Impact of soaking–drying cycles on gypsum sand roadbed soil

Transportation Geotechnics 2 (2015) 78–85

Contents lists available at ScienceDirect

Transportation Geotechnics journal homepage: www.elsevier.com/locate/trgeo

Impact of soaking–drying cycles on gypsum sand roadbed soil Sabah Said Razouki a,⇑, Bushra M. Salem b a b

Nahrain University, Baghdad, Iraq Columbus State Community College, OH, USA

a r t i c l e

i n f o

Article history: Received 4 July 2014 Revised 4 November 2014 Accepted 14 November 2014 Available online 25 November 2014 Keywords: California Bearing Ratio Cyclic soaking and drying tests Geotechnical engineering Gypsum sand Pavement design Roads and highways

a b s t r a c t A thorough laboratory investigation is carried out to study the effects on strength and deformation of cyclic soaking and drying tests on gypsum rich sand used for the construction of roadbeds. The changes in the properties were assessed by the use of California Bearing Ratio (CBR) tests. Each cycle consisted of 90 days of soaking followed by 90 days of drying at room temperature giving a cycle length of 180 days. The soil tested was poorly graded sand with gravel, (SP) soil according to the Unified Soil Classification System and A-1-b soil according to American Association of State Highway and Transportation Officials (AASHTO) Soil Classification System. It had a gypsum content of about 39%. Eleven pairs of CBR samples were prepared at the optimum moisture content and at 95% of the maximum dry density for modified AASHTO compaction, and subjected to a surcharge load of 45 lb (200 N) during soaking and drying. The CBR load penetration test was carried out at the end of each soaking and each drying phase of each cycle of the five cycles studied. The paper reveals that the CBR increased during drying and decreased during soaking. The CBR value decreased with increasing number of cycles reaching equilibrium at the end of the fifth cycle indicating that the soaked equilibrium CBR is about 83% of that for the commonly used four days soaking. During the first soaking phase, the soil swelled for the first three days and then underwent settlement that continued at a slower rate during the next drying phase. Thereafter, swelling during soaking and settlement during drying took place for the second and third cycle but led to an equilibrium condition after the third cycle. Ó 2014 Elsevier Ltd. All rights reserved.

Introduction The widespread use of gypsiferous sands as roadbed soils and as embankment fill in the Middle East, especially, Iraq (Razouki et al., 2008, 2011; Razouki and El-Janabi, 1999; Tomlinson and Boorman, 1996; Fookes, 1978) necessitated research into gypsiferous soils. Recent research on gypsum sand has focused attention on the development

⇑ Corresponding author. E-mail address: [email protected] (S.S. Razouki). http://dx.doi.org/10.1016/j.trgeo.2014.11.003 2214-3912/Ó 2014 Elsevier Ltd. All rights reserved.

of experimental techniques to provide more accurate characterization of such compacted soils as construction materials (Razouki et al., 2008, 2011). However, geotechnical engineers should be aware of the importance of saturation state in governing the strength and deformation properties of compacted gypsiferous fills and gypsiferous roadbed soils both immediately after construction as well as during its service life. The important effect of moisture content on both the strength and stiffness of gypsum-rich roadbed soil was reported by Razouki and Al-Azawi (2003). Their laboratory tests were carried out on well graded sand with silt

S.S. Razouki, B.M. Salem / Transportation Geotechnics 2 (2015) 78–85

79

Notation AASHTO American Association of State Highway and Transportation Officials ASTM American Society for Testing and Materials CBR California bearing ratio

(SW-SM), according to ASTM D2487-93, having a gypsum content of about 34%. Razouki and Al-Azawi (2003) pointed out that the laboratory long-term soaking tests for CBR soil specimens revealed a marked drop in both CBR and MR (resilient modulus) with soaking period. This decrease in strength and stiffness took place at a high rate within the first week and at a decreased rate thereafter so that the soil strength and stiffness became almost constant after about six months of continuous soaking in fresh water. In addition, the tests revealed that the soil swelled initially then it started to settle and the settlement process continued at a slow rate even after 180 days soaking. Unfortunately, the majority of previous studies on gypsiferous soils (Razouki and El-Janabi, 1999; Razouki and Kuttah, 2004a, 2006) were concerned with continuous soaking without drying cycles. However, roadbed soils in both cut and fill sections of roads are subject to changing environmental conditions during their service life. Such seasonal variations are reflected primarily by moisture content and temperature changes. Accordingly, in this study, samples are to be subjected to cyclic soaking and drying. This research is a part of a wider research aiming at a thorough understanding of the behavior of gypsiferous soils of wide occurrence in the Middle East especially Iraq, where the routes of various transportation systems pass through gypsiferous terrains. In addition, this paper aims at focusing attention on the evaluation of moisture sensitivity, risk of settlement, the proper CBR-value for pavement design and the need for introducing special specifications for such soils. The test methodology of this research program can be summarized as follows: From a region in Iraq with gypsum rich sand, sufficient amount of the soil is to be obtained for testing in the laboratory. For studying the strength and deformation behavior of this gypsum rich roadbed sand in the laboratory, the CBR (California Bearing Ratio) test was adopted. To study the effect of cyclic soaking and drying, a cycle length of 180 days was considered suitable for a Ph.D. research for about 2–3 years. Such a cycle length will allow sufficient cycles to be obtained for the study within the available research period. Although the CBR method is not a mechanistic method to estimate strength/stiffness characteristics of soils, but it has local/regional importance as reported by Razouki and Salem (2014), Razouki et al. (2014, 2008), Razouki and Kuttah (2006) and Huang (2003). The CBR soil specimen is sufficiently large to represent the gypsum-rich sand of this study. In addition, the repeated load triaxial testing machine for soil stiffness, which is both expensive and complex, was not available at the University of Technology

CBRt CBRun MR

California bearing ratio at any time t California bearing ratio for unsoaked condition Resilient modulus

in Baghdad at the time of this study. Therefore, the CBR testing method was considered suitable especially for Iraq and all developing countries of the Middle East. Properties of soil tested The soil under study is a gypsum sand obtained from a region near Baghdad, Iraq. In accordance with the Earth Manual of US Department of Interior (1980), the total soluble salt (TSS) content in the soil tested was determined as 42.6% at a soil:water ratio of 1:300. The gypsum content in the soil tested was 38.8% according to BSI (1990). Thus, the gypsum content is about 91% of TSS indicating that nearly all of the total soluble salts in the soil tested are mainly gypsum. In addition, X-ray diffraction analysis was carried out on the soil tested indicating that the components of the soil under study are quartz, gypsum, calcite, dolomite and feldspar. The particle size distribution for the soil tested, according to AASHTO (1993), was determined by Razouki and Salem (2014) indicating that the percentage passing no. 200 sieve is 3.65%. The liquid limit of the soil could not be determined and the soil is accordingly non-plastic. Using the British Standard density bottle method (BSI, 1990), with white spirit instead of water, the specific gravity of the gypsum sand studied was 2.47. According to the Unified Soil Classification System (ASTM, 1993, D 2487-93), the soil tested was found to be poorly graded sand with gravel (SP) and according to American Association of State Highway and Transportation Officials (AASHTO, 1986) soil classification system, the soil belongs to A-1-b soil group. The compaction curve of the soil, for each of standard and modified AASHTO compaction test (AASHTO, 1986), was obtained using six pairs of soil samples as shown in Fig. 1. Note that the maximum dry unit weight of the modified AASHTO compaction was 18.27 kN/m3 taking place at the optimum moisture content of 10.8%. It is necessary to note that regarding the moisture content determination, Horta (1989) reported that soil specimens containing gypsum should be dried to a constant weight at a temperature lower than 60 °C and preferably 40 °C. Razouki et al. (2008) recommended strongly this method of Horta (1989) and, therefore, it was adopted in this work. Preparation and testing of soil samples To study the effect of changing environmental conditions on the strength and deformation characteristics of gypsum sand subgrade soil, 11 pairs of California Bearing

80

S.S. Razouki, B.M. Salem / Transportation Geotechnics 2 (2015) 78–85

23 Gs = 2.47

22 modified AASHTO compaction curve

Dry unit weight (kN/m3)

21

Standard AASHTO compaction curve

20 19 18.27

18

0%

17

16.8

air

voi ds cur ve

16 5%

15

10 %

14 15 % OMC 10.8 %

13

OMC 12.8 %

20 %

12 2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22

moisture content (%) Fig. 1. Moisture–density relation of tested gypsum sand.

Ratio (CBR) soil samples were prepared according to ASTM (2007) D 1883. The samples were prepared at optimum moisture content of 10.8% and at 95% of the maximum dry density of the modified AASHTO compaction (29 blows/layer in CBR mold) and subjected to the same surcharge load of 45 lb (200 N) during soaking and drying. This surcharge load represents a pavement structure of about 57 cm total thickness above the roadbed soil (ASTM, 2007, D 1883). One pair of CBR samples was tested for unsoaked conditions, while the other ten pairs were soaked in soaking tanks, each containing generally 90 L of tap water for two CBR specimens. It is worth mentioning that the use of tap water was recommended by Ismael and Mollah (1998) due to its similarity to ground water in the field. Razouki and Kuttah (2006) and Razouki et al. (2008) made use of tap water for soaking too. This encouraged the use of tap water in this study. To avoid any possibility of saturation of soaking water with gypsum, the water in soaking tank was circulated at a rate of 6.5 L/day. After 90 days of soaking, the first soaked pair was subjected to the CBR load-penetration test, while nine of the other pairs were taken off from the water to be dried for 90 days. One of these nine pairs was tested at the end of the drying phase of the first cycle, while the others (eight pairs) were soaked again for another 90 days. This cyclic process continued so that the last pair was tested after 30 months (2.5 years for five cycles of soaking and drying) from the date of its compaction. It is worth mentioning that for the soaking phase of any cycle, the CBR samples in molds were soaked for 90 days in the mentioned tanks placed in the laboratory. Regarding the drying phase, the CBR samples in molds were taken

out from the tanks and placed in open air over a desk in the laboratory for 90 days for drying. The average temperature in the laboratory (average of three temperature readings daily) fluctuated over the year between about 17 °C and 35 °C.

California Bearing Ratio test results Due to the fact that one pair of CBR specimens was tested at each 90 days interval, the representative CBR value was obtained as the mean value of those two (the difference between them did not exceed 2%) of that pair and used throughout the whole paper. The results of the CBR values obtained for each phase of each of the five cycles studied are shown in Fig. 2. It is quite obvious from this figure that there is a decrease in CBR after each soaking and increases in CBR after each drying, but a decrease in CBR with progressing cycles. In terms of dimensionless ratios, Fig. 3 shows the effect of cyclic soaking and drying on the strength ratio CBRt/ CBRun and CBRt/CBR4 (where CBRt refers to the CBR at any time t, while CBR4 corresponds to four days soaking duration and CBRun refers to the CBR for unsoaked conditions) for 180 days cycle length. It appears from these figures that the CBR is reaching equilibrium at the end of the fifth cycle due to a decreasing rate of change and that the expected equilibrium soaked CBR is 75% of CBRun and 83% of CBR4. Thus, the maximum decrease in soaked CBR, due to soaking–drying cycles, becomes 17% of CBR4. This indicates that the use of the common four days of soaking (ASTM D1883-87) overestimates

81

S.S. Razouki, B.M. Salem / Transportation Geotechnics 2 (2015) 78–85

56 54 52

CBR (%)

50 48 46 44 42 40

First cycle

Second cycle

Third cycle

270

450

Fourth cycle

Fifth cycle

38 0

90

180

360

540

630

720

810

900

Time (days) Fig. 2. CBR versus time for cyclic soaking and drying of gypsum sand sample tested.

1.2

1.2 CBRt/CBRun

1.1

CBRt/CBR4

1.0

1.0

0.9

0.9

0.8

0.8

0.7

0.7 First cycle

Second cycle

Third cycle

270

450

CBRt / CBR4

CBRt / CBRun

1.1

Fifth cycle

Fourth cycle

0.6

0.6 0

90

180

360

540

630

720

810

900

Time (days) Fig. 3. Effect of cyclic soaking and drying on the strength ratios CBRt/CBRun and CBRt/CBR4 for tested gypsum sand samples.

the strength of the soil as cyclic soaking and drying reduces the soil strength. Thus, a gypsum-rich roadbed soil supporting a new pavement structure reaches an equilibrium CBR at about the end of the fifth cycle of cyclic soaking and drying. The required actual time period to achieve this equilibrium CBR depends upon the frequency or cycle length of the soaking–drying process. For a short cycle length (high frequency), this time (in days) is shorter than that for a long cycle length (low frequency). For example, for the 180 days cycle length treated in this work, about 900 days are required to achieve the equilibrium CBR. To compare the results of this work with published results, it was difficult to find any published paper related to this study except that by Razouki and Salem (2014). However, Razouki and Salem (2014) carried out their laboratory experimental work on the same soil of this work but

they restricted their study to three cycle lengths of cyclic soaking and drying, namely 8, 14 and 60 days only. Thus, it was decided to compare the results of the present work, dealing with 180 days cycle length, with Razouki and Salem’s results for the case of 60 days cycle length. According to Razouki and Salem (2014), the soaked equilibrium CBR (for 60 days cycle length) was 75.4% of unsoaked CBR and 83.4% of CBR4, while in the present work the soaked equilibrium CBR was 75% of unsoaked CBR and 83% of CBR4 indicating good agreement. Effect of cyclic soaking and drying on deformation of gypsum sand The study of gypsum-rich soil deformation is rather of great importance. Serious settlement and differential settlement problems of different structures founded on or in

82

S.S. Razouki, B.M. Salem / Transportation Geotechnics 2 (2015) 78–85

4 2 0 -2 -4 -6 -8 -10 -12 -14 -16 -18 -20 -22

under cyclic soaking and drying is about 0.2 mm or 0.16% of the initial sample height which is insignificant when compared with the acceptable value of about 3% as reported by Head (1982), AASHTO (1986) and ASTM (2007). Note that these results relate to the gypsum sand tested and such results are typical of roadbed materials in gypsiferous regions in Iraq. With reference to Fig. 4, there is a very limited post compaction of the tested soil leading to a denser material as time elapses. However, the final very low settlement measured, is mainly due to loss of soluble salts in the soil (Day, 2000) and not to the effect of the surcharge load of 45 lb (200 N). This load introduces very low pressure on a CBR soil specimen of 600 (15.24 cm) diameter. Thus, the possible slight increase in CBR due to the more compact state of the soil is overcome to a great extent by the increased dissolution of the cementing agent, gypsum, with soaking time. This in turn decreases the cohesion in the soil and hence the soil strength (CBR values) decreases with soaking time. Effect of number of cycles on time variation of moisture content In order to give an idea about the effect of cyclic soaking and drying on the time variation of moisture content at each of top, middle and bottom of CBR specimens, samples for moisture content determination were obtained from these locations after performing the load–penetration test. For the determination of moisture content, the method of Horta (1985) was adopted namely drying the soil sample at 40 °C for three to four days until it reached a constant mass. Fig. 6 shows the time variation of moisture content at the top, middle and bottom of the sample. It is obvious from this figure that the moisture content increased after each soaking and decreased after each drying and that the moisture content at the bottom is greater than that at the top while there is little fluctuation at the middle of the specimen. It is worth noting that the moisture content

0.02

swelling

0.00 settlement

-0.02 -0.04 -0.06 -0.08 -0.10 -0.12

Deformation S (%)

Vertical displacement * 0.01 (mm)

such soils have been reported by various authors (James and Lumpton, 1978; James and Kirkpatrik, 1980; Hawkins and Pinches, 1987; Razouki et al., 1994; Cooper and Saunders, 2002; Kota et al., 2007 and others). To study the effect of cyclic soaking and drying on the deformation (swelling and/or settlement) of the gypsum sand under study, a 0.01 mm dial gauge was used. This gauge was placed on the top of the stem of the swelling plate of CBR test to measure the vertical deformation of the CBR specimen during soaking and drying. Excepting the first soaking phase of the first cycle, only one measurement of deformation was made at each of the begin and end of the soaking phase of each cycle. However, due to rapid changes in deformation during the first soaking phase of the first cycle, seven measurements for swelling and settlement were taken for this phase. Fig. 4 shows the time variation of deformation during cyclic soaking and drying. It is clear from this figure that there is a very small amount of swelling at the first three days of the first soaking phase (see Fig. 5), thereafter the soil starts to settle gradually. This is expected as the gypsiferous soil under study is a granular soil different from the clayey gypsiferous soil with significant swelling treated by Razouki and Kuttah (2004b). At the start of the soaking process, the soil tries to suck water and swell. The swelling becomes maximum at the third day. It appears that the effect of dissolution of gypsum starts to be obvious after three days so that settlement starts to take place at a significant rate. However, during the drying phase of the first cycle, the settlement continued but at a slower rate due to the softening and/or dissolution of gypsum. In the soaking phase of the second cycle, the suction of water takes place again causing reduction in settlement (swelling). This process of oscillation in settlement behavior continues till the end of the third cycle. Thereafter no further movement is observed. It should be noted that the maximum observed value of settlement (the equilibrium deformation for the 180 days cycle length) of gypsiferous soil samples in CBR molds

-0.14 -0.16 90

0

270

180

450

360

630

540

810

720

900

Time (days) Fig. 4. Effect of cyclic soaking and drying on the time variation of swelling/settlement of CBR samples of gypsum sand tested.

83

Vertical displacement * 0.01 (mm)

S.S. Razouki, B.M. Salem / Transportation Geotechnics 2 (2015) 78–85

4 2

swelling 0 -2

settlement

-4 -6 -8 -10 -12 0

10

20

30

40

50

60

70

80

90

Time (days) Fig. 5. Enlarged image for the deformation during the first 90 days of soaking.

17 top

16

middle

Moisture content (%)

bottom

15 14 13 12 11 Fourth cycle

Third cycle

Second cycle

First cycle

Fifth cycle

10 0

90

180

270

360

450

540

630

720

810

900

Time (days) Fig. 6. Effect of cyclic soaking and drying on the moisture content at the top, middle and bottom of CBR soil specimens of tested gypsum sand.

continued increasing with progressing cycles reaching equilibrium at the end of the fifth cycle but with different values at the top, middle and bottom of the sample. Note that the moisture content values, even in a drying state, are pretty high compared to OMC value due to the fact that the top surface of any CBR specimen in the drying phase, is covered by the swelling plate and surcharge load. This cover decelerated the evaporation process from the specimens. Effect of changing environmental conditions on distribution of gypsum The distribution of gypsum content along the CBR soil sample for the first cycle is shown in Fig. 7. It is obvious

from this figure that at the end of the soaking phase of the first cycle, the gypsum content decreased at all depths due to dissolution of gypsum with soaking. However, after taking the CBR sample in mold out of soaking tank, it was placed in open air over a desk in the laboratory for drying for the 90 days drying phase. The gypsum content started to increase at the top and to decrease at the bottom of the CBR specimen due to the upward migration of gypsum that was encouraged by the evaporation process at the top of the specimen. The process of increased dissolution of gypsum at the bottom and increased migration of gypsum to the top of the CBR specimen is in full agreement with Blight (1976). This fact is obvious from Fig. 8 showing the distribution of gypsum at the end of the wet and dry phases of the fifth cycle.

84

S.S. Razouki, B.M. Salem / Transportation Geotechnics 2 (2015) 78–85

Top First cycle

In itia l g y p s u m c o n te n t = 3 8 .7 5 %

Depth ratio below top of CBR sample

0.00

0.25 90 days 180 days

0.50

end of dry phase

0.75

end of wet phase

Middle

Bottom

1.00 22

24

26

28

30

32

34

36

38

40

42

Gypsum content (%) Fig. 7. Distribution of gypsum content along CBR samples at the end of the wet and dry phases of the first cycle of soaking and drying of tested gypsum sand.

Top

0.00

0.25

720 days

Initial gypsum content = 38.75%

Depth ratio below top of CBR sample

Fifth cycle

810 days 900 days

0.50 end of dry phase end of wet phase

0.75

Middle

1.00

Bottom

22

24

26

28

30

32

34

36

38

40

42

Gypsum content (%) Fig. 8. Distribution of gypsum content along CBR samples at the end of the wet and dry phases of the fifth cycle of soaking and drying of tested gypsum sand.

The need to broaden the study It is quite obvious from the above study that it is interesting and necessary to broaden the present research in the future to other soils especially to fine-grained gypsum-rich soils. In addition, it is also strongly recommended to study

the influence of such characteristics as grading, percentage of gypsum and permeability on the behavior of such soils including both gypsum sand and fine-grained gypsiferous soils. Finally, it is recommended to apply other more fundamental laboratory methods available to study the

S.S. Razouki, B.M. Salem / Transportation Geotechnics 2 (2015) 78–85

characteristics of gypsum-rich soil and compare their results with those of the present study. At the same time it is interesting to study the effect of cyclic soaking and drying on the cohesion, friction angle and resilient modulus of gypsum sand subjected to cyclic soaking and drying.

Conclusions and recommendations For the case of cyclic soaking and drying of 180 days cycle length, the main conclusions of this work can be summarized as follows: (a) There is a decrease in California Bearing Ratio (CBR) after each soaking phase and an increase after each drying phase. The increase in CBR due to drying is less than the corresponding decrease due to soaking for the same cycle. (b) The maximum decrease in soaked CBR, due to soaking–drying cycles, was about 17% of CBR4 indicating that the use of the four days soaking overestimates the soil strength of the roadbed soil. (c) The moisture content at the top of the CBR samples is less than that at the bottom for each phase of each cycle. (d) The moisture content at the middle of the CBR sample after soaking and/or drying was the least as compared with the top and bottom. (e) The gypsum concentrates at the top of CBR soil samples in drying phases under the effect of migration due to evaporation. At the bottom of the samples, the gypsum content decreases due to progressive dissolution of gypsum. (f) There is little amount of swelling at the first three days of the first soaking phase. Thereafter, the soil starts to settle gradually. For the second and third cycles, slight decrease in settlement due to swelling during soaking followed by slight increase in settlement during drying, takes place. (g) The equilibrium settlement takes place almost at the end of the third cycle. (h) The issue of special specifications for the laboratory CBR test for gypsum-rich soils is strongly recommended. The specification should include, among others, a proper method of determining the moisture content and an adequate long-term soaking period for such soils. The introduction of cyclic soaking and drying with appropriate cycle length, for the site of interest, is highly recommended. (i) The pavement and geotechnical engineers should ensure the adequate compaction of gypsum-rich roadbed soil to avoid serious settlement and differential settlement and even soil collapse. (j) It is highly recommended to carry out similar studies but on fine-grained gypsum-rich soils subjected to cyclic soaking and drying. The study of the influence of such characteristics as particle size distribution, percentage of gypsum and permeability on the behavior of such soils is recommended too.

85

References AASHTO (American Association of State Highway and Transportation Officials). Standard Specification for Transportation Materials and Methods of Sampling and Testing. Part II, 14th ed. Washington, DC: AASHTO; 1986. AASHTO (American Association of State Highway and Transportation Officials) D T88–93: Particle Size Analysis of Soils. Standard Specifications for Transportation Materials and Methods of Sampling and Testing. 16th ed. Washington, DC: AASHTO; 1993. ASTM (American Society for Testing and Materials). D1883: standard test method for CBR (California Bearing Ratio) of laboratory-compacted soils. West Conshohocken, PA, USA: ASTM International; 2007. ASTM (American Society for Testing and Materials). D 2487–93: ASTM standard classification of soils for engineering purposes (Unified Soil Classification System). West Conshohocken, PA, USA: ASTM; 1993. Blight GE. Migration of subgrade salts damages thin pavements. Transp Eng J ASCE 1976;102(4):779–91. BSI (British Standard Institution). Methods of test for soils for civil engineering purposes—part 3. Chemical tests. Milton Keynes, BS: BSI; 1990. p. 1377. Cooper AH, Saunders JM. Road and bridge construction across gypsum karst in England. Eng Geol 2002;65(2–3):217–23. Day RW. Discussion on leaching effects on properties of cemented sands in kuwait. J Geotech Geoenviron Eng ASCE 2000;124(1):99–101. Fookes PC. Middle east-inherent ground problems. Quart J Eng Geol 1978;11(1):33–49. Hawkins AB, Pinches CM. Expansion due to gypsum growth. Sixth International Conference on Expansive Soils, New Delhi, India; 1987. p. 183–7. Head KH. Manual of soil laboratory testing 1982;vol. 2. London: Pentech Press Limited; 1982. Horta JC. Salt heaving in the Saharra. Geotechnique 1985;35(3):329–37. Horta JC. Carbonate and gypsum soils properties and classification. In: Proceedings of the 12th international conference on soil mechanics and foundation engineering, Rio de Janeiro, vol. 1; 1989. p. 53–6. Huang YH. Pavement analysis and design. 2nd ed. Englewood Cliffs, USA: Prentice-Hall Inc.; 2003. James AN, Lumpton AR. Gypsum and anhydrite in foundations of hydraulic structures. Geotechniques 1978;28(3):249–72. James AN, Kirkpatrik M. Design of foundations of dams containing soluble rocks and soils. Quart J Eng Geol 1980;13:189–98. Ismael NF, Mollah MA. Leaching effects on properties of cemented sands in Kuwait. J Geotech Geoenviron Eng ASCE 1998;124(10):997–1004. Kota PB, Hazlett D, Perrin L. Sulfate-bearing soils: problems with calciumbased stabilizers. Transp Res Rec 2007;1546:62–9. Razouki SS, Al-Omary RR, Nashat IH, Razouki HF, Khalid S. The problem of gypsiferous soils in Iraq. In: Proceedings of the Symposium on Gypsiferous Soils and their Effect on Structures, NCCL, Baghdad, Iraq; 1994. p. 7–33. Razouki SS, EL-Janabi OA. Decrease in the CBR of a gypsiferopus soil due to long-term soaking. Quart J Eng Geol 1999;32(1):87–9. Razouki SS, Al-Azawi MS. Long-term soaking effect on strength and deformation characteristics of a gypsiferous subgrade soil. Eng J Univ Qatar 2003;16:49–60. Razouki SS, Kuttah DK. Effect of soaking period and surcharge load on resilient modulus and California bearing ratio of gypsiferous soils. Quart J Eng Geol Hydrogeol 2004a;37(2):155–64. Razouki SS, Kuttah DK. Distress of Light Structures and Pavements over Swelling Gypsiferous Soils. In: Proceedings of the international conference on geotechnical engineering, October 3–6, Sharjah, U.A.E; 2004. p. 387–394. Razouki SS, Kuttah DK. Predicting long-term soaked CBR of gypsiferous subgrade soils. Proc Inst Civil Eng Transp 2006;159(3):135–40. Razouki SS, Kuttah D, Al-Damluji OA, Nashat IH. Using gypsiferous soil for embankments in hot desert areas. Proc Inst Civil Eng Construct Mater 2008;161(2):63–71. Razouki SS, Kuttah DK, Abood M. Design of gypsiferous fill for hot desert road pavements. Proc Inst Civil Eng Construct Mater 2011;164(1):3–11. Razouki SS, Salem BM. Soaking–drying frequency effect on gypsum-rich roadbed sand. Int J Pavement Eng 2014. http://dx.doi.org/10.1080/ 10298436.2014.893326. Razouki SS, Abood MH, Al-Abbasy KJ. Top and base California bearing ratio behaviour of soaked gypsum-rich roadbed soils. Proc Inst Civil Eng Construct Mater 2014;167(CM3):131–9. Tomlinson MJ, Boorman R. Foundation design and construction. 6th ed. Singapore: Longman; 1996. US Department of Interior. Earth Manual. Water and Power Resources Service, 3rd ed. Washington, DC; 1980.