Document Actions

IJEGE-11_BS-Ikeda-et-alii

background image
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
91
DOI: 10.4408/IJEGE.2011-03.B-011
DEFORMATION OF STREAM BED DEPOSIT AND RUNOFF PROCESS
AT DEBRIS FLOW INITIATION ZONE
a
kiHiko
IKEDA
(*)
, t
akaHisa
MIZUYAMA
(**)
, n
obuo
SUGIURA
(***)
& y
uJi
HASEGAWA
(***)
(*)
Sabo Technical Center
(**)
Graduate School of Agriculture, Kyoto University
(***)
Civil Engineering Research Laboratory
dune. The upper part of the dune had no surface water
but the lower part did. The dune flowed down gradu-
ally, finally reaching the downstream end with a con-
stant discharge. This dune started to move or deform
when the surface water reached and flowed over the
shoulder of the dune, forming a conspicuous peak. 3)
As the amount of supplied water discharge (q
in
) was
small and the moving layer flowed down, the runoff
discharge containing sediment at the downstream end
(q
out
) was constant and the ratio of q
out
/ q
in
was around
0.8. Over time, q
in
increased and the dune moved dy-
namically, forming a conspicuous peak. Values for q
out
were variable and the ratio of q
out
/ q
in
was around 2.0.
K
ey
words
: debris flow, initiation zone, deformation of stre-
am bed deposit, moving layer, dune, runoff process
INTRODUCTION
Hydraulic experiments to analyze the mechanism
of debris flow occurrence, flow properties and depo-
sition, experimental condition must be necessary to
make sure a debris flow as a sediment gravity flow.
Therefore, in many experiments the grains were
layered on the bed of a steep flume with a uniform
gradient and saturated it beforehand and the rap-
idly supplied large amount of water from upstream
end (k
aki
, 1955; y
ano
& d
aido
, 1959; t
akaHasHi
,
1977). In natural torrents, however, the riverbed gra-
dient usually becomes gentler from the upstream to-
ward the downstream, the stream bed deposit is not
ABSTRACT
Generally, debris flow occurs in response to heavy
rainfall. The way in which debris flow occurs may be
influenced by the conditions of the stream bed deposit,
such as its shape, gradient and deformation. In past
studies, a hydraulic experiment conducted to analyze
debris flow occurrence used a steep flume and uni-
form gradient, and then some water was supplied to
saturate the stream bed deposit before the debris flow
was occurred. But in a natural torrent, the riverbed
gradient becomes gentler downstream, the stream bed
deposit is not uniform thickness, and also the deposit
is sometimes unsaturated. We observed and analyzed
the condition and deformation of a stream bed deposit
and the runoff process of debris flow in the zone where
debris flow is initiated through hydraulic experiments.
The experiments were carried out on an experimental
flume whose bed slope was 30° at the upstream end
and 12° at the downstream end. We supplied a small
amount of water, nearly as much as the infiltration flow
at first, and rising as the surface water level rose. The
experimental results clarified the following: 1) In the
steep section with a gradient of 27°-30° and a small
amount of surface water, a moving layer was formed
that flowed downhill slowly. This moving layer started
to move when the layer was saturated by the infiltration
flow. When the moving layer flowed down in this sec-
tion, no surface water or infiltration flow was observed
at the downstream end. 2) As the gradient decreased to
21°-24°, the moving layer started to deposit and form a
background image
A. IkEDA, T. MIZUYAMA, N. SUGIURA & Y. HASEGAwA
92
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
EXPERIMENTAL SETUP AND PROCEDURE
As shown in Figure 1, the experimental flume is a
gradient changing flume (horizontal length is approxi-
mately 12 m and height is approximately 4.5 m) with
a width of 15 cm and a length of 1.8 m, with seven
connected flumes. The flume bed slope is 30°at the up-
stream end and 12°at the downstream end, continually
changing at intervals of 3° as 30°, 27°, 24°, 21°, 18°,
15°, and 12°. The depth of the flume is 30 cm. Smooth
glass was attached on the left lateral side, and a mesh
for measurement was also attached. Roughened bot-
tom surfaces were attached on the flume floor from
the 21° to the 30° gradients to prevent slide, which is
stuck by a wooden piece at 10 cm intervals, and a steel
mesh board of the same height as the thickness of the
deposit was installed at the end of the flume as a stop-
per. A water gauge was installed at the downstream
end and a tank and weight gauge at the downstream of
the flume to measure discharge and sediment volume.
Sand for the experiment is mixed with an average
grain size of 4.42 mm as shown in the grain size distri-
bution in Figure 2. The sand was laid on the bottom of
the flume to a thickness of 5 cm. Table 1 shows the ex-
perimental conditions, and Figure 3 shows the model
hydrograph used in the experiment. In the preliminary
experiment, the discharge which did not cause surface
water was found to be about 0.02 L/sec; thus 0.05 L/
sec was used as the minimum discharge assuming that
it resembled a normal stream flow in natural torrents.
Presuming that the discharges of stream flows in natu-
ral torrents gradually increase during precipitation,
uniform thickness and grain sizes, and also deposit
sometimes is unsaturated when there is no precipita-
tion even while stream flow is available. Furthermore,
the process of rainfall infiltration during precipitation
and the process of appearance of surface flow re-
mains unclear, the stream bed deposit sometime have
high permeability, infiltration flow become so much,
unsaturated condition seems to be present even in
during heavy rainfall (s
atofuka
& m
izuyama
, 2009).
In past, a few studies focused on consider a natural
torrent, experimental flume continually changing
slope gradient have been used (t
akaHasHi
, 1977;
i
keya
& m
izuyama
, 1982; m
izuyama
& H
u
, 1989),
and a small amount of water nearly as much as infil-
tration flow is supplied (Yano et alii, 1969), unfortu-
nately, and general studies involved saturated depos-
its and large amounts of water. While the occurrence
of debris flow is probably related with properties of
stream bed deposit (deposit gradient, thickness, grain
size, etc.) at the initiation zone of debris flow and its
time and spatial variations (deformations), few stud-
ies have focused on deposits and their runoff process.
This study thus attempts to clarify deformative
characteristics of torrential debris flow at initiation
zone in order to adopt a suitable measure to pre-
vent or to mitigate the disasters by discussing the
moving properties, deformation of stream bed de-
posit through hydraulic experiments, using a flume
in which the gradient gradually becomes gentler
(hereinafter referred to as “the gradient changing
flume”), presuming a natural torrent.
Fig. 1 - Experimental flume
background image
DEFORMATION OF STREAM BED DEPOSIT AND RUNOFF PROCESS AT DEBRIS FLOW INITIATION ZONE
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
93
which flowed down with a clear forefront was defined as
a “moving layer,” and the layer which formed a shoulder
after the moving layer had stopped and moved in a lay-
ered form as if extending downward as a “dune.”
1) In the steep section with a gradient of 27°-30°,
shortly after the small amount of water was di-
scharged, surface water appeared and eroded the
surface of the deposit, forming a moving layer, and
this flowed down above the deposit gradually. Infil-
tration flow occurred inside the deposit beneath the
surface water, and the deposit became saturated.
2) Although the moving layer appeared saturated, no
surface water was observed. Meanwhile, the depo-
sit at the downstream end of the moving layer was
unsaturated, and surface water was not observed.
3) The moving layer flowed down ahead of the infil-
tration flow. After the moving layer had stopped,
the infiltration flow reached the downstream end
of the moving layer, and the moving layer started
to move when the deposit became saturated.
4) When the gradient decreased to 21°-24° and the
dune began to form, the surface water infiltrated
into the dune from its upper end, and it flowed
down along with the infiltration flow in the deposit.
5) Although surface water was observed at the back sur-
face of the dune, and the dune appeared saturated, no
surface water was observed at the fore surface of the
dune. Also, the deposit at the lower part of the dune
was unsaturated, and no surface water was observed.
6) The dune began to move as if extending downward
when the surface water at the back surface of the
dune flowed over the shoulder of the dune, and the
infiltration flow reached the downstream end of the
dune and saturated the deposit beneath the dune.
7) Surface water appeared above the deposit at the
lower part of the dune, when the forefront of the
dune reached a gradient gentler than 18°.
RELATIONSHIP BETwEEN MOVING PROPER-
TIES AND RIVERBED GRADIENT
Two types of moving properties of debris flow were
observed: the moving layer of sediment which flowed
down above a stream bed deposit with a clear leading
edge; and the dune which formed a shoulder after the
moving layer had stopped and flowed down gradually
in a layered form with the stream bed deposit. There was
a tendency for the moving layer to form in steep gradi-
ent zones (about 30° to 27°) where the surface water
the discharge levels were set at 0.05 L/sec, 0.10 L/sec,
0.20 L/sec, and 0.30 L/sec (CASE 2 to 5). Water was
supplied until the sediment stopped moving in each
discharge level. In the preliminary experiment, the
sediment was observed to reach the end of the flume
at 0.30 L/sec; thus, the water was supplied at 0.30 L/
sec assuming it resembled debris flow in natural tor-
rents during a sudden onset of heavy rain (CASE 1).
RESULTS OF EXPERIMENTS
CHANGES IN THE DEFORMATION, SURFACE
wATER AND INFILTRATION FLOw
Figures 4 to 8 show stream bed deposit deformations
and their changes in each case using profile of bed surface.
As a result of detailed observations on the flow ranges
and changes in surface water and infiltration flow based
on the movement of deposit in each case, the following
phenomena are found as the deposits which moved from
upstream to downstream. Meanwhile, the sediment layer
Fig. 2 - Grain size distribution of the experimental sand
Tab. 1 - Experimental conditions
Fig. 3 - Model hydrograph
background image
A. IkEDA, T. MIZUYAMA, N. SUGIURA & Y. HASEGAwA
94
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
Fig. 4 - Profile of bed
surface (CASE-
1: 0.30L/sec)
Fig. 5 - Profile of bed surface
(CASE-2: 0.05L/sec)
Fig. 6 - Profile of bed
surface (CASE-
3: 0.10L/sec)
background image
DEFORMATION OF STREAM BED DEPOSIT AND RUNOFF PROCESS AT DEBRIS FLOW INITIATION ZONE
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
95
discharge was small and the deposit was less eroded.
Meanwhile, the dune tended to form and rise higher in
low-gradient zones (less than 24°) where the moving lay-
er stopped and the sediment supplied from the upstream
increased. Moving layers and dunes tended to stop at
zone of gradient shift in all cases. Thus, the first dune
formation often occurs at zone of gradient shift. The
thickness of the moving layer increases as it erodes
and flows down along sediment layers. On the con-
trary, the thickness of the dune shoulder increases
and the fore surface of the dune become steeper as
the sediment supplied from upstream accumulates at
the dune peak and its back surface.
TIME CHANGES IN THE TRAVELLING DI-
STANCE AND MEAN VELOCITY OF MO-
VING LAYERS AND DUNES
The distance between the points where the leading
edge of a moving layer and a dune stopped and started
moving again and its time changes in each case were
studied. Figure 9 shows the traveldistance curve to
clarify the average velocity (the velocity which takes
Fig. 7 - Profile of bed
surface (CASE-
4: 0.20L/sec)
Fig. 8 - Profile of bed
surface (CASE-
5: 0.30L/sec)
Fig. 9 - Travel-distance curve of moving layers and dunes
background image
A. IkEDA, T. MIZUYAMA, N. SUGIURA & Y. HASEGAwA
96
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
charge dynamically changed, forming a conspicuous
peak as in CASE-1. The peaks of both discharges and
sediment volume suddenly increased and then gradu-
ally decreased. The ratio of q
out
/ q
in
was around 1.3,
and the peak, which ratio of q
out
/ q
in
was around 2.0,
appeared when the dune collapsed and caused a sud-
den movement of the sediment as CASE-1. Actual
sediment volumes without air void exhibited good
adaptations to changes in discharges in all cases, and
the sediment concentration was approximately 25%.
DISCUSSION
The following deformation processes of stream bed
deposit and their factors are estimated based on the ob-
servations and analyses described in the above chapters:
1) stream bed gradient (gradient of a flume); 2) shape of
stream bed (fore surface slopes and back surface slopes
of moving layers and dunes); 3) flow ranges of surface
water (discharge) and infiltration flow, etc. In terms of
flow ranges of surface water and infiltration flow, the
surface gradient of the infiltration flow inside moving
layers and dunes were studied and the relationships of
each factor was discussed based on the fact that moving
layers and dunes start moving when the infiltration flow
reaches the lower end of the moving layer or the dune
or the surface water flow over the dune shoulder.
Figure 11 shows the relationship between shape of
stream bed deposit and stopping gradient (the fore sur-
face and back surface gradients of a moving layer and
dune). The back surface gradients are about the same
as or smaller than the stream bed gradients, and the
fore surface gradients are about 1.5 times steeper than
stream bed gradients in all cases. Both the fore sur-
account of the time of moving and stopping). Flume
gradient shift points and their gradients are shown on
the line in the figure. Travelling distances greatly vary
from several centimetres to several meters regardless
of the discharge volumes and gradients in all cases.
In terms of gradients, the mean velocity tends to be
faster in steep gradient zones and slower in low gradi-
ent zones. The velocity increases or decreases in pro-
portion to discharges. Moving layers and dunes tend
to take longer before starting to move again when they
reach areas where the gradient is gentler than 18°.
TIME CHANGES IN THE SUPPLIED wATER DI-
SCHARGE, RUNOFF DISCHARGE AND SEDI-
MENT VOLUME
Figure 10 shows time changes in supplied water
discharge (qin), runoff discharge at the downstream
end (clear-water discharge and discharge contain-
ing sediments (q
out
), and sediment volume (sediment
volume without void). In CASE-1, the deposit move-
ment, deformation of the deposit and discharges dy-
namically changed from the start to the end of the
flow. The peaks of both discharges and sediment vol-
ume appeared around 300 seconds and then gradually
decreased. The ratio of q
out
/ q
in
was around 1.3. The
peak, which ratio of q
out
/ q
in
was around 2.0, appeared
when the dune collapsed and caused a sudden move-
ment of the sediment. Meanwhile, CASE- 2 to 4 and
until the first 200 seconds in CASE-5, the discharge
change is small, and the ratio of q
out
/ q
in
was around
0.8 and the sediment discharge at the downstream end
was small, too. Immediately after the collapse of the
dune (after 200seconds) in CASE-5, however, the dis-
Fig. 10 - Relationship between supplied water discharge and runoff discharge, sediment volume
background image
DEFORMATION OF STREAM BED DEPOSIT AND RUNOFF PROCESS AT DEBRIS FLOW INITIATION ZONE
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
97
than stream bed gradients, as in the slopes while stop-
ping. Meanwhile, the fore surface gradients of moving
layers and dunes and the gradient of infiltration flow
surface show that while there are variations depend-
ing on discharges, the fore surface gradients before
moving are about the same as the infiltration flow sur-
face gradients. It probably indicates the condition in
which the infiltration flow surface gradient increases
and matches the fore surface gradient of moving lay-
ers and dunes, which is the condition that arises when
the surface of infiltration flow goes over the shoulders.
In terms of the supplied water discharge (q
in
), run-
off discharge containing sediments at the downstream
end (q
out
) and sediment volume, while a small dis-
face slopes and back surface slopes are proportional
to the gradients. While there is no major difference in
tendencies among discharges, they tend to show large
variations because dunes are formed in low-gradient
zones, and the fore surface gradients of dunes change
due to the sediments supplied from the upstream.
Figure 12 shows the relationship between the gra-
dient of stream bed deposit (the fore surface gradients
of moving layers and dunes) and infiltration flow sur-
face. In comparison to the gradient while stopping in
Figure 11, the fore surface gradients are steep in all
cases, and the forefront of moving layers tend to bulge
or dune shoulder tend to rise higher in them. Also,
the fore surface gradients are about 1.5 times steeper
Fig. 11 - Relationship between shape of stream bed deposit and stopping gradient
Fig. 12 - Relationship between gradient of stream bed deposit and imlfiltration flow surface
background image
A. IkEDA, T. MIZUYAMA, N. SUGIURA & Y. HASEGAwA
98
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
charge (0.20 L/sec or less) does not have a large influ-
ence on the downstream end, a discharge goes through
large fluctuations when it gets larger (0.30 L/sec), and
the ratio of q
out
/ q
in
was around 1.3. It is probably
because the movement is based on state of traction
when the discharge is 0.20 L/sec or smaller, and the
movement is based on debris flow when the discharge
becomes larger than 0.20 L/sec. The discharge and
sediment volume largely depend on dune shapes and
moving process at the 15° to 18° zones; when a dune
is formed and the surface water is retained at the back
surface of the dune, the discharge slightly decreases,
and the discharge also increases when it goes over the
dune shoulder and the dune start to move.
CONCLUSION
In this study, the following aspects were found in
experiments conducted based on gradient-changing
flumes and discharge conditions which mimicked
natural torrents.
● The stream bed deposit at the initiation zone of de-
bris flow is also moving (deforming) during a small-
scale discharge which is only slightly larger than an
infiltration flow. The deformation processes based
on gradients are roughly categorized as follows:
1) 30° to 27°: The moving layer with a clear leading
edge is formed. The moving layer flows down over
the stream bed deposit. The velocity is relatively fast.
2) 24° to 21°: The moving layer stops and forms a
dune. The moving sediment shifts from a moving
layer to dune. The entire dune moves gradually
with the development of a dune shoulder. The ve-
locity is slower than 1).
3) 18° to 12°: The dune is still the main morphology
type. The velocity is slow. The stream bed gets defor-
med instead of experiencing sediment movement.
● When a moving layer or a dune is formed, the surfa-
ce water infiltrates into the layer at its upper end or
the back surface. Thus, no surface water is observed
at the fore surface of the moving layer or the dune.
● Deposits tend to stop and accumulate at zone of
gradient shifts, and dunes are likely to form and
rise higher at the 24° gradient zone (deposits tend
to accumulate).
● Moving layers formed in relatively steep gradient
zones (equivalent of 1) in the previous descriptions)
start moving when the deposit gets saturated with
the surface water or infiltration flow. Also, dunes
formed in relatively low gradient zones (equivalent
of 2) and 3) in the previous descriptions) tend to
start moving when the deposit gets saturated with
the surface water or infiltration flow, or when the
surface water goes over the dune shoulder.
● Changes in discharge and sediment volume at the
downstream end become active when the supplied
water discharge is larger than 0.30 L/sec, and the ra-
tio of q
out
/ q
in
was around 1.3. Meanwhile, changes
in discharge and sediment volume become small
when the supplied water discharge is smaller than
0.30 L/sec, and the ratio of q
out
/ q
in
was around 0.8.
In actual settings, the former is estimated to be the
debris flow and the latter the state of traction.
● Changes in discharge and sediment volume at the
downstream end (the shape of hydrograph) is rela-
ted to the dune forming, deformation and moving
process at the 15° to 18° gradient zones.
CONCLUDING REMARKS
Based on this study, the stream bed deposit at the
initiation zone of debris flow is estimated to be mov-
ing (deforming) even during slight discharges. The
study also indicated the possibility that after going
through such moving (deforming) processes, deposits
are not necessarily in stable forms; rather, they accu-
mulate in unstable forms like dunes where fore surfac-
es become steep, or they accumulate in a large amount
at zone of gradient shifts. The study also indicated the
possibility that surface water occurs if a discharge that
is larger than the infiltration flow is supplied from the
upstream (including from sides and branch streams)
regardless of the saturation level of the deposit. In ad-
dition, the discharge and shape of stream bed deposit
at the initiation zone of debris flow are estimated to
have great influences on the hydrograph which flows
toward downstream. In general, slope collapses and
mobilization of stream bed deposit have been esti-
mated as cause of debris flow occurrence. According
to the forms of stream bed deposit and its deformation
at the initiation zone of debris flow conjectured in this
study, however, occurrence conditions of debris flows
might change. As a future study, it is necessary to in-
vestigate and observe the phenomena clarified in this
study in actual torrents, examine how temporal and
spatial moving (deformation) processes of deposits
at the initiation zone of debris flow affect conditions
which cause debris flow, and explore the conditions
background image
DEFORMATION OF STREAM BED DEPOSIT AND RUNOFF PROCESS AT DEBRIS FLOW INITIATION ZONE
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
99
ments in the experiment, and to measure cohesion and
internal friction of the sediment which may have a role
in affecting the formation and movement of the dunes.
which cause debris flow while taking account actual
torrent conditions. And also we intend to observe and
analyze the sediment movement using different sedi-
REFERENCES
i
keda
a., m
izuyama
t., s
uGiuRa
n. & H
aseGawa
y. (2009) - Study about deformation of stream bed deposit at initiation zone of
debris flow. Journal of the Japan Society of Erosion Control Engineering 62 (4): 46-51. (In Japanese)
i
keya
H. & m
izuyama
t. (1982) - Flow and Deposit Properties of Debris Flows. Report of Public Works Research Institute,
Ministry of Construction 157: 88-153. (In Japanese)
k
aki
t. (1955) - Experimental Study on the Mud Flow 1st Report. Journal of the Japan Society of Erosion Control Engineering
19: 1-6. (In Japanese)
m
izuyama
t. & H
u
, P. (1989) - Development and Deformation of Debris Flow. Proc. The Japan-China Symposium on Landslides
and Debris Flows, 1989, Niigata, Tokyo, Japan, pp.217-222
s
atofuka
y. & m
izuyama
t. (2009) - Numerical Simulation of Stony Debris Flow Developing on Unsaturated Deposit. Annual
Journal of Hydraulic Engineering, JSCE 53: 697-702. (In Japanese)
t
akaHasHi
t. (1977) - A Mechanism of Occurrence of Mud-Debris Flows and their Characteristics in Motion. Annuals of
Disaster Prevention Research Institute, Kyoto University, 20B-2: 405-435. (In Japanese)
y
ano
k. & d
aido
a. (1959) - Fundamental Study on Mud-Flow (1). Annuals of Disaster Prevention Research Institute, Kyoto
University, 3: 11-16. (In Japanese)
y
ano
k., d
aido
a. & s
umino
m. (1959) - Experiments on the Criterion for Occurrence of Mud Flow. Annuals of Disaster
Prevention Research Institute, Kyoto University, 12B: 323-331. (In Japanese)
Statistics