Introduction

Volatiles are known to be effective tools for directly observing the influence of the mantle on the Earth’s surface. As the asthenospheric mantle ascends and reaches the lithosphere, volatiles can be transported through volcanism or fault processes1,2,3,4. Through interaction with fluids containing mantle-derived volatiles, the lithosphere in intra-plate settings, such as incipient continental rift zones, undergoes thinning with physical and thermal weakening, eventually resulting in the development of new divergent boundaries5,6. Among volatiles, He that is the lightest noble gas and CO2 that is the second most abundant component of volcanic gas after water vapor7 are widely used to trace mantle-derived components in plate margins (e.g., Sano and Marty8; Tardani et al. 9). In particular, He isotopes are typically expressed as 3He/4He (1 Ra = 1.38 × 10−6, refs. 10,11), and this ratio allows for a clear distinction between mantle and crustal values, effectively discerning the mantle signals from rock or fluid samples. Sources of CO2 can be traced via carbon isotope compositions (δ13C), which can differentiate between the depleted upper mantle and recycled carbonate or organic matter-bearing slab materials (e.g., Tardani et al. 9 and Lee et al. 12).

While magmatism is predominant at plate boundaries, it also appears in the interiors of tectonic plates13,14,15. The intra-plate magmatism is associated with the break-up of the supercontinent, involving large igneous provinces (LIPs)16, which can trigger the development of continental rifts through the thermal weakening of the lithosphere17. In this context, it has been reported that in various modern continental rifting areas (e.g., East Africa and Rio Grande), the 3He/4He ratios and δ13C-CO2 values of the asthenospheric mantle origin predominate in fault-related springs from regions with relatively thinner lithosphere, as a result of weaken physical properties of the plates due to magmatism by asthenospheric mantle upwelling3,12,18.

Northeast (NE) Asia, including the Korean Peninsula and China, became an intra-plate environment following the separation of the Japanese arc from the Eurasian plate during the East Sea back-arc opening19,20. The late Cenozoic magmatism in this region has led to the emplacement of alkaline rocks across NE Asia (Fig. 1). 3He/4He ratios measured in olivine and clinopyroxene contained in these Cenozoic volcanic rocks range from 4.5 to 7.3 Ra, indicating potential mantle melting by lithosphere-asthenosphere interaction20,21. Moreover, 3He/4He ratios in the mantle value range (up to 6.0 Ra) have been observed in groundwater and fault-related spring samples within the southeastern Korean Peninsula22,23,24. Interestingly, seismic tomography has detected distinctive slow velocities in the upper mantle mainly beneath the central and eastern coastal margins of the Korean Peninsula, indicative of asthenospheric upwelling accompanied with partial melting25. Additionally, edge-driven convection (EDC) has been proposed as a possible mechanism for low velocity zone (LVZ) generation in the Northeast Asia region from evidence such as the heterogeneous distribution of LVZs controlled by the lithospheric thickness and the swelling that forms the mountain ranges present on the surface25,26. However, a geochemical interpretation to support the presence of partial melting has not yet been attempted.

Fig. 1: Helium isotope distribution on S-wave mantle tomography in the study area.
figure 1

a Distribution of corrected He isotope compositions (Rc/Ra) of the samples (gray scale circles) and the distribution of the Korean Cenozoic volcanic rocks (yellow and red triangles) overlaid on the relative S-wave velocity map that represents S-wave perturbations at a depth of 60 km. Sampling locations are elliptically represented in three regions: mwKp (midwestern Korean peninsula), meKp (mideastern Korean peninsula), and seKp (southeastern Korean Peninsula). The larger circles are Rc/Ra ratios reported in this study, and the smaller circles are reference data63,64,65,66. The region with velocity reductions exceeding −2% is delineated by dashed white lines. Most of the sampling locations are related to faults, and their distribution is shown in Supplementary Information. b The study area is boxed with red dashed lines in the NE Asia map, and the names of the Cenozoic volcanic areas in NE Asia are abbreviated as follows: Wudalianchi (WL), Dariganga (DG), Hannuoba (HN), Mt. Baekdu (BD), Datong (DT), Ulleung (UL), and Jeju (JJ). The satellite image is from Google Maps (2023) Northeast Asia, 22,500,000:1, https://goo.gl/maps/qC7f5ojgNtMLUdom9 [September 11, 2023].

Full size image

This study presents the comprehensive analytical results that combine geochemical and geophysical data from the Korean Peninsula, employing the following approaches. First, we assume that the fault-related springs distributed across South Korea are emitters of volatiles on the surface to test the existence of mantle upwelling. This assessment further confirms that the Korean Peninsula regions share a common mechanism with other continental areas that include intra-plate volcanism. Then, by integrating the measured He data with S-wave velocity tomography, we propose a potential magma distribution at the upper mantle level beneath the Korean Peninsula, addressing that mantle upwelling acts as a source of volatiles from the Cenozoic volcanic rocks and fluids in this region.

Results and discussion

Sources of volatiles

Based on the distribution of He isotope ratios, we found distinct areas with high 3He/4He ratios in the Korean Peninsula and divided it into three areas (Fig. 1 and Supplementary Information): the southeastern Korean Peninsula (seKp), the mideastern Korean Peninsula (meKp), and the midwestern Korean Peninsula (mwKp). In general, CO2 and N2 are the main components that make up 81 to 100 vol.% of the gases emitted from the fault-related springs (Supplementary Information). Among them, 18 samples have N2 concentrations greater than 10 vol.%, and 10 out of them have N2 concentrations greater than 75 vol.%. 22 samples show CO2 concentrations higher than 90 vol.%. The range of He concentration is similar between mwKp (0.03 to 676.0 ppm) and seKp (0.07 to 613.5 ppm), but a slightly narrower range between 0.02 and 163.5 ppm is observed in meKp. The average concentrations of N2, O2, Ar, and CO2 are similar between meKp (23.2%, 2.9%, 0.5%, and 74.2%, respectively), mwKp (15.5%, 2.9%, 0.2%, and 82.8% respectively), and seKp (36.0%, 1.7%, 0.6%, and 61.9%, respectively), with CH4 up to 0.6% in some samples of seKp. Our samples show a mixing relationship between atmospheric (Air-ASW) and endogenic (mantle-crustal) components based on the N2-Ar-He ternary diagram (Fig. 2), which is consistent with previous studies in the SE Korean Peninsula22,23. In addition, the subduction zone signatures (e.g., elevated N2/He and N2/Ar ratios in Fig. 2) were not identified in the study area as previously reported22,23.

Fig. 2: N2-He-Ar ternary plot.
figure 2

MORB and ASW refer to Mid-Ocean Ridge Basalts and Air-Saturated Water, respectively. Reference values are from Gyeonju, Ulsan, and Busan in South Korea22 (white squares), Pohang in South Korea23 (dark gray squares), and global continental rift gases12 (light gray squares). All of the following figures use the same symbols.

Full size image

The sources of He are investigated by using 3He/4He ratios and reciprocal 4He/20Ne ratios (Fig. 3). Overall, 3He/4He ratios of the fluid samples in the Korean Peninsula range from a maximum of 6.2 Ra to a minimum of 0.1 Ra, and 4He/20Ne ratios range from 0.3 to 342.8, with an average of 45.6. As 30 samples out of 39 samples have higher 4He/20Ne ratios than 1, the effect of air contamination on He isotopes is unlikely to be severe. Since LVZs are located in the deeper levels than the lithospheric thickness27, we selected the asthenospheric mantle as the original source of volatiles rather than subcontinental lithospheric mantle (SCLM). It seems that He emitted from the asthenospheric mantle beneath the Korean Peninsula is affected by various degrees of the contributions from the SCLM (6.1 ± 0.9 Ra, ref. 28) and continental crust (0.02 Ra, ref. 29) end- members. The highest 3He/4He ratio in each region is seKp (6.2 Ra), meKp (4.3 Ra), and mwKp (2.2 Ra) indicating that springs with an obvious mantle contribution are spatially well distributed, especially in the seKp area (Fig. 1).

Fig. 3: 3He/4He (R/Ra) vs. 20Ne/4He plot.
figure 3

Mixing lines between each end-member (MORB, SCLM, crust, and air) are shown with a percentage contribution from the MORB mantle source. Reference values are shown as gray diamonds63,64,65,66.

Full size image

The δ13C-CO2 values of the samples (vs. Vienna Pee Dee Belemnite) range from −20.2 to −3.7 ‰. When the CO2/3He ratio is considered together, two main trends appear (Fig. 4). First, it has a range similar to the δ13C-CO2 of the mid-ocean ridge basalts (MORB) mantle (−6.5 ± 2.5 ‰, ref. 8), but shows a wide range of CO2/3He ratios (up to 2.81 × 1013). This vertical trend suggests 1) crustal contamination or 2) the effect of degassing fractionation on gas components associated with the difference in solubility of CO2 and He in hydrothermal systems without considerable impact on isotopic change9,11. Another geochemical feature in our CO2 data is the tendency of CO2/3He ratios to decrease with δ13C-CO2 values. Although the δ13C-CO2 values of some of these samples alone can indicate organic CO2, previous studies of gas and rock samples around the Korean Peninsula suggested that slab-derived volatiles are not detectable20,21,22,23,24, which is also not observed in the N2-Ar-He systematics (Fig. 2). Instead, it is more likely that CO2 removal occurred due to precipitation of carbonate controlled by pH in water30,31,32. In support of this, the linear relationship between CO2 concentration and pH (Supplementary Information) exhibits a fairly valid correlation (R2 = 0.52), implying that CO2 loss may have occurred in some high pH springs in the Korean Peninsula, as previously reported in seKp23. Therefore, we propose that CO2 emitted through the fault-related springs of the Korean Peninsula is originally from the mantle, assuming that there were no effects of degassing fractionation and CO2 loss.

Fig. 4: CO2/3He vs. δ13C plot.
figure 4

The percentage of organic sediment contributions of the crustal end-members are displayed. Global continental rift gases are shown12 (open gray boxes). Trend (1) indicates degassing fractionation and trend (2) reflects the effects of calcite precipitation (brown dashed lines) at various temperatures and water-gas interaction (gray dashed lines) under various pH conditions at 15 °C.

Full size image

The origins of N2, another major gas component, were traced by δ15N-N2 (−8.0 to 3.3‰ vs. Air) values and N2/3He ratios (Fig. 5). If we used the conventional δ15N-N2/3He mixing diagram of the MORB mantle (−5 ‰; 8.9 × 105), air (0 ‰; 1.0 × 1011), and sediment (7 ‰; 1.4 × 1012) end-members33, some data points from the meKp and seKp samples are out of the mixing relationship. Unlike the increase in CO2/3He ratios due to degassing fractionation shown in the mixing diagram of CO2 (Fig. 4), for N2/3He, N2 is 72 times and He is 200 times less soluble in Henry’s law compared to CO2 in the low temperature water (0 °C) (refs. 34,35). Thus, as both N2 and He have very low solubility in similar orders, it is likely that the degassing fractionation effect is insignificant for the N2/3He ratios of the samples. To address this off-trend samples in three component mixing, we show the local SCLM-derived N2 source from the reported δ15N-N2 range of mantle xenoliths in Far Eastern Russia36. Since the δ15N-N2 heterogeneity in the lithospheric mantle has been suggested regionally (e.g., Fischer et al. 37), we attempted to account for the cause of some of the deviation values has been resolved by applying the local end-member. As chemical composition of the SCLM beneath the Korean Peninsula is not reported yet, spatially closest data of the SCLM (i.e. Far East Russia) would be the most plausible substitution. This could indicate that the SCLM interacting with the ascending asthenospheric mantle with partial melts, underwent metasomatism and became the potential N2 source. Although some data points show sediment contributions, the air contribution seems to be more striking. Regarding the formation of Cenozoic magmas in NE Asia, the involvement of the SCLM has been proposed to have driven by lithosphere-asthenosphere interaction20,21. In the same context, as the lowest δ15N (< 0‰) decreases when S-wave velocity decreases (Supplementary Information), implying that endogenic N2 from the SCLM or the asthenospheric mantle tends to be emitted from LVZs.

Fig. 5: N2/3He vs. δ15N plot.
figure 5

The SCLM range (1σ area around the average value) is from xenoliths (noted by gray squares) in Sveyagin, Far East Russia36. The solid lines connect each end-member (MORB, SCLM, sediment, and air). The dashed lines display the contribution of mantle to sediment for quantitative tracing of nitrogen sources, and the larger the air end-member contributes, the closer they are to the mantle-air mixing line.

Full size image

Distribution of He isotope ratios related to mantle upwelling

In NE Asia, lithosphere and asthenosphere boundary (LAB) depths typically range from 50 km to 250 km, and the Cenozoic volcanoes are mainly distributed in areas with LAB depths less than 75 km (ref. 38). Referring to this, we examined the spatial distribution of seismic velocity anomalies at a depth of 60 km (Fig.1a) and the averaged S-wave anomaly at depths from 40 to 100 km (Supplementary Information), which are in range of the lithospheric thickness in the Korean Peninsula and NE China27. In the 40 to 100 km depth range, the values at each depth and the average for the interval show similarity to each other (Supplementary Information), but the horizontal contrast of the physical properties at a depth of 60 km can be most clearly shown with respect to the resolution of the mantle tomography model. Considering this, the tomographic image with a depth of 60 km was selected as the representative (Figs. 1a and 6). To quantitatively visualize the correlation between mantle-derived helium and LVZs, we plotted Rc/Ra ratios and relative S-wave velocity values together (Fig. 6). The enveloping line between the maximum values of the Rc/Ra ratios and the relative S-wave velocity indicates a negative correlation. A majority of samples with lower 3He/4He ratios (e.g., Rc/Ra <1) are distributed from slightly negative (> −1%) to positive S-wave anomalies, which is attributed to the influence of Precambrian metamorphic rocks and Mesozoic granitoids with abundant U and Th39. Lee et al. 22 reported the results of Rc/Ra < 0.1 related to the distribution of granitoids in the seKp area. However, the elevated Rc/Ra ratios (e.g., > 1 Ra) and other mantle-derived volatiles (e.g., CO2 and N2) are displayed only within areas with negative S-wave anomalies, implying an obvious mantle contribution.

Fig. 6: 3He/4He ratios vs. relative S-wave velocity (at a depth of 60 km) plot.
figure 6

The dashed line represents the negative correlation of the maximum values between the Rc/Ra ratios and the relative S-wave velocities.Reference values are noted by gray diamonds63,64,65,66.

Full size image

Implications for potential magmatism

In the crustal and mantle 4He flux plot (Fig. 7), samples from volcanic region and tectonic strain processes can be clearly distinguished40. In addition, magmatism is thought to have higher crustal 4He fluxes as well as mantle 4He fluxes compared to the values related to the tectonic strain degassing41. Using the above method, the 4He flux values of crust and mantle calculated from our samples (Supplementary Information) were plotted together with those of Okinawa Trough, Taiwan, Taupo, Yellowstone, and Magadi-Natron, where volcanic 4He flux values have previously been reported12,40,41,42,43. Except for 4 samples with lower mantle 4He flux values, where thermal spa waters are highly affected by radiogenic 4He at the crustal level (Supplementary Information), most data points are plotted in the volcanic region (Fig. 7). In general, it has been suggested that mantle He could be emitted through faulting in non-volcanic areas (e.g., San Andreas Fault1). Likewise, it was proposed that mantle He is discharged through the active faults extending beyond the crustal thickness, considering no active volcanoes currently in the Korean Peninsula22,23,24. However, our calculated 4He fluxes refute the aforementioned model and can support that regions with negative S-wave anomalies are consistent with the distribution of Neogene and Quaternary volcanic rocks (Fig. 1b). Specifically, the Quaternary volcanic rocks in Ulleung and Jeju (Fig. 1b), which were suggested to be produced as a consequence of lithosphere-asthenosphere interaction, show similar R/Ra ratios (3.5 to 7.3 Ra) to our samples20,21. Furthermore, the Cenozoic volcanic rocks (< 0.1 to 30 Ma) in the central North China Craton region adjacent to the Korean Peninsula have neen addressed to originate from deep magma sources distributed at the LAB (90 and 130 km, ref. 44). In the North China Craton area, Rc/Ra ratios have been reported up to 2.08 in hot springs and geothermic wells with negative Vp anomaly mainly observed at a depth of 70 km, indicating that the mantle-derived volatiles are emitted from faults through asthenopheric mantle upwelling45.

Fig. 7: Crustal and mantle 4He flux plot.
figure 7

Flux values of the samples were calculated with a porosity value (Φ) of 0.01, and the gray area is a computed range for the samples with different porosity values from 0.005 to 0.2. Reference values are from Okinawa Trough42 (purple), Taiwan43 (gray), Yellowstone41 (black), Magadi-Natron12 (white with black outline), and other areas40 (green).

Full size image

Comprehensively, our results address the occurrence of partial melting or the existence of deep magma beneath the Korean Peninsula. By illustrating the He degassing characteristics of the Korean Peninsula (Fig. 1), we propose potential magmatism beneath the Korean Peninsula based on the following rationales; 1) the primary source of volatiles is inferred to be the mantle; 2) the presence of asthenospheric mantle signals at shallow upper mantle depths is evidenced by the negative S-wave anomaly, elevated Rc/Ra ratios, and the spatial distribution of Cenozoic volcanic rocks across the Korean Peninsula; and 3) the 4He flux values of the mantle and the crust show a volcanic region trend. Our findings propose that Cenozoic intraplate magmatism in NE Asia may not be limited to the distributed volcanic rocks but could exist more extensively as latent magmatism, which emits mantle-derived volatiles that can be observed on the Earth’s surface.

To explain N-S elongated cratonic root and low velocity zone, we suggest that EDC is the LVZ generation mechanism25,26. The eastern part of the Korean Peninsula underwent thinning during the East sea (Sea of Japan, Fig. 1a) back-arc opening. This thinning event generated a heterogeneous lithospheric thickness difference between thicker western cratonic and thinner eastern regions, inducing decompression melting perpendicular to the lithosphere thickness gradient27. Taking this into account, we hypothesized that EDC may occur in slow plate movement environments (< 1 cm/yr, Kaislaniemi and van Hunen46), such as the Korean Peninsula47. As a result, partial melting occurring in the asthenospheric mantle from this process supports that the volatiles have geochemical properties similar to those of gases emitted from other continental rifts. Since the stagnant Pacific slab physically separates the lower mantle and upper mantle layers in the mantle transition zone (410-660 km depth)48,49, it is difficult for any large-scale lower mantle plume to affect the Korean Peninsula. Additionally, the seismic velocity anomaly is concentrated in the upper mantle (50 to 120 km depth) than deeper regions (~300 km, ref. 25). Furthermore, as the opening of the East Sea (Sea of Japan) progressed, the source of volcanic rocks was mainly contributed by metasomatized SCLM from arc magma, and as the plate thinning continued, the asthenosphere component eventually prevailed without the lower mantle plume components50,51. Finally, the geochemical characteristics of volatiles, including helium derived from the upper mantle emitted from the Korean Peninsula, are particularly affected by the lithospheric thickness, indicating that EDC is a major mechanism for the Cenozoic magmatism in Northeast Asia. Furthermore, it suggests that larger-scale mantle degassing may occur globally and that the total flux of volatiles may be underestimated.

Methods

Sampling and gas analysis

30 samples (18 water samples, 3 thermal spa water samples, and 9 gas samples) from 26 sampling sites (Fig. 1; Supplementary Information) were measured to obtain gas composition, and isotope ratios of 3He/4He, δ13C-CO2, δ15N-N2, and 36Ar/40Ar. Additionally, 9 previously reported samples from the SE Korean Peninsula were discussed together22,23,24. Both bubbling and dissolved gases in water samples analyzed at the Atmosphere and Ocean Research Institute (AORI), the University of Tokyo (e.g., Lee et al. 22 and Kim et al. 23). For water samples, dissolved gases were extracted in the high vacuum system before analysis. Concentrations (CO2, N2, O2, CH4, Ar, and He) of gases were measured by a Pfeiffer Prisma QMS 200 quadrupole mass spectrometer (QMS). δ13C of CO2 and δ15N of N2 were measured by an isotope ratio mass spectrometer (Isoprime 100 by Elementar). To determine 3He/4He and 4He/20Ne ratios, dissolved gas samples were purified by titanium getters at 400 °C and charcoal traps at liquid nitrogen temperature (77 K). After measuring 4He/20Ne ratios via online QMS (Prisma QMS 200 by Pfeiffer), Neon was trapped by the cryogenic pump at 40 K. Then, purified He was injected into a noble gas mass spectrometer (Helix SFT by ThermoFisher) to measure 3He/4He ratios. He isotope ratios were calibrated using the internal He standard of Japan (HESJ, ref. 52), the air, and the air saturated water (ASW). In another vacuum line, argon was purified using a liquid nitrogen temperature U-trap and titanium getter at 600˚C, and analyzed by QMS (MKS Microvision-2). Measured 3He/4He ratios (R/Ra) were corrected to eliminate the air-derived He effect for each sample. Postulating 20Ne as an air-derived component53, the pristine 3He/4He ratios of samples in the absence of air contamination can be calculated as below54:

$${{\rm{Rc}}}/{{\rm{Ra}}}=[({\!\,}^{3}{{\rm{He}}}{/}^{4}{{\rm{He}}})_{{{\rm{measured}}}}-{{\rm{r}}}]/(1-{{\rm{r}}})$$
(1)

and

$${{\rm{r}}}=({\,\!}^{4}{{\rm{He}}}/^{20}{{\rm{Ne}}}){{\rm{ASW}}}/({\,\!}^{4}{{\rm{He}}}/^{20}{{\rm{Ne}}})_{{{\rm{measured}}}}$$
(2)

where Rc/Ra is the corrected 3He/4He ratio in R/Ra notation, and (4He/20Ne)ASW is the 4He/20Ne of ASW. In this study, we used the 4He/20Ne value of sample #8 as (4He/20Ne)ASW, where sample #8 shows a typical ASW signature (Supplementary Information).

S wave modelling and He flux estimation

We used a three-dimensional S-wave tomography model of the southern Korean Peninsula25. We focus on seismic velocities at depths ranging from 40 to 100 km, which correspond to the upper mantle. The geophysical methods, datasets, and model resolutions can be found in Song et al. 25. Through the velocity model, volumetric local velocity heterogeneities in the crust and upper mantle have been estimated in greater detail. Crustal thicknesses were derived from Moho depth datasets around the southern Korean Peninsula obtained from receiver function analyses55,56,57.

As mentioned in the previous sections, it was suggested that volatiles from mantle or crustal sources are transported mainly through permeable faults. Thus, the helium flux values emitted from the Korean Peninsula were calculated. First, the current 4He production from the crust, P(He), can be computed as follows:

$${{\rm{P}}}({{\rm{He}}})={{{\rm{\rho }}}}_{{{\rm{c}}}}\times (1.19\times {10}^{-13}\times [{{\rm{U}}}]+2.88\times {10}^{-14}\times [{{\rm{Th}}}])\times (1{{\rm{\hbox{-}}}}\Phi )/\Phi$$
(3)

where Φ is the porosity of the fluid path, and [U] and [Th] are the U and Th concentrations of the crust, respectively. 1.19 × 10−13 and 2.88 × 10−14 are the 4He production rates of U and Th, respectively (ref. 58). Using the computed P(He), the He flow rate (qHe) can be acquired as below:

$${qHe}=\frac{{H}_{c}{\rho }_{c}P\left({He}\right)}{{\rho }_{f}{\left[{He}\right]}_{F.m}}\times \left[\frac{{R}_{s}-{R}_{c}}{{R}_{m}-{R}_{s}}\right]$$
(4)

where Hc is the crustal thickness; ρc and ρf are the density of the crust and the fluid (water, 1 g/cm3); [He]F.m is the He concentration of the mantle-derived fluid; Rs, Rc, Rm are 3He/4He of the sample, the crust (0.02 Ra, ref. 29), and the MORB mantle (8 Ra, ref. 59), respectively. We used geophysical data for the crustal thickness as described above. For all samples, the average composition ([U] = 1.3 ppm and [Th] = 5.6 ppm) and the crustal density (2.8 g/cm3) were substituted60. [He]F.m was calculated from the sample He concentration and Rs, where postulating the 3He/4He ratio of sample helium is the simple mixture of 0.02 Ra of the crustal He (Rc) and 8 Ra of the mantle He (Rm). Using the acquired qHe and He concentration of each sample, the 4He flux of each sample site was determined. All variables in Eqs. (3) and (4) except Φ were defined or measured. Therefore, the He flow rate and flux are functions of Φ. To further specify processes of how mantle He is released, we estimated distinction in 4He fluxes between the mantle and crust contributions by using the conditions of 1) concentration of volatiles in water, 2) crustal thickness, 3) crustal porosity, and 4) Rc/Ra ratios1,40,61,62. We postulated Φ (0.01) and its possible range (0.005 to 0.2) from previous studies1,40,62. The He fluxes are from 1.13 × 102 to 1.48 × 106 cm/yr with Φ = 0.005 and from 2.26 × 100 to 2.97 × 104 cm/yr with Φ = 0.2. Total 4He fluxes are from 1.12 × 103 to 4.16 × 103 atoms/m2/sec × 1010 with Φ = 0.005 and from 2.25 × 101 to 8.37 × 101 atoms/m2/sec × 1010 with Φ = 0.2 (Supplementary Information).