EPA/600/2-37/050
July 1937
VERIFICATION OF THE HYDROLOGIC EVALUATION
OF LANDFILL PERFORMANCE (HELP) MODEL
USING FIELD DATA
by
P. R. Schroeder and R. L. Peyton
U.S. Army Engineer Waterways Experiment Station
Vicksburg, MS 39180-0631
Interagency Agreement Number DW96930236-01-1
Project Officer
D, C. Amnion
Landfill Pollution Control Division
Hazardous Waste Engineering Research Laboratory
Cincinnati, OH 45268
HAZARDOUS WASTE ENGINEERING RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
-------
NOTICE
This report was prepared by P. R. Schroeder and R. L. Peyton of the
U.S. Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, under
Interagency Agreement DW96930236-01-1. The U.S. Environmental Protection
Agency Project Officer was D. C. Amnion of the Hazardous Waste Engineering
Research Laboratory, Cincinnati, Ohio.
The research described in this document has been funded wholly or in part
by the United States Environmental Protection Agency, It has been subjected
to Peer and Administrative Review and has been approved for publication as an
EPA document. Mention of trade names or commercial products does not constitute
an endorsement or recommendation for use.
ii
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial prod-
ucts and practices frequently carry with them the increased generation of
solid and hazardous wastes. These materials, if improperly dealt with, can
threaten both public health and the environment. Abandoned waste sites and
accidental releases of toxic and hazardous substances to the environment also
have important environmental and public health implications. The Hazardous
Waste Engineering Research Laboratory assists in providing an authoritative
and defensible engineering basis for assessing and solving these problems.
Its products support the policies, programs and regulations of the Environ-
mental Protection Agency, the permitting and other responsibilities of State
and local governments and the needs of both large and small businesses in
handling their wastes responsibly and economically.
This report describes a study conducted to verify the Hydrologic Evalua-
tion of Landfill Performance (HELP) computer model using existing field data
from a total of 20 landfill cells at 7 sites in the United States, Simula-
tions using the HELP model were run to compare the predicted water balance
with the measured water balance. Comparisons were made for runoff, evapo-
transpiration, lateral drainage to collection systems and percolation through
liners. The report also presents a sensitivity analysis of the HELP model
input parameters.
Thomas R. Hauser, Ph.D.
Director, Hazardous Waste Engineering
Research Laboratory
iii
-------
ABSTRACT
Simulations of 20 landfill cells from seven sites were performed using
the Hydrologic Evaluation of Landfill Performance (HELP) model. Results were
compared with field data to verify the model and to Identify shortcomings.
Sites were located in California, Kentucky, New York, and Wisconsin and
included a wide variety of climates and landfill designs. Landfill descrip-
tions and soil properties were loosely defined, requiring much judgment in
selecting model input values and allowing significant variance in the simula-
tion results. The field measurements of the various water budget components
varied greatly from cell to cell despite some having identical designs. Con-
sequently, the precision of the verification effort is fairly low, but the
study demonstrates that the HELP model is a useful tool for realistically
estimating landfill water budgets. Simulation results generally fell within
the range of field observations. The results indicated that two modifications
in the HELP model may be warranted. Specifically, the computation of daily
temperatures for estimating snowmelt and the estimation of unsaturated
hydraulic conductivity for vertical drainage should be changed. Further study
is needed for verification of lateral drainage and percolation when the infil-
tration rate is small.
A sensitivity analysis of the HELP model was performed to examine the
effects of the major design parameters on components of the water budget for
landfills. Hydraulic conductivity values for the topsoil, lateral drainage
layers, and clay liners are the most important parameters in determining the
water budget components. These parameters are particularly important in esti-
mating the percolation through the landfill. Other design parameters tend to
affect the apportionment among runoff, evapotranspiration, and lateral drain-
age from the cover. This information, along with the verification results,
was used to evaluate RCRA landfill design guidance and regulation.
This report was submitted in partial fulfillment of Interagency Agreement
DW96930236-01-1 between the U.S. Environmental Protection Agency and the
U.S. Army Engineer Waterways Experiment Station. This report covers a period
from October 1984 to September 1986, and work was completed as of Septem-
ber 1986.
iv
-------
CONTENTS
Foreword ......... iii
Abstract ........ iv
Figures . vii
Tables . xiii
Acknowledgments xv
1. Executive Summary ............ 1
Purpose and scope 1
Field verification 2
Sensitivity analysis 6
Review of technical guidance ... 7
2. Introduction 8
Background .......... 8
Scope and purpose .................... 9
3. Model Description ......... ........ 11
Runoff and infiltration 11
Evapotranspiration 13
Vertical moisture flow 15
Percolation . 18
Lateral flow submodel 19
4. Sensitivity Analysis ..... 22
Landfill cover 23
Lateral drainage and barrier soil percolation ...... 44
5. Review of Landfill Design Regulation and Guidance . . 52
Regulation 52
Guidance ..... ..... 53
Evaluation and recommendations 54
6. Simulation of University of Wisconsin-Madison
Lysimeter Cells 56
Site description 56
Selection of model input values 56
Results of model simulations . 60
7. Simulation of Sonoma County Test Cells 86
Site description 86
Selection of model input values ............. 88
Results of model simulations 89
8. Simulation of Boone County Test Cell ..... Ill
Site description Ill
Selection of model input values ............. 113
Results of model simulations 113
v
-------
9. Simulation of Three County Landfills in Wisconsin 119
Brown County landfill 119
Eau Claire County landfill 127
Marathon County landfill ... ...... 131
10. Simulation of Chemical Waste Disposal Facility in
Niagara Falls, NY..., 143
Site description 143
Selection of model input values 145
Results of model simulations .... 146
11. Evaluation of Simulations ............. 154
Evaluation of field data .......... 154
Evaluation of model predictions 155
12. Summary and Conclusions ............ 159
References 162
vi
-------
FIGURES
Number Page
1 Vertical flow submodel for landfill cover . , . . . . 16
2 Configuration of lateral drainage model , ...... 20
3 Cover designs for sensitivity analysis ... .... 25
4 Bar graph for hazardous waste cover design
showing effect of surface vegetation, topsoil
type, and location 28
5 Bar graph for municipal cover design showing
effect of topsoil depth, surface vegetation,
and location 29
6 Effect of runoff curve number on hazardous
waste cover design 33
7 Effect of runoff curve number on municipal
cover design 34
8 Effect of evaporative depth on hazardous
waste cover design 35
9 Effect of evaporative depth on
municipal cover design ....... ... 36
10 Effect of drainable porosity on hazardous
waste cover design 37
11 Effect of drainable porosity on
municipal cover design . 38
12 Effect of plant available water on
hazardous waste cover design 39
13 Effect of plant available water on
municipal cover design ...... 40
14 Effect of and on lateral drainage
and barrier soil percolation .... ..... 47
vii
-------
Number Page
15 Effect of K^/Kp racio on lateral drainage
and barrier soil percolation , 48
16 Effect of L/a on the steady-state and
maximum average head above the liner 51
17 Effect of L/a on percolation 51
18 Cell dimensions for University of Wisconsin-
Madison cell 58
19 Field measurements for Cell 1 compared to
HELP simulation for covered cells;
cumulative comparisons . 62
20 Field measurements for Cell 2 compared to
HELP simulation for covered cells;
cumulative comparisons . 63
21 Field measurements for Cell 3 compared to
HELP simulation for covered cells;
cumulative comparisons 64
22 Field measurements for Cell 4 compared to
HELP simulation for uncovered cells;
cumulative comparisons 65
23 Field measurements for Cell 5 compared to
HELP simulation for uncovered cells;
cumulative comparisons .................... 66
24 Field measurements for Cell 6 compared to
HELP simulation for uncovered cells;
cumulative comparisons 67
25 Field measurements for Cell 7 compared to
HELP simulation for uncovered cells;
cumulative comparisons . 68
26 Field measurements for Cell 8 compared to
HELP simulation for covered cells;
cumulative comparisons 69
27 Field measurements for Cells 1, 2, 3, and 8
compared to HELP simulation for covered cells 70
28 Field measurements for Cells 4, 5, 6, and 7
compared to HELP simulation for uncovered cells 71
viii
-------
Number Page
29 Field measurements for Cell 1 compared to HELP
simulation for covered cells; monthly comparisons 72
30 Field measurements for Cell 2 compared to HELP
simulation for covered cells; monthly comparisons 73
31 Field measurements for Cell 3 compared to HELP
simulation for covered cells; monthly comparisons 74
32 Field measurements for Cell 4 compared to HELP
simulat x on s for uncovered cells^ month 1 y compa r xs on s • • « • . 73
33 Field measurements for Cell 5 compared to HELP
simulations for uncovered cells; monthly comparisons ..... 76
34 Field measurements for Cell 6 compared to HELP
simulations for uncovered cells; monthly comparisons ..... 77
35 Field measurements for Cell 7 compared to HELP
simulations for uncovered cells; monthly comparisons ..... 78
36 Field measurements for Cell 8 compared to HELP
simulations for uncovered cells; monthly comparisons ..... 79
37 Field measurements for Cell 8 compared to HELP
simulation for covered cells using CN=92 . . 83
38 Field measurements for Cells 4,5,6 and 7 compared
to HELP simulation for uncovered cells using
CN-86 85
39 Cell dimensions for Sonoma County cells ...... 87
40 Field measurements for Cells A, B, and E 92
41 Field measurements for Cell A compared to HELP
simulation; cumulative comparisons . . 93
42 Field measurements for Cell B compared to HELP
simulation; cumulative comparisons ..... .... 94
43 Field measurements for Cell E compared to HELP
simulation; cumulative comparisons .............. 95
44 Field measurements for Cell A compared to HELP
simulation; monthly comparisons ......... 96
45 Field measurements for Cell B compared to HELP
simulation; monthly comparisons 97
ix
-------
Number Page
46 Field measurements for Cell E compared to HELP
simulation; monthly comparisons .......... 98
47 Field measurements for Cells A and E compared to
HELP simulation using default soil texture 15
for topsoil ..... 101
48 Field measurements for Cells A and E compared to
HELP simulation using a barrier soil hydraulic
conductivity of 0.000094 in./hr ..... 102
49 Field measurements for Cells A and E compared to
HELP simulation using a runoff curve number of 60 103
50 Field measurements for Cells A and E compared to
HELP simulation using 25% compaction for topsoil ....... 104
51 Field measurement of leachate drainage for Cell C
compared to HELP simulation; cumulative comparison 106
52 Field measurement of leachate drainage for Cell D
compared to HELP simulation; cumulative comparison 107
53 Field measurement of leachate drainage for Cell C
compared to HELP simulation; monthly comparison . 108
54 Field measurement of leachate drainage for Cell D
compared to HELP simulation; monthly comparison . 109
55 Cell dimensions for Boone County cell 112
56 Field measurements for Boone County cell compared to
HELP simulation; cumulative comparisons .116
57 Field measurements for Boone County cell compared to
HELP simulation; monthly comparisons ....... 117
58 Field measurements of Boone County cell compared to
HELP simulation using hydraulic conductivity
of 7.0 x 10 cm/sec for topsoil 118
59 Cell dimensions for Brown County landfill 120
60 Field measurement of leachate drainage for
Brown County landfill compared to HELP
simulation for daily cover; cumulative comparison ....... 123
x
-------
Number
Page
61 Field measurement of leachate drainage for
Brown County landfill compared to HELP
simulation for daily cover; monthly comparison . 124
62 Field measurement of leachate drainage for Brown
County landfill compared to HELP simulation for
final cover; cumulative comparison L25
63 Field measurement of leachate drainage for Brown
County landfill compared to HELP simulation for
final cover; monthly comparison ....... 126
64 Cell dimension for Eau Claire County landfill 129
65 Field measurement of leachate drainage for Eau Claire
County landfill compared to HELP simulation for
daily sand cover; cumulative comparison 132
66 Field measurement of leachate drainage for Eau Claire
County landfill compared to HELP simulation for
daily sand cover; monthly comparison ..... 133
67 Field measurement of leachate drainage for Eau Claire
County landfill compared to HELP simulation for
uncompacted clay sludge cover; cumulative comparison ..... 134
68 Field measurement of leachate drainage for Eau Claire
County landfill compared to HELP simulation for
uncompacted clay sludge cover; monthly comparison . . 135
69 Field measurement of leachate drainage for Eau Claire
County landfill compared to HELP simulation for
clayey loam sludge cover; cumulative comparison . . 136
70 Field measurement of leachate drainage for Eau Claire
County landfill compared to HELP simulation for
clayey loam sludge cover; monthly comparison .... 137
71 Field measurement of leachate drainage for Eau Claire
County landfill compared to HELP simulation for
final cover; cumulative comparison . 138
72 Field measurement of leachate drainage for Eau Claire
County landfill compared to HELP simulation for
final cover; monthly comparison ..... 139
73 Cell dimensions for Niagara Falls cell 144
xi
-------
Number Page
74 Field measurement of leachate drainage for Niagara
Falls Cell 1 compared to HELP simulation;
cumulative comparison 148
75 Field measurement of leachate drainage for Niagara
Falls Cell 1 compared to HELP simulation;
monthly comparison 149
76 Field measurements of leachate drainage for Niagara
Falls Cells 2 and 3 compared to HELP simulation;
cumulative comparisons 151
77 Field measurements of leachate drainage for Niagara
Falls Cell 2 compared to HELP simulation;
monthly comparison 152
78 Field measurements of leachate drainage for Niagara
Falls Cell 3 compared to HELP simulation;
monthly comparison 153
xii
-------
TABLES
Kunber Page
1 Parameters Selected for Sensitivity Analysis 22
2 Climatological Regimes 24
3 Sensitivity of Water Budget Components to Location, Vegetation
and Topsoil Type and Depth 26
4 Sensitivity of Water Budget Components to Evaporative Depth and
Curve Number 31
5 Sensitivity of Water Budget Components to Drainable Porosity and
Plant Available Water 32
6 Sensitivity of Liner/Drain System Performance to Hydraulic
Conductivity of Lateral Drainage Layer and
Barrier Soil Layer , 45
7 Sensitivity of Liner/Drain System Performance to Lateral Drainage
Slope and Length 49
8 University of Wisconsin-Madison Cell Characteristics ....... 57
9 Input Data for Simulation of University of Wisconsin-
Madison Cells * 59
10 Difference between Cumulative HELP Model Predictions
and Cumulative Field Measurements for Covered Cells at
University of Wisconsin-Madison , , 80
11 Difference between Cumulative HELP Model
Predictions and Cumulative Field Measurements
for Uncovered Cells at University of Wisconsin-Madison ..... 81
12 Sonoma County Cell Operational Designs . 88
13 Input Data for Simulation of Sonoma County Cells 90
14 Differences Between Cumulative HELP Model
Predictions and Cumulative Field Measurements
for Sonoma County Cells 99
15 Input Data for Simulation of Boone County Cell . . , , » 114
xiii
-------
Number Page
16 Input Data for Simulation of Brown County Landfill 122
17 Difference Between Cumulative HELP Model
Predictions and Cumulative Field Measurements
for Wisconsin County Landfills 127
18 Input Data for Simulation of Eau Claire
County Landfill ..... ........ 130
19 Input Data for Simulation of Marathon
County Landfill . ...... 141
20 Comparison of HELP Simulations to Field
Measurements for Marathon County Landfill ..... 142
21 Input Data for Niagara Falls Landfill Simulation 147
xiv
-------
ACKNOWLEDGMENTS
The authors would like to express their sincere appreciation to
Ms. Anita Zitta, Ms. Charlotte Harness, Ms, Kathy Smart, Mr. Thomas E.
Schaefer, Jr. and Mr. Anthony Gibson of the Environmental Engineering
Division (EED) , U. S. Army Engineer Waterways Station (WES), for their many
contributions to the completion of this study. In addition, the authors
would like to acknowledge the support and general supervision of
Mr. F. Douglas Shields and Dr. Raymond L. Montgomery of the EED, WES. The
authors would also like to thank Dr. Robert Havis and Dr. Bruce McEnroe of
the EED, WES, for their technical review and the Publications and Graphics
Arts Division, WES, for their preparation of the final figures and their
editorial review.
xv
-------
SECTION 1
EXECUTIVE SUMMARY
PURPOSE AND SCOPE
This study was performed to verify Version 1 of the Hydrologic Evaluation
of Landfill Performance (HELP) model using existing field data. Mathematical
simulations of 20 landfill cells at seven sites across the United States were
made using the HELP model and were compared to measured field data. Measure-
ments of leachate drainage were available from all 20 landfill cells, while
data on runoff were available only from about half of the cells. Measurements
of percolation were available only from one cell and no data on evapotranspi-
ration were available. These landfills included a wide variety of conditions
for which the HELP model was tested. The cells ranged in size from 0.04 to 24
acres and the simulation periods ranged from 2.5 to 8 years. This report sum-
marizes the results of these simulations and evaluates the verification that
has been achieved. In addition, the report presents a sensitivity analysis
for the input parameters used in the HELP model and a review of landfill
design regulation and guidance in light of the results of the verification
studies and sensitivity analysis.
The HELP model was developed to help hazardous waste landfill designers
and evaluators estimate the magnitudes of components of the water budget and
the height of water-saturated soil above barrier soil layers (liners). This
quasi-two-dimensional, deterministic computer-based water budget model per-
forms a sequential daily analysis to determine runoff, evapotranspiration,
percolation, and lateral drainage for the landfill (cap, waste cell, leachate
collection system, and liner) and obtain estimates of daily, monthly, and
annual water budgets (1,2). The model does not account for lateral inflow or
surface runon.
The HELP model computes runoff by the Soil Conservation Service (SCS)
runoff curve number method (3). Percolation is computed by Darcy1s law, modi-
fied for unsaturated flow (1). Lateral flow is computed by a linearization of
the Boussinesq equation, and evapotranspiration is determined by a method
developed by Ritchie (4). The vertical percolation and evapotranspiration
components of the HELP model originated with the Chemical, Runoff, and Erosion
from Agricultural Management Systems (CREAMS) model (5). The development pre-
sented here, however, reflects a significant advance beyond the CREAMS model
in both of these areas.
In the course of the development of any model, provisions should be taken
to verify that the model accurately represents reality. Laboratory tests have
been performed to verify the lateral drainage portion of the HELP model, but
1
-------
prior to this study the other portions of the model had only been calibrated
and not verified. This report presents the results of efforts to verify the
model using existing data collected at landfills, test cells and lysimeters.
This study consists of three parts. The main part is an attempt to ver-
ify the model using existing data. The purpose of this part was to assess the
adequacy of the model to simulate reality and to validate the use of the model
to evaluate designs (generally with minimum information). The second part of
this study is a sensitivity analysis of the principal input parameters used in
the HELP model. The purpose of this part was to determine which parameters
need to be well-defined and what are the likely effects of a change in the
value of a parameter. This analysis also provided much insight for interpret-
ing and explaining the verification results. The third part of the study con-
sists of an evaluation of the technical guidance developed to support the
regulations regarding design and operation of a landfill. The purpose of this
part was to examine the results of the laboratory and field verification stud-
ies and the sensitivity analysis to determine whether the technical guidance
was practicable and achieved its objectives (particularly that of minimizing
potential percolation through the liner at the base of the landfill) in the
best practicable manner.
FIELD VERIFICATION
Comparisons were performed between simulations and measured field data
for seven sites:
(L) University of Wisconsin-Madison. From 1970 to 1977, eight
large lysioeter cells filled with either shredded or unprocessed refuse
were monitored for surface runoff and leachate production. The general
purpose of the study was to determine the effect on landfill performance
of shredding the refuse prior to placement and covering the refuse with a
soil layer.
(2) Sonoma County, CA. A solid waste stabilization project was
sponsored by the U.S. Environmental Protection Agency and Sonoma County,
California from 1971 to 1974. The purpose of the project was to investi-
gate the stabilization of solid waste in five municipal sanitary landfill
test cells by analyzing leachate, gas, temperature, and settlement param-
eters and to determine the effect on solid waste stabilization of apply-
ing excess water, septic tank pumpings, and recycled leachate. Leachate
production was measured for all five cells and runoff was measured from
three of the cells.
(3) Boone County, KY. Two field-scale test cells and three small-
scale cells were studied from 1971 to 1980 in Boone County, Kentucky
under the sponsorship of the U. S. Environmental Protection Agency. The
study objectives were to evaluate the amount and characteristics of
leachate, the composition of gases, the temperature conditions, the set-
tlement of the cells, the clay liner efficiency, and to compare the
behavior between the field-scale and small-scale cells. The data col-
lected from one of the field-scale cells was used in this study.
2
-------
(4-6) Brown, Eau Claire, and Marathon Counties» WI. The State of
Wisconsin Bureau of Solid Waste Management has reported on the geologic
setting, major design features, construction experience, leachate produc-
tion and operational performance of these three large landfills in Wis-
consin. These landfills were started between 1976 and 1980; however,
none of these landfills has yet been completely filled and capped so that
each data set reported thus far represents the conditions of a continu-
ously expanding landfill. For example, at any given time, the cover
could range from the daily cover of a 6-inch-thlck blanket of sand or
silty clay to a final cover of clay and topsoil,
(7) Niagara Falls, NY. Since 1976, a chemical waste management
company has filled and capped three landfill cells in Niagara Falls, NY.
The surface areas of the cells range from 2 to 5 acres. Records of
leachate pumpage have been kept from 1983 and indicate annual withdrawals
ranging from 1 to 11 inches. An evaluation of the performance of the
facility during 1984 was reported to the USEPA Region II by Recra
Research, Inc.
The data used in the simulations were obtained from a variety of sources.
In most cases, daily rainfall and monthly temperature data were obtained from
the nearest National Oceanic and Atmospheric Administration weather station.
Solar radiation values stored in the HELP model were used for all simulations.
Model input values for design data and soil and waste characteristics were
determined from published reports describing the construction and operation of
each landfill. In general, the information available on soil and waste char-
acteristics, surface vegetation, runoff curve numbers, or evaporative depths
was descriptive and sketchy, not quantitative; therefore, extensive use was
made of default values stored in the HELP model.
The measured data used for comparison with the HELP model simulations
were primarily lateral leachate drainage volumes. Measured runoff data was
available from 11 landfill cells. Barrier soil percolation was measured at
one landfill, although its suitability for model verification was limited.
The measured data from very similar cells at the same landfill varied greatly
from cell to cell. For four practically identical cells the range in total
runoff was about 50 percent of the mean total runoff, and for lateral drainage
the range was greater than 100 percent of the mean.
Runoff
Measured runoff data existed for eight cells at the University of Wiscon-
sin and for three cells at Sonoma County, CA. Runoff was overpredicted for
five cells by an average of 30 percent of the measured runoff, and underpre-
dicted for six cells by an average of 20 percent of the measured runoff.
Overall, runoff was overpredicted by 3 percent. Following these initial simu-
lations, the curve numbers were varied to determine their effect on the over-
all model prediction of landfill performance. Five simulations were improved
by a change in curve number—all had originally underpredicted runoff.
3
-------
For the three cells at Sonoma County, it was obvious that the evapotran-
spiration and/or soil characteristics were controlling runoff volume and not
the curve number. Because of this close interaction, it was difficult to
assess the accuracy of the curve number method in the HELP model based on the
field data in this report. However, the predicted runoff volumes appear over-
all to be in reasonable agreement with the measured results.
A comparison of measured and predicted runoff on a monthly basis for the
University of Wisconsin cells indicated that the assumptions used in the HELP
model for snowmelt runoff may not be appropriate. The model stores all pre-
cipitation on the surface when the mean daily temperature interpolated from
the mean monthly temperature is below freezing. When this mean daily tempera-
ture rises above freezing, the precipitation is allowed to either run off or
infiltrate. Since mean daily temperatures are computed In the HELP model
based on mean monthly temperatures which are generally below freezing in Wis-
consin for several consecutive months, no runoff was predicted by the HELP
model during the winter. Instead, a large runoff volume was predicted during
April of each year when temperatures warmed. This compared to measured
results which showed significant runoff throughout the winter without an
excessively large runoff in April. This discrepancy probably contributed to
the overprediction of runoff for several cells,
Evapotransp irat ion
No suitable evapotranspiration field data from landfill sites was found
for model testing. This was not unexpected due to the complexities involved
in collecting this type of data. Yet, evapotranspiration is typically the
single largest outflow component of the landfill system; therefore, small
changes in evapotranspiration can have major impacts on volumes of lateral
drainage and barrier soil percolation.
For those cells which had runoff data available, a surrogate variable for
evapotranspiration was identified, and comparisons were made between measured
and predicted results. The variable consisted of the sura of the water balance
components which were not directly measured. In the case of the University of
Wisconsin cells, the variable was the sum of evapotranspiration and change in
moisture storage, ET+DS. For the Sonoma County cells, it was the sum of evap-
otranspiration, change in moisture storage, and percolation, ET+DS+PERC, The
ET+DS variable was found to be underpredicted by an average of 4 percent of
the measured values, whereas the ET+DS+PERC variable was underpredicted by an
average of 25 percent. It is obviously rather complex to discern the meaning
of these results since evapotranspiration, change in moisture storage, and
percolation are all interrelated. The evidence suggests that values chosen
for evaporative depths may have been too small.
Lateral Drainage and Percolation
Since measurements of barrier soil percolation volumes and leachate pond-
ing depths were not available, the lateral drainage and barrier soil percola-
tion submodel could only be evaluated using measured leachate collection data.
One exception was the Boone County, KY cell where barrier soil percolation
volumes were measured. However, the configuration of the clay liner and
4
-------
percolation collection pipe was such that vertical percolation did riot actu-
ally occur; rather, the percolation flow paths were forced to converge radi-
ally toward the collection pipe. The attempt to simulate this percolation
using the HELP model resulted in an overprediction of approximately 35
percent.
Lateral drainage was overpredicted by 10 percent of the measured drainage
in two cells where very high leachate collection rates were observed. In
three cells where very small quantities of leachate were collected, lateral
drainage was underestimated by 97 percent of the measured drainage, although
this difference only amounted to 1.4 inches per year. Of the remaining nine
cells, lateral drainage was overpredicted by an average of 4 percent of the
measured drainage in five covered cells and overpredicted by an average of 53
percent of the measured drainage in four permanently uncovered cells with a
weathered waste surface that supported dense vegetation. Small errors in the
hydraulic conductivities of the cover soils can cause large differences in the
leachate production when the leachate production is small. Also the overpre-
dictions nay have been partially related to the manner in which the HELP model
estimates unsaturated hydraulic conductivities. To linearly relate unsatu-
rated hydraulic conductivity to moisture content between field capacity and
saturation tends to overpredict unsaturated hydraulic conductivity. Thus,
moisture is routed more quickly through the evaporative zone, contributing to
larger leachate volumes and smaller evapotranspiration volumes.
The poor reproductions of lateral drainage for the uncovered cells at the
University of Wisconsin and the three cells without subsurface liquid addition
at Sonoma County were probably caused by poor estimates of the hydraulic con-
ductivity of the surface layer. The field results could be reproduced by
adjusting only the hydraulic conductivity of the surface layer by less than a
factor of 10, within the range of its probable value. This result is under-
standable since cumulative lateral drainage is dependent on two main factors:
the rate of infiltration into the lateral drainage layer and the rate of per-
colation through the liner beneath the drainage layer. The rate of percola-
tion was very small in these cells; therefore, the rate of infiltration was
the source of error.
Summary
The lack of adequate site description and measured water budget compo-
nents affected the verification study in two ways. First, the lack of
descriptive landfill information required the frequent use of default values
in the HELP model which Introduced additional uncertainty into the verifica-
tion. Second, the lack of water balance outflow measurements limited the num-
ber of HELP outflow predictions that could be verified. These limitations
restricted the ability of the study to isolate and test mathematical charac-
terizations of specific physical processes, such as soil moisture storage and
routing, evapotranspiration demand and its distribution through the soil pro-
file, unsaturated vertical drainage, and details of the apportioning of leach-
ate production between lateral drainage to collection systems and vertical
percolation through the clay liner.
5
-------
In addition, the variable degree of field measurement precision and reli-
ability presented challenges in interpreting the data which did exist. None
of the field data used in this report were collected specifically for verify-
ing the HELP model; therefore, the field data were not always consistent with
the needs of this study. For instance, the data available for the three
largest landfills were collected while they were simultaneously undergoing
expansion. In other cases, there was large variability in measured results
between otherwise identical landfill cells. In general, the error in esti-
mates of water budget components were much smaller than the variability in the
field measurements for similar landfill cells. These results are very good in
light of the fact that the precipitation data used in this study, which is
known to be spatially highly variable, were not measured at most of the land-
fill sites. All of this required a significant amount of engineering judgment
in interpreting the data for the HELP model comparisons.
Although a detailed verification of specific model components was not
always possible, the data did confirm the model's overall utility in estimat-
ing a landfill water balance even without extensive knowledge of specific
landfill characteristics. This was an Important finding since the HELP model
is typically used without a large amount of detailed landfill information.
The following conclusions are made. The field data verified the utility
of the HELP model for estimating general landfill performance. However, not
all model components were well tested due to the limited field data available.
It is concluded that a laboratory and field monitoring program explicitly
designed for HELP verification would be necessary for further refinement of
specific model components. In addition, studies are needed to examine lateral
drainage and percolation for small infiltration rates and flow through syn-
thetic liners and in leakage detection or double liner systems.
The overall data base of long-term water budget measurements at landfills
is poorly organized and too small to continually advance the state of the art
in understanding landfill leachate generation and migration. More extensive
monitoring activities are required to fill this gap.
Improvements to the HELP model are warranted in the areas of snowmelt,
winter runoff, unsaturated hydraulic conductivities, and the selection of
evaporative depths based on the results of this study.
SENSITIVITY ANALYSIS
A sensitivity analysis of the HELP model was performed to examine the
effects of the major design parameters on components of the water budget for
landfills. The analysis examined the effects of cover design, topsail thick-
ness, topsoil characteristics, vegetation, runoff curve number, evaporative
depth, drainable porosity, plant available water capacity, hydraulic conduc-
tivity, drainage length, and liner slope on the water budget. Hydraulic con-
ductivity values for the topsoil, lateral drainage layers and clay liners are
the most important parameters in determining the water budget components.
These parameters are particularly important in estimating the percolation
through the landfill. Other design parameters tend to affect the apportion-
ment between runoff, evapotranspiration and lateral drainage from the cover.
6
-------
The interrelationship between parameters influencing the hydrologic per-
formance of a landfill cover in the HELP model is complex. It is difficult to
Isolate one parameter and exactly predict its effect on the water balance
without first placing restrictions on the values of the remaining parameters.
With this qualification in mind, the following general summary statements are
made.
The primary importance of the topsoil depth or thickness is in control-
ling the extent or existence of overlap between the evaporative depth and the
head in the lateral drainage layer. Surface vegetation has a significant
effect on evapotranspiration from soils with long flow-through travel times
(low hydraulic conductivity) and large plant available water capacities;
otherwise, the effect of vegetation on evapotranspiration is small. The gen-
eral influence of surface vegetation on lateral drainage and barrier soil
percolation is difficult to predict outside the context of an individual cover
design. Clayey soils yield greater runoff and evapotranspiration, and less
lateral drainage and barrier soil percolation. Simulations of landfills in
colder climates and in areas of lower solar radiation are likely to show less
evapotranspiration and greater lateral drainage and barrier soil,percolation.
An increase in the runoff curve number will increase runoff and decrease
evapotranspiration, lateral drainage, and barrier soil percolation. As evapo-
rative depth, drainable porosity or plant available water increase, evapotran-
spiration tends to increase while lateral drainage and barrier soil
percolation tend to decrease; the effect on runoff is varied.
The sensitivity analysis shows that the ratio of lateral drainage to per-
colation is a positive function of the ratio of and the average head
above the liner. However, the average head is a function of Q^/K^ and L/a.
The quantity of lateral drainage, and therefore also the average Head, is in
turn a function of the infiltration. Therefore, the ratio of lateral drainage
to percolation increases with increases in infiltration, and the ratio of
K^/Kp for a given drain and liner design. The ratio of lateral drainage to
percolation for a given ratio of K^/Kp increases with increases in infiltra-
tion and decreases in L/a. The percolation and average head above the liner
are positive functions of the term L/a,
REVIEW OF TECHNICAL GUIDANCE
The information from the sensitivity analysis and the verification
results were used to evaluate RCRA landfill design guidance and regulation.
This evaluation showed that saturated hydraulic conductivity is the most
important design parameter for minimizing percolation. Care should be taken
to recommend the highest hydraulic conductivity that Is commonly available for
drainage media. Similarly, the lowest saturated hydraulic conductivity prac-
tically obtainable should be used as guidance for soil liners. Changes in
other design parameters yield much smaller effects on percolation or leakage
volumes if the values of these parameters are kept in a reasonable range.
7
-------
SECTION 2
INTRODUCTION
BACKGROUND
Landfills have come to be a widely employed means for disposal of munici-
pal, industrial and hazardous solid wastes. Storage of any waste material in
a landfill poses several potential problems. Among these is the possible con-
tamination of ground and surface waters by the migration of water or leachate
from the landfill to adjacent areas. Given this potential problem, it is
essential that the liquids management technology perform as expected during
its development. It is also essential that the performance of the technology
can be simulated or modeled with sufficient accuracy to design landfills to
prevent migration of water from the facility. The modeling of the moisture
movement through landfills is also important to review landfill designs and to
determine the adequacy of the design and the limitations of the liquids
management technology.
The Hydrologic Evaluation of Landfill Performance (HELP) model was devel-
oped to help hazardous waste landfill designers and evaluators estimate the
magnitudes of components of the water budget and the height of water-saturated
soil above barrier soil layers (liners). This quasi-two-dimensional, deter-
ministic computer-based water budget model perforins a sequential daily analysis
to determine runoff, evapotranspiration, percolation, and lateral drainage for
the landfill (cap, waste cell, leachate collection system, and liner) and
obtain estimates of daily, monthly, and annual water budgets (1,2). The model
does not account for lateral inflow or surface runon.
The determination of the magnitude of each component of the water budget
is not a simple task. The interrelationships between climate, vegetation and
soil characteristics and their effects on runoff, evapotranspiration and ver-
tical drainage are very complex and not easily determined without a model.
Therefore, a model can be very useful to verify that liquid management
technology performs as anticipated, that regulations are practicable and
achieve their goal, and that the technical guidance for design and evaluation
is correct.
In the course of the development of any model, provisions should be taken
to verify that the model accurately represents reality. Laboratory tests have
been performed to verify the lateral drainage portion of the HELP model, but
prior to this study the other portions of the model had only been calibrated
and not verified. This report presents the results of efforts to verify the
model using existing data collected at landfills, test cells and lysimeters.
8
-------
SCOPE AND PURPOSE
This study consists of three parts. The main part is an attempt to
verify the model using existing data. The purpose of this part of the study
was to assess the adequacy of the model to simulate reality and to validate
the use of the model to evaluate designs (generally with minimum information).
Data sets were obtained at several landfill sites in different areas of the
United States—California, Wisconsin, Kentucky and New York. The landfills
varied in their designs, operating conditions and soil characteristics. In
general, the landfill characteristics are loosely defined and permit a lot of
latitude and uncertainty in the selection of values to describe the landfill.
The water budget data from the different sites varied in the number of compo-
nents measured—only leachate collection was measured at some while precipita-
tion, runoff, percolation and leachate collection were measured at others.
Consequently, the level of verification attempted varied from site to site,
but efforts were made in all cases to isolate processes (evapotranspiration,
runoff, percolation and lateral drainage) and to determine how accurately each
is modeled.
The second part of this study is a sensitivity analysis of the principal
parameters used in the HELP model. The purpose of this part of the study was
to determine the parameters that need to be well defined and the likely
effects of a change in the value of a parameter. This analysis also provided
insight to interpret and explain the verification results. Several levels of
detail and specificity were used in the sensitivity analysis. The lowest
level of specificity examined the general effects of cover design, cover soil
type, cover thickness and vegetation for three climates—cold and humid
(Schenectady, NY), hot and humid (Shreveport, LA) and hot and semiarid
(Santa Maria, CA). The next level of detail specifically examined the effects
of runoff curve number, evaporative depth, drainable porosity and plant
available water capacity for two cover designs and two climates using the same
cover soil type, cover thickness and vegetation. The highest level of specif-
icity examined only the proportioning of infiltrated water between the budget
components of percolation and lateral drainage. This last analysis compared
the effects of hydraulic conductivity of the barrier soil layer (liner),
hydraulic conductivity of the lateral drainage layer, slope of the base of the
lateral drainage layer and maximum lateral drainage length.
The third part of the study consists of an evaluation of the technical
guidance developed to support the regulations regarding design and operation
of a landfill. The purpose of this part of the study was to examine the tech-
nical guidance and regulations from the point of view of minimizing potential
percolation or leakage through the liner at the base of the landfill in the
best practicable manner. This evaluation examined the results of the labora-
tory and field verification studies and the sensitivity analysis to determine
whether the technical guidance was practicable and achieved its objectives.
The evaluation examined guidance on final covers, leachate detection, collec-
tion and removal systems and liners. Specifically, evaluation of the final
cover guidance included effects of vegetation, vegetated cover soil thickness,
drainage layer design and liner design. Evaluation of leachate detection,
collection and removal systems includes examination of the hydraulic
9
-------
conductivity, slope and maximum drainage length of the drainage layer. Evalu-
ation of liners included effects of liner thickness and hydraulic
conductivity.
10
-------
SECTION 3
MODEL DESCRIPTION
The HELP computer program models the effects of hydrologic processes on
the water budget for landfills by performing daily, sequential analysis using
a quasi-two-dimensional, deterministic approach. The dominant hydrologic con-
siderations include precipitation in any form, surface storage, interception,
surface evaporation, runoff, snowmelt, infiltration, vegetation, rooting
depth, plant transpiration, soil evaporation, temperature, solar radiation,
soil moisture storage, soil moisture potential, unsaturated flow, saturated
flow, percolation and lateral drainage. The program handles each of these
considerations, often in a simplified manner, in five main routines to esti-
mate runoff, evapotranspiration, vertical drainage to liners, percolation
through liners and lateral drainage from layers above liners. Several other
routines interact with these routines, and these routines are involved with
snowmelt, surface storage, synthetic flexible membrane liners and generation
of daily temperature, solar radiation and leaf area index values. This sec-
tion briefly describes the five major routines and the snowmelt and synthetic
flexible membrane liner routines to provide a basis for understanding the
approaches used in the sensitivity analysis and verification study, the
results of the studies, and the conclusions. A more detailed description of
the model is presented in the documentation report for the model (1).
The HELP model computes runoff by the Soil Conservation Service (SCS)
runoff curve number method (3). Percolation is computed by Darcy's law, modi-
fied for unsaturated flow (1). Lateral flow Is computed by a linearization of
the Boussinesq equation, and evapotranspiration is determined by a method
developed by Ritchie (4). The vertical percolation and evapotranspiration
components of the HELP model originated with the CREAMS (5) program. The
development presented here, however, reflects a significant advance beyond the
CREAMS model in both of these areas. Each of these model components will be
discussed in detail below.
RUNOFF AND INFILTRATION
The SCS curve number technique (3), as first implemented in the CREAMS
model, was selected as a model to partition incoming rainfall or snowmelt
between runoff and infiltration in the HELP model.
The SCS equation relates daily runoff, Q, to daily precipitation, P, and
a watershed retention parameter, S, as follows:
Li
-------
(P - 0.2S)2
(P + 0.8S)
where Q, P, and S are in inches.
The retention parameter, S, is a nonlinear function of soil moisture and
vegetative cover density. This function is described by a series of empirical
curves developed by SCS. An appropriate curve can be selected from the SCS
manual (3) to match the desired surface cover for a landfill. The curve num-
ber, CN, is related to the retention parameter as follows;
S - - 10 (2)
The actual value of S for any time during a simulation period is computed
by the procedure described in CREAMS (5):
5 ¦ I1 - (i^§)] <3'
where S » retention parameter value for a dry antecedent condition
max r
SM * a depth-weighted estimate of soil moisture, vol/vol
WP = the lower limit of soil water storage or wilting point,
vol/vol
11 = upper limit of soil water storage or total porosity of the
soil, vol/vol.
To account for the nonuniform distribution of moisture in the soil pro-
file, a weighting function which gives greater weight to moisture near the
surface is used. The soil profile in the evaporative zone is divided into
seven segments. These soil segments are used in the moisture routing scheme,
and their soil moistures are averaged to compute the effective soil moisture.
Soil segments vary in thickness as follows; the uppermost segment has a
thickness equal to 1/36 times the lesser of the depth from the surface to the
barrier soil layer and the depth of the evaporative zone profile; the second
layer has thickness equal to 5/36 of this profile, and the remaining five seg-
ments each have thicknesses equal to 1/6 of this profile. The depth—weighted
retention parameter is computed as follows:
max
SM. - WP
2 1
where = weighting factor for segment j
SM.
soil moisture content of segment j
(4)
12
-------
WPj = wilting point of segment j
UL^ = upper limit of soil moisture content of segment j.
The weighting factors are identical to those used In CREAMS (5): for segments
1 through 7, these weighting factors are 0.111, 0.397, 0.254, 0,127, 0.063,
0.032, and 0.016, respectively.
Infiltration is computed as the difference between the daily precipita-
tion and runoff computed by Equation 1, minus the daily surface evaporation.
If the mean daily temperature is below 32°F, the precipitation is stored on
the surface as snow and does not contribute to infiltration or runoff until
the mean daily temperature rises above 32°F. The amount of melted snow is
added to precipitation for use in Equation 1.
The amount of snowmelt on day i is computed as follows (3):
M. => 0 for T, < 32°F or SN0, , = 0
1 i i-1
M - 0.06 (T - 32) for T > 32°F (5)
i i i
and SN01_1 > 0.06 (T - 32)
M = SNO, , for T > 32°F and
i i-1 i
SNO, . < 0.06 (T, - 32),
i-1 i
where = amount of snow melted on day i, inches
T^ = mean temperature on day i, "F
SNO^ ^ = amount of snow water on surface at the end of day i-1, inches,
EVAPOTRANSPIRATION
The evapotranspiration (ET) from a landfill cover is a function of the
energy available, the vegetation, the soil transmissivity, and the soil mois-
ture content. The potential ET is computed in the HELP model by a modified
Penman method as described by Ritchie (4) and used in CREAMS (5):
E = i'+VftS (6)
o A + 0.68
where Eq = potential ET, mm
A = slope of saturation vapor pressure curve
H ¦ net solar radiation, langleys.
The daily potential ET is applied first to any free water available on
the surface, thereby reducing the computed infiltration or the amount of snow
on the surface. Any ET demand in excess of free surface water is exerted on
13
-------
the soil column for evaporation directly from the soil and for transpiration
through the surface vegetation.
Soil evaporation proceeds in two stages: stage one, where evaporation is
controlled by atmospheric demand, and stage two, where evaporation is limited
by soil transmissivity. The potential evaporation from soil is computed as
follows;
ES - E exp (-0.4 LAI) (7)
o 0
where ESq ¦ potential evaporation from soil, mm
LAI = leaf area index (on scale of 0 to 3).
During the nongrowing season, LAI in Equation 7 is replaced by a winter cover
factor that varies between 0 and 1.8 depending on the density of dead or dor-
mant vegetation.
In stage one evaporation, soil water evaporation is equal to the potential
soil water evaporation ES . An upper limit to the amount of stage one evapora-
tion is computed as follows:
U = 9 (a - 3)0-42 (8)
s
where U = the upper limit for stage one evaporation, mm
L/2
a = soil transmissivity parameter for evaporation, (mm/day) ,
s
When the accumulated stage one evaporation less the infiltration exceeds the
upper limit, U, any further soil water evaporation proceeds by the following
equation:
t1/2 - (t - 1)1/2| (9)
ES2 " as .
where ES2 - stage two evaporation from soil, mm
t ¦ days since stage one ended.
Whenever the infiltration exceeds the cumulative evapotranspiration less the
upper limit for stage one evaporation, evaporation reverts to the stage one
process.
Potential plant transpiration is computed as
E LAI
EP = ~ (10)
O 5
The potential ET may be satisfied entirely by evaporation from free water or
from stage one or stage two evaporation from soil. If there is remaining
atmospheric demand after these sources of ET have been exhausted, the demand
14
-------
will be satisfied by plant ET to a maximum of EP^. The actual ET varies as a
function of soil moisture and atmospheric demand as shown by Shanholtz and
Lillard (6), Saxton et al. (7), and Sudar et al. (8). The following empirical
relationship was developed from ET rate curves for no-till corn presented by
Shanholtz and Lillard (6):
EP = EPD ^1.20 - + FC - WP > (11)
where EP = actual plant ET, mm
EPD * actual plant ET demand, mm
- E - ES - ESS if EP > E - ES - ESS
0 o o
= EP if EP < E - ES - ESS
O 0 0
ES = actual soil water evaporation, mm
ESS = that portion of the total evapotranspiration met from precipi-
tation or snow, mm
SM - soil moisture content, vol/vol
WP - wilting point moisture content of the soil, vol/vol
FC = field capacity of the soil, vol/vol.
The soil moisture, wilting point, and field capacity values used in Equa-
tion 11 are depth-weighted using the same weighting coefficients used for
computation of the retention factor used in Equation 1.
VERTICAL MOISTURE FLOW
The vertical flow submodel is depicted in Figure 1 as it may be imple-
mented for simulation of moisture movement in a landfill cover. When applied
to an open landfill or to a landfill with multiple barrier soil layers, soil
or waste layers below the top barrier soil layer are represented by fewer
segments.
The input to the vertical flow model is the infiltration estimate. Out-
put from the model is divided between lateral flow and vertical percolation
through the barrier soil layer. Drainable moisture, that in excess of field
capacity, is transferred through the profile by storage routing procedures.
Storage routing through segments of the model consists of simultaneous
solution of Darcy's law and an equation of continuity. The model does not
account for water movement by capillary action. Darcy's law is represented as
Q = K ^ (12)
15
-------
i "
INFILTRATION
i 517
¦DRAINAGE FROM SEGMENT 1
<1 D5
DR2
DR •:
TOPSOIL LAYER
DR/
4
DRc
DR6
LATERAL DRAINAGE
LAYER
-> LATERAL DRAINAGE RATE
4, PERCOLATION RATE
DR7= LATERAL DRAINAGE RATE + PERCOLATION RATE
¦BARRIER SOIL
LAYER
PERCOLATION RATE
Figure 1. Vertical flow submodel for landfill cover.
16
-------
where Q = rate of vertical flow, inches/day
K = hydraulic conductivity, inches/day
h = gravitational head, inches
1 = length in the direction of flow, inches.
Free outfall is assumed from each segment such that dh/dl may be set
equal to unity, and Q is equal to the hydraulic conductivity. This assumption
is acceptable only if the conductivity of the profile is constant or increases
with depth. Because this assumption is not met at the interface between the
seventh segment and the barrier layer in Figure 1, a different procedure is
employed at the top interface of a barrier layer.
The rate of vertical moisture flow in soil varies with soil moisture con-
tent. When the soil is fully saturated, the flow rate is equal to the satu-
rated hydraulic conductivity, and when the moisture content is reduced to
field capacity, the flow rate is zero. In the HELP model the hydraulic
conductivity is taken to be a linear function of soil moisture:
K =
u
K
SM - FC
s UL - FC
(13)
where = unsaturated hydraulic conductivity, inches/day
Kg = saturated hydraulic conductivity, inches/day
SM
soil moisture content, vol/vol
FC - soil moisture content at field capacity, vol/vol
UL = total porosity or the soil moisture content at saturation,
vol/vol.
The change in storage in a segment from mid time period i-1 to mid time
period i may be written as the sum of one-half of the net drainage from the
previous time period and one-half of the net drainage at the end of the cur-
rent time period. Therefore,
SM1 - SMi_l + 1/2 NDRi + 1/2 NDR L (14)
where SM^ » soil moisture at the end of the current time interval (mid-
time period)
SMi 1 " soi^ moisture aC the end of the previous time interval (mid-
time period)
SDR = net drainage
and
17
-------
NDRi(j) = |DRi(j-l) - DRjLCj) - ET^j^At
where j = segment number
1 = time period
DR = drainage rate
At = the time interval.
The drainage rate from a segment is written as
'SM. + SM,
(15)
DR,
K
i-1
FC
UL - FC
(16)
Equations 14, 15 and 16 may be solved simultaneously for DR^(j), knowing SM^ ^
from the previous time interval. The resulting drainage from segment j is as
follows:
DR1(j)
2
SMi_1(j-l) + 1/2 NDR^Cj)
(17)
+ [l/2 DR1(j-l) At] - |l/2 ETt(j) AtJ- FC(j)|
/ 12 |lJL(j) - FC(j)J+[Ks(j) At]
The routing proceeds from the top segment, where incoming drainage is infiltra-
tion (rainfall less surface evaporation), to the seventh segment in Figure 1
where the drainage out is divided between percolation through the barrier soil
layer and lateral drainage above the barrier soil layer*
PERCOLATION
An accurate estimate of percolation rate through a barrier soil layer
cannot be obtained using the routing procedure described in the vertical flow
submodel because of the discontinuity at the interface with the barrier soil
layer. Percolation is modeled as Darcian flow where the percolation rate is
computed as:
TH
(18)
where Qt
the rate of percolation through the barrier soil layer
Kj, = the saturated hydraulic conductivity of the barrier soil layer
TH = the total head in the profile above the barrier soil layer
the thickness of the barrier soil layer.
18
-------
To reduce the complexity of the percolation model, the drainable porosity of
the barrier soil layer is considered negligible; that is, the barrier soil
layer is assumed to be saturated at all times. This assumption has the effect
that percolation stops as soon as there is no head above the barrier soil
layer. In fact, if the barrier soil is a typical clay, it would continue to
drain at a very low rate until the drainable water is depleted. If the drain-
able porosity of the clay were 1 percent and the thickness of the barrier soil
layer were 60 cm, 0.6 cm of water could drain from the layer after all drain-
able water was removed from the profile above the barrier soil laye^. Because
the hydraulic conductivity of clay is very low (on the order of 10 cm/sec),
it would require several months to deplete the drainable water from the clay.
Further, it must be considered that the percolation of water from the soil
profile into the drainable volume of the clay will also proceed at a slow rate
due to the low hydraulic conductivity of the clay. Therefore, under humid
conditions, the drainable porosity of the barrier soil layer assumption will
have a negligible effect; under arid conditions with occasional wetting, the
amount of percolation may be underpredicted. The underprediction, however,
would be largely compensated over extended simulation periods.
Percolation through barrier soil layers having a synthetic membrane liner
is also modeled by Darcy's law. The model uses the same equation (Equa-
tion 18) and routine to compute the quantity of percolation as in the absence
of a synthetic membrane but uses a different hydraulic conductivity for the
barrier soil. The model computes a resultant hydraulic conductivity of the
layer and synthetic liner by multiplying the hydraulic conductivity of the
soil by the leakage fraction of the synthetic liner. This resultant hydraulic
conductivity is used in Equation 18 to compute percolation through soil liners
with synthetic membranes.
LATERAL FLOW SUBMODEL
The lateral flow submodel of the HELP program is based on a linearization
of the steady-state Boussinesq equation performed by Skaggs (9) which yielded
the following equation:
2 S y h0
% - 2 <19>
Li
where = lateral drainage rate
Kp = hydraulic conductivity for lateral flow
y * average thickness of flow
h ¦ elevation of water surface at x = L
o
x = lateral distance measured from drain
L = maximum length to drain.
19
-------
The model configuration is shown in Figure 2 for a landfill having a drainage
length of L and a drainage slope of a.
DRAIN
Figure 2. Configuration of lateral drainage model.
The vertical flow submodel simulates vertical flow for the entire cross
section shown in Figure 2. Therefore, the thickness of the saturated zone
computed in the vertical flow submodel corresponds to y. Using a finite dif-
ference model, developed by Skaggs (9) to solve the Boussinesq equation,
lateral drainage rates and water profiles were generated for nine combinations
of a and L with y ranging from 0.2 to L.O m to obtain correction factors for
the linearized Boussinesq equation.
Equation 19 may be rewritten with the correction factor as follows:
2 Cj_ y hQ (20)
Qn = 7 '
where
h - y + a L (21)
o Jo
and yQ = depth of saturation at x - L
a = fractional slope at surface of liner
= correction factor
=¦ 0.510 + 0.00205 a L (22)
The variables, y and h , are unknown during the simulation; therefore, a
relationship between y and yQ was developed of the following form;
20
-------
y *= c~ v
J o 2 -
(23)
where
¦ (^) °'
16
(24)
Therefore, the final font of Equation 19 is;
*
. 1 - V-16
2^7 (0.510 + 0.00205 a L)
IrJ + ¦ '
(25)
where the units of v and L are in inches and a is dimension less.
21
-------
SECTION 4
SENSITIVITY ANALYSIS
This chapter examines the sensitivity of the HELP model to variations in
values of selected input parameters. This information is useful in a variety
of ways. It can aid the design engineer in selecting preliminary design
alternatives for specific soil characteristics, geographical regions, surface
vegetation, and layer thicknesses. It can serve as a basis for evaluating and
establishing technical guidelines for regulatory agencies. It can also pro-
vide additional insight to the HELP model user in understanding the importance
and interaction of specific variables in controlling the water balance within
a landfill. Finally, it can assist in evaluating the suitability of method-
ologies used in the HELP model.
The parameters included in this analysis are listed in Table 1. They are
grouped according to their role in either cover design or drainage and perco-
lation design. Each group will be examined separately.
TABLE 1. PARAMETERS SELECTED FOR SENSITIVITY ANALYSIS
Analysis of Cover Design
Topsoil type
Cover soil thickness
Surface vegetation
Runoff curve number
Evaporative depth
Drainable porosity
Plant available water
Geographical location
Municipal vs. hazardous waste cover design
Analysis of Percolation and Drainage Design
Hydraulic conductivity of barrier soil layer
Hydraulic conductivity of lateral drainage layer
Slope of lateral drainage layer
Drainage length
22
-------
LANDFILL COVER
This analysis for landfill covers is divided into two parts. Both parts
examine the effect of geographical location, municipal versus hazardous waste
cover design, and topsoil characteristics. In addition, Part 1 considers top-
soil thickness and surface vegetation, whereas Part 2 considers runoff curve
number, evaporative depth, drainable porosity, and plant available water. The
following discussion applies to Parts 1 and 2,
To determine the effect of various climatological regimes on cover per-
formance, three locations were studied—Santa Maria, CA; Schenectady, NY; and
Shreveport, LA. These locations represent a wide range in levels of precipi-
tation, temperature, and solar radiation as summarized in Table 2. Default
values for precipitation, temperature, solar radiation, and leaf area index
are stored in the HELP model for each site and were used for the sensitivity
analysis simulations. The period of record for the daily precipitation values
stored in the HELP model is 1974 through 1978.
In addition, two cover designs were examined as shown in Figure 3. One
was typical of hazardous waste landfills where topsoil overlies a 1-foot-thick
lateral drainage layer (hydraulic conductivity = 3 x 10 cm/sec, slope »
0.03, drainage length = 200 feet) which further overlies a 2-^oot-thick clay
liner or barrier soil layer (hydraulic conductivity - 1 x 10 cm/ sec). The
second was typical of municipal sanitary landfills where topsoil overlies a
two-foot-thick barrier soil layer (hydraulic conductivity = 1 x 10~ cm/sec).
Two types of topsoil were considered in the cover designs: sandy loam
and silty clayey loam. The sandy loam characteristics were those of the HELP
model default soil texture 8, which represents Unified Soil Classification
System (USCS) soil class SM and U.S. Department of Agriculture (USDA) soil
class SL. The silty clayey loam characteristics were those of the HELP model
default soil texture 15, which represents USCS soil class CL and USDA soil
class SICL.
Part 1
In Part 1, the vegetative cover was designated as either good grass or
poor grass. This selection dictated the values for runoff curve number and
evaporative depth, and influenced the value of the hydraulic conductivity of
the topsoil. For a given vegetative cover, the runoff curve number was
obtained from Figure 4 of the HELP Model User's Guide (2) using the minimum
infiltration rate given in Table 10 of the same reference for sandy loam or
silty clayey loam. The depth of the evaporative zone was chosen as 7 inches
for poor grass and 14 inches for good grass, based on recommendations con-
tained in the HELP model.
Table 3 summarizes the parameter combinations examined under Part 1 and
presents the results of the HELP model simulations as percentage of precipita-
tion. Simulations for the hazardous waste cover design were performed for
both soil types, whereas simulations for the municipal design were performed
only for sandy loam. The same results are presented in the form of bar graphs
23
-------
TABLE 2. CLIMATOLOGICAL REGIMES
Climatological
Variable
Location
Santa Maria, CA
Schenectady, NY
Shreveport, LA
Precipitation*
Mean Annual (in.)
14
48
44
Mean Winter
(Nov-Apr) (in.)
12
19
22
Mean Summer
(May-Oct) (in.)
2
29
22
Temperature
Mean Annual (°F)
57
49
66
Mean Jan. (°F)
51
23
47
Mean July (°F)
62
73
83
Days with Minimum
Below 32°F
24
129
37
Solar Radiation
Mean Daily (langleys)
450
290
410
* These mean values are for the period simulated by the HELP model in this
section, 1974-1978.
24
-------
VEGETATION
i.. 11,11 1111 ii, ! 1 I It. II i., 1 ij-i I n 111 I f.l ,. 11IV i I 1 in It (I If (f f I 11 w 11« 1 I!)
TO PS 01L
LATERAL DRAINAGE
BARRIER SOIL
18* or 36"
12'
24"
a) Hazardous Waate Landfill Cover Design
VEGETATION
i ii'i
TOPSOIL
BARRIER SOIL
18* or 36*
b) Municipal Landfill Covor Design
Figure 3. Cover designs for sensitivity analysis.
25
-------
TABLE 3. SENSITIVITY OF WATER BUDGET COMPONENTS TO LOCATION, VEGETATION,
AND TOPSOIL TYPE AND DEPTH
Average Annual Volume (Percent Precipitation)**
Description* Hazardous Waste Design Municipal Design
Soil
Soil Depth Lat.t Barriertf Barrierf
Site
Type
Veg.
(in.)
Runoff
ET
Drng.
Perc.
Runoff
ET
Perc
CA
SL
GG
18
0.0
53.4
43.4
4.2
8.8
58.1
34.2
CA
SL
GG
36
0.0
53.6
43.2
4.2
3.1
55.1
42.9
CA
SL
PG
18
3.0
52.0
41.4
4.2
11.2
52.4
36.9
CA
SL
PG
36
3.0
52.2
41.2
4.2
5.6
52.5
42.6
CA
SICL
GG
18
7.4
73.1
18.1
2.0
CA
SICL
GG
36
7.4
73.1
18.1
2.0
CA
SICL
PG
18
21.6
61.4
15.0
2.2
CA
SICL
PG
36
21.6
61.4
15.0
2.2
LA
SL
GG
18
0.2
52.9
43.8
3.1
2.2
66.3
31.6
LA
SL
GG
36
0.2
53.0
43.7
3.1
0.2
56.6
42.6
LA
SL
PG
18
4.4
51.6
40.6
3.1
7.5
56.2
35.6
LA
SL
PG
36
4.4
51.8
40.5
3.1
4.6
52.3
42.4
LA
SICL
GG
18
8.5
73.9
15.0
2.1
LA
SICL
GG
36
8.6
74.1
14.7
2•1
LA
SICL
PG
18
22.3
64.2
11.3
2.0
LA
SICL
PG
36
22.3
64.2
11.3
2.0
NY
SL
GG
18
0.04
50.9
45.6
2.5
9.5
60.6
28.9
NY
SL
GG
36
0.04
51.0
45.5
2.5
3.5
54.3
41.2
NY
SL
PG
18
2.2
50.1
44.0
2.5
13.4
53.3
32.1
NY
SL
PG
36
2.2
50.2
43.9
2.5
5.5
50.9
42.4
NY
SICL
GG
18
7.4
63.8
25.2
2.0
NY
SICL
GG
36
6.7
63.9
25.7
2.0
NY
SICL
PG
18
19.2
57.3
20.3
1.9
NY
SICL
PG
36
19.2
57.2
20.3
1.9
* CA = Santa Maria, CA; LA - Shreveport, LA; NY - Schenectady, NY;
SL = sandy loam (HELP default texture 8); SICL - silty clayey loam
(HELP default texture 15); GG = good grass; PG = poor grass. Curve
numbers were selected from Table 10 and Fig. 4 in Ref. 2. Evaporative
depths were 14 inches for GG and 7 inches for PG,
** Change in storage is not included in this table; therefore, the water
balance components shown do not always sura to 100.0%.
t Sloge = 3%; drainage length » 200 ft; hydraulic conductivity « 3 x
10 cm/sec.
ft Hydraulic conductivity - 10 g cm/sec.
| Hydraulic conductivity = 10 cm/sec..
26
-------
in Figures 4 and 5, Here, the height of each bar segment represents the
corresponding water balance component in mean annual inches.
Topsoil Depth—
As seen in Table 3, negligible differences existed between the 18- and
36-inch soil depth simulations for the hazardous waste design. For this rea-
son only the results for 18-inch depth are included in Figure 4, These small
differences are not surprising considering that both soil depths were suffi-
ciently great so that no overlap existed between the evaporative zone and the
head in the lateral drainage layer. This case is contrasted with the munici-
pal design where larger heads (due to the lack of lateral drainage) interacted
with the evaporative zone and increased evapotranspiration and runoff. At the
same time, the larger heads increased percolation through the barrier soil.
These increases in percolation over the hazardous waste design ranged from
800 to 1700 percent. The permeability of the barrier soil for the municipal
design was 10 times as large as that for the hazardous waste design, which
could account for a maximum of a 1000-percent increase in percolation.
For the municipal design, significant differences existed between the 18-
and 36-inch soil depth simulations. Runoff and evapotranspiration were
greater for the 18-inch depth, indicating that the head above the barrier soil
layer maintained higher moisture contents in the evaporative zone. However,
percolation was greater from the 36-inch depth due to its ability to accommo-
date higher heads and longer sustaining heads since evapotranspiration was
limited to the top 7 to 14 inches of the cover.
Surface Vegetation—
In comparing the effect of good and poor grass, one would expect good
grass to increase evapotranspiration and decrease runoff. Evapotranspiration
would be increased due to an Increased plant demand for moisture. Runoff
would be decreased due to a reduced runoff curve number, a drier topsoil due
to the greater evapotranspiration, and an increased hydraulic conductivity of
the evaporative zone corresponding to an Increased root density. The results
of the HELP model simulations in Table 3 and Figures 4 and 5 confirm these
effects. However, the influence of surface vegetation on the volume of lat-
eral drainage and barrier soil percolation is varied. For the hazardous waste
design, the increase in infiltration with good grass was greater than the
increase in evapotranspiration, resulting in a larger volume of lateral drain-
age and a negligible change in barrier soil percolation. For the municipal
design, the high heads above the barrier soil assisted evapotranspiration by
maintaining higher moisture levels for plant uptake and hindered infiltration
by maintaining higher antecedent moisture conditions; therefore, the increase
in evapotranspiration was greater than the increase in infiltration. This
resulted in a trend toward decreased barrier soil percolation for good grass.
It should be recognized that good grass could not be maintained in the
climate of Santa Maria, CA, by rainfall alone, as assumed in this analysis.
An additional sprinkler system water inflow component would be required to
make the conditions realistic. However, the careful regulation of this
surface-applied water would not change the results regarding lateral drainage
or barrier soil percolation.
27
-------
ho
00
60 -
55 -
~ 50 -
w
UJ
X
o
z
UJ
3
_l
<
>
_l
<
z
z
<
z
<
111
z
45 -
40 -
35 -
30 -
25 -
20 -
15 -
10
5
0
En3
RUNOFF
EVAPOTRANSPIRATION
LATERAL DRAINAGE
PERCOLATION
QQ GOOD GRASS
PQ POOR GRASS
SL 18"OF SANDY LOAM
SICL 18" OF SILTY
CLAYEY LOAM
0.0
7.6
6.2
*Lfi_
0.4
7.4
6.9
QJL
1.1
10.4
2.6
3.1
8.8
2.2
QJL
0.1
23.3
QQ PQ
SL
QQ PQ
SICL
SANTA MARIA, CA
19.3
LA.
2.0
n 3.8
22.9
f
18.0
1A.
32.8
6.6
±iQ_
9.9
28.4
6.0
QJL
0.0
~ 1.1
24.7
QQ PG
SL
GQ PQ
SICL
SHREVEPORT, LA
22.1
*2_
24.3
21.4
LL2_
QG PQ
SL
SCHENECTADY, NY
3.6
31.1
12.3
9.4
27.9
9.9
UL
GG PG
SICL
Figure 4. Bar graph for hazardous waste cover design showing effect of surface
vegetation, topsoil type, and location.
-------
M
vO
<
>
—I
<
3
Z
<
UJ
2
60 -
55 -
~ 50 -
45 -
(0
Hi
X
o
z
w 40 -
Hi
3
GG
PG
18
36
RUNOFF
EVA POTRANSPIR ATION
PERCOLATION
GOOD GRASS
POOR GRASS
18" OF SANDY LOAM
36" OF SANDY LOAM
35 -
30 -
25 "
20 -
I 5
1.3
10
5
0
/
/
/
8.3
I
4.9
1.6
7.5
5.3
^ 0.4
7.8
6.1
0.8
7.6
6.1
GG
PG
GG
PG
18
36
SANTA MARIA, CA
1.0
29.4
13.8
GG
3.3
26.0
15.8
0.1
26.1
18.9
PG
18
GG
2.0
23.3
16.9
PG
36
SHREVEPORT, LA
4.6
/
/
/
/
~
/
/
/
/
/
~
/
/
/
/
29.4
r
/
/
/
/
/
/
/
/
/
/
/
/
/
/
~
14.0
6.5
25.9
15.6
?
/
'
/
/
s
/
/
/
/
/
/
/
/
/
/
1.7
26.3
19.9
2.7
~
~
/
/
/
/
/
~
/
/
/
/
/
/
24.7
20.6
GG
PG
GG
PG
18
36
SCHENECTADY, NY
Figure 5. Bar graph for municipal cover design showing effect of topsoil depth,
surface vegetation, and location.
-------
Topsail Type—
The results show that the clayey topsoil increased both runoff and evapo-
transpiration, which in turn decreased lateral drainage and barrier soil per-
colation. The increase in runoff was due primarily to the larger runoff curve
number selected by the HELP model for clayey soils using the minimum infiltra-
tion rate method (2). The increase in evapotranspiration was due to: (a) the
lower hydraulic conductivity of the clayey soil which slowed the drainage rate
and maintained moisture levels above field capacity for longer periods of
time, thus allowing evapotranspiration to deplete larger volumes of moisture,
and, more Importantly, (b) the larger plant available water capacity (field
capacity minus wilting point) which left a larger, moisture reservoir available
for evapotranspiration losses after drainage ended.
Geographical Location-
Daily potential evapotranspiration as computed in the HELP model is a
function of daily temperature and daily solar radiation. Of the three loca-
tions considered in this analysis, Schenectady, NY, experiences the lowest
temperature and the lowest mean solar radiation. Consequently, for the haz-
ardous waste design, the HELP model simulations show the least evapotranspira-
tion (as percent precipitation) for Schenectady. However, the simulations
show a slightly greater evapotranspiration percentage at Shreveport for clay
soils even though Santa Maria experiences higher temperatures and greater
solar radiation. The long travel time through clay soils combined with the
higher average soil moisture levels at Shreveport (due to greater rainfall)
probably contributes to this increased evapotranspiration percentage. Rain-
fall is more uniformly distributed throughout the year in Shreveport, while
precipitation in Santa Maria occurs primarily during the winter months.
For the municipal designs, simulated evapotranspiration percentages for
Santa Maria were among the lowest. This was probably related to the greater
rainfall amounts at the other locations combined with the hydraulically
restrictive barrier soil layer which created higher heads in the evaporative
zone; thus, more water was subjected to potential evapotranspiration for a
longer period of time. Consequently, the evapotranspiration percentages
tended to be larger for Shreveport and Schenectady, In actuality, the depth
of root penetration would probably vary between the sites for the same degree
of vegetation due to effects of climate. If this were considered, evaporative
depths and therefore evapotranspiration would likely be somewhat greater than
indicated herein for the drier climate of Santa Maria.
Part 2
The effects of runoff curve number, evaporative depth, drainable poros-
ity, and plant available water are discussed below. For this analysis, the
vegetation was assumed to be fair grass. Tables 4 and 5 summarize the param-
eter combinations examined under Part 2 and present the results of the HELP
simulations as percentage of precipitation. Figures 6 through 13 illustrate
the same results, with values in mean annual inches.
Runoff Curve Number—
The runoff curve number was varied from 65 to 90 for the sandy loam and
from 75 to 95 for the silty clayey loam. The depth of the evaporative zone
30
-------
TABLE 4, SENSITIVITY OF WATER BUDGET COMPONENTS TO EVAPORATIVE DEPTH AND
CURVE NUMBER
Average Annual Volume (Percent Precipitation)**
Description* Hazardous Waste Design Municipal Design
Evap
Soil
Depth
Curve
Lat.t
Barrier 1t
Barrier*
Site
Type
(in.)
Number
Runoff
ET
Drng.
Perc.
Runoff
ET
Perc.
CA
SL
10
65
0.1
52.7
43.6
4.2
7.1
53.8
39.9
CA
SL
10
80
2.6
51.9
41.9
4.2
8.7
53.0
39.1
CA
SL
10
90
11.3
49.5
35.9
4.1
14.4
50.4
36.0
CA
SICL
10
75
5.5
70.8
22.1
2.2
CA
SICL
10
85
12.7
67.6
18.0
2.2
CA
SICL
10
95
34.4
57.3
6.4
1.6
CA
SL
4
75
1.1
41.3
53.3
4.5
8.9
42.9
48.5
CA
SL
10
75
1.1
52,4
42.9
4.2
7.8
53.4
39.6
CA
SL
18
75
1.3
61.9
34.1
3.9
6.9
63.8
30.6
CA
SICL
4
85
12.6
53.3
30.5
3.7
CA
SICL
10
85
12.7
67.6
18.0
2.2
CA
SICL
18
85
12.0
77.0
11.2
1.2
LA
SL
10
65
0.5
52.1
44.1
3.1
2.0
57.9
39.4
LA
SL
10
80
4.2
50.9
41,6
3.1
5.1
55.9
38.3
LA
SL
10
90
15.3
47,1
34.5
3.0
15.6
49.1
34.8
LA
SICL
10
75
5.8
71.2
20.3
2.3
LA
SICL
10
85
13.5
69.6
14.5
2.2
LA
SICL
10
95
36.5
59.0
3.0
1.4
LA
SL
4
75
2.0
38.8
55.7
3.2
8.2
45.1
45.2
LA
SL
10
75
2.1
51.6
43.0
3.1
3.3
57.0
39.0
LA
SL
18
75
2.3
62.4
32.0
3.0
3.0
66.5
30.2
LA
SICL
4
85
12.4
55.6
28.8
2.9
LA
SICL
10
85
13.5
68.1
14.4
2.1
LA
SICL
18
85
14.3
75.8
8.1
1.2
* CA = Santa Maria, CA; LA = Shreveport, LA; SL = sandy loam (HELP model
default texture 8); SICL = silty clayey loam (HELP model default
texture 15). Fair grass was used for all cases.
** Change in storage is not included in this table; therefore, the water
balance components shown do not always sum to 100.0%.
t Slope -3%; drainage length = 200 ft; hydraulic conductivity =
3 x 10 cm/sec. _j
tt Hydraulic conductivity ¦ 10 ^ cm/sec,
~ Hydraulic conductivity - 10 cm/sec.
31
-------
TABLE 5. SENSITIVITY OF WATER BUDGET COMPONENTS TO DRAINABLE POROSITY AND
PLANT AVAILABLE WATER
Average Annual Volume (Percent Precipitation)**
MHaaaaaaHHaM|BM||a|i|Miaaaaaaaaaat|B|aaaH|IBMB||aaaiaai|h|MB|^^
ft
Description Hazardous Waste Design Municipal Design
Lat.t Barrierff Barrier f
Site
DP
PAW
Runoff
ET
Drng.
Perc.
Runoff
ET
Perc.
CA
0
18
0.07
1.07
48.51
46.45
4.31
8.57
49.78
42.16
CA
0
18
0.13
1.14
52.54
42.83
4.22
7.87
53.55
39.41
CA
0
18
0,20
1.30
56.43
39.43
4.12
7.06
57.18
37.02
CA
0
10
0.13
1.17
48.87
47.38
4.33
10.48
50.40
40.02
CA
0
18
0.13
1.14
52.53
42.81
4.22
7.87
53.55
39.41
CA
0
27
0.13
1.1
55.8
39.6
4.1
5.22
57.34
38.20
LA
0
18
0.07
2.08
47.38
47.12
3.12
4.36
54.57
40.08
LA
0
18
0.13
2.15
51.74
42.86
3.08
3.45
57.05
38.84
LA
0
18
0.20
2.26
55.68
38.92
3.04
2.98
59.99
36.69
LA
0
10
0.13
2.10
46.93
47.66
3.12
6.63
55.24
37.65
LA
0
18
0.13
2.15
51.74
42.86
3.08
3.45
57.05
38.84
LA
0
27
0.13
2.2
55.7
38.8
3.0
2,32
59.60
37.49
* CA = Santa Maria, CA; LA = Shreveport, LA; DP = drainable porosity (vol/
vol); PAW = plant available water (vol/vol). All cases are for sandy
loam top soil (HELP model default texture 8); fair grass; evaporative
depth = 10 in.; and curve number = 75.
** Change in storage is not included in this table; therefore, the water
balance components shown here do not always sum to 100.0%.
"f Slope = 3a; drainage length = 200 ft; hydraulic conductivity =
3 x 10 cm/sec, ^
ft Hydraulic conductivity « 10 ^ cm/sec.
f Hydraulic conductivity = 10 cm/sec.
32
-------
3 CA. MAI. , SL
—OCA. HAZ., SICL
°LA. HAZ.. SL
®LA. HAZ.. SICL
68 88
S.C.S. RUNOFF CURVE NUMBER
34
32
38
2«
26
24
22
2d
16
16
H -
12
18
e
«CA, HAZ.„ SL ,
O—^CA. HAZ . SICL
•—°LA. HAZ.. SL
®LA. HAZ.. SICL
68
S.C.S. RUNOFF CURVE NUMBER
CA. HAZ.. SL
CA. HAZ.. SICL
LA. HAZ.. SL
LA. HAZ.. SICL
T
70
CO
Ul
Z
o
z
I s
z
o
o
0 I
a,
u
a.
<
Z 8.6 -|
Z
<
1
a—"C*. HAZ.. SL
"CA. HAZ.. SICL
•—°LA. HAZ.. SL
•—°LA. HAZ.. SICL
I
00
S.C.S. RUNOFF CURVE NUMBER S.C.S. RUNOFF CURVE NUMBER
Figure 6. Effect of runoff curve number on hazardous waste cover design.
(CA = Santa Maria, California; LA = Shreveport, Louisiana;
HAZ = hazardous waste design; SL = sandy loam topsoil;
SICL = silty clayey loam topsoil)
-------
S.C.S. RUNOFF CURVE NUMBER
S.C.S. RUNOFF CURVE NUMBER
S.C.S. RUNOFF CURVE NUMBER
Figure 7. Effect of runoff curve number on municipal cover design.
(CA = Santa Maria, California; LA = Shreveport, Louisiana;
MUN = municipal design; SL = sandy loam topsoil;
SICL = silty clayey loam topsoil)
-------
® CA, HAZ., SL
®—«CA, HAZ., SICL
•—°LA, HAZ.. SL
•—•la, haz.. sicl
I
~r
4 0 8 18 12 14 10
EVAPORATIVE DEPTH (INCHES)
~r
18
94
32
30
28
26
24
22
20
16
10
14
12
10
8
O
4
CA. HAZ., SL
CA, HAZ., SICL
LA. HAZ., SL
LA, HAZ., SICL
~r
6 8 10 12 14
EVAPORATIVE DEPTH (INCHES)
a—oca, haz.. sl
«CA, HAZ.. SICL
•—°LA. HAZ.. SL
•—®LA. HAZ., SICL
<0
Ul
Z
o
z
^ I .6
z
O
O
o
a.
ui
a.
<
D
I
O—®CA,
O—OCA.
•—©LA,
•—•LA.
HAZ.. SL
HAZ.. SICL
HAZ., SL
HAZ., SICL
~v
/
/
/
/
/
EVAPORATIVE DEPTH (INCHES) EVAPORATIVE DEPTH (INCHES)
Figure 8. Effect of evaporative depth on hazardous waste cover design.
(CA = Santa Maria, California; LA = Shreveport, Louisiana;
HAZ = hazardous waste design; SL = sandy loam topsoil;
SICL = silty clayey loam topsoil)
-------
EVAPORATIVE DEPTH (INCHES)
EVAPORATIVE DEPTH (INCHE9)
U)
o—OCA. MUH.. SL
Figure 9.
4 8 8 10 12 14 10
EVAPORATIVE DEPTH (INCHES)
Effect of evaporative depth on municipal cover design.
(CA = Santa Maria, California; LA = Shreveport, Louisiana;
MUN = municipal design; SL = sandy loam topsoil)
-------
CO
uj
x
o
z
u.
0
z
3
£E
-I
<
2 0.5 -
2
<
1
«CA. HAZ
°LA. MAZ.
SL
SL
~r
~r
.BE
*.t a.ib 0 2 g.;
DRAINABLE POROSITY (VOL./VOL.)
29
26
24
22 -
20
16
le -
M
12 •
*—UCA. HAZ.. SL
•—*LA. HAZ.. SL
-1 _1 J j
)1 0 1& 0 2 8.25
DRAINABLE POROSITY
-------
DRAINABLE POROSITY (VOL./VOL.)
DRAINABLE POROSITY (VOL /VOL.)
LO
00
z
o
z
z
o
K
<
_J
o
o
cc
Ul
a.
<
3
Z
Z
<
LA. HUN.. SL
I.I Q.IG 0 2 0 25
DRAINABLE POROSITY (VOL./VOL.)
Figure 11. Effect of drainable porosity on municipal cover design.
(CA = Santa Maria, California; LA = Shreveport, Louisiana;
MUN = municipal design; SL = sandy loam topsoil)
-------
III
I
o
z
o
z
=J
s
<
3
z as
z
o
5
O—OCA, HAZ.. SL
«—®LA. HAZ., SL
1 r™ 1 1
e& 0 1 0 16 0 2 0 2S
PLANT AVAILABLE WATER CAPACITY (VOL./VOL.)
26
26
24
22 -
28
18
ie
1 4
12
10
8
e h
4
CA. HAZ . SL
LA, HAZ. „ SL
T
1
0 lb
1—
0.2
0 as 8 I B lb 0.2 e 25
PLANT AVAILABLE WATER CAPACITY (VOL./VOL.)
U>
VD
£
O
z
Ul
(9
<
z
<
cc
o
.4
<
3
z
16
I 4
12
18
e
8
»—OCA, HA7 . SL
®—0LA, HAZ .. SL
"I
a 2
(A
Ul
X
o
2
o
o
o
cc
Ul
£L
z 05
z
s
<
CA, HAZ., SL
LA. HAZ.. SL
—I—
.05
~r~
a, i
—J—
0. 16
1—
0.2
PLANT AVAILABLE WATER CAPACITY (VOL./VOL.) PLANT AVAILABLE WATER CAPACITY (VOL./VOL.)
Figure 12. Effect of plant available water on hazardous waste cover design.
(CA - Santa Maria, California; LA = Shreveport, Louisiana;
HAZ
hazardous waste design; SL = sandy loam topsoil)
-------
PLANT AVAILABLE WATER CAPACITY (VOL./VOL.)
PLANT AVAILABLE WATER CAPACITY (VOL./VOL.)
PLANT AVAILABLE WATER CAPACITY (VOL /VOL.)
Figure 13. Effect of plant available water on municipal cover design.
(CA = Santa Maria, California; LA = Shreveport, Louisiana;
MUN = municipal design; SL = sandy loam topsoil)
-------
was 10 inches in all cases. Simulations for the hazardous waste cover design
were performed for both soil types, whereas simulations for the municipal
design were performed only for sandy loam. The results are presented in
Table 4 and Figures 6 and 7.
As expected, an increase in runoff curve number produced an increase in
runoff and a decrease in evapotranspiration, lateral leachate drainage, and
barrier soil percolation. The percent increase in runoff was less for the
municipal design than for the hazardous waste design, This result was due to
the higher average moisture content in the topsoil layer of the municipal
design caused by the restriction to vertical flow imposed by the barrier soil
layer. This limited the infiltration capacity of the topsoil so that excess
surface storage accumulated and contributed to the runoff volume. Thus, run-
off volume at low curve numbers was higher for the municipal design compared
to the hazardous waste design. This effect was not as great at high curve
numbers because infiltration for both designs was significantly reduced by the
curve number itself.
The effect of location on runoff was difficult to discern from the
results presented here. This is due to two unusually large storms at
Santa Maria in January and February 1978 which biased the results. For exam-
ple, in comparing runoff from Santa Maria and Shreveport, one would expect a
smaller percentage of precipitation to leave as runoff in Santa Maria due to
the higher evapotranspiration demand combined with lower total precipitation
and long periods of time between storms. However, for the municipal design
simulations at the lower curve numbers, the influence of the two large storms
in Santa Maria caused the runoff percentage to exceed that in Shreveport.
This would not be the case if the two storms were excluded.
Evaporative Depth—
The evaporative depth was varied from 4 to 18 inches for both sandy loam
and silty clayey loam. The runoff curve number was 75 for the sandy loam and
85 for the silty clayey loam. Simulations for the hazardous waste design were
performed for both soil types, whereas simulations for the municipal designs
were performed only for sandy loam. The results are presented in Table 4 and
Figures 8 and 9.
As evaporative depth increased, evapotranspiration increased while lat-
eral drainage and barrier soil percolation decreased; the effect on runoff
varied. The interrelationship between these variables in the HELP model is
complex and depends on many factors. The increase in evaporative depth allows
evapotranspiration to deplete soil moisture from greater depths, generally
Increasing the total volume of evapotranspiration. However, since the total
evapotranspiration demand remains constant, a smaller volume of water deple-
tion occurs per unit depth. Consequently, the antecedent moisture conditions
computed by the HELP model could be higher, resulting in larger runoff using
the curve number method. The higher moisture content might also reduce the
infiltration capacity of the surface soil segment so that excess surface mois-
ture might further Increase runoff volume. However, when the time period
between storms is sufficiently long, evapotranspiration demand is able to
deplete soil moisture to equal levels in either small or large evaporative
depths. In this case, runoff volume computed by the HELP model could decrease
41
-------
with increasing evaporative depth since antecedent moisture conditions would
remain the same and the increased storage volume in the deeper evaporative
zone would increase the infiltration capacity.
The effect of evaporative depth on the volume of lateral drainage and
barrier soil percolation is directly related to the composite effect on evapo-
transpiration and runoff. In the examples chosen for Table 4 and Figures 8
and 9, the increase in evapotranspiration with increased evaporative depth was
greater than any increase in infiltration; therefore, lateral drainage and
barrier soil percolation always decreased.
An increase in evaporative depth caused an increase in infiltration for
the municipal design compared to a slight decrease for the hazardous waste
design. This difference relates to the different mechanisms controlling
infiltration in these two cases. For the municipal design, the hydraulic con-
ductivity of the clay barrier soil was much less than the sandy loan topsoil.
Therefore, infiltration tended to saturate the topsoil layer, and the total
volume of infiltration was dependent primarily on the volume of empty storage
available in this layer. A larger evaporative depth increased the potential
for a larger volume of empty storage and thus for more infiltration. For the
hazardous waste design, the lateral drainage layer generally maintained a free
outfall condition at the topsail/lateral drainage layer interface. Infiltra-
tion was then controlled primarily by the hydraulic conductivity of the top-
soil and the empty storage volume in the top segment of the subprofile. As
explained above, this condition could result in either an Increase or decrease
in infiltration with an increase in evaporative depth.
Drainable Porosity—
This term is defined as the difference between porosity and field capac-
ity; that is, the amount of water that could be vertically drained from a sat-
urated soil by gravity. Values ranged from 0.100 to 0.270 in this study.
These values represent the volume of moisture storage capacity in excess of
field capacity, divided by the bulk volume of soil including voids. Values
for field capacity and wilting point remained constant at 0.263 and 0,133,
respectively. Only sandy loam soil was considered. The evaporative depth was
10 inches, and the runoff curve number was 75. Both hazardous waste and muni-
cipal designs were simulated. The results are presented In Table 5 and Fig-
ures 10 and 11.
An Increase in drainable porosity increases the moisture storage volume
above field capacity. Therefore, more water can be made available during ver-
tical drainage for evapotranspiration. An increase in drainable porosity
(with constant field capacity) will also decrease the height to which ponded
water can rise Into the top soil layer. The net result is shown to increase
the volume of evapotranspiration and decrease the volume of lateral drainage
and barrier soil percolation in Table 5 and Figures 10 and 11. However, the
effect of increased drainable porosity on runoff is varied. For the hazardous
waste design, runoff decreased slightly at Santa Maria and increased slightly
at Shreveport. For the municipal design, runoff decreased significantly at
both locations since the relative soil moisture is lower and the available
storage is greater. At Santa Maria, where evapotranspiration demand is high
and precipitation is low, the moisture storage volume above field capacity
42
-------
tends co remain empty. An increase in this storage volume results in larger
empty storage available for infiltration. Combined with low antecedent moi-
sture conditions due to climate, the increased drainable porosity tends to
reduce runoff. At Shreveport, where unsatisfied evapotranspiration demand is
lower and precipitation is higher, the moisture storage volume above field
capacity remains partially filled for longer periods of time compared to
Santa Maria. An increase in this storage volume without an accompanying
increase in hydraulic conductivity lengthens further the time required to
drain the moisture to field capacity. Therefore at Shreveport, an increase in
drainable porosity increased the potential for higher antecedent moisture con-
ditions and thus higher runoff, which is shown in the simulation results for
the hazardous waste design at Shreveport. For the municipal design, infiltra-
tion is controlled to a larger extent by the available storage volume in the
top soil rather than its hydraulic conductivity, as explained earlier. Thus,
an increase in drainable porosity resulted in greater infiltration and less
runoff for both municipal designs.
Plant Available Water Capacity—
This term is defined as the difference between field capacity and wilting
point, or the amount of water available for plant uptake after vertical drain-
age by gravity has ceased. Values ranged from 0.070 to 0.200 in this anal-
ysis. These values represent the volume of moisture storage capacity between
wilting point and field capacity, divided by the bulk volume of soil including
voids. The values for wilting point and drainable porosity remained constant
at 0.133 and 0.180, respectively. Only sandy loam soil was considered. The
evaporative depth was 10 inches, and the runoff curve number was 75. Both
hazardous waste and municipal designs were simulated. The results are pre-
sented in Table 5 and-Figures 12 and 13.
As plant available water increases, a greater volume of water is avail-
able for evapotranspiration after vertical drainage has ceased. This results
in larger evapotranspiration losses as indicated in Table 5 and Figures 12 and
13. However, it also results in an increased moisture content at field capa-
city and therefore a greater potential for higher antecedent moisture condi-
tions. In the hazardous waste designs, this effect was apparently influencing
the small increase in runoff with increased plant available water capacity.
Runoff decreased for the municipal designs so that, as before, available stor-
age volume appeared to be controlling infiltration in these cases. From
Table 5 it is seen that the increases in evapotranspiration were great enough
to offset any increases in infiltration; therefore, leachate drainage and bar-
rier soil percolation also decreased.
Summary of Sensitivity Analysis for Landfill Cover
The interrelationship between parameters influencing the hydrologic per-
formance of a landfill cover in the HELP model is complex. It is difficult to
isolate one parameter and exactly predict its effect on the water balance
without first placing restrictions on the values of the remaining parameters.
With this qualification in mind, the following general summary statements are
made.
43
-------
The primary importance of the topsoil depth or thickness is in control-
ling the extent or existence of overlap between the evaporative depth and the
head in the lateral drainage layer. Surface vegetation has a significant
effect on evapotranspiration from soils with long flow-through travel times
(low hydraulic conductivity) and large plant available water capacities;
otherwise, the effect of vegetation on evapotranspiration is small. The gen-
eral influence of surface vegetation on lateral drainage and barrier soil per-
colation is difficult to predict outside the context of an individual cover
design. Clayey soils produce greater runoff and evapotranspiration, and less
lateral drainage and barrier soil percolation. Simulations of landfills in
colder climates and in areas of lower solar radiation are likely to show less
evapotranspiration and greater lateral drainage and barrier soil percolation.
An increase in the runoff curve number will increase runoff and decrease evap-
otranspiration, lateral drainage, and barrier soil percolation. As evapora-
tive depth, drainable porosity or plant available water increase,
evapotranspiration tends to increase while lateral drainage and barrier soil
percolation tend to decrease; the effect on runoff is varied.
LATERAL DRAINAGE AND BARRIER SOIL PERCOLATION
This analysis examines how lateral drainage and barrier soil percolation
in the HELP model are affected by the slope, drainage length, and saturated
hydraulic conductivity of the lateral drainage layer, 1C, and by the saturated
hydraulic conductivity of the barrier soil layer, K^. Two types of vertical
inflows to the lateral drainage layer were considered. First, an inflow rate
of 50 inches/year was used to represent infiltration at an open landfill.
This inflow was distributed in time according to actual rainfall patterns at
Shreveport, LA. Second, an inflow rate of 8 inches/year uniformly distributed
in time was used to represent infiltration at a covered landfill. Xn the dis-
cussion that follows, the hydraulic conductivities are first investigated by
holding the slope and drainage length constant. Then, the slope and drainage
length are examined by holding the hydraulic conductivities constant. In all
cases, the thickness of the lateral drainage layer was greater than the maxi-
mum head, and the thickness of the barrier soil layer, T , was 24 inches.
c
i.he equation used in the HELP model to cosjpute the barrier soil percola-
tion rate, Q_, is Equation 18 where the average thickness of lateral flow, y,
is used for the total head above the barrier soil layer, TH, resulting in
y + T
Qp - Kp (26)
c
The equation used to compute the lateral drainage rate, Q , is Equation 25.
The two equations are solved simultaneously to compute Qp, Q^, and y.
Hydraulic Conductivities
The combinations of and values used in this analysis are listed in
Table 6 along with resulting average annual volumes of lateral drainage and
barrier soil percolation expressed as a percentage of annual inflow. Values
44
-------
TABLE 6. SENSITIVITY OF LINER/DRAIN SYSTEM PERFORMANCE TO HYDRAULIC
CONDUCTIVITY OF LATERAL DRAINAGE LAYER AND BARRIER SOIL LAYER
Hyd. Cond.
Lat. Drug.
*D
Hyd. Cond.
Barrier Soil
Kp
Avg.
(*r
\ '©
Annual Vol.*
Inflow)
Max Head
Annual**
*D
KP
BarrierTT
in Lat.
Infilt.
Lat .t
Soil
Drng Layer
(in.)
(cm/sec)
(cm/sec)
Drug.
Perc.
(in.)
50
io"}
io'l
lot
26.16
73.84
6,3
50
10":
10 \
10,
io i
io~;?
10 *
lot
81.44
18.56
7.9
50
50
50
50
io"
io-;
io-r
10-5
10?
105
10,
104
97.44
67.35
96.36
46.79
2.56
32.65
3.64
53.21
8.1
22.9
24.8
46,4
50
10"'
io"
10"8
10
93.18
6.82
60.6
50
10
io5
99.30
0.70
62.0
8
8
10"}
10I
10_2
10 ,
io";
10"'
-5
10 Z
10"^
105
106
0.58
5.55
99.42
94.45
<0.1
<0.1
8
8
8
8
io~i
10"?
10"'
io_5
10 8
io"8
10
io:
IO:!
10f
84.37
0.58
83.56
0.06
15.63
99.42
16.40
99.94
0.2
<0.1
1.4
<0.1
8
10,
lot
77.60
21.97
10.1
8
10~3
10
97.22
2.32
12.1
* Change in storage is not included in this table; therefore, the water
balance components shown do not always sura to 100.0%.
** Value of 50 in./year represents inflow through an open landfill; the
tenporal distribution is based on rainfall records for Shreveport, LA.
Value of 8 in./year represents Inflow through landfill cover; the temporal
distribution is uniform throughout the year,
f Slope = 3%; drainage length = 75 ft; porosity = 0.351 vol/vol; field
capacity = 0.174 vol/vol.
tf For 24-in. barrier soil layer.
45
-------
-3 -1 -8
of ranged from 10 to 10 cm/sec while those of Kp ranged from 10 to
10 cm/sec, The slope was 3 percent, and the drainage length was 75 feet.
For the large inflows (50 inches/year), only three cases produced barrier
soil percolation volumes less than S^percent of the inflow, all^with barrier
soil hydraulic conductivities of 10 cm/sec or less: 1C_ - 10„ and
Kp ¦ 10 g cm/sec; 1C ¦ 10 and Kp ¦ 10 cm/sgc; and K_ ¦ 10 and
KL ¦ 10 cm/sec. unly the case with Kp = 10 restricted percolation to less
than 1 inch/year. For the small inflows (8 inches/year), only one case
produced^barrier soil percolation volumes less than 5 percent of the inflow*,
Kp = 10 and K_ = 10 ca/sec. This was the only case which restricted
percolation to less than 1 inch/year.
Figure 14 shows how the hydraulic conductivities affected lateral drain-
age and barrier soil percolation. In particular, the curves for the small g
inflows show that almost all inflow leaves as percolation at K_ values of 10
cm/sec and greater. This is consistent with Equations 25 and 26 which indi-
cate that as y approaches zero, the percolation rate approaches the hydraulic
conductivity of the barrier soil while the lateral drainage rate approaches
zero.
The effect is also seen in Figure 15 which relates the ratio to
lateral drainage and barrier soil percolation. The curves represent steady-
state analytical solutions to Equations 25 and 26, while the data points
represent results of the HELP model simulations for the 50- and 8-inches/year
inflows. The trend is for percolation to dominate as heads decrease.
Since the inflow rate of 8 inches/year was applied uniformly in time, the
maximum heads listed in Table 6 for these inflows should approximate the
steady-state values of y in Figure 15. A comparison of these maximum heads
computed by the HELP model to the position of the data points in the figure
indicates that the two are consistent.
Since the HELP model does not print values of y, heads for the
50-inches/year inflows cannot be compared in the same manner. However, in
each case the maximum head for these larger inflows is greater than the corre-
sponding y on Figure 15.
Slope and Drainage Length
The combinations of slope and drainage length used in this analysis are
listed in Table 7 along with resulting average annual volumes of lateral
drainage and barrier soil percolation expressed as a percentage of annual
inflow. The table also contains the resulting maximum heads above the barrier
soil layer. The slope ranged from 0.01 to 0.09 (1 to 9 percent) while the
drainage length ranged from 25 to 225 ft. The hydraulic conductivities of the
lateral drainage and barrier soil layers were 10~ and 10 cm/sec, respec-
tively. The product of aL and the ratio of L/a ranged from 0.25 to 6.75 ft
and 280 ft to 22,500 ft, respectively.
The .term aL is a measure of head above the drain resulting from the
sloped barrier soil layer. While an increase in head generally increases the
46
-------
•Kj
o
C
#
HI
o
<
z
<
-------
90.9
SO in/yr 8 in/yr
Figure 15. Effect of K^/K^, ratio on lateral drainage and barrier soil
percolation (drainage slope = 3%; drainage length ¦
75 ft; thickness of barrier soil layer » 24 in,).
48
-------
TABLE 7. SENSITIVITY OF LINER/DRAIN SYSTEM PERFORMANCE TO LATERAL DRAINAGE
SLOPE AND LENGTH
Avg. Annual Vol.*
(%
Inflow)
Max. Head
Annual**
Slope
Length
BarrierTT
In Lat.
Infilt.
a
L
aL
L/a
Lat.t
Soil
Drng. Layer
(in.)
(ft/ft)
(ft)
(ft)
(ft)
Drng.
Perc.
(in.)
50
0.01
25
0.25
2500
96.71
3.29
13.8
50
0.01
75
0.75
7500
95.89
4.11
29.7
50
0,01
225
2,25
22500
93.42
6.57
58.2
50
0.03
25
0.75
830
96.85
3.15
12.3
50
0.03
75
2.25
2500
96.36
3.64
24.8
50
0.03
225
6.75
7500
95.10
4.90
42.3
50
0.09
25
2.25
280
97.37
2.63
8.5
50
0.09
75
6.75
830
96.86
3.13
16.2
8
0.01
25
0.25
2500
83.71
16.27
1.2
8
0.01
75
0.75
7500
82.23
17.71
3.4
8
0.01
225
2.25
22500
78.21
21.49
9.4
8
0.03
25
0.75
830
84.16
15.84
0.5
8
0.03
75
2.25
2500
83.55
16.41
1.1
8
0.03
225
6.75
7500
82.20
17.72
3.5
8
0.09
25
2.25
280
84.35
15.65
0.2
8
0.09
75
6.75
830
84.22
15.77
0.4
* Change In storage is not included in this table; therefore, the water
balance components shown here do not always sum to 100.0 percent.
** Value of 50 in,/year represents inflow through an open landfill; the
temporal distribution is based on rainfall records for Shreveport, LA.
Value of 8 in./year represents inflow through landfill cover; the temporal
distribution is uniform throughout the year,
t Hydraulic conductivity = 10 cm/sec; porosity • 0.351 vol/vol; field
capacity = 0.174 vol/vol. _j
tt For 24-in. barrier soil layer; hydraulic conductivity = 10 cm/sec.
49
-------
lateral drainage, an increase in aL resulting from an increase in drainage
length does not increase lateral drainage because the lateral drainage is also
a function of 1/L , as seen in Equation 25. An increase in aL resulting from
an increase in slope does increase the lateral drainage. At constant values
of aL, lateral drainage decreases as the drainage length increases.
For a low, uniform infiltration rate, y remains small so that from Equa-
tion 21, h Is approximately equal to aL. Therefore using Equations 20, 21
and 22, the steady-state average head in the lateral drainage layer was lin-
early related to L/a as follows;
(—) (k)
- . \V W (27)
y 1.02 + 0.0041 aL
where y and L are in inches and and IC have the same units. The results
for the 8-inches/year infiltration rate listed in Table ? are plotted in Fig-
ure 16 to show this relationship. With unsteady infiltration the maximum
average head in the lateral drainage layer was also a function of L/a, but the
relationship was no longer linear due to the storage available in the lateral
drainage layer.
Percolation through barrier soil increases with drainage length and
decreases with slope because the lateral drainage rate decreases. As a result
of the decrease in drainage rate, the head increases and is maintained at
greater depths for longer periods of time. Consequently, the percolation
increases as predicted by Equation 26. Since percolation is a function of y
and y is a function of L/a as shown in Equation 27, percolation increases
linearly with increases in L/a. The percolation results listed in Table 7 are
plotted in Figure 17 to show the effect of L/a on percolation.
Lateral drainage increases with slope and decreases with drainage length
in the opposite manner that percolation changes. The decrease in the net lat-
eral drainage is a result of the increase in percolation. In the absence of
significant increases in percolation, the head increases to yield only small
changes in the lateral drainage rate when large changes in slopes and drainage
lengths are used.
Summary of Sensitivity Analysis
for Lateral Drainage and Percolation
The sensitivity analysis shows that the ratio of lateral drainage to per-
colation is a positive function of the ratio of K^/K- and the average head
above the liner. However, the average head is a function of /K_ and L/a.
The quantity of lateral drainage, and therefore also the average head, is in
turn a function of the infiltration. Therefore, the ratio of lateral drainage
to percolation increases with increases in infiltration, and the ratio of
Kp/Kp for a given drain and liner design. The ratio of lateral drainage to
percolation for a given ratio of SC/iL, increases with increases in infiltra-
tion and decreases in L/a. The percolation and average head above the liner
are positive functions of the term L/a.
50
-------
10
8 IN/YR
«
«
JZ
o
c
~
<
Ui
X
Ul
©
<
E
Ul
6-
_t—
;n
¥"
~l—
23
25
L/oc (1000 feet)
Figure 16. Effect of L/a on the steady-state and maximum
average head above the liner.
i
o
a.
a
o
o
-------
SECTION 5
REVIEW OF LANDFILL DESIGN REGULATION AND GUIDANCE
This section evaluates the Resource Conservation and Recovery Act (RCRA)
landfill design regulation, as amended by the Hazardous and Solid Waste Amend-
ment s of 1984 (HSWA) (10), and Minimum Technology Guidance (KTG) (II) based on
the results of the laboratory and field verification studies and the sensitiv-
ity analysis. This evaluation examines whether the guidance is practicable
and achieves its objectives in the best practicable manner. The evaluation
examined guidance on final covers, liners and leachate detection, collection
and removal systems. Evaluation of the final cover guidance included effects
of vegetation, vegetated cover soil thickness, drainage layer design and liner
design. Evaluation of liners included effects of liner thickness and
hydraulic conductivity. Evaluation of leachate detection, collection and
removal systems examined hydraulic conductivity, slope and maximum drainage
length of the drainage layer. The goal of the designs is to minimize the
potential percolation through the liner at the base of the landfill.
REGULATION
New and laterally expanded RCRA landfills must have a double liner system
that is designed, constructed, and installed to prevent any migration of
wastes out of the landfill and into or out of the space between the two lin-
ers, The liners must be constructed of materials that prevent wastes from
passing out of the liners during the operating period of the facility. A leak
detection system must be designed, constructed, maintained, and operated
between the liners to detect any migration of liquid into the space between
the liners. The landfill must have a leachate collection and removal system
above the top liner to ensure that the leachate depth over the liner does not
exceed 30 cm (1 foot) (10). The Regional Administrator will specify design
and operating conditions in the permit to ensure that the leachate depth over
the liner does not exceed 30 cm (1 foot). The owner or operator must design,
construct, operate, and maintain a runon control system capable of preventing
flow on to the active portion of the landfill during peak discharge from at
least a 25-year storm and a runoff management system to collect and control at
least the water volume resulting from a 24-hour, 25-year storm. The Regional
Administrator will specify in the permit all design and operating practices
that are necessary to ensure that the requirements of this regulation are sat-
isfied (10).
At final closure of the landfill or any cell, the owner or operator must
cover the landfill or cell. The cover must be designed and constructed to
provide long-term minimization of migration of liquids through the closed
landfill, function with minimum maintenance, promote drainage and minimize
52
-------
erosion of Che cover, accommodate settling and subsidence, and have a perme-
ability less than or equal to the permeability of any bottom liner system.
GUIDANCE
RCRA Minimum Technology Guidance on Double Liner Systems (11) specifies
designs for liners, and leachate detection, collection and removal systems.
The RCRA Guidance Document for Landfill Design (12) specifies designs for
final covers. The guidance for liners recommends that the liners be con-
structed wholly above the seasonal high water table. The double liner system
should consist of a primary leachate collection and removal system; a flexible
membrane liner (FML) for the top liner; a secondary leachate collection and
removal system; and a low permeability soil liner or a composite (FML plus low
permeability soil) for the bottom liner.
The flexible membrane liner should be chemically resistant to the waste
and leachate. The FML should have a thickness of at least 30 mils; greater
thicknesses are necessary if the liner will be exposed to weather for an
extended period or unusual stresses during installation and operation. The
FML should be protected from damage from above and below by at least 12 inches
of bedding material that is no coarser than sand (soil type SP using the Uni-
fied Soil Classification System). The bedding material must be smooth and
free of rock, fractured stone, debris, cobbles, rubbish, and roots (11).
The low permeability soil liner should_^ave an in-place saturated
hydraulic conductivity not more than 1 x 10 cm/sec and a thickness of not
less than 90 cm (3 feet). The soil must be free of large clods, rock, frac-
tured stone, debris, cobbles, rubbish, and roots that would increase hydraulic
conductivity or serve to promote preferential leachate flow paths (11).
The primary leachate collection and removal system should have at least a
30-cm (1 foot) chemically^resistant drainage layer with a hydraulic conductiv-
ity not less than I x 10 cm/sec and a minimum bottom slope of 2 percent. A
graded granular or synthetic fabric filter should be placed above the drainage
layer to prevent clogging. A drainage pipe system of appropriate size and
spacing and a sump pump should be installed to efficiently remove leachate
from the drainage layer. The leachate collection system should cover the
entire bottom and sidewalls of the landfill unit (11).
The secondary leachate collection and removal system should permit rapid
detection, collection, and removal of any migration of liquid into the space
between the liners. The drainage layer should be chemically resistant to the
was|e and leachate, have a hydraulic conductivity not less than I x
10 cm/sec and a minimum bottom slope of 2 percent. A drainage pipe system
of appropriate size and spacing and a sump pump should be installed to effi-
ciently remove leachate from the drainage layer. The leachate collection sys-
tem should cover all areas between the double liners likely to be exposed to
waste and leachate (11).
The final cover should have a vegetated top cover, a middle drainage
layer and a liner as a minimum. The vegetated top cover should be at least
60 cm (2 feet) thick and support persistent, shallow-rooted,
53
-------
erosion-controlling vegetation. The cover should have a slope of 3 to 5 per-
cent to promote runoff collection without causing excessive erosion. The
drainage layer should be at least 30 en (1 footl thick with a saturated
hydraulic conductivity of not less than 1 x 10~ cm/sec and have a bottom
slope of at least 2 percent. Lateral drainage should be collected from the
drainage layer. The liner should be composed of at least a 20-mil synthetic
membrane overlying a 60-cm (2-foot) layer of compacted clay soil having a
hydraulic conductivity of not more than 1 x 10 cm/sec (12),
EVALUATION AND RECOMMENDATIONS
Liner Systems
Results of the studies show that soil (clay) liners restrict percolation
and migration of liquids from landfills but do not prevent it. Small quanti-
ties of percolation occur even with an overdesigned leachate collection and
removal system and a very low permeability soil liner according to the HELP
model simulations. Therefore, the use of synthetic membranes in conjunction
with soil liners appears prudent. Field data were not available to verify
this prediction.
Two design specifications are used for soil liners: thickness and satu-
rated hydraulic conductivity. Percolation is primarily controlled by the sat-
urated hydraulic conductivity and not thickness. Very large changes in
percolation rates can be obtained by the materials and construction methods
selected for the soil liner while only small improvements can be achieved by
using very thick liners. The thickness of the liner should be controlled by
the requirements for uniformity, physical strength and subsidence, A thick-
ness of 3 feet should be more than adequate to meet these requirements.
Based on field observations, the MTG (11) recommending that soil liners
have saturated hydraulic conductivities not more than 1 x 10 cm/sec is very
reasonable and practicable. Most landfill cells used in this study had
hydraulic conductivities that were very close to the guidance.
Use of low permeability soil for the bottom liner instead of a composite
liner may defeat the leak detection capabilities of the secondary leachate
collection and removal system. In addition, it would consequently violate the
intent of the HSWA (10) to prevent migration of any hazardous constituents
from the facility and the design criterion of the MTG (11) to minimize the
migration of any hazardous constituent through the bottom soil liner. The
HELP model predicts that small leakage from a FML would all pass through the
soil liner without detection if the leakage rate is less than the saturated
hydraulic conductivity of the soil liner. This would occur because sufficient
head would not build up in the secondary leachate collection and removal sys-
tem to divert the leakage to the sump.
Leachate Collection and Removal System
3y regulation (10), the leachate depth over the liner must not exceed 30
cm (1 foot), but based on HELP model simulations the existing technical guid-
ance does not appear to satisfy this during the active life of a landfill
54
-------
cell. The capability to maintain the leachate depth below 30 cm (1 foot)
under all conditions except rare storms requires extraordinary designs. The
drainage la^r must have very high saturated hydraulic conductivity, not less
than 2 x 10 cm/sec. The slope of the drainage layer base must be at least
10 percent, and the drain pipe spacing should be not more than 50 feet. With
this design the system should be able to accommodate infiltration rates up to
4 inches/day (a 1-year return period event in some regions).
Saturated hydraulic conductivity of the drainage media, slope of the
drainage layer base and drain spacing are the only design parameters of the
leachate collection and removal system that significantly affect percolation.
Increasing the leachate collection or lateral drainage rate reduces percola-
tion by lowering the leachate depth or head controlling the percolation rate
and reducing the quantity of leachate available for percolation. In practical
terms, slopes greater than 5 percent and drain spacings less than 50 feet are
not desirable. Therefore, only selection of drainage media having very high
hydraulic conductivity can significantly increase the drainage rate and lower
the leachate depth without adverse practical, consequences. Sands having
hydraulic conductivity greater than 2 x 10 cm/sec are commonly available as
are gravels and plastic drainage nets. Sands used as drainage media in the
laboratory and field studies had hydraulic conductivities that were much
greater than the recommended value of 1 x 10 cm/sec. Consequently, higher
hydraulic conductivities should be recommended.
Final Cover
Final covers affect percolation through the base of the landfill by con-
trolling the infiltration of water into the waste. The cover is composed of a
vegetated top cover and a liner and lateral drainage system that is comparable
to the bottom liner and primary leachate collection and removal system dis-
cussed above. The evaluation and recommendations presented above for those
systems also apply to the final cover. The remaining component of the final
cover to be evaluated for its effect on percolation is the vegetated top
cover.
The HELP model simulations Indicate that the vegetated top cover has very
little effect on percolation through the final cover. Decreases in runoff or
increases in infiltration are offset by Increases in evapotranspiration and
lateral drainage, yielding only small effects on percolation.
Summary
Saturated hydraulic conductivity is the most important design parameter
for minimizing percolation. Care should be taken to recommend the highest
hydraulic conductivity that is commonly available for drainage media. Simi-
larly, the lowest saturated hydraulic conductivity practically obtainable
should be selected as guidance for soil liners. Changes in other design
parameters yield much smaller effects on percolation if kept within a prudent
range. The bottom liner should include a FML to minimize potential migration
from the facility in the event that the top liner leaks. A composite bottom
liner also ensures that leakage from the top liner will be detected,
-------
SECTION 6
SIMULATION OF UNIVERSITY OF WISCONSIN-MADISON
LYSIMETER CELLS
From 1970 to 1977, eight large lysimeter cells filled with either
shredded or unprocessed refuse were monitored for surface runoff and leachate
production at the University of Wisconsin-Madison (13), The general purpose
of the study was to determine the effect on landfill performance of shredding
the refuse prior to placement and covering the refuse with a soil layer. The
HELP model was used to simulate the performance of these cells using climato-
logical data recorded near the site. Comparisons are shown between model pre-
dictions and field measurements.
SITE DESCRIPTION
The test site was located in Madison, WI, where the average temperature
is 45°F and the average annual precipitation is 31 inches. The minimum daily
temperature falls below freezing on 163 days per year. The average daily
solar radiation is approximately 330 langleys.
Each cell was 60 feet long by 30 feet wide and had depths of 4 to 8 feet.
Cells were underlain with a 4—inch layer of crushed granite over a 6-mil poly-
ethylene barrier. Bottom slopes of approximately 3 percent directed leachate
to a collection box at the center of the cell. The collection box was peri-
odically pumped, and the leachate volume was measured. The top surface of
each cell was sloped at 3 percent toward one of the 60-foot walls, A collec-
tion trough along this wall directed runoff to a tank outside the cell where
volume measurements were made. The depth of refuse, presence of soil cover,
and condition of refuse (shredded or unprocessed) varied within each cell.
Cell characteristics and dimensions are summarized in Table 8 and Figure 18.
SELECTION OF MODEL INPUT VALUES
The HELP simulations in this analysis used daily precipitation and mean
monthly temperature values taken from the National Oceanic and Atmospheric
Administration (NOAA) weather station at Madison, WI. Solar radiation values
used were the default values for Madison incorporated in the HELP model.
Other parameters were assigned values based on descriptions of the lysimeters
and are summarized in Table 9 (13). As indicated in this table, covered and
uncovered cells were modeled separately. However, no attempt was made to dif-
ferentiate between shredded and unprocessed waste.
56
-------
TABLE 8, UNIVERSITY OF WISCONSIN-MADISON CELL CHARACTERISTICS
Cell
Depth
(ft)
Cover
Refuse
Type
Period
of Test
Cell 1
4
6-in. Soil
Layer
Unprocessed
Sep 1970-
May 1976
Cell 2
4
6-in. Soil
Layer*
Shredded
Sep 1970-
May 1976
Cell 3
4**
6-in. Soil
Layer
Shredded
Sep 1970-
June 1977
Cell 4
4**
No Cover
Shredded
Sep 197 0—
June 1977
Cell 5
4
No Cover
Unprocessed
Aug 1972-
June 1977
Cell 6
4
No Cover
Shredded &
Unprocessed 1"
Aug 1972-
June 1977
Cell 7
8
No Cover
Shredded
Aug 197 2—
June 1977
Cell 8
8
6-in, Soil
Layer
Unprocessed
Aug 1972-
June 1977
* Covered 6 months after initial placement. Covered again in July 1975
immediately following the placement of an additional 4 feet of refuse.
** Depth increased to 8 ft. in July 1975.
t 66 tons of unprocessed refuse covered with 30 tons of shredded refuse.
57
-------
m
"1
leachate
collector
£D"«J
60'
A
o
m
6' soil cover
4* deep
crushed granite
4' waste layer
3% slope (approx)
1 ^
8-mil polyethylene
SECTION A-A
surface runoff collection trough
/_ ¦«— 3% slope
3% slope
(approx)
u
SECTION B-B
Figure 18. Cell dimensions for University of Wisconsin-Madison cells.
58
-------
TABLE 9. INPUT DATA FOR SIMULATION OF UNIVERSITY OF WISCONSIN-
MADISON CELLS*
Parameter
Covered Cell
Uncovered Cells
No. of layers
4
4
Layer 1
Thickness (in.)
6
6
Layer type
1
I
Soil texture
9
19
Is layer compacted?
Yes
Partial compaction**
Layer 2
Thickness (in.)
48 or 96
42 or 90
Layer type
4
4
Soil texture
L9
19
Layer 3
Thickness (in.)
4
4
Layer type
2
2
Soil texture
1
1
Is layer compacted?
No
No
Layer 4
Thickness (in.)
1
1
Layer type
5
5
Soil texture
21
21
Liner leakage fraction
0
0
Type of vegetation
Poor
Fair
SCS runoff curve no.
84.4
79
Evaporative depth (in.)
8
12
Surface area (sq ft)
1800
1800
Slope of lateral drainage (%)
3
3
Drainage length (ft)
25
25
* Input data terminology defined in the HELP model documentation (1) and
user's guide (2).
** Partial compaction was assumed for the uncovered cells due to effects of
weathering at the surface. This was simulated by using a hydraulic
conductivity equal to the default value for soil texture 19 divided by
5.0, Then, this resulting hydraulic conductivity underwent the normal
adjustment for type of vegetation.
59
-------
Since hydraulic parameters for the soil and waste were not available,
default characteristics were used. Soil texture 9 was chosen for the soil
cover, described as sandy silt. Soil texture 19 was chosen to describe the
waste.
The vegetation on the surface of the cells was described to be mixed vol-
unteer vegetation, comparable to meadow grass, which became established on
both covered and uncovered cells over a several-year period (13). This vege-
tation grew more quickly and more densely on the uncovered cells; therefore,
"fair grass" was chosen to describe the vegetation on the uncovered cells and
"poor grass" was chosen for the covered cells. Corresponding evaporative
depths were set at 12 and 8 inches, respectively, to be consistent with (yet
somewhat less conservative than) the suggested values in the HELP model.
Default leaf area index values and winter cover factors presented in the HELP
model User's Guide were adopted for these two types of vegetation (2).
SCS runoff curve numbers were selected by using observed precipitation
and runoff data for several storms to estimate separate curve numbers for the
covered and uncovered cells. Since only monthly runoff data were available,
this method was expected to produce only approximate values. The results were
generally consistent with the minimum infiltration rate method presented in
the HELP model user's guide (2). Calibration was required since curve numbers
are not available in the literature for municipal waste. Calibration in this
manner reduces the validity of the verification of the runoff component, but
only a small portion of the data was used for calibration. Also, the calibra-
tion procedure did not use the HELP model or the SCS curve number procedure as
applied in the HELP model. Therefore, the value of this verification study
should not be significantly diminished by this calibration.
The soil cover of the covered cells was assumed to be compacted based on
a description of cell construction activities. The top layer of the uncovered
cells was assumed to be partially compacted due to weathering at the surface,
thus influencing the hydraulic conductivity as outlined in Table 9.
RESULTS OF MODEL SIMULATIONS
The simulation period for the two HELP runs began with January 1970,
whereas the period of field monitoring began in September 1970 for Cells 1
through 4 and August 1972 for Cells 5 through 8, This simulation period did
not exactly match the period of field monitoring since HELP must start at the
beginning of a calendar year. Also, this initial simulation period and the
first 2 years of field monitoring were treated as periods of equilibration for
both the HELP model and the test cells and therefore are not plotted in the
accompanying figures. It was assumed that this period of equilibration was
necessary to bring internal moisture to normal levels and to establish a
weathered surface on the uncovered cells and volunteer vegetation on all the
cells.
The simulation results are presented in the form of cumulative and
monthly comparisons between model predictions and field measurements for the
following components of the water balance: (a) runoff, (b) evapotranspiration
60
-------
plus change in landfill moisture storage (ET+DS), and (c) leachate drainage.
Field "measurements" of ET+DS were actually computed values based on the fol-
lowing relationship:
ET+DS - PRCP - RNF - DRG (28)
where PRCP = measured precipitation, RNF = measured runoff, and DRG = measured
leachate drainage. It was assumed that all leachate was recovered by lateral
drainage due to the presence of an impervious synthetic liner below the lat-
eral drainage layer.
Comparisons of accumulated runoff, ET+DS, and leachate drainage for each
individual cell are plotted in Figures 19 through 26. The sane curves (for
the period of overlapping records) are plotted together in Figure 27 for
covered cells and in Figure 28 for uncovered cells. Monthly comparisons of
runoff, ET+DS, and leachate drainage for each individual cell are plotted in
Figures 29 through 36. Tables 10 and 11 summarize the differences between
measured and predicted results.
For the covered cells, the simulations overpredicted cumulative runoff
for two cells and underpredicted cumulative runoff for the remaining two
cells. The runoff simulations generally compare well prior to a large storm
in July 1975. As seen In Table 10, measured runoff accounted for an average
of 7.6 percent of the precipitation from the covered cells. The standard
deviation, s, of the measured runoff fraction was 2.2 percent. The HELP model
predicted 8.1 percent, yielding an error of 0.5 percent (s = 2.6 percent) of
the precipitation or a relative error of 6.8 percent of the measured runoff.
The measured ET+DS was 68.9 percent (s = 3.1 percent) of the precipitation,
while the predicted value was 67.1 percent. The model underpredicted by
1.8 percent (s =* 3.3 percent) of the precipitation or 2,6 percent of the mea-
sured ET+DS. Measured leachate drainage accounted for 23.5 percent
(s - 4,6 percent) of the precipitation while the model estimated 24,8 percent.
The average error was 1.4 percent (s - 3.7 percent) of the precipitation, and
the relative error was 6.0 percent of the measured leachate. In all cases the
mean error in the HELP model predictions was significantly less than the stan-
dard deviation of the measured values.
Simulation of the uncovered cells was not as successful, particularly for
leachate drainage. As seen in Table 11, measured runoff accounted for
3.4 percent (s = 0.5 percent) of the precipitation, while the HELP model esti-
mated 2.7 percent. The model underestimated by 0.8 percent (s = 0.5 percent)
of the precipitation, yielding a relative error of -23.5 percent of the mea-
sured runoff. The measured ET+DS was 83.9 percent (s = 9.1 percent) of the
precipitation, while the predicted value was 78.0 percent. The model under-
predicted by 5.9 percent (s = 10.1 percent) of the precipitation or 7.0 per-
cent of the measured ET+DS. Leachate drainage accounted for 12.7 percent
(s = 9.4 percent) of the precipitation. This large standard deviation shows
the highly variable nature of uncovered waste. The model predicted 19.4 per-
cent, producing an error of 6.7 percent (s « 10.2 percent) of the precipita-
tion and a relative error of 52.8 percent of the measured leachate. For both
covered and uncovered cells, the difference between predicted values and
61
-------
12.5 -
io.o 4
3 7-5l
2 p Q i
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
< 30
Q 20 -5
JAN
1976
DATE
Figure 19. Field measurements for Cell 1 compared to HELP simulation for
covered cells; cumulative comparisons.
62
-------
LEGEND
12.5
¦a HELP SIMULATION
- FIELD MEASUREMENT
z
u: 10.0
li-
es
H—
<
I
Z3
2
JAN
JAN
JAN
JAN
1973 1974 1975 1976
DATE
Figure 20, Field measurements for Cell 2 compared to HELP simulations for
covered cells; cumulative comparisons.
-------
110.0
LEGEND
HELP SIMULATION
- FIELD MEASUREMENT
s 2.5
125-
s 25
JAN
1975
DATE
Figure 21. Field measurements for Cell 3 compared to HELP simulations for
covered cells; cumulative comparisons.
64
-------
LEGEND
9—a HELP SIMULATION
• FIELD MEASUREMENT
z
5-
¦ I
li-
CS
z
o
125
>
=3
O
i 1
O
<
2
30 4
<
oc
o
i—
<
5
o
JAN
JAN
JAN
JAN
JAN
1973 1974 1975 1976 1977
DATE
Figure 22. Field measurements for Cell 4 compared to HELP simulation for
uncovered cells; cumulative comparisons.
65
-------
z" 2.0 j
u_"
u_
§ 1.5
ZZD
az
£ 1.0
0.5
o
LEGEND
-a HELP SIMULATION
-• FIELD MEASUREMENT
12.5
< 7.5
> 5.0
JAN
1976
DATE
Figure 23. Field measurements for Cell 5 compared to HELP simulation for
uncovered cells; cumulative comparisons.
66
-------
2.51
U-- 2.0 4
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
CO
o
3D
O
. 12.5
z
S 10.0
<£
z
ce
o
3
=>
o
JAN
JAN
1975 1976
DATE
Figure 24. Field measurements for Cell 6 compared to HELP simulations for
uncovered cells; cumulative comparisons.
67
-------
. 2.5
5*
t 2,0 1
ZD
= 1.5
l i i
>
5 1-0 -i
ZD
O
0.5
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
¦i mjm-
. 60-
z
od 50
Q
fc 40
UJ
= 30
<
= 20
S 10
JAN
1976
DATE
Figure 25, Field measurements for Cell 7 compared to HELP simulation for
uncovered cells; cumulative comparisons.
68
-------
LEGEND
o
9 HELP SIMULATION
• FIELD MEASUREMENT
c/> 40
a
CO
c
-------
LEGEND
HELP SIMULATION
-CELL 1 MEASUREMENT
-• CELL 2 MEASUREMENT
CELL 3 MEASUREMENT
- CELL 8 MEASUREMENT
CO
Q
• 15.0
Z.
3 12-5
z 10.0
<
rr
Q
JAN
JAN
1975 1976
DATE
Figure 27. Field measurements for Cells 1, 2, 3 and 8 compared to HELP
simulation for covered cells.
70
-------
LEGEND
2.5
fe 2.01
z
=3
« 1.5
yj
s
< 1.0 ^
0.51
04-
- HELP SIMULATION
- CELL 4 MEASUREMENT
-• CELL 5 MEASUREMENT
- CELL 6 MEASUREMENT
- CELL 7 MEASUREMENT
60
^ 50
cn
® 40
UJ
w 30
< 20
s 10
o
0
< 7.5
DATE
Figure 28. Field measurements for Cells 4, 5, 6 and 7 compared to HELP
simulation for uncovered cells.
71
-------
3,0'
S 2.5'
llT
£2.0^
z
QC 1,3-
>-
x 1.0-3
1 0.5'
s
0'
LEGEND
-3 HELP simulation
- FIELD MEASUREMENT
JjLI
iii
9 '
Ttlt.i T f X t J a -
i
i-~
z
o
34
s
3
2-|
3
3
0 *-
f D
i
*11-
IllITffl,
¥
All
J
t%
I '
1 i
JAN
1973
JAN
1974
11 tT??.?TuI
JAN
1975
JAN
1976
DATE
Figure 29. Field measurements for Cell 1 compared to HELP simulation for
covered cells; monthly comparisons.
72
-------
LEGEND
¦a HELP SIMULATION
FIELD MEASUREMENT
= 1,5
u
CD
<
z
-
1
X
4
3
24
9
i
il
9
i
i
i
i
l
I
$
i
i
i
i
i tfe
9
1
i
MlllllnrAjltlk.TTJlff
JAN
1973
JAN
1974
JAN
1975
JAN
1976
DATE
Figure 30. Field measurements for Cell 2 compared to HELP simulation for
covered cells; monthly comparisons.
73
-------
LEGEND
3.0 -1--^ HELP simulation
«J—• FIELD MEASUREMENT
• C.D ~i
z 2.0
CE
>- 1.5
0
5
S 4
CO
Q o
4- W
¦
2
>-
—I
zc
1
z
o
0
5
i a
JAN JAN JAN JAN JAN
1973 1974 1975 1976 1977
DATE
Figure 31. Field measurements for Cell 3 compared to HELP simulation for
covered cells; monthly comparisons.
74
-------
1.25
~ 1.00 -I
u_
§: 0 75 -=
ZD
rr
i 0.50 4
3=
| 0.25 \
0
LEGEND
-8 HELP SIMULATION
- FIELD MEASUREMENT
I2J1
nwi
* I
i I
J_lllfltii.li . t.lLllfjk ,[.t , rf.r
CO
o
<
ac
Q
4
3
2
p
w
i
gjQj
B I
i i
II
9
II
Oil
Jl
mmm'
til
iUJL
iii
9
¦rfit'f'tl
JAN
1975
DATE
JAN
1976
JAN
1977
Figure 32.
Field measurements for Cell 4 compared to HELP simulation for
uncovered cells; monthly comparisons.
75
-------
1.3
-1.0
Li-
Li-.
O
z5 0-8
QC
5 0.5
zsz
h—
Z
S 0.3
CO
o
+
LEGEND
—a HELP SIMULATION
— FIELD MEASUREMENT ?
I
•1 X—Ji iL—ii *—x 1—£
: 1 1
o
<
z
<
QC
Q
4
3
1
0
I , * T i
CD
I
m
i
> » t
t
i
_5L_j3Bl_
? to
s
_L
JAN
1975
JAN
1976
DATE
Figure 33. Field measurements for Cell 5 compared to HELP simulation for
uncovered cells; monthly comparisons.
76
-------
1.3
ti-
ll.
o
z
=3
oc
>-
1.0
0.8
_ 0.5
P
z
S 0.3
z
CO
a
+
o
<
z
<
cc
a
>-
—j
3=
4
3
2
1
0
LEGEND
1 -—3 HELP SIMULATION
! • FIELD MEASUREMENT
B
a
: i
; i
j i
: i
- ; i j 11. x! ? I
lii:,
. ll A
' i i
9
I
I
¦? I
$ ^ A.. ? * ^ i T f it
¦ ? b.
* ' *
JAN
1975
JAN
1976
DATE
Figure 34, Field measurements for Cell 6 compared Co HELP simulation for
uncovered cells; monthly comparisons -
77
-------
1.25-g
z
rtitjrr
O
o
u_
3
o
z
0.75 J
=3
¦
cr
:
>-
0.50 -1
X
o
0.25 -1
2
:
0 i
LEGEND
HELP SIMULATION
—• FIELD MEASUREMENT ?
f
X
JUL
I i "
iLu»j* I ? X. IXaxZ«l»iX< ^
to
Q
>-
JZ
o
4
3-
2 -
1 -i
11
I
-qB I
? T i ~
B
f
-1
CD
<
C
cc
o
>¦
i
¦JZ
i—
Z
o
2
.itII
<*
i
i i
-------
LEGEND
-o HELP SIMULATION ?
FIELD MEASUREMENT
,& l
S 2 5
<
| 2.0
cc
Q
1.5
1.0
I 0-5
0
JAN
JAN
JAN
1975 1976 1977
DATE
Figure 36. Field measurements for Cell 8 compared to HELP simulation for
covered cells; monthly comparisons.
79
-------
TABLE 10. DIFFERENCE BETWEEN CUMULATIVE HELP MODEL PREDICTIONS
AND CUMULATIVE FIELD MEASUREMENTS FOR COVERED CELLS AT
UNIVERSITY OF WI SCON'S IN-MAD I SON
Runoff
Cell No.
Measured
(% precip)
Predicted
(% precip)
Error
(% precip)
%
1
5.4
8.5
3.1
56
2
6.0
8.5
2.5
42
3
9.1
7.5
-1.6
-18
8
9.8
8.0
-1.8
-18
Mean
7.6
8.1
0.5
15
Std. Dev.
2.2
0.5
2.6
39
Mean Error as Percent
of Mean Measured Runoff ¦ 6.8%
Range in Measured Runoff as Percent
of Mean Measured Runoff = 71% to 129%
ET+DS
Measured
Predicted
Error
Cell No.
(% precip)
(% precip)
(% precip)
%
1
67.7
65.3
-2.4
-3.5
2
67.1
65.3
-1.8
-2.7
3
73.6
68.0
-5.6
-7,6
8
67.4
69.9
2.5
3.7
Mean
68.9
67.1
-1.8
-2.5
Std. Dev.
3. 1
2.2
3.3
4.7
Mean Error as Percent
of Mean Measured ET+DS = -2.6%
Range in Measured ET+DS as Percent
of Mean Measured ET+DS * 97% to 107%
Drainage
Measured Predicted Error
Cell No.
(% precip)
(% precip)
(% precip)
%
1
26.9
26.3
-0.6
-2.2
2
26.9
26.3
-0.6
-2.2
3
17.3
24.5
7.2
41.6
8
22.7
22.1
-0.6
-2.7
Mean
23.5
24.8
1.4
8.7
Std. Dev.
4.6
2.0
3.7
22.0
Mean Error as Percent
of Mean Measured Drainage ¦ 6.0%
Range in Measured Drainage as Percent
of Mean Measured Drainage = 74% to 114%
80
-------
TABLE II. DIFFERENCE BETWEEN CUMULATIVE HELP MODEL PREDICTIONS AND
CUMULATIVE FIELD MEASUREMENTS FOR UNCOVERED CELLS
AT UNIVERSITY OF WISCONSIN-MADISON
Runoff
Measured
Predicted
Error
Cell No.
(% precip)
(% precip)
(% precip)
%
4
3.6
2.2
-1.4
-39
5
3.4
3.0
-0.4
-12
6
3,9
3.0
-0.9
-23
7
2.8
2.4
-0.4
-14
Mean
3.4
2.7
-0.8
-22
Std. Dev.
0.5
0.4
0.5
12
Mean Error as Percent
of Mean Measured !
Runoff - -23.5%
Range in Measured
Runoff as Percent
of Mean Measured
Runoff = 82% to 115%
ET+DS
Measured
Predicted
Error
Cell No.
(% precip)
(% precip)
(% precip)
%
4
79.2
75.3
-3.9
-5
5
90. L
77.8
— 12.3
-14
6
92.7
77.8
-L4.9
-L6
7
73.5
81.0
7.5
10
Mean
83.9
78.0
-5.9
-6
Std. Dev
9.1
2.3
10.1
12
Mean Error as Percent
of Mean Measured
ET+DS = -7.0%
Range in Measured ET+DS as Percent
of Mean Measured
ET+DS = 88% to 110%
Drainage
Measured
Predicted
Error
Cell No.
(% precip)
C% precip)
(% precip)
%
4
17.2
22.5
5.3
31
5
6.5
19.2
12.7
195
6
3.5
19.2
15.7
449
7
23.7
16.6
-7.1
-30
Mean
12.7
19.4
6.7
161
Std. Dev
9.4
2.4
10.2
214
Mean Error as Percent
of Mean Measured Drainage = 52.8%
Range in Measured Drainage as Percent
of Mean Measured Drainage = 28% to 187%
81
-------
actual measurements was generally equal to about one-half the standard devia-
tion of the measured values, Indicating fairly good agreement.
The variability between the actual measurements and the HELP predictions
could be influenced by a number of factors in the field. Errors were present
in the field measurements of runoff and leachate drainage due to short periods
of pump and runoff system malfunction in the springs of 1973, 1974, and 1975.
Also, errors may have been introduced by using precipitation measurements col-
lected approximately 2 miles from the test site. This might explain the large
discrepancy during July 1975 seen in both the cumulative and monthly runoff
plots. Based on a comparison with nearby weather stations, the major storm
responsible for this runoff was highly localized, so that the measured rain-
fall may not accurately represent rainfall at the test site. Other uncertain-
ties were introduced in selecting parameter values to describe the extent of
vegetative growth, leaf area index, winter cover factor, evaporative depth,
soil and waste characteristics and degree of compaction.
As previously discussed, a significant amount of variability existed
within each set of covered and uncovered cells. The major difference in cell
construction within each set was depth of the waste layer and refuse condi-
tioning (shredded or unprocessed). The HELP model indicated that a negligible
difference in long-term water balance would result from a change in waste
layer depth. Similarly, the effect of refuse conditioning on the long-term
water balance was not expected to be significant, so the model did not attempt
to differentiate between these two types of waste, Other factors that could
influence this variability include random nonuniformities in soil character-
istics, degree of compaction, surface weathering, and vegetative growth.
Monthly plots in Figures 29 through 36 show measured field runoff occur-
ring throughout the winter compared to HELP predictions which show no runoff
during this period but excessive runoff in April. These computed results are
consistent with HELP methodology which stores all precipitation on the surface
as snow when average temperatures are below freezing. Thus, when Wisconsin
temperatures warmed in April, all precipitation stored by HELP at the surface
was allowed to either run off or infiltrate. The measured data suggest that
this methodology is not appropriate. Significant runoff did occur in the
field throughout the extended periods of daily average below-freezing tempera-
tures; therefore, the field runoff in April was significantly less than that
predicted by HELP.
To assess the runoff curve numbers (CN's) chosen for these cells, the
simulation runs were repeated numerous times, changing only the value of the
CN, For each year of measured cell data (excluding the first 2 years), the
HELP run which most closely matched the measured annual runoff volume was
determined. These CN's were averaged for each cell. The questionable pre-
cipitation and runoff from the large storm in April 1975 was excluded from
this analysis. The CN's determined in this manner for the first three covered
cells were 83, 82, and 84—all very similar to the originally selected value
of 84.4. However, the average best-fit CN for the fourth covered cell (Cell
8) was 92. A comparison between Cell 8 field measurements and the HELP pre-
diction for CN = 92 is shown in Figure 37, The average best-fit CN's for the
82
-------
s 12'51
t£ 10.0 \
© a
z i
« 7.5 4
5 5-0 1
O
CO
Q
4-
o
2.5 -i
0
50 1
40 |
^ 301
LU -
2! 2
S 201
10
3
o 4.
o
-------
uncovered cells were all close to 86, higher than the originally selected
value of 79. The field measurements from these cells are plotted with HELP
predictions for CN = 86 in Figure 38.
A similar analysis was conducted to assess the hydraulic conductivity
assumed for Layer 1 of the uncovered cells. This is the layer of waste that
was subjected to surface weathering and vegetative growth. The simulation run
for the uncovered cells (using CN - 86) was repeated numerous times changing
only the hydraulic conductivity. The value of hydraulic conductivity which
minimized the percent differences between field measurements and HELP predic-
tions for runoff, ET+DS, and leachate drainage was selected as the best-fit
hydraulic conductivity. These best-fit values were 0.14, O.LL, 0.11, and
0.08 inches/hour compared to the original value of 0.17 inches/hour.
84
-------
LEGEND
o
z
ZD
QC
LLI
>
I—
-------
SECTION 7
SIMULATION OF SONOMA COUNTY TEST CELLS
A solid waste stabilization project was sponsored by the U.S. Environmen-
tal Protection Agency and Sonoma County, CA, from 1971 to 1974 (14), The pur-
pose of the project was to investigate the stabilization of solid waste in
five municipal sanitary landfill test cells by analyzing leachate, gas, tem-
perature, and settlement parameters and to determine the effect on solid waste
stabilization of applying excess water, septic tank pumpings, and recycled
leachate. This section describes the simulation of these cells using the HELP
model and presents comparisons between model predictions and field
measurements.
SITE DESCRIPTION
The test site is located in the southwestern portion of Sonoma County
approximately 45 miles north of San Francisco. The mean temperature is 58°F»
with the daily minimum temperature falling below freezing on 39 days per year.
The mean annual precipitation is 31 inches. Ninety-five percent of this pre-
cipitation occurs from October to April. The mean daily solar radiation is
approximately 410 langleys.
The cells were excavated in an area of sandy clay containing occasional
thin layers of pervious waterbearing soils. When such pervious areas were
encountered during cell excavation, the areas were overexcavated 2 feet and
lined with compacted clay. Subsurface drains outside the cells were also
installed where needed.
Cell dimensions presented in the remainder of this section were deter-
mined by roughly scaling them from figures in the stabilization project report
(14) and therefore are approximate. The base of the cell in plan view is
50 feet square; 45-degree side slopes rise from each side of the base.
Household refuse was placed directly on the clay base and compacted by a D-7
dozer in a manner similar to normal operations at a sanitary landfill. The
depth of compacted refuse is 8 feet. All cells were capped with a 2-foot
layer of native sandy clay soil.
Other characteristics of cell configuration varied as shown in Figure 39.
Operational characteristics are summarized in Table 12. Cells A, B, and E
were constructed and operated as typical sanitary landfills except that the
moisture content of the refuse was brought to field capacity prior to capping
in Cell B by water and Cell E by septic tank pumpings. Cells C and D were
86
-------
70'
2' soil covar
8* waste layer
drain
2% slope
10'
40
a) CELLS A and E
10'
10'
10'
10'
b) CELL B
Inflow pipe network
2' soil cover
1' sand layer (Cell C)
1* gravel layer (Coll D)
c) CELLS C and D
Figure 39. Cell dimensions for Sonoma County cells.
-------
TABLE 12. SONOKA COUNTY CEIL OPERATIONAL DESIGNS
Cell
Initial Application
of Liquid
Continuous Application
of Liquid
A
None
None
B
Water added to
field capacity
None
C
None
Water added daily
(200 to 1000 gal/day)
D
None
Recycled leachate
added daily (500 to
1000 gal/day)
E
Septic tank pump-
ings added to
field capacity
None
constructed with an inflow pipe network, and a sand/gravel distribution medium
was installed between the waste layer and the soil cover. Water was continu-
ously added to the waste through this network in Cell C, whereas landfill
leachate was continuously recycled through the waste by the network in Cell D.
Leachate collection drains shown in Figure 39 diverted the leachate by
gravity to collection tanks (Cells B, C, D) or to the nearby county landfill
(Cells A and E). Standard house service water meters installed on the dis-
charge lines of Cells A, C, D, and E measured the volume of leachate col-
lected. Cell B leachate was measured with a graduated bucket as it was
discharged from a collection tank. The frequency of meter readings and volume
measurements varied but averaged about once per week,
SELECTION OF MODEL INPUT VALUES
Daily precipitation measurements made at the test site were used in the
HELP simulations. Mean monthly temperatures taken from the NOAA Environmental
Data Service for Santa Rosa, CA, (10 miles from the test site) were used. The
solar radiation values used in the simulations were the default values for
Sacramento incorporated in the HELP model.
Native clay at the site composed the landfill base, side slopes, and soil
cover. In the stabilization project report (14), this soil was described as
sandy clay and was identified as USCS soil class CL based on liquid and plas-
tic limit tests. The HELP model default soil texture 14 was selected for use
in the simulations since it represents CL soil and has a hydraulic conductiv-
ity similar to the USDA sandy clay classification as shown in Table 10 of the
88
-------
HELP model user's guide. For the soil cover, soil texture 14 was selected to
be treated as compacted during the simulations using the HELP model, since it
was reported that "the sandy clay was spread in one-foot lifts and compacted
by numerous passes of a D-7 dozer" (14). For the landfill base, or barrier
soil layer, soil texture 14 was again treated as compacted, plus the hydraulic
conductivity was changed to reflect the average of three in-place hydraulic
conductivity tests (1.95 * 10 cm/sec or 0.000277 in./hr) reported for this
barrier soil layer (14). The thickness of the barrier soil layer was set to
24 inches since this was the depth of clay liner replacement where pervious
lenses were encountered during excavation.
Surface vegetation was assumed to be absent since (a) summer precipita-
tion is inadequate to support significant vegetation, (b) no water balance
accounting was made in the project report to include surface irrigation or
sprinkling, and (c) the presence of surface vegetation was not addressed in
the project report (14). Therefore, the vegetation type used in the HELP sim-
ulation was "bare ground" with year-round leaf area index values of zero and a
winter cover factor of zero. The evaporative depth was chosen as 4 inches as
suggested for bare ground by the HELP model.
SCS runoff curve numbers were estimated in two ways. First, the minimum
infiltration rate method presented in Table 10 and Figure 4 of the HELP model
documentation report (1) was used to determine a CN of 95. Second, individual
storm rainfall and runoff data at the site were used to estimate CN values
ranging from 88 to 98 with an average value of 95. Therefore, a CN of 95 was
selected for the simulations. This calibration procedure is identical to that
used for the University of Wisconsin-Madison data.
As previously described, Cells C and D had continuous inflow introduced
into the landfill above the waste layer. To simulate these cells, the HELP
computer code was modified to account for these inflows at the interface
between the soil cover and the distribution medium. During operation of these
cells, standard house service water meters measured the volume of inflow and
were read approximately once per week. These volumes were input in the HELP
simulation assuming a uniform distribution in time and over the landfill area.
Since the evaporative zone depth was set to a value less than the soil cover
depth, these inflows were not subjected to evapotranspiration losses in the
HELP simulations. The parameter values chosen for the model are presented in
Table 13.
RESULTS OF MODEL SIMULATIONS
Cells A, B, and E
Field monitoring began in November 1971, but a negligible leachate volume
was produced in Cells A, B, and E until the following winter's rainy season.
Therefore in the present study, the initial 12 months were treated as an
equilibration period for the cells. A similar period was designated in the
HELP simulations. Since the model must begin its simulation at the beginning
of a calendar year, the period between January and November 1972 was used as
an equilibration period for the model. These initial periods are not included
in the following tables and figures.
89
-------
TABLE 13, INPUT DATA FOR SIMULATION OF SONOMA COUNTY CELLS*
Parameter
Cells A and E
Cell B
Cell C
Cell D
No. of layers
4
4
5
5
Layer 1
Thickness (in.)
24
24
24
24
Layer type
1
1
1
1
Soil texture
14
14
14
14
Is layer compacted?
Yes
Yes
Yes
Yes
Layer 2
Thickness (in.)
72
72
12
12
Layer type
4
4
1
1
Soil texture
19
19
5
i
Is layer compacted?
No
No
Layer 3
Thickness (in.)
24
24
72
72
Layer type
2
2
4
4
Soil texture
19
19
19
19
Layer 4
Thickness
24
24
24
24
Layer type
3
3
2
2
Soil texture
14**
14**
L9
19
Is layer compacted?
Yes
Yes
Layer 5
Thickness (in.)
24
24
Layer type
3
3
Soil texture
14**
14**
Is layer compacted?
Yes
Yes
Type of vegetation
Bare
Bare
Bare
Bare
SCS runoff curve no.
95
95
95
95
Evaporative depth (in.)
4
4
4
4
Surface area (sq ft)
4900
4900
4900
4900
Slope of lateral drainage (%)
2
2
2
2
Drainage length (ft)
40
5
5
5
* Input data terminology defined in the HELP model documentation (1) and
user's guide (2).
** Hydraulic conductivity = 0.000277 in./hr based on permeability tests.
90
-------
Cells A, B, and E were monitored for both surface runoff and lateral
leachate drainage. Therefore, comparisons between model predictions and field
measurements are presented for the following components of the water balance:
(a) runoff; (b) the sum of evapotranspiration, change in landfill moisture
storage, and barrier soil percolation (ET+DS+PERC); and (c) lateral leachate
drainage. Field "measurements" of ET+DS+PERC were actually computed values
based on the following relationship:
ET+DS+PERC = PRCP - ROT - DRG (29)
where PRCP = measured precipitation, RNF = measured runoff, and DRG - measured
leachate drainage. Runoff from Cells A and E flowed into a single collection
tank prior to measurement. It was assumed that Cells A and E contributed
equally to this total measured runoff.
The field measurements for Cells A, B, and E are plotted together in Fig-
ure 40. The only differences in construction and operation of these cells
were their initial refuse moisture content and the shorter lateral drainage
length of Cell B. Since the first 12 months of data were treated as a period
of equilibration and are not plotted, one would expect the remainder of the
data to be somewhat similar. This was generally the case except for the
leachate drainage measurements. The Sonoma County project report (14) was
unable to explain these differences. However, a possible contributing factor
was the manner in which soil cover shrinkage cracks were treated in the field.
Soon after construction, the study investigators concluded that no leachate
drainage would occur in Cells A, B, and E unless additional water was added.
So shrinkage cracks which appeared in the soil cover during the first summer
were deliberately left unsealed until the winter rains of 1972-73 infiltrated
and eventually sealed the soil cover by natural swelling. Random variations
in the size, timing, and patterns of the cracks may have influenced the
erratic leachate drainage response.
Comparisons of cumulative runoff, cumulative ET+DS+PERC, and cumulative
leachate drainage are plotted for Cells A, B, and E in Figures 41 through 43;
monthly comparisons are plotted in Figures 44 through 46; and comparisons
between predicted and measured values are summarized in Table 14. These
results indicate that on the average for Cells A, B, and E, runoff accounted
for 61.0 percent (s = 2.7 percent) of the precipitation. The HELP model pre-
dicted 71.6 percent, yielding an error of 10.6 percent (s =¦ 2.5 percent) of
the precipitation or a relative error of 17 percent of the measured runoff.
The measured ET+DS+PERC was 36.0 percent (s = 3.4 percent) of the precipita-
tion while the predicted value was 26,8 percent. The model underpredicted by
9,2 percent (s « 3,4 percent) of the precipitation or 26 percent of the mea-
sured ET+DS+PERC. Leachate drainage accounted for 2,94 percent (s " 1.56 per-
cent) of the precipitation while the model estimated 0.07 percent. The
average error was -2.87 percent (s = 1.55 percent) of the precipitation, and
the relative error was -98 percent of the measured leachate.
The obvious major discrepancy in these comparisons is related to lateral
leachate drainage. The model predicted that practically all leachate would
leave through barrier soli percolation rather than lateral drainage. This is
consistent with the assumptions that the barrier soil layer is always
91
-------
LEGEND
z 50
% 40
z
=>
"30-1
> :
<201
3
•CELL A
'CELL B
'CELL E
04.
o
cc
w>
a
25 -I
o_ 20
15
10
5 4
o
0
o _
< 3 -3
s=
«c
cc
O o
_j 1
3
2
ZD
o o
JAN
1973
JAN
1974
DATE
Figure 40. Field measurements for Cells A, B and E.
92
-------
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
z
U_ -Jw -
u.
z 40 -
=5 j
cc i
5 20
o
0
z 30
4
£ 20
Q
¦m
15
S 10
i—
<
3 5
0
1,0
uj 0.8
CD
<
z
< 0.6
ac
Q
£ 0 4
<
HELP PERCOLATION
^ 0.2
0
*3
JAN
JAN
1973 1974
DATE
Figure 4L. Field measurements for Cell A compared to HELP simulation;
cumulative comparisons.
93
-------
LEGEND
« HELP SIMULATION
FIELD MEASUREMENT
o
cc
CO
Q
+
>
H—
<
ZD
O
20
15 -i
10 -
5-
0
2.5
eg 2.0
-------
LEGEND
60 h
HELP SIMULATION
FIELD MEASUREMENT
z
u_
Li.
cc
z 25
o
a.
CO
<
=>
z
UJ
uj
s
5
HELP PERCOLATION
5
ID
JAN
JAN
1973 1974
DATE
Figure 43, Field measurements for Cell E compared to HELP simulation;
cumulative comparisons.
95
-------
CJ
cc
LU
Q_
+
CO
o
+
5^
z
o
CD
<
z
<
cc
o
o
0.5
0.4
0.3
i
0.2 4
0.1
0
LEGEND
-—HELP SIMULATION
—• FIELD MEASUREMENT
i I
{ i I I I i I
JAN
1973
JAN
1974
DATE
Figure 44. Field measurements for Cell A compared to HELP simulation;
monthly comparisons.
96
-------
12.5 -i
10.0 1
LEGEND
-a HELP SIMULATION
- FIELD MEASUREMENT
z 7.5 4
QC
5 5.0
x
o 2.5
-------
o
Z2
OC
15.0
12.5
10.0
7.5
>"j
? 5.0 4j
z
i 2.5
LEGEND
'HELP SIMULATION
'FIELD MEASUREMENT
1
¦ ?
HI
1 r
Hi
O
ac
o
C3
<
Z
c
cc
o
z
o
¦
JO »_
II
l L I I I I
JAN
1974
DATE
Figure 46. Field measurements for Cell E compared to HELP simulation;
monthly comparisons.
98
-------
TABLE 14. DIFFERENCE BETWEEN CUMULATIVE HELP MODEL PREDICTIONS AND
CUMULATIVE FIELD MEASUREMENTS FOR SONOMA COUNTY CELLS
Runoff
Measured
Predicted
Error
Cell
(% precip)
(% precip)
(% precip)
(%)
A
B
C
59.3
64.1
71.9
71.9
12.6
7.8
21
12
D
E
59.7
71.1
1L.4
11
Mean*
Std. Dev.*
61.0
2.7
71.6
0.5
10.6
2.5
17
5
Mean Error as Percent of Mean Measured Runoff =
17%*
Range in
Runo f f =
Measured Runoff as Percent of Mean Measured
97% to 105%*
ET+DS+PERC
Measured
Predicted
Error
Cell
(% precip)
(% precip)
(% precip)
(%)
A
B
C
39.5
32.7
26.9
26.9
-12.6
-5.8
-32
-18
D
E
35.9
26.7
-9.2
-26
Mean*
Std. Dev.*
36.0
3.4
26.8
0.1
-9.2
3.4
-25
7
Mean Error
as Percent of Mean Measured
ET+DS+PERC
= -26%*
Range In Measured ET+DS+PERC as Percent
ET+DS+PERC = 91% to 110%*
of Mean Measured
Drainage
Measured
Predicted
Error
Cell
(% precip)
(% precip)
(% precip)
(%)
A
B
C
D
E
1.26
3.21
314.49
996.53
4.35
0.03
0.15
380.25
989.85
0.03
-1.23
-3.06
65.76
-6.68
-4.35
-98
-95
21
-1
-99
Mean*
Std. Dev.*
2.94
1.56
0.07
0.07
-2.87
1.55
-97
2
Mean Error as Percent of Mean Measured Drainage = -98%*
Range in Measured Drainage as Percent of Mean Measured
Drainage = 43% to 148%*
* Computed for Cells A, B, and E only,
99
-------
saturated and that Darcy's law can be used to describe vertical percolation
through this layer. These assumptions are incorporated into Equation 26,
rewritten as
Qp =» A y + Kp (30)
where Qp is the vertical percolation; A is a constant; y is the head above the
barrier soil layer; and is the saturated hydraulic conductivity of the bar-
rier soil layer. Equation 30 is solved simultaneously with Equation 25 to
describe both vertical percolation and lateral flow through a lateral drainage
layer. When the slope, drainage length, and saturated hydraulic conductivity
of the drainage layer are specified, Equation 25 reduces to
- ? 16 -
Qd = B y -*iD + C y (31)
where CL is the lateral flow, and B and C are constants. Therefore when y
approaches zero, approaches zero while Qp approaches the saturated
hydraulic conductivity of the barrier soil layer. In the case of Cells A, B,
and E, the model input values described a system that produced very small
heads such that Qp >> Q^. The large discrepancy between predicted and mea-
sured Qp in these cells implies that either the HELP methodology for dividing
flows between Qp and may not be appropriate for very small volumes of
inflow or the model input values were not appropriate. Since no data exist
for evaluating Qp and Q with small inflows, the following discussion focuses
on the choice of model input values. Results are presented only for cells A
and E.
First, default soil characteristics for one additional CL soil, soil tex-
ture 15, are available in the HELP model. Figure 47 shows the results of
using this soil texture (compacted) for the soil cover and for the barrier
soil whose ^ydraulic conductivity remained 0.000277 in./hr
(1.95 * 10 cm/sec). This soil texture produced a greater deviation from
measured values, reducing predicted leachate drainage from an average of
1.5 percent to an average of less than 0.1 percent of the measured drainage.
Also, the average of the three barrier soil hydraulic conductivity tests
may not be representative, especially in view of the large percolation volumes
predicted by the model. Figure 48 shows the results using the lowegt of the
three measured hydraulic conductivities, 0.000094 in./hr (6,6 x 10 cm/sec),
for the barrier soil. Although this change reduced percolation and increased
drainage to a small degree, percolation still dominated, and drainage remained
only a small fraction of the measured value.
Additional runoff curve numbers were also used in an attempt to match
predicted and measured runoff. Figure 49 presents a plot of the model results
using a CN of 60 instead of 95. This change had a negligible effect on
predicted runoff, indicating that CN alone was not controlling runoff.
It was difficult to determine from the project report (14) the degree of
compaction of the soil cover. Figure 50 shows the results using soil
100
-------
60
z
t 50
z 40
3
QC
oj 30
>
< 20
I
i 10
0
30
£ 25
uj
Cl.
L£GEN0
HELP SIMULATION
CELL A MEASUREMENT
CELL E MEASUREMENT
o
to
o
4-
20 i
15-
10'
-•* /
e o
e> 0,
c o
HELP PERCOLATION
JAN
1974
DATE
Figure 47. Field measurements for Cells A and E compared to HELP simulation
using default soil texture 15 for topsoil.
101
-------
u_- 50
U-
O -
Z 40 -
cc
m 30
>
< 20
ZD
§ m
O
0
LEGEND
HELP SIMULATION
CELL A MEASUREMENT
CELL E MEASUREMENT
5 4
to
<
z
«c
QC
Oi
>
(—
C
-J
Z3
s
ZD
O
3
2
-HELP PERCOLATION
JAN
1973
JAN
1974
DATE
Figure 48. Field measurements for Cells A and S compared to HELP
simulations using a barrier soil hydraulic conductivity of
0,000094 in./hr.
102
-------
LEGEND
3—c help simulation
— CELL A MEASUREMENT
— CELL E MEASUREMENT
U-
z 40
ZD
ec
>
<20 ^
_i
3 :
s 10 -
ZD
O
a.
(***!
s 10-
c :
= 5:
Z3
O
2 4 -
HELP PERCOLATION
JAN JAN
1973 1974
DATE
Figure 49. Field measurements for Cells A and E compared to HELP simulation
using a runoff curve number of 60.
103
-------
40 i
30
20
s 10
o
0
LEGEND
HELP SIMULATION
CELL A MEASUREMENT
CELL E MEASUREMENT
CD
<
Z
-------
characteristics reflecting changes equal to 25 percent of the magnitude of
changes predicted by the HELP model due to compaction. That is, the wilting
point remained constant, the plant available water and drainable porosity
estimated for ur.compacted soil were both reduced by 6,25 percent, the
hydraulic conductivity was divided by 5.0, and the evaporation coefficient was
set to 3.7. In this case runoff decreased, allowing more rainfall to infil-
trate without an equally compensating increase in evapotranspiration so that
predicted leachate drainage exceeded the measured value. This implies that
perhaps the actual effects of compaction is between the 25 percent level and
full compaction as defined by the HELP model. This was explored further by
varying thg hydraulic conductivity of the soil cover between 0.00325 in./hr
(2.3 * 10 cm/sec) representing 100 percent of the effects of compaction and
0.06500 in./hr (4.6 * 10 cm/sec) representing no compaction. The hydraulic
conductivity that matched measured leachate drainage the closest was between
0.00727 and 0.01190 in./hr (between 5.1 x 10 and 8.4 * 10 cm/sec).
Finally, the evaporative depth was varied to determine its effect on the
predicted results. Increasing the evaporative depth from 4 to 24 inches
reduced the predicted runoff from about 120 percent to about 114 percent of
measured runoff, while predicted ET+DS+PERC increased from about 71 percent to
about 84 percent of measured ET+DS+PERC, Predicted leachate drainage was
reduced to zero.
Cells C and D
As previously described, Cells C and D contained a distribution pipe net-
work beneath the soil cover. A continuous inflow of liquid above the waste
layer was thus introduced which was assumed to be protected from evapotranspi-
ration losses. The average annual inflow was approximately 141 inches for
Cell C and 355 inches for Cell D. This inflow would be expected to control
the response of leachate drainage and barrier soil percolation since rainfall
alone produced only a few inches of leachate drainage per year in Cells A, B,
and E. Because of this large volume of Inflow, no equilibration period was
deleted from the comparisons that follow.
Inflow volume and lateral leachate drainage were the only variables mea-
sured in the field for Cells C and D. Measurements show that at least
18 percent of the inflow did not appear in leachate drainage for Cell C and at
least 4 percent did not appear in Cell D. This unaccounted volume could be
due to barrier soil percolation, leakage through a failed portion of the clay
liner, measurement errors, or evaporation through shrinkage cracks in the soil
cover. Of these sources, the HELP model can only simulate barrier soil
percolation.
Comparisons between predicted and measured leachate drainage are con-
tained in Figures 51 and 52 for cumulative volume and Figures 53 and 54 for
monthly volume. HELP model predictions for runoff and ET+DS+PERC are also
included. Percent differences are summarized in Table 14. The comparisons
show that the HELP model overpredicted leachate drainage by 21 percent for
Cell C and by 1 percent for Cell D. Since the Cell D simulation appears to be
very accurate and since construction plans for Cells C and D were identical,
it is assumed that most of the excess drainage predicted by the HELP model for
105
-------
LEGEND
=—= HELP SIMULATION
•—• FIELD MEASUREMENT
o
cc
20
O
0
i
o
cc
co
Q
>
^ 400 ~
300
<
oc
Q 200
>
HELP PERCOLATION
JAN
JAN
1973 1974
DATE
Figure 51. Field measurement or leachate drainage for Cell C compared to
HELP simulation; cumulative comparison.
106
-------
. 60-
S
a: 50-
LL-
CS
^ 40
cc
LU 30 -
J—
< 20
id
1 10
o
0
LEGEND
-a HELP SIMULATION
-• FIELD MEASUREMENT
z 40
z 80°
ui 700
1 600
< 500 "i
cc
® 400
| 300'
2 200
5 100
0
o
HELP PERCOLATION
JAN
1973
DATE
JAN
1974
Figure 52. Field measurement of leachate drainage for Cell D compared
HELP simulation; cumulative comparison.
107
-------
12.5
,-10,0
cc
>-
X
b—•
O
7.5
5.0
2.5
0
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
f
cc 2.0
mi
JAN
1971
DATE
Figure 53. Field measurement of leachate drainage for Cell C compared
HELP simulation; monthly comparison.
108
-------
12.5
— 10.0
O r
2 7.5
ZD
cc
5 5.0
x
o 2.5
LEGEND
--c HELP SIMULATION
— FIELD MEASUREMENT
_SLi
JE.
i
11
tTt?
TTTtt
C5
<
Z
<
cc
Q
>¦
_J
I
I—
o
A D
b fa
<9 m f * _ W
a •
JAN
1971
JAN
1972
JAN
1973
DATE
Figure 54. Field measurement of leaehate drainage for Cell D compared
HELP simulation; monthly comparison.
109
-------
Cell C was actually lost through sources other than lateral drainage or barrier
soil percolation.
Based on measurements of leachate drainage in Cells A, B, and E, rainfall
accounted for about 1 percent or less of the leachate drainage in Cells C and
D. This provides a unique test of the HELP model methodology for dividing
flow between lateral drainage and vertical percolation since inflows and
lateral drainage outflows are both essentially known. The excellent reproduc-
tion of Cell D drainage appears to confirm the appropriateness of the HELP
model methodology for large Inflows.
110
-------
SECTION 8
SIMULATION OF BOONE COUNTY TEST CELL
Two field-scale test cells and three small-scale cells were studied from
1971 to 1980 in Boone County, KY, under the sponsorship of the U.S. Environ-
mental Protection Agency (15,16). The study objectives were to evaluate the
amount and characteristics of leachate, the composition of gases, the tempera-
ture conditions, the settlement of the cells, and the efficiency of the clay
liner, and to compare the behavior between the field—scale and small-scale
cells. The data collected from one of the field-scale cells were determined
suitable for a HELP simulation study which is reported in this section.
SITE DESCRIPTION
The test site is located in Boone County approximately 20 miles south of
Cincinnati, Ohio, The mean annual temperature is 54°F, with the daily minimum
temperature falling below freezing on 111 days per year. The mean annual pre-
cipitation is 43 inches. The mean daily solar radiation is approximately
360 langleys.
The test cell shown in Figure 55 consisted of a 30-foot-wide by 150-foot-
long treneh with vertical walls and ramps on both ends sloping at 14 percent.
The middle 50 feet were sloped at 7 percent to the transverse centerline.
Since the base of the excavation consisted of fractured limestone, an addi-
tional depth of 0.5 foot was excavated and replaced with compacted native
clay. A 30-mil synthetic liner, 30 feet wide by 50 feet long, was centered
over the base of the cell. A leachate collection pipe embedded in gravel was
placed on the synthetic liner at the bottom of the cell. An 18-inch-thick
compacted clay liner was placed over the synthetic liner and collection pipe.
This line^ was found to have an average in-place hydraulic conductivity of
4.0 x 10 cm/sec at the conclusion of the test cell study. A second pipe was
embedded in a gravel-filled section of the clay liner directly above the lower
pipe as shown in Figure 55 to collect lateral leachate drainage above the clay
liner. A 6—mil polyethylene strip was placed beneath this pipe to prevent
leachate from short—circuiting to the lower pipe. Residential refuse was
placed and compacted above this liner system, and a 2-foot layer of cover soil
was deposited onto the completed waste layer. The cover soil was identified
as USCS soil class CL and was found to have an average in-place hydraulic con-
ductivity of 5.0 x 10 cm/sec at the conclusion of the test cell study.
Ill
-------
r
2' soil cover
18* clay liner
leachate drainage system
30-mll synthetic liner
SO'
18"
8% slope
clay linor
wi
L
8.8% alopo
natural clay soil
12
o
18"
8* slop®
6-mil synthetic liner
3"-diameter
PVC slotted pipe
r slope
"V
30-mll synthetic liner
Figure 55. Cell dimensions for Boone County cell.
112
-------
SELECTION OF MODEL INPUT VALUES
Precipitation at the test site was recorded once or more per week
throughout the study period. These values were used to adjust daily precipi-
tation records from the nearest NOAA weather station, located approximately
15 miles away at Covington, KY. The adjustments were made by multiplying each
daily value by the ratio of total monthly precipitation at the test site to
total monthly precipitation at Covington, Mean monthly temperatures were
taken from the Covington weather station. Solar radiation values were the
default values for Cincinnati incorporated in the HELP model.
Default soil texture 14 was chosen to describe the 2-foot topsoil layer.
This texture matches the CL classification for this topsoil and has a
hydraulic conductivity close to that measured for this layer.
To model the circuitous path of clay liner percolation, an equivalent
liner thickness of 12 feet was chosen. The area of the gravel packing/clay
liner interface surrounding the lower pipe was treated as a leakage opening
through a synthetic membrane. The leakage fraction was computed as the area
of this interface divided by the area of the clay liner/waste layer interface,
or about 0.03. The measured in-place hydraulic conductivity of the clay liner
was used in the HELP simulations along with the remaining values for default
soli texture 20.
The surface vegetation was assumed to be fair grass based on reported
observations at the end of the test cell study indicating abundant fine grass
roots in the top 7 to 8 inches of soil cover. An evaporative depth of
7.5 inches was selected to correspond with this observation. A default runoff
curve number of 85.7 was determined by the HELP model and used in the simula-
tions. Table 15 summarizes the input parameter values for the Boone County
test cell.
RESULTS OF MODEL SIMULATIONS
The test cell was monitored for both leachate drainage above the clay
liner and leachate percolation through the clay liner. However, a number of
factors could have influenced these measurements. First, the synthetic liner
beneath the clay base only extended to the edge of the 14-percent side slopes.
The project report (15) states that the 14-percent slope was steep enough to
encourage water to flow down the ramps onto the liner without appreciable
seepage loss. Nevertheless, this created the potential for losses, especially
if leachate was ponded to depths greater than about 2 feet above the drainage
pipe. Secondly, there appeared to exist a possibility of high water table
conditions. Although the project report (15) did not directly address water
table elevations, it discussed a second test cell located about 150 feet away
In which measured drainage exceeded precipitation by 50 percent. The possible
explanations given were surface water infiltration from outside the cell and
groundwater seepage. Since the vertical sides of the test cell were not
lined, any water table rise above the cell base could have resulted in ground-
water inflow.
113
-------
TABLE 15. INPUT DATA FOR SIMULATION OF BOONE COUNTY CELL*
Parameter Value
No. of layers 4
Layer 1
Thickness (in.) 24
Layer type 1
Soil texture 14**
Is layer compacted? No
Layer 2
Thickness (in.) 48
Layer type 4
Soil texture 19
Layer 3
Thickness (in.) 24
Layer type 2
Soil texture 19
Layer 4
Thickness (in.) 144
Layer type 5
Soil texture 20f
Liner leakage fraction 0.03
Type of vegetation Fair
Evaporative depth 7.5
Surface area (sq ft) 4653
Slope of lateral drainage (%) 8
Drainage length (ft) 50
* Input data terminology defined in the HELP model documentation (1) and
user's guide (2).
** Hydraulic conductivity - 0.071 in./hr or 5.0 * 10 jm/sec.
t Hydraulic conductivity * 0.00057 in./hr or 4.0 * 10 cm/sec.
114
-------
Measurable leachate was first produced in September 1971, 3 months after
construction was completed. Cumulative plots of measured drainage and perco-
lation and the HELP model predictions for the period beginning in September
1971 are shown in Figure 56. Monthly plots are presented in Figure 57. These
figures show very little measured leachate until the end of 1972. After that
time, leachate drainage volumes somewhat exceed those predicted by HELP.
Percolation volumes are negligible for both the field measurement and the
model prediction. Over the 7-year period, leachate drainage accounted for
28,8 percent of the precipitation, while the HELP model predicted 24.6 per-
cent, The model underpredicted the drainage by 4.2 percent of the precipita-
tion or 14.6 percent of the measured drainage.
The field assessment of the test cell (16) at the end of the Boone County
study indicated that secondary openings existed in the soil cover through
which relatively rapid infiltration could have occurred. Therefore, a second
HELP simulation was conducted using the highest of the three in-^lace
hydraulic conductivity measurements of the soil cover, 7.0 x 10 cm/sec.
These results, plotted in Figure 58, showed a predicted leachate drainage of
26.2 percent of the precipitation, yielding an error of -2.6 percent or
-9.2 percent of the measured drainage. Both HELP simulations appeared to rea-
sonably reproduce the measured results at the Boone County test cell.
115
-------
LEGEND
40 i
~—~ HELP simulation
•—- FIELD MEASUREMENT
Ll-
U-
CD
z
=3
cc
UJ
5»
I—
<
_l
3
o
z
"250
200
C/3
o
150
100
>
ZD
=3
O
o
<
Z
<
or
a
o
80 -3
60 J
3
401
201
0-Ll.
JAN
1971
AND FIELD PERCOLATION
JAN
1972
JAN
1973
JAN
1974
JAN
1975
DATE
JAN
1976
JAN
1977
JAN
1978
JAN
1979
Figure 56. Field measurements for Boone County cell compared to HELP
simulation; cumulative comparisons.
116
-------
LEGEND
HELP SIMULATION
- FIELD MEASUREMENT 0
fe 2.0 J
kjl
t
iilffli
JAN JAN JAN JAN
1974 1975 1976 1977
DATE
Figure 57. Field measurements for Boone County cell compared to HELP
simulation; monthly comparisons.
117
-------
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
uj 15
C3
<
Z
40-3
i— i
-------
SECTION 9
SIMULATION OF THREE COUNTY LANDFILLS
IN WISCONSIN
The State of Wisconsin Bureau of Solid Waste Management (BSWM) has
reported on the geologic setting, major design features, construction experi-
ence, and operational performance of four large landfills in Wisconsin (17).
However, none of these landfills has yet been completely filled and capped, so
that each data set collected thus far represents the conditions of a continu-
ously expanding landfill. For example, at any given time, the cover could
range from a 6-inch—thick blanket of sand to a final layer of clay and top-
soil, This section presents HELP simulation results for a range of these con-
ditions at three of the better documented sites—Brown County landfill, Eau
Claire County landfill, and Marathon County landfill,
BROWN COUNTY LANDFILL
Site Description
The development of the Brown County landfill began in 1976 and Is occur-
ring in seven major sequences that will ultimately encompass an area of
58 acres. The landfill is located near Green Bay, WI, where the mean annual
temperature is 44°F. The daily minimum temperature falls below freezing on
163 days per year. The mean annual precipitation is 27 inches, and the mean
daily solar radiation is approximately 330 langleys.
A cross-sectional view of the landfill upon completion is shown in Fig-
ure 59, The base of the landfill consists of a 4-foot-thick compacted clay
liner and a leachate collection system. Sequences 1 through 3 (approximately
17 acres) were designed with a i-percent base slope, a leachate flow distance
of 300 feet, and a 1-foot thick sand blanket on the base and sidewalls. The
leachate flow distance was shortened to 100 feet for Sequence 4 (approximately
7 acres). The remaining three sequences have not been developed.
The waste initially consisted of municipal and commercial refuse. More
recently, industrial waste consisting of flyash and water treatment plant
sludge has been added. Daily cover is a 6-inch thickness of silty clay soil
(USCS classification CL), During the period of leachate monitoring presented
in this section, much of the landfill area was overlaid only with daily cover.
Final cover consists of 2 feet of compacted clay and 6 inches of topsoil.
119
-------
NJ
o
Final Cover -
(6" topsoil
2' clay)
4 Clay Liner
Water_TabJ_e
0
l_
500
i
HORIZONTAL SCALE, feet
Figure 59. Cell dimensions for Brown County Landfill.
-------
Selection of Model Input Values
Daily precipitation and mean monthly temperature measurements were taken
from the NOAA weather station at Green Bay. Solar radiation values were the
default values in the HELP model for Madison, WI.
Since the field data were collected during various stages of cover place-
ment, two HELP simulations were conducted to encompass the range of cover con-
ditions. First, the 6-inch daily cover was simulated using default soil
texture 14 to represent the silty clay cover soil. Vegetation was assumed to
be absent from this cover, and the evaporative depth was set to 4 inches to
correspond to the recommendation in the HELP model for bare soil. Second, a
6-inch top soil layer (default soil texture 14) underlaid by a 24-inch com-
pacted clay layer (default soil texture 18) was used to simulate the final
cover conditions. Fair grass was assumed based on State of Wisconsin require-
ments for seeding the topsoil cover. An evaporative depth of 10 inches was
used based on recommendations in the HELP model for fair grass. Since this
evaporative depth penetrated into the top 4 inches of the compacted clay
layer, this 4-inch section was treated as a vertical percolation layer while
the remaining 20 inches was treated as a barrier soil layer. Runoff curve
numbers were selected by the HELP model based on surface vegetation and the
minimum infiltration rate of the topsoil.
An average depth of 75 feet was used for the waste layer. The waste
characteristics were simulated using default soil texture 19. The 12-inch
sand blanket below the waste layer was modeled as a lateral drainage layer
using default soil texture 5. The 4-foot-thick clay liner was represented by
default barrier soil textgre 20 except for the hydraulic conductivity, which
was taken to be 6.5 x 10 cm/sec, an average of in-place permeability tests.
Table 16 summarizes all parameter values chosen for the simulations.
Results of Model Simulations
The site began accepting waste in August 1976, with routine leachate gen-
eration beginning in July 1977. The HELP simulations began in January 1977,
The comparisons that follow represent the period from 1978 through 1983.
Figures 60 and 61 compare the field measurements to the results of the
HELP model for daily cover. Figures 62 and 63 compare the field measurements
to the results of the HELP model for the final cover. Table 17 summarizes
differences between measured and computed results. Over the 6-year period,
the daily cover simulation overestimated leachate drainage by 65 percent of
the measured drainage, while the final cover simulation underestimated leach-
ate drainage by 29 percent. Since the actual cover during the period was
partial daily cover and partial final cover, the bracketing of the measured
results by the two HELP simulations seems reasonable.
The difference between the two simulation results relates primarily to
evapotranspiration. Although the runoff curve number was greater for the
unvegetated daily cover, the small evaporative depth, relatively short
121
-------
TABLE 16. INPUT DATA FOR SIMULATIONS OF BROWN COUNTY LANDFILL*
Parameter
Dally Cover
Simulation
Final Cover
Simulation
No. of layers
4
6
Layer I
Thickness (in.)
6
6
Layer type
L
1
Soil texture
L4
14
Is layer compacted?
No
No
Layer 2
Thickness (in.)
900
4
Layer type
4
1
Soil texture
19
18
Is layer compacted?
No
Yes
Layer 3
Thickness (in.)
12
20
Layer type
2
3
Soil texture
5
18
Is layer compacted?
No
Yes
Layer 4
Thickness (in.)
48
900
Layer type
3
4
Soil texture
20**
19
Layer 5
Thickness (in.)
12
Layer type
2
Soil texture
5
Is layer compacted?
No
Layer 6
Thickness (in.)
48
Layer type
3
Soil texture
20**
Type of vegetation
Bare
Fair
Evaporative depth (in.)
4
10
Slope of lateral drainage
(%)
1
1
Drainage length (ft)
300
300
* Input data terminology
user's guide (2).
defined
in the HELP model documentation
(1) and
10~8 cm/sec.
** Hydraulic conductivity
=» 6.5 x
122
-------
40
o 30-
z
rr
>
_l
=3
s
o
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
125 1
JAN
1979
JAN
1981
DATE
JAN
1982
JAN
1983
Figure 60, Field measurement of leachate drainage for Brown County landfill
compared to HELP simulation for daily cover; cumulative
comparison.
123
-------
LEGEND
-a HELP SIMULATION
FIELD MEASUREMENT
1
o
<
az
o
>-
m
HHsrf?,
iii
ill
9o 1
i f a
i i «
i i 1
4 I 1
I I
J
9?«>
I USB
111
I'll
It
mm
9a
1
JAN
1979
JAN
1980
JAN
1981
DATE
JAN
1982
JAN
1983
Figure 61. Field measurement of leachate drainage for Brown County landfill
compared to HELP simulation for daily cover; monthly comparison.
124
-------
LEGEND
15.0
q—j HELP SIMULATION
•—• FIELD MEASUREMENT
z
12.5
QC
7.5
5.0
2.5
o
0
2 180 1
« 150
ZD
O
5 J -i
HELP PERCOLATION
1978 1979 1980 1981 1982 1983 1984
DATE
Figure 62. Field measurement of leachate drainage for Brown County landfill
compared to HELP simulation for final cover; cumulative
comparison.
125
-------
O < j-
z 1.5
ZD
cc
^ 1
o 0,5
«E
0
M
LEGEND
™0 HELP SIMULATION
— FIELD MEASUREMENT
O
+ 4
fc
^ 2
2 0
*!E
-2 A
ILI
K
M
1
tfL
2 3^
JAN
1978
JAN
1979
JAN
1980
JAN
1981
DATE
Figure 63. Field measurement of leachate drainage for Brown County landfill
compared to HELP simulation for final cover; monthly comparison.
126
-------
TABLE 17. DIFFERENCE BETWEEN CUMULATIVE HELP MODEL PREDICTIONS AND
CUMULATIVE FIELD MEASUREMENTS FOR WISCONSIN COUNTY LANDFILLS
Drainage
Measured
Predicted
Error
Simulation
(% precip)
(% precip)
(% precip)
(%)
Brown County
3.1
Daily cover
5.1
2.0
65
Final cover
2.2
-0.9
-29
Eau Claire County
7.7
Daily sand cover
18.3
10.6
138
Interim sludge cover
0.3
-7.4
-96
(Uncompacted clay)
Interim sludge cover
5.3
-2.4
-31
(Clayey loam)
Final cover
17.5
9.8
127
Marathon County
6.9
Daily cover
8.9
2.0
29
Final cover
4.6
-2.3
-33
travel time through the evaporative zone, and unrestricting hydraulic conduc-
tivity of the layer below the evaporative zone combined to allow large volumes
of infiltration to pass through the cover. The final cover design with fair
grass required a lower runoff curve number, but the greater evaporative depth,
greater plant demand for moisture, and the greater residence time in the evap-
orative zone due to the restrictive clay barrier soil underlying the evapora-
tive zone all combined to reduce volumes of infiltration through the cover,
A groundwater monitoring program at this site has detected small changes
in quality Immediately adjacent to and below the landfill. The percolation
through the clay liner predicted by the HELP model during this period was
approximately 8 inches, or 32 percent of the pore volume of the liner as esti-
mated using the porosity for HELP default soil textures 20 and 21. With the
limited information available, it is difficult to speculate whether this pre-
dicted percolation rate was sufficient to cause groundwater quality changes.
EAU CLAIRE COUNTY LANDFILL
Site Description
The Eau Claire County landfill opened in December 1978 in the vicinity of
Eau Claire, WI. The ultimate size will be 24 acres. The data presented in
this section cover the period of landfill expansion up to 14 acres.
127
-------
The mean annual temperature is 43aF with 172 days per year experiencing a
minimum temperature below freezing. The mean annual precipitation is
29 inches, and the mean daily solar radiation is approximately 330 langleys.
The base of the landfill consists of a 4-foot-thick compacted clay liner
overlaid with a 1-foot-thick sand blanket and a leachate collection system.
The liner slope is 1 percent, and the maximum leachate flow distance along the
base is approximately 130 feet. The waste is primarily municipal and commer-
cial refuse with minor amounts of industrial wastes.
Daily cover at this landfill is a 6-inch layer of sand. A 30-inch
interim layer of papermill sludge covered a large percentage of the site dur-
ing the period considered here. Final capping will include a 12-inch sand
blanket over the sludge and then 6 inches of topsoil. The ultimate maximum
fill thickness will be approximately 50 feet. A cross-sectional view of the
landfill upon completion is shown in Figure 64.
Selection of Model Input Values
Daily precipitation and mean monthly temperature measurements were taken
from the NOAA weather station at Eau Claire. Solar radiation values were the
default values in the HELP model for Madison.
The modeling approach used here was similar to that used for the Brown
County landfill simulations. Since the field data were collected during vari-
ous stages of cover placement, four different HELP simulations were conducted
to encompass the range of cover conditions. First, the 6-inch daily sand
cover was simulated using default soil texture 3. Vegetation was assumed to
be absent from this cover, so the evaporative depth was set to 4 inches to
correspond to the recommendation in the HELP model for bare soil. Second, the
30-inch-thick interim sludge layer was simulated as an unvegetated cover. Two
sets of default soil characteristics were used to describe this sludge cover.
These were chosen based on comments by the State of Wisconsin 3SWM (17) that
field performance indicated that this sludge cover was more permeable than the
typical clay cover. One simulation used uncompacted clay (soil texture 18) to
describe the sludge cover and another used clayey loam (soil texture 14).
Finally, the permanent capping was represented by a 6-inch layer of topsoil
(soil texture 11) underlain by a 12-inch layer of sand (soil texture 3) over
the 30-inch sludge layer (soil texture 14). Vegetation was assumed to be fair
grass with an evaporative depth of 10 inches. Runoff curve numbers were
selected by the HELP model based on surface vegetation and the minimum infil-
tration rate of the topsoil.
An average depth of 38 feet was used for the waste layer. The waste
characteristics were simulated using default soil texture 19. The 12-inch
sand blanket below the waste layer was modeled as a lateral drainage layer
using default soil texture 5. The 4-foot-thick clay liner was represented by
default barrier soil texture 20 except for the hydraulic conductivity, which
was taken to be 1.4 x 10 cm/sec, an average of in-place permeability tests.
Table 18 summarizes all parameter values chosen for the simulations.
128
-------
940
Final Cover —
(6" topsoil,
12" sand,
30" sludge)
s>
v£>
Sevenmile Cr.
4 Clay Liner
Water Table —
910 -
©
z
O
I-
<
>
Ul
_J
UJ
880
'850
300
HORIZONTAL SCALE, feet
Figure 64. Cell dimensions for Eau Claire County landfill.
-------
TABLE 18. INPUT DATA FOR SIMULATION OF EAU CLAIRE COUNTY LANDFILL*
Sludge
Sludge
Cover as
Cover as
Sand
Uncompacted
Clayey Loam
Final
Cover
Clay
Clay
Cover
No, of layers
4
4
4
6
Layer 1
Thickness (in.)
6
30
30
6
Layer type
L
1
1
1
Soil texture
3
18
14
11
Is layer compacted?
No
No
No
No
Layer 2
Thickness (in.)
450
450
450
12
Layer type
4
4
4
1
Soil texture
19
19
19
3
Is layer compacted?
No
No
No
No
Layer 3
Thickness (in.)
12
12
12
30
Layer type
2
2
2
3
Soil texture
5
5
5
14
Is layer compacted?
No
No
No
No
Layer 4
Thickness (in.)
48
48
48
450
Layer type
3
3
3
4
Soil texture
20**
20**
20**
19
Layer 5
Thickness (in.)
12
Layer type
2
Soil texture
5
Is layer compacted?
No
Layer 6
Thickness (in.)
48
Layer type
3
Soil texture
20*"
Type of vegetation
Bare
Bare
Bare
Fair
Evaporative depth (in.)
4
4
4
10
Slope of lateral drainage (%)
1
1
1
1
Drainage length (ft)
130
130
130
130
* Input data terminology defined in the HELP model documentation (1) and
user's guide (2), _j
** Hydraulic conductivity = 1,4 x 10 cm/sec.
130
-------
Results of Model Simulations
The landfill opened in December 1978, and leachate generation began in
the spring of 1979, The HELP simulations began in January 1979. The compari-
sons that follow represent the period from 1980 through 1984,
The results of the four simulations (monthly and cumulative) are pre-
sented in Figures 65 through 72, Table 17 summarizes differences between mea-
sured and computed results. The simulated leachate drainage compared to field
measurements was 138 percent greater for the daily sand cover, 96 percent less
for the uncompacted clay sludge cover, 31 percent less for the clayey loam
sludge cover, and 127 percent greater for the final cover design. The large
increase in leachate drainage for the daily sand cover is not surprising. The
sand was modeled with a low runoff curve number, a large hydraulic conductiv-
ity, a small evaporative depth, and no surface vegetation—all of which con-
tribute to high infiltration rates. The simulation of very small drainage
volumes using uncompacted clay for the sludge cover tends to confirm that the
sludge is more permeable than clay. The increase in leachate drainage for the
final cover simulation must be qualified by the uncertainty of the sludge
characteristics. However, the overall bracketing of the measured results by
these simulations tends to confirm the general adequacy of the model.
Groundwater monitoring at the site has shown that there has been no mea-
surable effect on groundwater quality due to the landfill. The percolation
through the clay liner predicted by the HELP simulations ranged from 7 to
13 inches, or from 28 to 52 percent of the pore volume of the liner as esti-
mated using the porosity for HELP default soil textures 20 and 21, These
predictions appear to be reasonably consistent with the groundwater
observations.
MARATHON COUNTY LANDFILL
Site Description
The landfill was opened in December 1980 in central Wisconsin near the
city of Wausau. Approximately 27 acres have been licensed for waste disposal
with 14 acres under development at the end of the data collection period
presented in this section.
The mean annual temperature is 42°F with 170 days per year experiencing a
minimum temperature below freezing. The mean annual precipitation is
31 inches, and the mean dally solar radiation is approximately 330 langleys.
The facility is designed with a 4-foot-thick compacted clay liner and a
leachate collection system. The liner is sloped at 1 percent toward 8-inch-
diameter perforated PVC pipes. The pipes are embedded in a 1-foot-deep trench
oriented at approximately 45 degrees to the slope of the liner. The clay
liner thickness increases to 5 feet in the vicinity of these trenches. The
maximum leachate flow distance is about 250 feet in Phases 1 and 2 (9 acres
total). Phase 3 {5 acres) was designed for a maximum flow distance of
100 feet. Following construction of the clay liner and installation of the
131
-------
. 5
z
u_ 4
it* J
o
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
CJ3
< 30 i
2*
HELP PERCOLATION-7
JAN
1980
DATE
Figure 65. Field measurement of leachate drainage for Eau Claire County
landfill compared to HELP simulation for daily sand cover;
cumulative comparison.
132
-------
ac
>_
1.251
1.00 -i
0.75'
0.50
0.25 -i
0
LEGEND
HELP SIMULATION
- FIELD MEASUREMENT
?
T , T
10.0 i
Z :
~ 7.5 i
CO
a
+ 5.0 1
5 2.5
x
o 0
-2.5
T
1
SI
of
T T Tit,Tl
1.25
itff
JAN
1983
JAN
1984
DATE
Figure 66. Field measurement of leachate drainage for Eau Claire County
landfill compared to HELP simulation for daily sand cover;
monthly comparison.
133
-------
LEGEND
« HELP SIMULATION
- FIELD MEASUREMENT
60
50
40
30
20
10
0
100 H
C/3
a
UJ
1X1
>
(—
<
_J
ZD
2
o
15.0
uJ 12.5 -
C3
i 100
-------
7 -f
o
z c
oc ,
1
0
LEGEND
"O HELP SIMULATION
— FIELD MEASUREMENT
m
\m
1 i i m
4 i
3
>- 2
5"t»ai
Oh
? IP
O
I
0.8
S 0.7 •
S 0.6 ¦
<
5 0.5 •
<
g 0.4
2 03
| 0.2
I 0,1 ^
0
JAN
1980
ti 8 21 h %
JAN
1981
Ssa
i4*"cii
M "X>,
Taiaffi.
JAN
1982
JAN
1983
JAN
1984
DATE
Figure 68. Field measurement of leachate drainage for Eau Claire County
landfill compared to HELP simulation for uncompacted clay
sludge cover; monthly comparison•
135
-------
o
z
Z3
er
UJ
>
t—
-------
z
oc
>-
re
5
4
3 I
2
1 I
0
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
aImbbSj
T T
-f . t
Eft
es o
ILU
si
7
6
CO
Q
4 -=
S 2
1 1
t— '
z
o 0
2
-1
-2
Li
™?r
? ep
i 0.3 t
S 0.2
i o.i
¦>' 'A
"li
JAN
1980
JAN
1981
JAN
1982
JAN
1983
JAN
1984
DATE
Figure 70. Field measurement of leachate drainage for Eau Claire County
landfill compared to HELP simulation for clayey loam sludge
cover; monthly comparison.
137
-------
2»
3.0 -
2.5 ^
az
2.0
1.5
1.0
I 0.5
ZD
t-3
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
¦¦-J3-'""—"
^ f
ii¦¦¦l
y
uu 75
30 i
25 -=
20
15 ^
10
JAN
1980
JAN
1981
JAN
1982
JAN
1983
JAN
1984
DATE
Figure 71. Field measurement of leachate drainage for Eau Claire County
landfill compared Co HELP simulation for final cover;
cumulative comparison.
138
-------
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
1.50
z
~ 1.25
LU
CD
I 1.00
|0.75
5 0.50
o 0.25
s
0
JAN
1980
1"?
! 9
f '
I* -
i M
h
MtliiM
JAN
1981
JAN
1984
DATE
Figure 72, Field measurement of leachate drainage for Eau Claire County
landfill compared to HELP simulation for final cover;
monthly comparison.
139
-------
collection pipe, a minimum i-foot-thick silty sand drainage blanket was placed
over the base and sidewalls.
The waste is 75 percent municipal refuse and 25 percent papermill sludge.
The thickness of the waste layer ranges from 50 to 80 feet. The daily cover
is 0.5 to 1,0 foot of sand. The final cover is 2 feet of clay covered by
6 inches of silty sand and 6 inches of topsoil.
Selection of Model Input Values
Daily precipitation and mean monthly temperature measurements were taken
from the KOAA weather station at Wausau. Solar radiation values were the
default values in the HELP model for Madison,
Two HELP simulations were performed—one for daily cover and the other
for the final cover design. A 6-inch daily cover was simulated using the
silty sand default soil texture 7 placed directly over the waste. Vegetation
was assumed to be absent, so the evaporative depth was chosen to be 4 inches,
corresponding to the recommended depth for bare soil in the HELP model. The
final cover was simulated with a 6-inch topsoil layer whose characteristics
were represented by the loam default soil texture 11. Below the topsoil, a
6-inch layer of silty sand was simulated using default soil 7. This was
underlain by a 24-inch barrier layer of clay represented by the compacted
characteristics of soil texture 18. Vegetation was assumed to be poor grass
based on reports that only sparse vegetation was established on the topsoil
during the period of record considered here. Runoff curve numbers were
selected by the HELP model based on the assumed vegetative conditions and on
the minimum infiltration rate of the topsoil.
An average depth of 60 feet was used for the waste layer which was simu-
lated using default soil texture 19. The 12-inch sand blanket below the waste
layer was modeled as a lateral drainage layer using default soil texture 7.
The 4-foot-thick clay liner was represented by default soil texture 21 with a
hydraulic conductivity of 3,0 x 10~ cm/sec, an average of in-place permeabil-
ity tests. Table 19 summarizes all parameter values chosen for the
simulations.
Results of Model Simulations
The landfill began operation in December 1980, and leachate generation
began approximately 5 months later. The HELP simulations began in Janu-
ary 1981. The comparisons that follow include 1981 through 1983 and are based
on total annual leachate volume since monthly field data are not presently
available.
Table 20 compares annual and cumulative leachate drainage volumes from
field measurements and HELP predictions. Table 17 summarizes percent dif-
ferences between the measured and computed results. For the initial year,
1981, field measurements significantly exceeded both HELP predictions. In the
following 2 years, the predicted volumes increased so that the total
cumulative measured volume fell between the daily cover and final cover
140
-------
TABLE 19. INPUT DATA FOR SIMULATION OF MARATHON COUNTY LANDFILL*
Parameter
Daily Cover
Final Cover
No. of layers
4
6
Layer I
Thickness (in.)
6
6
Layer type
1
1
Soil texture
7
11
Is layer compacted?
No
No
Layer 2
Thickness (in.)
720
6
Layer type
4
i
Soil texture
19
7
Is layer compacted?
No
No
Layer 3
Thickness (in.)
12
24
Layer type
2
3
Soil texture
7
18
Is layer compacted?
No
Yes
Layer 4
Thickness (in.)
48
720
Layer type
3
4
Soil texture
21**
19
Layer 5
.Thickness (in.)
12
Layer type
2
Soil texture
7
Is layer compacted?
No
Layer 6
Thickness (in.)
48
Layer type
3
Soil texture
21**
Type of vegetation
Bare
Poor
Evaporative depth (in.)
4
8
Slope of lateral drainage (%)
1
1
Drainage length (ft)
250
250
* Input data terminology defined in the HELP model documentation (1) and
user's guide (2). _g
** Hydraulic conductivity = 3.0 x 10 cm/aec.
141
-------
TABLE 20. COMPARISON OF HELP SIMULATIONS TO FIELD MEASUREMENTS
FOR MARATHON COUNTY LANDFILL
Annual Leaehate Cumulative Leaehate
Drainage (in.) Drainage (in.)
1981 1982 1983 1981 1982 1983
HELP simulation using 0,8 2,9 5.7 0.8 3.7 9.4
daily cover
HELP simulation using 0.6 1.5 2.7 0.6 2.1 4.8
final cover
Field measurements 2.2 2,4 2.7 2.2 4.6 7.3
predictions for the overall 3-year period of record. In this case, it appears
that the model and the field measurements are approaching equilibrium at dif-
ferent rates and that 3 years Is an insufficient equilibration period. How-
ever, the field measurements appear to be converging to the final cover
simulation. Overall, the daily cover simulation overestimated drainage by
29 percent of the measured value, while the final cover simulation underesti-
mated drainage by 33 percent.
142
-------
SECTION 10
SIMULATION OF CHEMICAL WASTE DISPOSAL FACILITY
IN KIAGARA FALLS, NY
Since 19 76, a chemical waste management company has filled and capped
three landfill cells in Niagara Falls, NY. The surface areas of the cells
range from 2 to 5 acres. Records of leachate pumpage have been kept from 1983
and indicate annual withdrawals ranging from 1 to 11 inches. An evaluation of
the performance of the facility during 1984 was reported to the USEPA
Region II by Recra Research, Inc. (18). Comparisons between field measure-
ments and HELP simulations are presented below.
SITE DESCRIPTION
The site is located in Niagara Falls where the mean annual temperature is
48°F and where the daily minimum temperature falls below freezing on approxi-
mately 142 days per year. The mean annual precipitation is 35 inches, and the
mean daily solar radiation is approximately 310 langleys.
The site is a former disposal area containing waste industrial slag rang-
ing in depth from 7 to 50 feet. Beneath the slag is a layer of marsh silt
ranging in thickness from a few inches to 5 feet. Below the silt is a 6-foot
layer of lacustrine clay underlain by a 5-foot layer of glacial till. Excava-
tion of the cells stopped at the lacustrine clay because of the high water
table which required dewatering. A 10-foot-thick clay liner, compacted in
6-inch lifts, was placed at the bottom and along the side slopes of ^ach cell.
The hydraulic conductivity of this layer was specified to be 1 x 10 cm/sec
or less. The bottom liner was sloped at 2 percent, and the sides were sloped
at 2 horizontal to 1 vertical. Above this liner on both the cell bottom and
side slopes was placed a "combination" liner consisting of 2 feet of compacted
clay topped by a 30-mil synthetic liner and an additional 2 feet of compacted
clay as shown in Figure 73. These cl^y layers were specified to have a maxi-
mum hydraulic conductivity of 1 x 10 cm/sec. The 30-mil synthetic liner was
installed in Cell 1 only.
Stacked barrels containing chemical wastes are located on this combina-
tion liner. The depth of the waste layer is approximately 50 feet. Above the
waste is a 3-foot layer of compacted clay, then two 6-mil synthetic liners,
and finally a 1.5-foot layer of uncompacted clay. An underdrain is included
in the uncompacted clay layer around the perimeter of the landfill. In Cell 1
143
-------
1.5* uncompacted clay
two 6-mil synthetic liners
3' compacted clay
2' compacted clay
30-mil synthetic liner (Cell 1)
2' compacted clay
10' compacted clay
Figure 73. Cell dimension for Niagara Falls cell.
-------
only, the underdrain feeds into the leachate sump at the bottom of the cell.
Surface areas are 1.93 acres for Cell 1, 2.50 acres for Cell 2, and 5,13 acres
for Cell 3. Figure 73 shows a typical cross section through a cell.
The top clay layer was seeded with grass and is now kept mowed. Recent
inspections indicate no signs of surface cracking, excessive subsidence and
ponding, or excessive infiltration. Localized subsidence areas are repaired
under an inspection program. The years of operation for Cells 1, 2, and 3
were 1976 to 1978, 1978 to 1981, and 1981 to 1983, respectively.
Cells 2 and 3 are divided into subcells by interior berms. Each subcell
contains two standpipes into which leachate drains directly since no interior
leachate collection pipes are contained within the cells. The leachate is
pumped from the standpipes and metered on a regular basis. Since Febru-
ary 1984, level-activated pumps have been used to maintain.accumulated leach-
ate at lower levels within the cells. Cell 1 operates in a similar manner but
it does not have interior berms. During the period January to June 1984, mal-
functions occurred in the metering equipment, and the accuracy of the leachate
volume measurements was questionable.
Recra Research, Inc. (18) presents measurements of groundwater levels
outside the cells during 1984. These results show that groundwater levels
remained above the base of the cells for significant periods of time,
SELECTION OF MODEL INPUT VALUES
Daily precipitation and mean monthly temperatures were taken from a XOAA
weather station in Lockport, NY, approximately 15 miles from the landfill
facility. Solar radiation values were the default values for Syracuse incor-
porated in the HELP model.
The clay soil cover was described using default soil texture 18, The top
18 inches was assumed to be uncompacted, while the underlying 36 inches was
assumed to be compacted. The synthetic liner between these two soil layers
was modeled by adjusting the hydraulic conductivity of the 36-inch layer based
on an estimated leakage fraction through the synthetic liner as discussed
later in this section. The drainage slope above this synthetic liner was set
at 2 percent. The drainage length was 150 feet, an average distance from the
center of the cover to the edge of the landfill.
The top 25-foot layer of waste was described using the default waste
characteristics in the HELP model (soil texture 19). The bottom 25-foot layer
was described using estimated characteristics for tightly packed stacked bar-
rels with voids between barrels filled with a loosely packed coarse sand.
This resulted in a composite porosity of 0,035 and field capacity of 0.017,
The clay liner at the base was represented by soil texture 20, which has a
hydraulic conductivity of 1 x 10 cm/sec.
The surface vegetation was assumed to be fair grass based on reports of
seeding and mowing. The corresponding evaporative depth was chosen to be
10 inches as suggested by the HELP model for fair grass. A default runoff
145
-------
curve number of 89.6 was determined by che HELP model and was used in the sim-
ulations. A summary of all input values is presented in Table 21.
RESULTS OF MODEL SIMULATIONS
Leachate pumping records were made available for the period January 1983
to March 1985. Measured leachate volumes for Cell 1 include drainage from the
perimeter underdrains in the soil cover. The HELP simulation curves for
Cell 1 also Include this additional component. Results for Cells 2 and 3 are
shown together since their construction designs were essentially identical for
the purposes of the HELP simulation.
The use of synthetic liners on the bottom, side slopes, and cover of
Cell 1 theoretically should eliminate all inflow into the cell except for
drainage from the perimeter underdrains, which is discharged directly into the
leachate sump of Cell 1. However, HELP simulations indicated that this perim-
eter drainage accounted for less than 1 percent of the measured leachate
volume. Other possible sources of inflow include surface runoff through
direct connections to perimeter underdrains, leakage of infiltrated surface
water through the clay and synthetic liners in the cover, and leakage of
groundwater into the cell through the clay and synthetic liners on the bottom
and side slopes. Of these possible sources, only leakage through the cover
liners can be simulated in the HELP model; therefore, a leakage fraction in
the cover was used in the Cell 1 simulations to account for additional inflow
from these possible sources. It was assumed that leakage out of the cell
through the bottom synthetic liner was negligible due to the high groundwater
table.
Figures 74 and 75 show the simulation results for Cell 1 assuming a leak-
age fraction of O.IG through the cover synthetic liner. The simulation rea-
sonably matched measured volumes through about February 1984. After that
time, the measured volumes significantly increased and deviated from the HELP
prediction. The match prior to February 1984 with a 0.10 leakage fraction
indicates either that significant inflows from other sources occurred after
February 1984 or that the HELP simulation was underpredicting lateral drainage
in the cover, or both.
The use of synthetic liners in the cover of Cells 2 and 3 theoretically
should eliminate all surface water inflow into the cells since the perimeter
underdrains are discharged offsite. Therefore, possible sources of the leach-
ate volumes pumped from these two cells include leakage of infiltrated surface
water through the clay and synthetic liners in the cover, and inflow of
groundwater into the cell through the clay liner on the bottom and side
slopes. As described for Cell 1, a leakage fraction through the cover syn-
thetic liner was used in the Cell 2 and 3 simulations to account for addi-
tional Inflow from these possible sources. As for Cell 1, it was assumed that
leachate percolation out of the cell through the bottom clay liner was negli-
gible due to the high water table.
146
-------
TABLE 21. INPUT DATA FOR NIAGARA FALLS LANDFILL SIMULATION*
Parameter Value
No. of Layers
6
Layer 1
Thickness (in.)
12
Layer type
1
Soil texture
18
Is layer compacted?
No
Layer 2
Thickness (in.)
6
Layer type
2
Soil texture
18
Is layer compacted?
No
Layer 3
Thickness (in.)
36
Layer type
3
Soil texture
18
Is layer compacted?
Yes
Layer 4
Thickness (in.)
300
Layer type
4
Soil texture
19
Layer 5
Thickness (in.)
300
Layer type
2
Soil texture
**
Is layer compacted?
No
Layer 6
Thickness (in.)
168
Layer type
3
Soil texture
20
Linear leakage fraction
0.10
Type of vegetation
Fair
Evaporative depth (in.)
10
Surface area (sq ft)
84,000
Slope of lateral drainage (%)
2
Drainage length (ft)
150
* Input data terminology defined in the HELP model documentation (I) and
user's guide (2).
** Porosity = 0.035; field capacity = 0.017; hydraulic conductivity =
10.0 in./hr or 0.007 cm/sec.
147
-------
LEGEND
- HELP SIMULATION
* FIELD MEASUREMENT
o
2
>
—J
=3
O
z
i 12
<
ce
c Q
JAN
1984
DATE
JAN
1983
JAN
1985
Figure 74, Field measurement of leachate drainage for Niagara Falls Cell 1
compared to HELP simulation; cumulative comparison.
148
-------
o
z
ZD
CC
>-
-J
H-
-g.
o
3
2
1
oi
LEGEND
--a HELP SIMULATION
— FIELD MEASUREMENT
T , T
to
c
>-
_J
X
o
-------
Figures 76 through 78 show the simulation results for Cells 2 and 3
assuming a leakage fraction of 0.10 through the cover synthetic liner. In
this case, the simulation reasonably matched measured volumes in Cell 2
through about February 1984 but underestimated measured volumes in Cell 3
throughout the period of record.
The period of metering equipment malfunction (January to June 1984) fell
within the period of greatly increased measured volumes of leachate (after
February 1984). However, the large measured volumes were sustained beyond the
period of equipment malfunction, implying that the equipment was not entirely
responsible for the reporting of increased volumes.
The date when the measured leachate volumes began to significantly
Increase was the same in all cells and corresponded to the date when level-
actuated pumps were first used to discharge leachate from the cells. This
fact, combined with the knowledge of high groundwater levels, strongly sug-
gests that the lower ponding depths maintained by the new pumps increased the
hydraulic gradient across the landfill liner, increasing the rate of ground-
water inflow into the cells. If this was the case, the 0.10 leakage fraction
which matched measured results prior to February 1984 was probably due more to
groundwater inflow than a grossly underpredicted lateral drainage rate In the
cover or a large leakage fraction through the cover synthetic liner.
150
-------
LEGEND
a—a HELP SIMULATION
FIELD MEASUREMENT CELL 2
• FIELD MEASUREMENT CELL 3
o 20
z
ZD
OC
UJ 15
10
ZD
0
60
z
50
CO
o
+ 40
30
10
0
JAN
1984
DATE
Figure 76, Field measurements of leachate drainage for Niagara Falls
Cells 2 and 3 compared to HELP simulation; cumulative
comparisons.
151
-------
LEGEND
« HELP SIMULATION
FIELD MEASUREMENT
GSLmj&Maap
03-
CD
1 0,2-
<
QC
O
0.1
04-1
JAN
1983
m sp
CI 13 C> (I
i) (>
" m
JAN
1984
DATE
JAN
1985
Figure 77. Field measurement of leachate drainage for Niagara Falls
Cell 2 compared to HELP simulation; monthly comparisons.
152
-------
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
1.8 :
ui i 5 -
CD 13 ^
<
1.2 :
0.9
<
cc
Q
>•
tE 0.6
*£.
i 0.3
t If t Tynti
JAN
1983
JAN
1984
DATE
(I 13
JAN
1985
Figure 78. Field measurement of leachate drainage for Niagara Falls
Cell 3 compared to HELP simulation; monthly comparisons.
153
-------
SECTION 11
EVALUATION OF SIMULATIONS
Mathematical simulations of 20 landfill cells at seven sites across the
United States were made, and the results were compared to measured field data.
These landfills included a wide variety of conditions for which the HELP model
was tested. This section summarizes the results of these simulations and
evaluates the level of verification that has been accomplished.
EVALUATION OF FIELD DATA
It was found that the extent of available data on landfill leachate pro-
duction is very limited, especially for periods of record which extend signif-
icantly beyond the initial water budget equilibration period, which may last
up to several years. The extent of available data on other important facets
of the water balance, such as runoff* evaporation, rainfall, soil moisture,
leachate ponding depths, percolation rates, and detailed soil characteristics
are even more limited. The scarcity of data is understandable since data col-
lection can be costly. Nevertheless, the level of field verification of the
HELP model and the resulting level of understanding of the important processes
involved in leachate production and migration are highly dependent on obtain-
ing this information.
For this study, runoff volumes were measured from II landfill cells at
two locations—all were test cells with surface areas less than 0.1 acre.
Evaporation was measured at one site, but was not documented sufficiently for
inclusion in this study. Daily rainfall was measured at one site. Soil mois-
ture, leachate ponding depths, and barrier soil percolation rates were not
adequately or consistently measured in any of the landfill cells. Most sites
had at least limited data on the hydraulic conductivity of the clay liner,
although the testing methods and the applicability of the results varied
widely. Data related to characteristics of the cover soils or to the extent
of surface vegetation were generally lacking.
The lack of adequate site description and measured water budget compo-
nents affected the verification study in two ways. First, the lack of
descriptive landfill information required the frequent use of default values
in the HELP model, which introduced additional uncertainty into the verifica-
tion. Second, the lack of water balance outflow measurements limited the
number of HELP outflow predictions that could be verified. These limitations
restricted the ability of the study to isolate and test mathematical
154
-------
characterizations of specific physical processes, such as soil moisture
storage and routing, evapotranspiration demand and its distribution through
the soil profile, unsaturated vertical drainage, and details of the
apportioning of leachate production between lateral drainage to collection
systems and vertical percolation through the clay liner.
In addition, the variable degree of field measurement precision and
reliability presented challenges in interpreting the data which did exist.
None of the field data used in this report were collected specifically for
verifying the HELP models therefore, the field data were not always consistent
with the needs of this study. For instance, the data available for the three
largest landfills were collected while they were simultaneously undergoing
expansion. In other cases, there was large variability in measured results
between otherwise identical landfills. All of this required a significant
amount of engineering judgment in interpreting the data for the HELP model
comparisons,
Although a detailed verification of specific model components was not
always possible, the data did confirm the model's overall utility in estimat-
ing a landfill water balance even without extensive knowledge of specific
landfill characteristics. This was an important finding since the HELP model
is typically used with a limited amount of detailed landfill information.
EVALUATION OF MODEL PREDICTIONS
Runo f £
Measured runoff data existed for eight cells at the University of Wiscon-
sin and for three cells at Sonoma County, CA. In all cases, an attempt was
made to calibrate the runoff curve number in advance of the simulation by
examining measured rainfall-runoff data. These calibrated curve numbers were
used for the simulations, but they were consistent with default values that
would have been selected by the HELP model based on surface vegetation and the
minimum infiltration rate of the topsoil. Runoff was overpredicted for five
cells by an average of 30 percent of the measured runoff, and underpredicted
for six cells by an average of 20 percent of the measured runoff. Following
these initial simulations, the curve numbers were varied to determine their
effect on the overall model prediction of landfill performance. Five simula-
tions were improved by a change in curve number—all had originally underpre-
dicted runoff.
For the three cells at Sonoma County, it was obvious that the evapotran-
spiration and/or soil characteristics were controlling runoff volume and not
the curve number. Because of this close interaction, it was difficult to
assess the accuracy of the curve number method in the HELP model based on the
field data in this report. However, the predicted runoff volumes appear over-
all to be in reasonable agreement with the measured results.
A comparison of measured and predicted runoff on a monthly basis for the
University of Wisconsin cells indicated that the assumptions used in the HELP
model for snowmelt runoff may not be appropriate. The model stores all
precipitation on the surface when the mean daily temperature interpolated from
155
-------
the mean monthly temperature is below freezing. When this mean daily tempera-
ture rises above freezing, the precipitation is allowed to either run off or
infiltrate. Since mean daily temperatures are computed in the HELP model
based on mean monthly temperatures which are generally below freezing in
Wisconsin for several consecutive months, no runoff was predicted by the HELP
model during the winter. Instead, a large runoff volume was predicted during
April of each year when temperatures warmed. This compared to measured
results which showed significant runoff throughout the winter without an
excessively large runoff in April. This discrepancy probably contributed to
the overprediction of runoff for several cells.
It should be noted that the measured runoff data examined in this report
were restricted to relatively flat surfaces. The effect of steeper slopes
typical of mound construction was not specifically studied.
Evapotranspiration
No suitable evapotranspiration field data from landfill sites was found
for model testing. This was not unexpected due to the complexities involved
in collecting this type of data. Yet, evapotranspiration is typically the
single largest outflow component of the landfill system; therefore, small
changes in evapotranspiration can have major impacts on volumes of lateral
drainage and barrier soil percolation.
Of particular importance and interest was the appropriate depth to assume
for the evaporative zone in the top subprofile. As shown in the sensitivity
analysis, an increase in evaporative depth from 4 to 18 inches can decrease
leachate production by more than 50 percent. However, the simulations in this
study were only able to indirectly assess the evaporative depth assumptions.
For those cells which had runoff data available, a surrogate variable for
evapotranspiration was identified, and comparisons were made between measured
and predicted results. The variable consisted of the sum of the water balance
components which were not directly measured. In the case of the University of
Wisconsin cells, the variable was the sum of evapotranspiration and change in
moisture storage, ET-HDS, For the Sonoma County cells, it was the sum of
evapotranspiration, change in moisture storage, and percolation, ET+DS+PERC.
The ET+DS variable was found to be underpredicted by an average of 4 percent
of the measured values, whereas the ET+DS+PERC variable was underpredicted by
an average of 25 percent. It is obviously rather complex to discern the
meaning of these results since evapotranspiration, change in moisture storage,
and percolation are all interrelated. The evidence suggests that values
chosen for evaporative depths may have been too small. However, for the
Sonoma County landfills, an increase in the evaporative depth from 4 to
24 inches had only a small effect on the ET+DS+PEEC results.
Lateral Drainage and Percolation
Since measurements of barrier soil percolation volumes and leachate
ponding depths were generally not avaxlable, the lateral drainage and barrier
soil percolation submodel could only be evaluated using measured drainage data.
One exception was the Boone County, KY, cell where barrier soil percolation
156
-------
volumes were measured. However, the configuration of the clay liner and per-
colation collection pipe was such that vertical percolation did not actually
occur; rather, the percolation flow paths were forced to converge radially
toward the collection pipe. The attempt to simulate this percolation using
the HELP model resulted in an overprediction of approximately 35 percent.
Lateral drainage was overpredicted by 10 percent of the measured drainage
in two cells where very high leachate collection rates were observed. In
three cells where very small quantities of leachate were collected, lateral
drainage was underestimated by 97 percent of the measured drainage, although
this difference amounted to only 1.4 inches per year. Of the remaining nine
cells, lateral drainage was overpredicted by an average of 4 percent of the
measured drainage in five covered cells and overpredicted by an average of
53 percent of the measured drainage in four permanently uncovered cells with a
weathered waste surface that supported dense vegetation. Small errors in the
hydraulic conductivities of the cover soils can cause large differences in the
leachate production when the leachate production is small. Also the over-
predictions may have been partially related to the manner in which the HELP
model estimates unsaturated hydraulic conductivities. To linearly relate
unsaturated hydraulic conductivity to moisture content between field capacity
and saturation tends to overpredict unsaturated hydraulic conductivity. Thus,
moisture is routed more quickly through the evaporative zone, contributing to
larger leachate volumes and lower evapotranspiration volumes.
The cells at Sonoma County provided two important tests of the lateral
drainage and barrier soil percolation submodel. First, the three cells with-
out liquid redistribution generated very small leachate volumes due to low
summer rainfall, high evapotranspiration, and a clay cover soil. As discussed
previously, the initial simulation significantly underpredicted lateral drain-
age under these conditions. Second, the cells that included the additional
inflow to the waste layer provided a case where infiltration rates to the
lateral drainage layer were very large and essentially known. Under these
conditions, the submodel very closely reproduced measured lateral drainage
volumes.
The poor reproductions of lateral drainage for three of the Sonoma County
cells could have been influenced by two assumptions incorporated in the HELP
model related to barrier soil percolation. First is the assumption of free
outfall conditions below the clay liner. In many cases, the hydraulic con-
ductivity of the underlying soils is indeed larger than that of the clay
liner, and the assumption that the head at the base of the liner is zero dur-
ing vertical flow may be reasonable. However, it is also possible that
groundwater conditions beneath the clay liner, such as saturated soil layers
under the liner, may increase this head and therefore reduce percolation
rates. If such conditions are known from soil borings or monitoring wells,
the thickness of the clay liner could be artificially adjusted in the HELP
model to account for the increased resistance to percolation.
Secondly, the model assumes that the barrier soil layer is saturated for
the purposes of vertical flow calculations. This would likely be the case
after long—term operation of the landfill. However, the time period for the
wetting front to move through a very thick, impermeable clay liner to create
157
-------
the saturated condition nay be lengthy. During this period, the percolation
rate will likely be reduced below that of the saturated hydraulic conductivity.
Therefore, modeling the early life of landfills might require the use of a
reduced hydraulic conductivity for barrier soil layers. There was insufficient
information from the Sonoma County cells to test these two assumptions.
The cells at Niagara, NY, presented special problems for the purposes of
HELP model verification. The cover included a synthetic liner that theoreti-
cally eliminated surface water infiltration into the cell. Yet significant
leachate drainage was measured. Therefore, leakage into the cell must have
occurred either from surface water through failed portions of the synthetic
liner in the cover or from groundwater through the clay liner (and synthetic
liner for Cell 1) on the sides and base. The modeling difficulty was twofold.
First, there was no basis from field information for assigning a leakage frac-
tion to the synthetic liner. Second, the HELP model does not have capability
to simulate groundwater seepage into the cell. However, the analysis (for the
first 15 months) did show that assigning a reasonable leakage fraction to a
synthetic liner can reasonably reproduce measured lateral drainage volumes in
a leaky landfill.
158
-------
SECTION 12
SUMMARY AND CONCLUSIONS
Existing field data from landfill sites across the United States were
evaluated for their suitability in verifying the HELP computer model. Seven
sites were selected for analysis in this report. A total of 20 landfill cells
were simulated, ranging in size from 0.04 to 24 acres. Simulation periods
ranged from 2.5 to 8 years. Measurements of leachate drainage were available
from all landfills, while data on runoff were available from about half of the
landfills.
In most cases, daily rainfall and monthly temperature data were obtained
from the nearest National Oceanic and Atmospheric Administration weather sta-
tion for use in the model. Solar radiation values stored in the HELP model
were used for all simulations.
Model input values were determined from published reports describing the
construction and operation of each landfill. In general, little detailed
information was available on soil characteristics, surface vegetation, runoff
curve numbers, or evaporative depths, so that extensive use was made of
default values stored in the HELP model.
The measured data used for comparison with the KELP model simulations
were primarily lateral leachate drainage volumes. Measured runoff data were
available from 11 landfill cells. Barrier soil percolation was measured at
one landfill, although its suitability for model verification was limited.
There was a high degree of variability in the data from similar landfill
cells.
Where runoff data were available, an attempt was made to calibrate the
runoff curve number in advance of the simulation by examining measured
rainfall-runoff data. These calibrated curve numbers were used for the sim-
ulations, but they were consistent with default values that would have been
selected by the HELP model based on surface vegetation and minimum infiltra-
tion rates. Runoff was overpredicted for five cells by an average of 30 per-
cent and underpredicted for six cells by an average of 21 percent.
No suitable evapotranspiration field data from landfill sites were found
for model testing. However, the results of this study raise the possibility
that the evaporative depths suggested by the HELP model are too small.
Lateral drainage was overpredicted by 10 percent of the measured drainage
in two cells where very high leachate collection rates were observed. In
159
-------
three cells where very small quantities of leachate were collected, lateral
drainage was underestimated by 97 percent. Lateral drainage was overpredicted
by an average of 4 percent in five covered cells and overpredicted by an aver-
age of 53 percent in four permanently uncovered cells. Percent deviations
were not computed for the remaining cells due to the nature of their simula-
tion analysis.
In addition to field verification, a sensitivity analysis of the HEL?
model was performed to examine the effects of the major design parameters on
components of the water budget for landfills. The analysis examined the
effects of cover design, topsoil thickness, topsoil characteristics, vegeta-
tion, runoff curve number, evaporative depth, drainable porosity, plant avail-
able water capacity, hydraulic conductivity, drainage length, and liner slope
on the water budget. Hydraulic conductivity values for the topsoil, lateral
drainage layers and clay liners are the most important parameters in deter-
mining the water budget components. These parameters are particularly impor-
tant in estimating the percolation through the landfill. Other design
parameters tend to affect the apportionment between runoff, evapotranspiration
and lateral drainage from the cover.
The information from the sensitivity analysis and the verification
results were used to evaluate RCRA landfill design guidance and regulation.
This evaluation showed that saturated hydraulic conductivity is the most
important design parameter for minimizing percolation. Care should be taken
to recommend the highest hydraulic conductivity that is commonly available for
drainage media. Similarly, the lowest saturated hydraulic conductivity prac-
tically obtainable should be used as guidance for soil liners. Changes in
other design parameters yield much smaller effects on percolation if the
values of these parameters are kept in a reasonable range.
The following conclusions are made. The field data verified the utility
of the HELP model for estimating general landfill performance. However, not
all model components were well tested due to the limited field data available.
It is concluded that a laboratory and field monitoring program explicitly
designed for HELP verification would be necessary for further refinement of
specific model components. In addition, studies are needed to examine lateral
drainage and percolation for small infiltration rates and flow through syn-
thetic liners and in leakage detection of double liner systems.
The overall data base of long-tern water budget measurements at landfills
is poorly organized and too small to continually advance the state of the art
in understanding landfill leachate generation and migration. More extensive
monitoring activities are required to fill this gap.
The HELP model was shown to simulate particularly well the leachate
drainage from landfills with relatively large infiltration rates. The model
did not simulate well the leachate drainage due to very small infiltration
rates, although this could have easily been due to the selection of values
describing the cover characteristics. Runoff predictions were found to be
within an average of plus or minus 25 percent of measured data. Evapotran-
spiration verification data were lacking and should be emphasized in future
studies due to their significant impact on the overall water balance. In
160
-------
general, the error in estimates of water budget components were much smaller
than the variability in the field measurements for similar landfill cells.
These results are very good in light of the fact that the precipitation data
used in this study, which Is known to be spatially highly variable, were not
measured at most of the landfill sites.
Improvement to the HELP model should be made in the areas of snowmelt,
winter runoff, unsaturated hydraulic conductivities, and the selection of
evaporative depths.
161
-------
REFERENCES
1. Schrceder, P. R., A, C. Gibson, and M, D. Smolen. The Hydrologic
Evaluation of Landfill Performance (HELP) Model. Vol II. Documentation
for Version I. EPA/530-SW-84-010, U.S. Environmental Protection Agency,
Office of Solid Waste and Emergency Response, Washington, DC, 1984.
256 pp.
2. Schroeder, P. R., J, M. Morgan, T. M. Walski, and A. C. Gibson. The
Hydrologic Evaluation of Landfill Performance (HELP) Model. Vol I.
User's Guide for Version I. EPA/530-SW-84-009, U.S. Environmental Pro-
tection Agency, Office of Solid Waste and Emergency Response, Washington,
DC, 1984. 120 pp.
3. USDA, Soil Conservation Service. National Engineering Handbook, Sec-
tion 4, Hydrology. U.S. Government Printing Office, Washington, DC,
1972.
4. Ritchie, J. T. A Model for Predicting Evaporation from a Row Crop with
Incomplete Cover, Water Resources Research, Vol 8, No. 5, 1972.
pp 1204-1213.
5. Knisel, W. J., Jr., ed. CREAMS, A Field Scale Model for Chemical Runoff
and Erosion from Agricultural Management Systems. Vols I, II, and III,
Draft Copy, USDA-SEA, ARS, Cons. Res. Report 24, 1980. 643 pp.
6. Shanholtz, V. 0., and J, B. Lillard. A Soil Water Model for Two Con-
trasting Tillage Practices, Bulletin 38, Virginia Water Resources
Research Center, VPISU, Blacksburg, VA, 1970, 217 pp,
7. Saxton, K. E., H. P. Johnson, and R. H. Shaw. Modeling Evapctranspira-
tion and Soil Moisture. In: Proceedings of American Society of Agricul-
tural Engineers 1971 Winter Meeting, No. 71-7636, St. Joseph, MI, 1971,
8. Sudar, R. A., K, E. Saxton, and R, G. Spomer. A Predictive Model of
Water Stress in Corn and Soybeans, Transactions of American Society of
Agricultural Engineers, 1981. pp 97-102.
9. Skaggs, R. W. Modification to DRAINM0D to Consider Drainage from and
Seepage Through a Landfill. Draft Report, U.S. Environmental Protection
Agency, Cincinnati, OH, 1982. 21 pp.
10. Code of Federal Regulations. Title 40, Part 264. Office of the Federal
Register, General Services Administration, Washington, DC, July 1985,
162
-------
11. U.S. Environmental Protection Agency. Minimum Technology Guidance on
Double Liner Systems for Landfills and Surface Impoundments—Design, Con-
struction, and Operation. EPA/530-SW-85-014, Washington, DC, May 24,
1985.
12. U.S. Environmental Protection Agency. RCRA Guidance Document for Land-
fill Design. Federal Register, General Services Administration, Washing-
ton, DC, July 26, 1982.
13. Ham, R. K. Decomposition of Residential and Light Commercial Solid Waste
in Test Lysimeters. SW-190c, U.S. Environmental Protection Agency,
Office of Solid Waste, Washington, DC, 1980. 103 pp.
14. EMCON Associates. Sonoma County Solid Waste Stabilization Study.
E?A/530-SW-65d.1, U.S. Environmental Protection Agency, Office of Solid
Waste Management Programs, Washington, DC, 1975. 283 pp.
15. Wigh, R, J. Landfill Research at the Boone County Field Site, Final
Report. Purchase Order No. C3016NASX. Municipal Environmental Research
Laboratory, Office of Research and Development, U.S. Environmental Pro-
tection Agency, Cincinnati, OH (NTIS PB 84-161546). 116 pp.
16. EMCON Associates. Field Assessment of Site Closure, Boone County, Ken-
tucky. Contract No. 68-03-2824/02. Municipal Environmental Research
Laboratory, Office of Research and Development, U.S. Environmental Pro-
tection Agency, Cincinnati, OH (NTIS PB 83-251629).
17. Gordon, M. E., Huebner, P. M., and Kmet, P. An Evaluation of the Perfor-
mance of Four Clay-Lined Landfills in Wisconsin. Bureau of Solid Waste
Management, Wisconsin Department of Natural Resources.
18. Recra Research, Inc. Response to USEPA Region II Administrative Order On
Consent, Index No. II RCRA-85-3013-50201, Item I.A.3. Amherst, New York,
April 9, 1985.
163
-------
TECHNICAL REPORT DATA
/Please read lntiruction% on she rtt ene before completing)
1 REPORT NO. 2,
EPA/600/2-87/050
3 RECIPIENT'S ACCESS ONO
PB87- 22 7 s i arm
4. title AND Subtitle
Verification of trie Hydrolcgic Evaluation of Landfill
Performance (HELP) Model Using Field Data
5 REPORT DATE
July 1987
6. PERFORMING ORGANIZATION CCOE
?. AUTHQRlSl
P.R. Schroeder and R. L. Peyton
8. PERFORMING ORGANIZATION REPORT NO.
9 PERFORMING ORGANIZATION NAME AND ADDRESS
U.S. Ar.-v Engineer Waterways Experiment Station
Vicksuurg, Mississippi 39180
io. program element no.
11. CONTRACT,GRANT NO.
DW-96930236
12. SPONSORING AGENCY NAME AND ADDRESS
Hazardous Waste Engineering Research Laboratory
Office of Hesearcn and Development
U.S. Environmental Protection Agency
Cinrina'i . n,-in d^?f-,P.
13. TYPE OF REPORT AND PERIOD COVERED
14, SPONSORING AGENCY CODE
EPA/600/12
* 5. SUPPLEMENTARY NOTES
Project. Officer: Douglas C. Airaaon
16. ABSTRACT
This report describes a study conducted to verify tiie Hydrologic Evalua-
tion cf Landfill Performance (HELP) computer model using existing field dara
fron a total of 20 lancfill cells at 7 sites in the United States. Simula-
tions using the HELP model were run to compare the predicted water balance
with the measured water balance. Comparisons were mace for runoff, evapo-
transpiration, lateral drainage to collection systems and percolation through
liners. The report also presents a sensitivity analysis of the HELP model
input parameters.
17, KEY WORDS AND DOCUMENT ANALYSIS
a. descriptors
b.lDENTIf'ERS/OPEN ENDED T E P V S
c, cosati Field-Group
13. D.STRIS'JTION STATEMENT
Release to Public
19. SECURITY CLASS (This Heporij
Unclassified
21. NC. OF PAGES
180
20. SECURITY CLASS (This page/
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |