Document Actions

IJEGE-11_BS-Baum-et-alii

background image
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
825
DOI: 10.4408/IJEGE.2011-03.B-090
ASSESSING SUSCEPTIBILITY AND TIMING OF SHALLOW LANDSLIDE
AND DEBRIS FLOW INITIATION IN THE OREGON COAST RANGE, USA
R
ex
L. BAUM
(*)
, J
onatHan
W. GODT
(*)
& J
effRey
A. COE
(**)
(*)
U.S. Geological Survey, Box 25046, M.S. 966, Denver, CO 80225-0046 USA
INTRODUCTION
Information on where and when debris flows are
likely to occur is greatly needed to reduce resultant
losses and deaths. Shallow, rainfall-induced landslides
that transform into debris flows commonly occur un-
der conditions of transient infiltration into initially un-
saturated soils. Application of mathematical models
of transient, unsaturated infiltration and slope stability
over broad regions to assess timing and potential loca-
tions of debris flow requires understanding of unsatu-
rated zone hydrology and soil mechanics as well as
information on rainfall, topography and the distribu-
tion and properties of hillside soils. In this paper, we
survey techniques for generating the spatial and tem-
poral input data for such models and present example
calculations for a debris-flow producing storm using
our spatially distributed mathematical model for Tran-
sient Rainfall Infiltration and Grid-based Slope Stabil-
ity (TRIGRS) (b
aum
et alii, 2008; 2010). The model
combines an analytical solution to assess the pore-
pressure response to rainfall infiltration into unsatu-
rated soil with an infinite-slope stability calculation
to estimate the timing and locations of slope failures.
Pore-pressures and factors of safety are computed on
a cell-by-cell basis and can be displayed or manipu-
lated in a grid-based geographic information system
(GIS). Input data are high-resolution topographic data
and simple descriptions of initial pore-pressure dis-
tribution and boundary conditions for a study area in
western Oregon, USA. Rainfall information was taken
ABSTRACT
Effective management of debris-flow hazard relies
on accurate assessments of debris-flow susceptibility.
Mathematical models of rainfall infiltration and slope
stability can be applied to predict the temporal and
spatial variation of debris-flow susceptibility. These
models require high-resolution (<10 m) topographic
data, as well as (ideally also high-resolution) data on
initial groundwater conditions, physical properties of
near-surface earth materials, depth to bedrock, and
rainfall. A case study from the Oregon Coast Range,
USA illustrates the use of generalized data from a soil
survey, limited field measurements, and simple mod-
els to parameterize a combined infiltration and slope
stability model for predicting debris-flow timing and
source-area locations. Although the model over-pre-
dicts the extent of debris-flow source areas, results are
consistent with mapping which shows channels to be
the preferred source areas. Simulation of a November
1996 storm that produced debris flows in the study
area indicates that instability probably developed near
the end of the period of most intense rainfall; how-
ever precise timing of debris flows in the study area is
unknown. Model results also indicate that differences
in rainfall interception between forested and logged
areas may account at least in part for the observed dif-
ferences in debris-flow susceptibility.
K
ey
words
: debris-flow susceptibility, rainfall infiltration,
rainfall interception, Oregon Coast Range
background image
R.L. BAUM, J.w. GODT & J.A. COE
826
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
and NIEM, 1990). The bedrock weathers to highly
permeable colluvium consisting of abundant sandstone
clasts in a matrix of sandy silt or silty sand, covered
with a layer of decomposed plant material 0 - 10 cm
deep (NCRS, 2010). Depth to bedrock is commonly
25-110 cm deep (NCRS, 2010) except in hollows,
where it can be greater than 300 cm deep (s
CHmidt
et
alii, 2001). Vegetation consists of Douglas Fir, Red Al-
der, and associated species of coniferous, hardwood,
and understory plants (s
CHmidt
et alii, 2001).
MODELING APPROACH
We apply the TRIGRS model to predicting the
shallow landslide source areas of debris flows in the
study area. In this section we briefly describe the theo-
retical basis of the TRIGRS model (b
aum
et alii 2008;
2010) and review required input data and strategies for
defining these inputs. Specific strategies applied to the
study area are described in more detail in a later section.
TRANSIENT VERTICAL GROUNDwATER FLOw
The characteristic time scales for lateral and slope
normal pore-pressure transmission in initially wet, ho-
mogeneous, isotropic hillslopes are A/D
0
, and H
2
/ D
0
,
respectively, where A is the upslope contributing area
above a given location, D
0
is the saturated hydraulic
diffusivity, and H is the depth of the failure surface in
the slope-normal direction. For areas of the landscape
where the ratio
, long- and short-
term pressure-head response to rainfall is dominated
by vertical flow implying that pore-pressure variation
in initially wet materials can be adequately described
by simplified, one dimensional forms of the govern-
ing equation for infiltration (IVERSON, 2000). The
Richards equation describes infiltration and vertical
flow through the unsaturated zone (f
Reeze
& C
HeRRy
,
1979). When transformed to account for the effects of
a sloping surface the equation is
where θ is the volumetric water content, ψ is the pres-
sure head, k(ψ) is the pressure head dependent hydrau-
lic conductivity, and δ is the slope angle. Following
s
Rivastava
& y
eH
(1991), we linearize equation 1 us-
ing G
aRdneR
s
(1958) exponential hydraulic parameter
models, k(ψ) = k
s
exp(αψ) and θ = θ + (θ
S
- θ
r
)exp(αψ),
where α is the air-entry head or inverse of the height of
the capillary rise, k
S
is the saturated hydraulic conduc-
from a nearby recording rain gage. Material strength
and hydraulic properties were gleaned from the lit-
erature and measured both in the field and laboratory.
Results are tested by comparison with an inventory
of shallow landslides that mobilized to debris flows.
SETTING AND CLIMATE
Debris flows typically initiate from shallow land-
slides in the Oregon Coast Range, where hillsides are
susceptible to shallow landslides and debris flows due
to steep topography, heavy rainfall, and land-use ac-
tivities associated with timber harvesting (e.g. b
Rown
& k
RyGieR
, 1971; P
ieRson
, 1977). Recent work in the
region on shallow landslide and debris-flow initiation
has focused on the influence of vegetation roots on
slope stability (s
CHmidt
et alii, 2001; R
oeRinG
et alii,
2003) and the role of pore-water response from rain-
fall in soil and shallow bedrock (m
ontGomeRy
et alii,
2009; e
bel
et alii, 2010). The topography of the area is
deeply dissected and steep (> 30º). Shallow landslides
tend to occur in bedrock hollows (d
eitRiCH
& d
unne
,
1978), where colluvium is several meters thick (b
en
-
da
, 1990). Average annual precipitation of about 2000
mm falls mainly as rain during the winter wet season
of November to March (d
aly
et alii, 1994). The focus
of this study is the North Charlotte Creek basin located
about 20 km ESE of Reedsport, OR (Fig. 1) in the
Elliott State Forest. The forest is administered by the
Oregon Department of Forestry (ODF). The basin is
underlain by the middle Eocene Tyee Formation; a tan,
medium- to coarse-grained thick-bedded sandstone
with interbedded layers of siltstone and shale (NIEM
Fig. 1 - Location of study area in Oregon, USA
(1)
background image
ASSESSING SUSCEPTIBILITY AND TIMING OF SHALLOW LANDSLIDE AND DEBRIS FLOW INITIATION
IN THE OREGON COAST RANGE, USA
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
827
landslide hazard mapping and modeling (e.g. s
CHulz
,
2007). These data are typically of very high spatial
resolution (<5 m) with small absolute elevation errors,
and can be processed to reveal the “bare earth” surface
beneath vegetation (s
latton
et alii, 2007).
Maps of the soil depth on steep hillsides are re-
quired for deterministic shallow landslide models
that include the effects of infiltration or soil cohesion
(G
odt
et alii, 2008a). Field observations of soil depth
in landslide-prone areas of landscapes with well-de-
veloped drainage networks indicate that colluvium
tends to collect in areas of topographic convergence
(hollows) (e.g. d
ietRiCH
& d
unne
, 1978; R
eneau
&
d
ietRiCH
, 1987). On steep slopes (> 20°) colluvium
depth tends to decrease exponentially with slope angle
(d
e
R
ose
et alii, 1991; d
e
R
ose
, 1996). Collecting suf-
ficient measurements of soil thickness to it map over
broad areas is a practical impossibility, and empirical
models have been used to create soil depth maps for
landslide susceptibility studies (e.g. d
e
R
ose
et alii,
1996; s
alCiaRini
et alii, 2006; G
odt
et alii, 2008b).
Such models can be constrained using information on
soil or colluvial depth from soil surveys, field investi-
gations, and geophysical mapping.
Linearized solutions for infiltration are gener-
ally very sensitive to initial water-table depths and
moisture conditions of the unsaturated zone and if ap-
plied to simulate actual landslide events require some
knowledge of initial conditions (G
odt
et alii, 2008a).
Where detailed information on groundwater condi-
tions is unavailable, parametric studies assuming a
range of initial conditions may indicate the conditions
that have caused shallow landslides in the past (C
Ros
-
ta
& f
Rattini
, 2003; s
alCiaRini
et alii, 2006).
Rainfall data are needed to determine the flux at
the ground surface for transient modeling of infiltration.
Attempts to recreate conditions for past landslide events
may benefit from rainfall data with high spatial (kilom-
eter scale) and temporal (hourly) resolution such as can
be obtained from weather radars (C
Rosta
& f
Rattini
,
2003). Empirical rainfall intensity-duration thresholds
can be used in applications designed to estimate the
likelihood of landslide occurrence for a given set of ini-
tial and rainfall conditions (G
odt
et alii, 2008b).
The strength and hydrologic properties of hillside
materials and an estimate of their spatial distribution can
be obtained using field and laboratory tests and soil sur-
veys or geologic mapping (G
odt
et alii, 2008a; 2008b).
tivity, θ is the moisture content, and θ
S
and θ
r
are the
saturated and residual moisture contents, respectively.
See b
aum
et alii (2008; 2010) for details of the solution
scheme and a discussion of limitations of this approach.
STABILITY OF INFINITE SLOPES
With some exceptions (e.g. R
eid
et alii, 2000),
models designed for assessing slope stability over
large areas using digital topography typically rely on
a statically determinate infinite-slope stability analy-
sis (e.g. m
ontGomeRy
& d
ietRiCH
, 1994; C
Rosta
&
f
Rattini
, 2003), which assumes slopes are infinitely
long and that the failure plane is parallel to the ground
surface (t
ayloR
, 1948). Stability of an infinite slope
is characterized by the factor of safety ratio, F
S
, of re-
sisting basal Coulomb friction to shear stress and is
calculated at an arbitrary depth Z by
where c' and f are soil cohesion and friction angle for
effective stress, respectively, γ
w
is the unit weight of
water, γ
S
, the unit weight of soil, and ψ is the ground-
water pressure head. For stability above the water
table where pressure heads are negative we use an
approximation of b
isHoP
s
(1959) effective stress
and calculate suction stress (l
u
& l
ikos
, 2004) as
the product of (θ - θ
r
)/(θ
S
- θ
r
) and ψ(Z,t)γ
w
(b
aum
et
alii., 2008; 2010). Slope failure is predicted where
F
S
<1, stability holds where F
S
≥1, and F
S
=1 indicates
a state of limiting equilibrium.
INPUT AND TEST DATA
Grid-based digital topography, soil thickness,
materials strength and hydrologic properties, initial
groundwater conditions, and time-varying rainfall
intensity and duration are required input data for ap-
plication of the model over broad geographic areas
(b
aum
et alii, 2008; G
odt
et alii, 2008a). Landslide
maps are needed to test modeling results and should
include information on sources, tracks, and deposits,
and where and when the landslides occurred.
Digital Elevation Models (DEMs) are regularly
spaced arrays of elevation values that are used to
calculate local topographic slope, δ. Of the emerg-
ing remote-sensing technologies used to generate
DEMs, Airborne Laser Swath Mapping (ALSM also
commonly referred to as LiDAR - Light Detection
and Ranging) has arguably had the most impact on
(2)
background image
R.L. BAUM, J.w. GODT & J.A. COE
828
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
The model data were managed in a geographic in-
formation system (GIS). Topography was represented
by a 3-m resolution DEM, obtained by resampling
a 1-m resolution “bare-earth” DEM derived from
LIDAR elevation data. We used a cubic convolution
resampling algorithm to smooth facets and other ar-
tifacts in the 1-m data caused by poor ground-point
densities in areas of thick vegetation. Slope is com-
puted using the standard D-8 algorithm (Fig. 2a).
Depth to bedrock was estimated using an expo-
nential function of slope angle (d
e
R
ose
, 1996), with
a modification to account for the effect of curvature.
Field measurements at a monitoring site several kil-
ometers to the southeast of the study area (Fig. 1)
revealed a strong secondary dependence of colluvial
thickness on topographic curvature. At locations of
equal slope, colluvium was thicker in concave areas
along axes of zero order drainage basins and thin-
ner on convex interfluves between the basins. Depth
of bedrock ranged from 0.47 m to 3.1 m and slopes
ranged from 26° to 41°. We computed depth of bed-
rock, d
LZ
(Fig. 2b), in the study area using (5.0-1.5
sgn(κ))exp(-0.04δ), in which δ is the slope angle and
κ is the curvature. Using this model, colluvium is 2.3
m in concave areas and 1.2 m in convex areas on
26° slopes and 1.3 m (concave) or 0.68 m (convex)
on 41° slopes. Slopes steeper than 45° were treated
Soil strength parameters (angle of internal friction, f'
and cohesion c') and saturated hydraulic conductivity,
k
S
, are typically obtained from standard geotechnical
tests (e.g. s
avaGe
& b
aum
, 2005), however, tests to de-
termine the hydraulic conductivity function k(ψ), or the
relation between pressure head and moisture content,
θ(ψ) are needed to estimate the air-entry head, α. Predic-
tion of these relations and parameters from other more
easily obtained properties such as particle-size distribu-
tions and bulk density is possible (e.g. l
eiJ
et alii, 2002).
Plant roots are thought to impart significant strength to
hillside soils (e.g. s
CHmidt
et alii, 2001); however, be-
cause the resisting forces imparted by plant roots do not
typically act normal to the failure plane, incorporating
their effects into one-dimensional infinite-slope stability
models (i.e. equation 2) is problematic.
MODEL APPLICTION
We applied the TRIGRS model (b
aum
et alii,
2008: 2010) to a 2.4 km² study area in the Oregon
Coast Range to study model parameterization and
performance in an area with a strongly developed
dendritic drainage pattern. Heavy rainfall during No-
vember 1996 produced numerous debris flows in the
central Oregon Coast Range (w
iley
, 2000), including
those mapped in the study area (Fig. 2). Specific tim-
ing of the debris flows during the rainfall is unknown.
Fig. 2 - Maps showing (a) topographic slope angle, δ, w is watershed boundary, L is logged area, D is debris flow inventory from
c
oe
et alii 2010, (b) depth to bedrock, d
LZ
, (c) distribution of ratio ε, (d) property zones, 1 is colluvium and 2 is bedrock
background image
ASSESSING SUSCEPTIBILITY AND TIMING OF SHALLOW LANDSLIDE AND DEBRIS FLOW INITIATION
IN THE OREGON COAST RANGE, USA
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
829
as bedrock outcrops even though thin discontinuous
patches of colluvium locally exist there. Assignment
of steep slope areas to bedrock outcrop results a
higher proportion of rock outcrop in the eastern one
third of the basin, consistent with NRCS (2010) soil
mapping in the study area (Fig. 2a).
The fall season initial water table depth for the
study area was approximated by assuming lateral flow
in the colluvium and weathered bedrock. We solved
Darcy’s law (f
Reeze
& C
HeRRy
, 1979) for the height
of flow above unweathered bedrock, H
f
,
H
f
= (A / w)(I
ZLT
/ k
S
)/(tanδ + 0.001),
in which A is the upslope contributing area, w is the
cell width (3 m), I
ZLT
is the average winter infiltration
rate, k
S
is the saturated hydraulic conductivity and,
slope of the ground surface, tanδ, approximates the
hydraulic gradient. The constant, 0.001, prevents divi-
sion by zero in flat areas. For this computation, both A
and tanδ were estimated using t
aRboton
s
(1997) D-∞
algorithm. We estimated initial water-table depth using
d = H
w
+ d
LZ
- H
f
; H
f
< H
w
+ d
LZ
- d
min
and
d = d
min
; H
f
H
w
+ d
LZ
- d
min
in which H
w
is the vertical thickness of weathered, per-
meable bedrock and d
min
is the minimum initial water
table depth. In this example, H
w
= 2.25 and d
min
= 0.75
m to ensure that F
S
> 1 at time t = 0 During fall and
winter months water commonly flows in many of the
channels; however, we lack data on the channel slope
angles where open channel flow begins.
Rainfall during the November 1996 storm was
estimated from the nearest available hourly record
at similar elevation and distance inland (RAWS rain
gauge, Fig. 1). The hourly data were approximated by
three time periods of constant rainfall intensity (Fig.
3). The forest canopy intercepts rainfall and modifies
the amount and intensity reaching the ground surface
(e.g m
ooRe
& w
ondzell
, 2005), and these effects
may reduce the potential for landslide generation
(k
eim
& s
kauGset
, 2003). Mature stands of Doug-
las fir may reduce the amount of rainfall reaching the
ground surface by as much as 40 percent (e.g. l
ink
et
alii, 2004). k
eim
et alii, (2004) present a stochastic
model that describes the reduction in rainfall inten-
sity for a given storm duration and return period. In
general, rainfall from high-intensity storms is affected
less than that from low-intensity storms. For the case
presented here, rainfall intensities were applied at
100% in logged areas (Fig. 2) and 90% in forested
areas to account for interception consistent with ob-
served storm intensity (k
eim
et alii, 2004).
We identified areas in the digital landscape (Fig 2c)
where vertical transmission of pore-pressure as imple-
mented in the TRIGRS model is dominant over lateral
transmission (i
veRson
, 2000) by comparing the slope
normal depth, H
D
, and the square root of the upslope
contributing area, A. Where the ratio
long and short-term pressure-head responses to rainfall
can be adequately described by 1-D linear and quasi-
linear approximations to the Richards equation (i
veRson
,
2000). We computed the upslope contributing area for
each grid cell using t
aRboton
s
(1997) D-∞ algorithm
and calculated slope-normal depths using the vertical
thickness (Fig. 2b) and topographic slope (Fig. 2a). Fig.
2c compares the distribution of mapped shallow land-
(3)
(4a)
(4b)
Fig. 3 - Rainfall and computed pressure head response (a)
Rainfall at the Goodwin Peak Remote Automated
weather Station (RAwS) during the November
1996 storm, A is average rainfall, P
1
P
2
, and P
3
,
are rainfall during first, second, and third, time
periods, respectively. (b) Computed pressure head
for a typical cell in the debris-flow source areas
H is hourly rainfall, A is Average rainfall, 3P is
hourly rainfall averaged over three time periods,
3P
90
is 90% of 3P, used in forested areas. (α=40°,
d=0.75 m, d
LZ
=1.3 m)
background image
R.L. BAUM, J.w. GODT & J.A. COE
830
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
water table remained unsaturated (Fig. 4). Expansion
of saturated areas up the channels into hollows is
consistent with observations in similar topographic
settings elsewhere (w
ilson
& d
ietRiCH
, 1987). Fac-
tor-of-safety gradually decreased in the steep chan-
nels, where debris flows occurred in 1996, as well as
in other similar channels. Due to higher infiltration
rates, pore pressures rose more rapidly in the logged
areas than in the forested areas (Fig 3b, 4). Conse-
quently the proportion of steep channel grid cells
where F
S
<1 by the end of the storm was greater in
the logged area than in the forested areas.
DISCUSSION
Several points related to data preparation, model
results, and interpretation of results merit further expla-
nation. In the case study presented, the TRIGRS model
produced results that are consistent with the observed
distribution of debris-flow source areas despite the
use of generalized and in some cases estimated data.
Although we are currently in the process of acquiring
more detailed data on physical properties and hydro-
logic response of debris-flow source areas in the Coast
Range, these results indicate that the TRIGRS model
can be applied successfully even where limited data are
available. The model over-predicts potential landslides
as a result of data uncertainty, particularly fine-scale
subsurface variations in the depth of bedrock, strength
and hydrologic properties, and initial groundwater and
soil moisture conditions. Consequently results must be
interpreted in a probabilistic sense.
GIS methods used in preparing some of the input
grids, notably the initial water table depth resulted in
artifacts that skew some of the model results. Both
the D-8 and the D-∞ methods of computing upslope
contributing area result in many cells adjacent to the
main channels having very small contributing areas.
Consequently (3), (4a) and (4b) computed very low
water tables at many grid cells near the valley bottom
(Fig. 4). A finite-difference or finite-element model for
steady saturated-unsaturated groundwater flow would
compute a more realistic initial water table, probably
at considerable computational expense.
Several factors may account for the much higher
density of debris flows in the logged area compared
to forested areas (Fig. 5). These include (1) greater
difficulty detecting debris flows beneath forest cov-
er (ODF, 1999), (2) the greater proportion of steep
slides (C
oe
et alii, 2010) and ε in the study area. Values
of ε are typically much less than unity along the steep
slopes and in the channel network where the mapped
shallow landslides (debris-flow source areas) lie. At most
points within the mapped landslides, ε = ≤ as suggested
by i
veRson
(2000) (Fig. 2c).
We used two property zones (table 1, Fig. 2d), bed-
rock outcrop (as defined previously), and colluvium.
Strength parameters of the colluvium were based on
published values for previous work in nearby areas
of similar geology (e
bel
et alii, 2010). Although the
contribution of root strength and pullout resistance
could be a further basis for distinguishing differences
between the logged and forested areas, implementation
using (2) is problematic, as noted previously. Hydrau-
lic properties of the colluvium were estimated from a
published soil survey (NRCS, 2010) and unpublished
monitoring data. Continuous unsaturated conditions
observed throughout the fall and winter rainy season
are consistent with α ≈ 3.0 m
-1
, rather than α ≈ 0.5 m
-1
as indicated by analysis of the soil survey data. Hydrau-
lic diffusivity, D
0
, was estimated as described by b
aum
et alii (2010) so that 2k
S
D
0
≤ 10k
S
.
Properties of weathered bedrock were estimated
by adjusting colluvial properties toward lower poros-
ity, permeability and diffusivity and higher strength
(b
eaulieu
& H
uGHes
, 1975; f
Reeze
& C
HeRRy
, 1979).
RESULTS
Simulated rainfall infiltration resulted in gradual
pore pressure rise and reduction in factor-of-safety.
Simulating infiltration using three periods of con-
stant rainfall rather than hourly data (35 periods)
preserves the major features and timing of pore pres-
sure rise (Fig. 3b) while reducing computational ef-
fort by about 90%. Infiltrating water percolated to
the shallow water table, causing it to rise gradually.
The water table rose above the top of weathered bed-
rock only in channel areas and colluvium above the
Tab. 1 - Material Strength and Hydraulic Properties for
the Two Map Units in Fig. 2d
background image
ASSESSING SUSCEPTIBILITY AND TIMING OF SHALLOW LANDSLIDE AND DEBRIS FLOW INITIATION
IN THE OREGON COAST RANGE, USA
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
831
Fig. 4 - Maps showing evolution of pressure head, ψ at base of colluvium at different stages of the three rainfall periods
in Fig. 3 and boundary and initial conditions depicted in Fig. 2. Rainfall after 20 h=85 mm, after 30 h =161 mm
and after 35 h = 201 mm. Rectangle in upper left panel (0 h) indicates approximate area shown in other panels, L
is logged area, D is boundaries of debris flows
Fig. 5 - Maps showing evolution of factor of safety. Rectangle in upper left panel (0 h) indicates approximate area shown
in other panels, L is logged area, D is boundaries of debris flows
background image
R.L. BAUM, J.w. GODT & J.A. COE
832
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
slopes in the logged area (Fig. 2), (3) differences in
evapotranspiration, which might result in drier ini-
tial conditions in the wooded areas, (4) loss of root
strength and pullout resistance in logged areas, and
(5) interception of rainfall by the forest as considered
in our model. We are confident that all debris flows
in the study area that resulted from the November
1996 rainfall are shown on the map (Fig. 4). Given
sufficient data the TRIGRS model could account for
the other four factors. Our model accounts for two of
these (distribution of steep slopes, and interception)
and our results indicate that even the modest effect of
interception explains at least some of the difference in
susceptibility between logged and wooded areas.
CONCLUSIONS
The TRIGRS model was used to simulate rainfall
infiltration and occurrence of shallow landslides from
a November 1996 storm that resulted in widespread
landslides and debris flows in the Oregon Coast Range.
• Modeling results are consistent with observations
that steep channels in this setting are more suscepti-
ble to shallow landslides and debris flows than either
hillsides or basin rims as observed in other settings.
• Parameter estimation for the TRIGRS model can
be based on generalized data and simple models of
soil depth distribution and initial water table, where
site-specific data are lacking. Availability of more
specific data would reduce model uncertainty.
• Storm rainfall interception can account for at least
part of the observed landslide susceptibility diffe-
rence between logged and forested areas.
ACKNOWLEDGEMENTS
Jon McKenna (USGS) shared observations about
the debris-flow scars and bedrock exposures in the
clear cut area. John Seward (ODF) granted access to
the study area. Joseph Gartner and Dianne Brien as
well as two anonymous reviewers made suggestions
that greatly improved the paper.
REFERENCES
b
aum
R.l., s
avaGe
w.z. & G
odt
, J.w. (2008) - TRIGRS - A Fortran program for transient rainfall infiltration and grid-based
regional slope-stability analysis, version 2.0. U.S. Geological Survey Open-File Report 2008-1159.
b
aum
R.l., G
odt
J.w. & s
avaGe
, w.z. (2010) - Estimating the timing and location of shallow rainfall-induced landslides using
a model for transient, unsaturated infiltration. Journal of Geophysical Research, doi:10.1029/2009JF001321.
b
eaulieu
J.d. & H
uGHes
P.w. (1975) - Environmental geology of western Coos and Douglas Counties, Oregon. Oregon
Department of Geology and Mineral Industries Bulletin 87.
b
enda
L. (1990) - The influence of debris flows on channels and valley floors in the Oregon Coast Range, U.S.A. Earth Surface
Processes and Landforms, 15: 457-466.
b
isHoP
A.W. (1959) - The principle of effective stress. Teknisk Ukeblad, 39: 859-863.
b
Rown
G.w. & k
RyGieR
J.T. (1971) - Clear-cut logging and sediment production in the Oregon Coast Range. Water Resources
Research, 7(5): 1189-1198.
C
oe
J.a., m
iCHael
J.a. & b
uRGos
m.m. (2010) - Map showing debris flows caused by precipitation in 1996 in parts of the
Reedsport and Deer Head Point quadrangles, southern Coast Range, Oregon. U.S. Geological Survey Open File Report.
C
Rosta
G.b. & f
Rattini
R. (2003) - Distributed modelling of shallow landslides triggered by intense rainfall. Natural Hazards
and Earth System Sciences, 3: 81-93.
d
aly
C., n
eilson
R.P. & P
HilliPs
d.l. (1994) - A statistical-topographic model for mapping climatological precipitation over
mountainous terrain. Journal of Applied Meteorology, 33: 140-158.
d
e
R
ose
R.C., t
RustRum
n.a., & b
lasCHke
P.m. (1991) - Geomorphic change implied by regolith-slope relationships on
steepland hillslopes, Taranaki, New Zealand. Catena 18 : 489-514.
d
e
R
ose
R.C. (1996) - Relationships between slope morphology, regolith depth, and the incidence of shallow landslides in
eastern Taranaki hill country. Zeitschrift für Geomorphologie Supplementband 105 : 49-60.
d
ietRiCH
w.e. & d
unne
t. (1978) - Sediment budget for a small catchment in mountainous terrain. Zeitschrift für Geomorphologie
Supplementband, 29: 191-206.
e
bel
b.a., l
oaGue
k., & b
oRJa
R.i. (2010) - The impacts of hysteresis on variably saturated hydrologic response and slope
failure. Environmental Earth Sciences, doi: 10.1007/s12665-101-0448-z.
f
Reeze
R.a. & C
HeRRy
J.a. (1979) - Groundwater. Prentice-Hall Inc., Englewood Cliffs, New Jersey.
background image
ASSESSING SUSCEPTIBILITY AND TIMING OF SHALLOW LANDSLIDE AND DEBRIS FLOW INITIATION
IN THE OREGON COAST RANGE, USA
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
833
G
aRdneR
W.R. (1958) - Some steady-state solutions of the unsaturated moisture flow equation with application to evaporation
from a water table. Soil Science, 85: 228-232.
G
odt
J.w., b
aum
R.l., s
avaGe
w.z., s
alCiaRini
d., s
CHulz
w.H. & H
aRP
,e.l. (2008a) - Transient deterministic shallow landslide
modeling: Requirements for susceptibility and hazard assessment in a GIS framework. Engineering Geology, 102: 214-226.
G
odt
J.w., s
CHulz
w.H., b
aum
R.l. & s
avaGe
w.z. (2008b) -Modeling rainfall conditions for shallow landsliding in Seattle,
washington. In: b
aum
R.l, G
odt
J.w. & H
iGHland
l.M. (eds.) Landslides and Engineering Geology of the Seattle,
washington, area. Geological Society of America Reviews in Engineering Geology, 20: 137-152.
I
veRson
R.m. (2000) - Landslide triggering by rain infiltration. Water Resources Research, 36(7) : 1897-1910.
k
eim
R.f. & s
kauGset
a.e. (2003) - Modelling effects of forest canopies on slope stability. Hydrological Processes, 17: 1457-1467.
k
eim
R.f., s
kauGset
a.e., l
ink
t.e. & i
Roume
a. (2004) - A stochastic model of throughfall for extreme events. Hydrology and
Earth System Sciences, 8(1): 23-34.
l
eiJ
P.J., s
CHaaP
m.G. & a
Rya
l.m. (2002) - Indirect methods. In: d
ane
, J.H. & t
oPP
, G.C. (eds.) Methods of Soil Analysis - Part
4 Physical Methods: 1009-1045, Soil Science Society of America, Madison, WI.
l
ink
t.e., u
nswoRtH
m. & m
aRks
d. (2004) - The dynamics of rainfall interception by a seasonal temperate rainforest.
Agricultural and Forest Meteorology, 124: 171-191.
l
u
n. & l
ikos
w.J. (2004) - Unsaturated soil mechanics. Wiley, New York.
m
ontGomeRy
d.R., s
CHmidt
k.m., d
ietRiCH
w.e. & m
C
k
ean
J. (2009) - Instrumental record of debris flow initiation during
natural rainfall: Implications for modeling slope stability, Journal of Geophysical Research, 114, F01031.
m
ontGomeRy
d.R., & d
ietRiCH
w.e. (1994) - A physically-based model for the topographic control on shallow landsliding.
Water Resources Research, 30: 1153-1171.
m
ooRe
R.d. & w
ondzell
s.M. (2005) - Physical hydrology and the effects of forest harvesting in the Pacific Northwest: A
review. Journal of the American Water Resources Association, 41(4): 763-784.
NRCS (2010) - Soil Survey Geographic (SSURGO) database for Douglas County Area, Oregon. U.S. Department of Agriculture, Natural
Resources Conservation Service, on-line at http://soildatamart.nrcs.usda.gov/Metadata.aspx?Survey=OR649&UseState=OR
(accessed 30 Aug. 2010).
ODF (1999) - Storm Impacts and Landslides of 1996: Final Report. Oregon Department of Forestry, Forest Practices Technical Report 4.
n
iem
a.R. & n
iem
w.a. (1990) - Geology and oil, gas, and coal resources, southern Tyee Basin, southern Coast Range, Oregon.
Department of Geology and Mineral Industries, Open-File Report O-89-3.
P
ieRson
T.C. (1977) - Factors controlling debris-flow initiation on forested hillslopes in the Oregon Coast Ranges. Ph.D. thesis,
University of Washington, Seattle, Washington.
R
eid
m.e., C
HRistian
s.b. & b
Rie
d.l. (2000) - Gravitational stability of three dimensional stratovolanco edifices. Journal of
Geophysical Research, 205: 6043-6056.
R
eneau
s.l. & d
ietRiCH
w.e. (1987) - The importance of hollows in debris flow studies: Examples from Marin County, California.
In: C
osta
J.e. & w
ieCzoRek
G.F. (eds.) - Debris flows/avalanches: Process, recognition, and mitigation. Geological Society
of America Reviews in Engineering Geology, 7: 165-658.
R
oeRinG
, J.J., s
CHmidt
, k.m., s
toCk
., J.d., d
ietRiCH
, w.e., & m
ontGomeRy
, d.R. (2003) - Shallow landsliding, root reinforcement,
and the spatial distribution of trees in the Oregon Coast Range, Canadian Geotechnical Journal, 40: 237-253.
s
alCiaRini
d., G
odt
J.w., s
avaGe
w.z., C
onveRsini
R., b
aum
R.l. & m
iCHael
J.a. (2006) - Modeling regional initiation of
rainfall-induced shallow landslides in the eastern Umbria Region of central Italy. Landslides, 3: 181-194.
s
avaGe
w.z. & b
aum
R.l. (2005) - Instability of steep slopes. In: J
akob
m. & H
unGR
O. (eds.) Debris-Flow Hazards and Related
Phenomena: 53-77, Springer, Berlin.
s
CHmidt
k.m., R
oeRinG
J.J., s
toCk
J.d., d
ietRiCH
w.e., m
ontGomeRy
d.R., & s
CHaub
t. (2001) - The variability of root
cohesion as an influence on shallow landslide susceptibility in the Oregon Coast Range. Canadian Geotechnical
Journal, 38: 995-1024.
s
CHulz
W.H. (2007) - Landslide susceptibility revealed by LiDAR imagery and historical records, Seattle, washington.
Engineering Geology, 89: 67-87.
s
latton
k.C., C
aRteR
w.e., s
HRestHa
R.l. & d
ietRiCH
w. (2007) - Airborne Laser Swath Mappping: Achieving the resolution
and accuracy required for geosurficial research. Geophysical Research Letters, 34: L23S10.
background image
R.L. BAUM, J.w. GODT & J.A. COE
834
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
s
Rivastava
R. & y
eH
t.-C.J. (1991) - Analytical solutions for one-dimensional, transient infiltration toward the water table in
homogeneous and layered soils. Water Resources Research, 27: 753-762.
t
ayloR
D.W. (1948) - Fundamentals of soil mechanics. Wiley, New York.
t
aRboton
D.G. (1997) - A new method for the determination of flow directions and contributing areas in grid digital elevation
models. Water Resources Research, 33(2): 309-319, doi: 10.1029/96WR03137.
w
iley
T.J. (2000) - Relationship between rainfall and debris flows in western Oregon. Oregon Geology 62(2): 27-43.
w
ilson
C.J. & d
ietRiCH
w.e. (1987) - The contribution of bedrock groundwater flow to storm runoff and high pore pressure
development in hollows. Erosion and Sedimentation in the Pacific Rim, International Association of Hydrological Sciences
Publication, 165: 49-59.
Statistics