Source and Evolution of Subduction–Related Hot Springs Discharged in Tengchong Geothermal Field, Southwest China: Constrained by Stable H, O, and Mg Isotopes

Abstract

The hydrothermal system plays a crucial role in material and energy cycling between the lithosphere and hydrosphere. In general, seafloor hydrothermal systems are one of important Mg sinks, but the situation may not be the same as it is in terrestrial hydrothermal systems. In addition, the behavior of Mg isotopes during hydrothermal circulation is still unclear. Thus, in this study, we determined the Mg isotopic compositions of the hydrothermal fluids discharged in the Tengchong region to understand better the fate of Mg in the continental hydrothermal system. The δ²H and δ¹⁸O values of the Tengchong hydrothermal fluids indicate that the recharge water sources are primary from meteoric water and influenced by the evaporation process. In contrast, the subduction–related volcanic water input is limited, except in for the Rehai area. The Mg in most of the samples is contributed by percolated meteoric water. The measured δ²⁶Mg values range from –0.969 to 0.173‰, which are enriched in light Mg compared to the volcanic rocks of Tengchong. Combined with the precipitation dissolution of carbonate, we calculated the δ²⁶Mg value for the endmember fluid before precipitation, which shows that the process of carbonate precipitation changes the Mg isotope of the fluid, substantially. The Shiqiang (SQ) vent is unique among all of the samples, characterized by an extremely a high δ²⁶Mg value and Mg concentration, and it is estimated that it could have been mixed with an upper crustal material. This also reveals the diversity of the hydrothermal fluid material sources in the subduction zone.

1. Introduction

Hydrothermal systems play a crucial role in the cycling of matter and energy between the lithosphere and hydrosphere [1,2]. Hydrothermal circulation usually requires the presence of a magma bodies to provide the heat source [3,4] and fractures to provide channels [5]. Generally, the chemical compositions of hydrothermal fluid are affected by the magmatic input, water–rock interaction, phase separation, mineral precipitation, and the mixing of it with shallow cold water [1,6,7,8]. The seafloor hydrothermal system is one of the most significant Mg sinks [9], whose Mg content is typically near zero [10]. However, the Mg geochemical behavior in the terrestrial hydrothermal system may differ. In contrast to submarine hydrothermal systems, the recharge water source for terrestrial hydrothermal systems is mainly from infiltrating meteoric water [11]. Compared with the stable chemical components of seawater, the chemical components of the infiltrating meteoric water are variable, and the temperature of the terrestrial hydrothermal reservoir is generally lower than that of the seafloor hydrothermal system, and Mg may not be completely removed, making the Mg content in the terrestrial hydrothermal system non-zero.

The Mg isotopic compositions of the mantle are relatively homogeneous, with the δ²⁶Mg values ranging from −0.35‰ to −0.18‰, and there being a mean value of −0.25 ± 0.07‰ [12]. This phenomenon indicates that high-temperature geological processes (e.g., dehydration process of subducted slabs, partial melt, and magma fractionation) have relatively minor effects on Mg isotope fractionation [2,12,13,14,15]. Nevertheless, the Mg isotopic composition of the continental crust and rivers differ from the mantle. They both have heterogeneous δ²⁶Mg [16], which suggests that the low-temperature geological processes remarkably fractionate the Mg isotopes. During silicate rock weathering, light Mg preferentially enters the fluid, thus resulting in a weathering residue with a relatively heavy Mg isotopic composition [17,18]. In contrast, light Mg isotopes are preferentially absorbed during the precipitation of carbonate rocks [19]. Huang et al. [20] measured the Mg isotopic compositions of altered oceanic crust, and they figured out the global Mg cycle and Mg isotopic budget. Therefore, deciphering the source and behavior of Mg in hydrothermal circulation is meaningful to understanding the Mg cycle. The Mg cycle is closely associated with the carbon cycle through chemical weathering and carbonate precipitation [21]. We can better understand the Earth’s carbon cycle by studying the magnesium cycle.

Tengchong is the only known high-temperature hydrothermal system that is affected by Cenozoic volcanic activity in China [8], among which the Rehai, Ruidian, and Panzhihua are three typical high-temperature hydrothermal systems. Thus, Tengchong plays a significant role in hydrothermal studies. Previous studies have suggested that the concentration of elements is affected by the recharge of magmatic fluids, mixing with cold water and the water–rock interaction [8,22,23,24,25]. Some scientists found that the Tengchong volcanic rocks exhibit homogeneous and light Mg isotopic compositions, and the values are lower than that of the average mantle [26]. They believed that Mg-rich carbonate dissolved by supercritical liquids in the oceanic subduction zone, which formed the low δ²⁶Mg volcanic rocks. Shi et al. [27] quantified mantle and hydrothermal endmembers’ contributions. However, the superficial Mg cycle associated with hydrothermal circulation is unknown. Is the composition of the Tengchong hydrothermal fluid influenced by the deeper material in the subduction zone? Tengchong’s hydrothermal system is related to the subduction of the Indian Ocean Plate beneath the Eurasian Plate, which provides a great research site for us to understand the geochemical behavior of Mg during hydrothermal circulation.

In this study, we determined the Mg isotopic composition and major metal elemental concentrations of hot springs discharged in Tengchong. Combined with H and O isotopes and trace elements, we try to explain the origin and evolution of Tengchong hydrothermal fluid. Furthermore, we attempt to illustrate the fate of Mg during hydrothermal circulation.

2. Geological Setting

The Tengchong volcanic field (98°05′ E~98°45′ E, 24°38′ N~25°52′ N) is located at the Southwest of Yunnan province, China, along the boundary between the Indian Ocean Plate and the Eurasian Plate, and it belongs to the Tengchong microplate [25,28,29]. High-resolution seismic tomography show that the eastward subduction of the Indian plate has reached the Burma–Tengchong massif [30,31,32]. Numerous geophysical studies show the existence of intra-crustal low-velocity bodies in the area, and the three-dimensional velocity structure studies show the existence of negative anomalies in the upper crust and upper mantle of the Tengchong volcanic region, and the existence of magma-related hot material in the upper crust from the upper mantle [33,34,35,36].

Volcanoes and faults structures are remarkable tectonic features in Tengchong [37]. Almost all of the eruptive dacitic volcaniclastics and lava erupted in the early Pleistocene. In the middle Pleistocene, the Tengchong volcanism reached its climax, and several cinder cones formed. In the late Pleistocene, a moderately quiet central eruption occurred in Tengchong [38]. Additionally, the latest mafic magma eruption occurred in the Holocene [39]. Volcanism continued from the Pliocene to the Holocene in Tengchong, which provided a heat source for hydrothermal activity. Since 55 Ma, the collision of the Indian-Eurasian plate caused the compressional tectonics in the southern part of the Tibetan Plateau, and this also caused the stretching of the tectonics in Tengchong [40]. The lithosphere of Tengchong is under NW extensional stress and NE extrusion [41], thus, this has generated abundant NE and NS extending faults [42,43,44]. These faults create favorable conditions for volcanic activities and hydrothermal activities.

Tong and Zhang [28] reported that the Tengchong outcrops cover an area of more than 1000 km², and this include basalt, andesitic basalt, andesite, and dacite. The petrography shows that the volcanic terrains in Tengchong are typical of basalt–andesite–dacite arc volcanic rock associations of the calc-alkaline series [45]. The isotopic data of the Tengchong volcanic rocks also indicate that there is obvious crust–mantle mixing in the source area, and this is influenced by the subduction material [26,46].

Driven by the underlying magma chambers and faults, hydrothermal systems are widely distributed along the three main faults (the Lianghe–Guyong Fault, Xiaolongchuan—the Tengchong–Ruidian Fault; Longchuanjiang—the Longlin–Ruili Fault) in Tengchong [43,45]. Most of the geothermal springs in Tengchong are alkaline or slightly alkaline, with an average temperature that is higher than 45 °C [47]. In addition, three boiling springs groups are distributed in Tengchogn. Rehai is the most active and largest hydrothermal system in Tengchong, whose reservoir consists of Yanshanian granite and Proterozoic metamorphic rocks [25].

3. Materials and Methods

3.1. Sampling

The sampling campaigns were conducted in Jan 2019 and Dec 2020, respectively. Fluid samples were collected from the Tengchong hydrothermal field using low-density polyethylene (LDPE) bottles with a 125–250 mL volume. The sampling sites are marked in Figure 1, and Figure 2 shows photographs of some sampling sites. They are the Datang (DT) vent, the Shiqiang (SQ) vent, the Banglazhang (BLZ) area, the Longwo (LW) vent, the Taihe (TH) vent, the Laxing (LX) vent, the Rehai area, the Reshuitang (RST) vent, the Dapingzi (DPZ) vent, the Shihuadong (SHD) vent, the Qingkou (QK) vent, the Xiaolaisong (XLS) vent, and the Heinitang (HNT) vent. Specifically, the Rehai area includes the Huaitaijing (HTJ) vent, the Huangguaqing (HGQ) vent, and the Zhenzhuquan vent, while the Banglazhang area, consists of the Banglazhang vent and the Dafeiquan (DFQ) vent. Two bottles of water samples were collected from each sampling site. The vent fluids were used to rinse the LDPE bottles several times before the sampling. Fluid samples were also collected from the Longchuangjiang river for compassion. All of the fluid samples were preserved as original without filtration or acidification for later liquid analyses.

The venting temperatures were monitored using a DTM-280 LCD thermocouple probe directly inserted into the venting area. The temperature data were recorded using a data logger connected to the probe after the readings were stabilized. The temperature of the samples in this study ranges from 48.6 °C to 99 °C. The measurement uncertainties were ±0.5 °C.

3.2. Analytical Methods

The total alkalinity (TA), pH, and major anions of the samples were measured at Ocean College, Zhejiang University. The pH values were determined using a PHS-3C pH meter (Leici, Shanghai, China) at 25 °C with an analyzing accuracy is ±0.01. The TA was analyzed using an AS—ALK2 total alkalinity titrator (Apollo, the US with a precision of 2‰. The chemical compositions of the samples were analyzed at Ocean College, Zhejiang University. The major anions (SO₄²⁻, Cl⁻) were analyzed by ICS-5000+ ion chromatography (Agilent, Santa Clara, CA, USA). The measurement uncertainties of SO₄²⁻ and Cl⁻ were ±10%. The SiO₂ contents were measured using a CLEVERCHEM 380plus (DeChem-Tech, Hamburg, Germany) with a test accuracy of 0.001 mg/L. The accuracy and precision of the analyses were controlled by using blanks, standards, and sampling replicates.

The concentrations of the major (Ca, Mg, Na, and K) and trace metals were analyzed using an Agilent 7700e ICP-MS (Thermo Fisher, Waltham, MA, USA) at the Wuhan Sample Solution Analytical Technology Company. The measurement uncertainties were 5% based on repeated analysis of standards (AGV-2, BHVO-2, BCR-2, and RGM-2).

The Mg isotopic compositions of the samples were determined using a Neptune Plus MC-ICP-MS (Thermo Fisher Scientific, Dreieich, Germany) at the Beijing Createch Testing Technology Company. Detailed chemical separation and measurement procedures were described in a previous study [50]. Specifically, a mixture of concentrated HF-HNO₃ (~3:1, v/v) was used for the digestion. The samples were then refluxed with concentrated HNO₃ to remove the residual fluoride. Finally, the samples were dissolved in 1 mL 2 N HNO₃ for chromatographic column chemistry. Mg purification was performed in Savillex microcolumns that were loaded with 2 mL Bio-Rad AG50W-X12 cation resin. An in-house IGGMg1 Mg standard solution was used as a bracketing reference standard. All of the Mg isotope results were reported relative to the commonly used reference standard DSM-3 by the following calculations:

$${\mathsf{\delta }}^{25}{\mathrm{Mg}}_{\mathrm{DSM}-3}\left(‰\right)={ \mathsf{\delta }}^{25}{\mathrm{Mg}}_{\mathrm{IGGMg}1}-0.904$$

(1)

The Mg isotopic compositions are reported using δ-notation relative to DSM-3:

$${\mathsf{\delta }}^{25}{\mathrm{Mg}}_{\mathrm{DSM}-3}\left(‰\right)=\left(\frac{{{(}^{25}\mathrm{Mg}{/}^{24}\mathrm{Mg})}_{\mathrm{sample}}}{{{(}^{25}\mathrm{Mg}{/}^{24}\mathrm{Mg})}_{\mathrm{DSM}-3}}-1\right)\times 1000‰$$

(2)

$${\mathsf{\delta }}^{26}{\mathrm{Mg}}_{\mathrm{DSM}-3}\left(‰\right)={\mathsf{\delta }}^{25}{\mathrm{Mg}}_{\mathrm{DSM}-3}\times \frac{{\mathsf{\delta }}^{26}{\mathrm{Mg}}_{\mathrm{IGGMg}1}}{{\mathsf{\delta }}^{25}{\mathrm{Mg}}_{\mathrm{IGGMg}1}}$$

(3)

The value of δ²⁶Mg was obtained by multiplying the value of δ²⁵Mg and the mass fractionation factor of the instrumental test. The precision and accuracy were assessed by replicated measurements of the standard materials (CAM-1 and GSB Mg). The δ²⁵Mg of the repeatedly analyzed CAM-1 and GSB Mg are −1.35 ± 0.04 (2 SD, n = 7) and −1.01 ± 0.03 (2 SD, n = 10), respectively.

The stable hydrogen and oxygen isotopic compositions of the samples were determined using a Thermo Flash HT—Delta V Advantage isotope ratio mass spectrometer (Thermo Fisher, Waltham, MA, USA) at Ocean College, Zhejiang University. The fluid was reduced to H₂ gas via the Zn reduction technique for analyzing the hydrogen isotopic composition on EA/HT-IRMS. The oxygen isotopic compositions were measured by the CO₂–H₂O isotope exchange technique on Gasbench-253Plus. Three certified reference materials (GBW04458, GBW04459, and GBW04460) were used for the calibration. The δ²H and δ¹⁸O values were reported to be relative to V-SMOW. The analytical precisions were ±1‰ for δ²H and ±0.1‰ for δ¹⁸O.

3.3. Calculations and Statistical Analyses

The concentrations of CO₃²⁻ and HCO₃⁻ were calculated from the measured TA and pH values. Due to the negligible low content of other weak acid anions in the sample, the TA is almost exclusively contributed by carbonate alkalinity.

$$\mathrm{TA}\approx \left[{\mathrm{HCO}}_3^{-}\right]+2\times \left[{\mathrm{CO}}_3^{2-}\right]$$

(4)

$${\mathrm{HCO}}_3^{-}+{\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{H}}_3{\mathrm{O}}^{+}+{\mathrm{CO}}_3^{2-}$$

(5)

$${\mathrm{K}}_{\mathrm{a}2}=\frac{\left[{\mathrm{H}}^{+}\right]\left[{\mathrm{CO}}_3^{2-}\right]}{\left[{\mathrm{HCO}}_3^{-}\right]}$$

(6)

$$\left[{\mathrm{CO}}_3^{2-}\right]=\frac{\mathrm{TA}\times {\mathrm{K}}_{\mathrm{a}2}}{{10}^{-\mathrm{pH}}+2\times {\mathrm{K}}_{\mathrm{a}2}}$$

(7)

$$\left[{\mathrm{HCO}}_3^{-}\right]\approx \mathrm{TA}-2\times \left[{\mathrm{CO}}_3^{2-}\right]$$

(8)

The K_a2 values vary as a function of temperature, pressure, and salinity. In this study, we determined the K_a2 at saturation pressures using R programming 3.6.3 with a package of CHNOSZ [51].

4. Results

4.1. Chemical Compositions of the Tengchong Hydrothermal Fluids

The chemical compositions of the geothermal fluids are highly variable across the Tengchong field (Table S1). The average temperature of the hydrothermal fluids in Tengchong is 73 °C. Some of the vents (HGQ, ZZQ, HTJ, DFQ, BLZ) have temperatures exceeding 90 °C, which are close to the local boiling point. The hydrothermal fluids are mostly medium alkaline, with an average pH of 7.40. HGQ and ZZQ are the few acidic vents in this study, with pH values of 3.23 and 3.28, respectively.

The Mg content varies greatly in the Tengchong hydrothermal fluid, ranging from 0 mmol/kg (HTJ) to 1.21 mmol/kg (SQ). The average Mg concentration was 0.347 mmol/kg, which is two times higher than it is in the river (0.146 mmol/kg). The average Mg content of the Tengchong hydrothermal fluids is 0.492 mM after removing several vents with extremely low Mg contents (<0.01 mM). The average calcium concentration of the hydrothermal fluids in Tengchong is 0.604 mmol/kg, with the highest content being in the HGQ (4.20 mmol/kg). The vent with the lowest calcium content is the HTJ, with a concentration of only 0.028 mmol/kg. The highest cation content in most of the samples is Na, with the highest content of 32.6 mmol/kg in the HTJ, and the lowest content of 0.212 mmol/kg in the HGQ. The average concentration of Na in the Tengchong hydrothermal fluid is 9.21 mmol/kg. Potassium and sodium have similar chemical properties, and among the fluid samples of the Tengchong hydrothermal fluid, the K content of HTJ is also the highest, at 2.66 mmol/kg. The XLS has the lowest K content of only 0.067 mmol/kg. The Li content of most of the hydrothermal fluids in Tengchong ranges from 2 to 50 μM, among which the HTJ exhibits extremely high Li concentrations up to 143 μM. In contrast, the Li content of HGQ is extremely low, with it being less than 0.5 μM.

Chlorine is one of the major anions in hydrothermal fluids. The average Cl concentration in the Tengchong hydrothermal fluid is 1.105 mM, but the Cl content varies greatly among the hot springs, with the highest Cl content in HTJ (8.13 mM), and the lowest Cl content in HGQ at 0.003 mM. Carbonate and bicarbonate are important components of the Tengchong hydrothermal fluid. The TA varies within 2.16–22.9 mM. Unlike most of the hydrothermal fluids in Tengchong, where Cl and HCO₃ are the main anions, the anions in ZZQ and HGQ are dominated by SO₄, by up to 1.61 mM.

4.2. Stable H and O Isotopes

The δ²H of the Tengchong hydrothermal fluid samples falls in a range from −91‰ to −51‰ (

Table 1. The stable hydrogen and oxygen isotope compositions of the Tengchong hydrothermal fluids.

Site	Sample	δ 2H	Stdev (‰)	δ 18O	Stdev (‰)
DT	015−DT−3	−83.74	0.07	−11.31	0.01
DT	015−DT−4	−84.17	0.16	−11.21	0.02
DT	2020−DT−L1	−87.02	0.88	−11.59	0.09
SQ	015−SQ−3	−81.40	0.12	−10.34	0
SQ	2020−SQ−L1	−84.17	0.25	−10.91	0.03
SQ	2020−SQ−L3	−85.22	0.33	−10.85	0.11
BLZ	2020−BLZB3−L1	−72.45	0.13	−8.62	0.16
BLZ	2020−BLZ−L1	−74.40	0.16	−9.23	0.05
DFQ	2020−DFQ−L2	−75.28	0.07	−10.10	0.07
TH	2020−TH−L2	−72.46	0.08	−8.93	0.12
LX	2020−LX−L1	−86.08	0.37	−10.72	0.04
HGQ	015−HGQ−1	−52.68	0.46	−3.35	0.19
HGQ	015−HGQ−2	−51.50	0.08	−0.47	0.12
HGQ	2020−HGQ−L2	−70.92	0.05	−9.38	0.08
HGQ	2020−HGQ−L4	−62.82	0.33	−7.87	0.06
ZZQ	015−ZZQ−1	−58.01	0.27	−4.20	0.06
HTJ	015−HTJ−2	−72.18	0.25	−7.45	0.07
RST	2020−RST−L2	−71.17	0.87	−9.04	0.04
DPZ	015−DPZ−1	−70.53	0.06	−8.63	0.04
LW	2020−LW−L2	−71.46	0.06	−9.72	0.05
SHD	2020−SHD−L3	−91.40	0.07	−12.35	0.13
QK	2020−QK−L1	−80.20	0.46	−10.65	0.08
XLS	2020−XLS−L1	−84.31	0.47	−11.47	0.14
XLS	2020−XLS−L3	−80.05	0.04	−11.01	0.08
HNT	015−HNT−3	−73.76	0.21	−9.72	0.10
HNT	015−HNT−5	−73.89	0	−9.74	0.08
river	2020−RLJ−L1	−75.96	0.13	−10.34	0.07

), with an average of −75.12‰. The δ¹⁸O values vary from −12.3‰ to −0.5‰, showing an average of −9.24‰. The isotopic compositions of the river samples overlap with these ranges with δ²H and δ¹⁸O values of −76‰ and −10.3‰, respectively. HGQ and ZZQ, as the few two acidic vents, show the highest δ²H and δ¹⁸O values among all of the samples.

4.3. Mg Isotopes

The Mg isotopic compositions of the Tengchong geothermal fluids are also highly variable between the sampling locations (

Table 2. The Mg isotopic compositions of the Tengchong hydrothermal fluids.

Site	Sample	δ25Mg DSM-3‰	2σ	δ26Mg DSM-3‰	2σ
DPZ	015−DPZ−1	−0.17	0.02	−0.33	0.01
HGQ	015−HGQ−1	−0.36	0.03	−0.70	0.04
HGQ	015−HGQ−3	−0.32	0.02	−0.62	0.02
HNT	015−HNT−1	−0.51	0.01	−0.97	0.04
LX	015−LX−1	−0.19	0.02	−0.37	0.03
SQ	015−SQ−1	0.09	0.05	0.17	0.01
QK	2020−QK−L1	−0.12	0.02	−0.23	0.05
river	2020−RLJ−L1	−0.67	0.05	−1.30	0.08
river	2020−RLJ−L1	−0.70	0.02	−1.36	0.04
RST	2020−RST−L2	−0.40	0.04	−0.67	0.02

). The river samples show a δ²⁵Mg value of −0.24‰, corresponding to a δ²⁶Mg value of −0.64‰. Both of the HGQ and RST demonstrate comparable δ²⁵Mg values of −0.32 ± 0.02 ~ −0.36 ± 0.03‰, corresponding to δ²⁶Mg values of −0.62 ± 0.02–−0.70 ± 0.04‰, respectively. DPZ, LX, and QK present similar Mg isotopic compositions with a δ²⁶Mg value of −0.23~−0.33‰. SQ manifests the most elevated δ²⁶Mg value of +0.17 ± 0.01‰. The previous studies show that the magnesium isotopic composition of the mantle is relatively homogeneous, with δ²⁶Mg ranging from −0.35‰ to −0.18‰, and a mean value of −0.25 ± 0.07‰ [12]. The Mg isotopic composition of seawater does not vary with location, depth, temperature, and salinity [52,53], with a mean value of −0.83 ± 0.09‰ for δ²⁶Mg. The δ²⁶Mg of the global rivers range from −2.52‰~−0.31‰, with a mean value of −1.09 ± 0.05‰ [54] and no significant fractionation of the Mg isotopes occurred during the river transport. Tian et al. [26] found that the Tengchong volcanic rocks exhibit homogeneous and Mg isotopic-like compositions (δ²⁶Mg = −0.51~−0.33‰). Thus, the δ²⁶Mg values of LX, QK and DPZ are close to the mantle. The Mg isotopic compositions of HGQ and RST are comparable to those of the Tengchong volcanic rocks. However, SQ exhibits extremely high δ²⁶Mg values.

5. Discussion

5.1. Source of Vent Fluid

Although meteoric water is the major source of recharge for hydrothermal systems [11], snowmelt and magmatic water also play a role in geothermal fluid recharge [55,56,57,58]. The water resulting from subduction–related volcanic vapors distribution (δ²H = −20 ± 10‰, δ¹⁸O = +8 ± 2‰) is plotted in Figure 3 [59,60]. The mantle−derived magmatic water has a constricted range of δ²H (−65 ± 20‰) and δ¹⁸O (+6 ± 1‰) [61,62]. Fan et al. [63] reported the snowmelt water isotopes (δ¹⁸O = −17‰, δ²H = −123‰). Meteoric water is typically characterized by a linear relationship between ²H and ¹⁸O [64].

However, the fluids’ hydrogen and oxygen isotopic compositions are also susceptible to evaporation and the water–rock interaction [65,66]. Usually, the δ¹⁸O value in rocks is higher than that in fluids. The hydrogen content in rocks is extremely low, so the water–rock reaction causes a positive drift of fluid oxygen isotopes [59]. During evaporation, the lighter water will preferentially form water vapor to leave the fluid, thus heavying the hydrogen-oxygen isotopic composition of the remaining fluid [64]. Thus, stable hydrogen and oxygen isotopes are effective tools to trace the potential water source of hydrothermal vent fluids, the water–rock interaction, and other hydrological cycle characteristics − [3,7,55,67].

Figure 3. Stable hydrogen and oxygen isotopic compositions of hydrothermal fluids collected from Tengchong. The Global Meteoric Water line (GMWL, δ²H = 8 × δ¹⁸O + 1, [64]). Volcanic water resulting from subduction–related volcanic vapor has δ²H values ranging from −10 to −30‰ and δ¹⁸O values from +6 to +10‰ [59,60]. The Evaporation Water Line (δ²H = 5.1 × δ¹⁸O − 26.7, R² = 0.81) was derived from the linear regression of the Tengchong samples, intercepting the GMWL at −89.76‰ (δ²H), −12.47‰ (δ¹⁸O), and it is comparable to previously suggested curves − [55,67]. The intersection points are represented by pentagrams in the figure and represent the initial water sources.

Figure 3. Stable hydrogen and oxygen isotopic compositions of hydrothermal fluids collected from Tengchong. The Global Meteoric Water line (GMWL, δ²H = 8 × δ¹⁸O + 1, [64]). Volcanic water resulting from subduction–related volcanic vapor has δ²H values ranging from −10 to −30‰ and δ¹⁸O values from +6 to +10‰ [59,60]. The Evaporation Water Line (δ²H = 5.1 × δ¹⁸O − 26.7, R² = 0.81) was derived from the linear regression of the Tengchong samples, intercepting the GMWL at −89.76‰ (δ²H), −12.47‰ (δ¹⁸O), and it is comparable to previously suggested curves − [55,67]. The intersection points are represented by pentagrams in the figure and represent the initial water sources.

All of the hydrothermal fluid and river samples were plotted close to the local meteoric and global meteoric water lines (GMWL), demonstrating their meteoric origin. Undeniably, the regression line and the GMWL have some deviations, which the mixing of endmember may cause. Alternatively, the stable hydrogen and oxygen isotopic compositions could be controlled by evaporation [68,69].

Considering the location of Tengchong, it is unlikely that seawater and snow meltwater are potential recharge sources for hydrothermal fluid. Previous studies have shown the presence of magma chambers below Tengchong, so magma water as a recharge source is possible. Figure 3 plots the mixing area of volcanic water and meteoric water. According to Figure 3, a small number of samples (HTJ, HGQ, and ZZQ) are in the mixing area, and these samples received the recharge of magma water. Most of the vents have a mixing ratio of less than 10%.

In contrast, the three vents (HTJ, ZZQ, and HGQ) from the Rehai hydrothermal field are subject to a significantly higher percentages of magma water mixing than the other vents are, with mixing ratios of 25%, 45%, and 54%, respectively. The Rehai hydrothermal field is also located above the magma chamber [70]. Additionally, the vent temperatures of all three are higher than 90°C, and the geological thermometer estimates also show that the Rehai hydrothermal field has the highest temperature of reservoirs in Tengchong, further proving that the hydrothermal vents extracted from the Rehai were affected by obvious magma mixing.

However, most of the samples from Tengchong do not fall in the mixing zone, which shows that these vents contain almost no magma water recharge indicators, and their hydrogen and oxygen isotopic compositions are mainly influenced by water vapor fractionation in the evaporation process. In addition, the equations of the regression lines (δ²H = 5.1 × δ¹⁸O − 26.7) are very close to those that are proposed for the evaporation lines (δ²H = 5.2 × δ¹⁸O − 38.7; δ²H = 5.5 × δ¹⁸O − 19.3) − [55,67].

Generally speaking, the δ¹⁸O and δ²H values vary significantly with elevation in a unique geographic district [71]. The elevation and hydrogen-oxygen isotopic composition of the other samples are shown in Figure 4. The values of δ²H and δ¹⁸O are significantly negatively correlated with altitude. This is mainly due to the fractionation of the hydrogen and oxygen isotopes by altitude effects, with the isotope values decreasing as the altitude increases [72], so we proposed the equation for the variation of the oxygen isotopes with altitude in the Tengchong area:

$$-{\mathsf{\delta }}^{18}\mathrm{O}=0.0038\times \mathrm{H}+4.86 \left({\mathrm{R}}^2=0.803\right)$$

(9)

$$-{\mathsf{\delta }}^2\mathrm{H}=0.0235\times \mathrm{H}+45.46 ({\mathrm{R}}^2=0.844)$$

(10)

where H represents the elevation of the recharge zone. Comparisons were made with Tibet ($-{\mathsf{\delta }}^{18}\mathrm{O}=0.0031\times \mathrm{H}+6.19$)) and the Hani terraces ($-{\mathsf{\delta }}^{18}\mathrm{O}=0.0033\times \mathrm{H}+6.84$), which are expressed in a similar way.

5.2. Magnesium Is Contributed by Percolated Meteoric Water

Based on the formation process of the hydrothermal systems, the potential sources and processes affecting the chemical and isotopic characteristics of Mg in the Tengchong fluids are (1) magmatic fluid input, (2) the water–rock interaction, (3) the leaching of the host rock, and (4) the dissolution and precipitation of Mg-bearing minerals [1,22,25,58].

In Section 5.1, we discussed the recharge water source of Tengchong hydrothermal fluid, and we found that magma water had a limited impact on Tengchong hydrothermal fluid. Moreover, the Mg content of the vents in the Rehai is not high; even the lowest value of Mg content occurs in the HTJ, which shows that the Mg content in the hydrothermal fluids is not mainly controlled by volcanic water recharge. The rivers could be a potential source of Mg in the hydrothermal fluids. However, in the Tengchong hydrothermal system, the Mg content of many hydrothermal vents was much higher than that of the rivers. Rivers are “Mg-poor”, which also indicates that the rivers cannot be an additional source of Mg for the Tengchong hydrothermal system. Mg could be leached from the host rocks by heated fluids during hydrothermal circulation [9]. During this process, lighter Mg isotopes are preferentially leached from the host rock [73]. Hence, the Mg isotopic composition of the hydrothermal fluid is lighter than that of the host rock. However, as the leaching continues, the δ²⁶Mg values of the rock and fluid tend to be close to each other, so the δ²⁶Mg of the fluid increases with the increase in the Mg content, which is not consistent with that shown in Figure 5. Thereout, the leaching of the rock is not the primary source of Mg in the fluid.

During hydrothermal circulation, the water–rock interaction is an essential process in determining the chemical composition of the hydrothermal fluids [74]. Mg is generally removed from percolated fluids by forming Mg-bearing secondary minerals, creating strongly acidic endmember hydrothermal fluids with a totally depleted Mg content [75,76].

According to Figure 6, the Mg content was slightly negatively correlated with the pH, which could be due to the following reasons: (1) the Mg contents of the percolated water were not high enough to decrease the pH significantly. However, the HGQ samples were acidic, which were formed via the absorption of H₂S–rich steam which had been separated from magma degassing [8]. (2) Some of the vents have high pH values, which may be related to serpentinization. Serpentinization is one kind of hydrothermal alteration of basic and ultramafic rocks, which could be expressed by the following reaction equations [77]:

$$2{\mathrm{Mg}}_{1.8}{\mathrm{Fe}}_{0.2}{\mathrm{SiO}}_4+3{\mathrm{H}}_2\mathrm{O}={\mathrm{Mg}}_{2.85}{\mathrm{Fe}}_{0.15}{\mathrm{Si}}_2{\mathrm{O}}_5{\left(\mathrm{OH}\right)}_4+{\mathrm{Mg}}_{0.75}{\mathrm{Fe}}_{0.25}{\left(\mathrm{OH}\right)}_2$$

(11)

$$57{\mathrm{Mg}}_{0.75}{\mathrm{Fe}}_{0.25}{\left(\mathrm{OH}\right)}_2+30{\mathrm{SiO}}_2\left(\mathrm{aq}\right)=15{\mathrm{Mg}}_{2.85}{\mathrm{Fe}}_{0.15}{\mathrm{Si}}_2{\mathrm{O}}_5{\left(\mathrm{OH}\right)}_4+23{\mathrm{H}}_2\mathrm{O}+4{\mathrm{H}}_2+4{\mathrm{Fe}}_3{\mathrm{O}}_4$$

(12)

However, in our previous study, no significant amount of H₂ was found in the hydrothermal gas [27], which showed that serpentinization was not the cause of the hydrothermal pH change in Tengchong. It can be seen that the incomplete removal of Mg is the primary reason for the alkaline nature of the hydrothermal fluid in Tengchong.

Furthermore, heavy Mg isotopes are preferentially incorporated into octahedral lamellae in clay minerals such as chlorite. Therefore, the newly formed clay minerals are enriched in heavy Mg isotopes, reducing both the δ²⁶Mg values and Mg concentrations of the residual hydrothermal fluids [19,78,79]. Nevertheless, the Tengchong hydrothermal fluids exhibit elevated δ²⁶Mg values with decreasing Mg concentrations (Figure 5), suggesting that the formation of secondary minerals does not control the Mg geochemistry in the Tengchong hydrothermal fluids.

In summary, we excluded magmatic input, fluvial recharge, and host rock leaching as potential sources of Mg. Meanwhile, Mg is usually depleted in the hydrothermal reaction zones, thus suggesting that Mg in the hydrothermal fluids of Tengchong is mainly derived from percolated meteoric water. Specifically, the SQ sample showed distinctive characteristics from the other samples, with high δ²⁶Mg values and Mg concentrations, indicating the presence of additional “Mg sources”.

5.3. Precipitation Dissolution Equilibrium of Mg

Moreover, the dissolution and precipitation of Mg-bearing minerals could further modify the chemical and isotopic compositions of Mg in the Tengchong hydrothermal fluids. The commonly occurring sinters evidence this phenomenon in the venting areas. Because carbonate and bicarbonate are the main anions in most of the Tengchong springs, the precipitation and dissolution of Mg-bearing minerals are predominantly dolomite (CaMg(CO₃)₂) and magnesite (MgCO₃). The solubilities of both of the minerals can be calculated using the following equations:

$$\mathrm{CaMg}{\left({\mathrm{CO}}_3\right)}_2\leftrightarrow {\mathrm{Ca}}^{2+}+{\mathrm{Mg}}^{2+}+2{\mathrm{CO}}_3^{2-}$$

(13)

$${\mathrm{K}}_{\mathrm{sp}}={\mathrm{m}}_{{\mathrm{Ca}}^{2+}}\times {\mathrm{m}}_{{\mathrm{Mg}}^{2+}}\times {\left({\mathrm{m}}_{{\mathrm{CO}}_3^{2-}}\right)}^2\times {\mathsf{\gamma }}_{{\mathrm{Ca}}^{2+}}\times {\mathsf{\gamma }}_{{\mathrm{Mg}}^{2+}}\times {\left({\mathsf{\gamma }}_{{\mathrm{CO}}_3^{2-}}\right)}^2$$

(14)

$${\mathrm{MgCO}}_3\leftrightarrow {\mathrm{Mg}}^{2+}+{\mathrm{CO}}_3^{2-}$$

(15)

$${\mathrm{K}}_{\mathrm{sp}}={\mathrm{m}}_{{\mathrm{Mg}}^{2+}}\times {\mathrm{m}}_{{\mathrm{CO}}_3^{2-}}\times {\mathsf{\gamma }}_{{\mathrm{Mg}}^{2+}}\times {\mathsf{\gamma }}_{{\mathrm{CO}}_3^{2-}}$$

(16)

where m is the molality and γ is the activity coefficient of the designated aqueous species. The activity coefficient is generated by the R package CHNOSZ, which was developed by [51].

Here, we calculated dolomite and magnesite solubility products (log Q) in the Tengchong hydrothermal fluids (Figure 7). The hydrothermal fluids from SQ, RST, and LX illustrate significantly higher log Q values of dolomite and magnesite with respect to their equilibrium constants. Therefore, massive precipitation of both of the minerals could have occurred at these sites. These results are consistent with the occurrence of carbonate deposits near the vent of these sites. In contrast, dolomite and magnesite are undersaturated in the hydrothermal fluids from SHD and XLS. Mg²⁺ may be released into the vent fluids if enough Mg-bearing minerals are retained in the vent fluids. The rest sites, QK, DT, HNT, TH, BLZ, and DFQ, manifest log Q values of dolomite and magnesite that are close to the precipitation and dissolution equilibria. As a result, Mg is balanced by the precipitation and dissolution of both of the minerals in these sites. According to the distribution of hydrothermal vents in Tengchong, the hydrothermal vents located on the south side of Tengchong reached precipitation-dissolution equilibrium, but the hydrothermal vents located on the west side of Tengchong (SHD, QK, and XLS) have unsaturated Mg-bearing carbonates, while the eastern and central hydrothermal vents (SQ, LX, and RST) show a supersaturation of Mg-bearing carbonates.

The precipitation dissolution of carbonate reached an equilibrium, indicating that the hydrothermal fluids experienced a long period and reached thermodynamic equilibrium. In comparison, the hydrothermal vents distributed on the south of Tengchong, whose surrounding rocks are Early Cretaceous granitoids, which are the oldest rocks in Tengchong. After a long period of the water–rock reaction, the fluid reached a stable state, thus containing magnesium carbonate precipitation and dissolution equilibrium. Some of the vents are supersaturated, which could have been caused by an increase in the endmember fluid temperature, a decrease in the pressure, or the presence of additional carbonate and Mg inputs. Unsaturated carbonate indicates that other reactions removed the carbonate or Mg. The difference in carbonate precipitation dissolution at the hydrothermal vents on the west and east sides was caused by the local tectonic activity, probably. From Figure 1, we could find that QK, SHD, and XLS were located between the two fractures, while SQ, LX, and RST were closer to the main fault. Taking SQ as an example, some of the vents appeared to be dry and inactive during the sampling period in 2019, but they were active again in 2020, which showed that the local tectonic activity in the short term had a greater influence on the hydrothermal activity. Additionally, the unsaturation of carbonate does not necessarily mean that carbonate precipitation will not be generated. Take SHD as an example, the local topographical change can induce travertine deposition despite the fluid’s low contents of Ca and Mg [49].

5.4. Mg Isotope Fractionation Caused by Precipitation of Carbonate

In Section 5.3, we discussed the precipitation and dissolution of Mg-containing carbonate in the Tengchong hydrothermal fluid, and previous studies have shown that the carbonate precipitation process preferentially utilizes light Mg [16,19], so as carbonate precipitation is generated, the Mg content of the fluid decreases while the δ²⁶Mg values increase, which is consistent with Figure 5. Li et al. [80] conducted simulations of dolomite formation under hydrothermal conditions and showed that the degree of fractionation showed a good correlation with temperature:

Δ²⁶Mg_{dolomite-solution} = δ²⁶Mg_dolomite − δ²⁶Mg_solution = −0.1554 (± 0. 0096) × 10⁶/T²

(17)

where T is the temperature in K. Pearce et al. [81] pointed out that the degree of magnesite fractionation was temperature dependent. At 150 °C, Δ²⁶Mg_{magnesite-solution} is −1.19‰; at 200 °C, Δ²⁶Mg_{magnesite-solution} is −0.88‰; at 25 °C, Δ²⁶Mg_{magnesite-solution} is −2.3‰. Dong and Zhu [82] inferred |Δ²⁶Mg_{dolomite−solution}|≈|Δ²⁶Mg_{magnesite-solution}| from experimental simulations, field observations and theoretical calculations. Therefore, in this paper, we only calculated the effect of dolomite precipitation on the fluid Mg isotopes. We calculated the δ²⁶Mg values of −2.32‰~−1.16‰ for dolomite. The results are consistent with the Mg isotopic composition of carbonates [16,83]. Furthermore, we estimated the initial δ²⁶Mg value of the fluid before the carbonate precipitation according to the following equations:

$${\mathsf{\delta }}^{26}{\mathrm{Mg}}_{\mathrm{dolomite}}={∆}^{26}{\mathrm{Mg}}_{\mathrm{dolomite}-\mathrm{solution}}+{\mathsf{\delta }}^{26}{\mathrm{Mg}}_{\mathrm{solution}}$$

(18)

$$\mathrm{f}=\frac{\left[\mathrm{Mg}\right]-{\left[\mathrm{Mg}\right]}_{\mathrm{hydrothermal} \mathrm{fluid}}}{\left[\mathrm{Mg}\right]}\times 100\%$$

(19)

$${\mathsf{\delta }}^{26}{\mathrm{Mg}}_{\mathrm{initial} \mathrm{solution}}=\mathrm{f}\times {\mathsf{\delta }}^{26}{\mathrm{Mg}}_{\mathrm{dolomite}}+\left(1-\mathrm{f}\right)\times {\mathsf{\delta }}^{26}{\mathrm{Mg}}_{\mathrm{solution}}$$

(20)

where f is the removal rate of Mg, and [Mg] denotes the Mg concentration of the hydrothermal fluid before the precipitation. If carbonate precipitation is the main cause controlling the Mg isotopic composition of the Tengchong hydrothermal fluid, then the fractionation effect caused by the formation of carbonate can be quantified. We calculated the hydrothermal endmember Mg contents of DPZ, HNT, LX, and RST to be 0.49 ± 0.07 mM, 0.68 ± 0.12 mM, 0.51 ± 0.08 mM, and 0.45 ± 0.17 mM, respectively, corresponding to removal rates of 71 ± 5%, 46 ± 8%, 67 ± 6%, and 31 ± 18%, respectively, and the hydrothermal endmember fluids (before precipitation of Mg-containing carbonate minerals) with δ²⁶Mg values corresponding to −1.03 ± 0.17‰, −1.47 ± 0.27‰, −1.08 ± 0.19‰ and −0.94 ± 0.37‰ (Figure 8). The calculated δ²⁶Mg values of the Tengchong hydrothermal fluids are slightly lower than those of seafloor hydrothermal fluids [9,17], which may be caused by two reasons: (1) The Mg of the Tengchong hydrothermal fluids mainly comes from the infiltrating meteoric water. Although the δ²⁶Mg values of meteoric water have not been reported yet, the δ²⁶Mg values of continental surface water (rivers, lakes, and groundwater) are all lower than those of the seawater [84], which may lead to the lighter Mg isotopic composition of the Tengchong hydrothermal fluids. (2) The Mg isotopic composition of the Tengchong volcanic rocks is lighter than that of the MORB and OIBs due to subduction [26,85], thus, during hydrothermal circulation, the fluid–rock interaction may result in fluid with a low δ²⁶Mg value.

However, although JK is located near the fitted line (Figure 5), according to Figure 7, the JK fluid is not saturated with Mg carbonate, indicating that the content of Mg in its fluid is not controlled by the precipitation and dissolution of carbonate. In fact, differences in the host rock, temperature, pressure, and tectonics of vents could cause variations in the Mg concentration of the hydrothermal endmember. The host rocks of Tengchong are dominated by granites (Figure 1b). The mean value of granite δ²⁶Mg in the orogenic belt is −0.21‰, and its variation range is also narrow (−0.26‰~−0.14‰) [86], which is comparable to the Mg isotopic composition of JK. Therefore, the lower Mg concentration with a higher δ²⁶Mg value exhibited by JK was not the result of adequate carbonate precipitation, but it was controlled by the dissolution of the host rock. Not only does SQ have extremely high δ²⁶Mg values, its Mg content is also much higher than that of the other vents. The unusually high δ²⁶Mg values suggest that SQ may have undergone intense carbonate precipitation, and the field observations also indicate substantial surrounding calcareous springing deposits, and the carbonate precipitation dissolution model also indicates that the vent is supersaturated with Mg carbonate, but the SQ deviates far from the fitted line in the figure, which further suggests that there may be additional Mg source input, and thus, SQ affects the Mg isotopic composition of hydrothermal fluids. According to the Mg isotopic compositions of major reservoirs (Figure 9), the continental crust may have provided an additional source of Mg for the SQ.

5.5. Implication for Water and Mg Cycles in the Tengchong Geothermal Systems

Meteoric water is the primary water source of Tengchong hydrothermal systems; however, the hydrothermal fluids’ hydrogen and oxygen isotopic compositions are slightly heavier than the meteoric water ones are. The isotopic difference between the meteoric water and the hydrothermal fluid suggests that other recharge sources and processes preferentially remove light H and O isotopes from the fluid, such as evaporation. The samples collected from the Rehai area were affected by subduction–associated volcanic water vapor, which shows that plate subduction has a non-negligible influence on the water circulation of the Tengchong hydrothermal fluids.

The Mg in the hydrothermal fluid of Tengchong is mainly from the infiltrating meteoric water, and the Mg content is much lower than that of seawater, so the Mg consumption is insufficient to change the pH of the fluid significantly, so the hydrothermal fluid of Tengchong is primarily alkaline. The influence of precipitation of the Mg-containing carbonate minerals on hydrothermal Mg isotopic composition is remarkable because carbonate precipitation preferentially uses light Mg, which can increase the hydrothermal Mg isotopic value by about 0.85‰. The Mg isotopic composition of the hydrothermal fluid at Tengchong proves that the hydrothermal circulation may cause Mg isotopic fractionation to enrich the fluid with light Mg. Furthermore, the SQ also shows the diversity of the Mg sources in terrestrial hydrothermal fluids associated with subduction zones, and continental crustal materials could also be involved in hydrothermal circulation.

6. Conclusions

In this study, we attempted to investigate the origin and evolution of the Tengchong hydrothermal fluids which are constrained by stable H, O, and Mg isotopes. The following conclusions are drawn:

(1)

The δ²H of the Tengchong hydrothermal fluid samples falls in a range from −91‰ to −51‰, with an average of −75.12‰. The δ¹⁸O values vary from −12.3‰ to −0.5‰, showing an average of −9.24‰. The hydrogen and oxygen isotopic compositions illustrate that meteoric water is the primary water source of the Tengchong hydrothermal fluids. However, the samples from the Rehai were affected by the mixing process of the subduction–related volcanic water. Additionally, the δ2H and δ¹⁸O values show a good correlation with the altitude, and the equation for the altitude effect of local hydrogen and oxygen isotopes in Tengchong is proposed: $-{\mathsf{\delta }}^{18}\mathrm{O}=0.0038\times \mathrm{H}+4.86$; $-{\mathsf{\delta }}^2\mathrm{H}=0.0235\times \mathrm{H}+45.46$.

(2)

We measured the Mg isotopic composition of the hydrothermal fluids in Tengchong for the first time. The δ²⁶Mg values (−0.97 ± 0.04‰~0.17 ± 0.01‰) varied considerably in different vents, and they showed lighter Mg isotopic compositions compared to host rock, indicating Mg isotopes fractionation during hydrothermal circulation. The relationship between the Mg concentration and the δ²⁶Mg values suggests that the water–rock reaction is responsible, which dominantly controls the Mg content in the Tengchong hydrothermal system. By simulating the precipitation dissolution of the Mg-containing carbonates, we found that this process can significantly change the δ²⁶Mg values. The δ²⁶Mg values and Mg contents of hydrothermal fluids before the carbonate precipitation were about −1.13 ± 0.25‰ and 0.533 ± 0.11 mM. The removal rates of Mg ranged from 31 to 71%. In contrast, SQ exhibits extremely high δ²⁶Mg values and Mg content, which suggests an input of additional Mg sources, which are possibly influenced by the upper crust.