Introduction

Global warming is characterized by the frequent occurrence of floods and droughts, which can have profound consequences for water resources [1, 2]. The quantity of groundwater remains constant despite climate change [3]; therefore, groundwater is important for mid- to long-range water resource planning for stable water supplement to cope with water scarcity [4, 5].

In 2018, groundwater in Korea was mainly utilized for domestic (1229 million m3/year; 42.5% of total usage) and agricultural (1485 million m3/year; 51.7%) applications [6]. Groundwater consumption in Korea is highest in Daejeon, Sejong, and Chungcheongnam-do, excluding Jeju Island [6]. Chungcheong-do, which encompasses Daejeon, Sejong, Chungcheongnam-do, and Chungcheongbuk-do, is located in the midwestern region of South Korea, facing the Yellow Sea. The Geum River is a major river in Korea that flows through Chungcheong-do; its estimated groundwater recharge rate reaches 3298 million m3/year, with about 91% (3012 million m3/year) occurring in the Chungcheong-do area [7]. Nevertheless, the accumulated summer (June–August) precipitation in this area was 365 mm in 2019, which was less that in 2018 (588 mm), as well as the overall average for Korea (493 mm) [8]. Salt accumulation and nitrate contamination are increasing in this region due to the development of small and medium cities, agricultural activities, and livestock facilities. Seawater intrusion by excessive pumping of groundwater in coastal areas also contributes to the degradation of groundwater quality in midwestern South Korea [6, 7, 9].

In this study, we used contamination indicators including ionic and isotopic species to identify groundwater sources. These include nitrate nitrogen (NO3-N) and chlorine (Cl) as anthropogenic inputs, fluoride (F–) as a geologic input, chloride (Cl–) and bromide (Br–) as seawater intrusion inputs. NO3-N is present in natural water at very low concentrations, where concentrations > 3 mg/L are interpreted as having artificial origin [10, 11]. Potential sources of NO3-N leaching include animal mature, domestic sewage, agricultural operations, and inflow from air pollution. The high solubility of this species facilitates inflow and migration into soil water and groundwater [12, 13]. High concentrations of NO3-N (> 10 mg/L) in drinking water can cause methemoglobinemia, which leads to acute toxic responses to NO2 exposure via oxygen transport [14]. Therefore, the NO3-N drinking water standard for Korea has been established at 10 mg/L (~ 44 mg/L as NO3) [15].

F– is generally present in natural water at concentrations < 1 mg/L; it is usually derived from rock minerals such as fluorite, apatite, amphiboles, and micas [10]. Low concentrations (< 0.7 mg/L) of F– in natural water can prevent tooth decay [16]. However, excessive F– can cause dental fluorosis and bone weakness; therefore, F– levels should be maintained within 1.5 mg/L [14].

Cl– and Br– concentrations can be influenced by seawater intrusion, domestic sewage, industrial waste water, de-icing salt dissolution, and bedrock mineral interactions. These ions represent stable species in the natural environment because they do not adsorb to other minerals or organic matter, or form non-soluble precipitations according to changes in the oxidation–reduction environment [17].

Tritium (3H; t1/2 = 12.43 years, Emax = 18.6 keV) is a natural radioactive nuclide that is produced by cosmic ray bombardment of nitrogen and deuterium in the upper atmosphere.3H enters the hydrological cycle in the form of tritiated water molecules (HTO), where it can be used as a conservative groundwater tracer as a component of the water molecule. When water infiltrates groundwater, it becomes isolated from the atmospheric 3H source and its 3H concentration decreases over time due to radioactive decay. Therefore, 3H is used as a groundwater dating tool for ages up to 100 years [18, 19].

Decreases in annual precipitation and active groundwater pumping can deplete water resources and lower its quality. For the sustainable use and efficient management of groundwater in midwestern South Korea, it is necessary to understand the recharge characteristics and spatial distribution of groundwater hydrochemistry. Therefore, we investigated the coastal and inland groundwater hydrochemistry of the Geum River basin regional water system. Specifically, we assessed the spatial distribution of groundwater hydrochemistry in the Chungcheong-do region, determined the type and sources of groundwater contaminants, and identified promising candidate sites for large-scale groundwater supply to cope with potential water shortages.

Site description

The average annual temperature of South Korea for the past 30 years has been 10.9 °C, and the average annual precipitation is 1295 mm, indicating a temperate climate [8]. Our study site (35°58′33.71″–37°03′40.10″N, 126°07′13.25″–128°04′06.03″E; Fig. 1) covers Sejong, Chungcheongnam-do, and Chungcheongbuk-do, but excludes Daejeon because it contains several industrial complexes, has a high population density, and therefore cannot meet our study objectives.

Fig. 1
figure 1

Map of nine hydrogeological units and sampling points for a groundwater and b isotope analyses. Left panels show the geomorphological location of the study area

Full size image

The area of the study site is approximately 12,294 km2 and is bounded to the east by the southern Sobaek mountain range and to the west by the Yellow Sea. Thus, the topography of the site slopes downward from the east to the west, and agricultural fields and urban districts are mainly distributed in the west, whereas forests are distributed in the east. The geology of the study site is mainly composed of the Archean–Proteorozoic Gyeonggi metamorphic complex, the Sobaegsan metamorphic complex, and Jurassic Daebo granitoids [20], in connection with the Mesozoic rift, as well as rocks from the Paleozoic Okcheon Supergroup (from Chungcheongbuk-do to Daejeon) and Great Limestone series (Chungcheongbuk-do). Large-scale plutonic rock has developed in the parallel Okcheon Belt, via the Jurassic Daebo orogeny. Because we examined groundwater over a large area in this study, we have simplified the bedrock geology to nine hydrogeological units: crystalline gneiss, intruded Cretaceous granite, intruded Jurassic granite, unconsolidated detrital deposits, metasedimentary rock (with preserved structural geometry), metasedimentary rock, (with depleted structural geometry), nonporous volcanic rock, clastic sedimentary rock, and carbonate rock.

Materials and methods

Field sampling and ion analyses

Most of the groundwater sampling sites were located adjacent to rice paddies or dry farms during the agricultural off-season to avoid direct contamination and observe background hydrochemistry. We collected 128 groundwater samples for hydrochemical identification (Table 4) and analyzed 47 samples for isotope distribution in April 2020 (Table 1 and Fig. 8).

Table 1 Values of groundwater pH, depth, and electrical conductivity (EC) measured at the study sites and isotope values including δ18O, δ2H, deuterium excess, and tritium (3H) concentrations of collected samples
Full size table

According to the field data, which included factors such as groundwater level and pumping rate, we selected four candidate sites to identify a promising site for a large-scale groundwater system: KS, KJ, NS, and BO. Hydrochemical, stable isotope (δ18O and δ2H), and 3H analyses were conducted to determine the groundwater recharge properties of these sites (Table 1).

In situ pH and electrical conductivity (EC) measurements were performed using a multi-parameter portable meter (WTW Multi 3620 IDS, Fisher Scientific, Sweden) and the uncertainty range of EC value was below ± 0. 5. Cation (Ca, K, Mg, and Na), anion (F, Cl, Br, NO3, and SO4), stable isotope (18O and 2H), and radioisotope (3H) analyses were performed in the laboratory. Alkalinity titration was performed to measure HCO3 ion levels using 0.02 N hydrochloric acid and a Rondolino DL50 Titroprocessor (Mettler Toledo, Australia). All water samples were filtered through a 0.45 µm pore size cellulose acetate membrane (Advantec), stored in polypropylene bottles, and kept in a refrigerator at 4 °C for later chemical analyses. Cation samples were acidified with concentrated nitric acid to pH < 2. Cation and anion concentrations were analyzed using inductively coupled plasma optical emission spectrometry (ICP–OES, OPTIMA 7300DV, PerkinElmer, USA) and ion chromatography (Dionex Aquion, Thermo Scientific, USA), respectively. The sample ion concentration charge balance error was within ± 10%.

Isotope analyses

Oxygen (18O) and hydrogen (2H) isotopes were analyzed using cavity ring-down spectroscopy (L2140-i for isotopic H2O; Picarro, USA) and the results were compared to those of standard mean ocean water (V-SMOW). The stable isotope data of 18O and 2H are reported with the δ in ‰ deviation relative to V-SMOW and the uncertainties of isotope ratios were generally below ± 0. 1‰ for δ18O and ± 0.5‰ for δ2H.

Two local meteoric water regression lines (LMWLs) were plotted to describe the isotopic data for different seasons: δD = 7.93δ18O + 8.11 for summer precipitation and δD = 7.79δ18O + 18.39 for spring, autumn, and winter precipitation [21]. The slope and intercept of the regression line for wet season precipitation were virtually identical to those of the empirical global meteoric water line (GMWL) [22]:

$$\delta {\text{D}} = 8 \, \delta^{18} {\text{O}} + 10.$$

For 3H analyses, groundwater samples were filtered through a 0.45 µm nuclear pore filter and distilled to remove impurities, followed by electrolytic enrichment as previously described [23]. The enriched sample was distilled and a 10 mL water sample was mixed with a 10 mL cocktail (Ultima Gold LLT, PerkinElmer) consisting of a mixture of scintillator, solvent, and surfactant. Subsequent 3H measurements were performed using a low-level liquid scintillation counter (Wallac Quantulus 1220, PerkinElmer).

Geographic information system (GIS) database

A map of East Asian administrative boundaries (National Spatial Data Infrastructure Portal; http://data.nsdi.go.kr/dataset/20171206ds00003) was used as the GIS base map, with rivers (K-water; http://www.wamis.go.kr:8081/WaterMap2013/watermap.aspx?paramTab=searchmap), hydrogeological units, and electrical conductivity (EC) as layers. Open source data were obtained from Arcgis Online. Hydrogeological units were classified according to geological age, lithofacies, topography, and porosity (Table 2) [24]. EC spatial distribution was analyzed using inverse distance weighting (IDW) using the IDW spatial analyst tool in the ArcGIS ver. 10.7 software. Interpolation was performed using a distance index of 2, influence radius of 12 wells per central well, and an output raster cell size of 30 m × 30 m.

Table 2 Classification of hydrogeological units (
Full size table

Results and discussion

Regional hydrochemistry distribution

Most of the groundwater samples in the Chungcheong-do region were distributed in fields 3, 5, and 6 of the Durov diagram (Fig. 2). The pristine groundwater in this region is characterized by the Ca-HCO3 water type, and is most influenced by simple dissolution of geologic materials or mixed natural/anthropogenic inputs (Table 5) [25]. Annual groundwater usage in this area has been stable since 2010, at about 39 million m3/year, but decreased significantly in 2018 to about 28 million m3/year [26]. Based on these data, seawater intrusion decreased with the change in pumping, and fresh groundwater increased the depth of the coastal aquifer causing reverse cation exchange. However, the amount of total dissolved solids (TDS) was highest in this region among all groundwater samples (Fig. 2); therefore, this site continues to experience seawater intrusion or effects from other pollutants.

Fig. 2
figure 2

Durov diagram of hydrochemical facies of groundwater samples collected in the study area and major interaction processes (red arrows)

Full size image

Groundwater samples from fields 8 and 9 (G14 and G65 of Table 4, respectively) were collected from wells that were 100 and 210 m deep, respectively. These samples had lower Na+ concentrations than the other samples, but the highest Na+  + K+ contribution, according to the major ion ratio data shown in Fig. 3A. The high concentrations of F– indicated alkaline, Na–HCO3 type water, developed through the dissolution of biotite from Mesozoic granitoids and Precambrian granitic gneiss [27].

Fig. 3
figure 3

Spatial distribution of (A) major ion species and electrical conductivity (EC) in the study area

Full size image

The six sampling sites with the highest EC values (> 625 μS/cm) are shown in Fig. 3B; half of these sites are distributed inland, and half in coastal areas. These EC values were affected by seawater intrusion as well as input from other pollutants. Site A (G32 of Table 4) is bounded by the sea on three sides, and is about 200 m from a livestock facility; the groundwater sample from this site had the highest EC value (3390 μS/cm), likely due to input by seawater and livestock manure. Site D (G82 of Table 4), which had the highest population density (1,057 people/km2) and an urban area ratio of 42% [28, 29] had the second highest EC value. Sites B and C (G74 and G60 of Table 4, respectively) are located in coastal areas with dense facilities (65 million in 2010) and very high groundwater usage (44 million m3/year in 2010) due to water supply shortages, with a 69.6% penetration rate in 2014 [28]. About 91% of all area in Sites E and F is non-urban (G88 and G65 of Table 4, respectively), with low population density (22 people/km2) [28]; however, these sites had high groundwater EC and high F– concentrations. The pristine groundwater at these sites likely reflects geologic characteristics.

Site-specific hydrochemistry

Bubble plots of each sampling site showing specific hydrochemical indicators are shown in Fig. 4. Most groundwater in the study area showed F– concentrations < 1.5 mg/L, which is the drinking water standard [15], except at two sites (Fig. 4a). The groundwater samples with the highest (6.86 mg/L, G65 plotted on the field 9) and second highest (3.36 mg/L, G92 plotted on the field 8) F– concentration also had high Na-HCO3 concentrations. The geology of these areas mainly comprises metasedimentary rocks from the Paleozoic Okcheon supergroup and Granitoid intrusions of Mesozoic age [20]. In these regions, F– was likely released through the reaction of groundwater with rock-forming minerals in granite, which promoted calcite precipitation and fluorite dissolution at high pH [30, 31].

Fig. 4
figure 4

Spatial distribution of natural and anthropogenic water components determined using a fluoride (F–), b chloride (Cl–), and c nitrate nitrogen (NO3-N) as indicators

Full size image

Cl– concentrations varied greatly along the coast, and were highest at an inland sampling site with dense residential population and industrial facilities (Fig. 4b). The average background concentration of Cl– in groundwater of South Korea is about 5–10 mg/L [32, 33]; about 27.3% of our samples (35 sites) were within the background range. About 7% of samples (nine sites) exceeded 50 mg/L Cl–, reflecting water pollutants from urban areas [34], and one site (G32 of Table 4) exceeded the World Health Organization (WHO) drinking water standard of 250 mg/L, at which a salty taste is detected [14]. NO3-N concentrations exceeded 3 mg/L at about 65% of sites (83 sites) (Fig. 4C) and exceeded the drinking water standard at about 16% of sites (20 sites), presumably under the influence of anthropogenic sources such as livestock facilities, which are found within a 500 m radius of 55.5% of all groundwater sampling sites, and residential and industrial facilities (23.4%). The detailed values are presented on the Table 4.

Source estimation using ion indicators

A number of our samples were identified as endmembers, including those indicated in the lower left and upper right of Fig. 5a, which have a background concentration of Cl– [32, 33, 35] and the highest natural (seawater intrusion) and anthropogenic (domestic/industrial sewage, animal manure) contamination among our samples. Cl– concentration and EC were positively correlated (R2 = 0.7932) among the groundwater samples; thus, this pair of samples are thought to have been similarly affected by contamination.

Fig. 5
figure 5

Relationships between a Cl– concentration and EC, b Cl concentration and Cl/Br molar ratio, and c NO3-N and Cl–. Axes are presented on a logarithmic scale

Full size image

Groundwater contamination factors can be identified using the relationship between Cl/Br and Cl in the range above background concentration (10 mg/L Cl–). Br– is introduced into groundwater aquifers as a by-product of industrial activity or decomposition of organic matter in the soil layer [36]. Samples from agricultural areas with Cl–concentrations > 100 mg/L (Fig. 5b) can be interpreted as pristine groundwater mixing with halite, sewage, animal waste (manure), basin brine, and landfill leachate [35]. Samples with Cl–concentrations > 10 mg/L and with Cl/Br molar ratios greater than that of seawater (655 ± 4) indicate seawater intrusion of a coastal area.

The correlation between Cl– and NO3-N concentrations (Fig. 5c) can be used to distinguish contamination sources for a range of pristine groundwater and that contaminated by synthetic fertilizer, road de-icer, manure, septic sources, and other mixed sources [37]. The thresholds of NO3– contamination for South Korea are 3.0 mg/L (0.68 mg/L NO3-N) for bedrock groundwater and 5.5 mg/L (1.24 mg/L NO3-N) for alluvial groundwater [38]. The background concentration of Cl– in shallow gravel aquifers is typically 10–20 mg/L in the USA and other countries [39]. About 14% of samples (18 sites) showed a background level of NO3– (< 5.5 mg/L), and only about 5.5% (seven sites) showed background levels of both NO3– and Cl– (< 10 mg/L), which indicates that most Geum River basin groundwater samples were affected by natural or anthropogenic pollutants.

Regional isotope distribution

Oxygen–hydrogen isotope pairs (δ18O–δ2H) in samples collected in the study area were correlated with altitude and latitude, and a stronger positive correlation with altitude was detected (Fig. 6a, red arrow). Samples from higher-altitude regions showed depleted isotopic composition, apparently affected by winter precipitation, based on the winter LMWL. Samples from lower-altitude regions showed enriched isotopic composition and were affected mainly by precipitation recharge in summer. Some groundwater appears to have undergone evaporation (Fig. 6a, black arrow); these samples showed no significant relationship with contamination by F, Cl, or NO3-N. The sample obtained closest to the coast (i.e., lowest altitude) showed the highest level of contamination in the study area and the greatest depletion of δ18O–δ2H, likely due to mixing with domestic sewage and livestock manure.

Fig. 6
figure 6

a Stable oxygen–hydrogen isotope composition of groundwater samples collected in the study area. LMWL, local meteoric water line [21]; GMWL, global meteoric water line [22]. b Relationship between Cl– concentration and isotopic distribution of groundwater samples; reference line indicates the background Cl– concentration for Korea, according to Cho et al. [32] and Park et al. [33]

Full size image

The Cl– concentration of fresh groundwater was at the background level, at about 0.28 mEq/L (~ 10 mg/L Cl–) [32, 33]; thus, about 27.3% of samples (35 sites) had background Cl– levels. As the groundwater Cl– concentration increased in this area, δ18O enrichment was observed (Fig. 6b), likely due to the influence of evapotranspiration during recharge [40]. Mixing due to agricultural activity, domestic sewage, and livestock manure can increase Cl– and NO3-N concentrations and affect δ18O enrichment [41]. The highest Cl– concentration (26.5 mEq/L, G32 of Table 4) was associated with depleted δ18O, likely due to proximity to a livestock facility rather than seawater intrusion.

The distribution of 3H concentrations in Geum River basin groundwater is shown in Table 3. Similar 3H residence times were observed among regions, with averages of 2.19 ± 0.59 TU (1.03–2.88 TU) at KS, 2.15 ± 0.55 TU (0.90–2.91 TU) at KJ, 2.38 ± 1.08 TU (1.01–4.87 TU) at NS, and 2.48 ± 0.69 TU (1.14–3.60 TU) at BO. The 3H concentration distribution was narrow at KJ but broad at NS; values were higher at BO than in other regions (Fig. 7). Therefore, BO groundwater appears to be younger, with faster recharge times than observed in other regions [42, 43]. Values of 3H concentration < 0.5 TU indicate recharge prior to the 1950s, indicating older groundwater, whereas those > 2 TU indicate younger groundwater. Most 3H values from the four candidate sites examined in this study are located between 1.5 to 3 TU, indicating recent recharge or slight mixing with old groundwater at these sites.

Table 3 Tritium (3H) concentration of four candidate regions for large-scale groundwater supply: KS, KJ, NS, and BO
Full size table
Fig. 7
figure 7

Tritium (3H) concentrations in the four candidate regions for a large-scale groundwater supply system. Left panel shows the target area (upper) and sampling sites (lower) used for 3H analyses

Full size image

Conclusion

We investigated regional groundwater hydrochemistry and isotopic distribution in 128 groundwater samples collected in Chungcheong-do (excluding Daejeon), which is highly dependent on groundwater for domestic and agricultural usage. Most of the sampling sites are located on rice paddies or dry farms, but are within a 500 m radius of livestock, residential, or industrial facilities. In the study area, groundwater is mainly recharged through precipitation and water is mainly of the Ca-HCO3 type due to simple dissolution and/or mixing processes.

About 65% (83 sites) of all samples showed NO3-N concentrations > 3 mg/L, indicating that the sampling aquifer was influenced by anthropogenic inputs, and about 16% (20 sites) exceeded the NO3-N drinking water standard of Korea. Cl– concentrations > 100 mg/L were observed at six sites in inland or coastal areas; these clearly indicated contaminant mixing. High F– concentrations were correlated with high Na+ and HCO3– concentrations due to interactions between the water and bedrock containing fluorine minerals.

Based on indicators of natural and anthropogenic contamination, the BO region is a promising candidate site for a large-scale groundwater system. The groundwater in this region has a relatively fast recharge time within a uniform time span, and contamination indicators were observed within background levels. Further groundwater sampling should be conducted at the BO site using a denser sampling strategy and more diverse indicators for the creation of a systematic preservation plan for long-term groundwater availability.