EM 1110-1-2908

30 Nov 94

Chapter 6 (S term in Figure 6-1) greater than four to five times the
Bearing Capacity width (B term in Figure 6-1) of the foundation. As a
rule, joints are so widely spaced that joint orientation and
condition are of little importance. Two types of failure
modes are possible depending on rock type. The two
6-1. Scope modes are local shear failure and general wedge failure
associated with brittle and ductile rock, respectively.
This chapter provides guidance for the determination of
the ultimate and allowable bearing stress values for foun- a. Brittle rock. A typical local shear failure is initi-
dations on rock. The chapter is subdivided into four ated at the edge of the foundation as localized crushing
sections with the following general topic areas: modes (particularly at edges of rigid foundations) and develops
and examples of bearing capacity failures; methods for into patterns of wedges and slip surfaces. The slip sur-
computing bearing capacity; allowable bearing capacity; faces do not reach the ground surface, however, ending
and treatment methods for improving bearing capacity. somewhere in the rock mass. Localized shear failures are
generally associated with brittle rock that exhibit signifi-
6-2. Applicability cant post-peak strength loss (Figure 6-1a).
a. Modes of failure, methods for estimating the ulti- b. Ductile rock. General shear failures are also initi-
mate and allowable bearing capacity, and treatments for ated at the foundation edge, but the slip surfaces develop
improving bearing capacity are applicable to structures into well defined wedges which extend to the ground
founded directly on rock or shallow foundations on rock surface. General shear failures are typically associated
with depths of embedments less than four times the foun- with ductile rocks which demonstrate post-peak strength
dation width. Deep foundations such as piles, piers, and yield (Figure 6-1b).
caissons are not addressed.
6-5. Jointed Rock Mass
b. As a rule, the final foundation design is controlled
by considerations such as deformation/settlement, sliding Bearing capacity failures in jointed rock masses are
stability or overturning rather than by bearing capacity. dependent on discontinuity spacing, orientation, and
Nevertheless, the exceptions to the rule, as well as pru- condition.
dent design, require that the bearing capacity be
evaluated. a. Steeply dipping and closely spaced joints. Two
types of bearing capacity failure modes are possible for
Section I structures founded on rock masses in which the predom-
Failure Modes inant discontinuities are steeply dipping and closely
spaced as illustrated in Figure 6-1c and 6-1d. Discon-
6-3. General tinuities that are open (Figure 6-1c) offer little lateral
restraint. Hence, failure is initiated by the compressive
Bearing capacity failures of structures founded on rock failure of individual rock columns. Tightly closed dis-
masses are dependent upon joint spacing with respect to continuities (Figure 6-1d) on the other hand, provide
foundation width, joint orientation, joint condition (open lateral restraint. In such cases, general shear is the likely
or closed), and rock type. Figure 6-1 illustrates typical mode of failure.
failure modes according to rock mass conditions as modi-
fied from suggested modes by Sowers (1979) and b. Steeply dipping and widely spaced joints. Bearing
Kulhawy and Goodman (1980). Prototype failure modes capacity failures for rock masses with steeply dipping
may actually consist of a combination of modes. For joints and with joint spacing greater than the width of the
convenience of discussion, failure modes will be described foundation (Figure 6-1e) are likely to be initiated by split-
according to four general rock mass conditions: intact, ting that eventually progresses to the general shear mode.
jointed, layered, and fractured.
c. Dipping joints. The failure mode for a rock mass
6-4. Intact Rock Mass with joints dipping between 20 to 70 degrees with respect
to the foundation plane is likely to be general shear
For the purpose of bearing capacity failures, intact rock (Figure 6-1f). Furthermore, since the discontinuity
refers to a rock mass with typical discontinuity spacing

6-1
EM 1110-1-2908
30 Nov 94

Figure 6-1. Typical bearing capacity failure modes associated with various rock
mass conditions

6-2
 EM 1110-1-2908
30 Nov 94

represents major planes of weakness, a favorably oriented Section II

discontinuity is likely to define at least one surface of the Methods for Computing Bearing Capacity
potential shear wedge.
6-9. General
6-6. Layered Rock Mass
There are a number of techniques available for estimating
Failure modes of multilayered rock masses, with each the bearing capacity of rock foundations. These tech-
layer characterized by different material properties, are niques include analytical methods, traditional bearing
complicated. Failure modes for two special cases, how- capacity equations, and field load tests. Of the various
ever, have been identified (Sowers 1979). In both cases methods, field load tests are the least commonly used for
the founding layer consists of a rigid rock underlain by a two reasons. First, as discussed in Chapter 4, field load
soft highly deformable layer, with bedding planes dipping tests, such as the plate bearing test, are expensive. Sec-
at less than 20 degrees with respect to the foundation ond, although the test provides information as to the load
plane. In the first case (Figure 6-1g), a thick rigid layer that will cause failure, there still remains the question of
overlies the soft layer, while in the second case (Fig- scale effects.
ure 6-1h) the rigid layer is thin. In both cases, failure is
initiated by tensile failure. However, in the first case, 6-10. Definitions
tensile failure is caused by flexure of the rigid thick layer,
while in the second case, tensile failure is caused by Two terms used in the following discussions require defi-
punching through the thin rigid upper layer. The limiting nition. They are the ultimate bearing capacity and allow-
thickness of the rigid layer in both cases is controlled by able bearing value. Definition of the terms are according
the material properties of each layer. to the American Society for Testing and Materials.

6-7. Highly Fractured Rock Masses a. Ultimate bearing capacity. The ultimate bearing
capacity is defined as the average load per unit area
A highly fractured rock mass is one that contains two or required to produce failure by rupture of a supporting soil
more discontinuity sets with typical joint spacings that are or rock mass.
small with respect to the foundation width (Figure 6-1i).
Highly fractured rock behaves in a manner similar to b. Allowable bearing capacity value. The allowable
dense cohesionless sands and gravels. As such, the mode bearing capacity value is defined as the maximum pres-
of failure is likely to be general shear. sure that can be permitted on a foundation soil (rock
mass), giving consideration to all pertinent factors, with
6-8. Secondary Causes of Failure adequate safety against rupture of the soil mass (rock
mass) or movement of the foundation of such magnitude
In addition to the failure of the foundation rock, aggres- that the structure is impaired. Allowable bearing values
sive reactions within the rock mineralogy or with ground will be discussed in Section III of this chapter.
water or surface water chemistry can lead to bearing
capacity failure. Examples include: loss of strength with 6-11. Analytical Methods
time typical of some clay shales; reduction of load bear-
ing cross-section caused by chemical reaction between the The ultimate bearing capacity may be implicitly estimated
foundation element and the ground water or surface water; from a number of analytical methods. The more con-
solution-susceptible rock materials; and additional stresses venient of these methods include the finite element and
imposed by swelling minerals. Potential secondary causes limit equilibrium methods.
should be identified during the site investigation phase of
the project. Once the potential causes have been identi- a. Finite element method. The finite element
fied and addressed, their effects can be minimized. method is particularly suited to analyze foundations with

6-3
EM 1110-1-2908
30 Nov 94

unusual shapes and/or unusual loading conditions as well qult cNc 0.5 γBNγ γDNq (6-12)
as in situations where the foundation rock is highly vari-
able. For example, the potential failure modes for the
layered foundation rock cases illustrated in Figures 6-1g where
and 6-1h will require consideration of the interactions
between the soft and rigid rock layers as well as between qult = the ultimate bearing capacity
the rigid rock layer and the foundation. The primary
disadvantage of the finite element method is that the γ = effective unit weight (i.e. submerged unit wt.
method does not provide a direct solution for the ultimate if below water table) of the rock mass
bearing capacity. Such solutions require an analyses of
the resulting stress distributions with respect to a suitable B = width of foundation
failure criterion. In addition to the method’s ability to
address complex conditions, the primary advantage is that D = depth of foundation below ground surface
the method provides direct solutions for deformation/
settlement. c = the cohesion intercepts for the rock mass

b. Limit equilibrium. The limit equilibrium method is The terms Nc, Nγ, and Nq are bearing capacity factors
applicable to bearing capacity failures defined by general given by the following equations.
wedge type shear, such as illustrated in Figures 6-1b,
6-1d, 6-1f, and 6-1i. The limit equilibrium method, as Nc 2 Nφ1/2 (Nφ 1) (6-2a)
applied to sliding stability, is discussed in Chapter 7.
Although the principals are the same as in sliding stability
solutions, the general form of the equations presented in 2 (6-2b)
Nγ Nφ1/2 Nφ 1
Chapter 7 needs to be cast in a form compatible with
bearing capacity problems. The ultimate bearing capacity
corresponds to the foundation loading condition necessary Nq Nφ
2 (6-2c)
to cause an impending state of failure (i.e. the loading
case where the factor of safety is unity).
Nφ tan2 (45 φ / 2) (6-2d)
6-12. Bearing Capacity Equations

A number of bearing capacity equations are reported in where

the literature which provide explicit solutions for the
ultimate bearing capacity. As a rule, the equations repre- φ = angle of internal friction for the rock mass
sent either empirical or semi-empirical approximations of
the ultimate bearing capacity and are dependent on the Equation 6-1 is applicable to failure modes in which both
mode of potential failure as well as, to some extent, mate- cohesion and frictional shear strength parameters are
rial properties. In this respect, selection of an appropriate developed. As such, Equation 6-1 is applicable to failure
equation must anticipate likely modes of potential failure. modes illustrated in Figures 6-1b and 6-1d.
The equations recommended in the following discussions
are presented according to potential modes of failure. b. General shear failure without cohesion. In cases
The appropriate equation number for each mode of failure where the shear failure is likely to develop along planes
is given in Figure 6-1. of discontinuity or through highly fractured rock masses
such as illustrated in Figures 6-1f and 6-1i, cohesion
a. General shear failure. The ultimate bearing cannot be relied upon to provide resistance to failure. In
capacity for the general shear mode of failure can be such cases the ultimate bearing capacity can be estimated
estimated from the traditional Buisman-Terzaghi (Terzaghi from the following equation:
1943) bearing capacity expression as defined by Equa-
tion 6-1. Equation 6-1 is valid for long continuous foun- qult 0.5 γBNγ γDNq (6-3)
dations with length to width ratios in excess of ten.

6-4
 EM 1110-1-2908
30 Nov 94

All terms are as previously defined. f. Splitting failure. For widely spaced and vertically
oriented discontinuities, failure generally initiates by
c. Local shear failure. Local shear failure represents splitting beneath the foundation as illustrated in Fig-
a special case where failure surfaces start to develop but ure 6-1e. In such cases Bishnoi (1968) suggested the fol-
do not propagate to the surface as illustrated in Fig- lowing solutions for the ultimate bearing capacity:
ure 6-1a. In this respect, the depth of embedment contrib-
utes little to the total bearing capacity stability. An For circular foundations
expression for the ultimate bearing capacity applicable to
localized shear failure can be written as: qult JcNcr (6-6a)

qult cNc 0.5γBNγ (6-4)

For square foundations

All terms are as previously defined. q 0.85 JcNcr (6-6b)

d. Correction factors. Equations 6-1, 6-3, and 6-4

are applicable to long continuous foundations with length For continuous strip foundations for L/B ≤ 32
to width ratios (L/B) greater than ten. Table 6-1 provides
correction factors for circular and square foundations, as qult JcNcr / ( 2.2 0.18 L / B ) (6-6c)
well as rectangular foundations with L/B ratios less than
ten. The ultimate bearing capacity is estimated from the
appropriate equation by multiplying the correction factor where
by the value of the corresponding bearing capacity factor.
J = correction factor dependent upon thickness of the
foundation rock and width of foundation.
Table 6-1
Correction factors (after Sowers 1979)
L = length of the foundation
Foundation Cc Cγ
Shape Nc Correction Nγ Correction
The bearing capacity factor Ncr is given by:
Circular 1.2 0.70
Square 1.25 0.85 2Nφ
2  
(cot φ) (S/B) 1
1
Ncr
Rectangular 1 Nφ  Nφ  (6-6d)
L/B = 2 1.12 0.90
L/B = 5 1.05 0.95 Nφ (cot φ) 2Nφ1/2
L/B = 10 1.00 1.00

All other terms are as previously defined. Graphical solu-

Correction factors for rectangular foundations with L/B tions for the correction factor (J) and the bearing capacity
ratios other than 2 or 5 can be estimated by linear factor (Ncr) are provided in Figures 6-2 and 6-3,
interpolation. respectively.

e. Compressive failure. Figure 6-1c illustrates a case g. Input parameters. The bearing capacity equations
characterized by poorly constrained columns of intact discussed above were developed from considerations of
rock. The failure mode in this case is similar to uncon- the Mohr-Columb failure criteria. In this respect, material
fined compression failure. The ultimate bearing capacity property input parameters are limited to two parameters;
may be estimated from Equation 6-5. the cohesion intercept (c) and the angle of internal friction
(φ). Guidance for selecting design shear strength parame-
ters is provided in Chapter 4. However, since rock
qqult 2 c tan (45 φ / 2) (6-5)
masses generally provide generous margins of safety
against bearing capacity failure, it is recommended that
All parameters are as previously defined.

6-5
EM 1110-1-2908
30 Nov 94

initial values of c and φ selected for assessing bearing

capacity be based on lower bound estimates. While inex-
pensive techniques are available on which to base lower
bound estimates of the friction angle, no inexpensive
techniques are available for estimating lower bound cohe-
sion values applicable to rock masses. Therefore, for
computing the ultimate bearing capacity of a rock mass,
the lower bound value of cohesion may be estimated from
the following equation.

qu (s)
c (6-7a)
 φ
2 tan 45 
 2

where

qu = unconfined compressive strength of the intact

rock from laboratory tests.
(RMR 100) (6-7b)
Figure 6-2. Correction factor for discontinuity spacing s exp
9
with depth (after Bishnoi 1968)

All other parameters are as previously defined.

6-13. Eccentric Load on a Horizontal Foundation

Eccentric loads acting on foundations effectively reduce

the bearing capacity. Figure 6-4a illustrates a typical
structure subjected to an eccentric load. In order to pre-
vent loss of rock/structure contact at the minimum stress
edge of the foundation (Figure 6-4a), the structure must
be designed so that the resultant of all forces acting on
the foundations passes through the center one-third of the
foundation. As indicated in Figure 6-4a, the stress distri-
bution can be approximated by linear relationship. Equa-
tions 6-8a and 6-8b define the approximate maximum and
minimum stress, respectively.

Q  6e  (6-8a)
q(max) 1 
B  B

Q  6e  (6-8b)
q(min) 1 
B  B

Figure 6-3. Bearing capacity factor for discontinuity

spacing (after Bishnoi 1968)

6-6
 EM 1110-1-2908
30 Nov 94

B = the foundation width

e = distance from the center of the foundation to

the vertical force component Q

The ultimate bearing capacity of the foundation can be

approximated by assuming that the vertical force com-
ponent Q is uniformly distributed across a reduced
effective foundation width as indicated in Figure 6-4b.
The effective width is defined by the following equation.

B′ B 2e (6-10)

The effective width (B′) is used in the appropriate bearing

capacity equation to calculate the ultimate bearing
capacity.

6-14. Special Design Cases

The bearing capacity equations discussed above are appli-

cable to uniformly loaded foundations situated on planar
surfaces. Frequently, designs suited to the particular
requirements of a project require special considerations.
Special design cases for which solutions of the ultimate
bearing capacity are readily available are summarized in
Figure 6-5. As indicated in Figure 6-5, these special
cases include inclined loads, inclined foundations, and
foundations along or near slopes. Guidance for these
special cases is provided in EM 1110-2-2502 and the
NAVDOCKS DM-7. Ultimate bearing capacity solutions
for special design cases should be in keeping with the
modes of failure summarized in Figure 6-1.

Section III
Allowable Bearing Capacity Value

6-15. General
Figure 6-4. Typical eccentrically loaded structure
foundation
The allowable bearing capacity value is defined in para-
graph 6-10b. In essence, the allowable bearing capacity is
where the maximum limit of bearing stress that is allowed to be
applied to the foundation rock. This limiting value is
q(max) = maximum stress intended to provide a sufficient margin of safety with
respect to bearing failures and deformation/settlement.
q(min) = minimum stress Nevertheless, a prudent design dictates that, once the
allowable bearing capacity value has been determined, a
Q = vertical force component of the resultant of separate calculation be performed in order to verify that
all forces acting on the structure

6-7
EM 1110-1-2908
30 Nov 94

to the unique problems of hydraulic structures. Finally,

allowable values may be obtained from empirical correla-
tions. As a rule, empirical correlations are not site spe-
cific and hence should be used only for preliminary
design and/or site evaluation purposes. Regardless of the
approach used, the allowable value selected for final
design must not exceed the value obtained from the factor
of safety considerations discussed in paragraph 6-16a.

a. Factor of safety. The allowable bearing capacity

value, qa, based on the strength of the rock mass is
defined as the ultimate bearing capacity, qult, divided by a
factor of safety (FS):

qa qult / FS (6-11)

The average stress acting on the foundation material must

be equal to or less than the allowable bearing capacity
according to the following equation.

Q / BL ≤ qa (6-12)

For eccentrically loaded foundations the B′ value (i.e.

Equation 6-10) is substituted for the B term in Equa-
tion 6-12. The factor of safety considers the variability of
the structural loads applied to the rock mass, the relia-
bility with which foundation conditions have been deter-
mined, and the variability of the potential failure mode.
For bearing capacity problems of a rock mass, the latter
two considerations are the controlling factors. For most
structural foundations, the minimum acceptable factor of
safety is 3 with a structural load comprised of the full
Figure 6-5. Special foundation design cases dead load plus the full live load.

b. Building codes. Allowable bearing capacity

the allowable differential deformation/settlement is not values that consider both strength and deformation/
exceeded. settlement are prescribed in local and national building
codes. Local codes are likely to include experience and
6-16. Determination geology within their jurisdiction while national codes are
more generic. For example, a local code will likely spec-
There are at least three approaches for determining allow- ify a particular rock formation such as “well-cemented
able bearing capacity values. First, the allowable value Dakota sandstone” while a national code may use general
may be determined by applying a suitable factor of safety terminology such as “sedimentary rock in sound condi-
to the calculated ultimate bearing capacity. The selection tion.” As a rule, allowable values recommended by the
of final allowable bearing values used in design of building codes are conservative.
hydraulic structures must be based on the factor of safety
approach in which all site specific conditions and unique c. Empirical correlations. Peck, Hanson, and
problems of such structures are considered. Second, Thornburn (1974) suggested an empirical correlation
allowable values may be obtained from various building between the allowable bearing capacity stress and the
codes. However, building codes, in general, apply only to RQD, as shown in Figure 6-6. The correlation is intended
residential or commercial buildings and are not applicable

6-8
 EM 1110-1-2908
30 Nov 94

Section IV
Treatment Methods

6-18. General

Treatment methods for satisfying bearing capacity require-

ments are essentially the same as those for satisfying
deformation/settlement requirements discussed in Chap-
ter 5. In addition to the previously discussed methods, an
examination of the general ultimate bearing capacity equa-
tion (i.e. Equation 6-1) indicates the importance of two
parameters not directly related to deformability. These
two parameters are the effective unit weight of the foun-
dation rock and the depth of the foundation below the
ground surface.

6-19. Effective Unit Weight

For foundations below the water table the effective unit

weight is the unit weight of the foundation rock minus the
unit weight of water (i.e. submerged unit weight of the
rock). Hence, foundations located above the water table
will develop significantly more resistance to potential
bearing capacity failures than foundations below the water
Figure 6-6. Allowable contact pressure on jointed rock table.

for a rock mass with discontinuities that “are tight or are 6-20. Foundation Depth
not open wider than a fraction of an inch.”
Foundations constructed at greater depths may increase
6-17. Structural Limitations the ultimate bearing capacity of the foundation. The
improved capacity is due to a greater passive resisting
The maximum load that can be applied to a rock foun- force and a general increase in rock mass strength with
dation is limited by either the rock’s ability to sustain the depth. The increased lithostatic pressure closes discon-
force without failure or excessive settlement, or the ability tinuities, and the rock mass is less susceptible to surficial
of the substructure to sustain the load without failure or weathering. Occasionally, deeper burial may not be
excessive deformation. In some cases the structural advantageous. A region with layers of differing rock
design of the foundation element will dictate the mini- types may contain weaker rock at depth. In such an
mum element size, and, consequently, the maximum con- instance, a strong rock might overlie a layer such as mud-
tact stress on the rock. For typical concrete strengths in stone, or, if in a volcanic geology, it might be underlain
use today, the strength of the concrete member is signifi- by a tuff or ash layer. In these instances, deeper burial
cantly less than the bearing capacity of many rock may even decrease the bearing capacity. The geologic
masses. investigation will determine this possibility.

6-9