Chapter 4: SUBSURFACE INVESTIGATION and • Reviewing soil maps developed for

CHARACTERIZATION agricultural purposes

• Reviewing old and new aerial photographs
Site investigation includes methods of defining the and topographic maps (may reveal previous
development or grading at the site)
soil profile and other relevant data and recovering soil • Reviewing water well logs (helps establish
samples. historic groundwater levels)
• Locating underground improvements, such
Laboratory testing includes testing the soil samples as utility lines, both onsite and immediately
in order to determine relevant engineering properties. offsite, along with locating foundations of
adjacent structures, especially those that
In situ testing includes methods of testing the soils might be impacted by the proposed
in- place, thus avoiding the difficulties associated with construction.
recovering samples.
Field Reconnaissance
SITE INVESTIGATION
Along with the background literature search, the
The objectives of the site investigation phase foundation engineer should visit the site and perform
include: a field reconnaissance. Often such visits will reveal
• Determining the locations and thicknesses obvious concerns that may not be evident from the
of the soil strata literature search or the logs of the exploratory borings.
• Determining the location of the
groundwater table as well as any other • Is there any evidence of previous
groundwaterrelated development on the site?
characteristics • Is there any evidence of previous grading on
• Recovering soil samples the site?
• Defining special problems and concerns • Is there any evidence of landslides or other
stability problems?
Background Literature Search •Are nearby structures performing
satisfactorily?
Before conducting any new exploration at a project • What are the surface drainage conditions?
site, gather whatever information that is already • What types of soil and/or rock are exposed
available, both on the proposed structure and the at the ground surface?
subsurface conditions at the site. • Will access problems limit the types of
subsurface exploration techniques that can be
• Its location and dimensions used?
• The type of construction, column loads, • Might the proposed construction affect
column spacing, and allowable settlements existing improvements? (E.g., a fragile old
• Its intended use building adjacent to the site might be
• The finish floor elevation damaged by vibrations from pile driving.)
• The number and depth of any basements • Do any offsite conditions affect the
• The depth and extent of any proposed proposed development? (E.g., potential
grading flooding, mudflows, rockfalls, etc.)
• Local building code requirements
Exploratory Borings
The literature search also should include an effort to
obtain at least a preliminary idea of the subsurface most common method of exploring the
conditions. subsurface conditions is to drill a series of
vertical holes in the ground.
• Determining the geologic history of the site, typically, 75 to 600 mm (3–24 in) in diameter
including assessments of anticipated and 3 to 30 m (10–100 ft) deep.
rock and soil types, the proximity of faults, It can be drilled with hand augers or with
and other geologic features portable power equipment.
• Gathering copies of boring logs and it is possible to drill an open hole using a
laboratory test results from previous flight auger or a bucket auger
investigations on this or other nearby sites
Caving if fill is present, the borings must extend
-is likely to be encountered in clean sands, through it and into the natural ground below.
especially below the groundwater table. if soft soils are present, the borings should
-the sides of the boring falling in extend through them and into firmer soils
below.
Squeezing
For heavy structures, at least some of the
-is likely in soft saturated clays. borings should be carried down to bedrock, if
-the soil moving inward, reducing the possible, but certainly well below the depth
diameter of the boring of any proposed pile foundations.

On large projects, the drilling program might be

*One method of dealing with caving or squeezing soils
divided into two phases:
is to use casing. 1. Preliminary phase
-This method involves temporarily lining ➢ to determine the general soil profile.
some or all of the boring with a steel pipe. 2. Final phase
➢ based on the results of the
Hollow- stem auger preliminary borings.
-The driller screws each of these augers into
the ground and obtains soil samples by *The conditions encountered in an exploratory boring
lowering sampling tools through a hollow are normally presented in the form of a boring log.
core.
Soil Sampling
Rotary wash boring
-These borings are filled with a bentonite The two categories of samples are:
slurry (a combination of bentonite clay and 1. Disturbed sample (sometimes called a bulk
water) or a polymer slurry to provide sample)
hydrostatic pressure on the sides of the boring ➢ is one in which there is no attempt to
and thus prevent caving. retain the in- place structure of the
soil.
*Engineers often use coring, which recovers intact ➢ suitable for many purposes, such as
cylindrical specimens of the rock. classification and Proctor
compaction tests.
In terms of number of borings: ➢ The driller might obtain such a
sample by removing the cuttings off
For International Building Code (ICC, 2012) the bottom of a flight auger and
➢ minimum of two borings for built- over areas placing them in a bag.
up to 465 m2 (5,000 ft2), and at least one 2. Undisturbed sample
additional boring for each additional 233 ➢ is one in which the soil is recovered
m2(2,500 ft2) of built- over areas up to completely intact and its in- place
1,860 m2 (20,000 ft2). structure and stresses are not
➢ For built- over areas larger than 1,860 m2 modified in any way.
(20,000 ft2), ICC (2012) suggests at least ➢ desirable for laboratory tests that
eight borings plus at least one additional depend on the structure of the soil,
boring for each additional 465 m2 (5,000 ft2). such as consolidation tests and shear
strength tests.
For AASTHO [10.4.2] (for bridge structures)
➢ minimum of one boring per substructure Groundwater Monitoring
which is less than 30 m (100 ft) wide.
➢ minimum of two borings for each
substructure which is wider than 30 m (100 - The position and movements of the
ft). groundwater table are very important factors
in foundation design.
*Borings generally should extend at least to a depth - This is often done by installing an
such that the change in vertical effective stress due to observation well in the completed boring to
the new construction is less than 10 percent of the monitor groundwater conditions.
initial vertical effective stress.
 - Once the groundwater level has stabilized, compression test, and the triaxial
we can locate it by lowering a probe into the compression test.
observation well.
- We usually compare observation well data • Compaction tests —Sometimes it is
with historic groundwater records, or at least necessary to place compacted fills at a site
consider the season of the year and the recent and place the foundations on these fills. In
precipitation patterns to determine the design such cases, we perform Proctor compaction
groundwater level. tests to assess the compaction characteristics
of the soil.
Exploratory Trenches • Corrosivity tests —When corrosion or
sulfate attack is a concern need to perform
- (also known as test pits) special tests such as resistivity tests and
- provide more information than a boring of sulfate content tests.
comparable depth (because more of the soil is
exposed) and are often less expensive.
• Disturbed samples can easily be Consolidation (Oedometer) Tests
recovered with a shovel.
• Undisturbed samples can be obtained - To predict the magnitude of consolidation
using hand- held sampling equipment. settlement, 𝛿 c, in a soil, we need to know its
stress- strain properties.
Two special precautions are in order when using - are usually performed on high quality
exploratory trenches: undisturbed samples.
- obtain these samples in soft to medium
1. Must be adequately shored before clays, but quite difficult in clean sands.
anyone enters them. - these tests are usually less critical because
2. Must be properly backfilled to avoid well- compacted fills have a low
creating an artificial soft zone that might compressibility.
affect future construction.
Test Procedure and Results
LABORATORY TESTING
- The test procedure consists of applying a
Soil samples obtained from the field are normally series of normal loads to the sample, allowing
brought to a soil mechanics laboratory for further it to consolidate completely under each load,
classification and testing, sometimes called ex- situ and measuring the corresponding vertical
testing. deformations.
- They can be plotted in the conventional way
• Classification, weight- volume, and index as stress versus strain with linear scales for
tests —Several routine tests are usually the two axes, giving a conventional stress-
performed on many of the samples. These strain curve.
include:
- Moisture content We can obtain the 1-D constrained tangent modulus
- Unit weight (density) M as the slope of the curve, which may vary with
- Atterberg limits (plastic limit and stress or strain:
liquid limit)
𝑑𝜎𝑧′
-Particle size distribution (sieve and M= (4.1)
𝑑𝜀𝑧
hydrometer analyses)

• Consolidation tests —Foundation designs where:

also require an assessment of soil
𝑑𝜎𝑧′ = vertical effective stress
compressibility, which provides the
𝜀𝑧 = vertical strain
necessary data for settlement analyses.

• Shear strength tests —Foundation designs

require an assessment of shear strength.
Common shear strength tests include the
direct shear test, the unconfined
The slope of this curve is defined as the coefficient of Laboratory Shear Strength Tests
compressibility, mv, which is the reciprocal of M: The shear strength parameters, c' and ∅′ (or cT and fT),
may be determined by performing laboratory or in situ
𝑑𝜀𝑧 1 tests.
𝑚𝑣 = = (4.2)
𝑑𝜎𝑧′ 𝑀
When selecting a test method, we must consider many
factors, including the following:
The slope of the virgin curve is called the compression • Soil type
index, Cc: • Initial moisture content and need, if any, to
(4.3) saturate the sample
𝐶𝑐 = −
𝑑𝑒 • Required drainage conditions (drained or
′
𝑑 log 𝜎𝑧 undrained)

Since the virgin curve is approximated as a straight Direct Shear Test

line (on a semilogarithmic plot), we can obtain a
numerical value for Cc by selecting any two points, a - The French engineer Alexandre Collin may
and b, on this line: have been the first to measure the shear
strength of a soil (Head, 1982).
- typically accepts a 60 to 75 mm (2.5–3.0 in)
(4.4) diameter cylindrical specimen and subjects it
to a certain effective stress.
- The test is normally repeated on new
specimens of the same soil until three sets of
Alternatively, if the data is plotted only in the vertical effective stress and shear strength
strain form (𝜀𝑧 − 𝜎𝑧′ ), then: measurements are obtained. A plot of this
data produces values of the cohesion, c, and
friction angle, f.
(4.5) - has the advantage of being simple and
inexpensive and it is an appropriate method
when we need the drained strength of
where: cohesionless soils.
e0 = initial void ratio (i.e., at beginning of test) - has the disadvantages of forcing the shear to
occur along a specific plane instead of
This slope, which we call the recompression index, allowing the soil to fail along the weakest
Cr, is defined in the same way as Cc. zone, and it produces nonuniform strains in
the sample, which can produce erroneous
Using void ratio data: results in strain softening soils.

Unconfined Compression Test

(4.6)
- uses a tall, cylindrical specimen of cohesive
soil subjected to an axial load.
Using strain data: - This load is applied quickly (i.e., only a
couple of minutes to failure) to maintain the
undrained condition.
(4.7) - this load and the stresses in the soil are both
equal to zero.

The cross- sectional area of the specimen increases as

*Another important parameter from the consolidation the test progresses, and the area at failure, Af, is given
test results is the stress that corresponds to the break- by:
in- slope this is called preconsolidation stress, 𝜎𝑧′ . That
(4.9)
represents the greatest vertical effective stress the soil
has ever experienced.
where:
𝐴𝑓 = cross-sectional area at failure
(4.8)
𝐴0 = initial cross-sectional area
𝜀𝑓 = axial strain at failure
Therefore, the undrained shear strength, su, is given and computing the effective
by: stresses.

(4.10) IN SITU TESTING

where - we can often circumvent these problems by

su = undrained shear strength using in situ (in- place) testing methods.
Pf = axial load at failure - It entails bringing the test equipment to the
field and testing the soils in- place.
Triaxial Compression Test - use empirical correlations and calibrations to
convert in situ test results to appropriate
- as an extension of the unconfined engineering properties for design.
compression test.
- allows more flexibility and greater control Advantage:
over the test. - They are usually less expensive, so a greater
- It can measure either the drained or undrained number of tests can be performed, thus
strength of nearly any type of soil. characterizing the soil in more detail.

The three most common types of triaxial compression Disadvantages:

tests are as follows: - Often no sample is obtained, thus making soil
classification more difficult.
1. The unconsolidated- undrained (UU) test - The engineer has less control over confining
(also known as a quick or Q test) stresses and drainage.
o Horizontal and vertical stresses,
usually equal to the vertical stress
that was present in the field. Standard Penetration Test (SPT)
o No consolidation is permitted, and
the soil is sheared under undrained - was originally developed in the late 1920s
conditions. The result is expressed and has been used most extensively in North
as an su value. and South America, the United Kingdom, and
2. The consolidated- drained (CD) test (also Asia.
known as a slow or S test) - it is also plagued by many problems that
o Horizontal and vertical stresses, affect its accuracy and reproducibility and is
usually equal to or greater than the slowly being replaced by other test methods.
vertical - N values may be obtained at intervals no
o it is sheared under drained closer than 450 mm (18 in), and typically the
conditions. test is conducted at depth intervals of 1.5 to
o performed at different confining 3.0 m (5–10 ft).
stresses to find the values of the
drained strength parameters c' and Advantages:
- it obtains a sample of the soil being tested.
∅′ . - it is very fast and inexpensive because it is
3. The consolidated- undrained test (also performed in borings that would have been
known as a rapid or R test) drilled anyway.
o The initial stresses are applied as - nearly all drill rigs used for soil exploration
with the CD test and the soil is are equipped to perform this test.
allowed to consolidate.
o shearing occurs under undrained
conditions.
o The results could be interpreted to
give the su as a function of the
effective consolidation stress
o to obtain the drained strength
parameters c′and ∅′ by measuring
pore water pressures during the test
Corrections to the Test Data Cone Penetration Test (CPT)

The variations in testing procedures may be at least - another common in situ test.
partially compensated by converting the measured N - used extensively in Europe for over half a
to N60 as follows (Skempton, 1986): century, and now is also widely used in North
America and elsewhere.
- have a cone- shaped tip with a 60 degrees
(4.11) apex angle.
- with the two most common sizes having
projected areas of 10 cm2 and 15 cm2.
where - A mini- cone and a large cone with projected
N60 = SPT N value corrected for field areas of 2 cm2 and 40 cm2.
procedures - all CPTs in use today are electronic cones
Em = hammer efficiency (from Table 4.3) [ASTM D5778].
CB = borehole diameter correction (from - The cone is mounted on a set of steel rods and
Table 4.4) a hydraulic ram then pushes this assembly
CS = sampler correction (from Table 4.4) into the ground while instruments measure
CR = rod length correction (from Table 4.4) the resistance to penetration.
N = measured SPT N value
The cone resistance (qc) is the total force acting on the
The SPT data also may be adjusted using an cone divided by its projected area.
overburden correction that compensates for the effects
of effective stress. The corrected value, N1,60, is (Liao The sleeve friction (fs) is the total frictional force
and Whitman, 1985): acting on the friction sleeve divided by the sleeve
surface area.
(4.12 SI)
It is common to express the side friction in terms of the
friction ratio, Rf :

(4.12 English)
(4.13)

where:
N1,60 = SPT N-value corrected for field
procedures and overburden stress o The CPT defines the soil profile with much
greater resolution than does the SPT.
𝜎𝑧′ = vertical effective stress at the test
o CPT rigs are often mounted in large three-
location (kPa or lb/ft2)
axle trucks.
N60 = SPT N-value corrected for field
o Typically capable of producing maximum
procedures
thrusts of 100 to 200 kN (10–20 tons).
*The overburden correction should be applied only
Development of piezocones with pore pressure
when the analysis procedure calls for an N1,60 value.
transducers in order to measure the excess pore water
pressures that develop while conducting the test.
Uses of SPT Data
CPT with pore pressure measurements or CPTu the
- The SPT N value, as well as many other test
enhance procedure and especially useful in saturated
results, is only an index of soil behavior.
clays.
- most of these correlations are very
approximate, especially those based on fairly
Uses of CPT Data
old data that were obtained when test
procedures and equipment were different
- is more repeatable.
from those now used.
- useful way to evaluate soil profiles.
- because of the many uncertainties in the SPT
- is able to detect fine changes in the
results, all of these correlations have a wide
stratigraphy.
margin of error.
- the SPT should be used only in - can be used to assess the engineering
cohesionless soils. properties of the soil through the use of
empirical correlations.
 - is much less prone to error due to differences * n depends on the soil type and stress history and that
in equipment and technique. the value of n can be taken to be 1.0 for clay- like soils,
- correlations intended for use in cohesionless and can be obtained for a general soil using an iterative
soils are generally more accurate and most procedure.
commonly used.
- less accurate in cohesive soils because of the Correlation with Soil Behavior Type
presence of excess pore water pressures.
- The CPT does not recover any soil samples
Corrected CPT Parameters during penetration.
- Soil behavior type is slightly different from
Because of the geometry of the cone penetrometer, the standard soil classification, such as that from
pore pressure just behind the cone, u2, will affect the qc the Unified Soil Classification System, but it
measured and qc must be corrected by the following could be argued that SBT is actually more
equation: useful than USCS.

(4.14) *If n = 1, Qtn and Fr can readily be determined using

Equations 4.15 and 4.16.

where: *If n ≠ 1, an iterative procedure is used to first

qt = corrected cone resistance compute n as follows (Robertson and Cabal, 2012):
a = net area ratio
1. Use a trial n value (use an initial value of 1)
*The value of a is measured in the laboratory and is and compute Qtn and Fr.
typically between 0.70 and 0.85. 2. Compute the SBTn index Ic using the
following equation:
* For sands, qt can be taken to be the same as qc
because qc is usually much larger than u2. (4.18)

Normalized CPT Parameters

3. Compute n using the following equation:
The corrected CPT parameters may be normalized to
obtain dimensionless parameters as follows (4.19)
(Robertson and Wride, 1998; Zhang et al., 2002;
Robertson, 2009; Robertson and Cabal, 2012):

4. Compare the computed n value in Step 3 to

(4.15) the trial n value in Step 1. If the absolute
change in n (Δn) is larger than 0.01, repeat
Steps 1 through 4. This iteration continues
(4.16) until Δn … 0.01, in which case the final n
value will be used to calculate final Qtn.

(4.17) Correlation with SPT N- Value

- is useful to convert results from one to the

other.
- This technique compounds the uncertainties
where: because it uses two correlations— one to
Qtn = normalized cone resistance convert to N, and then another to compute the
Fr = normalized friction ratio desired quantity.
Bq = Pore pressure ratio
𝜎𝑧 = vertical total stress before cone penetration * N60 does not include an overburden correction.
𝜎𝑧′ = vertical effective stress before cone penetration
𝑢0 = equilibrium water pressure
Δu = excess penetration pore pressure
Pa = atmospheric pressure (100 kPa or 1 tsf )
n = stress exponent
 Dilatometer Test Researchers have developed correlations between
these intermediate DMT parameters and certain
- is one of the newer in situ test devices, was engineering properties of the soil (Schmertmann,
developed during the late 1970s in Italy by 1988b; Kulhawy and Mayne, 1990; Marchetti et al.,
Silvano Marchetti. (known as flat dilatometer 2001), including:
or a Marchetti dilatometer). • Soil type
- consists of a 95 mm wide, 15 mm thick metal • Coefficient of at- rest lateral earth pressure,
blade with a thin, flat, expandable, 60 mm K0
diameter circular steel membrane on one • Overconsolidation ratio, OCR
side. • Undrained shear strength, su
- it measures the lateral stress condition and • Friction angle, f′
compressibility of the soil.
• Constrained modulus, M
- is a good way to evaluate soil strength,
whereas the dilatometer assesses
compressibility and in situ stresses.
Correlation with Soil Type
- dilatometer blade is most easily pressed into
the ground using a conventional CPT rig.
Marchetti (1980) correlates soil type with ID as
- is a relative newcomer but has become a
follows:
common engineering tool, especially in
Europe.

(4.20)

(4.21)
*This correlation generally gives reasonable
(4.22)
descriptions of “normal” soils.

Vane Shear Test (VST)

where:
ΔA and ΔB = corrections obtained from - The Swedish engineer John Olsson
membrane calibration developed the vane shear test (VST) in the
ZM = gage zero offset (ZM is the gage pressure 1920s to test the sensitive Scandinavian
when the gage is vented to the marine clays in situ.
atmosphere; use ZM = 0 if the same gage is - This test [ASTM D2573] consists of inserting
used to measure ΔA and ΔB) a metal vane into the soil and rotating it until
the soil fails in shear.
- is only strong enough to be used in soft to
The corrected readings are then used to compute the medium cohesive soils.
intermediate DMT parameters, the material index ID, - The test is performed rapidly (about 1 minute
horizontal stress index KD, and the dilatometer to failure) and therefore measures only the
modulus ED: undrained strength.
- has a cylindrical shape.
(4.23)
The data analysis neglects any shear resistance along
the top and bottom of this cylinder. Usually the vane
(4.24) height- to- diameter ratio is 2, which, when combined
with the applied torque, produces the following
theoretical formula:
(4.25)
(4.26)
where:
𝜎𝑧′ = vertical effective stress before
blade penetration where:
su = undrained shear strength
Tf = torque at failure
d = diameter of vane
* (knowing that the factor of safety was 1.0)

Therefore, an empirical correction factor, λ is applied Comparison of In Situ Test Methods

to the test results:

(4.27)

* An additional correction factor of 0.85 should be

applied to test results from organic soils other than peat
(Terzaghi et al, 1996).

Pressuremeter Test (PMT)

- In 1954, a young French engineering student

named Louis Ménard began to develop a new
type of in situ test.
- is a cylindrical balloon that is inserted into the
ground and inflated.
- Measurements of volume and pressure can be
used to evaluate the insitu stress, DETERMINATION OF SOIL PROPERTIES FOR
compressibility, and strength of the adjacent FOUNDATION DESIGN
soil and thus the behavior of a foundation.
- may be performed in a carefully drilled Unit Weight
boring or the test equipment can be combined
with a small auger to create a self- boring The in situ unit weight is typically determined in the
pressuremeter. laboratory using undisturbed samples. Typical unit
- produces much more direct measurements of weights of different soils are given in Table 3.2.
soil compressibility and lateral stresses than
the SPT and CPT.

Becker Penetration Test

- developed in Canada.
- One method of penetrating through these
very large- grained soils is to use a Becker
hammer drill.
- uses a small diesel pile- driving hammer and
percussion action to drive a 135 to 230 mm
(5.5–9.0 in) diameter double- wall steel
casing into the ground.
- been used successfully on very dense and
coarse soils.
- The cuttings are sent to the top by blowing air
through the casing.
- can be used to assess the penetration
resistance of these soils using the Becker
penetration test.
- The number of blows required to advance the
casing 300 mm (1 ft) is the Becker blow
count, NB.
Overconsolidation Ratio Coefficient of Lateral Earth Pressure at Rest

Correlation with SPT Data Correlation with CPT Data

Empirical correlations that relate OCR to the SPT N60 Kulhawy and Mayne (1990) present an equation that
can be used to estimate the OCR of a sand. For OCR gives a very rough estimate of the K0 value from CPT
in sands, silty sands, and sandy silts, Mayne (2007) data:
gives the following empirical correlation:
(4.32)
(4.28)
*this correlation gives a very rough estimate of the
value of K0.
where:
Pa = atmospheric pressure (100 kPa or 1 tsf) Correlation with DMT Data
m = 0.6 for clean quartzitic sands or 0.8 for
silty sands to sandy silts Marchetti (1980) presents an equation that gives an
estimate of the K0 value for uncemented clays using
Another empirical formula, intended for gravels, is DMT Data:
attributed to Kulhawy and Chen (2007):
(4.33)
(4.29)

Correlation with CPT Data For sands, Baldi et al. (1986) provide a correlation that
relates K0 to KD and qc from the CPT:
A simple way to estimate the OCR of cohesive soils is
through this empirical correlation to Kulhawy and
Mayne (1990) using CPT data: (4.34)

(4.30)
where:
K = 0.005 for “seasoned” sands and 0.002
for “freshly deposited” sands (Marchetti et
where: al., 2001)
k = 0.2 to 0.5 (use an average of 0.3)
Relative Density of Cohesionless Soils
*It is valid for Qt1 < 20, and higher values of k should
be used for aged, heavily overconsolidated clays. Correlation with SPT Data

Correlation with DMT Data The relative density is a useful parameter that
describes the consistency of cohesionless soils. It may
According to Marchetti (1980), the OCR of be determined from SPT data using the following
uncemented clays can be estimated through this empirical correlation (Kulhawy and Mayne, 1990):
empirical correlation with DMT Data:

(4.31) (4.35)

(4.36)
* for normally consolidated clays with OCR = 1, KD =2
and that this correlation does not apply to “cemented-
(4.37)
structured- aged” clays.
(4.38)
where
Dr = relative density (in decimal form)
N1,60 = SPT N-value corrected for field
procedures and overburden stress
CP = grain-size correction factor
CA = aging correction factor
COCR = overconsolidation correction factor
D50 = grain size at which 50 percent of the soil
is finer (mm)
t = age of soil (time since deposition) (years)
OCR = overconsolidation ratio