Document Actions

IJEGE-11_BS-Hurlimann-et-alii

background image
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
553
DOI: 10.4408/IJEGE.2011-03.B-061
DEBRIS-FLOW MONITORING STATIONS IN THE EASTERN PYRENEES.
DESCRIPTION OF INSTRUMENTATION, FIRST EXPERIENCES AND PRE-
LIMINARY RESULTS
MARCEL HÜRLIMANN
(*)
,CLAUDIA ABANCÓ
(*)
, JOSE MOYA
(*)
, CARLES RAÏMAT
(**)
& ROBERTO LUIS-FONSECA
(*)
Dept. of Geotechnical Engineering and Geosciences, Technical University of Catalonia, Spain
(**)
Geobrugg Ibérica SA, Spain
INTRODUCTION
Field observations of moving debris flows by
means of monitoring stations are of great impor-
tance to improve understandings of triggering, flow
behaviour and accumulation mechanisms. Upon
the knowledge of the authors, in Europe debrisflow
monitoring stations are only situated in the Alps:
Italy (b
eRti
et alii. 2000; m
aRCHi
et alii. 2002), Aus-
tria (H
uebl
& k
aitna
2010) and Switzerland (H
üR
-
limann
et alii. 2003a; b
adoux
et alii. 2009), while
other stations are located in China (z
HanG
1993), Ja-
pan (s
uwa
et alii. 2009), t
aiwan
(y
in
et alii. 2007)
and USA (l
a
H
usen
2005) among others. However,
no test site is located in a catchment affected by
Mediterranean climate.
In the Eastern Pyrenees, debris flows are no as re-
ported as in other mountain ranges, but can cause im-
portant damages as shown for example by the events
occurred in 2008 (P
oRtilla
et alii. 2010). In order to
improve the knowledge on the triggering conditions
and the dynamic behaviour, three monitoring systems
in the Eastern Pyrenees have been set up.
In the following, the monitoring systems in-
stalled in the test sites will be described and the first
results and experiences will be discussed. The results
presented here come from one of the sites, the Senet
catchment, which revealed a high activity during the
first year test phase.
ABSTRACT
Monitored observation stations represent a funda-
mental tool to properly investigate the initiation, flow
behaviour and accumulation of debris flows. In the
recent years, tree different monitoring stations have
been built up in the Eastern Pyrenees. The instrumen-
tation of all of them consists of four geophones and a
rain gauge, while two of them also have an ultrasonic
device and one site a video camera. First experiences
regarding the set-up and calibration of the different de-
vices indicate that debris-flow monitoring is a complex
task and requires knowledge of different research ar-
eas. In particular, qualified electronic skills are essen-
tial. The preliminary results show that especially the
Senet test site, with its initiation area in a steep and vo-
luminous glacial deposit, presents a high debris-flow
activity. Several events were recorded during the first
year test phase and the analysis of precipitation data
showed that most of the debris flows were triggered by
short duration-high intensity rainfalls. The interpreta-
tion of the monitoring data related to the flow behav-
iour was not easy, because only geophone measure-
ments, rainfall data and post-event field observations
were available for the process analysis in most of the
events. Thus, the visual information of a video camera
is very helpful to carry out the calibration of monitor-
ing data and to clarify doubts of interpretation.
K
ey
words
: monitoring, ground vibration, rainfall threshold,
Pyrenees
background image
M. HÜRLIMANN, C. ABANCÓ, J. MOYA, C. RAÏMAT & R. LUIS-FONSECA
554
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
SELECTION OF TEST SITES
As first step, a debris-flow database for the East-
ern Pyrenees was built up by the interpretation of aer-
ial photos and the research of different types of publi-
cations including technical reports and administration
archives. The locations of the debris-flows gathered in
this database are shown in Fig. 1.
Then, the activity of selected torrents have been
checked by aerial photographs of different years and
finally three test sites have been chosen to set up the
monitoring system (Table 1). Ensija site is situated in
the Pre-Pyrenees, where bedrock mostly consists of
marls, conglomerates and limestones that are covered
by a periglacial colluvium. The two other sites (Senet
and Erill) are located in the Axial Pyrenees, where
bedrock consists of slates and the superficial deposits
are glacial deposits (tills).
DESCRIPTION OF DEVICES INSTALLED
In Table 2, the principal devices installed in the
three monitoring stations are listed. Information and
evaluation of the devices available for debris-flow mon-
itoring can be found in several reviews (i
takuRa
et alii,
2005; l
a
H
usen
, 2005; a
Rattano
& m
aRCHi
, 2008).
GEOPHONE
Geophones measure the ground vibration gener-
ated by debris flows (e.g. a
Rattano
, 2000). In each
catchment, four geophones GEOSPACE 20 DX with a
natural frequency of 8 Hz and a standard coil resistance
of 395 Ohms have been installed. An electronic signal
conditioner transforms these vibrations into a number of
GENERAL SETTING
The Eastern Pyrenees limit the Spanish region of
Catalonia from France and include the Principality of
Andorra (Fig. 1). The highest peaks are located at the
axis range and reach almost 3000 m a.s.l.
From a geological point of view, two zones are
clearly distinguishable (e.g. m
uñoz
, 1992): i) the Ax-
ial Zone that corresponds to the paleozoic basement,
and ii) the covering layers that form the outer zones
at both, northern and southern sides of the range.
Our study concerns the Axial Zone and southern
outer zone (here called Pre-Pyrenees). The basement
consists almost entirely of igneous and metamorphic
Paleozoic rocks formed and tectonised during the
Hercinian orogeny and deformed again during the
Alpine orogeny. The cover is composed of sedimen-
tary sequences of Mesozoic and Paleogene age. Col-
luvial deposits reach a thickness of a few meters in
some low order catchments of the range and glacial
deposits are found in the upper reaches of the Axial
Pyrenees, locally presenting a thickness of several
tens of meters.
The climate of the Eastern Pyrenees is strongly
influenced by two factors: the vicinity of the Medi-
terranean Sea and the orographic effects of the
Pyrenean mountain range. There are two typical
rainfall patterns, which can trigger debris-flow ac-
tivity (H
üRlimann
et alii, 2003b): i) Short duration,
high intensity rainfalls related to convective summer
storms, and ii) moderate intensity rainfall during au-
tumn/winter lasting for several days or weeks and af-
fecting large areas.
Fig. 1 - Location of the three monitoring stations. white
dots show the debrisflow events gathered in the
database and used during the site selection
T
ab. 2 - Number of the devices installed in the three test
sites
Tab. 1 - Situation and general information on the three
test sites
background image
DEBRIS-FLOW MONITORING STATIONS IN THE EASTERN PYRENEES. DESCRIPTION OF INSTRUMENTATION, FIRST EXPERIENCES
AND PRELIMINARY RESULTS.
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
555
the effectiveness of such kind of protection measure-
ment. Detailed information on this topic can be found
in l
uis
-f
onseCa
et alii (2011).
LOGGING AND TRANSMISSION OF DATA
Two different clusters of devices can be distin-
guished in each test site: i) the meteorological station,
and ii) the “flow station”, which consists of all the oth-
er devices installed along a torrent reach (geophones,
ultrasonic sensor, video camera and load cells). Both
clusters incorporate a datalogger, and a GSM modem
for data transmission, all of them powered by a battery
which is recharged by means of a solar panel.
The meteorological station is controlled by a
Campbell Scientific CR200 datalogger powered by a
12V 7Ah battery connected to a 10W solar panel and
also contains a Wavecom Fastrack GSM modem.
The readings from the sensors connected to the
“flow station” are recorded in a Campbell Scientific
CR1000 datalogger. The datalogger is programmed
to control the frequency of sensors scanning and data
recording, as described in the following section. The
power supply is established by 12V 24Ah battery
which is recharged by a 30W solar panel. The video
camera is powered separately by an additional bat-
tery and a solar panel. Data transmission is completed
again by a Wavecom Fastrack GSM modem
impulses per second (IMP/sec), if a certain threshold of
acceleration is exceeded. Therefore, the output data of
the geophones provide vibration intensity of the pass-
ing debris flow, which can be used to estimate a mean
front velocity between the different geophones. In addi-
tion, the geophones trigger the other measuring devices
installed along the torrent. Geophone data can also be
used to determine the moment of debris-flow initiation,
which is necessary information in the rainfall analysis.
ULTRASONIC DEVICE
Ultrasonic devices measure the flow depth of
a passing debris flow. In two catchments, Senet and
Ensija, an ultrasonic device (UC6000-30GM-IUR2-
V15 manufactured by P
ePPeRl
+f
uCHs
) was installed.
The raw ultrasonic measurements, which depend on
the air temperature, are automatically and internally
corrected by means of a temperature sensor that is
connected to the device. Additional configurations on
the longitudinal range of measure, the opening or the
numbers of measures could be realised by the program
ULTRA3000 provided by P
ePPeRl
+f
uCHs
.
The ultrasonic data in combination with the geo-
phone data can be used to estimate a mean flow veloc-
ity and finally a discharge.
VIDEO CAMERA
Video cameras provide very helpful qualitative
information on the general flow behaviour. Video or
photographic images can also be used for detailed
processing (e.g. C
HanG
& l
in
, 2007). In Erill, a stand-
ard GANZ video camera is installed in combination
with a spot light.
METEOROLOGICAL STATION
The meteorological station includes a rain gauge
and a thermometer. The rain gauges are standard tip-
ping bucket devices with a resolution of 0.1 mm (RM
YOUNG 52203). Because rain gauges are unheated,
the temperature data is necessary to distinguish, if
precipitation has been rain or snow. The recording in-
terval of both sensors is defined constant as 5 minutes
FLEXIBLE RING-NET AND LOAD CELLS
In Erill test site, a flexible ring VX160 of GEO-
BRUGG was installed in combination with several
load cells of 500 kN capacity incorporated along the
horizontal cables. The goal of these devices is to test
Fig 2 - Oblique view of the Senet test site indicating the
position of the different devices installed. GEOn:
geophone number n. US: ultrasonic ic sensor
background image
M. HÜRLIMANN, C. ABANCÓ, J. MOYA, C. RAÏMAT & R. LUIS-FONSECA
556
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
hyperconcentrated flows). Thus, geophones provide
indirect information that must be correctly interpret-
ed. In contrast, the ultrasonic sensor supplies flow
depth and, therefore, contributes to the identification
of debris-flow. In Senet test site, many “monitoring
events” were recorded during the first year test phase.
Nevertheless, the ultrasonic sensor was most time not
operational due to technical problems and also due to
a rockfall that destroyed the device. Thus, geophone
data were the principal information for debris-flow
identification. However, geophone vibration not only
depends on the type of flow process, but also on the
properties of the channel bed (e.g. presence of soil or
rock, soil thickness, type of soil matrix) and on the
distance between the flow and the sensor. Finally, the
interpretation of monitoring data was based on both
the geophone records and the periodic field reconnais-
sance looking for morphologic changes.
The first step during the interpretation of the
“monitoring events” involved a determination of con-
fidence on the geophone data, which was expressed
as a qualitative likelihood. The “monitoring events”
were classified in four likelihood categories high,
medium, low and null. The different classes were as-
signed considering the maximum vibration intensity
(in IMP/sec), the event duration and the shape of the
impulse time-series. Our data and experiences from
other studies (e.g. a
Rattano
, 2000) indicate that a
mature debris flow seems to give a record with a con-
tinuous vibration, a quick-rising and slow-decreasing
shape, a minimum duration of ground vibration of at
DATALOGGER PROGRAMMING
A main characteristic regarding debris-flow mon-
itoring systems is the fact that it should distinguish
between a “no-event” mode, when no debris flow oc-
curs in the torrent, and an “event” mode in the case of
debris-flow occurrence. This means that an adequate
programming of the datalogger is of essential im-
portance. Figure 3 shows a flowchart of the program
mounted in the Senet datalogger controlling the four
geophones and the ultrasonic device. In “noevent”
mode the datalogger scans the four geophones at 1 Hz
(1 scan per second) and checks the number of impuls-
es of each geophone. If one of the geophones exceeds
the threshold of 20 IMP/sec, the program switches to
“event”-mode, herein called “monitoring event”.
If no geophone exceeds the defined threshold, the
information on vibration and an ultrasonic measure-
ment will be recorded in a “no-event” output file every
5 minutes. In the “event” mode, the four geophones
and the ultrasonic device record at 1Hz. After the de-
bris flow passage and the vibrations at the geophones
decrease below the threshold of 20 IMP/sec, the sys-
tem keeps in “event” mode during 2 additional min-
utes before switching back to “noevent” mode.
DEBRIS-FLOW EVENT IDENTIFICA-
TION
Geophone data include measurements of pass-
ing debris flows, but also of other phenomena that
induce ground vibration (e.g. rock fall from the steep
outcrop of till in the source zone, thunder during a
storm, boulder transport by torrential flows or by
Fig. 3 - Flowchart of the program installed in the datalog-
ger controllinggeophones and ultrasonic device in
Senet
Tab. 3 - Summary of maximum geophone vibration inten-
sity and corresponding likelihood of selected mon-
itoring events at Senet test site. Maximum rainfall
intensity related with the events is also indicated
background image
DEBRIS-FLOW MONITORING STATIONS IN THE EASTERN PYRENEES. DESCRIPTION OF INSTRUMENTATION, FIRST EXPERIENCES
AND PRELIMINARY RESULTS.
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
557
monitoring events with maximum intensities greater
than 80 IMP/sec. Second, events of medium likelihood
showed usually intensities greater than 40 IMP/sec..
In spite of all these drawbacks that have occurred
during the first year test phase, the summary of events
listed in Table 3 indicates that the Senet test site is
characterised by a rather high debris-flow activity.
EXAMPLES OF DEBRIS-FLOW EVENTS
RECORDED IN THE SENET MONITO-
RING SYSTEM
SENET: AUGUST 7, 2009
Just 2 weeks after installation of the monitoring
system, a first debris flow was recorded at Senet. At
that moment, a preliminary version of the datalogger
program was running and only two geophones (GEO1
and GEO4) were operational. Nevertheless, interesting
data and new experiences to improve the monitoring
system could be gathered.
The recorded data on rainfall, ground vibration
and flow depth are shown in Fig. 4. Geophone and ul-
trasonic measurements indicate two main peaks. These
peaks could be interpreted as two main debris-flow
surges and are clearly correlated to two peaks in the
rainfall intensity. The rainfall was a typical convective
summer storm and lasted about 5h. The total rainfall
amount was about 75 mm, but a clear maximum in-
least some tens of seconds and a peak greater than 50
IMP/sec. Additionally, a progression of the vibration
down the channel should be recorded by the different
geophones. Finally, we also incorporated the informa-
tion gathered during the periodic controls focussing
on geomorphologic changes in the field. This field
controls have especially been used to support events
of high likelihood.
PRELIMINARY RESULTS
In the following, three different types of data will
be described, all of them focussing on the Senet test
site. First, a summary of selected monitoring events
recorded at the Senet site is presented. Second, de-
tailed data of three specific cases will be shown. These
three events were on one side classified as a high
likelihood (Tab. 3) and on the other side the geomor-
phologic changes in the field indicated that a debris
flow has occurred. Third, the analysis of the rainfall
data will be described focussing on triggering condi-
tions for debris flows and analysing different rainfall
thresholds.
During the first year test phase between August
2009 and July 2010, a total amount of 280 “monitor-
ing events” have been gathered, although many are
clearly false alarms corresponding to null-likelihood
events (events with duration of one or two seconds).
In fact, after the initial filtering of these “monitoring
events” only 23 events could be classified with a like-
lihood higher than null. This means that almost three
events per month could be registered from August to
October 2009 and from March to July 2010. These
events were subsequently analysed more in detail and
classified with low, medium or sure likelihood.
Table 3 shows that 6 events could be classified
with high likelihood, 6 with medium and 11 with low
likelihood. The long distance to the Senet test site
(which requires a trip of two days) and the high fre-
quency of monitoring events hindered to carry out a
field inspection after each event. That’s why monitor-
ing events of medium and low likelihood could not be
directly checked by a field inspection
Although this classification of the events in differ-
ent likelihood categories involves many uncertainties,
some general patterns on the measured ground vibra-
tion could be observed. Two conclusions can be done
regarding the maximum intensity. First, events char-
acterised by high likelihood generally correspond to
Fig. 4 - Rainfall, geophone and ultrasonic device data
registered during the Senet debris flow on August
7, 2009
background image
M. HÜRLIMANN, C. ABANCÓ, J. MOYA, C. RAÏMAT & R. LUIS-FONSECA
558
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
tensity of about 6 mm/5min or 30 mm/h could be ob-
served around 13:30. Regarding the geophone and ul-
trasonic data, no approximation of mean flow velocity
between the devices could be done and especially the
measurements of flow depth may represent underesti-
mated values, because the defined recording frequency
of 10Hz (one record per decisecond) was leading to
some malfunction in the data gathering.
SENET: MARCH 25, 2010
An important debris-flow event with a volume
of about 2000 m3 was recorded at the end of March
2010. A rainfall with a small total amount of about 20
mm and a low intensity of less then 2 mm/5min and
about 8.7 mm/h triggered this rather important event
(Fig. 5). A major increase in temperature during the 10
days preceding the debris flow support the hypothesis
that snowmelt may have played a significant role in
the initiation mechanisms. This aspect was the topic of
another analysis (H
üRlimann
et alii, 2010) and will not
be discussed herein.
Finally, the rainfall causing the debris flow was
characterised by a low intensity and by a duration
of several hours, which represent a common pat-
tern for rainfalls occurring during spring season in
Eastern Pyrenees. The debris flow took place around
midnight and temperature was between 4-6 ºC. The
ground vibrations produced by the debris flow lasted
from ~1 minute at Geo1 and Geo2 to ~8 minutes at
Geo4 and illustrate interesting information on the
flow behaviour (Fig. 5). The quality of the data dif-
fers for each of the four geophones, but especially
geophone 4 (Geo4) shows a sharp increase of the
vibration intensity with a subsequent continuous de-
crease. This shape represents the debris-flow front
with maximum discharge and maximum concentra-
tion of boulders and stands for a common feature in
debris-flow behaviour (e.g. a
Rattano
, 2000).
Vibrations gathered by three geophones (Geo2,
Geo3 and Geo4) include a peak and thus some pre-
liminary velocity estimates could be carried out.
The resulting mean velocity of about 3 – 5 m/s is
rather low and can only be explained by a stop-and-
go mechanism. No ultrasonic device measurements
could be gathered, because a previous rockfall de-
stroyed the steel cables that fixed the sensor.
SENET: JULY 11, 2010
A rather different debris flow occurred in the
afternoon of July 11th 2010. The registered ground
vibrations lasted almost three hours from 12:43 to
15:23 and several surges represented by maximum
Fig.5 - Rainfall, and geophone data registered during the
Senet debris flow on March 25, 2010. watch dif-
ferent axes for the geophone data
Fig.6 - Rainfall, and geophone data registered during the
Senet debris flow on July 11, 2010
background image
DEBRIS-FLOW MONITORING STATIONS IN THE EASTERN PYRENEES. DESCRIPTION OF INSTRUMENTATION, FIRST EXPERIENCES
AND PRELIMINARY RESULTS.
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
559
facts that should be taken into account during the setup
of a monitoring station or a warning system.
Again, an approximation of the mean flow veloc-
ity between the geophones Geo2, Geo3 and Geo4 has
been carried out using the data of the peak at about
14:00. Finally, velocity estimates between 3 and 4
m/s could be calculated, which again represent rather
low values for a granular debris flow along a channel
bed with a slope of about 20º.
RAINFALL TRIGGERING CONDITIONS
Empirical rainfall thresholds are a common tool
in the research of rainfall-triggered slides and debris
flows and can be established using statistic analysis
of historic data (e.g G
uzzetti
et alii, 2007). There are
different types of thresholds, but the most common is
the one proposed by C
aine
(1980) comparing rainfall
duration and mean intensity.
The definition of such a rainfall threshold for
debris-flow initiation in the Eastern Pyrenees is one
of the main objectives of the three monitoring sys-
tems installed in the region. The data gathered dur-
ing the test phase of the observation stations provide
first important information on the rainfall conditions
triggering debris flows, but do not yet allow the de-
termination of a reliable threshold. In the following,
precipitation data are again related to the Senet test
site, since most of the information has been gathered
at that monitoring station. In this preliminary analy-
sis, only the rainfall data associated with monitoring
events, which have been classified as medium and
high likelihood for debris-flow occurrence, are taken
into account (Fig. 8).
A first conclusion regarding the general rainfall
pattern indicates that most of these monitoring events
were triggered by short duration-high intensity rain-
falls. However, other events were related to moderate
values of impulses per second can be observed (Fig.
6). The ground vibrations perfectly coincide with the
duration of the rainfall measured (about three hours).
The rainfall pattern represents the characteristics of a
short convective storm. The total rainfall registered
was 97.6 mm and the maximum intensity was about 9
mm/5min and 49.3 mm/h. The largest detected ground
vibrations were observed at about 14:00 coinciding
with the highest rainfall intensity. The field reconnais-
sance carried out 8 days after the event showed great
morphologic changes (channel erosion in some sec-
tions and accumulation in other ones, see Fig. 7).
In contrast to the spring event shown in Fig. 5, all
the geophones registered high vibration intensity. The
explanation that even the geophone 1 (Geo1) reflected
this time high ground vibration, was a change of the
flow path. During the July event, the debris flow passed
much closer at Geo1 reducing the distance between the
moving sediment and the sensor, which consequently
produced an increase of the ground vibration. Such al-
terations of the flow paths reducing or increasing the
distance between channel bed and sensor are important
Fig.7 - Morphologic changes in the higher part of the Senet fan caused by the debris flow occurred on July 11th 2010 (left:
before and right: after the event)
Fig.8 - Critical rainfalls recorded at Senet test site during
the first year test phase (only the events related
high likelihood are illustrated). Three different
thresholds are added for comparison
background image
M. HÜRLIMANN, C. ABANCÓ, J. MOYA, C. RAÏMAT & R. LUIS-FONSECA
560
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
data, however, showed that such a frequency was ex-
ceeding the capacity of the devices installed.
Additional problems have been related to the pow-
er supply, because not enough sunshine was available
to charge the battery during autumn. So, larger solar
panels and batteries have been installed subsequently.
Another unexpected problem arose from the rock-
fall occurrence in the Senet site. Several rockfalls in-
cluding large granitic boulders failed out of the glacial
deposit and caused considerable damages at some mon-
itoring devices located in the higher part. One rockfall
of unexpected volume even destroyed the ultrasonic
device situated initially near the geophone GEO2 posi-
tion. Both the ultrasonic device and the meteorological
station were dislocated down into an area not affected
by rockfalls. Minor damages had to be accepted by van-
dalism and animal activity.
Finally, a major problem was the correct inter-
pretation of the geophone data. False alarms due to
triggers not associated with fluvial processes could be
identified easily, but an accurate distinction between
different processes such as debris flows, hyperconcen-
trated flows or sediment transport was difficult. Thus,
all the events monitored during the test phase were
classified into different likelihood categories apply-
ing the approach previously explained. However, an
improvement of the visual control in test sites with
high activity is unavoidable. Therefore, in the Senet
test site, a high-resolution video camera MOBOTIX
M12D-Sec DNight D43N43 was set up in August
2010. The installation of a video camera is considered
as very helpful or even fundamental to improve the
analysis of the monitored debris flows. Image analy-
sis from video camera records will also contribute to
resolve uncertainties on flow velocity estimates and
doubts on the flow behaviour.
In spite of the problems encountered during this
test phase, several conclusions can be obtained. Moni-
toring results of the first year show that the Senet test
site situated in the Axial Pyrenees is characterised by
high activity. On one side, abundance of debrisflow
prone material and high slope angles in the initiation
zone, which consists of a thick till deposit, create per-
fect conditions for debris-flow initiation. On the other
side, frequent convective rainstorms are common in
this high-mountain area. However, not only convec-
tive summer storms have triggered debris flows in
Senet, but also a combination of moderate to small
intensity rainfalls and the one at the end of March 2010
probably associated with antecedent snowmelt. The
different rainfall data gathered at Senet can be com-
pared with other thresholds published. No comprehen-
sive rainfall threshold has been established for debris-
flow occurrence in the Eastern Pyrenees and only some
very general criteria for landslide initiation have been
published (C
oRominas
& m
oya
, 1999; C
oRominas
et
alii, 2002). Finally, three thresholds were selected for
comparison with the Senet data: i) the threshold de-
fined for the debris-flow monitoring station installed at
Illgraben (m
C
a
Rdell
& b
adoux
, 2007; b
adoux
et alii,
2009), ii) the threshold established for the Moscardo
test site (d
eGanutti
et alii, 2000), and iii) the well-
known global threshold defined by C
aine
(1980). The
rainfall events that triggered event classified with high
likelihood (see Tab. 3), are generally fitting the three
threshold conditions. However, the debris flow oc-
curred during spring 2010 is clearly situated below the
critical rainfall amount, which supports the hypothesis
that an additional water input may have been available
and that snowmelt may have played an important role
in the initiation of that event.
The conclusion of this preliminary analysis of the
critical rainfalls indicates that different thresholds must
be established for such kind of debris-flow triggering.
CONCLUSIONS
During the selection, the configuration and calibra-
tion of the different devices and also during the set-up
and testing of the monitoring system many experiences
could be gathered that will be subsequently discussed.
These comments may help other research groups in-
terested in debris-flow monitoring to build up a new
system or to improve an existing one.
Most of the problems regarding the device configu-
ration have been related to electronic tasks. In particu-
lar, the electronic schemes controlling the geophone
signals are rather sophisticated. Moreover, the selection
of the resistor value at the conditioner of the geophone
signal, which defines the critical acceleration to initiate
the impulse count, is of great importance. This value
depends on several factors such as underground lithol-
ogy or the distance to the torrent, all of them related to
the damping of vibration.
A first version of the datalogger program included
a registering frequency in the “event” mode of 10Hz
for the geophones and the ultrasonic device. Recorded
background image
DEBRIS-FLOW MONITORING STATIONS IN THE EASTERN PYRENEES. DESCRIPTION OF INSTRUMENTATION, FIRST EXPERIENCES
AND PRELIMINARY RESULTS.
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
561
calibration of all the devices and sensors.
The major conclusion obtained during the test
phase is the fact that debris-flow monitoring only with
geophones and ultrasonic devices is possible, but can
not provide enough data for a thorough analysis of the
flow behaviour. That’s why a video camera is recom-
mended in test sites with rather high activity.
ACKNOWLEDGEMENTS
This research was supported by the Spanish Min-
istry of Science and Innovation, contract CGL2008-
00299/BTE (project DEBRISCATCH) and the EC FP7
project IMPRINTS, contract ENV-2008- 1-226555.
Technical help from Swiss Federal Research Institute,
WSL, is also acknowledged.
rainfall with snowmelt in early spring.
In contrast, only minor debris-flow frequency
could be observed in the Ensija and Erill test sites. The
explanations of this are rather different. In Ensija, the
sediment availability is more limited due to the collu-
vium layer and only large rainstorms seem to generate
debris flows. In Erill, the initiation area is similar as in
Senet, but the bed slope of the channel is much lower.
The experiences gathered at the three test sites
show that debrisflow monitoring is a complex and
difficult task including many research areas. Apart
from the typical research fields like hydraulics and
geomorphology, also areas like geophysics, electron-
ics, telecommunications etc. are necessary. Especially
electronic skills are essential for a correct set-up and
REFERENCES
a
Rattano
M. (2000) - On the use of seismic detectors as monitoring and warning systems for debris flows. Natural Hazards,
20: 19-213.
a
Rattano
m. & m
aRCHi
l. (2008) - Systems and Sensors for Debris-flow Monitoring and warning. Sensors, 8: 2436-2452.
b
adoux
a., G
Raf
C., R
HyneR
J., k
untneR
R. & m
C
a
Rdell
b. (2009) - A debris-flow alarm system for the Alpine Illgraben catch-
ment: design and performance. Natural Hazards, 517-539.
b
eRti
m., G
enevois
R., l
a
H
usen
R., s
imoni
a. & t
eCCa
P.R. (2000) - Debris flow monitoring in the Acquabona watershed on
the Dolomites (Italian Alps). Physics and Chemistry of the Earth, 25: 707-715.
C
oRominas
J. & m
oya
J. (1999) - Reconstructing recent landslide activity in relation to rainfall in the Llobregat River basin,
Eastern Pyrenees, Spain. Geomorphology, 30: 79-93.
C
oRominas
J., m
oya
J. & H
üRlimann
m. (2002) - Landslide rainfall triggers in the Spanish Eastern Pyrenees. In 4th EGS Plinius
Conference "Mediterranean Storms". Editrice, Mallorca.
C
HanG
s.y. & l
in
C.P. (2007) - Debris flow detection using image processing techniques. In 4th Int. Conf. on Debris-Flow Haz-
ards Mitigation. Eds. m
aJoR
J.J. & C
Hen
C. 549-560. Millpress, Chengdu, China.
d
eGanutti
a.m., m
aRCHi
l. & a
Rattano
m. (2000) - Rainfall and debris-flow occurrence in the Moscardo basin (Italian Alps).
In: Debris-Flow Hazards Mitigation. E
ds
. w
ieCzoRek
G.f. & n
aeseR
N.D.: 67-72. Balkema.
G
uzzetti
f., P
eRuCCaCCi
s., R
ossi
m. & s
taRk
C.P. (2007) - Rainfall thresholds for the initiation of landslides in central and
southern Europe. Meteorology and Atmospheric Physics, 98: 239-267.
H
uebl
J. & k
aitna
R. (2010) - Sediment delivery from the Lattenbach catchment by debris floods and debris flows. In EGU
General Assembly. p. 10585.
H
üRlimann
m., a
banCó
C. & m
oya
J. (2010) - Debris-flow initiation affected by snowmelt. Case study of the Senet monitoring
site, Eastern Pyrenees. In Mountain Risks: Bringing Science to Society. in press, Florence, Italy.
H
üRlimann
m., R
iCkenmann
d. & G
Raf
C. (2003a) - Field and monitoring data of debris-flow events in the Swiss Alps. Canadian
Geotechnical Journal, 40: 161-175.
H
üRlimann
m., C
oRominas
J., m
oya
J. & C
oPons
R. (2003b) - Debris-flow events in the Eastern Pyrenees. Preliminary study on
initiation and propagation. In 3rd Int. Conf. on Debris-Flow Hazards Mitigation. e
ds
. R
iCkenmann
d. &C
Hen
C.: 115-126.
Millpress, Davos.
i
takuRa
y., i
naba
H. & s
awada
t. (2005) - A debris-flow monitoring devices and methods bibliography. Natural Hazards and
Earth System Science, 5: 971-977.
l
a
H
usen
R. (2005) - Debris-flow instrumentation. In: Debris-flow Hazards and Related Phenomena. e
ds
. J
akob
m. & H
unGR
o.: 291-304. Springer, Berlin.
l
uis
-f
onseCa
R., R
aïmat
C., H
üRlimann
m., a
banCó
C., m
oya
J. & f
eRnández
J. (2011) - Debris-flow protection in recurrent
background image
M. HÜRLIMANN, C. ABANCÓ, J. MOYA, C. RAÏMAT & R. LUIS-FONSECA
562
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
areas of the Pyrenees. Experience of the VX systems after the output results collected in the pioneer monitoring station in
Spain.
In 5th Int. Conf. on Debris-Flow Hazards Mitigation. e
ds
. G
enevois
R., H
amilton
d. & P
Restininzi
a. IJEGE, Padua,
Italy.
m
aRCHi
l., a
Rattano
m. & d
eGanutti
a.m. (200) - Ten years of debris-flow monitoring in the Moscardo Torrent (Italian Alps).
Geomorphology, 46: 1-17.
m
C
a
Rdell
b. & b
adoux
a. (2007) - Influence of rainfall on the initiation of debris flows at the Illgraben catchment, canton of
Valais, Switzerland. Geophysical Research Abstracts, 9: 08804.
m
uñoz
a. (1992) - Evolution of a continental collision belt: ECORS-Pyrenees crustal balanced cross-section. In Thrust Tecton-
ics. e
d
. m
C
C
lay
k.R.: 235-246. Chapman & Hall.
P
oRtilla
m., C
HevalieR
G. & H
üRlimann
m. (2010) - Description and analysis of major mass movements occurred during 2008
in the Eastern Pyrenees. Nat. Hazards Earth Syst. Sci., 10: 1635-1645.
s
uwa
H., o
kano
k. & k
anno
t. (2009) - Behavior of debris flows monitored on test slopes of kamikamihorizawa Creek, Mount
Yakedake, Japan. International Journal of Erosion Control Engineering, 2: 33-45.
y
in
H.y., H
uanG
C.J., C
Hen
C.y., y
eH
C.H., l
ee
b.J., f
anG
y.m. & C
HanG
y.H. (2007) - Monitoring ground vibrations generated
by debris flows. In 4th Int. Conf. on Debris-Flow Hazards Mitigation. e
ds
. m
aJoR
J.J. &C. C
Hen
C.: 625-633. Millpress,
Chengdu, China.
z
HanG
s. (1993) - A comprehensive approach to the observation and prevention of debris flows in China. Natural Hazards, 7:
1-23.
Statistics