Water: The Upwelling Water Flux Feeding Springs: Hydrogeological and Hydraulic Features

water
Article
The Upwelling Water Flux Feeding Springs:
Hydrogeological and Hydraulic Features
Francesco Fiorillo *, Libera Esposito, Giovanni Testa, Sabatino Ciarcia ID
and Mauro Pagnozzi
Department of Science and Technology, University of Sannio, 82100 Benevento, Italy;
libera.esposito@unisannio.it (L.E.); geologo.giannitesta@gmail.com (G.T.); sciarcia@unisannio.it (S.C.);
pagnozzi.mauro@libero.it (M.P.)
* Correspondence: francesco.fiorillo@unisannio.it; Tel.: +39-0824-305195

Received: 8 March 2018; Accepted: 13 April 2018; Published: 18 April 2018
Abstract: The upwelling groundwater flux has been investigated by deep piezometers in a spring area
characterized by alluvial deposits covering a karst substratum in Southern Italy. The piezometers are
of varying depth located in a flat area. They have been monitored for a long period (about 40 years),
and when measured, a good relationship between spring discharge and hydraulic head was observed.
The local upwelling groundwater flux has been deducted by the increasing of the hydraulic head
in depth, which allows the estimation of ascendant hydraulic gradient and groundwater velocity
during the dry and wet seasons. A specific analytical solution has been used to estimate the zone
involved by the ascendant flow, and could also be used in other spring areas. Some physical and
chemical characteristics of spring water have been collected, including the radon (222 Rn) activity, to
support the phenomenon of the ascendant flux. The man geological and hydrogeological features
leading to ascendant flux in karst environments is also discussed for some areas of Southern Italy,
where many springs are affected.
Keywords: upwelling flux; hydraulic head; ascendant hydraulic gradient; spring; Southern Italy
1. Introduction
Springs are among the essential features of the hydrological cycle, as their main characteristics
such as discharge, temperature and chemistry can be measured, giving fundamental knowledge of
groundwater flow.
The location of springs depend on the local topographic and hydrogeological features, and most
of them originate along geological contacts, faults and ground depressions [1,2].
Following the model of regional groundwater flow [3–5] the discharge areas are located in
depressed zones which increases the river discharge by diffused recharge or by springs. In these areas,
the vertical component of the flux can be higher than the horizontal one, and the water flux develops
in a typical ascendant path.
The ascendant flux occurring in the discharge area can lead to the start-up of a river, especially in
the karst mountain areas, where powerful springs are fed by a wide catchment which supports high
discharge for the entire hydrological year. In other cases, rivers increase their discharge when crossing
zones are characterized by ascendant groundwater flow [6].
Karst aquifers, characterized by high transmissivity and generally by capacity to store large
volumes of groundwater [7,8], can allow deep water circulation which leads to the start of the ascendant
flux in the discharge zone.
Although, spring discharge measurements can be easily or directly measured directly at the source
or along a river, the connected ascendant flux is not easily detected, because it needs the knowledge of
the vertical hydraulic gradient to be distinguished from the most common horizontal flow. Without the
Water 2018, 10, 501; doi:10.3390/w10040501 www.mdpi.com/journal/water

Water 2018, 10, 501 2 of 21
knowledge of the actual path of the spring flow, some hydrologic processes may be poorly understood
which could lead to inappropriate water management, as the delimitation of the protection zones
of springs.
In this study, we focused on the ascendant flux feeding karst springs, measuring the vertical
hydraulic gradient by monitoring deep piezometers. The spring area is located near the village of
Serino, along the Sabato River, Southern Italy, where long monitoring data are available for spring
discharge and piezometers.
A method used to estimate the wideness of the zone involved in the ascendant flux has
been provided, which can be used for other similar spring areas, characterized by ascendant flux.
The estimation is an important tool for delineating the protection zone of springs characterized by
ascendant flux.
To support this hydraulic phenomenon, deducted from the analysis of hydraulic head monitored
into several piezometers and located in the discharge area, specific geochemical measurements have
been carried out by monitoring the radon (222 Rn) activity of spring water.
In general, we found that the ascendant flux is a topic not regularly discussed in literature.
The aim of this study is to provide a detailed case history, which applies to other areas of Southern
Italy, and perhaps elsewhere.
2. Materials and Methods
2.1. Ascendant Hydraulic Gradient and Wideness of Influenced Zone

The ascendant flow has been investigated by deep piezometers, where the ascendant hydraulic
gradient has been estimated from measurements of the hydraulic head.
In the discharge area where springs are located, we assume that the piezometer depth can have
a strong control over the hydraulic head measured due to non-hydrostatic conditions, particularly,
under descendant or ascendant groundwater flux [9].
Figure 1a shows the estimation of the vertical hydraulic gradient by the hydraulic head
measurements in wells; here the difference in the hydraulic head measured in two close boreholes is
caused by the vertical (ascendant) flux. This estimation is based on the assumption that no horizontal
component of the flow occurs, and the hydraulic head is lost during the ascendant path of the flow.
The estimated ascendant hydraulic gradient has been considered as an hydraulic model, where
analytical solutions define some geometrical features of the phenomenon, particularly the wideness of
the zone involved by the ascendant flux feeding springs or drainage systems.
In literature, analytical solutions of discharge drained by deep drainage have been provided by
Hooghoudt [10]. Later on, other authors [11,12] improved the analytical procedure of Hooghoudt’s
method. In general, this model assumes steady drainage towards ditches (or tile drains) in homogenous
soil underlain by an impervious layer; it applies the radial flow equation to the region near the
drains and the ellipse equation to the region away from the drains to determine the water table drop
(or head loss) in each region. The water table drops on each region were added to find the total
drop. This method is a useful practice procedure for designing a drainage system, aimed at regulating
the water table of an unconfined aquifer, under a constant steady-state recharge. However, in the
Hooghoudt’s model, a prevalent horizontal flow through the porous media is induced by the drainage
system, and no ascendant flux toward the drainage is considered; moreover, an impervious layer is
fixed below the drainage system.
Other types of analytical solutions have been provided to calculate steady-state inflow into a
circular tunnel [13–16] useful for the design of the tunnel’s drainage systems [17]. In these models, the
discharge drained by the tunnel depends on some geometrical and hydraulic features, such as the depth
of the tunnel, the groundwater level, the tunnel radius and the hydraulic conductivity of the aquifer.
The transient conditions are not considered, as the aim is to obtain the groundwater inflow (volume of
water for unit tunnel length) generated by the introduction of a tunnel in an initial hydrostatic system.
Water 2018, 10, 501 3 of 21
The inflow converges into the tunnel (thus the flow is ascendant in the inverse zone of the tunnel), and
the differences between inflows at the crown and invert decrease with increasing depth of the tunnel
under the2018,
Water groundwater
10, x FOR PEER table.
REVIEWAs these models assume initial hydrostatic conditions, they cannot 3 of 21 take
into account the ascendant groundwater flow condition.

increasing depth of the tunnel under the groundwater table. As these models assume initial
To estimate the influenced zone of the aquifer by the drainage tunnel, a simple 2D model is
hydrostatic conditions, they cannot take into account the ascendant groundwater flow condition.
provided To under a non-hydrostatic
estimate the influenced initial
zone ofcondition,
the aquifer where
by theandrainage
ascendant hydraulic
tunnel, gradient
a simple 2D model causes
is the
upwelling
provided under a non-hydrostatic initial condition, where an ascendant hydraulic gradient causeswhich
flow. The aim is not to estimate the inflow or the discharge in steady-state conditions,
wouldtheneed a moreflow.
upwelling complex
The aimapproach
is not to(which
estimateisthe
beyond
inflowthe
or scope of this in
the discharge work), but to analyze
steady-state the role
conditions,
of ascendant
which would hydraulic
need a gradient
more complexon the wideness
approach of zone
(which involved
is beyond by the
the scope ascendant
of this flow,
work), but namely the
to analyze
the rolezone.
influenced of ascendant hydraulic
This aspect playsgradient on the wideness
a fundamental of zone involved
role in delimitating by the ascendant
the protection zone offlow,
springs
namely the influenced
and their tapping systems. zone. This aspect plays a fundamental role in delimitating the protection zone
of springs
To estimate andthetheir tapping
width systems.
of the influenced zone, with distance R from a drainage tunnel or from a
To estimate the width of the influenced zone, with distance R from a drainage tunnel or from a
spring, the analytical solution assumes: (i) constant hydraulic conductivity in all directions; (ii) high
spring, the analytical solution assumes: (i) constant hydraulic conductivity in all directions; (ii) high
aquifer thickness to guarantee vertical flow lines below a specific depth from the drainage system or
aquifer thickness to guarantee vertical flow lines below a specific depth from the drainage system or
spring; (iii) each
spring; flow
(iii) each lineline
flow follows
follows thethedirection
direction of of “maximum slope”.
“maximum slope”.
(a) (b)
Figure 1. (a)
Figure 1. Ascendant
(a) Ascendantflux
fluxconnected
connected to to artesian conditionsofof
artesian conditions karst
karst aquifer
aquifer covered
covered by alluvial
by alluvial
deposits.
deposits. Hydraulic
Hydraulic headobserved
head observed in in close
close boreholes
boreholeswith withdifferent depths
different allow
depths to estimate
allow to estimatethe the
ascendant
ascendant hydraulic
hydraulic gradient,
gradient, ia; (b)
ia ; (b) Sketchofofestimation
Sketch estimationof of the
the ascendant
ascendant hydraulic
hydraulicgradient,
gradient,ia, iand
a , and of
of the potential
the potential hydraulic
hydraulic gradient,
gradient, iT, the
iT , in in the pointP;P;HHP,, excess
point excess ofofhydraulic
hydraulichead of the
head point
of the P respect
point P respect
P
to undisturbed water table; D, elevation difference between undisturbed water table and drainage
to undisturbed water table; D, elevation difference between undisturbed water table and drainage
system; R, width of water table involved by drainage (influenced zone); LP, horizontal distance of P
system; R, width of water table involved by drainage (influenced zone); LP , horizontal distance of P
from drainage; ZP, depth of P from drainage; ZK, depth of the karst substratum.
from drainage; ZP , depth of P from drainage; ZK , depth of the karst substratum.
Under these assumptions, the drainage system induces a potential hydraulic gradient, iT, in each
generic these
Under point assumptions,
P of the aquiferthe (Figure 1b); asystem
drainage genericinduces
flow lineacontinues
potentialalong its vertical
hydraulic ascendant
gradient, iT , in each
path for i < i , and up to i = i . When the condition changes, i > i , the flow line
generic point P of the aquifer (Figure 1b); a generic flow line continues along its vertical ascendant
T a T a T a diverges from the
vertical and tends to reach the drainage system (or springs). From Figure 1b, the ascendant
path for iT < ia , and up to iT = ia . When the condition changes, iT > ia , the flow line diverges from the hydraulic
gradient, ia, and potential hydraulic gradient, iT, are computed as:
vertical and tends to reach the drainage system (or springs). From Figure 1b, the ascendant hydraulic
gradient, ia , and potential hydraulic ( + + )−( + )
= gradient, iT , are computed = as: (1)
+ +
( H + D + Z ) − (D + Z ) H
i a = = (P + + P ) − = ( P += ) + P (2) (1)
Z+P + D + D + ZP
where HP is the excess hydraulic P + of
( Hhead D the P ) −PZwith
+ Zpoint P i a ( D +toZthe
respect P) + D
undisturbed water table; D is
iT = =table q (2)
the elevation difference between the undisturbed water and drainage system; ZP is the depth of
q
L2P + ZP2 L2P + ZP2
point P with respect to the drainage system; and LP is the distance of point P from the drainage
system.
Water 2018, 10, 501 4 of 21
where HP is the excess hydraulic head of the point P with respect to the undisturbed water table; D is
the elevation difference between the undisturbed water table and drainage system; ZP is the depth of
point P with respect to the drainage system; and LP is the distance of point P from the drainage system.
If equations are related to the spring outlet, then the parameters D, ZP and LP have to be replaced
by D’, ZP ’ and LP ’ (Figure 1b).
2.2. Physical and Chemical Characteristics Measurements

Monthly temperature, electrical conductivity and radon (222 Rn) activity measurements have been
carried out in the spring water.
Temperature and electrical conductivity have been detected by multiparameter probe (Horiba
mod.U50), immersing the sensor directly in the spring water outlet.
The measurements of 222 Rn activity have been carried out in spring water; water samples were
taken directly into the drainage tunnel of the spring tapping system, from the depth of about 0.5 m
below the level of water surface, using 500 mL glass bottles. Completely filed bottles were tightly
closed when still submerged to prevent the entry of air and the release of radon from it. The samples,
immediately sealed by parafilms, were transported to the laboratory of Hydrogeology of the University
of Sannio for radon activity measurements with a minimum delay from collection time (30’). The 222 Rn
activity was performed using AlphaGUARD radon detector (Genitron Instruments GmbH, Germany):
a pulse-counting ionization chamber suitable for alpha spectroscopy. Water samples of 100 mL were
placed in the appropriate system of glass vessels. Detected 222 Rn activity levels were analyzed using
the software Data EXPERT (by GENITRON Instruments).
The main interest of 222 Rn activity in spring water is its possible connection with groundwater
velocity and, consequently, its variation during the high or low flow conditions of springs.
3. Geological and Hydrogeological Setting

The study area is located in Southern Italy, inland of about 40 km from Naples, on the
north-western boundary of the Picentini Mountains, including the high and middle valley of the
Sabato River. Locally, there are several different geological formation outcrops, forming part of the
Irpinia Apennine chain (Figure 2).
The outcropping sedimentary successions, as described in a recent geological map (ISPRA, 2017),
are grouped into several stratigraphic or tectonic units: (i) Mesozoic platform carbonates of the
Picentini Unit with a high permeability rate; (ii) Meso-Cenozoic deep-basin deposits of the Lagonegro
Unit with a low permeability rate; (iii) upper Tortonian/lower Messinian wedge-top basin clastic
sediments of the Castelvetere Fm. with a low permeability rate; (iv) deposits of Quaternary age with a
medium-high permeability rate.
The tectonic record highlights several main sequences of events, from the carbonate platform
involved in the wedge accretion and north/east-ward overthrust onto the Lagonegro basin deposits
during the lower-middle Miocene [18] to extensional tectonic movements of Plio-Pleistocene, where
the NW-SE left-lateral strike-slip main fault [19], crossing the Sabato River valley was generated.
Subsequently, the extensional movements, especially the Tyrrhenian side of the chain, induced the
creation of the Sabato River graben, filled by fluvio-lacustrine sediments.
These tectonic regimes induced a hydrogeological framework that has dislocated carbonate rocks
in high morphological element (horst) and low morphological element (graben), creating a ground
space between the recharge area and the discharge area.
In this context, the Terminio-Tuoro massif constitutes a wide karst system, feeding several karst
springs (Figure 3).
The entire karst area is estimated to be about 163 km2 and provides a total spring discharged
water volume of 170.5 × 106 m3 /year [20]. It is arduous to define each spring catchment, and
probably a unique water table feeds all basal springs. However, more than 500 faults have been locally
Water 2018, 10, 501 5 of 21
surveyed [21], and some of them are believed to limit the entire Terminio karst massif in four main
sectorsWater
[22], each
2018, 10, xone
FOR feeding several basal karst spring groups.
PEER REVIEW 5 of 21
Water 2018, 10, x FOR PEER REVIEW 5 of 21
Figure 2. Geological sketch of Southern Appennines (from [23], modified).

To achieve the aim of the current study, the lithologies involved in Terminio aquifer were
Togrouped
achieveinto
thethree
aim lithological
of the current study, the lithologies
Flyschinvolved in Terminio aquifer were grouped
To achieve the aim of thecomplexes: Alluvial,
current study, and involved
the lithologies Limestonein(Figure 3), characterized
Terminio aquifer were
into three
by lithological
different type complexes:
and degree ofAlluvial, Flysch
permeability. and Limestone (Figure 3), characterized by different
grouped into three lithological complexes: Alluvial, Flysch and Limestone (Figure 3), characterized
type and degree of
by different permeability.
type and degree of permeability.
Figure 3. Terminio massif hydrogeological map (location in Figure 2): (1) cover deposits (Quaternary)
with medium-high permeability (slope breccias and debris, pyroclastic, alluvial and lacustrine
Figure 3. Terminio massif hydrogeological map (location in Figure 2): (1) cover deposits (Quaternary)
3. Terminio
Figuredeposits); massif terrains
(2) no-karst hydrogeological map (location
with low permeability in Figurecomplex
(argillaceous 2): (1) cover deposits
and flysch (Quaternary)
sequences of
with medium-high permeability (slope breccias and debris, pyroclastic, alluvial and lacustrine
with medium-high permeability
Paleogene–Miocene); (slope
(3) karst terrains breccias and
(calcareous-dolomite debris,
seriespyroclastic, alluvial
of Jurassic–Miocene);
deposits); (2) no-karst terrains with low permeability (argillaceous complex and flysch sequences of
and
(4) lacustrine
Towns;
(5) main
deposits); karst springs;
(2) no-karst (6) elevationlow
terrains points (m a.s.l.); (7)(argillaceous
mountain peak; (8) AA’ section line; (9) S1
Paleogene–Miocene); (3) karstwith permeability
terrains (calcareous-dolomite complex and flysch
series of Jurassic–Miocene); sequences
(4) Towns;
monitored well. Dashed
of Paleogene–Miocene); (3) rectangle
karst area is(calcareous-dolomite
terrains detailed in Figure 5. series of Jurassic–Miocene); (4) Towns;
(5) main karst springs; (6) elevation points (m a.s.l.); (7) mountain peak; (8) AA’ section line; (9) S1
(5) main karst well.
monitored springs;
Dashed (6)rectangle
elevation areapoints (m in
is detailed a.s.l.);
Figure(7)5. mountain peak; (8) AA’ section line;
(9) S1 monitored well. Dashed rectangle area is detailed in Figure 5.
Water 2018, 10, 501 6 of 21
In alluvial deposit (Quaternary) slope breccias, debris pyroclastic, alluvial and lacustrine deposits
are embedded. Argillaceous deposits and flysch sequences (Paleogene–Miocene) constitute Flysch
complex bounding the karstified limestone complex (Jurassic–Miocene).
The Serino group is located in the valley of the Sabato River, along the northwestern boundary of
the Picentini massif, and is formed by the Acquaro-Pelosi springs (377–380 m a.s.l.) and the Urciuoli
springs (330 m a.s.l.). These springs are fed by the Terminio massif [24,25] with an overall mean annual
discharge of 2.25 m3 /s. Roman aqueducts (first century AD) were supplied by these springs, and the
Urciuoli spring was waste-tapped between 1885 and 1888 by the Serino aqueduct. The actual aqueduct
is composed of a gravity channel followed by a system of pressured conduits that is used to supply
water to the Naples area. Additionally, the Acquaro and Pelosi springs were also re-tapped in 1934 by
the Serino aqueduct.
The Cassano springs group is located along the eastern side of the Terminio karst massif, in the
Calore River basin, and it is formed by the Bagno della Regina, Peschiera, Pollentina and Prete springs
(473–476 m a.s.l.). Also these springs are primarily fed by the Terminio massif [24,26], with an overall
mean annual discharge of 2.65 m3 /s [20]. In 1965, these springs were tapped in order to supply the
Puglia region with water, and a gravity tunnel was joined to the Pugliese aqueduct.
Among the springs of the northern side of the Terminio-Tuoro system, the Sorbo Serpico springs
should be mentioned. This spring group involves the Lagno spring (486 m a.s.l. with a mean discharge
of 60 L/s), the Sauceto-Titomanlio spring (462 m a.s.l. with a mean discharge of 120 L/s) and the
Beardo spring (446 m a.s.l. with a mean discharge of 250 L/s). The latter originated after a tunnel
excavation for hydroelectrical purpose.
The Terminio massif can be considered a wide diffuse type karst system, where the karst network
is not well developed or interconnected, and the flow is controlled by small karst fissures and occurs
in the laminar regime. The shape of the spring hydrograph produced by this system is characterized
by very few or only one smoothed peak that occurs after a time lag with regard to the rainy season; the
presence of the pyroclastic mantling deposits limit the runoff and favors a diffuse infiltration [27,28].
The long spring discharge measurements and the relation to climate variable have been analyzed
by Fiorillo and Guadagno [27], and the hydraulic aquifer behavior during droughts has been described
by Fiorillo [28].
The Terminio karst massif is characterized by wide endorheic areas (poljes), which play
an important role in the recharge processes (Figure 3). The origin of these endorheic zones is
connected to tectonic activity during upper Pliocene-Pleistocene, which has caused a general uplift
by direct faults, and a formation of graben zones. During the following continental environment
(Pleistocene-Holocene), karst processes have transformed these zones into endorheic ones, allowing
the complete absorption of runoff [20]. The largest endorheic area is the Piana del Dragone (55.1 km2 );
several sinkholes drain this endorheic area, and hydraulic works were carried out to limit the flooding
during the wet period by connecting a drainage system to the Bocca del Dragone sinkhole.
The water table of the saturated zone inside the karst area can be measured in several wells and
piezometers located in the Piana del Dragone, which can be considered as the main recharge zone of
the aquifer. Locally, the water level has a mean of 530 m a.s.l. (period from 2002–2006, Figure 4), and
can be considered as the higher elevation of the basal water table in the Terminio massif; this level is
150 and 55 m higher than the Acquaro-Pelosi and Cassano springs, respectively. On the basis of the
measured data (water level has been measured every three months), the water table level increases
during the recharge phase (autumn and winter) and decreases in late spring/summer.
Water 2018, 10, 501 7 of 21
560
550
water table, m a.s.l.

Water 2018, 10, x FOR540
PEER REVIEW 7 of 21
530
560
520
550
510
water table, m a.s.l.

540
500
530
May-02
Aug-02
Nov-02
May-03
Aug-03
Nov-03
May-04
Aug-04
Nov-04
May-05
Aug-05
Nov-05
May-06
Aug-06
Feb-03
Feb-04
Feb-05
Feb-06
520
510
Figure 4. Water table500level measured every three months in well S1, located in the Piana del Dragone
May-02
Aug-02
Nov-02
May-03
Aug-03
Nov-03
May-04
Aug-04
Nov-04
May-05
Aug-05
Nov-05
May-06
Aug-06
Feb-03
Feb-04
Feb-05
Feb-06
Figure 4. Water table level measured every three months in well S1, located in the Piana del Dragone
(Figure 3, well head 686 m a.s.l.).
(Figure 3, well head 686 m a.s.l.).
Figure 4. Water table level measured every three months in well S1, located in the Piana del Dragone
The Serino Spring Area
The Serino(Figure
Spring3, Area
well head 686 m a.s.l.).
Between the villages of S. Lucia and S. Michele di Serino, many springs exist which rapidly
Between
The Serino the villages
Spring Area of S. Lucia and S. Michele di Serino, many springs exist which rapidly
increased the discharge of the Sabato River. After drainage works to tap these springs, only few
increased the discharge of the Sabato River. After drainage works to tap these springs, only
resurgence Between the villages
phenomena can be of S. Lucia and
observed S. Michele
along di Serino,
the river, whichmany springs exist
are negligible whichthe
during rapidly
low flow
few resurgence
increased phenomena
the discharge ofcan
the be observed
Sabato River. along
After the river,
drainage which
works to are these
tap negligible during
springs, only the low
few
season.
flow resurgence
season. phenomena can be observed along the river, which are negligible during the low flow
In the Urciuoli springs zone (approximately 330 m a.s.l.), there are three drainage tunnels (North,
In the Urciuoli springs zone (approximately 330 m a.s.l.), there are three drainage tunnels (North,
season.
West and Main channels) that convey the water towards a storage tank. From this point, a gravity
West andIn Main channels)
the Urciuoli thatzone
springs convey the water 330
(approximately towards a there
m a.s.l.), storage
are tank. From this
three drainage point,
tunnels a gravity
(North,
channel, followed by a system of pressured conduits, supply Naples urban area [25].
West and Main channels) that convey the water towards a storage
channel, followed by a system of pressured conduits, supply Naples urban area [25]. tank. From this point, a gravity
The Acquaro-Pelosi
channel, springs
followed by asprings
system of area is located
pressured 3 kmsupply
conduits, uphill Naples
of Urciuoli
urban springs
area [25].(Figures 3 and 5), in
The Acquaro-Pelosi area is located 3 km uphill of Urciuoli springs (Figures 3 and 5), in a
a flat areaThe along the Sabato River (370–380 m a.s.l.). Uphill this area, the Sabato River and
is completely
flat area alongAcquaro-Pelosi
the Sabato River springs area is located 3 km uphill of Urciuoli springs (Figures
(370–380 m a.s.l.). Uphill this area, the Sabato River is 3completely 5), in
dry,
dry, and
a flatitarea
is only
alongafter rainy River
the Sabato days (370–380
that the m runoff
a.s.l.).can occur.
Uphill this Before
area, thethe tapping,
Sabato River numerous
is completely spring
and it is only after rainy days that the runoff can occur. Before the tapping, numerous spring outlets
outlets existed
dry, and it locally; the construction
is only after rainy days that of the
tunnels
runoffincan1934 has Before
occur. tappedthe most of these
tapping, springs,
numerous and only
spring
existed locally; thelocally;
construction of tunnels in 1934 has tapped most of these springs, and only a few
a fewoutlets
springexisted the construction
points discharge of tunnels in 1934 has tapped most of these springs, and only
after wet periods.
spring points
a few discharge
spring after wetafter
points discharge periods.
wet periods.
Figure 5. Aerial view of discharge area of Acquaro-Pelosi springs (dashed rectangle area of Figure 3),
Figure 5. Aerial view of discharge area of Acquaro-Pelosi springs (dashed rectangle area of Figure 3),
currently tapped by the drainage tunnels, and piezometers location (P1–P12).
currently tappedview
Figure 5. Aerial by the
of drainage
dischargetunnels, and piezometerssprings
area of Acquaro-Pelosi location (P1–P12).
(dashed rectangle area of Figure 3),
currently tapped
The two by the
tunnels drainage
are tunnels,
excavated and piezometers
into alluvial deposits, location
they are (P1–P12).
400 m long, and 4 to 7 m deep.
Water
The is drained
two tunnelsthrough loopholesinto
are excavated in the concrete
alluvial structurethey
deposits, of tunnels along
are 400 their bases
m long, and 4and
to sides.
7 m deep.
The
The two
tunnelstunnels
convey are excavated
water to a into
storage alluvial
tank, then deposits,
toward they
Urciuoli are 400
springs. m long,
Water is drained through loopholes in the concrete structure of tunnels along their bases and and 4 to 7 m deep.
sides.
Water
The is drained
tunnels through
convey loopholes
water to a storageintank,
the concrete structure
then toward of tunnels
Urciuoli springs.along their bases and sides.
The tunnels convey water to a storage tank, then toward Urciuoli springs.
Water 2018, 10, 501 8 of 21
Waterwells
Ten 2018, 10,are
x FOR PEER REVIEW
located inside the protected area, around the two drainage tunnels (Figure 8 of 5),
21 and
have depths between 30 m and 93 m. They cross the alluvial deposits constituted by sands, gravels and
Ten wells are located inside the protected area, around the two drainage tunnels (Figure 5), and
clayshave
(Figure 6). Inside each well, a piezometer tube has been placed, which is slotted at the bottom
depths between 30 m and 93 m. They cross the alluvial deposits constituted by sands, gravels
part; and
this clays
means(Figure
that the 6). water
Insidelevel
each measured into the piezometer
well, a piezometer tube has been has to be connected
placed, to theathydraulic
which is slotted the
headbottom
at the bottom part.
part; this means that the water level measured into the piezometer has to be connected to the
The entirehead
hydraulic water discharged
at the bottom part.by the Serino springs (mean of 2.25 m3 /s) is linked to a deep karst
substratum Theconnected
entire water todischarged
a rechargebyareathe Serino
of the springs
Terminio (mean of 2.25
massif. In m 3/s) is linked to a deep karst
the Acquaro-Pelosi spring area,
substratum connected to a recharge area of the Terminio massif. In
the karst substratum has not been reached by wells, and thus it lies below the deepest the Acquaro-Pelosi spring area,
piezometer
the
(>93.40 m).karst substratum has not been reached by wells, and thus it lies below the deepest piezometer
(>93.40 m).
Following the Geological map of Italy (ISPRA 2017), the karst substratum is believed to lie about
Following the Geological map of Italy (ISPRA 2017), the karst substratum is believed to lie about
400 m deep in the Sabato Valley, but this depth is considered variable, reaching locally the minimum
400 m deep in the Sabato Valley, but this depth is considered variable, reaching locally the minimum
in correspondence
in correspondence to springs.
to springs.InInany
anycase,
case,all
all spring watercomes
spring water comesfrom fromthethe upwelling
upwelling flux flux inside
inside the the
alluvial deposits of the Sabato plain, and these characteristics are further
alluvial deposits of the Sabato plain, and these characteristics are further detailed. detailed.
Figure 6. Well logs located in Figure 5; the ground elevation (m a.s.l.) and the depth (m), are shown
Figure 6. Well logs located in Figure 5; the ground elevation (m a.s.l.) and the depth (m), are shown
above and below the lithostratigraphic column, respectively.
above and below the lithostratigraphic column, respectively.
4. Data Analysis
4. Data Analysis
4.1. Hydrological Data and Correlations
4.1. Hydrological Data and Correlations
Spring discharge measurements have been carried out by automatic hydrometers located along
the hydraulic
Spring discharge system of tapping. The
measurements discharge
have is directly
been carried out measured
by automaticat the end of the Pelosi
hydrometers andalong
located
Acquaro drainage tunnels; the discharge of Urciuoli spring is indirectly obtained
the hydraulic system of tapping. The discharge is directly measured at the end of the Pelosi and from the discharge
measurements
Acquaro carried out
drainage tunnels; downstream.
the discharge of Urciuoli spring is indirectly obtained from the discharge
Figure 7 shows the hydrographs of the Pelosi, Acquaro and Urciuoli springs for the period 1963–
measurements carried out downstream.
2003; the minimum and maximum values occur in spring and autumn-winter seasons, respectively;
Figure 7 shows the hydrographs of the Pelosi, Acquaro and Urciuoli springs for the period
during droughts, the spring hydrographs are characterized by a continuous decreasing trend for the
1963–2003; the minimum
entire hydrological and
year, as maximum
occurred in 2002values
[28]. occur in spring and autumn-winter seasons,
respectively; during droughts, the spring hydrographs
Fiorillo et al. [25] described the long time series areofcharacterized by a continuous
the Serino spring discharge anddecreasing
its
trendrelationship
for the entire hydrological year, as occurred in 2002 [28].
with climate (rainfall and temperature) variables. By cross-correlation analyses, they
Fiorillo et al. dependence
found strong [25] described the long
of spring time series
discharge of the rainfall
cumulative Serino tospring
be overdischarge
150 days,and itstime-lag
with relationship
with climate (rainfall and temperature) variables. By cross-correlation analyses, they found strong
dependence of spring discharge cumulative rainfall to be over 150 days, with time-lag of 2 months [25].
Water 2018, 10, 501 9 of 21
Besides, they focused on the minimum amount of rainfall needed to avoid a groundwater drought,
thereafter, their statistical occurrence was estimated [25].
Piezometer level data considered in this study spanned the period of 1963–2003 as shown in
Figureof 7.2 For the piezometer
months [25]. Besides,P1,theyP2,focused
P3, P7,on P8,the
P9 minimum
and P10, dataamountare ofalmost continuous,
rainfall needed to whereas
avoid a for
piezometers P4, P5 and P6 data series present many discontinuities.
groundwater drought, thereafter, their statistical occurrence was estimated [25]. In particular, these discontinuities
occurred during
Piezometer lowlevel
flowdata
periods when these
considered in thispiezometers
study spanned arethe
used to tap
period of the water as
1963–2003 conveyed
shown into the
Figure
storage tank.7. For the piezometer P1, P2, P3, P7, P8, P9 and P10, data are almost continuous, whereas for
piezometers
Besides, the P4, P5 and P6
piezometric level data
canseries present during
be artesian many discontinuities.
the high flow In particular,
period, these in
as observed
discontinuities occurred during low flow periods when
piezometers P2, P3, P4, P5 and P7; these piezometers are also the deepest. these piezometers are used to tap the water
conveyed to the storage tank.
The P4 and P5 piezometers are also under artesian conditions also during low flow periods except
Besides, the piezometric level can be artesian during the high flow period, as observed in
for a few years of extreme drought (for example 2002). In the shallower piezometers P1, P9 and P10,
piezometers P2, P3, P4, P5 and P7; these piezometers are also the deepest.
the hydraulic head is always below the ground level.
The P4 and P5 piezometers are also under artesian conditions also during low flow periods
Figure
except for a few in
8 shows detail
years the hydraulic
of extreme droughthead of piezometers
(for example 2002). Induring the period
the shallower 1963–1969,
piezometers during
P1, P9
whichandrecords
P10, thearehydraulic
available almost
head continuously.
is always below the Thegrounddeepest
level. wells (P2, P3, P4, P5, P7) have a higher
hydraulicFigure
head, 8especially during
shows in detail thethe high flow
hydraulic headperiod of springs
of piezometers (March–July).
during During the
the period 1963–1969, low flow
during
which records are available almost continuously. The deepest wells (P2,
period of springs (September–November), these wells present a drop in the hydraulic head, which P3, P4, P5, P7) have a higher
hydraulic
decreases below head,
thatespecially during thewells
of the shallowest high flow
(P1, period
P10). of springs (March–July). During the low flow
period of springs (September–November),
As can be deduced from Figure 8, the distribution these wells present
of athedrop in the hydraulic
hydraulic head into head,thewhich
alluvial
decreases below that of the shallowest wells (P1, P10).
deposits is not phreatic, as it generally increases in depth. In fact, the deepest piezometers show
As can be deduced from Figure 8, the distribution of the hydraulic head into the alluvial deposits
a higher hydraulic head than the shallower piezometers, especially during the April–June period.
is not phreatic, as it generally increases in depth. In fact, the deepest piezometers show a higher
Thesehydraulic
characteristics are shown
head than in Figure
the shallower 9, with data
piezometers, plotted as
especially meanthe
during values in the period.
April–June period 1963–1969,
These
during which a continuous record occurred for all piezometers.
characteristics are shown in Figure 9, with data plotted as mean values in the period The dependence of1963–1969,
the hydraulic
head during
on the which
piezometer depth is
a continuous obvious
record (Figure
occurred for 9a), as well as the
all piezometers. The hydraulic
dependence head of range (Figure 9b).
the hydraulic
The piezometer
head on the P8, which is
piezometer drilled
depth mainly(Figure
is obvious in the 9a),
clayey deposits
as well at the bottom
as the hydraulic head part
range(Figure
(Figure 6),
9b).is not
shown The inpiezometer
Figures 8 and P8, which
9. is drilled mainly in the clayey deposits at the bottom part (Figure 6), is not
shown
To assessin Figures 8 and 9.transmissivity, permeability and specific yield, a multi-purpose aquifer
the aquifer’s
test was carried out onaquifer’s
To assess the pumping transmissivity, permeability
well and monitored and specific
piezometer (P11 yield,
andaP12,
multi-purpose
Figure 5). aquifer
test was carried out on pumping well and monitored piezometer (P11 and P12, Figure 5).
The test was carried out at a constant pumping rate of 38.5 L/s, for a duration of 68 h; drawdown
The test was carried out at a constant pumping rate of 38.5 L/s, for a duration of 68 h; drawdown
data were processed using the method of Neuman [29], enhanced by Tartakovsky and Neuman [30].
data were processed using the method of Neuman [29], enhanced by − Tartakovsky and Neuman [30].
3 m/s and the effective porosity,
Results are shown
Results in Table
are shown 1; the
in Table hydraulic
1; the hydraulic conductivity,KK== 33 ×
conductivity, 10−3 m/s
× 10 and the effective porosity,
neff =n0.1, are in line with lithostratigraphic and granulometric features.
eff = 0.1, are in line with lithostratigraphic and granulometric features.
2000
Urciuoli Aquaro Pelosi
discharge, l/sec
1500
1000
500
0
385 P1 (51 m depth)
hydraulic head
380 ground level

375
370
365
385 P2 (75 m depth)
hydraulic head
380
375
370
365
385 P4 (77 m depth)
hydraulic head
380
375
370
365
Figure 7. Cont.
Water 2018, 10, 501 10 of 21
Water
Water 2018,
2018, 10,
10, xx FOR
FOR PEER
PEER REVIEW
REVIEW 10
10 of
of 21
21
385 P5 (74.5 m depth)
headhead
385
380 P5 (74.5 m depth)
385 P5 (74.5 m depth)
hydraulic
head 380
375
380
hydraulic
375
370
hydraulic
375
370
365
370
385
365 P6 (35 m depth)
365
385
380 P6 (35 m depth)
headhead
385 P6 (35 m depth)

380
375
hydraulic
380
head
375
370
hydraulic
375
hydraulic
370
365
370
365
365
Figure 7. Discharge of Urciuoli, Acquaro and Pelosi springs, and piezometric levels measured in P1,
Figure 7. Discharge of Urciuoli, Acquaro and Pelosi springs, and piezometric levels measured in P1,
Figure
P2, P4, 7.
Figure 7. Discharge
P5, P6 and P7,
Discharge of Urciuoli,
of during the
Urciuoli, Acquaro
period and
Acquaro and Pelosi
Pelosi springs,
1963–2003; and
and piezometric
values above
springs, levels
levels measured
the thick horizontal
piezometric in P1,
line (ground
measured
P2, P4,
P2,P5, P6 andand
P7, P7,
during thethe
period 1963–2003; valuesabove
above the thick horizontal linein(ground
P1,
P2, P4,
level) P5,
P5, P6
highlights
P4, P7, during
artesian
P6 and the period
period 1963–2003;
conditions.
during 1963–2003; values
values above the
the thick
thick horizontal
horizontal line
line (ground
(ground
level)level)
highlights artesian
highlights conditions.
artesian conditions.
level) highlights artesian conditions.
384
384
384
382
382
382
380 P2
P5
380 P2 P7
380
378 P2
P5
P5 P7
m a.s.l.
P10 P3
378 P7
378
376
a.s.l.
P10 P3P4
ma.s.l.
P10 P3
head,
376
376
374 P4
P4
head,m
hydrailic
head,
374
374
372
hydrailic
hydrailic
372
372
370
P1
370 P9
370
368 P1 P6
P1
P9
368 P9 P6
368
366 P6
Jan-63Jan-63
Jan-64Jan-64
Jan-65Jan-65
Jan-66Jan-66
Jan-67Jan-67
Jan-68Jan-68
Jan-69Jan-69
Jan-70Jan-70
366
366
Jan-63
Jan-64
Jan-65
Jan-66
Jan-67
Jan-68
Jan-69
Jan-70
Figure 8. Hydraulic head during the period of 1963–1969, characterized by almost continuous records.
FigureFigure
Figure 8.
8. Hydraulic
8. Hydraulic head
Hydraulic head
head during
during the
the
during period
period
the of
periodof 1963–1969,
1963–1969, characterized
of1963–1969, characterized
characterized by almost
byby continuous
almost
almost records.
continuous
continuous records.
records.
382 10
mean of annual maxima
382
a) 10
b)
382
380 mean
mean of
of annual
annual minima
maxima P2 10
mean of annual maxima a)
a) b)
b)
mean of annual minima P2 8
380 mean of annual minima P2P3 P7 P2 P3
m a.s.l.
380
378
mm m
8
y = 0,1439x + 366,43 P7 8 P2 P3 P7
range,
P5
a.s.l.
P3 P7 P2 P4
P3
378
mma.s.l.
R² = 0,79 P3 P4 6 y = 0,1173x - 2,4417

378
head,
376 y = 0,1439x + 366,43 P7

range,
R² = 0,83 P5 P4
y = 0,1439x + 366,43 P5 P7
headrange,
R² = 0,79 P4 P5 P4
6
head
P10 R² = 0,79 P4 y = 0,1173x - 2,4417

head,
376 6 y = 0,1173x - 2,4417

R² = 0,83
head,
376 P5
hydraulic
374 P5 R² = 0,83
P10 4
hydraulic
head
P10 P10
hydraulic
374 P9 P12 P7
hydraulic
374 P1 4
372
hydraulic
4 P10
hydraulic
P9 P7 P9
P1
P9 P1 P12 P10
372 P6 P1 P12 P7 2
372 y = 0,0266x + 368,87
370 P4 P6 P1 P9
P9
P1 R² = 0,24 2
P6 P9 P1 P5 y = 0,0266x + 368,87 2
370 P6
P6 P4 y = 0,0266x + 368,87 P6 P1
370 P4 R² = 0,24 P6 P1
368 P9 P5 R² = 0,24 0
0 20 P6
P6 40 P9 60 P5 80 100 120 0 20 40 60 80 100 120
368 0
368 0 20 40 well depth,
60 m 80 100 120 00 20 40 well depth,
60 m 80 100 120
0 20 40 60 80 100 120 0 20 40 60 80 100 120
well depth, m well depth, m
(a) depth, m
well (b)
well depth, m
(a)
(a) (b)
(b)
Figure 9. (a) Correlation between well depth and hydraulic head during high (mean of annual
Figure
maxima, 9. (a)
(a) Correlation between well depth and hydraulic head of during high (mean of
of annual
Figure 9. (a)9.period
Figure betweenbetween
Correlation
Correlation 1963–1969) anddepth
betweenwell
well low flow
depth andperiods
and (mean
hydraulic
hydraulic head
head annual
during
duringminima,
high period
(mean
high (mean 1963–
annual
of annual
maxima,
1969);
maxima,(b) period between
Correlation
period 1963–1969)
between
between well
1963–1969) and
depth low
and flow periods
hydraulic (mean
head rangeof annual
(mean minima,
values, period
period 1963–
between
maxima, period between 1963–1969) andand
lowlowflowflow periods
periods (mean
(mean of of annual
annual minima,period
minima, period1963–1969);
1963–
1969); (b)
1963–1969).
1969); (b) Correlation
Correlation between
between well
well depth
depth and
and hydraulic
hydraulic head
head range
range (mean
(mean values,
values, period
period between
between
(b) Correlation between well depth and hydraulic head range (mean values, period between 1963–1969).
1963–1969).
1963–1969).
Water 2018, 10, 501 11 of 21
Water 2018,
Table 10, x FOR
1. Main PEER REVIEWcharacteristics of soils (Unified Soil Classification System—USCS)
granulometry 11 of 21
and
some hydraulic parameters obtained by pumping test carried out in well P12. Diameter, φ of soil:
Table 1. Main granulometry characteristics of soils (Unified Soil Classification System—USCS) and
gravel, φ > 2 mm; sand, 0.06 mm < φ < 2 mm; silty and clay, φ < 0.06 mm. T, transmissivity; K,
some hydraulic parameters obtained by pumping test carried out in well P12. Diameter, φ of soil:
hydraulic conductivity; neff , effective porosity.
gravel, φ > 2 mm; sand, 0.06 mm < φ < 2 mm; silty and clay, φ < 0.06 mm. T, transmissivity; K,
hydraulic conductivity; neff, effective porosity. 2 (2) (2) (2)
Strata Soil Type USCS Thick. m Gravel % Sand % Silt & Clay % T m /s K m/s neff-
1 Strata Soil
OLType USCS 1.0Thick. m Gravel % Sand % Silt & Clay % T m2/s (2) K m/s (2) neff- (2)
2 1 GW OL
(1) 3.1 1.0
3 2 SPGW (1) 1.3 3.1 32 66 2
4 3 GW (1)SP 3.3 1.3 32 66 2
5 4 GWGW (1) 3.3 3.3 53 43 4
6 5 GW (1)GW 4.5 3.3 53 43 4
7 6 SWGW (1) 8.5 4.5 35 58 7
8 7 SM SW 5.0 8.5 10 35 78 58 12 7
8 SM 5.0 10 78 12 ≈5 × 10−2 ≈3 × 10−3 ≈0.1
(1)
Note: Laboratory test missing, classification deducted from analogy with ≈3 × 10
Results ≈0.1
from pumping × 10(2)
stratum≈5n.5; −2 −3
test inNote:
the well P12 (30 m depth)
(1) Laboratory test missing, classification deducted from analogy with stratum n.5; (2) Results
from pumping test in the well P12 (30 m depth)
4.2. Physical and Chemical Characteristics
4.2. Physical and Chemical Characteristics
The total dissolved solids values are 221 mg/Kg and 201 mg/Kg for Acquaro-Pelosi and Urciuoli
The total dissolved solids values are 221 mg/Kg and 201 mg/Kg for Acquaro-Pelosi and Urciuoli
respectively and it results in the poor mineralization of water. The main ions are the++calcium
respectively and it results in the poor mineralization of water. The main ions are the calcium ( Ca,
++
( Ca, 65 mg/L) and the hydrogen carbonate (− HCO3 , 240 mg/L) according to origin of water
65 mg/L) and the hydrogen carbonate (−HCO3, 240 mg/L) according to origin of water for karst
for karst aquifer.
aquifer.
A setAofsetmonthly temperature
of monthly temperatureandandelectrical
electrical conductivity measurements
conductivity measurements have
have beenbeen carried
carried out out
within the period of March 2015–July 2016, for Acquaro, Pelosi and Urciuoli springs (Figure
within the period of March 2015–July 2016, for Acquaro, Pelosi and Urciuoli springs (Figure 10a). 10a).
600 16 900 45
Elec. Cond. Acquaro discharge
550 Acquaro spring 800 40
15 radon
500 temperature
700 35
450 14
discharge (l/s)
400 600 30
Rn (Bq/l)
E.C. (μS/cm 20 °C))
13
Temperature, °C
350 500 25
300 12
400 20
250
222
11 300 15
200
150 10 200 R=0.91 10
100
9 100 5
50
0 8
0 0
600 16 900 45
550 Pelosi spring Elec. Cond. Pelosi discharge
15 800 radon 40
500 temperature
450 14 700 35
discharge (l/s)
400 600 30
E.C. (μS/cm 20 °C))
13
Rn (Bq/l)
Temperature, °C
350
500 25
300 12
250
400 20
222
11
200 300 15
150 10 200 10
100 R=0.82
9 100 5
50
0 8 0 0
600 16 2000 45
550 Elec. Cond. Urciuoli discharge
Urciuoli spring 15 1800 40
500 temperature radon
1600 35
450 14
400 1400 30
Discharge (l/s)
E.C. (μS/cm 20 °C))
13
Temperature, °C
350
222 Rn (Bq/l)
1200 25
300 12
1000 20
250
11
200 800 15
150 10
600 10
100 R=0.87
9 400 5
50
0 8 200 0
Feb-15
Feb-16
Aug-15
Aug-16
Feb-15
Feb-16
Aug-15
Aug-16
(a) (b)
Figure 10. (a)

Figure 10. Temperature and
(a) Temperature andelectrical
electricalconductivity, periodbetween
conductivity, period between March
March 2015–July
2015–July 2016,2016,
for for
Acquaro,
Acquaro, PelosiPelosi
andand Urciuoli
Urciuoli springs,period
springs, period between
between February
February2015–July 2016;
2015–July (b) (b)
2016; Radon (222 Rn)
(222Rn)
Radon
activity
activity and and spring
spring discharge,
discharge, periodbetween
period betweenFebruary
February 2015–July
2015–July2016.
2016.
Water 2018, 10, 501 12 of 21
The electrical conductivity appears almost constant for the Pelosi spring, where a value of
350 µS/cm is observed for the entire period. In the Acquaro and Urciuoli springs a range between 450
and 350 µS/cm is recorded in the same period. The temperature shows constant ◦
The electrical conductivity appears almost constant for the Pelosi spring,values
where around
a value10.5of C
(Acquaro and Pelosi) and 11 ◦
350 μS/cm is observed for C (Urciuoli);
the while
entire period. In during the recession
the Acquaro period
and Urciuoli (August–November)
springs a range between an
increase of 1–2 degrees is observed.
450 and 350 μS/cm recorded in the same period. The temperature shows constant values around
In the°C
10.5 same periodand
(Acquaro (February 2015–July
Pelosi) and 2016) a monthly
11 °C (Urciuoli); 222 Radon
while during the concentration
recession periodmeasurements
(August–
November)
were carried out anforincrease of 1–2 degrees
the Acquaro, Pelosi is observed.
and Urciuoli springs. Figure 10b shows the results obtained
from 222 RnInactivity
the same period (February 2015–July values,
2016) a monthly 222Radon concentration measurements
together with the discharge where a clear and similar trend can be observed.
were carried
A significantly out for linear
positive the Acquaro, Pelosi and
relationship Urciuoli
results fromsprings. Figure 10b
a comparison shows the
between results
spring obtained and
discharge
222 Rn from
222Rn activity together with the discharge values, where a clear and similar trend can be
activity, indicating the strong control of the hydrological cycle on the radon concentration.
observed. A significantly positive linear relationship results from a comparison between spring
discharge and 222Rn activity, indicating the strong control of the hydrological cycle on the radon
5. Results
concentration.
The correlations of Figure 9 indicate that the hydraulic behavior of the alluvial aquifer is characterized
by an 5.upwelling
Results flux during the high flow period, where the slope of the interpolation line (equation
y = 0.1439xThe + 366.43, Figureof9a)
correlations can be
Figure assumedthat
9 indicate as athefirsthydraulic
coarse estimation
behavior of of the
the hydraulic is ia .
gradient,
alluvial aquifer
Due tocharacterized
the non-isotropic behavior flux
by an upwelling of the alluvial
during the deposits,
high flow this hydraulic
period, where the gradient
slope ofhastheto be assumed as
interpolation
line value
the mean (equation y =ascendant
of the 0.1439x + flow 366.43, Figure
into 9a) deposits
alluvial can be assumed as ahigh
during the first flow
coarse estimation of the
period.
hydraulic gradient, i . Due to the non-isotropic behavior
During the low flow period, the behavior appears more complicated. Considering
a of the alluvial deposits, this hydraulic
the mean of
annual minima (Figure 9a), an ascendant hydraulic gradient can be roughly estimatedduring
gradient has to be assumed as the mean value of the ascendant flow into alluvial deposits by the theslope
high flow period.
of the interpolation line (equation y = 0.0266x + 368.43; Figure 9a), but the correlation is weaker. In
During the low flow period, the behavior appears more complicated. Considering the mean of
fact, during the recession period, the hydraulic head of some deep piezometers decreases up to values
annual minima (Figure 9a), an ascendant hydraulic gradient can be roughly estimated by the slope
lowerof than
the shallower
interpolationpiezometers
line (equation (Figure 8). Thus,
y = 0.0266x during
+ 368.43; the low
Figure flowthe
9a), but period, the flux
correlation could have
is weaker. In a
descendant component
fact, during into some
the recession period,portions of thehead
the hydraulic alluvial
of somedeposits; under such
deep piezometers conditions,
decreases up tothe alluvial
values
aquifer can than
lower feedshallower
the springs (drainage(Figure
piezometers tunnels).
8). Thus, during the low flow period, the flux could have
Another
a descendantmaincomponent
consideration into can
some beportions
done asofshown Figuredeposits;
the alluvial 9b: the hydraulic
under suchhead range appears
conditions, the
alluvial dependent
to be clearly aquifer can feed
on the the depth
springsof(drainage
the well,tunnels).
indicating a wider oscillation of the hydraulic head
into deepest Another main consideration can be done as shown Figure 9b: the hydraulic head range appears
deposits.
to be clearly dependent
To investigate the aquifer on the depth of and
behavior the well, indicating“reverse”
its possible a wider oscillation
behaviorof(descendant
the hydraulic flux),
head the
into deepest deposits.
difference in hydraulic head between P7 and P1 (P7−P1), and between P2 and P1 (P2−P1) have been
To investigate the aquifer behavior and its possible “reverse” behavior (descendant flux), the
plotted as shown in Figure 11; these piezometers have almost complete records in the period between
difference in hydraulic head between P7 and P1 (P7−P1), and between P2 and P1 (P2−P1) have been
1963–2003,
plottedand allowinfor
as shown Figure comparison a deep (P7have
11; these piezometers and almost
P2) and shallow
complete (P1) in
records piezometer. Observing
the period between
the difference between P7 and P1 (P7 − P1), it seems that the “reverse”
1963–2003, and allow for comparison a deep (P7 and P2) and shallow (P1) piezometer. Observing period occurs every year
thesince
1987; whereas, observing
difference between P7 the and difference
P1 (P7−P1), it between
seems that P2theand P1 (P2−
“reverse” P1), the
period “reverse”
occurs every yearperiod
sincewould
1987; be
limited to the drought
whereas, observing ofthe
2002 and to other
difference betweenfewP2andandshort periods.
P1 (P2−P1), the Figure
“reverse” 11 period
indicates
wouldthatbethe “reverse”
limited
to the
condition drought
would of 2002
be very limitedand to inother
time few
and and
space,short
andperiods. Figure 11
that probably indicates
it is that the “reverse”
only a misleading result due
condition would
to a comparison of thebe very limited
hydraulic head in measured
time and space,
into twoand distinct
that probably it is onlyand
piezometers, a misleading
not alongresult
the same
due to a comparison of the hydraulic head measured into two distinct
vertical. Observing Figure 10a, the slight temperature increase between August and November piezometers, and not along the could
same vertical. Observing Figure 10a, the slight temperature increase between August and November
indicate the possible mixing between deep (ascendant flow) and shallow water of alluvial aquifer, but
could indicate the possible mixing between deep (ascendant flow) and shallow water of alluvial
the electrical conductivity does not change.
aquifer, but the electrical conductivity does not change.
10
P7-P1
hydraulic head difference, m
8
ascendant flow
P2-P1
6
4
2
0
descendant
-2
-4
Jan-63
Jan-65
Jan-67
Jan-69
Jan-71
Jan-73
Jan-75
Jan-77
Jan-79
Jan-81
Jan-83
Jan-85
Jan-87
Jan-89
Jan-91
Jan-93
Jan-95
Jan-97
Jan-99
Jan-01
Jan-03
Figure 11. Hydraulic head difference between piezometers P7 and P1 (P7−P1) and piezometers P2
Figure 11. Hydraulic head difference between piezometers P7 and P1 (P7−P1) and piezometers P2
and P1 (P2−P1); positive and negative value means ascendant and descendant flow, respectively.
and P1 (P2−P1); positive and negative value means ascendant and descendant flow, respectively.
Water 2018, 10, 501 13 of 21
To reduce the ambiguity observed

To observed in Figure 11, the ascendantascendant hydraulic
hydraulic gradient
gradient hashas been
been
estimated by considering
estimated consideringtwo twoclose
closepiezometers,
piezometers,characterized
characterized byby
different
different depths. ForFor
depths. thisthis
reason, the
reason,
analysis
the is restricted
analysis to piezometer
is restricted to piezometer P5 P5
and P6;P6;
and P4P4could
couldreplace
replaceP5P5(it(it
is is
farfar
from
fromP6P6
asasP5,
P5,and
andhas
hasa
adepth
depthcomparable
comparable with
withP5)P5)
providing
providingsimilar results.
similar TheThe
results. estimated
estimatedascendant
ascendant hydraulic gradient
hydraulic into
gradient
alluvial
into deposits
alluvial is shown
deposits in Figure
is shown 12, where
in Figure it is well-correlated
12, where with thewith
it is well-correlated discharge. In this case,
the discharge. the
In this
hydraulic
case, gradient gradient
the hydraulic is alwaysispositive
always (ascendant flow), and
positive (ascendant it increases
flow), or decreases
and it increases during the
or decreases high
during
flowhigh
the and flow
low flow
and periods,
low flow respectively. It has to be specified
periods, respectively. It has to that the actualthat
be specified hydraulic gradient
the actual could
hydraulic
be up higher
gradient could with respect
be up higherto with
that computed because
respect to that of the uncertainty
computed because of the in the hydraulicinisolation
uncertainty of the
the hydraulic
piezometric
isolation tube.
of the piezometric tube.
1000 0.21
Aquaro-Pelosi spring discharge a) 0.18
800 vertical hydraulic gradient, P5-P6
0.15
hydraulic gradient
600
discharge; l/sec
0.12
0.09
400
0.06
200
0.03
0 0
(a)
16000
b)
12000
Q/ia; l/sec
8000
4000
0
Jan-63
Jan-64
Jan-65
Jan-66
Jan-67
Jan-68
Jan-69
(b)
Figure 12.
Figure (a)Acquaro-Pelosi
12.(a) Acquaro-Pelosispringspringdischarge
dischargeand
andvertical
vertical(ascendant)
(ascendant)hydraulic
hydraulicgradient,
gradient,computed
computed
between P5 and P6 (60 m distance from each other; Figure 5); (b) ratio between discharge
between P5 and P6 (60 m distance from each other; Figure 5); (b) ratio between discharge and hydraulic
and
gradient (Q/i a ).
hydraulic gradient (Q/ia).
Plotting
Plotting the
the ratio
ratio between
betweenthe
thespring
springdischarge
dischargeand
andhydraulic
hydraulicgradient,
gradient,Q/iQ/i
a (Figure 12b),
a (Figure a
12b),
constant path
a constant can
path canbebeobserved,
observed,which
whichisisinterrupted
interruptedby
byaapeak
peakduring
duringlow
low flow
flow periods.
periods. Following
Following
the
the Darcy
Darcy law:
law:
Q
=K×A (3)
ia = × (3)
with K being the hydraulic conductivity, and A the area, the plot of Figure 12b means that during low
with K being the hydraulic conductivity, and A the area, the plot of Figure 12b means that during low
flow periods, the area, A, has to increase proportionally to ratio Q/ia , as the hydraulic conductivity, K,
flow periods, the area, A, has to increase proportionally to ratio Q/ia, as the hydraulic conductivity,
can be considered constant. It has to be specified that the area, A, lies on the horizontal plane in this case,
K, can be considered constant. It has to be specified that the area, A, lies on the horizontal plane in
as the flow lines are vertical under the ascendant hydraulic flow into the aquifer. The area, A, increases
this case, as the flow lines are vertical under the ascendant hydraulic flow into the aquifer. The area,
or decreases under low and high ascendant hydraulic gradients, respectively; and to investigate this
A, increases or decreases under low and high ascendant hydraulic gradients, respectively; and to
behavior, Figure 1b shows the effect of the drainage tunnel on the water table. The natural ascendant
investigate this behavior, Figure 1b shows the effect of the drainage tunnel on the water table. The
hydraulic flow causes the start-up of springs on the ground level; the drainage tunnel excavation
natural ascendant hydraulic flow causes the start-up of springs on the ground level; the drainage
induces a water table drop; the distance R defines the wideness of the influenced zone of the aquifer.
tunnel excavation induces a water table drop; the distance R defines the wideness of the influenced
To estimate the R-value under an ascendant hydraulic gradient condition, Equations (1) and (2) have
zone of the aquifer. To estimate the R-value under an ascendant hydraulic gradient condition,
been used.
Equations (1) and (2) have been used.
Figure 13 shows the plot of Equation (2), where the potential hydraulic gradient, iT, depends on
the depth, Z, and distance, L, from the drainage/spring, distinguishing the condition under high
Water 2018, 10, 501 14 of 21
Figure 13 shows the plot of Equation (2), where the potential hydraulic gradient, iT , depends
on the depth, Z, and distance, L, from the drainage/spring, distinguishing the condition under high
(Figure 13a) and low (Figure 13b) ascendant hydraulic gradient, respectively; the hydraulic gradient
(Figure 13a) and low (Figure 13b) ascendant hydraulic gradient, respectively; the hydraulic gradient
values, i = 0.15 and i = 0.026, are the mean of the annual highest and lowest values in the period
values, iaa = 0.15 and iaa = 0.026, are the mean of the annual highest and lowest values in the period
1963–1969 (Figure 12a), and are similar to the slope of the interpolation line of Figure 9a.
1963–1969 (Figure 12a), and are similar to the slope of the interpolation line of Figure 9a.
(a) (b)
Figure 13. Potential hydraulic gradient, iTT,, induced

induced by by drainage
drainage system
system at depth D, computed along
the vertical, Z, with distance L from
from drainage (Figure 1b): (a) (a) aquifer
aquifer with
with natural
natural ascendant
ascendant hydraulic
hydraulic
gradient, i = 0.15;
gradient, iaa = 0.15; (b) aquifer with natural ascendant hydraulic gradient,
ascendant hydraulic gradient, iaa = i = 0.026. The drainage
toward the
toward the tunnel/spring
tunnel/springoccurs
occursininthe fieldiTiT> i>a; ifor
thefield iT <iTia<the
a ; for ia the ascendant
ascendant fluxflux is not
is not influenced
influenced by
by the
the drainage.
drainage.
As
As shown
shown in in Figure
Figure13, 13,the
thepotential
potentialhydraulic
hydraulic gradient,
gradient,iT, itend to ia when Z increases. The
T , tend to ia when Z increases.
drainage towards the tunnel/spring occurs for i > i ; in this case
The drainage towards the tunnel/spring occurs for iT > ia ; in this case the ascendant
T a the ascendant flowflow
converges to the
converges to
tunnel/spring. In the field i < i , the ascendant flow continues its vertical path.
the tunnel/spring. In the field iT < ia , the ascendant flow continues its vertical path. The horizontal
T a The horizontal line iT
=line
ia isiTa=boundary limit, splitting
ia is a boundary the zones
limit, splitting the descripted
zones descripted above;above;
along thisalong line
thisdoes
linethe condition
does L≡R
the condition
occur. In Figure 13a, for Z = 100 m the condition i = i occurs when L ≈ 100
L ≡ R occur. In Figure 13a, for Z = 100 m the condition ia = iT occurs when L ≈ 100 m; in Figure 13b,
a T m; in Figure 13b, for Z=
100
for Z m,=the100same condition
m, the occurs when
same condition 200when
occurs < L < 500 200 m; < Lthese
< 500theoretical
m; these results are in
theoretical accordance
results are in
with the actual
accordance with results
the actualof results
Q/i ratio (Figure
of Q/i ratio12b),
(Figure where 12b),the area,the
where A,area,
has A,to has
increase underunder
to increase a lowa
hydraulic
low hydraulic gradient.
gradient.
Equalizing
Equalizing Equations
Equations (1) and (2)
(1) and (2) by
by fixing
fixing iiaa == iiTT, ,the
thedistance
distanceLLPPwill willcoincide
coincidewithwithR,R,and
andthus:
thus:

1 1
s
≡ = 1
LP ≡ R = D 1 ++1 1
D 1++ 1 +2
+ 2ZP
(4).
(4)
ia ia
Figure 14 shows the dependence of the zone influenced by the drainage system (R-value) in the
Figure 14 shows the dependence of the zone influenced by the drainage system (R-value) in the
function of Z, for a different ascendant hydraulic gradient, ia, including the maximum and minimum
function of Z, for a different ascendant hydraulic gradient, ia , including the maximum and minimum
observed (0.15 and 0.026). This relationship highlights how the depth, Z, of the ascendant flow lines
observed (0.15 and 0.026). This relationship highlights how the depth, Z, of the ascendant flow lines
control the R-value. As the depth of the karst substratum can be used as the maximum Z value, it can
control the R-value. As the depth of the karst substratum can be used as the maximum Z value, it
be estimated by comparing two different scenarios in Equation (4). In particular, as shown in Figure
can be estimated by comparing two different scenarios in Equation (4). In particular, as shown in
12b, the area, A, increases n times from high flow to low flow condition, by the following:
Figure 12b, the area, A, increases n times from high flow to low flow condition, by the following:
( ) × ( ) (5)
R(ia )min ≈ n × R(ia )max (5)
and from Equation (4):

1 1
1+ 1+ +2
( ) ( )
(6)

1 1
= × 1+ 1+ +2
( ) ( )
Water 2018, 10, 501 15 of 21
and from Equation (4):

s h s h
i i
1 1 1 1
D 1+ i D 1+ (i a )min
+ 2ZP = n× D 1+ i (i a ) D 1+ (i a )max
+ 2ZP (6)
Water 2018, 10, x FOR(iaPEER
)min REVIEW max 15 of 21
the value Z ZP deducted provides the depth depth of

of the
the karst substratum,ZZKK ≡
karstsubstratum, ≡ZZPP.. For
For nn == 2.5, ZKK ==353
2.5, Z 353 m.
m.
This value ofZZKKisiscompatible
value of compatiblewithwith the geological and hydrogeological features of the study
the geological and hydrogeological features of the study area; the area;
the correspondent
correspondent values values R, 169
of R,ofare are 169
andand
422 422
m inmconditions
in conditions of high
of high flowflow
(ia = (i0.15)
a = 0.15)
and and low flow
low flow (ia =
(i a = 0.026)
0.026) as shown
as shown in Figure
in Figure 14. 14.
R, m
169 422
0 250 500 750 1000
0
200
ZK=353 m
400
Z, m
600
800 1 1
≡ = 1+ 1+ +2
D=5 m
1000
Figure 14. Width of the influenced zone, R, induced by drainage system at depth D, in relation to
Figure 14. Width of the influenced zone, R, induced by drainage system at depth D, in relation to
depth,
depth, Z,
Z, and for aa different
and for different ascendant
ascendant hydraulic gradient, iia.. Z
hydraulic gradient, ZK isisthe
thedepth
depthof
of karst
karst substratum.
substratum.
a K
6. Discussion
6. Discussion
The relationship between groundwater and surface water has been discussed by Sophocleous
The relationship between groundwater and surface water has been discussed by Sophocleous [31],
[31], and by Barthel and Banzhaf [32] at a regional scale. In general, the presence of an ascendant flux
and by Barthel and Banzhaf [32] at a regional scale. In general, the presence of an ascendant flux
could be very difficult to detect in many hydrogeological contexts. For example, in an alluvial plain
could be very difficult to detect in many hydrogeological contexts. For example, in an alluvial plain
characterized by sand deposits and ascendant water flux with vertical hydraulic gradient ia = 0.01,
characterized by sand deposits and ascendant water flux with vertical hydraulic gradient ia = 0.01, two
two distinct piezometers, close together, but having 10 m depth difference, would measure a
distinct piezometers, close together, but having 10 m depth difference, would measure a hydraulic head
hydraulic head difference of 0.1 m. This hydraulic head difference could be misleading of a
difference of 0.1 m. This hydraulic head difference could be misleading of a prevalently horizontal flow,
prevalently horizontal flow, and similar considerations can also be provided for other aquifer types,
and similar considerations can also be provided for other aquifer types, karst included. The possibility
karst included. The possibility of measuring and quantifying the ascendant hydraulic gradient
of measuring and quantifying the ascendant hydraulic gradient depends on its magnitude and on the
depends on its magnitude and on the depth of the piezometers.
depth of the piezometers.
In some areas the ascendant flux has been detected due to a deep geological survey. For example,
In some areas the ascendant flux has been detected due to a deep geological survey. For example,
in the Crni Timok valley, Serbia, geophysical surveys by the spontaneous potential method (SP) has
in the Crni Timok valley, Serbia, geophysical surveys by the spontaneous potential method (SP)
allowed for the identification of the zone characterized by an ascendant flux in the karstified
has allowed for the identification of the zone characterized by an ascendant flux in the karstified
limestone buried under alluvial deposits [33]. Besides, in the Duna-Tisza interfluves, Hungary,
limestone buried under alluvial deposits [33]. Besides, in the Duna-Tisza interfluves, Hungary, Szőnyi
Szőnyi and Tóth [34] reconstructed the ascendant flux from deep water and hydrocarbon wells,
and Tóth [34] reconstructed the ascendant flux from deep water and hydrocarbon wells, which is
which is responsible for severe problems of soil and wetland salinization.
responsible for severe problems of soil and wetland salinization.
In many karst aquifers of Southern Apennines the water table of the saturated zone is flat, and
In many karst aquifers of Southern Apennines the water table of the saturated zone is flat, and
it is generally characterized by hydraulic gradient lower than 1% [35]; these conditions favor a deep
it is generally characterized by hydraulic gradient lower than 1% [35]; these conditions favor a deep
water circulation, with flow lines converging to a spring outlet and characterized by vertical path
water circulation, with flow lines converging to a spring outlet and characterized by vertical path
(Figure 15a). The development of a conduit net into a karst aquifer leads to a hierarchical flow [36],
(Figure 15a). The development of a conduit net into a karst aquifer leads to a hierarchical flow [36],
which controls the water path; the ascendant flux can lead to the formation of subvertical conduits,
which controls the water path; the ascendant flux can lead to the formation of subvertical conduits,
and many karst springs are fed by a typical subvertical conduit, as the Vauclusian types. In these
and many karst springs are fed by a typical subvertical conduit, as the Vauclusian types. In these cases,
cases, tectonic shifts between karst terrains and aquiclude favor the formation of the subvertical
conduit [37]. Many karst springs of Southern Italy can be associated with the scheme of Figure 15a,
as Caposele spring of Picentini mountains or Boiano springs of Matese massif, where the conduit net
(not drawn) leads the flow lines.
Water 2018, 10, 501 16 of 21
tectonic shifts between karst terrains and aquiclude favor the formation of the subvertical conduit [37].
Many karst springs of Southern Italy can be associated with the scheme of Figure 15a, as Caposele
spring of Picentini mountains or Boiano springs of Matese massif, where the conduit net (not drawn)
leads the flow
Water 2018, lines.
10, x FOR PEER REVIEW 16 of 21
(a)
(b)
(c)
Figure
Figure 15. Ascendant flux
15. Ascendant fluxconnected
connectedtotodifferent
differenthydrogeological
hydrogeologicalconditions
conditionsofof the
the discharge
discharge zone
zone of
of karst
karst aquifers.
aquifers. (a) (a)
flowflow lines
lines tend
tend to vertical
to be be vertical as consequences
as consequences of deep
of deep groundwater
groundwater flowflow
andand
flat
flat
waterwater table;
table; (b) (b)
no no karst-terrains
karst-terrains between
between thetherecharge
rechargeand
anddischarge
discharge area
area favors
favors groundwater
groundwater
syphoning; the hydraulic head difference, h
syphoning; (c) the hydraulic head difference, h1, is needed to allow the rising of groundwater through
(c) 1 , is needed to allow the rising of groundwater through
the
the porous
porous medium
medium from from buried
buried karst
karst substratum.
substratum.
Many other
other karst
karstsprings
springsemerge
emergefrom
from carbonate
carbonate rock
rockblock tectonically
block joined
tectonically to a wider
joined karst
to a wider
area (Figure
karst 15b). The
area (Figure presence
15b). of no-karst
The presence of terrains,
no-karstwhich act as
terrains, aquiclude
which act asoraquiclude
aquitard between
or aquitardthe
main aquifer
between and the
the main rockand
aquifer block,
thefavors the development
rock block, of a deep groundwater
favors the development flow, converging
of a deep groundwater flow,
toward the toward
converging rock block, which
the rock is which
block, interested in ascendant
is interested groundwater
in ascendant flux. Influx.
groundwater Southern Italy,
In Southern
Capovolturno
Italy, spring
Capovolturno of Meta
spring mountains
of Meta or Grassano
mountains or Grassano spring of Matese
spring massif
of Matese is associated
massif withwith
is associated the
scheme
the schemeof Figure 15b.15b.
of Figure
In many other
other cases,
cases,the
thecarbonate
carbonaterock-block
rock-blockis is missing
missing in in
outcropping
outcropping andand lieslies
below the the
below no-
karst terrains
no-karst (Figure
terrains 15c);15c);
(Figure springs can can
springs originate through
originate ascendant
through fluxes
ascendant into into
fluxes porous deposits,
porous and
deposits,
and faults allow for crossing the aquiclude/aquitard above the karst aquifer. Serino springs can be
faults allow for crossing the aquiclude/aquitard above the karst aquifer. Serino springs can
associated with the scheme of Figure 15c, which is common common also to other other karst
karst springs
springs of of Southern
Southern
springs along
Italy as Cassano springs along the
the eastern
eastern side
side of
of Terminio
Terminiomassif
massif(Figure
(Figure3).3).
The hydrogeological cross-section of Figure 16 shows how the flux from the recharge area is
forced under the aquiclude of flysch sequences; in the Sabato plain, the presence of springs has been
explained by
explained byaaburied
buriedkarst
karstsubstratum,
substratum,covered
covered byby thick
thick alluvial
alluvial deposits.
deposits. In the
In the alluvial
alluvial deposits,
deposits, the
the flux is widespread and circulates in a porous medium; most of the upwelling flux outcrops along
the Sabato River, located on the shorter and vertical flow line of the upwelling flux, and thus, along
the buried fault system.
Inside the karst substratum, the limit between karstified and unkarstified terrains is unknown;
this limit has been marked by question marks, and it could be fixed at −130 m below sea level, which
Water 2018, 10, 501 17 of 21

flux is widespread and circulates in a porous medium; most of the upwelling flux outcrops along the
Sabato River,exposition
continental located onof
thecarbonate
shorter and vertical
reliefs, flow line
deeper of the
karst upwelling
levels flux, and
are lowered by thus, alongand
Pliocene the
buried fault faults.
Quaternary system.
Figure 16. Hydrogeological

Figure 16. Hydrogeological section
section crossing
crossing discharge
discharge and
and recharge
recharge area
area of
of the
the Acquaro
Acquaro and
and Pelosi
Pelosi
springs (section trace in Figure 3). Δh 1 is the variation in the water table level in the phreatic zone
springs (section trace in Figure 3). ∆h1 is the variation in the water table level in the phreatic zone
(well S1), which
(well S1), which becomes
becomesartesian
artesianininthe
theSabato
Sabatovalley.
valley.The
Thelimit
limit
ofof
thethe karstification/saturated
karstification/saturated zone
zone is
is unknown. R is the width of the influenced zone during high flow period (R = 169 m;
unknown. R is the width of the influenced zone during high flow period (R = 169 m; Figure 14). Figure 14).
In the Piana del Dragone, which is the main recharge area of the aquifer (Figure 3), the hydraulic
Inside the karst substratum, the limit between karstified and unkarstified terrains is unknown;
head varies between 506 m a.s.l and 554 m a.s.l. in the period between 2003–2008 (Figure 4); the
this limit has been marked by question marks, and it could be fixed at −130 m below sea level,
observed range, Δh1 = 48 m, has been indicated in Figure 16. The hydraulic head in this area controls
which was the base level during the last glaciation. However, as karstification processes began
the hydraulic head observed in the piezometers of the Serino plain, which have a range up to 8 m
during the continental exposition of carbonate reliefs, deeper karst levels are lowered by Pliocene and
(Figure 9b).
Quaternary faults.
Under drought conditions, a descendant hydraulic gradient could occur in alluvial deposits,
In the Piana del Dragone, which is the main recharge area of the aquifer (Figure 3), the hydraulic
which means (i) a poor hydraulic head into karst substratum connected to a poor recharge and (ii) a
head varies between 506 m a.s.l and 554 m a.s.l. in the period between 2003–2008 (Figure 4); the
temporary reverse flow into alluvial deposits which recharges the karst aquifer below. However, the
observed range, ∆h1 = 48 m, has been indicated in Figure 16. The hydraulic head in this area controls
above drought conditions occur rarely, and ascendant hydraulic gradient occurs during the entire
the hydraulic head observed in the piezometers of the Serino plain, which have a range up to 8 m
hydrological year and supports the fact that the spring never dries up (except Acquaro spring during
(Figure 9b).
in the year 2002). A minimum and maximum hydraulic gradient (0.026 and 0.15, respectively), as
Under drought conditions, a descendant hydraulic gradient could occur in alluvial deposits,
previously estimated, will now be considered.
which means (i) a poor hydraulic head into karst substratum connected to a poor recharge and
Figure 17 shows the hydraulic gradient estimated in depth, extrapolating the hydraulic head
(ii) a temporary reverse flow into alluvial deposits which recharges the karst aquifer below. However,
measured in P5 and P6 wells (Figure 12a); the top of the karst substratum can be estimated by
the above drought conditions occur rarely, and ascendant hydraulic gradient occurs during the entire
Equation (6), and thus the hydraulic head ranges into karst substratum would be Δh1 ≈ 40 m. This
hydrological year and supports the fact that the spring never dries up (except Acquaro spring during
value is slightly lower than 48 m found in well S1 located in the recharge zone (Figures 3, 4 and 16),
in the year 2002). A minimum and maximum hydraulic gradient (0.026 and 0.15, respectively), as
according to water path into karst aquifer, from the recharge zone to karst substratum below the
previously estimated, will now be considered.
alluvial deposits. However, as data sets cover different periods, a more specific link between water
Figure 17 shows the hydraulic gradient estimated in depth, extrapolating the hydraulic head
level recorded in the Piana del Dragone (2002–2006) and piezometers of Sabato valley (1963–1969)
measured in P5 and P6 wells (Figure 12a); the top of the karst substratum can be estimated by
could not be allowed.
Equation (6), and thus the hydraulic head ranges into karst substratum would be ∆h1 ≈ 40 m.
A relevant consequence of high values of the ascendant hydraulic flow is the spring water
This value is slightly lower than 48 m found in well S1 located in the recharge zone (Figures 3,
quality. As described, this water is cold (T = 11 °C) and it is poorly mineralized (E.C. = 370 μS/cm,
4 and 16), according to water path into karst aquifer, from the recharge zone to karst substratum below
20 °C), but it comes from a deep aquifer, located at 353 m depth below the Sabato River. We expected
the alluvial deposits. However, as data sets cover different periods, a more specific link between water
that water coming from so deep an aquifer should be more mineralized, as found in other
hydrogeological contexts [34,38,39]. A possible explanation can be found in the high velocity of the
ascendant flow into alluvial deposits directly connected to the high hydraulic gradient.
Water 2018, 10, 501 18 of 21
level recorded in the Piana del Dragone (2002–2006) and piezometers of Sabato valley (1963–1969)
could not be allowed.
A relevant consequence of high values of the ascendant hydraulic flow is the spring water quality.
As described, this water is cold (T = 11 ◦ C) and it is poorly mineralized (E.C. = 370 µS/cm, 20 ◦ C),
but it comes from a deep aquifer, located at 353 m depth below the Sabato River. We expected that
water coming from so deep an aquifer should be more mineralized, as found in other hydrogeological
contexts [34,38,39]. A possible explanation can be found in the high velocity of the ascendant flow into
alluvial
Water deposits
2018, 10, x FORdirectly connected to the high hydraulic gradient.
PEER REVIEW 18 of 21
Figure 17.
Figure Hydraulic head
17. Hydraulic head measured
measured into
into wells
wells P5
P5 and
and P6
P6 during
during high
high and
and low low flow
flow periods;
periods; the
the
extrapolation in depth allow to estimation of the hydraulic head range, ∆h in the karst substratum.
extrapolation in depth allow to estimation of the hydraulic head range, Δh1,1, in the karst substratum.
Using the theresults

resultsofof a pumping
a pumping testtest carried
carried out inout
theinP11theand P11
P12and P12(Figure
wells wells 5),
(Figure
which5), which
provided
provided a hydraulic conductivity, K−=3 3 × 10 −3 m/sec, a velocity of v = 1.62 m/h (i = 0.150, high flow)
a hydraulic conductivity, K = 3 × 10 m/sec, a velocity of v = 1.62 m/h (i = 0.150, high flow) and
and v = 0.28
v = 0.28 m/hm/h(i =(i0.026,
= 0.026,
lowlow flow)
flow) cancanbebe estimated.Considering
estimated. Consideringan aneffective
effectiveporosity
porosity ofof 0.1
0.1 of the
alluvial deposits
deposits (sand
(sandandandgravel),
gravel),the theactual
actualvelocity
velocity is v ≈ 16.2 m/h (high flow) and
is veff ≈ 16.2 m/h (high flow) and veff ≈ 2.8
eff v eff ≈ 2.8 m/h
m/h
flow). This means
(low flow). means that,
that, during
during the thehigh
highflow
flowand
andlow lowflow,
flow,water waterwould
wouldneed need≈≈1 dayor
1 day or≈ ≈55 days
to reach the springs from the karst substratum, respectively, and within this short period there is a
better understanding as to why the Serino spring water is is poorly
poorly mineralized
mineralized and and cold.
cold.
The Rn222
222
Rnactivity
activityfor
forAcquaro
Acquaro and and Pelosi
Pelosi springs
springs clearly
clearly indicate
indicate aa variation
variation in thethe concentration
concentration
in the monitored period (February
(February 2015–July
2015–July 2016);
2016); inin particular,
particular,as asshown
shownin inFigure
Figure10,10,222
222Rn activity
clearly increases
increasesasas spring
spring discharge
discharge increases,
increases, and decreases
and decreases as spring asdischarge
spring discharge
decreases; decreases;
maximum
maximum
and minimum andradon
minimum radon
activities occuractivities
during theoccurhigh andduringlow the flow high and low
conditions, flow conditions,
respectively.
respectively.
The high flow condition is sustained by high ascendant hydraulic gradient, under which higher
waterThe high flow
velocity condition
and higher radonis sustained by high
activity occur. ascendant
Thus, the highhydraulic
velocity ofgradient, under which
the upwelling highera
flow allows
water velocity
more rapid and higher
ascendant waterradon
which activity
favors occur. Thus,
a higher 222the
Rn high
activity. velocity
Duringof the
lowupwelling flowwhen
flow periods, allowsa
alower
moreascendant
rapid ascendant water 222
which favors a higher 222Rn activity. During low flow periods, when a
flow occurs, the Rn activity decreases as consequence of the longer period it takes
lower ascendant
for water to reachflow occurs, the 222Rn activity decreases as consequence of the longer period it takes
the springs.
for water to reach
Moreover, thethe springs. between the time needed to reach the spring from karst substratum
difference
during the high and low flow (between
Moreover, the difference ≈4 days)the time needed
coincides with the to half-time
reach theof the 222from
spring karst
Rn (3.8 substratum
days), and the
during the high and low flow (≈ 4 days) coincides with the half-time of the Rn (3.8 days), and the 222
radon activity measured during the high flow period almost double of that measured during the low
flow period. These results suggest that radon activity of groundwater stored in the saturated zone of
karst aquifer substratum can be considered almost constant during the year, as a consequence of the
long-term equilibrium between its accumulation (from the decay of uranium and radio contained in
Water 2018, 10, 501 19 of 21
radon activity measured during the high flow period almost double of that measured during the low
flow period. These results suggest that radon activity of groundwater stored in the saturated zone of
karst aquifer substratum can be considered almost constant during the year, as a consequence of the
long-term equilibrium between its accumulation (from the decay of uranium and radio contained in
the rocks) and its decay (with a half time of 3.8 days). In this case, the faster upwelling of spring water
occurring during high flow periods favor higher radon activity, whereas, the slower upwelling of
spring water occurring during low flow periods favor the radon decay and, thus, its reduced activity
in the spring water.
7. Conclusions
The ascendant water flux which feeds the springs belong to a wide hydraulic phenomenon
involving wide areas and deep portions of aquifers. It is always connected to specific hydrogeological
conditions, coming from tectonic and geomorphologic evolution. In a karst environment, the ascendant
flux can be particularly developed because of the carbonate dissolution along the shorter flow lines,
characterized by the higher hydraulic gradient and forming karst conduits, as well as allowing the
water flux to siphon under no-karst terrains. This phenomenon is still poorly quantified, as deep
hydraulic measurements are generally missing.
The Serino spring area is a useful example of detecting and estimating the ascendant water flux,
thanks to deep piezometers into alluvial deposits which cover a karst substratum and high values
of the ascendant hydraulic gradient. The estimation of the ascendant hydraulic gradient during the
high and low periods has provided an overall view of the phenomenon, allowing us to estimate the
depth of the karst substratum in the Sabato plain and have a deeper knowledge of groundwater
flow. During low flow conditions, a wider area appears to be influenced by the drainage systems
with respect to high flow conditions. These specific characteristics have been quantified by analytical
solutions to find the area being influenced by a drainage system under different geometrical and
hydraulic features.
As a consequence of the ascendant flux, the Serino alluvial plain can be considered outside of the
spring catchment, as it does not contribute in recharging the below karst aquifer. Besides, especially
during the high flow period, the spring water could be exposed to contaminations by perched water in
the alluvial plain, and ascendant flux appears to be a useful and natural condition for preserving deep
groundwater. However, drought conditions reduce the power of the ascendant hydraulic gradient
drastically, and spring discharge can decrease up to nil; thus, drought conditions could favor the
deterioration of water quality.
The ascendant water flux allows deep groundwater to reach the surface and start up a spring,
and thus it is a common condition for thermal springs and other mineralized waters. For the Serino
springs, which are cold and very low mineralized water, it is assumed that the estimated ascendant
velocity is high enough to avoid increases in temperature and mineralization. In particular, the time
needed to reach the spring outlet has been estimated at one (high flow) or a few days (low flow)
from the karst substratum, and this sheds light on the almost constant value of temperature and
electrical conductivity. Under such hydraulic conditions, characterized by a general fast velocity of
the groundwater, the 222 Rn activity seems to be a useful indicator of groundwater dynamic; in detail,
during high flow periods, the water is younger with respect to low flow periods, due to a higher
ascendant flow velocity.
The recognized upwelling flux in the Serino plain could lead to a better understanding also of
many karst springs of Apennine, when hydrogeological conditions can be schematized as shown
in Figure 15. Under such hydrogeological conditions, the ascendant water flux has not been fully
recognized. This could lead to a better vulnerability definition and, thus, a more efficient local planning
in order to correctly manage and preserve the quantity and quality of groundwater resources.
Water 2018, 10, 501 20 of 21
Acknowledgments: Hydrological and chemical data have been provided by aqueduct companies of Naples
(Acqua Bene Comune) and Avellino (Alto Calore Servizi). Thanks also to Domenico Cicchella, University of
Sannio, for his support during the Radon activity measurements.
Author Contributions: Francesco Fiorillo has written the text and coordinated the research with Libera Esposito;
Giovanni Testa has carried out is situ measurements and analyzed data series; Sabatino Ciarcia supported
the geological aspects of the area; Mauro Pagnozzi developed some figures/maps and review some aspects
of literature.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Todd, D.K.; Mays, L.W. Groundwater Hydrology; Wiley: Chichester, UK, 2005; 636p.
2. Stevanovic, Z.; Kresic, N. Groundwater Hydrology of Springs; Butterworth Heinemann-Elsevier: Oxford, UK,
2009; 592p.
3. Hubbert, M.K. The theory of groundwater motion. J. Geol. 1940, 48, 785–944. [CrossRef]
4. Toth, J. A theoretical analyses groundwater flow in small drainage basins. J. Geophys. Res. 1963, 68, 4795–4812.
[CrossRef]
5. Domenico, P.A.; Schwartz, F.W. Physical and Chemical Hydrogeology; John Wiley & Sons: Singapore, 1990;
824p.
6. Martínez-Santos, P.; Díaz-Alcaide, S.; Castaño-Castaño, S.; Hernández-Espriú, A. Modelling discharge
through artesian springs based on a high resolution piezometric network. Hydrol. Process. 2014, 28,
2251–2261. [CrossRef]
7. Ford, D.; Williams, P. Karst Hydrogeology and Geomorphology; Wiley: Chichester, UK, 2007; 562p.
8. Stevanovic, Z. Characterization of Karst Aquifer. In Karst Aquifers—Characterization and Engineering;
Stevanovic, Z., Ed.; Springer: Berlin/Heidelberg, Germany, 2015; 692p.
9. Blyth, F.G.H.; De Freitas, M.H. A Geology for Engineers, 7th ed.; Arnould Ed.: London, UK, 1984; 325p.
10. Hooghoudt, S.B. Bijdragen tot de Kennis van eenige Natuurkundige Grootheden van de Grond; Verslagen van
Landbouwkundige Onderzoekingen Algemeene Landsdrukkerij: The Hague, The Netherlands, 1940.
11. Van Schilfgaarde, J. Design of tile drainage for falling water tables. J. Irrig. Drain. Div. 1963, 89, 1–12.
12. Mishra, G.C.; Singh, V. A new drain spacing formula. Hydrol. Sci. J. Sci. Hydrol. 2007, 52, 338–351. [CrossRef]
13. Lei, S. An analytical solution for steady flow into a tunnel. Ground Water 1999, 37, 23–26. [CrossRef]
14. El Tani, M. Circular tunnel in a semi-infinite aquifer. Tunn. Undergr. Space Technol. 2003, 18, 49–55. [CrossRef]
15. El Tani, M. Helmholtz evolution of a semi-infinite aquifer drained by a circular tunnel. Tunn. Undergr.
Space Technol. 2010, 25, 54–62. [CrossRef]
16. Park, K.H.; Owatsiriwong, A.; Lee, J.G. Analytical solution for steady-state groundwater inflow into a
drained circular tunnel in a semi-infinite aquifer: A revisit. Tunn. Undergr. Space Technol. 2008, 23, 206–209.
[CrossRef]
17. Butscher, C. Steady-state groundwater inflow into a circular tunnel. Tunn. Undergr. Space Technol. 2012, 32,
158–167. [CrossRef]
18. Vitale, S.; Ciarcia, S. Tectono-stratigraphic and kinematic evolution of the Southern Apennines/
Calabria—Peloritani Terranesystem (Italy). Tectonophysics 2013, 583, 164–182. [CrossRef]
19. Cinque, A.; Patacca, E.; Scandone, P.; Tozzi, M. Quaternary kinematic evolution of the Southern Apennines.
Relationship between surface geological features and deep lithospheric structures. Ann. Geophys. 1993, 36,
249–260.
20. Fiorillo, F.; Pagnozzi, M.; Ventafridda, G. A model to simulate recharge processes of karst massifs.
Hydrol. Process. 2015, 29, 2301–2314. [CrossRef]
21. Calcaterra, D.; Ducci, D.; Santo, A. Aspetti geo-meccanici ed idrogeologici nel settore sud-orientale del,
M.te Terminio (Appennino Meridionale). Geol. Romana 1992, 30, 125–134.
22. Celico, P. Schema Idrogeologico dell’Appennino Carbonatico Centro-Meridionale; Memorie e Note Istituto di
Geologia Applicata: Napoli, Italy, 1978; Volume 14, pp. 1–43.
23. Ciarcia, S.; Vitale, S. Sedimentology, stratigraphy and tectonics of evolving wedge-top depozone: Ariano
Basin, southern Apennines, Italy. Sediment. Geol. 2013, 290, 27–46. [CrossRef]
24. Civita, M. Idrogeologia del Massiccio del Terminio-Tuoro (Campania); Memorie e Note Istituto di Geologia
Applicata: Napoli, Italy, 1969; Volume 11, pp. 5–102.
Water 2018, 10, 501 21 of 21
25. Fiorillo, F.; Esposito, L.; Guadagno, F.M. Analyses and forecast of the water resource in an ultra-centenarian
spring discharge series from Serino (Southern Italy). J. Hydrol. 2007, 36, 125–138. [CrossRef]
26. Celico, P.; Magnano, F.; Monaco, L. Prove di Colorazione del Massiccio Carbonatico del Monte Terminio-Monte
Tuoro; Notiziario Club Alpino Italiano: Napoli, Italy, 1982; Volume 46, pp. 73–79.
27. Fiorillo, F.; Guadagno, F.M. Long karst spring discharge time series and droughts occurrence in Southern
Italy. Environ. Earth Sci. 2012, 65, 2273–2283. [CrossRef]
28. Fiorillo, F. Spring hydrographs as indicators of droughts in a karst environment. J. Hydrol. 2009, 373, 290–301.
[CrossRef]
29. Neuman, S.P. Effect of partial penetration on flow in unconfined aquifers considering delayed gravity
response. Water Resour. Res. 1974, 10, 303–312. [CrossRef]
30. Tartakovsky, G.D.; Neuman, S.P. Three-dimensional saturated-unsaturated flow with axial symmetry to
a partially penetrating well in a compressible unconfined aquifer. Water Resour. Res. 2007, 43, W01410.
[CrossRef]
31. Sophocleous, M. Interactions between groundwater and surface water: The state of the science. Hydrogeol. J.
2002, 10, 52–67. [CrossRef]
32. Barthel, R.; Banzhaf, S. Groundwater and surface water interaction at the regional-scale—A review with
focus on regional integrated models. Water Resour. Manag. 2016, 30, 1–32. [CrossRef]
33. Stevanovic, Z.; Dragisic, V. An example of identifying karst groundwater flow. Environ. Geol. 1997, 35,
241–244. [CrossRef]
34. Szőnyi, J.M.; Tóth, J. A hydrogeological type section for the Duna-Tisza Interfluve. Hung. Hydrogeol. J. 2009,
17, 961–980. [CrossRef]
35. Fiorillo, F.; Petitta, M.; Preziosi, E.; Rusi, S.; Esposito, L.; Tallini, M. Long-term trend and fluctuations of
karst spring discharge in a Mediterranean area (central-southern Italy). Environ. Earth Sci. 2015, 74, 153–172.
[CrossRef]
36. Kiraly, L. Karstification and Groundwater Flow. In Evolution of Karst: From Prekarst to Cessation;
Gabrovsek, F., Ed.; Zalozba ZRC: Ljubljana, Slovenia, 2002; pp. 155–190.
37. Palmer, A.N.; Audra, P. Patterns of caves. In Encyclopedia of Cave and Karst Science; Gunn, J., Ed.; Fitzroy
Dearborn: London, UK, 2004; pp. 573–574.
38. Czauner, B.; Szőnyi, J.M.; Tóth, J.; Pogácsás, G. Hydraulic potential anomaly indicating thermal water
reservoir and gas pool near Berekfürdõ, Trans-Tisza Region, Hungary. Cent. Eur. Geol. 2008, 51, 253–265.
[CrossRef]
39. Goldscheider, N.; Szőnyi, J.M.; Erőss, A.; Schill, E. Review: Thermal water resources in carbonate rock
aquifers. Hydrogeol. J. 2010, 18, 1303–1318. [CrossRef]
© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Water: The Upwelling Water Flux Feeding Springs: Hydrogeological and Hydraulic Features

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Water: The Upwelling Water Flux Feeding Springs: Hydrogeological and Hydraulic Features

Uploaded by

Copyright:

Available Formats

water

Water 2018, 10, 501; doi:10.3390/w10040501 www.mdpi.com/journal/water

2. Materials and Methods

2.1. Ascendant Hydraulic Gradient and Wideness of Influenced Zone

into account the ascendant groundwater flow condition.

2.2. Physical and Chemical Characteristics Measurements

3. Geological and Hydrogeological Setting

Water 2018, 10, x FOR PEER REVIEW 5 of 21

Figure 2. Geological sketch of Southern Appennines (from [23], modified).

water table, m a.s.l.

water table, m a.s.l.

380 ground level

385 P6 (35 m depth)

R² = 0,79 P3 P4 6 y = 0,1173x - 2,4417

376 y = 0,1439x + 366,43 P7

P10 R² = 0,79 P4 y = 0,1173x - 2,4417

376 6 y = 0,1173x - 2,4417

Figure 10. (a)

To reduce the ambiguity observed

Figure 13. Potential hydraulic gradient, iTT,, induced

and from Equation (4):

the value Z ZP deducted provides the depth depth of

Water 2018, 10, x FOR PEER REVIEW 17 of 21

Figure 16. Hydrogeological

Using the theresults

You might also like