Risk Factors for Landslide in Penang, Malaysia l ...

39 downloads 0 Views 4MB Size Report
Penang Island is selected to be area of interest in landslide hazard analysis ... The study area for this paper is Penang Island (Figure 1) of Malaysia due to high.
Risk Factors for Landslide in Penang, Malaysia Md. Kamrul Hossainl, Habibah Latehl, Lea Tien Tat' afi,Anton Abdulbasah Kamill lschool of Distance Education, Universiti Sains Malaysia, 11800 USM, Penang,Malaysia. 2school of Electrical Electronic Engineering, Universiti Sains Malaysi a,l4300Nibong Tebal, Penang, Malaysia.

Abstract Landslide is a caustic natural threat in Malaysia. Geological factors, morphological factors, physical factors and factors associated with human activity are the causes of landslide. The knowledge of risk of the factors for happening landslide would be help to predict on landslide. The landslide causative factors - elevation, slope gradient, slope aspect, curvature, land cover, vegetation, distance from road, distance from stream, distance from fault line, geology, soil texture and precipitation with different classes are considered in the study. Risk of happening landslide is higher when the aspect is NorthEast and less when it is flat. Elevation in the range 400 to 480 is in the higher risk. The accnracy of the relative risk (RR) method is 75.08% in mapping thehazard. Keywords: Risk of Landslide, Penang, Prediction. Corresponding e-mail : kamrulstat@gmail. com

l.Introduction Landslides are geological phenomenon that involves movement of rock, earth or debris due to soil erosion. Landslides arehazardous as they destroyproperties and even claim human lives.

In Malaysia, landslides happen rather frequently due to heavy rainfall during annual monsoons mainly known as Southwest Monsoon from late May to September, and Northeast Monsoon from November to March. These recent years from 2000 to 2009, damages due to landslides have been particularly high [1]. Although it is difficult to predict a landslide event in space and time, an area may be divided into different ranks according to the degree of potential hazard due to mass movement [2]. The identification of high risk areas is important in landslide prediction and warning system. In this paper, Penang Island is selected to be area of interest in landslide hazard analysis as it has suffered numerous damages due to landslide in recent years. In recent years, there have been many studies on landslide hazard evaluation using GIS and various evaluation techniques. Relative risk is a method to measure the risk to occur a case in presence of a factor. In this paper, the method is used to measure the risk of occurring landslide in present of considered factors. From the available literature, relative risk approach is used for the first time to measure the risk factors of the landslide.

1i5

and dis

2. Study areas

in digit

of Malaysia due to high The study area for this paper is Penang Island (Figure 1) 5o15',N to 5o 30'N of latitudes frequency of landslide over the years. It is located about of the Island is varies from 290 C and 1000 10,E to 1000 20'E of l0ngitudes. Temperature to32o C and annual rainfall from2254 to 2903 mm'

Table ,

Fig 1: Penang Island 2.1. Data set sensing and GIS methods' Various Data was collected through the application of remote departments of Malaysia such as images of Penang lsland were obtained from different krigation and Drainage' Department, Department

of Malaysian Meteorological Department of Survey Department, Department of Agficulture, Minerals and Geoscience and Mapping Malaysia and Pe GIS Center'

Most landslides happed in the hilly area

in the Penang

Island are collected and

factors such as aspect' transformed into spatial database for analysis. Landslide-causative distance from drainage, and curvature, geology, soil texture, land use, elevation, slope, in calculating probability of distance from road, fault and vegetation were considered model (DEM) were constructed landslides. From topographic database, digital elevation Elevation was then with a resolution of 5-meter. From DEM, elevation was extracted' curvature' Distance from drainage used to compute slope angle, slope aspect and slope

116

Note:

and distance from road were calculated from drainage map and road map each available in digital map respectively. .;

Table

]S

C

l: Factors considered in the study

Factor

Class of factor

Factor

Flat

Factor

Forest

North

Feringgi granite (FG)

Vegetation

North-East

Clay sand granite Transport

(csc) Geology

East

Sungai ara grarile Settlement

(sAG) Batu maung granite

South-East South

(BMG)

Land

Tanjung bunga granite

use

Industry

(rBG)

West

Forest, plant

Plains, hills

North-West

Bushes

Religious area

Sandy clay (SC)

Swamp

Sea,lake.

Public utilify

Mixed farm

(scL) Vegetation

Sandy loam (SL)

river

Govemment

institution

Fruit farm

Silty clay (SiC)

lt8

Urban land (UL)

Vegetable farm

It 80

Others

ge 16 and,lt24

Nothing

ge24 andlt32

ge 80 and

ge 160

lt

ge 8 and

160

atdlt240

ge240 andlt320 ge 320 and

lt 400

ge 400 and

lt 480

ge 480 and

lt 560

Slope

Convex Curvafure

ge 32

Concave

lt

16

atdlt

ge 40 and

lt 48

ge 48 I

It 50

11100

lt 100 ge 100 lt 150

ge 200

ge 50

It 50

ge 100 It 200

lt 300

lt 100 ge 100 lt 150

ge 150 lt 200

ge 300 lt 400

ge 200 1t250

ge 400

ge 150 lt 200

ge 250

lt 300

ge 300

lt 350

ge 50

ge 200 1t250

Road

ge 250

lt 300

ge 350 lt 400

ge 300

lt 350

ge 400

ge 350 lt 400

ge 450

ge 400

40

Flat

ge 560

l,*.",.

Cemetery

South-West

Sandy clay loam

Elevation

Class of factor

(FMG)

Aspect

Soil texture

Class offactor Face to micro granite

lt 450

fault

lt 500

ge 500 lt 600 ge 600 lt 700 ge 700 lt 800 ge 800 lt 900 ge 900

lt

lt 450 ge 1100

ge 450

Notes. ge is refers to greater than and lt refers

117

to

less than

1000

3. Methodology

Relative risk (RR) is used frequently for statistical analysis when the outcome is binary and has low probability to occur. This statistical tool is suitable to use for estimating the risk of happening landslide, since this situation satisfied both conditions that is outcome is binary and low probability to occur. Let us consider the

following2X2

contingency table

Landslide occurred

Landslide did not occurred

Factor present

A

b

Factor absent

C

d

Then the relative risk (RR) of the factor to occur the landslide is

RR=

a/(a+b)

cl(c+d)

4. Results and Discussion

Figure 2 is showing the relative risk among the different direction of aspect. The slopes which are in the North-East direction have the maximum risk whereas flat type slopes have the minimum risk to occur landslide. kr Figure 3, the convex curvature that is upwards direction slopes are more on risk than the others types. _-?.

-iE-

1-?

-3E- 1-1 t.5E- i-3

lE-

13

1.5E- 13

lE-i3 -iE- i 4 0

$trltr EEfn{E+## [email protected]@@m

r Conver I Cotcave Flat

oH=b zaSz

Fig 2:

Relative

risk of

Fig 3: Relative risk of curvaturefor occuruing landslide

aspect

occurring landslide

118

?E-71

+E- 13 _r.

5E-

13

1.5E 11

_3E- 13

t. _iE- l3 1E-11

lE-1-3 1.-\E-13

5E-12

1E-1-3

-sE-11

0

0

F\{G

FG TBG SAGB\IG

Fig 4: Relative risk of

T-------

Geology

-------T-*-----------:

5L sic

CSG

UL

Fig 5: Relative risk of Soil texture for

for

occuruing landslide

occurring landslide

Figure 4 is to present the risk of occurring landslide according to the granite type. The Feringgi granite (FG) is the most risk area of Penang island than others t5pe's granite area. The relative risk of Sandy clay (SC), Sandy loam (SL), Silty clay (SiC), Urban land (UL) type soil texture is very low than the Sandy clay loam (SCL) type soil texture (Figure 5). 6E-13 4E-13

.!.rj-ts '

0

,'1 ./

'V:-', tE-11

f

J

G -j----------:------

------

r 8162+32404856

)r

l ()

i

s0

Fig

Fig 6: Relative risk of slope for

160 140 -310 .+00 -180 -i60 6.+rl

7: Relative risk of elevation

Jbr

occurring landslide

occuruing landslide

Figure 6 is showing that when the slope angle is less than or equal 80 the risk of happening landslide is low. As the degree of slope angle increased, the risk of occurring landslide also increased. Risk of occurring landslide has increased until the elevation increased to 480. After the point 480 of elevation, the risk of occurring landslide decreased very sharply. The shortage distance from drainage of a point has higher risk compare to a long but distance from road of a point has shown an inverse relation (Figure 8 and Figure 9).

119

l

fr

1f

0

L

)E-12

H8ff= Fa{

o vt=

ca,

+

350

=f,

Fig 8: Relative risk of distancefrom Drainage for occurring landslide

.1.

,5E-

450

Fig 9: Relative risk of distancefrom road for occurring lands lide

1i

1.{E-11

i.tE-

-tE-1-l

il

1E-11 8E-13

3.-iE-1-1 3E-1-?

6E-1-1

I iF-] t

,lE- 13

lE-1-1

lE-1-3

0

1.,5E- 13

.*.t-.*{:**tsl,":.** -:'

iE-13 a

c{

ra 'j-

t..t

Landslide hazar several categoni Highly hazardou 10% (80-e0%1.: the remaining 6tr

-\$.t:sS' -:i"

Fig 10: Relative risk offaultfor

Fig I l:

occurring landslide

occurring landslide

Relative risk of land usefor

5. Conclusion

In Figure 10, the relative risk of happening landslide is low when the point is not far from 100 meter from the fault line. Figure 11 shows that forest and plant area of the

Penang Island are the most risky zone for landslide occurred. This result is similar with Lee et al. [3].Using the risk scores of all the factors, the followinghazad,map (Figure 12) was created which has 75Yo accuracy to indicate landslide.

Due to the limita housing and der e

to warn people othe risk of indir hazardmapping: be used as a too^ urban infra-strucr. References

lll

Pradhan. ar Island, Mala

I2l t31

network moc Varnes, D.. practice. -\ a;; LEE, S,,

CH

verification 131. 120

-.-

u

Fig t2: Hazard Map of Penang Island Using kR Score Landslide hazard mapping is done by segregating LHI (landslide hazard index) into several categories of risks. In this study, the indexes are categorized into 4 groups: Highly hazardous for the highest 10% of the indexes (90-100%), Hazardous for the next 10% (80-90Yo), moderately hazardous for the next 20% (60-50%) andNot hazardous for the remainin g 60% (0-60%). 5. Conclusion Due to the limitation of flat land in Penang Island, more hill areas are being explored for housing and development. This creates a scenario where landslide alert system is needed to warn people of impending landslides on risky area. The relative risk methods provide the risk of individual factors as well as the scores from the method is also provided a

hazardmapping for the Penang Island. The accuracies show that relative risk method can be used as a tool for hazard mapping and these maps can be useful in planning of future urban infra-structures development. References

11]

Pradhan, and Lee, S. (2010). Delineation of landslide hazard areas on Penang Island, Malaysia, by using frequency ratio, logistic regression, and artificial neural network models, Environmental E arth Sciences, 60(5), 1 03 7- 1 0 5 4, 20L0.

t2l

Varnes, D.J. (1984). Landslide hazard zonation: a review of principles and practice. Natural Hazards, 3. LEE, S., CHOI, J., and Min, K. (2002), Landslide susceptibility analysis and verification using the Bayesian probability model, Environmentol Geology,40, L20-

t3l

131.

t2t