Document Actions

ijege-14_02-frollini-et-alii.pdf

background image
27
Italian Journal of Engineering Geology and Environment, 2 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
DOI: 10.4408/IJEGE.2014-02.O-03
E
leonora
Frollini
(*)
, e
va
Pacioni
(*)
& M
arco
PEtitta
(*,**)
(*)
Sapienza Università di Roma - Dipartimento di Scienze della Terra - P.le Aldo Moro 5 - 00185 Roma, Italy. E-mail eleonora.frollini@uniroma1.it
(**)
Sapienza Università di Roma - CERI Research Center - P.le Aldo Moro 5 - 00185 Roma, Italy
A numericAl groundwAter flow model of chienti river vAlley
(centrAl itAly): results And boundAry problems
eXtended AbstrAct
il basso bacino del fiume chienti, situato nella regione Marche, dagli anni ’90 è stato interessato da una diffusa contaminazione da
solventi clorurati usati dalle fabbriche di calzature presenti nell’area (P
acioni
et alii, 2010), per questo motivo dal 1997 l’area è stata sot-
toposta al monitoraggio da parte della USl e dell’arPaM, poi affiancate dall’Università di roma la Sapienza (arPaM, 2007). nel 2001
il basso bacino del chienti è stato inserito nei Siti di interesse nazionale, per poi essere declassato nel 2013 a Sito di interesse regionale.
tale bacino si instaura nel bacino periadriatico marchigiano, formato da argille, sabbie e conglomerati, che rappresenta il bedrock
su cui si attesta la valle del chienti, caratterizzata da quattro ordini di terrazzi dei quali i più antichi affiorano a monte mentre l’ultimo,
insieme alle alluvioni attuali, affiora nella zona più prossima alla costa (n
anni
& v
ivalda
, 1986). Proprio il iV ordine di terrazzi e le
alluvioni attuali rappresentano l’acquifero di subalveo del basso bacino del chienti, caratterizzato da ghiaie eterometriche in matrice
limoso-sabbiosa con spessori variabili lungo la valle, spesso intercalate con limi sabbiosi e sabbie argillose e limi argillosi che, per la
loro continuità verticale, rendono l’acquifero multifalda.
Sulla base del modello concettuale della circolazione idrica sotterranea (P
etitta
et alii, 2013) è stato realizzato, attraverso l’uso di Fe-
flow (DHi-WaSY GmbH, 2010) un modello numerico di flusso avente un’area di 68 km
2
e formato da un layer di copertura superficiale,
due acquiferi e un interposto livello a bassa permeabilità. Per quanto riguarda le condizioni al contorno sono stati imposti limiti a flusso
nullo a nord e a Sud, constant head pari a 0 m s.l.m. lungo la linea di costa, portate in entrata pari a 0.05 m
3
/g a monte e una ricarica di
227 mm/a. Dopo la validazione del modello, al fine di caratterizzare in dettaglio l’andamento della falda nella parte più bassa del bacino
del chienti e di identificare i possibili percorsi dei contaminanti al suo interno, ne è stato costruito uno di maggior dettaglio, sia in regime
stazionario che transitorio, in corrispondenza dell’area del campo pozzi di civitanova Marche.
il nuovo modello ha un’area di 19.5 km
2
, compresa tra Montecorsaro e civitanova Marche (in direzione W-E) e tra i terrazzi fluviali
a nord e il chienti a Sud. al fine di riprodurre un acquifero multifalda semiconfinato, il modello è stato diviso in 4 layers e 5 slices bas-
andosi sulla ricostruzione stratigrafica. le caratteristiche idrogeologiche sono state imposte ad ogni layer sulla base dei dati bibliografici
e dei risultati di prove di pompaggio. al primo acquifero è stata applicata una conducibilità idraulica orizzontale (k
x
) dell’ordine 10
-3
m/s,
mentre nel secondo k
x
varia da 1*10
-3
a 5*10
-4
m/s. la ricarica applicata, ricavata dalla precipitazione efficace ottenuta con il metodo
di thornthwaite (t
hornthwaite
, 1948), è di 95 mm/a, in accordo con i dati climatici. Per quanto riguarda le condizioni al contorno, sul
primo slice è stata applicata la condizione di cauchy lungo il chienti per simulare l’interazione con la falda, mentre a monte e a valle
è stata applicata la condizione di constant head. le stesse constant heads sono state applicate su tutte le restanti slices, dove è stata ap-
plicata la condizione di Dirichlet anche lungo il fiume. Sul limite nord è stato invece imposto un flusso nullo. infine sono stati simulati i
pompaggi dei pozzi dell’acquedotto e quelli vicino alle fabbriche presenti nell’area.
tale modello simula il reale andamento W-E della falda ottenendo un buon grado di correlazione dei livelli piezometrici con i valori
simulati (rMS 0.76 nello stazionario; rMS 0.69-0.9 nel transitorio). anche l’analisi di bilancio dimostra l’attendibilità del modello dal
momento che i quantitativi di acqua in entrata e in uscita si equiparano. Da questa analisi si evidenzia inoltre l’importanza del ruolo
delle condizioni di Dirichlet, mentre l’interazione falda/fiume appare limitata. Una situazione critica si osserva però nell’area SW, al
contatto tra i limiti orientale e meridionale. Qui infatti si riscontra uno scambio inatteso e forse irrealistico delle acque sotterranee do-
vuto al contatto a 90° di diverse condizioni al contorno, combinato con un più alto gradiente idraulico e con una k
x
dell’ordine di 10
-3
m/s. Per vedere poi quali siano i percorsi e i tempi di transito di una particella che si muova per advezione, è stata applicata la funzione
particle tracking, da cui emerge che una particella immessa direttamente nell’acquifero profondo arriva più velocemente al campo pozzi
rispetto a una immessa in quello superficiale. Da questi simulazione emerge quindi che, sebbene il modello possa essere considerato un
valido strumento per la pianificazione di interventi di bonifica, il suo potenziale utilizzo pratico richiede necessariamente la revisione
dell’interazione falda/fiume e della geometria del dominio.
background image
e. frollini, e. pAcioni & m. petittA
28
Italian Journal of Engineering Geology and Environment, 2 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
AbstrAct
Since ’90s the lower valley of chienti river has been inter-
ested by a diffused contamination by chlorinated solvents (mainly
PcE) used by local shoes companies. in order to analyze the fea-
sible paths and travel times of a pollulant in the aquifer and so the
possible problems that these contaminants can cause to the well
field of civitanova Marche, a detailed groundwater flow numeri-
cal model related to the drinking well field area has been devel-
oped, in steady and transient conditions, using Feflow 6.0 from
Wasy inc (finite elements code).
the model has four layers and reproduces a multilayer semi-
confined aquifer characterized by a shallow and a deep levels.
in the first aquifer the hydraulic conductivity (k) is ranging from
1*10
-3
m/s to 5*10
-3
m/s (storativity 0.20); in the second aquifer k
is ranging from 1*10
-3
to 5*10
-4
m/s (storativity 1.3*10
-3
); the in-
termediate local aquitard has k 10
-5
m/s and storativity 10
-2
. the re-
charge applied in steady model is 95 mm/y according with climatic
data. in the first slice, along the chienti river, a cauchy boundary
condition has been inserted. constant head conditions have been
applied along western (22 m a.s.l) and eastern (0 m a.s.l.) limits
of all slices and in correspondence with the river location, in the
slices deeper than shallow one. along the northern limit a no-flow
boundary condition inhibits flow entering or exiting from the hy-
drogeological basin. the model simulates the real W-E trend of
groundwater flow, obtaining a good correlation between simulated
and measured piezometric values (rMS 0.76 in the steady state
simulation; rMS 0.69-0.9 in the transient one). the whole flow
budget shows a comparable rate between entering and exiting
flow from the model, but a critical situation in the SW area, at
the contact between western and southern boundaries, is observed.
this contact, combined with a significant hydraulic gradient (8‰)
and with a hydraulic conductivity of about 10
-3
m/s, generates an
unexpected and perhaps unrealistic interchange of groundwater in
that area. Detailed analysis of this local situation reveals modeling
inaccuracy where different boundary conditions were applied in a
boundary area characterized by complex hydrogeological setting.
K
ey
words
: groundwater contamination, numerical model, domain
geometry
introduction
the lower valley of chienti river, located in the Marche re-
gion, has been interested since ‘90s by a diffused contamination by
chlorinated solvents (mainly PcE). During the 1980s and 1990s,
the main chlorinated compound used by shoe manufacturers locat-
ed in the study area was 1,1,1-trichloroethane (1,1,1-tca), which
was substituted by Perchloroethene (PcE) in the last 15 years.
Since 1997 to 2005 the study area has been monitored by USl
(local Public Health Unit) and arPaM (regional agency for En-
vironmental Protection) with the aim to verify the extent and the
concentration of the latest contamination by PcE and the residual
one by tca (P
etitta
& P
acioni
, 2010); and since 2009 the public
authority has been sustained by University of rome la Sapienza.
in 2001 the lower valley of chienti river has been proclaimed na-
tional interest Polluted Site and in 2003 the perimeter of the site
that includes a land area of about 26 km
2
and a marine area of 12
km
2
has been defined. in 2013 the study area has been downgraded
from national interest Site to regional interest Site.
With the aim to characterize in detail the groundwater flow in
the lower valley of chienti river and to identify possible paths of
contaminants in groundwater, a new detailed numerical groundwa-
ter flow model related to a drinking well field area has been devel-
oped, in steady and transient conditions, from a previous steady-
state wider flow model. the numerical code used is Feflow 6.0
(DHi-WaSY GmbH, 2010), which is a finite elements code that
can be efficiently used to describe the spatial and temporal distri-
bution and reactions of groundwater contaminants, to estimate the
duration and travel times of chemical species in aquifers, to plan
and design remediation strategies and capture techniques, and to
assist in designing alternatives and effective monitoring schemes.
study AreA
the lower valley of chienti river, located in the Marche
region between the districts of Macerata and Fermo, has a SW-
nE trend between trodica di Morrovalle and civitanova Marche
(Fig. 1). this valley is limited at north and South by Macerata
- Montecosaro ridge and by corridonia - Montegranaro ridge,
respectively. chienti river flows from adriatic side of Umbria-
Marche apennines to adriatic sea, after receiving Ete Morto
Fig. 1 - Location of lower valley of Chienti River (after N
aNNi
& V
iV
-
alda
, 1986, modified)
background image
A numericAl groundwAter flow model of chienti river vAlley
(centrAl itAly): results And boundAry problems
29
Italian Journal of Engineering Geology and Environment, 2 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
Stream, the only tributary of the study area.
the lower valley of chienti river is established in the Peri-
adriatic basin of Marche region, which is made up Plio-Pleis-
tocene clays, sands and conglomerates (Fig. 2). these sediments
are the bedrock of the chienti river Valley, which is character-
ized by four orders of fluvial terraces of which the oldest out-
crops upstream, and the latest outcrops downstream of the valley
(n
anni
& v
ivalda
, 1986).
the continental sequence includes sand and silty sand, inter-
bedded with silty-clay levels; more recent deposits are mainly
silty, covered by vegetated soil. these deposits are differently lo-
cated along the valley, and low permeability layers are variable in
thickness and location (Fig. 3). as a consequence, alluvial depos-
its of the chienti river Valley host a multilayer porous aquifer, by
gravels in a sand-silt matrix. Silty-clay low-permeability lenses
have locally created multilayer and perched aquifers. the aquifer
is semi-confined in some locations, while at the regional scale
groundwater flow is considered to be homogeneous (P
acioni
et
alii, 2010). the piezometric map (Fig. 4) shows predominant
groundwater flow from west to the coastline revealing a mean
hydraulic gradient of 0.5%. Evidences of river/groundwater in-
teractions are observed along the stream and a clear piezometric
depression coincides with a drinking water well-field.
as regards the climatology, the chienti river Valley can be
included in the first area described by a
Mici
& S
Pina
(2002) char-
acterized by climate from wet to semiarid, with rainfall between
600 and 850 mm/year (regione Marche, 2008).
methods
a hydrogeological conceptual model has been developed
for the alluvial aquifer taking into account the presence of low
permeability lenses, forming a multilayer semi-confined aqui-
fer, as shown by hydrodynamic tests (pumping and flowmeter
heat-pulse tests) coupled with standard and multilevel hydro-
chemical and isotopic samplings and physical-chemical param-
eter logs (P
etitta
et alii, 2013).
after the conceptual model of groundwater flow, a numeri-
cal flow model has been realised. this regional model has an
area of 68 km
2
. it is constituted by a soil layer, two aquifer lay-
Fig. 2 - Geological model of Periadriatic Basin of Marche Region. a)
upper sands and conglomerates; b) upper blue clays; c) lower
blue clays; d) pre-Pliocene deposits; e) faults; f) sector limit;
1) Ancona sector; 2) Macerata sector; 3) Fermo sector; 4)
Teramo sector (after C
aNtalamessa
et alii, 2002)
Fig. 3 - Stratigraphic distribution along the Chienti Valley (after P
etitta
& P
aCioNi
, 2010)
background image
e. frollini, e. pAcioni & m. petittA
30
Italian Journal of Engineering Geology and Environment, 2 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
ers and one interbedded low-permeability layer. the model do-
main is characterized by northern and southern no-flow limits,
corresponding to geological boundaries defined by terraces; a
constant head of 0 m a.s.l. has been imposed along the shore-
line; an inflow rate of 0.05 m
3
/s coming from upgradient and
rainfall recharge of 227 mm/y have been included as boundary
conditions (Fig. 5). after the validation of this model (Fig. 6),
a new local fine scale model has been carried out, related to
the drinking well field area in steady and transient conditions,
with the aim to analyze the feasible path and travel times of a
pollutant in the aquifer and so the possible problems that these
contaminants can cause to the drinking well field.
the fine scale model, realized with Feflow 6.0 (by DHi-
WaSY), covers an area of about 19.5 km
2
from Montecorsaro to
civitanova Marche (W-E direction) and from fluvial terraces at
north to chienti river at South (Fig. 7). in order to reproduce a
Fig. 4 - Piezometric map. In purple June-July 2009 piezometric contours; in blue November-December 2009 piezometric contours, in diamond monitor-
ing points (after P
etitta
& P
aCioNi
, 2010)
Fig. 5 - Numerical model domain and boundary conditions (after P
aCioNi
et alii, 2010)
background image
A numericAl groundwAter flow model of chienti river vAlley
(centrAl itAly): results And boundAry problems
31
Italian Journal of Engineering Geology and Environment, 2 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
multilayer semi-confined aquifer characterized by a shallow and a
deep levels, the model domain has been divided in four layers and
five slices: 1) layer 1 - soil; 2) layer 2 - shallow aquifer (gravels
and sands); 3) layer 3 - aquitard (silts and sandy-clay silts); 4)
layer 4 - deep aquifer (gravels and sands).
the topography of the slices has been reconstructed by P
a
-
cioni
et alii (2010) analyzing and reworking about 175 strati-
graphic logs available in the area and returned them to the four
layers listed above. in areas where one of these levels is not
present, it was still represented in the model by imposing a
thickness of 10 cm, so as to be sufficient for the proper func-
tioning of the model, but having low influence on the ground-
water flow of the aquifer.
the hydrogeological characteristics have been imposed at
each layer based on bibliographic data and on pumping test re-
sults. in the first aquifer the horizontal hydraulic conductivity
(k
x
) is ranging from 1*10
-3
m/s to 5*10
-3
m/s, vertical conductiv-
ity (k
z
) is one order of magnitude lower and storativity is 0.20;
in the second aquifer k
x
is ranging from 1*10
-3
to 5*10
-4
m/s, k
z
is one order of magnitude lower and storativity is 1.3*10
-3
(Fig.
8); the sandwiched local aquitard has k
x
10
-5
m/s, k
z
one order of
magnitude lower and storativity 10
-2
.
the recharge applied in steady model is 95 mm/y according
with climatic data; this value has been obtained from effective
precipitation through thornthwaite method (t
hornthwaite
,
1948), by averaging years since 2000 to 2008 and considering
a c.i.P. (potential infiltration coefficient) of 50% respect with
effective precipitation.
as regards the boundary conditions, in the first slice a
cauchy boundary condition has been applied along the chi-
Fig. 6 - Scatter plot of simulated vs measured hydraulic heads (after
P
aCioNi
et alii, 2010)
Fig. 7 - Area of fine scale numerical model (Google Maps)
Fig. 8 - Horizontal conductivity distribution of shallow (A) and deep (B) aquifers
background image
e. frollini, e. pAcioni & m. petittA
32
Italian Journal of Engineering Geology and Environment, 2 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
enti river, in order to simulate the existing interaction river/
groundwater observed by discharge measurements. constant
head conditions have been applied along western and eastern
limits of all slices and in correspondence with the river loca-
tion in the slices deeper than shallow one. imposed values are
obtained from measured piezometric levels. along the northern
limit a no-flow boundary condition inhibits flow entering or ex-
iting from the hydrogeological basin (Fig. 9). in the steady-state
model constant head at the western limits has been evaluated 22
m a.s.l., which corresponds to the hydraulic head measured in
Fig. 9 - Domain and boundary conditions of the fine scale model for each slice
Fig. 10 - Rates of pumping wells in steady and transient model; recharge
in the transient model
Tab. 1 - Constant head applied at western limit in the transient model
background image
A numericAl groundwAter flow model of chienti river vAlley
(centrAl itAly): results And boundAry problems
33
Italian Journal of Engineering Geology and Environment, 2 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
this section in May 2007 (arPaM, 2007); otherwise at the east-
ern limit the value is 0 m a.s.l. because this limit corresponds
with the shore line. in the transient model (9 time steps, each of
30 days) the constant head at eastern limit is always 0 m a.s.l.,
instead at western limit it varies monthly as shown in tab.1, be-
cause of the monthly fluctuation of measured piezometric level.
the model includes also the drinking well field and pumping
wells of the local manufacturers (E,F) (Fig. 10). the applied
flow rates in the steady and transient models are shown in figure
10, which also shows the monthly trend of applied recharge in
the transient model, calculated from thornthwaite method.
the calibration of steady-state model has been performed us-
ing the real piezometric distribution of May 2007; the transient
model has been calibrated using the real piezometric distribution
of May 2007, august 2007 and January 2008.
results And discussion
the numerical model simulates the real W-E trend of ground-
water flow with a progressive decrease of hydraulic heads from
upstream (about 22 m a.s.l.) to downstream (0 m a.s.l.), obtain-
ing a positive correlation between simulated and measured values
(rMS 0.76 in the steady state model; rMS 0.70, 0.9 and 0.69 in
the transient one, respectively at May 2007, august 2007 and Janu-
ary 2008) (Fig. 11). the whole flow budget shows a comparable
rate between entering and exiting flow from the model. in particu-
lar, as it is shown in figure 12, most of the inflows derives from
Dirchlet conditions (4.10*10
4
m
3
/d) and from recharge (5.07*10
3
m
3
/d), instead a small contribution (3.10*10 m
3
/d) is given by in-
teraction river/groundwater. this situation could be attributed to
the imposed condition of equal elevation between the water table
and the river, based on limited data and on the assumption of a
steady-state equilibrium between surface waters and groundwater.
as regards outflows, for the same reason, a small contribution is
given by interaction river/groundwater (8.70*10
2
m
3
) and instead,
major contributions are due to Dirichlet conditions (3.40*10
4
m
3
/d)
and Wells (1.06*10
4
m
3
/d). a critical situation in the SW area, at
the contact between western and southern boundaries, is observed
(Fig. 13). in fact, the higher rates of inflow groundwater are in the
SW area, where inflow about 24000 m
3
/d of water from the most
southern area of upstream, and suddenly exit about 12000 m
3
/d of
water from the most eastern area of southern limit (Fig. 13). this
unexpected and perhaps unrealistic interchange of groundwater in
that area is due to the model geometry, which shows a perpendicu-
lar contact between boundary conditions, combined with a higher
hydraulic gradient (8‰) and with a hydraulic conductivity of about
10
-3
m/s. Detailed analysis of this local situation reveals modeling
inaccuracy where different boundary conditions were applied in a
boundary area characterized by complex hydrogeological setting.
the particle tracking function was applied in order to ana-
lyze the possible path and travel times of a pollutant that moves
by advective transport. the results of this simulation show that
a particle input in the shallow aquifer travels at lower velocity
than a particle input in the deep aquifer. in fact, the particles input
in the shallow aquifer along a section located upstream of the
Fig. 11 - Measured and simulated hydraulic heads of the steady state
model (A) (R
2
=
0,9878) and of transient model (B) (R
2
=
0,9900) (January 2008)
Fig. 12 - Budget analysis of the steady state model
background image
e. frollini, e. pAcioni & m. petittA
34
Italian Journal of Engineering Geology and Environment, 2 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
civitanova Marche drinking well field, will take two and a half
years to get there; otherwise the particles input in the deep aquifer
along the same section will arrive at the same goal after one and a
half year (Fig. 14). this difference in the velocity is due to lower
hydraulic conductivity in the shallow aquifer and then because a
particle input in the shallow aquifer must pass through the aqui-
tard (k
x
=10
-5
m/s, k
z
=10
-6
m/s) between shallow and deep aquifer
before reaching the drinking well field (Fig. 15).
conclusion
the alluvial aquifer of lower valley of chienti river, in the
past polluted national interest Site and now polluted regional
interest Site, is contaminated by chlorinated solvents (mainly
PcE) and for this reason it has been continually monitored dur-
ing the time by USl, arPaM and University Sapienza of rome.
the analysis have carried to reconstruction of a conceptual model
of groundwater flow. according to it, in the area there is a main
aquifer constituted of gravelly sediments in a sandy matrix with
high values of trasmissivity. this aquifer is often interbedded by
variable thicknesses of fine sediments which locally can act as
aquitard and isolate perched aquifers.
Based on this conceptual model and on a large-scale numeri-
Fig. 13 - Map of the flow budget for the steady state model. Values in
m
3
/d. Red dots represent the inflows, blue dots represent the
outflows. The size of dots are proportional to the rates
Fig. 14 - Particle tracking in shallow (A) and deep (B) aquifers. The
drinking well field corresponds to the red box
Fig. 15 - Simulation of advective transport along a transect and related
section of hydraulic conductivity (A) and particle tracking (B).
3D Point Set#1 is particle input in shallow aquifer, 3D Point
Set#9 is particle input in deep aquifer
background image
A numericAl groundwAter flow model of chienti river vAlley
(centrAl itAly): results And boundAry problems
35
Italian Journal of Engineering Geology and Environment, 2 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
cal model, a new local fine-scale numerical model has been car-
ried out, both in steady and transient conditions, applied to the
lower valley of the chienti river, with the aim to analyze the
feasible path and travel times of a pollutant in the aquifer and so
the possible problems that these contaminants can make to the
well field of civitanova Marche.
Both models coherently reproduce the flowpath having W-E
direction, with an high correlation coefficient between simu-
lated and measured water table. Hydrological budget analysis
allows to confirm the important role of Dirichlet conditions,
while river/water table interaction appears to be very limited.
this situation could be attributed to the imposed condition of
equal elevation between the water table and the river, based on
limited data and on the assumption of a steady-state equilibrium
between surface waters and groundwater.
the perpendicular contact on the SW boundary, combined with
a higher hydraulic gradient (8‰) in this area and with a hydraulic
conductivity of about 10
-3
m/s, generates an unexpected and per-
haps unrealistic interchange of groundwater in that area. Detailed
analysis of this local situation reveals modeling inaccuracy where
different boundary conditions were applied in a boundary area
characterized by complex hydrogeological setting. all collected
data, coupled with the numerical model, can be used as manage-
ment tools for planning an effective remediation project and also
to ensure the protection of drinking well field, requiring a carefully
revision of the simulation model. in detail, the river/water table in-
teractions seem to have a significant uncertainty at the actual stage
of knowledge and also a revision of the geometry of the domain is
required in order to avoid problems linked to the intersection of dif-
ferent boundary conditions in a complex hydrogeological setting.
AcKnowledgements
the authors would like to thank the arPaM in the person
of Mr. Gianni corvatta and Mr. Marco Fanelli, and Fondazione
della cassa di risparmio della Provincia di Macerata (Marche
region - italy).
Received July 2014 - Accepted November 2014
references
a
Mici
M. & S
Pina
r. (2002) - Campo medio della precipitazione annuale e stagionale sulle Marche per il periodo 1950 - 2000. regione Marche.
arPaM (2007) - Risultati della fase II del Piano della Caratterizzazione del Sito di Interesse Nazionale del Basso Bacino del Fiume Chienti. Technical Report.
c
antalaMeSSa
G., c
entaMore
e., d
idaSkalou
P., M
icarelli
a., n
aPoleone
G. & P
otetti
M. (2002) - Elementi di correlazione nella successione marina
plio - pleistocenica del bacino periadriatico marchigiano. Studi Geol. camerti, nuova serie 1: 33-49.
DHi-WaSY GmbH (2010) - Feflow 6. Finite element subsurface flow & transport simulation system. Installation Guide & Demonstration exercise.
n
anni
t. & v
ivalda
P. (1986a) - Caratteri idrogeologici schematici delle falde di subalveo e delle unità arenacee plio - pleistoceniche delle Marche. Mem.
Soc. Geol. it., vol. 35: 957-978.
n
anni
t. & v
ivalda
P. (1986b) - Inquadramento idrogeologico ed influenza della tettonica sugli acquiferi di subalveo delle pianure marchigiane. Studi
Geol. camerti, vol. spec. “la Geologia della Marche”, pp. 105-133.
P
acioni
e., P
etitta
M., c
orvatta
G., F
anelli
M., c
arucci
v., S
aladini
v. & S
barbati
c. (2010) - Il sito contaminato di interesse nazionale della Bassa Valle
del Fiume Chienti (Italia): dal modello idrogeologico concettuale al modello numerico. EngHydroEnv Geology, vol 14: 195-206.
P
etitta
M. & P
acioni
e. (2010) - Caratterizzazione e bonifica dei siti contaminati in Italia: il caso di studio del SIN della Bassa Valle del Fiume Chienti (MC).
http://geologilazio.it/public/file/2011/08_agosto/12_MaGGioPetitta_1.pdf
P
etitta
M., P
acioni
e.,S
barbati
c., c
orvatta
G., F
anelli
M. & a
ravena
r. (2013) - Hydrodynamic and isotopic characterization of a site contaminated
by chlorinated solvents: Chienti River Valley, Central Italy. applied Geochemistry, 32: 164-174.
r
eGione
M
arche
(2008) - Piano di Tutela delle Acque. http://areeurbane.regione.marche.it/tra/pta/index.htm
t
hornthwaite
c. w. (1948) - An Approach toward a Rational Classification of Climate. Geographical review, Vol. 38(1): pp. 55-94.
background image
Statistics