1. The influence of increasing rainfall due to climate change on landslide slopes
Tetsuya Kubota Forest Environment Science Department, Faculty of Agriculture, Kyushu University, Japan
1. Objective
Increase in rainfall that supposed to be induced by the global climate change is obvious in western Japan, according to the analysis of rainfall data observed in various locations including mountainside that are not influenced by local warming due to urbanization. On this point of view, we elucidated the response of landslide slope to this increase in rainfall. Hence, its long term influence on the specific landslide slopes in this area was analyzed using numerical simulation method i.e. finite element method in order to evaluate the landslide slope stability.

2. 2. Method and target areas
Field investigation on landslides slopes and slope failures was conducted to obtain the geologic information, geo-structure, vegetation feature, soil samples and topographic data i.e. cross sections. Accordingly soil shear tests and soil permeability tests are also conducted. The rainfall data at the nearest rain observatory were obtained from the database of Japan meteorological agency. The long term influence on the slope stability in the area is analyzed by the finite element method (FEM) combined with rain infiltration and seepage analysis with the long term rainfall fluctuation data, obtaining factor of safety (Fs = resistance force/ driving force) on real landslide slopes.
The target areas are located in northern Kyushu district, western Japan where they are often suffered from severe landslide disasters. For the FEM analysis, 14 landslide slopes are selected from ones that occurred due to heavy rainfall caused by stationary front in July 2009~2017. The geology in research areas consists of Paleozoic and Mesozoic rocks (mainly schist, granite, slate, serpentine) and volcanic sediment. The vegetation consists of mainly Japanese cypress, cedar or bamboo.
The target area
Tokyo
Kyoto

3. Fig.1 Some examples of landslide occurred in the rainfall disaster in 2017

4. The rain infiltration and seepage analysis The finite elements
The slope stability analysis (Fs: factor of safety)
Fig. 2 FEM analysis examples (coupling analysis of infiltration-seepage and slope stability: Asakura-Kurokawa and Yame-Kuroki-Kasahara)

5. Fs=0.990, (Fs=1.00 with estimated rainfall in 2009)
Fig.3 The FEM analysis concerning with the long term rainfall increase at Aso-Teno, Kumamoto
3. Result and consideration Consequently, the long term rainfall increase in the region such as increment of approximately 20mm/hr for rain intensity Ri, or 50mm/day for daily rain Rd in 40 years is confirmed statistically using Kendall's rank correlation, and it is found that its impact on slope stability is oblivious and critical. In the sample landslide slopes, even the increase in rain of duration for only 10 years has severe impact on their stabilities in terms of Fs. The Fs calculated with rains in previous decade is higher than 1.0 such as 1.05 that corresponds to stable state, whereas the Fs with present rains is lower than 1.0 such as 0.98 which means unstable state.

6. The Aso-Hakoishi shallow landslide
The Yame-Kuroki, Kasahara deep-sheeted landslide
Fig.4 The case of Yame-Kuroki, Kasahara deep-sheeted landslide: “rainfall increased Fsno_rinc=0.9” and “rainfall in the disasters Fsrealrain=0.7” ; 22% safer if there were no rainfall increase. The case of Aso-Hakoishi shallow landslide: Fsno_rinc=1.02, Fsrealrain=0.990

7. With rainfall estimated for previous decade: Fs=1.02 (Stable)
With heavy rainfall in July 2009：Fs=0.980 ( landslide)
Without rain: Fs=1.10
Fig.4 The FEM slope stability analysis results in Sasaguri (Fs=1.02 →0.980)
With rainfall estimated for previous decade: Fs=1.0 (Stable) .
With heavy rainfall in July 2017：Fs=0.990 (landslide) with 24mm/hr increase.
Without rainfall: Fs=1.12
Fig.5 The FEM slope stability analysis results in Asakura

8. With rainfall estimated for previous decade: Fs=1.04 (Stable)
The hyetograph of heavy rain in July 2009 at Sasaguri：Fs=0.98 ( landslide)
The hyetograph of heavy rain in June 2001：Fs=1.02 (stable)
With rainfall estimated for previous decade: Fs=1.04 (Stable)
With heavy rainfall in July 2009：Fs=0.980 ( landslide)
(Without rainfall: Fs=1.05)
Fig.6 The examples of rainfall pattern (Sasaguri)
Fig.7 The slope stability analysis results in Fukuchi

9. ◇ The Fs calculated with estimated rainfall for previous 3 decade is Fs=1.01 （－9mm/hr）. ◇ The Fs calculated with estimated rainfall for previous decade is Fs = （－3mm/hr）. ( The Fs calculated with estimated uniform rainfall （8mm/hr） is Fs = The increment in maximum shear strain is larger than the one without rainfall stress.) Fig.8 A slope stability analysis for the landslide in Keisen area in July 2010

10. Fig.9 FEM analysis of Valle San Angel landslide in Nuevo Leon, Mexico
With the rainfall in Hurricane Alex in 2010: Fs=0.990
With assumed rainfall in 2000 (calculated with the long term rainfall fluctuation): Fs=1.0
Fig.9 FEM analysis of Valle San Angel landslide in Nuevo Leon, Mexico
Without rainfall: Fs=1.04

11. Results of the FEM slope stability analysis with rainfall infiltration/seepage analysis
FEM Fs slope stability Analysis concerning rainfall long term fluctuation due to the climate change
Fs = Resistance force/Driving force: Fs<1.0 corresponds landslide occurrence.
Landslide Case Name
(1)Fs without rain increase
(2)Fs with actual rainfall
(3)Fs without rain
Rain increased (mm/hr/40years)
Slope case
Fsno-r_inc
Fsrealrain
Fs_non
Rinc(mm/hr)
DFs=(1)-(2)
Aso-Teno Shallow
0.990
1.20
19.7
Aso-Teno Deep 1
1.05
0.950
1.10
0.100
Aso-Teno Deep 2
1.02
0.900
0.120
Aso-Takanasi Shallow
1.00
0.050
Aso Hakoishi Shallow
0.99
0.030
Yame-Kuroki Deep
0.700
2.13
19.5
0.200
Sasaguri Deep
0.980
20.0
0.040
Fukuchi Deep
1.04
5.5
0.060
Keisen Deep
1.01
0.970
12.0
Asakura-Kurokawa Deep
1.12
24.0
0.010
Asakura-IshizumeShallow
1.16
Asakura-Takagi Deep
0.985
Hita-Turukochi2 Deep
3.0
Valle San Angel, MX Deep
31.4

12. with rainfall infiltration/seepage analysisFs
Landslide Case No.
Fig.10 Results of the FEM slope stability analysis
with rainfall infiltration/seepage analysis
(The difference between “rainfall increased Fsno_rinc” and “rainfall in the disasters Fsrealrain” is obvious by statistical significance level=5%, P=0.0475)

13. 4. Conclusion
The influence of increasing rain in landslide occurrence i.e. slope instability was estimated by approximate average 5% (1% to 12%, in extremely case 22%) of reduction in Fs, and it deemed enough high on the slope stability point of view.
The increase of “rainfall rainfall intensity Ri”, “daily rainfall Rd” due to climate change with the increasing rate such as 20mm/hr or 50mm/day in 40 years, surely has strong influence on almost landslide slopes in aspect of slope stability.
Therefore, with this rain increase rate, it is possible for many forest slopes or natural slopes to become unstable and cause landslide disasters.

14. Thank you