Geotechnical Characterization of Suape Soft Clays , Brazil

Comprehensive research has been carried out by the Geotechnical Group (GEGEP) of the Federal University of Pernambuco in the soft clay deposits in Northeastern Brazil near the city of Recife. This paper presents the results of important geotechnical investigations of soft clays in two areas within the Suape Port and Industrial Complex. The geotechnical parameters were obtained from laboratory (classification, compressibility and strength) and in situ (SPT, vane and CPTU) tests, and were compared with regional empirical correlations and proposals presented in literature. The vane tests were performed to obtain undrained strength and overconsolidation ratio parameters. The classifications for soil behavior, together with flow characteristics, strength and overconsolidation ratio parameters, were determined by piezocone tests. The results are compared with results from reference tests, and discussed with results from the literature, including the results of Recife and other Brazilian clays. This study confirms that parameters can be obtained by means of in situ tests with correlations suited to the local/regional experience and the importance of having a joint laboratory and in situ test program. This prediction is fundamental for a proper geotechnical site characterization in studies and engineering projects.


Introduction
The Suape Port and Industrial Complex is a very complete and important industrial center in Northeast Brazil.The geographical location of Pernambuco State gives Suape Port an international status since it is located on the main international shipping routes.Large companies, a shipyard, refinery and other industries already exist or their facilities are under construction in this area.
The use of field tests to evaluate geotechnical parameters of soils has been increasing in recent years.Coutinho (2008) (see also Coutinho et al., 2008) published a study about the geotechnical parameters obtained from in situ investigations for practical projects.The ability to obtain parameters by means of in situ tests with correlations suited to the local/regional experiment is fundamental for proper geotechnical site characterization in studies and engineering projects.Soil stratigraphy, compressibility and rate of consolidation parameters, undrained shear strength (Su) and the overconsolidation ratio (OCR) of soft clays can be estimated based on an investigation program including in situ and laboratory testing (Schnaid, 2009;Mayne, 2007;Lunne et al., 1997).
This paper presents the results of important geotechnical investigations of soft clays carried out in two areas within the Suape Port and Industrial Complex.The geotechnical parameters were obtained from laboratory (classification, compressibility and strength) and in situ (SPT, water content measurement, vane and CPTU) tests, and then compared with regional and proposed correlations presented in literature.The vane tests were used for obtain-ing undrained strength and overconsolidation ratio parameters.The classifications concerning soil behavior, in addition to flow characteristics, strength and overconsolidation ratio parameters were determined by piezocone tests.The results were discussed after comparing the laboratory tests and regional and literature results.This study is part of a research program of the Geotechnical Research Group (GEGEP) of the Federal University of Pernambuco (UFPE).

Characteristics of the Study Area
The AE-1 and AE-2 study areas are within the Suape Industrial and Port Complex, in the town of Ipojuca, Pernambuco State, Brazil (Fig. 1).The coastal location is characterized by a complex geology, including low-lying plains, with soils featuring reduced load capacities, very soft organic clays, with the presence of peat, roots, shells and fine layers of sand and silt.This soft soil generally has a high water and organic content with very low penetration test values (N SPT ).Projects included construction of embankments of varying heights (some 17 m) on the subsoil of clays and peat material.
The AE-1 study area is part of an access route where a temporary embankment around 2.0 m in height, has been built.The AE-2 study area, which includes an important project, was divided into five (5) subdivisions, where 20 boreholes and 34 undisturbed Shelby samples were taken.Geotechnical site characterization included laboratory (characterization, oedometer and triaxial) and field (SPT, vane and CPTU) tests.Figures 2 and 3 illustrate typical soil parameter profiles for SUB-AREA A and C, respectively (AE-2 study area).Figure 4 presents a plasticity chart with laboratory test results on Suape soft soils, including organic soils.Results from two other Brazilian soft deposits are also presented (Recife-PE and Juturnaíba-RJ).Proposed ranges for inorganic and organic clays and peats are included.Further information about the geological and geotechnical characterization, together with parameters of the Suape study areas, can be found in Coutinho (2010) and Bello (2011).Robertson & Campanella (1983), Robertson et al. (1986) and Sully et al. (1988) were the first to present charts based on piezocone tests that include measurements for cone resistance (q t ) corrected for pore pressure (Eq.1).
where u 2 = pore pressure measured at the cone shoulder; a = the ratio between the shoulder area unaffected by the pore water pressure and the total shoulder area.

B u u q
where B q = pore pressure ratio; u 0 = in-situ pore pressure; s v = total vertical overburden stress at the depth z corresponding to the readings.Robertson (1990) proposed a refinement of the Robertson et al. (1986) profiling chart, plotting "normalized cone resistance", Q t , against "normalized friction ratio", F r against pore pressure ratio B q , and presented a nine-zone chart (Eqs.3 and 4 respectively), (Fig. 5).Normalization was proposed to compensate for q c dependency on the overburden stress, and when analyzing deep CPTU soundings (deeper than 30 m).Profiling charts developed for shallower soundings are not suitable for deeper sites.
where f s = sleeve friction.Robertson (2010) suggested updating the charts supplied by Robertson et al. (1986), and Robertson (1990).The updated charts, which are dimensionless and color coded for improved presentation, define nine (9) SBT zones that are consistent.Schneider et al. (2008) developed general soil classification charts using parametric studies (Fig. 6).The "zones" in the three charts are exactly the same, but the plots are shown in different formats: (1) coordinates log Q -log Du 2 /s' vo ; (2) coordinates Q -Du 2 /s' vo ; (3) coordinates semilog Q -Du 2 /s' vo .These formats are best used in cases of: (1) clays, clayey silts, silts, sandy silts, and sands without negative penetration pore pressures; (2) sands and transitional soils with small negative excess penetration pore pressures, and (3) clay soils with large negative excess penetration pore pressure.Long (2008) (see also Mollé, 2005;Coutinho, 2008) in a special study concluded that the Robertson et al. (1986) and Robertson (1990) charts for soil classification using CPTU data seem to work well in clays, clayey silts, silty sands and sands.They may have difficulties in using these charts in organic soils and peat, and in cases of complex stratigraphy.
size distribution and N SPT values were compared with the CPTU classification.

Robertson (1990) charts results
In the Robertson (1990) chart (log Q -B q ), it was observed that the points tend to rise when the overconsolidation ratio (OCR) value increases (Figs. 7 and 8).These points are relative to the surface layer, and the layer below the clay layer, where silty and sandy material (bands 5, 6 and 7) can be found.These results were concordant with OCR and sensitivity (S t ) values obtained from oedometer and vane field tests, respectively.The points indicating clay soils are situated in bands 3 and 4.
In the Robertson (1990) chart (log Q -F r ), the points are situated in bands 3 and 4 (clay and silty clay).The F r values were greater than 1.0.
The Robertson (1990) charts predict soil behavior, and are not directly related to soil classification criteria, using geological descriptions based on grain-size distribution.It has been verified that the greatest difference  (Robertson, 1990).
Figure 6 -Proposals for soils classification using CPTU testing for several forms of plotting (Schneider et al., 2008).
between the results of classification from grain-size distribution and CPTU tests occurs in the mixed soil regions and in the transition zones between layers of the profile.

Schneider et al. (2008) charts results
The Schneider et al. (2008) chart -Case 1 (Coordinates log Q -log Du 2 /s' vo ) features coordinates identical to the Robertson (1990) chart, allowing direct comparison of the results (Figs. 7 and 8).In general, soil classification of the study area by the Schneider et al. (2008) chart -Case 1 was in agreement with the classification from Robertson (1990).
In the Schneider et al. (2008) chart -Case 3 (Coordinates semi log Q -Du 2 /s' vo ), the plotted results only took into consideration material with positive pore pressure values, since the horizontal coordinate is a semi-log scale.
Therefore, this chart does not show materials with negative value of pore-pressure.
It is important to emphasize the care needed when identifying the points relating to SPT tests.Classification of layers indicating grain-size distribution can demonstrate differences in the values for N SPT and water content, making it necessary to divide classification into sub-layers to be plotted into separate charts.
Another major concern is the presence of transitional soils.These soils are characterized as mixtures of different materials.In the Schneider et al. (2008) charts, these soils can be clearly observed inside the determined boundaries (Figs. 7 and 8).In general, results from the Schneider et al. (2008) charts were also satisfactory for the Suape soft soils, consistent with classification using grain-size distribution and N SPT values.(Bello, 2011;Bello & Coutinho, 2012).

Compressibility Parameters
Laboratory incremental oedometer tests were performed on soil specimens with a diameter of 87 mm and height of 20 mm, using Bishop apparatus, with double drainage.Loads were doubled for each stage, beginning at 5-10 kPa until 640-1280 kPa and then decreased to 10 kPa.Each stage usually took 24 h.Typical curves of void ratio vs. effective stress obtained from oedometer tests are shown in Fig. 9 for the two research sites.It can be seen that the "virgin" portion is not linear.This is consistent with findings in many other investigations (Coutinho, 1976;Coutinho & Lacerda, 1987;Mesri & Choi, 1985).In this study, the virgin portion curve was simplified by two linear parts to obtain the compression index C c (C c1 and C c2 ).
The quality of samples was evaluated using the proposal presented in Coutinho (2007), which represents the Brazilian experiment, based on the proposal of Lunne et al. (1997).This criterion uses the ratio De/e 0 corresponding to the initial effective vertical stress (s' vo ), and can be described as: OCR = 1-2.5;De/e 0 < 0.05; -Very Good to Excellent; De/e 0 = 0.05-0.08-Good to Fair; De/e 0 = 0.08-0.14-Poor; and De/e 0 > 0.14 -Very Poor.Some samples in this study showed results with considerable disturbance.In the study, these curves e (Î v ) vs. log p or the compressibility parameters were corrected to obtain results equivalent to samples of very good to excellent quality.Methodology proposed by Oliveira (2002), Schmertmann (1955), Coutinho (2007) and Futai ( 2010) was used and are briefly presented in Coutinho & Bello (2012a).In this paper, the results are in general those effectively corrected or initially of very good quality.
Figures 2 and 3 show results of the overconsolidation ratio (OCR = s' vm /s' vo ) and the compressibility parameters: compression index C c (first part) and swell index C s for two sites studied.It can be observed that the deposits have a higher void ratio and very high compressibility.
Statistical correlations (C c vs. e o and C c vs. w (%)) for Suape soft clays have been developed using all data from the corrected laboratory test database.The results of C c vs. w are shown in Fig. 10).Results of Juturnaíba organic soils (Coutinho, 1986;Coutinho & Lacerda, 1987) are also shown, as are the results from Recife soft clays (Coutinho et al., 1998); Coutinho (2007); and Juturnaiba organic soils (Coutinho, 1986;Coutinho & Lacerda, 1987).The water content w (%), obtained from the SPT testing can be used to have the first estimate of C c .
The C c vs. w (%) correlations (Figs.10a and b) are quite similar for the deposit, particularly for clay soils.The Suape clays presenting a smaller inclination: Recife Soft Clay C c @ 0.0126w (%) and Suape Soft Clay C c = 0.0097w (%) C c @ 0.01w (%).In general, correlations for clay have higher correlation coefficients (r 2 ) and lower standard error (lower dispersion) than those for organic soil/peat.Probably for the more difficulty do obtain sample of very good quality.
The organic soft Juturnaíba research site has higher correlation coefficients than Recife and Suape deposits (less dispersion), due to better quality samples.The behavior of these clay deposits under one-dimensional consolidation is strongly affected by sampling disturbance.The general equation between C c and w (%) for Recife and Suape Soils and Rocks, São Paulo, 37(3): 257-276, September-December, 2014.263 Geotechnical Characterization of Suape Soft Clays, Brazil  clays is very similar to the equation presented by Bowles (1979) for organic silts and clays (C c = 0.0115w), and is also very similar to those presented by Djoenaidi (1985) (see Kulhawy & Mayne, 1990).Almeida et al. (2008) plotted C c vs. w values (%) of seven clays from Rio de Janeiro (C c = 0.013w), including Juturnaíba-RJ clays from Coutinho (1986).Koppula (1981) obtained a similar correlation (C c = 0.010w) for the normally densified clays with low sensitivity (S t < 1.5).In general, it has been found that soft clays and soft organic soils C c = 0.010 to 0.015w (%) (Coutinho, 2007).

Coefficient of Consolidation
The coefficient of consolidation measurement is one of the soil properties major challenges for geotechnical engineering.Parameters for the consolidation rate may be assessed using oedometer (laboratory) and piezocone in situ tests by measuring the dissipation or decay of pore pressure through time after penetration ceases (Lunne et al., 1997).
Field stress and pore-pressures around the piezocone can be calculated using the strain path method, according to formulas provided by Baligh & Lavadoux (1986), and Houlsby & Teh (1988).Baligh & Lavadoux (1986) concluded that dissipation is predominantly in a horizontal direction.The dissipation process can be expressed by a one-dimensional factor of time, as in Eq. 5: where R: piezocone radius; T * : dissipation time; I r : rigidity index (G/S u ); G: shear modulus.Houlsby & Teh (1988) presented the following procedure to determine c h : (a) calculate the difference between pore pressure at the beginning of dissipation, u i , and the hydrostatic pore pressure, u o ; (b) calculate the percentage of dissipation u 50% = (u iu o )/2, and use the experimental curve to determine the actual time taken for 50% of the dissipation to occur, t 50% ; (c) obtain the T* value in Table 1, and calculate c h by using Eq. 5. Robertson et al. (1992) propose a direct estimate of c h from the t 50 value using an abacus.This calculation is valid for I r values varying between 50 and 500, and for cone areas of 10-15 cm 2 .Values for c h obtained from these procedures correspond to soil properties in the preconsolidation band, due to the fact that during penetration, material surrounding the cone undergoes increased levels of deformation, and in this state behaves as soil in recompression (Baligh, 1986;Baligh & Levadoux, 1986).Jamiolkowski et al. (1985) proposed estimating c h in the normal compression band (N.C.) using Eq. 6.The authors presented experimental values for C s /C c that varied in the 0.13-0.15range.
The coefficient of consolidation in the vertical direction can be estimated using permeability in the horizontal and vertical planes from Eq. 7 (Lunne et al., 1997).
where k h and k v represent horizontal and vertical permeability, respectively.
In the study area, c h calculations were carried out according to a formula proposed by Houlsby & Teh (1988).In each CPTU vertical, c h values were calculated in relation to the depths of the dissipation tests in both study areas, considering I r = 50.Figure 11 shows an example of results of pore pressure of dissipation tests, including the procedure to determine t 50% , for the CPTU 120, Sub Area C -AE -2.Values of c h were also determined from laboratory trough radial oedometer tests.
Figure 12a shows value variation relating to the depths from the radial oedometer and piezocone tests for the AE-1 study area.In this figure , c (Houlsby & Teh, 1988).Leroueil & Hight (2003) evaluated the coefficient of consolidation determined by piezocone testing, in addition to other methods (Fig. 13).The in situ c h values within the normal consolidation range are typically 10 times higher than the values deduced from oedometer test results (using the Casagrande method).The in situ c h values in the normal consolidation range are also around two orders of magnitude less than values in the over-consolidated range, deduced from in situ observations.The piezocone dissipation tests appear to be somewhere between the field values obtained for the over-consolidated and the normally consolidated ranges (c v and c h ).This figure also shows the range determined for the laboratory and piezocone c h values of Suape clays.The ranges is situated in the limits shown by Leroueil & Hight (2003).

Undrained and Effective Shear Strength
No single undrained shear strength exists.The in situ undrained shear strength depends on the failure mode, soil anisotropy, strain rate and stress history (Lunne et al., 1997).
The triaxial compression, triaxial extension and simple shear laboratory tests with consolidation in the isotropic or anisotropic condition for effective field stress conditions have been used to obtain undrained strength values in geotechnical engineering studies and projects.
Undrained strength can be determined in the field by means of vane and piezocone tests.In order to use in a project, the strength obtained in a vane test must be corrected by Bjerrum's correction factor (1973).In the piezocone test, empirical correlations appropriate for the area under study based on field and laboratory tests should be used.
In the field vane test, undrained strength (S u ) can be determined from maximum torque obtained with the vane rotation (Eq.8).

S T D
where: T max is the maximum torque measured during the test; D is the vane diameter.
To estimate the value of S u through the piezocone test, three cone factors are normally used: N kt , N ke and N Du , load capacity, effective tip resistance and pore pressure, respectively (Lunne et al., 1997).Undrained shear strength is then  defined by Eq. 9.The q t , q e and Du values are determined from the piezocone tests results.The N kt , N ke and N Du , factors are based on theoretical work or, more often, on empirical experimental correlations using laboratory or in situ tests (Lune et al., 1997).
where q t is cone resistance corrected for pore pressure effects; q e is the effective cone resistance; Du is excess pore pressure; s vo is total vertical pressure.

Laboratory -Triaxial UU-C 2
The undrained strength profiles obtained in laboratory triaxial UU-C tests for the studied verticals in SUB-AREAS A and C are presented in Figs. 15 and 16.The mean values of S u obtained in the triaxial tests in SUB-AREA A were 10.8 ± 5.4 kPa for the first 2.0 m of depth, followed by around 7.4 ± 5.0 kPa to 7.0 m in depth.In SUB-AREA C the mean S u values were 7.8 ± 0.6 kPa throughout the profile.Some of the results obtained in the triaxial UU tests may be influenced by the disturbance of the sample (see Coutinho & Bellom, 2012a), so it is important to compare the results of the triaxial with field vane tests.

Vane tests
The undrained strength profiles obtained in the field vane tests for the studied verticals are presented in Figs. 15 and 16.The S uvane results in the verticals of the SUB-AREA A showed a similar trend in behavior, with a mean S uvane value of 7.5 ± 1.5 kPa to a depth of 2.0 m, and with mean values 11.8 ± 1.1 kPa at a depth of 7.0 m.The E110 vertical presented a high S uvane value at a depth of 4.0 m, possibly due to the presence of roots in the soil.Sensitivity had mean values of around 5. The S u results determined in SUB-AREA C presented mean S u values of 15.8 ± 3.6 kPa up to 2.5 m in depth, starting from this point, the S uvane values increase linearly reaching 21.0 kPa at 7.0 m.In general the S t presented mean values in the 5-10 range.
The mean S u values obtained from the field vane test were generally greater than the S u laboratory values.These results are similar to those obtained in Recife soft clays (Coutinho, 2007), but in the Suape study the laboratory results are more influenced by the sample quality.
The values of the ratio S uvane /s' vo vs. PI for soils in the areas AE-1 and AE-2 (SUB-AREAS A and C) are presented in Fig. 17.The area AE-1 was divided in two stretches of mangrove deposits.
The two deposits of AE-2 (SUB-AREAS A and C) revealed layers with a similar value range as S uvane /s' vo , situated above the curve proposed by Bjerrum (1973) for young clays and around the curve of old clays.The range of the mean S uvane /s' vo values found in the AE-2 study area was 0.45-0.68,much greater than the Recife values (Coutinho et al., 2000).The PI values obtained in the area under study were also much greater than the PI values obtained in Recife.
The mangrove 1 deposit (AE-1) presented values of S uvane /s' vo situated above the curve proposed by Bjerrum (1973) for young clays, while the mangrove 2 deposit (AE-1) had values of S uvane /s' vo slightly below the curve for young clays.
In order to draw up embankment projects it is necessary to use a correction factor according to Bjerrum's proposal (1973).Considering that the plasticity index results in Suape clays were generally high (values up to 150%), the correction factor soils have a value of about 0.6.

Piezocone tests
To estimate S u values using the piezocone test, three experimental parameters were adopted, N ku , N ke and N Du .The value of the experimental parameters was determined by the S u values from field vane tests.
Figure 19 shows the N kt variation of SUB-AREAS A and C. In general, it is observed that N kt varied between 6 and 14 in SUB-AREA A, with a mean value of 10.In SUB-AREA C, N kt varied between 5 and 16, with a mean value of 9.These results show that the two sub-areas have similar N kt variation ranges with a mean value between 9 and 10. Figure 20 presents the mean value of N kt (12 ±1.0) obtained from S uvane for the Recife research sites.From these results, S u for Recife soft clay can be estimated with reasonable confidence for practical purposes (Coutinho et al., Figure 17 -Resistance ratio-Suape, study areas AE-1 and AE-2 -S uvanea /s' vo and PI (Bello, 2011(Bello, ). 2000)).These mean values are about 20-25% greater than the N kt results obtained in Suape.Almeida et al. (2010) found a wide range of N kt mean values (3-20) for deposits of very soft soils in Rio de Janeiro.Schnaid (2009) found N kt values representative of soft clay deposits in Porto Alegre varying between 8 and 16, with a mean value of 11.In general, it can be observed that mean values of Brazilian clays vary around 9 to 12 (Coutinho & Schnaid 2010).
According to Lunne et al. (1997) N kt values tend to increase with an increase in plasticity, decrease with an increase in sensitivity and decrease as B q increases.Brazilian clay results confirm that it is recommended to evaluate in each deposit or at least local experience to obtain representative N kt values.values around 8.0.In general, N Du increases linearly with depth (Bello, 2011;Bello & Coutinho, 2012).In Recife and Suape clays, N Du values varied between 7.5 and 11.0 with mean values around 9.5 (Coutinho, 2007(Coutinho, , 2008;;Coutinho & Schnaid, 2010).La Rochelle et al. (1988) obtained for three Canadian clays N Du values between 7 and 9, using as reference S u values from field vane tests, where the overconsolidation ratio values varied between 1.2 and 50.The variation range of the N ke factor values obtained in the Suape study areas was between 4.0 and 9.0 with mean values around 5.0.In general, N ke increases linearly with depth (Bello, 2011;Bello & Coutinho, 2012).Figures 21a and b show the undrained strength profiles of the verticals E102 (SUB-AREA A) and E128 (SUB-AREA C), obtained through the piezocone using mean valves of the experimental parameters and field vane tests.The S u values derived from N kt showed agreements with S u values obtained from the vane tests.The S u values derived from N Du showed a greater difference in relation to S u vane values, because of the difficulty of an accurate Du measurement, including pore pressure negative values.The S u values deriving from N ke , showed the highest degree of dispersion, which can be explained by the small effective tip resistance value, q e , that was the basis for calculating N ke .

Effective shear strength
The effective shear strength envelope is usually determined by laboratory testing, such as, for example, the triaxial drained compression or excursion test (or undrained with pore pressure measurements).The effective friction angle (f') can be estimated by in situ testing, such as the piezocone, or by statistical correlation with plasticity index (PI) proposed by Bjerrum & Simons (1960).Figure 22a shows this statistical correlation with results from five Brazilian clays, including Recife soft clays.Figure 22b gives the results of f' from Suape clays (AE-1 and AE-2) obtained by triaxial CIU-C test.There is dispersion in the results but it is possible to see that the Bjerrum & Simons (1960) proposal is also satisfactory for a preliminary estimation of f' for Suape soft clays.A specific Suape clay statistical correlation is also shown in Fig. 22b.

Overconsolidation Ratio
The stress history of the soil can be indicated for the profiles of effective field stress (s' vo ), preconsolidation stress (s' vm ) and the overconsolidation ratio (OCR).It constitutes an indispensable factor for the analysis of behavior of clay deposits.Traditionally obtained in oedometer tests, OCR can be estimated from field vane (Chandler, 1988) and piezocone tests (Lunne et al., 1997).
Critical-state soil mechanics, as well as the SHANSEP method showed that normalized undrained strength (S u /s' vo ) increases with an increase in OCR (Eq.10).(11) Tavenas & Leroueil (1987) gathered the data used for AAS et al. (1986) and Chandler (1988) and plotted them using Bjerrum's curve (1973) as reference.The authors found the m value equal to 1 (with small dispersions) (Eq.12).Mayne & Mitchell (1988) developed a database with results of field vane and oedometer tests including index properties of 96 different clays, in order to define a general correlation that could be used to estimate OCR values from field vane tests (Eq.13).The deposits showed: 1 < OCR < 40; 3% < PI < 300%; 1.6 kPa < Suvane < 380 kPa and sensitivity varying from 2 up to high values.OCR values can be obtained from CPTU results through correlations in functions of Q t , where Q t = (q T -s vo ) /s' vo and Du/s' vo (Eqs.14 to 17).OCR (Lunne ., 1989

Oedometer tests
Figure 23 shows the OCR profiles of SUB-AREAS A and C obtained from conventional oedometer tests.In general, OCR oed.values were greater than 1.0 up to 2.0 m in depth and a tendency towards the unit value with depth is observed.Poor quality samples had their OCR oed values corrected (see Coutinho & Bello, 2012b;Bello, 2011).

Vane tests
To estimate the OCR value using the field vane tests performed in the two study areas (AE-1 and AE-2), Chandler (1988), Tavenas & Leroueil (1987) and Mayne & Mitchell (1988) proposals were used.Figure 23a presents the results obtained for the OCR profile of SUB-AREA A, including oedometer tests.The OCR results well differenti- ated along the depths but with similar trends in behavior are observed.
Figure 24 shows OCR results determined through the three proposals compared to OCR results obtained in conventional oedometer tests.Mayne & Mitchell (1988) proposal is the one that comes closest to the laboratory OCR values for the Suape study areas, always presenting greater values, as was the case with Recife clays (Coutinho et al., 2000;Coutinho, 2008).Mayne & Mitchell (1988)  Figure 24b shows the OCR profile obtained in a laboratory and OCR estimated by the Mayne & Mitchell (1988) proposal adapted for Suape clays (Eq.11).A good correlation can be seen between OCR vane and OCR lab values in good quality samples.These correlations can be useful when no good quality samples are available and to provide complementary results in an investigation.

Piezocone tests
The OCR values were estimated through the CPTU test using proposals from: Lunne et al. (1989) and Kulhawy & Mayne (1990).Results of Recife soft clays are also presented (Coutinho, 2007).Figure 25a shows OCR profiles obtained by the three proposals, together with the OCR results obtained through the oedometer test for SUB-AREAS A and C. It can be seen that the correlations of Lunne et al. (1989) and Kulhawy & Mayne (1990) show higher OCR CPTU values than the OCR lab but with similar behavior trends.Coutinho (2007Coutinho ( , 2008) ) showed OCR profile results from Recife soft clays obtained using oedometer tests and   2 shows the summary of results the comparative study and recommended correlations.
Figure 26 shows the study for obtaining the coefficient of the Kulhawy & Mayne (1990) proposal that is suitable for use in Suape clays.The coefficient obtained was equal to 0.173, almost half the original coefficient (0.32), and smaller than the value of 0.23 found for Recife clays by Coutinho (2007Coutinho ( , 2008)).Jannuzi (2009) and Baroni (2010) obtained the coefficient of 0.153 for Sarapuí-RJ soft clay and Barra da Tijuca deposits, respectively.(Coutinho, 2007(Coutinho, , 2008)).Figure 25b shows the OCR profile obtained in the laboratory and OCR estimated from the Kulhawy & Mayne (1990) proposal adapted for Suape clays (OCR = 0.173Q t ).A good correlation between OCR CPTU values and OCR lab values in good quality samples is observed for Suape Clays.This correlation can be useful when there are no good quality samples available, and for further research.

Conclusions
This paper has provided results of a geotechnical site characterization from laboratory and in situ tests performed in an important investigation of soft clays in two areas within the Suape Port and Industrial Complex, Brazil.
Soil classification results from the Robertson (1990) and Schneider et al. (2008) charts generally agreed with the grain-size distribution and N SPT values.Considering the Brazilian experiment and results obtained from the study areas, the Robertson (1990) proposal confirmed its efficiency for soil classification from piezocone tests.The Schneider et al. (2008) proposal was adequate for use in Suape soils; however, it would seem appropriate to carry out further studies elsewhere in Brazil.
Compressibility parameters are discussed, including the development of a statistical correlation C c = 0.0097 vs. w (%) -approximately C c @ 0.01 w (%).It is also very important to determine the water content in each SPT test.
The horizontal coefficient of consolidation values with depths from the oedometer and piezocone tests for the study areas revealed a wide range of variations considering the O.C. and N.C.ranges.The c h values in the N.C.range obtained from piezocone tests were higher (around 10 times more) than the c h values obtained from oedometer tests.These results are in agreement with results from Recife soft clays, along with other soils reported in the literature.
This study confirmed the potential of piezocone tests for use in obtaining adequate predictions for geotechnical classification/consolidation parameters in soft soil deposits.
The values of the ratio S uvane /s' vo vs. PI for Suape clays in general fell between the curves for young and old clays proposed by Bjerrum (1973).For Recife and Suape clays, and other Brazilians clays, the points of the ratio S uvane /s' vm vs. PI fall between the correlations of Larsson (1980) and Mesri (1975), with satisfactory prediction by Coutinho (2007) proposal S u /s' vm = 0.11 + 0.0037 PI (modified from Skempton, 1957).It is possible to make a satisfactory prediction of S u from CPTU using proposals in the literature (N kt , N Du and N ke ).In Suape, satisfactory S u results were obtained considering N kt of 9-10, in general, with Brazilian clays in the range (9-12), including Recife.
The Mayne & Mitchell (1988) proposal for obtain OCR vane adapted for Suape clays (OCR lab = 0.65.OCR vane ) presented values with a good correlation with the OCR lab values in good quality samples.Kulhawy & Mayne's (1990) proposal to obtain OCR from CPTU adapted to Suape clays (OCR= 0.173Q t ) provided OCR CPTU values with a good correlation with OCR lab values.The corrected coefficient (0.173) was almost half the original coefficient (0.32), and it was smaller than the value found for Recife clays (0.23) and similar for Rio de Janeiro clay (0.153).
This study confirms that parameters can be obtained through in situ tests with correlations suited to local/regional experiment; and it is also important to have a joint laboratory and in situ test program.This prediction is fundamental for a proper geotechnical site characterization in research and engineering projects.
h values are also obtained from piezocone tests in the N.C.band estimated

UFaceFigure 11 -
Figure 11 -Results of pore pressure of dissipation tests, including the procedure to determine t 50% , for the CPTU 120, Sub Area C -AE -2.

Figure 14
Figure 14 shows results for c v and c h values obtained from oedometer and piezocone tests in the O. C. and N. C. ranges.The c h (OC range) was much higher (three times as much) in the piezocone test in Recife and Suape clays.Recife soft clays have slightly higher values.The coefficient of consolidation values obtained from oedometer and piezocone tests are complementary.The oedometer tests are essential for obtaining appropriate compressibility parameters.Results concerning the coefficients of consolidation obtained in the study areas are satis-

Figure 14 -
Figure 14 -Values for the c v , c h and c h /c v relation vs. log s' vc -Recife and Suape (from Coutinho & Bello, 2008).

Figure 19 -
Figure 19 -Nkt values obtained from field vane test: Suape clays study area AE-2 (SUB-AREA A and C).
C. and O.C. is the normally consolidated and overconsolidated range, respectively.Chandler (1988) collected data from vane tests of 19 clay deposits, enclosed normally consolidated and overconsolidated clays with OCR up to 7.5, obtaining m value equal to 0.95(Eq.11 et al. (1997) stated that methods to derive OCR from piezocone tests data fall into three main categories: (a) methods based on undrained shear strength; (b) methods based on the shape of the CPTU profile; and (c) methods based directly on piezocone tests data.
proposal is being adapted for Suape clays (Eq.11), considering both Sub-Areas A and C.

Table 1 -
Time factor in function of percentage of pore pressure dissipation