Bull Eng Geol Environ (2009) 68:97–106 DOI 10.1007/s10064-008-0177-7

ORIGINAL PAPER

Engineering behavior of the Lisan Marl as a dyke foundation material: Dead Sea, Jordan Ziad Mansour Æ Mohd Raihan Taha Æ Zamri Chik

Received: 3 April 2008 / Accepted: 23 August 2008 / Published online: 8 November 2008 Ó Springer-Verlag 2008

Abstract Problems have occurred with dykes constructed on the Lisan Marl to retain brine pumped from the Dead Sea, which is chemically processed to produce potash. In this study the engineering behavior, collapse potential and compressibility of the Lisan Marl were assessed by undertaking laboratory tests with both natural, distilled water and brine as the media. The results showed that the liquid limit and fine particles increased when the soil was mixed with fresh/distilled water compared with brine water. The soil can be classified as highly compressible with a low undrained shear strength. In terms of collapsibility, the soil could be classified as slightly collapsible upon inundation with distilled water and moderately collapsible when soaked in Dead Sea brine. The results are important when determining the height and nature of the dykes. Keywords Dead Sea

Marl
Dyke
Brine
Mechanical properties


Re´sume´ Pour retenir les saumures pompe´es de la Mer Morte en vue d’en retirer de la potasse par traitement chimique. Dans cette e´tude, les proprie´te´s me´caniques, l’aptitude a` l’effondrement et la compressibilite´ de la marne de Lisan ont e´te´ e´value´es en re´alisant des essais de laboratoire, utilisant comme fluide de l’eau naturelle distille´e ou bien de la saumure. Les re´sultats montrent que la limite de liquidite´ et la teneur en particules fines augmentent lorsque le sol est me´lange´ avec de l’eau douce distille´e par comparaison avec de l’eau issue de la saumure. Le sol peut eˆtre classe´ comme un sol fortement Z. Mansour (&)
M. R. Taha
Z. Chik Department of Civil and Structural Engineering, UKM Universiti, 43600 Bangi, Selangor, Malaysia e-mail: [email protected]

compressible, avec une faible re´sistance au cisaillement non draine´. Par ailleurs le sol peut eˆtre caracte´rise´ par une aptitude faible a` l’effondrement suite a` une saturation rapide par une eau distille´e et par une aptitude moyenne a` l’effondrement suite a` une saturation rapide par une saumure. Ces re´sultats sont importants a` conside´rer pour dimensionner les digues du point de vue de leur constitution et de leur hauteur. Mots cle´s Marne
Digue
Saumure
Proprie´te´s me´caniques
Mer Morte

Introduction At the southern tip of the Dead Sea in Jordan, a series of evaporation pans or ponds have been established by the Arab Potash Company (Fig. 1), in which the brine pumped from the Dead Sea is concentrated. The carnallite salts (the raw compound of potash) are then harvested from the pans and refined in the plant to form potassium chloride (KCl), simply termed as potash. Potash is a commercial product and mainly used as fertilizer. Increasing the potash production requires the construction of more salt pans. These salt pans are surrounded by large watertight dyke structures built on a soft silty clay—the Lisan Marl. In recent years, the Arab Potash Company (one of the biggest industrial companies in Jordan) began experiencing serious engineering problems threatening the stability and integrity of the existing dykes, and consequently the potash production. In particular, a dyke failed during impounding (Clossen 2004), and partial emptying of another dyke was required as a result of foundation problems (Arab Potash Company limited 2002; Tabbal and van der Schrier 2005). There was also quick, intensive erosion at the downstream

123

98

Z. Mansour et al.

Fig. 1 Arab Potash Company salt pans system (aerial photos in 1989 and 2001)

toe of a third dyke. These dykes were built on an emerging platform as the Dead Sea regressed (Fig. 3). The salt pans experience sinkholes, cracks and seepage (Clossen 2004). Baer et al. (2002) mentioned that wide, shallow subsidence features (WSSFs) were noticed in the early 1990s and have been a major problem along the Dead Sea shores in the occupied Palestine/Israel area and in Jordan. WSSFs were observed in the Lisan peninsula and the along Dead Sea shores as circular or elongated depressions with lengths ranging from a few hundred meters to a few kilometers. Two mechanisms could be responsible for the formation of such features: dissolution of the upper surface of the 20 m deep salt layer by fresh water and/or they may be structurally controlled by tectonic movements. Geoelectrical sounding measurements showed that areas underlying the coastal aquifers that were occupied by brine water were gradually flushed out and occupied by freshwater. The intrusion of freshwater into the Lisan Marl was documented by Salameh and El-Naser (2000) who calculated the volume of water in the coastal Dead Sea aquifers which was replaced by freshwater due to Dead Sea drawdown (-392 to -411 m) and migration of the fresh–brine water interface as about 3.21 9 1011 m3 (Fig. 3). The migrating freshwater dissolved the salt in the Lisan Formation, causing collapses and sinkholes along the foreshore. This highlighted the importance of studying the effect of both the change in the pore water salinity and the removal of soluble minerals from the Lisan Marl on the behavior of these sediments. The lack of published information on the engineering properties and performance of Lisan Marl as a foundation material upon leaching by fresh water and consequently the long term stability and integrity of the existing and future constructions was the motive for this research. The effect of the fresh water percolation could be one of the factors

123

contributing to the problems at the Arab Potash Company site and may influence future engineering construction and foundation problems. Earth dams which satisfactorily retained reservoir salt water for long periods have been known to quickly develop piping problems when fresh water replaced the salt water. This is possibly a result of ion concentrations in the clay layer of the dam being diluted and consequently forming dispersive clays i.e. fine grained soils that will deflocculate in still water and erode if exposed to low velocity water (McCarthy 2007).

The study area The Dead Sea The Dead Sea is a closed lake (Fig. 2) with no outlet except evaporation from the surface, which results in a drop in the brine level amounting to around 0.8–1.0 m/year (Fig. 3). The climate of the Dead Sea is classified as arid with an annual rainfall of about 65 mm and temperatures exceeding 40°C in the dry season. The salinity of the Dead Sea is approximately 332 g/l (Asmar and Ergenzinger 2002) resulting in an average specific gravity for the brine of 1.23. The current area of the Dead Sea is more than 500 km2 with an average width of 12 km. The Dead Sea is situated at the deepest part of the Jordan Rift Valley between Jordan to the east and occupied Palestine/Israel to the west (Fig. 2). It has unique chemical properties, significantly different from those of other oceans, seas or lakes (Abed 1985). The Dead Sea brine has the following salts concentration: calcium chloride (CaCl2) 14.4%, potassium chloride (KCl) 4.4%, magnesium chloride (MgCl2) 50.8% and sodium chloride (common salt, NaCl) 30.4% (Table 1).

Dyke on Lisan Marl: Dead Sea

99

Fig. 2 Dead Sea— geographical location

Fig. 3 Dropping of the Dead Sea brine level, 1976–2005

Table 1 Average ions distribution in the Dead Sea and normal oceans (g/l) (Abed 1985; Mahasneh 2004) Na

K

Ca

Mg

Cl

Dead Sea

38.9

7.1

16.72

39.7

208.9

Oceans

10.56

0.38

0.4

1.27

Geology of the study area The Dead Sea is a pull-apart basin or graben which formed associated with the Jordan Rift around 25 million years ago. This rift runs from the spreading Red Sea, through the Gulf of Aqaba in Jordan, the Araba Valley, the Dead Sea, Jordan Valley, Lebanon, Syria and finally towards the Taurus Mountains in Turkey (Fig. 4). The formation of the rift is a matter of dispute between two schools of thought. One supports the theory of vertical displacement along parallel normal faults while the plate tectonic theory suggests left lateral strike slip movement

Br

18.98

SO4

Na/K

Cl/Br

5.0

0.50

4.6

35.1

0.065

2.65

27.9

289.7

TDS 332 30–40

with a shear extending over 100 km due to the spreading of the Red Sea (e.g. Lyakhovsky et al. 1994; Sneh 1996). The Lisan Marl is named after the 225 km long Lisan Lake (Abed 1985) which extended from the Tiberia Lake in Palestine/Israel in the north to about 35 km south of a previously existing southern basin of the Dead Sea—now dried up. This alternating sequence of white (dominated by aragonite) and grey (dominated by calcite) marl is about 120 m thick and includes gypsum, clay minerals (mainly kaolinite), quartz, halite, feldspar and diatom. Salt diapers or salt domes can be found at several locations beneath the Lisan peninsula. It is believed that the salt domes are

123

100

Z. Mansour et al.

Fig. 4 Tectonic setting of the study area (after Arab Potash Company)

moving upwards as a result of their low density compared with the overlying soil. This slow movement has resulted in the formation of boundary faults, drag folds and radial fractures in the overlying sediments (Clossen 2004) which could also enhance fresh water infiltration.

Experimental programme Three sampling locations were chosen based on proximity to where the foundation problems have occurred and availability of access roads. Twelve undisturbed samples were collected from 1 to 2 m below the existing ground surface using 600 mm long, 100 mm diameter PVC tubes. As seen in Fig. 5, an excavator was used to dig out the hard salt crust. The disturbed upper surface of clay and salt was then cleaned and the Shelby tubes pushed vertically into the clay, first by hand and then using the excavator bucket. The samples were finally collected from the excavator bucket, both ends of each sample was waxed and the samples put into boxes and shipped to the laboratory. Mineralogy X-ray analysis was undertaken using the Shimadzu-XRD6000 with a Cu tube. Representative samples from three

123

Fig. 5 Recovery of sample tubes using the excavator (Arrows show the Shelby tubes)

locations were dried at 50 ± 5°C and ground to a fine powder before conducting the random X-ray diffraction test. Particle size distribution Representative samples from three locations weighing about 150 g were wet sieved following Head (1992) and the hydrometer tests used to determine the details of the fine fraction following British Standard (BS1377 Part 2:1990:9.5). The 50 g of oven dried (60 ± 5°C) material

Dyke on Lisan Marl: Dead Sea

passing the 63 lm sieve was dispersed using a solution of 35 g sodium hexametaphosphate, 7 g sodium carbonate and distilled water to produce 1 l of solution. The resulting slurry was shaken overnight to ensure that the soil particles were completely dispersed. Another three representative 150 g samples of known moisture contents were wet sieved through the 63 lm sieve using Dead Sea brine. The part retained on the 63 lm sieve was collected and left to settle for about 30 min; a sponge was used to remove the extra brine to avoid the precipitation of the salt during drying. The samples were then oven dried at 60 ± 5°C, weighed and disaggregated using a rubber pestle before standard mechanical sieving was undertaken. The portion passing the 63 lm sieve was discarded. Consistency limits The test was conducted using the fall cone method (Head 1992). The soil was wet sieved through the 425 lm sieve using distilled water, and partially oven dried at 60 ± 5°C to reach the plastic state. For comparison, both distilled water and Dead Sea brine were used as the molding media. Specific gravity The specific gravity of solids (Gs) for samples from each location was determined according to British Standard using a small pycnometer. Distilled water was replaced by kerosene to avoid any possible dissolution of soluble minerals from the soil matrix. The specific gravity of the kerosene was determined in the laboratory as approximately 0.785. Unconsolidated undrained triaxial test The test was conducted according to BS 1377 (Part 7:1990, Clause 8). A set of three identical samples from each location were hydraulically extruded from the Shelby tubes into 38 mm diameter, 100 mm long tubes. After pressurizing the triaxial cell, the axial load was applied with a strain rate of 1.5 mm/min. The test was terminated at 20% strain. Evaluation of collapse potential Thirteen samples were hydraulically extruded from the Shelby tubes into a 50 mm diameter, 20 mm thick consolidation ring and both single and double oedometer tests undertaken. The single oedometer test was performed according to ASTM (D 5333-96) as detailed by Day (2001). The undisturbed soil specimens at natural moisture content were loaded in the conventional oedometer to a stress level of

101

200 kPa. Each load was kept on the specimen for 1 h to minimize loss of natural water content i.e. ‘‘quick consolidation’’, except for the 200 kPa load which was maintained for 48 h (a period considered sufficient for the settlement to have ceased). The deformation readings were recorded at the end of the loading period. Before inundation in distilled water to induce collapse (ASTM 2000, Knight 1963), the sample was covered with aluminum foil and damp cloth to inhibit water evaporation After soaking, each load was maintained for 24 h and the strain readings recorded according to the conventional oedometer test readings. Three double oedometer tests were conducted on three sets of ‘‘identical’’samples. One of the specimens was tested at natural in-situ water content; the second sample was soaked with distilled water before the test began while the third was inundated with Dead Sea brine. All the samples were subjected to an initial load of 12 kPa and inundation was applied with simultaneous deformation monitoring. Thereafter, the samples were loaded incrementally with identical applied loads. Each applied pressure was maintained for 24 h. After the maximum desired pressure was reached (400 kPa), the specimens were unloaded in three decrements to a stress level of 12 kPa. The unsoaked sample was covered with aluminum foil and damp cloth to inhibit water evaporation.

Results and discussion X-ray results Figure 6 shows a representative X-ray diffraction diagram. Calcite and aragonite are the predominant minerals, with gypsum, halite, dolomite, meionite (Ca4Al6Si6O24(CO3)) and quartz also present. Kaolinite was found to be the major clay mineral at the three sampling locations (Fig. 7). This conforms to the findings of Abed (1983) who conducted a mineralogical study of the white and grey laminae of the Lisan Marl. Traces of anhydrite in some samples were also reported by Abed (1983). Specific gravity The specific gravity (Gs) of the samples ranged from 2.21 to 2.31. This low value is thought to be due to the presence of gypsum which has a specific gravity of between 2.3 and 2.4 (Jones 1989). The presence of organic material (diatoms) noted by Abed (1985) could also play a role in decreasing the specific gravity. Observation of a soil column at one sampling location showed an alternation between laminated and unlaminated silty clay in which the specific gravity for the laminated part was 2.175.

123

102

Z. Mansour et al.

Fig. 6 Representative random X-ray diffraction

Fig. 7 Representative oriented X-ray diffraction treated with HCl

Particle size distribution Figure 8 shows the result of the sieving analysis for a representative sample. The percentage passing the 63 lm sieve (wet sieving) ranged between 95 and 99%. The corresponding values when brine was used as the washing media were 80–93%.The increase in the fines percentage is attributed to the dissolution of the soluble minerals such as halite and to a lesser degree gypsum. In addition, the dissolution of the cementing material encourages the

Fig. 8 Grain size distribution

123

degradation of larger particles. Frydman et al. (1977) found that mixing KCl (the main salt in Dead Sea brine) with heavy clay decreases the clay fraction (\0.005 mm) from 75 to 50%. Consistency limits A decrease in liquid limit occurred when brine was used as the molding agent. The maximum liquid limit was 35% and the minimum 27.5% using distilled water while the corresponding values when brine was used were 30.5 and 22.5%. The plastic limits were in the range of 21–25% when tested with distilled water and 16–23% using brine. The decrease in Atterberg limits is due to the formation of aggregates which reduces the available surface for interaction with water (Okagbue and Onyeobi 1999). The presence of high valence exchangeable cations (such as Mg?2 and Ca?2) in the Dead Sea brine decreased the distance between particles by decreasing the repulsive forces such that the van der Waal’s attractive forces became dominant, hence increasing the capillary stress between particle boundaries which affects the formation of the aggregates (Pandian et al. 1991). The presence of K?

Dyke on Lisan Marl: Dead Sea

Fig. 9 Unconsolidated undrained triaxial test results for representative sample

ions in the Dead Sea water (even monovalent) decreases the liquid limit and compressibility (Torance 1975; Rao et al. 1989). This is consistent with the findings of Frydman et al. (1977) who reported that mixing KCl from the Dead Sea with heavy clay decreased the activity of the clay with a consequential decrease of the liquid limit and plasticity index. When Dead Sea brine was used as a soil improvement agent, the salts decreased the double layer thickness and water content and consequently increased the soil stiffness (Mahasneh 2004). When washed with brine water, the Lisan Marl would be classified as CL–ML to CL (USCS soil classification) but when distilled water is used all samples would be classified as CL.

Unconsolidated undrained triaxial test (UU) Figure 9 shows the unconsolidated undrained triaxial test results for a representative sample. The stress-strain relationship shows that the yielding and failure points were identical and easily defined. The average calculated undrained shear strength (su) was 14 kPa. The subsequent material performance may be defined as strain-hardening. This behavior is typical for a flocculated clay structure. When the clay is subjected to shear, some interparticle bonds break down while new ones are formed continuously (Kezdi 1974). This compares with the results for hand shear vane strengths reported from the vicinity of the study area by Al-Homoud et al. (1999) of 10 kPa at 1.0 m depth and 20 kPa at 4 m depth.

Evaluation of collapse potential Single Oedometer test Figure 10 shows the collapse test results for a representative sample described as greenish grey calcite

103

Fig. 10 Consolidation curves for single oedometer test

interlaminated with white aragonite. Knight (1963) and subsequently Day (2001) pointed out that the collapse potential (CP) could be calculated (following ASTM D 5333-96, 2000) as: CPð%Þ ¼ ðDec =1 þ eo Þ  100 where Dec is the change in void ratio resulting from saturation at a stress level of 200 kPa, eo is the natural void ratio. From this equation, the collapse potential of the studied soil is about 0.34%. Day (2001) suggests such soils could cause some problems while Jennings and Knight (1975) would anticipate no problem. This low value of collapse potential is thought to be due to the dissolution of the readily soluble sodium chloride (NaCl) from the soil matrix during the inundation process. This increases the void ratio and leads to collapse as a result of ‘‘chemical piping’’. This is different from the hydrocollapse of the matrix under applied pressure which occurs in a porous honeycomb soil structure (Karakouzian et al. 1996). For soil with non-expansive clay minerals, the pore water chemistry has relatively little effect on the compression behavior after the initial fabric has formed and the structure has stabilized under moderate effective stress. Mitchell and Soga (2005) note that leaching of normally consolidated marine clay with a high water content may be sufficient to cause a small reduction in volume owing to changes in interparticle forces. The test results also show that the compression index (Cc) is approximately 0.8 and the swelling index (Cs) is 0.03 while the coefficient of secondary compression (Ca) is 0.008 for a stress range of 200–400 kPa. The high value of Cc would mean the soil could be classified as highly compressible soil according to Mitchell and Soga (2005). Al-Homoud et al. (1999) recorded Ca values ranging between 0.005 and 0.013 in different locations in the Lisan peninsula. This wide range may be related to a change in the nature of the soil, e.g. the presence, direction and continuity of laminations.

123

104

Z. Mansour et al.

Double-Oedometer test The double-oedometer test results are shown in Fig. 11 in the form of e–log p plots. Taking the curve of a natural unsoaked sample as reference, it can be seen that soaking with distilled water caused ‘‘chemical collapse’’ behavior due to the dissolution of NaCl. Inundation of the sample with Dead Sea brine results in an appreciable increase in compressibility which is thought to be due to the dissolution of gypsum. AlNouri and Saleam (1994) reported that gypsum possesses intermediate solubility (0.2%), but the amount dissolved is much greater if the water contains salt. Mielenz and King (1955) reported that for clayey loess, the electrolytic aqueous solutions decrease collapse in the following order H2O [ NaCl solution [ FeCl3 solution. The increase in compressibility upon using brine was also demonstrated by Azam (2000) in a study of calcareous soil. Furthermore, Petrukhin and Presnov (1989) stated that the dissolution of gypsum from the soil mass will cause one or more of the following three processes: collapse of soil structure as a result of the breaking of bonds between soil particles; consolidation; and/or leaching due to continuous water flow through the soil mass. The latter leads to progressive collapse of the soil structure as the removal of gypsum is continuous. In terms of collapsibility, Fig. 12 shows that the inundation of the sample in brine caused a collapse potential greater than 2% at 200 kPa stress level (moderately problematic soil). It can be observed that the two curves almost converge at stress levels of about 400 kPa. The converging of the two curves may be explained by the fact that the saturated curve has reached the limiting void ratio for particle packing, while the curve for natural soil could go beyond the yield point (Lutenegger and Saber 1988). This could also be the result of the breaking down of the natural soil structure at that stress level.

Fig. 11 Consolidation curves for double oedometer test

123

Fig. 12 Collapse potential of soil versus applied pressure

For collapse potential evaluation, as the two curves (unsoaked and soaked in distilled water) did not initiate from the same eo point, a correction has been made according to the procedure suggested by Jennings and Knight (1975) using a total overburden pressure of 20 kPa. As observed from Fig. 12 and according to Day (2001), the soil could be classified as a slightly problematic soil in terms of collapsibility (Table 2). The values of coefficient of consolidation (Cv) versus applied pressure are shown in Fig. 13. The Cv decreases as the vertical stress increases up to a stress level of 100 kPa, after which no further changes were noted with increased loads. The maximum recorded value for Cv was 112 m2/ year at a stress level of about 25 kPa for the sample soaked in distilled water. Al-Homoud et al. (1999) in a study conducted on the Lisan Marl reported values of Cv in the order 100 m2/year to 200 m2/year based on field data and constant strain oedometer tests. The high rate of compression at low pressure for the sample treated with

Table 2 Collapse potential versus severity of the problem—at 200 kPa stress level (after Day 2001) Collapse Potential (%)

Severity of problem

0

No problem

0.1–2

Slight

2.1–6

Moderate

6.1–10

Moderately severe

[10

Severe

Fig. 13 Coefficient of consolidation (Cv) versus applied pressure

Dyke on Lisan Marl: Dead Sea

distilled water may be due to the dissolution of readily soluble NaCl from the soil matrix, which increases the permeability and enhances the consolidation process before the breaking down of the soil structure at higher loads. The compression indices (Cc) for the tested samples were 0.81 for the sample soaked in distilled water, 0.92 for the natural unsoaked sample and 0.98 for the sample soaked in brine; i.e. highly compressible soil according to Mitchell and Soga (2005). Higher values (up to about 1.3) were recorded by Quigley and Ogunbadejo (1972) for soft varved clay. The relatively lower value of Cc for the sample soaked in distilled water could be attributed to the formation of lath-like anhedral gypsum crystals due to hydration. These aggregate in conjoined colonies and it is the consequential arching of the gypsum crystals which is responsible for the relatively low compressibility (Azam 2000). The steeper virgin portion (high Cc) for the sample inundated with brine is an indication of a sensitive material with an open structure (Quigley and Ogunbadejo 1972). Dissolution of gypsum increases the sensitivity of the soil (Azam 2000). The overconsolidation ratio (OCR) for the studied soil is in the range of 2.2–2.4, with an average preconsolidation pressure of 60 kPa. It is considered that the relatively high OCR in spite of the high Cc values could be attributed to the sensitive nature of the studied marl (Hossain and Ali 1988). Figure 14 shows the values of coefficient of compressibility (mv) against consolidation pressure. All the tested samples are located in the zone of high compressibility according to Head (1992), with the highest value of 1.46 m2/MN at a stress level of 100 kPa for the sample soaked in Dead Sea brine. Again mv sharply increases up to a stress level just beyond the pre-consolidation pressure, after which it starts to gradually decrease. In Fig. 15 the plots of secondary compression index (Ca) versus consolidation pressure show that Ca gradually increased as the pressure increased up to a stress level of 100 kPa (just above the pre-consolidation pressure, Pc & 60 kPa), beyond which it remained almost constant

105

Fig. 15 Change of coefficient of secondary compression with applied pressure

for all three conditions. This behavior is typical for inorganic clays for which Ca was found to be independent of applied stress when loaded above the pre-consolidation pressure (Head 1992). This could be due to the failure of the soil structure as a result of breaking down of soil bonds at that critical strain. The lower values of Ca for the sample soaked in brine compared with the other two samples is thought to be due to the higher rate of gypsum dissolution. This causes the major breakdown of the soil structure during primary consolidation, thus relatively few bonds remain to be broken by secondary compression. Subjecting the sample to distilled water may lead to partial dissolution of the gypsum and weakening of particle cementation. In addition, the weak strength and flexibility of needle-like gypsum crystals (Horta and De 1989) contribute to the particle re-arrangement manifested during secondary compression, explaining the higher values of Ca for the sample soaked in distilled water. In the light of the high compressibility, sensitivity and low strength of the Lisan Marl, it could be considered as a difficult soil for construction purposes. Although low collapse potential values were obtained from the testing, it is important to ensure a minimum brine level in the pans so that the brine can flow under gravity from one pan to another. The additional settlement referred to above should be taken into account when assessing post construction settlement and hence the height of the dykes required.

Conclusions

Fig. 14 Variation of coefficient of compressibility (mv) with applied pressure

The study presented in this paper focuses on the basic properties, collapse potential and compressibility of the Lisan Marl from Jordan. It was found that using Dead Sea brine as a testing medium instead of distilled water decreases the percentage of fines and the liquid limit of the soil, which slightly changes the soil classification. The studied soil exhibited strain hardening behavior in the unconsolidated undrained triaxial tests. In terms of collapsibility, the soil is classified as slightly collapsible. Soaking the Lisan Marl in Dead Sea brine in

123

106

the double oedometer test increased both the collapse potential and the soil compressibility compared with inundation in distilled water. Nevertheless both soaking media failed to produce significant soil collapse. The steep virgin portion (i.e. high Cc) for the studied soil in spite of the relatively high OCR is an indication of sensitive material with an open structure. Dissolution of gypsum increases the sensitivity, notably for the sample inundated with brine. The additional settlement resulting from the collapse behavior should be considered when assessing the height of the dykes. Acknowledgments The authors acknowledge Mr. Brent Heimann, Dr. Mah. Tabbal and site staff of The Arab Potash Company for permission to access some unpublished data and for logistical and financial support during the site visit. The authors also acknowledge Dr. Mustafa Al-Naddaf from Yarmouk University, Jordan for the Xray analyses. The contribution of Mr. Talal Abu Baker of Triple Corporation, Jordan and the Sinohydro Company, China is greatly appreciated.

References Abed AM (1985) Geology of the Dead Sea, 1st edn. Dar Al Arqam, Amman (in Arabic) Abed AM (1983) Paleo climates of the Jordan rift in the uppermost Pleistocene. 2nd international archeological congress of Jordan Al-Homoud AS, Makinson C, Tal AB, Penman JG (1999) Insitu geotechnical experimentation and stability evaluation of an instrumented 180 m 9 70 m 9 12.5 m fill dike constructed on laminated soft ground. Environ Geol 39(2):177–196 AlNouri I, Saleam S (1994) Compressibility characteristics of gypseous sandy soils. Geotech Test J 17(4):465–474 Arab Potash Company limited (2002) Invitation to bid for the retender for execution of remedial works, dike 18. Bid No. APC/ 001/2002 Asmar BN, Ergenzinger P (2002) Prediction of the Dead Sea dynamic behavior with the Dead Sea–Red Sea canal. Adv Water Resour 25:783–791 Azam S (2000) Collapse and compressibility behavior of arid calcareous soil formations. Bull Eng Geol Env 59:211–217 Baer G, Schattner V, Wach D, Sandwell D, Wdowinski S, Frydman S (2002) The lowest place on the earth is subsiding-An INSAR (interferometric synthetic aperture radar) perspective. Geol Soc Am Bull 114(1):12–23 Clossen D (2004) Structural control of sinkholes and subsidence hazards along the Jordanian Dead Sea coast (Env. Geol., issue: on line). Springer, Heidelberg Day RW (2001) Soil testing manual. McGraw-Hill, NY Frydman S, Ravina I, Ehrenreich T (1977) Stabilization of heavy clay with potassium chloride. Geotech Eng 8:95–108 Head KH (1992) Manual of soil laboratory testing, 2nd edn. Pentech Press, London Horta JC, De OS (1989) Carbonate and gypsum soils properties and classification. In: Proceedings of the 12th international

123

Z. Mansour et al. conference of soil mechanics and foundation engineering Rio de janerio 1:53–56 Hossain D, Ali KM (1988) Shear strength and consolidation characteristics of Obhor sabkha, Saudi Arabia. Q J Eng Geol Hydrogeol 21(4):347–359 Jennings JE, Knight K (1975) A guide to construction on or with materials exhibiting additional settlement due to collapse of grain structure. 6th regional conference for Africa on soil mechanics and foundation engineering Durban, South Africa. pp 99–105 Jones DR (1989) Gypsum/anhydrite: a brief review of geotechnical problems related to construction in north Kuwait. First symp. geotech. prob. In Kuwait, Kuwait University, 1:43–51 Karakouzian M, Pitchford A, Leonard M, Johnsan B (1996) Measurements of soluble salt content of soils from arid and semi-arid regions. Geotech Test J 19(4):364–372 Kezdi A (1974) Handbook of soil mechanics. Akademiai kiado, Budapest. p 294 Knight K (1963) The origin and occurrences of collapsing soil. In: Proceedings of the 3rd regional conference for Africa on soil mechanics and foundation engineering 1:127–130 Lutenegger AJ, Saber RT (1988) Determination of collapse potential of soils. Geotech Test J 11(3):173–178 Lyakhovsky V, Ben-Avraham Z, Achmon M (1994) The origin of the Dead Sea rift. Tectonophysics 240:29–43 Mahasneh BZ (2004) Dead Sea water as a soil improvement agent. Electron J Geotech Eng 9 McCarthy FD (2007) Essentials of soil mechanics and foundations, 7th edn. Pearson Prentice Hall, Ohio, p 850 Mielenz RC, King ME (1955) Physical–chemical properties and engineering performances of clays: in clays and clay technology. Calif Div Mines Bull 169:196–254 Mitchell JK, Soga K (2005) Fundamental of soil behavior. Wiley, USA, p 577 Okagbue CO, Onyeobi TUS (1999) Potential of marl dust to stabilize red tropical soils for road construction. Eng Geol 53:371–380 Pandian NS, Nagaraj TS, Babu GLS (1991) Effect of drying on the engineering behaviour of Cochin marine clay. Geotechnique 4(1):143–147 Petrukhin VP, Presnov OM (1989) Collapse deformation of gypsum sand. Soil Mech Found Eng 28(3):127–130 Quigley RM, Ogunbadejo TA (1972) Clay layer fabric and oedometer consolidation of a soft varved clay. Canadian Geotech J 9:165– 174 Rao SM, Sridharan A, Chandrakaran S (1989) Influence of drying on the liquid limit behavior of a marine clay. Geotechnique 39(4):715–719 Salameh E, El-Naser H (2000) The interface configuration of the fresh/Dead Sea water-theory and measurements. Acta hydrochemica et hydrobiologica 28(6):323–328 Sneh A (1996) The Dead Sea rift: lateral displacement and down faulting phases. Tectonophysics 263:277–292 Tabbal MA, Van Der Schrier JS (2005) Liquefaction assessment dike 18 Dead Sea, Jordan. The international earthquake engineering conference. Dead Sea – Jordan Torance JK (1975) On the role of chemistry in the development and behavior of the sensitive marine clays of Canada and Scandinavia. Canada. Geotech J 12:326–335