Jonathan Stewart Daniel Chu paper

advertisement
VALIDATION OF LIQUEFACTION-RELATED GROUND
FAILURE MODELS USING CASE HISTORIES FROM
CHI-CHI, TAIWAN EARTHQUAKE
by
Daniel B. Chu and Jonathan P. Stewart*
University of California, Los Angeles
Shannon Lee
National Chi Nan University, Puli, Taiwan
J. S. Tsai
National Cheng Kung University, Tainan, Taiwan
P.S. Lin and B.L. Chu
National Chung-Hsing University, Taichung, Taiwan
Robb E.S. Moss
Fugro Inc., Ventura, CA
Raymond B. Seed
University of California, Berkeley
S. C. Hsu
Chao-Yang University, Wufeng, Taiwan
M. S. Yu
Resources Engineering Services, Inc., Taipei, Taiwan
Mark C.H. Wang
Moh and Associates, Taipei, Taiwan
*Corresponding author
Submitted for presentation at:
U.S.-Taiwan Workshop on Soil Liquefaction
Hsinchu, Taiwan, November 2-5, 2003
ABSTRACT
The 1999 Chi Chi, Taiwan, earthquake provides case histories of ground failure and
non-ground failure that are valuable to the ongoing development of liquefaction
susceptibility, triggering and surface manifestation models because the data occupy
sparsely populated parameter spaces (i.e., high cyclic stress ratio and high fines content
with low to moderate soil plasticity). In this paper, we synthesize results from several
large site investigation programs conducted in Nantou, Wufeng and Yuanlin, Taiwan,
and compare the data to susceptibility, triggering and surface manifestation models.
With regard to liquefaction susceptibility, we find components of the well-known
Chinese criteria associated with liquid limit (LL) and water content/LL to be reasonably
well validated by the Taiwan data, but clay fraction criteria and CPT-based criteria to
not be effective. Triggering models are generally validated for ground failure sites, but
the data raise important questions regarding non-ground failure sites whose performance
is not well predicted. Problems were encountered with a liquefaction manifestation
model widely used in practice. Specifically, the model fails to predict non-ground failure
at most free-field sites with high fines content, and fails to predict ground failure for
several building sites with large static driving shear stresses. Accordingly, it appears that
this model, which was developed principally for clean sands, may not be applicable for
soils with high fines content.
Key words: Liquefaction, liquefaction triggering, liquefaction susceptibility, and
surface manifestation of liquefaction, Chi Chi earthquake.
1
INTRODUCTION
The 1999 Mw = 7.6 Chi-Chi Taiwan earthquake triggered numerous significant incidents
of liquefaction in inland alluvial areas and in several coastal hydraulic fills [1]. Due to
significant interest in the available case histories of liquefaction and non-liquefaction, a
series of site investigation programs were undertaken in 2000 by researchers with the
National Center for Research in Earthquake Engineering (NCREE) in Taiwan and in
2001-2002 by the authors with funding from the Pacific Earthquake Engineering
Research Center (PEER). Results of both site investigation programs are synthesized on
the web page http://www.cee.ucla.edu/faculty/Taiwanwebpage/Main.htm.
The objectives of this paper are to (1) document a number of the particularly
significant case histories to emerge from this earthquake, (2) compare the case histories
to established liquefaction susceptibility criteria (e.g., the Chinese criteria), (3) compare
these case histories to state-of-practice [2, 3, 4] as well as probabilistically-based stateof-the-art [5, 6] liquefaction triggering procedures, and (4) compare the case histories to
a widely used model that is used to predict whether liquefaction effects will be manifest
at the ground surface [7]. As will be shown subsequently, the Taiwan liquefaction data
has an important role to play in the ongoing development of empirical liquefaction
assessment methodologies for two principal reasons:

Many of the Taiwan case histories involve high cyclic stress ratios (CSR  0.40.6), where existing data is sparse.

The Taiwan case histories involve primarily high fines content soils, where the
existing data inventory is sparse (i.e., the Youd et al. [4] triggering model is
2
based on only 13 cases with  35% fines content, [2]), although a number of
other studies have been conducted in similar materials since that 1985 study
[e.g., 8, 9, 10].
SITE INVESTIGATION PROGRAM
The site investigation programs by NCREE and PEER resulted in a total of 92 Cone
Penetration Test (CPT) profiles (of which 63 were seismic CPTs) and 98 soil borings
with Standard Penetration Testing (SPT) (typically at 1 to 2 m spacing). The majority of
the NCREE work was performed in the city of Yuanlin, whereas the entirety of the
PEER work and some of the NCREE work was performed in the cities of Nantou and
Wufeng.
Most of the borings/CPTs were limited to depths of 10-30 m. CPT profiling was
performed according to standard techniques (ASTM D 5778-95). For SPT sampling, the
percentage of the total theoretical energy delivered to the split-spoon sampler, or energy
ratio, was controlled by following the procedures in ASTM D6066-98 and ASTM
D1586. We used a safety hammer with a rope/cathead release mechanism, two turns of
the rope around the cathead, standard AW rod, and a 12 cm borehole diameter. Hence,
the energy transmitted to the sampler would be assumed to be 60% if no short-rod
correction was applied. The actual delivered energy was measured for each blow of the
hammer using a rod section instrumented with accelerometers and strain gages [11].
Using the average energy ratio (ER) for each test, we computed the blow-count
normalized to 60% of the theoretical energy, N60.
All retrieved soil samples were subjected to a full suite of laboratory index tests
per ASTM standards including sieve, hydrometer, liquid limit, plastic limit, and water
3
content. Results are presented on boring logs on the aforementioned web page. These
test results were used for liquefaction susceptibility analysis, as discussed below.
Sites selected for subsurface exploration included lateral spread sites, locations of tilted
and/or settled buildings, and locations of no apparent ground failure based on postearthquake reconnaissance. A total of 53 sites were investigated by the authors (38 in
Wufeng and 15 in Nantou) as well as 137 sites by NCREE (20 in Wufeng, 22 in Nantou
and 95 in Yuanlin). Locations of the Wufeng, Nantou, and Yuanlin sites are overlaid on
damage locations in Figures 1, 2 and 3, respectively.
DATA ANALYSIS PROCEDURES
Seismic Demand in Nantou and Wufeng
Strong motion accelerographs (SMAs) are present in Nantou (Station TCU076), Wufeng
(TCU065) and Yuanlin (TCU 110 and TCU 138). In both Wufeng and Nantou, the
SMAs are located within about 1 km of the liquefaction/non-liquefaction sites
considered in this research, and are on generally similar site conditions (young
alluvium). In Yuanlin, the SMAs are located within approximately 5 km of the
liquefaction and non-liquefaction sites. There was no evidence of ground failure in the
immediate vicinity of the SMAs. All SMAs are on the footwall side of the ruptured
Chelungpu fault, as are the subject liquefaction/non-liquefaction sites. The geometric
mean peak horizontal accelerations of the two horizontal components of shaking were
PHA = 0.38g in Nantou and 0.67g in Wufeng. For Yuanlin, stations TCU110 and TCU
138 recorded geometric mean PHA = 0.18g and 0.21g, respectively (a representative
value of 0.2 g was used for subsequent analysis). These PHA values were used to
estimate cyclic stress ratios at the liquefaction/non-liquefaction sites, as discussed below.
4
Liquefaction Susceptibility Analysis
Most of the soils at the investigated sites in Nantou, Wufeng and Yuanlin contain
significant fines (i.e. > 35% passing the #200 sieve). Fine-grained soils require analysis
to evaluate their liquefaction susceptibility. A well-known specification for checking
liquefaction susceptibility is the “Chinese criteria,” originally presented by [12] and restated by [4]. The Chinese criteria specify that liquefaction can only occur if all three of
the following conditions are met: (1) weight fraction smaller than 5m (i.e., “clay
fraction,” CF) < 15%, (2) liquid limit (LL) < 35%, and (3) natural water content (wn) >
0.9LL. More recently, Ref. [13] stated that silty soils are susceptible to liquefaction if
both LL < 32% and the amount finer than 2 m < 10 %, whereas Ref. [10] found from
Adapazari, Turkey, case histories that the Chinese criteria are effective provided the CF
criterion is neglected. As described further below, in this study weight was given to the
LL and wn/LL components of the Chinese criteria during the identification of critical
layers at the subject sites.
Liquefaction susceptibility criteria based on CPT test results are not well
established, although in [3] it is proposed that the Ic parameter can distinguish relatively
granular soils from potentially plastic soils, with Ic = 2.6 being an approximate boundary
between the two. In Ref. [6], it was found from Bayesian analysis of case history data
that Ic correlates poorly with the “clean sand” correction factors needed for CPT-based
liquefaction triggering analysis, which implies that Ic may not be a suitable parameter for
liquefaction susceptibility analysis. Ref. [6] found that for soils with an overburdennormalized tip resistance of qc,1 >  1 MPa, traditional liquefaction is unlikely to occur if
friction ratio Rf >  3%, which implies that Rf may be a preferred parameter to Ic.
5
Liquefaction Triggering Analysis
The liquefaction triggering analysis procedures used here provide an estimate of cyclic
resistance ratio (CRR) based on a measure of penetration resistance (SPT blow count or
CPT tip resistance) normalized to 1.0 atm overburden pressure. The current standard of
practice for CRR analysis consists of well-known SPT and CPT procedures summarized
in [4]. The calculation of seismic demand in terms of a cyclic stress ratio (CSR) in these
procedures is performed using ground surface PHA, effective and total stresses at the
depth of interest, and stress reduction factors (rd) by [14] that are a function solely of
depth.
New liquefaction triggering procedures for SPT and CPT have been presented
by [5, 6]. These procedures differ from the procedure in [4] in that they are based on
different data sets (generally larger and more carefully screened) and are fully
probabilistic. These procedures also use different rd models, which are based on
statistical interpretation of ground response analysis results. These rd models are
sensitive to depth, depth to groundwater, shear wave velocity in the upper 12 m, and
earthquake magnitude.
In the back-analysis of liquefaction/non-liquefaction sites using SPT or CPT
procedures, it is necessary to identify a critical layer having the minimum seismic
resistance to liquefaction triggering. The identification of this “weakest strata” is ideally
performed based on careful study of CPT tip resistance and friction ratio, in conjunction
with a boring log with laboratory index testing to evaluate susceptibility. Shown in
Figure 4 is a data set used to evaluate the location of the critical layer at an example site
that showed evidence of ground failure in the form of lateral spreading. Beginning with
6
the CPT data on the left side of the figure, the critical layer is preliminarily identified as
indicated by the dashed lines based on a combination of low qc (indicating relatively low
density) and low Rf (indicating low plasticity). The index properties from the layer (right
side of figure) are compared to the LL and wn/LL components of the Chinese criteria,
which in this case suggest the layer is susceptible to liquefaction. In the example, a
susceptible zone occurs from 2 to 7.5 m deep, but the relatively high CPT tip resistance,
qc, along with high corrected SPT blow-count (N1)60 at the depth of 4.5 to 7.5 m
preclude it from liquefaction. Accordingly, the critical layer is determined to be at a
depth of 2 to 4.5 m, and data from this layer are used to represent this site for data
comparisons to liquefaction triggering models.
Figure 5 illustrates the identification of the critical layer for an example site that
did not show evidence of ground failure. The critical layer is identified by the dashed
lines based on a combination of low qc and low Rf. The index properties from the layer
(right side of figure) are compared to the LL and wn/LL components of the Chinese
criteria, which in this case suggest the layer is not susceptible to liquefaction. In such
cases, additional layers are sought that might be susceptible, and these layers are used
for subsequent analysis provided they are not at large depth. In the example, a
marginally susceptible zone is identified at about 15 m, but this depth is too great for use
with triggering models. Accordingly, since the critical layer is not susceptible, and
potentially susceptible layers are deep, data from this site would not be included in data
compilations for liquefaction triggering.
Liquefaction Surface Manifestation Analysis
The large amount of case histories and subsurface soil data gathered in this study
7
provide a good opportunity to examine models for surface manifestation of liquefaction.
These models were developed by Ishihara [7], who found that the occurrence of
liquefaction in some layer of a deposit is not necessarily associated with damage to
structures and disruption of the ground surface. Ishihara states: “Only when the
development of liquefaction is sufficiently extensive through the depth of a deposit and
shallow enough in proximity to the ground surface, do the effective of liquefaction
become disastrous, leading to sand boiling and ground fissuring with various types of
associated damage to structures and underground installations.”
Ishihara [7] investigated the conditions under which liquefaction effects are
manifest at the ground surface in terms of the thickness of liquefiable strata and
overlying non-liquefiable strata. Examining case histories from three earthquakes,
Ishihara identified locations with and without surface manifestations of liquefaction,
and borehole data from both regions was compiled. Using available SPT-based
triggering and susceptibility models similar to those described above, judgments were
made regarding the depth and thickness of liquefiable and non-liquefiable strata. In
some cases, the non-liquefiable strata were judged to be so on the basis of soil type (i.e.,
cohesive soils), while in other cases layers were considered not liquefiable because they
were above the ground water table. In all cases, the liquefiable layers were described by
Ishihara as “sand” or “fine sand.” We interpret these descriptions of the soil type as
applying to clean sands or sands with low fines content (and certainly not plastic fines).
A widely-used outcome of the aforementioned analyses is the boundary curves shown
in Figures 9 and 10. These curves suggest that liquefaction effects will generally not
be manifest at the surface (regardless of the thickness of the liquefiable layer) if the
8
thickness of the non-liquefiable layer is > 3 m for PGA  0.2g, and if the thickness is
> 9 m for PGA  0.4-0.5g. It should be emphasized, however, that these results apply
essentially for sandy soils. To our knowledge, their reliability for fine-grained
materials has not been verified.
To compare our case histories with Ishihara’s model, we selected 32 data sets
for which the non-CF components of the Chinese susceptibility criteria are satisfied.
The liquefiable layer, H2, as depicted on Figures 9 and 10, was identified on the basis
of liquefaction susceptibility and triggering criteria (i.e., the layer should be both
susceptible and likely to trigger, e.g., Figure 4). The non-liquefiable layer, H1, was
identified on the basis of soils being unsaturated (usually in the upper 1-2 m) or
clayey (e.g., Figure 5).
DATA SYNTHESIS AND MODEL VALIDATION
Synthesis of Data
Presented in Table 1 are critical layer depths, mean and standard deviation soil
properties within the critical layers, and derived quantities utilized in the triggering
analysis models for sites in Nantou, Wufeng, and Yuanlin. The statistics on soil
properties are evaluated considering both vertical and lateral variability (lateral
variability is considered when more than one boring/CPT are available for a site). Also
shown are brief descriptions of site performance as observed during post-earthquake
reconnaissance. In the following, sites are considered to have “ground failure” when
permanent deformations of the ground surface were observed. It should be noted that
sites without observed ground failure may have had localized liquefaction at depth that
9
was not manifest at the surface.
Validation of Susceptibility Models
Plotted in Figure 6 are the average soil index properties within the critical layers for the
sites in Table 1. The data are plotted as dots for sites with evidence of ground failure,
and as open circles for non-ground failure sites. Figure 6(a) shows the data in LL-wn
space along with the boundary curves associated with the Chinese criteria. In the space
marked as “susceptible” (i.e., LL < 35, wn/LL > 0.9), 71% of the case histories are
ground failure sites (15 of 21), while in the “not susceptible” space 64% (7 of 11) are
non-ground failure sites. While clearly imperfect, the non-CF components of the
Chinese criteria appear to be unbiased, and in general are validated by the Taiwan data.
Figure 6(b) shows the data in LL-CF space. In the susceptible space (CF < 15%),
76% of the case histories are ground failure sites (16 of 21). There are relatively few
data in the “not susceptible” space (CF > 20%), with 3 of 5 case histories being nonground failure sites. However, four additional ground failure sites occur between CF =
15-20%. Taken as a whole, these data suggest that the CF component of the Chinese
criteria is not effective for liquefaction susceptibility evaluations. Similar findings were
reached by [10] using ground failure data in similar soil types in Adapazari, Turkey.
Another important observation relates to the use of CPT-based indices for
analysis of liquefaction susceptibility. In Figure 7, we plot the Ic and Rf of 47 ground
failure and non-ground failure sites having sufficiently low tip resistance (parameterized
either by qc1N or qc1) that liquefaction would be likely to trigger in susceptible materials.
Figure 7(a) shows the data relative to the widely used Ic criteria for soil classification.
Most of the data fall below the Ic = 2.6 threshold, which might be taken to imply
10
liquefaction susceptibility. Within this zone, the data is roughly evenly distributed
between ground failure and non-ground failure sites (51% ground failure), and visual
examination of the data suggests that the Ic = 2.6 threshold has no value at distinguishing
ground failure and non-ground failure sites. The conclusions are similar when the data is
plotted against the Rf = 3% threshold.
The data from several individual sites are illustrative with respect to the poor
performance of the CPT-based indices for assessing liquefaction susceptibility. Consider
for example Site W-N-B8 in Table 1, which has soils with Ic < 2.6 and Rf < 3%
(indicating relatively granular soils and potentially high liquefaction susceptibility), but
which is not susceptible based on the Chinese criteria, and in fact did not show evidence
of ground failure. Conversely, Site W-N-B10 with Ic > 2.6 passes the non-CF
components of the Chinese criteria and had ground failure. In each of these cases it
should be noted that the ground failure involved settlement of rather tall structures (7
stories), and as discussed further below, the static stresses associated with these
structures can play an important role with respect to surface manifestation of
liquefaction.
Liquefaction Triggering
Plotted in Figure 8 is average CSR-penetration resistance data within the critical layers
for sites in Table 1 that pass the LL and wn/LL components of the Chinese criteria. Also
shown are the CRR models discussed previously. In the case of the probabilistic models,
the CRR curves shown apply for a 20% probability of liquefaction. Generally, the upper
band of results at CSR  0.6 are Wufeng sites, sites with CSR  0.4 are in Nantou, and
11
sites with CSR  0.2 are in Yuanlin. Note that most of the data plotted in Figure 8 are
dots (indicating ground failure sites). Most of the non-ground failure sites do not appear
on Figure 8 because the soils are not susceptible to liquefaction based on the LL and
wn/LL components of the Chinese criteria.
The ground failure sites (dots) are generally encompassed by the CRR models,
and it is not possible to judge the relative accuracy of these models based on the Taiwan
data. The non-ground failure sites (open circles) that plot to the left of the CRR lines in
Figure 8 merit additional discussion (i.e., sites W-P-AE, N-P-B1, Y-N-BH7, Y-NBH27). These sites also correspond to several of the open circles within the
“susceptible” space in Figure 6(a). Sites W-P-A-E, Y-N-BH7, and Y-N-BH27 consist
of a 2-3 m clay bed overlying a layer of relatively sandy soil. The critical layers in
these cases were taken as the upper section of sand. Site N-P-B1 consists of 1.5 m of
unsaturated soil overlying 6 m of silty sand, which was taken as the critical layer. All
sites have essentially “free-field” conditions – namely, the absence of tall structures
(local structures at these sites are light, single story buildings). All sites were observed
within 2 to 4 weeks of the earthquake by reconnaissance team members, who report no
evidence of ground failure in the area. Further discussion of these sites is provided
below.
Surface Manifestation of Liquefaction
Figures 9 and 10 show preliminary comparisons of the Ishihara surface manifestation
model to data from Yuanlin and Nantou/Wufeng, respectively. Data from 21 sites is
shown in the figures, each of which has a critical layer that satisfies the LL and wn/LL
components of the Chinese criteria and is likely to trigger based on penetration
12
resistance. In Figures 9 and 10, H1 is taken as the thickness of surface layers unlikely to
liquefy, usually because they are above the water table or consist of soils that are not
susceptible to liquefaction. Parameter H2 is the thickness of the critical layer and
adjacent layers judged to be susceptible and likely to trigger. In the figures, building
sites (triangles) are distinguished from free-field sites (circles). For buildings, the
number of stories and number of basement levels are indicated. As before, the symbols
are solid for ground failure sites and open for non-ground failure sites.
The model generally adequately encompasses the free-field ground failure sites,
which all plot to the left of Ishihara’s boundary lines. The model does not adequately
represent some free-field non-ground failure sites in each city (Y-N-BH27, N-P-B1, WP-A-E). Problems are also encountered with the building sites, many of which plot to the
right of the threshold curves, yet experienced foundation settlement/ground failure.
Discussion
Much of the Taiwan ground failure/non-ground failure data presented in Figures 6-10 is
consistent with existing knowledge of liquefaction phenomena as implemented in
models that are widely used in engineering practice. A significant exception to this
encouraging result is a series of free-field, non-ground failure sites that would be
expected to have liquefaction susceptibility, would be expected to trigger, and generally
would be expected to have liquefaction effects manifest at the ground surface. These
sites are labeled in Figures 8-10 as W-P-A-E, N-P-B1, Y-N-BH-27, and Y-N-BH-7.
An important question to ask regarding these non-ground failure sites is whether
they may have in fact experienced liquefaction. As shown in Figures 9-10, all except YN-BH-7 would be expected to have surface manifestation of liquefaction. We speculate
13
that the absence of ground failure does not mean that liquefaction within the critical
layers did not occur. Liquefaction in the critical layers may not have been manifest since
the layers are overlain by non-liquefiable strata, the sites lacked driving static shear
stresses that could mobilize significant ground failure effects through the overlying
strata, and the fines contents of the liquefiable soils at depth are considerably higher than
what appears to have been considered by Ishihara [7].
In attempting to understand what occurred at the aforementioned sites,
consideration of driving static shear stresses may be particularly important. Those nonground failure sites have level ground and free-field conditions, and hence lack driving
static shear stresses. However, a majority of the sites that did experience ground failure
had significant static shear stresses. Accordingly, questions are raised regarding the
appropriate site metric for assessing liquefaction manifestation. Ishihara’s existing
criteria are based on the relative thickness of liquefiable and non-liquefiable strata,
which are undeniably important factors. However, driving static shear stress may also be
significant, at least in high fines content materials similar to those present at the
investigated sites in Taiwan. This topic is being actively investigated at present.
The Taiwan data also raise corollary questions regarding how a site is classified
within the “ground failure” or “non-ground failure” categories as a result of observations
from post-earthquake reconnaissance, e.g., for the development of liquefaction
triggering models. While there is no question about assigning “dots” to sites with ground
failure, a “circle” denoting non-ground failure might actually have experienced ground
failure had static stress conditions been different (i.e., if a building was present). This is
further complicated by the fact that the existing CRR models presumably apply for a
14
zero static shear stress condition, but for the reasons discussed above, the development
of triggering models based on such protocols is likely unconservative for the high fines
content materials commonly encountered in Nantou, Wufeng, Yuanlin and many other
areas. These issues warrant further research in the development of next-generation
triggering models.
SUMMARY OF FINDINGS
Case histories of ground failure and non-ground failure from the Chi-Chi earthquake are
important for the ongoing development of liquefaction susceptibility and triggering
models because the effected soils have large fines contents and marginal plasticity
levels, and because the earthquake generated large CSRs in these soils. Prior data sets
had a paucity of data for such conditions. Comparisons of the data to models indicate
that:
1. The CF component of the Chinese criteria appears to be less reliable than the LL
and wn/LL components. This is similar to the findings of Sancio et al. [10] for
soils in Adapazari, Turkey.
2. CPT-based indices appear unreliable for evaluating liquefaction susceptibility. It
is noted that CPT indices for susceptibility have not been formally proposed in
Refs [3, 6]. Nonetheless, parameter Ic has been applied in the literature for this
purpose, and this practice is not recommended.
3. Ground failure sites from Wufeng, Nantou and Yuanlin are generally
encompassed by available CRR models (i.e., the CSR values generally plot
above the CRR line).
15
4. Some non-ground failure sites plot above the CRR lines, which raises questions
regarding the nature of the site performance (did the site liquefy in a manner that
was not manifest at the surface?) and the degree to which the apparently good
site performance may be an adequate predictor of future performance at other
sites. These questions remain unanswered, but are important for the ongoing
development of robust empirical liquefaction triggering models.
5. Problems were encountered with a liquefaction manifestation model widely used
in practice. Specifically, the model fails to predict non-ground failure at most
free-field sites with high fines content, and fails to predict ground failure for
several building sites with large static driving shear stresses. Accordingly, it
appears that this model, which was developed principally for clean sands, may
have limitations for soils with fines.
ACKNOWLEDGEMENTS
This project was sponsored by the Pacific Earthquake Engineering Research Center’s
Program of Applied Earthquake Engineering Research of Lifeline Systems supported by
the State Energy Resources Conservation and Development Commission and the Pacific
Gas and Electric Company. This work made use of Earthquake Engineering Research
Centers Shared Facilities supported by the National Science Foundation under Award
#EEC-9701568. The support of the California Department of Transportation’s PEARL
program is also acknowledged. In addition, data from the National Center for Research
on Earthquake Engineering (NCREE) through a memorandum of understanding
between NCREE and PEER was used in this research and is acknowledged.
16
REFERENCES
[1] Stewart, J.P.: coordinator. Chapter 4: Soil liquefaction. Chi-Chi, Taiwan Earthquake of
September 21, 1999 Reconnaissance Report. In: Uzarski J, Arnold C, editors.
Earthquake Spectra, Supplement A to Vol. 17. 2001, 37-60.
[2]
Seed, H.B., Tokimatsu, K., Harder, L.F., and Chung, R.M. The influence of SPT
procedures in soil liquefaction resistance evaluations. Journal of Geotechnical
Engineering, ASCE, 1985; 111(12): 1425-1445.
[3]
Robertson, P.K. and Wride, C.E. Evaluation cyclic liquefaction potential using he
cone penetration test. Canadian Geotechnical Journal, 1998, 35(3): 442-459.
[4]
Youd, T.L., Idriss, I.M., Andrus, R.D., Arango, I., Castro, G., Christian, J.T.,
Dobry, R., Finn, W.D., Harder, L.F., Hynes, M.E., Ishihara, K., Koester, J.P., Liao,
S.S.C., Marcuson, W.F., Martin, G.R., Mitchell, J.K., Moriwaki, Y., Power, M.S.,
Robertson, P.K., Seed, R.B., and Stokoe, K.H., II. 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, ASCE, 2001; 127 (10): 817-833.
[5]
Seed, R.B., Cetin, K.O., Moss, R.E.S., Kammerer, A.M., Wu, J., Pestana, J.M., and
Riemer, M.F. Recent advances in soil liquefaction engineering and seismic site
response evaluation. In: Proceedings of 4th International Conference of Recent
Advance in Geotechnical Earthquake Engineering and Soil Dynamics, 2001, Paper
SPL-2.
[6]
Moss, R.E.S. and Seed, R.B. Probabilistic evaluation of seismic soil liquefaction
17
potential using CPT. In: Proceedings of 8th US-Japan Workshop on Earthquake
Resistant Design of Lifeline Facilities and Countermeasures against Liquefaction.
2002, Tokyo, Japan.
[7]
Ishihara, K. Stability of natural deposits during earthquakes. In: Proceedings of the
11th Int. Conference of Soil Mechanics and Foundation Engineering, San Francisco,
CA, 1985, Vol. 1.: 321-376.
[8]
Holzer, T. L., Bennett, M. J., Ponti, D. J., and Tinsley, J. C. III. Liquefaction and soil
failure during 1994 Northridge earthquake. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, 1999; 125 (6), 438-452.
[9]
Boulanger, R. W., Mejia, L. H., and Idriss, I. M. Liquefaction at Moss Landing during
Loma Prieta earthquake. Journal of Geotechnical and Geoenvironmental Engineering,
ASCE, 1997; 123 (5), 453-467.
[10] Sancio, R.B., Bray, J.D., Stewart, J.P., Youd, T.L., Durgunoğlu, H.T., Önalp, A., Seed,
R.B., Christensen, C., Baturay, M.B., and Karadayılar, T. Correlation between ground
failure and soil conditions in Adapazari, Turkey. International Journal of Soil
Dynamics and Earthquake Engineering, 2002, 22 (9-12): 1093-1102.
[11] Abou-Matar, H. and Goble, G. SPT dynamic analysis and measurements. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE, 1997; 123 (10): 921-928.
[12] Seed, H.B. and Idriss, I.M. Ground motions and soil liquefaction during earthquakes.
Earthquake Engineering Research Institute Monograph, 1982, Oakland, CA.
[13] Andrews, D.C.A. and Martin, G.R. Criteria for liquefaction of silty soils. In:
Proceedings of 12th World Conference on Earthquake Engineering, New Zealand,
18
2000, Paper No. 0312.
[14] Seed H.B. and Idriss I.M. Simplified procedure for evaluating soil liquefaction
potential. Journal of Soil Mechanics and Foundations Division, ASCE, 1971, 97(9):
1249-1273.
19
Tsa
o
-Hu
R
iver
W-N-B2
(Route 127)
W-N-B5
W-N-B3
W-N-C9
W-N-B1
Ground Surface
Rupture
W-N-B13
kR
ive
r
W-N-B11
W-P-D
Dry
C
ree
W-N-C16
N
W-N-B7
0
W-P-E
100 200 300 400m
TCU 065
W-N-B8
W-N-B4
TCU 065
W-P-A
Strong Motion Accelerograph
Settled, Tilted Buildings
Lateral Spread
W-P-B
W-N-B9
Sediment Boils
W-P-C
Do
st Cre
ve N e
ek
Surface Fault Rupture
NCREE Investigation Site
PEER Investigation Site
Fig. 1. Map of Wufeng showing ground failure zones and locations of investigated
sites
20
Mao-Lo River
Chia-Ping River
Hwy 3
Hwy 3a
Mao-Lo River
TCU 076
LINE 2
Hwy 14b
N
Hwy 3a
Scale
0
100
200
300
400
500 m
LINE 1
Hwy 3
TCU 076
Strong Motion Station
Principal Ground Failure/
Subsidence Areas
Damaged Levees
Liquefaction and/or Sand Boils
Reported by NCREE
PEER investigation site
NCREE investigation site
Fig. 2. Map of Nantou showing ground failure zones and locations of investigated
sites
21
Railroad
0 100
500
1000M
Tk
1
Ha
Tk
Da-Tung
Ha
1
Ha
TCU 110
Yuanlin
River
Shih-Shuan
-Bay River
Yuan-Suai
Bar-Bo
River
Road
Tk
Ha
1
TCU 110
Central Weather Bureau's
Digital Accelerograph Station
Holocene Alluvial
Sediments
Tk
Ha
Pleistocene Toukoshan
Formation
Sand Boils (large scale)
Reported by MAA
Subsidence Area (liquefaction
induced settlement) Reported
by MAA and/or The Authors
Ha
Liquefaction Sites (with
settlement and/or sand boils)
Reported by NCREE
Limit of Yuanlin Township
NCREE investigation site
PEER investigation site
Fig. 3. Map of Yuanlin showing ground failure zones and locations of investigated
sites
22
WCS-1
(56.85 m MSL)
WCC-1
(56.85 m MSL)
R f (%)
qc (MPa)
0 5 10152025 0 2 4 6 8 10 2
0
Ic
3
Moisture
Content (%)
(N1 )60
0 1020304050 0 5 10 15 20 25 30 35 40
4
2.6
0
0
Fill
CL
Critical layer
PI
5
SM
5
5
10
10
15
15
20
20
90% LL
LL
SM-SC
Depth (m)
10
15
SM
SM
ML
20
CL-ML
LL
CL
25
SM
25
25
In-situ water contents
30
30
30
Fig. 4. Example data set from Wufeng- Site C showing identification of critical
layer for a liquefied site.
23
NAS-1
(90.39 m MSL)
NAC-1
(90.39 m MSL)
qc (MPa)
0 5 10 15 0
R f (%)
2 4
Ic
3
62
4
0
2.6
0
(N 1 )60
10
20
Moisture
Content (%)
30 0 5 10 15 20 25 30 35 40
0
0
Fill
PI
5
Depth (m)
ML
Critical
Layer
Critical
Layer
90% LL
5
5
10
10
15
15
20
20
ML
10
LL
ML
15
SM
20
In-situ water contents (wn)
Non-liquefaction site
50
50
40
40
Liquid Limit
Liquid Limit
Fig. 5. Example data set from Nantou showing identification of critical layer for a
non-liquefied site.
30
20
Not susceptible
30
20
Susceptible
Not susceptible
10
10
Susceptible
(a)
(b)
0
0
0
10
20
30
40
50
Clay Fraction (% < 5 m)
60
0
10
20
30
40
50
Wn (%)
non-ground failure
ground failure
Fig. 6. Ground failure and non-ground failure sites plotted relative to mean soil
index properties
24
60
10
120
100
Relatively
granular
Relatively
fine-grained
qc1 (MPa)
80
qc1N
Liquefaction
not likely
8
60
(a)
6
4
(b)
40
2
20
0
1.6
0
2
2.4
2.8
0
3.2
2
4
Rf (%)
Ic
Non-ground failure
Ground failure
Fig. 7. Ground failure and non-ground failure sites plotted relative to mean CPTbased soil index properties
25
6
Seed et al (2001)
Youd et al. (2001)
0.8
0.8
W-P-A-E
W-P-A-E
0.6
CSR
CSR
0.6
N-P-B1
0.4
N-P-B1
0.4
Y-N-BH27
0.2
Y-N-BH27
0.2
Y-N-BH7
Y-N-BH7
0
0
0
10
20
30
40
50
0
Corrected Blow Count, (N1)60,C.S.
10
20
30
40
50
Corrected Blow Count, (N1)60,CS
ground failure site
non-ground failure site
Moss and Seed (2002)
Robertson and Wride (1998)
0.8
0.8
W-P-A-E
W-P-A-E
0.6
CSR
CSR
0.6
0.4
0.2
0.4
0.2
0
0
50
100
150
200
0
250
0
Corrected CPT Tip Resistance, qC1N
5
10
15
20
qC,1mod
Fig. 8. Ground failure and non-ground failure sites plotted relative to existing
liquefaction triggering models
26
25
10
Not Susc.
H1
Susceptible
H2
Thickness of liquefiable sand layer, H2 (m)
PHA = 0.2 g
8
Not Susc.
Y-N-BH27
6
Yuanlin
4
Non-ground failure,
free-field
Ground failure,
building settlement
(No. floors/No. bsmt
levels)
Y-N-BH7
2
(9F/1B)
(12F/1B)
0
0
2
4
6
8
Thickness of surface layer, H1 (m)
Fig. 9. Ground failure and non-ground failure sites in Yuanlin plotted relative to
Ishihara [1] surface manifestation model
27
10
12
Thickness of liquefiable sand layer, H2 (m)
10
Not Susc.
H1
Susceptible
H2
Not Susc.
8
6
PHA = 0.4 g
N-P-B1
(7F/0B)
4
(5F/0B)
W-P-A-E
Wufeng and Nantou
(4F/0B)
2
0
(4F/0B) (4F/0B)
Non-ground failure,
free-field
Ground failure,
free-field
Ground failure,
building settlement
(No. floors/No. bsmt)
(3F/0B)
(3F/0B) (3F/0B)
0
2
4
6
8
10
12
Thickness of surface layer, H1 (m)
Fig. 10. Ground failure and non-ground failure sites in Nantou and Wufeng plotted
relative to Ishihara [1] surface manifestation model
28
Table 1. Inventory of data at selected sites in Nantou, Wufeng and Yuanlin (mean  standard deviation values given for data
fields)
Boring/CPT
No
No. Borings
No. CPT
Youd et al. (2001) procedures
W-P-A-E
1
3
3-6.5
3.1
0.5
W-P-A-W
3
6
10-15
7.0
2.1
W-P-B
1
5
2 - 5.3
W-P-C
2
13
0.5 - 6
W-P-D
1
1
1-3.3
W-P-E
1
1
N-P-A1
1
1
N-P-A2
1
3.8-5.5
15
N-P-A3
1
1-2
10
N-P-A4
1
1.5-2.5
5
N-P-B1
1
1.6-7.3
11
N-P-B2
1
1.7-3
8
N-P-B3
1
3-5.5
29
N-P-B4
1
2-6.3
15
N-P-B5
1
3.2-6.5
19
N-P-C
W-N-B1
2
1
W-N-B3
1
0.6-3.1
3
W-N-B4
1
1-2.5
6
W-N-B5
1
1.5-4.5
13
W-N-B6
1
0.2-4
11
W-N-B7
1
1
0.5-5
W-N-B8
1
1
5-9.5
W-N-B9
1
W-N-B10
1
W-N-B11
1
1.6 - 3.6
10
W-N-B12
1
8 - 14
17
W-N-B13
1
1 - 3.5
4
W-N-B14
1
1-2
15
3
ZC (m)
qC (Mpa)
 2.4
 4.7
2.2  1.3
2.6  2.1
Rf (%)
 0.5
 1.2
1.2  5.3
2.4  2.0
1.5-3.4
1.5-2.5
2-6
W-N-C9
1
1 - 2.8
W-N-C15
1
2.2 - 2.8
W-N-C16
1
2.5 - 5.5
 9.9 24.5  8.0
 22.8 22.2  2.5
28.5  8.2
17.6  8.5
24.5  4.9
35.3  0.6
wn/LL
9
4
 12
9 5
5 2
13  4
39.0
1.0
31.6
0.9
1.3
4.8
 0.2
 2.9
 3.1
5.7  2.4
4.1
1.9
1.2
 0.3
3
2


6

 10
1
9
0



6
7
4
7

1

4
1

66.7
 1.1
 0.6
1.3  0.7
1.6
1
1-6
8
13
32
11
16
7
1.4
 0.9
 0.3
0.7  0.2
2.2  1.6
1.4  0.4
0.8
1.8
 1.0
 0.5
1.9  0.9
2.8  1.1
5.1  0.7
13
1.0
 0.1
 0.2
 0.3
< 5 m (%)
CSR
(N1)60,CS
qC1N
 29.5
 39.7
29.8  17.3
39.2  28.5
CSR
IC
10.5
 0.7
 4.0
3.5  2.1
14.0  12.7
12.7  0.6
0.65
 0.03
 0.03
0.57  0.03
0.59  0.10
0.43  0.00
16
2.2
0.62
0.65
5
 13
15  4
8 3
20  5
40.3
6.7
27
58.4
2.4
0.49
10
N-N-B3
1
3 - 9.4
16
N-N-B4
1
N-N-B5
1
0.5 - 7.5
8
N-N-B6
1
1.5 - 11
20
N-N-B7
1
N-N-B8
1
0-2
4
N-N-B9
1
7 - 11
33
N-N-B10
1
0 - 5.5
10
N-N-B11
1
1.8 - 5.8
5
N-N-B12
1
1.5- 4
8
N-N-B13
1
1
1.8 - 5
26
N-N-B14
1
1
1 - 8.5
16
N-N-C2
1
3-5
N-N-C3
1
1.5 - 3.5
N-N-C13
1
10-11.5
0.8-11.5
7.1
 3.3
 1.3
2.4  0.9
3.8  0.4
4.8
37.4
50.0
4.0
21.3
19.0
10.0
10.0
22.0
76.0
67.5
65.3
34.5
54.7
38.0
45.5
45.0
40.5
34.0
21.0
30.0
20.0
26.7
22.0
21.5
38.0
25.0
0.7
1.0
0.9
1.0
1.0
0.9
1.0
0.8
0.8
 0.1
 0.1


 0.2

 0.2

 0.1
 0.0
 3.5


 6.7

 0.6
 4.7
 2.1
 2.2



 4.9
 6.4
 4.2
 7.0

 11.3

 1.6
 2.8

0.39
 0.02
 0.01


 0.04

 0.01
 0.04
 0.03
 0.02



 0.03
 0.03
 0.01
 0.01

 0.04

 0.01
 0.11

 0.06
 0.01
 0.04
 0.00

 0.02
 0.01
 0.05
 0.01
 0.01

 0.00
 0.03
 0.02
 0.02

 0.07
 0.00
 0.03
 0.00
15
4
1


7

 10
4
 11
3
19.6
8.0
3.5
4.0
9.0
8.0
13.0
1.3
7.0
4.5
3.8
4.0
6.0

27.4 
29.9  10.7
31.8

0.9 
0.9  0.2
0.8
26.0
17.5
30.5
7.0
26.7

0.6
16.0
9.0
23.8
 0.4
0.9
 0.3
12.0
9.0
24.5
 0.8
0.8
 0.0
16.0
9.0
7.0
0.7
1.2
 0.5
 0.9
 0.6
2.7  0.8
1.6  0.7
0.6
0.39
0.28
0.33
0.38
0.34
0.38
0.37
0.37
0.38
0.55
0.57
0.43
0.62
0.45
0.44
0.75
0.43
0.65
0.64
0.54
0.57
0.43
0.56
1 - 2.5
 2.5
35.0
23.0
 3.0
 1.4


 3.2

 3.3

 1.4
0.44
1
3.9
14.0
39.0
0.55
N-N-B2
2.4 - 8
10.0
 12.6
 9.9


 8.2

 1.0
 9.1
 0.6
 0.5



 31.8
 30.5
 20.5
 17.6

 33.2

 1.7
 5.7

0.5
1
1
LL
22
N-N-B1
1
FC (%)
 0.4
 0.4
2.4  0.4
2.5  0.4
0.52
0.62




(N1)60,CS
qC1 (Mpa)
DqC (Mpa)
5
 13
12  5
7 3
16  5
 3.8
 4.1
3.3  2.9
3.7  2.6
 0.4
 1.2
0.8  5.6
2.1  1.9
CSR- Cetin
qC1,mod (Mpa) Probabilistic Field Observations
 3.8
 3.9
4.1  8.0
5.8  2.5
 0.13
 0.17
0.52  0.13
0.60  0.12
0.36  0.11
0.01
13
4.6
0.2
4.9
0.61
No ground failure
0.04
26
6.0
1.6
7.6
0.49
Building settlement
0.11
0.12
Lateral spread
Lateral spread
No ground failure
Building settlement
1.5 - 7
0 - 2.3
1
(N1)60
Seed et al. (2001) and Moss and Seed (2002) procedures
7
7
17
1

 11
0
4
4
6

 35
7
1
6

4
94.7
26.0
17.8
33.5
65.0
25.5
40.3
40.0
27.7
24.7
92.5
87.0
76.0
19.8
 1.5

 1.7
 12.0
 39.0
 11.7
 30.5

 16.2
 1.2
 9.2
 4.2

 12.2
37
8
0.8
 0.1
57.7
6.0
8.8
9.0
15.0
6.7
13.5
24.9

0.9

10.0
10.3
6.3
3
24.7  2
21.7 
34.1
 0.0
1.0  0.1
0.8 
1.0
20.0
13.0
21.0
5.0
 5.1

 1.5
 2.8
 14.7
 2.3
 19.9

 7.6
 0.6
 7.1
 8.5

 1.8
0.25
0.25
0.30
0.25
0.31
0.37
0.43
0.27
0.23
0.28
0.30
0.26
0.30
0.35
0.24
0.30
0.37
29
17
12
11
18
14
29
19
23
13
7
12
20
18
17
24
13
19
17
22
9
20


7
8
2
9

0

5
2

 2.3
15
21
13
14
26
24
10
41
15
11
15
36
21
2

 11
1
4
5
7

 39
8
1
8

2
 0.0
0.40
 0.01
14
5
1.8
 0.2
1.4
 0.3
3.2
 0.5
0.39
0.40
11
8
15
12
30
17
21
73.3
 62.8
2.1
 0.5
0.36
0.52
 0.10
 0.11
11
4
11
19
16
 35.8
64.4  26.4
54.8
 0.3
2.2  0.3
2.3
0.43
0.66
 0.02
 0.03
14
21
10
19.3
 12.0
2.7
 0.2
0.56
 0.01
17
14
26
6
18
 5.7
10.5  3.6
30.1  21.7
18.2  7.7
13.1
13
2.7
 0.2
2.9  0.2
2.7  0.2
3.0  0.2
2.6
0.56
0.42
0.54
0.24
0.25
0.29
38.4
 22.1
2.2
 0.3
0.24
0.29
0.32
73.3
 31.0
2.1
 0.3
0.42
0.27
0.21
0.27
0.29
0.26
0.29
0.32
 12.0
39.7  13.7
34  4.2
53.9
 0.2
2.5  0.2
2.4  0.1
2.0

















0.06
0.01
0.03
0.01
14
12
0.02
18
0.00
10
0.03
14
0.02
23
0.01
22
7
0.01
38
0.03
12
0.02
12
0.02
15
38
0.05
19


7

 10
2
 11



5
6
2
7

2

5
2




2

 12
1
6
4
7

 38
8
1
9

3
0.42
0.33
0.40
0.34
0.37
0.39
0.39
7.3
 6.1
0.6
 1.0
7.9
 5.7
0.38
0.52
0.52
0.50
0.59
0.37
 3.3
6.2  2.5
5.2
 0.6
0.9  0.8
1.1
 3.4
7.1  2.3
6.4
0.38
0.66
0.31
1.9
 1.1
1.4
 1.1
3.3
 1.7
0.56
0.62
0.34
0.61
0.26
 0.9
1.0  0.3
2.7  2.0
1.7  0.6
2.1
 0.5
1.4  0.9
2.5  1.2
3.7  0.6
0.0
 0.8
2.4  0.9
5.2  2.1
5.4  1.0
2.1
0.57
0.42
0.54
0.19
0.25
0.29
9.0
 2.1
0.2
 0.4
3.9
 2.1
0.22
0.28
0.35
7.3
 3.1
0.7
 0.9
7.9
 3.1
0.96
0.25
0.20
0.26
0.27
0.24
0.31
0.35
 1.2
3.5  1.2
8.8  0.4
5.4
 0.5
1.9  0.7
8.6  0.6
0.1
 0.9
5.4  1.0
7.9  0.6
5.5
0.21
0.30
0.27
 0.15
 0.11
 0.14
 0.09
 0.15
 0.10
 0.11
 0.13
 0.12
 0.12
 0.11
 0.16
 0.12
 0.16
 0.15
 0.15
 0.19
 0.15
 0.18
 0.15
 0.12
 0.18
 0.09
 0.15
 0.08
 0.13
 0.08
 0.08
 0.10
 0.07
 0.13
 0.15
 0.02
 0.12
 0.06
 0.12
 0.09
 0.08
 0.10
 0.15
 0.06
 0.09
 0.09
No ground failure
Building settlement
Building settlement
Building settlement
No ground failure
Building settlement
Building settlement
No ground failure
No ground failure
Lateral spread
Building settlement
Sand boils
No ground failure
Building settlement
No ground failure
Sand boils
No ground failure
Sand boils
Sand boils
Sand boils
Building settlement
Building settlement
Sand boils
Sand boils
Sand boils
No ground failure
No ground failure
Sand boils
Sand boils
Sand boils
Sand boils
Lateral spread
Lateral spread
Sand boils
No ground failure
Sand boils
Building settlement
Sand boils
Building settlement
Sand boils
Sand boils
Sand boils
No ground failure
Table 1. Continue
ZC (m)
Y-N-C1
1
8.5-12.0
Y-N-C3
1 10.0-13.0
7.4
Y-N-C5
1
6.5-8.0
6.9
Y-N-C7
1
6.5-9.5
5.1
Y-N-C8
1 10.0-11.0
2.8
Y-N-C9
1
9.0-11.5
6.7
Y-N-C10
1
3.5-6
2.7
Y-N-C11
1
5.0-7.0
2.2
Y-N-C13
1
5.5-8.0
2.7
Y-N-C15
1
7.5-12.0
3.9
Y-N-C16
1
5.5-7.5
2.5
Y-N-C28
1
6.0-8.0
4.2
Y-N-C36
1
12.5-30
10.0
Y-N-C37
1
6.0-8.0
4.0
Y-N-C38
1 12.0-15.0 11.8
Y-N-C42
1 10.0-13.0
6.7
Y-NA-C44
1
11-14.5
9.5
Y-N-C2
1
2.5-3.5
2.1
Y-N-C4
1
3-5.5
3.1
Y-N-C19
1
4.5-5.5
2.2
Y-N-C22
1
2-8.5
3.5
Y-N-C23
1
3.0-5.0
3.1
Y-N-C24
1
2.0-8.0
2.7
Y-N-C25
1
2.5-4
2.6
Y-N-C31
1
2.-3.5
3.6
Y-N-C32
1
4.5-7.5
3.3
Y-N-C35
1
2.-3.0
2.9
Y-N-C43
1
6.5-8.5
6.4
2.5-4.0
qC (Mpa)
4.1
 1.5
 1.5
 1.9
 1.5
 1.0
 2.4
 0.6
 0.5
 1.0
 1.1
 0.6
 0.8
 1.8
 0.9
 3.4
 1.2
 3.0
 1.0
 0.7
 0.2
 2.0
 0.7
 1.2
 0.4
 0.5
 1.0
 0.5
 1.0
Rf (%)
0.9
0.5
0.5
0.3
1.2
0.4
0.8
1.3
1.2
1.3
0.9
0.5
0.5
1.1
0.3
0.3
0.7
0.7
1.3
1.1
0.1
1.1
0.1
0.2
0.1
0.7
0.4
0.4
(N1)60
FC (%)
LL
wn/LL
< 5 m (%)
 0.4
 0.7
 0.7
 0.2
 0.4
 0.2
 0.5
 0.4
 0.5
 0.6
 0.5
 0.3
 0.2
 0.4
 0.2
 0.2
 1.3
 0.7
 0.4
 0.2
 0.1
 0.3
 0.1
0
 0.1
 0.3
 0.3
 0.3
CSR
0.2
0.2
0.19
0.21
0.22
0.22
0.18
0.17
0.22
0.22
0.18
0.21
0.15
0.21
0.2
0.19
0.2
0.22
0.22
0.22
0.2
0.19
0.2
0.13
0.13
0.22
0.17
0.19
Y-N-BH-3
1
7
18
Y-N-BH-5
1
1-5.5
Y-N-BH-7
1
2.5-4.5
Y-N-BH-10
1
5.0-9.7
Y-N-BH-12
1
5.5-7.5
Y-N-BH-14
1
2.0-4.0
Y-N-BH-15
1
5.0-8.0
Y-N-BH-17
1
10.5-17
Y-N-BH-18
1
2-3.5
5
47
Y-N-BH-21
1
1-2.3
3
7
Y-N-BH-25
1
4.5-6.3
7
Y-N-BH-26
1
1.5-3.2
6
Y-N-BH-27
1
2.0-8.0
5
Y-N-BH-28
1
4.0-8.0
5
Y-N-BH-29
1
2.0-8.0
7
Y-N-BH-30
1
2.0-8.5
9
Y-N-BH-31
1
5.0-9.0
4
Y-N-BH-32
1
6.0-13.0
Y-N-BH-35
1
1-6.5
9
Y-N-BH-39
1
3-5.5
5
Y-N-BH-40
1
1-5.0
15
Y-N-BH-41
1
9.5-14
19
Y-N-BH-43
1
0-6.5
4
Y-N-BH-44
1
2.0-10
7
Y-N-BH-45
1
4-9.7
6
Y-N-BH-46
1
1.5-8
5
Y-N-BH-47
1
2.2-4.5
5
2
2
10  2
6 3
6
62.5
4
37.5
4
61.5
27
1
15  4
5
19
11.7
1
55.5
26
76.5
4.0
 19.1
 40.3 30.1
 4.73
 40.3

 0.71
 10.6
 27.6
63.5
45.3
44.3
33.4
35.3
17.5
49.3
 25.5 32.8  1.63
 14
 41.6
 24.3
 18.2
 2.08
 29.8
0.9
 0.0
6
1
2
2
0
2
1
28
10.7
88.8
51
61.3
75
60
54.9
27.4
66.4
35.2
25.2
32.6
40.4
28.3
49.2
70.3
46.2
101
60.4
85.1
36.6
46.6
29.7
43.9
43.7
34.7
33.3
48.9
41
49.2
67.6
11
1.9
1.8
1.9
2.5
1.8
2.2
2.5
2.4
2.3
2.4
2.0
1.9
2.2
1.6
1.9
1.8
2.2
2.3
2.4
2.2
2.2
2.3
2.2
1.8
2.2
2.0
1.8
CSR
6.5
7.4
5.5
2.7
6.4
3.6
2.5
3.2
3.9
2.8
5.0
6.3
4.5
8.7
5.7
8.0
4.5
4.5
3.0
4.6
4.3
3.6
3.4
6.9
4.2
5.8
6.7
0.2
0.16
0.13
0.13
4.0
0.16
 2.8
6.8  5.0
0.23
0
0.17  0.01
11
5.0
0.22
11
0.2
0.17
3
0.18
5.0
0.23
15
0.19
1.0
0.19
10
0.17
6.0
4.8
5.0
4.0
5.0
6.8
 2.3
 7.2
 2.2
 2.0
 2.0
 1.4
 5.7
2.0
 1.14
31.1  0.21
1.0
 0.0
1.0  0.0
17.3
33.7
1.0
11.0
35.3
1.0
13.3
37.8
0.9
15.0
11.6
0.21
0.18
0.17
0.2
0.16
0.2
0.15
9
 0.01
 0.01
 0.02
 0.01
 0.01
0
 0.02
0.21
 5.0
 1.0
 6.7
 9.4
 7.8
 10.4
 12.7
30
0.17
0.21
0.23
0.2
0.21
0.2
0.15
0.15
2
20  5
11
11
11
13
9
23
15
1
1
1
2
2
5
3
11
 0.03
 0.01
 0.03
 0.02
 0.01
 0.02
 0.02
20
20
10
12
12
11
10
0.17
0.12
0.18
0.15
0.15
0.17
0.12
0.13
0.14
0.19
5
1
2
1
0
3
1
qC1 (Mpa)
3.8
0.21
2
2
12  3
12  4
(N1)60,CS
 0.2
 0.3
 0.3
 0.2
 0.3
 0.2
 0.2
 0.1
 0.2
 0.2
 0.2
 0.2
 0.1
 0.2
 0.2
 0.1
 0.4
 0.4
 0.1
 0.1
 0.4
 0.1
 0.3
 0.1
 0.1
 0.2
 0.2
 0.2
8
5.7
37.6
74.5
IC
2.3
12
4.0
 16.1
 36.1
 9.14
 34.6
 30
 26.1
 42.4
68.9
 14.4
 13.9
 19.1
 13.8
 10.0
 22.1
 7.5
 5.4
 10.8
 11.4
 6.1
 8.4
 11.1
 10.5
 28.1
 10.0
 26.2
 17.5
 9.3
 2.9
 20.1
 8.7
 12.4
 4.2
 6.6
 10.9
 9.0
 11.2
3.0
8.7
35
qC1N
39
19.0
6.0
19
1
1
4
3
2
5
3
0.22
 19.8
0.2  0.04
 2.8
0.19  0.01
2.3  0.6
0.19  0
12.0  7.1
0.19  0.01
1.0
(N1)60,CS
0
0
0
0
0
0
 0.01
 0.01
0
0
0
0
 0.03
0
0
0
0
0
 0.01
0
 0.01
 0.01
 0.01
0
0
0
 0.01
0
Seed et al. (2001) and Moss and Seed (2002)
0.17
0.14
0.22
0.17
0.17
0.18
0.14
 0.06
 0.08
 0.05
 0.04
 0.04
 0.04
 0.05
 0.05
 0.06
 0.05
 0.06
 0.05
 0.07
 0.04
 0.06
 0.07
 0.04
 0.05
 0.05
 0.06
 0.06
 0.06
 0.11
 0.07
 0.06
 0.07
 0.04
9
 1.4
 1.3
 1.9
 1.3
 1.0
 2.0
 0.8
 0.5
 1.1
 1.1
 0.6
 0.9
 1.1
 1.0
 2.6
 1.0
 2.4
 2.5
 0.9
 0.3
 2.1
 0.9
 1.3
 0.5
 1.9
 1.1
 1.4
 1.1
DqC (Mpa)
0.3
0.1
0.1
0.0
0.4
0.0
0.2
0.4
0.5
0.5
0.3
0.1
0.0
0.4
0.0
0.0
0.2
0.2
0.6
0.4
0.4
0.4
0.3
0.2
0.0
0.2
0.1
0.1
 0.2
 0.3
 0.4
 0.1
 0.3
 0.1
 0.3
 0.2
 0.3
 0.3
 0.3
 0.1
 0.1
 0.3
 0.0
 0.0
 0.7
 0.4
 0.3
 0.1
 0.7
 0.2
 0.3
 0.2
 0.0
 0.2
 0.1
 0.1
qC1,mod (Mpa)
4.1
6.7
7.6
5.5
3.1
6.4
3.9
2.9
3.7
4.4
3.1
5.1
6.4
4.9
8.7
5.7
8.2
4.8
5.1
3.4
5.0
4.7
3.9
3.6
6.9
4.3
5.8
6.8
 1.4
 1.1
 1.6
 1.3
 0.8
 2.0
 0.7
 0.5
 1.0
 1.0
 0.5
 0.8
 1.1
 1.0
 2.6
 1.0
 2.0
 2.2
 0.7
 0.2
 1.7
 0.8
 1.1
 0.4
 1.9
 1.0
 1.3
 1.1
Ground
failure
No. CPT
No. Borings
Boring/CPT
No
Youd et al. (2001) procedures
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
1
0
11  2
10  7
10
Y
7
N
6
N
N
N
2
18  5
9
N
N
8
Y
3
Y
13
Y
7
9
8
10
12
6
21
13
Y
3
1
2
2
2
5
3
7
18
20
10
11
10
10
9
N
Y
Y
Y
N
N
Y
Y
6
1
3
3
2
2
2
Y
N
Y
Y
Y
Y
Y
Download