Landslide Risk Management of Patong City: Demonstration of

Geotechnical Engineering Approach

Suttisak Soralump

Geotechnical Specialist

Damrong Pungsuwan Researcher

Mananya Chantasorn GIS Analyst

Nattawuth Inmala

Graduate Student Geotechnical Engineering Research and development center, Civil Engineering Department, Kasetsart University N.M.S.I. Alambepola Asian Disaster Preparedness Center

Keywords: landslide, landslide susceptibility map, landslide risk map, landslide management, antecedent precipitation index

ABSTRACT: Patong city located in Phuket province which is in the southern part of Thailand, it is one ofmost desire tourist destination in Thailand. The development of the city occasionally caused flooding andlandslide. Landslide risk management program is introduced by Asian Diaster Preparedness Center (ADPC)under RECLAM II project. Geotechnical approach was used to produce landslide susceptibility map andlandslide risk map. Buildings in the landslide hazard area were classified into various risk classes. Man made factors such as the road, area of prohibition by law and others have been considered as a factor for landslidehazard mapping. Engineering practice handbooks were specification technique that will minimize the triggering of landslide. Critical API was calculation and the appropriate warning method was adapted for local personnel to use as critical for landslide warning. All the works mentioned were closely done togetherwith the Patong Municipality in order to customize these mitigations for the person who are in charge. Theproject is perhaps the first landslide risk mitigation project in Thailand.

1 INTRODUCTION

Landslide is one of natural hazard that affected Thailand. The direct economic lost due to landslide is calculated to be equal to 100 million Baht per year and the return period of large area landslide is once in every 3-5 years (Soralump, 2010).

Patong beach, located in Phuket province, is one of the famous tourist destinations in Thailand. This beautiful beach is surrounded by the mountain range. Since the city is expanding, the use of steep slope area on the mountain is unavoidable. This causes the disturbance to the environment and sometime triggers the landslide. Asian Disaster Preparedness Center (ADPC) together with The Norwegian Geotechnical Institute (NGI) had the responsibility for execution of Regional Capacity Enhancement for Landslide Impact Mitigation program (RECLAIM) which the funding was provided by the Royal Norwegian Embassy in Bangkok. Department of Mineral Resources and

Kasetsart University were asked by ADPC to take responsibility for the implementation which decided to demonstrate the landslide mitigation in Patong.

2 STUDY AREA

Patong Municipality is approximately 16 square kilometers in area and located on the west coast of Phuket Island. Population mostly consists of tourists and Muslim communities living in areas near the beach. However, after the tsunami event the development tends to move higher on the mountain. Changing of the land use and land cover caused the landslide in the area (Pungsuwan, 2006) such as shown in Figure 1.

Table 1 shows some event of landslide records gathered by Patong municipality, it shows that most of the landslide has triggered by excessive rainfall.

Figure 1: Na Nai roadside landslide events on October 25, 2007. Table 1: Landslide records gathered by Patong Municipality. No. date location Triggering factor

1 October 19, 2001

Various places

Heavy rainfall

2 October 21, 2003

50th anniversary

Rd.

Cut slope and Heavy rainfall

3 October 14, 2004

Na Nai village

Blockage of drainage and Heavy rainfall

4 October 14, 2004

Kalim village

Inappropriate drainage and Heavy rainfall

5 October 25, 2006 Na Nai Rd. Cut slope non protection

and Heavy rainfall

6 July 15, 2007

50th anniversary

Rd. Under construction

7 September 5, 2008

50th anniversary

Rd. Drainage

8 September 19, 2009

Kathu-Patong Rd. Heavy rainfall

3 GEOLOGICAL CHARACTERISTICS

Geological characteristics of Patong include formation of granite in Cretaceous Era and modern beach sediments in Quaternary period. Granite rock is found to be in moderately to highly weathered condition with various sets of joint and fracture. Left lateral strike-slip fault has found mostly with their strike lies between north-east to south-west direction. Department of mineral resources has interpreted the geologic structures of Patong as shown in Figure 2 (Department of Mineral Resources, 2006).

Figure 2: Geological characteristics, Patong (Department of Mineral Resources).

4 SLOPE STABILITY ANALYSIS

Multi-Stage Direct Shear Test were done using the residual soil samples collected in the study area as well as the data gathered from previous study (Soralump and Torwiwat, 2007). Finally, soil data from 12 locations were used for the analysis in this study. The drained direct shear tests were done to the soaked (almost saturated) samples. In order to ensure the drained behavior, pressure sensor, capable of measuring both positive and negative pressure, was embedded in the top cap to monitor the change in pore pressure. The excess pore pressure is only exist during consolidation stage but not the shearing stage. This can be concluded that the soil sample is sheared under fully drained condition. The results of strength parameter (c’,φ’) are clearly shown that cohesion value has tendency of increasing when the degree of saturation is getting higher (Figure 3).

Slope stability analyses were done by KUslope computer program developed by Kasetsart University (Isaroranit, 2001). The geometry of the mountain slope was studied to select the appropriate cross section for the analysis. The cross section of slope above Na Nai village shown in Fig.4 were used as a typical section for analyses.

The analyses were done with various angle of slope from 14 to 40 degrees of both natural (one bench) and cut slope. Since the strength parameters vary based on degree of saturation, therefore the slope analysis were done by trial a pair of strength

parameters along the left boundary line shown in Figure 3 to obtain the possible lowest F.S.. The results are summarized in Fig 5. The results show that the factor of safety of cut slope has lower value than the natural slope as expected. The critical slope angle that the cutting might trigger the landslide is 17.1 degree in which corresponding to FS equal to 1.3. Therefore, any construction cutting (1H:2V) in the natural slope with slope angle greater than 17.1 degree, those cut slopes need to be analyzed and the counter measures need to be considered.

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50

Friction, degree

Coh

esio

n, t

/m2

New Data

Previous Data91.85

94.99

91.34

94.18

86.51

85.01

86.96

89.77

63.89%

62.16%

86.19%

100

Cohesion 0.168 0.170 0.200 0.250 0.300 0.350 0.410

Degree 34.5 33.0 29.6 27.0 25.2 24.8 26.1 Figure 3: Boundary of strength parameters used for the analyses.

Figure 4: Typical section for analyses.

0.500.600.700.800.901.001.101.201.301.401.501.601.701.801.902.002.102.202.302.402.502.602.702.802.903.00

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

Natural Slope Angle (degree)

Fact

or o

f Saf

ety

Normal Slope cohesion = 0.35 degree = 24.8Cut Slope 2 : 1 cohesion = 0.17 degree = 33Cut Slope 3 : 1 cohesion = 0.17 degree = 33

Figure 5: Factor of safety of cut and natural slope.

5 LANDSLIDE SUSCEPTIBILITY MAP

The landslide susceptibility area is analyzed using weighting factor method. 8 major factors listed below were considered in the analysis including the safety factor calculated from geotechnical method explained in the previous section.

1. Factor safety and slope angle relationship 2. Rock mass structure & slope angle relationship (Later neglected) 3. Lineament zone 4. Rock type (Later neglected) 5. Distance from road 6. Elevation 7. Land use and land cover 8. Surface drainage zone The factor of rock mass structure and rock type

were neglected since the detail analysis found that there is less correlation of those factors with previous landslide. In determining the numerical rating for landslide potential of each 6 parameters, an area of 5x5 square meters grid cell has been employed for the analysis by GIS program. The weight-rating value of each parameter is determined in each square grid cell of each derivative map. Table 2 shows the weighting value of the parameters, considering weighted matrix method based on expert assessments as shown in Table 3. Finally the scores of weight-rating in each 5x5 square meters grid cell is obtained from the summation of weight-rating values of each derivative map to obtain the landslide susceptibility map. The map of each influencing factors is shown in Figure 6.

Left Boundary Line

Table 2: Weights and rating values used for the analyses.

Rating Value Parameter Weight

Value Description Rating (1-5)

1. Factor safety and slope angle relation-ship

1.875

A. F.S.≤ 1.3 (≥26 de-gree) B. 1.3 < F.S. ≤ 1.5 (22≤ slope< 26 degree) C. 1.5 < F.S. ≤ 1.8 (18≤ slope< 22 degree) D. F.S. >1.8 (< 18 de-gree)

5 3.66 2.33

1

2. Linea-ment zone

1.625 A. Area inside linea-ment zone B. Area outside linea-ment zone

5 1

3. Distance from road

1 A. Area inside road zone B. Area outside road zone

5 1

4. Eleva-tion

1 A. >80 m B. 40-80 m C. 0-40 m (Not include slope < 10O)

1 3 5

5. Surface drainage

1 A. Area inside surface drainage zone B. Area outside surface drainage zone

5 1

6. Land use and land cover

1 A. Agriculture area B. Urban and built-up area C. Other deforestation D. Forest area

5 3.66 2.33

1

Table 3: The weighing process.

Factors 1 2 3 4 5 6 Total score Weight

1 3 3 3 3 3 15 1.875 2 1 3 3 3 3 13 1.625 3 1 1 2 2 2 8 1 4 1 1 2 2 2 8 1 5 1 1 2 2 2 8 1 6 1 1 2 2 2 8 1

The results of the analysis are shown in Figure 7.

The landslide susceptibility classes are classified as shown in Table 4. It clearly shows in Figure 7 that the actual events are matched with the moderate to high landslide susceptibility area from the result of analysis with the matching of 78.2 percent. Table 5 shows the suggestion for land development act, proposed to Patong Municipal which will be used as a guideline for the future enforcement.

(a) (b)

(c) (d)

(e) (f) Figure 6: Factors used in GIS. a. Factor of safety classification b. Lineament zone c. Distance from road d. Elevation e. Surface drainage area f. Land use and land cover. Table 4 Landslide classification based on score.

Score Landslide Potentials Classes

pixel Area (m2) %

31.6-37.5 Very high potential 408 10,200.00 0.07

25.6-31.5 High potential 13,054 326,218.36 2.27

19.6-25.5 Moderate potential 102,359 2,555,706.73 17.76

13.6-19.5 Low potential 237,269 5,904,230.18 41.03

0-13.5 Very low to nil potential 224,914 5,594,591.66 38.88

Total 578,004 14,390,946.93 100

Figure 7: Landslide susceptibility map. Table 5: Recommendations for slope cutting at various level of landslide susceptibilities.

Landslide Hazard Level Action

High Medium Low Very Low

Geotechnical engineer required

√ √

Geologist required

√

Land cover control

√ √

Drainage management

√ √

Control of the angle of cut slope

√ √ √ √

6 LANDSLIDE RISK MAP

Since risk is a function of hazard and consequence. Therefore, in order to develop risk mapping, the consequence area from landslide need to be estimated. Buildings location and their data were obtained from taxation map provided by Patong City. The hazard levels were classified based on landslide susceptibility map produced as discussed earlier. The affected area at the toe slope area was estimated to be equal to the height of the slope above the toe based on Finlay et al. (1999). Fig. 8 shows the boundary of affected area at the toe of the slope. The buildings were classified into levels based on their vulnerability due to landslide. The number of population at risk (PAR) is estimated from the census data in the taxation map.

Figure 8: Part of landslide risk map in Patong.

7 API MAP

Landslide disaster management will be incomplete if lack of warning system. The past studies found that the accumulated rainfall and current precipitation has great influence in rainfall-triggered landslide. This is consistent with the use of the API concept (Antecedent Precipitation Index). The API represents the moisture of the soil at any time using the values measured by rain gauge. The critical API is determined by calculating the critical moisture content in soil layer that will cause the soil layer to fail. Factor of safety was analyzed by infinite slope stability and with soil strength parameter of c' = 0.35 ksc., φ′= 24.8o, the unit weight of soil ( dγ ) = 1.41 t/m3 , Gs = 2.65, void ratio (e) = 0.89, degree of saturation (Sr) = 93% and the porosity of the soil (n) = 0.471. The calculation was done according to equation 1 and the critical thickness is determined from Equation 2. The result is shown in Figure 9 and Figure 10. The critical API map is shown in Figure 11.

criu H

crSFγβββ

φβ′

−′

+−=cossin1

tantan)]tan1(1[.. 2 (1)

when ..SF = Factor of safety for infinite slope ur = Pore pressure ratio Hu γ/ β = Slope angle γ = Unit weight of sliding mass criH = Critical depth (m)

crcrrcr TSnAPI ⋅⋅= , (2) when crAPI = critical API (mm.)

n = porosity crrS , = percent critical of saturation

crT = critical depth (m)

RELATIONSHIP BETWEEN SLOPE AND THICKNESS

0

5

10

15

20

25

30

35

40

45

0 10 20 30 40 50 60 70 80 90CRITICAL SLOPE

CRIT

ICAL

THI

CKNE

SS, m Sr = 93%

Figure 9: The relationship between the critical thickness and slope angle.

RELATIONSHIP BETWEEN SLOPE AND CRITICAL API

0

500

1000

1500

2000

2500

3000

3500

4000

0 10 20 30 40 50 60 70 80 90CRITICAL SLOPE

CRI

TICA

L A

PI, m

m.

Sr = 93%

Figure 10: The relationship between critical API and critical slopes angles.

8 WARNING BY RAINFALL

Using critical API for landslide warning is required to install rain gauge stations to present monitor rainfall in the area for calculating the API at time t. (APIt) (Equation 3) (Soralump and Thowiwat, 2010). Figure13 to 16 compare between the warning method using 3 days accumulated rainfall data and API value in which the later giving a longer time period prior to landslide events. However, as seen in Fig.16, it shows that a landslide event occurred before the calculated critical API value, therefore more real time wireless rain gauge system was installed as shown in Figure 17 and the Alert-Alarm-Action criteria were set for practical purpose.

Figure 11: Critical API map.

tttt PAPIKAPI +×= − )( 1 (3) when tAPI = API on ‘t’ day (mm) 1−tAPI = API on ‘t-1’ day (mm) tP = Precipitation on ‘t’ day (mm) K = recession constant API constant ( K =< 1.0 and usually 0.85-0.98)

12

3

45

6

7

89

Figure 12: Position of rain gauge stations. (Landscape scenarios: GISTHAI).

Critical rainfall envelope (3-days)

000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000050

7060

40

1500

70

0010

0000000000

25

000000

6050

20

5040

000

70

4020

4050

70

020

0000000000020

40

70

0000000000

110

0020

40 40

00000000010

5030

100

20000

30

00

30

100 100

70

3020

10500

3040

30

020

1005

30

000000000000000000000000000000000000000000000000000000000000000000000000000004.50.25000320.25

0.25500000.500000000000000000000002000000006.250.2500002.50.2500000000000074.2530011.5

01.25000001000000.25005.2500010

37.75

0.250.25018

37

0.254.75

34.25

41.5000000000000008.25

31.7545.5

130.250030.254.7511.25

56.25

0

41

00000010.75

00.25000000004

31.5

0.251.54.512.75218.25

000007.500000001.59.25

36.25

00005.5000000000.50.25000000.25216

000000.25000.250.25001018.25

1.759.751.250.50000003.25

36.75

0.50.2500000

25

0.7512.2512.7521.75

0000.2502029.75

40.25

00.7504.75

55.75

011.5

54

0003.5

127.25

200.25

41.75

11

85

34.5

6.50.750.521

0.250000000005002.7521.5

00000000000000000000000000000000000.2500000000000000000000000

50

100

150

200

250

300

0 50 100 150 200 250 300 350 400

3-days accumulated rainfall (mm.)

Daily

Rai

nfal

l (m

m.)

2007(Tarawadee)2007(Sai Mon)2007(Tai Trang)200420032008

Oct 14,2004

Figure 13: 3 days accumulated rainfall and daily rainfall rela-tionship.

APIcr (mm.)

Antecedance Precipitation Index ( APIcritical )

000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000050

7060

40

1500

70

0010

0000000000

25

000000

6050

20

5040

000

70

4020

4050

70

020

0000000000020

40

70

0000000000

110

002040 40

00000000010

5030

100

20000

30

00

30

100 100

70

3020

10500

3040

30

02010

05

30

000000000000000000000000000000000000000000000000000000000000000000000000000004.50.25000320.25

0.25500000.500000000000000000000002000000006.250.2500002.50.2500000000000074.2530011.5

01.25000001000000.25005.2500010

37.75

0.250.25018

37

0.254.75

34.25

41.5000000000000008.25

31.7545.5

130.250030.254.7511.25

56.25

0

41

00000010.7500.25000000004

31.5

0.251.54.512.752

18.25000007.500000001.59.25

36.25

00005.5000000000.50.25000000.25216

000000.25000.250.25001018.25

1.759.751.250.50000003.25

36.75

0.50.2500000

25

0.7512.2512.75

21.750000.250202

9.75

40.25

00.7504.75

55.75

011.5

54

0003.5

127.25

200.25

41.75

11

85

34.5

6.50.750.5210.250000000005002.75

21.500000000000000000000000000000000000.2500000000000000000000000

50

100

150

200

250

300

0 50 100 150 200 250 300 350 400

Antecedance Precipitation Index API(t-1) (mm.)

Daily

Rai

nfal

l (m

m.)

2007(Tarawadee)2007(Sai Mon)2007(Tri Trang)20042003API critical/day line2008

Oct 14,2004

Figure 14: API value and daily rainfall.

Critical rainfall envelope (3-days)

3.24.4500.3000000000000000000000000000000000000000000000000000000

27.3515.6514.85

26.8

015.051.2

25.459.45

000010.7

0.950.60.409.15

22.818.5500

47.9

0.10.713.35

2.454.857.6520.75

2.9514.95

57.217.350

31.8518.9521.7

53.3

0.350.217.35

0.457.59.5503.23.411.154.2511.28

32.2

2.352.10000.53.39.10.20.350.450.153.85

55.55

19.60.300

32.1

10.60.24.20.970

16.116.757.4001.35

11.8

67.8

25.65

4.05004.23.34.3000000.700004.8

52.25

28.15

00.051.9000.14.85.850.94.850.21000009.4

0.61.8

98.4

0.200

65.2

42.6

0004.40.400020.4

34.2

77.6

0.2020.60.60000000032.2

25.2

0.62.401.40000

22.6

0.44.811.2013.2

87

17.402.8

24

94

31.8

7.47.816

1012.2

22

0.200

21.4 15.46.43.41.8

12.422.4

0.4

22.2

00015.4

5.6

40.6

0.400.8

23.25.64.24.20.40

31.8 23.6

52

1310.8

38

14.41.2000.400.4

47.2

18.424

0018.6

0.2

47.8

27

49.8

156.800

63

11.240.20.40.200.21.811.8

34.822.2

3.40.2

65.4

12

58.6

0.2000000000000000000000.8100012

0.41.80.40000008.2

0.2000000

101.4110.2

0

50

100

150

200

250

300

0 50 100 150 200 250 300 350 400

3-days accumulated rainfall (mm.)

Dai

ly R

ainf

all (

mm

.) KALIM 2009

Sept. 19, 2009

Figure 15: 3 days accumulated rainfall showing the landslide events in 2009.

Antecedance Precipitation Index ( APIcritical )

3.24.4500.3000000000000000000000000000000000000000000000000000000

27.3515.6514.85

26.8

015.05

1.2

25.459.450000

10.70.950.60.40

9.1522.818.55

00

47.9

0.10.713.352.454.857.65

20.752.95

14.9557.2

17.350

31.8518.9521.7

53.3

0.350.217.350.457.59.55

03.23.411.154.2511.28

32.2

2.352.10000.53.39.10.20.350.450.153.85

55.55

19.6

0.300

32.1

10.60.24.20.970

16.116.757.4001.35

11.8

67.8

25.65

4.05004.23.34.3000000.700004.8

52.25

28.15

00.051.9000.14.85.850.94.850.21000009.4

0.61.8

98.4

0.200

65.2

42.6

0004.40.4000

20.434.2

77.6

0.2020.60.60000000032.2

25.2

0.62.401.40000

22.6

0.44.811.2013.2

87

17.402.8

24

94

31.8

7.47.816

1012.2

22

0.200

21.4 15.46.43.41.8

12.422.4

0.4

22.2

00015.4

5.6

40.6

0.400.8

23.25.64.24.20.40

31.823.6

52

1310.8

38

14.41.2000.400.4

47.2

18.424

0018.6

0.2

47.8

27

49.8

156.800

63

11.240.20.40.200.21.8

11.8

34.822.2

3.40.2

65.4

12

58.6

0.2000000000000000000000.8100012

0.41.80.40000008.2

0.2000000

101.4

110.2

0

50

100

150

200

250

300

0 50 100 150 200 250 300 350 400

Antecedance Precipitation Index API(t-1) (mm.)

Daily

Rai

nfal

l (m

m.)

KALIM 2009API critical/day line

Sept. 19, 2009

Figure 16: API graph showing the landslide event in 2009.

9 CONCLUSION

Geotechnical approach is found to be useful in landslide risk management in Patong, Phuket. Hazard and risk mapping together with the landslide warning system using rainfall data were produced and set up based on geotechnical data and analyses.

Figure 17: Wireless rain gauge system.

Antecedence Precipitation Index ( APIcritical )

0

50

100

150

200

250

300

0 50 100 150 200 250 300 350 400

Antecedence Precipitation Index API(t-1) (มม.)

Daily

Rai

nfal

l (m

m.) mm./day

rainfall 2010

Figure 18: Alert Alarm Action criteria.

ACKNOWLEDGEMENT

The authors would like to thank Norwegian government and Patong Municipality.

REFERENCES

Soralump, S. 2010. Rainfall-Triggered Landslide: from re-search to mitigation practice in Thailand. The 17th South-east Asian Geotechnical Conference, May 10-13, 2010, Taipei, Taiwan.

Pungsuwan, D. 2006, Evaluation of Landslide Sensitive Area for Slope Development in Phuket. M.S. Thesis, Kasatsart University, Bangkok, Thailand.

Department of Mineral Resources (DMR), 2006. A Study of Prevention and Mitigation Landslide. Final Report, Bang-kok, Thailand.

Soralump, S. and W. Thowiwat, 2007. Shear Strength-Moisture Behavior of Residual Soils of Landslide Sensitive Rocks Group in Thailand. The Fifteenth National Conven-tion on Civil Engineering (NCCE15), 12-14 May 2010. Ubon Ratchathani, Thailand

Finlay, P.J., Mostyn, G.R., Fell, R., 1999. Landslide risk as-sessment: prediction of travel distance. Can. Geotech. J. 36, 556–562.

Soralump, S. and W. Thowiwat, 2007. Critical API Model for Landslide Warning. The Fifteenth National Convention on Civil Engineering (NCCE15), 12-14 May 2010. Ubon Ratchathani, Thailand

Isaroranit, R., 2001, Development for Slope Stability Program by Generalized Limit Equilibrium. M.S. Thesis, Kasatsart University, Bangkok, Thailand.