- Email: [email protected]
Seawater intrusion via surface water vs. deep shoreline salt-wedge: A case history from the Pisa coastal plain (Italy)
Groundwater for Sustainable Development 2-3 (2016) 73–84
Contents lists available at ScienceDirect
Groundwater for Sustainable Development journal homepage: www.elsevier.com/locate/gsd
Research paper
Seawater intrusion via surface water vs. deep shoreline salt-wedge: A case history from the Pisa coastal plain (Italy) F. Franceschini a,b,n, R. Signorini a a b
Earth Sciences Department of Pisa University, Via S. Maria, 53, 56126, Pisa, Italy Environmental Protection Agency of Tuscany (ARPAT), Via Vittorio Veneto, 27, 56127, Pisa, Italy
art ic l e i nf o
a b s t r a c t
Article history: Received 24 July 2015 Received in revised form 29 April 2016 Accepted 14 May 2016 Available online 10 June 2016
This study analyses data regarding the quality of surface water in the Pisan plain between the mouth of the Arno river, Pisa and Livorno in order to identify the origin of the high chloride content derived from the sea water in areas far from the coast. We used existing environmental monitoring data from the study area integrated with data from the present study. The study involved measuring the piezometric levels and electrical conductivity (EC), together with a chemical analysis of the surface water and phreatic aquifer (water table aquifer) samples. Data acquisition through sampling along coastline piezometers, was carried out in 2008. In the summer of 2012 we supplemented data regarding areas further inland with surveys of piezometric levels and EC in the hydrographic network and shallow wells. The distribution of salinity was compared with a three-year monitoring survey carried out at the water sampling points in the Tombolo forest, an area about 5 km from the coastline. All the collected data show the presence of significant salinisation generated by the sea water, which is carried towards the inner plain through an artificial watercourse (Navicelli channel) and its hydrographic network. In the dry season this channel contains no fresh water from upstream and sea water tends to flow up to the town of Pisa, about 7 km from the coast. The sea water carried by the channel infiltrates into the shallow aquifers directly or through the minor hydrographic network, thus polluting the internal areas of the Pisan plain. Potential hydraulic connections with the intensively exploited deeper aquifers might lead to the salinisation of the same aquifers. This phenomenon could be confused with seawater intrusion from the coast (deep shoreline salt-wedge). & 2016 Elsevier B.V. All rights reserved.
Keywords: Salinity Navicelli channel Salt pollution Land management Phreatic aquifer
1. Introduction The northern coastal plains of the Mediterranean are subject to significant human pressure, which often leads to the depletion and deterioration of water resources. One of the most recurring effects is the variation in the natural equilibrium between fresh and sea waters and the consequent advance of the sea water through the coastal aquifers. Many Italian coastal areas are affected by seawater intrusion (Barrocu, 2003). In Tuscany, several critical situations have been highlighted (Pranzini, 2002; Grassi et al., 2007). The salty waters in the multilayered aquifer system of the Pisan coast have been known about for at least 20 years, and numerous studies have highlighted how the problem affects the environment (Signorini, 2008; Doveri et al., 2010; Butteri et al., 2010; Giannecchini et al., n Corresponding author at: Environmental Protection Agency of Tuscany (ARPAT), Via Vittorio Veneto, 27, 56127, Pisa, Italy E-mail address: [email protected] (F. Franceschini).
http://dx.doi.org/10.1016/j.gsd.2016.05.003 2352-801X/& 2016 Elsevier B.V. All rights reserved.
2010). Seawater intrusion increases especially during the summer months due to the intensive exploitation of civil, industrial and agricultural uses with the consequent deterioration of the qualitative state of the groundwater. In addition to degrading the quality of groundwater and restricting the agricultural use of the land, increased salinity reduces the diversity of ecosystems (Amores et al., 2013) and leads to the development of halophilic species (Williams, 1987). Sea water contains approximately 35,000 mg/L of dissolved solids, which include about 20,000 mg/L of chloride. These chloride values are normally associated with EC values exceeding 45,000 μS/cm. Fresh groundwater in most coastal areas generally contains less than 100 mg/L of chloride. However, concentrations in excess of 100 mg/L are not conclusive evidence of seawater intrusion because they could be due to airborne sea spray in precipitation, well pumping, local sources of chlorides, including septic systems or animal manure, or to relic seawater in the aquifer. Chloride concentrations higher than a few hundred mg/L can cause adverse effects on many crops (USEPA, 1992; Mara, 1998). In order to provide criteria to assess the chemical status of groundwater
74
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
bodies according to the 2006/118 European directive, Italy (Legislative Decree No. 30/2009, 2009) defined the threshold values with respect to chloride and EC as 250 mg/L and 2500 μS/cm respectively. Seawater intrusion is the introduction of saltwater into inland freshwater systems. It can move into the freshwater system through both surface water and groundwater connections to the sea. Seawater intrusion via groundwater connection is normally directed along the shoreline as a deep salt-wedge. Human creation of channels and ditches for drainage of freshwater systems encourages saltwater intrusion where channels connect with a saltwater source (Turner, 1997; Day et al., 2000). This is particularly evident when river flows are low and or during high tides or storm surges (Anderson, 2002; Switzer, 2014). Storm surges and strong wind events can push saltwater into freshwater channels encroaching on freshwater coastal systems. Moreover, eustatic sealevel rise is projected to be 1–2 mm/yr, which is the change in sea level rise with a change in the ocean volume due to climate change (Milly et al., 2003). Seawater intrusion can easily propagate in-land along rivers and channels then infiltrate downward to the groundwater zone. In recent decades, the excessive overexploitation of fresh water resources has exacerbated the situation leading to periods when river flow is reduced to a threshold level beyond which it can no longer prevent sea water from moving inland. The salt water can go back to the lower courses of the rivers for miles, changing the chemical characteristics of the waters of the coastal plains (including groundwater) rendering it too saline for drinking or irrigation water. Coastal plains are underlain by alluvial aquifers which can be highly heterogeneous sequence of sands and silts. Seawater
naturally intrudes into coastal aquifers as a wedge due to density differences. It is very important to distinguish between natural saline wedges in the subsurface and saline intrusion via surface water flow followed by infiltration to the subsurface. How these effects are monitored is, in fact, very different as are the actions required to remove or mitigate the impacts. We studied the surface and subsurface waters in an area of the Pisan plain in terms of geological and hydrogeochemical aspects, with reference to the relationships of feeding and drain between the water table aquifer and the hydrographic network. The study area includes the river Arno and the Scolmatore channel to the north and south respectively, and the coastline to the west. The eastern boundary includes Pisa and the Coltano hills (Fig. 1). The study area includes the Migliarino-San Rossore-Massaciuccoli Regional Park and has a high environmental value, due to its great variety of natural habitats, ranging dunes to sandy shores and from hygrophilous forests to marshlands. The results provide a tool for managing a network of datasets to help local authorities to improve land management and the monitoring of water resources, and also to prevent possible environmental damage caused by salt pollution.
2. Geological setting The coastal plain is located in the west of Tuscany, north of Livorno (Fig. 1a). It is bounded by the Pisan Mountains to the NE, the Livorno hills to the south, and the Ligurian Sea to the west. The Arno valley consists of a succession of estuarine clays, subdivided into three, vertically-stacked transgressive-regressive millennial-
Fig. 1. (a) Location map of the study area (b) Lithological map, derived from the Provincial Administration of Pisa (Provincial administration of Pisa, 2005) geological map. The red lines indicate the section traces of Fig. 5.
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
scale depositional cycles, which are replaced at landward locations by coastal plain deposits (Amorosi et al., 2008). After two major transgressive cycles (Upper Miocene, Lower Pliocene), which led to the deposition of conglomerates, sands, clays and evaporitic sediments, an important regressive phase ensued during the Middle Pliocene. This led to land emergence and widespread erosion. Because of glaciations, significant eustatic fluctuations occurred during the Pleistocene. Initially (Lower Pleistocene), sediments of predominantly marine and saline environments were deposited in the area and then, during the Middle Pleistocene, gravels, sands, silts and clays followed due to the flooding of the Arno-Serchio water system. Then (Upper Pleistocene–Holocene), mainly fluvial and marshy sediments were deposited, due to reduced fluvial activity linked to both the end of the lowering of the eustatic sea-level and the drying of the climate. This climatic change also led to the deposition of aeolian sands, found both in depth and at the surface, where aeolian sands are partly linked to modern dunes during Holocene. The presence of dunes and the rise in the sea level hindered the outflow of waterways to the sea, with the consequent formation of marshes and lagoons, and the deposition of fine sediments along with the development of peat and lignite (Fancelli, 1984; Grassi and Cortecci, 2005). On the basis of pollen data from 19 continuously-cored boreholes drilled south of the Arno River, near Livorno, GallettiFancelli (Galletti-Fancelli, 1978) assigned the uppermost 30 m to the sub-Atlantic-Boreal phases (ca. last 10,000 cal yr BP). During these phases there were alternations in subsidence and a slight rise in sea level. Following a rapid rise in the sea level during the middle Holocene, the Arno paleovalley was flooded and the interfluves submerged by the sea. The subsequent evolution of the Pisa area saw the development of a wide, thick, clay-dominated, lagoonal succession, controlled by the inherited paleotopography (Rossi et al., 2011). At the time of maximum marine ingression (7.8 cal ka BP, (Amorosi et al., 2013)), the shoreline was located more than 7 km inland from its present position. From Pisa to the coastline, the plain (Fig. 1b) is currently composed of alluvial muddy and sandy deposits that develop seaward into a wide sandy strandplain system composed of several juxtaposed coastal dune ridges (CDRs) and narrower interposed depressions. The alignment of these ridges records the position of the shoreline over the last 2500–3000 years (Pranzini, 2001). According to Sarti et al. (2008) the genesis of CDRs is closely related to a classical model showing how a strandplain advances. The process begins with the formation of an embryonic dune which occurs when the whole foreshore-berm-backshore system migrates inland during a storm. At the end of the storm, a storm berm forms. This constitutes an obstacle in the backshore where the sand begins to accumulate forming an embryonic dune. There are also flooding deposits from the Arno river consisting mainly of fine sands in the proximal area and silts and clayey silts in the distal area, inter-dune and retro-dune fluvial-palustrine deposits consisting of moderately thick silts with peaty levels. This system of dunes contains a phreatic aquifer of regional importance in terms of the subsistence of the current ecosystems. It was previously used for irrigation with shallow wells, but it is no longer used due to the presence of high salt contents. Inland from the plain, where the sandy aquifer is confined by the floodplain deposits, the feed comes from the piedmont zone where it is in direct contact with the fan gravelly materials, and through the alluvial deposits of the Serchio River (Spandre et al., 1999). Below the sandy phreatic aquifer, the alternation of lithological layers with variable permeability has led to the creation of confined aquifers which are only partially independent of each other. These layers can be affected by direct hydraulic exchanges or by drainage induced by the various hydraulic gradients that characterize them. In
75
fact in this variable permeability, the lack of continuity and adequate thickness, do not guarantee a complete hydraulic separation between higher permeability levels, thus triggering interconnections and mutual exchanges (Butteri et al., 2010). The hydrogeological units have been classified as aquifers and aquitards after parameterization through estimates of permeability (Baldacci et al., 1994). In the inner portion of the Pisa plain, Baldacci et al. (1994) reported a complex multi-level confined aquifer that contains two major confined aquifers defined, from the top to the bottom as, “1st sandy confined aquifer” and “1st gravelly confined aquifer”. Moving towards the coast, the 1st sandy confined aquifer becomes the phreatic aquifer contained in the CDRs. The gravelly aquifer (GA) is represented by clastic fluvial deposits with high permeability, which originated during the migration of the Serchio and Arno riverbeds (Della Rocca et al., 1987). The roof of this aquifer (Baldacci et al., 1999; Della Rocca et al., 1987) is about 10 m thick with good permeability, and deepens northwards and westwards. Its minimum depth (30 m) are towards the southern border of the plain, whereas the maximum depth (180 m) are towards the north-western part of the plain. The gravelly aquifer is found at a depth of about 140 m in the southwestern zone of the city of Pisa, and seems to disappear northwards. In agreement with Baldacci et al. (1994), the GA is mainly recharged by the direct infiltration on the Livorno hills through the pebbly bodies that lie in the piedmont areas and in the intermountain valley of the Pisan hills. Isotopic content values of the water circulating in the confined gravelly aquifer indicate a recharge altitude higher than the plain (Grassi and Cortecci, 2005). The GA is a very important water resource for civil, industrial and irrigation uses. The overexploitation of the GA leads to water levels below sea-level in most of the area. There are two local and more accentuated piezometric depressions in San Piero a Grado and Calambrone, where the groundwater is heavily exploited. In particular, up to 1.5–2 km from the southern coastline (Calambrone area) and in a narrow area close to the coastline to the north (Arno mouth), the seawater intrusion develops directly in the confined gravelly horizon in contact with the sea floor. In the S. Piero a Grado area the mixing with seawater should be controlled by the seawater intrusion that affects the Arno River-phreatic sandy aquifer system and by the hydraulic connection between the sandy and gravelly aquifer (Butteri et al., 2010). The deposits found at Vicarello (Vicarello sands in Fig. 1b) are morphologically in relief on the plain (about 9 m). They emerge at the foot of the Pisan and Livorno hills (Fig. 1a). These deposits were previously related to an aeolian environment (Della Rocca et al., 1987; Lazzarotto et al., 1990), and were then reinterpreted as a fluvial environment (Carosi et al., 2010). As can be inferred from the stratigraphic sequence shown in this work, these deposits are likely to be in stratigraphic contact with the CDRs.
3. On-site surveys and analytical methods In order to identify the hydrogeochemical conditions of the CDRs we have evaluated the existing data and carried out onsite surveys. Since our aim was to highlight any criticality in the balance between the feeding and recharge of the aquifer contained in the CDRs, all data refer to the summer season. The geomorphological features and reconstruction of the hydrographic network were obtained using a high-resolution map made using 3D laser scanning (LIDAR) (Fig. 2a). For the Pisan plain a dense digital elevation model (DEM) with an accuracy of less than 10 cm is available at Pisa municipally authority. Landforms and hydraulic network identification was based on photointerpretation, morphometric analyses on the LIDAR digital elevation model (DEM) and a field survey.
76
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
Fig. 2. (a) LIDAR map of the study area. Dark shades indicate lowland (b) Hydraulic map derived from survey LIDAR. Natural (no colour) and mechanical (pink colour) drainage basins and some main drainage channels are reported in the map. The dewatering plants are numbered as table 1a.
Remote sensing analyses were carried out in order to detect traces of past drainage networks and wetlands (marshes, swamps, and ponds). The DEM available for this study came from a LIDAR survey carried out during 2008 and commissioned by the Ministry of the Environment and Protection of Land and Sea. Fig. 2a shows a raster map where each cell in the matrix has a spatial resolution of 1 metre and contains the elevation information. The plain hosts a large number of water-wells with depths varying from a few metres to nearly 250 m, which enabled the stratigraphic succession of the study area to be reconstructed. The stratigraphic reconstruction was carried out with stratigraphic well logs derived from the database of the Provincial administration of Pisa (2005) and various infrastructural engineering projects. Regarding the hydrogeochemical features, we used data by Signorini (2008) and Giannecchini et al. (2010) for the CDRs, and Grassi et al. (2007), Butteri et al. (2010) for the gravelly aquifer. The hydrogeochemical information on CDRs was integrated with two field surveys carried out in June and July 2012. With regard to the first survey, the piezometric level, temperature and EC were measured and the chloride and sulphate concentrations were determined from a selection of surface water samples from a hydraulic network and groundwater from existing piezometers and wells. In the subsequent surveys we measured the EC of the Lamone and Navicelli channels. Further EC and chloride concentration measurements were drawn from the ARPAT database (http://sira.arpat.toscana.it/sira/ acqua.php) concerning periodic monitoring performed in the
previous 10 years in the study area by selecting, to avoid excess data, those related to the summer season. To reconstruct the EC distribution, data by Silvestri and Gorreri (2008) were also used. The piezometric levels and variations in EC in the forest of Tombolo come from the monitoring of surface water and shallow groundwater as part of a about 3-year (March 2010–November 2013) study commissioned by the U.S. Army base at Camp Darby, which occupies a substantial portion of the area (Franceschini and Heusch, 2013). The monitoring concerned surveys of surface water (11 stations) and shallow groundwater (12 stations). The shallow groundwater monitoring stations consist of 14 mini-piezometers placed at a maximum depth of 3.5 m from ground level. In addition, measurements of EC, pH and temperature were taken from October 2012 until November 2013. These mini-piezometers intercept only the most superficial portion of the phreatic aquifer in order to separate them from the other groundwater samples considered here which represent the greater thickness of this aquifer. Temperature and EC (μS/cm) were measured in situ using a multiparametric probe (WTW Multi340i). The analyses to determine other chemical parameters reported in Table 2 were performed at the laboratories of Institute of Geoscience and Earth Resources (IGG) of the Italian National Research Council (CNR) and ARPAT laboratory in Pisa with regard to the data in Table 2b. Major anions were measured, such as chlorides, nitrates and sulphates, through an ion chromatography system (Dionex DX-120) using undiluted water samples.
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
4. The hydrographic network The coastal area of the Pisan plain is characterized by a Mediterranean climate with hot, dry summers and cool, wet winters. Annual rainfall ranges from 750 mm to about 850 mm. The annual evapotranspiration is 587 mm. The study area is characterized by an area of higher elevation (forest of Tombolo and coastal area) with altitudes that range between 2.5 and 0.5 m a.s.l., and depressed areas with altitudes of about 0.5 m a.s.l. and heights up to 1.5 m a.s.l. (Fig. 2b). Many depressed areas, especially around San Piero a Grado and along the Lamone channel, have been drained to create agricultural land. In winter the run-off rainwater flows towards ditches and channels in depressed areas (low waters) and are lifted by dewatering plants (mechanical drainage); in higher elevation areas (high waters) it is taken directly to the sea through the main watercourses (natural drainage). The hydraulic reclamation for mechanical drainage removes the excess and stagnant waters, but also prevents the rise of phreatic water, which is harmful to agricultural crops. The basins within the study area, identified through the DEM survey, are shown in Fig. 2b, with the main flow water channels and the mechanical drainage draining pump stations. In the study area there are five main mechanical drainage basins and five natural drainage basins (Table 1a and b). Of the most important land reclamation, carried out between the end of the 19th century and the first half of the 20th century, the Navicelli channel, is worth mentioning. This channel represents the main watercourse of low water from mechanical drainage. It is a construction which enables the natural drainage of part of the plain (basin in Table 1a and b) that houses important infrastructures (e.g. an international airport) and favours good transport by ship between the industrial area of Pisa and the harbour in Livorno. The Navicelli channel is about 16 km long and runs straight from southern Pisa to the northern border of the Livorno harbour where it flows into the Scolmatore channel immediately before its confluence into the sea. It was built with a steady curved section, which is almost trapezoidal with 13 m width and 3 m a.s.l. steady altitude of the bed channel, with a 1:2 slope (Pozzo and Todaro, 1923). Its construction was completed in 1924 after about three years of intensive work. The initial project was supposed to connect it with the river Arno, but it was not completed. The upper stretch of the Navicelli channel near Pisa, named “Incile”, is not directly connected with the main
77
channel. This stretch also has a wastewater input. Waters resulting from reclamation areas from the whole coastal plain south of the Arno flow into the Navicelli channel (Table 1a and b). The channel overlaps the previous artificial hydrographic network since the Middle Ages for reclaiming the depressed areas of south Pisa (Fig. 2b); the old Navicelli channel, built during Medici age, was the main flow channel of these areas. Other tributaries from natural drainage include Sanguinetto and Mezzanina channels from the Sanguinetto basin, Scolo di Pisa channel from the San Giusto basin and San Ermete channel from eastern Pisa. The tables highlight that the Navicelli channel is fundamental in that it provides the hydraulic balance of the whole study area in the rainy periods. During low precipitation periods, instead, the channel is in hydrochemical and hydrostatic continuity with the sea, conveying “pure” seawater into the inland plain. Fig. 3 shows the results of the survey (July 2012) where this situation is evident. The survey was carried out during a significant interval in the tidal oscillations in order to highlight the quality of the water during a period of minimum flow. Two measurements at different depths were carried out for each point. In Fig. 3 blue and red represent measurement at a depth of 20 cm and 50 cm, respectively. In the final stretch of the channel the contributions of channels fed by the highland with fresh water are clear. In the Fosso Chiara channel the rising salt water associated with the presence of fresh water inflows is particularly evident because this channel is connected to the continental drainage network. Significant contributions of fresh water are absent so the water in the entire length of the Navicelli channel is characterized by high salinity, very close to that of the sea.
5. Coastal dune ridges (CDRs) features The CDRs (the portion of the Pisan plain to the west of the section E–E’ in Fig. 1b) cover the entire area between the Arno river and Scolmatore to the north and south respectively and the coastline to the west. To the east the CDRs limit coincides with depressed areas characterized by fine flood and marsh plain deposits (Coltano and Stagno basins in Fig. 2b). South of the study area (sections A and B in Fig. 4), the sub-surface is composed of a 20–25 m thick layer of sandy deposits interpretable as upper and lower shoreface (CDRs). Below is a thick layer (30–40 m) of lagoon/ estuary deposits constituted by silt and clay (for example, TPW
Table 1 (a) Mechanical drainage basins in study area. (b) Natural drainage basins in study area. See Fig. 2b about location of watercourses and basins. (a) Name
Main drainage channel (high waters)
Dewatering pumping system (numbers in Fig. 2b)
Final Watercourse (low waters)
Arnino (North Lamone) South Lamone Stagno-Coltano South Pisa Vettola
“Collettore settentrionale” “Lama Larga” (Northern watercourse) “Lamone” “Lama Larga” (Southern watercourse) “Collettore Padule Maggiore” “Collettore dello Stagno” “Venticinque”, “Sofina”, “S.Ermete” Old Navicelli
Marina di Pisa (1)
Arno river
Calambrone (2) Ragnaione (3) Airport (4) La Vettola (5)
Scolmatore channel Navicelli channel Navicelli channel Navicelli channel
(b) Name
Main drainage channel
Final Watercourse
San Giusto Sanguinetto Tombolo Coltano (dune) Camp Darby
“Scolo di Pisa” “Sanguinetto”, “Mezzanina” “Collettore orientale” “Mandracchio” Depot area
Navicelli channel Navicelli channel Navicelli channel Navicelli channel Fossa Chiara channel
78
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
Fig. 3. EC trend along the Navicelli Channel. Survey of July 2012. Triangles represent the Navicelli water, squares represent the water of his tributaries. The tidal oscillation from Leghorn harbour tide gauge are showed in the box.
Fig. 4. CDRs geometry in study area as reconstructed after stratigraphic interpretation. Section trace is reported in Fig. 1b. In cross-sections CDR is drawn in yellow, gravelly aquifer is drawn in blue.
well). Then there are 5–10 m thick layers of fluvial channel and flood plain deposits, which are composed of sand/gravel and silt, respectively (GA). Moving towards the coast, the CDRs are reduced in thickness from 20–25 m to about 10–15 m. Towards inland, CDRs are replaced (to the east of the section E–E’ in Fig. 1b) by a layer of silty and clayey lagoon/estuary deposits (see section A).
Eastward the subsurface is composed of coastal plain/swamp and lagoon/estuary deposits, constituted by silt and clay (A, B and C sections in Fig. 4). North of the study area is a similar stratigraphic situation with CDRs of limited thickness (10–15 m) along the coast and 20 m in the eastern portion (section B and C). Inland, the CDRs are covered
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
by more recent flood plain deposits (silt and clay) and are connected with the sandy deposit of the inner Pisan plain (1st confined aquifer in Baldacci et al. (1994)) (see section C). The cross sections (D, E, F in Fig. 4) confirm the stratigraphic framework described above. In several zones there are discontinuous gravelly horizons (GA) at different depths (higher and lower than 50 m). Across the study area the gravelly level is separated from the overlying sandy aquifer, but locally (e.g. San Piero a Grado) there are some gravel levels in hydraulic connection, due to the absence of the siltyclayey aquiclude/aquitard that usually separates them, or due to their thickness reduction. In the San Piero a Grado zone, the discontinuous gravelly levels are interposed between the main sandy aquifer (CDRs) and the deeper gravelly aquifer (section E). In this section, the well P4 is characterized by the absence of clays since the thickness between CDRs and GA is made up of silt-sandy lithologies. Hydraulic connections between the two main aquifers take place by vertical exchange phenomena through the semi-permeable thin horizons which separate such aquifers, according to Baldacci et al. (1999) also. Fig. 4 shows the CDRs geometry via a stratigraphic interpretation of the cross sections. In section E it is clear that CDRs develops in depth, at times locally connecting to deeper permeable layers (gravelly aquifers).
79
6. Hydrogeology and hydrogeochemistry features Towards the coast, the first sandy aquifer is represented by the CDRs and changes from confined to phreatic at the alignment of section E. The CDRs are comprised of medium-fine grained sand and have a medium level of permeability (about 10 4–10 5 m/s) as can be seen from various permeability tests performed in previous explorations in this aquifer. The small strips between the dune alignments have a low permeability, due to the presence of silt and clay, but do not interfere with the main hydrodynamic characteristics due to their minimal thickness. To the east of section E (Fig. 1b) the thickness of the low permeability layers (silty clays and sometimes peaty, clayey silts and sandy silts with interbedded thin sandy layers) prevents any continuous phreatic aquifers. The CDRs can be regarded as an area of direct recharge by the infiltration of rainwater. This recharge occurs through infiltration when precipitation exceeds evapotranspiration, generally between May and September. This aquifer is directly connected to the hydrographic network; during rainy periods it recharged, while it is fed by the hydrographic network in the dry season. In the study area the recharge of the sandy aquifer is mainly occurs in the coastal dune area. Flow direction proceeds from inside the plain to the confined portion on the coast, while at the CDRs the direction of the flow is influenced by the depressed areas and thus goes from topographically elevated areas to areas below the sea level.
Table 2 EC, and chemical analysis of surface waters and phreatic aquifer (piezometers) carry out during 2012 survey (A) and 2008 survey (B). The values of June 2012 concerned the average of two surveys (6 and 15 June 2012). Locations of the sampling points are shown in Fig. 6a. (A) June 2012 Code
Depth (metres)/channel
CM_A1 CM_A2 CM_A3 CM_A4 CM_A5 CM_A6 CM_A7 CM_A8 P20 P17 New_CNR Mandria Bargagna_S Bargagna RMS63 RMS29 S.Guido Bigattiera
Collettore Settentrionale Collettore Orientale Lama Larga Northern Lama Larga Northern Lama Larga Northern Lama Larga Northern Collettore Settentrionale Collettore Settentrionale 7 10 15 10 3 90 15 3 3 3
EC μS/cm Surface water Surface water Surface water Surface water Surface water Surface water Surface water Surface water Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater
dry 39,000 2650 5900 3400 7800 4880 1003 940 1440 4650 742 dry 3360 46,850 dry 804 617
T °C
Cl mg/L
SO4 mg/L
NO3 mg/L
20.0 20.3 20.0 21.0 20.9 20.5 20.5 nd 16.3 16.05 14.4
14,180 520 1715 2109 2438 1476 100 52 160 1213 22
1958 12 132 200 218 109 56 117 79 65 44
o50 o2 o5 o2 o5 o5 o1 4 78 o5 o1
nd 16.6
545 18,080
508 o 50
o5 o50
nd 15.5
37 20
16 21
6 o1
Cl mg/L
SO4 mg/L
NO3 mg/L
17,130 3910 164 199 6928 732 815 2163 6751 2861 17,275 16,390 10,053 17,629
1031 190 64 101 128 81 53 74 66 10 1622 54 56 337
(B) September 2008 Code
Depth (metres)
RMS01 RMS04 RMS14 RMS20 RMS21 RMS23 RMS42 RMS43 RMS44 RMS45 RMS50 RMS51 RMS52 RMS53
15 15,2 14,5 11,75 11,9 13,2 15 15 15 15 15 10 10 10
EC μS/cm Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater
47,800 10,300 880 1614 18,270 1735 7320 6480 16,400 12,760 46,200 41,500 34,900 45,800
T °C 19.1 19.5 18.8 17.3 18.0 17.2 17.9 17.0 16.4 16.5 16.4 16.0 14.5 16.4
o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5
80
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
The seasonal average water table depth in the Tombolo area shows a trend from north-east to south-west in average depth, which is everywhere between 1.0 m and 2.0 m below ground level. On the whole, the salinity of the fresh water aquifer is mainly determined by HCO3. However, considering the individual aquifers, the salinity of the sandy aquifer (CDRs) is greatly affected by seawater intrusion affecting the Cl content, whereas the gravelly aquifer (GA) is significantly influenced by SO4 and HCO3 (Grassi and Cortecci, 2005). The waters sampled from the wells drilled in the CDRs along the coastal area (max 20 m in depth) have a relatively high salinity, poor quality and basically belong to Na-Cl facies, especially for
those samples located close to the mouths of the Arno and Scolmatore and those located close to the main lowland channels. High values of EC (frequently more than 3000 μs/cm) are generally observed in all wells close to the coastline (Table 2). The water samples taken from piezometers in the coastal area between Marina di Pisa and Calambrone are alcaline-chloride waters, with a prevalence of sodium to potassium (Signorini, 2008). These waters have chloride levels between 200 and 17,000 mg/l and EC that spans from a minimum of 1000 to a maximum of 12,000 μS/cm measured at 2 m depth, while at the maximum depth (15 m) EC increases up to values between 3400 μS/cm and 45,000 μS/cm. Very high values are detected into
Fig. 5. EC distribution in CDRs by Giannecchini et al. (2010).
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
some inland piezometers RMS50, RMS51, RMS52 and RMS53. The data available regarding the gravelly aquifer from Butteri et al. (2010) show that the study area is characterized by a water circulation with a prevalently Ca/HCO3-SO4 composition, which evolves towards Na/Cl composition caused by seawater mixing. With the exception of P2, P3 and P4 (San Piero a Grado area in Fig. 1b), almost every water sample collected by these Authors in the deep wells of the inland zones is characterized by relatively low EC, often less than 2000 μs/cm. In the water-wells belonging to San Piero and Calambrone zones, they recorded higher EC values (7000 μs/cm). The Bargagna well (corresponding to P4 in Butteri et al. (2010)), sampled during the June survey, present geochemical feature typical of gravelly aquifer with high sulphate and chlorides and EC values of 3360 μS/cm. All other groundwater in Table 2 are representative of the phreatic water into the CDRs. Very high chloride have also been found in RMS63 piezometers and CM_A2 channel. This last is directly connected to the Navicelli channel. Along the coastal (piezometers in Table 2a) the CDRs aquifer evidently have shallow shoreline salt-wedge. Further inland the channels have chloride values higher that piezometers with the exception of the RMS63.
7. Salinity distribution In 2009 a survey with accurate EC measurements and chemical analyses regarding the piezometers network installed by the municipality of Pisa led to the reconstruction of the EC distribution
81
within shallow, sandy aquifer (CDRs) in the study area (Fig. 5; Giannecchini et al. (2010). The values are not particularly low if we consider that Italian law regarding drinking waters states (Legislative Decree No. 31/ 2001, 2001) that the value should not exceed 2500 μS/cm; this value is similar to the threshold value of a later decree (Legislative Decree No. 30/2009, 2009) of 2500 μS/cm as an upper threshold for good quality groundwater. Fig. 6b shows a salinity distribution map created on the basis of surface water and shallow groundwater EC data. Data derived from the ARPAT database include monitoring data related to environmental emergencies and to periodic monitoring conducted in line with regulations from the 2000/60 European Framework Directive. Data from the ARPAT database were implemented with the monitoring data from this work together with data from Silvestri and Gorreri (Signorini, 2008). The processed data, 248 elements in total, were used to reconstruct the map in Fig. 6b using the Kriging method. Fig. 6b highlights a high EC zone overlapping the Navicelli channel, which links the sea to the area of San Piero a Grado and the suburbs of Pisa. The increase in EC connected to the conveyance of salt water by the Collettore Orientale channel is also evident. Monitoring and sampling points carried out in the summer of 2012 are shown in Fig. 7a; the values regarding major anions are shown in Table 2. Fig. 7 shows the results of surface water (Fig. 7a) and shallow groundwater (Fig. 7b) monitoring performed during the threeyear monitoring of surface water (2010–2013) and the annual monitoring (2012–2013) of the piezometers placed in the Tombolo
Fig. 6. (a) Location of sampling stations and (b) distribution of EC as obtained from the Kriging elaboration of 248 measures.
82
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
Fig. 7. (a) EC measurements concerning surface water in the three years period (2010–2013) in the sampling stations inside the Tombolo forest. (b) EC measurements concerning shallow water (phreatic aquifer) during 2012 in sampling stations in the same area. c) Rainfall patterns from three weather stations. Locations of the sampling points are shown in Fig. 6a.
forest. The locations of the sampling points are shown in Fig. 6 a. The chloride values and EC observed along the Collettore Orientale channel (CM_A2 in Table 2a and W1, W4 and W7 in Fig. 7a) confirm the presence of high salinity water from the south.
The plots also show the rainfall patterns during the reference period (red vertical bars in Fig. 7a and b). The rainfall patterns refer to the average daily cumulative rainfall of the three weather stations included in the study area (Bocca d’Arno, Calambrone and
Fig. 8. EC measurements concerning surface water and shallow water (phreatic aquifer) trends compared during the overlap period of the monitoring (Oct.’12–Dec.‘13).
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
Airport stations, data from Servizio Idrologico Regionale (2014). Stations W1, W4, W7, W8 and W10 are located in the sub-basin of Tombolo, which is characterized by natural drainage (Table 1b and Fig. 2b), while stations W2, W3, W5, W6 and W8 are located in the mechanical drainage basin (South Lamone in Table 1a and Fig. 2b). The surface water, and particularly W4, W1 and W7 relating to the Collettore Orientale channel, receives a feed of high salinity water directly from the Navicelli channel during low precipitation periods. This feed is very fast and takes the values of EC to values close to those of sea water. Particularly noticeable (Fig. 7a) are the hikes observed in the 2010 and 2011 springs and 2013 summer. Throughout the period April 2011–November 2012, the EC values of the Collettore Orientale channel remained above 40,000 μS/cm. This due to low amount of autumnal and winter rains that did not prevent the influx of seawater from Navicelli channel. The trend related to groundwater (Fig. 7b) is referred only to Oct.’12–Dec.’13 period. It shows a strong increase of salinity in the spring of 2013 for the piezometer PZ11 (EC values 410,000 μS/cm). In the piezometer PZ06 salinity increases instead since July 2013. In Fig. 8 the two trends are compared during the overlap period of the monitoring (Oct.’12–Dec.‘13). Sudden increases in salinity show that salinisation regarding the shallow groundwater comes from the circulation of surface water and not from feeds by the shallow aquifer. The highest values are found in the piezometers closest to the Collettore Orientale channel (Pz11 and Pz6 points) coinciding with the dry season. The increase of EC in PZ6 follows the salinisation of the Collettore Orientale according to surface infiltration. The increase of EC in PZ11 precedes instead the salinisation of the Collettore Orientale. This fact is actually not justified from available data and could be connected to a local anthropic pollution. The feed from surface water to groundwater is very important around the Collettore Orientale and probably around the Navicelli channel. This can justify the PZ10 values which remain relatively high (around 5000 μS/cm) over that monitoring period. The EC values greater than 40,000 μS/cm detected into RMS50 and RMS52 could be attributed to influx of seawater infiltration from Navicelli channel too.
83
The data evaluated in this work lead to an alternative hypothesis: seawater intruding via surface water could be the cause of this phenomenon and, in particular, the Navicelli channel. Without permanent feeding from upstream, the Navicelli channel has a fresh water flow towards the sea only when the dewatering pumps, which are used for lowland drainage, are being operated. Normally dewatering pumps are activated in conjunction with main rainfall events. During the summer months, when rainfall is usually absent or very reduced, the Navicelli channel is affected by rising seawater. In addition the channel bed runs through CDRs for a long stretch allowing seawater infiltration into the phreatic aquifer in the presence of unfavourable hydraulic gradients. The CDRs can thus be regarded as an area of direct recharge by infiltration of meteoric waters. Too much infiltration water is drained from the hydraulic network during the main rainy periods. On the other hand, during dry periods, salt waters rising from the Navicelli channel tend to infiltrate towards the phreatic aquifer. The alternation of increasingly longer drought periods tends to gradually extend the portions of the phreatic aquifer affected by salinisation. Due to the gravelly aquifer overexploitation, phreatic aquifer salinisation can lead to the transfer of the contamination to deeper levels resulting in the gradual salinisation of groundwater. Coastal areas close to the mouths of the Arno and Scolmatore are particularly critical because they are characterized by the high salinity of the groundwaters. The inland plain areas, especially close to the main drainage channels (e.g. Navicelli), are characterized by poor quality groundwaters, with high chloride concentrations. The data available indicate that the seawater intrusion takes place directly through the phreatic aquifer (along the coastline) and the rising of salt and brackish waters along rivers and channels with subsequent infiltration into the phreatic aquifer. The watercourses characterized by poor or no feeding upstream are most affected by rising seawater during dry periods. To sum up, the waters with chemical characteristics similar to those of the seawater in inland areas far from the coastline cannot be explained by deep seawater intrusion, but instead should be attributed to the intrusion of seawater along the Navicelli channel, which conveys the seawater to the Pisa suburbs, located about 7 km from the coastline, after a course of about 16 km.
8. Discussion A comparative analysis of all the data available indicates that the seawater intrusion may be due both to the direct input of seawater in the aquifers, especially in areas close to the coastline, and to the rise in salt water in surface streams, with subsequent infiltration into the first sandy aquifer (CDRs). Giannecchini et al. (2010) observed that CDRs have a salinisation effect, developing towards the inner plain with increasing values with the strong gradient from the coastline to the more inland area of the Navicelli channel. Fig. 5 shows that the EC values along the coastline are lower than the innermost area of the coastal plain by one order of magnitude. This is in contrast with the effects of typical seawater intrusion, since the salt wedge should deepen as it moves away from the coast. Giannecchini suggested that this effect may be related to the rising of the salt wedge induced by the strong exploitation of wells. The exploited wells in the area between the coast and the Navicelli channel however only intercept the gravel confined aquifer. It is not clear how the overexploitation of this aquifer could promote the development of a salt wedge in the phreatic aquifer at such distances from the coast. In our study area the previous data therefore show evidence of salinisation of the surface aquifer content within CDRs, which seems to increase inland.
9. Conclusion This investigation provides insights into the inland flow of seawater via surface water through the hydrographic network. Especially during the dry season, this flow leads to the loss of the hydrodynamic surface equilibrium between the groundwater table, which is contained within the coastal dune ridges (CDRs) aquifer, and the seawater level in surface water. This causes widespread salinisation of the phreatic aquifer starting from the hydrographic network, which carries seawater towards the inland plain. The geometrically complex geology and the overexploitation in the deep confined gravelly aquifer lead to the hydrogeological unbalance that propagates this salinisation towards the deep aquifer, which is currently used intensively. Given the actions required to remove or mitigate the pollution impacts are different in presence of salt intrusion due to the rise in seawater in hydrographic network or the deep salt wedge from shoreline, in large and thin coastal plains such as the Pisa plain, it is crucial to be able to clearly distinguish between these two phenomena in order to implement the appropriate monitoring actions. We suggest accurate monitoring by competent authorities in order to better identify intrusion via surface water also through
84
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
the hydrogeological modelling of the water table aquifer contained in the CDRs. We believe that our results constitute a tool for managing a network of datasets that local authorities can use to improve land management and water resources, and also to prevent possible environmental damage caused by salty pollution via surface water.
References Amores, M.J., Verones, F., Raptis, C., Juraske, R., Pfister, S., Stoessel, F., Antón, A., Castells, F., Hellweg, S., 2013. Biodiversity impacts from salinity increase in a coastal wetland. Environ. Sci. Technol. 47 (12), 6384–6392. Amorosi, A., Bini, M., Giacomelli, S., Pappalardo, M., Ribecai, C., Rossi, V., Sammartino, I., Sarti, G., 2013. Middle to late Holocene environmental evolution of the Pisa coastal plain (Tuscany, Italy) and early human settlements. Quat. Int. 303 (2013), 93–106. Amorosi, A., Sarti, G., Rossi, V., Fontana, V., 2008. Anatomy and sequence stratigraphy of the late quaternary Arno valley fill (Tuscany, Italy). Advances in Application of Sequence Stratigraphy in Italy. 1; , pp. 55–66. Anderson Jr., W.P., 2002. Aquifer salinization from storm overwash. J. Coast. Res. 18, 413–420. Baldacci, F., Bellini, L., Raggi, G., 1994. Le risorse idriche sotterranee della Pianura Pisana. Atti Soc. Tosc. Sci. Nat. Mem. Ser. A 101, 241–322. Baldacci, F., 1999. Struttura e piezometria del “primo acquifero artesiano in ghiaie” della Pianura Pisana. Atti della Soc. Tosc. Sci. Nat. Mem. Ser. A 106, 90–101. Baldacci, F., Pierotti, L., Spandre, R., Tessitore, S., 1999. I1 sistema acquifero freatico della pianura Pisana. Rapporti fiume-falda e acqua dolce-acqua salata. In: Procedings of the Conference Conoscenza e salvaguardia delle aree di pianura, Ferrara, 8 November. Barrocu, G., 2003. Seawater intrusion in the coastal aquifers of Italy. In: Calaferra, J., (Ed.), State of seawater intrusion in coastal aquifers of the Mediterranean Coast, SWIM-SWICA, Alicante, Spain, pp. 207–223. Butteri, M., Doveri, R., Giannecchini, P., 2010. Gattai Hydrogeologic-hydrogeochemical multidisciplinary study of the confined gravelly aquifer in the coastal Pisan plain between the Arno River and Scolmatore Channel (Tuscany). Mem. Descr. Carta Geol. d'It., XC, 105–120. Carosi, R., Montomoli, C., Pertusati, P.C., Sarti, G., Frassi, L., 2010. Note illustrative della Carta Geologica d’Italia alla scala 1:50.000. ISPRA. Day, J.W., Britsch, L.D., Hawes, S.R., Shaffer, G.P., Reed, D.J., Cahoon, D., 2000. Pattern and process of land loss in the Mississippi Delta: spatial and temporal analysis of wetland habitat change. Estuaries 23 (4), 425–438. Della Rocca, B., Mazzanti, R., Pranzini, E., 1987. Studio geomorfologico della Pianura di Pisa. Geogr. Fis. Dinam. Quat. 10, 56–84. Doveri, M., Giannecchini, R., Butteri, M., 2010. Seawater intrusion in the VersiliesePisan coastal aquifer system (northwestern Tuscany): results from a hydrogeologic-hydrogeochemical study. Presented to SWIM21 – 21st Salt Water Intrusion Meeting, Azores, Portugal, June 21–26. Fancelli, R., 1984. Tentativo di ricostruzione dell’andamento della falda artesiana tra i-30 ed i-50 m di profondità nel sottosuolo di Pisa e dintorni. Rapporto CNR-Ist. Inc. Ric. Geotermiche, Pisa. Franceschini, S., Heusch, G., 2013. Final Mitigations Project to support the projects “Repair/Upgrade Cotongrosso Road” and “Construction of three new Pads” in the Ammunition Storage Area (ASA). Migliarino-San Rossore-Massaciuccoli Regional Park, Nulla Osta no. 304/2013. Galletti-Fancelli, M.L., 1978. Ricerche sulla subsidenza della Pianura pisana analisi polliniche di sedimenti quaternari della pianura costiera tra Pisa e Livorno. Boll. Soc. Geol. It. 97, 197–245.
Giannecchini, R., Menichini, M., Raco, B., Viti, F., 2010. Hydrogeological and geochemical features of the sandy unconfined aquifer between Marina di Pisa and Calambrone (NW Tuscany) affected by seawater intrusion. Abstract/Poster 85° Congresso Soc. Geol. It. In: Rendiconti Online Societa Geologica Italiana, vol. 1, pp. 179–180. Grassi, S., Cortecci, G., Squarci, P., 2007. Groundwater resource degradation in coastal plains: the example of the Cecina area (Tuscany–Central Italy). Appl. Geochem. 22, 2273–2289. Grassi, S., Cortecci, G., 2005. Hydrogeology and geochemistry of multilayered confined aquifer of the Pisa plain. Appl. Geochem. No. 20, 41–54. Lazzarotto, A., Mazzanti, R., Nencini, C., 1990. Geologia e morfologia dei comuni di Livorno e Collesalvetti. Quad. Mus. Stor. Nat. Livorno 11, 1–82. Legislative Decree No. 30/2009, 2009. Attuazione della direttiva 2006/118/CE, relativa alla protezione delle acque sotterranee dall'inquinamento e dal deterioramento. Gazzetta Ufficiale No. 79, 4 April. Legislative Decree No. 31/2001, 2001. Attuazione della direttiva 98/83/ce relativa alla qualità delle acque destinate al consumo umano. Gazzetta ufficiale No. 52, 3 March. Mara, D., 1998. Waste stabilization ponds and wastewater storage and treatment reservoirs: the lowcost production of microbiologically safe effluents for agricultural and aquacultural reuse. In: Asano, T. (Ed.), Wastewater Reclamation and Reuse. Technomic Publishing Co., pp. 141–158. Milly, P.C.D., Cazenave, A., Gennero, M.C., 2003. Contribution of climate-driven change in continental water storage to recent sea-level rise. Proc. Natl. Acad. Sci. USA 100 (23), 13158–13161. Pozzo, A., Todaro, U., 1923. La bonifica idraulica di Coltano (Provincia di Pisa). Giornale del Genio Civile, LXI, fasc. IX. Pranzini, G., 2002. Groundwater salinization in Versilia (Italy). In: Proceedings of 17th SWIM, Delft, Holland, pp. 412–421. Pranzini, E., 2001. Updrift river mouth migration on cuspate deltas: two examples from the coast of Tuscany (Italy). Geomorphology 38 (2001), 125–132. Provincial administration of Pisa, 2005. La Geologia della Provincia di Pisa – Cartografia, Geositi e Banche Dati. Provincia di Pisa. Resource document 〈http://sit. provincia.pisa.it/SisterIms/html/geologia/indice_geologia.htm〉. Rossi, V., Amorosi, A., Sarti, G., Potenza, M., 2011. Influence of inherited topography on the Holocene sedimentary evolution of coastal systems: an example from Arno coastal plain (Tuscany, Italy). Geomorphology 135, 117–128. Sarti, G., Testa, G., Zanchetta, G., 2008. A new stratigraphic insight of the Upper Pliocene-Lower Pleistocene succession of Lower Valdarno (Tuscany, Italy). Geoacta 7, 27–41. Servizio Idrologico Regionale, 2014. Resource document 〈http://www.sir.toscana.it/ index.php〉. Signorini, R., 2008. Idrochimica delle acque sotterranee e superficiali come strumento di comprensione dei fattori idraulici, idrochimici e geologici dell’ingressione marina. Applicazione alla fascia costiera della Pianura di Pisa tra il Fiume Arno e lo Scolmatore (Thesis). Earth Science Department, Università degli studi di Pisa. Silvestri, N., Gorreri, L., 2008. Il monitoraggio delle acque ad uso irriguo nel parco regionale Migliarino San Rossore Massaciuccoli. Pacini (Ed.), Pisa. Spandre, R., Ceragioli, M., Spinicci, A., 1999. Realizzazione di una carta di vulnerabilità degli acquiferi della Pianura pisana compresa tra la linea di costa, il Canale Scolmatore, il Fosso della Bufalina e la città di Pontedera. Provincia di Pisa, Settore Pianificazione del Territorio. Switzer, A.D., 2014. Coastal hazard: storms and tsunamis. In: Masselink, G., Gehrels, R. (Eds.), Coastal Environments & Global Change. Wiley Publ., pp. 104–127. Turner, R.E., 1997. Wetland loss in the Northern Gulf of Mexico: multiple working hypotheses. Estuaries 20 (1), 1–13. USEPA, 1992. Guidelines for water reuse. Agency for International Development, NSFC, Morgantown, West Virginia, USA. Williams, W.D., 1987. Salinisation of rivers and streams: an important environment hazard. Ambio 16, 180–185.
Contents lists available at ScienceDirect
Groundwater for Sustainable Development journal homepage: www.elsevier.com/locate/gsd
Research paper
Seawater intrusion via surface water vs. deep shoreline salt-wedge: A case history from the Pisa coastal plain (Italy) F. Franceschini a,b,n, R. Signorini a a b
Earth Sciences Department of Pisa University, Via S. Maria, 53, 56126, Pisa, Italy Environmental Protection Agency of Tuscany (ARPAT), Via Vittorio Veneto, 27, 56127, Pisa, Italy
art ic l e i nf o
a b s t r a c t
Article history: Received 24 July 2015 Received in revised form 29 April 2016 Accepted 14 May 2016 Available online 10 June 2016
This study analyses data regarding the quality of surface water in the Pisan plain between the mouth of the Arno river, Pisa and Livorno in order to identify the origin of the high chloride content derived from the sea water in areas far from the coast. We used existing environmental monitoring data from the study area integrated with data from the present study. The study involved measuring the piezometric levels and electrical conductivity (EC), together with a chemical analysis of the surface water and phreatic aquifer (water table aquifer) samples. Data acquisition through sampling along coastline piezometers, was carried out in 2008. In the summer of 2012 we supplemented data regarding areas further inland with surveys of piezometric levels and EC in the hydrographic network and shallow wells. The distribution of salinity was compared with a three-year monitoring survey carried out at the water sampling points in the Tombolo forest, an area about 5 km from the coastline. All the collected data show the presence of significant salinisation generated by the sea water, which is carried towards the inner plain through an artificial watercourse (Navicelli channel) and its hydrographic network. In the dry season this channel contains no fresh water from upstream and sea water tends to flow up to the town of Pisa, about 7 km from the coast. The sea water carried by the channel infiltrates into the shallow aquifers directly or through the minor hydrographic network, thus polluting the internal areas of the Pisan plain. Potential hydraulic connections with the intensively exploited deeper aquifers might lead to the salinisation of the same aquifers. This phenomenon could be confused with seawater intrusion from the coast (deep shoreline salt-wedge). & 2016 Elsevier B.V. All rights reserved.
Keywords: Salinity Navicelli channel Salt pollution Land management Phreatic aquifer
1. Introduction The northern coastal plains of the Mediterranean are subject to significant human pressure, which often leads to the depletion and deterioration of water resources. One of the most recurring effects is the variation in the natural equilibrium between fresh and sea waters and the consequent advance of the sea water through the coastal aquifers. Many Italian coastal areas are affected by seawater intrusion (Barrocu, 2003). In Tuscany, several critical situations have been highlighted (Pranzini, 2002; Grassi et al., 2007). The salty waters in the multilayered aquifer system of the Pisan coast have been known about for at least 20 years, and numerous studies have highlighted how the problem affects the environment (Signorini, 2008; Doveri et al., 2010; Butteri et al., 2010; Giannecchini et al., n Corresponding author at: Environmental Protection Agency of Tuscany (ARPAT), Via Vittorio Veneto, 27, 56127, Pisa, Italy E-mail address: [email protected] (F. Franceschini).
http://dx.doi.org/10.1016/j.gsd.2016.05.003 2352-801X/& 2016 Elsevier B.V. All rights reserved.
2010). Seawater intrusion increases especially during the summer months due to the intensive exploitation of civil, industrial and agricultural uses with the consequent deterioration of the qualitative state of the groundwater. In addition to degrading the quality of groundwater and restricting the agricultural use of the land, increased salinity reduces the diversity of ecosystems (Amores et al., 2013) and leads to the development of halophilic species (Williams, 1987). Sea water contains approximately 35,000 mg/L of dissolved solids, which include about 20,000 mg/L of chloride. These chloride values are normally associated with EC values exceeding 45,000 μS/cm. Fresh groundwater in most coastal areas generally contains less than 100 mg/L of chloride. However, concentrations in excess of 100 mg/L are not conclusive evidence of seawater intrusion because they could be due to airborne sea spray in precipitation, well pumping, local sources of chlorides, including septic systems or animal manure, or to relic seawater in the aquifer. Chloride concentrations higher than a few hundred mg/L can cause adverse effects on many crops (USEPA, 1992; Mara, 1998). In order to provide criteria to assess the chemical status of groundwater
74
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
bodies according to the 2006/118 European directive, Italy (Legislative Decree No. 30/2009, 2009) defined the threshold values with respect to chloride and EC as 250 mg/L and 2500 μS/cm respectively. Seawater intrusion is the introduction of saltwater into inland freshwater systems. It can move into the freshwater system through both surface water and groundwater connections to the sea. Seawater intrusion via groundwater connection is normally directed along the shoreline as a deep salt-wedge. Human creation of channels and ditches for drainage of freshwater systems encourages saltwater intrusion where channels connect with a saltwater source (Turner, 1997; Day et al., 2000). This is particularly evident when river flows are low and or during high tides or storm surges (Anderson, 2002; Switzer, 2014). Storm surges and strong wind events can push saltwater into freshwater channels encroaching on freshwater coastal systems. Moreover, eustatic sealevel rise is projected to be 1–2 mm/yr, which is the change in sea level rise with a change in the ocean volume due to climate change (Milly et al., 2003). Seawater intrusion can easily propagate in-land along rivers and channels then infiltrate downward to the groundwater zone. In recent decades, the excessive overexploitation of fresh water resources has exacerbated the situation leading to periods when river flow is reduced to a threshold level beyond which it can no longer prevent sea water from moving inland. The salt water can go back to the lower courses of the rivers for miles, changing the chemical characteristics of the waters of the coastal plains (including groundwater) rendering it too saline for drinking or irrigation water. Coastal plains are underlain by alluvial aquifers which can be highly heterogeneous sequence of sands and silts. Seawater
naturally intrudes into coastal aquifers as a wedge due to density differences. It is very important to distinguish between natural saline wedges in the subsurface and saline intrusion via surface water flow followed by infiltration to the subsurface. How these effects are monitored is, in fact, very different as are the actions required to remove or mitigate the impacts. We studied the surface and subsurface waters in an area of the Pisan plain in terms of geological and hydrogeochemical aspects, with reference to the relationships of feeding and drain between the water table aquifer and the hydrographic network. The study area includes the river Arno and the Scolmatore channel to the north and south respectively, and the coastline to the west. The eastern boundary includes Pisa and the Coltano hills (Fig. 1). The study area includes the Migliarino-San Rossore-Massaciuccoli Regional Park and has a high environmental value, due to its great variety of natural habitats, ranging dunes to sandy shores and from hygrophilous forests to marshlands. The results provide a tool for managing a network of datasets to help local authorities to improve land management and the monitoring of water resources, and also to prevent possible environmental damage caused by salt pollution.
2. Geological setting The coastal plain is located in the west of Tuscany, north of Livorno (Fig. 1a). It is bounded by the Pisan Mountains to the NE, the Livorno hills to the south, and the Ligurian Sea to the west. The Arno valley consists of a succession of estuarine clays, subdivided into three, vertically-stacked transgressive-regressive millennial-
Fig. 1. (a) Location map of the study area (b) Lithological map, derived from the Provincial Administration of Pisa (Provincial administration of Pisa, 2005) geological map. The red lines indicate the section traces of Fig. 5.
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
scale depositional cycles, which are replaced at landward locations by coastal plain deposits (Amorosi et al., 2008). After two major transgressive cycles (Upper Miocene, Lower Pliocene), which led to the deposition of conglomerates, sands, clays and evaporitic sediments, an important regressive phase ensued during the Middle Pliocene. This led to land emergence and widespread erosion. Because of glaciations, significant eustatic fluctuations occurred during the Pleistocene. Initially (Lower Pleistocene), sediments of predominantly marine and saline environments were deposited in the area and then, during the Middle Pleistocene, gravels, sands, silts and clays followed due to the flooding of the Arno-Serchio water system. Then (Upper Pleistocene–Holocene), mainly fluvial and marshy sediments were deposited, due to reduced fluvial activity linked to both the end of the lowering of the eustatic sea-level and the drying of the climate. This climatic change also led to the deposition of aeolian sands, found both in depth and at the surface, where aeolian sands are partly linked to modern dunes during Holocene. The presence of dunes and the rise in the sea level hindered the outflow of waterways to the sea, with the consequent formation of marshes and lagoons, and the deposition of fine sediments along with the development of peat and lignite (Fancelli, 1984; Grassi and Cortecci, 2005). On the basis of pollen data from 19 continuously-cored boreholes drilled south of the Arno River, near Livorno, GallettiFancelli (Galletti-Fancelli, 1978) assigned the uppermost 30 m to the sub-Atlantic-Boreal phases (ca. last 10,000 cal yr BP). During these phases there were alternations in subsidence and a slight rise in sea level. Following a rapid rise in the sea level during the middle Holocene, the Arno paleovalley was flooded and the interfluves submerged by the sea. The subsequent evolution of the Pisa area saw the development of a wide, thick, clay-dominated, lagoonal succession, controlled by the inherited paleotopography (Rossi et al., 2011). At the time of maximum marine ingression (7.8 cal ka BP, (Amorosi et al., 2013)), the shoreline was located more than 7 km inland from its present position. From Pisa to the coastline, the plain (Fig. 1b) is currently composed of alluvial muddy and sandy deposits that develop seaward into a wide sandy strandplain system composed of several juxtaposed coastal dune ridges (CDRs) and narrower interposed depressions. The alignment of these ridges records the position of the shoreline over the last 2500–3000 years (Pranzini, 2001). According to Sarti et al. (2008) the genesis of CDRs is closely related to a classical model showing how a strandplain advances. The process begins with the formation of an embryonic dune which occurs when the whole foreshore-berm-backshore system migrates inland during a storm. At the end of the storm, a storm berm forms. This constitutes an obstacle in the backshore where the sand begins to accumulate forming an embryonic dune. There are also flooding deposits from the Arno river consisting mainly of fine sands in the proximal area and silts and clayey silts in the distal area, inter-dune and retro-dune fluvial-palustrine deposits consisting of moderately thick silts with peaty levels. This system of dunes contains a phreatic aquifer of regional importance in terms of the subsistence of the current ecosystems. It was previously used for irrigation with shallow wells, but it is no longer used due to the presence of high salt contents. Inland from the plain, where the sandy aquifer is confined by the floodplain deposits, the feed comes from the piedmont zone where it is in direct contact with the fan gravelly materials, and through the alluvial deposits of the Serchio River (Spandre et al., 1999). Below the sandy phreatic aquifer, the alternation of lithological layers with variable permeability has led to the creation of confined aquifers which are only partially independent of each other. These layers can be affected by direct hydraulic exchanges or by drainage induced by the various hydraulic gradients that characterize them. In
75
fact in this variable permeability, the lack of continuity and adequate thickness, do not guarantee a complete hydraulic separation between higher permeability levels, thus triggering interconnections and mutual exchanges (Butteri et al., 2010). The hydrogeological units have been classified as aquifers and aquitards after parameterization through estimates of permeability (Baldacci et al., 1994). In the inner portion of the Pisa plain, Baldacci et al. (1994) reported a complex multi-level confined aquifer that contains two major confined aquifers defined, from the top to the bottom as, “1st sandy confined aquifer” and “1st gravelly confined aquifer”. Moving towards the coast, the 1st sandy confined aquifer becomes the phreatic aquifer contained in the CDRs. The gravelly aquifer (GA) is represented by clastic fluvial deposits with high permeability, which originated during the migration of the Serchio and Arno riverbeds (Della Rocca et al., 1987). The roof of this aquifer (Baldacci et al., 1999; Della Rocca et al., 1987) is about 10 m thick with good permeability, and deepens northwards and westwards. Its minimum depth (30 m) are towards the southern border of the plain, whereas the maximum depth (180 m) are towards the north-western part of the plain. The gravelly aquifer is found at a depth of about 140 m in the southwestern zone of the city of Pisa, and seems to disappear northwards. In agreement with Baldacci et al. (1994), the GA is mainly recharged by the direct infiltration on the Livorno hills through the pebbly bodies that lie in the piedmont areas and in the intermountain valley of the Pisan hills. Isotopic content values of the water circulating in the confined gravelly aquifer indicate a recharge altitude higher than the plain (Grassi and Cortecci, 2005). The GA is a very important water resource for civil, industrial and irrigation uses. The overexploitation of the GA leads to water levels below sea-level in most of the area. There are two local and more accentuated piezometric depressions in San Piero a Grado and Calambrone, where the groundwater is heavily exploited. In particular, up to 1.5–2 km from the southern coastline (Calambrone area) and in a narrow area close to the coastline to the north (Arno mouth), the seawater intrusion develops directly in the confined gravelly horizon in contact with the sea floor. In the S. Piero a Grado area the mixing with seawater should be controlled by the seawater intrusion that affects the Arno River-phreatic sandy aquifer system and by the hydraulic connection between the sandy and gravelly aquifer (Butteri et al., 2010). The deposits found at Vicarello (Vicarello sands in Fig. 1b) are morphologically in relief on the plain (about 9 m). They emerge at the foot of the Pisan and Livorno hills (Fig. 1a). These deposits were previously related to an aeolian environment (Della Rocca et al., 1987; Lazzarotto et al., 1990), and were then reinterpreted as a fluvial environment (Carosi et al., 2010). As can be inferred from the stratigraphic sequence shown in this work, these deposits are likely to be in stratigraphic contact with the CDRs.
3. On-site surveys and analytical methods In order to identify the hydrogeochemical conditions of the CDRs we have evaluated the existing data and carried out onsite surveys. Since our aim was to highlight any criticality in the balance between the feeding and recharge of the aquifer contained in the CDRs, all data refer to the summer season. The geomorphological features and reconstruction of the hydrographic network were obtained using a high-resolution map made using 3D laser scanning (LIDAR) (Fig. 2a). For the Pisan plain a dense digital elevation model (DEM) with an accuracy of less than 10 cm is available at Pisa municipally authority. Landforms and hydraulic network identification was based on photointerpretation, morphometric analyses on the LIDAR digital elevation model (DEM) and a field survey.
76
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
Fig. 2. (a) LIDAR map of the study area. Dark shades indicate lowland (b) Hydraulic map derived from survey LIDAR. Natural (no colour) and mechanical (pink colour) drainage basins and some main drainage channels are reported in the map. The dewatering plants are numbered as table 1a.
Remote sensing analyses were carried out in order to detect traces of past drainage networks and wetlands (marshes, swamps, and ponds). The DEM available for this study came from a LIDAR survey carried out during 2008 and commissioned by the Ministry of the Environment and Protection of Land and Sea. Fig. 2a shows a raster map where each cell in the matrix has a spatial resolution of 1 metre and contains the elevation information. The plain hosts a large number of water-wells with depths varying from a few metres to nearly 250 m, which enabled the stratigraphic succession of the study area to be reconstructed. The stratigraphic reconstruction was carried out with stratigraphic well logs derived from the database of the Provincial administration of Pisa (2005) and various infrastructural engineering projects. Regarding the hydrogeochemical features, we used data by Signorini (2008) and Giannecchini et al. (2010) for the CDRs, and Grassi et al. (2007), Butteri et al. (2010) for the gravelly aquifer. The hydrogeochemical information on CDRs was integrated with two field surveys carried out in June and July 2012. With regard to the first survey, the piezometric level, temperature and EC were measured and the chloride and sulphate concentrations were determined from a selection of surface water samples from a hydraulic network and groundwater from existing piezometers and wells. In the subsequent surveys we measured the EC of the Lamone and Navicelli channels. Further EC and chloride concentration measurements were drawn from the ARPAT database (http://sira.arpat.toscana.it/sira/ acqua.php) concerning periodic monitoring performed in the
previous 10 years in the study area by selecting, to avoid excess data, those related to the summer season. To reconstruct the EC distribution, data by Silvestri and Gorreri (2008) were also used. The piezometric levels and variations in EC in the forest of Tombolo come from the monitoring of surface water and shallow groundwater as part of a about 3-year (March 2010–November 2013) study commissioned by the U.S. Army base at Camp Darby, which occupies a substantial portion of the area (Franceschini and Heusch, 2013). The monitoring concerned surveys of surface water (11 stations) and shallow groundwater (12 stations). The shallow groundwater monitoring stations consist of 14 mini-piezometers placed at a maximum depth of 3.5 m from ground level. In addition, measurements of EC, pH and temperature were taken from October 2012 until November 2013. These mini-piezometers intercept only the most superficial portion of the phreatic aquifer in order to separate them from the other groundwater samples considered here which represent the greater thickness of this aquifer. Temperature and EC (μS/cm) were measured in situ using a multiparametric probe (WTW Multi340i). The analyses to determine other chemical parameters reported in Table 2 were performed at the laboratories of Institute of Geoscience and Earth Resources (IGG) of the Italian National Research Council (CNR) and ARPAT laboratory in Pisa with regard to the data in Table 2b. Major anions were measured, such as chlorides, nitrates and sulphates, through an ion chromatography system (Dionex DX-120) using undiluted water samples.
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
4. The hydrographic network The coastal area of the Pisan plain is characterized by a Mediterranean climate with hot, dry summers and cool, wet winters. Annual rainfall ranges from 750 mm to about 850 mm. The annual evapotranspiration is 587 mm. The study area is characterized by an area of higher elevation (forest of Tombolo and coastal area) with altitudes that range between 2.5 and 0.5 m a.s.l., and depressed areas with altitudes of about 0.5 m a.s.l. and heights up to 1.5 m a.s.l. (Fig. 2b). Many depressed areas, especially around San Piero a Grado and along the Lamone channel, have been drained to create agricultural land. In winter the run-off rainwater flows towards ditches and channels in depressed areas (low waters) and are lifted by dewatering plants (mechanical drainage); in higher elevation areas (high waters) it is taken directly to the sea through the main watercourses (natural drainage). The hydraulic reclamation for mechanical drainage removes the excess and stagnant waters, but also prevents the rise of phreatic water, which is harmful to agricultural crops. The basins within the study area, identified through the DEM survey, are shown in Fig. 2b, with the main flow water channels and the mechanical drainage draining pump stations. In the study area there are five main mechanical drainage basins and five natural drainage basins (Table 1a and b). Of the most important land reclamation, carried out between the end of the 19th century and the first half of the 20th century, the Navicelli channel, is worth mentioning. This channel represents the main watercourse of low water from mechanical drainage. It is a construction which enables the natural drainage of part of the plain (basin in Table 1a and b) that houses important infrastructures (e.g. an international airport) and favours good transport by ship between the industrial area of Pisa and the harbour in Livorno. The Navicelli channel is about 16 km long and runs straight from southern Pisa to the northern border of the Livorno harbour where it flows into the Scolmatore channel immediately before its confluence into the sea. It was built with a steady curved section, which is almost trapezoidal with 13 m width and 3 m a.s.l. steady altitude of the bed channel, with a 1:2 slope (Pozzo and Todaro, 1923). Its construction was completed in 1924 after about three years of intensive work. The initial project was supposed to connect it with the river Arno, but it was not completed. The upper stretch of the Navicelli channel near Pisa, named “Incile”, is not directly connected with the main
77
channel. This stretch also has a wastewater input. Waters resulting from reclamation areas from the whole coastal plain south of the Arno flow into the Navicelli channel (Table 1a and b). The channel overlaps the previous artificial hydrographic network since the Middle Ages for reclaiming the depressed areas of south Pisa (Fig. 2b); the old Navicelli channel, built during Medici age, was the main flow channel of these areas. Other tributaries from natural drainage include Sanguinetto and Mezzanina channels from the Sanguinetto basin, Scolo di Pisa channel from the San Giusto basin and San Ermete channel from eastern Pisa. The tables highlight that the Navicelli channel is fundamental in that it provides the hydraulic balance of the whole study area in the rainy periods. During low precipitation periods, instead, the channel is in hydrochemical and hydrostatic continuity with the sea, conveying “pure” seawater into the inland plain. Fig. 3 shows the results of the survey (July 2012) where this situation is evident. The survey was carried out during a significant interval in the tidal oscillations in order to highlight the quality of the water during a period of minimum flow. Two measurements at different depths were carried out for each point. In Fig. 3 blue and red represent measurement at a depth of 20 cm and 50 cm, respectively. In the final stretch of the channel the contributions of channels fed by the highland with fresh water are clear. In the Fosso Chiara channel the rising salt water associated with the presence of fresh water inflows is particularly evident because this channel is connected to the continental drainage network. Significant contributions of fresh water are absent so the water in the entire length of the Navicelli channel is characterized by high salinity, very close to that of the sea.
5. Coastal dune ridges (CDRs) features The CDRs (the portion of the Pisan plain to the west of the section E–E’ in Fig. 1b) cover the entire area between the Arno river and Scolmatore to the north and south respectively and the coastline to the west. To the east the CDRs limit coincides with depressed areas characterized by fine flood and marsh plain deposits (Coltano and Stagno basins in Fig. 2b). South of the study area (sections A and B in Fig. 4), the sub-surface is composed of a 20–25 m thick layer of sandy deposits interpretable as upper and lower shoreface (CDRs). Below is a thick layer (30–40 m) of lagoon/ estuary deposits constituted by silt and clay (for example, TPW
Table 1 (a) Mechanical drainage basins in study area. (b) Natural drainage basins in study area. See Fig. 2b about location of watercourses and basins. (a) Name
Main drainage channel (high waters)
Dewatering pumping system (numbers in Fig. 2b)
Final Watercourse (low waters)
Arnino (North Lamone) South Lamone Stagno-Coltano South Pisa Vettola
“Collettore settentrionale” “Lama Larga” (Northern watercourse) “Lamone” “Lama Larga” (Southern watercourse) “Collettore Padule Maggiore” “Collettore dello Stagno” “Venticinque”, “Sofina”, “S.Ermete” Old Navicelli
Marina di Pisa (1)
Arno river
Calambrone (2) Ragnaione (3) Airport (4) La Vettola (5)
Scolmatore channel Navicelli channel Navicelli channel Navicelli channel
(b) Name
Main drainage channel
Final Watercourse
San Giusto Sanguinetto Tombolo Coltano (dune) Camp Darby
“Scolo di Pisa” “Sanguinetto”, “Mezzanina” “Collettore orientale” “Mandracchio” Depot area
Navicelli channel Navicelli channel Navicelli channel Navicelli channel Fossa Chiara channel
78
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
Fig. 3. EC trend along the Navicelli Channel. Survey of July 2012. Triangles represent the Navicelli water, squares represent the water of his tributaries. The tidal oscillation from Leghorn harbour tide gauge are showed in the box.
Fig. 4. CDRs geometry in study area as reconstructed after stratigraphic interpretation. Section trace is reported in Fig. 1b. In cross-sections CDR is drawn in yellow, gravelly aquifer is drawn in blue.
well). Then there are 5–10 m thick layers of fluvial channel and flood plain deposits, which are composed of sand/gravel and silt, respectively (GA). Moving towards the coast, the CDRs are reduced in thickness from 20–25 m to about 10–15 m. Towards inland, CDRs are replaced (to the east of the section E–E’ in Fig. 1b) by a layer of silty and clayey lagoon/estuary deposits (see section A).
Eastward the subsurface is composed of coastal plain/swamp and lagoon/estuary deposits, constituted by silt and clay (A, B and C sections in Fig. 4). North of the study area is a similar stratigraphic situation with CDRs of limited thickness (10–15 m) along the coast and 20 m in the eastern portion (section B and C). Inland, the CDRs are covered
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
by more recent flood plain deposits (silt and clay) and are connected with the sandy deposit of the inner Pisan plain (1st confined aquifer in Baldacci et al. (1994)) (see section C). The cross sections (D, E, F in Fig. 4) confirm the stratigraphic framework described above. In several zones there are discontinuous gravelly horizons (GA) at different depths (higher and lower than 50 m). Across the study area the gravelly level is separated from the overlying sandy aquifer, but locally (e.g. San Piero a Grado) there are some gravel levels in hydraulic connection, due to the absence of the siltyclayey aquiclude/aquitard that usually separates them, or due to their thickness reduction. In the San Piero a Grado zone, the discontinuous gravelly levels are interposed between the main sandy aquifer (CDRs) and the deeper gravelly aquifer (section E). In this section, the well P4 is characterized by the absence of clays since the thickness between CDRs and GA is made up of silt-sandy lithologies. Hydraulic connections between the two main aquifers take place by vertical exchange phenomena through the semi-permeable thin horizons which separate such aquifers, according to Baldacci et al. (1999) also. Fig. 4 shows the CDRs geometry via a stratigraphic interpretation of the cross sections. In section E it is clear that CDRs develops in depth, at times locally connecting to deeper permeable layers (gravelly aquifers).
79
6. Hydrogeology and hydrogeochemistry features Towards the coast, the first sandy aquifer is represented by the CDRs and changes from confined to phreatic at the alignment of section E. The CDRs are comprised of medium-fine grained sand and have a medium level of permeability (about 10 4–10 5 m/s) as can be seen from various permeability tests performed in previous explorations in this aquifer. The small strips between the dune alignments have a low permeability, due to the presence of silt and clay, but do not interfere with the main hydrodynamic characteristics due to their minimal thickness. To the east of section E (Fig. 1b) the thickness of the low permeability layers (silty clays and sometimes peaty, clayey silts and sandy silts with interbedded thin sandy layers) prevents any continuous phreatic aquifers. The CDRs can be regarded as an area of direct recharge by the infiltration of rainwater. This recharge occurs through infiltration when precipitation exceeds evapotranspiration, generally between May and September. This aquifer is directly connected to the hydrographic network; during rainy periods it recharged, while it is fed by the hydrographic network in the dry season. In the study area the recharge of the sandy aquifer is mainly occurs in the coastal dune area. Flow direction proceeds from inside the plain to the confined portion on the coast, while at the CDRs the direction of the flow is influenced by the depressed areas and thus goes from topographically elevated areas to areas below the sea level.
Table 2 EC, and chemical analysis of surface waters and phreatic aquifer (piezometers) carry out during 2012 survey (A) and 2008 survey (B). The values of June 2012 concerned the average of two surveys (6 and 15 June 2012). Locations of the sampling points are shown in Fig. 6a. (A) June 2012 Code
Depth (metres)/channel
CM_A1 CM_A2 CM_A3 CM_A4 CM_A5 CM_A6 CM_A7 CM_A8 P20 P17 New_CNR Mandria Bargagna_S Bargagna RMS63 RMS29 S.Guido Bigattiera
Collettore Settentrionale Collettore Orientale Lama Larga Northern Lama Larga Northern Lama Larga Northern Lama Larga Northern Collettore Settentrionale Collettore Settentrionale 7 10 15 10 3 90 15 3 3 3
EC μS/cm Surface water Surface water Surface water Surface water Surface water Surface water Surface water Surface water Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater
dry 39,000 2650 5900 3400 7800 4880 1003 940 1440 4650 742 dry 3360 46,850 dry 804 617
T °C
Cl mg/L
SO4 mg/L
NO3 mg/L
20.0 20.3 20.0 21.0 20.9 20.5 20.5 nd 16.3 16.05 14.4
14,180 520 1715 2109 2438 1476 100 52 160 1213 22
1958 12 132 200 218 109 56 117 79 65 44
o50 o2 o5 o2 o5 o5 o1 4 78 o5 o1
nd 16.6
545 18,080
508 o 50
o5 o50
nd 15.5
37 20
16 21
6 o1
Cl mg/L
SO4 mg/L
NO3 mg/L
17,130 3910 164 199 6928 732 815 2163 6751 2861 17,275 16,390 10,053 17,629
1031 190 64 101 128 81 53 74 66 10 1622 54 56 337
(B) September 2008 Code
Depth (metres)
RMS01 RMS04 RMS14 RMS20 RMS21 RMS23 RMS42 RMS43 RMS44 RMS45 RMS50 RMS51 RMS52 RMS53
15 15,2 14,5 11,75 11,9 13,2 15 15 15 15 15 10 10 10
EC μS/cm Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater Groundwater
47,800 10,300 880 1614 18,270 1735 7320 6480 16,400 12,760 46,200 41,500 34,900 45,800
T °C 19.1 19.5 18.8 17.3 18.0 17.2 17.9 17.0 16.4 16.5 16.4 16.0 14.5 16.4
o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5 o0,5
80
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
The seasonal average water table depth in the Tombolo area shows a trend from north-east to south-west in average depth, which is everywhere between 1.0 m and 2.0 m below ground level. On the whole, the salinity of the fresh water aquifer is mainly determined by HCO3. However, considering the individual aquifers, the salinity of the sandy aquifer (CDRs) is greatly affected by seawater intrusion affecting the Cl content, whereas the gravelly aquifer (GA) is significantly influenced by SO4 and HCO3 (Grassi and Cortecci, 2005). The waters sampled from the wells drilled in the CDRs along the coastal area (max 20 m in depth) have a relatively high salinity, poor quality and basically belong to Na-Cl facies, especially for
those samples located close to the mouths of the Arno and Scolmatore and those located close to the main lowland channels. High values of EC (frequently more than 3000 μs/cm) are generally observed in all wells close to the coastline (Table 2). The water samples taken from piezometers in the coastal area between Marina di Pisa and Calambrone are alcaline-chloride waters, with a prevalence of sodium to potassium (Signorini, 2008). These waters have chloride levels between 200 and 17,000 mg/l and EC that spans from a minimum of 1000 to a maximum of 12,000 μS/cm measured at 2 m depth, while at the maximum depth (15 m) EC increases up to values between 3400 μS/cm and 45,000 μS/cm. Very high values are detected into
Fig. 5. EC distribution in CDRs by Giannecchini et al. (2010).
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
some inland piezometers RMS50, RMS51, RMS52 and RMS53. The data available regarding the gravelly aquifer from Butteri et al. (2010) show that the study area is characterized by a water circulation with a prevalently Ca/HCO3-SO4 composition, which evolves towards Na/Cl composition caused by seawater mixing. With the exception of P2, P3 and P4 (San Piero a Grado area in Fig. 1b), almost every water sample collected by these Authors in the deep wells of the inland zones is characterized by relatively low EC, often less than 2000 μs/cm. In the water-wells belonging to San Piero and Calambrone zones, they recorded higher EC values (7000 μs/cm). The Bargagna well (corresponding to P4 in Butteri et al. (2010)), sampled during the June survey, present geochemical feature typical of gravelly aquifer with high sulphate and chlorides and EC values of 3360 μS/cm. All other groundwater in Table 2 are representative of the phreatic water into the CDRs. Very high chloride have also been found in RMS63 piezometers and CM_A2 channel. This last is directly connected to the Navicelli channel. Along the coastal (piezometers in Table 2a) the CDRs aquifer evidently have shallow shoreline salt-wedge. Further inland the channels have chloride values higher that piezometers with the exception of the RMS63.
7. Salinity distribution In 2009 a survey with accurate EC measurements and chemical analyses regarding the piezometers network installed by the municipality of Pisa led to the reconstruction of the EC distribution
81
within shallow, sandy aquifer (CDRs) in the study area (Fig. 5; Giannecchini et al. (2010). The values are not particularly low if we consider that Italian law regarding drinking waters states (Legislative Decree No. 31/ 2001, 2001) that the value should not exceed 2500 μS/cm; this value is similar to the threshold value of a later decree (Legislative Decree No. 30/2009, 2009) of 2500 μS/cm as an upper threshold for good quality groundwater. Fig. 6b shows a salinity distribution map created on the basis of surface water and shallow groundwater EC data. Data derived from the ARPAT database include monitoring data related to environmental emergencies and to periodic monitoring conducted in line with regulations from the 2000/60 European Framework Directive. Data from the ARPAT database were implemented with the monitoring data from this work together with data from Silvestri and Gorreri (Signorini, 2008). The processed data, 248 elements in total, were used to reconstruct the map in Fig. 6b using the Kriging method. Fig. 6b highlights a high EC zone overlapping the Navicelli channel, which links the sea to the area of San Piero a Grado and the suburbs of Pisa. The increase in EC connected to the conveyance of salt water by the Collettore Orientale channel is also evident. Monitoring and sampling points carried out in the summer of 2012 are shown in Fig. 7a; the values regarding major anions are shown in Table 2. Fig. 7 shows the results of surface water (Fig. 7a) and shallow groundwater (Fig. 7b) monitoring performed during the threeyear monitoring of surface water (2010–2013) and the annual monitoring (2012–2013) of the piezometers placed in the Tombolo
Fig. 6. (a) Location of sampling stations and (b) distribution of EC as obtained from the Kriging elaboration of 248 measures.
82
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
Fig. 7. (a) EC measurements concerning surface water in the three years period (2010–2013) in the sampling stations inside the Tombolo forest. (b) EC measurements concerning shallow water (phreatic aquifer) during 2012 in sampling stations in the same area. c) Rainfall patterns from three weather stations. Locations of the sampling points are shown in Fig. 6a.
forest. The locations of the sampling points are shown in Fig. 6 a. The chloride values and EC observed along the Collettore Orientale channel (CM_A2 in Table 2a and W1, W4 and W7 in Fig. 7a) confirm the presence of high salinity water from the south.
The plots also show the rainfall patterns during the reference period (red vertical bars in Fig. 7a and b). The rainfall patterns refer to the average daily cumulative rainfall of the three weather stations included in the study area (Bocca d’Arno, Calambrone and
Fig. 8. EC measurements concerning surface water and shallow water (phreatic aquifer) trends compared during the overlap period of the monitoring (Oct.’12–Dec.‘13).
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
Airport stations, data from Servizio Idrologico Regionale (2014). Stations W1, W4, W7, W8 and W10 are located in the sub-basin of Tombolo, which is characterized by natural drainage (Table 1b and Fig. 2b), while stations W2, W3, W5, W6 and W8 are located in the mechanical drainage basin (South Lamone in Table 1a and Fig. 2b). The surface water, and particularly W4, W1 and W7 relating to the Collettore Orientale channel, receives a feed of high salinity water directly from the Navicelli channel during low precipitation periods. This feed is very fast and takes the values of EC to values close to those of sea water. Particularly noticeable (Fig. 7a) are the hikes observed in the 2010 and 2011 springs and 2013 summer. Throughout the period April 2011–November 2012, the EC values of the Collettore Orientale channel remained above 40,000 μS/cm. This due to low amount of autumnal and winter rains that did not prevent the influx of seawater from Navicelli channel. The trend related to groundwater (Fig. 7b) is referred only to Oct.’12–Dec.’13 period. It shows a strong increase of salinity in the spring of 2013 for the piezometer PZ11 (EC values 410,000 μS/cm). In the piezometer PZ06 salinity increases instead since July 2013. In Fig. 8 the two trends are compared during the overlap period of the monitoring (Oct.’12–Dec.‘13). Sudden increases in salinity show that salinisation regarding the shallow groundwater comes from the circulation of surface water and not from feeds by the shallow aquifer. The highest values are found in the piezometers closest to the Collettore Orientale channel (Pz11 and Pz6 points) coinciding with the dry season. The increase of EC in PZ6 follows the salinisation of the Collettore Orientale according to surface infiltration. The increase of EC in PZ11 precedes instead the salinisation of the Collettore Orientale. This fact is actually not justified from available data and could be connected to a local anthropic pollution. The feed from surface water to groundwater is very important around the Collettore Orientale and probably around the Navicelli channel. This can justify the PZ10 values which remain relatively high (around 5000 μS/cm) over that monitoring period. The EC values greater than 40,000 μS/cm detected into RMS50 and RMS52 could be attributed to influx of seawater infiltration from Navicelli channel too.
83
The data evaluated in this work lead to an alternative hypothesis: seawater intruding via surface water could be the cause of this phenomenon and, in particular, the Navicelli channel. Without permanent feeding from upstream, the Navicelli channel has a fresh water flow towards the sea only when the dewatering pumps, which are used for lowland drainage, are being operated. Normally dewatering pumps are activated in conjunction with main rainfall events. During the summer months, when rainfall is usually absent or very reduced, the Navicelli channel is affected by rising seawater. In addition the channel bed runs through CDRs for a long stretch allowing seawater infiltration into the phreatic aquifer in the presence of unfavourable hydraulic gradients. The CDRs can thus be regarded as an area of direct recharge by infiltration of meteoric waters. Too much infiltration water is drained from the hydraulic network during the main rainy periods. On the other hand, during dry periods, salt waters rising from the Navicelli channel tend to infiltrate towards the phreatic aquifer. The alternation of increasingly longer drought periods tends to gradually extend the portions of the phreatic aquifer affected by salinisation. Due to the gravelly aquifer overexploitation, phreatic aquifer salinisation can lead to the transfer of the contamination to deeper levels resulting in the gradual salinisation of groundwater. Coastal areas close to the mouths of the Arno and Scolmatore are particularly critical because they are characterized by the high salinity of the groundwaters. The inland plain areas, especially close to the main drainage channels (e.g. Navicelli), are characterized by poor quality groundwaters, with high chloride concentrations. The data available indicate that the seawater intrusion takes place directly through the phreatic aquifer (along the coastline) and the rising of salt and brackish waters along rivers and channels with subsequent infiltration into the phreatic aquifer. The watercourses characterized by poor or no feeding upstream are most affected by rising seawater during dry periods. To sum up, the waters with chemical characteristics similar to those of the seawater in inland areas far from the coastline cannot be explained by deep seawater intrusion, but instead should be attributed to the intrusion of seawater along the Navicelli channel, which conveys the seawater to the Pisa suburbs, located about 7 km from the coastline, after a course of about 16 km.
8. Discussion A comparative analysis of all the data available indicates that the seawater intrusion may be due both to the direct input of seawater in the aquifers, especially in areas close to the coastline, and to the rise in salt water in surface streams, with subsequent infiltration into the first sandy aquifer (CDRs). Giannecchini et al. (2010) observed that CDRs have a salinisation effect, developing towards the inner plain with increasing values with the strong gradient from the coastline to the more inland area of the Navicelli channel. Fig. 5 shows that the EC values along the coastline are lower than the innermost area of the coastal plain by one order of magnitude. This is in contrast with the effects of typical seawater intrusion, since the salt wedge should deepen as it moves away from the coast. Giannecchini suggested that this effect may be related to the rising of the salt wedge induced by the strong exploitation of wells. The exploited wells in the area between the coast and the Navicelli channel however only intercept the gravel confined aquifer. It is not clear how the overexploitation of this aquifer could promote the development of a salt wedge in the phreatic aquifer at such distances from the coast. In our study area the previous data therefore show evidence of salinisation of the surface aquifer content within CDRs, which seems to increase inland.
9. Conclusion This investigation provides insights into the inland flow of seawater via surface water through the hydrographic network. Especially during the dry season, this flow leads to the loss of the hydrodynamic surface equilibrium between the groundwater table, which is contained within the coastal dune ridges (CDRs) aquifer, and the seawater level in surface water. This causes widespread salinisation of the phreatic aquifer starting from the hydrographic network, which carries seawater towards the inland plain. The geometrically complex geology and the overexploitation in the deep confined gravelly aquifer lead to the hydrogeological unbalance that propagates this salinisation towards the deep aquifer, which is currently used intensively. Given the actions required to remove or mitigate the pollution impacts are different in presence of salt intrusion due to the rise in seawater in hydrographic network or the deep salt wedge from shoreline, in large and thin coastal plains such as the Pisa plain, it is crucial to be able to clearly distinguish between these two phenomena in order to implement the appropriate monitoring actions. We suggest accurate monitoring by competent authorities in order to better identify intrusion via surface water also through
84
F. Franceschini, R. Signorini / Groundwater for Sustainable Development 2-3 (2016) 73–84
the hydrogeological modelling of the water table aquifer contained in the CDRs. We believe that our results constitute a tool for managing a network of datasets that local authorities can use to improve land management and water resources, and also to prevent possible environmental damage caused by salty pollution via surface water.
References Amores, M.J., Verones, F., Raptis, C., Juraske, R., Pfister, S., Stoessel, F., Antón, A., Castells, F., Hellweg, S., 2013. Biodiversity impacts from salinity increase in a coastal wetland. Environ. Sci. Technol. 47 (12), 6384–6392. Amorosi, A., Bini, M., Giacomelli, S., Pappalardo, M., Ribecai, C., Rossi, V., Sammartino, I., Sarti, G., 2013. Middle to late Holocene environmental evolution of the Pisa coastal plain (Tuscany, Italy) and early human settlements. Quat. Int. 303 (2013), 93–106. Amorosi, A., Sarti, G., Rossi, V., Fontana, V., 2008. Anatomy and sequence stratigraphy of the late quaternary Arno valley fill (Tuscany, Italy). Advances in Application of Sequence Stratigraphy in Italy. 1; , pp. 55–66. Anderson Jr., W.P., 2002. Aquifer salinization from storm overwash. J. Coast. Res. 18, 413–420. Baldacci, F., Bellini, L., Raggi, G., 1994. Le risorse idriche sotterranee della Pianura Pisana. Atti Soc. Tosc. Sci. Nat. Mem. Ser. A 101, 241–322. Baldacci, F., 1999. Struttura e piezometria del “primo acquifero artesiano in ghiaie” della Pianura Pisana. Atti della Soc. Tosc. Sci. Nat. Mem. Ser. A 106, 90–101. Baldacci, F., Pierotti, L., Spandre, R., Tessitore, S., 1999. I1 sistema acquifero freatico della pianura Pisana. Rapporti fiume-falda e acqua dolce-acqua salata. In: Procedings of the Conference Conoscenza e salvaguardia delle aree di pianura, Ferrara, 8 November. Barrocu, G., 2003. Seawater intrusion in the coastal aquifers of Italy. In: Calaferra, J., (Ed.), State of seawater intrusion in coastal aquifers of the Mediterranean Coast, SWIM-SWICA, Alicante, Spain, pp. 207–223. Butteri, M., Doveri, R., Giannecchini, P., 2010. Gattai Hydrogeologic-hydrogeochemical multidisciplinary study of the confined gravelly aquifer in the coastal Pisan plain between the Arno River and Scolmatore Channel (Tuscany). Mem. Descr. Carta Geol. d'It., XC, 105–120. Carosi, R., Montomoli, C., Pertusati, P.C., Sarti, G., Frassi, L., 2010. Note illustrative della Carta Geologica d’Italia alla scala 1:50.000. ISPRA. Day, J.W., Britsch, L.D., Hawes, S.R., Shaffer, G.P., Reed, D.J., Cahoon, D., 2000. Pattern and process of land loss in the Mississippi Delta: spatial and temporal analysis of wetland habitat change. Estuaries 23 (4), 425–438. Della Rocca, B., Mazzanti, R., Pranzini, E., 1987. Studio geomorfologico della Pianura di Pisa. Geogr. Fis. Dinam. Quat. 10, 56–84. Doveri, M., Giannecchini, R., Butteri, M., 2010. Seawater intrusion in the VersiliesePisan coastal aquifer system (northwestern Tuscany): results from a hydrogeologic-hydrogeochemical study. Presented to SWIM21 – 21st Salt Water Intrusion Meeting, Azores, Portugal, June 21–26. Fancelli, R., 1984. Tentativo di ricostruzione dell’andamento della falda artesiana tra i-30 ed i-50 m di profondità nel sottosuolo di Pisa e dintorni. Rapporto CNR-Ist. Inc. Ric. Geotermiche, Pisa. Franceschini, S., Heusch, G., 2013. Final Mitigations Project to support the projects “Repair/Upgrade Cotongrosso Road” and “Construction of three new Pads” in the Ammunition Storage Area (ASA). Migliarino-San Rossore-Massaciuccoli Regional Park, Nulla Osta no. 304/2013. Galletti-Fancelli, M.L., 1978. Ricerche sulla subsidenza della Pianura pisana analisi polliniche di sedimenti quaternari della pianura costiera tra Pisa e Livorno. Boll. Soc. Geol. It. 97, 197–245.
Giannecchini, R., Menichini, M., Raco, B., Viti, F., 2010. Hydrogeological and geochemical features of the sandy unconfined aquifer between Marina di Pisa and Calambrone (NW Tuscany) affected by seawater intrusion. Abstract/Poster 85° Congresso Soc. Geol. It. In: Rendiconti Online Societa Geologica Italiana, vol. 1, pp. 179–180. Grassi, S., Cortecci, G., Squarci, P., 2007. Groundwater resource degradation in coastal plains: the example of the Cecina area (Tuscany–Central Italy). Appl. Geochem. 22, 2273–2289. Grassi, S., Cortecci, G., 2005. Hydrogeology and geochemistry of multilayered confined aquifer of the Pisa plain. Appl. Geochem. No. 20, 41–54. Lazzarotto, A., Mazzanti, R., Nencini, C., 1990. Geologia e morfologia dei comuni di Livorno e Collesalvetti. Quad. Mus. Stor. Nat. Livorno 11, 1–82. Legislative Decree No. 30/2009, 2009. Attuazione della direttiva 2006/118/CE, relativa alla protezione delle acque sotterranee dall'inquinamento e dal deterioramento. Gazzetta Ufficiale No. 79, 4 April. Legislative Decree No. 31/2001, 2001. Attuazione della direttiva 98/83/ce relativa alla qualità delle acque destinate al consumo umano. Gazzetta ufficiale No. 52, 3 March. Mara, D., 1998. Waste stabilization ponds and wastewater storage and treatment reservoirs: the lowcost production of microbiologically safe effluents for agricultural and aquacultural reuse. In: Asano, T. (Ed.), Wastewater Reclamation and Reuse. Technomic Publishing Co., pp. 141–158. Milly, P.C.D., Cazenave, A., Gennero, M.C., 2003. Contribution of climate-driven change in continental water storage to recent sea-level rise. Proc. Natl. Acad. Sci. USA 100 (23), 13158–13161. Pozzo, A., Todaro, U., 1923. La bonifica idraulica di Coltano (Provincia di Pisa). Giornale del Genio Civile, LXI, fasc. IX. Pranzini, G., 2002. Groundwater salinization in Versilia (Italy). In: Proceedings of 17th SWIM, Delft, Holland, pp. 412–421. Pranzini, E., 2001. Updrift river mouth migration on cuspate deltas: two examples from the coast of Tuscany (Italy). Geomorphology 38 (2001), 125–132. Provincial administration of Pisa, 2005. La Geologia della Provincia di Pisa – Cartografia, Geositi e Banche Dati. Provincia di Pisa. Resource document 〈http://sit. provincia.pisa.it/SisterIms/html/geologia/indice_geologia.htm〉. Rossi, V., Amorosi, A., Sarti, G., Potenza, M., 2011. Influence of inherited topography on the Holocene sedimentary evolution of coastal systems: an example from Arno coastal plain (Tuscany, Italy). Geomorphology 135, 117–128. Sarti, G., Testa, G., Zanchetta, G., 2008. A new stratigraphic insight of the Upper Pliocene-Lower Pleistocene succession of Lower Valdarno (Tuscany, Italy). Geoacta 7, 27–41. Servizio Idrologico Regionale, 2014. Resource document 〈http://www.sir.toscana.it/ index.php〉. Signorini, R., 2008. Idrochimica delle acque sotterranee e superficiali come strumento di comprensione dei fattori idraulici, idrochimici e geologici dell’ingressione marina. Applicazione alla fascia costiera della Pianura di Pisa tra il Fiume Arno e lo Scolmatore (Thesis). Earth Science Department, Università degli studi di Pisa. Silvestri, N., Gorreri, L., 2008. Il monitoraggio delle acque ad uso irriguo nel parco regionale Migliarino San Rossore Massaciuccoli. Pacini (Ed.), Pisa. Spandre, R., Ceragioli, M., Spinicci, A., 1999. Realizzazione di una carta di vulnerabilità degli acquiferi della Pianura pisana compresa tra la linea di costa, il Canale Scolmatore, il Fosso della Bufalina e la città di Pontedera. Provincia di Pisa, Settore Pianificazione del Territorio. Switzer, A.D., 2014. Coastal hazard: storms and tsunamis. In: Masselink, G., Gehrels, R. (Eds.), Coastal Environments & Global Change. Wiley Publ., pp. 104–127. Turner, R.E., 1997. Wetland loss in the Northern Gulf of Mexico: multiple working hypotheses. Estuaries 20 (1), 1–13. USEPA, 1992. Guidelines for water reuse. Agency for International Development, NSFC, Morgantown, West Virginia, USA. Williams, W.D., 1987. Salinisation of rivers and streams: an important environment hazard. Ambio 16, 180–185.