Formation Mechanisms and Hydrogeochemical Evolution of a Metasilicate-Strontium Rich Mineral Water in a Subtropical Volcanic Terrain, East China

Abstract

Natural mineral waters hosted in volcanic terrains are globally significant, but the co-enrichment mechanisms of metasilicate and strontium remain poorly understood. Here we investigate a Jurassic volcanic-hosted mineral water source in eastern China using hydrochemical analysis, ¹⁴C dating, stable isotopes, and structural analysis. The groundwater is of Ca–Mg–HCO₃ type with slightly alkaline pH (7.44–7.63). Metasilicate (26.4–32.9 mg/L) and strontium (0.40–0.83 mg/L) co-enrichment is governed by plagioclase weathering in a bicarbonate-dominated, weakly alkaline environment where ${\mathrm{S}\mathrm{r}\mathrm{H}\mathrm{C}\mathrm{O}}_3^{+}$ ion pairs enhance strontium mobility. Pearson-corrected ¹⁴C ages of 3900–4900 years reveal that millennial-scale residence time is critical for sufficient water-rock interaction and attainment of regulatory thresholds. A conduit-barrier system formed by NW-trending extensional-shear and NNE-trending compressional-shear faults controls groundwater flow paths and residence times, leading to systematic inter-well hydrochemical differentiation. These findings provide a theoretical basis for the genetic identification, potential evaluation, and sustainable management of high-quality mineral water resources in volcanic terrains.

1. Introduction

Natural mineral water is a valuable groundwater resource containing specific chemical components beneficial to human health, and its formation results from the long-term coupling of particular geological and hydrological conditions [1,2,3]. Volcanic terrains, especially areas underlain by intermediate-acid volcanic and pyroclastic rocks, commonly host high-quality mineral waters enriched in trace elements such as metasilicate, strontium, and lithium [4,5,6,7]. Metasilicate and strontium are both regulated as threshold indicators for natural mineral water [8,9,10]. Although their simultaneous occurrence above required thresholds is not uncommon, the mechanisms underlying their co-enrichment, the necessary timescale of water-rock interaction, and the spatial differentiation patterns controlled by geological structures remain poorly understood [11,12].

Recent studies have investigated the formation mechanisms of metasilicate-rich mineral waters. In Tunisia, ¹⁴C ages of mineral waters range from 5000 to 20,000 years, indicating mixing between deep ancient water and modern water [13,14]. In the Cameroon Volcanic Line, soda springs exhibit highly depleted δ¹⁸O values and high total dissolved solids (TDS), suggesting paleo-recharge and intense water-rock interaction. In China, the Taoping water source has metasilicate concentrations of 22.0–67.0 mg/L and a groundwater age of ~60 years, with enrichment attributed to silicate weathering and leaching. The Zhaojue area shows metasilicate contents of 25.74–46.04 mg/L and a Ca–Mg–HCO₃ hydrochemical type, where formation is jointly controlled by silicate mineral distribution and fracture development. In the Antu area of Changbai Mountain, average metasilicate content is 50.78 mg/L, with an average groundwater residence time of 35.5 years [15,16]. Strontium isotopes (⁸⁷Sr/⁸⁶Sr) have also been successfully applied to trace water-rock interaction in volcanic terrains [17,18].

These studies have provided an important foundation for understanding the genesis of volcanic-rock hosted mineral waters. However, most have focused either on the source apportionment of single components or on the characterization of regional hydrochemical features [19,20]. Consequently, key scientific questions, such as the factors controlling the co-enrichment of multiple components within the same hydrogeochemical setting, the role of groundwater age in constraining water quality, and the influence of geological structures on hydrochemical differentiation patterns-remain underexplored [21,22,23]. In eastern China, widespread Jurassic-Cretaceous volcanic sequences host several well-known mineral water sources, yet genetic studies on these resources have largely remained at a qualitative descriptive stage, lacking high-precision dating data and quantitative hydrochemical modeling to elucidate the dynamic processes and key controlling factors involved in their formation.

To address these knowledge gaps, this study focuses on the Wawushan area in Liyang, Jiangsu Province, a representative volcanic-hosted mineral water source in eastern China. By integrating high-precision hydrochemical analysis, radiocarbon (¹⁴C) dating, and stable carbon isotope (δ¹³C) measurements, this study aims to elucidate the co-enrichment mechanisms of metasilicate and strontium in the volcanic aquifer, determine the groundwater residence time required for the formation of such compliant mineral water, and clarify the spatial differentiation of groundwater flow paths and chemical compositions controlled by fault structures. The findings will advance the understanding of the formation mechanisms of mineral water in fractured volcanic aquifers and provide a theoretical basis for the exploration, evaluation, and sustainable development of high-quality mineral water resources in similar geological settings.

2. Background

The study area is located in Shangxing Town, Liyang City, Jiangsu Province, China, at the intersection of the Maoshan Mountains and the remnant ranges of the Tianmu Mountains (Figure 1). The regional stratigraphy is dominated by volcanic rocks of the Upper Jurassic Longwangshan Formation (J₃1), which consist mainly of andesite, tuff, and volcanic breccia. This succession serves as the direct host horizon for the mineral water. The rocks are rich in silicate minerals such as plagioclase and K-feldspar, with strontium occurring as an isomorphic substitute in calcium-bearing minerals, providing the material basis for the enrichment of metasilicate and strontium.

Figure 1. Regional Hydrogeological Map.

Tectonically, the study area is situated within the Lower Yangtze Depression of the Yangtze Paraplatform and has experienced multiple phases of tectonic deformation since the end of the Triassic, resulting in well-developed fault systems. At the scale of the water source site, two main sets of faults are present: NW-trending extensional-shear faults and NNE-trending compressional-shear faults. Together, they form a “conduit-barrier” system, wherein the NW-trending faults serve as the dominant pathways for groundwater flow, while the NNE-trending faults act as local barriers. This structural configuration controls groundwater flow paths and residence times, providing the structural conditions for long-term water-rock interaction and spatial hydrochemical differentiation.

The study area has a subtropical monsoon climate, with an average annual precipitation of 1133.6 mm, providing a stable recharge source for groundwater. Groundwater in the area is classified into unconsolidated porous water and bedrock fissure water. The water relevant to the formation of mineral water is mainly hosted in the Jurassic volcanic fissure aquifer, which belongs to the hard rock category of clastic rock fissure water. This aquifer is characterized by hard and brittle lithology with well-developed structural fractures, exhibiting moderate water yield with a single-well yield of 300–500 m³/d. Low-permeability Quaternary fine-grained deposits surrounding the water source area act as a relatively impermeable boundary, facilitating deep circulation and enrichment of groundwater within the volcanic aquifer.

Three production wells (Well #0, Well #1, and Well #2) are present in the water source area, with depths ranging from 235 to 333 m, all extracting from the Jurassic volcanic fissure aquifer. Abundant precipitation combined with well-developed structural fractures collectively forms a complete hydrogeological cycle of “recharge-runoff-enrichment.”

3. Methodology

3.1. Hydrochemical Parameter Testing

Groundwater samples were collected from each of the three wells. The samples were immediately sealed, stored at low temperature, and transported to the laboratory. A total of 54 parameters were analyzed for the hydrochemical characterization.

Metal and metalloid elements (including potassium, sodium, strontium, etc.) were determined using inductively coupled plasma mass spectrometry (ICP-MS). In this method, argon serves as the working gas. A radio-frequency generator creates a high-frequency electromagnetic field that ionizes the argon to form an inductively coupled plasma torch at high temperature. The sample is converted into an aerosol by a nebulizer and then introduced into the plasma torch, where it is fully ionized to form positively charged ions under the high-temperature conditions. The ions pass through a sampling cone and a skimmer cone into the high-vacuum system, are focused by an ion lens, and then enter the mass analyzer. Separation and detection are achieved based on the mass-to-charge ratio (m/z) of different ions, with the signal intensity being proportional to the ion concentration. ICP-MS offers advantages such as high sensitivity, low detection limits, and a wide linear range, enabling simultaneous multi-element analysis, and is currently one of the most effective techniques for trace element determination.

Anions (including bicarbonate, fluoride, sulfate, etc.) were determined using ion chromatography (IC). This method is based on the principle of ion-exchange chromatographic separation, with the column packed with a quaternary ammonium-type strongly basic anion-exchange resin. A carbonate-bicarbonate mixed solution serves as the eluent, flowing through the column at a constant rate to establish a stable ion-exchange environment. After sample injection, target anions undergo competitive adsorption with the resin functional groups and are separated due to differences in their affinities for the stationary phase: ions with weaker affinity are eluted first, while those with stronger affinity are eluted later. The separated anions sequentially enter a conductivity detector, where the signal intensity is positively correlated with ion concentration, and quantification is achieved using a standard calibration curve. This method is characterized by good selectivity, high sensitivity, and the ability to detect multiple ions simultaneously.

It should be noted that the term “metasilicate” is used here in accordance with the Chinese national standard GB 8537-2018 and the official testing method GB 8538-2022 [24,25], where the concentration is expressed as mg/L H₂SiO₃. In the natural water environment at the measured pH of 7.44–7.63, the dominant dissolved silicon species is actually orthosilicic acid (H₄SiO₄). However, the reported values have been converted to H₂SiO₃ equivalents using the standard conversion factor (H₂SiO₃/SiO₂ = 1.30) as required by the regulatory framework.

3.2. Carbon-14 Isotopic Dating

To determine the groundwater residence time, dissolved inorganic carbon (DIC) in water samples was analyzed for accelerator mass spectrometry (AMS) ¹⁴C dating and stable carbon isotope (δ¹³C) analysis. The analyses were conducted by Beta Analytic Inc. in the United States. Sample pretreatment employed a vacuum extraction method: water samples were injected into an acid chamber in a vacuum extractor, and phosphoric acid was added to lower the pH to below 1, thereby converting DIC completely to CO₂ gas. The released CO₂ was cryogenically trapped to remove water vapor and then purified. The purified CO₂ was reduced with hydrogen at 550 °C using a cobalt catalyst, converting it to elemental carbon, which was deposited as graphite on an iron powder carrier to prepare graphite targets for AMS measurement, as shown in Figure 2.

Figure 2. Schematic diagram of the AMS.

¹⁴C measurements were performed using an accelerator mass spectrometer. The basic principle involves ionizing the graphite target in a cesium sputter ion source to produce a beam of negative carbon ions. After sequential stages of low-energy mass analysis, acceleration with gas stripping, and high-energy mass analysis, ¹⁴C ions were counted with a gas ionization detector, while the beam intensities of stable carbon isotopes (¹²C and ¹³C) were measured using Faraday cups. The ¹⁴C content of the sample was then calculated based on isotopic ratios. δ¹³C was measured using a Delta-Plus isotope ratio mass spectrometer (IRMS) manufactured by Thermo Fisher Scientific, with analyses performed independently on the same sample batch as the ¹⁴C determinations.

4. Results

4.1. Hydrochemical Characteristics

The hydrochemical parameters of groundwater from the three production wells in the study area are summarized in Table 1 and the piper three-line graph is shown in Figure 3. Overall, the groundwater is slightly alkaline, with pH values ranging from 7.44 to 7.63. This slightly alkaline environment is favorable for the stability of silicic acid and promotes the hydrolysis of silicate minerals. Total dissolved solids (TDS) range from 487 to 561 mg/L, indicating freshwater of moderate mineralization, which reflects the intensity of water–rock interaction. Inter-well TDS shows an increasing trend from Well #0 (487 mg/L) to Well #1 (554 mg/L) to Well #2 (561 mg/L).

Table 1. Hydrochemical analysis results (selected parameters).

Parameter	Unit	Well #0	Well #1	Well #2
pH	-	7.63	7.56	7.44
Metasilicate	mg/L	28.1	32.9	26.4
Strontium	mg/L	0.40	0.83	0.62
Bicarbonate	mg/L	322	354	361
Calcium	mg/L	70.6	73.5	79.6
Magnesium	mg/L	26.1	18.6	23.5
Sodium	mg/L	16.4	42.2	28.4
Potassium	mg/L	3.47	5.30	4.88
Total dissolved solids	mg/L	487	554	561
Lithium	mg/L	0.013	0.056	0.050
Barium	mg/L	0.17	0.19	0.08
Total alkalinity	mg/L	264	290	296
Chloride	mg/L	7.60	6.51	7.72
Sulfate	mg/L	19.2	28.4	35.5
Fluoride	mg/L	0.240	0.401	0.358
Total hardness	mg/L	284	260	296

Figure 3. Piper trilinear diagram of the water samples.

Anions are dominated by bicarbonate (${\mathrm{H}\mathrm{C}\mathrm{O}}_3^{-}$, with concentrations ranging from 322 to 361 mg/L, accounting for over 89% of the total anion milliequivalents, indicating a typical ${\mathrm{H}\mathrm{C}\mathrm{O}}_3^{-}$ type water. The high ${\mathrm{H}\mathrm{C}\mathrm{O}}_3^{-}$ concentration has dual genetic significance: it is a direct product of the weathering and leaching of silicate minerals in the presence of CO₂, and it acts as a ligand that can form soluble ion pairs (${\mathrm{S}\mathrm{r}\mathrm{H}\mathrm{C}\mathrm{O}}_3^{+}$) with Sr²⁺, significantly enhancing strontium solubility [26]. Inter-well variations in ${\mathrm{H}\mathrm{C}\mathrm{O}}_3^{-}$ content (Well #0 < Well #1 < Well #2) are consistent with the trend observed for TDS. Sulfate (19.2–35.5 mg/L) and chloride (6.51–7.72 mg/L) concentrations are extremely low, confirming that the groundwater resides in a relatively confined, deep-circulation environment and has not been affected by modern surface contamination.

Cations are dominated by calcium (70.6–79.6 mg/L) and magnesium (18.6–26.1 mg/L), accounting for over 80% of the total cation milliequivalents, while sodium (16.4–42.2 mg/L) and potassium (3.47–5.30 mg/L) are present in relatively low concentrations. This cation composition is consistent with hydrochemical characteristics dominated by the weathering of silicate rocks. The Ca/Mg ratio (2.71–3.95) exceeds 1, indicating a significant contribution from plagioclase weathering. Inter-well variations in the Na/K ratio (4.73–7.96) may reflect differences in the relative proportions of plagioclase and K-feldspar within the aquifer at the different well locations. Based on the overall ionic composition, the hydrochemical type of the groundwater can be classified as HCO₃-Ca·Mg.

For the characteristic components, metasilicate (H₂SiO₃) concentrations range from 26.4 to 32.9 mg/L, and strontium (Sr) concentrations range from 0.40 to 0.83 mg/L. All three wells consistently exceed the threshold values specified in the Chinese national standard for drinking natural mineral water (GB8537-2018: H₂SiO₃ ≥ 25 mg/L, Sr ≥ 0.2 mg/L). Metasilicate originates from the hydrolysis of silicate minerals, whereas strontium occurs as an isomorphic substitute in the crystal lattice of calcium-bearing minerals and is released into groundwater upon mineral weathering.

Strictly speaking, at the slightly alkaline pH of the groundwater (7.44–7.63), the dissolved silicon exists predominantly as orthosilicic acid (H₄SiO₄), with negligible concentrations of deprotonated species given the pKa₁ of H₄SiO₄ is approximately 9.8. However, following the Chinese national standard GB 8537-2018, the concentrations are reported as H₂SiO₃ equivalents.

Notably, the concentrations of metasilicate and strontium exhibit a co-variation trend: Well #1 has the highest values for both components (32.9 mg/L and 0.83 mg/L), Well #0 has moderate metasilicate but the lowest strontium (28.1 mg/L and 0.40 mg/L), and Well #2 has the lowest metasilicate but intermediate strontium (26.4 mg/L and 0.62 mg/L). This trend suggests that the two components may originate from the weathering and release of the same suite of mineral assemblages (notably plagioclase), with their enrichment likely facilitated by the high ${\mathrm{H}\mathrm{C}\mathrm{O}}_3^{-}$ environment that promotes strontium mobility.

Trace elements lithium (0.013–0.056 mg/L) and barium (0.08–0.19 mg/L) exhibit stable concentrations far below the relevant regulatory limits, further supporting the consistency of the water source and the closed nature of the system. Systematic chemical differentiation is observed among the three wells. When considered in conjunction with the regional fault framework, this differentiation suggests that variations in groundwater flow paths and the intensity of water–rock interaction are likely the primary factors controlling the inter-well hydrochemical characteristics.

4.2. Carbon Isotope Composition and Groundwater Age

Carbon-14 is widely present in both organic and inorganic carbon in groundwater and serves as an ideal tracer for estimating groundwater ages up to 40,000 years. The measured ¹⁴C activities of water samples from the study area range from 25.15 to 35.31 pMC, with δ¹³CDIC values ranging from −5.85‰ to −8.61‰ (Table 2).

Table 2. ¹⁴C and ¹³C contents and ¹⁴C ages of the water samples.

Sample Number	14C (pMC)	δ13C (‰)	${\mathit{a}}_0^{*}$	Corrected 14C Age
Well #0	35.31	−8.61	59.18	4300
Well #1	25.15	−5.85	40.21	3900
Well #2	29.35	−7.72	53.06	4900

The basis of ¹⁴C dating is the measurement of the reduction in radioactive nuclides in a given sample. The method assumes two main characteristics of the system: (1) the initial content of ¹⁴C in the parent material is known and has remained constant over time; and (2) aside from radioactive decay, the system is closed, with no loss or gain of the parent nuclide. Based on its half-life, the parent nuclide decreases exponentially, allowing for precise age determination. The decay equation is expressed as:

$$a_t=a_0·e^{-\lambda t}$$

(1)

where ${\mathrm{a}}_0$ is the initial radioactivity of the parent nuclide, and ${\mathrm{a}}_{\mathrm{t}}$ is the radioactivity after time. The decay constant λ is equal to ${\displaystyle \frac{\ln 2}{{\mathrm{T}}_{1/2}}}$, $T_{1/2}$ is the half-life of the nuclide. For ¹⁴C, the half-life $T_{1/2}$ is 5730 years. The above equation is thus simplified to:

$$t=-8267\cdot \mathit{ln}{\displaystyle \frac{a_t}{a_0}}$$

(2)

Although rainwater contains ¹⁴CO₂ derived from the atmosphere, it is the soil zone that supplies the radiocarbon signal to recharging groundwater. Carbon released from soil respiration accumulates in the soil zone, creating a large reservoir of ¹⁴C. Figure 4 illustrates the pathway by which atmospheric ¹⁴C enters groundwater through the soil zone. If the ¹⁴C originating from the soil were preserved in the groundwater during flow without being diluted, its decay could be used directly for dating. In reality, however, geochemical reactions both within the soil and along the flow path cause dilution and reduction in the ¹⁴C signal. The most typical reactions include: dissolution of carbonates; exchange with the aquifer matrix; oxidation of “old” organic matter and other biogeochemical reactions within the aquifer; and diffusion of ¹⁴C into the aquifer matrix, among others.

Figure 4. Evolution of ¹⁴C in groundwater.

Through these reactions, ¹⁴C becomes diluted. In the recharge environment, the ¹⁴C activity of DIC in groundwater undergoing carbonate dissolution changes, resulting in an adjusted initial ¹⁴C activity. The decay equation, after correction for this dilution, is thus expressed as:

$$t=-8267\cdot \mathit{ln}{\displaystyle \frac{a_t}{a_0^{*}}}$$

(3)

Therefore, to obtain accurate ages for the measured water samples, data correction is required. The Pearson isotopic mixing model, which considers the balance of carbon in an open system in the recharge area and the dissolution and exchange of carbonates in a closed system, is a reasonable and relatively simple ¹⁴C correction model [27]. ¹³C is an excellent tracer for the evolution of DIC in groundwater in both open and closed systems. The significant difference in ¹³C between soil-derived DIC and carbonate minerals in the aquifer provides a reliable means of determining ¹⁴C dilution through carbonate dissolution. The ¹³C mixing model accounts for the addition of ¹⁴C-bearing DIC during carbonate dissolution under open-system conditions, as well as the subsequent dilution of ¹⁴C that occurs under closed-system conditions.

Any addition, removal, or exchange of carbon in the source of DIC will alter the ¹⁴C concentration, and similarly affect the ¹³C concentration. The corrected initial ¹⁴C activity, $a_0^{*}$, obtained from carbon isotope mass balance, is given by:

$$a_0^{*}=a_{soil}{\displaystyle \frac{{\delta }^{13}{C}_{DIC}-{\delta }^{13}{C}_{carb}}{{\delta }^{13}{C}_{rech}-{\delta }^{13}{C}_{carb}}}$$

(4)

where ${\delta }^{13}{C}_{DIC}$ is the ¹³C content of DIC in the water, ${\delta }^{13}{C}_{carb}$ is the ¹³C content of the carbonate, ${\delta }^{13}{C}_{rech}$ is the ¹³C content of inorganic carbon in the recharge zone under equilibrium conditions, and $a_{soil}$ is the activity of soil CO₂, which is taken as 100 pMC. The value of ${\delta }^{13}{C}_{rech}$ can be obtained from the following equation:

$${\delta }^{13}{C}_{rech}={\delta }^{13}{C}_{soil}+\epsilon {\delta }^{13}{C}_{DIC-CO2\left(soil\right)}$$

(5)

where ${\delta }^{13}{C}_{soil}$ is the ¹³C content of soil CO₂, with a value of −23‰, and $\epsilon$ is the enrichment factor between soil CO₂ and DIC in water. Given that the groundwater in the study area is slightly alkaline and DIC is predominantly in the form of ${\mathrm{H}\mathrm{C}\mathrm{O}}_3^{-}$, $\epsilon {\delta }^{13}{C}_{DIC-CO2\left(soil\right)}$ is 8.5‰ at an average water temperature of 20.3 °C. Consequently, ${\delta }^{13}{C}_{rech}$ is calculated as −14.5‰.

Based on the above calculations, the groundwater ages in the study area are determined to range from 3900 to 4900 years, as summarized in Table 2. The term ‘groundwater age’ used throughout this paper refers to the corrected radiocarbon age of the dissolved inorganic carbon in the water sample, calculated using the Pearson isotopic mixing model, and represents the time elapsed since recharge.

4.3. Mineral Saturation Indices

To quantitatively assess the state of water-rock interaction in the Jurassic volcanic aquifer, speciation and saturation index (SI) calculations were performed using PHREEQC (version 3.7) with the LLNL thermodynamic database. The saturation index is defined as SI = log(IAP/K_t), where IAP is the ion activity product and K_t is the equilibrium constant at the given temperature. Values of SI < 0 indicate undersaturation (mineral dissolution thermodynamically favored), SI > 0 indicate supersaturation, and SI ≈ 0 indicate equilibrium [27,28]. Input data included the full hydrochemical parameters from Table 1 and the measured in situ water temperature (17.5–18.3 °C).

The calculated saturation indices for selected minerals are summarized in Table 3. The results show that all three groundwater samples are undersaturated with respect to primary aluminosilicate minerals such as plagioclase and K-feldspar (SI values ranging from −0.9 to −2.1), indicating that the dissolution of these minerals is thermodynamically favorable and ongoing. This is consistent with the continued release of metasilicate, Ca²⁺, Na⁺, and Sr²⁺ into the groundwater. The waters are near equilibrium with respect to quartz and chalcedony (SI values from −0.3 to +0.2), suggesting that silica concentrations are primarily controlled by the solubility of these SiO₂ phases. Notably, the samples are slightly supersaturated with respect to calcite (SI +0.1 to +0.4) and dolomite (SI +0.2 to +0.6), which may indicate that calcite and dolomite are not primary source minerals but could be precipitating as secondary phases due to CO₂ degassing or common-ion effects—a phenomenon commonly observed in bicarbonate-rich groundwaters. Strontianite (SrCO₃) shows near-equilibrium to slightly undersaturated values (SI from −0.3 to −0.1), indicating that strontium concentrations are approaching saturation with respect to this mineral but have not yet reached the precipitation threshold. These calculations provide direct thermodynamic evidence supporting the interpretations of active plagioclase dissolution and the ongoing evolution of the groundwater chemistry toward mineral equilibrium.

Table 3. Calculated saturation indices (SI) for selected minerals.

Mineral	Chemical Formula	Well #0	Well #1	Well #2
Plagioclase (albite)	NaAlSi3O8	−3.2	−2.8	−3.5
K-feldspar (K-spar)	KAlSi3O8	−1.8	−0.9	−2.1
Quartz	SiO2	−0.1	+0.2	−0.3
Chalcedony	SiO2	−0.4	−0.1	−0.5
Calcite	CaCO3	+0.2	+0.4	+0.1
Dolomite	CaMg(CO3)2	+0.4	+0.6	+0.2
Strontianite	SrCO3	−0.2	−0.1	−0.3

5. Discussion

5.1. Co-Enrichment Mechanism of Metasilicate and Strontium

The co-enrichment of metasilicate and strontium in this subtropical volcanic terrain follows a specific geochemical pathway. The key mechanism can be summarized as the synergistic interplay of three factors: material source, hydrochemical environment, and reaction timescale.

The andesite and tuff of the Jurassic Longwangshan Formation are rich in silicate minerals such as plagioclase and K-feldspar, along with minor strontium-bearing minerals (e.g., strontianite), providing a stable and abundant material source for the enrichment of metasilicate and strontium. Among these, plagioclase acts as a common reservoir for both silicon and strontium: silicon is present in the silicate network structure, while strontium occurs as an isomorphic substitute for calcium in the mineral lattice. During infiltration of CO₂-rich meteoric water, which creates a weakly acidic to weakly alkaline transitional environment, plagioclase undergoes weathering and hydrolysis, simultaneously releasing silicic acid (H₄SiO₄), calcium, sodium, and other cations, with strontium entering the solution as a result [29]. The weakly alkaline environment further promotes the release and migration of SiO₂, maintaining metasilicate concentrations at a stable level of 26.4–32.9 mg/L.

It should be noted that, while plagioclase is the primary source of calcium, sodium, and metasilicate, it contains negligible magnesium. The measured magnesium concentrations (18.6–26.1 mg/L) in the groundwater are instead derived from the incongruent dissolution of ferromagnesian silicate minerals, such as pyroxene, amphibole, and biotite, which are abundant in the andesite and tuff of the Longwangshan Formation [30,31]. Upon weathering under CO₂-charged meteoric water, these minerals release Mg²⁺, Fe²⁺, and additional silica, contributing to the overall cation budget [32]. This distinction is critical because the combined weathering of plagioclase and ferromagnesian minerals accounts for the full suite of major cations observed (Ca²⁺, Mg²⁺, Na⁺, and K⁺), consistent with the Ca–Mg–HCO₃ water type.

The hydrochemical environment of the study area is characterized by ${\mathrm{HCO}}_3^{-}$ as the dominant anion (322–361 mg/L, accounting for over 89% of total anion milliequivalents). This bicarbonate-rich water plays a dual facilitating role: on one hand, it supplies H⁺ to promote silicate mineral dissolution; on the other hand, the high ${\mathrm{H}\mathrm{C}\mathrm{O}}_3^{-}$ concentration enables the formation of soluble ${\mathrm{S}\mathrm{r}\mathrm{H}\mathrm{C}\mathrm{O}}_3^{+}$ ion pairs with Sr²⁺, significantly reducing strontium adsorption onto mineral surfaces and thereby greatly enhancing its mobility and final concentration in the water [33]. Additionally, regional fault structures facilitate the infiltration of CO₂ into the aquifer, helping to sustain this favorable hydrochemical environment.

Over the course of water-plagioclase interaction, the concentrations of SiO₂ and Sr in solution increase synchronously until approaching dynamic equilibrium. The co-variation trend between metasilicate and strontium concentrations further supports their common origin from the weathering of the same mineral assemblage. Thus, in volcanic aquifers, the concurrent attainment of both components above regulatory thresholds is likely not an independent event but rather a product of the same set of geochemical processes.

5.2. Implications of Groundwater Age for Genesis

The obtained groundwater ages of 3900–4900 years provide a critical temporal constraint for understanding the formation of this type of volcanic-hosted mineral water. Compared to younger groundwater (<1000 years), this millennial-scale residence time carries threefold genetic implications.

First, it represents a sufficient duration for water–rock interaction. Relatively slow geochemical reactions—such as the hydrolysis of silicate minerals and the leaching of strontium-bearing phases—can proceed to completion. If the groundwater circulation period were only a few decades, these reactions would be insufficient to accumulate characteristic components to regulatory levels. The fact that all three wells yield ages exceeding 3000 years directly supports the genetic logic of long-term water-rock interaction.

Second, it serves as an indicator of hydrogeochemical maturity. On this timescale, the hydrochemical type stabilizes, the saturation indices of major minerals approach equilibrium, and water quality parameters exhibit minimal temporal variation.

Third, it acts as a natural barrier against modern anthropogenic contamination. The groundwater was recharged long before the industrial era and has no hydraulic connection with modern surface water. These ages are comparable to those reported for high-quality mineral waters in other typical volcanic terrains worldwide, for example, 2000–5000 years in the Mount Etna region of Italy and 2500–4500 years in the Lake Chapala basin of Mexico [34]. This consistency suggests that the formation of mineral water with stable chemical compositions and sensory properties requires a hydrological cycle spanning several millennia.

5.3. Hydrogeochemical Differentiation Controlled by Fault Structures

The well-developed fault systems provide critical geological conditions for hydrogeochemical differentiation. In the study area, the NW-trending extensional-shear faults serve as high-permeability preferential flow pathways, governing rapid groundwater flow and regional recharge direction, whereas the NNE-trending compressional-shear faults act as hydraulic barriers that force groundwater to decelerate and stagnate, locally forming enrichment zones for hydrochemical evolution.

All three production wells are located near the intersection of these two fault sets; however, their specific positions relative to the fault zones result in systematic differences in flow paths, residence times, and the lithological assemblages encountered, leading to distinct patterns of hydrochemical differentiation. Well #2 is situated at the core of the intersection between the NW-trending and NNE-trending faults, where the influence of the barrier fault is greatest. It exhibits the longest flow path and residence time (4900 years), has undergone the most extensive water-rock interaction, and shows the highest total dissolved solids (561 mg/L) and bicarbonate (361 mg/L). Well #1 has the youngest age (3900 years), but its strontium (0.83 mg/L), sodium (42.2 mg/L), and metasilicate (32.9 mg/L) concentrations are markedly higher than those of the other two wells, suggesting that its flow path may have passed through lithological units enriched in easily weathered minerals such as albite, resulting in more efficient water-rock interaction and allowing characteristic components to reach elevated concentrations over a relatively shorter timescale. Well #0 has an intermediate age (4300 years) but the lowest TDS (487 mg/L) and strontium (0.40 mg/L) among the three wells, possibly because it lies on a relatively direct flow path where water-rock contact area or mineral reaction rates are comparatively limited.

This spatial differentiation pattern reveals the combined control of fault architecture on water-rock interaction intensity, flow paths, and geochemical evolution. The relationship between groundwater age and solute enrichment is not simply linear; the lithological composition along flow paths and the spatial configuration relative to faults also exert important influences on hydrochemical evolution.

From an engineering perspective, the “lithology-time-structure” model provides actionable guidance for wellfield design and risk mitigation. The systematic relationship between fault architecture and groundwater age suggests that new production wells should be sited at the intersection of NW- and NNE-trending faults to access the oldest, most mineralized water, while avoiding locations where younger, more variable water may intrude. The observed millennial-scale residence time also implies a very low recharge rate; thus, sustainable pumping rates must be carefully constrained to prevent drawdown that could induce vertical mixing with shallower, younger groundwater. From a risk assessment standpoint, the pristine nature of the water and its age indicate high vulnerability to surface contamination if the confining barrier is breached.

6. Conclusions

Through a systematic investigation of the volcanic-hosted mineral water in Wawushan, Liyang, the following main conclusions are drawn:

• The co-enrichment of metasilicate and strontium in the study area is governed by the synergistic interplay of three factors: material source, hydrochemical environment, and reaction timescale. Silicate minerals such as plagioclase in the Jurassic volcanic rocks provide a common material source for both components. The bicarbonate-dominated and weakly alkaline environment not only facilitates the hydrolysis of silicate minerals but also significantly enhances strontium solubility through the formation of soluble ${\mathrm{S}\mathrm{r}\mathrm{H}\mathrm{C}\mathrm{O}}_3^{+}$ ion pairs. Millennial-scale water-rock interaction allows the two components to accumulate synchronously until reaching dynamic equilibrium.
• Carbon-14 dating, corrected by the Pearson model, yields groundwater ages of 3900 years (Well #1), 4900 years (Well #2), and 4300 years (Well #0), indicating that the millennial-scale residence time of 3900–4900 years provides a critical temporal constraint for sufficient water-rock interaction and the attainment of regulatory thresholds for the characteristic components. This timescale ensures that relatively slow reactions—such as the hydrolysis of silicate minerals and the leaching of strontium-bearing phases—proceed to completion, allowing the hydrochemical type to stabilize and the saturation indices of major minerals to approach equilibrium, while also serving as a natural barrier against modern anthropogenic contamination, thereby forming a high-quality mineral water with mature hydrogeochemical characteristics.
• The conduit-barrier system formed by NW-trending extensional-shear faults and NNE-trending compressional-shear faults controls the inter-well hydrochemical differentiation. Well #2, located at the core of the fault intersection, exhibits the oldest corrected ¹⁴C groundwater age (4900 years) and the highest total dissolved solids and bicarbonate. Well #1 has the youngest groundwater age (3900 years) but the highest concentrations of strontium, sodium, and metasilicate. Well #0 shows an intermediate groundwater age (4300 years) but the lowest TDS and strontium. This differentiation pattern demonstrates that the relationship between groundwater age and solute enrichment is not simply linear; rather, residence time, the lithological composition along flow paths, and the spatial configuration relative to faults collectively determine the final hydrochemical composition.

Limitations in this study and the work in the future: the study is based on only three sampling wells located within a single water source area. This limited spatial coverage may not adequately represent the full hydrochemical heterogeneity of the Jurassic volcanic aquifer. And, only one set of samples was collected from each well. Consequently, the analytical reproducibility and potential temporal variability of the hydrochemical parameters and ¹⁴C ages have not been statistically assessed. While the systematic inter-well consistency in major ions, TDS, and ¹⁴C ages provides confidence in the overall dataset, future investigations should include replicate sampling to evaluate analytical precision and to capture possible temporal fluctuations in water chemistry. Such long-term monitoring would also help to confirm the stability of the groundwater ages and the sustainability of the mineral water resource under exploitation.