GEOLOGICAL SURVEY CIRCULAR 688
Liquefaction, Flow, and Associated Ground Failure
Liquefaction, Flow, and Associated Ground Failure
By T. Leslie Youd
G E 0 L 0 G I C A L S U R V E Y C I R C U L A R 688
., ... , United States Department of the Interior ROGERS C. B. MORTON, Secretary
Geological Survey V. E. McKelvey, Director
Free on application to the U.S. Geological Survey, Washington, D.C. 20244 CONTENTS
Page Abstract 1 Introduction ...... 1 Published definitions of liquefaction...... 1 Observed behavior and proposed definitions...... 2 Consequences of liquefaction...... 6 Flow landslides ...... 7 Lateral-spreading landslides ...... 8 Quick-condition failures ...... 9 Methods for evaluating liquefaction and ground-failure potential ...... 9 Surmnary ...... 10 References cited ...... 11
ILLUSTRATIONS
Page FIGURES 1-7. Graphs showing: 1. Stress-strain and pore pressure-strain curves for monotonically loaded triaxial compression test on undrained sample of Ottawa sand illustrating liquefaction, flow deformation, and solidification ...... 3 2. Typical stress-strain curves for three monotonically loaded triaxial compression tests on un- drained samples of Ottawa sand...... 4 3. Stress paths for the three triaxial compression tests plotted in figure 2 ...... 4 4. Liquefaction and unlimited flow generated by cyclic loading ...... 5 5. Repeated episodes of liquefaction and limited flow generated by cyclic loading...... 5 6. Test data illustrating the necessity of stress reversals for cyclic loading to produce repeated cycles of liquefaction and limited mobility ...... 6 7. Data showing that resistance to reliquefaction is reduced by a previous episode of liquefaction 7
liquefaction, Flow, and Associated Ground Failure
By T. Leslie Youd
ABSTRACT field and allowed the possibility that analyses of Ambiguities in the use of the term liquefaction and in liquefaction potential could be misinterpreted or defining the relation between liquefaction and ground misapplied. failure have led to encumbered communication between In this paper explicit definitions of liquefaction workers in various fields and between specialists in the and related concepts are given on the basis of same field, and the possibility that evaluations of lique behavior observed in laboratory tests. These con faction potential could be misinterpreted or misapplied. Explicit definitions of liquefaction and related concepts cepts are then used to define relations between are proposed herein. These definitions, based on observed liquefaction and the various types of ground fail laboratory behavior, are then used to clarify the relation ure commonly associated with liquefaction. Ex between liquefaction and ground failure. amples are given to illustrate the character of Soil liquefaction is defined as the transformation of a granular material from a solid into a liquefied state as each type of ground failure. Finally, the applica a consequence of increased pore-water pressures. This bility and limitations of existing methods for definition avoids confusion between liquefaction and pos evaluating liquefaction potential are briefly ex sible flow-failure conditions after liquefaction. Flow amined with the aid of the definitions formulated failure conditions are divided into two types: (1) unlim herein. ited flow if pore-pressure reductions caused by dilatancy during flow deformation are not sufficient to solidify the Appreciation is given to D. H. Gray, E. J. Rei material and thus arrest flow, and (2) limited flow if ley, Kaare Hoeg, K. L. Lee, H. W. Olsen, and they are sufficient to solidify the material after a finite H. B. Seed for comments and suggestions. deformation. After liquefaction in the field, unlimited flow commonly leads to flow landslides, whereas limited PUBLISHED DEFINITIONS OF flow leads at most to lateral-spreading landslides. Quick condition failures such as loss of bearing capacity form LIQUEFACTION a third type of ground failure associated with liquefac The standard technical dictionary definition of tion. liquefaction is "The act or process of transform INTRODUCTION ing any substance into a liquid" (Lange and Because of the critical effect liquefaction has on Forker, 1961, p. 1738). Several definitions of soil the safe performance of engineered construction liquefaction are in basic agreement with this con and the stability of certain geologic formations, cept. For example, Terzaghi and Peck ( 1948, p. considerable study has been devoted to this topic 100) defined spontaneous liquefaction as "the inrecentyears. Significant progress has been made sudden decrease of shearing resistance of a quick in understanding the liquefaction phenomenon sand from its normal value to almost zero withoo&" and the factors controlling it; however, ambiguity the aid of seepage pressure." A generally compa in present definitions of the term liquefaction and rable definition was published by the American lack of clear distinction between liquefaction and Society of Civil Engineers ( 1958, p. 1826-22) and ground-failure conditions associated with this quoted by the American Geological Institute phenomenon have produced some confusion in (1972, p. 410) as follows: "The sudden large usage and understanding of liquefaction and its decrease of shearing resistance of a cohesionless consequences. This confusion has encumbered soil, caused by a collapse of the structure by communication between workers in various dis shock or strain, and associated with a sudden but ciplines as well as between specialists in the same temporary increase of the pore fluid pressure [is liquefaction]. It involves a temporary transfor or amount of ground-failure movement, if any, mation of the material into a fluid mass." Gha that might follow the transformation; (2) a map boussi and Wilson ( 1973) gave the following defi of deposits that could lose strength and acquire nition: "The phenomenon of the loss of strength a degree of mobility sufficient to permit move of saturated granular soils during earthquakes is ments ranging from several feet to several thou generally referred to as liquefaction. The process sand feet; or (3) as a map of deposits that could of liquefaction transforms an element of soil from lose strength and fail in the form of flow land a state of saturated granular solid to a state of slides. In order to reduce this ambiguity and pro viscous fluid." vide for more precise usage, explicit definitions Several definitions for liquefaction and related are needed, and the relation between liquefaction phenomena have been proposed by Seed and his and ground failure needs clarification. coworkers on the basis of behavior observed in cyclically loaded shear tests on saturated samples OBSERVED BEHAVIOR AND PROPOSED of sand. Lee and Seed (1967a, p. 49) define (1) DEFINITIONS complete liquefaction as "when a soil exhibits no Laboratory test data illustrative of the behav resistance (or negligible resistance) to deforma ior of saturated, undrained granular soils during tion over a wide strain range, say a double ampli shear loading (monotonic and repetitive) have tude of 20 percent"; (2) partial liquefaction as been published by several investigators (Castro, "when a soil exhibits no resistance to deforma 1969; Finn and others, 1970; Lee, 1970; Lee and tion over a strain range less than that considered Fitton, 1968; Lee and Seed, 1967a, b; Peacock to constitute failure"; and (3) initial liquefaction and Seed, 1968; Seed and Lee, 1966; 1969). as "when a soil exhibits any degree of partial These data are used here as a basis for formula liquefaction during cyclic loading." In addition ting definitions of liquefaction and related terms, Seed and Idriss (1971, p. 1249) have used the clarifying the relation between liquefaction and term liquefaction to describe "a phenomenon in ground failure, and examining the character of which a cohesionless soil loses strength during an various types of ground failure associated with earthquake and acquires a degree of mobility suf liquefaction. ficient to permit movements ranging from several Some pertinent aspects of the behavior of sat feet to several thousand feet." Except for initial urated sands during conditions of undrained, liquefaction, both a flow deformation require monotonic shear loading are illustrated in the ment and a strength loss requirement are incor curves plotted in figure 1. These data were taken porated into each of the definitions proposed by from test 4-7 of Castro ( 1969), a saturated un Seed and his colleagues. drained triaxial compression test on moderately 1 Castro (1969, p. 7) used the term liquefaction dense (relative density; Dr==4 7 percent ) Ottawa to denote the following phenomenon: "The con sand. Three distinct phases, through which the ventional use of the term [liquefaction] as it will sample passed, are shown on the curves. ( 1) Dur be used throughout this thesis, refers to the phe ing initial loading (beginning of test to point l), nomenon which takes place in a mass of soil dur the sample behaved as a solid (did not percep ing flow slides. Liquefaction or flow failure of a tibly flow (Lange and Forker, 1961, p. 1794) ). sand is caused by a substantial reduction of its During this part of the test, pore pressures in shear strength." This usage in effect defines creased with applied load until a point of insta liquefaction as a condition of unrestrained flow bility was reached at point l. This rise in pore deformation. pressure was caused partly by the tendency of The incorporation or confusion of two different all but very dense samples to compact initially phenomena (strength loss leading to a material during loading (Youd, 1972a, p. 717). (2) Be- phase transformation and flow deformation) into lRelative density, Dr• in percent is defined by the following rela a single definition has led to ambiguities in usage. tionship: For example, a map of sand deposits susceptible e max -e to liquefaction could be interpreted as (1) a map Dr- emax-emin (100) where emax and emin are void ratios of a given san~ in its loosest of deposits that could be transformed into a and densest states, respectively, and e is the void ratio of the sand liquefied state, but with no assessment of type at the density in question.
') !/) agreement with the standard technical definition oz .. 1-0 1 X 105 N/m2 =2090 lb/ft2 (Lange and Forker, 1961, p. 1738), the defini wX~-- 6 za: tions of Terzaghi and Peck (1948, p. 100), the ~~.w 5 American Society of Civil Engineers ( 1958, p. ~:!: ww 1826-22), the American Geological Institute a: a: 4 (1972, p. 410), and Ghaboussi and Wilson (1973). !I):>~--~ a:O For cyclic loading conditions, the definition pro Q!/) 3 1-a: posed above is congruous with the definition of ~w :;:a.. 2 w initial liquefaction given by Lee and Seed (1967a, 0 p. 49).
!/) At the conclusion of the liquefied flow segment z QL---__L__ ___ L__ ___ .L.______, 0"' ~--~ 4.----,----,----.---~ of the test 4-7 (points on the curves), a process ~X wa: opposite to liquefaction occurred-the transfor Zw 3 Zl mation of a granular material from a liquefied -w w·:!: state into a solid state. This process is defined a:w 2 :>a: !/)~ here as solidification. Deformations in the solid !/):::>wo state are generally referred to as elastic strain or g:!/) wa: L_____ L__ ___ L__ ___ .L.______, plastic strain depending on whether deformations a:W 0 QQ. 5 10 15 20 a.. o are recoverable or not, respectively. Deformations AXIAL STRAIN, IN PERCENT in the liquefied state are referred to as flow defor FIGURE 1.-Stress-strain and pore pressure-strain curves mations. for monotonically loaded triaxial compression test on Flow deformation can be divided into two undrained sample of Ottawa sand illustrating lique types. ( 1) Limited flow (equivalent to Castro's faction, flow deformation, and solidification (after test 4-7 Castro, 1969). (1969, p. 19) limited liquefaction or, for cyclic loading conditions, Lee and Seed's (1967a, p. 49) yond the point of instability (point l), the sample partial liquefaction) refers to a condition in flowed in a liquefied state (rapidly deformed which liquefaction occurs and flow deformation under constant total stress). In this phase the ensues but is arrested by solidification after a sample deformed through an axial strain of 6 finite movement without significant change of percent in 1.0 sec. (wavy part of curves). During total stresses in the process. The sample in test the latter part of the flow deformation segment, 4-7 (fig.1) underwent limited flow. (2) Unlimited dilatant tendencies within the sample caused a flow (equivalent to Castro's (1969, p. 16) lique reduction of pore pressure and concomitant faction failure or, in essence, equivalent to Lee strengthening of the sample. This eventually led and Seed's (1967a, p. 49) complete liquefaction) to an arrest of the flow, and the sample was recon refers to a condition in which dilatancy-caused verted into a solid state (points on the curves). reductions in pore pressure are insufficient to (3) Beyond points the sample continued to be arrest flow; thus, flow deformation continues have as a solid. As additional load was applied, unabated until the applied shear stresses are the dilitant tendency continued to cause further reduced to a level less than the viscous shear re pore pressure reductions and greater sample sistance of the liquefied material. At that point strength. solidification occurs. The sample in test 4-4 (fig. On the basis of the behavior illustrated in test 2) underwent liquefaction and unlimited flow. In 4-7 (fig. 1), the following definitions for liquefac this test, flow deformation ceased only when the tion and related terms are proposed. Liquefaction deviator stress was relieved by stopblocks in the is defined as the transformation of a granular ma loading apparatus. terial from a solid state into a liquefied state as a Additional insight into the behavior of satu consequence of increased pore-water pressures. rated sands during undrained shear can be gained This transformation occurred during the initial from an examination of stress paths constructed loading phase of test 4-7 and was complete from published laboratory test data. For example, (liquefaction occurred) at point l on the curves. effective stress paths for the three triaxial com The proposed definition of liquefaction is in close pression tests of Castro ( 1969) listed in figure 2
~ J 8 8
1X 1 05 N/m2 =2090 lb/ft2 1X 1 05 N/m2 =2090 lb/ft2 (I) 7 z 0 o .. X 7 ~~ 6 a: wx w za: 1-w zw -I- 5 :E •W w 6 !2:E a: Ww 4 <( a: a: ::::1 1-<( a (I) ::::I (I) 3 a: a: a w 5 Ocn 0.. 1-a: (I)
red (point s1 ) as a result of a dilatancy-caused remained within the stable domain (path from 1 drop in pore pressure. The inferred stress path to 1u), thus preventing a recurrence of liquefac during this phase of the test is shown by the inter- tion. During the second loading (path from 1u to repeated occurrence of liquefaction and limited
~ 1X 105 N/m2 =2090 lb/ft 2 flow. X a: w The influence of a previous history of liq uefac 1-w tion on the susceptibility of a sand to reliquefac ::E w tion at a later period has been demonstrated by a: c( Finn, Bransby, and Pickering (1970). The ap :::::> 0 C/) proximate stress paths for two liquefaction tests a: w on the same Ottawa sand sample are plotted in a.. zC/) figure 7 (data from fig. 5 (test 40) of Finn and others, 1970). The sample was formed in a tri w~ z axial compression device, saturated, and then 5 ~ consolidated under an isotropic stress of 2.0 X 10 vj 2 2 C/) N/m (4,100 lb/ft ) to a relative density of 50 w a: 1- percent. Cyclically reversing loads were applied C/) .a: to the sample. Pore pressure increased with each 0 1- c( loading until liquefaction occurred during the w> extensional loading segment of the 25th cycle 0 .J (light-line stress path in fig. 7). The sample ~ flowed through an axial strain of 0.6 percent, at X c( which time deformation was arrested by solidifi cation. Several additional loading cycles were
EFFECTIVE SPHERICAL PRESSURE, IN NEWTONS applied (stress paths not shown). With each load PER SQUARE METER X105 reversal, the sample reliquified, flowed, and then FIGURE 6.-Test data illustrating the necessity of stress solidified at the opposite end of the strain excur reversals for cyclic loading to produce repeated cycles sion. After a total of 29 cycles the loading was of liquefaction and limited mobility. Stress path from stopped, leaving the sample stabilized at the com a cyclically loaded triaxial compression test without pressional end of the strain excursion. The sample stress reversals. Liquefaction and limited flow occurred only during first loading. (Data from test 99, Lee, was then reconsolidated under an isotropic stress 5 2 2 1970.) of 2.0 X 10 N /m ( 4,100 lb/ft ). This consolida tion yielded a relative density of 60 percent. 2), the applied shear stresses were resisted imme Cyclic loading was then resumed. This time liq ue diately by increased sample strength as a result faction occurred during the extensional loading of dilatancy-caused decreases in pore pressure. segment of the first cycle (heavy-line stress path This sample hardening eliminated any possibility in fig. 7). From this and similar tests, Finn, of liquefaction or flow during that segment of Bransby, and Pickering (1970) conclude that the test although some plastic straining did occur. resistance to reliquefaction is substantially re During unloading the stresses again traced a path duced by a previous episode of liquefaction. within the stable domain (from point 2 to about point 1u). This behavior was repeated during CONSEQUENCES OF LIQUEFACTION each additional cycle. The stress conditions at maximum load for several of these cycles are Liquefaction by itself poses no particular haz shown by the correspondingly numbered points ard. In fact, during seismic shaking a liquefied on the diagram. (The values decreased because layer at depth could act as an isolator, impeding the sample cross-sectional area increased during the transmission of vibrational energy from the test as a result of plastic straining.) In each underlying layers to s-tructures founded at the cycle the stress paths returned approximately to surface (Seed, 1968; Ambraseys, 1973). Only point 1u during unloading. Thus, other than the when liquefaction leads to some form of perma single episode of liquefaction and flow deforma nent ground movement or ground failure does it tion during initial loading, the sample did not become a serious problem. Three basic types of reliquefy as did those in tests with appreciable ground failures associated with liquefaction have stress reversal. These results clearly show that been identified (Seed, 1968)-flow landslides, shear-stress reversals are necessary to produce landslides with limited displacement, and quick- 0.8
1 X 10' N/m2 =20901b/ft2 c liquefaction Reliquefaction X a: w w~ ~ 0.4 w a: < ::J 0 Cl) a: w CL. zCl) 0 w~ 0.0 z ~ w~ a: ~ Cl) a: 0 ~ < -0.4 >w c _J x< <
EFFECTIVE SPHERICAL PRESSURE, IN NEWTONS PER SQUARE METER X 105
FIGURE 7.-Data showing that resistance to reliquefaction is reduced by a previous episode of liquefaction. Stress paths from cyclically loaded liquefaction and reliquefaction tests on the same Ottawa sand sample. Sample liquefied and flowed in limited flow in both tests. (Data from fig. 5, Finn and others, 1970.) condition failures. In addition to these types of Several flow landslides have occurred in the ground failure, ejection of water and sediments in estuary section of the Dutch province of Zeeland the form of sand boils has been a source of dam (Koppejan and others, 1948), including some age associated with liquefaction during earth that have caused disastrous breaches in the dikes quakes (Ambraseys and Sarma, 1969). of that region. Dike slopes before failure were typically about 36 percent (20°), with banks as FLOW LANDSLIDES high as 40 m ( 130 ft). After failure, many slopes Where conditions are favorable, liquefaction in were 7 percent ( 4 °) or less. Liquefaction of a the field may lead to a state of unlimited flow. In loose granular layer underlying the slides areas this situation, if the mobilized soil is unrestrained, was the apparent cause of the flow failures. sizeable masses of earth materials may travel A flow failure occurred in the Fort Peck Dam, long distances in the form of liquefied flows or Montana, during construction of a hydraulic-fill blocks of intact materials riding on liquefied embankment (Casagrande, 1965). A 250-m flows. In this type of failure, flow ceases only (1,700-ft)-long section of the upstream shell when the driving shear forces are reduced (such failed and moved 460 m (1,500 ft) upstream in as by slope reduction) to values less than the about 4 min. Before failure, the slope of the up viscous shear resistance of the flowing material. stream embankment was about 25 percent (14°); The dimensions and character of typical flow after failure slopes on the failed mass were gen landslides are illustrated in the few examples erally less than 5 percent (3°). The cause of cited below. failure is believed to have been liquefaction of a sand zone in the shell and possibly a natural Several important characteristics of lateral granular layer beneath the dam (Casagrande, spreading landslides associated with liquefaction 1965, p. 10). can be deduced from the laboratory behavior Flow failures commonly are triggered by seis described above. ( 1) Repeated episodes of lim mic shaking (Seed, 1968). Crandall (1908, p. ited flow can only develop if shear stress reversals 249) described one that occurred near San Fran occur during the loading sequence. These rever cisco during the 1906 earthquake sals are more easily accomplished beneath mild Northeast of Mount Olivet Cemetery there was an slopes where static shear stresses are small than earth-flow in the sandy soil at the base of the San Bruno beneath steeper ones. However, if the slope is too Mountains. The angle at which the materials slid was gentle, the gradient may not be sufficient to cause hardly more than 10 degrees [18 percent]. The sand and movement. Thus mild slopes should be most sus water forming this slide came out of a hole several hun ceptible to this type of failure. dred feet long and 150 feet [ 46 m] wide, flowed down the hill several hundred yards toward the cemetery, carried (2) With each occurrence of restabilization, away a pile of lumber, and knocked the power-house the mobilized soil in the failure zone receives a from its foundations. The front of the mud-flow piled up dilatant impulse that could cause loosening if in a bank when it reached the nearly level ground, and excess water were available to fill the created dammed up the mass behind it. voids. Excess water is often available during earthquakes as a result of compaction of other LATERAL-SPREADING LANDSLIDES granular materials in the section. Thus, it is pos Where conditions are favorable for liquefaction sible for loosening to occur which, in turn, would but the sediments are too dense to allow unlimited allow displacements to increase with each episode flow, limited flow may develop. On sloping ground of limited flow and could eventually lead to a this may allow finite downslope movements fol condition of unlimited flow. lowed by dilatancy-caused solidification. Other (3) At the conclusion of a series of limited-flow factors that may also limit downslope displace episodes, the soil in the failure zone is commonly ment include rapid drainage and liquefaction of left in a dilatant restabilized condition. The soil materials along only part of the potential failure itself may be denser or looser or the same as it plane (Seed, 1968). In many instances, however, was before the disturbance, depending on whether dilatancy-caused solidification appears to be the pore water migrated into or out of the liquefied most important factor in limiting displacement. soil during shear. In any case, and even after re One episode of limited flow may produce displace consolidation, the soil may be left in a particu ments of negligible importance; on the other larly vulnerable condition for reliquefaction dur hand, cumulative displacements generated by ing a subsequent disturbance. repeated episodes of limited flow, such as might ( 4) Repeated episodes of limited flow can con occur during an earthquake or other source of tinue only as long as strong ground shaking pro repeated loading, could be substantial. Descrip ducing stress reversals continues. Thus, lateral tion of limited displacement landslides generated spreading landslides associated with liquefaction by liquefaction are common in the earthquake would normally move only during periods of literature, but they are described under different strong shaking and restabilize immediately upon nomenclature. For example, "earth lurches" the cessation of shaking. (Richter, 1958, p. 124), land spreading (McCul The following few examples of lateral-spread loch and Bonilla, 1970), and lateral spreading ing ground failures illustrate their typical dimen (Oldham, 1899, p. 87-90; Youd, 1973) are, in sions and character. fact, limited displacement ground failures asso During the 1971 San Fernando earthquake, ciated with liquefaction. The term "lateral lateral-spreading ground failures occurred on spreading landslide" is used here to denote this both sides of Van Norman Lake (Youd, 1971; type of failure because it is commonly used, it 1973; Smith and Fallgren, 1973; Proctor and has been rigorously defined (Varnes, 1958, p. others, 1972). Northeast of the lake the land 29-32), and many failures described in the liter slide was tongue shaped in plan, 1.2 km ( 4,000 ature fit the description of lateral spreading given ft) long, and about 0.3 km (1,000 ft) wide (aver by Varnes. age). The mean slope from head to toe of the failure was about 1.5 percent (0.9°). Horizontal several lateral-spreading ground failures occurred displacements were as large as 1.9 m ( 5. 7 ft), in the city of San Francisco. One, in an area of whereas maximum vertical displacements were filled marshland, was approximately 2 blocks only about 0.15 m (0.5 ft). The surface layer was wide and extended over a 1.5-km (5,000-ft) dis fractured into several large blocks that slid down tance from Eighth and Mission Streets to the slope with very little tilting. Sand boils, indica vicinity of Fourth and Brannan Streets. Wood tive of increased pore pressures at depth, erupted ( 1908) described this area as follows at several points on and near the slide. The soil The fissuring and slumping and the buckling of block profile beneath the slide consisted of a firm sur and asphalt pavements into little anticlines and syn face layer overlying a soft saturated layer com clines (arches and hollows) , accompanied by small open cracks in the earth, characterize the land surface. This posed of sand and sandy silt. Liquefaction in the slumping movement or flow took place in the direction lower layer presumably led to the failure by of the length of the area [down about a 0.8 percent repeated episodes of limited flow. Relict preearth (0.5°) slope], and its amount was greatest near the quake fissures and sand boils that were found in center or channel, where the street lines were shifted postearthq uake exploratory trenches are evidence eastward out of their former straight courses by amounts varying from 3 ft to 6 ft [1 m to 2m]. that similar ground failures had occurred before These examples, which are only a few of many at this site, most likely during previous earth that could be cited, show that lateral-spreading quakes (Youd, 1972b). failures on mild slopes are very common during West of the lake, the surface ruptures broke moderate and strong earthquakes. roughly parallel to the long free face formed by the west shore of Van Norman Lake and severely QUICK-CONDITION FAILURES disrupted the newly constructed fill for the J en Seepage forces, caused by upward-percolating sen water-filtration plant. Maximum horizontal pore water, often reduce the strength of granular displacements were about 1 m ( 3 ft) near the free soils to a point of instability. This state is termed face and decreased with distance upslope. Sand a "quick condition." Artesian pressures are one boils were found near the toe of the landslide and cause of upward seepage; seismic compaction of on the fill in the southern part of the zone. A saturated granular materials at depth is another. layer of saturated fine sand was discovered in This condition is generally restricted to sand postearthquake borings at depths of 2-3 m (6-9 layers of significant thickness that extend from ft) below the original ground surface (Proctor below the water table to the surface. and others, 1972). The failure presumably was Loss of bearing capacity is the most common caused by liquefaction of this layer. type of failure produced by a quick condition. Reports from recent Alaskan earthquakes in Buoyant rise of buried tanks and other vessels is clude numerous descriptions of lateral-spreading another common effect. Shear deformations dur type ground movements down mild slopes toward ing such failures could be of either the unlimited free faces; these ground movements were accom or limited-flow types. Bearing-capacity failures panied by fissuring and ejection of subsurface illustrative of this type of failure were reported water. For example, McCulloch and Bonilla during the 1964 Niigata, Japan, earthquake. In (1970, p. D-1) report that during the March 27, the city of Niigata, several high-rise apartment 1964 earthquake buildings rotated and subsided into the liquefied a general loss of strength [was] experienced by wet, soil (Seed and Idriss, 1967). Railroad embank water-laid unconsolidated granular sediments (silt to ments, especially the Echigo line, subsided into coarse gravel that allowed embankments to settle and the "liquefied ground" (Kobayashi, 1969). enabled sediments to undergo fiowlike displacement toward topographic depressions, even in flat-lying areas METHODS FOR EVALUATING LIQUEFAC * * * Stream widths decreased, often about 20 inches [0.5 m] but at some places by as much as 6.5 ft [2.1 m], TION AND GROUND-FAILURE POTENTIAL and sediments moved upward beneath stream channels Existing methods for evaluating liquefaction * * * Ground cracks * * * commonly extended 500 ft [160 m], and occasionally 1,000 ft [320 m], back from and associated ground-failure potential are gen streams * * * Sediment-laden ground water was dis erally based on one of two criteria: ( 1) the num charged from the cracks. ber of cyclical loadings required to produce a During the 1906 San Francisco earthquake, liquefied condition or a certain amount of flow
9 deformation considered to constitute failure; or cycles of loading after liquefaction (or initial (2) the smallest void ratio (critical void ratio) at liquefaction) to produce significant ground-failure which unrestrained liquefied flow (unlimited movements in denser sediments (for example, flow) can develop. Both of these criteria are relative densities greater than 75 percent). In controlled largely by the same general factors; addition, ground slope, a factor of primary im however, the influence of these factors may be portance in evaluating ground-failure potential, considerably different for each situation. For ex is not incorporated in the simplified procedure. ample, relative density is a primary factor con Thus, while the simplified procedure may be valid trolling both liquefaction of sands under cyclical for evaluating liquefaction potential it is not loading conditions and unlimited flow failure. necessarily valid for evaluating ground-failure Liquefaction can be induced by cyclical loading, potential. More rigorous methods, such as those however, at considerably greater relative densi developed for earth dams (Seed and others, ties than those at which unlimited flow can occur. 1973), are necessary for evaluating the latter Thus different criteria have been developed for potential. each approach. The critical void ratio approach has been devel Seed and his coworkers (Lee and Seed, 1967a, oped by Casagrande (1940) and Castro (1969) b; Peacock and Seed, 1968; Seed andidriss, 1967, chiefly for monotonic loading conditions, the 1971; Seed and Lee, 1966, 1969; Seed and Pea principal application being the differentiation of cock, 1971) have developed the criteria for re conditions under which unlimited and limited peated loading. One application of this procedure flow may occur. With additional research, partic has been the analysis of stability and deformation ularly on the effect of repetitive loadings, this of earth dams during earthquakes (Seed, 1966; method could be made more general and could Seed and others, 1969, 1973). Similar procedures possibly be applied to evaluating type and should be generally applicable for analyzing lique amount of ground-failure movement after lique faction potential and ground-failure movements faction. With such development the two methods beneath natural slopes. would materially complement each other. In addition, Seed and Idriss ( 1971) formulated a "simplified procedure" for evaluating liquefac SUMMARY tion potential. This method incorporates several Explicit definitions for liquefaction, solidifica simplifying approximations and is valid only for tion, limited flow, and unlimited flow have been sediments with relative densities less than 80 given and the relation between these phenomena percent that lie beneath level surfaces. Although and ground failure examined in an attempt to the definition of liquefaction given by Seed and present a clear and integral description of lique Idriss (1971, p. 1249, cited earlier) seems to indi faction and its consequences. Also, methods for cate that the simplified procedure is generally evaluating liquefaction and ground-failure poten valid for predicting ground-failure potential as tial have been briefly examined. well as liquefaction potential, examination of cri Liquefaction is defined as the transformation teria used in formulating the procedure shows of a granular material from a solid state into a that this is not true. "Initial liquefaction" (Lee liquefied state as a consequence of increased pore and Seed, 1967a, p. 49) was the criterion used by water pressures. Solidification is defined as the Seed and Idriss to determine the point at which opposite process, that is, the transformation of a liquefaction occurred in laboratory tests utilized granular material from a liquefied state into a in formulating the procedure. (Initial liquefaction solid state. Once liquefied, a granular material is is defined as the point at which a granular mate free to flow, provided there is sufficient gradient, rial exhibits any degree of liquefied flow during until solidification occurs. If solidification occurs cyclic loading.) Although significant ground as a result of dilatancy with very little change in failure movements may occur almost simultan total stress, the condition is termed limited flow. eously with initial liquefaction in some loose If solidification occurs only when driving forces granular soils (for example, relative densities less are reduced to a level less than the viscous shear than 65 percent), this is not generally true for ing resistance of the material, the condition is denser soils. For example, it may require many termed unlimited flow. Three types of ground failure conditions com Crandall, Roderic, 1908, The San Francisco Peninsula, monly follow liquefaction in the field. (1) Un in The California Earthquake of 1906: Carnegie Inst. Washington, v.l, p. 246-254. limited flow commonly leads to flow landslides. Finn, W. D., Bransby, P. L., and Pickering, D. J., 1970, (2) Limited flow commonly leads to lateral Effect of strain history on liquefaction of sand: Am. spreading landslides. One episode of limited flow Soc. Civil Engineers Proc., Jour. Soil Mechanics may be inconsequential; however, a series of epi and Found. Div., v. 96, no. SM6, p. 1917-1934. sodes, such as might occur during an earthquake, Ghaboussi, Jamshid, and Wilson, E. L., 1973, Liquefac often leads to significant permanent ground tion and analysis of saturated granular soils: World Conf. on Earthquake Eng., 5th Rome 1973, Proc., movements. Many phenomena variously de (in press). scribed in the earthquake literature as lurching Kobayashi, Yoshimasa, 1969, Mechanism of earthquake or land spreading are, in fact, lateral-spreading damage to embankments and slopes: World Conf. landslides generated by liquefaction. (3) Quick on Earthquake Eng., 4th, Chile 1969, Proc., v. 3, p. condition failures such as loss of bearing capacity (A-5)74-(A-5)87. Koppejan, A. W., van Wamelan, B. M., and Weinberg, form the third type of ground failures attribut L. J. H., 1948, Coastal flow slides in the dutch prov able to liquefaction. ince of Zeeland: Internat. Conf. on Soil Mechanics Two approaches for evaluating liquefaction and Foundation Eng., 2d, Rotterdam 1948, v. 5, p. and ground-failure potential have been proposed. 89-96. Lange, N. A., and Forker, G. M., 1961, Handbook of The procedure developed by Seed and his col chemistry [lOth ed.]: New York, McGraw-Hill, leagues provides a method for evaluating lique 1969 p. faction and ground-failure potential under cycli Lee, K. L., 1970, Triaxial compressive strength of satu cal loading conditions. The procedures developed rated sands under seismic loading: Univ. California, by Casagrande and Castro were basically de Berkeley, Ph.D. thesis, 521 p. Lee, K. L., and Fitton, J. A., 1968, Factors affecting the signed to differentiate conditions under which cyclic loading strength of soil: Am. Soc. Testing and unlimited and limited flow occur during mono Materials Spec. Tech. Pub. 450, p. 71-95. tonic loading conditions. With further develop Lee, K. L., and Seed, H. B., 1967a, Cyclic stress condi ment these two approaches should become com tions causing liquefaction of sand: Am. Soc. Civil plementary and lead to improved understanding Engineers Proc., Jour. Soil Mechanics and Found. Div., v. 93, no. SMl, p. 47-70. of the phenomena involved. ---1967b, Dynamic strength of anisotropically con solidated sand: Am. Soc. Civil Engineers Proc., REFERENCES CITED Jour. Soil Mechanics and Found. Div., v. 93, no. SM5, p. 169-190. Ambraseys, N. N., 1973, Dynamics and response of foun McCulloch, D. S., and Bonilla, M. G., 1970, Effects of the dation materials in epicentral regions of strong earthquake of March 27, 1964, on the Alaska Rail earthquakes: World Conf. on Earthquake Eng., 5th, road: U.S. Geol. Survey Prof. Paper 545-D, 161 p. Rome 1973, Proc., (in press). Oldham, R. D., 1899, Report on the great earthquake of Ambraseys, N. N., and Sarma, S., 1969, Liquefaction of 12th June, 1897: India Geol. Survey Mem., v. 49, soils induced by earthquakes: Seismol. Soc. America 379 p. Bull., v. 59, no. 2, p. 651-664. Peacock, W. H., and Seed, H. B., 1968, Sand liquefaction American Geological Institute, 1972, Glossary of geol under cyclic loading simple shear conditions: Am. ogy: Washington, D.C., 857 p. Soc. Civil Engineers Proc., Jour. Soil Mechanics and Found. Div., v. 94, no. SM3, p. 689-708. American Society of Civil Engineers, 1958, Glossary of Proctor, R. J., Crook, R., Jr., McKeown, M. H., and Mor terms and definitions in soil mechanics: Am. Soc. esco, R. L., 1972, Relation of known faults to surface Civil Engineers Proc., Jour. Soil Mechanics and ruptures, 1971 San Fernando Earthquake, Southern Found. Div., v. 84, no. SM4, pt. 1, p. 1826-1-1826- California: Geol. Soc. America Bull., v. 83, no. 6, p. 43. 1601-1618. Casagrande, Arthur, 1940, Characteristics of cohesion Richter, C. F., 1958, Elementary seismology, San Fran less soils affecting the stability of slopes and earth cisco, W. H. Freeman, 768 p. fills: Contributions to Soil Mechanics 1925-1940, Schofield, Andrew, and Wroth, Peter, 1968, Critical state Boston Soc. of Civil Engineers, Boston, p. 257-276. soil mechanics: New York, McGraw-Hill, 310 p. --1965, Role of the 'calculated risk' in earthwork and Seed, H. B., 1966, A method for earthquake resistant foundation engineering: Am. Soc. Civil Engineers design of earth dams: Am. Soc. Civil Engineers Proc., Jour. Soil Mechanics and Found. Div., v. 91, Proc., Jour. Soil Mechanics and Found. Div., v. 92, no. SM4, p. 1-40. no. SMl, p. 13-41. Castro, Gonzalo, 1969, Liquefaction of sands: Harvard ---1968, Landslides during earthquakes due to soil Soil Mechanics Ser. no. 81, 112 p. liquefaction: Am. Soc. Civil Engineers Proc., Jour.
11 Soil Mechanics and Found. Div., v. 93, no. SM5, p. converter station, in San Fernando, California, 1053-1122. Earthquake of February 9, 1971: U.S. Dept. of Com Seed, H. B., and ldriss, I. M., 1967, Analysis of soil lique merce, Nat'l. Oceanog. and Atmospheric Adm. (in faction: Niigata earthquake: Am. Soc. Civil Engi press). neers Proc., Jour. Soil Mechanics and Found. Div., Terzaghi, Karl, and Peck, R. B., 1948, Soil mechanics in v.94,no.SM3,p.83-108. engineering practice: New York, John Wiley and ---1971, Simplified procedure for evaluating soil Sons, 566 p. liquefaction potential: Am. Soc. Civil Engineers Varnes, D. J., 1958, Landslide types and processes, chap. Proc., Jour. Soil Mechanics and Found. Div., v. 97, 3, in Eckel, E. B., ed., Landslides and engineering no.SM9,p.1249-1273. practice: Nat'l. Research Council, Highway Re Seed, H. B., and Lee, K. L., 1966, Liquefaction of satu searc Board Spec. Rept. 29, NAS-NRC Pub. 544, rated sands during cyclic loading: Am. Soc. Civil p. 20-47. Engineers Proc., Jour. Soil Mechanics and Found. Wood, H. 0., 1908, Distribution of apparent intensity in Div., v. 92, no. SM6, p.105-134. San Francisco, in The California earthquake of ---1969, Pore-water pressure in earth slopes under April 18, 1906: Carnegie Inst. Washington, v. 1, p. seismic loading conditions: World Conf. on Earth 220-246. quake Eng., 4th, Chile 1969, v. 3, p. (A-5) 1-(A-5)- 11. Youd, T. L., 1971, Landsliding in the vicinity of the Van Seed, H. B., Lee, K. L., and ldriss, I. M., 1969, Analysis Norman Lakes in The San Fernando, California, of Sheffield Dam failure: Am. Soc. Civil Engineers earthquake of February 9, 1~71: U.S. Geol. Survey Proc., Jour. Soil Mechanics and Found. Div., v. 95, Prof. Paper 733, p. 105-109. no. SM6, p. 1453-1490. ---1972a, Compaction of sands by repeated shear Seed, H. B., Lee, K. L., ldriss, I. M., and Makdisi, F. 1., straining: Am. Soc. of Civil Engineers Proc., Jour. 1973, Analysis of slides in the San Fernando Dam Soil Mechanics and Found. Div., v. 98, no. SM7, p. during the earthquake of February 9, 1971: Univ. 709-725. California, Berkeley, Rept. no. EERC 73-2, 150 p. ---1972b, Landslides in the vicinity of the Van Nor Seed, H. B., and Peacock, W. H., 1971, Test procedures man Lakes [abs.]: Seismol. Soc. America, Eastern for measuring soil liquefaction characteristics: Am. Sec., Earthquake Notes, v. 43, no. 1, p. 15. Soc. Civil Engineers Proc., Jour. Soil Mechanics and ---1973, Ground movements in the Van Norman Found. Div., v. 97, no. SM8, p. 1099-1119. Lake vicinity, in San Fernando, California, earth Smith, J. L., and Fallgren, R: B., 1973, Ground displace quake of February 9, 1971: U.S. Dept. of Commerce, ment at San Fernando Juvenile Hall and Sylmar Natl. Oceanog. and Atmospheric Adm. (in press).
,'; GPO 7!ll-564 12