The application of hydrological models in shallow landslides prediction 指導老師:李錫堤 教授 報告者: 李浩瑋 報告日期:2010/11/4 Outline Introduction Review Objective Method Data Preliminary results Introduction Taiwan has been vulnerable to shallow landslide disasters caused by heavy rainfalls. Mitigate.
Download ReportTranscript The application of hydrological models in shallow landslides prediction 指導老師:李錫堤 教授 報告者: 李浩瑋 報告日期:2010/11/4 Outline Introduction Review Objective Method Data Preliminary results Introduction Taiwan has been vulnerable to shallow landslide disasters caused by heavy rainfalls. Mitigate.
Slide 1
The application of hydrological models in
shallow landslides prediction
指導老師:李錫堤 教授
報告者: 李浩瑋
報告日期:2010/11/4
1
Slide 2
Outline
Introduction
Review
Objective
Method
Data
Preliminary
results
Slide 3
Introduction
Taiwan has been vulnerable to shallow landslide disasters
caused by heavy rainfalls.
Mitigate landslide disasters :
Evaluate the potential of slope failure events in space and time
Lee at al., 2008
Landslide susceptibility analysis in the Tachia Creek drainage basin
Slide 4
Introduction
Statistical approaches:
(Fuzzy Logic, Logistic Regression, and Neural Networks…)
Deterministic approaches:
Physically-based models
(Iverson, 2000)
Steady state model
SHALSTAB MODEL(Montgomery et al., 1994)
SIMMAP (Pack et al., 1998)
Transient model
TRIGRS (Baum et al., 2002)
Basic:
Modelling of slope the hydrological response
Topography is considered
Slide 5
Objective
To establish a slope-instability analysis and a hydrological
model for landslide prediction during heavy rainstorms.
FS?
θ
Soil
Slide 6
Application of hydrology model
Review
Method
Data
Result
SHALSTAB
(Dietrich and Montgomery, 1994)
- Steady- state hydrological conditions
- Fully saturated conditions
- Homogeneous soil
- Slop-parallel groundwater flow
- Impermeable basal boundary
Hydrological model
Critical Rainfall
Infinite slope
qcr
Slide 7
Application of hydrology model
TRIGRS
(Baum et al., 2002)
Review
Method
Data
Result
- Transient hydrological conditions
- Fully saturated conditions
- Homogeneous soil
- Slope-parallel watertable
- Impermeable or infinite basal boundary
Infiltrate model
solution of Richard’s equation
Infinite slope model
(Z , t )
FS ( Z , t )
Slide 8
Application of hydrology model
Review
Method
Data
Result
(Casadei., 2003)
- Transient hydrological conditions
- Unsaturated conditions
- Slope-parallel watertable
- Homogeneous soil
- Impermeable boundary
Coupled hydrological–slope stability model
Groundwater table above the slip surface
hw (t )
Infinite slope model
FS (t )
Slide 9
Flow chart
Input data
Rainfall
Data
DEM Data
Area
Data
Soil
Parameter
Groundwater level
Estimation model
Hydrologic
Parameter
Infinite Slope
Stability Model
Output data
Groundwater
Height
Safety
Factor
Landslide
susceptibility Map
Slide 10
Hydrologic Model
Review
Method
TOPMODEL(TOPgraphy based hydrological MODEL)(Beven.
et al., 1979)
Stream line
Data
Result
Contour line
Specific Catchment Area a = A/b
Unit contour
length b
Contributing area A
Slide 11
Topographic index
Review
Method
Data
Result
a
T ln
tan
a: Specific area
tan β :slope
Slide 12
TOPMODEL Structure
Review
q
total
q
overland
q
subsurface
Method
Data
Srz
Result
Suz
Groundwater table
qv
Zw
Srz : Root zone
Suz : Unsaturated zone
D:Soil Depth
Zw:Water table height
qv :垂直入滲率
Slide 13
TOPMODEL Assumption
q
total
q
overland
q
subsurface
Saturated defict, zj
Q j T0e
(
zj
m
This equation can be solved for z:
)
c tan
Steady-state
Q j RA
a
z z m[ ln(
)]
tan
j
j
z j m ln(
R
a
) j m ln(
)
T0
tan
z m ln(
r
) j ] m
T0
1
a
ln(
)j
A j
tan
Slide 14
Hydrologic Routing
Review
Method
Data
Result
a
z j z m[ ln(
)j]
tan
The recharge to the groundwater table
q v, j SUZ j /(t d Z j ) , Q v qv , j A j
j
The baseflow from the saturated zone
z
Q Q exp( )
m
b
0
z(t 1) z(t)
Q b (t ) Qv (t )
t
A
Slide 15
Hydrologic Routing
Review
Method
Data
Result
a
z j (t 1) z (t 1) m[ ln(
)j]
tan
The recharge to the groundwater table
q v, j SUZ j /(t d Z j ) , Q v qv , j A j
j
The baseflow from the saturated zone
z
Q Q exp( )
m
b
0
z(t 1) z(t)
Q b (t ) Qv (t )
t
A
Slide 16
Infinite Slope MODEL
Review
依據有效應力概念,土體之阻滯力 τr 表示如下
Data
C ( u ) tan
C ( Z cos Z cos ) tan
Result
土體之駟動力τd為飽和土體沿坡面之分力
Method
r
2
2
s
s
w
Z sin cos )
d
s
resistance force
FS
driving force
C ( Z Z ) cos tan
Z sin cos
2
s
w
s
w
FS>1 Stable
FS<1 Unstable
Slide 17
Infinite Slope MODEL
Review
resistance force C ( Z Z ) cos tan
FS
driving force
Z sin cos
2
Method
s
w
w
s
Saturation ,W
Data
Result
FS
C
Z tan
[1 ( )]
Z sin cos
Z tan
w
s
w
s
TOPMODEL calculate local
storage deficit Zi
C
Z Z tan
FS
[1 (
)]
Z sin cos
Z
tan
w
s
i
s
Distribution of the soil saturation induced slope instability
Slide 18
Study Area
Review
Method
Data
Result
Slide 19
Study Area
Review
Method
Data
Result
Slide 20
Geological Map
Review
Method
Data
Result
Slide 21
Geo-material Parameter
Review
參數值
Method
Data
Result
土壤單位重,γS (KN/m3)
說明
zone1
18.2
實驗
zone2
19.6
實驗
zone3
參數
(zone1+zone)/2
坡度,θ (。)
ArcView軟體計算
由DTM推估
土壤深度,Z (m)
經驗式
由坡度推求
有效凝聚力,C (KPa)
搭配NDVI推估
抗剪摩擦角,ψ (。)
有效範圍推估
Slide 22
TOPMODEL Parameter
Review
參數
參數值
說明
Method
坡度,θ (。)
ArcView軟體計算
由DTM推估
Data
比集水面積,a/b (m)
ArcView軟體計算
由DTM推估
Result
土壤導水率隨時間衰減,m
0.032
退水歷線推估
0.0000328
流量觀測
土壤飽和時導水率,T0(m2/h)
5
林地飽和滲透
係數推估
未飽和層下滲延遲時間,
TD(h)
50
假定值
根系層最大容許儲蓄水量,
Srmax(m)
0.05
林地最大蓄水
量
初始根系層含水量,Sr0(m)
0.002
假定值
初紿流量,q0(m/h)
Slide 23
Landslides induced by AERE Typhoon
Review
Method
Data
Result
Slide 24
Topographic index
Review
Method
Data
Result
Slide 25
Local saturation deficits
Review
Method
Data
Result
Given
• m=0.032
• .Z 0.317
T=0
Compute
• =5.43
a
Z Z m ln(
tan
i
i
)
Slide 26
Soil profile saturation
Review
Method
Data
Result
Z w D Zi
w
Saturation > 1.0, Set to 1.0
Saturation < 0 , Set to 0
Zi
Zw
Z
Slide 27
Soil profile saturation
Review
Method
Data
Result
Slide 28
Rainfall Process
Review
Method
Data
Result
T=36
Slide 29
Prediction
Review
T=36
Method
Data
Result
Distribution map of slope failures of actual vs predicted by model
Slide 30
Classification Error Matrix
Review
Method
X
Data
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Result
非山崩網格
山崩網格
X
X
預測的山崩網格,坐落
於非山崩網格中
預測的山崩網格,坐落
於山崩網格中
Slide 31
Classification Error Matrix
Review
Method
Data
Result
The results of both are quantitatively compared by using two
following evaluation indexes.
AR= ( N1 ) / ( N1+ N2+ N3+ N4 )
FAR= ( N3 ) / ( N1+ N2+ N3+ N4 )
(1)
(2)
Here, AR is accuracy rate (%), FAR is false alarm rate (%) and N1,
N2, N3 and N4 (see Table )
Predict
Actual
Unstable
(FS < 1)
Stable
(FS ≧1)
Unstable
N1
N2
stable
N3
N4
N1, N2, N3, N4 : number of cell on each category
Slide 32
Classification Error Matrix
Review
Method
The results of both are quantitatively compared by using two
following evaluation indexes.
Data
Predict
Unstable
(FS < 1)
Stable
(FS ≧1)
Unstable
613
451
stable
13402
107343
Result
Actual
N1, N2, N3, N4 : number of cell on each category
AR= 57.61%
Slide 33
崩塌地分析成果及評估
文獻回顧
研究方法
正確率的計算方式可以分類誤差矩陣表(classification
error matrix)進行表示,其計算方式如下:
Total=121809
預測結果
研究資料
1
山崩(14015)
(FS < 1)
非山崩
(FS ≧1)
山崩(1064)
613
451
非山崩
(120745)
13402
107343
初步成果
未來工作
原始資料
山崩正確率=57.61%
非山崩正確率=88.91%
總體正確率=88.62%
Slide 34
Future Work
Calibration parameter
Model validation
Comparing the results
34
Slide 35
Thank you for your attention
The application of hydrological models in
shallow landslides prediction
指導老師:李錫堤 教授
報告者: 李浩瑋
報告日期:2010/11/4
1
Slide 2
Outline
Introduction
Review
Objective
Method
Data
Preliminary
results
Slide 3
Introduction
Taiwan has been vulnerable to shallow landslide disasters
caused by heavy rainfalls.
Mitigate landslide disasters :
Evaluate the potential of slope failure events in space and time
Lee at al., 2008
Landslide susceptibility analysis in the Tachia Creek drainage basin
Slide 4
Introduction
Statistical approaches:
(Fuzzy Logic, Logistic Regression, and Neural Networks…)
Deterministic approaches:
Physically-based models
(Iverson, 2000)
Steady state model
SHALSTAB MODEL(Montgomery et al., 1994)
SIMMAP (Pack et al., 1998)
Transient model
TRIGRS (Baum et al., 2002)
Basic:
Modelling of slope the hydrological response
Topography is considered
Slide 5
Objective
To establish a slope-instability analysis and a hydrological
model for landslide prediction during heavy rainstorms.
FS?
θ
Soil
Slide 6
Application of hydrology model
Review
Method
Data
Result
SHALSTAB
(Dietrich and Montgomery, 1994)
- Steady- state hydrological conditions
- Fully saturated conditions
- Homogeneous soil
- Slop-parallel groundwater flow
- Impermeable basal boundary
Hydrological model
Critical Rainfall
Infinite slope
qcr
Slide 7
Application of hydrology model
TRIGRS
(Baum et al., 2002)
Review
Method
Data
Result
- Transient hydrological conditions
- Fully saturated conditions
- Homogeneous soil
- Slope-parallel watertable
- Impermeable or infinite basal boundary
Infiltrate model
solution of Richard’s equation
Infinite slope model
(Z , t )
FS ( Z , t )
Slide 8
Application of hydrology model
Review
Method
Data
Result
(Casadei., 2003)
- Transient hydrological conditions
- Unsaturated conditions
- Slope-parallel watertable
- Homogeneous soil
- Impermeable boundary
Coupled hydrological–slope stability model
Groundwater table above the slip surface
hw (t )
Infinite slope model
FS (t )
Slide 9
Flow chart
Input data
Rainfall
Data
DEM Data
Area
Data
Soil
Parameter
Groundwater level
Estimation model
Hydrologic
Parameter
Infinite Slope
Stability Model
Output data
Groundwater
Height
Safety
Factor
Landslide
susceptibility Map
Slide 10
Hydrologic Model
Review
Method
TOPMODEL(TOPgraphy based hydrological MODEL)(Beven.
et al., 1979)
Stream line
Data
Result
Contour line
Specific Catchment Area a = A/b
Unit contour
length b
Contributing area A
Slide 11
Topographic index
Review
Method
Data
Result
a
T ln
tan
a: Specific area
tan β :slope
Slide 12
TOPMODEL Structure
Review
q
total
q
overland
q
subsurface
Method
Data
Srz
Result
Suz
Groundwater table
qv
Zw
Srz : Root zone
Suz : Unsaturated zone
D:Soil Depth
Zw:Water table height
qv :垂直入滲率
Slide 13
TOPMODEL Assumption
q
total
q
overland
q
subsurface
Saturated defict, zj
Q j T0e
(
zj
m
This equation can be solved for z:
)
c tan
Steady-state
Q j RA
a
z z m[ ln(
)]
tan
j
j
z j m ln(
R
a
) j m ln(
)
T0
tan
z m ln(
r
) j ] m
T0
1
a
ln(
)j
A j
tan
Slide 14
Hydrologic Routing
Review
Method
Data
Result
a
z j z m[ ln(
)j]
tan
The recharge to the groundwater table
q v, j SUZ j /(t d Z j ) , Q v qv , j A j
j
The baseflow from the saturated zone
z
Q Q exp( )
m
b
0
z(t 1) z(t)
Q b (t ) Qv (t )
t
A
Slide 15
Hydrologic Routing
Review
Method
Data
Result
a
z j (t 1) z (t 1) m[ ln(
)j]
tan
The recharge to the groundwater table
q v, j SUZ j /(t d Z j ) , Q v qv , j A j
j
The baseflow from the saturated zone
z
Q Q exp( )
m
b
0
z(t 1) z(t)
Q b (t ) Qv (t )
t
A
Slide 16
Infinite Slope MODEL
Review
依據有效應力概念,土體之阻滯力 τr 表示如下
Data
C ( u ) tan
C ( Z cos Z cos ) tan
Result
土體之駟動力τd為飽和土體沿坡面之分力
Method
r
2
2
s
s
w
Z sin cos )
d
s
resistance force
FS
driving force
C ( Z Z ) cos tan
Z sin cos
2
s
w
s
w
FS>1 Stable
FS<1 Unstable
Slide 17
Infinite Slope MODEL
Review
resistance force C ( Z Z ) cos tan
FS
driving force
Z sin cos
2
Method
s
w
w
s
Saturation ,W
Data
Result
FS
C
Z tan
[1 ( )]
Z sin cos
Z tan
w
s
w
s
TOPMODEL calculate local
storage deficit Zi
C
Z Z tan
FS
[1 (
)]
Z sin cos
Z
tan
w
s
i
s
Distribution of the soil saturation induced slope instability
Slide 18
Study Area
Review
Method
Data
Result
Slide 19
Study Area
Review
Method
Data
Result
Slide 20
Geological Map
Review
Method
Data
Result
Slide 21
Geo-material Parameter
Review
參數值
Method
Data
Result
土壤單位重,γS (KN/m3)
說明
zone1
18.2
實驗
zone2
19.6
實驗
zone3
參數
(zone1+zone)/2
坡度,θ (。)
ArcView軟體計算
由DTM推估
土壤深度,Z (m)
經驗式
由坡度推求
有效凝聚力,C (KPa)
搭配NDVI推估
抗剪摩擦角,ψ (。)
有效範圍推估
Slide 22
TOPMODEL Parameter
Review
參數
參數值
說明
Method
坡度,θ (。)
ArcView軟體計算
由DTM推估
Data
比集水面積,a/b (m)
ArcView軟體計算
由DTM推估
Result
土壤導水率隨時間衰減,m
0.032
退水歷線推估
0.0000328
流量觀測
土壤飽和時導水率,T0(m2/h)
5
林地飽和滲透
係數推估
未飽和層下滲延遲時間,
TD(h)
50
假定值
根系層最大容許儲蓄水量,
Srmax(m)
0.05
林地最大蓄水
量
初始根系層含水量,Sr0(m)
0.002
假定值
初紿流量,q0(m/h)
Slide 23
Landslides induced by AERE Typhoon
Review
Method
Data
Result
Slide 24
Topographic index
Review
Method
Data
Result
Slide 25
Local saturation deficits
Review
Method
Data
Result
Given
• m=0.032
• .Z 0.317
T=0
Compute
• =5.43
a
Z Z m ln(
tan
i
i
)
Slide 26
Soil profile saturation
Review
Method
Data
Result
Z w D Zi
w
Saturation > 1.0, Set to 1.0
Saturation < 0 , Set to 0
Zi
Zw
Z
Slide 27
Soil profile saturation
Review
Method
Data
Result
Slide 28
Rainfall Process
Review
Method
Data
Result
T=36
Slide 29
Prediction
Review
T=36
Method
Data
Result
Distribution map of slope failures of actual vs predicted by model
Slide 30
Classification Error Matrix
Review
Method
X
Data
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Result
非山崩網格
山崩網格
X
X
預測的山崩網格,坐落
於非山崩網格中
預測的山崩網格,坐落
於山崩網格中
Slide 31
Classification Error Matrix
Review
Method
Data
Result
The results of both are quantitatively compared by using two
following evaluation indexes.
AR= ( N1 ) / ( N1+ N2+ N3+ N4 )
FAR= ( N3 ) / ( N1+ N2+ N3+ N4 )
(1)
(2)
Here, AR is accuracy rate (%), FAR is false alarm rate (%) and N1,
N2, N3 and N4 (see Table )
Predict
Actual
Unstable
(FS < 1)
Stable
(FS ≧1)
Unstable
N1
N2
stable
N3
N4
N1, N2, N3, N4 : number of cell on each category
Slide 32
Classification Error Matrix
Review
Method
The results of both are quantitatively compared by using two
following evaluation indexes.
Data
Predict
Unstable
(FS < 1)
Stable
(FS ≧1)
Unstable
613
451
stable
13402
107343
Result
Actual
N1, N2, N3, N4 : number of cell on each category
AR= 57.61%
Slide 33
崩塌地分析成果及評估
文獻回顧
研究方法
正確率的計算方式可以分類誤差矩陣表(classification
error matrix)進行表示,其計算方式如下:
Total=121809
預測結果
研究資料
1
山崩(14015)
(FS < 1)
非山崩
(FS ≧1)
山崩(1064)
613
451
非山崩
(120745)
13402
107343
初步成果
未來工作
原始資料
山崩正確率=57.61%
非山崩正確率=88.91%
總體正確率=88.62%
Slide 34
Future Work
Calibration parameter
Model validation
Comparing the results
34
Slide 35
Thank you for your attention