Document Actions

IJEGE-11_BS-Trentini-et-alii

background image
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
983
DOI: 10.4408/IJEGE.2011-03.B-107
AN INTEGRATED APPROACH FOR DEBRIS FLOW HAZARD ASSES-
SMENT - A CASE STUDYON THE AMALFI COAST - CAMPANIA, ITALY
m
aRia
n
iColina
PAPA
(*)
, G
iuliano
TRENTINI
(**)
, a
ntonio
CARBONE
(***)
& a
ntonio
GALLO
(***)
(*) University of Salerno, Dept. of Civil Engineering - Via Ponte don Melillo - 84084 Fisciano (SA), Italy
(**) ELEMENTI Associated, via La Marmora 51 - 50121 Firenze (FI), Italy
(***) GEORES - A. Carbone & A. Gallo Associated - Via M. Vernieri 119 - 84125 Salerno (SA), Italy,
INTRODUCTION
The governmental organizations in charge of pro-
tection against natural risks, usually develop the strat-
egies for risk mitigation on the basis of a risk map
displaying the different natural risks affecting the terri-
tories studied. A risk map is obtained through the over-
laying of a hazard map with a vulnerability map. The
realization of a hazard map, that requires the estimation
of the event intensity along with its probability, is the
most difficult task. In the case of debris flows, a com-
plete and trustworthy methodology has yet to be devel-
oped. The intensity of debris flows, in a particular point
of the land surface, may be evaluated by the means of
the depth and velocity of the flow. Therefore, a cor-
rect hazard estimation passes through the estimation,
in each point of the studied area, of velocity and depth
along with the associated probability of occurrence.
Debris flow phenomena are generally not frequent
enough to allow for a direct estimation of probability on
the basis of the observation of past events. Therefore,
the probability of an event, with a given intensity, has to
be derived from the estimation of the probability of the
variables contributing to the event formation.
Assuming the land characteristics do not change
over time, the triggering variable is the rainfall. The
rainfall with a given intensity and duration, and given
probability, should be coupled with the consequent
mobilization of soils and formation of debris flows
Simple land stability models, available in litera-
ture (m
ontGomeRy
& d
ietRiCH
, 1994) are “static” and
ABSTRACT
In this work, a multidisciplinary study is presented
in which potential debris flow events are studied from
their beginning to their end. The case study is located
on the Amalfi Coast where historical events of debris
flow are well documented in 1910, 1924 and 1954. An
integrated approach was used for the geomorphology
and geo-pedology of volcanic deposits. The matches
found between morphotypes, depositional and pedo-
logical processes and soil characteristics made it pos-
sible to develop a detailed map of the soil deposits
which was then used to estimate the geographical
distribution of the parameters relevant in the stability
analysis. The modeling of debris flow initiation proc-
ess is carried out through a stability analysis of the
soil after a rainfall characterized by a given intensity
and duration. The debris flow volume corresponding
to a given return period is subsequently obtained. The
propagation of this volume is then simulated through
a commercial, two dimensional model (FLO2D). Both
modeling approaches, for the initiation and the propa-
gation processes, are tested by comparing them with
a historical event in 1954. The integrated procedure
described makes it possible to draw up maps of debris
flow intensity corresponding to assigned return peri-
ods. These elements may be easily elaborated in order
to draw up debris flow hazard and risk maps.
K
ey
words
: debris flow, hazard map, modelling
background image
M.N. PAPA, G. TRENTINI, A. CARBONE & A. GALLO
984
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
and covers the municipalities of Ravello and Minori
(Amalfi coast, province of Salerno - Fig. 1).
The northern part of the basin is mountainous
and has a rough morphology with an elevation vary-
ing between 800 and 1000 meters above sea level.
Proceeding southward, towards the coastal area, the
maximum elevation of the relief gradually decreases
until it reaches 200 meters.
The mountainsides of the basin have very high
slopes and are often interrupted by sub-vertical walls,
the most widespread slope is between 30° and 45°.
These areas are used predominantly for forestry (hard-
wood forests of holm-oak and chestnut) with areas of
Mediterranean bush. The foothills and lower slopes
(with a gradient between 17° to 30°), are mostly ter-
raced for agriculture (vineyards, citrus orchards and
mixed orchards). The urbanized area coincides with
the final stretch of the T. Sambuco valley towards the
alluvial fan and the narrow coastal strip
METHODOLOGY KEY POINTS
CHARACTERIZATION OF THE SOIL DEPOSITS
The deposits covering the slopes of the studied
area are constituted by ashes, pumice and scoriae, and
are defined, from a geological point of view, “pyro-
clastic fall”. This type of deposits mainly derives from
the eruptive activity of the Somma-Vesuvius volcano
(eruptions immediately before, and after the in 79 A
D, until the last one in 1944).
These deposits are layered, rehandled and often
profoundly weathered. They overlay a large portion
of the reliefs especially in more morphologically de-
pressed areas. It is well known that in the Campania
region, as well as different Italian mountain systems
(i
amaRino
& t
eRRibile
, 2008), the weathering and the
pedogenetic processes of such deposits have produced
soils with peculiar characteristics, defined “andic
properties” (k
eys
to
s
oil
t
axonomy
, 2006).
Andic soils have a unique set of morphological,
physical and chemical properties that can be attrib-
uted to the presence of noncrystalline and/or poorly
ordered clay minerals such as allophone, imogolite,
ferrihydrite and Al/Fe- humus complexes (P
aRfitt
,
1990; s
HoJi
et alii, 1993). These soils, classified as
Andosols, are characterized by large porosity, elevat-
ed water retention, a friable structure, variable-charge
minerals, large organic matter and thixotropy proper-
ties. The andic characteristics are often distributed in
allow for the estimation of the effect of a rainfall with
a given intensity and indefinite duration. This means
that the association of a probability to the occurrence
of this rainfall is not feasible and as a consequence
the hazard estimation is not realizable. In this work,
an attempt has been made to overcome this difficulty.
Once the volumes of the soils mobilized by the
rainfall with a given probability have been estimated,
the propagation of the solid liquid mixture may be
simulated with a two dimensional mathematical model
(o'b
Rien
, 2007, m
edina
et alii, 2008) and the event in-
tensity distribution may be evaluated in the alluvial fan.
Thus, an intensity map can be developed for
each probability, with it subsequently being possible
to draw up a hazard map.
In order to allow for a simple and safe use of
the developed methodology, simulation models have
been chosen that have been widely tested in differ-
ent geological contests, that are largely documented
in literature and that are available as open source or
commercial software.
In order to perform a trustworthy estimation of
mobilized soil volume, land stability model requires
a detailed map of soil deposit characteristics. In the
present work a rapid and economic method for soil
deposits mapping is proposed
Both modeling approaches, for the initiation and
the propagation processes, are tested by the back
analysis of a past event.
STUDY AREA
The catchment area of Torrente Sambuco has an
areal extent of approximately 6.5 square kilometers
Fig. 1 - Location of the study area (Italy, Campania Re-
gion, Sambuco catchment in red)
background image
AN INTEGRATED APPROACH FOR DEBRIS FLOW HAZARD ASSESSMENT - A CASE STUDYON THE AMALFI COAST - CAMPANIA,
ITALY
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
985
In this study, the problem has been addressed by
the well known correspondence between surface geo-
morphology and the spatia variability of soils and de-
posits; it has been scientifically described by various
authors (C
onaCHeR
& d
alRymPle
, 1977; b
iRkeland
,
1984), opening the way for rapid and economic meth-
ods of mapping soils on large areas, not otherwise
achievable, especially in mountainous areas.
An integrated mapping (scale 1:2,000) of the
geo-pedological and geomorphological characteris-
tics of the area concerned was therefore carried out,
based on the observed correlation between the mor-
photypes, pedogenetic conditions and soils detected
(Fig. 2). The methodological approach used for the
integration of the geomorphology and pedology
follows, in large part, as indicated in G
uida
et alii
(2007), for similar near areas.
Once the territory has been divided into morpho-
logical units and associations of soils, with each one
being characterized by a profiletype representative
(stratified pyroclastic sequences where present), an at-
tempt was made to assign characteristic hydrological
and geotechnical parameters to the four most common
horizon profiles (see Table 2). This procedure was car-
ried out on the basis of detailed studies, made by some
authors, on the physical, mechanical and hydrological
characterization of the different layers that make up the
pyroclastic covers of the Campanian relief (b
ilotta
et
alii, 2005; b
asile
et alii, 2003).
a non-homogeneous way into the soil horizons, result-
ing in a marked vertical anisotropy in its physical and
chemical properties. Mainly due to some of their com-
mon properties, andic soils have high soil water con-
tents, often equal to or even higher than its dry weight
(b
asile
et alii, 2003).
The soils of Sambuco catchment has a complex
sequence, due to the deposition of volcanic materi-
als that have covered the carbonaterocks at different
times, which has lead to interrupting the previous
pedogenetic cycles (as witnessed by buried soils).
The typical horizons sequence of a pedological
profile in the study area is: A-Bw-C-2Bwb. However,
this sequence is often delayed by erosion and post-
depositional processes.
The contact between the limestone and volcanic
soil is generally abrupt. However, in some places,
between Andosols and limestone, the presence of
ancient soils strongly argillified (Alfisols) can be ob-
served, which fill pockets and fractures in carbonate
rocks. This arrangement results in a marked variabil-
ity of the soil properties as well as the absence of a
single soil stratigraphy reference. Consequently, one
major problem in drawing up a detailed map of the
deposits, and therefore a good prevision of possible
instable soil volume, is in identifying the most appro-
priate method for defining the spatial distribution of
the different soil typologies.
Fig. 2 - Integrated mapping of soils and morphotypes
(simplified label description in Table 1)
T
ab 1 - Outline of the correspondence between morpho-
types and typical soil profiles
background image
M.N. PAPA, G. TRENTINI, A. CARBONE & A. GALLO
986
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
SHALSTAB simulation of land stability condition
has been performed after a rainfall having duration
equal to 15 days.
The shape of the flow hydrograph has been as-
signed trough the following relation:
where Q
p
is the peak discharge of the hydrograph
and t
p
is a characteristic time length. The total debris
flow volume (W) and the peak discharge are related
as follow:
The total debris flow volume is assumed equal to
the instabilized soil volume, and tp has been estimat-
ed through taking into account the length path of the
farthest instabilized pixel from the control section,
and the mean velocity of the flow along this path.
The soil concentration of the mixture is estimat-
ed as the complement to 1 of the averaged instabi-
lized soil deposits porosity
SIMULATION OF DEBRIS FLOw PROPAGA-
TION
The propagation of debris flows has been simu-
lated through the widespread used commercial code
FLO-2D (o'b
Rien
, 2007). This model solves the
Saint-Venant equations in the two dimensional frame
(m
aHmood
& y
evJeviCH
, 1975). The FLO-2D model
makes it possible to simulate the presence of build-
ings, by limiting the flow that can cross the computa-
tional cell as well as the presence of roads, which are
instead preferentially paths
The flow resistance is described through a rheologi-
cal equation (o'b
Rien
& J
ulien
, 1988, 1993; o'b
Rien
,
2007) declinable for different kinds of mixture:
The first term (τ
y
) is the threshold stress that must be
overcome in order to give place to the flow. This term
takes into account both the cohesion, linked to the pres-
ence of fine sediments, as well as the Mohr-Coulomb
stress. The second term represents the stress due to vis-
cosity, which, according to Newton's law, has a trend
proportional to strain rate (du/dy) through a viscosity co-
ESTIMATION OF DEBRIS FLOw INPUT
HYDROGRAPH
The widespread used SHALSTAB model (m
ont
-
GomeRy
& d
ietRiCH
, 1994) has been employed to simu-
late the land instabilities caused by a given rainfall. The
SHALSTAB model assumes that the precipitation has
a constant intensity for a time interval sufficiently long
to reach a stationary underground water flow. This hy-
pothesis has two important consequences. The first is
that, when performing a back analysis, instability areas
are usually overestimated. This happens because the
triggering rainfall is usually shorter than the time nec-
essary for reaching the stationary conditions and there-
fore the effect of a longer rainfall is simulated instead
of the real one. The second consequence of the cited
hypothesis is the difficulty of defining the return period
associated to a given soil instability scenario.
In order to overcome these limitations, in the
present work, a rough estimation of the rainfall duration
is given, assuming that it is equal to time necessary for
the soil to reach steady state conditions (s). This time
is assessed by an extremely simplified volume balance
between input and output water flux through the basin
(P
aPa
et alii, 2010):
where n is the total number of computation cells, a
i
is
the contributing area to the cell i, b is the width of the
cell along the contour line, k is soil permeability, a is
the slope and s is the water content at saturation.
According to this equation, the interval time re-
quired to reach steady conditions, for the Sambuco ba-
sin closed at the outlet, is 75 days. For the sub-basins
closed at their junction with the main valley, the time
intervals are smaller, varying with the size of the sub-
basin and its average slope, and falling in a range be-
tween 15 and 51 days.
Debris flow origins mainly in the first order lit-
tle sub-basins; with reference to these basins, the
Tab. 2 - Hydrological and geotechnical characteristic pa-
rameters assigned to the horizons (typical profile
1, 2, 3, 4, 6)
(1)
(2)
(3)
(4)
background image
AN INTEGRATED APPROACH FOR DEBRIS FLOW HAZARD ASSESSMENT - A CASE STUDYON THE AMALFI COAST - CAMPANIA,
ITALY
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
987
shown in Figure 3.
Once given the regression lines of Figure 3, the
curves of equation (5) that best approximate the param-
eters can be estimated. The regression curves obtained are
shown in Figure 4.
The estimated values for the resistance parameters of
equation(5) are reported below:
α
1
= 0,0145 Pa
β
1
= 8,3
α
2
= 0,0134 Pa
β
2
= 14,1
The depth integrated law resistance in non-dimen-
sional form is (o'b
Rien
& J
ulien
, 1993):
where S
f
is the energy gradient, γ
m
is the specific
weight of the mixture, h is the flow depth, V is the
flow velocity, n
td
is a generalized Manning coefficient
(o'b
Rien
, 2007) and k is a coefficient that takes into
account the fact that under real scale conditions, the
flow resistance is increased by the presence of obsta-
cles and macro-roughness as well as by the effect of
efficient (η). The third term is the sum of the effects of
the collisions between particles and the turbulent stress.
Both these components depend on the square of the strain
rate (b
aGnold
, 1954; t
akaHasHi
, 1991; e
GasHiRa
et alii,
1997) and are expressed through the coefficient χ.
The shear strength (τ
y
) and viscosity (η) depend on
solid concentration (C) through the following equations
where α
1
, β
1
, α
2
e β
2
are parameters depending on the mix-
ture properties.:
This formulation can be specified to describe the
behaviour of the mixtures involved in the debris flows
of the Campanian Apennine. These mixtures have a
significant fine content of clay and silt (about 20%) and
at the same time, a high content of coarse sediments,
with sands and gravels. This particular composition is
responsible for a particular rheological behaviour that
is influenced both by the presence of fine particles, giv-
ing rise to a threshold-type cohesive effort and a high
viscosity, as well as the presence of coarser sediments
that give rise to the collisional behaviour
Laboratory tests (P
aPa
& m
aRtino
, 2007 m
aRtino
,
2003), were carried out on samples taken in areas affected
by debris flow phenomena in May 1998 in the town of
Sarno. The mixtures involved in these events have the
same main characteristics, genesis and granulometry of
the Amalfi Coast soils, and therefore the results of the
rheological tests performed on the samples collected in
the Sarno area may be used in characterizing the debris
flow mixtures of the Amalfi Coast basins, such as the
Sambuco basin.
The rheological tests showed a shear thickening be-
haviour, described in detail by a Herschel Bulkley equa-
tion (C
oussot
, 1996) with the exponent of strain rate be-
ing equal to 1.7..
In order to use the FLO-2D code, the rheological
equation should be put in the form of equation (4). This
is possible from experimental data, although this equa-
tion gives rise to a lower agreement with the experimen-
tal data than the Herschel Bulkley type. For each test, the
solid concentration is known and the shear stresses are
measured at different strain rate values. For each test, it
is then possible to estimate the quadratic equation of the
type of equation (4) that best approximates the experi-
mental trend. The curves obtained and their equations are
(5)
Fig. 4 - Regression curves of the resistance parameters
depending on the solid concentration of the mix-
ture
Fig. 3 - Regression curves of the experimentally mea-
sured shear stress and shear rate
(6)
background image
M.N. PAPA, G. TRENTINI, A. CARBONE & A. GALLO
988
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
the compact cross sections. As noted by m
aRtino
&
P
aPa
, 2008, the restraining effect that occurs in the in-
channel flux is responsible for a significant reduction
in flow velocity. Although the estimate of k, is crucial
for a correct characterization of the motion, a rational
criteria for its assessment is still not available.
The authors of the FLO-2D code propose empirical
relationships to calculate the parameters ntd and k de-
pending on the Manning's coefficient (o'b
Rien
, 2007).
In the present work, the Manning coefficient was esti-
mated according to the type of sliding surfaces, using
literature values (C
How
, 1959).
Being that all the rheological parameters had al-
ready been obtained from other tests or literature, it
was not necessary to calibrate the model in order to
estimate the significant parameters.
BACK ANALYSISI OF 1954 DEBRIS FLOW
EVENT
HISTORICAL DATA
On 25 October 1954, on the Amalfi Coast and
in the city of Salerno, several detachments from
the slopes occurred, resulting in waves of debris
flow that also invested the village of Minori, caus-
ing extensive damage and three deaths. The debris
flow covered part of the historical centre of the vil-
lage, damaged many houses, as well as carried large
amounts of debris onto the beach.
Table 3 shows a list of the documentary sources
that have been consulted in order to both estimate
the main characteristics of the event as well as ob-
tain useful data for the validation of the simulation
models. The documentary sources are classified as:
(1) photographs; (2) publication of historical studies
(b
uonomo
& G
ambaRdella
, 2004); (3) video docu-
mentaries (d
e
i
uliis
, 2004); (4) documents claiming
damages by individuals and companies; (5) list of
homeless families
The location of the different documentary sourc-
es is shown in Figure 7.
The reconstruction of the tracks of the landslide
is part of another study that is currently underway,
aimed at mapping all the landslides in the entire area
affected by the flooding of 1954.
COMPARISON AMONG OBSERVED AND SIMU-
LATED
TRIGGERING AREAS
The reconstruction of the meteorological event
made by the Servizio Idrografico e Mareografico Ital-
iano (SIMI) has provided the map of isohyets that in-
tersected with the Sambuco basin which gives, for the
daily precipitation of the 25th of October, an average
of 321 mm. In Figure 5, the unstable areas assessed
by the model SHALSTAB, are represented, as a re-
sult of a precipitation intensity of 321 mm/day. In the
same figure, the landslide scars reconstructed on the
basis of field investigations are also plotted. It is also
worth noting that the simulation greatly overestimates
the event. The simulated area affected by the landslide
is 33.6% of the total area of the basin, while the area
actually affected was equal to only 2.8%. This result
is not surprising given the consideration discussed
above. The simulation was repeated taking into con-
sideration the cumulative precipitation during the 15
days preceding the event, resulting in an average in-
tensity of 22 mm/day. The results are shown in Figure
6. In this case, the simulated landslides cover a total
area of the 2.8% of the basin.
However, there is a marked difference in the
shapes of the landslide, in the simulations, with the
instable areas being made up of many distributed pix-
els, while in the real case the areas are more compact.
These differences can be attributed to the fact that the
Tab. 3 - Evidence of the effects of the 1954 event
background image
AN INTEGRATED APPROACH FOR DEBRIS FLOW HAZARD ASSESSMENT - A CASE STUDYON THE AMALFI COAST - CAMPANIA,
ITALY
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
989
The simulations presented here were carried out on
a 10 m computational grid, with a higher resolution not
being possible due to numerical stability reasons.
A first simulation was carried out assuming that
the total volume mobilized contributed to form a
single surge (single impulse). In this case, the solid-
liquid input hydrograph is characterized by a total
volume of 300,000 m
3
and a peak flow of 180 m
3
/s.
Figure 7 shows the total area invaded which appears
to be higher than the observed one.
According to witnesses (b
uonomo
& G
am
-
baRdella
, 2004), therewere several waves. Therefore,
the propagation of the hydrograph resulting from the
landslide with the largest volume (63,000 m
3
) was
also simulated. In this scenario (several impulses), the
peak flow is 58 m
3
/s. The effects, in terms of maxi-
mum envelope of flow depths for each computational
cell, are shown in Figure 7.
In this case, the agreement between the simulation
and observations is very good. All the positions at which
debris flow effects have been documented result being
affected by the simulated debris flow (see Tab. 3).
The only position (ID 14) that is not touched by
the simulated debris flow, while affected by the real
one, is in a small road that cannot be reproduced on
the used10 m grid.
The simulation reproduces the volume of 1500 cu-
bic meters of material, which had filled the depressed
area at the archaeological site. The thickness of the de-
bris that according to witnesses had filled the ground
floor of buildings (ID 11) is also well simulated. The
single pixel, that is unstable in the simulation, is stable
in reality, due to the presence of adjacent stable pixels.
Moreover, an unstable cluster of pixels that begins to
move may involve other adjacent pixels, though these
appear to be stable in the simulations. These processes
are not reproduced by the procedure outlined here.
Further studies and investigations are needed in order
to establish the criteria to describe them. However,
when aiming to the simulation of the downstream
propagation of a debris flow only the total triggered
volume is necessary, regardless to the shape and dis-
tribution of the instabilized areas.
Based on the analysis carried out, it appears that
the total volume triggered by a steady rainfall with
intensity equal to the average of the cumulated rain-
fall of duration estimated by equation (1), despite the
limitations related to the extreme simplicity of the ap-
proach, gives significantly accurate estimates which
can be used for technical applications.
COMPARISON AMONG OBSERVED AND SIMU-
LATED FLOwS
At the time of the event the Sambuco stream flowed
into the centre of the village of Minori, along the houses
and the main street (position ID 4 in Fig. 7), the down-
stream part the river flowed into an underground canal
(beginning in ID 6) until reaching the sea outlet. In the
simulations, it has been assumed that the underground
canal was completely clogged with sediments, due to
its small section. This hypothesis is substantiated by
testimonies relating to it being blocked.
Fig. 6 - Landslide of 1954, comparison among observa-
tions and thesimulation obtained with a triggering
rainfall intensity of 22 mm/day
Fig. 5 - Landslide of 1954, comparison among observa-
tions and the simulation obtained with a trigger-
ing rainfall intensity of 321 mm/day
background image
M.N. PAPA, G. TRENTINI, A. CARBONE & A. GALLO
990
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
Many different surges may develop, as it was ob-
served in the 1954 event, but it is not possible to make
a reliable prevision of this behaviour. Thus, for the
sake of safety, a single surge event was considered.
All the relevant parameters of the calculation are
summarized in Table 4.
DISCUSSION AND CONCLUDING RE-
MARKS
Debris flow hazard mapping requires the develop-
ment of various models in order to assess the effects
of precipitation with assigned return period in terms
of flow depths and velocities in the flooded area. In
particular, the system considered in this study consists
of four parts: the mapping of soil characteristics, the
modeling of DF initiation, the modeling of DF propa-
flow velocities seem consistent with the magnitude
of effects observed. These indeed have values of ap-
proximately 2 m/s along Corso Vittorio Emanuele,
where the debris flow produced the largest amount of
damage to buildings, with values of the order of 1 m/s
in areas adjacent to the Corso Vittorio Emanuele, and
values under 0.5 m/s in the remaining part of village
where there are documented deposits but without any
damage to the structures.
HAZARD MAP
Several criteria for the classification of debris
flow intensity and hazard, are available in literature
(H
üRlimann
et alii, 2008). In the present work, the cri-
terion proposed by R
iCkenmann
(2005) has been used,
with a minor modification (see Fig. 8)
Once fixed the return period (T) and the rainfall du-
ration (15 days), the intensity of the triggering rainfall
can be easily estimated from the Intensity-Duration-
Frequency curves relevant for the Sambuco catchment.
The volumes of possible debris flows have been conse-
quently estimated, with the procedure described above.
Fig. 7 - Simulation of the total debris flow area in the
“single impulse” scenario and envelope of maxi-
mum flow depths in the ”several impulses” sce-
nario
Fig 8 - Hazard matrix adapted from Rickenmann (2005).
T is the return period of the event
Fig. 9 - Calculated hazard map
background image
AN INTEGRATED APPROACH FOR DEBRIS FLOW HAZARD ASSESSMENT - A CASE STUDYON THE AMALFI COAST - CAMPANIA,
ITALY
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
991
od needs to be developed if the delimitation of po-
tentially unstable slopes is required. But when, like
in the present case, the object of the study is the pre-
diction of extension and magnitude of invaded areas,
only the total triggered volume is necessary, regard-
less to the shape and distribution of the instabilized
areas. In these circumstances the employed model
seems to be appropriate.
The runout process simulation has provided fairly
good results. After the testing of the two modelling
methodologies, the entire system has been applied for
the drawing up of hazard maps
ACKNOWLEDGEMENTS
The work described in this publication was sup-
ported by the European Community Seventh Frame-
work Programme through the grant of the Collabora-
tive Project IMPRINTS (IMproving Preparedness and
RIsk maNagemenT for flash floods and debriS flow
events), Contract FP7-ENV-2008-1-226555
The simulation methodology here presented has
been employed for the development of a pilot study
for debris flow risk assessment, in the framework of
the hydrogeological risk assessment plan of Destra
Sele Basin Authority (Campania Region).
gation, the intersection of event probability (assessed
by the triggering rainfall return period) with the event
intensity (resulting from the propagation model)
With the aim of providing a debris flow hazard
mapping methodology for technical application, the
models selected for the simulation of triggering and
propagation processes are both widespread used mod-
els that have been tested in many different geological
contests. In order to allow for a wide use of the devel-
oped system, simulation models have been chosen that
are available as open source or commercial software.
The routines employed for the simulation of DF
formation and DF propagation have been previously
tested by comparison with an historical event, recon-
structed on the basis of historical documentation.
The model used for the prevision of land instabilities
gave simulated triggering areas very different from
the observed ones. Nevertheless the total triggered
volume was estimated accurately. Some other meth-
Tab. 4 - Relevant parameters of the estimated events used
to develop the hazard map
REFERENCES
b
aGnold
R. a. (1954) - Experiments on a gravity-free dispersion of large solid spheres in a Newtonian fluid under shear. Proc.
R. Soc. London, A 225: 49-63.
b
asile
a., m
ele
G. & t
eRRibile
f. (2003) - Soil hydraulic behaviour of a selected benchmark soil involved in the landslide of
Sarno 1998. Geoderma - Elsevier, 117: 331–346.
b
ilotta
e., C
asCini
l., f
oResta
v. & s
oRbino
G. (2005) - Geotechnical characterization of pyroclastic soils involved in huge
flowslides. Geotechnical and Geological Engineering, Springer Ed, 23: 365-402.
b
iRkeland
P.w. (1984) - Soils and geomorphology. Oxford University Press. New York. 372 pp.
b
uonomo
s. & G
ambaRdella
m. (2004) - Minori antica Rheginna Minor, eventi alluvionali tra il 1910 e il 1954. Terra del Sole
edizioni.
C
How
v.-t. (1959) - Open Channel Hydraulics. McGraw Hill, New York NY
C
onaCHeR
a.J. & d
alRymPle
J.b., (1977) - The nine-unit landsurface model: an approach to pedogeomorphic research. Geo-
derma, 18: 1-154.
C
oussot
P. & m
eunieR
m., (1996) - Recognition, classification and mechanical description of debris flows. Earth-Science Re-
views, 40: 209-227.
d
e
i
uliis
C. (2004)- Minori racconta. ricordi dell’alluvione del 1954. Video realizzato nell’ambito delle manifestazioni per il
cinquantenario dell’alluvione del 25 ottobre 1954.
e
GasHiRa
s., m
iyamoto
k. & i
toH
t. (1997) - Constitutive equations of debris flow and their applicability, Proceedings of the 2
nd
International Conference on Debris Flow Hazard Mitigation: Mechanics, Prediction, and Assessment, Balkema, 340-349.
G
uida
d., C
aRbone
a., C
estaRi
a., C
aRdiello
G., d
e
n
aRdo
a., G
allo
a., b
uonoConto
a., i
amaRino
m., l
anzaRa
R., s
ieRvo
v. (2007) - Interdisciplinary approaches to the flowslide-debris flow hazard assessment in the weathered pyroclastic soil-
mantled carbonate hillslopes
. Experiences in Campania Region (Southern Italy). In: EGU European Geosciences Union
background image
M.N. PAPA, G. TRENTINI, A. CARBONE & A. GALLO
992
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
- General Assembly -Vienna, Austria.
H
üRlimann
m., R
iCkenmann
d., m
edina
v. & b
ateman
a. (2008) - Evaluation of approaches to calculate debris-flow parameters
for hazard assessment. Engineering Geology, 102: 152-163.
m
aHmood
k. & y
evJeviCH
v. (1975) - Unsteady flow in open channels, water resources publications, Fort Collins, Colorado.
m
aRtino
R. (2003) - Experimental analysis on the rheological properties of a debris flow deposit. Proceedings of the 3
rd
Inter-
national Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction, and Assessment, Davos, Switzerland.
m
aRtino
, R. & P
aPa
, m. (2008) - Variable-concentration and boundary effects on debris flow discharge predictions. Journal Of
Hydraulic Engineering, 134: 1191-1404.
m
edina
v., b
ateman
a. & H
üRlimann
m. (2008) - FLATModel: a 2D finite volume code for debris-flow modelling. Application
to events occurred in the Eastern Pyrenees. Int. J. Sediment Research, 23 (4): 348-360.
m
ontGomeRy
d.R. & d
ietRiCH
w. e. (1994) - A physically-based model for topographic control on shallow landsliding. Water
Resources Research, 30(4): 1153-1171.
i
amaRino
m. & t
eRRibile
f. (2008) - The relevance of andic soils in mountain ecosystems in Italy: a pedological investigation.
European Journal of Soil Science, December 2008, 59: 1284-1292.
k
eys
to
s
oil
t
axonomy
(2006) - United States Department of Agriculture - Natural Resources Conservation Service. Tenth
Edition.
o’b
Rien
J.s.& J
ulien
P.y. (1988) - Laboratory analysis of mudflow properties. Journal of Hydraulic Engineering, 110: 877-887.
o’b
Rien
J. s., J
ulien
P. y. & f
ulleRton
w.t. (1993) - Two-dimensional water flood and mudflow simulation. Journal of Hydraulic
Engineering, 119 (2): 244-261.
o'b
Rien
J. s. (2007) - FLO-2D User’s Manual, Version 2007.06. FLO-2D, Nutrioso, AZ.
P
aPa
m.n. & m
aRtino
R. (2007) - A numerical model for heterogeneous and confined debris flows. Proceedings of the 4
th
Interna-
tional Conference on Debris-Flow Hazards Mitigation: Mechanics, Predictions, and Assessment, Chengdu, China, Millpress
Science Publishers, Rotterdam, Netherlands, 197-208.
P
aPa
m.n. & t
Rentini
G. (2010) - Valutazione della pericolosità connessa a fenomeni di correnti detritiche, Proceedings of the
XXXII Convegno Nazionale di Idraulica e Costruzioni Idrauliche, Palermo, 14-17 settembre 2010.
P
aRfitt
R.l. (1990). Soils formed in tephra in different climatic regions. Transactions of the 14
th
International Congress of Soil
Science, 7: 134-139.
R
iCkenmann
d. (2005) - Hangmuren und Gefahrenbeurteilung. kurzbericht für das Bundesamt für wasser und Geologie. Un-
published report, Universität für Bodenkultur, Wien, und Eidg. Forschungsanstalt WSL, Birmensdorf, 18 pp.
t
akaHasHi
t. (1991) - Debris flow. IAHR/AIRH monograph, Rotterdam (Balkema).
Statistics