22. measurements and estimates of engineering and other physical properties… · 2007-05-17 ·...
TRANSCRIPT
22. MEASUREMENTS AND ESTIMATES OF ENGINEERING ANDOTHER PHYSICAL PROPERTIES, LEG 19
Homa J. Lee, U. S. Naval Civil Engineering Laboratory, Port Heuneme, California
INTRODUCTION
In order to design structures for the sea floor it isnecessary to have some knowledge about how the bottomsediments will behave, in particular, how they will deformafter the structure has been placed. The parameters whichdescribe this behavior are identified as engineering prop-erties, and a new field of study, often termed "marinegeotechnique," has developed to investigate these prop-erties and their influence. An effort is currently underwayto measure engineering properties throughout the world sothat ultimately empirical correlations between sedimenttype and engineering behavior can be developed. The DeepSea Drilling Project provides a unique opportunity forobtaining data relative to this effort since the distributionof properties over large sediment depth ranges can bedetermined. These distributions are often easier to correlatewith sediment type than are the near-surface propertyvalues which can be obtained from tests on conventionalcores. As a complicating factor, however, the disturbanceproduced by drilling is so severe that each piece ofmeasured engineering property data must be criticallyevaluated in terms of how it relates to the in situengineering behavior. In most cases a large deviationbetween laboratory and in situ values will exist. During Leg19 the physical properties program was expanded toinclude the measurement of properties which relate toengineering behavior. This was done in full recognition ofthe role disturbance might play and attempts were made tocompensate for this disturbance.
The physical and engineering properties measured onshipboard during Leg 19 were as follows:
1) Natural gamma radiation.2) Density (by GRAPE and water displacement
methods).3) Water content (from 1-gram syringe samples).4) Vane shear strength.5) Residual negative pore water pressure.6) Acoustic velocity.At a shore-based laboratory (U.S. Naval Civil Engineer-
ing Laboratory) additional physical and engineering prop-erties measured were as follows:
7) Water content (from 10-gram samples).8) Grain density.9) Compressibility.10) Triaxial shear strength (one test).A brief description of the procedures followed in making
these measurements is provided below.
PROCEDURES
Natural Gamma Radiation
Natural gamma-ray measurements were made on1.5-meter core sections using a scan system based on adesign provided by the Marathon Oil Company. This systemis described in detail in the Initial Reports of the Deep SeaDrilling Project (Peterson, Edgar et al., 1970). The equip-ment was removed from the Glomar Challenger inYokohama to provide additional space in the core lab.Since these readings are no longer being made by DSDP, theLeg 19 measurements are not presented in this volume.
Shipboard Density Measurements
GRAPE DENSITY
The Gamma-Ray Attenuation Porosity Evaluator(GRAPE) was used onboard ship to estimate the wet-bulkdensity of the Leg 19 cores prior to splitting. Theequipment and its use have been described by Evans(1965), Harms and Choquette (1965), and Peterson, Edgaret al. (1970).
The device functions by measuring the attenuation ofthe intensity of a gamma-ray beam as it passes through acore. For most materials this attenuation is exponentiallyrelated to the number of electrons in the path of thegamma-ray beam. Also, for most materials of geologicsignificance, the ratio of electron density to mass density isapproximately constant. Therefore, the amount of gamma-ray attenuation is nearly proportional to the material massdensity as long as the material thickness does not vary.
With these general principles available, it is quite simpleto progress from measured gamma-ray attenuation toGRAPE density. The procedure followed on Leg 19 wasslightly more involved since it was desired to remove anydependance upon the absolute numbers produced by theequipment. Setting and electronic errors could thereby beremoved. The procedure that was followed required thatstandard aluminum and water samples be run before eachcore. Since it was known what density values thesematerials should produce, it was possible to use thesereadings to calibrate the readings made on cores. Theiterative procedure required to accomplish this wasexecuted onshore using a computer and digitized dataobtained from the original analog data acquired onboardship.
The GRAPE densities can be converted into porosities ifa grain density is measured or assumed. In this volume the
701
H. J. LEE
GRAPE densities provided in the site summary sectionsmay be converted into porosities using the porosity scaleprovided at the top of each plot.
WATER DISPLACEMENT DENSITIES
The firm material at moderate to great sediment depthsgenerally does not fill the core barrels. In some cases a"wafers and paste" structure results in which pieces ofrelatively undisturbed material are found in a matrix ofbadly disturbed, almost fluid material. In other cases the"paste" is missing and the core liner is filled with irregularlyshaped pieces of material. In these situations, GRAPEdensity measurements would be in error since the require-ment of constant material thickness is not met. It ispossible to compensate for these errors by measuring thethickness of the sample pieces and adjusting the GRAPEdensities accordingly. This procedure was not used on Leg19. Instead, large portions of samples (about 1 kg) wereweighed on a beam balance onboard ship to the nearest 5grams. These were then submerged in a water-filled grad-uated cylinder to measure volume to the nearest 5 ml. Theaccuracy resulting from this sort of procedure is about plusor minus 1.5 percent. These data are presented as "waterdisplacement densities" on the site summary density plots.
Shipboard Water Content
A gimbal-mounted Kahn balance was available aboardthe Glomar Challenger for accurately measuring weights lessthan 1 gram. Small (1 to 2 cc) samples were taken with acalibrated plastic syringe and weighed wet. The sampleswere then dried at 105°C for 24 hours and weighed dry.The water content was calculated using the usual procedurewithout salt correction. Volume measurements were alsomade so that porosity and density could be calculated. Acomparison of the various density and water contentmeasurements is given in a later section.
Vane Shear Strength
The U.S. Naval Civil Engineering Laboratory (NCEL)miniature vane shear device was used onboard ship duringLeg 19. The device functions in the usual manner (Brand,1967). A four-bladed vane is inserted into a sedimentsample and the vane is torqued until a peak reading isrecorded. The shear strength is calculated from the peaktorque assuming complete mobilization of shearing resist-ance over the surface of a cylinder inscribed about the vane.The NCEL vane has the capacity of operating at a variety ofrotation speeds, although for Leg 19, the rotation speed of83 degrees/min was chosen. This speed is significantlyhigher than the 6 degrees/min usually used in vane sheartesting and it has been shown (Migliori and Lee, 1971) thatvariations in rotation speed of this magnitude can causechanges in measured strength on the order of 10 percent.The faster speed was chosen because it permitted theperformance of many more tests and anticipated changes instrength resulting from disturbance (at least 100%) weremuch greater than any possible rate effects.
The vane tests were performed with a 1/2" × 1/2" vaneinserted into samples that had been split. The axis of thevane was perpendicular to the core axis, also a procedurewhich differs from the norm but which should introduce
errors small in comparison with disturbance effects. Apresentation and discussion of the vane shear test results isprovided in a later portion of this section.
Residual Negative Pore Water Pressure
For reasons which will be explained later, sedimentsamples develop pressures in their pore water which are lessthan atmospheric (negative gage). As will also be discussedlater, it is desirable to know the magnitude of thesepressures in order to assess the extent of disturbance towhich the samples have been subjected. These pressureswere measured during Leg 19 using a somewhat standardtechnique (Gibbs and Coffey, 1969) which required the useof a fine-grained, saturated, ceramic disk. One side of theceramic disk was connected to a sealed water-filled chamberwhose pressure could be monitored with a pressure trans-ducer. The other side of the disk was placed in contact witha sediment surface.
The principle of operation is as follows. The sediment,having an affinity for water by virtue of its negative porewater pressure, attempts to draw water out of the ceramicdisk. To achieve this, another fluid (e.g. air) would have tobe drawn into the disk to maintain continuity. Since thepore size of the disk is very small, the menisci at thewater-air interfaces are strong enough to prevent airintrusion. With water and air flow prohibited, the pressurein the ceramic disk and the sealed chamber are forced toequalize with the pressure in the sediment. It is thenpossible to make a pressure reading with the transducer.
Acoustic Velocity
Compressional wave velocities were measured with theHamilton Frame System which has been described pre-viously (Edgar, Saunders et al., in preparation). All testswere performed on portions of sediment which had beencompletely removed from the core liner.
Shore Laboratory Water Content
As a check on the other density measurements, a largenumber of approximately 10-gram samples were takenaboard ship, sealed, shipped to a shore laboratory (NCEL),and subjected to the usual water content test. Since thesamples were sealed in a container within another sealedcontainer, little change in moisture content prior to testingwould have been possible.
Grain Density
For a saturated sediment it is possible to calculate bulkdensity given only the water content and grain density(Bennett et al., 1971). In order to utilize the 10-gram watercontent samples for density determination, a number ofthese were subjected to grain density testing. Oven-driedsamples were weighed and then placed in a Beckmanair-comparison-type pycnometer to determine their volume(Hironaka, 1966). Only a few grain densities were measuredon samples from each.site. However, each major sedimenttype from each site was tested, and the results wereextrapolated to other samples of the same type.
702
MEASUREMENTS AND ESTIMATES OF ENGINEERING AND OTHER PHYSICAL PROPERTIES
Compressibility
Several short unsplit samples were obtained at most ofthe sites. About half of these were taken from the smallmetal sleeve which is inserted into the plastic core linerimmediately above the core retainers. The material fromthe sleeve appears to be less disturbed than that of theremainder of the core. The other half of the unsplit samplestaken were so-called zero sections. Occasionally drilling fornine meters produced more than nine meters of sample,possibly as a result of sample expansion during retrieval.The extra material is labeled the zero section and issomewhat more disturbed than the remainder of the core.However, since high-recovery cores tend to be less disturbedthan low-recovery cores, the disturbance of zero sections isapproximately average for the site. In general, of course, allDSDP cores are moderately to highly disturbed in anengineering sense.
The unsplit samples were taken so that consolidation1
tests could be performed at a shore laboratory (NCEL). Inall, eight of these tests were performed on samples fromfive sites (183, 184, 185, 188, 190). The procedures were asfollows. Cylindrical samples 5.91 cm in diameter and1.99 cm in height were cut from the unsplit sample sections.These were inserted into a rigid ring system so that nolateral sample deformation could occur. Porous stones wereplaced on the sample-free surfaces and the sample-stone-ring system was submerged in seawater. The sample wasthen loaded axially and the reduction in sample heightmeasured. Loading was achieved with a Karol-Warner airpressure loading frame without back pressure. Loadingfollowed the usual pattern; an initial load was applied forabout 24 hours. The load was then progressively doubleduntil the equipment limit (about 25 kg/cm2) was reached.The sample was allowed to compress under each load forabout 24 hours. After the peak load was reached, thesample was unloaded, generally in two large steps, so thatthe swelling potential could be determined.
Good discussions of the theoretical implications andpractical applications of consolidation testing are providedby Terzaghi (1943) and Taylor (1948).
Triaxial Shear Strength
NCEL has developed techniques for performingconsolidated-unchained triaxial shear tests with pore pres-sure measurements on sea floor sediments. The usualprocedures are followed (Bishop and Henkel, 1962) withmodification to allow seawater to be used as a pore fluid.The tests are run under back pressure (about 3.5 kg/cm2)developed with a mercury pot system.
RESULTS AND ANALYSIS
Introduction
One of the primary objectives of foundation engineers isto be able to predict whether a proposed foundation oranchoring system will perform satisfactorily during the
operational life of the structure it supports. "Satisfactory"performance generally implies that the structure does notmove more than a predetermined amount as a result ofmotions in the underlying soil. It is desirable then to beable to predict these motions so that the adequacy of theproposed foundation design can be evaluated. To do thisrigorously it would be necessary to determine a generalstress-strain-time function for every element of soil affectedby the structure. This is currently impossible even on land,although great progress is being made as a result of thedevelopment of the finite element method on analysis. Onthe sea floor, the state of the art is much more primitivebecause of the difficulty involved in measuring the sedi-ment properties. Here it is usually adequate to be able topredict whether the soil will rupture completely under thestructural loading or whether an extremely large amount ofsediment compression will occur. To predict these events itis necessary to know the sediment shear strength (shearstress at which rupture occurs) and the sediment compress-ibility (change in volume which results from a change innormal stress). Even these parameters are quite complex inthat they are interrelated and both very dependent upontime. During Leg 19 only two sediment engineeringproperties were investigated in detail: the short term(undrained) shear strength and the intermediate (drainedbut without secondary effects) compressibility.
Vane Shear Strength
BACKGROUND
The term "consolidation" will be used in this report to refer tothe compression of sediments under increasing pressures. This isapproximately analogous to the geologic term, "compaction."
Hvorslev (1949) suggested a number of corer designparameters which would be expected to lead to goodquality (in an engineering sense) sediment samples andwhich have generally been accepted by the engineeringprofession. As an example of one of these parameters,Hvorslev suggested that the corer length should not exceedits inside diameter by more than 20. The Deep Sea DrillingProject length to diameter ratio is about 140. The otherparameters of Hvorslev are similarly violated and drillingintroduces even more disturbance. Therefore, the Deep SeaDrilling Project cores are highly disturbed in an engineeringsense. Engineering property measurements are practicallymeaningless unless this disturbance is considered.
In order to compensate for disturbance it is necessary todevelop a means of measuring it quantitatively. A commonprocedure for doing this is to utilize the residual stressesretained by a sediment after sampling. In situ, the sedi-ments are under a state of compressive stress. Most of thisstress is attributable to water pressure, u, and may besubstracted from the total stress, σ, to yield what is referredto in soil mechanics as the effective stress, ~σ.
~ö=o - u
It has been shown that soil response is almost alwaysdetermined by the effective rather than the total stress. Foran element of sediment in situ the vertical effective stress,övo, is simply equal to the cumulative buoyant weight ofsediment overlying it. The in situ horizontal effective stress,cFßo, is usually less than the vertical effective stress, oftenby a factor of about 50 percent. As a result, it may be said
703
H. J. LEE
that sediment in situ is under a state of anisotropic (stressesvary with direction), compressive, effective stress. When asample is taken, there is a tendency for the sample toexpand as a result of the removal of stresses. However, inorder for any significant amount of expansion to occur, it isnecessary for additional fluid to be drawn into the sedimentpores. This occurrence will be prevented if the sediment isfine grained and is handled very carefully. Menisci will formin the pores at the sediment surface and prevent fluidintrusion and volume change. However, the residual tend-ency toward expansion will be reflected in the developmentof negative gage pressures in the pore water. Referring toEquation 1, it may be seen that a state of zero total stress,σ, and negative pore water pressure, u, produces a positive(compressive) effective stress, cf. Since water cannot main-tain shear stresses, the state of stress in the sample isnecessarily isotropic (i.e., equal in all directions). In general,therefore, sampling causes a transformation in the state ofeffective stress within the sediment from one of anisotropiccompression to one of isotropic compression. If nothingelse occurs in the way of disturbance, the situation isreferred to as "perfect sampling." It is possible to approachperfect sampling in a se mi the ore tic al fashion and arrive atthe following equation for the residual pore water pressure,ups, which would develop if perfect sampling could beaccomplished (Ladd and Lambe, 1963): ~
Ups = ~σvo [Ko + Au (1 ~KO) (2)
where
Ko = Coefficient of lateral earth pressure at rest
Au = "perfect sampling" pore pressure parameter.
Both Au and Ko are related to sediment type and varyfrom about —0.1 to +0.3 and 0.4 to 0.7 respectively (Laddand Lambe, 1963). Typical values for average cohesive soilsare about +0.1 for^4u and 0.5 foxK0. Inserting these valuesinto Equation 2 yields
lps = -0.55 σvo (3)
This relation indicates approximately what pore pres-sures would develop under perfect sampling conditions. Ofcourse, perfection is never achieved and all samples aresomewhat disturbed. Disturbance causes a reorientation ofgrains, breaking of interparticle bonds, and an overalldecrease in the tendency of a sample to expand. Theresulting residual pore water pressure, Ur, is, therefore, lessnegative than ups. It has been suggested by Ladd andLambe (1963) and others that the ratio ur/ups is a goodmeasure of the amount of disturbance to which the samplehas been subjected. This hypothesis was tested by Lee (inpress) on several terrigenous seafloor soils and it was foundthat the ratio of measured in situ to laboratory vane shearstrengths could be correlated relatively well with urjups.
RESULTS
About 700 measurements of vane shear strength weremade on shipboard during Leg 19. In addition, about 60measurements of residual negative pore pressure were made
at locations of vane tests. The sediments which appeared tobe less disturbed were vane shear tested. The residual porepressure tests were performed on the highest qualitysamples available for each depth range.
The measured laboratory vane shear strengths, Si, areplotted versus depth in Figures 1 through 10. The measuredresidual pore pressures are given on the plots adjacent tothe appropriate shear strength measurements.
0
100
200
inn
1.
•
t
: *•
«•
•
i
•
•
• •
* . t
• *
SITE 183
_
•
0 . 5 1.0
UNCORRECTED VANE STRENGTH (Kg/cm2);
Figure 1. Uncorrected, measured vane shear strength versussubbottom depth, Site 183.
1.0 2 . 0
UNCORRECTED VANE STRENGTH (Kg/cm2)
Figure 2. Uncorrected, measured vane shear strength versussubbottom depth, Site 184.
704
MEASUREMENTS AND ESTIMATES OF ENGINEERING AND OTHER PHYSICAL PROPERTIES
0.5 1.0
UNCORRECTED VANE STRENGTH (Kg/cm2)
Figure 3. Uncorrected, measured vane shear strength versussubbottom depth, Site 185.
•--(.0054)y\ » H.0068)
\.0068) ^ .0272)
^ - ( . 0 4 4 2 ) .
( . 0 2 1 7 ) ' • ( . 0558 )
^ ( . 0 6 6 9 )
(.141)
NUMBERS BESIDE DATA POINTSREPRESENT THE ABSOLUTE VALUEOF THE MEASURED RESIDUALPORE PRESSURES in Kg/cm2 .
0 .0 0.5 1.0
UNCORRECTED VANE STRENGTH (Kg/cm2)
Figure 4. Uncorrected, measured vane shear strength versussubbottom depth, Site 186.
A number of conclusions may be drawn from thesecurves:
1) Strength usually increases with depth.2) At about the same depth, higher strengths corre-
spond to more negative residual pore pressures.3) Uncorrected strength varies strongly with sediment
type. For example, samples from the highly diatomaceous
*~S-(.00B5)
^(•014)
^(.014)
-
(.012). . .
-
•
1
•(•027)
..
.(.013)
I
1 '
NUMBERS BESIDE DATA POINTSREPRESENT THE ABSOLUTE VALUEOF THE MEASURED RESIDUALPORE PRESSURES IN Kg/cm2.
•(.056)
I0.5 1.0
UNCORRECTED VANE STRENGTH (Kg/cm2)
Figure 5. Uncorrected, measured vane shear strength versussubbottom depth, Site 188.
/(.0O70)^(.O211)
(.0086) / 1*(.0155/ K.0493)
^(.0246)
•
*
• *
C -J 27)
1
SITE 189
•
NUMBERS BESIDE DATA POINTSREPRESENT THE ABSOLUTE VALUEOF THE MEASURED RESIDUALPORE PRESSURES IN Kg/cm2.
*
1 *0.5 1.0
UNCORRECTED VANE STRENGTH (Kg/cm2)
Figure 6. Uncorrected, measured vane shear strength versussubbottom depth, Site 189.
Site 188 are much weaker than those from the more clayeySite 189. Residual pore pressure varies similarly withsediment type.
4) There is a great deal of scatter in the data.The data of Figures 1 through 10 are somewhat valuable
from a qualitative point of view. However, there is noreason to believe that any of the data accurately represent
705
H. J. LEE
(.015)x(
('.063)^
•
-
006)
(;014)
I" J*j
• X
- *•
•
1
/(.051)
(.084K
•
•s •*
• ••
SITE 1901
NUMBERS BESIDE DATA POINTSREPRESENT THE ABSOLUTE VALUEOF THE MEASURED RESIDUALPORE PRESSURES IN K g / c m 2 .
.190)
• * . '(.162)
.(.225)
X . 0 0 9 )
0.5 1.0
UNCORRECTED VANE STRENGTH (Kg/cm2)
Figure 7. Uncorrected, measured vane shear strength versussubbottom depth, Site 190.
^ N 0098)
\ .012)
(.014)
NUMBERS BESIDE DATA POINTSREPRESENT THE ABSOLUTE VALUEOF THE MEASURED RESIDUALPORE PRESSURES IN Kg/cm2.
(.139)
— 0.0 0.5 1.0
UNCORRECTED VANE STRENGTH (Kg/cm2)
Figure 8. Uncorrected, measured vane shear strength versussubbottom depth, Site 191.
in situ behavior. In order to deduce in situ characteristics, itis necessary to quantitatively utilize the residual porepressure measurements. This was done by first calculatingthe in situ vertical effective stress, ~övo, for each sedimentdepth at which a pore pressure test was run. The shorelaboratory densities were used in these calculations
* \ \£v- ( • 0 2 1 ) . .
' \ , * ,\ .016)
(.023). .
-
1
.039)̂
(-183) . . . .
070) * ','\
,
1NUMBERS BESIDE DATA POINTSREPRESENT THE ABSOLUTE VALUEOF THE MEASURED RESIDUALPORE PRESSURES IN K g / c m 2 .
-
N.098)
* . '
10.5 1.0
UNCORRECTED VANE STRENGTH (Kg/cm 2
Figure 9. Uncorrected, measured vane shear strength versussubbottom depth, Site 192.
(.010)* *
.(.008)
•
(.007)'
(.006)
( . 0 1 9 ) .
-
. 1
, . ( . 0 1 2 )
• , #
• , ' ( . 0 4 6 )
( . 0 8 4 ) ' "
1
1
NUMBERS BESIDE DATA POINTSREPRESENT THE ABSOLUTE VALUEOF THE MEASURED RESIDUALPORE PRESSURES IN Kg/αn 2 .
. ( . 0 3 7 )
i
0.25 0.5
UNCORRECTED VANE STRENGTH (Kg/cm2)
Figure 10. Uncorrected, measured vane shear strengthversus subbottom depth, Site 193.
although the GRAPE densities would have yielded com-parable values. These stresses were multiplied by —0.55 toobtain an estimate of the residual pore pressure whichwould result from perfect sampling, ups. The measuredresidual pore pressures, ur, were divided by these values toobtain ratios, ur/ups, which are measures of disturbance.
706
MEASUREMENTS AND ESTIMATES OF ENGINEERING AND OTHER PHYSICAL PROPERTIES
The next step was to determine how these pore pressureratios would relate to the ratios of laboratory to in situstrength. Ladd and Lambe (1963) provide the series ofcurves shown in Figure 11. In their original report the axeswere labeled somewhat differently and the data wereobtained in an entirely different manner. However, Laddand Lambe (1963) suggest that the data be used in basicallythe same fashion as given by Figure 11, i.e., as a plot ofstrength ratio versus pore pressure ratio.
One problem immediately apparent from Figure 11 isthat different soils produce different curves on a plot suchas this. In the earlier NCEL study (Lee, in press) it wasfound that terrigenous sediments from the Santa BarbaraChannel area followed the Wealt clay curve relatively well.However, there is no reason to believe that Bering Sea andNorth Pacific sediments will do likewise.
In order to approach this problem, one of the commonlyaccepted concepts of soil mechanics was relied upon. In anormally consolidated soil (one in which the in situ verticaleffective stress, ~σvo, has never exceeded its present value),the in situ strength, Sf, increases linearly w i t h α ^ . That is,the ratio, Sy/σvo, is a constant.
It was observed that the shapes of the curves in Figure11 were all approximately the same. To simplify matters,four characteristic curves, having the same shape as theactually measured curves and spanning the complete rangeof values, were constructed and are shown in Figure 12.
/KAWASAKI CLAY 1
LAGUNILLAS
B.B.C.
AMUAY CLAY
OSLO CLAY
5 10 20Pore Pressure R a t i o , u /u
Figure 11. Proposed plot of strength ratio versus porepressure ratio-based on previous tests - after Ladd andLambe, 1963.
Pore Pressure Ratio, u /u
Figure 12. Idealized curves developed to representinfluence of pore pressure changes on strength. Curve 3is recommended for Leg 19.
Given these curves and a set of Si,ur, and ups data, it waspossible to calculate four values of in situ strength, Sf.These values of Sf were in turn divided by the appropriate~vo to yield four values of the ratio, Sf/ ~vo, for each set of
measured values of Si and ur. These data were thenanalyzed statistically. The variances of the logarithms ofeach set of Sf/σ o data (below a sediment depth of about40 m) for each site-Figure 12 curve combination werecalculated. The logarithms were used to eliminate purescaling effects. Data above about 40 meters were notconsidered in this particular analysis because there is oftenconsiderably more variation in Sf/σ o in this range. Thevariances were then averaged over the sites to yield fouraverage variances for each of the curves in Figure 12. As aresult, curve 3 was found to yield the lowest variance withcurve 2, second; curve 4, third; and curve 1, fourth. Curve3, therefore, yields values of Sf/ ~vo which most nearlysatisfy the criterion that Sf/σvo should be constant withdepth. It is recommended therefore that curve 3 of Figure12 be used to correct the vane shear strengths measuredduring Leg 19. A further justification for doing this will begiven below in the discussion of the one triaxial testperformed.
The estimated in situ vane shear strengths (based oncurve 3) for all the points at which residual pore pressuretests were performed are plotted versus depth in Figure 13.
ESTIMATED IN SITU VANE STRENGTH, Sp (Kg/cm
2)
Figure 13. Estimated in situ vane strength versus depthLeg 19 (based on Curve 3, Figure 12).
707
H. J. LEE
Several important points are apparent from these curves:1) The estimated in situ strength values are much higher
(by a factor of 2 to 3 or more) than the measuredlaboratory strengths (Figures 1-10).
2) The data scatter among the corrected strength valuesfor each site is much less than among the laboratory values.This would appear to indicate that much of the variation inmeasured values is a result of disturbance rather than actualchanges in in situ character.
3) There is not a great deal of variation in estimated insitu strength among the sites themselves for any givensediment depth. Some of the variation that does exist maybe a result of errors in the correction technique. It wouldbe possible to construct an average curve through the datafor use in design which would be correct within about plusor minus 50 percent.
4) There are three basic types of shear strength-depthplots illustrated. At Sites 186, 188, and 191, the strengthincreases consistently with depth from a value of near zeroat the surface. At 189 and 192, the strength increases veryrapidly over the first 10 to 20 meters, passes through analmost uniform zone to 50 to 60 meters, and then begins toincrease at about the same rate as the strength at Sites 186,188, and 191. At Site 190 the strength increases rapidly forthe first 20 meters. There is then a change in slope with thestrength below that point increasing at about the same rateas at the greater depths of the other sites. These variationscould be a result of lithologic changes or errors. However,the possibility that there could be two components ofstrength also exists. One component would be strictlyfrictional (increasing linearly with normal stress, ~övo) andthe other component would be nonfrictional or cohesive(independent of σ"vo). Sites 186, 188, and 191 would belacking any significant cohesion component. The cohesivecomponent at Sites 189 and 192 would build up (as a resultof aging, chemical alteration, etc.) over the upper 10 to 20meters and then be destroyed over the next 40 meters (as aresult of bond breaking caused by increased overburden).Below 60 meters strength would be primarily frictional. AtSite 190 the cohesive strength would build up in the upper20 meters and not be destroyed below that point (strongerbonds).
It is, of course, difficult to determine how valid thesehypotheses are at present. Additional research should showwhether they are valid and if so, what environmental andsediment characteristics lead to the different types ofbonding.
The data are presented in a somewhat different mannerin Figure 14. Here the ratio of estimated in situ strength,Sf, to vertical effective stress, σ"vo, is plotted versussediment depth. As discussed previously, this ratio wasartifically forced to have as little variation as possible belowa sediment depth of 40 meters. No constraints were placedon the data above this depth, however. As may be seen, thisshallow material always produced a higher Sf/σ~vo ratio thanthe deeper material. The point of divergence is at asediment depth of about 30 meters, above which, very highvalues (greater than 1.0) of Sfjö~vo may occur. Below thispoint, the ratio values are similar to those which have beendetermined for normally consolidated soils on land (0.2 to0.4).
Triaxial Shear Strength
One consolidated-undrained triaxial test was performedon a sample cut from the zero section of Core 5 from Site192. The in situ vertical effective stress, ~övo, would havebeen about 1.35 kg/cm2. However, this sample wasreconsolidated isotropically (horizontal and vertical stressesequal) to an effective stress level of 5.64 kg/cm2. Con-solidating to stresses much higher than the in situ level isone recommended (Ladd and Lambe, 1963) technique forreducing the effects of disturbance on friction angle andSf/(TVO ratio. Following consolidation, the sample was failedby increasing the axial load. This was done withoutdrainage (no changes in volume allowed) and the pore waterpressures which developed were measured. Using theeffective stress relation (Equation 1) it was possible tocalculate the effective vertical and horizontal stresses, σ~v
and σ^.
One way of expressing these results is in the form of astress path for the undrained loading portion of the test. Atypical stress path is a plot of the quantity,
σv+σf
versus°h
for the loading.
In mechanics terms, this is a plot of the top points of theMohr circles representing all of the intermediate stressstates. This stress path is shown in Figure 15. A good dealof important information can be drawn from a plot of thissort: however, for the purposes of this report, only a fewpoints need to be noted. First the point of failure should bedetermined. By the common failure criteria (Max ~σv — on
or σ^/σ^), this is determined to be:
°V °k = 2.11 kg/cm2
°v+σh= 3.63 kg/cm2
A line drawn through this point and the origin representsa failure envelope assuming no cohesion intercept. This linemakes an angle, a., with the horizontal axis which is relatedto the usual angle of internal friction Φ by the equation:
Sin0 = Tanα
Tanα = |4T3.63
= 0.58
(Lambe & Whitman, 1969)
Therefore, 0 = 35.5°.
This is a typical value for a slightly plastic, cohesive soil(Bjerium and Simons, 1960). The sediment was a dia-tomaceous silty clay.
Another parameter which can be determined from thesedata is an estimate of the strength to overburden pressureratio, Sf/Wvo. The initial vertical consolidation stress forthis test was 5.64 kg/cm2 and the measured shear strength(assumed equal to ~ÖV — σ^/2) was 2.11 kg/cm2. Thestrength to pressure ratio was then 2.11/5.64, or 0.37.
708
MEASUREMENTS AND ESTIMATES OF ENGINEERING AND OTHER PHYSICAL PROPERTIES
100
SUB-BOTTOM DEPTH (m)
Figure 14. Variation of estimated in situ strength - vertical effective stress ratio (S•pjGV0) with subbottomdepth.
4 -
3 -
(Kg/cm2)
1 -
1 1
Estimated
Tan <x.
f
I i
y
Failure Envelope
= 0.58
= 35.5°
i i
I 1
y
\
1 I
2 -
V G h (Kg/cm2)2
Figure 15. Stress path for consolidated-undrained triaxial test Site 192 - Core 5 - Zero station.
709
H. J. LEE
Consolidating to 5.64 kg/cm^ approximately simulated asediment which has that vertical effective stress in situ. AtSite 192 this stress would correspond to a depth of 130meters. The vane shear strength and residual pore pressurewere measured on a sample from 156 meters. Using thefour curves of Figure 12 to predict in situ strengths, thefollowing values of Sf/σ~vo were obtained.
Curve1234
0.170.220.340.69
Once again curve 3 appears to yield the best results witha strength to pressure ratio of 0.34, versus a measured 0.37.
Compressibility
BACKGROUND
The compressibility of sediments has traditionally beenapproached in two ways:
1) A sample is taken and compressed rather rapidly(usually 24 hours per load increment). The compressibilityfor each loading is obtained and generally used to predicthow much a structure will settle. Relatively good agreementbetween predicted and observed results has been obtained.
2) The in situ porosity or void ratio is measured as afunction of depth. The change in void ratio with increasingoverburden stress is taken as the field compressibility. Thisvalue relates to how the sediment compacts as additionalmaterial is deposited on the surface.
The difference between these two approaches is clearlythe time factor, but it is difficult to determine exactly whatinfluence it may have. Schmertmann (1955) presents dataon a number of river bottom deposits which show that thefield and laboratory consolidation data are virtually identi-cal if the laboratory data are corrected for disturbance andthe sediment is normally consolidated. Leonards andAltschaeffl (1964), however, present data on laboratoryconsolidation tests run at radically different rates whichshow that sediment compressibility at low loading rates(still high in a geologic sense) is much lower than at highrates. The authors attribute this effect to the formation ofstrong interparticle bonds (resulting from the reorientationof water molecules in the vicinity of contact points) duringlong-term loading. Hamilton (1964), in analyzing the resultsof tests on samples from the experimental Mohole (Guada-lupe Site), developed a conclusion similar to that ofLeonards and Altschaeffl. However, he attributed the lowrate of field consolidation to the development of chemicalcementation in the vicinity of interparticle bonds. Unpub-lished Naval Civil Engineering Laboratory data indicate aninverse trend—greater field than laboratory compressibility.This can be rationalized in terms of a great deal oflong-term (secondary) compression occurring in the fieldbut not occurring in a short-term laboratory test.
The situation is a complex one and it is difficult todetermine which factors dominate under which conditions.Sample disturbance and sediment type variation with depthcompound the problem.
RESULTS
During Leg 19, measurements of both field and labora-tory compressibility were attempted. Sample densities werecalculated using the shore laboratory water contents, graindensities, and the tables provided by Bennett et al. (1971).These were assumed equal to the in situ densities. It isrecognized that slight changes in density occur duringsampling. However, aside from the slight (1 to 2%) amountof change resulting from direct expansion of pore water, itis almost impossible to quantitatively evaluate thesechanges. In any case, density is perhaps the most reliablephysical property which can be obtained from disturbedsamples. These densities were converted to void ratios, e,which were plotted versus the logarithms of calculated insitu vertical effective stress, ~övo• These field compressibilitydata are shown in Figures 16 through 25.
\ -LAMBE AND WHITMAN(1969) PLOT FORHIGHLY COLLOIDALCLAY
ESTIMATED "UNDISTURBED"CONSOLIDATION BEHAVIOR
•SAMPLE VOID RATIOAS A FUNCTION OFIN SITU VERTICAL EFFECTIVESTRESS
1.0 10
VERTICAL EFFECTIVE STRESS (Kg/cm2) (LOG SCALE)
Figure 16. Sediment compression data, Site 183.
\ LAMBE AND WHITMAN (1969)PL0T F0R HIGHLY COLLOIDALCLAYS
1.0 10VERTICAL EFFECTIVE STRESS (Kg/cm ) (LOG SCALE)
Figure 17. Sediment compression data, Site 184.
710
MEASUREMENTS AND ESTIMATES OF ENGINEERING AND OTHER PHYSICAL PROPERTIES
£ 2 . 0
-S ILTY-CLAY
\LAMBE AND WHITMAN (1969)PLOT FOR HIGHLY COLLOIDALCLAYS
VOID RATIO - IN SITU VERTICALEFFECTIVE STRESS (185-6-0)
CONSOLIDATION TEST(185-6-0)
•-SAMPLE VOID RATIO AS AFUNCTION OF IN SITU VERTICALEFFECTIVE STRESS
1.0 10
LOG VERTICAL EFFECTIVE STRESS (Kg/cm )
Figure 18. Sediment compression data, Site 185.
• N > T ^ LAMBE AND WHITMAN (1969)\ PLOT FOR HIGHLY COLLOIDAL
\\
\
\
1.0 10
LOG VERTICAL EFFECTIVE STRESS (Kg/cm2)
Figure 21. Sediment compression data, Site 189.
2 2 • ° "
ç LAMBE AND WHITMAN (1969)/r PLOT FOR HIGHLY COLLOIDAL\ CLAYS
\ *\
Ss \
-SAMPLE VOID RATIO AS AFUNCTION OF IN SITU VERTICALEFFECTIVE STRESS
1.0 10
LOG VERTICAL EFFECTIVE STRESS (Kg/cm
• SITE - 186
• SITE - 187
Figure 19. Sediment compression data, Sites 186 and 187.
VOID RATIO - IN SITU VERTICAL# EFFECTIVE STRESS (188-7 -0 )
VOID RATIO - IN SITU VERTICALEFFECTIVE STRESS (188-8-0 )
> - SAMPLE VOID RATIO AS AFUNCTION OF IN SITU VERTICAL LAMBE AND WHITEMAN (1969) -EFFECTIVE STRESS PLOT FOR HIGHLY COLLOIDAL
CLAY
1.0 10
LOG VERTICAL EFFECTIVE STRESS ( K g / α /
Figure 22. Sediment compression data, Site 190.
VOID RATIO - IN SITU VERTICAL. EFFECTIVE STRESS ( 1 8 8 - 7 - 0 )
VOID RATIO - IN SITU VERTICALEFFECTIVE STRESS ( 1 8 8 - 8 - 0 )
ESTIMATED UNDISTURBEDCONSOLIDATION BEHAVIOR
•– SAMPLE VOID RATIO AS A
FUNCTION OF IN SITU VERTICAL LAMBE AND WHITEMAN ( 1 9 6 9 ) - _EFFECTIVE STRESS PLOT FOR HIGHLY COLLOIDAL
CLAY
1.0 10
LOG VERTICAL EFFECTIVE STRESS (Kg/cm2)
"xAV- LAMBE AND WHITMAN (1969)
#X PLOT FOR HIGHLY COLLOIDAL
\\
\
N\\
\\
1.0 10
LOG VERTICAL EFFECTIVE STRESS (Kg/cm2)
Too
Figure 20. Sediment compression data, Site 188. Figure 23. Sediment compression data, Site 191.
711
H. J. LEE
T(CLAYSTONE)
SITE - 192
\(^- LAMBE AND WHITMAN (1969)X PLOT FOR HIGHLY COLLOIDAL CLAYS
1.0
LOG VERTICAL EFFECTIVE STRESS (Kg/cm")
Figure 24. Sediment compression data, Site 192.
SAMPLE VOID RATIO AS AFUNCTION OF IN SITU VERTICALEFFECTIVE STRESS
LOG VERTICAL EFFECTIVE STRESS (Kg/cm )
Figure 25. Sediment compression data, Site 193.
Also plotted on these figures are the results of thelaboratory consolidation tests. Each test is presented as aloading curve followed by an unloading curve. For each testthe point which represents the estimated in situ sample voidratio and vertical effective stress is denoted by a specialsymbol. Under ideal conditions the loading portions of theconsolidation curves would pass through these points. Asmay be seen, the curves pass well below the pointsindicating a great deal of disturbance.
A curve taken from Lam be and Whitman (1969) isdrawn on each plot. This curve was developed on the basisof a considerable amount of data and is typical of thecompression behavior of a highly colloidal clay.
ANALYSIS
Maximum Past Pressures
Casagrande (1936) suggested a graphical technique forestimating the maximum past pressure (largest vertical
effective stress to which a sediment has ever been sub-jected), σvm, on the basis of standard consolidation testresults. This technique was applied to most of the consoli-dation test curves to yield the following results:
Site Core
In-situ VerticalEffective Stress,
Section "σvo (kg/cm2)
Estimated MaximumPast Pressure,ö v m (kg/cm2)
183183184185188188190190
311467879
CCCC0000CCCC
1.15.68.33.85.16.94.16.0
1.24.6
?3.0
9
5.62.56.0
The graphical constructions are shown on Figures 16through 25. The Casagrande construction could not beapplied to two of the curves representing highly diatom-aceous material for reasons which will be given later.
As may be seen, the estimated maximum past pressuresdo not vary greatly from the current in situ verticaleffective stresses. The few differences which do exist can berationalized in terms of disturbance. It is difficult to applythe Casagrande construction accurately on a curve based ona disturbed sample. In general, therefore, there is noindication that any of the samples tested are anything otherthan normally consolidated, i.e., that the actual maximumpast pressure is different from the current in situ verticaleffective stress.
Comparison of Field and Typical Highly Colloidal Clay Data
The sediments sampled during Leg 19 consisted almostentirely of silty clay, diatom ooze, or a mixture of silty clayand diatom ooze. The subbottom depths over which eitherdiatom ooze or silty clay predominated are delineated inFigures 16 through 25. As may be seen, there is a clearcorrelation between field compression data and sedimenttype. The silty clay data compares favorably with theLambe and Whitman (1969) curve for a highly colloidalclay. The diatom ooze data deviates significantly (muchhigher void ratios) from this curve.
Estimation of Undisturbed Consolidation Behavior
Schmertmann (1955) observed that virtually all samplesof the same sediment, ranging from completely remolded toalmost undisturbed, consolidate to the same void ratiounder the same stress if the stress is high enough tocompress the sample to a void ratio less than 42 percent ofits initial void ratio. That is to say, all consolidation testcurves converge at around QA2eo, where e0 is the initialvoid ratio.
It is generally observed that the void ratio variesapproximately linearly with the logarithm of effective stressonce the maximum past pressure has been exceeded. TheLeg 19 samples appeared to be normally consolidated sothese linear variations should begin at the points withcoordinates of in situ vertical effective stress, ~övo, and insitu (assumed equal to sample) void ratio, eo. UsingSchmertmann's observation, these lines should intersect thelaboratory consolidation curves at 0.42eo and represent an
712
MEASUREMENTS AND ESTIMATES OF ENGINEERING AND OTHER PHYSICAL PROPERTIES
estimate of how the sediment would consolidate if it werecompletely undisturbed. These estimates of undisturbedconsolidation behavior are shown in Figures 16 through 25.
Unfortunately, all but one of the consolidation testswere performed on samples of diatom ooze. The oneexception (Site 183, Core 3, core catcher sleeve sample)was performed on a silty clay. As may be seen (Figure 16),the estimated undisturbed consolidation curve for thissediment agrees relatively well with the Lambe andWhitman (1969) curve for highly colloidal clay. Since thislatter curve was found to agree with the field consolidationdata for silty clay, there is some reason to believe that thefield data may be estimated using short-term consolidationtest results. This might imply that neither the conclusionsof Leonards and Altschaeffl (1964) nor Hamilton (1964)regarding cementation and time effects are applicable to theNorth Pacific and Bering Sea silty clays.
The estimated undisturbed consolidation curves for thediatom oozes are much more difficult to evaluate becausethe field data on these sediments are extremely scattered. Adiscussion of the implications of these follows.
Consolidation Behavior of Diatom Oozes
The diatom oozes which were sampled displayedextremely high void ratios under substantial overburdenpressures. Consider, for example, Site 188 where void ratiosas high as 5.0 were found at subbottom depths in excess of150 meters. There are at least two ways in which a materialsuch as this could have been formed: (a) It could havebeen deposited at a much higher void ratio and thencompressed at the rate suggested by the estimated undis-turbed consolidation curves of Figures 16 through 25; or(b) it could have been deposited and maintained at aboutthe same void ratio as it has now. This would imply a nearlyincompressible material.
To determine which of these processes is most probable,a sample from Core 8, Site 188, was suspended in seawaterand allowed to resediment over a period of 24 hours. Avoid ratio of 4.2 was obtained. This sample originally had avoid ratio of 3.1 and was taken from a subbottom depthwhich would produce a vertical effective stress of 6.9kg/cm2. Compressing from a void ratio of 4.2 to one of 3.1implies a volume change of only 20 percent, a very slightamount for a stress increment of this magnitude. It maythen be tentatively concluded that diatom oozes have highvoid rates at large embedment depths because they arealmost incompressible and not because they have com-pressed to that point from much higher void ratios.
There is still the problem that the consolidation testresults, especially the estimated undisturbed curves, indi-cate a relatively great amount of compressibility. Onepossible explanation for this may be seen by observing theshape of the consolidation curves for several of the tests.For the test on the Site 188, Core 7 sample and the one teston a Site 184 sample, the consolidation curves never reach alinear portion as would generally be expected. Instead, theslope of the curve increases steadily with the logarithm ofeffective stress. This could easily be a result of graincrushing which begins to occur at around 5-7 kg/cm2 andthen continues to occur with increasing frequency at higherstress levels. A hypothesis of this nature would allow for
the ooze to appear to be very compressible at high stresseswhile behaving almost incompressibly at low to moderatestress levels. The hypothetical general model of diatomooze consolidation which results is as follows. The indi-vidual diatoms originally deposit themselves relativelyclosely although the resulting void ratio is high by virtue ofthe high porosity of each individual grain. As stressincreases with increasing overburden, the grains roll andpack slightly more closely. However, since they wererelatively closely packed originally, the net compression issmall. Finally, the crushing strength of the grains isexceeded at high stress levels and the material begins tobehave very compressibly.
This problem of continuing curvature of the consolida-tion curves at high stress levels made both Casagrande(1936) and Schmertmann (1955) type constructionsimpossible.
Comparison of Density and Water ContentMeasurement Techniques
A variety of different techniques were used to measuredensities and water contents. Of these the most accurate(and also most time-consuming) is that involving takingsamples on shipboard, carefully sealing, and measuringwater content and grain density onshore. These measure-ments can then be used as a basis for checking the accuracyof the shipboard measurements (GRAPE and syringe). Allof the possible comparisons of the shore laboratory densityand GRAPE density and shore laboratory water contentwith shipboard syringe water content are given in Table I2 .Table 2 presents average values of each quantity measuredby each method at each site. Also given are the averageabsolute differences between shipboard and shore measure-ments for each site and for the leg as a whole.
As may be seen, the GRAPE densities deviate from theshore densities by about 5 percent on the average while theshipboard syringe water contents differ from the shorelaboratory water contents by about 10 percent on theaverage. Much of the GRAPE density variation was accu-mulated at Sites 186 and 191, however, both locations atwhich the samples contained a good deal of gas. For thissituation there is a basic difference in the GRAPE methodof measurement and the shore laboratory technique: theGRAPE measures the density of the sample which the shorelaboratory technique estimates what the density would beif the sample were completely saturated. Obviously, if thegas content is high, the two values differ. Ignoring Sites 186and 191, the difference between GRAPE and shorelaboratory measurements is an easily tolerable 2.7 percent.
Water content is not affected by gas content; therefore,there is no way to reduce the average deviation betweensyringe and shore laboratory measurements. The 10 percenterror introduced by the syringe is probably intolerable foranything other than a rough approximation.
2Water contents are defined by the engineering convention asweight of water divided by weight of solids.
713
H. J. LEE
TABLE 1Comparison of Density and Water Content Measurements—Individual Values
Hole
183
184
184B
185
185
186
Core
456677
8899
1111
121213171819
212223262728
3138
114679
111216171819
202122
55
45678
10
1719
1923
234678
89
Section
242424
242424
244644
442243
12
231466
422222
232
33
621224
22
22
131221
36
Interval
129-13217-20
133-13610-13
141-144135-138
112-11590-92
111-11495-97
133-136114-11770-72
112-115103-10535-38110
83-87
39-4296-9810175
95-9889-92
130-1336-8
65-6729-3262-65
131-13488-97
104-107
112-11540
89-92130-133100-10368-70
134-137118-12192-95
50100
127-129128-131131-13483-85
113-116131-133139-14141-44
5075
66-6939-4231-33
128-13164-6676-79
100-102138-141
Shore Lab.Density
(g/cc)
1.671.471.501.481.411.46
1.661.501.571.651.581.451.561.391.311.301.311.311.341.851.801.841.781.931.861.96
1.45IAA1.371.491.411.52
1.591.531.381.371.381.44
1.381.491.51
1.991.97
1.491.411.531.421.491.481.631.68
1.691.65
1.501.521.551.631.601.611.741.65
GRAPEDensity(g/cc)
1.621.551.521.541.401.551.641.611.621.401.541.571.621.521.431.301.311.34
1.351.851.711.871.821.881.88
-
1.441.521.381.521.381.56
1.601.451.401.441.381.46
1.401.491.37
_
1.641.461.44
_1.481.501.521.341.471.431.511.50
_1.521.421.591.561.621.40
Shore Lab.Water Content
(% dry wt.)
149.83
120.9886.59
87.47
78.23
79.5577.35
SyringeWater Content
(% dry wt.)
114.5
121.081.1
84.2
85.9
78.876.1
714
MEASUREMENTS AND ESTIMATES OF ENGINEERING AND OTHER PHYSICAL PROPERTIES
TABLE 1 - Continued
Hole
187
188
189
190
Core
10111718
191920212123
242425252627
27272727272828
224
123333466777
8899
334455
67
10
112233
344455
Section
2422
244233
111121
2234644
122
111234323126
1412
143434
231
121223
423445
Interval
66-6936-39
119-12125
81-8499-10159-62
110-11230
102-10520
40-4250
50-53117-119
30
40143-144
3010075
109-112130
89-9187-90-
128-130128-13038-4035-3838-41
114-11641-4361-6392-9535-3731-3418-2086-8915-1870-7354-5610-12
36-3814-1636-3947^9
117-11984-87
2-6140-143123-126
137-14068-71
140-14215-17
139-142124-12643-4527-29
122-12560-63
116-119100-103
Shore Lab.Density
(g/cc)
1.761.751.711.73
1.791.841.761.831.851.75
1.801.881.761.811.841.86
1.881.971.891.831.791.901.89
1.761.631.90
1.321.371.431.451.451.391.301.361.441.231.261.28
1.361.321.261.20
1.581.561.571.561.651.55
1.581.581.94
1.471.531.531.571.511.571.551.551.471.471.621.46
GRAPEDensity(g/cc)
1.641.671.281.38
_
1.751.541.611.541.41
1.301.191.491.491.671.53
1.411.441.191.521.491.581.60
1.701.431.72_
1.37_
1.601.40_
1.461.471.261.291.34
1.401.321.311.29
1.551.551.531.581.311.41
1.591.491.87_
1.561.471.531.501.571.531.501.461.381.511.45
Shore Lab.Water Content
(% dry wt.)
53.90
150.59126.15
94.82
160.79
92.63205.28
118.82
95.5882.3783.472.586.8
77.0
99.17
SyringeWater Content
(% dry wt.)
51.3
144.6149.4
91.2
173.0
100.0200.0
97.3
96.082.374.369.587.3
77.2
87.7
715
H. J. LEE
TABLE 1 - Continued
Hole
191
191A
191B
192
Core
566777
788889
91010111111
1112131414
244445
555567
89
122244
111
122333
344444
455555
Section
645123
423452
323234
54226
213461
234522
21
412356
236
124123
412345
612345
Interval
43-4579-8130-3326-28
137-13927-30
97-9911-1349-5234-3690-9274-76
17-19110-11381-8479-82
109-11257-59
81-8380-8249-5173-7569-71
22-2333-3579-8191-93
140-14297-9975-77
101-10494-96
116-119103-105141-142
98-10190
137-138114-11773-76
113-11532-3557-60
131-13437-3998-101
68-70137-140104-10682-8565-6753-55
115-11852-5546-4885-8851-5423-2620-2289-9132-3416-1899-102
7-10
Shore Lab.Density
(g/cc)
1.521.531.531.511.641.54
1.541.571.511.611.591.62
1.631.411.421.521.511.46
1.371.491.341.361.36
1.351.491.441.461.531.711.741.691.741.761.671.80
1.711.78
1.571.571.431.501.531.65
1.441.411.49
1.411.461.501.481.561.57
1.491.481.571.481.551.561.541.451.551.471.511.51
GRAPEDensity(g/cc)
1.511.491.521.511.621.60
1.591.571.501.641.571.55
1.611.391.451.531.541.46
1.371.561.631.421.41
__
1.331.321.331.471.66_
1.651.681.661.601.62
1.651.321.52_
1.381.601.431.53
1.421.431.53
_
1.441.431.411.561.60_
1.541.611.461.581.541.501.511.551.411.531.50
Shore Lab.Water Content
(% dry wt.)
82.17
79.11
66.38
137.25
48.11
71.75
80.09
115.57
91.12
92.90
77.14
83.95
SyringeWater Content
(% dry wt.)
88.5
71.6
78.4
132.6
47.9
91.7
58.1
119.2
84.3
88.0
81.0
89.9
716
MEASUREMENTS AND ESTIMATES OF ENGINEERING AND OTHER PHYSICAL PROPERTIES
TABLE 1 - Continued
Hole
192A
193
Core
677778
89999
10
121212121313
131515171921
2125
555555
111223
333333
33
Section
212345
634563
123423
544322
12
112345
112131
223344
55
Interval
143-145134-13788-9187-89
104-10635-38
88-9085-8749-52
9-11120-123
13-16
77-7941-4357-5944-4666-6844-46
119-12129-3193-95
139-14160-63
131-13312570
401304013514035
114-116136-13896-98
107-108107-109131-134
95-97105-10848-5196-9870-73
124-126
71-74125-128
Shore Lab.Density
(g/cc)
1.391.491.571.551.491.51
1.521.481.471.551.53
1.391.391.391.371.301.37
1.361.411.411.411.411.39
1.361.44
2.162.822.612.722.632.61
1.431.401.491.391.491.49
1.451.501.381.491.501.44
1.431.54
GRAPEDensity(g/cc)
1.511.591.551.501.49
1.461.451.471.541.55
1.411.431.411.381.261.38
1.361.401.411.451.441.39
1.281.39
1.802.312.332.132.112.31
1.421.401.421.391.511.42
1.501.561.421.571.521.49
1.441.51
Shore Lab.Water Content
(% dry wt.)
109.04
71.08
81.63
74.90
104.67
109.03
106.55
105.53
115.4688.85
119.2887.49
86.36
85.33
105.68
SyringeWater Content
(% dry wt.)
112.1
75.0
77.1
71.4
104.9
105.7
101.2
96.9
97.493.5
107.091.2
83.4
90.2
86.2
CONCLUSIONS
1. All Deep Sea Drilling Project cores are disturbed froman engineering point of view. Engineering properties (withthe possible exception of density) obtained from samples ofthese cores are not representative of the in situ condition.Correction procedures must be used to estimate in situbehavior.
2. Previous research has shown that shear strength maybe corrected for disturbance using the residual pressuremaintained by the sample pore water. These residual porepressures were measured during Leg 19 and an approximatetechnique for using them to estimate in situ strength wasdeveloped. The estimated in situ strengths increase with
overburden at rates which are typical of normally consoli-dated soils (around 0.3). There is only moderate variation(plus or minus 50%) among the sites and among thesediment types.
The estimated strength of the material more shallowthan 30 meters was always significantly greater relative tooverburden than that of the more deeply embeddedmaterial. Possibly a form of interparticle bonding which isindependent of overburden pressure dominates over thisregion.
3. One triaxial test performed on a diatomaceous siltyclay sample yielded a typical drained friction angle of 35degrees. The strength to overburden pressure ratio obtainedfrom the triaxial test was 0.37. Using the proposed residual
717
H. J. LEE
TABLE 2Comparison of Density and Water Content Measurements-Average Values
Site
183184185186187188189190191192193
AverageShore Lab.
Density(g/cc)
1.5751.4841.5491.7691.7631.3321.6191.5191.5141.4691.459
AverageGRAPEDensity(g/cc)
1.5781.4641.4611.4941.6171.3761.5421.5151.4351.4671.469
Average AbsoluteDifference,
GRAPE and ShoreDensities
(g/cc)
0.0590.0560.1190.2740.1460.0540.0830.0390.1030.0270.036
Average AbsoluteDifference,% Average
Shore Density
3.73.87.7
15.58.34.15.12.66.81.82.5
AverageShore Lab.
Water Content(%)
149.83103.5982.8570.27
158.18
82.4566.6594.0998.35
AverageSyringe
Water Content(%)
114.5101.0585.0568.7
159.25
81.2865.992.8292.7
Average AbsoluteDifference,
Shore and SyringeWater Contents
(%)
35.332.7555.481.53
13.21
5.1314.574.44
19.04
Average AbsoluteDifference
% Average ShoreWater Content
23.62.76.62.2
8.3
6.221.95.1
19.4
Average for Leg 19 4.57Average for Leg 19 (excluding Sites 186 and 191). 2.74
9.10
pore pressure method of estimating in situ shear strength, avalue of 0.34 was obtained. This close compliance lendsadditional credibility to the proposed procedure.
4. In general, the observed shear strength behavior wasnot unusual for a normally consolidated material. Onepossible engineering problem which was observed, however,was the ease with which the diatom oozes could beliquified.
5. Consolidation (equivalent to compaction in geologicterms) was investigated through laboratory testing ofsamples and consideration of void ratio variation with insitu overburden pressure (field curve). For the silty clays itwas found that the field data closely approximated pre-viously obtained data on highly colloidal clays. The onelaboratory test on silty clay produced a consolidation curvewhich, when corrected for disturbance, also agreed wellwith the field data and the previously obtained data. Nocementation or time effects were evident.
6. The diatom ooze data differed from any previouslyobtained data. Extremely high void ratios existed at greatembedment depths. The laboratory consolidation test dataindicated that the materials was compressible at very highstresses. However, an artificially sedimented sample studyshowed that the material was almost incompressible at lowstresses. Grain crushing at the higher stresses was hypo-thesized as a possible explanation for this behavior.
7. Using the standard Casagrande (1936) construction,it was determined that all of the samples which weresubjected to consolidation tests were normally consoli-dated, i.e., had never been subjected to stresses greater thanthe present in situ stresses.
8. The Gamma Ray Attenuation Porosity Evaluation(GRAPE) provides a good estimate (plus or minus 2.5%) ofthe sample density. If the sample contains gases, careshould be taken in extrapolating results back to the in situcondition. The shipboard syringe produces results whichmay be in error by an average of 10 percent.
REFERENCES
Bennett, R. H., Lambert, D. N. and Grim, P. J., 1971.Tables for determining unit weight of deep sea sediments
from water content and average grain density measure-ments. NOAA Tech. Mem. ERL ADML-13.
Bishop, A. W. and Henkel, D. J., 1962. The Measurement ofSoil Properties in the Triaxial Test. London (EdwardArnold Ltd.). 2nd Ed.
Bjerrum, L. and Simons, N., 1960. Comparison of shearstrength characteristics of normally consolidated clays.ASCE Con. Shear Strength of Cohesive Soil. Boulder,Colorado. 711.
Brand, E. W., 1967. The vane shear test and its use forstrength measurement of cohesive soils. RILEM Bull. 36,191.
Casagrande, A., 1936. The determination of the pre-consolidation load and its practical significance. Proc.First Intern. Conf. Soil Mech. Found. Engineering.Cambridge, Massachusetts. 60.
Edgar, N. T., Saunders, J. B. et al., in preparation. InitialReports of the Deep Sea Drilling Project, Volume XV.Washington (U. S. Government Printing Office).
Evans, H. B., 1965. GRAPE — A device for continuousdetermination of material density and porosity. Trans.6th Ann. SPWLA Logging Sympo. Dallas, Texas. II, Bl.
Gibbs, H. J. and Coffey, C. T., 1969. Application of porePressure Measurements to Shear Strength of CohesiveSoils. U. S. Bur. Reclamation Rept. No. EM-761.Denver, Colorado.
Hamilton, E. L., 1964. Consolidation characteristics andrelated properties of sediments from ExperimentalMohole (Guadalupe Site), J. Geophys. Res. 69 (20),4257.
Harms, J. C. and Choquette, P. W., 1965. Geologicevaluation of a gamma-ray porosity device. Trans. 6thAnn. SPWLA Logging Sympo. Dallas, Texas. VII, Cl.
Hironaka, M. C, 1966. Engineering Properties of MarineSediments Near San Miguel Island, California. U. S.Naval Civil Engineering Lab. Tech. Rep. R-503.
Hvorslev, M. J., 1949. Subsurface exploration and samplingof soils for civil engineering purposes. Rep. Comm.Sampling and Testing, Soil Mech. and Found. Div., Am.Soc. Civil Engrs.
Ladd, C. C. and Lambe, T. W., 1963. The strength of"undisturbed" clay determined from undrained tests.ASTM Standard Tech. Pub. 361, 342.
Lambe, T. W. and Whitman, R. V., 1969. Soil Mechanics.New York (John Wüey and Sons), 127.
718
MEASUREMENTS AND ESTIMATES OF ENGINEERING AND OTHER PHYSICAL PROPERTIES
Lee, H. J., in press. In-situ strength of sea floor soildetermined from tests on partially disturbed cores. U. S.Naval Civil Engineering Lab. Tech. Rept.
Leonards, G. A. and Altschaeffl, A. G., 1964. Compressi-bility of clay. J. Soil Mech. and Founda. Div., Am. Soc.Civil Engrs.
Migliore, H. J. and Lee, H. J., 1971. Seafloor PenetrationTests: Presentation and Analysis of Results. U. S. NavalCivil Engineering Lab. Tech. Note N-l 178.
Peterson, M. N. A., Edgar, N. T. et al., 1970. Initial Reportsof the Deep Sea Drilling Project, Volume II. Washington(U.S. Government Printing Office).
Schmertmann, J. M., 1955. The undisturbed consolidationof clay. Trans. Am. Soc. Civil Engrs. 120, 1201.
Taylor, D. W., 1948. Fundamentals of Soil Mechanics. NewYork (John Wiley and Sons).
Terzaghi, K., 1943. Theoretical Soil Mechanics. New York(John Wiley and Sons).
719