Determination of Recharge by Means of Isotopes and Water Chemistry in Shaqlawa-Harrir Basin, Kurdistan Region, Iraq

Stable isotopes (2H, 18O), 14C determination, and chemical analysis of deep groundwater and surface waters (river and springs) were used to investigate the sources of groundand surface water, groundwater recharge mechanisms as well as possible sources of ions in groundwater in the semi-arid Shaqlawa-Harrir basin in Kurdistan. One hundred water samples were taken during wet and dry season. The d-excess varies significantly depending on temperature and humidity at the vapor sources. The means of the δ18O and δ2H values in the groundwater samples are -6.8 and -36.8‰, for the spring samples -6.3 and -34.5‰, and for the river samples -9.2 and -51.3‰, respectively. The depletion in the δ18O content of some water samples is due to the higher altitude of some recharge areas (altitude effect). 14C data ranges from 3.4, 71.4 and 82.7 pmC and shows that recharges rates and means residence times vary greatly in the study area. Groundwater was mainly classified as Ca-Mg-HCO3 and Mg-Ca-HCO3 type; only five well water samples belong to the Na-Ca-HCO3 type. All water samples investigated are suitable to be used as both drinking water and irrigation water.


Introduction
Stable isotopes in groundwater hydrology give a direct insight into the movement and distribution processes within the aquifer. Environmental isotope techniques in hydrogeological investigations are very helpful for studying, utilizing, managing and developing water resources [1,2]. These techniques are suitable for solving water resource problems, especially in semi-arid areas where water resources are rare and the importance of groundwater increases. Environmental isotopes provide additional information about the origin of the water, storage properties, water dynamics (relations between surface and groundwater), and groundwater contamination [3][4][5].
Stable isotopes can be useful for distinguishing the sources of surface water and groundwater discharges. Depending on the altitude where the recharge occurs, groundwater may have special isotopic signatures [5]. Some regional environmental isotope studies were undertaken in the study area related to hydrogeological and hydrological conditions; Mawlood determined a Local Meteoric Water Line (LMWL) for Haji Omaran-Erbil city according to the data of two meteorological stations located in Syria (Halab) and in Turkey (Adana) [6]. Also, Al Manmi investigated the stable isotopic composition of major aquifers in the Rania area, Sulaymania city, to distinguish the sources of the groundwater and to determine the relation between altitude and δ 18 O [7]. Hamamin and Ali determined an LMWL by continuous monitoring and taking regular monthly samples, they established a δ 18 O-altitude relationship based on groundwater and surface water samples taken in the Basara basin, Sulaymania, Kurdistan [8].
In Kurdistan region of Iraq demand for freshwater has risen dramatically because of the growth of the population and the establishment of agricultural and industrial projects. Therefore, groundwater utilization has become more important for the Shaqlawa-Harrir area. The main objectives of this work were to investigate the ground and surface water sources, the groundwater recharge mechanism, the possible sources of ions in the groundwater, and to understand the hydrogeological processes and the hydrochemical characteristics of the groundwater from the different aquifers and surface water to discuss the possibility of using groundwater for drinking (human and animal consumption), agricultural and industrial purposes. This study aims as well to determine the groundwater age, flow paths, and to quantify the mixing of groundwater between aquifers by using stable isotopes (δ 18 O and δ 2 H) and radiocarbon ( 14 C).

Study area description
The Shaqlawa-Harrir basin is located in the northeast of Erbil city, extending between 44º 2' -44º 34' E and 36º 11' -36º 39' N. The area covers about 1150 km 2 with elevations ranging from 350 to 1500 m above sea level within the high folded zone [9,10]. The area is bounded by the Greater Zab River in the northwest, the Harrir anticline in the northeast and the Safin anticline in the southwest. The Shaqlawa-Harrir basin is divided into two major basins, Shaqlawa and Harrir basin, and three small basins: Tawska, Hiran and Harash ( Figure 1).
The climate in the basin belongs to the semi-arid Mediterranean type. It is characterized by cold, rainy winters on the one hand and long, hot, dry summers on the other hand. Meteorological data obtained from the ground meteorological station in Salahaddin district/Erbil Governorate (Pirmam meteorological station) for the period between 1992 and 2012 show that the annual precipitation is about 589 mm, with maximum and minimum mean monthly relative humidity of 71.7% in January and 33.9% in July, respectively. The maximum monthly temperature is about 35.5ºC in July and the minimum is about 8.1ºC in January (Figure 2).

Geological and stratigraphic framework
The exposed geological units are represented by 11 formations which date from Early Cretaceous to Pliocene and Quaternary deposits, represented by alluvial deposits of Holocene age and river terraces of Pleistocene age. The oldest unit of Lower Cretaceous age is the Qamchuqa formation which is predominantly carbonate and forms the core of most of the anticlines in the area. The Upper Cretaceous formations are represented by carbonate rocks of the Bekhme formation and impure carbonate rocks of the Shiranish formation. The sequence is followed by alternating flysch and carbonate sediments of the Kolosh, Khurmala, Gercus and PilaSpi formations ( Table 2 and Figure 3). The geological cross section is shown in Figure 4.    Relatively impermeable rocks represented by the Kolosh, Gercus, Fatha and Injana formations and in some cases fine alluvium sediments, which cover the central and northwestern parts of the basin, have a great impact on impeding the infiltration of water to the groundwater within the basin. The development of the basin is attributed to the structural, stratigraphic and geomorphological setting of the area. The formation of numerous large anticlines and synclines with a NW-SE axis parallel to the main structural setting are good examples of intensive uplifting and folding of geological formations during the Alpine Orogenic phases. This uplift process has significantly influenced the recharge and discharge rates to and from the basin [11].

Aquifer system
Three types of aquifers exist in the Shaqlawa-Harrir basin: karstic, fissured-karstic and porous aquifers. Karst aquifers are represented by the Bekhme formation which forms an inhomogeneous anisotropic aquifer containing large groundwater reserves. This aquifer is characterized by its very high permeability and transmissivity and by its turbulent water-flow regime. Wells in it have high specific yields and the drawdown values are very small. Most freshwater in the Shaqlawa-Harrir basin is produced from this aquifer.
Fissured-karstic aquifers are represented by the PilaSpi, Kolosh and Gercus formations containing medium to large groundwater reserves. These aquifers are characterized by high permeability and transmissivity, but to a lesser extent than those in the karstic aquifer mentioned before [12]. They are developed in limestone, dolomites, marly limestones, and dolomitic limestones and are most important for the irrigation and water supply of large areas. The flow regime in the PilaSpi formation is turbulent.   The Injana, Fatha, and Kolosh formations contain groundwater in limited and varying quantities. Impermeable layers like claystones or marls partly alternate with permeable rocks, such as fractured limestone, resulting in low permeability and limited groundwater presence. This rock complex represents either aquitards or even a fully impermeable barrier to groundwater flow (aquiclude).

Water sampling and analysis
One hundred water samples were collected during the wet season (spring 2012) and the dry season (fall 2012) to investigate seasonal variations. A total of forty groundwater samples (4 deep well samples from confined aquifers and 36 from unconfined aquifers), nine spring samples and one river sample were taken. pH, eH, T, EC, O 2 were read until constant values were achieved before the samples were collected. Samples for cations and anions were filtered through cellulose acetate filters (0.2 µm) into 50 ml polyethylene bottles. Samples for total inorganic carbon and dissolved organic carbon were collected in 100 ml glass bottles; samples for stable isotopes were collected in 30 ml polyethylene bottles and filtered in the laboratory; and three samples were collected for 14 C in 1 L polyethylene bottles.
Analysis of the stable isotopic composition (δ 2 H and δ 18 O) was undertaken using a Liquid-Water Isotope Analyzer (LWIA, model 908-0008-3001) of Los Gatos Research (LGR). Using the high performance mode, precision for δ2H and δ18O was ± 1‰ and ± 2‰, respectively. The values are reported according to V-SMOV.
Rainwater samples could not be taken during the study. Therefore, we refer to the LMWL of Hamamin and Ali [8], which was obtained from sampled rainwater of 55 events during the period from December (2) 14 C was determined by converting the dissolved inorganic carbon to CO 2 by adding concentrated phosphoric acid to a 1L water sample which was constantly stirred. The released gas was sampled in an evacuated quartz ampoule. The CO 2 gas ampoules were sent to the Poznań Radiocarbon Laboratory, Poland, to determine the percentage of modern carbon (pmC) using an accelerator mass spectrometer (1.5 SDH-Pelletron Model "Compact Carbon AMS" of the National Electrostatics Corporation, Middleton, USA).
Major cations (Ca 2+ , Mg 2+ , Na + , K + and Li + ) and anions (SO    PHREEQC was used to check the charge balance, compute the aqueous speciation and the saturation indices of certain minerals by using the database wateq4f.dat [14]. The ionic charge balance was less than ± 5%.

Interpolation and statistics
Spatial interpolation by means of Kriging was used to plot the distribution of isotopes and major ions (Figures 7,10,12 and 13). Kriging assumes that the distance or direction between sample points reflects a spatial correlation which used to explain variation in the surface. This method is a powerful geostatistical interpolation technique based on the spatial correlation of sampled points [15]. The Kriging tool fits a mathematical function to a specified number of points to determine the output value for each location. Kriging is most appropriate when you know there is a spatially correlated distance or directional bias in the data. The interpolation was performed by means of ArcGIS version 10.1. Statistical analysis (e.g. Kruskal-Wallis test) was done by SPSS version 16.

Isotopic composition of surface and groundwater
In semi-arid regions water might have rather long residence times in the top few meters of soil. The kinetic effect of water vapor diffusion from the unsaturated zone might have a greater impact on the isotopic composition of the water than fractionation due to evaporation from surface water might have [16,17]. Therefore, evaporation from the unsaturated zone is predominantly characterized by stronger evaporative enrichment with a lower slope of the isotopic composition.
The mean of the δ 18 O and δ 2 H values for the wet season in the groundwater samples were -6.6 and -36.5‰, in the spring samples -6.3and -34.3‰, and in the river water -9.4 and -52.4‰, respectively. During dry season, δ 18 O and δ 2 H in groundwater were -6.9 and -37.2‰, in the springs -6.3 and -34.7‰, and in the river water -8.7 and -50.1‰, respectively (Table 3).
All samples plot above the GMWL, which reflects the significant impact of precipitation of water vapor originating from the closed basins. The regression lines of the samples have different slopes, which either means evaporation effect or water-rock interaction [18] ( Figure  6). The d-excess of the LMWL (14.4) is larger than the d-excess of the GMWL (10). This can be explained as the effect of mixing with vapor originating from the Mediterranean Sea cyclone, which is characterized by higher deuterium excess values due to the elevated relative humidity of the atmosphere. Some groundwater samples plotting above the LMWL indicate recharge from local rain events and high elevation areas with different wind speeds and temperatures. Spring samples and some groundwater samples plot on or near the LMWL, because of the altitude effect. Few groundwater samples plot below the LMWL and close to the GMWL which indicates another source of precipitation in the area originating from the Persian Gulf. The river samples plot above the LMWL because the source of the river originates from the Taurus Mountains, Turkey. The lower d-excess for the LMWL than the MMWL indicates a less evident evaporation effect in the study area. Indeed, the d-excess varies significantly depending on the temperature and humidity of the vapor's source region. The isotopic composition is depleted in the groundwater from karstic and fissured-karstic aquifers, while it is enriched in the porous aquifer. These isotopically heavier groundwater samples from the porous aquifer as compared to the karstic and fissured-karstic aquifers may be attributed to the impact of evaporation before or during infiltration.  rainfall within the plain area, rainwater accumulates and is exposed to evaporation resulting in an enrichment of heavier isotopes and then infiltration into the porous aquifer. The karstic and fissured-karstic aquifers receive the recharge from the surrounding highly elevated area. The high altitude contributes to more depletion in heavy isotopes in rainwater, and the joints and fractures of these aquifers supply good paths for transporting this infiltrated rainwater to the lower area. Stable isotope data can be used to estimate the flow of groundwater from adjacent aquifers because of the hydraulic interconnection between aquifers with contrasting compositions.
The difference in isotopic composition supports the existence of different recharge sources in the area and the altitude effect. The distribution of the isotopic composition shows that in the mountain areas heavy isotopes are more depleted than in the plain area (Figure 7).

Altitude effect
The mean altitude effect of the different aquifers was determined by assessing the δ 18 O content of the water samples and the altitude. The mean δ 18 O of the karstic aquifer samples is -6.2‰, that of the fissuredkarstic aquifer samples is -7.3‰, of the porous samples is -5.79‰, and that of the carbonate spring samples is -6.61‰. The result of the regression line was estimated and shows that the δ 18 O content of the water samples decreases with the increase in altitude (Figure 8).
By using the non-parametric Kruskal Wallis test it was shown that the differences between the aquifers are significant on a significance level of ≤ 0.001 (Figure 9). Interpolation of the 18 O data of the basin shows that the effect of high elevation of the recharge is clearly represented by more negative isotopic signature along the southwestern border with the Safin and Pirmam Mountains, and Harrir Mountain in the eastern part of the basin as compared to the less negative values in the northern parts toward the Greater Zab River. 14 C was analyzed in two water samples, well sample (S-49) and spring sample (S-38) in the karstic aquifer. A third 14 C value (well sample S-23) in the fissured-karstic aquifer was provided by (Ahmed M, Salahaddin University, Kurdistan region, personal communication, 2013). The differences in δ 13 C between soil-derived TIC and carbonate minerals in the aquifer can provide a reliable measure of 14 C dilution by carbonate dissolution. The δ 13 C mixing model allows for incorporation of 14 C-active TIC during carbon dissolution under open system conditions, and subsequent 14 C dilution under closed system conditions ( Table 4).

Radiocarbon ( 14 C)
The q-factor according to Pearson [19] and Pearson and Hanshaw [20] was obtained from carbon isotope mass balance as: 13 13 13 13

TIC carb
Soil carb where δ 13 C TIC = measured 13 C in groundwater, δ 13 Csoil=δ 13 C of the soil CO 2 (usually close to -23‰), and δ 13 C carb =δ 13 C of the calcite being dissolved (usually close to 0‰).    The 14 C activity of well S-23 in the unconfined aquifer (fissuredkarstic) with a corrected 14 C age of 1667 ± 43 y BP and the artesian well S-49 in the confined karstic aquifer with a corrected 14 C age of 970 ± 35 y BP indicate either a rather low flow velocity or has to be interpreted as biased by additional chemical reactions. This is likely and proof once more that 14 C is anything but an ideal tracer to determine mean residence times in limestone aquifers.
The very low 14 C activity of S-38 (carbonate spring) in the karstic aquifer with a corrected 14 C age of 24782 ± 410 y BP would indicate an extreme long residence time of groundwater in the fractures of the bedrock. But this is contradiction with the data from stable isotopes, which do not show any evidence for recharge during cold climate (fossil water recharged during the last ice age). Reasons for the low 14 C value are either admixture of volcanic CO 2 or oxidation of organic matter contained in the aquifer. Volcanic CO 2 contains no 14 C and will dissolve carbonate with no 14 C and thus significantly bias the 14 C pattern. In case of oxidation of organic matter the 14 C signal of the organic matter is one further unknown variable depending on the age of the organic matter. Depending on the oxidation process (O 2 , NO 3 , SO 4, CH 4 ) and the origin of the organic matter a δ 13 C correction is often not applicable in such cases [21,22]. Because volcanic CO 2 is rather unlikely in the given geologic setting it is assumed that organic matter is the reason for the very low 14

Hydrochemistry
Field parameters and major cations and anions for both wet and dry season samples are shown in Tables 5 and 6. The spatial TDS distribution shows that low TDS values and low salinities are found in the northeast and south of the basin and high salinities in the middle part of the basin for the three aquifers according to the groundwater flow from the mountains toward the center of the basin and then to the Greater Zab River (Figures 10 and 11).
The high TDS values in the Shaqlawa-Harrir basin are resulting from the long residence time and enhanced water-rock interaction. In general, TDS and EC in the wet period are lower than in the dry period due to dilution by rainfall in the wet period. The pH of the spring samples increased from the mountains toward the Greater Zab River. The groundwater and surface water samples have pH values ranging between 7 and 8.3 with a mean value of 7.6. The average Ca 2+ and Mg 2+ concentrations are 70 mg/L and 26.6 mg/L, respectively. These which are the most abundant cations in the water samples; the average K + concentration is very low (about 1.2 mg/L).
The most abundant anion in the water samples is HCO 3¯. In groundwater generally most of the hydrogen carbonate ions are derived from the carbon dioxide in the atmosphere, soil gas and the dissolution of carbonate rocks. The average hydrogen carbonate concentration of the samples, calculated according to TIC and pH by using PHREEQC program, is 305 mg/L. Sulfate comes from the atmosphere and from the dissolution of sulfate minerals in the sedimentary rocks. Shales also may release a large amount of sulfate through the oxidation of marcasite and pyrite. The average SO 4 2¯ concentration of 34.6 mg/L is rather moderate. The most important of the Cl sources in the near-surface water seems to be chloride transported in the atmosphere and subsequently falling to earth by rain or snow. The average chloride concentration is 16.5 mg/L. Most nitrates in natural water originate from organic sources or from industrial and agricultural activities; the average concentration in the samples is 10 mg/L (Figures 12 and 13).      Most groundwater samples were classified as to be of Ca-Mg-HCO 3 and Mg-Ca-HCO 3 type; only five well-water samples belong to the Na-Ca-HCO 3 type. The physicochemical parameters for all the samples suggested that the groundwater is generally good for domestic use, irrigation and industrial purposes (Tables 7 and 8).
The relationship between the Na + and Cl¯ concentrations has been used to identify the form of salinity [23][24][25]. The high concentration of Na + and Cl¯ detected in water sample 37 in porous aquifer during dry season proposes the dissolution of chloride in the form of halite ( Figure 14).

Conclusions
The present study examined the stable isotopes and radiocarbon in groundwater from the three main aquifers in the Shaqlawa-Harrir basin. The fissured-karstic aquifer was more depleted in heavy isotopes than the karstic aquifer, while the porous aquifer is enriched. The water vapor source in the area originated from two sea cyclones; Mediterranean Sea and Persian Gulf. More negative isotopic signature along the southwestern border (Safin and Pirmam Mountain) and eastern part (Harrir Mountain) in comparison to the northern part of the basin indicates the altitude effect on the isotopic composition.
According to 14 C analysis the recharge in the area occurs in the Safin and Pirmam mountains in the south and southwest and the Harrir Mountain in the eastern part of the basin. Hydrochemical investigations indicated that the high TDS values of the groundwater in the Shaqlawa-Harrir basin result from a long residence time and enhanced water-rock interaction. The type of the water samples is mainly hydrogen carbonate and the water quality is generally suitable for drinking, agriculture, irrigation, and industrial purposes.