Process-based groundwater salinisation risk assessment methodology: Application to the Akrotiri aquifer (Southern Cyprus)

https://doi.org/10.1016/j.jhydrol.2010.12.032Get rights and content

Summary

Groundwater salinisation is a major groundwater contamination issue world-wide and can be caused by different processes, such as seawater intrusion, agrochemical pollution, geogenic contamination and irrigation-induced salinisation. In many areas, several salinisation processes are superimposed. Since remedial measures vary for different salinisation processes, correct identification is fundamental for adequate design of management strategies: different strategies may be required in one and the same aquifer, depending on which salinisation process is active where in the domain.

A simulation-based salinisation risk assessment methodology is proposed, based on the principle of linear superposition of total dissolved solutes in groundwater. In a first step, the measured bulk salinity distribution is used to calibrate a numerical groundwater flow and transport model, accounting for all identified salinisation processes. Then, the bulk salinity distribution is decomposed into different salinity components by adapting the boundary conditions, running a simulation for each salinisation process separately. These simulation results yield the necessary components to calculate the risk index distributions, which are a measure of the respective future potential salinity increase. Overlaying the risk index distributions with a defined threshold concentration reveals risk areas requiring remediation or conservation measures with respect to each process. The risk area maps resulting from this methodology are a promising tool for the design of groundwater management schemes. They condense relevant information from complex dynamic processes obtained from numerical simulations and visualise the results in simple and static maps, accessible to decision makers who are not familiar with groundwater dynamics.

The different steps of the salinisation risk assessment procedure are first described and illustrated on a synthetic example and then applied to a real aquifer system in Southern Cyprus (Akrotiri), where three major salinisation processes are superimposed.

Research highlights

► Superimposed salinisation processes are simulated separately. ► Risk indices are obtained from simulated salinity distributions. ► Salinisation risk maps are established based on the risk indices. ► Overlay with defined threshold concentration yields risk area. ► Different management actions are attributed to risk areas.

Introduction

Salinisation is a wide-spread threat to fresh groundwater resources and can be considered as one of the most prominent global groundwater pollution problems. In contrast to specific pollutants, salinity as such is not a contaminant but is merely a measure of the content of dissolved solutes. Its contamination potential and toxicity (e.g. to plants) is related to an overall concentration threshold of dissolved solutes and not to specific constituents of the water composition. Salinity may have many different origins and can either be caused by primary salinisation processes that actually add solutes to the system, such as seawater intrusion, dissolution of geogenic salt deposits or agricultural inputs (e.g. Konikow and Person, 1985, Custodio, 1997, Sites and Kraft, 2000, Pearce and Schumann, 2001). Or, it can be induced by secondary salinisation, such as solute recycling from irrigation or by evaporative processes: secondary processes do not add any solutes to the system, but lead to salinisation by redistribution or concentration of solutes already present in the system (Milnes and Renard, 2004, Milnes and Perrochet, 2006).

In areas affected by groundwater salinisation, correct identification of the spatial distribution and dynamic interplay of the superimposed salinity components is crucial, since the respective remedial or conservation measures may be entirely different. To avoid seawater intrusion, as an example, the water table has to be kept high, while to avoid irrigation-induced salinisation the groundwater table should be kept low. In many coastal irrigated aquifers, in particular in the Mediterranean region, these two salinisation processes are often superimposed.

Identification of different origins of salinisation can be considered a separate field in hydrogeological research and is efficiently addressed by means of various hydrogeochemical techniques (e.g. Richter and Kreidtler, 1993, Vengosh and Rosenthal, 1994, Custodio, 1997, Vengosh et al., 1999). As opposed to simple salinity measurements, sophisticated techniques that yield clear information about the origin of salinity are time-consuming and costly, and usually only yield snap-shots in time and space. Hence, combining information on the origins of salinity from such investigations with numerical simulations is a promising strategy, not only to investigate the dynamic interplay of different salinisation processes, but also to decompose the bulk salinity evolution into different salinity components induced by different processes.

Vulnerability and risk assessments are becoming a standard approach in groundwater management when dealing with water quality and contamination issues. Vulnerability maps are designed to identify areas of greatest potential for groundwater contamination on the basis of hydrogeological criteria, while groundwater risk assessments additionally consider the presence of potential contamination sources or polluting land-use activities (Gogu and Dassargues, 2000). The most commonly used vulnerability mapping procedures are based on empirical point rating systems, such as DRASTIC (Aller et al., 1987), EPIK (Doerfliger and Zwahlen, 1997), GOD (Foster, 1987) or SINTACS (Civita and de Rigibus, 1995), bringing together key factors which influence the solute transport process. However, it has been found that the vulnerability of aquifers to groundwater contamination can rarely be predicted with these key factors (e.g. Ruppert, 2001). Gogu and Dassargues (2000) therefore emphasize the need for physically based risk and vulnerability assessments. Such an example is presented by Stewartd and Loague (1999), who developed a regional-scale vulnerability assessment methodology to estimate the impact of non-point source groundwater contamination, using a generalised type transfer function. Sophocleous and Ma (1998) used numerical modelling combined with sensitivity analysis multiple regression analysis and classification procedures to develop a decision support model to assess spatial vulnerability to groundwater salinisation in the Great Bend Prairie aquifer of Kansas. Another example of a physically-based approach is given in Connell and van den Daele (2003), who investigated the use of analytical solutions for unsaturated solute migration, in order to calculate contaminant transport to groundwater in view of combining it with a geographic information system (GIS) to establish vulnerability maps. The spatial overlay principle of different information is common to most approaches, yielding vulnerability or risk maps, which can be used by decision makers. The overlay principle is also used in the present approach. However, the information which is overlain is entirely deduced from numerical flow and transport simulations: the spatial distribution of the information is therefore continuous.

Hence, the aim of this work is to present a physically based salinisation risk mapping procedure and to apply it to a real-case site. The calibrated observed (bulk) salinity distribution is decomposed, which means ‘picked apart’, in contrast to the common usage of the word decomposed, into several salinity components, by adapting the boundary conditions in a flow and transport model, assuming linear superposition of total dissolved solids with different origins. The decomposed salinity distributions are then used in a simple overlay principle to calculate the respective spatial risk index distributions, with which areas are identified that are prone to further salinisation. The salinisation risk index mapping procedure is first illustrated on a schematic 2D model and then applied to a real-case site in Southern Cyprus (Akrotiri aquifer) using a 3D finite element model, revealing the practical outcomes of the method, but also the limitations.

Section snippets

Definition of the risk index Ri(x)

The basis of risk assessments is the comparison of a present state of the environment with an adverse state which may potentially occur in the future, a state which will have a negative impact on human interests (e.g. Helm, 1996). Placing this general definition into the context of groundwater salinisation requires first of all definition of an ‘adverse state’ for each salinisation process in an aquifer system.

For any salinisation process we may be dealing with, the ‘adverse state’ with respect

Salinisation risk assessment components

As can be seen from Eq. (1), the salinity components required to obtain the risk index Ri(x) of the ith salinisation process are the salinity distributions Ci(x, tp) and Ci(x, ), at ‘present state’ tp and at steady-state (t = ∞), respectively, as well as the bulk salinity distribution at steady-state CB(x, ). The simulated salinity distribution Ci(x, tp) is the salinity component of the ith salinisation process at ‘present state’ tp, being the time corresponding to the measured bulk salinity

Risk mapping procedure

To identify areas that require remediation and conservation with respect to different salinisation processes, the risk indices have to be confronted with a threshold salinity CT, e.g. the salt tolerance of cultivated crops in an area. The definition of such a threshold salinity CT has to be based on criteria related to different fields, such as agronomy, irrigation science and economy and may vary in different contexts. Since definition of a threshold concentration CT is beyond the scope of

Application to the Akrotiri aquifer (Southern Cyprus)

In this section, the described salinisation risk assessment methodology is applied and tested on a coastal irrigated aquifer system in Southern Cyprus (Akrotiri aquifer), where hydrochemical and hydrogeological investigations showed that at least three salinisation processes are superimposed (Meilhac, 2003). Groundwater salinisation is caused by seawater intrusion, irrigation-induced salinisation and by evapo(transpi)ration from the groundwater table. Hence, seawater intrusion is the only

Discussion and conclusions

The presented framework for a salinisation risk assessment methodology allows process-based identification of the spatial variability of different salinisation processes, provided the salinisation processes have been correctly identified by field investigations and a numerical flow and transport model has been well calibrated.

The risk area maps resulting from this methodology are a promising tool for the design of groundwater management schemes. They condense relevant information from complex

Acknowledgements

The field work carried out and the data retrieved to elaborate the numerical model could only be done due to the close collaboration with the Cyprus Water Development Department, in particular Dr. A. Christodoulides. The author thanks Pierre Perrochet, Fabien Cornaton and Philippe Renard for fruitful discussions. The constructive comments from two anonymous reviewers were very much appreciated and considerably improved the manuscript. The research described in this paper was sponsored by the

References (30)

  • N. Doerfliger et al.

    ‘EPIK: a new method for outlining of protection areas in karstic environments’

  • Foster, S.S.D., 1987. Fundamental concepts in aquifer vulnerability, pollution risk and protection strategy. In: Van...
  • R.C. Gogu et al.

    Current trends and future challenges in groundwater vulnerability assessment using overly and index methods

    Environmental Geology

    (2000)
  • P. Helm

    Integrated risk management for natural and technological disasters

    Tephra

    (1996)
  • Jackovides, J., 1982. Southern Conveyor Project, Feasibility study. Groundwater Resources (3), Cyprus Water Development...
  • Cited by (40)

    • Hydrogeochemical characteristics of groundwater in the Mediterranean region: A meta-analysis

      2023, Physics and Chemistry of the Earth
      Citation Excerpt :

      The collected data was extracted from 123 articles conducted from 2000 till 2019 (Fig. 1). These studies focus on the following Mediterranean countries: Algeria (9 articles) (Aouidane and Belhamra, 2017; Belfar et al., 2017; Bettahar et al., 2017; Bouderbala and Gharbi, 2017; Bouderbala et al., 2016; Chemseddine et al., 2015; Belkhiri and Mouni, 2014, 2012; Fehdi et al., 2009), Croatia (1) (Terzić et al., 2007), Cyprus (2) (Milnes, 2011; Neal and Shand, 2002), Egypt (5) (Eissa et al., 2018; Salem and Osman, 2017; Salem et al., 2017; Eissa et al., 2016; Gomaah et al., 2016), France (8) (Santoni et al., 2016; Maréchal et al., 2014; Bicalho et al., 2010; Charmoille et al., 2009; De Montety et al., 2008; Huneau and Blavoux, 2000; Bourg and Richard-Raymond, 1994; Bosch et al., 1986), Greece (4) (Filippidis et al., 2016; Tziritis et al., 2016; Voutsis et al., 2015; Skordas et al., 2013), Italy (15) (Biddau et al., 2019; Preziosi et al., 2019; Allocca et al., 2018; Busico et al., 2018; De Caro et al., 2017; Critelli et al., 2015; Corniello and Ducci, 2014; Pierotti et al., 2013; Brozzo et al., 2011; Giménez-Forcada et al., 2010; Ghiglieri et al., 2009; Cucchi et al., 2008; Manno et al., 2007; Barbieri and Morotti, 2003; Capaccioni et al., 2001), Lebanon (25) (Chaza et al., 2018; Hanna et al., 2018; Khadra and Stuyfzand, 2018; Khadra et al., 2017; Samad et al., 2017; Assaker, 2016; Kalaoun et al., 2016; Koeniger et al., 2016; Bakalowicz, 2015; Youssef et al., 2015; Bakalowicz, 2014; Khadra and Stuyfzand, 2014; Daou et al., 2013; Saadeh et al., 2012; Awad, 2011; Assaf and Saadeh, 2009; Bakalowicz et al., 2008, 2007; Korfali and Jurdi, 2007; El Moujabber et al., 2006; El Hakim, 2005; Metni et al., 2004; Acra and Ayoub, 2001; Ghannam et al., 1998; Lababidi et al., 1987), Libya (2) (Alfarrah et al., 2016; Sadeg and Karahanoðlu, 2001), Morocco (6) (Chafouq et al., 2018; Mountadar et al., 2018; Karroum et al., 2017; Re et al., 2014, 2013; De Jong et al., 2008), Palestine (7) (Abu-alnaeem et al., 2018; Jebreen et al., 2018; Aliewi and Al-Khatib, 2015; Jabal et al., 2015; Da’as and Walraevens, 2013; Qahman and Larabi, 2006; Al-Agha, 2005), Portugal (8) (Freitas et al., 2019; Carreira et al., 2017; Marques et al., 2017; Barroso et al., 2015; Andrade and Stigter, 2011; Carreira et al., 2011; Fernandes et al., 2006; Van der Weijden and Pacheco, 2006), Spain (15) (Argamasilla et al., 2017; Merchán et al., 2015; Vallejos et al., 2015; Merchán et al., 2014; Molina-Navarro et al., 2014; Acero et al., 2013; Daniele et al., 2013; González-Ramón et al., 2012; Hidalgo et al., 2010; Lentini et al., 2009; Giménez Forcada and Morell Evangelista, 2008; Navarro and Carbonell, 2007; Gómez et al., 2006; López-Chicano et al., 2001; Pulido-Bosch et al., 1995), Syria (2) (Abou Zakhem, 2016; Abo and Merkel, 2015), Tunisia (10) (Hamdi et al., 2018; Mejri et al., 2018; Telahigue et al., 2018; Houatmia et al., 2016; Ben Cheikh et al., 2014; Farid et al., 2013; Hamed and Dhahri, 2013; Moussa et al., 2012; Hamzaoui-Azaza et al., 2011; Zghibi et al., 2014) and Turkey (4) (Varol and Şekerci, 2018; Somay, 2016; Tayfur et al., 2008; Demirel and Güler, 2006). For each study the key information that was systematically collected and analyzed include:

    • Identifying the origin of groundwater salinization in the Bokoya massif (central Rif, northern Morocco) using hydrogeochemical and isotopic tools

      2021, Groundwater for Sustainable Development
      Citation Excerpt :

      These reasons make karst aquifers particularly difficult to exploit, explore, protect, and to manage sustainably in terms of water quality and availability. Groundwater quality degradation can be attributed mainly to natural and anthropogenic factors such as seawater intrusion, impact of adjacent aquifers, hydrodynamic conditions, dissolution of evaporitic rocks, upwelling of deep saline water levels and uncontrolled use of fertilizers (Milnes, 2011; Elgettafi et al., 2012; Giambastiani et al., 2013; Ledesma-Ruiz et al., 2015; Ouhamdouch et al., 2021). However, knowledge of the sources and understanding of salinization processes remains a requirement to prevent further deterioration of groundwater quality.

    • SWOT risk analysis towards sustainable aquifer management along the Eastern Mediterranean

      2021, Journal of Environmental Management
      Citation Excerpt :

      Efforts towards the identification, assessment, and management of saltwater intrusion (SWI) into coastal aquifers have relied on multiple techniques and methods including geochemical and geophysical analysis, environmental tracers, laboratory experiments, hydrodynamic techniques, as well as physical modeling (Zghibi et al., 2014; Batayneh et al., 2013; Mondal et al., 2010; Werner et al., 2013; Duque et al., 2008). In parallel, groundwater vulnerability assessment models (e.g. GALDIT, DRASTIC and EPIC) have been increasingly used (Moazamnia et al., 2020; Momejian et al., 2019; Fijani et al., 2013) to quantify, assess and map the vulnerability of aquifers to SWI based on data layers covering hydrogeological and hydrochemical characteristics of the aquifers (Milnes 2011; Li et al., 2018; Shirazi et al., 2012; Klassen and Allen 2017; Solangi et al., 2019). Yet, almost all techniques are data intensive and require significant resources for this purpose.

    View all citing articles on Scopus
    View full text