Document Actions

ijege-13_bs-margarint-et-alii.pdf

background image
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
275
DOI: 10.4408/IJEGE.2013-06.B-25
LARGE LANDSLIDE RISK ASSESSMENT IN HILLY AREAS.
A CASE STUDY OF HUŞI TOWN REGION (NORTH-EAST OF ROMANIA)
M
ihai
C
iprian
MARGARINT
(*)
, D
oru
T
oaDer
JuraVLe
(*)
, a
Drian
GroZaVu
(*)
,
C
risTian
V
aLeriu
paTriChe
(**)
, M
ăDăLina
pohriB
(***)
& i
uLian
C
ăTăLin
sTÂnGă
(*)
(*)
"Alexandru Ioan Cuza" University of Iaşi - Department of Geography - Blvd. Carol I, 20 A - Iaşi, Romania
(**)
Romanian Academy, Department of Iaşi - Geography Group - Blvd. Carol I, 8, Iaşi, Romania
(***)
“Gh. Asachi” Technical University of Iaşi - Blvd. D.Mangeron, 67 - Iaşi, Romania
K
ey
words
: landslides susceptibility assessment, logistic re-
gression, elements at risk, Huşi town, Romania.
INTRODUCTION
Landslides are landscape modeling processes,
often with risk character and important potential eco-
logical and economic consequences. Their occurrence
is controlled both by a series of favoring factors (e.g.
geological, climatic, hydro-geological characteristics,
geomorphometry) and triggering factors (neotecton-
ics, heavy rainfalls, human activities).
In the last three decades, landslides investigations
have passed from qualitative to quantitative approaches
and from purely geological or geomorphological investi-
gations to hazard and risk assessment (V
ARNES
, 1984) at
different time or spatial scales. Although there are some
landslide susceptibility and hazard studies since the
1970s years (F
ELL
et alii, 2008), the scientific approach
still faces many methodological difficulties especially
for quantifying and multi-scale landslide risk mapping
(V
AN
W
ESTEN
et alii, 2006). Therefore, new techniques
and methods have been developed and continuously
improved on various conceptual and methodological
frameworks: geophysical models to assess slope stability
and landslide dynamics (B
OGOSLOVSKY
& O
GILVY,
1977;
G
ALLIPOLI
et alii, 2000; J
ONGMANS
& G
ARAMBOIS,
2007);
susceptibility and hazard prediction (G
UZZETTI
et alii,
2005; B
RENNING
, 2005; G
UZZETTI
et alii, 2006; G
ÜNTHER
et alii, 2012); vulnerability assessment (G
LADE
, 2003;
U
ZIELLI
et alii, 2008; K
AYNIA
et alii, 2008; P
ASCALE
et alii,
ABSTRACT
Large landslides are a common geomorphologi-
cal characteristic of the Moldavian Plateau (Romania),
their presence and manifestation being favored mainly
by geological and climatic conditions. In numerous
cases, the spatial pattern of the large landslides is under
the form of amphitheatres, with impressive dimensions,
reaching sizes of hundreds of hectares. Due to their
defensive characteristics, these landforms often con-
stituted sites for settlements location during the Mid-
dle Ages. The constant growing of the human pressure
(19
th
, 20
th
century) had led to the settlements extension
in areas of high slope instability. A representative case
of settlements location on these landforms is the Huşi
town, with a population of 25.000 inhabitants, situated
in the central part of the Moldavian Plateau. The land-
slide risk assessment for the Huşi town area (approxi-
mately 79 km
2
) was performed on the basis of landslide
susceptibility and exposed elements maps. Landslide
susceptibility was assessed using the logistic regression
approach, taking into account ten quantitative and qual-
itative factors. The cartographic base was represented
by topographic maps at scale 1:5.000 and the high reso-
lution orthorectified aerial image (2010). The elements
exposed to risk were digitized from the same primary
cartographic materials and the General Urban Plan pro-
vided by local administration. The results indicate high
level of landslide susceptibility on the southern side of
the Huşi town and this should be seriously considered
by the decision makers in the land planning projects.
background image
M.C. MARGARINT, D.T. JURAVLE, A. GROZAVU, C.V. PATRICHE, M. POHRIB & I.C.STÂNGĂ
276
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
since the physical scientist have a clear preference for
a hazard-based point of view, while the social scientists
and human geographers rather adopt a structural and
human-centered perspective (S
MITH
& P
ETLEY
, 2008).
To counter the landslide risk in urban areas, four
approaches have been employed by landslide manag-
ers and urban planners: (1) restricting development in
landslide-prone areas; (2) implementing and enforcing
excavation, grading, and construction codes; (3) pro-
tecting existing developments by physical mitigation
measures and (4) developing and installing monitoring
and warning systems (S
CHUSTER
& H
IGHLAND
, 2007).
In Romania, during the last years, several studies
and modern approaches can be mentioned, including
applications of GIS techniques and statistical analysis
methods, evaluation and mapping of the inherent risk
associated with these geomorphological processes
and of landslide susceptibility, in particular (M
ICU
&
B
ĂLTEANU
2009; B
ĂLTEANU
et alii, 2010; G
ROZAVU
et
alii, 2010, 2012; M
ĂRGĂRINT
et alii, 2011; Ş
ANDRIC
et
alii, 2011; A
RMAŞ
, 2011, 2012; N
ICORICI
et alii, 2012).
The current study aims to assess the landslide risk
in terms of susceptibility and exposure appraisal, with-
out considering the temporal variability of landslides or
the economic or functional dimension of vulnerability.
The analysis is applied to the southern part of Huşi town
territory, situated within the Moldavian Plateau, where
landslides represent defining geomorphological feature.
Landslide susceptibility assessment constitutes a manda-
tory step for landslide hazard and risk evaluation (C
ARDI-
NALI
et alii, 2002; G
UZZETTI
et alii
,
2008), the prognosis
and mapping of future landslide locations being possible
only through a better understanding and evaluation of
the importance of favoring and triggering factors.
2010); complex approaches intended to quantify land-
slide risk (G
UZETTI
, 2000; D
AI
et alii, 2002; K
O KO
et alii,
2003; B
ELL
& G
LaDe
, 2004; C
ROZIER
& G
LADE
, 2005).
The last direction became increasingly important
aiming to meet the needs of the present-day society that
records important financial, material and even human
losses caused by these geomorphological processes. In
landslide risk assessment, risk is most often viewed as
a product between hazard, exposure and vulnerability.
Due to the high conceptual complexity, it is not surpris-
ing that there many points of view, with some overlap-
ping between hazard and risk or with understanding
of exposure either as element of hazard or as part of
vulnerability. Susceptibility is seen as the probabil-
ity of landslide occurrence in a given area, taking into
account all preparatory conditions and any triggering
factor, whose action is considered permanently act-
ing. It still remains unknown and aleatory (the hazard)
the moment of landslide triggering, its dynamics and
spreading or the release rate (G
LADE
et alii, 2005). The
exposure reflects a spatial dimension of human struc-
tures (peoples, dwellings and utilities, road or special
networks, other social and economic objectives) and
identifies their presence in areas that may be affected
by the considered probable landslides. Vulnerability is
perhaps the most differently defined component in risk
assessment, varying from the limitations to exposure
or the simply value and percentage of the economic
losses to exposure, preparedness and prevention, cop-
ing ability, adaptive capacity and recovering (S
TÂNGĂ
& G
ROZAVU
, 2012). These different meanings evolved
over time, the role of the intrinsic characteristics of the
society being widely recognized (H
AQUE
& B
URTON,
2005), but the disciplinary context seems to be decisive,
Fig. 1 - Geographical position of study area. The relief, hydrography and the limit of Huşi town
background image
LARGE LANDSLIDE RISK ASSESSMENT IN HILLY AREAS. A CASE STUDY OF HUŞI TOWN REGION (NORTH-EAST OF ROMANIA)
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
277
landslides. In fact, within the Moldavian Plateau,
there are many cases with landslides having the as-
pect of large amphitheaters, with impressive dimen-
sions, reaching sizes of hundreds of hectares. Due to
their defensive characteristics, these landforms often
constituted sites for settlements location during the
Middle Ages. The constant augmentation of human
pressure (19
th
, 20
th
century) has led to the extension
of settlements in areas of high slope instability.
The landslides recognition is facilitated by their
general morphology, their evident scarps, and the
rolling aspect of slope deposits. Still, only about 15%
of the slopes presents evidences of active landslides.
These activations of deluviums are closely related
to the multiannual and annual precipitations regime
and also to some rainfall events when the precipita-
tions are concentrated during the spring or summer
months (P
UJINĂ
, 2003). In the study area, the mean
multiannual precipitations are about 500 mm/year,
the altitude induced variations ranging from 460 to
560 mm/year. About 70% of the annual quantities fal
between April and September (I
RIMIA
et alii, 2011).
In the larger area of the Moldavian Plateau, a se-
ries of periods were identified, for the last 50 years,
STUDY AREA
Huşi town is situated in North-East of Romania
(46°40’30”N, 28°03’30”E), in the central-eastern part of
the Moldavian Plateau (Fig. 1). This is a platform region,
in which the sedimentary layers display a faint inclina-
tion (4-7 m/km) from NNW to SSE. The town is located
in the southern part of a depression area developed along
the right side of Prut River, which marks the border be-
tween Romania and the Republic of Moldova. The el-
evation ranges from 35 m a.s.l., in the valley of the Huşi
River (tributary to Prut River) to 365 m a.s.l., on top of
the hills from the south-western part of the area.
GEOMORPHOLOGICAL SETTING
The overall morphology is characterized by the
dominance of monocline relief, with north facing
cuesta escarpments and south facing dip slopes, con-
sequences of river network adaptation to geological
structure. Along the escarpments, landslides present
an almost continuous distribution. In the southern
part of Huşi town, along the right side of the homo-
nym river, several large, semicircular shaped land-
slides are developed (Fig. 2). These landslide basins
formed through a succession of deep, rotational
Fig. 2 - Satellite image (Google Earth®) of Huşi town: 1-2 and 3-4 represent the directions of geological profiles described
in Fig. 4; A, B, C are the different evolution stages of landslide basins
background image
M.C. MARGARINT, D.T. JURAVLE, A. GROZAVU, C.V. PATRICHE, M. POHRIB & I.C.STÂNGĂ
278
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
SOCIO-ECONOMIC FEATURES
Huşi Depression is predominantly an agricul-
tural area, in which the arable lands occupy about
50% of the region. The western half is dominated
by orchards, while the southern part (the study
area) in mostly covered by vineyards (Fig. 3). The
region of Huşi represents one of the main vine-
yards centres form eastern Romania, with a total
surface of about 2130 ha. The relief soil and cli-
mate conditions are suitable especially for white
wines production (I
RIMIA
et alii, 2011). The popu-
lation of Huşi town has evolved from 500-600
inhabitants, in the 16
th
century, to approximately
2800 inhabitants, at the beginning of the 19
th
cen-
tury, 13,400 in 1899, 15,000 in 1946 and 18,400
in 1956 (G
UGIUMAN
, 1959). In 2004, the town had
29,510 inhabitants, while the preliminary data of
2011 census points out a significant falling down to
25,000 inhabitants.
Initially extended over a low declivity perim-
eter, without landslides, the town grew constantly
during the 20
th
century and occupied gradually
when landslides activity increased for at least 2 con-
secutive years with precipitation excess (1968-1972,
1978-1980, 1996-1998 etc.), but also for single
years, such as 1991 (S
TÂNGĂ
, 2012). These were the
periods when landslide activation was recorded in
Huşi area as well.
Another characteristic of this slope is that the
development degree of landslide basins is more and
more evident from East towards West (A, B and C
in Fig. 2). Along this direction a gradual decrease
of landslides activity is noticed. This situation is in
relation with the different evolution stages of the
area, with the base level and with the development
of the hydrographic network (M
ONTGOMERY
& D
IET-
RICH,
1988).
Generally, it should be noticed that the land-
slide processes have almost a permanent character,
aligning in the normal evolution of this cuesta es-
carpment. The occurrence and intensity of periods
with excessive rainfall, the rapid snow melting or
the earthquakes play a vital role in the activation of
old deluviums.
Fig. 3 - Land use in Huşi region: 1 - arable land; 2 - built area; 3 - forest; 4 - pastures; 5 - vineyards; 6 - complex (arable,
gardens, vineyards and orchards); 7 - complex (forest and pastures); 8 - orchards; 9 - other categories; 10 - lakes
background image
LARGE LANDSLIDE RISK ASSESSMENT IN HILLY AREAS. A CASE STUDY OF HUŞI TOWN REGION (NORTH-EAST OF ROMANIA)
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
279
ic column reveals the lithological products of three sedi-
mentogenetic phases: in the first phase (Middle Bessara-
bian), a sedimentary epiclastic sequence accumulated;
the second one (Middle-Late Bessarabian) was defined
by a mixt sedimentation, predominantly carbonate with
epiclastic episodes; in the third phase, the epiclastic sed-
imentation was again generalized (Fig. 4).
The sedimentary succession of the Middle Bes-
sarabian (Bârnova-Muntele Formation) outcrops bel-
low the altitude of 90 m in the western part of the area
and below 80 meters in the eastern part. It begins with
amalgamated sedimentary bodies with lutites, siltic
lutites and siltites. In the upper levels, there is an
alternation of sandy bodies with subordinated inter-
calations of lutites and siltites, the succession being
finalized by a silto-lutitic body. Lutites in the bottom
of the succession were intercepted with geotechnical
drilling, while the middle and upper strata outcrop in
the southern part of the area.
The sedimentary deposits accumulated in the
Middle-Upper Bessarabian (Şcheia Formation) out-
crop in the southern part of Huşi Depression, between
the altitudes of 90-240 m in the West and 80-150 m
in the East. From the bottom of the suite, the severe
erosion exposed the oolitic limestone and sandstone
stratified deposits with sands intercalations, hav-
ing a thickness of 20-25 m in the western part and
3-5 m in the eastern one (Pietrăria Member). Above
these, there is a prograding sedimentary succession
(Muncelu Member), consisting of a sequence of
sandy-siltic and sandy bodies with intercalations of
lumachelle limestone. After their accumulation, in the
Late Bessarabian, an interruption of sedimentation al-
lowed the formation of the IInd Moldovalah paleore-
lief (i
onesi
& i
onesi
, 1994; i
onesi
et alii, 2005). This
paleorelief was confirmed by the sands with silicified
wood scraps, found at 240 m in the western part of
the area (G
uGiuMan
, 1959), and by the exogenous
clays of the O-Bt horizon, mentioned by p
ohriB
et
alii (2012) to the south of the Voloşeni-Rusca align-
ment. Above the IInd Moldovalah discordance, the
Khersonian-Meotian epiclastic sequence formed,
consisting of Huşi Formation and Nuţaşca-Ruseni
cinerites. This sequence is predominantly sandy with
some intercalations of sandstones, siltites and lutites.
The Meotian cinerites and sands end the lithological
column of the area.
more and more instable terrains, which is a common
evolution pattern for other settlements within the Mol-
davian Plateau as well (M
ĂRGĂRINT
et alii, 2010). Cur-
rently, the infrastructure is relatively scarce both in
terms of roads and railways (the town being located
at the end of Crasna-Huşi trail). Nevertheless, the de-
velopment potential is high, the town being situated
along the IX European route, connecting Bucharest and
Chişinău capitals, along which the construction of a
highway is foreseen.
GEOLOGICAL SETTING
Geologically, the region belongs to the south-
eastern part of the East-European Platform. For the
Romanian territory, this morphostructural unit is
known as the Moldavian Platform.
The geological description of the study area is based
on the data provided by G
UGIUMAN
(1959), J
EANRENAUD
(1971), I
ONESI
& I
ONESI
(1994), I
ONESI
et alii (2005), P
O-
HRIB
et alii (2012) and also on the data obtained from
our own observation in the southern part of the Huşi De-
pression. The granofacial classification of the geological
deposits was done according to the ternary diagram and
the Romanian Standard STAS 1243-83. The granulom-
etric analyzes were performed in the laboratory of geo-
technical research within the Faculty of Engineering of
“Gh. Asachi” Technical University in Iaşi).
In the study area, the outcropping deposits of Bes-
sarabian-Meotian belong to the last cycle of marine
sedimentation in the area of the Moldavian Platform
The geological description of the study area is based
on the data provided by G
UGIUMAN
(1959), J
EANRENAUD
(1971), I
ONESI
& I
ONESI
(1994), I
ONESI
et alii (2005), P
O-
HRIB
et alii (2012) and also on the data obtained from
our own observation in the southern part of the Huşi De-
pression. The granofacial classification of the geological
deposits was done according to the ternary diagram and
the Romanian Standard STAS 1243-83. The granulom-
etric analyzes were performed in the laboratory of geo-
technical research within the Faculty of Engineering of
“Gh. Asachi” Technical University in Iaşi).
In the study area, the outcropping deposits of
Bessarabian-Meotian belong to the last cycle of marine
sedimentation in the area of the Moldavian Platform.
The lithological peculiarities were determined by
the eustatic sea level variations recorded in the Late Bes-
sarabian in the East-Carpathian platform domain (I
ONESI
et alii, 2005; P
OHRIB
et alii, 2012). Thus, the stratigraph-
background image
M.C. MARGARINT, D.T. JURAVLE, A. GROZAVU, C.V. PATRICHE, M. POHRIB & I.C.STÂNGĂ
280
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
METHODOLOGY
The starting point of this study was the drawing
up of landslide inventory based on the high resolution
orthorectified aerial image (2010 edition, pixel size:
0.5 x 0.5 meters), high resolution images available
from Google Earth
®
, completed with field surveys
and mapping. Because of some shortcomings as the
lack of multitemporal data and dating or the ambi-
guities in defining the landslide age (G
UZZETTI
et alii,
2012), in this approach, there were mapped and ana-
lyzed all landslides, regardless the age and the type.
The next phase aimed to realize the landslide sus-
ceptibility map for a region that includes the Huşi town
territory, delineated based on a rectangular cutout of
10.5 x 7.5 km (Fig. 1). Methodologically, there was
chosen one of the statistical methods frequently used
in the international scientific literature, including for
landslide study: logistic regression (A
TKINSON
& M
AS-
SARI,
1998; A
YALEW
& Y
AMAGISHI
, 2005; B
RENNING,
2005; M
EUSBURGER
& A
LLEWELL
, 2009; A
KGÜN
, 2012).
This method has some important advantages: it offers
a greater computational simplicity (F
ALASCHI
et alii
,
2009), GIS software having implemented different
facilities for this kind of analysis (D
AI
& L
EE
, 2002);
it gives freedom to integrate the variables which can
be either continuous or discrete (categorical), or any
combination of both types, and they do not necessar-
ily have normal distributions (G
ORSEVSKI
et alii, 2000;
M
ATHEW
et alii
,
2007); it has the capability to eliminate
unrelated causative factors and to evaluate the signifi-
cance of the related ones, providing more detailed and
reliable outcome (Y
ESILNACAR
& T
OPAL
, 2005; F
ALAS-
CHI,
2009; C
HAUHAN
et alii, 2010; G
HOSH
et alii, 2011);
it offers the possibility to realize models based on a
Fig. 4 - The architecture of sedimentary bodies separated on granofacial criteria. 1-2 and 3-4 – geological section, according
to Fig. 2. O-Bt - deposits with aspect of Bt illuvial soil horizon, a and b – slidings and slumps; A and B - lithological
column (corresponding to 1-2 and 3-4 sections respectively): ks-m - Chersonian-Meotian epiclastic cover, BMF -
Bârnova-Muncelu Formation, ŞF - Şcheia Formation (PM - Pietrăria Member, MM - Muncelu Member ), HF – Huşi
Formation, NRC - Nuţaşca-Ruseni Cinerites, SB - sandy body with reduced intercalations of siltic sands and clayey
sands, SiSB - sandy silt, CaSB - sandy with lumashelic limestone, CSiB
- clayey silt body, with transitions towards
siltic clays, MCSiB - mixed silt clayey body, with transitions from clayey silts to siltic clays and siltic mud,
CMB - clayey mud body, MMB - mixed mud body with transitions towards siltic mud and clayey mud, OC
- oolithic limestones, lumashelic limestone, with sandstones and sands; C - the granulometric projection sedi-
mentary bodies of lithological column A and D - of lithological column B (according to BOGGS, 2009; grain
size distribution % according to STAS 1913-5-85): 1 - clay (claystone) (C); 2 - silt (siltstone) (Si); 3 - sand
(sandstone) (Sa); 4 - sandy clay (sandy claystone) (SaC); 5 - silty clay (silty claystone) (SiC); 6 - clayey silt
(clayey silstone) (CSi); 7 - sandy silt (sandy silstone) (SaSi); 8 - silty sand (silty sandstone) (SiSa); 9 - clayey
sand (clayey sandstone) (CSa); 10 - clayey mud (clayey mudstone) (CM); 11 - silty mud (silty mudstone)
(SiM); 12 - sandy mud (sandy mudstone) (SaM)
background image
LARGE LANDSLIDE RISK ASSESSMENT IN HILLY AREAS. A CASE STUDY OF HUŞI TOWN REGION (NORTH-EAST OF ROMANIA)
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
281
georeferenced environment provided by TNTMips
7.3, ArcGIG 9.3 and SAGA 2.0.8 software packages
while the statistical analysis was performed using
Excel 2003 and XLSTAT 2010 trial version.
RESULTS AND DISCUSSIONS
The logistic regression generated the landslide
susceptibility map, as a graphical output material-
ized by continuous values between zero and one.
Fig. 5 shows classified landslide susceptibility map
performed on the basis of natural breaks method
(Jenks), which identifies the class breaks that best
group similar values and maximizes the differences
between classes. According to the standardized re-
gression coefficients (Tab. 1), landslides occurrence
is best explained by slope inclination, land use and
slope aspect classes, similar outcomes with other ob-
tained in Moldavian Plateau, using the same method
(G
ROZAVU
et alii, 2010, 2012; M
ĂRGĂRINT
et alii,
2011). Secondary positions are occupied by distance
to drainage network, slope height and plan curvature.
The influence of mean annual precipitations is less
significant (error probability equals 0.07) and more
uncertain, since the upper bound of the standard-
ized coefficient is positive, while the lower bound
and the coefficient in itself are negative (Tab. 1). The
stepwise procedure of the logistic regression method
eliminated 3 variables from the analysis: terrain alti-
tude, mean and profile curvature.
The model validation was realized through the
ROC (Receiver Operating Characteristic) analysis, a
very useful method for evaluating the predictive accu-
racy of the logistic regression model (G
ORSEVSKI
et alii,
2006). For the cut-off value of 0.5, the area under the
ROC curve has the value of 0.891, which means a high
degree of accuracy (Fig. 6). In addition, 84.6% of the
landslide area was correctly classified by the logistic re-
gression model, the overall accuracy being of 81.64%.
Over the landslide susceptibility map, there were
limited dataset (V
an
D
en
e
eCkhauT
et alii, 2012) and
gives possibility to evaluate predictive accuracy (G
OR-
SEVSKI
et alii, 2006). In literature, this method was con-
sidered even the most useful for landslide susceptibil-
ity assessment at regional scale (O
HLMACHER
& D
AVIS
,
2003; C
HAU
& C
HAN,
2005).
Logistic regression links the presence or the ab-
sence of landslides to a set of quantitative or qualita-
tive variables, generating a continuous spatial prob-
ability model:
(which varies from 0 to 1 on a shaped curve), com-
puted on the basis of a linear combination (z) of inde-
pendent variables (x
1
, x
2
… x
n
): z = b
0
+ b
1
x
1
+ b
2
x
2
+
… + b
n
x
n
, where b
o
is the intercept of the model and
b
1
, b
2
… bn are the regression coefficients.
In order to extract predictors’ values from a
raster layer, a total number of 2483 equally distanced
grid points were generated for the landslide and
landslide-free areas. To preserve the relative equality
of the two points’ samples, required by the nature
of the statistical analysis, the density of points in-
side the landslide area is markedly higher than in the
landslide-free area (G
ROZAVU
et alii, 2010).
Landslide causative factor database consists of
several environmental data layers. As qualitative
variable was considered the land use (Fig. 3) and as
quantitative continuous variables were considered
topographical parameters (terrain altitude, slope an-
gle, mean curvature, plan curvature, profile curvature
and slope height, obtained from the digital elevation
model - DEM - with 10 x 10 m resolution, derived
from 1:5,000 Romanian topographic plans in stereo-
graphic projection), distance to the drainage network
and mean annual precipitations. The classified slope
aspect (horizontal, N, E, S, W) and land use were
converted into quantitative variables by computing
the respective landslide densities.
To obtain the elements exposed to landslide risk,
we used the General Urban Plan (GUP), at 1:5,000
scale, 1996 edition, provided by local administra-
tion. The reconstitution of Huşi town extending dur-
ing the 1920s was made on the basis of the Shooting
Directory Plans in the Lambert-Cholenski conic pro-
jection at 1:20,000 scale (C
RĂCIUNESCU
, 2010).
The data integration was accomplished in the
Tab. 1 - Standard regression coefficients of the logistic re-
gression model
background image
M.C. MARGARINT, D.T. JURAVLE, A. GROZAVU, C.V. PATRICHE, M. POHRIB & I.C.STÂNGĂ
282
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
overlapped the exposed elements, initially acquired
in vector format and subsequently converted in raster
format: the intravilan area (according to the General
Urban Plan for 1996 and 2012), the area occupied
with buildings and gardens in 1920 (according to
the Shooting Directory Plans), the current built area
(based on the orthorectified aerial image of 2010).
Also, several linear vector elements were added and
analyzed: roads, railways, aerial and underground
electrical network, the city ring road project (Tab. 2)
Analyzing the exposed elements and their weight
on susceptibility classes reveals that the town extended
continuously on lands with an increasingly higher risk
to landslides. While in 1920, only 3.4% of the town cor-
responded to the classes with high and very high land-
slide susceptibility, in 1996, the summed weight of the
two classes rises to 11.2% and reaches today 13%. It
can be clearly noted that only the very low susceptibil-
ity class recorded decreasing percentage, while all the
other four classes had an ascending trend.
Tab. 2 - Elements exposed to the landslide risk in Huşi
town. The weight within the susceptibility classes
Fig. 6 - ROC curve and area under de curve value
Fig. 5 - Landslide susceptibility classes (Jenks method) in Huşi town area: 1 - very low; 2 - low; 3 - medium; 4 - high and
5 - very high class. 6 - landslides bodies. 7 - Huşi town
background image
LARGE LANDSLIDE RISK ASSESSMENT IN HILLY AREAS. A CASE STUDY OF HUŞI TOWN REGION (NORTH-EAST OF ROMANIA)
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
283
ard gauge in 1940, the line does not benefited from ma-
jor investment and it follows sinuously the fragmented
land with radii up to 120 m and slope gradients of 30
m/km. Along this line, the classes with high and very
high landslide susceptibility totalize 59% (Fig. 7).
Much more relevant are the landslide susceptibil-
ity values for the lands along the future ring road that
should deflect the heavy traffic (Fig. 7). This foreseen
ring road would insinuate along the southern slope
of the city, dominated by large landslides: our results
show that 58% of the projected road pass through ar-
eas that fall within the classes of high and very high
susceptibility. Therefore, we consider that the route of
this objective to be realized must be carefully rede-
signed and moved on the northern part of the city. This
is the appropriate solution in long term, regardless the
higher momentary economic costs of lands.
CONCLUSIONS
The example of the Huşi town reveals once more
that the growing trend of landslide risk is largely re-
lated to the expansion of human structures (peoples,
dwellings and utilities, road or special networks, other
social and economic objectives) in areas that may be
affected by the landslides.
Assessing the probability of landslide occurrence
in some places (landslide susceptibility) and, implic-
itly, identifying the exposed elements that could be af-
Between 1920 and 1996, the town extended both
to East and West, through districts with industrial and
transport function, on lands with low susceptibility
and to the southern slope with high susceptibility
(Fig. 7). There have been built here especially indi-
vidual houses at a time when, generally, there was
no legal interdiction (the legislation vacuum during
the 1990s years). Also here, in the same period, there
have been established some winemaking units. In
the eastern extremity of the town, on lands with high
susceptibility to landslides, there have been designed
recreational areas, but without important technical
urban facilities.
After 1996, the intravilan extending was regu-
lated by adopting in 1999 the General Urban Plan.
The layer with the built area in 2010 (including both
the construction itself and the courtyards) reveals
that only 4.6% of them overlap the classes of high
and very high susceptibility.
In the case of linear elements, there are some
contrasting situations. The less exposed are the elec-
trical networks (51% of the 20kV underground lines,
in the central town) and the roads (regardless the
level of importance or modernization).
A special situation is the one of the railway that
connects Huşi to the national railway system. It was
originally constructed with narrow gauge of 1000 m
and became operational in 1890. Extended to a stand-
Fig. 7 - Territorial evolution of Huşi town and some exposed elements to landslide risk, overlapped to the landslide suscep-
tibility classes. (A - town limit in 1920; B - town limit in 1996; C - town limit in 2012; D - railways; E - ring road
project. Susceptibility classes: 1 - very low; 2 - low; 3 - medium; 4 - high; 5 - very high)
background image
M.C. MARGARINT, D.T. JURAVLE, A. GROZAVU, C.V. PATRICHE, M. POHRIB & I.C.STÂNGĂ
284
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
fected are mandatory phase for an adequate landslide
risk management.
Identifying the significant factors that favor or trig-
ger the sliding processes is the most important for a cor-
rect assessment of landslide susceptibility. In this con-
text, the logistic regression proves to be a useful method
and a tool that guarantee the objectivity of the evaluation.
The identification of susceptible areas and of the
exposed elements can serve to decision makers for a
sustainable territory planning. This current regional
approach must be continued and completed at a much
higher detail level, including with the estimation of
the potential socio-economic and environmental costs.
REFERENCES
a
rMaş
i. (2011) -
An analytic multicriteria hierarchical approach to assess landslide vulnerability.
Case study: Cornu
village, Subcarpathian Prahova Valley/Romania. Zeitschrift für Geomorphologie, 55: 209-229.
a
rMaş
i. (2012) - Weights of evidence method for landslide susceptibility mapping. Prahova Subcarpathians. Romania. Natural
Hazards, 60 (3): 937-950.
a
kGün
A. (2012) - A comparison of landslide susceptibility maps produced by logistic regression, multi-criteria decision, and
likelihood ratio methods: a case study at Izmir, Turkey. Landslides, 9: 93-106.
a
Tkinson
p.M. & M
assari
r. (1998) - Generalised linear modeling of susceptibility to landsliding in the Central Apennines,
Italy. Computers & Geosciences, 24: 373-385.
a
yaLew
L. & y
aMaGishi
h. (2005) - The application of GIS-based logistic regression for landslide susceptibility mapping in the
Kakuda-Yahiko Mountains, Central Japan. Geomorphology, 65: 15-31.
B
ăLTeanu
D., C
henDeş
V., s
iMa
M. & e
nCiu
p. (2010) - A country-wide spatial assessment of landslide susceptibility in Romania.
Geomorphology, 124 (3-4): 102-112.
B
eLL
r. & G
LaDe
T. (2004) -
Quantitative risk analysis for landslides
- examples from Bíldudalur, NW-Iceland. Natural
Hazards and Earth System Sciences, 4: 117-131.
B
oGosLoVsky
V.a. & o
GiLVy
a.a. (1977) - Geophysical methods for the investigation of landslides. Geophysics, 42 (3): 562-571.
B
oGGs
S. JR. (2009) - Petrology of sedimentary rocks. 2
nd
Ed., Cambridge University Press, 600 pp.
B
renninG
A. (2005) - Spatial prediction models for landslide hazards: review, comparison and evaluation. Natural Hazards and
Earth System Sciences, 5: 853-862.
C
arDinaLi
M., r
eiChenBaCh
p., G
uZZeTTi
F., a
rDiZZone
F., a
nTonini
G., G
aLLi
M., C
aCCiano
M., C
asTeLLani
M. & s
aLVaTi
p.
(2002) - A geomorphological approach to the estimation landslide hazards and risks in Umbria, Central Italy. Natural
Hazards and Earth System Science, 2: 57-72.
C
hau
k.T. & C
han
J.e. (2005) - Regional bias of landslide data in generating susceptibility maps using logistic regression: Case
of Hong Kong Island. Landslides, 2: 280-290.
C
hauhan
s., s
harMa
M. & a
rora
M.k. (2010) - Landslide susceptibility zonation of the Chamoli region, Garhwal Himalayas,
using logistic regression model. Landslides, 7: 411-423.
C
răCiunesCu
V. (2010) - eMap: Shooting Directory Plans. http://earth.unibuc.ro/articole/. (in Romanian)
C
roZier
M.J. & G
LaDe
T. (2005) - Landslide Hazard and Risk: Issues, Concepts and Approach. In: G
LaDe
T., a
nDerson
M.G. &
C
roZier
M.J. (e
Ds
.) - Landslide Hazard and Risk. Wiley, 1-40.
D
ai
F.C. & L
ee
C.F. (2002) - Landslide characteristics and slope instability modeling using GIS, Lantau Island, Hong Kong.
Geomorphology, 42: 213-228.
D
ai
F.C., L
ee
C.F. & n
Gai
y.y. (2002) - Landslide risk assessment and management: an overview. Engineering Geology, 64 (1),
64-87.
F
aLasChi
F., G
iaCoMeLLi
F., F
eDeriCi
p.r., p
uCCineLLi
a., D’a
MaTo
a
VanZi
G., p
oChini
a. & r
iBoLini
a. (2009) - Logistic
regression versus artificial neural networks: landslide susceptibility evaluation in a sample area of the Serchio River valley,
Italy
. Natural Hazards, 50: 551-569.
F
eLL
r., C
oroMinas
J., B
onnarD
C., C
asCini
L., L
eroi
e. & s
aVaGe
w.Z. (2008) - Guidelines for landslide susceptibility, hazard
and risk zoning for land-use planning. Engineering Geology, 102: 99-111.
G
aLLipoLi
M., L
apenna
V., L
orenZo
p., M
uCCiareLLi
M., p
errone
a., p
isCiTeLLi
s. & s
Dao
F. (2000) - Comparison of geological
and geophysical prospecting techniques in the study of a landslide in southern Italy. European Journal of Environmental
background image
LARGE LANDSLIDE RISK ASSESSMENT IN HILLY AREAS. A CASE STUDY OF HUŞI TOWN REGION (NORTH-EAST OF ROMANIA)
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
285
and Engineering Geophysics, 4: 117-128.
G
hosh
s., C
arranZa
e.J.M., V
an
w
esTen
C.J., J
eTTen
V.G. & B
haTTaCharya
D.n. (2011) - Selecting and weighting spatial
predictors for empirical modeling of landslide susceptibility in the Darjeeling Himalayas (India). Geomorphology, 131: 35-56.
G
LaDe
T. (2003) - Vulnerability assessment in landslide risk analysis. Die Erde, 134: 121-138.
G
LaDe
T., a
nDerson
M.G. & C
roZier
M.J. (2005, Eds.) - Landslide hazard and risk. Wiley. 824 pp.
G
orseVski
p.V., G
essLes
p. & F
oLTZ
r.B. (2000) - Spatial Prediction of Landslide Hazard Using Logistic Regression and GIS.
Proceedings of the 4
th
International Conference on Integrating GIS and Environmental Modeling (GIS/EM4): Problems,
Prospects and Research Needs. Banff, Alberta, Canada, September, 2000.
G
orseVski
P.V., G
essLes
P., F
oLTZ
R.B. & e
LLioT
W.J. (2006) - Spatial prediction of landslide hazard using logistic regression
and ROC analysis. Transaction in GIS, 10: 395-415.
G
roZaVu
a, M
ărGărinT
M.C. & p
aTriChe
C.V. (2010) - GIS applications for landslide susceptibility assessment: a case study
in Iaşi County (Moldavian Plateau, Romania). In: B
reBBia
C.A. (e
D
.) - Risk Analysis VII & Brownfields V, WIT Press:
393-404.
G
roZaVu
a, M
ărGărinT
M.C. & p
aTriChe
C.V. (2012) - Landslide susceptibility assessment in the Brăieşti-Sineşti sector of Iaşi
Cuesta. Carpathian Journal of Earth and Environmental Sciences, 7 (1): 39-46.
G
uGiuMan
I. (1959) - Huşi Depression. Study of physical and economic geography. Scientific Press, Bucharest, 218. (in
Romanian)
G
ünTher
a., r
eiChenBaCh
p., M
aLeT
J.-p., V
an
D
en
e
eCkhauT
M., h
erVás
J., D
ashwooD
C. & G
uZZeTTi
F. (2012) -
Tier-based
approaches for landslide susceptibility assessment in Europe
.
Landslides, DOI 10.1007/s10346-012-0349-1.
G
uZZeTTi
F. (2000) - Landslide fatalities and the evaluation of landslide risk in Italy. Engineering Geology, 58 (2): 89-107.
G
uZZeTTi
F., r
eiChenBaCh
p., C
arDinaLi
M., G
aLLi
M. & a
rDiZZone
F. (2005) - Probabilistic landslide hazard assessment at the
basin scale. Geomorphology, 72: 272-299.
G
uZZeTTi
F., r
eiChenBaCh
P., a
rDiZZone
F., C
arDinaLi
M. & G
aLLi
M. (2006) - Estimating the quality of landslide susceptibility
models. Geomorphology, 81 (1-2): 166-184.
G
uZZeTTi
F., a
rDiZZone
F., C
arDinaLi
M., G
aLLi
M., r
eiChenBaCh
P. & r
ossi
M. (2008) - Distribution of landslides in the Upper
Tiber River basin, central Italy. Geomorphology, 96 (1-2): 105-122.
G
uZZeTTi
F., M
onDini
a.C., C
arDinaLi
, F
ioruCCi
F., s
anTanGeLo
M. & C
hanG
K.-T. (2012) - Landslide inventory maps: new tools
for an old problem. Earth-Science Reviews, 112: 42-66.
h
aque
C.e. & B
urTon
i. (2005) - Adaptation options strategies for hazards and vulnerability mitigation: an international
perspective. Mitigations and Adaptation Strategies for Global Changes, 10: 335-353.
i
onesi
B. & i
onesi
L. (1994) - Limite Bessarabien-Chersonien sur les unités de plate-forme de la Roumanie. The Miocene from
the Transylvanian Basin-Romania. Cluj-Napoca, 187-192. (in French)
i
onesi
L., i
onesi
B., r
oşCa
V., L
unGu
a. & i
onesi
V. (2005) - Middle and Upper Sarmatian on Moldavian Platform. Romanian
Academy Press, Bucharest, 558. (in Romanian)
i
riMia
L
., p
aTriChe
C.V. & q
uenoL
h. (2011) - GIS applications in viticulture. The spatial distribution of climatic suitability for
grape growing in Huşi wine-growing centre (Romania). Agricultural Research in Moldova, Iaşi, Romania, 44(3): 51-59.
J
eanrenauD
P. (1971) - Geologic map of Central Moldavia between Siret and Prut valleys. Scientific Annals of Alexandru Ioan
Cuza University of Iaşi, Geological Section, 17: 65-78. (in Romanian)
J
onGMans
D. & G
araMBois
s. (2007) - Geophysical investigation of landslides: a review. Bulletin of Geological Society of
France, 178(2): 101-112.
k
aynia
a.M., p
apaThoMa
-k
öhLe
M., n
euhäuser
B., r
aTZinGer
k., w
enZeL
h. & M
eDina
-C
eTina
Z. (2008) - Probabilistic
assessment of vulnerability to landslide: application to the village of Lichtenstein, Baden-Württemberg, Germany.
Engineering Geology, 101 (1-2): 33-48.
k
o
k
o
C., F
LenTJe
p. & C
howDhury
r. (2003) - Quantitative landslide hazard and risk assessment: a case study. Quarterly
Journal of Engineering Geology and Hydrogeology, 36 (3): 261-272.
M
aThew
J., J
ha
V.k. & r
awaT
G.S. (2007) - Application of binary logistic regression analysis and its validation for landslide
susceptibility mapping in part of Garhwal Himalaya, India. International Journal of Remote Sensing, 28 (10): 2257-2275.
M
ărGărinT
M.C, G
roZaVu
a., C
onDoraChi
D, p
LeşCan
s. & B
oaMFă
i. (2010) - Geomorphometric features of the built areas of
background image
M.C. MARGARINT, D.T. JURAVLE, A. GROZAVU, C.V. PATRICHE, M. POHRIB & I.C.STÂNGĂ
286
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
the localities along Iaşi Cuesta, Geographia Technica, 10: 79-89.
M
ărGărinT
M.C., G
roZaVu
A., p
aTriChe
C.V., T
oMaşCiuC
A
.-M.I., u
rDea
R. & u
nGurianu
I. (2011) - Évaluation des risques
de glissements de terrain par la méthode de la régression logistique: application à deux zones basses de Roumanie.
Dynamiques Environnementales, 28: 41-50, Université Michel de Montaigne, Bordeaux 3. (in French)
M
eusBurGer
k. & a
LeweLL
C. (2009) - On the influence of temporal change on the validity of landslide susceptibility maps.
Natural Hazards and Earth System Sciences, 9: 1495-1507.
M
iCu
M. & B
ăLTeanu
D. (2009) - Landslide hazard assessment in the Curvature Carpathians and Subcarpathians, Romania.
Zeitschrift für Geomorphologie. Supplementbände, 53 (2): 31-47.
M
onTGoMery
D.r. & D
ieTriCh
w.e. (1988) - Where do channels begin? Nature, 336: 232-234.
n
aTionaL
C
ounCiL
For
s
CienCe
anD
T
eChnoLoGy
- Romanian Standardization Institute - Classification and identification of
earths. STAS 1243-83. (in Romanian).
n
aTionaL
C
ounCiL
For
s
CienCe
anD
T
eChnoLoGy
- Romanian Standardization Institute - Determination of grain size distribution.
STAS 1913-5-85. (in Romanian).
n
iCoriCi
C., G
ray
J.,
i
MBroane
A.M. & B
arBosu
M. (2012) - GIS susceptibility maps for shallow landslides: a case study in
Transylvania, Romania. Carpathian Journal of Earth and Environmental Sciences, 7 (2): 83-92.
o
hLMaCher
G.C. & D
aVis
J.C. (2003) - Using multiple logistic regression and GIS technology to predict landslide hazard in
northeast Kansas, USA. Engineering Geology, 69: 331-343.
p
asCaLe
s., s
Dao
F. & s
oLe
A. (2010) - A model for assessing the systemic vulnerability in landslide prone areas. Natural
Hazards and Earth System Sciences, 10: 1575-1590.
p
ohriB
M.D., J
uraVLe
D.T., n
iaCşu
L., u
rsu
a., s
TanCiu
a. & p
LăTiCă
D. (2012) - Paleogeography of the Chersonian to
Meotian in the north of Fălciu Hills (Moldavian Platform) based of sedimentological data. Carpathian Journal of Earth and
Environmental Sciences, 7 (2): 23-36.
p
uJină
D. (2003) - Risk to landslides of slopes in Moldavian Plateau. 17
th
National Conference of Soil Science, Timişoara,
Romania. (in Romanian)
s
ChusTer
r.L. & h
iGhLanD
L.M. (2007) - The Third Hans Cloos Lecture. Urban landslides: socioeconomic impacts and overview
of mitigative strategies. Bulletin of Engineering Geology and the Environment, 66: 1-27.
s
MiTh
k. & p
eTLey
D.n. (2009, e
Ds
.) - Environmental hazards. Assessing risk and reducing disaster. 5
th
Ed., Routledge, Taylor
and Francis Group, London and New York, 383.
s
TÂnGă
I.C. (2012) - Tutova Basin. Natural risks and territory vulnerability. Alexandru Ioan Cuza University Press, Iaşi,
Romania, 290. (in Romanian)
s
TÂnGă
I.C. & G
roZaVu
A. (2012) - Quantifying human vulnerability in rural areas: case study of Tutova Hills (Eastern
Romania). Natural Hazards and Earth System Sciences, 12: 1987-2001.
ş
anDriC
i., C
hiţu
Z., M
ihai
B. & s
ăVuLesCu
I. (2011) - Landslide susceptibility for the administrative area of Breaza, Prahova
County, Curvature Subcarpathians, Romania. Journal of Maps, 2011: 552-563.
u
ZieLLi
M., n
aDiM
F., L
aCasse
s. & k
aynia
a. (2008) - A conceptual framework for quantitative estimation of physical
vulnerability to landslides. Engineering Geology, 102(3-4): 251-256.
V
an
D
en
e
eCkhauT
M., h
erVás
J., J
aeDiCke
C., M
aLeT
J.-p., M
onTanareLLa
L. & n
aDiM
F. (2012) - Statistical modelling of
Europe-wide landslide susceptibility using limited landslide inventory data. Landslides, 9: 357-369.
V
an
w
esTen
C.J., V
an
a
sCh
T.w.J. & s
oeTers
r. (2006) - Landslide hazard and risk zonation. Why is it still so difficult? Bulletin
of Engineering Geology and the Environment, 65: 167-184.
V
arnes
D.J. (1984) - Landslide hazard zonation: a review of principles and practice. United Nations International, Paris.
y
esiLnaCar
e. & T
opaL
T. (2005) - Landslide susceptibility mapping: a comparison of logistic regression and neural networks
methods in a medium scale study, Hendek region (Turkey). Engineering Geology, 79: 251-266.
Statistics