For the first time a comprehensive investigation has been carried out to quantify the possible effects of dredging a navigable canal on the hydrogeological system underlying a coastal lagoon. The study is focused on the Venice Lagoon, Italy, where the port authority is planning to open a new 10 m deep and 3 km long canal to connect the city passenger terminal to the central lagoon inlet, thus avoiding the passage of large cruise ships through the historic center of Venice. A modeling study has been developed to evaluate the short (minutes), medium (months), and long (decades) term processes of water and pollutant exchange between the shallow aquifer system and the lagoon, possibly enhanced by the canal excavation, and ship wakes. An in-depth characterization of the lagoon subsurface along the channel has supported the numerical modeling. Piezometer and sea level records, geophysical acquisitions, laboratory analyses of groundwater and sediment samples (chemical analyses and ecotoxicity testing), and the outcome of 3-D hydrodynamic and computational fluid dynamic (CFD) models have been used to set up and calibrate the subsurface multi-model approach. The numerical outcomes allow us to quantify the groundwater volume and estimate the mass of anthropogenic contaminants (As, Cd, Cu, Cr, Hg, Pb, Se) likely leaked from the nearby industrial area over the past decades, and released into the lagoon from the canal bed by the action of depression waves generated by ships. Moreover, the model outcomes help to understand the effect of the hydrogeological layering on the propagation of the tidal fluctuation and salt concentration into the shallow brackish aquifers underlying the lagoon bottom.
Coastal lagoons are transient ecosystems highly sensitive to changes in sedimentation, sea level rise, and land subsidence. In many cases, their evolution over the last centuries has been strongly impacted by human activities. The use of these peculiar ecosystems, for fish and shell farming, tourism, and transportation of people and goods, has usually contrasted with the preservation and protection of habitat and biodiversity (Kennish and Paerl, 2010). One typical intervention in coastal systems is dredging of canals and inlets, which may be performed to increase the water volume exchanged with the sea (Gong et al., 2008) or for navigation purposes (Fortunato and Oliveiram, 2007; Healy et al., 1996). Elsewhere, dredging has been used as a source of fill material for adjacent upland development and land reclamation (López et al., 2013).
The opening of waterways in shallow coastal waterbodies and lagoons has facilitated navigation for centuries, providing sheltered routes and permitting safe access to inland ports and harbours. However, the progressively increasing tonnage of vessels and sediment dynamics requires port authorities worldwide to implement dredging programmes to guarantee navigability and to open new shipping channels to allow larger traffic.
In many cases, this practice has led to environmental deterioration, by changing the flushing efficiency of the canal system, aggravating salinity stratification, re-suspending fine sediments, pollutants, and nutrients, which are responsible for eutrophication, hypoxic events, and increasing contamination and release of pollutants (e.g., Newell et al., 1998). For example, the combined impacts of increased turbidity and physical removal or burial during dredging caused a loss of approximately 81 % of the seagrass in Tampa Bay, Florida (Erftemeijer and Lewis, 2006). Moreover, canal dredging has been responsible for significant hydro-morphological impacts in coastal lagoons. A significant example is the case of Aveiro Lagoon in Portugal, where 2 centuries of channelization, jetty breakwater construction, and dredging have led to a progressive shift from the original fluvially dominated system into the present tidally dominated one. The associated stresses imposed by increased tidal currents have led to important changes in the sedimentary regime (Duck and da Silva, 2012).
In the Venice Lagoon, Italy, engineers and administrations have planned
dredging works for centuries, creating a series of canals for navigation and
reclaimed land for urban expansion and industrial settlement (Balletti, 2006;
D'Alpaos, 2010). The last major navigable canal, the Malamocco-Marghera
Industrial Canal (MMIC), was excavated in 1970 to connect the Porto Marghera
Industrial Zone (PMIZ) on the mainland with the Adriatic Sea through the
Malamocco inlet (Fig. 1a). A large number of studies developed over the last
decade has demonstrated that the MMIC and the navigation of large vessels
through the lagoon shallows have likely been the main causes of the
morphological deterioration observed in the central lagoon as deepening of
the tidal flats, marshland erosion, and sediment loss (e.g., Amos et al.,
2010; Carniello et al., 2009; Ferrarin et al., 2013; Marani et al., 2011;
Molinaroli et al., 2009; Tambroni and Seminara, 2006). The Venice Port
Authority recently planned the excavation of a new approximately 3 km long
and 10 m deep navigation canal (called the Marghera-Venice Canal, MVC in the
following) to reroute vessels along the MMIC and reach the passenger terminal
located in the southwestern part of the historic center (Fig. 1b). The
intervention should avoid the transit of large cruise liners though the
historic center of Venice. At present, more than 500 cruise ships enter the
lagoon each year (
Despite the large research effort dedicated to the understanding of the freshwater–groundwater exchange in coastal aquifers (e.g., Li et al., 1999; Michael et al., 2005; Nakada et al., 2011; Qu et al., 2014), studies developed in the past have never addressed the evaluation of possible effects of excavating navigable canals through tidal flats on the underlying hydrogeological system. However, in-depth investigations using direct measurements (isotopes, benthic chambers), geophysical surveys, and modeling simulations revealed that submarine groundwater discharge (SGD) may provide considerable freshwater inputs to coastal waterbodies (e.g., Rapaglia et al., 2010; Wang et al., 2015) and may be the primary pathway for nutrients and other contaminants to enter coastal lagoons (e.g., Rapaglia, 2005; Rocha et al., 2016; Santos et al., 2008; Tait et al., 2013).
The primary objectives of this study are to investigate how the construction
of a new large navigable canal through tidal flats affects (i) the
groundwater flow and quality of the shallow aquifers underlying the lagoon
bottom and (ii) the exchange of water and chemicals from the subsurface to
the surface waterbodies. Our research focuses on the Venice Lagoon as a
representative case study. Based on the available knowledge on the surface
and subsurface lagoon environment, the following issues had to be considered
in the context of the study.
The quality of the surficial water, mainly its salinity, with respect to the
groundwater. Can an eventual cut of impervious layers enhance saltwater
leakage beneath the lagoon bottom? The presence of chemicals in the groundwater below the lagoon bottom and the
sediment toxicity due to leakage from the industrial and urban centers
located in the lagoon surroundings. Are contaminants present also along the
MVC designed path? Which is their mobility? And can the MVC excavation
determine their release into the lagoon, also favored by SGD from exposed
sub-surficial heterogeneities? The evolution of water level in the lagoon canals and flats due to the
transit of large vessels. How do the solitary waves associated with the
passage of large vessels in the navigation channel influence the flow and
contaminant transfer between the subsurface and surficial systems?
In this study, for the first time, we explore in detail these issues,
improving the understanding of the interaction between the subsurface and
surface waters in coastal systems, and providing quantitative evaluations for
the specific case study. This is carried out through an accurate
investigation of the lagoon environment along the MVC trace and the use of
uncoupled and coupled density-dependent groundwater flow and transport
simulators.
The paper is organized as follows. Section 2 presents the dataset available to characterize the subsurface system of the Venice Lagoon along the MVC and the factors forcing its dynamics. The numerical approach used to perform the hydrogeological modeling study is revised in Sect. 3 together with the description of the model setup. The results obtained by the computations are presented in Sect. 4, pointing out the effect of the canal dredging by comparing the model solutions in the present (i.e., without the canal) and planned (i.e., with the canal) scenarios. A discussion (Sect. 5) and a concluding section (Sect. 6) close the paper, evaluating the main outcomes in a general context and summarizing the principal results of the study.
About 40 km of very-high-resolution seismic (VHRS) lines (Fig. 1b) were collected by a boomer system equipped with an electro-dynamic plate and a single-channel streamer. The latter consisted of eight equidistant piezoelectric elements housed in an oil-filled tube and connected in series with a 2.8 m active array section (Tosi et al., 2009). The frequency bandwidth produced by the plate ranged from 0.4 to 9 kHz, thus allowing a decimeter resolution. Suitable floaters kept the streamer as shallow as possible to avoid destructive interference between reflected signals and multiple events from the air–water interface. Because the investigated area is characterized by shallow water (< 1.5 m) and the conventional acquisition geometry (streamer towed behind a source) generates poor results using a single-channel streamer, a transverse geometry was applied to collect more coherent events (Baradello and Carcione, 2008). The seismic data were processed by a conventional sequence, including initially spherical divergence removal, secondly a time-variant gain, and finally a time-variant band-pass filter. The marine and boat waves degraded the reflection signal in a number of profiles. This effect was mitigated by computing a mean trace in a given interval, cross-correlating it with the single traces, and applying the corresponding time shift as a static correction.
Hydrogeological setting of the subsurface of the Venice Lagoon along
The interpretation of the seismic units with the support of stratigraphic
data obtained through 10 continuous 10 m long cores (Fig. 1b) specifically
drilled for the study allowed us to sketch the hydro-stratigraphic setting of
the lagoon subsoil along the MVC designed trace. In addition, a variety of
geophysical, lithological, sedimentological, and geotechnical information
available from previous investigations (Fabbri et al., 2013; Madricardo and
Donnici, 2014; Teatini et al., 2011; Tosi et al., 2007, 2011; Zecchin et al.,
2011, 2014) was reprocessed to characterize the architecture of the deeper
deposits (down to a depth of approximately 50 m b.m.s.l.) and contextualize the investigated area within a regional
hydrogeological framework. Figure 2 shows two interpreted VHRS lines, namely
Section-1 and Section-2, orthogonal to the MVC trace. The seismic survey
revealed a high heterogeneity of the lagoon subsoil due to a number of buried
paleo-channels, whose presence would not have been highlighted only by the
core samples. This study is focused on these two sections, each of them
crossing a 10 m depth borehole. The borehole coordinates are the following:
286 467
Combining the new hydro-stratigraphic information and that available from
previous studies, it has been possible to characterize the
hydro-stratigraphic system and identify three main permeable units down to
about 50 m depth beneath the lagoon bottom (Fig. 2). In the easternmost
Section-1, the shallowest sandy unit (Aqf-1) is 7 to 10 m thick and lies
below a silty-muddy layer a few meters thick (Aqt-1). Aqf-1 is almost
continuous in the central and southern parts of the investigated area and
decreases northward where silty-clay deposits prevail. This aquifer
represents a buried paleo-channel, whose direction is from the industrial
zone to the historical center of Venice. The top of Aqf-1 represents the
Holocene–Pleistocene limit. This is marked by an erosional unconformity
generally made by a metric cemented clayey layer (Tosi et al., 2007), known
in the Venice area as “caranto”. A second sandy unit (Aqf-2) is generally
confined below a 2–3 m thick silty-clayey layer (Aqt-2). In the
central–western part of the study area where Section-2 is located, Aqt-2 is
lacking because of paleo-channel incisions, and Aqf-1 and Aqf-2 are
undifferentiated. At the bottom of Aqf-2, a second quite continuous aquitard
(Aqt-3) confines a third aquifer (Aqf-3), which has a regional extent. The
Aqf-3 depth ranges between
Despite the large effort carried out in the past to define the
hydrogeological setting of the shallow lagoon subsurface, scarce information
is available in the study area because the characterization was mainly
concentrated along the littoral strips. A 100 km long airborne
electromagnetics (AEM) survey carried out in 2009 between Venice and the
industrial area pointed out the important hydrogeologic function played by
the caranto (Teatini et al., 2011). The AEM investigation provided
resistivity information from the lagoon bottom down to about 120–140 m
depth and clearly showed that the caranto reduces or precludes the downward
leakage of seawaters. Groundwater with a salt concentration comparable with
the marine waters (resistivity
Groundwater and hydrogeologic properties in the upper 10 m in depth have
been investigated by a Casagrande piezometer installed at the bottom of the
boreholes along Section-1 and Section-2. Each monitoring station was
instrumented by two CTD-Divers, one placed within the borehole and connected
to the Casagrande cell and the other fixed outside the borehole casing at the
lagoon bottom. The configuration allowed the simultaneous monitoring of
electrical conductivity (EC), temperature (
Figure 3 shows an example of the recorded Aqf-1 pressure head and lagoon
level during a few days in March 2016 at the Section-2 station. As expected,
the fluctuation of the groundwater level is phased on the semi-diurnal tidal
regime, with a gentle (10–15 %) reduction of the wave height and a delay
of 10–20 min on the maximum/minimum occurrence. Similar values were
obtained at the Section-1 borehole. Concerning EC, the records are
characterized by a negligible variability in time. With reference to a
temperature
Among the various pollutants, trace elements are of particular concern since they are taken up by biota and may have toxic effects. Some trace elements are on the priority list and regulated by European directives, e.g., the 2000/60 EC (Water Framework Directive), and their national transpositions, e.g., Ministerial Decree 260/2010 in Italy. Furthermore, laws and decrees regulate the presence of trace elements in specific relation to dredging. Before any dredging, the chemical characterization together with an eco-toxicological evaluation must be carried out. Although in the past much interest has been focussed on the total concentration of trace elements, it has more recently been accepted that assessing the mobility, the bioavailability, the bioaccessibility, and the toxicity of metals is fundamental (Schintu et al., 2016; Zhang et al., 2017).
Measured tidal level and piezometric head at the Section-2 borehole over the period between 7 and 10 March 2016.
Within this study, both the total concentration of trace elements and the geo-speciation, defined according to Ure et al. (1993), were carried out on samples collected from the two reference boreholes along Section-1 and Section-2 (Fig. 1b). The geo-speciation was performed via the sequential extraction procedure (SEP) proposed by Tessier et al. (1979) and harmonized by Corami et al. (2009). The SEP allows an operational classification of metals into four geochemical fractions with different mobility, bioavailability, and bioaccessibility: labile, bound to oxyhydroxides of iron and manganese, bound to organic matters and sulfides, residual. Mobility decreases from the first fraction, i.e., labile, to the residual, which is inert. The labile fraction is bioaccessible and bioavailable, with the second and third fractions that may become accessible to biota according to their mobility, e.g., in the case of anoxic events, or during dredging operation. The labile fraction is composed of a readily exchangeable portion and a portion bound to carbonates, which are readily available for the uptake by the benthic fauna, i.e., the biota living in and on the sediments and, due to the sediment resuspension, for the uptake by the biota in the water column (Qiao et al., 2013; Lee et al., 2017).
A detailed description of the preparative and analytical methods employed to assess the total concentration and to study the geo-speciation is besides the aim of this paper and the reader can refer to DAIS (2016). Briefly, a hydraulic corer was used to collect the samples. Dried aliquots of sediments, previously homogenized for each meter interval, were assessed for the total concentration of the 12 trace elements (As, Be, Cd, Cr, Cu, Hg, Ni, Pb, Sb, Se, V, Zn), in compliance with EU Directive 2013/39 and the Memorandum of Understanding dated 8 April 1993 between the Italian Ministry of the Environment, Veneto Region, Province of Venice, Venice and Chioggia Municipality, on the “Criteria for the safety of the environment during excavation, movement, and reemployment of muds dredged from the Venice channels”, also known as “Protocollo 93”. Unaltered aliquots were analyzed by the harmonized SEP to quantify the labile and bioaccessible fraction of some trace elements, namely As, Cd, Cu, Cr, Hg, Pb, and Se. Figure 2c provides an example of the results obtained by the chemical characterization in terms of labile Cr concentration versus depth measured in the B1 (Section-1) and B2 (Section-2) boreholes. As a general feature, the labile and bioaccessible fraction increases with depth. This trend has been confirmed at Section-1 for every trace element considered by this study and, to a lesser extent (Cd and Pb), in Section-2.
Ecotoxicity is of great interest in sediment assessment and management,
providing an integrated response related to the bioavailable and
bioaccessible fraction of contaminants within the checked matrix (whole
sediment, pore water, and elutriates). The sediment samples collected in
Section-1 and Section-2 were homogenized and sieved at 2 mm (in a N
The results showed a toxicity range from absent/low (acute tests) to very
high (i.e., all embryotoxicity tests) considering as ranking tools the
toxicity scales set up on a species-by-species basis for Venice Lagoon
sediments (Losso et al., 2010). According to the worst-case scenario
approach, sediment presented very high levels of toxicity independently of
the core depth (from 0 to
Apart from the natural tidal regime, a certain effect on the hydrogeological system in the surroundings of deep channels is expected to be driven by long inverse solitary waves associated with the passage of large vessels in the navigation channel and known as depression wakes, or Bernoulli wakes (Rapaglia et al., 2015). Ship wakes were characterized by means of water level measurements made with pressure sensors and turbidity meters deployed along a profile on the channel side and the surrounding mudflat together with a modeling chain capable of reproducing the hydrodynamic patterns in the channel around the hull of the moving ship, and the propagation of the depression wake on the tidal flat.
Tidal level and ship-induced depression wakes as well as short-period boat
wakes were measured with a pressure sensor with a logger (Solo D/Wave, RBR,
Canada) immersed at a depth of approximately 4 m on the eastern side of the
MMIC. Pressure was recorded by the instrument at a sampling frequency of
16 Hz and converted into depth data. The experimental setup also included an
electromagnetic current meter deployed at the bottom of the navigation
channel which recorded water level, current speed, and direction at an acquisition frequency of
2 Hz. Simultaneously, an automatic identification system (AIS) receiver
permitted us to acquire traffic data for the area, relating every observed
event to the specific ship in transit in the measurement section. Tidal
levels were referred to the local datum, while depression wakes were
calculated as the difference between maximum and minimum levels at the
passage of a ship. Figure 4 shows the ship wake generated by the passage of a
large commercial vessel on 6 April 2016. A relatively small rise in the water
level (
The simulation of the pressure and velocity fields around the hull of the ship was carried out using the uRaNSe-Xnavis simulator (Broglia et al., 2014; Di Mascio et al., 2007, 2009). It is a finite volume solver based on the discretization of free surface, incompressible, viscous, high-Reynolds-number fluid equations (unsteady Reynolds averaged Navier–Stokes equations). The fluid-dynamical field is discretized using the overlapping grid approach, with an increase in spatial resolution close to the hull and the free surface, and then degrading in the channel close to the tidal flat. The channel and the lateral zone where the bathymetry is deeper than 2 m represent the uRaNSe-Xnavis domain. An example of the computational grid is shown in Fig. 5a. The model was tested and calibrated using the data recorded along the MMIC and then applied to forecast the movement of a typical cruise along the planned MVC.
Water level (m a.m.s.l.) recorded on the bottom of the MMIC and
computed by SHYFEM on the tidal flat 300 m from the channel caused by the
transit on 6 April 2016 of the Cargo-Hazard A. The commercial vessel, which
was used as a reference, is 280 m long, 40 m wide, and characterized by a
gross tonnage of 66 433 t. The ship speed
Water level behavior at various distances from the center of the MVC
as computed by the hydrodynamic model for a liner ship moving at
The high-resolution CFD steady-state dimensionless results, in terms of water
level and velocity, were interpolated on a regular
0.7
A set of four scenarios was produced, considering a typical liner ship with
the geometrical characteristics provided in Table 1 and moving in the system
at
Short (from minutes to hours) and medium (i.e., months) term simulations addressing the effects of the tidal fluctuations and ship wakes have been carried out by a flow and transport uncoupled approach using subsurface modules FLOW3D and TRAN3D of the finite element CATchment HYdrology Flow-Transport (CATHY_FT) model (Camporese et al., 2010; Weill et al., 2011). The mixed hybrid finite element–finite volume COUPHYB simulator (Mazzia and Putti, 2006) for the solution of density-dependent flow and transport has been used to perform long-time (i.e., decades) analyses of seawater leakage into the aquifer system below the lagoon bottom.
Geometry of the typical cruise vessel used in the uRaNSe-Xnavis and SHYFEM simulations.
Section-1: computed maximum depression (m) in the MVC and percentage attenuation at various distances from the channel center for different speeds of the cruise ship. Similar values are obtained in Section-2.
The numerical simulations were carried out on 2-D vertical sections, in
particular along Section-1 and Section-2 (Fig. 1b), which can be considered
representative of the lagoon hydrogeological setting with respect to the MVC
excavation. Indeed, the MVC bottom planned at
The evolution of the pressure and velocity fields in the shallow subsurface due to water level fluctuations in the lagoon was simulated by FLOW3D, neglecting the possible effects of different groundwater and surface water salinity. Indeed, density-driven processes are characterized by a much longer characteristic time than those typical of tidal regimes and ship wakes.
FLOW3D solves the groundwater flow equation in saturated conditions:
FLOW3D was initially used to calibrate the hydrogeologic properties of the
upper units. This was carried out by running the model in the present
configuration and matching the pressure records available at the two 10 m
deep piezometers. Dirichlet conditions representing the observed tidal regime
over spring 2016 were imposed on the top boundary, which constitutes the
lagoon bottom; a constant head
Hydrogeological parameters obtained by the model calibration and used in the numerical simulations. Aqt-2 is lacking in Section-2.
The calibrated model was then used to quantify the effect of the MVC in terms
of pressure and flow field on the subsurface in relation to
tidal regime: the same boundary conditions used for the model calibration
were applied on the domain with the MVC and the results obtained with the two
configurations were compared. The simulations were carried out using a time
step Ship wakes: we investigate the effects induced by both the commercial
vessel of Fig. 4, which represents an extreme of the perturbations possibly
stressing the system, and a cruise vessel with
Transport processes in the subsurface of non-reactive chemicals are
described by the classical advection–dispersion equation (Bredehoeft and
Pinder, 1973):
Similarly to FLOW3D, TRAN3D solves Eq. (2) using a Galerkin FE approach and a
weighted finite difference time integration scheme (Gallo et al., 1996).
TRAN3D was used to quantify the possible exchange of the contaminants
detected in the subsurface between the groundwater and the surface waters
along the MVC trace. In particular, attention was focused at the ship wakes
that, being strongly asymmetric, favor the contaminant outflow into the
lagoon waters much more than the tide fluctuations, which are almost
symmetric with respect to the mean sea level. The velocity field at each time
step is provided by the outcome of FLOW3D. Based on the general outcome of
the chemical characterization, the following simplifying assumptions were
adopted in the modeling setup.
The Table 4 provides the
The simulations were carried out by normalizing the actual
The transport process was investigated over a multiple ship passage.
Specifically, the transit of
Average contaminant concentration (labile fraction) in the depth
range between
As reported above, the available hydrogeological investigations revealed that the subsurface of the Venice Lagoon is characterized by significant stratification in terms of water salinity, reflecting the layering of the sedimentary sequence. Dredging of new canals in such an environment can cut impermeable units, producing a certain mixing of the salt concentration between the shallowest contaminated units and the fresher underlying layers. Due to the lack of a significant pressure gradient between the various geologic layers, the difference in groundwater density is likely the main driver of salt transport.
Longitudinal (
Following Bear (1979), the mathematical model of density-dependent flow in
aquifer systems can be written using the equivalent freshwater pressure head
In the COUPHYB simulator (Mazzia and Putti, 2006), the system (5) is solved
numerically using a mixed hybrid finite element scheme for the flow equation
and a mixed hybrid finite element–finite volume time-splitting-based scheme
for the transport equation. This approach is computationally effective and
accurate, introducing minimal numerical diffusion even in the absence of
physical dispersion, and when the process is advection dominated or density
changes yield instabilities in the flow field. COUPHYB was applied to the
same triangulation shown in Fig. 7b, with the solutions in terms of
Behavior versus time of the pressure at a depth of 13 m b.m.s.l.
in
Based on the hydrogeological information, the simulations were carried out
starting from an initial
Section-1: computed pressure distribution
Section-2: computed pressure distribution at the times T1–T6 highlighted in Fig. 4 during the transit of the Cargo-Hazard A. The vertical exaggeration is 8.
The dredging of a new relatively deep channel in a tidal environment can perturb the natural pressure and flow fields in the shallow subsurface. Quantification is obtained by comparing the results provided by the calibrated FLOW3D for the two simulated sections in the present condition and after the MVC excavation. Figure 8 shows the behavior of the pressure at a depth of 13 m b.m.s.l. in correspondence to the MVC symmetry axis. The point is located within Aqt-2 and Aqf-1/2 in Section-1 and Section-2, respectively. The effect of the different stratigraphic sequence is obvious, with an approximately 85 % reduction of the pressure fluctuation within the clayey layer with respect to the oscillation of the lagoon level (Fig. 8a). The MVC dig reduces the time lag and the attenuation of the perturbation at depth. These effects are quantitatively negligible and develop only in the surroundings of the channel (Fig. 9).
Computed
Although a depression wave caused by a ship transit develops over a period of a couple of minutes (Figs. 4 and 6), which usually is a very short time for hydrogeological processes, its height is sufficient to affect significantly the groundwater pressure and flow fields in the proximity of the channel bottom.
Figure 10 provides the pressure distribution in the surroundings of the MVC,
Section-2, computed by FLOW3D at the significant time steps highlighted in
Fig. 4 during the transit of the Cargo-Hazard A. The pictures clearly show
how the pressure gently rises before the ship passage, significantly
decreases soon after the transit, and then recovers with a gradient that
changes its sign during each phase. The pressure change affects the portion
of the subsoil down to the top of the first clay layer below the channel
bottom, and extends laterally up to about 30 m from the channel slope
(Fig. 11a and b for Section-1 and Section-2, respectively). The typically
short duration of such events precludes the propagation of the pressure
change far from the channel edges. The velocity field in correspondence to
the maximum depression is provided in Fig. 11c and d. The ship generates a
sort of “piston effect” with an efflux distributed along the whole channel
bottom and slope. The maximum values of the velocity amount to
Behavior of
The seepage from the MVC bottom and slope (between times T2 and T4 in
Fig. 10) amounts to
The same computation was carried out for the ship wake caused by a cruise
ship moving at 7.7 knots. The results provided by the hydrodynamic model
(Fig. 6) were used to force FLOW3D. The computed subsurface pressure and
velocity fields are qualitatively similar to those previously described
(Figs. 10 and 11), with smaller values determined by the lower wave height
used as a forcing factor. The average efflux along the whole MVC decreases to
The effect of the transit of several ships along the MVC in terms of
contaminant release from the subsurface into the lagoon was investigated
through TRAN3D. The behavior of the released mass
Ratio between the reference mass
Combining the actual initial concentration of the various anthropogenic
contaminants (Table 4) and the TRAN3D outcomes allows estimation of the real
mass
Section-2: behavior of
Mass of various chemicals expelled from the lagoon subsurface
through the MVC bottom versus the transit number of
Relative salt concentration in
Figure 15 shows the outcome of COUPHYB in terms of relative concentration at
the end of the simulation period, i.e., 10 years after the inception. The
results are presented for both the sections addressed by the study. The
effect of the MVC excavation is pointed out by comparing the two settings,
i.e., the present condition and that where the MVC is dredged. Cutting the
top clayey layer, the excavation favors the propagation at depth of the
seawater, with an increase in
We are aware that the analyses presented here rely on a number of simplifications, with results that can be affected by (1) the approximated modeling approach used for the simulations, (2) the representativeness of the hydrologic and geological information used to calibrate the model parameters, and (3) the boundary conditions and the factors forcing the system (Tsang, 2005).
In this study, we have elected to use a 2-D modeling approach along vertical sections instead of a 3-D analysis. The choice is warranted by the shape of the domain possibly influenced by the excavation, which is much more elongated along the MVC trace than in the orthogonal direction. Moreover, the groundwater flow induced by a ship moving along the canal, which represents the main factor forcing the system in the short and medium term, is characterized by a net component along the direction orthogonal to the ship track only. The two selected sections are representative of the main variability characterizing the hydrogeological architecture of the Venice Lagoon subsurface. Several boreholes drilled in the past and the extensive seismic survey carried out during the initial phase of the study provided an accurate characterization of the shallowest 50 m thick depositional sequence and revealed that, although local sedimentary anomalies are frequently encountered, the presence of a buried large channelling system cutting Aqt-2 in a specific portion of the study area is the main feature to be accounted for in the modeling investigations.
A critical issue is related the calibration of the hydrogeological parameters
(
Concerning the transport model, a sensitivity analysis of
Finally, the simulations have been carried out assuming a null natural flow due to the lack of specific information. Because of the shallow depth of investigation and the position of the study area at the edge of the flat Po Plain, the natural hydraulic gradient is certainly very small. Therefore, the results of the short- and mid-term simulations are almost unaffected by the assumption.
Lagoons are natural environments that undergo a continuous increase in anthropogenic pressure. On the other hand, their present hydraulic and morphologic equilibrium is in some cases artificially preserved by human interventions, which are aimed at contrasting the combined effects of sea level rise associated with global warming and land subsidence of natural and anthropogenic origin.
The Lagoon of Venice is a paradigm of the complexity in the interactions among economic, social, and environmental needs (Rinaldo, 2001). This holds for both the surface and subsurface environments. Investigations carried out over recent years (Teatini et al., 2011; Viezzoli et al., 2010) revealed that fresh groundwater resources are found at quite low depths, i.e., 30–40 m and even less than 10 m beneath the lagoon bottom. Like the hydrogeological settings of other lagoons (e.g., Santos et al., 2008), the silty-clayey layer marking the boundary between the marine Holocene and continental Pleistocene deposits precludes or at least reduces the vertical leakage of the salt waters downward into the underlying freshwater aquifers. However, a large petrochemical industrial district, the PMIZ, has been in operation since the 1950s at the lagoon–mainland interface, representing the main source of soil and water pollution around the area (e.g., Zonta et al., 2007). Despite an almost 50 km long cut-off wall built up along the canal banks of the PMIZ to prevent discharge of contaminated waters into the lagoon (Paris et al., 2011), results from chemical analyses provided evidence of a high content of Hg, Zn, and other metals in the bottom sediments and pore water not only in front of the industrial site (Gieskes et al., 2015), but also at distance. Although quite gentle in the shallower subsurface, the natural groundwater flow from the mainland seaward has likely transported the contaminants to the lagoon ecosystem over the last decades.
The quality of an aquatic ecosystem is set by the quality of its sediments. Sediments are a sink for pollutants and nutrients, but can also act as a long-term source as well, with the groundwater playing a key role in the redistribution of hazardous substances in other environmental compartments, such as the biota, upon changes in the physical–chemical conditions. Similarly to other coastal lagoons with an inner port and/or an industrial zone, for example the Maryut Lagoon, Nile Delta (Oczkowsly and Nixon, 2010), or Lake Macquarie, New South Wales, Australia (Thomsen et al., 2009), anthropogenic contaminants have been detected in the Venice Lagoon subsurface.
Cutting of the clayey layers that characterize the shallower Pleistocene and
Holocene deposits in coastal zones can significantly increase the exchange
between groundwater and surficial water bodies and the anthropogenic and/or
natural contaminants transported with the waters. For example in the
Mangueira Lagoon, which is a large (90 km long), shallow (
The evaluation of the possible impacts of the MVC excavation must be investigated in this context. The contaminants in the labile and bioaccessible fraction along the MVC designed path and depth range might be released into the lagoon because of ship wakes, with a considerable amount in the mid-term. The potential bioavailability and bioaccessibility of contaminants were confirmed by the high ecotoxicity levels shown by elutriates obtained from sediment samples collected along the MVC trace. They generate concern as it represents an easily exchangeable fraction which can move from sediment to water. Proper measures should then be planned to limit the risk of contamination of the lagoon water during the years following dredging. Moreover, the modeling study provides a first evaluation of how the interruption of the caranto aquitard by the MVC digging favors the saltwater flow deepward in a medium to long time interval, in the range of a few decades. Electromagnetic surveys and marine electric topographies carried out in the part of the lagoon between Venice and Chioggia clearly pointed out that groundwater with a salt content similar to the marine waters is found beneath 5–15 m b.m.s.l. only where the caranto layer is cut, generally by natural erosion or channel excavation (Tosi et al., 2009; Zecchin et al., 2014). The modeling results suggest that the salt contamination remains localized around the incision, with an important role in controlling the depth of percolation played by the actual layering of the sedimentary deposits below the channel bottom.
This study presented the results of a systematic modeling investigation of the possible hydrological processes activated by digging a large and deep navigable channel through a shallow lagoon in the hydrogeological system below the lagoon bottom. Although the practice of dredging channels is quite common in coastal lagoons to facilitate the movement of goods and people from the outer sea to inner ports, a specific investigation of these effects had never been carried out before this study.
Here we focus on the specific case of the Venice Lagoon, where the new 10 m
deep, 3 km long, and 150 m wide MVC canal is under planning. Results from
the modeling approach show that a significant influence on the
groundwater–surficial water exchange is expected to be produced by the
excavation and the transit of cruise vessels along the channel. Each large
ship in transit can produce a depression wake on the order of 1 m, thus
pumping out the groundwater from the shallow deposits around the excavation.
Although for a given channel section the ship wake lasts a couple of minutes
only, the large groundwater velocity induced in the surroundings of the
excavation combined with the length of the MVC are responsible for an efflux
on the order of 50–100 m
Proper measures should then be planned to limit the risk of contamination of the lagoon water during the years following dredging. Moreover, considering the importance and the fragility of the Lagoon of Venice, if the MCV is to become a real project, a variety of new and more detailed information will necessarily have to be collected to provide a more accurate quantification of the possible environmental impacts of the canal dredging on the subsurface system of the Venice Lagoon. For example, additional piezometers should be placed along the MMIC to verify in advance the ship-wake effects on the subsoil and along the planned MVC path to characterize the natural flow regime; groundwater age through isotope analyses of water samples should be determined to evaluate the groundwater origin and fate; and pumping and tracer tests should be planned to characterize the hydrogeological properties of the shallow aquifers below the lagoon bottom.
Data are available upon request to Pietro Teatini (pietro.teatini@unipd.it).
The authors declare that they have no conflict of interest.
The research was funded by the Venice Port Authority, Italy, and partially supported by the Flagship Project RITMARE – The Italian Research for the Sea, CNR-MIUR, National Research Program 2011–2013, “Linea SOLVE”. Edited by: Alberto Guadagnini Reviewed by: two anonymous referees