Environmental Science and Pollution Research

Environmental Science and Pollution Research Large scale groundwater flow and hexavalent chromium transport modeling under current and future climatic conditions: The case of Asopos River Basin --Manuscript Draft-Manuscript Number:

ESPR-D-15-02748R2

Full Title:

Large scale groundwater flow and hexavalent chromium transport modeling under current and future climatic conditions: The case of Asopos River Basin

Article Type:

Research Article

Corresponding Author:

Zoi Dokou GREECE

Corresponding Author Secondary Information: Corresponding Author's Institution: Corresponding Author's Secondary Institution: First Author:

Zoi Dokou

First Author Secondary Information: Order of Authors:

Zoi Dokou Vasiliki Karagiorgi, Ms George P. Karatzas, PhD Nikolaos P. Nikolaidis, PhD Nicolas Kalogerakis, PhD

Order of Authors Secondary Information: Funding Information:

European Union (LIFE program) (LIFE10 ENV/GR/000601)

Abstract:

In recent years, high concentrations of hexavalent chromium, Cr(VI), have been observed in the groundwater system of the Asopos River Basin, raising public concern regarding the quality of drinking and irrigation water. The work described herein focuses on the development of a groundwater flow and Cr(VI) transport model using hydrologic, geologic, and water quality data collected from various sources. An important dataset for this goal comprised of an extensive time series of Cr(VI) concentrations at various locations that provided an indication of areas of high concentration and also served as model calibration locations. Two main sources of Cr(VI) contamination were considered in the area: anthropogenic contamination originating from Cr-rich industrial wastes buried or injected into the aquifer and geogenic contamination from the leaching process of ophiolithic rocks. The aquifer's response under climatic change scenario A2 was also investigated for the next two decades. Under this scenario it is expected that rainfall, and thus infiltration, will decrease by 7.7% during the winter and 15% during the summer periods. The results for two sub-scenarios (linear and variable precipitation reduction) that were implemented based on A2, show that the impact on the study aquifer is moderate, resulting in a mean level decrease less than 1 m in both cases. The drier climatic conditions resulted in higher Cr(VI) concentrations, especially around the industrial areas.

Response to Reviewers:

Comment 1. Harmonize notation µg/d or µg d-1 vs. nM/h or nM h-1

Dr. Zoi Dokou

Reply: the notation was harmonized (µg/d and nM/h). Comment 2. Not very clear. Any sensitive parameter should be considered in the calibration process. Even if less uncertain, sorption parameter cannot be considered as Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation

"known". Unfortunately the authors have not included the sorption parameter as a calibration parameter. This could be done in the future if more measurements become available and the model is re-calibrated. Comment 3. Is the measured head reliable for those 2 outliers ? Reply: A comment regarding the reliability of the 2 outliers was provided in the text. Comment 4. Rephrase : e.g. Four outliers (shown in Figure 6) were observed during the Cr(VI) model calibration. Reply: The sentence was rephrased according to the reviewer’s suggestion. Comment 5.What other sources? Aquifer management can be more than that + reduce water demand / losses during transport etc. Reply: A comment regarding water resources management options was added in the text on the results and discussion (475-479 lines) and the conclusion section (513-515 lines).

Comment 6. Potential remediation technologies that were suggested and evaluated under the LIFE10 ENV/GR/000601 project. References ? References on the remediation technologies were added in the text. Additional Information: Question

Response

Are you submitting to a Special Issue?

No

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation

Manuscript

Click here to download Manuscript manuscript_Asopos_Basin_revision_FINAL_R2.doc

Click here to view linked References

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

Large scale groundwater flow and hexavalent chromium transport modeling under current and future climatic conditions: The case of Asopos River Basin Zoi Dokou,* Vasiliki Karagiorgi, George P. Karatzas, Nikolaos P. Nikolaidis, Nicolas Kalogerakis Technical University of Crete, School of Environmental Engineering, University Campus, 73100 Chania, Greece. *corresponding author (email: [email protected], Phone: +302821037793, Fax.:+302821037860)

Abstract In recent years, high concentrations of hexavalent chromium, Cr(VI), have been observed in the groundwater system of the Asopos River Basin, raising public concern regarding the quality of drinking and irrigation water. The work described herein focuses on the development of a groundwater flow and Cr(VI) transport model using hydrologic, geologic, and water quality data collected from various sources. An important dataset for this goal comprised of an extensive time series of Cr(VI) concentrations at various locations that provided an indication of areas of high concentration and also served as model calibration locations. Two main sources of Cr(VI) contamination were considered in the area: anthropogenic contamination originating from Cr-rich industrial wastes buried or injected into the aquifer and geogenic contamination from the leaching process of ophiolithic rocks. The aquifer’s response under climatic change scenario A2 was also investigated for the next two decades. Under this scenario it is expected that rainfall, and thus infiltration, will decrease by 7.7% during the winter and 15% during the summer periods. The results for two sub-scenarios (linear and variable precipitation reduction) that were implemented based on A2, show that the impact on the study aquifer is moderate, resulting in a mean level decrease less than 1 m in both cases. The drier climatic conditions resulted in higher Cr(VI) concentrations, especially around the industrial areas. Keywords: hexavalent chromium; ophiolithic rocks; groundwater modeling; climate change; Asopos River Basin; heavy metals

1. Introduction In Greece, several regions are facing problems related to either water quantity, usually due to excessive water abstraction and reduced precipitation due to climate change (Giannakopoulos 2011), or water quality, due to the leaching of various contaminants in water bodies (Daskalaki and Voudouris 2008; Karavoltsos et al. 2008), or a combination of the two (Mimikou et al. 2000). One of the most threatened water bodies in the country is the Asopos River Basin, located in the Sterea Ellada Region and the River Basin District of Eastern Sterea Ellada. The surface and especially the groundwater system of the Asopos River Basin present high concentrations of trivalent (Cr(III)) and hexavalent (Cr(VI)) chromium, a situation that has generated considerable public concern. The groundwater is used mostly for irrigation purposes, and to a lesser extent for drinking water supply. The problem has been perpetuated for the past decades mainly due to lax environmental enforcement. However, in the past few years it rose to prominence as one of the most important environmental issues in Greece, pressuring the Greek government for stricter environmental enforcement (Moraetis et al. 2012). Chromium (Cr) is a metal used primarily as a coating or alloy by the metal finishing industry, because of its unique properties that include high resistance to corrosion and hardness. Cr can also be found in refractory bricks in furnaces and fireplaces and is used in leather tanning, wood treatment and pigment processing (Jacobs and Testa 2005). Cr exists under various oxidation states, the most important being its trivalent and hexavalent forms. Chromium geochemical behavior and toxicity in these two oxidation states are profoundly different. Cr(III) is almost immobile in natural systems, while Cr(VI) exhibits high toxicity, mobility and water solubility,

1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108

thus facilitating the contamination of soil, surface water, and groundwater reserves (Moraetis et al. 2012; Hellerich and Nikolaidis, 2005). The different toxicity and mobility of the two oxidation states complicates the assessment of potential human health risks. The extensive use of Cr in industrial activities has resulted in thousands of contaminated sites in Europe and North America. This has rendered Cr(VI) the focus of scientific discussion, regulatory concern, and legal posturing (Nikolaidis and Shen 2000). Still, despite the recognized severe toxicity of Cr(VI), no parametric value has yet been established for drinking water; instead the 50 μg/L limit for total chromium is still being used. However, due to recent pressure by grassroots organizations, scientists and the public, a parametric value specifically for Cr(VI) is expected to be instituted in the near future. The frequent occurrence and high toxicity of Cr(VI) has also motivated research on the fate and transport of chromium compounds in the environment (Hellerich and Nikolaidis 2005). Until recently, high levels of Cr(VI) in the environment were always attributed to anthropogenic pollution. However, it is now recognized that relatively high levels of Cr(VI) can be attributed to natural geogenic processes. These processes are especially prevalent in areas where the sediment is naturally rich in Cr(III) or Cr(VI) and natural processes that can convert Cr(III) to Cr(VI) are present. Demonstrative cases have been reported in California, USA (Oze et al. 2004; Morrison et al. 2009), Zimbabwe (Cooper 2002), Mexico (Villalobos-Aragón et al. 2012), Italy (Fantoni et al. 2002), New Caledonia (Becquer et al. 2003), on the Tibetan plateau in China (Sheng et al. 2012), Czech Republic and Poland (Novak et al. 2014) and Greece (Moraetis et al. 2012). The Cr(VI) contamination in the Asopos River Basin contamination affects several compartments of the environment: groundwater (used for human consumption and irrigation), river water and soils. The geogenic origin of chromium contamination in groundwater poses a very complex and unique contamination problem in the region (Moraetis et al. 2012). The extended Asopos River Basin has been subject of increasing research in the past few years. Vasilatos et al. (2008) found significant Cr(VI) concentrations in wells used for the urban water supply of Oropos (up to 80 μg/L) and Inofyta (up to 53 μg/L) exceeding the limit of 50 μg/L for total Cr in drinking water. Lower concentrations up to 33 μg/L and 40 μg/L were detected in the groundwater used for the urban water supply of the towns of Thiva and Schimatari, respectively. Economou-Eliopoulos et al. (2011) examined groundwater samples from the Asopos aquifer that ranged from 200m could only be attributed to deep injection of industrial wastes rather than originating from the Asopos river or Cr leaching rocks. Economou-Eliopoulos et al. (2012) studied the soil, plant-crops and groundwater system at the Thiva Basin. Regarding Cr contamination in groundwater, they found that the samples from domestic and irrigation wells throughout the Thiva Basin exhibited relatively low (8–37 μg/L) Cr(VI). The authors concluded that the low Cr(VI) concentrations in the Thiva wells may be related to their large depth. Tziritis et al. (2012) investigated Cr(VI) concentrations in groundwater from the wider area of Thiva Basin. They detected two distinctive groups of samples: the first group, located northwards of Thiva town, with concentrations ranging from 13 to 212 μg/L and the second, located near Mouriki village, with lower concentrations (9 – 10 μg/L). A recent work conducted by Panagiotakis et al. (2014) investigated the contribution of geogenic and anthropogenic Cr(VI) sources in the wider area of Thiva. Groundwater samples collected from two Cr(VI) plumes, supported the geogenic source hypothesis. Pangiotakis et al. (2015) performed a groundwater sampling campaign at wells across the wider Thiva area that revealed the presence of two hydrologically independent Cr(VI) plumes: one in the north (with maximum concentration of 160 μg/L) and one in the south of the study area (with maximum concentration of 75 μg/L). Kaprara et al. (2015) tested the tap water (and groundwater from few wells) of around 600 sites in Greece, including the Asopos River Basin. They found high Cr(VI) concentrations in the Basin only in private wells, whereas Cr concentrations in tap water of most municipalities (Thiva, Inofyta and Oropos) were very low. This finding was explained by the change of water sources that supply the area with drinking water. Economou-Eliopoulos et al. (2013) found a seasonal variation of Cr(VI) concentrations in the groundwater wells of the Oropos area, with the maximum Cr concentrations recorded during the dry period and the lowest during the wet period.

2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161

Recently, the remediation focus of Cr contaminated sites has shifted from expensive pump and treats methods to in-situ methodologies such as: chemically enhanced pump and treat, permeable reactive barriers, redox manipulation, microbial reduction, phytoremediation and natural attenuation (Fruchter 2002). Predicting Cr(VI) fate and transport is one of the critical elements in remediating the multitude of sites. Properly calibrated numerical models, in spite of being a simplified representation of the natural system, can be useful tools towards this goal (Dragoni and Sukhija 2008). Recent case studies investigating Cr fate and transport in groundwater using modeling tools include applications in a valley near the Mojave River, California, USA (Andrews and Neville, 2003), near a chromite mine at the Sukinda Valey in India (Tiwary et al., 2005; Dhakate et al., 2008), in Ranipet, Tamil Nadu, India (Rao et al., 2011), at a landfill site in Thailand (Tantemsapya et al., 2011) and at the landfill area of Eskisehir in Turkey (Bakis and Tuncan, 2011). A smaller part of the Asopos River Basin (Oropos plain), located downstream of the largest industrial park of Greece (Inofyta municipality), was already studied by Moraetis et al. (2012). They combined a finite element groundwater model with groundwater, surface water and soil field surveys, laboratory experiments, and geochemical modeling in order to elucidate the origin and mobility of Cr(VI) in the area. The groundwater sampling they performed in the Oropos area revealed Cr(VI) concentrations up to 120 μg/L, with a seasonal variation observed in several wells. The majority of the wells exhibited maximum concentrations during the dry period, whereas only one well showed the opposite behaviour. Their main conclusion was that geogenic sources are capable of producing high concentrations of Cr in groundwater, a result contradicting the hypothesis of Economou-Eliopoulos et al. (2011) of only deep well injection of industrial wastes. In addition, they concluded that industrial contamination transported by surface waters is able to contaminate only the upper layers of the subsurface in the vicinity of the river. The study presented here builds upon the aforementioned research by focusing on a much larger area - the entire Asopos River Basin (ca. 703 km²), and investigating both groundwater flow and Cr(VI) fate and transport. To the best of our knowledge this is the first Cr(VI) flow and transport modeling study conducted on such a scale. The objectives of this work were i) to investigate the fate and transport of Cr(VI) in the area, ii) to associate high Cr(VI) levels in groundwater to potential sources (anthropogenic and geogenic) and iii) to provide guidance in the estimation of natural background levels and threshold values of Cr(VI). The groundwater flow field and Cr(VI) plume’s response under ICCP (Intergovernmental Panel on Climate Change) scenario A2 for two sub-scenarios were also assessed using the calibrated model for the next two decades. 2. Study Area The Asopos River Basin has an area of approximately 703 km2 and its water bodies (surface and groundwater) face serious water quantity and quality problems. The former are predominantly related to the use of water in agriculture and abstraction for drinking water purposes, whereas the latter are related to the significant industrial activity in the area. According to the River Basin Management Plan for the District of Eastern Sterea Ellada (GR07), the total water needs for the Asopos River Basin are estimated at 60.5 hm3 that correspond to 50 hm3 for irrigation, 4.5 hm3 for drinking purposes and 6 hm3 for industrial use. The industrial sector bloomed during the last decades to represent approximately 20% of the total national industrial production, mainly due to the proximity of the area to the capital city of Athens. An industrial park was established during the late 1970s and early 1980s around the area of Inofyta that consists of more than 450 industrial units, including metal finishing and metal manufacturing units. Despite the fact that most industries have their own wastewater treatment units, extensive wastewater disposal into the groundwater and the Asopos River has been documented over the years (Botsou et al. 2011). This industrial growth resulted in significant deterioration of the water quality in the area. The most important environmental issue in the area is related to the high concentrations of Cr(VI) both in surface and groundwater. Public concern has been increasing and the Hellenic Ministry of Environment and Climate Change recently established an Environmental Quality Standard (EQS) for the surface water bodies of the Asopos River and Emission Limit Values (ELVs) for all industries in the catchment area (Ministerial Decree 20488/2010, Official Journal of the Hellenic

3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215

Republic 749/B/31-5-2010), including limits for total Cr and Cr(VI) with maximum permissible concentrations of 110 μg/L and 11 μg/L, respectively. Due to complex geological settings in the study area, a detailed description of the type, location and depth of the geological formations was a very important step in the development of an accurate groundwater and transport model. To this end, six maps from the Institute of Geology and Mineral Exploration (IGME) were georeferenced and combined into a single map and, then, a digitized version of the main formations (polygons) encountered in the upper layers of the subsurface of the region was generated (Figure 1). In Figure 1, the grey curve marks the greater area of the Asopos River Basin while the blue lines depict the river reaches. The main geological formations encountered in the Asopos River Basin are limestones and dolomitic limestones, neogenic deposits (marls, sands, clay, conglomerates, etc.), quaternary deposits (clay, alluvial, talus cones, etc.), shales, sandstones and ultrabasic rocks. Ultrabasic rocks (peridotites and pyroxenites) are the main geogenic origin of Cr(VI). Common minerals that host Cr as Cr(III) are spinels (chromite and magnetite) and silicate minerals (pyroxene and olivine) (Moraetis et al. 2012). The process of serpentination (alteration by hydrothermal fluids) in mantle rocks like ultramafic rocks introduces new mineral phases to ultramafic rocks. Some of these minerals are serpentine (lizardite, crysotile and antigorite), chlorite, talc, and actinolite, all of which exhibit high Cr(III) content (Oze et al. 2004). Several locations with ultrabasic rocks have been observed in the Asopos River Basin, especially in its eastern part (Figure 1), providing evidence that supports the geogenic origin of chromium contamination in the area. The locations of all the wells for which vertical geological information exists are shown in Figure 2. Boring logs were created for all of these wells. Subsequently, some of them were combined to create a large number of cross-sections. For brevity, only three cross-sections are presented (Figure 3). From the vertical cross-sections it is observed that an upper layer of clay to sandy clay is prominent in many parts of the basin, underlain by interchanging layers of sand, gravel or conglomerates. In some cases, a second layer of clay is observed between these layers. 3. Materials and Methods 3.1 Data As a first step, all available data (geological information from boreholes, groundwater levels and Cr(VI) concentrations at wells) pertinent to the groundwater flow and chromium transport modeling of the area were organized into maps, using geographical information systems and geotechnical graphics software. The data were collected from various sources such as field campaigns, existing reports (Thiva and IGME reports), and data logs (Dokou et al. 2013) as categorized in Figure 1. Note that in Figure 1 the available documented wells in the extended area of Asopos River Basin are shown, but not all contained information regarding groundwater levels, pumping rates and Cr(VI) concentrations, resulting in a lot of missing data. For the modeling work presented here, only the wells located inside the river basin were used. Part of the available data include detailed information regarding the geology of the area of study, based on existing geological maps as well as on-well profiles (boring logs) as described above. This information was utilized for the determination of the geological stratification and 3-D representation of the physical system, including the determination of aquifer properties (e.g. porosity, hydraulic conductivity) using parameter ranges obtained from the literature based on the geological formation type. A potentiometric map was also created using hydraulic head data that correspond to the initial conditions of the model. Another important source of data was comprised of time series of Cr(VI) concentrations at various locations. All of the above data were incorporated in a three-dimensional finite element groundwater flow. The model enabled the prediction of groundwater flow and Cr(VI) fate and transport in the Asopos River Basin. 3.2 Groundwater flow and Cr(VI) transport model 3.2.1. Model description In order to describe the hydrologic and chemical phenomena of the area, the PTC (Princeton Transport Code) numerical simulation model was employed. PTC is a three-dimensional groundwater flow and contaminant transport simulator that uses a combination of finite element

4

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

216 217 218

219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263

and finite difference methods to solve a system of partial differential equations that represents the groundwater flow, velocity and contaminant mass transport of the simulated physical system (Babu et al. 1997):   (K  h)  0 (1) (2) c    (D  c)    ( v c)  0 t (3) K v  h n In these equations, h is the hydraulic head, K the hydraulic conductivity, D the hydrodynamic dispersion, c the solute concentration, n the effective porosity, and v the average linear velocity of groundwater. The above system of equations includes: the steady state flow of water through a porous medium (Equation 1), the transport equation that describes the contaminant concentration migration with time (Equation 2) and Darcy’s Law for the average linear velocity (Equation 3). PTC employs a splitting algorithm for solving the fully three-dimensional equations which provides considerable savings both in computer memory requirements and computational effort. The domain is discretized into approximately parallel horizontal layers. Within each layer a finite element discretization is used allowing for accurate representation of irregular domains. The vertical connection of the layers is accomplished using finite difference discretization. This hybrid coupling of the finite element and finite difference methods provides the opportunity to divide the computational effort into two steps during a given time iteration (splitting algorithm). In the first step, all the horizontal equations are solved using finite elements while in the second step, the vertical equations which connect the layers are solved using finite differences (Babu et al. 1997). The PTC model has been successfully used in several previous studies (Aivalioti and Karatzas 2006; Dokou and Pinder 2011; Moraetis et al. 2012; Karatzas and Dokou 2015). 3.2.2 Model development The model simulates a period of 18 years, from September 1995 (when the first hydraulic head and Cr(VI) concentration data are available) to September 2013 (when the latest data are available). Initial conditions were used for hydraulic heads and Cr(VI) concentrations based on measurements collected and reported in 1995. For the simulation of groundwater flow and Cr(VI) transport in the Asopos River Basin, the model domain was discretized using a triangular mesh with 4747 nodes and 8992 elements. The mesh is denser at locations where pumping wells exist, allowing for higher accuracy when the flow equations are solved. The model was discretized vertically in three layers based on the boring log information and the vertical cross-sections created for the area (note that only AA’ and BB’ are located inside the model domain). The complex geological information presented in the crosssections of Figure 3 had to be simplified in order to be used as input to the simulator. Based on the boring logs of a majority of wells (not shown here for brevity) after a certain depth the main geological formation encountered is limestone with lenses of different geological formations. Thus in the model, the bottom (100 m thick) and middle layers (60 m thick) are assumed to consist mainly of limestone with occasional lenses of different hydraulic conductivities. The top layer initially followed the topography and geology of the area presented in Figure 1. During the calibration procedure variations in the hydraulic conductivities of the geological formations, attempted to capture the presence of more than one geological formation (e.g. when clay and sand are both present, the hydraulic conductivity of this formation was lower than the one for clean sand material etc.) 3.2.3 Flow and transport boundary conditions A first type boundary condition (i.e. specified head) with a zero hydraulic head value was set along the coastline. Asopos River was incorporated as a first type boundary condition in the model only in the top layer of the model. Pumping wells in the area of interest were defined as second type boundary conditions (i.e. specified flow) and their screens were set at various depths, depending

5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317

on the information on their boring logs (when available). A total of 199 wells (or well clusters) are active in the area (Bottom layer: seven wells with pumping rates 480-9312 m3/d, middle layer: 32 wells with pumping rates 216-840 m3/d and top layer: 160 wells with pumping rates 120-9504 m3/d). Out of the 199 wells, 9 cases are clusters of two-four closely located wells that had to be grouped due to meshing difficulties. This explains the large pumping rates in some cases reaching 9500 m3/d, which represents the sum of all pumping rates of the cluster. It is assumed that drinking wells (about 50 wells) are active during the whole year, while irrigation wells pump only during the summer months according to local irrigation schedules. In addition, a time dependent influx was defined on the southern model boundary (as a first type boundary condition), based on information from the River Basin Management Plan for the District of Eastern Sterea Ellada (GR07), that showed that most of the influx in the basin occurs in the southern part. In the northern part, boundary conditions of 1st and 2nd type were defined only locally during the flow calibration process (the default boundary condition of the model is no-flow). Additionally, based on the available meteorological data for the area at meteorological stations of Tanagra and Kalithea, an average monthly infiltration rate, equal to 30% of the average monthly rainfall rate, was included as an inflow parameter on the top layer of the model. First type boundary conditions for Cr(VI) transport were imposed at various locations during the calibration process that correspond to industrial areas (30-4000 μg/L) based on the available Cr(VI) measurements of the plumes resulting from these sources (later shown in Figure 6c). For the geogenic sources, second type (mass flux) boundary conditions were applied in deeper layers that were consistent with laboratory experiments for soils from the region collected in previous work (Moraetis et al. 2012). According to these experiments, the rate of Cr leaching to water from the serpentinite rock/soil area and the aquifer sediments was 0.108 μg/d (or 4.24 nM/h) and 0.0192 μg/d (or 0.77 nM/h), respectively. These initial values were fine tuned during the calibration process, in order to match the observation data more closely with final values of 0.12 μg/d and 0.015 μg/d, respectively. 3.2.4 Μodel parameters The parametrization of hydraulic conductivity (K) varied depending on the different geological formations encountered in the area. The specific values used were determined during the flow calibration process starting from a range of literature values (Table 1). The hydraulic conductivity values for the same geologic formation varied also from area to area, to account for the existence of small lenses inside the formation as well as interchanging layers of formations vertically inside the same model layer. This was necessary due to the geological complexity of the area. A spatial map of the final calibrated K values is presented as supplementary material (Figure S1). The calibrated porosity values for each geological formation along with the corresponding ranges from the literature are also presented in Table 1. For the longitudinal and transverse dispersivities values from 1-20 m and 0.1-2 m respectively were tested and the best results were obtained using the values of 5 m and 0.5 m, respectively. A sorption value of 4.7 (linear isotherm) that was estimated by sorption experiments previously conducted using geologic materials from the region (Moraetis et al. 2012) was also included in the model. Since this value was calculated by specific experiments was considered less uncertain than dispersivity thus it was not used as a calibration parameter, despite the fact that it is an important model parameter for Cr(VI) transport and fate. Dispersivities and sorption parameters were considered homogeneous throughout the model domain. 3.2.5 Climate change and groundwater modeling Groundwater recharge is a sensitive function of the climatic factors, local geology, topography and land use. It is a process greatly affected by variations in the regime and quantity of precipitation, temperature and evapotranspiration (Dragoni and Sukhija 2008). The climate changing conditions are expected not only to influence aquifer recharge and discharge, but also to affect the quality of groundwater, which is an effect that is just beginning to be studied, especially in groundwater (Delpla et al. 2009). For example, water recharged during a dry period may have higher contaminant concentrations, while during a wet period the opposite might occur (Sukhija et al.

6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370

1998). Such behavior has been observed by Economou-Eliopoulos et al. (2013) and Moraetis et al. (2012) in the Oropos plain, which is a subarea of the Asopos River Basin. However, to appreciate such changes long-term monitoring of rainfall and groundwater quality is required. Here, an attempt to associate the occurrence of Cr(VI) in groundwater due to water–rock processes and climate change process was made. Although different global change climate models often provide different and conflicting results, most scientist agree that the precipitation pattern in Europe is trending towards wetter conditions in the North and drier conditions in the South and Central regions (IPCC 2014). Tolika et al. (2012) analyzed the results from various regional climate models (RCMs) in regards to the potential regional future changes in seasonal (winter and summer) temperature and precipitation over the 21st century, under A2, A1B and B2 future emission scenarios of IPCC for the country of Greece. In order to reduce the uncertainties in climate-related applications, multi-regional climate models are generally found to perform better than single-model forecasts (Tebaldi and Knutti, 2007). Their results for the A2 scenario that is of interest in this work, showed that Greece is expected to experience a general precipitation decrease during winter and summer. This reduction is more pronounced in the summer and becomes progressively more intense towards the south of the country. Specifically for the Asopos River Basin, a 30% decrease is expected in the winter and a 50% decrease in the summer by the end of the century combining the results of all models (Tolika et al. 2012). Two sub-scenarios were implemented regarding the precipitation reduction under emission scenario A2. In sub-scenario 1 the reduction was applied linearly each year starting from 2013 onwards (this means that each year is considered drier than the previous, consistently). If the above percentages are translated to linear yearly reduction the corresponding reductions are 0.4% for the wet and 0.8% for the dry periods; thus, in the climate change sub-scenario 1, groundwater infiltration was reduced by the above percentages each wet and dry period. These reductions were applied each year up to 2033, which corresponds to the end of the simulation period of 20 years (2013-2033) modeled here, leading to a total decrease of about 7.7% during the winter periods and 15% during the summer periods by 2033. Under sub-scenario 2, the statistical characteristics of precipitation (and thus infiltration) of the past 20 years were kept, thus a reduction of 7.7% (winter periods) and 15% (summer periods) was applied on the infiltration time-series. Under this sub-scenario the precipitation variability was conserved, which is a more realistic representation of future conditions. For both sub-scenarios, the model boundary conditions that simulate lateral inflows to the basin were also reduced by the same percentages likewise, assuming that infiltration in connected basins is affected in a similar manner. As water resources in the area are expected to decrease due to climate changing conditions, the aquifer is expected to be more exploited in the future. For this reason, the current pumping rates in the model, occurring mainly during the summer months, were increased by 15%, for both sub-scenarios. Regarding the Cr(VI) boundary conditions that describe the leaching process of ophiolithic rocks, they were kept the same as it is assumed that the leaching rates will not be affected by the climatic variability, since only the saturated zone is modeled here. Although the leaching rate of Cr through the ophiolithic rocks is kept the same, Cr(VI) concentrations are expected to increase in several regions, due to drier conditions, as discussed above. 4. Results and discussion 4.1 Current conditions 4.1.1. Flow model calibration In this work, manual calibration was used for both the flow and transport calibration. Although automated calibration is an easier method, using the “trial and error” technique the user is able to utilize his/her experience and knowledge of the physical system in order to produce more realistic results. By changing parameter values and analyzing the results, the modeler develops a better understanding of the model and the design on which it was based. In addition, PTC does not have the capability of automatic calibration.

7

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423

Flow calibration was performed by varying a) the hydraulic conductivity values of the model within pre-specified ranges for each geologic formation obtained from the literature (as described above and shown in Table 1) and b) the lateral inflow boundary conditions. The results are shown in Figure 4a. The data used for the flow calibration were collected in September 2011 at 22 selected wells. Unfortunately, hydraulic head measurements were available only for this (dry) season. The model would have been more complete of measurements were also available for the wet season. The well locations and their corresponding hydraulic head values used for the calibration process are shown in Figure 4c. In most cases, the flow calibration results show a good agreement between measured and modeled values, with the exception of two wells (that exhibit a large difference between field and model values). These outliers (GW408 and G81) are shown in light blue color in Figure 4a. Their measured hydraulic heads are very low (12 m and 32 m, for outliers 1 and 2, respectively) while the wells located close to them have significantly higher values (100-200 m and around 50 m, for outliers 1 and 2 respectively). It is possible that these values have high measurement errors, thus cannot be considered very reliable. The R2 value for the comparison between data and model results is 0.96, indicating a very good fit and the RMSE is 8.85 m. The NRMSE is 0.0424 (or 4.24%) which corresponds to a 95.76% flow model accuracy. One major drawback of the flow calibration process is that the availability of calibration data is limited to the N-E part of the study area. For most of the wells in the Basin, only measurements taken in 1995 were available and were used as initial hydraulic head values for the model. Thus, the accuracy of the model in other parts of the study area cannot be evaluated. Nevertheless, the NE part of the study area is more important for Cr(VI) transport as most of the industrial areas and geogenic sources are located close to this part of the basin. 4.1.2. Flow model validation The flow model validation was based on the rest of the available hydraulic head measurements (15 wells) in September 2011, taken at locations shown in Figure 4c. In general, the flow validation results are acceptable, although in some cases large discrepancies between modeled and measured values exist. The R2 value for the comparison between data and model results is 0.8, the RMSE is 17.23 m and the NRMSE is 0.16 (or 16%) which corresponds to 84% flow model accuracy for the validation test. 4.1.3. Hydraulic head results The hydraulic head contours for the model area are presented in Figure 5. The results concur with the hydraulic head map created using kriging interpolation in GIS (Figure S2 – supplementary material). The general direction of groundwater flow is from the south to the north, with local variations in flow direction with the highest hydraulic head values, observed in the southern part of the basin, where the inflow from adjacent basins occurs. The model temporal variability, due to the effect of dry and wet periods, is illustrated by presenting hydraulic head time-series for three selected wells (GW444, G117 and G83 located in the northern part of the Basin, in the west, central and eastern parts respectively (Figure S3 – supplementary material). The variability for wells GW444 and G117 is small (about 0.5 m) while for well G83 it reaches 1 m in some time periods. The larger fluctuations in the beginning of the time-series are due the initial condition effect. Groundwater velocities calculated by the model are relatively small for all layers, suggesting generally slow transport processes. At the middle and bottom layers, their values do not exceed 0.1 m/d inside the model domain and their maximum values are encountered at the southern border (reaching 0.7 m/d locally). For the top layer, velocities are larger but still remain at low levels. Their values do not exceed 0.3 m/d inside the model domain. 4.1.4. Transport model calibration The mass calibration was performed by varying a) the values of longitudinal and transverse dispersivities and b) the source boundary conditions in an effort to match the model results to Cr(VI) data from 111 wells collected in 2008 (January, May, and October), 2010 (July, September, and October), and 2013 (April and May). The transport model calibration was challenging due to the large number of Cr(VI) measurements. It is important to note that a significant fraction (~15%) of the water quality measurements of Cr(VI) fell below the detection limit and were therefore

8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477

reported as non-detect (ND). In this work, the common practice of setting these values at the detection limit (5 μg/L) was followed. In most cases, the mass calibration results show a good agreement between measured and simulated values (Figure 6a). The R2 value for the comparison between data and model results is 0.85, indicating a good fit, and the RMSE is 139 μg/L. The NRMSE of this dataset was 0.0421 (or 4.21%) which corresponds to a 95.79% model accuracy. The NRMSE metric is used in order to evaluate how the model performs without giving more importance to the high concentration values, since geogenic contamination is of great importance in this work in addition to contamination caused by industrial activities. Four outliers (shown in Figure 6) were observed during the Cr(VI) model calibration. Outliers 1 and 2 are non-detects that are located very close to locations with high (140-180 μg/L) measured Cr(VI) concentrations. Thus, their modeled values are also very high, very far from the detection limit. Outliers 3 and 4 on the other hand, correspond to locations where the measured values are high while the calibrated results present low concentration values. This is due to the fact that they are not located very close to a source zone. The Cr(VI) locations used for the mass transport calibration model cover a significantly larger percent of the entire Asopos River Basin, as opposed to the flow calibration data, thus the results can be considered more reliable for the entire study area. 4.1.5. Transport model results The Cr(VI) contours of the area are presented in Figure 7. There are two major sources of Cr(VI) contamination in the Asopos River Basin: i) anthropogenic contamination emanating from the direct injection or burial of industrial wastes in various industrial sites scattered in the area with the highest concentrations reaching up to 3300 μg/L at the area of Inofyta and 200 μg/L in other industrial areas (these values are within the limits expected for anthropogenic source concentrations leaching from buried industrial wastes), and ii) geogenic contamination from the leaching process of ophiolithic rocks with concentrations reaching 40 μg/L (Figure 7). The results showed that Cr(VI) migration is slow because of the relatively low hydraulic conductivity of the main geological formations found in the region that results in low groundwater velocities. The process of adsorption also plays a significant role in retarding the transport of Cr(VI) in the aquifer system. Nevertheless, the extremely high toxic concentrations of Cr(VI) found in many areas pose a serious threat to public health. 4.2 Future conditions under climate change The groundwater flow and Cr(VI) transport model was run from 2013-2033 (20 years) in order to evaluate the impact of climate change, using the sub-scenarios 1 and 2, described in section 3.2.5. The resulting water level decrease, as compared to the current conditions, induced by the dryer conditions is shown in Figure 8a for sub-scenario 1 and Figure 8b for sub-scenario 2, for 27 locations scattered within the Asopos River Basin. A mean decrease less than one meter (0.87 m and 0.84 m, respectively) is expected in both cases, with the larger water level decrease occurring at the southwestern part of the basin (maximum value of 5.38 m and 5.52 m, respectively). For sub-scenario 1 the lowest water level decrease occurs at the western part of the basin (0.008 m at point 4) while for sub-scenario 2 at the eastern part of the basin (0.034 m at point 27) although point 4 also has a low reduction for this sub-scenario as well. The spatial variation in the results is due to a combination of different local geology, more extensive pumping in regions with active wells, lower inflows from the bottom of the basin and lower infiltration values. According to the above results, under both sub-scenarios, the aquifer is affected moderately by the dryer conditions that are expected to occur in the next 20 years. A pronounced effect is only seen locally in regions located close to the inflow boundary conditions (that occur mainly in the south-western part). The results for the points located very close to the model boundary are more uncertain, as they are greatly affected even by small variations in the boundary condition values. The variation between the results for the two sub-scenarios is due to the different application of infiltration values and boundary conditions (linear reduction and reduction with conserved variability). The above results suggest that if pumping activity in the area is sustained (or increased) this will have a long-term effect on the aquifer. This gives rise to the need for sustainable aquifer management plans, which will reduce or change the spatial patterns of pumping activity and focus on alternative water sources (surface water, treated brackish or contaminated groundwater for reuse, desalination of

9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531

seawater etc), managed groundwater recharge using treated wastewater, water conservation regulation enforcement and methods to reduce water losses during transport. With regards to Cr(VI) concentrations, Figure 9 shows the concentration plume as predicted by 2033 for sub-scenario 1. A very similar plume is observed for sub-scenario 2, thus it is not shown here for brevity. For both sub-scenarios, a larger area of high concentrations on the eastern part of the basin, where the majority of industries are located, is observed as the plumes observed in 2013 have merged into larger plumes. In this areas the Cr(VI) concentrations exceed the value of 50 μg/L, which is the parametric value for total chromium concentration, since no parametric value has been instituted for Cr(VI) yet. In the rest of the basin the concentrations remain at low levels with only small variations. The Cr(VI) concentration increase under drier conditions is consistent with the findings of Moraetis et al. (2012) and Economou-Eliopoulos et al. (2013), who observed higher concentrations during dry periods. 5. Conclusions The groundwater system of the Asopos River Basin presents high concentrations of Cr and Cr(VI). As a result, there is an increased public concern in the region regarding the drinking water quality. There are two main sources of Cr(VI) contamination in the region of the Asopos River Basin: i) anthropogenic contamination emanating from the direct injection or burial of industrial wastes in various industrial sites scattered in the area with the highest concentrations reaching up to 3300 μg/L at the area of Inofyta and 200 μg/L in other industrial areas, and ii) geogenic contamination from the leaching process of ophiolithic rocks with concentrations reaching 40 μg/L. In the present work, a groundwater flow and transport model for the region was constructed for a period of 18 years (1995-2013), in order to estimate the extent of Cr(VI) plumes from natural and anthropogenic sources. The simulations revealed that geogenic sources are capable of producing high concentrations of Cr in groundwater, building on laboratory work by Moraetis et. al. (2012) to determine geogenic leaching rates. The model results can also provide guidance for the estimation of natural background levels of Cr(VI) and consequently in the determination of threshold values. Using the Cr(VI) isoconcentration map of Figure 7, a Cr(VI) concentration of approximately 20 μg/L can serve as a background level for the whole area. Climate change effects on the aquifer were also investigated under the most severe A2 climate change scenario, regarding both water quantity and quality. An infiltration decrease was assumed (linear reduction and reduction with conserved variability) as well as more intense pumping that is expected as a result of reduced water resources. The impact on the aquifer is moderate, resulting in a mean level decrease less than 1 m with a maximum decrease of about 5 m, locally. This result indicates the need for sustainable aquifer management planning, which will reduce the pumping activity by relying on alternative water sources, water conservation regulation enforcement and methods to minimize water losses during transport. The climate change effect on Cr(VI) concentrations was also investigated. A larger area of high concentrations on the eastern part of the basin, where the majority of industries are located (consistent with the findings of Moraetis et al. (2012) and Economou-Eliopoulos et al. (2013) regarding increased Cr(VI) concentrations during drier periods). The above observations underline the need for immediate action in order to restore the aquifer in many parts of the basin and ensure public safety. The simulation model developed here can assist in the planning and management of such remediation plans, as a tool to investigate alternative scenarios and remediation measures. Potential remediation technologies for the removal of Cr(VI) from groundwater that were suggested and evaluated under the LIFE10 ENV/GR/000601 project include a) use of zero valent iron (ZVI) filings (Lilli et al. 2013) b) removal by polyphenol coated nano ZVI iron (Mystrioti et al. 2015) c) anoxic-anaerobic biological treatment systems (Panousi et al. 2013) and d) phytoremediation methods using endophytic bacteria (Dimitroula et al. 2015). Acknowledgments This work was co-funded by the European Union in the LIFE10 ENV/GR/000601 project: CHARM - Chromium in Asopos Groundwater System: Remediation Technologies and Measures.

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585

The authors would like to thank the anonymous reviewers for their constructive suggestions that greatly improved the original version of the manuscript. References Aivalioti MV, Karatzas GP (2006) Modeling the flow and leachate transport in the vadose and saturated zone - A field application. Environ Model Assess 11(1):81-87. Andrews CB, Neville CJ (2003) Ground Water Flow in a Desert Basin: Challenges of Simulating Transport of Dissolved Chromium. Groundwater 41(2):219–226. Babu DK, Pinder GF, Niemi A, Ahlfeld DP, Stothoff SA (1997) Chemical transport by threedimensional groundwater flows. Princeton University USA84-WR-3. Bakis R, Tuncan A (2011) An investigation of heavy metal and migration through groundwater from the landfill area of Eskisehir in Turkey. Environ Monit Assess 176:87–98. Becquer T, Quantin C, Sicot M, Boudot JP (2003) Chromium availability in ultramafic soils from New Caledonia. Sci Total Environ 301:251-261. Botsou F, Karageorgis AP, Dassenakis E, Scoullos M (2011) Assessment of heavy metal contamination and mineral magnetic characterization of the Asopos River sediments (Central Greece). Mar Pollut Bull 62:547–563. Cooper GRC (2002) Oxidation and toxicity of chromium in ultramafic soils in Zimbabwe. Appl Geochem 17:981–986. Daskalaki P, Voudouris ΖK (2008) Groundwater quality of porous aquifers in Greece: a synoptic review. Environ Geol, 54: 505–513. Delpla I, Jung A-V, Baures E, Clement M, Thomas O (2009) Impacts of climate change on surface water quality in relation to drinking water production. Environ Int 35:1225–1233 Dermatas D, Vatseris C, Panagiotakis I, Chrysochoou M (2012) Potential contribution of geogenic chromium in groundwater contamination of a Greek heavily industrialized area. Chem Eng Trans 28:217–222. Dhakate R, Singh VS, Hodlur GK (2008) Impact assessment of chromite mining on groundwater through simulation modeling study in Sukinda chromite mining area, Orissa, India. J Hazard Mater 160:535–547. Dimitroula H, Syranidou E, Manousaki E, Nikolaidis NP, Karatzas GP, Kalogerakis N (2015) Mitigation measures for chromium-VI contaminated groundwater – the role of endophytic bacteria in rhizofiltration. J Hazard Mater 281:114–120. Dokou Z, Pinder GF (2011) Extension and field application of an integrated DNAPL source identification algorithm that utilizes stochastic modeling and a Kalman filter. J Hydrol 398(34):277-291. Dokou Z, Karatzas GP, Nikolaidis NP, Kalogerakis N (2013) Mapping of chromium in the greater area of Asopos river basin. 13th International Conference on Environmental Science and Technology (CEST), September 5-7, Athens, Greece. Domenico PA, Schwartz FW (1990) Physical and Chemical Hydrogeology. John Wiley & Sons New York 824. Dragoni W, Sukhija BS (2008) Climate change and groundwater: a short review. Geological Society, London, Special Publications 288:1-12. Economou-Eliopoulos M, Megremi I, Vasilatos C (2011) Factors controlling the heterogeneous distribution of Cr(VI) in soil, plants and groundwater: Evidence from the Assopos basin, Greece. Chem Erde-Geochem 71:39-52. Economou-Eliopoulos M., Antivachi D., Vasilatos C. and Megremi I. (2012) Evaluation of the Cr(VI) and other toxic element contamination and their potential sources: The case of the Thiva basin (Greece). Geosci Frontiers 3(4):523–539. Economou-Eliopoulos M, Megremi I, Atsarou C, Theodoratou C, Vasilatos C (2013) Spatial Evolution of the Chromium Contamination in Soils from the Assopos to Thiva Basin and C. Evia (Greece) and Potential Source(s): Anthropogenic versus Natural Processes. Geosciences 3:140158. Fantoni D, Brozzo G, Canepa M, Cipolli F, Marini L, Ottonello G, Zuccolini MV (2002) Natural in groundwaters interacting with ophiolitic rocks. Environ Geol 42:871-882.

11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638

Giannakopoulos C, Kostopoulou E, Varotsos KV, Tziotziou K, Plitharas A (2011) An integrated assessment of climate change impacts for Greece in the near future. Reg Environ Change 11:829– 843. Fruchter J (2002) In situ treatment of chromium-contaminated groundwater. Env Sci Tech 36:464472. Heath RC (1983) Basic ground-water hydrology: U.S. Geological Survey. Water-Supply Paper 2220 86. Hellerich LA, Nikolaidis NP (2005) Sorption studies of mixed chromium and chlorinated ethenes at the field and laboratory scales. J Environ Manage 75:77-88. IPCC (2014) Climate Change 2014 Impacts, Adaptation, and Vulnerability Part A: Global and Sectoral Aspects, Working Group II Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change V. R. B. Christopher B. Field, ed., New York, NY 10013-2473, USA: Cambridge University Press. Jacobs JA, Testa MS (2005) Overview of chromium (VI) in the environment: background and history. Chromium Handbook. Kaprara E, Kazakis N, Simeonidis K, Coles S, Zouboulis AI, Samaras P, Mitrakas M (2015) Occurrence of Cr(VI) in drinking water of Greece and relation to the geological background. J Hazard Mater 281: 2–11 Karatzas GP, Dokou Z (2015) Optimal management of saltwater intrusion in the coastal aquifer of Malia, Crete (Greece), using particle swarm optimization. Hydrogeol J 23 (6): 1181-1194 Karavoltsos S, Sakellari A, Mihopoulos N, Dassenakis M, Scoullos MJ (2008). Evaluation of the quality of drinking water in regions of Greece. Desalination 224(1): 317-329. Lilli MA, Nikolaidis NP, Karatzas GP, Kalogerakis N, Mueller M (2013) Mitigation measures for hexavalent chromium contaminated groundwater and soil using zero valent iron technologies. 13th International Conference on Environmental Science and Technology (CEST), September 5-7, Athens, Greece, Mimikou M, Baltas E, Varanou E, Pantazis K (2000) Regional Impacts of Climate Change on Water Resources Quantity and Quality Indicators. J Hydrol 234: 95 – 109. Moraetis D, Nikolaidis NP, Karatzas GP, Dokou Z, Kalogerakis N, Winkel LHE, PalaiogianniBellou A (2012) Origin and mobility of hexavalent chromium in North-Eastern Attica, Greece. Appl Geochem 27: 1170-1178. Morrison JM, Goldhaber MB, Lee L, Holloway JM, Wanty RB, Wolf RE, Ranville JF (2009) A regional-scale study of chromium and nickel in soils of northern California, USA. Appl Geochem 24:1500-1511. Mystrioti C, Sparis D, Papasiopi N, Xenidis A, Dermatas D, Chrysochoou M (2015) Assessment of Polyphenol Coated Nano Zero Valent Iron for Hexavalent Chromium Removal from Contaminated Waters. Bull Environ Contam Toxicol 94:302–307. Nikolaidis NP, Shen H (2000) Conceptual site model for evaluating contaminant mobility and pump and treat remediation. Global Nest 2(1):67-76. Novak M, Chrastny V, Cadkova E, Farkas J, Bullen TD, Tylcer J, Szurmanova Z, Cron M, Prechova E, Curik J, Stepanova M, Pasava J, Erbanova L, Houskova M, Puncochar K, Hellerich LA (2014) Common occurrence of a positive δ53Cr shift in Central European waters contaminated by geogenic/industrial chromium relative to source values. Environ Sci Technol 48:6089−6096. Oze C, Fendorf S, Bird DK, Coleman RG (2004) Chromium geochemistry in serpentinized ultramafic rocks and serpentine soils from the Franciscan Complex of California. Am J Sci 304:67-101. Panagiotakis I, Dermatas D, Vatseris C, Chrysochoou M, Papassiopi N, Xenidis A, Vaxevanidou K (2014) Forensic investigation of a chromium(VI) groundwater plume in Thiva, Greece. J Hazard Mater 281:27–34. Panousi E, Mamais D, Noutsopoulos C, Bertoli K, Koutoula K (2013) Biological treatment of groundwater with high nitrates content for hexavalent chromium removal. 13th International Conference on Environmental Science and Technology (CEST), September 5-7, Athens, Greece.

12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692

Psarropoulou ET, Karatzas GP (2014) Pollution of Nitrates - Contaminant transport in heterogeneous porous media: a case study of the coastal aquifer of Corinth, Greece. Global Nest J, 16(1): 9-23. Rao GT, Rao VVSG, Ranganathan K, Surinaidu L, Mahesh J, Ramesh G (2011) Assessment of groundwater contamination from a hazardous dump site in Ranipet, Tamil Nadu, India. Hydrogeol J 19:1587–1598. Sheng J, Wang X, Gong P, Tian L,Yao T (2012) Heavy metals of the Tibetan top soils. Level, source, spatial distribution, temporal variation and risk assessment. Environ Sci Pollut Res 19:3362–3370 Sukhija BS, Reddy DV, Nagabhushanam P (1998) Isotopic fingerprints of paleoclimates during last 30 K years in deep confined groundwaters of southern India. Quaternary Res 50: 252–260. Tantemsapya N, Naksakul Y, Wirojanagud W (2011) Mathematical modeling of heavy metals contamination from MSW landfill site in Khon Kaen, Thailand. Water Sci Technol 64(9):1835– 1842. Tebaldi C, Knutti R (2007) The use of the multi-model ensemble in probabilistic climate projections. Philosophical Trans R Soc A 365(1857): 2053-2075. Tiwary RK, Dhakate R, Rao VA, Singh VS (2005) Assessment and prediction of contaminant migration in ground water from chromite waste dump. Environ Geol 48:420–429. Todd DK, Mays LW (2005) Groundwater Hydrology. Third Edition, Wiley, 636. Tolika CK, Zanis P, Anagnostopoulou C (2012) Regional Climate Change Scenarios For Greece: Future Temperature And Precipitation Projections From Ensembles Of RCMs. Global NEST Journal 14(4):407-421. Tziritis E, Kelepertzis E, Korres G, Perivolaris D, Repani S (2012) Hexavalent chromium contamination in groundwaters of Thiva basin, central Greece, Bull Environ Contam Toxicol 89:1073–1077. Vasilatos C, Megremi I, Economou-Eliopoulos M, Mitsis I (2008) Hexavalent chromium and other toxic elements in natural waters in Thiva-Tanagra-Malakasa basin, Greece. Hell J Geosci 43:57–66. Villalobos-Aragón A, Ellis AS, Armienta MA, Morton-Bermea O, Johnson TM (2012) Geochemistry and Cr stable isotopes of Cr-contaminated groundwater in León valley, Guanajuato, Mexico. Appl Geochem 27:1783-1794. Figure captions Figure 1. Digitized geological polygons of the Asopos River Basin area and existing well location in the extended area (different colors correspond to different data sources). The location of the Industrial park is also shown. Figure 2. Locations of log data and selected cross-section locations Figure 3. Selected geological cross-sections of the extended Asopos River Basin. Note that the vertical hatching corresponds to well screens. Figure 4. a) Measured vs modeled hydraulic heads (in m) for model calibration b) Measured vs modeled hydraulic heads (in m) for model validation and c) well locations with hydraulic head data (in m) used for the calibration (orange dots) and validation (blue dots) processes Figure 5. Calibrated hydraulic head contours Figure 6. Cr(VI) calibration results for a) all values, b) values < 200 μg/L, c) calibration well locations Figure 7. Calibrated Cr(VI) concentrations for the year 2013 Figure 8. Climate change impact on groundwater levels for the year 2033 at 27 locations (blue dots) for sub-scenario 1 (green bars) and b) for sub-scenario 2 (magenta bars) Figure 9. Climate change sub-scenario 1 results for Cr(VI) concentrations (plume) for the year 2033 Table captions Table 1. Representative ranges and calibrated hydraulic conductivity (m/d) and porosity values for the geological formations encountered at the study area

13

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

693 694 695 696 697 698

Supplementary material Figure S1: Calibrated hydraulic conductivity values (m/d) Figure S2: Hydraulic head map created using kriging interpolation in GIS for comparison reasons Figure S3: Temporal variability of hydraulic heads (in m) for three selected wells (GW444, G117 and G83).

14

Figure1

Click here to download Figure Figure1.tif

Figure2


Figure3


Figure4

Click here to download Figure Figure4.TIF

Figure5


Figure6


Figure7


Figure8


Figure9


Table1

Table 1. Representative ranges and calibrated hydraulic conductivity (m/d) and porosity values for the geological formations encountered at the study area

Geological Unit Neogenes (mixture of marls, clay, sands and conglomerates) 1 Limestones (varying degrees of fracturing)2 Mixture of peat, loam and conglomerates3 Clay1 Alluvial deposits1 Conglomerates3 Talus cones3 Ultrabasic rocks3 Shales – sandstones1 Clay – sandstones1 1Health

K range (m/d) 5 x 10-5– 100

Calibrated K (m/d) 0.5-30

10-4 – 10

10-3– 1.5

0.15-0.3

0.2

0.1-500

5-500

0.4-0.8

0.5

5 x 10-7 – 5 x 10-3 5 – 600 1-150 10-2-10 10-8– 10-3 5 x 10-5 – 10-2 5 x 10-5 – 10-3

10-6 5-200 10 5 10-3 5 x 10-4 5 x 10-4

0.4-0.5 0.35-0.45 0.25-0.35 0.2-0.3 0.25-0.35 0.3-0.4 0.35-0.45

0.5 0.35 0.25 0.3 0.35 0.4 0.4

(1983), 2Domenic and Schwarz (1990), 3Todd and Mays (2005)

Porosity range 0.2-0.5

Calibrated porosity 0.3

Supplementary Figure S1

Click here to access/download

Supplementary Material FigureS1.TIF



Supplementary Material FigureS2.tif



Supplementary Material FigureS3.tif