This is the author’s version of a work that was submitted/accepted for pub-

lication in the following source:

Suryo, Eko Andi, Gallage, Chaminda, & Trigunarsyah, Bambang

(2015)
A method for predicting rain-induced instability of an individual slope. In
Barnes, Paul H. & Goonetilleke, Ashantha (Eds.)
The 9th Annual International Conference of the International Institute for
Infrastructure Renewal and Reconstruction. (8-10 July 2013), Queensland
University of Technology, Brisbane, Australia, pp. 118-127.

This file was downloaded from: https://eprints.qut.edu.au/61759/

c Copyright 2013 Please consult the authors

Notice: Changes introduced as a result of publishing processes such as

copy-editing and formatting may not be reflected in this document. For a
definitive version of this work, please refer to the published source:

http://digitalcollections.qut.edu.au/2213/
 A Method for Predicting Rain-induced Instability of an
Individual Slope
Eko Andi Suryo1, Chaminda Gallage2, and Bambang Trigunarsyah3
1Eko Andi Suryo, Brawijaya University Malang, Indonesia
E-mail: andisuryo@gmail.com
2Chaminda Gallage, QUT Brisbane, Australia
E-mail: chaminda.gallage@qut.edu.au
3Bambang Trigunarsyah, QUT Brisbane, Australia
E-mail: bambang.trigunarsyah@qut.edu.au

Awareness to avoid losses and casualties due to rain-induced landslide is increasing in regions that rou-
tinely experience heavy rainfall. Improvements in early warning systems against rain-induced landslide
such as prediction modelling using rainfall records, is urgently needed in vulnerable regions. The existing
warning systems have been applied using stability chart development and real-time displacement meas-
urement on slope surfaces. However, there are still some drawbacks such as: ignorance of rain-induced
instability mechanism, mislead prediction due to the probabilistic prediction and short time for evacuation.
In this research, a real-time predictive method was proposed to alleviate the drawbacks mentioned above.
A case-study soil slope in Indonesia that failed in 2010 during rainfall was used to verify the proposed
predictive method. Using the results from the field and laboratory characterizations, numerical analyses can
be applied to develop a model of unsaturated residual soils slope with deep cracks and subject to rainwater
infiltration. Real-time rainfall measurement in the slope and the prediction of future rainfall are needed. By
coupling transient seepage and stability analysis, the variation of safety factor of the slope with time were
provided as a basis to develop method for the real-time prediction of the rain-induced instability of slopes.
This study shows the proposed prediction method has the potential to be used in an early warning system
against landslide hazard, since the FOS value and the timing of the end-result of the prediction can be
provided before the actual failure of the case study slope.

Key Words : numerical analysis, rain-induced slope instability, real-time prediction.

1. INTRODUCTION
The growth of the world population has created an intensive demand for residential and agricultural
development1). The limited availability of suitable residential land has forced people to live in areas
proven to be subjected to natural disasters, such as landslides and floods. It has been reported that most
landslides occur in the rainy season, potentially causing damage to infrastructure and human lives 2), 3), 4),
5), 6), 7), 8), 9), 10), 11)
. The effects of rainfall have to be considered in rain-induced landslide hazard assess-
ments12). In Indonesia which is located in a tropical region with high seismic activity and receives a very
high annual rainfall, rainfall-induced landslides are acknowledged as one of the major natural disasters.
In-depth studies of the stability of these slopes, together with the related programs to increase the safety
awareness of people living in these areas, has become important and challenging geotechnical engi-
neering task.
The existing warning systems against rain-induced slope failures are mainly based on the following:
• Charts developed correlating past measured rainfall and observed slope failures associated with
rainfall (e.g. 13), 14), 15), 16)). These charts can be used with the real-time measured rainfall data to
predict slope failures. The failure of individual slopes cannot be predicted accurately using this
method due to its probabilistic nature. Further, these charts cannot be used to assess the stability
gain of a slope after a rainfall event.
• Real-time displacement measurement on slope surfaces (using GPS, extensometers) or/and in the
sub-soil (using inclinometers) (e.g. 17), 18), 19), 20)). Although this method can be applied to an indi-
vidual slope, it may give short time for possible evacuations, as the displacement is measured when
the slope moves. The displacement measurement of the slope cannot be used to assess the stability
gain of the slope after a rainfall event.
Note: Slopes become less stable during and just after rainfall but sometimes do not fail, these slopes
gain the stability with time after the rainfall event.
1
 Considering the drawbacks of the existing methods, there is a real need to develop a landslide pre-
dictive method that can be used for predicting the rain-induced failure of individual slopes in real-time.
This paper presents a method for predicting rain-induced slope instability. The method is based on
coupled transient seepage and stability analyses that capture the mechanism of rain-induced slope in-
stability. This method can be used in real-time, applied to an individual slope, and to predict the stability
gain of the slope after a rainfall even. The applicability of the proposed method is also presented by
using a case study slope that failed on the 31st of October 2010.

2. METHOD TO PREDICT RAIN-INDUCED INSTABILITY OF A SLOPE

When applying the proposed method to a slope, ‘critical slope’ that refers to soil slope that highly
susceptible to landslide or slope failure that ranges in its potential effects from being a temporary nui-
sance as a result of the partial closure of a roadway, to destroying human lives and physical structures,
or being potentially catastrophic has to be identified. The proposed method of predicting rain-induced
instability of a critical slope consists of the steps briefly described in the following sections.

(1) Land Survey and field observation

Field surveys and tests are conducted to obtain geometrical details of the critical slope, soil samples
for laboratory testing, water table location, water content of the soil, deep crack locations and orienta-
tion. The main equipment used in field survey includes a total station, a levelling apparatus, and a global
positioning system (GPS) device. A number of bore-holes are drilled along the slope to identify soil
profile, to take undisturbed/disturbed soil samples for laboratory soil tests, to identify water table lo-
cation. The electrical resistivity tomography (ERT) survey is conducted in the subsoil to identify the
location and geometrical information relating to subsoil cracks 21).

(2) Laboratory investigations

Laboratory tests are conducted on undisturbed/disturbed soil samples collected from the field
bore-holes to determine soil classification parameters, unit weight, water content, saturated/unsaturated
shear strength parameters, and saturated/unsaturated hydraulic properties of the soil in each soil layer.
These parameters are needed for the modelling and analyses of the slope.

(3) Rainfall data collection & prediction

It is recommended to collect rainfall records at the selected location for at least five years are col-
lected for this research. Past and current rainfall data would then be used for predicting future annual,
daily, and hourly rainfall. Most meteorological institutions provide long-term history rainfall records
for research purposes. For instance, The Indonesian Agency for Meteorology, Climatology and Geo-
physics (Badan Meteorologi, Klimatologi, dan Geofisika - BMKG) provides forecasts of monthly
rainfall for up three months ahead. Simple statistically method also can be used to predict hourly
rainfall. By calculating deviation of hourly rainfall in every single hour compare to previous hour, and
then comparing the results over several years, the maximum deviation in each hour in a year can be
determined. Regardless of its moderate level of accuracy, the maximum deviations can be applied as a
threshold for predicting the hourly rainfall for slope stability analysis purposes. This study uses the
combination of statistical methods and monthly rainfall forecast from climatology institution to predict
future rainfall.

(4) Transient seepage and stability analyses

Based on the field and laboratory test results, the slope is modelled for transient seepage (using
SEEP/W) and stability (using SLOPE/W) analyses. The analyses are conducted following the steps
listed below. The results of these analyses are used for warning (real-time) against rain-induced slope
instability of an identified and investigated critical soil slope.
• The slope is modelled in SEEP/W and SLOPE/W, based on slope geometrical information
obtained from the field survey, sub-soil layers and soil cracks identified by bore-hole data,
geophysical surveys, and laboratory soil testings.

2
• The relevant boundary conditions are defined for both models in SEEP/W and SLOPE/W, and
necessary material properties are assigned for each layer. For the transient seepage analysis
using SEEP/W, the hydraulic conductivity function for each soil needs to be given. For the
stability analysis using SLOPE/W, shear strength parameters (c, φ, and φb) and unit weight
should be given for each layer.
• The transient seepage analysis of the slope are performed using SEEP/W by giving initial
pore-water pressure condition in the soil based on water table and rainfall (unit flux) as a
function of time on the surface boundary of the model.
• The results of the transient seepage analysis (the variation of pore-water pressure conditions in
the soil slope with time) can be used in SLOPE/W to get the pore-water pressure in the slope to
calculate the stability of the slope at a given time.
• The above steps are schematically illustrated in Figure 1.

Fig 1: Illustration of coupled analysis of SEEP/W and SLOPE/W

• Following the procedure outlined in Figure 1, the stability of the slope (FOS) can be calculated
for a year in advance (365 days), based on the predicted rainfall pattern given in Figure 2. One
year prediction of rainfall is based on the past rainfall records. If the calculated FOS of the slope
reaches or goes below a critical FOS anytime during the forecast 365 days (Figure 3), it is
recommended that a real-time stability analysis using an hourly or daily (24 hours) basis of
recorded rainfall (Stage 2) be conducted commencing from the time when FOS = 1.1 *
FOS critical . FOS critical has to be determined based on criticality of the consequences of the slope
failure.
• Figure 4 shows that the FOS can be calculated near real-time in daily basis (one day ahead of
the current time) based on the predicted maximum possible rainfall for the next day. If the next
day FOS reaches or goes below the FOS critical, a warning should be issued for the evacuation
of people in the potentially affected area.

3
 Fig 2: Typical measured and predicted rainfall patterns

Fig 4: Illustration of FOS using predicted daily rainfall

Fig 3: Illustration of measured and predicted FOS

3. THE CASE-STUDY SLOPE

(1) Location of the case-study slope

To verify the applicability of the proposed method, a residual soil slope located slopes in the hilly terrain in
East Java Province in Indonesia were selected. As shown in Figure 5, the chosen slope located is located in a
high seismic active region where three tectonic plates meet. The field investigation which consists of geo-
metrical survey, surface observation, three bore–holes, and Electrical Resistivity Tomography (ERT) was
conducted on the slope in May and June 2010. This slope failed on the 31st of October 2010 during rainy
season.

Fig 5: (a) Map of Tectonic Plate in the Indonesian Archipelago; (b) Indonesian Region;
(c) Selected slope location for this study

4
Figure 6 shows the elevation contour map of the area where the selected slope is located. The land survey
conducted on the slope gave the maximum average slope angle of about 300 along line BB’ (Figure 6). The
ERT survey was undertaken along the line BB’ to detect the sub-soil cracks and the cracks observed on the
surface are shown in Figure 6. As shown in Figure 6, three bore-holes (BH1, BH2, BH3) were conducted along
line BB’ to get undisturbed soil samples for laboratory tests, to identify sub-soil profile, and to obtain ground
water table location.

Fig 6: Topographical map of the case-study slope

(2) Rainfall records and prediction in the area where the slope is located
Figure 7 shows the monthly rainfall records for five years from 2007 to 2011. These data were collected
from rain gauges located not more than 5 km from the investigated slope. Though the rainfall data presented in
Figure 7 are in monthly basis, both daily and hourly rainfall data were also collected for these five years (2007
to 2011).
1000
2007 2008 2009 2010 2011
800
Rainfall (mm)

600

400

200

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Time Elapse (month)
Fig 7: Mothly rainfall record form 2007 to 2011 at case-study slope

Several methods can be used to predict the rainfall for one year in advance by using the past recorded rainfall
data. One of them is statistical method 22), 23) that is used by meteorological institutions to forecast the rainfall.
The simple average or the maximum rainfall over the past five years (at least) on daily or monthly basis can be
used as the predicted daily or monthly rainfall for the next 365 days.
When it is necessary to predict the stability in near real-time (one day in advance), then the rainfall has to be
predicted only one day in advance. One method is to use the next day rainfall forecast by the meteorological
institute for the analysis. Another method is to calculate the maximum increase in rainfall for a given day (e.g.:
2nd of January) with respect to the rainfall of the previous day (e.g.: 1st of January) over the period where the
past rainfall records are available (e.g.: five years) and then the maximum rainfall increase in the next day is
added to the present day rainfall (measured) to predict the next day rainfall.

5
4. MODELLING of CASE STUDY SLOPE
Based on the field investigations and the laboratory soil testing, the slope can be modelled with two-layer
soil profile with few deep cracks a shown in Figure 8.

Fig 8: Illustration of soil layers in case-study slope with crack zone and material

The slope has a length of 140 m and the maximum elevation difference of 35 m. The angle of the slope
surface was up to 30o. The location of the ground water table found on the day of investigation (12th June 2010)
is shown in Figure 8. A stream was found at the downstream of the slope with its water level at the elevation of
660 m at the time of the field investigation on 12th June 2010.
Based on the laboratory soil tests, properties and classification of soil in each layer is presented in Tables 1.
These properties are used in transient seepage and stability analysis of the slope. The deeps cracks were
modelled as “weak zones” which are zones of the silty soil that was being disturbed by direct rainwater infil-
tration through the soil deep crack. Since deep cracks cause the direct infiltration of rainwater into soil slope 24),
25)
, a thin layer (less than 20 cm) with very high hydraulic (k sat = 0.1 cm/sec) was introduced in to the weak
zone to facilitate the direct infiltration process.

Table 1: Parameter of Soil Layers of Case-study slope

Parameter Layer 1 Layer 2 Weak zone
USCS type ML-MH ML-MH ML-MH
γ (kN/m3) 16 17 15
W (%) 36 38 43
C (kPa) 17 23 5.4
φ (degree) 20 32 17.5
k sat (cm/sec) 3.87x10-5 1.83x10-6 0.001

Some of the properties of the weak zone were not measured, but were estimated based on some studies24), 25),
26), 27)
. Soil water characteristic curve (SWCC) is needed in transient seepage and stability analysis of
unsaturated soil condition. Since SWCC was not measured for any soil type in the case study slope, therefore,
the sample of SWCCs available in SEEP/W were assigned to the materials in the slope based on the results of
the grain-size distribution test 28) that is available in SEEP/W. Due to the absence of grain-size distribution data
of the weak zone soil, the SWCC of the weak zone at case-study slope was assigned by using a similar SWCC
value as for the weak zone21), since both slopes were in close proximity. For determining φb in this case study,
the prediction method suggested in previous research was used29).
To apply the proposed method of prediction of rain-induced slope stability and warning against the slope
failure, of the case-study slope which failed on the 31st October 2010, the following steps would be followed:
• The ‘current day’ for prediction purposes was assumed to be 12th June 2010 when field investigations
were completed. The case-study slope shown in Figure 8 was modelled in SEEP/W, with the initial
pore-water pressure in the slope being a based on water table location that was observed during the field
investigation.

6
• Using average value of the past three years rainfall records and three monthly rainfall prediction from
BMKG as mentioned in a previous section in this paper, the daily rainfall for the next 365 days from the
12th of June 2010 (13th June 2010 to 13th June 2011) was predicted and shown in Figure 9. For this pre-
diction, measured rainfall records for the last 3 years period (June 2007 to June 2010), were used.

0.070
0.060
Rainfall (m/day)

0.050
0.040
0.030
0.020
0.010
0.000
13-Jun-10

7-Nov-10

19-Dec-10

3-Apr-11

15-May-11
9-Jan-11

5-Jun-11
4-Jul-10

15-Aug-10

24-Apr-11
5-Sep-10

17-Oct-10

28-Nov-10

30-Jan-11

13-Mar-11
25-Jul-10

26-Sep-10

20-Feb-11
Time Elapse (Daily)

Fig 9: Predicted annual rainfall data from 13th June 2010 to 12th June 2011

Figure 10 shows the maximum increase in daily rainfall with respect to the previous day over one year period
(13th June 2010 to 13th June 2011) that was calculated using the daily rainfall records for the past years (June
2007 to June 2010).

Fig 10: Deviation chart of maximum increase of daily rainfall from 13th June 2010 to 13st June 2011

In the modelling of case-study slope using finite element in SEEP/W, as shown in Figure 8, the surface was
defined as flux boundary where rainfall for the period of analysis was assigned. The left vertical boundary and
the bottom boundary were considered as “no flow”. The right vertical boundary was defined as a constant head
boundary as there is a stream in the area. The initial conditions (pore-water pressures) in the slope were given
based on the location of the water table that was monitored during the field investigation. The pore-water
pressure below the water table is positive and increase linearly with the depth at a rate of 9.8 kPa/m. It is
negative above the water table and considered to be decreased linearly with the height (above the water table)
at the negative rate of 9.8 kPa/m.
The time-dependent pore water pressure distribution determined from SEEP/W analysis was then coupled
with the stability analysis using SLOPE/W to obtain the variation of the FOS with time. The general limit
equilibrium method available in SLOPE/W was used to analyse the stability of the slope during rainfall. By
applying auto locate method to define 2000 iterations of trail failure surfaces for each scenario, the minimum
safety factor of the slope under given conditions was obtained.

7
5. RESULT AND DISCUSSION
Transient seepage and stability analyses using the predicted daily rainfall for one year in advance (from 13th
June 2010 to 12th June 2011) was performed and the variation of stability (Factor of safety -FOS) for this
period is shown in Figure 11.

Fig 11: Factor of safety of case-study slope with predicted rainfall from 13th June 2010 to 12th June 2011

It can be seen in the Figure 11 that case-study slope is in near failure condition (FOS = 1.04) on the current
date (13th June 2010) and FOS of the case-study slope will eventually go below unity at time step #155 (14th
November 2010) if the case-study slope receives the predicted rainfall. Since FOS is 1.04 on 12th June 2012,
the people living in the area that could be affected by the possible failure of the case-study slope can be in-
formed to be ready for possible evacuation from the start of the next rainfall event. For this slope, it is im-
portant to conduct near real-time stability analysis (daily) from 13th June 2010, to closely observe the stability
of the slope for accurate warning and safe evacuation of the people.
It can be decided to warn and evacuate the people in the possible affected area when the real-time FOS is
equal to, or just below 1.02. This evacuation could be possible, as it is not raining at present. If it is raining, an
FOS = 1.04 could be critical for evacuation of the people
With the start of rainfall at 1 pm on 22nd September 2010, the real-time stability can be calculated. The de-
viation chart as shown in Figure 10 was used. The result of the real-time stability analysis using daily rainfall is
shown in Figure 13. In this real-time daily analysis, all analyses were started at 1 am in every single day during
the observed time.

Fig 13: FOS at day #141 (31st October 2010) after assigned with predicted rainfall

8
 After predicted rainfall using deviation chart was assigned, it can be seen in Figure 13, the FOS on 31st
October 2010 had decreased below 1.02. The FOS value on 31st October 2010 was very near to unity. Based on
this daily prediction, it can be stated that the landslide at the case-study slope was likely to have occurred, after
being triggered by the rainfall event on 31st October 2010.
This verification result shows that the predictive method could be used to prevent losses or casualties (e.g.
closing of roads, movement of the people) in the area of the landslide disaster on 31st October 2010, by giving
an immediate warning for evacuation. Since this near real-time analysis can be done in short time, the analysis
can be started at the beginning of rainfall with the result (FOS) able to be determined even before the rainfall
stop. However, the safer warning should have been given before any rainfall event occurred from 13th June
2010 since the FOS was very near to unstable conditions from the beginning of the analysis (FOS=1.04).
By using long-term stability analysis, the landslide on 31st October 2010 at case-study slope can be esti-
mated to occur between 13th June to 14th November 2010. The near real-time stability analysis can predict a
landslide at the beginning of the triggering rainfall, by detecting the decrease of FOS. However, in this study, it
could be possible to have a discrepancy between calculated FOS and the actual stability due to the uncertainty
of soil parameter and the accuracy of the rainfall prediction. It is recommended to use the probabilistic stability
analysis to minimize above uncertainties.

6. CONCLUSION
By coupling transient seepage analysis using the finite element method with slope stability analysis using
the limit equilibrium method, a predictive method for rain-induced slope instability of natural residual soil
slopes has been proposed in this paper. To apply the proposed method to a critical slope, field investigations,
laboratory soil testing, and rainfall prediction have to be performed. The proposed method was applied to a
natural slope in Indonesia which was investigated (field and laboratory testing) in June 2010 and failed on 31st
October 2010. The results of the analyses suggested that the application of the proposed method in near re-
al-time (one day in advance) would have prevented the human casualties and minimized damage to movable
properties.

ACKNOWLEDGMENT: The authors would like to express their gratitude to Perum JASA TIRTA I, Malang,
Indonesia for their contribution for providing the rainfall data records.

REFERENCES
1) United Nation. Food and Agriculture Organization Statistic, 2009, Available
at http://data.un.org/Data.aspx?q=forest&d=FAO&f=itemCode%3a6661, Accesed on August 11st , 2010
2) Sweeney DJ and Robertson PK: A Fundamental Approach to Slope Stability problems in Hong Kong. Hong Kong, pp.35-44, 1979.
3) Chipp, P.N., Henkel, D.J., Clare, D.G., and Pope, R.G: Field measurement of suction in colluvium covered slopes in Hong Kong. In
Proceedings of the Seventh Southeast Asian Geotechnical Conference, November 22-26 Hong Kong: pp.49-62, 1982.
4) Pitts, J. : An investigation of slope stability on the NTI campus, Singapore. Nanyang Technological Institute, Applied Research
Project, March, 1985.
5) Brand, E.W. : Landslides in southeast Asia: a state-of-the-art report. In Proceedings of the 4th International Symposium on
Landslides, Toronto, Canada, 1, pp.17 – 59, 1984.
6) Tan, S.B., Lim, T.L., Tan, S.L., Yang, K.S.: Landslide Problems and their Control in Singapore. In: 9th Southeast Asian
Geotechnical Conference, Bangkok, Thailand, 1987.
7) Johnson, K.A. and Sitar, N. : Hydrologic conditions leading to debris-flow initiation. Canadian Geotechnical Journal, 27, pp.789 –
801, 1990.
8) Fredlund, D.G. and Rahardjo, H.: Soil Mechanics for Unsaturated Soils. New York: John Wiley & Sons, Inc., 1993.
9) Brand, E.W. : Keynote Paper: Slope Instability in Tropical Areas. Proc. Of 7th International Symposium on Landslides, Trondheim,
pp.2031 – 2051, 1996.
10) Lim, T.T., Rahardjo, H., Chang, M.F. and Fredlund, D.G.: Effect of Rainfall on Matric Suction in a Residual Soil Slope. Canadian
Geotechnical Journal, 33(2), pp.618 – 628, 1996.
11) Ng, C.W.W. and Shi, Q.: A Numerical Investigation of the Stability of Unsaturated Soil Slopes Subjected to Transient Seepage.
Computer and Geotechnics, 22 (1), pp.1 – 28, 1998.
12) Chowdhury, R., Flentje, P., and Bhattacharya, G.: Geotechnical Slope Analysis. Taylor & Francis Group. London, 2010.
13) Lumb, P.: Slope failures in Hong Kong. Quarterly Journal of Engineering Geology, 8, pp.31-65, 1975.
14) Keefer, D.K., Wilson, R.C., Mark, R.K., Brabb, E.E., Brown, W.M., Ellen, S.D., Harp E.L., Wieczorek, G.F., Alger, C.S., and
Zatkin, R.S.: Real-time Landslide Warning during Heavy Rainfall. Science, 238, pp.921-925, 1987.

9
15) Wilson, R. C., Torikai, J. D. and Ellen, S. D.: Development of Rainfall Warning Threshold for Debris Flows in Honolulu District,
Oahu. Open-file Report, US. Geological Survey, Washington, D.C., pp.92-521, 1992.
16) Slosson, J. E. and Larson, R. A.: Slope Failures in Southern California: Rainfall Threshold, Prediction, and Human causes. Envi-
ronmental & Engineering Geosciences, 1(4), Winter, pp.393-401, 1995.
17) JLS: Landslide in Japan. The Japan Landslide Society, n.d. , Available at : http://www.tuat.ac.jp/~sabo/lj, Accessed on August
27th, 2010.
18) Terzis, A., Anandarajah, A., Moore, K., and Wang, I.-J.: Slip surface localization in wireless sensor networks for landslide
prediction. Proceedings of the 5th International Conference on Information Processing in Sensor Networks. Nashville, Tennessee,
USA, 2006.
19) Frasheri, A. , Kaplani, L., and Dhima, F.: Geophysical landslide investigation and prediction in the hydrotechnical work. Journal of
the Balkan Geophysical Society, 1(3), pp.38-43, 1998.
20) Wilkinson, P.B., Chambers, J.E., Meldrum, I., Gunn, D.A., Ogilvy, R.D. and Kuras, O. : Predicting the Movements of Permanently
Installed Electrodes on an Active Landslide using Time-lapse Geoelectrical Resistivity data only. Geophysical Journal
International, 183, pp.543-556., 2010.
21) Suryo, E.A., Gallage, C., Trigunarsyah, B., Mochtar, I.B., and Soemitro, R.A.A. : Application of electrical resistivity method to
detect deep cracks in unsaturated residual soil slope. Proc., 5th Asia-Pasific Conf. On Unsaturated Soils (AP-UNSAT 2011), Ka-
setsart University, Pattaya, Thailand, pp. 901-906, 2011.
22) SPSS, IBM SPSS Forecasting 19, IBM, 2010.
23) Restu Tresnawati, Tri Astuti Nuraini, Wido Hanggoro: Prediksi Curah Hujan Bulanan menggunakan Metode Kalman Filter
dengan Prediktor SST NINO 3.4 Diprediksi, 2010. (transl. EA. Suryo). Available at : http://www.bmkg.go.id/Puslitbang/filePDF.
Accessed on April 4th, 2013.
24) Gofar, N., Lee, L.M., Asof, M.: Transient Seepage and Slope Stability Analysis for Rainfall-Induced Landslide: A Case Study.
Malaysian Journal of Civil Engineering 18(1), pp.1 – 13, 2006.
25) Wang, Z.F., Li, J.H. and Zhang, L.M.: Influence of Crack on the Stability of a Cracked Soil Slope. Proc., 5th Asia-Pacisif Conf. on
Unsaturated Soils (AP-UNSAT 2011), Kasetsart University, Pattaya, Thailand, pp.721-728, 2011.
26) Zhang, J., Jiao, J.J. and Yang, J.: In Situ Rainfall Infiltration Studies at a Hillside in Hubei Province, China. Engineering Geology,
57, pp.31 – 38, 2000.
27) Das, B.M.: Fundamentals of Geotechnical Engineering (Second Edition). Toronto: Thomson, 2010.
28) Aubertin, M., Mbonimpa, M., Bussière, B.and Chapuis, R.P, A model to predict the water retention curve from basic
geotechnical properties. Canadian Geotechnical Journal, 40(6), pp.1104-1122, 2003.
29) Rahardjo, H., Lim, T.T., Chang, M.F., and Fredlund, D.G., Shear-strength characteristics of a residual soil. Canadian
Geotechnical Journal, 32(1), pp. 60–77, 1995.

10