Conference PaperPDF Available

Recent advances in soil liquefaction engineering: A unified and consistent framework

April 2003

April 2003

Conference: Proceedings of the 26th Annual ASCE Los Angeles Geotechnical Spring Seminar: Long Beach, CA

Authors:

Kemal Onder Cetin

Middle East Technical University

Annie Kammerer

Annie Kammerer Consulting

Show all 10 authorsHide

Over the past decade, major advances have occurred in both understanding and practice with regard to assessment and mitigation of hazard associated with seismically induced soil liquefaction. Soil liquefaction engineering has evolved into a sub-field in its own right, and engineering assessment and mitigation of seismic soil liquefaction hazard is increasingly well addressed in both research and practice. This rapid evolution in the treatment of liquefaction has been pushed largely by a confluence of lessons and data provided by a series of major earthquakes over the past dozen years, as well as by the research and professional/political will engendered by these major seismic events. The overall field of soil liquefaction engineering is now beginning to coalesce into an internally consistent and comprehensive framework, and one in which the various elements are increasingly mutually supportive of each other. Although the rate of progress has been laudable, further advances are occurring, and more remains to be done. As we enter a “new millenium”, engineers are increasingly well able to deal with important aspects of soil liquefaction engineering. This paper will highlight a number of important recent and ongoing developments in soil liquefaction engineering, and will offer insights regarding research in progress, as well as suggestions regarding further advances needed.

Content uploaded by Annie Kammerer

Content may be subject to copyright.

Content uploaded by Juan M. Pestana

Content may be subject to copyright.

EARTHQUAKE ENGINEERING RESEARCH CENTER

COLLEGE OF ENGINEERING

UNIVERSITY OF CALIFORNIA, BERKELEY

RECENT ADVANCES IN SOIL LIQUEFACTION

ENGINEERING: A UNIFIED AND CONSISTENT

FRAMEWORK

R.B. Seed

K.O. Cetin

R.E.S. Moss

A.M. Kammerer

J. Wu

J.M. Pestana

M.F. Riemer

R.B. Sancio

J.D. Bray

R.E. Kayen

A. Faris

REPORT NO.

EERC 2003-06

26th Annual ASCE Los Angeles Geotechnical Spring Seminar,

Keynote Presentation, H.M.S. Queen Mary,

Long Beach, California, April 30, 2003.

RECENT ADVANCES IN SOIL LIQUEFACTION ENGINEERING:

A UNIFIED AND CONSISTENT FRAMEWORK

R. B. Seed1, K. O. Cetin2, R. E. S. Moss3, A. M. Kammerer4, J. Wu5, J. M. Pestana1,

M. F. Riemer1, R.B. Sancio1, J.D. Bray1, R. E. Kayen6, and A. Faris1

ABSTRACT

Over the past decade, major advances have occurred in both understanding and practice with regard to assessment and mitigation of

hazard associated with seismically induced soil liquefaction. Soil liquefaction engineering has evolved into a sub-field in its own

right, and engineering assessment and mitigation of seismic soil liquefaction hazard is increasingly well addressed in both research

and practice. This rapid evolution in the treatment of liquefaction has been pushed largely by a confluence of lessons and data

provided by a series of major earthquakes over the past dozen years, as well as by the research and professional/political will

engendered by these major seismic events. The overall field of soil liquefaction engineering is now beginning to coalesce into an

internally consistent and comprehensive framework, and one in which the various elements are increasingly mutually supportive of

each other. Although the rate of progress has been laudable, further advances are occurring, and more remains to be done. As we

enter a “new millenium”, engineers are increasingly well able to deal with important aspects of soil liquefaction engineering. This

paper will highlight a number of important recent and ongoing developments in soil liquefaction engineering, and will offer insights

regarding research in progress, as well as suggestions regarding further advances needed.

1 Dept. of Civil and Environmental Engineering, University of California, Berkeley.

2 Dept. of Civil Engineering, Middle East Technical University, Ankara, Turkey.

3 Fugro Engineering, Santa Barbara, California.

4 Arup, San Francisco, California.

5 URS Corporation, Oakland, California.

6 U.S. Geological Survey, Menlo Park, California.

1.0 INTRODUCTION

Soil liquefaction is a major cause of damage during

earthquakes. “Modern” engineering treatment of liquefaction-

related issues evolved initially in the wake of the two

devastating earthquakes of 1964; the 1964 Niigata (Japan) and

1964 Great Alaskan Earthquakes. Seismically-induced soil

liquefaction produced spectacular and devastating effects in

both of these events, thrusting the issue forcefully to the

attention of engineers and researchers.

Over the nearly four decades that have followed, significant

progress has occurred. Initially, this progress was largely

confined to improved ability to assess the likelihood of

initiation (or “triggering”) of liquefaction in clean, sandy soils.

As the years passed, and earthquakes continued to provide

lessons and data, researchers and practitioners became

increasingly aware of the additional potential problems

associated with both silty and gravelly soils, and the important

additional issues of post-liquefaction strength and stress-

deformation behavior also began to attract increased attention.

Today, the area of “soil liquefaction engineering” is emerging

as a semi-mature field of practice in its own right. This area

now involves a number of discernable sub-issues or sub-

topics, as illustrated schematically in Figure 1. As shown in

Figure 1, the first step in most engineering treatments of soil

liquefaction continues to be (1) assessment of “liquefaction

potential”, or the risk of “triggering” (initiation) of

liquefaction. There have been major advances here in recent

years, and some of these will be discussed.

Once it is determined that occurrence of liquefaction is a

potentially serious risk/hazard, the process next proceeds to

assessment of the consequences of the potential liquefaction.

This, now, increasingly involves (2) assessment of available

post-liquefaction strength, and resulting post-liquefaction

overall stability (of a site, and/or of a structure or other built

facilites, etc.). There has been considerable progress in

Seed et al. (2003) 2

Fig. 1: Key Elements of Soil Liquefaction Engineering

1. Assessment of the likelihood of “triggering”

or initiation of soil liquefaction.

2. Assessment of post-liquefaction strength and

overall post-liquefaction stability.

3. Assessment of expected liquefaction-induced

deformations and displacements.

4. Assessment of the consequences of these

deformations and displacements.

5. Implementation (and evaluation) of engineered

mitigation, if necessary.

evaluation of post-liquefaction strengths and stability over the

past fifteen years. If post-liquefaction stability is found

wanting, then deformation/displacement potential is large, and

engineered remediation is typically warranted.

If post-liquefaction overall stability is not unacceptable, then

attention is next directed towards (3) assessment of anticipated

deformations and displacements. This is a very “soft” area of

practice, and much remains to be done here with regard to

development and calibration/verification of engineering tools

and methods. Similarly, there are few engineering tools and

guidelines regarding (4) the effects of liquefaction-induced

deformations and displacements on the performance of

structures and other engineered facilities, and criteria for

“acceptable” performance are not well established.

Finally, in cases in which the engineer(s) conclude that

satisfactory performance cannot be counted on, (5) engineered

mitigation of liquefaction risk is generally warranted. This,

too, is a rapidly evolving area, and one rife with potential

controversy. Ongoing evolution of new methods for

mitigation of liquefaction hazard provides an ever increasing

suite of engineering options, but the efficacy and reliability of

some of these remain contentious, and accurate and reliable

engineering analysis of the improved performance provided by

many of these mitigation techniques continues to be difficult.

It is not possible, within the confines of this paper, to fully

address all of these issues (a textbook would be required!)

Instead, a number of important recent and ongoing advances

will be highlighted, and resultant issues and areas of

controversy, as well as areas in urgent need of further

advances either in practice or understanding, will be noted.

2.0 ASSESSMENT OF SUSCEPTIBILITY

2.1 Liquefiable Soil Types:

The first step in engineering assessment of the potential for

“triggering” or initiation of soil liquefaction is the

determination of whether or not soils of “potentially

liquefiable nature” are present at a site. This, in turn, raises

the important question regarding which types of soils are

potentially vulnerable to soil liquefaction.

It has long been recognized that relatively “clean” sandy soils,

with few fines, are potentially vulnerable to seismically-

induced liquefaction. There has, however, been significant

controversy and confusion regarding the liquefaction potential

of silty soils (and silty/clayey soils), and also of coarser,

gravelly soils and rockfills.

Coarser, gravelly soils are the easier of the two to discuss, so

we will begin there. The cyclic behavior of coarse, gravelly

soils differs little from that of “sandy” soils, as Nature has

little or no respect for the arbitrary criteria established by the

standard #4 sieve. Coarse, gravelly soils are potentially

vulnerable to cyclic pore pressure generation and liquefaction.

There are now a number of well-documented field cases of

liquefaction of coarse, gravelly soils (e.g.: Evans, 1987;

Harder, 1988; Hynes, 1988; Andrus, 1994). These soils do,

however, often differ in behavior from their finer, sandy

brethren in two ways: (1) they can be much more pervious,

and so can often rapidly dissipate cyclically generated pore

pressures, and (2) due to the mass of their larger particles, the

coarse gravelly soils are seldom deposited “gently” and so do

not often occur in the very loose states more often encountered

with finer sandy soils. Sandy soils can range from very loose

to very dense, while the “very” loose state is relatively

uncommon in gravelly deposits and coarser soils.

The apparent drainage advantages of coarse, gravelly soils can

be defeated if their drainage potential is circumvented by

either; (1) their being surrounded and encapsulated by finer,

less pervious materials, (2) if drainage is internally impeded

by the presence of finer soils in the void spaces between the

coarser particles (it should be noted that the D10 particle size,

not the mean or D

50 size, most closely correlates with the

permeability of a broadly graded soil mix), or (3) if the layer

or stratum of coarse soil is of large dimension, so that the

distance over which drainage must occur (rapidly) during an

earthquake is large. In these cases, the coarse soils should be

considered to be of potentially liquefiable type, and should be

evaluated accordingly.

Questions regarding the potential liquefiability of finer,

“cohesive” soils (especially “silts” and “silty clays”) are

increasingly common at meetings and professional short

courses and seminars. There is considerable new field data

regarding this issue from recent major earthquakes, and this is

an area in which major changes in both understanding and

practice are occurring.

Seed et al. (2003) 3

Figure 2 illustrates the “Modified Chinese Criteria” (Wang

(1979), and Seed and Idriss (1982)), which represent the

criteria most widely used for defining potentially liquefiable

soils over the past two decades. According to these criteria,

fine (cohesive) soils that plot above the A-line are considered

to be of potentially liquefiable type and character if: (1) there

are less than 15% “clay” fines (based on the Chinese

definition of “clay” sizes as less than 0.005 mm), (2) there is a

Liquid Limit of LL ≤ 35%, and (3) there is a current in-situ

water content greater than or equal to 90% of the Liquid

Limit.

Andrews and Martin (2000) re-evaluated the liquefaction field

case histories from the database of Wang (1979), as well as a

number of subsequent earthquakes, and have transposed the

“Modified Chinese Criteria” to U.S. conventions (with clay

sizes defined as those less than about 0.002 mm). Their

findings are largely summarized in Figure 3. Andrews and

Martin recommended: (1) that soils with less than about 10%

clay fines (< 0.002 mm), and a Liquid Limit (LL) in the minus

#40 sieve fraction of less than 32%, be considered potentially

liquefiable, (2) that soils with more than about 10% clay fines

and LL ≥ 32% are unlikely to be susceptible to classic

cyclically-induced liquefaction, and (3) that soils intermediate

between these criteria should be sampled and tested to assess

whether or not they are potentially liquefiable.

Over the period from 1994 to 1999, a group of approximately

two dozen leading experts worked to achieve concensus

regarding a number of issues involved in the assessment of

liquefaction potential. This group, referred to hereafter as the

NCEER Working Group, have published many of their

consensus findings (or at least near-consensus findings) in the

NSF-sponsored workshop summary paper (NCEER, 1997),

and the summary article in the ASCE Journal of Geotechnical

and Geoenvironmental Engineering (Youd et al., 2001). The

NCEER Working Group addressed this issue, and it was

agreed that there was a need to reexamine the “Modified

Chinese Criteria” for defining the types of fine “cohesive”

soils potentially vulnerable to liquefaction, but no improved

concensus position could be reached at that time, and more

study was warranted.

Two major earthquakes in 1999 then dramatically altered the

picture. Widespread soil liquefaction occurred throughout

much of the city of Adapazari in the 1999 Kocaeli (Turkey)

Earthquake, and widespread liguefaction-induced damages

also occurred in the cities of Wu Feng, Yuan Lin and Nantou

in the 1999 Chi-Chi (Taiwan) Earthquake. In all four of these

cities, significant liquefaction-type damages (including

settlements and/or partial or complete bearing failures of

shallow-founded structures) occurred at sites where the soils

responsible appear to be more “cohesive” than would be

expected based on the Modified Chinese Criteria.

There is significant ongoing research with regard to the field

performance of increasingly cohesive soils in Adapazari; work

is in progress both at U.C. Berkeley (Sancio, 2003) and at the

Middle East Technical University in Ankara (Cetin, 2003),

and more detailed publications can be anticipated in the very

near future as these efforts are completed. Similarly, studies

are also in progress by a number of research teams (including

Stewart, et al., 2003) regarding performance of increasingly

cohesive soils in Wu Feng, Yuan Lin and Nantou during the

Chi-Chi Earthquake.

In the “new” field performance cases in these four cities, it is

often difficult to reliably discern whether or not soils with

cohesive fines “liquefied”. Soils with large fines contents do

not generally exude excess pore pressures rapidly, and so are

less prone to produce surface boil ejecta than are “cleaner”

cohesionless soils.

As a result, soils with significant (and plastic) fines have been

sampled and then subjected to cyclic testing in the laboratory

by a number of researchers. This laboratory testing, much like

Liquid Limit

< 32

Liquid Limit ≥ 32

Clay Content2

< 10%

Susceptible Further Studies

Required

(Considering plastic

non-clay sized

grains – such as

Mica)

Clay Content2

≥ 10%

Further Studies

Required

(Considering non-

plastic clay sized

grains – such as mine

and quarry tailings)

Not Susceptible

Notes:

1. Liquid limit determined by Casagrande-type percussion apparatus.

2. Clay defined as grains finer than 0.002 mm.

Fig. 3: Liquefaction Susceptibility of Silty and Clayey

Sands (after Andrews and Martin, 2000)

SAFE

TEST

IQU

I D

I MI T ,L

,(%

)

NATURALWATER CONTENT,W(%)

100

0.9XLL

1.PercentFinerthan 0.005mm 15%

2.Liquid Limit(LL)35%

3.WaterContent(W)0.9xLL

Fig. 2: Modified Chinese Criteria (After Wang (1979)

and Seed and Idriss (1982)

Seed et al. (2003) 4

the observed field performance, suggests that: (1) soils of

higher plasticity may be susceptible to significant cyclic pore

pressure increase and consequent loss of strength than is

suggested by the Modified Chinese Criteria, and (2) the

transition in behavior to soils of even higher plasticity, which

do not appear to be prone to similarly severe cyclic pore

pressure generation and strength loss, is gradual (rather than

abrupt).

Some of the confusion here is related to the definition of

liquefaction. In this paper, the term “classic” cyclic

liquefaction will refer to significant loss of strength and

stiffness due to cyclic pore pressure generation, in contrast to

“sensitivity” or loss of strength due to monotonic shearing

and/or remolding as a result of larger, monotonic (uni-

directional) shear displacements. By making these

distinctions, we are able to separately discuss “classic”

cyclically-induced liquefaction and the closely-related (but

different) phenomenon of strain-softening or sensitivity.

Sandy soils, and silty soils of very low plasticity, tend to

experience “triggering” of cyclically induced soil liquefaction

at relatively low shear strains (typically on the order of 3% to

6%), and the loss of strength can be severe. Soils of higher

plasticity, on the other hand, may also exhibit loss of strength

and stiffness, accompanied by increased pore pressures, but

the pore pressure ratios achieved may be somewhat lower than

those associated with more “classically” liquefiable soils, and

the loss of strength and stiffness becomes pronounced at

somewhat larger shear strains. In other words, there is a

transition in behaviors; as soils’ behaviors become controlled

by fines of increasing plasticity their cyclic behavior becomes

more “ductile”, and the boundary between soils which are

potentially susceptible to “classic” cyclic liquefaction and

those that are not is not a sharp transition.

It is recommended herein that the Modified Chinese Criteria

be relegated to history, and that we move forward to broader

consideration of potentially liquefiable soil types. One

element of the Modified Chinese Criteria has been clearly

shown to be flawed, and that is the “percent clay fines” rule

(e.g.: Bray et al. 2001; Sancio et al.; 2002, 2003). Percent clay

fines is less important than the overall contribution of the fines

to plasticity, and there are numerous cases of liquefaction of

soils with more than 10 or 15% clay-sized fines. The other

elements of the Modified Chinese Criteria (Liquid Limit, and

water content as a fraction of the liquid limit) both appear

better directed, but warrant some revision as well.

Post-earthquake reconnaissance efforts (e.g. Bray and Stewart,

2000) and follow-on studies (e.g. Sancio et al., 2002), clearly

found ample evidence of liquefaction and ground softening at

sites where critical soil layers contained more than 15%

particles finer than 5 mm. As suggested in Bray et al. (2001),

Sancio et al. (2002), and Sancio et al. (2003), the percent

"clay-size" criterion of the Chinese criteria and Andrews and

Martin (2000) criteria is misleading, because it is not the

percent of "clay-size" particles that is important. Rather, it is

the percent of clay minerals present in the soil and their

activity that are important. Fine quartz particles may be

smaller than either 2 or 5 mm, but if largely nonplastic, these

soils respond as a cohesionless material in terms of

liquefaction under cyclic loading. Accordingly, use of the

percent "clay size" criterion as is commonly done in current

engineering practice (e.g.: "Guidelines for Analyzing and

Mitigating Liquefaction Hazards in California"; edited by

Martin and Lew, 1999), can be unconservative, because soils

that are susceptible to liquefaction can be incorrectly classified

as non-liquefiable.

Figure 4 represents interim

recommendations regarding

“liquefiability of soils with

significant fines contents. This may

evolve further, based on work in

progress, but is a good summary of

what we know to date. For soils

with sufficient fines content that the

fines separate the coarser particles

and control overall behavior: (1)

Soils within Zone A are considered

potentially susceptible to “classic”

cyclically induced liquefaction, (2)

Soils within Zone B may be

liquefiable, and (3) Soils in Zone C

(not within Zones A or B) are not

generally susceptible to “classic”

cyclic liquefaction, but should be

checked for potential sensitivity

(loss of strength with remoulding or

monotonic accumulation of shear

deformation).

0 10 20 30 40 50 60 70 80 90 100

LL (Liquid Limit)

PI (Plasticity Index)

U-line

A-line

37 47

CL -ML ML

Applicable for:

(a) FC ≥ 20% if PI>12%

(b) FC ≥ 35% if PI<12%

Zone A: Potentially Liquefiable

Zone B: Test if wc ≥ 0.85(LL)

Fig 4: Recommendations Regarding Assessment of “Liquefiable” Soil Types

if wc > 0.8(LL)

Seed et al. (2003) 5

Both experimental research and review of liquefaction field

case histories show that for soils with sufficient “fines”

(particles finer than 0.074 mm, or passing a #200 sieve) to

separate the coarser (larger than 0.074 mm) particles, the

characteristics of the fines control the potential for cyclically-

induced liquefaction. This separation of the coarser particles

typically occurs as the fines content exceeds about 15% to

35%, with the precise fines content required being dependent

principally on the overall soil gradation and the character of

the fines. Well-graded soils have lesser void ratios than

uniformly-graded or gap-graded soils, and so require lesser

fines contents to fill the remaining available void space and

thus separate (or “float”) the coarser particles in a matrix of

the fines. Similarly, clay fines carry higher void ratios than

silty particles and so are more rapidly effective at over-filling

the void space available between the coarser (larger than

0.074mm) particles; a lesser weight (or percentage) of clay

fines is required than would be required if the fines were lower

plasticity silty particles.

In soils wherein the fines content is sufficient as to separate

the coarser particles and control behavior, cyclically-induced

soil liquefaction appears to occur primarily in soils where

these fines are either non-plastic or are low plasticity silts

and/or silty clays (PI ≤ 12%, and LL ≤ 37%), and with high

water content relative to their Liquid Limit (wc > 0.85·LL). In

fact, low plasticity or non-plastic silts and silty sands can be

among the most dangerous of liquefiable soils, as they not

only can cyclically liquefy; they also “hold their water” well

and dissipate excess pore pressures slowly due to their low

permeabilities.

Soils with sufficient fines that the fines control their behavior,

and falling within Zone A in Figure 4, are considered

potentially susceptible to “classic” cyclically-induced soil

liquefaction. Soils within Zone B fall into a transition range;

they may in some cases be susceptible to liquefaction

(especially if their in situ water content is greater than about

85% of their Liquid Limit), but tend to be more ductile and

may not “liquefy” in the classic sense of losing a large fraction

of their strength and stiffness at relatively low cyclic shear

strains. These soils are also, in many cases, not well suited to

evaluation based on conventional in-situ “penetration-based”

liquefaction hazard assessment methods. These types of soils

usually are amenable to reasonably “undisturbed” (e.g.: thin-

walled, or better) sampling, however, and so can be tested in

the laboratory. It should be remembered to check for

“sensitivity” of these cohesive soils as well as for potential

cyclic liquefiability. Soils in Zone C are generally not

susceptible to “classic” cyclically-induced soil liquefaction,

but they may be “sensitive” and vulnerable to strength loss

with remoulding or large shear displacements.

This is a step forward, as it extends the previous “Modified

Chinese” criteria to encompass important new field

performance data (and corollary laboratory test data) from

recent earthquakes. It should also be noted that there is a

common lapse in engineering practice inasmuch as engineers

often tend to become distracted by the presence of potentially

“classically” liquefiable soils, and then often neglect cohesive

soils (clays and plastic silts) that are highly “sensitive” and

vulnerable to major loss of strength if sheared or remolded.

These types of “sensitive” soils (which can exist in Zones B

and C) often co-exist in close proximity with potentially

liquefiable soils, and can be similarly dangerous in their own

right.

Appropriate sampling and testing protocols for soils of Zone B

are not yet well established, and further research is needed

here. Issues of sample disturbance, and sample densification

during reconsolidation, and the potential applicability of

“SHANSEP-like” laboratory reconsolidation approaches to

offset these potential problems, are not yet well studied.

Accordingly, sampling and testing of these types of soils may

produce important qualitative data regarding likely soil

performance, but it is difficult to rigorously quantitatively

assess the levels of seismic loading necessary to “trigger”

liquefaction in these soil types at present. It should also be

noted that soils of Zone B may sometimes exhibit relatively

innocuous behavior under cyclic loading in the absence of

“static” driving shear stresses, but may exhibit much more

significant softening and pore pressure increase if cyclically

loaded while also subjected to significant “static” driving

shear stresses. Accordingly, it appears that cyclic testing of

these types of soils with non-zero static driving shear stresses

(á > 0) is adviseable if this is potentially applicable to field

conditions.

The criteria of this section do not fully cover all types of

liquefiable soils. As an example, a well-studied clayey sand

(SC) at a site in the southeastern U.S. has been clearly shown

to be potentially susceptible to cyclic liquefaction, despite a

clay content on the order of 15 %, and a Plasticity Index of up

to 30% (Riemer et al., 1993). This is a highly unusual

material, however, as it is an ancient sand that has weathered

in place, with the clay largely coating the exterior surfaces of

the individual weathered grains, and the overall soil is

unusually “loose”. Exceptions must be anticipated, and

judgement will continue to be necessary in evaluating whether

or not specific soils are potentially liquefiable.

Finally, two additional conditions necessary for potential

liquefiability are: (1) saturation (or at least near-saturation),

and (2) “rapid” (largely “undrained”) loading. It should be

remembered that phreatic conditions are variable both with

seasonal fluctuations and irrigation, and that the rapid cyclic

loading induced by seismic excitation represents an ideal

loading type for initiation of soil liquefaction.

3.0 ASSESSMENT OF TRIGGERING POTENTIAL

Quantitative assessment of the likelihood of “triggering” or

initiation of liquefaction is the necessary first step for most

projects involving potential seismically-induced liquefaction.

There are two general types of approaches available for this:

(1) use of laboratory testing of “undisturbed” samples, and (2)

use of empirical relationships based on correlation of observed

field behavior with various in-situ “index” tests.

Seed et al. (2003) 6

The use of laboratory testing is complicated by difficulties

associated with sample disturbance during both sampling and

reconsolidation. It is also difficult and expensive to perform

high-quality cyclic simple shear testing, and cyclic triaxial

testing poorly represents the loading conditions of principal

interest for most seismic problems. Both sets of problems can

be ameliorated, to some extent, by use of appropriate “frozen”

sampling techniques, and subsequent testing in a high quality

cyclic simple shear or torsional shear apparatus. The

difficulty and cost of these delicate techniques, however,

places their use beyond the budget and scope of most

engineering studies. In addition, frozen sampling can be

infeasible in soils with significant fines content, as the low

permeability of these can lead to ice expansion completely

disturbing the soils rather than preventing disturbance.

Accordingly, the use of in-situ “index” testing is the dominant

approach in common engineering practice. As summarized in

the recent state-of-the-art paper (Youd et al.; 1997, 2001), four

in-situ test methods have now reached a level of sufficient

maturity as to represent viable tools for this purpose, and these

are (1) the Standard Penetration Test (SPT), (2) the cone

penetration test (CPT), (3) measurement of in-situ shear wave

velocity (Vs), and (4) the Becker penetration test (BPT). The

oldest, and still the most widely used of these, is the SPT, and

this will be the focus of the next section of this paper.

3.1 SPT-Based Triggering Assessment:

3.1.1 Existing SPT-Based Correlations

The use of the SPT as a tool for evaluation of liquefaction

potential first began to evolve in the wake of a pair of

devastating earthquakes that occurred in 1964; the 1964 Great

Alaskan Earthquake (M = 8+) and the 1964 Niigata

Earthquake (M ≈ 7.5), both of which produced significant

liquefaction-related damage (e.g.: Kishida, 1966; Koizumi,

1966; Ohsaki, 1966; Seed and Idriss, 1971). Numerous

additional researchers have made subsequent progress, and

these types of SPT-based methods continue to evolve today.

As discussed by the NCEER Working Group (NCEER, 1997;

Youd et al., 2001), one of the most widely accepted and

widely used SPT-based correlations is the “deterministic”

relationship proposed by Seed, et al. (1984, 1985). Figure 5

shows this relationship, with minor modification at low CSR

(as recommended by the NCEER Working Group; NCEER,

1997). This familiar relationship is based on comparison

between SPT N-values, corrected for both effective

overburden stress and energy, equipment and procedural

factors affecting SPT testing (to N1,60-values) vs. intensity of

cyclic loading, expressed as magnitude-weighted equivalent

uniform cyclic stress ratio (CSReq). The relationship between

corrected N

1,60-values and the intensity of cyclic loading

required to trigger liquefaction is also a function of fines

content in this relationship, as shown in Figure 5.

Although widely used in practice, this relationship is dated,

and does not make use of an increasing body of field case

history data from seismic events that have occurred since

1984. It is particularly lacking in data from cases wherein

peak ground shaking levels were high (CSR > 0.25), an

increasingly common design range in regions of high

seismicity. This correlation also has no formal probabilistic

basis, and so provides no insight regarding either uncertainty

or probability of liquefaction.

Efforts at development of similar, but formally

probabilistically-based, correlations have been published by a

number of researchers, including Liao et al. (1988, 1998), and

more recently Youd and Noble (1997), and Toprak et al.

(1999). Figures 6(a) through (c) show these relationships,

expressed as contours of probability of triggering of

liquefaction, with the deterministic relationship of Seed et al.

from Figure 5 superimposed (dashed lines) for reference. In

each of the figures on this page, contours of probability of

triggering or initiation of liquefaction for P

L = 5, 20, 50, 80

and 95% are shown.

The probabilistic relationship proposed by Liao et al. employs

a larger number of case history data points than were used by

Seed et al. (1984), but this larger number of data points is the

Fig. 5: Correlation Between Equivalent Uniform Cyclic

Stress Ratio and SPT N1,60-Value for Events of

Magnitude MW7.5 for Varying Fines Contents,

With Adjustments at Low Cyclic Stress Ratio as

Recommended by NCEER Working Group

(Modified from Seed, et al., 1984)

Seed et al. (2003) 7

result of less severe screening of points for data quality, and so

includes a number of low quality data. This relationship was

developed using the maximum likelihood estimation method

for probabilistic regression (binary regression of logistic

models). The way the likelihood function was formulated did

not permit separate treatment of aleatory and epistemic

sources of uncertainty, and so overstates the overall variance

or uncertainty of the proposed correlation. This can lead to

over-conservatism at low levels of probability of liquefaction.

An additional shortcoming was that Liao et al. sought, but

failed to find, a significant impact of fines content on the

regressed relationship between SPT penetration resistance and

liquefaction resistance, and so developed reliable curves

(Figure 6(a)) only for sandy soils with less than 12% fines.

The relationship proposed by Youd and Noble employs a

number of field case history data points from earthquakes

which have occurred since the earlier relationships were

developed, and excludes the most questionable of the data

used by Liao et al. The basic methodology employed,

maximum likelihood estimation, is the same, however, and as

a result this correlation continues to overstate the overall

uncertainty. The effects of fines content were judgmentally

prescribed, a priori, in these relationships, and so were not

developed as part of the regression. This correlation is

applicable to soils of variable fines contents, and so can be

employed for both sandy and silty soils. As shown in Figure

6(b), however, uncertainty (or variance) is high.

The relationship proposed by Toprak et al. also employs an

enlarged and updated field case history database, and deletes

the most questionable of the data used by Liao et al. As with

the studies of Youd et al., the basic regression tool was binary

regression, and the resulting overall uncertainty is again very

large. Similarly, fines corrections and magnitude-correlated

duration weighting factors were prescribed a priori, rather than

being regressed from the field case history data, further

decreasing model “fit” (and increasing variance and

uncertainty).

Overall, the four prior relationships presented in Figures 5 and

6(a) through (c) are all excellent efforts, and are among the

best of their types. It is proposed that more can now be

achieved, however, using more powerful and flexible

probabilistic tools, and taking fullest possible advantage of the

currently available field case histories and current knowledge

affecting the processing and interpretation of these.

3.1.2 Proposed New SPT-Based Correlations:

This section presents new correlations for assessment of the

likelihood of initiation (or “triggering”) of soil liquefaction

(Cetin, et al.; 2000, 2003). These new correlations eliminate

several sources of bias intrinsic to previous, similar

correlations, and provide greatly reduced overall uncertainty

and variance. Figure 6(d) shows the new correlation, with

contours of probability of liquefaction again plotted for PL = 5,

20, 50, 80 and 95%, and plotted to the same scale as the earlier

correlations. As shown in this figure, the new correlation

provides greatly reduced overall uncertainty. Indeed, the

uncertainty is now sufficiently reduced that the principal

uncertainty now resides where it belongs; in the engineer’s

ability to assess suitable CSR and representative N1,60 values

for design cases.

Key elements in the development of this new correlation were:

(1) accumulation of a significantly expanded database of field

performance case histories, (2) use of improved knowledge

and understanding of factors affecting interpretation of SPT

data, (3) incorporation of improved understanding of factors

affecting site-specific ground motions (including directivity

effects, site-specific response, etc.), (4) use of improved

methods for assessment of in-situ cyclic shear stress ratio

(CSR), (5) screening of field data case histories on a

quality/uncertainty basis, and (6) use of higher-order

probabilistic tools (Bayesian Updating). These Bayesian

methods (a) allowed for simultaneous use of more descriptive

variables than most prior studies, and (b) allowed for

appropriate treatment of various contributing sources of

aleatory and epistemic uncertainty. The resulting relationships

not only provide greatly reduced uncertainty, they also help to

resolve a number of corollary issues that have long been

difficult and controversial, including: (1) magnitude-correlated

duration weighting factors, (2) adjustments for fines content,

and (3) corrections for effective overburden stress.

As a starting point, all of the field case histories employed in

the correlations shown in Figures 5 and 6(a) through (c) were

obtained and studied. Additional cases were also obtained,

including several proprietary data sets. Eventually,

approximately 450 liquefaction (and “non-liquefaction”) field

case histories were evaluated in detail. A formal rating system

was established for rating these case histories on the basis of

data quality and uncertainty, and standards were established

for inclusion of field cases in the final data set used to

establish the new correlations. In the end, 203 of the field

case histories were judged to meet these new and higher

standards, and were employed in the final development of the

proposed new correlations.

A significant improvement over previous efforts was the

improved evaluation of peak horizontal ground acceleration at

each earthquake field case history site. Specific details are

provided by Cetin et al. (2001, 2003). Significant improve-

ments here were principally due to improved understanding

and treatment of issues such as (a) directivity effects, (b)

effects of site conditions on response, (c) improved attenuation

relationships, and (d) availability of strong motion records

from recent (and well-instrumented) major earthquakes. In

these studies, peak horizontal ground acceleration (amax) was

taken as the geometric mean of two recorded orthogonal

horizontal components. Whenever possible, attenuation

relationships were calibrated on an earthquake-specific basis,

based on local strong ground motion records, significantly

reducing uncertainties. For all cases wherein sufficiently

detailed data and suitable nearby recorded ground motions

were available, site-specific site response analyses were

Seed et al. (2003) 8

(a) Liao et al., 1988 (b) Youd et al., 1998

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

(N1)60

CSRN

FC ≥≥ 35% 15% ≤≤5%

95% 5%50%80% 20%

Liao, et al. (1988)

Deterministic Bounds,

Seed, et al. (1984)

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

(N1)60,cs

CSRN

15%

95% 5%50%80% 20%

Deterministic Bounds,

Seed, et al. (1984)

Youd, et al. (1998)

FC ≥≥ 35% ≤≤ 5%

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

(N1)60,cs

CSRN

15%

95%

50%80% 20%

Deterministic Bounds,

Seed, et al. (1984)

Toprak et al. (1999)

FC ≥≥ 35% ≤≤ 5%

Fig. 6: Comparison of Best Available Probabilistic Correlations for Evaluation of Liquefaction Potential

(All Plotted for MW=7.5, σσv’=0.65 atm, and Fines Content

≤

5%)

Seed et al. (2003) 9

performed. In all cases, both local site effects and rupture-

mechanism-dependent potential directivity effects were also

considered.

A second major improvement was better estimation of in-situ

CSR within the critical stratum for each of the field case

histories. All of the previous studies described so far used the

“simplified” method of Seed and Idriss (1971) to estimate

CSR at depth (within the critical soil stratum) as

( )

peak r

CSR ⋅











′

⋅













=σ

max (Eq. 1)

where

amax = the peak horizontal ground surface acceleration,

g = the acceleration of gravity,

σv = total vertical stress,

σ′v = effective vertical stress, and

rd = the nonlinear shear mass participation factor.

The original rd values proposed by Seed and Idriss (1971) are

shown by the heavy lines in Figure 6(a). These are the values

used in the previous studies by Seed et al. (1984), Liao et al.

(1988, 1998), Youd et al. (1997), and Toprak et al. (1999).

Recognition that rd is nonlinearly dependent upon a suite of

factors led to studies by Cetin and Seed (2000) to develop

improved correlations for estimation of rd. The numerous light

gray lines in Figures 7(a) and (b) show the results of 2,153

seismic site response analyses performed to assess the

variation of r

d over ranges of (1) site conditions, and (2)

ground motion excitation characteristics. The mean and +1

standard deviation values for these 2,153 analyses are shown

by the heavy lines in Figure 7(b). As shown in Figures 7(a)

and (b), the earlier r

d proposal of Seed and Idriss (1971)

understates the variance, and provides biased (generally high)

estimates of rd at depths of between 10 and 50 feet (3 to 15 m.)

Unfortunately, it is in this depth range that the critical soil

strata for most of the important liquefaction (and non-

liquefaction) earthquake field case histories occur. This, in

turn, creates some degree of corresponding bias in

relationships developed on this basis.

Cetin and Seed (2000, 2003) propose a new, empirical basis

for estimation of rd as a function of; (1) depth, (2) earthquake

magnitude, (3) intensity of shaking, and (4) site stiffness (as

expressed in Equation 2).

Figure 8 shows the values of rd from the 2,153 site response

analyses performed as part of these studies sub-divided into 12

“bins” as a function of peak ground surface acceleration (amax),

Fig. 7: Rd Results from Response Analyses for 2,153 Combinations of Site Conditions and Ground

Motions, Superimposed with Heavier Lines Showing (a) the Earlier Recommendations of Seed

and Idriss (1971), and (b) the Mean and + 1 Standard Deviation Values for the 2,153 Cases

Analyzed (After Cetin and Seed, 2000).

(a) (b)

Seed et al. (2003) 10

site stiffness (VS,40ft), earthquake magnitude (Mw), and depth

(d). [VS,40ft is the “average” shear wave velocity over the top

40 feet of a site (in units of ft./sec.), taken as 40 feet divided

by the shear wave travel time in traversing this 40 feet.]

Superimposed on each figure are the mean and + 1 standard

deviation values central to each “bin” from Equation 2. Either

Equation 2, or Figure 8, can be used to derive improved (and

statistically unbiased) estimates of rd.

It is noted, however, that in-situ CSR (and rd) can “jump” or

transition irregularly within a specific soil profile, especially

near sharp transitions between “soft” and “stiff” strata, and

that CSR (and rd) are also a function of the interaction between

a site and each specific excitation motion. Accordingly, the

best means of estimation of in-situ CSR within any given

stratum is to directly calculate CSR by means of appropriate

site-specific, and event-specific, seismic site response

analyses, when this is feasible. As the new correlations were

developed using both directly-calculated rd values (from site

response analyses) as well as rd values from the statistically

unbiased correlation of Equation 2, there is no intrinsic a priori

bias associated with either approach.

This represents an important improvement over all previous

SPT-based “triggering” correlations. All prior correlations

had been based on use of the “simplified” rd of Seed and Idriss

(1971) for back-analysis of field performance case histories,

and were as a result unconservatively biased relative to actual

case-specific seismic response analysis. These previous

methods could be used in forward engineering so long as the

“simplified” r

d was used to assess CSR, but could be

unconservative if used in conjunction with (1-D or 2-D or 3-

D) seismic response analyses (as they often are for

“important” projects such as dams and other critical facilities.)

The new correlations, on the other hand, can be safely used in

conjunction with project-specific dynamic response analyses

without introducing bias.

In the new correlations proposed herein, in-situ cyclic stress

ratio (CSR) is taken as the “equivalent uniform CSR” equal to

65% of the single (one-time) peak CSR (from Equation 1) as

peakeq CSR)65.0(CSR ⋅= (Eq. 3)

In-situ CSReq was evaluated directly, based on performance of

full seismic site response analyses (using SHAKE 90; Idriss

and Sun, 1992), for cases where (a) sufficient sub-surface data

was available, and (b) where suitable “input” motions could be

developed from nearby strong ground motion records. For

cases wherein full seismic site response analyses were not

performed, CSReq was evaluated using the estimated amax and

Equations 1 and 2. In addition to the best estimates of CSReq,

the variance or uncertainty of these estimates (due to all

contributing sources of uncertainty) was also assessed (Cetin

et al., 2001).

At each case history site, the critical stratum was identified as

the stratum most susceptible to triggering of liquefaction.

Only one critical stratum was analyzed at any one site, and in

many cases two or more SPT borings were combined jointly

to characterize a single critical stratum. When possible,

collected surface boil materials were also considered, but

problems associated with mixing and segregation during

transport, and recognition that liquefaction of underlying strata

can result in transport of overlying soils to the surface through

boils, limited the usefulness of some of this data.

The N

1,60-values employed were “truncated mean values”

within the critical stratum. Measured N-values (from one or

more points) within a critical stratum were corrected for

d < 65 ft:

*04,s

)888.24V0785.0(104.0

*04,swmax

)888.24V0785.0d(104.0

*04,swmax

*04,smaxwd

e201.0258.16

V016.0M999.0a949.2013.23

e201.0258.16

V016.0M999.0a949.2013.23

)V,a,M,d(r ε

+⋅⋅

′

+⋅+−⋅

′

′σ±













⋅+

⋅+⋅+⋅−−













⋅+

⋅+⋅+⋅−−

′

′ (Eq 2)

d ≥≥ 65 ft:

*04,s

)888.24V0785.0(104.0

*04,swmax

)888.24V0785.065(104.0

*04,swmax

*04,smaxwd)65d(0014.0

e201.0258.16

V016.0M999.0a949.2013.23

e201.0258.16

V016.0M999.0a949.2013.23

)V,a,M,d(r ε

+⋅⋅

′

+⋅+−⋅

′

′σ±−⋅−













⋅+

⋅+⋅+⋅−−













⋅+

⋅+⋅+⋅−−

′

where

0072.0d)d(850.0

rd⋅=σε [for d < 40 ft], and 0072.040)d(850.0

rd⋅=σε [for d ≥ 40 ft]

Seed et al. (2003) 11

(a) Mw≥6.8, amax≤0.12g, Vs,40 ft. ≤525 fps (b) Mw≥6.8, amax ≤0.12g, Vs,40 ft. >525 fps

Fig. 8: Rd Results for Various “Bins” Superimposed with the Predictions (Mean and Mean ±±1σσ) Based

on Bin Mean Values of Vs,40 ft , Mw, and amax (continued…)

Seed et al. (2003) 12

(e) Mw≥6.8, 0.12< amax ≤0.23g, Vs,40 ft. ≤525 fps (f) Mw≥6.8, 0.12< amax ≤0.23g, Vs,40 ft. >525 fps

(g) Mw<6.8, 0.12< amax ≤0.23g, Vs,40 ft. ≤525 fps (h) Mw<6.8, 0.12< amax ≤0.23g, V.s,40 ft. >525 fps

Fig. 8: Rd Results for Various “Bins” Superimposed with the Predictions (Mean and Mean ±±1σσ) Based

on Bin Mean Values of Vs,40 ft , Mw, and amax (continued…)

Seed et al. (2003) 13

(i) Mw≥6.8, 0.23< amax, Vs,40 ft. ≤525 fps (j) Mw≥6.8, 0.23< amax, Vs,40 ft. >525 fps

(k) Mw<6.8, 0.23< amax, Vs,40 ft. ≤525 fps (l) Mw<6.8, 0.23< amax, Vs,40 ft. >525 fps

Fig. 8: Rd Results for Various “Bins” Superimposed with the Predictions (Mean and Mean ±±1σσ) Based

on Bin Me an Values of Vs,40 ft , Mw, and amax

Seed et al. (2003) 14

overburden, energy, equipment, and procedural effects to N1,60

values, and were then plotted vs. elevation. In many cases, a

given soil stratum would be found to contain an identifiable

sub-stratum (based on a group of localized low N1,60-values)

that was significantly more critical than the rest of the stratum.

In such cases, the sub-stratum was taken as the “critical

stratum”. Occasional high values, not apparently

representative of the general characteristics of the critical

stratum, were considered “non-representative” and were

deleted in a number of the cases. Similarly, though less often,

very low N1,60 values (very much lower than the apparent main

body of the stratum, and often associated with locally high

fines content) were similarly deleted. The remaining,

corrected N

1,60 values were then used to evaluate both the

mean of N1,60 within the critical stratum, and the variance in

N1,60.

For those cases wherein the critical stratum had only one

single useful N1,60-value, the coefficient of variation was taken

as 20%; a value typical of the larger variances among the

cases with multiple N

1,60 values within the critical stratum

(reflecting the increased uncertainty due to lack of data when

only a single value was available).

All N-values were corrected for overburden effects (to the

hypothetical value, N

1, that “would” have been measured if

the effective overburden stress at the depth of the SPT had

been 1 atmosphere) [1 atm. ≈ 2,000 lb/ft2 ≈ 1 kg/cm2 ≈ 14.7

lb/in2 ≈ 101 kPa] as

N1CNN ⋅= (Eq. 4(a))

where CN is taken (after Liao and Whitman, 1986) as

CN = 1

σ’v

___

0.5

(Eq. 4(b))

where σ’v is the actual effective overburden stress at the depth

of the SPT in atmospheres.

The resulting N1 values were then further corrected for energy,

equipment, and procedural effects to fully standardized N1,60

values as

EBSR160,1 CCCCNN ⋅⋅⋅⋅= (Eq. 5)

where CR = correction for “short” rod length,

CS = correction for non-standardized sampler

configuration,

CB = correction for borehole diameter, and

CE = correction for hammer energy efficiency.

The corrections for CR, CS, CB and CE employed correspond

largely to those recommended by the NCEER Working Group

(NCEER, 1997; Youd et al., 2001).

Table 1 summarizes the correction factors used in these

studies. The correction for “short” rod length between the

driving hammer and the penetrating sampler was taken as a

nonlinear “curve” (Figure 9), rather than the incremental

values of the NCEER Workshop recommendations, but the

two agree reasonably well at all NCEER mid-increments of

length.

CS was applied in cases wherein a “nonstandard” (though very

common) SPT sampler was used in which the sampler had an

internal space for sample liner rings, but the rings were not

used. This results in an “indented” interior liner annulus of

enlarged diameter, and reduces friction between the sample

and the interior of the sampler, resulting in reduced overall

penetration resistance (Seed et al., 1984 and 1985). The

reduction in penetration resistance is on the order of ~10 % in

loose soils (N1<10 blows/ft), and ~30 % in very dense soils

(N1>30 blows/ft), so CS varied from 1.1 to 1.3 over this range.

Borehole diameter corrections (CB) were as recommended in

the NCEER Workshop Proceedings (NCEER, 1997; Youd et

al., 2001).

0.7 0.8 0.9 1

Rod Length (m)

Fig. 9: Recommended CR Values (rod length from point

of hammer impact to tip of sampler).

Seed et al. (2003) 15

Table 1: Recommended Corrections for SPT Equipment Energy and Procedures

CR (See Fig. 9 for Rod Length Correction Factors)

CS For samplers with an indented space for interior liners, but with liners omitted during sampling,

(Eq. T-1)

With limits as 1.10 ≤ CS ≤1.30

CB Borehole diameter Correction (CB)

65 to 115 mm 1.00

150 mm 1.05

200 mm 1.15

E ER

60%

(Eq. T-2)

where ER (efficiency ratio) is the fraction or percentage of the theoretical SPT impact hammer

energy actually transmitted to the sampler, expressed as %

• The best approach is to directly measure the impact energy transmitted with each blow.

When available, direct energy measurements were employed.

• The next best approach is to use a hammer and mechanical hammer release system that has

been previously calibrated based on direct energy measurements.

• Otherwise, ER must be estimated. For good field procedures, equipment and monitoring,

the following guidelines are suggested:

Equipment Approximate ER (see Note 3) CE (see Note 3)

-Safety Hammer1 0.4 to 0.75 0.7 to 1.2

-Donut Hammer1 0.3 to 0.6 0.5 to 1.0

-Donut Hammer2 0.7 to 0.85 1.1 to 1.4

-Automatic-Trip Hammer 0.5 to 0.8 0.8 to 1.4

(Donut or Safety Type)

• For lesser quality fieldwork (e.g.: irregular hammer drop distance, excessive sliding

friction of hammer on rods, wet or worn rope on cathead, etc.) further judgmental

adjustments are needed.

Notes: (1) Based on rope and cathead system, two turns of rope around cathead, “normal” release

(not the Japanese “throw”), and rope not wet or excessively worn.

(2) Rope and cathead with special Japanese “throw” release. (See also Note 4.)

(3) For the ranges shown, values roughly central to the mid-third of the range are more

common than outlying values, but ER and CE can be even more highly variable than the

ranges shown if equipment and/or monitoring and procedures are not good.

(4) Common Japanese SPT practice requires additional corrections for borehole diameter

and for frequency of SPT hammer blows. For “typical” Japanese practice with rope

and cathead, donut hammer, and the Japanese “throw” release, the overall product of

CB x CE is typically in the range of 1.0 to 1.3.

1,60

C= 1+

100

Seed et al. (2003) 16

Corrections for hammer energy (CE), which were often

significant, were largely as recommended by the NCEER

Working Group, except in those cases where better

hammer/system-specific information was available. Cases

where better information was available included cases where

either direct energy measurements were made during driving

of the SPT sampler, or where the hammer and the

raising/dropping system (and the operator, when appropriate)

had been reliably calibrated by means of direct driving energy

measurements.

Within the Bayesian updating analyses, which were performed

using a modified version of the program BUMP (Geyskens

et al., 1993), all field case history data were modeled not as

“points”, but rather as distributions, with variances in both

CSR and N

1,60. These regression-type analyses were

simultaneously applied to a number of contributing variables,

and the resulting proposed correlations are illustrated in

Figures 6(d) and 10 through 12, and are expressed in

Equations 6 through 12.

Figure 10 shows the proposed probabilistic relationship

between duration-corrected equivalent uniform cyclic stress

ratio (CSReq), and fines-corrected penetration resistances

(N1,60,cs), with the correlations as well as all field data shown

normalized to an effective overburden stress of σ’v = 0.65 atm.

(1,300 lb/ft2). The contours shown (solid lines) are for

probabilities of liquefaction of P

L=5%, 20%, 50%, 80%, and

95%. All “data points” shown represent median values, also

corrected for duration and fines. These are superposed

(dashed lines) with the relationship proposed by Seed et al.

(1984) for reference.

As shown in this figure, the “clean sand” (Fines Content 

5%) line of Seed et al. (1984) appears to corresponds roughly

to PL≈50%. This is not the case, however, as the Seed et al.

(1984) line was based on biased values of CSR (as a result of

biased r

d at shallow depths, as discussed earlier). The new

correlation uses actual event-specific seismic site response

analyses for evaluation of in-situ CSR in 53 of the back-

analyzed case histories, and the new (and statistically

unbiased) empirical estimation of rd (as a function of level of

shaking, site stiffness, and earthquake magnitude) as presented

in Equation 2 and Figure 8 (Cetin and Seed, 2000) for the

remaining 148 case histories. The new (improved) estimates

of in-situ CSR tend to be slightly lower, typically on the order

of ∼ 5 to 15% lower, at the shallow depths that are critical in

most of the field case histories. Accordingly, the CSR’s of the

new correlation are also, correspondingly, lower by about 5 to

15%, and a fully direct comparison between the new

correlation and the earlier recommendations of Seed et al.

(1984) cannot be made.

It should be noted that the use of slightly biased (high) values

of rd was not problematic in the earlier correlation of Seed et

al. (1984), so long as the same biased (rd) basis was employed

in forward application of this correlation to field engineering

works. It was a slight problem, however, when forward

applications involved direct, response-based calculation of in-

situ CSR, as often occurs on analyses of major dams, etc.

It was Seed’s intent that the recommended (1984) boundary

should represent approximately a 10 to 15% probability of

liquefaction, and with allowance for the “shift” in (improved)

evaluation of CSR, the 1984 deterministic relationship for

clean sands (<5% fines) does correspond to approximately PL

≈ 10 to 30%, except at very high CSR (CSR > 0.3), a range in

which data were previously scarce.

Also shown in Figure 10 is the boundary curve proposed by

Yoshimi et al. (1994), based on high quality cyclic testing of

frozen samples of alluvial sandy soils. The line of Yoshimi et

al. is arguably unconservatively biased at very low densities

(low N-values) as these loose samples densified during

laboratory thawing and reconsolidation. Their testing provides

potentially valuable insight, however, at high N-values where

reconsolidation densification was not significant. In this range,

the new proposed correlation provides slightly better

agreement with the test data than does the earlier relationship

proposed by Seed et al. (1984). Improvement of the new

correlation at high CSR values is due, in large part, to the

availability of significant new data (at high CSR) from recent

earthquakes that had not been available in 1984.

The new correlation is also presented in Figure 6(d), where it

can be compared directly with the earlier probabilistic

relationships of Figures 6(a) through (c). Here, again, the new

correlation is normalized to σ’v = 0.65 atm. in order to be fully

compatible with the basis of the other relationships shown. As

shown in this figure, the new correlation provides a

tremendous reduction in overall uncertainty (or variance).

3.1.3 Adjustments for Fines Content:

The new (probabilistic) boundary curve for PL = 15% (again

normalized to an effective overburden stress of σ’v = 0.65

atm.) represents a suitable basis for illustration of the new

correlation’s regressed correction for the effects of fines

content, as shown in Figure 11. In this figure, both the

correlation as well as the mean values (CSR and N1,60) of the

field case history data are shown not corrected for fines (this

time the N-value axis is not corrected for fines content effects,

so that the (PL=20%) boundary curves are, instead, offset to

account for varying fines content.) In this figure, the earlier

correlation proposed by Seed et al. (1984) is also shown (with

dashed lines) for approximate comparison.

In these current studies, based on the overall (regressed)

correlation, the energy- and procedure- and overburden-

corrected N-values (N1,60) are further corrected for fines

content as

N1,60,CS = N1,60 * CFINES (Eq. 6)

where the fines correction was “regressed” as a part of the

Bayesian updating analyses. The fines correction is equal to

1.0 for fines contents of FC < 5%, and reaches a maximum

(limiting) value for FC > 35%. As illustrated in Figure 11,

Seed et al. (2003) 17

Fig. 10: Recommended Probabilistic SPT-Based Liquefaction Triggering

Correlation (for M

w=7.5 and σσv′′ =0.65 atm), and the Relationship

for “Clean Sands” Proposed by Seed et al. (1984)

Fig. 11: Recommended “Deterministic” SPT-Based Liquefaction Triggering

Correlation (for M

w=7.5 and σσv′′ =0.65 atm), with Adjustments for

Fines Content Shown

N1,60

0 10 20 30 40

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

MW=7.5

σV'=0.65 atm

Seed et al. (1984)

LiquefiedMarginalNon-liquefied

“Old”Data(Pre-1985)

“New”Data

FC5%

FC15%

FC35%

35%

15%

N1,60,CS

0 10 20 30 40

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

MW=7.5 σV'=0.65 atm

80% 20%

95% 50% 5%

Seed et al. (1984)

Yoshimi et al. (1994)

LiquefiedMarginalNon-

liquefied

Pre-1 985 Data

“New”Data

Seed et al. (2003) 18

the maximum fines correction results in an increase of N-

values of about +6 blows/ft. (at FC > 35%, and high CSR). As

illustrated in this figure, this maximum fines correction is

somewhat smaller than the earlier maximum correction of

+9.5 blows/ft proposed by Seed et al. (1984).

The regressed relationship for CFINES is

( )













⋅+⋅+= 60,1

05.0004.01 N

FCCFINES (Eq. 7)

lim: FC ≥ 5% and FC  35%

where FC = percent fines content (percent by dry weight finer

than 0.074mm), expressed as an integer (e.g. 15% fines is

expressed as 15), and N1,60 is in units of blows/ft.

Magnitude-Correlated Duration Weighting:

Both the probabilistic and “deterministic” (based on PL=20%)

new correlations presented in Figures 10 and 11 are based on

the correction of “equivalent uniform cyclic stress ratio”

(CSReq) for duration (or number of equivalent cycles) to

CSRN, representing the equivalent CSR for a duration typical

of an “average” event of MW = 7.5. This was done by means

of a magnitude-correlated duration weighting factor (DWFM)

CSRN = CSReq,M=7.5 = CSReq, M=M / DWFM (Eq. 8)

This duration weighting factor has been somewhat

controversial, and has been developed by a variety of different

approaches (using cyclic laboratory testing and/or field case

history data) by a number of investigators. Figure 12

summarizes a number of recommendations, and shows

(shaded zone) the recommendations of the NCEER Working

Group (NCEER, 1997). In these current studies, this

important and controversial factor could be regressed as a part

of the Bayesian Updating analyses. Moreover, the factor

(DWFM) could also be investigated for possible dependence

on density (correlation with N

1,60). Figures 13 shows the

resulting values of DWFM, as a function of varying corrected

N1,60-values. As shown in Figure 13, the dependence on

density, or N1,60-values, was found to be relatively minor.

The duration weighting factors shown in Figures 12 and 13

fall slightly below those recommended by the NCEER

Working group, and very close to recent recommendations of

Idriss (2000). Idriss’ recommendations are based on a

judgmental combination of interpretation of high-quality

cyclic simple shear laboratory test data and empirical

assessment of “equivalent” numbers of cycles from recorded

strong motion time histories, and are the only other values

shown that account for the cross-correlation of r

d with

magnitude. The close agreement of this very different (and

principally laboratory data based) approach, and the careful

(field data based) probabilistic assessments of these current

studies, are strongly mutually supportive.

4.5 5.5 6.5 7.5 8.5

4.5

3.5

2.5

1.5

0.5

Seed and Idriss (1982)

Idriss (2000)

Ambraseys (1988)

Arango (1996)

Andrusand Stoke(1997)

Youd and Noble(1997)PL<32

Youd and Noble(1997)PL>50

Range ofrecommended DWFM

fromNCEERWorkshop (1997)

This

Study

Fig. 12: Previous Recommendations for Magnitude-

Correlated Duration Weighting Factor, with

Recommendations from Current Studies

4.5 5.5 6.5 7.5 8.5

3.5

2.5

1.5

0.5

N1,60=40

Fig. 13: Recommended Magnitude-Correlated Duration

Weighting Factor as a Function of N1,60

Seed et al. (2003) 19

3.1.5 Adjustments for Effective Overburden Stress:

An additional factor not directly resolved in prior studies

based on field case histories is the increased susceptibility of

soils to cyclic liquefaction, at the same CSR, with increases in

effective overburden stress. This is in addition to the

normalization of N-values for overburden effects as per

Equation 4.

The additional effects of reduction of normalized liquefaction

resistance with increased effective initial overburden stress

(σ’v) has been demonstrated by means of laboratory testing,

and this is a manifestation of “critical state” type of behavior

(soils become less dilatant at increased effective stress).

Figure 14 shows the recommendations of the NCEER

Working Group (Youd et al., 2001) regarding the correction

factor Kσ to be used to correct to the normalized resistance to

liquefaction at an initial effective overburden stress of 1 atm.

(CSRliq,1atm) as

CSRliq = CSRliq,1atm. Kσ (Eq. 9)

These current studies were not very sensitive to Kσ, as the

range of σ’v in the case history data base was largely between

σ’v = 600 to 2,600 lb/ft2, but it was possible to “regress” Kσ as

part of the Bayesian updating. The results are shown in Figure

15, over the range of σ’v  600 to 3,600 lb/ft2 for which they

are considered valid. These are in good agreement with the

earlier recommendations of Figure 14, and it is recommended

that Kσ can be estimated as

Kσ = ( v

′

)f-1 (Eq. 10)

where f ≈ 0.6 to 0.8 (as N1,60,cs varies from 1 to 40 blows/ft.)

The field case history data of these current studies are not a

sufficient basis for extrapolation of Kσ to much higher values

of σ’v, and the authors recommend use of Figure 14 for σ’v >

2 atm.

The earlier relationships proposed by Seed et al. (1984), Liao

et al. (1988, 1998), Youd and Noble (1997), and Toprak et al.

(1999) were all stated to be normalized to an effective

overburden stress of approximately σ’v = 1 atm (2,000 lb/ft2).

The correlation of Seed et al. (1984) was never formally

corrected to σ’v = 1 atm., however, as it was noted that the

field case histories of the database were “shallow”, and

approximately in this range. The database was, however, not

centered at σ’v = 1atm., but rather at lesser overburden (Mean

σ’v 1,300 lb/ft2or 0.65 atm), and this proves to render this

earlier relationship slightly unconservative if taken as

normalized to σ’v = 1 atm. (The same is true of all of the

previous relationships discussed.) It should be noted,

however, that this unconservatism is minimized if the

correlations are applied at shallow depths.

(

)

(

)

( ) ( )

























+⋅+′

⋅−⋅−⋅−⋅+⋅

−Φ=

′70.2

97.4405.0

ln70.3ln53.29

ln32.13004.01

),,,,(

60,1

CSRFCN

FCMCSRNP

vwL

σ (Eq. 11)

where

PL = the probability of liquefaction in decimals (i.e. 0.3, 0.4, etc.), and

Φ = the standard cumulative normal distribution.

___________________________________________________________________________________________________

Also the cyclic resistance ratio, CRR, for a given probability of liquefaction can be expressed as:

(

)

(

)

( ) ( )























Φ⋅++⋅+

′

⋅−

⋅−⋅+⋅

′−

32.13

70.297.4405.0ln70.3

ln53.29004.01

exp),,,,,(1

60,1

60,1Lv

LvwPFC

MFCN

PFCMCSRNCRR σ

σ (Eq. 12)

where

Φ-1(PL) = the inverse of the standard cumulative normal distribution (i.e. mean=0, and standard deviation=1)

note: for spreadsheet purposes, the command in Microsoft Excel for this specific function is “NORMINV(PL,0,1)”

Seed et al. (2003) 20

For correctness, and to avoid ambiguity, both the earlier

relationship of Seed et al. (1984), and the correlations

developed in these current studies, need to be formally

normalized to σ’v = 1 atm. Accordingly, in these present

studies, all data are corrected for K σ-effects (by Equations 9

and 10); not just those data for which σ’v was greater than 1

atm. A recommended limit is K

σ < 1.5 (at very shallow

depths.) Figures 16 and 17 show the proposed new

correlations, this time for σ’v =1 atm, and these figures

represent the final, fully normalized recommended

correlations.

The overall correlation can be expressed in parts, as in the

previous sections (and Equations 6 - 12, and Figures 7 - 17).

It can also be expressed concisely as a single, composite

relationship as shown in Equation 11.

Recommended Use of the New SPT-Based Correlations:

The proposed new probabilistic correlations can be used in

either of two ways. They can be used directly, all at once, as

summarized in Equations 11 and 12. Alternatively, they can

be used “in parts” as has been conventional for most previous,

similar methods. To do this, measured N-values must be

corrected to N

1,60-values, using Equations 3, 4 and 5. The

resulting N1,60-values must then be further corrected for fines

content to N1,60,cs-values, using Equations 6 and 7 (or Figure

17). Similarly, in-situ equivalent uniform cyclic stress ratio

(CSReq) must be evaluated, and this must then be adjusted by

the magnitude-correlated Duration Weighting Factor (DWFM)

using Equation 8 (and Figure 13) as

CSRN = CSReq,M=7.5 = CSReq / DWFM (Eq. 13)

The new CSReq,M=7.5 must then be further adjusted for

effective overburden stress by the inverse of Equation 9, as

CSR* = CSReq,M=7.5,1atm = CSReq,M=7.5 / Kσ (Eq 14)

The resulting, fully adjusted and normalized values of N1,60,cs

and CSReq,M=7.5,1atm can then be used, with Figure 16, to assess

probability of initiation of liquefaction.

For “deterministic” evaluation of liquefaction resistance,

largely compatible with the intent of the earlier relationship

proposed by Seed et al. (1984), the same steps can be

undertaken (except for the fines adjustment) to assess the fully

adjusted and normalized CSReq,M=7.5,1atm values, and

normalized N

1,60 values, and these can then be used in

conjunction with the recommended “deterministic”

relationship presented in Figure 17. The recommendations of

Figure 17 correspond to the new probabilistic relationships

(for P

L = 15%), except at very high CSR (CSR > 0.4). At

these very high CSR; (a) there is virtually no conclusive field

data, and (b) the very dense soils (N1,60 > 30 blows/ft) of the

boundary region are strongly dilatant and have only very

limited post-liquefaction strain potential. Behavior in this

region is thus not conducive to large liquefaction-related

VerticalEffectiveStress,σv’(atmunits)

Kσ’

Dr40%(f=0.8)

Dr~60%(f=0.7)

Dr80%(f=0.6)

Kσ’=(σv’)f-1

Fig. 14: Recommended Kσσ Values for σσ’v >> 2 atm.

200

600

1000

1400

1800

2200

2600

3000

3400

3800

4200

Number of Case Histories

0 1000 2000 3000 4000

1.2

1.0

0.8

0.6

0.4

Kσ

σv’(psf)

ThisStudy

Recommended byNCEER

Working Group (1998)

Fig. 15: Values of Kσσ Developed and Used in These

Studies, NCEER Working Group

Recommendations (for n=0.7, DR ≈≈ 60%)

for Comparison

Seed et al. (2003) 21

Fig. 16: Recommneded “Probabilistic” SPT-Based Liquefaction Triggering

Correlation (For MW=7.5 and σσv’=1.0 atm)

Fig. 17: Recommneded “Deterministic” SPT-Based Liquefaction Triggering

Correlation (For MW=7.5 and σσv’=1.0 atm) with Adjustments for

Fines Content Shown

N1,60,CS

010 20 30 40

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

MW=7.5 σV'=1.0 atm

80% 20%

95% 50% 5%

LiquefiedMarginalNo n-

liquefied

Pre-1985 Data

“New”Data

N1,60

0 10 20 30 40

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

MW=7.5 σV'=1.0 atm

LiquefiedMarginalNon-liquefied

“Old”Data(Pre-1985)

“New”Data

FC5%

FC15%

FC35%

Seed et al. (2001) 22

displacements, and the heavy dashed lines shown in the upper

portion of Figure 17 represent the authors’ recommendations

in this region based on data available at this time.

3.1.7 Summary

This section of this paper has presented the development of

recommended new probabilistic and “deterministic”

relationships for assessment of likelihood of initiation of

liquefaction. Stochastic models for assessment of seismic soil

liquefaction initiation risk have been developed within a

Bayesian framework. In the course of developing the

proposed stochastic models, the relevant uncertainties

including: (a) measurement/estimation errors, (b) model

imperfection, (c) statistical uncertainty, and (d) those arising

from inherent variability were addressed.

The resulting models provide a significantly improved basis

for engineering assessment of the likelihood of liquefaction

initiation, relative to previously available models, as shown in

Figure 5(d). The new models presented and described in this

paper deal explicitly with the issues of (1) fines content (FC),

(2) magnitude-correlated duration weighting factors (DWFM),

and (3) effective overburden stress (Kσ effects), and they

provide both (1) an unbiased basis for evaluation of

liquefaction initiation hazard, and (2) significantly reduced

overall model uncertainty. Indeed, model uncertainty is now

reduced sufficiently that overall uncertainty in application of

these new correlations to field problems is now driven

strongly by the difficulties/uncertainties associated with

project-specific engineering assessment of the necessary

“loading” and “resistance” variables, rather than uncertainty

associated with the correlations themselves. This, in turn,

allows/encourages the devotion of attention and resources to

improved evaluation of these project-specific parameters. As

illustrated in Figures 6(d), 16 and 17, this represents a

significant overall improvement in our ability to accurately

and reliably assess liquefaction hazard.

The new correlations also eliminate a bias intrinsic in all prior,

similar relationships when using actual dynamic response

analyses to directly calculate in-situ CSR, as all prior

relationships were based on an unconservatively biased

“simplified” (rd-based) assessment of CSR. This was not a

major problem when using these previous correlations in

conjunction with the same r

d for “forward” engineering

analyses, but it was a problem when using prior correlations in

conjunction with direct calculation of in-situ CSR. The new

correlations are unbiased in this regard, and can be used either

in conjunction with “simplified” CSR assessments (based on

the new r

d recommendations presented herein), or in

conjunction with direct dynamic response analyses for

calculation of in-situ CSR. The new correlations cannot,

however, be used in conjunction with assessment of CSR

based on the “old” (Seed and Idriss, 1971) rd relationship.

3.2 CPT-Based Triggering Correlations:

3.2.1 Introduction

In addition to SPT, three other in-situ index tests are now

sufficiently advanced as to represent suitable bases for

correlation with soil liquefaction triggering potential, and

these are (a) the cone penetration test (CPT), (b) in-situ shear

wave velocity measurement (VS), and (c) the Becker

Penetration Test (BPT).

Up to this point in time, the SPT-based correlations have been

better defined, and have provided lesser levels of uncertainty,

than these other three methods. CPT, however, is approaching

near parity, and newly developed CPT-based correlations now

represent nearly co-equal status with regard to accuracy and

reliability relative to SPT-based correlations.

CPT-based correlations have, to date, been based on much less

numerous and less well defined earthquake field case histories

than SPT-based correlations. This is changing, however, as a

number of research teams are working on development of

improved CPT-based “triggering” correlations. This includes

the authors of this paper, and the next sections will present a

much-improved basis for CPT-based assessment of

liquefaction initiation (or “triggering”) potential.

It is important to develop high quality CPT-based correlations

to complement and augment the new SPT-based correlations

presented herein. The authors are often asked whether SPT or

CPT is intrinsically a better test for liquefaction potential

evaluation. The best answer is that both tests are far better

when used together, as each offers significant advantages not

available with the other.

SPT-based correlations are currently ahead of “existing” CPT-

based correlations, due in large part to enhanced data bases

and better data processing and correlation development. The

new SPT-based correlations described in this paper are

currently more accurate and reliable, and provide much lower

levels of uncertainty or variance. An additional very

significant advantage of SPT is that a sample is retrieved with

each test, and so can be examined and evaluated to ascertain

with certainty the character (gradation, fines content, PI, etc.)

of the soils tested, as contrasted with CPT where soil character

must be “inferred” based on cone tip and sleeve friction

resistance data.

The CPT offers advantages with regard to cost and efficiency

(as no borehole is required). A second advantage is

consistency, as variability between equipment and operators is

small (in contrast to SPT). The most important advantage of

CPT, however, is continuity of data over depth. SPT can only

be performed in 18-inch increments, and it is necessary to

advance and clean out the borehole between tests.

Accordingly, SPT can only be performed at vertical spacings

of about 30 inches (75cm) or more. As a result, SPT can

completely miss thin (but potentially important) liquefiable

strata between test depths. Similarly, with a 12-inch test

Seed et al. (2001) 23

height and allowance for effects of softer overlying and

underlying strata, SPT can fail to suitably characterize strata

less than about 3 to 4 feet in thickness.

CPT, in contrast, is fully continuous and so “misses” nothing.

The need to penetrate about 4 to 5 diameters into a stratum to

develop full tip resistance, to be at least 4 to 5 diameters from

an underlying softer stratum, and the “drag length” of the

following sleeve, cause the CPT test to poorly characterize

strata of less than about 12 to 15 inches (30 to 40cm) in

thickness, but this allows for good characterization of much

thinner strata than SPT. Even for strata too thin to be

adequately (quantifiably) characterized, the CPT at least

provides some indications of potentially problematic materials

if one examines the qc and fs traces carefully.

3.2.2 Existing CPT-Based Correlations

Owing to its attractive form and simplicity, as well as its

endorsement by the NCEER Working Group, the CPT-based

correlation of Robertson and Wride (1998) is increasingly

used for liquefaction studies. This correlation is well

described in the NCEER summary papers (NCEER, 1997;

Youd, et al., 2001).

Robertson and Wride had access to a much smaller field case

history database than is currently available, and so their

correlation represents a valuable interim contribution as

development of new correlations taking advantage of the

wealth of new earthquake field case history data now available

now proceeds.

Figure 18 shows the “baseline” triggering curve of Robertson

and Wride for “clean” sandy soils. Adjustments for fines are

based on combinations of sleeve friction ratios and tip

resistances in such a manner that the “clean sand” boundary

curve of Figure 18 is adjusted based on a composite parameter

IC. IC is a measure of the distance (the radius) from a point

above and to the left of the plot of normalized tip resistance

(qc,1) and normalized Friction Ratio (F) as indicated in Figure

19. Tip resistance is corrected for increasing fines content and

plasticity as

qc,1,mod = qc,1 · KC (Eq. 15)

The recommended “fines” correction is a nonlinear function of

IC, and ranges from a multiplicative factor of KC = 1.0 at IC =

1.64, to a maximum value of KC = 3.5 at IC = 2.60. A further

recommendation on the fines correction factor is that this

factor be set at KC = 1.0 in the shaded zone within Area “A” of

Figure 19 (within which 1.64 < IC < 2.36, and friction ratio F <

0.5).

Based on cross-comparison with the new SPT-based

correlation, it appears that the CPT-based correlation of

Robertson and Wride is slightly unconservative for clean

sands, especially at high CSR, and that it is very

unconservative for soils of increasing fines content and

Fig. 18: CPT-Based Liquefaction Triggering

Correlation for “Clean” Sands Proposed

by Robertson and Wride (1998)

Zone A: Cyclic liquefaction possible – depends on size and duration of cyclic loading

Zone B: Liquefaction unlikely – check other criteria

Zone C: Flow/cyclic liquefaction possible –

depends on soil plasticity and sensitivity as well

as size and duration of cyclic loading.

Fig. 19: Fines Correction as Proposed by Robertson and

Wride (1998)

KC=1.0

KC=3.5

Seed et al. (2003) 24

plasticity. This, as it turns out, is verified by comparison with

the new CPT-based correlations presented and described in the

section that follows.

Additional researchers have been and are continuing to

develop CPT-based correlations, but rather than discuss all of

these we will, instead, present a recommended new CPT-

based correlation with many of the attributes and strengths of

the new SPT-based correlation presented previously.

3.3 Recommended New CPT-Based Triggering Correlation:

3.3.1 Introduction

The approach followed in development of the new CPT-based

correlation presented herein was similar in many ways to that

followed in development of the SPT-based correlation

presented previously. As a result, the new CPT-based

relationship shares many of the same strengths.

Key elements in the development of this new correlation were:

(1) accumulation of a significantly expanded database of field

performance case histories, (2) use of improved knowledge

and understanding of factors affecting interpretation of CPT

data, (3) incorporation of improved understanding of factors

affecting site-specific ground motions (including directivity

effects, site-specific response, etc.), (4) use of improved

methods for assessment of in-situ cyclic shear stress ratio

(CSR), (5) screening of field data case histories on a

quality/uncertainty basis, and (6) use of higher-order

probabilistic tools (Bayesian Updating). Once again, detailed

review of the processing and back-analyses of the field

performance case histories by a group of leading experts, and

establishment of concensus (or at least near-concensus)

regarding all resulting critical parameters and variables, is a

key feature of this effort.

These new correlations are not yet quite complete, as iterative

review of some of the case history interpretations is still

underway. The correlations are far enough along that they are

nearly final, however, and as they already incorporate far more

data (and of higher overall quality) than previous correlations,

they represent a significant advance. The resulting

relationships not only provide greatly reduced levels of

uncertainty, they also help to resolve a number of corollary

issues that have long been difficult and controversial,

including: (1) adjustments for fines content, and (2)

corrections for effective overburden stress.

3.3.2 Improved Treatment of Normalization of CPT Tip and

Sleeve Resistance for Effective Overburden Stress

In development of optimally improved CPT-based

correlations, it was desirable to go after each of the issues that

have historically contributed to the uncertainty (or variance) of

previous correlations. One particularly significant issue was

the approach used to normalize CPT tip resistance (qc) and

sleeve resistance (fs) for effective overburden stress effects.

Approaches have differed significantly here. Olsen and

Mitchell (1995) presented the most comprehensive set of

recommendations in this regard, and their recommendations

(along with their recommended approximate soil classification

scheme) are presented in Figure 20. This figure’s axes

(normalized CPT tip resistance qc,1 on the vertical axis, and

sleeve friction ratio R

f on the horizontal axis) will provide a

useful template for much of the rest of this section. [Friction

Ratio is taken as Rf = fs/qc · 100.]

In these current studies, a suite of four different cavity

expansion models, each used for the soil type and density (DR)

or overconsolidation ratio (OCR) range for which it is best

suited, were used to study variation of CPT tip resistances

with changes in effective overburden stress (óv'). The model

of Salgado & Randolph (2001) was used for dense (dilatant)

cohesionless soils. The model of Boulanger (2003) was used

for very high overburden stress conditions for the same dense

(dilatant) cohesionless soils. The model of Yu (2000) with

Ladanyi and Johnston (1974) was used for loose to medium

dense cohesionless soils. The model of Cao et al. (2001) was

used for overconsolidated cohesive (clayey) soils and the

model of Yu (2000) was used for normally consolidated

cohesive (clayey) soils. Each of these models was both

constrained and calibrated using significant bodies of

laboratory calibration chamber test data. The results of these

laboratory and analytically based methods were then

augmented using actual field data regarding variation of tip

resistance vs. effective overburden stress. Details of all

analyses, as well as field data summaries, will be presented in

Moss (2003). The combined data was then judgmentally

interpreted, and used to develop recommendations for

normalization of CPT tip resistance to develop normalized qc,1

values as

cqcqCq

⋅

1, where

C











='σ (Eq. 16)

The normalization exponent (c) is a function of both

normalized tip resistance and friction ratio (Rf) as shown in

Figure 21. Also shown, for purposes of comparison, are the

earlier recommendations of Olsen and Mitchell (1995).

Cavity expansion models are not able to provide insight

regarding similar normalization of sleeve friction (fs) for

effective overburden stress effects, so a more approximate

assessment was made, based largely on laboratory calibration

chamber test data and field data, to develop similar corrections

for sleeve resistance as,

sfsfCf

⋅

1, where

C











='σ (Eq. 17)

Seed et al. (2003) 25

The recommended normalization exponent (s) is shown, as a

function of normalized tip resistance and friction ratio, in

Figure 22, along with the recommended tip normalization

exponents (c) from Figure 21.

Figure 22 thus shows the recommended normalization for both

tip and sleeve resistances. These are not identical to each

other, but it should be noted that they appear to be sufficiently

similar that “normalized” friction ratio [Rf,1 = (fs,1 / qc,1)*100]

is very similar to non-normalized friction ratio [Rf = (fs /

qc)*100]. Limited iteration is necessary to make the

recommended adjustment of qc to qc,1, because Rf and Rf,1 vary

only slightly.

3.3.3 Thin Layer Corrections

A second source of potential uncertainty is the adjustment of

measured CPT tip resistances for finite stiff layers. The

effects of initial penetration into a stronger (e.g., less cohesive,

potentially liquefiable) layer prior to achieving sufficient

penetration into the layer to develop a “fully developed” tip

resistance can result in a reduced tip resistance reading, with a

similar reduction occurring as the cone approaches and exits

the bottom of a stronger layer (“sensing” the approach of the

softer underlying layer before actually reaching it).

Several approaches have been proposed for adjustment of

measured tip resistances in “thin” layers (e.g.; Robertson and

Fig. 20: Recommended CPT Tip Normalization

Exponents, and Approximate Soil

Characterization Framework

(After Olsen & Mitchell, 1995)

0.1

100

0.1 1

qc,1 (MPa)

Rf (%)

1.00 0.75 0.55

0.35

Olsen & Mitchell (1995)

Proposed Tip Exponent

Fig. 21: Recommended CPT Tip Normalization

Exponents, and Previous Recommendations

of Olsen & Mitchell (1995)

0.1

100

0.1 110

qc,1 (MPa)

Rf (%)

1.00

0.75

0.55

0.35

Tip Exponent, c

Sleeve Exponent, s

cqcqCq⋅=1,

sfsfCf⋅=1,

C











='σ

C











='σ

Fig 22: Recommended CPT Tip and Sleeve

Normalization Exponents

Seed et al. (2003) 26

Wride, 1997; Youd et al., 2001). In this current effort, the

elastic solution for “thin layer effects” proposed by

Vreugdenhil et al. (1995) was calibrated against both available

laboratory calibration chamber test data as well as field data

(Moss, 2003), and the resulting recommended correction of

CPT tip resistances for thin layer effects is presented in Figure

23. Also shown, for comparison, are the earlier

recommendations of the NCEER Working Group (Youd et al.,

2001). The new recommendations are largely compatible with

the NCEER Working Group’s recommendations, but serve to

extend the approach to higher contrasts in measured tip

resistances between the (stiffer) “thin” layer and the (softer)

overlying and underlying layers. In this procedure, the ratio of

the final, corrected CPT tip resistance in the “thin” layer (qcA),

relative to the average tip resistance of the softer overlying

and underlying layers (qcB) serves as a proxy for the ratio of

the stiffnesses of these layers. It should be noted that field

cases with ratios of q

cB/qcA of greater than about 5 are

relatively uncommon.

The correction of CPT tip resistances for “thin layer effects” is

then

qcB,corrected = qcb,thin = Cthin · qcB (Eq. 18)

where Cthin is as shown in Figure 23.

Given the intrinsic uncertainty involved in this type of “thin

layer” adjustment, it is recommended that adjustment factors

(Cthin) of greater than 1.8 not be used for engineering

applications. A more severe limit was employed in back

analyses of field case histories for liquefaction correlation

development. Only a small number of cases incorporated in

development of the final correlations required any thin layer

adjustments, and none required an adjustment factor of greater

than 1.5.

3.3.4 Field Performance Case Histories

A total of more than 600 field performance case histories were

acquired and evaluated as a part of these studies. Some of

these had been used by previous researchers in similar efforts,

but many are new. This is, to date, the largest set of CPT-

based field cases assessed for purposes of development of

liquefaction triggering hazard correlations. The cases

considered were from the 1964 Niigata (Japan), 1968

Inangahua (New Zealand), 1975 Haicheng (China), 1976

Tangshan (China), 1977 Vrancea (Romania), 1978 Izu-

Oshima-Kinkai (Japan), 1979 Imperial Valley (USA), 1980

Mexicali (Mexico), 1981 Westmorland (USA), 1983 Borah

Peak (USA), 1983 Nihonkai-Chubu (Japan), 1987 Edgecumbe

(New Zealand), 1989 Loma Prieta (USA), 1994 Northridge

(USA), 1995 Hyogoken-Nambu (Japan), 1995 Dinar (Turkey),

1999 Kocaeli (Turkey), 1999 Duzce (Turkey), and 1999 Chi-

Chi (Taiwan) Earthquakes.

Length constraints do not permit a full treatment of the details

involved in back-analyses and processing of these case

histories, but evaluation methods used were similar to those

employed by Cetin (2000) and Seed et al. (2003) for SPT-

based studies. At each site, only the single most critical

stratum was considered. Cyclic stress ratios (CSR) were

evaluated using the recently proposed rd relationships by Cetin

and Seed (2000), except in cases where site-specific site

response analyses could be performed. These new r

relationships represent an improved basis for estimation of

CSR, and are statistically unbiased with respect to site-specific

response analyses. It should be noted that the earlier r

recommendations of Seed and Idriss (1971) cannot be used

with the new correlations proposed herein, as these earlier rd

recommendations provide generally higher estimates of CSR

than do site-specific response analyses (or the new r

recommendations of Cetin et al.) for strong levels of shaking,

and are thus not compatible with the new correlations

proposed herein.

For each field case history, the variances or uncertainties in

both the critical loading and in-situ soil and index parameters

were evaluated, and the cases were systematically rated for

overall quality on the basis of uncertainty of key parameters.

Only the most highly rated cases were then used for

development of the new correlations; cases of lesser quality

(with unacceptably high uncertainty or poorly defined

parameters) were deleted from further consideration. At this

time, a total of 201 cases were selected for incorporation in the

correlations presented herein. This is the largest number of

high quality cases (based on unusually high screening

standards) used to date in development of these types of CPT-

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.1

2.2

0500 1000 1500 2000 2500

3000

qcB/qcA =10

qcB/qcA =5

qcB/qcA =2

NCEER Recommendations (qcB/qcA=2)

Recommended Limit

Layer Thickness, h (mm)

Thin Layer Correction Factor, Cthin

cBthinthincB, qCq⋅=

B(d=35.7mm, A=10cm

2 Cone)

0 500 1000 1500 2000 2500

3000

2.2

2.1

2.0

1.9

1.8

1.7

1.6

1.5

1.4

1.3

1.2

1.1

1.0

Fig 23: Recommended Thin Layer Correction for CPT

Tip Resistance, and Earlier NCEER Working

Group Recommendations

Seed et al. (2003) 27

qC,1,mod

0 5 10 15 20

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

MW=7.5 σV'=1.0 atm

80% 20%

95% 50% 5%

Fig. 24: Contours of 5%, 20%, 50%, 80% and 95% Probability of

Liquefaction as a Function of Equivalent Uniform Cyclic

Stress Ratio and Fines-Modified CPT Tip Resistance (Mw =

7.5, óv' = 1atm.)

Fig. 25: Contours of 5%, 20%, 50%, 80% and 95% Probability of

Liquefaction for ó

v' = 1 atm. and Duration Associated with

Mw = 7.5 As a Function of Fines Corrected SPT Penetration

Resistance (Seed et al., 2001)

N1,60,CS

010 20 30 40

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

MW=7.5 σV'=1.0 atm

80% 20%

95% 50% 5%

LiquefiedMarginalNon-

liquefied

Pre-1985 Data

“New”Data

Seed et al. (2003) 28

based correlations. As with the largely parallel SPT-based

efforts, the back-analyses, processing, and selection of these

cases was subjected to iterative review by an accomplished

group of international experts with excellent prior experience

in this area, in order to establish concensus evaluations of

critical parameters.

3.3.5 Correlation Development

Length constraints again do not permit a full discussion of all

details involved in development of the new recommended

CPT-based probabilistic soil liquefaction triggering

correlations. The details are largely parallel to those described

by Cetin (2000) and Seed et al. (2001) in development of SPT-

based correlations. A Bayesian updating methodology was

employed to develop probabilistic correlations. This is

essentially a high order probabilistic regression method well

suited to this problem, and capable of dealing with the various

types of contributing sources of aleatory and epistemic

uncertainty involved. A discussion of development of a

(similar) Bayesian treatment of the SPT-based correlations

was presented by Cetin et al. (2002).

Figure 24 presents one view of the new recommended

correlation, in this case a plot of contours of probability of

liquefaction (for P

L = 5%, 20%, 50%, 80% and 95%) as a

function of equivalent uniform cyclic stress ratio (CSReq) and

modified normalized CPT tip resistance (qc,1,mod). In this

figure, equivalent uniform CSR has been corrected for

duration effects based on the magnitude-correlated duration

weighting factor (DWFM) proposed in Seed et al. (2001) for

SPT-based correlations, as the regression for this CPT-based

correlation resulted in essentially equivalent magnitude-

correlated duration scaling. This results in an expression of

the correlation appropriate for events of M

w  7.5. The

correlation in Figure 24 is also normalized to an effective

overburden stress of óv' = 1atmosphere.

In Figure 24, the solid dots represent the centroids of

probabilistic distributions of the individual case histories for

cases wherein liquefaction was judged to have been

“triggered”, and open circles represent centroids of

distributions of field cases wherein liquefaction did not occur.

These distributions quantify each individual field case history

and its distributed variance in both the horizontal and vertical

axes. The CPT tip resistances of Figure 24 are adjusted for

effects of fines (fines content and plasticity) to values of

qc,1,mod as described subsequently.

Figure 25, presented adjacent to Figure 24 for comparison

purposes, represents a similar view of the corollary new

recommended SPT-based relationship, also with contours for

probabilities of liquefaction of PL = 5%, 20%, 50%, 80% and

95%, also normalized for M

w = 7.5, = 1atm., and with N-

values corrected for fines content to N1,60,cs.

The horizontal axis of Figure 24 represents modification of

normalized CPT tip resistances (qc,1 values) for the frictional

effects of apparent fines content and character. Values of qc,1

are adjusted as

qc,1,mod = qc,1 + Äqc (Eq. 19)

where Äqc is a function of qc,1, Rf, and c, as shown in Figure

26. Figure 27 repeats the recommended fines adjustment of

Figure 26, and also shows for comparison the “fines

correction” factors recommended by Robertson and Wride

(1997). Robertson and Wride recommended adjustment by a

multiplicative factor (KC) as presented previously in Equation

15, where KC is a function of both tip resistance and friction

ratio as shown in Figure 27. Robertson and Wride also

recommended, however, that K

C be taken as 1.0 (a null

adjustment of qc,1 ) in the shaded region of Figure 27. It is

interesting to note that the new recommended correction

contours for fines content and character proposed herein also

reflect a null adjustment in this shaded zone, yet provide a

smoother variation of the adjustment (Äqc) as it transitions to

other areas of Figure 27. (The Robertson and Wride

recommendations jump very sharply at the base and right edge

of the shaded “null correction” zone.) The new contours also

provide for much smaller overall adjustments of qc,1 for fines

content and character than did the earlier curves proposed by

Robertson and Wride, suggesting that these earlier

recommendations were unconservative at high fines contents.

Figure 28 presents an alternate view of the new correlation, in

this case contours of 15% probability of liquefaction

triggering, but for three different values of Äqc spanning the

full available range of Äqc. P

L = 15% represents the new

recommended “deterministic” threshold, analogous to the

“deterministic” recommendations of most prior relationships.

The adjacent Figure 29 similarly presents a view of the new

SPT-based correlation, with 15% probability of liquefaction

contours shown for three different levels of percent fines

(fines content) spanning the full available range of fines

corrections for the SPT-based relationship. The similarity

between the relationships shown in Figures 28 and 29 is both

interesting and important, as it represents strong mutual

support between the new proposed SPT- and CPT-based

correlations.

Finally, Figure 30 provides a comparison between the new

proposed CPT-based correlation, and the previous correlations

proposed by Robertson and Wride (1997) and Suzuki et al.

(1995). The new correlation is represented in this figure by

contours of 15% probability of liquefaction, as that is the level

of probability recommended by the authors herein for use as a

reasonable “deterministic” threshold. The correlations of

Robertson and Wride, and Suzuki et al., are “deterministic”

correlations, as they do not explicitly address probability (or

uncertainty).

As shown in Figure 30, the “clean sand” (Äqc = 0) line for the

new correlation falls between the similarly based “clean sand”

(Rf < 0.5%) line proposed by Suzuki et al., and the “clean

sand” (KC = 1.0) line proposed by Robertson and Wride. The

range of fines-corrected lines for the new correlation,

Seed et al. (2003) 29

Fig 26: Recommended CPT Tip Resistance Modification for “Fines Content

and Character” as a Function of qc,1 and Rf

Fig 27: Recommended CPT Tip Resistance Modification for “Fines Content

and Character” as a Function of qc,1 and Rf, Compared with the

Earlier Recommendations of Robertson and Wride (1997)

0.1

100

0.1 1 10

Rf (%)

qc,1 (Mpa)

0.1

100

0.1 1 10

Rf (%)

qc,1 (Mpa)

∆∆qc =0 ∆∆qc =1.7 ∆∆qc =4.2

Ic=1.30

Kc=1.0

Ic=1.64

Kc=1.0

Ic=2.36

Kc=2.14

Ic=2.60

Kc=3.31

KC=1.0

Seed et al. (2003) 30

Fig. 28: Contours of 15% Probability of Triggering of Liquefaction for

“Clean Sands” and Two Higher Levels of Required Fines

Adjustment of CPT Tip Resistance (Spanning the Full Range of

Fines Adjustments)

Fig. 29: Contours of 15% Probability of Triggering of Liquefaction for

“Clean Sands” and Two Higher Levels of Required “Fines

Correction” of SPT N-values (Spanning the Full Range of Fines

Corrections) (Seed et al., 2001)

N1,60

010 20 30 40

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

MW=7.5 σV'=1.0 atm

LiquefiedMarginalNon-liquefied

“Old”Data(Pre-1985)

“New”Data

FC5%

FC15%

FC35%

qc,1

0 5 10 15 20

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Rf=0.5%,∆qc=0

Rf=5%,∆qc=4.2

Rf=2%,∆qc=1.7

Seed et al. (2003) 31

however, represent much smaller adjustments for fines than

the relationship proposed by Robertson and Wride, suggesting

that the fines adjustment of the older relationship was

unconservative. The fines adjustment proposed by Suzuki et

al. (1995) was also smaller than that proposed previously by

Robertson and Wride, but did not extend to high friction

sleeve ratios (Rf > 1.0).

3.3.6 Summary and Conclusions

The new CPT-based correlation presented herein represents a

significant advance over previously available CPT-based

correlations for assessment of seismically induced soil

liquefaction hazard. These correlations are probabilistically

posed, and a “deterministic” correlation based on PL = 15% is

also recommended.

The new correlations employ a much larger database of high

quality field performance case histories than was available to

previous researchers, and the processing of these cases

involved resolution of issues that had historically added to

overall uncertainty including (1) normalization of CPT tip and

sleeve resistances for effective overburden stress effects, and

(2) development of improved “thin layer” corrections.

Overall, the new correlations are in very good overall

agreement with previous, similar CPT-based efforts with

regard to “clean sands”. The earlier “clean sand” liquefaction

boundary curve proposed by Robertson and Wride (1997) is

only slightly unconservative relative to the new relationship,

and the “clean sand” boundary curve proposed by Suzuki et al.

(1995) is slightly more conservative than the recommended

new relationship.

It is principally when dealing with silt and silty, sandy, clayey

soils that the new correlations differ significantly from earlier

and widely used CPT-based correlations. The new

relationships reflect a much smaller adjustment (increase) in

modified CPT tip resistance (qc,1,mod) as apparent fines content

and plasticity increase than the earlier relationship of

Robertson and Wride (1997), suggesting that the earlier

relationship can be significantly unconservative for these soils.

The fines adjustment of Suzuki et al. (1995) is in closer

agreement with the new relationship proposed, but does not

extend to high friction ratios and so is incomplete.

Overall, the new CPT-based relationships appears to be

largely compatible with the similarly improved SPT-based

relationships proposed by Seed et al. (2001), and the new

CPT-based relationship appears to have similar levels (only

qC,1,mod

0 5 10 15 20

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Shibata&Teparaska(1995)

Topraketal.(1999)PL=50%

Juang etal.(2003)PL=50%

Robertson &Wride (1998)

Fig. 30: Comparison Between the Recommended New CPT-Based Correlation and Previous “Clea

n Sand”

Boundary Curves from Relationships Proposed by Shibata and Teparaska (1995), Roberson and Wride

(1998), Toprak et al. (1999), and Juang et al. (2003) (MW=7.5, σσv’=1 atm.)

Seed et al. (2003) 32

marginally higher) of uncertainty (or variance) associated with

assessment of liquefaction triggering potential as the new

SPT-based relationship. This does not mean that the SPT-

based relationships are intrinsically “better”; the use of CPT

offers important advantages with regard to continuity of

penetration data, and also the ability to discern and

characterize thinner strata, than SPT (while the SPT offers

increased certainty as to soil type and character, especially

invariably stratified soils.) Accordingly, both methods have

significant relative advantages, and both are likely to be of

continued significant value to working engineers.

3.4 Vs-Based Triggering Correlations:

Liquefaction triggering correlations based on measurements of

in situ shear wave velocity (VS-based correlations) are very

attractive because: (1) VS can be measured with non-intrusive

methods (e.g. Spectral Analysis of Surface Waves (SASW)),

and (2) VS can be measured in coarse soils (gravelly soils and

coarser) in which SPT and CPT can be obstructed by

interference with coarse soils particles. V

S-based correlations

can provide both a potentially rapid screening method, and a

method for assessment of coarse, gravelly soils which cannot

be reliably penetrated or reliably characterized with small

diameter penetrometers (SPT and CPT).

At this time, the best VS-based correlation available is that of

Andrus and Stokoe (2000). Figure 32 presents the core of this

correlation, which is based on overburden stress-corrected VS,1

vs. magnitude-correlated equivalent uniform CSR. This V

S-

based correlation is well described in the NCEER Workshop

summary papers (NCEER, 1997; Youd et al., 2001.)

Although it is certainly the best of its type, this correlation is

less well-defined (more approximate) than either SPT- or

CPT-based correlations. This is not due only to lack of data

(though the V

S-based field case history database is

considerably smaller than that available for SPT and CPT

correlation development). It is also because V

S does not

correlate as reliably with liquefaction resistance as does

penetration resistance because V

S is a very small-strain

measurement and correlates poorly with a much “larger-

strain” phenomenon (liquefaction). Small amounts of

“ageing” and cementation of interparticle contacts can cause

VS to increase more rapidly than the corollary increase in

liquefaction resistance. Accordingly, the relationship between

VS and the CSR required to induce liquefaction varies

significantly with the geologic age of the deposits in question.

An additional problem with V

S-based correlations is

uncertainty regarding appropriate normalization of V

S for

effects of effective overburden stress. In view of these

uncertainties, current V

S-based correlations for resistance to

qc,1

0 5 10 15 20

CSR*

0.0

0.1

0.2

0.3

0.4

0.5

0.6

MW=7.5 σV'=1.0 atm

Rf=0.5%,∆qc=0

Rf=5%,∆qc=4.2

Rf=2%,∆qc=1.7

Suzukietal.(1995)

0.5%Rf1%

Rf0.5%

Robertson &Wride (1998)

Ic=2.59

Ic=2.07

Ic=1.64

Fig. 31: Comparison Between the Recommended New CPT-

Based Correlation and Previous

Relationships Proposed by Suzuki et al. (1995) and Robertson and Wride (1998)

[MW=7.5,

σσ

v’=1 atm.]

Seed et al. (2003) 33

“triggering” of liquefaction are best employed either

conservatively, or as preliminary (and approximate) screening

tools to be supplemented by other methods.

Efforts are underway to improve the resolution and reliability

of V

S-based correlations. Dr. Rob Kayen of the U.S.

Geological Survey recently spent a year travelling through

Asia and making V

S measuremernts (by means of SASW) at

many of the field performance case history sites employed in

the new SPT- and CPT-based correlations. This augments

recent V

S measuremernts at U.S. case history sites, and has

resulted in a tremendous increase in the number of VS-based

case histories now available, and at sites where the critical soil

stratum can be cross-identified by SPT- and/or CPT-based

methods. This new data, along with previously available VS-

based data, is currently being processed and back-analyzed,

and new V

S-based correlations are under development using

these data (using largely the same types of procedures as those

used to develop the new SPT- and CPT-based correlations.)

In the interim, the relationship of Andrus and Stokoe is the

best available, and should be used conservatively and with

understanding of the considerable uncertainties involved.

3.5 Evaluation of Liquefaction Potential in Coarse,

Gravelly Soils:

Coarse, gravelly soils can be especially problematic with

regard to evaluation of resistance to “triggering” of

liquefaction, as large particles (gravel-sized and larger) can

impede the penetration of both SPT and CPT penetrometers.

As large-scale frozen sampling and testing are too expensive

for conventional projects, engineers faced with the problem of

coarse, gravelly soils generally have three options available

here.

One option is to employ V

S-based correlations. V

measurements can be made in coarse soils, either with surface

methods (e.g. SASW, etc.) or via borings (cross-hole V

Fig. 32: Vs-Based Liquefaction Triggering Correlation (Andrus & Stokoe, 2000)

Seed et al. (2003) 34

down-hole V

S, or “OYO Method” V

S (suspension logger)).

As discussed in the previous section, VS-based correlations are

somewhat approximate, however, and so should be considered

to provide conclusive results only for deposits/strata that are

clearly “safe” or clearly likely to liquefy.

3.5.1 Short-Interval SPT

A second option is to attempt “short-interval” SPT testing.

This can be effective when the non-gravel (finer than about

0.25-inch diameter) fraction of the soil represents greater than

about half of the overall soil mix/gradation. (Note that it is

approximately the D

30 and finer particle size range that

controls the liquefaction behavior of such soils.)

Short-interval SPT involves performing the SPT in largely the

standard manner, but counting the blow count (penetration

resistance) in 1-inch increments rather than 6-inch increments.

When penetration is more than 1-inch for a single blow, a

fractional blow count of less than 1 blow/inch is credited. The

resulting history of penetration (in blows/inch) is then plotted

for each successive inch (of the 12-inches of the test). When

values (per inch) transition “significantly” from low to high, it

is assumed that a coarse particle was encountered and impeded

the penetrometer. High values (rapid increases in blowcount)

are discarded, and the low values are summed, and then scaled

to represent the equivalent number of blows per 12-inches.

(e.g.: If it is judged that 7 of the inches of penetration can be

“counted”, but that 5 of the inches must be discarded as

unrepresentatively high, then the sum of the blows per the 7

inches is multiplied by 12/7 to derive the estimated overall

blow count as blows/12 inches.)

Figure 32 illustrates this approach. This figure shows

“correction” of two of the more than 400 SPT performed as

part of the investigation of seismic stability of Calaveras Dam

in California (Olivia Chen Consultants and Geomatrix

Consultants, 2003). Large portions of the embankment dam

fill were comprised of hydraulically placed excavated

colluvium, and so represented an unusually complex and

heterogeneous mix of weathered silty and clayey fines, sands,

and variably high fractions of coarse (gravel-sized and

coarser) particles.

Figure 32(a) shows an example in which the SPT apparently

encountered significantly increased resistance after about the

8th inch of the 12-inch interval (neglecting the sacrificial first

6-inch interval which drives to test depth). For this test, the

slope of the unimpeded drive was extrapolated to develop the

estimated “corrected” blowcount of 30 blows/ft. Figure 32(b)

shows the cumulative blowcount for a second SPT at the same

site. In this case, the blowcounts increased a bit towards the

end of the test, but were judged not to have exhibited a sudden

increase, and the blowcount was therefore simply summed

over the 12-inch driving interval in the conventional manner.

(It should be noted that blowcount can often tend to increase

slightly, but not suddenly, as penetration progresses. In these

cases, judgement is required as to whether or not to impose a

“correction” to the measured full 12-inch blowcount.)

This approach has been shown to correlate well with N

BPT

values from the larger-scale Becker Penetrometer for soils

with gravel-plus sized fractions of less than about 40 to 50%.

It is noted, however, that the corrected short-interval SPT

blowcounts can still be biased to the high side due to

unnoticed/undetected influence of coarse particles on some of

the penetration increments used, so that it can be appropriate

to use lower than typical enveloping of the resulting blow-

counts to develop estimates of “representative” N-values for a

given stratum (e.g.: 25 to 40-percentile values, rather than 35

to 50-percentile values as might have been used with regular

SPT in soils without significant coarse particles).

3.5.2 The Becker Penetrometer Test

When neither VS-based correlations nor short-interval SPT can

sufficiently reliably characterize the liquefaction resistance of

coarse soils, the third method available for coarse soils is the

use of the large-scale Becker Penetrometer. As illustrated

schematically in Figure 34, the Becker Penetrometer is

essentially a large-diameter steel pipe pile driven by a diesel

pile hammer (while retrieving cuttings pneumatically up

through the hollow pipe.) The Becker Penetrometer Test

(BPT) resistance can be correlated with SPT to develop

“equivalent” N-values (NBPT). Care is required in continually

monitoring the performance of the BPT during driving, as

corrections must be made for driving hammer bounce chamber

Cumulative Blows

Penetration (inches)

Cumulative Blows

N = 30 blows/ft. N = 42 blows/ft.

’ 9480 11,280

CN: 0.46 0.42

CE: 0.87 0.87

CR: 1 1

CS: 1 1

(N1)60: 14 18

Fig. 33: Adjustment of Short-Interval SPT for Effects

of Coarse Particles

(Olivia Chen Consultants, 2003)

Seed et al. (2003) 35

pressures, etc. (see Harder, 1997 and Youd et al., 2001). The

best current BPT correlation (with SPT) for purposes of

liquefaction engineering applications is described by Harder

(1997), NCEER (1997), and Youd et al. (2001), and is

presented in Figure 35 (based on energy-corrected BPT

resistance.)

BPT has been performed successfully for liquefaction

evaluations in soils with maximum particles sizes (D100) of up

to 1 m. and more, and to depths of up to 80 m. The BPT is a

large and very noisy piece of equipment, however, and both

cost and site access issues can be problematic.

Another problem with the BPT is the question of “casing

friction”. The cross-correlation between BPT resistance and

equivalent N-values (NBPT) is based largely on relatively

shallow data. As the Becker penetrometer drives the pipe

“pile” (penetrometer) to progressively greater depths, there is

progressively more side wall area available upon which side

wall friction and adhesion can act to impede penetration (in

addition to the resistance at the penetrating tip.) It is primarily

tip resistance that we seek to measure.

Three approaches have been taken to deal with this issue. The

simplest is to note that there was at least some casing drag in

the data used to establish the correlation between Becker

blowcounts and equivalent NBPT, and then to simply neglect

potential casing drag. This involves further noting that casing

drag is minimized by driving relatively continuously (without

prolonged pauses or breaks) so that the casing does not “set

up”.

The second approach was proposed by Sy et al. (1997) and

involves the use of instrumented driving, and then application

of dynamic wave equation analysis (as for regular pile driving)

to attempt to analytically separate tip and side wall resistances.

This approach suffers the same problem as does dynamic pile

driving analysis; the analysis is very sensitive to the “J-factor”

Fig. 34: Schematic Illustration of the Becker

Penetrometer (Harder, 1997)

(_________________, 19__)

Fig. 35: Cross-correlation Between Becker Penetrometer Blowcounts and Equivalent SPT Blowcounts

(Harder, 1997)

Seed et al. (2003) 36

used to represent damping along the exterior casing surface,

and this factor cannot be readily determined. The result is a

volatility in the analysis; high BPT blowcounts can tend to go

a bit higher, and low blowcounts can tend to go a bit lower,

after correction by this approach.

Figure 36 illustrates typical results. This figure shows BPT

resistance vs. depth as measured in a deep glacial till in

western Canada. The soil being penetrated is a coarse,

gravelly fill underlain by a deep deposit of glacial till. The till

is highly variable, but can be generally characterized as very

broadly well-graded, with variable fines content and coarser

particles ranging up to 1m. and more in size. As shown in this

figure, the tip resistance after correction (the solid dots) using

the method of Sy et al. ranges both slightly higher and slightly

lower than the uncorrected resistance (the solid and dashed

lines), and follows approximately the same mean trend.

In view of the uncertainties involved in either neglecting

casing drag, or analytical adjustment by dynamic wave

equation analysis, a third approach has also been developed.

This involves attempting to directly measure “casing drag”,

and then correcting BPT driving resistance for this effect.

This approach was used in investigations at the Calaveras

Dam in California. The Dam was initially constructed mainly

by hydraulic placement of colluvial fill, an unusual mix of

materials and methods that resulted in unusually complex sub-

stratification, and an unusually challenging mix of variable

soils ranging from low to very high fines contents, and with

coarse, gravelly fractions ranging from a few percent to well

over 50%. The initial embankment failed during construction

in 1918, and was subsequently completed using both dumped

and rolled fill sections. Figures 37(a) and (b) illustrate some

of the complexity of the resulting internal embankment and

foundation zonation and geometry. The final result was lower

elevation fills and underlying alluvium that required

investigation, topped by more competent rolled fills that were

potentially obstructions (with regard to casing drag) to the

planned use of BPT to investigate the variably coarse lower

soils.

Both BPT and short-interval SPT were used in this

investigation. Seventeen rotary wash borings, with more than

400 short-interval SPT, were performed, and for zones and

strata considered to be of potentially liquefaction-susceptible

soil “type” the short-interval SPT were processed as in Section

3.5.1 to develop “corrected” SPT N-values. In addition,

eleven Becker penetrometer probes were performed (a total of

more than 1600 feet of BPT). The Becker probes were driven

in 10-foot continuous lengths, and were then halted and

withdrawn 5 feet, and then re-driven 5 feet. The first foot of

re-driving was taken as “re-seating” of the penetrometer, and

material ravelled and/or squeezed into the hole during re-

driving, so that the last 2 or 3 feet had significant tip resistance

as well as casing drag. The second foot of re-drive was,

however, taken as representing almost entirely casing drag.

The casing drag from the 2nd foot of re-driving was then used

as a basis for “correcting” the total driving resistance to

account for casing drag.

Casing drag was found to typically represent between 5% to

45% of the total measured BPT resistance, with an average of

about 19%.

Because of the complex geometry, and the complex internal

sub-stratification within apparent “zones” (see Figure 37) an

unusually large number of sub-zones and sub-strata were

evaluated with regard to liquefaction resistance. Without

specifically identifying these, Table 2 presents a summary of

the characterization of these various zones using both

“corrected” short-interval SPT and casing-corrected BPT

based on subtraction of averaged casing drag measurements as

described above. Large numbers of both types of data were

available in most of the zones of interest, and both median and

35-percentile resultant “corrected” equivalent clean sand

blowcounts (N1,60,cs-values) were developed by both

approaches. As shown in Table, there is a generally good

level of agreement between the results of the short-interval

SPT and the corrected NBPT data, suggesting that these two

methods can both be used in variable soils of these types with

some reliability.

Fig. 36: Comparison Between Uncorrected BPT Driving

Resistances and Corrected Tip Resistances

Based on Dynamic Wave Equation Analysis

Seed et al. (2003) 37

(a) Cross-Section Up to Time of 1918 Upstream Slope Failure During Construction [Hazen, 1918]

(b) Current Cross-Section (as Re-Built), [Olivia Chen Consultants, 2003]

Fig. 37: Cross-Sections of Calaveras Dam Showing Embankment and Foundation Geometry

Seed et al. (2003) 38

3.6 Non-Zero Static “Driving” Shear Stresses:

An additional consideration in evaluation of the liquefaction

potential of saturated soils is the presence of non-zero static

“driving” shear stresses. These are shear stresses induced by

gravity loading (and geometry) that are present both before

and during the earthquake. Because they are gravity-induced,

they are essentially “following” loads and continue to act

during seismic (cyclic) excitation.

Figure 38 presents the “classic” representation of static,

driving shear stresses (After Seed, 1979). Figure 38(a)

illustrates the effective normal and shear stresses acting on

horizontal and vertical planes in an element of soil at some

depth below a level ground surface. There is a non-zero shear

stress within the element, as the vertical effective stress and

the “At-Rest” horizontal effective stress are not equal, but the

soil element is kinematically constrained and has no strong

intrinsic desire to deform in shear. The shear stress on the

horizontal planes is zero, and the “driving” shear stresses are

taken as zero for this case.

Figure 38(b) illustrates a second element of soil, this time at

some depth below a sloping ground surface. At this point

within (or beneath) a slope, the shear stress acting on the

horizontal planes is not zero, and the element has non-zero

“driving” static shear stresses, and a resultant desire to deform

Table 2: Selection of Representative (N1)60,CS Values for Embankment and Foundation

Zones and Subzones, Calaveras Dam and Foundation

Zone Zone

Description Subzone 30th

Percentile

(N1)60 SPT

30th

Percentile

(N1)60 BPT

50th

Percentile

(N1)60 SPT

50th

Percentile

(N1)60 BPT

Representative

Fines Content ∆NCS

(for

fines)

I Rock Berm

(Placed In The

1970s)

N/D 22 N/D 29 15 (F) N/A

II(M) 17 19 (B) 21 23 14 1.5

II(TD) 9 8 12 8 7 1

II Dumped

Weathered

Rock Fill II(US) 23 21 22 20 10 1

III Cobbly Gravel

Fill /ND 7 N/D 8 20 (F) 1.5

IV 17 23 (L) 22 (L) 25 48 N/A IV Rolled Fill IV(R) 24 12 (L) 32 (L) 16 12 (F) 1

V 13 19 16 23 20 1.5

V(F) 12 17 17 23 15 (F) 1.5

V Mixed

Dumped and

Sedimented

Hydraulic Fill V(G) 17 17 20 22 19 1.5

V(R) Mixed

Hydraulic and

Rolled Fill

21 14 (L) 24 18 15 (F) 1.5

VI 10 N/D 17 N/D 11 1

VI(F) 11 22 (L) 18 36 (L) 59 N/A

VI(G)-

Res 7 N/D 8 N/D 11 1

VI(G)-

Emb 27 22 40 31 11 1

VI Disturbed and

Mixed

Hydraulic,

Dumped, and

Rock Fill

VI(R) 12 (L) N/D 12 (L) N/D 15 1.5

VII Sedimented

Hydraulic Fill 10 N/D 13 N/D 62 N/A

VIII Base Alluvium 19 20 30 26 8 1

X Mixed Fill 12 17 13 26 19 (F) 1.5

XI Rocky

Colluvium 32 36 34 43 N/D 0

(L): Limited penetration data available

(B): Based on data at bottom of zone

(F): Calibrated field-estimated fines contents were also considered

N/A: Not Applicable (High CL content)

N/D: Not Determined

Seed et al. (2003) 39

in shear in a downslope direction. The measure of the relative

importance of these non-zero driving shear stresses is

routinely expressed as the ratio (á) of “static, driving” shear

stress acting on a horizontal plane divided by (normalized by)

the effective vertical stress acting on that plane as

á = ôhv / óv´ [Eq. 20]

Increasing levels of static driving shear stresses can have an

increasing effect on the vulnerability of the soil to cyclic

generation of pore pressures and triggering or initiation of

liquefaction. For very loose soils (soils that are contractive

under monotonic shearing), the presence of initial static

driving shear stresses (á > 0) significantly increases

vulnerability to liquefaction, as initial cyclic pore pressure

induced softening leads to monotonic accumulation of shear

deformations, and these, in turn, lead to further pore pressure

increases.

For very dense soils (soils that are dilatent under monotonic

shearing), however, the presence of non-zero initial static

driving shear stresses can lead to reduction in the rate of

generation of pore pressures during cyclic loading. As each

cycle of loading produces an incremental increase in pore

pressure, and some resultant reduction in strength and

stiffness, the driving shear stresses then act to produce shear

deformations that cause dilation of the soil, in turn reducing

pore pressures.

Figure 39 presents one of the best “simplified” representations

of the effects of non-zero static driving shear stresses on the

vulnerability of soils to “triggering” of liquefaction under

cyclic loading (after Harder and Boulanger, 1997). This figure

presents an adjustment factor (Ká) that represents the relative

increase in liquefaction due to the presence of non-zero

driving shear stresses. This factor is usually applied to scale

the equivalent uniform cyclic shear stress ratio required to

“trigger” liquefaction as

CSRliq,á >0 = CSRliq,á =0  Ká [Eq. 21]

As shown in Figure 39, the CSR required to induce

liquefaction increases with increasing á for high SPT N-

values, and Decreases with increasing á for low N-values.

It should be noted that Figure 39 is appropriate only for soils

at an initial effective overburden stress of less than or equal to

approximately 3 atmospheres. At higher initial effective

overburden stresses, the high effective stresses suppress

dilation of dense soils, and exacerbate contraction of loose

soils, so that the Ká values of Figure would rotate clock-wise

(in an adverse manner.) There is currently no widely accepted

guidance as to the degree or rate of this rotation with increased

effective stress, but work is in progress (by several research

teams) and improved guidance here can be expected in the

next year and two.

It should be noted that the Ká valus do not have to be applied

as a multiplier of the CSR required to trigger liquefaction

(they do not have to be applied to the “resistance” term.) They

can, instead, be applied as inverse multipliers of the loading

term by scaling the earthquake-induced CSR as

CSReq,á >0 = CSReq,á =0 / Ká [Eq. 22]

This has significant potential advantages with regard to

prediction of liquefaction-induced displacements and

deformations, as discussed later in parts of Section 5.

Finally, it should be noted that “slopes” are not the only

source of non-zero static driving shear stresses. Non-zero á

Fig. 38: Stress Conditions on Horizontal Planes beneath

Level and Sloping Ground Surfaces

(Seed et al., 1979)

Dr=55~70%

(N1)60~14-22

Kα

α=(τhv/σ’v)

σvo’<3tsf

1.5

1.0

0.5

00 0.1 0.2 0.3 0.

Dr=45~50%

(N1)60 ~8-12

Dr=35%

(N1)60 ~4-6

Fig. 39: Recommended Values of Kαα as a Function of

SPT N-Values for Effective Vertical Stresses of

Less Than 3 atmospheres

(After Harder and Boulanger, 1997)

Seed et al. (2003) 40

conditions can also arise due to bearing loads of shallow

foundations, due to loading of piles and other deep foundation

elements, due to “free” faces of excavations, due to grade

changes constrained by walls, etc. It has largely been

conventional to neglect the non-zero á conditions near the

edges of shallow-founded structures in performing

liquefaction triggering assessments, and this will be examined

a bit further in Section 5.4.2.

4.0 ASSESSMENT OF POST-LIQUEFACTION

OVERALL STABILITY

Once it has been determined that initiation or “triggering” of

liquefaction is likely to occur, the next step in most

liquefaction studies is to assess “post-liquefaction” global

stability. This entails evaluation of post-liquefaction strengths

available, and comparison between these strengths and the

driving shear stresses imposed by (simple, non-seismic)

gravity loading. Both overall site stability, and stability of

structures/facilities in bearing capacity, must be evaluated. If

post-liquefaction stability under simple gravity loading is not

assured, then “large” displacements and/or site deformations

can ensue, as geometric rearrangement is necessary to re-

establish stability (equilibrium) under static conditions.

The key issue here is the evaluation of post-liquefaction

strengths. There has been considerable research on this issue

over the past two decades (e.g.: Jong and Seed, 1988; Riemer,

1992; Ishihara, 1993; etc.). Two general types of approaches

are available for this. The first is use of sampling and

laboratory testing, and the second is correlation of post-

liquefaction strength behavior from field case histories with

in-situ index tests.

Laboratory testing has been invaluable in shedding light on

key aspects of post-liquefaction strength behavior. The

available laboratory methods have also, however, been shown

to provide a generally unconservative basis for assessment of

in-situ post-liquefaction strengths. The “steady-state” method

proposed by Poulos, Castro and France (1985), which used

both reconstituted samples as well as high-quality “slightly”

disturbed samples, and which provided a systematic basis for

correction of post-liquefaction “steady-state” strengths for

inevitable disturbance and densification that occurred during

sampling and re-consolidation prior to undrained shearing,

provided an invaluable incentive for researchers. The method

was eventually found to produce post-liquefaction strengths

that were much higher than those back-calculated from field

failure case histories (e.g.: Von Thun, 1986; Seed et al., 1989).

Reasons for this included: (1) the very large corrections

required to account for sampling and reconsolidation

densification prior to undrained shearing, (2) sensitivity to the

assumption that the steady-state line (defining the relationship

between post-liquefaction strength, Su,r vs. void ratio, e) which

was evaluated based on testing of fully remolded

(reconstituted) samples provides a basis for “parallel”

correction for this unavoidable sample densification, (3) use of

C-U triaxial tests, rather than simple shear tests, for field

situations largely dominated by simple shear, (4)

reconsolidation of samples to higher than in-situ initial

effective stresses, and (5) the failure of laboratory testing of

finite samples to account for the potentially important effects

of void redistribution during “undrained” shearing in the field.

It has now been well-established that both simple shear and

triaxial extension testing provide much lower undrained

residual strengths than does triaxial compression (e.g.:

Riemer, 1992; Vaid et al., 1990; Ishihara, 1993; etc.), often by

factors of 2 to 5, and simple shear tends to be the predominant

mode of deformation of concern for most field cases.

Similarly, it is well-established that samples consolidated to

higher initial effective stresses exhibit higher “residual”

undrained strengths at moderate strains (strains of on the order

of 15 to 30%), and this range of strains represents the limit of

accurate measurements for most testing systems.

These issues can be handled by performing laboratory tests at

field in-situ initial effective stress levels, and by performing

undrained tests in either simple shear or torsional shear. The

remaining unresolved issues that continue to preclude the

reliable use of laboratory testing as a basis for assessment of

in-situ (field) post-liquefaction strengths are two-fold. The

first of these is the difficulty in establishing a fully reliable

basis for correction of laboratory test values of S

u,r for

inevitable densification during both sampling and laboratory

reconsolidation prior to undrained shearing. The correction

factors required, for loose to medium dense samples, are

routinely on the order of 3 to 20, and there is no proven

reliable basis for these very large corrections. Use of frozen

samples does not fully mitigate this problem, as volumetric

densification due to reconsolidation upon thawing (prior to

undrained shearing) continues to require large corrections

here.

The second problem is intrinsic to the use of any laboratory

testing of finite samples for the purpose of assessment of in-

situ (field) post-liquefaction strengths, and that is the very

important issue of void redistribution. Field deposits of soils

of liquefiable type, both natural deposits and fills, are

inevitably sub-stratified based on local variability of

permeability. This produces “layers” of higher and lower

permeability, and this layering is present in even the most

apparently homogenous deposits. During the “globally

undrained” cyclic shearing that occurs (rapidly) during an

earthquake, a finite sublayer “encapsulated” by an overlying

layer of at least slightly lower permeability can be largely

isolated and may perform in a virtually undrained manner,

remaining essentially at constant volume. Although the

sublayer loses no volume, however, there is a progressive

rearrangement of the solids and pore fluid within the sublayer

as the soils cyclically soften and/or liquefy. This progressive

rearrangement, which causes the solid particles to settle

slightly and thus increase the density in the lower portion of

the sub-layer, while simultaneously reducing the density of the

top of the sublayer, is “localized void redistribution” during

globally undrained shearing.

Seed et al. (2003) 41

Owing to the very sensitive relationship between post-

liquefaction strength (Su,r) and void ratio (e) for loose to

medium density soils, even apparently minor amounts of

increase in void space (reduction in dry density) at the top of a

sub-layer can result in large reductions in S

u,r. In extreme

cases, water attempting to escape from the sublayer can be

temporarily trapped by the overlying, less pervious layer, and

can form a “film” or water-filled “blister” at the interface

between the two layers (in which case the shear strength, Su,r,

is reduced fully to zero along this interface.)

An interesting early example of this behavior was produced in

a centrifuge test performed by Arulanandan et al. (1993), as

illustrated in Figure 40. In this experiment, an embankment

was constructed with a sand “core” and a surrounding clay

“shell” to prevent drainage during cyclic loading. The sand

core was marked with layers of black sand so that localized

changes in volume (and density) could be tracked during

globally undrained shearing. When subjected to a model

earthquake, cyclic pore pressure generation within the sand

occurred, and the embankment suffered a stability failure.

During the “undrained” earthquake loading, the overall

volume of the saturated sand “core” remained constant,

satisfying the definition of globally undrained loading.

Localy, however, the lower portions of the sand “core”

became denser, and the upper portions suffered corollary

loosening. The top of the sand layer suffered the greatest

loosening, and it was along the top of this zone of significantly

reduced strength that the slope failure occurred.

Given the propensity for occurrence of localized void

redistribution during seismic loading, and the ability of Nature

to selectively push failure surfaces preferentially through the

resulting weakened zones at the tops of localized sub-strata

(and water blisters in worst-cases), the overall post-

liquefaction strength available is a complex function of not

only initial (pre-earthquake) soil conditions (e.g. density, etc.),

but also the scale of localized sub-layering, and the relative

orientations and permeabilities of sub-strata. These are not

qualities that can be reliably characterized, at this time, by

laboratory testing of soil samples (or “elements”) of finite

dimensions.

Accordingly, at this time, the best basis for evaluation of post-

liquefaction strengths is by development of correlations

between in-situ index tests vs. post-liquefaction strengths

back-calculated from field case histories. These failure case

histories necessarily embody the global issues of localized

void redistribution, and so provide the best indication

available at this time regarding post-liquefaction strength for

engineering projects.

Figure 41 presents a plot of post-liquefaction residual strength

(Su,r) vs. equivalent clean sand SPT blow count (N1,60,cs). This

was developed by careful back analyses of a suite of

liquefaction failures, and it should be noted that these types of

back analyses require considerable judgement as they are

sensitive to assumptions required for treatment of momentum

and inertia effects. The difficulties in dealing with these

momentum/inertia effects (which are not an issue in

conventional “static” stability analyses) are an important

distinction between the efforts of various investigators to

perform back-analyses of these types of failures. In this

figure, the original correction for fines used to develop N1,60,cs

is sufficiently close to that of Equations 6 and 7, that

Equations 6 and 7 can be used for this purpose.

Stark and Mesri (1992), noting the influence of initial

effective stress on Su,r, proposed an alternate formulation and

proposed a correlation between the ratio of Su,r/P and N1,60,cs,

as shown in Figure 42, where P is the initial major principal

effective stress (σ’1,i). This proposed relationship overstates

the dependence of Su,r on σ’1,i, and so is overconservative at

shallow depths (σ’1,i < 1 atmosphere) and is somewhat

unconservative at very high initial effective stresses (σ’1,i > 3

atmospheres).

It is also true, however, that the relationship of Figure 41

understates the influence of σ’1,i on Su,r. Figure 43 shows an

excellent example of this. Figure 43(a) shows the stress paths

for a suite of four IC-U triaxial tests performed on samples of

Monterey #30 sand, all at precisely the same density, but

initially consolidated to different effective stresses prior to

undrained shearing. (The sample void ratios shown are post-

consolidation void ratios.) As shown in this figure, the

samples initially consolidated to higher effective stresses

exhibited higher undrained residual strengths (Su,r). The ratio

between Su,r and P was far from constant, however, as shown

in Figure 43(b).

The influence of σ’1,i on Su,r (and on the ratio of Su,r/P) is a

function of both density and soil character. Very loose soils,

and soils with higher fines contents, exhibit Su,r behavior that

is more significantly influenced by σ’1,i than soils at higher

densities and/or with lower fines content. At this time, the

authors recommend that the relationship of Figure 41 (Seed &

Harder, 1990) be used as the principal basis for evaluation of

Fig. 40: Post-Failure Configuration of Centrifuge

Model Dam with Sand “Core” and Clay Shells

Showing Shear Localization Along Top of Sand

(Arulanandan et al., 1993)

Seed et al. (2003) 42

in-situ Su,r for “relatively clean” sandy soils (Fines Content <

12%). For these soils it is recommended that both

relationships of Figures 41 and 42 be used, but that a 5:1

weighting be employed in favor of the values from Figure 41.

Similarly, a more nearly intermediate basis (averaging the

results of each method, with 3:1 weighting between the

relationships of Figures 41 and 42) is recommended for very

silty soils (Fines Content > 30%). For fines contents between

12% and 30%, a linear transition in weighting between the two

proposed relationships can be used.

It must be noted that engineering judgement is still required in

selection of appropriate post-liquefaction strengths for specific

project cases. Consideration of layering and sub-layering,

permeability/drainage, and potential void redistribution, and

the potential for confluence of alignment of layering interfaces

with shear surfaces must all be considered. For most “typical”

cases, use of S

u,r values in the lower halves of the ranges

shown in Figures 41 and 42 (with due consideration for

weighting of these) appears to represent a suitably prudent

range for most engineering purposes at this time, but lower

overall average post-liquefaction strengths can be realized

when layering and void redistribution combine unusually

adversely with potentially critical failure modes.

Finally, a common question is “what happens at N1,60,cs values

greater than about 15 blows/ft.?” The answer is that the

relationships of Figures 41 and 42 should be concave upwards

(to the right), so that extrapolation at constant slope to the

right of N

1,60,cs=15 blows/ft should provide a conservative

basis for assessment of Su,r in this range. As these projected

values represent relatively good strength behavior, this linear

extrapolation tends to be sufficient for most projects. It should

be noted, however, that values of Su,r should generally not be

taken as higher than the maximum drained shear strength.

Values of S

u,r higher than the fully-drained shear strength

would suggest significant dilation. Dilation of this sort tends

to rapidly localize the shear zone (or shear band), and so

reduces the drain path length across which water must be

drawn to satisfy the dilational “suction”. As these distances

can be small, rapid satisfaction of this dilational demand is

possible, and “undrained” (dilational) shear strengths higher

than the drained strength can persist only briefly.

Accordingly, for most engineering analyses the use of the

fully drained shear strength as a maximum or limiting value is

prudent. Similarly, the maximum shear strength cannot

exceed the shear strength which would be mobilized at the

effective stress corresponding to “cavitation” of the pore water

(as it reaches a pore pressure of −1 atmosphere). The above

limit (to not more than the fully-drained strength) is a stronger

or more limiting constraint, however, and so usually handles

this problem as well.

5.0 EVALUATION OF ANTICIPATED

LIQUEFACTION-INDUCED DEFORMATIONS

AND DISPLACEMENTS

5.1 Introduction:

Engineering assessment of the deformations and

displacements likely to occur as a result of liquefaction or

pore-pressure-induced ground softening is a difficult and very

challenging step in most projects, and this is an area where

further advances are needed.

5.2 Assessment of “Large” Liquefaction-Induced

Displacements:

For situations in which the post-liquefaction strengths are

judged to be less than the “static” driving shear stresses,

deformations and displacements can be expected to be “large”;

generally greater than about 1m., and sometimes much greater.

Figure 44 shows examples of global site instability

corresponding to situations wherein post-liquefaction strengths

are less than gravity-induced driving shear stresses. These are

schematic illustrations only, and are not to scale.

Fig. 41: Recommended Relationship Between Su,r and

N1,60,CS (Seed and Harder, 1990)

Fig. 42: Relationship Between Su,r/P vs. N1,60,CS as

Proposed by Stark and Mesri (1992)

Seed et al. (2003) 43

For most engineering projects, the “large” deformations

associated with post-liquefaction “static instability” are

unacceptably large, and engineering mitigation is thus

warranted. It is often, therefore, not necessary to attempt to

make quantified estimates of the magnitudes of these “large”

deformations. Exceptions can include dams and embankments,

which are sometimes engineered to safely withstand

liquefaction-induced displacements of more than 1m.

Estimates of the “large” deformations likely to occur for these

types of cases can often be made with fair accuracy (within a

factor of about + 2). “Large” liquefaction-induced

displacements/deformations (> 1m.) are usually principally the

result of gravity-induced “slumping”, as geometric

rearrangement of the driving soil and/or structural masses is

required to re-establish static equilibrium. A majority of the

deformations, for these cases, usually occur after strong

shaking has ceased so that cyclic inertial forces are not very

important in “driving” the deformations (though they are very

important in “triggering” the liquefaction-induced ground

softening.)

Three general types of approaches can be used to estimate

expected “large” liquefaction-induced ground deformations,

and these are: (1) fully nonlinear, time-domain finite element

or finite difference analyses (e.g.: Finn et al., 1986; Beaty et

al., 1998; France et al., 2000; etc.), (2) statistically-derived

empirical methods based on back-analyses of field earthquake

case histories (e.g.: Hamada et al, 1986; Youd et al., 2002;

etc.), and (3) simple static limit equilibrium analyses coupled

with engineering judgement. When applied with good

engineering judgement, and when the critical

deformation/displacement modes are correctly identified and

suitable post-liquefaction strengths are selected, all three

methods can provide reasonable estimates of the magnitudes

of expected displacements.

Finite element and finite difference analyses are the most

complex of the three approaches, and we cannot reasonably

discuss these in detail within the confines of this paper. These

methods have, to date, principally been employed mainly for

relatively critical (and well-budgeted) studies, but growing

comfort with these methods (coupled with decreasing

computing costs) can be expected to bring these types of

analyses more into the mainstream. The principal difficulty

associated with these methods is the difficulty of evaluating

the model “input” parameters necessary for the relatively

complex behavioral and/or constitutive models used. These

models are usually “sensitive” to relatively minor variations in

one or more parameters, and assessment of this type of

parameter sensitivity is a vital element of such studies. (A

slightly more extensive discussion of these methods is

presented in Section 5.5.)

The second type of methods available are the “Hamada-type”

empirical methods for estimation of lateral displacements due

to liquefaction-induced lateral spreading. These methods are

based on back-analyses of lateral spreading case histories, and

involve probabilistically and/or statistically derived empirical

equations for estimation of expected lateral spreading

displacements. Currently, the most widely used such method

in the western U.S. is that of Bartlett and Youd (1995), as

recently updated by Youd et al. (2002). This method

addresses two types of cases: cases where there is a “free face”

towards which lateral spreading can occur (e.g.: Figures 44(a)

and 44(b)), and cases without a free face but with a sloping

ground surface (e.g.: Figures 44(c) and 44(d)). Two different

empirical equations are provided, one for each of these two

situations.

Figure 45 shows the results of this approach (both equations,

as applicable.) Figure 45(a) shows a plot of predicted

displacement magnitude vs. the actual observed displacement

(a) Stress Paths

0.2

0.4

0.6

0200 400 600 800 1000

P (kPa)

Su,r/P

(b) Ratio of Su,r/P vs. P

Fig. 43: Results of IC-U Triaxial Tests on Monterey

#30/0 Sand (After Riemer, 1992)

Seed et al. (2003) 44

- Liquefied zone with low residual undrained strength

(a) Edge Failure/Lateral Spreading by Flow

(b) Edge Failure/Lateral Spreading by Translation

(d) Translational Displacement

(e) Rotational and/or Translational Sliding

Fig. 44: Schematic Examples of Liquefaction-Induced Global Site Instability

and/or “Large” Displacement Lateral Spreading

Seed et al. (2003) 45

for the case histories studied. For (measured) displacements

greater then approximately 1.5m., the ratio of

predicted:measured displacements was generally in the range

of 0.5:1 to 2:1, and this is a reasonable band of accuracy for

engineering purposes in this range of displacements.

For displacements of less than about 1m, however, the

predictive accuracy is much poorer, reflecting the difficulty of

predicting displacements and deformations in this “small to

moderate” displacement range within which cyclic shearing

and cyclic shear stress reversal, as well as dilatent strength

with each reversal of cyclic load, gives rise to very complex

stress-strain and cyclic pore pressure behaviors.

The third method for estimation of expected “large”

liquefaction-induced displacements is based on evaluation of

the deformations/displacements required to re-establish static

equilibrium. This requires careful assessment of the most

critical mode of failure/deformation. An important issue in

this approach is the progressive acceleration and then

deceleration of the displacing soil (and/or structural) mass.

The deformations are not arrested when the geometry is

sufficiently rearranged as to produce a “static” Factor of

Safety of 1.0 (based on post-liquefaction strengths, as

appropriate.). Instead, shear strength must be employed to

overcome the momentum progressively accumulated during

acceleration of the displacing mass, so that the deforming

mass comes to rest at a “static” Factor of Safety of greater

than 1.0 (FS ≅ 1.05 to 1.25 is common, depending on the

maximum velocity/momentum achieved before decelleration).

For many problems, simply estimating the degree of geometry

rearrangement necessary to produce this level of Factor of

Safety (under “static” conditions, but with post-liquefaction

strengths) can produce fair estimates of likely displacements.

Alternatively, incremental calculations of (1) overall stability

(excess driving shear stresses), (2) acceleration (and then

decelleration) of the displacing mass due to shear stress

imbalance (vs. shear strength), (3) accrual and dissipation of

velocity (and momentum), and (4) associated geometry

rearrangement, can produce reasonable estimates of likely

ranges of displacements for many cases (Moriwaki et al.,

1998).

Finally, it should be noted that these three types of approaches

for estimation of expected “large” liquefaction-induced

displacements and deformations can be used to cross-check

each other. For example, it is prudent to check the final

geometry “predicted” by the results of finite element or finite

difference analyses for its “static” Factor of Safety (with post-

liquefaction strengths.)

5.3 Assessment of “Small to Moderate” Liquefaction-Induced

Displacements:

Although it is feasible to make reasonably accurate estimates

of post-liquefaction deformations and displacements for cases

of “large” displacements, we currently do not have tools for

accurate and reliable estimation of “small to moderate”

liquefaction-induced displacements (displacements/deforma-

tions of less than about 0.75m). Unfortunately, it is this

“small to moderate” range of 0 to 0.75m. that is most

important for most conventional buildings and engineered

facilities.

Unlike the case of “large” liquefaction-induced displacements,

which are dominated by displacements “driven” principally by

gravity forces after the cessation of strong shaking, “small to

moderate” displacements are very strongly affected by cyclic

inertial forces produced by strong shaking. In addition, “small

to moderate” displacements are usually controlled in large part

by complicated cyclic, pore pressure-induced softening

followed by dilation and corollary reduction in pore pressures

(and consequent re-establishment of strength and stiffness).

(a) All case histories

(b) Cases of less than 2m. displacement

Fig. 45: Predicted vs. Measured Displacements from Lateral

Spreading Case Histories (Youd et al., 2002)

Seed et al. (2003) 46

This softening and re-stiffening behavior is relatively complex

and difficult to predict with good accuracy and reliability.

Figures 46 through 48 illustrate the complicated types of

mechanical behaviors that control cyclic deformations in this

“small to moderate” displacement range. Figure 46 presents

the results of an undrained cyclic simple shear test of

Monterey #0/30 sand at a relative density of Dr = 50%, and an

initial vertical effective stress of σ’v,i = 85 kPa. These

conditions correspond roughly to a soil with an N1,60,cs value of

about 10 blows/ft. In this figure, (a) the bottom left figure

presents evolution of cyclically-induced pore pressures

(expressed as reduction in σ’v,i), (b) the bottom right figure

shows increasing shear strains with increasing numbers of

cycles, (c) the top right figure shows shear stress vs. shear

strain behavior, and (d) the top left figure presents the

effective stress path followed during this test. All four sub-

figures are scaled so that the axes of the figures to the side

and/or above and below each share commonly scaled axes.

As shown in Figure 46, shear strains are relatively small for

the first 25 cycles, until significant cyclically-induced pore

pressures have been generated. At that point (after about 25

cycles), there is a rapid increase in cyclic shear strains,

representing “triggering” of liquefaction. Examining the

stress path plots (and also the stress-strain and cyclic pore

pressure generation plots) shows clearly that pore pressures

are generated upon initial reversal of cyclic shear stresses

during each half-cycle of loading, but that dilation ensues later

in each cycle as shear strains begin to increase in the new

direction of loading. This process of cyclic softening and then

re-stiffening during each cycle is now well understood, but

remains difficult to model reliably for non-uniform (irregular)

cyclic loading, as in earthquakes.

Figure 47 similarly shows the same suite of plots for an

undrained cyclic simple shear test on a sample of the same

sand, but this time at an initial relative density of Dr = 75%.

This corresponds roughly to an in situ N1,60,cs value of about 25

to 30 blows/ft. Denser soils in this range exhibit very different

behavior than the looser sample of Figure 46.

The cyclic stress ratio of the test presented in Figure 47 is 1.8

times higher than that of the previous figure. The denser

sample (Figure 47) is more strongly dilatant with each half-

cycle of loading, and instead of relatively “suddenly”

beginning a rapid rate of increase of shear strains (as in the

previous test), this denser sample exhibits a more moderate

(and less dramatically accelerating) rate of increase of cyclic

shear strains. Indeed, as there is no sudden transition in

behaviors, it is difficult to identify a singular point at which

“triggering” of liquefaction can be said to occur. At this time,

it is recommended that “triggering” or initiation of

liquefaction be considered to have occurred when a soil has

experienced significant cyclic pore pressure generation (and

attendant softening and loss of strength), and has reached a

cyclic shear strain (in either single direction) of γ ≅ 3%. At

this level of shear strain, subsequent performance (including

“post-liquefaction” strength and stress-deformation behavior)

will be controlled largely by the soil’s contractive or dilational

behaviors.

Further complicating the issue of prediction of liquefaction-

induced deformations is the fact that, for most cases of

engineering interest, there is a directionally preferential

“driving” shear stress due to gravity loading (in addition to

cyclic inertial stresses induced by the earthquake). Figure 48

presents the results of an undrained cyclic simple shear test

with these initial "driving” shear stresses. In this test, the

“driving” shear stresses are aligned in the same direction as

the (reversing) cyclic shear stress loading, and the initial

(constant) driving shear stresses are equal to 0.08 times the

initial vertical effective stress (of 85 kPa). In addition to the

types of cyclic softening and dilatent re-stiffening shown in

the two previous figures, this test (Figure 48) also exhibits

cyclic “ratcheting” or progressive accumulation of shear

strains in the direction of the driving shear force. It is this

type of complex “ratcheting” behavior that usually principally

controls “small to moderate” liquefaction-induced

deformations and displacements (displacements in the range of

about 2 to 75 cm. for field cases.)

This problem is further complicated in field cases by the

occurrence of cyclic shear stresses “transverse” (not parallel

to) the direction of the (static) driving shear stresses.

Boulanger et al. (1995) clearly demonstrated that cyclic shear

stresses transverse to driving shear forces can, in many cases,

represent a more severe type of loading for “triggering” of

liquefaction than cyclic shear stresses aligned “parallel” with

driving forces. It is only in the last few years, however, that

high quality laboratory data with “transverse” as well as

“parallel” cyclic simple shear loading (and driving shear

stresses) has begun to be available (e.g.: Kammerer, 2002;

Wu, 2003; etc.), and development and calibration of improved

analytical and constitutive models for this type of behavior are

currently still under development. Additional complications

involved in attempting to predict “small to moderate”

liquefaction-induced deformations and displacements include:

(1) the irregular and multi-directional loading involved in field

situations, representing a complex and multi-directional

seismic response problem, and (2) the many types and

“modes” of deformations and displacements that can occur.

Figures 49 and 50 illustrate a number of “modes” or

mechanisms that can result in “small to moderate” lateral and

vertical displacements, respectively. These figures are

schematic and for illustrative purposes only; they are not to

scale.

Figure 49 illustrates three examples of modes of deformation

that can produce “small to moderate” liquefaction-induced

lateral displacements (of less than about 1m.) It should be

noted that these can also produce much larger deformations, if

the liquefiable soils are very loose, and geometry is

sufficiently adverse. Figure 49(a) shows an example of limited

lateral spreading towards a free face, and Figure 49(b) shows

an example of limited lateral spreading downslope or

downgrade. These modes can also give rise to large

Seed et al. (2003) 47

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.300.2 0.4 0.6 0.8 1

Norm. Shear Stress, (τ/σc )

Normalized Effective Vertical Stress,

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

-10 -5 0 5 10

Shear Strain, γ (%)

00.20.40.60.81

Excess Pore Pressure Ratio, ru (%)

Number of Cycles

-10 -5 0 5 10

Shear Strain, γ (%)

Fig. 46: Undrained Cyclic Simple Shear Test on Monterey #30/0 Sand (Test No. Ms15j)

Dr=50%, σσv,i’=85 kPa, CSR=0.22, αα=0

-0.4

-0.2

0.0

0.2

0.4

00.2 0.4 0.6 0.8 1

Norm. Shear Stress, (τ/σc )

Normalized Effective Vertical Stress, σ'v/σc'

-0.4

-0.2

0.0

0.2

0.4

-10 -5 0 5 10

Shear Strain, γ (%)

00.20.40.60.81Excess Pore Pressure Ratio, ru (%)

Number of Cycles

-10 -5 0 5 10

Shear Strain, γ (%)

Fig. 47: Undrained Cyclic Simple Shear Test on Monterey #30/0 Sand (Test No. Ms30j)

Dr=75%, σσv,i’=85 kPa, CSR=0.4, αα=0

Seed et al. (2003) 48

displacements, but when the liquefiable soils have limited

shear strain potential (the shear strain required for dilatent re-

stiffening), then displacements are limited.

Figure 52 (Shamoto, Zhang and Tokimatsu, 1998) presents

engineering estimates of limiting (post-liquefaction) shear

strains, as a function of SPT N-values. As shown previously

in Figures 46 through 48, the shear strain required for

dilational re-stiffening decreases with increased initial density

(or increased N-value). Although there is not yet a well-

established (or well-defined) basis for selection of the precise

shear strain corresponding to the “limiting” shear strain (see

for examples, Figures 46 through 48), the types of values

presented in Figure 52 represent suitable approximate values

for many engineering purposes. An updated set of

recommendations of this type will be presented and discussed

in Section 5.4.

These “limiting” shear strains are not, by themselves, an upper

bound to displacement potential; rather they are a basis for

estimation of resistance to shear deformations. In field cases

in which significant and adverse static “driving” shear stresses

occur (e.g. slopes, free faces, etc.) actual deformations can be

as much as twice the values of these “limiting” shear strains,

and even more when post-liquefaction residual strength is low

relative to the static driving shear stresses.

The two general types of lateral spreading deformations

illustrated in Figures 49(a) and (b) correspond to the two types

of lateral spreading addressed by the empirical corelation

proposed by Youd et al. (2002). As shown in Figure 45(a),

this approach provided reasonable estimates of expected

displacements for cases with displacements of greater than

about 1.5m. However, as shown in Figure 45(b) (which is an

enlarged view of part of Figure 45(a)), this approach does not

provide accurate or reliable estimates of lateral displacements

for cases where measured displacements are less than about

1m. (the range within which complex cyclic inertial loading

and cyclic softening and dilational re-stiffening largely control

displacements). There are, at present, no well-calibrated and

-0.2

0.0

0.2

0.4

0.60 0.2 0.4 0.6 0.8 1

Norm. Shear Stress, (τ/σc )

Norm. Effective Vertical Stress, σ'v/σc

-0.2

0.0

0.2

0.4

0.6

-2 0 2 4 6 8 10 12

Shear Strain, γ (%)

00.20.40.60.81 Excess Pore Pressure Ratio, ru (%)

Number of Cycles

-2 0 2 4 6 8 10 12

Shear Strain, γ (%)

Fig. 48: Undrained Cyclic Simple Shear Test on Monterey #30/0 Sand (Test No. Ms10k)

Dr=55%, σσv,i’=85 kPa, CSR=0.33, αα=0.18

Seed et al. (2003) 49

verified engineering tools for accurate and reliable estimation

of lateral displacements in this range. This is an area of urgent

need for further advances, and research to fill this gap is

underway in several countries.

Figure 45(c) shows another mechanism which can produce

“limited” lateral displacements; in this case, liquefaction of

soils beneath a non-liquefied surface “crust”, and laterally

constrained against large lateral spreading towards a free face.

When the surface “crust” is thin relative to the thickness of the

underlying layer, and when the liquefied soils have low

density (low N

1,60,cs values), the “crust” can separate into

distinct sections or “blocks”, and these crustal sections can

move differentially with respect to each other. This can

produce shearing, compression and tensile separations at the

edges of surface blocks. This, in turn, can be damaging to

structures and/or utilities that are unfortunate enough to

straddle the block boundaries.

There are no good means to predict where the inter-block

boundaries will occur, and there are no reliable methods at

present to predict the magnitudes of localized differential

block displacements that are likely to occur. Ishihara (1985)

provides some insight into this “pie crust” problem, as shown

in Figure 52. Ishihara suggests, based on empirical

observations from a number of Japanese earthquakes, that

surface manifestations of liquefaction will not be significant if

(1) the site is relatively level, (2) the edges are constrained so

that lateral spreading towards a free face is prevented, and (3)

the ratio of the thickness of the non-liquefied surface “crust”

(H1) to the thickness of the liquefied underlying soils (H2) is

greater than the values indicated in Figure 52 (as a function of

peak ground surface acceleration, as shown.) It should be

noted, however, that these recommendations are useful only

up to surface peak accelerations of up to 0.4 to 0.5g, and that

these have not been verified in many earthquakes as yet.

Preliminary field data from the recent 1999 Kocaeli (Turkey)

and 1999 Chi-Chi (Taiwan) Earthquakes suggests that these

criteria may not always provide fully satisfactory

performance.

- Liquefied zone

(a) Spreading Towards a Free Face

(b) Spreading Downslope or Downgrade

Fig. 49: Schematic Examples of Modes of “Limited” Liquefaction-Induced Lateral Translation

Seed et al. (2003) 50

(a) Ground Loss Due to Cyclic Densification

and/or Volumetric Reconsolidation (b) Secondary Ground Loss Due to Erosion

of “Boil” Ejecta

Displacement (d) “Slumping” or Limited Shear Deformations

(e) Lateral Spreading and Resultant Pull-

Apart Grabens (f) Localized Lateral Soil Movement

(g) Full Bearing Failure (h) Partial Bearing Failure

or Limited “Punching” (i) Foundation Settlements

Due to Ground Softening

Exacerbated by Inertial

“Rocking”

Fig. 50: Schematic Illustration of Selected Modes of Liquefaction-Induced Vertical Displacements

Seed et al. (2003) 51

Given the potential risk associated with localized differential

movements at crustal block boundaries, it is recommended

herein that these criteria should, at a minimum, always be

supplemented by well-reinforced and laterally continuous

foundations to constrain lateral differential displacements and

to reduce differential vertical displacements at the bases of

structures at such sites, especially when the liquefied layer

contains soils with low equivalent N1,60,cs values (N1,60,cs ≤ 15),

or when the ratios of H1/H2 are near the boundaries of Figure

52. Shallow-founded structures and facilities should always

be checked for bearing capacity (with post-liquefaction

strengths) at such sites, and the possibility of lateral site

translations (lateral spreading) should also be checked.

In addition to differential lateral displacements, engineers

must also deal with the hazard associated with both total and

differential potential vertical displacements. There are a

number of mechanisms that can produce vertical

displacements of sites and/or structures and other engineered

facilities. Figure 50 presents schematic illustrations of a

number of these. Again, this figure is schematic and for

illustrative purposes only; it is not to scale. The modes of

vertical displacement illustrated in Figure 50 can be grouped

into three general catagories. Figures 50(a) and (b) illustrate

settlements due to reduction or loss of soil volume. Figures

50(c) through (f) illustrate modes of settlement due to

deviatoric ground movements. Figures 50(g) through (i)

illustrate structural settlements due to full or partial bearing

failures.

Figure 50(a) shows “ground loss” or settlement due to cyclic

densification of non-saturated soils and/or due to volumetric

reconsolidation of liquefied (or partially liquefied) soils as

cyclically-induced pore pressures escape by drainage. The

overall magnitude of these types of settlements can be

reasonably well predicted by several methods (e.g.: Tokimatsu

and Seed, 1987: Ishihara and Yoshimine, 1992), but these

methods cannot reliably predict the magnitude and distribution

of locally differential settlements. Overall settlement

estimates are generally accurate within + a factor of about 2 to

3, so long as suitable adjustments are made for fines content

(as both methods are for “clean” sands.) The fines adjustment

recommended here is that of Equations 6 and 7. Section 5.4.1

will present a recommended improved procedure for

estimation of these types of “post-liquefaction

reconsolidation” ground settlements.

Figure 50(b) illustrates a second mechanism of ground loss;

secondary ground loss as a result of erosion of soil particles

carried by water escaping through cracks and fissures (often

referred to as “sand boils”) as excess pore pressures are

dissipated. Boil ejecta (transported soils) can be carried to the

ground surface, or they can be carried to accessible buried

voids (e.g.: basements, buried culverts and sewers, etc.)

Secondary ground loss due to erosion of boil ejecta is usually

localized, and so can be locally differential. It is also

essentially impossible to predict. The best defense here is

usually to ensure sufficient lateral continuity of foundations as

to be able to “bridge” or cantilever over localized subsidences.

Another alternative is deep foundation support (piles or piers)

extending beneath the depth of potential ground loss.

Fig. 52: Proposed Boundary Curves for Site

Indentification of Liquefaction-Induced

(Surface) Damage (After Ishihara, 1985)

Fig. 51: Estimates of Limiting Shear Strains for Sandy

Soils with ~ 10% Fines (Shamoto et al., 1998)

Seed et al. (2003) 52

Figures 50(c) and (d) illustrate rotational and “slumping”

(distributed shear) types of ground movements that produce

settlements at the crests or heels of the slopes or

embankments. Although these types of potential liquefaction-

induced deformations and displacements are relatively

amenable to engineering prediction when they are “large”

(>1m.), there are at present no accurate and reliable (or well-

calibrated) methods for estimation of expected displacements

when displacements will be small to moderate (D ≅ 0.05 to

0.75m.) Accordingly, significant judgement is currently

required to assess the likely deformations, and their impact on

structures and other engineered facilities. The lack of reliable

and well-calibrated analysis tools here often results in the need

for conservative assumptions, and often leads to

implementation of conservative hazard mitigation measures.

Figures 50(e) and (f) illustrate closely related mechanisms that

can produce surface settlements. Figure 50(e) illustrates

lateral spreading producing grabens, or settlements, in zones

of locally differential extension (pull-apart zones). Figure

50(f) illustrates localized lateral soil movement producing

both heaving and settlement as overall soil volume is largely

conserved. These types of potential movements are also

difficult to predict, and again conservative assumptions and/or

conservative steps to mitigate this type of hazard are often

called for when these types of movements are judged to

represent potentially serious hazards for a site, structure, or

other engineered facility.

Finally, in addition to liquefaction-induced soil (or site)

displacements, another class of potential concerns are those

associated with potential differential movements of structures

relative to the ground. Figures 50(g) through (i) illustrate

several subsets of these types of movements.

Figure 50(g) represents the case in which liquefaction-induced

loss of strength and stiffness is sufficiently severe that full

bearing failure occurs. This type of full bearing failure occurs

when overall bearing capacity, based on post-liquefaction

strengths (Su,r) as appropriate, is insufficient for static

equilibrium under gravity loading. This can produce very

large “punching” settlements (many tens of centimeters or

more), and can even lead to toppling of structures when they

are narrow relative to their height.

Figure 50(h) represents partial bearing failure or limited

“punching” settlements. These limited punching types of

settlements can occur at isolated footings, or can occur with

mat and raft foundations (especially at corners and edges.)

Limited punching settlements are generally associated with

situations in which post-liquefaction strengths are sufficient to

prevent full bearing failure, and they are the result of cyclic

softening and attendant deformations required to generate

sufficient dilational re-stiffening as to arrest movements.

Estimation of these “limited” punching/bearing settlements

can be further complicated by the interaction of increased

cyclic vertical loads due to inertial “rocking” of structures

with cyclic softening (and cyclic dilational re-stiffening), as

illustrated schematically in Figure 50(i). There are, at present,

no reliable and well-calibrated engineering/analytical tools for

estimation of likely limited punching settlements. This is a

major gap in practice, as it is limited punching settlements ( in

the range of about 0.05 to 0.75 m.) that represent one of the

principal liquefaction-related hazards for many buildings and

engineered structures. Preliminary results of studies to

develop, and to field calibrate such analytical methods, will be

presented and discussed briefly in Section 5.4.2.

Widespread liquefaction in the city of Adapazari in the recent

1999 Kocaeli (Turkey) Earthquake and in the city of Duzce in

the 1999 Duzce (Turkey) Earthquake produced differential

foundation/soil punching types of settlements in this range for

hundreds of buildings, and many additional buildings suffered

similar ranges of settlements in the cities of Wu Feng, Nantou,

and Yuan Lin during the 1999 Chi-Chi (Taiwan) Earthquake.

These two events thus provided both strong incentive, as well

as large numbers of potential field case histories, and as a

result considerable research efforts are currently underway to

develop methods for estimation of these types of “limited”

punching/bearing displacements.

5.4 Engineering Assessment of Small to Moderate

Liquefaction-Induced Displacements (Selected Modes):

There is a need for improved “simplified” analytical methods

for engineering assessment of expected liquefaction-induced

deformations and displacements. For most civil projects, the

engineer needs a basis for estimation of likely resulting lateral

and vertical displacements of the ground and/or the base of the

structure or other engineered facility. These methods need to

be both adequately accurate and reliable, and thus must be

well-calibrated against field performance case histories.

Significant research efforts are underway, by multiple teams

of researchers and in several countries, to develop improved

analysis tools for these purposes. This section will briefly

comment on some of these evolving methods.

5.4.1 Site Settlements Due to Post-Liquefaction Volumetric

Reconsolidation

Estimation of expected site settlements due to post-

liquefaction volumetric reconsolidation (as cyclically

generated excess pore pressures are dissipated by expulsion of

water; see Figures 50(a) and (b)) is the simplest of the vertical

displacement mechanisms to analyze, and several good

methods already exist for this (Tokimatsu and Seed, 1987;

Ishihara and Yoshimine, 1992; Shamoto et al., 1998). All of

these methods produce reasonably good predictions of actual

field case history observations of post-liquefaction site

settlements for sites where lateral site displacements were

small.

Figure 53(a) presents new recommendations regarding

expected volumetric reconsolidation strains after liquefaction,

or after at least significant cyclically-induced pore pressure

Seed et al. (2003) 53

(a) Volumetric Reconsolidation Strains

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 5 10 15 20 25 30 35 40

N1,60, cs

Cyclic Stress Ratio (CSR)

Cetin et al.(2000). 50% PL

This study(2002), limiting strain

50% 35% 20%

10%

5% 3%

(b) Shear Strain Potential Index

Fig. 53: Recommended Relationships for (a) Volumetric Reconsolidation Strains and (b) Shear Strain Potential Index

as a Function of Equivalent Uniform Cyclic Stress Ratio and N1,60,CS for Mw = 7.5 (Wu, 2003)

Cetin et al. (2002)

Seed et al. (2003) 54

generation. The solid line in this figure is the “triggering”

boundary for P

L = 50% from Figure 16, and represents the

approximate boundary for “triggering” of liquefaction. The

strain contours represent expected values of volumetric strain

due to post-earthquake dissipation of cyclically generated

excess pore pressures. This is based on recent laboratory

cyclic simple shear testing data, as well as previously

available laboratory and field data from other researchers (Wu,

2003).

The horizontal axis of Figure 53(a) represents fines-adjusted,

normalized SPT penetration resistance, using the same fines

corrections that were employed previously in the new

“triggering” relationships presented in Section 3.1 (Equations

6 and 7). The vertical axis represents equivalent uniform

cyclic stress ratio adjusted for: (1) magnitude-correlated

duration weighting (DWFM), and (2) effective overburden

stress (Kó). In using this figure, the earthquake-induced

CSReq must be scaled by both DWFM and Kó, using

Equations 13 and 14.

To estimate expected site settlements due to volumetric

reconsolidation, the recommended procedure is to simply

divide the subsurface soils into a series of sub-layers, and then

to characterize each sub-layer using SPT data. Volumetric

contraction (vertical strain in “at-rest” or K0 conditions) for

each sub-layer is then simply summed to result in total site

settlements.

Figure 54 presents a summary of the results of application of

this procedure to back-analysis of field performance case

histories during a number of earthquakes (Wu, 2003). As

shown in this Figure, predicted settlements are typically within

a factor of + 2 relative to those actually observed. All of the

sites represented in Figure 54 experienced lateral site

dsiplacements of less than 1.5 m. Sites that experienced

lateral displacements of less than 0.3 m are represented by

solid symbols, and sites that experienced maximum lateral

displacements of between 0.3 to 1.5m are represented with

open symbols.

For sites experiencing “small to moderate” lateral site

displacements (displacements of between 0.3 to 1.5 m), the

vertical site settlement estimated based on summation of the

volumetric reconsolidation strains of Figure 52(a) were

increased by an additional term representing 10% to 20% of

the lateral site translation (with a mean of 15%). The vertical

bars of Figure 54 represent this 10% to 20% augmented range.

For sites expected to experience maximum lateral translations

of greater than about 1.5m, these types of “simplified”

predictions of vertical settlements should not be considered

reliable.

It should be noted that non-saturated soils (above the water

table) can also suffer volumetric contraction during strong

shaking (“shake down”). This volumetric contraction is

usually significantly less severe than that experienced by

saturated soils generating significant cyclically-induced pore

pressures (typically on the order of less than 0.5% volumetric

strain in all but the loosest of soils), but when significant

depths of non-saturated soils with low to moderate SPT

penetration resistance are present, it is adviseable to also add

prediction of non-saturated “shake-down” to estimates of site

settlements. The best published method for prediction of non-

saturated shake-down for “liquefiable” types of soils is that of

Tokimatsu and Seed (1987), and non-saturated shake-down

predictions by their method are included in the preditions

presented in Figure 54. These were only significant (greater

than about 10% of the total predicted settlements) at 5 of the

case sites studied; at several sites with very low ground water

tables and thus significant depths of non-saturated alluvial

soils with relatively low blowcounts, non-saturated shake-

down accounted for up to 20% to 30% of the total predicted

settlements.

Finally, it should be noted that deposits of cohesionless soils,

and low plasticity cohesive (“silty”) soils, can be notoriously

heterogeneous in nature. As a result, interpretation of the

results of predictions of expected site settlements due to

volumetric reconsolidation should be leavened by an

understanding of the variance or uncertainty (which appears to

be a factor of about + 2), as well as by the understanding that

these are “average” setlements, and that local differential

settlements can be expected.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Measured ground settlement (m)

Estimated ground settlement (m)

Lateral displacement < 0.3m

Lateral displacement  0.3 to 1.5m

Fig. 54: Predicted vs. Actually Observed Liquefaction-

Induced Ground Settlements (Wu, 2003)

Seed et al. (2003) 55

5.4.2 Engineering Assessment of Liquefaction-Induced

Settlements of Shallow-Founded Structures

The quest for relatively “simple” and reliable methods for

prediction of liquefaction-induced settlements of shallow-

founded structures has been one of the most important and

elusive objectives of research to fill “holes” in our analytical

repertoire. Numerous research efforts are currently underway,

by diverse teams of researchers in at least several different

countries, many inspired by the widespread damages resulting

from differential or “partial punching” settlements of many

hundreds of structures in the recent 1991 Luzon (Phillipines),

1999 Kocaeli (Turkey), 1999 Duzce (Turkey) , and 1999 Chi-

Chi (Taiwan) Earthquakes.

One approach under development by the authors (as well as a

very talented team of research students at Middle East

Technical University in Ankara, Turkey, working with Prof.

Onder Cetin) is nearing completion, and is showing very

promising results when predictions are compared with

observed field performance. This “simplified” method is not

all that simple, and space limitations do not permit a full

treatment of this approach in this paper. (Besides, it is still

under development; expected to be completed within the

calendar year.) Instead, a brief description of the approach,

and of the results to date, will be presented.

This method is being developed and field-calibrated using

field performance case histories selected for study from

among the many hundreds of structures that suffered

liquefaction-induced settlements and/or partial punching

failures in the city of Adapazari, Turkey during the 1999

Kocaeli Earthquake, and in the city of Duzce, Turkey, during

the 1999 Duzce Earthquake.

Figure 55 illustrates typical conditions considered. In both of

these cities, relatively stiff, monolithic reinforced concrete

buildings of 2 to 6 stories were routinely founded at shallow

depths (usually on thick mat foundations) over potentially

liquefaible soils and with a shallow ground water table.

Performance varied from relatively minor foundation

settlements of less than 10cm (relative to the adjacent ground),

to measured settlements of more than 1m. Several buildings

(of tall, narrow aspect ratio) in Adapazari suffered sufficient

partial bearing failures that they toppled over.

The approach under development involves first assessing post-

liquefaction stability (bearing capacity relative to post-

SPT-F1

CPT-F1

CPT-F2

CPT-F3

Clay

CL to ML

ML to SM

with CL

CL to ML

with SM

Dense SM with CL

0 m

5 m

10 m

15 m

East West

20 m

HORIZONTAL SCALE

0 5 10

q (MPa)

10.8 m

Fig. 55: Example of Foundation Soil Conditions for a Four-Story Reinforced Concrete

Structure in Adapazari, Turkey (After Bray et al., 2003)

Seed et al. (2003) 56

liquefaction residual strengths). All of the field cases studies

“passed” this screening (as, apparently, did all but a few of the

structures in these two cities), and consideration next

progressed to assessment of expected structural settlements.

Settlements of the structures had two principal contributing

source mechanisms; “volumetric” settlements arising

principally from volumetric reconsolidation, and “deviatoric”

settlements arising from cyclic loading in combination with

static “driving” shear stresses due to foundation bearing loads.

Total settlement, ÄZtotal, at either the corner or the edge of a

shallow-founded structure was then estimated as

ÄZtotal = ÄZvolumetric + ÄZdeviatoric [Eq. 23]

In this equation, ÄZvolumetric can be calculated in much the

same manner as was described in the prevoius section, except

that the cyclic loading (and the resultant CSR) in each sub-

layer beneath the building foundation is exacerbated by

soil/structure interaction (SSI) which induces additional cyclic

loading of the ground near the edges due to both differential

lateral inertial forces between the structure and the ground

(“hockey-puck-like” kinematic and inertial interaction) and

vertical loading pulses due to “rocking” forces from the

structure.

These increased (SSI-induced) cyclic loadings have been

analyzed by means of extensive 3-dimensional, nonlinear

dynamic SSI analyses, and one of the great challenges in

development of “simplified” methods is boiling down all the

results of these SSI analyses to develop simplified estimates of

exaccerbated CSR in layers near and below the foundations.

An additional, and relatively minor, issue is the increase in

vertical effective stress beneath the foundations (relative to the

adjacent free field ground) which produces a minor and

adverse Kσ effect.

With regard to ÄZvolumetric, the result of SSI-exacerbated CSR

(and of Kσ effects) is some minor increase in settlements

relative to the adjacent “free field” ground, but these were

relatively minor; typically on the order of 5 to 10 cm. As the

adjacent free field ground also experienced some non-zero

ÄZvolumetric, and as the “Observed Settlements” of the cases

presented in Figure 56 represent differential settlement of the

structure relative to the adjacent free filed ground surface, the

contribution of ÄZvolumetric to the “Estimated (predicted)

Settlements” in Figure 56 was relatively minor.

The second term of Equation 23 (ÄZdeviatoric) is more

complicated, and was the principal contributor to observd

building settlements in the two cities studied. ÄZdeviatoric

represents shear deformations in the general direction of

“partial bearing failure” though often with much smaller

displacements) and is a function of: (1) the SSI-exacerbated

cyclic loading (CSR) in each soil sub-layer beneath the

foundation, and (2) the static “driving” shear stresses due to

the bearing loads of the foundation.

Figure 53(b) presents recommended values of Shear Strain

Potential Index (SPI). SPI is the maximum shear strain

developed in 15 cycles of uniform cyclic loading, without

significant static “drivivng” shear forces (á = 0 conditions.)

The limiting shear strain indices of Figure 53(b) were used as

the principal index of resistance to shear deformations.

These are only an “index” as the actual shear strains

developed are a function of the interaction of CSR with static

“driving” shear stresses. As these driving shear stresses are

very significant near the bases of the edges of the structures,

the complex interaction of these driving loads with the

earthquake-induced cyclic loads is a critical issue. There is

some insight that can be gleaned from laboratory cyclic testing

data, but in the end the final characterization of the interaction

between CSR and á was developed by regression of field case

histories. The form of the calculation of ÄZdeviatoric is to

perform analyses of each soil sub-layer beneath the corner or

edge of the structure as in Equation 24, and then sum the

settlements. For each sub-layer

ÄZdeviatoric = f (CSRfree-field, CSRSSI, SPI, á, Ká,å) [Eq. 24]

where Ká,å is a strain-based factor that characterizes the effects

of non-zero driving shear stresses on the accumulation of

shaer strain in the driven direction. (This is somewhat

analogous to the Ká factor discussed earlier in Section 3 for

“triggering” evaluations, but is not at all the same.)

Total Building Settlement

0 10 20 30 40 50 60 70 80 90

Observed Settlement (cm)

Estimated Settlement (cm)

Fig. 56: Predicted vs. Actually Observed Liquefaction-

Induced Building Settlements in Duzce and

Adapazari (Cetin, et al., work in progress)

Seed et al. (2003) 57

Figure 56 presents a comparison between predicted

liquefaction-induced building settlements and those actually

observed for 26 buildings in the cities of Adapazari and

Duzce. Both “Estimated” and “Observed” settlements in this

figure represent settlements of the building relative to the

adjacent “free field” ground.

Most of the cases presented in Figure 56 are from the city of

Duzce, as many of the cases studied in Adapazari are

complicated by the presence of cohesive soils from “Zone B”

of Figure 3; soils that are vulnerable to cyclic strain softening

(especially when á is non-zero). These cohesive soils appear

to have contributed significantly to overall building

settlements in many of these cases, but the analytical methods

described abve are applicable only to soils of Zone A of

Figure 4.

Additional cases are being studied and analyzed, and these

analytical tools are still being refined. It is hoped that this

work will be completed by late Summer or early Fall, and

more complete presentations of this method, as well as its

development and calibration against an increasing number of

field case histories, should be available soon.

5.4.3 Engineering Assessment of “Small to Moderate”

Lateral Site Displacements

A number of researchers have investigated the phenomenon of

permanent deformation due to liquefaction-induced lateral

spreading, beginning with the seminal early work of Hamada

et al. (1986). Hamada et al. began by assembling a database

of case histories, consisting of sites where lateral spreading

occurred during three earthquakes (the 1964 Niigata, 1983

Nihonkai-Chubu, and 1971 San Fernando Earthquakes). The

case histories were divided into three main types based on

topographic conditions, which are illustrated in Figure 57: (A)

slightly inclined ground conditions (“gently sloping”), (B)

horizontal ground surface with a vertical discontinuity (a

“free-face”), and (C) horizontal ground surface and a

liquefiable layer with an inclined lower boundary. Each case

history consisted of a “segment” where the sliding could be

regarded as one block. The geotechnical characteristics and

measured displacements were then averaged across the

segment. An empirical regression technique was applied to the

database, the variable components of which were based on

topographic and geologic descriptors (e.g. thickness of

liquefiable layer, H, gradient of ground surface, θ, etc.). The

result was a very simple predictive equation for lateral spread

displacement as

DH = 0.75 H0.5θ0.33 [Eq. 25]

Bartlett and Youd (1992) built on Hamada’s empirical

approach by: (1) adding additional case histories to the

database, (2) changing the definition of a case history from the

previously described “segments” to each individual measured

displacement vector, and (3) adopting an expanded set of input

variables into the predictive equation. Cases were divided into

two types: (1) “gently sloping ground” cases, and (2) cases

with a “free face”. Separate predictive equations for each of

these two types of cases were then developed by multiple

linear regression (MLR).

Youd et al. (2002), the most recent update to Youd’s body of

work on lateral spread displacement prediction, corrects

several errors in the original database and attempts to further

optimize the variables used in the MLR predictive equations,

as shown Equations 26(a) and 26(b).

(a) For free-face conditions:

Log DH = -16.713 + 1.532M – 1.406 log R* - 0.012R + 0.592 log W

+ 0.540 log T15 + 3.413 log (100 - F15) – 0.795 log (D5015 + 0.1 mm) [Eq. 26(a)]

(b) For gently sloping ground conditions:

Log DH = -16.213 + 1.532M – 1.406 log R* - 0.012R + 0.338 log S + 0.540 log T15

+ 3.413 log (100 - F15) – 0.795 log (D5015 + 0.1 mm) [Eq. 26(b)]

Fig. 57: Types of Permanent Lateral Ground Displacements

(after Hamada et al., 1986)

Seed et al. (2003) 58

For a description of the input variables, the reader is referred

to Youd et al. (2002). Figure 45 shows the measured vs.

predicted displacements from the revised Youd et al. work

(using both equations, as appropriate to each individual case).

Despite efforts to refine the MLR equations, the predictive

capacity is largely within a factor of two for displacements of

greater than about 1.5m., but is less accurate and reliable for

smaller displacements. Because it is displacements of

considerably less than 1m. that are of principal interest for

most engineering applications, further developments are

needed.

Bardet et al. (1999) built upon the Youd et al. corrected

database, and used largely the same lateral spread case history

database to develop a probabilistic model. This is a potentially

valuable step because it casts the predictive equations in a

probabilistic format, such that an analysis of lateral spreading

could be folded into a probabilistic seismic hazard analysis

framework. Bardet et al. also performed an additional

regression on only that portion of Youd’s database that

represents displacements of less than two meters, reframing

the problem of lateral spreading deformations to focus on

“small to moderate” displacements.

Rauch and Martin (2000, 2001) took a good look at the

phenomenon of liquefaction-induced lateral spreading,

resulting in a fundamental step “backward” to the original

Hamada work. Rauch and Martin built a new database of case

histories where each lateral spread feature was represented as

a single data point, characterized by a maximum and mean

horizontal and vertical displacement, rather than using

multiple individual displacement vectors at a single “feature”

as independent data points. While this substantially reduces

the number of case histories within the database, it is a

fundamentally more sound approach, as adjacent displacement

measurements are not statistically independent as simple

statistical regression techniques would require.

Current research at UC Berkeley continues the advancement

of the “Hamada”-type approach, i.e. empirical treatment of

liquefaction-induced lateral spreading displacements. Building

upon the viewpoint that a lateral spread is a single case, but

internally addressing the variation of displacements across a

given spreading feature, ongoing research at UC Berkeley

breaks away from MLR statistical techniques and adopts the

Bayesian methodology previously described in Section 3.1 (b)

of this paper. This methodology allows for appropriate

treatment of uncertainty and variability in the data, as well as

in the modelling.

This research effort will also tailor the predictive equation

form (an ability granted by the Bayesian methodology) to

better account for principal factors affecting lateral spread

displacements: (a) magnitude and duration, as represented by

the magnitude-corrected duration-weighted cyclic shear ratio,

CSReq, (b) distributed strain within the potentially liquefiable

layer(s), as characterized and indexed to limiting shear strain

potential index (SPI), and (c) cyclic strain accumulation

attributed to interaction of cyclic loading with static driving

stress. Efforts will be made to combine the “free-face” and

“gently sloping ground” conditions into one condition

represented by a statically-induced shear strain normalized by

effective overburden stress, similar to the treatment given to

analyses of dams. The resulting model is expected to represent

an improvement on several fronts through: (1) the utilization

of engineering parameters that represent the principal factors

affecting the problem of liquefaction-induced lateral

spreading, (2) appropriate treatment of inherent variability and

statistical/model uncertainty, and (3) sound calibration against

field case histories of liquefaction-induced lateral spreading as

represented by the case history database (estimated completion

in about one year).

6.6.3 Finite Element and Finite Difference Analyses:

There are an increasing number of finite element (FEM) and

finite difference (FDM) programs available, including both

commercial and proprietary codes, for analysis of

liquefaction-related problems. Both relatively simple, and

more advanced and complex, constitutive and behavioral

models continue to evolve for these applications. FEM and

FDM analyses are increasingly being applied to significant

projects including analyses of major earth and rockfill dams as

well as complex soil/structure interaction problems including

harbor frontages, quay wall systems, levees, bridge abutments

and foundations, and pile and pier foundations in liquefiable

ground.

As analytical models become more powerful and more

complex, there is an increased need to “check” and calibrate

these analyses against field case histories and against both

simpler approximate analyses and engineering judgement on

individual engineering projects.

These types of analyses are typically “sensitive” to variations

in one or more modelling parameters, and often to variations

within the range of accuracy with which the key parameter(s)

can be defined. It is important to check for these parameter

sensitivities, and to account for the resulting range of

analytical outcomes in engineering use of the results.

The San Fernando Dam case histories are a key suite of

studies, and should be back-analyzed with any program

intended for subsequent “forward” application to analyses of

current dams. (These are also an important suite of case

histories for calibration of more “simplified” analytical

methods.) The San Fernando Dams essentially represent a

suite of four case histories, as there are two dams (the Upper

and Lower San Fernando Dams), and each dam has both an

upstream and a downstream face. Performance of the dams in

the earthquake is well-documented, and embankment and

foundation soil conditions are also well-studied.

Figure 58 illustrates the use of finite difference analyses in

back-analysis of the liquefaction-induced upstream slope

stability failure of the Lower San Fernando Dam in the 1971

San Fernando Earthquake. The code used was a modified,

proprietary version of the commercially available code FLAC

Seed et al. (2003) 59

a) FLAC Mesh of Initial Conditions

b) Final Configuration with Velocity Vectors

Fig. 58: Finite Difference Analyses of the 1971 Liquefaction-Induced Upstream Slope Failure in the Lower San Fernando Dam (Beaty and Byrne: Beaty,2001)

Seed et al. (2003) 60

(a) Post-Failure Cross-Section

(b) Reconstruction of Conditions Prior to the Earthquake

Fig. 59: Cross-Sections Through the Lower San Fernando Dam Showing Conditions Before and After the Upslope Slope Failure (Seed et al., 1988)

Seed et al. (2003) 61

(Beatty, 2001). The original code (Itasca Consulting Group,

Inc., 2000) was modified to implement a new constitutive

model and to facilitate improved treatment of post-liquefaction

stress-deformation and residual strength behaviors.

The Lower San Fernando Dam was initially constructed by

hydraulic fill methods, and was subsequently topped and

buttresses by lesser rolled fill sections. Figure 59(a) shows a

cross-section of soil conditions prior to the earthquake, and

Figure 59(b) shows the deformed/displaced configuration after

the upstream slope failure. It is well-known that the upstream

face of the Lower Dam suffered a liquefaction-induced

stability failure that resulted in large displacements (of up to

150 feet) back into the reservoir, and significant crest loss

(~40 feet) as well.

Many analysis methods have successfully concluded that the

upstream face would fail in this manner. What is more

difficult, however, is to use the same analysis method to

demonstrate that the downstream face does not fail at the same

time. Differences between the soil properties and geometries

of the upstream and downstream faces are relatively subtle,

and many analyses either predict failure of both, or successful

performance of both. The “correct” answer was a massive

upstream failure, and limited movements of less than several

feet on the downstream side.

Even more challenging, is to also use the same analytical tools

to predict the performance of the Upper Dam as well. The

Upper Dam was built with similar methods and materials, but

with different geometry. The Upper Dam remained “stable”,

suffering relatively minor lateral bulging at the lower faces on

both the upstream and downstream sides, and attendant crest

settlement of approximately 3 feet.

Figure 58(a) shows the original, pre-earthquake mesh used to

back-analyze the performance of the Lower Dam. Figure

58(b) shows the deformed mesh, and displacement velocity

vectors, at the point that the analysis was discontinued due to

excessive mesh distortion. These results indicate large

deformations and displacements to the upstream side, and only

limited displacements to the downstream side, in very good

agreement with observed behavior.

It should be noted that the analysis of the Lower dam can be

extended to larger total deformations by re-meshing, and then

continuing the analysis. This can impose some degree of

approximation in the analysis, but can also produce both

useful and reasonable results.

Figure 60 shows analyses of the performance of the Upper

Dam using the same FDM code and methods. The dashed

lines show the pre-earthquake mesh configuration, and the

solid lines show the final (post-earthquake) deformed mesh.

Maximum movements in both the upstream and downstream

directions, as well as the “predicted” crest slumping, were

again in good general agreement with those actually observed.

When using these types of (FEM or FDM) procedures to

analyze “expected” embankment deformations and displace-

ments, it is impotant to recognize the strengths and short-

comings of these approaches, and also to cross-check the

analyses against simpler analytical approaches. FEM and

FDM analyses tend to be unable to adequately “localize” shear

to a narrow shear band or slip surface in many cases of very

large displacement. On the other hand, they can suitably

model inertial and gravity “driving” forces as well as general

strength/resistance to deformations, and can produce very

reasonable predictions of the general magnitude and

distribution of displacements. The details of shear

displacements may be off, but these predictions can provide

very useful engineering insight.

One important check of predictions of “large” liquefaction-

induced displacements is to check the “static” factor of safety

based on post-liquefaction undrained residual strengths (as

discussed in Section 4.0). Because the slumping/deforming/

sliding failure masses accumulate some velocity as they move,

Fig. 60: Finite Difference Mesh Showing Final Deformed Shape of Upper San Fernando Dam

(Beaty and Byrne: Beaty, 2001)

Seed et al. (2003) 62

the resultant accumulated momentum must be reversed as they

are brought back to rest. As a result, these masses come to

rest at an apparent “post-liquefaction” Factor of Safety of

greater than 1.0, and values of FS = 1.05 to 1.2 are common

for cases of large displacements. Checking the apparent "post-

liquefaction” Factor of Safety of the final (predicted)

deformed geometry can provide important insight regarding

the reasonableness of the analytical results in such cases.

It should also be noted that even the most advanced FEM and

FDM analysis methods cannot reliably predict the degree of

local differential displacements of adjacent “blocks” along the

embankment, and cannot therefore reliably serve to predict

expected longitudinal and transverse cracking; these can be

critical issues in evaluation of the ability of the embankment

to safely retain the reservoir. Similarly, some judgement must

also be applied in assessment of the likely height of the single

lowest point of the crest after a design earthquake, as it is the

single lowest point that defines available freeboard.

These same types of considerations apply to use of these

methods for analysis of other types of problems and

geometries, including soil/structure interaction applications.

Limitations of the analytical models in terms of their accuracy

and reliability, and their ability to model key details, must be

assessed, and (1) the results need to be checked by simpler

analytical approaches, and (2) significant judgement needs to

be applied in evaluation and use of the results of advanced

FEM or FDM analyses.

This paper cannot possibly go much deeper into this subject,

so we will summarize by noting that the increased availability

and power of FEM and FDM analysis tools does not reduce

the importance of either (1) “simplified” analytical tools, or

(2) engineering judgement. Instead, the power and complexity

of these evolving analysis tools places an increased premium

on judgement and cross-checking of the results of these more

advanced analyes. Used with prudence and judgement,

advanced (FEM and/or FDM) analysis tools can provide

significant improved insight for many engineering problems.

7.0 MITIGATION OF LIQUEFACTION HAZARD

7.1 General:

When satisfactory performance of structures and/or other

engineered facilities cannot adequately reliably be assured,

engineered mitigation of the unacceptable liquefaction hazard

is generally required. There are many methods, and variations

on methods, currently available for this, and more are under

development.

Table 3 presents a brief list of selected major mitigation

methods available. It should be noted that these do not have to

be employed singly; it can often be optimal to use two or more

methods in combination.

It is not reasonable, within the constraints of this paper, to

attempt a comprehensive discussion of all available mitigation

methods. Instead, limited comments will be offered regarding

various aspects of some of these. It should be noted that

mitigation of liquefaction hazard is an area subject to

considerable controversy, and that our understanding of the

efficacy of some of these methods is still evolving. It is

suggested that key issues to be considered in selection and

implementation of mitigation methods are: (1) applicability,

(2) effectiveness, (3) the ability to verify the reliability of the

mitigation achieved, (4) cost, and (5) other issues of potential

concern (e.g.: environmental and regulatory issues, etc.). More

comprehensive treatments of many of the mitigation methods

listed in Table 3 are available in a number of references (e.g.:

Mitchell et al., 1995; Hausmann, 1990).

The first class or category of methods listed in Table 3 involve

surface compaction. When this is the case, potentially

liquefiable soil types should be placed in layers and

compacted, using vibratory compaction, to specifications

requiring not less than 95% relative compaction based on the

maximum dry density (γd,max) as determined by a Modified

AASHTO Compaction Test (ASTM 1557D).

The second group of methods listed in Table 3 involves in-situ

ground densification. It is recommended that these methods

be coupled with a suitably comprehensive post-treatment

verification program to assure that suitable mitigation has

been achieved. CPT testing is particularly useful here, as it is

rapid and continuous. When CPT is to be used for post-

densification verification, it is a very good idea to establish

pre-densification CPT data, and to develop site-specific cross-

correlation between SPT and CPT data.

In addition, it should be noted that ageing effects (including

establishment of microbonding and even cementation at

particle contacts) is disrupted by in situ densification. These

ageing effects increase both resistance to liquefaction, and also

resistance to penetration (as measured by SPT, CPT, etc.)

Immediately after in-situ densification, despite increased

overall density of the soils, it is not unusual to find that

penetration resistances have not increased nearly as much as

expected, and in some cases they have even been observed to

decrease slightly. Over subsequent weeks and months,

however, as ageing effects re-establish themselves, penetration

resistances generally continue to increase. A large fraction of

ageing effects usually occur over the first 6 to 12 weeks after

treatment, and penetration tests performed sooner than this can

be expected to provide conservatively biased results.

In-situ vibrodensification, or compaction by means of

vibratory probes, has been employed to depths of 70m.

Difficulties in penetrating to depth through dense and/or

coarse soils, and failure to deliver sufficient vibrational energy

as to achieve adequate densification in the face of high

overburden stresses, can limit the efficacy of these methods at

the deepest of these depths. Vibroflotation, using vibroflots

whose vibrational source is at the lower tip of the vibrating

probe (within the ground), can generally deliver higher

vibrational energy to greater depths than most other vibratory

probe systems.

Seed et al. (2003) 63

Vibrodensification is generally very effective in soils with less

than about 5% clay fines, but can be ineffective in soils with

larger fractions of clay fines. It had long been thought that the

diffficulty in vibrodensification of soils with high fines

contents was related to the inability of water to escape, and

indeed some improvement in densification of soils with high

fines contents has been observed with the use of pre-installed

wick drains to assist in allowing egress of water. It is noted,

however, that the clay contents at which vibrodensification

begins to be ineffective are at least somewhat similar to the

clay contents at which classic cyclically-induced liquefaction

ceases to occur (see Figures 2 and 3). It appears likely that, as

vibrodensification essentially works by liquefying and

densifying the soils, the limit of “treatable” soil types is at

least somewhat coincident with the types of soils that are

“liquefiable”, and thus in need of treatment.

Some of the vibrodensification methods also result in

installation of dense gravel columns through the treated

ground (vibro-replacement). It has been suggested that these

General Category Mitigation Methods Notes

I. Excavation and/or

compaction (a) Excavation and disposal of liquefiable soils

(b) Excavation and recompaction

II. In-situ ground

densification (a) Compaction with vibratory probes (e.g.:

Vibroflotation, Terraprobe, etc.)

(b) Dynamic consolidation (Heavy tamping)

(d) Deep densification by blasting

(e) Compaction grouting

-Can be coupled with

installation of gravel

columns

-Can also provide

reinforcement

III. Selected other

types of ground

treatment

(a) Permeation grouting

(b) Jet grouting

(d) Drains

- Gravel drains

- Sand drains

- Pre-fabricated strip drains

(e) Surcharge pre-loading

(f) Structural fills

-Many drain installation

processes also provide

in-situ densification.

IV. Berms, dikes,

sea walls, and

other edge

containment

structures/systems

(a) Structures and/or earth structures built to

provide edge containment and thus to prevent

large lateral spreading

V. Deep foundations (a) Piles (installed by driving or vibration)

(b) Piers (installed by drilling or excavation)

-Can also provide ground

densification

VI. Reinforced shallow

foundations (a) Grade beams

(b) Reinforced mat

(d) “Rigid” raft

Table 3: List of Selected Methods for Mitigation of Seismic Soil Liquefaction Hazard

Seed et al. (2003) 64

dense gravel columns, which have high shear moduli relative

to the surrounding (treated) soils, will attract a large share of

the earthquake-induced cyclic shear stresses propagating

through the composite treated ground, and thus partially shield

the softer surrounding soils from cyclic loading. This, in turn,

would produce the added benefit of reducing the cyclic shear

stress ratios (CSR) to which the treated soils would be

subjected during an earthquake.

Estimates of the level of shear stresses borne by the dense

gravel columns are sometimes computed by estimating the

contributions of the stiffer columns and the softer surrounding

soil, based an assumption of a simple shear mode of

deformation, and using contributory areas of the dense gravel

columns and the surrounding soils and their respective shear

moduli. Unfortunately, for column height to diameter ratios of

greater than about three, the deformations of the gravel

columns are dominated by flexure, rather than simple shear,

and this renders them much “softer” than the above-described

analyses would suggest. Indeed, the gravel columns generally

provide relatively little “shielding” of the surrounding soils,

and this hypothesized shielding effect can usually best be

conservatively neglected.

Vibrodensification-installed gravel columns are also

sometimes credited as serving as “drains” to rapidly dissipate

seismically-induced excess pore pressures. This will be

discussed a bit later under “Drains”.

Dynamic consolidation (or heavy tamping) involves raising a

large mass to great height (with a crane), and then dropping it,

producing both surface impact and vibrational compaction.

The depth to which this can be effective is principally a

function of the weight that can be raised, and the height from

which it can be dropped. Good results can usually be

achieved to depths of up to about 6 to 8m. with “conventional”

equipment, and special purpose equipment has been built to

extend these depths somewhat for individual, large projects.

Dynamic consolidation is generally less expensive (per treated

volume) than vibrodensification, but cannot reach to the same

depths and is progressively less effective as depth increases.

Other issues, including treatable soil types and post-treatment

verification (including ageing effects) are largely as discussed

previously for vibrodensification.

Compaction piles provide improvement by three mechanisms;

(1) by densification due to driving installation, (2) by

increasing lateral stresses, and (3) by providing structural

reinforcing elements. This method is only rarely used,

however, due to its cost. It is generally employed in unusual

situations where other methods cannot reliably be

implemented.

Blasting can be used to achieve deep densification of

potentially liquefiable soils. This method, however, tends to

produce less uniform densification than vibrodensification,

and generally cannot reliably produce densities as high as

those that can be obtained with high energy vibrodensification

methods that effectively transmit high vibrational energy to

soils at depth (e.g. Vibroflotation, etc.) Blasting also raises

environmental concerns, issues regarding propagation of

vibrations across neighboring sites, and issues regarding noise

and safety.

Compaction grouting is the last of the “in-situ ground

densification” methods listed in Table 3, and is also the first of

three “grouting” methods listed in this table. Compaction

grouting involves injection of very stiff (low slump) cement

grout into the ground at very high pressure, ideally forming

“bulbs” of grout and displacing the surrounding soils.

Compaction grouting works both by densifying soils, and by

increasing in-situ effective lateral stresses. The degree of

densification that can be achieved by the monotonic (non-

cyclic) loading imposed by the growing grout mass is

dilationally limited, however, and recent research suggests that

the increased lateral stresses can relax over time. An

additional drawback is the difficulty in verifying improvement

by means of penetration testing. Compaction grouting

performed well at one site in San Francisco during the 1989

Loma Prieta Earthquake, but the site was subjected to only

moderate levels of shaking (amax ~ 0.2g., and a relatively short

duration of shaking). This method remains unproven at higher

levels of shaking.

Permeation grouting involves injection of a grouting agent in a

fluid form into the void spaces between the soil grains. A

limitation of this method is the inability of even the most

finely ground cement grouts to reliably penetrate into the

voids of soils with greater than about 6 to 10% fines. As this

can include silty fines, this leaves most silty soils potentially

vulnerable to liquefaction. This is also problematic in sandy

and silty soil deposits of variable fines content, a common

situation. Chemical grouts are available that can more reliably

penetrate into finer soils, but these are increasingly

problematic with regard to environmental and regulatory

issues. Another significant drawback with permeation

grouting is the inability to know, with certainty, just where the

grout has actually gone. This is exacerbated by the inability to

“check” conditions after treatment, except by means of

expensive borings, as the hardened grout impedes penetration

of CPT. Finally, cost is usually very high.

Jet grouting is an attempt to achieve grout penetration by

jetting at very high pressure from a rotating probe, as the

probe is withdrawn. Ideally, this produces a cylindrical

column of treated soil (or soil cement). Penetration of the jet

varies with soil density and character, however, so that the

diameter of the treated column can be uncontrollably variable.

Coarse particles (gravelly and coarser) can fully deflect the jet,

leaving untreated slivers in the treated column. As with

permeation grouting, post-treatment “checking” is rendered

difficult and expensive by the hardened treated column. This

method is also expensive, and it is not economical to attempt

to treat the full volume of liquefiable soil. Accordingly,

treatment of overlapping columns is employed, as described

below for deep soil mixing. Overall, jet grouting can be an

uncertain process in variable cohesionless soils, and has been

Seed et al. (2003) 65

supplanted to some extent by the more certain process of deep

mixing for liquefaction applications.

Deep mixing involves the use of large augers both to introduce

cement grout and to mix it with the soil, producing treated soil

cement columns. This is essentially a brute force method, and

it has a significant advantage over both permeation and jet

grouting inasmuch as the injection and mixing process

provides reliable treatment of a known volume of soil. The

problem with deep mixing is that it is not economical to treat

the full liquefiable soil volume. Accordingly, rows of slightly

overlapping treated columns are used to create “walls”, and

these are arranged in a cellular pattern (in plan), surrounding

“cells” of untreated soil. The soils within the cells can still

liquefy, however, especially when the “treatment ratio” (the

ratio between treated soil volume, and the untreated volume

within the cells) is low. Soils within the cells can also settle,

producing differential settlements. This can, clearly, be an

effective method, and performance was good at one site

during the recent 1995 Kobe Earthquake. It is not known with

any assurance, however, exactly what treatment ratios are

required for various situations, and as the cost of treatment is

relatively high, selection of treatment ratios has a tremendous

impact on overall cost.

Drains are a very interesting and challenging method for

mitigation of liquefaction hazard. An important potential

drawback of this method is that it poses a “brittle” solution; it

is effective only if it successfully promotes sufficiently rapid

dissipation of pore pressures as to prevent the occurrence of

liquefaction. If pore pressure dissipation is not sufficiently

rapid during the relatively few critical seconds of the

earthquake, however, this method does relatively little to

improve post-liquefaction performance. An additional

drawback is that, although it may prevent liquefaction, this

method only reduces (but does not eliminate) settlements due

to cyclic densification and reconsolidation after partial cyclic

pore pressure generation.

A major difficulty in the use of drains is the need to assess the

in-situ permeability of the soils to be drained. It is usually

difficult to reliably assess the in-situ permeability of soils with

an assured accuracy of better than about plus and minus one to

two orders of magnitude, and this type of uncertainty can have

a tremendous effect on the required spacing of drains. This is

routinely exacerbated by the intrinsic in-situ variability in

character (e.g.: fines content, etc.) of liquefiable soil deposits.

It should also be noted that concerns regarding potential

“plugging” of drains, either by formation of an external “skin”

of transported fines, or by infiltration of transported fines into

soil drains, is a risk that is difficult to quantify. When drains

are installed by vibro-probes, without external filters,

significant mixing of the coarse (and ostensibly free draining)

drain soils and the (finer) surrounding soils routinely occurs,

and this greatly reduces the drains’ ability to rapidly pass large

volumes of water over the critical few seconds of an

earthquake.

Drains, alone, can represent a difficult and uncertain

mitigation approach. Many of the drain installation techniques

employed also provide in-situ vibrodensification, however,

and this can be a very attractive combination. As discussed

previously, in-situ vibrodensification can be an effective

mitigation method, and can be checked to verify post-

treatment conditions. When coupled with drains, the drains

can be useful in retarding the formation of “loose” zones

and/or water blisters at the interfaces between layers of

differing vertical permeabiltiy.

Surcharge pre-loading (Method III(e) in Table 3) induces

increased vertical and horizontal effective stresses. When the

surcharge is then removed, the resulting overconsolidation

leaves the soil somewhat more resistant to triggering or

initiaion of liquefaction. The degree of increased liquefaction

resistance that can be achieved is only moderate, however, and

this is not generally an effective method in regions of high

seismicity.

Structural fills can be used to increase the thickness of a non-

liquefiable “crust” overlying potentially liquefiable soils (see

Figures 26(c) and 29). These can be further improved by

inclusion of horizontal layers of high-strength and ductile

reinforcing mats, to minimize differential movements at the

edges of “blocks” of intact crust and/or structural fill (see

Figure 26(c)).

Structural fills can also be used to buttress free faces towards

which lateral spreading otherwise might occur, and this leads

naturally to the suite of methods in Group IV of Table 3.

These methods involve creating secure containment of

“edges” or free faces towards which liquefaction-induced

lateral spreading might otherwise occur. The key here, of

course, is to ensure that the containment system itself does not

fail during the earthquake. These methods serve primarily to

prevent “large” lateral spreading deformations; they are often

less effective at reducing localized differential lateral and

vertical movements and/or bearing settlements, so the

acceptability of expected localized deformations after

remediation must be checked.

The next two groups of mitigation methods in Table 3 are

“structural” methods, and the first of these is the use of deep

foundations (piles or piers). Piles or piers, safely bearing at

depths below the occurrence of liquefaction (or significant

cyclic softening due to partial liquefaction), can provide

reliable vertical support and so can reduce or eliminate the risk

of unacceptable liquefaction-induced settlements. Pile or pier

foundations do not, however, necessarily prevent damages that

may occur as a result of differential lateral structural

displacements, so piles and/or piers must be coupled with

sufficient lateral structural connectivity at the foundation as to

safely resist unacceptable differential lateral displacements.

An additional concern, which prior to this past decade had

been routinely neglected, is the need to ensure that the piles or

piers themselves are not unacceptably damaged during seismic

excitation. Numerous field cases of damage to piles during

Seed et al. (2003) 66

earthquakes, dating back as far as the 1964 earthquakes in

Alaska and Niigata (Japan), and continuing through the recent

Kobe (Japan) and Chi Chi (Taiwan) Earthquakes, continue to

emphasize the importance of this topic. Significant research

efforts over the past 15 years have led to the development of a

range of analytical methods for this problem, ranging from

fully nonlinear, time domain, fully integrated soil/pile/

superstructure interaction analyses to considerably simpler

analyses based on separate assessment of expected site

response and resultant pile (or pier) loadings (e.g.: Pestana,

2001). These types of methods, complemented with

appropriate conservatism, can provide a suitable basis for

analysis of this issue, and for the design and detailing of piles

(or piers) and pile/cap connections.

The second group of “structural” mitigation methods in Table

3 involves the use of very stiff, reinforced shallow foundations

to resist differential lateral and vertical displacements.

Japanese practice has increasingly employed both grade beams

and continuous reinforced foundations for low to moderate

height structures, and performance of these types of systems in

earthquakes has been good. The strength and stiffness of both

grade beams and reinforced continuous foundations used in

Japan for this purpose are higher than those often used in U.S.

practice, however, and standards for design of these are

lacking in the U.S., so that engineering judgement is required

here.

Stiff, shallow foundations can be designed to adequately resist

unacceptable flexure and resultant “wracking” of the structure,

but it should be noted that differential settlements can still

result in rotational “tilting” of the structure. A number of

methods have been developed to re-level such structures after

earthquake-induced settlements, including careful micro-

underexcavation (extraction of soil by horizontal borings), and

successful re-levelling of a pair of large (12-story) reinforced

concrete apartment buildings in Nantou, Taiwan after the 1999

Chi-Chi Earthquake suggests that these methods are more

adaptable than had previously been expected.

7.2 Assessment of Mitigation:

It is important to assess the expected pereformance of the

mitigated situation. This involves returning to the top of the

framework illustrated in Figure 1, and again progressing

through the various steps to assess the expected performance

of the mitigated site and/or system, and the adequacy of this

expected performance. It is no longer acceptable practice to

simply implement mitigation; the adequacy of the mitigation

must also be evaluated.

8.0 SUMMARY AND CONCLUSIONS

There have been major advances in seismic soil liquefaction

engineering over the past decade. These advances have been

spurred in no small part by lessons and data provided by

earthquakes that have occurred over the past 15 years, as well

as by the research efforts and professional will borne of these

events. The advances achieved have, importantly, affected

practice as well as research, and soil liquefaction engineering

has now grown into a semi-mature field in its own right.

As important and heartening as the recent advances in this

field are, however, more needs to be done. Major recent, and

ongoing, advances are significantly improving our ability to

predict the probability of “triggering” or initiation of soil

liquefaction, but major gaps continue to persist with regard to

our ability to accurately and reliably assess the likely

consequences of liquefaction. This is particularly true for

situations in which structural and/or site displacements and

deformations are likely to be “small to moderate” (≤ 0.75m.).

Improved analytical and design tools, and improved

understanding of what constitutes “acceptable” performance,

are urgently needed here.

The rapid rate of progress in liquefaction engineering can be

confidently expected to continue in the years ahead.

Significant research efforts are currently underway, literally

around the world, to address all of these urgent needs. Over

the next 3 to 5 years, engineers can expect to see the results of

these efforts begin to make their way into practice.

We can also expect a need to provide improved assessments of

expected performance in response to the evolving new

questions being raised in the name of “performance-based”

engineering. Performance-based predictions are not new to

geotechnical engineers, but the levels of refinement (in terms

of increased accuracy and increased reliability) beginning to

be sought are new to the general area of liquefaction

engineering, and will continue to pose a new set of challenges.

In summary, the past decade has seen a laudable rate of

improvements in practice, and more of the same can be

expected over the next 3 to 5 years. Indeed, further advances

will be needed to keep pace with the increased demands being

generated by the ongoing shift in practice towards increasingly

performance-based design.

ACKNOWLEDGEMENTS

The authors wish to acknowledge the many dedicated and

insightful individuals who have collaborated in fomenting

many of the advances chronicled in this paper. We are

grateful to Prof’s. Armen Der Kiureghian, Khoji Tokimatsu

and Jon Stewart as well as Dr. Les Harder for their work on

the new SPT-based triggering correlation, and Prof’s Les

Youd, Kohji Tokimatsu and Jon Stewart as well as Mr. Daniel

Chu for their assistance in development of the new CPT-based

correlation. We are also grateful for discussions and sharing of

ideas with many researchers; it is often the sharing of ideas,

and even debate, that leads to important progress. Finally, we

are also grateful for continued support for much of this work

through the Pacific Earthquake Engineering Research Center

(PEER) Lifelines Research Program, without which much of

this would not have been possible.

Seed et al. (2003) 67

REFERENCES

Ambraseys, N. N. (1988). “Engineering Seismology.”

Earthquake Engineering and Structural Dynamics, Vol. 17, pp.

1-105.

Andrews, D. C. A. and Martin, G. R. (2000). “Criteria for

Liquefaction of Silty Soils.” 12th World Conference on

Earthquake Engineering, Proceedings, Auckland, New

Zealand.

Andrus, R. D. (1994). "In situ characterization of gravelly

soils that liquefied in the 1983 Borah Peak Earthquake." Ph.D.

Dissertation, University of Texas at Austin.

Andrus, R.D. and Stokoe, K.H (1997). “Liquefaction

Resistance Based on Shear Wave Velocity.” Proc., NCEER

Workshop on Evaluation of Liquefaction Resistance of Soils,

NCEER-97-0022.

Andrus, R.D. and Stokoe, K.H., (2000). “Liquefaction

Resistance of Soils from Shear-Wave Velocity.” Journal of

Geotechnical and Geoenvironmental Engineering, Vol. 126,

No. 11, pp. 1015-1025.

Arango, I. (1996). “Magnitude Scaling Factors for Soil

Liquefaction Evaluations.” Journal of Geotechnical

Engineering, Vol. 122, No. 11, pp. 929-936.

Arulanandan, K. and Symbico, J. (1993). "Post-Liquefaction

Settlement of Sands." Predictive Soil Mechanics: Proceedings

of the Wroth Memorial Symposium held at St. Catherine's

College, Oxford, July 27-29, Thomas Telford, London, pp. 94-

110.

Bardet, J.P., Mace, N., Tobita, T., and Hu, J. (1999a). “Large-

Scale Modeling of Liquefaction-Induced Ground Deformation

Part I: A Four-Parameter MLR Model.” Proceedings of the

Seventh U.S.-Japan Workshop on Earthquake Resistant

Design of Lifeline Facilities and Countermeasures Against

Soil Liquefaction, Technical Report No. MCEER-99-0019,

Mid-America Center for Earthquake Engineering Research,

Buffalo, N.Y.

Bardet, J.P., Mace, N., Tobita, T., and Hu, J. (1999b). “Large-

Scale Modeling of Liquefaction-Induced Ground Deformation

Part II: MLR Model Applications and Probabilistic Model.”

Proceedings of the Seventh U.S.-Japan Workshop on

Earthquake Resistant Design of Lifeline Facilities and

Countermeasures Against Soil Liquefaction, Technical Report

No. MCEER-99-0019, Mid-America Center for Earthquake

Engineering Research, Buffalo, N.Y.

Bartlett, S. F. and Youd, T.L. (1992). “Empirical Analysis of

Horizontal Ground Displacement Generated by Liquefaction-

Induced Lateral Spread.” Technical Report No. NCEER-92-

0021, National Center for Earthquake Engineering Research,

Buffalo, N.Y.

Bartlett, S.F. and Youd, T.L. (1995). “Empirical Prediction of

Lliquefaction-Induced Lateral Spread.” Journal of

Geotechnical Engineering, Vol. 121, No. 4, pp. 316-329.

Beaty, M. H. and Byrne, P.M. (1998). “An Effective Stress

Model for Predicting Liquefaction Behaviour of Sand.”

Geotechnical Special Publication No. 75, Proceedings of a

Specialty Conference on Geotechnical Earthquake

Engineering and Soil Dynamics III, pp. 766-777.

Beaty, M. H. (2001). “A Synthesized Approach for Estimating

Liquefaction-Induced Displacements of Geotechnical

Structures.” Ph. D. Thesis. University of British Columbia,

Vancouver, Canada.

Boulanger, R. W. and Seed, R. B. (1995). "Liquefaction of

Sand Under Bi-directional Monotonic and Cyclic Loading."

Journal of Geotechnical Engineering, Vol. 121, No. 12, pp.

870-878.

Boulanger, R. W. (2003). "High Overburden Stress Effects in

Liquefaction Analyses." Journal of Geotechnical and

Geoenvironmental Engineering, accepted in press.

Bray, J.D. and Stewart, J.P., coordinators (2000). "Damage

Patterns and Foundation performance in Adapazari." Kocaeli,

Turkey Earthquake of August 17, 1999 Reconnaissance

Report, T.L. Youd, J.P. Bardet, and J.D. Bray,

eds., Earthquake Spectra, Supplement A to Vol. 16, 163-189.

Bray, J. D., Sancio, R. B., Durgunoglu, H. T., Onalp, A., Seed,

R. B., Stewart, J. P., Youd, T. L., Baturay, M. L., Cetin, K. O.,

Christensen, C., Karadayilar, T., and Emrem, C.

(2001). "Ground Failure In Adapazari, Turkey." Proceedings

of Earthquake Geotechnical Engineering Satellite Conference

of the XVth International Conference on Soil Mechanics &

Geotechnical Engineering, Istanbul, Turkey, August 24-25.

Bray, J. D., Sancio, R. B., Durgunoglu, T., Onalp, A., Youd,

T. L., Stewart, J. P., Seed, R. B., Cetin, K. O., Bol, E.,

Baturay, M. B., and Christensen, C. (2003). "Subsurface

Characterization at Ground Failure Sites in Adapazari,

Turkey.” Journal of Geotechnical and Geoenvironmental

Engineering, submitted.

Cao, L. F., Teh, C. I., and Chang, M. F. (2001). "Undrained

Cavity Expansion in Modified Cam Clay I: Theoretical

Analysis." Geotechnique, Vol. 51, No. 4, pp. 323-334.

Cetin, K. O. (2000). "Reliability-Based Assessment of

Seismic Soil Liquefaction Initiation Hazard." Dissertation in

partial fulfillment for the degree of doctor of philosophy,

University of California, Berkeley.

Cetin, K.O. & Seed, R.B. (2000). “Earthquake-Induced

Nonlinear Shear Mass Participation Factor (rd).”

Geotechnical Engineering Research Report No. UCB/GT-

2000/08, University of California, Berkeley.

Seed et al. (2003) 68

Cetin, K. O. and Seed, R. B. (2001). "Nonlinear Shear Mass

Participation Factor (Rd) For Cyclic Shear Stress Ratio

Evaluation." Research Report No. UCB/GT-2000/08,

University of California, Berkeley.

Cetin, K.O., Der Kiureghian, A., & Seed, R.B. (2002).

“Probabilistic Models for the Initiation of Seismic Soil

Liquefaction.” Structural Safety, Vol. 24, pp. 67-82.

Chang, W. S., Bray, J. D., Gookin, W. B., and Riemer, M. F.

(1997). "Seismic Response Of Deep Stiff Soil Deposits in the

Los Angeles, California Area During The 1994 Northridge,

Earthquake." Geotechnical Research Report No. UCB/GT/97-

01, University of California, Berkeley.

Evans, M. D. (1987). "Undrained Cyclic Triaxial Testing of

Gravels : the Effect of Membrane Compliance." Ph.D. Thesis,

University of California, Berkeley.

Finn, W. D. L., Ledbetter, R. H., and Beratan, L. L. (1986).

“Seismic Soil-Structure Interaction: Analysis and Centrifuge

Model Studies.” Nuclear Engineering and Design, Vol. 94,

No. 1, pp. 53-66.

Finn, W. D. L., Ledbetter, R. H., and Wu, G. (1994).

“Liquefaction in Silty Soils: Design and Analysis.” Ground

Failures under Seismic Conditions, Geotechnical Special

Publication 44, ASCE, New York, pp. 51-76.

Finn, W. D. L. (1998) “Seismic Safety of Embankment Dams

Development in Research and Practice 1988-1998.”

Geotechnical Special Publication No. 75, Proceedings of a

Specialty Conference on Geotechnical Earthquake

Engineering and Soil Dynamics III, pp. 812-853.

France, J. W., Adams, T., Wilson, J., and Gillette, D. (2000).

"Soil Dynamics and Liquefaction", Geotechnical Special

Publication No. 107, ASCE.

Geyskens, P., Der Kiureghian, A., Monteiro, P. (1993),

"Bayesian Updating of Model Parameters", Structural

Engineering Mechanics and Materials Report No.

UCB/SEMM-93/06, University of California at Berkeley.

Hamada, M., Yasuda, S., Isoyama, R., and Emoto, K. (1986).

"Study On Liquefaction Induced Permanent Ground

Displacements", Report of the Association for the

Development of Earthquake Prediction in Japan, Tokyo Japan.

Hamada, M., O'Rourke, T. D., and Yoshida, N. (1994).

"Liquefaction-Induced Large Ground Displacement."

Performance of Ground and Soil Structures during

Earthquakes: Thirteenth International Conference on Soil

Mechanics and Foundation Engineering, New Delhi,

Proceedings, Japanese Society of Soil Mechanics and

Foundation Engineering, Tokyo, p. 93-108.

Harder, L. F. Jr. (1977). "Liquefaction of Sand under Irregular

Loading Conditions." M.S. Thesis, University of California,

Davis.

Harder, L. F. Jr. (1988). "Use of Penetration Tests to

Determine the Cyclic Loading Resistance of Gravelly Soils

During Earthquake Shaking." Ph.D. Thesis, University of

California, Berkeley.

Harder, L. F. Jr. (1997). “Application of the Becker

Penetration Test for Evaluating the Liquefaction Potential of

Gravelly Soils.” Proc., NCEER Workshop on Evaluation of

Liquefaction Resistance of Soils, NCEER-97-0022.

Harder, L. F. Jr. and Boulanger, R. (1997). “Application of

Kσ and Kα Correction Factors.” Proc., NCEER Workshop on

Evaluation of Liquefaction Resistance of Soils, NCEER-97-

0022.

Hausmann, M. R. (1990). Engineering Principals of Ground

Modification, McGraw Hill, pp. 632.

Hazen, A. (1918). “ A Study of the Slip in the Calaveras

Dam.” Engineering News Record, Vol. 81, No. 6, pp. 1158-

1164.

Hynes, M. E. (1988). "Pore Pressure Generation

Characteristics of Gravel under Undrained Loading." Ph.D.

Thesis, University of California, Berkeley.

Idriss, I. M., Sun, J. I., (1992). "Users Manual for SHAKE91,

A Computer Program for Conducting Equivalent Linear

Seismic Response Analyses of Horizontally Layered Soil

Deposits, Program Modified Based on the Original SHAKE

Program Published in December 1972, by Schnabel, Lysmer

and Seed, August.

Idriss, I. M. (2000). "Personal Communication"

Idriss, I.M. (2002). Notes from CDMG-sponsored

professional continuing education short-course on “Evaluation

and Mitigation of Seismic Soil Liquefaction and Seismic

Slope Instability and Slope Deformation Hazard.” Univ. of

Calif. Extension, Berkeley, Calif., August 8 – 10.

Ishihara, K. (1985). "Stability Of Natural Deposits During

Earthquakes." 11th International Conference on Soil

Mechanics and Foundation Engineering, Proceedings, San

Francisco, Vol. 1, pp. 321-376.

Ishimara, K. and Yoshimine, M. (1992). “Evaluation of

Settlements in Sand Deposits Followiing Liquefaction During

Earthquakes.” Soils and Foundations, Vol. 32, No. 1, March,

pp. 173-188.

Ishihara, K. (1993). “Liquefaction And Flow Failure During

Earthquakes: Thirty-Third Rankine Lecture.” Geotechnique,

Vol. 43, No. 3, pp. 351-415.

Seed et al. (2003) 69

Itasca Consulting Group, Inc., (2000) "FLAC Version 4.0-Fast

Lagrangian Analysis of Continua- User's Manual."

Jong, H.L. and Seed, R. B. (1988). "A Critical Investigation of

Factors Affecting Seismic Pore Pressure Generation and Post-

Liquefaction Flow Behavior of Saturated Soils." Geotechnical

Engineering Research Report No. SU/GT/88-01, Stanford

University.

Juang, C.H., Yuan, H., Lee, D.-H., and Lin,P.-S. (2003).

“Simplified Cone Penetration Test-based Method for

Evaluating Liquefaction Resistance of Soils.” Journal of

Geotechnical and Geoenvironmental Engineering, Vol. 129,

No. 1, pp. 66-80.

Kammerer, A., Pestana, J., and Seed, R. (2002). “Undrained

Response of 0/30 Sand Under Multidirectional Cyclic Simple

Shear Loading Conditions.” Geotechnical Engineering

Research Report No. UCB/GT/02-01, University of

California, Berkeley, July, 2002.

Kishida, H. (1966), "Damage to Reinforced Concrete

Buildings in Niigata City with Special Reference to

Foundation Engineering", Soils and Foundations, Vol. VII,

No. 1.

Koizumi, Y (1966). "Change in Density of Sand Subsoil

caused by the Niigata Earthquake", Soils and Foundations,

Vol. VIII, No. 2, pp. 38-44.

Ladanyi, B. and Johnston, G. H. (1974). "Behavior of Circular

Footings and Plate Anchors Embedded in Permafrost."

Canadian Geotechnical Journal, 11, 531-553.

Liao, S. S. C., and Whitman, R. V. (1986). "Overburden

Correction Factor for SPT in Sand." Journal of Geotechnical

Engineering, Vol. 112, No. 3, March 1986, pp. 373-377.

Liao, S. S. C., Veneziano, D., and Whitman, R.V. (1988).

"Regression Models for Evaluating Liquefaction Probability."

Journal of Geotechnical Engineering, Vol. 114, No. 4, pp.

389-409.

Liao, S. S. C., and Lum, K. Y. (1998). "Statistical Analysis

and Application of the Magnitude Scaling Factor in

Liquefaction Analysis." Geotechnical Earthquake Engineering

and Soil Dynamics III, Vol. 1, 410-421.

Martin, G.R. and Lew, M., eds. (1999) "Recommended

Procedures for Implementation of DMG Special Publication

117: Guidelines for Analyzing and Mitigating Liquefaction

Hazards in California," Southern California Earthquake

Center, Univ. of Southern Calif., March.

Mitchell, J. K., Baxter, C. D. P., and Munson, T. C. (1995).

"Performance of Improved Ground during Earthquakes." Soil

Improvement for Earthquake Hazard Mitigation, ASCE, New

York, 1-36.

Moss, R.E.S. (2003). “CPT-Based Probabilistic Assessment

of Seismic Soil Liquefaction Initiation.” Dissertation in

partial fulfillment of the requirements for the degree of Doctor

of Philosophy, University of California, Berkeley, in

preparation.

Muraleetharan, K.K., Seed, R.B., Kabilamany, K. (1993).

“Centrifuge Study of Volume Changes and Dynamic Stability

of Earth Dams.” Journal of Geotechnical Engineering, Vol.

119, No. 11, pp. 1717-1731.

NCEER (1997). "Proceedings of the NCEER Workshop on

Evaluation of Liquefaction Resistance of Soils." Edited by

Youd, T. L., Idriss, I. M., Technical Report No. NCEER-97-

0022, December 31, 1997.

Ohsaki, Y. (1966). "Niigata Earthquakes, 1964, Building

Damage and Soil Conditions", Soils and Foundations, Vol. 6,

No. 2, pp. 14-37.

Olivia Chen Consultants, Inc. (2003). “Report on the Seismic

Stability of Calaveras Dam”, report prepared for the San

Francisco Public Utilities Commission, Utilities Engineering

Bureau.

Olsen, R. S. and Mitchell, J. K. (1995). "CPT Stress

Normalization and Prediction of Soil Classification."

International Symposium on Cone Penetration Testing, CPT

95 Linkoping, Sweden, pp. 257-262.

Poulos, S. J., Castro, G., and France, J. W. (1985).

"Liquefaction Evaluation Procedure." Journal of Geotechnical

Engineering, Vol. 111, No. 6, pp. 772-792.

Rauch, A.F. and Martin II, J.R. (2000). “EPOLLS Model for

Predicting Average Displacements on Lateral Spreads.”

Journal of Geotechnical and Geoenvironmental Engineering,

Vol. 126, No. 4, pp. 360-371.

Rauch, A.F. and Martin II, J.R. (2001). “Prediction the

Maximum and Distribution of Displacements on Liquefaction-

Induced Lateral Spreads.” Proceedings: Fourth International

Conference on Recent Advances in Geotechnical Earthquake

Engineering and Soil Dynamics, San Diego, California.

Riemer, M. F. (1992). "The Effects of Testing Conditions on

the Constitutive Behavior of Loose, Saturated Sands under

Monotonic Loading." Ph.D. Thesis, University of California,

Berkeley.

Riemer, M. F., Seed, R. B., and Sadek, S. (1993). “The

SRS/RFT Soil Evaluation Testing Program.” University of

California, Berkeley Geotechnical Report No. UCB/GT-93/01.

Riemer, M.F., and Seed, R.B. (1997). “Factors Affecting

Apparent Position of Steady-State Line”, Journal of

Geotechnical and Geoenvironmental Engineering, Vol 123,

No. 3, pp. 281-288.

Seed et al. (2003) 70

Robertson, P. K. and Wride, C. E. (1997). "Cyclic

Liquefaction and its Evaluation Based on the SPT and CPT."

NCEER-97-0022, Proceedings of the NCEER Workshop on

Evaluation of Liquefaction Resistance of Soils.

Robertson, P. K. and Wride, C. E. (1998). "Evaluating Cyclic

Liquefaction Potential Using The Cone Penetration Test."

Canadian Geotechnical Journal, Vol. 35, No. 3, pp. 442-459.

Salgado, R. and Randolph, M. F. (2001). "Analysis of Cavity

Expansion in Sand." International Journal of Geomechanics,

Vol. 1, No. 2, pp. 175-192.

Sancio, R. B., Bray, J. D., Stewart, J. P., Youd, T. L.,

Durgunoglu, H. T., Onalp, A., Seed, R. B., Christensen, C.,

Baturay, M. B., and Karadayilar, T. (2002). "Correlation

between ground failure and soil condition in Adapazari,

Turkey." Soil Dynamics and Earthquake Engineering, 22,

1093-1102.

Sancio, R. B. (2003). “Ground Failure and Building

Performance in Adapazari, Turkey”, Ph.D. Dissertation (in

progress), supervised by Prof. J.D. Bray, University of

California, Berkeley.

Sancio, R. B., Bray, J. D., Riemer, M. F., and Durgunoglu, H.

T. (2003). "An Assessment of the Liquefaction Susceptibility

of Adapazari Silt." Paper 172, Pacific Conference on

Earthquake Engineering, Proceedings, New Zealand

Seed, H. B. and Idriss, I. M. (1971). “Simplified Procedure for

Evaluating Soil Liquefaction Potential”, Journal of the Soil

Mechanics and Foundations Division, ASCE, Vol. 97, No

SM9, Proc. Paper 8371, September 1971, pp. 1249-1273.

Seed, H.B. (1979). “Considerations in the earthquake-resistant

design of earth and rockfill dams”, Geotechnique, Vol. XXIX,

No. 3, pp. 213-263.

Seed, H.B. and Idriss, I.M. (1982). “Ground motion and soil

liquefaction during earthquakes”, Monograph, Earthquake

Engineering Research Institute, Oakland, Ca.

Seed, H. B., Tokimatsu, K., Harder, L. F., Chung, R. M.

(1984). "The Influence of SPT Procedures in Soil Liquefaction

Resistance Evaluations", Earthquake Engineering Research

Center Report No. UCB/EERC-84/15, University of

California at Berkeley, October, 1984.

Seed, H. B., Tokimatsu, K., Harder, L. F., and Chung, R. M.

(1985). "Influence of SPT Procedures in soil liquefaction

resistance evaluations." Journal of Geotechnical Engineering,

ASCE, 111(12), 1425-1445.

Seed, H. B., Seed, R. B., Harder, L. F., and Jong, H. L. (1988).

"Re-evaluation of the Slide in the Lower San Fernando dam in

the 1971 San Fernando earthquake", Earthquake Engineering

Research Center Report No. UCB/EERC-88/04, University of

California, Berkeley.

Seed, R. B. and Harder, L. F. (1990). "SPT-Based Analysis Of

Cyclic Pore Pressure Generation And Undrained Residual

Strength." H.B. Seed Memorial Symposium, Berkeley, Ca.,

BiTech Publishing, Ltd., v. 2, p. 351-376.

Seed, R. B., Chang, S. W., Dickenson, S. E., and Bray, J. D.

(1997). “Site-Dependent Seismic Response Including Recent

Strong Motion Data.” Proc., Special Session on Earthquake

Geotechnical Engineering, XIV International Conf. On Soil

Mechanics and Foundtion Engineering, Hamburg, Germany,

A. A. Balkema Publ., Sept. 6-12, pp. 125-134.

Seed, R. B., Cetin, K. O., Moss, R. E. S., Kammerer, A. M.,

Wu, J., Pestana, J. M., and Riemer, M. F. (2001). "Recent

Advances in Soil Liquefaction Engineering and Seismic Site

Response Evaluation." 4th Int. Conf. Recent Advances in

Geotechnical Earthquake Engineering and Soil Dynamics, San

Diego, California, March.

Seed, R. B., Cetin, K. O., Der Kiureghian, A., Tokimatsu, K.,

Harder, L. F. Jr., Kayen, R. E., and Moss, R. E. S. (2002).

"SPT-Based Probabilistic and Deterministic Assessment of

Seismic Soil Liquefaction Potential." Journal of Geotechnical

and Geoenvironmental Engineering, accepted, in-press.

Shamoto, Y., Zhang, J.-M., and Tokimatsu, K. (1998).

"Methods for evaluating residual post-liquefaction ground

settlement and horizontal displacement." Soils and

Foundations, Special Issue, 69-83.

Shibata, T. and Teparaksa, W. (1988). “Evaluation of

Liquefaction Potentials of Soils using Cone Penetration

Tests.” Soils and Foundations, Vol. 28, No. 2, pp. 49-60.

Stark, T. D. and Mesri, G. (1992). "Undrained Shear Strength

of Liquefied Sands For Stability Analysis." Journal of

Geotechnical Engineering, ASCE, 118(11), 1727-1747.

Suzuki, Y., Tokimatsu, K., Koyamada, K., Taya, Y., and

Kubota, Y. (1995). "Field Correlation of Soil Liquefaction

Based on CPT Data." International Symposium on Cone

Penetration Testing, CPT 95 Linkoping, Sweden.

Sy, A. and Campanella, R. G. (1994). “Becker and Standard

Penetration Tests (BPT-SPT) Correlations with Consideration

of Casing Friction”, Canadian Geotechnical Journal, Vol. 31,

No. 3, pp. 343-356.

Sy, A. (1997). “Recent Developments in the Becker

Penetration Test; 1986-1996”, 20th Canadian Geotechnical

Colloquim, Richmond, Canada, pp. 952-973.

Tokimatsu, K., and Seed, H. B. (1987). “Evaluation of

Settlements in Sands due to Earthquake Shaking.” Journal of

Geotechnical Engineering, Vol. 113, No. 8, pp. 861-878.

Toprak, S., Holzer, T. L., Bennett, M. J., Tinsley, J. C. (1999).

"CPT- and SPT-based Probabilistic Assessment of

Liquefaction Potential, Proceedings of Seventh U.S.-Japan

Seed et al. (2003) 71

Workshop on Earthquake Resistant Design of Lifeline

Facilities and Countermeasures Against Liquefaction.

Vaid, Y.P., Chung, E.K.F., and Kuerbis, R.H. (1990). “Stress

path and steady state.” Canadian Geotechnical Journal, 27, 1-

Von Thun, J. L. (1986). "Analysis of Dynamic Compaction

Foundation Treatment Requirements, Stage I, Jackson Lake

Dam", Technical Memo No. TM-JL-230-26, Bureau of

Reclamation, Engineering and Research Center, Embankment

Dams Branch.

Vreugdenhil, R., Davis, R., and Berrill, J. R. (1994).

"Interpretation of Cone Penetration Results in Multilayered

Soils." International Journal for Numerical and Analytical

Methods in Geomechanics, 18, 585-599.

Wang, W. (1979.) “Some Findings in Soil Liquefaction”,

Research Report, Water Conservancy and Hydroelectric

Power Scientiific Research Institute, Beijing, August.

Wu, J. (2003). "Liquefaction Triggering and Post

Liquefaction Deformations of Monterey 0/30 Sand Under Uni-

Directional Cyclic Simple Shear Loading." Dissertation in

partial fulfillment for the degree of doctor of philosophy,

University of California, Berkeley.

Yoshimi, Y., Tokimatsu, K., Ohara, J. (1994). "In-situ

Liquefaction Resistance of Clean Sands Over a Wide Density

Range", Geotechnique, Vol. 44, No. 3, pp. 479-494.

Youd, T. L., Idriss, I. M., Andrus, R. D., Arango, I., Castro,

G., Christian, J. T., Dobry, R., Finn, W. D. L., Harder, L. F.

Jr., Hynes, M. E., Ishihara, K., Koester, J. P., Liao, S. S. C.,

Marcuson, W. F. III., Martin, G, R., Mitchell, J. K., Moriwaki,

Y., Power, M. S., Robertson, P. K., Seed, R. B., and Stokoe,

K. H., II. (1997). Summary Paper, Proc., NCEER Workshop

on Evaluation of Liquefaction Resistance of Soils, NCEER-

97-0022.

Youd, T. L., Noble, S. K. (1997). "Liquefaction Criteria Based

on Statistical and Probabilistic Analyses", Proceedings of the

NCEER Workshop on Evaluation of Liquefaction Resistance

of Soils, December 31, 1997, pp. 201-205.

Youd, T. L. (2000). Personal Communication.

Youd, T.L., I. M. Idriss, Ronald D. Andrus, Ignacio Arango,

Gonzalo Castro, John T. Christian, Richardo Dobry, W. D.

Liam Finn, Leslie F. Harder Jr., Mary Ellen Hynes, Kenji

Ishihara, Joseph P. Koester, Sam S. C. Liao, William F.

Marcuson III, Geoffrey R. Martin, James K. Mitchell,

Yoshiharu Moriwaki, Maurice S. Power, Peter K. Robertson,

Raymond B. Seed, and Kenneth H. Stokoe II. (2001).

“Liquefaction Resistance of Soils: Summary Report from the

1996 NCEER and 1998 NCEER/NSF Workshops on

Evaluation of Liquefaction Resistance of Soils.” Journal of

Geotechnical and Geoenvironmental Engineering, 124(10).

Youd et al. (2001). Personal Communication.

Youd, T. L., Hansen, C. M., and Bartlett, S. F. (2002).

“Revised Multilinear Regression Equations for Prediction of

Lateral Spread Displacement”, Journal of Geotechnical and

Geoenvironmental Engineering, Vol. 128, No. 12, pp. 1007-

1017.

Yu, H. S. (2000). "Cavity Expansion Methods in

Geomechanics." Kluwer Academic Publishers.

Multidimensional characterization of problematic soils linked to foundation and building failures in parts of Southeast Nigeria

Article

Full-text available

Apr 2024

Investigating the causes of foundation failures and building collapses can have far-reaching implications for the safety and sustainability of human lives, buildings and infrastructure. The present paper addressed a significant and critical issue related to construction and infrastructure failure in Enugu, Southeast Nigeria by integrating geotechnical testing, statistical computing, geochemical testing, and spatial analysis. In all analyses, standard procedures were adopted. In this detailed research, the tested geotechnical properties include wet bulk density (1.37–1.88 g/cm³), specific gravity (2.59–2.84), moisture content (19.54–44.04%), Atterberg limits (liquid limit 44.40–83.00%, plastic limit 22.50–43.80%, plasticity index 18.40–41.00%), shrinkage limit (12.90.00-21.40%), optimum moisture content (16.60–26.40%), maximum dry density (1.55–1.83 g/cm³), triaxial shear strength characteristics (cohesion 24.13–47.87 kPa, frictional angle 4.40-10.74°), ultimate bearing capacity (230.32-377.07 kN/m²), allowable pressure (92.13-150.83 kN/m²), rate of settlement (0.058–1.995 m/yr), and amount of settlement (2.058–4.616 mm). These tests revealed crucial insights into the soil behavior, indicating that the soils are not excellent foundation materials. While spatial statistical dendrograms categorized the vulnerable zones of failure in the area, Spearman’s and Pearson’s correlation matrices and principal component analysis highlighted the key associations existing among the geotechnical properties. The clay mineralogy was determined via X-ray diffraction and the results indicated the presence of minerals such as kaolinite, quartz, magnetite, hematite, muscovite, goethite, dickite, and feldspar, with quartz and kaolinite being the most predominant. These findings hold significant implications for regions worldwide faced with similar soil challenges. This study emphasizes the importance of interdisciplinary collaboration and continuous assessment in addressing complexities of foundation soils.

Response Comparisons of Prefabricated and Cast-in-Place Subway Station Structures in Liquefiable Soil Foundation with the Ground Surface Slight Inclined

Article

Full-text available

May 2024

In recent years, the prefabricated subway station structure (PSSS) has become a hot spot of underground structure research. In this paper, the numerical model of a soil–subway station structure in a slowly inclined liquefiable site at the surface is established by using FLAC3D finite difference software. And the applicability of the PSSS under the gently inclined liquefiable site is investigated through the foundation pore water pressure, lateral movement of liquefied soil, and dynamic response and uplift characteristics of the subway station structure. It is found that under the gently inclined liquefiable site conditions, the PSSS exhibits tilting and floating behaviors and has an anti-liquefaction effect within a certain range of surrounding soil layers. Compared with the same type of cast-in-place subway station structure (CIPSSS), it has better resistance to overturning and uplift, and the structure has less stress. Under the premise of ensuring static waterproofing, the PSSS can be applied to surface inclined liquefiable sites.

EXPERIMENTAL INVESTIGATION OF MECHANICAL BEHAVIOR OF SAND-SILT MIXTURES

Preprint

Full-text available

May 2024

Airat Iunusov

A comparative analysis of ensemble learning algorithms with hyperparameter optimization for soil liquefaction prediction

Article

Full-text available

May 2024
ENVIRON EARTH SCI

Accurate prediction of soil liquefaction potential is crucial for evaluating the stability of structures in earthquake regions. This study focuses on predicting soil liquefaction using a dataset that included historical liquefaction cases from the 1999 Turkey and Taiwan earthquakes. The dataset was divided into three subsets: Dataset A (fine-grained), Dataset B (coarse-grained), and Dataset C (all samples). Through the analysis of these subsets, the study aims to assess the performance of machine learning algorithms in predicting soil liquefaction potential. This study applied ensemble machine learning algorithms, including extreme gradient boosting, adaptive boosting, extra trees, bagging classifiers, light gradient boosting machine, and random forest, to accurately classify the liquefaction potential of fine-grained and coarse-grained soils. A comparison between the genetic algorithm approach for hyperparameter optimization and traditional methods such as grid search and random search revealed that genetic algorithms outperformed both in terms of average test and train accuracy. Specifically, the light gradient boosting machine yielded the best predictions of soil liquefaction potential among the algorithms tested. The study demonstrated that Dataset B achieved the highest learning performance with accuracy of 0.92 on both the test and training sets. Furthermore, Dataset A showed a training accuracy of 0.88 and a test accuracy of 0.84, while Dataset C exhibited a training accuracy of 0.87 and a test accuracy of 0.87. Future studies could build on these findings by evaluating the performance of genetic algorithms on a wider range of machine learning algorithms and datasets, thus advancing our understanding of soil liquefaction prediction and its implications for geotechnical engineering.

Liquefaction phenomena following the February 6th, 2023, Turkiye earthquakes: observation and lessons

Article

Full-text available

Apr 2024
NAT HAZARDS

Post-earthquake scientific investigation is considered as one of the pillars supporting earthquake engineering. On the 6th of February, 2023, two deadly strong earthquakes, which magnitudes were Mw7.8 and Mw7.5, respectively, shook Southern-Central Turkiye, caused significantly large casualties and tremendous economy loss. Through on-site field survey, liquefaction phenomena and liquefaction-induced damage to buildings were observed. The observations are: (1) the consequences of soil liquefaction included sandboils, lateral spreading, ground subsidence and ground failure caused by loss of bearing capacity; (2) in two liquefied areas, lateral spreading was investigated and the spreading displacement ranged from several centimeters to meters, resulting in damage or demolishing of buildings; (3) in Golbasi town, many 6 to 10-story buildings significantly subsided and tilted due to liquefaction-induced loss of ground bearing capacity. Buildings subsided by tens of centimeters to 2 ~ 3 m, and tilted by several degrees to tens of degrees; (4) ground subsidence of tens of centimeters with respect to adjacent buildings was detected. The liquefaction phenomena were compared with those triggered by the 2008 Wenchuan, China, earthquake which maintained similar in magnitude and focal depth. The findings and lessons learnt will enhance the understanding of liquefaction hazard, challenge the current liquefaction countermeasures, and eventually facilitate to improve liquefaction mitigation techniques.

Kum-Silt Karışımlarının Mekanik Özelliklerinin Deneysel Olarak İncelenmesi

Preprint

Full-text available

May 2024

Airat Iunusov

Optimization of Soil Behavior Type Index for Evaluating Liquefaction Potential of Fine-Grained Soils in Adapazarı

Preprint

Full-text available

Apr 2024

This study investigates soil liquefaction potential, specifically focusing on fine-grained soils. We analyze the soil behavior index (Ic) values, emphasizing their role in distinguishing between soils susceptible and resistant to liquefaction. Multiple Ic formulas from the existing literature undergo rigorous examination and refinement to improve accuracy in differentiating soil types. Data from Adapazarı, Türkiye, post-1999 earthquake, and cone penetration test (CPT) results from an acoustic CPTU machine inform the study. Various scenarios, considering 11 literatures from physical properties of the soil, evaluate the potential for liquefaction. The threshold Ic values that best distinguish liquefied and non-liquefied fine-grained soils from each other are determined. Performance assessment identifies the most suitable approach for accurate discrimination. To enhance the chosen scenario, an optimization process determines the optimal coefficients for different Ic formulas. With these optimized Ic formulas, liquefied and non-liquefied soils are analyzed again and the most appropriate threshold values are found. The results show improved accuracy in assessing the liquefaction potential. This study significantly advances liquefaction analysis techniques, introducing novel recommendations, refining criteria, optimizing Ic formulas, and offering insights into nonliquefiable soil classification based on CPT data. The findings aim to improve the safety and effectiveness of geotechnical engineering practices, particularly during seismic events.

Liquefaction at two petrochemical facilities during the February 6, 2023, Kahramanmaras-Turkiye earthquake sequence

Article

May 2024
ENG GEOL

Two petrochemical facilities, located in the Dortyol district of Hatay metropolitan city along the Mediterranean coastline in Turkiye were shaken by the February 6, 2023, M7.8 Pazarcik and M7.6 Ekinozü-Elbistan earthquakes in Kahramanmaras. This study investigates their seismic performances from a seismic soil liquefaction engineering perspective, addressing several key aspects: the i) identification of surface manifestations of seismic soil liquefaction, including soil ejecta, ground deformations, and displacements, ii) examination of subsurface soil characteristics, iii) evaluations of soil liquefaction susceptibility and triggering, iv) analyses of post-liquefaction ground settlements, lateral deformations, and displacements, and v) comparisons of these observed outcomes with predicted results. The assessment results indicated that liquefaction triggering evaluations, based on the SPT- and CPT-based semi-empirical methods of Cetin et al. (2018) and Moss et al. (2006), respectively, were consistent with observed field manifestations of liquefaction or their absence. However, estimated settlements obtained from the liquefaction-induced volumetric settlement procedures of Tokimatsu and Seed (1984), Shamoto et al. (1998) and Cetin et al. (2009), underestimated field settlements. This underestimation was partially explained by the additional contribution of lateral spreading deformations to field settlements, which were not considered due to one-dimensional nature of these volumetric settlement assessments. Additionally, semi- empirical correlations of Hamada et al. (1986) and Youd et al. (2002) consistently overestimated observed lateral displacements. The discrepancies in settlement and lateral spreading displacements are believed to stem from the presence of liquefaction susceptible fine-grained and/or gravelly units, which may have higher residual shear strengths and/or produce higher volumetric strains than typical clean sand layers in case history sites used for developing these semi-empirical methods. The documented field performances and assessment results are hoped to contribute to further developing and refining/calibrating increasingly accurate and reliable empirical and/or analytical methods for assessing hazards related to liquefaction triggering, and liquefaction-induced ground deformations and displacements.

The performance of Karasu Bridge following the February 6, 2023 Kahramanmaras-Turkiye earthquake sequence

Article

Apr 2024
ENG GEOL

LIQUEFACTION-INDUCED DAMAGE IN THE CITIES OF ISKENDERUN AND GOLBASI AFTER THE 2023 TURKEY EARTHQUAKE

Article

Full-text available

Apr 2024

On February 6th 2023 two major earthquakes struck southeast Turkey, M7.7 Kahramanmaras and M7.6 Elbistan, respectively. Unfortunately, due to the impact of these catastrophic events more than 50 000 casualties and 35 000 collapsed buildings have been reported since then. The aim of the study is to demonstrate preliminary site response analysis and assessment of re-liquefaction potential of sites which have been affected by the earthquakes – especially the cities of Iskenderun and Golbasi. Both site-specific areas have clear evidences of liquefaction and lateral spreading events which imply the focus of the presented paper. A series of geophysical MASW and microtremor tests have been performed in order to determine shear wave velocities up to depth of 30 m as well as the fundamental natural frequency of the soil deposits. Moreover, samples have been collected from sand and silt ejecta in order to evaluate some basic physical properties – grain-size curves, specific gravity and plasticity parameters. On the basis of the obtained data seismic classification of the investigated sites according to current design codes has been made and in-depth distance to relatively stiff layer has been assumed. For the sake of evaluating risk of re-liquefaction the widely-used simplified stress-based approach to triggering assessment has been adopted considering some rules of the thumb (e.g., sieve analysis and plasticity properties evaluation). Lastly, post-liquefaction reconsolidation settlement and lateral displacement have been determined in terms of future earthquakes

CPT stress normalization and prediction of soil classification

Conference Paper

Full-text available

Jan 1995

An updated CPT soil classification chart is presented based on a new cone penetrometer resistance normalization techniques

T.4 Earthquake-induced nonlinear shear mass participation factor, rd

Technical Report

Jun 2000

SPT-based probabilistic and deterministic evaluation of seismic soil liquefaction potential

Article

Jan 2003

SPT-based probabilistic and deterministic assessment of seismic soil liquefaction potential. ASCE J. of Geotech. and Geoenvir. Eng. 130(12): 1314-1340, 2004.

BPT-SPT correlations with consideration of casing friction

Article

Jan 1993

The Becker penetration test (BPT), through correlations with the Standard penetration test (SPT), is widely used for foundation design and liquefaction assessment in gravelly soils. Most of the existing correlations, however, do not adequately account for the variable energy output of the diesel hammer used in the Becker system, and shaft friction on the Becker casing is ignored. An alternative and more rational approach to BPT-SPT correlations is presented, based on experimental and numerical studies conducted at the University of British Columbia. The research involves performing SPT, BPT and other in-situ tests at several sites in British Columbia, and includes dynamic measurements of energy transfer in the SPT and BPT. Stress wave measurements and wave equation analyses are used to evaluate the effect of friction on the BPT blow count. New BPT-SPT correlations are proposed which consider the energy transfer in both tests and which explicitly consider casing friction in the BPT. A recommended procedure for estimating equivalent SPT N60 from BPT blow count is presented.

An effective stress model for predicting liquefaction behaviour of sand

Article

Jan 1998

An elastoplastic constitutive model is presented that simulates the liquefaction response of sands in a relatively uncomplicated manner. The model, UBCSAND, is based on the characteristic behaviour of the soil skeleton as observed in laboratory element tests. The model has several key features, including a hyperbolic relationship between stress ratio and plastic shear strain, a flow rule for estimating plastic volumetric strain from plastic shear strain, and the ability to handle anisotropy. The simple framework which describes the observed soil response and forms the basis for the modal is presented. Monotonic and cyclic results are computed using the model and shown to be in good agreement with laboratory element tests. The model is also applied to the Wildlife Site in California and the predictions compared with field measurements from the 1987 Superstition Hills Earthquake.

Liquefaction in silty soils: design and analysis

Article

Jan 1994

The conditions for triggering liquefaction in silty soils and assessing the consequences of liquefaction are reviewed. Recent findings on the evaluation of residual strength and the effects of effective stress on the resistance to liquefaction are presented. The major impact of these findings on the potential costs of remediation are demonstrated. The procedures for estimating the consequences of liquefaction and for evaluating the effectiveness of remedial measures are described by means of case histories.

Changes in Density of Sand Subsoil Caused by the Niigata Earthquake

Article

Jan 1966

Yasunori KOIZUMI

Methods for evaluating residual post-liquefaction ground settlement and horizontal displacement

Article

Jan 1998

During the 1995 Hyogoken-Nambu earthquake, large vertical residual ground settlements and horizontal displacements extensively occurred in liquefied soils behind caisson-type quay walls where lateral movements of the walls towards sea were led by strong seismic shaking. Case studies further show that both significant vertical residual ground settlements and horizontal displacements were also induced by this earthquake in liquefied level sandy ground with a sufficiently large lateral extent where no significant initial shear stresses were present. The mechanism regarding the occurrence of the above phenomena is revealed through new post-liquefaction stress-strain constitutive analysis. Consequently, the residual post-liquefaction ground settlement and horizontal displacement are found to be interdependent of each other, so that neither cannot be determined separately in principle. A new methodology with pragmatic charts is developed for concurrently evaluating the residual post-liquefaction ground settlement and horizontal displacement in liquefiable ground. The prediction results using the proposed method agree with actual observations during the 1995 Hyogoken-Nambu earthquake.

A study of the slip in the Calaveras Dam

Article

A. Hazen

Seismic safety of embankment dams developments in research and practice 1988-1998

Article

Jan 1998

Liam Finn

Developments in the procedures for evaluating the seismic response and safety of embankment dams since the last decennial conference, Earthquake Engineering and Soil Dynamics II, in 1988 are reviewed. There are three major developments. First is the extensive research on the factors controlling the residual strength of potentially liquefiable sands and the selection of strengths for use in design. The second is the use of displacement as a criterion for assessing the seismic performance of embankment dams and planning remediation measures, especially when there is a potential for liquefaction in the dam itself or in the foundation. This requires large strain analysis. The third development is the emergence of seismic risk and reliability analysis as an aid to determining the dominant failure modes of the dam, the probability of occurrence of unacceptable damage, and the associated probabilities of both economic losses and loss of life.

Recent advances in soil liquefaction engineering: A unified and consistent framework

Abstract

Recommended publications

Regional earthquake loss estimation : role of transportation network, sensitivity and uncertainty, a...

C.9 Recent Advances in Soil Liquefaction Engineering and Seismic Site Response Evaluation

T.8 Recent advances in soil liquefaction engineering: a unified and consistent framework

LABORATORY STUDY OF LIQUEFACTION TRIGGERING CRITERIA

SHEAR STRAIN DEVELOPMENT IN LIQUEFIABLE SOIL UNDER BI-DIRECTIONAL LOADING CONDITIONS