Degassing of Mantle‐Derived Helium From Hot Springs Along the India‐Asia Continental Collision Settings: Origins, Migration Velocity and Flux

Mantle‐derived volatile degassing lacks quantitative evaluation in continental regions without active magmatism, such as the Tibetan Plateau. Ten new gas abundance and helium isotope data points combined with 286 hydrothermal volatile literature data points in India–Asia continental collision settings demonstrate widespread mantle‐derived volatiles across the thick (∼70 km) crust. The mantle‐derived 3He is best explained by direct mantle volatile input from subcontinental lithospheric mantle (1%–36.5%) or the asthenospheric wedge (1%–27.8%) instead of fossil residual magmatic fluids or Quaternary mantle‐derived melt intrusion into crustal depth. Mantle‐derived 3He is transported from the deep mantle to the surface at an upflow rate of 30–11,700 mm/year based on a newly developed steady‐state one‐dimensional flow model, corresponding to a mantle‐derived 3He flux of 17 to 1.5 × 107 atoms m−2 s−1 (81.7%–99.4% of the total 3He flux) and a mantle‐derived 4He flux from 2.0 × 106 to 1.8 × 1012 atoms m−2 s−1 (1.4%–36.5% of the total 4He flux). The mantle‐derived helium fluxes in the study area are comparable to those of other nonvolcanic hydrothermal systems in the tectonically active regions but lower than those of volcanic fields. The helium transit time ranges from 5.3 ka to 2.3 Ma, indicating that the spatial pattern of 3He/4He ratios in the India‐Asia continental collision settings can provide a snapshot of the state of the Indian mantle lithosphere between those revealed by potassic‐rich mafic rocks (25–8 Ma) and seismic methods (present).


Introduction
Mantle-derived fluids (e.g., He, CO 2 , N 2 ) provide a key means of learning about the mass and energy transfer between the surface of the Earth and its interior (Hilton, 2007), reveal fundamental geological processes, that is, volcanic eruptions, earthquakes (Sano et al., 2014) and volatile recycling, and pinpoint new geothermal resources (Kennedy et al., 1997).Among mantle-derived volatiles, helium isotopes ( 3 He/ 4 He) provide an unambiguous measure of the mantle degassing signature due to dominantly primordial 3 He in the mantle and radiogenic 4 He produced by 235,238 U and 232 Th decay in the crust (Ballentine & Burnard, 2002).Although 3 He/ 4 He values in different mantle sources differ, for example, 8 ± 1 R A (where R A is the air 3 He/ 4 He ratio of 1.39 × 10 6 ) for depleted mid-ocean ridge basalt (MORB)-type mantle (DMM, Graham, 2002), 6.1 ± 2.1 R A for subcontinental lithospheric mantle (SCLM, Day et al., 2015), and >25R A for ocean island basalt (OIB)-type mantle (Day, 2022), they are significantly higher than those in the crust (∼0.01-0.05R A , with a typical value of 0.02 R A ). Regions with mantle volatile degassing can be clearly identified according to the sensitivity of 3 He/ 4 He ratios to the presence of even <1% mantle helium (O'Nions & Oxburgh, 1988).In Cenozoic volcanic areas, such as mid-ocean ridges (Jean-Baptiste et al., 2004), volcanic arcs in subduction zones (the Central American system in Nicaragua and Costa Rica, Shaw et al., 2003), and hot spots (Hawaii, Sano & Fischer, 2013), volatile escape with the help of ascending magma makes it straightforward to identify helium sources and flow paths (Bekaert et al., 2021;Sano & Fischer, 2013).However, the occurrence of mantle-derived helium in continental regions lacking recent volcanism is less straightforward (Hilton, 2007).Several studies suggest that the occurrence of mantle-derived helium is closely related to extensional tectonics with mantle melt intrusion to shallow crust (Torgersen, 1993), enhanced permeability induced by an increase in dextral shear strain (Kennedy et al., 1997), and lithospheric strike-slip faults (Klemperer et al., 2013), which could provide migration pathways for mantle-derived helium transport from deep mantle-depth to surface.From the aspect of the mantle-derived He release, Hiett et al. (2021) suggested that slab-derived fluids could mobilize helium from the subcontinental lithospheric mantle in an active flat-slab setting (Hiett et al., 2021).Quantitative evaluation of these processes, such as the timescale of mantle-derived fluid migration through the crust and the flux of mantle-derived He, is critical for developing new models or testing existing models for the geodynamic evolution of Earth's continental lithosphere (Hiett et al., 2021;Kennedy et al., 1997;Torgersen et al., 1995;Zhang et al., 2022).
The Tibetan Plateau (TP) was formed by the India-Asia collision since ∼55 Ma and subsequent continental lithosphere subduction (Nábĕlek et al., 2009;Shi et al., 2015;Tapponnier et al., 2001;Taylor & Yin, 2009).The widespread hot springs release large quantities of hydrothermal volatiles (e.g., He, CO 2 , N 2 ) (Hao et al., 2023;Liao, 2018;Wang et al., 2022).Numerous studies of hydrothermal volatiles have detected 3 He/ 4 He ratios >0.1 R A and suggest that mantle contributions are unequivocally present in the TP (Becker, 2006;Hao et al., 2023;Hoke et al., 2000;Klemperer et al., 2013Klemperer et al., , 2022;;Marty et al., 1996;Newell et al., 2008;Sun et al., 2020;Walia et al., 2005;Yokoyama et al., 1999;Zhang, Zhang, et al., 2021;L. Zhang et al., 2017;M. Zhang et al., 2017;Zhao et al., 2022).However, due to the lack of recent volcanic activity (<8 Ma), there is a large controversy regarding the mantle-derived He origin, that is, earlier magmatism, Quaternary mantle-derived melt intrusion into the shallow crust (Hou & Li, 2004;M. Zhang et al., 2017), partial melting and degassing of carbonate-enriched mantle induced by slab rollback (Zhao et al., 2022), and direct input from the asthenosphere (Klemperer et al., 2022).In addition, previous studies have endeavored to use mantle-derived He as another method distinct from geophysical and petrologic studies to obtain a snapshot of the lithospheric structures and constrain the geodynamic mechanisms governing the formation and evolution of the TP (Hoke et al., 2000;Hou & Li, 2004;Klemperer et al., 2022;M. Zhang et al., 2017).However, less attention is focused on the timescales for mantlederived He transiting the thick Tibetan crust, which are critical for the interpretation of the lithospheric structure revealed by 3 He/ 4 He ratios and for comparing them with those revealed by petrologic studies (>8 Ma) and geophysical studies (present).No studies have attempted to systematically quantify the flow rate, and mantlederived 3 He flux of geothermal fluids in this typical tectonic region before using 3 He/ 4 He ratios to reveal geodynamic mechanisms, except for the estimates of hydrothermal helium fluxes in the Simao block with Quaternary volcanoes, Southeast Tibetan Plateau margin (Zhang et al., 2022).
In this study, 10 newly acquired data and 286 previously published data of helium concentrations and isotopes in hot springs were used to constrain the helium origin and migration pathways by linking the 3 He/ 4 He distribution to geological, geophysical and petrological data.The migration velocity, transit time and fluxes of mantle-derived helium in southern Tibet and the Himalayas were systematically estimated based on a one-dimensional steadystate flow model, and compared to other tectonically active fields and volcanic regions.Combining these qualitative and quantitative estimations with petrologic studies (>8 Ma) and geophysical studies (present), we provide geochemical constraints on the formation time of shape of the Indian lithospheric structure.This would also contribute to a better quantitative understanding of the mantle-derived helium degassing in continental regions lacking recent volcanism and its role in the global helium budget.

Study Area
The study area extends from southern Tibet to the Himalayas and across the Indus-Yarlung suture zone (IYS) (Figure 1).Southern Tibet (Lhasa terrane) has a southern boundary of the IYS and a northern boundary of the Bangong-Nujiang suture.The Precambrian crystalline basement is locally distributed in the northern Lhasa terrane, dominated by metamorphic rocks of the Nyainqentanglha Group and Amdo gneiss (Yin & Harrison, 2000).The sedimentary cover in the Lhasa terrane consists of shallow sea clastic sedimentary sequences (Yin & Harrison, 2000).The Gangdese magmatic belt formed during the India-Asia continental collision extends 2,000 km from east to west, and 60% is concentrated in the southern Lhasa terrane (Mo, 2020).The Himalayan orogen to the south of the IYS consists of the Lesser Himalayan Sequence, the Greater Himalayan Sequence and the Tethyan Himalayan Sequence from south to north (Figure 1).

Sampling and Methods
Eleven free gas samples were collected from natural springs in the form of warm springs to boiling springs and geysers in southern Tibet and the Himalayas (Figure 1).Springs exhibit a wide range of temperatures from 40 to 87°C.A multiparameter water quality analyzer (HANNA HI-98194) was used to measure the electrical conductivity (EC) and pH of the hot spring waters on site.The gas samples were collected by the gas drainage method (Sano & Fischer, 2013).An inverted funnel with a tube was immersed in the vent of the hot spring and completely washed with geothermal fluids for >30 min.Then, the outlet of the tube was inserted into 50 mL lead glass bottles, which had been filled with hot spring water.The detailed procedures of gas collection are described in Hao et al. (2020).
The gas compositions and He-Ne isotopic compositions were measured at the Oil and Gas Research Center, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences.The bulk gas compositions of the major gases (N 2 , CO 2 , CH 4 , O 2 , and Ar) were measured using a MAT 271 mass spectrometer.The detection limit and relative standard deviation were 0.0001% and <2%, respectively.The relative abundance and isotopic compositions of He and Ne were analyzed by a Noblesse noble gas mass spectrometer (Nu Instruments, UK) with an atmospheric standard from the Gaolan Hill area south of Lanzhou (Cao et al., 2018(Cao et al., , 2020;;Zhang, Guo, et al., 2021;Zhang, Xu, et al., 2021).The detailed analytical procedures are described in Cao et al. (2018) and Li et al. (2020).The 4 He, 20 Ne and 22 Ne were determined by a Faraday collector. 3  The tectonic framework, lithology, and gas sample sites of the study area.The gas sample collection sites in this study and from the literature (Becker, 2006;Hao et al., 2023;Hoke et al., 2000;Klemperer et al., 2013Klemperer et al., , 2022;;Marty et al., 1996;Newell et al., 2008;Sun et al., 2020;Walia et al., 2005;Yokoyama et al., 1999;Zhang, Zhang, et al., 2021;L. Zhang et al., 2017;M. Zhang et al., 2017;Zhao et al., 2022)

Results
The entire dataset includes a total of 315 gas samples, of which 11 new data points in the Himalayas were collected in this study (Figure 1, Tables 1 and 2), and 304 data points in the southern Tibet and Himalayas were collected from the literature (Tables S1 and S2 in Supporting Information S2; Yokoyama et al., 1999; Hoke He ratios of samples are normalized to the 3 He/ 4 He ratio of air (R A = 1.39 × 10 6 ).c X-factor = ( 4 He/ 20 Ne) measured /( 4 He/ 20 Ne) air × ß Ne /ß He ; ß represents the Bunsen coefficient from Weiss (1971), assuming a groundwater recharge temperature of 10°C.(Hilton, 1996).e Samples with 3 He/ 4 He ratios < 0.42 R A and an X-factor < 10 or samples with 3 He/ 4 He ratios > 0.42 R A and an X-factor < 4 are too contaminated by air-derived helium to correct and are not included in further discussion.et al., 2000;Newell et al., 2008;Klemperer et al., 2013;Zhang, Guo, et al., 2021;Zhang, Xu, et al., 2021;Zhang, Zhang, et al., 2021;L. Zhang et al., 2017;M. Zhang et al., 2017;Sun et al., 2020;Zhao et al., 2022;Klemperer et al., 2022;Hao et al., 2023).
The gas mixtures from this study and the literature in the study area are dominated by CO 2 and N 2 , with 78% of the total gas samples being CO 2 -rich (CO 2 /N 2 > 1, Figure 2a).He concentrations exhibit a wide range from 0.05 ppm to 2.31 vol.%.The measured 3 He/ 4 He ratios (R M ) range from 0.02 R A to 0.37 R A for this study (Table 2) and from 0.009 R A to 2.23 R A for literature (Table S2 in Supporting Information S2).Considering that the input of airderived helium during sampling or analysis may change the original 3 He/ 4 He values, the X-factor calculated by air-normalized measured 4 He/ 20 Ne ratios and Bunsen coefficients of Ne (ß Ne ) and He (ß He ) (Weiss, 1971) at an assumed groundwater recharge temperature of 10°C (X-factor=( 4 He/ 20 Ne) measured /( 4 He/ 20 Ne) air × ß Ne /ß He ) is used to recognize and correct air contamination (Hilton, 1996).The X-factor of our samples (except sample G10) is higher than 10, indicating an insignificant air-derived helium contamination in our geothermal fluids (Figure 3, Zhang, Xu, et al., 2021).Eighteen samples with 3 He/ 4 He ratios <0.42 R A and an X-factor <10 or with 3 He/ 4 He ratios >0.42 R A and an X-factor <4, collected from the literature, are too contaminated by air-derived helium to correct and are not included in further discussion (Table S2 in Supporting Information S2).The 3 He/ 4 He ratios corrected by air contamination (R C ) of gas samples are between 0.02 and 0.17 R A for this study (n = 10), within the range of previously reported data (0.003 R A and 2.24 R A , n = 286, Figure 3).The air-corrected helium concentration ([He] C ) of gas samples ranges from 0.7 to 2,823 ppm for this study and from 0.06 ppm to 2.31 vol.% for the total dataset.Because the amounts of 4 He produced by the 235,238 U and 232 Th decay in the crust are approximately 8 orders of magnitude higher than those of nucleogenic 3 He produced by the reaction 6 Li(n,α) 3 H (β-) 3 He, samples with R C > 0.1 R A generally reflect an unambiguous addition of mantle helium (Ballentine et al., 2002).Based on the 3 He/ 4 He ratios, total gas samples can be divided into the "mantle-derived 3 He domain"  2b).Combined with previously reported data, the CO 2 / 3 He ratios of gas samples exhibit a wide range from 2.13 × 10 6 to 1.50 × 10 14 (9.35 × 10 9 to 4.47 × 10 12 for this study), extending beyond the values of reference materials of MORB-type mantle (7.5 × 10 8 to 3 × 10 9 ) and crust (10 12 to 10 14 ) (Sano & Marty, 1995).The 3 He/ 4 He ratios exhibited no clear relationship with CO 2 /N 2 ratios (Figure 2a) and CO 2 / 3 He ratios (Figure 2c).However, the MG group has higher proportions of CO 2 -rich type gases and CO 2 / 3 He values than those of the CG group.

Mantle-Derived Helium Sources
Helium isotopes indicate that ∼36% of the total samples have unambiguous mantle helium contributions (R C > 0.1R A , Figure 3).However, the lack of magmatism since 8 Ma (Guo et al., 2015) makes the mantle source ambiguous.The Indian lithosphere is regarded as an implausible source for mantle degassing due to its old (Paleoproterozoic to Archean in age) and cold properties (<700°C) (Klemperer et al., 2022).Possible candidates exist for mantle-derived helium in the TP with thick crust (50-80 km) according to geological conditions and previous studies, including residual mantle-derived He from earlier magmatism, Quaternary mantle-derived melt intrusion into the shallow crust, partial melting and degassing of carbonate-enriched mantle induced by slab rollback and direct input from asthenosphere (Hoke et al., 2000;Kennedy et al., 1997;Klemperer et al., 2022;Zhao et al., 2022).

"Fossil" Helium Sources
"Fossil" helium sources are residual mantle-derived He in earlier magmatism that is preserved in (a) magmatic rocks during rock crystallization in the crust and (b) residual magmatic fluids in the crust (Burnard et al., 2012;Méjean et al., 2020).Scenario (a) has been systematically discussed and ruled out by Zhao et al. (2022) based on the magma-aging model (Figure 3).Here, we further discuss the scenario (b) in this article.If the residual mantle-derived He is preserved as residual magmatic fluids in the shallow crust (i.e., CO 2 reservoirs containing mantle-derived He, Burnard et al., 2012;Gilfillan et al., 2009), the initial 3 He/ 4 He ratios of the magmatic fluids will be diluted by radiogenic 4 He and nucleogenic 3 He released from the rocks into the fluids since the emplacement time (Ballentine & Burnard, 2002): where R f is the final 3 He/ 4 He ratio in the fluid after the addition of in situ nucleogenic 3 He and radiogenic 4 He produced in the crust since the emplacement time (t).R i,mantle and [ 4 He] i,mantle are the initial 3 He/ 4 He ratio and 4 He concentration (cm 3 STP g 1 H 2 O) in the mantle-derived fluid at t = 0. A( 3 He) and A( 4 He) are the accumulation rates of nucleogenic 3 He and radiogenic 4 He in the fluid (cm 3 STP g 1 H 2 O year 1 ), which can be calculated by Equations 2-4 (Andrews & Lee, 1979;Castro, 2004;Zhou & Ballentine, 2006): where ρ s and ρ f represent the crust and fluid densities in g cm 3 , respectively; φ is porosity; P( 3 He) and P( 4 He) are the 3 He and 4 He production rates of crustal rocks in cm 3 STP g 1 year 1 , respectively; [Li], [U] and [Th] are the Li, U, and Th concentrations in the crustal rock in ppm, respectively.Given the geological conditions that allow magmatic fluids to be isolated and immoveable for at least 8 Ma, a generous porosity of 10% is used to represent the middle to upper crust (Hiett et al., 2021), although the higher porosities of the upper crust are found in the upper most crust in the TP (Shen, 1992).The rock density and fluid density are 2.8 g cm 3 and 1 g cm 3 (Kulongoski et al., 2013), respectively.Based on an average crustal composition with Li content of 16 ppm, U content of 1.3 ppm, Th content of 5.6 ppm, and Th/U of 4.3 (Rudnick & Gao, 2003), the average 3 He production and 4 He production rates in the crust are 2.54 × 10 21 cm 3 STP g 1 rock year 1 and 3.17 × 10 13 cm 3 STP g 1 rock year 1 , respectively.Accordingly, 3 He and 4 He accumulation rates in fluids calculated from Equation 2 are of 6.4 × 10 20 cm 3 STP g 1 H 2 O year 1 and 8.0 × 10 12 cm 3 STP g 1 H 2 O year 1 , respectively.The initial 3 He/ 4 He [ 4 He] i,mantle = [ 4 He] s × F 4 He) (5) where F( 4 He) is the mantle 4 He fraction; R m , R c , and R s are the 3 He/ 4 He ratios of the mantle (8 ± 1 R A ), crust (0.02 R A ) and samples (R s ), respectively; and [ 4 He] s is the 4 He concentration in the geothermal water (cm 3 STP g 1 H 2 O).Details of dissolved 4 He concentrations in hot springs, and mantle-derived 4 He fraction and concentration calculations are available in Table S3 in Supporting Information S2.Considering the gas samples could have experienced some degassing (Hiett et al., 2021), the highest mantle-derived 4 He concentration of 1.6 × 10 6 cm 3 STP g 1 H 2 O from ZDX14 ZP01-ZK4001 (R = 0.26 R A , [He] = 3,476 ppm) is adopted for [ 4 He] i, mantle as a more conservative option.These calculations yield present-day 3 He/ 4 He ratios of 0.06-0.23R A for a magmatic fluid preserved in the crust since the latest magmatic episode in the southern Tibet (from 25 Ma to 8 Ma, Guo and Wilson, 2019) (Figure 3a).Although a portion of the data could plausibly be explained by this mechanism, it cannot explain the larger dataset, including samples with 3 He/ 4 He values higher than 0.23 R A .Especially, the highest 3 He/ 4 He value of 2.24 R A observed in the hot springs is approximately one order of magnitude higher than the highest estimated 3 He/ 4 He ratio in the residual magmatic fluid (Figure 3b).Furthermore, these are conservative estimates since the mixing of crustal fluids and magmatic crustal assimilation will further lower the 3 He/ 4 He ratios in the residual magmatic fluid (Hiett et al., 2021;Hu et al., 1998).Thus, we generally discount this mechanism as the dominant source of mantle-derived He.

Quaternary Mantle-Derived Melt Intrusion Into the Shallow Crust
The intrusion of mantle-derived melts at shallow depths in the crust (Hoke et al., 2000) was proposed based on geophysical observations of low-velocity zones (bright spots) at depths of 15-20 km in the crust in the northern Yadong-Gulu rift (Makovsky & Klemperer, 1999), where a mantle helium signature occurs (Figure 4a).The lowvelocity zones generally reflect the existence of magma chambers/aqueous fluids (Wei et al., 2001).However, with more geophysical investigations, the low-velocity zones were found to be widely scattered throughout  (1996).Crustal thickness data are from the interpolated crustal thickness grid of uniformly spaced data points, each 0.25°× 0.25° (Li et al., 2014).
southern Tibet and the Himalayas (Hetényi et al., 2015), covering both the "mantle-derived 3 He domain" and "crust-derived 4 He domain" regions (Figure 4a).No clear spatial relation was found between high 3 He/ 4 He ratios (R C > 0.1R A ) and the low-velocity zone observations in the region between 86 and 92°E.In addition, the increase in 3 He/ 4 He ratios northward along the northern segment of the Yadong-Gulu rift cannot be explained by the depth trends of the Angang (ABS), Yangbajing (YBS), Nyingzhong (NBS), and Damxung (DBS) bright spots (e.g., the depths of bright spots in the ABS and YBS are shallower than those in the NBS and DBS) (Figure 4b).Therefore, magmatic melts in the shallow crust cannot account for mantle-derived helium.Gas samples with crustal helium isotopes (R C ≤ 0.1 R A ) are widely distributed throughout southern Tibet and the Himalayas, suggesting that the magma chamber in the shallow crust was most likely formed from crustal partial melting.
Consequently, the occurrence of mantle-derived helium signatures in the study area may be attributed to the direct influx of mantle fluids from the mantle depths (>80 km), as suggested by Klemperer et al. (2022) and Zhao et al. (2022).Considering the uncertainty regarding the relative importance of an asthenospheric mantle versus Tibetan lithospheric mantle contributions, the mantle He proportions in the study area are 0%-27.8%as estimated by mixing asthenospheric source (DMM, 8 ± 1 R A ) and crust; alternatively, they are 0%-36.5% as estimated by assuming mixing of crust and the subcontinental lithospheric mantle (SCLM, 6.1 ± 2.1 R A ; Day et al., 2015) (Figure 3, Table 2; Table S2 in Supporting Information S2).

Migration Pathways
The widespread presence of mantle-derived He in the India-Asia continental collision settings suggests the existence of extensive pathways for fluids to be transported from the mantle depth through the crust, which may be related to the widespread deformations in the boundary and interior of the TP (i.e., east-west extension and strikeslip faults; Li & Song, 2018;Royden et al., 1997).On a regional scale, the mantle degassing feature of the MG is partly associated with deformation in Tibet.Samples with >10% DMM-derived (or 13% SCLM-derived) He contributions are located within 15 km of rifts (3.6 km for ZNMA23, 6.4 km for JQ1501, 3.3 km for 20WB01, 0.4 km for ZNMA24) (Figures 5a and 5b) or strike-slip faults (∼13 km for ZGR13) (Figure 5c).In particular, three samples with extremely high 3 He/ 4 He ratios (R C > 1R A ) are near the intersection of Yadong-Gulu rifts and the right-slip fault, the intersection of Tangra Yum Co rift and the right-slip fault, and the Karakorum fault (Figure 4).These results are comparable to those of other fault systems in continental regions lacking recent volcanism, that is, enrichment of mantle-derived He within 0-30 km of the San Andreas Fault System (SAFS) (Kennedy et al., 1997;Kulongoski et al., 2013) and North Anatolia Fault System (NAFZ) (Dŏgan et al., 2009).These fault regions are associated with higher permeability, thereby increasing fluid vertical transport velocities and resulting in less dilution of mantle-derived helium by crustal radiogenic 4 He addition.

The Upward Fluid Flow Rates and Transit Times for Mantle-Derived Helium
The observed 3 He/ 4 He ratios of MG samples lower than those of DMM or SCLM are generally caused by dilution of radiogenic 4 He produced in the thick crust as the mantle-derived fluids transport the continental crust or mix with 4 He-rich crustal fluids at the shallow crustal depths (Hiett et al., 2021;Kennedy & van Soest, 2006;Kennedy et al., 1997).Shallow secondary processes occur in the geothermal waters, as indicated by most samples beyond the mixing zone of CO 2 / 3 He and 3 He/ 4 He (Figure 2c).The lower CO 2 / 3 He ratios are generally attributed to calcite precipitation, dissolution reactions, and the input of He-rich N 2 -type crustal fluids in the shallow crust, which have been discussed in detail in previous studies (Hao et al., 2023;M. Zhang et al., 2017;Zhao et al., 2022).The lack of correlations between the 3 He/ 4 He ratios and the chemical characteristics of geothermal fluids (e.g., CO 2 / 3 He and CO 2 /N 2 ) suggests that mixing of 4 He-rich fluids at shallow crustal depths is not the dominant cause of the lower 3 He/ 4 He ratio.Instead, lower 3 He/ 4 He ratios of MG samples than the mantle endmembers and the variations in 3 He/ 4 He ratios among different samples likely reflect the transit times for mantle-derived helium through the continental crust, which can be estimated by Equations 1 and 2. In this scenario, the final 3 He/ 4 He ratios of the geothermal fluid are those of the samples.Then, the transit time (t) can be expressed as Equation 7: Geochemistry, Geophysics, Geosystems 10.1029/2023GC011297 HAO ET AL.
When the fluids transit the crust thickness (H C ), the time-averaged upward flow rate (q) can be calculated based on a one-dimensional steady-state flow model (Kennedy et al., 1997) (named Equation 8 method 0 to distinguish it from the other methods in the next section): Due to the uncertainty regarding the relative importance of DMM and SCLM, the starting fluid is assumed to have the 3 He/ 4 He ratios of SCLM (6.1 ± 2.1 R A ) as a conservative option.The initial 4 He concentration in the mantle fluid ([ 4 He] i,mantle ) is 2.7 × 10 7 cm 3 STP g 1 H 2 O, which is calculated from the air-corrected 3 He/ 4 He ratios and helium concentrations of our 20 least degassed samples with the highest He concentrations (1.0 × 10 6 to 5.4 × 10 5 cm 3 STP g 1 H 2 O) (Hiett et al., 2021;Kennedy et al., 1997; Table S3 in Supporting Information S2).
The crustal thickness of each sampling site is obtained from a crustal thickness map of the Chinese mainland generated from teleseismic receiver functions (Li et al., 2014), ranging from 58 to 77.9 km (Figure 4a).Two reasonable porosity parameters of 1% (an extreme upper limit for the porosity in the deep crust) and 10% (a generous porosity in the middle to upper crust) were used to model the crustal conditions (Hiett et al., 2021;Kennedy et al., 1997).The calculated upward flow rates of mantle-derived fluids for southern Tibet and the Himalayas range from 332 to 11,700 mm/year with porosities of 1% and from 30 to 1,064 mm/year with porosities of 10% (Table S4 in Supporting Information S2).It should be noted that the fluid migration through the ductile lower crust is much slower than the migration through the upper crust; estimates based on 1% show greater possibility than those based on 10%.The high 3 He/ 4 He ratios in the study area are related to the high upward flow rates (Figure 6d; Figure S1b in Supporting Information S1), which correspond to transit times ranging from 5.3 ka-2.3Ma.As the model here only considers continuous steady-state He release, and the actual flow rates could be significantly higher if the He is episodically released (Kennedy & van Soest, 2006;Kennedy et al., 1997; Zhang et al., 2022).As such, the calculated results based on the model are conservative.In addition, there is a fair degree of uncertainty due to the lack of consideration of mixing with other radiogenic 4 He-rich fluids from the shallow crust in the model, which indeed occurs as mentioned above but is challenging to quantify (Hao et al., 2023;Zhang et al., 2022).

Differences in Different Methods for Calculation of Upward Fluid Flow Rates
Before comparing the flow rates in the study area with other similar regions, it is essential to assess the discrepancies produced using different methods.Two methods were developed for the estimation of the flow rates and volatile flux based on a steady-state one-dimensional flow model in previous studies (Hiett et al., 2021;Kennedy et al., 1997).Method 1 expressed as Equation 9 (Kennedy et al., 1997) has been widely used to calculate upward rates in the San Andreas Fault System in California (2-147 mm/year, Kennedy et al., 1997;Kennedy and van Soest, 2006;Kulongoski et al., 2003Kulongoski et al., , 2005Kulongoski et al., , 2013)) 6a and 6b).Method 2, expressed as Equation 10, was used by Hiett et al. (2021), and yields an upward flow rate of 49-771 mm/year for the Peruvian flat slab (Table S5 in Supporting Information S2; Figures 6a and 6b).(Kennedy et al., 1997;Kulongoski et al., 2003Kulongoski et al., , 2005Kulongoski et al., , 2013)); NAFS: North Anatolia Fault System, Turkey (de Leeuw et al., 2010); SB: Simao block, southeastern Tibetan Plateau margin (Zhang et al., 2022); DV: Dixie Valley, Basin and Range Province, USA (Kennedy & van Soest, 2006); DFS: Denali fault system, Alaska, USA (Newell et al., 2023); PFS: Peruvian flat slab, South America (Hiett et al., 2021); AF: Alpine Fault, New Zealand (Menzies et al., 2016). Method To illustrate the influences of adopting different methods, the calculated upward flow rate is plotted as a function of 3 He/ 4 He ratios assuming scenarios with two different crustal porosities (1%, Figures 6d and 10%, Figure S1a in Supporting Information S1).The calculations are based on a typical crustal thickness of 30 km, a 4 He crustal production rate (P( 4 He)) of 3.17 × 10 13 cm 3 STP g 1 rock year 1 (Rudnick & Gao, 2003), a bulk rock density (ρ s ) of 2.8 g cm 3 , a fluid density (ρ f ) of 1 g cm 3 , R i,mantle of 8 R A and R c of 0.02 R A .The initial mantle 4 He concentration ([ 4 He] i,mantle ) is 7.8 × 10 7 cm 3 STP g 1 H 2 O using the average value of those in the literature (Table S5 in Supporting Information S2).The flow rates calculated based on different methods with the same input parameters exhibit an apparent difference, which could be explained by ignoring different model parameters in the previously developed methods, for example, porosity for method 1, the density of fluid and rock for method 2. The results calculated using method 0 (this study) are approximately 1-2 orders of magnitude higher than those obtained from method 1, and the differences decrease with increasing porosity.The discrepancy between the results calculated using method 0 and method 2 is relatively small, within 1 order of magnitude, and related to the bulk rock density.Compared with method 1 and method 2, the method 0 developed in this study takes into account the porosity of crust and the density of fluid and crust, which can reflect the natural geological conditions more accurately.Therefore, the previously estimated flow rates of the mantle-derived fluids by different methods were calibrated by method 0 with porosities of 1% (Figures 6d) and 10% (Figure S1b in Supporting Information S1) for purposes of comparison.Taking the results of a porosity of 1% as an example, the highest flow rate of mantlederived fluids in southern Tibet and the Himalayas (11,700 mm/year) is comparable to those estimated for SAFS (14,553 mm/year), and higher than those estimated for the Dixie Valley, North Anatolian Fault System (DV, 693 mm/year).If the flow rate of the mantle-derived fluids in the Tibetan Plateau is estimated using the widely used method 1, it is up to 118 mm/year.It is close to the reported speed limit of mantle-derived fluids through lithospheric-scale Denali fault system estimated using method 1 (160 mm/year, Newell et al., 2023) and exhibits a similar comparison pattern between different regions as those estimated using method 3.This implies that the 3 He/ 4 He values in Tibet lower than those of SAFS without recent volcanic activity are mainly attributed to the greater thickness of the crust, as the crustal thickness of the study area is nearly twice that of the latter (Figure 6b).In addition, the highest flow rates of mantle-derived fluids in southern Tibet and the Himalayas are lower than those estimated for the Simao block in the southeastern Tibetan Plateau margin (56,133 mm/year) with mantle helium fed by recent magma recharge beneath Quaternary volcanoes (Zhang et al., 2022).
A diagram of continental 4 He degassing fluxes proposed by Torgersen (2010) noted increased mantle 4 He flux with an increase in crustal 4 He flux but lower mantle contributions of the "tectonically active field" relative to those of the "volcanic field" (Figure 7b; Figure S2b in Supporting Information S1; Torgersen, 2010).However, this diagram was largely built on the basis of 4 He degassing fluxes from lakes and groundwater basins worldwide (Ballentine et al., 2002;Kipfer et al., 2002;Torgersen, 2010).In this study, we further compiled the helium flux of hydrothermal fluids in volcanic regions (e.g., Yellowstone, Lowenstern et al., 2014;Southeast Tibetan Plateau margin, Zhang et al., 2022;North Anatolian Fault System in Turkey, de Leeuw et al., 2010) and fault tectonically active regions (e.g., San Andreas Fault System in USA, Kennedy et al., 1997;Kulongoski et al., 2003Kulongoski et al., , 2005Kulongoski et al., , 2013;; Alpine Fault in New Zealand, Menzies et al., 2016;Southeast Tibetan Plateau margin, Zhang et al., 2022; Dixie Valley in USA, Kennedy and van Soest, 2006; Peruvian flat slab in South America, Hiett et al., 2021; Denali fault system in USA, Newell et al., 2023; southern Tibet and the Himalayas, this study) (Tables S4 and S6 in Supporting Information S2; Figure 7b; Figure S2b in Supporting Information S1).Samples from hydrothermal systems in southern Tibet and the Himalayas as well as other nonvolcanic hydrothermal systems span 5 orders of magnitude in both axes and plot within the upper part of the typical "tectonically active field" and the lower part of the "volcanic field" in the previous diagram proposed by Torgersen (2010), suggesting higher mantle helium  (Torgersen, 2010) and natural gas (Sano et al., 1986).The modified "tectonically active field" (the yellow area) is built by the addition of continental hydrothermal data from volcanic regions (e.g., Yellowstone, Lowenstern et al., 2014;Southeast Tibetan Plateau margin, Zhang et al., 2022;North Anatolian Fault System in Turkey, de Leeuw et al., 2010) and tectonically active regions (e.g., Peruvian flat slab in South America, Hiett et al., 2021; San Andreas Fault System in USA, Kennedy et al., 1997;Kulongoski et al., 2003Kulongoski et al., , 2005Kulongoski et al., , 2013 contributions for the hydrothermal systems in the "tectonically active field."Thus, the "tectonically active field" in the diagram proposed by Torgersen (2010) was modified to cover the helium fluxes of nonvolcanic hydrothermal systems (Figure 7b; Figure S2b in Supporting Information S1).As Figure 7 shows, mantle 3 He and 4 He fluxes in southern Tibet and the Himalayas are comparable to those of lakes and groundwater (Torgersen, 2010) and other similar nonvolcanic hydrothermal systems in tectonically active regions (10 1 -10 7 atoms m 2 s 1 for mantle 3 He flux and 10 6 -10 12 atoms m 2 s 1 for mantle 4 He flux), but lower than those measured in lakes (Torgersen, 2010) and hydrothermal systems in volcanic fields (10 4 -10 9 atoms m 2 s 1 for mantle 3 He flux and 10 9 -10 14 atoms m 2 s 1 for mantle 4 He flux) (Figure 7b; Figure S2b in Supporting Information S1).The mantle 4 He flux in southern Tibet and the Himalayas is much higher than that of tectonically stable continental regions (≤3 × 10 7 atoms m 2 s 1 , O'Nions & Oxburgh, 1988) (Figure 7b; Figure S2b in Supporting Information S1).All of these results show that hydrothermal fluids in southern Tibet and the Himalayas as well as other nonvolcanic continental regions are characterized by important mantle helium outgassing.

Constraints on the Lithospheric Structure and Tectonic Processes
Based on the boundary between the "mantle-derived 3 He domain" and the "crustal-derived 4 He domain," Klemperer et al. (2022) determined the position of the mantle suture, where a cold underplated Indian lithosphere is juxtaposed at >80 km depth against a sub-Tibetan incipiently molten asthenospheric mantle wedge (Figure 8a).The mantle suture location is also confirmed and slightly modified by the data of the region between ∼86°E and ∼95°E in this study (Figure 8a).However, the interpretations of 3 He/ 4 He ratios lack consideration of the timescales for mantle fluids transiting the thick Tibetan crust (Klemperer et al., 2022), which reflects the time of the structure state revealed by 3 He/ 4 He ratios.The high flow rates of mantle-derived helium in the study area correspond to transit times ranging from 5.3 ka-2.3Ma, which suggests that the mantle fluids in southern Tibet and the Himalayas can present a state of the Indian lithosphere between those revealed by potassic-rich mafic rocks (25-8 Ma) (Guo & Wilson, 2019) and seismic imaging techniques (present day) (Chen et al., 2015;Kosarev et al., 1999;Li & Song, 2018;Nábělek et al., 2009;Shi et al., 2015) (Figure 8a).
The southern boundary of the "mantle-derived 3 He domain" is stepwise, located 33-115 km north of the IYS for 86-91°E and 135-185 km north of the IYS for 81-86°E, while it is ∼100 km south of the IYS at ∼91°E: this is close to the suggested mantle suture locations based on geophysical observations (Figure S3 in Supporting Information S1), that is, at ∼31°N at ∼85°E (∼175 km north of the IYS, Nábělek et al., 2009) and ∼30°N at ∼90°E (∼50 km north of the IYS, Kosarev et al., 1999) along the INDEPTH profile of central Tibet, ∼50 km south of the IYS at 92°E proposed based on a passive-source seismic experiment on eastern Tibet (Shi et al., 2015).The distinct west-east variations in mantle suture locations revealed by spatial variations in 3 He/ 4 He values suggest a fragmented Indian mantle lithosphere, which may be attributed to tearing of the Indian mantle lithosphere along the longitudinal boundary at ∼86°E and ∼92°E and undergoing flatter subduction in the middle and steeper subduction in the east (Chen et al., 2015;Klemperer et al., 2022;Li & Song, 2018).Because the kinematic characteristics of the subducted lithosphere are mainly controlled by the dip angle (Chen et al., 2015), the different locations of the mantle suture may reflect the distinct subduction angle and advancement of the slab beneath the middle (84-91°E) and eastern (>91°E) regions of the plateau (Figure 8b).The mantle suture beneath the central plateau reaching 33-185 km north of the IYS corresponds to flatter subduction of the Indian mantle lithosphere slab (which subducts with a shallow angle or subhorizontal underthrusting beneath southern Tibet), as movement at a low angle has a major horizontal velocity component and extends farther north (Chen et al., 2015).The mantle suture beneath the eastern plateau located ∼100 km south of the IYS may reflect that the Indian mantle lithosphere subducts steeply into the mantle, which causes a major vertical movement component and extends a shorter distance to the north.The low coverage of 3 He/ 4 He ratios in the region west of 84°E cannot give a single interpretation of lithologic structures, which have been mentioned by Klemperer et al. (2022).But mantle-derived 3 He can be traced at ∼13 south of the IYS, which seems to support seismic interpretations of a tearing of the Indian lithosphere and/or the Indian mantle lithosphere with a steeper subduction (Zhang et al., 2014).Potassic and ultrapotassic magmatism (25-8 Ma) is thought to have been triggered by Indian slab breakoff, detachment and tearing based on its spatiotemporal distribution, geochemical data and modeling simulations (Bian et al., 2020;Guo & Wilson, 2019;Guo et al., 2013Guo et al., , 2015)).The southern boundary of the mantle-derived 3 He overlaps with the northern limit of the K-rich mafic rocks in the region between 82°E and 90°E (Figure 8a), which may indicate that the current subduction state (tearing and nonuniform subduction) of the middle part of the Indian mantle lithosphere was formed before ∼8 Ma.Although the 3 He/ 4 He values in Tibet are lower than those in the San Andreas Fault System without recent volcanic activity, they have a similar mantle-derived 3 He migration velocities and mantle-derived 3 He fluxes.This suggests that the similar mechanisms control the mantle-derived 3 He migration in both Tibet and the San Andreas Fault System, while the crustal thickness of Tibet (∼70 km) is nearly twice that of the latter (∼30 km), causing lower 3 He/ 4 He ratios.In the southern Tibet, which lacks active magmatism, the widespread east-west extension and eastward motion (Royden et al., 2008) may induce enhanced permeability extending to the mantle depths, sustaining mantle fluids that can penetrate the ductile lithosphere (Hilton, 2007).The migration velocities of mantle-derived 3 He span 3 orders of magnitude and reflect the heterogeneity of the subsurface permeability or the different scales of faults (Figure 8b).In particular, the intersection between the Yadong-Gulu rifts and Tangra Yum Co rift and right-slip faults or the Karakorum fault has higher permeability in the deep crust, as indicated by the highest 3 He/ 4 He ratios and high migration velocities.

Conclusions
We constrained the origin of mantle-derived helium and quantitatively evaluated upward fluid flow rates, the timescale of mantle-derived fluids through the crust, and the flux of mantle-derived He.These quantitative estimations provide insights into the lithospheric structure and activity intensity of tectonic geodynamics along the India-Asia continental collision settings.The following points are emphasized.
Mantle-derived volatiles in the southern Tibet and the Himalayas are more likely attributed to the direct injection of mantle-derived He from mantle depths rather than the residual magmatic fluids or Quaternary mantle-derived melt intrusion into the shallow crust.Mantle-derived volatiles in southern Tibet and the Himalayas correspond to 1%-27.8% asthenospheric or 1%-36.5% subcontinental lithospheric mantle volatiles.
There are large discrepancies between different methods for helium flux calculations.Based on the steady-state one-dimensional flow model developed in this study, we estimate a mantle-derived 3 He flux of 17 to 1.5 × 10 7 atoms m 2 s 1 , accounting for 81.7%-99.4% of the total 3 He flux, and a mantle-derived 4 He flux of 2.0 × 10 6 to 1.8 × 10 12 atoms m 2 s 1 , accounting for 1.4%-36.5% of the total 4 He flux.The mantle-derived helium fluxes in southern Tibet and the Himalayas are comparable to those of hydrothermal systems without recent volcanic activity and lower than those of volcanic fields.However, the lower 3 He/ 4 He values in Tibet than in other hydrothermal systems without recent volcanic activity are mainly attributed to the crustal thickness of the study area being nearly twice that of the latter.Hydrothermal fluids in southern Tibet and the Himalayas as well as other nonvolcanic continental regions are characterized by important mantle helium outgassing.
The upward flow rate of mantle-derived fluids ranges from 30 to 11700 mm/year, corresponding to transit times ranging from 5.3 ka-2.3Ma.The timescale of helium transport suggests that the distribution of 3 He/ 4 He ratios in the India-Asia continental collision settings can represent the state of the Indian mantle lithosphere between those revealed by potassic-rich mafic rocks (25-8 Ma) and seismic imaging techniques (present day).
and [He] C concentrations of 265 ± 645 ppm, and the "crustal-derived 4 He domain" group (CG, R C ≤ 0.1 R A ), with average R C values of 0.04 ± 0.02 R A (1σ, n = 188) and [He] C concentrations of 2,458 ± 4,078 ppm (Figure

Figure 5 .
Figure 5. (a) 3D plot of 3 He/ 4 He ratios versus distance from rifts and strike-slip faults; (b) 3 He/ 4 He ratios plotted as a function of distance from the rifts and (c) strike-slip faults.

Figure 7 .
Figure 7.Comparison of mantle and crustal helium fluxes in the study area with those in the volcanic field and tectonically active field.(a) Mantle 3 He flux calculated by porosities of 1% plotted against 3 He/ 4 He ratios.(b) Mantle 4 He flux plotted against crustal 4 He flux by adopting a porosity of 1%.The original mantle 4 He flux and crustal 4 He flux of the volcanic field (pink area) and tectonically active field (black dashed area) were determined based on the continental helium flux in the lake, groundwater(Torgersen, 2010) and natural gas(Sano et al., 1986).The modified "tectonically active field" (the yellow area) is built by the addition of continental hydrothermal data from volcanic regions (e.g., Yellowstone,Lowenstern et al., 2014; Southeast Tibetan Plateau margin, Zhang et al., 2022; North Anatolian Fault  System in Turkey, de Leeuw et al., 2010)  and tectonically active regions (e.g., Peruvian flat slab in South America,Hiett et al., 2021; San Andreas Fault System in USA,Kennedy et al., 1997;Kulongoski et al., 2003Kulongoski et al., , 2005Kulongoski et al., , 2013; Alpine Fault in New Zealand, Menzies et al., 2016; Southeast Tibetan Plateau margin, Zhang et al., 2022; Dixie Valley in USA, ; Kennedy & van Soest, 2006; Denali fault system in USA, Newell et al., 2023; southern Tibet and the Himalayas).
Figure 7.Comparison of mantle and crustal helium fluxes in the study area with those in the volcanic field and tectonically active field.(a) Mantle 3 He flux calculated by porosities of 1% plotted against 3 He/ 4 He ratios.(b) Mantle 4 He flux plotted against crustal 4 He flux by adopting a porosity of 1%.The original mantle 4 He flux and crustal 4 He flux of the volcanic field (pink area) and tectonically active field (black dashed area) were determined based on the continental helium flux in the lake, groundwater(Torgersen, 2010) and natural gas(Sano et al., 1986).The modified "tectonically active field" (the yellow area) is built by the addition of continental hydrothermal data from volcanic regions (e.g., Yellowstone,Lowenstern et al., 2014; Southeast Tibetan Plateau margin, Zhang et al., 2022; North Anatolian Fault  System in Turkey, de Leeuw et al., 2010)  and tectonically active regions (e.g., Peruvian flat slab in South America,Hiett et al., 2021; San Andreas Fault System in USA,Kennedy et al., 1997;Kulongoski et al., 2003Kulongoski et al., , 2005Kulongoski et al., , 2013; Alpine Fault in New Zealand, Menzies et al., 2016; Southeast Tibetan Plateau margin, Zhang et al., 2022; Dixie Valley in USA, ; Kennedy & van Soest, 2006; Denali fault system in USA, Newell et al., 2023; southern Tibet and the Himalayas).

Figure 8 .
Figure 8.(a) Comparisons of the mantle sutures revealed by potassic-rich mafic rocks (25-8 Ma), seismic imaging techniques and 3 He/ 4 He ratios.Seismic profiles A-A′, B-B′, C-C′, and D-D′ and the inferred mantle suture locations are from Zhang et al. (2014), Nábělek et al. (2009), Kosarev et al. (1999), and Shi et al. (2015), respectively.(b) Illustration of mantle-derived helium degassing in the India-Asia collision zone.The lithospheric structure is inferred on the basis of 3 He/ 4 He ratios from Klemperer et al. (2022) and this study.The variability in the 3 He/ 4 He ratios, migration velocity and flux of mantle-derived 3 He is associated with the permeability of the deep crust.
He and 21 Ne isotopes were analyzed by an electron multiplier.Mass peaks of 40 Ar ++ and 44 CO 2 ++ were also monitored during Ne isotope analysis to correct for 40 Ar ++ and 44 CO 2 ++ interferences on 20 Ne and 22 Ne, respectively (Hao et al., 2023; Li

Table 1
Sample Locations and Physical and Chemical Properties of Hydrothermal Fluids in Southern Tibet and the HimalayasPositive and negative values of distance to the Indus-Yarlung Suture (IYS) represent the sampling sites located at the north and south of the IYS, respectively. a

Table 2 Gas Compositions and He and Ne Isotope Values of Hydrothermal Fluids
a Air-corrected helium concentration ([He] C