Geotechnical and Geological Engineering (2006) 24: 1259–1269 DOI 10.1007/s10706-005-1560-9

Ó Springer 2006

Liquefaction evaluation discrepancies in tropical lagoonal soils PETER G. NICHOLSON Associate Professor, and Graduate Chair, Department of Civil and Environmental Engineering, University of Hawaii at Manoa, 2540 Dole Street, Honolulu, Hawaii (e-mail: [email protected]) (Received 21 December 2004; accepted 28 July 2005) Abstract. Field penetration tests and shear wave velocity measurements are both established and accepted methods for evaluating liquefaction potential in soils. The results produced by the two methods are generally well correlated. However, recent studies have shown that when investigating tropical lagoonal deposits, the same accepted methods for evaluating liquefaction potential often produce significant discrepancies in results. This discrepancy is most apparent in saturated lagoonal deposits of calcareous gravelly sand (or sandy gravel), which tend to exhibit low penetration resistance values but relatively high shear wave velocities. These disparate test results can suggest different soil classifications under current building codes. Ambiguity in the code may allow for a potentially unconservative classification, which may in turn allow for the use and construction of less costly, lighter weight foundation systems than warranted. Equally as important, the potential for unconservative design as related to liquefaction appears to be high when shear wave velocity measurements are used as a basis for evaluation in these types of lagoonal deposits. Because of this, it is strongly recommended that caution should be excercised when determining seismic design parameters in these types of geologic environments. A hypothesis to explain the discrepancies in the results of evaluation methods and a suggested design protocol is proposed. Key words. building codes, calcareous soils, liquefaction, penetration tests, shear wave velocity.

1. Liquefaction evaluations Several methods exist for evaluating the susceptibility of a soil deposit to liquefaction under dynamic loading conditions. These include in situ field penetration testing and shear wave velocity measurements, as well as a variety of dynamic laboratory tests. While each type of evaluation method has advantages and disadvantages, these methods have been shown to provide acceptably reliable (and/or conservative) results individually and/or in conjunction with one another for a wide range of soil types. Liquefaction evaluation using field penetration test data is well established and is used worldwide. These methods are generally based on historical field penetration test data for both sites that have liquefied and those that have not liquefied in past earthquakes. Most commonly, penetration resistance values correspond to various levels of earthquake shaking represented as a function of cyclic stress ratio (CSR)

1260

PETER G. NICHOLSON

applied from the earthquakes. Through the use of widely published curves like the one shown in Figure 1, liquefaction potential can be evaluated based solely on the field test data obtained and the loads applied by a design earthquake. In looking at the liquefaction evaluation curve shown in Figure 1, it can be seen that the cyclic resistance to liquefaction is usually sufficiently high for corrected penetration resistance values (blow counts) above 15 (or greater than about 20 for high levels of cyclic loads). Similar curves have been developed using cone penetration test (CPT) data, as depicted in Figure 2. CPT tip resistance values greater than approximately qc1N = 100 (or qc1N = 150 for high levels of cyclic loads) are usually considered sufficiently high to represent adequate liquefaction resistance. More recently, the use

Figure 1. Liquefaction resistance based on SPT values based on field performance data after Seed et al. (1985), modified by NCEER (Youd and Idriss, 1997. Reprinted by permission from MCEER).

LIQUEFACTION EVALUATION DISCREPANCIES

1261

Figure 2. Liquefaction resistance criteria based on CPT field data (Youd and Idriss, 1997. Reprinted by permission from MCEER).

of in-situ shear wave velocity measurements has gained wide acceptance as being similarly reliable. Certain shear wave velocity limits have been proposed as cut-off values for liquefaction resistance (shown in Figure 3), again as a function of cyclic stress ratio. Classification of soil type under the building codes currently used in Hawaii (and practiced throughout much of the world) is based on either penetration resistance values (usually N1.60) or shear wave velocities (Vs1). According to Table 16-J from the Uniform Building Code (UBC, 1997) (shown in Table 1), soils with blow counts (N1.60) greater than 15 blows/0.3 m or shear wave velocities (Vs) greater than 180 m/s (600 ft/sec) fall into the category of soil type SD or better. These classifications refer to stiff or dense soils or better. For most soil types, these ‘cut-offÕ levels indicate adequate liquefaction resistance. For most soils, either penetration resistance or shear wave velocity evaluation would be acceptable and design values based on results of the different test methods would not be expected to vary greatly. However, current research indicates that this is not the case with some types of soil deposits. A recent study of liquefaction evaluation practices for calcareous soils has shown that different accepted test methods can provide very different and, in some cases, contradictory results, particularly for certain geologic regimes. Because the two testing methods can produce very different evaluations of liquefaction resistance

1262

PETER G. NICHOLSON

Figure 3. Limiting shear wave velocities for liquefaction resistance after Andrus and Stokoe, 2000 (Youd and Idriss, 1997. Reprinted by permission from MCEER).

for certain types of tropical soils described herein, it is not surprising that the two testing methods can also result in disparate soil classifications. However, the code does not require that different methods of evaluations need to be used or that the results of those evaluation methods should be compared. This has led some engineers to choose to use evaluation results and soil classifications based on Vs1 values, even when penetration test data may suggest classifications as lower capacity soils with significantly lower liquefaction resistance. This result may lead to unconservative designs and an analysis of liquefaction hazard potential that is lower than may actually exist.

1263

LIQUEFACTION EVALUATION DISCREPANCIES Table 1. UBC, 1997 – Table 16-J

Soil profile type

Soil profile name generic description

SPT, N1 (blows/0.3 m)

Shear wave velocity, Vs (m/s)

Undrained shear strength, Su (Pa)

SA SB SC SD SE SF

Hard rock Rock Very dense soil and soft rock Stiff soil profile Soft soil profile Soil requiring site-specific evaluation*

– – >50 15 to 50 <15

>1525 760–1525 365–760 180–365 <180

– – >100 50 to 100 <50

2. Previous studies on calcareous soils There have been some studies conducted to evaluate differences between calcareous soils and their siliceous counterparts using both cyclic resistance testing in the laboratory (Ross and Nicholson, 1995; Flynn, 1997) and penetration resistance (Morioka and Nicholson, 1999; McLemore, 2002). Similarly, seismic cone testing has also produced significantly different shear wave velocities for calcareous and siliceous soils (Campanella, personal communications, 2001). Despite the differences in test results, it has been suggested that a reasonable evaluation of the liquefaction resistance of calcareous soils can be made based on either of these testing methods so long as consideration of the differences is accounted for through adjustments to established liquefaction resistance correlations. There has been limited research and published findings regarding the cyclic behavior of calcareous soils and there continues to be some debate as to the actual liquefaction potential of these types of calcareous deposits. However, cases of liquefaction in calcareous soils induced by earthquakes as well as remedial construction and field blasting have previously been reported (Mejia and Yeung, 1995; Morioka personal communications, 2001; Rollins et al., 2004).

3. Field test cases A number of sites on Oahu and Maui with a variety of tropical soils were explored using both penetration resistance (SPT and/or CPT) and shear wave velocity measurements. The sites investigated contained both layers of relatively uniform sands as well as highly variable deposits, including gravelly lagoonal deposits and buried reef formations. Some locations contained both uniform and non-uniform deposits. Some of the penetration test data used was taken from previous site investigation reports, but was deemed reliable with knowledge of the local drilling operators with whose work the author is familiar. For the most part, shear wave measurements were made using seismic cone penetrometers. The approximate locations of these sites are depicted in Figure 4. Brief logs and pertinent test data are provided in Table 2.

1264

PETER G. NICHOLSON

Figure 4. Map of testing sites on Oahu and Maui.

An example of the grain size distributions typical for the two disparate soil types compared in this study (uniform and non-uniform sands) are provided in Figure 5. Photographs of these soil types are depicted in Figure 6. The uniform sands represent sediments typically deposited as shoreline beach sands deposited by wave action. The non-uniform sands, which contain significant coarse and gravelly grain sizes, represent the ‘lagoonal’ sediments, which are believed to have been deposited

1265

LIQUEFACTION EVALUATION DISCREPANCIES Table 2. Data from selected test sites on Oahu and Maui

Depth (ft) Soil description

Case study – Honolulu–Ward 1–5 Med. dense, gravel/sand (coral fill) 5–20 Loose, silty-sand w/gravel (lagoonal) 25–30 Dense, gravel/sand 30–38 Med to dense, gravel/sand Case study – Keehi 1–10 Dense, gravel/sand fill 10–60 Loose, silty-sand w/gravel (lagoonal) Case Study – Kailua 0–14 Med. dense, fine sand 14–20 Med. dense, coarse sand 20–35 Loose, silty-sand w/gravel (lagoonal) Case Study – Ala Wai 0–7 Med. dense sand/gravel 0–14 Loose silty sand/gravel (lagoonal) Case Study – Punaluu 0–10 Med. dense sand w/gravel (fill & beach sand) 10–22 Loose silty sand/gravel (lagoonal) 22–66 V. loose silty sand/gravel Case Study – Olowalu, Maui 0–5 Silty clay (topsoil) 5–12 Med-fine sand 12–25 Med. coarse sand w/gravel & shells (lagoonal)

CPT, Shear wave Correlation SPT, qc (MPa) N1 velocity, Vs1 (m/s) (blows/0.3 m) 10–20

50–150

2–12

3–30+

16–60 16–48

N/A N/A

150–300 Poor (avg 240) 65–200 550–760 Fair/Good 10–50 425–550 Fair

20–25 2–6

5–40 6–20

170 Fair 200–300 Poor

15–30 4–13 2–3

85–160 20–40 6–20 (28)

2–30 1–6

N/A N/A

20–50

40–140

300–450 Fair/Good

5–12

10–50

120–150 Fair/Good

2–5

8–23

140–155 Fair/Good

16 6–12 3–11

N/A N/A N/A

120–150 Fair 150–190 Fair/Good 190–280 Poor

240–300 Fair/Poor 250–300 Poor

N/A N/A N/A N/A 220 Poor

in a back reef environment. These back reef lagoons were low energy environments where coarse fragments of coral reef deposited from the sea side became mixed with finer grained material that settled gently through the calm waters of the lagoons. More detailed investigations of these types of deposits have indicated that some gravel sized coral may also have grown in these low energy environments along with deposition of the loosely deposited sediments. Analysis of the data for sites where relatively uniform sands were encountered (e.g., Kailua 0–20 ft and Punaluu 0–10 ft), show that there is relatively good agreement between measurements indicative of density (or test values used as indicators of liquefaction susceptibility) from both penetration tests and shear wave

1266

PETER G. NICHOLSON

Grain Size Distributions 100 Lagoonal Soil A

90

Lagoonal Soil B

80

Uniform Sand A

% Passing

70

Uniform Sand B

60 50 40 30 20 10 0 100

10

1

0.1

0.01

Garin Size (mm) Figure 5. Examples of typical gradation for uniform and non-uniform sands sampled.

velocity measurements. Column 6 of Table 2 indicates the correlation (agreement) of liquefaction evaluations made using the data from the different types of tests. In contrast, the data for sites where lagoonal deposits were found (e.g., Kailua 20–35 ft, Honolulu–Ward 5–20 ft, Keehi 10–60 ft and Olowalu 12–25 ft) showed significant discrepancies between the relative density indices and resulting evaluation of liquefaction susceptibility depending on the testing/measurement method employed. In each of these locations, corrected blow counts were less than 12 (in all but one case, in low single digits) while shear wave velocities were consistently above 200 m/s. For each of these sites would be classified as either ‘safe’ or highly susceptible, depending on the evaluation method employed. For the design engineer, this poses a real dilemma. Furthermore, the corroborating data from a number of different sites where these types of soils are found all follow the same trend. Thus, the discrepancies in testing results from the four sites are not an exception, but rather are a recurring and potentially dangerous flaw in the codes used in current practice.

4. Suggested mechanism for field test discrepancies – skeletal interconnectivity After evaluating many borings in lagoonal sediments and having examined a number of grouted lagoonal deposits, the author has developed a hypothesis that suggests an explanation for the large discrepancies between liquefaction evaluations, or more directly, the density indices derived from the results of various field tests described. These deposits were typically formed in the relatively calm lagoons found behind (landward of) shallow coral reefs. The materials deposited in these lagoons consisted

LIQUEFACTION EVALUATION DISCREPANCIES

1267

Figure 6. Photographs of typical uniform and non-uniform sands sampled.

of a mixture of coralline detritus (typically gravel sized fragments broken from the pre-existing coral reef) and finer-grained silt and sand-sized calcareous sediments. In the quiet lagoonal setting, these materials were deposited in a very loose configuration analogous to pluviation through water. While the soils may appear to be

1268

PETER G. NICHOLSON

predominantly gravel when inspected visually, grain size analyses usually show the gradations to be gravelly sand with non-plastic silt fines. As previously noted, gravelly lagoonal soils tend to exhibit relatively high shear wave velocities while showing very low penetration resistance values, in contrast to most categories of soils with similar gradation and USCS classification. A theoretical explanation is that even in a very loose condition the gravel particles could be in contact, forming a continuous skeleton supported by a matrix of saturated silty sand. In this condition, shear waves could travel through the interconnected gravel skeleton at velocities more representative of a medium dense deposit.

5. Proposed protocol When different evaluation methods produce such disparate results, the current building codes become ambiguous. In these types of soils, it may be prudent to require that if shear wave velocities are used to determine the code soil classification, they should also be compared to classifications resulting from field penetration values. An additional question arises when hard intact reef formations are encountered within a subsurface profile. A pleasant surprise was that shear wave velocity measurements made by seismic cone were found to be clear and well defined even at significant depths below and between reef formations. Shear wave velocity profiles correlated well with boring log interpretations in this type of environment. The derivation of the soil classification to be used according to building code specifications then becomes an issue. One interpretation is that the code allows for reporting and using an ‘average’ soil type for any site location rather than a more conservative lower bound. Obviously, the hard reef layers will significantly increase the average shear wave velocity of a deposit (as well as increase average penetration resistance). It is suggested that for these cases, where a significant layer(s) of low quality soils exist, a site-specific analysis should be required.

6. Conclusions The results of these studies clearly show a problem with the use of alternative accepted methodologies in determining liquefaction potential in certain geologic regimes. In particular, where loose, gravelly, calcareous, lagoonal sediments are found, shear wave velocity testing may provide unconservatively high values while penetration resistance tests show soil conditions susceptible to liquefaction. However, the building code allows the use of the less conservative methodology for classification of soils even when there is poor agreement between alternative evaluation methodologies. The resulting soil classification may allow the use of lesser design values, which can result in considerable savings for structural foundations, but at the same time may be introducing considerable risk. This should be a concern to building owners, planning officials, insurers and ultimately to the design professionals and should be addressed in future codes and/or amendments.

LIQUEFACTION EVALUATION DISCREPANCIES

1269

Acknowledgements Drilling for the testing performed as part of this study was provided by Geolabs Hawaii, Inc. and Ernest K. Hirata & Assoc., Inc. and this support is gratefully acknowledged. The shear wave velocity testing was performed by R.G. ‘‘Dick’’ Campanella, Professor Emeritus at University of British Columbia (Oahu) and James Bay of Utah State University (Maui) and this valuable contribution is appreciated.

References Andrus, R.D. and Stokoe, K.H. II (2000) Liquefaction resistance of soils from shearwave velocity, Journal of Geotechnical and Geoenvironment Engineering, ASCE, 126(11), 1015–1025. Flynn, W.T. (1997) A Comparative Study of Cyclic Loading Responses and Effects of Cementation on Liquefaction Potential of Calcareous and Silica Sands, M.S. Thesis, University of Hawaii, Manoa. McLemore, T.B. (2002) Cementation Processes of Naturally Aged Hawaiian Calcareous Sands, M.S. Thesis, University of Hawaii, Manoa. McLemore T.B. and Nicholson, P.G. (2004) Effects of Light Cementation on Static and Cyclic Loading Response of Calcareous Soils, Research Report No. UHM/CEE/Report No. 94–08. Mejia, L.H. and Yeung, M.R. (1995) Liquefaction of coralline soils during the 1993 Guam Earthquake, In: Earthquake-Induced Movements and Seismic Remediation of Existing Foundations and Abutments, Geot. Spec. Pub. No. 55, ASCE, 33–48. Morioka, B.T. and Nicholson, P.G. (1999) Evaluation of the Static and Cyclic Properties of Calcareous Sands in a Calibration Chamber Study, Research Report to Hawaii Department of Transportation and Federal Highways Administration. Morioka, B.T. and Nicholson, P.G. (2000) Evaluation of the Liquefaction Potential of Calcareous Sand, Proceedings 10th International Offshore and Polar Engineering Conference, Seattle, WA, May 2000. Nicholson, P.G. and Morioka, B.T. (1999) Liquefaction Potential of Calcareous Sand Using Cone Penetrometer Testing, Transportation Research Board, National Research Council, Washington, D.C. Rollins, K.M., Nicholson, P.G., Lane, J.D. and Rollins, R.E. (2004), Liquefaction Hazard Assessment Using Controlled-Blasting Techniques, In: Proceedings 3rd Intl. Conference on Earthquake Geotechnical Engineering, Balkema. Ross, M. and Nicholson, P.G. (1995) Liquefaction Potential and Cyclic Loading Response of Calcareous Soils, Research Report No. UHM/CE/95–05. Uniform Building Code, (1997) Prentice Hall, Inc. Youd, T.L. and Idriss, I.M. (1997) In: Proceedings of the NCEER Workshop on Evaluation of Liquefaction Resistance of Soils, Technical Report NCEER 97–0022.