Synthesis of past isotope hydrology investigations in the area of Ljubljana, Slovenia Pregled preteklih izotopskih hidroloških raziskav na območju Ljubljane, Slovenija

Water isotope investigations are a powerful tool in water resources research as well as in understanding the impact that humans have on the water cycle. This paper reviews past hydrological investigations of the Ljubljansko polje and Ljubljansko barje aquifers that supply drinking water to the City of Ljubljana, with an emphasis on hydrogen, oxygen and carbon stable isotope ratios. Information about the methods used and results obtained are summarised, and the knowledge gaps identified. Overall, we identified 102 records published between 1976 and 2019. Among them, 41 reported stable isotope data of groundwater, surface water and precipitation and were further analysed. Isotope investigations of the Ljubljansko barje began in 1976, while groundwater and surface water investigations of the Ljubljansko polje and along the Sava River began as late as 1997. Isotope investigations of carbon started even later in 2003 in the Ljubljansko polje and in 2010 in the Ljubljansko barje. These investigations were performed predominantly in the frame of short-term groundwater research projects at five main wellfields and sites along the Sava River. Almost no large-scale, long-term stable isotope studies have been conducted. The exceptions include groundwater monitoring by the Union Brewery in Ljubljana (20032014) and precipitation in Ljubljana since 1981. Since 2011, more detailed surveys of the Ljubljansko barje were performed, and in 2018, the first extensive investigation started at wellfields and objects that form part of the domestic water supply system. Given the number of available studies, we felt that publishing all the numerical data and appropriate metadata would allow for a better understanding of the short and long-term dynamics of water circulation in the urban environment. In the future, systematic long-term approaches, including the appropriate use of isotopic techniques, are needed.


Introduction
As Bowen et al. (2019) states "Earth's water cycle links solid Earth, biological, and atmospheric systems, and it is both pivotal to the fundamental understanding of our planet and critical to our practical well-being." In nature, water is bound in different compartments of the hydrosphere (ice, groundwater, surface water, lakes, soil moisture reservoirs, oceans, and biomass), biosphere, lithosphere and the atmosphere, which form part of a global hydrological cycle. The rapid growth in population, coupled with an increased demand for water by agriculture and industry, are putting pressure on water resources (Mook, 2001). Although the impact that humans are having on the water cycle is indisputable, there is still a lot unknown about how water usage alters regional and global water budgets (Bowen et al., 2019). One of the prerequisites for efficient management of water resources is having reliable information about the quantity and the quality of the resource that is being exploited (Dansgaard, 1954;Craig, 1961).
Stable water isotopes ( 1 H, 2 H, 16 O, 17 O and 18 O) and carbon isotopes ( 12 C and 13 C) in the dissolved inorganic carbon (DIC) occur naturally. They can be measured using isotope-ratio mass spectrometry (dual-inlet or continuous-flow) (de Groot, 2004), laser spectroscopy (Wassenaar et al., 2018), or by spectrometric imaging methods (Bowen et al., 2019). An isotope abundance of an element is generally reported in ‰ (per mill = parts per thousand = 10 -3 ) deviations relative to the known isotope abundance of a standard, δ: (Gat, 1996): δ (‰) = (R sample /R standard −1) × 10 3 were R sample and R standard present isotope ratios ( 2 H/ 1 H, 18 O/ 16 O, 13 C/ 12 C, 15 N/ 14 N, 34 S/ 32 S) of a heavy isotope to a light isotope in a sample and an international standard, respectively. Because the numerical values obtained by this equation are small they are expressed in delta notation (δ). Delta values can be negative or positive numbers meaning that the isotope ratio of the sample is lower or higher relative to a standard (Gat, 1996;Meier-Augenstein & Schimmelmann, 2019).
Isotopes are an important tool for studying the water cycle and can be divided into two main categories: environmental isotopes (isotope variations in waters by natural processes) and artificial radioactive isotopes (radioactive isotopes that are injected into the system under investigation) (Kendall & Doctor, 2003). δ 18 O, δ 2 H and δ 13 C DIC values are important in different applications (Gat, 1996;Clark & Fritz, 1997;Ehleringer et al., 2008;Clark, 2015;Bowen et al., 2019): δ 18 O and δ 2 H can be used as conservative tracers if the isotope signature is unmodified within a study system, i.e., to identify water sources contributing to water sampled at a given place; δ 18 O, δ 2 H and δ 13 C DIC and their variations can enable the identification of important water and carbon cycle processes overlooked by other methods; δ 18 O and δ 2 H can link information on the history of water as it moves through the hydrological cycle. Isotope methods were introduced into catchment hydrology research to help scientists to understand better the geographical origin of water, recharge and discharge processes, biogeochemical processes and the sources and mechanisms of pollution (Clark & Fritz, 1997;Aggarwal et al., 2005;Bowen et al., 2005;Ehleringer et al., 2008;2016;Jameel et al., 2016;Du et al., 2019).
Concerns over climate change and the increasing demand for water in urban areas has focused research on water supplies and dynamics within the urban system in order to gain a better understanding of the connections between human populations, climate, and water extraction Zhao et al., 2017;Tipple et al., 2017).
Water circulates in nature differently than in urban environments, where the world's population is expected to increase to more than 60 % by 2050. Supplying large urban areas with high-quality drinking water and providing water resources in the long term is a major challenge Ehleringer et al., 2016). In Slovenia, drinking water supply is mainly based on groundwater (around 97 % of the drinking water supply is from groundwater resources) (Uhan & Krajnc, 2003) and in the capital city, Ljubljana, it provides an invaluable drinking water resource (Trček, 2017).
In Slovenia, only tritium and radon analyses are prescribed by drinking water legislation (Official Gazette, No. 74/15), however, if the parametric value for tritium is exceeded, it must be investigated to see if the cause is the presence of artificial radionuclides. Parametric values for specific basic ions, e.g., NO 3 -, SO 4 2and trace elements, e.g., Se, Sb, Pb, Ni, Fe, Cu, Cd, Al, As, B in drinking water have also been established (Official Gazette,Nos. 19/04,35/04,26/06,92/06,25/09,74/15,and 51/17), while the regular monitoring of stable isotopes of H, O in water and C and N in different compounds (e.g., HCO 3 -, NO 3 -) is not required by legislation. Despite quite a large number of isotope analyses performed in the past, to date, there has been no comprehensive research in the use of environmental isotopes in urban water management systems in Slovenia.
Here, we review and synthesize past research involving δ 18 O, δ 2 H and δ 13 C DIC to advance our understanding of the groundwater characteristics of the Ljubljana aquifers, which can be used as the basis for future investigations. We focus on work conducted over the past 40 years. The main aims of this review were the following: -make a synthesis of past urban hydrology investigations of the Ljubljansko polje and Ljubljansko barje aquifers with emphasis on the use of δ 18 O, δ 2 H and δ 13 C DIC until 2019; -collect information about sampling (location, time, type of sampling site) and the analytical methods used; -identify the main gaps in the previous investigations and propose future activities.
Two rivers bound the LP aquifer ( Fig. 1) -the Ljubljanica River to the south and the Sava River to the north Ogrinc et al., 2018). Because of the high velocities (10 m/day) and quite plunder groundwater flow (3-4 m 3 /s), the quality of groundwater is good Jamnik & Žitnik, 2020). Hydrological conditions in the area are characterized by strong interactions between surface water and groundwater and by the high velocities of groundwater flow and pollutant transport: that is, up to 20 m/ day (Andjelov et al., 2005;Janža et al., 2005). The LP is located in the eastern part of the Ljubljana basin (Ljubljanska kotlina). It was formed by tec- Fig. 1. Locations of the studied area with the main wellfields (Kleče, Hrastje, Brest, Jarški prod and Šentvid) and corresponding water supply areas in the Municipality of Ljubljana (wellfield Hrastje does not represent a unique water supply area). Source of topography: Geodetska uprava RS. tonic subsidence in the early Pleistocene together with the main neotectonic fault system that runs in an east-west direction. The basin is composed of Permian and Carboniferous slate claystone and sandstone (Žlebnik, 1971). The Pleistocene and Holocene sediments, accumulated by the Sava River, form highly permeable of partially conglomerated sand and gravel.
The thickness of these fluvial sediments increases towards the centre of the LP, where it even exceeds 100 m (Andjelov et al., 2005). The aquifer system has an intergranular porosity, and an unconfined groundwater table, located on 20-25 m below the surface  and can fluctuate up to 10 m (source archive JP VOKA SNAGA d.o.o.). The main recharge of the aquifer comes from infiltration of precipitation and the Sava River, which recharge the aquifer mainly in its north-western part and drains the eastern part of the LP. The LP is also recharged via lateral inflow from the LB multi-aquifer system in the south Vižintin et al., 2009;Vrzel et al., 2018) as well as from the Kamniško-Bistriško polje (Jamnik & Urbanc, 2000).
Groundwater is exploited at LP from four wellfields: Kleče, Hrastje, Jarški prod and Šentvid where drinking water is pumped from 16, 10, 3 and 3 wells, respectively (Fig. 1). Anthropogenic conditions of the aquifer are characterized by significant pressures of urbanization, industry, traffic, agriculture and old environmental burdens , which occur within the aquifer recharge area (Trček, 2017). To date, several different sources of pollutants have been detected and investigated. These include dispersed pollution sources where pollutants are consistently present (nitrates from agriculture and sewerage losses, new emerging contaminants in traces -pesticides from agriculture, plasticizers, corrosion and fire inhibitors, pharmaceuticals from sewage system losses (Jamnik et al., 2009) while others originate from past agricultural and industrial activities (atrazine, desethyl-atrazine, chromium (VI), trichloroethene, tetrachloroethene). Also, the characteristics of plumes and multipoint pollution contamination sources were recognized (Brilly et al., 2003;Karahodžič, 2005;Prestor et al., 2017).
The LB aquifer ( Fig. 1) extends from the southern part of Ljubljana to the Krimsko-Mokrško hills. The Barje is a depression with a stone bedrock that consists in the southern, western and central parts of Upper Triassic dolomite and Jurassic limestone, and in northern and eastern parts of Triassic and Permo-Carboniferous shaly mudstone, quartz sandstone and conglomerate, characterized by low hydraulic conductivity. The gravel fans are present on the borders of the basins (Mencej, 1988/89;Cerar & Urbanc, 2013). The basin was formed by a tectonic depression and filled by alluvial, marshy and lacustrine sediments during the Pleistocene and Holocene (Mencej, 1988/89). The Ljubljanica River contributes to groundwater storage as well as the Krimsko-Mokrško hills (ARSO, 2012;Cerar & Urbanc, 2013). The wellfield at Brest (Fig.  1) is an important source of drinking water for the southern part of the city of Ljubljana (Bračič Železnik & Globevnik, 2014). It consists of 13 wells of different depths (Bračič Železnik, 2016). Water resources in the area are under significant pressure, and environmental problems include water pollution, increasing water demand, flood and drought risk, reduction in retention capacity, decreasing groundwater levels and terrain subsidence (Bračič Železnik & Globevnik, 2014). However, desethyl-atrazine represents the most severe problem for the further development of the Brest water source (Prestor et al., 2017).

The Ljubljana drinking water supply system
The central Ljubljana water supply system consists of five water supply facilities with altogether active 44 wells and more than 1,100 km long water supply network supplying 330,000 users through 43,000 connections. Water supply network includes different objects (i.e., reservoirs, water treatment locations, pumping stations) (Jamnik & Žitnik, 2020). In the central system, some settlements are continuously supplied with drinking water from a single wellfield (water supply areas A, C, D and E in Fig. 1), and others from two or more wellfields (water supply areas F, G, H and I2 in Fig. 1), depending on water consumption and pressure conditions in the system. Wellfield Hrastje (B) does not represent a unique water supply area (Jamnik & Žitnik, 2020).
The water from the wells is pumped directly to consumers or a reservoir for the short-term, from where it is distributed to the users. Water disinfection devices are built-in into the system; however, water does not undergo technical treatments. It is only chlorinated occasionally. For the Brest wellfield UV disinfection is used (Jamnik & Žitnik, 2020).

Methods
Studies related to the characterization of aquifers important for the domestic water supply in the municipality of Ljubljana were reviewed, with a focus on those studies that used δ 18 O, δ 2 H, and δ 13 C DIC values for the characterization of water sources.

Study selection criteria
First, we considered articles and reports related to the water cycle and domestic water supply investigations for the LP and LB published from 1976 to the present (Fig. 2). In the scope of the review, a comprehensive search of journals was completed based on several keywords related to the Ljubljana aquifers (Ljubljana/Ljubljansko polje, Ljubljansko barje, Ljubljana groundwater, Ljubljana water, Ljubljana water supply). The search included all studies containing information about i) sampling, ii) analytical methods, iii) the parameters determined, and iv) isotope data.
In the second step, we focused on studies reporting the use of δ 18 O and δ 2 H to measure, describe or establish the characteristics of the LP and LB aquifers. Additionally, we also collected studies involving δ 13 C DIC . Articles on the modelling of LP and LB and other groundwater parameters, e.g., toxic metals in the groundwater and spring waters, electrical conductivity, and pharmaceuticals, and the quantity and quality conditions of groundwater in the Ljubljana aquifers were beyond the scope of this review (Fig. 2).

Search methods
The databases were searched for relevant literature published before November 2019 and included Google, Google Scholar, Science Direct, Co-operative Online Bibliographic System, and Service -COBISS. Included were national and international journals, conference papers, PhD and Master Theses, reporting data on δ 18 O, δ 2 H and δ 13 C DIC in an urban water system, precip-itation, and the Sava River. Also, the reference section of the articles was searched to identify additional sources. We also inspected the working reports for JP VOKA SNAGA d.o.o. available at Jožef Stefan Institute (JSI) including isotope data. Studies published in both Slovene and English were considered.
Information about i) sampling including location coordinates, type of sampling location (groundwater, spring water, precipitation, river) and sampling period; ii) the analytical methods used for δ 18 O, δ 2 H and δ 13 C DIC analysis, and iii) δ 18 O, δ 2 H and δ 13 C DIC data were collected and summarised.

Results and discussion
The initial combined search retrieved 102 records ( Fig. 2). After removing 41 non-relevant records, the 61 articles remaining were assessed for eligibility. Of these, 24 records were used to summarize site characteristics, while 41 records containing δ 18 O, δ 2 H and δ 13 C DIC data (Table 1) were reviewed in detail. Some articles were used in both categories. Information about sampling is summarised in subchapter Sampling, followed by Analytical methods used for determining δ 18 O, δ 2 H and δ 13 C DIC . Finally, a summary of the isotope research and the important findings relating to the Ljubljana aquifers is presented.

Sampling
Information collected about the sampling area, sampling locations and type of samples collected in different investigations for isotope analysis is presented in Table 1. Isotope investigations of groundwater were first performed in 1976 at LB (Breznik, 1984) while groundwater and surface water investigations at LP and on the Sava River in Tacen began in 1997 (Urbanc & Jamnik, 1998). The isotope composition of precipitation in Ljubljana has been regularly monitored since 1981 (Pezdič, 1999;Vreča et al., 2008).
To assess the usefulness of environmental isotopes, scientists have been performing systematic monitoring of the Ljubljana drinking water supply system since 2018. The first detailed sampling campaign was carried out between 06/09/18 and 29/11/18 at 103 locations; 41 wells in five water supply facilities, seven joint exits from the water pumping station, 22 reservoirs, two water treatment locations, 13 fountains, and 19 taps (see Table 1). In addition, samples were collected on the Sava River at Brod, Črnuče and Šentjakob (Vreča et al., 2019a;Vreča et al., 2019b). The first 24-hour experiment was performed in the basement of the main building at the Jožef Stefan Institute in Ljubljana with emphasis on the hourly isotope variability of tap water in April 2019 (Vreča et al., 2019d).

History of the stable isotope research in the catchment area of Ljubljana aquifers
Here we present a summary of the 41 records (Table 1) related to the past stable isotope investigations in the area of LP and LB aquifers. Articles usually report the use of δ 18 O and δ 2 H in water resources investigations; however, it is interesting, that the δ 13 C DIC was determined in only 13 records.

Ljubljansko barje
The first isotope investigations in the area of Ljubljana aquifers were performed in 1976 (Breznik, 1984), as part of the hydrological research into the Brest wellfield between 1974 and 1976. Water samples were collected at the LB aquifer, from the Iška River and other springs in the vicinity. No precise sampling locations with coordinates were reported, and no information was given about the collection of the samples or where the analyses were performed. They reported values for δ 18 O between -9.94 and -8.90 ‰ and -65.8 and -58.9 ‰ for δ 2 H. From the tritium isotope data, Breznik (1984) concluded that the recharge rate of the lower aquifer is very low.
Samples from the southern part of LB were collected in early spring and autumn in 1993. Nineteen sampling points for groundwater and river base flow measurements were established for the determination of groundwater recharge and storage capacity (Pezdič, 1998). Unfortunately, the sampling locations are presented only graphically, and the author gives no exact coordinates or location names. Precipitation was collected in Ljubljana for the determination of δ 18 O and δ 2 H values. Pezdič (1998) reported δ 18 O values of springs and surface river water of -9.65 and -8.82 ‰, while δ 2 H values ranged from -67.4 to -61.2 ‰. The weighted means of δ 18 O and δ 2 H in precipitation for the year 1993 were -8.07 ‰ and -55.6 ‰, respectively. The author concluded that the contribution of local precipitation was small and infrequent; however, local precipitation could recharge nearby aquifers (Pezdič, 1998).
After 1997, Urbanc & Jamnik (2002) performed more detailed investigations of the LB in which the chemical and isotope composition of groundwater was studied. Isotope investigations combined with hydrogeochemical methods were used to obtain hydrogeological data on the properties of water in individual aquifers: the Holocene aquifer and the upper and the lower Pleistocene aquifers. The authors, however, do not provide any sampling information or at which institute the analyses were conducted. Also, location names are shown only on maps. Surface water and groundwater in wells, piezometers and boreholes (Table 1) were sampled between November 1999 and February 2002. The authors report mean values for δ 18 O in surface waters and based on the isotope data, the mean altitude of individual water recharge areas (exact numbers were not provided). The δ 18 O values of groundwater in the Holocene aquifer were -8.9 to -8.6 ‰, -9.6 to -8.6 ‰ in the upper Pleistocene aquifer, and -9.5 to -9.2 ‰ in the lower Pleistocene aquifer. Again, values were mainly presented graphically, and numerical values were given only for the lower Pleistocene aquifer (Urbanc & Jamnik, 2002).
Since 2010, many isotope investigations at wellfield Brest were performed. In 2011, δ 18 O, δ 2 H and δ 13 C DIC values were determined in water samples collected during a pumping test from a 200 m deep well (VD Brest-3a) to determine the recharge dynamics, origin and age of groundwater in the dolomite. The investigation began on the 23/05/11 when a step-test was performed, followed by a one-month-long pumping test. In the third step, the rising of water was investigated. Testing finished on 24/06/11 (Brenčič, 2011). The δ 18 O, δ 2 H and δ 13 C DIC were also determined in seven wells at Brest and in one observation well (P-23/10). The values of δ 18 O ranged between -9.98 and -9.61 ‰ and δ 2 H between -64.9 and -61.1 ‰. δ 13 C DIC values were between -12.8 and -11.8 ‰. The isotope composition of springs near wellfield Brest was also determined. Isotope values were between -9.56 and -6.21 ‰ for δ 18 O, between -64.4 and -58.8 ‰ for δ 2 H and between -9.42 and -18.65 ‰ for δ 13 C DIC (Brenčič, 2011). By performing the pumping test, mixing of water from different aquifers, namely, shallow water from the upper Holocene aquifer and a lower Pleistocene aquifer in well VD Brest-3a, was confirmed. A certain amount of deep-water was also present; however, the exact amount was unknown, and its characteristics were not determined. The isotope composition of the water also varied during the pumping test, indicating that the fraction of water of different origin had changed (Brenčič, 2011;Vreča et al., 2011;Bračič Železnik et al., 2017). In 2013 (from 21/05/13 to 31/05/13), the pumping test was repeated in well VD Brest-3a. The δ 18 O, δ 2 H and δ 13 C DIC values ranged from -9.46 and -9.05 ‰, -65.9 and -63.4 ‰, and -14.5 and -12.3 ‰, respectively (Vreča et al., 2013;Bračič Železnik et al., 2017).
To conclude, the data shows a broad range of δ 18 O, δ 2 H, and δ 13 C DIC values in groundwater in the LB. Historically, isotope investigations were rare. In the last years, the δ 18 O, δ 2 H, and δ 13 C DIC are used more often but still sporadic. Also, different wells in the wellfield Brest yield different isotope compositions. This variation is because the depths of the wells are not consistent, and the groundwater is captured from different aquifers. Therefore, careful consideration about how to implement isotope techniques in the future is needed for better water resource management of the wellfield Brest.

Ljubljansko polje
According to available data, isotope investigations of groundwater from the LP were not performed until 1997. The first samples were collected between October 1997 and September 1998 at 13 pumping wells in the wellfields Kleče, Hrastje, Jarški prod and Šentvid (Urbanc & Jamnik, 1998). Samples were collected only for δ 18 O analysis. A more extensive set of observations (October 1997to September 1999 is presented by Andjelov et al. (2005). From this data, the authors estimated the proportion of locally infiltrated precipitation and water from the Sava River, but only reported the mean values of all measurements obtained during the sampling period for selected wells. Reported δ 18 O values in the groundwater were between -9.0 and -8.6 ‰ in Kleče (7 wells), -9.1 and -9.0 ‰ in Jarški prod (2 wells), and -8.9 and -8.8 ‰ in Hrastje (3 wells). In Šentvid, the mean value of several measurements from a single well was -8.8 ‰ (Urbanc & Jamnik, 1998). However, from the figures, it is possible to read the values for specific wells for the entire sampling period (Urbanc & Jamnik, 1998;Andjelov et al., 2005). At the same time, samples from the Sava River at Tacen were collected  those wells where the recharge zone extends under the city (Urbanc & Jamnik, 1998). In July and October 2003, the Institute for Public Health in Maribor collected samples at following locations: Yulon, Hrastje 1a, Kleče 17, GeoZS, Kleče 11, Šentvid 2A, Kleče 8a, Hrastje 3, Navje, Petrol-Šmartinska cesta, L.P. Vodovodna, HMZ Hrastje, for the δ 13 C DIC and alkalinity measurements. The δ 13 C DIC values were ranged from -14.7 to -12.2 ‰. The δ 13 C DIC results from LP were graphically presented in Kanduč (2006), together with δ 13 C DIC values of samples from the Sava River to indicate possible biogeochemical processes in the groundwater-river water system.
From March 2010 to December 2011, monthly samples were collected for δ 18 O and δ 2 H analyses from seven wells at three wellfields: Kleče, Hrastje, and Jarški prod, and from the Sava River at Šentjakob . Based on δ 18 O and δ 2 H results, the authors determined the proportion of the Sava River in groundwater resulting from periods of low and high precipitation in 2010 and 2011. Numerical values are reported in the Supplementary Data and are presented here as a box plot (Fig. 3). The authors found that both sources directly influence the groundwater: infiltration of local precipitation and recharge from the Sava River. Based on average δ 18 O and δ 2 H values, it was apparent that groundwater from Kleče 11, Hrastje 3, and Hrastje 8 contained only a low amount of the Sava River water (up to 14 %) and was mostly composed of recently infiltrated local precipitation. For comparison, a higher percentage of the Sava River water (up to 86 %) is present in the groundwater in wells Jarški prod 1, Jarški prod 3, Kleče 8 and Kleče 12. Findings were similar to that reported by Urbanc & Jamnik (1998).
More detailed investigations (from 2000 to 2014) in LP were performed in the area of Union Brewery where groundwater in Pleistocene fluvial sediments and the lower gravel aquifer is exploited by the Brewery (Trček 2005;2017). The Union Brewery's lysimeter was ideal for studying urban water infiltration and to make accurate measurements of water flow and water balance parameters. It consisted of 42 boreholes drilled into the right and left walls of the construction (Juren et al., 2003;Trček, 2005). As part of its sustainable groundwater management plan, extensive studies of groundwater flow and solute transport were performed from 2003 to 2014 to predict groundwater flow and contaminant transport through the unsaturated and saturated zone of the urban intergranular aquifer (Trček, 2017).
Actual stable isotope monitoring began in July 2003 (Trček, 2005) with the aim to obtain information about mixing processes and groundwater residence times in the unsaturated zone and to determine the risk of contamination of drinking water. From July 2003 to August 2004, monthly groundwater samples were collected, and δ 18 O and δ 2 H values determined. Trček (2005) reported δ 18 O groundwater values between -14.7 ‰ and -4.5 ‰. All other δ 18 O values were presented as boxplots, and no values for δ 2 H are reported. A synthesis of one-years' worth of data revealed two types of flow: lateral flow, which has an essential role in the protection of groundwater of the Pleistocene alluvial gravel aquifer, and vertical flow, which is the main factor controlling contaminant transport towards the saturated zone (Trček, 2005).
From July 2003 to June 2004 and from July 2004 to June 2005, δ 18 O and δ 2 H values in 16 observation wells (piezometers) were measured next to the Union Brewery. The mean values from a single sampling site for δ 18 O varied between -9.21 and -8.70 ‰ (Trček, 2006). During the same period (from July 2003 to June 2005) monthly oxygen isotope measurements of groundwater (lysimeter) ranged from -14.7 to -4.4 ‰, while the means of single sampling points were between -10.7 and -8 ‰ (Trček, 2005). In 2017, Trček published the results of the 2004 to 2014 investigation (Trček, 2017). Water samples were collected daily, weekly or at monthly intervals, although only seasonal monitoring was performed after 2010. Samples were collected from 18 observation points on the right side of the Union Brewery lysimeter, while precipitation was collected near the entrance to the lysimeter. The δ 18 O values in groundwater from 2004 to 2010 ranged from -16 to -6 ‰. In precipitation, δ 18 O values ranged from -18 to -3 ‰. Trček studied the weighted averages of the lysimeter water δ 18 O values for the period 2005-2009 to get a better insight into the lysimeter drainage system. Reported values varied between -9.82 and -7.62 ‰. Again, Trček emphasised the importance of lateral flow and that the goal for future investigations should be directed towards vertical transport studies of contaminant loads (Trček, 2017).
Isotope investigations of groundwater were also performed at the pumping station LMV-1 (located near the Kleče wellfield) from 2009 to 2011 (Mezga, 2014). The three-year sampling campaign covered three annual season cycles: groundwater at each sampling location was sampled twice, in spring (March-July) and autumn (August-November). The samples were collected as part of an extensive survey looking at the origin of groundwater in Slovenia. For the LMV-1, the authors reported mean values of δ 18 O of -8.59 ± 0.33 ‰, δ 2 H of -60.4 ± 0.6 ‰ and δ 13 C DIC of -12.7 ± 1.3 ‰ .

Ljubljansko polje and Ljubljansko barje simultaneous investigations
Simultaneous isotope investigations of both aquifers are rare. Cerar & Urbanc (2013) studied their interactions during two sampling campaigns in autumn 2010 and spring 2011. They aimed to obtain a better understanding of how the aquifers interact in order to improve a hydrogeological conceptual model of the aquifers. In total, they collected 138 samples at 69 locations from 28 wells from the five main wellfields, five industry wells, two private wells, 29 boreholes, and five samples of surface water. Based on the hydrogeological and the geographical position of the aquifers they divided LB into three areas: the northern part, middle part and southern part, including the area of Brest and Iški vršaj (Cerar & Urbanc, 2013). The δ 18 O in the groundwater of the northern part of LB varied between -9.0 and -8.6 ‰. Groundwater from this part of the aquifer is enriched in 18 O isotope compared to the other parts of the aquifers. This enrichment is due to the higher influence of local precipitation on the open aquifer. δ 18 O values in the middle part of the aquifer were from -10.0 to -9.1 ‰, while δ 18 O values in the southern part (including Brest and Iški vršaj) were -9.6 to -9.2 ‰. In their final report, Urbanc et al. (2012) report the range of δ 18 O values for groundwater from Brest to vary between -9.6 and -9.4 ‰ (tabulated values not given). For LP, δ 18 O values in Kleče wells varied from -9.1 to .0 ‰ in Jarški prod. Jamnik & Urbanc (2000) were the first to study the connections between LB and LP. They found that LP is partially recharged with groundwater from LB. However, Cerar & Urbanc, (2013) also showed that based on the hydrochemical composition (Ca/Mg molar ratio and HCO 3 concentration) of water, the contribution of groundwater from LB is of minor importance. The minimal contribution was detected near the boundary between the two aquifers. By measuring tritium activity, they classified groundwater in LP as "modern waters" with a residence time of up to 10 years, at the interface between the aquifers as "submodern waters" with a residence time of more than 50 years and in LB as "older waters" with residence time between 10 and 50 years.
However, increased tritium activities also indicated "bomb tritium" from nuclear experiments in the 1960s (Cerar & Urbanc 2013). Vrzel et al. (2018) confirmed "modern" water was mainly present in LP and also estimated, using the 3 H/ He method, that 10 % of groundwater in Kleče is very old, but additional analyses are needed for precise determinations.
In the period from March 2010 to October 2010 δ 13 C DIC was measured monthly along with alkalinity and pH at LP in the following wells: Hrastje 3, 8 (average -12.6 ‰, n = 12), Kleče 8,11,n = 22), Jarški prod 1, 3 (average -11.3 ‰, n = 13), and the Sava River at Dolsko (average -10.6 ‰, n = 7) (Kanduč, unpublished data). At LB sampling was performed only in June 2010 at wells Brest 1a, Brest 2a and Brest 4a with δ 13 C DIC values ranging from -11.3 ‰ to -10.8 ‰ (Kanduč, unpublished data). To our best knowledge, this was for the first time δ 13 C DIC was measured at LB. Vreča et al., (2019c) were the first to perform a stable isotope survey (June and July 2014) of tap water covering Slovenia according to our best knowledge. The authors determined δ 18 O and δ 2 H values in nine tap water samples collected in Ljubljana and its vicinity. The δ 18 O and δ 2 H values varied between -9.74 and -9.06 ‰, and between -65.2 and -60.1 ‰, respectively. The most negative values were in tap water from wellfield Brest and the most positive from Kleče.
A more detailed investigation within the Ljubljana water supply system started in 2018. The δ 18 O, δ 2 H and δ 13 C DIC values of all objects in the system (wells, joint exits from water pumping station, water reservoirs, water treatment locations, fountains and taps) ranged from -9.53 and -8.68 ‰, -63.6 and -57.8 ‰ and -15.3 and -9.38 ‰, respectively. Also, δ 2 H and δ 18 O values in samples from Šentvid were less negative, while samples from Brest had on average lower δ 13 C DIC values (Vreča et al., 2019a;2019b). The results for wells Kleče, Hrastje, Jarški prod and the Sava River are presented in Fig. 3 together with data from Vrzel et al., (2018). The values for 2018 are lower and less spread, which is a result of a shorter sampling period (September to November).
The first 24-hour analysis of tap water was performed from 9:00 on 24/04/19 until 9:00 on 25/04/19, with an emphasis on the hourly variability (Vreča et al., 2019d). The tap water was sampled in the basement of the main building of the JSI where water from two wellfields (Kleče and Brest) is mixed. The diurnal variations of δ 18 O, δ 2 H and δ 13 C DIC were small. However, 24hour differences in isotope and major and trace elemental composition suggest that the proportion of groundwater from Kleče and Brest water fields changed over 24 hours.
Based on the past investigations of LP and LB, especially 2018 -2019, the authors selected a systematic multi-analytical approach that started in 2020. Monthly monitoring of δ 18 O and δ 2 H and multi-element composition in groundwater in five wellfields (Kleče (4 wells), Brest (4 wells), Hrastje (2 wells), Jarški prod (2 wells), and Šentvid (1 well)) was established. Also, samples from the Sava River (Brod and Šentjakob) are collected on the same day and additional tap water investigations are planned.

The Sava River
Numerous isotope investigations have been performed along the Sava River basin (e.g., Kanduč, 2006;Ogrinc et al., 2008;Torkar et al., 2016;Vrzel et al., 2018;Ogrinc et al., 2018). However, only sampling locations close to Ljubljana (Tacen, Brod, Dolsko, Šentjakob and Črnuče) are relevant for this review (Table 1 and 2). Among these studies, ten reported δ 18 O, δ 2 H, δ 13 C DIC values ( Isotope investigations of the Sava River near Ljubljana began in October 1997, when the first sampling in Tacen was performed (Urbanc & Jamnik, 1998 The authors used data to provide information on hydrological flow paths and to estimate the water residence times. The data (Ogrinc et al., 2008) also forms part of the long-term the Global Network of Isotopes in Rivers database (GNIR; IAEA, 2020), managed by IAEA. The mean residence times at Tacen and Dolsko of 1.54 and 1.09 years, respectively, were estimated by using an exponential model in which precipitation inputs are assumed to mix rapidly with resident water. It was also observed that the Sava River responds quickly to precipitation, which is reflected in the isotope composition of the Sava River water (Ogrinc et al., 2008). Vrzel et al. (2018)   .

Conclusions
The use of isotopes to characterize water resources and to track the movement of water in the LP and LB over the past 40 years has significantly improved our understanding of groundwater quality and hydrological processes affecting its recharge and the distribution. Despite this, most isotope data are a result of intermittent shortterm studies, and only a few represent long-term monitoring programmes. From all of the analysed articles and reports, it is evident that limited sampling and coverage of monitoring of well networks presents a high risk of, e.g., not detecting contamination events .
The first δ 18 O and δ 2 H investigations of groundwater in the LB began in 1976, and only later in 1997 in LP. Also, in 1997 investigations at the Sava River in Tacen started. The first time δ 13 C DIC was systematically measured at LP was in 2003, while at LB it was only in 2010. Historically, isotope studies were performed in the LP; however, since 2011, isotope data are used more frequently, but still sporadically in the LB. These investigations mainly involve sampling from wells -sampling was most often performed in Kleče, while other objects in the water supply system were not well sampled. Five locations on the Sava River near Ljubljana were identified. Also, precipitation was monitored for δ 18 O and δ 2 H at six different locations.
To our knowledge, 102 relevant records were found and analysed; however, only 41 records published O, H and C isotope data and underwent a detailed review. The highest number of publications contained δ 18 O data (40 records), followed by δ 2 H (32 records), while δ 13 C DIC investigations were rarely implemented (13 records). Also, longterm systemic approach with more frequent (e.g., seasonal) monitoring of relevant environmental isotope tracers is missing. In the scope of this review, we would also like to point out that many investigations contain an insufficient description of sampling times and exact locations (missing coordinates), analytical methods, and reporting of raw data. In this regards, better use of supplementary material, which should include all appropriate metadata would be beneficial and necessary for proper comparison in time and space and would enable tracing isotope changes in water resources.
The first stable water isotope survey of tap water in the City of Ljubljana was performed in 2014. In order to assess the usefulness of environmental isotopes more systematically, monitoring has been performed on the drinking water supply system of Ljubljana since 2018.
Based on all of the results from previous investigations of LP and LB, monthly monitoring of δ 18 O and δ 2 H in groundwater in five water supply facilities was established in January 2020. Besides, also the Sava River is sampled at two locations monthly and additional more detail sampling of tap water is planned. The results will be used to prepare guidelines for future isotope monitoring that will provide a better overall understanding of water interactions of domestic supply important for water managers.