Document Actions

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

background image
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
541
DOI: 10.4408/IJEGE.2013-06.B-52
HYDROGEOLOGICAL SPRING CHARACTERIZATION
IN THE VAJONT AREA
P
aolo
FABBRI
(*)(**)
, M
irta
ORTOMBINA
(*)
, l
eonardo
PICCININI
(*)
,
d
ario
ZAMPIERI
(*)
& l
uca
ZINI
(***)
(*)
Università degli Studi di Padova - Dipartimento di Geoscienze - Via G. Gradenigo, 6 - Padova, Italy
(**)
Consiglio Nazionale delle Ricerche (C.N.R.) - Istituto di Geoscienze e Georisorse - Padova, Italy
(***)
Università degli Studi di Trieste - Dipartimento di Matematica e Geoscienze - Via Weiss, 2 - Trieste, Italy
Gallina Valley on the south, the Piave River on the west
and the Zemola and Mesazzo valleys on the east (Fig. 1).
The landscape of Vajont valley was severely al-
tered by the catastrophe of 9 October 1963, when a
vast deposit of the Mt. Toc landslide completely blan-
keted the bottom of the valley producing an enormous
wave of at least 30 million cubic metres of water
which flooded the Piave Valley below.
Despite a large body of geological and geome-
chanical literature mainly concerning the Vajont
disaster (M
uller
, 1964, 1968, 1987; S
eMenza
, 1965;
c
orbin
, 1982; H
endron
& P
atton
, 1985; b
elloni
&
S
tefani
, 1987; n
onveiller
, 1987; G
Hirotti
, 1994;
t
ika
& H
utcHinSon
, 1999; S
eMenza
& G
Hirotti
,
2000; k
ilburn
& P
etley
, 2003; M
antovani
& v
ita
-
f
inzi
, 2003; H
elMStetter
et alii, 2004; G
enevoiS
&
G
Hirotti
, 2005; v
eveakiS
et alii, 2007; W
ard
& d
ay
,
2011), hydrogeological investigations are few (b
e
-
Sio
, 1986). This is most likely due to the relatively
shallow and poor groundwater circulation before the
disaster, currently shown by the absence of springs
in the Mt. Toc area and by the very low spring flow
rates in the Vajont Valley.
The bedrock of the study area belongs to the east-
ern South Alpine structural unit, which represents the
Neogene-Present back-thrusted (south vergent) part of
the Alpine chain. The Vajont Valley coincides with the
core of an Alpine syncline (Erto syncline) with an axis
trending approximately E-W and gently plunging to
the east (r
iva
et alii, 1990). The exposed deformed
rocks are Liassic to Eocene carbonates and marls (S
e
-
ABSTRACT
The Vajont Valley is mainly known for the cata-
strophic event of 9 October 1963, when a vast landslide
occurred on the southern slope of the Vajont dam reser-
voir, causing a giant wave of water that flooded the Pi-
ave Valley below. Since then, many studies of the geo-
logical and geomechanical aspect of the landslide have
been carried out, while very few studies have focused
on the hydrogeological characteristics of this area. This
paper proposes a hydrogeological conceptual model for
the carbonate aquifers of the Vajont area, based on the
continuous monitoring of two springs and environmen-
tal isotope investigations. Cross-correlation functions
between time series and the VESPA index were used to
delineate groundwater flow systems and the degree of
karstification. This model has been confirmed by analy-
ses of the amounts of stable isotopes, such as
18
O and
3
H, in precipitation and groundwater.
K
ey
words
: Vajont, spring, continuous monitoring, cross-
correlation function, VESPA index, environmental isotopes
GEOLOGICAL AND HYDROGEOLOGI-
CAL SETTINGS
This paper focuses on the hydrogeological charac-
teristics of the Vajont area. The study site is located in the
municipalities of Erto-Casso (Friuli Venezia Giulia re-
gion) and Longarone (Veneto region) in the southeastern
part of the Dolomite Mountains, northeastern Italy. This
area is bounded by Mt. Salta (2,039 m above sea level
[a.s.l.]) and Mt. Borgà (2,215 m a.s.l.) on the north, the
background image
P. FABBRI, M. ORTOMBINA, L. PICCININI, D. ZAMPIERI & L. ZINI
542
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
anca and Col Tramontin faults on the east and the west
branch of the Col delle Erghene fault on the west), while
the landslide crown was constrained by E-W-trending
structures (Col delle Erghene fault; r
iva
et alii, 1990).
The succession of strata and their relative perme-
abilities are shown in Tab. 1.
From a hydrogeological point of view, only the
Zemola Valley contains several significant springs,
whereas the area around Mt. Toc is characterized by less
surface water and fewer springs, most of which display
low levels of discharge. This situation, we believe, is
due to karstic groundwater circulation. This hypothesis
is based on direct observation of sinkholes on the upper
slope of the Vajont landslide; in this area, the majority of
meteoric waters infiltrate without leading to a significant
surface flow. The presence of karst phenomena in this
Menza
, 1965); the Upper Triassic Dolomia Principale
crops out only in the neighboring valleys. The north-
ern limb of the Erto syncline is reversed and stretched,
lying just at the foot of the Mt. Borgà thrust (Mt. Salta
thrust in r
iva
et alii, 1990) with its Spesse thrust
splay, which consists of two older thrusts passively
transported in reverse style by the Belluno thrust.
The Vajont landslide occurred on the Erto syncline's
south limb, which dips 30° to 50° to the north-northeast
(western sector) and north (eastern sector). The land-
slide rupture surface is localized within the Middle-
Upper Jurassic Fonzaso Formation, a sequence of thinly
stratified limestones with thin (0.1-5 cm) intercalations
of clays (H
endron
& P
atton
, 1985). The sliding lay-
ered sequence was laterally constrained by a system of
southward-converging subvertical faults (the Croda Bi-
Fig. 1 - Geographic setting of the Vajont area (coordinates and elevations are in meters, datum Roma 1940, Gauss Boaga - Est)
background image
HYDROGEOLOGICAL SPRING CHARACTERIZATION IN THE VAJONT AREA
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
543
210 mg/l), cold (6.7 < T < 10°C) and calcium magnesi-
um bicarbonate-type waters; furthermore, all the springs
have a slightly alkaline pH ranging between 8 and 8.4,
Eh ranging from 182 to 274 mV and electrical conduc-
tivity ranging from 162 to 281 μS/cm. These physico-
chemical data suggest the presence of young waters with
short travel paths. The locations of five of these springs
are shown in Fig. 2, and their corresponding Schoeller
diagrams are plotted in Fig. 3.
After a reconnaissance of the geology of the study
area is well reported in the geologic literature (H
endron
& P
atton
, 1985; S
eMenza
& G
Hirotti
, 2000).
In the study area, twenty springs were identified.
Most of them have very low discharge (< 1 L/s) and
similar physico-chemical parameters. Chemical analy-
ses were carried out on water from eight springs (so-
dium, calcium, potassium, magnesium, total hardness,
total dissolved solids [TDS], iron, manganese, methyl
orange alkalinity, chloride, sulphate, carbonate), and
they could be classified as oligomineral (135 < TDS <
Fig. 2 - Geological sketch map of the study area (partly from R
iva
et alii, 1990); red dots are the monitoring springs
Tab. 1 - Main lithologic units
of the Vajont area and
their relative perme-
ability (units symbols
correspond to label in
the Fig. 7)
background image
P. FABBRI, M. ORTOMBINA, L. PICCININI, D. ZAMPIERI & L. ZINI
544
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
slope of Mt.. Toc, on the east side of the Piave Valley,
at an elevation of 515 m a.s.l., at Dogna (municipality of
Longarone). The Vajont Limestone (high permeability)
underlies the Ega Nass spring recharge area. The Vajont
Limestone overlies the Igne Formation (very low perme-
ability). The spring emerges at the exposed contact be-
tween these two units. The Vajont Limestone is charac-
terized by a high degree of karstification. The location of
this spring may also be controlled by a system of faults
parallel and orthogonal to the slope (Figs. 6b and 7).
The measured discharge of the Ega Nass spring
does not represent the total flow emerging because
spring water emerges from several points beyond our
monitoring station at the main discharge location.
DATA MONITORING ANALYSIS
LE SPESSE
An analysis of Figure 8 shows a relationship be-
tween the rainfall (P) and the spring flow rate (Q). More-
over a cross-correlation investigation has helped us to
quantify the lag times over short time periods. Lagged
correlation refers to the correlation between two time
series at different lags, or offsets in time. The cross-cor-
relation function (CCF) is the correlation, related to the
two time series, as a function of lag. The sample cross
area, two springs with appropriate features were chosen
for continuous monitoring; Ega Nass spring, on the west
side of Mt. Toc (Dogna, Longarone), and Le Spesse
spring, on the opposite side of the Vajont landslide (Le
Spesse, Erto-Casso). In July and September 2010, two
data-loggers (Diver) were installed to measure the hourly
discharge, electrical conductivity (EC) and temperature
(T) of the spring waters. These data were compared with
rainfall from two weather stations (ARPAV and Servizio
Idrografico Regione Friuli Venezia Giulia) to identify
correlations between spring behavior (discharge, tem-
perature and electrical conductivity) and rainfall events.
HYDROGEOLOGICAL BEHAVIOR OF LE
SPESSE AND EGA NASS SPRINGS
LE SPESSE SPRING
Le Spesse spring (Figs. 2 and 4) is situated at
an elevation of approximately 810 m a.s.l. close to
a small hamlet in the municipality of Erto Casso
(Pordenone). This spring is located west of Erto
Casso, on the north side of the Vajont Valley and
east of the Vajont dam, at the foot of the southern
slope of Mt. Borgà. The spring emerges from the
Quaternary deposits; however, after evaluating the
cross section (Fig. 7), in particular the formations
next to the spring, it is evident that this spring water
originates in the Vajont Limestone (high permeabil-
ity), which is tectonically juxtaposed against the
Scaglia Rossa Formation (low permeability), and
flows through and emerges from the Quaternary
deposits (Fig. 6a). According to the hydrogeologi-
cal classification proposed by c
ivita
(1972), which
takes into account the point of emergence, this
spring can be classified as a spring of contact be-
tween units of highly contrasting permeability.
EGA NASS SPRING
Ega Nass spring (Fig. 5) is located on the west
Fig. 5 - Ega Nass spring
Fig. 3 - Schoeller diagrams of five springs
Fig. 4 - Le Spesse spring
background image
HYDROGEOLOGICAL SPRING CHARACTERIZATION IN THE VAJONT AREA
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
545
where N is the series length,
ū
and
y
are the sample
means and k is the lag. The sample cross-correlation is
the cross-covariance of the two series divided by their
standard deviations:
The cross-correlation function (CCF; P vs Q) in
Figure 9 shows a significantly positive correlation at
lag 1, which means that the most important response in
flow rate (output) occurs beginning 1 day after a rainfall
event (input). Moreover, a less evident positive cross-
correlation at lag 0 also indicates a certain temporal
association between rainfall and the spring flow rate.
In a situation of a quick flow-rate response to rain-
fall events, it can be useful to compare spring water
temperature (Ts) with atmospheric temperature (Ta)
to ascertain whether the flow-rate increase is due to a
“piston” effect or to a local rainfall infiltration.
covariance function is given by the following equation:
Fig. 8 - Comparison between daily rainfall (light blue his-
togram) and Le Spesse flow rate (red line) from 15
July 2010 to 4 June 2012
Fig. 6 - H y d ro g e o l o g i -
cal sketch map of
the Le Spesse (a)
and Ega Nass (b)
springs. The eleva-
tions of the topo-
graphic contour
lines are in meters
Fig. 7 - Hydrogeological cross sections
through Le Spesse and Ega Nass
springs. Cross section line locations
are shown in Figures 6a and 6b, re-
spectively. Symbols representing geo-
logic units are explained in Table 1
background image
P. FABBRI, M. ORTOMBINA, L. PICCININI, D. ZAMPIERI & L. ZINI
546
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
Analysis of Figure 10a shows a positive cross-
correlation between variations in Ta and variations in
Ts at lag 0 and lag 1, indicating that the water tem-
perature variations in the short term are strictly linked
to the atmospheric temperature variations. Moreover
the temperature trend comparison in Figure 10b shows
a reverse trend: in winter the Ta is lower and the Ts
is higher, while the reverse was observed in summer.
During the warm season, the temperature of spring
water is colder than that in the cold season, meaning
that the water discharge is in partial equilibrium with
the reservoir temperature instead of with the atmos-
pheric temperature. Evidently, the travel time through
the aquifer is insufficient for water to reach a thermal
equilibrium with the rock matrix: in the winter, warm
water discharges represent infiltration from the previ-
ous summer, and in the summer, outflowing cold water
represents infiltration from the previous winter. Only
Fig. 9 - Cross-correlation of rainfall and spring flow rates
Fig. 10 - (a) Cross-correlation between Ta and Ts; (b) com-
parison between atmospheric temperature (black
line) and water spring temperature (red line)
Fig. 11 - (a) Cross-correlation between spring flow rate and electrical conductivity; (b) comparison between spring flow rate
(red line) and electrical conductivity (black line)
background image
HYDROGEOLOGICAL SPRING CHARACTERIZATION IN THE VAJONT AREA
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
547
during rainy days (short time periods) are
Ts variations influenced by variations in
atmospheric temperatures, due to the ef-
fects of local rainfall infiltration, as seen
in the cross-correlations in Figure 10a.
A comparison between the spring flow
rate (Q) and electrical conductivity (EC;
Fig. 11) shows a negative cross correla-
tion at lag 0 (e.g., when Q increases then
the EC decreases and vice versa). Thus
the relationship shown in Figure 11 con-
firms the effect of local rainfall infiltration,
which temporarily decreases the EC of the
spring water.
This hypothesis is also confirmed by a
negative cross-correlation at lag 1 between
rainfall (P) and EC (Fig. 12). In fact the
EC decrease in the spring water at lag 1
corresponds to a maximum positive cross-
correlation between flow rate and rainfall,
i.e., the dilution of spring water is due to
the local rainfall infiltration.
The variations in CCF found in the analyses sug-
gest that local short-term rainfall increase spring flow
rates. Other analytical results also show some charac-
teristics of a deeper spring water components, coming
from the corresponding aquifer reservoir. Figure 11a
shows a visible a positive cross-correlation between Q
and EC at lag 1, i.e., an increase in Q produces an in-
crease in EC. This trend can be explained as more-con-
ductive coming from the reservoir; this signal is not as
pronounced as that from lag 0 (superficial infiltration)
but is nevertheless statistically significant. In addition,
a positive cross-correlation was seen between Q and Ts
at lag 0 (Fig. 13a); i.e., an increase in Q is positively
correlated with an increase in Ts. Figure 13b shows an
at lag 0 and 1 (Fig. 15a). However, Figure 15b shows
a similar temperature trend in winter, while in sum-
Fig. 13 - (a) Cross-correlation of spring flow rate (Q) vs spring water
temperature (Ts); (b) comparison between water temperature
(black line) and flow rate of the spring (red line)
Fig. 12 - (a) Cross-correlation between rain-
fall (P) vs electrical conductivity
(EC); (b) comparison between rain-
fall (red histogram) and electrical
conductivity (black line)
increase in Ts in winter that is positively correlated to
an increase in Q, while in summer the reverse is seen.
This reversal is most likely because, during the dry
season, the low flow rate allows a partial equilibrium
with the rock matrix.
EGA NASS
Analysis of the data from EgaNass spring tends to
be less informative (than that from Le Spesse spring)
because of problems related to functioning of the data-
logger during the monitoring period.
Analysis of Figure 14a seems to indicate a rela-
tionship between P and Q, but CCF analysis yields no
significant cross-correlations (Fig. 14b).
Analyses of atmospheric and spring water tem-
peratures confirm a local influence of shallow infiltrat-
ing waters associated with a positive cross-correlation
background image
P. FABBRI, M. ORTOMBINA, L. PICCININI, D. ZAMPIERI & L. ZINI
548
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
at lag 1 means that an increase of Q causes
a decrease of Ts and vice versa, the increase
in Q is essentially due to rainfall infiltration
after one day, after which the decrease in Q
is associated with water from the reservoir
and an increase in Ts.
VESPA INDEX
Because we collected monitoring data
spanning more than one year, the VESPA
(Vulnerability Estimator for Spring Protec-
tion Areas) index was calculated to quan-
tify the vulnerability of Le Spesse spring.
This index is an estimate of vulnerability
based on an analysis of spring hydrographs.
The VESPA index (G
alleani
et alii, 2011)
takes into account the discharge (Q), water
spring temperature (T) and electrical con-
ductivity (EC), based on 1 year of monitor-
ing data from spring discharge.
It is well known that different hy-
drographs represent different aquifers
(a
Mit
et alii, 2002; M
alvicini
et alii,
2005). Every aquifer drainage system
has an impulse function transforming an
input, for example rainfall, into output
parameters such as discharge, temperature
and EC variations. Such an analysis may
provide information regarding ground-
water network connectivity (P
laGneS
&
b
akaloWicz
, 2001; v
iGna
, 2007; k
reSic
& S
tevanovic
, 2009).
The equation for calculating the
VESPA index is as follows:
V
=
c
(ρ)
βy
where c(ρ) is the correlation factor, represented by
c(ρ) = [u(-ρ) + 0.5u(ρ)] |ρ|, ρ is the correlation coef-
ficient between Q and EC related to a time interval of
one year and u(ρ) is the Heavside step function:
The factor β, the temperature variability factor, is
defined as follows:
where T
max
and T
min
are the maximum and minimum
temperature, respectively, during the monitored year.
The factor γ, the discharge variability factor, is
mer a partial inversion is seen, as in the corresponding
diagram from Le Spesse spring.
Analyses of EC and Q from Le Spesse spring
shows a negative cross-correlation but at lags 1 and
2 instead of at lag 0 (Fig. 16a), which means that an
increase of Q is associated with a decrease in EC,
indicating a superficial component of dilution with
a greater lag time. Regrettably, the monitoring pe-
riod was shorter than that of Le Spesse spring, and
the relationship observed might only be an apparent
or spurious one.
Problems that are related to the changing of in-
struments during the monitoring period do not allow
a more detailed comparison.
In Figure 16b, where a negative cross-correlation
Fig. 14 - (a) Rainfall (black histogram bars) plotted against spring flow
rate (red line); (b) CCF between rainfall and spring flow rates
Fig. 15 - (a) CCF between atmospheric temperature (Ta) and spring wa-
ter temperature (Ts); (b) comparison between atmospheric and
spring water temperature
background image
HYDROGEOLOGICAL SPRING CHARACTERIZATION IN THE VAJONT AREA
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
549
defined as follows:
where Q
max
, Q
min
and Q
m
are the maximum,
minimum and average discharges, respec-
tively, during the monitored year.
Based on the correlation coefficient,
G
alleani
et alii (2011) proposed three
broad behavioral categories (Tab. 2) rep-
resenting various levels of network effec-
tiveness and, consequently, various styles
of response to an infiltration input.
A Type A spring is one associated
with a highly effective drainage system,
with well-developed karst conduits,
quick, strong discharge, and discharge
that is rapidly depleted. This type of
system has also been called a hypersensitive karst
(H
obbS
& S
Mart
, 1986). The value of the annual
discharge variability index (M
einzer
, 1923) is quite
high, with values upwards of 100%.
A Type B spring is associated with a moderately
effective drainage system. In these types of aquifers,
a certain storage volume is present, and the infiltra-
tion produces a typical “piston effect” for mobilizing
the resident groundwaters. Groundwater is in thermal
equilibrium with its reservoir, and its EC is primarily
than that of freshly infiltrated waters. The piston effect
produces an increase in Q, EC and T.
A Type C spring is associated with a relatively
ineffective drainage system, and the piston effect is
barely visible or absent. The discharge displays low
fluctuations, with delays of up to several months asso-
ciated with significant changes in rainfall. The electri-
cal conductivity and temperature show similar trends,
with low variations. A large saturated zone is present,
and water infiltrations reach equilibrium with the aq-
uifer and resident groundwater; external output due to
the infiltrative process is strongly reduced.
In this case, only our data from Le Spesse
spring were collected over a sufficiently long
monitoring period. The monitoring period spanned
from 1 October 2010 to 30 September 2011, during
which time the discharge, temperature and EC data
were monitored hourly.
The results show a correlation coefficient (ρ) of
-0.1. Thus, Le Spesse spring falls into type C (homog-
enization; Tab. 2).
This result is in agreement with previous analyses,
indicating that the initial increase of discharge was due
to local infiltrating waters, and only after a period of
time did the groundwater arrive from the deep reservoir.
A β factor of 2, a γ factor of 2.8 and a c(ρ) factor
of 0.1 produce a VESPA index of 0.83, which places
Le Spesse spring in the category of medium vulner-
ability (0.1 < V < 1 ; Tab. 3).
ISOTOPIC ANALYSIS
In this study, an analysis of stable isotopes in wa-
ters, such as oxygen-18 (δ
18
O) and deuterium (δD),
was also performed. Le Spesse and Ega Nass springs
were sampled monthly from April to August 2011.
Three rain samplers were installed, the first in Cava
Buscada (municipality of Erto-Casso, at an altitude
of approximately 1,750 m a.s.l.), the second in Le
Spesse (in the same municipality, at an altitude of 810
m a.s.l.) and the third in Longarone (at approximately
450 m a.s.l.; Fig. 1). Beginning in April 2011, eleven
samples of rainfall waters and sixteen samples of
Fig. 16 - (a) CCF between spring flow rate and electrical conductivity;
(b) CCF between spring flow rate and spring water temperature
Tab. 2 - Spring categories (from G
alleani
et alii, 2011)
Tab. 3 - VESPA index (from G
alleani
et alii, 2011)
background image
P. FABBRI, M. ORTOMBINA, L. PICCININI, D. ZAMPIERI & L. ZINI
550
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
spring water were collected. The samples were then
analyzed at the Geosciences Department of Trieste
University to measure their stable isotopic contents.
The derived (empirical) relationship between lo-
cal δ
18
O and δD is as follows:
δD
=
7.9
δ
18
O
+
12.24
The highly correlative plot of the experimental
data (coefficient of correlation of 0.98) and a high
similarity between own experimental line and those
from the literature (G
at
& c
arMi
, 1970; c
raiG
, 1961;
l
onGinelli
& S
elMo
, 2003; Fig. 17) indicate that
evaporation in the Vajont area does not affect the
amount of D and
18
O in local rainfall and mixing with
connate or very old waters may be excluded.
From Figure 17, it may be seen that the values
of spring samples are less scattered/dispersed. An at-
tenuation phenomenon of seasonal variations in
spring waters is well known. As groundwater en-
ters confined conditions, it is isolated from further
seasonal and rainfall contribution and its isotopic
compositions are attenuated to values represent-
ing a weighted mean of meteoric water inputs. The
spring values show variation with respect to time,
but while the variation in rainfall samples show a
direct correlation with respect to air temperature,
the spring samples show a shift in time. This phe-
nomenon is observable both in the δ
18
O and δD
contents (Fig. 18).
An explanation for this trend is that during sum-
mer the spring discharge mainly consists of water that
infiltrated during the previous winter, as suggested by
the comparison between discharge and water tempera-
ture. In fact, during the dry season the bulk discharge
is from groundwater storage (fissure/porous matrix
network; c
lark
& f
ritz
, 1997), whereas in the wet
Fig. 17 - Isotopic diagram. Local Va-
jont line (dashed red line) in
comparison with GMWL (solid
blue line; C
RaiG
, 1961), MMWL
(solid black line; G
at
& C
aRmi
,
1970) and North Italy (solid
light blue line; l
onGinelli
&
S
elmo
, 2003)
season the groundwater is mixed with waters of local
infiltration.
Even if few isotopic data are present, it is possible
to assume that the phase shift of isotopic values in the
spring water samples were due to the residence time
of rainfall in the aquifer. This phenomenon is also
known in the literature as seasonal isotopic inversion
(l
onGinelli
& S
elMo
, 2010).
Afterwards, the average value of rainfall isotopic
composition was calculated to estimate a relationship
between δ
18
O in local rainfall and elevation.
We were aware of the limited number of data and
anomalous isotopic values found in the Longarone rain-
fall samples (elevation 450 m a.s.l.), which suggested
that we eliminate these data from the estimation. Nev-
ertheless, we estimated a theoretical meteoric local line
of δ
18
O vs elevation (notwithstanding the evident limit
of two available points). The δ
18
O and δD values found
in the Longarone rainfall samples were more negative
than those in Le Spesse rainfall samples (750 m a.s.l.).
However, these anomalous values are common in the
Alpine areas, and they are mainly related to the posi-
tion of the mountains with respect to the direction of
the prevailing wind and the prevailing trajectory of the
rain-cloud front (l
onGinelli
& S
elMo
, 2010).
When including only the Le Spesse and Cava
Buscada δ
18
O average contents vs altitude, it was pos-
sible to estimate an approximate theoretical meteoric
local line (Fig. 19), as follows:
δ
18
O
=
-
0.0028
h
-
5.0393
where h is the elevation (m a.s.l.).
The line shows an average gradient of -0.28% for
every 100 m of elevation gain. This gradient is very
similar to the medium line over the Adriatic slope pre-
sented by z
uPPi
et alii (1974), as shown in Figure 19.
background image
HYDROGEOLOGICAL SPRING CHARACTERIZATION IN THE VAJONT AREA
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
551
By calculating the average values of δ
18
O in
spring water, it becomes possible, using the medium
local line, to calculate the average elevation of wa-
ter that recharges the aquifers supplying the springs
in our study area. The calculated elevations of aquifer
recharge water are approximately 1,700 m a.s.l. for
Le Spesse spring, 1,600 m a.s.l. for Ega Nass spring,
2,200 m a.s.l. for Cava Buscada spring and 1,900 m
a.s.l. for Del Cristo spring. The results show recharge
rates that are in agreement with local hydrogeological
and morphological conditions.
Fig. 19 - Diagram of δ
18
O versus al-
titude: medium local line in
comparison with medium
line of the Adriatic slope
(Z
uppi
et alii, 1974)
Fig. 18 - (a) Comparison between air temperature (T) and δ
18
O in rainfall at Le Spesse; (b) comparison between T and δD in
rainfall at Le Spesse; (c) comparison between T at Le Spesse and Longarone and δ
18
O of Le SPesse and Ega Nass
springs; d) comparison between T at Le Spesse and Longarone and δD of Le SPesse and Ega Nass springs
HYDROGEOLOGICAL
CONCEPTUAL
MODEL
In a general conceptual model, the karstic reser-
voir may be thought of as two interconnected parallel
flow systems, one associated with high-conductivity
karstic channels and the other with a low-hydraulic-
conductivity but high-storage-capacity, fissured, po-
rous aquifer (b
eniScHke
et alii, 1988; S
eiler
et alii,
1989; M
aloSzeWSki
et alii, 2002). These two different
parts of the aquifer hold different volumes of water.
The first reservoir consists of a fissured, porous me-
dium and contains a relatively large volume of water,
whereas the second reservoir consists of karstic drain-
background image
P. FABBRI, M. ORTOMBINA, L. PICCININI, D. ZAMPIERI & L. ZINI
552
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
the VESPA index. The absence of a piston effect con-
firms the low level of karstification of this aquifer and a
misleading karst response due to local infiltration.
The rainfall water enters the aquifer system at the
surface of the catchment area and infiltrates down to
karstic channels. From these channels, a portion of the
water enters the fissure-porous aquifer. Water entering
the karstic-fissured system from these channels contrib-
utes to the mean transit time of water in that system. The
water in the channels finally discharges at the springs.
The composition of the spring water is a weighted
combination of water fluxes from both the karstic and
fissured/porous sources. Thus, the karst is not well de-
veloped and its channels are of lesser hydrodynamical
relevance than the fissured part of the reservoir.
ACKNOWLEDGMENTS
This research was supported by a grant, GEO-RISKS
- Geological, morphological and hydrological process-
es: monitoring, modeling and impact in the north-east-
ern Italy
, funded by the University of Padova (Project
Number STPD08RWBY_001). The authors would like
to thank an anonymous reviewer for helpful comments
that contributed to improvements in the paper.
age channels containing little water.
In our springs, the conceptual model derives
from analyses of diagrams displaying discharge,
temperature and EC, different cross-correlations and
the VESPA index, which are partially based on iso-
topic analytical results. In particular, Le Spesse spring
shows a slow circulation of the deep reservoir water,
and the high increases in discharge are essentially due
to local infiltration rather than infiltration in the true
(larger) recharge area (increases in discharge are as-
sociated with decreases in EC and T; there are positive
cross-correlations between Ta and Ts). The compari-
son between discharge and temperature shows a de-
crease in Ts during the summer and an increase during
the winter, indicating a discharge of deep reservoir
water in partial thermal equilibrium with the rock ma-
trix, and this seasonal trend was confirmed by the iso-
topic analytical results (seasonal isotopic inversion).
A residence time of six months for the total spring
flow is a relatively low value for karst aquifers.
The conceptual model is one of a regular discharge
(base flow of deep waters) along with a local circula-
tion directly linked to infiltration through the debris,
which was confirmed by the correlation coefficient of
REFERENCES
a
Mit
H., l
yakHovSky
v., k
atz
a., S
tarinSky
a. & b
urG
a. (2002) - Interpretation of spring recession curves. Ground Water,
40: 543-551.
b
elloni
l.G. & S
tefani
r. (1987) - The Vaiont slide: instrumentation-Past experience and the modern approach. Engineering
Geology, 24: 445-474. doi: 10.1016/0013-7952(87)90079-2.
b
eniScHke
r., z
ojer
H., f
ritz
P., M
aloSzeWSki
P. & S
ticHler
W. (1988) - Environmental and artificial tracer studies in an Alpine
karst massif (Austria). In y
uan
d. & X
ie
c. (
edS
.). Karst hydrogeology and karst environment protection. Proceedings IAH
21
st
Congress: Beijing, Geological Publishing House, 938-947.
b
eSio
M. (1986) - Hydrogeological notes regarding mount Toc and vicinity. In: S
eMenza
e. & M
elidoro
G. (
edS
.). Proceedings of
the meeting on the 1963 Vaiont landslide. Convegno sulla frana del Vaiont, 133-155, Ferrara, Italy, 17-19- September-1986,
c
ivita
M. (1972) - Schematizzazione idrogeologica delle sorgenti normali e delle relative opere di captazione. Mem. e Note Ist.
Geol. Appl., 12, Napoli.
c
lark
i. & f
ritz
P. (1997) - Environmental isotopes in hydrogeology. Lewis Publishers, CRC Press, Boca Raton - New York,
328 pp.
c
orbyn
j.a. (1982) - Failure of a partially submerged rock slope with particular references to the Vajont rock slide. International
Journal Engineering and Mining Sciences, 19: 99-102. doi: 10.1016/0148-9062(82)91635-7.
c
raiG
, H. (1961) - Isotopic variations in meteoric waters. Science, 133: 1702-1703.
G
alleani
l., v
iGna
b., b
anzato
c. & l
o
r
uSSo
S. (2011) - Validation of a vulnerability estimator for spring protection areas:
the VESPA index. Journal of Hydrology, 396: 233-245.
G
at
j.r. & c
arMi
i. (1970) - Evolution of the isotopic composition of atmospheric waters in the Mediterranean Sea area. J.
Geoph. Res., 75: 3039-3048.
G
enevoiS
r. & G
Hirotti
M. (2005) - The 1963 Vaiont Landslide. Giornale di Geologia Applicata, 1: 41-53. doi: 10.1474/
GGA.2005-01.0-05.0005.
background image
HYDROGEOLOGICAL SPRING CHARACTERIZATION IN THE VAJONT AREA
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
553
G
Hirotti
M. (1994) - Nuovi dati sulla frana del Vaiont e modellazione numerica. Geologica Romana, XXX: 207-215.
H
elMStetter
a., S
ornette
d., G
raSSo
j.r., a
nderSen
j. v., G
luzMan
S. & P
iSarenko
v. (2004) - Slider block friction model
for landslides: application to Vaiont and La Clapière landslides. Journal of Geophysical Research, 109: 1-15. doi:
10.1029/2002JB002160.
H
endron
a.j. & P
atton
f.d. (1985) - The Vaiont slide, a geotechnical analysis based on new geological observations of the
failure surface. Tech Rep GL-85-5, 2 vols. Department of the Army, US Corps of Engineers, Washington, DC.
H
obbS
S.l. & S
Mart
P.l. (1986) - Characterization of carbonate aquifers: a conceptual base. Proceedings 9
th
Congress of
Speleology, 1: 43-46.
k
ilburn
c.r.j. & P
etley
d. n. (2003) - Forecasting giant, catastrophic slope collapse: lessons from Vajont, Northern Italy.
Geomorphology, 54: 21-32. doi: 10.1016/S0169-555X(03)00052-7.
k
reSic
n. & S
tevanovic
z. (e
dS
.) (2009) - Groundwater hydrology of springs. Elsevier Inc., 565 pp. ISBN: 978-1-85617-502-9.
l
onGinelli
a. & S
elMo
e. (2003) - Isotopic composition of precipitation in Italy: a first overall map. Journal of Hydrology, 270:
75-88.
l
onGinelli
a. & S
elMo
e. (2010) - Isotope geochemistry and the water cycle: a short review with special emphasis on Italy.
Mem. Descr. Carta Geol. d’It., XC: 153-164.
M
aloSzeWSki
P., S
ticHler
W., z
uber
a. & r
ank
d. (2002) - Identifying the flow systems in a karstic-fissured-porous aquifer, the
Schneealpe, Austria, by modelling of environmental
18
O and
3
H isotopes. Journal of Hydrology, 256 (1-2): 48-59.
M
einzer
o.f. (1923) - The occurrence of groundwater in the United States. Government Printing Office, USGS, Water Supply
Paper 489. Washington D.C. (USA).
M
alvicini
c.f., t
aMMo
S.S., W
alter
M.t., P
arlanGe
j. & W
alter
M.f. (2005) - Evaluation of spring flow in the uplands of
Matalom, Leyte, Philippines. Adv. Water Resour., 28: 1083-1090. doi:10.1016/j.advwatres.2004.12.006.
M
antovani
f. & v
ita
-f
inzi
c. (2003) - Neotectonics of the Vajont dam site. Geomorphology, 54: 33-37. doi: 10.1016/S0169-
555X(03)00053-9.
M
üller
L. (1964) - The rock slide in the Vajont Valley. Rock Mechanics and Engineering Geology, 2: 148-212.
M
üller
L. (1968) - New considerations on the Vajont slide. Rock Mechanics and Engineering Geology, 6: 1-91.
M
üller
L. (1987) - The Vaiont catastrophe: a personal review. Engeneering Geology, 24: 513-523.
n
onveiller
E. (1987) - The Vajont reservoir slope failure. Engineering Geology, 24: 493-512, doi: 10.1016/0013-7952(87)90081-0.
P
laGneS
v. & b
akaloWicz
M. (2001) - May it propose a unique interpretation for karstic spring chemographs? In: Proceedings
of the 7
th
Conference on Limestone Hydrology and Fissured Media, Besançon, France (20-22 September 2001), 293-298.
S
eMenza
e. & G
Hirotti
M. (200) - History of the 1963 Vaiont slide: the importance of geological factors. Bull. Eng. Geol. Env.,
59: 87-97
r
iva
M., b
eSio
M., M
aSetti
d., r
occati
f., S
aPiGni
M. & S
eMenza
e. (1990) - La geologia delle valli Vaiont e Gallina (Dolomiti
orientali). Ann. Univ. Ferrara Sez. Sci Terra, 2 (4): 55-76.
S
eiler
k.P., M
aloSzeWSki
P. & b
eHrenS
H. (1989) - Hydrodynamic dispersion in karstified limestones and dolomites in the Upper
Jurassic of the Franconian Alb, F.R.G. Journal of Hydrology, 108: 235-247.
S
eMenza
E. (1965) - Sintesi sugli studi geologici sulla frana del Vajont dal 1959 al 1964. Mem. Mus. Tridentino Sci. Nat., 16: 52 pp.
t
ika
t.H.e. & H
utcHinSon
n. (1999) - Ring shear tests on soil from the Vaiont landslide slip surface. Geotecnique, 49: 59-74.
v
eveakiS
e., v
ardoulakiS
i. & d
i
t
oro
G. (2007) - Thermoporomechanics of creeping landslides: The 1963 Vaiont slide,
northern Italy. Journal of Geophysical Research, 112: 1-21. doi: 10.1029/2006JF000702.
v
iGna
B. (2007) - Classification and operation of aquifer systems in carbonate rocks. Memorie dell’Istituto Italiano di
Speleologia, XIX: 21-26. ISBN: 978-88-89897-03-4.
W
ard
S.n. & d
ay
S. (2001) - The 1963 Landslide and Flood at Vaiont Reservoir Italy. A tsunami ball simulation. Ital. J. Geosci.,
130 (1): 16-26.
z
uPPi
G.M., f
onteS
j.c. & l
eotolle
r. (1974) - Isotopes du milieu et circolations d’eaux sulfureès dans le Latium. Proc. Symp.
Isot. Techn. in Groundwater Hydrology, I.A.E.A., Vienna.
background image
Statistics