Helium isotopes as a tool for detecting ... - Wiley Online Library

Geochemistry Geophysics Geosystems

3

G

Research Letter Volume 10, Number 8 14 August 2009 Q08010, doi:10.1029/2009GC002501

AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES Published by AGU and the Geochemical Society

ISSN: 1525-2027

Helium isotopes as a tool for detecting concealed active faults Koji Umeda and Atusi Ninomiya Tono Geoscientific Research Unit, Geological Isolation Research and Development Directorate, Japan Atomic Energy Agency, 959-31, Jorinji, Izumi, Toki 509-5102, Japan ([email protected])

[1] A magnitude (Mj) 7.3 crustal earthquake occurred in the western Tottori area, southwest Japan, on

6 October 2000. However, there was no obvious prefaulting indication at surface of a fault corresponding to the western Tottori earthquake in 2000. This study was undertaken to elucidate the geographic distribution of 3He/4He ratios around the seismic source region using helium isotope data obtained from groundwater samples from drinking water wells. The maximum 3He/4He ratio observed was from the water well nearest to the epicenter of the main shock. In addition, there appears to be a clear trend of decreasing 3He/4He ratios with distance away from the main trace of the probable fault segments. The observations provide significant evidence that the source fault of the earthquake in 2000 is associated with leakage of mantle volatiles through the crust to the Earth’s surface. We suggest that helium isotopes can be regarded as a tool for investigating and/or mapping concealed active faults with no surface expression. Components: 5574 words, 5 figures, 1 table. Keywords: helium isotope; active fault. Index Terms: 1031 Geochemistry: Subduction zone processes (3060, 3613, 8170, 8413); 7230 Seismology: Seismicity and tectonics (1207, 1217, 1240, 1242); 8118 Tectonophysics: Dynamics and mechanics of faulting (8004). Received 13 March 2009; Revised 19 June 2009; Accepted 1 July 2009; Published 14 August 2009. Umeda, K., and A. Ninomiya (2009), Helium isotopes as a tool for detecting concealed active faults, Geochem. Geophys. Geosyst., 10, Q08010, doi:10.1029/2009GC002501.

1. Introduction [2] Even in regions of present-day low to moderate seismic activity, historically, large earthquakes have caused extensive damage [e.g., Johnston, 1996]. For example, on 6 October 2000, the western Tottori earthquake (Mj = 7.3, Mw = 6.6 6.8) occurred in western Honshu, Japan, where very few large earthquakes have occurred since the 1943 Tottori earthquake [Kanamori, 1972]. Although this was a large, shallow, intraplate earthquake, there appear to be three characteristics that distinguish it from common large earthquakes: (1) any surface expression of fault rupture is not clearly observed;

Copyright 2009 by the American Geophysical Union

(2) active faults are absent or unknown; and (3) the maximum shear strain rate, determined using the nationwide GPS network, is lower than in other areas. This is one of the reasons why the earthquake fault is considered to be an ‘‘immature fault,’’ a fault at an early stage of evolution, on the basis of geomorphological, geological and seismological evidence [e.g., Inoue et al., 2002]. From the viewpoint of seismic hazard and geological risk reduction, it is imperative to recognize potentially active faults lacking clear surface expression, especially in areas with a low density of active faults but where there is potential for earthquakes with magnitude larger than would be

1 of 10


3

G

umeda and ninomiya: helium isotopes and concealed active faults

expected because of the lack of any geomorphological signature. [3] Noble gases and their isotopes are excellent natural tracers for elucidating mantle-crust interactions in different geotectonic provinces because they are chemically inert and thus conserved in crustal rock-water systems. Helium is the lightest noble gas and both isotopes, 3He and 4He, are produced in the crust at a ratio of 0.02 RA (RA denotes the atmospheric 3He/4He ratio of 1.4 10 6). Values higher than the atmospheric ratio are an indication of helium from a reservoir enriched in 3 He. The most viable possibility for such a reservoir is the mantle, which stored 3He captured during planetary accretion [Ozima and Podosek, 2002]. Therefore, helium isotopic variations are considered to provide not only potentially useful information for indicating of the addition of mantle volatiles, but also provide a way to determine migration pathways from the subcrustal lithosphere [e.g., Umeda et al., 2008]. [4] The enrichment with a wide variety of terrestrial gases along active faults suggests that active faults may be major leaks or pathways in the crust for the gases, where materials are more porous and the permeability is relatively high [King, 1986]. Several researchers have made efforts previously to elucidate the geographic distribution of the 3 He/4He ratios around active faults and their relationship to seismically active areas. Kennedy et al. [1997] attributed elevated 3He/4He ratios of up to 4 RA to transport of mantle helium through the fault structure along the San Andreas fault system. Kulongoski et al. [2005] insisted that mantle helium found in groundwaters of the Morongo Basin moved via deeply penetrating active faults, and that episodic seismicity and associated hydrofracturing drive volatile transfer from the mantle to the crust. Gu¨leç et al. [2002] also suggested that fault activity might trigger the transport of mantle helium in the western part of the North Anatolian Fault Zone. In this paper, we describe the geographic distribution of 3He/4He ratios in and around the source region of the 2000 earthquake and interpret the relationship between helium isotope variations and geophysical structures in the earthquake source regions, specifically with regard to the results of seismic and magnetotelluric observations in the crust and mantle. In addition, we propose the potential use of helium isotope ratios as a tool to investigate immature faults at an early stage of fault evolution, i.e., unknown active

10.1029/2009GC002501

faults with no clear indication of previous faulting, but with the potential to cause large earthquakes.

2. The 2000 Western Tottori Earthquake [5] The geology of the western Tottori area is mainly composed of Late Cretaceous to Paleogene granitic rocks which are intruded into Late Triassic to Middle Jurassic metamorphic rocks. The Miocene sedimentary rocks, unconformably overlying the granitic rocks, are distribution on the northern part of the source region. The Quaternary, but not active, Daisen volcano is located to northeast of the western Tottori area, and a small amount of basaltic rocks of early Pleistocene age (Yokota volcanic rocks) are distributed to west of the area [Hattori and Katada, 1964]. [6] The 2000 western Tottori earthquake was the first large intraplate earthquake recorded by the recently developed nationwide seismic network in Japan and thus provides a unique set of near-field data. Detailed fault structures were determined on the basis of a combination of relocated hypocenters and moment tensor solutions of the aftershocks by broadband waveform inversion [Fukuyama et al., 2003]. The conceptual fault model resolved 11 individual fault segments, most of which are leftlateral strike-slip faults oriented in the NW-SE direction. Rupture during the main shock was principally confined to the three large fault segments in the southern part of the seismogenic region. The northern part was activated later with a large number of aftershock events concentrated in the region that did not slip during the main shock, but which is reflected in the postseismic deformation determined by analysis using GPS data [Sagiya et al., 2002]. [7] As indicated above, the earthquake in 2000 did not exhibit any clear preearthquake evidence at surface of faulting (Figure 1). However, after the earthquake, many detailed field surveys in and around the aftershock region revealed five surface fractures oriented to NW-SE trending lines in an area 6 km long and 1 km wide [Fusejima et al., 2001] and along with NW-SE trending lineaments with a left-lateral offset of valleys and ridges [Inoue et al., 2002], which are interpreted as active faults. However, the fault length of these surface fractures and lineaments is significantly shorter than the expected length (about 30 km) estimated by the empirical law on the basis of the relationship 2 of 10


3

G


10.1029/2009GC002501

Figure 1. Topographical map of southwestern Honshu and distribution of active faults (red lines) [Nakata and Imaizumi, 2002] and an active volcano (red triangle).

between magnitude of an earthquake and length of the earthquake fault [Matsuda, 1975]. Inoue et al. [2002] stated that these are ‘‘immature faults’’ on the basis of the short fault lengths with narrow fracture zones. Seismic and geoelectric evidence [Fukuyama et al., 2003; Yamaguchi et al., 2007], and the results of sandbox analog experiments

[Ueta et al., 2000] suggest that these faults are in an early stage of fault development. In addition, simulation by Dalguer et al. [2003] indicated that shear slip occurring on a preexisting fault at least at a depth of 2.0 km could cause the generation of surface raptures under a tensile stress regime

3 of 10


3

G


10.1029/2009GC002501

Figure 2. Simplified geological map of the western Tottori area based on Geological Survey of Japan [1995]. Localities of hot springs sampled in this study are indicated with sample numbers that correlate with Table 1. Pattern a shows Holocene sediments, pattern b shows debris, pattern c shows Pleistocene sediments, pattern d shows Quaternary volcanic rocks, pattern e shows Middle to Late Miocene sedimentary rocks, pattern f shows Early Miocene sedimentary rocks, pattern g shows Paleogene mafic igneous rocks, pattern h shows Paleogene felsic igneous rocks, pattern i shows Mesozoic mafic igneous rocks, pattern j shows Mesozoic felsic igneous rocks, pattern k shows Mesozoic metamorphic rocks, and pattern l shows active faults.

around the aftershock region, consistent with the field observations.

3. Analytical Procedures and Results [8] A total of 18 groundwater samples were collected from drinking water wells above the source region of the 2000 western Tottori earthquake and in their surrounding areas. They were collected in October 2008. The water samples were pumped from wells with depths of several meters to several tens of meters. Most of water samples were collected from drinking water wells located on Paleogene granitic rocks (Figure 2). [9] Stable isotope ratios (D/H and 18O/16O) of sampled waters were determined with a Micromass Optima isotope ratio mass spectrometer (IRMS),

using the zinc metal reduction method of Coleman et al. [1982] and the CO2-H2O equilibrium method of Yoshida and Mizutani [1986]. Values of hydrogen and oxygen isotopes of samples are expressed in d% versus SMOW (standard mean ocean water). Estimated total uncertainty of measurement is ±1.0% for dD, and ±0.1% for d 18O. The dD and d 18O value of hot spring waters show 54 to 46% and 9.1 to 7.7% versus SMOW, respectively (Table 1). The isotopic trend of groundwaters is similar to the local meteoric water line [Mizota and Kusakabe, 1994] (Figure 3). This suggests that either the water had a short residence time and little chance to interact with the local geology, or that the waters have not reached high enough temperatures to initiate significant water/ rock interaction. 4 of 10


3

G


10.1029/2009GC002501

Craig et al. [1978]. The corrected 3He/4He ratios are calculated to range from 0.15 to 4.7 RAcor. Contribution of relative mantle helium can be estimated to represent from 0.5 to 56% by a simple three-component mixing model [Sano and Wakita, 1985].

4. Discussion and Conclusions

Figure 3. The dD-d18O relationship of groundwaters around the seismic source region. The local meteoric water line on the coast of the Japan Sea (dD = 7.2 d 18O + 12 [Mizota and Kusakabe, 1994]) is shown.

[10] The dissolved gases in water samples were expelled from solution by ultrasonic agitation and collected in glass bottles. Helium and neon isotopes were measured using a modified VG5400 mass spectrometer installed in the Laboratory for Earthquake Chemistry, University of Tokyo. The 3 He and 4He ion beams were detected on a double collector system: 3He by axial counting and 4He by the high Faraday collector, (feedback resistor = 10 GW). A resolving power of 600 allowed the complete separation of the 3He+ beam from the H3+ and HD+ beams. Measured 3He/4He ratios are normalized to a standard 3He/4He gas (3He/4He = 1.71 10 4) prepared and stored in a stainless steel container on the inlet line. Neon was released from a cryogenic trap at 45 °K, and blank levels of 4 He and 20Ne determined were about 2 10 10 cm3 STP and 3 10 10 cm3 STP, respectively. These blank levels represent less than 0.1% of the amount of sampled gases, so blank correction was not required. Mass spectrometric details, including purification procedures, have been published elsewhere [e.g., Aka et al., 2001]. The 3He/4He and 4 He/20Ne ratios of the samples range from 0.55 10 6 to 6.2 10 6, and from 0.28 to 4.8, respectively (Table 1). Assuming that the 4He/20Ne ratios of mantle and radiogenic helium are significantly larger than that of ASW (0.254 at 15°C), samples can be corrected for atmospheric helium contamination in accordance to the principle reported by

[11] The geographic distribution of 3He/4He ratios in terrestrial gases or waters may reveal mantlecrust interactions in different geotectonic provinces, and the occurrence and distribution of conduits for the mantle volatiles within the crust [e.g., Hilton, 2007]. Figure 4 shows the spatial distribution of the air-corrected 3He/4He ratios of gases from water samples around the source region of the 2000 western Tottori earthquake. Generally, the crustal helium component is expected to be dominant in water samples from wells in regions away from active volcanoes [e.g., Sano and Wakita, 1985; Umeda et al., 2009]. That is, the 3He/4He ratios would be lower than the atmospheric value owing to the addition to groundwater of radiogenic helium derived from decay of U and Th in crustal rocks during geologic time. In contrast, near active volcano, primordial helium supplied from a magma system appears to have spread through the aquifer. However, the magmatic influence on the 3He/4He ratio in groundwater could be limited to several kilometers or 10 odd kilometers radius from the centers of active volcanoes [e.g., Sano et al., 1990; Sakamoto et al., 1992]. The epicenter of the main shock is located more than 20 km from the center of the Quaternary Daisen volcano, suggesting a relatively small contribution of magmatic helium. Exceptionally, an elevated 3He/4He ratio came from the Morogi well (Table 1, site 1), which is dug into debris avalanche derived from the Daisen volcano, suggesting the addition of fluids with rather high 3He/4He ratios in the erupted Quaternary dacite. [12] Figure 4 reveals that except for the sample from the Shimoenoki well (Table 1, site 14), water samples collected in the aftershock zone following the 2000 earthquake are characterized by 3He/4He ratios higher than the atmospheric value. It should be noted that the sample with 3He/4He ratios about four times higher than the atmospheric value came from the Sasabata well (site 10), the nearest well to the epicenter of the main shock. Although the Yokota volcanic rocks of early Pleistocene age are found near the Akaya well (site 6) and the Otani well (site 7) far from the aftershock distri5 of 10

35.374 35.327 35.338 35.346 35.356 35.295 35.285 35.247 35.295 35.279 35.264 35.241 35.259 35.227 35.224 35.198 35.191 35.229

133.372 133.359 133.326 133.301 133.276 133.256 133.231 133.238 133.311 133.346 133.475 133.412 133.384 133.408 133.389 133.384 133.348 133.316

Latitude Longitude dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved dissolved

Feature 5.98 1.26 1.71 1.41 1.55 1.12 0.94 1.15 1.28 6.20 1.47 2.50 3.12 0.55 1.70 2.66 1.08 1.61

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.08 0.03 0.03 0.03 0.04 0.02 0.02 0.02 0.02 0.10 0.06 0.04 0.04 0.01 0.04 0.04 0.03 0.04

3 He/4He ( 10 6) (±1s)

4.71 0.43 0.43 0.38 0.42 0.40 0.52 0.39 0.40 4.80 0.28 0.98 0.60 2.27 0.31 1.46 0.35 0.31

He/20Ne

4

4.49 0.77 1.56 1.05 0.77 0.45 0.37 0.50 1.28 4.66 1.64 2.08 3.15 0.32 2.27 2.11 0.15 1.88

He/4Hecor 3(R/RA)

3

4

267 8.64 7.16 7.77 13.1 9.98 26.0 7.34 20.1 118 3.08 14.3 9.98 46.8 4.9 34.2 8.46 4.41

He ( 10 8) (cm3STP/g)

20

56.6 20.1 16.7 20.5 31.3 25.2 49.8 18.7 50.8 24.6 11.0 14.7 16.5 20.6 15.9 23.4 24.5 14.2

Ne ( 10 8) (cm3STP/g) 49 46 48 48 48 46 48 50 48 49 49 53 52 54 53 53 53 52

8.4 7.7 7.9 8.1 8.2 7.9 8.6 8.6 8.4 8.4 8.5 9.2 8.8 9.1 9.0 8.9 8.7 8.6

dD d 18O (%) (%)

17.8 19.1 19.3 15.3 14.9 17.5 14.3 15.8 21.1 14.7 16.0 15.9 13.9 14.8 15.1 14.9 13.3 12.2

7.2 5.6 6.2 6.7 7.1 5.7 8.4 5.6 6.2 6.9 6.2 8.3 6.6 7.2 6.5 7.7 7.8 6.0

30.1 12.1 15.2 22.2 21.3 10.3 16.3 5.4 12.0 15.9 10.1 11.8 11.9 11.8 9.6 14.5 15.5 10.0

Temperature EC (°C) pH (mS/m)

The RAcor is calculated by subtracting atmospheric components using the method described by Craig et al. [1978]. The analytical error for 4He/20Ne is 15% of the values given.

Morogi Basara Hosshoji Yoichidani Haradai Akaya Otani Oriwatari Irikura Sasabata Sugasaki Yasuhara Fukuoka Shimoenoki Shimokurosaka Nakasuge Kamisuge Akibara

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

a

Site Name

G

3

Site Number

Table 1. Isotopic Compositions of the Well Water Sampled Around the Source Region of the 2000 Western Tottori Earthquakea

Geochemistry Geophysics Geosystems umeda and ninomiya: helium isotopes and concealed active faults

10.1029/2009GC002501

6 of 10


3

G


10.1029/2009GC002501

5.7 T.U. from the precipitation in Japan between 1970s and the 2000s [Yabusaki et al., 2003], the accumulated amount of tritiogenic helium by complete decay of 5.7 T.U. would produce 1.4 10 14 cm3 STP/g H2O [e.g., Schlosser et al., 1989]. Thus, tritiogenic helium could not be a significant contributor to the total 3He of hot spring gases, considering that 4He concentration is 4.9 10 8 cm3 STP/g H2O in air-saturated water [Ozima and Podosek, 2002]. Therefore, the observations described above are considered to provide significant evidence that the source fault of the 2000 earthquake is associated with leakage of mantle volatiles within the crust to the Earth’s surface.

Figure 4. Geographic distribution of the air-corrected 3 He/4He ratios in groundwaters around the seismic source region. Red stars and small black dots denote the epicenters of the main shock and aftershock events (30 days after each main shock), respectively. Also shown are fault segments (orange lines) activated by the main shock, which was resolved by Fukuyama et al. [2003].

bution, the 3He/4He ratios are markedly lower than the atmospheric value, indicating an insignificant contribution of mantle helium released from basaltic volcanic rocks. Figure 5 shows the 3He/4He ratios plotted as a function of distance from the main trace of the source fault. There is a clear trend of decreasing 3He/4He ratios with distance from the source fault of the 2000 earthquake. [13] The potential interpretation of significant 3He enrichment is thought to be due to either nucleogenic 3He produced by the decay of higher than average concentrations of 6Li, or contamination of groundwater by the decay of tritium from nuclear bomb tests. The 3He/4He production rate in the crust is expected to vary according to the abundance of the a-producing isotopes (232Th, 235U, 238 U) and those nuclides involved in 6Li (n, a), 3H (b ) and 3He reactions. Thus, the production rate was estimated to be 10 7 (0.1 RA) for average granitic crust [Morrison and Pine, 1955]. Consequently, nucleogenic 3He from the decay of 6Li cannot account for the high 3He/4He ratios in the groundwaters. With respect to tritiogenic helium, assuming an average tritium concentration of

[14] Geophysical findings of crustal heterogeneities may provide a clue to the mechanism for the generation of large intraplate earthquakes. In order to determine the crustal structure in the seismic source region, many geophysical studies have been carried out in the western Tottori area. Seismic tomography defined the details of the seismic velocity structure and the Vp/Vs ratio structure beneath the estimated earthquake fault [Zhao et al., 2004]. The results demonstrate that high P wave velocity anomalies occur in the upper crust, and low P and S wave velocity anomalies are detectable in the lower crust under the main shock hypocenter. Detailed magnetotelluric imaging also revealed that a conductive zone at depths greater than 15 km was imaged below the seismic source region [Oshiman, 2002]. In addition, deep low-

Figure 5. Helium isotopic composition plotted as a function of approximate distance from the main trace of the estimated fault segments. The thin cross represents sample from the Morogi well (Table 1, site 1) which was dug into debris avalanche derived from the Daisen volcano, suggesting the addition of fluids with rather high 3He/4He ratios in the erupted Quaternary dacite. 7 of 10


3

G


frequency (DLF) earthquakes occurred at depths of about 30 km in the seismic source region [Ohmi and Obara, 2002]. Waveform analyses indicated that a single-force source mechanism is more likely than a double-couple source mechanism for those DLF earthquakes, indicating that DLF events may be attributed to the transfer of fluids such as magma or aqueous fluid [Ukawa and Ohtake, 1987]. Therefore, these findings are considered to confirm that anomalously low P and S wave velocities and high electrical conductivities below the seismic source region away from volcanoes can be attributed to infiltration of deep fluids from the subcrustal lithosphere. This is most likely due to mantle fluids derived by the dehydration of the subducting Philippine Sea slab [Zhao et al., 2004] and/or the upwelling asthenosphere below the Philippine Sea slab [Nakajima and Hasegawa, 2007]. [15] In general, mantle fluid movement through the lower crust to the Earth’s surface remains obscure in a compressional tectonic regime because the ductile lower crust is usually considered a barrier to mantle volatiles [e.g., Hilton, 2007]. This raises the question of whether the seismically active source fault causing the 2000 earthquake acts as a preferential conduit for the transfer of mantle fluids through the ductile lower crust to the brittle upper crust. Recent reflection analysis of the aftershock waveforms in the source region of the 2000 earthquake shows that the strengths of the reflected waves are different on opposite sides of fault planes at depths of 10 to 25 km, suggesting that this source fault extends vertically down to the lower crust, i.e., deeper than the seismogenic upper crust [Doi and Nishigami, 2007]. Hypocenters of the DLF earthquakes are distributed on the downward extension of the seismogenic fault. Therefore, aqueous fluids, indicated by the DLF earthquakes, affect aseismic slip processes in the deeper portion of the source fault within the lower crust, and control the occurrence of the earthquake on the seismic portion of the fault within the upper crust [Ohmi et al., 2004]. Similar downward extension into the lower crust was also shown for the San Andreas fault and the Hayward fault in the northern California [Parsons and Hart, 1999]. In addition, the San Andreas fault transects the entire crust, on the basis of receiver function analysis [Zhu, 2000], that indicated the Moho is disrupted under the Eastern California Zone. Moreover, DLF events were observed along the San Andreas fault [Nadeau and Dolenc, 2005].

10.1029/2009GC002501

[16] As mentioned previously, crustal fluids with high 3He/4He ratios were recognized along the main trace of the source fault estimated by aftershock events related to the 2000 earthquake. Likewise, in the central to southern San Andreas fault system, crustal fluids collected from springs and wells have anomalously high 3He/4He ratios, indicating a significant contribution of mantle fluids to the near-field hydrologic system [Kennedy et al., 1997]. Therefore, it is expected that source faults extending down to the lower crust facilitate the effective transfer of mantle fluid to the Earth’s surface. If mantle fluids ascend in and/or along source faults within a ductile lower crust, the emanations from groundwaters with high 3He/4He ratios could be interpreted as a contribution of mantle helium derived from the subcrustal lithosphere. [17] Geological studies of natural ductile shear zones, such as mylonite zones, have established that there is clear tendency for fluid to flow into and along shear zones [e.g., Kerrich et al., 1984]. For example, the presence of anhydrous granulites in the exhumed lower crustal rock and xenoliths derived from the lower crust indicate that the lower crust is generally dry. In contrast, hydrous minerals are localized in the vicinity of ductile shear zones developed in the host anhydrous lithology [Yoshino, 2002]. It is indicated that at depths below the brittle-ductile transition, an approximate balance between rates of porosity destruction and porosity creation in shear zones can generate quasi-steady state permeability that are higher than those in the surrounding, less rapidly deforming rocks, leading to continuous fluid flow along ductile shear zones [Cox, 2002]. Numerical experiments studying pressure distributions in plastic and viscous media demonstrated that brittle shear zones have lower pressure compared to the surrounding rocks, whereas ductile zones have higher pressure [Mancktelow, 2006]. These results lead to the conclusion that ductile shear zones, connected to deep zones with near-lithostatic, pressurized fluids related to the DLF earthquakes, facilitate migration of fluids upward to the base of the seismogenic regime. Earthquakes occurring in the crust are generally thought to arise from the frictional instability of existing faults induced by coupling of increased hydraulic pressure as well as tectonic shear stress within the crust [Sibson, 1992]. The upwelling of fluids to the seismogenic upper crust could contribute to triggering the 2000 western Tottori earthquake by weakening the shear strength of the source fault. 8 of 10


3

G


[18] Accordingly, whether or not mantle fluids are involved with the dehydration of the subducting Philippine Sea slab or the upwelling asthenosphere below the Philippine Sea slab remains unclear, the transfer of mantle fluids to the shallow aquifer is considered to be efficient along the highly permeable rapture zone. Near the Earth’s surface, these fluids then mixed with groundwater derived from meteoric water, but without significant dilution by crustal radiogenic helium, resulting in the emanation of groundwaters with high 3He/4He ratios along the trace of the source fault. Our study indicates the clear relationship between helium isotope variations and seismogenic faults with downward extension to the ductile lower crust. In addition to geophysical observations, despite water samples collected from shallow drinking water wells, the helium isotope provides an excellent geochemical tracer to detect immature faults at an early stage of their evolution, i.e., concealed active faults without geomorphological and geological expression, but nevertheless, faults with the potential to cause large earthquakes.

Acknowledgments [19] We would like to thank K. Nagao and T. Iwanaga for helping with the helium isotope analyses. This manuscript was improved by the careful review of V. J. M. Salters, J. T. Kulongoski, G. F. McCrank, and anonymous reviewers.

References Aka, F. T., M. Kusakabe, K. Nagao, and G. Tanyileke (2001), Noble gas isotopic compositions and water/gas chemistry of soda springs from the islands of Bioko, Saõ Tome´ and Annobon, along with Cameroon Volcanic Line, West Africa, Appl. Geochem., 16, 323 – 338, doi:10.1016/S08832927(00)00037-8. Coleman, M. L., T. J. Shepherd, J. J. Durham, J. E. Rouse, and G. R. Moore (1982), Reduction of water with zinc for hydrogen isotope analysis, Anal. Chem., 54, 993 – 995, doi:10.1021/ac00243a035. Cox, S. F. (2002), Fluid flow in mid- to deep crustal shear systems: Experimental constraints, observations on exhumed high fluid flux shear systems, and implications for seismogenic processes, Earth Planets Space, 54, 1121 – 1125. Craig, H., J. E. Lupton, and Y. Horibe (1978), A mantle helium component in circum-Pacific volcanic gases: Hakone, the Marianas and Mt. Lassen, in Terrestrial Rare Gases, edited by E. C. Alexander and M. Ozima, pp. 3 – 16, Jpn. Sci. Soc. Press, Tokyo. Dalguer, L. A., K. Irikura, and J. D. Riera (2003), Simulation of tensile crack generation by three-dimensional dynamic shear rupture propagation during an earthquake, J. Geophys. Res., 108(B3), 2144, doi:10.1029/2001JB001738. Doi, I., and K. Nishigami (2007), Three-dimensional distributions of S wave reflectors in the source region of the 2000 western Tottori earthquake, Geophys. Res. Lett., 34, L20312, doi:10.1029/2007GL031169.

10.1029/2009GC002501

Fukuyama, E., W. L. Ellsworth, F. Waldhauser, and A. Kubo (2003), Detailed fault structure of the 2000 western Tottori, Japan, earthquake sequence, Bull. Seismol. Soc. Am., 93, 1468 – 1478, doi:10.1785/0120020123. Fusejima, Y., T. Yoshioka, K. Mizuno, M. Shishikura, R. Imura, T. Komataubara, and T. Sasaki (2001), Surface ruptures associated with the 2000 Tottori-ken Seibu earthquake (in Japanese with English abstract), Annu. Rep. on Active Fault and Paleoearthquake Res., 1, pp. 1 – 26, Geol. Surv. of Jpn., Tsukuba. Geological Survey of Japan (1995), Geological map of Japan, Digital Geosci. Map G-1, 3rd ed., scale 1:1,000,000, Geol. Surv. of Jpn., Tsukuba. Gu¨leç, N., D. R. Hilton, and H. Mutlu (2002), Helium isotope variations in Turkey: Relationship to tectonics, volcanism and recent seismic activities, Chem. Geol., 187, 129 – 142, doi:10.1016/S0009-2541(02)00015-3. Hattori, H., and M. Katada (1964), Explanatory text of the geological map of Japan, 49 pp., Geol. Surv. of Jpn., Tsukuba. Hilton, D. R. (2007), Geochemistry—The leaking mantle, Science, 318, 1389 – 1390, doi:10.1126/science.1151983. Inoue, D., K. Miyakoshi, K. Ueta, A. Miyawaki, and K. Matsuura (2002), Active fault study in the 2000 Tottori-ken seibu earthquake area (in Japanese with English abstract), J. Seismol. Soc. Jpn., 54, 557 – 573. Johnston, A. C. (1996), Seismic moment assessment of earthquake in stable continental regions III. New Madrid 1811 – 1812, Charleston 1886, Lisbon 1755, Geophys. J. Int., 126, 314 – 344, doi:10.1111/j.1365-246X.1996.tb05294.x. Kanamori, H. (1972), Determination of effective tectonic stress associated with earthquake faulting, the Tottori earthquake of 1943, Phys. Earth Planet. Inter., 5, 426 – 434, doi:10.1016/ 0031-9201(72)90114-8. Kennedy, B. M., Y. K. Kharaka, W. C. Evans, A. Ellwood, D. J. DePaolo, J. Thordsen, G. Ambats, and R. H. Mariner (1997), Mantle fluids in the San Andreas fault system, California, Science, 278, 1278 – 1281, doi:10.1126/science. 278.5341.1278. Kerrich, R., T. E. Latour, and L. Willmore (1984), Fluid participation in deep fault zones: Evidence from geological, geochemical, and 18 O/ 16O relations, J. Geophys. Res., 89(B6), 4331 – 4343, doi:10.1029/JB089iB06p04331. King, C. Y. (1986), Gas geochemistry applied to earthquake prediction: An overview, J. Geophys. Res., 91(B12), 12,269 – 12,281, doi:10.1029/JB091iB12p12269. Kulongoski, J. T., D. R. Hilton, and J. A. Izbicki (2005), Source and movement of helium in the eastern Morongo groundwater Basin: The influence of regional tectonics on crustal and mantle helium fluxes, Geochim. Cosmochim. Acta, 69, 3857 – 3872, doi:10.1016/j.gca.2005.03.001. Mancktelow, N. S. (2006), How ductile are ductile shear zones?, Geology, 34, 345 – 348, doi:10.1130/G22260.1. Matsuda, T. (1975), Magnitude and recurrence interval of earthquakes from a fault (in Japanese with English abstract), J. Seismol. Soc. Jpn., 28, 269 – 284. Mizota, C., and M. Kusakabe (1994), Spatial distribution of dD-d 18O values of surface and shallow groundwaters from Japan, south Korea and east China, Geochem. J., 28, 387 – 410. Morrison, P., and J. Pine (1955), Radiogenic origin of the helium isotopes in rocks, Ann. N. Y. Acad. Sci., 62, 71 – 92, doi:10.1111/j.1749-6632.1955.tb35366.x. Nadeau, R. M., and D. Dolenc (2005), Nonvolcanic tremors deep beneath the San Andreas Fault, Science, 307, 389, doi:10.1126/science.1107142.

9 of 10


3

G


Nakajima, J., and A. Hasegawa (2007), Tomographic evidence for the mantle upwelling beneath southwestern Japan and its implications for arc magmatism, Earth Planet. Sci. Lett., 254, 90 – 105, doi:10.1016/j.epsl.2006.11.024. Nakata, T., and T. Imaizumi (2002), Digital Active Fault Map of Japan [DVD-ROM], 60 pp., Univ. of Tokyo Press, Tokyo. Ohmi, S., and K. Obara (2002), Deep low-frequency earthquakes beneath the focal region of the Mw 6.7 2000 western Tottori earthquake, Geophys. Res. Lett., 29(16), 1807, doi:10.1029/2001GL014469. Ohmi, S., I. Hirose, and J. Mori (2004), Deep low-frequency earthquakes near the downward extension of the seismogenic fault of the 2000 western Tottori earthquake, Earth Planets Space, 56, 1185 – 1189. Oshiman, N. (2002), Conductivity structure beneath the western part of Japan (in Japanese), Earth Mon., 38, 82 – 90. Ozima, M., and F. A. Podosek (2002), Noble Gas Geochemistry, 2nd ed., 286 pp., Cambridge Univ. Press, Cambridge, U. K. Parsons, T., and P. E. Hart (1999), Dipping San Andreas and Hayward faults revealed beneath San Francisco Bay, Calif. Geol., 27, 839 – 842. Sagiya, T., T. Nishimura, Y. Hatanaka, E. Fukuyama, and W. L. Ellsworth (2002), Crustal movements associated with the 2000 western Tottori earthquake and its fault models (in Japanese with English abstract), J. Seismol. Soc. Jpn., 54, 523 – 534. Sakamoto, M., Y. Sano, and H. Wakita (1992), 3He/4He ratio distribution in and around the Hakone Volcano, Geochem. J., 26, 189 – 195. Sano, Y., and H. Wakita (1985), Geographical distribution of 3 He/4He ratios in Japan: Implications for arc tectonics and incipient magmatism, J. Geophys. Res., 90, 8729 – 8741, doi:10.1029/JB090iB10p08729. Sano, Y., H. Wakita, and S. N. Williams (1990), Heliumisotope systematics at Nevado del Ruiz volcano, Colombia: Implications for the volcanic hydrothermal system, J. Volcanol. Geotherm. Res., 42, 41 – 52, doi:10.1016/0377-0273(90) 90068-Q. Schlosser, P., M. Stute, C. Sonntag, and K. O. Mu¨nnich (1989), Tritiogenic 3He in shallow groundwater, Earth Planet. Sci. Lett., 94, 245 – 256, doi:10.1016/0012-821X(89)90144-1.

10.1029/2009GC002501

Sibson, R. H. (1992), Implications of fault-valve behaviour for rupture nucleation and recurrence, Tectonophysics, 211, 283 – 293, doi:10.1016/0040-1951(92)90065-E. Ueta, K., K. Tani, and T. Kato (2000), Computerized X-ray tomography analysis of three-dimensional fault geometries in basement-induced wrench faulting, Eng. Geol., 56, 197 – 210, doi:10.1016/S0013-7952(99)00143-X. Ukawa, M., and M. Ohtake (1987), A monochromatic earthquake suggesting deep-seated magmatic activity beneath the Izu-Ooshima volcano, Japan, J. Geophys. Res., 92, 12,649 – 12,663, doi:10.1029/JB092iB12p12649. Umeda, K., A. Ninomiya, and G. F. McCrank (2008), High 3 He emanations from the source regions of recent large earthquakes, central Japan, Geochem. Geophys. Geosyst., 9, Q12003, doi:10.1029/2008GC002272. Umeda, K., A. Ninomiya, and T. Negi (2009), Heat source for an amagmatic hydrothermal system, Noto Peninsula, Central Japan, J. Geophys. Res., 114, B01202, doi:10.1029/ 2008JB005812. Yabusaki, S., M. Tsujimura, and N. Tase (2003), Recent trend of tritium concentration in precipitation in Kanto Plane, Japan (in Japanese with English abstract), Bull. Terr. Environ. Res. Cent. Univ. Tsukuba, 4, 119 – 124. Yamaguchi, S., H. Murakami, H. Iwamoto, K. Takemoto, K. Kitada, I. Shiozaki, N. Oshiman, and S. Katoh (2007), Two-dimensional resistivity structure of the fault associated with the 2000 western Tottori earthquake, Earth Planets Space, 59, 1211 – 1217. Yoshida, N., and Y. Mizutani (1986), Preparation of carbon dioxide for oxygen-18 determination of water by use of a plastic syringe, Anal. Chem., 58, 1273 – 1275, doi:10.1021/ ac00297a071. Yoshino, T. (2002), Role of water in conductive anomalies and seismic reflections in the lower crust, Bull. Earthquake Res. Inst. Univ. Tokyo, 76, 479 – 500. Zhao, D., H. Tani, and O. Mishra (2004), Crustal heterogeneity in the 2000 western Tottori earthquake region: Effect of fluids from slab dehydration, Phys. Earth Planet. Inter., 145, 161 – 177, doi:10.1016/j.pepi.2004.03.009. Zhu, L. (2000), Crustal structure across the San Andreas Fault, southern California from teleseismic converted waves, Earth Planet. Sci. Lett., 179, 183 – 190, doi:10.1016/S0012821X(00)00101-1.

10 of 10