Engineering Geology 188 (2015) 77–87

Contents lists available at ScienceDirect

Engineering Geology

j ourna l homepage: www.e lsev ie r .com/ locate /enggeo

Evaluation of undrained shear strength of Swedish fine-grainedsulphide soils

Bo Westerberg a,⁎, Rasmus Müller b, Stefan Larsson c

a Swedish Geotechnical Institute, Linköping, Swedenb Tyréns AB, Borlänge, Sweden/KTH Royal Institute of Technology, Stockholm, Swedenc KTH Royal Institute of Technology, Stockholm, Sweden

⁎ Corresponding author.E-mail address: bo.westerberg@swedgeo.se (B. Weste

http://dx.doi.org/10.1016/j.enggeo.2015.01.0070013-7952/© 2015 The Authors. Published by Elsevier B.V

a b s t r a c t

a r t i c l e i n f o

Article history:Received 8 September 2014Received in revised form 11 November 2014Accepted 18 January 2015Available online 24 January 2015

Keywords:Undrained shear strengthEmpirical factorsOrganic soilFieldLaboratoryTest

In Swedish practice, there is a long tradition of evaluating undrained shear strength from fall-cone tests and fieldvane tests. During the last 20 years cone penetration tests have also become widely used. However, the resultsfrom all these test methods have to be evaluated using empirical factors. The factors generally used for Swedishclays are related to liquid limit and overconsolidation, but they are not applicable to all types of fine-grained soilsand can often be improved by local calibration for the particular type of soil in the area of current interest. Forthis calibration, the results of direct simple shear tests and/or triaxial tests in the laboratory are normally used.This paper presents an evaluation for Swedish fine-grained sulphide soils, for which a general correction factor of0.65 for field vane tests and fall-cone tests, a cone factor Nkt of 20.2 for cone penetration tests and a relationcu,DSS/(σ′cOCR−0.2) of 0.28 have been found. No correlations were found between these empirical factors and theclay content, liquid limit or organic content, but a relationship was found between the overconsolidation ratioand both the cone penetration test and the field vane test. The sulphide soils in question are found in northernSweden along the coast of theGulf of Bothnia. They aremostly classified as organic silt or organic clay,which is nor-mally silty.

© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

A major task in geotechnical engineering is to establish geologicaland geotechnical models on the basis of empirical knowledge and alimited but sufficient number of tests. The geotechnical model forms abasis for idealized geomechanical models. For relatively young, saturat-ed glacial and post-glacial fine-grained sediments, the main parameterin the ultimate limit state design or stability assessment of existingstructures or slopes is the undrained shear strength (cu). However, cuis not a fundamental soil property but depends on factors such as thestress state, strength anisotropy, and strain rate,making it difficult to as-sess for failure domains where these factors vary. Advanced laboratorytests can be performed, but the number of tests may be extensive andcostly. For most geotechnical applications, it is considered sufficient toevaluate strength parameters from in-situ soundings and relatively sim-ple laboratory tests. Theoretical methods for the evaluation of cu fromcone penetration tests (CPTs), field vane tests (FVTs) and fall-conetests (FCs) on undisturbed samples have been developed (e.g. Teh andHoulsby, 1991; Su and Liao, 2002; Morris and Williams, 1994, 2000;Houlsby, 1982; Koumoto and Houlsby, 2001). However, theoretical

rberg).

. This is an open access article under

approaches generally suffer from a number of limitations making theo-retical evaluations from different test methods difficult (e.g. Low et al.,2010; Casey and Germaine, 2013) and there is still a need for empiricalcorrelations between test results and cu evaluated from well-definedreference tests. Like other empirical interpretation of geotechnical infor-mation, such correlations are still useful especially for low to moderaterisk projects (Robertson, 2012).

Sulphide soils predominantly occur in tropical areas around theworld but they also occur in the coastal areas around the Gulf of Bothniain Sweden and Finland, Fig. 1a. These sulphide soils are fine-grained andhave a relatively high organic content. According to Swedish practice, cuof inorganic clays is evaluated from CPT, FVT and FC through empiricalrelations according to Table 1. Sulphide soils have previously also beenevaluated using these relations, but the results were then often incon-sistent and differed from experience andmore accurate determinations(Westerberg et al., 2005). Alternatively an empirical factor of 0.5, whichis a conservative and lower boundary value for organic soils (Larsson,1990), has been used for FCs and FVTs. There is thus a need for reliablecorrelations between a reference method and the commonly used CPT,FVT and FC.

The purpose of this study is to determine the relationships betweencu evaluated from direct shear tests (DSSs) and results evaluated fromCPTs, FVTs, FCs and the preconsolidation pressure (σ′c) assessed from

the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Fig. 1. a) Approximate location of sulphide soils around the Gulf of Bothnia (from Schwab, 1976). b) Location of the five sites. The cities of Kalix and Umeå are marked in both figures.

78 B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

constant rate of strain tests (CRSs) of Swedish fine-grained sulphidesoils. Results from five main sites were investigated, but results fromthree earlier investigations were also included in the analyses. Thestudy was performed with the assumption that the evaluated cu fromDSSs can be considered as a reference value, which is common practicein Sweden. Many researchers have proposed empirical correlations re-lated to various geotechnical parameters, and the results in this studywere therefore analysed with respect to clay content, liquid limit (wL),overconsolidation ratio (OCR), and organic content. Thefirst three prop-erties may affect the relations for inorganic clays and often also for or-ganic soils.

Table 1Evaluation of cu for FCs, FVTs, CPTs and σ′c assessed from CRSs according to Swedishpractice for inorganic clays.

No Equation References and comments

1 cu = μτc Larsson et al. (1984), Larsson et al. (2007b). ForFC and FVT respectively, where τc and τv are theuncorrected values from FC and FVT.

2 cu = μτv3

μ ¼ 0:43wL

� �0:45≥0:5

4cu ¼ τv 0:43

wL

� �0:45OCR1:3

� �−0:15 Larsson et al. (1984), Larsson et al. (2007b). Inoverconsolidated soils (OCR N 1.3), a correctionfor OCR is included for FVT.

5 cu ¼ qnetNkt

¼ qnet13:4þ6:65wL

Larsson et al. (2007b). For CPT in normallyconsolidated soils.

6 cu ¼ qnetNkt

OCR1:3

� �−0:2 Larsson et al. (1984), Larsson et al. (2007b). ForCPT in overconsolidated soils (OCR N 1.3)

7 cu = aσ′cOCR− (1 − b) Larsson et al. (2007b). For σ′c assessed fromCRS. The factor a is dependent on the soil typeand on the mode of shear. The factor b typicallyvaries between 0.7 and 0.9.

8 a = 0.125 + 0.205wL / 1.17 Larsson et al. (2007b). Factor a for direct shear.

2. Historical background

The FVT is traditionally used to determine cu in many countries, e.g.Canada, USA, Japan, Norway, and Sweden. More recently, the CPThas also become widely used. In Sweden, the FC has been used in thelaboratory for almost 100 years (SJ, 1922) and FVT has been usedsince the 1940s. Different reference methods have been used to cali-brate empirical factors, e.g. the correction factors (μ) for the evaluationof cu from FCs and FVTs and the cone factors (Nkt) from CPTs. Ideally,the results from full-scale field tests and failures are used as references(e.g. Schwab, 1976; Larsson, 1980; Azzouz et al., 1983), but this isnot usually possible and, in most cases, the correction must be basedon results from reference tests in the laboratory. The results from triax-ial tests and DSSs have been considered to be reliable and relevant (e.g.Larsson, 1980). The results from triaxial compression tests are relevantfor the special case of active shear, whereas the results from DSSs areconsidered to be relevant for both direct shear in a horizontal planeand the average stress along a failure surface including approximatelyequal parts of active and passive shear (e.g. Ladd and Foott, 1974;Graham, 1979; Larsson, 1980; Ladd, 1991). DSSs are therefore oftenused as reference tests in Sweden.

The evaluation of cu from FCs in Sweden has been revised manytimes over the years, mainly in connection with the introduction ofnew samplers yielding a better sample quality. When the Swedishstandard piston sampler was developed in the 1950s (and finally fin-ished in 1961), the FC was calibrated to yield the same cu as the FVT(Hansbo, 1957). Empirical factors based on wL were introduced by theSwedish Geotechnical Society, SGF, in 1960 (SGF, 1960). More recently,both types of tests have been calibrated against DSSs (Larsson et al.,2007b). The correction of measured strength from FVTs has been fur-ther developed including the relation between the in situ effective

79B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

vertical stress σ′v0 and cu (Morris and Williams, 1994; Aas et al., 1986;Chung et al., 2007, 2010, 2012), and OCR (Larsson and Åhnberg, 2003,2005).

Values of cu evaluated from DSSs have also been used for the calibra-tion of Nkt used to derive cu from measured net cone resistance (qnet)assessed from CPTs (e.g. Larsson and Mulabdic, 1991; Larsson andÅhnberg, 2003, 2005). For inorganic clays in Sweden, the relation be-tween cu and qnet (qnet = qt − σv0), where qt is total cone resistance andσv0 is total overburden pressure, has been found to be dependent on wL

and OCR, as proposed by Larsson and Åhnberg (2003, 2005), Table 1.Strain rate, creep, yield stresses, and anisotropy affect the results

from FVTs as well as from CPTs (e.g. Chung et al., 2007; Larsson et al.,2010). However, as long as the tests are performed at a standard shearrate and with a standard mode of shear, the evaluation inherently takesthese effects into account.

Fig. 2. Soil properties, stresses, net cone resistance and uncorrected strength at the sites:

The assessment of cu from σ′c, also requires an empirical factortaking into account several parameters such as the shear rate, themode of shear, and the type of soil material. It is commonlysuggested that cu is related to σ′c and OCR (e.g. Ladd and Foott,1974; Jamiolkowski et al., 1985; Larsson and Åhnberg, 2003),Table 1.

Organic soils are usually included in the common evaluationmethods for fine-grained soils (e.g. Bihs et al., 2010; Bajda andSkutnik, 2010; Mlynarek et al., 2006; Lechowicz and Szymanski,2001). However, experience and a more general investigation ofthe behaviour of organic soils (e.g. Larsson, 1990) have shown thatsome organic soils and particularly sulphide soils often do not fitinto this pattern. Schlue et al. (2010) have proposed new shearrate dependent factors of FVT to evaluate cu of organic mud. Howev-er studies related to organic soils are considerably fewer than those

a) Gammelgården, b) Sunderbyn, c) Hjoggböle, d) Västerslätt and e) Umeå bangård.

Fig. 2 (continued).

80 B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

on inorganic soils, and site-specific correlations are needed for organicsoils (Saye et al., 2013).

3. Methods

The test results from the FCs, the FVTs, the CPTs, and CRSswere com-pared to values of cu obtained from DSSs on soil samples taken at thetest sites. The samples were taken with a Swedish standard piston sam-pler (SGF, 1996). Empirical factors (μ, Nkt and a) fitting the relationsstated in Table 1 were assessed from these comparisons. The resultswere then analysed in order to investigate the occurrence of correla-tions between these empirical factors and the properties clay content,wL, OCR, and organic content.

At the sites, CPTs and FVTs were performed in situ and laboratorytests in the form of DSSs, FCs, and CRSs (in order to determine σ′c)were made on undisturbed samples taken at the sites. All testswere made according to Swedish standard practice (SS 027125,1991; SS 027126, 1991; SGF, 1993a, 1993b; SGF, 1996, and SGF,2004). Swedish routine testing on clay samples, to determine bulk

Table 2Properties and classified soil types for the sulphide soils at the sites.

Site Gammel-gården Sunderbyn Hjogg

Main investigated depth (m) 2–12 2–5 2–5Clay content (%) 23–35 33–34 14–2St ≈8 ≈10 ≈14OCR 1.1–3.1 1.3–2.4 5.0–6Classified soil type Organic sulphide clay/clayey

sulphide gyttjaOrganicsulphide clay

Organsulph

density (ρ), natural water content (wN), wL and sensitivity (St), and soiltype classification was also performed. The organic content was deter-mined by analyses of the total organic carbon content (TOC), and thetotal sulphur and total iron contents were also determined.

In theDSSs, the test specimenswere consolidated for 0.85σ′c follow-ed by unloading to the in situ stresses before they were sheared inundrained conditions. This is thenormal procedure in Sweden, to recon-solidate the samples and to avoid seating errors without the risk ofincreasing the strength of the samples. The DSSs were conducted atroom temperature (about+20 °C), and evaluated according to Swedishstandard (SS 027127, 1991), and cu was evaluated as the shear stress atan angle of distortion of 0.15 rad (unless a peak-value was obtained at alower shear strain). The limit of 0.15 rad is in accordance with Swedishpractice. In Sweden, CRSs are performed according to SS 027126 (1991).The normal rate of compression of the 20 mm high test specimen is0.6%/h, but the rate is decreased in cases where the pore pressure atthe undrained end exceeds 15% of the applied vertical pressure and itmay be increased if the pore pressure is too low to evaluate the perme-ability and coefficient of consolidation.

böle Västerslätt Umeå bangård

2–10 2–83 10–25 6–19

4–11 ≈9.7 1.7–4.9 2.8–5.1ic sulphide clay/organicide silt

Organic sulphide silt (and as organicsulphide clay below 10 m depth

Organicsulphide silt

81B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

4. Sulphide soils and test sites

4.1. General properties of sulphide soils

Sulphide soils are common along the coast of the Gulf of Bothnia, i.e.in the north-eastern parts of Sweden over a distance of about 900 kmand also in the north-western parts of Finland, Fig. 1a. These sulphidesoils were formed during the last 8000 years, as fine-grained sedimentssettled at river mouths and at areas near the coast in brackish/freshwater, where the environment was deficient in oxygen (Mácsik, 1994).The environment prevented a complete decomposition of the organic

Fig. 3. Plots of cu,DSS versus data obtained from a) FCs, b) FVTs, c)

material and iron monosulphide (FeS) and pyrites (FeS2) were formed.Sulphide soils generally consist of organic matter and iron sulphides inaddition to clay and silt particles (Pusch, 1973; Larsson, 1990). Accordingto the Swedish classification, a soil is designated as an inorganic mineralsoil when the organic content is less than 2%, as an organic mineral soilwhen it is between 2 and 6%, as a mineral organic soil when it is 6–20%,and as an organic soil when it is 20% or higher (Karlsson and Hansbo,1984). The investigated sulphide soils were classified from organic siltto organic clay and, in cases with higher organic contents, as silty gyttjaor clayey gyttja. Sulphide soils found in other parts of theworldmaydifferin origin, deposition and composition from the sulphide soils found along

CPTs, d) CPTs with correction for OCR and e) σ′c from CRS.

82 B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

the Gulf of Bothnia, and the mechanical properties, including the empiri-cal factors suggested in the present study, may therefore differ.

Typical ranges for some properties of the sulphide soils found inSweden around the Gulf of Bothnia are: total sulphur content 0.1–2%of dry weight; total iron content 2–5% of dry weight; organic content2–7% of dry weight; wN = 50–150%; and ρ = 1.3–1.7 t/m3. Typicallycu is 10–20 kPa. Sulphide soils are in general very compressible andshowsignificant creep behaviour. The organicmatter and iron sulphidesare believed to contribute to the open structure of the material, hence;the relatively low ρ and high wN. Due to the varved stratification ofthe sulphide soils, the properties may vary significantly, even within adepth of centimetres. The sulphide soils are usually black in colour orshow varved with black bands originating from iron monosulphides.Deposits of sulphide soils can be over 20m thick and are typically slight-ly overconsolidated to overconsolidated.

4.2. Test sites

Investigations were performed at five sites within a distance ofabout 350 km along the north-eastern coast line of Sweden, fromKalix in the north to Umeå in the south, Fig. 1b. The properties of thetested soils vary but their geological histories are similar. The sites are des-ignated, with the nearest city in parentheses, Gammelgården (Kalix),Sunderbyn (Luleå), Hjoggböle (Skellefteå), Västerslätt (Umeå) andUmeå bangård (Umeå), Fig. 1b. Gammelgården is located in sloping ter-rain in the Kalix river valley, whereas Sunderbyn is located in an areawith level ground in the Luleå river valley, close to Road97. TheHjoggbölesite is located in a slightly inclined terrain near Road 364, far from anyriver, and the Västerslätt and Umeå bangård sites are located in an areawith a level ground surface in the Umeå river valley, about 1 km apart.

Fig. 4. Evaluated cu,DSS/τc from FCs as functions of a) cl

4.3. Geotechnical soil properties

4.3.1. General soil propertiesA selection of results from the site investigations of the sulphide soils

found at the five sites is presented in Fig. 2, where the soil properties ρ,wN,wL, organic content, iron content and sulphur content are presented.Vertical preconsolidation pressures (σ′vc) assessed from CRSs, σ′v0, qnet,τc, τv and cu assessed fromDSSs are also included. Table 2 shows the claycontent, St, OCR and the soil classification for the five sites.

Variations in the properties with depth at the site are generally quitelarge and the properties may also vary at a given depth due to thevarved stratification. As seen in Fig. 2, the values of ρ and wN are influ-enced by the organic content, a high organic content giving rise to a lowvalue of ρ and a high value of wN. The iron content and sulphur contenttogether are generally slightly higher than the organic content.

At the five sites, the soil is overconsolidated with respect to σ′v0throughout the soil profiles, the lowest OCR (=σ′vc/σ′v0) being 1.1–1.2 at Gammelgården and the highest being 6.7 at Hjoggböle. The trendsfor σ′vc and σ′v0 with depth vary between the sites, the highest OCRvalues of about 3–6 being in the upper parts of the soil profiles. WhentheCRS results are comparedwith the criteria for the evaluationof sampledisturbance proposed by Lunne et al. (1997), the results range from “verygood to excellent” to “good to fair”. The rate of compression in the CRSsvaried between 0.6 and 1.3%/h.

4.3.2. Strength dataAt all the sites except Hjoggböle, the cu values obtained from the

DSSs (cu,DSS) are significantly lower than the corresponding values ofτc and τv, and the variations of these properties with depth are similarat each site. The variation of qnet with depth is similar to that of cu

ay content, b) wL, c) organic content, and d) OCR.

83B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

from the DSSs at the five sites. The cu value was evaluated as the shearstress at an angle of distortion of 0.15 rad for most of the DSSs. For afew tests, a peak-value was obtained at a distortion angle of between0.10 and 0.15 rad.

5. Results and discussion

In Fig. 3, τc, τv,qnet (with and without correction for OCR), andσ′cOCR−0.2 are plotted against corresponding values of cu,DSS. In eachdiagram, the best fit linear regression lines (black lines) are presentedtogether with the upper and lower 95% confidence intervals for theseregression lines (dashed lines, assuming Student's t distributions) andwith ρc as the coefficient of correlation. Taking the averages of all datapoints for each method, cu,DSS/τc = 0.65 was obtained for the FCs,cu,DSS/τv = 0.65 for the FVTs, qnet/cu,DSS = 20.9 for the CPTs, qnet(OCR/1.3)−0,2/cu,DSS = 20.2 for the CPTs with correction for OCR, and cu,DSS/(σ′cOCR−0.2) = 0.28 (assuming b = 0.8) for σ′c assessed from CRSs. Inthe figures, the dashed dotted lines show these relationships.

As seen in Fig. 3, the confidence intervals are relatively narrow andthe dashed dotted lines for CPTs (with and without correction forOCR) and for σ′c assessed from CRS lie within these intervals, whilethe dashed dotted lines representing FC and FVT lie, at least in part, out-side the confidence intervals. This indicates that the empirical factorsgiven above for the assessment of cu,DSS based on qnet obtained fromCPTs and on σ′c assessed from CRSs (i.e. Nkt = 20.9, Nkt = 20.2 anda = 0.28) are valid over a large stress interval, whereas the empiricalfactors for FC and FVT (i.e. μc = 0.65, μv = 0.65) show less consistencywith stress level indicating that these latter factors may be dependenton the stress level.

Fig. 5. Evaluated cu,DSS/τv from FVTs as functions of a) c

Figs. 4 to 8 show the empirical factors, cu,DSS/τc for FC, cu,DSS/τv forFVT, qnet/cu,DSS for CPTs, qnet(OCR/1.3)−0,2/cu,DSSfor CPTs with correctionfor OCR, and cu,DSS/(σ′cOCR−0.2) plotted against clay content,wL, organiccontent, and OCR, in order to investigate whether there are any correla-tions with these properties. In Figs. 3 to 8 evaluated data reportedin Larsson et al. (2007a) from three other sites with sulphide soils(Gideåbacka, Teg and Rödäng) are also included. In the figures, thebest fit linear regression lines (black lines) are presented togetherwith the upper and lower 95% confidence intervals for these regressionlines (dashed lines, assuming Student's t distributions) and dashed dot-ted lines representing the empirical factors suggested above. For FC(Fig. 4), no significant correlations were found and the hypothesis of aconstant value of μc = 0.65 cannot be rejected at the 95% level of confi-dence. The same holds for FVT (Fig. 5) with the exception of OCR, andthe hypothesis of a constant value of μv = 0.65 cannot be rejected atthe 95% level of confidence. For CPT without correction for OCR(Fig. 6), there seem to exist (albeit rather vague) correlations with wL,organic content and OCR. However, when CPT has been corrected forOCR (Fig. 7) no significant correlation with any of the studied parame-ters can be observed, and the hypothesis of a constant value of Nkt =20.2 cannot be rejected at the 95% level of confidence. For σ′c assessedfrom CRS there are no obvious correlations with any of the studiedproperties and the hypothesis of a constant value of a = 0.28 cannotbe rejected at the 95% level of confidence. In Figs. 4b–8b, dotted linesrepresenting the suggestions for inorganic clays (Table 1), includingcorrelations with wL, are also presented. It is obvious that these linesare not suitable for the sulphide soils.

The results shown in Fig. 9 have been evaluated both in accordancewith Swedish practice for fine-grained soilswith a correction forwL (de-noted Corr wL; Table 1) and in accordance with the correction factors,

lay content, b) wL, c) organic content, and d) OCR.

Fig. 6. Evaluated qnet/cu,DSS from CPTs as functions of a) clay content, b) wL, c) organic content, and d) OCR.

84 B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

cone factors and empirical factors proposed for sulphide soil; for FCs0.65, for FVTs 0.65, for CPTs 20 with a correction for OCR, and for σ′cassessed from CRSs with a = 0.28. The results of the FCs, FVTs, CPTsand σ′c assessed from CRSs agree better with the results of the DSSsthan the results obtained following Swedish practice for fine-grainedsoils, Fig. 9a–d. Since there is no correlation with wL for the empiricalfactors for FCs, FVTs, CPTs and σ′c assessed from CRSs, Figs. 4–8, it canbe expected that the empirical equations for inorganic soils (Table 1)will not be applicable to sulphide soils. For cu-values higher thanabout 25 kPa, a correction factor of 0.65 for FVTs does not fit the DSS re-sults, suggesting that the correction factor depends on the stress level.Theremay also be other factors, such as stratification and sample distur-bance, which affect the correlations between strength evaluated fromDSS and strength from FCs, FVTs, CPTs and σ′c assessed from CRSs.There is thus a need for more detailed experimental data in order to as-sess and explain possible correlations with other factors in the case ofsulphide soils.

Experience from Swedish inorganic clays says that for OCR correc-tions should be made for strength values obtained from CPTs and FVTsbut not for values from FCs, and that the correction is larger for CPTs(Larsson and Åhnberg, 2003, 2005). The results of the present studyfor sulphide soils also indicate that there are dependence on OCR forstrength values from CPTs and from FVTs.

A possible explanation of the correlation between empirical factorsobtained from CPTs and from FVTs and OCR could be the differentmodes of deformation and effective stress paths developed during theshearing in the tests. For the CPT, the penetration of the cone causesthe soil to be displaced (predominantly) horizontally (Larsson et al.,2010; Löfroth, 2008), and the mass displacement is large. This leads toincreases in the total horizontal stress and in the excess pore pressure.

The pore pressure increase is atfirst lower than the increase in total hor-izontal stress, and this leads to an increase in the effective horizontalstress.With increasing deformation, the effective horizontal stressfinal-ly reaches the horizontal preconsolidation pressure. From this stage on,the increase in pore pressure corresponds to the increase in total hori-zontal stress, since the effective yield stress cannot be exceeded in un-drained conditions, and the effective horizontal stress remains equalto the horizontal preconsolidation pressure (Larsson, 1977; Larssonet al., 2010). The vertical stresses below the cone also increase, but thehorizontal preconsolidation pressures govern the behaviour of the soilat cone penetration. In CPTs, the effective horizontal stresses thus cor-respond to those in a normally consolidated state and the total cone resis-tance depends on this. The strength evaluated from cone resistancemeasured in a normally consolidated state should therefore be correctedfor the change in strength that occurs when the soil is unloaded and be-comes overconsolidated. The evaluated cu value should thus be adaptedto correspond to the prevailing stress state in situ.

In the FVT, the mass displacement is less than in the CPT when thevane is inserted. In inorganic clays, the cu values also have to becorrected for OCR. In organic soils such as sulphide soils, the soil is lessrigid and this leads to a relatively smaller increase in horizontal stressduring installation of the vane than with inorganic clay. The test istherefore performed at the natural state of consolidation and the needfor correction is less significant. For the FC, there is so far no researchshowing that a correction should be made for OCR in any type of soil.

For the studied sulphide soils, none of the three constituent parts(i.e. the soil particles consisting of clay minerals, organic matter andiron sulphides) seem to significantly affect the empirical factors neededfor the assessment of cu from the results of the methods investigated.Nor is there any obvious relation towL or the organic content. However,

Fig. 7. Evaluated qnet(OCR/1.3)−0,2/cu,DSS from CPTs (correcting for OCR) as functions of a) clay content, b) wL, c) organic content, and d) OCR.

85B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

both the organic content and the clay content probably influence theempirical factors, and the sulphide soil can be expected to behave likea soil intermediate between a mineral (inorganic) soil and an organicsoil, which is supported by the following comparisons: The calibratedNkt of 20.2 for CPTs when correcting for OCR, is approximately the aver-age of the typical values of around 16 for inorganic clays and around 24for gyttja (Larsson, 2007). The resulting μc= μv=0.65 for FCs and FVTs isalso approximately the average of the typical values of about 0.8 for inor-ganic clays and 0.5 for soils with a high organic content (Larssonet al., 2007b; Larsson, 1990). The factor cu,DSS/(σ′cOCR−0.2) = 0.28is approximately the average of the typical values of about 0.22 forinorganic clays and 0.4 for organic soils (Larsson, 1990). The defor-mation at failure for the DSSs on sulphide soils was larger thanthat normally obtained for inorganic clays, which is a typical un-drained behaviour for soils containing organic material and a highersilt content (Yu, 1993).

6. Conclusions

For Swedish fine-grained soils classified as sulphide soils, the mea-sured strength values from FCs and FVTs should be corrected by a factorμ = 0.65 according to

cu ¼ 0:65τC;V ð1Þ

where cu is the undrained shear strength corresponding to a direct shearmode, and τc,v is the strength value obtained from FCs (τc) or FVTs (τv).However, care should be takenwhen using this relationship for FVTs forhigher stresses since the appropriate correction factor can be dependenton the stress level.

For CPTs, cu (corresponding to a direct shear mode) should be eval-uated using a cone factor of Nkt = 20.2 ≈ 20 and a correction for theoverconsolidation ratio OCR according to

cu ¼ qt−σv0

20OCR1:3

� �−0;2ð2Þ

where qt is the total cone resistance and σv0 is the total overburdenpressure.

For a vertical preconsolidation pressure, σ′vc = σ′c, assessed fromCRSs, cu (corresponding to a direct shear mode) should be evaluated as

cu ¼ 0:28σ 0cOCR

−0:2: ð3Þ

The results of the FCs, FVTs, CPTs and σ′c assessed from CRSs, evalu-ated in accordance with the factors proposed for sulphide soils, agreebetter with the results of the DSSs than the results obtained followingSwedish practice for fine-grained soils with a correction for wL.

No correlations were found between the empirical factors for FCs,FVTs, CPTs and σ′vc assessed from CRSs, and the clay content,wL, or or-ganic content, but there was a relation to the OCR for the CPTs and theFVTs.

Acknowledgements

The authors wish to express their gratitude to the following organi-zations that funded this study: the Swedish Transport Administration,the Sven Tyréns Foundation, the Swedish Geotechnical Institute, theLuleå University of Technology, and the Swedish Civil ContingenciesAgency.

Fig. 8. Evaluated cu,DSS/(σ′cOCR−0.2) as functions of a) clay content, b) wL, c) organic content, and d) OCR.

86 B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

References

Aas, G., Lacasse, S., Lunne, T., Höeg, K., 1986. Use of in situ tests for foundation design onclay. ASCE Geotech. Spec. Publ. 6, 1–30.

Azzouz, A.S., Baligh, M.M., Ladd, C.C., 1983. Corrected field vane strength for embankmentdesign. J. Geotech. Eng. ASCE 109 (5), 730–734.

Bajda, M., Skutnik, Z., 2010. Determination of mechanical parameters of organic soil usingCPT and SCPT. 14th Danube-European Conference on Geotechnical Engineering:From Research to Design in European Practice, Bratislava, June 2010. Proceedings(8 pp.).

Bihs, A., Long, M., Marchetti, D., Ward, D., 2010. Interpretation of CPTU and SDMT in or-ganic, Irish soils. 2nd International Symposium on Cone Penetration Testing,CPT'10, Huntington Beach, CA, May, 2010. Proceedings vol. 2, pp. 257–264.

Casey, B., Germaine, J.T., 2013. Stress dependence of shear strength in fine-grained soilsand correlations with liquid limit. J. Geotech. Geoenviron. Eng. ASCE 139 (10),1709–1717.

Chung, S.G., Kim, G.J., Kim, M.S., Ryu, C.K., 2007. Undrained shear strength from field vanetest on Busan clay. Mar. Georesour. Geotechnol. 25 (3–4), 167–179.

Chung, S.G., Chung, J.G., Jang, W.Y., Lee, J.M., 2010. Correlations between CPT and FVT, re-sults for Busan clay. Mar. Georesour. Geotechnol. 28 (1), 49–63.

Chung, S.G., Hong, Y.P., Lee, J.M., Min, S.C., 2012. Evaluation of the undrained shearstrength of Busan clay. KSCE J. Civ. Eng. 16 (5), 733–741.

Graham, J., 1979. Embankment stability on anisotropic soft clays. Can. Geotech. J. 16 (2),295–308.

Hansbo, S., 1957. A new approach to the determination of the shear strength of clay bythe fall cone test. Royal Swedish Geotechnical Institute, SGI. Proceedings No. 14 (46pp. Stockholm).

Houlsby, G.T., 1982. Theoretical analysis of the fall cone test. Geotechnique 32 (2),111–118.

Jamiolkowski, M., Ladd, C.C., Germaine, J.T., Lancelotta, R., 1985. New developments infield and laboratory testing of soils. Theme lecture. 11th International Conferenceon Soil Mechanics and Foundation Engineering, San Francisco, CA, Aug. 1985. Pro-ceedings vol. 1, pp. 57–153.

Karlsson, R., Hansbo, S., 1984. Jordarternas indelning och benämning. Geotekniskalaboratorieanvisningar 2. Byggforskningsrådet, Stockholm (T21:1982, In Swedish).

Koumoto, T., Houlsby, G.T., 2001. Theory and practice of the fall cone test. Geotechnique51 (8), 701–712.

Ladd, C.C., 1991. Stability evaluation during staged construction. J. Geotech. Eng. ASCE 117(4), 537–615.

Ladd, C.C., Foott, R., 1974. New design procedure for stability of soft clays. J. Geotechn. Eng.Div. ASCE 100 (GT7), 763–786.

Larsson, R., 1977. Basic Behaviour of Scandinavian Soft Clays. Swedish Geotechnical Insti-tute, SGI, Linköping (Report, No 4, 108 +/21/pp.).

Larsson, R., 1980. Undrained shear strength in stability calculation of embankments andfoundations on soft clays. Can. Geotech. J. 17 (4), 591–602.

Larsson, R., 1990. Behaviour of Organic Clay and Gyttja. Swedish Geotechnical Institute,SGI. Report No. 38 (153 pp. Linköping).

Larsson, R., 2007. CPT-sondering. Statens geotekniska institut, SGI, Linköping (Informa-tion Nr 15. 71 pp, 2 rev utg., in Swedish).

Larsson, R., Åhnberg, H., 2003. Long-Term Effects of Excavations at Crests of Slopes; PorePressure Distribution — Shear Strength Properties — Stability — Environment. Swed-ish Geotechnical Institute, SGI, Linköping (Report No. 61, 368 pp.).

Larsson, R., Åhnberg, H., 2005. On the evaluation of undrained shear strength andpreconsolidation pressure from common field tests in clay. Can. Geotech. J. 42 (4),1221–1231.

Larsson, R., Mulabdic, M., 1991. Piezocone Tests in Clay. Swedish Geotechnical Institute,SGI, Linköping (Report 42, 240 pp.).

Larsson, R., Bergdahl, U., Eriksson, L., 1984. Utvärdering av skjuvhållfasthet ikohesionsjord. Statens geotekniska institut, SGI, Linköping (Information Nr 3, 28pp. in Swedish).

Larsson, R., Westerberg, B., Albing, D., Knutsson, S., Carlsson, E., 2007a. Sulfidjord –

geoteknisk klassificering och odränerad skjuvhållfasthet (Sulphide soil — geotechni-cal classification and undrained shear strength). Swedish Geotechnical Institute. SGI,Linköping (Report No. 69/Luleå University of Technolog. 135 pp., in Swedish).

Larsson, R., Sällfors, G., Bengtsson, P.-E., Alén, C., Bergdahl, U., Eriksson, L., 2007b.Utvärdering av skjuvhållfasthet i kohesionsjord. Statens geotekniska institut, SGI,Linköping (Information Nr 3, 28 pp., in Swedish).

Larsson, R., Åhnberg, H., Löfroth, H., 2010. CPT data in relation to a general model for theundrained behaviour of homogeneous Swedish clays. 2nd International Symposiumon Cone Penetration Testing, CPT'10, Huntington Beach, CA, May, 2010. Proceedingsvol. 2, pp. 145–152.

Lechowicz, Z., Szymanski, A., 2001. Evaluation of undrained shear strength of cohesiveand organic soils from in situ tests. 15th International Conference on Soil Mechanicsand Geotechnical Engineering: The First International Conference of the ThirdMillenium, Istanbul, August, 2001. Proceedings vol. 1, pp. 445–448.

Löfroth, H. (2008). Undrained shear strength in clay slopes — Influence of stress con-ditions. A model and field test study. Diss. Chalmers University of Technology.Department of Civil and Environmental Engineering. Division of GeoEngineering.

Fig. 9. Plots of the corrected strength cu,corr in accordance withwL for inorganic clays (Table 1) and in accordance with empirical factors proposed for sulphide soil versus cu,DSS for a) FCs0.65, b) FVTs 0.65, c) CPTs 20 (with correction for OCR), and d) σ′c assessed from CRSs with a = 0.28.

87B. Westerberg et al. / Engineering Geology 188 (2015) 77–87

Geotechnical Engineering 2846/Swedish Geotechnical Institute, SGI. Report 71,195 pp. Göteborg.

Low, H.E., Lunne, T., Andersen, K.H., Sjursen, M.A., Li, X., Randolph, M.F., 2010. Estimationof intact and remoulded undrained shear strengths from penetration tests in softclays. Geotechnique 60 (11), 843–859.

Lunne, T., Berre, T., Strandvik, S., 1997. Sample disturbance effects in soft low plastic Nor-wegian clay. Symposium on Recent Developments in Soil and Pavement Mechanics,Rio de Janeiro, June, 1997, pp. 81–102.

Mácsik, J., 1994. Risken för utfällning av ferriföreningar ur dräneringsvatten från anaerobaoch aeroba sulfidjordar. Licentiate thesis, Luleå University of Technology Division ofSoil Mechanics and Foundation Engineering, Luleå, p. 10 (in Swedish).

Mlynarek, Z., Tschuschke, W.,Wierzbicki, J., Marchetti, S., 2006. Interrelationship betweenshear and deformation parameters of gyttja and peat from CPT and DMT tests. 13thDanube-European Conference on Geotechnical Engineering: Active Geotechnical De-sign in Infrastructure Development, Ljubljana, May, 2006. Proceedings (6 pp.).

Morris, P.H.,Williams, D.J., 1994. Effective stress vane shear strength correction factor cor-relations. Can. Geotech. J. 31 (3), 335–342.

Morris, P.H., Williams, D.J., 2000. A revision of Blight's model of field vane testing. Can.Geotech. J. 37 (5), 1089–1098.

Pusch, R., 1973. Influence of Organic Matter on the Geotechnical Properties of Clays.Byggforskningsrådet, Stockholm (Document D11:1973, 63 pp.).

Robertson, P.K., 2012. James K. Mitchell Lecture: interpretation of in-situ tests— some in-sights. International Conference on Geotechnical and Geophysical Site Characteriza-tion, Porto de Galinhas, Brazil, September 2012. Proceedings vol. 1, pp. 3–24.

Saye, S.R., Lutenegger, A.J., Santos, J., Kumm, B.P., 2013. Assessing overconsolidation ratiosin soil with piezocone: referencing soil index properties. J. Geotech. Geoenviron. Eng.ASCE 139 (7), 1075–1085.

Schlue, B.F., Moerz, T., Kreiter, S., 2010. Influence of shear rate on undrained vane shearstrength of organic harbor mud. J. Geotech. Geoenviron. Eng. ASCE 136 (10),1437–1447.

Schwab, E., 1976. Bearing Capacity, Strength and Deformation Behaviour of Soft OrganicSulphide Soils. (Ph.D. Thesis). Royal Institute of Technology. Department of Soil andRock Mechanics, Stockholm.

SGF, 1960. Recommended correction of shear strength values from FCs and FVTs. MinutesFrom aMeeting Held by the Swedish Geotechnical Society at the Swedish Geotechni-cal Institute. Stockholm.

SGF, 1993a. Recommended Standard for Cone Penetration Tests. Swedish GeotechnicalSociety, SGF, Linköping (Report 1:93E. 26 pp.).

SGF, 1993b. Recommended Standard for Field Vane Shear Test. Swedish Geotechnical So-ciety. SGF, Linköping (Report 2:93E. 11 pp.).

SGF, 1996. Geoteknisk fälthandbok. Allmänna råd och metodbeskrivningar. SvenskaGeotekniska Föreningen, SGF, Linköping (Rapport 1:96. 147 pp., in Swedish).

SGF, 2004. Direkta skjuvförsök – en vägledning. Svenska geotekniska föreningen,Linköping (SGF Notat 2:2004, 27 pp., in Swedish).

SJ, 1922. Statens järnvägars geotekniska kommission 1914–22. Slutbetänkande angivettill Kungl. Järnvägsstyrelsen den 31 maj 1922. Geotekniska Meddelanden 2. StatensJärnvägar, Stockholm (180 pp., in Swedish).

SS 02 71 25, 1991. Geotekniska provningsmetoder – Skjuvhållfasthet – Fallkonförsök –

Kohesionsjord (Geotechnical tests — shear strength — fall-cone test — cohesivesoil). Swedish Standard SS 027125. Swedish Building Standards Institution,Stockholm (11 pp., in Swedish).

SS 02 71 26, 1991. Geotekniska provningsmetoder – Kompressionsegenskaper –Ödometerförsök – CRS-försök (Geotechnical tests — compression properties —

oedometer tests — CRS tests). Swedish Standard SS 027126. Swedish Building Stan-dards Institution, Stockholm (11 pp., in Swedish).

SS 02 71 27, 1991. Geotekniska provningsmetoder. Skjuvhållfasthet - direkta skjuvförsök,CU- och CD-försök – kohesionsjord. Geotechnical tests— shear strength— direct sim-ple shear tests, CU and CD tests — cohesive soil. Swedish Standard SS 027127. Swed-ish Building Standards Institution, Stockholm (13 pp., in Swedish).

Su, S.F., Liao, H.J., 2002. Influence of strength anisotropy on piezocone resistance in clay.J. Geotech. Geoenviron. Eng. ASCE 128 (2), 166–173.

Teh, C.I., Houlsby, G.T., 1991. An analytical study of the cone penetration test in clay.Geotechnique 41 (1), 17–34.

Westerberg, B., Albing, D., Larsson, R., 2005. Research on strength and deformation prop-erties of Swedish fine-grained sulphide soils. 16th International Conference on SoilMechanics and Geotechnical Engineering, Osaka, September, 2005. Proceedings vol.2, pp. 623–626.

Yu, Y., 1993. Testing and Modelling of Silty and Sulphide-Rich Soils. (Ph.D. Thesis). LuleåUniversity of Technology, Department of Civil Engineering, Luleå.