The Ararat Valley is a Cenozoic basin south of the Lesser Caucasus, shared by the countries Armenia and Turkey (Fig. 1a). Elevations within the valley range from 800 to 900 m asl, with the lowest elevations in the southeast and the highest in the northwest. The valley occupies an area of about 1,300 km2 (ME&A, Armenian branch 2014). The Araks River divides the Ararat Valley into a northern and a southern part. The northern part belongs to Armenia and contains the study area.
From a hydrogeological perspective, the Ararat Valley is a typical closed intermountain artesian basin. The major watercourses are the Araks River and its tributaries. Generally, the basin is characterized by a high heterogeneity of lithological and hydrogeological parameters. The strongly variable components, in particular, are (1) thickness and composition of the different lithological units, (2) hydraulic conductivities of the different aquifer domains, and (3) groundwater temperature and mineralization. This overall heterogeneity is due to the fact that the entire basin was divided by folding into five substructures (ME&A, Armenian branch 2014).
In spite of this heterogeneous structure, the subsurface succession can be subdivided into nine lithological units. These mainly consist of interbedded dense clays, gravels, sands, basalts, and andesite deposits. Four of the nine units are considered to be water-bearing. While the upper water-bearing unit 2 is unconfined, the lower three water-bearing units (4, 6, 8) are confined, due to a clay layer or sandy clay with an average thickness of 14 m between units 2 and 4 (Valder et al. 2018; Fig. 1b).
The upper unconfined aquifer has a mean thickness of 32 m and consists of alluvial deposits, primarily boulder or gravel deposits with coarse-grained sand, sandy clay, or clay. The three confined aquifers, units 4, 6 and 8, have average thicknesses of 27, 17, and 23 m, respectively. While unit 4 has a similar sedimentological setup as unit 2, units 6 and 8 are mainly characterized by highly fractured basalt, as well as gravel deposits with coarse grained-sand and boulders. The heterogeneous nature of the depositional environments is also reflected by the maximum thicknesses of the units, which are 89 m for unit 2, and 110, 148, and 119 m for units 4, 6 and 8, respectively (Valder et al. 2018). The reported hydraulic conductivities of these units also show appreciable variability. They range from about 1 m/day to about 150 m/day in the unconfined unit 2, and from about 0.9 m/day to about 115 m/day in the three confined units (ME&A, Armenian branch 2014).
The general groundwater flow direction coincides with the terrain inclination and is mostly directed towards the central axis of the valley, i.e., the Araks River. The groundwater generally flows northwest to southeast in the western part of the valley and from northeast to southwest in its western part. The paleo-valleys of the Araks tributaries, which are filled with interbedded clays, sands, gravels, basalts, and andesites (Valder et al. 2018), serve as major groundwater migration pathways between the recharge areas in the surrounding mountains (up to 5,137 m asl) and the centre of the valley. The Araks River itself is a losing stream in the western part of the valley, resulting in a deviating flow direction from south to north in unit 2 in this part of the valley. In the southeast, it is a gaining stream—(Valder et al. 2018; Fig. 1b; Figs. S1–S3 in the electronic supplementary material (ESM)).
The confined aquifers are artesian in the central part of the valley, however, due to the heavy groundwater abstraction many wells have ceased to flow and the area with artesian conditions substantially decreased (ME&A, Armenian branch 2014; Valder et al. 2018; see also Fig. 3).
Sampling and analyses
Two sampling campaigns were carried out in the Ararat Valley as part of this study, the first in October and the second in December 2019. In total, 28 wells were sampled (W01–W28; for detailed information, see Table S1 of the ESM). Out of these 28 wells, one was located in the unconfined unit 2 (W01) outside of the valley boundaries, three in the confined unit 4 (W15, W16, W18), and the rest in the confined unit 6. Well W02 is also situated outside the valley boundaries (Fig. 1a).
This study focused on revealing the age of groundwater sampled from all 28 wells, based on radioisotope tracer results. Revealing the water age, i.e., the groundwater residence time, is essential for evaluating the vulnerability of the tapped aquifer domains. Furthermore, water age-distribution information can be used to constrain groundwater recharge areas. The conclusions regarding the groundwater origin are backed by stable isotope data, physical-chemical parameters, and chemical characteristics of the water samples.
The groundwater samples taken during the two campaigns were analysed for the radionuclides radiosulphur (35S) and tritium (3H), the stable isotope signatures δ18O and δ2H, the field parameters temperature, electrical conductivity (EC), pH and O2, as well as a range of ions. The following sections introduce briefly the suitability of the used environmental tracers regarding the aim of the study.
Age of the groundwater
Although at least partly artesian, several of the wells in the central part of the valley were known to have elevated nitrate concentrations (Avetisyan and Tadevosyan 2017), which might indicate rather young waters. A suitable approach for investigating the presence of very young groundwater (subannual ages) is the use of 35S as an age tracer. 35S is continuously produced in the upper atmosphere by spallation of 40Ar. From there it is transferred with the rain to the earth’s surface and finally to the groundwater. As soon as the water enters the subsurface, its 35S activity concentration decreases with a half-life of 87.4 days, which makes 35S suitable for investigating subannual groundwater residence times (Schubert et al. 2019, 2020).
For 35S detection, 20-L water samples were taken at five selected wells located along the central axis of the valley (W24–W28). 35S was (together with total sulphate) extracted from the samples in a local laboratory by means of an adsorption resin. Subsequently, the resin samples were shipped to Germany and analysed for 35S in the UFZ radionuclide laboratory by low-background liquid scintillation counting (Quantulus GCT 6220; PerkinElmer, Inc., Waltham, MA, USA). The sample preparation and measurement process is described in detail in Schubert et al. (2019).
The radionuclide tritium (3H, half-life 12.32 years (a)) is used as a tracer in a range of settings, including column experiments (Stephens et al. 1998), and investigation of surface waters (Mundschenk and Krause 1991; Schmidt et al. 2020) and the unsaturated zone (Dincer et al. 1974; Jiménez-Martínez et al. 2013). Yet, its main application is the tracing of young groundwaters (Clark and Fritz 1997; Stadler et al. 2012). While the isotope is continuously formed naturally in the stratosphere by cosmic radiation, thermonuclear bomb tests conducted in the 1950s–1980s (mainly during the so-called “fallout peak” era between the late 1950s and 1963, i.e. prior to the Partial Test Ban Treaty in 1963) triggered a remarkable rise of 3H concentrations in the atmosphere. Since then, washout and decay have caused a drop of atmospheric 3H concentrations, which have now reached natural background levels in most regions (Clark and Fritz 1997; Schmidt et al. 2020).
Samples (1 L each) were distilled and 3H was electrolytically enriched. Concentrations were measured by low-background liquid scintillation counting (Quantulus GCT 6220; PerkinElmer, Inc., Waltham, MA, USA) with a detection limit of about 0.08 Bq/L at the laboratory for environmental isotopes (LfU) of the BfG. All results are conventionally reported in tritium units (TU).
Origin of the groundwater
Stable water isotopes (δ18O, δ2H)
The stable isotope ratios of oxygen and hydrogen in the water molecule, 18O/16O and 2H/1H, represent a powerful fingerprinting tool in hydrogeology, complementing water ages. Generally, applications include the assessment of water sources (Schmidt et al. 2011; Stadler et al. 2012), (paleo-)climatic effects (Stadler et al. 2012), recharge elevations (Koeniger et al. 2017), biases of recharge towards a certain season (Jasechko et al. 2014) or precipitation type (Müller et al. 2020), or the estimation of evaporative water losses (Skrzypek et al. 2015).
Stable isotope ratios were measured by Laser Cavity Ring-Down Spectroscopy (L2130-I; Picarro, Santa Clara, CA, USA) at the LfU (BfG). Results are expressed in per mil (‰) using the conventional δ notation relative to Vienna Standard Mean Ocean Water (V-SMOW). The external precisions (±1 σ), determined by repeated analyses of a control sample, were ± 0.15 and ± 0.6‰, for δ18O and δ2H, respectively.
During sampling, field parameters (temperature, pH, electrical conductivity, oxygen content) were measured with field probes (HQ40d, Hach, Loveland, CO, USA). Bicarbonate concentrations were also determined on-site with a field kit (HI775, Hanna Instruments, Woonsocket, RI, USA). Furthermore, two samples were collected at each well in 50-ml polyethylene bottles for ion analyses. These samples were filtered using 0.45 μm membrane filters. Samples for cation analyses were acidified by adding concentrated HNO3. All major ions were analysed by ion chromatography (882 compact ICplus equipped with a Metrosep A Supp 5-250 column for anions and a Metrosep C 4-250 column for cations, Metrohm, Herisau, Switzerland) at the Institute of Applied Geosciences of the Technical University of Darmstadt. The precision of the applied method, expressed as relative standard deviation, is better than ±3%. The charge balance errors were typically within ±5%.