National Academies Press: OpenBook
« Previous: 2 A Primer on Earthquake-Induced Soil Liquefaction
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

3
Case Histories

Development of methods to assess liquefaction and its consequences relies on field case histories. Geotechnical engineers have compiled databases that contain hundreds of these case histories, the largest of which focus on whether liquefaction was triggered and the associated site conditions. Other databases contain field examples of the consequences of liquefaction, such as

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

lateral spreading and flow slides. Those databases are important in developing and validating methods for predicting liquefaction and its consequences (see Chapters 4-6).

A useful case history provides information about a soil’s in situ properties, its response to seismic shaking, and the nature of that shaking. Ideally, the information would include in situ properties prior to the earthquake and earthquake-induced ground motions measurements at the site. Most often, however, soil profile characteristics are determined after they may have been altered by the earthquake, and the seismic loading is estimated for the site rather than having been recorded directly.

Most observations in the databases are of surface manifestations, not of the soil response at depth (e.g., the extent of subsurface liquefaction and excess porewater pressure distribution). As a result, use of case history data for development, calibration, or validation of liquefaction analyses typically involves considerable judgment. Not surprisingly, varying interpretations of the same field observations are not uncommon.

Despite these limitations, there is recognition within the profession of the value of well-documented case histories. Post-earthquake surveys, including those under the auspices of the National Science Foundation–sponsored Geotechnical Extreme Events Reconnaissance (GEER) Association, provide important information before sites are altered by natural or anthropogenic activities.

This chapter considers opportunities to strengthen the geologic and geotechnical basis for describing and interpreting field case histories of liquefaction and related phenomena. The opportunities described include enhancing site characterization with new technologies and using new methods to improve case history protocols. The chapter focuses largely on case histories in which an earthquake triggers, or fails to trigger, liquefaction. But the chapter also considers case histories of two of the main consequences of liquefaction: lateral spreading, and the shear strength of a soil that undergoes large deformation subsequent to liquefaction triggering.

SITE CHARACTERIZATION FOR CASE HISTORY ASSESSMENTS

One of the most important steps in documenting a liquefaction case history is to document a site’s geologic and geotechnical conditions. In this site characterization, geologic descriptors include tectonic setting, stratigraphy, landforms and land use, groundwater conditions, geologic age, and prior earthquake history. Geotechnical observations include measurements related to dynamic response characteristics of the soil and the liquefaction resistance of various deposits. In liquefaction case histories, site characterization sets the stage for inferring whether and where liquefaction has occurred within a profile and how its consequences evolved.

Geologic Characterization

Seismic shaking at a site depends on the fault-rupture location, mechanism, and dimensions; on the rock structures through which the seismic waves pass; and on the geology directly beneath the site. Whether and how extensively a saturated material liquefies can be influenced by depositional environment, grain size distribution, lithology, weathering, age, and prior earthquake history. Local stratigraphy and landforms affect how liquefaction at depth is manifested at the ground surface.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

Local geologic controls on liquefaction were recognized at least a half century ago, when railroad damage from the March 27, 1964 Prince William Sound region Alaska earthquake (M 9.2) was related to “the underlying physiography and the materials of which the landforms are composed” (McCulloch and Bonilla, 1970, p. D95). Studies of site geology have been important in the interpretation of liquefaction case histories from several subsequent earthquakes in California, including the November 24, 1987 earthquakes (M 6.5 and M 6.7) on the Superstition Hills fault west of Westmorland (Holzer and Youd, 2007) and the October 17, 1989, Loma Prieta earthquake (M 6.9) (Bennett and Tinsley, 1995; Boulanger et al., 1998; Mejia, 1998), and 1994 Northridge (Holzer et al., 1999).

Lateral variability in stratigraphic units can complicate the geologic and geotechnical characterization of a case history site, and the lateral variability of an alluvial fan or meandering riverbed deposits may differ significantly from those in man-made hydraulic fills. For example, in one Loma Prieta earthquake liquefaction case history, the earthquake triggered liquefaction in a channel fill but not in the sand just outside the channel fill (Bennett and Tinsley, 1995). Similarly, the 2010 Darfield earthquake preferentially liquefied the fills of abandoned river channels in Christchurch, New Zealand (Wotherspoon et al., 2012).

Protocols for documenting geologic site conditions could include attention to what geologists call sedimentary architecture—the geometries of the various types of deposits laid down by water and wind. Geologic clues to these geometries can be combined with surface and subsurface investigations to aid estimating the extent of liquefiable layers through interpolating between and extrapolating beyond the places where geotechnical observations are available.

Geotechnical Characterization

Most liquefaction case histories include data from field tests for measuring soil properties. In situ testing avoids disturbance of the soil from sampling, particularly for nonplastic soils, and it obviates the need to reestablish the in situ state and fabric in the laboratory. Three in situ methods used most widely to evaluate soil properties for case histories are the standard penetration test (SPT; see Box 2.4), the cone penetration test (CPT; see Box 2.5), and shear wave velocity (Vs) testing (see Box 2.6). Table 3.1, updated from Youd and colleagues (2001), summarizes the advantages and disadvantages of these tests for assessing liquefaction triggering potential. The main advance in liquefaction potential assessment since Table 3.1 was originally published is the increased size of the Vs liquefaction triggering database. The liquefaction triggering database now contains more case histories with documentation of Vs than of either SPT or CPT resistance (Kayen et al., 2013). Nonetheless, the number of documented case histories using SPT and CPT has also increased since 2001, and more importantly, some of the databases include quality ratings and parameter uncertainties assigned to their case histories. The number of high-quality case histories continues to increase with recent additions from earthquakes offshore Bio-Bio, Chile (February 27, 2010; M 8.8); near Honshu, Japan (March 11, 2011; M 9.1); and the series of earthquakes in or near Christchurch, New Zealand, in 2010 and 2011.

Other in situ test methods used to characterize soil behavior and stratigraphy for liquefaction case histories include the piezocone penetration test (CPTu). The CPTu can be particularly useful for identifying dilatant zones, providing information regarding groundwater elevations, and identifying very thin silt and clay seams missed by other in situ test methods (Mayne, 2007). More esoteric techniques include an electrical probe to assess the anisotropy in soil fabric (Arulmoli et al., 1985; Arulmoli and Arulanandan, 1994; Arulanandan, 2008) and the vision cone penetrometer

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

test (VisCPT) to identify dilatant zones, to provide information regarding groundwater elevations, and to identify very thin silt and clay seams missed by other in situ test methods (Mayne, 2007). The VisCPT may also provide information on fines content (Raschke and Hryciw, 1997). These methods need to be more fully developed, however, and their efficacy demonstrated to add value for documenting and interpreting case histories, before they are embraced by the engineering community.

The presence of coarse gravel-sized and larger particles in soil requires use of methods other than the SPT and the CPT. Large particles physically interfere with the advancement of the SPT sampler and the CPT cone into the ground, resulting in measured resistance values that are too high. The main in situ methods used for gravelly soils are the Becker penetration test (BPT) (Harder and Seed, 1986), the Chinese dynamic penetration test (DPT) (Cao et al., 2013), and various large penetration tests (LPTs) (Daniel et al., 2003). BPT and LPT efforts have mainly focused on developing correlations relating their penetration resistance to that of the SPT. These equivalent SPT soil resistance correlations are then used to assess liquefaction triggering potential. Recent advances have been made in the development of improved SPT-BPT correlations by instrumenting the BPT to directly measure the energy transferred from the hammer to the driving shoe, avoiding the need to estimate the energy loss due to the frictional resistance between the drill string and the surrounding soil (Dejong et al., 2014; Ghafghazi et al., 2014). Efforts with the DPT have focused on direct correlation between DPT penetration resistance and liquefaction resistance. The DPT case history database is relatively small, though, and the quality of the cases has not been formally assessed.

Given that other methods are less mature, Vs testing may provide the most reliable data for evaluating liquefaction triggering resistance in coarse-grained soils not amenable to use of the SPT or the CPT. Development of the database of case histories of Vs measurements in coarse- grained soils in which liquefaction triggering is suspected should continue. In addition, however, the instrumented BPT and the DPT both show significant promise, and their further development would benefit the technical community.

There are far fewer case histories that document in situ soil properties at sites where lateral spreading and flow sliding have occurred than are found in liquefaction triggering databases. The SPT has been the primary method of characterizing the former phenomena, where these case studies have been used to develop correlations to estimate the post-triggering strength of the liquefied soil. Of the few case histories of lateral spreading or flow sliding characterized by the CPT, many of the data points are from older and less reliable CPT probes. In the residual shear strength of liquefied soil case history database compiled by Robertson (2010), only 6 of the 34 case history sites were characterized using the modern electronic CPT. Data collection for residual shear strength of liquefied soil databases need to include testing with modern CPT probes and other new and improved in situ test methods to enrich the available databases. Triggering may also be identified with additional investigation, including a search for dikes and sills underground, as illustrated in Box 3.1.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

TABLE 3.1 Advantages and Disadvantages of Various Field Tests for Assessment of Liquefaction Triggering

Attribute Test SPT CPT Vs
Cumulative use in liquefaction case histories Abundant Abundant Abundanta
Stress-strain conditions during test Partially drained, large strain Drained, large strain Small strain
Quality control and repeatability Poor to good Very good Good
Detection of vertical stratification of soil deposits Good for closely spaced tests Very good Depends on testing procedure, but generally fair to goodb
Soil types in which test is recommended (i.e., is expected to provide a reliable index of liquefaction resistance) All but gravelly soils All but gravelly soils All, with due consideration of particle size effectsc
Soil sample retrieved Yes No for standard configurationd No
Test measures index or engineering property Index Index Engineering property

a Kayen et al. (2013).

b Stokoe and Santamarina (2000).

c Committee comment.

d Robertson and Cabal (2014).

NOTE: Committee updates are in italics.

SOURCE: Modified from Youd, T.L., I.M. Idriss, R.D. Andrus, I. Arango, G. Castro, J.T. Christian, R. Dobry, W.D.L. Finn, L.F. Harder, M.E. Hynes, K. Ishihara, J.P. Koester, S.C.C. Liao, W.F. Marcuson, III, G.R. Martin, J.K. Mitchell, Y. Moriwaki, M.S. Power, P.K. Robertson, R.B. Seed, and K.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 127:817-833. With permission from ASCE.

LIQUEFACTION TRIGGERING CASE HISTORIES

The largest databases for liquefaction triggering are those assembled for use in developing the simplified, stress-based liquefaction evaluation procedures discussed in Chapter 4. Central to these procedures is use of the cyclic resistance ratio (CRR). In a simplified procedure, CRR is correlated to one of the in situ test parameters described above (e.g., SPT blow count, CPT tip resistance, or Vs). The evolution of liquefaction triggering case histories compiled to develop CRR curves is summarized below. The databases used to develop the most common triggering curves, as well as the challenges in interpreting and the opportunities to improve liquefaction triggering case histories, are also discussed.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

Overview of Triggering Databases

Published collections of triggering case histories date back to at least 1971. Whitman (1971) assembled 13 different cases from 8 earthquakes in Chile, Japan, Mexico, the Philippines, and the United States. Entries for each case included the depositional environment of the soil inferred to have liquefied; the estimated peak ground acceleration (PGA) at the ground surface; the depth of the groundwater table; the depth of the soil inferred to have liquefied (i.e., the “critical layer”); a representative SPT blow count (or N-value) for the critical layer; and the estimated duration of strong earthquake shaking. Seed and Idriss (1971) compiled 35 cases from 12 earthquakes in Chile, Japan, and the United States. Entries in that database included earthquake magnitude, distance from the earthquake source, and soil type in addition to water-table depth, critical-layer depth, representative SPT N-value for the critical layer, and estimated values of PGA and strong shaking duration. The Seed and Idriss (1971) database included 23 cases where liquefaction is known to have occurred and 12 cases in which no liquefaction is inferred.

Many other liquefaction triggering databases have been assembled since 1971, chiefly to support the development and updating of CRR curves. The databases vary in their levels of documentation. Examples listed chronologically and grouped by in situ test method include:

Other liquefaction triggering databases have been compiled in support of the development of CRR curves for other types of in situ tests, including the flat plate dilatometer test (Monaco et al., 2005) and the Chinese dynamic penetration test (Cao et al., 2013).

Table 3.2 summarizes the databases used to establish five liquefaction triggering curves used commonly in engineering practice, grouped by the in situ test method used to characterize soil resistance. The table reports the number of data points for cases in each database where (1) liquefaction is known to have occurred; (2) liquefaction is not believed to have occurred; and (3) borderline cases (“yes/no” cases). Other attributes in this table include the ranges of the depth to the center of the layer, with the lowest factor of safety against triggering (i.e., the critical layer) and the effective overburden pressure associated with the critical layer; the fines content (percent by weight passing the #200 sieve); and the normalized value of the penetration resistance or small strain Vs (N1,60cs, qc1Ncs, Vs1) for each database.

Figure 3.1 illustrates the range of the normalized value of the penetration resistances and small strain shear wave velocities for three of these databases (one for the SPT, one for the CPT, and one for Vs). Figure 3.1 indicates that the case history database is deficient in case histories for depths greater than 12-15 meters depending on the in situ test method; for normalized and standardized clean sand blow count, N1,60cs, in excess of 25; for normalized CPT tip resistance, qc1Ncs, greater than 125 atm (12.7 MPa); and for normalized Vs (Vs1) in excess of 225 m/s. Histograms of the ranges of the various parameters tabulated in Table 3.2 for the recently assembled SPT, CPT, and Vs databases are shown in Appendix C. Those histograms also suggest

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

that databases may be deficient with respect to magnitudes less than 5.9 and greater than 7.8-8.2, again depending on which in situ test is under consideration.

A significant issue when evaluating these databases is the criteria used to determine the data points and attributes summarized in Table 3.2. Subjective assessments of which data points to include for the “yes,” “no,” and “yes/no” cases and selection of their associated critical layers have been applied during database development. Issues associated with these assessments are described in the next section. Given the importance of data point and attribute selection in the development of CRR triggering curves, standardized protocols need to be applied in the development of future databases. There is some movement in this direction within the profession (e.g., Stewart et al., 2015), but no liquefaction case history database has been compiled to date using standardized procedures developed by consensus within the profession. Where possible, such procedures need to emphasize use of multiple in situ test techniques to characterize the soil layer expected to have liquefied, including primary wave (p-wave) velocity measurements to confirm saturation of the layer (e.g., for soil profiles where the groundwater table fluctuates regularly or, possibly, where decaying organics may release gases into the soil layer that result in unsaturated conditions). This will facilitate comparison and correlation among the different in situ test techniques.

TABLE 3.2 Summary of Recently Compiled Liquefaction Triggering Case History Databases for Level-Ground Conditions Showing Ranges in Values of the Parameters

Parameter SPT CPT Vs
Cea04a BI14b Mea06c BI14 Kea13d
“yes” cases 109 133 139 180 287
“no” cases 88 118 44 71 124
“yes/no” cases 3 3 0 2 4
Critical depth (m) 1.1-20.5 1.8-14.3 1.4-14.0 1.4-11.8 1.1-18.5
Effective overburden stress σ’vo (kPa) 8.1-198.7 20.3-170.9 14.1-145.0 19.0-147.0 11.0-176.1
Fines content (% by weight) 0-92 0-92 0-85
N1,60cs (blows/30 cm), qc1Ncs (atm), or Vs1 (m/s) 2.2-66.1e 4.6-63.7 11.2-252.0 16.1-311.9 81.7-362.9
Cyclic stress ratio CSRM7.5 0.05-0.66 0.04-0.69 0.08-0.55f 0.06-0.65 0.02-0.73
Earthquake magnitude M 5.9-8.0 5.9-8.3 5.9-8.0 5.9-9.0 5.9-9.0

a Cea04: Cetin et al. (2004).

b IB14: Boulanger and Idriss (2014).

c Mea06: Moss et al. (2006).

d Kea13: Kayen et al. (2013).

e N1,60 values listed for Cea04, as opposed to N1,60cs.

f CSR values listed for Mea06 and Kea13, as opposed to CSRM7.5.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Image
FIGURE 3.1 Plots of normalized in situ test values for case histories as a function of depth for the most recently compiled databases: (a) SPT (Boulanger and Idriss, 2014); (b) CPT (Boulanger and Idriss, 2014); and (c) Vs (Kayen et al., 2013). COURTESY: R. Green.
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

Liquefaction of Fine-Grained Soils

Soils with greater than 50% of their constituents passing the #200 sieve (i.e., fine-grained soils with grain sizes smaller than 0.075 millimeters) were once considered non-liquefiable, but several investigators have reported liquefaction of fine-grained soils in earthquakes. Wang (1979), as cited by Bray and colleagues (2004a,b) and Boulanger and Idriss (2004b), reported on liquefaction of “silty soils” in seven earthquakes in China. Based on data provided in Wang (1979), Seed and Idriss (1982) proposed criteria for assessing the susceptibility of fine-grained soils to liquefaction, commonly referred to as the “Chinese criteria.” These criteria are a function of percent clay (less than 15% by weight of particles < 0.005 mm), liquid limit (LL < 35%), and with an in situ water content greater than 0.9 multiplied by the LL. Subsequent studies, however, prompted by field observations of the cyclic response of fine-grained soils following more recent earthquakes, raised questions about the validity of the Chinese criteria (e.g., Seed et al., 2003; Bray et al., 2004a,b; Bray and Sancio, 2006). Those studies indicated that the percentage of clay was less important than was the plasticity index (PI) of the soil and that the ratio wc/LL was more important than was the LL alone in assessing the liquefaction susceptibility of fine-grained soils. Furthermore, the characteristics of the cyclic behavior of the soil and whether the soil’s strength and compressibility were more aligned with clay than sand were identified as being important in evaluating the potential for ground failure of fine-grained soil deposits during earthquakes (Boulanger and Idriss, 2006). In the absence of more detailed laboratory or in situ data, PI alone of the fine-grained soil was judged by Boulanger and Idriss (2006, 2008) to be a suitable proxy for determining whether the soil behaved as “clay-like” or “sand-like.”

Although the recommended PI threshold among the various studies (i.e., Boulanger and Idriss, 2004b, 2006, versus Bray et al., 2004a,b, and Bray and Sancio, 2006) at first appearance seem inconsistent, in reality the purpose of the recommendations made in these studies differs. Specifically, Bray and Sancio (2006) provide the results from a detailed laboratory study focused on determining the characteristics of fine-grained soils that are and are not susceptible to liquefaction. In contrast, Boulanger and Idriss (2008) focus on the more pragmatic issue of whether the simplified liquefaction evaluation procedure is suitable to evaluate the liquefaction potential of a fine-grained soil or whether laboratory testing needs to be performed. Seed and colleagues (2003) attempt to address both the fundamental issue of the susceptibility of fine-grained soils to liquefaction and the pragmatic issue of how to evaluate soil liquefaction potential.

All of the studies mentioned above focused on soils with relatively high fines contents (i.e., fines contents above that required to fill the voids of the coarse-grain fraction of the soil, resulting in the coarse grains “floating” in the fine-grain matrix). As a result, there is uncertainty about the applicability of the recommendations from these studies for assessing liquefaction susceptibility of soils having lower fines contents. Furthermore, interpretation of published field data is complicated by the general failure of the investigators to report Atterberg limits for fine-grained soils that were subject to strong shaking but did not liquefy. Additional field case history data on the performance of fine-grained soils subject to strong shaking, including Atterberg limits on soils that did not liquefy, supplemented by sampling and laboratory testing of the fine- grained soils reported to have liquefied and not liquefied, are needed to develop comprehensive criteria regarding the liquefaction susceptibility of fine-grained soils and the suitability of various procedures for evaluating liquefaction potential. These data needs include soils with liquid limits greater than 35% and PIs greater than 15%, which, to date, represent the limits of soils reported to have liquefied in the field.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

Interpretative Issues in Liquefaction Triggering Case Histories

To determine whether or not soils at a site have liquefied, and which deposits are likely to have liquefied if liquefaction is inferred to have occurred, representative geotechnical properties must be established for the soils. More often than not, these properties are determined using data collected after an earthquake, despite the fact that properties may have changed as a result of the earthquake. Earthquake ground motions at case history sites are usually estimated based on nearby recordings, event-specific contours of strong ground motions, or generalized ground motion prediction equations, as cases are rare where motions are recorded at a site known (or suspected) to have liquefied.

Determining Whether Liquefaction Was Triggered

With a few exceptions, whether a case history site is classified as having liquefied or not is based on the presence or absence of post-earthquake surface manifestations of liquefaction (e.g., sand boils and ejecta, cracking, or settlement). When surface manifestations of liquefaction are observed, it is assumed that at least one deposit in the soil profile has liquefied. Changes to the surface (or lack thereof) are not necessarily accurate indicators of liquefaction having occurred (or not), however. For example, there may be settlements due to dissipation of excess porewater pressures in sandy soils that occur when the factor of safety against liquefaction is upwards of two, particularly in looser deposits (Ishihara and Yoshimine, 1992). In such a circumstance, it is possible that case histories could be classified erroneously as having liquefied. Similarly, liquefaction may not manifest itself at the ground surface due to the depth or density of the liquefied layer or the integrity and thickness of the overlying non-liquefied crust. In most databases, the “yes/no” cases are those with marginal or minor evidence of liquefaction. Discrepancies in interpretation do occur even in identifying the occurrence of liquefaction: one such example is Smyrou and colleagues (2011) versus Wood and colleagues (2011). And, there is little to no consistency among databases in how to define those cases in which evidence of liquefaction is ambiguous. Protocols for documenting and categorizing triggering case histories could result in more consistency among databases.

Identifying Which Soils Liquefied

The soil layer inferred to have liquefied in a triggering case history is commonly called the “critical layer,” and the depth to this soil is called the “critical depth” (zcrit). The critical depth for “no” cases is taken generally as that at the center of the soil layer judged to have the highest liquefaction potential. The criteria used to select the critical layer in case studies, however, is often not provided. This could result in inconsistencies among studies and uncertainty in establishing a triggering relationship from the data. Early case history databases sometimes contained multiple critical layers extracted from a single boring or sounding (e.g., Tokimatsu and Yoshimi, 1983; Yegian and Vitelli, 1983; Seed et al., 1984; Stark and Olson, 1995). This practice has been avoided in the development of more recent databases, as confirmed by study of the databases by members of this study committee, in which the “weakest-link-in-the-chain” concept is used to select a single critical layer in a soil profile (e.g., Cetin et al., 2000; Moss, 2003; Idriss and Boulanger, 2008; Kayen et al., 2013; and Boulanger and Idriss, 2014).

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

In “yes” cases, the critical layer has sometimes been determined by comparing soil ejecta from surficial liquefaction manifestations to samples obtained from borings (e.g., Liao and Whitman, 1986; Green et al., 2011). In such cases, there is a possibility that the soil ejecta include soil from the overlying strata. Water expelled from a liquefied soil, with or without material from that soil itself, can entrain material from overlying strata as it flows from the critical layer to the ground surface. For example, the 1989 Loma Prieta earthquake produced boils of mud that erupted in an area where liquefiable sand underlies estuarine mud. Estuarine mud may have liquefied, as suggested by Mejia (1998), or the mud may merely have been fluidized by waters derived from the underlying sand, as suggested by Boulanger and colleagues (1998). It is rarely simple to use trenches and borings to trace subsurface intrusions to the source of the expelled water. Along the Columbia River, for instance, sediment peels from large rectangular cores were used in an attempt to trace surficial dikes formed during the 1700 Cascadia earthquake downward to their roots (Takada and Atwater, 2004). The attempt was confounded both by the apparent dissipation of subsurface dikes in cross-bedded sand and by the likely presence of dikes and sills from earthquakes prior to the one responsible for the surficial diking (see Figure 2 in Box 3.1).

General criteria for selecting the critical layer for paleoliquefaction as well as modern case histories have been proposed in a few studies wherein consistency between the observed surface manifestations and the depth, density, and thickness of the critical layer is emphasized (Green et al., 2005, 2014; Olson et al., 2005). Considerable judgment is still required to implement these criteria, however, suggesting that differing interpretations of the same case histories is likely (e.g., Boulanger and Idriss, 2014; Green et al., 2014). Use of multiple lines of evidence to identify the critical layer and protocols for documenting and interpreting the critical layer and its representative properties could reduce subjective interpretation and its associated uncertainties.

Representative Properties of Soils That Liquefied

The soil in a triggering case history database is generally characterized by a single normalized in situ test parameter (e.g., corrected SPT N-value). There is no standard protocol for selecting this value where multiple in situ test parameter values are measured in the critical layer. Some investigators have taken the smallest measured test parameter value as the representative value of the critical layer (Liao and Whitman, 1986). Other investigators have used the average value of the in situ test parameter within the critical layer (e.g., Cetin et al., 2000; Idriss and Boulanger, 2008), in which case the selected thickness of the critical layer becomes important because the measured test parameter values often vary with depth within a given stratum (e.g., Robertson, 2009; Boulanger and Idriss, 2014; Green et al., 2014).

Where averaging is used, the resulting representative test parameter can differ depending on whether the measured test indices are averaged first and then corrections are applied, or whether the corrections are applied to the measured values and the corrected values are then averaged (Green et al., 2014). The differences between the two averaging approaches can be large for soils with high fines contents. The committee is not aware of any justification for averaging before an in situ resistance parameter is applied.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

Pre- Versus Post-Earthquake Soil Properties

It has already been stated that most case histories in liquefaction triggering databases rely on in situ test results obtained after liquefaction occurred. The inherent assumption in this approach—namely, that characteristics of the critical layer do not change significantly as a result of the earthquake—is supported to a limited extent by data from sites in California and New Zealand, where both pre- and post-earthquake data are available (Holzer and Youd, 2007; Orense et al., 2011). Some investigators, however, have reported large differences between pre- and post-earthquake SPT N-values (Ohsaki, 1966). In those cases, soil layers with low pre-earthquake N-values tended to densify (contract) while soils with high pre-earthquake N-values tended to loosen (dilate).

It is difficult to reconcile the effects of an earthquake on the critical layer penetration resistance. Therefore, liquefaction databases need to report the measured values of soil resistance parameters and whether those measurements were made pre- or post-earthquake. Individual investigators may then decide whether or not to make a correction to post-earthquake values (and document any such correction) when interpreting a case history.

Identifying the Seismic Demand Imposed on the Critical Layer

In the absence of a downhole accelerometer array at the site, a detailed site-specific site response analysis generally is the most accurate means to estimate the seismic demand imposed on the critical layer (e.g., the shear stresses induced in the critical layer). Such analysis, however, is possible only if the soil profile is well characterized down to bedrock, the Vs of the bedrock is known, and representative bedrock input motions are available.

In lieu of detailed site response analyses, simplified procedures often are used to estimate the seismic demand imposed on the critical layer. Several potential sources of ground motion intensity data of varying degrees of uncertainty are used to estimate the intensity data for the case histories. These include on-site ground surface motion recordings; near-site ground surface motion recordings (preferably at sites with known Vs profiles) with intra-event residuals for conditional PGA (Bradley, 2013) estimation; ShakeMap intensity measure values;1 event-specific ground motion predictive equation (GMPE) values; values back-estimated from observed structural performance (e.g., overturning of headstones); and region-specific as well as general GMPE values.

Case history databases need to include enough information, such as the magnitude, source mechanism, and site-to-source distance of the earthquake, to permit independent assessment of demand value by the user. When the database includes a demand parameter (e.g., the PGA at the ground surface or the CSR at the critical depth), the method used to establish this parameter should be documented.

Limitations of Case History Database Coverage

While there are hundreds of liquefaction triggering case histories (see Table 3.2), only a few dozen plot near the curves that separate zones of liquefaction and no liquefaction: that is,

___________________

1 For example, http://earthquake.usgs.gov/research/shakemap.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

deterministic plots of CRR versus soil liquefaction resistance parameter (N1)60cs, Qc1Ncs, or Vs1 (see Chapter 4). This limited number of influential case histories increases the uncertainty in the location of the CRR curve. In addition to the limited case histories near the deterministic CRR boundary line between “yes” and “no” cases, there is also a paucity of data for depths of less than 2 meters and greater than 15 meters, for soils with nonplastic fines content greater than 35% by weight, for gravelly soils, and at sloping ground sites.

A maximum depth at which liquefaction can be triggered is an important, yet controversial, factor when evaluating the risk to projects from liquefaction. As may be observed from Table 3.2, the plots shown in Figure 3.1, and in Appendix C, the maximum depth to the center of the critical layer listed in recently compiled case history databases is 20.5 meters, but some researchers place the maximum value for that same case history at considerably shallower depth.2 The next deepest cited cases for liquefaction are approximately 18 meters for two sites (Kayen et al., 2013), but it is uncertain whether these two sites are included in the other databases. The maximum depth to the center of the critical layer is approximately 15 meters in most databases. The absence of “yes” case histories with a critical layer depth of greater than 15 to 20 meters does not mean liquefaction cannot be triggered at such depths. It may just indicate that surface manifestations of liquefaction are unlikely for level ground conditions (i.e., where the initial static shear stress has little to no influence on the triggering or manifestation of liquefaction) if liquefaction occurs at those or greater depths. For example, it is believed that soils liquefied at depths up to 25 meters or greater in the Lower San Fernando Dam during the 1971 San Fernando earthquake (Castro et al., 1992), resulting in the upstream slide of the dam. It is unknown whether surficial manifestations of liquefaction would have been evident in the absence of the upstream slide, or whether the soil at this depth would have liquefied in the absence of the high initial static shear stress conditions. Data from instrumented sites where strong ground shaking is expected and liquefiable layers at depths of greater than 15 meters are present could indicate that liquefaction can be triggered at depth. Additionally, centrifuge tests may provide insights into the maximum depth at which liquefaction can occur.

Limitations of the case history databases are expected to decrease given the growing size and quality of field case history databases, especially when data from the 2010-2011 earthquakes in Chile, Japan, and New Zealand and other recent earthquakes3 are included. Emphasis is needed on documenting quality case histories with parameter ranges beyond those of case histories currently in the database. Case histories that approach and extend the bounds of the data with respect to grain size (i.e., low plasticity silts and gravelly soils), depth (i.e., less than about 2 meters and greater than about 15 meters), earthquake magnitude (i.e., greater than magnitude 7.8), and from sloping ground sites are of particular interest. Continued collection of case history data within the ranges covered by existing databases is also warranted to reduce associated uncertainties.

___________________

2Cetin and colleagues (2004) interpreted the critical depth to liquefaction at the Kushiro Port Seismograph Station site during the 1993 Kushiro-Oki, Japan, earthquake to be 20.5 m, but Kayen and colleagues (2013) and Boulanger and Idriss (2014) reinterpreted the critical depth for this case history as 4.5 m and 3.8 m, respectively.

3 Including the 1999 events in Turkey and Taiwan and the 2004 and 2007 events in western Japan.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

Balance Between Liquefied and Non-Liquefied Sites

Imbalance in the number of “yes” and “no” cases can be a significant source of bias when establishing CRR curves. While recently compiled SPT databases do balance the numbers of “yes” and “no” cases fairly well, recently compiled CPT and Vs databases contain significantly more “yes” than “no” cases (see Table 3.2). Information from profiles where liquefaction is known to have occurred—for example, where surface manifestations of liquefaction are observed—often attracts more attention than that from nearby areas where no surficial evidence of liquefaction is observed. As a result, data are not collected for those nearby profiles that may have liquefied without surface manifestations; that may be susceptible to liquefaction but did not liquefy under the specific earthquake loads; or that are not liquefiable. To minimize bias, the numbers of “yes” and “no” cases need to be balanced when developing liquefaction resistance correlations (e.g., CRR curves).

Minimum Earthquake Magnitude

Although M 5.9 is the smallest earthquake magnitude represented in Table 3.2 and Appendix C, sand boils in the dry lake bed of Soda Lake, a quarry-tailing deposit, were observed following an M 4.6 aftershock of the 1989 Loma Prieta earthquake (Sims and Garvin, 1995)—the minimum magnitude earthquake reported to have triggered liquefaction. In this case, ejecta were vented through existing dikes formed during the main shock and an earlier aftershock. It is possible that no surface manifestations would have occurred if ejecta pathways had not already existed. The lowest magnitude main shock event documented to have triggered liquefaction is an M of about 5 (Ambraseys, 1988).

The absence of case histories for earthquakes with an M less than 5.9 in the databases represents a significant shortcoming because the potential for surface manifestations and damage to fragile structures due to liquefaction in events having magnitudes less than about 5.9 cannot be completely discounted. Developers of liquefaction triggering analyses typically provide magnitude scaling factors for earthquakes as small as M 5.5. The variability among the M 5.5 magnitude scaling factors proposed by various investigators (discussed in Chapter 4) is greater than the variability at larger magnitudes and is a major source of uncertainty in liquefaction analysis. Because magnitude scaling factors often are extrapolated in practice to magnitudes even smaller than 5.5, the uncertainty is exacerbated. Therefore, case histories of the performance of liquefiable soil deposits subject to earthquakes of magnitudes 5.5 or smaller, whether or not surface manifestations of liquefaction or liquefaction-induced damage are observed, are valuable, if not essential, for reducing the uncertainty in predicted liquefaction hazard due to small magnitude events.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

Predominance of Plate-Boundary Settings

Shallow ruptures on faults along or near tectonic plate boundaries, as may occur in California, and shallow intra-slab events associated with subduction zones, as may occur in Japan, are the source mechanisms represented most often in the liquefaction case history databases. Comparatively few case histories are available for great earthquakes—that is, of magnitude 8 to 9—on the plate-boundary thrust zones. But as they become available, case histories from the 2010 Maule, Chile, and 2011 Tohoku, Japan, earthquakes will provide more examples from subduction zones.

Also in short supply are well-documented case histories of liquefaction from earthquakes in stable continental regions (e.g., the New Madrid and Charleston seismic zones in the United States). Because the bedrock is denser, colder, and less fractured in these regions than in near plate boundaries, bedrock earthquake motions are expected to be higher in amplitude, shorter in duration, and richer in higher frequencies than in shallow crustal or subduction zone earthquakes. It has been suggested that these differences in ground motions may lead to difference in liquefaction triggering behavior (Semple, 2013). These interior earthquakes happen so rarely, however, that few cases histories from this tectonic setting are likely to emerge in the coming decades. This predicament adds to the importance of laboratory and analytical studies of the impact on liquefaction behavior of the nature of ground motions in these regions.

Influence of Soil Age on Liquefaction Potential

The vast majority of case histories in the liquefaction triggering databases represent either man-made fills or Holocene alluvial and fluvial sediments (Robertson and Wride, 1998; Youd et al., 2001). Based on geologic evidence, Youd and Hoose (1977) and Youd and Perkins (1978) postulated a generalized ranking of the liquefaction susceptibility of sedimentary deposits by age and sedimentary environment. In this ranking, sedimentary deposits become less susceptible to liquefaction as they age. Deposits less than 500 years old are generally ranked most susceptible, and deposits more than about 2 million years old are ranked least susceptible. Revisiting case studies containing Holocene soils and fills may provide beneficial information on the rate of strength gain with age. Most deposits of Pleistocene age (between about 10,000 and 2 million years) are ranked low or very low in liquefaction susceptibility, but Pleistocene deposits are known to have liquefied in historic earthquakes. Alluvial fan deposits estimated to be 10,000-15,000 years old (Andrus, 1986) liquefied during the October 28, 1983 Borah Peak, Idaho, earthquake (M 6.9; Andrus and Youd, 1987). Far older Pleistocene deposits liquefied during the September 1, 1886 Charleston, South Carolina, earthquake (M 7.3; Martin and Clough, 1994).

The Charleston area contains geologic evidence for repeated liquefaction at intervals averaging 500-600 years over the past 6,000 years (Talwani and Schaeffer, 2001). As a result, even though these deposits are relatively old from a depositional point of view, from a geotechnical perspective (i.e., time since last significant disturbance), they are much younger (see, e.g., Andrus et al., 2009). The Charleston data suggest that liquefaction will “reset the clock” with respect to the age of a soil deposit for a liquefaction assessment. Nevertheless, it is not clear if “near liquefaction” (i.e., significant pore-pressure generation without liquefaction) may also reset the clock, and studies are therefore warranted on this point.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

Case History Quality Rating

Cetin and colleagues (2000) made a major advance in case history compilation in their SPT database, wherein procedures for quantifying uncertainties and assigning an overall quality rating for case histories were proposed. They define four case history quality classes, and the criteria for each relates to the quality and thoroughness of subsurface characterization and the quantification of the seismic demand—information needed to interpret a case history properly. The quality ratings allow the model developers to include, exclude, or weight cases based on the certainty in their interpretation. A quality rating system of the type proposed by Cetin and colleagues (2000) would enhance the value of liquefaction case history databases.

SITE-SPECIFIC GROUND MOTION RECORDINGS

Installation of strong motion instruments has increased in the last two decades and has resulted in ground motion recordings at sites underlain by liquefied soil. Recorded motions at those sites reflect the occurrence of liquefaction and can be used to better understand liquefaction triggering and the response of liquefied profiles. Most of the sites were equipped with only ground surface instruments, but a few of those locations also include downhole instruments in vertical arrays.

Vertical Arrays

Vertical ground motion instrument arrays, in which strong motion instruments are placed at the ground surface and at one or more depths in nearby boreholes, can provide excellent opportunities to understand the seismic response of soil profiles and of individual soil layers between pairs of instruments at adjacent depths.4 They also provide data on the ground shaking intensity at the time liquefaction is triggered. Vertical arrays (see Box 3.2) installed specifically to better understand liquefaction behavior are still relatively rare, but such arrays may include pore-pressure transducers and slope inclinometers to provide information on liquefaction behavior, reducing uncertainties on the location of the critical layer and the intensity of ground motions necessary to induce liquefaction. Shape-acceleration arrays (SAAs) (Danisch et al., 2004; Zeghal et al., 2007) offer a flexible and economical means of measuring soil response in vertical arrays. Installation of additional vertical arrays at sites with a high potential for liquefaction would provide valuable case history data and detailed ground motion data for calibration of numerical models when one or more of these vertical array sites actually liquefied.

LATERAL SPREADING CASE HISTORIES

Compared to case history databases for liquefaction triggering, few lateral spreading case history databases have been assembled (see, e.g., Hamada et al., 1986; Bartlett and Youd, 1995; Rauch and Martin, 2000; and Youd et al., 2002). Cases that are available have been characterized almost exclusively using SPT. The database of 484 case histories compiled by Youd and colleagues (2002), a revision of the Bartlett and Youd (1995) database, is among the most recently

___________________

4 The records obtained from such arrays provide data for development, calibration, and validation of empirical, semiempirical, and mechanistic models of liquefaction triggering.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

published. The majority of the cases in this database are from the 1964 Niigata, Japan, and the 1983 Nihonkai-Chubu, Japan, earthquakes. As is true for empirical relationships to predict liquefaction triggering, the validity of empirical relationships for predicting the intensity of lateral spread displacements is limited by the parameter ranges of the case histories from which the relationships are derived. The geometric parameters used to characterize lateral spreads by Youd and colleagues (2002) are defined in Figure 3.2, and the ranges of all of the parameters used in the Youd and colleagues (2002) case histories are listed in Table 3.3.

Sources of bias in lateral spread case history databases are similar to those in liquefaction triggering databases (e.g., tectonic setting, geologic age and setting, pre- versus post-earthquake characterization of soil profiles, interdependence of case histories), but they are more significant in some regards. For example, multiple displacement vectors at a lateral spread site are treated as separate case histories by Bartlett and Youd (1995) and Youd and colleagues (2002). Rauch and Martin (2000) modified the Bartlett and Youd (1995) database by grouping vectors from the same spreading area, oriented in the same general direction, into a single case history. Using this approach, the 467 cases listed in the Bartlett and Youd (1995) database were reduced to

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

approximately 70 case histories. Some case histories in the Rauch and Martin (2000) database, however, were from adjacent sites and therefore may not be completely independent. Consequently, further efforts to account for case history interdependence are needed when developing lateral spread predictive relationships, such as using mixed effects regression techniques (Pinheiro and Bates, 2000).

TABLE 3.3 Parameter Ranges for Case History Database

Parameter Range
Depth of range (M; m) 6.4-9.2
Site-to-Source Distance (R; km) 0.2-100
Peak ground accelerationa (g) 0.19-0.55
Cumulative thickness (m) of liquefiable soils in upper 20 m of soil (N1,60 < 15 blows/30 cm; T15) 0.01-19.7
Average fines content of soil comprising T15 (D50; mm) 0-70
Median particle size of soil comprising T15 (D50; mm) 0.036-12
Slope angle from horizontal (S; degrees) 0.05-11
Ratio of height of free face to distance from free face (W; %: see Figure 3.2) 1-56.8
Permanent horizontal deformation observed at the point of interest (DH; m) 0.01-10.16

a PGAs from Bardet et al. (2002).

NOTE: Abbreviations used in Figure 3.2 provided.

SOURCE: Youd et al. (2002).

Lateral spread case history databases have not been subject to the same level of scrutiny as have liquefaction triggering databases. Additionally, the lateral spreading phenomenon is complex, and the resulting spatial spreading deformations and patterns are often characterized simply (e.g., as a single lateral displacement). Often only a few in situ field tests are used to characterize the soils across a lateral spread. This combination of factors results in fewer high-quality, well-documented case histories. Protocols are needed for documenting and interpreting lateral spreading case histories, particularly to address the issue of case history interdependence. Formal quality rating schemes for lateral spread case histories need to be developed and clearly stated by model developers that use the case histories. These ratings need to be taken into consideration when developing, calibrating, or validating lateral spreading models.

Remote sensing technologies such as terrestrial and airborne Light Detection and Ranging (LiDAR) and satellite imagery are being employed more commonly in post-earthquake reconnaissance (e.g., Bray and Frost, 2010). Such technologies allow for more comprehensive documentation of the displacements associated with lateral spreading. For example, optical image correlation (Leprince et al., 2007) can be applied to pre- and post-earthquake satellite imagery to develop lateral displacement measurements at high spatial resolution (e.g., Martin and Rathje, 2014) as shown in Figure 3.3. Pre- and post-earthquake LiDAR-derived digital elevation models

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

can be used to compute settlement and lateral displacements associated with liquefaction and lateral spreading (e.g., van Ballegooy et al., 2014b). Pre-earthquake data are required, however, to use these approaches, and investing in baseline data collection in areas of interest (e.g., Holocene deposits in seismically active areas) in anticipation of earthquakes would make use of such methods possible. Additionally, such new approaches and technologies as digital elevation models derived from digital photogrammetry and real-time processing of LiDAR data collected from unmanned aerial vehicles are being explored for documenting and interpreting case histories (Rathje and Franke, 2015).

POST-LIQUEFACTION SHEAR STRENGTH CASE HISTORIES

Compilations of field case histories for post-liquefaction residual shear strength (i.e., large deformation) of granular soils are significantly smaller and contain larger uncertainties than are those for liquefaction triggering and lateral spreading. One of the first, if not the first, field case history databases for residual shear strength consisted of just 12 case histories and included earthquake-induced failures of several dams, dikes, embankments, natural slopes and lateral spreads, an earthquake-induced bearing capacity failure of a four-story apartment building, and the failure of two hydraulic-fill dams under construction (Seed, 1987). Normalized and fines-corrected SPT N-values were provided for each case, as were estimates of the residual strength from back-analyses based on initial and final geometries of the sliding mass system.

Subsequent researchers have added, reinterpreted, and filtered case histories based on deformation mode, and they have applied quality control ratings to residual shear strength case histories (e.g., Seed and Harder, 1990; Stark and Mesri, 1992; Olson and Stark, 2002; Idriss and Boulanger, 2007; Robertson, 2010; and Kramer and Wang, 2015). Table 3.4 summarizes the ranges of the variables in the case history databases for the most commonly used correlations. As may be seen from this table, there are no case histories that have a corrected SPT N-value greater than 15 blows/30 cm or a corrected CPT tip resistance greater than 85 atm. The validity of empirical relationships for predicting residual strengths is limited by the parameter ranges of the case histories from which the relationships are derived.

Image
FIGURE 3.2 Parameters used to characterize lateral spread case histories. NOTE: DH = horizontal ground displacement; H = height of the free-face; L = horizontal distance from the freeface; S = ground slope; and freeface ratio W = H/L. SOURCES: Bardet, J.P., T. Tobita, N. Mace, and J. Hu. 2002. Regional modeling of liquefaction-induced ground deformation. Earthquake Spectra 18(1):19-46; Bartlett and Youd, 1992. With permission from the Earthquake Engineering Research Institute.
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

TABLE 3.4 Variable Ranges in Post-Liquefaction Residual Strength Case History Databases

Seed (1987) Seed and Harder (1990) Stark and Mesri (1992) Olson and Stark (2002)a Idriss and Boulanger (2007) Robertson (2010) Kramer and Wang (2015)a
Number of cases 12 17 20 33 18 34 31
N1,60-cs (blows/30c m) 3-15 3-15 3-15 0-11.5 3-15 - 1.1-12.6
qc1-cs (atm) - - - 0-60 25-85 34-77 -
Fines content (%) 3-80 3-80 0-80 0-100 0-85 0-100 0-100

a The listed penetration values for Seed (1987), Seed and Harder (1990), Stark and Mesri (1992), Idriss and Boulanger (2007), and Robertson (2010) include fines content corrections (i.e., N1,60-cs or qc1-cs). The listed penetration values for Olson and Stark (2002) and Kramer and Wang (2015) do not include fines content corrections (i.e., N1,60 or qc1).

Image
FIGURE 3.3 Amplitudes of horizontal displacement from optical image correlation displayed along with observed crack locations from field reconnaissance of the 2011 Christchurch, New Zealand, earthquake. This is an example of what can be derived from satellite imagery in situations where both pre- and post-earthquake imagery is available at high resolution. SOURCE: Martin, J.G., and E.M. Rathje. 2014. Lateral spread deformations from the 2010-2011 New Zealand earthquakes measured from satellite images and optical image correlation. In Proceedings of the 10th U.S. National Conference in Earthquake Engineering, 21-25 July, Anchorage, Alaska. Oakland, CA: Earthquake Engineering Research Institute. With permission from the Earthquake Engineering Research Institute.
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

Evolution of the databases has not been linear: some databases exclude case studies that are included in subsequent databases. For example, Idriss and Boulanger (2007) explicitly removed lateral spread case histories from their residual shear strength database. Olson and Johnson (2008), on the other hand, stated that their analyses showed that residual shear strengths back-calculated from lateral spread cases were in good agreement with those from the Olson and Stark (2002) residual shear strength case history database. Additionally, Stark and Mesri (1992), Olson and Stark (2002), and Kramer and Wang (2015) express the residual shear strength as the strength ratio (i.e., ratio of residual shear strength to the initial vertical effective stress), while the other databases do not. The residual shear strength ratio of Kramer and Wang (2015) is a function of the magnitude of the vertical effective stress. Also, Olson and Stark (2002) and Kramer and Wang (2015) state that the strength ratios correlate just as well or better to penetration resistances that are not corrected for the influence of fines content, while other databases use corrected penetration resistances. Seed and Harder (1990), Olson and Stark (2002), Idriss and Boulanger (2007), and Kramer and Wang (2015) considered dynamic (or kinematic) effects in estimating the residual shear strength. Variability among databases complicates data collection and case history interpretation. Accordingly, protocols need to be established for documenting and interpreting post-liquefaction residual shear strength case histories.

The quality and reliability of case histories can range, and quality rating of case histories has varied in databases in which they have been used. For example, Idriss and Boulanger (2007) included only those cases with enough information to allow dynamic (or kinematic) effects to be incorporated into the calculation of the residual shear strength. Robertson (2010) rated the case histories from Class A to E based on details of the in situ characterization of soil properties. Class A cases were characterized by the modern electronic CPT prior to failure, and subsequent classes have less reliable characterizations (e.g., estimated soil properties based on judgment, the SPT, mechanical CPT). Kramer and Wang (2015) applied a weighting factor of 1.0 to thoroughly investigated and well-documented case histories and lower weighting factors (between 0.8 and 0.2) to other case histories. Applying a rating system to case histories in a database to be used for establishing a post-liquefaction shear strength correlation is a welcome improvement. Because all quality rating systems are inherently subjective, the rationale for the rating system and the weighting factors used within it need to be documented in such a way that allows potential users of the database and resulting regressions to be fully informed about the system employed.

NONTRADITIONAL SOURCES OF DATA

Recent developments in instrumentation and sensing, imaging technologies, and data acquisition, processing, and display may provide new and unique opportunities to collect and record data during and after strong earthquake shaking. For example, mobile phones with video and Global Positioning System capabilities allow liquefaction features to be captured on video and accurately located. While engineers often think of “response” as detailed measurements of displacements, velocities, accelerations, porewater pressures, and other physical reactions, videos of earthquake response by laypersons are more commonly available, and crowdsourcing of earthquake response data is becoming possible. Coupled with detailed post-event deformation measurements, crowdsourced data such as videos can aid the understanding of a case history.

While most efforts to date have focused on emergency response or humanitarian actions (Starbird and Palen, 2013), opportunities to extract technical information such as locations of liquefaction features and lateral spreading deformations also exist. Nontraditional types and

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

sources of data for documenting and interpreting liquefaction case histories may be a means to expand existing databases.

ENHANCING DATABASE DEVELOPMENT

This chapter describes existing liquefaction case history databases and their limitations. The impact of these limitations could be reduced if future case histories meet heightened and consistent standards for documentation. Strong, community-based collaboration will be necessary to establish standards to allow a more holistic understanding of liquefaction and its consequences. Table 3.5 is an example draft checklist adapted from the Next-Generation Liquefaction (NGL) project (see Box 3.3) for documenting liquefaction and related phenomena case histories. This table represents a good starting point for a standard case history documentation protocol, but it lacks certain important information such as the availability of remote sensing data and whether data quality was rated.

TABLE 3.5 Example Checklist for Documenting Liquefaction and Related Phenomena Case Histories

Data Type Desired Documentation Comments
Site overview Longitude and Latitude of Site
Topography, Ground Slope Proximity to topographic irregularities
Presence/Proximity of Structures Building structures, buried infrastructure, embankments, etc.
Geologic setting and site geology Natural or Man-Made Fill Construction records for fills (dates, methods of placement)
Depositional Environments
Age of Deposits
Spatial Variability Both lateral and vertical variability, potential for layering (and void redistribution)
Evidence of Liquefaction Description of Sand Boils Severity of liquefaction/dimensions of boil/amount of ejecta, color/classification of ejecta, etc.
Ground Movement Ground crack widths, lengths, orientations; LiDAR images, remote sensing measurements
Ground Settlement Amount and method of measurement; reference point for measurements
Ground Motion Characteristics E.g., dilational spikes in acceleration time history
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Data Type Desired Documentation Comments
Damage to Structures and Other Infrastructure Post-triggering consolidation settlement, bearing capacity failure, broken buried pipelines, etc.
Visual Accounts Photographs of before and after, videos from observers or security cameras, interviews with eye witnesses, etc. For lateral spreading and residual shear strength case histories, LiDAR scans and high-resolution aerial photographs for both pre- and post-event or topographic surveys.
Ground Motions Evidence of rapid changes in frequency content, dilation-induced acceleration spikes
Site Characterization Borings Number, locations relative to liquefaction features, drilling data (i.e., rotary wash, hollow stem, casing)
SPT System Hammer type, rod type and length, drill bit, lifting system, sampler diameter (inside and outside), energy measurements
SPT Data 6/6/6 blow counts, soil description, sample recovery
CPT Soundings Number, locations relative to liquefaction feature, availability of CPT samples
Vs Data Type of measurement (downhole, seismic CPT, spectral analysis of surface waves, multichannel analysis of surface waves, etc.); locations
Vp Data For determination of depth to saturation
Groundwater Data Depth at time of field testing and at time of earthquake, daily and seasonal fluctuations
Laboratory Data Grain Size Distribution Fines content
Atterberg Limits PI, LL
Soil Mineralogy Silicate, feldspar, calcareous, etc.
Angularity of Particles Water Content, Unit Weight Tests of Mechanical Behavior Consolidation, undrained strength, cyclic tests, etc.
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Data Type Desired Documentation Comments
Ground Motions Recorded Motions PGA, acceleration time history; location of nearby strong motion stations; type of instrument/housing; links to recorded motions
Source/Path/Site Parameters Moment magnitude, source-to-site distance (Rjb, Rrup, etc.), style of faulting, Vs30, presence of basin, basin depth, hanging wall, etc.; event type (shallow crustal in active region, subduction, stable continental region)
ShakeMaps Event-specific
Ground Shaking Intensity Structural effects or damage caused by ground vibratory motion (e.g., overturning and sliding of rigid objects or yielding of moment frames)
Ground Motion Prediction Equations Regionally applicable where available; event terms, mapped within-event residuals

SOURCE: Personal communication, S. Kramer.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×

In order for the NGL database to become the model for documenting case histories, the final draft needs to be thoroughly vetted by the geotechnical community. When establishing a model for documenting case histories, however, it should be noted that there are established geotechnical data collection protocols (e.g., those by the Geotechnical Extreme Event Reconnaissance organization)5 that need to be considered. Although it is unlikely that all of the items described in a final protocol will be available for any given case history, providing as much of this information as possible could increase the value and quality of the case histories. Factors necessary to establish the quality of the case history (if not a quality rating system) in the database are important. Engineering practice will improve when case history documentation consistently includes factors that describe the quality of the data, and when case history quality ratings are considered during development, calibration, and validation of post-liquefaction residual shear strength models. Development of the quality rating can be left to individual investigators developing correlations, provided their case history documentation includes those factors necessary to develop a rating system.

___________________

5 See http://www.geerassociation.org/reconnaissance.htm.

Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 49
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 50
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 51
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 52
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 53
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 54
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 55
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 56
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 57
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 58
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 59
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 60
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 61
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 62
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 63
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 64
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 65
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 66
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 67
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 68
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 69
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 70
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 71
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 72
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 73
Suggested Citation:"3 Case Histories." National Academies of Sciences, Engineering, and Medicine. 2021. State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences. Washington, DC: The National Academies Press. doi: 10.17226/23474.
×
Page 74
Next: 4 The Simplified Stress-Based Approach to Triggering Assessment »
State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences Get This Book
×
Buy Paperback | $75.00 Buy Ebook | $59.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Earthquake-induced soil liquefaction (liquefaction) is a leading cause of earthquake damage worldwide. Liquefaction is often described in the literature as the phenomena of seismic generation of excess porewater pressures and consequent softening of granular soils. Many regions in the United States have been witness to liquefaction and its consequences, not just those in the west that people associate with earthquake hazards.

Past damage and destruction caused by liquefaction underline the importance of accurate assessments of where liquefaction is likely and of what the consequences of liquefaction may be. Such assessments are needed to protect life and safety and to mitigate economic, environmental, and societal impacts of liquefaction in a cost-effective manner. Assessment methods exist, but methods to assess the potential for liquefaction triggering are more mature than are those to predict liquefaction consequences, and the earthquake engineering community wrestles with the differences among the various assessment methods for both liquefaction triggering and consequences.

State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences evaluates these various methods, focusing on those developed within the past 20 years, and recommends strategies to minimize uncertainties in the short term and to develop improved methods to assess liquefaction and its consequences in the long term. This report represents a first attempt within the geotechnical earthquake engineering community to consider, in such a manner, the various methods to assess liquefaction consequences.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!