Document Actions

IJEGE-11_BS-Kaitna-et-alii

background image
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
351
DOI: 10.4408/IJEGE.2011-03.B-040
ON THE DEVELOPMENT OF
AN UNSATURATED FRONT OF DEBRIS FLOWS
R. KAITNA
(*)
, l. HSU
(**)
, d. RICKENMANN
(***)
& w.e. DIETRICH
(****)
(*)
University of Natural Resources and Life Sciences, Austria (Boku)
(**)
University of California at Santa Cruz, USA
(***)
Swiss Federal Institute for Forest, Snow and Landscape Research (WLS)
(****)
University of California at Berkeley, USA
INTRODUCTION
Debris flows are commonly described as surges
of rapidly flowing mixtures of unsorted sediment and
water, which are characterized by a granular front fol-
lowed by a more dilute body (e.g. s
tiny
, 1910; s
uwa
,
1989; C
osta
, 1984). P
ieRson
(1986), for example, pro-
vides a detailed description of the geometry of debris
flow events at Mount St. Helens, Washington, USA.
The flow fronts were steep and ‘typically composed
of the coarsest particles available for transport’. The
coarse particles were transported on the surface to the
front and tumbled down the steep leading edge where
they partly accumulated (leading to an increasing
boulder front length with flow duration) or were partly
shouldered aside by the slurry pushing from behind
(contributing to the formation of lateral levees). P
ieR
-
son
(1986) as well as other authors (e.g. s
uwa
1988)
report that the pore spaces of the front are not filled
with the matrix slurry, resulting in an unsaturated front
with assumed high internal friction. The formation of
an unsaturated front lacking significant pore pressures
was repeatedly measured in large scale debris flow
experiments (i
veRson
et alii, 2010), in natural flows
(m
C
a
Rdell
et alii, 2007, m
C
C
oy
et alii, 2010) as well
as in small-scale flume experiments (s
Cotton
& d
eGa
-
nutti
, 1997). The damming action of the coarse front
is assumed to cause peak discharge and flow depth to
increase dramatically and friction concentrated at flow
margins is considered to impede debris flow move-
ment at lower gradients and may therefore be highly
ABSTRACT
A visibly granular debris flow front, where large
boulders accumulate, is often observed in nature. Al-
though there is abundant evidence of particle sorting
and solid-fluid segregation processes, little is known
about the specific segregation mechanisms and what
factors control the relatively dry coarse snout. To
investigate the conditions associated with the devel-
opment of an unsaturated front, experiments have
been conducted with grain-fluid mixtures of differ-
ent compositions. To create long-lived, accessible,
stationary flows we used two vertically rotating drum
setups with a diameter of 2.4 m and 4 m respectively.
The presence of an unsaturated front was detected by
pore pressure transducers installed at the base of the
flows. Additionally normal stress and flow depth are
measured. We carried out a series of runs with varying
fluid viscosity, sediment concentration, channel bed
roughness, and mean velocity. The unsaturated fronts
developed in faster flows with higher sediment con-
centration. The dry fronts formed even in well-sorted
coarse particles (hence differential particle inertia is
not necessary). High fluid viscosity may favor dry
front formation due to the tendency for the fluid to
stick to the boundary but also reduce the front forma-
tion by retarding segregation processes at the front.
K
ey
words
: granular front, pore fluid pressure, rotating
drum
background image
r. K
AitNA
, l. h
Su
, D. r
icKeNmANN
& w.e. D
ietrich

352
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
sieving represents a possible mechanism to cause
inverse grading in granular flows, but may be less
important in flows where the size ratio of particles is
large (t
Homas
, 2000). The magnitude of this effect in
grain-fluid mixtures needs to be assessed.
Frontal focusing is a result of a force balance be-
tween the downslope component of the gravity force
of a particle, the frictional force on the channel bot-
tom and the drag force of the surrounding flow (s
uwa
,
1988). In steep reaches of a channel large boulders
reach a higher terminal velocity compared to smaller
ones due to a higher downslope component of gravity
force. Consequently large boulders tend to migrate to
the front. The proposal of s
uwa
(1988
)
is based on a
simplified analysis of a particles dragged by a fluid
and is supported by small-scale laboratory experi-
ments of a bore of water entraining glass particles of
different sizes. Effects due to high overall sediment
concentration and the presence of other particles are
not taken into account. A concentration of boulders
at the front has also been observed at lower gradients
(m
C
a
Rdell
et alii, 2007).
In summary, current explanations for the forma-
tion of a typical debris flow front concentrate on dif-
ferent mechanisms for the frontal focusing of large
boulders rather than solid-fluid segregation processes.
In this contribution we aim to investigate the condi-
tions that favor the formation of an unsaturated front.
Is segregation by large particles (relative to the me-
dian size) necessary to develop an unsaturated front
of a flowing grain-fluid mixture? Is the formation of a
typical debris flow front a combination of the separate
processes of particle sorting (accumulation of larger
particles at the front) and phase segregation (particles
preceding the fluid)? What is the role of the viscos-
ity of the intergranular fluid (water and fines)? Is it
possible to create an unsaturated front in grain-fluid
mixtures composed of uniform sediment? We explore
these questions by conducting laboratory experiments
in a rotating drum apparatus, where it is possible to
establish stationary flows of grain-fluid mixtures of
varying composition at different mean flow velocities.
METHODS
We conducted a first series of experiments in a
rotating drum with a diameter of 2.4 m and a channel
width of 0.45 m (‘small drum’, Fig. 1). The curved
channel bottom was roughened with a synthetic mesh
relevant for debris flow runout and deposition (m
aJoR
& i
veRson
, 1999). Considering the large porosity of a
front composed of large boulders it is interesting that
the liquid matrix doesn’t drain at the front in the course
of an event, which typically has duration of several
minutes (a debris flow front has often been described
as a “moving dam” with more liquid material “push-
ing from behind”). For debris flows containing a sig-
nificant fraction of clay, silt and fine sand, this can be
explained by a relatively high concentration of fines
in the body of the flow resulting from to particle seg-
regation. Here the low hydraulic conductivity of the
liquid matrix retards drainage over long time scales.
For debris flows lacking a significant amount of fines
post-depositional drainage within seconds to minutes
was observed (m
aJoR
& i
veRson
, 1999).
Engineering simulation tools model debris flows
as homogeneous mixtures associated with an intrin-
sic rheologic flow behavior (e.g. o’b
Rien
et alii,
1993,;m
C
d
ouGall
& H
unGR
, 2004; n
aef
et alii,
2006; b
eGueRia
et alii, 2009). Sorting effects like the
accumulation of large boulders at the front or the seg-
regation of the fluid and solid phase are not taken into
account. Newer concepts treat debris flows as two-
phase mixtures; that is, the flow resistance of the solid
and the fluid components are taken explicitly into ac-
count and are coupled via the pore pressure of the fluid
(i
veRson
& d
enlinGeR
, 2001) or buoyancy and drag
(b
eRzi
& J
enkins
, 2008). This approach allows simu-
lating varying fluid pressure along a debris flow surge,
but does not account for particle sorting.
Different mechanisms have been suggested to ex-
plain the formation of a typical debris flow front:
Dispersive pressure, that causes large particles
to move vertically upward (”inverse grading”, e.g.
b
aGnold
, 1968; t
akaHasHi
, 1991). Due to higher ve-
locities in the upper layers of the flow profile these
particles are regularly transported to the front where
they accumulate. l
eGRos
(2002) presents a theoretical
analysis that dispersive pressure is less important for
inverse grading in grain flows and might be therefore
not the dominant mechanism responsible for boulder
accumulation at the front of debris flows.
In an agitated granular mass large particles tents
to migrate to the upper regions of the flow profile by
the mechanism of “kinetic sieving”, and are subse-
quently transported to the front (e.g. s
avaGe
& l
un
,1988; P
ouiQuen
et alii, 1997). The effect of kinetic
background image
ON THE DEVELOPMENT OF AN UNSATURATED FRONT OF DEBRIS FLOWS
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
353
eter of fine components of the mixture was determined
to be 0.008 mm (Hsu, 2010).
In both setups the fluid pressure was monitored
by two pressure transducer installed at the centerline
of the rotating channel bed. The transducer measured
the pressure in an oil filled reservoir, which was con-
nected to the channel bed by a flexible membrane. The
membrane was protected from being affected by parti-
cle impacts by a steel mesh (2 mm spacing). Addition-
ally both drums were instrumented with load plates
to measure bed normal total stress in the centerline of
the flow. The measurement frequency was set to 400
Hz for the runs in the small drum and 1000 Hz for
the runs in the big drum. The experiments were con-
tinuously monitored with digital video cameras from
the top and with digital photo cameras through the
transparent side walls. To measure the flow geometry,
ultra-sonic sensors were used for the experiments in
the small drum and a 2-D laser profiler was employed
for the runs in the big drum. Details about the experi-
mental setups and the instrumentation can be found
in k
aitna
& R
iCkenmann
(2007) and in H
su
(2010).
In the following, all information about the presence
and length of an unsaturated front are based on the
comparison between normal stress and pore fluid pres-
sure data. Since we investigate only uniform particle
mixtures with relatively large pore spaces, we assume
tension saturation (suction pressure within the pores)
to be negligible. In most cases it was possible to quali-
tatively confirm the measurements with observations
through the transparent side walls.
RESULTS
PVC-FLUID MIXTURES
The tests with this artificial material showed that
it is possible to establish stationary flows for all tested
mixtures in the small drum. Though the sidewalls
were rather smooth (stainless steel on one side and
acrylic glass on the other) inducing less friction than
the bottom of the drum, the effect of the sidewall was
not negligible and a symmetric, three-dimensional
flow pattern developed. At high velocities this flow
pattern sometimes turned asymmetric. The direction
of asymmetry varied. Strong asymmetric flows were
excluded from further analysis. Some fraction of the
fluid was stripped from the bulk mixture due to adhe-
sive forces on the boundary. It is noted that this effect
was enhanced with increasing rotation velocity and by
of 10x10 mm spacing and a height of around 5 mm. In
an initial step we tested artificial grain-fluid mixtures
of varying volumetric concentration (Cv = volume
solids / total volume
= 0.26 to 0.62), fluid viscosities,
bulk volume (V = 0.028 to 0.046 m³), and mean veloc-
ity (v
m
= 0.1 to 2.8 m/s). The particles used were cylin-
drical PVC grains of 10 mm in diameter and 10 mm
height, with a specific density ρ
s
of 1.42. The fluid was
either water (µ ~ 0.001 Pa.s) or a transparent synthetic
polymer as used in chemical industry (see k
aitna
&
R
iCkenmann
, 2007 for details). The polymer exhibited
a shear thinning flow behavior which may be roughly
characterized by a Bingham model over a limited
range of shear rates. The viscosities (due to polymer
in the water) were varied over one order of magnitude
(µ = 0.02 to 0.2 Pa.s) and the yield stress was kept
relatively small (τ
y
= 0.5 to 20 Pa). Bulk volumes in
our experiments ranged between 0.028 and 0.046 m³,
corresponding to a total mass of 19 to 55 kg.
To compare the results of these artificial mixtures
with more realistic material we conducted another set
of experiments in a larger rotating drum using gravel
and water and gravel and mud, respectively. The sec-
ond drum had a diameter of 4 m and a channel width
of 0.8 m (‘big drum’, Fig. 1). The material tested com-
prised uniform gravel of 4, 10, and 13 mm (ρ
s
= 2.65).
For all experiments the total mass of solids was 455
kg. For the gravel-mud experiments we added a clay-
silt-fine sand mixture. The densities of the muddy
fluid ranged between 1136 and 1205 kg/m³, represent-
ing volumetric sediment concentrations between 0.09
and 0.14 in the fluid. Fluid viscosities measured in a
co-axial cylinder rheometer (Bohlin Visco88) ranged
from 0.008 to 0.09 Pa.s over the estimated shear rate
range of relevance (0-150 s
-1
). We estimated the maxi-
mum shear rate based on surface velocity and mean
flow depth during our experiments. The mean diam-
Fig. 1 - Photographs showing the drum setup of 2.5 m
diameter (left) and 4 m diameter (right)
background image
r. K
AitNA
, l. h
Su
, D. r
icKeNmANN
& w.e. D
ietrich

354
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
mixtures with Cv – values of 0.48±0.05 and volumes
between 0.028 and 0.048 m³ (Tab. 1). Though fluid
viscosity was varied over two orders of magnitude (µ
= 0.001–0.2 Pa.s) a dry front formed depending on
mean flow velocity v
m
to some extent in all mixtures.
.
GRAVEL-wATER MIXTURES
The experiments in the big drum reveal that
solid-fluid segregation processes are also an impor-
tant feature of natural grain-fluid mixtures. Figure 3
depicts an example of the longitudinal stress profiles
of a mixture with a well sorted gravel with median
size of 7 mm and water (C
v
= 0.63) at (1) 0.41 m/s,
(2) 0.83 m/s, and (3) 1.46 m/s. The load plate sen-
sor and the pore fluid pressure sensor are located in
different quadrants in the drum, but given the steady
flow conditions we shifted the data by 86.3° to allow
direct comparison. Since we are not interested in the
fluctuating components of the signal, the load cell and
fluid pressure data were low-pass filtered at 5 Hz and
averaged over a window size of one degree for three
rotations. Therefore the error bars do not represent the
fluctuation around a mean value but the standard de-
viation of the mean values for three rotations.
It can be seen that flow velocity strongly controls
the formation and length of an unsaturated front in the
experiments as in the small drum. Figure 3a shows
that at low velocities water drains out at the front and
a bore of fluid precedes the not fully saturated mass of
gravel (note dark color of the unsaturated gravel). The
fluid pressure in the bore exceeds total normal stress
measured by the load cell at the front, which is not
possible for a fluid flow (normal stress should equal
fluid pressure). This disagreement (also observed in
the faster flow shown in Fig. 3b) is primarily due to
the fact that the mesh attached to roughen the bed in-
creased the specific channel surface area significantly.
Using the bulk volume calculated form geometric data
and the known volume of the solids and fluid dumped
into the drum, we were able to calculate the Cv value
for each run separately. Cv-values varied by ± 0.05 for
a given mixture.
We observed an unsaturated front for mixtures
with Cv values larger than 0.35. The onset of the for-
mation as well as the length of a dry front was directly
related to mean flow velocity. For mixtures with Cv
values smaller than 0.35 the front region was fluidized
at all speeds.
Figure 2 gives an example of normal stress and
pore fluid pressure data for a PVC – fluid mixture (µ =
0.09 Pa.s, Cv = 0.58) at flow velocities of 0.1 m/s and
1.1 m/s, respectively. In this study we concentrate only
on the averaged values, thus the data was smoothed
and represent values averaged over ten rotations (i.e.
the sensors which are installed at the channel bed are
passing underneath the surge ten times). It can be seen
that the mixture is well saturated throughout the flow
length at a very low velocity, whereas the front region
is clearly unsaturated at a velocity of 1.1 m/s. The
length of the unsaturated front increased gradually
with flow velocity. This pattern was observed for most
of our tested mixtures.
The sample volume had some influence on the de-
velopment of a dry front because fluid loss due to ad-
hesive forces at the boundary is more pronounced for
small volumes, but is less relevant for larger volumes.
For this reason we kept the sample volume constant
for most experiments.
We tried to assess the influence of fluid viscos-
ity on the formation of a dry front by comparing four
Fig. 2. - Longitudinal normal stress and pore fluid pres-
sure profiles for a PVC – fluid mixture (µ =
0.09 Pa.s, Cv = 0.58) at Vm = 0.1 m/s (left)
and 1.1 m/s (right)
Tab. 1 - Presence of an unsaturated front for mixtures
of varying fluid viscosity at different mean flow
velocities v
m
background image
ON THE DEVELOPMENT OF AN UNSATURATED FRONT OF DEBRIS FLOWS
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
355
reduced in this region and explains some inconsisten-
cies in the different datasets at the front and the tail
of our experimental flows. Figure 3b shows total nor-
mal stress and pore fluid pressure rising together, but
soon total stress significantly exceeds fluid pressure.
The mean flow velocity v
m
in this situation was 0.83
m/s, which is around twice the velocity of (a). We con-
clude that the mixture is relatively well mixed over the
whole length of the flow. As can be seen in the photos,
only the uppermost region close to the flow surface
at the deepest part of the flow is not fully saturated.
Further increase of flow velocity induces a progres-
sive development of an unsaturated front. At v
m
= 1.46
m/s the first ~ 20 cm of the front were clearly not satu-
rated but saturation increased linearly within the flow
following the unsaturated front (Fig. 3c). Visually the
deepest part of the flow became completely saturated.
Experiments with gravel-water mixtures of mean
diameter of 4 mm and 13 mm and varying water con-
tent resemble measurements with the 7 mm gravel and
support the results from the small drum: for a given
mixture the onset and the length of a dry front is close-
ly related to flow velocity.
To investigate the effect of fines on the formation
of a dry front we conducted some runs with gravel-
mud mixtures of varying mud sediment concentration.
The results from the experimental runs showed a simi-
lar pattern as the gravel-water runs, however, the ef-
fect of increased fluid viscosity was more pronounced
than found in the experiments in the small drum: an
unsaturated front formed in all mixtures above a criti-
cal velocity, but developed over only a short distance
(limited length) and did not increase proportional to
an increase of mean flow velocity as observed with the
PVC-fluid mixture.
DISCUSSION AND CONCLUDING RE-
MARKS
Our experiments show that segregation processes
are an important feature of grain-fluid mixtures. Since
we used only uniform grain-fluid mixtures we con-
clude that neither the mechanism of kinetic sieving
nor the effect of dispersive pressure is necessary for
the development of an unsaturated front. However,
we cannot exclude these mechanisms of being im-
portant for frontal focusing of large boulders. Three
different regimes can be identified for our uniform
particle mixtures of a given water content and fluid
the fact that the load cell and the fluid pressure sensor
are not located at the position in the drum and data
had to be shifted for comparison. It is noted that in the
vicinity of the normal stress load cell some roughness
elements had to be taken out to guarantee an undis-
turbed measurement.
As a consequence the flow surface was slightly
Fig. 3 - Stress distribution and photographs
through the transparent side walls of a
gravel-water mixture (Cv = 0.62, d
m
=
7 mm) at rotation velocities 0.41 m/s (a),
0.83 m/s (b), and 1.46 m/s (c). Sensor
position 0° represents 6 o’clock position
in the drum. Photos are flipped for com-
parison. Black bar delineates the “dry
front”. Lines of pore fluid pressure have
been smoothed
background image
r. K
AitNA
, l. h
Su
, D. r
icKeNmANN
& w.e. D
ietrich

356
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
viscosity, which are separated only by different mean
velocities of the flow (Fig. 4): (1) at low velocities
and low viscosities the fluid precedes the solid par-
ticles which are concentrated in the body and tail of
the flow; (2) at ‘medium’ velocity the bulk mixtures
are visually homogeneously mixed; and (3) at ‘high’
velocities an unsaturated front develops. All these re-
gimes formed rapidly (duration of seconds) after the
start of the experiment and stayed relatively constant
for the time of observation. Regime (2) and (3) may
resemble dynamics in natural debris flows. Regime
(1) on a straight slope would cause the flow to drain
and stop motion.
An unsaturated front has been reported in large-
scale debris flow experiments (e.g. m
aJoR
& i
veRson
1999; i
veRson
et alii, 2010) and field measurements
(m
C
a
Rdell
et alii, 2007, m
C
C
oy
et alii, 2010). One
of the main differences between our laboratory ex-
periments and natural flows is that we impose a mean
flow velocity through drum rotation. The flows have
to adjust flow height and surface slope accordingly.
For this reason it is possible to keep a granular suspen-
sion in constant motion, though the fluid drains out
at the front (regime 1). In a stationary flow driving
forces have to balance resisting forces. Therefore the
mean surface slope increases as the flow resistance of
the experimental mixtures increases. Comparison of
the mean surface slope based on flow depth measure-
ments showed that the partially drained condition of
the flow body in regime (1) leads to increased flow
resistance compared to regime (2) and (3). The loss
of water from the bulk mixture in regime (1
)
results
in an increase of effective normal stress and subse-
quently to an increase of Coulomb grain shear stress.
Regime (1) may represent the final phase of a debris
flow event, when the mass reduces its speed due to
reduction channel slope and/or spreading on the fan.
Observation of grain-fluid segregation processes
raises the question of grain velocity vs. fluid veloc-
ity. Most two-phase debris flow models assume that
grains travel with the same velocity as the fluid (e.g.
i
veRson
, 1997). Regime (1) and (3) clearly show that
the fluid and particles must follow different trajecto-
ries.
Fluid viscosity can affect unsaturated front devel-
opment in two opposing ways: boundary resistance
due to the no slip condition and drag resistance to
particle motion. The fluid experiences adhesive forces
on the boundary. This non-slip condition applies for a
fluid, but is not necessarily expected for an intensively
sheared grain flow, since particle slip occurs in granu-
lar experiments (e.g. l
ouGe
& k
east
, 2001; b
aRtelt
et alii, 2005) including in our big drum experiments.
High fluid viscosity could enhance the bottom flow
to ‘stick’ to the base even as the inertia of the coarse
particles carries them past the flow to the front.
High fluid viscosity could, however, also reduce
the tendency for coarse particles to segregate from
the flow. Consider a situation as in regime (2), where
the grain-fluid mixture is relatively well mixed. The
density ratio between the solid (ρ
PVC
= 1.42, ρ
gravel
=
2.65) and the fluid components (ρ
water
= 1.0, ρ
mud
~ 1.2)
is larger than unity, resulting in higher inertia of the
solids than the fluid. Particles and fluid of the upper
layers where velocity is larger than the absolute mean
velocity (which is zero in case of the drum) travel to-
wards the front at a relative high speed. At the front the
bulk mass has to turn towards the bed, and particles –
owing to their high inertia – may have the tendency
to escape from the fluid. High fluid viscosities would
dampen this effect because of higher drag resistance.
Similar to the analysis of s
uwa
(1988), larger particles
Fig. 4 - Three regimes of grain-fluid flow observed in the rotating drums
background image
ON THE DEVELOPMENT OF AN UNSATURATED FRONT OF DEBRIS FLOWS
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
357
big drum which has larger roughness elements, but a
low specific surface area.
Taken together, our drum experiments show that
both flow velocity and fluid viscosity influence the
degree of development of the unsaturated front in de-
bris flows. Even in well sorted material, where there is
no size dependent differential behavior of the coarse
particles, a dry front develops in sufficiently viscous
and rapidly moving flows. The fluid and particle paths
differ. Losses to the bed and walls in a debris flow
can contribute to drying of the flow, but the unsatu-
rated front requires neither losses nor segregation of a
coarse fraction to occur.
ACKNOWLEDGEMENTS
The authors would like to thank Alexander
Krawtschuk, Fritz Zott and Stuart A. Foster for their
contribution in realizing the experiments. This re-
search was financially supported by the Austrian
Science Fond (J2837-N10) and by the STC program
of the National Science Foundation via the National
Center for Earth-surface Dynamics under the agree-
ment Number EAR-0120914.
should be preferentially expelled from the grain-fluid
assembly. The relative importance of these processes
should depend on mean velocity of the flow.
The experiments in the small drum show that
the development of a dry front depends both on the
velocity and the viscosity of the intergranular fluid.
This was also observed in small-scale flume experi-
ments of s
Cotton
& d
eGanutti
(1997) and the exper-
iments in the big drum support this observation. The
runs with natural material in the big drum showed
in addition, that the presence of fines dampens the
formation of a dry front. An additional influence on
dry front formation is the net loss of material during
the flow, which in the case of drums is influenced
by the bed roughness and specific surface area cre-
ated. The effect of stripping of fines from the leading
edge and sequestering them into the flume bed has
been discussed by P
aRson
et al. 2001. They conclude
that this effect may contribute to the concentration
of coarse sediment at the front, but is not sufficient
to explain the rapid formation of a granular snout in
their experiments. This effect may be more signifi-
cant in our experiments in the small drum than in the
REFERENCES
b
aGnold
R.a. (1968) – Deposition in the process of hydraulic transport. Sedimentology, 10: 45-56.
b
aRtelt
P., b
useR
o. & k
eRn
m. (2005) - Dissipated work, stability and the internal flow structure of granular snow avalanches.
Jounal of Glaciology, 51(172): 125-138.
b
eGueRía
s., v
an
a
sCH
t.w.J., m
alet
J.P. & G
RöndaHl
s. (2009) - A GIS based numerical model for simulating the kinematics
of mud and debris flows over complex terrain. Nat Hazards Earth Syst Sci, 9: 1897-1909. Disponible en el siguiente registro:
http://digital.csic.es/handle/10261/20650.
b
eRzi
d., J
enkins
J.t. (2008) - A theoretical analysis of free-surface flows of saturated granular-liquid mixtures. Journal of Fluid
Mechanics, 608: 393-410 (doi:10.1017/S0022112008002401).
C
osta
J.e. (1984) - Physical Geomorphology of Debris Flows. In C
osta
J.e.; f
leisCHeR
P.J. (
eds
.): Developments and Applica-
tions of Geomorphology. Berlin: Springer: 268-317.
H
su
l. (2010) - Bedrock Erosion by Granular Flow. Dissertation submitted at the University of California at Berkeley, USA.
i
veRson
R.m. (1997) - The physics of debris flows. Reviews of Geophysics, 35(3): 245-296.
i
veRson
R.m.; d
enlinGeR
R.P. (2001) - Flow of variably fluidized granular masses across three-dimensional terrain: 1. Coulomb
mixture theory. Journal of Geophysical Research, 106(B1): 537-552.
i
veRson
R.m., l
oGan
m., l
a
H
usen
R.G. & b
eRti
m. (2010) - The perfect debris flow? Aggregated results from 28 large-scale
experiments. Journal of Geophysical Research, 115 (doi:10.1029/2009JF001514).
k
aitna
R., R
iCkenmann
d. (2007) - A new experimental facility for laboratory debris flow investigation. Journal of Hydraulic
Research, 45(6): 797-810.
l
eGRos
M.Y. (2002) – Can dispersive pressure cause inverse grading in grain flows? Journal of Sedimentary Research, 72(1):
166-170.
l
ouGe
m.y. & k
east
s.C. (2001) - On dense granular flows down flat frictional inclines. Physics of Fluids, 13(5): 1213-1232.
m
aJoR
J.J. & i
veRson
R.m. (1999) - Debris-flow deposition: Effects of pore-fluid pressure and friction concentrated at flow
background image
r. K
AitNA
, l. h
Su
, D. r
icKeNmANN
& w.e. D
ietrich

358
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
margins. Geological Society of America Bulletin, 111(10): 1424-1434.
m
C
a
Rdell
b.w., b
aRtelt
P. & k
owalski
J. (2007) - Field observations of basal forces and fluid pressure in a debris flow. Geo-
physical Research Letters, 34, L07406 (doi: 10.1029/2006GL029183)
m
C
C
oy
s.w., k
ean
J.w., C
oe
J.a., s
taley
d.m., w
asklewiCz
t.a. & t
uCkeR
G.e. (2010) - Evolution of a natural debris flow:
In situ measurements of flow dynamics, video imagery, and terrestrial laser scanning. Geology, 38: 735-738 (doi: 10.1130/
G30928.1)
m
C
d
ouGall
s. & H
unGR
o. (2004) - A model for the analysis of rapid landslide motion across threedimensional terrain. Can
Geotech J, 41: 1084-1097.
n
aef
d., R
iCkenmann
d., R
utsCHmann
P. & m
C
a
Rdell
b.w. (2006) - Comparison of flow resistance relations for debris flows
using a one-dimensional finite element simulation model. Natural Hazards and Earth System Sciences, 6: 155-165.
o’b
Rien
J.s., J
ulien
P.y. & f
ulleRton
w.t. (1993) - Two-dimensional water flood and mudflow simulation. J. Hydraulic Engi-
neering, 119(2): 244-261.
P
aRson
J.d., w
HiPPle
k.x. & s
imoni
a. (2001) - Experimental study of the grain-flow, fluid-mud transition in debris flows.
Journal of Geology, 109: 427-447.
P
ieRson
t.C. (1986) - Flow behavior of channelized debris flows, Mount St. Helens, washington. In a
bRaHams
a.d. (
ed
.): Hills-
lope Processes. Boston: Allen and Unwin: 269-296.
P
ouliQuen
o., d
elouR
J. & s
avaGe
s. (1997) - Fingering in granular flows. Nature, 386: 816–817.
s
avaGe
s.b. & l
un
C.k.k. (1988) - Particle size segregation in inclined chute flow of dry cohesionless granular solids. Journal
of Fluid Mechanics, 189: 311-335.
s
Cotton
P. & d
eGanutti
a.m. (1997) – Phreatic line and dynamic impact in laboratory debris flow experiments. In C
Hen
C.
(edt.): Debris-flow hazard mitigation: Mechanics, Prediction, and Assessment. ASCE, USA.
s
uwa
H. (1988) - Focusing mechanism of large boulders to a debris flow front. Trans. Japan. Geomorph. Union, 9: 151-178.
s
uwa
H. (1989) - Field observation of debris flow. Proceedings of the Japan-China (Taipei). Joint Seminar on Natural Hazard
Mitigation, Kyoto, Japan, July 16-20: 343-352.
s
tiny
J. (1910) - Die Muren [Debris flows]. Wagner. Innsbruck.
t
akaHasHi
t. (1991) - Debris Flow. IAHR Monograph Series, Balkema Publishers, The Netherlands.
t
Homas
n. (2000) - Reverse and intermediate segregation of large beads in dry granular media. Physical Review E(62): 961-974..
Statistics