Introduction

Continued population growth, the development of mining industries, and recent climate change are putting water resources under pressure in arid and semi-arid regions (e.g., Ragab and Prudhomme 2002; Bahir et al. 2018; Ouhamdouch et al. 2016, 2018). The Gobi Desert is one of the largest deserts in the world, which expands across northern China and the southern part of Mongolia (Fig. 1a). The largest mineral deposits of Mongolia are located in the Southern Gobi Region (SGR).

Fig. 1
figure 1

Map showing: a location of the study area and its relationship to the Khanbogd district and Gobi Desert, b topography of the Gunii Khooloi basin and sampling locations. The contour lines represent elevation (m asl)

Full size image

The groundwater in the SGR is in a critical situation. In this region, shallow groundwater serves as the only source of water supply for drinking, domestic and livestock watering; while deep groundwater (Gunii Khooloi aquifer) is largely utilized by exploration and mining companies such as the Oyu Tolgoi Mine (southern part of the Gunii Khooloi basin, Fig. 1b). The Gunii Khooloi basin has recently been the focus of attention due to increasing water demand by the Oyu Tolgoi Mine and its resulting environmental impact. Local communities are concerned that the groundwater that they always use may dry up or be contaminated, thereby restricting their domestic usage and negatively impacting their livelihoods as herders.

Over the past few decades, the Mongolian government and the Oyu Tolgoi company have conducted hydrogeological exploration for assessment and development of the water resources in the Gunii Khooloi basin, and for understanding the hydrogeological conditions of the study area (Munkhbaatar et al. 2004; Tuvdendorj et al. 2008). Analysis of the aquifer system and hydrogeological conditions together with geochemical and isotopic techniques can provide information on the groundwater sources and their associated recharge processes (Coplen et al. 2000; Mahlknecht et al. 2004; Glynn and Plummer 2005; Ma et al. 2009).

Therefore, although significant isotopic and hydrogeochemical surveys have been conducted over the past decade in the middle and southern part of the Gobi Desert (northern China) such as the Shiyang River basin (Ma et al. 2005), the Badain Jaran Desert (Gates et al. 2008), the Heihe River basin (Yang et al. 2011), the Dunhuang basin (Ma et al. 2013) and the Hexi Corridor (Wang et al. 2017) the northern part of the Gobi Desert (Southern Gobi Region, Mongolia) is still not well understood. Comprehensive investigations to understand groundwater geochemical evolution, recharge mechanisms, and the origin of groundwater and groundwater age that combine chemistry and isotopic indicators are also few or nonexistent in Mongolia.

Through a hydrogeological survey of shallow and deep aquifers in the Gunii Khooloi basin using geochemical and isotope methods, the current study sought to clarify (1) shallow and deep groundwater characteristics, (2) groundwater age, and (3) the groundwater origin through assessment of a combination of isotopic data and hydrogeological information.

Study area

General setting

The Gunii Khooloi basin is located in the Khanbogd district, SGR (northern part of the Gobi Desert), between 106°10′E to 108°05′E and 43°05′N to 43°55′N with an area of approximately 7,100 km2 (Fig. 1a). The area is surrounded by the Khanbogd Mountains (peak elevation 1,350 meters above mean sea level (m asl) to the south, the Zogdor Mountains (peak elevation 1,581 m asl) to the west, and the Ikh Shankh range (peak elevation 1,790 m asl) and Gurvan Kharaat Mountains (peak elevation 1,264 m asl) to the northwest and northeast (Fig. 1b). Surface runoff is ephemeral in the basin, with some permanent springs and temporary surface streams originating from the surrounding high mountains during summer and autumn rain events. Runoff from all of these flows traverse the surface above the Gunii Khooloi aquifer footprint.

Yang and Scuderi (2010) concluded that a cold and wetter climate prevailed between about 41 and 30 ka before present (BP), and generally drier conditions prevailed from 24 to 18 ka BP, in the Gobi Desert. From the beginning of the Holocene, wetter conditions prevailed until about 3.5 ka BP, interrupted by three shorter dry and cold reversals (Felauer et al. 2012). Most studies from northern China and southern Mongolia have concluded that in the late Holocene (from 3.5 ka BP to present), the climate changed to dry and warm conditions (Chen et al. 2016; Stauch 2016; Lehmkuhl et al. 2018).

At present, the region is climatically characterized by an extremely arid desert climate (Kottek et al. 2006) that is hot in the summer and cold in the winter. According to observational data from 1977 to 2016 at the Khanbogd meteorological station, the mean annual temperature is approximately 7.5 °C with an average monthly temperatures of –11.6 °C (January) and 24.9 °C (July). Precipitation is influenced by Asian monsoon, with 80% of precipitation from June to September. The mean annual rainfall measured at the Khanbogd station was only 98.8 mm (during the last 40 years), and potential evaporation is approximately 2,900 mm/year.

Geology and hydrogeology

The study area is situated in the Gurvansaikhan island arc terrane, which is one of the 44 terranes of Mongolia, in the center of the Central Asian Orogenic Belt (CAOB; Badarch et al. 2002). The CAOB is one of the largest orogens in the world (Fig. 2a), and it is surrounded by the Siberian and East European (Baltica) cratons to the north and the Tarim–North China (Sino–Korean) cratonic blocks to the south (Şengör et al. 1993; Jahn et al. 2000; Windley et al. 2007). The Gurvansaikhan terrane is formed through a complex process of island arc accretion and rifting of paleo–ocean. From the Neoproterozoic to the Silurian, the area was a Paleo–Asian Ocean margin setting (Fig. 2b), which was later subjected to subduction, rifting and finally accretion to form a continental environment until the Cretaceous period (Fig. 2c).

Fig. 2
figure 2

a Simplified tectonic map showing the location of the study area and the main tectonic components. The yellow area indicates the Central Asian Orogenic Belt (CAOB). Blue areas indicate neighboring Archean to Mesoproterozoic cratons (modified after Jahn et al. 2000). b–c Simplified evolutionary model of terrane accretion and amalgamation in Mongolia and neighboring regions (modified after Badarch et al. 2002). d Geological sketch map of the study area. Legend: (1) Khanbogd alkali granite pluton, (2) segments of rift zones overlapped with younger sediments, (3) Hercynides and complexes of active continental margin therein, (4) Mesozoic and Cenozoic Basins, and (5) faults (modified after Kovalenko et al. 2006)

Full size image

In terms of the geotectonic characteristics of the Gurvansaikhan terrane, geological units can be described: volcanic complex, alkaline granites and Cretaceous sedimentary basin. The uplift of the Ikh Shankh Mountain range, Duulga Mountain, Gurvan Kharaat and Ulaan Tolgoi ridges (volcanic complex) and Khanbogd Mountains (alkaline granite) occurred from the Devonian to Permian period and created the embryonic form of the Cretaceous sedimentary basin (Fig. 2d). As a result of this system activity, intensive denudation and erosion from the mountains led to a significant transfer of clastic material to the depressions of the Gunii Khooloi basin. This deposition, mainly from the Cretaceous to the Quaternary period, of thick alluvial, diluvia–proluvial sediments, and some aeolian and lacustrine deposits, led to the formation of the aquifer systems.

The Gunii Khooloi basin contains two main aquifer systems, a shallow streambed aquifer and a main deep confined aquifer (Gunii Khooloi aquifer, Fig. 3a). The shallow streambed aquifer system in this region is formed from highly permeable Quaternary alluvium, diluvial–proluvial sand, cobbles and gravel sediments with a total thickness from less than 8 to 20 m in some places. This aquifer system constitutes a single layer of unconfined aquifer and the water table ranges from 0.5 to 7 m below the ground surface (m bgs). The specific capacity is approximately 0.5–3.5 L/s m (Munkhbaatar et al. 2004). Most of the herders’ hand-dug wells are located within these dry river channels, which are mainly used for husbandry and domestic purposes.

Fig. 3
figure 3

a Hydrogeological map of the study area showing sampling locations (modified after Koldisheva et al. 1986 and Burenkhuu et al. 1995). b Piezometric map of Gunii Khooloi aquifer (Tuvdendorj et al. 2008)

Full size image

The Gunii Khooloi aquifer, in the Bayanshiree series, consists of Upper Cretaceous sedimentary formations and is distributed in the north of the Khanbogd Mountains, spreading out within the Gunii Khooloi basin. It is composed of a stratified sequence of sand, gravel, weakly consolidated sandstone, siltstone, gravelstone and conglomerate, which are interbedded with thin clay layers. The confining layers, Tsogt-Ovoo and Ulaan Gobi, are impermeable or weakly permeable layers formed by clayey sandstone and conglomerate with some lenses of clay, which indicates there are no vertical flows existing between the shallow and deep aquifers. The series has less thickness at the edges of the valley than at the central part of the deep aquifer, where it reaches a thickness of 160–260 m. Therefore, these formations make this Gunii Khooloi aquifer a confined aquifer. The Gunii Khooloi aquifer thickness and permeability increase continuously from the southwest to the northeast. Sandstone and conglomerate associated lithology of the Sainshand series are underlain by the deep aquifer and are gradually dipping in the same direction. The Gunii Khooloi aquifer is located at 115–287 m bgs, and the thickness is generally from 55 to 243 m. The total hydraulic head varies from 66 to 175 m, and the transmissivity ranges from 112 to 2,800 m2/day (Tuvdendorj et al. 2008). There are also two groundwater flow directions observed (Tuvdendorj et al. 2008), which are southwest to northeast and northwest to southeast (Fig. 3b).

Near the central part of the Gunii Khooloi aquifer (left side of point 7 in Fig. 4), where the basement rocks have been uplifted due to the geological setting of the basin, many thin clayey layers are found in the aquifer, which makes it more clayey overall. This narrow uplift may also cause discontinuity in the center of the basin, dividing the deep aquifer into two different parts with various hydrogeological parameters. The average thickness of the deep aquifer in the southwestern part is consequently 88 m, and 153 m in the northeastern part. In the southwestern part, the hydraulic head is 980 m asl and transmissivity is 327 m2/day. In contrast, in the northeastern part of the deep aquifer, the hydraulic head and transmissivity are 930 m asl and 957 m2/day, respectively (Tuvdendorj et al. 2008). This allows thickness and transmissivity of the aquifer at the western part to be lower than those in the eastern part; however, the recharge area of the deep aquifer is still unknown.

Fig. 4
figure 4

Hydrogeological cross-section of the Gunii Khooloi basin along transects I–I′ and II–II′ shown in Fig. 3 (modified after Tuvdendorj et al. 2008). The numbers on the surface indicate the sampling sites, as shown in Fig. 1

Full size image

Materials and methods

Wells, springs and precipitation in the Gunii Khooloi basin were sampled twice (in September, 2016 and in June, 2017) for chemical and isotope analysis. Major chemical composition and stable isotope samples were collected from late September to the beginning of October 2016. These included one precipitation sample (labeled ‘Rain’ in Table 1) and 49 groundwater samples (2 springs, 19 shallow hand-dug wells and 28 deep production wells). The rainfall sample was collected from Khanbogd soum center (sample No. 0 in Fig. 1).

Table 1 Field data, major ion concentrations, and stable isotopic compositions of precipitation and groundwater samples in the study area. Locations are shown later in the figures
Full size table

Two spring water samples (labeled ‘Spring’ in Table 1, samples 48 and 49 in Fig. 1) were collected from the Khanbogd Mountain. Nineteen shallow-aquifer groundwater samples were collected from hand-dug wells at depths of less than 10 m (wells used for domestic purposes, labeled ‘Shallow’ in Table 1). Most of the shallow groundwater wells are located in the recharge area of the Gunii Khooloi aquifer. The Gunii Khooloi aquifer water samples were obtained from 28 deep production wells (>245 m bgs, serving as water supply for the Oyu Tolgoi mine, labeled ‘Deep’ in Table 1). To insure consistent in situ conditions of representative samples, all production wells had been pumping continuously for several hours or days prior to sampling. All water taken from the shallow hand-dug wells was not stagnant inside the wells for long because the wells are used for livestock every day.

Water pH (DKK–TOA HM–30, three-point calibration), electrical conductivity (EC; DKK–TOA CM–31P; calibrated with 0.1 N sodium chloride standard solution) and dissolved oxygen (DO; DKK-TOA DO-31P; calibrated with sodium sulfite solution) were determined in the field. Alkalinity (sampled in 100-ml polyethylene bottles, free of bubbles, measured each sampling day, by 0.1-N-sulfuric-acid-titration method with digital titrator (Model 16900, HACH) at an accuracy of ±1%) were also determined in the field.

All sampling equipment was disposable and the sampling procedures for major ions and stable isotopes were performed following the Australian Geoscience method (Sundaram et al. 2009). Samples for laboratory analysis were filtered through 0.45-μm cellulose acetate membrane filters and only those analyzed for cations were stabilized by the addition of 3% HNO3 in the field. All water samples stored in 50-ml polypropylene, gamma-sterilized centrifuge tubes were free of bubbles and transported at temperatures below 4 °C.

Major anions (Cl–, SO42–, and NO3–) were determined using an ion-chromatography (881 Compact IC pro, Metrohm) with 1% accuracy, and major cation concentrations (Ca2+, Mg2+, Na+, and K+) were analyzed using an inductively coupled plasma-atomic emission spectrometer (ICP–AES; Thermo Scientific iCAP 6000 series) with 3% accuracy at the Graduate School of Environmental Studies, Tohoku University, Japan. The analytical error for the measurement of major ions was determined by calculating the ionic charge balance, which was within ±5% for all samples.

All water samples taken for stable isotopes (δ18O and δD) determination were analyzed using a liquid water isotope analyzer (Picarro, L2120-i) at the Groundwater Research Group in the Geological Survey of Japan (GSJ), National Institute of Advanced Industrial Science and Technology (AIST). The measured results are reported relative to the Vienna Standard Mean Ocean Water (VSMOW) in per mille (δ‰). The analytical precisions for δ18O and δD are within ±0.1 and ±1‰, respectively.

Fifteen tritium (3H) samples and 10 radiocarbon isotope samples were collected from the end of June 2017. Water samples for 3H analysis were collected in 1-L polyethylene bottles. The 3H compositions were measured using a liquid scintillation counter (Hitachi ALOCA, LSC-LB7) after electrolytic enrichment and samples were analyzed at GSJ. The results were expressed as tritium units (TU) and the analytical error and detection limit were 0.23 and 0.3 TU, respectively.

Samples for radiocarbon analysis were collected in 1 L PAN (BAREX) techno bottles and measured at Beta Analytic Inc, in Miami, Florida, USA. Analysis to determine the δ13C and 14C values of total inorganic carbon (TIC, ΣCO2) were measured using isotope ratio mass spectrometry (IRMS) and accelerator mass spectrometry (AMS) by the gas-strip method, respectively. The results for δ13C are expressed in the conventional δ-per mille (δ‰) notation referenced to the Vienna Pee Dee Belemnite (VPDB) international standard (Craig 1957) and results for 14C are reported as percent modern carbon (pMC). The analytical error was ±0.3‰ for δ13C and ±0.1 pMC for 14C.

Results

The results of hydrochemical, stable isotopic and tritium analyses of groundwater from the Gunii Khooloi basin are presented in Table 1, and sampling sites are shown in Figs. 1 and 3. Groundwater samples were separated into three representative groups according to the hydrogeological conditions of the Gunii Khooloi basin: springs, shallow groundwater (<10 m bgs, Quaternary streambed aquifer), and deep aquifer or Gunii Khooloi aquifer (>100 m bgs, Cretaceous confined aquifer).

During the investigations conducted in September–October 2016, the groundwater temperature ranged from 10.1 °C (sample 49) to 19.7 °C (sample 10). The mean temperature of the deep groundwater (16.8 °C) was higher than that of shallow groundwater (14.4 °C) and spring water (13.8 °C). The groundwater pH ranged from 7.3 to 8.3 with an average value of 7.8 that indicates a neutral to a slightly alkaline character. The mean pH of the spring water (8.0) was higher than that of the shallow (7.7) and deep groundwater (7.9). Electrical conductivity (EC) increases along the flow directions, with an average of 480 μS/cm in springs and 970 μS/cm in shallow groundwater, with the exception of three samples with EC 4,040 μS/cm (sample 37), 3,150 μS/cm (sample 43), and 2,000 μS/cm (sample 45). Deep groundwater samples exhibit high variability of EC, which ranges from 2,680 μS/cm (sample 8) to 6,500 μS/cm (sample 23) and an average value of 4,465 μS/cm. All groundwater samples contain dissolved oxygen (DO >2 mg/L).

At first inspection, the classification of water quality was conducted using a Piper trilinear diagram (Piper 1944), which is a graphical interpretation of the chemistry of water samples and is extensively used to evaluate the geochemical evolution of groundwater (O’Shea and Jankowski 2006). The shallow groundwater plots were dispersed in the Piper diagram (Fig. 5). In the case of the Gunii Khooloi deep aquifer, water samples were predominantly chloride type, whereas samples from the western part of the aquifer were more sulfate-enriched than the eastern part of the aquifer. The concentrations of major cations in the Gunii Khooloi basin were in the order of Na + K > Ca > Mg. Overall, the trilinear diagram shows that almost all water samples fall into zone 2, which indicates that groundwater in the Gunii Khooloi basin is dominated by alkalis, which exceed alkaline earths.

Fig. 5
figure 5

Piper diagram for the chemical analysis of groundwater collected from the Quaternary and Cretaceous aquifer in the Gunii Khooloi basin

Full size image

To examine the relationship between sampling sites and water quality, Stiff diagrams (Stiff 1951) were prepared and are shown on the map in Fig. 6. Spring water samples show Ca(HCO3)2 or Na(HCO3) type, while shallow groundwater is mainly characterized by Na(HCO3) and NaCl type; however, these diagrams are small, which indicates low EC. Although the springs and shallow groundwater are dispersed in the Piper diagram, it is clear that the dispersion reflects a slight change of ion concentrations due to the local conditions of the sampling sites. Only the shallow groundwater samples (Nos. 37, 43, and 45) are NaCl type with high EC; these are located at an outcrop of Cretaceous and Paleozoic-Mesozoic units. In contrast, all deep groundwater samples are classified as NaCl type, with different total dissolved solids (TDS) concentrations between the western and eastern parts. TDS in deep groundwater ranges from 1,660 to 4,460 mg/L. The average TDS concentrations in the southern and northern parts of the deep aquifer are 2,440 and 3,400 mg/L, respectively. Combined with the difference in the TDS, a detailed discussion is necessary to consider the groundwater flow from the western to eastern parts.

Fig. 6
figure 6

Shallow and deep groundwater quality represented by Stiff diagrams shown on the hydrogeological map of the study area (refer to Fig. 3 for the hydrogeological map legend). Orange Stiff diagrams indicate the shallow groundwater chemistry, while blue and red Stiff diagrams indicate the deep aquifer chemistry in the southwestern and northeastern part of the Gunii Khooloi aquifer, respectively

Full size image

The Na and Cl concentrations in groundwater show a strong linear relationship (Fig. 7). Such a relationship would be derived from the components of the paleo-ocean (fossil lake) that existed in the Gunii Khooloi basin in geologic age (section ‘Geology and hydrogeology’); however, the majority of samples fall above the seawater dilution line, indicating that Na has other sources. Discussion is necessary to consider for groundwater geochemical evolution.

Fig. 7
figure 7

Relationship between Na and Cl in groundwater samples in the study area

Full size image

The stable isotope ratio (δD and δ18O) can be a tracer of groundwater origin. Figure 8 shows the relationship between δ18O and δD for groundwater and precipitation in the study area. The values for all groundwater samples show a range from –11.6 to –8.1‰ for δ18O and from –97 to –73‰ for δD. The shallow groundwater samples are distributed somewhat sparsely, but seem to concentrate around –77‰ for δD and –10.5‰ for δ18O. The δD content of the deep groundwater tends to be lighter than that of the shallow groundwater (except sample Nos. 37, 44 and 45). The Gunii Khooloi aquifer can be divided into two different groups, depending on the location of sampling sites; the northeastern part of the aquifer (mean values of –95‰ δD and –11.1‰ δ18O) is more enriched than the southwestern part in respect to δD (mean values of –90‰ δD and –10‰ δ18O). The δ18O and δD values of the precipitation sample were 1.9 and 4‰, respectively.

Fig. 8
figure 8

Plot of δ18O and δD content for groundwater from the Gunii Khooloi basin, with plots of the GMWL, LMWL and EL

Full size image

Tritium (3H) is a radioactive isotope with a half-life of 12.43 years and has been used to distinguish groundwater recharged during the pre-bomb time from younger water (Fontes 1983; Clark and Fritz 1997). 3H activities (Table 1) for spring water and shallow groundwater samples range from 3.1 (sample 45) up to 28.2 TU (sample 36). No significant tritium content could be detected in the deep groundwater samples (<0.3 TU). By simple calculation, approximately 50 years are necessary for 3H to decrease below the quantitative limit from an initial 28.2 TU by radioactive decay.

In the deep groundwater samples, 14C activities (Table 2) were between 0.74 and 3.3 pMC, except for sample Nos. 22 (30.1 pMC) and 24 (14.9 pMC). The δ13C values mostly range from –6.4 to –8.8‰, except for sample No. 3 (–12.0‰).

Table 2 Radiocarbon results and corrected 14C age for the Gunii Khooloi aquifer
Full size table

Discussion

Groundwater origin

The local meteoric water line (LMWL) in the Gunii Khooloi area has not been established due to a lack of stable isotope data for precipitation. According to data from several meteorological stations around the study area since 1986, including Ulaanbaatar station, Baotou station, Yinchuan station and Zhangye station, taken from the International Atomic Energy Agency (IAEA) database (IAEA/WHO 2018), the δD and δ18O values in precipitation are highly variable, from −237‰ (Ulaanbaatar) to 5‰ (Yinchuan), and from −30.5‰ (Ulaanbaatar) to 3.9‰ (Yinchuan), respectively. Therefore, it is assumed that the LMWL for the Southern Gobi Region in Mongolia is δD = 7.29δ18O + 0.48 (Fig. 8), which is close to the global meteoric water line (GMWL) as determined by Craig (1961). From the weighted mean values obtained from these stations (–7.8‰ for δ18O and –56‰ for δD) and the precipitation data from the present study, this LMWL seems to be reasonable.

The springs and shallow groundwater samples (except sample Nos. 37, 40, 42, 44–47) fall below the LMWL, which is because the evaporation processes modify the original δD-δ18O relationship of the rainfall that results in slopes lower than 8, as reported in many arid regions (Gat 1980; Clark and Fritz 1997; Benjamin et al. 2004; Pang et al. 2011). The deuterium excess (Dansgaard 1964) shows a significant difference between the shallow and deep aquifers, ranging from –15.5 to +10.0‰. Some shallow groundwater samples have low deuterium excess values, which may be influenced by evaporation before or during infiltration. The degree of this influence could be the reason for the relatively high EC of the spring water.

It is noted that deep groundwater in the Gunii Khooloi basin has significantly lighter δD composition than the shallow groundwater (except for sample Nos. 37, 40, 42 and 44– 47). The depleted isotopic signatures are considered to be originated from paleowater (during the late Pleistocene; definition after Edmunds 2001), when the climate was colder and wetter than the present day. Such phenomena have been reported in previous studies on the Gobi Desert, Northern China, the Badain Jaran Desert (Gates et al. 2008), Dunhuang basin (Ma et al. 2013), the Ejina basin (Su et al. 2009), Heihe River basin (Zhu et al. 2008), Minqin basin (Edmunds et al. 2006) and Shiyang River basin (Ma et al. 2010). The clusters of deep groundwater seem to form a line with a regression equation of δD = 3.9δ18O – 51.1‰ (correlation coefficient: r = 0.89). Although the slope of 3.9 is indicative of an evaporation line, the relationship of the two clusters is still not clear. These isotopic compositions are typical for waters that have been affected by evaporation before recharging in a closed basin (Kamel et al. 2008) and in paleo lakes where the isotope composition remains in the pore water of the aquitard (Ortega-Guerrero et al. 1997).

Groundwater age

The main recharge of shallow groundwater in the Gunii Khooloi basin is via precipitation, which is also the main input source of 3H in groundwater. The global network of stations that monitor isotopes in precipitation does not include the Gunii Khooloi area and IAEA monitoring stations are far from the study area; however, 3H concentrations are available from 1986 to 2003 at Ulaanbaatar station (in Mongolia), Baotou station (in China), Yinchuan station (in China) and Zhangye station (in China). Detailed data are not available for these regions before 1986; thus, it is imperative to reconstruct missing data to infer what the historic distribution of 3H has been in precipitation in the study area. The earlier tritium data for the Baotou, Yinchuan and Zhangye stations were reconstructed for northern China (Li et al. 2008; Ma et al. 2008; Liu et al. 2014). These reconstructed 3H data are very informative and can be used to assist in the interpretation of the relative age dating in the study area.

According to the reconstructed data, the 3H values for shallow groundwater (15-30 TU) in the study area clearly characterizes these waters as natural background, derived from very new surface runoff in the stream network. For sample Nos. 47 and 42, lower 3H concentrations of 3.1 and 7.1 TU were measured. They can be classified as modern groundwater, as defined by Edmunds (2001), but seem somewhat older. By simple calculation, their ages are estimated as more than 25 years.

Groundwater in the Cretaceous deep aquifer with relatively low tritium content was supposed to be recharged before 1952, and therefore the 14C method was applied to these waters. To obtain adjusted 14C ages, it is necessary to know the initial 14C activity of recharge water. However, estimation of the initial 14C activity is a complicated task and several correction models have been applied to calculate it (Ingerson and Pearson 1964; Tamers 1967; Vogel 1970; Mook 1976; Fontes and Garnier 1979).

In the current study, soil CO2, organic matter and carbonate minerals are considered as the carbon sources in the groundwater. The Gunii Khooloi aquifer does not seem likely to contain much organic matter because of high dissolved oxygen, as shown in Table 1 (Winograd and Robertson 1982); therefore, the soil CO2 and carbonate minerals are assumed to be the primary sources of carbon in the groundwater. Detailed information about 14C in the unsaturated zone and in saturated part of the aquifer is not available; thus, it is for these reasons that the Tamers (1967) model and Ingerson and Pearson (1964) model were used to determine the 14C age. In both models, the 14C content in the initial water (in this case, recharging water in the past) is assumed to be formed by mixing of soil CO2 and carbon dissolved from carbonate minerals. The 14C content in the initial water is obtained from the concentrations of carbon species in the Tamers model, while the same is obtained from δ13C in the Ingerson and Pearson model. Calculations were performed using NETPATH (Plummer et al. 1994).

Both ages are very similar, which supports the preceding assumption. The average age of groundwater in the Cretaceous aquifer increases from approximately 22,500 years old (sample Nos. 2 and 3) to more than 33,500 years old (sample No. 10) along the groundwater flow direction up to the central part of the deep aquifer (Table 2). Samples collected from the central part of the Cretaceous aquifer are 27,000 years of age (sample Nos. 12 and 19). The age of the central part of the aquifer is highly influenced by younger groundwater flow which is distributed in the northeastern part of the aquifer. The average age of the groundwater in the northeastern part of the Cretaceous aquifer is much younger (except sample No. 25) than the southwestern part of the aquifer and increases from northwest to southwest. These results support that deep groundwater was recharged during the last glacial age. The deep groundwater shows a very old age, although the EC increases along the groundwater flow path.

The same tendency has been observed in semi-arid and arid basins around the world such as the northern Sahara sedimentary basin in Algeria (Guendouz et al. 2003), the Chahardouly alluvial basin in West Iran (Tizro and Voudouris 2008), as well as other groundwater basins in the Gobi Desert (northern China part) such as the Minqin Basin (Zhu et al. 2007), the Heihe River Basin (Yang et al. 2011), the Dunhuang Basin (Ma et al. 2013) and the Ejina Basin (Wang et al. 2013).

The 14C ages from sample Nos. 2 and 10 can be used to estimate the maximum groundwater flow velocity. The older sample (33,500 years) is approximately 12.2 km further along the flow path from the younger sample (22,500 years), which indicates an approximate flow velocity of 1.11 m/year. Assuming an average porosity of 0.3 (Tuvdendorj et al. 2008), the groundwater flux can be calculated as 0.33 m/year. On the other hand, using Darcy’s Law, the estimated groundwater velocity between sample Nos. 2 and 10, assuming hydraulic conductivity of 3.7 m/day and a hydraulic gradient of 3.3 × 10−4 (Tuvdendorj et al. 2008), is 1.48 m/year. The result is basically concordant with that calculated by 14C dating, and indicates that the Cretaceous deep confined aquifer flow is relatively slow and the water resources are essentially non-renewable in the Gunii Khooloi aquifer.

Conceptual model

Based on this first analysis of the isotopic and hydrogeochemical data, a conceptual model of groundwater recharge, flow and geochemical evolution is proposed to further understand the hydrogeological conditions in the Gunii Khooloi basin (Fig. 9). Quaternary unconsolidated sand, cobbles and gravel sediments form the shallow streambed aquifer, which is recharged by local precipitation in the Gunii Khooloi basin; however, in arid regions with low precipitation and high evaporation, the diffuse recharge is limited. In the Elgen and Gurvan Kharaat Mountain area, the isotopic compositions of the shallow groundwater tend to conform to patterns similar to the deep confined aquifer, which indicates that those areas could be the recharge area of the deep confined Cretaceous aquifer. In contrast, the shallow groundwater stable isotope ratio and tritium data in the Duulga Mountain area are different from the others. It should be noted that the shallow aquifer and deep aquifer are totally different aquifer systems in the Gunii Khooloi basin.

Fig. 9.
figure 9

Conceptual hydrogeological and hydrogeochemical model of the Gunii Khooloi basin.

Full size image

In the deep confined Cretaceous aquifer, the groundwater recharge mechanism is relatively complex. According to climatic conditions, it is difficult to reason that recharge occurs in the deep aquifer today, even though approximately 1.5 m/year of groundwater flow was calculated by 14C and the hydrogeological conditions. Therefore, the groundwater must be stagnant today but flowed during the last glacial age, which is supported by the light δD and δ18O content in most of the confined groundwater samples, and by their 14C age. Furthermore, the water quality distribution, 14C and groundwater flow direction indicate that groundwater flows from southwest to northeast in the southwestern part of the aquifer, while groundwater mainly comes from the northwest and flows into the southeastern part of the aquifer. The piezometric map (Fig. 3b) supports these results. Between these areas, groundwater is mixed and goes through the study area in the southeast direction. Therefore, according to geological evolution, the study area was of a Paleo-Asian Ocean margin setting (Fig 2b) and stable isotope results showed deep groundwater is affected by evaporation (paleowater). This is one of the pieces of evidence that deep groundwater originated from a fossil lake, non-renewable and with no genetic relationship to modern recharge.

Conclusions

Geochemical and isotopic tracers were used to understand the groundwater origin and hydrogeochemical characteristics of groundwater in the Gunii Khooloi basin, which is located in the Southern Gobi Region of Mongolia. Spring water in the Gunii Khooloi basin is classified as Ca-Na-HCO3 or Na-Ca-HCO3 type and shallow groundwater is characterized by Na-Ca-HCO3, Na-mixed and Na-Cl-SO4 water types. In contrast, deep groundwater is characterized by Na-Cl-SO4 to Na-Cl water types.

Based on the stable isotope compositions, groundwater in the Gunii Khooloi basin has a meteoric water origin. The deep confined aquifer is significantly depleted in heavy isotopes compared with the Quaternary shallow aquifer and modern rainfall. Thus, the stable isotope results clearly imply that there is no direct recharge connection between the deep confined aquifer and direct modern recharge. Deep groundwater is also mainly paleowater that was formed under colder and wetter climate conditions during the late Pleistocene and Holocene period. This phenomenon is confirmed by the 14C age (2.8–33.5 ka) and it is common in arid regions.

The results of this study have significant implications with respect to developing a conceptual model, to water resources management, and to the identification of groundwater origin and age in the Gunii Khooloi basin and other water-stressed basins in the Southern Gobi Region of Mongolia. These results will facilitate the management of water usage and the design groundwater monitoring systems, to confirm assessment of the groundwater resources in this region. Further studies within the Gunii Khooloi basin are currently in progress, including geochemical evolution and groundwater quality.