The Japanese Geotechnical Society

Soils and Foundations

Soils and Foundations 2013;53(5):671–679

0038-0http://d

nCorE-m

hino@ihamadayoshim

1Tel.2Tel.3Tel.4Tel.Peer

806 & 201x.doi.org/

respondinail addrelt.saga-u.a@cc.sagaura@soila/fax: þ81: þ81 952/fax: þ81: þ81 6 6review un

www.sciencedirect.comjournal homepage: www.elsevier.com/locate/sandf

Interpretation of density profile of seabed sediment from nuclear densitycone penetration test results

Rui Jiaa,n, Takenori Hinoa,1, Jinchun Chaib,2, Takaharu Hamadaa,3, Mitsugu Yoshimurac,4

aInstitute of Lowland and Marine Research, Saga University, 1 Honjo-machi, Saga City, Saga 840-8502, JapanbDepartment of Civil Engineering and Architecture, Saga University, 1 Honjo-machi, Saga City, Saga 840-8502, Japan

cSoil and Rock Engineering Co., Ltd., 2-21-1 Shonaisakae-machi, Toyonaka City, Osaka 561-0834, Japan

Received 1 February 2013; received in revised form 24 May 2013; accepted 15 June 2013Available online 24 September 2013

Abstract

The nuclear density cone penetration test (ND-CPT) is used to investigate the wet density (ρ) of soft soil deposits under the sea. In the ND-CPT,the density is determined by the count rate ratio (Rρ) of the gamma rays detected by the detector to the gamma rays from the source. Thus, the ND-CPT measures the average wet density of the material within a spheroid volume around the central point of the source and the detector, which isdesignated as ρc. The measured ρc values are considerably different from the actual ρ of the seabed sediment because the density varies significantlywith depth. We designate the layer of seabed sediment in which the density increases rapidly with depth as the “upper layer” and the layer below itas the “lower layer”. A method is proposed to deduce the actual ρ profile from the measured ρc profile. The upper and the lower boundaries of theupper layer of the seabed sediment can be directly determined from the measured ρc profile as well as the distance (2b) between the gamma raysource and the detector. Based on the results of a back analysis of the measured ρc profile at Isahaya Bay, Ariake Sea, Japan,the density distribution in the upper layer of the seabed sediment can be approximated as a square root function of depth. Comparisons ofthe measured ρc profiles, obtained using the ND-CPT, and the actual ρ profiles, deduced by the proposed method, show that the actual ρ areobviously different from the measured ρc near the upper and the lower boundaries of the upper layer of the seabed sediment. The thickness of theupper layer of the seabed sediment at Isahaya Bay lies in the range of 0.05–0.55 m.& 2013 The Japanese Geotechnical Society. Production and hosting by Elsevier B.V. All rights reserved.

Keywords: Nuclear density cone penetration test (ND-CPT); Seabed sediment; Density profile

3 The Japanese Geotechnical Society. Production and hosting by10.1016/j.sandf.2013.08.005

g author. Tel.: þ81 952 28 8582; fax: þ81 952 28 8189.sses: jiarui@ilt.saga-u.ac.jp (R. Jia),c.jp (T. Hino), chai@cc.saga-u.ac.jp (J. Chai),-u.ac.jp (T. Hamada),ndrock.co.jp (M. Yoshimura).952 28 8612.28 8580; fax: þ81 952 28 8190.952 28 8494.331 6031; fax: þ81 6 6331 6243.der responsibility of The Japanese Geotechnical Society.

1. Introduction

Seabed sediments generally exhibit significant variations inphysical properties near the water–sediment interface (Ross andMehta, 1989; Winterwerp and Van Kesteren, 2004; Holland et al.,2005). The investigation of seabed sediment properties, such aswet density (ρ), is important in predicting sediment transport, indetermining nautical depth and in planning dredging projects(McAnally et al., 2007a, b). However, there are few reliablemethods to adequately measure the density profile of seabedsediments (Ha et al., 2010). Density measurements based on thecore samples are costly and sometimes unreliable because of thedifficulty involved in obtaining undisturbed samples. It is hard toobtain a useful density profile by acoustic methods because the

Elsevier B.V. All rights reserved.

Fig. 1. Nuclear density cone penetrometer.

R. Jia et al. / Soils and Foundations 53 (2013) 671–679672

reflection strength is proportional to the density gradient, not aspecific density (McAnally et al., 2007b). The application of atuning fork requires complex and extensive calibration, and it islimited to low-density fluid mud (Ha et al., 2010).

Nuclear transmission and backscatter techniques have beendemonstrated to measure density accurately; they are the firstchoice in several areas (McAnally et al., 2007b). The techniqueshave been used for many years in geotechnical engineering to mea-sure the wet density of soil (Homilius and Lorch, 1958; Ruygrok,1988). The bulky instrument employed in the nuclear backscattertechnique was developed and incorporated into the piezoconepenetrometer, which is the most versatile and widely used tool forin situ soil exploration (Ledoux et al., 1982; Lunne et al., 1997;Mimura et al., 1995; Shibata et al., 1992, 1994). The nuclear den-sity cone penetration test (ND-CPT) has been used for soil investi-gations on land (Dasari et al., 2006; Karthikeyan et al., 2007;Mimura and Shrivastava, 1998; Shibata et al., 1994) and overclosed water (Umezaki et al., 2005, 2009), and it can provide relia-ble ρ profiles. However, there are few data regarding the measure-ment of the density profile of seabed sediments using the ND-CPT.

In the ND-CPT, the density is evaluated in a spheroid volumearound the central point of the gamma ray source and thedetector; it is called the “measuring volume” (Dasari et al., 2006;Karthikeyan and Tan, 2008). Thus, the density measured by theND-CPT is the average wet density of the material (ρc) within aspheroid volume, which is not exactly the actual ρ at the centralpoint of the gamma ray source and the detector. If the densitydoes not change much with depth, the measured ρc profile canapproximate the actual ρ profile. However, there may be aconsiderable difference between the measured ρc profile and theactual ρ profile of the seabed sediment due to the significantvariation in density with depth.

In this study, a method is proposed to deduce the actual ρprofiles from the measured ρc profiles, and the proposed method isused to interpret the ND-CPT measurements at Isahaya Bay,Ariake Sea, Japan. Firstly, the ND-CPT equipment and the methodused to investigate the seabed sediment are described. Then, thetypical density profile measured by the ND-CPT is presented.Based on the theory that the ND-CPT measurement provides theaverage wet density of the material (ρc) within the spheroidmeasuring volume, the difference between the measured and theactual density profiles is illustrated. The backscatter model and theaverage model that can calculate ρc from a known actual densityprofile were used to interpret and back analyze the measureddensity profiles. Finally, the actual ρ profiles were deduced fromthe measured ρc profiles and compared with the measured ρcprofiles, and the differences between them are discussed. Thededuced actual ρ profiles were adopted to determine the thick-nesses of the upper layer (where the density increases rapidly withdepth) of the seabed sediment at Isahaya Bay.

2. ND-CPT equipment and test method

2.1. Description of nuclear density cone penetrometer

Fig. 1 shows the nuclear density cone penetrometer. Thelower part, which is for piezocone testing, consists of a cone

with an apex angle of 601 and a base diameter of 35.7 mm(10 cm2 cross-sectional area), a pore pressure filter at theshoulder of the cone and a 150 cm2 friction sleeve behind thefilter. The upper part, which is for the density measurement,consists of an inclinometer, a gamma ray source, a lead shield,a density count gamma ray detector and a background countgamma ray detector; it has a diameter of 48.6 mm and a lengthof 1124.5 mm. The distances of the gamma ray source, thedensity count gamma ray detector and the background countgamma ray detector from the apex of the cone are 0.721 m,0.986 m and 1.4535 m, respectively.

2.2. Density measurement principle and measuring volume

The principle of the density measurement in the ND-CPTrelies on the Compton scattering of gamma rays, whose energylies in the range of 0.3–2.0 MeV, passing through matter. Thegamma rays emitted from the source collide and scatterrepeatedly with the atomic electrons in the material. In thisprocess, parts of the gamma rays are absorbed by the material,while others reach the detector. The amount of gamma rays

R. Jia et al. / Soils and Foundations 53 (2013) 671–679 673

reaching the detector is related to the density of the material;thus, the density of the material can be determined.

To obtain an accurate count rate of the gamma rays from thesource, the background count rate (BCR) measured by thebackground count gamma ray detector (the amount of naturalgamma rays) must be subtracted from the density count rate(DCR) measured by the density count gamma ray detector.Then, the accurate count rate of the gamma rays from thesource is normalized by the standard count rate (SCR) toobtain the count rate ratio (Rρ), which is used to calculate thedensity, i.e.,

Rρ ¼DCR�BCR

SCRð1Þ

To reduce the statistical error in the amount of gamma raysemitted from the source and, at the same time, to ensure the preci-sion of the test results, the average BCR and DCR over a depthof 100 mm were used in this study. SCR can be measured using astainless steel chamber, with a diameter of 0.7 m and a heightof 1 m, filled with pure water before the in situ investigations(Jia et al., 2013; Karthikeyan, 2005). The equation used to calcu-late the ρ of the soils from the Rρ value is determined based on acomparison of the in situ measurements by the ND-CPTs to thelaboratory measurements of thin-wall tube samples of clay orfrozen sand samples (Nobuyama, 2000), and the proposed rela-tionship is as follows:

Rρ ¼ Aρ2�BρþC ð2Þwhere A, B and C are constants whose values are 0.6264, 4.0954and 6.8422, respectively, over a density range of 1.0–2.2 g/cm3.

The density measurement is evaluated in an extendedspheroid volume around the central point of the gamma raysource and the density count gamma ray detector, which is

Fig. 2. Test l

Table 1Index properties of soil at locations c2, g2, c5 and g6.

Location Clay content (%) Silt content (%) Sand content (%) D

c2 61.7 34.1 3.1 2g2 58.8 27.8 10.3 2c5 61.6 31.9 5.6 2g6 54.8 30.9 7.7 2

called the “measuring volume” (Fig. 1). In the figure, the valuefor 2b is 26.5 cm, which is the same as the distance betweenthe gamma ray source and the detector of the cone penetrom-eter used in this study. The value for a decreases with anincreasing density of the surrounding material, and it can beestimated based on the theory of gamma ray scattering(Homilius and Lorch, 1958) as follows:

a

2b¼ 6:67

ðρU2bÞ0:61 ð3Þ

where ρ is the wet density of the material in g/cm3. The valuefor a is approximately 240 mm if the surrounding material iswater, which is identical to the value obtained from the labo-ratory measurements (Karthikeyan, 2005). For clay, loose sandand dense sand, with densities of 1.4 g/cm3, 1.8 g/cm3 and2.0 g/cm3, the values for a are approximately 200 mm,170 mm and 160 mm, respectively.

2.3. Locations of field investigation and test method

The locations investigated are shown in Fig. 2. A total of 48ND-CPTs (30 outside and 18 inside the dike), simultaneouslymeasuring the density and the undrained shear strength (su), wereconducted to investigate the properties of the seabed sedimentaround the dike at Isahaya Bay, Ariake Sea, Japan. Table 1 showssome of the index properties of the soil at locations c2 and g2(inside the dike) and locations c5 and g6 (outside the dike). Themain clay mineral of the sediment in Ariake Sea is smectite(Ohtsubo et al., 1995; Mizota and Longstaffe, 1996).Since the tests performed in this study were conducted at sea,

and seabed sediments are very soft, the ND-CPT equipment wasset on a boat and the cone penetrometer was penetrated by a dead

ocations.

ensity of solid particles (g/cm3) Plasticity index Water content (%)

.644 97.5 189.3

.676 81.6 147.7

.677 94.8 175.8

.699 101.1 147.4

R. Jia et al. / Soils and Foundations 53 (2013) 671–679674

load (Fig. 3). The weight of the cone penetrometer was approxi-mately 15 kg and the additional weight was 60 kg. A winch wasused to control the up and down movements of the cone pene-trometer, and the penetration speed was approximately 1–2 cm/s.The cone might be inclined during testing due to the motion of theboat caused by the waves. The measured maximum inclination ofthe cone from the vertical was approximately 31, which is withinthe acceptable error range.

Fig. 4. Typical ND-CPT results at location a1.

3. Density profile measured by ND-CPT

3.1. Typical ND-CPT results

The cone resistance (qc), the pore pressure at the shoulder of thecone (u), the sleeve friction (fs), the background count rate (BCR)and the density count rate (DCR) can be simultaneously measuredby the ND-CPT. The wet density is calculated according to themeasured BCR and DCR by Eqs. (1) and (2). Typical ND-CPTresults for ρc, the corrected cone resistance (qt) and the porepressure (u), at location a1 (see Fig. 2), are shown in Fig. 4. In thefigure, u0 is the hydrostatic pore water pressure.

The measured density profiles vary significantly around thewater/clayey deposit interface. The profiles can be divided intothree sections, namely, an initial section close to the region ofconstant density for water, a middle curved section and a finallinear section. The measured values of density are sometimessmaller (or larger) than 1.0 g/cm3 in the region of constantdensity for water. This is due to a statistical error in the amountof gamma rays emitted from the source, although it is minimizedby applying the average BCR and DCR over a depth of 100 mm.

Fig. 3. ND-CPT equipment and method for seabed sediment investigation.

3.2. Relationship between measured and actual densityprofiles

Since the ND-CPT measures the average wet density of thematerial within a measuring volume, when the density of thematerial to be measured changes significantly with depth, there willbe a considerable difference between the actual density profile andthe measured density profile, as illustrated in Fig. 5. Point A is theintersection of the initial constant density section and the middlecurved section, while point B is the intersection of the final linearsection and the middle curved section. We designate the layer ofseabed sediment in which the density increases rapidly with depthas the “upper layer” and the layer below it as the “lower layer”.The measured ρc will increase once the gamma ray sourceenters the upper layer of the seabed sediment, while the centralpoint between the source and the detector still does not enterthe sediment. Then, when the center enters the lower layer ofthe seabed sediment, the detector is still in the upper layer andthe measured ρc will still be in the middle curved region. Thus, thethickness of the curved region is equal to the thickness of the upperlayer of the seabed sediment plus the gamma ray source-detectorseparation distance (2b¼26.5 cm for the device used). Based onthis argument, it is assumed that the actual upper layer starts atpoint C (distance b below point A) and ends at point D (distance babove point B).In the field, above the seabed sediment, there are suspended

solids (SS) in the seawater. Normally, the SS concentration is afew kg/m3. The field measurements indicate that the ND-CPTdoes a poor job of estimating the concentration of SS. Theobjective of this study was to develop a method to estimate theactual density of sediments and/or soil–water mixtures with solidconcentrations above 10 kg/m3. An SS concentration of 10 kg/m3

was adopted as the lower limit, because this concentration oftencorresponds to the lutocline at the top of a fluid mud layer(Kineke et al., 1996; Kirby and Parker, 1977; Ross and Mehta,1989), and it corresponds to a density increment of approximately0.006 g/cm3. For simplicity, we propose that the density at pointC (ρC) can be approximated to equal the density at point A (ρA),and that the density at point D (ρD) can be approximated to equalthe density at point B (ρB) minus γb (where γ is the slope of the

Fig. 5. Relationship between measured and actual density profiles.

Fig. 6. Laboratory-measured density of seabed sediment.

R. Jia et al. / Soils and Foundations 53 (2013) 671–679 675

final linear region). We propose the following method to deter-mine the locations of points C and D.

(1)

Location of C. Considering that when the gamma raysource enters the upper layer of the seabed sediment, themeasured density curve will start to diverge from thevertical line (line with a density of 1.0 g/cm3); the locationof point A is identified as the intersection of the verticalline and the measured density curve. Then, the location(depth) of C is the location of A plus the b value.

(2)

Location of D. Assuming that when both the gamma raysource and the detector are in the lower layer of the seabedsediment, the measured density profile will increase in a linearfashion, and when the detector is in the upper layer and thesource is in the lower layer, the measured profile will becurved; the location of point B can be identified as theintersection of the measured density curve and the linearregression line of the measured density lying in the lowerlayer. Then, the location (depth) of D is the location of Bminus the b value.

The remaining problem is to determine the distributionbetween points C and D in Fig. 5. The method adopted toidentify the distribution is as follows: the actual profile isfirstly assumed; then, ρc is back calculated for the soil–watermass within the spheroid measuring volume of the ND-CPT;finally, if the calculated ρc profile fits the measured one, theassumed profile is deemed the actual profile.

Undisturbed samples with a diameter of 0.075 m and a heightof 1 m were obtained from the test area at Isahaya Bay forlaboratory density measurements. The samples were cut intoslices, 50 mm in height for the top 100 mm and then 100 mm inheight for the remaining soil; the average wet density of eachslice was measured immediately after the sampling. Typical testresults at locations c1 and g6 (see Fig. 2) are shown in Fig. 6. Byreferring to the laboratory density measurements, it is assumedthat the possible density distributions in the upper layer of theseabed sediment can be (1) linear or (2) a square root function ofdepth. Their suitability is investigated in the next sections.

4. Theoretical models for interpreting measured densityprofile

4.1. Theoretical models

4.1.1. Backscatter modelThe backscatter model was developed to predict the labora-

tory performance of the nuclear backscatter probe (Parker et al.,1975). The three assumptions adopted are as follows.

(1)

The volume affecting the measured Rρ is a spheroid, asshown in Fig. 7. The contribution of the count rate ratiofrom a small volume in the spheroid to the total count rateratio is proportional to its volume.

(2)

The value of 2b is a constant equal to the gamma raysource-detector separation distance of the cone penetrom-eter. For simplification, the value for a is also assumed tobe a constant. In this study, the values for 2b and a are26.5 cm and 24 cm.

(3)

The relationship between Rρ and ρ is expressed by Eq. (2)based on a comparison of the in situ measurements ofND-CPTs to the laboratory measurements on soil samples.

Based on the above assumptions, the total count rate ratiowithin the measuring volume (Rρc) can be calculated by thefollowing equation:

Rρc ¼3

4b3

ZðAρðzÞ2�BρðzÞþCÞ½b2�ðz�ziÞ2�δz ð4Þ

where z is depth and zi is the depth corresponding to themidpoint of the measuring volume. The value for ρc, which isreferenced to the midpoint of the spheroid, can be calculated asfollows:

ρc ¼4:0954� ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

2:5056Rρc�0:3715p

1:2528ð5Þ

Then, the ρc profile can be obtained as the measuring volumemoves with depth.

4.1.2. Average value modelThe method is based on the assumption that the ND-CPT

measurement provides an average density of the soil within themeasuring volume. The assumed measuring volume is the

R. Jia et al. / Soils and Foundations 53 (2013) 671–679676

same as that shown in Fig. 7. Therefore, the average valuemodel can be applied to directly calculate ρc by the followingequation:

ρc ¼3

4b3

ZρðzÞ½b2�ðz�ziÞ2�δz ð6Þ

As in the backscatter model, the ρc profile can be obtained asthe measuring volume moves with depth.

4.2. Demonstration of difference between actual andcalculated density profiles

Two different density profiles are assumed in Fig. 8. Eachprofile has three layers. The first and the third layers have constantdensity values, while the density varies with depth in the second(middle) layer as a linear or square root function. The backscatterand average value models are used to calculate the average densitywithin the measuring volume, assuming that the ND-CPT passesthrough the profiles. The calculated density (ρcc) profiles do notshow the same transition as the assumed actual density profiles atthe interfaces of the layers. For the calculated profile, there is atransition zone near the interface between the two layers with athickness of b (half of the source-detector separation distance). Theresults in Fig. 8 confirm the argument made in Section 3.

Fig. 7. Spherical coordinates for density measurement using ND-CPT.

Fig. 8. Comparison of assumed

The calculated results show that the ρcc profiles calculated bythe two models are almost the same. This is because ρ changesapproximately linearly with Rρ in the range of 1.0–2.2 g/cm

3 basedon Eq. (2). It is suggested, therefore, that the simple average valuemodel can be used to back interpret the actual ρ profile of theseabed sediment from the measured ρc profile by the ND-CPT.

5. Interpretation of density profile of upper layer of seabedsediment

5.1. Back analysis of the measured density profile

As shown in Fig. 9, the density profile for the water/suspen-sions is assumed to be constant, ρz ¼ ρ0, while the density profilefor the lower layer of the seabed sediment is assumed to be linearwith depth, ρz ¼ ρDþγðz�zDÞ. The density profile for the upperlayer of the seabed sediment is assumed to be a linear (①) or asquare root function of depth (②), respectively, by the followingequations:

ρðzÞ ¼ βðz�zCÞþρC ðlinearÞ ð7Þ

ρðzÞ ¼ αðz�zCÞ1=2þρC ðsquare rootÞ ð8ÞThe values for β and α can be calculated by the followingequations:

β¼ ðρD�ρCÞ=ðzD�zCÞ ð9Þ

α¼ ðρD�ρCÞ=ðzD�zCÞ1=2 ð10Þwhere zC and zD are depth values corresponding to the upper andthe lower boundaries of the upper layer of the seabed sediment,and ρC and ρD are density values corresponding to the upper andthe lower boundaries of the upper layer of the seabed sediment.Their values can be determined according to the measureddensity profile, as shown in Fig. 5.The value for ρc at any depth zi is calculated by Eq. (6) for z

varying between (zi�b) and (ziþb), as shown in Fig. 7.Fig. 10 compares the back-calculated average density (ρcc)with the ρc profiles measured by the ND-CPT at four selectedlocations at Isahaya Bay, Japan (indicated as a1, c7, f5 and g7in Fig. 2). The results show a better agreement between ρc

ρ and calculated ρcc profiles.

Fig. 9. Probable actual density profiles.

Fig. 10. Comparison of measured ρc and calculated ρcc profiles.

Fig. 11. Comparison of measured ρc and actual ρ profiles.

R. Jia et al. / Soils and Foundations 53 (2013) 671–679 677

and ρcc with the square root distribution assumption than withthe linear assumption. Based on the above results, we proposethat the actual ρ distribution between points C and D in Fig. 5is a square root function of depth, and that it can be calculatedby Eqs. (8) and (10).

5.2. Interpreted actual density profiles

For the four selected points in Fig. 2 (a1, c7, f5 and g7), theinterpreted actual ρ profiles are compared with the measured ρcprofiles in Fig. 11. The interpreted actual ρ are clearly differentfrom the measured ρc in the upper layer of the seabed sediment.For locations c1 and g6, the laboratory-measured densities areavailable (Fig. 6), and the corresponding interpreted ρ profilesand measured ρc profiles are compared with the laboratory data inFig. 12. The interpreted actual ρ profiles show a better agreementwith the laboratory results. The interpreted actual ρ profile shouldbe adopted to determine the depth of the water and the thicknessof the upper layer of the seabed sediment.

5.3. Thickness distribution of upper layer of seabed sedimentat Isahaya Bay

The thickness distribution of the upper layer of the seabedsediment, determined from the interpreted actual ρ profiles atIsahaya Bay, is shown in Fig. 13. The thicknesses lie in therange of 0.05–0.55 m. The thicknesses of the upper layer of theseabed sediment outside the dike are thicker than those insidethe dike, and the thicknesses of the upper layer of the seabedsediment in the coastal area are thicker than those in the deepwater area outside the dike.

Fig. 12. Comparison of interpreted actual ρ profiles and measured ρc profileswith laboratory results.

Fig. 13. Thickness distribution of upper layer of seabed sediment at IsahayaBay (cm).

R. Jia et al. / Soils and Foundations 53 (2013) 671–679678

6. Conclusions

The nuclear density cone penetration test (ND-CPT) hasbeen widely used to investigate the wet density (ρ) of soft soildeposits on land and under water. The ND-CPT measures theaverage wet density of the material (ρc) within a spheroidvolume around the central point of the gamma ray source andthe detector. When the actual ρ varies significantly with depth,the measured ρc will differ considerably from the actual ρ.The density distribution of seabed sediments is a typicalexample of such a situation. We designate the layer of theseabed sediment in which the density increases rapidly withdepth as the “upper layer” and the layer below it as the “lowerlayer”. A method has been proposed to interpret the actual ρprofile from the measured ρc profile by the ND-CPT, and theproposed method has been used here to interpret the ND-CPTmeasurements at Isahaya Bay, Ariake Sea, Japan. The follow-ing conclusions can be drawn from this study.

(1)

Proposed method. Based on the measured ρc profile, aswell as a back analysis, it has been proposed that (a) theupper (C) and the lower (D) boundaries of the upper layerof the seabed sediment can be directly determined from theND-CPT-measured ρc profile as well as the distance (2b)

between the gamma ray source and the detector (see Fig. 5)and (b) the density distribution of the upper layer of theseabed sediment (between C and D) follows a square rootfunction with depth (Eq. (8)).

(2)

Interpreted ρ profile. Comparisons of the ρc profilesmeasured by the ND-CPT and the interpreted actual ρprofiles by the proposed method show that the actual ρprofiles are clearly different from the measured ρc profilesnear the upper and the lower boundaries of the upper layerof the seabed sediment. The thickness of the upper layer ofthe seabed sediment at Isahaya Bay, determined from theactual ρ profile, lies in the range of 0.05–0.55 m.

References

Dasari, G.R., Karthikeyan, M., Tan, T.S., Mimura, M., Phoon, K.K., 2006. Insitu evaluation of radioisotope cone penetrometers in clays. GeotechnicalTesting Journal 29 (1), 45–53.

Ha, H.K., Maa, J.P.Y., Holland, C.W., 2010. Acoustic density measurementsof consolidating cohesive sediment beds by means of a non-intrusive‘Micro-Chirp’ acoustic system. Geo-Marine Letters 30 (6), 585–593.

Holland, C.W., Dettmer, J., Dosso, S.E., 2005. Remote sensing of sedimentdensity and velocity gradients in the transition layer. Journal of theAcoustical Society of America 118 (1), 163–177.

Homilius, J., Lorch, S., 1958. On the theory of gamma ray scattering inboreholes. Geophysical Prospecting 6 (4), 342–364.

Jia, R., Hino, T., Hamada, T., Chai, J.C., Yoshimura, M., 2013. Density andundrained shear strength of bed sediment from ND-CPT. Ocean Dynamics63 (5), 507–517.

Karthikeyan, M., 2005. Application of Radioisotope Cone Penetrometer toCharacterize a Lumpy Fill. Dissertation. National University of Singapore.

Karthikeyan, M., Tan, T.S., Mimura, M., Yoshimura, M., Tee, C.P., 2007.Improvements in nuclear-density cone penetrometer for non-homogeneoussoils. Soils and Foundations 47 (1), 109–117.

Karthikeyan, M., Tan, T.S., 2008. Profiling of heterogeneous soil usingnuclear-density cone penetrometer. Geotechnical Testing Journal 31 (6),513–525.

Kineke, G.C., Sternberg, R.W., Trowbridge, J.H., Geyer, W.R., 1996. Fluidmud processes on the Amazon continental shelf. Continental ShelfResearch 16, 667–696.

Kirby, R., Parker, W.R., 1977. The Physical Characteristics and EnvironmentalSignificance Fine Sediment Suspensions in Estuaries. Estuaries, Geophy-sics and Environment, National Research Council, Academy, Washington,DC, pp. 110–120.

Ledoux, J.L., Menard, J., Soulard, P., 1982. The penetro-gammadensimeter. In:Proceedings of the 2nd European Symposium on Penetration Testing,ESOPT-II, vol. 2, Amsterdam, Balkema, Rotterdam, pp. 679–681.

Lunne, T., Robertson, P.K., Powell, J.J.M., 1997. Cone Penetration Testing inGeotechnical Practice. Blackie Academic & Professional, London.

McAnally, W.H., Friedrichs, C., Hamilton, D., Hayter, E., Shrestha, P.,Rodriguez, H., Sheremet, A., Teeter, A., 2007a. Management of fluidmud in estuaries, bays, and lakes. I: present state of understanding oncharacter and behavior. Journal of Hydraulic Engineering 133 (1), 9–22.

McAnally, W.H., Teeter, A., Schoellhamer, D., Friedrichs, C., Hamilton, D.,Hayter, E., Shrestha, P., Rodriguez, H., Sheremet, A., Kirby, R., 2007b.Management of fluid mud in estuaries, bays, and lakes. II: measurement,modeling, and management. Journal of Hydraulic Engineering 133 (1),23–38.

Mimura, M., Shrivastava, A.K., 1998. RI-cone penetrometers experience innaturally and artificially deposited sand. In: Proceedings of the 1stInternational Conference on Site Characterization, vol. 1, Atlanta,pp. 575–580.

Mimura, M., Shrivastava, A.K., Shibata, T., Nobuyama, M., 1995. Performance ofRI cone penetrometers in sand deposits. In: Proceedings of the International

R. Jia et al. / Soils and Foundations 53 (2013) 671–679 679

Symposium on Cone Penetration Testing, CPT'95, vol. 2, Linkoping, Sweden,pp. 55–60.

Mizota, C., Longstaffe, F.J., 1996. Detrital origin for smectite in soils andsediments from the coastal plain of Ariake Bay, Northern Kyushu, Japan.Geoderma 73 (1-2), 125–130.

Nobuyama, M., 2000. The Result of the RI Cone Penetration Tests. NUSInternal Report No. 6004998/264. Soil and Rock Engineering Co., Ltd.,Japan.

Ohtsubo, M., Egashira, K., Kashima, K., 1995. Depositional and post-depositional geochemistry and its correlation with geotechnical propertiesof the marine clays in Ariake bay. Geotechnique 45, 509–523.

Parker, W.R., Sills, G.C., Paske, R.E.A., 1975. In situ nuclear densitymeasurement in dredging practice and control. In: 1st InternationalSymposium on Dredging Technology, University of Kent, England,pp. 25–42.

Ross, M.A., Mehta, A.J., 1989. On the mechanics of lutoclines and fluid mud.Journal of Coastal Research, SI 5, 51–62.

Ruygrok, P.A., 1988. Evaluation of the gamma and neutron radiation scatteringand transmission methods for soil density and moisture determination.Geotechnical Testing Journal 11 (1), 3–19.

Shibata, T., Mimura, M., Shrivastava, A.K., Nobuyama, M., 1992. Moisturemeasurement by neutron moisture cone penetrometer: design and applica-tion. Soils and Foundations 32 (4), 58–67.

Shibata, T., Mimura, M., Shrivastava, A.K., 1994. Use of RI cone penetrom-eter data in foundation engineering. In: Proceedings of the 13th Interna-tional Conference on Soil Mechanics and Foundation Engineering, vol. 1,New Delhi, India, pp. 147–150.

Umezaki, T., Kawamura, T., Yoshimura, M., 2005. Evaluation of sedimenta-tion and consolidation properties using RI-density log for dredging andreclamation. In: Proceedings of the International Workshop on theMitigation and Countermeasures of Ground Environment (IW-SHIGA2005), pp. 101–106.

Umezaki, T., Kawamura, T., Yoshimura, M., 2009. Investigation of sedimentenvironments in closed water body using RI-density Log. In: Proceedingsof the 17th International Conference on Soil Mechanics and GeotechnicalEngineering, Alexandria, Egypt, pp. 1050–1053.

Winterwerp, J.C., Van Kesteren, W.G.M., 2004. Introduction to the Physics ofCohesive Sediment in the Marine Environment. Elsevier B.V., Amsterdam,the Netherlands.