Natural Hazards 12: 101-118, 1995. © 1995 Kluwer Academic Publishers. Printed in the Netherlands.

101

Landslides and Stability of Coastal Cliffs of Cox's Bazar Area, Bangladesh MD. H A M I D U R

R A H M A N and Y O U N U S A H M E D K H A N

Department of Geology and Mining, University of Rajshahi, Rajshahi 6205, Bangladesh (Received: 20 January 1993; in final form: 21 December 1994) Abstract. Successive applications of two-dimensional circular slip limit equilibrium analysis of slices

have been made to the three sections of the Cox's Bazar Coastal Cliff. These three sections have been found to be stable (S), quasistable (QS), and unstable (U) with factors of safety (Fs) of 1.269 -+ 0.01, 1.13-+ 0.0504, and 1.01-+ 0.058, respectively. In the quasistable and unstable sections, the slopes suffered shallow slidings with slip circles parallel to the surface of slopes. The main causes of these slidings are steepness of slopes, lack of vegetation cover, erosion of toe accumulations by run-off water and waves during high tides. The influence of the slope geometry and the material's strength properties of slopes on safety factors are evaluated. The steeper cliffs reach an unstable condition faster than the more gentle cliffs as their heights change by an equal amount. Again, the high cliffs reach an unstable condition laster than the lower cliffs as slope inclinations change by an equal amount. The safety factor increases with a corresponding increase of friction angle (0B°) and vice versa. Key words. Landslides, slope stability, Cox's Bazar coastal cliff, strength properties, factor of safety, charas and khals, erosions, slope geometry. Abbreviations. BK, Bahar chara to Kalatali chara; KR, Kalatali chara to Reju khal; RI Reju khal to

Boro Inani khal; Fs, Factor of Safety; S, Stable; QS, Quasistable (semistable); US, Unstable.

1. Introduction Bangladesh occupies a major part of the Bengal Basin and landslides of different magnitudes occur along its coastal regions. The Cox's Bazar coastal cliff is one of these regions. The different sections of the 20 km coastal cliff of Cox's Bazar suffer many landslide problems. The studied area (Figure 1) belongs to the folded flank region and is comprised of loosely consolidated sandstone. The Cox's Bazar coastal cliff suffer from weathering and erosion which cause many shallow failures of the loosely consolidated sandstones. These are threatening buildings and roads in the Bahar chara area. The agricultural land and houses of local villagers have been badly affected by these slidings. North of Kalatali village, the boundary wall of the local power supply station was partially damaged by landslide activity. Houses and prawn culture farms are affected by slides and incur considerable financial losses. Every year, the houses and farms must be rebuilt as the cliff materials slide over the temporary structures. Himchari and Sonapara villages are the main affected area. Between the areas of Kalatali and Reju khal a cliff section of nearly 10 km

MD. H A M I D U R R A H M A N A N D Y O U N U S A H M E D K H A N

102 91"145 '

--

.

! 92" 00'

^'lChitla9ong

i 92"15'

s

I"

(D-,,---'5 ~ù,ù~ù,oV]" L h

Z2°OO'

, ,,"

,~--.

~'~5 ,.:~

[

/ Boh«rCMi;l:mZOrX,~"~~j:

",

[

h,

Section-BKl "'~~,.~:'-" ~°gRT~~, "I..) LKol.to,i Chor~ -"4~~, " Section

-KR -

Section--RI

,-"

"21°

{o'~ox's~_ ° R a m u

Im«horl eh,01

I~~ . . . . r~ "

._

."~NO,A

0

~.

~_~

e~e° r-

~AùGLAO~S"

« J

.F,

• ' Moheshkholi) ] 2|°30 '

I.,.."%

.~~.~

%0 °

92 °

O"

N

~

Re]u Kha

. .

oroInoniKho|

|

0 'Z~OO'

°o

Area of study Boundary Line

0I

II0

20 I

30! Km

\t'" ù~cTeknaf~kl, (

-20*30'

B, ST. M A R T I N » S " ; ~ I

Fig. 1.

I

ISLANDS

]

'~"

Map showing location of the study area.

suffered 23 shallow slides during the one year of investigation. The number of landslides increased further south and 31 slides were encountered in the next 10km section from Reju khal to Boro Inani khal. About 20% of the agricultural land was affected by landslides near Himchari chara, although farmland in the Bahar chara area was not affected during the investigation. These unexamined cliff sections deserve detailed study of both their stability and landslide activities.

LANDSLIDES AND STABILITY OF COASTAL CLIFFS

103

In America, Edil and Haas (1980) proposed two separate criteria for interpreting the stability of Chicago's Lakeshore Bluff. First, the critical circle criterion, which defines the failures in terms of a circle having the smallest safety factor, and evaluates processes causing successive small rotational slides which are retrograde from toe to cliff top. This approach is applicable to low inclination slopes (/3° < &) or slopes of material with significant cohesion intercept. Second, the unstable circle criterion, which is applicable to steeper slopes (/~o > ~bo)or slopes of material with low cohesion and often predicted deeper slips with a limited sequence of retrogressive failures.

2. Environmental Context The study area is situated within the Cox's Bazar Sadar and Ukhia thana in the south-western part of the Cox's Bazar district. Geographically, the area lies between latitude 21°12'N and 21°30'N and longitude 92°0'E and 92°05'E. In the Cox's Bazar cliff section, the Neogene rock sequences are well developed and due to the predominance of sandstones, the whole vertical stratigraphic section is termed 'Cox's Bazar Sandstone' (Hossain, 1978). Rocks throughout the cliff sections are more or less similar in lithology, consisting mostly of loosely consolidated massive sandstones. The maximum elevation of the area is about 75 m above mean sea level near the Reju khal. The average elevation ranges between 35 and 75 m. The drainage system of the study area is characterized by a number of seasonal and intermittent tributaries and distributaries locally known as 'Khal' and 'Chara', respectively. The different khals and charas, which include Kalatali chara, Himchari chara, and Boro Inani khal, transversely cross the cliff. Most of them dry up during the winter season but in the rainy season they are full of water due to heavy rainfall (the average rainfall in the rainy season of 1990 (March to September) was 416 mm per day). The 25 km cliff area is situated on the western side of the low hill ranging between the Baghkhali River on the east and the Bay of Bengal on the west. The two main streamlets, Kalatali Chara and Reju Khal, cut the cliff into three sections: Bahar Chara-Kalatali Chara (BK), Kalatali Chara-Reju Khal (KR), and Reju Khal-Boro Inani Khal (RI). The distance between the coast line and the base of the cliff at the Bahar chara area is about 200 m and decreases southward to 10 m at Boro Inani khal. The cliff slope angle varies from place to place. In the northern part, from Bahar chara to Kalatali chara (section BK), the slope angle varies from 30° to 40 °, but from Kalatali to Reju khal (section KR) it ranges between 35 ° to 60 °. In the southern part, from Reju khal to Boro Inani khal (section RI), the slope angle varies from 50° to 70 °. In the section BK, from Bahar chara to Kalatali chara, landslide activity was not noticeable. Landslide activity causes several slope failures in the section KR and it increases towards the south in the section RI. Most landslides are shallow failures occurring along a plane parallel to the slope surface.

104

MD. H A M I D U R R A H M A N AND YOUNUS A H M E D K H A N

At the Boro Inani khal, waves reach the toe of the cliff slope during high tide level and wash out toe accumulations. The cliff area of section BK is covered by thick tropical evergreen types of forest and dense undergrowth which are mainly shrubs and trees. But the southern part of the cliff, sections KR and RI, support little or no vegetation.

3. Aim of Study The slopes of the studied cliff section of the Cox's Bazar coast are stable in terms of large failure. However, shallow failures of the cliff slopes in the area between Kalatali chara and Boro Inani khal (sections KR and RI) were available to study. The northern part of the cliff from Bahar chara to Kalatali chara (section BK) suffered almost no failure. These three sections of the cliff were unexamined in terms of stability. Therefore, the purpose of this work is to identify the stability status of these three cliff sections, in terms of safety factor, by the simplified Bishop method of slices, and to identify the causes of landslides. In addition, it aims to explain the relationship between slope failure activities and the variations of the different geotechnical properties of cliff materials and as weil as slope height (H) and slope angle (/3°).

4. Methodology 4.1. Field Investigations Geotechnical investigations were made in the study area during the period from March 1990 to February 1991. Information about lithology, morphology, and other conditions associated with cliff processes were also collected. The lithological information of different sections of cliff showed that the cliff materials consisted of loosely consolidated massive sandstones (Hossain, 1978). The morphological and geometrical characteristics such as height and inclination of cliff slopes, were later used in the stability analysis. The other important work of the field study was to collect samples of cliff materials for the laboratory determination of the geotechnical properties used in the safety factor calculations. Sliding phenomena and their causes were not uniformly active on the entire 25 km of exposed cliff, and they varied from one part to another. About 71 slope locations from three sections of cliff were selected for study. The slopes were selected randomly from each section. There were 18 slopes in section BK, 29 slopes in section KR, and 24 in section RI. The geometry of the slope, including slope angle, height, and toe accumulation was measured in the field by means of measuring tape, clinometer, and altimeter. In the case of failed slopes, the geometry of toe accumulation and height of the slope were scaled in a graph paper from which the slope angle of before and after the failure were estimated. The vegetation cover on the slopes was assessed by visual estimation. This vegetation

LANDSLII)ES AND STABILITY OF COASTAL CLIFFS

105

cover decreases significantly towards the south, i.e, section RI. As a result, the erosional and weathering activity were more intense in section RI than in the other two sections. This observation was supported by the presence of loose disintegrated soil over the slopes of section RI. Erosion of toe accumulations of the failed slopes was observed (Figure 2) on rainy days in sections KR and RI. The toe accumulations were also washed away by the bay waves during high tide level. 4.2. Laboratory Methods Hand-cut and bag samples were collected and were wrapped in polyethylene sheets and then given a wax coating. Hand-cut samples were used to determine strength and porosity, while the bag samples were used to determine the bulk density, water content and unit weight. Undisturbed samples were used in the triaxial compression cell and disturbed samples were used in the direct shear box tests to determine the angle of internal friction and cohesion. A total of 210 samples from 71 slopes were used in the above laboratory tests. All tests were done using the U.S. Corps of Engineers' methods, as described in Bowles (1978). The results were calculated by statistical measurements of mean and standard deviation for proper applications of the data in the analysis. 4.3. Stability Analysis Slope stability analysis of cliff is a task involving searching out a critical circle for an individual slope. The dimensionless parameter (At6) defined by Janbu (1954) is A«~ = yH(tan O)/c, where y = unit weight of the soil, H = height of the slope, and ~b,c are the strength parameters of the soll. Since the cliff consists of cohesionless (c = 0 kN/m 2) massive sandstone, therefore A«+ tends to infinity. Since A«+ equals infinity, the critical slip surface is a plane parallel to the surface of the slope and the circle passes through the toe of the slope (Duncan and Wright, 1980). Furthermore, the field observations show shallow surfacial slides along the cliff sections. Therefore, a computer analysis technique based on the modified Bishop method (Bishop, 1955) was used to find the critical slip circle parallel to the slope surface. A series of shallow slip circles was selected and a critical slip circle, the one having the smallest safety factor (Fs), was identified. The factor of safety is defined as the ratio of the available shear strength (tl) to the shear stress (7,) which must be mobilized to maintain a condition of limiting equilibrium, i.e., Fs = ~'I/"Cm.In Figure 3, the potential failure circle AC is divided by vertical planes into a series of slides of width 'b'. For any slice, the inclination

106

MD, HAMIDUR RAHMAN AND YOUNUS AHMED KHAN

U

O

.o "a ~

O

m 3

>"

oJg3~.

o

|1

.11-

~2

ùd i13 ù~

I

»

re" e,,,

~0

o

I

õ

._o ùE -.

. &-..\

,> I|

«

"% %

c4

"...'~..

i

~,.a:

~'...",~ ù~

||

."

~ o'-

%

~

•

O '%%% %

e'L

LANDSLIDES AND STABILITY OF COASTAL CLIFFS

107

Y

Centre of O.._«li_g , u r f a c «

--

.

rsin~

--

-

~ ] I

I

\\, ù "~.. , \ \, %

]

\\,\ \

Fig. 3.

Simplified Bishop method of slices.

of the base to the horizontal is ' a ' and the height measured on the center line is 'h'. Considering m o m e n t s about 'o' the sum of the m o m e n t s of the shear force z on the failure arc A C must be equal to m o m e n t of the weight (W) of the soil mass A B C D . Therefore, Ezr = EWr sin a. Now -c = rfl/Fs

(l is the length of the slice base)

or

E~'/I/Fs = E W s i n a. Therefore

108

MD. H A M I D U R R A H M A N AND YOUNUS A H M E D K H A N

Fs-

Z.rfl

_ Z ( c l + N ' tan ~b)

ZWsin «

E W s i n o~

(1)

(putting ~'j = c + o- tan ~b and o-I = N' = effective normal stress). Again, total shear forces on the base of the slice are T :---- "l'ml

= (cl + N ' tan ~b)/Fs.

Resolving the forces on the vertical direction, we have W = N cos c~ + ~-sin a, where N = total normal force = N ' + ul (u = pore pressure) and ~- = total shear on base. W = N' cos a + ul cos a + (c//Fs) sin a + (N'/Fs) tan ~bsin a. Therefore, N ' = W - (c//Fs) sin a - ul cos a cos a + (tan ~bsin cdFs)

(2)

Putting the values of N' from Equation (2) and l = b/cos a, in Equation (1) we have, Fs=

1 a~[ {cb+(W-ub) tanth} ] . ZWsin l ( 1 + tän--~äalä ~ ~ cös a ]

(3)

An initial estimate of Fs = 1 was used in the solution given in the right-hand side of Equation (3) and the calculäted Fs were used repeatedly until the difference between the values of the two successive Fs was less than 0.001. This calculation was performed with a computer program.

5. Results and Analysis The properties of eliff materials differ significantly from one section to another (Table I). Water content, density, and unit weight gradually increase towards section RI. In section BK, the water content is 10.55% with a variation of 1.05. On the other hand, it is 17.06% in section K R and 18.61% in section RI with standard deviations of 0.69 and 0.72, respectively. Density in BK is 2 -+ 0.03 mg/m 3 and in K R and RI is 2.10 +- 0.07 mg/m 3 and 2.40 +- 0.05 mg/m 3, respectively. Unit weight also varies from 17 -+ 0.66 kN/m 3 in B K to 20.00 +- 0.15 kN/m 3 in RI. The average height and angle of slopes in K R and RI is higher than in B K (Table

109

L A N D S L I D E S A N D S T A B I L I T Y O F C O A S T A L CLIFFS Table I.

Properties of cliff materials

Properties

Section B K (n = 54)

Section K R (n = 87)

Section RI (n = 72)

W a t e r content (w%) Density (mg/m 3) Unit weight (kN/m 3) Porosity (n%)

"10.55 -+ 1.05 2.00 -+ 0.03 17.00 -+ 0.66 28.20 +- 2.10

17.06 +- 0.69 2.10 -- 0.07 18.20 -+ 0.72 30.32 -+ 1.92

18.61 -+ 0.72 2.40 -+ 0.05 20.00 -+ 0.15 34.10 -+ 4.57

*Mean value + standard deviation. n = n u m b e r of samples.

Table II.

G e o m e t r y of slopes and factor of sa•ety (Fs) of cliff

Properties of slope

Section B K (N = 18)

Section K R (N = 29)

Section RI (N = 24)

Height (meter) Slope angle (/3°) Friction angle (4,°) Factor of safety (Fs)

*40.0 -+ 4.20 31.5 ° -- 3.82 40.0 ° --+ 1.20 1.269 6 0.01

56.3 -+ 6.6 38.5 ° -+ 2.5 38.5 ° -+ 0.51 1.13 +- 0.05

58.6 -+ .31 65.0 ° -+ 6.39 40.2 ° -+ 0.33 1.01 +--0.058

*Mean value -+ standard deviation. N = n u m b e r of slopes.

II). The standard deviation of the slope angle in RI is more than 5, which implies the presence of steeper slopes. On the other hand, in the northern part, BK consists of slopes with angles of 31.5 ° --+ 3.82 which is less than the angle of internal friction (40° - 1.2) of the cliff material in this section. This indicates that the slopes of KB are stable. The factor of safety (Fs) of the three sections is depicted in Table II and shows a greater variability in KR (Fs = 1.13 -+ 0.0504) and RI (Fs = 1.01 -+ 0.058) than that BK (Fs = 1.26 _+ 0.010). Variations in the values Of variables like slope height, slope angle, and unit weight of cliff materials, create relative variations of the factor of safety and affect the stability of the cliff. The factor of safety largely depends on slope angle (/3°) of the cliff and the friction angle (&) of the cliff materials. The friction angle and density of the cliff materials are interrelated. An increase in density causes a proportional increase in the friction angle of the cliff sandstones. In response to an increase in porosity from 28.20% (BK) to 34.10% (RI), the watet content increases from 10.55% (BK) to 18.61% (RI). Again unit weight is directly proportional to the density of the cliff materials, since density increases from 2 to 2.4 mg/m 3, the respective values of unit weight also increase from 17 to 20 kn/m 3 (Table I). These increased values of water content and unit weight are orte of the causes in reducing the factor of safety (Fs) of the cliff RI. Detailed interpretations of these results and analysis are discussed next.

110

MD. HAMIDUR RAHMAN AND YOUNUS AHMED KHAN

6. Discussion Results of the stability analysis, supported both by computed safety factors (Fs) and field observations, indicate that the slopes of KR and RI are unstable in shallow surfacial slidings. The 'critical circle' criterion, developed by Edil and Haas (1980), which is defined as the failure circle having the smallest safety factor, indicates the most favorable surface of sliding. In KR, after sliding, the landslide debris accumulation which provided support to the toe of the cliff was quickly removed by run-off watet and erosion. This process altered the slope geometry and reduced the factor of safety. The new geometry is orten proved to be unstable causing the new slope to be subjected to sliding along the next critical circle. Consequently, the cliffs tend to retrogress from toe to top by means of successive rotational slides. These rotational, shallow small slides continue until a safety factor greater than 1.07 is obtained and a stable slope configuration is reached with Fs > 1.25. In the case of BK, slopes of 80% vegetation cover with safety factors less than 1.25 were observed as stable and this was possible because the vegetation cover suppressed erosion by run-off water on the slopes. On the other hand, the slopes of KR and RI with less than 30% vegetation cover suffered many successive shallow slidings, even though the calculated safety factors were greater than 1. Across the whole coastal area, stable slopes have a safety factor greater than 1.25 and for semistable to unstable slopes the factor of safety is equal to or less than 1.25. On the basis of the critical circle analysis, the three studied sections, namely BK, KR, and RI, of the cliff were termed as in Figure 4: Stable sections have an average Fs = 1.269 --- 0.010, quasistable sections have an average Fs = 1.13 + 0.0504, and unstable sections have an average Fs = 1.01 + 0.058. The key to the cliff stability was an absence of erosion of the toe accumulation after sliding. This suppressed further successive slidings which then allows the slope angle to become gentler (<30 °) and to reach a stable state with average Fs = 1.269 +- 0.010.

6.1. Effects of Different Parameters on Cliff Failures 6.1.1. Effects of slope geometry. The influence of cliff height on safety factor is largely dependent on the angle of slope (Edil and Villejo, 1980; Kachugin, 1963). High elevation cliffs are relatively unstable compared to low elevation cliffs with the same inclination (Figure 5). It is evident that cliffs of more than 60 m elevation are quite unstable at any slope angles greater than 40° (Figures 5 and 6). Cliffs of less than 60 m height with 45 to 65 ° inclination are quasistable (1.07 < Fs < 1.25), which indicates that there exists several slip surfaces along which the failures may or may not occur. Similar slopes of more than 65° are no longer quasistable, and they are entered into the unstable section (Fs ~< 1.07). Slopes of less than 50 m in height are stable up to an inclination of 45 ° (Figure 6).

LANDSLIDES AND STABILITY OF COASTAL CLIFFS

111

92"°00 '

N

T

BAZAR

I t

I I I« I S S

% % S.

.20 ~

i5onapara • • t

i i

t

I

-

- -

I

-15' %

k

I

Boro

S'[ABL E SECT.ION

\,

QUASI STABLE SECTIO ,

~ ' ~ UNSTABLE SECTION

/ ~,v, I o

21

0

I MIle

t l

Fig. 4.

•

•

I

=

%

%

% I t

Map showing three sections in terms of stability of the Cox's Bazar coastal cliff.

L'.

112

MD. HAMIDUR RAHMAN AND YOUNUS AHMED KHAN

1.55

C:O KN/ m2 (~"40 o

1.5

Failuresnever occur (s~oble) ~

V:2OKN / m3

1.25 1,2

u. 1.15 a

doocc~ur (~ H% i $~obl Q ~~ u~e~)o s =:50m

Foilures

I.I

o

o 1.05 O

o h_

Foiluresore c

o

m

~

0.95 0.9 0.85

I

35

I

I

I

I

l

|

45 55 65 Slope angle(B} in degree$

|

¢

75

«

Fig. 5. Influence of slope height on cliff stability.

So, the slope height and inclination are dependent on each other in evaluating the cliff safety factor. The steeper cliffs can reach an unstable condition faster than m o r e gentle cliffs as their heights change by an equal amount. Again, the higher cliffs reach an unstable condition faster than lower cliffs as slope inclinations change by equal amounts.

113

L A N D S L I D E S A N D STABILITY O F C O A S T A L CLIFFS

1.55[

C=0 KN/m2 ~;=20KN/m5 ~~'=40=

~ ~

never occur( sloble )

1.3 I.?.5

\

1.2 03 ~.

Failuresd

1.15

\

I.|

o

Io

°

, o~~o\ ~ ~ o

50

*

40

O

="

0

1.05

0

O b.

Foil=ure 0.95

•

0.9' 0.85| Fig. 6.

i

20

l

i

I

i

40 60 SIope height(H] in meter

m

80

Influence of inclination on cliff stability.

6.1.2. Effects of strengthparameter. Figure 7 demonstrates the influence of friction angle (~b°) on the factor of safety with different values of slope inclinations (/3°) and height (H). In all cases, the unit weights (y kN/m 3) are the same (20 kN/m3). The factor of safety (Fs) increases from 0.97 to 1.0 for a 70-m high slope and 40 ° inclination, with a corresponding increase of ~ values from 33 to 40 °. Therefore, the safety factor increases with a corresponding increase of friction angle (~b°) and vice versa.

114

MD. H A M I D U R R A H M A N AND Y O U N U S A H M E D K H A N

1.55

1.5

Failufresneveroccur,(stable)

~

C=OKN/m2 Y=2OKN/m5

1.25 =

1.2 0"J

Faitures do occur (Quasi s'mble}

IL

\.\\

1.15 0

H=40m

I.I

0 t-

4?.O 1.05

o 0 14-

~

40o

0

Failures are common(Uns~oble)

H=60m

0.95

0.9 0.85

I

:55

I

I

45 Slope angle

Fig. 7.

I

I

55

(B) in

I

65

I

l

•

75

degren

lnfluence of angle of internal friction of cliff materials on stability.

6.1.3. Effects of material properties. Besides strength properties, the stability is also influenced by material properties like unit weight (y), density, and porosity

(n%). Figure 8 displays a graph of a unit weight of 17 kN/m 3 and of a saturated unit weight of 20 kN/m 3 with constant ~ßo values for slope angles differing from 40 to 80°. For low slope angles, the influence of unit weight may be negligible. However, for high angles, it certainly influences the slopes. A slope of 70 ° inclination with

115

LANDSLIDES AND STABILITY OF COASTAL CLIFFS

1.55

N~

1.3

C = 0 KN/m 2 (~ = 4 0 =

Foilures never occur (stoble)

1.25

_'~ u) b.

\\ Foitures do occur(Quosi-s

"-" 1,15 q.0

ù..

I.I

0

õ ü

1.05

p =~C

i Foilures or 0.95

p =70 ° 0.9

0.85 ' ZO

'

40 Slope

Fig. 8.

'

6 0'

80

h e i g M ( H ) in m e t e r

Influence of unit weight of materials on stability.

materials of unit weight of 17 kN/m 3 is stable at less than 30 m in h e i ß t . But slopes of the same height are no longer stable with a unit weight of 20 kN/m 3, The sandstone cliff of the Cox's Bazar is more or less porous (porosity 27 to 35%) in nature. The upper cliff is more pervious than the lower cliff when the cliff tops are weathered due to the absence of sufficient vegetation cover. In a weathered top, the water easily infiltrates and, during infiltration, the rainwater

116

MD. H A M I D U R R A H M A N AND YOUNUS A H M E D K H A N

removes the cementing particles from the upper weathered cliff zone and gradually reduces its strength. 6.1.4. Effects of environmental parameters. Here in the study, the environmental parameters include weathering, vegetation cover, seepage of water and tidal streams in sections KR and RI. The weathered slopes, devoid of sufficient vegetation (less than 30%), have comparatively loose materials especially at the slope's top. Therefore, after rainfall, the loose materials wash out down slope with the run-off watet. Infiltrating water increases the pore pressure and decreases the shear strength (Rahman, 1988, 1989, 1990; Hossain and Dutta, 1986; Hasan, 1981) and the internal friction of the materials (J. Rupke et al., 1988). The rainwater easily percolates along the fractures and tensional cracks causing a reduction of strength and shallow slides occur. On the other hand, in BK and the northern part of KR, slopes with more than 80% vegetation cover give additional strength to hold the sandstone intact from the roots of trees and grasses. Seepage forces within the lower or basal part of the cliff are sufficient to dislodge mineral grains from such poorly cemented sandstones, especially in KR. This seepage gradually reduces the support offered by the layers of overlying sandstones which, as a result, commonly collapse as some sott of slide. The tidal streamlets have eroded shallow but steep slopes in sandstones, which are mainly uniform in grade. The resulting drainage of water stored in this sand is sometimes sufficient, between high and low tide, to initiate movement in the cliff which has been over-steepened by erosion. Such types of movements have been observed in most of the Charas in the southernmost cliff near Reju Khal and Inani Khal.

7. Conclusion On the basis of slope stability analysis with critical circle and field observations of landslides, the three studied sections of the Cox's Bazar coastal sandstone cliff are evaluated in terms of factor of safety. The three sections, BK, KR, and RI, of the cliff are found to be stable with a factor of safety Fs = 1.269 --- 0.010, quasistable with Fs = 1.13 + 0.0504, and unstable with Fs = 1.01 --- 0.058, respectively. (a) Stable section: The cliff area from Bahar chara to Kalatali chara is found to be stable with a factor of safety of 1.269 ± 0.010. The key to the cliff stability is the absence of erosion of the toe accumulation after sliding. This suppresses further successive slidings and finally allows the slope angle to become more gentle (<30 °) and to reach a stable state with Fs = 1.269 +_ 0.010. Presently, the slope surfaces of this part are carpeted with vegetation, allowing little or no scope to erosional agents like run-off water and weathering. (b) Quasistable section: The 10 km long area from Kalatali chara to Reju khal is quasistable with a Fs value of 1.13 +- 0.0504. This section contains a few surfacial

LANDSLIDES AND STABILITY OF COASTAL CLIFFS

117

slidings near Kalatali chara. The stability of this section may fail with devegetation and oversteepness of the slope. (c) Unstable section: The area from Reju khal to Boro Inani khal in the south, about 10 km long, is found to be an unstable part of the cliff. The safety factor of this part is 1.01 + 0.058. This part is currently suffering shallow sliding parallel to the surface of the slopes. These are successively occurring from the toe to the top, while the slide masses of the slopes are washed away by run-off rainwater and/of ocean waves. As a result, steepening of the slopes continues allowing further sliding in the area. The main factors of the instability of the southern part of the cliff are steepness of slopes, lack of vegetation cover, erosion of toe accumulations by run-off water and by bay waves (during high tide level). Steeper cliffs reach an unstable condition faster than more gentle cliffs, as their heights change by an equal amount. Again, high cliffs reach the unstable condition faster than lower cliffs as slope inclinations change by an equal amount. Finally, it is suggested that the proper utilization of land associated with the cliff demands mass revegetation programs along with earthworks to reduce the cliff slopes, especially in the semistable and unstable sections.

References Bowles, J. E.: 1978, Engineering Properties of Soll and their Measurements, 2nd edn. McGraw-Hill, New York. Bishop, A. W.: 1955, The use of slip circle in the stability analysis of slopes, Geotechnique 5(I), 711. Duncan, J. M. and Wright, S. G.: 1980, The accuracy of equilibrium methods of slope stability analysis, in: S. L. Koh (ed.), 'Mechanics of Landslides and Slope Stability', Eng. Geol. 16(1), 5-7. Edil, T. B. and Haas, B. J. : 1980, Proposed critefia for interpreting stability of lakeshore bluffs, in: S. L. Koh (ed.), 'Mechanics of Landslides and Slope Stability', Eng. Geol. 16, 97-110. Edil, T. B. and Villejo, L. E.: 1980, Mechanics of coastal landslides and the influence of slope parameters, in: S. L. Koh (ed.), 'Mechanics of Landslides and Slope Stability', Eng. Geõl. 16, 8396. Farquhar, O. C.: 1980, Geologic processes affecting the stability of rock slopes along Massachusetts highways, in: S. L. Koh (ed.), 'Mechanics of Landslides and Slope Stability', Eng. Geol. 16, 135145. Hasan, A. K. M. S.: 1981, Slope instability and construction damages at Mercantile Marine Academy, Chittagong district, Bangladesh, Records Geol. Surv. Bangladesh 3(1), 1-21. Hossain, K. M.: 1978, A study of the lithology and condition of deposition of the sediments of Cox's Bazar cliff, Abstract, 3rd Bangladesh Science Conference. Section V, Geology-Geography. Hossain, K. M. and Dutta, D. K.: 1986, Slope stability problems of the Chittagong University Campus, Dhaka Univ. Stud. B34(1), 77-89. Janbu, N.: 1954, Stability of Slopes with Dimensionless Parameters, Harvard Soll Mech. Ser. 46, 81. Kachugin, E. G.: 1963, Some patterns in the reworking process of reservoir bank, in: I. V. Popov and F. V. Kotlov (eds.), The Stability of Slopes, Consultants Bureau, New York, Vol. 35, pp. 30-38. Quigley, R. M., Gelinas, P. J., Bou, W. T., and Packer, R. W.: 1977, Cyclic erosion-instability relationships: lake Erle north shore bluffs, Can. Geotech. J. 14, 310-323. Rahman, M. H.: 1988, Landslides and slope stability problem of the non-homogeneous embankment slope along the Padma river bank in Rajshahi city, J. Inst. Bangladesh Studies 11, 189-202. Rahman, M. H.: 1989, Analysis of slope stability problem of the non-homogeneous embankment slope

118

MD. HAMIDUR RAHMAN AND YOUNUS AHMED KHAN

by finite element method (FEM), Proc. of the Internationalconf. on Eng. Geol. in Tropical Terrains, June 26-29, 1989, University of Kebangsaan, Malaysia, Bangi, Malaysia. Rahman, M. H.: 1990, Analysis of slope stability problem of the heterogenous slope by finite element method. Proc. of the 6th International congress of IAEG, August 6-10, 1990, Amsterdam, The Netherlands, pp. 2279-2285. Rupke, J. et al.: 1988, Engineering geomorphology of the Widentobel catchment, Appenzell and Sankt Gallen, Switzedand. A Geomorphological inventory system applied to geotechnical appraisal of slope stability, Eng. Geol. 26, 33-68.