Hydrogeology and numerical modeling of coastal groundwater resources focusing on salinization risk in the Metaponto plain (Basilicata, southern Italy)

Coastal aquifers, defined as the domains where fresh continental groundwater and seawater meet (Post, 2005), represent a valuable source of water supply, but they are highly vulnerable to natural and anthropogenic changes. The phenomenon of seawater intrusion (SWI), defined as the movement of salt water invading freshwater aquifers, represents the most typical hydrogeological problem for coastal aquifers. SWI is one of the primary causes of groundwater being unsuitable for drinking and irrigation. It also affects coastal environments by changing the soil chemistry, reducing fertility, and impacting local ecosystems. The most common causes of SWI are aquifer overutilization and, locally, overpumping, but this phenomenon can be further worsened by the effects of climate change. With sea levels rising, the saline front (the interface between fresh water and salt water) can advance landward and further upstream into coastal estuaries. This process can be aggravated by drought and reduced rainfall. The resulting increase in water salinity can threaten coastal areas’ flora and fauna and damage delicate habitats such as marshes and wetlands. Around 680 million people live in low-lying coastal zones - that is expected to increase to a billion by 2050 (2022 UN Ocean Conference ). Furthermore, coastal areas, especially coastal plains, are generally very relevant from an economic, social, and environmental point of view due to the presence of agricultural and productive activities, residential and tourism settlements, wooded areas, and wetlands of high ecological value (Vespasiano, et al., 2019; Erostate, et al., 2020). SWI is one of the primary genetic sources of groundwater salinization, which affects 16% of the world’s land area (Polemio & Zuffianò, 2020). The phenomenon of SWI as a result of groundwater over-exploitation is a significant concern in many European aquifers (Scheidleger et al., 2004; EEA, 2012). In the Mediterranean regions, coastal aquifers are important reservoirs of water resources intended for drinking, but at the same time, they are hydrogeological systems extremely vulnerable to pollution phenomena due to anthropogenic pressure and the effects of climate change. Italy is one of the countries most affected by SWI due to overexploitation and climate change (Romanazzi & Polemio, 2013). The risk of groundwater's qualitative degradation highlights the importance of developing and applying efficient methods for preventing and mitigating salinization processes. The detailed knowledge of coastal aquifer systems and the assessment of vulnerability and risk to the phenomenon of SWI can contribute to more effective management and protection of groundwater resources. This thesis represents the conclusion of a three-year research activity conducted as part of the Ph.D. program under the Joint Research Agreement signed between the University of Basilicata, ENI, and CNR (Italian National Research Council). The scholarship, funded by ENI, belongs to the curriculum Methods and Technologies for Environmental Monitoring and Protection about the pre-assigned topic “Optimal management of coastal groundwater resources focusing on salinization and subsidence risks”. The main objectives were to evaluate and model the SWI process in the Metaponto coastal aquifer (Basilicata, southern Italy) by analysing the intrinsic geological and hydrogeological characteristics of the coastal plain and external factors such as excessive withdrawals and climate change causing recharge modification. This area was chosen among the various Italian coastal aquifers subject to SWI because, in addition to the effects of climate change, other anthropogenic aspects also play a significant role in facilitating the SWI process. The Metaponto coastal plain is intensively cultivated, and groundwater resources are important for the economic growth linked to tourism and agriculture (Polemio et al., 2005). The plain represents an important area for the entire Basilicata region, both for its marked agricultural vocation and for the presence of residential settlements, tourist facilities, wooded areas and wetlands. Five protected sites located near the river mouths are included in the Natura 2000 network thanks to the high ecological value of their flora and fauna. During the 20th century, the anthropogenic impact, mainly linked to the development of modern irrigation systems, land reclamation works, the overexploitation of wells, and agricultural and industrial activities, has significantly modified the plain. The land reclamation works of the marshy areas started in the 1930s have helped to lower the water table and, at the same time, the excavation of the drainage channels has facilitated the circulation of salt water also far from the coast. The coastline modifications through the construction of ports, and the phenomenon of coastal regression due to climatic variations and the progressive human changes to the beach, must also be considered. These land use changes negatively impacted the hydrogeological system threatening groundwater availability and quality along the plain and intensifying the risk of aquifer pollution. These modifications also magnified soil salinization and SWI risks. Groundwater resources are particularly exposed to quantitative degradation due to the historically unfavourable climatic conditions worsened by climate change and the growing water demand, and to qualitative degradation caused by the SWI. Low rainfall (mean annual value of 538 mm) and moderate temperature (mean annual value equal to 16.3 °C), typical of a Mediterranean climate, cause low aquifer recharge (Muzzillo et al., 2021a; Muzzillo et al., 2022). The surface and sub-surface structure of the study area, characterized by the presence of marine terraces, coastal plains, and paleo-riverbeds, is the result of evolutionary processes of sedimentation, erosion, and sea level variations due to the action of marine and fluvial morphological agents. From a geological point of view, the area belongs to the sedimentary succession of the Fossa Bradanica, made up, from bottom to top, of the Sub-Apennine Clays Formation (Upper Pliocene?-Middle Pleistocene) (Vezzani, 1967; Parea, 1986), passing upwards to the Terraced Marine Deposits (Middle-Upper Pleistocene) and the alluvial and coastal deposits (Upper Pleistocene?-Holocene), in discordance on the Sub-Apennine Clays (Tropeano et al., 2002; Pescatore et al., 2009; Cilumbriello et al., 2010; Sabato et al., 2018). The hydrogeology of the plain is characterized by the presence of a complex coastal aquifer system. The application of the inverse hydrogeological water balance (Celico, 1988; Cotecchia et al., 1990; Civita, 2005; Canora et al., 2018; Canora & Sdao, 2020) and the characterization of the hydrogeological complexes allowed us to calculate the effective infiltration, equal to about 34 Mm3/year. The investigations conducted and the data collected allowed the reconstruction of the geolithological and hydrogeological structure of the various aquifers present in the study area and the definition of the chemical-physical characteristics of the groundwater (Cilumbriello et al., 2010; Muzzillo et al., 2021a; Muzzillo et al., 2021b). In particular, the hydrochemical characteristics and the spatial distribution of the groundwater electrical conductivity showed that the coastal aquifer is partially affected by the phenomenon of SWI (Polemio et al., 2002; Muzzillo et al., 2021a; Muzzillo et al., 2021b). Firstly, in this Ph.D. research project, the proneness to SWI was explored with a multidisciplinary approach based on geophysical and hydrochemical investigations. Defining the main hydrostratigraphic features can help prevent the worsening of SWI: detailed knowledge of the aquifer bottom is important to support groundwater management. For this purpose, a significant portion of the coastal plain was selected between the Agri and Cavone Rivers. Geoelectrical measurements recorded along three Electrical Resistivity Tomography (ERT) profiles showed where the aquifer bottom pattern is deeply incised by paleovalleys. The hydrochemical study focused on 49 groundwater samples for which on‐site chemical-physical parameters (electrical conductivity at 25 °C, temperature, and pH) and main ions (Na+, K+, Ca2+, Mg2+, Cl-, NO3-, SO42-, and HCO3-) were determined. The hydrochemical data highlighted the areas with higher SWI proneness, especially where the aquifer bottom is very deep below sea level, even far from the coast. Subsequently, the chemical-physical parameters of 53 groundwater samples collected between Cavone and Bradano Rivers were detected on-site using a multiparametric probe (electrical conductivity at 25 °C, temperature, and pH), and the main ions (Na+, K+, Ca2+, Mg2+, Cl-, NO3-, SO42- and HCO3-) were determined using ion chromatography methods. The concentrations of the main constituents were reported in a Piper diagram, which made it possible to represent the geochemical facies and the types of analysed groundwater. Despite a significant dispersion of the representative points of groundwater chemistry, it was possible to recognize two main types: the bicarbonate-alkaline-earthy facies (typical of the groundwater of marine terraces and alluvial deposits) and the sulphate-chlorinated alkaline facies (referable to samples of wells located in coastal deposits). The distribution of the main dissolved ions in groundwater showed, in most cases, the enrichment of Ca2+ together with the depletion of Na+ and K+, indicating a progressive mixing with seawater. After the description of the geological and hydrogeological setting (with the inverse hydrogeological water balance) and the hydrochemical characterization, the groundwater vulnerability assessment of the Metaponto coastal plain was conducted. Firstly, the assessment of the intrinsic vulnerability of the coastal aquifer was carried out by the GIS-based application of the SINTACS method (Civita & De Maio, 1997). It considers seven aquifer parameters: water table depth, effective infiltration, unsaturated conditions, soil media, aquifer media, hydraulic conductivity, and topography. Furthermore, to consider the anthropogenic influence in the study area, the SINTACS method was modified by adding the parameter of land use (LU). The SINTACS and SINTACS-LU vulnerability indexes were provided by summing the product of ratings and weights assigned to each parameter. The intrinsic vulnerability maps showed three classes ranging from low to high vulnerability. In both cases, the south-eastern part of the coastal plain, closest to the sea, is characterized by the high vulnerability class, indicating that it is the most vulnerable to contamination due to the intrinsic hydrogeological factors. The wide central part of the study area shows the moderate vulnerability class, whereas the low one is scattered in small portions in the northern part of the plain, which represent the areas less contaminable in space and time in the case of potential pollution. In the SINTACS-LU map, some areas classified as highly vulnerable in the SINTACS method show a minor vulnerability class. These areas are localized in natural and wooded sectors of the Metaponto plain, which are less populated and where the human impact on the groundwater is minimal. Then, the groundwater vulnerability to SWI was evaluated through the GALDIT method application, also conducted using the QGIS open-source geographic information system. GALDIT is an overlay-index method with scores and weights specific for the vulnerability to SWI assessment in coastal aquifers (Chachadi & Lobo Ferreira, 2001; Chachadi & Lobo Ferreira, 2005; Lobo Ferreira et al., 2005). It allows the determination and zoning of the vulnerability index by taking into consideration the intrinsic parameters of the aquifer, such as type of aquifer (G), hydraulic conductivity (A), groundwater level (L), distance from the coastline (D), the impact of the existing status of SWI in the area (I) and aquifer thickness (T). The vulnerability map produced has three classes: low, moderate, and high, covering 70.40%, 22.65%, and 6.95% of the study area, respectively. The high vulnerability to SWI is detected along the coastal strip, within 500 m from the coast, covering an extension of about 20 km2. The high class is located close to the coast due to the proximity to the sea, the greater thickness of the aquifer, and the shallow freshwater-seawater interface. Areas characterized by moderate vulnerability extend, on average, up to 4 km in correspondence with the Agri and Cavone Rivers, with a width of about 7 km between the Basento and Bradano Rivers. In these zones, the aquifer thickness is greater than 10 m, and the highest values of groundwater electrical conductivity are found. The low vulnerability covers most of the study area and is detected where the piezometric altitude is highest, and the electrical conductivity of the groundwater has the lowest values. The vulnerability zoning reveals that the propensity to SWI is accentuated from SW to NE, widening from the coast towards the inland. The sensitivity analysis was conducted to evaluate the influence of the parameters and weights on the final vulnerability indexes obtained with the SINTACS, SINTACS-LU, and GALDIT methods. In the last part of the thesis the research performed to conceptualize and numerically model the Metaponto coastal aquifer is described. Numerical simulations were performed for the portion of the coastal plain characterized by a greater propensity to the SWI phenomenon based on the results of hydrochemical analyses and the evaluation of intrinsic vulnerability with the GALDIT method. Once the study area was selected, the geological, hydrogeological, hydrological, climatic, and hydrochemical data acquired and processed became the inputs for the numerical modeling of groundwater flow and variable-density transport. The hydrogeological and hydrochemical characterization of the aquifer was necessary to define an accurate conceptual model and the subsequent numerical simulations of the groundwater flow and SWI phenomenon. The MODFLOW and SEAWAT codes were used within the Visual MODFLOW Flex 7.0 software (© 2021 by Waterloo Hydrogeologic). Groundwater flow was reconstructed in the steady-state simulation under natural conditions. Subsequently, transient simulations were carried out for the years 1997-1999. Then three scenarios were simulated by considering changes in pumping rate and the impact of climate change in terms of different recharge values. With pumping rates above 100 l/s, SWI will not be negligible in the future, causing significant effects on coastal groundwater. The overall results highlight that the aquifer requires the definition of effective management criteria to avoid the progress of the SWI phenomenon. The outcomes, in addition to improving the knowledge of the investigated area, can provide effective support for optimal groundwater management, protection, and exploitation planning. The groundwater flow modeling and the understanding of the mechanisms that determine salinity variations can support the implementation of management criteria facing SWI, climate change, and water demand in future scenarios linked to qualitative and quantitative changes in groundwater. Research on this aquifer system could be continued and expanded to further improve knowledge of the salinization process. In particular, in situ hydrogeological and geognostic surveys could be essential for the consolidation and validation of the numerical model.