£EPA
United States
Environmental Protection
Agency
Risk Reduction
Engineering Laboratory
Cincinnati OH 45268
EPA/600/2-90/025
June 1990
Research and Development
Relationship of
Laboratory- and Field-
Determined Hydraulic
Conductivity in
Compacted Clay Layer
-------
EPA/600/2-90/025
June 1990
RELATIONSHIP OF LABORATORY- AND FIELD-DETERMINED HYDRAULIC
CONDUCTIVITY IN COMPACTED CLAY LAYER
by
A. S. Rogowski
U.S. Department of Agriculture, ARS
Northeast Watershed Research Center
University Park, Pennsylvania 16802
Interagency Agreement No. DW-12930303
Project Officer
Walter E. Grube, Jr.
Waste Minimization, Destruction and Disposal Research Division
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
The information in the document has been Bunded wholly or in part by the
United States Environmental Protection Agency under assistance agreement
number DW-12930303 to the United States Department of Agriculture, ARS. It
has been subject to the Agency's peer and administrative review and has been
approved for publication as a U.S. EPA document. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
11
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial
products and practices frequently carry with them the increased generation of
materials that, if improperly dealt with, can threaten both public health and
the environment. The United States Environmental Protection Agency is charged
by Congress with protecting the Nation's land, air, and water resources.
Under a mandate of national environmental laws, the Agency strives to
formulate and implement actions leading to a compatible balance between human
activities and the ability of natural systems to support and nurture life.
These laws direct the U.S. EPA to perform research to define our environmental
problems, measure the impacts, and search for solutions.
The Risk Reduction Engineering Laboratory is responsible for planning,
implementing, and managing research, development, and demonstration programs
to provide an authoritative, defensible engineering basis in support of the
policies, programs, and regulations of the U.S. EPA with respect to drinking
water, wastewater, pesticides, toxic substances, solid and hazardous wastes,
and Superfund-related activities. This publication is one of the products of
that research and provides a vital communication link between the researcher
and the user community.
This report documents the spatial variability of hydraulic conductivity,
measured by both infiltration and seepage, throughout an area of clayey soil
compacted according to engineering specifications for landfill liners. The
data emphasize the need for clear design specifications and high quality
construction of such earthen barriers in waste management facilities. This
report will be useful to scientists, engineers, and regulatory staffs who are
concerned with the actual hydraulic performance of soil liners and soil covers
constructed to protect the Nation's ground water.
E. Timothy Oppelt
Director
Risk Reduction Engineering Laboratory
iii
-------
ABSTRACT • • >. •
A study was begun in 1983 to characterize the areal variation in hydrau-
lic conductivity of a compacted clay liner. A field-scale research facility
was constructed, consisting of a 30' x 75' area of clay soil compacted in
three layers to specifications commonly used in constructing clay liners. The
facility was fully instrumented to measure infiltration, drainage, and soil
properties at numerous data collection points. Preliminary studies were
initiated using sections of small barrels and larger caissons to verify the
performance of monitoring systems. Results from these preliminary (prototype)
studies have shown that any perforations of the compacted soil, such as wells,
or access tubing for detectors to monitor wetting front advance, may result in
preferential water movement by gravity down the walls of these perforations.
To avoid this situation in the field-scale facility, access tubes were placed
horizontally to accommodate the nuclear probes used to measure changes in clay
density and porosity. Underdrains were imbedded in the concrete support
structure to collect outflow, infiltration cylinders were installed to monitor
infiltration, and metal pedestals were placed on the clay surface to assess
swelling by measuring elevation changes. Quality control observations
collected during the construction showed that on the average water content and
dry density of the compacted clay were close to design specifications, but the
spatial variability in these values was large. Measured infiltration rates
and outflow rates obtained following ponding the field-scale facility were
poorly predicted by the prototype data from small barrels and larger caissons.
Initial data showed rapid breakthrough of percolate near the confining walls,
a feature that was also observed earlier in prototype studies. The extent of
clay liner integrity and observed travel times reflect the effectiveness of a
field-scale clay liner in preventing possible ground water contamination.
Proper evaluation of flux rates and their distribution in time and space is
necessary to characterize the system.
Flux values, computed from observed infiltration and outflow measurements
at 18M locations in a layer of compacted clay subsoil, were compared to effec-
tive flux values based on breakthrough time distributions for water and Br~
tracer over the same area. Results suggested that both water and tracer move
at similar rates, but considerably faster than expected, on the basis of the
outflow flux alone, and that only a small fraction of the total pore space is
involved in active transport. The ramifications of these findings are
explored against the background of effective porosity, degree of compaction,
and observed changes in bulk density with time.
The experimental clay liner was ponded for 1 year. During that time
inflow, outflow, and changes in density were routinely monitored at 250
locations. Results suggested that initial increase and final leveling off of
the density values could be associated with water passing into the clay matrix
and attainment of the steady state. Observed increases in density initially
were accompanied by increases in outflow which subsequently declined and
leveled off at twelve months. These changes were associated with leaching and
precipitation of Fe and Mn. Tracer breakthrough times were consistently
faster than water flow rates, although initial water breakthrough times
following ponding were similar to tracer breakthrough times. Results suggest
that water and solutes moved in the clay through only a small portion of total
porosity.
iv
-------
Hydraulic conductivity distribution based on laboratory measurements
underestimated field measured hydraulic conductivity distribution by a~factor
of five. However, comparisons of individual values at the same location could
differ by several orders of magnitude. Tt was found that the distribution of
water content and density in the compacted clay was adequately described by
core samples and nuclear surface moisture-density probe data. However, the
water content and density data appeared to have little relationship to average
values of spatially distributed hydraulic conductivity.
This report was submitted in fulfillment of Interagency Agreement
No. DW-12930303 between the United States Environmental Protection Agency and
the United States Department of Agriculture, Agricultural Research Service, k
This report covers a period from September 1983 to August 1988 and work was
completed as of August 1988.
-------
CONTENTS
Page
Foreword ill
Abstract . iv
Figures viii
Plates xiv
Tables xv i
Acknowledgment ........ xx
1. Introduction 1
Objectives and Approach 2
Literature Review (Phase I) 3
2. Conclusions 11
3. Recommendations 13
4. Preliminary Studies 14
Design Criteria 14
Clay Liner Materials 20
Barrel Studies 25
Caisson Prototypes 32
5. Field Scale Studies ty
Testing Facility 49
Soil Materials 54
Liner Construction 55
Experimental Procedures 56
Summary of Installation . 63
Site, Scale, and Spatial Relationships 66
Surface Moisture and Density 68
Experimental Results . 72
Preponded Stage '72
Density 72
Total Available Pore Space 73
Ponded Stage 77
Sampling Plan 77
Average Values ..... 78
Gradient 82
Density 82
Inflow and Outflow 92
Initial infiltration 95
Flow at a point 95
Tracer Studies 104
Water and tracer breakthrough ...... 107
Effective porosity 110
Macropore flow 112
Tracer distributions 115
Swelling 116
Percolate Quality 120
VI
-------
Page
6. Concluding Studies 131
Postponded Stage 131
Laboratory Hydraulic Conductivity 131
Flexible Membrane Method 131
Comparative Analysis 134
Distribution of Bulk Density and Water Content. . . 136
Veihmeier Tube Samples 137
Core Data 137
7. Data Quality 148
Data Acquisition 148
Analytical Procedures 152
Density 152
CSE 152
Leachate Drains 153
Infiltration Rings 153
Liner 153
Evaporation 153
Data Reduction, Validation and Reporting 154
Internal QC Checks 156
Final Density Check 157
Inflow/Outflow Balance 161
Statistical Stability Tests 161
Number of Samples ..... 165
Appendices
A. Literature. ..... 169
B. Individual Readings Disc
C. Statistics Disc
D. Tracers Disc
E. Laboratory Disc
Publications - Clay Liner 203
Vll
-------
FIGURES
Number Page
1.1 Schematic representation of a variogram Y(h) .......... 7
4.1 A platform for testing hydraulic properties of a 30 cm thick
clay liner .......................... 14
4.2 Standard proctor test on sieved material ............ 22
4.3 Standard proctor test on nonsieved material ........... 23
4.4 Moisture characteristic of the clay liner material ....... 24
4.5 Permeability and infiltration measurements: schematic diagram
of the barrel, infiltration ring, inner and outer constant
head devices ......................... 29
4.6 Infiltration rate x 10"? (m/s) in the inner ring of barrel
#\ plotted as a function of time (days) ............ 29
4.7 Cumulative infiltration in the inner ring of barrel #1
plotted as a function of time (days) ............. 30
4.8 Cumulative outflow from the inner bottom compartment
of the barrel #1 plotted as a function of time (days) ..... 30
4.9 Cumulative inflow in the inner ring (0.05 m^) of barrel #1 as a
function of the total pore volume below the infiltration ring
plotted as a function of time (days) ............. 31
4.10 Cumulative outflow as a function of the total pore volume abpve
the bottom inner compartment plotted as a function of time . . 31
4.11 Schematic diagram of the caisson study ............. 32
4.12 Swelling of clay liner given as percent of liner thickness and
plotted as a function of time ................. 37
4.13 Comparison between cumulative infiltration in barrel #1 and
caisson rings ......... . ............. . . 37
4.14 Typical water content with depth (access tube #1), caisson
was flooded 41 days after being compacted. ... ....... 38
4.15 Improved version of clay liner prototype ............ 41
4.16 Comparison of observed swelling on prototype caisson 1 and 2
studies ............................ 45
viii
-------
Number (Figures continues)
5.1 Clay liner testing facility, similar to Figure 4.1 but in SI
units showing additional detail on horizontal access tubes . . 49
5.2 Experimental measuring grid for (a) bulk density,
(b) infiltration and drainage, (c) swelling and
evaporation 50
5.3 Standard proctor compaction test on till material used in
the liner 55
5.4 Values of water content and dry density measured with the nuclear
surface probe for the sheepsfoot roller, and with Eley
volumeter, gravimetric samples for jackhammer and small
vibratory roller; solid lines represent laboratory measured
standard proctor compaction and saturation curves, dashed
vertical lines give optimum water content (17.8$) and dashed
horizontal lines represent 90$ of the maximum density
(1754 kg/m3) 64
5.5 Distribution of water content (by wt) and dry densities (DD)
on three compacted lifts measured with Troxler surface probe
in the backscatter mode, shown against the background of
proctor density at water content (opt), lines of maximums and
line of saturation (.zero voids) 65
5.6 Geometry and scale considerations of Troxler dual gamma probe,
Eley volumeter, and Troxler surface moisture-density probe
in backscatter mode 67
5.7 Distribution of water content in lifts #1, #2 and #3 of the
experimental clay liner based on small gravimetric grab
samples, optimum (OPT) moisture content is 17.6$ by weight . . 68
5.8 Distribution of water content in lifts #1, #2 and #3 of the
experimental clay liner based on nuclear surface moisture
gauge, optimum (OPT) moisture content was 17.8$ by weight. . . 69
5.9 Distribution of dry bulk density in lifts #1, #2 and #3 of
compacted clay liner in terms of standard proctor (PROCT)
density (1754 kg/m3, 110 PCF) 71
5*10 Average density distribution over the clay liner 73
5.11 An illustration of why the center portion of the clay liner
may have received more passes than the sides 74
•5.12 Contours of average density (in kg/m3) adjusted to surface
probe density values (a), and of surface probe moisture
content (in kg/m3) (b), before ponding 75
ix
-------
Number (Figures continued)
5.13 Interpolated isopleths of the amount of water needed to
saturate the liner (in kg/m3) at ponding ........... 76
5.11* Schematic representation of constant head Mariotte bottle
assemblage .......................... 78
5.15 Average wet density measured with dual gamma probe before and
after ponding ......................... 79
5.16 Average ring infiltration rate during a one year study period. . 80
5.17 Average outflow rate during a one year study period ....... 81
5.18 Average infiltration and outflow rates for a ponded liner as
a whole during the one year study period ........... 82
5.19 Schematic representation of gradient parameters for compacted
clay liner at Klingerstown, Pa ................ 83
5.20 Interpolated distribution of hydraulic gradient over the
study area .......................... 83
5.21 Interpolated contours of density two days (top), one week
(middle), and one month (bottom) after ponding ........ 86
5.22 Distribution of standard counts for the dual density probe
with time ........................... 87
5.23 Net change in density for consecutive readings at the site
M#H, I#M, G#U and J#U, plotted as functions of time,
dashed lines indicate possible extent of the error of
measurement (±8.7 kg/m3) ........ .... ....... 90
5.2U Cumulative change in density at the sites M#4, I#U, G#4 and
J#4 plotted as functions of time, each value ( •) has an
associated uncertainty of ± 8.7 kg/ra3 ............. 91
5.25 Spatial distribution of dual probe density values (kg/m3)
before ponding (a), and 2 days (b), 10 days (c), 1 month
(d), 3 months (e), 6 months (f), 9 months (g), and 1 yr
(h) after ponding ................. . ..... 92
5.26 Distribution of infiltration rings which showed major, minor
or slight leaks when fluorescein was added to each ring. ... 93
5.27 Illustration of Philip's (1957) method for calculation of
sorptivity S, and A-values . . ................ 96
x
-------
Number (Figures continued) Page
5.28 Ring inflow rate (R), drain outflow rate (D) and associated
bulk density distribution to the east (Dg) and west (Dy)
of the ring-drain, for the sites A3 (a), N8 (b) ........ 98
5.29 Ring inflow rate (R), drain outflow /ate (D) and associated
bulk density distribution to the east (Dg) and west (Dy)
of the ring-drain, for the sites H5 (a), FT (b) ........ 99
5.30 Comparison of hydraulic conductivity distributions computed
from (ring) inflow and (drain) outflow at 3 months and 6
months ............................ 102
5.31 Comparison of hydraulic conductivity distributions computed
from ring inflow and drain outflow at 9 months and 1 year
(13 months) ..... . .................... 103
5.32 Comparison of hydraulic conductivity distributions computed
from ring inflow and drain outflow averaged over time
(6/27/85 to 4/30/86) . . . . .................
5.33 Location of 15 infiltration rings and respective sampling
areas (shaded) to which 1M Br~ solution had been added .... 105
5.34 Logarithm (base 10) of relative (C/Co) breakthrough
concentrations (logRC) of Br" pulses in hours (sumtime)
at a particular drain (ID-B:AA6) and arrival times
after applying Br" tracer to central ring (ID-A:AA7) ..... 107
5.35 Relative concentration of Br" in leachate from AA7 and
surrounding drains (a), and from A3 and surrounding
drains (b) .......................... 108
5.36 Frequency histograms of breakthrough times (B-T time) for
water (a) and Br" tracer (b) ................. 109
5.37 Distribution of breakthrough (first arrival) time (days) for
water (a) and Br" tracer (b) in compacted, spatially
variable clay liner .................... .. 110
5.38 Distribution of effective porosity (Pe) in a compacted and
ponded, spatially variable clay liner given as the percent
of cross sectional -area (A), based on first arrival times
of ponded water ........................ 113
5.39 Distribution in time of ring inflow (R), drain outflow (D),
and bulk density in the east (Dg) and west (D^) of the
primary ring and drain locations for the slow-flowing site
F1 (a) and fast-flowing site G5 (b) .............. 114
xi
-------
Number (Figures continued) Page
5.40 Relative concentration of Br~ as leachate from site F1 and
surrounding drains (a) and site G5 and surrounding
drains (b) 115
5.41 Relative distribution of Br" tracer in soil around the sites
F1 (a) and G5 (b) to which tracer has been applied 117
5.42 Relative distribution of Br" tracer in soil around the sites
D7, L3, M7, and 01 to which the tracer has been applied. . . . 118
5.43 Relative distribution of Br~ tracer in soil around the sites
Q5, S1, V8 and T5 to which the tracer has been applied .... 119
5.41 Relative distribution of Br~ tracer in soil around the sites
JO and 09 to which the tracer has been applied 120
5.45 Elevation changes in millimeters (mm) for the compacted clay
liner 9 months after ponding; positive symbols indicate
swelling, negative symbols are indicative of shrinkage.
AAAA -2.50 to -2.00 mm; BBBB -2.00 to -1.50 mm; CCCC -1.50
to -1.00 mm; DDDD -1.00 to -0.50 mm; EEEE -0.50 to 0.00 mm;
FFFF 0.00 to 0.50 mm; GGGG 0.50 to 1.00 mm; HHHH 1.00 to
1.50 mm; IIII 1.50 to 2.00 mm; JJJJ 2.00 to 2.50 mm; KKKK
2.50 to 3.00 nun; LLLL 3.00 to 3.50 mm 121
5.46 Number of pore volumes leached through the clay liner at nine
months, starred points indicate locations where leachate
has been sampled monthly ..... 121
5.47 Spatial distribution of electrical conductivity (EC) in
vimhos/cm (a), pH (b), and SOij mg/J, (c) in leachate from
drains (250) at 9 months 125
5.48 Spatial distribution of electrical conductivity (EC) in
ymhos/cm (a), pH (b), and SOij mg/!l (c) in leachate from
drains (250) at 12 months 126
5.49 Spatial distribution of K (a) and Na in mg/H (b) in leachate
from drains at 9 months 127
5.50 Spatial distribution of K (a) and Na in mg/8, (b) in leachate
from drains at 12 months 128
5.51 Spatial distribution of Ca (a) and Mg in rag/i (b) in leachate
from drains (250) at 9 months 129
5.52 Spatial distribution of Ca .(a) and Mg in mg/A (b) in leachate
from drains (250) at 12 months 130
xii
-------
Number (Figures continued)
6.1 Chronological distribution of activities after the clay liner
was drained ..................... ..... 132
6.2 Point to point comparison of laboratory hydraulic conductivity
(K) with field derived values based on last (upper plots)
and average (lower plots) observed ring and drain flux .... 135
6.3 Rank to rank comparison of laboratory hydraulic conductivity
(K) with field observed values based on last (upper plots)
and average (lower plots) observed ring and drain flux .... 136
6.4 Comparison between dual probe readings on 8/1/86 just after
the clay liner was drained and on 8/18/86 just prior to
when 3" core samples were taken ................ 138
6.5 Spatial distribution of the 3" core/dual probe density ratio
on 8/18/86 in the compacted clay liner ............ 140
6.6 Contours of density (a) before ponding, (b) two days after
ponding and (c) just prior to being drained
6.7 Spatial distribution of the amount of water needed (in kg/m3)
to saturate the clay liner before ponding, 2 days after
ponding and just before being drained ............. 142
6.8 Changes in the degree of saturation at 3" (8cm) and water
content by volume at 6" (25cm) in the drained clay liner
with time ........................... 145
6.9 An east to west transect of wet (a) and dry (b) density in 1"
(2.5cm), depth increments as measured with the dual gamma
probe on 3" (91cm) centers .................. 145
6.10 An east to west transect of water contents by weight in 1"
(2.5cm) depth increments measured gravimetrically ....... 146
7.1 Gravimetric balance performance data .............. 151
7.2 Data processing flow chart ................... 154
7t3 Distribution of dual density probe standard counts during the
study period ......................... 165
7.4 Standard error of estimate as a function of the number of
observations for ring infiltrometers ............. 166
7.5 Relative values of the standard error as a function of the
number of observations for density, water content and
hydraulic conductivity measurements .............. 168
xiii
-------
PLATES
Number Page
4.1 Elevated platform construction: reinforced concrete footers . . 15
4.2 Elevated platform construction: finished platform 16
4.3 Installation of the lower access tube prior to pouring of
concrete floor on the platform 16
4.4 The 10m long access tubes were supported by brackets to keep
them level 17
4.5 Location of lower access tubes after concrete floor has been
poured; spacers show where upper tubes will go after the
installation of the clay liner 17
4.6 Measurement of density in horizontal access tubes with dual
gamma gauge 18
4.7 Distribution of infiltration rings on compacted clay liner, a
few pedestals to monitor swelling can be seen in the
background 18
4.8 Distribution of outflow ports equipped with moisture blocks
(hanging wires) to indicate early arrival of breakthrough
water 19
4.9 To test double ring infiltrometer geometry clay liner material
was compacted in barrels (half drums) 26
4.10 Clay liner material was compacted using a scaled up version of
Proctor compaction mold and drop hammer 27
4.11 An experimental set up to measure infiltration rate on
compacted clay liner material 28
4.12 A bank of vertical access tubes for measurement of moisture
and density with depth, and a ring for measuring
infiltration rate in flooded caisson 33
4.13 Additional details of optocator positioning and of the
optocator pedestal for measurement of clay swelling,
infiltratioin rings and access tubes shown in a
flooded caisson 34
4.14 Constant head devices (cylinders) for infiltration rings and
for the caisson itself (carboy) were placed on the outside
of the caisson 35
4.15 Prototype liner being installed in caisson #2 study using
mechanized compaction, neutron surface moisture/density
gauge was used to monitor compaction 42
xiv
-------
Number (Plates continued)
4.16 The prototype liner concrete box, just showing, burlap covered
bottom, gypsum blocks located over drains and removable wooden
wedges on the sides; the bottom horizontal access tube is
barely visible in the center; raised metal ridge separates
outer wall flow from flow in the inner compartment; constant
head devices to be used with infiltration rings are shown
along the walls 43
5.1 Infiltration rings, a pedestal and a wooden walkway supported
by access tubes were installed in compacted clay liner .... 51
5.2 The floor of the platform was sealed, and the bead of the
bentonite was placed 3' away from the sidewalls,
subsequently the floor was covered with burlap and a
thin layer of sand 52
5.3 Bentonite panels are being placed against sealed platform
sidewalls, to minimize wall effects 53
5.4 Test plot is being compacted using a sheepsfoot roller 56
5.5 Clay material was brought in by trucks and spread on the
platform with a large backhoe 57
5,6 After the required amount of water was added, it was
incorporated into the clay by rototilling 58
5.7 Following several passes with the bulldozer and sheepsfoot
roller, each lift was smoothed out by large vibratory
roller prior to measurement of density 59
5.8 Near the walls clay was compacted using a "nervous turtle"
roller (shown above) and jackhammer (not shown) 60
5.9 Water content and density were measured using a nuclear
moisture-density probe ... 61
5.10 Prior to installation of the next lift the clay surface was
roughened by bulldozer treads 62
xv
-------
TABLES
Number
4.1 Average properties of the B-horizon of Hubblersburg
cherty silt loam (Typic Hapludult, illitic or mixed
mesic) developed on limestone, from PSU GP-10 file 21
4.2 Average surface and depth moisture and density readings. ... 35
4.3 Results of gravimetric moisture sampling of the compacted
clay liner in caisson 1 40
4.4 Electrical resistivity of gypsum blocks buried by respective
drains (1 to 6) under the compacted clay ponded (time 0)
on 10/9/84 46
4.5 Wet and dry bulk density (WD and DD) of compacted clay
before ponding, measured with nuclear probes and
gravimetrically 47
4.6 Water content measured gravimetrically and with nuclear
probes on compacted clay prior to ponding 47
4.7 Wet bulk density (WD) of compacted clay as measured with
horizontal access tubes and computed values of total
available pore space (taps) before (0) and following
(1, 3, 14 days) ponding 48
4.8 Outflow rates measured in inner (#5 and #6) and outer (#1, #2,
#3 and #4) compartment drains 48
5.1 Selected properties of the till used as liner material as
determined by supplier, our (NWRC), and EPA (HWERL) soil
testing laboratories • 54
5.2 Clod size analysis of two large samples. 55
5.3 Cross sectional areas and volumes associated with monitoring
of water flow, water content, and bulk density in the
compacted clay liner ...... 66
5.4 Distribution of water content (gravimetric) in clay liner
computed from nuclear gauge data and point grab samples. . . 70
5.5 Dry bulk density of the three lifts of compacted clay liner
as a function of standard proctor compaction test 71
5.6 Statistics of 13 bulk density (wet) data sets taken
continuously and consecutively following Installation
on 11/1/84 and prior to ponding on 3/26/85 72
xvl
-------
Number (Tables continued) Page
5.7 Particle density of ground fraction of clay liner material . . 72
5.8 Average values of dry bulk density and water content
by weight , 84
5.9 Selected statistics for the density distributions before
and after the clay liner was ponded 85
5.10 Individual values of bulk density obtained using dual gamma
, probe on 9/25/85 (6 months after ponding) 88
5.11 Summary of statistics for 9/25/85 bulk density data (6 months
after ponding) ..... 89
5.12 Average infiltration rates for selected times uncorrected
(all) and corrected for ring area (area correction) and
leaking rings (leak rings omit) 94
5.13 Sorptivity, A-values, and saturated hydraulic conductivity
based on initial (1000 sees) infiltration rates for
.selected double rings (613 m2 area infiltrometers using
Philip's (1957) method) 97
5.14 Selected statistics for the hydraulic conductivity
distributions in time computed from nontransformed
: (a) and log-transformed (b) ring infiltration data
for the ponded clay liner 100
•5.15 Selected statistics for the hydraulic conductivity
distributions in time computed from nontransformed
(a) and log-transformed (b) drain outflow data for
the ponded clay liner 101
5.16 Amounts of 1M Br"1 diluted to 2000mil and applied to rings
in 1, 2 or 3 (used) increments of 2000mX,; applied
concentration (Co) in ppm and grams (g) 106
5.17 Laboratory values of hydraulic conductivity compared with
average field values of ring inflow and drain outflow 111
5.18 Breakthrough times (T^) for Br" tracer and water, and
cumulative tracer concentration (C/CO), recovered and
computed effective flux density (Qe) values for selected
sites based on tracer breakthrough times and average 30
cm thickness of clay 112
5.19 Water quality changes as a result of passing through
experimental liner 122
xvii
-------
Number (Tables continued)
5.20 Means and standard deviations, 'arid coefficient of variation
(CV) of leachate quality parameters sampled at 9 and 12
months after- ponding 123
5.21 Correlation matrix of leachate ^quality parameters a"t 9 and 12
months following poridtng 124
6.1 Summary statistics for laboratory and field evaluated hydraulic
conductivity (x 10~9 m/sec) 133
6.2 Field values of hydraulic conductivity near the sites where
defective cores were taken . 134
6.3 Statistics of the 3" core data compared with final dual
probe (8/18) values 139
6.4 Statistics of the wet density values obtained using 3" cores
and Troxler surface moisture density probe during construction
(initial) and after the liner was drained (final) 139
6.5 Comparative statistics for wet density values (kg/m2) obtained
using 3" cores, 2" cores, 6" holes, excavations and K-cores. . 139
6.6 Statistics of the ratio: (3" core density/dual probe
density) 140
6.7 Comparative statistics for the amount of water needed to
saturate the clay liner 143
6.8 Average swelling of clay 143
6.9 Statistics for the 2" cores, average and by layer 144
6.10 Comparative statistics for pesthole data 147
7.1 Methods used and classical data quality indicators for
infiltration, evaporation and leachate rates, bulk
density, and change of surface elevation (CSE) 150
7.2 Percent error as a function of volume of leachate 150
7.3 Density data selected for QC/QA validation 158
7.4 Example of density data record 159
7.5 Grouping summary 160
7.6 Summary density flagged validation values 160
7.7 Mass balance of inflow and outflow of clay liner water 162
xviii
-------
Number (Tables continued) Page
7.8 Statistical stability and drift test - surface moisture gauge. . 163
7.9 Personnel check, inflow measurements 16M
7.10 Comparison of the relative values of standard error for density,
moisture, and hydraulic conductivity measurements. . 167
xix
-------
ACKNOWLEDGMENTS
Special thanks and recognition go to the USDA-ARS personnel involved in
different aspects of this project, E. L. Jacoby, Jr., D. E. Simmons, W. M.
Yazujian, J. E. Donley, B. J. Chamberlin, C. W. Artz, P. J. Dockey and
students D. E. Gedon and R. M. Petery, T. Knerr and F. K. Reeser. Mr.
Yazujian's, Mr. Jacoby's, and Mr. Simmons' contributons to the success of this
project were particularly outstanding. The author acknowledges the support of
this study by the Land Pollution Control Division, Hazardous Waste Engineering
Research Laboratory, U. S. Environmental Protection Agency, Cincinnati, Ohio,
through Interagency Agreement No. DW129-303-03-01-0 with the Northeast
Watershed Research Center, Agricultural Research Service, U. S. Department of
Agriculture, University Park, Pennsylvania; D. Walter E. Grube, Jr. is the U.
S. EPA Project Officer. This report has not been subjected to the EPA review
and therefore the contents do not necessarily reflect the views of the Agency
and no official endorsement should be inferred.
XX
-------
SECTION 1
INTRODUCTION
Installation of liners at sanitary landfills and hazardous waste sites
has been one of the commonly recommended methods of containment and control.
The primary function of a liner is to prevent, or limit the amount of leachate
that might ultimately reach the groundwater. Thus, liners must have appropri-
ate properties to restrict, delay or dilute leachate migration so that a given
site may ideally provide an ultimate containment. Desirable properties of
linear materials include (1) low permeability, (2) high adsorption capacity,
and (3) resistance to chemical, biological and mechanical breakdown. These
desirable attributes work to enhance several physical and chemical processes
which will lessen the environmental consequences of contaminant migration.
The primary mitigating factors include dilution, time delay, and retardation.
The choice of liner material should be based on the extent to which these
factors operate to meet the desired performance criteria.
Various treated and untreated soil mixtures have been utilized as liners
for waste sites: compacted mixed clay soils, pure montmorillonite, montmoril-
lonite mixed with concrete, bentonite with an added polymer, as well as other
commercially available products. Of these, the clay liners are the most
common. Typically, clay liners are constructed in one of the following ways:
(1) A commercial refined clay product is added to and mixed with
the top few centimeters of native soil.
(2) A nearby deposit of clay soil is excavated, hauled to, and
compacted in place at the disposal site.
(3) A native clay soil is compacted in place.
Generally, a liner is covered with a layer of sand or soil to minimize
drying, and waste is placed on top of the soil varying in depth from 2 to 8
meters. For municipal waste the density varies between 297-53^ kg/m3, for
many hazardous waste materials it is likely to be higher. Taking 1000 kg/m3
as a likely number and adding to it the weight of cover cap, overburden
pressures on the liner might be on the order of 5 to 10 T/m^. Since good
construction begins with adequate specifications, proper design is the single
most important factor in a successful liner installation. Design procedures
need to take into account the Federal and State regulations and performance
standards. While State requirements must meet the minimum Federal standards,
many go beyond to accommodate local conditions. For example, the Pennsylvania
Department of Environmental Resources (DER) requires the following specifica-
tions be met if a native clay is to be used as a primary layer.
-------
Permeability: < 10"? cm/sec
Thickness : 60 cm
Compaction : 95% standard proctor
Clay content: greater or equal to 25%
In addition to the liners being compatible with the waste, the site must be
inspected and approved by the Department. Quality testing is mandatory during
and upon completion of the installation.
The least water content (by weight) at which the soil just flows under
its own weight is called the liquid limit (LL), while the smallest water
content (also by weight) at which the soil can still be rolled out into thin
(- 3mm) threads without crumbling is called the plastic limit (PL), the
difference between the tow (LL-PL) is taken as the plasticity index (PI) and
is given in units of water content by weight. In general, liners should be
constructed of inorganic clays with a liquid limit greater or equal to 30 and
plasticity index greater or equal to 15, and with their respective cation
exchange characteristics carefully considered. The liners should be compacted
wet of optimum water content using a sheepsfoot roller, and their clod size
should also be controlled. Frequently inadequate installation results in clay
liner failure. Thus, a strict quality control of materials, equipment and
workmanship appears essential to ensure compliance with the engineering
specifications.
Assuming a constant head of ponded water above a compacted clay liner the
flow rate through the liner should be proportioned to its hydraulic
conductivity. As far back as 1856 Darcy established experimentally that the
volume of water (Q) flowing through a unit cross sectional area of a sandy
formation in a unit of time, was proportional to the difference in hydraulic
head (Ah) between the top and bottom of the formation. Since the head at the
bottom of the formation of thickness L is - zero, Ah is numerically equal to
the height of water ponded above the bottom, Ah/L is referred to as the
dimensionless gradient (i), and K the proportionality coefficient with units
of length/time, is the hydraulic conductivity. Combining the above we get an
expression Q = Ki, known as Darcy's Law.
OBJECTIVES AND APPROACH
The purpose of this study was to evaluate hydraulic conductivity of a
field scale clay liner and to compare the field observed values of the
hydraulic conductivity with the laboratory determined values of the hydraulic
conductivity. Supplemental data were also to be gathered to support research
into factors potentially affecting leaching of the contaminants into the
groundwater.
The specific objectives of this study were:
(1) To document the present state-of-the-art for determining the in situ
hydraulic conductivity of compacted clay soils.
(2) To construct a field-scale test plot composed of recompacted clay
soil, to install appropriate monitoring devices for measuring water
-------
flow rate in compacted clay, and to determine the in situ hydraulic
conductivity using selected permeants, verifying that accurate data
can be obtained.
(3) To compare the hydraulic conductivity values obtained on a field
test site with those obtained in the laboratory on core samples,
and if significant lack of agreement exists, to evaluate the
factors responsible.
The project was to proceed in three discrete phases. The initial phase
(Phase I) was to include a literature review and methods evaluation to recom-
mend the most appropriate field scale procedure for measuring the hydraulic
conductivity in a compacted clay soil liner. Particular emphasis was to be
placed on available methods addressing cohesive soils, undisturbed sites, and
large surface areas, as well as those that could provide data within a reason-
able time frame, remembering that hydraulic conductivity of the clay liner was
likely to be - 1 x 10~9 m/sec. The output from Phase I was to be a literature
review report, with recommended approaches to construction of the field-scale
hydraulic conductivity measuring facilities. In Phase II the field apparatus
for measuring flow and the prototype clay liner were to be designed and tested
for accuracy and performance. In Phase III field-scale facility was to be
constructed and clay liner compacted and ponded according to standard industry
methods and procedures. After the completion of the field phase, core samples
were to be removed from the site for laboratory determinations of hydraulic
conductivity. This phase was to include long-term data collection, analysis
of field obtained results, and comparison with the laboratory values of deter-
mined hydraulic conductivity.
LITERATURE REVIEW (PHASE I)
It is generally assumed that the denser the clay liner the lower its
permeability. When constructing clay liners for containment of water the soil
material at a certain water content is compacted to a prespecified density.
The amount of compaction will vary with the water content. For high values of
water content a test involving a standard number of blows will eliminate the
few air filled pores present. Resulting density will be low and near to the
saturation density value. If, however, the soil is quite dry, no amount of
compaction will substantially reduce the air filled porosity because there is
not enough water to provide lubrication and allow soil particles to pack
closer together. Thus, the final dry density will also be low. At some water
content in between these extremes there is a point, known as optimum water
content, when a compaction test such as Standard Proctor test will result in a
maximum dry density for a given amount of effort (i.e., number of blows,
height of fall, mass of weight) (Cooper and Cassie, 1978).
The use of Standard Proctor test as a compaction measure assumes the same
aggregate size distribution for lab and field materials. This need not be
necessarily so and large differences in water content between the lab and
field can occur (Cox, 1978). Further modifications will take place as clay
adjusts to overburden stresses imposed on it by a combined load of waste and
cover cap. Thus, the equilibrium water content in the field may be different
from Standard Proctor test "optimum," and may also vary sufficiently over the
-------
area to cause differences in observed flow. Although Cox (1978) investigated
the behavior of clays under extreme wet or dry conditions, his work shed
little light on clays compacted at or above the optimum water content. The
optimum water content occurs at water contents less than saturation, with only
part of the void space occupied by water and the rest by air. It is not known
whether under the conditions of optimum moisture and maximum density infiltrat-
ing water moves through the still empty, connected air voids or by displace-
ment through water-filled pores.
According to Jumikis (1965) volume of solids, volume of water, and volume
of air, stabilize at the optimum water content and maximum dry density and
stay constant with total void space at the minimum. Just above the optimum a
10$ increase in the relative volume of water may be accompanied by as much as
5Q% decrease in the air voids with little change in total « ^%) volume of
voids. Considering results presented by Gary et al., (19^3) and accompanying
discussion by Kellog and Creager (Gary et al., 19^3) large decrease in perme-
ability can occur over the same range (just above optimum), suggesting that
primary pathways of water movement may be through relatively large air filled
and connected pores, while water in smaller pores of the clay matrix is more
or less immobilized. This question has not been specifically addressed in
literature except indirectly by Gary et al., (19^3), Anderson and Low (1958),
Lambe (1955) and Mitchell et al., (1965). It needs to be clarified whether
the pores through which water moves following compaction are the ones which
are predominantly air filled, or the ones which are predominantly water filled.
Gary et al., (19^3) state that optimum compaction means realignment of clay
platelets, while Lambe (1955) and Cary et al., (19^3) suggest that there is
little or no subsequent swelling as water is applied to the surface of remold-
ed clay. Thus, the primary water conducting pores could well be the larger
pores which may initially be filled with air. In either case, the flow would
be taking place only through a fraction of the total void space and in effect
constitute an unsaturated flow regime. Some support for this view comes from
the work of Anderson and Low (1958) who suggest that the structure of adsorbed
water may also be different (i.e., less dense) from that of the free water.
The concept of field capacity (Burrows and Kirkham, 1958) in agricultural
soils, or nature of flow through coarse mine spoil (Rogowski and Weinrich,
1981) carries a similar connotation: only certain pores conduct water.
Recent attempts to measure large pores in field soils (Clothier and White,
1981) may be a likely approach to clay liner permeability evaluation.
Generally, a quantitative knowledge of clay liner properties is required
for prediction of hydrologic behavior at the hazardous waste sites. The
choice of an approach may be dictated by the magnitude of spatial variability
and the distribution of hydrological properties. Which properties should be
measured, what sample volumes should be taken, what locations and what sam-
pling frequency should be considered, are some of the aspects which need to be
resolved in characterizing a given clay liner. Of prime importance in this
context are the objectives and the desired accuracy for which hydrologic
predictions are needed; this will influence the level of sophistication and
detail at which a site is sampled and the data analyzed.
One source of information to evaluate the suitability of a soil as clay
liner material is soil classification data. Soil classification is based on
the premise that soil properties vary in space. Soil surveys are then used to
-------
identify and delineate the soil boundaries and predict extent and properties
of individual horizons. But in most classification schemes these boundaries
remain imprecisely defined. The soil survey classification is based on the
broad morphological features of the landscape correlated to sampled profile
properties such as color, horizon, depth, structure, and texture.
Unfortunately, the extent and nature of variability within a soil unit and
associated mapping purity are not always recorded. As a first approximation,
properties of soil series may be used to select clay liner material. However,
it must be realized that the criteria used in the classifying soils may not
coincide with those affecting hydrologic response. Furthermore, appreciable
spatial variability in soil hydrological properties has been observed within a
soil series (e.g., Rogowski, 1972; Nielsen et al., 1973; Sharma et al., 1980),
and this may affect the areal response of remolded material. Thus, under most
conditions field characterization of an in place liner is considered important
and for this, field-oriented methods are needed. These methods should be
simple, rapid, and reliable so that a large number of measurements can be made.
For the ease of handling and analyzing of data, grid or transect sampling
schemes are preferred.
The choice of hydrological properties needed to be determined and the
extent of detail of their characterization, depend largely on objectives and
the choice of model to be used. For detailed prediction of water distribution
within a clay liner, a physically-based deterministic model such as a
three-dimensional water flow equation for swelling soils could be used. This
however, would require detailed knowledge of the spatial distribution of the
soil water retention ij;(0) as a function of water ratio
-------
determination of approximate K(ijj) and 6(ip) values for evaluation of spatial
variability and estimation of redistribution of applied water in a field soil.
Because K values in the compacted clay liner are expected to be very low (-
10~9 m/sec), it is doubtful to what extent this approach could be used.
Considerable effort has been made in developing empirical methods to esti-
mate hydrological properties based on particle size analysis data. Broad
scale hydrological classification of agricultural soils of the USA is being
attempted by estimating parameters appropriate to the Brooks and Corey model
of water retention (e.g., Rawls et al., 1982) as well as parameters suitable
for use in the Green and Ampt (1911) infiltration equation (Swartzendruber,
1987). In the absence of soil hydrologic data, these approaches are likely to
have a wide appeal since they are usually based on readily available
information. The limitations of such approaches to clay liner construction
should however be realized, particularly for soils with predominant structural
features.
In general, soil systems, such as a remolded clay liner, are extremely
complicated and highly variable at a scale of individual aggregate, but such
complexity can be bypassed by measuring hydrological properties at a larger
scale. Usually the variance of a property decreases with an increase in
volume of a sample. The smallest volume above which the variance no longer
decreases significantly defines the representative volume for that property
(Bear, 1972). This is a theoretical concept, and in real world situations, it
may be difficult to define. Ideally, the representative volume should encom-
pass components of variability at several scales, in practice, however, hydro-
logic properties are measured on much smaller samples and they are considered
points of a continuum.
For comparative purposes and assuming normality the magnitude of variabil-
ity is sometimes represented by the coefficient of variation (CV). In soils
coefficient of variation is found to be highest (CV > 1.0) for transport
coefficients (hydraulic conductivity and diffusivity), medium (CV = 0.15-0.5)
for properties such as water contents at selected water potential, and least
(CV < 0.15) for properties such as bulk density and total porosity (Warrick
and Nielsen, 1980). The variability of a property can also be described by a
cumulative distribution function (CDF), which (if we assume a Gaussian model)
contains information about the mean and other moments, and these permit estima-
tion of confidence intervals for the property. Assuming a CDF for a parameter
has important implications in computing the number of observations required to
estimate the mean with a specified precision (Rogowski, 1972; Sharma, 1983),
and in determining the integrated hydrologic response of an area (Sharma and
Luxmoore, 1979; Warrick and Amoozegar-Fard, 1979).
In traditional parametric analysis of variability a Gaussian or normal
model is usually adopted. Properties measured within an area are assumed
spatially independent of one another and observations are represented by their
mean, standard deviation and other moments. The assumption of spatial
independence, at least for points close by, seldom holds for natural systems.
Geostatistical techniques (e.g., Journel and Huijbregts, 1978) and nonpara-
metric methods can then be employed to evaluate the degree of interdependence.
The spatial dependence of neighboring observations of a property Z measured at
-------
all points x and x+h as functions of the distance vector h can be expressed by
the seraivariogram Y(h),
N
Y(h) = ^
- Z(Xi+h)]2
(1.1)
where N is the number of pairs [Z(xi), Z(x^+h)] for a particular distance (or
time) increment h. Thus, a semivariogram of a property describes the average
rate of change of Y(h) with distance and shows the variance structure of
observations. These observations may or may not follow any type of CDF, but
should be additive (Journel and Huijbregts, 1978). An an idealized semivario-
gram is shown in Figure 1 . The sill is an upper limit of a variogram model
where it levels off, while the distance at which the variogram model begins to
level off is known as the range in most applications. The sill usually
approaches a priori sample variance, while the range defines a neighborhood
where the variable is continuous. With increasing separation distance h, Y
may increase and approach a constant value (sill) beyond some separation
distance a (range). This indicates the extent of spatial dependence. If Y
continues to increase with increasing h, information at larger separation
distances may be required. Often Y does not pass through the origin and at
Spatial
Variance
Y (h)
Random
Variance
Variance = Sill
Nugget Effect
I
a.
Distance (h) Between Samples
Figure 1.1. Schematic representation of a variogram Y(h).
7
-------
h=0 has some positive finite value called a nugget effect. Such a nugget
effect usually suggests a spatial structure on a scale smaller than the sam-
pling interval h. Once the spatial structure of the variable is defined by a
semivariogram, the values can be kriged to interpolate the distribution of a
property over an area or in space.
Extensive reviews and discussion of laboratory and field techniques of
measuring permeability and related parameters are readily available in mono-
graphs (Black, 1965; Hagan et al.f 1967; USGS, 1977; Nielsen et al., 1972),
books (Bear et al., 1968; Childs, 1968; Baver et al., 1972; Kirkham and
Powers, 1972; Hillel, 1971 and 1980), journals (Bianchi and Haskell, 1970;
Ritchie et al., 1972; Alemi et al., 1976; Clothier and White, 1981). Most
writings are concerned with permeability measurements in agricultural soils,
and the requirement that a flow rate be less than 10~9 m/sec, as specified for
clay liners, is of little practical interest. We have therefore, to turn to
engineering literature (Olson and Daniel, 1979), particularly literature on
clay cores in earth dams and embankments as well as to novel approaches in
soil mechanics (Mitchell et al., 1965; Lee, 1974; Zimmie and Riggs, 1981), and
to experimental and theoretical studies of shrinking and swelling systems
(Kemper, 1960, 196la,b; Sposito, I975a,b,c) to get some idea what may be
involved. These studies add new insights in a practical sense, to what has
been written before by Gary et al., (1943), Lambe (1955), and Babcock (1963).
Swelling clay soils do not behave in the same manner as do nonswelling
materials. Swelling soils, in addition to matric and gravitational potential
flows, will be subject to overburden potential and probably varying degree of
chemical potential depending on waste compatibility. This, of course, compli-
cates the situation considerably. The effect of gravity on flow will be less
and vertical moisture profiles are likely to vary depending on clay thickness,
amount of swelling, and location with respect to the water table. In shallow
profiles volume of water per unit volume of soil may increase, decrease, or
remain constant with depth, while, for deep profiles volume of water per unit
volume of soil will decrease. These developments may markedly affect profile
permeability. It is well known (Philip, 1969a,b) that surface macrotopography
does have a considerable influence on the equilibrium water content, but it is
not known to what extent microtopography of an essentially level clay liner
will affect its hydrology. Much work has been done recently to predict the
behavior of swelling soils in response to water using the fluid mechanics
approach (Philip, 1969a,b; Youngs and Towner, 1970; Philip, 1970; Talsma,
1974, 1977a,b; Talsma and Flint, 1958; Talsma and Lelij, 1976) or thermodynam-
ics approach (Babcock, 1963; Sposito, 1972, 1973, 1975a,b,c; Sposito et al.,
1976; Chu and Sposito, 1980. Work done by Kemper and coworkers (i.e., Kemper
and Evans, 1963; Kemper and Rollins, 1966, etc.) addresses movement of water
across clay membranes as a function of concentration gradients. When salt
solution is forced through a compacted clay a portion of electrolyte is
excluded from water films surrounding clay particles and remains behind at the
high pressure side increasing the concentration of salts and possibly affect-
ing permeability. To what extent work on swelling systems is applicable to
our studies remains to be seen.
/
When discussing aspects of structural stability Quirk (1978) pointed out
that in a soil containing illitic clay, the dominant pore peak occurred at
-------
3.5 nm (10~9 m). However, volume changes during wetting usually lead to forma-
tion of planes of discontinuity and development of pores two to three orders
of magnitude larger than the original peak. Recent work of Smalley (1978)
suggests that compaction at moistures slightly greater than optimum leads to
structural rearrangement of clay particles. Particle realignment on wetting
and compaction in Leda/Champlain clays of eastern Canada was shown by Smalley
(1978) to result in considerable local variations in density. Although ini-
tial open structure of these clays tended to be preserved by cementation of
short range bonds, bond breakage could subsequently result in complete struc-
tural collapse.
Changes in hydraulic conductivity can be brought about by chemical and
physical reactions within the clay matrix. Frenkel and Rhoades (1978) pointed
out that hydraulic conductivity can increase appreciably because of
dispersion, provided bulk density of the material and electrolyte level of
percolate are sufficiently low, flow rates are sufficiently high, and exchange-
able sodium is between 10 and 20%. Under these conditions dispersion begins
to affect mechanics of transport and may lead to piping failures. On the
other hand, if sufficient clay is present and exchangeable sodium is greater
than 25%, swelling will most likely reduce hydraulic conductivity. At low
electrolyte levels in the percolate, if the bulk density and clay content are
sufficiently high, dispersion will lead to blocking of pores and reduction in
hydraulic conductivity.
Thus, a system, such as a clay liner, may react in different ways and
require not only the specifications depending on location, clay type, amount,
and water content, but also the knowledge of the particular waste chemistry.
A basic experimental tool in field infiltration research is the ring
infiltrometer. There are many different types and sizes that have been used
and modifications range from a double ring infiltrometer on one end of the
scale to an enclosed air entry permeameter and double tube permeameter at the
other. Infiltration rate I within the rings is assumed to take place under
unit gradient conditions and calculations may involve fitting the S
(sorptivity) and A (coefficients) in Philip (1967) infiltration equation,
I = St1/2 + At (1 .2)
where t is the time. Straight line plot of I/t1/2 with respect to t^/2 gives
S as a y-axis intercept and coefficient A as a slope of the line. A working
estimate of hydraulic conductivity (Ks) can then be obtained as,
Ks = 3A (1.3)
This relatively rapid method has been used by Talsma and Lelij (1976) to
evaluate in situ infiltration rate and water movement on a swelling rice paddy
soil during prolonged ponding. Measured values of hydraulic conductivity
ranged from 5.8 x 10"? to 1.12 x 10~5 cm/sec, certainly within the range of
interest for clay liner studies. Based on their average sorptivity of - 25
mm/day1/2 and average Ks value of - 2 mm/day (2.3 x 10~5 cm/sec) the plot of
-------
I/t1/2 vs t1 /2 was expected to be linear for about 3 days according to the
relation,
t £ 0.02 (S/K)2 (1.4)
It might therefore be anticipated that a linear behavior of I/t1/2 vs t1/2 for
a clay liner with K - 10~7 cm/sec is likely to be even longer. If a ring is
large enough (> 30 cm diameter), or enclosed in a manner similar to an air
entry parameter, or a falling head infiltrometer (Daniel and Trautwein, 1986)
measurable infiltration rates can be readily calculated from the fall of water
level in a capillary, or by weight.
Following ponding, and after the water drains a disc infiltrometer
(Clothier and While, 1981) can be used to check for the presence of large
pores. Disc infiltrometer will allow water movement into the profile under a
known negative head and is used together with ponded values to see what propor-
tion of the larger pores are likely to be conducting water. In agricultural
soils as much as 80 or 90% of flow can take place through the larger pores.
The "larger" pores in clay matrix are likely to be much smaller necessitating
the use of different membrane materials with higher impedance to induce higher
values of negative head.
10
-------
SECTION 2
CONCLUSIONS
The initial literature Review ('Phase I) has¥aised several questions.
The foremost among these was the evidence "that -flow''through compacted clay may
be very sensitive to the remolding water content and accompanying uncertainty
about the representativeness-and: effectiveness of standard co'mpaction
techniques. Equally important was the potential'Existence and distribution of
preferential flow pathways.
•,- '" •'*•:.'•*' < ,;,'',' . •' ; '•!-, ; '" " • '
Preliminary studies (Phase I!) proved useful' in 'desi'gn of the''field
experiment and instrumentation. ' Resultsj suggested 'poterit'ial for leakage along
the walls of containment and sides of instruments.^ Some Swelling was also
observed; however, swelling infiltration and drainage rates from1;prototype
studies were"poor predictors of fleld'scale facility behavior. "!
The large 10 x 25m, bridgelike• "platform'facility an'd a cbttipacted 0.3m
thick clay line (Phase 3) provided satisfactory data about the performance of
clay liners and relationship of field hydraulic conductivity to lab values.
Clay liner was constructed from a B-horizon of a typical soil meeting the
EPA specifications. In the course of analysis the soil was found to contain a
larger than expected number of coarse fragments. However, no evidence was
found of preferential layering, or preferential distribution of these rock
fragments either because of natural tendencies or mechanized compaction of the
three lifts. Sand layer on top of the clay liner acted as a moisture barrier
and prevented rapid drying. Spatial measurements of evaporation needed to be
made to correct infiltration flux and for mass balance purposes.
The considerable variability observed in the spatially distributed
infiltration, outflow and outflow chemistry on the compacted clay liner may
have been affected by the presence of preferential pathways of leachate flow
through the clay liner.
Higher flow rates which originated in apparently unsaturated areas
suggested the presence of the preferential flow pathways. Such pathways will
potentially pose a grave threat to underlying ground water quality even in the
presence of a clay liner.
There was no correlation between laboratory and field derived values of
hydraulic conductivity on the point to point basis when values from the same
locations were compared. However, when the distribution of laboratory values
was compared with the distribution of field values the results appeared to be
linearly correlated. Laboratory distribution underestimated the field distri-
bution by approximately a factor of 5, despite the fact that individual obser-
vations varied as much as four orders of magnitude.
11
-------
Laboratory hydraulic conductivity values were not a good indicator of the
clay liner behavior. Ring infiltrometers, despite problems, appeared to pro-
vide better estimates of potential outflow from below the compacted clay.
Unfortunately, using of the ring infiltrometers is time consuming. Best esti-
mates of clay liner performance were obtained by following the conservative
tracer (Br~) movement, and breakthrough history.
We have found little change in clay liner wet density with time suggest-
ing a very limited movement of water into the clay matrix. Initial change,
however, just after flooding could be indicative of the extent of
macroporosity.
Finally, we have found a surface moisture-density probe with capability
for direct transmission from shallow depths to be a quick and satisfactory
method of determining field distribution of moisture and density for individ-
ual lifts during construction. Unfortunately, there appeared to be no
relationship between density and water content of the clay and the observed
values of hydraulic conductivity and the flow regime, as perceived by changes
in wet density, did not appear to be continuous but consisted of concurrent
alternating, filling and draining episodes distributed in space.
The amount of available pore space in well compacted clay was very small,
and even a small change may be disproportionately large. In our case minimal
swelling of - 2.4-mm could have accounted for - 20% increase in available pore
space in compacted clay.
Far fewer (1/10 as many) samples were needed to characterize the compact-
ed clay liner density and water content compared with the number of samples
needed to characterize hydraulic conductivity with the same degree of
precision. Considering that hydraulic conductivity per se did not appear to
be the primary controlling factor in the flow and breakthrough of water and
tracers in the compacted clay liner more effort is needed to characterize
potential distribution of the preferential flow pathways.
12
-------
SECTION 3
RECOMMENDATIONS
Field constructed clay liners should have a layer of sand on the top and
the bottom, because sand acts as a capillary barrier for clay water, minimiz-
ing drying and outflow components.
Specifications for liner construction should include an upper permissible
level of clods, aggregates and stones to be contained in the candidate
materials. Clay liner materials are often derived from deeper horizons of
ordinary field soils. Such materials, particularly in the northeastern USA,
contain many coarse fragments.
It is recommended that the integrity of compacted clay liners be tested
using a conservative tracer, ponded conditions and an underdrain catchment
system. Because, despite adequate moisture and compaction, the moisture and
density readings did not appear to be correlated on a field scale with
observed hydraulic conductivity or the distribution of critically important
preferential flow paths. However, the nuclear surface moisture-density probe
with a shallow direct transmission capacity appeared adequate for evaluating
distribution of moisture and density within individual compacted lifts of clay.
In designing clay liner studies particular attention needs to be given to
potential flow along the walls of the containment facility or instrumentation
access ports.
Judicious selection of liner materials (specific guidelines) and exten-
sive quality control of moisture and density (specific guidelines) during con-
struction are recommended. At present many samples of hydraulic conductivity
are needed to adequately characterize the potential flux and transport through
the clay. Because far fewer samples of water content and density are required
to characterize the clay liner compared with the number of samples required to
characterize hydraulic conductivity with the same degree of precision,
increased quality control and use of homogeneous materials at a reasonably
constant water content may help decrease the number of samples needed to
characterize hydraulic conductivity.
13
-------
SECTION 4
PRELIMINARY STUDIES (PHASE II)
DESIGN CRITERIA
Design criteria and plans for the liner testing facility and associated
instrumentation included a 10 x 25m elevated platform and a 30cm thick liner.
Figure 4.1 and Plates 4.1 and 4.2 show the schematic plan and different
construction stages of the platform.
Plate 4.1 shows the reinforced concrete footers, and in Plate 4.2 we show
the finished (b) elevated platform. The open ramps on either side were subse-
quently filled with compacted soil and gravel to provide a drive on access to
the platform for liner construction and removal. Plates 4.3 and 4.4 give
details of drain locations and reinforcing grid. They also show the method of
installation (4.3) and support (4.4) for lower access tubes. In Plate 4.5
lower access tubes protrude above the just poured concrete floor with spacers
for the upper access tubes in the sidewall in the background. Plate 4.6
illustrates how density was to be measured after the clay was compacted and
upper access tubes installed. The gamma source was to be placed in the lower
Figure 4.1. A platform for testing hydraulic properties of a 30cm thick
clay liner.
14
-------
access tube and the detector, connected to the scalar (foreground), was to be
placed in the upper access tube. Density changes were to be computed from
routine measurements at the same locations. Additional instrumentation was to
include infiltration rings (Plate 4.7) and drain ports (Plate 4.8) equipped
with moisture blocks to detect early arrival of breakthrough front (note wires
hanging from drains), square 10 x 10cm pedestals (several can be seen in Plate
4.7) to monitor swelling, and evaporation pans (same size as infiltration
rings) to correct infiltration and outflow for evaporation. The instrumenta-
tion at the site was to be installed so as not to interfere with the struc-
tural integrity of the compacted liner. The clay material was to be trucked
in from an actual commercial facility. Installation and compaction were to be
carried out according to the USEPA standards (i.e., USEPA, 1988) and also were
to be quality tested during installation. A special effort was to be made to
take many more moisture samples and make many more measurements of bulk densi-
ty than may strictly be necessary, so as to establish the minimum sampling and
testing criteria needed for adequate quality control.
Plate 4.1. Elevated platform construction: reinforced concrete footers.
15
-------
Plate 4.2. Elevated platform construction: finished platform.
16
-------
Plate ^.3. Installation of the lower access tube prior to pouring
of concrete floor on the platform.
16a
-------
Plate 4.4. The 10m long access tubes were supported by brackets to
keep them level.
17
-------
Plate M.5. Location of lower access tubes after concrete floor has
been poured; spacers show where upper tubes will go after
the installation of the clay liner.
17 a
-------
Plate 4.6. Measurement of density in horizontal access tubes
with dual gamma gauge.
18
-------
Plate H.Tt Distribution of infiltration rings on compacted clay liner, a few
pedestals to monitor swelling can be seen in the background.
18 a
-------
Plate 4.8. Distribution of outflow ports equipped with moisture
blocks (hanging wires) to indicate early arrival of
breakthrough water. ' .
It was anticipated that 30cm diameter infiltration rings (Plate 4.7) and
complementary network of outflow drains (Plates 4.3 and 4.8) will provide
sufficiently rapid response to measure infiltration, compute hydraulic
conductivity and estimate the distribution of macroporosity for a clay liner.
Because inflow rates were anticipated to be slow, control of evaporation and
maintaining of relatively constant temperature was thought to be important.
Many more measurements were to be taken than are necessary, so that enough
samples were available to estimate the minimum number of samples required to
characterize variability. The design called for closing off both ramp ends of
the liner testing platform and flooding the whole area. The final output of
the study was to be a grid of inflow and outflow measurements, calculated
conductivities and changes in bulk density and porosity over the area with
depth, and with time. If areas of high permeability occurred a 1 x 1 m2 grid
of drains was to intercept the leachate.
Shrinkage and swelling of the liner system were to be monitored using
square 10 x 10cm pedestals and a laser beam technique. Evaporation was to be
measured with a class A pan and a number of smaller pans the same size as
infiltration rings.
Following the ponding stage the clay liner was to be core sampled and
hydraulic conductivities of cores, evaluated in the laboratory, were to be
compared with those obtained from field ring and drain measurements and with
overall flow rates through the liner.
19
-------
Clay Liner Materials
The clay liner material consisted of commercially available B-horizon
subsoil (10" to 60", 25 to 250cm) of Hubblei
the 3 to 8% slopes north of Roaring Springs
Pennsylvania. Typically, the subsoil exten<
•sburg cherty silt loam found on
and Highway 36 and 164 in central
is to a depth of 60" (150cm) or
more. It is yellowish red, friable cherty isilty clay loam, silty clay, and
cherty silty clay to a depth of 35" (90cm). At a depth of more than 35"
(90cm) it is yellowish red, friable silty clay loam.
Normally included in this soil (5 to 1
Mertz, Clarksburg, Morrison, and Wharton Va
natural state is considered to be moderate
)$) are small areas of Opequon,
•iant soils. Permeability in
and available water capacity is
high. The soil is very strongly acid throughout. The unified classification
of the subsoil ranges from CL to CH (10" to
less than 5% of fragments > 3" (76mm), with
60-95$ passing #10, 55-95$ passing #40, and
ranges from 35 to 55% and plasticity index
ranges from 20 to 45/5, bulk density from 12DO to 1600 kg/m3 and in place
permeability from 6 to 2 inches per hour (15 to 50mm, 40 to 140 x 10"? m/sec).
Available water is 12 to 16? by volume with
60", 25 to 150cm). It contains
35 to 100$ passing sieve #4,
55-85$ passing #200. Liquid limit
from 12 to 30$. Clay (< 2mm)
pH within 4.5 to 5.5 range.
Shrink swell potential is moderate and erosion potential is quite large. In
general, organic matter is low (1 to 3$). Potential for frost action is
moderate and risk of corrosion for steel and concrete is high to moderate,
respectively. The soil is classified as clayey, illitic, mesic Typic
Hapludult (USDA, 1981), Table 4.1 shows average properties of the B-horizon.
The site where the soil came from was
own analyses have shown it to be a CL type
Liquid Limit:
Plastic Limit:
Plasticity Index:
mined commercially for clay.
clay with,
water by weight
water by weight
water by weight
Our
which compares reasonably well with values given by supplier. The material
appears to contain quartz, K-feldspar kaolinite, illite and vermiculite. It
is 47$ clay, 45$ silt, and 8$ sand as giver
The Standard Proctor test results and 95$ c
by the sedigraph measurements.
ionfidence intervals results are
shown for sieved and unsieved material in Figures 4.2 and 4.3. The optimum
water content (by weight) was 18$ by weight for both, and the maximum density
is 111.5 pcf (1786 kg/m3) for sieved and 114.5 pcf (1834 kg/m3) for nonsieved
material, respectively. Comparison with the moisture characteristic curve of
the sieved material (Figure 4.4) shows that the optimum water content (at
least in loose soil) occurred between 2 and 3 bars tension. What this means
is that when initially water is ponded on the surface of a clay liner a very
high water content gradient will be imposed across the clay-water interface.
20
-------
TABLE 4.1. AVERAGE PROPERTIES OF THE B-HORIZON OF HUBBLERSBURG CHERTY SILT
LOAM (TYPIC HAPLUDULT, ILLITIC OR MIXED MESIC) DEVELOPED ON
LIMESTONE. FROM PSU GP-10 FILE
Property
Total coarse fragments
Coarse fragments
Less then 2 mm
Bulk density (clod)
Bulk density (< 2 mm)
Cole
Uncor 1 /3 bar core
15 bar, fragment
pH water: soil, field
pH KCL:soil, field
Organic carbon (titratin)
Calcium
Magnesium
Sodium
Potassium
Total bases
Aluminum
CEC:Ex acidity
Base saturation
Fe203
Kaolinite
Illite
Vermiculite
Chloride
Int.
Talc
Unit
% by wt
% by vol
% by wt
kg/m3
kg/m3
cm/ cm
% by wt
% by wt
-
-
% by wt
% by wt
% by wt
% by wt
% by wt
% by wt
% by wt
me/100 g
% by wt
% by wt
% by wt
% by wt
% by wt
% by wt
% by wt
% by wt
Mean Value
14.0
9.5
86.0
1510.9
1488.1
0.017
25.0
17.7
5.1
4.2
0.16
2.1
0.9
0.1
0.2
3.4
4.8
9.9
25.5
5.1
24
52
8
5
5
10
Standard
deviation
10.7
11 .0
10.7
132.0
125.3
0.009
5.4
3.4
0.4
0.4
0.19
1.9
0.7
0.0
0.1
2.2
2.1
2.6
15.7
1.0
18
18
3
-
-
6
21
-------
I 15.0
12.5
10.0
107.5
H
55
HI
o
105.0
102.5
100.0
97.5
95.0
-i 1
SIEVED
12 14 16 18 20 22
MOISTURE (PERCENT)
24 26
Figure U.2. Standard Proctor test on sieved material.
22
-------
CW
c
CD
BULK DENSITY PCF
<» r
|