Hydrochemistry and Nitrate Pollution of Groundwater in the Alluvial Aquifer of the Eastern Mitidja (Algeria)

. The study carried out in the eastern Mitidja (Algeria) using groundwater chemistry data is aimed at clarifying the geochemical behavior of the alluvial groundwater as well as assessing the degree of nitrate contamination. The aquifer is characterized by a heterogeneous lithology marked by alternating levels of alluvium, gravel, sand and clay. The chemical facies encountered are: Na - Cl type with 52.7%, Mixed Ca - Mg - Cl type with 29.1%, Mixed Ca - Na - HCO 3 type with 9.1% and Ca – HCO 3 type with 9.1%. The interpretation of the analytical data shows that water mineralization is controlled by many ions Na+, SO42-, Cl-, Ca2+ and HCO3-. The Gibbs diagram suggests that weathering rock is the determining factor in the current chemical composition of groundwater. Agricultural activity is the main source of groundwater contamination by nitrates. Most of the water points tested have NO3 values very close to the critical threshold of 50 mg/l, the world health organization (WHO) drinking water reference value. By comparing this with the "human affected value" of 13 mg/l, 95% of the analyzed waters are considered contaminated. This water degradation is linked to the widespread use of nitrogen fertilizers, agricultural practices and increasing urbanization. mg/l to 1407 mg/l with an of 835.61 ± results reflect relatively water [40]. Calcium (Ca 2+ ) concentrations vary between 26 mg/l and 135 mg/l and magnesium (Mg 2+ ) concentrations range between 0.7 mg/l and 74 mg/l. The average concentrations are respectively 60.16 mg/l ± 16.79 mg/l and 29.41 mg/l ± 18.18 mg/l. These values are related to the alluvial nature of the aquifer containing limestone fragments and the clay matrix responsible for the basic exchanges. The minimum sodium (Na +) and potassium (K + ) contents vary between 60 mg/l and 0 mg/l respectively and the highest reach 300 mg/l and 19.5 mg/l respectively. The average Na + content is 128.05 mg/l ± 41.56 mg/l. The high Na + contents come from cationic exchanges with clays and sea spray from the Mediterranean Sea. Chloride (Cl - ) concentrations range from 15.60 mg/l to 66.30 mg/l and with an average of 34.53 mg/l ± 10.58 mg/l. Chlorides would probably come from marine sources through aerosols, as the geology has not revealed the existence of evaporative type salt rocks and the existence of marine intrusion. Chlorides may come from anthropogenic sources due to sewage discharges be increased. On the other hand, carbonate minerals show that the IS of calcite, aragonite and dolomite range between -0.77 and 1.89, between -0.91 and 1.74 and between -2.32 and 4.39 with respective averages of -0.17; -0.31 and -0.39. These results suggest that these minerals are slightly under saturation for the majority of water points except for a few samples that indicate a tendency towards saturation. As a result, groundwater can still dissolve carbonate minerals and thus increase water mineralization.

The climate of the Mitidja plain is Mediterranean with hot and dry summers and mild and humid winters. The proximity of the Mediterranean Sea influences climatic parameters, in particular precipitation and humidity [25,26]. Proximity to the Mediterranean Sea and presence of relief often give rise to both deep and shallow convection events. The collected samples are acidified with sulphuric acid H 2 SO 4 2at pH =2. Field measurements (pH and T) are performed by a pH/EC/TDS/T° Tester, the EC is determined by means of a field conductivity meter. Calcium, sodium and potassium are evaluated by flame photometry while the total hardness is determined by the titrimetry method with respect to the EDTA standard solution [27], magnesium is deduced. Bicarbonates are determined by the titrimetric method in the presence of methyl orange. However chloride concentrations are determined by a ionometer. Nitrate and phosphorus contents are determined by UV spectrophotometry and sulphates by conductivity method. The quality of the hydrochemical data is assessed by the ion balance establishing the relationship between all cations (Ca 2+ , Mg 2+ , Na + , K + ) and the total anions (Cl -, SO 4 2-, HCO 3 -, NO 3 and PO 4 3-) for each complete analysis. Validation of analytical results is based on error calculation and a threshold of 10% is tolerated [28].

 Hydrochemistry Analysis
Some indicators and indices were used to represent hydrochimical characteristic such as Saturation index, chloro Alkaline index, Gibbs diagram. The representation of chemical facies makes it possible to identify the order of cations and dominant anions present in water. We opted for the Piper diagram because it allows us to project several water points, compare them and determine any groups formed and monitor their evolution, especially when mineralization increases. This diagram is produced using the Diagrams software [29] which is composed of two triangles, allowing to represent the cationic and anionic facies, and a diamond synthesizing the global facies.
The saturation index is an essential parameter in hydrogeological and geochemical studies because it makes it possible to identify the behaviour of certain key minerals in groundwater. In this study, the saturation indices used are those for calcite, dolomite, aragonite, gypsum, anhydrite and halite. The saturation index is calculated according to relationship 1.
Where IAP is ion activity product and K T is the equilibrium constant of the reaction considered at the sample temperature (SI is saturation index of a mineral, IAP = ion activity product of the dissociated mineral, KT = equilibrium solubility at mineral temperature). The geochemical code Phreeqc (Parkhurst and Appelo, 2013) is used to evaluate saturation indices. When groundwater is saturated with minerals, the SI values are zero; for positive SI values, the state is oversaturated, and for negative values, it indicates under-saturated [30,31]. Schoeller [32]. proposed chloroalkaline indices to highlight cation exchanges between groundwater and the solid matrix of the hydrogeological reservoir. These indices are widely used in hydrochemistry to identify changes in the chemical composition of water as it flows through the aquifer.
The chloro-alkaline indices called CAI-I and CAI-II are calculated according to the relationships (2) and (3) where all ions are expressed in meq/l: When the indices are negative, the Ca 2+ or Mg 2+ cations of the groundwater are exchanged for the Na + or K + of the surrounding rock. Thus, the aquifer waters are enriched with Na + and K + . On the other hand, if the chloroalkaline index values are positive, the opposite reaction occurs, i.e. the Na + or K + cations of the groundwater are exchanged for the Ca 2+ or Mg 2+ ions of the reservoir rock. As a result, groundwater is enriched with alkaline earth.
The Gibbs diagram is widely used in hydrogeological studies to identify the explanatory processes of water salinization. [33] recommends two diagrams to assess the dominant phenomenon between precipitation, weathering and evaporation in the geochemical evolution of groundwater.

 Nitric Pollution
The standard of potability required by the corresponding to the maximum permissible limit of 50 mg/l and the threshold of 13 mg/l [34], is used to evaluate the level of nitric pollution in samples, when, the correlation matrix is used to highlight the significant relationships at the probability threshold (0.05) between nitrate contents and the other parameters analyzed. The groundwater pollution was mapped using kriging to estimate the values of the nitrate in unsampled locations using the points around it. While, the average Cl/NO 3 ratio of 4.9 of river water is used as a reference point [35,36], for estimating dnitrification effectiveness.A field survey is conducted on same water points covers type and use of fertilizers, type of crops grown and irrigation methods.

 Groundwater Hydrochemistry
The minimum and maximum pH values of groundwater vary between 6.7 and 10.5 respectively. These data reflect a pH ranging almost from neutral to strong alkalinity. The latter can be linked to high evaporation. The average pH value is 7.17 ± 0.52.
The electrical conductivity (EC) varies between 837 µS/cm and 3450 µS/cm with an average of 1565.94 µS/cm ± 424.62 µS/cm. This variation in EC data is due to the different geochemical processes that prevail in this field of study. TDS values range from 583 to 1407mg/l with an average of 835.61 ± 196.76mg/l. These values reflect relatively fresh and mineralized waters [37]. The total water hardness (HT), in CaCO 3 equivalent, ranges from 583 mg/l to 1407 mg/l with an average of 835.61 mg/l ± 196.76 mg/l. These results reflect relatively hard water [40]. Calcium (Ca 2+ ) concentrations vary between 26 mg/l and 135 mg/l and magnesium (Mg 2+ ) concentrations range between 0.7 mg/l and 74 mg/l. The average concentrations are respectively 60.16 mg/l ± 16.79 mg/l and 29.41 mg/l ± 18.18 mg/l. These values are related to the alluvial nature of the aquifer containing limestone fragments and the clay matrix responsible for the basic exchanges. The minimum sodium (Na +) and potassium (K + ) contents vary between 60 mg/l and 0 mg/l respectively and the highest reach 300 mg/l and 19.5 mg/l respectively. The average Na + content is 128.05 mg/l ± 41.56 mg/l. The high Na + contents come from cationic exchanges with clays and sea spray from the Mediterranean Sea. Chloride (Cl -) concentrations range from 15.60 mg/l to 66.30 mg/l and with an average of 34.53 mg/l ± 10.58 mg/l. Chlorides would probably come from marine sources through aerosols, as the geology has not revealed the existence of evaporative type salt rocks and the existence of marine intrusion. Chlorides may come from anthropogenic sources due to sewage discharges or livestock.
Sulphates (SO 4 2-) range from 78.8 mg/l to 501.40 mg/l (average of 212.21 mg/l ± 55.22 mg/l). These relatively high levels are related to a natural (gypsum, pyrite oxidation,) or anthropogenic (fertilizer, air pollution) presence. Bicarbonates (HCO 3 -) with an average of 321.54 mg/l ± 102.68 mg/l have values ranging from 153.5 mg/l to 793 mg/l. They result from carbonate weathering and the dissolution of carbonic acid in aquifers The partial CO 2 pressures (pCO 2 ) of the groundwater in the alluvial aquifer range from -5.00 to -1.13 atm and have an average of -1.70 atm. This last value is much higher than the atmospheric reference set at -3.5 atm, Table (1). Only one sample has data less than -3.5 atm. This high value suggests that groundwater is confined in a hydrogeological system open to CO 2 from the soil. Groundwater receives CO 2 from the respiration of the root system (rhizosphere) and the decomposition of the soil.
The Pearson correlation matrix in table (1), highlights the various chemical processes that occur in the aquifer between water and rock. These relationships are more or less complex. Among them, we will be able to link certain ionic associations (cation -anion). Ca 2+ is correlated respectively to Cl -(0.58), SO 4 2-(0.39) while Mg 2+ is linked only to HCO 3 with a value of 0.60. The correlations between Ca 2+ and SO 4 2on the one hand and Mg 2+ and HCO 3 on the other hand suggest the dissolution of gypsum and dolomite followed by the departure of Ca 2+ and Mg 2+ by base exchanges. Na + is significantly correlated to SO 4 2-, HCO 3 and Clwith the values of (0.73), (0.57) and (0.55) respectively. The high correlation between Na + and SO 4 2-suggests the alteration of Glauber's salt (Na 2 SO 4 , 10H 2 O). The mean relationship between Na+ and Climplies the phenomenon of halite dissolution. The EC of groundwater in the alluvial aquifer is controlled significantly in descending order by HCO 3 -, Ca 2+ , Cl -, SO 4 2and Na + . Summary statistics of the physical and chemical parameters of groundwater are provided in Table 1. According to correlation matrix (Table. 2), it is observed that NO 3 is significantly correlated to Ca 2+ , K + and PO 4 3with respective values of 0.27, -0.62 and -0.47. The correlation between NO 3 and Ca 2+ is positive and relatively low compared to other significant coefficients. The relationship between NO 3 and K + is negative and relatively high (-0.62). These two parameters evolve in an inversely proportional way. The mechanism for this behaviour is probably explained by the mobile nature of NO 3 whereas K + is easily fixed in the clay-humic complex of the unsaturated zone. The correlation matrix revealed a negative and significant correlation coefficient between NO 3 and PO 4 3-(-0.47). Phosphorus with low solubility in organic or inorganic particulate form binds to sediments (sources). The phosphate mineral fertilizer inputs, which would represent 58%, are much higher than other inputs, atmospheric, livestock manure and crop residues [38]. Phosphate is not as soluble as nitrate and ammonia and tends to adsorb to soil particles and enter water bodies through soil erosion [39].  In bold, significant values, at probability threshold 0.05 (two-tailed test) The representation of water points according to the Piper diagram shows that in the cation triangle, zone D (typical potassium sodium facies) is well represented followed by zone C (not dominant type). As for the triangle representing the anions, the large number of water points are in zone G (not dominant type) and very few samples in zones F (standard bicarbonate facies) and E (standard sulphate) (Fig. 3). In order of importance, the areas relating to the chemical facies encountered after projection of cations and anions are Na -Cl type with 52.7%, Mixed Ca -Mg -Cl type with 29.1%, Ca -HCO 3 type with 9.1% and Mixed Ca -Na -HCO 3 type with 9.1%. The representation of the results of analyses of the water analysed according to the Gibbs diagram (Fig. 5) shows that only one major phenomenon is found in the evolution of the geochemical processes of the groundwater in the study area. Indeed, the points highlight the influence of rock weathering. From the diagram, we can see that the phenomenon of rock alteration is responsible for the current water chemistry. The alluvial plain and the Astian sandstones in hydraulic communication, are sedimentary rocks, easily degradable and altering in particular the clays. In addition, the relatively flat relief and high temperatures are parameters that favour the degradation of rocks by chemical means, particularly the exchange of bases. We also highlight the human action that through agricultural practices and domestic discharges in the absence of sanitation networks, in some places, contribute to the modification of groundwater chemistry. The results of the Gibbs diagram highlighted the importance of aquifer rock in geochemical groundwater processes. In order to better identify these mechanisms, we refer to some graphs widely used in the literature. A Na + / Clratio approximately equal to one is generally attributed to the dissolution of halite, while a ratio greater than one indicates that sodium is released by silicate alteration [40]. Dissolved ionic elements in groundwater are the product of weathering of rock minerals with a minor contribution from atmospheric precipitation and human activities. The contribution of atmospheric sources to dissolved salts in water bodies can be assessed by considering rainwater chemistry or by taking the ratios of elements to Cl - [41]. The average ratios Na + / Cland K + / Clof the eastern Mitidja alluvial groundwater are 6.54 and 0.07 respectively. These ratios are higher than marine aerosols (Na + / Cl -= 0.85 and K + / Cl -= 0.0176) [43]. These results suggest that the contribution of atmospheric precipitations is limited in the region. The engine of chemical contribution is the one governed by the interaction of water with rock. This observation is confirmed by the Gibbs diagram (Fig.5).

 Geochemical Processes Between Water and Solid Matrix
The results of the Gibbs diagram highlighted the importance of aquifer rock in geochemical groundwater processes. In order to better identify these mechanisms, we refer to some graphs widely used in the literature.
A Na + / Clratio approximately equal to one is generally attributed to the dissolution of halite, while a ratio greater than one indicates that sodium is released by silicate alteration [42]. Figure 6, illustrates the relationship between Na + and Cl -, the line represents the dilution line of seawater. The water samples show a clear enrichment in Na + . As the water points are far from the seawater dilution line, we claim that the effect of the known salt bevel in the region did not affect the groundwater in the study area. These results corroborate the chloro alkaline indices that showed the passage of Na + and K + from the aquifer matrix to groundwater in exchange for Ca 2+ and Mg 2+ . Na ions would come from cationic exchanges, alteration of silicate minerals and probably pollution. The relationship between Na + and Cl -, the line represents the dilution line of seawater. The water samples show a clear enrichment in Na + . As the water points are far from the seawater dilution line, we claim that the effect of the known salt bevel in the region did not affect the groundwater in the study area. These results corroborate the chloro alkaline indices that showed the passage of Na + and K + from the aquifer matrix to groundwater in exchange for Ca 2+ and Mg 2+ . Na ions would come from cationic exchanges, alteration of silicate minerals and probably pollution.
The representation of the water points studied according to the graphical (Ca 2+ +Mg 2+ ) versus (SO 4 2-+HCO 3 -) ( fig. 9) highlights perfectly the positions with respect to the equiline line (1:1). Indeed, all the points are well below this reference line. Virtually no sample is aligned to the right or above it. The dominance of SO 4 2-+HCO 3 over Ca 2+ +Mg 2+ is an indicator of silicate alteration. These graphical representations suggest that the main source of Ca 2+ , Mg 2+ , SO 4 2and HCO 3 in groundwater is away from the dissolution of carbonate (calcite, dolomite) and sulphate (gypsum, anhydrite) minerals but governed by silicate minerals.
The continuous supply of fertilizers to irrigated crops necessarily results in a source of NO 3 which feeds the groundwater through soil leaching. This migration of NO 3 to the aquifer can be facilitated by stormy rainfall events. Considering the average Cl/NO 3 ratio of 4.9 of river water, the majority of water points in the study area have Cl/NO 3 <4.9 values meaning ineffective denitrification because NO 3 accumulates in wells. Only four water points have ratio values above the reference showing effective denitrification in NO 3 reduction.
The results of the Gibbs diagram highlighted the importance of aquifer rock in geochemical groundwater processes. In order to better identify these mechanisms, we refer to some graphs widely used in the literature.
A Na + / Clratio approximately equal to one is generally attributed to the dissolution of halite, while a ratio greater than one indicates that sodium is released by silicate alteration.
The projection of the points underlines that the majority of the samples (Fig. 7 and Fig. 8) are located mainly in the silicate-altering zone. The alteration process of silicate minerals (feldspar), is accompanied by high concentrations of HCO 3 - [43]. The relatively high amounts of Na + and HCO 3 can be explained by the presence of CO 2 in groundwater. We have already pointed out that almost all the samples studied have higher pCO 2 values than the atmospheric pCO 2 . The plot of the pH versus log pCO 2 graph ( fig. 10) shows that the pCO 2 values are negatively correlated with those of the pH. CO 2 partial pressure values decrease with increasing pH values [46]. This situation may be related to an aquifer where the residence time of the waters is relatively long. It is likely that the chemical reactions affecting the feldspar of the sandstone formations of the Astian are largely responsible for this chemical composition. The chemical reaction affecting feldspar is indicated by the relationship (4):

 Nitrate Contamination
Agricultural activity often leads to deterioration in water quality caused by the massive use of fertilizers, pesticides and various chemicals. In comparison with the results, the minimum and maximum values of NO 3 are 4.9 mg/l and 53 mg/l respectively. The arithmetic mean of the 55 samples is 48.1 mg/l ± 9 mg/l. The different values of the quartiles (1st, median, 3rd) are 40.2 mg/l, 42.9mg/l and 45.5mg/l. These results indicate a trend towards large values (Fig. 11) with a more or less homogeneous distribution. When comparing with the 50 mg/l standard recommended by the WHO for drinking water. Compared to the limit of 20 mg/l, considered as a low value, 3 water points (5.4%) have a lower value. On the other hand, 50 samples (91%) have NO 3 contents ranging from 20 mg/l to 50 mg/l. Compared to the "human affected value" of 13 mg/l, about 95% of the points are contaminated. From these results, we can expect a general increase in NO 3 levels, exceeding the standard of 50 mg/l, in groundwater, if measures are not taken in time to reduce this contamination. The nitrate concentration map (Fig. 12) shows values of NO 3 exceeding 42 mg/l. Within this band, we observe well defined zonalities where NO 3 concentrations are 44 mg/l or even higher. These localities are probably contaminated, not only by nitrogen fertilizers but also by domestic discharges because they are located near urban areas. Moreover, it is known that urban agriculture is often the cause of nitrogen pollution of groundwater from wastewater and solid waste discharges [44]. The rest of the study area is relatively less affected but nitrate levels remain well above the value of 40 mg/l, which is close to the 50 mg/l standard. These relatively high nitrate levels in groundwater are the result of the migration of these particles from fertilizer use to the groundwater table, due to rainwater leaching, especially with bare soils between crops [45]. The isovalent curves of NO 3 contents are more or less closed in shape, indicating a rather localized or even punctual pollution. The curves do not correspond to the piezometry where the latter shows a flow towards the sea.