Soil Modulus Correlations

J. Michael Duncan
1
, P.E., Dist.M.ASCE and Andrew Bursey
2
, P.E., P.G., M.ASCE

1
Distinguished ProIessor Emeritus, Department oI Civil and Environmental Engineering, Virginia Tech,
Blacksburg, VA 24061
2
Senior Tunnel Engineer, Jacobs Engineering Group, Norcross, GA, 30071

ABSTRACT: This paper Iocuses on procedures Ior estimating modulus values Ior soils that are
useable with simple elastic solutions and linear Iinite element analyses Ior stresses and
deIormations. Relationships among Young`s modulus, shear modulus, constrained modulus and
bulk modulus are deIined so that correlations Ior any oI these various measures oI soil stiIIness
can be adapted to the purpose at hand. Because the stress-strain behavior oI soil is nonlinear and
stress-dependent, soil modulus values depend on conIining pressure and load magnitude, as well
as soil type, strength, stress history, and drainage conditions. Modulus correlations Ior gravel,
sand, silt and clay are discussed, and best practices Ior their use in geotechnical engineering are
suggested, together with methods oI judging their reliability.

INTRODUCTION

Soil modulus values provide a convenient measure oI soil stiIIness, which, together with
suitable elastic solutions or Iinite element analyses, provide simple Iirst-order estimates oI
settlement and other movements in soil masses. In order that calculations based on elastic theory
provide useIul results, it is imperative that the modulus values used in such calculations are
representative oI actual conditions, including soil type and density, stress and strain magnitude,
and whether loading occurs under drained or undrained conditions.
Evaluation oI soil modulus by means oI laboratory tests is diIIicult and oIten inaccurate
because disturbance has a signiIicant eIIect on soil modulus, oIten several times as great as the
eIIect oI disturbance on soil strength. As a consequence, correlations oI soil modulus values with
results oI laboratory or in situ tests are oIten more useIul and eIIective than direct measurements
oI modulus. Correlations are also useIul Ior comparison with directly measured values oI soil
modulus, to provide a quality control 'second opinion. The objective oI this paper is to
summarize what are believed to be some oI the most accurate and useIul modulus correlations Ior
gravels, sands, silts and clays, and to discuss methods oI judging their reliability and accuracy.

LINEAR AND NONLINEAR STRESS-STRAIN BEHAVIOR

Although all soils exhibit nonlinear stress-strain behavior, modulus values are based on
linear elasticity. Nonlinear behavior is approximated by using modulus values that reIlect not
only the properties oI the soil, but also the stress and strain conditions involved in the application
to which the modulus values will be applied.
The Iollowing paragraphs discuss Iirst the elastic stress-strain relationships and modulus
values, and then the methods used to approximate nonlinear behavior as equivalent linear
behavior.
321
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


The relationships between stress and strain in the theory oI elasticity can be expressed as
Iollows (Jaeger, 1962):
x x y z
1
( )
E
Ac ! Ao vAo vAo (1)
y y z x
1
( )
E
Ac ! Ao vAo vAo (2)
z z x y
1
( )
E
Ac ! Ao vAo vAo (3)
xy xy
1
G
A ! At (4)
Where x, y, and z are three mutually perpendicular directions, Ac an increment oI normal
strain, A an increment oI shear strain, A an increment oI shear stress, A an increment oI
normal stress, E Young`s modulus, v Poisson`s ratio, and G shear modulus.
Although Equations 1 through 4 can be written Ior stresses and strains rather than increments
oI stress and strain, Ior soils it is usually more appropriate to consider increments oI stress and
strain. We almost always deal with conditions involving the eIIect oI a change in stress on a soil
mass that is already under stress due to its own weight, and sometimes other loads. Thus we are
almost always interested in the eIIect oI an increment oI stress.

TYPES OF SOIL MODULI

Young`s Modulus, Shear Modulus, Bulk Modulus, and Constrained Modulus
Table 1 shows deIinitions oI these types oI modulus:

a
a
change in axial stress
Young's modulus E
change in axial strain
Ao
Ac
(5)

Young`s modulus is deIined Ior 'uniaxial loading, where the stress changes in only one
direction, as Ior example in a triaxial compression test during application oI the deviator stress.
The deIinition oI E stems Irom Eq. 1. II the x-direction is the axial direction, and y and z are the
lateral directions, it can be seen that:

x x y z
1
when and are zero
E
Ac Ao Ao Ao (6)

Although Young`s modulus is useIul in many circumstances, the other parameters in Table 1
are suited Ior particular cases:

zx
zx
change in shear stress
Shear modulus, G
change in shear strain
t

Ior 'pure shear loading (7)

0
v
change in all-around stress
Bulk modulus, B
change in volumetric strain
o
c
(8)

where A
0
isotropic stress change, and Ac
v
volumetric strain.

FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 322
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


a
a
change in axial stress
Constrained modulus, M
change in axial strain
(9)

Ior conditions where there is no strain in the lateral direction.

Table 1. Types oI soil modulus.
(AIter Lambe & Whitman, 1969; Duncan & Wong, 1999).

Two elastic constants are required to describe the behavior oI an elastic material. Although
Equations 1 through 4 contain three parameters (E, v, and G), only two are independent, because
shear modulus can be expressed in terms oI Young`s modulus and Poisson`s ratio, as shown in
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 323
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


Table 2. Reasonable values oI Poisson`s ratio (v) can be estimated using the empirical
relationship v (1-sin|)/(2-sin|)

Ior either total or eIIective stress analyses.
Relationships among the various types oI modulus are shown in Table 2. These relationships
are useIul when data is available Ior one type oI modulus, and a diIIerent modulus is needed Ior
the purpose at hand.

Table 2. Relationships among types oI soil modulus.
(AIter Lambe & Whitman, 1969; Duncan & Wong, 1999).

Relationship
between E and
other stress-strain
parameters
E 2G (1 ) ! " v E 3B (1 2 ) ! v
(1 ) (1 2 )
E M
(1 )
"v v
!
v

Relationship
between G and
other stress-strain
parameters
E
G
2(1 )
!
"v

3B(1 2 )
G
2 (1 )
v
!
" v

(1 2 )
G M
2(1 )
v
!
v

Relationship
between B and
other stress-strain
parameters
E
B
3(1 2 )
!
v

2G (1 )
B
3 (1 2 )
" v
!
v

M (1 )
B
3 (1 )
" v
!
v

Relationship
between M and
other stress-strain
parameters
E(1 )
M
(1 )(1 2 )
v
!
" v v

2G (1 ) (1 )
M
(1 ) (1 2 )
" v v
!
" v v

(1 )
M 3B
(1 )
v
!
" v


Initial Modulus, Secant Modulus, Tangent Modulus, and Unload-Reload Modulus
Figure 1 shows a typical stress-strain curve Ior a soil sample subjected to triaxial
compression, Iollowed by an unloading and reloading cycle. Although some segments oI the
curve are well represented by linear relationships (i.e. at low strains and Ior unload-reload), a
signiIicant proportion oI the stress-strain curve is nonlinear.

Figure 1. Initial modulus, tangent modulus, secant modulus, and unload-reload modulus.
(AIter Duncan & Chang, 1970; Duncan & Wong, 1999; Duncan & Bursey, 2007).
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 324
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.



Dashed lines on Figure 1 illustrate how the nonlinear stress-strain behavior can be
approximated using various moduli. For this purpose, Iour moduli are deIined:

# E
i
, the initial tangent modulus, or slope oI the stress-strain curve at the origin.
# E
s
, the secant modulus, or slope oI a line between two points on the curve. The secant
modulus is most oIten based on a line connecting the origin to some point (like A) on the
stress-strain curve, as shown in Figure 1.
# E
t
, the tangent modulus, or slope oI the stress-strain curve at some point (like A).
# E
ur
, the unload-reload modulus, or slope oI an unloading or reloading stress-strain curve.
For most practical purposes the slopes oI unloading and reloading curves are the same.

As a result oI the curvature oI the stress-strain curve, the Iarther up the stress-strain curve is
the point A, the smaller the values oI E
s
and E
t
become.
Each oI the Iour values (E
i
, E
s
, E
t
, and E
ur
) can be deIined Ior drained or undrained loading
conditions. Drained modulus values are usually indicated using the prime symbol (e.g. E'
s

indicates the drained secant Young`s modulus), although the subscript 'd has also been used.
Undrained modulus values are usually indicated by the subscript 'u (e.g. E
ut
indicates the
undrained tangent Young`s modulus). Drained modulus values should be used Ior long-term
loading conditions in all types oI soils, and all but extremely rapid loading conditions in sands or
gravels.
Numerical subscripts have been used to indicate the percent oI strength mobilized Ior which
the modulus applies. For example, E
50
to indicate the modulus value appropriate when 50 oI the
strength is mobilized. Symbols such as E
25
, E
50
, E
75
, E
90
, etc. have been used Ior various purposes,
and have been used to indicate the values oI secant modulus or tangent modulus Ior the
corresponding points on the stress-strain curve. E
ur
is commonly used to approximate either
unloading or reloading modulus, ignoring hysteresis eIIects.
The signiIicant diIIerences among the various types oI modulus shown in Table 1, and the
various moduli shown by Figure 1 illustrate the importance oI determining which modulus is
applicable Ior the problem at hand. UnIortunately, speciIic terminology has not always been
applied to modulus parameters in geotechnical engineering, potentially leading to conIusion about
which modulus is intended. In some cases, ambiguous terms such as 'the elastic modulus have
been used. In others, the symbol Ior Young`s Modulus, 'E, has been used when the constrained,
or 'one-dimensional modulus, 'M is intended.
An important Iirst step in solving problems involving soil modulus is to determine which
modulus is required Ior the analysis, and to careIully assess which modulus is provided by a
particular correlation. ThereIore, we have included speciIic deIinitions oI the various types oI
modulus in this section.
The term 'modulus oI subgrade reaction (or the 'subgrade modulus) is oIten used in
Ioundation and pavement design to reIer to the ratio oI pressure applied to a speciIic area divided
by the resulting displacement. This parameter is deIined on the basis oI a speciIic loading area,
and has units oI pressure per unit oI displacement or Iorce per length unit cubed. This
terminology is common, but has led to conIusion among uninitiated engineers. For this reason, it
is probably better to reIer to such so-called 'moduli as 'coeIIicients oI subgrade reaction.

Variation of Modulus with Confining Pressure
As shown in the leIt side oI Figure 2, stress-strain curves Ior soils are dependent on the
conIining pressure (
3
). As the conIining pressure (
3
) increases, the strength oI the soil and the
steepness oI the stress-strain curve increases.
The right side oI Figure 2 shows that values oI tangent modulus increase with increasing
conIining pressure (o
3
), and decrease with increasing strain. Tangent modulus values decrease to
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 325
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


zero where the stress-strain curve becomes horizontal. Secant modulus values also decrease, but
not to zero.
















Figure 2. EIIect oI conIining pressure and strain on modulus values Ior soil subjected to triaxial
loading. (AIter Duncan & Wong, 1999).

Modulus Degradation with Strain
Decrease in modulus values with increasing strain is oIten called 'modulus degradation.
Shear modulus degradation is an important consideration Ior dynamic geotechnical engineering
problems, which are oIten concerned with small shear strain increments. Modulus degradation
relationships can also be useIul Ior certain static problems requiring adjustment oI Young`s
Modulus or Bulk Modulus appropriate Ior one strain level to approximate values Ior another
strain, as described in this section.
Seed & Idriss (1970) described the eIIect oI increased shear strain on dynamic shear
modulus (G). They noted that the stress-strain behavior oI soils subjected to symmetrical cyclic
loading is hysteretic, as shown in Figure 3, and that the shear modulus can be expressed as the
slope oI the line drawn between the extreme points on the hysteresis loop. Values oI G deIined in
this way are secant modulus values. By comparing the average slopes oI the hysteresis loop Ior
smaller shear strain (
1
), with the loop Ior larger shear strain, (
2
), it can be seen that larger strain
amplitude results in smaller shear modulus.
'Modulus degradation Iactors are used to relate modulus to strain magnitude. Table 3
shows shear modulus degradation Iactors developed by Seed & Idriss (1970), Vucetic & Dobry
(1991), Ishibashi (1992), Darendeli (2001), and Stokoe et al. (2004). Modulus degradation Iactors
are expressed as G/G
max
, where G shear modulus at a particular value oI shear strain (), and
G
max
shear modulus at 10
-4
. Values oI G
max
can be obtained using shear wave velocity
measurements.

G
max
v
s
2
(10)

where is the mass density oI the soil,
t
/g, in which
t

total unit weight, g the
acceleration oI gravity and v
s
shear wave velocity.

FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 326
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.



Figure 3. Hysteretic stress-strain relationships at small and large strain amplitudes.
(From Seed & Idriss, 1970).

Table 3. Modulus degradation Iactors, G/G
max
.

Soil type and reIerence
Shear strain ()
10
-4
10
-3
10
-2
10
-1
1.0
Sand
Seed & Idriss (1970)
Minimum 1.00 0.94 0.62 0.22 0.04
Maximum 1.00 1.00 0.84 0.36 0.08
Non-plastic soil with PI 0
Stokoe et al.(2004), Darendeli (2001)
o
0
25 kPa 1.00 0.94 0.68 0.48 0.03
o
0
100 kPa 1.00 0.96 0.77 0.37 0.04
o
0
400 kPa 1.00 0.98 0.83 0.27 0.06
o
0
1,600 kPa 1.00 0.99 0.89 0.19 0.10
Non-plastic soil with PI 0
Ishibashi (1992)
o
m
1 kPa 1.00 0.92 0.50 0.10 0.02
o
m
50 kPa 1.00 1.00 0.80 0.37 0.08
o
m
200 kPa 1.00 1.00 0.92 0.57 0.14
o
m
400 kPa 1.00 1.00 0.99 0.72 0.20
Fine-grained soil with PI 50
Ishibashi (1992)
o
m
1 kPa 1.00 1.00 0.94 0.65 0.17
o
m
50 kPa 1.00 1.00 0.98 0.70 0.18
o
m
200 kPa 1.00 1.00 1.00 0.73 0.19
o
m
400 kPa 1.00 1.00 1.00 0.74 0.20
Fine-grained normally consolidated
soils
(OCR 1, o
0
100 kPa)
Stokoe et al.(2004), Darendeli (2001)
PI 0 1.00 0.97 0.76 0.27 0.03
PI 5 1.00 0.97 0.79 0.30 0.04
PI 20 1.00 0.98 0.82 0.36 0.06
PI 80 1.00 0.99 0.90 0.52 0.12
Saturated Iine-grained soils with
OCR 1 to 15
Vucetic & Dobry (1991)
PI 0 1.00 0.96 0.70 0.24 0.02
PI 15 1.00 1.00 0.82 0.40 0.10
PI 30 1.00 1.00 0.90 0.52 0.17
PI 50 1.00 1.00 0.94 0.68 0.26
PI 100 1.00 1.00 0.98 0.82 0.37
PI 200 1.00 1.00 1.00 0.90 0.48
o
m
average conIining pressure, o
0
isotropic conIining pressure

FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 327
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


Considering the relationships between shear modulus and the other types oI soil modulus
shown in Table 2, it can be seen that the other modulus values will also decrease with increasing
strain. Figure 2 and Table 3 show how both Young`s modulus and shear modulus decrease with
increasing strain in the same manner.
The consequence is that shear modulus degradation relationships such as those presented in
Table 3 can be used to adjust Young`s Modulus values Irom one strain level to obtain
approximate modulus values at another. Stokoe et al. (2004) presented examples illustrating the
use oI the relationships presented in Table 2 to estimate E
max
Irom G
max
values obtained Irom
shear wave velocity measurements, and the use oI modulus degradation Iactors to obtain modulus
values applicable to solving Ioundation settlement problems.
Using shear wave velocity measurements as a means oI determining values oI G
max
has the
advantage that the measurements are made in situ, under ambient stress conditions, with
essentially no disturbance oI the soil. The disadvantage oI this method is that it is necessary to
make quite signiIicant corrections Ior modulus degradation to derive values corresponding to
larger values oI strain, thereby introducing a measure oI uncertainty.

MODULUS CORRELATIONS FOR GRAVELS, SANDS, AND SILTS

A correlation between Standard Penetration Test (SPT) blow count and secant values oI
constrained modulus developed by D`Appolonia et al. (1970) is shown in Figure 4. These values
were determined by back analysis oI the settlement oI bridge piers and Iootings. They are thus
most directly applicable to Iairly shallow depths, approximately one Iooting width beneath a
Ioundation. The correlations are valuable because the measurements on which they are based
involved signiIicant volumes oI soil in its undisturbed condition in situ.

Figure 4. Correlation between M'
s
and SPT blow count Ior sand and gravel.
(AIter D`Appolonia et al., 1970; Tan et al., 1991).
(Note: 1 tsI 0.096 MPa; 1 It 0.305 m)

Correlations between constrained modulus, cone penetration resistance q
c
, and relative
density Ior normally consolidated and overconsolidated sands based on data Irom calibration
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 328
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


chamber tests are shown in Figure 5. The advantage oI this method oI modulus determination is
that it is possible to control all oI the variables in calibration chamber tests closely and produce
specimens oI uniIorm, repeatable relative density. The disadvantage oI such correlations is that
the relative density oI soils in situ must be known to apply these correlations to Iield conditions,
and estimating in situ relative density requires use oI a separate correlation. The standard
deviations oI the trend lines in Figures 4 and 5 are shown in the Iigures, which is helpIul Ior
estimating the reliability oI the correlations.

Figure 5. Relationship between M'
dt
/q
c
and D
r
Ior sands Irom calibration chamber tests.
(AIter Kulhawy & Mayne, 1990).

Correlations between E'
s
and the corrected blow count (N
1
)
60
are shown in Table 4. The
advantages oI these correlations are that they cover a variety oI soil types, including silts, sands,
and gravels, and they are very simple. However, no data are shown, which makes it impossible to
judge the reliability oI the correlations. It seems logical only to use such correlations in
conjunction with other correlations or measurements, to provide a 'second opinion as a means oI
veriIying their validity.
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 329
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.



Table 4. Relationship between E'
s
and (N
1
)
60
Ior silts, sands, and sandy gravels.
(From Sabatini et al., 2002).











MODULUS CORRELATIONS FOR SILTS AND CLAYS

Modulus data and correlations Ior silt are not as well represented in geotechnical literature as
are data and correlations Ior other types oI soil. Table 4 includes one correlation between E'
s
and
(N
1
)
60
Ior silts. Janbu (1963, 1967, 1985) proposed the dimensionless 'modulus number m,
which he related to porosity Ior silt. The relationship between constrained modulus Ior silt and the
modulus number m is expressed as Iollows.
0.5
'
va
M' m p
t a
p
a
| |
|
|
\
(11)
where M'
t
drained tangent constrained modulus, m Janbu`s dimensionless modulus number,
p
a
atmospheric pressure 100 kPa, o'
va
average vertical eIIective stress beIore and aIter
loading, expressed in the same units as p
a
. Values oI m Ior silt are given in Table 5.

Table 5. Values oI Janbu`s dimensionless modulus number m Ior silts.
(From Duncan & Wong, 1999, aIter Janbu 1963, 1967, 1985).

Porosity (n)
Percent
Value oI m (dimensionless)
Normally loaded silt Preloaded silt
50 25 50 120 240
40 60 120 300 600
30 100 200 500 1,000

Porosity can be computed Irom dry density and speciIic gravity. For saturated soils, n can be
computed using the Iollowing relationship:
wG
s
n
(1wG )
s
(12)
where w water content and G
s
speciIic gravity. For unsaturated soils, the porosity is larger
than the value computed using Eq. (12).
Kulhawy & Mayne (1990) developed the correlation between the drained secant constrained
modulus M'
s
Ior clays, and net cone tip resistance q
T
shown in Figure 6. They suggested that the
correlation was appropriate only Ior Iirst order approximations. Similar advice would apply to all
Soil Type E'
s
(kPa)
Silts, sandy silts, slightly cohesive
mixtures
400 (N
1
)
60

Clean Iine to medium sands and
slightly silty sands
700 (N
1
)
60

Coarse sands and sands with little
gravel
1,000 (N
1
)
60

Sandy gravels 1,200 (N
1
)
60

FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 330
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


oI the correlations discussed in this paper. The standard deviation oI the correlation is shown in
the Iigure, making it possible to make an inIormed judgment regarding the degree oI accuracy
inherent in use oI the correlation. The advantage oI this correlation is that it is based on high
quality data Irom 12 sites, and includes clays oI varying sensitivity.
The net cone tip resistance q
T
q
c
(1 - a)u
2
, where q
c
measured cone tip resistance, a
net area ratio which is dependent on the design oI the cone, and u
2
the pore pressure behind the
cone tip. The horizontal and vertical axes in Figure 6 are dimensionless, with the constrained
modulus and the net cone tip resistance divided by atmospheric pressure, which, as noted above,
is approximately 100 kPa.

Figure 6. Approximate relationship between M'
s
( M
ds
) and q
T

Ior clays, based on high-quality tests Irom 12 sites.
(AIter Kulhawy & Mayne, 1990).

The small-strain shear modulus (G
max
) Ior clays can be estimated based on cone penetrometer
tip resistance using the correlation shown in Figure 7, which was developed by Mayne & Rix
(1993). The advantages oI this correlation are that both q
c
and G
max
are determined in situ, and
that a considerable amount oI data Irom a variety oI sites is included.

Figure 7. Relationship between G
max
and cone tip resistance (q
c
) Ior 31 clays.
(AIter Mayne & Rix, 1993, with permission Irom ASTM International,
100 Barr Harbor Drive, West Conshohocken, PA 19428).
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 331
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


The drained tangent one-dimensional compression modulus, M'
t
, Ior clays can be estimated
using Figure 8 and the Iollowing relationships:

va
c
2.3 '
M'
t
c
c
o
(15)
where c
cc
c
c
/(1e
0
) compression ratio (dimensionless), c
c
slope oI e-log p curve, and o'
va

average oI pressures beIore and aIter loading.


Figure 8. Relationship between c
cc
and water content Ior
normally consolidated and overconsolidated clays.
(Developed by Ignacio Arango, used with permission).

Values oI c
cc
can be estimated using the correlation shown in Figure 8. This correlation was
developed by Ignacio Arango based on data Irom projects he was involved with beginning in the
1960s. An early version oI this Iigure, with Iewer data points, was published by Lambe &
Whitman (1969). The correlation, which applies to both normally consolidated and
overconsolidated clays, has surprisingly little scatter Ior values oI water content above 50. The
standard deviation oI the average trend line is 0.033.
The drained and undrained initial tangent Young`s modulus (E
ui
, and E'
i
), Ior lean clays at
depths oI roughly 2 to 5 Ieet, compacted between 95 and 100 oI Standard Proctor, can be
estimated using the correlations in Table 6.
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 332
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


Table 6. Undrained (short term) and drained (long term) initial tangent Young`s modulus Ior CL
clays compacted to RC 95 to 100 oI Std Proctor max dry density at depths oI 0.6 to 1.5 m.
(From Duncan & Mokwa, 2001, aIter Duncan et al., 1980).

Compaction water
content
Short-term modulus
(MPa)
Long-term modulus
(MPa)
w opt. 2 E
ui
19.2 57.5 E'
i
5.7 19.2
w optimum E
ui
9.6 28.7 E'
i
5.7 19.2
w opt. 2 E
ui
4.8 14.4 E'
i
5.7 19.2

USE AND MISUSE OF SOIL MODULUS CORRELATIONS

EIIective use oI elastic solutions and linear Iinite element analyses in geotechnical
engineering requires that the calculations be done using appropriate values oI soil modulus. To
achieve useIul results, it is important that careIul consideration be given to the modulus values
used in the analyses, which requires consideration oI the reliability oI the correlation used and its
suitability Ior the purpose at hand.
Selection oI appropriate soil modulus correlations requires Iirst identiIying the type oI
modulus required Ior the problem at hand. For example, to evaluate the dynamic response oI a
Iooting on clay supporting a vibrating machine, values oI G
max
Irom Figure 7 would be
appropriate. To estimate long-term settlements oI the same Iooting, however, modulus values
calculated using data Irom Figure 8 would be more suitable.
In order to assess the applicability oI a particular correlation, it is worthwhile to consider the
conditions under which the correlation was developed, and how these conditions compare to
those at the site. Numerous site-speciIic Iactors aIIect soil modulus, and contribute to diIIerences
between the soil modulus estimated using a correlation and the modulus indicated by in situ
behavior. Factors that have been shown to aIIect soil modulus values include:

# Loading history (as illustrated in Figures 4 and 5, Ior example).
# Aging.
# Cementation.
# Residual rock structure (in the case oI residual soils).

The loading history oI clays can be assessed using OCR. However, the other site-speciIic Iactors
mentioned above can usually only be assessed in a qualitative sense, and they are not reIlected in
engineering soil classiIications. This illustrates the value that knowledge oI geological conditions
at the site can provide Ior guiding selection and use oI modulus correlations.
One simple but useIul tool Ior judging reliability oI a correlation is its standard deviation.
Using the value oI standard deviation, conIidence limits can be constructed to provide guidance
with regard to modulus values determined using the correlation. II a correlation is assumed to
have either a normal or log-normal distribution, conIidence limits vary with standard deviation as
shown in Table 7. Thus selecting a modulus value Irom the average minus one standard deviation
line in Figure 7 will reduce the possibility that the selected value could be too large Irom 50
(Ior the average line) to 16.
The method oI determining the correlated modulus values should also be considered. The
greatest reliability will be achieved when modulus values back-calculated Irom Iield
measurements are used to compute movements Ior the same or similar type oI loading. An ideal
example would be use oI modulus values Irom Figure 4 to estimate settlements oI Iootings or
piers on sand or gravel. The reliability oI the settlements calculated in this case would be
expected to be consistent with the standard deviation oI the data in Figure 4 (o 46 tsI |4.4
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 333
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


MPa|, Ior preloaded sand, or o 64 tsI |6.1 MPa|, Ior normally loaded sand or sand and gravel).
For analyses oI the dynamic response oI the same Iooting supporting a vibrating machine,
however, these modulus values require adjustment Ior the smaller strains involved in the dynamic
response, which would be possible using modulus degradation Iactors Irom Table 3. However,
this adjustment would introduce additional uncertainty.

Table 7. Relationship between standard deviation and conIidence limits.








Using more than one correlation always provides valuable additional inIormation Ior
selecting soil modulus values, and a basis Ior judging overall reliability. Thus, although no data
are shown in Tables 4, 5, or 6, they provide useIul second opinions Ior quality control.
A Iurther consideration regarding correlations is the degree oI uncertainty involved in the
value oI the parameter used as the basis Ior the correlation, which in Figure 4 is the Standard
Penetration Test blow count. It is well understood that the Standard Penetration Test is in Iact not
'standard. Harr (1984) and Kulhawy (1992) have shown that the coeIIicient oI variation oI SPT
blow count can range Irom 15 to 45. This variation should also be considered when Figure 4
is used to estimate modulus values Ior sand and gravel. Similarly, with other correlations, the
reliability oI whatever parameter is used as the basis Ior the correlation should be considered, as
well as the eIIect on the degree oI conIidence in the estimated modulus value.

CONCLUSIONS AND RECOMMENDATIONS

Soil modulus correlations are useIul Ior geotechnical engineering because they can provide a
meaningIul basis Ior comparison against directly-measured values, and in many situations yield
Iirst-order approximations oI modulus values that are suitable Ior simple elastic solutions and
linear Iinite element analyses. Obtaining useIul soil modulus values Irom correlations requires
reasoned assessment oI the basis, limitations, reliability and accuracy oI each correlation used.
Correlations can never be considered to be precise, because oI the inevitable diIIerences
between the soils and conditions on which the correlations are based and those to which they are
applied. As a result, a range oI modulus values should always be estimated, rather than a single
value. This can be achieved by using two or more correlations rather than just one, by increasing
and decreasing the estimated modulus value by a multiple oI the standard deviation in cases
where the standard deviation is known, or by using experience and judgment to estimate how
much higher or lower than the best estimate the modulus value might be Ior the particular soil and
conditions oI interest.
Even though precision can never be expected, the thought process involved in making the
estimate is worthwhile, and having some idea oI what is a reasonable value is better than having
no idea at all. In the end, estimating modulus values based on correlations is an art, not a science,
and this art should involve the use oI all available data, experience and judgment.

Number oI standard deviations
above or below the trend line
Corresponding conIidence limit
$ 1 84 conIidence limit
$ 2 98 conIidence limit
$ 3 99.9 conIidence limit
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 334
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


REFERENCES

Arango, I. (2012). Personal communication.
Darendeli, M.B. (2001). 'Development oI a new Iamily oI normalized modulus reduction and
material damping curves. Doctoral dissertation, University oI Texas, Austin, TX.
D'Appolonia, D.J., D'Appolonia, E. & Brissette, R.F. (1970). 'Settlement oI spread Iootings on
sand (Closure). J. Soil Mech. Found. Div., 96(SM2), 754-761.
Duncan, J.M., & Bursey, A. (2007). 'Soil and Rock Modulus Correlations Ior Geotechnical
Engineering, Center Ior Geotechnical Practice and Research, Virginia Tech, 71 pp.
Duncan, J.M., Byrne, P., Wong, K.S. & Mabry, P. (1980). 'Strength, Stress-Strain, and Bulk
Modulus Parameters Ior Finite Element Analysis oI Stresses and Movements in Soil
Masses. Report UCB/GT/80-01, Department oI Civil Engineering, University oI CaliIornia,
Berkeley, CA.
Duncan, J.M., & Chang, C.-Y. (1970). 'Nonlinear analysis oI stress and strain in soils. J. Soil
Mech. Found. Div., 96(5), 1629-1653.
Duncan, J.M. & Mokwa, R.L. (2001). 'Passive earth pressures: theories and tests. J. Geotech. &
Geoenv. Eng. 127(3), 248-257.
Duncan, J.M. & Wong, K.S. (1999). 'User's Manual Ior SAGE: Volume II - Soil Properties
Manual. Report oI a study sponsored by the Center Ior Geotechnical Practice and Research,
Virginia Polytechnic Institute and State University, Blacksburg, VA.
Harr, M.E. (1984). Reliability-based design in civil engineering.`` 1984 Henry M. Shaw
Lecture, Dept. oI Civil Engineering, North Carolina State University, Raleigh, N.C.
Ishibashi, I. (1992). 'EIIect oI Soil Plasticity on Cyclic Response (Discussion). J. of Geotech.
Eng, 118(5), 830-832.
Jaeger, J.C. (1962). 'Elasticity, Fracture and Flow. John Wiley & Sons, Inc., 208 pp.
Janbu, N. (1963). 'Soil compressibility as determined by oedometer and triaxial tests. Proc.,
European Conference of Soil Mechanics and Foundation Eng., Wiesbaden, Vol. 1, 19-25.
Janbu, N. (1967). 'Settlement Calculations Based on Tangent Modulus Concept. Bulletin No. 2,
Soil Mechanics and Foundation Engineering Series, The Technical University oI Norway,
Trondheim, Norway.
Janbu, N. (1985). 'Soil models in oIIshore engineering. Geotechnique, 25th Rankine Lecture,
35(3).
Kulhawy, F.H., & Mayne, P.W. (1990). 'Manual on Estimating Soil Properties Ior Foundation
Design. Final Report for Profect 1493-6, EPRI EL-6800, Electric Power Research Institute,
Palo Alto, CaliIornia, Prepared by Cornell University, Geotechnical Engineering Group,
Ithaca, NY.
Kulhawy, F.H. (1992). On the evaluation oI soil properties.`` GSP No. 31, Stabilitv and
Performance of Slopes and Embankments-II, A 25-Year Perspective , R.B. Seed and R.W.
Boulanger, eds.,95115.
Lambe, T.W. & Whitman, R.V. (1969). Soil Mechanics, John Wiley and Sons, Inc., New York.
Mayne, P.W. & Rix, G.J. (1993). 'G
max
and q
c
relationships Ior clays. ASTM Geotech. Test. J.,
16(1), 54-60.
Sabatini, P. J., Bachus, R.C., Mayne, P.W., Schneider, J.A. & Zettler, T.E. (2002). 'Geotechnical
Engineering Circular Number 5: Evaluation oI Soil and Rock Properties. Report FHWA-IF-
02-034, GeoSyntec Consultants report sponsored by U.S. DOT, OIIice oI Bridge
Technology, Federal Highway Administration, Washington, D.C.
Seed, H.B. & Idriss, I.M. (1970). 'Soil Moduli and Damping Factors Ior Dynamic Response
Analyses.` Report EERC-70-10, Earthquake Engineering Research Center, University oI
CaliIornia, Berkeley, CA.
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 335
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.


Stokoe, K.H., II, Joh, S.H. & Woods, R.D. (2004), 'Some Contributions oI In Situ Geophysical
Measurements to Solving Geotechnical Engineering Problems. 2
nd
International
Conference on Site Characteri:ation (ISC-2), Porto, Portugal, September.
Tan, C.K., Duncan, J.M., Rojiani, K.B. & Barker, R.M. (1991). 'Engineering Manual Ior Shallow
Foundations. National Cooperative Highwav Research Program (NCHRP Profect 24-4).
Sponsored by American Association oI State Highway and Transportation OIIicials and
Federal Highway Administration, Washington, DC..
Vucetic, M. & Dobry, R. (1991). 'EIIect oI soil plasticity on cyclic response. J. Geotech. Eng.,
117(1), 89-107.
FOUNDATION ENGINEERING IN THE FACE OF UNCERTAINTY 336
Foundation Engineering in the Face of Uncertainty
D
o
w
n
l
o
a
d
e
d

f
r
o
m

a
s
c
e
l
i
b
r
a
r
y
.
o
r
g

b
y

I
N
D
I
A
N

I
N
S
T

O
F

T
E
C
H
N
O
L
O
G
Y
-

M
u
m
b
a
i

(
I
I
T
M
)

o
n

0
9
/
0
3
/
1
3
.

C
o
p
y
r
i
g
h
t

A
S
C
E
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d
.