Anthropogenic forcings on shallow landslides triggering Maria Cristina Rulli,
Politecnico di Milano
Factors which influence or trigger mass movement
Slope
Antrophogenic
and natural
activities
Water
0 < z < hi1
hi1 < z < d1
z = d1
d1 < z < d1+hi2
d1+hi2<z<d1+d2
z = d1+d2
[ ]( ) ααγ
φααγsenWzWzcccFs
d
dvA
⋅+⋅⋅⋅+⋅⋅+++
=cos
tancoscos
1
12
11
( )[ ]{ }αααγγ
φααγγγsenWsenhzh
WhzhccFssatidi
wsatidiv
⋅+⋅⋅⋅−+⋅⋅⋅+⋅−⋅−+⋅++
=cos])([
tancoscos)(
1111
12
11111
( )[ ]{ }αααγγγ
φααγγγγsenWsendzhh
WdzhhccFsdsatSdi
dwsatSdiA
⋅+⋅⋅⋅−+⋅+⋅⋅⋅+⋅⋅−+−⋅+⋅++
=cos])([
tancoscos)(
211111
22
2111112
( ) ( )[ ]{ }αααγγγγ
φααγγγγγγsenWsenhdzhhh
WhdzhhhcFssatidisatSdi
wsatidiwsatSdi
⋅+⋅⋅⋅−−+⋅+⋅+⋅⋅⋅+⋅−⋅−−+⋅+−⋅+⋅+
=cos])([
tancoscos)(
221221111
22
2212211112
αφ
tantan brFs =
( )[ ]{ }αααγγ
φφααγγγsenWsenhh
WhhccFssatSdi
wsatSdi
⋅+⋅⋅⋅+⋅⋅⋅+⋅−⋅+⋅+
=cos][
)),tan(min(coscos),min(
1111
212
111121
Precipitation induced landslides at basin scale
The problem is usually solved in term of SF by coupling an hydrological model with a geomechanical model
Rosso R, Rulli M.C. Vannucchi G., WRR 2006
Roads effect on shallow landslides triggering
Roads Effects
Road surface
and Ditches
Road Cut Culverts
Add new Channel to
stream network
Increase Overland
Flow
Intercept Subsurface Flow
Increase quick discharge
Increase and anticipation of peak discharge
Channel Network Extension ( )
=∆
+=
cutroad
rainquick
quickbase
Q
QQtQ
Volumetric and Timing Effect (From Wemple, et al.
1996)
With road
Without road
Time
Hydrological Effects:
a. Outsloped
b. Insloped
c. Crowned
a b c
Rainfall Effect on the road
roadrrain AIQ ⋅=
Runoff generation on the road and on the ditch
Hp: Impervious Road Surface Small roughness
Negligible vegetation transpiration and interception
Hydraulic Analogy: Road - Channel
Flow Interception by the ROAD
Flow Interception by the ROAD
Effect of the road cut
Overland and Subsurface Flow interception
subsurfaceoverlandroadcut QQQ +=
hint
hwt
d
( )
( ) wt
int le subsurface subsurface
wt
wt int Overland-road subsurface
h h d Q
h h d h Q
− ∝
− − ∝
→
→
Drainage Density
AL
D Sd∑= ( )
ALLLL
D CeRgRcSd∑ +++
=′
river
Ephemeral Channel
LC
e
LRg
LRc
LS
Culvert not connected
Culvert connected to ephemeral channel
Culvert connected to the actual drainage network
Drainage paths Changes
The intercepted discharge by the road is routed downslope trough the culvets
Flow Interception by the ROAD
The presence of ROADS can modify
contributing area at
selected point
OUTLET
Road
Road
Sovraccarico dovuto al terreno di riporto
Cut slope landslides triggering
Fillslope landslides triggering
Debris flow
Increase of shallow landslides
Road Influence
Geometric Effect
Geo-mechanic Effect
Hydrologic Effect
Increase of slope in
cutslope and fillslope areas
Load due to fillslope
Increase of soil water content in
downslope culvert areas.
Loss of vegetation
Debris flow are not taken in account in the present analysis
Geomechanical Effects:
Shallow landslides typology
Hillslope Slides (infinite slope analysis) Cutslope Slides (Stability maps) Fillslope Slides (Stability maps)
Hillslope SLIDES Infinite Slope Analysis
0 < z < hi1
hi1 < z < d1
z = d1
d1 < z < d1+hi2
d1+hi2<z<d1+d2
z = d1+d2
[ ]( ) ααγ
φααγsenWzWzcccFs
d
dvA
⋅+⋅⋅⋅+⋅⋅+++
=cos
tancoscos
1
12
11
( )[ ]{ }αααγγ
φααγγγsenWsenhzh
WhzhccFssatidi
wsatidiv
⋅+⋅⋅⋅−+⋅⋅⋅+⋅−⋅−+⋅++
=cos])([
tancoscos)(
1111
12
11111
( )[ ]{ }αααγγγ
φααγγγγsenWsendzhh
WdzhhccFsdsatSdi
dwsatSdiA
⋅+⋅⋅⋅−+⋅+⋅⋅⋅+⋅⋅−+−⋅+⋅++
=cos])([
tancoscos)(
211111
22
2111112
( ) ( )[ ]{ }αααγγγγ
φααγγγγγγsenWsenhdzhhh
WhdzhhhcFssatidisatSdi
wsatidiwsatSdi
⋅+⋅⋅⋅−−+⋅+⋅+⋅⋅⋅+⋅−⋅−−+⋅+−⋅+⋅+
=cos])([
tancoscos)(
221221111
22
2212211112
αφ
tantan brFs =
( )[ ]{ }αααγγ
φφααγγγsenWsenhh
WhhccFssatSdi
wsatSdi
⋅+⋅⋅⋅+⋅⋅⋅+⋅−⋅+⋅+
=cos][
)),tan(min(coscos),min(
1111
212
111121
0
1
2
3
4
5
6
7
8
9
10
0
0.2
0.4
0.6
0.8 1
1.2
1.4
1.6
1.8 2
2.2
2.4
2.6
2.8 3
3.2
3.4
3.6
3.8 4
Profondità z rispetto al piano campagna (m)
Fs
most probable
sliding surfaces
Hypothesis: • Circular failure surface • hillslope simple geometry • Soil having friction and cohesion • Soil homogeneous and isotropic
Friction circle method
Mohr-Coulomb: Safety Factor: with and
φστ ′′+′= tanc
d
CFS P
cNF′
=
ctgPe
C ′′
=φλ φ
tWq
WWd
HqHPµµµγγ −+
=Wq
WWe
HqHP'
'µµγγ −+
=
HcNγ
=
Taylor, 1948
Cutslope and Fillslope SLIDES: Stability Charts analysis
BASIN PARTITIONING: state of the art The model is a contour-based model. It derives at first the drainage network
starting from the highest contour, then proceeds downslopes follow the steepest lines (Menduni et al., 2000). Two type of topographic elements are considered in
flow accounting and routing:
OUTLET
Cell : polygon having two vertexes on two adjacent contours and two vertexes on two steepest lines. It can be triangular when 2 steepest lines meet.
Channel: mono-dimensional topographic element starting at the junction of 2 steepest lines
The effect of presence of ROAD on basin partitioning
Rulli M.C., Rosso R., 2008
Hydrological Fluxes
RFt RFt+1
Rock Rock
Saturated Saturated
Saturated Saturated
Unsaturated
Unsaturated
BIN
ST2
LIN
ST1
INt
BIN
SF2t
EX2t+1
IE2t
SF1t
SE2t+1 EX1t+1
OFt
IE1t SE1t+1
OFt+1
SF1t+1
SF2t+1
Zi1
Zs1
Zi2
Zs2
L1
L2
Simplified Bucket Model with Layers
• Saturated Hydraulic Conductivity: dove • Water Content:
• Flux unsaturated -saturated:
• Subsurface Flow:
• Infiltration:
• Percolation in the cracked Rock: con
fzS eKK −= 0 m
f dryS θθ −=
( ) ( )drySis zzS θθ −⋅−=
( ) SzS sdrySd −⋅−= θθ
( ) SidrySd IzI −⋅−= θθ
dS III −=
d
SS S
IKst =
( ) mS
S
d
eKsf−
⋅⋅= αtan
2
1
di I
SKlin =
tKbin ar ∆⋅= mfar KKbeK +=
Rulli M.C., Rosso R., AWR 2007
STUDY AREA Study area is watershed 3 (WS3) at the H. J.
Andrews Experimental Forest, Oregon. Location of roads, culverts, and instrumentation are shown.
• Elevation 470 – 1050 m a.s.l.
• Area 1 km2
• Clay Soil
• Macropores in the soil
• Ks =10-3 ÷ 10-5 m/s
• Mean Annual Precip. 2300 mm
• Mean rainfall Intensity 4 mm/h
• Mean Discharge 5*10-2 m3/s
• In 1959 3 orders of roads were built.
• Rainfall event 3-9 February
1996: max rain 12 mm/h Total rain 340mm • 16 shallow
landslides: 8 hillslope 8 fillslope/cutslope
Rulli M.C., Rosso R., 2008
Good Model reproduction
Downslope road: landslides triggering is mainly due to the hydrologic effect.
Upslope road:
landslides triggering is mainly due to the geometric effect.
Scenario 1: increase of the number of culverts
Observed slides:
The increase of the number of culverts decrease the slides because decrease the contributing area at any single culvert
By increasing the number of roads dramatically increase the number of slides.
Scenario 2: increase of the number of roads
Fire Forcing on Hydrological Processes
Increased overland flow, sediment yield and shallow landslides
susceptibility
Vegetation cover destruction & canopy cover
& roots at 10-20 cm of depth
Effects on soil properties k increase in erodibility
k decrease in infiltration capacity k decrease of cohesion
Formation of a water repellent layer ? formation mechanism ? thickness position duration soil type temperature vegetation cover water content
FIRE INDUCED SHALLOW LANDSLIDES TRIGGERING
Rulli M.C.,Bocchiola D., Rosso R., JoH 2006
0
100
200
300
400
500
0 1000 2000 3000 4000 5000Time [minutes]
Rai
nfal
l [m
m]
18/07/2005 debris-flow19-20/09/200028-29-30/09/200012-13-14-15/10/200014-15-16/11/2002
THE EVENT OF 18 th OF JULY, 2005
In the period from the beginning of the year 2000 to the day of the forest fire, 47 rainfall events
occurred having or intensity or total greater than the rainfall event
triggering the mass movement of July 2005.
It means that the role of forest fire in triggering mass movement should be
taken in consideration.
Severe rainfall event. About 40 mm of rainfall dropped in less than 30 minutes (TR= 10 years ) on a previously burned soil.
On 16 th March 2005 a severe forest fire occurred in the area of Rio Casella burned more than 70% of the study area.
Mass Movement Triggering (30000 m3)
1.6 km2 catchment (Rio Casella) located
in the Piedmont, Italy
Major Features Geology: granitic gneiss; in the highest area, large zones with exposed granitic formations
Climate: mediterranean Rain: mainly in May and
October (MAP is 1300 mm) Vegetation: Chestnut
Hydrography: ephemeral streams
Topografy: steep slopes, elevation range 1600-250 a.s.l.
The Study Area
Rulli M.C., Rosso R., GRL 2006
SBM Parameters
POST-FIRE
TRANSIENT/ 16 MONTHS after the forest fire
PRE-FIRE
Soil depth upslope [m] 1 1.5 2.5 Soil depth downslope [m] 1.5 2 3
Ks upslope [m/s] 0.000001 / 0.00005
0.000001 / 0.00007
0.000001 / 0.00015
Ks downslope [m/s] 0.000001 / 0.00005
0.000001 / 0.00007
0.000001 / 0.00015
Multipling factor for horizontal Ks [-] 1 1 1
Rate of decay in Ks with depth [-] 0.3 0.3 0.3
Satured soil water content θS [-] 0.4 0.4 0.4
Residual soil water content θN [-] 0.1 0.1 0.1
Initial satured store of the bucket [%] 10 - 30 10 - 30 10 - 30
Initial soil water content - upslope [-] 0.1 - 0.2 0.1 - 0.2 0.1 - 0.2
Initial soil water content - downslope [-] 0.1 - 0.2 0.1 - 0.2 0.1 - 0.2
POST-FIRE
TRANSIENT/16 MONTHS
PRE-FIRE
Manning coefficient - Slopes [s m-1/3] 0.04/0.015 0.04/0.025 0.04/0.04
Manning coefficient - Channels - upslope [s m-1/3 ] 0.015 0.025 0.04
Manning coefficient - Channel –downslope[ s m-1/3 ] 0.015 0.025 0.04
Critical Support Area [-] 50000 50000 50000 Channel Width Scaling Factor [-
] 0.26 0.26 0.26 Outlet Flow Width [m] 3.7 3.7 3.7
Time Step for the Kinematic Wave [s]
60
60
60
Field Data Vegetation Survay (Pre; Post and Transient)
Soil depth measuremts
WDPT test
Soil type
Pedology
Geology
Hydraulic conductivity measurements (Pre; Post and Transient)
Modeling and predicting hydrological and sedimentological
response of burned areas
To investigate the influence of changes in
land cover and soil characteristics, from
pre- to post-fire conditions, on runoff production and shallow landslides triggering, during a rainstorm
Deterministic model
To provide a deterministic framework for studying
shallow landslides
susceptibility in the
different temporal scenarios
Simulation Runs Historical
Rainfall Records
POST-
FIRE
POST-FIRE/ 16
MONTH PRE-FIRE
Friction Angle [°] 42 42 42 Saturated Soil Spec. Weight
[KN/m³] 21.5 21.5 21.5 Cohesion [KPa] 0 4 10
Apparent Cohesion [Kpa] 0 0 0
Temporal Scenarios
- pre-fire: the period of time before the forest fire occurred
- post-fire: the period from the time when the fire estinguished to the end of the spring in the subsequent year
- transient: the period of time required for the restoration of the soil, following fire induced disturbance
POST FIRE
TRANSIENT
PRE FIRE POST FIRE
Rulli M.C., Rosso R., AWR 2007
Conclusions: Anthropogenic and natural activities like roads and forest fires can strongly influence the hydrologic response and shallow landslides susceptibility of upland catchments!
For studying anthropogenic and natural activities influence on shallow landslides triggering it is necessary correctly evaluate the effects of these activities on the hydrological fluxes (generation and direction).
For studying anthropogenic and natural activities influence on shallow landslides triggering it is necessary correctly select the gomechanical model!
THANK YOU!
( )[ ]( )[ ] θ
φωω
tan'tan
11
⋅−⋅⋅+⋅+⋅+⋅−⋅+
=rrs
rrsSeSeGSeSeGFS
−−>
−−≤
−−
=;*1lnif,1
;*1lnif,1exp1*
1
11
ppA
ppA
App
ξ
ξξω
aTbp θsin* =
( )rSe
eA −⋅+
= 111
ω = h/z ξ = (p*/z)t
Precipitation threshold for slope instability Coupling hillslope hydrology with geomechanics yields landslide triggering by precipitation
( )( )
−+
−−
−+
−−
−⋅−⋅−+
−⋅⋅+
=
taz
TbeSee
taz
TbeSee
zh
See
SeG
abT
tp
r
r
i
r
rs
CRθ
θ
φθ
φθ
θ
sin1exp1
sin1exp
'tantan111
'tantan1)(
sin
Temporal scale of hillslope evolution The rainfall rate pF that can be exceeded with a probability of (1 – F) in a year ( ) )1(
1n
FF txmtp −−=
pCR = pCR(t) ( ))1(
1sin1exp1
sin1exp
'tantan111
'tantan1)(
sin
nF
r
r
i
r
rs
txmt
azTb
eSee
taz
TbeSee
zh
See
SeG
abT
−−=
−+
−−
−+
−−
−⋅−⋅−+
−⋅⋅+
θ
θ
φθ
φθ
θ
10
100
0.1 1 10 100
t [d]
pC
R a
nd p
F [m
m/d
]
a/b = 200 m
TR = 500 yeaTR = 70 yearsTR = 15 years
a/b = 50 m
a/b = 100 m
a/b = 200 m
10
100
0.1 1 10 100
t [d]
pC
R a
nd p
F [m
m/d
]
a/b = 200 m
10
100
0.1 1 10 100
t [d]
pC
R a
nd p
F [m
m/d
]
a/b = 200 m
TR = 500 yeaTR = 70 yearsTR = 15 years
a/b = 50 m
a/b = 100 m
a/b = 200 m
Maps of Mettman Ridge catchment showing shallow landsliding prone areas in term of return period of potential failure considering initial condition of stable piezometric at the depth of bedrock, h(0) = 0 and h(0) = 0.15, φ’ = 45°, T = 65 m2/d, ρs = 1600 kg/m3 and Gs = 2.60 .