Benson Et Al-1999

FIELD PERFORMANCE
OF
COMPACTED CLAY LINERS
Downloaded from ascelibrary.org by INDIAN INST OF TECHNOLOGY - CHENNAI on 08/06/13. Copyright ASCE. For personal use only; all rights reserved.
By Craig H. Benson,1 David E. Daniel,2 and Gordon P. Boutwell,3 Members, ASCE

ABSTRACT: A database consisting of 85 full-scale compacted clay liners was assembled to evaluate field
hydraulic conductivity KF. Large-scale field hydraulic conductivity tests were conducted on each liner. All of
the clay liners were intended to achieve KF 1 107 cm/s, but only 74% succeeded in meeting this objective.
The important factors affecting KF are the soil wetness relative to the line of optimums and the number of lifts
or the thickness of the liner. Poor correlation exists between KF and hydraulic conductivities measured in the
laboratory (KL) on field-compacted samples (especially for liners with KF > 107 cm/s compacted at lower water
contents relative to the line of optimums), percent compaction, and index properties representative of composition
(e.g., Atterberg limits and particle size fractions). The findings indicate that (1) compacted clay liners having
KF 107 cm/s can be constructed with a broad variety of clayey soils; (2) the primary emphasis should be
ensuring compaction is primarily wet of the line of optimums; (3) less emphasis should be placed on other
traditional measures such as percent compaction, KL, and index properties; and (4) liners that are thicker or have
a greater number of lifts tend to have lower KF.
INTRODUCTION
In the past decade, dozens of papers have been written on
factors that influence the hydraulic conductivity of compacted
clays used for liners and covers in waste containment systems
[e.g., Boynton and Daniel (1985), Kleppe and Olson (1985),
Benson and Daniel (1990), Daniel and Benson (1990), Benson
and Othman (1993), Daniel and Wu (1993), Shelley and Daniel (1993), and Othman et al. (1994)], laboratory and field
hydraulic conductivity testing methods [e.g., Daniel (1989,
1994), Sai and Anderson (1990), Benson et al. (1994a), and
Trautwein and Boutwell (1994)], the correlation of laboratory
and field hydraulic conductivities [e.g., Day and Daniel
(1985), Elsbury et al. (1990), Reades et al. (1990), Benson and
Boutwell (1992), Benson et al. (1994a)], and methods of construction and construction quality assurance [e.g., Benson et
al. (1994b), Daniel and Koerner (1995), and Benson et al.
(1997)]. Despite this wealth of information, the hydraulic performance of compacted clay liners in the field is largely undocumented. In fact, only a few studies have characterized the
field performance of in-service compacted clay liners constructed using standards typical of industry [e.g., Gordon et al.
(1990), Reades et al. (1990), and Benson and Boutwell
(1992)].
There are two principal reasons for the lack of performance
data. First, compacted clays are frequently used in combination with geomembrane liners, and in such applications it is
difficult to separate the performance of the compacted clay
component from that of the geomembrane-clay composite.
Second, few landfills have lysimeters or other instrumentation
installed directly beneath the lining system that enable documentation of field performance. Although lysimeters have been
used to document field performance of several clay liners in
landfills (Gordon et al. 1990; Reades et al. 1990; Benson and
Boutwell 1992), not many clay liners (9% in this study) include lysimeters. Future opportunities to document field performance of clay liners using lysimetry will be limited because
clay liners are now infrequently used in lining systems without
a geomembrane.
1
Assoc. Prof., Dept. of Civ. and Envir. Engrg., Univ. of Wisconsin,

Madison, WI 53706. E-mail: chbenson@facstaff.wisc.edu
2
Prof. and Head, Dept. of Civ. Engrg., Univ. of Illinois, Urbana, IL
61801. E-mail: dedaniel@uiuc.edu
3
Pres., Soil Testing Engrs., Inc., Baton Rouge, LA 70884.
Note. Discussion open until October 1, 1999. To extend the closing
date one month, a written request must be filed with the ASCE Manager
of Journals. The manuscript for this paper was submitted for review and
possible publication on March 6, 1998. This paper is part of the Journal
of Geotechnical and Geoenvironmental Engineering, Vol. 125, No. 5,
May, 1999. ASCE, ISSN 1090-0241/99/0005-0390 0403/$8.00 $.50
per page. Paper No. 17884.
The purpose of this study was to collect as much information as possible on the field performance of compacted clay
liners. In view of the dearth of information on actual performance of in-service clay liners, the next best source of data
was also used: large-scale hydraulic conductivity tests on fullscale field test pads. The test pads were constructed with materials, methods of construction, and quality assurance procedures that are typical of industry practice for landfill liners.
Test pads constructed for research purposes were not included.
This paper has two objectives: (1) Presentation and documentation of the database in sufficient detail to permit independent analysis by others; and (2) discussion of the key factors affecting KF that are particularly important to practicing
engineers trying to achieve KF 107 cm/s.
DATABASE
Data Sources
The database was assembled primarily from unpublished
data obtained from the writers files (85% of the database) and
supplemented by information in the literature (15% of the database). Most of the data were obtained from engineering reports documenting results from test pads. The data collection
process is thought to have captured the results of 50 75% of
all test pads in North America where large-scale field hydraulic
conductivity tests have been conducted.
Information was collected and screened from more than 120
sites. The requirements for inclusion in the database were (1)
construction methods in general accord with industry practices
for full-scale liners; (2) construction quality assurance in general accord with industry practices; (3) construction with the
objective of achieving KF 1 107 cm/s; (4) reasonably
complete documentation of construction test data; and (5)
availability of results from large-scale field hydraulic conductivity tests. Clay liners constructed from soil-bentonite
mixtures were not included in the database.
The database obtained using these requirements consists of
85 compacted clay liners, of which eight are actual in-service
liners for landfills and 77 are test pads. The database is summarized in Tables 1 and 2. Many of the numbers in Tables 1
and 2 are averages of multiple measurements (geometric mean
for KF and KL, arithmetic mean for all other data). Specific
sites are not identified due to potential sensitivities. The sites
are distributed geographically throughout the United States
and Canada. As a result, the soils included in the database
comprise a broad spectrum of material types. For example, the
average liquid limit (LL) varies from 21 to 101, the fines content varies from 48 to 99%, and the 2-m clay content varies
from 16 to 57%. Nearly all of the soils classify as either CL
390 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999
J. Geotech. Geoenviron. Eng. 1999.125:390-403.
TABLE 1.
Index Properties and Compaction Characteristics

dm
Compaction
(kN/m3)
effort
(7)
(8)
w
(%)
(9)
d
(kN/m3)
(10)
Po
(11)
10.3
19.8
44
26.6
13.8
14.4
19.4
28
98
SP
21.3
16.0
80
18.6
MP
17.3
17.3
95
13.5
19.5
MP
13.8
19.0
32
27
14.1
18.6
SP
17.2
17.7
88
75
45
14.5
18.8
MP
15.3
17.7
45
12.7
18.6
MP
19.6
17.0
90
21
29
16.6
18.7
MP
17.8
16.9
50
57
30
39
21.7
17.3
MP
25.4
16.0
75
12
55
28
33
23.0
16.6
MP
26.0
16.1
78
13
37
15
78
37
18.0
17.0
SP
20.7
16.7
100
14
40
20
70
25
16.2
16.7
SP
17.0
16.8
78
15
85
58
99
57
25.8
14.6
SP
30.8
14.1
98
16
41
22
77
38
15.8
17.0
SP
19.8
16.1
91
17
50
34
95
47
20.3
16.4
SP
23.3
15.7
100
18
30
18
52
16
13.0
18.7
SP
16.6
17.4
85
19
32
14
85
44
10.5
20.1
MP
13.6
19.0
81
20
21
49
51
23
26
94
90
43
36
18.5
11.8
17.2
18.5
SP
MP
17.6
19.5
16.9
16.9
8
80
22
23
63
39
42
18
96
73
30
20.5
20.0
16.3
16.5
SP
SP
22.0
16.4
89
24
67
46
94
53
81
41
88
36
18.9
16.9
71
26
33
19
85
37
15.5
17.6
17
27
31
18
74
26
13.5
18.0
28
35
19
89
41
16.2
17.7
57
29
27
10
76
28
13.9
18.8
84
30
31
32
40
19
24
58
23
MP
SP
MP
SP
MP
SP
RP
MP
SP
RP
MP
SP
RP
MP
SP
RP
MP
SP
15.8
53
18.4
16.3
19.8
18.0
19.3
17.7
17.1
19.5
17.7
16.8
19.4
17.5
17.0
20.5
19.1
18.6
18.6
19.3
23.6
25
16.0
21.5
11.5
16.1
12.2
17.5
18.5
12.1
16.7
18.9
11.5
16.6
18.5
9.0
13.0
14.4
14.0
12.4
16.2
13.1
18.6
19.1
65
75
32
45
27
99
42
11.0
19.9
MP
13.9
19.2
92
33
29
15
87
40
13.3
18.9
MP
13.4
18.7
80
34
44
16
96
17.3
17.1
SP
17.8
17.1
45
35
39
19
97
22.2
16.4
SP
20.7
16.8
78
36
36
17
74
30
13.2
18.3
SP
17.6
77
37
36
17
48
16
12.4
19.0
SP
15.5
15.5
14.1
18.2
45
Site
number
(1)
LL
(2)
PI
(3)
Fines
(%)
(4)
Clay
fraction
(5)
wo
(%)
(6)
24
10
65
37
2
3
58
25
29
10
85
85
50
22
10.2
9.0
26.8
12.3
20.1
21.3
14.6
19.0
SP
MP
SP
SP
50
34
95
47
17.9
16.8
43
26
87
32
14.3
32
19
88
35
33
13
77
35
22
55
31
10
43
11
Compaction
criterion
(12)
w > wo
PC > 90% dm MP
None
w > wo
PC > 95% dm SP
w > wo 2, <8
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo 2
PC > 90% dm MP
w > wo 2 to 4
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo 2, <5
PC > 90% dm MP
w > wo 2, <5
PC > 90% dm MP
w > wo
PC > 100% dm SP
w > wo 2, <5
PC > 90% dm MP
w > wo 2, <6
PC > 90% dm SP
Si > 78.5
PC > 90% dm MP
w > wo
PC > 90% dm MP
Si > 82.0
w > wo
PC > 95% dm SP
Si > 85.0
w > wo
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo 2, 5
PC > 90% dm MP
PC > 90% dm MP
w > wo
PC > 95% dm SP
w > wo, <6
PC > 90% dm MP
w > wo
PC > 90% dm MP
w > wo
PC > 95% dm SP
w > wo
PC > 95% dm SP
w > wo, <6
PC > 95% dm SP
w > wo
PC > 95% dm SP
JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999 / 391
TABLE 1.
(Continued )
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
38
21
60
26
10.3
20.4
SP
11.5
20.4
39
21
60
26a
10.3
20.4
SP
11.6
17.9
10
31.6
13.4
SP
35.5
100
w > wo
PC > 90% dm MP
w > wo 0.7, <wo 1.7
PC > 90% dm SP
w > wo, 5
PC > 92% dm SP
PC > 95% dm SP
w > wo 1
PC > 95% dm SP
w > wo 1
PC > 90% dm SP
w > wo, <2
PC > 95% dm SP
w > wo 1, <3
PC > 95% dm SP
w > wo 1, <5
PC > 90% dm SP
w > wo, <3
PC > 95% dm SP
w > wo, <3

PC > 95% dm SP
w > wo, <3

PC > 95% dm SP
w > wo, <3

PC > 95% dm SP
w > wo, <3

PC > 95% dm SP
PC > 95% dm SP
PC > 91% dm MP
w > wo 1, <5
PC > 95% dm SP
w > wo 1, <5
PC > 95% dm SP
w > wo 1, <5
PC > 95% dm SP
w > wo 1, <5
PC > 95% dm SP
w > wo 2, <4
PC > 90% dm MP
w > wo 2, <4
PC > 90% dm MP
w > wo, <6
PC > 95% dm SP
w > wo, <SL
PC > 95% dm SP
w > wo, <SL
PC > 95% dm SP
w > wo 3, <6
PC > 95% dm SP
w > wo 3, <5
PC > 95% dm SP
w > wo 0.4
PC > 98% dm SP
w > wo 1.5
PC > 94% dm SP
w > wo
PC > 95% dm SP
w > wo
PC > 95% dm SP
w > wo 4
40
101
71
98
49
41
42
43
47
69
62
30
45
42
66
79
86
49
19.5
23.4
22.4
16.3
15.1
15.4
SP
SP
SP
21.9
25.0
23.4
12.8
12.8
16.0
15.1
15.4
44
62
42
86
22.4
15.4
SP
24.2
15.0
47
45
44
28
70
19.5
16.4
SP
19.8
16.3
71
46
35
16
98
22
23.3
15.4
SP
27.3
15.4
100
47
39
24
70
14.6
17.7
SP
16.5
17.7
100
48
41
23
86
75
22
86
18.9
16.7
86
50
43
24
86
18.6
16.9
84
51
40
22
86
17.8
17.0
73
52
37
18
73
38
RP
SP
MP
RP
SP
MP
RP
SP
MP
RP
SP
MP
SP
17.0
42
16.2
16.7
18.7
16.2
16.7
18.7
16.2
16.7
18.7
16.2
16.7
18.7
16.5
17.8
49
20.0
18.0
13.3
20.0
18.0
13.3
20.0
18.0
13.3
20.0
18.0
13.3
19.9
21.2
16.1
67
53
54
55
56
57
58
59
60
61
62
63
64
65
66
54
66
66
69
69
69
68
68
51
51
47
47
50
31
35
35
38
38
38
35
35
20
20
30
31
29
93
93
98
98
98
95
95
73
73
66
66
75
40
19.9
27.4
27.4
26.8
26.8
26.8
26.6
26.6
20.2
20.2
19.5
13.5
19.0
16.4
14.5
14.5
14.6
14.6
14.6
14.6
14.6
15.9
15.9
16.3
19.2
16.1
SP
SP
SP
SP
SP
SP
SP
SP
SP
SP
SP
MP
SP
21.6
27.0
30.6
29.6
30.7
29.4
26.8
29.8
24.6
22.7
21.6
17.2
21.7
15.5
15.0
14.2
14.0
14.3
14.4
15.1
14.4
15.4
15.4
16.0
17.4
17.2
71
88
0
95
67
49
27
62
19.3
16.1
SP
21.4
17.2
98
68
35
17
67
22
100
50
16
11.5
19.4
75
70
42
26
88
45
SP
MP
SP
MP
MP
18.0
22
17.7
19.0
19.9
21.4
18.7
17.6
69
14.8
11.5
10.0
8.5
14.9
20.6
16.1
60
71
29
19
83
34
12.2
19.6
MP
14.3
18.0
64
72
36
20
85
35
18.0
16.5
SP
23.7
15.5
100
73
76
53
21.0
15.5
25.2
14.8
97
74
56
40
64
18.0
16.9
19.6
16.1
47
75
21.0
15.6
SP
25.4
15.2
100
76
37
17
92
19.2
16.6
SP
21.8
15.9
86
77
32
13
19
9.9
19.7
SP
11.0
19.2
51
78
32
16
25
11.5
19.6
SP
12.4
18.8
63
79
62
41
82
35
84
23.1
14.9
81
47
22
SP
MP
SP
MP
SP
16.2
52
14.9
16.5
15.9
18.0
15.3
28.2
80
25.0
17.8
19.6
14.4
25.0
28.0
14.2
92
81
63
TABLE 1.
(Continued )
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
82
83
84
85
39
16
84
81
54
48
18.2
17.6
SP
17.8
15.7
17.1
18.4
(12)
PC
PC
PC
PC
>
>
>
>
96%
96%
90%
90%
dm
dm
dm
dm
SP
SP
MP
MP
Note: LL = liquid limit; PI = plasticity index; w = compaction water content; wo = optimum water content; PC = percent compaction; dm = maximum
dry unit weight; MP = modified Proctor (ASTM D 1557); SP = standard Proctor (ASTM D 698); and RP = reduced Proctor (Daniel and Benson 1990).
a
4-m clay fraction.
TABLE 2.
Site
number
(1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Compactor
weight
(Mg)
(2)
32.4
30.0
19.8
36.0
32.4
32.4
12.6
32.4
39.0
19.8
25.0
32.4
32.4
32.4
18.9
18.9
18.9
18.9
18.9
32.4
27.0
32.4
19.8
19.8
32.4
17.1
19.8
19.8
7.2
7.2
19.8
10.9
19.8
19.8
19.8
19.8
19.8
19.8
19.8
19.8
19.8
Construction Information and Hydraulic Conductivity

Hydraulic Conductivity (cm/s)
Passes
per lift
(3)
Lift
thickness
(cm)
(4)
Number
of lifts
(5)
6
6
4
6
5
4
4
6
4
5
8
6
8 12
8 12
46
46
46
6
4
6
8
4
12
12
40
16
8
6
8
6 10
58
3
3
45
12
6 (lifts 1 3)
21 (lift 4)
6
15
15
15
15
15
15
15
15
15
15
15
15
15
17
20
17
15
15
15
15
15
15
15
15
15
15
15
17
15
30
15
15
15
15
6
6
15
13
15
15
15
15
15
15
20 25
8.5
15
15
15
15
15
16
13
13
14
6
5
8
4
10
6
8
6
10
10
10
10
8
6
7
5
4
5
10
6
6
10
5
5
6
6
6
9
6
2
8
8
3
6
6
5
6
6
6
4
4
5
5
4
5
10
4
4
4
4
4
4
4
4
4
4
Thin-wall
tube
(6)
3.2
3.6
8.0
5.0
8.7
2.4
8.4
9.0
1.0
8.0
2.0
3.0
1.3
4.8
4.4
3.7
3.0
1.5
1.9
3.0
3.1
2.4
1.5
9.0
2.3
2.9
3.0
1.9
2.2
3.0
1.6
3.0
1.3
1.5
3.0
9.1
4.9
2.6
3.5
5.5
2.4
2.4
5.8
1.5
1.1
5.1
7.4
4.1
1.7
8.1
2.8
2.8
3.4
108
109
109
109
109
108
108
109
108
109
109
109
108
108
109
108
109
108
108
109
107
108
108
109
109
109
108
108
108
108
108
108
108
108
108
109
108
108
109
109
109
109
109
108
108
108
108
108
108
108
108
108
108
SDRI
(7)
2.8 107
1.5 107
1.1 108
9.0 109
2.7 107
5.8 108
1.2 107
1.3 108
2.0 108
3.3 109
3.0 108
9.8 109
8.0 107
2.5 107
2.0 108
1.5 108
8.0 109
2.0 107
1.8 107
9.0 108
1.7 108
1.1 107
6.0 108
3.9 108
3.9 108
4.0 107
3.7 108
3.0 108
1.3 108
<3.6 108
3.0 108
2.2 108
1.0 107
8.0 108
7.0 108
2.0 107
3.7 108
2.0 108
5.0 108
4.0 108
5.0 108
2.6 107
3.0 107
1.1 107
2.2 108
7.0 108
1.3 107
2.4 108
2.4 108
5.6 108
Lysimeter
(8)
9.0
7.0
3.0
3.0
2.0
6.0
4.4
1.4
109
109
108
109
109
109
108
108
TSB
(9)
4.3
1.4
1.6
5.0
9.2
4.7
1.6
2.1
3.2
7.5
1.1
1.2
108
108
108
109
109
108
108
108
107
108
107
108
300-mm
block
(10)
2.6 107
4.0 108
1.4 108
2.2 107
1.1 108
6.0 109
1.8 107
1.5 107
1.7 107
1.7 108
3.5 107
4.1 109
4.8
7.7
3.1
5.3
108
108
106
107
TABLE 2.
(1)
(2)
(3)
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
19.8
19.8
19.8
19.8
19.8
19.8
19.8
19.8
19.8
19.8
19.8
19.8
32.4
19.8
19.8
19.8
19.8
19.8
19.8
19.8
10.2
10.2
19.8
19.8
16.2
32.4
32.4
19.8
19.8
19.8
19.8
12
12
12
12
12
12
40
80
2
2
6
6
7 10
7 10
8
22
22
22
22
6
10
4
8
8
4
6
6
74
75
76
77
78
79
80
81
82
83
84
85
1
6
1
6
1
3
1
3
1
3
(4)
(5)
15
17
17
19
15
23
15
15
10
10
15
15
15
15
15
at 20
at 10
at 20
at 10
at 20
at 15
15
10
10
15
15
15
15
15
at 30
at 20
at 30
at 20
4
4
4
4
4
4
4
4
10
11
4
4
6
6
4
7
7
7
7
4
4
4
9
9
8
8
4
4
4
4
4
4
4
or CH under ASTM D 2487 and span the clay range of the

plasticity chart defined in that standard (i.e., between the A
and U lines and above the hatched zone).
(Continued
(6)
2.5
2.7
3.4
4.3
1.6
1.7
5.5
3.7
3.0
7.8
2.1
2.0
2.0
1.4
4.7
3.3
1.8
4.2
1.5
1.7
108
108
108
108
107
107
109
108
108
109
108
108
108
108
108
109
109
108
108
108
(7)
5.0
9.4
1.2
3.7
3.1
3.9
2.3
1.8
1.2
8.3
2.3
1.3
4.0
8.3
2.0
8.0
8.0
9.0
9.0
5.0
5.0
3.0
3.0
3.0
4.5
4.0
1.5
3.0
4.5
1.3
1.3
2.8
2.8
108
108
107
108
107
107
107
107
108
108
108
108
108
108
108
108
108
108
108
108
108
108
108
108
108
108
107
108
108
107
107
108
108
(8)
(9)
108
108
108
108
1.1
8.5
2.6
5.6
(10)
4.1 109
hole tests (Trautwein and Boutwell 1994) and/or laboratory

tests on large (300-mm-diameter) undisturbed block specimens
(Benson et al. 1994a; Trast and Benson 1995). Results from
these tests are also in the database for completeness.
Field Hydraulic Conductivity

A variety of methods can be used to measure KF, including
surface infiltration measurements, borehole techniques, lysimetry, and tests on large block specimens (Daniel 1989; Benson et al. 1994a; Trautwein and Boutwell 1994). Benson et al.
(1994a) indicate that an area of 0.1 m2 is the minimum area
required to obtain a representative field hydraulic conductivity
measurement. For most of the sites in the database, only lysimeters and sealed double-ring infiltrometers (SDRIs) met
this criterion, but at some sites large-scale laboratory tests
were conducted on block specimens of sufficient size. However, because the lysimeters and SDRIs tested the largest volume of soil, they were selected to define KF. Most SDRIs
permeated an area of 2.3 m2, but a few were as small as 1.4
m2. One lysimeter covered an area of 0.37 m2, but the rest
covered 64 225 m2.
Hydraulic conductivity was calculated from SDRI data using the wetting-front method with a wetting-front suction of
zero (Trautwein and Boutwell 1994). No correction was made
for swelling of the soil, but for nearly all of the SDRI tests
swelling had ceased by the latter part of the test period when
KF is determined. These assumptions result in a measured KF
that can be slightly larger (less than a factor of 2) than the true
KF when the depth of the wetting front Df penetrates at least
half the thickness of the liner L (Wang and Benson 1995). For
94% of the sites in the database, Df /L was >1/2. Hydraulic
conductivity was computed from lysimeter data following the
method in Daniel (1989). If unknown, the leachate depth was
assumed to be 300 mm (the common maximum depth stipulated by regulations).
At several sites, KF was also measured using two-stage bore-
OVERVIEW OF HYDRAULIC CONDUCTIVITY DATA

A histogram of the logarithms of KF is shown in Fig. 1. The
geometric mean KF is 4.5 108 cm/s, whereas the highest
KF is 8.0 107 cm/s and the lowest is 2.0 109 cm/s. All
test pads and clay liners in the database were intended to have
KF 107 cm/s. Sixty-three of them (74% of the database)
achieved this objective, whereas 22 (26% of the database) did
not. Of the 22 sites that failed to achieve KF 107 cm/s,
most just barely failed; 12 had KF 2 107 cm/s and 17
had KF 3 107 cm/s. Nevertheless, a hydraulic conductivity of 1 107 cm/s is the maximum typically allowed by
FIG. 1.
Histogram of Field Hydraulic Conductivities
regulation, and failure to meet this requirement is usually

grounds for denying a construction permit.
The database includes information from eight lysimeters
that were located beneath actual compacted clay landfill liners.
The geometric mean KF from the lysimeters is 9 109 cm/s,
which is smaller than the geometric mean KF from the 77
SDRI tests (5 108 cm/s). Comparison of the average KF
obtained from lysimeters and the average KF obtained from
the SDRI tests using a t-test showed that KF from the lysimeters is significantly smaller at the 5% significance level.
However, there was no statistically significant difference between the water content relative to the line of optimums, percent compaction, or classification data for the sites tested with
SDRIs and lysimeters. Higher average KF was probably obtained from the SDRIs because they are conducted at lower
effective stress than exists in an in-service liner at a landfill.
For example, KF measured with the SDRIs and other tests used
to measure KF at low effective stress [two-stage borehole
(TSB) tests, tests on large block specimens] are similar (Fig.
2). Moreover, Trast and Benson (1995) show that KF can decrease an order of magnitude when the effective stress increases as a landfill is filled. Capillary barrier effects may also
have resulted in lysimeters yielding KF that is lower than the
true field hydraulic conductivity, particularly for those lysimeters using a gravel backfill (Chiu and Shackelford 1994).
Nevertheless, there are numerous applications where the effective stress is low, and the results from the SDRIs are directly applicable (e.g., cover systems, lagoons, and surface impoundments).
One of the most contentious issues in quality assurance testing is the degree to which hydraulic conductivities obtained
from laboratory tests (KL) conducted on undisturbed, 75-mmdiameter specimens provide an accurate indication of KF. Several investigators have shown that for poorly built liners, KL
can be orders of magnitude lower than KF, whereas the two
values are often similar for well-built liners (Daniel 1984; Day
and Daniel 1985; Reades et al. 1990; Benson and Boutwell
1992; Benson et al. 1994a). A similar conclusion can be drawn
from the database. The relationship between KF and KL is
shown directly in Fig. 3, and the ratio KF/KL is shown as a
function of KF in Fig. 4. In both cases, KF is shown with closed
circles for liners where the percentage of the field compaction
data falling wet of the line of optimums Po was greater than
80%. The significance of this factor will be illustrated later. In
Figs. 3 and 4, the field hydraulic conductivity is usually greater
than KL when KF is high (i.e., >107 cm/s), whereas KF and
KL are comparable when KF is low (i.e., 108 cm/s). For the
FIG. 2. Hydraulic Conductivities Measured with SDRIs versus

Hydraulic Conductivities Measured with TSBs and on 300-mmBlock Specimens
FIG. 3. Field Hydraulic Conductivity versus Hydraulic Conductivity Measured in Laboratory on Small Undisturbed Specimens. Percentage of Water Content-Dry Unit Weight Points Unknown for Sites Marked Po = ?
FIG. 4. Ratio of Field to Laboratory-Hydraulic Conductivity

versus Field Hydraulic Conductivity. Percentage of Water Content-Dry Unit Weight Points Was Unknown for Sites Marked
Po = ?
reported data, the ratio KF/KL is 11.2 on average when KF >

107 cm/s, whereas KF/KL is 0.78 on average when KF < 108
cm/s.
The reasons for differences between KF and KL are many,
as discussed in the literature [e.g., Benson and Boutwell
(1992) and Benson et al. (1994a)]. The general belief is that
KF and KL are similar when evaluated at similar effective stress
and that the clay liner is devoid of macroscopic features (e.g.,
macropores, cracks, fissures, and heterogeneous materials) or
is at effective stresses high enough to close macroscopic features (Boynton and Daniel 1985). Liners compacted wet of the
line of optimums are usually devoid of macroscopic features
and thus have low KF. Under such conditions, only microscale
pores are present to conduct flow, and they are readily represented in small laboratory-scale specimens. In contrast, macroscopic defects such as interclod voids, macropores, cracks,
fissures, etc. are not accurately represented in small laboratoryscale specimens or are closed during sampling or testing. Macroscopic defects are common when construction practice is
poor. Their presence results in high KF and KF >> KL. Not
surprisingly, the worst correlation between KF and KL occurs
for one of the highest KF in the database (Site 20, Fig. 4 and
Table 1). The sites with higher KF/KL were generally compacted drier relative to the line of optimums (i.e., data shown
as Po < 80% in Figs. 3 and 4) a condition known to result
in macroscopic features and scale-dependent KF (Benson and
Daniel 1990; Benson and Boutwell 1992; Benson et al.
1994a).
Laboratory tests on 75-mm-diameter specimens obtained
from compacted clay liners are a routine part of many construction quality assurance programs, and the results of these
tests often form the principal basis for pass/fail decisions.
Twenty-two clay liners in the database failed to achieve a KF
107 cm/s. Of the 18 failing clay liners for which KL data
are available, the laboratory tests incorrectly indicate that 15
of them pass the 1 107 cm/s requirement. That is, KL provided a false positive (pass) in 15 of 18 (83%) of the failing
test pads. The database clearly shows that laboratory hydraulic
conductivity tests have a strong tendency to yield unconservative (passing) values for failing clay liners (KF > 107 cm/s).
Consequently, they have limited value for assessing construction quality, especially for detecting failing liners. The best
that can be said for small-scale laboratory tests is that a failing
KL is occasionally measured and that in all three such instances
in the database KF was also >107 cm/s (i.e., there were no
false negatives in the KL values). Thus, laboratory tests are
occasionally successful in identifying unsatisfactory KF in
those rare cases where KL > 107 cm/s.
position as described by LL, PI, and percent fines; redundancy

of the liner in terms of lift thickness, number of lifts, and liner
thickness; wetness of the soil at compaction in terms of percentage of field compaction data wet of the line of optimums
Po, average percent wet of an optimum water content wo, and
the average initial degree of saturation Si (i.e., the average
degree of saturation at compaction); and dry unit weight d in
terms of percent compaction. Gravel content, clay content,
compactor weight, and number of passes were not included in
the regression analysis because these variables were unavailable for more than 20 sites. Some of these regression variables
are correlated, as will be discussed subsequently.
The initial Si defines the average wetness at compaction relative to the line of optimums. Lines of constant Si typically
fall parallel to the line of optimums, and the line of optimums
typically corresponds to a contour of Si 85% (Boutwell and
Hedges 1989; Benson et al. 1994b; Blotz et al. 1998). The
parameter Po describes the average wetness and the amount of
scatter relative to the line of optimums. An example illustrating
how Po is determined is shown in Fig. 5. Liners with higher
Po are usually constructed at higher average water content and
dry unit weight relative to those combinations of water content
and dry unit weight comprising the line of optimums.
STATISTICAL ANALYSIS
Various studies including those cited in this paper indicate
that the factors influencing KL are soil composition, water content relative to the line of optimums, compactive effort, and
effective stress. For KF, the thickness or number of lifts must
also be considered (Boutwell and Rauser 1990; Benson and
Daniel 1994). Statistical analyses were therefore conducted to
identify if these variables had a significant effect on KF.
The first analysis consisted of t-tests to determine which, if
any, of the aforementioned variables were statistically different
in the 53 clay liners where KF was 107 cm/s relative to
those for the 22 clay liners where KF was >107 cm/s. Neither
soil composition [in terms of LL, plasticity index (PI), and
percent fines] nor dry unit weight was significantly different,
even at a significance level of 0.30. Relative wetness (in terms
of average saturation or percent of compaction data falling wet
of the line of optimums) and thickness (or number of lifts)
were significant, even at the 0.005 significance level.
A more comprehensive analysis was subsequently conducted using stepwise regression. The analysis was similar to
that used by Benson et al. (1994b) using principles described
in Draper and Smith (1981). In short, stepwise regression identifies independent variables (e.g., compaction parameters, index properties, etc.) having a statistically significant effect on
a dependent variable (in this case log KF) by examining the
linear correlation between the independent and dependent variables using the partial-F statistic. At the common significance
level of 0.05, statistically significant variables have partial-F
4. Borderline significant variables have partial-F < 4 and
>2, whereas statistically insignificant variables have partial-F
< 2. The regression proceeds in steps, identifying the most
significant independent variable in the first step and the less
significant variables in subsequent steps after removing the
correlation attributed to the most statistically significant independent variable. The base-10 logarithm of KF was used in the
regression analysis, because it is nearly normally distributed
(Fig. 1). More detail on stepwise regression can be found in
Benson et al. (1994b) and Draper and Smith (1981).
The first regression analysis was conducted with the entire
database, using the following independent variables: soil com-
FIG. 5.
TABLE 3.
Example Showing How Po Is Computed

Results of Stepwise Regression Analysis
Entire Databasea
Independent
variable
(1)
Percent wet line of
optimums
Liner thickness
Liquid limit
Plasticity index
Percent fines
Lift thickness
Percent wet of optimumc
Percent compaction
Initial saturation
a
Sites Wet of Line

of Optimums
(Po 80%)b
Partial-F
statistic
(2)
Significant?
(3)
Partial-F
statistic
(4)
Significant?
(5)
11.3
4.7
0.63
0.64
3.10
1.11
Yes
Yes
No
No
Borderline
No
5.43
0.31
3.13
0.88
0.37
Yes
No
Borderline
No
No
No
No
No
0.47
0.32
0.15
No
No
No
0.74
0.65
0.46
5 sites in regression analysis.

29 sites in regression analysis.
c
Partial-F = 6.8 when Po removed from analysis on entire database.
b
A second stepwise regression was conducted on the portion

of the database with sites having Po 80% (i.e., those clay
liners compacted wet of the line of optimums). Stepwise regression was conducted using the same independent and dependent variables (except Po) previously used. Results of both
regression analyses are in Table 3.
IMPORTANCE OF COMPACTION SPECIFICATIONS
Line of Optimums
The first regression analysis indicates that the only two independent variables are statistically significant (Table 3): Po
and liner thickness. The relationship between KF and Po is
illustrated in Fig. 6(a); KF decreases with increasing Po , and
KF is almost always <107 cm/s when Po > 80%. The data
indicate that lower KF should be expected when a greater portion of a test pad or liner is compacted wet of the line of
optimums. Discussion of the importance of thickness is in a
subsequent section.
Numerous laboratory studies over the past 40 years have
shown that the hydraulic conductivity of compacted clay varies orders of magnitude from dry to wet of the line of optimums (Lambe 1954; Mitchell et al. 1965; Garcia-Bengochea
et al. 1979; Acar and Oliveri 1990; Benson and Daniel 1990),
whereas composition (e.g., index properties) usually has a
much smaller effect on hydraulic conductivity (Daniel 1987;
Benson et al. 1994b). Because compaction water content is so
influential, it masks the importance of the other variables [e.g.,
Benson and Trast (1995)].
A statistical inference regarding the relationship between KF
and Po can be made from box plots of KF obtained from test
FIG. 7. Maximum Dry Unit Weight versus Optimum Water

Content for Soils in Database
pads having similar percentage Po [Fig. 6(b)]. The central horizontal line of a box corresponds to the median, whereas the
upper and lower ends of the box correspond to the 75th and
25th percentiles. The outermost horizontal lines correspond to
5th and 95th percentiles; that is, the outermost lines delineate
the range of nearly all of the data. The circles represent outliers. The box plots provide a means to describe the distribution of KF for particular ranges of Po. The box plots show that
the distribution of KF shifts downward nearly an order of magnitude when Po increases beyond 60%. Moreover, the box
plots and the data in Tables 1 and 2 indicate that <33% of the
liners have KF 107 cm/s when Po < 60%, whereas when
Po > 80% the percentage of liners with KF 107 cm/s is
>90%.
The data shown in Fig. 6 suggest that specification writers
should consider requiring a minimum Po. Based on this analysis, the writers suggest a minimum Po of 70 80%, which is
slightly lower than the threshold recommended by Benson and
Boutwell (1992) (Po 85%). This criterion is relatively easy
to establish because the line of optimums is practically unique
for most clayey soils (Blotz et al. 1998). For example, Fig. 7
shows that the optimum water content-maximum dry unit
weight points for the sites in the database can be represented
with a simple equation nearly identical to the relationship reported by Blotz et al. (1998).
Ensuring the compaction water content is wet of the line of
optimums is not a panacea. For example, the four test pads at
Sites 48 51 were constructed with the same soil and methods
and at essentially the same water content and Po. Nevertheless,
their KF varied by more than an order of magnitude. Differences in hydration time were the apparent cause for the variation in KF (Benson et al. 1997). Thus, seemingly minor construction variables can have a profound influence on KF, even
if the gross measurements of water content and dry unit weight
indicate that compaction is wet of the line of optimums.
Percent Compaction/Wet-of-Optimum Specifications
FIG. 6. (a) Field Hydraulic Conductivity versus Percentage of

Compaction Data Wet of Line of Optimums; (b) Box Plots of
Field Hydraulic Conductivity for Various Percentages of Data
Wet of Line of Optimums
To facilitate construction of clay liners with low hydraulic

conductivity, construction specifications typically require that
the water content fall within a specified range (often 0 4%
wet of optimum, Table 1) and that the percent compaction
equal or exceed a specified minimum (typically 90 or 95%,
Table 1). The objective of this percent compaction type of
specification is to achieve compaction wet of optimum water
content. Unfortunately, this type of specification may result in

compaction dry of the line of optimums even if compaction is
wet of a particular optimum water content, because optimum
water content varies with compactive effort (Daniel 1984;
Daniel and Benson 1990; Benson and Boutwell 1992). Nevertheless, nearly all specifications have been written in this
manner (Table 1, column 12).
This pitfall is illustrated in the compaction data for Site 27
(Fig. 8). The compaction specification required that the compaction water content exceed optimum water content based on
the modified Proctor test and that the percent compaction exceed 90%. Nearly all of the field compaction data fall within
specification (94% of data), but only two data points (6%) fall
wet of the line of optimums. The field hydraulic conductivity
for this test pad was 1.8 107 cm/s, and thus the test pad
failed. However, tests conducted by Benson and Trast (1995)
on laboratory-compacted specimens prepared from the soil
from Site 27 show that the field hydraulic conductivity would
probably have been about 2 108 cm/s had the soil been
compacted wet of the line of optimums. Had these laboratory
tests been conducted as part of a preconstruction testing program for Site 27, a different compaction specification ensuring
compaction wet of the line of optimums probably would have
been used, and the outcome would probably have been different. This example suggests that a comprehensive preconstruction test program should be conducted to define combinations
of w and d likely to yield acceptably low hydraulic conductivity in the field. A similar problem is responsible for KF >
107 cm/s at 13 other sites.
Fig. 9(a) shows KF versus percent wet of optimum water
content wo. A trend of decreasing KF with increasing wo
exists. However, a broad range of KF exists even for water
contents about 3 4% above optimum, in part because compaction can be dry or wet of the line of optimums depending
on the compactive effort employed. The data indicate that KF
107 cm/s is probable when very high water contents are
used (i.e., >4% above optimum water content). This occurs
because requiring a high compaction water content effectively
ensures compaction is wet of the line of optimums.
No trend is evident between KF and percent compaction
[Fig. 9(b)]. Percent compaction is a poor indicator of KF because compaction can be wet or dry of the line of optimums
for a given percent compaction, resulting in KF that can vary
orders of magnitude (e.g., see Site 27 in Fig. 8). In fact, the
FIG. 9. Field Hydraulic Conductivity versus: (a) Compaction

Water Content Relative to Optimum Water Content; (b) Percent
Compaction
average KF for sites that used 95% of standard Proctor (ASTM

D 698) maximum dry unit weight as a criterion is lower than
the average KF for sites that used 90% of modified Proctor
(ASTM D 1557) maximum dry unit weight as a criterion. In
addition, the sites employing a standard Proctor specification
have an average Si = 90.1% and average Po = 77%, whereas
those sites employing a modified Proctor specification have an
average Si = 84.3% and average Po = 59%. That is, clay liners
constructed using a modified Proctor specification were compacted at lower water contents relative to the line of optimums
than sites constructed with a standard Proctor specification.
This discussion does not suggest that compaction is unimportant; compaction is essential because it remolds clods and
eliminates interclod voids, as well as increases shear strength.
Thus, requiring a minimum percent compaction is prudent because it ensures a minimum level of compactive effort as well
as a minimum shear strength, even though percent compaction
alone is not a good indicator of KF.
Initial Saturation Specifications
FIG. 8. Compaction Curves, Compaction Data, and Acceptable Compaction Zone for Site 27
Some engineers prefer to use a minimum initial degree of

saturation Si to define acceptable compaction conditions because Si contains both compaction water content and dry unit
weight, and contours of constant Si are approximately parallel
to the line of optimums [e.g., Boutwell and Hedges (1989),
Benson and Boutwell (1992), and Benson et al. (1994b)].
Thus, lower hydraulic conductivity should be obtained at
higher Si.
The data shown in Fig. 10 exhibit a trend of decreasing KF
with increasing Si. Because the specific gravity of solids (Gs )
was unknown for most sites, it was back-calculated from the
optimum water content and maximum dry unit weight assuming Si at optimum is 85%. Consequently, the Si shown in Fig.
FIG. 10. Field Hydraulic Conductivity versus Initial Degree of

Saturation
FIG. 11. Field Hydraulic Conductivity versus Liner Thickness

with Trend Lines from Benson and Daniel (1994)
10 are estimates. There is significant scatter in the KF-average

Si relationship because the true Gs values were not available,
and subtle changes in Gs cause larger changes in Si. Had sitespecific Gs been available, a better correlation between KF and
Si probably would have been obtained.
Fig. 11 shows KF versus liner thickness for the database

compiled here, with the trend lines from Benson and Daniel
(1994). A nearly identical graph can be made by plotting KF
versus number of lifts, because lifts are typically about 150
mm thick. Although significant scatter exists, the average KF
decreases with increasing thickness. For example, the average
KF = 8 108 cm/s for liners 0.6 m thick and 1 108
cm/s for liners 1.5 m thick. Moreover, KF is <107 cm/s for
all liners at least 1 m thick.
A stronger trend in the field data may have been found if
thinner liners had been included in the database, such as those
described by Benson and Daniel (1994). These thin liners were
not included in the database because they were constructed
with methods atypical of modern practice. In the database described here, only five clay liners were thinner than 0.6 m,
and none were thinner than 0.3 m. The lack of data for thin
liners suggests that thicker liners now predominate practice or
are required by regulation.
CONSTRUCTION FACTORS
The regression analyses conducted on the entire database
and that portion of the database with Po > 80% showed that
lift thickness was not statistically significant. Compactor mass
and number of passes were not included in the regression analysis due to lack of data (Table 2). For brevity, graphs of these
variables have not been included, but neither variable significantly influenced KF. Compaction conditions (e.g., Po and
wo) appear to be much more important than the details of
how the soil was compacted. For at least 95% of the sites,
four or more compactor passes were used. In addition, heavy
compactors (180 MN) were used for 90% of the sites (Table
2). Field hydraulic conductivity may have been significantly
related to these variables if a fewer number of passes or lighter
compactors had been used at a greater number of sites. The
data in Table 2 do show, however, that there is no apparent
advantage of using very heavy compactors (e.g., 315 MN)
relative to the moderately heavy compactors (weight 195
MN) that are commonly used.
Examples exist, however, where the number of compactor
passes was the only variable that differed substantially between two similar test pads having different KF. For example,
Sites 43 and 44 are two test pads that were constructed from
the same soil at essentially the same water content and compacted with the same compactor. The only variable was the
number of compactor passes: 16 for Site 43 and eight for Site
44. KF for Sites 43 and 44 are 7 108 and 2 107 cm/s,
respectively, illustrating that increasing the compactive effort
can reduce KF. The soil compacted with 16 passes had Po =
63%, whereas Po was 47% when eight passes were used. The
additional compactive effort used for Site 43 resulted in more
w-d points falling above the line of optimums.
Liner thickness was found to be statistically significant,
which is consistent with the conclusion Benson and Daniel
(1994) made in their study regarding an appropriate minimum
thickness for clay liners. Thicker liners typically have a greater
number of lifts, which results in a lower probability that a
continuous permeable pathway penetrates the entire thickness
of the liner (Boutwell and Rauser 1990; Benson and Daniel
1994).
SOIL COMPOSITION
Engineers usually specify minimum values for LL and/or PI
(LL typically is 25 30, PI typically is 12 15), fines (typically 50% passing the No. 200 sieve), and sometimes percentage clay (e.g., 20 25% 2-m clay) to ensure that suitable soils are used for construction of clay liners (Daniel 1990;
Benson et al. 1994b). These specifications are made based on
the belief that these variables influence KF (Benson et al.
1994b; Benson and Trast 1995). In their analysis of KL from
laboratory-compacted specimens, Benson and Trast (1995)
found that KL correlates reasonably well with these index properties when compaction is wet of the line of optimums. Boutwell and Hedges (1989) and Benson et al. (1994b) found similar results for KL of field-compacted specimens.
The second regression analysis (sites where Po 80%) was
conducted to assess whether relationships exist between KF
and index properties when compaction is wet of the line of
optimums. Results of the analysis (Table 3) show, however,
that liner thickness (or number of lifts) is the only statistically
significant variable even when Po 80%, and only PI is borderline significant. LL was found to be borderline significant
when a similar regression was conducted without PI as an
independent variable. Similar information is in the LL and PI,
because these variables are correlated.
Graphs of KF versus LL and PI are shown in Fig. 12. Slight
trends of decreasing KF with increasing LL and PI exist that
FIG. 12. Field Hydraulic Conductivity versus: (a) Liquid Limit;

(b) Plasticity Index
FIG. 13. Field Hydraulic Conductivity versus: (a) Percent
Fines; (b) Clay Content
are consistent with trends reported by Benson et al. (1994b)

for KL. One factor affecting the trends is that the clay used for
eight sites from Georgia (Sites 55 63) was nearly pure
kaolinite a low activity clay mineral. Removal of these sites
from the database increases the partial-F for LL and PI beyond
4. Thus, LL and PI may be more significant for clays comprised of a broader mixture of more active minerals.
Relationships between KF and percent fines and 2-m clay
content are shown in Fig. 13. Trends in the field data are weak,
in part because the portion of the database shown in Fig. 13
is not broad (i.e., the percentage fines is >50%, the 2-m clay
content is >20%). Nevertheless, the data do indicate that KF
107 cm/s can be achieved for a broad variety of clayey
soils that have sufficient fines and a plastic clay fraction. In
addition, the aforementioned criteria for index properties are
likely to provide suitable soils for construction of clay liners.
However, clear cutoffs for unacceptable index properties are
not evident in the data.
A small number of sites failed to achieve the hydraulic conductivity objective because the soil was not suitable. For example, the kaolinitic soil used at Sites 62 63 could not be
compacted to achieve KF 107 cm/s even with thorough
processing of the soil, moistening 2.5 4.4% above optimum,
and compaction with heavy kneading compactors. Unfortunately, Po could not be computed for these sites because only
the average w and d were reported for each lift. In this case,
KL of field-compacted specimens also indicated that the hydraulic conductivity was >107 cm/s. No laboratory testing
was conducted before the field-testing program. In retrospect,
a laboratory-testing program would have been advisable and
probably would have shown that the soil was not suitable even
though the classification data suggest the soil is adequate. Similar behavior has been observed by the writers at other sites
constructed with kaolinitic soils.
PRACTICAL LESSONS FOR ENGINEERS

Compacted clay liners are nearly always constructed with
the objective of achieving KF 107 cm/s, and all of the liners
and test pads in this database were constructed for this purpose. Despite all that has been written and learned about compacted clay liners, KF was 107 cm/s at 26% of the sites in
the database. What are the reasons for such poor success?
At a few sites the soil composition precluded achieving sufficiently low hydraulic conductivity, and in each of these cases
a test pad was constructed without the benefit of a comprehensive preconstruction laboratory-testing program.
Lesson 1: Always conduct laboratory hydraulic conductivity
tests on representative soils prior to construction [e.g., following the procedures recommended by Boutwell and Hedges
(1989) or Daniel and Benson (1990)]. This should eliminate
unsuitable soils from further consideration. In addition, preliminary screening can be conducted using index properties and
empirical equations used to predict hydraulic conductivity
[e.g., as in Boutwell and Hedges (1989) and Benson et al.
(1994b)] before laboratory testing is conducted. If the predicted hydraulic conductivities are too high, the soil can be
considered unacceptable and unnecessary laboratory testing
can be avoided. However, in no circumstance should a soil be
considered acceptable based solely on empirical predictions.
Such predictions are not an acceptable surrogate for laboratory
hydraulic conductivity testing.
The single biggest design problem is failure to recognize
that conventional compaction specifications based on a mini-
mum percent compaction and minimum water content often

lead to difficulties because that type of specification does not
ensure that compaction will be wet of the line of optimums.
Despite widespread publication of procedures that will avoid
this problem [e.g., Daniel and Benson (1990) and Daniel and
Koerner (1995)], many designers and specification writers
continue to repeat the mistake. The type of specification that
is not recommended (but is commonly used) is shown in Fig.
14(a); the more appropriate and recommended approach is
shown in Fig. 14(b).
Sites 64 and 65 provide an excellent example of the futility
of conventional specifications. Test pads were constructed of
the same material but with a minimum dry unit weight of 95%
of standard Proctor d,max (Site 64) or 91% of modified Proctor
d,max (Site 65). Both test pads were compacted several percentage points wet of their respective optimums, but neither
achieved KF 107 cm/s. Moreover, the index properties of
the clay used suggest that KF 107 cm/s should have been
achieved had compaction been wet of the line of optimums.
Site 65 is particularly interesting because it was compacted
3.7% wet of optimum to d > 91% of d,max using an astonishing 80 passes/lift of a heavy footed compactor. Despite this,
the liner failed to achieve the desired KF for one simple reason:
not a single w-d point was on or above the line of optimums.
The third lesson follows directly from lesson two and the
data shown in Fig. 6(b).
Lesson 2: Avoid the conventional percent compaction specification (minimum d and wet of a single optimum water content) and craft a specification that ensures compaction on or
above the line of optimums.
CONCLUSIONS
FIG. 14. Schematics of: (a) Percent Compaction Specification; (b) Recommended Compaction Specification
Lesson 3: The key to achieving KF 107 cm/s is to ensure

at least 70 80% of the field measured w-d points fall on or
above the line of optimums.
Although this paper has focused on hydraulic conductivity,
other important factors must be considered (e.g., bearing capacity, trafficability, internal and interface shear strengths,
compressibility, and resistance to desiccation cracking). Generally an increase in water content adversely affects these factors (Leroueil et al. 1992; Daniel and Wu 1993). An essential
trade-off in clay liner design is achieving combinations of water content and dry unit weight that result in soil sufficiently
wet to attain low KF, while maintaining sufficient strength and
acceptable compressibility and cracking potential. Engineers
must give proper consideration to these factors and avoid the
temptation to add too much water to the clay (to reduce KF)
at the expense of compromising the other critical engineering
properties of the clay liner. Suitable compaction conditions
meeting each of these criteria can be achieved by adjusting
the compaction water content and the compactive effort, as
outlined in Daniel and Wu (1993).
A database consisting of 85 full-scale compacted clay liners

and field test pads was assembled to evaluate field performance based on field hydraulic conductivity KF. Large-scale
field hydraulic conductivity tests were conducted on each liner.
Nearly all regulations for clay liners in the United States require that the hydraulic conductivity does not exceed 1 107
cm/s. The database indicates that 26% of the clay liners had
KF 1 107 cm/s despite all of the clay liners being constructed for the explicit purpose of achieving this level of low
KF. Although some may believe that achieving a hydraulic
conductivity 1 107 cm/s is routine, the database indicates otherwise, with a failure rate of roughly one in four. This
suggests that constructing clay liners to this level of impermeability requires careful planning when selecting the materials, crafting the compaction specifications, and evaluating the
methods of construction. Test pads provide an ideal means to
evaluate how these factors affect KF.
In most cases the primary cause for KF > 107 cm/s is that
compaction was dry of the line of optimums. None of the
liners compacted with Po > 90% had KF > 107 cm/s. Even
today, many construction specifications for clay liners employ
a percent compaction specification, where a range of water
content (usually wet of an optimum water content) and a minimum percent compaction are specified. Although laboratory
data have demonstrated that percent-compaction specifications
can result in compaction dry of the line of optimums, the use
of this type of specification is pervasive in practice. Analysis
of the information in the database points to the clear advantage
of examining compaction results not in terms of water content
relative to an optimum or percent compaction, but rather in
terms of the position of the field compaction data relative to
the line of optimums. When most of the field compaction data
points lie on or above the line of optimums, there is a high
probability that KF will be 107 cm/s. Engineers are strongly
encouraged to evaluate their specification writing procedures
in light of these findings and to consider supplementing the
usual requirements with a requirement that at least 70 80%
of the compaction data lie on or wet of the line of optimums.
In addition, the specification writer must ensure that the soil
has adequate strength and that the clay liner is constructable
for the range of water contents and dry unit weights that are
specified. The data also show that details, such as adequate

hydration time, can affect KF even if a carefully crafted compaction specification is employed. Indeed, specifications for
each project should be carefully designed taking into account
site-specific details.
Liner thickness (or equivalently number of lifts) is also a
statistically significant variable affecting KF, with thicker liners
having more lifts that typically have lower KF. The commonly
used range of thickness (0.6 0.9 m) seems to be adequate at
least from the perspective of obtaining sufficiently low KF. In
more critical applications (e.g., hazardous waste disposal facilities), however, thicker liners (i.e., 1 m) should be considered.
Analysis of the database also shows that relationships between KF and soil properties representative of soil composition
(particle size characteristics and Atterberg limits) are not statistically significant. The existing data show that clay liners
having KF 107 cm/s can be constructed with a broad range
of clayey soils and that values currently being used to define
acceptable ranges for the Atterberg limits and particle size
characteristics yield liners having sufficiently low hydraulic
conductivity. Regardless of the soil being used, however, a
comprehensive preconstruction laboratory hydraulic testing
program should be conducted to determine if a soil is suitable
and to define an acceptable range of compaction conditions
that yield the desired hydraulic conductivity. The laboratory
program should be followed by construction and hydraulic
conductivity testing of a test pad to demonstrate that acceptably low hydraulic conductivity can also be achieved in the
field.
ACKNOWLEDGMENTS
The information presented in this paper has been developed over a
period of many years and in cooperation with individuals who are too
numerous to mention individually. The writers sincerely thank all persons
involved in developing the data for the cases reported herein. The writers
also acknowledge the help of Owen Michaelis, who assisted in compiling
the database. Funding for Bensons effort was provided primarily by the
National Science Foundation under Grant No. CMS-9157116. Work by
Daniel on developing this database has been funded primarily by the U.S.
Environmental Protection Agency (EPA) under cooperative agreement
CR-821448. The EPAs Project Officers on this project were Robert E.
Landreth (retired) and David A. Carson. STE, Inc. funded Boutwells
effort. EPA has not subjected this paper to peer and administrative review.
APPENDIX.
REFERENCES
Acar, Y., and Oliveri, I. (1990). Pore fluid effects on the fabric and
hydraulic conductivity of laboratory-compacted clay. Transp. Res.
Rec. 1219, Transportation Research Board, Washington, D.C., 144
159.
Benson, C., and Boutwell, G. (1992). Compaction control and scaledependent hydraulic conductivity of clay liners. Proc., 15th Annu.
Madison Waste Conf., University of Wisconsin, Madison, Wis., 62
83.
Benson, C., and Daniel, D. (1990). Influence of clods on hydraulic conductivity of compacted clay. J. Geotech. Engrg., ASCE, 116(8),
1231 1248.
Benson, C., and Daniel, D. (1994). Minimum thickness of compacted
soil liners: II. Analysis and case histories. J. Geotech. Engrg., ASCE,
120(1), 153 172.
Benson, C., Gunter, J., Boutwell, G., Trautwein, S., and Berzanskis, P.
(1997). Comparison of four methods to assess hydraulic conductivity. J. Geotech. and Geoenvir. Engrg., ASCE, 123(10), 929 937.
Benson, C., Hardianto, F., and Motan, E. (1994a). Representative specimen size for hydraulic conductivity assessment of compacted soil liners. Hydraulic conductivity and waste contaminant transport in soils,
ASTM STP 1142, D. Daniel and S. Trautwein, eds., ASTM, West Conshohocken, Pa., 3 29.
Benson, C., and Othman, M. (1993). Hydraulic conductivity of compacted clay frozen and thawed in place. J. Geotech. Engrg., ASCE,
119(3), 276 294.
Benson, C., and Trast, J. (1995). Hydraulic conductivity of thirteen compacted clays. Clays and Clay Minerals, 43(6), 669 681.
Benson, C., Zhai, H., and Wang, X. (1994b). Estimating the hydraulic
conductivity of compacted clay liners. J. Geotech. Engrg., ASCE,
120(2), 366 387.
Blotz, L., Benson, C., and Boutwell, G. (1998). Estimating optimum
water content and maximum dry unit weight for compacted clays. J.
Geotech. and Geoenvir. Engrg., ASCE, 124(9), 907 912.
Boutwell, G., and Hedges, S. (1989). Evaluation of water-retention liners by multivariate statistics. Proc., 12th ICSMFE, Balkema, Rotterdam, The Netherlands, 815 818.
Boutwell, G., and Rauser, C. (1990). Clay liner construction. Proc.,
Geotech. Engrg. in Todays Envir., ASCE, Reston, Va., 1 7.
Boynton, S., and Daniel, D. (1985). Hydraulic conductivity tests on
compacted clay. J. Geotech. Engrg., 111(4), 465-478.
Chiu, T.-F., and Shackelford, C. (1994). Practical aspects of the capillary
barrier effect for landfills. Proc., 17th Annu. Madison Waste Conf.,
University of Wisconsin, Madison, Wis., 357 375.
Daniel, D. (1984). Predicting hydraulic conductivity of clay liners. J.
Geotech. Engrg., ASCE, 110(2), 285 300.
Daniel, D. (1987). Earthen liners for land disposal facilities. Proc.,
Geotech. Pract. for Waste Disposal 87, R. Woods, ed., ASCE, Reston,
Va., 21 39.
Daniel, D. (1989). In situ hydraulic conductivity tests for compacted
clays. J. Geotech. Engrg., ASCE, 115(9), 1205 1227.
Daniel, D. (1990). Summary review of construction quality control for
earthen liners. Proc., Waste Containment Sys., R. Bonaparte, ed.,
ASCE, Reston, Va., 175 189.
Daniel, D. (1994). State-of-the-art: Laboratory hydraulic conductivity
tests for saturated soils. Hydraulic conductivity and waste contaminant transport in soils, ASTM STP 1142, D. Daniel and S. Trautwein,
eds., ASTM, West Conshohocken, Pa., 30 78.
Daniel, D., and Benson, C. (1990). Water content-density criteria for
compacted soil liners. J. Geotech. Engrg., ASCE, 116(12), 1811
1830.
Daniel, D., and Koerner, R. (1995). Waste containment systems: Guidance
for construction, quality assurance, and quality control of liner and
cover systems. ASCE, Reston, Va., 354.
Daniel, D., and Wu, Y. (1993). Compacted clay liners and covers for
arid sites. J. Geotech. Engrg., ASCE, 119(2), 223 237.
Day, S., and Daniel, D. (1985). Hydraulic conductivity of two prototype
clay liners. J. Geotech. Engrg., ASCE, 111(8), 957 970.
Draper, N., and Smith, H. (1981). Applied regression analysis. Wiley,
New York.
Elsbury, B., Daniel, D., Sraders, G., and Anderson, D. (1990). Lessons
learned from compacted clay liner. J. Geotech. Engrg., ASCE,
116(11), 1641 1660.
Garcia-Bengochea, I., Lovell, C., and Altschaeffl, A. (1979). Pore distribution and permeability of silty clays. J. Geotech. Engrg. Div.,
ASCE, 105(7), 839 856.
Gordon, M., Huebner, P., and Mitchell, G. (1990). Regulation, construction and performance of clay lined landfills in Wisconsin. Proc.,
Waste Containment Sys., R. Bonaparte, ed., ASCE, Reston, Va., 14
27.
Johnson, G., Crumley, W., and Boutwell, G. (1990). Field verification
of clay liner hydraulic conductivity. Proc., Waste Containment Sys.,
R. Bonaparte, ed., ASCE, Reston, Va., 226 245.
Kleppe, J., and Olson, R. (1985). Desiccation cracking of soil barriers.
Hydraulic barriers in soil and rock, ASTM STP 874, ASTM, West
Conshohocken, Pa., 263 275.
Lambe, T. (1954). The permeability of compacted fine-grained soils.
ASTM STP 163, ASTM, West Conshohocken, Pa., 56 67.
Leroueil, S., Bouchard, R., and Bihan, J. (1992). Discussion of Water
content-density criteria for compacted soil liners, by D. Daniel and C.
Benson. J. Geotech. Engrg., ASCE, 118(2), 963 965.
Mitchell, J., Hooper, D., and Campanella, R. (1965). Permeability of
compacted clay. J. Soil Mech. and Found. Div., ASCE, 91(4), 41
65.
Othman, M., Benson, C., Chamberlain, E., and Zimmie, T. (1994). Laboratory testing to evaluate changes in hydraulic conductivity of compacted clays caused by freeze-thaw: State-of-the-art. Hydraulic conductivity and waste contaminant transport in soils, ASTM STP 1142,
D. Daniel and S. Trautwein, eds., ASTM, West Conshohocken, Pa.,
227 254.
Reades, D., Lahti, L., Quigley, R., and Bacopoulos, A. (1990). Detailed
case history of clay liner performance. Proc., Waste Containment Sys.,
R. Bonaparte, ed., ASCE, Reston, Va., 156 174.
Sai, J., and Anderson, D. (1990). Field hydraulic conductivity tests for
compacted soil liners. Geotech. Testing J., 13(3), 215 225.
Shelley, T., and Daniel, D. (1993). Effect of gravel on hydraulic con-
tests for compacted soil liners and caps. Hydraulic conductivity and
waste contaminant transport in soils, ASTM STP 1142, D. Daniel and
S. Trautwein, eds., ASTM, West Conshohocken, Pa., 184 223.
Wang, X., and Benson, C. (1995). Infiltration and field-saturated hydraulic conductivity of compacted clay. J. Geotech. Engrg., ASCE,
121(10), 713 722.
ductivity of compacted soil liners. J. Geotech. Engrg., ASCE, 119(1),

54 68.
Trast, J., and Benson, C. (1995). Estimating field hydraulic conductivity
at various effective stresses. J. Geotech. Engrg., ASCE, 121(10), 736
740.
Trautwein, S., and Boutwell, G. (1994). In-situ hydraulic conductivity

Benson Et Al-1999

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Benson Et Al-1999

Uploaded by

Copyright:

Available Formats

FIELD PERFORMANCE

COMPACTED CLAY LINERS

By Craig H. Benson,1 David E. Daniel,2 and Gordon P. Boutwell,3 Members, ASCE

Assoc. Prof., Dept. of Civ. and Envir. Engrg., Univ. of Wisconsin,

390 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

Index Properties and Compaction Characteristics

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999 / 391

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

w > wo, <3

w > wo, <3

w > wo, <3

w > wo, <3

392 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

Construction Information and Hydraulic Conductivity

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999 / 393

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

or CH under ASTM D 2487 and span the clay range of the

hole tests (Trautwein and Boutwell 1994) and/or laboratory

Field Hydraulic Conductivity

OVERVIEW OF HYDRAULIC CONDUCTIVITY DATA

Histogram of Field Hydraulic Conductivities

394 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

regulation, and failure to meet this requirement is usually

FIG. 2. Hydraulic Conductivities Measured with SDRIs versus

FIG. 4. Ratio of Field to Laboratory-Hydraulic Conductivity

reported data, the ratio KF/KL is 11.2 on average when KF >

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999 / 395

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

position as described by LL, PI, and percent fines; redundancy

Example Showing How Po Is Computed

Sites Wet of Line

5 sites in regression analysis.

396 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

A second stepwise regression was conducted on the portion

FIG. 7. Maximum Dry Unit Weight versus Optimum Water

FIG. 6. (a) Field Hydraulic Conductivity versus Percentage of

To facilitate construction of clay liners with low hydraulic

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999 / 397

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

content. Unfortunately, this type of specification may result in

FIG. 9. Field Hydraulic Conductivity versus: (a) Compaction

average KF for sites that used 95% of standard Proctor (ASTM

Some engineers prefer to use a minimum initial degree of

398 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

FIG. 10. Field Hydraulic Conductivity versus Initial Degree of

FIG. 11. Field Hydraulic Conductivity versus Liner Thickness

10 are estimates. There is significant scatter in the KF-average

Fig. 11 shows KF versus liner thickness for the database

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999 / 399

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

FIG. 12. Field Hydraulic Conductivity versus: (a) Liquid Limit;

are consistent with trends reported by Benson et al. (1994b)

PRACTICAL LESSONS FOR ENGINEERS

400 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999

J. Geotech. Geoenviron. Eng. 1999.125:390-403.

mum percent compaction and minimum water content often

Lesson 3: The key to achieving KF 107 cm/s is to ensure

A database consisting of 85 full-scale compacted clay liners

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / MAY 1999 / 401