variables;
moreover they do not have the limitations of the
empirical models, but may have disadvantages in
their complexity and use of extensive data sets.
Furthermore, they may not be fully adequate, be-
cause of lack of knowledge about processes, e.g.
sedimentation, resuspension and P-release from
263
sediments. For a review of model types, their
steady state solutions, advantages, disadvantages
and results see Golterman (1980), Jorgensen
(1983), Reckhow & Chapra (1983a, 1983b), Van
Straten (1983), Kamp-Nielsen (1985), Ahlgren
et al. (1988) and Prairie (1989).
The model choice depends on several criteria
such as model use, complexity of the system (e.g.
occurrence of stratification and the hydrodynamic
character of the aquatic ecosystem), availability
of data sets, etc. Empirical models are often ap-
plied if only mean annually values are of interest,
which is not the case in the Frisian situation.
Moreover, the steady state approach is not real-
istic for modelling the P levels in the Frisian lakes
because of the seasonal patterns of loading.
Therefore a dynamic model was preferred.
An important criterion is the use of the model.
Will the model be used for research or for man-
agement purposes (or possibly both)? Models for
research purposes may have objectives including
identification and hypotheses testing. Manage-
ment models generally focus more on the predic-
tion of one or two variables. Therefore, on aver-
age, research models are more complex than the
(practical) management models. Modelling the P
dynamics in the Frisian lakes is important for
management reasons. Hence I used a practical
management model approach.
Because of the short residence times and the
influence of the wind on the water transport in the
lakes (Van Huet, a, in press), it is assumed that
the lakes are ideally mixed. Thus, the CSTR-
approach (continuously stirred tank reactor) was
used. Another reason for choosing a simple dy-
namic model is that the uncertainty in the data
does not support detailed modelling (see Fig. 3,
and the section 'Representativeness ofphosphorus
concentrations').
The simple dynamic management model IS
based on the principle of mass conservation:
d(VP)
-- = Input - Ouput - Net loss, (1)
dt
,where V = the lake volume; P = the lake TP con-
centration; t = the time.
264
Net loss occurs due to sedimentation, resus-
pension and release of P and is influenced by
several factors, such as the water residence time,
the P loading, lake morphology, oxygen concen-
tration, the pH at the sediment-water interface
and wind-induced turbulence. It is very difficult to
describe net loss (Kamp-Nielson, 1985; Bolin
et al., 1987; Ahlgren et al., 1988 and Herman
et al., 1989). At this point I will not go into the
several aspects of net loss and will use the ap-
parent settling approach, since it is a physically
realistic description (Reckhow & Chapra, 1983a).
Net loss is assumed to be a first order process,
that is, the apparent settling rate is directly pro-
portional to the TP concentration in the lake.
Model description
For each lake (Tjeukemeer, Groote Brekken and
Slotermeer) the following basic equations were
used:
d (VP)
--=1:Q, ,p, ,
dt m.) m.)
d (VP) = V dP + P dV (3)
dt dt dt '
d (V)
~ = 1:Qin.} - 1:Qout,} - Qe' (4)
where, V = lake volume (m 3 ); P = lake TP con-
centration (mg 1- }); Pin.} = TP concentration in
influent j (mg 1- }); Qin,} = inflow discharge j
(m 3 d - 1); Qout,} = outflow discharge j, except
Qe = evaporation (m 3 d - }); Vs = apparent settling
rate (m 3 d - 1); A = lake surface area (m 2). Qe was
described separately because in this way the sim-
ulations led to a realistic increase of lake TP con-
centrations if only evaporation occurs.
The discharges Qin,} and Qout,} are defined to
be positive. For example, if there is water trans-
port through Follegasloot (Fig. 1) from Groote
Brekken in the direction ofTjeukemeer (discharge
QFo}), then for Tjeukemeer Qin. Fo} = QFol and
Qout, Fo} = 0 is valid, while for Groote Brekken
Qin. Fal = 0 and Qaut. Fal = QFa} is valid. This ex-
ample also shows that the TP concentrations in
one lake can be influenced by the concentrations
in the two other lakes.
From equations (2), (3) and (4), the rate of
change of the P mass can be expressed as:
d (P)
V -- = 1: {Qin )' (Pin )' - P)}
dt ,.
+ QeP- vsAP. (5)
Model inputs
The P dynamics in the three lakes could be mod-
elled, because the data on inflow and outflow
discharges were available from a hydrodynamic
model during three periods (August 8, 1985-
January 31, 1986; April I-December 31, 1986;
March I-December 20, 1987). The periods are
separated because ice cover during February 1-
March 31,1986 and January I-February 28,1987
led to unreliable data. Altogether, the hydrody-
namic model and consequently the mass balance
model covered about 25 months. Table 3 shows
the model sources (1:Qin,) and sinks (1:Qaut,)
and concentrations in sources. For a description
of the hydrodynamic model and contributions to
the water balance, see Van Huet (in press). Mostly
daily input data were available, while for not daily
available inputs and concentrations calculations
were made with interpolated values. For the con-
centrations in sinks the model concentrations
were used, according to the model definition. The
model calculations started each simulation period
with the measured concentrations in the lakes
(initial conditions).
Modelling results and discussion
Assessment of a chloride model
Prior to the P mass balance model, simulations
were made with a similar 0- model, for verify-
ing if the daily mean discharges, based on the
 265

Table 3. Model sources and sinks and concentrations in sources. (D = daily available; B = biweekly available; T = available every
decade; DPWE = Department of Public Works and Environment of the Province of Friesland; RNMI = Royal Netherlands
Meteorological Institute).

QLi = inlet of Lake IJ sselmeer water near Lemmer;

QT = inlet of Lake IJ sselmeer water near Tacozijl;
Qip = polder water pumped into the system;
Qp = precipitation;
Qw = contribution from waste water treatment plant near Lemmer;
Qs = inlet of Lake IJ sselmeerwater via sluices near Lemmer;
QLo = outlet of boezem water (punped into Lake IJ sselmeer);
Qop = outlet of boezem water into the surrounding polders;
Qe = evaporation;
Qi = infiltration;
Qc = inflow and outflow through the system boundary canals Broeresloot, Pier Christiaansloot, Scharsterrijn, Koevorder Meer
and the canals north of Woudsend.

Source/sink Range Comment Concentr. (ranges) Comment

106 m 3 'mo- 1

TP mg I-I Cl- mg 1-1

Qc (D) - 56.7-67.6 Hydrodynamic model 0.06-0.92 28-250 Measured

QLi (D) 0-36.0 Obtained from DPWE 0.08-0.38 57-250 Measured
QT (D) 0-20.4 Obtained from DPWE 0.08-0.38 57-250 Measured
Qs (D) 0-0.2 Obtained from DPWE 0.08-0.38 57-250 Measured
Qw (B) 0.1-0.4 Obtained from DPWE 0.6-4.0 150 Obtained (DPWE)
Qp (D) 0.9-5.9 Obtained from RNMI (station Lem- 0.04 10 Measured at station Span-
mer, Heeg and Joume) nenburg (Fig. 1)
Qip (D) 1.7-16.1 * 0.06-1.1 150 **
QLo(D) - 62.4-0 Obtained from DPWE Model Model
Qo p (D) - 5.7-0 It was assumed that inflow occurred Model Model
only during April-September and
that inflow was directly proportional
to the precipitation deficit
Qi (D) -0.3 Estimated (Broers, 1987) Model Model
Qe (T) -0.1-4.9 Obtained from RNMI (stations
Lelystad and Leeuwarden, Fig. 1)

* Obtained from Water Boards. Water quantities pumped from the Echtener Veenpolder into Tjeukemeer. However, only monthly
discharges were available, from polders surrounding Groote Brekken and Siotermeer. On average, these discharges agreed well
with the monthly discharges of the Echtener Veenpolder. Therefore, for Groote Brekken and Siotermeer the daily Echtener
Veenpolder discharges were used (for Groote Brekken estimated at 25% of the Tjeukemeer values and for Siotermeer at 50%).
** For Tjeukemeer, the measured TP concentrations in polder water of the Echtener Veenpolder were used for all three simu-
lated periods. For Groote Brekken and Slotermeer the mean concentrations of the polders surrounding these lakes. However,
these polders were not sampled during the first simulated period. Therefore, for this period and for these lakes the Echtener
Veenpolder values were used.

hydrodynamic model, were accurate enough for

model input. CI- was modelled as a conservative
material: no net loss was assumed. Therefore, the
CI- model corresponds to the P model with an
apparent settling rate Vs = O. Thus the CI- model in which C/- = C/- concentration in the lake
equation is: (mg 1- 1) and q;;,j = in the influent j (mg 1- 1).
There was a fair similarity between the simu-
266

lated and observed CI- concentrations in the
lakes during the three periods (Fig. 2). The
boundary discharges were about correct or at least
sufficiently accurate for the purpose of model
input. Hence, these discharge values were also
used as forcing functions for the P model. The
omission of a dispersion term in equations (5) and
(6) is discussed by Van Huet (l991a). The phys-
A Tjeukemeer
250 r---------------------------,
ical dispersion (dispersion due to gradients in the
water velocity) in the Follegasloot, as well as the
so-called equivalent dispersion (due to the fluc-
tuating flow between Tjeukemeer and Groote
Brekken) in this canal were estimated and com-
pared with the advective water transport. It was
concluded that the dispersion terms can be ne-
glected.

200

150

100

50 ... ..".,..
-.
. ~
_.~­
v~

B
oL--------------'------------'------'
AUG 19B5 B AUG 19B6 B AUG 19B7

• ass (FF) SIM

B Groote Brekken
250 r---------------------------,
..
200

150 .... ....... .

. _._-
100

50 .. ~-.- .

0
B AUG 19B5 B AUG 19B6 B AUG 19B7

• ass (FF) SIM

C Siotermeer
250

200

150
---.
~
0 100

50
--- . .- ~ ~

0
B AUG 19B5 B AUG 19B6 B AUG 19B7

• ass (FF) SIM

Fig. 2. Observed and simulated Cl- concentrations in the three Frisian lakes during three periods.
 267

Representativeness of phosphorus concentrations samples (. FF: mixing water samples of 3 lake

To examine the reliability of the data sets, two
available different data sets of lakes were com-
pared. The TP concentrations in Tjeukemeer dur-
ing the three simulated periods differed (Fig. 3A),
probably due to differences in (a) mixing of the
stations as described; + LI: mixing water sam-
ples of 10 stations (De Haan, 1982) and to (b) the
storage, handling and analyses in different labo-
ratories. The TP concentrations in Groote Brek-
ken and Slotermeer (Figs 3, Band C) differed
only in the mixing programme (. FF: a mixed

A Tjeukemeer
1.00 .
0.80

.
. . .
.

hi
§
0.60

OAO
.
- .-. . -_.
~+.- . .. ~
..·. . -.-:.;...
. .-.
0..

.-. ..... . .. .-.... .....

0.20
.'. . . . . :
. . : . + .......

.
: ...

0.00
8 AUG 1985 8 AUG 1986 8 AUG 1987

-+- ass (LI) • ass (FF) -- SIM

B Groote Brekken
1.00

0.80 ·
§ 0.60

fw1
.

. .. . ~
0.. OAO

0.20

0.00
. .
.. ~ .--. .
...... . ...
. .. .....
...
·
'.
.
..
--....
.-
--. -.- ,

8 AUG 1985 8 AUG 1986 8 AUG 1987

-+- ass (PF)

• ass (FF) -- SIM

C Siotermeer
1.00 r---------------------------

0.80

0.60

. .
0.. OAO

~.
'.
.

.. - ..... - . _. ~.' .........'- ....:.-

~
.

0.20 ._..
.. +~.

0.00 ' - - - - - - - - - - - - - - " - - - - - - - - - - - - ' - - - - - -

8 AUG 1985 8 AUG 1986 8 AUG 1987

ass (PF) • ass (FF) -- SIM

Fig. 3. Observed and simulated TP concentrations in the three Frisian lakes during three periods.
268
sample of 3 stations; + PF: a sample of a single
station). Although the sampling dates also dif-
fered, the figures show that different sampling
procedure might lead to a different interpreta-
tion of TP levels in a lake. This is important in
calibrating the model: the mean deviation be-
tween model and data cannot be expected to be
lower than the mean difference between the data
sets.
Calibration of the phosphorus model
The most important parameter for calibration
of the P model is Vs for each lake. With an ap-
parent settling rate Vs = 0.015 m.d -1, the simu-
lated TP concentrations in the three lakes (Fig. 3)
agreed rather well with the measured concentra-
tions. The model vs-sensitivity was highest for
the TP concentrations in Slotermeer, moderate
in Tjeukemeer and lowest in Groote Brekken
(Fig. 4). This result can be explained by analyz-
ing the steady state solution of equation (5) (dPj
dt= 0):
1 fication and only a mathematical tool (Leenen,
(7)
where Qe~:EQin, H=lake depth (= VjA) and
T = water residence time (= Vj:EQin). The water
residence times in Groote Brekken are short (Ta-
ble 1) and consequently T.VsIH ~ 1. The TP con-
centrations in this lake will be dominated by the
concentrations in inflowing water (Pin), but it will
be less so for the other two lakes because oflonger
water residence times.
Van Straten (1989) used a similar model for a
long-term simulation (12 years) of the TP con-
centrations in the S.W. Frisian lakes, with decade
input discharges from a stationary flow model
and different TP data sets. He also used different
Vs values for each lake, which were computed by
a parameter estimation procedure. Use of other Vs
values or different values for each lake do not lead
to a substantial improvement in model results
(Fig. 4), because the curves are hardly discernible
and the range is small as compared with the data
range (compare Figs 3 and 4). Also, changing the
vs-value for one lake does not greatly influence the
TP concentrations in the other two lakes.
By use of the P mass balance model the TP
balances of the three lakes during three periods
could be quantified, (see Van Huet, 1992).
Discussion and conclusions
The apparent settling approach is a simplifica-
tion, because P removal in lakes will in fact be the
end result of opposing processes such as sedi-
mentation, resuspension by wind and release.
These processes depend, in their turn upon the
fraction particulate P and thus upon algal mass,
which all vary seasonally. In general, there is lack
of information about these processes and more
research is needed to justify a detailed model.
Also, the complete mixing approach is a simpli-
1982b). An n-segment approach, however, was
not possible because of lack of spatial variation
in the P data sets. Detailed information on inputs,
e.g. polder water concentrations and discharges,
is necessary for a reliable estimation of the P-
load, but is less important for the calculation of
lake TP concentrations, because detailed input
will be smoothed due to the complete mixing ap-
proach.
Altogether, the simulations show that a simple
model approach leads to acceptable results. There
is no need for a differentiation in net loss veloc-
ities between lakes. The high frequency fluctua-
tions in the data as compared with the model are
partly due to measured uncertainty as shown by
comparison of data collected by different insti-
tutes. If real, they could only be captured by a
much more detailed model, but this is unneces-
sary from a management point of view. The model
was used to simulate management scenarios (see
Van Huet, 1992).
 269

A Tjeukemeer
0.60

OAS

0.36

Q. 0.24

0.12

0.00 L .l.- -"- ---'

8 AUG 1985 S AUG 1986 S AUG 19S7

--- 1 II III IV

B Groote Brekken
0.60 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,

OAS

0.36

Q. 0.24

0.12

0.00 L.... --I. ~ ~

B AUG 19a5 8 AUG 19S6 S AUG 19S7

--I II III IV

C Siotermeer
0.60 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,

OAS

0.36

Q.

0.00 L.... --I. ~ ---'

8 AUG 1985 S AUG 19S6 S AUG 19S7

--- 1 - - II III IV

Fig. 4. Simulated P concentrations during three periods. I: \'s = 0.015 m d - I for all lakes (basic line, see also Fig. 3);
II: I's = 0.03 m d - I for all lakes (bottom line); III: V = 0.0075 m d - 1 for all lakes (top line); IV: from Van Straten (1989):
S

I's = 0.Ql5 m d - I (Tjeukemeer), Vs = 0.063 m d - I (Groote Brekken), V

s = 0.0075 m d - 1 (Slotermeer).

Acknowledgements De Haan for critically reading the manuscript. I thank Steven

This study was supported by a grant from the Dutch Minis-
try of Housing, Planning and Environment and the Depart-
ment of Public Works and Environment of the Province of
Friesland. I would like to thank Gerrit Van Straten and Henk
Visser, Diederik Visscher, Aukje Stoeten and Koos Swart for
technical assistance, Lammie Van de Lende and Hans Hainje
for the chemical analyses and Meta Donk for the improve-
ment of the English text.
270
References
Ahlgren, I., T Frisk & L. Kamp-Nielsen, 1988. Empirical and
theoretical models of phosphorus loading, retention and
concentration vs. lake trophic state. Hydrobiologia 170:
285-303.
Beattie, D. M., H. L. Golterman & J. Vijverberg, 1978. An
introduction to the limnology of the Friesian lakes. Hydro-
biologia 58: 49-64.
Bolin, S. B., T J. Ward & R. A. Cole, 1987. Phosphorus
models applied to New Mexico reservoirs. J. Wat. Res.
Planning Mngt. 113/3: 323-335.
Broers, H. P., 1987. Groundwater flow systems in southern
Friesland (in Dutch). Province of Friesland. Limnological
Institute and Free University of Amsterdam: 86 pp.
Brinkman, A. G., H. J. W. J. Van Huet & G. Van Straten,
1989. Eutrophication research in Friesland (in Dutch).
Twente University and Limnological Institute: 55 pp.
Claassen, T. H. L., 1986. Eutrophication and algal growth in
the Frisian surface water (in Dutch). H 20 12: 268-275 and
279.
De Haan, H., 1982. Physico-chemical environment in Tjeu-
kemeer with special reference to speciation of algal nutri-
ents. Hydrobiologia 95: 205-221.
De Haan, H. & J. R. Moed, 1984. Phosphorus, nitrogen and
chlorophyll-a concentrations in a typical Dutch polder lake,
Tjeukemeer, in relation to its water regime between 1968
and 1982. Wat. Sci. Tech. 17: 733-743.
De Haan, H. & J. Voerman, 1988. The chemistry of a typi-
cal Dutch reservoir, the Tjeukemeer, in relation to its water
management between 1970 and 1986. Freshw. BioI. 20:
1-10.
Golterman, H. L., 1980. Phosphate models, a gap to bridge.
Hydrobiologia 72: 61-71.
Golterman, H. L., J. Voerman & H. W. De Nie, 1980. Phos-
phorus loads to Tjeukemeer (in Dutch). H 2 0 6: 116-
121.
Herman, G., A. Nishri & T Berman, 1989. A long-term pre-
diction model for total phosphorus concentrations in Lake
Kinneret. Wat. Res. 23, 1: 61-66.
Jorgensen, S. E., 1983. Ecological modeling oflakes. In: G. T
Orlob (ed.). Mathematical modelling of water quality.
Streams, lakes and reservoirs. John Wiley: 337-394.
Kamp-Nielsen, L., 1985. Modelling of eutrophication pro-
cesses. In: R. Vismara, R. Marforio, V. Mezzanotte & S.
Cernuschi (eds), Proceedings congress: Lake pollution and
recovery: 57-68.
Klink, F. J. & Claassen, T H. L., 1988. Eutrophication and
the Frisian Lakes (in Dutch). H 2 0 17: 490-493.
Lammens, E. H. R. R., 1986. Interactions between fishes and
the structure of fish communities in Dutch shallow, eutro-
phic lakes. Ph.D. Thesis. Agricultural University of Wa-
geningen: 100 pp.
Lammens, E. H. R. R., 1990. The relation of biotic and abi-
otic interactions to eutrophication in Tjeukemeer, The
Netherlands. Hydrobiologia 191: 29-37.
Leenen, J. D., 1982a. Hydrology of Tjeukemeer. Hydrobio-
logia 95: 199-203.
Leenen, J. D., 1982b. Wind induced diffusion in a shallow
lake, a case study. Hydrobiol. Bull. 16: 231-240.
Lijklema, L. & G. Van Straten, 1975. Static and dynamic
simulation of water transport in a complex network system.
In: G. C. Vansteenkiste (ed.), Computer simulation of water
resources systems. North Holland Publications Comp.,
Amsterdam, The Netherlands: 661-670.
Lijklema, L., J. H. Janse, R. M. M. Roijackers & M. L.
Meijer, 1988. Eutrophication in The Netherlands (in
Dutch). H 2 0 17: 462-467.
Moed, J. R. & H. L. Hoogveld, 1982. The algal periodicity in
Tjeukemeer during 1968-1978. Hydrobiologia 95: 223-234.
Moed, J. R., H. L. Hoogve1d & H. De Haan, 1988. A study
of factors regulating the succession of cyanobacteria in Lake
Tjeukemeer, The Netherlands. Verh. int. Ver. Limnol. 23:
1894-1987.
Murphy, J. & J. P. Riley, 1962. A Modified single-solution
method for the determination of phosphate in natural wa-
ters. Analyt. Chim. Acta 27: 31-36.
Prairie, Y. T, 1989. Statistical models for the estimation of
net phosphorus sedimentation in lakes. Aquat. Sci. 51/3:
192-210.
Reckhow, K. H. & S. C. Chapra, 1983a. Engineering
approaches for lake management. Butterworth Publishers.
Reckhow, K. H. & S. C. Chapra, 1983b. The need for sim-
ple approaches for the estimation of lake model prediction
uncertainty. In: M. B. Beck & G. Van Straten (eds), Un-
certainty and forecasting of water quality. Springer Verlag:
293-303.
Van Huet, H. J. W. J., 1990. Phosphorus eutrophication re-
search in the lake district of south western Friesland, The
Netherlands. Preliminary results of abiotic studies. Hydro-
biologia 191: 175-185.
Van Huet, H. J. W. J., 1991a. Modelling water transport and
phosphorus eutrophication in an interconnected lake sys-
tem. A scenario study. Ph.D. Thesis. Agricultural Univer-
sity Wageningen.
Van Huet, H. J. W. J., 1991b. Phosphorus loads from peaty
polders in the S.W. Frisian lake district, The Netherlands.
Wat. Air Soil Pollut. 55: 321-335.
Van Huet, H. J. W. J., in press. Application of a wind-driven
hydrodynamic model for the quantification of transport in
an open boundary canal-lakes system in S.W. Friesland,
The Netherlands. Aquatic Sciences.
Van Huet, H. J. W. J., 1992. Phosphorus eutrophication in
the SW Frisian lake district. 2. Phosphorus balances and
simulation of reduction scenarios. Hydrobiologia 233: 271-
281.
Van Huet, H. J. W. J. & De Haan, in press. Horizontal and
vertical distribution of phosphorus in sediments of inter-
connected eutrophic polder lakes in S.W. Friesland, The
Netherlands. Archiv fUr Hydrobiologie.
Van Straten, G., 1983. Lake eutrophication models. In: L.
Somly6dy, S. Herodek & J. Fischer (eds), Eutrophication
of shallow lakes: modelling and management. Proceedings
IIASA-workshop: 175-206.
Van Straten, G., 1989. Frifos, a simple model for the calcu-
lation of total phosphorus contents in S.W. Friesland, in
the case of different management options (in Dutch).
Twente University: 120 pp.
Vijverberg, H. & W. L. T Van Densen, 1984. The role of the
fish in the food web of Tjeukemeer. Verh. int. Ver. Limnol.
22: 891-896.
Hydrobiologia 233: 271-281, 1992.
L. Van Liere & R.D. Gulati (eds), Restoration and Recovery of Shallow Eutrophic Lake Ecosystems in The Netherlands. 271
© 1992 Kluwer Academic Publishers.

Phosphorus eutrophication in the SW Frisian lake district

2. Phosphorus balances and simulation of reduction scenarios

Harry J. W. J. Van Huet*

Limnological Institute, Tjeukemeer Laboratory, De Akkers 47, 8536 VD Oosterzee, The Netherlands;
* Present address: University of Nijmegen, Department of Environmental Sciences, Toernooiveld 1, 6525
ED Nijmegen, The Netherlands

Key words: eutrophication, phosphorus balance, modelling, polders, scenarios, static gain, restoration

Abstract

The lakes and interconnecting canals in the S.W. Frisian lake district are highly eutrophic. In the
mid-1980's a project on eutrophication and lake restoration research was started. This project was aimed
at modelling water transport and phosphorus (P) dynamics and at simulating management scenarios.
A simple dynamic P-balance model was used to calculate total phosphorus (TP) balances and to sim-
ulate three TP reduction scenarios in each of three lakes: Tjeukemeer, Groote Brekken and Slotermeer.
The model covered three periods in 1985, 1986 and 1987. The external loads to Tjeukemeer were highest,
moderate to Groote Brekken, and lowest to Slotermeer. The major P sources in the area were discharges
from the surrounding polders, used mainly for agriculture, and from IJsselmeer.
Despite a 75 % TP-reduction in water from the surrounding polders the 0.07 mg 1- I target level could
be reached only occasionally in Tjeukemeer, while in the other two lakes this level was not even
approached. The effect of a 75% reduction in water from IJsselmeer was greatest in Groote Brekken
(but again approached the target only occasionally), moderate in Tjeukemeer and least in Slotermeer.
The simulations showed that only a 75% reduction in both external loads (from polders and from
IJ sselmeer) will lead to achieving the target level in Tjeukemeer and Groote Brekken during the sum-
mer periods. In Slotermeer, a relatively isolated lake, other measures are necessary to reach the target
level. The results are confirmed by an approximate theoretical analysis of the effects of load reduction.

Introduction Densen et aI., 1986; Andersson, 1988; Van Donk

Restoration of eutrophic aquatic ecosystems
mainly involves reducing nutrient concentrations
and algal biomass. Measures used or proposed
include dredging (Ryding, 1982; Bjork, 1985;
Andersson, 1988), chemical treatment (Foy,
1985; Foy & Fitzsimons, 1987; Balmer & Hult-
man, 1988), control of discharges and reducing
point sources (Ahlgren, 1988; Cullen & Fors-
berg, 1988), biomanipulation (Richter, 1986; Van
et al., 1989), artificial mixing and input of oxygen
(Imboden, 1985), sewage effluent diversion (Ed-
mondson, 1985), isolation (Burden et al., 1987;
De Haan et al., 1988), aeration (Bjork, 1985) and
flushing (Hosper, 1984). By the application of
these techniques, deterioration might be success-
fully halted. Phosphorus (P) is often considered
as the critical nutrient in determining the degree
of lake eutrophy and therefore, many restoration
programmes concentrate on P control.
272
By the mid 1970's the problem of eutrophica-
tion became evident in the S. W. Frisian lake dis-
trict. Reducing P concentrations was the main
objective to combat the eutrophication problem.
Therefore, in 1979 dephosphorization of waste
water in the area started, but this measure alone
did not lead to the desired result. In 1984 a re-
search project started that focused on a more
integrated approach. Investigation of the role of
sediments and P-Ioads from surrounding peaty
polders were part of this project, while modelling
the water transport and P dynamics was seen as
tool for achieving a better understanding of the
eutrophication process. The sedimentary P dis-
tribution, polder P-Ioads, a water transport model
and a P dynamics model, were previously de-
scribed by Van Huet and De Haan (in press), Van
Huet (1991), and Van Huet (in press; 1992), re-
spectively. In this paper this set of tools is used
to assess effects of various options for P-Ioad
control.
Materials and methods
Area description
Figure 1 shows the three main lakes (Tjeukemeer,
Slotermeer, and Groote Brekken, depth 1-2 m),
canals (depth 1-4 m, width 20-50 m) and system
boundaries of the S.W. Frisian lake district. This
area is part of the 'boezem': the canal-lake system
in the province of Friesland, used for water table
regulation and flushing. The man-made water re-
gime is strongly influenced by climatological con-
ditions, resulting in a complex hydrology in the
system. Two different situations occur. Firstly, if
a precipitation deficit occurs, IJ sselmeer water is
let in near Lemmer and Tacozijl into the boezem
(Fig. 1) and from there inlet into the surrounding
polders takes place (generally in the summer pe-
riod, April-October). Secondly, if precipitation is
in excess, polder water is pumped towards the
lake system and from there into IJ sselmeer near
Lemmer (generally in the winter period, October-
April). For further information about this sea-
sonal variation, hydrology, eutrophication and
limnology of the lakes, see Beattie et al. (1978),
Moed & Hoogveld (1982), De Haan & Moed
(1984), Van Huet (1990; 1992).
Application of a phosphorus model
For the simulations a dynamic P balance model
was used (Van Huet, 1992; see also Eg. (1)). The
model focuses on phosphorus, assuming a signif-
icant correlation between TP concentrations and
chlorophyll-a concentrations. Chlorophyll-a, a
measure of phytoplankton biomass, can be used
as an approximate criterion of biological response
(Lambouet aI., 1983). A survey of about 35 Dutch
lakes indicates that at TP concentrations below
0.07 mg 1- 1 blue-green algae are not likely to
dominate (Lijklema et al., 1988). The model in-
puts are inflow and outflow discharges from a
hydrodynamic model (Van Huet, in press) and
TP concentrations in inflow discharges. In the
model, an overall net loss term is used to describe
the combined effect of sedimentation, resuspen-
sion and release of P. The simulations were run
over 2.5 years, with interruption in periods of ice
cover.
From the mass balance calculated using the
model (Table 1), it can be concluded that the loads
to Tjeukemeer are highest, moderate to Groote
Brekken and lowest to Slotermeer. Moreover, two
external loads dominate: the loads discharged
through the Broeresloot and the Pier Christiaans-
loot or both into Tjeukemeer (Fig. 1) are high
(third period, mean value 875 kg d - 1). These
loads can be considered as indirect polder loads
(i.e. loads from polders east of the study area).
Secondly, in the second period the load from IJ s-
selmeer into Groote Brekken is high (mean value
190 kg d - 1).
In the system itself the mass flow rates in the
Follegasloot and the Prinses Margrietkanaal
dominate. The contributions from precipitation,
inlet via sluices and from the waste water treat-
ment plant near Lemmer are relatively low. The
contribution of direct polder TP loads is relatively
highest for Slotermeer, although the loads into
Slotermeer are only half as high as the loads into
Tjeukemeer. Because of the very high indirect
 273

A
WADDENZEE

ECHTENER VEENPOLOER

2.5

Fig. 1. A. The Frisian 'boezem'. B. Its position in The Netherlands. C. The S.W. lake district (e-: boundaries research area;
RWZI: waste water treatment plants; RW50: Motorway 50; from Van Huet, 1990).
274

Table I. TP balances of three lakes during three periods;

Tjeukemeer Groot Brekken Siotermeer

Period 2 3 2 3 2 3
(kg d - I) (kg d - I) (kg d - 1)

IN IN IN
Follegasloot 170 129 111 Follegasloot 121 100 198 Ee 54 97 81
Pier Christiaansloot 151 68 41 Langesloot 31 83 46 Siotergat 22 31 36
Broeresloot 281 443 834 Kromme Ee 27 44 35
Scharsterrijn 104 57 28 Prins. Margrietkan. 222 65 114
Polders 37 14 33 Polders 9 6 8 Polders 19 7 14
Precipitation 2 3 3 Precipitation 0 4 I Precipitation 1 2 2
Waste water plant 9 1 1
Inlet IJ sselmeer 1 190 36
Inlet slucies I 5 7
Total 745 714 1050 Total 421 498 446 Total 96 137 133

OUT OUT OUT

Follegasloot 121 100 198 Follegasloot 170 129 111 Ee 21 21 25
Pier Christiaansloot 170 288 484 Langesloot 4 5 12 Siotergat 45 54 56
Broeresloot 286 124 46 Kromme Ee 15 15 24
Scharsterrijn 32 53 154 Pmrins. Margrietkan. 49 249 228
Polders 4 22 15 Polders I 3 2 Polders 2 11 6
Infiltration 2 2 3 Infiltration 0 0 0 Infiltration I I I
Pumping into IJsselm. 168 75 54
Net loss 95 115 145 Net loss 13 16 20 Net loss 35 49 48
Total 710 704 1045 Total 420 492 451 Total 104 136 136
Storage 35 10 5 Storage 6 -5 Storage -8 -3

polder loads Tjeukemeer can be considered as the
typical polder lake in the area, and Siotermeer
less so. The contribution of net loss is relatively
highest for Siotermeer (Van Huet, 1992).
Phosphorus load reduction scenarios
In scenario studies two approaches are possible.
The first approach is to define a scenario and
evaluate whether under this scenario a target level
is attained. The second starts at a target level and
then seeks a combination of scenarios that satis-
fies this level. For the second method an opera-
tional definition on what is meant by 'satisfies a
level' is needed, while the set of satisfactory com-
binations can be very large. Because a limited set
of reasonable scenarios can be formulated rela-
tively easily, and the main objective is to illustrate
a response that can be achieved by P reduction,
the more practical first approach was followed.
The simulated scenarios focus on the reduction
of the two main external P-Ioads: inlet water from
IJ sselmeer and water pumped from the polders
into the system. Peak loads from IJ sselmeer oc-
curred in the summer of 1986, and from polders
mainly in the winters (especially the indirect
polder contribution, see above).
Preliminary simulations showed that small re-
ductions in TP concentration from the main
sources and 100 % reductions in, for example the
waste water effluent, did not result in perceivable
changes in P-Ievels in the system (see also Ta-
ble 1). Apparently, only large reductions were
likely to have effects, therefore the reduction lev-
els in the various scenarios were set arbitrarily at

75% throughout. The model is approximately lin-
ear concerning TP concentrations and therefore
allows a crude estimate of the effects of larger or
smaller reduction levels from the effects. The sim-
ulated scenarios are a 75% TP reduction in:
(1) indirectly discharged polder water;
(2) both indirectly discharged and directly
pumped polder water;
(3) inlet water from IJ sselmeer;
(4) water both from all polders and from IJ ssel-
meer.
The results of the scenario simulations (Figs 2-4)
show the following lines: (a) a line that corre-
sponds with the actual basic water quality stan-
dard in The Netherlands for the summer period
(0.15 mg 1- I), (b) a line that corresponds with the
target TP concentration of0.07 mg 1- 1, (c) a basic
line that corresponds with simulation of the ac-
tual TP concentrations during the three periods
(Van Huet, 1992), and (d) one or more lines that
correspond with the 75% reduction scenarios.
Results and discussion
The three simulation periods cover both a normal
situation (P-Ioads from polders in winter and from
IJ sselmeer in summer), and an unusual situation,
such as high polder P-Ioads in summers of 1985
and 1987, due to precipitation excess.
A 75% reduction of TP concentrations in
indirectly discharged polder water, i.e. from
Broeresloot and Pier Christiaansloot (scenario 1;
Fig. 2) has, as expected, the greatest effect on the
P-Ievel in Tjeukemeer. In this lake, TP concen-
trations below the actual standard of 0.15 mg 1- I
can be achieved, while the 0.07 mg 1- I target
level is approximated. In Groote Brekken the
0.15 mg 1- I standard is approximated occasion-
ally. In this lake the effect is highest in the third
period due to high loads discharged through
Follegasloot (198 kg d - 1). In Siotermeer the ef-
fects of scenario 1 are poor, as expected, because
the distance from Broeresloot and Pier Chris-
tiaansloot is long. A further reduction ofTP con-
275
centrations in polder water directly pumped into
the system (scenario 2) does not substantially in-
crease the effect of scenario 1 in Tjeukemeer and
Groote Brekken, while the effect is relatively high-
est in Siotermeer (see also the TP-balance, Ta-
ble 1).
A reduction of TP concentrations in inlet
water from IJ sselmeer (Fig. 3, scenario 3) has,
as expected (see Table I), the greatest effect in
Groote Brekken, as this lake is situated close to
IJ sselmeer. In Groote Brekken the 0.15 mg 1- I
standard is achieved, and occasionally the
0.07 mg 1- 1 target level is approximated. In Tjeu-
kemeer only in the summer period of 1986 the
0.15 mg 1- 1 standard is achieved, while the effect
in Siotermeer is relatively low.
The effects of all reductions together (sce-
nario 4; Fig. 4) satisfies the 0.15 mg 1- 1 standard
in Tjeukemeer and Groote Brekken, while occa-
sionally the 0.07 mg 1- 1 target level can be
reached. Siotermeer responds only to a small ex-
tent to these measures.
The simulations started in each period with the
actual (measured) TP concentrations in the lakes,
while the 75% reductions were assumed to come
into effect at the beginning of each period. Con-
sequently, the results represent the short-term
transient effects for each simulated period sepa-
rately. In order to get an impression of the mea-
sures already in effect before, Fig. 4 also shows
the results of the simulations where the initial TP
concentrations in each lake are set to 50% of their
original value. As expected, the initial condition
effect is overruled by the discharged loads within
a period of some months.
The dynamic behaviour can be explained by
considering the dynamic mass balance equation
for each lake, derived from Van Huet (1992):
dP
AH - = L {Qin J(P in J - P)} + Qe P
dt "
(1)
where A = lake surface area (m 2 ); H = water
depth (m); P = lake TP concentration (mg 1- 1);
Pin,} = TP concentration in influent j (mg 1- I);
Qin,} = inflow discharge j (m 3 d - 1); Qe = evapo-
276

A Tjeukemeer
0.60 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,

0.50

0.40

0.30

0.20

0.10
t----------...::::..::::~---~--'~-::!---=---~
0.00 ' - - - - - - - - - - - - - - - ' - - - - - - - - - - - - ' - - - - - - - - - '
8 AUG 1985 8 AUG 1986 8 AUG 1987
-- Actual 2

B Groote Brekken
0.60 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,

0.50

0.40

0.30

0.20

0.10
t----------------------1
0.00 ' - - - - - - - - - - - - - - - - ' ' - - - - - - - - - - - - - - - ' - - - - - - - '
8 AUG 1985 8 AUG 1986 8 AUG 1987

-- Actual 2

C Siotermeer
0.60 r----------------------------,

0.50

0.40

0.30

0.20

0.10
~----------------------__t
0.00 '-- '-- -----''-- ---J

8 AUG 1985 8 AUG 1986 8 AUG 1987

-- Actual -----. 1 2

Fig. 2. Effects of 75% TP-reduction in three Frisian lakes in (1) indirectly discharged and (2) both indirectly discharged and di-
rectly pumped polder water, during three simulated periods. Actual: simulation results of actual TP concentrations (_).

ration (m 3 d - 1; Vs = apparent settling rate
(m d -1).
The transfer function from load discharges and
load concentrations to in-lake concentration is
generally difficult to evaluate from Eq. (1), be-
cause both Qin,}, Pin,} as well as P and Hare
functions of time, and consequently Eq. (1) is
non-linear. Also, for the three lakes the equations
are coupled. A full analysis would require the
linearization of the equations around a suitable
 277

A Tjeukemeer
0.60 , - - - - - - - - - - - - - - - - - - - - - - - - - - - _ - - - - ,

0.50

0.40

0.30
a.
0.20

0.10 ~--------------------------i
0.00 L.. --' ---' -.J

8 AUG 1985 8 AUG 1986 8 AUG 1987.

-- Actual

B Groote Brekken
0.60 r---------------------------,

0.50

0.40

§ 0.30
a.
0.20

:...'.'::+
0.10 ~---------..:.;f-···:....t·,-· -"- -1
·~V
0.00 L.. --'L- ---' --'

8 AUG 1985 8 AUG 1986 8 AUG 1987

-- Actual

C Siotermeer
0.60 r-----------------------------,

0.50

0.40

0.30
a.
0.20

0.10 ~-------------------------~
0.00 L.. L- --'L- ---.I

8 AUG 1985 8 AUG 1986 8 AUG 1987

-- Actual

0
Fig. 3. Effect of 75 0 TP-reduction in three Frisian lakes in (1) inlet water from IJsselmeer, during three simulated periods. Ac-

tual: simulation results of the actual TP concentrations (_).

point, e.g. the period-wise average, followed by
calculation of the deviation in output P concen-
trations vs. deviations in Qin and Pin in the La-
place domain. A crude appraisal, however, is pos-
sible by observing that AH = V = volume is
approximately constant in the long term, and by
combining all different loads into one single term:
(2)
Treating each lake independently, Eg. (1) can be
written in the standard form of a first order dy-
namic mass system:
278

A Tjeukemeer

0.60 I---------------------------~

0.50

0.40

0.30

0.20

0.10
r---------.-.:~-=::::,_....L----~_==d_~~--.:....-
0.00 ' - - - - - - - - - - - - - ' - - - - - - - - - - - - - ' - - - - - _
e aug 1985 8 AUG 1986 8 AUG 1987

--- Actual 2

B Groote Brekken
0.60 I----------------------------~

0.50

0.40

0.30

0.20

0.10
r-------.:..:::..-.....!......:....:..--\+-------:..:~i--.:.---___l
0.00 L- -'- -'- ---.J
8 AUG 1985 8 AUG 1986 8 AUG 1987

--- Actual 2

C Siotermeer
0.60 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,

0.50

0.40

0.30

0.20

0.10

0.00
t-------------------------
L- ----' ---'- _

8 AUG 1985 8 AUG 1986 8 AUG 1987

--- Actual 2

Fig. 4. Effects of 75%-reduction in three Frisian lakes in (I) both polder water and water from IJsselmeer and (2) the same ex-
pressed in the case of halving the initial TP concentrations (2), during three simulated periods. Actual: simulation results of the
actual TP concentrations (-).

dP 1
8- + P = k L (3) 8= !--- (4)
dt P' Vs
+- !
where 8= the effective time constant: H
and ! = the adapted hydraulic residence time:
 279

v This can occur if (a) there are strong fluctuations

T=-----
~Qin,j - Qe
and k p = the static gain:
8 fects, i.e. intermittent discharges of polder water.
kp =-.
V
Equation (3) shows that each lake roughly fol-
lows a first order dynamics. The response time of
the TP concentration in the lake to a step change
in external load is directly related to 8 (95 % of the
ultimate response is reached in about 38 time
units). Equation (4) shows that the effective time
constant is at most equal to the adapted hydraulic
residence time, but usually smaller because of the
net loss to the sediment (if Vs = 0 then 8= T, in all
other cases with positive Vs : 8< T). The final
concentration effect is k p units per unit step load-
ing change (i.e. mg 1- 1 concentration effect per
kg d - 1 load reduction).
The values of the effective time constant and
static gain were calculated for each of the lakes
for each of the three periods (Table 2). These re-
sults aid an interpretation of the graphical output.
Of all lakes, Groote Brekken has the fastest dy-
namics with effective time constants < 5 days.
This explains why P concentrations in this lake
show many more high frequency variations than
in the other two lakes. Load variations are not as
strongly attenuated as in Tjeukemeer and Sioter-
meer. In particular the P-behaviour of the model
for Siotermeer is very smooth, due to the effective
time constants in the order of 50 days. It is in-
teresting to note that in contrast to the model the
real data show appreciable fluctuations in time.
(5) in load, or (b) the assumption of complete mixing
is occasionally violated. In view of the location of
the sampling points, there are, in fact, reasons to
believe that the data are influenced by local ef-
(6) The static gains of the three lakes also differ.
The final concentration effect in Siotermeer is
about 4 times as large per kg d - 1 load reduction.
That this cannot be observed in Fig. 4 is because
a 75% load reduction does not correspond to the
similar amount of mass reduction for the three
lakes. The mass balance data (Table 1) show that
the load to Siotermeer is much less than to Tjeu-
kemeer, and consequently a 75% reduction has
much less effect despite the larger static gain. A
direct comparison of the load effect calculated
from the static gain with that seen in the figures
is not easy, because the lakes are interconnected.
A true first order response can only be expected
if the lake acts as the upstream end of the inter-
connected system. This is the case with Tjeuke-
meer, when there is no IJ sselmeer inlet, and with
Groote Brekken when there is inlet from IJ ssel-
meer. For example, considering the independent
loads due to indirectly and directly discharged
polder water into Tjeukemeer in period 2, the
mass balance data show that these loads amount
to 525 kg d - 1 in the untreated situation. A 75 %
reduction would thus be about 394 kg d - 1, lead-
ing to a concentration effect of 0.22 mg 1- 1, based
on the calculated static gain. Comparing this value
with the 75% reduction line in Fig. 2 shows that
the true effect is smaller, because of the capacity
of the connected lakes. Similarly, for Groote
Brekken in the second period, when there is con-

Table 2. Calculated values for the adapted hydraulic load, the effective time constant and the static gain.

V H I' s T 0 kp
(10 6 m 3 ) (m) (md- 1)
(d) (d) (10- 3 mg 1-1)(kg d -I)

Period 2 3 2 3 2 3

Tjeukemeer 34A· 1.76 0.015 18 23 19 16 19 16 0.45 0.56 0.47

Gr. breken 6.0 2.01 0.015 4 4 6 4 4 6 0.65 0.65 0.96
Slotermeer 18.1 1.64 0.015 54 59 58 36 38 38 2.00 2.12 2.10
280
siderable inlet of IJ sselmeer water, the 75 % load
reduction in the inlet water amounts to about
143 kg d - 1, leading to a calculated response of
0.1 mg 1- I, quite similar to the result obtained by
the simulation. With more complex hydrologic
conditions the simple analysis above is not ap-
plicable, and the simulation model must be used
instead.
The calculated effects of external load reduc-
tion, in reality will probably partly be counter-
acted by an increase of internal loading, because
at lower TP levels in the lakes P release is likely
to increase (Marsden, 1989). Effectively Vs will
decrease. Should this occur, even larger load
reduction would be needed. Thus, the calcula-
tions only roughly indicate the minimal require-
ment for an appreciable improvement in water
quality.
The role of the bottom sediment, long-term ef-
fects and manipulation of water transport (hydro-
logical scenarios) could not be simulated with the
model. These scenarios were roughly simulated
by Brinkman et al. (1989), who used a similar
model, covering one long-term period (1974-
1986). They concluded that reduction of sedimen-
tary P release, as well as hydrological measures,
such as increase or decrease of inlet from IJssel-
meer, did not lead to reduction in TP concentra-
tions to below 0.15 mg 1- 1. Only P-Ioad reduc-
tion measures similar to those in my study led to
an appreciable reduction in P levels: moreover,
long-term scenarios agreed well with the short-
term prospects as described in this paper.
Simulations for the whole boezem network
using other models (Delft Hydraulics Laboratory,
1986; Brinkman et al., 1987) indicated that flush-
ing the system with water from IJsselmeer (e.g.
during 1976 and 1977) positively affected the shift
in dominance from cyanobacteria towards green
algae due to the shorter residence times. How-
ever, the results, especially for the northern part
of the boezem, were not satisfactory, probably
because the equations used did not include the
effect of sediment-water P-exchange processes.
Moreover, 1976 was a very unusual year: because
of extremely dry conditions there was a large pre-
cipitation deficit in the summer period.
Concluding remarks
The model study shows that only a combination
of measures can lead to acceptable TP-Ievels in
the canal-lake ecosystem: large (at least 75 %)
TP-reductions in water from IJ sselmeer and from
polders are necessary to achieve a algae response
and, finally, restoration. The effects in the two
large lakes in the area, Tjeukemeer and Sioter-
meer, differ. In Tjeukemeer the 0.07 mg 1- 1 target
level can be approximated by the 75% TP-
reduction scenario, while in Siotermeer only the
0.15 mg 1- 1 standard is achieved. Tjeukemeer is
influenced largely by the indirect polder loads,
Groote Brekken by the IJ sselmeer loads, while
Siotermeer seems a relatively isolated lake in the
area. The results are very similar to the results of
a long-term model.
Acknowledgements
This study was supported by a grant of the Dutch
Ministry of Housing, Planning and Environment
and the Department of Public Works and Envi-
ronment of the Province of Friesland. I would like
to thank Gerrit Van Straten and John Rijnsdorp
for critically reading the manuscript. I am greatly
indebted to Meta Donk for the improvement of
the English text.
References
Ahlgren, I., 1988. Nutrient dynamics and trophic state re-
sponse of two eutrophicated lakes after reduced nutrient
loading. In: Balway, G. (ed.), Eutrophication and lake res-
toration. Water quality and biological impacts. Thonon-les-
Bains, 1988.
Andersson, G., 1988. Restoration of Lake Trummen, Swe-
den. Effects of sediment removal and fish manipulation. In:
Balway, G. (ed.), Eutrophication and lake restoration.
Water quality and biological impacts. Thonon-les-Bains,
1988.
Balmer, P. & B. Hultman, 1988. Control of phosphorus dis-
charges: present situation and trends. Hydrobiologia 170:
305-319.
Beattie, D. M., H. L. Golterman & J. Vijverberg, 1978. An
introduction to the limnology of the Friesian lakes. Hydro-
biologia 58: 49-64.

Bjork, S., 1985. Scandinavian lake restoration activities. In:
Vismara, R., R. Marforio, V. Mezzanotte & S. Cernuschi
(Eds), Lake pollution and recovery. Proc. EWPCA Con-
gress, Rome, 1985.
Brinkman, A. G., H. J. W. J. Van Huet & G. Van Straten,
1989. Eutrophication research in Friesland (in Dutch).
Final Report. Twente University and Limnological Insti-
tute: 55 pp.
Brinkman, J. J., P. S. Griffioen, S. Groot & F. L. Los, 1987.
A water quantity and water quality model for the evaluation
of water management strategies in the Netherlands - Ap-
plication to the province of Friesland -. In: Beck, M. B.
(ed.), Systems analysis in water quality management. Per-
gamon Press 1987: 23-30.
Burden, D. G., R. F. Malone & C. E. Mericas, 1987. Isola-
tion as a restoration strategy for nutrient reduction in a
small hypertrophic lake. Wat. Res. 4: 385-393.
Cullen, P. & C. Forsberg, 1988. Experiences with reducing
point sources of phosphorus to lakes. Hydrobiologia 170:
321-336.
De Haan, H. & J. R. Moed, 1984. Phosphorus, nitrogen and
chlorophyll-a concentrations in a typical Dutch polder lake,
Tjeukemeer, in relation to its water regime between 1968
and 1982. Wat. Sci. Tech. 17: 733-743.
De Haan, H., H. L. Hoogveld, T. De Boer, J. Voerman, J. R.
Moed, H. A. Kramer & J. Schrotenboer, 1988. Manipula-
tion of chemistry and phytoplankton by hydrological inter-
vention: a whole lake experiment in the northern Nether-
lands. Freshw. BioI. 20: 395-406.
Delft Hydraulics Laboratory, 1986. PAWN-II. Water quality
and models. Application to Friesland (in Dutch): 153 pp.
Edmondson, W. T., 1985. Recovery of Lake Washington from
eutrophication. In: Vismara, R., R. Marforio, V. Mezzan-
otte & S. Cernuschi (eds), Lake pollution and recovery.
Proc. EWPCA Congress, Rome, 1985: 308-314.
Foy, R. H., 1985. Phosphorus inactivation in a eutrophic lake
by the direct addition of ferric aluminium sulphate: impact
on iron and phosphorus. Freshw. BioI. 15: 613-629.
Foy, R. H. & A. G. Fitzsimons, 1987. Phosphorus inactiva-
tion in a eutrophic lake by the direct addition of ferric
aluminium sulphate: changes in phytoplankton populations.
Freshw. BioI. 17: 1-13.
Hosper, S. H., 1984. Restoration of Lake Veluwe, The Neth-
erlands, by reduction of phosphorus loading and flushing.
Wat. Sci. Tech. 17: 757-768.
Imboden, D. M., 1985. Restoration of a Swiss lake by internal
measures: can models explain reality. In: Vismara, R., R.
281
Marforio, V. Mezzanotte & S. Cernuschi (eds), Lake pol-
lution and recovery. Proc. EWPCA Congress, Rome, 1985:
91-102.
Lambou, V. W., W. D. Taylor, S. C. Hern & L. R. Williams,
1983. Comparisons of trophic state measurements. Wat.
Res. 11: 1619-1626.
Lijklema, L., J. H. Janse, R. M. M. Roijackers & M. L. Me-
ijer, 1988. Eutrophication in The Netherlands (in Dutch).
H 2 0 17: 462-467.
Marsden, M. W., 1989. Lake restoration by reducing external
phosphorus loading: the influence of sediment phosphorus
release. Freshwat. BioI. 21: 139-162.
Moed, J. R. & H. L. Hoogveld, 1982. The algal periodicity
in Tjeukemeer during 1968-1978. Hydrobiologia 95: 223-
234.
Richter, A. F., 1986. Biomanipulation and its feasibility
for water quality management in shallow eutrophic water
bodies in the Netherlands. Hydrobiological Bull. 20: 165-
172.
Ryding, S. 0., 1982. Lake Trehorningen restoration project.
Changes in water quality after sediment dredging. Hydro-
biologia 92: 549-558.
Van Densen, W. L. T., C. Dijkers & R. Veerman, 1986. The
fish community of the Loosdrecht Lakes and the perspec-
tive for biomanipulation. Hydrobiol. Bull. 20: 147-163.
Van Donk, E., R. D. Gulati & M. P. Grimm, 1989. Food-web
manipulation in Lake Zwemlust: positive and negative ef-
fects during the first two years. Hydrobiol. Bull. 23: 19-34.
Van Huet, H. J. W. J., 1990. Phosphorus eutrophication re-
search in the lake district of south western Friesland, The
Netherlands. Preliminary results of abiotic studies. Hydro-
biologia 191: 175-185.
Van Huet, H. J. W. J., 1991. Phosphorus loads from peaty
polders in the peaty S.W. Frisian lake district, The Neth-
erlands. Wat. Air Soil. Pollut. 55: 321-335.
Van Huet, H. J. W. J., in press. Application ofa wind-driven
hydrodynamic model for the quantification of transport in
an open boundary canal-lakes system in S.W. Friesland,
The Netherlands. Aquatic Sciences.
Van Huet, H. J. W. J., b. Phosphorus eutrophication in the
SW Frisian lake district. 1. Monitoring and assessment of
a dynamic mass balance model. Hydrobiologia. 233: 259-
270.
Van Huet, H. J. W. J. & De Haan, H., in press. Horizontal
and vertical distribution of phosphorus in sediments of in-
terconnected eutrophic polder lakes in S.W. Friesland, The
Netherlands. Archiv fOr Hydrobiologie.
Hydrobiologia 233: 283-287, 1992.
L. Van Liere & R.D. Gulati (eds). Restoration and Recovery oj Shallow Eutrophic Lake Ecosystems in The Netherlands. 283
© 1992 Kluwer Academic Publishers.

Restoration and recovery of shallow eutrophic lake ecosystems in The

Netherlands: epilogue

Louis Van Liere 1• 2 & Ramesh D. Gulati 2

1 National Institute of Public Health and Environmental Protection, Postbox 1, 3720 BA Bilthoven, The

Netherlands; 2 Limnological Institute, Rijksstraatweg 6, 3631 AC Nieuwersluis, The Netherlands

Key words: eutrophication, lake restoration, recovery, water management

Abstract

During the symposium 'Restoration and recovery of shallow lake ecosystems in The Netherlands' studies
on restoration of eutrophic lakes were addressed and discussed. Many Dutch shallow lakes have received
high external loadings of phosphorus through supply water that is influenced by the River Rhine and
loadings in The Netherlands. Two important Action Plans (the Rhine Action Plan, the North Sea Ac-
tion Plan) are now in operation to reduce nutrient emissions. The targets set are not likely to be fully
reached, so that supplementary reduction of phosphorus supplied to inland fresh waters will be required.
In several shallow lakes such a reduction has been achieved recently, but without leading to discern-
ible recovery. The main causes of delay are phosphorus storage and its subsequent release from sedi-
ments and foodweb; however, the remaining extraneous phosphorus supply is often still too high.
Supplementary actions are, therefore, called for. A further reduction of phosphorus inputs is suggested,
besides supplementary measures proposed, viz. dredging, flushing, biomanipulation, chemomanipulation.
Restoration to the past situation via upwelling groundwater appeares to be feasible in some cases. There
is a common consensus that each lake behaves differently depending to its morphology, hydrology and
history of eutrophication. Therefore each lake has to be studied before restoration measures can be
applied. Besides, the ecosystem should not only be studied as a separate entity, but as a part of sys-
tems of a higher integration level.

Introduction The Netherlands added in 1989 14.4 106 kg P and

The River Rhine on entering The Netherlands at
Lobith contains high concentrations of phospho-
rus (0.3 mg P 1- 1) and nitrogen (6.0 mg N 1- 1).
These high concentrations in the incoming water
results in an 'external load' of 20 106 kg phos-
phorus y - 1 and 352 106 kg nitrogen y - 1, being
80 % and 88 %, respectively, of the total external
nutrient load to The Netherlands by border-
crossing rivers (averages for 1989; Van Der Gaag
et al., 1991). The discharges to the River Rhine in
40.6 106 kg N, resulting in concentrations of
0.4 mg P 1- 1 and 5.0 mg N 1- 1 at Maassluis, i.e.
just before the river water is discharged into the
North Sea (Van Der Gaag et aI., 1991). Thus the
North Sea has had been receiving, besides phos-
phorus, a considerable amount of nitrogen, the
main algal growth-limiting nutrient in this sea,
resulting in frequent phytoplankton blooms in the
coastal area. These blooms in both 1989 and 1990
lasted for nine weeks (Van Der Gaag et al., 1991).
Dutch inland waters, especially those in the low
284
lying areas, mostly below sea level, use water from
the River Rhine, to regulate water level or to
counteract the salt water intrusion from the sea.
Thus, they receive considerable amounts of phos-
phorus, the limiting nutrient in freshwater sys-
tems of the temperate zone, apart from various
other sources and discharges. The colour of water
in these waters is often compared with diluted pea
soup. The year-round dominance of cyanobacte-
ria blooms is a rule rather than an exception.
After the Sandoz accident (Switzerland) in
November 1986 the countries in the catchment of
the River Rhine drew up the Rhine Action Plan
to restore the heavily polluted river and hasten its
recovery. One of the actions envisaged in the plan
is to reduce the phosphorus load to the river by
50% in 1995 compared with that in 1985 (ICRP,
1987). The countries surrounding the North Sea
decided to reduce nitrogen load by 50 % in 1995
compared with that in 1985 (NAP, 1990). The
Dutch ministries have planned at measures for a
further 50 % reduction to that planned in RAP/
NAP demands for phosphorus, and intend to di-
minish the load of nitrogen by 70% compared
with 1985 (V&W, 1989; VROM, 1989). These
targets are planned to be reached by 2010, but
preferably earlier.
With the measures taken the targets will not be
reached in 1995: only a 44 % emission reduction
in phosphorus will be achieved, and only a 23 %
reduction in that of nitrogen. In 2010, both the
nutrient emissions will be reduced only by 45 %
(Van Der Gaag et a!., 1991). An important hur-
dle is the intensive and continuing agricultural
activities, which have loaded the soil, and in the
case of nitrogen also the groundwater (Willems &
Hoogervorst, 1991). The share of agricultural
waste discharge in the phosphorus- and nitrogen
loadings of surface waters in the Netherlands is
21 % and 68 %, respectively (Van Der Gaag et al.,
1991). In this regard the situation is similar in the
other West European countries (Isermann, 1990).
Leaching of phosphorus saturated grounds to the
surface water may continue for many years to
come (Van Der Gaag et al., 1991) such that fur-
ther measures are needed if the targets are to be
met. However, even if the targets of the Rhine
Action Plan are met supplementary measures will
be needed for many inland freshwater lakes, the
external phosphorus load of which at present var-
ies between 0.05 and> 5 g Pm - 2 y- I (Van Der
Does et al., 1992).
Resume, recovery of shallow inland waters
Measures to reduce external phosphorus loading
in Loosdrecht lakes have been implemented al-
ready, and their effects on the lake ecosystem
have been presented in great detail. It was ob-
served and explained that a reduction of external
phosphorus from 0.69 to 0.35 g Pm - 2 Y- I will
not result in a conceivable change for a very long
time, (Van Liere & J anse, 1992; J anse et a!., 1992).
A further phosphorus reduction in external
sources to c. 0.15 g P m - 2 Y- I, which is techni-
cally possible, was proposed by Van Liere & J anse
(1992). The Province of Utrecht plans to realise
this additional reduction in Loosdrecht lakes in
near future (De Ruiter, 1992). According to model
calculations these measures will reduce chloro-
phyll a concentration (lanse et al., 1992). This
will mark the beginning of recovery, since pro-
duction of phytoplankton, thus also of the newly-
formed detritus, will be reduced (GOllS et al.,
1992; Otten et al., 1992). After reduction of ex-
ternal phosphorus and response of the ecosys-
tem, the supplementary measures are more feasi-
ble (lanse et al., 1992).
Although the research of the working group
Water Quality research Loosdrecht lakes inves-
tigated and quantified many basic mechanisms
involving the functioning of shallow lake ecosys-
tems, the findings cannot right away be extrapo-
lated to other lake ecosystems. For example, the
seepage which results in phosphorus losses
(Keizer & Sinke, 1992), although quite a common
feature in The Netherlands, does not occur in
every lake. Furthermore, the rates of internal and
external phosphorus loading and ratios of these
loadings greatly differ among lakes (Van Der Does
et a!., 1992; Verstraelen et al., 1992). Should the
management objectives encompass both abate-
ment of eutrophication and re-establishment of

the former plant vegetation, the ionic composition
of the supply water has to be considered, for ex-
ample by return through upwelling waters, char-
acteristic of the lake ecosystems some 100 years
ago (Verstraelen et al., 1992; Barendregt et al.,
1992). Even assigning other functions to certain
polders, or parts thereof, is an alternative mea-
sure in order to restore ecosystems to their pre-
vious states (Verstraelen et al., 1992).
All participants agreed that the reduction of the
external phosphorus sources needs to be achieved
first. The next question is: which target needs to
be set to attain and maintain the desired situa-
tion? At present the standard for the summer
averaged total phosphorus concentration in stag-
nant freshwaters in The Netherlands is 0.15 mg
Pl- 1 (V&W, 1989; VROM, 1989). This value is
too high. At such total phosphorus concentration
chlorophyll a summer averaged concentrations
generally far exceeding 100 J1g 1- 1 will be attained
(e.g. Van Liere et ai., 1992). Van Der Does et al.
(1992) consider, therefore, 'regional' standards,
depending on the function allocated to the eco-
system: e.g. recreation, drinking water produc-
tion, nature conservation. They suggest a general
standard of < 0.08 mg P 1- I, and for nature con-
servation purposes propose < 0.05 mg P 1- 1 as a
desirable concentration. Deep lakes and ditches
may have higher values.
Van Huet (1992a, b) simulated a target value of
0.08 mg P 1- 1 to be accomplished if the external
load to Frisian lakes were to decrease by 75%.
Van Der Does et al. (1992) observed rapid im-
provement in Nieuwkoop lakes in the first year
after reduction of the external phosphorus load
by 90%. However, in these lakes the internal
loading was found to be low, compared with the
external load. In the lake Geerplas, a small lake,
the internal phosphorus loading was quantita-
tively so high that the lake was dredged. This gave
rise to serious technical problems, such that ad-
ditional dredging was necessary with special at-
tention for the extremely loose sediments to re-
move it and prevent its dispersal during dredging
(Van Der Does et ai., 1992). Therefore, experi-
mental dredging on a pilot scale was recom-
mended to study its effects and possibilities (Ver-
285
straelen et al., 1992). Dredging in small systems
was found to be more practicable (Hovenkamp-
Obbema & Fieggen, 1992): it improved signifi-
cantly the trophic status of the peaty ditch system
in the polder Wormer, Jisp and Nek. However,
fish stock management in this system did not re-
sult in additional improvement, presumably be-
cause large-bodied crustacean zooplankton failed
to develop (Hovenkamp-Obbema & Fieggen,
1992). This was also true for the Lake Klein,
Vogelenzang (Van Der Vlugt et al., 1992). Fila-
mentous cyanobacteria, which dominate in these
systems are apparently difficult for the crusta-
ceans to consume (Dawidowicz et ai., 1990). Van
Der Vlugt et al. (1992) reported that the remain-
ing refractory detritus, resuspended frequently by
wind action created high turbidity and thus hin-
dered growth of waterplants needed to balance
the system after the biomanipulation. This con-
clusion is supported by the research outcome of
Loosdrecht lakes (Gons & Otten, 1991; Gons
et al., 1992; Otten et al., 1992). It still needs to be
demonstrated whether fixation with iron(III)chlo-
ride will stabilize the extremely loose sediment
particles (Boers et al., 1992). Although it seems to
be hard to use biomanipulation in large shallow
lakes, exposed to wind action, there again may be
an exception. For example, in the Lake Wolder-
wijd (27 km 2 ) the fish-stock management mea-
sures gave rise to a rapid increase in Secchi disc
transparency. Large bodied zooplankton, espe-
cially Daphnia spp., flourished, and controlled
phytoplankton growth (Meijer et al., 1991). How-
ever, after a six-week-Iong clear-waterphase, the
temperature rise favoured increased development
of algae again. The optimistic outlook is an even
longer clear-water phase in the years to come, and
a steady increase in the role of the waterplants,
which have already moderately established (Van
der Gaag et al., 1991).
Flushing resulted in lower chlorophyll a con-
centration in Lake Veluwe. However, because of
the prominent role of detrital particles in the light
scattering, the water transparency did not in-
crease (Jagtman et al., 1992). If flushing were to
be used to control eutrophication, large volumes
of water of good quality would be needed, besides
286
the problem where to discharge the superfluous
eutrophic water used for flushing. Furthermore,
to make flushing possible a technical infrastruc-
ture is needed. For example, in the case of Loos-
drecht lakes, in which inlet and discharge are sit-
uated close to each other, flushing is not likely to
be successful (Van Liere & Janse, 1992) If sedi-
mented detritus is important, flushing will mainly
control new formation of detritus by controlling
algal growth, and may speed up recovery.
Conclusions
Most of the lakes in The Netherlands for which
restoration was attempted were highly eutrophic.
They had large amounts of phosphorus stored in
sediments or in the food web, or both. The inter-
nal loading in such lakes frustrated the restora-
tion efforts. It is quite likely that relatively less
eutrophic lakes can recover faster.
The lakes were eutrophicated for several de-
cades. Thus, an important quantity of refractory
detritus governed underwater light climate in the
shallow lake ecosystems by settling and wind-
induced resuspension. It is the loss rate of this
particulate matter by mineralization and dilution
that will regulate the rate of improvement in the
underwater light regime.
However, the fact remains that reduction of the
incoming phosphorus is the most important fea-
ture in lake restoration. It is not possible to state
the minimum permissible phosphorus load to
shallow lakes. However, the lower the external
input, the faster the recovery. When the phospho-
rus input is reduced drastically supplementary
measures will succeed.
Depending on their morphology, hydrology and
history of eutrophication, the lake ecosystems will
respond differently to restoration measures. Each
lake, that is to be restored, needs to be studied
beforehand in order to optimize the rehabilitation
measures. The aquatic ecosystem should not only
be studied as a separate entity, but as an part of
systems of a higher integration level.
References
Barendregt, A., S. M. E. Starn & M. J. Wassen, 1992. Res-
toration of fen ecosystems in the Vecht River plain: cost-
benefit analysis of hydrological alternatives. Hydrobiologia
233: 247-258.
Boers, P. C. M., J. Van Der Does, M. Quaak, J. C. Van
Der Vlugt & P. A. Walker, 1992. Fixation of phosphorus in
lake sediments using iron(III)chloride: experiences, expec-
tations (extended abstract). Hydrobiologia 233: 211-212.
Davidowicz, P., Z. M. Gliwicz & R. D. Gulati, 1990. Can
Daphnia prevent a blue-green algal bloom? A laboratory
test. Limnologica 19: 21-26.
De Ruiter, M. A., 1992. Water quality research in the Loos-
drecht lakes: proposals for the follow-up restoration mea-
sures. Hydrobiologia 233: 161-163.
Gons, H. J., T. Burger-Wiersma, J. H. Otten & M. Rijkeboer,
1992. Coupling of phytoplankton and detritus in a shallow,
eutrophic lake (Lake Loosdrecht, The Netherlands). Hy-
drobiologia 233: 51-59.
Gons, H. J., J. H. Otten & M. Rijkeboer, 1991. The signifi-
cance of wind resuspension for the predominance of fila-
mentous cyanobacteria in a shallow, eutrophic lake. Mem.
1st. ita!. Idrobio!. 48: 233-249.
Hovenkamp-Obbema, I. R. M. & W. Fieggen, 1992. The ef-
fects of dredging and fish stocking on the trophic status of
shallow, peaty ditches. Hydrobiologia 233: 225-233.
ICRP, 1987. International Commission for the Protection of
the Rhine against pollution: 'Rhine Action Plan'. Ed:
Technisch-wissenschaftliches Sekretariat, Koblenz, 24 pp.
Isermann, K., 1990. Share of agriculture in nitrogen and phos-
phorus emissions into the surface waters of Western Eu-
rope against the background of their eutrophication. Fer-
tilizer Res. 26: 253-269.
Jagtman, E., D. T. Van Der Molen & S. Vermij, 1992. The
influence offlushing on nutrient dynamics, composition and
densities of algae and transparency in Veluwemeer, The
Netherlands. Hydrobiologia 233: 187-196.
Janse,J. H., T. Aldenberg& P. R. G. Kramer, 1992. A math-
ematical model of the phosphorus cycle in Lake Loosdrecht
and simulation of additional measures. Hydrobiologia 233:
119-136.
Keizer, P. & A. J. C. Sinke, 1992. Phosphorus in the sediment
of the Loosdrecht lakes and its implications for lake res-
toration perspectives. Hydrobiologia 233: 39-50.
Meijer, M-L., E. Jagtman, M. P. Grimm, S. H. Hosper &
E. M. Blauw, 1991. Uitdunning visstand Wolderwijd is vol-
tooid. Eerste veelbelovende resultaten (Reduction of the
fish stock in Lake Wolderwijd has been completed. Pre-
liminary promising results). H 2 0 24: 368-369.
NAP, 1987. Ministers of the North Sea countries. 'Final dec-
laration of the Third International Conference on the Pro-
tection of the North Sea'. The Hague, 36 pp.
Otten, J. H., H. J. Gons & M. Rijkeboer, 1992. Dynamics of
phytoplankton detritus in a shallow, eutrophic lake (Lake
Loosdrecht, The Netherlands). Hydrobiologia 233: 61-68.

Van Der Does, J., P. J. T. Verstraelen, P. C. M. Boers, J. Van
Roestel, R. Roijackers & G. Moser, 1992. Lake restoration
with and without dredging of phosphorus-enriched upper
sediment layers. Hydrobiologia 233: 197-210.
Van Der Gaag, M. A., K. C. J. Van Der Ende, L. Van Liere
& J. H. Woudstra, 1991. Fluvial environmental problems.
In: R. J. M. Maas (ed), National Environmental Outlook
1990-2010, National Institute of Public Health and Envi-
ronmental Protection.
Van Der Vlugt,J. C., P. A. Walker,J. Van Der Does & A. J. P.
Raat, 1992. Fisheries management as an additional lake
restoration measure: biomanipulation scaling-up problems.
Hydrobiologia 233: 213-224.
Van Huet, H. J. W. J., 1992a. Phosphorus eutrophication in
the SW Frisian lake district. 1. Monitoring and assessment
of a dynamic mass balance model. Hydrobiologia 233:
259-270.
Van Huet, H. J. W. J., 1992b. Phosphorus eutrophication in
the SW Frisian lake district. 2. Phosphorus balance and
simulation of reduction scenarios. Hydrobiologia 23: 271-
281.
287
Van Liere, L. & J. H. Janse, 1992. Restoration and resilience
to recovery of the Lake Loosdrecht ecosystem in relation
to its phosphorus flow. Hydrobiologia 233: 95-104.
Verstraelen, P. J. T., J. Wisserhof, Lj. Rodic & R. Eijsink,
1992. Eutrophication control strategies for three shallow
Vecht lakes in the province of North Holland. Hydrobio-
logia 233: 235-245.
VROM, 1989. Nationaal Milieubeleidsplan, Ministerie voor
Volkshuisvesting, Ruimtelijke Ordening en Milieubeheer
(N ational Environmental Policy Plan, Ministry of Housing,
Physical Planning and Environment). SDU, The Hague.
V & W, 1989. Derde Nota Waterhuishouding, Ministerie van
Verkeer en Waterstaat (Third Water Policy Document,
Ministry of transport and Waterworks). SDU, The Hague.
Willems, W. J. & N. J. P. Hoogervorst, 1991. Eutrophication
of soil and groundwater. In: R. J. M. Maas (ed.), National
Environmental Outlook 1990-2010, National Institute of
Public Health and Environmental Protection.
Working group Water Quality Research Loosdrecht Lakes:
its history, structure, research programme, and some re-
sults. Hydrobiologia 233: 1-9.