Vegetation and Hydrology of Floating Rich-Fens

Vegetation and hydrology of floating rich-fens
Postscript
This is (printed from) a searchable reprint in portable document format. Figures, tables, and
the start of paragraphs are found on pages with the same numbers as in the original book, but
text flow at the bottom and top of some pages slightly changed due to reformatting. When citing,
refer to the original book reference:
Van Wirdum, Geert, 1991. Vegetation and hydrology of floating rich-fens. Datawyse,
Maastricht, 310 p. (ISBN 90-5291-045-6).
Comments and questions are welcomed by geert@jolicoeur.nl
The book was peer reviewed and accepted for publication by
Prof. Dr. P.J.M. van der Aart, University of Utrecht,
Prof. Dr. Ir. J.C. van Dam, Technical University of Delft,
Prof. Dr. P.J. Jungerius, University of Amsterdam,
Prof. Dr. P.J.C. Kuiper, University of Groningen and
Prof Dr. Ir. I.S. Zonneveld, International Institute for Aerospace Survey and Earth Observation (ITC), Enschede.
It was successfully defended as a PhD thesis February, 8th, 1991 with promotors
Prof. Dr. T. van der Hammen, University of Amsterdam and
Prof. Dr. W. H. van der Molen, University of Wageningen.
The following corrections and changes were made to avoid confusion:

Chapter 3:
p.37: footnote added;
p.43: last sentence corrected.
Chapter 5:
p.60: reference to Fig.4.4 added to explain shading;
p.75: legend of Fig.5.7 corrected.
Chapter 6:
p.99: note added.
Chapter 7:
p.116, 118: confusion between baulks and ditches in figures 7.3 and 7.4 removed.
Chapter 8:
p.135: formula corrected: z = 0 bz;
p.146-147: instances of the word reliable replaced by credible, the solution for being within the range
concluded in the section on heat capacity and thermal diffusivity, p.136-138;
p.150-152: some instances of variable I replaced by vl to reduce confusion between volume in the overall
budget and the related lateral velocity in the preferential flow channel;
p.154: footnote added.
Chapter 9:
p.164: confusion between baulks and ditches in figure 9.6 removed;
p.174: symbol error in graph pointed out.
Chapter 11:
p.217: reference corrected;
p.220: order of column headers in table 11.3 corrected.
Appendix C:
P240: Campylopus fragilis (subsp. pyriformis) now named C. pyriformis.
Appendix D:
p.253: Figure D.1 redone with the computer programme MAION to correct misplacement of At-W80;
p.250-251: formulae explaining mol(a) and its relation to mol(c) corrected;
p.260: enumerator in formula near the bottom of the page corrected: {1 + (0.33 aiI)};
p.264: last text block above section on Ionic Ratio corrected.
ACADEMISCH PROEFSCHRIFT
ter verkrijging van de graad van doctor

aan de Universiteit van Amsterdam,
op gezag van de Rector Magnificus
prof.dr. P.W.M. de Meijer
in het openbaar te verdedigen in de Aula der Universiteit
(Oude Lutherse Kerk, ingang Singel 411, hoek Spui),
op vrijdag 8 februari 1991, te 15.00 uur

door
Geert van Wirdum
geboren te Amsterdam
FACULTEIT DER BIOLOGIE
promotores
Prof. dr. T. van der Hammen
Prof. dr. W.H. van der Molen
CIP-DATA KONINKLIJKE BILIOTHEEK, DEN HAAG
Wirdum, Geert van

Vegetation and hydrology of floating rich-fens / Geert van
Wirdum. - Maastricht : Datawyse. - Ill.
Thesis university of Amsterdam. - With ref., - With
summary in Dutch
ISBN 90-5291-045-6
SISO 586.6 UDC 58:556.56(043.3) NUGI 825
Subject headings: eco-hydrology / fen vegetation / seepage.
Produktie: Datawyse Maastricht, Ruud Leliveld

Technisch tekenwerk: Arjan Griffioen, Ruut Wegman (RIN)
Druk: Datawyse Maastricht / Krips Repro Meppel
Het veldonderzoek in de periode 1973-'75 werd mogelijk gemaakt door een
subsidie van de Nederlandse Organisatie voor Zuiver Wetenschappelijk
Onderzoek, onderzoeksbijdragen van vele studenten, diverse diensten verleend
door het Hugo de Vries-Laboratorium van de Universiteit van Amsterdam, en
vergunningen voor veldonderzoek verleend door de Vereniging tot behoud van
Natuurmonumenten en het Staatsbosbeheer.
Het Rijksinstituut voor Natuurbeheer stelde mij in de periode 1975-1990 in de
gelegenheid een gedeelte van mijn tijd te besteden aan verdiepend onderzoek,
uitwerking van gegevens en rapportage.
What is obviously essential to something, but not well understood,

is often denoted as its `structure'
(cf Rolf Lohberg & Theo Lutz (1968), `Keiner weiss, was Kybernetik ist')
Human interest in nature is probably because of its structure
Quagfens have a notable, but weak and unstable structure
Preface
Early in 1969 Dr.S.Segal introduced me to the study of quagfens in North-West

Overijssel. Dr.J. van Donselaar and Dr.L. de Lange supported my deeper inquiry
from 1973 to 1975. With a sum total of some 35 inventive students around, the
main lines of quagfen eco-hydrology were drawn in those years. At my
appointment at the Research Institute for Nature Management (RIN) in 1975 I was
not prepared to practice the distance necessary to enjoy and report what I learned,
rather than aching for more. After twenty years I obviously can no longer continue
this without blame, and this period has fortunately been just long enough to get
some important clues as to what sort of natural devices quagfens are. In this thesis
I essentially report these clues on the basis of detailed observations in one of the
investigated complexes. Other quagfens contributed no less to what I learned, but
the interests of the family whose accomodation I joined withhold me from a
treatment now.
I am much indebted to Professors Meeuse, who replaced Dr.Van Donselaar, and
Van der Molen. It was impressive to experience their continued interest and
patience at each delayed rendez-vous in the coffee-shop "De Plantage" in
Amsterdam. Prof. Van der Hammen kindly accepted to replace Meeuse as
promotor when this became necessary.
Contents
CHAPTERS 1-6:
General introduction
Published evidence of seepage in rich-fen quagmires (quagfens)
Site properties indicated by the flora of quagfens
The area of North-West Overijssel
Aspects of the hydrology of De Weerribben
The distribution of seepage indicators in De Weerribben
15
19
29
47
57
89
CHAPTERS 7-10:
The quagfens of De Stobbenribben and their vegetation
Peat temperature and the estimation of vertical water flow
Lateral water flow in longitudinal transects
Environmental and vegetational processes in De Stobbenribben
113
131
155
177
CHAPTER 11:
Summary and general discussion
201
APPENDICES A-F:
Explanation of some special terms
Classification of quagfen vegetation
Indicator list of fen-mire species
Evaluation of the major-ionic composition of natural waters
Evapo-transpiration from lysimeters with fen vegetation
Data reports
223
229
239
247
285
291
REFERENCES
SAMENVATTING: Vegetatie en waterhuishouding van trilvenen
SUMMARY
295
303
307
Detailed contents
1.
GENERAL INTRODUCTION
15
2.
2.1
2.2
2.3
2.4
PUBLISHED EVIDENCE OF SEEPAGE IN RICH-FEN QUAGMIRES (QUAGFENS)

Introduction
Definitions of seepage
The seepage hypothesis for quagfens
The source of the hypothesis; main arguments and evidence
Phytosociology
Hydrology
Ecophysiology
Indicator species reported in studies of Dutch quagfens
Special remarks with regard to bryophytes
Liverworts
Mosses
Seepage and calcidity
Conclusions and re-formulation of the seepage hypothesis
19
19
20
20
21
21
22
22
23
24
24
24
25
27
SITE PROPERTIES INDICATED BY THE FLORA OF QUAGFENS

Introduction
Species and associations of species as indicators
Response models and Associations in phytosociology
The OR assumption: an additional explanation of species association
The individuality of mire elements
Different scales of aggregation
The compilation of a list of indicator species (Appendix C)
The phytosociological groups
Ecological indications according to the Central-European tradition
The Finnish mire types
The ecological significance of the phytosociological groups
29
29
30
30
32
32
32
36
36
37
39
43
THE AREA OF NORTH-WEST OVERIJSSEL

Introduction
Surface structure, land-use, and water management
The original mire
Human occupation and peat industry (Table 4.1)
Water management
Reclamations
Present land-use in the mire area
Geology
47
47
49
49
49
51
51
53
55
ASPECTS OF THE HYDROLOGY OF DE WEERRIBBEN

Introduction
The hydraulic potential distribution
Available data
The equipotential pattern in the top of the sand-bed
The potential distribution in the B-transect
The decrease of the hydraulic head in the aquifer, 1935-1975
The alteration of the hydraulic head gradient, 1935-1975
Chemical composition and age of groundwater
B1 (Paaslo morainic area)
B2 (Border between De Weerribben mire and IJsselham polder areas)
57
57
59
59
59
59
61
62
64
68
68
2.5
2.6
2.7
2.8
3.
3.1
3.2
3.3
3.4
3.5
4.
4.1
4.2
4.3
5.
5.1
5.2
5.3
5.4
5.5
5.6
5.7
6.
6.1
6.2
6.3
6.4
6.5
6.6
6.7
7.
7.1
7.2
7.3
7.4
7.5
8.
8.1
8.2
LM187 (Pierikken area)

B3 (Centrally in De Weerribben mire area)
B4 (Border between De Weerribben mire and Blankenham polder areas)
B5 (Blankenham polder area)
B6 and B7 (Noordoost-Polder)
Conclusion
The chemical composition of boezem water in the 1970s and '80s
Introduction
The dominant pattern of the time series
Spatial patterns and asynchronous variations
Surface water flow through the area (Fig.5.10a, 5.11)
Inflow of surplus polder water (Fig.5.10b)
The lithotrophic influence of the Steenwijk-Ossenzijl canal (Fig.5.10c)
The lithotrophic influence of discharging groundwater
An extension of the time series into the 1981-'87 period
Summary of water quality influences during the 1970s and '80s (Fig.5.13)
Comparison of surface water composition 1960-'82
Data-sets OLD and WRNET
Discussion: Did anything really change?
70
70
70
71
71
71
72
72
74
74
79
81
81
81
81
81
83
83
86
THE DISTRIBUTION OF SEEPAGE INDICATORS IN DE WEERRIBBEN

Introduction
Seepage indicators in the vegetation of De Weerribben
Is the distribution of seepage indicators restricted to an area of groundwater discharge?
Scorpidium scorpioides
Other species
Is the distribution of seepage indicators restricted to any particular area?
Scorpidium scorpioides and Liparis loeselii
Menyanthes trifoliata, Utricularia intermedia, and Parnassia palustris
Do stands of vegetation with seepage indicators indicate a particular type
of environment?
Has the distribution of seepage indicators changed in time?
Other species
A new hypothesis on the local behaviour of Scorpidium scorpioides and
associated species
Further corroborative distributional information
The distribution of salt indicators
The distribution of Stratiotes aloides
The distribution of some other species
89
89
90
94
94
96
98
98
99
100
107
108
109
111
THE QUAGFENS OF DE STOBBENRIBBEN AND THEIR VEGETATION

Introduction
Topography, petgaten and kraggen
The vegetation cover
Available data
The vegetation map (1973)
On the stability of the vegetational gradient
The vegetational zones
Description of the vegetational zones
113
113
115
117
117
119
120
121
125
PEAT TEMPERATURE AND THE ESTIMATION OF VERTICAL WATER FLOW

Introduction
Theory of soil temperature as a function of seepage
The general model
131
131
133
133
101
102
105
105
8.3
8.4
8.5
8.6
9.
9.1
9.2
9.3
9.4
9.5
9.6
10.
10.1
10.2
10.3
10.4
10.5
10.6
Heat capacity and thermal diffusivity of very wet peat soils

A wave analysis of annual temperature fluctuations
The estimation of seepage in De Stobbenribben
The gauges for temperature measurements
Simple implementations of the Doppler analogy method (DOPPSOL)
Summary of results with varieties of DOPPSOL
Implementations of the Stallman model (FOUSOL)
Results obtained with the FOUSOL method
Lateral heat flow: a disturbing factor
The temperature regime in the root zone
Conclusions
136
138
142
142
144
145
145
147
150
153
154
LATERAL WATER FLOW IN LONGITUDINAL TRANSECTS

Introduction
Data acquisition in longitudinal transects
Hydraulic head (water manometers)
Kragge movement
Conductivity and temperature sounding
Measuring schemes and data processing
The general pattern found
The hydraulic head gradient
The longitudinal conductivity gradient
The seasonal movement of bodies of groundwater
The conductivity map
Temperature gradients in longitudinal sections
Isopleth patterns
The causes of the spatial patterns of temperature data
The possible incidence of density currents
The chemical identity of different bodies of mire water
Methods of sampling and analysis
Analyses used
Method of interpretation
1980-1983 analytical results
Conclusions
155
155
156
156
157
158
159
159
159
160
160
163
165
165
167
167
169
169
170
171
171
174
ENVIRONMENTAL AND VEGETATIONAL PROCESSES IN

DE STOBBENRIBBEN
Introduction
Flow rate and hydraulic conductivity in the preferential flow channel
QUAGSOLVE: the mixing of water in the preferential flow channel
Deviating concentrations of non-conservative constituents
P and K
Inorganic nitrogen
Calcium
Conclusion
Gradients in plant biomass and nutrient state in De Stobbenribben
Changes in De Stobbenribben and their possible causes
177
11. SUMMARY AND GENERAL DISCUSSION

11.1 Introduction
11.2 The seepage hypothesis for Dutch quagfens re-formulated
Seepage
Base state as a nodal parameter
11.3 Relations between environment and vegetation
177
178
179
188
189
191
191
192
193
197
201
201
202
202
203
204
Recurrent patterns of heterogeneity

Indication by species
Base state and quagfen vegetation
11.4 The study area
11.5 Hydrology of De Weerribben
Interpretation of water analyses
Groundwater
Surface water
The changing eco-hydrological state
11.6 The distribution of seepage indicators in De Weerribben
11.7 The quagfens of De Stobbenribben and their vegetation
11.8 Peat temperature and the estimation of vertical water flow
11.9 Lateral flow in longitudinal transects in De Stobbenribben
11.10 Environmental and vegetational processes in De Stobbenribben
The QUAGSOLVE model
The nutrient balance
Non-steadiness of the environment
11.11 Management and the rich-fen environment in zoned mires
204
206
207
208
209
209
210
211
211
213
215
215
216
217
217
218
220
221
A.
A.1
A.2
EXPLANATION OF SOME SPECIFIC TERMS

Specific terms related to the mire type concerned
Specific terms related to the water management system
223
223
226
B.
CLASSIFICATION OF QUAGFEN VEGETATION

Westhoff & Den Held (1969)
Ellenberg (1978) and Oberdorfer (1979)
Dierssen (1982)
Wheeler (1975, 1980, 1982)
229
229
232
234
236
C.
INDICATOR LIST OF FEN-MIRE SPECIES
239
D.
EVALUATION OF THE MAJOR-IONIC COMPOSITION OF

NATURAL WATERS
Processing of water quality data
Notational and conceptual conventions
Electrical conductivity
Concentration units
The mole concept
Partial molar(c) fractions
Special ratios and similarity coefficients
Reliability of analytical data
The use of partial molar(c) fractions as measures of ionic composition
Triangular diagrams according to Piper
Radial diagrams according to Maucha
The use of special ratios for the comparison of water analyses
The use of coefficients of similarity in relation to chosen analyses
247
Definition of MAION functions and related procedures

The MAION program
The ionic balance and the electroneutrality test
The conductivity test
Methods provided by the pertaining literature
The activity-based method used in MAION
Conductometric activity coefficients
259
259
259
260
260
260
261
D.1
D.2
247
247
250
250
250
250
251
251
251
252
252
253
255
256
Compensation for temperature differences

Method of evaluation
The ionic ratio and related quantities
The use of total hardness to approximate IR, yielding IR*
The Cl- and Ca-based conductivity ratios EClR and ECaR, yielding IR Cl and IRCa
The MAION similarity coefficient
Saturation with respect to calcite
The LAT framework
Introduction
Statistical evidence for the importance of EC and IR
Determinant analysis
The LAT framework
The LAT framework and the hydrological cycle
Series formed by actual water analyses
Applicability at the global scale
MAION similarity: an extension of the EC-IR characterization
Visualization of the MAION feature vector
Visualization of similarities in the LAT framework
Inferences from the TH-LI diagram
A comparison with conventional statistical methods
262
264
264
265
265
268
269
270
270
270
270
272
275
275
275
278
278
279
280
283
E.
EVAPO-TRANSPIRATION FROM LYSIMETERS WITH FEN VEGETATION
285
F.
DATA REPORTS
291
D.3
REFERENCES
SAMENVATTING: Vegetatie en waterhuishouding van trilvenen
SUMMARY
295
303
307
Reprinted from: Geert van Wirdum, 1990. Vegetation and hydrology of floating rich-fens. Datawyse, Maastricht, 310 p. (ISBN
90-5291-045-6). Known errors corrected.
CHAPTER 1
This thesis deals with the ecological relation between the vegetation of rich-fen quagmires (quagfens)
in abandoned turbaries in North-West Overijssel (The Netherlands), and hydrological factors. The
seepage hypothesis for Dutch quagfens was a starting point for a broader definition of the problem area
to be explored (Chapter 2). That hypothesis aims at an ecological explanation of the rare occurrence of
certain (extreme) rich-fen species that have drawn the attention of nature conservation in many
North-West European countries. Most of these species are characteristic of the phytosociological
Association Scorpidio-Caricetum diandrae. In the low-lying part of The Netherlands their occurrence
is almost strictly limited to terrestrializing former turbaries to which a mowing regime is applied.
Elsewhere they are notable but uncommon plants in some dune slacks and valley fens, often also under
a mowing or light grazing regime.
This distributional pattern seems to apply to other countries at the same latitude also: Germany
(Ellenberg 1978), England (Wheeler, pers. comm.), and Poland (Tomaszewska 1988). In the absence
of a vegetation management (mowing) the stands in terrestrializing turbaries usually develop into a carr
vegetation. Mowing alone, however, is not sufficient to evoke the development of the type of
vegetation involved. In the terrestrializing turbaries in The Netherlands a striking difference exists
between quagfens developing quickly into Sphagnum reeds and quagfens with stands of
Scorpidio-Caricetum diandrae and related vegetation. These differences fit into the general scheme of
mire types developed in North-West Europe (compare the treatment by Moore & Bellamy 1973),
suggesting a rather broad relevance of the underlying ecological questions, even to the study of bog
formation.
Introduction
15
In many countries the re-generation of bog vegetation on cut-over bog surfaces is stimulated, but it
has appeared that such a re-generation is often much less easy to achieve than the formation of initial
bog vegetation on floating rafts, both in fenlands and in turf ponds excavated in a bog surface. It is
indeed remarkable to observe the development of dense carpets of Sphagnum papillosum and,
occasionally, even S. imbricatum and S. fuscum, upon rafts floating over bodies of even slightly
brackish water. At least two sets of factors appear to be decisive for the development of various sorts of
mire: the fluctuations of the water level (inclusive of the frequency of flooding), and the water
composition, depending on water flow. Quagmires offer an opportunity to investigate the influence of
water flow and water composition under almost `controlled' conditions as regards water level
fluctuations, still representative for the development of a broad range of types of mire vegetation. The
main line of this research leads from the milieu as it is indicated by plant species, via a hydrological
description of the environment, into an attempt quantitatively to understand the relationships between
the three.
In the first part of this thesis (Chapters 2,3) I analyse the seepage hypothesis for Dutch quagfens,
resulting in a reformulation of this hypothesis and the derivation of an ecological reference system (the
species-wise list of indications, Appendix C), and of certain questions that can be investigated. This is
followed by an introduction into the study area and the prevailing hydrological environment (Chapters
4,5). The water management in the study area appears to provide boundary conditions for the
development of terrestrializing quagfens, especially as regards the water supply. On the whole-area
scale this is indicated by the changing patterns of distribution of various species (Chapter 6). The
attention is drawn to the possible importance of processes that were not yet considered before,
especially a gradual desalting of parts of the area formerly influenced by slightly brackish water.
In reporting my results I was in the comfortable position that I had to make a choice out of the case
studies done and of various analyses of quantitative data. This choice was made in favour of a,
hopefully clear, expos of the rle of seepage in one single quagfen complex, De Stobbenribben, rather
than an in-depth treatment of quagfen vegetation in the larger mire area of North-West Overijssel. So
much has been written and orally communicated about the seepage problem that only a careful
treatment of a key case could clear away some of the confusion left. This key case is introduced in
Chapter 7. This case is a de facto representative one, since De Stobbenribben is one of the very last
areas with a well-developed vegetation of seepage indicators in The Netherlands. In the Vechtplassen
area in the province of Utrecht no such extensive carpets of Scorpidium scorpioides have been left, and
almost all examples of quagfen vegetation nowadays include species indicating a more eutrophic
milieu.
The choice of methods and detail of the hydrological investigations were specifically tuned in to
the relevant ecological problems. While such an eco-hydrological approach has, in The Netherlands,
arrived at a mature stage in the application to agriculture (agro-hydrology), it is probably only in the
years of growing pains as regards natural `ecosystems' and the application to nature protection. As a
consequence, along with the results obtained in De Stobbenribben, quite some attention is paid to the
development of suitable methods (Chapters 8,9, Appendix D).
16
Introduction
In the course of the research, and upon comparison with results of other investigations carried out in
the mean time, it occurred to me that the importance of seepage, in North-West Overijssel usually as a
lateral inflow of slightly calcareous water, related to the quagfen vegetation through the base state of
the peat. The hydrological part of this relation is quantitatively studied in Chapter 10 and the results
seem to be in line with the apparent nutrient economy of the local stands of vegetation, but the base
state itself, and its influence upon the availability of nutrients, had to stay a lacuna in the data collected.
In sofar the results of this project indicate the importance of a more direct investigation of the base
state, and of the rle of exchange processes between the peat and the water, in future projects. Such
investigations are also demanded to explain the occurrence of seepage indicators as a result of other
mechanisms than seepage that can maintain a high base state in the uppermost horizons of the peat.
A detailed summary with a general discussion of the results obtained is provided in Chapter 11. In
order not to exceed some acceptable length for the main text of this thesis, certain background data
have been provided as appendices. Among these are an explanation of terms (Appendix A) and a
detailed treatment of the method developed for the interpretation of chemical analyses of natural waters
(The MAION method, Appendix D), which has found applications far beyond the scope of the present
report, but was never published in full before. Although I have tried to include the data necessary to
check my results throughout this thesis, no complete listing of all basic data is included. I am willing to
provide such data on an `as is' basis upon specified requests. Appendix F more-over lists data reports,
many of which resulted from student's projects that formed part of the present investigation. Several of
these reports have not been explicitly referenced in this thesis.
Introduction
17
Reprinted from: Geert van Wirdum, 1990. Vegetation and hydrology of floating rich-fens. Datawyse, Maastricht, 310 p. (ISBN
90-5291-045-6). Known errors corrected.
CHAPTER 2
Published evidence of seepage in rich-fen quagmires

(quagfens)
2.1 Introduction
The word seepage, as a translation of (Dutch) kwel, has often been used in ecological
descriptions of rich-fen quagmires (quagfens) in The Netherlands. This was stimulated by
the description of similar stands of vegetation in Central Europe by authors referring, in
German, to Quell as an important site factor. Quell, kwel, and seepage can be used in a
variety of technical and colloquial meanings. The meaning intended by Segal (1966), i.e.,
the oozing out of groundwater as a result of artesian pressure, is taken as a starting point
here in the presentation of the seepage hypothesis for quagfens that was widely accepted
by Dutch ecologists associating the occurrence of the plant species and types of
vegetation involved with a regional discharge of groundwater. It will be shown that this
supposition is only weakly supported by the pertaining literature. Nonetheless there are
several indications of seepage in a more general sense, i.e., of the percolation of
allochthonous water, in whatever direction, through the quagfen root zone. The supply of
calcium seems to play a vital rle here.
Seepage in quagfens
19
2.2 Definitions of seepage

According to Hooghart (1986) the Dutch word kwel is equivalent to the English seepage and
exfiltration. Two definitions are listed: (1), the outflow of groundwater, and, (2), more specifically, the
outflow of groundwater under the influence of a larger hydraulic head outside the seepage area, i.e., a
regional groundwater discharge. Hence, the intensity of seepage is defined as the volume of water
seeping out per unit of time and per unit of horizontal surface area of the region considered.
In The Netherlands, a regional groundwater discharge is associated with streamlines often reaching
depths of over one hundred metres, and travel times of hundreds or even thousands of years.
Accordingly, the discharging groundwater is usually saturated with calcium and (bi)carbonate ions, at a
carbon dioxide tension above the atmospheric one, and it is unpolluted, although it may have been
enriched with chloride and other ions from deeper strata.
At variance with these definitions, Thomson & Ingram (in Ivanov 1981, p.252) note, (3):
By the seepage or intensity of water exchange in a mire we understand the total quantity of water
flowing per unit of time through a volume of peat 1 m2 in area and equal in height to the depth of
the peat deposit in that part of the mire massif.
Such a quantification is well in line with current colloquial meanings of the intransitive verb to seep:
ooze out, percolate slowly (The Concise Oxford Dictionary of current English, sixth edition, 1976).
The Dutch kwel and German Quell are more specifically associated with the first of these meanings,
i.e., the oozing out or exudation of groundwater.
Next to Quell, Sickerung is often encountered in German texts relevant to the seepage hypothesis.
This term is more or less equivalent to seepage and percolation in their colloquial meaning. In the
hydrological literature Sickerung and percolation are mostly used to indicate the (downward)
infiltration of water through the unsaturated zone and the capillary fringe, recharging the body of
phreatic groundwater (Hooghart 1986, Brockhaus 1961). Throughout this text percolation is used in
the colloquial meaning as a synonym of seepage in the general sense (see below).
I will not use the word seepage when I explicitly mean a groundwater outflow or, more
specifically, a regional groundwater discharge, as in the kwel definitions (1) and (2), respectively.
Hence, seepage is used for percolation in general, regardless of the possible causes and any
preferential direction of water movement. The word is italicized in the designation of the seepage
hypothesis for quagfens and in associated expressions. Although the primary authors reporting seepage
in Dutch quagfens specifically envisaged a regional groundwater discharge, the non-specific term
seepage is maintained here in view of the eco-hydrological conditions disclosed in this investigation.
2.3 The seepage hypothesis for quagfens
De Wit (1951), Meijer & De Wit (1955), Kuiper & Kuiper (1958), and Segal (1966) have drawn the
attention to the occurrence in quagfens in The Netherlands of a vegetation cover supposed to be
associated with an exudation of groundwater (kwel). This supposed relation was based on (1) visual
indications of groundwater outflow in some of the quagfens involved and (2) on the similarity of the
floristic composition of the vegetation to phytosociological associations previously described from
Quellsmpfe in Central Europe. Accordingly, Westhoff & Den Held (1969) denote the
Scorpidio-Caricetum diandrae as a stage of terrestrialization in fen mires influenced by outflowing
groundwater. This also concerns the Scorpidio-Utricularietum, which is said to have its optimum in
depressions in mesotrophic quagfens, often alternating with the Scorpidio-Caricetum diandrae. This
20
Seepage in quagfens
statement implies that the presence of a Scorpidio-Caricetum diandrae and a Scorpidio-Utricularietum

in quagfens indicate an outflow of groundwater: the seepage hypothesis for quagfens.
Segal (1965, p.12; 1966, p.135) and Gonggrijp et al. (1981, p.73-79) leave but little doubt that the
seepage hypothesis as applied to quagfens in North-West Overijssel specifically refers to a regional
discharge of groundwater. This specific form of the hypothesis was widely accepted, and,
consequently, one has (vainly!) attempted to isolate quagmires which had been indicated as seepage
sites, so as to exclude the influence of (possibly) polluted surface water while maintaining the
presumed upward flow of groundwater through the subsoil into the mire.
2.4 The source of the hypothesis; main arguments and evidence
Phytosociology
With regard to the indicative significance of the vegetation cover, the seepage hypothesis as developed
in ecological studies in quagfens in The Netherlands was almost certainly inspired by the notion of
calcareous spring and seepage mires (Kalk-Quellsmpfe) in the Central European literature (cf.
Ellenberg 1978, p.421, 427-433). It is noteworthy that the Dutch authors dealing with seepage sites
based their phytosociological studies largely on the methods of the Zrich-Montpellier school, which
enabled them to compare their results more readily with those accumulated in the central European
countries. There were also extensive personal contacts and discussions with leading authorities of the
above-mentioned school. In his article of 1951 De Wit, who then worked in Montpellier, explicitly
mentions that Braun-Blanquet agreed with him as regards the feasibility of classing certain
communities with Carex diandra in the Caricion davallianae alliance, thus stressing the floristic
kinship of these communities with those characterized by Schoenus nigricans described from
calcareous, wet dune slacks in The Netherlands. The Scorpidio-Caricetum diandrae was first described
by Koch (1926, p.83) from Switzerland, who also noted character species shared with communities
dominated by Schoenus nigricans. With regard to the latter he remarks:
Den kologischen Ansprchen des dominierenden Schoenus nigricans gemss, stellt der Typus
der Assoziation hohe Anforderungen an den Kalkgehalt des Bodens. Ihr Vorkommen ist deshalb
beschrnkt auf den Rand sehr mineralreicher Gewsser, auf alte Seebecken mit Unterlage von
Seekreide, bei uns vor allem auf die Nhe sehr kalkhaltiger Quellen, sowohl in der Talebene, als an
den Hngen.
In view of these phytosociological relations, ecologists became interested in the question whether the
presence of Scorpidio-Caricetum diandrae stands in The Netherlands could be attributable to seepage
phenomena: the phytosociological scheme of classification indicated the direction of prospective
ecological studies (De Wit 1951, p.352). Since the rhizosphere of quagmire vegetation has no direct
contact with the mineral subsoil, the possible occurrence of a regional groundwater discharge was
emphasized to explain the presumed high calcidity in the environment.
Seepage in quagfens
21
Hydrology
Until the beginning of the presently reported survey (Van Wirdum 1973), further work with regard to
the seepage hypothesis had not been accompanied by any detailed geohydrological investigations. The
conventional argument in this field was the proximity of presumed seepage sites to more elevated areas
(De Wit 1951, p.346-347; Meijer & De Wit 1955, p.50; Kuiper & Kuiper 1958, p.361; Segal 1966,
p.110). In several case studies concerning the seepage hypothesis the visual presence of iron
compounds in seepage pools is mentioned, especially in Het Hol near Kortenhoef where the hypothesis
was first formulated. De Graaf (in Meijer & De Wit 1955, p.69) noted the presence of several active
iron wells and thoroughly documented the chemical composition of the water. A visible flow and
bubbling of the water in presumed seepage pools in quagmires is commonly mentioned. Such
phenomena, however, may well be associated with, (1), the forces caused by the weight of the
observers on the weak kragge, squeezing out water from below, and, (2), the escape of gases under the
influence of a lowering barometrical pressure. They obviously do not prove the incidence of seepage of
groundwater from elevated areas through the underlying mineral soil.
Ecophysiology
It was long realized that plants are not sensitive to seepage as such. In view of this awareness, Kuiper
& Kuiper (1958, p.363) mention two ecophysiological factors which may be the cause of seepage sites
bearing a characteristic association of plant species, namely, (1), a somewhat lower water temperature
during summer, and, (2), the binding of phosphate by the formation of insoluble iron compounds.
Segal (1966, p.135) also mentions, (3), the possible input of calcium and bicarbonate ions, and he
states that seepage may occur intermittently and may be surprisingly localized.
The significance of these factors in quagfens in The Netherlands is partially supported by
observations, but these observations relate to the phenomena as such, rather than to their physiological
effects. Segal (pers. comm. 1969) made continuous temperature recordings in De Stobbenribben and in
some other places. He found remarkable differences between shallow pools at a small distance from
one another. Within De Stobbenribben and other quagfens, seepage pools were thus supposed to exist
next to non-seepage pools. Using the same measuring equipment, I was unable reliably to reproduce
such results during 1969 and 1970 (see also Chapter 8).
The presence of iron is readily observed, both in the form of an oily film on the surface of the
water, and, in some cases, in the form of a rusty brown precipitate on macrophytes and the muddy
bottom. Especially the oily films are very common in many quagfens. The binding of phosphate seems
to have been proved by De Graaf (in Meijer & De Wit 1955, p.69) to occur in the Kortenhoef area.
As regards the supply of calcium and bicarbonate ions, Segal (1966, p.136) reports:
... the periodic fluctuations in the recorded environmental factors may be considerable ... In the
course of the succession in seepage areas the general trend is a gradual decrease in the specific
conductivity, in pH, in the hardness and bicarbonate concentration, and in the chloride and calcium
content, and an increase in the organic ammonium and phosphate, only the sulphate content
showing a peak in the Pellia phase, all this in spite of the fact that, generally speaking, percolated
water enriches the environment and normally causes a local increase in the specific conductivity,
and in the pH values, the hardness, and the chloride, bicarbonate, ammonium, iron and calcium
concentrations, concomitant with a decrease in the phosphate and the sulphate content.
22
Seepage in quagfens
Apparently, Segal compared the seepage environment with more oligotrophic environments which are
mainly fed by rain water. It is worthy of note that Kuiper & Kuiper apparently considered seepage sites
less eutrophic than other ones (p.363), especially with regard to phosphorus compounds. Segal hints at
an opposite trend in the nutrient and the base states, respectively, of seepage sites as compared with
other ones. Any systematic discussion of water analyses is lacking in the cited preliminary report,
however. De Graaf (in Meijer & De Wit 1955, p.70) pays some attention to the calcium and
bicarbonate concentrations in water samples from the area near Kortenhoef, but his report is not very
conclusive with regard to the specific relation of these factors with seepage or seepage sites. The
intermittent occurrence of seepage phenomena is not well documented either. There is even an element
of contradiction in the statements regarding the supposed significance of, (1), a constant temperature,
and, (2), a supposed intermittent activity of wells, resulting in considerable fluctuations of the
concentration of solutes in the water.
2.5 Indicator species reported in studies of Dutch quagfens
Although most authors seem to agree that certain plant associations are more characteristic of seepage
sites than individual plant species, some have ventured to list seepage indicators. Only some of these
species are unanimously considered so, at least when they occur in quagfens in The Netherlands (Table
2.1). Most of them are character species of the Caricion davallianae, the Scorpidio-Caricetum
diandrae, or the Scorpidio-Utricularietum, respectively. Since the above-mentioned associations are
the only ones of the Caricion davallianae present in quagfens in The Netherlands, and since the
Scorpidio-Utricularietum in quagfens is strongly associated with the Scorpidio-Caricetum diandrae,
which is seen as a seepage community par excellence, these species, when growing in quagfens, are
almost bound to be considered seepage indicators.
Liparis loeselii, although mentioned as a seepage indicator by Kuiper & Kuiper (1958, p.373), and
Carex diandra are, according to the same authors (p.395) not as characteristic of seepage sites as we
originally considered them. The taxonomic reliability of records of Philonotis fontana is doubtful. At
the sites from which this species was recorded by Kuiper & Kuiper I only found P. marchica, which
was not mentioned by Kuiper & Kuiper but is represented in a record by De Wit (1951, p.355) from
most probably one of the same sites. The presence of Riccardia chamedryfolia in quagfens in
North-West Overijssel has, until now, not been confirmed. I have seen several samples of R. multifida,
however, which had mistakenly been identified as R. chamedryfolia (compare also Mller 1954,
p.500). However, since less typical specimens, which cannot be easily identified, are frequently met
with, the occurrence of R. chamedryfolia cannot be excluded altogether (Van Wirdum 1983).
According to Kuiper & Kuiper (1958, p.395), Menyanthes trifoliata and Carex lasiocarpa are
abundant in the vegetation cover of transitional sites between seepage environments and more
eutrophic ones, but entirely absent from the more eutrophic environments proper.
Seepage in quagfens
23
Table 2.1
Seepage indicators in quagfens in The Netherlands
Spermatophyta
Calamagrostis stricta
Carex buxbaumii
Carex diandra
Carex lasiocarpa
Dactylorhiza incarnata
Eriophorum gracile
Liparis loeselii
Menyanthes trifoliata
Parnassia palustris
Sagina nodosa
Utricularia intermedia
K
W
*W
K
WC
KD
*KC
K
WC
KC
KU
Bryophyta
Bryum pseudotriquetrum
Campylium elodes
Campylium stellatum
Scorpidium cossoni
Drepanocladus revolvens
Fissidens adianthoides
Philonotis fontana
Riccardia chamedryfolia
Riccardia multifida
Aneura pinguis
KC
SC
KC
SC
WC
KC
*K
*KC
KC
K
KC
K: according to Kuiper & Kuiper 1958, p.373, 395; W: additions from Westhoff et al. 1971, p.80-81; S: additions from Segal
(pers. comm. 1969); C: character species Caricion davallianae; D: character species Scorpidio-Caricetum diandrae; U: character
species Scorpidio-Utricularietum; (Character species from Westhoff & Den Held 1969); *: questionable, see text
2.6 Special remarks with regard to bryophytes

As far as Angiosperms are concerned, the pertaining literature does not abound in references to
particular species as seepage indicators. The character species of seepage communities are usually
considered seepage indicators. Additional information, however, is being provided by the bryological
literature. This information is marred with confusion, as appears from the following, incomplete,
survey, restricted to the species mentioned in the aforegoing section.
Liverworts
Riccardia chamedryfolia, R. multifida: auf feuchten Boden in der Nhe von Quellen und
Wasserrinnen (Boros 1968). Aneura pinguis: associated with kalkhaltigen Quellen und anderen
feuchten Stellen (Mller 1954).
Mosses
(Note that Drepanocladus revolvens is synonymous with Scorpidium cossoni plus Scorpidium revolvens, as explained below.)
Mnkemeyer (1927) uses Quell-words in the description of the environments preferred by: Bryum
pseudotriquetrum, Fissidens adianthoides, and Philonotis fontana.
Boros (1968) reports a preference of Quell-environments for all mosses in Table 2.1, with the
exception of Scorpidium scorpioides (in Schlenken torfiger Seggenmoore) and Philonotis marchica
(but: an feuchten, berieselten, kalkhaltigen Stellen, besonders an Wassermhlen, gern an Thermen, auf
Kalktuff, seltener an Quellen).
Several of the seepage indicators among the bryophytes are being regarded as glacial relict species
in Central Europe, surviving in the usually cool environment of mires, especially in seepage mires.
Amann (1928, p.352), for example, enumerates, among other quagfen species, Drepanocladus spec.(D.
24
Seepage in quagfens
revolvens is given as an example in the text), Philonotis fontana, P. marchica, and Scorpidium
scorpioides as reliquats nordiques des marais.
Both Amann (p.89-100) and Boros systematically treat the relation of the occurrence of bryophytes
with the calcium content. Philonotis marchica, which Boros associates with lime (see the earlier
quotation), is considered to be a tolerant, yet calcifugous species by Amann, as is P. fontana, but the
latter species is also taken for a calcifugous one by Boros (p.337: in kalkmeidenden Quellfluren).
Drepanocladus revolvens occurs an kalkhaltigen, quelligen Stellen according to Boros (p.368), but is
mentioned among the espces calcifuges plus ou moins tolrantes by Amann.
These are just a few examples of slightly different to rather opposed opinions, which can partially be
attributed to the different geograpical areas considered by Boros (Hungary) and Amann (Switzerland),
respectively. A similar divergence of opinions is met in the literature concerning the types of
vegetation in seepage environments, as will be shown in the following section.
The habitat descriptions by Touw & Rubers (1989) probably reflect the wide acceptance of the seepage
hypothesis in The Netherlands, rather than representing independent information. Although kwel is not
regarded a strict requirement for any of the moss species considered here, it is mentioned for all but
Bryum pseudotriquetrum and Fissidens adianthoides, as is a preference, or tolerance at least, of
calcareous environments.
Drepanocladus revolvens ssp. revolvens and D.r. intermedius are considered synonymous by Touw
& Rubers, who place the species in the genus Scorpidium as S. revolvens. Hedens (1989) includes
Drepanocladus revolvens in the genus Scorpidium, and he ranks the sub-units D. r. revolvens and D. r.
intermedius at the species level as S. revolvens and S. cossoni, respectively. Dutch quagfen samples
certainly belong to the latter. It is worthy of note that, whatever the taxonomic rank, Hedens' samples
of S. cossoni from Southern Sweden mostly originated from calcium-rich areas apparently avoided by
S. revolvens.
2.7 Seepage and calcidity
It is obvious that the singular character of seepage quagfen sites, as conceived by Dutch ecologists, is
not only distinct from that of more oligotrophic, rain-fed bog or boggy transitional sites, but also from
that of the majority of more eutrophic fens. A survey of the pertaining literature confirms that most
Central-European authors regard seepage as one way to invoke and preserve particular environmental
conditions, rather than as a direct cause of the ocurrence of seepage indicators (cf Schmidt 1969,
p.245). The various types of seepage sites distinguished in Central Europe, each accompanied by
characteristic types of vegetation are listed in Table 2.2.
Ellenberg comments that the Kleinseggenrieder der Quellsmpfe owe their preservation to human
influences, in casu, mowing. According to Ellenberg this type of vegetation closely resembles that of
other kinds of fen not characteristic of seepage. In the absence of mowing the cover of vegetation
would become overgrown by Phragmites australis or replaced by willow and alder carr.
Seepage in quagfens
25
Table 2.2
Types of seepage sites and their characteristic vegetation

kalkarm
kalkreich
berrieselte
Quellflur
Weichwasser-Quellflure
Cardamino-Montion
Quelltuff-Fluren
Cratoneurion-commutati
durchfeuchteter
Quellsumpf
saure Kleinseggenrieder
Caricetalia nigrae
Caricion canescenti-nigrae
Kalk-Kleinseggenrieder
Tofieldietalia
Caricion davallianae
Cited from Ellenberg 1978, p. 421
Seepage is usually considered a prime factor capable of preserving a calcareous, yet oligotrophic fen
environment (cf Braun 1968, p.8). Westhoff & Den Held (1969, p.203), on the other hand, note a
relevant divergence in the floristic composition of the Caricion davallianae:
With regard to The Netherlands, the Caricion davallianae can probably be divided into two
subunits reflecting the fluctuation of the groundwater level and the nutrient status, especially the
calcium content of the environment. We would thus obtain a syntaxon comprising the associations
of environments with a constantly high groundwater table and usually low calcium contents in the
groundwater, and a syntaxon comprising the associations on mostly calcareous soils with a
fluctuating groundwater level.
There is no doubt that the Scorpidio-Caricetum diandrae and the Scorpidio-Utricularietum would have
to be placed in the first-mentioned syntaxon. The diagnosis comes close to the one given by Rybn ek
(1974) for the Scorpidio-Utricularietum.
Rybn ek distinguishes the Caricion demissae Rybn ek 1964 as a separate alliance in the order
Tofieldietalia and classifies the Scorpidio-Utricularietum association in this alliance. The presence of
certain acidophilous species, such as Drosera rotundifolia, Oxycoccuspalustris, and Sphagnum
contortum, is mentioned as a typical feature of this alliance which encompasses meistens montane
Parallelassoziationen des Caricion davallianae Klika 1934 in kalkarmen oder Silikatgebieten.
Although the three above-mentioned acidophilous species do occur in close proximity of, or even
within, stands of seepage-fen vegetation in Dutch quagmires, Westhoff & Den Held (1969, p.203) do
not accept the Caricion demissae for The Netherlands. Rybn ek's (1974, p.34) comment about the
influence of base ions in both alliances is probably also relevant to the subunits suggested by Westhoff
& Den Held, however:
Fr beide Verbnde (Caricion demissae, Caricion davallianae) ist aber die hhere Gehalt an
basischen Ionen im Grundwasser und die Hhe des Sattigungsgrades des Sorptionskomplexes der
Torfsubstrate kennzeichnend. In unserem Gebiet der Silikatgesteine der Bhmisch-Mhrischen
Hhe sind die Gesellschaften dieses Verbandes dort vorhanden, wo eine Mglichkeit des
ununterbrochenen Nachsttigen des Sorptionskomplexes mit einem an basischen Ionen reichen
Grundwasser besteht, d.h. meistens auf den Hangquellmooren.
26
Seepage in quagfens
2.8 Conclusions and re-formulation of the seepage hypothesis

Although the pertaining literature emphasizes the importance of seepage phenomena for the type of
vegetation studied, there is insufficient evidence for the attribution of seepage to a regional
groundwater discharge, as stated in the seepage hypothesis for Dutch quagfens. The rejection of the
original seepage hypothesis may lead to a re-formulation based on the literature cited in this chapter.
This is still a seepage hypothesis, but seepage is used here in the wider sense of percolation, rather than
of an outflow of groundwater specifically.
The re-formulated seepage hypothesis considers that the seepage sites under discussion derive their
very characteristic vegetation cover from the interaction of the substratum, the atmospheric
precipitation, and percolating water, as modified by the microrelief. In some cases there may be an
influence of a regional groundwater discharge, or of the exudation of groundwater along a slope, while
in other instances surface water from nearby canals, rivers, or lakes may seep through the fens
involved. The sites are characterized by a relatively high activity of calcium ions. In the long run a
somewhat calcareous type of seeping water is required in order to prevent a succession towards a more
acidophilic type of vegetation.
The precise species composition will, among other things, depend on the chemical composition and
the intensity of the atmospheric precipitation and the seepage, respectively, and on the microrelief. The
higher soil strata, and especially the hummocks, will show more acid conditions, both as regards the
soil and the interstitial water, while the lower strata, including terrain depressions, are more calcareous,
especially as regards the water.
Although the original seepage hypothesis cannot be maintained as a general one, the survey of the
literature does not allow for a rejection of the hypothesis that a discharge of groundwater occurs in
quagfens in The Netherlands or in North-West Overijssel in particular. This point is therefore included
in the present investigation. The ensuing questions can be formulated as follows:
1) Are the quagfens that are taken for seepage sites usually restricted to the outflow of groundwater
from the mineral subsoil into the overlying mire? Obviously, not only the present conditions, but
also the historical situation must be taken into account.
2) If this is not the case, which input terms are the most significant in the water balance of the
quagmires involved?
3) Are the calcium and bicarbonate contents possibly more important in such quagmires than they are
in other ones, and can this be explained in view of the water balances?
Seepage in quagfens
27
Reprinted from: Geert van Wirdum, 1990. Vegetation and hydrology of floating rich-fens. Datawyse, Maastricht, 310 p.
(ISBN 90-5291-045-6). Known errors corrected.
CHAPTER 3
Site properties indicated by the flora of quagfens
3.1 Introduction
The seepage hypothesis for quagfens analysed in the aforegoing chapter was to a
large extent based on the presence of seepage indicators in the local flora. The
occurrence of the species involved might be explained by the presence of an
appreciable amount of calcium and bicarbonate ions. In the following the
significance of indicator species in North-West Overijssel quagfens will be further
explored. The problem of ecological indication is first discussed from a theoretical
point of view. It is argued that the relation between the presence of individual plant
species and measured environmental factors is rather indirect. The concepts of
plant Associations, from phytosociology, and mire complexes, from mire ecology,
seem to fit the scales of resolution of field observations better than do the
individual species with their physiological requirements. These concepts do not
easily allow, however, for the recognition of delimited natural entities in the study
area. Different levels of environmental variety are proposed as a basis for the
understanding of the patterns of species aggregation in quagfens. A statistical
survey of the indicator species of seepage sites suggests that these are low-lying,
base-rich, yet nutrient-poor sites where fen peat has developed under the continued
influence of an external water supply.
Site indication by quagfen flora
29
3.2 Species and associations of species as indicators

Response models and Associations in phytosociology
The relation between the presence of plant species and environmental factors that can be
assessed with physical or chemical methods has led to the formulation of so-called response
models. Most response models rely on empirical data in a single geographic region and they are
not necessarily valid outside that region. The indicator lists used here illustrate a factor-wise
indication of the optimum or the range for each species, derived from empirical data.
A site is usually considered homogeneous in response models: all species in the local stand
of vegetation are supposed to respond to the very same environmental state, thus representing a
logical-AND Association as distinguished in the French-Swiss School of phytosociology (Van
Wirdum 1987). The milieu in the sites of the stands that belong to such an Association is, by
definition (Int. Bot. Congress 1910, Brussels), uniform and it must be considered a realization of
the intersection (logical product) of the ecological ranges of the various species present
(Fig.3.1a). An Association is itself conceived as the product of an evolution of biological
adaptations to the coexistence of species in such realizations (Ellenberg 1978 and a discussion
about this subject in Westhoff & Den Held 1969). The implied individuality at the Association
level has become the central theme of phytosociology.
The individuality of Associations enables one to determine the completeness or degree of
saturation of a given stand of vegetation. Since any plant individual requires some physical
space, and since the frequency and abundance in the stands of an Association differ among the
various species, a certain minimum area is required for each Association to be represented by its
complete species assembly, even under ideal environmental conditions. In practice personal
judgement cannot be avoided in the assessment of the environmental uniformity of sites. This
has led to different plot sizes for phytosociological field work in mire vegetation (Table 3.1).
Table 3.1
Plot sizes used in phytosociological field work in fen mires
Dierssen 1982
Den Held & Den Held 1973 (p.10)
Wheeler 1980
Mueller-Dombois & Ellenberg 1974 (p.48)
Braun 1968
1
4-10
10
10-25
25-100
m2
m2
m2
m2
m2
Dierssen (p.10-11) comments that mostly a plot size of 1 m2 is quite satisfactory in

peat-mires, with the exception of reed fens, which require 4-6 m2 for a representative sampling.
In species-rich calcareous fens with a considerable micro-relief, Dierssen was probably unable
to meet the minimum area requirement but he used 1 m2 plots for reasons of homogeneity. Some
of the divergence among various schemes of classification (Appendix B) is attributable to the
implicit use of different concepts of uniformity. It is relevant to the problem of hydrological
ecology of quagfens that Dierssen distinguishes Subassociations with Scorpidium scorpioides
30
Fig.3.1
The concept of AND and OR Associations

According to the OR theorem (B) the milieu of an Association is the union of the milieus of the various
associated species rather than their overlap area, as prescribed by the AND theorem (A); hypothetical
example.
in seventeen Associations in five Alliances and three Orders, while Westhoff & Den Held (1969,
p.202) mention this species as a character species of the Order Tofieldietalia only and of the
Alliance Caricion davallianae within that single Order. Dierssen obviously recognized
environmental differences relevant to the moss layer and found the same moss micro-coenon
associated with different vascular plants. The evidence from the work of other authors, using
larger plot sizes, is that these different vascular plants are also associated among each other.
This problem is partially overcome by the introduction of the concept of different synusiae,
or micro-coena, thus allowing for the difference in scale at which the various groups of plants,
according to their growth forms and life strategies, explore the available room
(Mueller-Dombois & Ellenberg 1974). This principle allows for different micro-habitats within
the French-Swiss uniform site of any stand of an Association. Similar moss synusiae can be
found within stands of different Associations of vascular plants, but the reverse may also occur.
So far no generally accepted approach to the problems of classification in this situation
seems to have been proposed (Segal 1968).
31
The OR assumption: an additional explanation of species association

The question arises whether the AND assumption is of exclusive validity, or if the association of
species may be partially due to recurrent patterns of environmental variety within the sites of
plant Associations. If the second alternative obtains, the relatively large environmental
amplitude of individual species, as inferred from vegetation studies, may be due to
environmentally inhomogeneous sites. Such sites may nevertheless represent a uniform habitat
of an Association when the pattern of environmental variety is a recurrent one, comprising
different micro-habitats. This will be called the OR assumption here, since it implies the
application of the logical OR proposition (union, logical sum, Fig.3.1b; Van Wirdum 1986,
1987). According to this explanation the recurrent environmental pattern, rather than species
interaction, is the primary cause of different species being found together. Conceivably the AND
and OR assumptions explain different aspects of the association of plant species in nature.
OR-type association emphasizes the individuality of species rather than that of whole
communities.
The individuality of mire elements
Scandinavian and Estonian authors in particular (Eurola et al. 1984, Masing 1984, Sjrs 1983,
see also Dierssen 1982) have stressed the individuality of mire complexes, rather than of plant
communities. A mire complex of a particular type is characterized by a recurrent pattern of mire
elements, such as hummocks and hollows, the various levels of aggregation each contributing to
the emerging local milieus of plant individuals. This concept is supplemented by a theory based
on the availability of nutrients (Ruuhijrvi 1983). The applicability to Dutch mires is hampered
by the fact that the net result in terms of the operational plant environment has not, or only
weakly, been defined. The principle of mire individuality is especially useful in situations where
natural processes are orderly expressed in the ecology, extent and morphology of mires. The
relatively young quagfens in The Netherlands have developed within the confined space of
petgaten in abandoned turbaries, and they have been subject to a variety of management regimes
and hydrological influences, even within a few decades. For this reason they cannot be
considered to represent steady states that primarily reflect the natural order. Under these
conditions species may become associated, or dissociated, without obeying the empirical rules
established elsewhere.
3.3 Different scales of aggregation
Although I do not reject the individuality concept at the mire complex and plant Association
levels, individual plant species, rather than complexes and Associations, are taken here as a basis
for the study of ecological relations between the vegetation cover and the hydrology of
quagfens. In doing so it must be accepted that the physiologically operational milieu of the
plants is usually not measured and, thus, stays ill-defined. This is not considered a significant
shortcoming in view of the fact that, for most quagfen species, the physiological requirements
are unknown anyway.
The following site aggregation model, based on the areal extent of the aggregates (synusiae),
is proposed for the understanding of species aggregation in the quagfens under study (Fig.3.2):
32
Fig.3.2
Different scales of environmental homogeneity in a quagfen

The picture shows a kragge adapted to a certain type of water supplied from below. Within the kragge
three hydro-environmental zones have developed according to the decreased influence of flooding and
the increased influence of rain water as the distance from the body of open water increases. Hummocks
and hollows are present in each zone. Micro-zonation is not shown.
Level 1 (Kragge synusiae): The quagfen kragge in a former petgat, 5-30 x103 m2, is ecologically characterized by the
magnitude of various terms in the water balance, especially the inflow of water from other sources than local
precipitation. It develops during the terrestrialization of the petgat from open water to, sometimes within a period of
some decades, initial bog or carr vegetation. Some species with a long lifetime may still be present when the environment
is no longer suitable for their renewed settling, as is obviously the case for such notorious rhizome builders as Phragmites
australis, Typha angustifolia, Equisetum fluviatile, Cladium mariscus, Menyanthes trifoliata, Nymphaea alba, and
various species of Carex. Their shoots can be considered deciduous parts of few long-lived and extremely extensive
individuals. The rhizomes are supplied with nutrients from the body of water underneath the kragge. The chemical
composition in this body of water is rather constant in some cases, but in other ones it varies considerably, both within
and between years. The kragge is in fact the most important storage organ of a quagfen and the extent of the rhizomes
of several species make it a more or less homogeneous base upon which other patterns become superimposed (levels
2-4). The kragge-forming plants contribute to a translocation of elements from the environment underneath the kragge to
the fen surface, and they provide shade, micro-relief, and substratum for smaller plants. Depending on the hydrological
situation large groups of more or less similar kragges may cover 103 to even well over 105 m2. Vegetation maps at scales
below 1/10 000 usually rely on kragge synusiae only (see also Chapter 6).
Level 2 (Hydro-environmental zones): As the kragge becomes thicker the influence of an external water supply
decreases, especially in the more isolated parts of a quagfen. This decrease is most obvious near the surface of the
kragge. With the exception of sites with discharging groundwater (Chapter 2), any external water is supplied along one or
several sides of a quagfen. Along its route into the fen a gradient develops in which some arbitrarily delimited zones,
each about 0.1-2 x103 m2, can be distinguished. In the most isolated parts patches of rainwater-fed, low-herbaceous and
moss vegetation can be assessed with a plot size of about 10 m 2. The various hydro-environmental zones can be
botanically characterized by the frequency and extent of such patches and by the abundance of the relevant indicator
species. An example is provided in the case study of De Stobbenribben (Chapters 7-10).
33
34
Fig.3.3
Hummock-and-hollow mosaic in quagfen vegetation
October 10, 1969, De Wobberibben; complex of low hummocks and shallow hollows. Sixteen additional species were not mapped. In
the course of the 1980s the moss cover at this precise location was almost entirely replaced by Sphagnum flexuosum, S.palustre,
and S.papillosum.
Level 3 (Hummock-and-hollow mosaic): The origin of hummock-and-hollow patterns is diverse. In the quagfens under
study the hummocks are mostly formed by the growth of such species as Carex paniculata, C.elata, Molinia caerulea,
and Sphagnum spec. div., while the hollows often originate from the removal of trees and shrubs of Alnus glutinosa,
Betula pubescens, Populus tremula, Salix cinerea, Salix aurita, and Myrica gale. Although hummocks may be a starting
point for the formation of larger patches characteristic of the level-2 pattern, the hummock-and-hollow mosaic is
essentially superimposed on that pattern. The typical extent of hummocks and hollows is 0.1-1 m2. Especially Characeae,
bryophytes, and such species as Utricularia spec. div., Drosera rotundifolia, Liparis loeselii, Vaccinium oxycoccus,
Valerianadioica, and Cirsium palustre are associated with this pattern. The species composition may differ according to
the hydro-environmental zones (level 2). Hummock-and-hollow patterns are illustrated in Fig.3.3.
Level 4 (Micro-zonation): Even within the hummock-and-hollow mosaic, obvious micro-zonations as regards the species
composition can be seen, both around the centres of the hummocks and the hollows and along the living or decaying
stems of larger plants. Especially mosses and several liverworts, such as species of Pellia, Aneura, Riccardia,
Cephalozia, and Cephaloziella are involved. The extent of each zone is measured in cm2. It may provide a suitable
environment for the germination of various plant species. An example of micro-zonation is provided by Fig.3.4.
The third and fourth levels of aggregation are not studied in any detail in this report, since the
patterns involved are largely inherent to the existence of aggregates appropriate to the other,
higher levels. In Chapters 6-10 it is attempted to relate the vegetational patterns at these higher
levels to hydrological factors. In the remaining part of the present chapter some representative
systems of botanical indication are analysed in more detail in order to provide a background for
the other chapters.
Fig.3.4
Example of micro-zonation in quagfen vegetation

August 6, 1970, Boonspolder; centre of a low hummock of Carex elata. The surrounding moss
vegetation was dominated by Calliergonella cuspidata and Campylium stellatum.
35
3.4 The compilation of a list of indicator species (Appendix C)

From reports by Kuiper & Kuiper (1958), Kuiper & Lapr (1956), Segal (1966), De Wit (1951),
and Bergmans (1975), an indicators list was compiled comprising all species recorded from
supposed seepage sites in North-West Overijssel. The list was later extended with most other
fen-mire species of The Netherlands. The list is used here to see what specific information about
the environment can be borrowed from the occurrence of seepage indicators as compared to the
majority of other fen species. In the course of time the list was updated as more recent data
became available. All tables and diagrams in the present publication are based on the state of the
list and the selection of attributes given in Appendix C.
Phytosociological indications from Westhoff & Den Held (1969), Ellenberg (1978),
Oberdorfer (1979), Dierssen (1982), Rybn ek (1974, 1984), Baltov-Tul kov (1972), Braun
(1968), Wheeler (1980), and Zijlstra (1981) were compared. The species could be arranged in
seven groups as treated in the next section.
Ecological indications from Ellenberg (1978), Kruijne, De Vries & Mooij (1967), Eurola et
al.(1984), Landwehr (1966, 1980), Pietsch (1982), and Boros (1968), were compared and
compiled into two different schemes: one according to acidity and productivity, and the other
one according to base state, mire water level, and inherent and supplementary nutrient effects.
These schemes reflect the Central European and Finnish traditions, respectively.
From the point of view of nature protection it is important to know whether the sort of
environment a certain species requires is threatened by certain human activities. Such species
will be called threatened species here, although several of them do not yet directly face
extinction. The qualification threatened or not threatened, at the national level (The
Netherlands), was originally based on data from Londo (1975), Van der Meijden et al. (1983)
and Margadant & During (1982), but the data for Pteridophyta and Spermatophyta were
replaced by those reported in the Red Data List by Weeda et al. (1990). All bryophytes marked !
or # in column T, and all species marked 0-5 in column R are considered in the group of
'threatened or 'Red-List species here.
The phytosociological groups
The basis of the phytosociological grouping is formed by Zijlstra's (1981) species groups. For
the present purpose, some of the groups were joined, and they were provided with short names
for easy reference. The assignment of species not listed by Zijlstra was derived from the
phytosociological literature mentioned above. The following groups were distinguished:
BOG species, including Oxycocco-Sphagnetea and Nardo-Callunetea indicators,
characteristic of extremely nutrient-poor conditions;
FEN species, including Parvocaricetea and Caricion fuscae indicators, characteristic of
oligo-mesotrophic mire. Species with an optimum in Associations characteristic of base-rich
sites are not comprised in this group;
LASFEN species, coinciding with Zijlstra's Caricion lasiocarpae indicators. Zijlstra classes
the Caricion lasiocarpae under the Caricion fuscae. At the Order and Class levels, it is thus
separated from the Caricion davallianae (see DAVFEN below). The LASFEN group includes
species which have been associated with seepage phenomena in The Netherlands, such as
Scorpidium scorpioides, but it excludes the DAVFEN species often thriving in a slightly dryer,
base-rich environment;
DAVFEN species, comprising Zijlstra's Caricion davallianae indicators and species of her
combined Caricion lasiocarpae and Caricion davallianae group, namely: Parnassia palustris,
36
Campylium stellatum, Fissidens adianthoides, Scorpidium cossoni + S.revolvens, Dactylorhiza

incarnata, and Drepanocladus lycopodioides;
MOLFEN species, including species characteristic of Cirsio dissecti-Molinietum, Molinion,
Juncion acutiflori or Calthion, Molinietalia, Molinio-Arrhenateretea, and Arrhenateretalia
vegetation. The group is associated with mown and grazed, but undunged sites;
LITFEN (litter fen) species, including the Filipendulo-Petasition or Soncho-Euphorbion
palustris indicators in Zijlstra's table and several species of her Agropyro-Rumicion crispi
group, and species indicating disturbance. This group combines species characteristic of
Associations of sites with a periodically fast turn-over of nutrients;
SMP (swamp) species, including Zijlstra's Phragmitetea indicators and various indicators of
early successional stages.
At later occasions some newly added species were marked AQU (aquatic species) or SMA
(salt marsh species characteristic of more or less brackish fens). This concerns few species, and
the categories are not separately dealt with here1.
According to the phytosociological literature (especially Ellenberg 1978, Dierssen 1982) the
following ecological relations may be expected between the phytosociological groups:
- When a mire is strongly influenced by surface water, including inundations, it is often
dominated by SMP species;
- When, locally, the influence of rain water and the processes in the root zone become
important enough to change the environmental conditions appreciably, FEN species are
more frequently encountered;
- When, in such places, the acidity is not much increased, LASFEN species constitute a
considerable part of the flora;
- In more or less calcareous environments DAVFEN species join the other ones, especially
when the sites are somewhat drier in the summer and used for hay-making. The same
situation may locally obtain on hummocks in generally wetter mire parts;
- A slight further drainage will lead to MOLFEN sites, especially under a strict mowing
regime, when the uppermost soil horizons start to become poorer in bases;
- The natural continuation of an increasing rain-water influence under water-saturated
conditions will cause a development into BOG. This development is most pronounced in the
case of FEN sites;
- When the vegetation is in direct contact with slightly eutrophic or fluctuating surface water,
or when it is burned, mown, or grazed at irregular intervals, or slightly manured or disturbed
in any other way, it may progressively obtain a LITFEN character.
Ecological indications according to the Central-European tradition
A first scheme of ecological indications was based on the indicator lists published by Ellenberg
(1974, 1978). An analysis of the frequency distribution (Fig.3.5) of 215 vascular plants present
both in my species list and in Ellenberg's revealed that not much information could be expected
from Ellenberg's light, temperature, and continentality factors. In the case of the moisture factor
(F-figure), an accumulation of species can be noticed in the wettest classes covering
permanently moist, wet, frequently inundated, and aqauatic sites (F8-F12). Many remaining
species are indifferent with regard to either of the factors soil reaction (pH), and nitrogen.
They are treated as a rest group (DIV) in Table 3.2.
37
Fig.3.5
Frequency distribution of 212 quagfen species according to Ellenberg's indicator

scales
The x-axis reflects an ordinal arrangement of classes from low (1) to high (9, in case of moisture: 9-12)
values of the appropriate factor. The correspondence between Ellenberg's alkalinity and nitrogen
indications and the water type and nutrient state classes, respectively, as introduced here is also shown.
The class X includes indifferent species.
With regard to soil reaction (R-figure), 86 species are indifferent, while the remaining ones are
distributed in a somewhat bimodal manner:
A, species of mostly acid sites (Ellenberg: R-figure 1-4);
B, species with an optimum in circumneutral or alkaline sites (R-figure 6-9);
In the quagfen environment the soil reaction and base state are determined by the dominant
water type in the hydro-environmental zones.
The frequency distribution in relation to the inorganic nitrogen (N-figure) supply does not show
any clear pattern, although there is a strong tendency towards a preference of poorer sites. Since
the nitrogen figure is well correlated with the preference of species to phosphorus and potassium
as listed by Kruijne, De Vries & Mooij (1967) (Van Wirdum & Van Dam 1984), it is regarded
here as an indication of the general nutrient state. The species with N=7-9 , when occurring in
38
mires, are mostly indicative of early successional stages of only marginal interest in the present
study.
Some vascular plants, and all bryophytes, are lacking in Ellenberg's list. Even if only five groups
are formed on the basis of the frequency distributions, the numbers of species in each class are
too small to draw reliable conclusions. The cited literature was studied in order to estimate the
missing indications, to replace certain less representative values, and to further reduce the
number of neutrals. Even a species with a conceivably wide amplitude may be indicative of an
ecological tendency when it is reported to avoid one of the extreme types. Since the list is
intended for use in wet mire environments only, the possibility of a different behaviour of
species in other environments was disregarded. The following classes were used:
1) Water type (reaction):
- (UNK): Unknown or indifferent, considered undefined;
- (ATM): Atmotrophic; Base state low, acid, usually ombrotrophic environments (Ellenberg:
R 1-4);
- (CIR): Circumneutral; Intermediary base state, mostly weakly acid sites (Ellenberg: R
(5-)6-7);
- (LTH): Lithotrophic; High base state, often calcareous, slightly or strongly alkaline sites,
pronounced influence of groundwater or surface water (Ellenberg: R 8-9);
2)
-
Nutrient state, especially with regard to productivity:

(UNK): Unknown or indifferent, considered undefined;
(OLI): Oligotrophic (Ellenberg: N 1-3);
(MES): Mesotrophic (Ellenberg: N (4-)5-6);
(EUT): Eutrophic (Ellenberg: N 7-9).
The distribution of the 308 species over these classes is given in Table 3.2 and discussed later in
this chapter.
The Finnish mire types
Since the beginning of this century students of Finnish mires have developed and applied a
system of mire types. Details concerning the most recent version of this system have been
published by Eurola et al. (1984), who claim that the system can be used in an extensive
geographical area. The various ecological types recognized in this system are inferred from the
species composition.
The theoretical basis is formed by a consideration of the ease with which a plant, according
to its physiological properties, is able to obtain nutrients at a site with a particular wetness,
nutrient and base state, and inherent and supplementary nutrient effects.
According to the definition of the classes one may expect rather strong correlations between
the various factors. This was investigated for 255 species in an early version of the indicators
list, of which 145, including both vascular plants and bryophytes, are also mentioned by Eurola
et al. (1984). Owing to the fact that these authors often assigned a species to more than one class
of a single factor, thus in fact creating new, combined classes, their indicative system allows for
more than 5000 combinations. A close analysis of the distribution of the North-West Overijssel
39
Table 3.2
The distribution of species in the systems of indicators
Species group
Nr of Species
ALL
SEP
RED
BOG
DAV
LAS
FEN
MOL
LIT
SMP
DIV
308
21
78
31
16
18
41
46
67
73
16
Water type:
ATM
CIR
LTH
UNK
71
110
60
67
25
20
25
8
24
6
12
3
1
8
9
2
2
10
2
20
13
3
5
10
23
4
9
12
25
13
17
2
36
18
17
3
3
10
Nutrient state:
OLI
MES
EUT
UNK
113
104
61
30
16
5
52
23
24
3
11
5
10
8
26
18
19
19
1
2
17
20
20
10
6
27
33
7
4
7
5
Co-occurrence of indications according to the Central-European tradition (see Fig.3.6):

LTHOLI
20
9
16
8
5
1
LTHMES
24
3
9
2
4
2
3
ATMOLI
51
20
21
1
11
9
CIROLI
31
6
23
1
5
6
12
REST
182
3
20
10
4
4
22
21
4
4
8
5
46
2
7
1
2
61
14
33
11
18
4
3
1
1
3
Finnish system:
Nr of Species
Base state:
OMB
POR
TRL
WMS
XRC
162
16
48
19
22
9
59
57
15
10
3
8
17
10
14
2
2
8
4
41
45
76
4
12
9
14
25
Supplementary nutrient effects:

SEP
36
FLD
68
ANY
33
NON
25
3
7
4
2
Inherent nutrient effects:

MOS
16
NVA
17
RFN
26
NON
103
4
9
3
Groundwater level:
HUM
INT
FLK
15
13
35
17
26
4
8
16
7
5
9
1
15
9
1
2
5
7
3
8
2
13
4
2
2
7
6
2
11
10
8
17
4
9
4
10
8
8
2
6
25
14
13
9
12
5
1
1
12
8
4
3
1
6
5
1
4
22
5
4
5
6
4
2
6
10
8
2
6
21
6
6
10
17
15
9
5
1
4
3
7
3
4
8
2
21
1
7
7
5
10
3
21
4
1
1
2
1
1
32
RED: Red-list species and threatened bryophytes; SEP: Seepage indicators; DIV: SMA + AQU;
other abbreviations: see text
40
species led to a reduction to a four-dimensional 5x4x3x4 representation. Of these 240

combinations only 47 were indeed occupied by species. The following definitions were used to
revise the classification of the species represented in the original scheme published by Eurola et
al.:
1)Base state, reflecting the sum total of physical and chemical growth factors of the peat,
including the influence of climate, pH, electrolyte content, the amount of individual nutrients,
peat thickness, water movement, and the height of the water table. Note that, although the term
nutrient state was used by Eurola, the calcium content and pH are key factors in the actual
estimation, justifying the name base state (see below). The naming of the factor states below
reflects the distinction of poor, transitional, rich (here named wide-range mesotrophic), and
extremely rich fen by Scandinavian authors. The following classes are distinguished:
- (OMB): Ombrotrophic (relying on rain water). The species which, in the original Finnish
system, are listed in the ombrotrophic class, as well as those exclusively mentioned for the
oligotrophic class, were all grouped into this new ombro class;
- (POR): Base-poor (with a low base state). This class holds the species marked both in the
oligotrophic and in the mesotrophic class by Eurola et al.
- (TRL): Transitional. This class is identical to Eurola's mesotrophic-only group. As explained
below, it appears that this class holds a number of species which are considered indicators of
an unstable environment in The Netherlands, as well as species characteristic of so-called
transitional fen;
- (WMS): Wide-range mesotrophic (with an intermediate or moderately high base state). This
class holds the species marked both in the mesotrophic and in the eutrophic class of the
original scheme;
- (XRC): Extremely rich (with a definitely high base state). Identical with Eurola's class
eutrophic-only;
2) Groundwater level. The groundwater level in mires is especially important with regard to the
adaptation of plants to anaerobic conditions. Plants which are adapted to anaerobic conditions
are supposed to have a greater difficulty in obtaining nutrients, and thus generally indicate a
higher nutrient state than species of aerobic sites. The following levels are distinguished:
- (HUM): Hummock level (water level lying, in a normal summer, more than 20 cm below the
site surface level). The class includes some species also occurring in the intermediate level
class of the Finnish scheme;
- (INT): Intermediate level (water level lying, in a normal summer, only 5-20 cm below the
site surface level). Identical to Eurola's intermediate level;
- (FLK): Flark level (water level lying, in a normal summer, above or only a few centimetres
below the site surface level). The class includes some species that were also scored in the
intermediate level by Eurola;
3) Supplementary nutrient effect, defined in accordance with the question whether or not the
surface peat is continually being supplied with additional nutrients from other sources than rain
water. The stronger this influence, the more eutrophic the site will usually be. According to an
increasing effect on the availability of nutrients for plant growth, the following supplementary
nutrient effects are distinguished:
- (SEP): Seepage influence (moving groundwater with a more or less constant temperature,
and a relatively high oxygen content). This influence is called groundwater influence in the
Finnish system. The authors indicate a type of groundwater that is relatively rich in oxygen
which is, in The Netherlands, not typical of discharging groundwater. I have, for this reason,
41
chosen the term seepage influence, in the general sense of seepage. The spruce mire
effect tabulated in the Finnish scheme is included here;
(FLD): Flooding influence (fluctuating surface water, in particular derived from streams,
rivers, and lakes). This influence is named surface water influence by Eurola et al., and
renamed flooding influence here in order to distinguish it from the influence of surface water
that reaches the mire sites by seeping through the peat. The latter is more properly included
in the seepage influence;
(ANY): Any supplementary nutrient effect. This class holds all combinations of supplementary nutrient effects;
(NON): None: no supplementary nutrient effects were mentioned for the species in this
class;
4) An inherent nutrient supply, produced by the autochthonous nutrient state of the site under
consideration, and, consequently, associated with the latter:
- (MOS): Moss influence. Species marked under hummock-level bog influence in Eurola's
table. The influence is renamed here in order to avoid confusion with other indications. A
hummock-level bog, according to the Finnish system, is a type of complex; it may well
include flarks;
- (NVA): Neva influence (typical of intermediate and, usually not eutrophic, flark-level mire
sites). Only exclusive neva-influence species are included in this class;
- (RFN): Rich-fen influence (typical of intermediate and flark-level mire sites with a high base
state). Includes species that combine this indication with others;
- (NON): None: no inherent nutrient effects were mentioned for the species held in this class.
Table 3.3
The indicative significance of eutraphentous (extremely rich) sites in the Finnish

system
Species
Base
Level
Suppl.
Inher.
Triglochin maritima
Solanum dulcamara
Carex panicea
Fraxinus excelsior
Sphagnum contortum
Carex paniculata
Cratoneuron filicinum
Plagiomnium elatum
Campylium stellatum
Carex appropinquata
Eleocharis quinqueflora
Epipactis palustris
Scorpidium cossoni
XRC
XRC
XRC
XRC
XRC
XRC
XRC
XRC
XRC
XRC
XRC
XRC
XRC
XRC
XRC
INT
FLK
FLK
INT
HUM
FLK
FLK
FLK
FLK
INT
INT
FLK
INT
INT
FLK
NON
FLD
ANY
NON
ANY
ANY
SEP
SEP
SEP
SEP
SEP
SEP
SEP
NON
FLD
NON
NON
RFN
RFN
NON
RFN
NON
NON
NON
RFN
NON
RFN
NON
RFN
RFN
Water
Nutr.
Type
ATM
CIR
CIR
LTH
LTH
LTH
LTH
LTH
LTH
LTH
LTH
LTH
EUT
MES
OLI
EUT
MES
EUT
MES
MES
OLI
OLI
OLI
OLI
OLI
OLI
SMA
SMP
DAV
MOL
SMP
LAS
SMP
DAV
DAV
DAV
SMP
DAV
DAV
DAV
LAS
Extract from Appendix C, sorted on nutrient state within water type indications
42
The overall distribution of the species revealed an accumulation in the classes representing the
intermediate and flark levels in base-poor and transitional (or unstable, see below) sites with
flooding influence. Supplementary nutrient effects are indicated by 137 species, while only 59
species are indicative of any inherent nutrient supply. A relatively large proportion (Table 3.2:
17 out of 26 species) of rich-fen indicators appears to be threatened. Threatened species are
also over-represented (10 out of 15, against 78 out of 308 for all species) among the indicators of
an extremely rich base state. However, only a minority (8 out of 59) of species which are
classed as transitional are threatened ones.
At first sight this result seems somewhat anomalous, since it is generally agreed that, in The
Netherlands, eutraphentous species are less likely to become threatened than (meso- and)
oligotraphentous ones. Apparently, the Finnish nutrient state is quite different from Ellenberg's
nitrogen figure. Eutrophic sites, in the sense of the Finnish definition, usually have both a high
alkalinity and a high base content. Especially phosphorus is not always easily available to plants
at such sites. This is apparent from Table 3.3, listing the eutrophic-only species (present class:
extremely rich) in comparison with the indicatory values assigned to them in the
Central-European system: they are mostly threatened DAVFEN and LASFEN species with a
clear tendency towards base-rich (lithotrophic), yet oligotrophic environments. For this reason I
have preferred the name base state over nutrient state for this factor.
A relatively large number of species in the transitional class of base state is known in
phytosociological circles as indicators of disturbance in The Netherlands, and is not threatened.
They are associated with weakly acid sites where the availability of nutrients largely depends
upon a fluctuating supplementary nutrient effect. This has been expressed by the mention of
unstable environments in addition to truly transitional fen.
3.5 The ecological significance of the phytosociological groups
The phytosociological groups can be relatively well defined in terms of the occurrence of
ecological indicators, as shown in Table 3.2 and in Fig. 3.6:
According to the condensed Central-European system of indications, the BOG-FEN-LASFEN-DAVFEN series corresponds to a gradient from atmo- to lithotrophic sites. The nutrient
state in this series, however, shows a (mesotrophic) maximum in the FEN stage. Species
reported as seepage indicators by Dutch authors are mostly LASFEN and DAVFEN species;
they strongly indicate an oligotrophic nutrient state, combined with a distinctly lithotrophic
water type (see also Table 3.4). MOLFEN species indicate about the same nutrient state as do
LASFEN species, but their base state appears to be much less lithotrophic. LITFEN and
SWAMP species include indicators of eutrophic sites.
This presumed response of the phytosociological species groups to the factors water type
and nutrient state is highlighted in so-called radar diagrams in Fig.3.6. The highest score in each
species group was used as the response unit, reflected by one radius length. In the ALL-species
diagram absolute species numbers were entered, but in the other diagrams a correction was
applied for the uneven distribution in the ALL group. This means that the circle (or, more
exactly, the inscribed regular pentagon) serves as the ALL-species reference in those diagrams.
The differences relative to the ALL-species distribution are significant at the 0.99% level (chisquare test) for all groups but the FEN, LIT, and DIV ones.
43
Fig.3.6
Radar diagrams highlighting differences between species groups as regards the

indicated water type and nutrient state in combined classes (see Table 3.2)
Note that radars representing specific group distributions have been corrected for the inequality in the
ALL-species distribution given in the first radar.
The Finnish nutrient state (base state) shows the same pattern (Table 3.2) as the water type in
the Central-European system, confirming that it is indeed an indication of base state, rather than
one of nutrient state proper. The wetness factor shows a difference between the LASFEN and
DAVFEN groups, the latter indicating a slightly deeper water table. The dominance of flark
level indicators, both in the LASFEN and SWAMP groups, and among the Dutch supposed
seepage indicators, is obvious.
The inherent nutrient effect is not easily interpretable in the Dutch situation. The pattern is
remarkable, however, especially in the general absence of indications for this type of effects
(high numbers in row None). Relatively low figures are found in this column for the BOG,
LASFEN, and, less so, DAVFEN, MOLFEN, and FEN groups. These figures, and the extremely
low figure for the group of Dutch seepage indicators, suggest that the locally developed peat
substratum might play an important rle in the explanation of the occurrence of these species. It
could play this rle by its exchange capacity for ions or by the release of nutrients and carbon
44
dioxide through decomposition processes (remember the indication of the kragge as a storage
organ in Section 3.3!).
The general importance of flooding influences in North-West Overijssel is very obvious
from Table 3.2. The attention is drawn, therefore, to the absence of indicators for this effect in
the DAVFEN group, which includes a large number of indicators of seepage influence. The
latter is described by Eurola et al. (p.21-22) as groundwater influence, a phenomenon
distinguished by moving groundwater with a more or less constant temperature or, in the
case of a less pronounced groundwater influence, a fluctuating temperature and oxygen
content of the water and peat of 6-8 mg/l with a saturation percentage of over 50 and a fairly
high reduction potential (Eh=200-400mV)...Groundwater influence includes not only the
immediately visible influence as in the case of springs, but also the minerotrophic effect
brought about by the seepage of water from the mineral soil (seepage effect).
Table 3.4
The seepage indicators according to Dutch authors

Centr.Eur.
Water
Nutr
Finnish system
Base
Level
Suppl
Inhrt
FEN species
Carex lasiocarpa
UNK
CIR
OLI
OLI
POR
POR
FLK
FLK
FLD
FLD
NVA
NVA
LASFEN species
Riccardia multifida
Carex diandra
Liparis loeselii
Aneura pinguis
Eriophorum gracile
Sagina nodosa
Drepanocladus lycopodioides
CIR
CIR
CIR
LTH
LTH
LTH
LTH
LTH
LTH
LTH
LTH
OLI
OLI
OLI
OLI
OLI
OLI
OLI
OLI
MES
MES
MES
WMS
TRL
FLK
FLK
ANY
FLD
RFN
NON
TRL
TRL
XRC
TRL
FLK
FLK
FLK
FLK
SEP
FLD
FLD
FLD
RFN
NVA
RFN
NVA
WMS
FLK
ANY
RFN
DAVFEN species
Parnassia palustris
Philonotis marchica
Scorpidium cossoni
Campylium stellatum
UNK
CIR
LTH
LTH
LTH
LTH
MES
OLI
OLI
OLI
OLI
OLI
XRC
WMS
WMS
WMS
XRC
XRC
FLK
INT
INT
FLK
INT
INT
ANY
NON
ANY
SEP
NON
SEP
RFN
RFN
NON
NON
NVA
RFN
MOLFEN species
Campylium elodes
Carex buxbaumii
UNK
CIR
MES
OLI
WMS
FLK
FLD
RFN
Extract from Appendix C as regards the species comprised in Table 2.1; Riccardia chamedryfolia excluded, Philonotis
fontana included as P. marchica
45
The seepage influence is much less strongly indicated, however, by the explicit group of Dutch
seepage indicators.
At this point it is possible to conclude that existing schemes of indicators suggest that the Dutch
seepage indicators are mostly LASFEN and DAVFEN species of the flark level, associated with
base-rich, yet nutrient-poor sites, where rich-fen or neva peat has developed under the continued
influence of seepage or flooding, as summarized in Table 3.4. It is worthy of note that, in the
quagfen situation, surface water usually reaches the root zone by penetration below the floating
kragge, where it should be able to acquire a physico-chemical character similar to the
groundwater described by Eurola et al.
46
CHAPTER 4
The area of North-West Overijssel
4.1 Introduction
The north-western part of the province of Overijssel (The Netherlands) is a portion

of the North Sea tectonic basin and forms a section of the former marshy fringe of
the North-West European lowland plain. It is located at 6 E long., 52 45' N lat.
(Fig.4.1).
The area lies within the major climatic region of the Marine West Coast, as a
constantly moist subtype characterized by a rainy climate with mild winters (The
Times Atlas of the World 1980, comprehensive edition). The climate graph,
Fig.4.2, summarizes 1931-1960 data (KNMI 1972, sheets 4, 16, and 24). The mean
annual precipitation for North-West Overijssel is about 765 mm. There is an
excess of water in winter and a shortage in summer (April-July).
North-West Overijssel is a polder-and-wetland landscape. The present mire surface
is about 0.5 m below NAP, whereas the water level in the boezem area is
maintained at 0.7-0.8 m below NAP. The polders are kept drained to various levels
(Fig.4.3). The wetland area, which is part of the boezem, comprises the nature
reserves De Weerribben in the north and De Wieden in the south. The quagfens
investigated in this study developed in abandoned turf ponds originating from
extensive peat dredging until the beginning of the 20th century.
The underlying strata consist of sandy infillings of a Pleistocene (Saalian)
melt-water valley, deeply eroded in older Pleistocene river sands. Impermeable
layers are not expected to occur above a depth of about 200 m below NAP.
North-West Overijssel
29
Fig.4.1
Location of North-West Overijssel (arrow)

A: Amsterdam; H: Hamburg; L: London; P: Paris;
1: Southern and eastern boundary of area with mainly Quarternary deposits
2: Southernmost extension of the Saale glaciation
Fig.4.2
Climate diagram for the North-West Overijssel area, based on 1931-1960 records
The smooth curve reflects the average monthly temperature. Bars represent rainfall, the dotted part after
subtraction of the evapo-transpiration (data from KNMI 1972)
30
4.2 Surface structure, land-use, and water management

The original mire
It is generally agreed that the North-West Overijssel area was virtually uninhabited mire till a
few hundred years BC. Both climatic and hydrological influences, and possibly also land-use by
early inhabitants, led to the cessation of the active growth of peat at about 200-300 AD. From
soil surveys (Veenenbos 1950, Haans & Hamming 1954, 1962) it appears that two expanses of
ombrotrophic mire, i.e., of bog, existed at the time, roughly coinciding with the central parts of
the present nature reserves De Weerribben and De Wieden (Fig.4.3e,g). Although the bogs
were presumably of a level type (Veenenbos, p.122), their expanses may have been up to 6 m in
thickness at the end of the growing stage. A sand bottom is reached at a depth of 2-4 m below
NAP.
The bogs were surrounded and separated by fen mire in the valleys of small rivers
originating from the Drenthian and Frisian bog areas, and, in the southern part of the area,
connected to branches of the Rhine-and-IJssel system. Sand dunes flanking the former river beds
are found in several places. The growth of the fen peat was influenced by flooding with water
from these rivers, and by the discharge of the local bogs. As a result, the peat is locally very rich
in iron, and clay and loam deposits are represented as well. Veenenbos (p.121) noted the
presence of low sandy ridges under the mire area, which might have protected the growing bog
area against the direct influences of river water.
Human occupation and peat industry (Table 4.1)
The present surface structure of the area was determined by the expansion of Lake Flevo (the
later Zuyderzee) from the north-west, and by human occupation and land-use. In the early
Middle Ages, goat- and cattle-grazing, some form of cropping, and shallow peat-cutting must
have constituted the primary forms of land-use in the mire area, including some superficial
drainage. The general tendency of a rising sea level, combined with subsidence and oxidation of
peat as an effect of land-use, made the local mire surface unsuitable for further habitation, thus
forcing the people to shift to the east, ultimately to the more elevated moraine area. Such
movings of settlements have been traced in several places, such as Giethoorn, Wanneperveen,
and IJsselham (Kroes & Hol 1979, p.97, 279).
During transgressions clay was deposited over the peat along the western margin of the area
(e.g., Fig.4.3, i). As the expansion of the Zuyderzee became a threat to the area, dike building
was undertaken since about 1000 AD. Dike bursts can be traced in the morphology of the area,
e.g., by the occurrence of pot-holes, and also in written documents, such as those about the
devastating floods of 1776 and 1825 AD.
Peat excavation for commercial purposes became of some importance in the late Middle
Ages. Although the minable peat mass extended to 2-4 m below the groundwater level, only the
superficial peat could be extracted. Later, (manual) dredging techniques were developed and
used from the 17th century onward, at first in the southern bog area (Slicher van Bath 1957).
Extensive peat excavations were still going on by the beginning of this century in the northern
bog area and in parts of the area where the peat had been covered by clay deposits. Peats with a
high ash content in the former river beds were not mined, but used as pastures for hay-making
and cattle-grazing. The peat was dredged in rectangular parcels, about 30 m wide and a few
hundred m long, the so-called petgaten. Narrow standing baulks were left in between, but still
31
Fig.4.3
Hydrographic setting of the North-West Overijssel mire reserves

from Van Wirdum 1979
Lettering
a Lake Venematen; b Vollenhove height; c Noordoost-Polder; d Lake Giethoornsche Meer; e nature reserve De Weerribben; f Paaslo morainic ridge;
g nature reserve De Wieden; h IJsselham polder area; i Blankenham polder area; k Kalenbergergracht (canal); B Beukerssluis locks (main boezem inlet up to
1973); D Meppelerdiep (Pleistocene discharge in wet seasons); O Ossenzijl (connexion to inlet from Frisian boezem); R Zwarte Water river (locked from river
IJssel); S Steenwijker Aa rivulet (Drenthe Pleistocene discharge); IJ Connects to IJsselmeer.
32
the area proved very sensitive to water erosion, especially during gales. The main broads
(wieden) of North-West Overijssel result from this form of erosion. Later phases of the
peat-dredging industry proceeded according to strict regulations prescribing the dimensions of
petgaten and baulks. This prevented the formation of large lakes in the northern part of the area.
In the beginning of our century, the area consisted mainly of a complex of broads, petgaten
and standing baulks, bordered and interspersed by grassland in the river valleys and protected by
dikes along the coast of the Zuyderzee. Fishing was an important means of subsistence in the
former mire area, next to some continued peat dredging and the beginning exploitation of reeds
(Phragmites australis). Locally, rushes (Scirpuslacustris s.l.) were cut for matting, but the main
area for rushes was the delta of the river IJssel and its branches, rather than the peat district
itself.
Water management
In the present century agricultural improvement became a goal for the development of the area.
This was accompanied by a progressively increasing rate of organization and planning, resulting
in large reclamations and a general improvement of the water management for this purpose.
The institutions responsible for the care of dikes, going back to 1363 AD, evolved into the
water board Waterschap Vollenhove in 1889. The water-level in the area was maintained by
sluicing the excess water off at low tide to the Zuyderzee, the Meppelerdiep, and the Linde.
Although a water level of 0.3 m below the average sea level was strived for, prolonged
high-water levels in the Zuyderzee and the rivers frequently inhibited adequate drainage. Winter
and spring inundations were a common phenomenon. It was only after the extremely high
internal water levels of 0.1 m above NAP, in 1910, that the responsible authorities were forced to
prepare a plan for a pumping station (Stroink), which was put into operation in 1920. The target
water-level was 0.5 m below NAP. The pumping station Stroink was provided with an additional
pump in 1929, so as to maintain acceptable water-levels also in the bordering Meppelerdiep
area. The water board was, by official regulation, obliged to take Meppelerdiep water in
whenever this exceeded a certain level. From 1930 on the official level in the Vollenhove
boezem area was 0.7 m below NAP, and from 1942 onwards the winter water-level was lowered
to 0.8 m below NAP. The obligations with respect to the Meppelerdiep area lapsed in 1972,
when the new pumping station Zedemuden took care of that area.
The pumping station Stroink was the main factor enabling the agricultural development of
North-West Overijssel, which was pushed ahead by state initiatives, especially since the world
wars during this century stressed the importance of a self-supporting production of food at the
national level.
Reclamations
The realization of an appropriate water-level control system resulted in 1928 in the foundation of
a reclamation company, in which State, Province, and Water Board participated. Its task was to
drain the lower parts of the area and to render the conditions for agricultural use as favourable as
possible. Much of the work was carried out as unemployment relief. In this way several large
and deeply drained polders (ca 2 m below boezem water level) were reclaimed (Fig.4.3) and
subsequently used for farming. Most of these polderlands are used as pastures to-day.
33
Table 4.1
250
1100 -
The development of the North-West Overijssel mire area
800
Cessation of active peat growth by the combined influence of climate,

hydrology, and early land-use. Formation of shallow lakes in the Zuyderzee area, erosion of mires, deposition of a clay cover in a brackish marsh
environment over the peat in the west;
1500
Human occupation of the bog surface. Increased transgressive influences

from the Zuyderzee and subsidence and peat loss resulting from land-use
force the inhabitants to shift to the eastern mire margin and the morainic
area;
ca. 1300
The building of dikes against the expansion of the Zuyderzee begins as a

first co-operative effort in water management;
1400 -
1600
Extraction of superficial peat layers for fuel;
1600 -
1900
Large-scale fuel peat excavation, shifting from the south to the north
during the 18th century. Formation of large broads by wind erosion and
floods. Inhabitation by peat workers, fishermen, and some farmers;
ca. 1825
The lower coarse of the Linde river is blocked from the open sea by the
construction of sluices, so enabling an improved drainage, for agriculture,
of the mire area;
1920
The pumping station Stroink becomes operational. The water level in the
area is lowered and held under control;
1930 -
1945
Closure of the Zuyderzee from the open Wadden Sea. The NoordoostPolder is reclaimed, increasing the water demand of the mire area. The
influence of brackish water is now excluded;
1928 -
1968
Further reclamations for agriculture (polders) and realization of a canals

plan. The boezem becomes a reservoir for temporary storage of excess
water from polders. Fen vegetation develops in the terrestrializing
petgaten and is mown for thatching (reed) and for hay making;
1945 -
1970
Biological inventories. The value of the area for nature conservation is

discovered. Increased supply of water from the IJsselmeer (the former
Zuyderzee), which has become a storage reservoir for Rhine water;
1970 -
now
Reed farming, recreation and nature conservation. Discontinuation of

management for economic reasons. Carr vegetation develops. Water
quality problems arise: the vegetation of aquatic macrophytes almost
disappears.
34
The ultimate extent of the polders has been a much-debated issue (Het Oversticht 1939,
Haans 1951, and other sources), until economical changes towards the end of the 1950s
necessitated the government to discontinue the reclamation activities. The reclamation company
was liquidated in 1968, when the value of the remaining mire area for nature conservancy and
for recreational purposes had become recognized.
The reclamation also resulted in the need for a canals plan to ensure and maintain the
communication between different parts of the area left as a boezem, primarily for watermanagement purposes. As the boezem area decreased in size, the temporary storage of excess
water from the increasing polder area was more clearly defined as a critical boezem function.
In the meantime other large undertakings involving water management in The Netherlands
were important for the area under investigation. In 1932 the Zuyderzee was dammed in and
renamed Lake IJsselmeer, nowadays a freshwater lake with a controlled water level. In 1942 the
Noordoost-Polder (Fig.4.3, c) was reclaimed within the IJsselmeer. This polder extends inland
as far as the former Zuyderzee dike, thus including a narrow strip of forelands. The average
water-level in the Noordoost-Polder is about 4 m below that in the boezem waters of North-West
Overijssel. This caused a lowering of the average groundwater-levels in part of the area and it
put an end to the seepage of brackish groundwater underneath the Zuyderzee dike.
During many years parts of the area that had not been excavated for peat were mainly used
for cattle grazing. Most of these have been polders for a considerable length of time, but their
water-levels were only 0.1-0.3 m below that in the boezem. In the last 25 years the target levels
have been lowered for most of these shallow polders, and their area has been increased.
Present land-use in the mire area
The impoverishment of the local economy, the hardships of fishermen and reed cutters, and the
discontinuation of reclamations brought large areas of mire in the hands of organizations for the
conservation of nature. The whole remaining mire area, including the lakes, has been declared a
nature protection area with different degrees of strictness. This change in land-use made
additional demands on the water management (Jol & Laseur 1982).
The reed production was severely affected during the 1960s by the import of cheap reed
from the Danube area. At present reed cutting in North-West Overijssel is recognized as an
important factor in nature management. The cropping of reed often involves extensive irrigation
of the reed beds by means of small wind-pumps and tractor-driven motor-pumps. More than
75% of the area in De Weerribben and a smaller part in De Wieden has been under irrigation for
a longer or shorter period (Muis 1974).
The importance of the boezem of North-West Overijssel for water recreation has grown
enormously since 1970, and this has been accompanied by an increasing amount of provisions
and regulations.
During dry summers the natural inflow of water into the boezem of Vollenhove is
insufficient to maintain the target water-level. Additional supply, during such periods, is
provided by the intake of water through sluices. This was formerly done from the Meppelerdiep
at the Beukerssluis (Fig.4.3, D, B), but that canal suffered from an increasing pollution,
especially during the dry season. At present the main water intake is at the Linthorst-Homan
sluices which connect the area with the Frisian water-storage system (Fig.4.3, near O).
35
36
Fig.4.4
The schematic profile across the valley includes approximate depth figures in m below NAP; the locations of borings have been indicated by the
codes of the appropriate piezometers (Chapter 5)
Diagram of the chrono- and lithostratigraphy of deposits in the ice-marginal valley of the Oer-Vecht, North-West Overijssel
4.3 Geology
Several ecological reports dealing with the mire area in North-West Overijssel suggest the
possibility of seepage of groundwater into the mire as a result of artesian pressure in the deeper
aquifer which was supposed to be overlain by a loamy morainic deposit. Meanwhile, however,
evidence has accumulated that the mire was formed in the deeply eroded (20-40 m deep) valley
of a Pleistocene (Saalian) melt-water river, the Oer-Vecht, where the sand infillings immediately
overlie lower-Pleistocene river sands (Faber 1960, p.463, 495-499; Ter Wee 1962, Jelgersma &
Breeuwer 1975, p.91, profile A-A'). The profiles in Geologische Dienst (1979) suggest a
fluvio-glacial origin for the lowermost 3-5 m of the sediment in the original river valley.
Impermeable layers are not expected to occur above a depth of about 200 m below NAP
(Fig.4.4).
The Late Saalian period contributed considerably to the sedimentation of sands in the Oer-Vecht
valley, but this deposition came to an end in the Eemian period. In some places an alteration of
layers of fresh-water clay and sandy clay rich in humus can still be found at a depth of 16-20 m,
often with peat and wood remains. The extent of these less transmissive Eemian deposits is
uncertain. They are certainly missing at several places due to strong erosion during the
Weichselian, preceding a new period of deposition of fluviatile sands (Ter Wee 1966). There is
no evidence of impermeable layers of any appreciable horizontal extent in the Oer-Vecht
sediments in North-West Overijssel.
During the Weichselian the Oer-Vecht changed into a local river system filling-up with
materials of mainly local origin. These deposits include fluvio-periglacial sands and loam layers
of relatively small horizontal extent. The top of the Twente Formation forms the basis of the
Holocene peat deposits in the mire landscape. The Twente strata cannot be considered
impermeable either, except for a low-permeability zone at their upper boundary with the
Holocene peat deposits.
37
CHAPTER 5
Aspects of the hydrology of De Weerribben
5.1 Introduction
The hydrology of De Weerribben has been strongly influenced by reclamations
and by the changing water management. Although in the past some outflow of
groundwater will have occurred, De Weerribben is now an area of groundwater
recharge. The chemical composition of the underlying body of groundwater is
largely determined by infiltrating surface water. Interestingly, older records of
groundwater composition show slightly brackish influences, and it seems that no
conspicuous upward flow of groundwater existed even before the reclamation of
the deep Noordoost-Polder in 1941. Still, a lithotrophic (akin to groundwater, see
Appendix D) character of the surface water is obvious from water samples taken
between 1960 and 1969. Although such a character is still found at some places in
the area, and at some instances, in more recent records the influence of the intake
of polluted surface water from the adjacent boezem of Friesland has become a
dominant feature from 1970 onwards. The associated type of water was not found
in earlier records. A detailed study of the chemical composition of surface water in
De Weerribben has shown a considerable lag of processes within more isolated
parts of the local boezem system as compared to the large canals. It is concluded
that the surface water management is a key factor for the water quality in the area.
57
Hydrology of De Weerribben
Fig.5.1
Isopotential pattern and location of piezometers in De Weerribben

a April, 18, 1975 (after a very wet winter); b August, 18, 1975 (extreme drought period)
58
5.2 The hydraulic potential distribution

Available data
The distribution of the hydraulic potential in the area can be studied on the basis of two sets of
numerical data. The first set comprises the hydraulic head in piezometers along a transect
through De Weerribben, the B-transect (Fig.5.1), installed in order to study the effects of the
reclamation of the Noordoost-Polder (Rijksbureau voor Drinkwatervoorziening 1938, Volker
1948). Data from this transect from 1940 onwards are contained in the archives of the TNO
organization and the Rijkswaterstaat. The 1974-1976 measurements were carried out as part of
the present study in close cooperation with the ICW Institute (presently the Winand Staring
Centre; Bon 1975). In this joint study additional piezometers were installed and regularly
controlled. The records obtained constitute the second set of numerical data.
A correction was applied for the greater density of brackish water in the calculations of the
hydraulic head from the observed water levels in the piezometers. This was done by reference to
complete chemical analyses and analyses of the chloride concentration. Errors due to an
expected inaccuracy of this correction are 10 cm for the B5 c and d filters, 4 cm for the B5,
B6, and B7 b filters, 1 cm for the B5, B6, and B7 a filters. No correction was applied for the
other filters, where the correction was less than 0.5 cm.
Observational errors are in the range of 1-2 cm. Systematic errors due to inaccurate absolute
height references are probably smaller than 2 cm for the various gauges at one location, but they
may be as large as 5 cm between locations. Most data refer to NAP, the Dutch datum level,
which roughly equals mean sea level.
The equipotential pattern in the top of the sand-bed
Fig.5.1 reflects the equipotential distribution in the uppermost part of the sandy subsoil below
the former mire area and in the phreatic aquifer in the morainic land. Fig.5.1a (April 1975)
reflects an extremely wet period, Fig.5.1b (August 1975) an extremely dry one. Note that even
in April the hydraulic head in the sandy subsoil of the mire area is well below the boezem water
level (0.79 m below NAP): this is an area of groundwater recharge. In the south, the east, and the
west, the groundwater potential decreases sharply towards the various polder areas. Only along
the borderline with the IJsselham polder area there may have been some discharge in April. It
must be borne in mind, however, that several wind-pumps and motor pumps used for irrigation
in spring raise the water level 20-60 cm over a considerable area of reed fields and so locally
influence the hydraulic head in the underlying mineral soil.
The potential distribution in the B-transect
The hydraulic potential distribution in the B-transect, at the same measuring dates, is shown in
Fig.5.2. The equipotential lines indicate a general E-W flow of water. The recharge by rainfall,
in the Paaslo area, and, by boezem-water, in the mire area are clearly demonstrated. It is
confirmed that discharge of groundwater into the IJsselham polder area and the neighbouring
margin of the mire area can only involve a rather shallow groundwater flow: below ca. 10 m the
general pattern is not interrupted.
59
Fig.5.2
Distribution of the hydraulic potential (cm) in the B-transect

a April, 1975 (after a very wet winter); b August 1975 (during extreme drought);
The numbers above each diagram refer to the water level in the surface water system (cm);
Shading applied to indicate lithology as in Fig.4.4 (in the blank area between 5 and 13 m the
transition between lower and upper strata is not accurately known)
60
Table 5.1
period:
The hydraulic potential in the boezem and in several piezometers from 1940 to
1975
1935-'37 <>
1942-'44 <>
-9
73
-4
1965-'67 <>
64
B1a-10
B1b-25
2
21
-29
-25
31
46
28
49
B2a-12
B2b-25
62
60
-7
-13
69
73
65
77
B3a-16
B3b*-22
B3b-29
63
-18
81
63
-18
81
B4a-14
B4b-30
50
50
-82
-79
132!
129!
-12
-14
144
143
-7
-9
151
152
B5a-12
B5b-30
B5c-75
B5d-120
54
43
31
16
-87
-109
-109
-108
141!
152!
140!
124!
-33
-38
-44
-21
174
190
184
145
-16
-11
-16
-28
190
201
200
173
-8
-11
-9
-17
198
212
209
190
IJsselmeer
5
215
250
+18
214
232
+63
+11
151
221
-15
-9
77
1974-'75
boezem
B6a-10
B6b-29
1953-'55 <>
77
28
47
-22
-21
50
68
80
86
-7
-10
87
96
-16
-19
B7a-17
B7b-27
*97
*100
154
158
435
393
See Fig.5.1 and 5.2 for locations of piezometers; All potentials in cm below NAP; Significant decreases or increases
between periods in italics; Piezometer codes: location-depth, the latter being the depth of the filter in m below NAP;
! significantly decreasing during 3-year period; * new filter
5.3 The decrease of the hydraulic head in the aquifer, 1935-1975

The B-transect allows for a historical account of the decrease of the hydraulic head in the
aquifer, as shown in Table 5.1. The numerical data for 1935-'37 reflect the situation before the
reclamation of the Noordoost-Polder in 1940-'42. The rainfall does not show a significant trend
during the 1935-'76 period. The following can be concluded:
(1) The reclamation of the Noordoost-Polder had a strong effect on the hydraulic head in the
B4 and B5 filters, and a weaker effect in the B3 ones. The decrease in the B1 and B2 filters is
due to other causes, such as the digging of the Steenwijk-Ossenzijl canal along the borderline
61
between De Weerribben and the IJsselham polder area, and the reclamation of polders within
North-West Overijssel and the adjacent Frisian area. Between 1935-'37 and 1940 (not shown in
the table), the hydraulic heads in B1a and B1b already fell to 27 and 45 cm below NAP,
respectively.
(2) All filters show a trend towards a decreasing hydraulic head from 1952 to 1975. This
trend probably results from combinations of various causes, such as the reclamation of polders
within North-West Overijssel, the lowering of the water table in existing polders, and the
increase of groundwater extraction.
(3) Since 1945 the B1, B2, and B3 filters all reflect a recharge profile, i.e., the gradient in
the hydraulic head enables a downward flow of groundwater towards the deeper parts of the
aquifer. At the B4 location, such a situation only exists since ca. 1965. At the B2, B3 and B4
locations a discharge profile was frequently observed until 1941. The difference between the
hydraulic heads at ca 15 and ca 30 m below NAP was remarkably small, however. The hydraulic
properties of the soil layer between the depths of the a and b filters having remained unaltered,
any upward flow of water through that layer, before 1941, must have been less conspicuous than
the present downward flow. At B5 a discharge profile existed before the reclamation of the
Noordoost-Polder. The upper part of the profile became recharging afterwards. The hydraulic
head decreased by more than 1 m, even at 120 m, however. An increase of the hydraulic head is
seen in the B6 filters after 1953-'55. This was probably due to the introduction of water supply to
the sandy soils in the border area of the Noordoost-Polder in 1949-'51 (Lindenbergh 1956).
5.4 The alteration of the hydraulic head gradient, 1935-1975
In order to summarize the alteration of the horizontal and vertical hydraulic gradients, the
1935-'37 and 1974-'75 data from the B-transect have been represented together in the diagram of
Fig.5.3. Next to the terms discharge and recharge of the aquifer, exfiltration and
infiltration are used here. The former denotes the transfer from groundwater to surface water,
the latter the reverse. Before the reclamation of the Noordoost-Polder, the drainage control level,
i.e., de boezem water-level, in De Weerribben was slightly below the hydraulic head at 15 m
below NAP, which, in turn, was slightly lower than the hydraulic head at 25 m below NAP. This
situation, indicating a modest discharge, is shown by the 1935-'37 records in Fig.5.3. The
gradient of the hydraulic head along the transect was remarkably level.
The occurrence and rate of exfiltration or infiltration of groundwater in De Weerribben not
only depends on the hydraulic gradients, but also on the presence of resisting layers. Such a
layer is notably present in the Schut- en Grafkampen area, where a clay cover had to be removed
before the peat could be dredged. The clay was disposed of in already existing petgaten. As a
result, the bottom of many petgaten in this area is now lined with clay. It is noteworthy that the
boezem water level was 0.5 m below NAP until 1930, and even 0.3 m below NAP before 1920.
Before 1920 it could only be controlled by discharge through sluices into the Zuyderzee at low
tide.
The very level gradient before 1940 suggests that small local flow systems may have existed
within the mire area, involving the transfer of water between the mire and the uppermost few
meters of the underlying sand bed. From a survey of the chemical composition of the
groundwater (Section 5.5) it appears that the B3 filters at 12 and 16 m below NAP contained
water with a chloride content of 169-208 mg/l from 1935 to 1950 (10 samples), while a variation
between 24 and 30 mg/l was found in the filters at 22 and 29 m below NAP.
62
Fig.5.3
The alteration of the hydraulic head gradient in the B-transect, 1935-1975

The distance of the piezometers from the centre of the mire area has been indicated along the axis
This almost conclusively proves that brackish water, originating from floodings, was able to
infiltrate into the sandy subsoil due to its higher density, and that the deeper fresh groundwater
did probably not discharge in any substantial quantity into the mire area at the B3 location. The
B2 and B4 filters contained fresh water in the same period.
In the 1974-'75 situation the gradient of the hydraulic head is steeper. Only in the NoordoostPolder and in the IJsselham polder area is the drainage control level kept below the hydraulic
head of the groundwater at a depth of ca 15 m below NAP, so that exfiltration may occur. As
mentioned before, the hydraulic head in the IJsselham area is lower at 25 m below NAP than at
15 m: the exfiltration profile is a local phenomenon, superimposed on a deeper recharge profile
which extends from the Paaslo morainic land to the Noordoost-Polder.
63
Extending conclusion (3) in Section 5.3, it is of some interest to include the much increased
difference between the hydraulic head in the boezem and that in the body of groundwater in the
underlying sand bed. It appears that any exfiltration of groundwater into the boezem in the early
period must have been of a much smaller magnitude than the present infiltration. It is almost
sure that the influx of fresh groundwater into the boezem, in historical times, was less important
quantitatively than the discharge of surface water from neighbouring areas.
Table 5.2
Chloride and isotope analyses of groundwater in the B-transect, 1935-'82
Nm-dpth
'35
'37
'38
'40
'41
'42
'43
'44
B1a-10
B1t-17
B1b-25
52
33
29
55
52
39
32
28
26
29
31
32
32
30
B2a-12
B2b-25
20
24
20
23
22
24
21
23
18
22
'47
'50
'58 '61
23
26
29
38
'63
'74
'75
'79
'80
'80
'81 '82
(26)
29
48
30
21
68.7
(31)
27
37
50
35
25.1 76.1
37
52
54
49
51
56
47
60.1
76.9 112.5
69
29
LM187a-10
LM187b-14
LM187c-29
B3t-12
B3a-16
B3t-22
B3b-29
B4a-14
B4a*-22
B4b-30
B4b*-50
B5a-12
B5t-26
B5b-30
B5t-39
B5t-54
B5t-69
B5c-75
B5t-90
B5d-120
3H
14C
115 110
330 335
540 510
11.9
73
63
56.6
45.1 71.8
<1.1 59.2
172
29
27
169
177
169
175
176
195
195
202
208
26
27
26
25
27
28
24
29
30
29
26
28
31
28
32
34
47
32
28
32
34
33
33
27
44
-new filter B3a-:

-new filter B3b-:
59
73
74
72
73
62
-new filter B4a-: 74 43.4

-new filter B4b-: 28 <1.0 50.0
3656 3729
1424
1577
1618
1547
1862
2127
2662
3251 2978
3727 3507 3564 3325
973
760
153
93
71
84
119
160 130
1597 1620 1630 1520 1480 1534 1275 1067
251
156
124
120
78
2130 2143 2146 2137 2100 2087 1917 1815
1075
428
250
180
3337 3351 3318 3332 3243 3302 2995 3235
3150 3044
B6a-10
B6b-29
B6t-31
2876 2955 3195 2245 1055 564

1250 1345 1333 (99 1540) 2060
1278
B7a-17
B7b-27
B7t-34
1048
122
233
909
119
335
115
129
177
284
212
576
334
130
56.7
76 32.3 79.6
3000 2750
2790
<1.6 67.7
115 94
820 339
100
360
42.9
50.5 95.2
1330
1210
Depth in m below NAP; Cl- in mg/l; 3H in TU (Tritium Units); 14C in % of recent concentration.
Notes: t temporary test-filter; * new filter; isotope analyses apply to this sample; () uncertain data.
5.5 Chemical composition and age of groundwater

The chloride content of the groundwater in the various filters has been rather frequently
analysed. The pertaining data have been collected in Table 5.2 and in Fig.5.4, together with the
results of 3H and 14C determinations made around 1980. W.G.Mook (pers.comm.) reported
64
Table 5.3
The chemical composition of groundwater in North-West Overijssel
smpl
pH
date
yymmdd
Ca
mg/l
Mg Na
mg/l mg/l
K
mg/l
Cl HCO3
mg/l mg/l
SO4
mg/l
EC25
mS/m
IR
%
x
%
y
%
pHsat
rLI rAT
% %
rTH rRH Ca
% % %
Mg
%
Cl Note
%
B1a (16D8): 10 m below NAP

B1a
370122 6.0? 8.0
B1a
790905 6.0? 12.0
B1a
80....
6.0? 18.0
B1a
801119 5.9 13.0
B1a
810827
9.5
3.9
3.6
12.0
9.0
40.7i
22.0
26.0
16.0
7.0?
7.9
8.0?
8.0
55.3
29.0
48.0
30.0
21.0
19.5
18.9!
35.0
19.0
28.8
44.8!
53.0
60.0
26.8
25.4
38.0
27.6
26.5
20
42
40
43
45
0 -11
(0) (2)
3
4
-2
-3
9.45
9.29
8.87
9.26
-20
-4
12
-8
49
62
33
58
8
14
46
16
51
53
75
38
16
29
28
28
13
15
31
33
63
40
45
35
1#
2
4
3
5
B1b (16D8): 25 m below NAP

B1b
370123 6.2?
B1b
790905 6.2?
B1b
80....
6.2?
B1b
801119 6.2
B1b
810827
4.1
5.5
9.0
12.0
29.9i
21.0
18.0
25.0
10.0?
10.0
10.0!
13.0
29.0
27.0
37.0
50.0
35.0
25.6
25.0
26.0
34.0
38.7
48.4!
49.0
68.0
21.1
22.1
28.0
37.5
38.6
33
42
40
42
48
0 -16
(0) -17
(0) -6
1
-6
9.32
9.21
9.09
8.84
-15
-10
-1
6
59
63
49
43
-5
-5
17
38
36
32
51
66
20
25
28
29
17
21
30
29
40
35
42
42
1#
2
4
3
5
5.5
5.2
9.9
18.0
19.6i
21.0
19.0
24.0
2.0?
2.2
2.0?
2.0
20.4
37.0
52.0
49.0
56.0
191.5
218.2!
400.0
374.0
tr.
4.0?
<3.0
4.0
32.9
45.3
72.0
71.8
74.0
81
76
78
79
78
-1 -12
(0) -3
-8
1
2
-2
7.71
7.54
7.12
7.14
95
97
98
98
-51
-49
-54
-56
4
21
42
43
30
48
58
58
65
70
75
67
12
9
12
19
15
22
18
18
1
2
4
3#
5
8.4 12.0i
9.4 22.0
11.0 26.0
1.0?
1.0
1.0
150.0
366.0
352.0
14.8
<3.0
3.0
7.86
7.17
7.16
94 -49
98 -53
98 -53
3
39
40
30 64
57 74
59 72
21 19
12 20
13 20
228.0
3.0?
77
77
79
81
70
-11
-2
-2
1.4
30.9
68.0
68.5
67.3
56.4
-0
-6
0
5.5 33.0
22.8
54.0
51.0
47.0
57.0
7.47
94 -44
37
65 66
8 30
1
4
3
5
8
8.1
11.0
14.0
20.0
18.5

B2a
370122 7.2? 48.0
B2a
790905 7.2? 66.0
B2a
80....
7.2? 103.0
B2a
800918 7.2 105.0
B2a
810827
115.0
B2b
370122 7.2? 43.7
B2b
80....
7.2? 100.0
B2b
800918 8.0 105.0
B2b
810827
112.0
B2b
880719 7.1 75.0
LM187a (16D41): 10 m below NAP
LMa
801119 6.7 108.0
LMa
810309 6.8 108.0
LMa
820308 6.8 100.0
LMa
851206 6.6 127.0
13.5
13.0
17.0
17.2
83.0
88.0
78.0
61.4
3.0
3.0
3.0
2.8
115.0
110.0
102.0
83.0
440.0
471.0
446.0
514.0
9.0
7.0
8.0
2.0
97.2
108.2
102.8
101.7
63
64
64
73
-2
-3
-2
-1
-5
4
5
1
7.08
7.05
7.11
6.95
82
78
80
89
-40
-39
-41
-48
67
75
74
66
84
86
85
77
LM187b (16D41): 14 m below NAP

LMb
801119 6.7 137.0
LMb
810309 6.8 135.0
11.5 180.0
10.0 190.0
4.1
5.0
330.0 415.0
335.0 423.0
13.0
2.0
154.6
169.1
42
42
-2
-1
-7
1
7.04
7.04
48 -16
46 -16
88
91
LM187c (16D41): 24 m below NAP

LMc
801119 6.7 275.0
LMc
810309 6.8 262.0
LMc
820308 6.8 246.0
LMc
851206 6.7 282.0
20.0
18.0
25.0
19.4
4.1
4.0
5.0
3.7
540.0
510.0
535.0
541.0
15.0
1.0
6.0
4.3
215.3
243.1
227.7
237.6
47
48
45
48
-1
-2
-1
2
-9
8
-1
0
6.75
6.75
6.80
6.75
46
44
43
44
-17
-19
-17
-18
B3a 1937 (16D14): 16 m below NAP, *B3a new piezometer (16D37): 15 m below NAP
B3a
370121 6.8? 163.0
5.7 98.4i
4.0? 169.0 494.0
2.3 123.8
*B3a
790905 6.8? 90.0
5.3 47.0
1.6
59.0 322.1!
4.0? 68.5
*B3a
80....
6.8? 103.0
6.6 40.0
1.5? 74.0 368.0
<3.0
75.0
*B3a
801119 6.8 98.0
5.8 36.0
1.4
73.0 300.0
7.0
65.1
*B3a
810827
95.0
73.0
67.3
63
73
71
70
70
-0
(0)
-4
-1
-3
-0
-2
-8
6.88
7.26
7.16
7.26
76
94
94
95
B3b 1937 (16D14): 29 m below NAP, *B3b new piezometer (16D37): 22 m below NAP
B3b
370121 7.0? 109.0 15.2 37.5i
2.0? 26.4 403.0
tr.
*B3b
790905 6.8? 114.0
6.5 46.0
3.2
73.0 377.2!
4.0?
*B3b
80....
6.8? 105.0
7.7 51.0 13.0? 72.0 434.0
<3.0
*B3b
801119 6.8 117.0
7.2 47.0
2.4
63.0 410.0
13.0
*B3b
810827
113.0
62.0
65.0
79.5
83.0
77.3
78.4
88
73
72
77
76
6 -13
(0) -2
-4
-3
-1
-8
7.10
7.11
7.08
7.06
99
94
93
96
B4a 1937 (16D15): 14 m below NAP, *B4a new piezometer (16D51): 22 m below NAP
B4a
370216 7.0? 106.0
6.8 25.3i
1.0? 26.3 380.0
tr.
*B4a
821026 6.9 122.0
8.7 51.0
1.5
74.0 435.0
5.0
*B4a
880719 7.1 140.0
9.1 34.0
1.3
57.0 418.0
5.0
61.1
88.3
88.4
88
75
81
-0
-1
4
7.12
7.03
6.98
165.0
150.0
176.0
178.0
455.0
476.0
447.0
453.0
-7
-0
4
53
52
51
60
11
10
14
14
30
28
28
22
5
6
6
6
98 44
97 42
6 57
5 57
5
6
92
96
95
95
96
93
95
95
61
62
55
60
7
7
9
7
66
65
67
67
5
6
6
6
-34
-46
-47
-44
77
45
48
39
91
68
69
65
63
64
69
70
4
6
7
7
37
24
26
29
1
2
4
5
5
-57
-47
-49
-48
29
51
53
45
47
71
70
65
65
68
62
68
15
7
8
7
10
25
22
20
1
2
4
5
5
99 -54
93 -47
95 -50
26
57
55
45 76
73 67
70 76
8 11
8 22
8 19
1
5
8
65
smpl
date
pH
yymmdd
Ca
mg/l
Mg
mg/l
Na
mg/l
K
mg/l
Cl HCO3
mg/l mg/l
SO4
mg/l
EC25
mS/m
IR x
% %
y
%
62.9
61.8
87 -0
87 -1
-1
-2
7.15 100 -56

7.16 100 -55
35.4 1282.6
38.5
66.6
35.0?
79.5
<3.0
91.0
30.0
87.2
88.3
6.0
67.4
12 0
59 0
44 0
39 -4
46 -3
46
57 1
8
-9
-2
-6
-4
6.62
7.55
7.67
7.49
7.48
29
85
60
60
66
-15
-43
-25
-31
-29
100
58
76
73
75
-1
7.53
81 -44
53
74 39
23 31
99
89
84
90
83
87
97
99
91
83
21
35
39
51
56
36
7
14
19
17
85
44
42
24
15
1
1a
2
4
3#
5
B4b 1937 (16D15): 30 m below NAP, *B4b new piezometer (16D51): 50 m below NAP
B4b
370126 7.0? 105.1 8.3
16.7
1.0?
28.2 360.0
0.8
*B4b
821026 7.0 106.0 8.0
12.5
1.3
28.0 350.0
7.0
pHsat
rLI rAT
% %
rTH rRH Ca
% % %
31
30
Mg
%
Cl Note
%
48 78
47 81
10 12
10 12
1
5#
85
77
94
90
92
6
18
19
22
17
1
1a
2
4
3#
5
8

B5a
370128 7.5? 293.0 91.7 2286.0i
B5a
630327 7.9
58.0 15.0
67.0i
B5a
790905 7.8? 53.0 17.0
66.0
B5a
80....
7.5? 57.0 24.0
81.0
B5a
801119 7.5
63.0 17.0
75.0
B5a
810827
62.0
B5a
880719 7.5
55.0 19.5
58.0
40.0? 3729.0 965.0

30.0?
71.0 256.0
28.0
119.0 217.9!
25.0? 160.0 314.0
25.0
130.0 290.0
130.0
2.6
74.0 282.0

B5b
370126 7.5?
B5b
630327 7.3
B5b
790905 7.5?
B5b
80....
7.0?
B5b
801119 7.0
B5b
821026 6.9
222.0
112.0
65.0
152.0
155.0
183.0
30.0? 1577.0 398.0 73.3

30.0? 251.0 427.0 94.3
37.0
124.0 164.8! 100.0?
4.0? 120.0 356.0 238.0
4.0
78.0 332.0 350.0
76.0
560.2
149.5
86.1
138.0
127.0
127.0
20 0 6
44 0 -10
48 0 -6
69 2 -5
78 -4 -14
81
7.02
7.12
7.71
7.08
7.10
29 -15
47 -14
50 -7
58 -15
58 -5
B5c (16D1): 75 m below NAP

B5c
370126 7.0?
B5c
630327 7.2
B5c
80....
7.0?
B5c
801119 6.9
350.0
272.0
143.0
130.0
30.0? 2127.0 376.0 168.5

10.0? 1075.0 531.0 12.2
3.0? 250.0 448.0 <3.0
3.0
180.0 442.0 30.0
728.6
393.4
126.0
121.4
23 0
31 0
50 -7
56 -4
3
-3
-9
-3
6.87
6.75
6.98
7.02
29
32
66
70
-13
-13
-27
-30
100
98
80
82
87
91
95
93
25
34
57
54
5
7
9
11
86
77
49
39
1#
1a
4
3#
B5d (16D1): 120 m below NAP

B5d
370126 7.0?
B5d
630327 7.3
B5d
80....
7.0?
B5d
801119 6.9
B5d
810827
211.0
208.0
300.0
305.0
295.0
11 1 18
10 1 0
12 -9 -19
16 -2 4
16
7.28
7.25
6.99
6.97
29
28
27
29
-15
-14
-13
-14
100
100
99
100
84
85
87
86
11
10
15
18
16
16
15
13
87
88
92
86
1#
1a
4
3
5
1
-9
-3
1
7.34
7.36
7.50
7.56
28 -13
27 -7
72 -35
68 -21
100
96
64
79
86 8
92 16
86 40
94 44
10
22
14
13
89
61
41
34
1
1a
4
3
5
-1 -7
-0 -5
-1 -17
-11 -22
-1 3
6.79
6.78
6.73
7.19
7.21
31
32
29
28
39
-12
-12
-12
-7
-16
98
98
99
92
94
92
91
90
94
94
3
3
5
13
16
85
84
87
75
54
1
1
1a
4
3
5
Oss: irrigation well in reed bed, 15 m below NAP (geohydrological code not yet available), test filter 6 m below NAP
Oss-06
880427 6.2
12.0 2.0
64.0
1.6
48.0 98.0 12.0
40.9 31 6 12
Oss-15
880613 6.8 105.0 7.7 120.0
2.9
170.0 415.0 14.0
114.9 52 -3 -1
Oss-15
880720 6.6 125.0 8.2 130.0
1.7
165.0 450.0
5.0
119.4 57 2 -3
8.59
7.13
7.02
28
5
66 -26
69 -29
30
79
79
64 17
94 47
93 49
5 42
6 40
5 38
8
8
8
225.0
14.4
14.0
33.0
28.0
525.0i
215.0i
67.0
100.0
85.0
44.8 1113.0i
33.6 532.0i
13.0
98.0
16.0
94.0
192.0
196.8
175.0
140.0
1652.0i
1744.0i
1550.0
1350.0

B6a
421111 7.3? 149.0 111.0 1709.0i
B6a
581103 7.3
83.5 67.4 375.0i
B6a
80....
7.4? 60.0 13.0
68.0
B6a
801007 7.4
66.0 12.0
62.0
B6a
810827
107.0
40.0?
40.0?
40.0?
42.0
2978.0
3150.0
3800.0
2750.0
2790.0
388.0
347.5
158.0
300.0
1174.4
1005.3
905.0
913.1
828.1
40.0? 2876.0 324.0 247.0

30.0? 564.0 380.0 196.3
20.0
115.0 279.0
3.0
20.0
94.0 224.0 73.0
100.0
923.1
248.5
77.0
81.7
103.8
8
21
48
55
66
408.6
414.1
574.0
244.0
185.5
171.1
33
34
27
18
31
30
B6t (16D2): test filter 31 m below NAP, B6b (16D2): 29 m below NAP
B6t
421021 7.3? 349.0 17.0 516.0i 30.0? 1278.0
B6b
421022 7.3? 368.0 16.7 496.0i 30.0? 1250.0
B6b
581103 7.1 426.5 38.3 947.0i 30.0? 2060.0
B6b
80....
7.2? 100.0 40.0 360.0 30.0? 820.0
B6b
801007 7.2
86.0 33.0 220.0 31.0
339.0
B6b
810827
86.0
360.0
275.0
305.0
387.0
380.0
386.0
376.0
427.0
465.0
460.0
tr.
16.1
75.4
<3.0
27.0
-0
-0
-2
-2
12
41
35
32
37
42
44
32
20
25
86
29
44
46
41
Next to the piezometer `B' codes, the codes used in the Geohydrological Archive have been indicated, along with the approximate depth of each filter in m below NAP
Explanation of items: Appendix D; x electro-neutrality error; y conductivity error; pHsat pH at which satured with respect to calcite, at 10 C; Ca, Mg, Cl (%) f(c)1/2Ca2+, f(c)1/2Mg2+, and f(c)Cl-, respectively; () Results
from assumptions made to compute missing values; ? Assumed value; ! Calculated from electro-neutrality or conductivity, or from both; i (Na+K), given as mg/l of Na; tr trace
Notes: 1 By RID for Dienst Zuiderzeewerken (archives); pH and K assumed, Na includes K (determined by analysis); 1a As 1, but pH computed from free CO2 and HCO3; 2 By Waterlaboratorium Zwolle for
Zuiveringschap West-Overijssel (pers.comm.); pH and either one of HCO3 and SO4 assumed, the other one computed; 3 By Waterlaboratorium Oost for Dienst Waterhuishouding en Waterbeweging and for RIN
(present survey); 4 By Oranjewoud Consulting Engineers for DGV-TNO (Uil & De Heer 1985); pH and K assumed; 5 By Waterleidingbedrijf Midden-Nederland for RIN (present survey); 6 By RIVM, groundwater
monitoring network; 8 By Waterlaboratorium Zwolle for Zuiveringschap West-Overijssel (pers.comm.); # Maucha diagram in Fig.5.5
66
Table 5.4
Sample
B1a
B1b
B2a
B2b
LM187a
LM187b
LM187c
B3a
B3b
B4a
B4b
B5a
B5b
B5c
B5d
B6a
B6t
B6b
Oss-06
Oss-15
Oss-15
Additional analytical data for groundwater in North-West Overijssel
Date
yymmdd
NO3 NH4 orgNH4

KMnO4 PO4
<----------N, mg/l--------->mg/l <---P, mg/l---->
801119
801119
800918
800918
880719
801119
810309
820308
851206
801119
810309
801119
810309
820308
851206
801119
801119
821026
880719
821026
630327
801119
880719
630327
801119
630327
801119
630327
801119
421111
581103
801007
421021
421022
581103
801007
880427
880613
880720
0.07
2.03
<0.02
<0.02
<0.05
0.14
<0.1
0.1
<0.3
0.09
<0.1
<0.02
<0.1
<0.1
<0.3
0.07
0.02
0.02
<0.05
0.05
0.0
0.16
<0.05
0.0
0.18
tr.
0.18
tr.
0.27
0.05
0.02
3.42
0.93
0.6
6.92
(8.2
(8.9
10.2
6.07
(7.2
3.11
(2.9
(4.3
4.07
3.19
2.41
4.05
3.3
1.17
1.24
2.18
3.2
2.88
2.80
4.20
2.18
3.03
3.50
0.0
0.77
0.0
0.59
0.09
<0.05
<0.05
0.19
0.08
0.68
0.09
0.7
12
4
119
38
= 8.2)
= 8.9)
57
62
= 7.2)
52
<0.01
<0.01
0.15
0.10
0.27
0.43
Ptot SiO2
mg/l mg/l
17
14
1.40
6.69
1.40
0.26
118
18
1.33
0.70
6.85
7.78
3.6
7.9
8.0
0.46
1.0
2.1
0.9
41
68
0.50
1.00
0.48
0.26
18
38
35
38
6
14C
N,P
68.7
25.1 76.1
60.1
76.9 112.5
Conversions
KMnO4
x
x
x
x
x
11.9
20
19
27
17
x
x
x
x
0.20
0.01
0.19
48
35
40
31
43
42
60
50
20
0.20
0.45
0.49
0.40
3H
0.46
0.30
0.04
0.01
0.06
<0.01
0.40
0.90
0.95
0.25
0.25
0.28
0.30
0.47
0.47
0.52
0.9
1.7
0.07
18
10
CO2
TU
0.83
0.14
0.9
0.30
0.1
15
17
27
27
Mn
mg/l
0.28
0.02
= 2.9)
= 4.3)
Fe
mg/l
0.08
19
26
10
13
21
<1.0
59.2
56.6
45.1
43.4
71.8
<1.0
50.0
x
x
x
x
x
0.27
19
26
19
0.84 0.27
0.65 0.58
8
14
56.7
x
x
x
1.4
33
24
24
30
30
26
39
21
26
34
7.5
11
26.4
12
15.0
17
1.6
14.3
13.4
30
8.4
44.8
16
0.33
0.71
0.62
0.56
0.80
1.9
49
60
83
86
39
86
32.3
79.6
<1.6
67.7
0.47
0.39
46
15
42.9
1.9
0.51
91
50
50.5
x
x
x
x
x
x
x
x
95.2
x
x
0.53
0.32
0.30
x marks conversions applied

Laboratories: see notes in Table 5.3; 3H and 14C by Isotope Physics Laboratory, University of Groningen, for RIN
NO2 concentrations (mg/l) all <0.02 (Waterlaboratorium Oost), <0.01 (Waterleidingbedrijf Midden-Nederland), "0" (RID), or included in NO3 (Waterlaboratorium
Zwolle)
the age of the relevant groundwater samples. Further data with respect to the chemical
composition of the groundwater in the B transect are available for 1937, 1980-'82, and some
other years. They have been summarized in Tables 5.3 and 5.4 and in the EC-IR, rTH-rLI, and
Maucha diagrams of Fig.5.5-5.6. But little attention is being paid here to the 1979 results. Their
position is often more or less intermediate between the 1937 and 80-'82 ones, but this may be
due to difficulties met while emptying the piezometers prior to sampling.
67
Fig.5.4
Variation of the chloride concentrations at some locations in the B-transect

Note the logarithmic scale for the chloride concentrations
The results of this part of the survey are summarized below:

B1 (Paaslo morainic area)
Of all samples, the B1 ones are the most similar to rain water, although the IR and EC values are
somewhat higher. The samples are definitely unsaturated as regards calcite. The chloride content
at 10 m below NAP decreased from ca 50 mg/l in 1935-1938 to 20-30 mg/l afterwards. This may
be due to a decreased chloride content in the precipitation since the IJsselmeer became a
freshwater lake in 1932. The filter at 25 m below NAP shows a variation of the chloride content
around 30 mg/l. The 3H figures indicate that in 1980 both filters contained water that must have
infiltrated after 1960.
The sulphate concentrations in the B1 samples are much higher than those in most other
samples, and they almost doubled between 1937 and 1980. The magnesium concentration also
showed a notable increase. This possibly reflects the application of manure and fertilizers.
B2 (Border between De Weerribben mire and IJsselham polder areas)
The 1937 analytical results reflect a lithotrophic type of groundwater, possibly a stage of
development between freshly infiltrated rain water and aged groundwater, such as found in the
B3b (1937) and B4 piezometers (Fig.5.5). The recent samples reveal a notable increase of the
overall concentration, which is mainly due to the calcium and bicarbonate concentrat-ions. The
chloride content in both B2 filters was 20-25 mg/l until 1944. The present values (1981-1982)
are ca 50 mg/l. In 1980 both filters contained water that must have infiltrated after 1960. The
increased chloride content may be due to local influences including the infiltration of boezem
water.
68
Fig.5.5
Maucha diagrams of representative groundwater analyses arranged in an EC-IR

diagram
LI, AT, TH: benchmark samples for litho-, atmo-, and thalassotrophic water, respectively (see Appendix
D)
Fig.5.6
EC-IR (a) and rTH-rLI (b) diagrams of groundwater analyses above ca 40 m below
NAP, showing the decreasing variation between 1937 and ca 1980
L, A, T, M: litho-, atmo-, thalasso-, and molunotrophic benchmark samples; Mixing contours and the
line IR(%)=EC25(mS/m) added to Fig.5.6a for convenience (see Appendix D)
69
The local infiltration involves a steady supply of carbonic acid, produced in the root zone of
the pastures in this area, which facilitates the solution of calcium in the sand bed. The recent B2b
samples are only slightly supersaturated in calcium and bicarbonate, especially due to the high
pH. In situ the calcium concentrations should possibly be considered to be just saturated.
LM187 (Pierikken area)
These filters were installed in 1980. The chloride contents of the water were consistently high:
ca. 110, 330, and 500 mg/l at depths of 10, 14, and 24 m, respectively. According to the isotope
determinations, which were only carried out in the 10 and 24 m filters, the deepest filter contains
water that may be about 2000 years old. The water in the filter at 10 m has probably in part
infiltrated before 1960. It is rather similar to recent Rhine water (see rRH in Table 5.3), but there
is certainly no genetic relation with the latter. There may be some connexion with a body of
fossil brackish groundwater which was mapped by Senden (1980) on the basis of geo-electrical
surveys. The well Oss drilled in 1988 in the Schut- en Grafkampen part of the mire area, near
Ossenzijl, also provides slightly brackish water. The situation shows, anyway, that there was
probably no discharge of fresh groundwater from a regional groundwater flow system at this
location in historical times.
B3 (Centrally in De Weerribben mire area)
As referred to before (Section 5.4), the B3 filter at 16 m below NAP contained water with a
chloride content of 169-208 mg/l from 1935-1950, while the chloride content at 29 m below
NAP varied from 24-30 mg/l. The higher values are confirmed by the analysis in the test filter at
12 m, 1935. The slightly brackish B3a sample from 1937 is not very different from the recent
LM187a samples. The same final conclusion seems justified: there was probably no regional
groundwater discharge in historical times. The B3b sample is very similar to the lithotrophic
water found in the B4a piezometer in 1937 and in the B4b piezometers in 1937 and 1982. The
latter must be considered to be aged fresh groundwater. Note, however, that the magnesium
concentration in B3b is higher than in B4.
The filters were not maintained after 1950. In 1974, however, new filters were installed at a
distance of ca 200 m from the original location. The variation of the chloride content in the new
filters was 59-73 mg/l, and, in 1980, both filters contained water that must have infiltrated after
1960. The origin of this infiltration water is probably the boezem.
B4 (Border between De Weerribben mire and Blankenham polder areas)
The chloride content of the water in the B4 filters varied around 30 mg/l from 1935 to 1944. The
1950 values are slightly higher: ca. 45 mg/l. The filters were removed in 1977. New filters were
installed nearby in 1982. The 1982 water samples from 22 and 50 m below NAP contained 74
and 28 mg/l of chloride, respectively. The water in the 22 m filter had very recently infiltrated
(after 1960). The deeper groundwater, however, has a low 14C content (50.0 %), and its age was
estimated as slightly less than 4000 years. It is a strongly lithotrophic type of water in
equilibrium with calcium carbonate.
70
B5 (Blankenham polder area)

As shown in Fig.5.4 the chloride content in the B5 filters exhibits a very regular variation. The
original brackish water is gradually being replaced by fresh water. At a depth of 12 m below
NAP the freshwater stage was reached around 1950. At 30 m below NAP it may have been
reached in 1981, while the desalinization process is still under way at 75 m below NAP. The
chloride content in the groundwater at a depth of 120 m below NAP only shows a slight
decrease. The isotope concentrations in the 1981 sample of this deep water indicate an age
between 30 and 1000 years; this brackish water obviously infiltrated from the former Zuyderzee.
Samples from the 12 and 30 m filters were also subjected to isotope analyses. The results reveal
that the brackish groundwater is being replaced by locally infiltrated water of recent (post-1960)
origin, rather than by aged groundwater. The B5b samples are especially remarkable for their
high and still increasing sulphate concentrations.
B6 and B7 (Noordoost-Polder)
The variations of the chloride concentration in water samples from the B6 and B7 filters, with
the exception of the B6 10 m filter, demonstrate an increase following the sharp decrease
immediately after the reclamation of the Noordoost-Polder. In the B6 30 m filter this increase is
in turn followed by a decrease, which may be due to the supply of fresh water in this part of the
Noordoost-Polder. A further discussion of the chemical composition of the B6 and B7 samples
lies outside the scope of the present study.
Conclusion
It is apparent from these data that the fall of the hydraulic head in the body of groundwater under
the mire area has caused a local, ongoing infiltration of boezem water to a considerable depth.
This water is mainly of external origin, being supplied to the boezem via the surface water
system. Only the B4 50 m filter contained aged freshwater which may have infiltrated during the
growth phase of the virgin mire, in the Subboreal period. The body of groundwater under De
Weerribben comprises one or more local pockets of brackish water which may have been more
important before the Noordoost-Polder was reclaimed.
From the young age of the groundwater at 20-25 m below the present mire surface the rate of
infiltration can be estimated at more than 1 mm/d, assuming a porosity of 30% in the sand bed.
Since, from Table 5.1, the average difference between the hydraulic heads in De Weerribben
and its surroundings (Paaslo, Noordoost-Polder) is ((0.5-0.75)+(4.5-0.75))/2 = 1.75 m, while it
was roughly -0.5 m from 1920 until 1941, the average rate of exfiltration may have been about
0.3 mm/d in that period. This corresponds with a 6 m upward movement of the groundwater in
20 years. Similarly, for the period 1889-1920 (hydraulic head difference -0.35 m) the upward
movement may also be estimated at 6 m. These very rough estimates show that the earlier
influence of slightly brackish water may indeed have long persisted before the deep polders
were reclaimed. The polder Wetering-West, bordering De Weerribben in the southeast, still
discharges water with a raised chloride concentration.
With respect to the question of exfiltration of groundwater into the mire, the calcium and
bicarbonate concentrations provide another clue. The fresh groundwater is saturated with
calcium ions at a concentration of 100-120 mg/l. Since the bicarbonate concentrations in the
groundwater are consistently high, the carbonic acid content at the prevailing pH values are also
high and must drop considerably when the water becomes exposed to the atmosphere. This
71
should have resulted in the precipitation of calcium carbonate, but no traces of such calcareous
deposits have ever been reported from the peat in this area. This seems to confirm that
exfiltration of groundwater will not have been a significant phenomenon within the area as a
whole.
The increased local infiltration of boezem water is clearly demonstrated in Fig.5.6.
Obviously atmotrophic, lithotrophic, and thalassotrophic samples are rare in the collection of
recent samples. This is partially due to the changed composition in the B1 and B5 piezometers.
When only B2, B3, and B4 are considered, however, it is still obvious that the plotting field for
the recent samples is narrowed to a rather uniform type somewhere between the lithotrophic and
the molunotrophic (Rhine water) benchmark. Its still relatively high IR is due to the dissolution
of calcium from the sand bed.
5.6 The chemical composition of boezem water in the 1970s and 80s
Introduction
The surface water in the main canals of the boezem was repeatedly sampled and analysed in a
sampling program Weerribbennet (WRNET) from 1972 to 1982. The aims were:
1) A characterization of the chemical composition of the boezem water;
2) An identification of the origin of the boezem water;
3) More specifically: an answer to the question in how far groundwater influx could modify the
boezem water composition;
4) During the course of the investigations a description of the influence of water supply during
dry seasons was identified as an additional aim.
The Weerribbennet program was undertaken as a joint effort of the Research Institute for Nature
Management (RIN, field sampling) and the Hugo de Vries Laboratory of the University of
Amsterdam (HdV-Lab, chemical analyses). Originally 20 sampling stations were selected for
bimonthly sampling (Fig.5.8). Since the sampling stations nrs. 3 and 19 were abandoned during
the course of the study, and since the bimonthly scheme could not be strictly followed, 801
analyses became available, covering 46 dates. The analyses will be referred to by a number s.d,
where s is the station number (WRN1-WRN20), and d is the sampling day number (1-46). A
selection of results is presented in Tables 5.5 and 5.6. They were processed with the MAION
program (Appendix D). It appears that the station-average values (Table 5.6) are fairly similar
although remarkable differences may exist on single sampling dates. For this reason the spatial
comparison made hereafter is primarily based on patterns of variation with time.
The analysis of calcium concentrations often failed in 1973-'74. Only 614 samples passed
the MAION electroneutrality test with a rejection limit of 8% for the absolute value of (cations anions) as compared to (cations + anions). Values for the electrical conductivity (EC25) and the
chloride, phosphorus and nitrogen concentrations have been regarded correct for all samples; all
other analytical results have only been considered for the 614 above-mentioned samples. In
addition to the calculation of the ionic ratio IR for these samples, an estimation of IR was made
for all 801 samples available, using the MAION relation:
IRCl(%) = 100 - 1380 ([Cl-] / EC25)
72
(Appendix D).
Table 5.5
Average analytical results for WRNET sampling dates
n=
date
yymmdd
pH
Ca Mg
mg/l mg/l
1
2
4
5
7
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
(18)
(17)
(18)
(18)
(2)
(15)
(6)
(9)
(16)
(17)
(6)
(18)
(15)
(18)
(9)
(18)
(18)
(12)
(17)
(18)
(7)
(15)
(14)
(16)
(9)
(15)
(15)
(18)
(17)
(18)
(17)
(18)
(15)
(7)
(13)
(11)
(17)
(17)
(16)
(18)
(17)
(18)
720500
720700
721130
730100
730529
740128
740402
740809
741107
750314
750514
750800
751020
760121
760405
760600
760800
761000
770223
770426
770929
780309
780512
780724
781103
790412
790613
790831
791024
800306
800412
800716
800912
801128
810313
810604
810730
811210
820318
820527
820918
821021
7.0
7.0
6.8
7.0
7.5
7.1
7.5
7.5
7.4
7.9
8.0
7.6
7.4
7.2
7.3
8.0
7.8
7.4
7.2
7.3
7.5
7.3
7.7
6.9
8.1
6.7
6.9
7.3
7.3
7.1
8.4
6.8
7.5
7.0
6.8
7.0
7.1
6.6
6.9
7.0
6.7
7.7
66.2
47.6
51.2
60.4
34.5
25.6
33.5
45.6
48.6
40.6
39.5
50.4
51.1
56.9
44.9
47.9
56.9
52.2
72.4
55.3
40.3
46.9
43.4
43.8
52.4
44.3
47.9
73.9
72.6
54.1
74.5
48.3
69.1
51.4
39.6
80.4
75.2
40.3
45.3
75.7
57.8
65.8
Na K
mg/l mg/l
Cl
mg/l
HCO3
mg/l
SO4
mg/l
EC25
mS/m
IR x y pHsat
% % %
rLI
%
rAT
%
8.6 31.9 4.4 57.6

8.5 29.4 3.4 50.8
12.7 33.2 5.5 51.1
12.1 38.3 4.9 62.7
10.0 30.5 4.0 50.5
17.3 41.4 8.5 61.1
13.7 41.0 6.2 62.0
10.8 48.6 4.4 83.0
14.2 41.3 5.1 68.1
14.2 42.2 7.2 57.9
10.3 28.8 4.3 44.3
13.3 83.8 7.6 150.9
13.5 69.9 5.6 128.7
17.0 46.7 6.3 75.7
13.7 37.3 5.2 61.2
15.3 87.6 7.8 155.6
18.0 143.2 9.4 262.2
13.2 87.0 5.4 148.3
26.2 65.8 7.8 111.8
15.9 48.1 9.3 80.5
11.3 44.6 4.0 78.9
13.6 43.6 5.0 73.0
12.4 41.4 9.0 68.4
12.9 61.0 6.8 108.8
10.8 42.6 4.6 81.4
12.1 33.5 7.6 59.7
11.1 34.7 4.2 54.3
9.4 34.3 2.5 65.7
9.0 31.7 3.4 64.0
12.3 31.3 6.8 48.9
8.9 32.2 4.2 59.2
10.1 35.0 5.4 63.6
8.2 31.8 3.7 63.1
12.7 29.0 8.2 45.7
8.6 22.1 7.8 33.2
10.2 34.5 2.9 65.2
7.3 28.8 2.9 52.8
11.9 26.3 6.6 42.2
12.8 27.6 7.9 43.0
8.9 32.9 3.8 60.2
14.2 79.5 8.0 143.7
11.7 55.3 5.9 108.2
188.8
148.9
88.8
138.3
152.5
89.1
132.2
149.8
130.4
116.3
136.2
122.3
146.0
102.0
152.5
119.6
119.0
186.1
99.0
142.3
156.9
140.3
130.7
119.3
175.5
95.2
119.6
204.4
196.6
115.2
195.9
126.1
192.4
94.1
72.3
204.1
179.1
87.2
86.2
204.0
133.8
169.4
33.3
37.1
93.6
59.1
32.0
99.1
78.0
41.0
69.9
84.2
59.3
64.8
54.9
103.8
71.8
76.6
89.8
50.6
165.9
93.2
44.9
74.0
77.4
63.1
45.0
78.3
56.0
32.6
29.8
79.3
44.7
39.8
35.1
76.6
70.2
34.5
33.6
61.4
76.4
32.2
59.8
46.7
56.3
48.1
53.0
59.0
47.3
60.3
67.3
59.2
60.3
60.7
50.7
86.2
80.6
70.7
62.8
91.2
126.3
87.9
92.6
71.3
59.3
59.7
63.1
68.4
63.4
52.8
48.3
54.7
56.9
53.5
60.8
49.2
56.5
49.4
37.3
60.3
52.8
34.6
50.0
60.2
78.3
67.1
68
62
64
63
55
43
49
51
55
56
62
37
43
57
57
36
28
39
54
56
48
53
53
42
60
57
61
67
67
66
69
57
66
67
68
69
72
63
65
69
42
52
90
86
61
75
83
34
55
66
64
58
81
36
48
50
71
33
24
47
39
61
67
65
62
44
84
56
75
89
90
69
89
75
89
68
62
89
91
64
63
90
42
65
-35
-27
3
-17
-29
14
-7
-16
-10
-2
-18
3
-6
2
-15
4
4
-8
5
-8
-16
-5
-6
3
-23
7
-12
-33
-35
-4
-32
-18
-32
-6
9
-35
-34
0
-2
-36
1
-14
1
-2
4
5
-7
-5
-7
-3
0
-1
-6
-1
-2
4
-5
-2
-2
-5
4
-1
-7
-3
-4
-2
-6
0
3
2
2
4
1
2
0
7
4
5
5
5
6
3
2
0
-1
?
-4
0
4
7
17
0
2
6
6
3
5
4
7
5
3
2
0
2
3
-3
6
0
3
-2
-7
-12
-5
-3
-1
-5
-4
-5
-11
-6
-6
-30
1
0
-6
-9
7.62
7.86
8.06
7.80
7.97
8.35
8.07
7.87
7.92
8.04
7.96
7.94
7.86
7.98
7.87
7.98
7.93
7.75
7.93
7.84
7.90
7.90
7.96
8.00
8.71
8.12
7.95
7.56
7.58
7.96
7.57
7.93
7.61
8.03
8.27
7.52
7.58
8.15
8.14
7.55
7.84
7.69
rTH rRH
%
%
44
40
51
58
45
74
81
57
64
71
55
80
76
76
71
83
88
77
84
74
60
63
74
71
62
56
43
35
41
53
49
44
42
45
24
44
35
9
53
47
72
59
72
71
76
85
73
85
91
86
88
89
79
99
96
91
89
99
97
97
91
90
87
90
93
97
94
83
77
66
70
77
74
79
71
71
55
71
65
49
76
73
98
90
Ca Mg Cl
% % %
60
53
50
52
44
27
37
43
44
39
47
34
38
44
44
31
26
35
41
44
41
43
42
36
54
46
48
61
62
51
62
49
61
51
51
63
66
47
48
62
37
48
13
16
20
17
21
30
24
17
21
22
21
15
16
22
22
17
14
14
24
20
19
21
20
18
18
21
19
13
13
20
13
17
12
20
19
13
11
23
22
12
15
14
30
31
30
34
31
33
31
40
36
31
27
56
50
36
30
55
66
50
38
34
38
35
34
48
41
34
33
31
31
28
28
38
31
29
26
31
29
31
29
29
54
45
The values in the columns IR-Cl(%) represent averages too, they differ slightly from values computed on the basis of the average concentrations
Explanation of items: Appendix D; x electro-neutrality error; y conductivity error; pHsat pH at which satured with respect to calcite, at 10 C; Ca, Mg, Cl (%)
f(c)1/2Ca2+, f(c)1/2Mg2+, and f(c)Cl-, respectively
In the 811210 samples EC measurements were systematically in error
The data from the water analyses show much redundancy. A separate analysis of correlations
was carried out for time series and for spatial patterns. In order to extend the analysis of the time
series over the period 1982-'87 chloride and conductivity data reported by the regional water
quality board (Zuiveringschap West-Overijssel 1981-'88) for the sampling station Kalenberg,
have been used.
In Section 5.7 comparison is made between the WRNET results and earlier analyses,
selected from reports and RIN archives.
73
Table 5.6
Average analytical results for WRNET sampling stations
station
n=
pH
Ca
mg/l
Mg
mg/l
Na K
mg/l mg/l
Cl
mg/l
HCO3
mg/l
SO4
mg/l
EC25
mS/m
(27)
(35)
(35)
(30)
(31)
(35)
(36)
(38)
(37)
(37)
(35)
(37)
(33)
(35)
(33)
(32)
(36)
(32)
7.2
7.2
7.2
7.2
7.3
7.3
7.2
7.1
7.2
7.1
7.3
7.2
7.2
7.5
7.3
7.4
7.3
7.5
62.8
54.4
51.9
64.9
65.7
56.8
52.1
51.6
53.1
51.3
49.1
48.8
52.2
56.1
53.7
56.1
46.7
59.9
12.0
12.2
11.8
12.7
13.4
12.5
12.6
12.2
12.8
13.4
12.6
12.8
13.8
13.3
13.1
13.0
11.6
8.7
50.0
48.5
47.9
50.1
50.1
49.5
48.2
45.4
45.6
44.7
47.3
45.8
48.0
46.7
46.7
47.3
44.5
35.6
89.9
84.1
83.2
90.5
89.5
88.2
83.8
78.7
79.6
77.2
81.8
79.3
81.0
77.0
79.4
81.1
76.9
65.6
165.2
146.1
139.3
168.8
173.2
148.3
140.6
136.3
133.5
117.4
120.8
115.9
122.0
143.3
128.3
136.9
121.8
155.4
52.2
55.4
56.9
58.8
65.4
62.3
62.5
60.8
67.5
76.0
68.6
71.5
78.1
71.9
71.2
69.4
57.6
46.3
67.5
63.9
62.7
69.2
70.1
65.6
64.2
61.5
62.4
62.4
62.5
61.3
64.1
64.2
62.7
64.6
58.1
58.0
1
2
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
20
5.1
5.8
6.0
5.5
5.8
5.9
5.8
5.6
5.7
6.1
6.0
6.2
6.7
5.9
6.5
6.5
5.6
5.3
IR x y pHsat
% % %
rLI
%
rAT
%
56
55
54
56
58
55
54
55
56
56
54
55
57
59
58
57
54
66
69
66
65
68
68
65
64
63
63
59
59
60
57
66
63
65
63
81
-17
-13
-12
-16
-15
-11
-11
-9
-9
-5
-5
-5
-5
-14
-9
-11
-9
-27
1
1
0
1
0
0
0
0
0
1
0
1
1
1
1
1
1
0
-1
0
0
-1
-2
-2
0
-1
-1
0
0
0
-1
-1
-2
-1
-1
2
7.75
7.85
7.89
7.74
7.73
7.83
7.88
7.91
7.91
7.98
7.97
7.99
7.96
7.84
7.90
7.86
7.98
7.76
rTH rRH
%
%
60
59
59
61
62
60
61
57
57
58
60
58
58
56
57
60
54
46
86
85
85
85
84
85
85
82
81
81
84
82
79
78
80
83
81
70
Ca Mg Cl
% % %
49
46
45
49
49
46
45
46
46
45
44
45
45
49
47
47
45
59
16
17
17
17
17
17
18
18
18
19
19
19
19
18
18
18
18
13
39
38
39
38
36
38
38
37
36
36
38
37
36
34
36
36
38
31
The values in the columns IR-Cl(%) represent averages too, they differ slightly from values computed on the basis of the
average concentrations
Explanation of items: Appendix D; x electro-neutrality error; y conductivity error; pHsat pH at which satured with
respect to calcite, at 10 C; Ca, Mg, Cl (%) f(c)1/2Ca2+, f(c)1/2Mg2+, and f(c)Cl-, respectively
Although the time series reveal differences between the various stations, one dominant pattern is
found in all of them. This dominant pattern is discussed first.
The dominant pattern of the time series
According to the high correlations, the dominant pattern of fluctuation of the chemical
composition is sufficiently reflected by the average for all stations (Table 5.5). Fig.5.7a
illustrates the fluctuation of EC25 and IRCl, as an approximation of IR. EC25 was especially high
during the drought of 1975-'76. IRCl was low in this period. In general, EC25 and IRCl are
negatively correlated during water intake, while they show a slight positive correlation in other
periods. Sulphate, phosphorus, and nitrogen show an irregular seasonal rhytm, which dominates
over other variations.
Spatial patterns and asynchronous variations
Spatial patterns are different at the various sampling dates. A study of the (a)synchronity of
variations enables an ordering of the stations on the basis of the locally observed time series. The
spatial distribution of the average values of some variables (Table 5.6) is shown in Fig.5.8a-d.
For chloride and sulphate the patterns are more or less complementary: the chloride
concentrations are high in the south-east (near WRN5), where polder surplus water enters the
area, while high sulphate values are especially found in the north (near WRN15), where water is
taken in during droughts. This water originates from the Frisian boezem, which in turn is
supplied with IJsselmeer water, for which the river Rhine is the main source (see Chapter 4 for a
description of the area).
74
Fig.5.7
Fluctuation of EC and IRCl (a), and of [Ptot], [Ntot], and [SO4=] in De Weerribben
1972-'82, average for all stations; Note the different scales for the various constituents (concentrations
and logarithms of concentrations in mg/l); The highlighted areas in Fig.5.7a, where EC25(mS/m)>IR(%),
indicate the presence of molunotrophic water, and they systematically co-occur with an inlet of
substantial quantities of water from elsewhere (> 7 106 m3 at each occasion)
75
Fig.5.8
Mean values of [Ntot], [Ptot], [Cl-], and [SO4=] at the various sampling stations in
De Weerribben (nr)
The patterns for phosphorus and nitrogen are different. There is a tendency for high values in the
main canals, but this is not consistent.
The spatial pattern of mean values of EC25 is equal to that of chloride. This factor is highest
at WRN5 and WRN6, situated near the entry point of the Wetering, a canal which receives
76
Fig.5.9
Correlations between time series of EC25, with WRN5 (a) and WRN15 (b) as
leading points, and standard deviation of EC25 (c)
the discharge of some large polders to the south-east of De Weerribben. The lowest mean value
was recorded at WRN20. The standard deviation of EC25 (Fig.5.9c) varies but little, but it is
lowest at the WRN5 end of the area and highest at the WRN15 end.
A cartographic representation of the correlations between time series (Fig.5.9a) reveals that
the variation of EC25 at sampling stations in the neighbourhood of WRN5 (WRN6,1,7,2,4) is
strongly correlated with that at WRN5, i.e., EC25 varies synchronously in this area with WRN5
as a possible source of the pattern of variation. The lower coefficients of correlation at the
other end of the area (WRN15) suggest that the variation is of a different kind there: the value of
the coefficient of correlation decreases with the distance from WRN5. Fig.5.9b, with WRN15 as
a point of reference, consequently shows a complementary pattern: high correlations in the
neighbourhood of WRN15, and low ones near WRN1,5,6. No indications were found that this
could have been caused by a simple time-lag.
77
Fig.5.10
Time series of EC25 and IRCl at WRN15 (a), WRN5 (b), and WRN20 (c)
Highlighted areas, where EC25(mS/m)>IR(%), indicate a presence of molunotrophic water due to inlet
or polder discharge; compare Fig.5.7a for the all-stations mean
78
The variation of total phosphorus (Ptot) and total inorganic nitrogen (Ntot) at the various
sampling stations is less strongly correlated than that of the other parameters and it is as yet
unsufficiently understood.
It is concluded that the variation of EC25 and most major ions is ruled by regional processes
while the variation of Ptot and Ntot rather seems to be governed by local processes. Since, in the
present report, the emphasis is on the regional processes, the Ptot and Ntot results will not be
treated in any greater detail here. The time series observed at WRN 5, 15, and 20 are shown in
Fig.5.10 for EC25 and IRCl as a basis for the following discussion of distinct regional influences.
Surface water flow through the area (Fig.5.10a, 5.11)
Apparently the water entering De Weerribben near WRN15 is an important source of variation
in EC25. The strong overall correlations of EC25 are probably caused by the circulation of
substantial quantities of water with a characteristic EC25 through the area. This circulation is
driven by pumping, water inlet, and winds (Lyklema & Van Straten 1977, Jol & Laseur 1982),
and it also penetrates deeply into secundary canals and locally residing bodies of water.
The overall effect of such a circulation is that an incoming body of water with a strongly
different EC25 can usually be traced nearly everywhere in De Weerribben at the next WRN
sampling day, i.e., within 1-3 months. This effect is shown in Fig.5.11, which also includes
analyses of samples collected in the southern part of the boezem area, the mire area De Wieden.
Apparently, at this date the sampling stations in De Wieden had not or not yet been reached by
the inlet water. Note that this may be partially due to the inlet strategy. When, shortly after water
intake, it starts raining, the pumping station Stroink is used to remove the water surplus, so
attracting the water towards the pumping station, which is located in between De Weerribben
and De Wieden (Chapter 4). Shortly after pumping the water level near the pumping station is
still somewhat lower than elsewhere. An eventual intake at such a moment will meet the same
attraction. In the early Seventies it was common practice to overshoot the target levels with the
pumping and intake measures. In feed-back terms, the authorities, as it were, prepared for a
continuation, or even worsening, of the threat to the target level, thus diminishing the retention
time of water in the boezem area. This does, of course, not mean that all water in the main canals
and ditches is entirely replaced: the analysis of time series shows a considerable lag in the
changes in chemical composition of the water in more isolated parts of De Weerribben.
At present a more prudent strategy is followed. Rather than overshooting the target levels,
just enough water is removed or let in to conserve the highest or lowest allowed water levels
during pumping or intake periods, respectively, so conserving the local body of water as much
as possible. Of course a still greater effect could be achieved if the discharge and inlet works
were located at the same side of the area.
The inlet water, flowing in through WRN15, originates from the Frisian boezem system. The
latter contains surplus water from several polders in the province of Friesland and inlet water
from lake IJsselmeer, i.e., ultimately, Rhine water. It can be recognized at WRN15 by an EC25
value (mS/m) numerically exceeding IR (%), as shown in Fig.5.10a. Especially during drought
periods the inlet water may be even slightly brackish, and it was so in the summers of 1975 and
1976.
79
80
Fig.5.11
The inlet water (March 1975, in De Weerribben) can be traced by its lower similarity to the lithotrophic
benchmark sample (rLI, a), and higher similarity to the thalassotrophic (rTH, b) and molunotrophic (rRH, c)
benchmarks, as compared to the mire water still present in De Wieden
Inflow of surplus polder water (Fig.5.10b)

The relatively high and constant values at WRN5 are due to the inflow of discharge water from
the polders in the south-east. This discharge water can often be recognized by its opaque,
greenish-brown appearance, and strongly resembles, both chemically and visually, the polder
water at the intake of the pumping station concerned (Rengersen 1979, Hack 1973). It contains a
fair amount of groundwater, mixed with boezem water which seeps into the deep polders from
where it is recycled into the boezem. In the inlet canals of the pumping station of these polders
it could be noticed that the surplus water of the polder Wetering-West is characterized by high
chloride concentrations, which must be due to the influence of slightly brackish local
groundwater (Section 5.5). This water is characterized by EC25 (mS/m) in excess of IR (%). At
WRN5 its presence alternates with that of the inlet water (Fig.5.10b).
The lithotrophic influence of the Steenwijk-Ossenzijl canal (Fig.5.10c)
The Steenwijk-Ossenzijl canal (WRN20) is usually filled with water rich in calcium and
bicarbonate. Most of this water originates from the Drenthian catchment area of the Steenwijker
A rivulet, along with a share of surplus water from the polders to the north-east of De
Weerribben. Most of these are shallow polders. The influence of the WRN20 water reaches De
Weerribben mainly via WRN15, but it is only an important source of water for the boezem
during moderately wet periods. In still wetter periods the boezem-discharge pumping station
Stroink causes a flow of canal water in the opposite direction (Jol & Laseur 1982, Lyklema &
Van Straten 1977). During extreme droughts the inlet water at WRN15 also reaches WRN20, as
can be traced by EC25 (mS/m) exceeding IR (%) (Fig.5.10c).
The lithotrophic influence of discharging groundwater
There is no evidence of any significant influx of groundwater into the De Weerribben mire area
proper. This conclusion is well in line with the conclusions from the groundwater survey.
An extension of the time series into the 1981-'87 period
While the WRNET monitoring system was in existence, the water quality board West-Overijssel (ZWO) set up a sampling scheme in a large area, including North-West Overijssel. The
sampling station K158 in Kalenberg is very close to the WRNET station WRN17.
The time series for EC25 and IRCl from both stations have been combined in Fig.5.12, where
K158 data were used from 1981 onwards. It is apparent that, since 1982, EC25 has stabilized at
about 50 mS/m, while IRCl slightly increased and stabilized at about 60%. Short periods of
considerable water intake, however, are clearly visible by a higher EC25 and a lower IRCl, as, for
example, in the summer of 1986.
Summary of water quality influences during the 1970s and '80s (Fig.5.13)
A summary of water quality influences in the surface water system of the boezem is given in
Fig.5.13. The extent of the average spheres of influence of waters from different sources was
calculated on the basis of a principal component analysis for 1972-'78 WRNET data. This
picture not only depends on the weather conditions and total amounts of inflow, supply, and
81
Fig.5.12
Fluctuation of EC25 and IRCl at Kalenberg, 1972-'81 (WRNET, a), and 1981-'87
(Zuiveringschap West-Overijssel, b)
Highlighting as in Fig.5.7a and Fig.5.10
discharge of water during that period, but also on the hydraulic parameters of the canals network
and operational aspects of the water management.
It must be concluded that the 1972-'78 period is strongly influenced by the dry summers of
1975-'76. Consequently, the processes and mechanisms seen during this period can only be used
in a more general way if it is possible to formulate a functional relation between the observations
and the relevant conditions. Among these conditions the resistance in the various canals (aquatic
macrophytes, strongly decreased in abundance since 1972!), the amount of water lost by seepage
from the boezem (polders, drinking water supply), and the inlet strategy (prudent since ca 1980
versus overshooting before) must be considered, together with the weather (droughts, rainy
periods, wind influence). In the absence of more and, preferably, longer series of observations it
is difficult to interpret the changes noticed during the 1980s.
82
Fig.5.13
Summary of water quality influences in the surface water system of De

Weerribben, 1972-'78
From Van Wirdum 1989, based on a 1978 advisory note to Zuiveringschap West-Overijssel
5.7 Comparison of surface water composition 1960-'82

Data sets OLD and WRNET
A comparison was made between the WRNET records and earlier analyses from the boezem
water in De Weerribben. The earliest data from which IR could be reliably calculated and whose
EC values are known are from 1960. In the archives of RIN and HdV-Lab 93 such analyses
were found (data set OLD). Since these analyses do not fit in time series the comparison with
WRNET results is based on the distribution of similarity values and on EC-IR diagrams.
From Fig.5.14 it becomes apparent that lithotrophic water is dominant in the OLD set.
Although the frequency distribution of rLI (the similarity to calcareous groundwater, see
83
Fig.5.14
84
The frequency distribution of the similarity to the litho-, atmo-, thalasso-, and
molunotrophic benchmarks; data sets old (93 1960-'70 analyses) and new (614
1972-'82 analyses)
Fig.5.15
EC-IR and rTH-rLI diagrams showing boezem water analyses from De

Weerribben
The analyses demonstrate a litho-thalassocline in 1972-'82 (a,c), and a litho- to slightly molunotrophic
cluster (b,d) in 1960-'70; L, A, T, M: Litho-, atmo-, thalasso-, and molunotrophic samples, respectively;
Mixing contours and the line IR(%)=EC25(mS/m) added to Fig.5.15a,b for convenience (Appendix D)
Appendix D) also shows a peak between 75% and 100% in the WRNET set, lower values are
more common in the latter. These analyses with a low rLI value cause the increased presence of
high values of rTH (similarity to sea water) and rRH (similarity to polluted Rhine water) in the
WRNET set, as compared to the OLD analyses (Fig.5.14a,b). This is partially due to the
presence of the extremely dry seasons 1975 and 1976 in the WRNET sampling program. The
higher rTH and rRH values in WRNET were especially found between 1975 and 1978.
However, many of the WRNET samples not considered here for reasons of electroneutrality
discrepancies, would, according to their IRCl and EC values, also fall in the high rTH and rRH
classes, especially the analyses of 1973-'74. It can be concluded that the molunotrophic water
type, resembling Rhine water (high rRH), is rare in the data set OLD, while it is abundant in the
WRNET set. This is due to the fact that only since 1972 the water from the Frisian boezem
system is supplied to the boezem system under investigation (see Chapter 4). This increased the
influx of water from the IJsselmeer in De Weerribben.
85
The plotting area for boezem water in the EC-IR and rTH-rLI diagrams has been indicated in
Fig.5.15a,c on the basis of all WRNET sampling dates which yielded reliable analyses. The
boundary line EC25(mS/m) = IR(%) was also drawn in this diagram since this condition can be
easily scanned in the time series shown for WRNET results. The OLD analyses are shown in
Fig.5.15b,d. The same trend is reflected by these diagrams: molunotrophic (RH) water was
considerably less abundant in De Weerribben before 1972.
Discussion: Did anything really change?
As introduced before, the main aim of this part of the study was to describe the prevailing
situation as regards the hydrology. The chemical composition and the origin of the boezem water
have been described, and no exfiltration of groundwater from the underlying aquifer was found.
Chemically, the surface water was akin to groundwater (lithotrophic) during the 1960s. In the
WRNET monitoring scheme the Steenwijk-Ossenzijl canal (WRN20) could be indentified as a
source of lithotrophic water, mainly carried from the bordering Drenthe Pleistocene area.
Only in the course of the investigation, ca 1974, it became obvious that a major change
might have occurred while the observations were carried out. Lithotrophic surface water was no
longer common in the area after 1973. In 1973-'74, many water analyses showed large
deviations from electroneutrality, so they must be considered unreliable, but the trend could be
confirmed in 1975 and 1976 with reliable analyses. While the monitoring program WRNET was
continued, the question arose, ca 1979, whether anything really had changed, or whether the
extreme conditions of a very wet winter (1974-'75), followed by two extremely dry seasons
(1975, 1976) had interfered with the normal situation.
First of all, the aquatic macrophyte vegetation in De Weerribben changed considerably. It is
not easy to date this change exactly, but, estimating from my own observations and from aerial
photographs this must have been after 1971 and before 1975. This may have been an effect of
some environmental change, possibly relating to the hydrology. In itself it has diminished the
hydraulic resistance in the canals network, and so facilitated the penetration of inlet water. This
change seems to have occurred somewhat earlier in De Wieden.
Also the change of the water intake from Beukerssluis, far south from De Weerribben (see
Fig.3.3) to the Linthorst-Homan sluices, right north of De Weerribben near Ossenzijl, is not well
documented, but it is generally agreed that the Linthorst-Homan sluices gradually became the
main intake point in the period 1972-'74. This has been a major change for De Weerribben, as is
shown in this chapter. During the 1960-1972 period the water quality at the intake point
Beukerssluis had badly worsened, providing a possible explanation for the earlier vegetational
effects in De Wieden, which is immediately influenced by water intake at Beukerssluis.
Also in the 1972-'75 period a land reallotment program was executed in the IJsselham polder
area, where lower drainage levels were installed, and in the bordering Paaslo morainic area. The
effects were not investigated, but it is almost certain that this program has diminished the
summer discharge of lithotrophic water in the Steenwijk-Ossenzijl canal. Increased drinking
water withdrawals may have contributed effects too.
These changes will inevitably have increased the chance that, during droughts, large
amounts of molunotrophic water had to be taken in and could penetrate far into the mire area.
So, although such events as shown in the analyses between 1977 and 1987 could have been
easily missed by the 1960-'70 water samples, and however extreme the 1975-'76 summers were,
the suggested difference probably exists. The deep penetration of molunotrophic water from the
boezem into the underlying body of groundwater seems to prove this difference.
86
All this does not prove that the 1960-'70 analyses are representative for a long period. Until
1920 brackish influences occasionally spread over De Weerribben. It is possible that no kragge
was present at that time in most petgaten which are now grown with a vegetation characteristic
of a lithotrophic environment, or that intervening periods with fresh water were long enough to
determine the nature of the vegetation. As discussed before (Chapters 4 and 5.3-5.5), the
installation of the pumping station Stroink and the construction of polders in the 1930-'55 period
have strongly changed the hydrology of the area. The Wetering polders may well have
discharged a substantial amount of lithotrophic water into the boezem during the late 1950s and
the 1960s. These polders were reclaimed from a large mire area, most probably comprising a
body of supposedly lithotrophic surface water and groundwater. A changed chemical
composition of the polder discharge water can be due to the exhaustion of this body of water, the
attraction of brackish groundwater from greater depths, and the increased agricultural
fertilization.
It must be concluded that much has changed, but that, during the development of the present
vegetation, a hydrological steady environment, if any, never was present longer than some ten
or twenty years. A comparison between a new and an old steady state is therefore not
realistic. Yet, during the development of the area as it is now, the decreased availability of
unpolluted lithotrophic water in the surface water system since about 1970 must be
considered a significant change. The hydrology suggests that this change became
dramatically evident by the coincidence of the extreme droughts of 1975-'76, the changed
water intake point, and the strategy of overshooting target values with the water
management.
87
CHAPTER 6
The distribution of seepage indicators in De Weerribben
6.1 Introduction
The reputed seepage indicators are not especially indicative of groundwater
outflow as such, but rather of low-lying, base-rich, yet nutrient-poor sites where
fen peat has developed under the continued influence of an external water supply
(Chapter 3). In Chapter 5 it appears that groundwater outflow is no characteristic
feature of De Weerribben at present. Chemically, the water in the surface system
determines most of the variation, both in time and in space. It was shown that inlet
and discharge, as related to weather conditions and human decision, resulted in
drastic fluctuations of the water quality. Slightly brackish conditions must have
obtained in part of the mire area in the beginning of the present century. In the
1960s, a base-rich, lithotrophic type of water became dominant in the area, but this
was followed in the 1970s by an increased supply of polluted, molunotrophic
(Chapter 5, Appendix D) water originating from the river Rhine. In the late 1970s,
and especially in the 1980s, more favourable weather conditions and a more
careful operation of discharge and inlet stations led to a return of the lithotrophic
character, possibly, however, with a higher nutrient load now prevailing in the
reserve than was observed during the 1960s.
Seepage indicators in De Weerribben
89
Although this information weakens the original theory of groundwater outflow and
seepage indicators, a further environmental explanation of the distribution of these
indicators requires a formal analysis given in this chapter. After an introduction to
the overall vegetation cover and to the various factors influencing it, this analysis
is carried out by raising some relevant questions. As a seepage indicator, the
greatest attention is paid to Scorpidium scorpioides. Firstly, the basic question:
Does the distribution of seepage indicators in De Weerribben coincide with the
reputed former seepage area?, is tackled and this is denied. In the following section
it is argued that the distribution of seepage indicators is partially restricted to a
particular area, for which the geographical evidence suggests a relatively high base
state and certain other favourable factors. The species are almost absent from the
central part of De Weerribben, i.e., the part of the former bog peat area that was
never subject to clay deposition, has reached a far stage of terrestrialization, and
missed regular management by cutting or mowing.
In the next section the environment of stands of vegetation with seepage indicators
is analysed with the help of the indicator table (Appendix C) treated in Chapter 3.
It appears that this indication is similar to what was already concluded in Chapter 3
for the seepage indicators themselves, though less extreme. It is shown that, even
when seepage indicators themselves are not counted as ecological indicators,
litho-oligotrophic environmental conditions are slightly more important in stands
of vegetation with seepage indicators, while stands without seepage indicators
point to a more atmo-oligotrophic milieu.
So far the dynamic behaviour and the possible significance of environmental
change was not taken into account. From an analysis of the distributional pattern of
seepage indicators in different periods, the suggestion emerges that these patterns
have altered. A correlation is found with a complex of factors, in which the
desalinization of the south-western part of De Weerribben and the progressing
succession of the mire vegetation stands candidate for a partial explanation. This is
corroborated by an analysis of the distribution of some other plant species. The
seepage indicators are ranked according to their suggested response to the changes
involved, from the relatively fast colonizers Liparis loeselii and Scorpidium
scorpioides to the more prudent Utricularia intermedia and Parnassia palustris,
Menyanthes trifoliata being intermediate. Equisetum fluviatile and Pedicularis
palustris can possibly be ranked between Scorpidium and Menyanthes.
The analysis of the dynamic behaviour of Stratiotes aloides additionally yields an
indication of the role of the molunotrophic type of water invading De Weerribben
in the mid-1970s. The recent recurrence of the lithotrophic water features may
explain the present resettlement and expansion of Stratiotes aloides. The same
factor, viz., an increased lithotrophy, may also influence quagfen vegetation,
although it remains to be seen whether the nutrient state of the mire water is still
low enough for a full development of the Scorpidio-Caricetum diandrae.
6.2 Seepage indicators in the vegetation of De Weeribben
The vegetation of De Weerribben largely consists of fen vegegation in terrestrializing petgaten.
Especially stands dominated by Phragmites australis and, to a less extent, Typha angustifolia,
and by Sphagnum species are abundant. Carr vegetation and brownmoss reeds, the latter often
with species of Carex and with Juncus subnodulosus co-dominant in the herb layer, are also
90
wide-spread. The local dominance of species is strongly influenced by the management regime.
The species composition of the stands of vegetation reflects gradual differences in the local
milieus, as produced by the almost unlimited variation in the combinations of hydrological,
managerial, historical, and other factors. Accordingly, ordination of more than 900 records of
the vegetation cover has not led to the separation of clearly distinct types, but gradients could be
derived and the presence of seepage indicators appeared to be correlated with the main
gradients. In respect of the present problem I have therefore grouped the records in accordance
to the presence or absence of seepage indicators (see Section 6.5).
Reed cutting in the winter season favours Phragmites australis, while mowing in summer, or
cutting in early autumn, suppresses the growth of Phragmites, in favour of species of Carex.
Irrigation of reed beds eventually leads to a firm kragge, growing up to the flooding water level,
and to a substantial increase of Carex elata. Irregular mowing and burning of the vegetation
cover benefit the growth of Molinia caerulea, and further abandonment ultimately leads to carr
vegetation. Locally, vegetation with hummocks of Sphagnum species and Polytrichum
juniperinum ssp. alpestre, rich in ericaceous shrubs, has developed, but it is still uncertain
whether such vegetation is stable or not, and under what conditions it may develop.
Pure stands of the Scorpidio-Caricetum diandrae are uncommon in the quagfens of De
Weerribben, although, with De Wieden and the Vechtplassen mire area, this reserve holds the
best and most extensive representatives of the Association left in The Netherlands. Almost
certainly De Weerribben is the main area of Scorpidium scorpioides in The Netherlands today.
Apart from the seepage factor, the full development of the Scorpidio-Caricetum diandrae
depends on a strict regime of mowing in summer. The stage of terrestrialization is also
important, but it has been shown that the stands are connected in a quagfen vegetational series.
The earlier stages described by Meijer & De Wit (1955) from the Vechtplassen area
(Kortenhoefse Plassen) are only rarely met in De Weerribben and De Wieden, since, in these
areas, quagfen vegetation develops more or less synchronously in the petgaten. Locally,
however, the succession can be observed through the years along the shore of terrestrializing
broads. Examples have been given by Segal (1966), by De Wit (1951) and by Kuiper & Kuiper
(1958). In time, stands of the Scorpidio-Caricetum diandrae may develop into a species-rich
type of grassland, reckoned to the Cirsio-Molinietum Association, or into a community of
Sphagnum and ericaceous shrubs. The character species of the Scorpidio-Caricetum diandrae
also occur in stands where Phragmites australis is dominant, but it is uncertain whether this
dominance of reed should not be regarded as an aberrant feature due to disturbance. Especially
in the northern part of De Wieden and in De Weerribben, Phragmites australis and some other
helophytes atypical of the Scorpidio-Caricetum diandrae proper are often found.
The ups and downs of economic returns from reed culture and changes in nature
management often lead to a local alternation of different management regimes: a once
abandoned reed bed, overgrown with Molinia caerulea and groups of young trees, may be
redressed by the cleaning of ditches, by sod cutting, removal of trees, deposition of ditch mud,
and by irrigation. Needless to say such measures drastically change the species composition and
dominance relations, and they complicate an accurate ecological explanation of the momentary
features of any actual stand of vegetation.
Overall vegetation maps of De Weerribben have been made by Van Zon-Van Wagtendonk
(1969) together with H.N. Leijs in 1968, and by Staatsbosbeheer (1975). These maps have a
physiognomic basis; the types of vegetation recognized are based on the dominance of structural
groups and single species, such as Phragmites australis, Juncus subnodulosus, Molinea
caerulea, scrub, aquatic macrophytes, etc.. Rare species and bryophytes have not been given any
diagnostic importance, or have even not been recorded at all. Such maps are quite indicative of
91
the management situation, and they are useful for the tracing of such processes as the increase of
stands of carr in time. They do not yield an accurate picture of, for instance, the occurrence of
the types of vegegation with alleged seepage indicators. The Staatsbosbeheer report (1988)
contains a somewhat more detailed vegetation map, based on field work by H.A.J. Koenders,
G.J.R. Allersma, J. Schreurs, R. Veldkamp, and J.H. van Slogteren in 1985-'86; several rare
species and mosses were included in the survey. This map refers to the situation about ten years
after the investigation reported here, and its further use would require a separate study,
especially since the map was reproduced in the form of 13 not easily surveyable leaves. Both
Van Zon-Van Wagtendonk and Staatsbosbeheer include distribution maps of selected species,
which have been used to improve the interpretation of distributional patterns given below.
A more detailed description of the vegetation in two NO-SW strips across De Weerribben,
and in some additional complexes of petgaten was made as part of the present study (e.g.,
Oosterbroek & Post 1977, Verschoor 1978). In contrast to the traditional French-Swiss method
of field survey for vegetation mapping, the vegetation was recorded in the form of maximum
areas, rather than minimum areas, in order to let the kragge-level of vegetational variation
prevail over more fine-scaled levels of species aggregation (Chapter 3). The maximum area
method provides a complete field discription of every map element. Map elements are
predefined on the basis of aerial photo analyses, but they may be subdivided or united if the
homogeneity, defined by a recurrence of vegetational patterns within each map element, renders
this necessary. Due to time limitations, and to the extremely complex vegetational patterns in the
study area, we were unable hierarchically to structure these mapping surveys by primarily
defining aggregate types at the finer scales, subsequently to describe the map elements as
composites of these aggregate types, as is done, e.g., in the traditional Finnish and Russian mire
survey and in the method described by Succow (1988). For this reason, species association in
our records of mapping elements may be due to species aggregation at any level described in
Chapter 3.
These more detailed surveys have provided a good, but not easily transferable, insight into
the vegetational variation. The vegetation records are especially important since they represent a
description of the whole statistical population of stands of vegetation in the strips mapped, and,
so, can be considered a vegetation sample of De Weerribben as a whole, which is unbiased as
regards the sympatric occurrence of seepage indicators and other species. This feature will be
used in this chapter to find out whether these other species indicate a difference between the
milieu of stands with and without seepage indicators, respectively.
More direct inferences about the type of environment preferred by the supposed seepage
indicators could be made by mapping the vegetation units with and without seepage indicators,
within single types of vegetation (Verschoor 1978). The seepage indicators appear to occur
most frequently in a zone of vegetation in the proximity of open surface water, but not in the
immediate reach of the water of the larger canals.
This point will be discussed in more detail in the context of a case study in De Stobbenribben
(Chapters 7-10), where several, possibly relevant environmental factors have been measured.
The mapping surveys, and observations during various terrain visits, raised doubts about the
restricted distribution of accepted seepage indicators as reported by Kuiper & Kuiper (1958).
92
Fig.6.1
Some important physiographic boundaries in De Weerribben, including the

seepage boundary after Kuiper & Kuiper (1958), and the generalized pattern of the
stage of terrestrialization in 1951 (after Haans & Hamming 1962)
The areas traversed for detailed mapping of the vegetation in the present project have been indicated by
hatching
Therefore, this distribution is covered in more detail in the next section. It is of course possible
that the distributional limit suggested by Kuiper & Kuiper held true in the 1950s, but if it can be
proven that the present areal extent of the supposed seepage indicators reaches far beyond their
seepage area limit, the present information on the hydrology of De Weerribben (Chapter 5)
renders a relation with any local groundwater outflow most improbable at least. Below, the
distribution of selected seepage indicators is reported, first with respect to the groundwater
outflow question, and then in order to formulate alternative hypotheses regarding the ecology of
these species. Some inferences are made on the basis of a survey of the correspondence between
the occurrence of species and the successional stage in 1951 (Haans & Hamming 1962; see
Fig.6.1 for the general pattern).
The country-wide distributional pattern of the species involved is not treated here.
Distribution maps are available of the vascular plants and the Musci, with the exception of the
genus Sphagnum. For the Musci, however, the information is incomplete, since it is based on
herbarium specimens only. The general pattern is that most of the seepage indicators are
93
especially found in the valleys of small rivers in Pleistocene areas, in dune slacks, and in fen
areas bordering on the Pleistocene plateaus. While this seems in line with the seepage
hypothesis, there are at least two weakening points. Firstly, many of the species involved are
relatively rare and, so, provide too few data for any statistical treatment, especially not so since
the environmental information from the locations is poor. Secondly, there are but few fen areas
without brackish influences outside the preferred distributional zones, so that the importance of
the seepage factor can not be tested there, and several of the seepage indicators have been
locally observed in the few other freshwater fen mires, e.g., in the Nieuwkoopse Plassen mire
area in the Province of South-Holland.
6.3 Is the distribution of seepage indicators restricted to an area of groundwater
discharge?
Published evidence of seepage in rich-fen quagmires (quagfens, Chapter 2) shows that many
authors carried out their research in mire sites in the proximity of more elevated areas, thus
suggesting a positive relationship between seepage indicators and a possible discharge of
groundwater from the mineral subsoil into the overlying mire. One way to shed some light on
this question is a mapping of the distribution of these seepage indicator species (Chapter 2) over
De Weerribben, an area shown (Chapter 4) to comprise large parts without a former
groundwater outflow, as well as a roughly indicated area of supposed groundwater discharge
(seepage area, Kuiper & Kuiper 1958). This seepage area includes the De Stobbenribben mire
complex, mentioned as a representative example of the influence of groundwater outflow on the
vegetation by Segal (1966). Although De Stobbenribben is an area of groundwater recharge at
present (Chapters 5 and 7-10), there is no proof that groundwater did not discharge into it
formerly. The local vegetation stand could still reflect a former environmental state. The area
delimited by the borderline of supposed former groundwater outflow phenomena in Kuiper &
Kuiper will be referred to as (supposed former) seepage area in the following treatment.
From a comparison with the soil map by Haans & Hamming (1962), it appears that the
border of the supposed former seepage area is located in the former Carex peat area at a distance
of some hundreds of metres from the transition to the former Sphagnum peat area. In the
application of formal statements about the seepage hypothesis, I have considered sites of
seepage indicators in the in-between area as belonging to the supposed former seepage area,
thus giving the seepage hypothesis the benefit of the doubt.
During this survey, Scorpidium scorpioides was selected as a guide species for the seepage
factor, and its distribution was mapped according to:
-
94
The results of a "random test" in 1969: starting at 19 places, more or less evenly distributed
over De Weerribben and lying outside the supposed former seepage area (Kuiper & Kuiper
1958), a one half-hour search for the species was made. It was eventually recorded at 13 of
these sites, several of which only yielded scarce stems;
All relevs with Scorpidium recorded in the vegetation mapping surveys by Oosterbroek &
Post in 1973 (1977) and by Verschoor in 1974 (1978). Scorpidium was present in a sum total
of 67 out of 922 records;
Fig.6.2
Former and present distribution of Scorpidium scorpioides in De Weerribben

The observations taken from Staatsbosbeheer (1988) apply to 1984-'85, the others to 1969-'75;
Hatching applies to the earlier period
The results of a bryological inventory in 1972 by members of the KNNV Bryology Working
Group (Van Wirdum 1983): Scorpidium was found in 9 out of 55 parcel groups visited,
partially coinciding with positive scores from the other sources;
Occasional records up to 1975.
The distribution of Scorpidium scorpioides is also shown in a map in Staatsbosbeheer (1988)

covering the whole area of the nature reserve. This map gives 153 locations for the species.
Fig.6.2 combines the results of both inventories. It is obvious that Scorpidium scorpioides is not
restricted to any limited seepage area. The moss is abundant, or even dominant in several former
petgaten immediately bordering upon the Blankenham polder area in the South-West, which is
known for the incidence of water losses towards the Noordoost-Polder (Veenenbos 1950, see
Chapter 5). Much of the area where Scorpidium has so far not been recorded consists of open
petgaten or of carr vegetation (see below).
95
Fig.6.3
Distribution of Liparis loeselii in De Weerribben

Records by Van Zon-van Wagtendonk that roughly coincide with those by Staatsbosbeheer have not
been separately indicated
Other species
Staatsbosbeheer (1988) also reports the distribution of Parnassia palustris, Menyanthes
trifoliata, Liparis loeselii, and Utricularia intermedia (Fig.6.3-5), which have been considered
seepage indicators also (Chapter 3). Among these species, Parnassia palustris is the only one
restricted to the supposed former seepage area.
From the other reputed seepage indicators, Eriophorum gracile (only reported for De
Weerribben by Kuiper & Kuiper 1958), Philonotis marchica (not reported for De Weerribben),
Scorpidium cossoni (= Drepanocladus revolvens ssp. intermedius, one collection from De
Stobbenribben by A.M.Kooijman and H.J.During, at the occasion of a very detailed mapping of
the moss cover of that quagfen complex), Calamagrostis stricta (mentioned by Kuiper & Kuiper
(1958) and observed by the present author in De Stobbenribben; see discussion by Corporaal
1984), Drepanocladus lycopodioides (three locations, all outside the supposed seepage area),
Sagina nodosa (observed, with Linum catharticum, in De Wobberibben by the present author),
Dactylorhiza incarnata (distribution in De Weerribben insufficiently known, but occurrence in
the supposed seepage area certain), and Carex buxbaumii (possibly one anonymous observation
96
Fig.6.4
Distribution of Menyanthes trifoliata in De Weerribben

The distribution roughly coincides with that of Equisetum fluviatile
from De Weerribben) are too rare or too difficult to find to draw any valuable information from
their distribution in respect to the present question.
Carex lasiocarpa, C.diandra, Aneura pinguis, Riccardia multifida, Bryum pseudotriquetrum,
Fissidens adianthoides, Campylium stellatum, and Campylium elodes are more common than
Scorpidium scorpioides in De Weerribben. Their pattern of distribution was not mapped, but
from my own observations and from vegetation records, especially those by Oosterbroek & Post
(1977) and Verschoor (1978), their distribution supports the conclusion drawn already for
Scorpidium scorpioides.
Hence, the distribution of seepage indicators in De Weerribben is not restricted to an area with
groundwater outflow, neither is it restricted to an area which could have been influenced by
groundwater outflow in the past, and which has been indicated as such by Kuiper & Kuiper
(1958).
97
Fig.6.5
Distribution of Utricularia intermedia and Parnassia palustris in De Weerribben
6.4 Is the distribution of seepage indicators restricted to any particular area?

Scorpidium scorpioides and Liparis loeselii
The distribution of Scorpidium scorpioides (Fig.6.2) and several other seepage indicators
(Fig.6.3-5) reveals some general patterns. Scorpidium scorpioides and Liparis loeselii are almost
absent from the central part of the reserve. This is a relatively far terrestrialized area with much
carr and scrub, and only few parts are regularly mown in summer. Conceivably, the stage of
terrestrialization and the management regime are unfavourable there for these species.
Both species are also absent from the part of the border zone of the Pleistocene sands
northeast of De Stobbenribben, called De Wanden. Although this zone is only partially within
the supposed seepage area indicated by Kuiper & Kuiper (1958), the groundwater outflow is
less improbable here than in the Schut- en Grafkampen region, where both species are frequently
met with and often abound in the local stands of vegetation. De Wanden consists of many open
petgaten, interlaced with intensively cultivated reed beds and strips of pasture. The pastures
were formerly, and partially still are, used by local farmers and fertilized for agricultural
98
production, and this is reflected on the soil map by Haans & Hamming by the abundance of
kragge-grassland. Presumably, the early stage of terrestrialization and the absence of relatively
nutrient-poor quagfen zones has kept the species away from this area.
Menyanthes trifoliata, Utricularia intermedia, and Parnassia palustris
Menyanthes trifoliata, Utricularia intermedia, and Parnassia palustris are almost absent from
the southeastern part of De Weerribben. In the northwestern part, their distribution shows a gap
around the village of Kalenberg. Parnassia palustris has not been observed south of this gap.
The three species are more or less restricted to the former Carex peat area.
It is possible that the distribution of Utricularia intermedia is limited by the fact that this
species, in The Netherlands, only rarely flowers (flowers only observed ca 1900 in the North of
the province of Limburg; Weeda et al. 1988). Turions are often abundant in quagfens, however.
Most of the known Dutch locations where the species has disappeared have suffered from
eutrophication, pollution, reclamation of mires, or a lowering of the groundwater table. The
general pattern is one of an obvious retreat from the southern part of The Netherlands during this
century. Since the Dutch sites of Utricularia intermedia are near its southern distributional limit
in North-Western Europe, site quality may be especially critical here.
Menyanthes trifoliata spreads both vegetatively and by means of seeds eaten by birds. The
limited distribution in De Weerribben must be due to a limited availability of good settling
environments (i.e., borders of shallow bodies of open water with a suitable base and nutrient
state). The great majority of Menyanthes sites in De Weerribben is in the proximity of the course
of an old rivulet, where the peat was not cut or dredged for reasons of a high mineral content,
and where the land has since long been used as hayfields and pastures (Chapter 4). The
distribution gap near Kalenberg may be explained either by a more intensive agricultural use of
the peatland (which also appears from the distribution of kragge-grasslands on the soil map by
Haans & Hamming), or by the influence of polluted surface water, flowing in from the main
canals (see Chapter 5).
Parnassia palustris occurs in two genetically different taxa in The Netherlands, one being
abundant in the coastal and dune areas, and the other being a now rare and notoriously
demanding species of inland mires and wet heathlands, including the North-West Overijssel
mires2. I have seen the species in some quagfens in the De Wieden area, and it was recorded for
De Weerribben ca 1965 by Leijs (pers. comm.) from a site where it was also recorded by
Staatsbosbeheer (1988), but where I never found it in the 1969-1975 period. Leijs described a
site that was cleared from scrub a few years before his observations, and this also holds for the
sites on the distribution map by Staatsbosbeheer (1988; see Fig.6.2). Next to this management
factor, it may be noted that several species of more or less calcareous milieus, such as
Parnassia, were relatively abundant in the mid-1980s, possibly due to an improvement of
hydrological factors, including the base state of the water. The present sites are all within or
nearby the supposed former seepage area, but, as shown in Chapters 5 and 7-10, they are now
part of an area of definite and strong groundwater recharge. Especially the recent increase of the
species in this single area does not support the original seepage theory.
Gadella, T.W.J. & E. Kliphuis 1968. Parnassia palustris in The Netherlands. Acta Botanica Neerlandica 17, 165-172.
99
In conclusion, the distribution of some of the seepage indicators is mainly restricted to some
delimited areas. This areal restriction coincides with various combinations of environmental
factors. The stage of succession (not too early or too late), summer mowing, isolation from the
main canals, associated with the absence of obvious pollution and eutrophication, and proximity
to a former rivulet, associated with both mineral and agricultural influences, seem to determine
the distribution of Scorpidium scorpioides, Liparis loeselii, Menyanthes trifoliata, Utricularia
intermedia, and Parnassia palustris. The occurrence of other seepage indicators, as far as traced
in De Weerribben, especially Carex diandra, Carex lasiocarpa, and several bryophytes, is
probably not restricted to any delimited area.
6.5 Do stands of vegetation with seepage indicators indicate a particular type of
environment?
The vegetation records in Oosterbroek & Post (1977) and Verschoor (1978) were grouped
according to the presence or absence of seepage indicators. For each group the frequency
distribution of ecological indications, from the list in Appendix C, was calculated on the basis of
species presence also. The seepage indicators were not taken into account in this calculation,
since they would possibly only confirm what was already deduced in Chapter 3, viz., that the
seepage indicators are associated with base-rich, yet nutrient-poor sites with rich-fen (or neva)
peat, and influenced by any sort of seepage or flooding. Indicated ecological differences
between the groups were tested with a chi-square test.
Fig.6.6 and Table 6.1 show the ecological spectra for the groups of records with and without
seepage indicators. The differences are not large, as can be understood from the fact that the
records obtain to vegetation stands at the kragge level of species aggregation Chapter 3). The
seepage indicators are characteristic of hydro-environmental zones which may or may not occur
within such stands. Moreover, they represent an ecological extremum in the spectrum (Chapter
3, Fig.3.6); the omission of their indication necessarily renders a vaguer picture, still the only
independent one that can be drawn on the basis of species indication. The data sets used here are
vast (17100 scores of non-seepage indicators in 920 records) and they include all stands of
vegetation in the areas covered, providing a description of the whole statistical population in
these areas, not biased by any preselection of supposed seepage and definitely non-seepage
sites, respectively.
The 568 records without seepage indicators comprise fewer litho-oligotrophic and more
atmo-oligotrophic species than expected on the basis of all-records distribution and the number
of species present, whereas the 352 records with seepage indicators have more litho-oligotrophic
species. The differences illustrated in Fig.6.6 are all significant at the 0.995 level. The records
with seepage indicators also comprise significantly more species and more red-list species,
whereas the records without seepage indicators have less.
More or less similar differences are found for the separate data sets. The Verschoor records
cover a strip right through the middle of De Weerribben. A relatively large area of this strip is in
a far stage of terrestrialization, including carr and ericaceous vegetation. The number of records
with seepage indicators is smaller than in the Oosterbroek & Post data set, and the average
number of species per record is also smaller. The Oosterbroek & Post data were collected in
another strip through De Weerribben, more to the South-East (see Fig.6.1), which is in an earlier
stage of terrestrialization. The general characterization of the subsets with and without seepage
indicators is similar to that in the Verschoor data, but the species frequency indicates a
somewhat higher base and nutrient state.
100
Fig.6.6
Ecological spectra of vegetation without (b) and with (c) seepage indicators,
corrected for the all-records species distribution (a)
Records from Verschoor (1978) and Oosterbroek & Post (1977); note that the seepage indicators
themselves were not included in the analyses
6.6 Has the distribution of seepage indicators changed in time?

The distribution of several supposed seepage indicators has been mapped, from various points of
view, upon more than one occasion in De Weerribben. This allows for some conclusions
regarding the dynamic behaviour of the respective distributional patterns.
101
Table 6.1
Frequency distribution of ecological indicators in de vegetation of De Weerribben
Subset
Data set Oosterbroek & Post
Nr of records
Sum of species scores
Seepage indicators
Red-list species
(without seepage indicators:)
Litho-oligotrophic species
Litho-mesotrophic species
Circumneutral-oligotrophic species
Atmo-oligotrophic species
Other species
Red-list species
Data set Verschoor
Nr of records
Seepage indicators
Red-list species
Other species
Red-list species
Absolute scores
All
No-sep Sep
Average per record

All
No-sep Sep
346
9465
444
976
161
3657
0
166
185
5808
444
810
27.4
1.3
2.8
22.7
0
1.0
31.4
2.4
4.4
9021
196
853
990
1722
5260
532
3657
73
336
368
795
2085
166
5364
123
517
622
927
3175
366
26.1
0.57
2.47
2.86
4.98
15.20
1.5
22.7
0.45
2.09
2.29
4.94
12.95
1.0
29.0
0.66
2.79
3.36
5.01
17.16
2.0
574
7635
279
931
407
4442
0
226
167
3193
279
705
13.3
0.5
1.6
10.9
0
0.6
19.1
1.7
4.2
7356
231
827
842
1683
3773
652
4442
99
468
429
1088
2358
226
2914
132
359
413
595
1415
426
12.8
0.40
1.44
1.47
2.93
6.57
1.1
10.9
0.24
1.15
1.05
2.67
5.79
0.6
17.4
0.79
2.15
2.47
3.56
8.47
2.6
568
8099
0
392
352
9001
723
1515
18.6
0.8
2.1
14.3
0
0.7
25.6
2.1
4.3
8099
172
804
797
1883
4443
392
8278
255
876
1035
1522
4590
792
17.8
0.46
1.83
1.99
3.70
9.82
1.3
14.3
0.30
1.42
1.40
3.32
7.82
0.7
23.5
0.72
2.49
2.94
4.32
13.04
2.3
Data set Verschoor plus Oosterbroek & Post

Nr of records
920
17100
Seepage indicators
723
Red-list species
1907
16377
427
1680
1832
3405
Other species
9033
Red-list species
1184
No-sep: records without seepage indicators; Sep: records with seepage indicators
For Scorpidium scorpioides a comparison can be made between the 1969-1975 period and the
situation reported by Staatsbosbeheer (1988). Although the distribution map for the 1969-1975
period is not at all exhaustive, I have quite frequently visited some of the areas where
Scorpidium was found ca 1985, but where it was not indicated for 1969-1975. This especially
holds true for the
102
areas marked 1 and 2 on the distribution map (Fig.6.2). I am almost certain that Scorpidium
was absent or almost so from these areas in the earlier period. Also, Kuiper & Kuiper (1958)
must have had enough information available to conclude that Scorpidium was rare or absent to
the south-west of the supposed former seepage area, and all further information I have leads me
to believe that the present abundance of Scorpidium in the Schut- en Grafkampen is attributable
to an expansion in the last decades. The distributional scores in the mapping area 3
(Oosterbroek & Post 1977) are also suggestive of such a shift: the records of isolated stems
resulting from the inventory in 1970, the findings from Oosterbroek & Post (1978), and those
from Staatsbosbeheer (1988) represent a southwestward movement of Scorpidium towards the
reserve margin.
In conclusion, there is a trend that the species has disappeared from the central area and adjacent
parts of the reserve, while hitherto unknown occurrences are found especially in the more
marginal areas. On the whole the species has probably expanded even since 1975, although the
recorded data do not provide any conclusive evidence.
The disappearance in De Wobberibben coincides with a substantial increase of Sphagnum
palustre, S.papillosum and S.flexuosum ssp. fallax, and almost certainly indicates a dramatic
acidification (unpublished vegetation map by G.J.M. Ruitenburg 1974, Calis & Van Wetten
1983), most probably due to the observed blocking of a ditch leading into the quagfen complex.
In general, the withdrawal of the species from the central part of the area can be attributed to
successional progress, including carr formation, abandonment of management, clogging of
ditches and the associated acidification or desiccation of quagfens.
The new appearance in area 1 coincides with the clearance of this area from scrub, and the
digging of some ditches in order to decrease the influence of rain water and, during droughts, of
a possible lowering of the mire water table. I have not revisited area 2 in recent years, but ca
1970 this area comprised a large areal of reed beds with, among other species, Campylium
stellatum, C.elodes, Carex lasiocarpa, and C.diandra, but without species more
characteristically bound to fen vegetation mown in summer or early autumn. It is worth testing
whether the management regime has changed here too.
In general, there are two local factors which could explain the expansion of Scorpidium
scorpioides towards the southwestern border of De Weerribben. Firstly, this area has long been
influenced by slightly brackish water. It is uncertain whether this was caused by local seepage
from the former Zuyderzee, as reported for the neighbouring polder area (but not very probable
at this greater distance from the former Zuyderzee dike), by a discharge of slightly brackish
water from these polders, or by resident salt in the local peat and clay deposits. The somewhat
brackish strain in the character of this part of De Weerribben is reflected in the occurrence of
such species as Scirpus lacustris ssp. tabernaemontani, Scirpus maritimus, Ophioglossum
vulgatum, and, possibly, Hippuris vulgaris. As shown in Chapter 5, the influence of lithotrophic
surface water was paramount in this area in the late 1960s, but this was only the case after the
construction of the Noordoost-Polder in 1941, and possibly even later, and facilitated by the
hydrological impact of reclamations southeast of De Weerribben in the mid 1950s. Clearly it is
an area subject to (further) desalinization during the 1940-1970 period, and the expansion of
Scorpidium could be a lagging response. This could also explain a possible absence of
Scorpidium and other seepage indicators, many of which avoid even slightly brackish water,
during the inventory reported by Kuiper & Kuiper (1958).
Secondly, the involved part of the nature reserve comprised the major part of relatively
young successional stages, used as productive reed beds, and so cut in winter, ca 1970. In the
103
Table 6.2
Frequency distribution of the stage of terrestrialization, in 1951, of 1985-'86 sites of

Scorpidium scorpioides in De Weerribben
Site group:
Open water
Aquatic macrophytes
Very weak and thin kragge
Firmer and thicker kragge
Firm and thick kragge
Not mapped by Haans & Hamming
5
50
37
43
0
18
1
6
2
11
0
18
4
44
35
32
0
18
A: all sites
B: sites north-east of the Kalenbergergracht (decrease area)
C: sites south-west of the Kalenbergergracht (increase area)
Numbers of Scorpidium sites; Data from Haans & Hamming 1962 and Staatsbosbeheer 1988
course of the 1970s, the reed culture went through some ups and downs, and nature conservancy
took care of the management of several parcels in the Schut- en Grafkampen area, partially
advancing the mowing or cutting time, and so favouring some of the supposed seepage
indicators prefering the vegetational physiognomy provided by summer-mown stands. For the
153 sites recorded in Staatsbosbeheer (1988), it was possible to quantify the relation with a
successional trend on the basis of the corresponding stage indication by Haans & Hamming
(1962). The results (Table 6.2) show that the majority of sites south-west of the
Kalenbergergracht were open water and very weak kragge stages in 1951, whereas the area
where Scorpidium is now decreasing mainly comprised kraggen that were already firmer in
1951. This result compares well with the general picture of the distribution map.
Scorpidium scorpioides is too rare a species in Dutch fen mires to expect its occurrence on the
basis of a preferential stage of terrestrialization and the absence of brackish influences alone.
There are no indications that the species was ever wide-spread and abundant in the central part
of De Weerribben, which was already in a relatively far stage of terrestrialization in 1951. This
central part not only differs from the marginal parts by the prevailing stage of terrestrialization
and by the absence of brackish influences, but also by the absence of mineral influences such as
provided by the former flooding with river water, in the Carex peat area in the eastern part of De
Weerribben, and by flooding with brackish water from the former Zuyderzee in the Schut- en
Grafkampen dredged-out bog peat area. These mineral influences are still traceable by the
presence of a local cover of clay, ca 1-4 dm thick, on undisturbed baulks and peat grassland.
A good correlation exists with the dominance of tree species in quagfens in the three relevant
areas, as mapped by Van Zon-Van Wagtendonk (1969):
- Expansion area (former bog, marine influences): Salix;
- Decrease area (former Carex peatland, fluvial influences): Alnus;
- Absence area (former bog, no marine or fluvial influences): Betula.
Next to these local processes and factors the expansion of Scorpidium scorpioides after 1975
was possibly facilitated by a general return towards lithotrophy of the surface water (Chapter 5).
104
Other species
The distribution of Liparis loeselii, Menyanthes trifoliata, and Utricularia intermedia was also
mapped by Van Zon-Van Wagtendonk (1969), and I have added the relevant information to the
distribution maps. I also included my own occasional observations during 1969-'72 and
information from vegetation records in Verschoor (1978) for Utricularia intermedia. Hack
(1973) mapped the occurrence of, inter alia, Menyanthes trifoliata along canals and ditches in
De Weerribben in 1969. Although that map is difficult to read the pattern appears to be similar
to that reported by Van Zon-Van Wagtendonk.
As regards Liparis loeselii (Fig.6.3), the inventory by Staatsbosbeheer (1988) shows much
more occurrences than were observed by Van Zon-Van Wagtendonk, although the general area
of distribution is similar. Since this species is only readily observable in a limited period and by
trained observers, and since Van Zon-Van Wagtendonk and H.N. Leijs, who contributed much
to her survey, largely worked on the basis of vegetational physiognomy and quick field
inspection, and since they did not visit all parcels, no conclusions can be drawn as to the
abundance and distribution of this species.
The distribution of Menyanthes trifoliata (Fig.6.4) has remained almost unaltered, although
some of the scattered occurrences in the former bog peat area were not noticed by Van Zon-Van
Wagtendonk and Hack.
The available data about the distribution of Utricularia intermedia have been summarized in
Fig.6.5. Although the abundance of this species may vary among years, and the individual data
sets may therefore be somewhat biased, the map suggests an outward move of Utricularia
intermedia, similar to that observed for Scorpidium scorpioides. Utricularia intermedia is much
less common, however, and it is especially rare outside the former Carex peat area. It would be
interesting to see whether, in due time, this species will or will not expand into the former bog
peat area.
A new hypothesis on the local behaviour of Scorpidium scorpioides and associated species
Summarizing, the following hypothesis is raised:
Possibly Scorpidium scorpioides, and some associated species, were restricted to the former
freshwater Carex peat area, roughly the supposed former seepage area, until 1940. This situation
was reported by Kuiper & Kuiper (1958). The Schut- en Grafkampen area was too brackish for
the species (and also too early a stage of terrestrialization), and the central, former bog peat area
had too low a base state. After the enclosure of the Zuyderzee and the reclamation of the
Noordoost-Polder (1941) and of the polders south-east of De Weerribben (mid 1950s), the
brackish water influence strongly decreased and the inflow of freshwater from the bordering
Pleistocene, mainly through the surface water system, eventually filled the mire area with a
lithotrophic water type, thus facilitating an expansion of the species over all parts of the terrain
that were both in a suitable successional stage and endowed with a high base state. Of course it
is uncertain whether Scorpidium might have settled in any appreciable amount in the central bog
peat part if that would have been in an earlier stage of succession. Especially in the 1973-'78
period, however, the expansion was slowed down, or even held up by a general deterioration of
the water quality, characterized by an increased similarity to polluted Rhine water and a
decreased similarity to lithotrophic groundwater (Chapter 5). In the last decade, the improved
water quality and a substantial increase in the area of quagfen mown in summer or early autumn
opened the Schut- en Grafkampen area for Scorpidium.
105
Simultaneously, successional progress and associated factors, such as the clogging of

ditches, leading to an increased similarity of the local mire water to atmotrophic rain water, and
the abandonment of management, caused the disappearance of Scorpidium from the supposed
former seepage region.
Considering the environmental requirements of Liparis loeselii, Scorpidium scorpioides,
Menyanthes trifoliata, Utricularia intermedia, and Parnassia palustris as belonging to a singular
class, it appears that these requirements differ in strictness in the order given. Or, in terms of a
human analogy, Liparis is the first to become conversant with a newly arisen environment,
whereas Utricularia and Parnassia are the most tardy. In a steady-state approach of
environmental response, such a difference is often colloquially named a difference in site
maturity, and it can be interpreted in connexion with the inherent nutrient supply of Finnish
authors, defined in Chapter 3. Neither of these terms explain much; they are just names for a
notion, which should be specified in order environmentally to test any site on its suitability for
the species.
Species-environment response relations are dynamic rather than static relations, however. In
the case of quagfens with seepage indicators, some sites, such as, in De Weerribben, those in the
expansion area of Scorpidium scorpioides, are apparently subject to a relatively fast succession,
whereas others, such as the older summer-mown quagfens on a firmer kragge, do not change
much during a certain phase of their existence. In this respect, the species are probably adapted
to a particular rate of change by physiological and regenerative capabilities. Liparis loeselii and
Scorpidium scorpioides seem to be more tolerant of relatively high rates of change than
Utricularia and (the quagfen form of) Parnassia palustris. This concept is probably similar to
the concept of milieu dynamics introduced by Van Leeuwen (1966, compare Van Wirdum
1982, 1985, 1986). Again, not much is explained, but the domain for further specification is
narrowed somewhat, and it has been made clear that the whole process must be studied, rather
than single states alone.
By adding the information about Pedicularis palustris and Equisetum fluviatile presented in the
next section, and my own, undocumented, impression with regard to the other seepage
indicators, the ranking of these species relative to their capability of quickly invading changing
environments can be completed as shown in Table 6.3. This capability is more or less related to
the withstanding of an increasing rate of change in the sites, such as occurs from disturbances,
acidification, or eutrophication.
Drepanocladus lycopodioides has not been included in this table, since I have no good
indication whether or not it belongs to the supposed class of seepage phenomena. It should be
noticed that the various species differ in their preference for various phases of the succession and
sometimes for environmental variants of the seepage sites, and that some of them may behave
differently (and possibly also belong to genetically different taxa) in other types of environment.
It is also important that all of the seepage indicators are slow invaders as compared to most
other species in the Dutch mire flora.
As to the application of this hypothesis in nature management, we could infer from Table 6.3
that, since Scorpidium scorpioides has expanded in De Weerribben, the 5 and 6 species groups
may possibly follow. If not, we might expect the group-4 species to decrease again, possibly
followed by group-3 taxa. Although we do not know in physical terms what happens exactly in
reponse to a certain management strategy, this would at least help to ascertain whether or not
management is progressing on its way to restore favourable conditions for the
106
Table 6.3
Ranking of seepage indicators and associated species according to their capability

of quickly invading changing environments, from strong (1) to weak (9)
1 Bryum pseudotriquetrum
Riccardia multifida
6 Utricularia intermedia
2 Aneura pinguis
Carex diandra
Campylium stellatum
7 Parnassia palustris
3 Campylium elodes
Carex lasiocarpa
Liparis loeselii
8 Eriophorum gracile
Scorpidium cossoni
Carex buxbaumii
4 Scorpidium scorpioides
Pedicularis palustris
9 Philonotis marchica
Sagina nodosa
5 Equisetum fluviatile
Scorpidio-Caricetum diandrae. With the ecological information presented throughout this study
such a diagnosis may be of help in the search for better strategies, if need be.
Although any testing and further environmental specification of the new hypothesis does lie
outside the scope of the present study, some corroborative information is given below. The
gathering of this information is based on the idea that several points in the hypothesis are not
restricted to Scorpidium and other seepage indicators. Especially the vegetation of aquatic
macrophytes is supposed to reflect the hydrological factors in the surface water system more
directly than do the quagfen species, since the local peat in the kraggen of a quagfen is an
additional factor determining the operational milieu of the species and the surface water quality.
An analysis of the distribution of other species may, therefore, contribute to the search for
ecological factors involved.
6.7 Further corroborative distributional information
The data presented in the above sections do not support the hypothesis that the distribution of
seepage indicators in De Weerribben is determined by the outflow of groundwater under the
influence of larger hydraulic heads in the bordering Pleistocene area. The distributional patterns,
and their dynamic behaviour, suggested a relation with (1) desalinization in the formerly slightly
brackish southwestern part of the nature reserve, (2) the general succession of the mire
vegetation, being in a different phase in the central and marginal parts of the reserve,
respectively, (3) management factors, possibly linked with the successional factor, (4) local
differences in the nutrient and base states of the fenland, and (5) a general tendency of increasing
lithotrophic influences in the whole mire area, both after the reclamation of polders and, in the
last decade, after a temporary deterioration of the water quality in the mid 1970s. These relations
were summarized in a new hypothesis about the changing distributional patterns of, especially,
Scorpidium scorpioides and Utricularia intermedia. Especially the factors 1, 2, and 5 above
107
Fig.6.7
Former and present distribution of some salt indicators in De Weerribben
refer to processes rather than states, and I have tried to rank the seepage indicators according to
the rate of change that they require and tolerate, respectively. The research carried out does not
allow any strict testing of this hypothesis and the factors involved, but some further
corroborative information will be presented below.
The distribution of salt indicators
The former brackish character of the southwestern marginal area of the De Weerribben nature
reserve neatly follows from geological, pedological, historical, and hydrological information, but
no water or soil analyses are available to document it. Staatsbosbeheer (1988), Van Zon-Van
Wagtendonk (1969) and Hack (1973) mapped the distribution of some salt indicators in De
Weerribben (Fig.6.7). Scirpus maritimus and S. lacustris ssp. tabernaemontani are salt
indicators beyond doubt. Hippuris vulgaris prefers clay soils in The Netherlands, and it is
considered a weak salt indicator. These three species occur in or nearby bodies of open surface
water, but their root systems may hold contact with a different micro-environment in the soil.
108
Ophioglossum vulgatum occurs in reeds and litter fen, and in The Netherlands it is less rare in
the coastal district with brackish influences, than it is in other parts of the country. The general
distribution of these four species together coincides with the present expansion area of
Scorpidium scorpioides, and also with the area where marine clay was once deposited (Fig.6.1
and 6.7, Haans & Hamming 1962) and disposed of in the petgaten in the course of peat
dredging.
Although the number of observations is relatively small, both species of Scirpus seem to
have disappeared from all but the very soutwestern margin of De Weerribben. The distribution
of Ophioglossum vulgatum was only mapped by Staatsbosbeheer (1988); I just added some
occasional observations in order to complete the picture somewhat. Hippuris vulgaris has drawn
the attention of various researchers through the years. The distribution maps are probably quite
reliable and show a substantial decrease in the presence of this species between 1968-'69 and
1985-'86. The finds northeast of the Hamsgracht canal which were not noticed by Hack were not
inspected by that author either and most probably do not represent new settlements. Hippuris
does not show any clear movement; it just gradually disappears.
The distribution of Stratiotes aloides (Fig.6.8)
The ecology of Stratiotes aloides has been the subject of other researchers (M.C. Groenhart,
largely unpublished, Bloemendaal & Roelofs 1988, Roelofs & Cals 1989), and it is only touched
upon here as far as the observations made for the present study provide a link between the
surface water system and quagfen vegetation. This is especially so since the aquatic vegetation is
expected to react more directly on the surface water quality than does the quagfen vegetation,
and Stratiotes aloides has shown a dramatically dynamic behaviour in the 1970-'89 period. The
species has recently been associated with groundwater outflow phenomena (Roelofs & Cals
1989).
Stratiotes aloides was especially abundant in the northeastern part of De Weerribben from ca
1940 to ca 1970, as can be inferred from maps, aerial photography, and information provided by
local inhabitants (Fig.6.8). Some of this information seems to indicate that the species was less
abundant before 1940, and that it reached a maximum abundance in the 1950s. Around 1970, it
was still so abundant in De Weerribben that several waterways were no longer navigable, and
the State Forestry Service purchased a mowing boat in order to remove this vegetation from
many places. On aerial photographs and from my own observations it could be noted, however,
that ca 1970 a gap existed between the Stratiotes zones in many lakes and petgaten and a band
of nymphoid vegetation, which gap I am now inclined to interpret as a sign of a retreat of the
stands of Stratiotes vegetation (Van Wirdum 1979). Indeed, in several places where Stratiotes
has now entirely disappeared, the nymphoid band remained present and only slowly widened in
a shoreward direction.
From 1972 onwards, and within 3-5 years Stratiotes disappeared from most places in the
nature reserves of North-West Overijssel, and ca 1977 it was almost completely absent from De
Weerribben. In De Wieden this process started some three years earlier. This has been ascribed
to eutrophication and turbulence (compare Segal & Groenhart 1967, for Broadland, U.K.: Moss
1978), to a fungal disease (Zonderwijk, pers. comm.), and to too high concentrations of
bicarbonate from the boezem system (Bloemendaal & Roelofs 1988, see discussion in Van
Wirdum 1989). In Van Wirdum (1979) I related the disappearance of Stratiotes to a combination
of local nutrient dynamics and the way the relevant processes were governed by the macro-ionic
109
Fig.6.8
Former and present distribution of Stratiotes aloides in De Weerribben

The areas 1 and 2 are specifically referred to in the text
composition of the surface water. More or less contrary to Bloemendaal & Roelofs, I hold the
decreased lithotrophy, indicated by a decreased Ionic Ratio [1/2Ca2+]/{[1/2Ca2+]+[Cl-]}
(Appendix D) and an increased electrical conductivity, responsible for the occurrence of peaks
and lows in the availability of nutrients, which fluctuations I suppose to be unfavourable for
Stratiotes aloides. Moss' publication provides a description of mechanisms that could be
triggered by the nutrient peaks involved. In my hypothesis, Stratiotes aloides was related to
lithotrophic, yet stable nutrient-rich environments, often in the neighbourhood of farmland. Such
a relation is also suggested by the distribution of Stratiotes in De Weerribben, where most of the
1968-'72 sites are in the immediate neighbourhood of grassland, and where an interesting
aggregation of sites is found near the strip of undredged mineral-rich peat traversing the area.
Through a detailed inventory in 1976 (Lumkes unpublished, Verschuren unpublished) of the
areas marked 1 and 2 in Fig.6.8 it was confirmed that the species had almost disappeared
from both areas. In area 1 two sites were still present, comprising eight plants and a small but
healthy stand, respectively. In area 2 six ditches and six petgaten still had Stratiotes, four
petgat sites and one ditch site containing a small number of unhealthy-looking plants only. The
sites where the species maintained itself longest were all enclosed in grassland and not easily
110
accessible for boezem water, which had changed in character from lithotrophic to
molunotrophic, but was not yet hypertrophic (Chapter 5).
During an occasional terrain inspection in 1979 I noticed a few healthy Stratiotes plants at
two sites from where the species had been absent in the years before. During the 1980s it
gradually became clear that Stratiotes aloides made its way back in De Weerribben, although
Staatsbosbeheer (1988), on the basis of the inventory given in Fig.6.8, did not yet believe this
would continue. Healthy Stratiotes was, however, in 1985-1989 found scattered all-over the
reserve in increasing numbers. From consecutive aerial and field observations, I noticed several
small, and at first submerged, flocks expanding to cover whole petgaten during this period.
Although no causal relation has been demonstrated, this correlates well with the observed
stabilization of the lithotrophic strain in the water quality (Chapter 5). It appears from Fig.6.8
that Stratiotes now occurs much more scattered throughout De Weerribben than it did in the
1968-'72 period. While lithotrophic conditions were observed all-over the area in the late 1960s,
this suggests that the nutrient state was only sufficiently high in the earlier period in the
neighbourhood of agricultural land use, and that such a relatively high nutrient state, i.e., a
consistently high supply, is now more wide-spread.
A similar recovery as recorded in De Weerribben (and in some other parts of North-West
Overijssel) has, until 1988, not been observed in the lakes Venematen and Duinigermeer in the
Wieden area. Water quality data are available for lake Venematen, and these data show that the
Ionic Ratio has not yet recovered there either.
These data, while not yielding any demonstrative proof of whatever hypothesis, would fit in the
hypothesis suggested by the expansion of Scorpidium scorpioides: an expansion towards the
southwest of De Weerribben by desalinization after the construction of the Noordoost-Polder,
and a general recovery by the presently lithotrophic water quality. It remains uncertain whether
the higher nutrient state possibly indicated by Stratiotes will become a negative factor of any
importance to the quagfen vegetation, as suggested by the absence of Scorpidium from the main
area of the former distribution of Stratiotes in De Wanden (compare Fig.6.2 and 6.8).
The distribution of some other species
Among the other species for which the distribution has been mapped in Staatsbosbeheer (1988),
Van Zon-Van Wagtendonk (1969), or Hack (1973) and which are relevant to the question of
(ground)water flow and vegetation are Pedicularis palustris and Equisetum fluviatile.
Pedicularis palustris (Fig.6.9) is often associated with quagfen vegetation including seepage
indicators. On the national scale, Pedicularis is considerably less rare than Scorpidium
scorpioides. Its distribution in De Weerribben is more or less similar to that of Scorpidium and
Liparis loeselii, but it is less common. Pedicularis palustris is a half-parasite and it thrives well
on graminoid species, such as Juncus subnodulosus, in quagfens. If the vegetation is not mown
during the growing season, its occurrence is not persistent. There is no evidence of the species
showing a strongly dynamic behaviour, and it most probably occurred at most sites in the
expansion area of Scorpidium scorpioides before the latter species settled there, which indicates
that the overlapping part of their requirements, especially the mowing regime and the local base
state, was already fulfilled.
Equisetum fluviatile (Fig.6.4) is also often associated with seepage indicators, and it has
been described as a dominant species in earlier phases of the quagfen succession in supposed
111
Fig.6.9
Distribution of Pedicularis palustris in De Weerriben
groundwater outflow areas (see Kuiper & Kuiper 1958, Segal 1966, Meijer & De Wit 1955). I
have observed similar zonations in parts of De Wieden also. In De Weerribben this is far less
commonly found, and Equisetum fluviatile is restricted to the Carex peat area, like Menyanthes
trifoliata. Its distribution was not studied after 1969, so that no conclusions can be drawn
concerning its dynamic behaviour.
112
CHAPTER 7
The quagfens of De Stobbenribben and their vegetation
7.1 Introduction
De Stobbenribben is a complex of former petgaten in the north-eastern part of De

Weerribben (Fig.7.1). Kuiper & Kuiper (1958, also Kuiper & Lapr 1956) and
Segal (1966) gave particular attention to this complex and mentioned it as a good
example of quagfen under the supposed influence of outflowing groundwater. As
shown in Chapter 6, De Stobbenribben is well within the area of distribution of
various seepage indicators in De Weerribben, and it may be considered a locus
classicus for the seepage hypothesis introduced in Chapter 2. From Chapter 5
onward it is clear that, since the reclamation of the nearby polder Wetering-Oost
about 1955, the hydraulic head of the groundwater in De Stobbenribben became so
much lowered that it is now an area of infiltration, supplied with water from the
boezem system and the local rainfall.
The vegetation cover of De Stobbenribben, however, did not change considerably
since about 1955, and several seepage indicators are still present in appreciable
quantities. De Stobbenribben was therefore selected for a descriptive case study
into the relations between hydrological factors and the occurrence of seepage
indicators. In this chapter I will provide the relevant information about De
Stobbenribben, the local stands of vegetation, and the indicated environmental
conditions. Some special methods concerning the eco-hydrology, and the results of
their application, will be reported in Chapters 8 and 9. In Chapter 10, I try to
combine the various recorded data into a functional hypothesis that can be used as
a basis for further research in De Stobbenribben and other mire complexes.
De Stobbenribben
113
Fig.7.1
Location of De Stobbenribben in North-West Overijssel

See Fig.4.3 for further details
Fig.7.2
Diagrammatic representation of De Stobbenribben

Relative width of baulks and ditches exaggerated; A-D: Parcels and sections referred to in the text of
Chapters 7-10
114
De Stobbenribben
7.2 Topography, petgaten and kraggen

The four Stobbenribben quagfen parcels studied in this Chapter are about 30 x 200 m each,
oriented in a southwest-northeast direction. They will be indicated here as parcel (former petgat)
A..D from north-west to south-east. (Fig.7.2). The original Carex peat was probably dredged
away between 1880 and 1920. The depth to the sand bottom varies between 1.6 and 3.6 m
(Fig.7.3). The original peat was formed under the influence of a rivulet and it contained many
old tree logs; hence the name Stobbenribben (Du. stobbe: tree stump). Before the peat was
dredged, the area had for some centuries been used as agricultural land, probably primarily as
pastures. Due to these factors an inferior quality of fuel peat was produced and tree stumps as
well as the grassland sod were partially dumped in the petgaten. Some remaining peat was dug
away during World-War II. During the terrestrialization, Typha angustifolia and Menyanthes
trifoliata probably were the pioneering helophytes, growing partially on sods and old peat
materials floating up from the bottom of the petgaten (Havinga 1957). The kragge thus became
an admixture of newly formed rhizomes and peat, and older materials. Where old peat was
raised partially above the water level, it may have given rise to a fast acidification and the
formation of irregular patches of Sphagnum and Molinia vegetation which can still be
recognized.
Most probably the 1.5-3 m wide baulks in-between the terrestrialized petgaten were formed
by the disposal of sods in old ditches during peat dredging. The kraggen are fixed to these
baulks about 0.5 m below the baulk top. Away from the baulks they float up and down with the
movement of the water table almost freely. During the present survey the baulks were partially
grown with Myrica, Salix and Alnus scrub and Molinia tussocks, but they have recently been
cleared to provide acess for tractors used for the transportation of hay and reed. The wider strips
of peat bordering De Stobbenribben to the south-east and south-west partially consist of the
original peat.
The thickness of the kragge roughly varies between 0.4 and 0.7 m, but at some
places the removal of scrub has produced scars providing a window to the underlying body of
mire water. Similar windows (called pulk-holes in Norfolk) have in the past been interpreted as
seepage windows, but there is no evidence of the correctness of this interpretation. Phragmites
rhizomes form the main network in the kragge, and vertical stems extend far below it into a
body of relatively clear water. The upper part of the kragge is mostly composed of Carex
rhizomes, covered with bryophyte material, but locally other species may have contributed
substantially to kragge formation. The peat in De Stobbenribben has been dredged to an average
depth of 2.5 m. The remaining peat is covered with a ca 1 m thick layer of sapropelic broth,
leaving a "sheet" of relatively clear water between 0.7 and 1.2-1.5 m below the kragge surface.
Between the remaining peat and the underlying sand bottom a 3-15 cm thin so-called gliede
layer is often observed. This layer consists of an amorphous, strongly humified organic matter
inclusive of illuvial humic substances, which may contain clay minerals and sand grains, and it
has a low permeability to water.
Nature management in De Stobbenribben started about 1959 and some scrub of Myrica and
Alnus was removed from the kraggen shortly after 1965, as can be inferred from a comparison
of vegetation maps. During a preliminary survey in 1969-'70 I raised strong doubts about the
presence of groundwater outflow (see Chapters 8 and 9), and the alternative hypothesis
emanated from studies of a quagfen complex loosing substantial amounts of water towards the
body of groundwater, especially under the influence of the low-lying polder Wetering-Oost at a
distance of only 250 m. De Stobbenribben was chosen for a case study for this reason and also
since (1) it is a locus classicus in the original seepage theory and (2) the entrance of surface
De Stobbenribben
115
Fig.7.3
Sand depth (cm) in De Stobbenribben; location of some research sites mentioned

in Chapters 7-10
water into the petgaten is almost entirely forced through the open connexion with a ditch at one
short end of each petgat. The study was directed towards the detection, description, and analysis
of a conceivably resulting length gradient in the petgaten.
116
De Stobbenribben
7.3 The vegetation cover

Available data
The vegetation cover of De Stobbenribben has repeatedly been recorded. A general description
was given by Kuiper & Lapr (1956). The stands of two permanent quadrats were reported by
E.H. Krijger, H. Gaasenbeek, S. Segal, A.C. Adam, W.N. Ellis, and M.C. Groenhart, in various
combinations, in 1959, '60, '61, and '64 (unpublished records in State Forestry Archives,
Zwolle). These permanent quadrats were abandoned after 1964 and their exact location was not
recorded. From a sketch map and additional data on the data sheets it is certain that they were
situated in quagfen parcel D, in a wet hollow with a weak kragge (Fig.7.3). More detailed
investigations were carried out from about 1964 to 1970 under the supervision of S. Segal
(University of Amsterdam). The present author was also introduced to the area and the seepage
problem as a student of Segal in 1969. From 1973 onwards various students took part in the
study reported here under his supervision.
The first vegetation map was produced by Van Zon-Van Wagtendonk (1965). The moss
layer was not analysed in detail during that study, and the boundaries between different types of
vegetation were sketched in, rather than accurately determined. Stegeman (1968) made a
detailed survey of the micro-zonation of bryophytes, especially in De Stobbenribben, but her
report does not provide a synopsis of the stands of bryophyte vegetation, and she obviously
wrongly identified some species. Van Zon-Van Wagtendonk and Stegeman also report relevs
of representative vegetation stands.
During the present survey, a new map of the vegetation cover was prepared in 1973 by
Bergmans (1975). Although some species may have been overlooked by him, this map provides
a rather precise and well-documented picture of the vegetation cover, based both on the higher
plants and the bryophytes. A simplified version of this vegetation map is discussed below and
presented as Fig.7.4.
Bredenbeek, Gerrits & Van Loon (1979), under the supervision of H.N. Leijs (RIN)
published another vegetation map, according to the legend developed by Van Zon-Van
Wagtendonk. The drawback of this legend, which is strongly based on the abundance of some
higher plant species with a rapid vegetative spread, is that the resulting maps probably reflect the
effects of drier and wetter years and of the time of mowing at least as strongly as the basic
pattern of water relations (cf Chapter 6). Boeye (1983) compared the various vegetation maps
and made an additional, though short, survey as part of the present study. Kooijman (1985)
studied some important factors of nutrient dynamics in De Stobbenribben at three locations in
one of the parcels and she recorded the vegetation stands at these locations various times
according to a detailed sampling in 20 x 20 cm plots. Her study was supervised by
J.T.A.Verhoeven (University of Utrecht) and by the present author.
Aerial photography and multispectral scanning were applied to De Stobbenribben as part of
a survey into the application of remote sensing techniques to the mapping of vegetation (Van
Wirdum 1977, Jalink 1990, and various unpublished results). Although these remote sensing
studies are not reported here, the results have provided additional information concerning the
vegetational pattern through the years (see also Chapter 10), and one aerial photo is reproduced
in Fig.7.5 in order to provide a picture of De Stobbenribben intermediate between the real
situation and the abstract vegetation map.
De Stobbenribben
117
Fig.7.4
118
Vegetation map of De Stobbenribben (generalized after Bergmans 1974)
De Stobbenribben
The vegetation map (1973)

In order to render the vegetational pattern of 1973, reported by Bergmans (1975), in a
surveyable diagram, several types were combined into four that especially emphasize the
composition of the moss layer and the presence of differential species. By doing so, the
abundance coding of species had to be adapted and was scaled down to four classes with a
numerical value used for quantitative interpretation:
5 Dominant, and alone determining the visual impression of either the herb layer or the moss
layer in substantial parts of the mapped area;
3 Frequent, i.e., found without much searching, and often locally abundant or even patchwise
dominant;
1 Infrequent, i.e., either rare, or in fact not rare but easily overlooked when the actual sites of
occurrence are not specially searched for;
0 Not observed.
The estimates given in Table 7.1 at the end of this chapter apply to the whole area mapped under
the relevant type, rather than to any arbitrary number of vegetation samples. The description is,
therefore, equivalent to that given for the map elements in the maximum-areas method
introduced in Chapter 5, although the reliability and accuracy of the descriptions for De
Stobbenribben are much greater, due to the more elaborate method of field recording. Still some
species must have been overlooked or misidentified.
Since the main types of the vegetation cover are different in their spectral reflection also, the
pattern can be recognized from the air, as shown in the remote sensing image of Fig.7.5. The
vegetation cover obviously exhibits a length gradient, both in its floristic composition and in its
spectral reflection.
Table 7.1 includes a summary of the vegetation records of the permanent quadrats, re-scaled to
the new scale, although strictly not applicable to single relevs. The records by S.Segal and
H.Gaasenbeek, dated 1961 (July, 25th, and August, 3rd) were taken as the main reference, since
these records appear most reliable. Additional species mentioned for any of the years 1959-'64,
in these permanent quadrats have been marked with a plus-sign in the appropriate column. All
relevs, especially of the moss layer, are biased by the experience and feeling for less
conspicious species of the various investigators, and the differences should not be overemphasized.
Some species were not mentioned by Bergmans (1975). Most of these species are still rather
abundant in De Stobbenribben, and must have been overlooked. The abundance of
Drepanocladus lycopodioides mentioned by Segal is almost certainly a mistake. The species
may have been present, but forms of Scorpidium scorpioides may have been taken for it in the
abundance coding. Collections I saw invariably contained only Scorpidium scorpioides.
Calliergon giganteum is difficult to distinguish from C.cordifolium. Plants usually considered to
belong to C.giganteum certainly occur (or occurred) in De Stobbenribben, but unquestionable
C.cordifolium is as abundant locally. Sparganium minimum is a very rare species in De
Weerribben anyway.
The permanent quadrats were situated in a very wet part of zone B in quagfen parcel D
(Fig.7.3). A comparison of the records made ca 1960 with the 1973 zone B description (Table
7.1) does not reveal any striking differences, the less so since the exact location is not known.
De Stobbenribben
119
Fig.7.5
Aerial photograph (near-infrared) of De Stobbenribben (1975, May, 30th)

Photography: CNES/NIWARS; the deviating field in parcel C (Fig.7.4) represents a Cladium mariscus
stand
On the stability of the vegetational gradient

A comparison of the various vegetation maps (Bergmans 1975, Boeye 1983) proves that the
general length gradient observed in 1973 was also present in 1965, 1979 and 1983. There are
120
De Stobbenribben
certain striking differences, however, within the main vegetational zones. These differences are
largely due to management factors, as dicussed below for the three major points.
In parcel A, zone B, a stand dominated by Typha angustifolia was mapped by Bergmans
(1975). This stand was also observed by Van Zon-Van Wagtendonk (1965), but it had
disappeared in 1979 (Bredenbeek et al. 1979). It was replaced by quagfen vegetation dominated
by Juncus subnodulosus and, on hummocks, by Sphagnum species. This may be due to the
continued mowing in late summer. Mowing in summer not only reduces the growth of large
helophytes but also exerts an influence on the terrain conditions: the weak upper part of the
kragge, with hollows and hummocks, is re-modelled by the machine tracks, and some of the
weakest spots become filled with plant remains. Possibly, the obstruction of the ditch, discussed
in Chapters 8 and 9, contributed to a changing water quality underneath the kragge, thus also
worsening the milieu for Typha angustifolia.
In parcel C, zone B, Bergmans (1975) found an extensive stand of Cladium mariscus. This
stand was formed by the extension of Cladium vegetation from smaller patches indicated by Van
Zon-Van Wagtendonk (1965). This was almost certainly caused by a less intensive mowing due
to the local inaccessibility of the weak kragge. After 1983 the stand of Cladium mariscus was
cleared and it is now replaced by another zone B ,and partially zone C, vegetation.
In parcel D, zone B, vegetation dominated by Juncus subnodulosus and Sphagnum species in
1965-'79 was partially replaced by a type dominated by Carex elata and Scorpidium scorpioides
in 1983. This seeming regression may well be due to the combined influences of ditch cleaning
and mowing with machines. It was also noticed in other parcels. It is possible that this particular
effect is rather due to weather and wetness conditions in specific years than representing an
ongoing trend, however. I have seen relatively strong fluctuations in the dominance of various
plant species through the years anyway.
As regards the present survey, the following conclusions can be drawn from the comparison of
vegetation maps:
The length gradient in the vegetation of De Stobbenribben was stable between ca 1965 and
ca 1983. Additional reference to aerial photography (1949, 1959), and own observations
suggests that this stability holds true for the whole period 1950-'90.
In spite of possible changes in hydrological conditions, the local environment was still
suitable for an expansion of Scorpidium scorpioides and other associated supposed seepage
indicators at the end of the period mentioned. Any significant expansion of zone A vegetation
can not be proved on the basis of the vegetation maps.
7.4 The vegetational zones
In order to quantify the similarities between the vegetational zones in 1973, the similarity
coefficient of Srensen (S) is evaluated in Table 7.2:
(Srensen Formula)
where
a,b:
c:
a+b-c:
S = 2c / (a+b)
sum total of species in zones a and b, respectively;

species common to both zones compared;
species unique to either of the zones compared.
De Stobbenribben
121
Table 7.2
Evaluation of the Srensen similarity coefficient for the vegetational zones in De

Stobbenribben (left-hand table, S given as %).
A
A 100
B
C
D
Srensen
C
D
55
55
65
68
100
78
100
B
53
100
A
B
70
63
36
65
47
54
44
53
common
unique
D
72
49
42
90
C
76
57
100
74
The right-hand table shows the number of species common to both zones compared (below diagonal), and the number of
unique species (above the diagonal). The diagonal has the sum total of species in each zone.
Table 7.3
Ecological indicators in the vegetational zones A-D in De Stobbenribben.

N
D TOT
ALL
133 150
149
Two-factor indications:
LTHOLI
8
4
20
LTHMES
6
9
13
REST
79 106
75
ATMOLI 24
13
22
CIROLI
16
18
19
Seepage, Red-List species:
SEP
10
11
27
RED
27
22
37
Water Type:
WT?
21
21
22
LTH
25
30
42
ATM
32
17
25
CIR
55
82
60
Nutrient State:
NU?
10
11
13
EUT
23
40
12
MES
47
60
54
OLI
53
39
70
Phytosociological Type:
BOG
8
0
1
DAV
4
5
7
FEN
30
28
41
LAS
10
9
19
LIT
30
33
27
MOL
22
20
16
SMP
29
55
38
228
170
697
10
10
116
66
26
3
8
90
44
25
37
40
387
145
88
21
63
4
25
63
147
40
29
83
76
31
17
54
68
16
16
81
115
14
14
60
82
114
118
179
286
ANY
54
82
255
306
16
10
72
22
47
39
22
7
3
52
8
45
33
22
24
25
193
58
152
108
137
Finnish System:
FN?
52
54
33
63 53
FIN
81
96 116
165 117
Base State:
OMB
7
1
4
21 10
POR
6
4
7
16
9
TRL
32
40
51
62 57
WMS 30
44
39
48 37
XRC
6
7
15
18
4
Mire Level:
FLK
41
71
81
70 45
HUM 18
7
13
44 29
INT
22
18
22
51 43
Supplementary Nutrient Effect:
NO
8
1
8
28 14
16
23
21
40
24 108
FLD
39
50
67
66 62
SEP
18
22
20
31 17
Inherent Nutrient Effect:
NO
51
68
69
90 76
MOS
6
1
4
16 10
NVA
10
11
18
21 11
RFN
14
16
25
38 20
TOT
203
494
36
36
210
168
44
267
93
134
51
245
90
303
31
61
99
N Number of species involved; A..D Frequency scores per zone; TOT Sum total of scores
(Phytosociological type SMP includes type AQU)
Explanation of terms: Appendix C, Chapter 3.
122
De Stobbenribben
The greatest floristic similarity exists between zones C and D, B and D, and B and C,
respectively. The zones are considered more or less arbitrarily defined in a continuous gradient.
The transition from zone A to B is somewhat more abrupt.
An ecological characterization of the zones is made on the basis of species indication, as
treated in Chapter 3 (see Table 7.1 also) in order to obtain an independent reference for
conclusions on the basis of ecological investigations. Several avenues are open for such a
characterization. The one treated below is based on the frequency of indicators in the various
vegetational zones, evaluating dominant, frequent, and infrequent species by counts of 5, 3, and
1, repectively. The result of this counting is shown in Table 7.3 and will be referred to as a
spectrum. The distribution of frequencies in De Stobbenribben as a whole results from the sum
total of the values found for the separate zones and this all-zones spectrum is diagrammatically
represented in Fig.7.6. The spectra for the separate zones have been corrected for the all-zones
distribution in order to obtain the corrected spectra given in Fig.7.6. Note that, as in Chapters 3
and 6, only species indicative of a co-occurrence of litho- and oligotrophic, litho- and
mesotrophic, atmo- and oligotrophic, and circumneutral and oligotrophic conditions,
respectively, were separately counted in these diagrams. The differences between the zones are
significant at the 0.995 level (chi-square test).
Seepage indicators and red-list species are especially found in zones B and C. The
phytosociological spectra in Fig.7.6 show that zone A especially differs from the other zones in
the abundant representation of SMP indicators. Zones B and C are somewhat different from the
others in their slightly higher presence of LAS and DAV indicators, the transition from B to C
being marked by a shift from SMP to BOG species. Although the frequency of BOG species is
also higher in zone D than in De Stobbenribben as a whole, this zone is especially characterized
by FEN, MOL, and LIT species. Ecologically, zone A is characterized by a strong dominance of
indicators for the CIR water type plus a wide range of trophic state indicators (Table 7.3). The
spectrum in Fig.7.6 draws the attention to the relatively poor representation of litho-oligotrophic
and atmo-oligotrophic conditions. In zones B..D the oligotrophic nutrient state is dominant and
eutrophic conditions (in the REST group) are indicated as unimportant. As regards the water
type, the circumneutral indications prevail, but not as strongly as in zone A. The spectra for the
zones B and C in Fig.7.6 suggest a relatively great importance of litho-oligotrophic and atmooligotrophic conditions, respectively.
The results of the application of the Finnish system of indicators have been included in Table
7.3. These results, while generally in line with the above discussion, are suggestive of some
further points. Flark level species indicative of an external nutrient supply by flooding are about
equally distributed over the zones A..C, such species being less abundant in zone D. Zone C
shows a peak in the presence of indicators of an inherent nutrient supply typical for bog (group
MOS) and rich fen (RFN). Hummock-level species are especially more abundant in this zone
suggesting a possible successional relationship between zones B and C, facilitated by local
factors. The importance of such local factors as the access of ditch water, the depth of peat
removal, and the proximity to baulks, is strongly suggested by the deviating pattern for zones
B..D in the narrow parcel B (Fig.7.4). The transition of zone A to zone B is not marked by a
reduced abundance of flark level indicators, and so does not yield evidence for the suggestion
that zone B follows upon zone A in the vegetational succession; the pattern is probably based on
spatial relations.
In conclusion, seepage indicators are especially abundant, in De Stobbenribben, in an extended
vegetational zone between the more swampy and eutrophic ditch side and the litterfen at the
opposite side of the quagfen parcels. This extended zone can be roughly divided into a part
De Stobbenribben
123
Fig.7.6
Ecological and phytosociological spectra of the vegetation stands in zones A-D in

De Stobbenribben, corrected for the all-zones distribution given in the first radar
diagrams
Radar diagrams have been explained in Chapter 3
124
De Stobbenribben
dominated by amblystegiaceous mosses and a part dominated by Sphagnum species, which may
be considered a successional phase due to the increased influence of peat accumulation, rain
water and acidification. The micro-sites unique to zones B and C are indicated as litho- and
atmotrophic, repectively, and oligotrophic in both.
7.5 Description of the vegetational zones
In this section a primarily floristic description of the vegetational zones A..D is given along with
references to mapping units used by various authors and to phytosociological units according to
Westhoff & Den Held (1969). A more elaborate discussion of phytosociological schemes of
classification is given in Appendix B, and further ecological data are presented in Chapter 10.
Zone A - Referred to as: Eu-mesotrophic
reed
Bergmans 1975:
Van Zon-Van Wagtendonk 1965:
Bredenbeek et al. 1979:
reed swamp, reed, swamp, eutrophic zone, Calliergonella-Phragmites

Type 6;
Type 30;
Types 35 and R31
Occurring along the place where ditch water enters the quagfen parcels in the north-east.
Phragmites australis reaches a height of ca 1.8 m. Near the ditch and the baulks the stems are even taller. It is
usually cut in late autumn, but cutting in winter and mowing in summer have occurred in certain years. Carex elata and
C.paniculata form tussocks, C.paniculata being the more abundant of the two. Phalaris arundinacea, Alisma plantagoaquatica, Ranunculus lingua, and Brachythecium rutabulum may have been overlooked. These species, and those listed
in the appropriate groups in Table 7.1, are quite characteristic of the zone. At the ditch side Glyceria maxima and
Sparganium erectum are also found. Most species that are frequent in this type are more vigorous here than elsewhere in
De Stobbenribben. The difference in types 35 and R31, both also different from 30, is not considered significant in
respect of the overall heterogeneity of the type and of the influence of the varying mowing season.
Classification: Class Phragmitetea, Order Nasturtio-Glycerietalia, Alliance Cicution virosae; several species of the
Orders Phragmitetalia and Magnocaricetalia of the same Class and species of the Alliances Calthion and Filipendulion
(Order Molinietalia, Class Molinio-Arrhenateretea) are also present. In view of the small areal extent, at a meeting place
of ditches, baulks, and kraggen, and their associated influences and border effects, the heterogeneity of the stands can
easily be explained.
Zone B - Referred to as: Litho-oligotrophic swampy brownmoss fen, brownmoss (quag)fen phase, ScorpidiumCarex fen
Bergmans 1975:
Types 7, 2, 3, 4, (1);
Types 52, 53, 32, 37;
Types 53, 55, 37, 52, (51)
Occurring in the large central parts of the parcels, in parcel A forming a fine-scaled mosaic with types A and C,
respectively.
Types 52, 53, and 55 represent quagfen vegetation characterized by Juncus subnodulosus with Scorpidium
scorpioides (52) and local hummocks of Sphagnum subnitens (53), and Carex elata-dominated quagfen (55). Most of the
areal extent of these types is covered by Bergmans' type 7: a quagfen vegetation with Scorpidium scorpioides and
Campylium stellatum dominant in the moss layer and with Carex elata and various other species in the herb layer.
Campylium elodes and Bryum pseudotriquetrum should be mentioned for this zone also, although missing from
Bergmans' table. The growth form of Carex elata is similar to that discribed as forma dissoluta by Braun (1968), but this
is most probably a modification stimulated by the mowing regime.
Type 51 (Bergmans: 1) above refers to a narrow strip of vegetation dominated by Menyanthes trifoliata and Carex
lasiocarpa near the baulks of parcel A. Type 50 (Bergmans: 2), a very weak part of the kragge in parcel D, where the
vegetational aspect is determined by Carex lasiocarpa, is a large hollow, with ca 0.2 m standing water over a mud
bottom in the central part, grading into a Scorpidium moss layer with, locally, Drepanocladus sendtneri towards the
De Stobbenribben
125
margins. Types 32 (Bergmans: 4) and 37 (Bergmans: 3) refer to dense stands of Typha angustifolia and Cladium
mariscus, respectively.
Classification: Class Parvocaricetea, Order Tofieldietalia, Associations Scorpidio-Caricetum diandrae and
Scorpidio-Utricularietum, mixed with elements of the Order Caricetalia nigrae, Alliance Caricion curto-nigrae,
Association Sphagno-Caricetum lasiocarpae. The overall composition of the stands also reflects a relationship to the
Magnocaricion Alliance (Order Magnocaricetalia, Class Phragmitetea; see also Appendix B), which seems to apply
especially to the kragge synusia (Chapter 3). The hydro-environmental zone of the Scorpidio-Caricetum diandrae with
Scorpidio-Utricularietum hollow communities seems to be superimposed on the Magnocaricionkragge character.
Zone C: Referred to as: Atmo-oligotrophic (Sphagnum) fen, Sphagnum (quag)fen phase, Sphagnum-Carex fen
Bergmans 1975:
Types 9 and 11;
Types 53, 54, 56, 57;
Types 54, 55, 57
Occurring at a distance of some 100 m from the ditch towards the dead ends of the petgaten.
A closed Sphagnum cover, typically consisting of S.subnitens and S.flexuosum, and, locally, S.papillosum and
S.palustre, is characteristic of this zone. Juncus subnodulosus, Carex elata, and Phragmites australis are the most
abundant helophytes, although their stand is usually more open than in zone B. Carex panicea, Carex curta, Cirsium
dissectum, and Succisa pratensis, along with several species extending further into zone D are locally abundant. Carex
tumidicarpa, which is also abundant in zone B, has shown a substantial increase in numbers in De Stobenribben since the
late 1970s. Dwarfshrubs and Molinea coerulea invade the hummocks in zone C.
Classification: This zone includes elements of a range of phytosociological units. Although these elements are
probably linked by successional relationships, the pattern is often fine-scaled and expressed as a hummock-hollow
pattern. The phytosociological units include:
Class Parvocaricetea, Order Caricetalia nigrae, Alliance Caricion curto-nigrae, Associations Caricetum curtonigrae (weakly represented), Sphagno-Caricetum lasiocarpae, Pallavicinio-Sphagnetum (weakly represented);
Class Oxycocco-Sphagnetea, Order Sphagnetalia magellanici, Alliance Erico-Sphagnion, Association Sphagnetum
palustri-papillosi.
Zone D: Referred to as: Litterfen
Bergmans 1975:
Types 10, 12, 13;

?;
57, 58, 59
This zone typically consists of zone C stands with various species indicating the effects of more frequent access by
people walking in from the baulks, and of disruptions of the moss cover due to machine damage. The zone is liable to
such damage since it is located where the machines turn around during mowing and since especially Molinea and certain
other phanerogam species provide hummocks of a firm peat not very resilient under machine attacks. It is possible that
this zone also suffers from slightly greater fluctuations of the water level and an associated superficial drying in summer.
Calamagrostos canescens, Rubus fruticosus and Lysimachia vulgaris are more abundant in this zone than elsewhere in
De Stobbenribben.
Classification: Any accurate classification seems difficult, but elements of the Class Molinio-Arrhenateretea, Order
Molinietalia, Alliance Filipendulion are weakly characteristic of this zone.
Zones B and C together will sometimes be referred to as the (spatially) intermediate quagfen
zone.
126
De Stobbenribben
Table 7.1
The distribution of species over zones A-D in the vegetation of De Stobbenribben

(De Weerribben)
species
xy
ABCD
321
+1
11
55
33
13
11
++
+
3333
3333
3333
3333
3333
3333
3333
3333
1111
112
000
201
455
100
112
200
101
1111
55
33
13
1
13
5333
5333
5331
3101
3111
3010
3010
3001
3001
3001
3000
3000
3000
3000
3000
3000
1000
1000
1000
1000
1000
1000
1000
1000
3030
1010
1010
wat
nut
typ
ATM
OLI
OLI
EQUALLY FREQUENT IN ALL ZONES (COM)
Agrostis canina
Cirsium palustre
Galium palustre
Juncus subnodulosus
Lysimachia thyrsiflora
Potentilla palustris
Thelypteris palustris
Viola palustris
Angelica sylvestris
Plagiothecium denticulatum
var. undulatum
ESPECIALLY IN ZONE A (A)
Carex elata
Phragmites australis
Calliergonella cuspidata
Caltha palustris
Eupatorium cannabinum
Hypericum tetrapterum
Rhizomnium pseudopunctatum
Berula erecta
Iris pseudacorus
Rorippa amphibia
Acorus calamus
Lemna gibba + L. minor
Lemna trisulca
Nasturtium microphyllum
Poa trivialis
Rumex hydrolapathum
Drepanocladus aduncus
Epilobium hirsutum
Eurhynchium praelongum
Hydrocharis morsus-ranae
Lophocolea heterophylla
Myosotis palustris
Oenanthe aquatica
Rhytidiadelphus squarrosus
Chiloscyphus pallescens
Funaria hygrometrica
Lycopus europaeus
ESPECIALLY IN ZONE B (B)
Lythrum salicaria
Peucedanum palustre
Valeriana officinalis
Chara spec.
Calystegia sepium
Liparis loeselii
Nymphaea alba
Typha angustifolia
Utricularia vulgaris
Aneura pinguis
Cladium mariscus
Dactylorhiza majalis
ESPECIALLY IN ZONE C (C)
Sphagnum contortum
Sphagnum subnitens
Alnus glutinosa
Carex panicea
De Stobbenribben
1
+1
11
+
1
+1
53
11
13
11
33
11
11
1
15
11
+
1
522
544
422
000
200
000
000
100
422
CIR
LTH
CIR
ATM
CIR
ATM
LTH
OLI
OLI
MES
MES
OLI
LIT
MOL
SMP
LIT
SMP
FEN
FEN
FEN
LIT
CIR
EUT
LIT
CIR
LTH
MES
MES
MES
MES
EUT
OLI
MES
EUT
EUT
EUT
EUT
CIR
CIR
LTH
EUT
EUT
EUT
EUT
EUT
EUT
EUT
EUT
MES
MES
MES
MES
MES
EUT
MES
FEN
SMP
LIT
MOL
LIT
MOL
DAV
SMP
SMP
SMP
SMP
AQU
AQU
SMP
SMP
SMP
SMP
LIT
SMP
AQU
LIT
MOL
SMP
MOL
LIT
LIT
SMP
LTH
CIR
CIR
LTH
LTH
OLI
MES
OLI
OLI
MES
LAS
LIT
FEN
LAS
LIT
EUT
OLI
EUT
MES
MES
OLI
OLI
OLI
LIT
LAS
AQU
SMP
SMP
LAS
SMP
MOL
MES
MES
EUT
OLI
LAS
LAS
LIT
MOL
CIR
LTH
CIR
CIR
CIR
CIR
CIR
CIR
LTH
CIR
LTH
CIR
LTH
LTH
CIR
CIR
ATM
LTH
1531
1311
1311
1310
1301
1300
0100
0100
0311
0311
0311
0310
0310
0301
000
CIR
LTH
CIR
LTH
CIR
LTH
LTH
CIR
1031
0353
0131
0131
011
050
000
000
CIR
CIR
CIR
ATM
100
010
#
#
100
100
#
!
!
!
127
species
Sphagnum papillosum
Agrostis stolonifera
Carex rostrata
Cirsium dissectum
Riccardia multifida
Pinus sylvestris
Atrichum undulatum
Filipendula ulmaria
Prunella vulgaris
Quercus robur
Sphagnum russowii
Sphagnum teres
Sphagnum flexuosum
Aulacomnium palustre
Carex curta
Cephalozia connivens
Succisa pratensis
ESPECIALLY IN ZONE D (D)
Rubus fruticosus
Calamagrostis canescens
Lysimachia vulgaris
Carex disticha
Lotus uliginosus
Ranunculus flammula
Rumex acetosa
Lathyrus palustris
Scutellaria galericulata
xy
33
+
+
ABCD
321
0050
0030
0030
0030
0030
0010
0010
0010
0010
0010
0010
0010
0053
0031
0031
0031
0031
004
333
1013
0113
0113
0001
0001
0001
0001
0101
0101
0101
33
1
33
1
3311
3311
3311
3300
33
1
33
1330
1330
1331
0110
0110
0331
+1
11
11
wat
nut
typ
ATM
OLI
MES
OLI
OLI
OLI
ATM
CIR
ATM
ATM
ATM
ATM
ATM
MES
OLI
MES
MES
OLI
OLI
OLI
BOG
LIT
FEN
MOL
LAS
BOG
FEN
LIT
MOL
BOG
BOG
FEN
FEN
FEN
FEN
FEN
MOL
ATM
CIR
CIR
CIR
CIR
CIR
LTH
CIR
CIR
EUT
OLI
OLI
MES
MES
OLI
MES
OLI
OLI
MES
LIT
LIT
LIT
LIT
MOL
LIT
MOL
LIT
LAS
SMP
CIR
CIR
CIR
CIR
OLI
MES
MES
MES
FEN
SMP
FEN
SMP
LTH
OLI
MES
OLI
OLI
EUT
OLI
DAV
DAV
FEN
BOG
SMP
FEN
#
#
#
MES
MES
OLI
MES
OLI
OLI
OLI
OLI
FEN
MOL
LIT
FEN
LIT
DAV
MOL
BOG
ATM
CIR
MES
MES
OLI
OLI
OLI
MES
OLI
OLI
ATM
ATM
OLI
OLI
ATM
ATM
CIR
ATM
ATM
LTH
024
002
000
000
000
000
100
EUT
MES
OLI
!
#
!
!
!
!
ESPECIALLY IN ZONES A AND B (AB)
Carex lasiocarpa
Mentha aquatica
Utricularia minor
Carex acutiformis
000
100
331
ESPECIALLY IN ZONES B AND C (BC)
Campylium stellatum
Hypnum jutlandicum
Carex pseudocyperus
Carex echinata
200
000
ATM
LTH
ATM
ESPECIALLY IN ZONES C AND D (CD)
Calypogeia fissa
Holcus lanatus
Mnium hornum
Sphagnum fimbriatum
Thalictrum flavum
Betula pubescens
Luzula multiflora multiflora
Anthoxanthum odoratum
Cephalozia bicuspidata
var. lammersiana
Dicranum bonjeanii
Erica tetralix
Molinia caerulea
Polytrichum commune
Polytrichum longisetum
Potentilla erecta
Salix repens
Sphagnum palustre
Luzula campestris
Campylopus fragilis
Sorbus aucuparia
Stellaria palustris
LESS FREQUENT IN ZONE D (-D)
Cardamine pratensis
128
1133
1033
1033
1033
1033
0133
0133
0033
010
000
313
010
012
000
002
CIR
ATM
CIR
ATM
CIR
1+
0033
0033
0033
0033
0033
0033
0033
0033
0033
0011
0011
0011
0011
000
CIR
MES
FEN
FEN
BOG
MOL
FEN
LIT
MOL
MOL
FEN
BOG
LIT
LIT
FEN
33
3331
301
CIR
MES
MOL
001
CIR
ATM
ATM
ATM
ATM
024
000
!
!
De Stobbenribben
species
xy
ABCD
321
wat
nut
typ
Carex diandra
Carex paniculata
Dryopteris carthusiana
Epilobium palustre
33
11
1
1+
3331
3331
3331
1110
002
401
CIR
LTH
ATM
LTH
OLI
EUT
OLI
MES
LAS
SMP
FEN
FEN
000
LESS FREQUENT IN ZONE C (-C)
Vicia cracca
1101
CIR
OLI
MOL
MES
SMP
ATM
CIR
CIR
CIR
CIR
OLI
OLI
MES
EUT
MES
LIT
MOL
LAS
LIT
MOL
LTH
ATM
MES
OLI
OLI
OLI
MES
LESS FREQUENT IN ZONES B AND C (-BC)
Carex riparia
1001
LESS FREQUENT IN ZONE B (-B)
Juncus conglomeratus
Valeriana dioica
Pellia neesiana
Sphagnum squarrosum
Lychnis flos-cuculi
+
1
3033
3033
3031
3133
1011
000
021
!
!
!
LESS FREQUENT IN ZONE A (-A)
Salix aurita + cinerea

Carex tumidicarpa
Drosera rotundifolia
Dryopteris cristata
Hydrocotyle vulgaris
Equisetum fluviatile
Frangula alnus
Myrica gale
1
++
+1
1333
0333
0333
0333
0333
0111
0111
0111
045
212
000
MES
SMP
MOL
FEN
FEN
LIT
FEN
FEN
FEN
CIR
LTH
EUT
EUT
MES
MES
OLI
MES
MES
EUT
SMP
LIT
LAS
LAS
FEN
MOL
LIT
SMP
LTH
CIR
ATM
CIR
LTH
LTH
CIR
ATM
CIR
MES
OLI
EUT
EUT
MES
EUT
OLI
OLI
OLI
LAS
DAV
LIT
LIT
SMP
SMP
SMP
FEN
LIT
CIR
CIR
CIR
CIR
EUT
MES
OLI
MES
SMP
SMP
MOL
SMP
ATM
CIR
LTH
CIR
NOT MENTIONED BY BERGMANS (?)
Alisma plantago-aquatica
Brachythecium rutabulum
Calliergon giganteum
Calliergon stramineum
Campylium elodes
Campylium polygamum
Cicuta virosa
Cladophora fracta
Pellia epiphylla
Plagiomnium affine
Ranunculus lingua
Sium latifolium
Sparganium minimum
Eriophorum angustifolium
Hierochlo odorata
Taraxacum cf. T. palustre
Calliergon cordifolium
Sparganium erectum
Galium uliginosum
Riccia fluitans
++
11
33
33
1
31
11
11
11
31
11
3
3
33
++
+
100
CIR
000
LTH
LTH
ATM
001
000
000
000
000
000
001
#
!
#
#
#
!
!
!
At the right-hand side indicator values have been listed as treated in chapter 3; the species have been grouped according to their preference for specific zones.
The occurrence of the species in permanent quadrats 1 and 2 ca 1960 (columns x, y) and in samples reported by Kooijman (1985, observations 1984) has been
indicated also (1-3, where 3 is closest to the ditch).
wat water type; nut nutrient state; typ phytosociological group; r red-list (! red-list species; # red-list species and seepage indicator)
De Stobbenribben
129
130
Peat temperature
CHAPTER 8
Peat temperature and the estimation of vertical

water flow
8.1 Introduction
Conventional hydrological survey methods are especially useful for the assessment
of the average characteristics of a whole area. In order to explain the vegetational
pattern within single petgaten, an additional, more detailed assessment of water
flow was deemed desirable. One appropriate method uses a heat transport model
for the estimation of seepage, the use of which was inspired by the frequent
mentioning of reduced temperature fluctuations as a result of upward seepage in
quagfens in North-West Overijssel. S.Segal (pers. comm.) carried out temperature
recordings in various shallow pools and he thought that the results demonstrated
differences in seepage intensity. While I was unable to reproduce such
measurements with enough accuracy to allow for any positive conclusion (Van
Wirdum 1972), the quantitative treatment of the problem suggested a different
approach. Instead of measuring in the immediate environment of plants, the
temperature sensors were moved deeper into the peat, where the effect of seepage
would be measurable in the characteristics of the sinusoidal annual temperature
wave. Although many complicating factors were discovered, the results
contributed to the understanding of the water flow through De Stobbenribben and
other terrestrialized petgaten.
Peat temperature
131
Daily and annual fluctuations of the temperature at the soil surface cause a
flow of heat between this surface and deeper strata, where the temperature is
more constant. Any vertical flow of water contributes to this heat transportation,
so that accurate measurements of temperature profiles in the soil enable an
estimation of the vertical component of water flow. Such estimations were made
with an increasing accuracy and reliability from 1969 onward. Van Wirdum
(1984) provides a summary of the development of gauges, physical models, and
methods of computation regarding the heat transportation method for the
estimation of seepage.
In the first section of this chapter a tentative theory is presented, including
studies of physical models, a survey of thermal properties of wet peat soils, and
an investigation of the periodicity of temperature fluctuations slightly above the
soil surface. The main questions to be answered in the next sections are:
Can seepage in De Stobbenribben and similar quagfens be assessed with
thermal methods?
If so, are there appreciable local differences, and can these differences have
an ecologically significant influence on the temperature regime at the soil
surface and in shallow pools?
I focus on the vertical component of water flow in this chapter. The lateral
flow of water will be considered in some detail in Chapter 9.
Table 8.1
Symbol
A
a
b
c
k
T
t
v
z
List of symbols, dimensions and units

Dimension
L-1
L-1
L2 T-2 -1
L M T-3
T
L T-1
L
L2 T-1
L-3 M
T
T-1
-1
Units used
K or 0C
m-1
m-1
J kg-1 K-1
W m-1 K-1
0
C
s
m s-1
m
m2 s-1
kg m-3
s
rad
rad s-1
Comment
temperature amplitude
damping coefficient
phase-shift coefficient
specific heat
thermal conductance
temperature
phase, relative to the wave minimum
gross velocity of water
depth
thermal diffusivity
density
period of the wave
phase, relative to the wave minimum
angular velocity of wave
A list of symbols, dimensions and units used in this chapter is provided in Table
8.1. In the text, v is given in mm/d for convenience. A positive sign indicates a
downward flow. Also for convenience, t is mostly given in days. The product c is
known as the heat capacity. The attribute 0 refers to water in the case of thermal
132
Peat temperature
properties. Elsewhere, 0 refers to the soil surface; is used to indicate an

infinite depth. The temperature gradient at great depth has not been considered,
however, and the geothermal heat flux is disregarded. The values presented for the
thermal properties of soil constituents in this chapter hold for an ambient
temperature of 10 0C, which is close to the mean annual air and groundwater
temperatures in The Netherlands.
8.2 Theory of soil temperature as a function of seepage
In this section, a general model of temperature fluctuations in wet peat soils is described. Two
extensions of the model are presented for the assessment of the flux density of water along the
vertical axis. The first uses an analogue of the Doppler effect and was developed during this
investigation. Although it appears to be in conflict with the continuity principle, experience
suggests that this model is less sensitive to some errors which result from the non-ideality of
actual situations. The simplest implementations of the model, however, are probably very
sensitive to a lateral flow of water and to the occurrence of extremely warm or cold periods. At
the lowest level of accuracy, the Doppler analogy can be applied already to a single temperature
profile comprising three measured temperatures.
The alternative extension of the model, borrowed from Stallman (1965), is consistent with the
physics of heat transfer. The procedure used for the derivation of values for the damping and
phase-shift coefficients, and the computational solution of the thermal diffusivity of the peat,
respectively, were newly developed.
The general model
Under the influence of a varying soil-surface temperature and a constant soil temperature at
infinite depth (z ), the temperature of the soil at some intermediate depth z will reflect the
fluctuations at the surface (z0) with a decreased amplitude and a phase lag. The theory of the
attenuation and phase shift with depth of the annual and daily temperature waves at the soil
surface by heat conduction alone has been dealt with by several authors (see Rose 1969),
resulting in
(Formula 8.1)
Tt,z = Tm - A0 e-bz cos( - bz), with b = ( /( ))1/2 (and
= t = 2 t/ ),
where Tm is the mean annual temperature, which is almost constant for the depths studied. Rose
(1969) mentioned the influence of convective heat transport and proposed the determination of
an effective value of , which should then be regarded an apparent diffusivity only. When the
thermal parameters of the soil medium are known and when they are nearly equal to those of
water, the gross velocity of the vertical flow of water can be estimated from the difference
between the apparent and real thermal diffusivities. Van Wirdum (1972) used the analogy
with the Doppler effect in this connection, by considering the penetration of the temperature
wave in a moving, semi-infinite body of water, yielding:
(Formula 8.2)
v = (1/b' - 1/b) 2 / ,
where b' results from an apparent . In this formula, v is the gross velocity of the medium as a
Peat temperature
133
Fig. 8.1
The relation between v and a, b, and b, respectively, in very wet peats

( = 1.35 10-7 m2 s-1)
whole, which, in the case of the very wet peat soils investigated here, contains more than 95%
by volume of water. The remaining part of the medium is the organic matrix, which may hold a
large volume of water in place by physical bounds or in closed (dead) pores. The actual
velocity of the moving fingers of water in preferential flow channels may well be much larger
than the calculated v. The gross velocity calculated here should be nearly equal to the flux
density or Darcy velocity as used in hydrological models. Note that the Doppler method is
based on the observed wave characteristics, rather than on the laws of conservation of mass and
energy and the relevant principles of continuity. The relation between v and b' in very wet peat
soils has been depicted in Fig.8.1.
A report by De Crook (1979) directed my attention to work by Stallman (1965) and Suzuki
(1960), who showed that, in order to satisfy continuity requirements, equation 8.1 should be
adapted to the case of combined conduction and convection of heat:
(Formula 8.3)
Tt,z = Tm - A0 e-az cos( - bz),
where a=b only if there is no convective heat transport. This means that the respective effects of
moving water on the amplitude and the phase of the temperature wave are different. The
following relations were derived by Stallman:
134
Peat temperature
(Formula 8.4)
and (Formula 8.5)
a = ((K2 + V4/4)1/2 + V2/2)1/2 - V,

b = ((K2 + V4/4)1/2 - V2/2)1/2,
where K = /( ) (= /(2 )), and V = (v
0c0)
/ (2
c).
This physical model holds true under the following conditions:

1) The flow of water in the soil layer considered is steady and uniform along the z axis;
2) The thermal properties of the water and the medium are constant in space and time;
3) All components of heat and fluid flow occur only along the z axis;
4) The temperature of the water at every point in the interstices equals the temperature of the
adjoining soil particles at all times.
Stallman suggests a graphical representation of measured temperatures to find the pertaining
values of a and b. Note that b is not sensitive to the sign of v, i.e., to the direction of the vertical
flow component. Stallman provides a table and graphs for the solution of 2V(K-1/2) as a function
of a/b, followed by the substitution of the value of (b2-a2)/a for 2V in order to find K and,
accordingly, . V is then solved from the formula defining V (see under Formula 8.5), where c
should be known. The relation between v and a and b, respectively, in very wet peat soils has
been depicted in Fig.8.1. An alternative, computational method to solve is derived below.
Equation 8.4 can be rewritten, using 2V = (b2-a2)/a, in the following form:
(Formula 8.6)
a + (b2- a2) /2a = ((K2 + V4/4)1/2 + V2/2)1/2.
Multiplication of (8.5) by (8.6), using the reverse of the well-known expansion formula
(p2-q2)1/2 = (p+q)1/2 (p-q)1/2, yields
K = (1/2) ab (1 + b2/a2).
This enables the solution of from the defining formula for K (see under Formula 8.5). In the
working scheme proposed here, a and b are determined by means of exponential and linear
regression, respectively, for
Az = A0 e-az, and
0-
bz
with Az and z data regarding the fundamental wave as obtained by Fourier analysis of Tt,z field
data. Now v may be found if c is known. A fixed value c = 4 106 J m-3 K-1 was used here.
Values of and c for wet peat soils are derived in the next subsection.
In summary, may be determined if
1) Sufficient Tt,z data are available reliably to determine the phase and amplitude of the
fundamental component of the annual temperature wave at depth z, for instance by graphical
or Fourier analysis;
2) Such data are available for at least two different depths whose distance z is known,
Peat temperature
135
and v may be determined if

1) c is known or may be estimated with sufficient accuracy from the volumetric composition
of the soil.
Problems may arise when, among others,
1) A considerable convective heat transport is present in directions other than vertical;
2) Variation of the moisture content with depth causes variation of ;
3) Sources or sinks of heat within the soil profile may not be disregarded;
4) Higher harmonics of the temperature wave may not be disregarded.
The conditions 1, 2, and 3 are considered the main drawbacks for the application of this physical
model. In certain cases the actual flow may be so much restricted to singular preferential
channels that the temperature of the flowing water does not equal that of the surrounding soil
matrix. Such cases have probably not been very important in the present investigation, but have
been suggested as a cause of error when the same method was applied to more consolidated bog
peat soils elsewhere in The Netherlands, where the results obtained with the Doppler analogy
method seemed to make more sense.
The first source of error is investigated from a theoretical point of view in Section 8.4. The
calculations indicate a considerable downward flow of water in De Stobbenribben, so that a
lateral inflow must be large enough to evoke serious errors in the quantitative interpretation at
the depth where this flow occurs. The presence of a lateral flow in the expected order of
magnitude is obvious from the independent data presented in Chapter 9. The use of time series,
as in this investigation, reduces the chance that large errors result. As shown in the next section,
the second source of error may be disregarded in the very wet peat soils considered here. Error
type 3 may also be left unconsidered as far as it is not caused by the first-mentioned type of
problem. A discussion of error source 4 is given later in the present section. The results provided
an argument for the application of Fourier analysis to the measured time series.
In places where the hydraulic conductivity of the porous medium is very large, density
currents can be generated during winter. This problem was investigated in connection with the
study of longitudinal transects as described in Chapter 9. It does probably not constitute a
serious problem in De Stobbenribben.
Heat capacity and thermal diffusivity of very wet peat soils
The rate of thermal diffusion (thermal diffusivity) through wet peat soils is considered a known
parameter in the Doppler analogy model discussed before. In the Stallman model, the thermal
diffusivity is resolved with temperature data and a heat capacity value as inputs, and it would be
desirable to have at one's disposal any reference value for the thermal diffusivity in order to
detect anomalous results. Such values are provided here. In order to avoid confusion, a concise
summary of the thermal characteristics of the porous medium and of what they depend on is
given in Table 8.2. The values of the various parameters also depend on the ambient
temperature. The values selected here are approximately correct at 10 0C. Many values found in
the literature appear to apply to ambient temperatures well above those normally found in soils
in The Netherlands. Especially in the case of thermal conductance this may cause serious errors.
136
Peat temperature
Table 8.2
Thermal characteristics of porous media
specific heat (c)

density ( )
heat capacity ( c)
thermal conductance (k)
thermal diffusivity ( )
damping coefficient (a)
phase shift coefficient (b)
depends on medium composition

depends on medium composition
product of specific heat and density
a property of the various soil components
depends on thermal conductance and heat capacity
depends on thermal diffusivity and wave period; in case of mass
flow also on heat capacity of flowing material and on flow
velocity
as damping coefficient
For dimensions and units: see Table 8.1
The thermal diffusivity of a medium of heat transport can be calculated from

(Formula 8.7)
= k/( c).
Since the type of medium under discussion (soil) is composed of solids of different kinds, and of
interstitial water and, possibly, air, c is a weighted sum:
(Formula 8.8)
c = 1/V (Vi ici),
where V is the volume considered and where the specifier i denotes the different materials of
which the soil is composed: sand, clay, organic matter, water, and air. Taking sand and clay
together as minerals, and substituting the accurately known values for and c this formula
reduces to (De Vries 1963):
c = 1/V (1.93 V1 + 2.5 V2 + 4.19 V3 + 0.0013 V4) 106,
where the specifiers 1, 2, 3, 4 denote minerals, organic matter, water, and air, respectively. The
term with V4 may be deleted due to its very small value. De Vries (l.c.) provides the following
formula for the calculation of the thermal conductance k:
(Formula 8.9)
k= (
Vi ki) / (
Vi),
where the specifier i denotes each particular combination of form and material (each type of
granule) present, and where is a parameter which depends on the type of granules and on the
medium in which they are embedded. According to De Vries the value of can be reliably
estimated by
(Formula 8.10)
= 1/3 (1 / (1 + gk (ki / k0 - 1)),
where 0 specifies the embedding medium (the volumetrically dominant component), and the
summation is over three components specified by k, the soil particles being conceived as
ellipsoidal bodies with a so-called factor of depolarization gk associated with each of the three
axes. The factors of depolarization have the following characteristic values:
Peat temperature
137
spherical grains
laminae
terete fibres
g1 = g2 = g3 = 1/3
g1 = 1, g2 = g3 = 0
g1 = g2 = 1/2, g3 = 0.
The thermal conductance of various composite soils can now be calculated by means of the
material properties and resulting values for listed in Table 8.3.
It is apparent from this table that an average value, = 1.31, will do for very wet peats.
Fig.8.2 shows the variation of and c for such soils, depending on the volumetric composition.
Since sand and clay have not been found in any appreciable amount in the peat in De
Stobbenribben, and the peat is of a very soft and water-saturated type, = 1.35 10-7 m2 s-1
7.5%. The value c = 4 106 J m-3 K-1 is also quite reliable.
Table 8.3
Thermal conductance of single soil components and values of

various types of granules
Type of granules
Thermal conductance
grains of air
laminae of peat
fibres of peat
grains of peat
grains of clay
grains of sand
0.025
0.251
0.251
0.251
2.9
8.7
W m-1 K-1
W m-1 K-1
W m-1 K-1
W m-1 K-1
W m-1 K-1
W m-1 K-1
associated with
1.47
1.43
1.26
1.23
0.42
0.17
embedding medium: water, k = 0.574 W m-1 K-1
A wave analysis of annual temperature fluctuations

The driving force of the variation in soil temperature is the temperature fluctuation at the soil
surface, which can be described as a complex wave. Of all components of this wave the annual
and daily ones have the largest amplitudes. In order to simplify matters, only one component
was included in the formulas presented before. In this section it is investigated which component
is most useful for the present purpose, and whether problems may arise from the exclusion of
other components.
As follows from formula 8.3, the amplitude of the temperature wave is damped according to
Az = A0 e-az,
which enables a calculation of the attenuation. Results of such calculations have been collected
in Fig.8.3 for very wet peat soils. It is obvious from this figure that the annual fluctuations
penetrate to a much greater depth than the daily ones. Assuming an amplitude of 10 K around
138
Peat temperature
Fig. 8.2
Representative values of c and

composition of the soil
in wet peats, depending on the volumetric
the mean at the surface, Az is <0.01 K for the daily wave component at z=0.6 m and 5.97 K for
the annual one. The daily component can thus not be assessed on the basis of soil temperature
measurements in very wet peat soils, unless these measurements are performed close to the soil
surface, but such measurements would require an undue precision. Alternatively, when the
annual wave component is used, fluctuations with a short period will not cause any appreciable
noise if the measurements are not taken too close to the surface. Taking 0.6 m as a
representative depth, and considering an amplitude of 0.5 K critical, any wave component with a
period shorter than 20 days may be disregarded unless, for such wave components, A0 is >5 K.
The daily and annual amplitudes are rather constant between years, but the other components
vary largely, since they result from relatively warm or cold periods lasting only for several days
or weeks, i.e., from the irregularities of the weather. In order to get an idea of the occurrence of
such irregularities and their possible effects on the soil temperature below 0.6 m, a Fourier
analysis was made of the decade temperatures measured at De Bilt and Eelde by the Royal
Meteorological Service (KNMI). These temperatures are measured with standard equipment in
meteorological boxes at 1.5 m above the soil surface. According to the KNMI (1972) records the
temperature in North-West Overijssel is about the mean of the values for Eelde and De Bilt. The
mean and amplitude of the wave components with periods longer than a few days, measured at
the soil surface and in a meteorological box, respectively, probably do not differ very much, as
Peat temperature
139
Fig. 8.3
Damping of temperature oscillations with depth in wet peats

( =1.35 10-7 m2 s-1)
Numbers denote the period ( ) for each line or mark in days; note that the y-axis is logarithmic
is confirmed by a comparison of the data in this section with those in the next. The mean
temperature in each decade has been used here as the value measured on the fifth day of that
decade. The use of decade values restricts the analysis to wave components with a period of 20
days or longer. A summary of the most relevant results is presented in Table 8.4.
The last column of the table gives the value of A0 required for any singular wave component
in order to cause a deviation of the temperature at 0.6 m which is equal to or larger than 10% of
the amplitude of the annual component at the same depth. This may be used as a criterion to
judge whether there is a fair chance that the neglect of overtones in the formulae causes
erroneous results.
Since the various overtones have different phases, the real noise must be calculated for each
year separately by the superposition of the various wave components and a comparison with the
annual one. This has been done by extending formula 8.3, with a=b, to incorporate all 18
components listed in Table 8.4. Note that, for each period, a (and b) have different values (see
formula 8.1). The results have been summarized by calculating the standard deviation of the
contribution of the overtones to the decade temperatures at a depth z, where the decade
temperatures are expressed by (Tt,z - Tm,z). These contributions can be considered normally
140
Peat temperature
Table 8.4
Summary of Fourier analyses of the annual wave of the ten-days mean air
temperature, 1.5 m above the soil surface (averaged values from KNMI stations
De Bilt and Eelde).
Period 1970
(days)
1971 1972 1973 1974 1975
1976
1977
1978
1979
mean
365
(fundamental tone)
Tm 8.68 9.28 8.61 8.98 9.33 9.48 9.50 9.39 8.69 8.10 9.00
L 37
35 29
32
29
37
34
36
34
36
34
A0 8.75 7.38 7.22 7.81 5.52 7.47 8.57 6.48 7.14 8.57 7.49
A0 of overtones:
183
1.06 0.42
122
0.41 0.68
91
0.99 0.86
73
0.35 0.36
61
0.33 0.85
52
0.91 1.45
46
0.87 0.59
41
0.52 0.60
37
0.46 0.43
33
0.37 0.64
30
0.30 0.17
28
0.52 1.22
26
0.89 0.71
24
0.76 0.25
23
0.50 0.15
21
0.09 0.75
20
0.37 0.44
0.31
1.30
0.48
0.39
0.14
0.72
0.33
0.26
0.93
0.58
0.87
0.39
0.33
0.60
0.49
0.11
0.29
1.51
0.48
0.37
0.56
0.74
0.38
0.90
0.68
0.28
0.45
0.34
0.40
0.65
0.24
0.29
0.46
0.37
1.33
1.13
0.53
1.06
0.48
0.61
0.65
0.74
0.24
0.22
0.63
0.14
0.28
0.61
0.26
0.60
1.14
2.28
0.92
0.46
0.16
0.52
1.44
0.17
0.54
0.12
0.32
0.54
0.73
0.28
0.73
0.76
0.37
0.35
0.79
0.52
0.73
0.33
0.88
1.00
1.90
0.91
1.06
0.17
0.28
0.45
0.42
0.89
0.83
0.03
0.17
0.66
0.98
0.22
0.82
1.08
0.26
0.93
1.12
0.16
0.16
0.14
0.39
0.38
0.37
0.66
0.15
0.63
0.67
0.41
0.32
1.45
0.78
0.96
0.81
0.14
0.48
0.95
0.69
0.47
0.37
0.76
0.36
0.32
0.09
1.02
0.77
1.37
0.59
1.34
0.89
1.06
0.77
0.83
0.37
0.42
0.58
0.25
0.87
0.92
0.33
0.32
1.01
0.76
0.63
0.61
0.71
0.86
0.82
0.63
0.50
0.42
0.44
0.53
0.46
0.61
0.52
0.32
0.42
0.9
1.1
1.2
1.4
1.6
1.7
1.9
2.1
2.2
2.4
2.6
2.8
3.0
3.2
3.5
3.7
3.9
Tm: mean temperature (0C); L: phase lag relative to January, 1st (days); A0: amplitude (K); N: A0 required to cause a
serious chance of 10% effects on the temperature at 0.6 m below the soil surface (K; see text); Values of A0 in excess of
N have been printed in bold face.
distributed around 0, thus enabling a probability computation. By taking 0.6 m as a critical

depth, the probability of these contributions to exceed 10% and 20% of Az is listed in Table 8.5.
A possible incidence of 8.3% at the 20% contribution level compares to one out of 12
monthly temperatures having a deviation exceeding the one expected according to formula 8.3 at
0.6 m depth by 0.9 K or more. At greater depths the probability of such deviations decreases
strongly, as is illustrated for the year 1979 in Fig.8.4. This figure illustrates the expected
temperature according to formula 8.3 next to the one according to the complete Fourier series,
where it was assumed that the decade temperatures in the meteorological box equal the
temperatures at the soil surface in the same decade, and where the calculations were performed
for a wet peat soil with conductive heat transport only. The differences between the measured
decade temperatures and the Fourier series at z = 0 are due to numerical errors, since the Fourier
Peat temperature
141
Table 8.5
The possible incidence of overtones contributing more than 10% (P10) or 20%
(P20) of Az to the measured temperatures Tt,z at z=0.6 m
Year
Az
K
s( T)
K
P10
%
P20
%
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
mean
5.23
4.41
4.31
4.66
3.30
4.46
5.12
3.87
4.26
5.12
4.47
0.61
0.52
0.50
0.63
0.68
0.91
0.60
0.54
0.54
0.71
0.62
39
40
39
46
62
62
40
47
43
47
47
9
9
9
14
33
33
9
15
11
15
15
Az is listed for the annual wave component at 0.6 m

s( T): standard deviation of contribution of overtones to actual temperature at z=0.6 m
series provides an exact description of the measured data. This figure also shows that, as the
depth increases, the remaining deviations from the fundamental sine wave occur in a grouped
manner. This is due to the fact that only deviations resulting from overtones with relatively long
periods can remain of importance at greater depths. Especially measurements performed during
a short period of up to about two months may thus suggest an annual wave different from the
actual one.
It can be concluded that the use of only one wave component (viz., the annual one) in this
investigation provides a fair approximation of the real situation, although it is not entirely
reliable. The single-profile implementations of the Doppler analogy method are very sensitive to
local deviations of the measured temperature. For this reason time series are being used here,
and the annual component is singled out by Fourier analysis in the implementation of the
Stallman method. The results given in the following sections do not suffer from serious errors of
this kind, therefore, and the values derived for a and b must be considered more reliable than
those obtained with graphical methods as proposed by Stallman (1965).
8.3 The estimation of seepage in De Stobbenribben
The gauges for temperature measurements
The gauges used for measuring temperature profiles will not be discussed in detail here. The first
was provided by the TFDL (Technical and Physical Engineering Service for Agricultural
Research, Wageningen) in 1969: a sounding rod with a thermistor mounted in the tip. An
improved version was built at RIN in 1977, and this one was regularly calibrated. The sensor tip
was replaced by a specially designed air chamber in 1979 at RIN. This reduced the reaction
142
Peat temperature
Fig. 8.4
Data and simulated penetration of annual course of meteorological-box

temperature
Mean temperatures of ten-day periods from De Bilt and Eelde meteorological stations averaged
time of the instrument considerably, and thus the errors associated with too short an
equilibration in the soil at each depth The measurements described in this report were made with
this instrument. The accuracy in the laboratory was 0.02 K, in the field it will have been between
0.05 and 0.1 K. The further development of the instruments has been summarized by Van
Wirdum (1984).
Peat temperature
143
Simple implementations of the Doppler analogy method (DOPPSOL)

In 1969 and 1970 the soil temperature was measured at several places in De Weerribben. The
data were entered in diagrams and compared with standard graphs, produced by means of
goniometrical methods in accordance with formulas 8.1 and 8.2. The standard diagrams were
based on a variety of expected values of Tm, A0, t, and b', assuming a fixed value b = 0.86 m-1
in case v = 0. It was concluded (Van Wirdum 1972) that these observations indicated a
downward water flow rather than an upward one, and this conclusion was backed by analyses of
the chemical composition of the mire water at several depths. Essentially the same method was
used in 1973-'74. From 1975 to 1980 the comparison of graphs was preceded by a computation
on an HP-25 programmable pocket calculator. This method, which still proves useful in several
situations, has been reprogrammed for various other computing devices.
It runs as follows3:
1) Select three depths z1, z2, and z3 and their temperatures for one record;
2) Choose a value for b' in formula 8.2 and use this for b in the formulae below;
3) Calculate the following parameters:
-bz
cos(-bz) for z1, z2, and z3;
z=e
-bz
=
e
sin(-bz) for z1, z2, and z3;
z
C1 = z2 - z1, C2 = z3 - z2;
S1 = z1 - z2, S2 = z2 - z3;
= S1C2 - C1S2;
(Note that, for the same set of three depths, and for b'=b, these calculations need only be
performed once!)
4) Calculate
D1 = C2(Tt,z1 - Tt,z2) - C1(Tt,z2 - Tt,z3), and
D2 = S1(Tt,z2 - Tt,z3) - S2(Tt,z1 - Tt,z2) (the formula cos(
) = cos( ) cos( ) sin( ) sin( ) is
used);
5) Solve A0 from A02 = (D12 + D22)/ 2 (since sin2 ) + cos2 ) = 1);
6) Solve t= /(2 ) from tan( ) = D1/D2 (since tan( ) = sin( )/cos( ));
(Note that there are two solutions, one half period apart!)
7) Solve Tm from Tm = Tt,z + zD2/ - zD1/ ;
8) With the obtained values, compute Tt,z for several depths according to formula 8.1 and draw
a graph of the results for comparison with the observed values.
Mathematically this basic DOPPSOL method yields exact solutions. The reliability of these
solutions depends on the choice of b', on the accuracy of the measurements, on the correctness
of the model in the given situation, and, to some extent, on the dates of the measurements
relative to the phase of the annual temperature wave. It is possible to vary b' in an iterative
approach in order to find limiting values between which the solutions of A0, Tm, and t,
respectively, cannot be rejected. It is now possible to use formula (8.2) in Section 8.2.1 to solve
v, since b = 0.86 m-1 will approximately hold true for most soft and very wet peat soils, such as
those in the area investigated here. A final check is possible by drawing a calculated temperature
profile, which can easily be calculated from formula 8.1.
In a non-Doppler approach, the same recipe can be used with a and b (formula 8.3), scenarios for v supplying the values to
be used in steps 2 and 3.
144
Peat temperature
Summary of results with varieties of DOPPSOL

Many varieties of the DOPPSOL method were developed in the course of years. Especially
some of these varieties were designed to use several sets of three temperatures out of one profile
so as to find a fit and to evaluate an average solution. The general use of 0.6, 1.0, and 1.6 m for
z1 through z3 yielded satisfactory results. The Doppler analogy method was also extended so as
to take account of measurements at several dates. In this extension it was possible to compute a
best fitting value for b', provided enough data were available, and somewhat depending on the
time of the year. Some results for De Stobbenribben have been summarized in Table 8.6, under
the heading DOPPSOL findif, indicating a finite-differences solution. A downward water
movement with a gross velocity of 4.8 mm/d (mean) was found. Similar results were obtained
by graphically estimating the attenuation and the time-lag of the temperature wave in time series
of temperature profiles, but this procedure consistently suggested different values for b' as
derived from the amplitude damping and the phase shift, respectively. This difference was
confirmed by the application of Fourier analysis to the time series, and initiated a further
literature search leading to the implementation of the Stallman model.
Implementations of the Stallman model (FOUSOL)
The work of Suzuki (1960) and Stallman (1965) led to the development of the method
introduced in Section 8.2.1. In 1981 a set of programs was written in order to implement this
method on an HP-41C programmable calculator with a cassette drive and a printer. The
implementation is called FOUSOL here after the name of the main program involved. The
FOUSOL method runs as follows:
1) Enter monthly temperature profiles into a data file. In our case each station yields 12 (data)
times 16 (depths) numbers, which refer to temperature with a precision of 0.01 K;
2) Run a Fourier analysis of these data, and save the mean temperature, the amplitude, and the
phase of the fundamental tone for each depth in a second data file;
3) Run the main program FOUSOL, which requires the second data file. FOUSOL needs
instructions with regard to the possible subdivision of the soil profile into layers. With 16
depths represented, the soil may be subdivided into 1 to 15 layers, which is specified by the
number of depths to be used per layer (2 n 16). The solution is thus given for a window
which moves downward along the z axis. The solution of a and b uses exponential and linear
regression, respectively, whenever n>2. A fixed value may be specified for c.
The problem of estimating the sensitivity and accuracy of this method has not been completely
solved. Stallman suggests that the value of a must be affected as much as, or even by more than,
5% by the liquid flow to ensure a satisfactory observation of v. Referring to Fig.8.1, this
compares, in the case of very wet peat soils, to v 1 mm/d. How realistic this is depends on
the errors in a and b, which are difficult to assess. From the range of the thermal diffusivity
derived in Section 8.2 it can be concluded that the maximum value for b, for the annual wave,
falls between 0.83 and 0.89 m-1, with a typical value of 0.86. The results reported below show
that the maxima actually found there do not extend beyond this range. Assuming opposite errors
of 1-3% in a and b, the error in v would be 0.8-2.3 mm/d in the critical region where a
approximately equals b. Taking account of the coefficients of determination of the regressions
for a and b, I expect that this estimation is on the safe side, and that violation of model
Peat temperature
145
Table 8.6
petgat:
D
C
Comparison of results obtained by the application of different methods of solution

of v from soil temperature profiles in De Stobbenribben.
B
station nr:
11
16
17
12
13
8
9
18
14
10
1
2
3
4
5
6
171
171
171
123
73
123
73
123
73
25
25
25
183
159
139
99
57
20
method: FOUSOL, time series 790521- 800422

v (mm d-1):
date =0
(in February):
10.6
2.8
2.7
-1.1
-0.9
4.4
4.5
3.4
1.3
0.7
1.6
0.9
2.1
6.3
3.3
2.1
4.1
6.4
17
10
16
20
12
8
15
17
18
16
20
A0 (K):
11
15
15
16
16
15
7.7
8.5
7.6
9.2 7.5
10.0 7.3
8.9
9.5
8.9
9.5
9.4
A0 (K):
10.0
5.8
3.8
8.4
8.9
7.4
1.9
2.0
4.0
5.0
3.9
5.5
9.1
9.3
7.6
8.0
8.7
9.2
2.8
6.8
1.4
8.9
9.7
9.6
26
25
30
16
17
25
29
25
17
15
19
14
26
28
17
12
17
20
1.43
1.87
1.32
1.68
1.42
1.34
1.55
1.31
1.78
1.54
2.26
1.37
1.60
1.42
1.94
1.75
1.89
1.65
Tm (0C):
method: DOPPSOL-findif, time series 790710- 800513

v (mm d-1):
date =0
(in January!):
5.4
7.1
5.1
3.0
6.5
6.4
method: FOUSOL 790521- 800422

*107 (m2 s-1):
distance from ditch (m):
15
8.7
8.3
8.8
9.3
8.7
9.1
9.5
9.4
9.2
9.7
9.8
8.2
8.4
9.3
9.6
9.9
10.4
9.8
9.7
9.5
9.3
9.6
9.6
9.7
Tm (0C):
8.7
8.7
9.7
10.4
9.3
9.8
9.6
9.4
9.2
9.7
9.9
8.1
8.4
9.4
9.6
9.9
10.3
9.7
9.6
9.4
9.2
9.5
9.6
9.8
The table should suggest schematic maps of De Stobbenribben, where drawn lines represent the baulks. Note that the
DOPPSOL-findif method was applied to a slightly different time series. A further explanation is given in the text. Where
a credible value was found for with FOUSOL, the values obtained for v and are printed in bold face. The DOPPSOL
method was applied with a fixed value
= 1.35 10-7 m2 s-1.
assumptions, including any disturbing factors may become of greater importance than small
errors in precision.
The main problems are those caused by a lateral flow of water and heat (see Section 8.4),
and those caused by the uncertainty with respect to the depth z. Unlike DOPPSOL, FOUSOL
out of necessity uses numerical data collected in all seasons. Since the kragge rises and sinks
with the water level, the question arises what reference should be used for the depth z = 0. The
fixed null-surface is a model construct that, by visual inspection in the terrain, can only be
approximately located. In this investigation, the average kragge surface at the time of
146
Peat temperature
Table 8.7
petgat:
D
C
FOUSOL results based on monthly temperature profiles from May, 1979 to April,
1980.
B
Tm (0C):
stand. dev. Tm (K):
9.5
9.4
9.2
9.7
9.8
8.2
8.4
9.3
9.6
9.9
10.4
9.8
9.7
9.5
9.3
9.6
9.6
9.7
A0 (K):
0.3
0.1
0.1
0.2
0.2
0.1
0.2
0.1
0.1
0.1
0.2
0.1
8.5
7.6
9.2
10.0
7.5
7.3
8.9
9.5
8.9
9.4
9.5
date (
17
10
16
20
12
8
15
17
18
16
20
8.7
8.3
8.8
9.3
8.7
9.1
= 0 ) in February:
11
15
15
16
16
15
0.2
0.2
0.4
0.3
0.4
0.1
2.1
6.3
3.3
2.1
4.1
6.4
v (mm d-1):
*107 (m2 s-1):
a (m-1):
7.7
1.43 1.87
1.32
1.68 1.34
1.42 1.55
1.31
1.78
1.54 2.26
1.37
1.60
1.42
1.94
1.75
1.89
1.65
10.6
2.8
2.7
-1.1
-0.9
4.4
4.5
3.4
1.3
0.7
1.6
0.9
r2 for a (%):
0.45
0.64
0.75
0.81
0.87
0.67
0.64
0.72
0.71
0.78
0.62
0.81
0.71
0.59
0.62
0.68
0.60
0.56
b (m-1):
99.8 99.2
99.7
98.7 99.7
99.5 99.2
99.4
99.7
99.1 97.7
99.5
99.5
99.5
99.6
99.0
98.9
99.5
r2 for b (%):
0.78
0.73
0.86
0.77
0.84
0.85
0.79
0.86
0.75
0.80
0.66
0.85
0.79
0.82
0.71
0.75
0.72
0.76
100
99.8
99.9
99.8 99.8
99.2 99.8
99.9
99.9
99.6 98.9
99.8
99.8
99.9
99.7
99.6
99.8
99.9
No separate layers were distinguished (n=16), and a heat capacity of 4 106 J m-3 K-1 was assumed for the peat. The
arrangement of the table is as in Table 8.6. Numbers in bold face refer to very credible solutions for .
measurement was referred to, and the sounding rod was inserted in the profile through a hole in
a 15 cm diameter disc, which was slightly pressed down onto the kragge.
Results obtained with the FOUSOL method
The complete results shown in Table 8.7 were already summarized in Table 8.6 for comparison
with the results obtained by the application of the Doppler analogy. The mean of the values for v
is 3.1 mm/d (all values used) or 3.9 mm/d (less credible values omitted), which is in the same
order of magnitude. The solution of shows that the model yields less credible data for 11 out
of the 18 sites. Subdivision of the soil profile into layers yields curves of v as depending on z
(Fig.8.5). The shape of these curves varies, but there is a slight general tendency of a decrease of
v above a depth of, approximately, 1.0 m. This may be due to the influence of
evapotranspiration causing an upward component in the flow of water through the root zone. As
will be shown in Chapter 9, compensation for water loss is supplied by a lateral water flow
underneath the kragge, so that the downward flux due to seepage is less important or even
absent in the uppermost part of the profile.
Peat temperature
147
Fig. 8.5
Variation of estimated v with depth

Each record refers to a soil layer with a thickness of 0.7 m. Layers overlap by 86%. Numbers indicate
measuring stations (see Table 8.6). Solutions satisfying = 1.35 10-7 m2 s-1 7.5% have been marked
with a dot.
Other explanations, such as the influence of several disturbing factors, cannot be excluded,
however. An illustration of FOUSOL output for station 7, with n=8, is provided by Table 8.8.
Each of the overlapping layers, i.e., each window, is 0.7 m thick here. A decreasing n results in a
more fluctuating pattern of v, as illustrated in Fig.8.6 by showing the result for n=3 together
with those for n=8 and n=16. When n=3, the thickness of each layer is 0.2 m.
148
Peat temperature
Table 8.8
FOUSOL solution for station nr.7 (De Stobbenribben), nine overlapping soil layers
of 0.7 m thickness being distinguished.
Layer
from (m)
to (m)
0.3
1.0
0.4
1.1
0.5
1.2
0.6
1.3
0.7
1.4
0.8
1.5
0.9
1.6
1.0
1.7
1.1
1.8
Tm
stand. dev.
(0C)
(K)
9.2
0.1
9.2
0.1
9.2
0.1
9.2
0.1
9.1
0.02
9.1
0.02
9.1
0.03
9.2
0.04
9.2
0.1
A0
a
r2
(K)
(m-1)
(%)
7.7
0.78
98.9
7.6
0.75
98.9
7.2
0.70
99.7
7.3
0.72
99.8
7.5
0.75
99.8
7.7
0.76
99.9
7.7
0.77
99.9
7.4
0.74
99.9
7.4
0.73
99.9
(Feb)
(m-1)
(%)
10
0.88
99.6
8
0.85
99.7
7
0.83
99.9
8
0.85
99.9
1.28
2.3
1.36
2.4
1.43
3.3
1.35
3.4
date
b
r2
0=0
*107
v
(m2 s-1)
(mm d-1)
10
10
9
8
8
0.87
0.87
0.87
0.86
0.85
<--------------------------100.0-------------------------->
1.30
2.9
1.29
2.6
1.31
2.3
1.35
2.8
1.36
2.9
Temperatures were measured at eight depths (n=8) in every layer.
Fig. 8.6
The effect of subdividing the soil into layers at station 7

The thickness of each layer (window) is 0.3, 0.7, and 1.5 m at n= 3, 8, and 16, respectively. Solutions
satisfying = 1.35 10-7 m2 s-1 7.5% have been marked with a dot.
Peat temperature
149
When comparing these results with computations on the basis of measurements of the
hydraulic head, one should be aware of the uncertainty with regard to the hydraulic conductivity,
which is probably not particularly smaller than that with regard to the thermal diffusivity. In fact,
the results obtained here may be considered quite reliable in view of the values found for .
Note, however, that the areal extent to which the thermal method applies is small in comparison
to that in a hydraulic survey. An interesting result is the upward flow which was found at the
stations 16 and, quite reliably, 17. It appears that a mineral bottom is present here at a level 0.5
m higher than elsewhere underneath De Stobbenribben (see Fig.7.3). This elevation of the
mineral bottom was also traced under the nearby ditch situated to the north-east, and it is
supposed that leakage of water from that ditch is possible either through or over this elevation.
The electrical conductivity measurements presented in Chapter 9 provide support for this
hypothesis. In any case, there is no reason to suppose that groundwater is upwelling here from a
greater depth.
8.4 Lateral heat flow: a disturbing factor
The vertical heat flow through the mire was used in Section 8.3 as a tracer for water flow. It was
assumed there that no appreciable lateral heat flow was present. This assumption is not justified
by the following quantitative approximation of the importance of lateral heat flow in De
Stobbenribben. This approximation is only a rough one.
The water budget of one petgat in De Stobbenribben can be written as:
P E + I S = 0,
where
P precipitation gain in m3;
E evapotranspiration loss in m3;
I surface water influx in m3;
S seepage loss in m3,
the volume of water in the petgat supposed to remain constant.
P, E, and S, are inputs and outputs through the surface and the bottom of the petgat,
respectively, with an approximate area of 30 (width) times 200 (length) m2. The flux density of
S was estimated at 3 (-5) mm/d in Section 8.3, while the flux density of P-E may reach values of
-2 mm/d (in the extremely dry year 1976 even -3 mm/d) during prolonged dry periods (see
Appendix E on evapotranspiration). Over a period of 100 days, (P-E-S) may thus amount to -0.5
times 30 times 200 m3.
I is an input through one narrow side of the petgat. Although the depth of the petgat to the
mineral bottom is about 2-3 m, I may be channelled through a somewhat thinner layer with an
area of, say, 30 (width) times 1 (height) m2. The flux density of I during this period of 100 days
would thus be 1 m/d. This is probably an over-estimation, since a considerable drop of the water
table in the mire can be observed during such dry periods. The flux density moreover decreases
linearly with the distance from the ditch. On the other hand, the downward seepage may be up to
twice as large, and the main lateral flow channel may be still narrower and thinner.
150
Peat temperature
From the value 1 m/d near the ditch, occurring during a dry period, a reasonable estimate of
the mean lateral flux density within the preferential channel in the whole petgat is 0.5 m/d. This
value can be used in connexion with formulas 8.3 and 8.4 in Section 8.2 to resolve the amplitude
of temperature fluctuations which may be ascribed to lateral heat flow at a distance L from the
ditch. For this purpose formula 8.3 is applied along the longitudinal, rather than the vertical axis,
and z and v are replaced by L and the lateral water velocity vl, respectively. A0 now represents
the amplitude of the temperature variation in the ditch. Since the drought influx has a short
duration, the overall contribution to the temperature fluctuation in the mire will probably show
up as one with a period considerably shorter than 1 year, say = 91 d. A value of 5 K will be
used for the amplitude of the temperature fluctuations in the ditch during this period. The same
values as before will be used for (1.35 10-7 m2 s-1) and c (4 106 J m-3 K-1).
From formula 8.3 it can be seen that
L = -ln(AL/A0) /a,
where L is the distance from the source of fluctuation in the direction of heat flow. In order to
find a reasonably critical value for AL, the variation of Az, for z=1.0 m, with the flux density of
downward seepage, according to the vertical application of the Stallman model, is investigated
first. From Fig.8.1 it is apparent that a is almost linear in v when v is only some mm/d; the
decrease of a is roughly 0.04 m-1 when v increases by 1 mm/d. For z =1.0 m, Az=3.81 K when
A0= 9 K and v = 0 mm/d. Accordingly, when v increases by 1 mm/d, Az increases by
approximately 0.16 K. Taking this as a critical value AL= 0.16 K in the lateral application of the
model, the values listed in Table 8.9 are found for the distance from the ditch where the
amplitude remains this large.
This means that, if the assumptions of the Stallman model apply, the measured temperature
may be more than 0.16 K in error at the critical depth of 1.0 m almost everywhere in the petgat.
This would cause problems when it happens to coincide with other deviating values so as to
suggest a different annual amplitude in the Fourier analysis. I have not numerically investigated
the chance that this occurs.
The previously used Doppler analogy is much more optimistic at such large values of vl.
According to formula 8.2 in Section 8.2, with b= 1.72 m-1 ( = 91 d!) and vl = 1.0 m/d, b'= 0.066
m-1, so AL= 0.16 K at a distance L= 52 m from the ditch.
These calculations are, needless to say, only tentative, but they show that there is a fair
probability that lateral flow of water and heat may seriously interfere with the use of the
DOPPSOL and the FOUSOL methods. The FOUSOL method is less sensitive to such
interferences since it isolates the annual wave component. The distance from the ditch at which
interference may be disregarded varies according to the assumptions made. In an optimistic
scenario this distance is some decametres, in a pessimistic one it reaches up to the dead ends of
the petgaten, where vl approaches 0. This point will be reviewed in a discussion of observed
temperature gradients along the length axis of the petgaten in Chapter 9 (Section 9.4).
Apparently the qualitative conclusion that no groundwater is discharged into De Stobbenribben
is not affected by this problem. However, if groundwater were welling up, this would be
discharged into the ditch, at least in winter, and the temperature profiles would reflect the
upward flux, at least in the dead ends of each petgat.
From Tables 8.6 and 8.7 it appears that the FOUSOL method resulted in anomalous values
for the thermal diffusivity of the peat at most sites in the terrain parts within ca 80 m from the
ditch.
Peat temperature
151
Table 8.9
Critical distances from a source of lateral heat flow for some values of vl (with
=91 d) according to the Stallman model.
vl
Fig. 8.7
m/d
m-1
1.0
0.5
0.2
0.1
6.4 10-5
5.1 10-4
8.0 10-3
6.2 10-2
5.4 104
6.7 103
4.3 102
5.6 101
Isotherms in mire profiles during the year

a: Computed with Formula 8.3 (Tm=10 0C, A0=8 K, =1.35 10-7 m2 s-1, c=4 106 J m-3 K-1,
February, 15th)
b: Measured at stations 11-14 (parcel C, De Stobbenribben; see Table 8.6).
Station 14 is near the ditch.
152
0=0
on
Peat temperature
Table 8.10
The period in 1979 when the temperature at 0.3 m below the kragge surface
exceeded 14 0C.
Parcel D
from
to
days
Parcel C
from
to
Jun, 6 Sep, 26
112 Jun, 11 Oct, 1
Parcel B
days from
to
days
112 Jul, 21 Sep, 10
51
Jun, 7
Jun, 10
Oct, 2
Sep, 30
117
112
Aug, 1
Jul, 27
Aug, 26
Sep, 3
25
38
Jun, 9
Jun, 5
Sep, 14
Sep, 22
97
115
Jun, 14
Sep, 24
92
May, 14
Oct, 4
143
Jun, 7
Sep, 13
104
Parcel A
from
to
days
Jun, 3 Oct, 5 124

Jun, 1
Jun, 3
Jun, 1
May, 25
Jun,71
Sep, 30
Sep, 29
Oct, 3
Sep, 29
Oct, 3
121
118
124
127
118
The measuring stations have been arranged as in Tables 8.6 and 8.7; see also Fig.8.7b
In conclusion it is observed that the uncertainty with respect to the quantitative value of the
flux density increases as the flux density along the vertical axis increases, and so causes an
important lateral inflow.
8.5 The temperature regime in the root zone
The formulae presented in this chapter also enable a calculation of the temperature in the root
zone and the influence of seepage on this ecologically relevant factor. The available data can be
used to describe the actual temperature regime in the kraggen of De Stobbenribben. This has
been illustrated in Fig.8.7. The theoretical approach (Fig.8.7a) reveals that a 1 mm/d downward
seepage may prolong the duration of the period for which the soil temperature at a depth of ca
0.3 m exceeds a certain biologically critical temperature, e.g., 14 0C, with ca 3 days. At v=0
mm/d this period lasts 101 days according to the Stallman model with Tm=10 0C and A0=8 K. It
appears (Fig.8.7b, Table 8.10) that the duration of this period varies to an even greater degree
among sites in De Stobbenribben. According to these differences, the apparent start and length
of the growing season will also vary. While this may in itself result in different plant
communities, it also controls microbial activity governing the nutrient state. The possible
consequences in De Stobbenribben have not been investigated, however.
It is worthy of note that the results shown in Fig.8.7b and Table 8.10 are not only, and
probably even not primarily, due to hydrological differences between sites. The temperature
regime at the soil surface is influenced by various other factors, which will be partially
mentioned in Chapter 9. An obvious anomaly is shown in Fig.8.7b for the measuring stations 13
an 14 in the period November 1979 - January 1980, when a thermal discontinuity has developed
at the lower boundary of the kragge. Apparently the local unmown stand of Cladium vegetation
strongly delays the cooling of the kragge, while the body of water underneath is cooling through
a lateral influx of colder water. The short duration of the warm season at these stations is also
largely due to the combined insulation by the unmown vegetation and a relatively strong mixing
of water underneath the kragge. In quagfen parcel B (Table 8.10), the measuring stations far
from the ditch also exhibit a relatively short growing season. This is caused by a vegetation
cover with much Polytrichum and Erica. Although this vegetation is low, it forms a relatively
dry and thermally insulating layer.
Peat temperature
153
Even more local factors become important in the upper 1-3 dm of the kragge, where
short-term fluctuations of the air temperature and radiative exchange of heat substantially
contribute to the temperature regime. In De Stobbenribben and similar quagfens the local
variation in such factors are no-doubt decisive for the daily temperature regime in shallow pools.
8.6 Conclusions
The conclusions from the material treated in this chapter can be divided over two main groups:
those with respect to the thermal assessment of seepage, and those with respect to seepage in De
Stobbenribben.
The following conclusions belong to the first group:
1 The thermal assessment of seepage in wet peat soils requires accurate measurements of soil
temperature below a depth of 0.4 m;
2 The Doppler analogy method may be applied to single temperature profiles. It is suggested
that these profiles comprise at least 3 values, and that the depths be chosen so that these
values are definitely different. This method can only be used when v is small and when the
thermal diffusivity of the porous medium is nearly equal to that of water, as in the case of
very wet peat soils. A further analysis of the reliability of this method is desirable;
3 The Stallman method is based on a physically more precise model, but the method is also
more demanding. It is suggested that measurements be performed at least monthly4. The
concept of the model is most efficiently used when each temperature profile comprises
several values, e.g, 0.1 m apart. The sensitivity to violation of model assumptions deserves
more attention (see also 4);
4 It is necessary to consider the possible influence of lateral flow whenever the results have to
be quantitatively interpreted;
5 In De Stobbenribben both methods yielded similar results which could not be rejected by
reference to results of any other method and which are in accordance with hydro-geological
considerations.
The following conclusions refer to De Stobbenribben:
6 A downward seepage in the order of magnitude of some mm/d could be confirmed. The
spatial variety of the flux density appears to be considerable, however;
7 In some places about a 1 mm/d upward seepage was found, which could be in the range of
reliably determined values. There is some corroborative evidence for these anomalous results
from other studies. This seepage originates probably from a nearby ditch at a higher level;
8 The Stallman model assumptions are not satisfied always and everywhere in De
Stobbenribben;
9 At different depths in the profile the flux density of seepage may have different values.
There is some indication that this coincides with water loss through the root zone to the
atmosphere;
10 The measured data and the heat transport model enable a description of the temperature
fluctuations in the root zone. Seepage may alter the annual temperature regime to an
ecologically significant degree, but the measurements suggest that the structural aspects of
the local stands of vegetation are even more important in this respect.
4
Single gradients can be solved with the same formulae used in the Doppler method, however. In this case a range of
scenarios for v, with known thermal properties providing values for a and b, can be tested.
154
Peat temperature
CHAPTER 9
Lateral water flow in longitudinal transects
9.1 Introduction
The pattern in the cover of vegetation of De Stobbenribben shows an obvious
gradient in the longitudinal direction. Since there appears to be a downward flow
of water (Chapter 8) this could be caused by a compensating influx of surface
water from the ditch at one side, while, towards the other side, an increasing
influence of rainwater is indicated by the local stand of vegetation (Chapter 7). It
will be shown that the hydraulic head of the phreatic water in the quagfen parcels,
during drought, decreases as the distance from the ditch increases. Only heavy
rainfall is capable to invert this pattern temporarily. A descriptive approach
suitable for routine surveys was developed and used to monitor the presence of
chemically different types of water in longitudinal sections through the petgaten.
The survey method includes a detailed study of the temperature and the electrical
conductivity of the peat, and, in addition, chemical analysis of the mire water.
This chapter shortly describes the method and then summarizes the results
obtained in De Stobbenribben. Only a small, but representative, selection of the
various maps is shown here in order to deliver a de facto proof of the existing
longitudinal gradient and the influx of ditch water through a preferent flow channel
Lateral water flow
155
just underneath the kragge. The chemical identity of the mire water could be
assessed on the basis of chemical analyses treated here and in Chapter 10.
In the long run the water level in De Stobbenribben remains nearly constant. Since
the downward loss of water, in the order of magnitude of 3 mm/d, or 1 m/a, is far
greater than the estimated precipitation surplus (0.2 m/a), there must be a
considerable lateral influx of water. The topography of the area suggests that the
influx originates mainly from the ditch at the north-eastern narrow ends of the
petgaten. Since in this area rainwater and surface water have quite different solute
concentrations (Appendix D), conceivably the different types of water may be
traced in the mire by their electrical conductivities. The influx of ditch water is
controlled by a hydraulic head gradient in the quagfen parcels and it has appeared
that this gradient is large enough to be measured.
It is worthy of note that conductivity as such is not a measure of the concentrations of nutrients in the water. I will therefore not denote high-conductivity and
low-conductivity waters as nutrient-rich and nutrient-poor, respectively. A high
conductivity indicates a relatively concentrated solution and a high base state; in
some geochemical literature it is referred to as "strongly mineralized", suggesting
the solution of minerals from the lithosphere. This term is not very adequate in the
present context and I have not used it. A low conductivity indicates dilute and
base-poor waters.
In order to compare the conductivity values accurately account was taken of
the ambient temperature. The availability of temperature data also allowed for an
assessment of longitudinal temperature gradients which shed some light on the
problem of lateral heat flow interfering with the computation of vertical water flow
by the heat transport method (Chapter 8).
Water samples with a similar electrical conductivity may still be different with
regard to the concentrations of the various solutes. Chemical analysis of water
samples were used to confirm the identification of coherent bodies of mire water.
9.2 Data acquisition in longitudinal transects
Hydraulic head (water manometers)
Hydraulic head differences at first appeared difficult to assess due to the small differences and
various problems associated with the measurement in piezometers in quagfens. Firstly, it is no
trivial matter to fix the usual pvc pipes in the mineral subsoil, so as to guarantee a fixed
reference height which is independent of the vertical movement of the kragge. Of course the
lower part of such pipes should be separated from the upper part containing a filter in the body
of phreatic water above the mineral soil. Secondly, the weight of an observer on the kragge
exerts a local pressure possibly altering the water table in the gauge. Thirdly, the installation of
many semi-permanent observation gauges was inacceptable from the point of view of nature
management: it attracts occasional visitors, damaging the vegetation (and frequently also the
installed water gauges!), and it hampers the mowing and cutting of the vegetation. A classical
water manometer, used by J.Bon in studies of the hydraulic gradient in rivulets (pers.comm.)
was adapted for the use in quagfens and appeared to provide a suitable solution (Fig.9.1).A
comparison of measurements of the hydraulic head gradient with different methods has shown
that the water manometer method is superior to piezometer methods.
156
Lateral water flow
Fig.9.1
Water-manometer for the measurement of water-level differences in quagfens
Since both lengths of flexible tubing in this water manometer are about 20 m long, it is
possible to obtain the readings at a considerable distance from the measuring points (pools or
open pits in the kragge). When possible, the readings were performed on the baulks between the
petgaten. A series of measurements can be executed in a closed loop, beginning and ending with
the same water level in part of the boezem system. The highest accuracy, which was not reached
at all occasions, was about 0.5 cm over a distance of 30 m, or ca 0.00017%.
Kragge movement
In order to check for a possible variation of the water level in the root zone, several experiments
were performed to measure the vertical movement of the kragge. This was done with small tins,
firmly attached in the kragge and moving around an iron post driven into the mineral bottom
(Fig.9.2). The results for De Stobbenribben have been reported by Touber (1973). It appeared
that the movement of most parts of the kragge is about 30-90% of that of the phreatic water
level. The kraggen are attached to standing baulks, and near these baulks their movement is
strongly reduced. A similar reduction is seen in certain firmer parts of the kragge. The
movement of the floating mat is also reduced at extreme water levels, which may result in a
local flooding. At the level of accuracy of the present treatment and for the periods chosen for
water balance calculations the varying storage of water in the kragge itself could be disregarded.
Lateral water flow
157
Fig.9.2
Gauge for measuring kragge movement
Conductivity and temperature sounding

The electrical conductivity probes were mounted in the same sounding rods as used for the
assessment of temperature profiles. The earliest models had two platinum rings as electrodes,
and were connected with conventional conductivity meters. The platinum rings were later
replaced by stainless steel ones, of which several sets were mounted in a single sounding rod, so
as to enable the simultaneous assessment of conductivity profiles. The temperature was noted
separately, and the conductivity at 10 C was calculated afterwards according to the formula
presented in Appendix D. The code EC10 is used. Since the measurement of the electrical
conductivity hardly requires any equilibration, it was, without any appreciable time loss,
performed in one run together with the temperature measurements discussed in Chapter 8. The
instruments were frequently re-calibrated in the laboratory. From time to time a comparison was
made between the in situ measurements with the sounding rods, and measurements with a
standard conductivity meter in water samples from the same place and depth. These samples
were collected according to the method described in Section 9.5. As a consequence, these
measurements do not refer to exactly the same bodies of water. Moreover, the sounding method
is influenced by the presence of the peat matrix. The results should, therefore, not be regarded as
straightforward conductivities of water. Since the deviation reflects a soil factor which is more
or less the same everywhere in De Stobbenribben, it was decided to use the results as relative
numbers, showing the conductivity patterns in the sections, rather than trying to convert the
numerical data into an absolute conductivity of water. The difference is not large in this
extremely wet substance, but it has appeared that differences are flattened out somewhat.
Thus, the numerical data presented here reflect the electrical conductance of the medium in
mS measured with the specific sort of cell used. They may be interpreted as electrical
158
Lateral water flow
conductivity values, which should be multiplied by a variable c close to 1 in order to get

electrical conductivity of water at 10 C in mS/m.
Measuring schemes and data processing
Occasional observations in 1969-'70 suggested a clear longitudinal gradient in the vegetation, in
the electrical conductivity of the water at various depths, and in the chemical composition of this
water. Periods of more systematic measurements followed in 1973-'75 and 1978-'83. In 1978 the
program was reduced to a sounding at fixed measuring stations and the assessment of water
level differences between the closed ends of each quagfen parcel and the boezem system. A
quantitative description of the relation between the hydraulic head and conductivity gradients,
which would have required the continuation of time consuming levellings, was not aimed at.
Representative data were selected on a basis of comparison of conductivity patterns as seen
in hand-drawn isopleth maps. As the amount of collected data increased it became desirable to
develop a database on a computer and to program the production of isopleth diagrams. This was
done by C.H.van Leeuwen for the computers and plotters used at RIN in those days. The
hundreds of diagrams produced have been filed in databooks at RIN.
It appeared that the construction of isopleths on the basis of the calibrated, but further
unprocessed numerical data yields rather irregular patterns as a result of very local deviations of
the conductivity. This problem was reduced by averaging three values in a moving window from
the upper part of the soil down to 1.8 m. In one profile, each value is thus replaced by the mean
of the three values provided by itself and by the 0.1 m upper and lower neighbouring ones,
respectively. This variable is shortly called EC10(3) here. The intersection with each isopleth is
then found by linear interpolation. Similarly, linear interpolation along the length axis through
each measuring depth is applied, and the various points of each isopleth are connected by
smooth lines. The isopleth maps were visually compared. Their general pattern is discussed in
Section 9.2.4 with some representative diagrams, which were redrawn in block schemes.
9.3 The general pattern found
The hydraulic head gradient
Some typical results of the water levelling program are shown in Fig.9.3. They show the
development of a longitudinal gradient in all four petgaten during summer. This gradient is less
pronounced during winter, but inversions, i.e., a hydraulic head surplus in the body of phreatic
mire water in the quagfens, are only observed during and after heavy rainfall. Such inversions
rarely exist longer than two or three days. A very interesting feature is the difference among the
four quagfen parcels: as the drought continues, parcel D, which is closest to the drained polder
area to the south-east, presents the lowest phreatic levels. In time this may vary, however,
according to various local circumstances. In various summer seasons the gradient from the ditch
towards the closed end of the petgaten reached a value of 13-21.8 cm over 160-200 m. In the
transverse direction almost ten times steeper gradients are found near the canal running along the
outer baulk of parcel D, causing small local seeps as suggested by the thermal measurements
presented in Chapter 8 and also by the conductivity measurements.
Lateral water flow
159
Fig.9.3
Hydraulic head gradients measured in De Stobbenribben, parcels A, B, D

from Vromen, Klamer & de Vries 1974
The longitudinal conductivity gradient

A longitudinal gradient of the electrical conductivity was apparent in all four petgaten (Fig.9.4).
Near the dead ends of the petgaten the conductivity is low, while it is higher towards the ditch.
In petgat B the higher values are confined to a narrow zone near the ditch, but they reach much
farther towards the dead ends in the other ones. The mean values for the period 1979-1980 show
a pattern very similar to the one of May, 1979. Next to the longitudinal gradient there is also
some difference along the vertical axis. Especially near the surface of the mire, the values tend
to be lower. On average the conductivity values have decreased somewhat between 1973 and
1984.
The general pattern may be explained by:
1) An influence of precipitation in the uppermost part of the mire profile;
2) An influence of ditch water penetrating to just below the kragge over a length which
depends on the flux density of the downward seepage and on the lateral hydraulic
conductivity of the mire.
Petgat B has a thicker and firmer kragge than the other ones, as is easily observed by the way it
yields underfoot. Also the vegetational succession appears to reflect a later stage of
terrestrialization. This petgat is somewhat narrower than the other ones, which might have
caused an earlier overgrowth. It cannot be excluded that it was even dredged earlier, since there
is a trend of early dredgings being narrower and shallower than later ones.
Many isopleth maps show "fingers" which extend from a body of water with a high electrical
conductivity into one with a lower conductivity, or the reverse. The higher conductivity fingers
are typically confined to a layer between, very roughly, 0.7 and 1.2-1.6 m, which is just
underneath the kragge. This layer is the most transmissive part of the profile, the preferent flow
channel for the inflow of ditch water.
The seasonal movement of bodies of groundwater
The successive isopleth diagrams suggest a seasonal movement of the transitional zone between
a body of high-conductivity water, originating from the ditch, and a body of low-conductivity
160
Lateral water flow
Fig.9.4
Fluctuations of the electrical conductivity in sections A-D through De Stobbenribben, viewed from the ditch towards the closed ends
Lateral water flow
161
Fig.9.5
The main flow patterns of inflowing ditch water (I) and precipitation surplus (P) in
De Stobbenribben
water originating from local precipitation (Fig.9.4). The measuring scheme and the interpolation
method used in the construction of isopleth diagrams obscure any sharp boundary between the
bodies of water along the length axis. Occasional checks suggested that the transitional zone is
indeed rather wide, however, and that both types of water become mixed in this zone. This can
be understood from the fact that the net position of this zone results from dispersion occurring
during the displacement of the transition zone.
It appears that the closed ends of the petgaten are nearly always under the influence of the
body of low-conductivity water, while the zones bordering the ditch are mostly in the reach of
more concentrated water. During the dry season a tongue of water with a high conductivity
penetrates the transitional area in a zone just underneath the kragge. During the cold season it is
expelled again by more dilute water. The expansion of the latter exhibits a remarkable pattern. It
is first observed in the superficial part of the kragge, where it indicates an increasing
precipitation surplus. This is for a short time followed by a finger which extends underneath the
tongue of high-conductivity water still present, as schematically indicated in Fig.9.5. It can also
be seen in the isopleth diagrams of Fig.9.4. This pattern suggests the peaty matrix outside the
most fluid part of the profile as the preferred medium for the expansion of this body of water.
There are several weaker spots in the kragge, some of which may be remains of the open-water
stage. Others originate from the uprooting of young trees and shrubs to favour reed cultivation
and hay-making. Such weak spots have formerly been taken as evidence for small groundwater
wells; they are often filled with a high-conductivity water amongst a kragge containing a type of
water poorer in solutes.
As a result of the rather irregular form of the isopleths, it was not so easy to quantify the
velocity of the displacements. From May or June to October there is a net inflow of ditch water,
with a conductivity of over 40 mS/m. The tips of the fingers of the bodies of water enclosed by
representative isopleths travel about 60 m forward from their original position in some 120 days.
162
Lateral water flow
During winter, from October to April, the 25 mS/m isopleth may move about 100 m in the
direction of the ditch. In the course of years, however, the movement seems to depend on many
singular circumstances, such as ditch dammings, cloggings, and cleanings, and infiltration works
in neighbouring reedbeds. In the period 1973-'82 the fluctuations due to these causes and to the
extremely dry summers of 1975 and 1976 appeared to be of greater importance than any other
trends.
A quantitative interpretation of the isopleth patterns is further interfered with by the vertical
movement of the kragge, thus changing the thickness and volume of the most transmissive layer
just underneath it. Although the porosity of this layer is generally assumed 100 %, the actual
situation is probably one of a subsurface channel with irregular lumps of less transmissive peat
and root masses. The effective porosity might therefore be well below 80 or even 50 %, so that
the velocity which is associated with the displacement of electrical conductivity isopleths could
be much higher than the Darcy velocity. Locally, displacements of even 2-20 m/d along the
length axis, and of 5 cm/d along the depth axis, were quite clearly noted in isopleth diagrams
with an interval of about one week. It should be kept in mind that the movement of electrical
conductivity isopleths results from a combination of the movement of whole bodies of water,
and such processes as diffusion, mixing, and concentration by evaporation. Local and short-term
disturbances may also originate from ephemeral pressures exerted on the kragge, such as when
people or machines go over it, as happens especially during mowing and hay-making.
The conductivity map
The presence of baulks at relatively short distances improves the suitability of De Stobbenribben
for such investigations as reported here, but the same factor introduces several influences in
gradients perpendicular to the length axis. The longitudinal transects discussed above were
chosen just in between and parallel to the baulks. Occasional measurements in perpendicular
transects proved that the inflow of ditch water underneath the kragge is strongest in the middle
of the petgaten. Along its edges the kragge is fixed to the baulks and prevented from rising with
the water level. Here occasional floodings may penetrate further across the mire surface. The
same marginal areas convey a part of the precipitation surplus from the kragge towards the ditch
as surface flow.
Unfortunately, conductivity mapping to visualize these points was not systematically applied
in De Stobbenribben. However, sounding data obtained in 1973, 1979, and 1980 enabled the
reconstruction of an approximate conductivity map for 1979-'80 (Fig.9.6). Reference was also
made to soundings reported by Boeye (1983) and Kooijman (1985). A striking difference
appears to exist between various places as regards the range of fluctuations. This range is
especially large in the middle part of the individual quagfen parcels.
One deviation of the general pattern became obvious in petgat D. Along the southeastern
baulk, at the left in the map, the conductivity was mostly intermediate to low throughout the
profile from the dead end to about halfway the distance to the ditch. Further towards the ditch,
however, a high conductivity was found close to the baulk, where a very weak part of the kragge
suggests a direct connexion with the ditch. Several longitudinal sections, such as the May and
September ones in Fig.9.4, show a high-conductivity plume near the end of this supposed
connexion, which itself lies outside the plane of these sections. This plume is fairly well seen in
the isopleth diagram from October 8, 1975. In that very dry period there was no direct access of
ditch water along the supposed connexion. From the conductivity map it appears that the
high-conductivity area in the kragge extends far beyond the underlying body of highLateral water flow
163
Fig.9.6
164
The average distribution of conductivity values in and underneath the kragge in De

Stobbenribben, ca 1980
Lateral water flow
conductivity water. The pattern suggests an influx of "ditch-type" water near the baulk, about 75
m from the ditch.
The same part of the terrain frequently yielded somewhat anomalous temperature soundings,
which even led to the solution of an upward flow velocity in Chapter 8. Further investigations
revealed that a sand ridge runs underneath De Stobbenribben in this place (Fig.7.3). These
observations might be explained as indications of some seepage through or over this sand ridge
originating from a ditch a few meters to the south-east, parallel to the length axis of the petgaten.
At this place some water samples from a depth up to 1.6 m in the baulk were drawn and
analysed in 1973. They showed an electrical conductivity of ca. 50 mS/m, which I am now
inclined to interpret as an indication of ditch water seeping through the baulk. Since the baulks
exceed the kraggen in height by about 0.5 m, this level corresponds to a depth of 1.1 m with
respect to the top of the kragge. This seepage may be caused by the transverse hydraulic
gradient already mentioned. Incidentally, the plume area in petgat D was recognized as an area
of seepage pools by botanists already before 1960, and the permanent quadrats mentioned in
Chapter 7 were situated in it. Seeping parts of baulks were frequently observed elsewhere in De
Weerribben during conductivity mapping.
It is concluded that a clear longitudinal gradient in the electrical conductivity is developed in the
quagfen parcels due to the influx of ditch water. This influx is mainly channelled through a very
transmissive layer in the profile just underneath the kragge. A combination of processes, rather
than just the mass movement of whole bodies of water, determines the chemical composition of
the mire water, and this results in a considerable variation of the amplitude of fluctuations in the
water quality.
9.4 Temperature gradients in longitudinal sections
Isopleth patterns
The isopleth patterns of the annual mean and the amplitude of the temperature, derived by
Fourier analysis of monthly profile measurements during May, 1979 through April, 1980, are
shown in Fig.9.7 and 9.8. These maps typically show a gradient along the length axis, which is
different in each petgat. Similar patterns were found in isotherm maps. It appears, from these
patterns, that the annual mean temperature increases slightly in the direction of the ditch and in
the upper part of the profile.
Several other intriguing patterns can be noted. The mean temperature is considerably lower
than elsewhere in the middle of petgat C. In A and C it increases somewhat towards the dead
ends, especially underneath the kragge.
The temperature amplitude isopleths strongly reflect the attenuation of the amplitude with
depth. In the petgaten A and C this attenuation decreases as the ditch is approached, but in C the
amplitude is small in the same part of the petgat where the mean temperature is low. The
attenuation of the amplitude in B increases towards both ends of the petgat. Parcel A shows a
gradual decrease from the closed end towards the ditch, as is apparent from the slope of the 5 K
isopleth.
The amplitude of the temperature fluctuation at the soil surface, as computed with the
FOUSOL method, is also given in the block diagram of Fig.9.8. The values show a variation
which is not easily interpretable. Very high values are not found near the dead ends of the
petgaten. The low values in the middle of C, and near the dead ends of B and of C, respectively,
Lateral water flow
165
Fig.9.7
The annual mean temperature according to Fourier analysis of frequent

measurements in longitudinal transects through De Stobbenribben (May 1979April 1980)
The ditch from which water penetrates into the parcels is at the front side of the block diagrams
Fig.9.8
The amplitude of the annual temperature fluctuations around the mean temperature
in longitudinal transect according to Fourier analysis as in Fig.9.7
The numbers written into the diagrams represent FOUSOL-computed amplitudes at the mire surface in
C
are somewhat unexpected, but they can be explained from the characteristic structure of the
vegetation (cfChapter 8). In the middle of parcel C an extensive and unmown stand of Cladium
mariscus is found, while the vegetation at the closed end of B is rich in the mosses Polytrichum
commune and P.juniperinum, and in Molinia caerulea and Erica tetralix. These types of
vegetation provide an insulating blanket over the soil.
166
Lateral water flow
The causes of the spatial patterns of temperature data

A longitudinal temperature gradient may be caused by the fact that the surface water in the
boezem area, and consequently in the ditch, has a different temperature as compared to the mire
water, owing to (1) a different radiation budget, and (2) the mixing of water in the ditch. The
annual fluctuation of the open water temperature is typically smaller than that at the mire
surface, but larger than the variation underneath the kragge. Since ditch water penetrates into the
petgaten during summer, when the ditch water is warmer than the mire water underneath the
kragge, the influx of ditch water will probably enlarge the amplitude of the temperature in the
mire profile. A similar effect, i.e., a deeper position of the isopleths, is to be expected for places
with a stronger downward water flow. It is doubtful whether these effects can be separated from
other influences in the amplitude diagrams, but in those of the mean temperature the influx of
ditch water is probably the main cause of the patterns seen in the petgaten A, B, and D. The
higher annual mean near the dead end of parcel A may have been caused by the incidental
working of a replica of an ancient small wind-pump (tjasker) placed there as a sight-seeing
object.
Next to these there are several other causes of variation. The main one is certainly the nature
of the surface, especially its vegetation structure and the height of the water table with respect to
the top of the kragge. The middle part of petgat C was overgrown with a dense cover of
Cladium mariscus during this investigation. This stand was never mown, and the water level
nearly always lay a few centimetres above the top of the kragge. These factors may modify the
heat budget considerably by reducing both radiative and advective heat transfers. In this place,
where the kragge is only weakly developed, the possible incidence of density currents cannot be
excluded, although this is not very probable (see below). There is also an obvious variation in
the overshading of various parts of the petgaten, and in the protection against the prevailing
winds.
Temperature recordings and freezing observations during the winter 1969-'70 yielded several
unexpected results, which could not be explained with hydrological parameters.
All in all the main conclusion from the isopleth patterns of temperature data is that there are
many factors that cause variation which may be both relevant and attributable to plant growth
and that the hydrological factors cannot always be readily singled out.
The possible incidence of density currents
The interpretation of isopleth patterns and the application of the thermal method for the
assessment of seepage heavily depend on the absence of density currents. Density currents, or
convective currents, may occur when the water in the superficial layers of the profile, while
cooling, acquires a higher density than the deeper water. Its starting depends on the temperature
gradient and on the hydraulic conductivity of the porous medium. Although an accurate
prediction seems a tall order, it is possible to determine under which conditions density currents
will not be generated.
According to the pertaining theory (Bear 1972, p.653-660), the generation of density currents as
a result of a temperature gradient in a medium with about 100 % porosity can be described by
Lapwood's non-dimensional convection parameter, which is a modified Rayleigh number:
Lateral water flow
167
Y = (T1-T0)
(z1-z0) K
-1
where the subscripts 0 and 1 indicate the upper and the lower boundary depths between which the
currents may develop; is the coefficient of thermal volume expansion of water (approximately
2 10-4 K-1), is the thermal diffusivity (approximately 1.35 10-7 m2 s-1), T is the temperature
( C, the difference between temperatures being expressed in K), z is depth (m), and K is the
(saturated) hydraulic conductivity of the porous medium (m s-1). The critical value of Lapwood's
convection parameter above which density currents may be generated is
Y = 4 2, or: Y 40.
Converting K to the more convenient units m/d, it follows that the hydraulic conductivity above
which density currents are started, obeys the relation:
Kmax = 2400 (T1-T0)-1 (z1-z0)-1.
Table 9.1 lists Kmax for a representative range of values of the other variates5.
Table 9.1
The hydraulic conductivity above which density currents may be generated at

typical temperature gradients
T1-T0
(K)
z1-z0
(m)
Kmax
(m/d)
2
4
8
1
1
1
1200
600
300
The hydraulic conductivity in the petgat profiles varies with depth and with the firmness and
thickness of the kragge. Although the electrical conductivity profiles prove that a quagfen as a
whole is an anisotropic medium, in which the hydraulic conductivity in the horizontal direction
by far eceeds that in the vertical direction, it is assumed that the medium in the preferential flow
channel, where the conductivity may approach the listed critical values, is approximately
isotropic. Boundary values are derived from the water balance and the hydraulic gradient in the
longitudinal direction, yielding ca 500 <K< ca 1000 (m/d) (see Chapter 10).
In conclusion I believe that the hydraulic conductivity of the profile only very locally, if at all,
reaches such high values as are needed for the generation of density currents. Apparently, the
conditions for the generation of density currents are usually not satisfied in De Stobbenribben,
except in some places with a very weak or even nearly absent kragge, such as occur in small
spots in the petgaten A and D, and in a larger area in C. The latter is characterized by a dense
vegetation cover of Cladium mariscus. Remarkable isothermal patterns were obtained here, but
a detailed check of all relevant temperature profiles did not provide support for the occurrence of
regular temperature inversions and density currents.
5
The symbol Kmax refers to the maximum of the range of values for which density currents are not expected
168
Lateral water flow
9.5 The chemical identity of different bodies of mire water

Bodies of mire water were distinguished in Section 9.3 by means of the electrical conductivity.
The electrical conductivity is a parameter of overall ionic concentration. It is neither sufficient to
asess the origin of the water, nor is it in itself an important growth factor, at least not in the range
of values observed in De Stobbenribben. More complete chemical analyses of water samples are
needed to assess the chemical identity of the bodies of mire water that were traced by the
electrical conductivity soundings. In this section I will treat the chemical identity of the mire
water on the basis of a selection of water samples analysed for this purpose.
Methods of sampling and analysis
Free surface water was sampled in ditches and pools by immersing the sampling bottle in the
body of water concerned. Permanent sampling sites were installed in order to sample the
uppermost layer of mire water in the kragge. The installation involved the digging or cutting of a
small pit, which was always emptied the day before sampling took place.
Water samples from below the kragge were drawn with samplers (Fig.9.9), consisting of an
inner and outer length of p.v.c. (polyvinylchloride) pipe. The inner length is provided with a
filter wrapped with a nylon filter cloth (during the first years of this study a cotton cheese-cloth
was used). The inner pipe is sealed underneath with a plug or cap that protrudes just enough to
seal the outer pipe (jacket) when that one is pushed down over the filter. The inner pipe diameter
is 3 cm. The samplers were installed the day before sampling by pushing them down through the
kragge. With the filter at the desired depth, the jacket was raised so as to open the filter. The
next day the filter was closed again and the whole sampler was carefully removed and tilted to
pour out the sample. By using these samplers no instruments were permanently left to the
curiosity and, possibly, disturbance by occasional visitors; the samplers could be thoroughly
cleaned after and before use, and a total of 40 samplers sufficed for the investigation of several
mire areas during the same period.
A first determination of pH and electrical conductivity was done before the samples were
bottled. Two methods of bottling were followed, depending on the requirements of the chemical
laboratory:
1) (Hugo de Vries-Laboratory of the University of Amsterdam) Two subsamples were made.
The first one was passed through a coarse filter into a 0.5 l glass bottle. The second one was
passed over a 0.45 m micropore filter with the help of a foot pump. This appeared to be a
very time-consuming job and it was therefore often put off till the same evening. This
subsample was poured into a 0.25 l glass bottle for the analysis of P and N. Occasionally, a
similarly treated 100 ml subsample was collected in a glass bottle with a drop of HCl for the
analysis of Fe. Clean medical glass bottles were provided by the laboratory.
2) (Laboratory of the water supply company Midden Nederland) A 1 l sample was put in a
clean p.v.c. bottle provided by the laboratory, after passing it through a coarse filter.
Whenever possible, some surplus of the sample was used to rinse the bottles, and the sample
was enclosed free of air. Samples were stored at 5 C in the dark. During sampling and
transportation on warm days a crate with a wet cloth was used to avoid the warming-up of
bottled samples. In some cases it was not possible to collect the prescribed quantity of sample.
Lateral water flow
169
Fig.9.9
Water sampler used to draw samples from underneath the kragge
This may have reduced the number and representation of subsamples for each analysis in the
laboratory.
In either laboratory the samples were allowed to settle and the supernatant was used in the
analyses. Alkalinity, electrical conductivity and pH were mostly determined within two days
after arrival of the samples. All analyses were performed according to laboratory standards
(N.E.N., the Dutch organization for standardization) by laboratory personel.
At least the concentration of the major ions (see Appendix D) was consistently determined in
the majority of the samples. A variety of N- and P-containing constituents was analysed in most
of them. Occasionally Si, Fe, Al, and Cu were determined.
Analyses used
Two groups of water samples were collected from De Stobbenribben in order to identify the
chemical constitution of the various types of water detected with the electrical conductivity
method.
During the first period of the present investigation, 1973-'74, 5 sampling sites were selected
in parcel A, covering the supposed gradient described with conductivity sounding. The same
sites were sampled on the dates 730724, 730920, 731125, and 740403, where dates are coded as
(19)xx, month number, day number. The analyses were performed by the Hugo de
Vries-Laboratory. Many of these analyses did not satisfactorily pass the MAION (Appendix D)
electroneutrality and conductivity tests. For this reason I have given priority, in this discussion,
to the analyses resulting from the second group. The main conclusions which have earlier been
derived from the interpretation of the 1973-'74 analyses (Bergmans 1975, Van Wirdum 1982)
170
Lateral water flow
appeared to be valid on the basis of the more reliable analyses in the second group, and some of
the analytical results are used in Chapter 10.
From 1980 through 1983 a sampling scheme was followed in connexion with the
temperature and conductivity soundings in longitudinal transects. Samples were approximately
collected bimonthly, and a total of 22 seres of analyses covering four surface stations and three
deeper ones was made available. The samples were analysed in the laboratory of the water
supply company "Midden Nederland".
Method of interpretation
The water analyses were subjected to the calculations associated with the MAION method
described in Appendix D. The electroneutrality test was interpreted as follows:
If the difference between the sums of the priciple cations and anions was less than 8% of the
total concentration of these ions, on a moles-of-charge basis, the sample was accepted.
Otherwise account was taken of N, P, Fe and all additionally analysed components, and the
calculation repeated. The same criterion was applied again, but an additional requirement was
that the calculated conductivity must likewise not differ more than 8% from the measured
conductivity. The remaining analyses are rejected or at least not allowed to play a major role in
the interpretations. The experience with this method suggests that the additional elements do not
predominantly occur in an ionic form in the samples considered. As appears from Malmer
(1963) and Gorham et al. (1985) it is quite normal for interstitial water from peaty sites to show
electroneutrality errors if only inorganic ions are being considered.
Special attention will be given to the similarity of the chemical composition of each sample
to the groundwater (rLI), rain water (rAT) and sea water (rTH) benchmark samples, respectively
(see Appendix D). The results are presented in rLI-rTH and EC-IR diagrams, where
IR = [1/2Ca2+] / ([1/2Ca2+] + [Cl-]).
Special attention was paid to the carbonate equilibrium. The interpretation of this equilibrium is
not free of ambiguity, since it is uncertain which factors may be involved in the natural situation.
The calculations are based on a relatively simple model situation (Kelts & Hs 1978; see also
Appendix D). Moreover, artifacts may have been introduced due to a changing
physico-chemical equilibrium in the samples between the times of, (1), sampling, (2), analysis of
pH, conductivity, and alkalinity, and, (3), analysis of calcium, respectively. Over-saturation with
respect to CaCO3 was frequently met with in the ditch site a (see below). It is possible that some
of the elements involved were present in a non-ionic form in the sampled water, as is reported
by Kelts & Hs. The clouding phenomenon described there was frequently seen in surface
waters in North-West Overijssel. Various mechanisms may explain the formation of colloidal
forms of calcium, especially in the presence of phosphorus and humic substances, acting as
surface inhibitors by masking the charged particles. Since the ionic balance of most samples was
in order, there is no particular reason to reject the results.
1980-1983 analytical results
The results of the analysis of 149 water samples from June, 1980, to June, 1983, are summarized
in Tables 9.2 and 9.3. Site a is a shallow ditch. Sites b, c, and d are kragge sites where the
uppermost layer of mire water was sampled. At these sites water was also sampled underneath
Lateral water flow
171
Table 9.2
a
a
b
b
c
c
d
d
bb
bb
cc
cc
dd
dd
m
s
m
s
m
s
m
s
m
s
m
s
m
s
Mean and standard deviation of analytical results and computations concerning

water samples from De Stobbenribben
pH
Ca Mg
mg/l mg/l
Na
mg/l
21
21
20
20
21
21
21
21
22
22
22
22
22
22
7.60
0.24
7.14
0.58
6.51
0.66
5.22
0.75
6.67
0.17
6.34
0.25
5.80
0.18
70.23 10.17 41.38

17.58 0.96 9.74
44.67 6.91 30.40
11.79 2.01 7.89
20.46 3.89 17.52
9.68 1.65 5.83
5.05 1.80 8.12
3.00 1.06 3.75
64.80 9.63 39.27
7.73 1.19 5.42
49.79 7.56 32.95
8.06 1.29 4.42
16.59 4.11 18.02
2.74 0.79 4.09
K
mg/l
Cl
mg/l
4.70
1.64
4.07
2.17
4.06
3.50
4.08
2.39
5.85
6.34
4.67
0.82
2.50
0.74
76.39
22.25
57.35
17.00
32.84
11.24
14.43
6.71
76.01
13.94
62.01
10.62
24.97
7.24
HCO3
mg/l
185.24
52.17
117.25
37.92
51.62
29.43
11.01
15.73
201.58
33.01
154.17
29.92
68.77
12.30
SO4
mg/l
EC25
mS/m
IR
%
x
%
y
%
46.70
14.22
31.75
11.35
21.95
9.38
15.76
9.68
22.73
12.81
19.32
6.36
10.45
4.32
60.75
10.20
42.43
9.54
23.76
7.99
10.68
4.80
57.45
6.59
45.80
5.98
20.38
3.05
61.95
5.58
58.05
4.10
51.29
10.06
37.19
10.52
60.41
2.87
58.77
3.45
54.59
6.13
1.05
2.22
0.35
1.04
-0.38
0.71
-1.48
2.19
-0.23
0.53
-0.05
0.82
-0.95
0.95
sat
-7.76 7.63
0.23 10.83
-6.95 8.01
0.23 0.26
-5.14 8.78
2.25 0.62
-8.38 9.34
6.55 3.26
-7.14 7.60
0.12 3.39
-7.82 7.82
0.15 6.30
-8.18 8.59
1.31 0.15
rLI
%
rAT
%
rTH
%
rRH
%
Ca
%
Mg
%
Cl
%
80.90
0.89
76.35
8.38
51.38
29.60
-11.71
23.85
85.95
7.17
83.36
7.52
67.86
10.39
-23.00
8.89
-15.80
2.94
6.05
23.04
52.33
31.76
-33.64
5.16
-29.77
6.08
-20.50
3.49
49.76
7.38
29.75
13.13
1.43
11.83
-22.14
4.70
43.68
2.37
30.68
1.80
-7.55
1.54
79.57
6.90
70.70
9.64
51.05
12.45
19.14
12.23
75.27
1.05
68.55
1.29
37.86
8.08
55.38
2.65
52.55
2.72
45.00
9.35
27.52
9.41
55.00
3.45
53.09
4.70
41.09
3.39
13.86
5.28
13.45
1.23
14.38
2.22
15.95
4.77
13.59
34.43
13.36 37.50
17.00 34.00
1.69 6.63
a-dd sampling stations; m mean values; s standard deviations; n number of samples
Table 9.3
a
a
b
b
c
c
d
d
m
s
m
s
m
s
m
s
Total concentrations of inorganic N, P and Fe in De Stobbenribben water samples;

means and standard deviations as in Table 9.2
Ptot
Ntot
10-2 mg/l
Fetot
10-1 mg/l
21
21
20
20
21
21
21
21
3.29
3.91
1.85
3.57
5.90
13.58
6.62
13.68
12.76
11.38
2.00
2.15
6.57
15.79
11.76
14.84
81.90
49.18
16.55
12.39
24.95
25.81
21.19
11.50
bb
bb
cc
cc
dd
dd
m
s
m
s
m
s
22
22
22
22
22
22
Ptot
Ntot
10-2 mg/l
5.09
5.96
5.55
5.40
8.50
7.79
40.68
19.46
24.73
8.74
39.18
22.42
Fetot
10-1 mg/l
13.73
15.03
9.18
7.10
81.09
188.46
a-dd sampling stations; m mean values; s standard deviations; n number of samples
the kragge at a depth of 1.2 m. These deeper sampling sites have been coded bb, cc, and dd,
respectively.
The electroneutrality and conductivity tests (x and y in Table 9.3) were successfully passed
in 90% of the cases. The d and dd samples, 830208, show a considerable deviation in both x and
y. These deviations disappear when the concentration of Fe is accounted for as Fe++. Note,
however, that there are several samples with appreciable concentrations of Fe where the same
procedure would only violate the electroneutrality found. Two of the remaining three large
deviations in x ( x >8%) are from site d where most concentrations are low and where the
numerical inaccuracy of the reported concentrations can explain the deviations found. The
deviation of 10% in a, 801215, coincides with high concentrations of NO3- and K+, but a formal
172
38.30
4.66
41.95
6.25
44.90
9.68
35.91
Lateral water flow
explanation was not found. There are 10 further cases where y >8%. These are all associated
with samples with a low EC25. If the test and the EC measurements are both reliable here, this
may indicate some change in the sample composition during the sojourn in the laboratory.
The following points become apparent from the analyses:
1) Mineral N concentrations (Table 9.3) show strong fluctuations, especially at sites a and bb.
The highest values are mostly found in winter. Guessing from the concentrations, the boezem
might act as a source of nitrogen. The time series for dd shows a puzzling similarity to that of
bb, but this cannot be explained by the influence of boezem water. Probably most of the mineral
nitrogen originates from local processes in the mire; the accumulation of N in the boezem water
may be due to the receiving function of the boezem in winter for surplus water from the mire and
from adjacent polders. Mineral N is not used by the vegetation in any appreciable amount during
the winter season.
The kragge sites b, c, and d feature more constant and also lower concentrations of mineral
N.
Nitrate is mostly the dominant mineral-N component, but occasionally ammonium reaches a
higher concentration.
As regards nitrogen, the differences between open surface water, water in the kragge, and
water from underneath the kragge dominates over a longitudinal gradient, if such exists.
2) The phosphate concentrations show a remarkable pattern. They have been low during 1981
and 1982 and became higher at the end of that period. These higher concentrations do not
necessarily reflect a trend. There is probably no relation through transport of dissolved
phosphate from a to d, however. The source of phosphate is probably mainly a local one.
The samples from underneath the kragge display higher concentrations than those from
within the kragge and from the boezem. This may be due to the physico-chemical conditions
which govern the solubility of phosphate, especially to the redox conditions. As for nitrogen,
any longitudinal gradient is obscured by these differences.
3) The EC-IR and rLI-rTH diagrams (Fig.9.10) reveal a clear gradient from atmotrophic water
at site d to polluted (molunotrophic) water, with a lithotrophic component, in the boezem. The
water underneath the kragge features less pronounced fluctuations than the water in the kragge
and in the boezem. According to their position in the atmo-molunocline, the sites can be
arranged as: d-(c,dd)-(b,cc)-bb-a. This arrangement reflects the influence of the boezem water
penetrating underneath the kragge and then into the kragge itself, where it becomes mixed with
rain water and is changed by various processes taking place in the root zone.
Note the position of the boezem water samples in dry periods approaching RH LOB (Rhine
water) as a result of the inlet of water from Friesland, which is in accordance with the results
obtained in Chapter 5 for the boezem system.
There is no trace of any different, especially lithotrophic, influences in the mire.
Lateral water flow
173
Fig.9.10
EC-IR (a,b) and rTH-rLI (c,d) diagrams of water analyses from De Stobbenribben,
1980-'83
a,c Kragge samples a-d shown in graphs at the left side; b,d samples from flow channel bb-dd shown in
graphs at the right side; L, A, T, M: litho-, atmo-, thalasso-, and molunotrophic benchmark samples;
Mixing contours and the line IR(%)=EC25(mS/m) added to Fig.9.10a for convenience (see Appendix D).
In the upper right graph dd samples are erroneously shown with a solid block rather than a solid disc.
9.6 Conclusions
The investigations treated in this chapter lead to the following conclusions:
1 The electrical conductivity sounding method is suitable for the detection and demonstration
of a longitudinal gradient in the chemical composition of the mire water;
174
Lateral water flow
2 The demonstrated gradient is caused by a different macro-ionic composition of the mire

water rather than by differences in the contents of the nutrients nitrogen and phosphorous;
3 The IR-EC and rLI-rTH diagrams are suitable to summarize certain differences in the
macro-ionic composition of the mire water, and they strongly suggest a gradient from atmoto litho-molunotrophic water caused by the interaction of rain water and boezem water;
4 Boezem water penetrates the quagfen parcels from the ditch at one side through a preferent
flow channel underneath the kraggen, and it exerts an obvious influence upon the chemical
composition of the water in the kragge, even at a distance of more than 100 m from the ditch
in parcel A. At the closed end of the petgaten the influence of rain water is predominant;
5 Conductivity mapping allows for an extension of the sounding results so as to characterize
the sites of the vegetation in terms of hydro-chemical parameters;
6 The middle part of the longitudinal gradient displays strong fluctuations of the electrical
conductivity and chemical composition of the water;
7 The influence of the boezem water can be traced in the temperature regime underneath the
kragge, but this influence can not be separated from influences associated with the varying
structure of the local stands of vegetation (see below);
8 The structure of the vegetation, as it results from the species composition and management,
has a strong influence upon the local temperature regime. This influence could be
ecologically significant (cf. Chapter 7);
9 Although the hydraulic conductivity, in the horizontal direction, in the preferent flow
channel underneath the kragge is large, it is probably not large enough that temperature
inversions may generate density currents;
10 In spite of the high hydraulic conductivity in the preferent flow channel a hydraulic head
gradient in the order of magnitude of 0.001% establishes during the summer season; under
these conditions the flow remains laminar;
11 The discharge of mire water towards the underlying body of groundwater in De
Stobbenribben is large enough to prevent a discharge towards the ditch from continuing for
more than a few days;
12 The layered and anisotropic nature of quagfen complexes as porous media is fundamental to
the understanding of the inflow of water.
Lateral water flow
175
CHAPTER 10
Environmental and vegetational processes in

De Stobbenribben
10.1 Introduction
It has been shown in Chapter 7 that a significant longitudinal gradient exists in the
vegetation of De Stobbenribben. The floral composition indicates an increase of
atmotrophic water effects at the cost of lithotrophic ones as the distance from the
ditch at one side of the quagfen complex increases, although the effects are
somewhat obscured by edge effects near both ends of the parcels. Typically
eutrophic species are only present in any substantial number near the ditch.
Elsewhere oligotrophic species dominate.
It was also shown (Chapters 8 and 9) that the quagfen parcels lose a considerable
amount of water through evapotranspiration and through seepage towards the
underlying body of groundwater. The water loss is only partially compensated by
rainfall, even during the cold season. As a result, a gradient in the hydraulic head
of the phreatic water in the semi-floating quagfen develops, which drives an influx
of ditch water into the various parcels. The flux density decreases as the distance
from the ditch increases. A distinct spatial pattern of bodies of mire water of
various types results, which is related to the characteristic nature of the mire as a
porous medium. The pattern changes with the season.
Processes in De Stobbenribben
177
Apparently, the influx of boezem water is an important ecological factor, both

chemically and physically, and it therefore deserves quantitative attention. In the
present chapter it will be ascertained whether the supposed influx of ditch water is
quantitatively consistent with the hydraulic head gradient in the flow channel and
the apparent hydraulic conductivity of the medium in this channel. Next a model
to explain a fundamental aspect of the water composition gradient will be derived.
Finally it will be discussed how various processes may act together in the
differentiation of the functionally operational environment of the local stands of
vegetation. The basis of that discussion will be formed by a rough balance drafted
for water and chloride.
10.2 Flow rate and hydraulic conductivity in the preferential flow channel
The flow rate underneath a quagfen kragge can be derived from the precipitation surplus and a
possible recharge or discharge of water through the mire bottom. It also depends on the
hydraulic conductivity of the porous medium and on the hydraulic head gradient in the
preferential flow channel. Although but few values for the hydraulic conductivity in quagfen
profiles have been published, an attempt is made to check whether the assumed flow rates are
consistent with the observed hydraulic gradients and with realistic values for the hydraulic
conductivity. Such values were already used in Chapter 9. In this discussion it will be assumed
that the water flow is of the laminar type and Darcy's law may be applied. This is justified when
Reynolds' number Re< ca 10. With an average pore size in the flow channel of ca 0.1 m and a
flow rate of the order of magnitude of 10-5 m/s Re will have a value of ca 1, and the condition is
satisfied.
The hydraulic conductivity in the kraggen in De Stobbenribben can be estimated from the
hydraulic head difference measured between the phreatic level in the closed ends of the parcels,
and the water level in the ditch. In various summer seasons this difference reached a maximum
value between 13 and 21.8 cm, the highest values never pertaining to more than one parcel.
During rainless periods in summer it stabilizes at ca 15 cm on average. Near the ditch the
gradient of the hydraulic head is steeper than farther off (compare Chapter 9, Fig.9.3). At ca 160
m from the ditch the gradient approaches zero. The values for the hydraulic gradient in Table
10.1 have been chosen so as to be representative for a stable, but not extreme, summer situation.
The duration of such periods is about one month. Since, in such periods, the sum of the
precipitation deficit and the seepage towards the body of underlying groundwater can be
estimated (Appendix E, Chapter 8), a balance sheet can be made up for a 1 m wide strip of a
quagfen parcel, divided into three compartments (see Table 10.1).
A preferential flow channel with an average height of 1 m is considered. The influence of the
somewhat lower position of the kragge as the hydraulic head decreases will not exceed 10-20%
of this height and is, therefore, disregarded here. Three regions are distinguished:
1)The first 30 metres near the ditch, where plant growth is more vigourous than elsewhere;
2)The remainder of the quagfen parcel;
3)The baulk at the end of the fen parcel.
Conservation of mass and the Darcy equation yield
Qi = Li (E+D-P) and Qi = -Ki fi A, with A=1,
178
Table 10.1 Hydraulic conductivity and flow rate in a 1 m wide strip of a quagfen parcel
distance from fixed water level (m)
hydraulic head in flow channel (cm)

head gradient (%)
L (m2)
height of flow channel (m)
solutions with (E+D-P)=0.004 m/d

Qi (m3/d)
Ki (m/d)
Qi (m3/d)
Ki (m/d)
Qi (m3/d)
Ki (m/d)
15
30
95
-8
160
baulk
-15
-0.27
145
1
-0.054
65
1
+2.14
7
2
0.58
218
0.26
483
0.16
7.5
1.08
404
0.455
845
0.28
13.1
1.45
54
0.65
1207
0.40
18.7
where
Q the volume of water flowing laterally through the middle of the i-th region (positive for
inflow from the ditch side; m3/d);
A the cross-sectional area of the flow channel (m2)
L the surface area of the quagfen strip between the middle of the i-th region and the assumed
dead end at 160 m from the ditch (the hinterland; m2);
E the evapotranspiration (m/d);
P the precipitation (m/d);
D the discharge to the underlying groundwater (m/d);
K the hydraulic conductivity in the preferential flow channel and in the flow direction (m/d);
f the gradient of the hydraulic head in the preferential flow channel (f is positive at an
increasing hydraulic head, negative at a decreasing one).
The hydraulic conductivity can now be solved for the first 30 m and the next 130 m of the
quagfen length assuming a typical hydraulic head gradient and chosen values for (E+D-P) as
listed in Table 10.1. The hydraulic conductivity is smaller near the ditch, probably as a result of
the deposition of ditch dredge on the banks, a more vigorous growth of plant roots, and a
blocking of the porous medium with bacteria and peaty material. Also a solution is given for the
hydraulic conductivity of the approximately 7 m wide peat baulk separating the closed end of
the quagfen parcel from another ditch, which has the same water level as the ditch at the open
end of the parcels. Accordingly, for this dam, f=0.15/7. Assuming that the seepage through the
dam occurs over a height of 2m, the hydraulic conductivity of the peat dam is listed in Table
10.1. Vegt (1978) found values in the range 0.4-2.9 for peat baulks in this area with the
piezometer method, and Van der Perk & Smit (1975) reported a value of 5.5 m/d, solved from
the Darcy equation, for De Wieden. The present value is apparently at the high end of the range,
but this can easily be explained in view of the rough assumptions made, the uncertainty of the
reference values, and the incidental water
179
Table 10.2 Typical values for the hydraulic conductivity in kragge profiles in De
Stobbenribben
layer
from (m)
to (m)
K (m/d)
field test (Vegt 1978):

kragge
just underneath kragge
peat mud
firm peat, mineral soil
0
0.5
1.5
2.5
0.5
1.5
2.5
-
> 75
>400
1-10
ca 1
derived from hydraulic gradient and balance sheet (see text):

underneath kragge
0.5
1.5
500-1000
supply through the tjasker wind-pump.

Since the chosen hydraulic head gradient is more or less representative for summer
conditions, the highest values for Q and K, obtained with the extreme value (E+D-P)=0.01 m/d,
are improbable. For the non-extreme situation to which the chosen hydraulic gradient applies,
the precipitation deficit must be ca 2 mm/d, so that, for (E+D-P) between 0.004 and 0.007 m/d,
the discharge of water from the quagfen towards the underlying groundwater must be 2-5 mm/d.
Note that the solution for K depends on the assumed height of the preferential flow channel,
while this is not so for Q.
The hydraulic conductivity in quagfen profiles varies with depth and with the firmness and
thickness of the kragge. Typical values obtained with piezometers in field tests in various layers
in the profile of De Stobbenribben are listed in Table 10.2. Extreme values of up to 1500 m/d
have been found in certain other quagfen complexes with a very weakly developed kragge on
the basis of hydraulic head gradients (Van der Perk & Smit 1975). Much lower values were
obtained by De Boer et al. (1977) and by Huijsmans & Zwietering (1980), but their results are
based on field tests with unsuitable piezometers prone to stoppage by the fine and humic peat
muck. Koerselman (1989) derived an average hydraulic conductivity of 64.5 m/d from balance
sheets for quite similar quagfens in the Vechtplassen area (Province of Utrecht, The
Netherlands). His results provide an average for various layers in the profile, with a considerably
thinner preferential flow channel than in De Stobbenribben. Seepage tube tests yielded an
average of 0.19 m/d for the flow channel in his investigations, but the use of such values in his
water balance calculations would strongly violate the mass conservation law for water.
According to his description the method used was not especially suitable for the determination
of the hydraulic conductivity in the horizontal direction. In conclusion, the hydraulic
conductivity in the preferential flow channel in De Stobbenribben is probably between 500 and
1000 m/d. This value is found on the basis of the hydraulic gradient and the water balance, but it
is consistent with local field tests with suitable piezometers. This result is probably also
consistent with Koerselman's, if the different nature of the profile to which it applies is taken
into account.
Since, under the assumptions made, acceptable solutions for K are found, it is concluded that the
influx of water underneath the kragge, with the assumed flow rate, does not violate reasonable
180
assumptions as regards the hydraulic properties of the porous medium. This justifies the
treatment of solute transport in the following section.
10.3 QUAGSOLVE: the mixing of water in the preferential flow channel
In the long run, the movement of water in the preferential flow channel depends on an
alternation of movements in shorter periods. Different water compositions will result from a
complex of processes. These include the movement of whole bodies of water, the mixing of
waters with a different composition, absorption of substances to, or release from, the peat, plant
uptake and release, and dissolution and precipitation. In a simple approach, I will consider
mixing as the dominant process and try to describe the chemical composition of the water in a
quagfen as a result of hydrological relations. Obvious anomalies will the indicate the importance
of processes other than mixing. I will tackle the problem for chloride first, since chloride is not
much involved in the other processes. With regard to chloride the mixing concept includes the
mixing of waters with different concentrations and the diffusion of ions. The quantification will
be treated on an annual basis and checked with the results of water analyses. Quite a few
assumptions have to be made in the resulting model, called QUAGSOLVE. The model
assumes a steady-state. Most quagfens in The Netherlands are not in a steady state, but the
equilibrium case will assist to find out what factors may cause a change, and whether a
particular true quagfen is likely to be involved in change processes. It has appeared (see Chapter
7 as regards De Stobbenribben) that the vegetation of quagfens can remain in a near-equilibrium
state for several decennia at least.
In the initial solution for chloride concentrations in De Stobbenribben the following parameters
are being used:
D
E
Li
P
Qi
Qi+1
Ti
ci
c0
Annual discharge to the underlying groundwater (m/a, variable);

Annual evapotranspiration: 0.35 m/a (Appendix E);
Surface area of i-th compartment of quagfen strip (m2; width is 1 m);
Annual precipitation: 0.8 m/a (KNMI 1972);
Annual lateral inflow into i-th compartment in flow channel (m3/a, variable);
Annual lateral outflow from i-th compartment (m3/a, variable);
Annual exchange between floating mat and flow channel in i-th compartment
(m/a, solved variable); in time a downward transfer will alternate with an upward
transfer as a result of the asynchronous incidence of precipitation and
evapotranspiration and a variety of other processes. Over the year the upward and
downward components are equal in magnitude, but the upward movement
transports salts with the concentration of the flow channel, whereas the downward
flow has a salt concentration equal to that in the kragge;
Chloride concentration in i-th compartment (g m-3, numerically equal to mg/l);
Chloride concentration in precipitation (3 g m-3, Appendix D);
Prefixed subscripts denote concentrations in different bodies of water:
ck upper layer (variable);
cz average in preferential flow channel (variable);
cl lateral inflow in preferential flow channel (variable; cl,1= 75 mg/l).
181
Fig.10.1 QUAGSOLVE balance sheets per compartment

a: Water balance for compartment i
b: Chloride balance for compartment i
A diagram for the concentrations and the balance sheets for water and chloride is given in
Fig.10.1. A steady state is assumed (no change in water storage and chloride concentrations). It
is also assumed that the kragge layer is well-mixed in all directions, while the flow channel is
well-mixed in the vertical direction and over the full width of the quagfen strip, but not in the
longitudinal direction. Note that the electrical conductivity and temperature soundings prove that
a small vertical gradient also exists in the flow channel. The well-mixed assumption is a
simplification justified only by the importance given to the gradient in the longitudinal direction
and to the concentration difference between the floating mat and the flow channel.
In the steady state (no change in water storage and chloride concentrations) the following
formulae apply:
(1)
Pc0 + Ticz,i = (Ti+P-E)ck,i;
(relation between floating mat and flow channel)
(2)
Qicl,i + LiPc0 = Qi+1cl,i+1 + LiDcz,i;
(relation within flow channel, in the longitudinal direction)
(3)
Qicl,i + Li(Ti+P-E)ck,i = Qi+1cl,i+1 + Li(Ti+D)cz,i;
(combination of (1) and (2))
(4)
(mass conservation of water)
182
Qi + Li(P-E) = Qi+1 + LiD.
Fig.10.2 Chloride concentrations underneath a kragge as a function of downward leakage of

water (a), and of chloride concentrations in the supplying ditch (b)
With the exception of the values mentioned in the diagram, L=160 m, P=0.8 m, E=0.35 m, c1=75 mg/l,
D=2 m/a, and c0=3 mg/l
By adhering to the probably reliable assumption that the influx of ditch water does not reach
much further than Li=160 m into the quagfen, the overall equilibrium chloride concentration in
the flow channel can be calculated. Since the chloride content in the floating mat is considered
constant, no matter at what level, the annual inputs into the flow channel equal the sum of the
inputs with P and Q1 with the respective concentrations c0 and cl,1. The output is entirely
determined by D with the concentration cz, so that this can be solved for given values of D: for
D=1 and D=2, with cl,1=75, cz is 44 and 59 mg/l, respectively. Fig.10.2a illustrates the variation
of cz with D for P=0.8 m/a and for P=0.5 m/a. It appears that cz decreases sharply when D falls
below some critical value. An increase of E has a similar effect as a decrease of P: the curve for
E=0.5 m/a would lie in between the curves for P=0.8 and 0.5 m/a, with E=0.35 m/a. At this
183
instance it can be concluded that relatively high chloride concentrations at a considerable

distance from the ditch can only be explained, in our climate and in the absence of upwelling
groundwater, by a substantial downward discharge of mire water, provided that a compensating
lateral influx of water is activated. In a reverse application, D can also be solved on the basis of
observed cz and cl,1 values, where cz=( Licz,i)/( Li). Such applications with the average
concentrations for 1980-'83 and those for 1984, with cl,1= 90 and 75 mg/l, respectively, yield a
range of 2-3 m/a for D, but still higher values would result if the observed lower chloride
concentrations in the ditch were used (see Table 10.3 and later comments). The linear
dependence of cz on the concentration cl,i in the ditch is shown in Fig.10.2b. The influence of
deviations in c0 is small and disregarded here.
The basic QUAGSOLVE formulation only requires the boundary conditions that the
concentrations at the ditch end equal the ditch water and that at the opposite end of the flow
channel, where the inflow is zero, an atmotrophic quagfen concentration is found. The latter is
the equilibrium chloride concentration in an entirely atmotrophic system, ck=cz = Pc0/(P-E) =
5.3 mg/l under the assumptions made. When the quagfen is partitioned into segments, analogous
boundary conditions apply to each segment, and a pattern can be generated for the chloride
concentrations.
The QUAGSOLVE model will now be used to compute the chloride concentrations with Q1,
cl,1, c0, E, P, and D as input data. In order to do so the compartments are taken very small, and it
will be assumed that cz,i=cl,i+1. This means that complete mixing is assumed in the flow channel
in each compartment. The floating mat is not considered, since ck can be derived from cz when T
is given.
With (5)
cl,i+1 = cz,i,
the essential formulae now become

(2a)
cz,i = (Qicl,i + LiPc0) / (Qi-Li(E-P)), and
(4a)
Qi+1 = Qi-Li(E+D-P).
D has vanished from the solute balance sheet (2a) due to the mixing assumption (5). Fig.10.3
shows the resulting gradient for chloride concentrations, for a variety of input data, along with
some observed values from Table 10.3. Q was derived from P, E, D, and L, the distance from
the ditch over which the lateral flow is present (the distance to the dead end). L, P, D, and cl,1
were varied. The influence of variation of D, noticed already in Fig.10.2, is obvious: when the
downward discharge of mire water decreases, the gradient in the chloride concentrations in the
flow channel changes from convex with a sharp drop at the dead end of the parcel to concave
and gradually approaching a stable atmotrophic state at a relatively short distance from the ditch
(Fig.10.3). A smaller value for P (or a larger value for E) results in a more pronounced
concentration drop at the dead end (Fig.10.3). Due to a variation of these factors between years,
a relatively strong variation of the chloride concentrations in the back end of a quagfen parcel
may be expected. The graph also shows that it will not always be easy exactly to locate the sharp
drop at the dead end in field situations. For this reason it will also be difficult to estimate L
within 30%. A variation of the chloride concentrations in the ditch will be of relatively great
influence, especially when D and E are large and P is small, since this may cause high
184
Fig.10.3 Chloride concentrations underneath a kragge as a function of the distance from the
supplying ditch
With the exception of the values mentioned in the diagram, L=160 m, P=0.8 m, E=0.35 m, c1=75 mg/l, D=2
m/a, and c0=3 mg/l
concentrations at a large distance from the ditch.

Chloride concentrations in parcel A have been observed in 1970, 1973-'74, 1979, 1980-'83
(chapter 8), and 1984 (Kooijman 1985). A tabular summary of these data is given in Table 10.3.
It appears that the concentrations have varied considerably. Noticeably, at distances farther than
ca 100 m from the ditch a decreasing trend is exhibited. I have especially considered the
1980-'83 and 1984 data, since these represent estimates of the average for each period. The
1980-'83 values are each based on 20-22 samples. For the 1984 ck-means nk=22, for the czvalues nz=44, since both the 0.7 and 2.5 m depth values were used to estimate the concentration
in the flow channel proper. Differences between cz values within each data set were all
significant (t-test, level 0.01), with the exception of the 1980-'83 pair (0 m, 55m).
In addition to the calculations reported above, QUAGSOLVE can be used to check the
consistency of these observed values. This is done for separate compartments with all chloride
concentrations given and no restrictions on Q and with D=1 m/a (2.74 mm/d) and D=2 m/a
(5.48 mm/d). Three compartments are defined, for which the parameters and variables, including
model solutions, are listed in Table 10.4. The annual exchange T is included in this model
solution, although it does not interfere with the gradient of concentrations in the flow channel as
long as a steady state is assumed. The results for T will be discussed later in this section.
It appears that the solution with D=2 m/a (5.5 mm/d) almost satisfies the condition Qi+1=0 at
the end of the last compartment, at 198 m from the ditch, rather than at 160 m. This solution also
nearly satisfies the condition that the lateral outflow from a certain compartment must equal the
lateral inflow into the next one and it is very close to the L=200 case in Fig.10.3. Solutions
with substantially different values for D, at the chosen situation in De Stobbenribben, obviously
violate the continuity of water flow required by the direct linkage of compartments. Apparently
185
Table 10.3 Chloride concentrations observed in De Stobbenribben, parcel A

Distance
from ditch
(m)
701118/23
730724
730920/21
731125
740403
791017
791115
1980-'83
1984
22
35
55
102
110
135
158
66
73
175
ck
cz
84
ck
cz
96
57
78
41
79
48
91
40
60
ck
cz
84
10
31
49
86
64
85
37
69
ck
cz
82
67
91
24
84
44
98
22
61
ck
cz
93
74
23
92
42
68
38
76
ck
cz
93
80
41
64
ck
cz
55
73
43
56
ck
cz
76
57
76
33
62
ck
cz
67
69
73.5
5
64
18
57.5
13
46.5
mg/l; in several cases average values have been given
the seepage rate D=2 is consistent with the results found in Chapter 8: the DOPPSOL and
FOUSOL methods yielded average values of 5.6 and 4.1 mm/d, respectively, for parcel A.
Further calibration does not seem to be very useful, since tests are difficult for this rough model,
which does not include any kinetic details. An extension of QUAGSOLVE to cover
non-steady-state situations, in relation to more detailed measurements of conductivity profiles
and hydraulic gradients, seems possible, but lies beyond the scope of the present project. It is
worthy of note that the balance sheet approach followed in the QUAGSOLVE model only
considers the overall results of processes. The details of these processes remain untreated.
186
Table 10.4 Application of the model QUAGSOLVE with assumed values for De Stobbenribben, parcel A.
Run
D (m/a)
Compartment (i)
L (m2)
1
110
I
1
2
48
3
40
42
70
75
58
16
54
58
46
8
35
46
30
ck (mg/l)
cz (mg/l)
cl,i (mg/l)
cl,i+1 (mg/l)
Solutions:
Ti (m/a)
Qi (m3/a)
Qi+1 (m3/a)
0.59
0.13
231
105
171
79
0.04
40
18
1
110
II
2
2
48
3
40
42
70
75
58
16
54
58
46
8
35
46
30
0.59
0.13
308
137
138
63
0.04
53
-9
Compartment 1 starts at the ditch
One other solved variable, viz., the annual exchange T of water between the floating mat and the
flow channel, can be evaluated (Table 10.4). The decrease of T with the distance from the ditch
suggests that the kragge at a greater distance is more capable of temporarily storing a
precipitation excess. This corresponds to the images produced on the basis of conductivity
soundings (Chapter 9), and it might be explained by the firmer kragge structure,by an obviously
lower density and smaller area of weak spots and pools providing windows to the flow channel
below, and by the opulent growth of Sphagnum species. Sphagnum vegetation is capable of
storing a substantial amount of water above the phreatic level and it considerably reduces the
hydraulic conductivity in the vertical direction. The highest value for T, 0.59 m/a, in the first
compartment, says that about 75% of the rain falling on the kragge is, after some mixing with
water already present there, discharged into the flow channel, thus urging an additional supply
from that channel during drought. T apparently reflects the asynchronous incidence of
precipitation and evapotranspiration as this is controlled in the kragge, but the higher values are
obviously caused by additional processes, such as a forced exchange under the influence of
pressures exerted on the kragge during mowing.
The general conclusion is that the gradient of the chloride concentrations observed in 1980-'84 is
approximately consistent with the quantified influx of ditch water and a reasonable variance of
the exchange rate between the kragge and the underlying flow channel. The kragge structure
governs this exchange rate, possibly due to the varying frequency of pools and weak spots and to
the varying abundance and characteristics of the moss layer. Apparently the exchange rate may
approach zero, thus allowing the floating mat to become atmotrophic. However, deeper-rooting
species may still be able to reach the underlying body of mire water in the flow channel, which
is maintained at higher concentrations of solutes. The rate of discharge from the flow channel
towards the body of groundwater is possibly some 3-5 mm/d. Substantially different
assumptions would definitely lead to unrealistic results. At the divide between the influx zones
the quagfen and the underlying body of mire water may become atmotrophic. In De
Stobbenribben the entirely atmotrophic situation is not properly reached in parcel A, but almost
187
so in parcel B, where the kragge is very thick and firm, and the transmissivity of the flow
channel is, therefore smaller. This is in accordance with chloride concentrations measured in the
flow channel below the kragge in parcel B between 1970 and 1983 at a sampling site 145 m
from the ditch. They were ca 25 mg/l. The quagfen surface is also slightly dryer here in summer,
however.
10.4 Deviating concentrations of non-conservative constituents
The QUAGSOLVE exercise showed that the chloride concentrations at various locations and
depths in the quagfen can be explained by the mixing of ditch water flowing in at one side of the
quagfen parcels with rain water penetrating from above through the kragge. Most other ions,
however, are involved in various other processes and it seems worth-wile to check the deviations
found. This was done by computing, on the basis of the relative volumes of rainwater and ditch
water as expected from chloride concentrations, the concentrations of other constituents as these
would follow from mixing, and checking this against analytical results. Note that the EC-IR and
rTH-rLI diagrams presented in Chapter 9 already suggest a general applicability of mixing
processes as regards the major ions, although the IR decreased a bit more than was expected as
the rain water influence increased. Roughly three groups of additional processes are to be
considered. These are summed up below with an indication of the constituents expected to be
involved to such an extent that the deviations can be large enough in regard of the actual
concentrations to be observed.
1) Seasonal uptake or release by the soil and the biocoenosis, governed by biological processes
and possibly different in magnitude according to local factors reflected in the structure of the
vegetation. Expected to hold for phosphorus, nitrogen, potassium, and carbon;
2) Exchange between the solution and the peat matrix. Expected to hold for calcium and
possibly for the other cations;
3) Transformations due to the very different environment in the flow channel and the kragge, as
compared to the surface water system. Expected to hold for sulphur, nitrogen, phosphorus,
and carbon.
These processes are characterized by different time scales and parameters that have not been
quantified, so that only tentative suggestions can be raised as regards their relative importance.
In some cases, e.g., when transformations are involved, ditch water and rain water analyses are
no appropriate sources for the calculation of mixing processes within the quagfen. This was
checked by replacing rain water and ditch water, in the calculations, with analyses of kragge
water at the same location and analyses pertaining to the nearest ditchward site in the flow
channel, respectively. I will refer to the fundamental assumption of this calculation as local
mixing below.
Some initial problems must be solved: reference analyses for ditch water and rain water, and
a representative set of analytical data must be chosen.
For statistical reasons the 1980-'83 data-set will be the main one to be used. The calculations
have also been applied to mean analytical results reported by Kooijman (1985) for 1984, and I
will comment on the comparison when relevant. The ditch water reference for the 80-'83 data
was taken from the data-set itself. The ditch water samples in the 1984 set exhibit slightly lower
chloride concentrations than samples at a distance of 22 m from the ditch. The 1980-'83 ditch
water reference could not serve the same purpose for the 1984 set since especially the calcium
concentrations were more than proportionally lower in the 1984 data. I therefore considered the
188
ditch water itself a mixture of rain water an a more extreme, unmixed ditch water, for which I
computed the concentrations. The benchmark analysis AT-W80 (Appendix D) was used as a
reference analysis for rain water with the 1980-'83 data. For 1984 rain-water analyses were
available in the data-set. This RAIN-84 reference is more concentrated for all constituents than
AT-W80. This is of significant influence when the deviations for K and P are calculated, as
commented below. The various reference analyses are summarized in Table 10.5. Fe was not
analysed in AT-W80 (the value 0 was used in the calculations).
Table 10.6 gives the results of the calculations for the 1980-'83 data set. Included in the table is a
site in parcel B, 145 m from the ditch (see chapter 9). Since parcel B has a much firmer and
thicker kragge, this site is more or less representative for the dead end of that quagfen parcel. It
was selected to replace a site at ca 190 m from the ditch in parcel A that was disturbed by edge
effects due to the proximity of the peat baulk at the closed end proper, the frequent visitation by
people, and the occasional influence of a replica of a small local type of wind-pump (tjasker).
The means in the data-set were considered as estimates of the time average, and apparent
gains and losses (negative gains) for each constituent were computed, referring to the expected
concentration in a mixed sample of ditch water and rain water, of which the ratio was derived
from chloride concentrations. Two (null-)hypotheses are tested for the various constituents with
a t-test (one-sided, 0.05 level of significance).
The first hypothesis is that the means for the sites do not differ significantly from the ditch
mean. The hypothesis is rejected for almost all data. The exceptions are: 1.2 m depth at 55 m
from the ditch (only SO4, pH, N-inor different), Fe, K, and P at all sites and depths except at 1.2
m, 145 m from the ditch in parcel B (Fe is also different at 0.0 m depth and 55 m from the
ditch). Apparently there are no depth and longitudinal gradients, at this level of significance, for
P, K, and Fe, and, at its entrance in the flow channel, SO4, pH, and N-inor are the first variables
marking the changing ditch water composition.
The second hypothesis is that the means for the sites do not differ significantly from what
would be expected on the basis of a mixing of ditch water and rain water according to the
method used. The rejection of this hypothesis is shown in bold face in Table 10.6 (right-hand
side), expressing, among others, significant losses of Ca and N-inor at all sites and depths, a
reasonably good match with the mixing assumption for Na, and a loss of SO4 in the flow
channel. As far as this, similar results were obtained with the 1984 data set.
Below the tested behaviour of some constituents in the quagfen is reviewed, and reference
is made to a possible ecological impact of this behaviour. In some cases an additional t-test was
applied to see whether the concentrations differed significantly also within the quagfen.
P and K
P and K concentrations in the quagfen do not differ significantly from the ditch-water
concentration, thus appearing to be relatively constant. When the sites are grouped, it appears
that the concentrations for P actually show an increase as the distance from the ditch increases. P
and K are important plant nutrients. At low supplies they may prevent a dominance of large
helophytes in the quagfen and thus favour a high species-richness. P is involved in many
transformations, and it is unclear how these interfere with the dissolved fractions represented in
the analyses.
189
Table 10.5 The composition of the reference samples used to compute gains and losses of
various constituents in the mire water in De Stobbenribben
NAME
pH
Ca
mg/l
AT-W80
RAIN84
DITC84
UNMIX
DITC80
4.2 0.4
5.8 3.4
7.6 43.3
47.7
7.6 70.2
Mg
mg/l
Na
mg/l
0.2
1.6
0.78 2.7
8.50 38.7
9.36 42.7
10.17 41.4
K
mg/l
0.23
1.76
3.47
3.66
4.70
Cl HCO3
mg/l
mg/l
3
5
67
74
76
0
7
116
128
185
SO4 EC25 NO3

mg/l mS/m mg/l
6
8
24
26
46
5.0
5.4
61.1
67.3
60.8
NH4
mg/l
3.41
3.87
1.34
1.06
1.70
2.42
0.43
0.21
P
mg/l
0.01
0.05
0.02
0.02
0.03
Fe
mg/l
0.08
0.10
0.10
1.28
Ninor
mg/l
2.09
2.76
0.64
0.40
0.82
Table 10.6 Concentrations (left side), percentage rain water and deviations from the
expectedconcentration in the mixture (right side) in samples from De
Stobbenribben (1980-'83)
CONCENTRATIONS
MIXING WITH RAIN WATER, GAINS AND LOSSES
Parcel
<---A--->

<---A--->

<---A--->

<---A--->

<---A--->

<---A--->

Distance from
ditch (m)
55
135 145
55
135 145
55
135 145
55
135 145
55
135 145
55
135
145
_____________________________________________________
_______________________________________________________
mg/l
mg/l
mg/l
% rain water
mg/l
mg/l
[Cl]
[Ca]
[Mg]
based on [Cl]
gain of [Ca]
gain of [Mg]
0.0 m
57.4 32.8 14.4
44.7 20.5 5.1
6.91 3.89 1.80
26
59
84
-7.5 -8.4 -6.3
-0.67 -0.37 +0.04
1.2 m
76.0 62.0 25.0
64.8 49.8 16.6
9.63 7.56 4.11
1
20
70
-5.1 -6.8 -4.8
-0.49 -0.66 +0.92
________________
_________________ _______________
______________
________________
__________________
mg/l
mg/l
mg/l
mg/l
mg/l
mg/l
[Na]
[K]
[HCO3]
gain of [Na]
gain of [K]
gain of [HCO3]
0.0 m
30.4 17.5 8.1
4.07 4.06 4.08
117 52
11
-0.7 -0.3 +0.3
+0.53 +2.01 +3.15
-20
-24
-18
1.2 m
39.3 33.0 18.0
5.85 4.67 2.50
202 154 69
-1.9 -0.6 +4.5
+1.17 +0.85 +0.93
+17
+5
+13
________________
_________________ _______________
______________
________________
__________________
mg/l
mg/l
mg/l
mg/l
mg/l
mg/l
[SO4]
[Fe]
[P]
gain of [SO4]
gain of [Fe]
gain of [P]
0.0 m
32
22
16
0.20 0.66 1.18
0.02 0.06 0.07
-4
-0
+4
-0.75 +0.14 +0.98
-0.01 +0.04 +0.05
1.2 m
23
19
10
1.37 0.92 8.11
0.05 0.06 0.09
-24
-19
-8
+0.10 -0.10 +7.73
+0.02 +0.03 +0.07
________________
_________________ ________________
______________
_________________
__________________
mg/l
mS/m
mg/l
[N-inor]
EC25
pH
gain of [N-inor]
0.0 m
0.17 0.25 0.21
42.4 23.8 10.7
7.1
6.5
5.2
-0.98 -1.32 -1.68
1.2 m
0.41 0.25 0.39
57.5 45.8 20.4
6.7
6.3
5.8
-0.49 -0.82 -1.32
Significant results (t-test, one-sided, significance level 0.05) have been printed in bold face
The absence of any significant differences in the P and K concentrations in the quagfen leads to
different conclusions depending on the assumed inputs with rain water. If the concentration in
rain water is low, as in the AT-W80 reference, a net mobilization of K and P within the quagfen
is found. Assuming a larger atmospheric input of P and K, as in the RAIN-84 data, which could
easily result from local factors, would lead to the conclusion that P and K are immobilized in the
quagfen. The 1984 data, with reference samples for rain water and ditch water from that data-set,
indeed yield this result.
An analysis of the variation of K concentrations in time shows relatively strong fluctuations.
Occasionally a decrease of K concentrations during the growing season was noticed, especially
190
in dense stands of helophytes, whereas burning, thawing after frost periods, and the activities
associated with mowing sometimes led to extremely high peak concentrations for both P and K.
In conclusion it is not possible to attribute a gain or loss in P and K concentrations to the
vegetation cover. Their concentrations in the quagfen seem to result from local processes. The
total stock in the peat and the vegetation is probably large, but mobilization and diffusion rates
may limit the supply of these elements to vigorously growing vegetation.
Inorganic nitrogen
Sulphate and inorganic nitrogen concentrations and pH differ between the surface water in the
ditch and the mire water in the quagfen. They are determined by processes that in turn depend
on the redox state of the medium, as this is controlled by barriers for the exchange of gases with
the atmosphere. The assumptions on which the QUAGSOLVE model was based are irrelevant
as regards these constituents. While, at both depths, pH and SO4 differ along the gradient, there
is no such significant trend in the case of N. Upon its entrance into the flow channel underneath
the kragge nitrate is replaced by ammonium, and a considerable amount of N is lost. According
to a test on local mixing the concentrations do not change very much within the quagfen.
As are P and K, N is often investigated as a possibly limiting nutrient. Next to the supply of
nitrogen in various inorganic and organic forms with (rain-) water, a fixation of atmospheric
nitrogen has been frequently reported for mires (Dickinson 1983, Koerselman et al. in press) and
this will also occur in De Stobbenribben, where Myrica gale and Alnus glutinosa are the most
important species with associated microbial N-fixation. The possible importance of N-fixation
by blue-green algae in quagfens is unknown. As is the case for P and S, N is strongly involved in
many transformations. The concentration in rain water is 2-3 times as high as it is in the ditch
water, and this explains the greater apparent losses seen in samples with a greater percentage of
rain water. The concentrations do not show any clear time dependence in De Stobbenribben, so
there is no sound basis to attribute losses from the dissolved phase to plant uptake. Rather, the
nitrogen input from the atmosphere becomes immediately involved in uptake and
transformations. A check with the assumption of local mixing only shows a substantial loss
(0.40 mg/l) at 55 m from the entrance of the flow channel. If this were due to an uptake by the
vegetation between the site proper and the ditch, the average supply of dissolved inorganic N
from the inflowing water over this distance would be, somewhat depending on assumptions
concerning a possible gradient, 1.5-2.5 g m-2 a-1. This will be shown to be a relevant supply later.
Calcium
Calcium losses from the dissolved phase in the quagfen are appreciable. This must be due to
processes here collectively named exchange with the peat matrix, since there is no evidence of
any substantial precipitation of calcareous minerals. The differences between sites are highly
significant, but a test on local mixing reveals that the deviations from the mixing assumption
especially apply to the kragge compartment and the first length of the flow channel. These are
the parts of the quagfen where the peat matrix reaches its greatest bulk density, and where,
accordingly, exchange processes may be expected to be more pronounced.
191
The exchange of cations in peats is a notoriously difficult matter (Clymo 1983), and only
few data are available for the interpretation of the Stobbenribben data. According to Sikora &
Keeney (1983) and to Clymo (1983) at pH near 6-7 a peat may be expected to have a cation
exchange capacity of ca 1-2 mmol(c)/g. With a bulk density of 80 g/l in the kragge, the resulting
capacity to hold and exchange cations would be 80-160 mmol/l. Moreover, there is some
evidence (Malmer 1962, Clymo 1983) that the distribution of the various cations between the
peat and the water exhibits a general pattern, such that about 1-2 % of the calcium, 3.5% of the
magnesium, 10-15 % of the potassium, and 40% of the sodium remain in the water at
equilibrium. If this is applied to the sampling site at 135 m from the ditch, it would appear that
some 80 mmol(c) were held at the exchange sites of the peat, so there is a possibility that there is
room for more. The concentrations at 1.2 m are more than twice as high, at this sampling site,
and the bulk density is certainly lower, so that the peaty matrix in the flow channel is probably
saturated with cations and an exchange of ions may occur depending on the ionic composition of
the water. On average, the calcium concentration in this part of the flow channel will not change
much. While this is in accordance with the results obtained from the local mixing calculation,
the significant decrease of the calcium concentration at 55 m from the ditch can not be similarly
explained. The peat matrix near the entrance of the flow channel would, in spite of its greater
bulk density, be saturated with cations at the high ambient concentration, so it would be unable
to extract more calcium from the incoming ditch water. None of the other cations (except
ammonium, for which the calculations are less reliable due to irregular fluctuations) show a
possibly compensating increase here.
The 1984 data-set reveals a similar picture, although the ditch-water reference concentration
was much lower for calcium than in 1980-'83 (see Table 10.5). The lower concentration and
lower proportion of calcium in the solution apparently still suffice for a continued transfer
towards the peat matrix, raising the question whether this may have discharged calcium during
the winter. The available data are inadequate to answer this question.
On rare occasions in the 1970-'74 sampling period, and very clearly and systematically at
158 m from the ditch in 1984, samples taken from below 1.5 m underneath the kragge surface
showed high Mg concentrations (12-20 mg/l). The calculations with the 1984 data set revealed
that the relevant cases could be attributed to an exchange of magnesium against calcium. This
would mean that the local peat (possibly inclusive of some clay) has formerly been in
equilibrium with a type of water with a higher proportion of dissolved magnesium ions,
probably due to the former influence of brackish water (see Chapter 5).
Whatever processes are involved, the evidence is that the quagfen is a substantial sink for
calcium flowing in with the ditch water in the flow channel underneath the kragge. According to
Kelts & Hs (1978), calcium and magnesium molecules provide a surface for the sorption of
phosphorus. Kemmers (pers. comm., 1990) found a statistical correlation between sorption of
phosphorus at peats and the saturation of the exchange sites with calcium. While no precise
information is available, the role of the base state of fen mires, as indicated by their flora and
vegetation, could relate to the sorption capacity for phosphorus.
Conclusion
Summarizing the above results, it appears that the availability of phosphorus and potassium in
the quagfen cannot be related to a direct supply with inflowing ditch-water. For nitrogen the
situation is slightly different. In the quagfen zone bordering the ditch the additional supply of
nitrogen with inflowing water may be of some importance. The atmospheric input is
quantitatively important for all three of these ions and the possibility that a substantial amount
192
goes into biomass production even before the main body of mire water is reached cannot be
excluded. The uncertainty with regard to the magnitude of the atmospheric inputs of P and K
renders any conclusion concerning these elements ambiguous, however. Calcium is accumulated
in the peat matrix of the mire, but it is not clear to what extent various parts of this matrix are
presently saturated, and whether, and at which rate, a discharge of calcium may occur. There can
be no doubt, however, that the supply of calcium is governed by the influx of relatively
calcareous ditch-water underneath the kragge.
10.5 Gradients in plant biomass and nutrient state in De Stobbenribben
In the preceding sections it was attempted to describe processes that possibly maintain the
observed longitudinal gradient in the base state of the quagfen parcels in De Stobbenribben. This
gradient concurs with a gradient in the species composition (Chapter 7), and the present question
is whether this concurrence can be partially understood from quantitative relations with plant
growth.
The first relevant question regards the biomass production. Judging from the height and density
of the local stands of vegetation a gradient in plant biomass and production in De Stobbenribben
has been presumed from the beginning of this research project. Indirect measurements of plant
biomass became available through remote sensing images. Since near-infrared radiation is
partially transmitted by plant leaves, its reflection increases with the number of leaf layers in the
radiation path, until saturation is reached in fairly dense stands of vegetation. Radiation in the
visible wavelengths, especially the red, is strongly absorbed by green leaves. The reflection
therefore increases according as a greater part of the ground surface is covered with green plants.
The ratio between near-infrared and visible reflection is accordingly correlated to the green
biomass of vegetation (Hoffer & Johannsen 1969, Knipling 1969). Adequate numerical results
not requiring the processing of remote sensing images were obtained in the field with hand-held
radiometers by Brand & Leemburg (manuscript 1978) and by Boeye (1983). Brand & Leemburg
incorporated a calibration through direct measurements of biomass by cutting, drying and
weighing of small plots of vegetation, and by doing so recorded the development of living aerial
biomass of vascular plants from near zero in April to ca 656, 341, 246, and 267 g m-2 in July in
the vegetational zones A, B, C, and D, respectively (Fig.10.4). In remote sensing images the
reflection ratio shows a sharp drop at the border of the Calliergonella-Phragmites reed zone
(Chapter 7) bordering the ditch. This drop is also manifest from the measurements with a hand
radiometer by Boeye (1983, Fig.10.5). These measurements confirm that the longitudinal pattern
that has been observed on aerial photography including the near-infrared band, from 1971
onwards, and that was floristically identified as a pattern of vegetation types, also reflects a
standing crop pattern.
A more detailed quantification of plant biomass in relation to nutritional factors was realized
by Kooijman (1985, see also Verhoeven et al. 1988) with direct methods. Her results,
summarized in Table 10.7, were obtained at three sites in quagfen parcel A in De Stobbenribben
at distances of 22, 102, and 158 m from the ditch, in the vegetational zone AB (transition
between Calliergonella-Phragmites reed and Scorpidium-Carex fen), B Scorpidium-Carex fen),
and BC (transition between Scorpidium-Carex and Sphagnum-Carex fen), respectively. The
maximum above-ground living biomass, determined from monthly measurements, was reached
in the beginning of August. The biomass in zone BC was slightly higher than that in zone B, due
to the higher moss biomass, but still markedly smaller than in zone AB. Moss biomass in zone
AB was only about 50% of that in zone BC, however, with zone B in an intermediate position.
193
Fig.10.4 Relation between the near-infrared to red reflection ratio and measured above-ground
biomass of phanerogams in various stands of vegetation in De Stobbenribben
Data from Brand & Leemburg (mscr.); The larger symbols indicated with a letter represent mean values
(Note that apparently some shifts in the vegetational pattern occurred between the 1970s and
1984 (see Chapter 7). At 102 m from the ditch the vegetation type indicated in Table 10.7 had
replaced a vegetation dominated by Typha angustifolia and Scorpidium scorpioides.)
Most of the N and a substantial part of P released is not incorporated in the above-ground
biomass in zones B and BC. Conversely, the stand of vegetation in zone AB seems to use as yet
unidentified additional sources of nitrogen and phosphorus. As regards phosphorus this can not
be ascribed to net inputs of rain and ditch water. For nitrogen the atmospheric input (Table 10.8)
and the input with the surface water, which appears to apply to this zone only (cf Section 10.4),
194
Fig.10.5 Decrease of the biomass (dry weight) with the distance from the ditch, De
Stobbenribben, parcel A
Very roughly estimated on the basis of reflection measurements by Boeye (1983)
are substantial. If it is assumed that, within this zone, the stands of vegetation nearest to the ditch
can benefit most from the dissolved inorganic nitrogen, this source might be of the order of
magnitude of 4 g m-2 a-1 at 22 m from the ditch. At an atmospheric input of 4.8 g m-2 a-1 and a
biological fixation of 2 g m-2 a-1 (see below) this would reduce the overall deficit to (14.3-2.810.8) g m-2, or 5%. This is negligible in view of the probable errors in various terms and in view
of the relocation of nutrients from the root system into the aerial shoots.
In conclusion, there is no straightforward relation between the local net mobilization (release) of
nutrients and plant growth in any of the vegetational zones, but there possibly is a relation
between (1) the input of dissolved inorganic nitrogen in the inflowing ditch water and plant
growth in the zone near the ditch, and, qualitatively assessed only, between (2) the inflow of
ditch water and the reduced release of nutrients from the soil in zone AB. Other factors
apparently determine the rate at which the local stand of vegetation uses the resources in the
quagfen.
These data and the results obtained in the preceding section can be combined into a, of necessity
rough, nutrient balance sheet as presented in Table 10.8. The input with surface water was
computed from the concentration in the feeding ditch and a 300 m/a inflow supplying 200 m of
quagfen length (see section 10.3). Outflow with groundwater was calculated according to
concentrations in the flow channel and a downward discharge of 2 m/a. The dry deposition
figure was loosely derived from Heij & Schneider (in press), biological fixation is guessed on
the basis of data in Dickinson (1983) and Etherington (1975), considering that the values found
by Koerselman et al. (in press) might be too low in regard of the contribution by blue-green
algae, and the amounts of nutrients in biomass were based on the average for the three zones in
Table 10.7 (Kooijman 1985). Note that 50-75% of this is harvested as hay. A near-equilibrium
exists for N, while P and K are lost at relatively high rates. For these elements the internal
195
Table 10.7 Ecological characteristics of three measuring stations in De Stobbenribben, 1984

Vegetational zone
Distance from ditch
AB
22
B
102
BC
158
Dominant mosses
Scorpidium
scorpioides
Calliergonella
cuspidata
Sphagnum
subnitens
Sphagnum
papillosum
S. flexuosum
Dominant herbs
Phragmites
australis
Carex elata
Juncus
subnodulosus
Juncus
subnodulosus
Moss species (0)

Vascular plants (0)
15
39
13
24
17
37
Above-ground living biomass

herbs (1)
1065
mosses (2)
98
234
154
243
228
Water composition
upper 0.5 m
below 0.5 m
similar to ditch
similar to ditch
atmotrophic
similar to ditch
atmotrophic
varying
Soil K release (3)

Soil N release (3)
Soil P release (3)
N in biomass
P in biomass
K in biomass
-0.9
2.8
-0.07
14.3
0.89
12
4.5
15.0
0.32
3.8
0.16
4.6
8.5
16.8
1.30
4.6
0.23
5
(m)
(g m-2)
(g m-2)
(g m-2)
(g m-2)
(g m-2)
(g m-2)
(g m-2)
(g m-2)
Source: Kooijman 1985, Verhoeven et al. 1988

(0) In 8 times 10 squares of 400 cm2; (1) Maximum value, reached between July, 24, and August, 24; (2) Average of 8
determinations between April and October; (3) April-September, upper 0.4 m of kragge
mobilization must be the main source, also at the whole-quagfen scale, and a possible relation
with the inflow of surface water should be explained by means of processes governing the
internal availability.
This means that, for P and K, the quagfen largely exhausts its own formerly accumulated
resources, whereas the processes involved are determined both by the state of the quagfen and
by its surrounding environment, such as here analysed for the water flow system and the harvest.
If it is considered that atmospheric inputs of N, and concentrations of this element in surface
water, are generally agreed to have been (much) lower in the first half of the present century,
then it appears that internal cycling of N may also have been more important in the past.
Intuitively, the time of harvesting will be very important, since especially the large helophytes
196
Table 10.8 Rough balance sheet for nutrients in De Stobbenribben (all values in g m-2 a-1)
N
Atmospheric deposition
(wet)
(dry)
Surface water inflow

Biological fixation
+
A
2.2
2.0
1.2
2.0
2.8
2.0
1.0
2.0
Groundwater outflow
Sum total (mean of A and B)
Net input
P( 10)
A
0.7
7.6
Above-ground vegetation
0.7
6.9
7.6
+
A
0.1
0.5
0.2 1.8
0.5
0.3
7.1 5.2
0.8
1.3
0.7
1.2
-0.5
4.3
+
A
1.2
8.7 6.0
7.2
-0.1
7.3
7.2
A: estimated for 1980-'83, B: for 1984
may relocate considerable amounts of the essential nutrient elements in the root system in
autumn, thereby reducing the discharge of these elements when the harvest falls in the late
autumn or winter.
10.6 Changes in De Stobbenribben and their possible causes
There is no clear record of change in De Stobbenribben, but a certain degree of atmotrophication is indicated, and a similar process in the vegetation cover has been observed very clearly
in De Bollemaat (Molenaar et al. 1990) and De Wobberibben (G.J.M. Ruitenburg, unpublished,
Calis & van Wetten 1983), two other quagfen complexes in North-West Overijssel. I will shortly
review some possible causes of a changing base-state in De Stobbenribben.
It is apparent from Table 10.3 (Section 10.3) that the chloride concentrations in De
Stobbenribben have varied substantially. There are some still older, additional observations. The
present author collected eight water samples in De Stobbenribben in August, 1969. Upon
analysis by the Hugo de Vries Laboratory the chloride concentrations appeared to vary between
91 and 99 mg/l. Coesel-Wouda (1967) and Stegeman (1968) reported analyses of water samples
collected in De Stobbenribben between August and October 1966. The first-mentioned
authoress sampled small pools at quite some distance from the ditch. Chloride concentrations
ranged from 67.5 to 125 mg/l. Stegeman sampled from between moss vegetation and found
18-64 mg/l chloride. In RIN files I found some other water analyses pertaining to De
Stobbenribben and collected between 1961 and 1968. These fall within the same range found for
the other analyses. Leentvaar (1960) in 1960 found chloride concentrations up to 114 mg/l in
open surface water of a petgat nearby De Stobbenribben.
All this suggests that the surface water has, at least temporarily, been slightly brackish in the
past. From the 1980-'83 and 1984 analytical results it is clear that a strong decline of the chloride
concentrations has taken place in this segment of the quagfen. While brackish influences are by
no means necessary for the development of a base-rich type of quagfen, such influences
definitely may faciltate the process. The case history for De Stobbenribben in this respect fits in
197
with the ongoing process of areal expansion of Scorpidium scorpioides and some other species,
as documented in Chapter 6.
Several possible causes for the observed changes can be mentioned. A further quantitative
treatment with models simulating the relevant processes and situations is necessary, however, to
check the relevant importance of each of these possible causes, and I will, for this reason, just
sum up some of the most obvious ones:
1) The hydrological situation of the area changed dramatically in the mid 1950s, when the
polder Wetering-Oost, at a distance of only 250 m from De Stobbenribben, was reclaimed. This
must have started a strong downward discharge of water from the mire area and so gradually
removed most traces of the earlier brackish water. Note that such traces have been found in the
groundwater system (Chapter 5) at a little distance from De Stobbenribben. More recently, a
variation of the water composition in the boezem system, as also reported in Chapter 5, also
involves the ditch feeding De Stobbenribben.
2) De Stobbenribben was probably dredged for peat by the end of the last century. Personal
observations and older descriptions suggest that the kragge has grown definitely firmer in De
Stobbenribben since 1960. Especially the number of weak spots and pools in the vegetational
zone B in parcel A has decreased. According to the QUAGSOLVE model these processes,
whatever be their causes, must have led to a decreased influence of the boezem water.
3) Irregular patterns in the time series may have been produced by the varying state of ditch
management and an alternating incidence of ditch blocking. In the autumn of 1972 the ditches
feeding the wider area of which De Stobbenribben is a part were separated from the boezem
system in order to keep polluted and eutrophicated surface water out of this area which was then
supposed, by the management authorities, to receive discharging groundwater from underneath.
This management experiment was abolished when the water level had fallen by more than 20
cm (written comm. J. ter Hoeve). This incidence must have caused a relatively strong
momentary inflow of ditch water. Similar ditch blockings have been frequently deliberately
apllied since then in order to improve the accessibility of the terrain during mowing. Depending
on the weather during such drainage periods (lasting 2-4 weeks) the influence on the water
composition will not always have been the same.
4) As mentioned in Chapter 5, the management of the boezem water level has also changed
during the 1970s and 1980s. Especially the sudden raising of the water level by 10-15 cm on the
first of April, and a corresponding lowering in October, was replaced by the maintenance of a
more constant level. At present water conservation strategies are applied which allow the water
level to fall during the summer until a critical level is reached at ca 10 cm below the target level.
Although these effects have not been quantitatively analysed, QUAGSOLVE explains how they
could have contributed to a certain atmotrophication of De Stobbenribben. Especially the
situation in zone B of parcel A, where an extensive Typha angustifolia stand was replaced by an
at present species-rich quagfen vegetation raises doubts as regards the future of this area.
Observations concerning the growth of Sphagnum vegetation and the changing water
composition in the preferential flow channel suggest a continuing process which would
ultimately lead to the disappearance of species depending on environmental conditions
maintained through the local, but not necessarily constant, influence of a lithotrophic type of
water. Although this would suggest a disturbance of the environment, it may well be considered
a final stage in a continued process, starting from an open water system with slightly brackish
water. The first, desalting phase coincides with the terrestrialization, as is still seen in other parts
of De Weerribben (see Chapter 6). The apparent stability of the base-rich conditions during this
198
phase originally seemed to confirm the theory of discharging groundwater. The present
investigation has proved that in fact no discharging groundwater, but rather an inflow of
lithotrophic surface water recharged the quagfen system with bases. The application of
QUAGSOLVE shows that the remarkably zoned pattern in the area results from the directional
supply of ditch water, governed by the particular lay-out of the parcels with one closed end at a
length which suffices to exhaust the inflow. The zoned pattern is reinforced by the succession of
the local vegetation cover, including structural changes in the kragge compartment and a
decreasing exchange rate T between it and the underlying flow channel. When this exchange
rate has become low enough a strong extension of Sphagnum vegetation is possible and
accelerates the further atmotrophication. Recently ditch cleaning and influences exerted on the
kragge by mowing and harvesting machines and by groups of volunteers seem to have partially
counteracted these processes.
199
CHAPTER 11
Summary and general discussion
11.1 Introduction
Two main types of mires are usually distinguished. Bogs are rainwater-fed and
hence poor in bases and nutrients. Fens, in contrast, are fed by groundwater and
surface water, and richer in nutrients and, especially, bases. They derive their
higher base and nutrient states from water seeping through the peat. Quagmires
form a less common category of mires usually developing as floating rafts over
bodies of open water and strongly yielding underfoot. In the moist climate of
North-Western Europe such rafts usually develop into quaking bogs, unless
hydrological conditions favour a continued inflow of water from elsewhere. Such
hydrological conditions apply to the quagfens in North-West Overijssel discussed
here.
The distribution of quagfens in North-Western Europe is not very accurately
known. The occurrence of extensive quagfens has been reported for fluvial plains
with riverine environments and lakes, such as found in Norfolk, North-West
Overijssel, the Danube-delta, and the Baltic Coast. Quagfens are usually characterized by a dominance of species indicating a eutrophic and base-rich environment,
and they seem to be adapted to strong fluctuations of the water level as determined
by a fluvial regime. Quagfens along lake shores often include less eutrophic zones.
The latter also applies to the quagfens in Norfolk and in North-West-Overijssel
North-West Overijssel quagfen eco-hydrology
201
which are due to the terrestrialization of bodies of open water in old turbaries. In
such less eutrophic, but still base-rich zones slender sedge species and extensive
moss carpets determine the structure of the vegetation, rather than tall reeds.
Similar types of vegetation are presently known to occur in extensive non-floating
rich-fens and remains of such fens are still found in North-West Overijssel also.
While possibly stable in the absence of human influences at a colder climate, such
less eutrophic rich-fen vegetation seems to have developed especially under a
mowing or grazing regime in our region (Ellenberg 1978).
Until the 1950's the relevant quagfens in North-West Overijssel were still highly
valued by local farmers for their fodder yield. Presently the interest is almost
entirely based on nature protection: substantial populations of many elsewhere
threatened organisms occur in these quagfens. The number and extent of the stands
of vegetation involved is strongly decreasing, however, not only in the Dutch
quagfens, but also in rich-fens in Germany (Braun 1968, Succow 1988), Poland
(Tomaszewska 1988), the Jura (Gallandat 1982), Czechoslovakia (Rybn ek 1974,
Baltov-Tul kov 1976), and Norfolk (Wheeler 1980a, Wheeler & Giller 1982).
Apart from the changed agricultural interests (decreased mowing and grazing of
natural fen vegetation, reclamation of mires), this seems to be due to a general
eutrophication of the environment. On the whole a loss of environmental and
biological diversity is reported, calling for a more detailed analysis of the hydroecological factors responsible for the occurrence of the species involved. The
relatively young quagfens in North-West Overijssel offer a suitable case for such
an analysis since inherent influences of old peat and historical plant populations
are less important here than in non-floating rich-fens with a long developmental
history. The existing seepage hypothesis for Dutch quagfens is taken as a starting
point in the present report.
In this rsum the course of the overall inquiry will be reviewed and some major
choices of approach mentioned, to be compared with data not already referred to in
the text. Note that frequent reference is made to key figures placed in other
chapters.
11.2 The seepage hypothesis for Dutch quagfens re-formulated
Seepage
Chapter 2 reviews the hypothesis that many plant species and their associations in quagmires of
the rich-fen type (quagfens) in The Netherlands depend on seepage. The hypothesis is gleaned
from the original publications (Meijer & De Wit 1955, Kuiper & Kuiper 1958, Segal 1966, and
others), and the arguments are checked with the pertaining international literature.
By the seepage or the intensity of water exchange in a mire Thomson & Ingram (in Ivanov
1981) understand the total quantity of water flowing per unit of time through a volume of peat 1
m2 in area and equal in height to the depth of the peat deposit in that part of the mire massif.
Hydrological definitions more specifically emphasize the exudation of groundwater or, even
narrower, such an exudation under the influence of a larger hydraulic head outside the seepage
area (CHO 1986).
The hypothesis of seepage in Dutch quagfens states that the occurrence of certain indicated
species (Table 11.1) is bound to an artesian discharge of groundwater. Many of these seepage
indicators were said to be characteristic of the phytosociological units Scorpidio-Caricetum
202
diandrae and Scorpidio-Utricularietum. The supposed discharge was never measured, however,
and various opinions are held as regards its bearing on the immediate environment of the plants.
North-West Overijssel was frequently mentioned as a seepage mire area par excellence (Segal
1966, Westhoff et al. 1971, Gonggrijp et al. 1981). The seepage hypothesis for quagfens was
widely accepted and various Dutch ecologists embraced the hypothesis and associated the
occurrence of the plant species and associations involved with an artesian discharge of
groundwater (Weeda et al.1988, Touw & Rubers 1989).
In spite of the missing evidence the seepage hypothesis cannot be rejected a priori. The
following questions arise:
1) Is there indeed an artesian discharge present at the classic locations? This is a somewhat
formal question of especially local importance, and subordinate to the second problem:
2) What ecological principle explains the rare and threatened occurrence of the species and
communities involved? Although the research is carried out in The Netherlands, the
findings may apply elsewhere if the effects of a different climate and other geographical
factors can be accounted for.
As regards this second question, the working hypothesis is formulated that the sites under
discussion derive their very characteristic cover of vegetation from the interaction of the
substratum, the atmospheric precipitation, and through-flowing water, as modified by the
microrelief. This re-formulation thus uses Thomson & Ingram's wider definition of seepage
rather than a more specific hydrological one, while it is more restrictive in adding a
(geographical) dependence on the climate and a dependence on the substrate. In the case of
mires, substrate characteristics depend on the stage of development in time.
Base state as a nodal parameter
Any further work starting from this hypothesis obviously requires a more precise specification
of the interactions mentioned. As suggested by the common elements in the reviewed treatises,
the base state was chosen as a major parameter for the characterization of the influence of these
interactions on the mire. In spite of the absence, in this investigation, of facilities for a direct
measurement of the base state of the peat, in this way the problem can be bisected. Rather than
considering the complex relation
(1)
vegetation = f(climate, substratum, seepage)
in one go, I chose the base state of the mire as a nodal parameter simultaneously serving for the
evaluation of the environmental interactions:
(1a)
base state = g(climate, substratum, seepage),
and to determine their ecological significance:

(1b)
vegetation = u(base state).
This construction facilitates a correction of estimates, both for errors and for deviations from the
ceteris paribus principle under which the relations found will hold true. The solution of the
seepage problem was probably retarded by the fact that such a step had not been taken earlier.
203
This point was envisaged by earlier authors, however, and I must credit S. Segal who charged
me, as a student, with the question what bearing seepage could have upon seepage indicators.
The nodal parameter can also be used in applications involving the scanning of impacts of
proposed measures concerning the water management, where the effects are evaluated according
to the expected variation in the nodal parameter and to an agreed relation between the latter and
some vegetational criterion fitting the field of application (e.g., nature protection).
In the present treatment, the term seepage is used for percolation in general, regardless of
the possible causes and any preferential direction of water movement. The word is italicized in
the designation of the seepage hypothesis for quagfens and in associated expressions. Although
the primary authors reporting seepage in Dutch quagfens specifically envisaged a regional
groundwater discharge, the non-specific term seepage is maintained here in view of the ecohydrological conditions disclosed in this investigation. In addition to a seepage of water,
seepage sites (italicized) apparently have some more specific ecological character in common,
which possibly relates to their base state.
11.3 Relations between environment and vegetation
Chapter 3 deals with an anticausal approach (Waldhauer 1982) according to formula (1b)
above: what information about the environment, especially its base state, can be borrowed from
the supposedly depending plants? The seepage indicators were compared with the majority of
other species reported for the classical seepage theory locations and for non-seepage mire sites.
The pertaining literature was searched for indications of the favoured environment attributed to
these species. The interpretation required a rough model of the spatial organization of the mire
environment and its variability. The first part of this chapter therefore concerns a reconciliation
of a pragmatic approach with ecological theory.
Recurrent patterns of heterogeneity
Most applied research in this direction relies on one of the basic concepts of the ZrichMontpellier school of phytosociology, i.e., the concept of species association in an ecologically
uniform habitat resulting from overlapping ecological amplitudes of the species concerned
(Shimwell 1971, Ellenberg 1978, Kltzli 1981). Rather than exclusively accepting this logical
AND-concept, I will consider the co-occurrence of species primarily as a result of ecological
heterogeneity of a site providing the union of the possibly different requirements of the species
present. This principle is accordingly called OR-association (Fig.3.1, Van Wirdum 1986, 1987).
In a more precise theory the season, frequency and duration of the presence of each microenvironment, and the life history phases of the species involved should be taken into account
(e.g., Connell & Slatyer 1977, Grubb 1977, Huston 1979).
In the present investigation the presence of species is considered as solely depending on the
environment, including other species, according to some determined biological program (Van
Wirdum 1982), and it is assumed that the set of available species is constant. A corollary of
this assumption is that the frequency distribution of species within a stand of vegetation
statistically reflects the frequency distribution in space and time of suitable micro-environments.
The recurrent co-occurrence of species must so be explained by non-random patterns in the
environment for the OR-concept to work, rather than by direct species interactions. This lines up
with recent approaches of succession theory (Grime 1979, Tilman 1982), but it differs from the
position chosen by ecologists studying the evolution of species as a result of abiotic processes
and competition (see Grubb 1977, Den Boer 1986). At the time scale of years it is assumed that
204
sites including their stands of vegetation (ecodevices) can be considered to represent a nearsteady state.
Recurrent patterns of heterogeneity in the environment have long been recognized in
northern European schools of vegetation and mire ecology. Since the relatively young (ca
70-100 years old) quagfens in The Netherlands, developing within the confined space of
abandoned turbaries, exhibit a quite different pattern of heterogeneity than large and old mires
do, the empirical rules of species association developed in northern Europe do not necessarily
apply here. Somewhat intuitively leaning against different life-history characteristics and
ecological strategies of species, such as reflected by their root systems and phenology, it is
argued that the terrestrializing petgaten with their developing quagfen typically provide a pattern
of heterogeneity suitable for the co-occurrence of the following synusiae (Fig.3.2):
1) Kragge synusiae, often 103 to well over 105 m2 in extent, and consisting of few plant
species with extensive rhizomes, exploring the nutrient resources underneath the main body of
the floating mat (kragge). Theoretically, the individual shoots can be considered to be the
deciduous parts of few long-lived and extremely extensive individuals. Managerial measures can
cause a dramatic reversal of dominance relations;
2) The synusiae adapted to the scale of hydro-environmental zones within single quagfen
parcels. The distinction of hydro-environmental zones builds upon lateral inflow of water into
many quagfens. The increasing density of the floating quagfen raft apparently gives rise to the
development of a more isolated, rainwater-fed top-layer, deepening where the lateral inflow
fades out. In case of an artesian water inflow, such a spatial pattern is replaced by a presumably
slow process related to the accrual of peat, and displaying the relevant synusiae as hydroenvironmental phases rather than zones. These synusiae consist of species rooting in the main
body of the kragge, especially in the upper 2-4 dm. In the quagfens with any substantial lateral
flow, the extent is of the order of magnitude of 0.1-2 103 m2. The species involved are especially
the slender species of Carex, herbaceous species, and amblystegiaceous and sphagnaceous
mosses;
3) Hummock-and-hollow synusiae, thriving upon the substratum provided by hummockbuilding species, and also upon deformations of the quagfen surface resulting from faunal
activity and management. The typical extent of hummocks and hollows in quagfens is 0.1-1 m2.
Several of the species concerned are short-lived and their abundance may vary from year to
year;
4) A micro-zonation of synusiae of seedlings, certain mosses, hepatics (such as species of
Riccardia, Pellia, Cephalozia, Cephaloziella, and Calypogeia) and blue-green algae, is found at
the cm2 scale. These synusiae often develop around decaying parts of large helophytes, at their
stem bases, or in hummocks dying back from their centre.
This distinction means that the environmental conditions to which the involved species respond
are to be measured in different compartments of the quagfen ecosystems, and at different spatial
scales. The emphasis is laid on the second and third groups. The first group is also considered
since it provides distinct light and litter conditions for the settlement and growth of species of the
other groups. A preliminary indication of the synusial group to which the seepage indicators
belong is given in Table 11.1.
205
Table 11.1 Seepage indicators according to Dutch authors and the state of their distribution in
the study area De Weerribben
Species
state in De Weerribben
synusiae
Carex buxbaumii
Carex diandra
Carex lasiocarpa
Eriophorum gracile
Liparis loeselii
Parnassia palustris
Sagina nodosa
rare
uncertain
abundant
abundant
insufficiently known
uncertain
distinct local distribution
rare
2
2
2
2
2/3
2/3
3
1
2/3
3
3
Campylium elodes
Campylium stellatum
Philonotis marchica
Riccardia multifida
Aneura pinguis
Scorpidium cossoni
abundant
abundant
abundant
outlying species
abundant
absent
abundant
abundant
rare
2/3
3/4
2
3
3
3
4
3/4
3
2/3
The synusial groups are explained in Section 11.3
Indication by species
The various traditions in vegetation ecology have led to different attempts to use plant species as
indicators of their environment. I have tried to use the empirical information included in the
emerged systems of indication, thereby considering that the OR-concept possibly requires a
different interpretation. In the central European phytosociological tradition the AND-concept
prescribes that the syntaxon to which a stand of vegetation belongs is more indicative than single
species. Only character species of the syntaxon are held to have a narrow ecological amplitude,
almost coinciding with that of the syntaxon as a whole. Compiling a list of various indications
per species for the study area, I assigned each species to a preferred syntaxon which I defined
wider as it was less specific for quagfen vegetation. As far as species of the synusial groups 3
and 4 are concerned, an interpretation according to the OR-concept implies that these species
may respond to different micro-environments apparently frequently associated in the particular
macro-environment specified for the syntaxon. A corollary of this interpretation holds that such
macro-environments favour, or at least tolerate, the development of the different microenvironments. As a result, conflicts arising from the combined application of phytosociological
indication and ecological indication are accepted as a possible indication of a heterogeneity of
206
micro-environments within seemingly homogeneous sites. It follows from the distinction of

synusial groups that various plants may themselves change the environment, provided that their
required settling conditions are met.
Ellenberg (1974, 1978) summarized the indicated nutrient, base, and soil-moisture state for
most vascular plant species occurring in Central Europe according to a multidimensional ordinal
system. I have been advised that such lists are insufficiently accurate since the classes are wide
and the lists do hardly account for the varying widths of ecological amplitudes, and it has been
claimed that the situation should be improved by drafting lists of local validity only. I am
opposed to these arguments, since I do not believe that our concepts of ecological relations, and
the available survey methods, do provide facilities to improve precision without further reducing
reliability. I am rather concerned about the fact that, where many species can be scored, the
statistical evidence of a multitude of somewhat extremely expressed preferences provides a
clearer picture than few precise indications plus many broad or indifferent indications. For this
reason I assigned each species, including mosses and hepatics, to an extreme class, keeping the
middle category for species that really do not seem to show any specific preference. Care was
taken to base this on the pertaining literature and not on my personal and partially local
experience, since that would clash with the independent character of the list. Due to the
relatively broad concept of the environment in this indicative system, the indication will in some
cases apply to the micro-environment, and in other cases to the macro-environment alone. The
resulting list is presented in Appendix C.
Next to these entries reflecting the central European tradition, the complex system of Finnish
indications of the nutritional relations of mire sites (Eurola et al. 1984) was analysed and used in
a slightly modified form. It was hoped that this system would enable a separation of micro- and
macro-environmental indications, but this could not be realized in account of the limited number
of species for which indications are given in the Finnish system and to the problems I met with
when applying the definitions to the specific quagfen situation.
Base state and quagfen vegetation
As might be expected from the basis of the original seepage hypothesis, a strong correlation
appeared to exist between unanimously supposed seepage indicators as recognized by Dutch
authors and certain phytosociological groups, especially the Caricion davallianae. The seepage
indicators differ from other species by their strong association with indications of a low nutrient
and high base state, and a preferential occurrence in terrain depressions (Fig.3.6). More weakly,
the Finnish system revealed an association with inherent influences of the local peat type, which
I interpret as a greater capacity of the peat to hold and exchange physiologically important
elements. The vegetation of seepage sites includes several species indicating more acid and
more eutrophic conditions, however, and this possibly reflects a micro-environmental
heterogeneity promoted by a wet macro-environment with a high base state. This can be
understood from the fact that, in the Dutch climate, characterized by an annual precipitation
surplus, slightly elevated micro-sites will be prone to acidification unless base-rich water
frequently floods them. Under these conditions a seepage of base-rich water promotes site
heterogeneity. In the absence of base-rich water even from the lower terrain parts, the area of
acidification will ultimately comprise the whole macro-environment. Of course the attribution of
species to extreme categories in the indicators system could have led to a separation of
ecologically related species, so artificially suggesting OR-association where they grow side-byside. A check with the original Ellenberg indicatory values in relevant cases confirmed real ORassociation by a range of several units on his scales for the base and nutrient states, however
207
(cfVan Wirdum 1986). Apparently most species scored on either side of the middle categories in
my system show a distinctly different ecological behaviour.
This analysis thus justifies the choice of the base state as a nodal parameter: there is no other
parameter in the accumulated empirical evidence for which especially the seepage indicators are
so unequivocally sensitive. While I can not claim this as a new idea -the indications were all
derived from the pertaining literature- the corollary that supposed seepage vegetation in
permanently wet quagfens could well rely upon any mechanism providing a high base state for a
sufficiently long spell, brought some clarity in the myriad of existing causal explanations of the
influence of seepage, and it opened a domain of more specific hypotheses, especially those
regarding the possible influence of flooding and surface water inflow underneath the kragge.
This has resulted in the development of some specific methods for field survey and data
interpretation, treated in other chapters and in Appendix D.
11.4 The study area
Chapter 4 provides a summary description of the study area North-West Overijssel. It is mainly
based on literature and on summarizing publications by others. The mire area was formed on
sand infillings in the deeply eroded valley of a Pleistocene melt-water river, the Oer-Vecht. Ter
Wee's (1962, 1966) geological study of this valley, in addition to a thorough analysis of the
descriptions of geological borings and one additional boring, sufficed to debunk the once, in
circles of ecologists, prevailing idea that the mire area of De Weerribben in North-West
Overijssel were underlain by an aquiclude at a depth of some 15 m. In fact the underlying strata
mainly consist of transmissive sands to a depth of more than 100 m. These results have been
confirmed by later borings (Geologische Dienst 1979, Ter Wee pers.comm.). This sand bed
could transmit a groundwater flow from the higher morainic area to the East into the mire area if
the hydraulic head differences between these areas would allow this.
The development of the original mire, as opposed to the recently terrestrializing turbaries,
was described in detail by Veenenbos (1950) and Haans & Hamming (1962). It appears that the
distribution of original mire types within the larger area of North-West Overijssel was strongly
determined by the courses of a deltaic river system in an undulating plain. The central part of the
area was formed by an extensive bog, with some draining streams contributing to river branches
flowing through from the Pleistocene area to the east. Interestingly, a type of peat most similar
in botanical composition to the recent quagfens was then formed especially in the zones of
presumed incidental flooding, thus providing a similarity with the situation as it has been
described from Poland (Pa zy ski 1984) and East Anglia (Godwin 1978).
Man has strongly interfered with the development of the mire since the early Middle Ages
and it is believed that the influence of incursions from the then developing Zuyderzee is
interwoven with this historical development. These incursions, and the flooding of the western
part of the area introduced brackish influences into the area, which could temporarily increase
when the areal extent of open water had grown as a result of peat dredging. A large flood from
the then brackish Zuyderzee destroying whole villages in the area occurred in 1825. Many more
recent floodings before 1932 (enclosure of the Zuyderzee), and once during war conditions
(1944), were possibly primarily due to the ponding up of freshwater at high sea levels.
The recent state of the area is especially tied up with the intensive management by the water
authority and with social and demographic developments since the beginning of this century.
The reclamation of some large polders in the 1928-1960 period, with a phreatic level maintained
at ca 2-4.5 m below the boezem water level has been of great importance to the hydrological
situation (Fig.4.3). It is more or less by coincidence that such a large area of terrestrializing
208
ponds and broads still existed in a semi-natural state here when the interests of nature protection,
environment and landscape were politically recognized.
11.5 Hydrology of De Weerribben
Chapter 5 treats the hydrology of De Weerribben. The main question concerns the possible
incidence of discharging groundwater in this area, now and in the past five decades. It appears
that no groundwater outflow of any importance is found at present: on the contrary, De
Weerribben is an area of considerable groundwater recharge (Fig.5.3). The chemical
composition of the underlying body of groundwater is largely determined by infiltrating surface
water. This present infiltration of water includes the location of a classical seepage site, De
Stobbenribben. Here the recharge must even be more substantial due to the proximity of a deep
polder reclaimed ca 1955.
Interpretation of water analyses
Next to the measurement of water levels and the hydraulic head according to traditional
methods, this part of the study included an extensive use of water composition parameters to
trace water flow. Chloride concentrations have often been used as a means to investigate flow.
Since influences of sea water, rain water, and groundwater were simultaneously expected to
occur in the study area, a simple model for the mixing of groundwater, rain water and sea water
was developed. Since three sources of water were concerned, it was necessary to add at least one
parameter, which is derived from the base concentrations. The transformation of rainwater into
calcareous groundwater is empirically included. The development of the relevant model and the
associated computation of water composition parameters is treated and compared with existing
graphical and statistical methods in Appendix D. The simplest of these procedures (Fig.D.9)
consists of the construction of EC-IR diagrams, where EC is the electrical conductivity at 25 C,
plotted logarithmically, and IR is the Ionic Ratio defined as:
IR = [1/2Ca2+] / {[1/2Ca2+]+[Cl-]} (concentrations in moles of charge per litre).
A very similar procedure appears to have been developed earlier by Gibbs (1970) in order to
understand river water compositions at a global scale. Before returning to the application in
North-West Overijssel, it suffices here to recall some terms introduced in the present study:
lithotrophic water: a calcium-bicarbonate type of water, usually owing its characteristic
composition to a contact with soil;
atmotrophic water: a type of water with low concentrations of most constituents, usually
owing its characteristic composition to atmospheric precipitation;
thalassotrophic water: a saline sodium-chloride type of water as found in the oceans;
molunotrophic water: polluted water as presently found in the Rhine.
Waters forming a series between two of these types, or a wide cluster obviously derived from
one type are denoted as clines, e.g., an atmo-lithocline, a molunocline, a litho-thalassocline.
Benchmark samples for the mentioned water types are given in Appendix D. They can be
plotted in the diagrams, together with curves describing their mixing ratios (Figs.D.9, D.12).
209
Groundwater
In 1937 water was sampled at 10-15 and 20-30 m below the soil surface in a transect of 5 sets of
piezometers from the morainic area through De Weerribben to the former Zuyderzee. The
analyses reveal an atmo-lithocline and a litho-thalassocline, respectively (Fig.5.5). Interestingly
the records of the groundwater composition in the central part of De Weerribben show slightly
brackish influences at 12 and 16 m below the mire surface, but not at the depth of 29 m
(Fig.5.4). Any upward flow of groundwater before the reclamation of the deep NoordoostPolder in 1941 had apparently not annihilated these traces of brackish influence. Differences
between the surface water level and the hydraulic head in deeper layers were very small in 1936'40 and varying conditions may have caused some local recharge or discharge. When
discharging in the surface water system calcite should precipitate from the groundwater found in
the mire area in 1937. An indirect clue for the absence of any substantial exfiltration of this
water therefore comes from the absence of calcareous deposits. This result was somewhat
surprising, since it was expected that lithotrophic water, either originating from groundwater
discharge, or from an inflow of fresh surface water from the morainic area, had been the
predominant type in the mire basin in the beginning of the 20th century. In the pertaining Dutch
literature, seepage indicators are usually assumed to avoid even slight thalassotrophic
influences.
The reclamation of various polders caused a dramatic fall of the hydraulic head in the body
of groundwater, resulting in a conspicuous infiltration of molunotrophic surface water to a
considerable depth. In 1980 an aged lithotrophic type of groundwater was only found at one
location at a depth of 50 m below the mire surface. An isotope analysis suggests that this water
infiltrated during the growth phase of the virgin mire, in the Subboreal period. Local pockets of
very slightly brackish groundwater were still found at a depth of 10-30 m. The raised chloride
contents (169-540 mg/l, compared to 18-74 mg/l in lithotrophic and molunotrophic samples)
may be due to flooding with brackish water in various periods. One sample, originating from a
depth of 29 m, was analysed for tritium and 14C, and its possible age accordingly estimated at
2000 years. The slightly brackish pockets of groundwater are considered remains of a more
extensive body of such water now largely washed away by the strong infiltration of surface
water. On the whole, the chemical variation among samples analysed in the 1980s is
substantially smaller than that among the 1937 samples.
High tritium contents in most recent groundwater samples and a consideration of hydraulic
head gradients led to an estimation of the rate of exchange between the body of groundwater and
the overlying mire area, and of the corresponding upward and downward movement of
groundwater for three periods, assuming a porosity of 30% in the sand bed:
1889-1920 exfiltration 0.2 mm/d; upward movement 6 m in 30 years;
1920-1941 exfiltration 0.3 mm/d; upward movement 6 m in 20 years;
1941-1980 infiltration >1 mm/d; downward movement >45 m in 40 years.
These very rough estimates show that the earlier influence of slightly brackish water may indeed
have long persisted before the deep polders were reclaimed. The reclamation of deep polders
had a large impact on the groundwater flow system. The groundwater survey does not support
the a priori attribution of a lithotrophic character to the mire area at present or in the past, but it
210
anyhow emphasizes a substantial presence of bases: no traces of atmotrophic groundwater were

found underneath the mire area.
Surface water
The surface water in the main canals in the mire area was repeatedly sampled and analysed
between 1972 and 1982. A reasonably large number of analyses covering the 1960-'72 period,
and the ongoing monitoring program of the water quality authority Zuiveringschap WestOverijssel enabled an analysis of the surface water composition over a period of more than 25
years.
A lithotrophic character of the surface water was obvious in the 1960-'70 samples. This is
attributed to the influence of an inflow of surface water from ditches, canals, and former rivulets
draining the morainic area. At some isolated places Leentvaar (1960) found raised chloride
concentrations, and there is some evidence that these were more abundant before 1960. After
1972 the influence of the controlled supply of polluted surface water from the adjacent Frisian
area has become a dominant feature (Fig.5.14). The lithotrophic type of water is now solely
associated with the inflow of water from the Steenwijk-Ossenzijl canal, which serves as a
catchwater for the discharge from the morainic area. The lithotrophic character of the water
entering the mire basin from the morainic area is obviously not associated with the infiltrating
groundwater in the elevated parts of this area, since that appeared to be of an atmotrophic type in
1937, and it is presently polluted-atmotrophic.
A statistical analysis of time series at various places in the canals network within De
Weerribben revealed different spheres of influence of the various sources of water (Fig.5.13). It
appeared that variations in the operation of surface water supply and discharge facilities are of
paramount importance to the distribution of these influences within the mire area. A precise
explanation was not possible, however, due to the complex and changing nature of the canals
network, to wind-generated circular currents, and to interactions with the mire in the meshes of
the canals network.
The changing eco-hydrological state
The long-term variations of surface water composition stress the shortcomings of a spatial
comparison on the basis of data that cover only a couple of years. During the development of the
present quagfen vegetation a hydrologically steady environment, if any, was probably never
present for longer than some ten or twenty years. Base-rich seems to apply, but that in itself is
not a feature unique enough to explain the predisposition of this area for the development of
seepage sites.
A comparison between a new and an old steady state is not realistic either. Yet, the
decreased availability of unpolluted lithotrophic water in the surface water system since about
1970 must be considered a significant change which became dramatically evident by an
increased water supply during the extreme droughts of 1975-'76 (Fig.5.15). Due to subsequent
changes in the execution of water management and to more favourable weather conditions the
1980s have brought a gradual return of lithotrophic, low-chloride surface water. Recent
observations in De Weerribben (1989) have shown that aquatic macrophytes, such as Stratiotes
aloides and species of Potamogeton, once considered to have disappeared, also returned.
211
Fig.11.1
EC-IR (a,b) and rTH-rLI (c,d) diagrams demonstrating a natural range of mire
water composition
Diagrams a, c: Mire types according to Moore & Bellamy (1973), ranging from extreme rich-fen (1) to
bog (7), and analyses pertaining to the Norfolk Broads (Giller 1982): a rivers and broads; b
Peucedano-Phragmitetum caricetosum with seepage indicators; c
Peucedano-Phragmitetum
schoenetosum with salt indicators; d other sites;
Diagrams b, d: Mire sites in Mecklenburg (Succow et al. unpublished): e rivers, ditches; f
Drepanoclado-Caricetum diandrae-lasiocarpae with seepage indicators; g other sites.
L,A,T,M: Litho-, atmo-, thalasso-, and molunotrophic benchmark samples (Appendix D); mixing
between these samples is indicated with contours in the EC-IR diagrams
It remains to be demonstrated whether the vegetation reported for the 1930-'60 period was
also determined by a decrease of brackish influences to be replaced by an inflow of lithotrophic
surface water. At this point it is of some interest to review the international literature for a
comparison of the water composition at sites of seepage indicators. Various such analyses have
been collected in Fig.11.1. Although the general pattern seems to confirm that (extreme) rich fen
sites, the preferential habitat for seepage indicators, are at the lithotrophic side of a lithoatmocline (Fig.11.1a,c), data from the Norfolk Broads (Giller 1982, Giller & Wheeler 1986, see
Fig.11.1a,c) demonstrate a thalasso-lithocline. According to Giller the sites involved probably
do not receive discharging groundwater, but they have been influenced by flooding with
212
brackish water in the past, and the underlying clays and peats are obviously brackish. Giller
suggests that the sites are influenced by a diffusion of salts from these clays into the overlying
mire. Unpublished data from sites near the Baltic in Mecklenburg (Eastern Germany), collected
by the author and some colleagues under the guidance of M.Succow and L.Jeschke also show
slight thalassotrophic influences (Fig.11.1b,d).
A more detailed analysis of data reported for other Dutch mire areas (Van Wirdum 1987,
Van Wirdum et al. in prep.) has revealed that seepage indicators have been recorded in the
Nieuwkoopse Plassen area and in mires in the province of North-Holland. Presently these areas
exhibit a stage of acidification not yet reached in the greater part of the classic seepage areas
North-West Overijssel and Vechtplassen. There is no doubt that thalasso-lithotrophic gradients
existed and partially still exist in these mire areas, but a clear relation with the quagfen
vegetation discussed here has not been assessed. The point will be reconsidered in Section
11.11.
11.6 The distribution of seepage indicators in De Weerribben
Chapter 6 treats the distribution of the seepage indicators in De Weerribben in order to ascertain
whether this coincides with a supposed seepage area, or with any area possibly distinct for its
base state. As indicated in Table 11.1, eight seepage indicators are relatively abundant in De
Weerribben, and until now no distinct distributional pattern has been assessed for them. It has
appeared that these species prefer a distinct hydro-environmental zone within the quagfen
parcels where they occur, however: they are usually absent from zones immediately bordering
ditches and canals, and also from obviously atmotrophic zones at a greater distance from the
surface water channels. For eight other species the distributional information is not considered
since these species are rare, possibly absent, or easily overlooked. The five remaining seepage
indicators appear to have a distinct distributional pattern within De Weerribben as a whole.
They more-over show the same preference for an intermediate hydro-environmental zone as
reported for the eight abundant indicators. It is usually believed that their absence near the
surface water channels is due to too high a nutrient state at such places.
The distribution of the five seepage indicators with a distinct distributional pattern in the
whole area includes recently occupied sites and has no correlation with any predisposition for
groundwater discharge in the last five decades. The occurrence of Menyanthes trifoliata,
Utricularia intermedia, and Parnassia palustris is almost restricted to the inundation zone of a
disappeared rivulet. The local presence of old Carex, Phragmites, and Alnus peats and scattered
clay deposits indeed suggests a relatively high base state and certain other favourable factors
associated with the former land use. All five are almost absent from the central part of De
Weerribben, where the old peat remains consist of bog peat and clay was never deposited.
At present, Liparis loeselii and Scorpidium scorpioides are fairly common in the southwestern part of De Weerribben (Schut- en Grafkampen). A comparison of the data about the
distribution of Scorpidium scorpioides between 1969 and 1985, and the supposed seepage area
indicated by Kuiper & Kuiper (1958) suggests that this species has only recently expanded over
that part of the area, while it is decreasing in the north-eastern part of the nature reserve. It
appears that Scirpus maritimus, S. lacustris tabernaemontani, Ophioglossum vulgatum, and
Hippuris vulgaris, all indicating possibly brackish influences, have decreased in the same area
over which Scorpidium is presently expanding. Liparis loeselii was already present in that area
in 1967. The expansion area also coincides with a dominance of Salix cf. S.cinerea in carr
213
vegetation. In the decrease area Alnus glutinosa dominates in carr vegetation, and in the former
bog peat area, where seepage indicators are anyway rare, Betula pubescens does so. Although
this relation cannot be quantitatively confirmed with the available data, the following hypothesis
is raised:
Until ca 1940 Scorpidium scorpioides was probably restricted to the freshwater Carex peat area
almost coinciding with the supposed seepage area indicated by Kuiper & Kuiper (1958), and
representing the lithotrophic end of an atmo-lithocline developed in the old mire landscape. The
Schut- en Grafkampen area was probably still slightly brackish then, and in too early a stage of
succession for the seepage indicators to settle. The peat in this area was saturated with bases,
however, due to the brackish water. The central bog area had too low a base state for most of
these species. Between 1920 and 1955 the improvement of the water control in the area by the
installation of a large discharge pumping station, the enclosure of the former Zuyderzee, and the
reclamation of several deep polders led to an increased influence of lithotrophic water rich in
bases in the whole mire area, and so facilitated an expansion of the seepage indicators,
especially Liparis loeselii, over those parts of the area where the stage of succession was suitable
and where a high exchange capacity of the soil buffered the local base state. In the 1973-'78
period the expansion was temporarily stopped by a dramatically increased influx of
molunotrophic water. This influx was caused by a particular strategy to operate the inlet and
discharge stations of the mire area and by the extreme drought of 1975-'76. Since 1978 the water
quality became more lithotrophic again, and the increased interest in nature protection resulted
in an increase of the area of reeds mown in summer and autumn. These factors restarted the
expansion of seepage indicators, and the conditions apparently now became favourable for
Scorpidium scorpioides also. Utricularia intermedia and Parnassia palustris possibly follow in
this line, but the presently available data are not convincing. Their physiological and
regenerative capacities possibly require that the above-mentioned influences have continued for
a longer spell of time, as is suggested by the common, but not very revealing, notions that they
require relatively mature sites or, according to the Finnish indicatory system, sites with a
distinct inherent nutrient supply. This hypothesis leads to a preliminary ranking of species as
regards their capability of quickly invading a new quagfen site, which can be put to the test by
future observations. The retreat of seepage indicators from the north-eastern part of the area is
explained by the fact that vegetation management is only partially continued here, and that
natural succession and a strongly decreased ditch management have led to a strong
atmotrophication of some traditional quagfen sites.
If the surface water composition plays such an important rle in preparing the site quality for
the settlement of seepage indicators, it is expected that certain aquatic macrophytes also respond
to it. Indeed the distributional behaviour of Stratiotes aloides between ca 1940 and 1989 seems
to support the hypothesis. This species formed increasingly dense stands until it almost
disappeared from De Weerribben in the early 1970s. A slow come-back since ca 1978 has been
noticed to accelerate in the mid-1980s, and in 1989 several extensive and healthy stands could
be found.
The hydrological investigation and the study of the distribution of seepage indicators in De
Weerribben seem sufficient to reject the original seepage hypothesis of quagfens as far as an
artesian discharge of groundwater is considered a prerequisite condition. Although they
contribute to some specification along the lines of the revised hypothesis (Section 11.2), the
conditions indicated are broader than was expected.
214
11.7 The quagfens of De Stobbenribben and their vegetation

Chapters 7-10 of this thesis deal with a case study in De Stobbenribben, one of the few
complexes of quagfen mentioned as a representative site by the authors of the seepage theory.
De Stobbenribben was chosen for this survey for still another reason: the quagfen parcels are
isolated from the surface water system at three sides, only one narrow end providing an open
connexion with a ditch. This topography provides an excellent opportunity to observe a linear
gradient as a result of a possible lateral inflow of water from the ditch (Fig.7.2).
The cover of vegetation indeed exhibits a linear gradient. Somewhat arbitrarily, four main
types of vegetation can be distinguished in hydro-environmental zones in De Stobbenribben
(Fig.7.4):
A:
B:
C:
D:
Calliergonella-Phragmites reed near the ditch;

Scorpidium-Carex fen in the central part of the parcels;
Sphagnum-Carex fen between the central parts and the far ends;
Calamagrostis canescens litter-fen at the far ends.
They can be characterized according to the indicatory information in the species list (Chapter
3, Appendix C). Seepage indicators are especially abundant in the zones B and C between the
more swampy and eutrophic ditch side and the litter-fen at the opposite side of the quagfen
parcels. Zone B is dominated by amblystegiaceous mosses and zone C by Sphagnum species in
the bryophyte layer. Zone C can thus be considered a successional stage due to the increased
influence of peat accumulation, rain water, and acidification. The micro-sites unique to zones B
and C are indicated as litho- and atmotrophic, respectively, and oligotrophic in both (Fig.7.6).
The length gradient in the cover of vegetation of De Stobbenribben was rather stable in the
1950-'90 period, although some remarkable changes were recorded in the Scorpidium-Carex
zone B. Following more frequent mowing an extensive bed of Cladium mariscus in one of the
parcels changed into a more typical representative of zone B vegetation. In another parcel
(parcel A investigated in considerable detail in Chapters 9 and 10) a similar change involved a
bed of Typha angustifolia and Scorpidium scorpioides, but this change, started between 1975
and 1979, continued to the effect that Juncus subnodulosus and hummock synusiae with
Sphagnum subnitens and S.flexuosum had become abundant in 1983-'84, thus indicating a
relatively fast transition towards the zone C type.
Mowing prevents a formation of alder and sallow carr in De Stobbenribben. At some
relatively dry fen sites ericaceous dwarfshrub vegetation with a moss cover of Sphagnum and
Polytrichum species and many birch seedlings has developed.
11.8 Peat temperature and the estimation of vertical water flow
Although the information reported in Chapter 5 was quite decisive as regards the absence of any
groundwater discharge into the Stobbenribben quagfens, the rate of vertical water flow still
remained to be assessed. A precise observation of hydraulic heads in the sand bed underlying the
Stobbenribben quagfens was hampered by difficulties associated with the installation of
piezometers in this sand bed, the top of which lies some 3-4 m below the quagfen surface. S.
Segal (pers.comm.) had observed differences in the temperature regimes between supposed
seepage pools and non-seepage pools, and so suggested that seepage could be quantified by
temperature measurements relevant to heat transport, and, simultaneously, that the temperature
regime could be a decisive ecological factor for seepage indicators. A method was designed to
215
effectuate temperature measurements and their interpretation, and to describe the temperature
regime in the quagfens.
It appeared that daily fluctuations of the temperature at shallow depths can not be held to be
very informative if the flow rate is less than some centimetres per day, as it is in De
Stobbenribben. An analysis of the annual temperature regime at depths below 0.4 m yielded
useful results, however. Two schemes were used for a quantitative interpretation, one of which
was based on Stallman (1965). They gave similar results in De Stobbenribben which could not
be rejected by reference to results of any other method and which were in accordance with
hydro-geological considerations (Table 8.6). In both methods it is assumed that no lateral heat
transport of any significance is present. This assumption is obviously violated at the observed
flow rates (up to 10 mm/d), since the water loss is compensated by a lateral influx of water from
the ditch at one end of the parcels. No accurate estimation of this disturbing effect could as yet
be made. Since the inflow must be most important in the very transmissive layer underneath the
kragge, as confirmed by an inspection of depth gradients in the temperature, it can be suppressed
by omitting values obtained in that layer, at the cost of some statistical accuracy. In my
implementation of the Stallman method I used a Fourier analysis so as to single out the effects of
non-annual temperature fluctuations. This probably also suppresses the effects of lateral flow.
All results indicate a downward seepage in the order of magnitude of some mm/d, but the spatial
variety was considerable.
The measuring method designed has appeared suitable for reconnaissance in various other
mire areas. The more detailed interpretation with the heat transportation models requires very
precise and, preferably, repeated measurements throughout a full year. Accordingly, that
application is attractive in more detailed investigations rather than in reconnaissance surveys.
The measurements and the heat transport models enabled an estimation of the duration of
temperatures in excess of some threshold value in the kragge. Surprisingly large differences
appeared to exist and demand future observations concerning the phenology of plant species in
the vegetational gradient (Fig.8.7). It should be emphasized, however, that, other factors being
the same, the actual temperature regime is determined by a combination of hydrological factors
and the structure of the local vegetation itself.
11.9 Lateral flow in longitudinal transects in De Stobbenribben
The study of the temperature and the heat transport in the Stobbenribben quagfens resulted in a
preliminary quantification of the vertical water exchange between the body of mire water and
the underlying sand bed. According to this quantification a considerable influx of ditch water is
necessary in order to compensate for water losses. In Chapter 9 the resulting longitudinal
gradient in the mire water composition and temperature is investigated in detail. Hydraulic head
differences in the horizontal direction were accurately measured with a water manometer.
Sounding rods with sensors for the temperature and the electrical conductivity were used to
assess the gradient and its changes through the seasons. The associated water composition was
determined with analyses of water samples taken at a selection of sites, depths, and dates, and
the similarity to litho-, atmo-, and thalassotrophic water was used as a basis for the ordination of
water samples. This method proved to be very useful for the survey of mires in general and it
was also applied elsewhere. The surprising vertical and horizontal differences found in the
present project (cf. Van Wirdum 1972, 1973, 1982,
1984) appear to be key features of many Dutch mires (see, among others, Grootjans 1985,
216
Wassen et al. 1989, Koerselman et al. 1990), and ongoing work in mires elsewhere seems to
confirm a wider applicability (see also Giller 1982, Giller & Wheeler 1986).
In De Stobbenribben a clear longitudinal gradient was demonstrated to exist (Fig.9.4). This
could be attributed to a different macro-ionic composition of the mire water rather than to
differences in the contents of nutrients. The gradient in the water composition can be described
as a gradient from litho-molunotrophic water in the ditch to atmotrophic water at a greater
distance, especially in the uppermost layer of the kragge (Fig.9.10). The discharge of mire water
towards the underlying body of groundwater in De Stobbenribben gives rise to the establishment
of a 0.001-0.002 hydraulic head gradient during the summer. Even in winter an inversion of the
hydraulic head gradient seldom continues for more than a few days. Ditch water penetrates the
quagfen through a preferential flow channel underneath the kragge and it exerts an obvious
influence upon the chemical composition of the water in the kragge even at a distance of more
than 100 m from the ditch. At the closed end of the quagfen parcels the influence of rain water is
predominant. The middle part of the longitudinal gradient displays strong fluctuations.
As regards De Stobbenribben this definitely proves that there exists a considerable
downward seepage from the mire into the underlying sandbed, and that this discharge generates
an inflow of base-rich ditch water. This inflow fades away in the closed ends of the quagfen
parcels, where seepage indicators become far less abundant in the cover of vegetation. It is not
yet clear what prevents the seepage indicators from occurring in the Calliergonella-Phragmites
vegetation near the ditch, although the vigorous stand suggests that this may be due to an
additional nutrient effect.
A general conclusion is that the layered and anisotropic nature of quagfen complexes as
porous media is fundamental to the understanding of the inflow of water.
11.10 Environmental and vegetational processes in De Stobbenribben
In Chapter 10 some elements are presented for the construction of a mechanistic model for the
understanding of the inflow of ditch water and its impact on the salinity and the base state in the
quagfen and the nutrient budget available to the vegetation.
First, the estimated flow rates in De Stobbenribben are checked and found in accordance
with the hydraulic conductivity in the preferential flow channel, ca 800 m/d. Near the ditch the
hydraulic conductivity is only half as large, probably due to the deposition of ditch dredge, a
more vigorous growth of plant roots, and a possible blocking of pores with micro-organisms and
peaty material. This check yields a justification for the range of the rate of downward discharge
of water from the quagfen found earlier. Precipitation and evapo-transpiration are estimated
from standard weather data and from an investigation with lysimeters (reported in Appendix E),
respectively.
The QUAGSOLVE model
A balance model QUAGSOLVE is constructed for the distribution of chloride in the quagfen.
This is a steady-state model. State transitions were not programmed for lack of suitable data
concerning the process kinetics. As regards the longitudinal relations in the flow channel this is
not regarded an important shortcoming. The vertical distribution of chloride ions between the
kragge and the underlying flow channel, however, appears to depend also on short-term water
exchange. The model provides for this by an exchange term. In excess of the precipitation
surplus, a downward flow T of water with the ambient chloride concentration in the kragge
balances an equal upward flow with the ambient concentration in the flow channel. In reality, T
217
is influenced by other factors too. A large value of T signifies that the kragge functions as an
integral part of the main flow channel. The model is calibrated with observed chloride
concentrations. An essential element is the fact that a boundary condition can be derived from
the situation in the dead end of the quagfen, which is not reached by the inflow of ditch water.
A fair simulation of the conditions observed in 1980-'84 is obtained with a downward seepage of
2 m/a (Fig.10.3), which corresponds surprisingly well with the results of heat transfer
calculations. The distinctly lower concentrations of chloride in the kragge can be explained by
assuming different exchange values T. Near the ditch this value is ca 1-1.2 m/a, decreasing to
0.6, 0.15 and 0.05 at distances of 55, 135, and 170 m from the ditch, respectively. This exchange
is caused by various processes, including the asynchronous incidence of evapo-transpiration and
precipitation, and a forced exchange due to pressures exerted on the kragge during mowing and
other incidences of access. The high values of T suggest the inclusion of lateral exchange and,
near the ditch, flooding. It seems that the quantitative effect of these factors on T depends on
such parameters as pool density, kragge weakness, and, if present, the extent of a Sphagnum
cover.
The distribution of most other constituents depends not only on the processes simulated in
the QUAGSOLVE model, but also on uptake and release processes intermediated by
vegetational and microbiological processes, and on an exchange with the peaty matrix. The
effect of such processes is assessed by comparing a faked QUAGSOLVE result with results of
water analyses. Since QUAGSOLVE essentially relies on the mixing of rain water and ditch
water, a faked QUAGSOLVE result is obtained by calculating the amounts of ditch water and
rain water needed to produce a mixture with the actual chloride concentration found in each
sample, and accordingly calculating an estimated concentration for the other constituents. Means
of 21 bi-monthly analyses in the 1980-'83 period were used for the comparison. Significant
differences were especially large for calcium (Table 10.5), which can be explained by the
known, large exchange capacity of fen peats for bases. Apparently the inflow of ditch water
increases the base state of the quagfen.
Concentrations of phosphorus and potassium in the quagfen do not differ significantly from
those found in the ditch water. The variation of their concentrations in atmospheric deposition
hampers an attribution of gains or losses of these constituents to any processes within the
quagfen. The total stocks in the peat and the stand of vegetation are large, but locally
mobilization and diffusion rates may limit the availability of these elements to vigorously
growing vegetation. If the greater productivity of the Calliergonella-Phragmites vegetation near
the ditch is to be attributed to a greater availability of phosphorus and potassium, then this must
be due to occasional floodings.
Inorganic nitrogen concentrations are decidedly lower underneath the kragge than in the
ditch. Within the flow channel, however, any significant loss appears to be restricted to the zone
bordering the ditch. This might be due to an uptake by the vegetation at a 1.5-2.5 g/m2 per year
rate from the inflowing water, so that it is possible that the nitrogen content of the surface water,
in combination with the high flow rate, supports an environment suitable for tall helophytes and
swamp species but not for Scorpidium-Carex fen.
The nutrient balance
With the data now available a balance can be drafted for phosphorus, potassium, and nitrogen,
and this can be compared to similar balances drafted for quagfens in another Dutch mire area by
Koerselman (1989). In addition, the amounts sequestered in the vegetation are estimated
218
Table 11.2 The nutrient balance of quagfens (kg per ha per year)
Stobbenribben
N
P
Out
Westbroek
Molenpolder
K
In Out
N
In Out
P
In Out
K
In Out
In
N
Out
In
P
Out
In
K
Out
In
Out
In
Precipitation
Wet
Dry
Fixation
25
(20)
(20)
0.3 -
10 - - -
24 18 13 -
0.7 -
6
-
26
18
2
0.6
-
6
-
Surface water
Groundwater
11
0
0
8
0.4 0.0
0.0 1.3
62 0
0 74
1 21
20 0
0.1 0.7
0.5 0.0
4
6
13
0
7
0
9
1
0.5
-
1.0
0.1
19
-
22
2
Harvest
57
44
38
3.9
32
Sum total
76
65
0.7 4.5
72 128
76 88
1.3 6.3
16
58
53
49
1.1
5.0
25
56
Mineralization
115
5.2 -
67 -
3.4 -
315 -
43.8
3.2
54
66
5.6
(Westbroek and Molenpolder: from Koerselman et al. in press)
on the basis of biomass assessments, and the release of nutrients in the kragge was determined
(Verhoeven et al. 1988). The comparison is made in Table 11.2. It appears that the external
inputs of nitrogen are roughly sufficient in all cases to support the production of biomass, but
phosphorus and potassium are probably released from the quagfen and discharged at harvest
time, provided that surface inputs have not been largely underestimated.
Mineralization has been included in the table for the sake of completeness. The error is
probably large, however, since in all cases it is measured at a depth of 20 cm during the growing
season and generalized for the upper 0.4 m of the kragge (Verhoeven & Arts 1987, Verhoeven
et al. 1988). However, on the whole-quagfen scale the amounts of P, K, and N released in the
kragge seem to be in excess of those incorporated in above-ground vegetation. According to
data in Ellenberg (1978), Dykyjov & Kv t (1978), and Van der Linden (1980, 1986)
Phragmites australis returns significant amounts of the nutrient stock into the root system from
July-August onwards. This probably also holds for the other helophytes in quagfen vegetation.
Apparently the date of mowing is very important here as regards the nutrient economy of the
quagfen as a whole, and it may significantly favour or disfavour the growth of singular plant
species.
A comparison of the contents of phosphorus, nitrogen, and potassium in the plant biomass to
standard data summarized by Kinzel (1982) reveals that these elements are represented in
smaller amounts in De Stobbenribben in all vegetational zones sampled (Table 11.3). This
possibly indicates a limiting availability of these elements in critical growth periods. According
to data reported by Kooijman (1985), Phragmites australis and Carex elata show a decreasing
relative content of potassium in the Calliergonella-Phragmites reed during the growing season.
Her data also suggest lower potassium concentrations in more superficially rooting species in the
Scorpidium-Carex and Sphagnum-Carex fen zones, possibly indicating a lower availability of
these ions near the kragge surface. Obviously any further work in this direction should include a
study of the different layers of the kragge in relation to the development of the root systems of
various species. Moreover, the choice of potassium, nitrogen, and phosphorus, prompted by
nutritional economy models of plant growth, should be extended with estimations of the
219
Table 11.3 Concentrations of some elements in plant tissues
Dicotyledoneous herbs
Cyperaceae and Juncaceae
De Stobbenribben mosses
vascular plants
% dry weight
Mg
Ca
3.8-3.9
2.7-3.0
0.4-1.0
0.9-3.1
1.3-1.6
0.3-0.5
0.31-0.38
0.13-0.20
2.8-3.1
1.6-2.1
0.7-1.8
0.6-2.4
0.25-0.31
0.19-0.25
0.01-0.09
0.03-0.20
According to Duvigneaud & Denaeyer-De Smet and Hhne (cited from Kinzel 1982) and found in De Stobbenribben
(Kooijman 1985)
quantities of metal cations in order to investigate the relation between the base state of the
quagfen and the spontaneous presence or absence of species in more detail, possibly by some
physiological effect upon nutritional economy. The present investigations were not aimed at a
detection of the physical and chemical aspects of the regulation of the availability of phosphorus
as a plant nutrient. The attention paid to the ionic ratio was inspired, however, on the hypothesis
that the cationic composition of the mire water, by governing the exchange of cations with the
binding sites on the soil complex, has relevance to the capacity of the organic soil to hold
phosphorus. While I have not contributed any further data about these matters, Kemmers
(pers.comm., Van Wirdum & Kemmers 1990) recently found statistical indications of such
relations.
All in all the results of this analysis suggest that the water management is probably
especially important as regards the supply of nitrogen in a relatively narrow zone with eutrophic
swamp vegetation bordering the ditch, and as regards the maintenance of a high base state in the
quagfen. The latter possibly has an influence on the local processes governing the
(im)mobilization and uptake of nutrients and on the selection of species that may settle. A
lowering of the base state (atmotrophication) soon leads to an extension of Sphagnum vegetation
at the cost of seepage indicators. Stands of quagfen vegetation with seepage indicators so
appear to be restricted to a zone or stage sandwiched in-between atmotrophic and nutrient-rich
ones, respectively. Harvesting of the vegetation plays its rle in the suppression of otherwise
dominant helophytes and invading shrubs, and in the overall nutrient budget of the quagfen. The
importance of the latter to the vegetation cannot be demonstrated, however, as long as no
quantification is available of the effect of the harvesting operation on the release of nutrients that
would otherwise have remained in living plant tissues.
Non-steadiness of the environment
Analyses of water samples from De Stobbenribben in the 1960-'84 period suggest a gradual
decrease of the salinity and calcium content of the water. Occasional samples also present
indications of a local exchange of calcium from the water for magnesium adsorbed at the peat
matrix in the deeper parts of the quagfen body. It is, therefore, possible that the hypothesis of a
gradual replacement of slightly thalassotrophic influences for lithotrophic ones also applies to
De Stobbenribben. Presently an atmotrophication seems to be ongoing.
220
Next to the studies in De Stobbenribben also some other, less detailed case studies of shorter
duration were carried out in this investigation (Raeymaekers 1978), but not yet reported in full
here. It appeared that the areas involved (De Provincie, Het Maatje, and De Vlakte, in the
neighbouring mire reserve De Wieden) are subject to similar hydrological influences as De
Stobbenribben. An inflow of water from surface water channels penetrates the quagfens
underneath the kragge up to a point were the influence of rain water becomes dominant. The
same has been found for De Wobberibben (Ruitenberg, unpublished data, Calis & Van Wetten
1983), a complex of quagfens close to De Stobbenribben. These complexes constitute the
majority of classic seepage sites in North-West Overijssel. An atmotrophication that has led to a
changed cover of vegetation is well documented for De Wobberibben, where extensive carpets
of Scorpidium have been replaced by Sphagnum vegetation. The atmotrophication in De
Wobberibben can be attributed to ditches becoming clogged, this reducing the access of baserich water to the quagfen parcels. Similar phenomena have also been observed in the other
investigated quagfen complexes, and they have recently been documented for the complex De
Bollemaat in De Wieden (Molenaar et al. 1990). The eco-hydrological system described for De
Stobbenribben is, therefore, held to be representative for seepage sites in North-West Overijssel.
11.11 Management and the rich-fen environment in zoned mires
Clearly, the water composition in the larger mire area, as dictated by the water management,
weather conditions, and the availability of resources for water supply, has an appreciable effect
upon the quagfens by providing regulating conditions for their base state. Although the present
investigation has not yielded any definitive conclusion as regards what determines the unique
environment required by seepage indicators, it has appeared that the old hypothesis must be
rejected.
Support is provided for the specification of an alternative hypothesis. In North-West
Overijssel water management, and especially water quality control, may be of primary
importance maintaining quagfen environment. As stated before this type of environment is not
unique to quagfen: quagfens just provide favourable conditions for the establishment of an
extreme rich-fen environment. While such an environment may be due elsewhere to discharging
lithotrophic groundwater it is also generated in landscape zones where atmotrophic and
lithotrophic water interact laterally, as in De Stobbenribben. Van Wirdum (1979) presents a
qualitative scheme of such a zoned mire landscape (Fig.11.2) according to the botanical
composition of old peats in North-West Overijssel reported by Veenenbos (1950). He explains
that the natural tendency of atmotrophication at hummock sites is balanced by base-rich water in
hollows in a zone intermediate between bog and areas that are occasionally entirely flooded by
base-rich water.
In our climate, and at the present state of eutrophication of the environment the flooded areas
are usually eutrophic and unsuitable for typical seepage fen vegetation. The analysis of water
composition in the intermediate zone (the poikilotrophic zone, Van Wirdum 1979) reveals a
similarity to water mixtures comprising a relatively large volume of atmotrophic, and a smaller
amount of lithotrophic water. The present study indicates that a very small amount of
thalassotrophic water may also contribute to the establishment of typical seepage mire
environments. Since atmotrophic and thalassotrophic water are widely available in The
Netherlands the maintenance of this species-rich mire zone is only possible if lithotrophic
influences can be guaranteed. In our climate this leads to site heterogeneity in a lithoatmocline
and to OR-association of species.
221
Fig.11.2
Water composition (vertical) in a zoned mire (horizontal)

1 Atmotrophic water; 1-2 Atmotrophic zone (ombrotrophic mire, O); 2 Ca 80% rain water, admixed
with 20% groundwater (lithotrophic); 2-3 Poikilotrophic zone (P); 3 Ca 40% groundwater, admixed
with rain water; 3-4 Flooding or seepage zone (rheo-eutrophic, R); 4 100% groundwater
From Van Wirdum (1979)
A similar zonation appears to exist in grasslands, where the central atmotrophic area is
represented by heathlands (Van Wirdum 1981).
Where the underlying strata are relatively impermeable to water, as in Nieuwkoopse Plassen
and in various mires in the province of North-Holland, the inflow of lithotrophic water
underneath the kragge is insufficient to maintain the rich-fen environment in the long run. In
North-Holland and in Norfolk (see Section 11.5) a resident body of brackish mire water may
prevent a deep atmotrophication.
In quagfens the deep-rooting kragge synusia usually become dominant if this is not
prevented by grazing or mowing. Natural succession therefore leads to carr and, ultimately, bog
vegetation, with the exception of small patches grazed by the natural fauna. This probably also
applies to non-quaking fens at our latitude. Human influences so extend the influence of the
natural fauna in establishing and maintaining a type of environment otherwise limited in
occurrence to more northern areas and, in The Netherlands, to earlier geologic periods. It is a
matter of policy to what extent the fundamental ecological variety should be maintained, so as to
allow for biological variety, and to what extent vegetation management should replace former
faunal and agricultural activities, so as to maximize the actual biological variety in selected
nature reserves. The present abundance of species threatened elsewhere indicates the potential
success of such measures.
222
APPENDIX A
Explanation of some specific terms
A.1 Specific terms related to the mire type concerned

The present study concerns a well defined type of mire, called trilveen in Dutch. Even this word
trilveen has different local meanings in The Netherlands and the translations found in English
writings may be confusing. I here define the type of mire studied in close agreement with recent
works on mire terminology (Stanek & Worley 1983, Int. Peat Soc. 1984, Bick et al. 1976, Gore
1983). The main concern is to explain in general terms what a trilveen (quagfen), a drijftil
(flotant), a kragge, and a petgat are. Meanwhile some related terms are defined. Note, especially,
that a rather broad definition is being used for the word "mire".
Wetland
Land which has the water table at, near, or above the land surface, or which is saturated for long
enough periods to promote wetland or aquatic processes as indicated by poorly drained soils,
hydrophilic vegetation, and various kinds of biological activity which are adapted to the wet
environment. Wetlands include peatlands and areas that are influenced by excess water but
which, for climatic, edaphic, or biotic reasons, produce little or no peat. Shallow open water,
generally less than 2 m deep, is also included in wetlands (Stanek & Worley 1983). Often, the
emphasis is on wetlands comprising a considerable area of such bodies of open water. For this
reason I have not taken "wetland" as an appropriate term to indicate what is defined as mire
below.
Terms
223
Mire
"Land" with a more or less permanently water-logged substratum, excluding open waters. Mires
are usually peat-producing ecosystems, but not necessarily so (Wheeler 1980a). Note that I have
preferred this definition over those that confine the mire concept to peat-producing ecosystems
(Stanek & Worley 1983, Gore 1983a). In doing so, the natural relatedness of all kinds of
"morass" (cf Rijksinstituut voor Natuurbeheer 1979, p.99-101, 107-109) is preserved in a
technical term. There are two main types of mire: undrained peatlands and marshes,
respectively.
Peatland
Any mire, or former mire, where peat has been produced to such an extent that the uppermost
layers of the profile, say 0.3 m, consists of soil materials comprising more than 50 %, by
volume, of peat. Undrained ("virgin") peatlands are mires.
Marsh
Mire with predominantly inorganic soil materials, little peat accumulation, and, usually, a grassy
vegetation cover (for a more complete treatment see: Stanek & Worley 1983).
Swamp
Mire with an average summer water table more than 0.15 m above the surface of the substratum
(Tansley, modified by Wheeler 1980). Note that the Americans usually define swamps as
forested wetlands on mainly inorganic soils (though rich in humus; Gore 1983). Swamps may be
either marshes or undrained peatlands.
Bog
Undrained peatland with an influx of rainwater only ("ombrotrophic")
Fen
Undrained peatland with an influx of groundwater or surface water ("minerotrophic", i.e.,
supplied with (nutrient) minerals from neighbouring areas)
Carr
Fen with scrub or woodland (Gore 1983a)
Quagmire
Related terms: quaking bog, quivering bog, quagfen
Equivalents: Schwingmoor (D); trilveen (NL)
Floating (quaking) mire, being a stage in hydrarch (hydroseral) succession resulting in
pond-filling; yields underfoot (Stanek & Worley 1983). Ombrotrophic types of quagmire may
be called quaking bog (quivering bog). Minerotrophic types can be named with the new term
quagfen. According to the botanical tradition in The Netherlands, trilveen is usually preserved
for quagfen with a cover of vegetation belonging to the Parvocaricetea class, to its associations
Sphagno-Caricetum lasiocarpae and Scorpidio-Caricetum diandrae in particular (Westhoff &
Den Held 1969). Compare also: Peucedano-Phragmitetum caricetosum and Acrocladio
(gigantei)-Caricetum diandrae, especially in the Norfolk Broads (U.K., Wheeler (1980);
moderately calcitrophic brown moss fen (Rybn ek in Moore 1984, p.182), rich fen (Moore &
Bellamy 1973, p.55). Note that this
224
Terms
particular type of quagfen has been called, among other things, "quaking bog" and "quivering
bog" by some Dutch authors (e.g., Segal 1966). According to Moore & Bellamy (p.10)
quagmires should be considered a primary, i.e., basin-filling, mire type, in spite of the fact that
even bog vegetation may occur within such a mire (Van Wirdum 1979, p.9-10).
Kragge (plur.kraggen)
Related terms: scragh, scraw (E, GB), hove(r) (Norfolk), plaur, plav? (R); Schwingrasen (D);
heve, zodde, zudde (NL)
(Dutch; no equivalent term found in technical English)
A more or less solid raft held together by the intertwining root systems of rhizomatous plants,
such as Phragmites australis, Typha angustifolia,
Carex lasiocarpa, Equisetum fluviatile, and Menyanthes trifoliata. Although pieces of kragge
may become detached to form floating islets, typical kragge is a sedentary formation, often
attached to firm land and formed by species invading more or less eutrophic water from a
riparian station. The terms plaur and plav are quoted for the reedbeds along the Danube (Gore
1983, p.28; Van der Toorn 1972, p.33; Pallis 1916, Sculthorpe 1967, p.423). A Phragmites reed
kragge is nicely demonstrated by Van der Toorn (1972, p.37), a Carex one (as Schwingrasen) by
Schmidt (1969, p.240).
Flotant
Related terms: drijftil (NL); sudd, sadd (Sudan); embalsado (Argentina); batume (Paraguay);
calamotale (?)
Free floating mass of plants, drifted together into sometimes extensive mats by wind and current
action. Often originating from mats of more or less free-floating aquatic macrophytes, such as
Eichhornia crassipes and Stratiotes aloides, invaded by certain herbs (in The Netherlands
Cicuta virosa, Menyanthes trifoliata, and other ones) and graminoids (Carex pseudocyperus, in
the Sudan Cyperus papyrus; Sculthorpe 1967, p.423, 479-483), but sometimes even bearing
trees (in The Netherlands Alnus glutinosa, Salix cinerea, and other ones). A description of the
Southern American flotants was provided by Bonetto (1975, p.184, 192). In excellent
photography, flotants or sudds are demonstrated by Sculthorpe (1967, p.468).
Petgat (plur.petgaten)
Synonym: trekgat (NL)
(Dutch; no equivalent term found in technical English)
A pond, originated by excavation of peat from below the water level with dredging tools.
Typical dimensions are 30-50 m width, 1.5-3.0 m depth, and a few hundreds of meters length.
The name petgat is also used for the already overgrown kragge stage. "Peat hole", "dredging
lake", "peat pond", or "peat working" might be used as English equivalents, although neither of
these terms seems to be as specific as the Dutch words petgat and trekgat. "Pet" has probably
common ethymological roots with the English "pit", while "gat" literally means "hole". The
word petgat might thus be a pleonasm; it is commonly used in several peat working areas,
especially in the Northern part of The Netherlands.
Terms
225
Fig. A.1
Diagram showing baulks and petgaten, and also kragge and flotants formed in
petgaten
(Standing) balk, baulk

Equivalent: legakker (NL)
A strip of land, in The Netherlands typically 2-5 m wide, between petgaten. The Dutch
legakkers may be composed of untouched peat, or they may have been constructed by filling
earlier drainage grubs or ditches with otherwise less useful materials, mostly a peaty clay. In
technical Dutch, especially in soil science, the untouched legakkers are mostly called "ribben",
while the other type of baulks is called "zetwal(len)". The English term seems to be more or less
indigenous in the Norfolk Broads area (cf Lambert et al. 1960).
Broad
Equivalent: wiede, wijde (NL)
Typically a lake that came into being by wind- and water-erosion of the baulks between
petgaten. Also applied to natural widenings of rivers. "Broad" and "wiede" ("wijde") have the
same ethymological range of meanings. Baulks, petgaten, kraggen, and flotants have been
illustrated in Fig.A.1
A.2 Specific terms related to the water management system
The water management system in North-West Overijssel is of a type very common in The
Netherlands. It also occurs in other parts of the world. Recently, TNO Committee on
Hydrological Research (CHO-TNO 1986) published a concise introduction to "Water in The
Netherlands", which well covers this subject. Some of the most common terms related to the
water control system, including boezems and polders, are explained below.
226
Terms
The average soil surface level in North-West Overijssel is slightly below sea level, but some
parts of the area are more low-lying, or at least drained to a lower level, and may have
groundwater levels
Fig. A.2
Diagram showing lay-out of boezem and polder, for the situation in NorthWest
Overijssel, with surface in- and outlets
1 primary pumping station; 2 secundary pumping station; l discharges sluices (low end); h inlet sluices
(high end)
maintained a few meters below sea level. (The present reference level for heights in The
Netherlands, Normaal Amsterdams Peil (NAP), almost coincides with the average sea level.)
Formerly, the area was bordered at one side by somewhat higher land and at the other side by
the Zuyderzee (Fig.A.2), with a separating dike acting as a defence against high seas. During
low tides, sluices could be opened for the discharge of surplus water. This became increasingly
annoying by the end of the last century, when the dredging of peat had left large areas of
economically nearly useless wetland with human settlements very susceptible to flooding and
storm damage. Since 1920 the primary pumping station "Stroink" maintains a nearly constant
water level by discharging excess water into the Zuyderzee.
Since the Zuyderzee became separated from the North Sea by a dam in 1932, the freshwater lake
IJsselmeer thus formed has a somewhat lower water level and it is not subject to tides. The water
level in the IJsselmeer is maintained by discharge of the surplus into the North Sea through
sluices and by control of the influx of water through the river IJssel. All this has furthered the
poldering of wetlands, in order to increase the agricultural production in the area.
The polders are the most low-lying parts of the area, separated from the main area by
embankments, or polder dams (dikes). There drainage levels are controlled by a number of
secundary pumping stations discharging into a network of canals and lakes which, together with
the mire and land areas freely draining into it, form the main water reservoir or basin, called
Terms
227
boezem in Dutch. Shallow polders, i.e., those which are not very deeply drained, also use the
boezem as a water supply during dry periods, whereas the boezem itself is replenished by intakes
through the sluices separating it from systems with a higher water-level, as diagrammatically
shown in Fig.A.2. The boezem with its in- and outlets must have a certain minimum capacity
related to the amounts of water to be retained, without causing unacceptable flooding, in the
worst possible case. In this way the water-level can be accurately regulated, so that it rarely
deviates more than ca. 0.1 m from the prescribed target value.
The local regulation of the water management is mainly in the hands of two institutional bodies:
the water board "Waterschap Vollenhove", which is in charge of the water-level control, and the
water-quality board "Zuiveringschap West-Overijssel". The name Waterschap Vollenhove is
frequently also applied to the area which is controlled by the water board, including the boezem,
several polders, and some freely draining land.
228
Terms
APPENDIX B
The syntaxonomical treatment of the "Scorpidio-Caricetum diandrae (W.Koch 1926) Westhoff

nom. nov." (Westhoff & Den Held 1969), according to Zijlstra (1981) synonymous with the
Caricetum lasiocarpae Koch 1926, varies considerably from author to author. A summary is
presented below.
Westhoff & Den Held (1969)
Association:
Alliance:
Order:
Class:
Scorpidio-Caricetum diandrae (W.Koch 1926) Westhoff 1969

Caricion davallianae Klika 1934
Tofieldietalia Preising apud Oberdorfer 1949
Parvocaricetea (Westhoff 1961 mscr.) Den Held & Westhoff 1969
The class Parvocaricetea is placed in the rankless group of "Vegetation of Fens and Bogs",
together with the classes Scheuchzerietea Den Held, Barkman & Westhoff 1969 and
Oxycocco-Sphagnetea Br.Bl. et R.Tx. 1943. The Parvocaricetea Class comprises fen vegetation
in minerotrophic environments. The herb layer is dominated by graminoids, and the moss layer
by pleurocarpic mosses, especially Amblystegiaceae, and by Sphagnum species. The
environment is described as meso- to eutrophic, with a low nitrogen status, and as weakly acid
to alkaline. Other authors have classified the pertaining communities, together with the Order
Scheuchzerietalia palustris, in a Class Scheuchzerio-Caricetea nigrae (compare Ellenberg,
Dierssen, below).
229
Within the Parvocaricetea Class the Orders Tofieldietalia and Caricetalia nigrae W.Koch
1926 em. Nordh. 1936 denuo em. R.Tx. 1937 (sub nom. Caricetalia fuscae) are distinguished.
The Association Sphagno-Caricetum lasiocarpae (Gadeceau 1909) Steffen 1931 em. Westhoff
1969 is placed in the Caricetalia nigrae, which, in The Netherlands, is considered to be
calciphobe. The Sphagno-Caricetum lasiocarpae is commonly considered to be a stage of
succession following upon the Scorpidio-Caricetum diandrae, which is, in turn, usually
preceded by vegetation of the Phragmitetea Class.
The differential species of both Associations, with respect to each other, are listed in Table
B.1. The list of the Scorpidio-Caricetum diandrae comprises all character species of the
Tofieldietalia Order and of the Caricion davallianae Alliance, although several of these species,
especially those which have not been found in North-West Overijssel, are more characteristic of
other Associations of the Caricion davallianae, such as the Parnassio-Caricetum pulicaris. The
character species of the Scorpidio-Caricetum diandrae, Eriophorum gracile and Juncus
subnodulosus, could be added to this list, and so might Utricularia intermedia and, perhaps,
Utricularia minor. These two Utricularia species are characteristic of the Association
Scorpidio-Utricularietum (Ilscher 1959 mscr.) Th.Mll. et Grs 1960 em. Den Hartog et Segal
1964 denuo em.Den Held et Westhoff 1969, which is characteristically found in, often small,
hollows in the meshes of stands of the Scorpidio-Caricetum diandrae. According to the
OR-assumption (Chapter 3) the Scorpidio-Utricularietum may be considered a part of the
Scorpidio-Caricetum diandrae, although they are not singularly, nor necessarily, associated.
It is evident from Table B.1 that the difference between the Sphagno-Caricetum lasiocarpae
and the Scorpidio-Caricetum diandrae, according to Westhoff & Den Held, is especially
expressed by bryophytes. The vascular plants that would indicate that a stand belongs to the
Scorpidio-Caricetum diandrae also present problems: Eriophorum gracile, Carex pulicaris,
Sagina nodosa, Epipactis palustris, Parnassia palustris, and Utricularia intermedia are by no
means found in all stands of the Association. With the exception of Eriophorum gracile and
Utricularia intermedia, they appear to be more characteristic of certain Associations of the
Alliance Molinion, which is placed in the Molinio-Arrhenateretea Class, or of transitional
vegetation between quagmire and Molinion. Carex curta, C. echinata, Drosera rotundifolia,
and, to some extent, Oxycoccus palustris, which would characterize the Sphagno-Caricetum
lasiocarpae, on the other hand, are often found locally in stands which would otherwise be
indicated as Scorpidio-Caricetum diandrae.
This can simply be explained from the fact that both Associations are linked in the supposed
successional sere. It remains to be seen whether such a situation is entirely different from the one
that brought Rybn ek (1964) to the preliminary proposal of an Alliance Caricion demissae with
Carex tumidicarpa, C. dioica, C. pulicaris, Triglochin palustre, Riccardia pinguis, Fissidens
adianthoides, Scorpidium scorpioides and, but unknown from quagfens in The Netherlands,
Trichophorum alpinum, Scirpus cespitosus ssp. cespitosus, and Calliergon trifarium as character
species (see also Chapter 2). The differential species with regard to the Caricion davallianae are
Drosera rotundifolia, Oxycoccus palustris, Sphagnum contortum, and Sphagnum subsecundum.
Juncus subnodulosus often occurs in a facies, as dense stands with few, if any, other vascular
plants, which is considered a Scorpidio-Caricetum diandraefacies by Westhoff & Den Held. In
North-West Overijssel Juncus subnodulosus is frequently found, without contributing to any
easier
230
Classification of quafgen vegetation
Table B.1Differential species of the associations Sphagno-Caricetum lasiocarpae (LAS) and

Scorpidio-Caricetum diandrae (DIA) according to Westhoff & Den Held (1969)
Vascular plants
LAS
Carex pulicaris
Sagina nodosa
Epipactis palustris
Liparis loeselii
Parnassia palustris
Carex curta
Carex echinata
Oxycoccus palustris
+
+
+
+
Bryophytes
LAS
Drepanocladus revolvens2) s.l.
Campylium elodes
Campylium polygamum
Campylium stellatum
Plagiomnium affine
Riccardia multifida
Pellia endiviaefolia
Pellia neesiana
Sphagnum contortum
Polytrichum commune
var. uliginosum
Sphagnum subsecundum
Sphagnum nemoreum
var. subnitens
Sphagnum flexuosum
var. flexuosum
Sphagnum fimbriatum
Sphagnum palustre
Calypogeia fissa
DIA
Not in the study area
+
+
+
+
+
+
Carex lepidocarpa
Carex flava s.s.
Juncus alpino-articulatus
ssp. arthrophyllus
Taraxacum limnanthes
Eleocharis quinqueflora1)
Equisetum variegatum
Eriophorum latifolium
Pinguicula vulgaris
DIA
Not in the study area
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Drepanocladus vernicosus
Fissidens osmundoides
Camptothecium nitens
Bryum ovatum
Bryum marratii
Mnium cinclidioides
Cinclidium stygium
Catoscopium nigritum
Riccardia chamaedryfolia
Sphagnum platyphyllum
Sphagnum cuspidatum
LAS
DIA
+
+
+
+
+
+
+
+
LAS
DIA
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
1) Recently found at one location in the study area (Piek et al., pers. comm.).,; 2) = Scorpidium revolvens + cossoni
(The species in the right-hand column have not, or not with certainty, been recorded from North-West Overijssel)
231
discrimination between the Sphagno-Caricetum lasiocarpae and the Scorpidio-Caricetum

diandrae.
Although the list of differential bryophytes seems impressive, it does not really clarify the
problem. In The Netherlands especially Calliergon giganteum and Scorpidium scorpioides form
dense stands which are in practice considered to belong to the Scorpidio-Caricetum diandrae,
especially when the Scorpidio-Utricularietum is present in the hollows. Many of the other
species seem to be more characteristic of particular microrelief elements in both Associations,
and also in certain other ones, especially in the Alliances Magnocaricion and Molinion. Dierssen
obviously treated the moss synusiae as ones that may occur in different Associations, and he
made them decisive for the presence of particular Subassociations (see below).
On the whole, the Associations Sphagno-Caricetum lasiocarpae and Scorpidio-Caricetum
diandrae seem to be closely interwoven in The Netherlands, and it is difficult to find pure
stands.
Ellenberg (1978) and Oberdorfer (1979)
The classification schemes by Ellenberg and Oberdorfer are similar in outline. Ellenberg's
scheme is followed here down to the Alliance level. The Associations are cited from Oberdorfer.
No attention is paid to bryophytes in these schemes.
Associations:
Alliance:
Order:
Class:
Caricetum lasiocarpae W.Koch 1926

Caricetum diandrae Jon. 1932 em. Oberd.1957
Caricion lasiocarpae Vanden Bergh. in Lebr. et al. 1949 (synonymous
with Eriophorion gracilis Preising apud Oberd.1957)
Scheuchzerietalia palustris Nordh.1937
Scheuchzerio-Caricetea nigrae (Nordh.1936) Tx.1937
The Order Tofieldietalia is also recognized, with, among other Alliances, the Alliance
Caricion davallianae. The Associations placed within that Alliance, however, are usually
characterized by the presence of Schoenus nigricans, S. ferrugineus, or Carex davalliana.
Triglochin palustre, as a character species of the Tofieldietalia, and Liparis loeselii,
characteristic of the Caricion davallianae, are frequently met with in Dutch quagmires.
Epipactis palustris and Taraxacum palustre, both listed for the Caricion davallianae, might be
added, although they are by no means common. Possibly they do not especially occur in
Westhoff's Scorpidio-Caricetum diandrae. The pertaining stands comprise several species
which, in the sense of Ellenberg and Oberdorfer, justify a classification under the Scheuchzerietalia rather than the Tofieldietalia, however, and some of these are even more characteristic
of the Alliance Caricion fuscae than of the Caricion lasiocarpae (Table B.2).
It is of some interest that certain communities with Carex buxbaumii, Juncus subnodulosus,
and Carex lasiocarpa, respectively, are placed under the Magnocaricion Alliance of the
Phragmitetalia Order in the Class Phragmitetea (Oberdorfer 1979, p.36). The similarity with
certain Magnocaricion Associations is especially obvious from Braun's (1968)
phytosociological study of calcareous fens in the Alpine foothills of Southern Bavaria. As
described by Braun, the Associations in the Magnocaricion and Caricion lasiocarpae (syn.
Eriophorion gracilis) are each primarily characterized by differential species groups that play a
similar role in several Associations. Examples of species of such groups are Utricularia
intermedia and Scorpidium scorpioides, which
232
Table B.2 Character species of certain Scheuchzerio-Caricetea nigrae communities according

to Ellenberg (1978, p. 902, 568).
Class:
Order:
Scheuchzerio-Caricetea nigrae
Carex dioica
Carex hostiana
Carex panicea
Carex pulicaris
(Dactylorhiza traunsteineri)
Parnassia palustris
Scheuchzerietalia
Alliances:
Caricion nigrae
Agrostis canina
Hierochlo odorata
Ranunculus flammula
Eriophorion
gracilis
Carex diandra
Carex lasiocarpa
Eriophorum gracile
(Additional species mentioned on p.568:)

Carex curta
Carex echinata
Carex nigra
Viola palustris
Tofieldietalia
Triglochin palustre
Caricion
davallianae
Liparis loeselii
Epipactis palustris
Taraxacum palustre
(Caricion
juncifoliae)
Carex tumidicarpa
Campylium stellatum
Scorpidium cossoni
Only species found in Dutch quagmires have been included here.
are thus no longer characteristic of any singular Association, but rather of similar
Subassociations or Varieties in different Associations.
The Scorpidio-Caricetum dissolutae Braun 1961 in the Magnocaricion is of particular
interest. It was named after a loose, not so pronouncedly tussocky modification of Carex elata,
mod. dissoluta sensu Braun, which is probably the modification found in quagmires in
North-West Overijssel also. The wetter stands of the mentioned Association are distinguished as
Subassociation utricularietosum, which has a variety with Campylium stellatum, Drepanocladus
revolvens var. intermedius (= Scorpidium cossoni), Carex panicea, and Carex lepidocarpa. The
drier stands, Subassociation molinietosum, lack the Utricularia synusia, but, in addition to the
species in the Campylium stellatum group, are characerized by several species that are strongly
reminiscent of the Scorpidio-Caricetum diandrae sensu Westhoff: Fissidens adianthoides,
Parnassia palustris, and certain Molinion species. The affinity to the Phragmitetea Class is
shown by the presence of such species as Phragmites australis, Equisetum fluviatile, Lycopus
europaeus, Peucedanum palustre, Galium palustre, Scutellaria galericulata, and Lysimachia
thyrsiflora.
233
The Associations Caricetum lasiocarpae and Caricetum diandrae, as they occur in Braun's
tables, are both subdivided in a scorpidietosum, a drepanocladetosum vernicosi, and a
polytrichetosum stricti. Apart from the "wet" scorpidietosum Subassociations, they all contain
several Magnocaricion and Molinion species, such as Galium palustre, Peucedanum palustre,
and Molinia caerulea, which are also present in the Scorpidio-Caricetum dissolutae
molinietosum. The drepanocladetosum vernicosi Subassociations are characterized by a group
of differential species of which only Drepanocladus vernicosus is, at least at present, absent
from the quagmires in North-West Overijssel, but the other taxa are not especially typical of
Parvocaricetea quagmires: Cardamine pratensis, Anthoxanthum odoratum, Calliergonella
cuspidata, Juncus articulatus, Ranunculus flammula, and Climacium dendroides. The
polytrichetosum stricti of both Associations is somewhat poorly documented, but it notably
includes Drosera rotundifolia and Polytrichum strictum.
In summary, it appears that few dominant species are decisive for the classification into
Associations and larger units in these schemes of classification, although they all share a core of
Scorpidio-Caricetum diandrae species, sensu Westhoff & Den Held.
Dierssen (1982)
At the higher levels of classification, Dierssen's scheme resembles Ellenberg's, although
several character species have been evaluated in a somewhat different way (Table B.3). Below
the Association level Dierssen has systematically distinguished Subassociations according to the
prevailing moss synusiae. A scorpidietosum Subassociation is recognized for each of the
following seventeen Associations in five Alliances and three different Orders:
Order:
Scheuchzerietalia palustris
Alliance:
Rhynchosporion albae
Associations:
Caricetum limosae
Sphagno tenelli-Rhynchosporetum albae
Caricetum rotundatae
Alliance:
Caricion lasiocarpae
Associations:
Caricetum lasiocarpae
Caricetum rostratae
Calamagrostio-Caricetum lyngbyei
Drepanoclado exannulati-Caricetum aquatilis
Caricetum diandrae
Drepanoclado revolventis-Caricetum chordorrhizae
Order:
Caricetalia nigrae
Alliance:
Caricion nigrae
Associations:
Caricetum magellanicae (as a rankless variety)
Calliergono sarmentosi-Caricetum saxatilis
Order:
Tofieldietalia
Alliance:
Associations:
Schoenetum nigricantis
Campylio-Caricetum dioicae
Drepanoclado revolventis-Trichophoretum cespitosi
Eleocharitetum quinqueflorae
Drepanoclado revolventis-Caricetum adelostomae
Alliance:
Caricion bicolori-atrofuscae
Association:
Caricetum microglochinis
234
Table B.3 Character species of certain Scheuchzerio-Caricetea nigrae communities according

to Dierssen (1982, various places).
Class:
(Additional from Fig.33:)
Orders:
Scheuchzerietalia palustris
Sphagnum subsecundum s.l.
(Additional from Fig. 33:)
Sphagnum riparium
Carex rostrata
Scheuchzerio-Caricetea nigrae
Carex nigra
Carex panicea
Carex rostrata (locally only?)
Carex lasiocarpa
Sphagnum flexuosum var. fallax
Sphagnum riparium
Sphagnum subsecundum s.l.
Polytrichum commune
Drepanocladus revolvens
Equisetum palustre
Caricetalia nigrae
Carex curta
Carex echinata
Epilobium palustre
Viola palustris
Juncus filiformis
Polytrichum commune
var. uliginosum
Sphagnum teres
Sphagnum palustre
Sphagnum flexuosum s.l.
Ranunculus flammula
Tofieldietalia
Parnassia palustris
Drepanocladus revolvens1)
Campylium stellatum
Campylium polygamum
(Additional from Fig.33:)
Carex flava agg.
Carex dioica
Alliances (Rhynchosporion albae and Caricion bicolori-atrofuscae disregarded):

Caricion nigrae
Carex rostrata
(as for the Order)
Carex hostiana
Carex lasiocarpa
Carex tumidicarpa
Carex pulicaris
Epipactis palustris
Eriophorum gracile
Liparis loeselii
Campylium elodes
Triglochin palustre
(Juncus subnodulosus)
Differential in respect to
(Carex buxbaumii)
Rhynchosporion albae:
Taraxacum palustre
Peucedanum palustre
Only species which occur in Dutch quagmires have been included here. Some additional species have been mentioned by
Dierssen in his Fig.33 (p.94); these have been included here as indicated in the table. 1) Scorpidum revolvens + S. cossoni
235
Although several of these Associations do not occur in The Netherlands, it is at once

demonstrable that the presence of Scorpidium scorpioides is only diagnostic at the Class and
Subassociation levels in Dierssen's scheme. Carex elata is, according to Dierssen's records,
especially characteristic of the Caricetum elatae, in the Alliance Magnocaricion, Order
Phragmitetalia. This Alliance has been narrowed somewhat by excluding the Caricetum
rostratae, which is now placed in the Caricion lasiocarpae, Order Scheuchzerietalia palustris.
Menyanthes trifoliata, Potentilla palustris, Carex rostrata, and Equisetum fluviatile thus become
slightly more characteristic of the Caricion lasiocarpae, especially in comparison to the
Magnocaricion.
The stands of quagmire vegetation in North-West Overijssel would, according to this
proposal, also become more strongly related to the Caricion lasiocarpae. The affinity to
Magnocaricion vegetation is nonetheless retained, since all but one of the character species of
the Magnocaricion sensu Dierssen, Scutellaria galericulata, Galium palustre ssp. elongatum,
Carex acutiformis, Lysimachia thyrsiflora, Carex disticha, and Peucedanum palustre, belong to
the most constant associate species in these stands of quagmire vegetation. Lysimachia
thyrsiflora and Peucedanum palustre are also recognized by Dierssen as differential species of
the Caricion lasiocarpae with regard to the Rhynchosporion albae.
The Caricetum dissolutae sensu Braun is not mentioned by Dierssen, but, on p.110, he refers
to Baltov-Tul kov (1972), who includes both the Caricetum lasiocarpae and the Caricetum
diandrae, together with the Caricetum rostratae and certain other Associations in the
Magnocaricion, especially referring to the vigirous growth of Menyanthes trifoliata, Potentilla
palustris, Carex rostrata, and other species. Records from North-West Overijssel contain
several species which prove a relation with the Caricion nigrae and Caricion davallianae (Table
B.3). It is noteworthy that Dierssen might have overcome this problem (or: missed this
resemblance) through the use of much smaller quadrats in his records.
The Juncetum subnodulosi is placed in the Caricion davallianae by Dierssen. In conclusion
there is strong evidence that the stands of vegetation characterized by the abundance of Carex
lasiocarpa, C. diandra, and Juncus subnodulosus cannot be easily classified in the schemes of
the Zrich-Montpellier school, regardless of the attention paid to the moss layer by different
authors.
Wheeler (1975, 1980, 1984)
Wheeler's scheme is probably not of the same kind as the other Zrich-Montpellier
classification systems. Although Wheeler (1980) refers to a comparison of Zrich-Montpellier
tabulations with statistical clustering techniques, his ultimate scheme reflects the results of the
latter approach. Formally, therefore, the characteristic species mentioned in this scheme are not
character and differential species in the Zrich-Montpellier sense. The Dutch quagmire
vegetation can probably be classed under two Associations in Wheeler's system. Down to the
level of Alliances the scheme largely coincides with that of Westhoff & Den Held (see before).
Association 1:
Alliance:
Order:
Class:
236
Peucedano-Phragmitetum Wheeler 1978

Magnocaricion elatae Koch 1926
Magnocaricetalia Pignatti 1953
Phragmitetea Txen et Preising 1942
Association 2:
Alliance:
Order:
Class:
Acrocladio-Caricetum diandrae Klika 1934

Caricion davallianae Klika 1934
Tofieldietalia Preising apud Oberdorfer 1949
Parvocaricetea Den Held et Westhoff 1969
Several Subassociations and Varieties have been distinguished by Wheeler, such as the
Carex lasiocarpa variety of the Peucedano-Phragmitetum cicutetosum. His work is of special
interest since it is based upon, among others, many vegetation records from The Broads in
Norfolk, an area which has much in common with North-West Overijssel. From his
descriptions, it is obvious that Carex lasiocarpa and Carex diandra, and, less typically so, the
amblystegiaceous mosses, are especially present in Norfolk in stands of vegetation somewhat
intermediary between the Phragmitetea and Parvocaricetea Classes. The Peucedano-Phragmitetum typically includes Carex elata, Juncus subnodulosus, Lysimachia vulgaris, Peucedanum
palustre, Thelypteris palustris, and other taxa, and it is usually dominated by Phragmites
australis, Cladium mariscus, or, less often, Calamagrostis canescens.
Very similar stands are present in North-West Overijssel, but in most typical quagfens, that
can be classed under Wheeler's Acrocladio-Caricetum diandrae, regular hay-making prevents
the dominance of these tall helophytes. The Acrocladio-Caricetum diandrae is, according to
Wheeler, typically characterized by mixed stands of Carex diandra, C. lasiocarpa, and C.
rostrata, usually with much Menyanthes trifoliata and Potentilla palustris, and sometimes
Carex limosa. Carex limosa is absent from the quagfens in North-West Overijssel, but it was
formerly associated with similar types of quagmire elsewhere in The Netherlands. Brown
mosses are frequently abundant in the Acrocladio-Caricetum (notably Calliergon giganteum),
but Sphagnum is also present, especially in less base-rich examples.
237
APPENDIX C
Indicator list of fen-mire species
In this appendix the species have been listed according to their scientific name (Van der Meijden
et al. 1983, Dirkse et al. 1989) with the indications attached to them (Chapter 3). A short
explanation of abbreviations is given below. For any details, reference should be made to
Chapter 3. The list is an extract of a larger database.
T
!
#
('threatened')
Especially occurring in protected areas
Especially occurring in protected areas, also a 'seepage' indicator
R
0
1
2
3
(Red Data List, Weeda et al. 1990)

Extinct and presumed extinct
Endangered
Most vulnerable
Vulnerable
4
5
6
Rare
Strongly declining, but not yet 1-4
Legal protection proposed, but not 1-5
(not considered 'threatened' in Chapter 3)
Typ
AQU
BOG
DAV
FEN
LAS
(Phytosociological group)
Communities of aquatic macrophytes
Oxycocco-Sphagnetea, Nardo-Callunetea
Other Parvocaricetea, Caricion fuscae
LIT
MOL
SMA
SMP
Litter fen
Molinion
Salt-marsh, brackish fen communities
Various swamp communities
Wat (Water type, acidity)

ATM Atmotrophic
CIR
Nut (Trophic state)

EUT Eutrophic
MES Mesotrophic
Circumneutral
LTH Lithotrophic
OLI
Oligotrophic
Indications according to the Finnish tradition (Eurola et al. 1984):

Bas Base state, as found in the following milieus:
OMB Ombrotrophic (very low)
POR Poor (low base state)
WMS Wide-range mesotrophic
XRC Extreme rich-fen
Lev Groundwater level as in:
HUM Hummocks
Ext
SEP
NO
INT
Intermediate mire parts
Supplementary nutrient effect associated with:

Seepage in general
FLD Flooding
No supplementary nutrient effect
Inh Inherent nutrient effect associated with:

MOS Moss (bog) peat
NVA Neva (intermediate mire)
NO No inherent nutrient effects
TRL Transitional
FLK Flarks (hollows)
ANY Either or both
RFN Rich-fen peat
Indicator list
239
Species Name
Author
Achillea ptarmica
Acorus calamus
Agrostis canina
Agrostis stolonifera
Ajuga reptans
Alisma lanceolatum
Alisma plantago-aquatica
Alnus glutinosa
Althaea officinalis
Amblystegium riparium
Amblystegium serpens
Amblystegium varium
Andromeda polifolia
Aneura pinguis
Angelica sylvestris
Anthoxanthum odoratum
Aronia x prunifolia
Aster tripolium
Atrichum undulatum
Atriplex prostrata
Azolla filiculoides
Berula erecta
Betula pubescens
Bidens cernua
Bidens connata
Bidens frondosa
Bidens tripartita
Brachythecium rutabulum
Briza media
L.
L.
L.
L.
L.
Withering
L.
(L.) Grtner
L.
! 3
(Hedw.) Schimp.
(Hedw.) Schimp.
(Hedw.) Mitt.
L.
! 5
(L.) Dum.
#
L.
L.
(Marshall) Rehder
L.
(Hedw.) P.Beauv.
Boucher ex DC.
(Hedw.) Schwgr.
Lamk.
(Hudson) Coville
Ehrhart
L.
Mhlenb. ex Willdenow
L.
L.
(Hedw.) Schimp.
L.
! 3
(Hedw.)Grtn., Meyer et Scherb.
#
(Weber) Roth
(Timm) Koeler
# 4
L.
! 6
(Hedw.) Kindb.
(Schimp.) Kindb.
!
(Brid.) Kindb.
(Hedw.) Loeske
Ktzing
(L.) Hull
L.
(L.) Raddi
(L.) R.Br.
(Lindb.) Kindb.
#
(Schimp.) J.Lange et C.Jens.
(Hedw.) J.Lange et C.Jens. #
(K.F.Schultz) Brid.
L.
L.
Ehrhart
Schumacher
! 3
Wahlenberg
# 4
Goodenough
Schrank
# 5
Calamagrostis canescens
Calla palustris
Calliergon cordifolium
Calliergonella cuspidata
Callitriche platycarpa
Calluna vulgaris
Caltha palustris
Calypogeia fissa
Calystegia sepium
Campylium elodes
Campylium polygamum
Campylium stellatum
Campylopus pyriformis
Cardamine pratensis
Carex acuta
Carex acutiformis
Carex appropinquata
Carex buxbaumii
Carex curta
Carex diandra
240
T R Typ Wat Nut

MOL
Bas
Lev
Ext
Inh
TRL
INT
FLD
NO
OLI
SMP
CIR
LIT
ATM
LIT
EUT
OLI
MES
MOL
LTH
MES
SMP
CIR
MES
SMP
CIR
EUT
TRL
FLK
FLD
NO
LIT
CIR
EUT
WMS
HUM FLD
NO
LIT
SMP
CIR
EUT
LIT
CIR
EUT
LIT
LTH
EUT
BOG
ATM
OLI
OMB
INT
NO
NVA
LAS
LTH
OLI
WMS
FLK
SEP
RFN
LIT
LTH
BOG
SMA
FEN
WMS
HUM SEP
RFN
TRL
INT
NO
OMB
HUM NO
MOS
POR
HUM ANY
NVA
TRL
HUM ANY
NO
ANY
EUT
ATM
EUT
ATM
MES
SMA
FEN
OLI
OLI
EUT
AQU
EUT
SMP
CIR
EUT
DAV
ATM
OLI
SMP
EUT
SMP
EUT
SMP
EUT
SMP
EUT
LIT
EUT
MOL
OLI
LAS
LTH
MES
WMS
FLK
ANY
RFN
LIT
ATM
OLI
TRL
FLK
FLD
NO
NO
LAS
CIR
OLI
TRL
FLK
FLD
SMP
CIR
MES
TRL
FLK
FLDNO
SMP
CIR
EUT
TRL
FLK
ANY
NO
LAS
LTH
MES
WMS
FLK
ANY
NO
FEN
ATM
LIT
OLI
POR
FLK
ANY
NVA
MES
WMS
FLK
ANY
NO
SMP
BOG
ATM
OLI
OMB
HUM NO
MOS
MOL
CIR
MES
TRL
FLK
NO
MES
OMB
HUM NO
MOS
CIR
EUT
CIR
MES
XRC
INT
SEP
RFN
WMS
FLK
ANY
NO
FEN
LIT
MOL
LIT
FLD
MES
DAV
LTH
OLI
LIT
ATM
OLI
MOL
CIR
MES
SMP
CIR
MES
SMP
CIR
MES
WMS
FLK
SEP
NO
SMP
LTH
OLI
XRC
INT
SEP
NO
MOL
CIR
OLI
WMS
FLK
FLD
RFN
FEN
ATM
OLI
POR
FLK
FLD
NO
LAS
CIR
OLI
WMS
FLK
ANY
RFN
Indicator list
Species Name
Author
T R Typ Wat Nut
Bas
Lev
Carex dioica
Carex disticha
Carex echinata
Carex elata
Carex hostiana
Carex lasiocarpa
NVA
Carex nigra
Carex panicea
Carex paniculata
Carex pseudocyperus
Carex pulicaris
Carex riparia
Carex rostrata
Carex tumidicarpa
Carex vesicaria
Centaurea jacea
Cephalozia bicuspidata
var. lammersiana
Cephalozia connivens
Cephaloziella hampeana
Cephaloziella spinigera
Chiloscyphus pallescens
Cicuta virosa
Cirsium dissectum
Cirsium palustre
Cladium mariscus
Climacium dendroides
Cochlearia officinalis
Cratoneuron filicinum
Dactylorhiza maculata
Dactylorhiza majalis
L.
Hudson
Murray
Allioni
DC.
Ehrhart
! 1
WMS
HUM NO
(L.) Reichard
L.
L.
L.
L.
Curtis
Stokes
N.J.Anderson
L.
L.
! 6
OLI
CIR
Inh
RFN
MES
FEN
ATM
OLI
TRL
FLK
SEP
NO
FEN
CIR
MES
WMS
FLK
FLD
NO
POR
FLK
FLD
! 2
# 5
BOL
ATM
OLI
FEN
CIR
OLI
FEN
ATM
OLI
TRL
INT
FLD
NO
! 6
MOL
ATM
OLI
XRC
INT
NO
RFN
! 2
SMP
LTH
EUT
XRC
FLK
SEP
NO
SMP
LTH
EUT
WMS
FLK
FLD
NO
MOL
ATM
SMP
!
!
! 6
(Hb.) Breidl.
!
(Dicks.) Lindb.
!
(Nees) Schiffn.
!
(Lindb.) Jrg.
!
(Ehrh.) Dum.
L.
(l.) Hill
! 2
(L.) Scopoli
(L.) Pohl
! 3
(Hedw.) Web. et Mohr
L.
3
(Hedw.) Spruce
!
(L.) Soo
# 3
(L.) Soo
! 3
(Reichenb.)Hunt & Summerhayes
!
Danthonia decumbens
(L.) DC.
Deschampsia cespitosa
(L.) Beauv.
Deschampsia flexuosa
(L.) Trinius
Dicranella cerviculata
(Hedw.) Schimp.
Dicranella heteromalla
(Hedw.) Schimp.
Dicranum bonjeanii
De Not.
Dicranum scoparium
Hedw.
Drepanocladus aduncus
(Hedw.) Warnst.
Drepanocladus fluitans
(Hedw.) Warnst.
Drepanocladus lycopodioides (Brid.) Warnst.
#
Drepanocladus sendtneri
(H.Mll.) Warnst.
!
L.
! 5
Dryopteris carthusiana
(Villars) H.P.Fuchs
Dryopteris cristata
(L.) A.Gray
!
Dryopteris dilatata
(Hoffmann) A.Gray
Echinodorus ranunculoides
(L.) Engelmann ex Ascherson
! 2
Eleocharis palustris
ssp. palustris
ssp. uniglumis
(Link) Hartman
Eleocharis quinqueflora
(Hartmann) Schwarz
! 2
Indicator list
DAV
LIT
Ext
FEN
ATM
OLI
MES
WMS
FLK
FLD
NO
OLI
POR
FLK
FLD
NVA
TRL
FLK
FLD
NO
MOL
LTH
MES
FEN
ATM
MES
LIT
FEN
CIR
MES
FEN
ATM
OLI
LAS
ATM
MES
BOG
ATM
OLI
LIT
CIR
MES
TRL
FLK
SEP
NO
SMP
LTH
EUT
WMS
FLK
FLD
NO
MOL
ATM
WMS
INT
ANY
NO
TRL
INT
FLD
NO
MOL
SMP
OLI
OLI
LTH
MOL
OLI
MES
SMA
DAV
LTH
MES
XRC
FLK
SEP
NO
DAV
CIR
OLI
WMS
INT
NO
RFN
BOG
ATM
OLI
TRL
INT
SEP
RFN
ANY
MOL
CIR
OLI
BOG
ATM
OLI
LIT
ATM
MES
TRL
INT
BOG
ATM
OLI
TRL
HUM SEP
NO
OMB
HUM NO
NO
LIT
ATM
OLI
BOG
ATM
OLI
FEN
ATM
MOL
NO
MES
WMS
INT
SEP
RFN
MES
OMB
HUM SEP
MOS
SMP
LTH
EUT
WMS
FLK
FLD
NO
BOG
ATM
MES
OMB
FLK
NO
NVA
LAS
LTH
MES
DAV
LTH
OLI
FEN
ATM
OLI
OMB
HUM NO
MOS
FEN
ATM
OLI
TRL
HUM SEP
NO
OLI
TRL
INT
NO
CIR
EUT
FEN
LIT
FLD
SMP
SMP
CIR
TRL
FLK
FLD
NO
SMP
CIR
TRL
FLK
FLD
NO
DAV
LTH
XRC
FLK
SEP
RFN
OLI
241
Species Name
Author
Empetrum nigrum
Epilobium hirsutum
Epilobium palustre
Epilobium parviflorum
Epipactis palustris
Equisetum arvense
Equisetum palustre
Erica tetralix
Eriophorum gracile
Eriophorum vaginatum
Eupatorium cannabinum
Euphrasia stricta
Eurhynchium praelongum
Eurhynchium striatum
Festuca arundinacea
Festuca ovina
Festuca rubra
Filipendula ulmaria
Fontinalis antipyretica
Frangula alnus
Fraxinus excelsior
Funaria hygrometrica
Galeopsis bifida
Galium palustre
Galium uliginosum
Gentiana pneumonanthe
Glyceria maxima
Hammarbya paludosa
Heracleum sphondylium
Hierochlo odorata
Holcus lanatus
Humulus lupulus
Hydrocharis morsus-ranae
Hypericum tetrapterum
Hypnum jutlandicum
Impatiens noli-tangere
Iris pseudacorus
Juncus acutiflorus
Juncus articulatus
Juncus bulbosus
Juncus conglomeratus
Juncus effusus
Juncus filiformis
Juncus gerardii
Juncus subnodulosus
Kurzia pauciflora
Lathyrus palustris
Lemna gibba + Lemna minor
Lemna trisulca
Leontodon autumnalis
Leucobryum glaucum
L.
L.
L.
Schreber
(L.) Crantz
L.
L.
L.
L.
Honckeny
Roth
L.
L.
Wolff ex J.F.Lehmann
(Hedw.) Schimp.
(Hedw.) Schimp.
Schreber
L.
L.
(L.) Maximowicz
Hedw.
Hedw.
Miller
L.
Hedw.
Bnninghausen
L.
L.
L.
(Hartman) Holmberg
(L.) O.Kuntze
L.
(L.) Beauv.
L.
L.
L.
L.
Fries
Holmen et Warncke
L.
L.
Ehrhart ex Hoffmann
L.
L.
L.
L.
L.
Loisel.
Schrank
(Dicks.) Grolle
L.
L.
L.
L.
(Hedw.) ngstr.
242
T R Typ Wat Nut

BOG
LIT
!
! 3
OLI
LTH
FEN
LTH
MES
LTH
MES
Ext
OMB
HUM NO
MOS
TRL
FLK
ANY
NO
SEP
DAV
LTH
OLI
XRC
INT
LIT
LTH
OLI
TRL
HUM ANY
FEN
CIR
Inh
NO
NO
MES
TRL
FLK
FLD
NVA
OLI
WMS
INT
ANY
NO
BOG
ATM
OLI
FEN
ATM
OLI
POR
FLK
FLD
NVA
LAS
LTH
OLI
WMS
FLK
FLD
NVA
OMB
HUM NO
BOG
ATM
OLI
LIT
LTH
EUT
CIR
MES
MES
MOS
BOG
SMP
MOL
#
!
Lev
EUT
LIT
SMP
5
5
# 1
! 3
Bas
EUT
LIT
CIR
LIT
ATM
MOL
CIR
LIT
LTH
DAV
AQU
LTH
FEN
LTH
MES
WMS
HUM SEP
RFN
WMS
INT
RFN
FLD
MES
WMS
INT
ANY
NO
MES
XRC
FLK
ANY
RFN
SEP
MES
SMP
CIR
EUT
LIT
CIR
EUT
TRL
FLK
WMS
HUM ANY
NO
NO
XRC
HUM ANY
NO
LIT
6
! 5
! 1
! 6
!
!
!
! 6
SMP
CIR
MOL
CIR
MOL
ATM
OLI
SMP
LTH
EUT
FEN
CIR
OLI
LIT
LTH
MES
LIT
CIR
OLI
MOL
CIR
MES
SMP
CIR
EUT
AQU
CIR
EUT
LIT
ATM
OLI
MOL
CIR
OLI
TRL
FLK
ANY
NO
TRL
INT
FLD
NO
WMS
FLK
FLD
RFN
WMS
INT
FLD
BOG
ATM
OLI
SMP
CIR
MES
WMS
FLK
SEP
NO
SMP
CIR
EUT
WMS
FLK
FLD
NO
TRL
FLK
FLD
NO
WMS
INT
FLD
NO
MOL
CIR
OLI
LIT
CIR
OLI
SMP
ATM
OLI
LIT
ATM
OLI
LIT
ATM
OLI
MOL
CIR
OLI
SMA
!
!
! 6
OLI
MES
LIT
LTH
BOG
ATM
OLI
LIT
LTH
OLI
LTH
EUT
ATM
OLI
AQU
AQU
LIT
BOG
Indicator list
Species Name
Author
T R Typ Wat Nut
Linum catharticum
Liparis loeselii
Lonicera periclymenum
Lophocolea bidentata
Lophocolea heterophylla
Lotus uliginosus
Luzula campestris
Luzula multiflora
ssp. multiflora
Lychnis flos-cuculi
Lycopus europaeus
Lysimachia vulgaris
Lythrum salicaria
Marchantia polymorpha
Mentha aquatica
Mnium hornum
Moehringia trinervia
Molinia caerulea
Myosotis laxa
Myosotis palustris
Myrica gale
Nardus stricta
Nasturtium microphyllum
Nuphar lutea
Nymphaea alba
Oenanthe aquatica
Oenanthe fistulosa
Ophioglossum vulgatum
Osmunda regalis
Oxycoccus macrocarpos
Oxycoccus palustris
Pallavicinia lyellii
Parnassia palustris
Pellia endiviifolia
Pellia epiphylla
Pellia neesiana
Peucedanum palustre
Phalaris arundinacea
Philonotis marchica
Phragmites australis
L.
(L.) Richard
L.
(L.) Dum.
(Schrad.) Dum.
Schkuhr
(L.) DC.
! 3
# 2
Pinus sylvestris
Plagiomnium affine
Plagiomnium elatum
Plagiomnium ellipticum
Plagiothecium denticulatum
var. undulatum
Plantago lanceolata
Platanthera bifolia
Pleurozium schreberi
Poa palustris
Poa pratensis
Indicator list
Ruthe ex Geh.
L.
(L.) Richard
(Brid.) Mitt.
L.
L.
LTH
OLI
LTH
OLI
FEN
ATM
LIT
L.
L.
L.
L.
L.
L.
L.
L.
#
Hedw.
(L.) Clairville
(L.) Mnch
Lehmann
(L.) L.
L.
!
L.
(Bnninghausen) Airy Shaw
(L.) J.E.Smith
L.
(L.) Poiret
L.
L.
!
L.
!
(Aiton) Pursh
!
Persoon
!
(Hook.) S.F.Gray
!
L.
#
L.
!
(Dicks.) Dum.
!
(L.) Corda
(Gottsche) Limpr.
!
(L.) Mnch
!
L.
(Hedw.) Brid.
#
(Cavanilles) Trinius ex Steudel
L.
(Bland.) T.Kop.
(Bruch et Schimp.) T.Kop.
(Brid.) T.Kop.
MOL
LAS
5
6
CIR
MES
CIR
MES
BOG
ATM
OLI
Inh
MOL
CIR
OLI
MOL
CIR
MES
WMS
INT
SEP
NO
SMP
LTH
MES
WMS
FLK
FLD
NO
SMP
CIR
OLI
TRL
FLK
FLD
NO
LIT
CIR
OLI
TRL
FLK
FLD
NO
TRL
FLK
FLD
NO
FLD
LIT
CIR
MES
SMP
LTH
EUT
SMP
CIR
FEN
ATM
MES
OLI
POR
FLK
OLI
TRL
HUM FLD
NO
WMS
INT
RFN
SMP
CIR
EUT
MOL
ATM
OLI
ANY
NVA
MES
MOL
ATM
MES
FEN
CIR
MES
TRL
HUM FLD
NO
BOG
ATM
OLI
WMS
INT
NO
EUT
SMP
CIR
AQU
CIR
AQU
CIR
EUT
SMP
LTH
MES
SMP
CIR
MES
SEP
TRL
FLK
FLD
NO
TRL
FLK
FLD
NO
OMB
FLK
FLD
NVA
MOL
LIT
CIR
BOG
3
3
Ext
MES
LIT
SMP
Lev
EUT
MOL
LIT
Bas
MES
OLI
BOG
ATM
OLI
FEN
ATM
MES
DAV
LTH
OLI
WMS
INT
ANY
NO
LAS
CIR
OLI
WMS
FLK
FLD
NVA
DAV
LTH
OLI
TRL
INT
ANY
NO
LIT
ATM
EUT
TRL
INT
ANY
NO
LAS
CIR
MES
TRL
INT
ANY
NO
FEN
CIR
OLI
TRL
FLK
FLD
NO
SMP
CIR
MES
TRL
INT
FLD
NO
DAV
LTH
OLI
WMS
FLK
SEP
NO
SMP
LTH
MES
WMS
FLK
FLD
BOG
ATM
OMB
HUM NO
RFN
MOS
LIT
CIR
EUT
WMS
FLK
ANY
NO
DAV
LTH
MES
XRC
FLK
SEP
NO
SMP
CIR
EUT
WMS
FLK
ANY
NO
LIT
CIR
EUT
TRL
HUM SEP
NO
LIT
MOL
CIR
MOL
ATM
OLI
OMB
HUM NO
MOS
SMP
CIR
MES
TRL
INT
NO
LIT
FLD
MES
243
Species Name
Author
Poa trivialis
L.
Pohlia nutans
(Hedw.) Lindb.
Polygonum amphibium
L.
Polygonum hydropiper
L.
Polytrichum commune
Hedw.
Polytrichum juniperine
Hedw.
Polytrichum longisetum
Swartz ex Brid.
Populus tremula
L.
Potentilla anglica
Laicharding
Potentilla anserina
L.
Potentilla erecta
(L.) Ruschel
(L.) Scopoli
Potentilla reptans
L.
Prunella vulgaris
L.
Pseudoscleropodium purum
(Hedw.) Fleisch. ex Broth.
Quercus robur
L.
Ranunculus acris
L.
Ranunculus ficaria
L.
Ranunculus flammula
L.
Ranunculus lingua
L.
Ranunculus repens
L.
Ranunculus sceleratus
L.
Rhinanthus angustifolius
C.C.Gmelin
(Bruch et Schimp.) T.Kop.
Rhizomnium punctatum
(Hedw.) T.Kop.
Rhytidiadelphus squarrosus
(Hedw.) Warnst.
Riccardia multifida
(L.) S.F.Gray
Riccia fluitans
L.
Ricciocarpos natans
(L.) Corda
Rorippa amphibia
(L.) Besser
Rubus fruticosus
L.
Rumex acetosa
L.
Rumex hydrolapathum
Hudson
Rumex palustris
J.E.Smith
Sagina nodosa
(L.) Fenzl
Sagina procumbens
L.
Salix aurita
L.
Salix cinerea
L.
Salix pentandra
L.
Salix repens
L.
Samolus valerandi
L.
Sanguisorba officinalis
L.
Scirpus lacustris
ssp. lacustris
ssp. tabernaemontani
(C.C.Gmelin) Syme
Scirpus maritimus
L.
Scorpidium cossoni
(Schimp.) Hedens
Scorpidium lycopodioides
(Brid.) Paul
Scorpidium revolvens (pro parte)
(Swartz) Rubers
(Hedw.) Limpr.
Scutellaria galericulata
L.
Senecio paludosus
L.
Sium latifolium
L.
244
T R Typ Wat Nut
Lev
Ext
SEP
SMP
LTH
EUT
WMS
FLK
LIT
ATM
MES
OMB
HUM NO
ATM
EUT
SMP
SMP
Bas
Inh
NO
MOS
EUT
FEN
ATM
OLI
POR
HUM SEP
NO
BOG
ATM
OLI
OMB
HUM NO
MOS
LIT
MES
BOG
TRL
INT
FLD
NO
TRL
HUM SEP
NO
LIT
LIT
EUT
MOL
ATM
OLI
WMS
INT
SEP
RFN
FEN
ATM
OLI
TRL
FLK
FLD
NVA
LIT
CIR
WMS
HUM SEP
NO
TRL
FLK
NO
MES
MOL
OLI
BOG
MES
BOG
MOL
CIR
MES
LIT
LTH
EUT
LIT
CIR
OLI
SMP
LTH
MES
SMP
CIR
EUT
! 5
MOL
CIR
OLI
!
!
DAV
CIR
MES
WMS
FLK
SEP
RFN
DAV
ATM
MES
TRL
INT
SEP
NO
CIR
OLI
WMS
INT
SEP
NO
! 5
LIT
EUT
MOL
#
!
!
LAS
ANY
MES
SMP
CIR
MES
SMP
LTH
MES
SMP
CIR
LIT
EUT
EUT
MOL
CIR
MES
SMP
CIR
EUT
SMP
# 3
!
LAS
LTH
MES
LIT
CIR
MES
FEN
CIR
OLI
POR
HUM ANY
NO
FEN
CIR
OLI
TRL
HUM FLD
NO
!
FEN
!
MOL
!
SMA
! 6 MOL
LTH
MES
WMS
INT
FLD
NO
CIR
OLI
TRL
HUM ANY
NO
CIR
OLI
!
!
SMP
LTH
MES
SMP
LTH
SMA
LTH
MES
WMS
FLK
FLD
NO
WMS
FLK
FLD
NO
XRC
INT
NO
RFN
MES
#
DAV LTH OLI
see Drepanocladus l.
see S. cossoni
#
LAS
LTH
OLI
XRC
FLK
FLD
RFN
SMP
CIR
MES
TRL
INT
FLD
NO
LTH
EUT
SMP
SMP
MES
Indicator list
Species Name
Author
Solanum dulcamara
Sonchus palustris
Sorbus aucuparia
Sparganium emersum
Sparganium erectum
Sparganium minimum
Sphagnum contortum
Sphagnum fimbriatum
Sphagnum flexuosum
Sphagnum fuscum
Sphagnum magellanicum
Sphagnum palustre
Sphagnum papillosum
Sphagnum riparium
Sphagnum rubellum
Sphagnum russowii
Sphagnum squarrosum
Sphagnum subnitens
Sphagnum subsecundum
ssp. subsecundum
Sphagnum teres
Spirodela polyrhiza
Stachys palustris
Stellaria palustris
Succisa pratensis
Symphytum officinale
Thalictrum flavum
Thelypteris palustris
Trifolium dubium
Trifolium pratense
Trifolium repens
Triglochin maritima
Triglochin palustris
Typha angustifolia
Typha latifolia
Urtica dioica
Utricularia minor
Utricularia vulgaris
Vaccinium vitis-idaea
Valeriana dioica
Valeriana officinalis
Veronica anagallis-aquatica
Veronica beccabunga
Veronica scutellata
Viburnum opulus
Vicia cracca
Viola palustris
L.
L.
L.
Rehmann
L.
Wallroth
K.F.Schultz
Wils.
Dozy & Molk.
(Schimp.) Klinggr.
Brid.
L.
Lindb.
ngstr.
Wils.
Warnst.
Crome in Hoppe
Russ. & Warnst.
Indicator list
Nees
(Schimp.) ngstr.
(L.) Schleiden
L.
Retzius
Mnch
L.
L.
Schott
Sibthorp
L.
L.
L.
L.
L.
L.
L.
Hayne
L.
L.
L.
L.
L.
L.
L.
L.
L.
L.
L.
T R Typ Wat Nut

SMP
LIT
EUT
CIR
EUT
CIR
MES
LIT
SMP
! 3
!
SMP
CIR
MES
SMP
CIR
OLI
LAS
CIR
FEN
!
!
Bas
Lev
Ext
Inh
XRC
FLK
FLD
NO
TRL
HUM SEP
NO
MES
XRC
FLK
ANY
RFN
MES
TRL
INT
FLD
NO
FEN
ATM
MES
OMB
HUM NO
MOS
BOG
ATM
OLI
OMB
HUM NO
MOS
BOG
ATM
OLI
OMB
HUM NO
MOS
FEN
!
!
!
!
BOG
ATM
OLI
OMB
HUM NO
NVA
BOG
ATM
MES
OMB
HUM SEP
MOS
LIT
CIR
EUT
TRL
INT
FLD
NO
LAS
CIR
MES
WMS
INT
NO
RFN
!
!
LAS
CIR
OLI
WMS
FLK
FLD
RFN
FEN
CIR
OLI
WMS
INT
ANY
NO
CIR
MES
TRL
INT
FLD
NO
WMS
FLK
FLD
NO
BOG
ATM
OLI
OMB
HUM NO
NVA
FEN
ATM
MES
TRL
FLK
NO
AQU
LIT
6
! 5
EUT
FEN
CIR
MES
MOL
ATM
OLI
CIR
OLI
SMP
LIT
EUT
FEN
CIR
MES
MOL
CIR
MES
MOL
LTH
MES
LIT
CIR
EUT
LIT
LTH
OLI
SMP
LTH
SMA
SMP
# 1
!
XRC
INT
NO
NO
WMS
FLK
SEP
RFN
MES
WMS
FLK
FLD
NO
EUT
WMS
FLK
FLD
NO
SMP
CIR
EUT
WMS
INT
SEP
NO
LAS
LTH
OLI
TRL
FLK
FLD
NVA
FEN
CIR
MES
TRL
FLK
FLD
NVA
SMP
CIR
MES
TRL
FLK
FLD
NO
OMB
HUM NO
MOS
WMS
HUM ANY
NO
TRL
INT
NO
BOG
ATM
OLI
! 5
MOL
CIR
OLI
LIT
LTH
SMP
LIT
SMP
! 6
FLD
MES
MES
CIR
MES
SMP
CIR
MES
MOL
CIR
OLI
FEN
ATM
MES
FLD
245
APPENDIX D
Evaluation of the major-ionic composition of

natural waters
D.1 Processing of water quality data

The ionic composition of natural waters plays an important role in ecological
studies for the protection of nature. A method for the mutual comparison of water
analyses on the basis of their similarity to certain chosen reference waters was
developed for use in such studies. The calculations involved in this method can be
done with a computer program called MAION, and this name MAION was also
chosen for the method as a whole. The MAION method was specially developed
during the present study of the hydrological ecology of quagfen vegetation in
North-West Overijssel. It is separately dealt with in this appendix, together with
several auxiliary functions of the MAION program.
The MAION method combines several well-known procedures from the literature
with some new ideas. The main additions to be discussed are: (1) the EC-IR
diagram with the LAT framework and the MIX model, and (2) the development of
a similarity formula for water analyses. Moreover, existing solutions for common
problems were re-formulated in order to facilitate their application: (3) a new
recipe for the construction of Maucha diagrams, (4) an activity-based procedure
for the calculation of electrical conductivity, with some alternative formulae for the
MAION: water composition
247
activity coefficients, and (5) the introduction of pHsat as an indicator of the

carbonate equilibrium state.
The analysis of a water sample yields a list of numerical data, most of which
represent the amount of a chemical species present in a unit volume of the sample.
Some other numbers apply to physical features of the sample, such as its electrical
conductivity, or its pH. In order to supply a reliable picture of the water quality, the
analysis should at least comprise all major ions (Ca2+, Mg2+, Na+, K+, Cl-, SO4=,
HCO3-) and the other most important plant nutrients (P, N). The MAION method
is suitable for the processing of data concerned with the major ions, pH, and
electrical conductivity. It will be shown that analyses comprising only some of
these parameters can still be profitably used in connexion with a limited number of
more elaborate analyses and even without such a framework.
The first section of this appendix (D.1) comments on current notational
conventions and introduces various aspects of data processing in MAION, among
which a new recipe for the construction of Maucha diagrams. The formulae used in
MAION are presented and dicussed in detail in Section D.2, including a discussion
of activity coefficients, the derivation of similarity coefficients, and the
introduction of pH as a practical indicator of the carbonate equilibrium state.
Section D.3 deals with the interpretation of MAION results in the light of
differences in water chemistry observed in North-West Overijssel and supposed to
be significant also for ecological studies elsewhere. Section D.3 includes an
account of the use of the EC-IR diagram in connexion with the LAT benchmark
analyses and the MIX model for an ecological interpretation of hydro-analytical
data.
Since this appendix deals especially with the MAION method, no attention will be
paid here to several other important aspects of processing and interpretation of
water analyses, such as the presence of N and P compounds and toxic substances,
and the dissolution and precipitation of minerals, with the exception of the calcite
equilibrium.
It must be emphasized that several of the major ions are quite often deliberately
disregarded in water studies (e.g., in most of the official water-quality monitoring
programs in The Netherlands). Often correlations are sought between
concentrations of single constituents and other phenomena. Especially in
biological ecology this correlative approach is often followed, with the intention of
constructing "response models" and "indicator lists", as in Segal (1966b), De
Lange (1972), Seddon (1967), and Hutchinson (1975, Chapter 30). This type of
correlative studies served as the basis for water-quality standards, as in the Dutch
"IMP" policy program for water-quality management. This train of thought will
not be followed in this appendix.
To improve the readability of this appendix,some sample analyses have been
used to exemplify the various aspects of data processing discussed. These
analyses are presented in Table D.1 together with the outcome of the MAION
calculations to be discussed later on. The sample analyses used represent types of
water regarded as very diverse from the ecological point of view. LI-ANG is a
relatively calcium-rich groundwater; AT-W80 is precipitation caught in a
relatively unpolluted inland area of The Netherlands,
248
Table D.1
The water analyses of the samples spanning the LAT framework LI-ANG,
AT-W80, and TH-N70, and the composition of Rhine water at Lobith in 1975
(RH-LOB)
Sample code
Date of sampling 801208
LI(ANG)
1980
AT(W80)
820727
Analytical data:
pH
7.3
4.22
Ca2+
mg/l
115
0.42
Mg2+
mg/l
8
0.20
Na+
mg/l
12
1.60
K+
mg/l
2
0.23
alkalinity
mmol(c)/l
6.6
0.00
Clmg/l
11
2.97
mg/l
13
5.79
SO4=
EC25
mS/m
65.1
5.01
Ionic balance and partial molar(c) fractions:
x= (cat-an)/(cat+an) (%)
-1.2
-8.4
f(c)H+
%
0.0
34.7
f(c)Ca2+
%
82.3
12.1
f(c)Mg2+
%
9.4
9.5
f(c)Na+
%
7.5
40.3
f(c)K+
%
0.7
3.4
%
91.9
0.0
f(c)HCO3f(c)Cl%
4.3
41.0
%
3.8
59.0
f(c)SO4=
Ionic strength and calculated EC25:
Ionic Strength mol(c)/l
0.010
0.00027
ECc
mS/m
65.8
4.27
y= (ECm-ECc)/ECm (%)
-1.0
14.8
Special ratios:
IR
%
94.9
20.0
ECaR
%
52.4
2.5
EClR
%
3.6
12.8
Coefficients of similarity:
rLI
%
100.0
-55.6
rAT
%
-55.6
100.0
rTH
%
30.1
-17.0
rRH
%
43.8
-1.0
Additional data:
mg/l
1.32
N in NH4+
mg/l
0.77
N in NO3-
TH(N70)
1975
RH(LOB)
8.3
420
1400
10480
390
2.0
19100
2640
5200
7.8
82
10
96
7
2.6
178
80
99.6
0.5
0.0
3.5
19.1
75.7
1.7
0.3
90.4
9.2
-0.1
0.0
44.1
8.9
45.1
1.9
28.0
54.1
17.9
0.70
4782
8.0
0.013
101
-1.7
3.7
2.4
79.1
44.9
24.4
38.5
30.1
-17.0
100.0
82.1
43.8
-1.0
82.1
100.0
Analytical data and results of the application of MAION (Section D.2). Further details about these samples are given in
Section D.3.
249
acquired a particular meaning in the MAION method, since they constitute the socalled LAT framework introduced and discussed in Section D.3. RH-LOB
represents water of the river Rhine at Lobith. This water is badly polluted by
human activities. Rhine water is the main source of water in many parts of The
Netherlands during periods of drought.
Notational and conceptual conventions
Electrical conductivity
The electrical conductivity (EC) will be expressed in the SI unit mS/m and always refers to the
standard temperature of 25 C (EC25, see Section D.2). For a comparison with the earlier current
unit S/cm (or mho/cm) it may be noted that 1 mS/m = 10 S/cm. No reduction was applied
for the contribution of [H+].
Concentration units
Most of the analyses used in the present publication have been reported as mass per unit volume
of the pure ionic form of the constituents (1 mg/l = 1 g/m3). For calculations such as those used
in MAION, the concentrations have been expressed as moles of charge per unit volume (1
mmol(c)/l = 1 mol(c)/m3). Ionic strength, however, is used as a dimensionless number, which is
always calculated on the basis of mol/l (see below: 1 mol(c)/l = 1000 mmol(c)/l).
The mole concept
A few words on the "mole" concept may be appropriate here. In older writings the concept is
often restricted to atoms, molecules, or ions. Thus, in a solution of CaCl2, twice as many moles
of Cl- will be present as there are moles of Ca2+. For solutions of ionizing substances, the
"equivalent" concept was introduced to facilitate charge-related computations. Each mole of an
ion was now defined to comprise as many equivalents as to equal its charge number. One mole
of Ca2+ ions in an aqueous solution thus equalled two equivalents of Ca2+. This concept is no
longer recognized in the SI (Systme International d'Units, International System of Units), and
it will not be used here.
The SI defines a mole of a substance as the amount of that substance which contains
Avogadro's number of elementary entities (atoms, molecules, ions, charges, etc.), i.e., 6.02 x
1023. This definition requires a precise statement of the elementary entities the moles refer to
(Schurer & Rigg 1980, p.6). In line with this recommendation the following notations will be
used:
-
mol(c) for moles of charge, i.e., the earlier equivalents. Molar concentrations in mol(c)/m3
are symbolically indicated by the specification of the elementary entities involved, which
indication is surrounded by square brackets:
[Na+], [Ca2+], ..., [1/nXn+], [1/nYn-] mol(c)/m3
Throughout this publication, moles of charge will be used whenever applicable; Ca2+
(halfCalcium) so is a symbol, not an arithmetical formula;
250
mol(a) for atomic, molecular or ionic moles, i.e., the "old" moles. Molar concentrations in
mol(a)/m3 are symbolically indicated by the specification of the elementary entities involved,
which indication is surrounded by square brackets:
[Na+], [Ca2+], [CaCl2], ..., [Xn+], [Ym-], [XmYn] mol(a)/m3
Quantitatively, the relation between concentrations in mol(c) and mol(a) per unit volume is:
[1/nXn+] mol(c)/m3 = n[Xn+] mol(a)/m3
Note that formally the parameter "ionic strength" (Section D.2) is also a molar concentration
(mol/l). The ionic strength moles are neither straightforward moles(a), nor moles(c). In formulae
used in this publication, the ionic strength will be treated as a dimensionless number. The form
of the formula for the computation of ionic strength as presented in Section D.2 requires mol(c)/l
input values for the ionic concentrations.
Partial molar(c) fractions
The ionic composition of a water sample can be expressed by calculating the share of individual
ionic species as a percentage of the total concentration of moles of the same type. These shares
will be called partial molar fractions here. For partial molar fractions, the mol(c) scale has been
used throughout this publication. For cationic species the sum of cations has been used as a total,
and for anionic species the sum of anions, and each fraction is expressed as a percentage:
f(c)Ca2+ = 100 [Ca2+]/[cat] and f(c)Cl- = 100 [Cl-]/[an] (%)
where:
f(c)1/nXn+/[cat]
[an]
[Ca2+]
[Cl-]
is the partial molar(c) fraction of X in %

is the sum of cations in mol(c)/m3
is the sum of anions in mol(c)/m3
is the concentration of Ca2+ in mol(c)/m3
is the concentration of Cl- in mol(c)/m3
Partial molar fractions are used in several types of diagrams, such as triangular diagrams
according to Piper and radial diagrams according to Maucha.
Special ratios and similarity coefficients
The special ratios introduced in this appendix are expressed as percentages with numerical
values between 0 and approximately 100. The numerical value of the similarity coefficient
defined in Section D.2 ranges from -1.00 to +1.00. Similarity coefficients have often been
expressed as percentages also.
Reliability of analytical data
One must bear in mind that the analytical concentrations found in a water sample may well
differ from the representative, "real" concentrations in the water sampled. The following
arguments may account for possible discrepancies:
- The sample may be not representative. Cost factors commonly prevent statistically "safe"
numbers of samples to be collected and analysed;
251
Not all constituents present in the sampled solution are determined. Especially the
concentrations of naturally occurring organic solutes, such as humic acids, are not
specifically assessed;
Changes during transport and storage of samples sometimes considerably alter their physicochemical and biological properties. Improved filtering procedures, specific preservatives and
careful planning of storage can diminish such changes, but they still modify the solution
(Hem 1970, p.86-88);
Errors may still occur in spite of careful procedures;
Results of chemical analyses are given as concentrations of a chemical entity. Several
elements may occur in different forms, however, and it is not always easy to infer the
concentrations of each of them (compare Hem 1970, p.234).
A first stage in data processing, therefore, should involve checks for consistency. MAION
incorporates checks of electroneutrality and of electrical conductivity (Section D.2). When these
checks do not reveal major discrepancies, it is generally assumed that:
- No major errors have been made;
- A solution prepared with the analytical results as a recipe would, in the same reactive
environment of nature, act in the same way as does the sampled solution.
Of course, the checks do not permit any conclusion regarding the sample being representative of
the body of water studied.
The use of partial molar(c) fractions as measures of ionic composition
When a number of water analyses is available, it is possible to compare the analyses and to infer
their possible relationships. Such relationships may exist when the samples reflect different
stages in processes which alter bodies of water with initially equal concentrations, or when
individual portions of such bodies of water come into contact with different environments.
Water samples may also be related when they originate from mixing of the same water types.
Several of these processes may well occur simultaneously and some concentrations may be
more strongly altered by any such process than other ones. Exact calculations may become fairly
complex, and graphical representations can help to arrive at valuable hypotheses. The
computation of partial molar fractions provides the input parameters for the graphical
representation in triangular diagrams according to Piper and in radial diagrams according to
Maucha. These are shortly discussed below.
Triangular diagrams according to Piper
A triangular diagram according to Piper (Hem 1970, p.264-270) comprises three parts: the
cations triangle, the anions triangle, and a quadrangle combining both (Fig.D.1). The overall
concentration of the samples is not reflected in the diagram, although this may be achieved by
choosing symbols with a different surface area to represent different overall concentrations. An
advantage of the Piper diagram is that it enables a quick comparison of many samples in one
diagram. When the surface area of the symbols is made proportional to the overall
concentrations of the water samples, the readibility of the diagram becomes poorer, however.
Each of the three parts of the Piper diagram may be used separately.
In Fig.D.1, the points representing the three "extreme" samples LI-ANG, AT-WTV, and
TH-N70 have been connected to form a contour which encloses the plotting area of all possible,
252
simple mixtures of the waters represented by these analyses. RH-LOB was added as a fourth
point.
Note:
The original graph (upper left corner) showed
At-W80 in a wrong position.
It should be noted that, in Piper diagrams, the
partial molar fraction of f(c)H+ is not included
in the sum of cation concentrations.
The symbols L, A, T, M are used for LI-ANG,
AT -W80, TH-N70 and RH-LOB, respectively.
Fig.D.1
Piper diagram, showing the LAT analyses (LI-ANG, AT-W80, TH-N70), the
analysis of RH-LOB, and certain mixtures
The triangular contours LI-AT-TH-LI delineate the plotting fields for mixtures of LI, AT, and TH. The
dotted triangular areas represent the plotting fields for mixtures of 20% by volume of RH with 80% of
the other samples.
The dotted plotting field represents samples which are composed of 20% by volume of RH-LOB
and 80% of any combination of the three other waters. Note that in the quadrangle and in the
left-hand triangle the very different samples TH an AT plot at nearly the same positions since
the overall concentrations are not involved in the Piper diagram.
Radial diagrams according to Maucha
The radial diagram according to Maucha (1932, p.87-89) is illustrated in Fig.D.2. Maucha refers
to this diagram as a modification of an unpublished suggestion by I. Telkessy, and Schoeller
(1962, p.321-322) calls it the "Telkessy diagram". I will follow Hutchinson (1975, p.557-559),
Moore & Bellamy (1973, p.60), and other authors calling it a Maucha diagram.
253
Fig.D.2
Maucha diagrams for the analyses LI-ANG (c), AT-W80 (d), TH-N70 (e), and
RH-LOB (f), including illustrations of the construction of a Maucha diagram (a,b)
The radius of each diagram is proportonal to the square root of 10log EC25
Each water sample is represented by one individual diagram, thus enabling the visual
comparison of features of the waters concerned. One may either use different radial lengths to
represent different overall concentrations, or use equal radial lengths for all diagrams. Maucha
diagrams are very distinct and suitable for use as map symbols. Fig.D.2 shows the Maucha
diagrams for the four analyses LI-ANG, AT-W80, TH-N70, and RH-LOB, and also
demonstrates the construction of such diagrams, as explained below. This method has generally
been regarded as computationally cumbersome, which has probably held up its wide application.
To overcome this problem, the recipe has been modified, so as to enable manual drawing. This
recipe has been implemented for computer application also.
254
Recipe for Maucha diagrams:

-
If the concentration is not to be reflected, choose an arbitrary radius length r;

Otherwise, define the radius length r of a circle, in such a way that the area will be
proportional to a measure of overall ionic concentration C:
r=kC
where k is a constant. Ionic strength, electrical conductivity, or total moles(c)/m3 may be
used for C. When the overall concentrations are very different, as for the samples
represented in Fig.D.2, it is often more convenient to use the logarithm of the overall
concentration, i.e.,
r=k(10logC)
This has been done in Fig.D.2 and elsewhere in this publication. For one set of diagrams, the
same rule should be used for the choice of r;
Draw the 16 radii of a "compass card", as in Fig.D.2a, and draw the circle with radius r;
Reserve the NNE, ENE, ..., NNW radii for fX= f(c)K+, f(c)Na+, f(c)Ca2+, f(c)Mg2+,
f(c)SO4=, f(c)Cl-, f(c)HCO3-, and f(c)H+, respectively;
Using 1%=r/25, plot the values of fX on the respective radii at the distance
d=(fX)(r/25)
from the centre;

Connect the marks on the radii with the points where the intervening radii (N, NE, ..., NW)
intersect the circle with the radius r (Fig.D.2b) and fill the obtained symbol in with black;
Since different authors may define the radial axes differenty, it is necessary to add indicators
for the meaning of the axes. Especially the NNW compass point is often deliberately used
for different parameters, but not so in this publication.
In the symbol obtained, the area of each black sector is proportional to the respective fX value,
whereas the total black area in different symbols is either a constant, or proportional to the
overall concentration, or proportional to the logarithm of the overall concentration (as in
Fig.D.2), depending on the choice of r. The total black area also equals the area of the inscribed
regular 16-angle in the circle.
The use of special ratios for the comparison of water analyses
The graphical methods described until now can only be used if all major ions have been
analysed. Cost factors often prevent this from being done. Moreover, until about 1965 it was
common practice to determine several major ion concentrations by computation, rather than by
analysis, according to the following procedure:
The anionic concentrations were determined separately. Total hardness in German degrees
( D), and Ca2+ concentration were the only measures of cationic concentration determined
analytically. Since the total hardness (HD) is almost completely attributable to the Ca2+ and
Mg2+ ions, and since 1 D refers to the same amount of moles(c) as there are moles(c) of Ca2+
in a solution of 10 mg/l of CaO,
255
[Ca2+]+[Mg2+] = 0.356 HD (concentrations in mol(c)/m3, HD in D)

This formula enables the computation of the Mg2+ content from the total hardness and the Ca2+
content.
The Na+ and K+ contents in mol(c)/m3 will often approximately equal the apparent surplus of
anions:
[Na+]+[K+] = [SO4=]+[Cl-]+[HCO3-]-[Ca2+]-[Mg2+]
(concentrations in mol(c)/m3)
Because Na+ is present in much larger concentration than K+ in most natural waters, the sum of
the Na+ and K+ contents is reported as mg/l of Na+ in this procedure.
Since it was thought desirable to have a means for the comparison of analyses of unequal
completeness, some special ratios were developed, which summarize the supposedly most
important features of water quality with respect to their concentrations of major ions. These
ratios, called IR, IR*, EClR, and ECaR, will be discussed in detail in Section D.2. They are well
correlated with f(c)Ca2+ or with its complement. Especially the ionic ratio IR, defined as
IR= 100[Ca2+]/{[Ca2+]+[Cl-]} (concentrations in mol(c)/m3, IR in %)
is often used in this publication, in combination with EC25 as a measure of the overall
concentration. A graphical representation was adopted which has 10log(EC25) as abscissae and
IR as ordinates. This EC-IR diagram (Van Wirdum 1978, 1980) enables a quick comparison of
many water samples on the basis of both ionic composition (for which IR is a measure) and
overall concentration (logEC). Fig.D.3 examplifies the EC-IR diagram. As in the Piper diagram
(Fig.D.1), contours have been added which enclose the plotting area of all possible, simple
mixtures of LI-ANG, AT-W80, and TH-N70. Also RH-LOB has been added for comparison. It
may be seen that the EC-IR diagram plots these mixed samples in a much wider area than does
the Piper diagram. This permits the addition of some further characteristic mixing isopleths as
will be discussed in detail in Section D.3.
The use of coefficients of similarity in relation to chosen analyses
Sets of water analytical data are multidimensional data sets. Some of the graphical procedures
discussed before are aimed at selecting representative combinations of parameters enabling the
use of a two-dimensional diagram to represent supposed (dis)similarities between water samples
by the distance between points in the diagram. The MAION similarity coefficients (Section D.2)
were developed to obtain numerical data which reflect such similarities.
Water analyses from a particular region often do not show any appreciable difference, the
less so when the types of water from which they possibly derive their quality are not native to
the region, and thus may have escaped from being sampled. In such cases it is advisable to
include analyses from elsewhere which represent these water types. When available, analyses
256
Fig.D.3
EC-IR diagram, showing the LAT-samples

The connecting lines represent simple mixtures (see also Fig.D.9)
should be taken from the source areas. It has appeared to be more convenient, however, to use a
few typical analyses for reference. If carefully chosen, the similarity to two or three of such
benchmark samples will efficiently summarize most of the features of the sample under study, as
will be explained later. The samples LI-ANG, AT-W80, and TH-N70, none of which represents
water native to North-West Overijssel, have been chosen as such benchmarks. They constitute
the so-called LAT framework explained in Section D.3.
The similarity to these selected analyses can be used in a graphical representation, e.g., a
rTH-rLI diagram, of which Fig.D.4 is an example. This diagram will be explained in detail in
257
Fig.D.4
rTH-rLI diagram, showing the LAT analyses

The connecting lines represent simple mixtures (see Fig.D.12)
Section D.3, after the MAION similarity coefficient has been treated in Section D.2. The
abscissa of any point in the diagram gives rTH, the similarity to TH-N70, while the ordinate
represents rLI, the similarity to LI-ANG. Since, according to the MAION calculations (Table
D.1), LI-ANG and TH-N70 have 30% similarity to each other, TH itself is plotted at the location
(+100, +30) in the diagram, while LI is at (+30, +100). AT-W80 has rTH=-17% and rLI=-50%.
The same contours and admixtures have been indicated as in the examples of the Piper diagram
(Fig.D.1) and the EC-IR diagram (Fig.D.3). The LAT framework has proved to be useful for
water-quality studies in different regions. Values of similarity coefficients can be easily shown
on maps.
258
D.2 Definition of MAION functions and related procedures

The MAION program
The MAION program is the result of a development from manual calculations and graphical
representations towards the use of programmable calculators and computers. At present, the
program is available in the programming code for Hewlett-Packard series 41 programmable
calculators (MAION41), and in the FORTRAN 77 (MAIONF) and Turbo-Pascal 3 (MAIONTP)
source codes for microcomputers. The core functions of the MAION programs are treated in
detail in this section.
The core functions of MAION are:
- Computation of the ionic balance, electroneutrality test;
- Computation of the electrical conductivity, conductivity test;
- Computation of the special ratio IR;
- Computation of the coefficients of similarity rLI, rAT, rTH, and rRH;
- Computation of the pH of saturation with respect to calcium carbonate;
The ionic balance and the electroneutrality test
The most frequently used check on the reliability of water analytical results is the
electroneutrality test, based on the calculation of the ionic balance (Golterman et al. 1978, p.91;
Hem 1970, p.233-234). In MAION, the ions H+, Ca2+, Mg2+, Na+, K+, Cl-, SO4=, and HCO3- (or
alkalinity) are taken into account, and it is tacitly assumed that no other ions are present in any
appreciable amounts. According to Golterman et al., this is justified when pH<9.5. [H+] is
inferred from the pH. Some versions of MAION can take account of other ions. The notation H+
will be used in this publication, rather than H3O+ or still different notations.
The basic concept of the electroneutrality test is that of equivalence of cations to anions.
MAION presents the result of the test by the ratio
x= (cat-an)/(cat+an)
where:
cat
an
x
is the sum total of the cations in mol(c)/m3

is the sum total of the anions in mol(c)/m3
is usually reported as a percentage.
One of the reasons for a deviation of the electroneutrality may be the presence of other ions,
such as NO3-, NH4+, and forms of Fe and Al. As it is often difficult to establish the formulae and
charges of such constituents, their inclusion is not always unbiased. This especially holds true
for solutions containing partly dissociated or complex-forming organic substances and microorganisms, some of which may pass through micropore (0.45 m) filters. In this study, a
somewhat conservative procedure was followed, by using only the quantities of the abovementioned eight ions. If no near-equivalence (-8% <x< 8%) was found, the possible presence of
other ions was investigated, using analytical results (if available), or the measured conductivity
as a criterion. No analyses were discarded if a sound reason was found for the deviation from the
balance. In Table D.1, AT-W80 has x= -8.4%. In this analysis, inclusion of NH4+ and NO3would yield the more satisfactory value x= 1.5%.
259
The conductivity test

Methods provided by the pertaining literature
Several authors dealing with aquatic chemistry have suggested procedures for the calculation of
the electrical conductivity. Some of these procedures are entirely empirical, whereas other
methods are based on physico-chemical theories. Several procedures of both types are
comprised in a method proposed by Stuyfzand (1983). Stuyfzand selected a best fitting
procedure for each of many types of water, and added regression constants in order to obtain an
even better fit. A more theoretical approach has been maintained in MAION, however.
The theory for the calculation of electrical conductivity is reasonably well developed,
especially for very dilute solutions. At conductivities below 10 mS/m, the electrical conductivity
(EC) at a given temperature can be expressed as:
EC= i(Ci
where:
EC
Ci
i
i)
is the electrical conductivity of the solution in mS/m

is the molar concentration of ionic species i in mol(c)/m3
is the molar conductivity of ionic species i at infinite dilution in mS/m per
mol(c)/m3
is the sum total over all ionic species in solution
Consequently, Golterman et al. (1978, p.91-92) advise dilution of water samples with a known
amount of demineralized water (with a known, but very low conductivity) until the electrical
conductivity falls below 10 mS/m. This method is computationally simple, but it has the
disadvantage of an extra manipulation of the sample in the laboratory, introducing at least three
additional numerical data with inherent uncertainties: the volume and the conductivity of the
water added, and the volume of the sample. Certain samples will, moreover, change in
composition upon dilution.
The activity-based method used in MAION
The electrical conductivity of less diluted solutions is influenced by interactions among the ions
in solution, which may vary for different ionic species. Although this is theoretically not quite
permissible, earlier versions of MAION used the stoichiometric Debije-Hckel-Gntelberg
activity coefficients (Stumm & Morgan 1970, p.83-84) to correct for these effects, which results
in the following formulae, which can be used at 25 C and when I < 0.1:
I= 0.5 10-3 i(Ci zi )
10
log(fi)= -0.5 zi2 I /{1+(0.33 ai I)} with ai=3 zi

EC= i(fi Ci
260
i)
where:
I
Ci
fi
ai
EC
zi
i
is the ionic strength (mol/l) of the solution, treated as a dimensionless number

is the concentration of ionic species i in mol(c)/m3
is the activity coefficient applicable to ionic solute species i
is a constant which depends on the effective diameter of ions of species i
is the electrical conductivity of the solution in mS/m
is the signed charge number of ionic solute species i
is the molar(c) conductivity of ionic species i at infinite dilution in mS/m per
mol(c)/m3
is the sum total over all ionic constituents
The values for i at 25 C used in the actual calculations in the computer program MAION have
been expressed as an array of constants (MolConduct[H..halfSO4]) in mS/m per mol(c)/m3 in
the program source code.
The non-extended form of the Debije-Hckel formula for log(f) differs from the one
presented here in the absence of the denominator, i.e., ai=0. The extended formula expands the
applicability to solutions with a greater ionic strength. Gntelberg's formula substitutes 1 for
0.33ai in the extended formula, i.e., ai=3. In more precise applications of the Debije-Hckel
theory published values of ai are used, although they must be regarded as calibration constants.
Most often the values listed by Kielland (1937, see Hem 1970) are used, but slightly different
values are also found in the literature (see Table D.2). Since the errors in single values for ai will
rarely reinforce each other in solutions of mixed salts, a simple and nearly as accurate method is
to substitute zi for 0.33ai, i.e., ai= zi /0.33, or ai=3 zi . In doing so, no new parameters are
introduced, while the accuracy is greater than in case ai=3 is assumed. This enables a
comparison of formulae on the basis of the charge number of ions. The assumption is seriously
in error only for H+. In nature, that ion is important only in very dilute solutions, so that the term
in which ai occurs has no significant influence.
Conductometric activity coefficients
The stoichiometric activity coefficients apply to co-ordinate chemistry. Onsager (see Levine
1978, p.465-466) applied the Debije-Hckel theory to conductivity problems. The application of
the theory to solutions of mixed salts is difficult, however. Onsager's formula was derived for
very dilute solutions of single salts. No activity coefficients are used, but it is possible to rewrite
the formula in order to have explicit conductometric activity coefficients. This enables a
comparison of the effects of ionic strength on the activity coefficients for stoichiometric and
conductometric problems, respectively. The derived conductometric activity coefficients for
conductivity (f') can be calculated as follows:
f'i= 1+ (log fi){2B1+ B2/( zi
i)}
with B1=0.23, B2=6065
where fi is the stoichiometric activity coefficient from the earlier formula with ai=0. The values
B are derived from physical constants relating to the solvent. Since, on average, i=6250 mS/m
per mol/l (note that I is based on mol/l, not mol/m3!), B2/( zi
1/ zi , and the formula can
i)
be simplified into:
f'i= 1+ (log fi)(0.46 + 1/ zi )
261
Table D.2 Numerical values of ai in the Debije-Hckel theory

Solute species
Formula, author
Debije-Hckel:
Extended Debije-Hckel,
Gntelberg:
Kielland:
various authors:
proposed here:
K+
ClNO3NO2-
Na+
HCO3-
3
3
3-4
3
3
4
3-4.5
3
SO4=
Ca2+
Fe2+
Mg2+
Al3+
Fe3+
H+
3
4
4
6
3
6
5-8
6
3
8
8
6
3
9
9
9
3
9
3
Mn2+
Some species not considered in standard MAION have been included for reference
Robinson & Stokes (1970) suggest the following solution:

i=
and
i-
{(B1zi2
i+B2
zi )(I)/(1+0.33aiI)}
EC= i( iCi)
where i is the actual molar(c) conductivity of species i in the solution involved, (mS/m per
mol(c)/l).
It appears that this formula complies with the conductometric activity coefficients if the
stoichiometric activity coefficients are determined according to the extended Debije-Hckel
formula with tabular values (Table D.2) for ai. It is assumed that the extended formulations for fi,
specifying the effect of solute interactions, may be applied here as well, since fi is the only place
in the formula where such interactions are considered. Fig.D.5 shows the effects of some
alternative formulations.
From Fig.D.5 it appears that the application of stoichiometric activity coefficients suggests a
stronger concentration effect than appears from the use of conductometric ones, the difference
being significant for ionic species with charge number 2 and higher. In recent versions of
MAION conductometric activity coefficients are used for the calculation of electrical
conductivity, while stoichiometric activity coefficients are maintained for the computation of the
carbonate equilibrium state.
Compensation for temperature differences
The electrical conductivity varies with the temperature. A useful formulation (Levine 1978,
p.468) is:
EC25 = (ECt) eb(25-t)
where:
EC25
is the electrical conductivity at 25 C in mS/m
ECtis the electrical conductivity at t C in mS/m
b
is a temperature coefficient in K-1
262
Fig.D.5a-c
Activity coefficients for ions with charge numbers 1-3
263
The value b= 0.02 K-1 is correct for most ionic compositions. Laboratory experiments not
reported here suggest that this value is also sufficiently accurate for very dilute and acid
samples.
Method of evaluation
As a standard procedure, the eight ions involved in the electroneutrality test were included in the
calculation of electrical conductivity. Calculations on the basis of the above-mentioned ions
allow for a reasonable check on the substantial presence of other ions. A formula is used in
MAION for the evaluation of the conductivity test by a quantity
y= (ECm-ECc) /ECm
where:
ECm
ECc
is the measured EC25 (mS/m)

is the calculated EC25 (mS/m)
For ECm, laboratory values are preferred to field measurements, since the ionic concentrations
are also determined in the laboratory. Due attention is paid to possible causes of large values for
y. If y is positive, other possibly determined ions may be taken into account. Allowing for NH4+
and NO3- would reduce y for AT-W80 in Table D.1 from 14.8% to 6.0%. Even for TH-N70
(I=0.7!), which is at the high end of the range of applicability of the Debije-Hckel theory, the
use of the conductometric activity coefficients results in a relatively small y=8%. (P.S.: In this
paragraph, the original publication suggested big values for y, not corresponding with those
given in Table D.1. That text had been falsely retained from the 1987 prepublication, which did
not yet use the conductometric coefficients of activity.)
The Ionic Ratio and related quantities
A special function of MAION is the calculation of the Ionic Ratio (IR), defined by (Van
Wirdum 1980):
IR= 100[Ca2+] /{[Ca2+]+[Cl-]}
(concentrations in mol(c)/m3, IR in %)
As shown before (Fig.D.3), IR together with 10 log(EC25) provides a useful coordinate system for
the arrangement of water samples. The IR concept was developed as a composite parameter to
characterize the ionic composition of the type of water concerned. The purpose was to obtain a
parameter, preferably well-correlated with f(c)Ca2+, and computable on the basis of a
minimum number of parameters. Especially Cl- was taken into account, as it can be easily
determined and it is hardly influenced by in situ processes. Further considerations (Van Wirdum
1978) were that, in The Netherlands:
-
264
Both f(c)H+ and f(c)K+ are usually small;

f(c)Mg2+ is fairly constant (17 5%);
f(c)SO4= and f(c)HCO3+ may be strongly affected by the action of biochemical processes
and by sampling conditions, so that their concentrations are less readily interpretable;
Variation of [Na+] is strongly correlated with variation of [Cl-];
In older analyses, [Na+] and [K+] have often not been individually determined (see Section
D.1). Their sum often more or less equals [Cl-].
The partial molar(c) fraction of Ca2+,

f(c)Ca2+= 100[Ca2+] /{[Ca2+]+[Mg2+]+[Na+]+[K+]} (all concentrations in mol(c)/m3)
is approximately equalled by the following expression:
f(c)Ca2+ (100-a)[Ca2+] /{[Ca2+]+[Cl-]} with a=17
f(c)Ca2+ 0.83 IR and IR
or:
1.2 f(c)Ca2+
Fig.D.6a shows the nearly linear relationship between the partial mole(c) fraction of Ca2+ and
IR for a variety of water samples from different Dutch localities. These samples will be used
further in Section D.3, where they are listed in Table D.3. It appears that the value used for a
does not exactly describe the relation in this limited example. A RIN archive, comprising ca
5000 water analyses, yields the result shown in Fig.D.6e.
The use of total hardness to approximate IR, yielding IR* and IRha
Although Cl- is almost always included as a parameter in water analyses in The Netherlands,
Ca2+ is not. There remain some possibilities to approximate IR by means of other data. Two
different cases will be considered. The first especially applies to most of the older water
analyses, where Ca2+ and Mg2+ have been determined together as "total hardness" (HD),
expressed in The Netherlands in German degrees ( D). As stated in Section D.1,
[Ca2+]+[Mg2+] = 0.356 HD (concentrations in mol(c)/m3, HD in D)
By accepting that the partial mole fraction of Mg2+ is rather constant and approximately 17%,
a ratio IR* can be defined as follows:
IR*= 35.6 HD /{0.356 HD + [Cl-]}
or:
IR*
100a + (1-a)IR with a=0.17
17 + 0.83 IR and IR IRha = 43 HD /{0.36 HD + [Cl-]} -20
The relationship between IR* and IR is illustrated in Fig.D.6b and Fig.D.6f.

The Cl-- and Ca2+-based conductivity ratios EClR and ECaR, yielding IR Cl and IRCa
The second case to be considered is a real poor man's approach for cases where only Cl- or Ca2+,
and electrical conductivity have been determined. A ratio EClR is defined as:
265
Fig.D.6a-d Diagrams, showing the correlation between IR and f(c)Ca2+ (a), IR* (b), EClR
(c), and ECaR (d), respectively
The samples used for these diagrams are documented in Table D.3. The relations shown with straight
lines are the ones derived in the text, rather than regression lines. Note that, in the IR-EClR diagram (c),
acid samples of low concentration (bog, poor fen) have lower EClR than described by the "theoretical"
relation, because of the contribution of [H+] to EC25; rather concentrated samples show higher EClR,
since the "theoretical" relation does not take the effects of concentration into account.
266
Fig.D.6e-h Diagrams, showing the correlation between IR and f(c)Ca2+ (e), IR* (f), EClR
(g), and ECaR (h), respectively
More than 5000 samples from archives of the Research Institute for Nature Management are represented
in these diagrams. Most of these samples were taken from surface waters, many of which are influenced
by Rhine water (see Fig.D.10, D.14), so that they do not represent an unbiased group. The diagrams
strongly indicate, however, that there is no reason to reject the relations derived in the text.
267
EClR= 100[Cl-]
where
EClR
EC25
[Cl-]
/EC25
is a dimensionless ratio, expressed as a percentage

is the measured electrical conductivity at 25 C (mS/m)
is the concentration of Cl- (mol(c)/m3)
is the molar(c) conductivity of Cl- at infinite dilution and at 25 C (7.6 mS/m per
mol(c)/m3)
In dilute solutions, this ratio is approximately the share of Cl- ions in the total conductivity. It is
expressed as %. EClR is less precisely correlated with IR than is IR*. As a rough approximation
EClR
c(100-IR) with c=0.55, and IR
IRCl = 100- 1380 [Cl-]/EC25
The relation between IR and EClR has been illustrated in Fig.D.6c and Fig.D.6g.
EClR has a Ca-based counterpart:
ECaR= 100[Ca2+]
/EC25
where
is the molar(c) conductivity of Ca2+ at infinite dilution and at 25 C (5.95 mS/m
per mol(c)/m3)
For ECaR the approximate relations
ECaR p IR with p=0.45, and IR IRCa = 1320 [Ca2+]/EC25
can be derived. The relation between IR and ECaR has been illustrated in Fig.D.6d and
Fig.D.6h.
It must be emphasized that IR and related parameters are especially informative in combination
with a measure of the total concentration, for which the electrical conductivity (EC25) is
recommended.
The MAION similarity coefficient
The MAION similarity coefficient is aimed at a simple numerical comparison of analyses, very
much like a visual comparison of graphical representations.
In MAION, each analysis is represented by a set of numerical values. Let X and Y be two
water analyses, and xi and yi the numerical values associated with the i-th of n features in
samples X and Y, respectively. The general correspondence between both samples is expressed
by a quantity r(X,Y) which is similar to the product-moment coefficient of correlation as
produced by the computational formula (Spiegel 1972, p.245)
r(X,Y)= (n ixiyi- ixi iyi) /{(n ixi2-( ixi)2) (n iyi2-( iyi)2)}
268
The different features should be scaled in such a way that none of them dominates the others. In
MAION, the concentrations of the eight ions mentioned before are used in mol(c)/m3. To this set
of data, the square of the electrical conductivity in mS/m at 25 C is added, divided by 1000, as a
scaled measure of the total concentration. Hence n=9. Note that r(X,Y) does not change when
every xi is replaced by k1xi and every yi by k2 yi, k1 and k2 being non-zero constants. As a
consequence, the nine above-mentioned features may be divided by a constant k, which is
different for each analysis, and proportional to EC25:
k= {(EC25)2 /1000} = EC25 /32
This leaves an array of features comprising eight different ionic concentrations, each relative to
the measure of overall concentration k, and k itself. This array, or feature vector, is discussed
further in Section D.3.
One must bear in mind that the similarity coefficient (r(x,y), as defined here, differs from the
correlation coefficient in statistics in requiring that the n observations (features) relate to the
same parameters for all computations. Each of these parameters has its own chemical meaning.
The use of this similarity coefficient in connexion with the pre-selected analyses LI-ANG,
AT-W80, and TH-N70, the LAT framework, will be explored in the next section, where this use
is introduced on the basis of the EC-IR diagram and on the basis of statistical and theoretical
considerations.
Saturation with respect to calcium carbonate
The carbonate equilibrium is an important factor as a pH-buffer system in natural waters,
determining the availability of carbon as CO2 and HCO3-. Moreover, there is some evidence that
the state of the carbonate system is indicative of the availability of various forms of phosphate in
the solution. The formulation in MAION was derived from the treatment by Kelts & Hs (1978)
and it is consistent with Kleijn (1986).
Calcite is considered to be the most important calcium carbonate mineral in the present
context. The saturation of an aqueous solution with respect to calcite can be determined by the
temperature, the pH, and the activity of Ca2+ and HCO3- ions. The activity of Ca2+ and HCO3ions is computed from the analytical concentrations and the ionic strength of the solution with
the Debije-Hckel-Gntelberg activity coefficients applicable to chemical equilibria, as
discussed earlier in relation to the MAION conductivity test. The state of the solution can be
characterized in several ways. A saturation quotient, or its logarithm, relating the saturated
concentration of calcium ions to the actual one, is often used for this purpose. Also, the
equilibrium temperature is a good candidate, since then it is possible to see whether saturation
conditions might apply during a day-and-night cycle. Usually, however, pH is the factor that
varies most and that, according to its logarithmic nature and serious problems for field
determination, is the weakest point in the formulation. For this reason the saturation pH at T C is
reported, where, in MAION, T=10:
pHsatT= K-log{f2[Ca2+]}-log{f1[HCO3-]}-0.015(T-10)
where
T
K
f1
f2
is the temperature of the solution in C

is a constant defined below; its numerical value is 8.43
is the activity coefficient for ions with charge number 1
is the activity coefficient for ions with charge number 2
269
[X]
is the molar(c) concentration of X in mol(c)/m3
The constant K combines some conversion factors to accomodate the formula for mol(c)/m3
units, with the equilibrium constant for calcium carbonate and the dissociation constant for the
bicarbonate ion, according to values cited by Kelts & Hs (1978, p.300). For an increase of the
temperature with 1 C, in the range from 0 to 30 C, the value of pHsat decreases 0.015. The pH
is understood as the negative logarithm of the hydrogen ion activity, rather than of the
concentration.
D.3 The LAT framework
Introduction
The purpose of this section is to explain the LAT framework, which has been mentioned in the
foregoing sections and which consists of the benchmark analyses LI-ANG, AT-W80, and
TH-N70, introduced in Table D.1, and graphically depicted in various figures in this chapter.
For illustrative purposes, several further analyses are used in this section (Table D.3). These
analyses were selected to show a wide variety of chemical compositions of natural waters,
mainly from North-West Overijssel. The LAT reference analyses have been included. It was
only while using and further developing the MAION method that the need for a single set of
well-defined benchmark analyses arose. The choice of such benchmarks, of cause, relies on the
definition of similarity and dissimilarity between water analyses and the weight given to the
various physico-chemical parameters. Although the illustrations in the Sections D.1 and D.2
empirically show some of the decisions made, more detailed arguments in favour of the EC-IR
and similarity methods are given below.
Statistical evidence for the importance of EC and IR
The fundamental question is here whether there is any statistical evidence for EC and IR to be
chosen as representative parameters for the characterization of water analyses. This question has
been investigated with a variety of statistical means, such as, among others, principal component
analysis (Van Wirdum 1981). For the present purpose, a simpler approach is followed, which
leads to the same conclusions. This is the approach by means of determinant analysis.
Determinant analysis
Sheffield (1985) has shown that selection of two or three parameters which explain most of the
statistical variation in multivariate data sets, can be done by determinant analysis of sub-sets of
elements of either the matrix of covariances, or the matrix of correlations. Such selection will
only be successful when there is much redundancy in the numerical data. This method provides
a formalized and quantitative supplement to the method of manually rearranging the relevant
matrices in order to group correlated parameters (McIntosch in Whittaker 1973, p.157-191),
although it is only practicable for relatively small submatrices.
The formulae for the computation of the determinant of a 2x2 covariance matrix, and a 2x2
correlation matrix, respectively, are simple:
Detcov = var(x)var(y) - (covar(x,y))2 and detcor = 1- (cor(x,y))2
270
Table D.3 Analytical data and MAION results for several water samples, mainly from NorthWest Overijssel
Sample date
pH
yymmdd
LI-ANG
LI-HDU
AT-W80
TH-N70
RH-LOB
B1a
B1b
B2a
B2b
B5a
B5b
B5c
B5d
*B4a
*B4b
13.13
26.10
30.12
30.14
30.9
39.9
4.14
46.13
LM187a
LM187b
LM187c
M02-04
M05-04
M09-16
M18-04
STOB-a
STOB-b
STOB-b
STOB-b
STOB-b
STOB-b
STOB-c
STOB-c
STOB-c
STOB-c
STOB-c
STOB-d
STOB-d
STOB-d
STOB-d
STOB-d
STOB-d
STOBdd
801208
1969
80mean
820727
75mean
801119
801119
800918
800918
801119
801119
801119
801119
821026
821026
741107
780309
790412
790412
790412
810313
721130
821021
801119
801119
801119
811021
811021
811021
811021
820617
800620
800821
801016
801215
810217
800620
800821
801016
801215
810217
800620
800821
801016
801215
810217
820617
820617
7.3
8.3
4.2
8.3
7.8
5.9
6.2
7.2
8.0
7.5
7.0
6.9
6.9
6.9
7.0
7.6
7.0
6.7
6.4
6.7
6.6
7.0
7.9
6.7
6.7
6.7
3.4
3.6
4.6
3.8
7.7
7.4
7.0
6.5
6.5
6.5
4.7
6.5
6.0
6.7
6.0
4.1
4.4
4.5
5.4
5.5
4.0
5.7
Concentrations of ions
Ca
Mg
Na
mg/l
mg/l
mg/l
K
mg/l
Cl
HCO3
mg/l mg/l
115.0
8.0
12.0
2.0
11.0
33.0
4.1
11.5
1.0
12.0
0.4
0.2
1.6
0.2
3.0
420.0 1400.0 10480.0 390.0 19100.0
82.0
10.0
96.0
7.0 178.0
13.0
9.0
16.0
8.0
30.0
20.0
12.0
25.0 13.0
50.0
105.0
18.0
24.0
2.0
49.0
105.0
11.0
26.0
1.0
51.0
63.0
17.0
75.0 25.0 130.0
155.0
28.0
85.0
4.0
78.0
130.0
16.0
94.0
3.0 180.0
305.0 140.0 1350.0 42.0 2750.0
122.0
8.7
51.0
1.5
74.0
106.0
8.0
12.5
1.3
28.0
32.0
10.0
35.0
4.0
64.0
39.0
14.0
42.0
5.0
75.0
39.0
11.0
30.0
7.0
53.0
44.0
18.0
41.0 11.0
62.0
38.0
11.0
32.0
7.0
52.0
31.0
7.0
15.0
6.0
22.0
60.0
24.0
31.0
7.0
37.0
56.0
12.0
58.0
7.0 114.0
108.0
13.5
83.0
3.0 115.0
137.0
11.5
180.0
4.1 330.0
275.0
20.0
165.0
4.1 540.0
1.5
2.3
4.5
1.2
9.0
1.0
1.6
4.0
1.1
7.5
0.2
3.0
18.5
6.1
37.0
1.0
0.8
5.5
0.8
9.0
81.0
9.9
39.0
2.7
74.0
48.9
7.7
34.0
5.8
73.1
42.4
6.2
27.0
3.2
52.1
67.9
10.4
38.0
3.7
77.1
30.9
3.9
16.5
3.5
33.1
28.9
4.2
19.0
2.4
31.0
3.5
1.3
21.0 15.0
40.0
26.9
4.6
14.5
2.8
29.0
20.5
3.5
16.5
1.8
30.0
6.0
1.1
6.0
2.7
9.5
14.0
2.2
10.5
1.9
19.0
2.0
0.8
14.0
5.1
25.0
2.5
1.0
5.0
4.9
11.0
3.0
1.1
11.0
4.0
16.0
1.0
0.3
4.0
1.6
5.5
5.5
1.4
6.0
3.6
10.5
3.0
1.8
9.5
2.6
14.5
13.5
3.5
15.0
2.5
16.0
SO4
mg/l
400.0
13.0
119.0
10.9
0.0
5.8
122.0 2640.0
158.6
80.0
19.0
60.0
34.0
68.0
374.0
4.0
352.0
3.0
290.0
30.0
332.0 350.0
442.0
30.0
380.0 300.0
435.0
5.0
350.0
7.0
103.7
27.0
91.5
76.0
67.1
79.0
54.9 125.0
79.3
72.0
42.7
76.0
42.7 220.0
134.2
51.0
440.0
9.0
415.0
13.0
455.0
15.0
0.0
23.0
0.0
15.0
4.0
11.0
0.0
8.0
220.0
39.0
125.0
28.0
105.0
27.0
200.0
21.0
82.0
13.0
69.0
35.0
2.0
30.0
76.0
19.0
36.0
30.0
18.0
9.0
34.0
14.0
0.1
14.0
0.1
13.0
1.0
18.0
1.0
5.0
8.0
19.0
0.0
18.0
62.0
7.0
EC25
mS/m
65.2
22.5
5.0
5200.0
99.6
27.6
37.5
71.8
68.5
87.2
127.0
121.4
913.1
88.3
61.8
45.5
51.0
45.0
59.0
45.0
30.0
67.2
64.0
97.2
154.6
215.3
19.3
13.3
18.2
8.6
60.7
46.5
38.2
57.4
28.7
28.1
21.1
23.9
19.3
9.6
15.5
13.5
8.4
9.9
5.3
8.8
11.0
16.0
IR
%
x
%
Similarities
y pHsat LI AT TH
%
r% r% r%
RH
r%
95 -1 -1 7.07 100 -56 30 44

83 0 -13 8.06 92 -46
-6 22
19 -8 13 ?.?? -56 100 -17 -2
4 1 8 7.66 30 -17 100 82
45 0 -2 7.64 44 -1 82 100
43 -2 -3 9.26 -8 58 16 38
42 1 -6 8.84
6 43 38 66
79 2 -2 7.14 98 -56 43 58
79 0 -2 7.16 98 -53 40 59
46 -3 -4 7.48 66 -29 75 92
78 -4 -14 7.10 58 -5 83 83
56 -4 -3 7.02 70 -30 82 93
16 -2 4 6.97 29 -14 100 86
75 -1 -0 7.03 93 -47 57 73
87 -1 -2 7.16 100 -55 30 47
47 -0 3 8.16 63 -12 47 85
48 -1 -12 8.15 46 16 51 87
57 1 -8 8.28 49 20 45 77
56 5 -5 8.33 29 27 64 79
56 1 -7 8.21 56 13 46 80
71 1 -12 8.55 42 30 10 35
74 2 -7 8.33 24 29 56 56
47 0 -11 7.84 54 -3 60 94
63 -2 -5 7.08 82 -40 67 84
42 -2 -7 7.04 48 -16 88 98
47 -1 -9 6.75 46 -17 92 96
23 10 -14 ?.?? -45 75 34 17
19 10 -11 ?.?? -60 89 11 6
1 -4 7 11.70 -40 47
3 39
16 12 -24 ?.?? -60 84 -14 12
66 1 -12 7.47 90 -33 43 73
54 0 -10 7.91 74 -18 35 77
59 2 -8 8.03 79 -18 23 67
61 1 -8 7.58 87 -35 43 75
62 2 1 8.25 84 -29 11 54
62 -1 -7 8.36 74 -2
8 54
13 -5 -6 10.78 -43 63
4 38
62 -1 -12 8.34 81 -23
-3 42
55 1 -24 8.77 42 26 -13 42
53 -2 10 9.56 49 10 -26 28
57 0 -2 8.94 61 1 -14 40
12 -1 -11 12.30 -45 65
-7 32
29 1 -5 12.19 -51 78 -22 11
25 1 -18 11.12 -44 76 -19 20
24 3 29 11.56 -39 66 -21 22
48 -5 -12 9.95 -6 65 -30 11
27 6 -22 ?.?? -51 84 -17 18
60 2 -10 8.70 72 -30 -15 26
fractions
Ca Mg Cl
% % %
82
66
11
3
44
28
29
67
72
37
56
54
18
67
81
39
38
45
38
43
53
46
43
53
44
61
8
8
1
9
61
52
54
56
58
54
11
55
49
42
50
10
21
18
17
37
17
40
10
14
10
19
9
33
29
19
13
17
17
11
13
8
10
21
23
21
26
21
20
31
15
11
6
7
22
21
20
12
12
14
13
14
12
13
7
16
14
13
13
7
14
11
9
16
17
17
4
13
41
90
54
35
42
18
20
41
15
39
86
22
12
44
41
35
33
34
21
16
50
30
57
66
35
40
78
60
32
44
39
37
37
32
63
33
41
36
39
71
53
54
56
36
52
28
Coding of samples: Chapter 5 (groundwater samples, WRNET samples), Chapter 9 (Stobbenribben), the present chapter
(benchmarks); M-codes are from the Meerstalblok bog area
271
The selection of the 2x2 submatrix with the largest determinant, i.e., of the two "best"
parameters can be done by a visual inspection of the matrices. Determinant analysis and manual
matrix rearrangement in most cases provide a sound basis for a reasonable decision with regard
to the selection of especially informative parameters. Both methods suffer less from statistical
uncertainties than such alternatives as are principal component analysis and factor analysis.
The application of these methods is not reported in detail here. EC, or Cl-, and Ca2+ do not
always form the statistically most informative combination. If such factors as pH and K+, which
vary within a small range, and SO4= and alkalinity, which are often less representative for the
body of water sampled (Section D.2), are disregarded, Ca2+ and Cl- are always indicated as one
of the most powerful pairs of parameters, however. When IR is used instead of Ca, the
indication is still stronger.
As a result of increasing interest in the method described, also Hoogendoorn (1983, part 4)
discussed several aspects of the EC-IR method and its applicability to some 3000 analyses of
ground waters from the eastern part of The Netherlands. Judging from information obtained
through a factor model, Hoogendoorn concludes that the choice of Ca2+ and Cl- would have been
a very good one for the characterization of his analyses.
The LAT framework
From the use of the EC-IR diagrams, it appeared that most water analyses plot within the area
bounded by the curved lines LI-AT-TH-LI in Fig.D.7 (after Van Wirdum 1980). This result
suggested not only that certain combinations of EC and IR are rare or absent in The Netherlands,
but also that the cluster of analyses can be described by three more or less extreme points. The
locations of these points have EC and IR values which, in the diagrams, appear to be
characteristic of rain water (both EC and IR low), sea water (EC high, IR low), and a certain
type of groundwater (IR high, EC 50-100 mS/m). If these points are accepted as a framework, it
would be possible to characterize water analyses with respect to EC and IR as more or less
resembling atmotrophic rain water (AT), lithotrophic high-IR groundwater (LI), or
thalassotrophic sea water (TH).
The first two characters of the codes of these analyses represent the type of water:
- LI: Lithotrophic, i.e., borrowing its chemical character from an intensive contact with the
lithosphere;
- AT: Atmotrophic, i.e., borrowing its chemical character from atmospheric water;
- TH: Thalassotrophic, i.e., borrowing its chemical character from oceanic water (Gr.
(thalassa): ocean, sea).
After some try-outs, a first set of reference points was chosen which was later replaced by the
analyses LI-ANG, AT-W80, and TH-N70.
LI-ANG was selected from the analyses for the groundwater monitoring program by the
National Institute of Public Health and Environmental Hygiene (R.I.V.M.) as an example of an
accurate analysis of water without any manifest sign of pollution, and with a very high IR value,
combined with an average EC25. LI-ANG was sampled at Angeren (Gelderland), station nr. 272
in the groundwater monitoring program, 24 m below soil surface (date: December, 8, 1980). It
replaced LI-HDU which is less extreme with regard to IR. This analysis has been included in
several illustrations in the present chapter.
272
Fig.D.7
A first reconnaissance of the area of interest in EC-IR diagrams: contour of the

plotting field for all samples in Van Wirdum (1980), as compared with the MIX
contour
AT-W80 is the weighted mean composition of rain water collected at Witteveen (Drenthe)
during 1980. It was chosen as a representative example after the chemical characteristics of rain
water at this station had been investigated over the period 1972-'80, as reported by the Royal
Netherlands Meteorological Institute (K.N.M.I.). The earlier benchmark AT-WTV represented
the weighted mean rain water composition at Witteveen during 1973-'74, but in those years
certain factors could only be assessed less accurately.
The analytical data of TH-N70 were kindly supplied by Rijkswaterstaat, North Sea
Directorate, as a representative analysis from the North Sea monitoring program, 70 km from
the coast at Noordwijk, station nr. N-70 (date: July, 27, 1982). Several parameters were
especially analysed for the present purpose. Its predecessor in MAION, TH-XXX was derived
from various data in the literature.
273
Fig.D.8
EC-IR diagram for the samples documented in Table D3

Note that the WRNET samples (w) depicted were selected to fill in some of the otherwise open space in
the diagram, rather than to represent the composition of the boezem water in "De Weerribben"
A fourth analysis often used in connexion with the LAT framework is RH-LOB, the mean
composition of Rhine water at Lobith during 1975. This analysis was included since it is
representative of the river water which is distributed over a large area in The Netherlands during
periods of drought. This water also reaches the boezem waters of North-West Overijssel in many
summers.
Both atmospheric and lithospheric water may be quite different from AT-W80 and LI-ANG,
but up to now, such different types of water have appeared to fall more or less in between the
LAT reference points in EC-IR diagrams. The analyses presented in Table D.3, which include
the LAT analyses, have been plotted in an EC-IR diagram in Fig.D.8.
274
The LAT framework and the hydrological cycle

The LAT framework can be related to a model describing the changes in water chemistry during
the hydrological cycle. Two very simple models were introduced by Van Wirdum (1980,
p.135-137) which lump several processes and parameters. Of these, the mixing model (MIX)
was used already in Fig.D.1, D.3, and D.4. This model will be considered in somewhat greater
detail here.
Due to coastal effects and very varying atmospheric pollution rain water samples differ in
composition according to the location of the measuring station. The composition of groundwater
depends on the properties of soils and substrata. Groundwater samples differ considerably in
total hardness, and may be influenced by such processes as dissolution or precipitation of
CaCO3, ion exchange, diffusion of connate salt from deeper marine formations, etc. In general,
the three benchmarks LI-ANG, AT-W80, and TH-N70, and their combinations, embrace almost
all water samples from North-West Overijssel and from elsewhere in The Netherlands
(Fig.D.10). An exception are deep brackish waters from Tertiary and older formations, which
are influenced by remnants of their original (connate) salinity and by the dissolution of solid
gypsum present in these formations (Fig.D.7). Moreover, some water samples have somewhat
higher EC25 than the mixtures of LI-ANG and TH-N70 with the same IR. Most of these samples
originate from places which have been under the influence of brackish waters in historical times.
These samples may have higher Cl-, SO42- or Mg2+ concentrations than a "typical" fresh
groundwater with similar IR.
Series formed by actual water analyses
It may be concluded that the LAT framework can be used to delineate the plotting area for actual
water analyses from The Netherlands in the EC-IR diagram. Repeated sampling at the same
locations, or sampling of different bodies of water at different locations in an interconnected
hydrological system, may yield more or less linearly extended clusters of points in the EC-IR
diagram, as shown in Van Wirdum (1980). By reference to the LAT framework, such clusters
may suggest developmental or mixing relations, which can be studied further with a more
complete set of analytical data and with models of water chemistry. As proposed by Van
Wirdum (1981, p.115), the clusters can be indicated as:
- Atmocline for clusters of samples that only slightly differ from AT;
- Atmo-lithocline for clusters of samples which stretch along the AT-LI contour line in the
diagram;
- Various other clines of this kind.
If one wants to distinguish between the influence of thalassotrophic water and the influence of
pollution, as traced in the MIX model by the mixing with RH-LOB, the clusters which are
supposed to have RH-LOB as an end point or as a centre of gravity, can be named molunoclines
(Gr.
(moluno): to defile, pollute, become vile). A cluster with the line AIR (Fig.D.9) as
a main axis would thus be named an atmo-molunocline. The dotted area in Fig.D.9 may
represent a molunocline.
Applicability at the global scale
An interested reader of earlier publications about the EC-IR method has drawn my attention to a
paper by Gibbs (1970), who considered the mechanisms controlling world water chemistry.
Gibbs plotted the weight of total dissolved salts (logarithmic scale) in various rain, river, lake,
275
Fig.D.9
EC-IR diagram, showing the results of application of the MIX model

Explanation in the text
and ocean samples from all over the world, against the weight ratio [Na]/{[Na]+[Ca]}, and, as
an alternative, against the weight ratio [Cl]/{[Cl]+[HCO3]}. This method yields results which
are very similar to the ones obtained with the EC-IR method, as can be easily understood from
the description of the ionic ratio and related quantities in Section D.2. The LAT framework
would fit Gibbs' analyses quite well, although LI-ANG is somewhat more concentrated than his
"lithotrophic" river waters. It is also apparent, from Gibbs' paper, that most samples of surface
waters, at the global scale, fit in either an atmo-lithocline, or a litho-thalassocline, and this is
explained as follows (Gibbs 1970):
276
Fig.D.10
EC-IR diagram, showing the plotting position of more than 5000 analyses of
(mainly surface) waters in The Netherlands (RIN archives)
The compositions of the world's surface waters plot as two diagonal lines. This ordered
arrangement can serve as a basis for discussion of the several mechanisms that control world
water chemistry.
The first of these mechanisms is atmospheric precipitation. The chemical composition of
low-salinity waters are controlled by the amount of dissolved salts furnished by precipitation.
...
The precipitation control zone is one end-member of a series. The opposite end-member of
this series is comprised of waters having as their dominant source of dissolved salts the rocks
and soils of their basins. This grouping defines the second mechanism controlling world
water chemistry as rock dominance. The waters of this rock-dominated end-member are
more or less in partial equilibrium with the materials in their basins. ...
The third major mechanism that controls the chemical composition of the earth's surface
waters is the evaporation-fractional crystallization process. ..., this mechanism produces a
series extending from the Ca-rich, medium salinity (freshwater), "rock source" end-member
277
grouping to the opposite, Na-rich, high-salinity end-member. The rivers and lakes in this
group are usually located in hot, arid regions. A number of these rivers ... show evolutionary
paths ... starting near the Ca or "rock source" end-member with changes in composition
toward the Na-rich, high-salinity end-member as the rivers flow toward the ocean. This
change in composition and concentration along the length of these rivers is due to
evaporation, which increases salinity, and to precipitation of CaCO3 from solution, which
increases the relative proportion of Na to Ca. ...
The Na-rich, high-salinity, end-member components are the various seawaters of the earth
whose compositions cluster near the Na-rich axis.
Gibbs even concludes that, in his diagrams, the axes showing data for total dissolved salts, could
be replaced with precipitation or runoff data without materially altering the interpretation.
According to him, second-order factors, such as relief, vegetation, and composition of material
in the basin dictate only minor deviations within the zones dominated by the three prime factors
(atmospheric precipitation, rock dominance, and the evaporation-crystallization process).
It is apparent that, in North-West Overijssel, the litho-thalassocline is not caused by the
evaporation-crystallization process, but rather represents a litho-molunocline caused by water
pollution. This can be proved with older analytical data, e.g. those by Gunning in Vereniging
(1870). The composition of Rhine water about 1860 was characterized by an IR of 88% and an
EC25 (calculated) of about 30 mS/m, compared to 45% and 100 mS/m, respectively, in 1975.
MAION similarity: An extension of the EC-IR characterization
Since the variety of EC and IR values found in natural waters can be conveniently delimited by
the LAT framework, the latter may also be capable of spanning the associated variety of other
parameters of water quality. The MAION similarity coefficient (Section D.2) was used in the
mixing model with the LAT analyses in order to investigate this point. Before the results of this
investigation are being presented, some attention will be paid to the way the MAION similarity
formula handles a water analysis.
Visualization of the MAION feature vector
The nine parameters used in the MAION similarity computation were defined in Section D.2.
These parameters may be looked upon as a vector, the feature vector, in a multi-dimensional
space.
The MAION similarity coefficient was defined with the product-moment formula known
from correlation analysis. This formula incorporates a centering of the numerical data around 0,
and a standardization of the variances to the value 1, i.e.,
xi'= (xi-xav)/s(x)
with xav = ixi/n and s(x)= { i(xi- xav)2/n}
278
Fig.D.11
Diagram of feature vectors of LAT analyses as used in MAION similarity

computations
The name of each feature indicates the analytical parameter it is associated with
These transformations of the feature vector have also been accounted for in Fig.D.11, which
thus shows the centred and standardized feature vectors. It is clear that there is hardly any
similarity between the three benchmark samples.
Visualization of similarities in the LAT framework
If the above-mentioned centering and standardizing transformations were carried out in advance
of application of the MAION similarity formula,
r(X,Y)= (n ixiyi- ixi iyi) /{(n ixi2-( ixi)2) (n iyi2-( iyi)2)}
the latter would reduce to
279
r(X,Y)= ixiyi/n
with ixi/n = iyi/n =0 and
2
ixi /n
= iyi2/n =1
For analysis X, r(X,Y) thus is a weighted mean of the elements of its standardized and centred
feature vector. The weighting coefficients are the elements of the standardized and centred
feature vector of analysis Y, for instance one of the LAT analyses. In this respect, the method
can be considered a method of comparative filtering, with L, A, and T as numerical filters.
It was shown already in Section D.1 that the MAION similarities to two of the LAT analyses
can be used as a basis for a graphical representation. Such a representation can be imagined as a
projection onto a plane in the multi-dimensional feature space. This plane is chosen in such a
way, that the feature vectors of two of the LAT analyses lie in it. It follows from the similarities
of the LAT analyses to one another, that the rTH-rAT and rTH-rLI diagrams will comprise less
redundant information than the rAT-rLI diagram, since r(AT,LI) deviates more from zero than
do the other similarities. The rTH-rLI diagram is given in Fig.D.12 which shows the same
simulated analyses as Fig.D.9.
Since the similarities of AT-W80 to TH-N70 and LI-ANG, respectively, are -17% and -56%
(Table D.1), the sample AT-W80 is plotted at the location (-17, -56) in the diagram. Likewise,
TH-N70 and LI-ANG, respectively, are plotted at (100, 30) and (30, 100). Note that, while TH
and LI lie in the plane of the diagram, AT lies in another plane in the multidimensional space
considered, and it is projected onto the plane of the diagram, as are other water samples.
As the similarity to both TH and LI decreases, the projected plotting position for a particular
sample is closer to the one of AT. Yet, rAT for such a sample may be low. It is advisable to
check rAT for samples which have both rLI and rTH <50%. The interpretation of such samples
as being more similar to AT, should be backed by rAT>50%. Meanwhile, one must be aware
that samples, which have approximately the same EC and IR values, and which will thus be
plotted close to one another in the EC-IR diagram (Fig.D.9) may, likewise, differ with respect to
other parameters.
Inferences from the TH-LI diagram (Fig.D.12)
The hatched fAT>90% area ABC in Fig.D.12 will now be considered as a first example of how
to read the TH-LI diagram, and so as to mention some conclusions with regard to the application
of MIX, the mixing model. It appears that an admixture of only 10% LI water in AT water will
shift the plotting position to C, where rLI=90%. Further addition of LI water will move the
position from C to L, thus mainly increasing rTH. This is due to the fact that LI has a positive
similarity of 30% to TH. When TH water is added to AT water, 10% of volume of TH water
will render the mixture very similar to TH (point B). The large shifts to C and B upon admixture
of only 10% of LI or TH can be explained by the much higher concentrations of LI, and
especially of TH, as compared to AT. The hatched area is an illustration of the plotting areas of a
litho-atmocline (ACKA) and a thalasso-atmocline (ABKA).
280
Fig.D.12
rTH-rLI diagram, showing the results of application of the MIX model

Explanation in the text
The diagram also shows that the addition of 1% of TH to AT (point F), or to LI (point G),
will have a considerable effect on the position of the mixtures in the diagram. Mixing 40% of
AT in LI, on the other hand, will only slightly shift the plotting position of the mixture away
from LI to E. Further points in Fig.D.12 are marked to indicate the same mixtures as plotted in
Fig.D.9.
It can be concluded that additions of very small amounts of thalassotrophic water will render
any type of water very much like TH in this diagram. Quite substantial amounts of LI are needed
to have the same effect in waters originally resembling atmotrophic water. Only very large
amounts of atmotrophic water can contribute appreciably to rendering samples atmotrophic.
281
Fig.D.13
rTH-rLI diagram for the samples in Table D.3

Note that the WRNET samples (w) were selected to fill in some of the otherwise open space in the
diagram, rather than to represent the composition of the boezem water in "De Weerribben"
The dotted area HIJH in the diagram represents the plotting position of mixtures of TH, LI,
and AT, containing more than 20% by volume of RH-LOB, i.e., Rhine water. Clearly this
addition will make the water more similar to thalassotrophic water, although especially
lithotrophic water will still remain quite lithotrophic after an addition of up to 20% of RH. Most
of the dotted area represents mixtures with much AT, as can be traced by means of the dotted
line AIR indicating mixtures of AT and RH only. A cluster of samples occupying the dotted area
might be regarded as a molunocline. As stated before, this is due to human activity: Rhine water
analysed until the beginning of this century plots very close to LI in the diagram.
Fig.D.13 shows the analyses from Table D.3 in a TH-LI diagram. This diagram illustrates
the general trend that only few analyses of natural waters plot in the central area. A comparison
of the TH-LI diagrams with the EC-IR diagrams suggests that their use will not lead to very
different conclusions. This is confirmed by Fig.D.14, showing the plotting positions of more
than 5000 water analyses in RIN archives (compare Fig.D.10).
282
Fig.D.14
rTH-rLI diagram, showing the plotting position of more than 5000 analyses of
(mainly surface) waters in The Netherlands (RIN archives)
A comparison with conventional statistical methods

Grootjans (1985) and Hoogendoorn (1983) and others have used conventional statistical
methods to arrange or classify water analyses. The relation between such methods and the
MAION similarity method is shortly described here.
"Cluster analysis" is mostly applied in a standard form which has not been adapted to the
treatment of the numerical data resulting from water analysis. With the MAION similarity
coefficient, it has the comparison of feature vectors in common, and the computation of
coefficients of similarity is a central item in clustering techniques. The main difference lies in
the (dis)similarity formula chosen, in the definition of the feature vector to be used, and in the
application of particular transformations to such vectors. Some aspects can be illustrated on the
basis of the MAION similarity coefficient.
When, in MAION, EC25 would have been omitted, or just not squared, the result of the
similarity calculations would be independent of the overall concentration of water samples, as
should be clear from the discussion in Section D.2. In that case, the similarity coefficient would
reflect differences in ionic composition of the water only, as does the Piper diagram (Fig.D.1,
Section D.1).
283
Use of the covariance instead of the correlation as a measure of similarity would correspond
to a centering, rather than a centering and standardizing of the feature vector and the subsequent
application of the simple formula presented before:
r(X,Y)= i(xiyi /n) with i(xi /n) = i(yi /n) =0
The usually undesired effect is that the method becomes especially sensitive to the overall
concentration. The values thus found, moreover, have no upper and lower limit.
This type of problem may have been of influence in the cluster analysis reported by
Grootjans (1985, p.111-115). The clusters found, all show a wide variety of IR values, while
they are strongly correlated with EC25 (written communication to the author by A.P. Grootjans,
R. van Diggelen, M. Wassen & W. Wiersinga, April 1983). This may be attributable to the
clusters being implicitly defined mainly on the basis of overall concentrations.
Hoogendoorn (1983, part 1, p.47-48,146) recommends Q-mode factor analysis for the study
of the multivariate relations between water samples. Mathematically, factor analysis is not very
different from principal component analysis (Pielou 1977, p.332-340). The MAION similarity
method, as well as the study of relations between water samples in graphical representations
(Section D.1) may be considered a simpler alternative. Where the statistical methods depend on
the frequency distributions of the numerical data within the data set considered, the LAT
analyses provide an independent framework. Like EC and IR do for the parameter variability,
the similarity to the LAT samples, for being "extreme" with respect to these very parameters will
"explain" most of the variability of the analyses.
In conclusion, conceivably advanced statistical techniques might be modified in such a way
that they yield better solutions to the problem of comparing and evaluating water samples than
do the MAION techniques presented here. Yet, these statistical methods require a well-balanced
data set and an appropriate definition of the feature vector to be used for water samples. For the
time being, the MAION processing has proved to be able to treat the numerical data of water
analyses in a satisfactory manner, especially when no a priori-knowledge is available with
respect of the representation of specific types of water in the area studied.
284
APPENDIX E
Evapo-transpiration from lysimeters with fen vegetation
In order to relate evapo-transpiration from fen vegetation to the evaporation from an open water
surface, and thus to standard data provided by meteorological stations (Penman 1963), five
lysimeters were installed in De Weerribben and followed in the 1975-1976 period (Bot 1975,
Brandsma 1975, Straathof 1976, Vegt 1978). Oil- drums (210 l, diameter 58 cm, height 80 cm)
were filled with a kragge cylinder and inserted in the holes so formed (Fig.E.1). The lysimeters
were allowed to float up and down with the kragge and the water level was adjusted to a fixed
level once every second day. For a precise adjustment water was added or removed, with a
calibrated vessel, until the meniscus just drew level with the tip of a nail in the middle of the
water surface. As demanded by the relative movement of the surrounding water level, the fixed
water level in the lysimeters was reset, to which purpose the contact nail could be vertically
adjusted with a micrometer slide.
The five lysimeters contained vegetation dominated by Phragmites australis, by Sphagnum
flexuosum and Juncus subnodulosus, by Scorpidium scorpioides and Carex lasiocarpa,
C.diandra and C.elata, by Cladium mariscus, and by Polytrichum commune, respectively, and
they were inserted in exactly the same place where the kragge cylinders had been cut out. It
must be mentioned that these places differed in their exposure to wind and sun. Large lysimeters
were used in order to reduce border effects and to avoid damage to the plants. The installation
proved difficult, however, due to the large volume and heavy weight of the lysimeters. The
vegetation fragment in the Cladium and Polytrichum lysimeters shortly exhibited reduced
growth, and the Polytrichumkragge even drowned in the lysimeter, the moss not recovering
afterwards.
Evapo-transpiration from lysimeters
285
Fig.E.1
Lysimeter installed in a quagfen kragge

a adjustable pin for water level control; b one of three posts forming a float-shaft, and used as support of
walking planks
The Phragmites lysimeter stood in a dense stand of reed. The vegetation in the lysimeter did not
grow as high as the surrounding reed, but it was similar to other natural stands in the middle of
kraggen where the supply of nutrients is reduced (Table E.1).
The results of the lysimeter study have been summarized in Table E.2, presenting regression
constants for the equations
(1)
(2)
Ea = f(E0)
Ea = a(E0) +b
where
Ea: actual evapotranspiration (mm/d)
E0: evaporation from an open water surface according to Penman (1963)
f:
multiplication factor (vegetation factor)
a,b:constants in linear regression equation
It should be noticed that the water supply to the vegetation fragments in the lysimeters has
always been unlimited, as it is in the natural situation. For this reason the measured actual
evapo-transpiration (Ea) may be taken as representative of the potential value (Ep).
286
Table E.1 Average characteristics of the lysimeters

Lysimeter:
Sphagnum
Polytrichum
Cladium
Scorpidium
Phragmites
equal
50 to 80 cm
lower at end
of season
(Phragmites)
equal
reduced at
Cover of various types in lysimeter as seen from above:

Open water:
5%
25%
50%
Dead plants:
0%
50%
30%
Mosses:
90%
25%
<5%
Helophytes:
5%
1%
15-20%
10%
0%
85%
5%
10%
0%
5%
85%
Kragge thickn.:
45 cm
65 cm
Height of vegetation as compared to the surrounding stand:

up to 5 cm
15 cm lower
40 cm lower
higher
(mosses,
(Cladium)
(mosses)
died due to
drowned kragge)
Growth of the vegetation as compared to the surrounding stand:
vital
strongly
equal
reduced
55 cm
end of season
>65cm
45 cm
lower part
of kragge
torn off
and lost
sheltered
due to high
standing dead
Cladium around
Remarks:
sheltered
due to vital,
1.3-2 m high
reed around
The correlations are rather poor, as is shown by the low values for r2, especially in 1976.
There may be various reasons why the evapotranspiration from lysimeters did not show a strict
correlation with Penman's E0:
- Any active role of the vegetation cover is empirically treated by the introduction of the
vegetation factor;
- The vegetation fragment in the lysimeters is isolated from the subsurface water flow system in
the mire, and may have suffered physiological damage by the installation;
- The whole procedure of frequent level adjustments and associated calculations is somewhat
inaccurate, and further errors are introduced by technical deficiencies in the installed lysimeters;
- Lysimeters may not be expected to receive the same amount of rain as do well-installed raingauges;
- The lysimeters are subject to locally different micrometeorological environments;
- Penman's E0 values, as calculated for data obtained at the weather stations in Leeuwarden and
Lelystad, are only estimates, and these localities are at a considerable distance from the
lysimeter site.
It is obvious, from Fig.E.2, that the evapotranspiration from the various lysimeters in both
1975 and 1976 shows a similar pattern (correlations with the Sphagnum lysimeter vary between
0.78 and 0.96), although the correlation with Penman's E0 is poorer, especially in 1976. Both
287
Table E.2 Regression results for evapotranspiration from lysimeters, as a linear function of
Penman's E0
Lysimeter
Sphagnum
Polytrichum
Cladium
Scorpidium
Phragmites
750520-751031
n=16
f
a
b
0.59 0.60 -0.02
0.40 0.54 -0.44
0.49 0.62 -0.44
0.61 0.77 -0.50
0.47 0.63 -0.48
r2
0.79
0.69
0.77
0.68
0.73
760320-760915
n=18
f
a
b
0.73 0.59 0.49
0.49 0.25 0.91
0.59 0.48 0.42
0.52 0.22 1.11
0.47 0.16 1.11
r2
0.41
0.18*
0.39
0.19*
0.09*
both periods
n=34
f
a
b
0.67 0.63 0.14
0.45 0.42 0.10
0.55 0.58 -0.12
0.56 0.49 0.23
0.47 0.41 0.21
r2
0.56
0.44
0.58
0.44
0.40
Entry data: mm/d for periods of ten days; f: ratio Ea/E0; b: offset (mm/d) in linear regression equation;
r2: coefficient of determination; * not significant at the 0.01 level (t-test, two-sided)
years were extremely dry, 1976 even more so than 1975. This may have influenced the
processes involved in the evapotranspiration. One should note, among other things, that the
evapotranspiration from the Polytrichum, Scorpidium and Phragmites lysimeters in June and
July 1976 was low. The upper surface of the moss plants in the Polytrichum and Scorpidium
lysimeters dried out during this summer period and may have had a reduced evapotranspiration.
The Phragmites lysimeter became sheltered, during the growing season, by the up to 0.8 m
higher surrounding reed, which kept the atmosphere moister.
If extremes are removed, a better correlation with E 0 can be obtained, but I cannot visualize
any method to obtain results that can safely be applied to other periods and places.
In 1986, Koerselman & Beltman (1988) also used lysimeters to assess the evapotranspiration
of quagfen vegetation. Their results suggest a multiplication factor of 0.75 for the types of
vegetation involved, and they found high correlations with E0 at a small quagfen site near
Utrecht, using meteorological data for De Bilt, at a distance of about 8 km from the lysimeter
site. A general perusal of their results shows that the pattern for E0 was quite similar to that
found in the present study. In view of the strong correlations between lysimeters in both studies,
I do not think that possible differences in the accuracy of measurements and calculations can
fully explain the different value found for the vegetation factor. The following factors may
explain the differences recorded:
The lysimeters used in the present study had a 1.66 times larger surface area and a 2.3 times
larger depth, possibly resulting in a more representative experimental system in the Phragmites,
Sphagnum, and Scorpidium lysimeters. This concerns especially the height of the water table,
relative to the kragge level, and the growth of the vegetation fragment. A rise or fall of only 1 or
2 cm of the relative height of the water table may strongly interfere with the evapotranspiration
through a changing relative area of surface water in the lysimeters. From the study by
Koerselman & Beltman, who included open water pans, it appeared that pan evaporation is only
0.4-0.5 times E0, which suggests sheltered conditions. The evapotranspiration from the
lysimeters in the present study is apparently closer to open water pan evaporation than to
evapotranspiration measured in quagfen lysimeters with a relative water level 3 cm below the
kragge surface by Koerselman & Beltman. Indeed, surface water covered a substantial part of
the lysimeters in the present study (see Table E.1).
288
Fig.E.2
Precipitation, Penman's E0, and evapo-transpiration from lysimeters in De

Weerriben
Average values per decade have been used; the start of each month is marked on the axis
289
In the present study, the aerial extension of the plants rooting in the lysimeter was essentially
confined to the cilindrical body of air just above the lysimeter, and above-ground herb biomass
was probably lower than in Koerselman's lysimeters.
Koerselman & Beltman conducted their study in small quagfen parcels in an open grassland
area. This may have increased the importance of advection of dry air and so increased the
evapotranspiration as compared to the large mire area, with much open water, in North-West
Overijssel.
In conclusion, during the growing season the evapotranspiration from quagfen in De
Stobbenribben is probably ca 0.6 times Penman's E0 as calculated on the basis of weather data
from Leeuwarden and Lelystad, but a variation of ca 30% may exist. Insufficient data are
available for any precise explanation and calculation of this variation.
290
APPENDIX F
Data reports
This appendix lists data reports and manuscripts resulting from the present project, and is
primarily intended to acknowledge the contributions of many students not explicitly referred to
in other parts of this book, rather then to provide a complete list.
Anonymus 1973. Waarnemingen verricht tijdens de veldcursus oecologie door studenten van het
3e studiejaar, richting B3 en B4. Mscr. Hugo de Vries-Laboratorium, UvA, Amsterdam.
Bergmans, W. 1975. Synoekologisch onderzoek in enige suksessiereeksen in het C.R.M.-reservaat De Weerribben (N.W.-Overijssel). Int. Rapp. 16. Hugo de Vries-Lab., UvA,
Amsterdam, 67p., bijl.
Blanksma, W.J. 1974. Hydrologisch Onderzoek in een kraggecomplex in het C.R.M.-Reservaat
"De Weerribben". Int. Rapp. 11. Hugo de Vries-Lab., UvA, Amsterdam, 17p., bijl.
Boeye, D. 1983. Verslag van een ecohydrologische stage in "De Weerribben" (N.W.-Overijssel,
NL). Vrije Universiteit Brussel, Dienst Hydrologie, Brussel, 22p., bijl.
Bon, J. 1975. Hydrologisch overzicht van het natuurreservaat De Wheerribben en omgeving.
Nota 865. Instituut voor Cultuurtechniek en Waterhuishouding, Wageningen, 13p.
Bos, C. & E. Moolhuizen. 1975. Enige zanddiepte-metingen in De Weerribben. Mscr. SBB,
Zwolle.
Bot, A.K. 1975. Hydrologisch onderzoek in drie proefterreinen in het C.R.M.-Reservaat "De
Weerribben". Int. Rapp. Inst. v. Cultuurtechniek en Waterhuishouding, Wageningen, 31p.,
bijl.
Brand, L.J. & J.A. Leemburg 1978. Een onderzoek naar de relatie tussen reflektie-eigenschappen
en hoeveelheid groene biomassa bij enige halfnatuurlijke vegetaties. Mscr. Lab. v.
Plantenoecologie RUG/Rijksinst. v. Natuurbeheer, Haren/Leersum, 61p.
Data reports
291
Brandsma, J. 1975. Voortgezet hydrologisch onderzoek in "De Weerribben". Int. Rapp. Inst. v.
Cultuurtechniek en Waterhuishouding/Staatsbosbeheer, Oldemarkt, 37p., bijl.
Brijker, H. & J. Hartevelt 1976. De vegetatie van het Woldlakebos, een verdroogd moerasgebied
in Noord-West Overijssel. Int. Rapp. 28. Hugo de Vries-Lab., UvA, Amsterdam, 28p., bijl.
Calis, J.N.M. & J.C.J. van Wetten 1983. Onderzoek van successie en hydrologie, in het
trilveen-complex "De Wobberibben" (De Weerribben N.W.-Overijssel). Int. Rapp. 153.
Hugo de Vries-Lab., UvA, Amsterdam, 91p.
Claassen, T.H.L. 1976. De beoordeling van oppervlaktewater in Noordwest-Overijssel op basis
van biologische, chemische en fysische gegevens. Int. Rapp. 26. Hugo de Vries-Lab., UvA,
Amsterdam, 29p., bijl.
De Boer, J., R, J. Holtland & I. Lucassen 1977. Kan hoogveen ontstaan in de Polletjesgaten?. Int.
Rapp. Rijksinstituut voor Natuurbeheer, Leersum, 53p., bijl.
De Waard, J. 1979. Een onderzoek naar het voorkomen van boomopslag op enkele percelen in
het CRM-reservaat "De Weerribben". Int. Rapp. 72. Hugo de Vrieslab., UvA, Amsterdam,
58p.
Heringa, R.B. & P.G.M. van der Sloot 1973. Manuscript vegetatiekaart. Mscr.
Holtland, J. & I. Lucassen 1976. De bereikbaarheid van de kraggen voor boezemwater in een
gedeelte van "De Weerribben". Mscr. Staatsbosbeheer, Zwolle, 18p.
Jalink, M. 1990. Een evaluatie van digitaal beeldmateriaal en vegetatiekundige gegevens van de
Stobbenribben, NW-Overijssel. Int. Rapp. 90/3. Rijksinstituut v. Natuurbeheer, Arnhem,
140p.
Klijn, J.A. 1973. Interpretatie false color luchtfoto's De Weerribben. Mscr. ITC/Hugo de
Vries-Lab., UvA, Enschede/Amsterdam.
Kooijman, A.M. 1985. Een onderzoek naar de nutrintenhuishouding van een trilveen in "De
Weerribben". Rapp. Vakgr. Bot. Oec., Sectie Landschapsoecologie, RUU, 77p.
Lumkes, M. 1976. Kartering van Watervegetaties in verband met milieu-eisen van krabbescheer.
Mscr., 16p.
Muis, A. 1974. Windmoleninventarisatie in "De Weerribben". Rapp. BCS, Velp, 34p, map.
Mulder, M.A.A. 1985. Geohydrologische modelstudie Weerribben. Rapp. T. H. -Delft, Afd.
Civiele Techniek, Vakgr. Waterbeheersing, 76p.
Noot, J. 1974. Ecologisch onderzoek van het Duiningermeer. Concept-verslag. Hugo de VriesLab., UvA.
Oosterbroek, P. & B.J. Post 1977. Een vegetatiekartering m. b. v. false-colour luchtfoto's in het
C.R.M.-reservaat "De Weerribben" (N.W.-Overijssel). Int. Rapp. 40. Hugo de Vries-Lab.,
UvA, Amsterdam, 45p., bijl.
Raeymaekers, G. 1975. Eem vegetatiekundig onderzoek in verband met de waterhuishouding
van enige kraggegebieden in het Natuurmonument "De Wieden" (N.W.-Overijssel). Int.
Rapp. 14. Hugo de Vries-Lab., UvA, Amsterdam, 21p., bijl.
Raeymaekers, G. 1977. Het verband tussen waterhuishouding en vegetaties in waterrijke
gebieden. Dumortiera 7-8, 38-47.
Rengersen, J. 1980. Een verspreidingsmodel voor conservatieve stoffen in het oppervlaktewater
van NW-Overijssel. Report. T. H. -Twente, Afd. Chemische Technologie, Onderzoeksgr.
Techn. Milieubeh., 122p.
Ruitenburg, G.J.M. 1974. Vegetatiekartering van De Wobberibben. Concept-verslag. Hugo de
Vries-Lab. UvA.
Schoof, A. J. G. 1973. Een gedetailleerde vegetatiekartering op een groeiplaats van enige
Ericaceae in het C.R.M.-Reservaat "De Weerribben" (N.W.-Overijssel). Int. Rapp. 5. Hugo
de Vries-Lab., UvA, Amsterdam, 39p., bijl.
292
Data reports
Straathof, N. 1976. Praktijkverslag over hydrologisch onderzoek in de Weerribben. Int. Rapp.

Staatsbosbeheer/Rijksinstituut voor Natuurbeheer, Zwolle, 11p., bijl.
Touber, L. 1973. Hydrologisch Onderzoek in enige verlande petgaten in het C.R.M.-Reservaat
"De Weerribben", N.W.-Overijssel. Int. Rapp. 3. Hugo de Vries-Lab., UvA, Amsterdam,
34p., bijl.
Van Leeuwen, C.H. 1978-'83. Temperatuur en electrische geleiding in De Stobbenribben.
Rijksinstituut voor Natuurbeheer, Leersum (ongepubliceerd).
Van Opbroek, G. 1976. Verslag van de vegetatie-opnamen en kaartering van de hoogveenontwikkeling op het terrein bij Tietema's Vaart. Mscr.
Van der Perk, J.C. & M.J. Smit 1975. Een hydrologisch onderzoek ten behoeve van het
natuurbeheer in de "Wieden". Int. Rapp. 17. Hugo de Vries-Lab., UvA, Amsterdam, 56p.,
bijl.
Vegt, J.J. 1978. Herkenning van natuurlijke vegetaties met behulp van multi-spectrale Remote
Sensing. Int. Rapp. Rijksinstituut voor Natuurbeheer, Leersum, 48p., bijl.
Vegt, J.J. 1978. Verdamping, berging en indringing van boezemwater in het moerasgebied "De
Weerribben". Int. Rep., RIN, Leersum, 51p., bijl.
Verhoeven, J.T.A., A.M. Kooijman & G. van Wirdum 1988. Mineralization of N and P along a
trophic gradient in a freshwater mire. Biogeochemistry 6, 31-43.
Verschoor, A. 1978. Vegetatiekartering met behulp van luchtfoto's in het CRM-reservaat "De
Weerribben" (N.W.-Overijssel). Int. Rapp. 56. Hugo de Vries-lab., UvA, 31p.
Verschuren, B. 1976. Een krabbescheerinventarisatie in Noordwest-Overijssel en Zuid-Friesland.
Int. Rapp., Hogere Tuinbouwschool ('s Hertogenbosch).
Vromen, H, L. Klamer & J. de Vries 1974. Voortgezet Hydrologisch onderzoek in enige
verlande petgaten in het C.R.M.-reservaat "De Weerribben". Int. Rapp., HBCS/SBB
(Zwolle).
Data reports
293
REFERENCES
Amann, J. 1928. Bryogographie de la Suisse. Matriaux pour la flore cryptogamique suisse VI(2). Fretz, Zrich, 453p.
Anonymus 1982. Makrophytische Indikatoren fr die kochemische Beschaffenheit der Gewsser. Ausgewhlte
Methoden der Wasseruntersuchung, Band II, Biologische, mikrobiologische und toxikologische Methoden, 1.1.2.
Gustav Fischer, Jena, 67-88.
Baltov-Tul kov, E. 1972. Flachmoorwiesen im mittleren und unteren Opava-Tal. Vegetace SSR A(4). Praha, 201p.
Baltov-Tul kov, E. 1976. Rieder- und Sumpfwiesen der Ordnung Magnocaricetalia in der Zhorie-Tiefebene und
dem nrdlich angrenzenden Gebiete. Vegetcia SSR B(3). Veda, Bratislava, 257p.
Bear, J. 1972. Dynamics of fluids in porous media. Elsevier, New York etc., 764p.
Bergmans, W. 1975. Synoekologisch onderzoek in enige suksessiereeksen in het C.R.M. reservaat De Weerribben
(N.W.-Overijssel). Int.Rapp.16. Hugo de Vries-Lab., UvA, Amsterdam, 67p., bijl.
Bick, W., A. Robertson, R. Schneider, S. Schneider & P. Ilnicki 1976. S ownik torfoznawczy Niemiecko-PolskoAngielsko-Rosyjski (Glossary for bog and peat, German-Polish-English-Russian). Biblioteczka Wiadomo ci IMUZ
nr 51. Instytut Melioracji i U ytkw Zielonych, Pa stwowe wydawnictwo rolnicze i le ne, Warszawa, 178p.
Bloemendaal, F.H.J.L. & J.G.M. Roelofs 1988. Waterverharding. In: F.H.J.L. Bloemendaal & J.G.M. Roelofs (red.),
Waterplanten en Waterkwaliteit. Nat.Hist.Bibl. KNNV 45. Kon. Ned. Nat.-Hist. Ver., Utrecht, 147-158p.
Boeye, D. 1983. Verslag van een ecohydrologische stage in "De Weerribben" (N.W.-Overijssel, NL). Vrije Universiteit
Brussel, Dienst Hydrologie, Brussel, 22p., bijl.
Bon, J. 1975. Hydrologisch overzicht van het natuurreservaat De Wheerribben en omgeving. Nota 865. Instituut voor
Cultuurtechniek en Waterhuishouding, Wageningen, 13p.
Bonetto, A.A. 1975. Hydrologic regime of the Paran River and its influence on ecosystems. In: A.D. Hasler (ed.),
Coupling Land and Water Systems. Springer, Berlin etc., 175-197.
Boros, A. 1968. Bryogeographie und Bryoflora Ungarns. Akadmiai Kiad, Budapest, 466p.
Bot, A.K. 1975. Hydrologisch onderzoek in drie proefterreinen in het C.R.M.-Reservaat "De Weerribben". Int.Rapp. Inst.
v. Cultuurtechniek en Waterhuishouding, Wageningen, 31p., bijl.
Brand, L.J. & J.A. Leemburg 1978. Een onderzoek naar de relatie tussen reflektie-eigenschappen en hoeveelheid groene
biomassa bij enige halfnatuurlijke vegetaties. Mscr. Lab. v. Plantenoecologie RUU/Rijksinst.v.Natuurbeheer,
Haren/Leersum, 61p.
References
295
Brandsma, J. 1975. Voortgezet hydrologisch onderzoek in "De Weerribben". Int.Rapp. Inst.v.Cultuurtechniek en

Waterhuishouding/Staatsbosbeheer, Oldemarkt, 37p., bijl.
Braun, W. 1968. Die Kalkflachmoore und ihre wichtigsten Kontaktgesellschaften im Bayerischen Alpenvorland.
Dissertationes Botanicae 1. Cramer, Lehre, 134p., bijl.
Bredenbeek, P., P. Gerrits & V. van Loon 1979. Vegetatiekartering van de Stobbenribben in het natuurreservaat "De
Weerribben" te Ossenzijl (N.W.-Overijssel). Sticht. Opleiding Leraren, Utrecht, 46p.
Brockhaus 1961. Die Entwicklungsgeschichte der Erde. Brockhaus, Leipzig, 772p.
Calis, J.N.M. & J.C.J. van Wetten 1983. Onderzoek van successie en hydrologie, in het trilveen-complex "De
Wobberibben" (De Weerribben N.W.-Overijssel). Int.Rapp.153. Hugo de Vries-Lab., UvA, Amsterdam, 91p.
CHO-TNO 1986. Water in The Netherlands. Proc. and Inf. 37. TNO Committee on Hydrological Research (CHO), The
Hague, 70p.
Clymo, R.S. 1983. Peat. In: A.J.P Gore (ed.), Mires: swamp, bog, fen and moor, General studies. Ecosystems of the world
4a. Elsevier, Amsterdam etc., 159-224.
Coesel-Wouda, M.J. 1967. Onderzoek naar de oecologie van de Europese Utricularia-soorten. Rapp. Hugo de
Vries-Laboratorium/RIVON, Amsterdam/Leersum, 48p.
Connell, J.H. & R.O. Slatyer 1977. Mechanisms of succession in natural communities and their role in community
stability and organization. The American Naturalist, vol. 111(982),1119-1144.
Corporaal, A. 1984. Calamagrostis x gracilescens (Blytt) nieuw voor Nederland (Engl. Summ.: Calamagrostis x
gracilescens (Blytt) new for the Netherlands). Gorteria 12(5),109-111.
De Boer, J., R,J. Holtland & I. Lucassen 1977. Kan hoogveen ontstaan in de Polletjesgaten?. Int.Rapp. Rijksinstituut voor
Natuurbeheer, Leersum, 53p., bijl.
De Crook, Th. 1979. Het berekenen van de verticale grondwatersnelheid uit grondtemperatuurprofielmetingen. Scientific
Rep. WR 79-11. Koninklijk Nederlands Meteorologisch Instituut, De Bilt, 35p.
De Graaf, F. 1955. De micro-organismen van plankton, lasion en bodem van het plassengebied Het Hol. In: W. Meijer &
R.J. De Wit (red.), Kortenhoef, Een veldbiologische studie van een Hollands verlandingsgebied. Commissie voor de
Vecht en het Oostelijk en Westelijk Plassengebied, Amsterdam, 67-91.
De Lange, L. 1972. An ecological study of ditch vegetation in The Netherlands. Thesis. University of Amsterdam, 112p.
De Vries, D.A. 1963. Thermal properties of soils. In: W.R. van Wijk (ed.), Physics of plant environment. North-Holland,
Amsterdam, 210-235.
De Wit, R. 1951. De draadzegge-gemeenschap in Noordwest-Overijssel. Kruipnieuws 13(1/2):3-6, also in: J.C.
Smittenberg (red.), Plantengroei in enkele Nederlandse landschappen (1973). Bondsuitgeverij van de Jeugdbonden
voor Natuurstudie (N.J.N, C.J.N. en K.J.N.), Amsterdam, 345-357.
Den Boer, P.J. 1986. The Present Status of the Competitive Exclusion Principle. Tree 1(1). Elsevier, Amsterdam, 25-28.
Den Held, J.J. & A.J. den Held 1973. Beknopte handleiding voor vegetatiekundig onderzoek. Wetenschappelijke
Mededelingen van de KNNV 97. Koninklijke Ned. Natuurhistorische Ver., Hoogwoud, 39p.
Dickinson, C.H. 1983. Micro-organisms in peatlands. In: A.J.P. Gore (ed.), Mires: swamp, bog, fen and moor, General
studies. Ecosystems of the world 4a. Elsevier, Amsterdam, 225-245.
Dierssen, K. 1982. Die wichtigsten Pflanzengesellschaften der Moore NW-Europas. Conservatoire et Jardin botaniques,
Genve, 382p.
Dirkse, G.M., H.M.H. van Melick & A. Touw 1989. Checklist of Dutch bryophytes. Lindbergia 14(3), 167-175.
Dykyjov, D. & J. Kv t (eds.) 1978. Pond Littoral Ecosystems. Springer, Berlin etc., 464p.
Ellenberg, H. 1974. Zeigerwerte der Gefsspflanzen Mitteleuropas. Scripta Geobotanica 9, Gttingen, 97p.
Ellenberg, H. 1978. Vegetation Mitteleuropas mit den Alpen in kologischer Sicht. Ulmer, Stuttgart, 981p.
Etherington, J.R. 1975. Environment and Plant Ecology. Wiley, London etc., 347p.
Eurola, S., S. Hicks & E. Kaakinen 1984. Key to Finnish Mire Types. In: P.D. Moore (ed.), European Mires. Academic
Press, London etc., 11-119.
Faber, F.J. 1960. Aanvullende hoofdstukken over de geologie van Nederland, IV. Noorduijn, Gorinchem, 607p.
Gallandat, J.D. 1982. Prairies marcageuses du Haut-Jura. Matriaux pour le lev gobotanique de La Suisse.
Flck-Wirth, Teufen, 327p.
Geologische Dienst 1979. Geologisch onderzoek natuurgebied "De Weerribben". Rapp. 10312. Rijks Geologische Dienst,
Haarlem, 3p., bijl.
Gibbs, R.J. 1970. Mechanisms controlling world water chemistry. Science 170, 1088-1090.
Giller, K.E. 1982. Aspects of the plant ecology of a flood-plain mire in Broadland, Norfolk. Thesis. University of
Sheffield, Sheffield, 239p.
Giller K.E., & B.D. Wheeler 1986. Peat and peat water chemistry of a flood-plain fen in Broadland, Norfolk, U.K.
Freshwater Biology 16, 99-114.
Godwin, H. 1978. Fenland, its ancient past and uncertain future. Cambridge University Press, Cambridge etc., 196p.
Golterman, H.L., R.S. Clymo & M.A.M. Ohnstad 1978. Methods for physical and chemical analysis of fresh waters. IBP
Handbook 8. Blackwell, Oxford etc., 213p.
Gonggrijp, G., V. Langenhoff & W. Schroevers 1981. Ontdek N.W.-Overijssel. Nederlandse Landschappen. IVN,
Amsterdam, 288p.
296
References
Gore, A.J.P. (ed.) 1983. Mires: Swamp, Bog, Fen and Moor, A: General Studies, B: Regional Studies. Ecosystems of the
world 4. Elsevier, Amsterdam etc., 440, 479.
Gorham, E. S.J. Eisenreich, J. Ford & M.V. Santelmann 1985. The chemistry of bog waters. In: W. Stumm (ed.),
Chemical processes in lakes. Wiley, New York etc., 339-363.
Grime, J.P. 1979. Plant strategies and vegetation processes. Wiley, Chichester etc., 222p.
Grootjans, A.P. 1985. Changes of groundwater regime in wet meadows. Thesis. University of Groningen, Groningen,
146p.
Grubb. P.J. 1977. The maintenance of species-richness in plant communities: the importance of the regeneration niche.
Biol.Rev. 52, 107-145.
Gunning 1870. See Vereniging 1870, p.125.
Haans, J.C.F.M. 1951. De rietgroei in de Schut- en Grafkampen onder de gemeente Oldemarkt. Rapp. 265. Stichting voor
Bodemkartering, Wageningen, 14p.
Haans, J.C.F.M. & C. Hamming 1954. De bodemgesteldheid van het ruilverkavelingsgebied Vollenhove. Int. Rapp. 392.
Stichting voor Bodemkartering, Wageningen.
Haans, J.C.F.M. & C. Hamming 1962. Over de bodemgesteldheid in het veengebied in het Land van Vollenhove. Int.
Rapp. 563. Stichting voor Bodemkartering, Wageningen, 58p.
Hack, J. 1973. Kartering en oecologie van een aantal wateren oeverplanten in De Weerribben te Oldemarkt. Int. Rapp.
Hugo de Vries-Lab., UvA/Rijksinst.v.Natuurbeheer, Amsterdam/Leersum, 48p., bijl.
Havinga, A.J. 1957. Bijdrage tot de kennis van het rietland in Noordwest-Overijssel. Boor en Spade 8, 131-140.
Hedens, L. 1989. The genera Scorpidium and Hamatocaulis, gen. nov., in northern Europe. Lindbergia 15(1), 8-36.
Heij, G.J. & T. Schneider (eds.) in press. Acidification research in The Netherlands. Elsevier, Amsterdam.
Hem, J.D. 1970. Study and Interpretation of the Chemical Characteristics of Natural Water. Water-Supply Paper 1473.
Geological Survey. U.S. Government Printing Office, Washington, 363p.
Heringa, R.B. & P.G.M. van der Sloot 1973. Manuscript vegetatiekaart. Mscr.
Het Oversticht 1939. Bezwaren tegen de verdere inpoldering in het Land van Vollenhove. Rapp. Ingenieursbureau voor
bouw- en waterkunde Het Oversticht, Deventer, 35p.
Hoffer, R.M. & C.J. Johannsen 1969. Ecological potentials in spectral signature analysis. In: P.L. Johnson, Remote
Sensing in Ecology. University of Georgia Press, Athens, 1-16.
Hoogendoorn, J.H. 1983. Hydrochemie Oost-Nederland. Rapport OS 83-38, 4 dln. Dienst Grondwaterverkenning TNO,
Delft-Oosterwolde, 157p.
Hooghart, J.C. (red.) 1986. Verklarende hydrologische woordenlijst. Rapp. en Nota's 16. CHO-TNO,'s-Gravenhage, 130p.
Huijsmans, T.M.F., & F.W. Zwietering 1980. De waterhuishouding van "De Weerribben". Rapp. Anonymus, 102p.
Huston, M. 1979. A general hypothesis of species diversity. The American Naturalist 113(1), 81-101.
Hutchinson. G.E. 1975. A Treatise on Limnology, III, Limnological Botany. Wiley, New York etc., 660p.
International Peat Society 1984. Russian-English-German-Finnish-Swedish Peat Dictionary. IPS, Helsinki, 595p.
Jalink, M. 1990. Een evaluatie van digitaal beeldmateriaal en vegetatiekundige gegevens van de Stobbenribben,
NW-Overijssel. Int. Rapp. 90/3. Rijksinstituut v. Natuurbeheer, Arnhem, 140p.
Jelgersma, S. & J.B. Breeuwer 1975. Toelichting bij de geologische overzichtsprofielen door Nederland. In: W.H.
Zagwijn & C.J. van Staalduinen (red.), Toelichting bij de geologische overzichtskaarten van Nederland. Rijks
Geologische Dienst, Haarlem, 91-93.
Jol, C. & J. Laseur (ed.) 1982. Waterkwaliteitsaspecten in het proefgebied Nationaal Landschap Noordwest-Overijssel.
Rapp. Zuiveringschap West-Overijssel, Zwolle, 43p.
Kelts, K. & K.J. Hs 1978. Freshwater Carbonate Sedimentation. In: A. Lerman (ed.), Lakes: chemistry, geology, physics.
Springer, New York etc., 295-321.
Kielland 1937. See Hem 1970.
Kinzel, H. 1982. Einfhrung. In: H. Kinzel (ed.), Pflanzenkologie und Mineralstoffwechsel. Ulmer, Stuttgart, 13-31.
Kleijn, H.F. 1986. Agressiviteit en de oplosbaarheidsconstante van calciumcarbonaat. H2O 19(13), 309-310.
Kltzli, F. 1981. Vgtation. In: C. Roth et al. (ed.), tangs naturels - comment les projeter, les amnager, les recrer.
Office fdral des forts, Division de la protection de la nature et du paysage, Berne, 27-37.
Knipling, E.B. 1969. Leaf reflectance and image formation on color infrared film. In: P.L. Johnson, Remote Sensing in
Ecology. University of Georgia Press, Athens, 17-29.
KNMI 1972. Klimaatatlas van Nederland. Koninklijk Nederlands Meteorologisch Instituut. Staatsuitgeverij,'s-Gravenhage, 18p., maps.
Koch, W. 1926. Die Vegetationseinheiten der Linthebene unter Bercksichtigung der Verhltnisse in der Nordostschweiz.
Jahrbuch St.-Gallischen Naturwissensch. Gesellsch. 61(2).
Koerselman, W. 1989. Groundwater and surface water hydrology of a small groundwater-fed fen. Wetlands Ecology and
Management 1, 31-43.
Koerselman, W., S.A. Bakker & M. Blom in press. Nitrogen, phosphorus and potassium mass balances for two small fens
surrounded by heavily fertilized pastures. J.Ecol.
Koerselman, W. & B. Beltman 1988. Evapotranspiration from fens in relation to Penman's potential free water
evaporation (Eo) and pan evaporation. Aquatic Botany 31, 307-320.
References
297
Koerselman, W., H. de Caluwe & W.M. Kieskamp in press. Denitrification and dinitrogen fixation in two quaking fens in
the Vechtplassen area, The Netherlands. Biogeochemistry.
Koerselman, W., D. Claessen, P. ten Den & E. van Winden in press. Dynamic hydrochemical gradients in fens in relation
to the vegetation. J.Ecol.
Kooijman, A.M. 1985. Een onderzoek naar de nutrintenhuishouding van een trilveen in "De Weerribben". Rapp. Vakgr.
Bot. Oec., Sectie Landschapsoecologie, RUU, 77p.
Kroes, J. & T. Hol 1979. Het Land van Vollenhove; een historisch-geografische studie van het Noordwest-Overijsselse
cultuurlandschap. Report. Anonymus, Zwolle, 296p.
Kruijne, A.A., D.M. de Vries & H. Mooi 1967. Bijdrage tot de oecologie van de Nederlandse graslandplanten (Engl.
summ.: Contribution to the ecology of the Dutch grassland plants). Versl. Landbouwk. Ond. 696, Med. 338 IBS.
Pudoc, Wageningen, 65p.
Kuiper, P. & Kuiper, C. 1958. Verlandingsvegetaties in Noordwest-Overijssel. Kruipnieuws 20(1): 1-19, also in J.C.
Smittenberg (red.), Plantengroei in enkele Nederlandse landschappen. Bondsuitgeverij Jeugdbonden voor
Natuurstudie, Amsterdam, 357-401.
Kuiper, P. & C. Lapr 1956. Verslag excursies 1956, I, II. Mscr. Staatsbosbeheer, Utrecht, 34p., krtn.
Lambert, J.M., J.N. Jennings, C.T. Smith, C. Green & J.N. Hutchinson 1960. The making of the Broads. A
reconsideration of their origin in the light of new evidence. Research Series 3. The Royal Geographical Society,
London, 153.
Landwehr, J. (m.m.v. J.J. Barkman) 1966. Atlas van de Nederlandse bladmossen. Bibliotheek KNNV 15. Koninklijke
Nederlandse Natuurhistorische Vereniging, Amsterdam, 504p.
Landwehr, J. (m.m.v. S.R. Gradstein & H. van Melick) 1980. Atlas van de Nederlandse levermossen. Bibliotheek KNNV
27. Koninklijke Nederlandse Natuurhistorische Vereniging, 287p.
Leentvaar, P. 1960. Hydrobiologische waarnemingen in het plassengebied van N.W.-Overijssel. Rapp. Rijksinstituut voor
Veldbiologisch Onderzoek t.b.v. het Natuurbehoud, Zeist, 19p., bijl.
Levine, I.N. 1978. Physical Chemistry. McGraw-Hill, New York etc., 847p.
Lindenbergh, A.G. 1956. De watervoorziening van de lichte gronden in de oostelijke helft van de Noordoostpolder (Engl.
Summ.: The irrigation of the light soils in the eastern half of the North-Eastern polder). Van Zee tot Land 17.
Tjeenk Willink, Zwolle, 51p.
Londo, G. 1975. Nederlandse lijst van hydro-, freato- en afreatofyten; See G. Londo 1988. Rapp. Rijksinstituut voor
Natuurbeheer, Leersum.
Londo, G. 1988. Nederlandse freatofyten (Engl. Summ.: Dutch freatophytes). Pudoc, Wageningen, 108p.
Lumkes, M. 1976. Kartering van Watervegetaties in verband met milieu-eisen van krabbescheer. Mscr., 16p.
Lyklema, L. & G. van Straten 1977. De waterbeweging in het plassengebied van N.W.-Overijssel. H2O 10(16), 360-363.
Malmer, N. 1962. Studies on mire vegetation in the Archaean area of South-Western Gtaland (South-Sweden), II,
Distribution and seasonal variation in elementary constituents on some mire sites. Opera Botanica 7(2), 1-67.
Malmer, N. 1963. Studies on mire vegetation in the Archaean area of Southwestern Gtaland (South Sweden), III, On the
relation between specific conductivity and concentrations of ions in the mire water. Botaniska Notiser 116(2), Lund,
149-256.
Margadant, W.D. & H. During 1982. Beknopte flora van de Nederlandse Blad- en Levermossen. Bibliotheek KNNV 28.
Thieme, Zutphen, 517p.
Masing, V. 1984. Estonian bogs: Plant cover, succession and classification. In: P.D. Moore (ed.), European mires.
Academic Press, London etc., 119-148.
Maucha, R. 1932. Hydrochemische Methoden in der Limnologie. Die Binnengewsser 12. Schweizerbart, Stuttgart, 173p.
McIntosch, R.P. 1973. Matrix and plexus techniques. In: R.H. Whittaker (ed.), Ordination and Classification of
Communities, Handbook of Vegetation Science 5. Junk, The Hague, 157-191.
Meijer, W. & R.J. de Wit (eds.) 1955. Kortenhoef, Een veldbiologische studie van een Hollands verlandingsgebied. Cie.
voor de Vecht en het O. en W. plassengebied, Amsterdam, 128p.
Meijer, W. & R.J. de Wit 1955. De verlanding in het kwelgebied van het Hol. In: W. Meijer & R.J. de Wit (red.),
Kortenhoef, Een veldbiologische studie van een Hollands verlandingsgebied. Cie. voor de Vecht en het O. en W.
plassengebied, Amsterdam, 45-49.
Molenaar, W.J., R. van Diggelen & A.M. Kooijman 1990. Vegetation Succession and Hydrology in De Bollemaat, De
Wieden. Acta Botanica Neerlandica 39(3), 318-319.
Mnkemeyer, W. 1927. Die Laubmoose Europas, Ergnzungsband, Andreaeales-Bryales. Dr. L. Rabenhorsts
Kryptogamen-Flora von Deutschland, sterreich und der Schweiz IV. Akad. Verlagsgesellsch., Leipzig, 960.
Moore, P.D. (ed.) 1984. European Mires. Academic Press, London etc., 367p.
Moore, P.D. & D.J. Bellamy 1973. Peatlands. Elek Science, London, 221p.
Moss, B. 1978. The ecological history of a mediaeval man-made lake, Hickling Broad, Norfolk, United Kingdom.
Hydrobiologia 60(1), 23-32.
Mueller-Dombois, D. & H. Ellenberg 1974. Aims and methods of vegetation ecology. Wiley, New York etc., 547p.
Muis, A. 1974. Windmoleninventarisatie in "De Weerribben". Rapp. BCS, Velp, 34p., map.
298
References
Mller, K. 1954-'57. Die Lebermoose Europas, eine Gesamtdarstellung der europischen Arten. Dr. L. Rabenhorst's
Kryptogamen-Flora von Deutschland, sterreich und der Schweiz VI(1,2). Akad. Verlagsgesellsch., Geest &
Portig, Leipzig, 1365p.
Oberdorfer, E. 1979. Pflanzensoziologische Exkursionsflora. Ulmer, Stuttgart, 997p.
Oosterbroek, P. & B.J. Post 1977. Een vegetatiekartering m.b.v. false-colour luchtfoto's in het C.R.M.-reservaat "De
Weerribben" (N.W.-Overijssel). Int.Rapp. 40. Hugo de Vries-Lab., UvA, Amsterdam, 45p., bijl.
Pa czy ski, A. 1984. Natural differentiation of plant communities in relation to hydrological conditions of the Biebrza
valley. Polish Ecological Studies 10(3-4), 347-385.
Pallis, M. 1916. The structure and history of plav: the floating fen of the Danube. J.Linn.Soc.(Bot.) 43, 233-290.
Penman, H.L. 1963. Vegetation and Hydrology. Techn. Comm. 53. Commonwealth Bureau of Soils, Harpenden;
Commonwealth Agricultural Bureaux,Farnham Royal, Bucks, 124p.
Pielou, E.C. 1977. Mathematical Ecology. Wiley, New York etc., 385p.
Pietsch, W. 1982. See Anonymus 1982.
Raeymaekers, G. 1975. Een vegetatiekundig onderzoek in verband met de waterhuishouding van enige kraggegebieden in
het Natuurmonument "De Wieden" (N.W.-Overijssel). Int.Rapp.14. Hugo de Vries-Lab., UvA, Amsterdam, 21p.,
bijl.
Raeymaekers, G. 1977. Het verband tussen waterhuishouding en vegetaties in waterrijke gebieden. Dumortiera 7-8,
38-47.
Rengersen, J. 1980. Een verspreidingsmodel voor conservatieve stoffen in het oppervlaktewater van NW-Overijssel.
Report. T.H.-Twente, Afd. Chemische Technologie, Onderzoeksgr. Techn. Milieubeh., 122.
Rijksbureau voor Drinkwatervoorziening 1938. Rapport inzake een geo-hydrologisch onderzoek ten behoeve van den
Noordoostpolder. Rapp. Rijksbureau voor Drinkwatervoorziening.
Rijksinstituut voor Natuurbeheer 1979. Levensgemeenschappen. Natuurbeheer in Nederland. Pudoc, Wageningen, 392p.
Robinson, R.A. & R.H. Stokes 1955-'70. Electrolyte Solutions (revised edition). Butterworth, London, 560.
Roelofs, J.G.M. & M.J.R. Cals 1989. Effecten van de inlaat van gebiedsvreemd water op de waterkwaliteit en
vegetatie-ontwikkeling in laag- en hoogveenplassen. In: J.G.M. Roelofs (red.), Aanvoer van gebiedsvreemd water:
omvang en effecten op oecosystemen. Faculteit Natuurwetenschappen, KUN, 72-85.
Rose, C.W. 1969. Agricultural physics. Pergamom, Oxford etc., 320p.
Ruuhijrvi 1983. The Finnish mire types and their regional distribution. In: A.J.P Gore (ed.), Mires: swamp, bog, fen and
moor, Regional studies. Ecosystems of the world 4b. Elsevier, Amsterdam etc., 47-67.
Rybn ek, K. 1964. Die Braunmoorgesellschaften der Bhmisch-Mhrischen Hhe (Tsjechoslowakei) und die
Problematik ihrer Klassifikation. Preslia 36, 403-415.
Rybn ek, K. 1974. Die Vegetation der Moore im sdlichen Teil der Bhmisch-Mhrischen Hhe. Vegetace SSR A6.
Academia, Praha, 243p., apps.
Rybn ek, K. 1984. The vegetation and development of Central European mires. In: P.D. Moore (ed.), European Mires.
Academic Press, London etc., 177-201.
Schmidt, G. 1969. Vegetationsgeographie auf kologisch-soziologischer Grundlage. Teubner, Leipzig, 396p.
Schoeller, H. 1962. Les eaux souterraines. Masson, Paris, 642p.
Schurer, K. & J.C. Rigg 1980. Grootheden en eenheden in de landbouw en de biologie. Pudoc, Wageningen, 121p.
Sculthorpe, C.D. 1967. The Biology of Aquatic Vascular Plants. Arnold, London, 610p.
Seddon, B. 1967. The lacustrine environment in relation to macrophytic vegetation. E.J. Cushing & H.E. Wright, Jr.
(eds.), Quarternary Palaeoecology. Proceedings of the VII Congress of the Int. Ass. for Quarternary Res., 7. Yale
University Press, New Haven and London, 205-215.
Segal, S. 1965. Een vegetatieonderzoek van de hogere waterplanten in Nederland. WM 57. Kon. Ned. Nat.-hist. Ver.,
Hoogwoud, 80p.
Segal, S. 1966. Ecological studies of peat-bog vegetation in the North-Western part of the province of Overijssel (The
Netherlands). Wentia 15, 109-141.
Segal, S. 1966. Oecologie van hogere waterplanten. Vakblad voor Biologen 46, 138-149.
Segal, S. 1968. Schwierigkeiten bei der Systematik von Moorgesellschaften. In: R. Txen (ed.), Pflanzensoziologische
Systematik. Int. Symp. Stolzenau/Weser 1964. Junk, Den Haag, 220-229.
Segal, S. & M.C. Groenhart 1967. Het Zuideindigerwiede, een uniek verlandingsgebied. Gorteria 3, 165-181.
Senden, W. 1980. Geo-elektrisch onderzoek Stavoren/Steenwijk. Rapport GF-124. DGV-TNO, 27p., bijl.
Sheffield, C. 1985. Selecting band combinations from multispectral data. Photogr.Engineering and Remote Sensing 51(6),
681-687.
Shimwell, D.W. 1971. Description and Classification of Vegetation. Sidgwick & Jackson, London, 322p.
Sikora, L.J. & D.R. Keeney 1983. Further aspects of soil chemistry under anaerobic conditions. In: A.J.P. Gore (ed.),
Mires: swamp, bog, fen and moor, General studies. Ecosystems of the world 4a. Elsevier, Amsterdam etc., 247-256.
Sjrs, H. 1983. Mires of Sweden. In: A.J.P. Gore (ed.), Mires: swamp, bog, fen and moor, Regional studies. Ecosystems
of the world 4b. Elsevier, Amsterdam etc., 69-94.
Slicher van Bath, B.H. 1957. Een samenleving onder spanning. Van Gorcum, Assen, 768p.
References
299
Spiegel, M.R. 1972. Theory and problems of statistics in SI units. Schaum's Outline Series. McGraw-Hill, New York etc.,
359p.
Staatsbosbeheer 1974-'75. Vegetatiekaart van De Weerribben (see Heringa & van der Sloot 1973). Unpublished, Zwolle.
Staatsbosbeheer 1988. Beheersplan voor de periode 1988-1998 (De Weerribben). Rapp. Staatsbosbeheer, Utrecht/Zwolle,
273p., bijl.
Stallman, R.W. 1965. Steady one-dimensional fluid flow in a semi-infinite porous medium with sinusoidal surface
temperature. Journal of Geophysical Research 70(12), 2821-2827.
Stanek, W. & I.A. Worley 1983. A terminology of virgin peat and peatlands. C.H. Fuchsman & S.A. Spigarelli (eds.),
Proc. Int.Symp. on Peat Utilization. Bemidji State University, Bemidji, 75-102.
Stegeman, J.H.J. 1968. Onderzoek van de Bryophyten in verlandingsreeksen in Noord-West Overijssel in afhankelijkheid
van milieu en vegetatie. Rapp. Hugo de Vries-Laboratorium/RIVON, Amsterdam, 31p., bijl.
Straathof, N. 1976. Praktijkverslag over hydrologisch onderzoek in de Weerribben. Int.Rapp. Staatsbosbeheer/Rijksinstituut voor Natuurbeheer, Zwolle, 11p., bijl.
Stumm, W. & J.J. Morgan 1970. Aquatic Chemistry. Wiley, New York etc., 583p.
Stuyfzand, P.J. 1983. De berekening van het elektrisch geleidingsvermogen van natuurlijke wateren: een zeer
nauwkeurige methode met voorbeelden van toepassing. KIWA rapport SWE-83.001, Nieuwegein, 40p.
Stuyfzand, P.J. 1983. Een zeer nauwkeurige berekening van het elektrisch geleidingsvermogen ter controle en aanvulling
van wateranalyses. H2O 16(16), 358-361, 363.
Succow, M. 1988. Landschaftskologische Moorkunde. Gustav Fisher, Jena, 340p.
Suzuki, S. 1960. Percolation measurements based on heat flow through soil with special reference to paddy fields. Journal
of Geophysical Research 65(9), 2883-2885.
Ter Wee, M.W. 1962. The Saalian glaciation in The Netherlands. Meded. Geol. St., N.S. 15, 57-77.
Ter Wee, M.W. 1966. Toelichting bij de geologische kaart van Nederland, blad Steenwijk Oost. Geologische Stichting,
Afd. Geol. Dienst, Haarlem, 106p., maps.
Thomson, A. & H.A.P. Ingram 1981. Notes. In: K.E. Ivanov, Water movement in mirelands (transl. by A. Thomson &
H.A.P. Ingram). Academic Press, London etc, 252-257.
Tilman, D. 1982. Resource competition and community structure. Princeton University Press, Princeton.
Tomaszewska, K. 1988. Plant cover of peatland in the Biebrza river valley and its changes determined on the basis of
aerial photographs. Proc. VIII Int. Peat Congress (1). International Peat Society, Helsinki, 199-207.
Touber, L. 1973. Hydrologisch Onderzoek in enige verlande petgaten in het C.R.M. Reservaat "De Weerribben",
N.W.-Overijssel. Int.Rapp.3. Hugo de Vries-Lab., UvA, Amsterdam, 34p., bijl.
Touw, A. & W.V. Rubers 1989. De Nederlandse Bladmossen. Bibliotheek KNNV 50. KNNV, Utrecht, 532p.
Uil, H. & E. de Heer (1985). Grondwaterkaart van Nederland, Stavoren/Steenwijk. Rep. GWK 34. DGV-TNO, 91p.,
maps.
Van der Linden, M.J.H.A. 1980. Nitrogen economy of reed vegetation in the Zuidelijk Flevoland Polder. Acta
Oecologica/Oecol. Plantarum 1(15), 219-230.
Van der Linden, M.J.H.A. 1986. Phosphorus economy of reed vegetation in the Zuidelijk Flevoland Polder (The
Netherlands): seasonal distribution of phosphorus among shoots and rhizomes and availability of soil phosphorus.
Acta Oecologica/Oecologia Plantarum 7(21),4, 397-405.
Van der Meijden, R., E.J. Weeda, F.A.C.B. Adema & G.J. de Joncheere 1983. Heukels/Van der Meijden Flora van
Nederland. Wolters-Noordhoff, Groningen, 583p.
Van der Perk, J.C. & M.J. Smit 1975. Een hydrologisch onderzoek ten behoeve van het natuurbeheer in de "Wieden".
Int.Rapp.17. Hugo de Vries-Lab., UvA, Amsterdam, 56p., bijl.
Van der Toorn, J. 1972. Variability of Phragmites australis (Cav.)Trin. ex Steudel in relation to the environment. Van Zee
tot Land 48. Staatsuitgeverij, 's-Gravenhage, 122p.
Van Leeuwen, C.G. 1966. A relation theoretical approach to pattern and process in vegetation. Wentia 15, 25-46.
Van Wirdum, G. 1972. Beknopt verslag over de eerste fase van een onderzoek naar de waterhuishouding in het
kraggelandschap van De Weerribben en de betekenis daarvan voor de vegetatie. Int.Rapp. Hugo de Vries-Laboratorium, UvA,Amsterdam, 30p.
Van Wirdum, G. 1973. Het verband tussen de successie en enige veranderingen in de eigenschappen van het water in de
Weerribben. Samenvatting van een voordracht gehouden op de 81e dag voor het vegetatie-onderzoek van de
Koninklijke Nederlandse Botanische Vereniging op 19-12-1972 te Wageningen ("laagveendag"). Jaarboek KNBV
1972, 49-52.
Van Wirdum, G. 1977. Natuurgebieden Noord-West Overijssel. In: N.J.N. Bunnik et al., Onderzoek naar de
toepassingsmogelijkheden van multispectrale scanning. Publ.44, Nederlandse Interdepartementale Werkgroep voor
het Applicatieonderzoek van Remote Sensing Technieken (NIWARS), Delft, 314-331.
Van Wirdum, G. 1978. Een landschapsecologische basis voor de normering van de waterkwaliteit. Int.Not. Rijksinstituut
voor Natuurbeheer, Leersum, 8p.
Van Wirdum, G. 1979. Dynamic aspects of trophic gradients in a mire complex. Proc. and Inf. 25. CHO-TNO, The
Hague, 66-82.
300
References
Van Wirdum, G. 1980. Eenvoudige beschrijving van de waterkwaliteitsverandering gedurende de hydrologische

kringloop. J.C. Hooghart (ed.) Waterkwaliteit in grondwaterstromingsstelsels. CHO-TNO, rapporten en nota's 5,
Den Haag, 118-143.
Van Wirdum, G. 1981. De waterkwaliteit in de natuurgebieden in het waterschap Vollenhove. Mscr. Rijksinstituut v.
Natuurbeheer, Leersum, 72p.
Van Wirdum, G. 1981. Linking-up the natec subsystem in models for the watermanagement. Water resources
management on a regional scale, Proc. and Inf. 27. CHO-TNO, The Hague, 108-128.
Van Wirdum, G. 1982. Design for a land-ecological survey of nature protection. S.P. Tjallingii & A.A. de Veer (eds.)
Perspectives in Landscape Ecology, contributions to research, planning and management of our environment.
Pudoc, Wageningen, 245-251.
Van Wirdum, G. 1982. The ecohydrological approach to nature protection. Annual Report 1981. Research Institute for
Nature Management, Arnhem etc., 60-74.
Van Wirdum, G. 1983. De mosseninventarisatie van De Weerribben. Buxbaumiella 14. Bryologische en Lichenologische
Werkgroep van de KNNV, 10-47.
Van Wirdum, G. 1984. Development of techniques for ecohydrological research. Annual Report 1983. Research Institute
for Nature Management, Arnhem, Leersum and Texel, 21-25.
Van Wirdum, G. 1985. Verschil moet er blijven (Eng. Summ.: There must be difference). De Levende Natuur 86(3,5),
97-101, 193.
Van Wirdum, G. 1986. Water-related impacts on nature protection sites. Proc. and Inf. 34. CHO-TNO, The Hague, 27-57.
Van Wirdum, G. 1987. The rle of spatial variety in the application of ecology to nature protection and ecology. Annual
Report 1986. Research Institute for Nature Management, Arnhem, Leersum and Texel, 31-33.
Van Wirdum, G. 1989. Ecohydrologische aspecten van waterinlaat in laagvenen. Salient features of terrestrializing fens in
The Netherlands. In: J.G.M. Roelofs (red.), Aanvoer van gebiedsvreemd water: omvang en effecten op
ecosystemen. Faculteit Natuurwetenschappen, KUN, Nijmegen, 52-71.
Van Wirdum, G., A.J. den Held & M. Schmitz in press. Salient features of terrestrializing fens in The Netherlands In:
J.T.A. Verhoeven (ed.) Fens and bogs in The Netherlands. Kluwer.
Van Wirdum, G. & R.H. Kemmers 1990. Hydrologie, bodem en natuurontwikkeling. In: F. Berendse (red.),
Natuurontwikkeling en landbouw. Agrobiologische Thema's 1. Cabo, Wageningen, 45-65.
Van Wirdum, G. & D. van Dam 1984. Standplaats en plant; bewerking ecologische indicatiewaardenlijsten.
SWNBL-rapport 1. Studiecommissie Waterbeheer Natuur, Bos en Landschap, Utrecht, 116p.
Van Zon-van Wagtendonk, A.M. 1965. Vegetatiekartering van een gedeelte van het natuurreservaat "de Weerribben" te
Oldemarkt (NW-Overijssel). Rapport. Hugo de Vries-Laboratorium/RIN, Amsterdam, 40p., bijl.
Van Zon-van Wagtendonk, A.M. 1969. Vegetatiekartering van het natuurreservaat "De Weerribben" in NW-Overijssel.
Rapp. Hugo de Vries-Laboratorium/RIN, Amsterdam, 39p., bijl.
Veenenbos, J.S. 1950. De bodemgesteldheid van het gebied tussen Lemmer en Blokzijl in het randgebied van de
Noordoostpolder (Engl.Summ.: Soil conditions of the area between Lemmer and Blokzijl in the border area of the
Noordoost-Polder). Verslagen van landbouwkundige onderzoekingen 55.12, De bodemkartering van Nederland, V.
Staatsuitgeverij, 's-Gravenhage, 162p.
Vegt, J.J. 1978. Verdamping, berging en indringing van boezemwater in het moerasgebied "De Weerribben". Int.Rep.,
RIN/LH, Leersum, 51p., bijl.
Vereniging voor de Statistiek in Nederland 1870. Algemeene Statistiek van Nederland. Sijthoff, Leiden.
Verhoeven, J.T.A., & H.H.M. Arts 1987. Nutrient dynamics in small mesotrophic fens surrounded by cultivated lands, II,
N and P accumulation in plant biomass in relation to the release of inorganic N and P in the peat soil. Oecologia
(Berlin) 72, 557-561.
Verhoeven, J.T.A., A.M. Kooijman & G. van Wirdum 1988. Mineralization of N and P along a trophic gradient in a
freshwater mire. Biogeochemistry 6, 31-43.
Verschoor, A. 1978. Vegetatiekartering met behulp van luchtfoto's in het CRM-reservaat "De Weerribben"
(N.W.-Overijssel). Int. Rapp. 56. Hugo de Vries-lab., UvA, 31p.
Volker, A. 1948. Geo-hydrologische gesteldheid van het oude land langs de Noordoostpolder tussen Lemmer en Blokzijl.
Rapport. Anonymus, 22p., bijl.
Vromen, H, L. Klamer & J. de Vries 1974. Voortgezet Hydrologisch onderzoek in enige verlande petgaten in het
C.R.M.-reservaat "De Weerribben". Int.Rap., Staatsbosbeheer Zwolle, 20p., bijl.
Waldhauer, F.D. 1982. Feedback. Wiley, New York etc., 651p.
Wassen, M.J., A. Barendregt, M.C. Bootsma & P.P. Schot 1989. Groundwater chemistry and vegetation of gradients from
rich fen to poor fen in the Naardermeer (The Netherlands). Vegetatio 79, 117-132.
Weeda et al. 1988. See Weeda et al. 1985-'88, 3.
Weeda, E.J., R. van der Meijden & P.A. Bakker 1990. Rode lijst van de in Nederland verdwenen en bedreigde planten
(Pteridophyta en Spermatophyta) over de periode 1.I.1980-1.I.1990 (Engl. summ.: Red Data List of the extinct,
endangered and vulnerable plants in the Netherlands in the period 1960-1990. Gorteria 16(1), 2-26.
Weeda, E.J., R. Westra, Ch. Westra & T. Westra 1985-'88. Nederlandse oecologische flora, 1,2,3. IVN, 910p.
References
301
Westhoff, V., P.A. Bakker, C.G. van Leeuwen, & E.E van der Voo 1971. Wilde planten. Flora en vegetatie in onze
natuurgebieden 2, Het lage land. Natuurmonumenten, Amsterdam, 304.
Westhoff, V. & A.J. den Held 1969. Plantengemeenschappen in Nederland. Bibliotheek KNNV 16, RIVON Verhandeling
5. Thieme, Zutphen, 324p.
Wheeler, B.D. 1978. The wetland plant communities of the River Ant Valley, Norfolk. Transactions of the Norfolk &
Norwich Naturalists' Society 24(4), 153-187.
Wheeler, B.D. 1980. Plant communities of rich-fen systems in England and Wales. I. Introduction. Tall sedge and reed
communities. J.Ecol.68, 365-395.
Wheeler, B.D. 1980. Plant communities of rich-fen systems in England and Wales. II. Communities of calcareous mires.
J.Ecol.68, 405-420.
Wheeler, B.D. 1984. British Fens: A Review. In: P.D. Moore (ed.), European Mires. Academic Press, London etc., 237281.
Wheeler, B.D. & K.E. Giller 1982. Species richness of herbaceous fen vegetation in Broadland, Norfolk in relation to the
quantity of above-ground plant material. Journal of Ecology 70, 179-200.
Zijlstra, G. 1981. Some remarks on the Cirsio-Molinietum and the Caricion davallianae. Proc. Kon.Ned.Akad.v.Wetensch. C84(1), 89-106.
Zuiveringschap West-Overijssel, Technologische Dienst 1981-'88. Kwaliteit oppervlaktewater in het gebied van het
zuiveringschap West-Overijssel. (Half)jaarrapporten. Zuiveringschap West-Overijssel, Zwolle.
302
References
SAMENVATTING
Vegetatie en waterhuishouding van trilvenen
In de verlanding van petgaten in het laagveengebied van Nederland neemt het kraggestadium
een belangrijke plaats in. In sommige gevallen vindt een snelle ontwikkeling plaats van
rietlanden naar door veenmossen of houtige gewassen gedomineerde vegetaties, terwijl in
andere gevallen zeer langdurig slaapmossen, waaronder Schorpioenmos, en kleine zeggen het
aspect bepalen. Dit laatste vegetatietype en verschillende daarvoor karakteristieke plantesoorten
worden veelal als kwelindicatoren beschouwd. Die veronderstelling vormt in dit proefschrift de
leidraad voor een nader onderzoek naar de waterhuishoudkundige factoren die bepalend zijn
voor de ontwikkeling van deze trilvenen par excellence. Het onderzoek voerde van een analyse
van het door de vegetatie gendiceerde milieu, via een hydrologische beschrijving van de
omgeving, naar een poging tot kwantitatief begrip hoe het kennelijke milieu van de vegetatie
onder invloed van de waterhuishouding wordt bepaald.
Een gedetailleerde analyse van de bronliteratuur (hoofdstuk 2) bracht aan het licht dat het
optreden van kwel op de "typelokaties" en het veronderstelde verband met botanische
"kwelindicaties" niet overtuigend zijn bewezen. In het thans uitgevoerde onderzoek kon op
verschillende "klassieke" lokaties geen kwel, maar juist wegzijging worden aangetoond. Voor
n gebied, "De Stobbenribben", is dit in de hoofdstukken 7-10 in detail beschreven.
Uit de internationale ecologische literatuur blijkt dat de "kwelindicatoren" niet eenduidig op
kwel wijzen, maar wel op een goede basenverzorging van het veen, in het bijzonder op
Samenvatting
303
kalkrijkdom (hoofdstuk 3). In dit proefschrift wordt een en ander nader besproken op grond van
onderzoek in Noordwest-Overijssel, een sleutelgebied voor de "kweltheorie".
In hoofdstuk 4 wordt een korte geografische en historische beschrijving van dit gebied gegeven.
De waterhuishouding van het noordelijk deelgebied "De Weerribben" wordt in detail
behandeld in hoofdstuk 5. Hier komt thans geen noemenswaardige uitstroming van grondwater
voor. Vr de drooglegging van de Noordoostpolder in 1941 zou er plaatselijk enige kwel
kunnen zijn geweest, maar de uitstroming van grondwater in die periode moet toch
ondergeschikt zijn geweest aan de toestroming van oppervlaktewater en ook aanzienlijk geringer
dan de huidige wegzijging. Belangrijke nadere aanwijzingen komen voort uit een analyse van de
samenstelling van het grondwater sinds 1935. In het centrum van het gebied bevonden zich toen
nog restanten zwak brak water in het bovenste gedeelte van het zandpakket onder De
Weerribben. Ook thans zijn er nog belangrijke brakwatervoorkomens, maar kalkrijk zoet
grondwater is door de inzijging tot grote diepte teruggedrongen. De aanwezigheid van zulk
water in het trilveenmilieu moet in de laatste eeuw voornamelijk bepaald zijn geweest door de
aanvoer van oppervlaktewater.
Gegevens over de oppervlaktewaterkwaliteit zijn pas vanaf ca. 1960 beschikbaar. In de
periode 1960-1987 varieerde de samenstelling van het oppervlaktewater ten gevolge van de
kunstmatige beheersing van het boezempeil door inlaat en uitmalen. Vooral in de zeventiger
jaren is de invloed van verontreinigd inlaatwater, uiteindelijk grotendeels afkomstig uit de Rijn,
sterk toegenomen. De detail-uitvoering van de waterbeheersing was in die periode oorzaak van
de aanvoer van zeer grote hoeveelheden inlaatwater, in het bijzonder gedurende de "droge jaren"
1975-'76. Door opeenvolgende inlaat-episodes raakte het water tussen 1973 en 1976 in vrijwel
de gehele boezem sterk vermengd met dit inlaatwater. Dit werd pas zeer geleidelijk
teruggedrongen door bij het peilbeheer naar minimalisering van inlaat te streven en door het
uitblijven van extreme droogte. Bij het onderzoek naar de samenstelling van het water werd
gebruik gemaakt van een tijdens dit onderzoek ontwikkelde "fenomenologische" methode die
berust op een kwantificering van de gelijkenis van het water met de extreme typen uit de
waterkringloop, regenwater, kalkrijk grondwater, en zeewater (appendix D).
In hoofdstuk 6 is het optreden van de zogenaamde kwelindicatoren in de vegetatie van De
Weerribben besproken. Ook wanneer de ecologische indicatie van deze kwelindicatoren zelf
buiten beschouwing wordt gelaten wijst het optreden van vegetatie met dergelijke soorten op een
betere basenverzorging. n van de meest eenduidige "kwelindicatoren", Schorpioenmos, blijkt
zich sinds 1960 in De Weerribben nadrukkelijk te hebben uitgebreid over een deel van het
gebied waar zeker geen kwelinvloeden zijn. Er zijn echter duidelijke aanwijzingen dat in dit
deelgebied in dezelfde periode de invloed van zwak brak water is teruggedrongen, terwijl
voldoende jonge verlandingsstadia aanwezig waren voor de vestiging van Schorpioenmos en
enkele andere "kwelindicatoren". Vermoed wordt dat de vroegere aanwezigheid van brak water
nog steeds in de basenverzorging van het veen doorwerkt, terwijl het chloride-gehalte is gezakt
tot een niveau dat voor zoetwater-planten geschikt is.
De hoofdstukken 7-10 gaan over een case-study in De Stobbenribben, een trilveen-complex dat
in de bronliteratuur over de "kweltheorie" wordt genoemd als een goed voorbeeld van het
optreden van kwel en kwelindicatoren. Door de ligging en opbouw was dit complex bovendien
bijzonder geschikt om de invloed van de waterhuishouding na te gaan. Sinds de aanleg van een
polder op zeer korte afstand, omstreeks 1955, is er in dit gebied sprake van een sterke
wegzijging. Vanuit een sloot aan n zijde stroomt als gevolg hiervan oppervlaktewater toe. Er
bevindt zich een zonering van een productieve rietvegetatie bij de aanvoersloot, via een
trilveenvegetatie met veel "kwelindicatoren", naar een door veenmossen gedomineerde en door
randeffecten benvloede vegetatie. De ecologische indicatie van het middengedeelte met
304
Samenvatting
"kwelindicatoren" wijkt van de rest af door een nadrukkelijk voedselarm-basenrijke component

(hoofdstuk 7).
In hoofdstuk 8 is een methode beschreven om met behulp van het warmtetransport in het
veen de eventuele verticale waterbeweging te kwantificeren. Hoewel aan deze methode
duidelijke bezwaren kleven door de vermoedelijk sterke laterale waterbeweging, kon toch een
consistent beeld worden verkregen door de metingen van de veentemperatuur regelmatig te
herhalen en het verloop tot een diepte van 1,8 m per decimeter te beschouwen. Over korte
afstand varieert de gevonden waarde voor de wegzijging van 2-10 mm/d met een gemiddelde
van ongeveer 5 mm/d.
Door herhaald meten van de soortelijke electrische geleiding in het veen kon worden
vastgesteld dat de laterale waterbeweging vanuit de aanvoersloot geconcentreerd is in een
preferent kanaal juist onder de kragge (hoofdstuk 9). De "kwelindicatoren" komen voor in een
zone waar gedurende de winter het aangevoerde slootwater weer enigszins wordt teruggedrongen.
In hoofdstuk 10 wordt een poging gedaan de wateraanvoer en het effect daarvan op de
chemische samenstelling van het veenwater te kwantificeren. Hiervoor is een eenvoudig model
opgesteld dat met de waarnemingen in overeenstemming is. Volgens dit model worden de
kalkrijk-voedselarme omstandigheden in het middengedeelte van het trilveen in stand gehouden
door de aanvoer van basen met het instromende water. De voedselrijkdom van het slootwater
wordt in de productieve rietvegetatie tot een "achtergrondsniveau" teruggedrongen. Aan het
andere einde van de gradient is de invloed van regenwater overheersend. De doorlatendheid van
het preferente stroomkanaal, de structuur van de kragge, de intensiteit van de inzijging, en de
aanvoercapaciteit van de sloot zijn aldus bepalend voor de zonering in de vegetatie. Waarschijnlijk vindt in De Stobbenribben thans een toename plaats van de invloed van regenwater. Dit
wordt veroorzaakt door het dichtraken van de aanvoersloot en door de voortschrijdende
verlanding, waardoor de uitwisseling tussen de kragge en het preferente stroomkanaal daaronder
vermindert. Elders in het onderzoeksgebied heeft zich een dergelijke "atmotrofiring" reeds
volledig voltrokken, vooral onder invloed van het dichtraken van sloten.
In hoofdstuk 11 wordt een uitgebreide samenvatting en bredere discussie van de bereikte
resultaten gegeven. De basenverzorging van het veen is vermoedelijk bepalend voor de
verschillen tussen de belangrijkste typen laagveenvegetatie. De waterhuishouding speelt hierbij
een sturende rol. In Noordwest-Overijssel is het kalkrijk-voedselarme type laagveenvegetatie,
met de zogenaamde kwelindicatoren, niet gebonden aan het uittreden van grondwater, maar
veelal juist aan de combinatie van wegzijging en laterale instroming van oppervlaktewater.
Wellicht schept de vroegere aanwezigheid van brak water een bijzonder gunstig uitgangsmilieu.
Naarmate de wegzijging geringer is kan een snellere verzuring optreden. De Stobbenribben is
als voorbeeldgebied representatief omdat de "kwelvegetatie" hier, ondanks de geconstateerde
wegzijging, nog steeds vrijwel optimaal voorkomt. Elders in Noordwest-Overijssel en in het
Vechtplassengebied is onder andere de karakteristieke moslaag met Schorpioenmos nog slechts
fragmentair ontwikkeld, en treden verschillende plantesoorten van meer voedselrijk milieu op de
voorgrond.
Hoewel het onderzoek geheel betrekking heeft op trilvenen, spelen de beschreven mechanismen
ongetwijfeld ook elders een belangrijke rol. De "kwelindicatoren", waarvan hier is aangetoond
dat zij niet noodzakelijkerwijs op kwel in strikte zin wijzen, en die dat in Noordwest-Overijssel
meestal ook niet doen, zijn gebonden aan een combinatie van vrij voedselarme (blijkend uit een
geringe biomassaontwikkeling van de vegetatie) en toch basenrijke omstandigheden. Deze
Samenvatting
305
komen voor waar in de bovenste 0-30 cm van het veen regenwater-invloeden tenminste
periodiek belangrijk zijn, terwijl een definitieve uitspoeling van basen verhinderd wordt door de
aanvoer van basenrijk grond- of oppervlaktewater. Dit treedt in ons klimaat van nature op in een
zone tussen grote voedsel- en basenarme landschapselementen (hoogvenen, heiden) en rivieren
of beken, eventueel zee-invloeden. Wanneer geen afvoer van biomassa plaats vindt (door
maaien of begrazen) ontwikkelt de vegetatie in deze zone zich bij ons echter tot broekbos,
waarin slechts een beperkt deel van de aangetroffen soorten meer dan uiterst lokaal kan
voorkomen.
306
Samenvatting
SUMMARY
Floating fens, or quagfens, are common in former turbaries in The Netherlands. Some are
subject to a fast development from reed-beds into sphagnaceous vegetation or carr; others, with
a strongly yielding kragge go through a long-continued stage dominated by amblystegiaceous
mosses, such as Scorpidium scorpioides, and slender sedge species (Carex lasiocarpa, C.
diandra). The occurrence of this type of vegetation and the species involved are usually
considered seepage indicators in The Netherlands. That hypothesis was taken as the starting
point for the present research into the hydrologic factors involved. The main line leads from the
environment indicated by plant species, via an empirical analysis of the ambient hydrologic
conditions, into some quantitative understanding of the relationship between the two.
A detailed analysis of the source literature (Chapter 2) revealed that the occurrence of seepage,
and especially of discharging groundwater, had never been convincingly proved, even not for
the type locations, and that the concept of seepage indicators must be considered hypothetical.
In the present survey an infiltration of mire water towards the underlying body of groundwater
was found at several classic locations, as is reported in detail with regard to De
Stobbenribben (see below, Chapters 7-10).
The European literature indicates that the presumed seepage indicators are especially
indicative of a high base state and calcidity of the local peat. A deeper inquiry was made in
Summary
307
North-West Overijssel, a key area for the seepage hypothesis. A geographical and historical
description of this area is presented in Chapter 4.
Hydrological details about the northern part, the nature reserve De Weerribben (inclusive of
De Stobbenribben), are given in Chapter 5. No groundwater discharge of any importance is
presently found in this area. Although this may have been different before the reclamation of the
Noordoost-Polder in 1941, any local discharge of groundwater even in that period must have
been less important than the inflow of surface water, and also considerably less important than
the present infiltration towards the underlying body of groundwater. Important clues come from
an analysis of the groundwater composition in the area since 1935. Slightly brackish water
remaining from former sea influences was present in the underlying sand bottom in the central
part of the area in 1935. Even today slightly brackish water is found at several places and any
calcareous fresh groundwater is only present at a depth of 50 m or more. The high base state
indicated by the quagfen flora must therefore be due to an inflow of surface water.
Not much is known about the chemical composition of the surface water before 1960. Since
then a fluctuation has occurred as a result of artificial discharge and supply for the boezem water
management. The influence of the polluted inlet water, largely originating from the Rhine
system, increased dramatically during the 1970s. The actual supply strategy followed in those
years led to an unnecessarily large influx of polluted water, especially during the droughts of
1975-'76. Between 1973 and 1976 virtually the whole body of boezem water became admixed
with this water. Through a more prudent water management, and favoured by the absence of
extreme droughts, a gradual improvement was realized in the following years. The water
composition was investigated with a variety of methods, inclusive of a quantification of the
similarity of water composition to the extreme types in the hydrologic cycle: rain water,
calcareous groundwater, and sea water (Appendix D). It appears that the surface water before ca
1960 at least temporarily was slightly brackish. Large reclamations of polders and other
influences on the water management brought a type of freshwater into the area which was very
similar to calcareous groundwater in the 1960s and early 70s. The 1970s were dominated by the
effects of the increased inlet of then polluted Rhine water.
The distribution of seepage indicators in the vegetation of De Weerribben is concerned in
Chapter 6. Even when the seepage indicators themselves are not counted, stands of vegetation
comprising such species floristically indicate a higher base state than other stands. Scorpidium
scorpioides, an undebated seepage indicator, appears to have recently invaded a part of the area
where certainly no groundwater discharges. This part of the area has long been slightly brackish
and so provided with a high base state, whereas the desalting apparently has now progressed far
enough to render the area suitable for true freshwater species. This apparently coincided with a
young stage of terrestrialization, favourable for the settlement of quagfen species in the 1960s.
In Chapters 7-10 a case study of De Stobbenribben is concerned. De Stobbenribben is one of the
classic seepage sites, and its topography is especially suitable for eco-hydrologic investigations.
Substantial quantities of mire water appear to leak away from De Stobbenribben into the
underlying strata, especially since a nearby polder was reclaimed ca 1955. As a result of this
leakage an inflow of surface water is generated from a ditch bordering the quagfen parcels of De
Stobbenribben at one narrow end, the open end. This has led to a zonation from eutrophic reed
swamp vegetation near the ditch, via a brownmoss-sedge zone characterized by an abundance of
seepage indicators,
308
Summary
to sphagnaceous vegetation with dwarfshrubs, and a zone with edge effects at the dead end of
the parcels where the inflow of ditch water is small or absent. The species composition of the
middle zone with seepage indicators points towards a notably meso-oligotrophic environment
with a high base state.
Chapter 8 deals with a method to trace the vertical water flow in the quagfen body by means
of observations relating to heat transportation. Although the strong lateral inflow hampers an
exact quantification, consistent results were achieved by frequent measurements of peat
temperature profiles and by calculations pertaining to layers of 10 cm thickness. The apparent
downward seepage so determined roughly varies between 2 and 10 mm/d with an average value
of ca 5 mm/d.
Similarly detailed measurements of the electrical conductivity in the quagfen body showed
that the lateral inflow is concentrated in a preferential flow channel just below the floating
kragge (Chapter 9). The seepage indicators are especially abundant in the middle zone of the
quagfen parcels were the ditch water flowing in in summer is pushed back somewhat by rain
water in winter.
Chapter 10 presents an attempt to quantify the inflow of ditch water and its consequencies in
regard of the chemical composition of the mire water. This is done with a relatively simple
computational model, according to which the inflow of ditch water is essential to maintain a
high base state in the middle part of the quagfen parcels. The nutrient load of the ditch water is
strongly decreased in the reed swamp zone at the open end near the ditch. The dead ends of the
parcels are subject to a dominant influence of rain water. The hydraulic transmissivity of the
preferential flow channel, the structure of the kragge, the intensity of the downward seepage,
and the supply capacity of the ditch thus determine the zonation of the vegetation. It is indicated
that the influence of rain water in De Stobbenribben is presently increasing as a result of
clogging of the ditch and ongoing peat accrual and terrestrialization in the quagfen, decreasing
the water exchange between the kragge and the underlying flow channel. A positive feedback
comes from the increased growth of Sphagnum species. Elsewhere in the investigated area such
an atmotrophication has already lead to entirely atmotrophic vegetational stages, especially
through the terrestrialization of ditches.
Chapter 11 contains a comprehensive summary and broader discussion of the results obtained.
The base state of the peat, as governed by hydrological processes is held responsible for the
main differences found in the vegetation of Dutch quagmires. In North-West Overijssel the
occurrence of meso-oligotrophic, alkaliphilic fen vegetation with many seepage indicators is
mostly not due to discharging groundwater, but often to a combination of downward seepage
with a lateral inflow of surface water. The former presence of slightly brackish water in the area
may have favoured a high base state already in early successional stages. As the seepage
decreases, acidification is accellerated. De Stobbenribben is a representative case since the
seepage vegetation is still almost optimally represented here in spite of the downward seepage
which has been occurring here for more than 35 years already. It is one of the best sites for
Scorpidium scorpioides in The Netherlands at present, since the species has disappeared from
most other quagfen areas, such as De Vechtplassen in the province of Utrecht and several
locations in North-West Overijssel.
Although this report pertains to quagfens, the mechanisms shown must be important in other
types of fen mire as well. The seepage indicators apparently don't necessarily indicate a
groundwater discharge, as they indeed mostly don't in North-West Overijssel, but rather a
combination of relatively nutrient-poor (indicated by a small standing crop), yet base-rich
Summary
309
conditions. Such conditions prevail where rain-water influences are at least periodically
important in the upper 30 cm of the peat, while an ongoing leaching of bases to a greater depth is
prevented by a supply of base-rich ground- or surface water. In our climate this specific
environment may be stable in-between large oligotrophic land units (bogs, heathlands), and
(small) rivers, or even land units with sea-influences. When no biomass is harvested (mowing,
grazing) the vegetation in this intermediate zone mostly develops into carr, where the majority
of the populations of presently abundant red-list species can only locally persist.
310
Summary

Vegetation and Hydrology of Floating Rich-Fens

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Vegetation and Hydrology of Floating Rich-Fens

Uploaded by

Copyright:

Available Formats

Vegetation and hydrology of floating rich-fens

The following corrections and changes were made to avoid confusion:

Vegetation and hydrology of floating rich-fens

ter verkrijging van de graad van doctor

op vrijdag 8 februari 1991, te 15.00 uur

FACULTEIT DER BIOLOGIE

CIP-DATA KONINKLIJKE BILIOTHEEK, DEN HAAG

Wirdum, Geert van

Produktie: Datawyse Maastricht, Ruud Leliveld

What is obviously essential to something, but not well understood,

Human interest in nature is probably because of its structure

Quagfens have a notable, but weak and unstable structure

Early in 1969 Dr.S.Segal introduced me to the study of quagfens in North-West

PUBLISHED EVIDENCE OF SEEPAGE IN RICH-FEN QUAGMIRES (QUAGFENS)

SITE PROPERTIES INDICATED BY THE FLORA OF QUAGFENS

THE AREA OF NORTH-WEST OVERIJSSEL

ASPECTS OF THE HYDROLOGY OF DE WEERRIBBEN

LM187 (Pierikken area)

THE DISTRIBUTION OF SEEPAGE INDICATORS IN DE WEERRIBBEN

THE QUAGFENS OF DE STOBBENRIBBEN AND THEIR VEGETATION

PEAT TEMPERATURE AND THE ESTIMATION OF VERTICAL WATER FLOW

Heat capacity and thermal diffusivity of very wet peat soils

LATERAL WATER FLOW IN LONGITUDINAL TRANSECTS

ENVIRONMENTAL AND VEGETATIONAL PROCESSES IN

11. SUMMARY AND GENERAL DISCUSSION

Recurrent patterns of heterogeneity

EXPLANATION OF SOME SPECIFIC TERMS

CLASSIFICATION OF QUAGFEN VEGETATION

INDICATOR LIST OF FEN-MIRE SPECIES

EVALUATION OF THE MAJOR-IONIC COMPOSITION OF

Definition of MAION functions and related procedures

Compensation for temperature differences

EVAPO-TRANSPIRATION FROM LYSIMETERS WITH FEN VEGETATION

Published evidence of seepage in rich-fen quagmires

2.2 Definitions of seepage

statement implies that the presence of a Scorpidio-Caricetum diandrae and a Scorpidio-Utricularietum

Seepage indicators in quagfens in The Netherlands

2.6 Special remarks with regard to bryophytes

Types of seepage sites and their characteristic vegetation

Cited from Ellenberg 1978, p. 421

2.8 Conclusions and re-formulation of the seepage hypothesis

Site properties indicated by the flora of quagfens

Site indication by quagfen flora

3.2 Species and associations of species as indicators

Plot sizes used in phytosociological field work in fen mires

Dierssen (p.10-11) comments that mostly a plot size of 1 m2 is quite satisfactory in

Site indication by quagfen flora

The concept of AND and OR Associations

Site indication by quagfen flora

The OR assumption: an additional explanation of species association

Site indication by quagfen flora

Different scales of environmental homogeneity in a quagfen

Site indication by quagfen flora

Site indication by quagfen flora

Hummock-and-hollow mosaic in quagfen vegetation

Example of micro-zonation in quagfen vegetation

Site indication by quagfen flora

3.4 The compilation of a list of indicator species (Appendix C)

Site indication by quagfen flora

Campylium stellatum, Fissidens adianthoides, Scorpidium cossoni + S.revolvens, Dactylorhiza

They are treated as a rest group (DIV) in Table 3.2.

Site indication by quagfen flora

Frequency distribution of 212 quagfen species according to Ellenberg's indicator