Available online at www.sciencedirect.

com

Engineering Geology 99 (2008) 85 90

www.elsevier.com/locate/enggeo

Engineering properties of loess in Algeria

M.S. Nouaouria , M. Guenfoud, B. Lafifi
Civil Engineering and Hydraulics Laboratory, Guelma University, Algeria
Received 3 February 2007; received in revised form 10 November 2007; accepted 20 January 2008
Available online 16 February 2008

Abstract

Loess in North Africa has been investigated using samples from Algeria. The specific gravity, Atterberg limits, grain size distribution and dry
density were determined. The hydro-collapsibility properties, due to wetting under different stress levels were measured in single-oedometer tests.
The results of this investigation indicate that the properties of Algerian loess are similar to those of loess from many parts of the world, such as
Iowa and Libya; they can be classified as silty loess.
2008 Elsevier B.V. All rights reserved.

Keywords: Algeria; Grain size; Engineering properties; Hydrocollapse; Loess; Oedometer tests

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
2. Materials and experimental programme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3. Test results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.1. Index properties of Algerian loess . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4. Hydrocollapse characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

1. Introduction
The Sahara is one of the world's major sources of minerogenic
dust (Prospero et al., 2002), many of the silty deposits in adjacent
areas being attributed to aeolian processes. Loess is found in many
parts of the worlds. In China, for example, it has been estimated
that the area covered by loess exceeds 630,000 km2, of which
about 60% is collapsible (Derbyshire et al., 1995). Silt
constituting 5070% by weight is the predominant grain size
fraction in loess, less than 5% of particles being greater than
1 mm, and clay particles making up only a few percent. Quartz
Corresponding author.
E-mail address: nouaouria@yahoo.com (M.S. Nouaouria).
0013-7952/$ - see front matter 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.enggeo.2008.01.013
and feldspar are the major mineral constituents of the coarser
grades, while illite is the principal clay mineral. In addition,
collapsible loess may contain relatively large amounts of calcium
carbonate, as well as various soluble salts distributed on the
surface of particles in solid or semi-solid state.
Loess deposits have certain distinctive characteristics (Rogers
et al., 1994) that define the material. Two major defining features
are particle size distribution and the susceptibility of the soil
structure to collapse when loaded and wetted. Chen (1992) stated
that the defining physical characteristics of loess in its natural state
are low water content and high porosity.
It has been estimated that loess covers nearly 10% of the
surface area of the earth (Pecsi, 1968). It is found in continental
drylands on all continents, with the notable exception of Africa.
86 M.S. Nouaouria et al. / Engineering Geology 99 (2008) 8590

response to wetting at different stress levels. For this purpose,

undisturbed soil samples representing two types of loess
were collected from two locations: Daia Bendahoua and Metlili,
Ghardaia (southern Algeria).

2. Materials and experimental programme

An experimental investigation was carried out to determine the

index properties and the hydro-collapsibility characteristics and
behaviour of the samples. Undisturbed block samples were
obtained from depths of between 0.6 and 1 m, and then carefully
trimmed, waxed and placed in boxes. The two samples sites lie
45 km apart in the Ghardaia region, some 600 km south of Algiers
(Fig. 1), some soil assessment has already been undertaken in this
region with a view to dam construction.
Specific gravity, and Atterberg limits were determined ac-
cording to British Standard Procedure (BSI, 1990). Particle size
Fig. 1. Map of Algeria showing sampling location, Ghardaia. analysis of material passing the 63 m sieve was undertaken using
Sedigraph Technique. The microstructure of loess particles was
Scheidig's (1934) map of world loess distribution recognizes only examined using a scanning electron an X-ray scanning system
one loess region in Africa (North Libya). However, this (Sedigraph). The collapsibility characteristics and behaviour of the
conclusion was based on limited information, owing much to
the work of Rathjens (1928) who described loess-like materials in
Libya but found Tripolitanian loess to have a grain size N 100 m
for loess. More recent work by Coud-Gaussen (1987) has shown
that North African loess may have larger modal sizes than better
known loess deposits. In the Techine field-section (Matmata,
Southern Tunisia), for examples, three series occur in a 17 m thick
succession (Coud-Guassen et al., 1982): an upper grey-beige
loess, an intermediate series with altered reddish horizons; and a
lower ochre loess. This work showed that Tunisian loess is coarser
in granulometric composition than most other loess. Assallay
et al. (1996) have recently justified Scheidig's classification of the
North Libyan silty deposits as loess. They demonstrated that it has
index properties and shows aspects of hydrocollapse behaviour
similar to those found in loess from many parts of the world. They
concluded that this material can be classified as silty loess in the
Tripoli region and clayey loess in the Ghat area. Abdrabbo et al.
(2000) have reported an extent of loess throughout the Western
Desert of Egypt, especially at certain locations including Sidi
Baranee, New El-America city, and El-Boustan. Despite the huge
surface area of Algeria, no serious study has tackled the question
as to whether or not loess is present in this vast area. In their study
of Algerian duricrusts, Smith and Whalley (1982) concluded that
these residual deposits contain quartz and feldspar grains, which
exhibits evidence of limited aeolian abrasion and chemical
weathering; they suggested a Holocene age for these deposits.
Because of the potential hazards posed to engineering struc-
tures with foundations on collapsible soils that undergo high
volume change, it is important to identify such soils and to
understand clearly their engineering properties. This paper is a
first attempt to fill the gap in knowledge of the known loess-like
silts of Algeria. Given the clear need for more careful analysis of
North African silty sediments as a means of determining whether
they constitute true loess, the main objective of this paper is to
investigate the collapsibility characteristics and behaviour of Fig. 2. Scanning electron micrograph of loess samples showing clay-coated silt
selected Algerian materials with particular reference to their grains of natural structure of (a) yellow loess and (b) grey loess.
 M.S. Nouaouria et al. / Engineering Geology 99 (2008) 8590 87

Table 1
Collapse settlement criteria for KansasNebraska loess (after Holtz and Hilf,
1961)
Dry density Condition
(mg m 3)
b1.28 Loess is considered loose and highly susceptible to settlement.
1.281.44 Loess is medium dense and is moderately susceptible to
settlement, particularly for critical or heavily loaded structures.
N1.44 Loess is quite dense and may be capable of supporting
ordinary structures without serious settlement.
1.36 A more general criterion has been used for earth dams: this
density is used as the division between low and high-density
loess or the division where special foundation treatment is
required for lower densities. Fig. 3. Plasticity properties of Algerian loess as compared with loess types
defined by Gibbs and Holland (1960). ( Grey loess; Yellow loess).
materials with particular reference to their response to wetting at
different stress levels was assessed by subjecting undisturbed soil Bendahoua) and Yellow (from Metlili) samples were prepared
specimens (70 mm in diameter and 19 mm thick) to one- and tested in natural conditions, water contents at the time of
dimensional compression in an oedometer. testing being 5% and 6% respectively.
Two methods are currently used to evaluate and determine the
collapse parameters, namely the single and double oedometer 3. Test results and discussion
tests. These have been shown to be reliable in investigating
collapsibility properties. According to Lawton et al. (1989), and 3.1. Index properties of Algerian loess
Basma and Tuncer (1992), these techniques give similar results.
The single-oedometer test consists of loading the specimens Index properties for the Yellow and Grey Loess are summarised
incrementally to a specific state of vertical stress and allowing the in Table 2. Despite the close similarity of their general
sample to come to equilibrium under the applied pressure. The characteristics, there are some slight differences in index proper-
sample is then flooded with water, and the deformation is ties, although both soils are classified as silty loess. The Grey
measured. The deformation induced by the addition of water is Loess consists of 16% sand, 75% silt and 9% clay. The more silty
divided by the initial height of the specimen, expressed in percent, Yellow loess consists of 2% sand, 86% silt and 12% clay. The
from which the magnitude of the hydrocollapse can be defined. Yellow loess has a slight higher dry density and higher specific
The single oedometer test method was performed on un- gravity (2.73). This may be explained by the fact that, as shown by
disturbed soil specimens from Ghardaia. Grey (from Daia X-ray analysis, its composition is dominated by the feldspar
group (Potassium Aluminium Silicate K (Si3Al)O8), and
Table 2
Calcium Magnesium Carbonate (CaMg(CO3)2), while Calcium
Index properties of Libyan (After Assallay et al., 1996) and Algerian loess Magnesium Carbonate is dominant in the Grey loess, which
contains only 4% of K(Si3Al)O8) and has a specific gravity of 2.68.
Property Tahala Gharyan Khoms Garaboli Grey Yellow
loess loess loess loess loess loess The specific gravity values of the Algerian samples are very
(Libya) (Libya) (Libya) (Libya) (Algeria) (Algeria) similar to those for loess from many other parts of the world. For
Natural moisture example, Assallay et al. (1996) report values for Libyan loess of
Content (%) 3 6 3 2 5 6 between 2.66 and 2.73. Values for Chinese loess of between 2.65
Specific gravity 2.73 2.66 2.68 2.67 2.68 2.73
(Gs)
Dry density 1.36 1.40 1.42 1.43
(Mg/m3) 1.42 1.46
Bulk density 1.49 1.52
(Mg/m3)
Void ratio 0.87 0.84 0.89 0.91
0.96 0.92
Degree of 15 18
saturation (%)

Mechanical analysis
Sand (%) 10 38 18 39 16 2
Coarse silt (%) 52 30 39 31 51 61
Fine silt (%) 25 21 30 22 24 25
Clay (%) 13 11 13 8 9 12

Atterberg limits
LL (%) 41 27 31 25 30 33
PL (%) 24 19 20 17 23 22
PI (%) 17 8 11 8 7 11
Fig. 4. Particle size distribution of Algerian yellow and grey loess.
88 M.S. Nouaouria et al. / Engineering Geology 99 (2008) 8590

Fig. 5. Hydrocollapse-wetting pressure relationship for undisturbed specimens

from Algeria (Grey loess and Yellow Loess) tested in single Oedometer test.

and 2.70 are reported by Bell (1992), those from Kent, U.K. Vary Fig. 6. Hydrocollapse test results for undisturbed grey loess flooded at different
from 2.68 to 2.70 (Derbyshire and Mellors, 1988), those in parts of stress levels. (Pw = Wetting pressure).
the United States lie between 2.57 and 2.79 (Sheeler, 1968), while
those for the loess of Iowa State ranges from 2.68 to 2.73. The appropriate in the light of the view of Lutenegger and Hallberg
microstructure studies show that both samples have an open (1988) that this is the most common method used to identify a
structure (Fig. 2). collapsible soil, and one with the most meaning to geotechnical
Using the collapse settlement criteria developed by Holtz and engineers. The degree of hydrocollapse of each sample was then
Hilf (1961) shown in Table 1, the Algerian loess can be considered calculated as follows.
as medium dense and moderately susceptible to settlement,
H
particularly beneath critical or heavy structures. Table 2 shows Hydrocollapse k
comparative index properties for Algerian and Libyan loess. H0
Loess soils are slightly to moderately plastic, the plasticity
Where
increasing with increases in clay content. Based on plasticity
characteristics, Gibbs and Holland (1960) subdivided loess into
H = change in sample height due to saturation
three types: sandy loess, silty loess and clayey loess. They
H0 = Sample height immediately prior to saturation
indicated that Liquid limits for silty loess varied from 25 to 35%
and values for clayey loess ranged up to 45%. All plasticity values
For the double oedometer test, the change in sample height
for the Algerian loess samples plot close to the A-line and below
due saturation is taken as the difference between the unsaturated
the 50% Liquid Limit on the Casagrande plasticity chart (Fig. 3),
sample heights at any given pressure. These definitions are
indicating that the material is inorganic soil of low plasticity. Such
standard throughout the literature Rogers et al. (1994).
soils are generally very sensitive to changes in moisture content.
The curves shown in Fig. 5 illustrate the effect of wetting
The test results show that all of the Algeria samples considered
pressure on the collapsibility of the loess, and indicate that the
here plot within the silty loess envelope of Holtz and Gibbs
magnitude of the hydrocollapse increases as the applied pressure
(1951). Particles size results for both samples are shown in Fig. 4.
It is evident that the dominant size fraction in these samples is silt
(260 m). Silt makes up some 86% of the Yellow loess and 75%
of the Grey, and clay percentages are 12 and 9, respectively.

4. Hydrocollapse characteristics

Undisturbed soil specimens from Ghardaia were prepared for

testing. The single Oedometer test was considered the most

Table 3
Identification of hydrocollapse problems (after Jennings and Knight, 1975)
Collapse potential (CP) Severity of problem
0%1% No problem
1%5% Moderate problem
5%10% Trouble
10%20% Severe trouble
Fig. 7. Hydrocollapse test results for undisturbed yellow loess flooded at different
N20% Very severe trouble
stress levels. (Pw = Wetting pressure).
 M.S. Nouaouria et al. / Engineering Geology 99 (2008) 8590
of wetting increases up to a certain value, at which the maximum
collapse will be obtained due to saturation.
Based on the values given by Jennings and Knight (1975) and
shown in Table 3, it can be taken that the Algerian loess con-
sidered here poses severe collapse problems. For a stress range up
to 1600 kPa, the measured magnitudes of hydrocollapse of the
specimens vary over a range of 1 to 21% and 0.8 to 20% for the
Grey and Yellow loess, respectively, depending on the level of the
vertical pressure applied to the specimens at the time of wetting.
Figs. 6 and 7 show the post-wetted relationship between specimen
height and applied pressure for all specimens that were flooded at
different stress levels.
5. Discussion
On the basis of the results presented here, it is evident that
there are similarities between the loessic soils in southern
Algeria and those found in many others parts of the world.
This includes the demonstrable collapsibility of the Algerian
soils described, collapsibility generally being predicted on the
soil having void space in its natural state that is sufficient to
sustain its liquid limit moisture at saturation.
As far as the Algerian loess is concerned, the high void ratios
of the Algerian loess result in a hydro-collapse magnitude of
about 13% for both Grey and Yellow loess for a wetting
pressure of 200 kPa. Based on the identification of the severity
of collapsibility given by Jennings and Knight (1975), the
Algerian loessic soils tested in this study are regarded as likely
to pose severe problems, having a susceptibility to settlement
particularly in respect of critical or heavily loaded structures.
The Atterberg limits, grain size characteristics, and specific
gravity of these Algerian loess samples were found to be very
similar to those of loess deposits from many parts of the world.
Based on the plasticity and particle size test results, and according
to the definition of loess types given by Holtz and Gibbs (1951) and
Gibbs and Holland (1960), the Algerian material can be classified
as silty loess. The amount collapse depends on the magnitude of
vertical stress at the time of wetting. As loess and loess-like
sediments are always unsaturated, the reduction in matric suction is
one of the major causes of collapse (Lloret and Alonso, 1980;
Maswoswe, 1985). It seems vital that in parallel with the continued
accumulation and comparative analysis of the bulk behaviour of
loess, greater effort should be devoted to the study of the behaviour
of collapsible loess in the context of the theory of unsaturated soils
in which the constitutive relations for volume change in
unsaturated soils could be used to explain more comprehensively
collapsible soil (Fredlund and Morgenstern, 1976).
6. Conclusion
(a) The southern Algerian loess is of two types, named here
Grey and Yellow loess. Lithologically and mineralogi-
cally, the two types are quiet similar, although there are
some differences composition that explain the colour
difference.
(b) The results of this investigation indicate that the Index
Properties of southern Algerian loess are similar to those
89
found in loess from many parts of the world, such as Iowa
(USA) and Libya.
(c) Both the Grey and Yellow loess soils are rich in silt and
can be classified as a silty loess.
(d) The amount of hydrocollapse is dependent on the magnitude
of vertical stress acting on specimen at the time of wetting.
The degree of hydrocollapse increases as the applied pres-
sure increases up to a certain level of stress, at which point
maximum collapse is reached (20% and 21% for yellow and
Grey loess, respectively).
(e) A unique post-wetted specimen height applied pres-
sure relation was observed in the Algerian samples tested.
Flooding at any applied pressure caused the material to
collapse down to a certain specimen height prescribed by the
unique post-wetted specimen height/applied pressure curve.
References
Abdrabbo, F.M., Hansy, R.M., Hamed, K.M., 2000. Collapsibility of Egyptian
loess soil. unsaturated soils for Asia. In: Toll, Leong (Eds.), Rahardjo.
Balkema, Rotterdam, pp. 595606.
Assallay, A.M., Rogers, C.D.F., Smalley, I.J., 1996. Engineering properties of
loess in Libya. Journal of Arid Environment 32, 373386.
Basma, A.A., Tuncer, E.R., 1992. Evaluation and control of collapsible soils.
Journal of Geotechnical Engineering Division, ASCE 118, 14911504.
Bell, F.G., 1992. Engineering Properties of Soils and Rocks, 3rd ed. Butterworth-
Heinemann, Oxford, U.K. (345pp.).
British Standards Institution, 1990. Bs 1377: British Standards Methods of test
for soils for civil engineering purposes. B.S.I., London.
Chen, Z.Y., 1992. Practical solutions to problems of collapsible loess in China.
Proceedings of the 7th International conference on expansive soils. Dallas,
Texas, pp. 112.
Coud-Gaussen, G., 1987. The Peri-Saharan Loess: sedimentological character-
ization and paliclimatical significance. Geological Journal 15, 177183.
Coud-Gaussen, G., Mosser, C., Rognon, P., Tourenq, J., 1982. Une accumulation de
loess du Plistocne suprieur dans le Sud-Tunisien: la coupe de Tchine.
Bulletin de la Societe geologique de France 7 (t. XXIV), No. 2, 283292 pp
(in French).
Derbyshire, E., Mellors, T.W., 1988. Geological and geotechnical characteristics of
some engineering loess and loessic soils from China and Britain: a comparison.
Engineering Geologist 25, 135175.
Derbyshire, E., Meng, X.M., Wang, J.T., Zhou, Z.Q., Li, B.X., 1995. Collapse
loess on the Loess Plateau of China. In: Derbyshire, E., Dijkstra, T., Smalley,
I.J. (Eds.), Genesis and properties of Collapsible Soils. Kluwer, Dordrecht,
pp. 267293.
Fredlund, D.G., Morgenstern, N.R., 1976. Constitutive relations for volume
change in unsaturated soils. Canadian Geotechnical Journal 13 (3), 261276.
Gibbs, H.H., Holland, W.Y., 1960. Petrographic and engineering properties of
loess. US Bureau of Reclamation, Engineering Monograph, vol. 28 (37 pp.).
Holtz, W.G., Gibbs, H.J., 1951. Consolidation and related properties of loessial
soils. American Society for Testing Materials. Special Technologies 126, 933.
Holtz, W.G., Hilf, J.W., 1961. Settlement of soil foundations due to saturation.
Proceedings of the 5th International Conference on Soil Mechanics and
Foundation Engineering, vol. 1, pp. 673679.
Jennings, J.L., Knight, K., 1975. A guide to construction on or with material
exhibiting additional settlement due to collapse of grain structure. Sixth
Regional Conference for Africa on Soil Mechanics & Foundation Engineering,
Durban, South Africa, September.
Lawton, E.C., Fragaszy, R.J., Hardcastle, J.H., 1989. Collapse of compacted
clayey sand. Journal of Geotechnical Engineering, (ASCE) 115, 12521267.
Lloret, A., Alonso, E.E., 1980. Consolidation of unsaturated soil including
swelling and collapse behaviour. Geotechnique 30 (4), 449477.
Lutenegger, A.J., Hallberg, G.R., 1988. Stability of loess. Engineering Geology
25, 247261.
90 M.S. Nouaouria et al. / Engineering Geology 99 (2008) 8590
Maswoswe, J., 1985. Stress path for a compacted soil during collapse due to
wetting. Ph.D thesis, Imperial College, London.
Pecsi, M., 1968. Loess. In: Fairbridge, R.W. (Ed.), The Encyclopaedia of
Geomorphology. Reinhold, New York, pp. 674678.
Prospero, J.M., Ginoux, P., Torres, M., Nicholson, S.E., Gill, T.E., 2002.
Environmental characterization of global sources of atmospheric soil dust
identified with the Nimbus 7 Total Ozone Mapping Spectrometer (TOMS)
absorbing aerosol product. Reviews of Geophysics 40, 1.
Rathjens, C., 1928. Loess in Tripolitanien Zeit. Ges. Evdkunde Zu Berlin,
pp. 211228.
Rogers, C.D.F., Dijkstra, T.A., Smalley, I.J., 1994. Hydro-consolidation and
subsidence of loess: Studies from China, Russia, North America and Europe.
Engineering Geology 37, 83113.
Scheidig, A., 1934. Der Loss und seine geotechnischen Eigenschaften. Steinkopf,
Dresden. 133 pp (in German).
Sheeler, J.J., 1968. Summarization and comparison of engineering properties of
loess in the United States. Highway Research Record (212), 19.
Smith, B.J., Whalley, W.B., 1982. Observations on the composition and mineralogy
of an Algerian duricrust complex. Geoderma 28, 285311.