HYDRAULIC PROPERTIES OF SOUTH
FLORIDA WETLAND PEATS
RALEIGH D. MYERS
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ENGINEERING
UNIVERSITY OF FLORIDA
This research was funded by the South Florida Water Management District. I
would like to thank Dr. William Wise, my supervisory committee chairman, for
providing me with the research opportunity and with a pleasant and productive graduate
experience. I would also like to thank my other committee members, Dr. Michael
Annable and Dr. Jennifer Jacobs, for their support and guidance. Research in wetlands is
not performed by one person, and I would like to thank everyone who helped me collect
samples in the field: Dr. Wise, Chris Martinez, Rick Roberts, and Randy Switt. The peat
sampler was designed by Dr. Joe Prenger and Mark Clark and tested by Joe, Mark, and
Jenny Gruentzel. Jenny also lent her skills in plant identification. Without Chris
Martinez and his carpentry and plumbing skills, I would have had a much more difficult
time designing and building the test apparatus. Finally, I would like to thank my family
and all my friends at the University of Florida for providing moral support during my
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ....................................................................................................ii
L IST O F TA B LE S .............. ................................................... ............... v.. ..... .... .v
L IS T O F F IG U R E S .............................................................................................................v i
A B S T R A C T ........................................................................................................ ........ .. ix
1 INTRODUCTION .................................. .. ......... ............ .............. 1
2 LITER A TU R E R EV IEW ... ...................................................................... .............. 3
In tro d u ctio n ......................................................... ................................................ . 3
Physical Properties of Peat.. ................................................................ .............. 4
Botanical Composition..................................................... .............. 4
M in eral C ontent............................................................................................... 5
D egree of H um ifi cation .................................................................... .............. 5
F ib er C o n ten t................................................................................................... 5
B u lk D en sity ........................................................................... .. .............. 6
P o ro sity .......................................................... ................................................ . 7
W after C content ......................................... ............ . ............. ... .......... 8
Interrelationships and Relative Importance of Different Properties .................. 10
Peat Classification and Terminology .............................................................. 10
Hydraulic Conductivity Of Saturated Peat............................................................... 11
Hydraulic Properties of Unsaturated Peat........................................................ 17
Effect of a Changing Water Table on Hydraulic Properties .............................. 17
Significance of Unsaturated Peat within the Hydrologic Cycle......................... 37
Irreversible Changes Due to Prolonged Drying .............................................. 42
Plant Growth and Soil Moisture in Peat................................................................ 46
Methods and Difficulties in Experimental Design................................................ 50
C o n c lu sio n s ............................................................................................................... .. 5 2
3 METHODS, MATERIALS, AND DATA ANALYSIS......................................... 55
S ite D e sc rip tio n ........................................................................................................... 5 5
U ndisturbed P eat Sam pling ......................................... ......................... .............. 55
G rain Size D distribution for Sand.................................... ...................... .............. 58
Peat Physical Properties Analysis .............................. ..................... 58
Tempe Pressure Cell Water Retention Test For Sand........................................... 60
Tw o-Tube Peat Testing A pparatus.......................................................... .............. 63
Saturated H ydraulic C onductivity........................................................... .............. 64
Unsaturated Column Water Balance Test............................................................. 68
Unsaturated Column Physical Analysis................................................................ 71
4 RESULTS AND DISCUSSION .......................................................................... 73
S an d ............................................................................................................... 7 3
G rain Size D distribution ......................................... .......................... ............. 73
Saturated H ydraulic Conductivity..................................................... .............. 74
U nsaturated B ehavior.. ................................................................... .............. 75
P e at ................ ............................................................ . . ............................ . ........... 7 8
Physical Properties ............................................................ .................. .. 78
Saturated H ydraulic Conductivity..................................................... .............. 82
U nsaturated B ehavior.. ................................................................... .............. 88
5 C O N C L U SIO N S ........................................................................................................... 94
A Derivation of the Two-Tube Permeameter Hydraulic Conductivity Expression.......... 98
B S statistical T ests ........................................................................................................... 10 0
C R aw D ata .................................................................................................................... 1 16
REFERENCES ............................................... .......... ........................ .. 141
BIOGRAPHICAL SKETCH................................................................. .............. 146
LIST OF TABLES
2.1. Summary of von Post Method of Determining Degree of Humification (ASTM D
2974-87) ................................................................................ ..........................6
2.2. Horizontal and Vertical Conductivities Measured in the Field and in the Laboratory
b y B o elter (19 6 5). ................................................................................................... 17
2.3. Fitting Parameters for Vorob'ev Soil Moisture Equation (Ivanov, 1981). ................29
2.4. Fitting Param eters for Heliotis (1989) Equation..................................... ................ 40
3.1. Coordinates of U ndisturbed Sam ple Cores............................................. ................ 58
3.2. Saturated Hydraulic Conductivity Tests on Sand .............. .................................... 65
3.3. Saturated Hydraulic Conductivity Tests on Peat..................................... ................ 66
3.4. Example of Calculations Required to Predict Cumulative Discharge Values.............71
4.1. Hydraulic Conductivity Values and Sums of Squared Residuals for all Sand Tests...74
4.2. Mean Hydraulic Conductivity Values for FP5 Sand ..............................................76
4.3. Physical Properties of Fresh Peat Sam ples ............................................. ................ 79
4.4. Hydraulic Conductivity Values for all Saturated Peat Tests and Sums of Squared
Residuals for Fitting of the Mathematical Model. .............................................84
4.5. Mean Hydraulic Conductivities of FP5 and SV5 Peat Samples ................................86
4.6. Relationships Between Hydraulic Conductivity, Time, and Initial Head in the
R eserv oir T u b e ........................................................................................................ 8 7
LIST OF FIGURES
2.1. Effect of Peat Fraction in a Peat-Sand Mixture on the Total Porosity of the Mixture...8
2.2. Phase D iagram s for Five Sedge Peats....................................................... ...............9...
2.3. Summary of Hydraulic Conductivity Values from the Literature.............................. 13
2.4. Decrease in Saturated Hydraulic Conductivity with Increasing Humification on the
v on P o st S cale .............. ................................................ .. ................. .. 14
2.5. Relationship of Hydraulic Conductivity to Fiber Content and Bulk Density ..............15
2.6. Measured Pore Tension as a Function of Water Table Depth.................................19
2.7. D ecrease in Specific Y ield W ith D epth .................................................. ................ 21
2.8. Moisture Characteristic Curves for Five Peat-Sand Mixtures ...............................23
2.9. Moisture Characteristic Curves for Three Peat Types...........................................24
2.10. M oisture Characteristic Curves For Two Peats ............... .............. ..................... 25
2.11. Moisture Characteristic Relationship for a Sedge and Sphagnum Peat with
Interm ediate H um ifi cation ....................................... ....................... ................ 26
2.12. Moisture Characteristic Curves for Several Peat Layers ........................................27
2.13. Moisture Characteristic Curves for a Cultivated Organic Soil in the Everglades........28
2.14. Moisture Characteristic Curve Models Tested by Weiss et al.( 1998)...................31
2.15. Relationship Between Water Content, Tension, Bulk Density, and von Post Degree
of Humification ........................ .... ....... ............... 33
2.16. Moisture Characteristic Relationship for Different Layers in a Finnish Peat..............35
2.17. Time Required to Reach Equilibrium for Different Peat Samples at Different
T e n sio n s .............................................................................................................. . . 3 6
2.18. Relationship Between Soil Moisture Deficit and Water Table Depth for Three
S u m m ers ............................................................................................................... .. 3 8
2.19. Rainfall Required for Flooding Given Water Table Depth.....................................40
2.20. Moisture Characteristic Curves for Air-Dried vs. Undried Samples........................43
2.21. Moisture Characteristic Curves for Similar Drained and Undrained Wetlands at
V ariou s D ep th s ........................................................................................................ 4 4
2.22. Water Table Elevation and Soil Water Content as a Function of Distance from a
D itch ...................................................................................................... . ........ .. 4 5
2.23. Plant-Available Water Superposed on Water Content-Bulk Density Relationship .....47
2.24. Sphagnum Seedling Coverage (a) Three Months and (b) Six Months After Planting
in Peat with Different W ater Table Depths ................ .................................... 51
3.1. Locations of Tw o W wetlands Sam pled ..................................................... ................ 56
3.2. Photographs of FP5 (left) and SV 5 (right).............................................. ................ 57
3.3. Diagram of the Two-Tube Peat Testing Apparatus. ......................... ..................... 63
3.4. Photograph of the Two-Tube Peat Testing Apparatus and Support Structure ..........64
4.1. FP5 Sand G rain Size D distribution .......................................................... ................ 73
4.2. Two-Tube Permeameter Observed and Predicted Values for FP5 Sand ..................75
4.3. Moisture Characteristic Curves for FP5 Sand Derived from Tempe Pressure Cell
Data and Fit with the van Genuchten (1980) Model..........................................77
4.4. Sand Column Water Release Data With Integrated van Genuchten (1980) Fit...........77
4.5. Peat Physical Properties of FP5-3 as Measured after Completion of all Other
T e stin g .................................................................................................................. ... 8 3
4.6. Two Examples of Raw Data from Peat-Sand Hydraulic Conductivity Experiments. .85
4.7. Relationship Between Hydraulic Conductivity and Time for SV5-3 and Downward
F lo w ....................................................................................................... . ....... .. 8 7
4.8. Unsaturated Column Experim ent Raw Data........................................... ................ 89
4.9. Volume Released by Unsaturated Sample FP5-3 as the Water Table Drops..............90
4.10. Volume Released by Unsaturated Sample FP5-4 as the Water Table Drops ............91
4.11. Volume Released by Unsaturated Sample FP5-2 as the Water Table Drops ............92
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Engineering
HYDRAULIC PROPERTIES OF SOUTH
FLORIDA WETLAND PEATS
Raleigh D. Myers
Chairman: William R. Wise
Major Department: Environmental Engineering Sciences
This study examined the hydraulic properties of peat samples taken from isolated
wetlands in southern Florida. The nature of peat soils affects wetland hydrology. Due to
rapid development in the region, water resources managers are challenged to conserve
wetlands while meeting human water needs. An extensive literature review was
conducted to determine the extent of existing knowledge of peat, possible experimental
methods, and procedures for data analysis. The physical properties of the samples, the
transmission of water in the saturated condition, and the retention and release of water in
the unsaturated condition were studied in laboratory experiments.
Both sample collection and laboratory experiments employed unique equipment.
Undisturbed peat samples were collected with a rotating device developed at the
University of Florida. Experiments on saturated and unsaturated hydraulic properties
were carried out using a two-tube apparatus designed by the author. Data from hydraulic
conductivity experiments conducted in this apparatus fit an analytical model very closely
and suggest that experimental error was very small. The apparatus also allowed the
establishment of a water table at any level in order to conduct water balance experiments
on an unsaturated sample.
Because peats differ from mineral soils and from each other, the physical
properties of the samples were characterized. Bulk density and ash content increased
with depth, while effective porosity decreased with depth. Although the samples studied
had ash contents ranging from about 25-90%, they behaved hydrologically as peats.
Saturated hydraulic conductivities for the samples ranged from 7.3x10-6 cm/s to
3.0x10-4 cm/s. These values fall within ranges reported in the literature for pure peats.
There was some evidence that the samples displayed a higher hydraulic conductivity in
the upward flow direction than in the downward flow direction. Such behavior may
enhance the ability of wetlands to maintain standing surface water when the groundwater
table is low.
Under unsaturated conditions, the ability of peat to hold water within its structure
may help support wetland ecosystems. The peat samples released some water as the
water table dropped, even at low tension heads. The height of capillary rise in the sand
was an important factor affecting peat moisture. Although the peat maintained a water
content of 30-40% even with a very deep water table, the residual water is not necessarily
available to plants and would decrease further in the field due to evaporation. Following
a change in water table height, soil moisture in the peat layer reached a new equilibrium
distribution within a week or less. Pumping during drought periods should be managed
so that the capillary fringe does not fall below the peat-sand interface any longer than this
This study of the hydraulic properties of saturated and unsaturated peat was part
of a larger study of the hydrology of small, isolated wetlands in south Florida. Because
land development and population growth are occurring at a rapid pace in this region, a
major challenge facing water resources managers is to achieve a balance between human
water needs and wetland conservation. It is reasonable to assume that depression of the
water table underlying a wetland affects that wetland's hydrology and ecology.
However, managers may be able to avoid long-term effects by wisely choosing the
intensity and frequency of groundwater pumping.
The nature of wetland soil is a major factor determining the hydrologic behavior
of a wetland. Because dead plant matter decomposes slowly in the low-oxygen aquatic
environment, many wetland soils contain a high fraction of organic matter. Peats, or soils
consisting of mostly or all organic matter, help create hydrologic conditions that support
wetland ecosystems. Under both saturated and unsaturated conditions, peat serves a
hydrologic buffering function, helping to support wetland vegetation as the water table
drops. Under saturated conditions, organic soils help to maintain standing surface water
following a rain event. There is some evidence that they may also allow greater
transmission in the upward flow direction than in the downward flow direction, further
increasing the amount of surface water. During dry periods long enough to exhaust
surface water, organic soils retain a high degree of soil moisture as the water table drops.
This soil moisture may increase the period of time wetland plant communities are able to
survive dry conditions.
The hydrologic buffering effect has implications for decisions regarding
groundwater pumping. When surface water exists, the water table can be lowered for a
period of time, depending on surface water depth and water table depth, without causing
all the surface water to drain or evaporate. When surface water does not exist, a
maximum long-term water table drawdown can be specified that will not cause major
changes in wetland hydrology. This drawdown may be exceeded for periods of time
short enough to prevent soil moisture from reaching a new equilibrium.
The study met two main objectives. First, it examined the ability of peat samples
to transmit water in the saturated condition. The hydraulic conductivity of peat, while not
a perfect measure of its water transmission ability, allows prediction of the change in the
level of surface water depending on the water table depth. Comparison of the hydraulic
conductivity during upward and downward flow provides further information on the
ability of wetlands to retain surface water when the water table drops. Second, the study
examined the water retention and release of unsaturated peat as the elevation of the water
table beneath it changed. The importance of the capillary fringe and the time required for
moisture in the peat to reach an equilibrium condition were studied.
This study was limited in scope in that it examined the problem in one dimension,
in the absence of evapotranspiration, and for two specific wetland systems in southern
Florida. The results provide one piece of a total understanding of the hydrology of peat-
Organic wetland soils, often referred to as peats, play an important role in
determining the impact of groundwater pumping on isolated wetland ecosystems. When
the soil is saturated, the hydraulic conductivity determines how much surface water is
present and how long it can persist. When the soil is unsaturated, the capillary properties
of the soil pores determine the ability of the soil to maintain sufficient moisture in order
to support plant life. The hydraulic properties of a soil are determined by that soil's
The results of research concerning the hydraulic properties of peats can be applied
to determine how to responsibly extract groundwater for human use. By maintaining
some minimum water table depth, water management authorities can ensure that soil
moisture is adequate over the long term to maintain isolated wetlands in their historical
The existing body of knowledge concerning peat is not nearly so large as that
concerning mineral soils. The physical properties of peats have been fairly well
documented because they are of interest to a range of scientists and industries. A smaller
amount of work has been performed on the hydraulic properties of peat. Of the studies
available, most deal with hydraulic conductivity and flow in the saturated zone. The
number of studies concerning the unsaturated zone in peat is small. This literature review
briefly surveys concepts and studies related to the physical properties of peat and to the
hydraulic properties of saturated peat. It contains a detailed discussion of the hydraulic
properties of unsaturated peat, citing a large fraction of the studies available on the
subject. Some of the studies discussed are several decades old but are included because
the existing body of knowledge is so small. The basic concepts and ideas related to
processes in the unsaturated zone of soils have not changed drastically in this time.
Most of the studies cited in this literature review were performed on temperate
peatlands in the northern United States, Canada, Europe, and Russia. While peats in
northern temperate climates are more extensive and more widely studied, peats also occur
in subtropical and tropical climates (Clymo, 1983). Peat forms when the rate of organic
matter deposition exceeds the rate of decomposition, a condition often met in topographic
lows with standing water. In temperate climates, precipitation usually exceeds
evapotranspiration (Verry and Boelter, 1978) and microorganisms are less active than in
tropical climates, leading to deeper and more extensive deposits. Tropical and
subtropical peat deposits are typically shallow compared to their northern counterparts
Physical Properties of Peat
The botanical composition of peat plays a role in determining the structure of the
peat as it decomposes. Therefore, gathering some information on the approximate
composition of the peat is important to the hydrologist. Peats are often divided into those
derived from mosses, sedges and grasses, and woody materials (Clymo, 1983). Provided
the ecosystem has not changed dramatically, the living plants in a system can provide
clues about the botanical composition of the peat.
The mineral content of peat is the percentage of inorganic matter present on a
weight basis. The mineral content is sometimes called the ash content because it can be
estimated by burning off the organic matter at a high temperature. Because most
inorganic matter is much more dense than plant remains, the weight percentage may tend
to overstate its influence on hydraulic properties in the soil matrix. In order for a material
to be considered a peat, it must contain no more than an arbitrarily determined maximum
inorganic content. This percentage varies somewhat depending on the reason for
studying the peat. 20% is a typical value, although some soil scientists allow up to 35%.
A peat containing up to 55% inorganic matter may be viable for commercial purposes
Degree of Humification
In studies of northern temperate peatlands, the degree of humification is often
estimated using a somewhat subjective system developed by von Post and Granlund
(1926). In this system, a sample of peat is squeezed in the hand, allowing liquids and soft
solids to ooze out. The appearance of the material squeezed out and the material retained
determines the degree of humification on a ten-point scale (Landva and Pheeney, 1980).
Table 2.1 summarizes the method. Although the method may seem extremely subjective,
it has a scientific basis. The maximum pressure applied in squeezing the human hand is
sufficient to expel free and most capillary water but not chemically-bound water.
The fiber content of peat provides an idea of the size of peat particles. The term
fiber is defined as "a fragment or piece of plant tissue, excluding live roots, that is large
enough to be retained on a 100-mesh sieve (openings 0.15 mm in diameter) and that
Table 2.1. Summary of von Post Method of Determining Degree of Humification
(ASTM D 2974-87).
H Nature of Material Extruded on Squeezing Nature of Plant Structure in Residue
1 Clear, colorless water; no organic solids Unaltered, fibrous, undecomposed
2 Yellowish water; no organic solids Almost unaltered, fibrous
3 Brown, turbid water; no organic solids Visibly altered but identifiable
4 Dark brown, turbid water; no organic solids Easily identifiable
5 Turbid water and some organic solids Recognizable but vague, difficult to identify
6 Turbid water; 1/3 of sample squeezed out Indistinct, pasty
7 Very turbid water; 1/2 of sample squeezed Faintly recognizable; few remains identifiable,
out mostly amorphous
8 Thick and pasty; 2/3 of sample squeezed Very indistinct
9 No free water; nearly all of sample No identifiable remains
10 No free water; all of sample squeezed out Completely amorphous
retains a recognizable cellular structure of the plant from which it came" (USDA Soil
Survey Staff, 1975, p. 66). Some researchers define fibers as greater than 0.10 mm in
length (e.g., Boelter, 1969). A crude grain size analysis provides an estimate of the fiber
content. The researcher soaks the peat in a dispersing agent overnight, stirs it vigorously,
and washes it with a gentle stream of water through a sieve or set of sieves. The fiber
content is the amount retained on the 100-mesh and higher sieve as a percentage of total
weight (ASTM D 1997-91).
Because they contain such a large proportion of water, peats have low bulk
densities. Typical values are in the range of 0.10 g/cm3 for pure, undecomposed peats to
0.20 g/cm3 for pure, well decomposed peats (Clymo, 1983). In this paper, the term bulk
density will refer to the dry bulk density as defined below:
Pb =- (2.1)
where pb = dry bulk density [M/L3],
Ms = mass of dry solids [M], and
VT = total volume of moist soil [L3].
Porosity in peat is high compared to that of mineral soils. Figure 2.1 shows that
as the percentage of peat in a particular horticultural peat-sand mixture increases from
zero to 100%, the porosity increases from 44% to 95%. In pure peats, porosity ranges
from close to 100% for freshly deposited plant remains to about 80% for very well-
decomposed matter (Boelter, 1969). Porosity can be described either as total porosity or
as effective porosity. Effective porosity is the more useful parameter, particularly for
swelling soils such as clays and peats. The phase diagrams in Figure 2.2 provide a visual
indication of how significant void spaces are in the structure of peat. In each case, solids
make up less than 15% of the total volume. Because peat consists of the remains of plant
materials, the shapes of peat particles and of the void spaces between particles are highly
irregular. Thus, traditional ideas about porosity and soil moisture must be applied with
caution because they are often based on simplifying assumptions about grain shapes.
E 60 -
0 25 50 75 100
Figure 2.1. Effect of Peat Fraction in a Peat-Sand Mixture on the Total Porosity of the
Mixture (based on Boggie, 1970).
The water content of mineral soils can be described on a volumetric or a
gravimetric basis as follows:
S= xV 100% (2.3)
where w = gravimetric water content (fraction),
Mw= mass of water [M],
Ms = mass of dry solids [M],
0 = volumetric water content (%),
Vw = volume of water [L3], and
VT = total volume [L3].
The gravimetric water content is most appropriate for geotechnical applications
because the mass of solids remains constant before and after a load is applied. The
volumetric water content is favored in hydrology because the total volume usually
1 SPHAGNUM SOLIDS SV SPHAGNUM INTRAPARTICLE VOIDS
ZFZFFZ SEDGE SOLIDS CV : SEDGE INTRAPARTICLE VOIDS
NUMBERS ALONG DIAGRAMS V: INTERPARTICLE VOIDS
ARE % BY VOLUME a GAS/AIR
6 G 6 G 6 G 7 G 6 G
29 V V
37 V 37 V 36 V
34 CV CV 37 CV 45 CV
14 II SV 12 SV II SV 12 SV
SI I MMMM I mmm I mmmml
n 0.86 0.88 0.89 0.90 0.92
w % 381 454 502 558 713
I Ir I IS E
Figure 2.2. Phase Diagrams for Five Sedge Peats (Landva and Pheeney, 1980). n is
porosity and w is the gravimetric water content expressed as a percent.
remains constant before and after changes in water content. However, neither method of
describing water content is ideal for peat. Gravimetric water contents for peat are often
as high as 3000-5000% and are therefore somewhat meaningless (Boelter and Blake,
1964). Because peat is mostly water, the gravimetric water content is highly sensitive to
small changes in the density of solids. The volumetric water content, on the other hand,
is insensitive to density changes and produces values within a narrow range of 80% to
100%. Nevertheless, Boelter and Blake (1964) strongly recommend the use of
volumetric water content to describe moisture in peat. This paper will not distinguish
between the terms water content and moisture content, and the term water content will
refer to the volumetric water content unless otherwise specified.
Interrelationships and Relative Importance of Different Properties
The von Post degree of humification, fiber content, porosity, and bulk density of a
sample all provide information on the extent of decomposition. Although several
researchers (e.g., Boelter, 1969) have studied the interrelationships of these properties, a
detailed discussion of these interrelationships is beyond the scope of this paper. Malterer
et al. (1992) compared a large number of methods of gauging the degree of
decomposition, including fiber content and the von Post scale. They concluded that the
von Post method is consistently more accurate than the others. They recommend its use
because it is both accurate and simple.
Peat Classification and Terminology
Researchers and professionals in various fields have proposed classification
schemes for peat. A brief survey of the most common terminology is worthwhile in order
to avoid confusion when relating physical and hydraulic properties. Geotechnical
engineers usually classify peat based on the relative contents of organic and inorganic
matter. Soil scientists, who serve mainly the agricultural and forestry industries, have a
more detailed classification system. Hydric organic soils, formally called Histosols, are
divided into three categories. Fibrists consist of two-thirds or more fibers and correspond
to about HI to H5 on the von Post humification scale (Table 2.1). Hemists consist of
more than one-third but less than two-thirds fibers and correspond to about H6 to H8.
Saprists consist of less than one-third fibers and correspond to about H9 to H10 (Clymo,
Soil scientists use the word peat to describe fibrists and the term muck to describe
saprists. Hemists are either mucky peat or peaty muck. However, many authors refer to
organic soils across the full range of decomposition as peat, and this convention will be
followed throughout this paper. The term peat will refer to any wetland soil with an
organic content great enough to control its hydraulic properties. The degree of
decomposition will be referred to using terms such as undecomposed, slightly
decomposed, moderately decomposed, and highly decomposed. The terms
decomposition and humification will be used interchangeably unless referring
specifically to the von Post humification scale.
Hydraulic Conductivity Of Saturated Peat
A number of studies exist concerning the applicability of Darcy's Law and the
hydraulic conductivity concept to saturated organic soils. In any soil, the flow rate of
water increases along with the magnitude of the hydraulic gradient applied to it. In
mineral soils, this relationship has been shown to be essentially linear, an assumption of
Darcy's Law. While the relationship may not be linear for organic soils, it may be
approximately linear within a certain range of gradients or for a particular peat type.
A number of studies suggest that Darcy's Law is applicable only to the upper
layer and only to slightly decomposed peat (e.g., Hemond and Goldman, 1985; Ingram et
al., 1974; Rycroft et al., 1975; Romanov, 1968). They identify two possible causes of a
departure from Darcy's Law in the deeper layers. First, although the structure of the
medium is constant, the flow rate may vary nonlinearly with hydraulic gradient or with
the absolute magnitude of head applied. Second, the structure itself may vary with
hydraulic gradient or with absolute head, leading to a nonlinear variation in hydraulic
properties. Hemond and Goldman (1985) recommend that Darcy's Law be applied only
in cases of small hydraulic gradients and fairly constant effective stress. They suggest
that the Richards equation, a generalized form of Darcy's Law in which hydraulic
conductivity varies as a function of hydraulic gradient, may be applicable to saturated
peats with a high degree of humification.
Although the applicability of the hydraulic conductivity concept to peats
continues to generate controversy, many researchers have attempted to measure the
saturated hydraulic conductivity of peats using traditional methods. Figure 2.3 compares
the results of a large number of tests by different groups. Values range from 10-1 to 10-7
cm/s, with most of the values falling between 10-3 and 10-5 cm/s. Because hydraulic
conductivity depends on the pore size distribution of peat, it is related to the degree of
decomposition. Figure 2.4 shows the relationship of hydraulic conductivity to the von
Post humification scale as measured by six different authors. The relationship shows a
decrease in conductivity with increasing humification. Moss peats have the lowest
hydraulic conductivity at all humifications, while sedge and reed peats have the greatest.
The conductivities of the different peat types converge as they reach a high degree of
humification. Boelter (1969) performed linear regressions to relate hydraulic
conductivity to fiber content and bulk density in a Minnesota bog peat. Figure 2.5 shows
that hydraulic conductivity increases approximately logarithmically (r2=0.54) with
increasing fiber content and with decreasing bulk density. Hydraulic conductivity ranges
from about 10-2 cm/s for undecomposed peat to less than 10-5 cm/s for very well
Some evidence exists that vertical hydraulic conductivity may vary depending on
the direction of flow. Such behavior has implications for the hydrology of isolated
3 4 1 R 7 a 10
SREF 3 4 EREN6 7
S 9 I10
Figure 2.3. Summary of Hydraulic Conductivity Values from the Literature (Chason and
Siegel, 1986). (a) Field values: 1, Baden and Egglesman (1961, 1963, 1964); 2,
Egglesman and Makela (1964); 3, Boelter (1965); 4, Ingram (1967); 5, Galvin and
Hanrahan (1968); 6, Romanov (1968); 7, Sturges (1968); 8, Dowling (1969); 9, Irwin
(1970); 10, Yamamoto (1970); 11, Knight et al. (1971); 12, Dai and Sparling (1972); 13,
Ingram et al. (1974); 14, Paivenen (1973); 15, Galvin (1976); 16, Dasberg and Neuman
(1977); 17, Chason and Siegel (1986); (b) Laboratory values: 1, Malstrom (1925); 2,
Sarasto (1961); 3, Boelter (1965); 4, Bazin (1966); 5, Irwin (1970); 6, Korpijaako and
Radforth (1972); 7, Bartels and Kunze (1973); 8, Galvin (1976); 9, Dasberg and Neuman
(1977); 10, O'brien (1977); 11, Chason and Siegel (1986).
;l i t i 1 1 i in .. I
S5 0-3- ll Sphagnum peats
E Brown Moss and
Brown Moss-Sedge peats
Reed and Sedge peats
0 I 2 3 4 5 6 7 8 9 10
Figure 2.4. Decrease in Saturated Hydraulic Conductivity with Increasing Humification
on the von Post Scale (Rycroft et al., 1975).
wetlands; if upward flow exceeds downward flow, the ability of a wetland to hold water
during dry periods is enhanced. Marshall (1968) showed that drag forces significantly
lowered the hydraulic conductivity of a sandy loam during downward flow. He attributed
this behavior to a combination of two factors. First, individual particles can be moved or
reoriented by friction forces. Downward flow may have a tendency to transport fine
particles into larger pores, blocking the pores and increasing resistance. Second,
compression of the soil matrix may reduce the total volume and porosity of the sample.
BULK DENSITY Ig/cm3)
10.04 .195 .075
: 1 ~Log Y = -6.539 + .0566X1
r2 = .54 I
S10Log Y = -1.589 16.068X2
S .r 54
SAPRIC HEMIC FABRIC
0 33 67 100
FIBER CONTENT1>0 1mmi. PERCENT OVENDRY WEIGHT
Figure 2.5. Relationship of Hydraulic Conductivity to Fiber Content and Bulk Density
Rycroft et al. (1975) present evidence that these effects are significant in peat.
The evidence suggests that these processes are largely reversible, indicating that
reversible compression effects are more important than presumably irreversible pore-
Horizontal hydraulic conductivity may in some cases be different from vertical
hydraulic conductivity. One possible explanation is that the orientation of live or dead
plant parts allows greater flow in one direction. Boelter (1965) expected horizontal
conductivity to be greater than vertical conductivity in his measurements. However, his
results did not demonstrate a statistically significant difference. Weaver and Speir (1960)
found the vertical conductivity of undisturbed layers of Everglades peat to be about three
times greater than the horizontal conductivity.
Rycroft et al. (1975) and Paivenen (1973) describe two methods of measuring
saturated hydraulic conductivity in peat, the auger hole method and the seepage tube
method. In both cases, researchers bore a hole, apply a hydraulic gradient, and observe
the resulting flow of water. The hole is unlined in the auger hole method, while the hole
is lined by a tightly fitting tube in the seepage tube method. An unlined cavity is bored
below the seepage tube. Both the unlined auger hole and the unlined cavity below the
seepage tube are intended to allow radial flow from all directions, providing an estimate
of horizontal conductivity in the peat. Boelter (1965) employed a method similar to the
seepage tube method; however, he did not bore a cavity below the tube and considered
the resulting value to represent vertical hydraulic conductivity. Boelter also describes a
piezometer method, in which a tube with a small diameter (approximately 3.2 cm) is
installed in a manner similar to that of the seepage tube method.
Hydraulic conductivity values measured in the field and in the laboratory may
differ significantly. Rycroft et al. (1975) consider the advantages of field methods as
compared to laboratory methods. In the laboratory, the investigator has control over
variables such as environmental conditions, fluid properties, rectilinear flow, hydraulic
gradient, and absolute pressures. In the field, the investigator has the ability to minimize
disturbance to the sample and to experiment on a much larger sample. Thus, the effects
of system edges and heterogeneous elements are averaged over the larger volume.
Table 2.2 compares values obtained in the laboratory with those obtained in the
field by Boelter (1965). Horizontal conductivities measured in the lab are roughly ten
times greater than those measured in the field. This difference is confirmed statistically.
Boelter attributes this difference to leakage around the edges of sample cores in the
laboratory. Vertical conductivities measured in the lab and in the field were of the same
order of magnitude for less decomposed samples, but the disparity increased sharply for
the well decomposed samples. Paivenen (1973) reports a similar trend, attributing it to
sample disturbance during collection and transport. Romanov (1968) recommends that
samples used for determination of hydraulic conductivity in the laboratory be as large as
possible in order to avoid distortions caused by edge effects.
Table 2.2. Horizontal and Vertical Conductivities Measured in the Field and in the
Laboratory by Boelter (1965).
Peat Type Field Laboratory
Horizontal Vertical Horizontal Vertical
undecomposed moss peat 3.81x10-2 6.20x10-2 1.50x10-1 9.59x10-2
partially decomposed moss peat 1.39x104 5.08x10-4 1.32x10 5.65x10-
decomposed moss peat 1.LlxlO-5 8.50x10-6 1.47x10-4 3.86x10-'
sedge and reed peat 7.50x10 7.50x10-6 3.11x10 2.98x10-
Hydraulic Properties of Unsaturated Peat
Effect of a Changing Water Table on Hydraulic Properties
Types of soil moisture in peat
The water present in unsaturated peat can be separated into two categories: free
water and bound water. Free water is defined as that water which can be displaced
through gravity. In the unsaturated zone, free water includes water that infiltrates after a
rain event and percolates through larger pores down to the water table. In contrast, bound
water does not seep out when a relatively small pressure gradient is applied (Ivanov,
1981). Some of the bound water is held by osmotic forces within intact plant cells
(Romanov, 1968). A small portion is strongly bound to soil particles by chemical and
electrostatic forces, constituting a weight equal to 50%-100% of the solid matrix. A
larger portion of bound water is held in sealed pores or held in capillary pores by tension
forces. This capillary water determines the water yield and the water-retention capacity
of the peat (Ivanov, 1981).
Chemical and electrostatic forces may play a significant role in the amount of
water retained in peat. In a manner similar to that displayed by clays, organic matter in
soils is known to interact with water on a microscopic level through electrostatic forces.
Like clay particles, the particles of organic matter are usually negatively charged,
attracting polar water molecules to their surfaces. The layer of water molecules adjacent
to a soil particle is tightly bound, and other nearby water molecules are attracted to this
first layer by hydrogen bonding. Cations also tend to adsorb to the negatively charged
particles and may themselves immobilize water molecules through hydration. Some
researchers believe that tightly bound water molecules develop a crystal structure and that
their density and viscosity differ significantly from liquid water (Hillel, 1998). Paivenen
(1973) measured pore tension of a Finnish peat as a function of distance to the
groundwater and found it to be nonlinear at groundwater depths greater than about 60 cm.
The dramatic pore tension increase at greater depths is shown in Figure 2.6. At high
tensions, such effects clearly cannot be ignored without significant error.
Capillary rise, specific yield, and equilibrium moisture characteristic curves
Capillary rise, specific yield, and equilibrium moisture characteristic curves all
provide information about the hydraulic behavior of unsaturated peat. Several
researchers have measured the capillary rise for peats in the upper layer of Russian fen
30 40 50 60 70 80
Distance to ground water table, cm
Figure 2.6. Measured Pore Tension as a Function of Water Table Depth (Paivenen,
deposits and found typical values in the range of 50 to 110 cm, with maximum values of
about 175 cm (Ivanov, 1981). The capillary fringe can have a significant effect on the
ability of a wetland to store water. It is not uncommon for the capillary fringe to extend
up to the ground surface in a wetland system. When estimating the total water storage
capacity of the soil, failure to take this initial soil moisture into account can lead to errors
in estimating the additional water storage capacity (Heliotis and Dewitt, 1987). A
number of researchers have expressed doubts about the position of the water table alone
as an indicator of the water storage capacity in a wetland (e.g., Boelter, 1964b; Heliotis
and Dewitt, 1987; Munro, 1984).
The specific yield of a soil provides information on the release of soil water upon
lowering of the water table. Larger pore spaces correspond to a smaller capillary rise and
a higher specific yield. The magnitude of the pore spaces corresponds primarily to the
degree of humification or to the fiber content of the peat. Fibrous peat, containing greater
than two-thirds fibers, may release more than 45% of its stored water immediately upon
water table lowering (Verry and Boelter, 1978). Heikurainen et al. (1964) reported that the
water content of a sedge-moss peat of intermediate humification decreased by about 5%
when the water table was lowered by 10 cm. Boelter (1965) found specific yields of
0.10-0.15 for highly humified peat; 0.19-0.33 for woody peats and deep, undecomposed
moss peats; and 0.52-0.79 for undecomposed moss peats near the surface.
Four major factors affect the water yield of a peat soil (Ivanov, 1981):
1. The initial moisture content of the peat
2. The distance the water table declines
3. The distribution of active porosity above the water table
4. The compressibility of the solid matrix
While specific yield is often assumed to have a constant value in mineral soils, it
may vary significantly in peat along with variations in physical properties. Within a
particular peat type, the specific yield may vary as a function of depth. Heliotis (1989)
found that the specific yield of peat in a northern Michigan cedar swamp decreased
approximately exponentially with depth (Figure 2.7), from 0.4-0.7 near the surface to
0.05-0.15 at a 0.5 m depth.
20 400 BOO.
INITIAL WATER-TABLE DEPTH- DI1
2 '. *
200 400 00 800
INITIAL WATE.-TABLE DEPTH-DI
200 40 60000O
INITIAL WATER-TABLE DEPTH-DI
(mam INITIAL WATER-TABLE DEPTH-DI
Figure 2.7. Decrease in Specific Yield With Depth (Heliotis, 1989).
Lundin (as reported by Ivanov, 1981) provides an equation for specific yield of
the upper layer as a function of the maximum height of capillary rise.
Sy =0.20 1 0.T7 .hma (2.4)
where Sy = specific yield (fraction),
he max = maximum height of capillary rise [L], and
z = depth of water table below the surface [L].
When the water table depth is less than the maximum capillary rise, the value of Sy varies
from 0.06 to 0.10, depending on the peat type and its physical properties. The maximum
possible mathematical value of Sy calculated by this method is 0.20. Typical
experimental values for a deeper water table vary from 0.13 to 0.26 (Ivanov, 1981).
In the extensive temperate peatlands of the northern U.S., Canada, and Europe,
the upper peat layer is sometimes removed for use in horticulture or for use as a fossil
fuel. Measurements of specific yield in similar disturbed and undisturbed systems allow
insights both into the differences between layers and into irreversible changes induced by
drying and shrinking. Without long-term measurements, however, it may be difficult to
distinguish between these two separate processes. Price (1996) compared a specific yield
of 0.35-0.55 in an intact wetland to a specific yield of 0.04-0.06 in a similar system in
which the upper layer had been harvested. Such a drastic difference indicates a
significant difference in pore sizes. Price concluded that while the water balance of a
disturbed site can be restored effectively to its original condition, the changes to the soil's
physical structure cannot easily be reversed.
While the specific yield provides information on the total volume of water
released upon drainage, the distribution of soil moisture within the soil profile is also of
interest. According to Ivanov (1981, p. 68), "the equilibrium distribution of capillary
and immobilized water above the water table is determined by the distribution of active
porosity through...the zone of aeration and the size of the pores..., and is therefore a
function of the position of the water table."
Moisture characteristic curves have been measured for a variety of different peat
types. In most cases, experiments used pressure cell and pressure plate apparatus.
Virtually all studies done to date have considered only the desorption or drainage curve.
Figure 2.8 shows the moisture characteristic curves for five different mixtures of a sand
and a horticultural moss peat. Because the water contents were measured on a weight
basis, it is difficult to make direct comparisons between the different mixtures. However,
the figure shows that the shape of the curve changes with increasing peat content. The
curves for pure sand, 25% peat, and 50% peat display the shape typical of sandy soils,
with a sudden drop in equilibrium water content between about 10 cm and 500 cm of
tension. The samples containing 75% and 100% peat show a more gradual decrease in
equilibrium water content. Thus, a soil containing a large proportion of this peat will
retain a greater fraction of its water as the water table beneath it drops than a soil
containing mostly sand.
100,000 [-0-- 100% Sand
10,000 -25% Peat/
1 ,000 ----50% Peat/
o 75% Peat/
0 -- 100% Peat
1.0 10.0 100.0 1000.0
gravimetric water content (%)
Figure 2.8. Moisture Characteristic Curves for Five Peat-Sand Mixtures (Boggie, 1970).
In one of the most extensive studies performed on pure peat to date, Paivenen
(1973) produced moisture characteristic curves for a variety of peat types. Three of his
curves are shown in Figure 2.9. In his words, "of the peats studied, Sphagnum peat
contains the greatest quantity of water at saturation, but it gives up its water more readily
with increasing matric suction. In the case of peats which have reached a more advanced
stage of decomposition, the water contents at saturation were lower, but the loss of water
with increasing matric suction was also smaller" (Paivenen, 1973, p. 42).
0 20 40 60 80 100
Water Content (%)
--Sphagnum -*--Sedge -A-Woody
Figure 2.9. Moisture Characteristic Curves for Three Peat Types (Paivenen, 1973).
Clymo (1983) presents two moisture characteristic curves measured by other
authors, reproduced in Figure 2.10. In the unhumified to slightly humified peat, most of
the water drains between potentials corresponding to circular pores of 1 mm to 20 |tm
diameter. This relationship appears to be nearly linear in this suction range. The more
humified peat, in contrast, retains a water content near saturation up to about a 20 jtm
diameter. Similar to a sandy soil, it exhibits a break point at which a sudden drop in
water content occurs with increasing suction.
Heikurainen et al. (1964) examines the statistical correlation between the depth of
the water table and water content at the surface. His results are nearly linear, with r2
values of 0.842 and 0.943, as shown in Figure 2.11. The peat tested was a sedge and
i- \-x_ p r nm pore
-104 ---20pm pore
Sphagnum peat with an intermediate degree ofhumification (H4-5). Both field and
f- 0 5 10 15
SWater content per unit dry matter, 4,d (g g-1)
Figure 2. 10. Moisture Characteristic Curves For Two Peats (Clymo, 1983).
Sphagnum peat with an intermediate degree of humification (H4-5). Both field and
laboratory experiments were carried out with similar results. It is difficult to determine
the range of water contents over which this study was conducted because Heikurainen
does not report the bulk density of the peat. Thus, the apparently linear fit may pertain
over a fairly small range. However, Heikurainen's results suggest that a linear
approximation may be reasonable for systems with shallow water tables.
Before measuring moisture characteristic curves for a Sphagnum moss peat in
northern Minnesota, Boelter (1964b) carefully documented the physical properties of the
peat. The peat was nearly undecomposed for the first 30 cm depth, then increased in
decomposition to a depth of about 60 cm. The deepest layer showed a low degree of
decomposition and consisted of reed and sedge materials. The surface undecomposed
peat had a bulk density of 0.20 g/cm3 and a saturated water content of 95-100% while the
deeper decomposed peats had a bulk density of about 0.24 g/cm3 and a saturated water
content of 80-90%.
A 10.6 -
8.0 ?s62, $ = 0.6 0,051 X r = --0.842 X
7.8 1-963, y 10.30 0. 053 X, r = 0.943
B 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48
Ground wa/er //bl/e from surface, cm
Figure 2.11. Moisture Characteristic Relationship for a Sedge and Sphagnum Peat with
Intermediate Humification (Heikurainen et al., 1964).
The results of Boelter's soil moisture experiments are shown in Figure 2.12.
Water retention varies both with the degree of decomposition and between different peat
types. Among the moss peats, the lowest water contents are found in the least
decomposed moss layer, while the highest are found in the highly decomposed layer.
The botanical components are not easily identifiable in this layer. At pore tensions
greater than about 200 cm, water contents in the relatively undecomposed sedge peat fall
in between the curves for moss peats of different degrees of decomposition, indicating
that the effect due to humification was greater than the effect due to botanical
composition in this tension range. However, differences due to botanical composition
appear to be more pronounced at lower suctions. In undecomposed moss peat, a water
table deeper than about 20 cm might not supply adequate soil moisture to plants near the
surface. Where layered conditions exist, the position of the water table alone may not be
an effective indicator of soil water retention behavior.
10000 -9 1 G --
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Volumetric Water Content
-*--LUMP 0-15 cm -*--UMP 15-30 cm -*-PDMP 30-45 cm
DP 45-61 cm ----SRP 70-80 cm ---SRP 90-100 cm
Figure 2.12. Moisture Characteristic Curves for Several Peat Layers (Boelter, 1964b).
LUMP = live undecomposed moss peat; UMP = undecomposed moss peat; PDMP =
partially decomposed moss peat; DP = decomposed peat; SRP = sedge and reed peat
The few studies of the water retention behavior of Florida organic soils have been
carried out on cultivated organic soils in the Everglades region. Although these soils are
disturbed and may have compositions different from those found in undisturbed systems,
they can provide insights into the hydraulics of organic soils of this region. Weaver and
Speir (1960) performed soil column and pressure plate desorption experiments on an
organic Everglades soil which had been under cultivation for more than twenty years.
The samples were taken at various depths and tested individually. The results are
presented in Figure 2.13. The deeper, presumably undisturbed layers have a higher
porosity than the shallow, cultivated layers. However, the shallow layers retain more
water at the greatest pore tensions. The break point of the curves seems to occur at only
about 10 cm depth. All of the layers lose water most rapidly as the suction increases
from about 10 cm to about 100 cm. However, the water content at 100 cm is still greater
than 50% for all layers. The bulk density of this soil decreases with depth (Weaver and
Speir, 1960), contrary to what might be expected from the results of the water retention
experiments. While these results are useful for study, they should be applied to
undisturbed systems only with great caution.
is" .. I I:!'
Everglades (Weaver and Speir, 1960)
mt m6---- m ,- iN ,--------------------" - --- -- r- "----'.
Everglades (Weaver and Speir, 1960).
Mathematical models of soil moisture in peat
A number of mathematical models have been suggested for the distribution of soil
moisture in unsaturated peat. Vorob'ev (as reported by Ivanov, 1981) presents an
empirical function for the equilibrium water content as a function of height above the
S= pexp(n-k log h) (2.5)
where 0 = volumetric moisture content (%),
p = bulk density (g/cm3),
n, k = coefficients representing humification and botanical composition, and
h = height above the water table (cm).
Examples of the fitting coefficients n and k are given in Table 2.3. These values
are applicable to the upper layer.
Table 2.3. Fitting Parameters for Vorob'ev Soil Moisture Equation (Ivanov, 1981).
Botanical Composition n k
Sphagnum fuscum and angustifolium 8.33 0.87
S. cuspidatum and S. dusenii 8.96 1.165
A hyperbolic relationship for peat soil moisture as a function of water table depth
was suggested by Romanov (1968):
0 =a (2.6)
where 0 = volumetric water content (%),
a, m = curve-fitting parameters, and
h = height above the water table (cm).
Romanov gives a wide variety of different a and m values determined empirically
for different samples. This curve is valid for a tension range on the order of 10 cm to 100
cm, depending on the sample. It does not take into account the two extreme regions of
the moisture characteristic curve. Because the fitting parameters take on such a wide
range of values, such a simple model is probably most useful for descriptive rather than
predictive purposes. In using such a model, it is important to keep in mind the range of
water contents and tension values for which the equation is valid.
Weiss et al. (1998) attempted to fit two soil moisture models commonly applied to
mineral soils and one model of their own design to Finnish peats. The models were as
van Genuchten (1980) model:
0= 0 +(,- r)[1 +( (,h)n m (2.7)
Zhang and van Genuchten (1994) model:
r (0s r) 1 + c, (a 2h)
1 + ( +(-h) + C2)(h) (2.8)
Weiss et al. (1998) semiempirical model:
O=exp ln(Os)-[1n(Os)-In(Owt)] log1(h) k (2.9)
where 0 = volumetric water content (%),
Or = residual water content (%),
Owilt = water content at the wilting point (%),
h = water table depth (cm),
0s = saturation water content (%),
u1, n, m = van Genuchten shape parameters,
U-2, ci, c2 = Zhang and van Genuchten shape parameters, and
k = Weiss et al. shape parameter.
Shape parameters were determined by a nonlinear regression technique. The
results of the fits are shown in Figure 2.14 for different layers and for different types of
peat. Based on their results, Weiss et al. (1998) recommend the van Genuchten model for
situations in which the behavior at relatively low suctions is of interest. They
recommend their own semi-empirical model for situations in which extensive statistical
manipulation is desirable. The greatest advantage of this model is that it has only one
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4
0 1 1 1 1 1 1 1 O
Sphagnum, LS Sphagnum, ErS -, Carex, LC lignin, SL
80 Layer = 1 Layer = 1 Layer = 1 Layer = 1 8
p = 0.058 p=0110 p = 0.099 p= 0.087
S=65 S=51 S=21 \ S=28
60 \ C=0 C=0 C=55 C=0 60
L=20 L=22 L L=19 L= 57
Er =25 Er=20 Er= 1 Er=3
A C E G .
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4
100 I + 1 1 i | | 100
Sphagnum, S Sphagnum, ErS Carex, LC lignin, ErL
80 Layer = 2 Layer = 4 Layer = 4 Layer = 4 80
p p=0057 p = 0.150 p = 0.099 p = 0.015
S=87 S=33 S=0 S=8
C=0 C=14 C=75 C=0 60
L=8 L= 22 L=15 L= 47
40- Er=6 Er=30 Er 2 Er=23 40
B D F H
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4
Logo(h) Loglo(h) Logo1(h) Loglo(h)
Figure 2.14. Moisture Characteristic Curve Models Tested by Weiss et al. (1998). Solid
lines represent the van Genuchten model, while dotted lines represent the semi-empirical
model developed by Weiss et al. p = bulk density (g/cm3); S, C, L, and Er represent
botanical components by weight (S = Sphagnum, C = Carex, L = lignin, Er =
Relationship of physical and hydraulic properties
The physical properties of a particular peat determine its hydraulic properties.
With knowledge of the physical properties of a sample, prediction of its moisture
characteristic curve becomes a worthwhile goal. Such predictions will become more
accurate as more moisture characteristic data become available.
Paivenen (1973) studied the relationship between the moisture retention
characteristics of a range of Finnish peats and their bulk densities and degrees of
humification. Figure 2.15 shows his data on the water content-bulk density and water
content-humification relationships at different values of bulk density. The difference in
behavior between large and small pore tensions is evident. For tensions in the range of
10 to 100 cm, the water content increases with increasing bulk density up to a certain
critical bulk density on the order of 0.10 to 0.15 g/cm3. The relationship for humification
is similar to that for bulk density. Paivenen (1973) concluded that bulk density is
superior to humification as a means of predicting water retention behavior, arguing that
bulk density captures more of the range of differences in physical properties in peats. He
also maintained that bulk density reflects changes in the pore structure immediately after
they occur, whereas the degree of humification takes a longer period of time to adjust.
Hysteresis in mineral soils is primarily a physical process attributed to the pore
size distribution of the soil. While several researchers have measured moisture
characteristic curves for peat upon drainage, very few have studied hysteresis effects.
Weaver and Speir (1960) attempt to provide information on the water gained or released
by a cultivated Everglades peat both under rising and under falling water table conditions.
However, they do not present their results clearly.
As in clays, shrinking and swelling behavior in organic soils may add to hysteretic
behavior. In the case of peat, shrinking and swelling effects may have both a reversible
and an irreversible component. Reversible shrinking increases the specific yield of the
peat beyond that expected without such effects. Shrinking and swelling are difficult to
observe on small cores in the laboratory, leading to an underestimation of water release
(Youngs et al., 1989). Romanov (1968) reports a maximum height increase of 2-4% due
to swelling of a variety of peat cores in the laboratory. Silins and Rothwell (1998) report
a volume shrinkage upon drying of 4.5% in previously drained peat and 6.5% in
previously undrained peat. They estimated that this shrinkage resulted in underestimation
of water contents by 2.6% and 2.4%, respectively. Due to the great uncertainties in the
field, such effects can most likely be neglected.
970- 7'0 pe Q*
A A a m FU 2
40- 4 [ ;70--
lUlk density, g M Degree of humilication
SymbolspF1 and F4 pea Smbo F and F4 S Peal
L peaty L peat*
Figure 2.15. Relationship Between Water Content, Tension, Bulk Density, and von Post
Degree ofHumification (Paivenen, 1973). pF is the log of tension head in cm.
Effects of layering
Layering can have a significant effect on the water retention behavior of peat.
Layering may occur both within the peeaat and another soil type. In
deep peat formations, physical properties may vary considerably throughout the depth.
The degree of decomposition generally increases with depth and with the age of the
material. Where an abrupt change in the living plant composition has occurred in the
past, a distinct change in peat type may occur at the corresponding depth. In subtropical
systems, such as those in south Florida, peat formations are usually not deep. A
maximum of a few meters of peat often overlies a mineral soil. In this case, the interface
between the mineral soil and the peat has an effect on the moisture characteristics of the
peat. In the situation when the water table drops below this interface, the effect on soil
moisture is largely unstudied.
Figure 2.16 is an example of the variation in water retention characteristics
between different peat layers in a Finnish peat. In this study (Heikurainen et al., 1964),
the tension-water content relationships are reasonably linear within the range of weights
studied. The deeper layers, with a greater degree of humification, retain a greater weight
of water as the groundwater elevation decreases. The slopes of the lines become less
negative with increasing depth, indicating that the less-decomposed surface peat layer
releases proportionately more water as the water table drops.
Time to equilibrium
In addition to the equilibrium moisture characteristic curve, the time required to
reach equilibrium is of interest. After a change in the pressure difference applied to a soil
sample, equilibrium occurs when the water content reaches a constant value. The time
required to reach equilibrium may be considerably greater in peats than in mineral soils.
This time appears to depend on the peat type, the degree of humification, and the
magnitude of the pressure difference and absolute pressures applied. The times reported
by different researchers range from 24 hours to more than one month, with times of about
5 to 10 days being typical for tensions on the order of 10 to 100 cm.
2.17 indicates that more humified peat retains a greater proportion of its water at
saturation and that's it reaches equilibrium more quickly than less humified peat.
Evxperimglades, reperforted by Pthat 85% of the total amount of water released was ime required inreased in th
Figure 2.17. Time Required to Reach Equilibrium for Different Peat Samples at
Different Tensions (Paivenen, 1973). The bulk densities of the samples are as follows:
A, 0.056 g/cm3; B, 0.091 g/cm3; C, 0.155 g/cm3.
first 24 hours after adjustment. Their samples reached equilibrium in about 5 days. They
attributed this extended period to "retarded release from hydrophyllic substances and
slow movement through the walls of intact plant cells (Weaver and Speir, 1960, p. 2)."
Boelter (1964a) found times of about 4 days for pressure plate apparatus and about 8 days
for membrane extractor apparatus necessary to reach equilibrium moisture conditions.
Silins and Rothwell (1998) report times of 2 days at 10 cm tension and low bulk density
and 10 days at 100 cm tension and higher bulk density. For these results, the trend
between bulk density and time to equilibrium is the opposite of that reported by
Significance of Unsaturated Peat within the Hydrologic Cycle
Even when the hydrology of a system is in a steady state long enough for the soil
moisture to reach an equilibrium condition, the measured moisture characteristic curve
applies only to the specific conditions prevalent at the time. Seasonal changes and
changes from year to year affect the behavior of peat systems. Munro (1984) collected
soil moisture data and other hydrologic data in a Canadian deciduous swamp during three
consecutive summers. The relationship between soil moisture, expressed as a deficit, and
water table depth varies for the three summers as shown in Figure 2.18. The data from
each summer can be fit by a straight line, but the slopes of these lines differ. In contrast,
Heikuraninen et al. (1964) found that data for a Finnish peat from two different summers
displayed parallel linear behavior (Figure 2.11). While moisture characteristic curves
provide valuable information about the hydraulic behavior of peat, a single curve should
not be assumed to capture the full range of that behavior. Many factors discussed in this
paper, both reversible and irreversible, may help account for such differences.
Capillary action is the main mechanism by which phreatic water reaches the root
zone and the atmosphere. The pore size distribution determines the effectiveness with
which the peat transmits water from the saturated zone upward. When "the proportion of
the effective capillary pore system in the total pore volume increases,...the capillary rise
of water is capable of compensating for the loss of water due to evaporation in the
topmost peat layer by transferring water from the ground water table to the surface
(Paivenen, 1973, p. 56)." Evapotranspiration is extremely important in subtropical and
tropical systems, where it often exceeds average rainfall.
20- / /"
0.1 0.2 0.3 0.4 1
Figure 2.18. Relationship Between Soil Moisture Deficit and Water Table Depth for
Three Summers (Munro, 1984). Soil moisture deficit, Sd, is the difference between
saturated water content and measured water content. z' is the water table depth. The data
are presented as follows: 1976 (squares, solid line), 1977 (triangles, broken line), 1978
(circles, dashed line).
Heikurainen et al. (1964) found that capillary action can keep the surface water
content of peat fairly constant during dry periods. During a two week dry period, they
measured a 1.6% change in water content at the surface. A water balance analysis
showed that 16 times more water had evaporated than was lost by the unsaturated layer
alone. This result indicates that capillary action transports phreatic water to the surface
very efficiently without great losses of soil moisture in the vadose zone. Romanov
(1968) confirmed that ET does not significantly depress the upper limit of the capillary
fringe, given a constant water table. Eventually, however, water loss to evaporation will
depress the water table. The length of the drought period is important in determining
how much soil moisture is lost.
Youngs et al. (1989) numerically model the relationship between ET and water
table depth in an English peat using drainage theory. They state, "the flux of water
passing through the water table, that determines the water table height, has a component
due to changes in the volume of water held in the unsaturated soil-water zone as well as a
component due to the flux through the soil surface (Youngs et al. 1989, p. 301)." They
demonstrate good agreement between the model and data, but the model itself is too
complex to present here.
Infiltration of rainfall is an important part of the hydrologic cycle, particularly for
isolated wetland systems with no other inputs. The total amount and the rate of
infiltration are important in determining the recovery of a system when rainfall returns
after a dry period. One important factor in controlling infiltration is pore size. Rainfall
may quickly infiltrate the upper layers of only slightly decomposed peat, reaching the
capillary zone nearly instantaneously. This process can lead to the phenomenon of rapid
water table rise (Romanov, 1968), discussed later in the paper.
Heliotis (1989) provides an equation intended to predict the amount of rainfall
necessary to raise the water table a given increment, assuming an exponential decrease in
specific yield with depth.
P A e-kDu ekDl) (2.10)
where P = amount of precipitation necessary to raise the water table [L3],
A, k = fitting parameters,
Du = final upper depth of the water table [L], and
Dl = initial lower water table depth [L].
Examples of the fitting parameters determined by Heliotis for the four sampling
stations shown in Figure 2.7 are listed in Table 2.4. This model fits the observed data
closely, with r2 values of 0.904 to 0.952. The model is graphed for the four stations in
Figure 2.19. Unfortunately, because Heliotis does not provide detailed information on
the physical properties of the soil at the four sampling stations, it is difficult to use these
fitting parameters for predictive purposes. This model is valid for equilibrium conditions.
Where short but frequent rainfall events are common, the system may never reach a
steady-state condition in which equilibrium theories can be applied.
Table 2.4. Fitting Parameters for Heliotis (1989) Equation.
Station A k
A 0.69 0.0048
B 0.35 0.0045
C 0.22 0.0028
D 0.27 0.0031
Rainfall m) ---- Station C
Go. ..-- --.~~~ -. Station C
o i A
Hater-table Depth (mi)
Figure 2.19. Rainfall Required for Flooding Given Water Table Depth (Heliotis, 1989).
The model is graphed for the fitting parameters in Table 2.4.
Rapid water table rise
The equilibrium water storage of peat depends on water stored in pore spaces and
water bound by other physical and chemical forces. However, the capillary behavior of
the unsaturated zone in peat can lead to a rapid rise in the water table during heavy rain
events. The water table may temporarily rise above its equilibrium level, storing an
amount of water in the short term that is not accounted for in the equilibrium theory.
This behavior depends on soil pore size, water table depth, and rainfall magnitude
(Heliotis and Dewitt, 1987). Rapid water table rise accounts for 10-35% of the
magnitude of water table rises, as determined in a northern Michigan cedar swamp by
Heliotis and Dewitt (1987). This phenomenon is of little importance in studies of long-
The horizontal hydraulic conductivity of unsaturated peat has been the focus of
some recent research, but a detailed discussion of it is beyond the scope of this paper.
Baird (1997) measured the unsaturated conductivity of a humified fen peat in England
and compared it to that predicted by traditional theory of unsaturated flow in mineral
soils. He determined that flow occurs in larger pores near the soil surface in which no
flow is predicted by the traditional theory. He concluded that "classical soil water models
based solely on the Richards equation may be inappropriate for modeling soil water
movement in this soil type" (Baird, 1997, p. 292). Thus, the use of traditional models
based on the Richards equation may underestimate the total flow in the unsaturated zone
of flow-through systems. Such effects are not of great concern in isolated wetlands with
no surface inflows or outflows.
Irreversible Changes Due to Prolonged Drying
In addition to causing physical hysteresis effects, drying of peat causes
irreversible changes to the physical properties. Thus, the moisture characteristic curve
may change between one drying and wetting cycle and the next, even when both cycles
begin at complete saturation. When peat is dry for a long period of time, considerable
changes in soil water conditions occur (Rothwell et al., 1996).
Boelter (1964a) compared the water retention of an undisturbed, previously air-
dried sample with that of an undisturbed, undried sample. Both samples were of the
same undecomposed Sphagnum moss peat. Figure 2.20 shows that the undried sample
retained more water at every tension value studied. Boelter provides two possible
explanations for this difference. First, shrinkage upon drying causes irreversible changes,
reducing the total porosity of the sample and shifting the pore size distribution in favor of
smaller pores. Thus, the water retention for dried and undried samples is similar for large
values of tension where small pores are dominant. However, the dried sample has lost
much of its ability to hold water at low tensions, where larger pores are dominant.
Boelter also suggests that drying may alter the structure of organic particles, reducing
their ability to hold water through sorption forces.
Silins and Rothwell (1998) provide results for a similar study (Figure 2.21).
However, their results seem to show the opposite trend of that found by Boelter. The
previously drained samples retain more water at all levels of tension than the previously
undrained samples. For peat samples taken near the surface, the difference between
undried and previously dried samples is largest at about 10 cm of tension and decreases
with increasing tension. The overall water retention appears to increase with depth, while
the difference between previously drained and undrained samples decreases with depth.
1 looo -- \------
0 5 10 15 20 25 30 35
Water Content (%)
Figure 2.20. Moisture Characteristic Curves for Air-Dried vs. Undried Samples (Boelter,
One possible explanation for the difference between these two samples is that
Boelter's results show primarily reversible changes due to drying, while Silins and
Rothwell's results show both reversible and irreversible changes. Silins and Rothwell's
measurements were collected at a forested site in Canada that had been drained several
years earlier, while Boelter's samples were air-dried in the laboratory.
After a prolonged period of high tension, as when peat is drained thoroughly for
agriculture, the water content after drainage may be higher than that for the same tension
in an undisturbed peat. The dry period must be long enough for irreversible changes to
occur. Rothwell et al. (1996) studied the long-term soil moisture response to drainage
imposed by evenly spaced ditches in sedge and moss peats in Alberta. Contrary to their
expectations, the water content did not follow the hyperbolic shape of the water table in
between two ditches (Figure 2.22). In fact, the water content appeared to be greatest at
the ditch edges and to decrease with increasing distance from the ditch edge. The spacing
100 (a) 0-10 cm depth (b) 020
75 \ -- Drained i 0.15
-- Undrained F
10-20 cm depth
75 -- 0.15
50 -- 0.10 00
0 : V
25 0.05 a.
.S 0 i .. ....... l I l| -- 0.00 **
0 -20-30 cm depth
75 0.15 (
50 0,10 j'
0 .. ....... 0.00
30-40 cm depth
0 ...| .ll| l ...l -- 0.00
Sat 10 100 1000 10000 Undr Dr
Water potential (- cm pressure head) Peat Type
Figure 2.21. Moisture Characteristic Curves for Similar Drained and Undrained
Wetlands at Various Depths (Silins and Rothwell, 1998).
of ditches was shown to have very little effect on the water content at the midpoint
between two ditches. The researchers attributed this unexpected result to spatial variation
in bulk density.
Irreversible shrinking and subsidence complicate the analysis of peat when it is
subjected to a fluctuating water table. Changes to the porosity and to the pore size
6 50 -
5> Mean 0-30 cm depth
20- ""I11 l11 1 11111
E 0.12 Mean 0-30 cm depth
-40 .,I , ,I I
-9 0 I 'l- I I I '
0 5 10 15 20 25
Distance from ditch edge (m)
Figure 2.22. Water Table Elevation and Soil Water Content as a Function of Distance
from a Ditch (Rothwell et al., 1996).
distribution bring about these irreversible changes. Silins and Rothwell (1998) report that
drainage and subsidence are accompanied by a reduction in pores of 600 |tm diameter or
greater and an increase in pores between 3 and 30 |tm diameter.
Rothwell et al. (1996) studied the elevation loss in Alberta peatlands two to five
years after drainage. The mean elevation loss was 7.5 cm, and elevation loss appeared to
increase with increasing ditch spacing. In the term investigated by Rothwell, subsidence
and the accompanying increase in bulk density may help to buffer the effects of water
table lowering. As the pore size distribution changes, the soil can hold more plant-
available moisture at a given height above the water table. Over a longer time period,
however, the increased mineralization of peat under aerobic conditions can lead to
significant soil loss.
Plant Growth and Soil Moisture in Peat
During periods in which an organic wetland soil is unsaturated, the ability of the
soil to hold water has an effect on the health of plants and the health of the wetland
ecosystem as a whole. Plants take up soil moisture when the tension in the unsaturated
zone is not greater than the ability of their roots to extract it. Small fluctuations in the
height of the capillary fringe can have significant ecological effects (Romanov, 1968).
By convention, plant-available water is considered to be that water residing in soil
with a water content less than the field capacity but greater than the permanent wilting
point (Fetter, 1994). Paivenen (1973) defines an alternative upper limit of available
water in peat at a water content corresponding to 10% air space, based on the belief that
10% air space is necessary for the long-term health of many root systems. The actual
wilting point of most wetland plants is not well-documented. At water contents below
the wilting point, intracellular water retention and chemical adsorptive forces add to the
capillary tension, insuring that water is unavailable for uptake (Clymo, 1983). For
convenience, a tension head on the order of 15,000 cm is often taken as the wilting point
(Paivenen, 1973). In Figure 2.23, Paivenen superimposes the different zones of plant-
available water on his measured relationship between peat bulk density and water
content. In general, the range in which soil moisture is available to plants decreases with
increasing bulk density.
9 0 ...i.... ....
READILY AILABLE WATER
50 / AVAILABLE
1 .O NENT
.04 .08 .12 .16 .20
Bulk density, g/cm3
Figure 2.23. Plant-Available Water Superposed on Water Content-Bulk Density
Relationship (Paivenen, 1973).
Two Russian researchers have measured the wilting point water contents of barley
seedlings on a range of peats. They chose barley seedlings because of the difficulty of
observing wilting in tree seedlings. Pyatetskiy (1976) determined that the wilting water
content fell within the range of 7-30%, with 13% being a typical value. He found typical
values for sand to be less than 5%. For peats containing Sphagnum material, the wilting
point increased with an increasing fraction of Sphagnum and decreased with an
increasing degree of humification. These results suggest that intracellular forces present
in recently deposited Sphagnum material are more important than electrostatic forces,
which might be expected to increase with an increasing degree of humification.
Varfolomeyev (1978) performed similar experiments on a range of mixtures
containing unhumified peat, humified peat, and mineral soil. Mixtures of highly
humified peat and mineral soil showed a strong positive correlation between the content
of organic matter and the wilting point, while mixtures of unhumified peat and mineral
soil did not follow a discernible trend. These results agree with those of Pyatetskiy,
suggesting that the dominant forces are intracellular at a low degree of humification and
electrostatic at a high degree of humification. The peat studied by Varfolomeyev
contained sedge and reed material in addition to Sphagnum material.
Data on the wilting point of plants growing on Florida peats are difficult to find.
Lucas (1982) reports that the water content of Everglades peat at 15,000 cm tension
ranges from 0.88 to 0.92 by weight. He does not report the specific bulk densities of the
samples studied. However, he reports typical values for Everglades peat to range from
0.18 g/cm3 to 0.35 g/cm3. Assuming that the bulk densities of the samples studied fell
within this range, the wilting point water content was between 16% and 32% on a volume
In wetlands, transpiration plays a large role in the amount of total
evapotranspiration. Plants that spread roots only within the upper layer are often adapted
to drought-like conditions. When the water table drops below the minimum level at
which it is supplying moisture to these roots, total evapotranspiration decreases sharply
(Romanov, 1968). Solonevich reported this depth to be about 30-35 cm in a Russian bog
with a capillary rise of approximately 20 cm (as cited by Romanov, 1968).
The effects of a falling water table are somewhat paradoxical. For plants with
very shallow roots, such as Sphagnum species, a long-term drop in the water table causes
dessication and death. However, for higher plants with extensive root systems, a drop in
the water table may be beneficial. The effect of moisture in unsaturated peat on tree
growth has been studied by foresters. The moisture conditions in drained peatlands are
thought to be ideal for tree growth. As the peat decomposes and subsides, its pore spaces
decrease in size and retain more water for a given water table depth, helping to mitigate
the effects of drought or overdrainage (Rothwell et al., 1996). Silins and Rothwell (1998)
report an approximately threefold increase in available water in a forested Alberta
peatland after about five years of drained conditions. Aerobic conditions also alleviate
the problems associated with slow oxygen diffusion and buildup of toxins in wet soils.
Because they are non-vascular plants, live Sphagnum mosses rely entirely on
capillary moisture provided by a shallow water table. Sphagnum mosses have difficulty
obtaining moisture in humified peats with small pore sizes, as in systems in which the
upper layer has been harvested (Price, 1996). In undisturbed systems, living mosses
inhabit the remains of recently deceased mosses. Thus, the surface layer always consists
of undecomposed to slightly decomposed material with large pore spaces, even as the
peat layer accumulates. Soil moisture is ample and easy for plants to extract, provided
the capillary fringe extends near the surface.
The pore spaces in living and undecomposed Sphagnum consist of the spaces
between leaves and branches and the capacity within cells able to hold water internally.
These cells can retain water up to a tension of approximately 100 cm (Price, 1997).
Thus, their effect is likely significant in wetland systems with shallow water tables.
Campeau and Rochefort (1996) studied the effect of water table depth on
Sphagnum regeneration in a greenhouse experiment. Figure 2.24 shows the surface area
covered by young plants three months and six months after planting. For five of the six
species studied, the amount of growth depends strongly on water table depth. After six
months, the covered area is approximately 90-100% for a 5 cm water table depth and
approximately 50-60% for a 15 cm water table depth. Growth is minimal for a water
table depth of 25 cm. One of the six species produces poor growth at all water table
depths. The soil mixture is a horticultural peat mix. Schouwenaars (1988) indicates that
a water table depth of 40 cm or less is necessary for successful regeneration of harvested
European bogs. Water table depth alone is not a good indicator of soil moisture
conditions unless soil physical properties are well understood.
Methods and Difficulties in Experimental Design
Moisture characteristic curve measurement
There are two approaches to measuring the soil moisture characteristic
curve in the laboratory. The first is to apply a pressure difference across a small sample
and measure the release of water as that pressure difference increases. This method
simulates the increasing pore tension at a particular point in the soil as the water table
drops. The advantage of this method is that a wide range of tension values can be
examined. The second method is to perform a careful water budget experiment on a soil
sample in a vertical soil column. The advantage of this method is that it more closely
simulates conditions in the field. Heterogeneous elements cause less error in the larger,
less disturbed sample, and layered conditions present in the field can be simulated. The
A5G FAL FUS CAP MAG PAP
80Water evel: 25 m 15cm 5 m
Planting in Peat with Different Water Table Depths (Campeau and Rochefort, 1996).
ANG = S. angustifolium; FAL = S. fallax; FUS = S. fuscum; CAP = S. capillifolium;
MAG = S. magellanicum; PAP = S. papillosum
disadvantage of this method is that the range of tension values is limited by the height of
Compression of soft peat creates a problem for attempts to generate a moisture
characteristic curve using pressure cell apparatus. Boelter (1964a) reports that a 4 psi
pressure difference across a fibrous peat sample compresses the sample to one-quarter of
its original height. While the test generates a curve under these conditions, a significant
error clearly results.
Tensiometers are the most common apparatus for taking unsaturated
measurements in the field. Hillel (1998) discusses tensiometers and a variety of other
methods employed for this purpose.
Disturbance due to the sampling process complicates the measurement of the
hydraulic properties of peat. Landva et al. (1983) provide three reasons to collect an
undisturbed sample when possible. First, the presence of gases causes a volume change
when stress is relieved upon sampling. Second, extrusion and preparation result in
moisture loss. Third, forcing a very soft soil into a container causes significant changes
to its structure. Various investigators have proposed devices for obtaining undisturbed
peat samples (e.g., Landva et al., 1983; NRC Canada, 1979).
Not all studies have indicated that undisturbed and disturbed samples have
significantly different properties. Landva and LaRochelle (1983) found that geotechnical
tests such as shear and consolidation yielded similar results for undisturbed and for
completely remolded samples. They suggest that, where a thorough characterization of a
site is necessary for geotechnical purposes, a large number of easily-collected disturbed
samples might be preferable to a smaller number of undisturbed ones. A large number of
samples is more likely to capture the range of heterogeneity in the soil. Errors due to
heterogeneity can be minimized by collection of multiple samples, by collection of
samples of maximum practical size, and by performing in situ measurements.
The physical properties of a peat determine its pore structure and thus its
hydraulic behavior. The pore structure depends most significantly on the degree of
decomposition and on the botanical composition. The degree of decomposition can be
indexed inexactly by the von Post scale, fiber content, or bulk density. A greater degree
of decomposition generally indicates a lower hydraulic conductivity under saturated
conditions. Under unsaturated conditions, less-decomposed peats generally hold more
water at low tensions and less water at high tensions than more-decomposed peats.
Botanical composition is most important at a low degree of decomposition and at low
tensions. Ideally, a single indicator property or a few easily measured properties would
allow for accurate prediction of the saturated and unsaturated behavior for any peat type.
Such predictions are not yet possible, but an accurate physical characterization of a peat
can provide some expectations about its behavior.
The time scale of interest is important when studying a system that includes a peat
layer. In the short term (days to months), peat has a hydrologic and ecological buffering
effect. In a saturated system, it retards the drainage of water. In an unsaturated system, it
transports water upward from the water table to the vadose zone, supplying soil moisture
to plants. When the water table drops, the soil moisture does not respond fully for several
days to a month, preventing temporary fluctuations from causing ecological harm. In the
medium term (months to years), drying leads to oxidation of organic matter, causing an
irreversible shift in pore structure. The finer pore structure results in an even lower
hydraulic conductivity and an increased ability to transport water through capillarity at
high tensions. In the long term (years to decades), this same oxidation of organic matter
causes irreversible soil loss and ecological damage.
The effects of layering both within peat and between peat and mineral layers are
very important and require more study. These effects are of most concern in tropical and
subtropical systems. These systems may be more fragile than northern systems because
of their shallow peat layers and underlying mineral soils.
For a given peat system, there exists some maximum drawdown that will preserve
desirable ecological conditions. This drawdown will differ for each system based on
climate, ecology, and peat properties. Because of the hydrologic buffering effect, it may
be possible to exceed this drawdown for short periods of time without causing ecological
Future research into peat should focus on improvements and modifications to
existing methods and on the development of new methods of study. Traditional measures
of soil properties, including water content, porosity, and bulk density, have both
advantages and disadvantages when applied to peat. A more accurate and objective
measure of the degree of decomposition is needed. Poorly-understood chemical and
sorption effects on total pore tension in the unsaturated zone complicate analysis.
Similarly, existing field and laboratory methods have both advantages and disadvantages
when applied to peat. Few mathematical models are intended to describe peat behavior.
Darcy's Law and models of soil moisture are useful but may require modification in order
to accurately predict the occurrence and movement of water in peat.
METHODS, MATERIALS, AND DATA ANALYSIS
Samples were collected at two sites in south Florida, shown in Figure 3.1. Figure
3.2 contains photographs of the sites. Three samples were collected at FP5, a small,
isolated cypress dome in Lee County in southwest Florida. This wetland is
approximately round with a diameter of about 60 m. The thickness of the peat layer
ranges from about 0.3 m to 2 m. The dominant tree species is Taxodium distichum (bald
cypress). Emergent and floating vegetation includes Panicum hemitomon (maidencane),
Pontederia cordata pickerelweedd), Salvinia minima (water fern), and Limnobium
spongia (frog's bit).
Two samples were collected at SV5, a small, isolated herbaceous marsh in Martin
County in southeast Florida. This site is also approximately round, with a diameter of
about 60 m. Dominant species at the site include Hypericumfasticulatum (St. John's
wort), Xyris elliottii (yellow-eyed grass), Blechnum serrulatum (saw fern), Woodwardia
virginica (chain fern), Amphicarpum muhlenbergianum (blue maidencane), Panicum
hemitomon (maidencane), and La,/ /mn/ caroliniana redroott) (Walser, 1998).
Undisturbed Peat Sampling
Sand was collected from the field by digging just below the leaf layer in the forest
near the edge of FP5 and transported back to the laboratory in five-gallon buckets. A
total of five peat cores was collected from the locations listed in Table 3.1. Sampling
locations were chosen to represent a range of conditions. Samples FP5-2 and FP5-3 were
SCALE IN MILES
09 10 20 30 40
.. "- -rr q 4
Figure 3.1. Locations of Two Wetlands Sampled.
Figure 3.2. Photographs of FP5 (left) and SV5 (right).
collected near the edge of standing water, under the cypress canopy. The peat near the
edge seemed to be sandier and contained a large proportion of cypress needles. FP5-4
was collected nearer to the center of the wetland. Near the center, there was a break in
the cypress canopy, and the vegetation consisted of emergent herbaceous species. SV5-1
and SV5-3 were both collected on the western side of SV5 in an area dominated by
L(/L I/ Iathe //' caroliniana.
The cores were obtained with a peat sampler designed by Dr. J.P. Prenger of the Center
for Wetlands and M.W. Clark of the Department of Soil and Water Science, both at the
University of Florida. The device consists of a hollow aluminum tube with steel teeth at
the bottom to cut through the peat layer and handles at the top to allow rotation. A clear
polycarbonate tube with a 3.75 inch (9.5 cm) diameter fits tightly inside the barrel. A
rubber stopper inserted in the upper end of the polycarbonate tube applies a suction
pressure during extraction of the sample. The polycarbonate tube has square teeth cut
into its bottom end in order to hold roots still for cutting by the steel teeth. The rotating
and cutting process allows collection of samples without application of a large downward
pressure. Thus, this method minimizes compaction of the peat structure. Additional
information on the sampling device and its use will be published by its designers at a
Table 3.1. Coordinates of Undisturbed Sample Cores. The coordinates represent
distance from the centers of the wetlands in meters, with north and east as positive
directions. Each wetland had a diameter of approximately 60 m.
Core Date Coordinates
FP5-2 2/27/99 (-11, 11)
FP5-3 2/27/99 (-8, 12)
FP5-4 2/27/99 (-3, 10)
SV5-1 7/1/99 (-11,2)
SV5-3 7/1/99 (-9,2)
Grain Size Distribution for Sand
The grain size distribution of the sand samples was determined according to the
standard method described by Das (1992). Approximately 600 g of sand was dried for 24
hours at 105C. The sand was sifted through five sieve sizes: 20, 40, 60, 140, and 200.
The sieves were shaken for 15 minutes before weighing.
Peat Physical Properties Analysis
The physical properties of fresh peat samples, including bulk density, saturated
volumetric water content (effective porosity), and ash content, were measured in order to
characterize the structure of the peat. From each sampling location listed in Table 3.1,
one small brass ring core was taken at the surface and one was taken at a depth of
approximately 20 cm. These cores were collected from the sides of the holes created by
extraction of the undisturbed peat samples. The bulk density and saturated water content
were determined by first saturating and then drying and weighing a known volume of
soil. The ash content was determined by igniting a known weight of oven-dry soil (based
on Das, 1992; ASTM D 2974-87; ASTM D 4511-92; ASTM D 4531-86).
brass ring, 5.3 cm diameter, 3.0 cm length (intended for Tempe pressure cells)
plastic soaking pan
ceramic drying crucible
balance, sensitive to 0.01 g
aluminum drying tins
1. After collecting an undisturbed peat core in the field (see Undisturbed Peat Sampling
method), measure the depth of the hole. Gently work a peat sample at the desired
depth into a brass ring. Apply as little pressure as possible and use a sharp knife to
cut into the peat if necessary. Use a sharp knife to trim excess material after
collecting the sample. Label the ring with an identifiable marking. Place the ring
and sample in a labeled plastic bag and seal. (These steps may be repeated for any
number of samples.)
2. In the laboratory, position two pieces of window screening in the bottom of the
soaking pan. Place the sample carefully on top of a piece of cheese cloth folded to
create four layers. Fill the pan slowly with water to a depth just below the top of the
samples. Soak for 72 hours (based on ASTM D4511-92).
3. Weigh an aluminum drying tin. Place a small amount of material in the tin,
disturbing it as little as possible to maintain its original water content. (Measuring
the gravimetric water content based on a small amount of material in a tin provides
more accurate data than measuring it based on the material in the crucible. Excess
water may be transferred when the sample is transferred from the drying pan to the
drying crucible.) Weigh a drying crucible. Extrude the sample into the crucible.
Weigh the tin and the crucible containing the saturated peat.
4. Dry the tins and crucibles at 105C for 24 hours. Record the weight of the tin and
dry sample and the weight of the crucible and dry sample.
5. Ignite the sample in the crucible for 24 hours in the muffle furnace with the
temperature at a minimum of 440C (based on ASTM D 2974-87).
6. Record the weight of the crucible and ignited sample.
The dry bulk density, gravimetric and volumetric water contents, and ash content
can all be calculated as functions of the measured raw data.
wg M M M (3.2)
S= vw W(3.3)
ash content = ignited (3.4)
where pb = dry bulk density [M/L3],
wg = gravimetric water content (fraction),
0 = volumetric water content, approximate effective porosity (fraction),
Ms = mass of dry solids [M],
VT = total volume [L3],
Mw = mass of water [M],
MT = total mass [M],
Vw = volume of water [L3],
pw = density of water at a given temperature [M/L3], and
Mignited = mass of ignited sample [M].
Tempe Pressure Cell Water Retention Test For Sand
Tempe pressure cells provide a method of measuring the moisture characteristic
curve of a small soil sample by applying a pressure difference across its length. Pressure
cell determinations were performed on two samples of sand from FP5. Each soil sample
was first saturated and placed in a cell. With each pressure difference applied, the
amount of water released was recorded by weight.
Tempe pressure cell (SoilMoisture Equipment Corp., Goleta, CA)
pressure difference equipment, and tubes (SoilMoisture Equipment Corp., Goleta, CA)
vibrating instrument (e.g., electric back massager)
one graduated cylinder for each sample and one extra
1. Weigh the cell and the sample ring dry. (These directions may be extended to any
number of cells.)
2. Submerge the porous stone for approximately 15 minutes in deionized water.
3. Place the porous stone and brass ring on the bottom of the Tempe cell.
4. Add soil to the brass ring in several shallow layers, saturating each layer. Smooth
the top with a flat knife. Vibrate to remove air bubbles. Repeat as needed.
5. Assemble the top of the cell. Cover the outlet to keep water in.
6. Weigh the cell. Weigh any water remaining on the scale.
7. Weigh and take an initial volume reading from the receiving graduated cylinder and
from the evaporation cylinder.
8. Measure the initial pressure. Make sure the tube is open to the atmosphere at both
9. Push the tube onto the cell. Adjust valves. Measure the temperature.
10. Allow to equilibrate four hours or until a constant weight is reached in the receiving
cylinder. Dislodge large hanging drops into the receiving cylinder.
11. Measure the volume and weight of the graduated cylinders.
12. Increase the pressure difference 1 to 2 inches.
13. Repeat steps 9 through 11 until no more water is discharged with increasing pressure
14. Weigh the apparatus.
15. Oven dry and weigh the soil sample.
Note: These steps were followed by the author. Additional information on the
recommended use of this apparatus is available from the manufacturer.
1. Graph head difference vs. cumulative volume released.
2. Graph head difference vs. volumetric water content.
Calculation of the volumetric water content at a given head difference consists of
several steps. First, the initial mass and volume of water present in the sample are
calculated. The total volume of the sample is calculated based on ring geometry. Finally,
the volumetric water content is calculated as a function of the initial volume of water, the
total volume, and the total water released at a particular head difference.
O(Ah) = x 100%
wI Pw(T) (3.7)
M .=M. -M
wi Ti s (3.8)
where O(Ah) = volumetric water content at a particular head difference (%),
Vwi = initial volume of water [L3],
Vr(Ah) = cumulative volume of water released at a particular head difference [L3],
VT = total sample volume [L3],
d = sample diameter [L],
Az = sample length [L],
Mwi = initial mass of water [M],
p,(T) = density of water as a function of temperature [M/L3],
MTi = initial total mass [M], and
Ms = mass of oven-dried soil [M].
Two-Tube Peat Testing Apparatus
A two-tube peat testing apparatus was designed and used to conduct saturated and
unsaturated experiments. With the exception of covers to prevent evaporation, Figure 3.3
shows the parts necessary to construct the two-tube peat testing apparatus. The clear
PVC tubing was manufactured by NewAge Industries Inc. of Willow Grove,
Pennsylvania and purchased through Aquatic Ecosystems Inc. of Apopka, Florida. The
clear polycarbonate tubing was manufactured by the Excelon Corporation. Except for the
clear PVC and clear polycarbonate tubing, all parts were readily available at a local
building supply store. Parts were connected using PVC glue in all cases except the 2"
coupler connecting the peat and sand tubes. This coupler was glued in place with
removable hot melt adhesive. To prevent evaporation, the tubes were covered by plastic
film with a small opening to allow pressure equilibration. The apparatus was supported
by the wooden structural frame seen in the photograph in Figure 3.4.
Diia, Clear Polycarbonate Tubing
-ndlscurloed Pea- Sample
1,5" DIan Clear PVC Pipe- 112" Leng-h, 4 DI Clear PVC Pipe
_-Repacked Sand Sample
4"-1.5" PVC Adoapter
.75" FlexibLe Vinyl Tubing
1.5" PVC T-Flt-tIng
/ 1,5"-0,75" Bushing on Each End
.75" Nozzle on Right, 0.75" Valve on Left
Figure 3.3. Diagram of the Two-Tube Peat Testing Apparatus.
Figure 3.4. Photograph of the Two-Tube Peat Testing Apparatus and Support Structure.
Saturated Hydraulic Conductivity
Studies of the saturated hydraulic conductivity were performed on sand and on
two-layer peat-sand systems in the two-tube peat testing apparatus. By saturating the
peat layer and creating a difference in water level between the two tubes, flow was
induced through the sample. A range of initial head differences was applied in order to
account for possible compression effects. Higher absolute heads may compress the soft
soil structure of peat, reducing pore sizes and hydraulic conductivity. The water level in
each tube was recorded at at least five different times. Table 3.2 and Table 3.3 list all the
hydraulic conductivity tests conducted on peat, along with the initial water levels in each
of the two tubes.
Table 3.2. Saturated Hydraulic Conductivity Tests on Sand. hl is the water level in the
sample tube, and h2 is the water level in the reservoir tube.
Test Sample Start Date Flow Direction Initial hI Initial h2
1 FP5 Sand 5/27/99 Down 109 61
2 FP5 Sand 5/27/99 Down 109 25
3 FP5 Sand 5/27/99 Down 109 48
4 FP5 Sand 5/27/99 Down 109 77
5 FP5 Sand 5/27/99 Down 109 11
6 FP5 Sand 5/27/99 Up 97 154
7 FP5 Sand 5/27/99 Up 97 153
8 FP5 Sand 5/27/99 Up 97 130
9 FP5 Sand 5/27/99 Up 97 177
10 FP5 Sand 5/27/99 Up 97 165
two-tube peat-testing apparatus
carpenter's water level
1. Saturate the sample by adding water to the reservoir tube. Wait for the water level in
the sample tube to rise above the level of the top of the sample.
2. Raise or lower the water level in the reservoir tube in order to establish the desired
difference between water level heights. For an upward flow test, the water level in
the reservoir tube should be greater than that in the sample tube. For a downward
flow test, the water level in the reservoir tube should be less than that in the sample
tube. For different tests, establish a variety of initial head differences ranging from
about 15 cm to 100 cm. When performing upward flow tests, the length of the tube
above the top of the peat sample limits the water level difference that can be applied.
3. Record the time and the initial heights of the two water levels.
4. Record the water levels at a minimum of five different times. The change in water
level in between readings should be about an inch.
Table 3.3. Saturated Hydraulic Conductivity Tests on Peat. hl is the water level in the
sample tube, and h2 is the water level in the reservoir tube.
Test Sample Start Date Flow Direction Initial h(cm) Initial h2
1 FP5-2 4/7/99 Up 169 197
2 FP5-2 4/7/99 Up 169 204
3 FP5-2 4/7/99 Down 173 142
4 FP5-2 4/7/99 Down 174 144
5 FP5-3 6/26/99 Up 145 180
6 FP5-3 6/26/99 Up 148 203
7 FP5-3 6/26/99 Down 154 101
8 FP5-3 6/26/99 Down 152 118
9 FP5-3 3/5/99 Up 143 198
10 FP5-3 7/6/99 Down 150 11
11 FP5-4 4/6/99 Up 160 200
12 FP5-4 4/6/99 Up 162 200
13 FP5-4 4/6/99 Down 162 138
14 FP5-4 4/6/99 Down 163 138
15 SV5-1 7/6/99 Up 159 201
16 SV5-1 7/6/99 Up 160 195
17 SV5-1 7/6/99 Up 154 189
18 SV5-1 7/6/99 Up 154 177
19 SV5-1 7/6/99 Up 150 205
20 SV5-1 7/6/99 Up 150 197
21 SV5-1 7/6/99 Up 151 205
22 SV5-1 7/6/99 Down 165 130
23 SV5-1 7/6/99 Down 165 87
24 SV5-1 7/6/99 Down 159 122
25 SV5-1 7/6/99 Down 157 79
26 SV5-1 7/6/99 Down 158 102
27 SV5-1 7/6/99 Down 157 86
28 SV5-3 8/4/99 Up 168 199
29 SV5-3 8/4/99 Up 181 205
30 SV5-3 8/4/99 Up 180 205
31 SV5-3 8/4/99 Up 177 194
32 SV5-3 8/4/99 Up 176 204
33 SV5-3 8/4/99 Down 185 155
34 SV5-3 8/4/99 Down 183 140
35 SV5-3 8/4/99 Down 182 125
36 SV5-3 8/4/99 Down 178 121
37 SV5-3 8/4/99 Down 178 120
Appendix A contains a complete derivation of equation 3.9, the expression that
describes the two-tube permeameter. The form of the derivation and the final equation
differ slightly from those presented by Daniel (1989), but the numerical results are
identical. The equation predicts the water level in the reservoir tube at any time.
Constants include the initial water level in each tube, the cross-sectional area of each
tube, the sample length, and the hydraulic conductivity. The initial water level in the
sample tube is incorporated in the P3 parameter. Equation 3.9 can be used to predict the
water level in the reservoir tube if the hydraulic conductivity is known. When a set of
experimental data is available, an optimum value of hydraulic conductivity can be
determined by least squares regression.
(P+yh )exp --KAI t
2 2h LA2 (3.9)
where K = average hydraulic conductivity of the total sample [L/T],
L = sample length [L],
A1 = sample tube cross-sectional area [L2],
A2 = reservoir tube cross-sectional area [L2],
h2 = reservoir tube water level at any time [L],
t = time [T],
h = initial sample tube water level [L],
h = initial reservoir tube water level [L],
= cah h (3.10)
y 1-c, and (3.11)
In the case of a column containing a single layer, the optimum value of hydraulic
conductivity resulting from the method above represents that of the material in the
column. In the case of a two layer column, the value represents an average value for the
two layers. When the hydraulic conductivity of one layer is known, the hydraulic
conductivity of the other layer can be determined using a harmonic average. In the case
of a sand layer and a peat layer, the following equation applies:
peat K Lsand (3.13)
where K = average column hydraulic conductivity [L/T],
Kpeat = peat hydraulic conductivity [L/T],
Ksand = sand hydraulic conductivity [L/T],
L = total column length [L],
Lpeat = peat sample length [L], and
Lsand = sand sample length [L].
Upward and downward hydraulic conductivity values were compared by
calculating the ratio of their mean values. When the sample size was large enough, one-
tailed hypothesis tests were performed for samples of equal variance using the t-statistic
at a significance level of 5%. Appendix B contains details of all statistical tests.
Unsaturated Column Water Balance Test
The two-tube apparatus provided a means of studying the water retention and
water release of unsaturated samples when subjected to changes in water table depth. For
falling water table experiments, the peat was first saturated by adding water to the
reservoir tube until the peat layer was completely saturated. The reservoir water level
was then lowered by increments of approximately 8 cm. After a change in reservoir
water level, the system was allowed to equilibrate until no measurable change in water
level occurred over a twelve-hour period (three hours for sand). The recovery of the
water level in the reservoir tube after each adjustment represented the amount of water
released from the soil samples at that water table depth.
two-tube peat testing apparatus
carpenter's water level
100 mL graduated cylinder
1. Add water to the reservoir tube to raise the water level in the sample tube to the top
of the peat sample. Wait for the water levels to equilibrate. (This step may take
2. Record the water level in the reservoir tube and in the sample tube to the nearest
1/16". Use a carpenter's level to measure the sample tube water level accurately.
3. Lower the water level in the reservoir tube 2-3 inches by opening the bottom valve
and discharging water into a graduated cylinder. Record the volume of water
discharged. (A discharge of 100 mL corresponds to a water level change of about 3
inches.) Record the reservoir water level. Record the physical appearance of the
sand and peat layers, including any apparent capillary fringe. Repeat measurements
every 12 hours for peat-sand columns or every three hours for sand columns. When
the water level in the reservoir tube does not change measurably over a sampling
interval, consider the system to be in equilibrium. Repeat steps 2 and 3 until
equilibrium is established.
Note: To perform a rising water table test, change step 3 by adding a known volume of
water to the reservoir tube during each adjustment rather than releasing water
from the bottom valve. Be sure to replace the evaporation cover after each
1. Graph the cumulative amount of water released as a function of water table depth.
2. When the sample tested contains a sand layer only, fit the integrated van Genuchten
equation to the data as discussed below.
Data obtained from column experiments on unsaturated sand were fit by a
numerical integration with respect to depth of the van Genuchten (1980) moisture
characteristic curve function. Van Genuchten proposed to describe moisture
characteristic curve data using the following equation:
0 -r 1 + (3.14)
0 1 + (ah)n
where 0 = dimensionless scaled water content (fraction),
0 = water content (fraction or percentage),
Or = residual water content (fraction or percentage),
0s = saturated water content (fraction or percentage),
h = tension head or water table depth [L], and
uc, n = shape parameters dimensionlesss).
This form of the equation fits data from Tempe pressure cells. For column
experiments in which the sample is never completely drained, integration of the van
Genuchten equation provides a means of predicting the total amount of water released as
the sand above the water table desaturates. The expression is difficult to integrate
analytically, but the trapezoidal rule provides a sufficiently accurate numerical
approximation. A least squares fit of an experimental data set generates optimum values
of the two shape parameters.
Figure 4.3 in the Results and Discussion section contains an example of the
integrated van Genuchten function. The calculations required to generate this figure are
summarized in Table 3.4. For these calculations, a = 0.0128 cm-1 and n = 7.507. Column
1 contains values of water table depth in cm at which data points were taken in the
laboratory. Column 2 contains values of the scaled dimensionless water content
subtracted from one. Numerical integration of Column 2 yields Column 3, a
representation of the amount of water remaining at a given depth. These values represent
the water released by the unsaturated sand above the water table in units of cm. Finally,
values in column three are multiplied by the cross-sectional area and by the effective
porosity of the sample to determine the total water released in cm3 when the water table
falls to a certain depth.
Table 3.4. Example of Calculations Required to Predict Cumulative Discharge Values.
(1) (2) (3) (4)
h 1 Int. (1 -) Cum. Vol.
20 0.0000 0.0001 0
25 0.0002 0.0012 0
30 0.0008 0.0051 0
37 0.0035 0.0298 1
43 0.0096 0.0819 3
47 0.0205 0.1762 6
53 0.0448 0.4179 14
58 0.0883 0.8947 29
63 0.1434 1.5322 50
66 0.1944 2.1495 70
68 0.2387 2.7180 88
70 0.2774 3.2465 105
74 0.3502 4.4139 143
75 0.3899 5.0948 165
78 0.4447 6.1537 199
80 0.5093 7.6091 247
83 0.5644 9.0427 293
84 0.5939 9.8912 320
87 0.6343 11.2002 363
88 0.6634 12.2534 397
89 0.6827 13.0120 422
Unsaturated Column Physical Analysis
After completion of saturated and unsaturated testing in each column, an
equilibrium water table was established at a depth of about 125 cm below the top of the
peat. The columns were dismantled, and small ring cores were taken at six evenly-spaced
depths along the sample length. Water content, bulk density, effective porosity, and ash
content were determined by the methods described in the Peat Physical Properties
See Peat Physical Properties Analysis Method.
1. Lower the water table to establish it at an equilibrium elevation of between 5 and 6
inches (about 13 and 15 cm; measuring stick is graduated in inches only).
2. Remove the 4" PVC coupling by cutting it with a hacksaw at the level where the two
sample tubes meet. Do not damage the sample tubes. Set the peat sample tube
gently on its sides and cover both ends to prevent evaporation. Cover the upper end
of the sand tube to prevent evaporation.
3. Obtain brass ring samples at at least six depths throughout the peat profile. At each
sampling depth, obtain a small amount of soil and immediately weigh it in a drying
tin for gravimetric water content determination. Obtain samples every 10 cm (if
reachable) within the sand profile.
4. Obtain physical properties measurements according to the Peat Physical Properties
RESULTS AND DISCUSSION
Grain Size Distribution
Figure 4.1 is a grain size distribution for a sand sample taken at FP5. The sample
appears to be well sorted, with most grains having diameters between 0.1 and 0.3 mm.
This range of grain sizes corresponds to a fine sand in most soil classification systems
(Holtz and Kovacs, 1981). The median grain diameter is approximately 0.17 mm. The
coefficient of uniformity (D60/D10) is approximately 1.5, confirming that the sand is well-
8 0 -'*-------- ---------- ---
-: --- - ~ ~- .^ C - _ _
0.01 0.1 1
Grain Diameter (mm)
Figure 4.1. FP5 Sand Grain Size Distribution.
Saturated Hydraulic Conductivity
The hydraulic conductivity of sand collected at FP5 was investigated using the
two-tube permeameter apparatus. For each test, the optimum value of hydraulic
conductivity was determined by a least-squares fit of the mathematical model. Table 4.1
contains hydraulic conductivity values and the sum of squared residuals for each of the
tests conducted on sand. The sum of squares provides a measure of how well the
mathematical model fits the experimental data. Predictions of the reservoir tube water
level, h2, appear to fit the observed data closely, indicating that the mathematical model
accurately described the physical processes taking place in the column. The
mathematical model is discussed in the Data Analysis section and derived fully in
Appendix A. Figure 4.2 is an example of the raw data collected during one of the
experiments on FP5 sand.
Table 4.1. Hydraulic Conductivity Values and Sums of Squared Residuals for all Sand
Test Sample Start Date Flow Direction K K Sum of Squares
(cm/s) (m/day) (cm2 )
1 FP5 Sand 5/27/99 Down 6.11x10-3 5.28 2.908
2 FP5 Sand 5/27/99 Down 5.98x10-3 5.16 0.662
3 FP5 Sand 5/27/99 Down 5.91x10-3 5.10 0.209
4 FP5 Sand 5/27/99 Down 5.83x10-3 5.04 0.021
5 FP5 Sand 5/27/99 Down 5.95x10-3 5.14 0.729
6 FP5 Sand 5/27/99 Up 5.96x10-3 5.15 0.271
7 FP5 Sand 5/27/99 Up 6.15x10-3 5.31 2.580
8 FP5 Sand 5/27/99 Up 6.07x10-3 5.25 0.033
9 FP5 Sand 5/27/99 Up 6.11xl0-3 5.28 0.113
10 FP5 Sand 5/27/99 Up 5.95x10-3 5.14 0.043
0 10 20 30 40 50 60
Observed h2 Predicted h2 Predicted h1
Figure 4.2. Two-Tube Permeameter Observed and Predicted Values for FP5 Sand. h2 is
the water level in the reservoir tube, while hl is the water level in the sample tube.
Table 4.2 contains mean values of hydraulic conductivity determined for upward
and downward flow through FP5 sand and a mean value for all the two-tube tests taken
together. Initial inspection of the data shows that the mean value for upward conductivity
is greater than that for downward conductivity. However, the difference is small and
statistically insignificant at the 5% level. Thus, there is no reason to consider different
values for upward and downward conductivity in the sand. The mean value for overall
conductivity, approximately 6.0x10-3 cm/s or 5.2 m/day, is a representative value to use
when analyzing the behavior of the layered peat-sand system. This value falls within the
upper portion of the range that might be expected for a fine, well-sorted sand (Fetter,
The water retention and release of sand collected at FP5 was investigated both in
Tempe pressure cells and in vertical columns. Figure 4.3 presents two moisture
Table 4.2. Mean Hydraulic Conductivity Values for FP5 Sand. Values are given in
m/day for convenience and in cm/s for comparison to previously published material.
Values for upward and downward flow represent five independent tests for each flow
direction. The overall values are for all ten values taken together.
Downward Flow Upward Flow Overall
Mean (cm/s) 5.95x10-3 6.05x10-3 6.00x10-3
Mean (m/day) 5.14 5.23 5.19
St. Dev. (m/day) 0.0880 0.0777 0.0895
characteristic curves derived from pressure cell trials. The best-fit curves are plotted
according to the model proposed by van Genuchten (1980). The van Genuchten model
includes two shape parameters, a and n. In cell 1, a = 0.01752 cm-1 and n = 16.24. In
cell 2, a = 0.01749 cm-1 and n = 14.38. The height at which the water content changes
abruptly corresponds to the capillary rise and occurs at a water table depth of
approximately 45-50 cm.
In addition to the Tempe pressure cell tests, sand desaturation data were collected
in the two-tube apparatus. Figure 4.4 displays data obtained from an experiment
conducted in an unsaturated sand column in March 1999. Very little water is released
until the tension head reaches a value for which the capillary fringe would no longer
reach the sand surface. As the water table falls further, the sand releases water quickly,
eventually displaying a linear release with increasing depth. The data are fit by a
numerically integrated form of the van Genuchten equation as described in the Methods
and Materials section. While this model seems to fit the data well both conceptually and
0 20 40 60 80 100 120
Tension Head (cm)
Cell 1 Data Cell 1 Fit Cell 2 Data Cell 2 Fit
Figure 4.3. Moisture Characteristic Curves for FP5 Sand Derived from Tempe Pressure
Cell Data and Fit with the van Genuchten (1980) Model.
0 20 40 60 80 100
Water Table Depth (cm)
Data Column Fit Tempe Fit
Figure 4.4. Sand Column Water Release Data With Integrated van Genuchten (1980) Fit.
The Tempe Fit curve represents the shape parameters determined from the Tempe
pressure cell fits shown in Figure 4.3 (a = 0.01749, n = 14.38). The Column Fit
represents parameters obtained from a least-squares fit on column data. (a = 0.0128, n =
in practice, it provides a slightly more gradual transition between the two linear portions
of the curve than the one suggested by the data.
The two curves in Figure 4.4 compare shape parameters obtained from Tempe
pressure cells to those obtained from column experiments. Clearly, the Tempe cell
behavior does not correspond to the behavior observed in the columns. The sharp release
corresponding to the capillary fringe depth occurs at about 50 cm in the Tempe cell and at
about 60 cm in the column. This difference suggests that, contrary to expectations, the
Tempe cell samples may contain larger void spaces than those in the columns. While
data from Tempe cells are more convenient to collect and more reproducible than those
from column experiments, they did not accurately represent the behavior of the sand in
Observation of the peat cores collected from FP5 and SV5 reveals that the upper
5-10 cm of each core contains nearly pure peat with only a slight degree of
decomposition. Individual plant parts are visible, and in many cases the plant type is still
identifiable. Below about 10 cm, however, the organic material is well-decomposed and
unidentifiable. Each time an undisturbed core sample was removed, small samples were
collected in brass rings from the side of the hole created. These samples were tested
upon returning to the lab to yield the results in Table 4.3. The data for both wetlands
present several expected trends. Bulk density increases with depth, porosity decreases
with depth, and ash content increases with depth.
Table 4.3. Physical Properties of Fresh Peat Samples.
Sample Depth Bulk Density Porosity Ash Content
(cm) (g/cm3) (%) (%)
FP5-2 0 0.19 97 32
FP5-2 35 0.85 69 90
FP5-3 0 0.17 96 26
FP5-3 20 0.89 58 91
FP5-4 0 0.21 101 39
FP5-4 30 0.73 79 81
SV5-1 0 0.22 82 49
SV5-1 20 1.09 59 91
SV5-2 0 0.13 78 35
SV5-2 20 1.15 59 92
SV5-3 0 0.19 88 41
SV5-3 20 0.74 67 83
Measured values of bulk density and porosity correspond to ranges reported in the
literature. The bulk densities of pure peats range from about 0.1 to 0.2 g/cm3, depending
on the degree of humification (Clymo, 1983). The values for surface peat measured in
FP5 and SV5 fall within the upper portion of this range. Although the surface peat
samples were mostly unhumified, they contained some mineral material and a vegetative
composition different from those of the temperate wetlands most represented in the
literature. At greater depths, the samples have bulk densities intermediate between values
expected of pure peat and pure sand. These values are reasonable because the sand
content of the peat increases from 25-50% at the surface to 80-90% at a depth of 20 cm or
more. Porosity values given in the literature typically range from 80-100% for pure
peats, depending on the degree of humification (Boelter, 1969). Surface samples
collected from FP5 and SV5 have porosities of 95-100%, corresponding well to ranges
given for unhumified to slightly humified peats. The porosity decreases to 60-80% at a
depth of 20 cm or more, most likely due to the increasing fraction of sand.
Both the ash content and the degree of decomposition are responsible for a
portion of the change in porosity and bulk density with depth. At the surface, the peat
consists of undecomposed to slightly decomposed plant matter. This material has a low
density and a high porosity. As the degree of decomposition increases with depth, the
organic matter becomes more decomposed, denser, and less porous. However, the
presence of such a large proportion of mineral matter likely has a dominant effect on bulk
density and porosity.
The presence of such a large proportion of mineral matter raises questions about
whether the soils studied can properly be termed peat. However, the existing systems of
peat classification were developed primarily for agricultural and fuel-related applications
and are not completely applicable to hydrology. In addition, the data obtained from the
ash content procedure may tend to overstate the influence of mineral matter for several
reasons. First, because the bulk density of the mineral matter, primarily sand, in these
wetland soils is an order of magnitude greater than the bulk density of the organic matter,
expressing the mineral content on a weight basis tends to overstate its importance.
Second, the ash content procedure does not provide a perfect measure of the mineral
content of a highly organic soil. Visual inspection of the material remaining after
ignition at 5500C indicates that it is not all sand. This situation is particularly true for
samples collected near the surface, where much of the remaining material may be
incombustible plant matter. Even with an ash content of 90%, the soil may still contain
enough organic matter to behave hydrologically as a peat.
No meaningful results were obtained from tests intended to estimate the degree of
humification because the most widely cited tests are not applicable to soils with an
appreciable mineral content. The von Post humification test is somewhat subjective
under any circumstances and did not yield useful results for the samples studied. The
fiber content test is intended to separate peat particles based on their dimensions, but it
requires a negligible mineral content. In addition, the grain size typically used to separate
fibrous from non-fibrous materials is very close to the median diameter of the sand
Heterogeneity and experimental uncertainty contribute to the high degree of
variability in measuring peat properties. For example, the variability in the bulk density
of surface samples is affected by the sampling procedure. It is difficult to determine at
what point living plant material and roots give way to undecomposed peat. Any
disturbance caused by the sampling method affects both the bulk density and porosity
measurements. The 101% porosity measurement for sample FP5-4 most likely represents
uncertainty introduced by the laboratory procedure and demonstrates the difficulty
associated with measuring porosities very close to 100%.
In addition to being measured in the field, peat physical properties were measured
in each laboratory soil column after the completion of saturated and unsaturated testing.
The system was allowed to reach equilibrium with a water table approximately 125 cm
below the top of the peat. Figure 4.5 contains an example of these results obtained from
sample FP5-3. The bulk density of the sample exhibits a strong positive linear trend with
increasing depth, with an r2 of 0.94. The bulk density ranges from a low of about 0.1
g/cm3, typical of unhumified peat, to a high of about 1.4 g/cm3, close to the bulk density
of a pure sand. The ash content of the sample increases abruptly from about 20% at the
surface to about 80% at a depth of only 6 cm. It then appears to approach 100%
asymptotically with increasing depth. The effective porosity of the sample exhibits a
weak negative linear trend with depth, with an r2 of only 0.40. Given a constant bulk
density and porosity, the water content would be expected to increase with depth,
reaching 100% at the water table elevation. However, variation in the other properties
makes it impossible to discern any water content trend in the sample studied.
Saturated Hydraulic Conductivity
Table 4.4 contains hydraulic conductivity values for all tests conducted on peat.
For FP5 samples, the values range from 5.4x10-6 cm/s to 5.8x10-5 cm/s and correspond to
the range reported in the literature for well-humified peats (e.g., see Figure 2.5). For SV5
samples, the values range from 4.6x105 cm/s to 2.4x10-4 cm/s and correspond to the
range reported in the literature for peats with an intermediate degree of humification (e.g.,
see Figure 2.5). Table 4.4 also contains the sum of the squared residuals for each test.
The sum of squares provides a measure of how well the mathematical model fits the
experimental data. Figure 4.6 contains two examples of raw data collected during
saturated peat-sand experiments. The top figure represents the first downward hydraulic
conductivity test for FP5-3. Even though this experiment had the highest sum of squares
value of all the tests (9.56), the model still appears to fit the data closely. The bottom
figure represents the first upward test for SV5-3. This test had a low sum of squares
(0.061), indicating a close fit.
Mean values of hydraulic conductivity in the upward and downward directions
are shown in Table 4.5 for the undisturbed peat samples. For all five samples, average
upward hydraulic conductivity exceeded average downward hydraulic conductivity by
12-30%. For three of the five samples, the number of tests conducted was great enough
to determine whether the difference between upward and downward conductivity was
0 5 10 15 20 25 30 35
0 5 10 15 20
-*- Porosity (%)
---Water Content (%)
25 30 35
-A-Ash Content (%)
Figure 4.5. Peat Physical Properties of FP5-3 as Measured after Completion of all Other
Testing. The water table is in the sand layer at a depth of approximately 125 cm below
the top of the peat. The trend line for bulk density is linear and has an r2 of 0.94.
significant. One-tailed hypothesis tests at a significance level of 5% yielded mixed
results; the difference was significant for SV5-1 but not for FP5-3 or for SV5-3.
Table 4.4. Hydraulic Conductivity Values for all Saturated Peat Tests and Sums of
Squared Residuals for Fitting of the Mathematical Model.
Test Sample Start Date Flow Direction K K Sum of Squares
(cm/s) (m/day) (cm2)
1 FP5-2 4/7/99 Up 5.76x10-' 0.0498 2.027
2 FP5-2 4/7/99 Up 5.18x10-' 0.0447 2.780
3 FP5-2 4/7/99 Down 4.90x10-' 0.0423 1.091
4 FP5-2 4/7/99 Down 4.34x10-' 0.0375 0.202
5 FP5-3 6/26/99 Up 1.87x10-' 0.0162 3.614
6 FP5-3 6/26/99 Up 1.71x10-' 0.0148 3.226
7 FP5-3 6/26/99 Down 2.67x10-' 0.0231 9.562
8 FP5-3 6/26/99 Down 1.61x10-' 0.0139 2.062
9 FP5-3 3/5/99 Up 3.30x10-' 0.0285 1.868
10 FP5-3 7/6/99 Down 1.86x10-' 0.0160 9.461
11 FP5-4 4/6/99 Up 6.68x10-6 0.0058 0.363
12 FP5-4 4/6/99 Up 7.89x10-6 0.0068 2.656
13 FP5-4 4/6/99 Down 5.42x10-6 0.0047 0.248
14 FP5-4 4/6/99 Down 5.78x10-6 0.0050 0.046
15 SV5-1 7/6/99 Up 3.50xl0-4 0.3024 0.496
16 SV5-1 7/6/99 Up 3.16x10-4 0.2730 0.651
17 SV5-1 7/6/99 Up 2.83x10-4 0.2443 0.121
18 SV5-1 7/6/99 Up 3.02x10-4 0.2610 0.521
19 SV5-1 7/6/99 Up 2.91x10-4 0.2511 1.161
20 SV5-1 7/6/99 Up 2.70x10-4 0.2333 0.607
21 SV5-1 7/6/99 Up 2.78x10-4 0.2402 0.647
22 SV5-1 7/6/99 Down 3.01xl0-4 0.2601 0.453
23 SV5-1 7/6/99 Down 2.73x10-4 0.2356 5.340
24 SV5-1 7/6/99 Down 2.62x10-4 0.2268 0.299
25 SV5-1 7/6/99 Down 2.55xl0-4 0.2207 0.829
26 SV5-1 7/6/99 Down 2.57x10-4 0.2216 2.840
27 SV5-1 7/6/99 Down 2.42x10-4 0.2091 1.521
28 SV5-3 8/4/99 Up 3.69x10-5 0.0319 0.061
29 SV5-3 8/4/99 Up 4.57x10-' 0.0395 0.566
30 SV5-3 8/4/99 Up 4.45x10-' 0.0384 0.681
31 SV5-3 8/4/99 Up 3.50x10-' 0.0302 0.009
32 SV5-3 8/4/99 Up 3.40x10-5 0.0294 0.421
33 SV5-3 8/4/99 Down 4.62x10-5 0.0399 2.519
34 SV5-3 8/4/99 Down 3.89x10-' 0.0336 1.804
35 SV5-3 8/4/99 Down 2.96x10-' 0.0256 1.215
36 SV5-3 8/4/99 Down 3.01x10-' 0.0260 1.167
37 SV5-3 8/4/99 Down 2.62x10-5 0.0226 4.457
Test 7, Sum of Squares = 9.56
0 20 40 60
* h2 Observed h2 Predicted -- h Predicted
Test 28, Sum of Squares = 0.061
0 10 20 30 40 5C
Elapsed Time (hrs)
h2 Observed h2 Predicted hl Predicted
Figure 4.6. Two Examples of Raw Data from Peat-Sand Hydraulic Conductivity
Experiments. hl is the water level in the sample tube, and h2 is the water level in the
Table 4.5. Mean Hydraulic Conductivities of FP5 and SV5 Peat Samples. N/A indicates
a result that was not determined due to a small sample size.
Upward Flow____ Downward Flow
Sample Mean Mean St. Dev. Mean Mean St. Dev. Up/Down Sample Significant
(cm/s) (m/day) (m/day) (cm/s) (m/day) (m/day) (%) Size Difference?
FP5-2 5.47x10-'5 4.72x10-2 N/A 4.62x10-'5 3.99x10-2 N/A 118 2 up, 2 down N/A
FP5-3 2.30x10-' 1.99x10-2 7.50x10-3 2.06x10-' 1.78x10-2 4.88x10-3 112 3 up, 3 down No
FP5-4 7.29x10-6 6.30x10-3 N/A 5.60x10-6 4.84x10-3 N/A 130 2 up, 2 down N/A
SV5-1 2.98x10-4 2.58x10-1 2.37x10-2 2.65x10-4 2.29x10-1 1.75x10-2 113 7 up, 6 down Yes
SV5-3 3.92x10-' 3.40x10-2 4.73x10-3 3.42x10-' 3.00x10-2 7.08x10-3 115 5 up, 5 down No
The overall results and the statistical results for SV5-1 provide some evidence that
hydraulic conductivity was greater in the upward direction. There are several possible
explanations for these results. Frictional forces caused by water flow may have
compressed the soft soil matrix or reoriented plant parts so that they provided more or
less resistance to flow. Fine particles may have been transported into pores, blocking
them during downward flow (Marshall, 1968; Rycroft et al., 1975). Third, some of the
absolute heads applied during upward flow experiments were greater than those applied
during downward flow experiments. Any gas pockets present in the samples may have
been more compressed during upward flow, causing them to provide less flow resistance.
Such an effect would tend to increase the hydraulic conductivity in the upward flow
Linear regressions were performed on the data from the SV5 samples to
determine whether certain trends existed. The SV5 data were chosen because they
contained more points than the FP5 data. For both SV5-1 and SV5-3, the measured
hydraulic conductivities appeared to decrease with time and with the number of tests.
The trends were strong for downward flow. Figure 4.7 shows the results graphically for
SV5-3 and downward flow. Each value of hydraulic conductivity is the result of an
independent test, and the times represent the starting times of the individual tests. The
trends are not easy to explain. Decomposition of the peat layer is unlikely to have had an
effect on pore sizes during the time taken to conduct a series of experiments. The peat
particles may have become more tightly packed by each successive downward flow test,
increasing flow resistance.
Table 4.6. Relationships Between Hydraulic Conductivity, Time, and Initial Head in the
Sample Flow Direction Relationship Equation R2
SV5-1 Up K (cm/s) vs. time (days) K = -3x10-6t + 3x10-4 0.67
SV5-1 Down K (cm/s) vs. time (days) K = -3x10-6t + 3x10-4 0.85
SV5-3 Up K (cm/s) vs. time (days) K = -4x10-7t + 4x10'5 0.20
SV5-3 Down K (cm/s) vs. time (days) K = -xl0-6 t + 4x10' 0.90
SV5-1 Down K (cm/s) vs. initial h2 (cm) K = 6x10-4 h + 0.1714 0.45
SV5-3 Down K (cm/s) vs. initial h2 (cm) K = 5x10-7 h 4x10'5 0.98
0 5 10 15
Starting Time (days)
Figure 4.7. Relationship Between Hydraulic Conductivity and Time for SV5-3 and
Table 4.6 contains the results of two regressions between hydraulic conductivity
and initial head in the reservoir tube. In theory, higher absolute heads may compress the
peat matrix, reducing hydraulic conductivity and causing a deviation from Darcian
behavior. The SV5 samples were chosen for the number of data points available, and
downward flow was chosen because the range of initial heads was greater for downward
than for upward flow. While the data for SV5-1 do not contain a strong trend, the
hydraulic conductivity measured for SV5-3 appears to display a strong positive linear
trend with increasing initial head in the sample tube. However, upon inspection of the
initial conditions in Table 4.4, the initial head in the reservoir tube was decreased for
each successive downward flow test on SV5-3. Therefore, the apparent trend is simply
the same one measured with increasing time and number of tests. There was no apparent
hydraulic conductivity trend with respect to initial head for the range of heads studied.
Because the range of heads present in the field is similar to the range studied in the
laboratory, the assumption of Darcian flow appears to be justified.
Water balance experiments carried out in the two-tube apparatus provide
information on the behavior of the unsaturated peat. For each equilibrium data point
desired, the water table was lowered a particular increment and then allowed to recover to
a constant level. Figure 4.8 shows the raw data obtained from an unsaturated experiment
on FP5-3. The time required to reach equilibrium is about a week when the equilibrium
point occurs within the peat layer. While the water table is in the sand but the capillary
fringe still reaches the peat-sand interface, the time required to reach equilibrium is about
3-5 days. The time required to reach equilibrium appears to decrease to approximately
one day after the capillary fringe falls below the peat-sand interface. Because the sand is
mostly or completely responsible for the release below this level, the release is much less
gradual than it is when the peat layer is releasing water.
| 100 -' ^ _-------------
0 20 40 60 80
Water Level - Interface Sand CF
Figure 4.8. Unsaturated Column Experiment Raw Data. The interface represents the
level at which the sand and peat layers meet. Sand CF denotes the approximate water
table elevation for which the capillary fringe is at the interface. The sample is FP5-3.
Figure 4.9 contains equilibrium data obtained from an unsaturated column
experiment on peat sample FP5-3. The 32 cm peat sample was underlain by a 95 cm
sand sample. The round data points represent the total volume released from both layers
as the water table dropped below the peat surface. The solid line indicates the calculated
relationship for the release from the sand layer, based on the van Genuchten expression,
as the water table drops. This relationship is based on separate column experiments
conducted with sand only. The square data points indicate the volume released from the
peat layer calculated as the difference between the total release and the sand release. This
calculation depends on the assumption that the releases from each layer can be
superposed to give the total release.
As the water table drops, the curves go through three distinct stages. The first is
an approximately linear portion accounted for almost entirely by release from the peat
0 20 40 60 80 100 120 140
WT Depth (cm)
Peat+Sand Sand -*-Peat - Interface
Figure 4.9. Volume Released by Unsaturated Sample FP5-3 as the Water Table Drops.
layer. The fact that this release begins when the water table is shallow suggests that large
void spaces exist within the peat. In order for any release to occur immediately as the
water table begins to drop through the peat profile, these void spaces must be large and
continuous enough to be virtually independent of capillary forces. The initial linear
portion extends to approximately 60 cm below the peat-sand interface, the minimum
depth at which the capillary fringe can reach the peat-sand interface. This depth
corresponds to the observed height of capillary rise of about 60 cm.
Between depths of about 90 and 100 cm below the top of the peat, both the peat
and sand layers are responsible for a portion of the release. The exact shape of the sand
release is inferred from separate column experiments. While its general nature is correct,
the change in slope may be sharper than indicated. A sudden emptying of the sand pores
in the upper layer may account for the abrupt total release at about 95 cm. Such an
abrupt release from the sand layer may also create a pore tension great enough to cause
an abrupt release from the peat layer.