MT Study of Mt Fugi Hydro Thermal System AizawaEPSL05

8/8/2019 MT Study of Mt Fugi Hydro Thermal System AizawaEPSL05

1/13

Hydrothermal system beneath Mt. Fuji volcano inferred from

magnetotellurics and electric self-potential

K. Aizawaa,*, R. Yoshimuraa, N. Oshimana, K. Yamazakia, T. Utoa, Y. Ogawab,S.B. Tankb, W. Kandac, S. Sakanakad, Y. Furukawad, T. Hashimotoe, M. Uyeshimaf,

T. Ogawaf, I. Shiozakig, A.W. Hursth

aResearch Center for Earthquake Prediction, Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japanb

Volcanic Fluid Research Center, Tokyo Institute of Technology, Ookayam 2-12-2, Meguro-ku, Tokyo 152-8551, JapancSakurajima Volcano Research Center, Kyoto University, Yokoyama Tsurusaki 1722-19, Sakurajima, Kagoshima 891-1419, Japan

dInstitute of Applied Earth Sciences, Faculty of Engineering and Resource Science, Akita University, Tegata-gakuen 1-1, Akita 010-8502, JapaneInstitute of Seismology and Volcanology, Graduate School of Science, Hokkaido University, Sapporo Kita 10 Nishi 8 cho-me,

Sapporo 060-0810, JapanfEarthquake Research Institute, University of Tokyo, Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-0032, Japan

gDepartment of Civil Engineering, Faculty of Engineering, Tottori University, Koyama-cho minami 4-101, Tottori 680-8552, JapanhInstitute of Geological and Nuclear Sciences, P.O. Box 30368, Lower Hutt, New Zealand

Received 28 May 2004; received in revised form 10 December 2004; accepted 8 March 2005

Available online 17 June 2005

Editor: V. Courtillot

Abstract

Wideband magnetotelluric (MT) soundings were carried out on Mt. Fuji volcano along a northeast to southwest axis. It was

found by two-dimensional inversion using the highest quality data (in the frequency range 1300 Hz) that a good conductor

(resistivity of approximately a few ohm m) was located beneath the summit with a lateral extent of approximately 4 km. It begins

approximately 1 km below the ground surface; however, its depth cannot be resolved. In our previous study, an intense positive

self-potential (SP) anomaly (approximately 2000 mV), was found around a summit crater having a diameter of approximately 3

km. We interpreted the presence of the good conductor and positive SP anomaly as a strong indication of an active hydrothermal

system. Subsequently, we searched for conduction current sources to explain the SP distribution on the surface by using the

resistivity structure determined by the MT inversion. The results obtained were that a positive conduction current source of the

order of 1000 A should be located at the top of the conductor. From these results, we deduced that the conductor represents ahydrothermal system in which single-phase (liquid) convection is taking place. Since the resistivity at a distance from the good

conductor can be explained by the effect of cold groundwater, the hydrothermal system does not seem to extend throughout the

entire body of the volcano, but seems to be confined to the area beneath the summit crater. Finally, an estimate of the order of

magnitude of the subsurface hydrothermal flow was performed using a relation between the fluid volume flux and electric current

0012-821X/$ - see front matterD 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.epsl.2005.03.023

* Corresponding author. Volcanic Fluid Research Center, Tokyo Institute of Technology, Ookayam 2-12-2, Meguro-ku, Tokyo 152-8551,

Japan.

E-mail address: [email protected] (K. Aizawa).

Earth and Planetary Science Letters 235 (2005) 343355

www.elsevier.com/locate/epsl


2/13

density in the capillary model. The result suggested that there exists fairly low permeability within the shallow part of Mt. Fuji.

We speculate that the low permeability in the volcano has a correlation with the confinement of the hydrothermal system and

quiescence of volcanic activities, such as low seismicity, no gas emanations, and no natural hot springs.

D 2005 Elsevier B.V. All rights reserved.

Keywords: self potential; magnetotellurics; hydrothermal system; resistivity structure; volcano; permeability

1. Introduction

In recent times, the importance of investigating the

upper portion of the structure of volcanoes has been

recognized. This is because the interaction between

the magma and groundwater within the top few kilo-

meters might be the cause of precursory phenomenaand can control the type of eruption [1]. Most active

volcanoes have a hydrothermal system near the sur-

face, which transports heat from its depths to the

surface by convection. Hydrothermal systems in the

case of active volcanoes have been studied by various

geophysical and geochemical methods [26]. On the

other hand, hydrothermal systems are less understood

in the case of apparently dormant volcanoes, which

have no significant geothermal manifestations and

low seismicity. For a comprehensive understanding

of volcanic systems, it is important to know whether

the shallow part of apparently dormant volcanoes

differs from that of more active volcanoes. Further,

since apparently dormant volcanoes can erupt abrupt-

ly [7], the study of these volcanoes might be useful forgreater accuracy in prediction of eruptions.

Mt. Fuji, which is the highest (3776 m above sea

level) and largest stratovolcano in Japan, might be an

ideal volcano for studying such problems because of

its large eruption potential and present quiescence.

Mt. Fuji has had many historic eruptions, erupting at

least 17 times since 781 AD, and the most recent

eruption occurred in 1707 when approximately 1.7

Fig. 1. Self-potential distribution of Mt. Fuji [10] with self-potential (SP) and magnetotelluric (MT) observation sites. Topographic contours are

also shown. The black dots and star show the survey points and reference point of the SP surveys, respectively. The white triangles with

characters show MT survey sites and the site names. The thick lines indicate the locations of the summit crater and Hoei crater.

K. Aizawa et al. / Earth and Planetary Science Letters 235 (2005) 343355344


3/13

km3 ash was emitted and the Hoei crater was formed

on the southeastern flank 3 km away from the summit.

It is reported that the last eruption from the summit

occurred more than 2200 yrs ago [8]. Throughout itsentire history of approximately 100,000 yrs, more

than 99% of the erupted products of Mt. Fuji was

found to be basaltic. From geological study, it is

known that an old Quaternary volcanic body (10ka), which is 2400 m high above sea level, with high

SiO2 content (5160%) is superposed by Fuji ejecta

[9]. At present, Mt. Fuji is dormant. In recent years,

neither fumaroles nor geothermal manifestation can be

seen anywhere on the surface. Natural hot springs do

not exist on the outer areas of the mountain. Shallow

seismicity under Mt. Fuji is also low. Low-frequencyearthquakes are regularly observed at depths of 15 km

beneath the summit; however, no other related volca-

nic activity has been observed.

Recently, self-potential (SP) surveys revealed an

intense positive anomaly around the summit and

suggested the presence of a hydrothermal system

within Mt. Fuji [10]. Fig. 1 shows the SP distribution

using a reference point that is to the west of the Hoei

crater. A positive SP anomaly of approximately 3 km

diameter is centered on the summit crater and has an

amplitude of approximately 2000 mV. A remarkable

btopographic effectQ [11], in which SP has a negative

correlation of approximately 0.5 to 1 (mV/m)with an elevation due to the electrokinetic effect of

downward groundwater flow, was also observed at

elevations lower than 2000 m [10].

In this study, we investigate the shallow resistivity

structure obtained by the magnetotelluric (MT) sound-

ing and discuss the hydrothermal system of Mt. Fuji

using the SP distribution. Since the resistivity structure

might affect the SP distribution, a combined analysis is

more realistic for studying the dynamics of the hydro-

thermal system within the volcano.

2. Shallow resistivity structure of Mt. Fuji

2.1. Data acquisition and processing

In order to image the hydrothermal system that has

been suggested by the previous SP surveys [10], we

performed wideband MT soundings on a survey line

trending northeast to southwest (Fig. 1). We also

acquired data of the Hoei crater (H0 in Fig. 1). The

data was acquired during a period from 9 September

to 19 September 2002. The typical recording duration

for one site was 5 to 10 days. Natural electric andmagnetic fields in the frequency range of 300 to

0.0005 Hz were measured at the surface using Phoe-

nix MTU5 and MTU2E systems. Since the leakage

electric current from railway lines seriously contam-

inates the data in the lower frequency range, we used

remote reference processing [12] to calculate the

noise-free impedance by using geomagnetic data

that was recorded in northern Hokkaido (1000 km

north of Mt. Fuji). This procedure modified the im-

pedance processed at a single site in the frequency

range below 10 Hz. In this study, we used the highest-quality impedances in the frequency range between 1

and 300 Hz. At higher elevation sites (over 2200 m),

we did not record geomagnetic data; therefore, we

used geomagnetic data from other sites to calculate

the impedances.

2.2. Dimensionality and distortion analysis

Fig. 2 shows the impedance skew [13] for all sites.

The skews are less than 0.1 except at the sites M0,

M1, and M2, which are located at a high elevation on

the mountain; this suggests a regional 2D or 1D

resistivity structure. However, in the summit area,

the presence of a local three-dimensional distortion

or 3D structure is suggested.

We investigated the dimensionality of the data

using the tensor decomposition technique of Groom

and Bailey [14]. In order to investigate whether the

Fig. 2. The impedance skews for all sites. The curves for the summit

sites, M0, M1, and M2 (as shown in Fig. 1) are individually labeled.

K. Aizawa et al. / Earth and Planetary Science Letters 235 (2005) 343355 345


4/13

structure is regionally 2 D with local shallow 3D

anomalies, distortion parameters were first determined

with no constraints (i.e., site-dependent and frequency-

dependent constraints). Fig. 3 shows the histograms ofthe estimated regional strike in two frequency ranges.

The cumulative estimated strikes slightly scatter at

northeastern sites; however, they basically converge

in one direction suggesting that there exists a NWSE

trending regional 2D resistivity structure within the

shallow part of Mt. Fuji. The distortion parameters

(twist and shear) were also determined to be frequency

independent (maximum fluctuation is less than 38) on

each site.

2.3. Two-dimensional analysis

Mt. Fuji has numerous parasitic cones trending a

NWSE direction [9], suggesting that the subsurface

structure is two dimensional with a NW SE trend.

The estimated regional strike directions (Fig. 3) are

in agreement with this geological trend. In this study,

we assumed the regional strike direction (parallel to

the 2D structure) to be N45W and conducted a 2D

inversion using the code of Ogawa and Uchida [15].

After fixing the N45W regional strike direction, all

the impedances were decomposed using site-depen-

dent and frequency-independent distortion para-

meters. Subsequently, we inverted the apparent

resistivity and the phase in the decomposed TM

mode (electric current flowing across the structure)

because the TM data is more robust in the presence

of three-dimensional anomalous structures [16]. The

starting model had a uniform resistivity of 100 V m

as well as 2D topography. The different locations of

electric and magnetic fields at certain observation

sites were taken into account in the inversion. Fig.

4 shows comparisons between the observed and

calculated data sets. The overall fit was good forall sites.

Fig. 5 shows the best-fit resistivity model. A good

conductor with a lateral extent of approximately 4 km

is located approximately 1 km beneath the summit.

We performed two forward calculations (sensitivity

tests) in order to assess the uniqueness of the conduc-

tor with modifications to the best-fit resistivity. First,

we changed the resistivity value of the conductor to a

uniform 30 V m value. Secondly, we placed the top of

the conductor at a depth of 700 m. Since the calcu-

lated sounding curves do not fit the observed sound-ing curves in the sites around the summit in either case

(Fig. 6), our model of a conductor with a resistivity of

approximately 10 V m located 1 km beneath the

summit is confirmed. Due to attenuation of the elec-

tromagnetic field penetrating into the body of the

conductor, the downward extent of this conductor is

uncertain.

Since the topography of Mt. Fuji is 3D and the SP

data suggest the presence of a 3D local structure

beneath the summit, the 2D modeling is not strictly

valid. However, when the 3D conductive structure is

located below a survey profile and has a horizontal

extent greater than one-half of a skin depth, it is found

that the effects of finite horizontal extent of the 3D

conductor in a 2D analysis are not significant [17].

The shallow resistivity structure of Mt. Fuji is inter-

preted as a regional 2D structure with an imbedded

3D conductor beneath the summit. The SP distribu-

tion, which suggests the presence of a hydrothermal

Fig. 3. Histograms for the regional strikes from Groom and Bailey [14] tensor decompositions in two frequency ranges. Distortion parameters

were set to be site-dependent and frequency-dependent. Note that the 908 ambiguity is included in each histogram. Site names are shown above

each histogram.



5/13

system beneath the summit, might support this inter-

pretation. Therefore, we interpret the best-fit 2D re-

sistivity structure as a first-order cross section of a 3D

structure for which some features might have a finite

horizontal extent. Although we modeled the 3D to-

pography as 2D, this will not enhance the resistivity

value of the good conductor because we used only the

TM mode in the 2D model and replaced air with finite

resistivity blocks. The 3D modeling will be the sub-

ject for future study.

2.4. Features of the resistivity structure

Besides the summit area, the shallow resistivity

structure of Mt. Fuji is essentially characterized by 3

layers. The first layer is a surface layer with a

thickness of several hundred meters and a resistivity

of approximately 1000 V m. The result of the sci-

entific drilling [18], which was carried out on the

northeast flank (1 km northwest of Y3) in 2003,

showed that the first layer is a porous and unsatu-

Fig. 4. Comparison between calculated and observed sounding curves. Only the decomposed TM mode data was used for 2D inversion. The

solid circles indicate observed responses with one standard deviation. The solid lines indicate calculated responses from the best-fit resistivity

model. Site names are shown above each panel.



6/13

rated layer. Core samples collected at a depth of 650

m revealed that the composition was primarily mud-

flow deposits with three sheets of lava flows. The

well-logging data showed that the porosity was from

10% to 40% and the resistivity was from 500 to

1000 V m. Further, the results revealed that aquifers

did not exist in the first layer because the water

pressure did not rise significantly during the drilling.

The second layer is located at depths between

several hundred meters and 2 km and is relatively

conductive (resistivity of approximately 50 to 200 V

m). Since the resistivity of the spring water at 18 8C

is between 50 and 200 V m at any place on the outer

surface of Mt. Fuji [19], the second layer is inter-

preted as an aquifer where a large quantity of ground-

water saturates the host rock. The third layer is the

layer that is deeper than 2 km and it shows resistivity

from 500 to 1000 V m. This layer might represent a

water poor region in which the cracks were closed by

pressure.

The first and second layers are more conductive by

one order of magnitude within a radius of about 2 km

around the summit area than in other places at the

same depths. The positive SP anomaly, whose radius

is about 1.5 km, is located above this conductor ( Fig.

1). This relation between a conductive zone and pos-itive SP anomaly is similar to that for the Miyakejima

volcano [20,21], where changes of the hydrothermal

system are suggested to precede eruption activity [21].

Since anomalous crustal movements have not been

detected around Mt. Fuji and the seismicity is low,

there is no reason to believe that the good conductor is

a magma body. In this study, we interpret the presence

of a good conductor and a positive SP anomaly as an

indication of an active hydrothermal system. The

existence of active fumaroles at approximately 80

Fig. 5. Best-fit resistively model obtained by the 2D inversion of the MT data. The decomposed TM mode data were used for inversion. The 2D

regional strike direction (parallel to the 2D structure) is assumed to be the NWSE direction. The inverted triangles with characters indicate the

measurement sites. Note that the good conductor beneath the summit prevents the penetration of the electromagnetic field into the region beneath the summit.

Fig. 6. The results of sensitivity tests at site M0. The solid black dot

shows observed data, and the solid line represents a sounding curve

calculated from the best-fit resistivity structure (Fig. 5). The heavy

dashed line shows the sounding curve calculated from the structure

on which the resistivity of the good conductor, which is located at a

depth from 1 to 2.4 km and has a horizontal extent of 4 km, was

changed to a uniform 30 V m. The fine dashed line represents the

sounding curve calculated from the structure on which the conduc-

tors top was made 700 m deep with 5 V m resistivity blocks.



7/13

8C on the summit 70 yrs ago [9] might support the

theory of the presence of the hydrothermal system.

Although several SP-generating mechanisms (e.g.,

electrochemical, thermoelectric, electrokinetic, andrapid fluid disruption) have been proposed [22,23],

it is widely believed that only the electrokinetic (EK)

effect can be the source of large SP anomalies of the

order of volts [24]. Hereafter, we interpret the SP

anomaly on the basis of the EK effect.

3. The modeling of SP sources

3.1. Electrokinetic potential

The flow of a fluid through a porous medium gen-

erates an electrical potential gradient (referred to as

belectrokineticQ orbstreaming potentialQ). Sill [25] pre-

sented a method for the investigation of self-potential

based on induced current sources. The general relations

between the electric current density J (A m 2), the

fluid volume flow density (Darcy velocity) U (m3

m2 s1), the electric potential U (V), and the hy-

draulic potential n (Pa) are expressed by the coupled

equations [26] in the case of static conditions as

follows:

J L11jU L21jn; 1

U L21jU L22jn; 2

where r=L11, L21, and L22 represent the electrical

conductivity (S m1), the cross coupling coefficient

(m2 s1 V1), and the hydraulic conductivity (m2

Pa1 s1), respectively. When the effects of the sec-

ondary electric potentials on the fluid flow are small,

the primary flow equation is decoupled and the result-

ing equations are as follows:

U L22jn Darcy0s law ; 3

and

J rjU L21jn: 4

The first term on the right-hand side of (4)

(rjU) represents a conduction current (Jcond) driv-en by the electric potential, while the second term on

the right-hand side of (4) ( L21jn) represents aconvection current (Jdrag) driven by fluid flow. The

electric potential difference that is measured on the

surface is related only to the conduction current. Since

external current sources can be neglected in static

conditions,

jd J 0; 5

and by substituting this in Darcys law results, we

obtain

jd Jcond jdJdrag jL21

L22

dU

L21

L22jdU: 6

Thus, sources of conduction current exist wherever

there are gradients of the cross-coupling coefficient or

permeability parallel to the fluid flow or wherever

there are external or induced sources of the fluid

flow. This mechanism is schematically illustrated inFig. 7. Since it can be concluded that the self-potential

measured on the surface is generated from conduction

current sources, an analysis that focuses on the fea-

tures of conduction current sources would provide

information on the subsurface fluid flow.

3.2. Estimation of the SP sources

The fundamental purpose of the SP and MT sur-

veys is to elucidate the subsurface hydrothermal flow.

However, achieving this is difficult because of the

Fig. 7. Schematic illustration of the generation of a streaming

potential. A situation of uniform fluid flow perpendicular to plane

boundaries in a static condition is shown. In the porous media,

electric charge is conveyed by convection (convection current);

however, no charge is conveyed outside the porous media. There-

fore, the charge accumulation occurs on both sides of the porous

media, and the conduction current flows in the direction opposite

the convection current to cancel the charge separation. The voltage

difference is caused by the conduction current.



8/13

large uncertainties in the physical parameters, such as

permeability, cross-coupling coefficient, resistivity,

and phase of fluids. However, In this study, the resis-

tivity structure is known, and relatively realistic mod-eling can be performed. We conducted an approximate

estimate of the conduction current sources by forward

modeling that included the resistivity structure. The

purpose is to deduce the spatial relationship between

the location of the SP source and resistivity structure

and to investigate the subsurface electric charge accu-

mulation process.

The forward code used in this study is based on the

finite-difference schemes proposed by Dey and Mor-

rison [27]. In forward modeling, we used the simple

structure, as shown in Fig. 8, representing the primaryfeatures of the best-fit resistivity model. The good

conductor was expressed as 8 V m blocks and was

located at depths greater than 1 km. We interpreted the

best-fit 2D resistivity structure (Fig. 5) as the first-

order cross section of the 3D structure that has a finite

horizontal extent and assumed that the 3D structure is

symmetric around the summit. Although theseassumptions are inexact, they might be sufficient for

an approximate estimation of the SP sources. The

mesh consisted of 73 73 horizontal points (onegrid point interval is 150 m) and 100 vertical points

(one grid point interval is 100 m) and air was defined

as a highly resistive zone (resistivity of 3 1013 V m).For simplicity, the conduction current sources are

represented by a pair of point sources of positive and

negative polarities that is located in a vertical line and

their locations, depths, and intensity were changed in

order to fit the observed SP data on the surface. Wealso incorporated the position of the SP reference

point and the btopographic effectQ (TE) coefficient

as model parameters. The TE coefficient, as noted in

Fig. 8. 3D electrical resistivity model used in the forward modeling for the SP source estimation. X, Y, and Z axes denote the east, north, and

vertical direction, respectively. The mesh consisted of 73 73 horizontal points (one grid point interval is 150 m) and 100 vertical points (onegrid point interval is 100 m). The color scale bar is identical to the one in Fig. 5.



9/13

the introduction, is assumed to be a linear function of

elevation, and the calculation was performed with four

TE coefficients (Cases 1 to 4) 0, 0.5, 1, and 1.5

(mV/m). We also performed the calculation withoutthe constraint of the TE coefficient (Case 5) in order

to estimate a plausible range for the TE coefficient. In

each case, the appropriate model parameters are de-

termined by a grid-search technique in order to min-

imize the RMS-misfit between the observed and

calculated SP on the surface. The searched space

was the range that extends 1500 m horizontally and

5300 m vertically from the center of the summit crater.

The modeling procedure is similar to the analytical

approach of Kanda and Mori [28], which assumed that

the resistivity structure is uniform and current sourcesare located just beneath the fumarolic vent in the

summit. However, this study considers the resistivity

structure, and the horizontal source locations are not

constrained.

Table 1 shows the determined parameter range and

RMS misfit for each case. We interpreted those para-

meters among the results that produced an equivalent

misfit level (RMS in the best-fit model + 0.002 for

each TE coefficient) as the reasonable parameter

range. The smallest RMS is obtained in Case 3, and

this suggests that a TE coefficient of 1.0 mV/m isreasonable in the case of Mt. Fuji. Since the deter-

mined horizontal source locations were within a circle

with a diameter of 300 m and a center located 400 m

northwest of the summit crater center in all cases, only

the depth of the source is shown in the table. Consid-

ering that the RMS in Case 1 is relatively large, the

depth of the positive source is determined to be

between 1.0 and 1.5 km. This is an important result

because this depth corresponds to the top of the good

conductor (Figs. 5 and 8). In addition, the strength of

the sources was determined to be of the order of 1000

A, and the negative source was determined to be

deeper. With regard to the depth of the positive source

and strength of the sources, the results did not varywhen we placed two sources obliquely instead of in a

vertical line and searched the current sources that

express an asymmetric positive SP anomaly on the

surface by trial and error.

4. Discussion and conclusion

When the presence of a hydrothermal system on

Mt Fuji was suggested, it seemed reasonable that it

would merely be a single-phase (liquid) hydrothermalsystem because there is no gas emanation on the

surface. The laboratory experiments show that the

zeta potential, which primarily contributes to convec-

tion currents, depends on the temperature [26,29]. By

numerical modeling, Ishido and Pritchett[30] showed

that positive charges accumulate due to this tempera-

ture dependence of zeta potential in the zone where

the upwelling fluid begins to move sideways and

cools. The positive charges produce a positive SP

anomaly on the surface in a single-phase (liquid)

case. The results of our forward modeling that the

conduction current sources are primarily accumulated

at the top of the conductor are consistent with these

results of Ishido and Pritchett [30]. By considering

that either a rock with a large quantity of hot water or

a hydrothermal altered rock is highly conductive

[31,32], we interpret that the good conductor beneath

the summit represents the hot-water saturated and

altered zone in which single-phase convection is

occurring. The schematic model for the hydrothermal

system beneath Mt. Fuji is shown in Fig. 9. At the top

Table 1Estimated source parameter range by forward calculation in which two conduction current sources in a vertical line are imposed on the

resistivity structure shown in Fig. 8

Increment Case 1 Case 2 Case 3 Case 4 Case 5

Topographic effect (mV/m) 0.1 0 (fix) 0.5 (fix) 1 (fix) 1.5 (fix) 0.8 to 1.2Source intensity (A) 10 70330 2901230 8701500 13401387 7801880

Depth of a positive source (km) 0.1 0.51.1 0.91.5 1.11.5 1.21.5 1.11.7

Depth of a negative source (km) 0.1 1.55.3 3.35.3 4.44.9 4.55.1 4.15.3

Best-fit RMS (V) 0.3092 0.3035 0.3002 0.3024 0.3002

The results that produced the equivalent misfit level (RMS bRMS in best-fit model+0.002) were shown in each case. The grid search was

performed in five cases that assume the topographic effect (TE) coefficients as 0, 0.5, 1, 1.5 (mV/m), and unconstrained (Case 5). In thecalculation without the constraint of TE coefficients, 1.0 mV/m was the best-fit TE coefficient.



10/13

of the hot-water saturated zone, the upwelling flow

turns sideways and cools; hence, the positive current

sources accumulate and generate the positive SP

anomaly on the Earths surface. The hydrothermal

system probably does not extend within the whole

volcano because the resistivity value outside the good

conductor can be explained by cold (18 8C) ground-

water content. It should be noted that the current

discussion holds only when the upwelling zone does

not have a large lateral extent because we expressed

the charge accumulation process using point currentsources.

Under static conditions, we can conduct an order of

magnitude estimation of the fluid mass flux by making

some more assumptions. When fluids are incompress-

ible and dehydration from magma does not occur,

there is no divergence of fluids and the current source.

Hence, Eq. (6) reduces to the following equation:

jdJcond jL21

L22

d U: 7

When the capillary model [26] is considered, the

equation becomes the following:

jdJcond j gt2enk

dU; 8

where g, t, and k denote porosity, tortuosity, and

intrinsic permeability (m2) of the porous media, re-

spectively. e denotes the dielectric constant of the

fluids (F m1), and f denotes the zeta potential (V).

Ifg, t, k, e, and U are constant in the region of charge

accumulation, we obtain

Jscond gt2eDf

kU; 9

where Jscond and U denote source intensity (A) and

fluid volume flux (m3 s1). Df denotes the difference

in the zeta potential through the direction of fluid flow.

When we assume the order of magnitude of each

parameter as gt2=0.1, e= 1010, k= 1014 [28],

Df= 0.1 [26], and Jscond=1000 (Table 1), the fluidmass flux is calculated to be of the order of 106 (ton/

day). If we assume the area of the upwelling path to

be from 104 to 106 (m2), the speed of the fluid

upwelling is calculated to be approximately 1 to

100 (m/day). Although the large uncertainty of phys-

ical parameters, especially in permeability, makes

this estimation somewhat uncertain, the speed of

the fluid seems to be extremely high. In the case

of Mt. Fuji, a considerably lower permeability might

be more realistic.

Even in the apparently dormant Mt. Fuji, the resis-

tivity structure seems to require hydrothermal circu-

lation as is suggested in the case of certain active

volcanoes; however, there is a difference in its struc-

ture. The difference lies in the horizontal extent of the

conductor beneath the summit crater. In the case of

active volcanoes, there are examples in which theshallow conductor is not limited to only the region

beneath the main summit, but extends to the entire

region beneath the volcano [3136]. A possibility that

explains this feature of the resistivity structure of Mt.

Fuji is the permeability that was discussed previously.

It is speculated that the hydrothermal system cannot

develop and extend to the whole volume of Mt. Fuji

due to the rather low permeability in the volcano. The

quiescence of volcanic activities, such as low seismic-

ity, no gas emanations, and no natural hot springs,

Fig. 9. Schematic model of the hydrothermal system of Mt. Fuji.

The thick arrows show the movement of hot fluids. The thin arrows

show the movement of relatively cool fluids. The good conductor

imaged by the magnetotelluric sounding was interpreted as a hot-

water saturated zone. The circles with plus and minus signs repre-

sent positive and negative conduction current sources, respectively.

The positive conduction current sources are accumulated at the top

of the hydrothermal upwelling.



11/13

might be related to the structure of and permeability in

the volcano.

The SP distribution (Fig. 1) suggests certain fea-

tures of the hydrothermal system. A remarkable SP positive anomaly exists only around the summit

crater and not around the Hoei crater. The Hoei

eruption in 1707 was the latest eruption, whereas

the last eruption on the summit occurred more than

2200 yrs ago [8]. Since the sounding curve obtained

in the Hoei crater does not show any significant

difference from those of the other sites at the same

elevation, the good conductor might not exist direct-

ly beneath the Hoei crater (Fig. 10). Therefore, the

hydrothermal system does not seem to be active

directly beneath the Hoei crater at present; this sug-gests that the growth process of the hydrothermal

system differs between the summit and the flank.

The heat that drives the hydrothermal system might

have been supplied only beneath the summit crater

for a long duration.

In Mt. Fuji, it is found that the good conductor is

located beneath the positive SP anomaly. The goodconductor has a lateral extent of approximately 4 km

and extends downward from 1 km below the surface.

The positive SP anomaly has a diameter of approxi-

mately 3 km. Based on the SP modeling using a pair

of positive and negative conduction current sources in

a vertical line imposed on the axisymmetrical resis-

tivity structure, the positive source was determined to

be located at the top of the conductor. These results

suggested the presence of a single-phase (liquid) hy-

drothermal system even in the apparently dormant

volcano. However, there are some volcanoes wherea dominant positive SP anomaly is not observed

around the peak; instead, only a strong btopographic

effectQ is observed [24,37]. The knowledge regarding

Fig. 10. Comparison between the sounding curves that are obtained at the Hoei crater and other sites at the same elevation. Site names and

elevations are shown above each panel. Regional strikes are assumed to be N45W and un-decomposed data are shown. The impedance skew of

H0 (at the Hoei crater) is approximately 0.2 for all frequencies. Note that the good conductor is not seen directly beneath the sites S1 and M2.

The curves for H0 are not significantly different; therefore, probably, a good conductor does not exist directly beneath the Hoei crater.



12/13

the extent to which good conductors exist within

apparently dormant volcanoes is sparse. In order to

better understand the SP generating mechanism and

hydrothermal systems in volcanoes, it would be important to study the relationship between SP anoma-

lies and subsurface conductors by conducting surveys

on both active and quiescent volcanoes.

Acknowledgments

We thank T. Kagiyama and N. Osada for providing

support for the arrangement of our field survey. We

also thank the Fujinomiya-Sengen Shrine and Yama-

nashi and Shizuoka Prefectures for permitting the fieldobservations. T. Mogi allowed us to use the MT data

sets of his own survey in Hokkaido for remote-refer-

ence processing in this study. Critical and useful

comments from Dr. Dominique Gilbert and an anon-

ymous reviewer greatly improved the manuscript. The

field operation was partly supported by the new Pro-

gram of the Study and Observation for Earthquake

Prediction. This research was partly supported by

special coordination funds for promoting science

and technology.

References

[1] T. Kagiyama, H. Utada, T. Yamamoto, Magma ascent beneath

Unzen Volcano, SW Japan, deduced from the electrical resis-

tivity structure, J. Volcanol. Geotherm. Res. 89 (1999) 3542.

[2] E. Fujita, Y. Ida, J. Oikawa, Eigen oscillation of a fluid sphere

and source mechanism of harmonic volcanic tremor, J. Volca-

nol. Geotherm. Res. 69 (1995) 365378.

[3] T. Hashimoto, Y. Tanaka, A large self-potential anomaly on

Unzen volcano, Shimabara peninsula, Kyusyu island, Japan,

Geophys. Res. Lett. 22 (1995) 191194.

[4] A.W. Hurst, H.M. Bibby, B.J. Scott, M.J. McGuiness, The

heat source of Ruapehu Crater Lake; deductions from the

energy and mass balances, J. Volcanol. Geotherm. Res. 46

(1990) 120.

[5] G. Chiodini, F. Frondini, C. Cardellini, D. Granieri, L. Marini,

G. Ventura, CO2 degassing and energy release at Solfatara

volcano, Campi Flegrei, Italy, J. Geophys. Res. 106 (2001)

1621316221.

[6] H. Kumagai, M. Nakano, B.A. Chouet, Temporal evolution of

a hydrothermal system in Kusatsu-Shirane Volcano, Japan,

inferred from the complex frequencies of long-period events,

J. Geophys. Res. 107 (B10) (2002) 2236, doi:10.1029/

2001JB000653.

[7] E.W. Wolfe, The 1991 eruptions of Mount Pinatubo, Philip-

pines, Earthq. Volcanoes 23 (1) (1992).

[8] H. Miyaji, History of younger volcano, J. Geol. Soc. Japan 94

(1998) 433452 (in Japanese).

[9] H. Tsuya, Explanatory of Text Geologic Map 1 : 50,000 Scale,

vol. 24, Geological Survey of Japan, 1968.

[10] K. Aizawa, A large self-potential anomaly and its changes on

the quiet Mt. Fuji, Japan, Geophys. Res. Lett. 31 (5) (2004)

L05612, doi:10.1029/2004GL019462.

[11] J. Zlotnicki, Y. Nishida, Review on morphological insights of

self-potential anomalies on volcanoes, Surv. Geophys. 24

(2003) 291 338.

[12] T.D. Gamble, W.M. Goubau, J. Clarke, Magnetotellurics

with a remote magnetic reference, Geophysics 44 (1993)

5368.

[13] C.M. Swift Jr., A magnetotelluric investigation of an electri-

cal conductivity anomaly in the southwestern United States,

in: K. Vozoff (Ed.), Magnetotelluric Methods, Geophys. Re- print Series, vol. 5, Soc. of Explor. Geophys, Tusla, OK,

1986, pp. 156 166 (extracted from a thesis submitted to

Dep. Geol. Geophys. M. I. T., 1967).

[14] R.W. Groom, R.C. Bailey, Decomposition of magnetotellurics

impedance tensors in the presence of local three-dimensional

galvanic distortions, J. Geophys. Res. 94 (1989) 1913 1925.

[15] Y. Ogawa, T. Uchida, A two-dimensional magnetotelluric

inversion assuming Gaussian static shift, Geophys. J. Int.

126 (1996) 6976.

[16] P.E. Wannamaker, S.H. Ward, G.W. Hohmann, Magnetotellu-

ric responses of three dimensional bodies in layered Earth,

Geophysics 49 (1984) 15171533.

[17] J. Ledo, P. Queralt, A. Marti, A. Jones, Two-dimensional

interpretation of three-dimensional magnetotelluric data: andexample of limitations and resolution, Geophys. J. Int. 150

(2002) 127 139.

[18] S. Nakada, M. Yoshimoto, T. Kaneko, T. Shimano, T. Fujii,

Outline of scientific drilling at Fuji Volcano, Chikyu Monthly

S48 (2004) 8288 (in Japanese).

[19] S. Yamamoto, Fujisan, in: A. Suwa (Ed.), Dobun-shoin Co,

Tokyo, 1992, pp. 197217 (in Japanese).

[20] Y. Sasai, J. Zlotnicki, Y. Nishida, P. Yvetot, P. Morat, H.

Murakami, Y. Tanaka, T. Ishikawa, S. Koyama, W. Sekiguchi,

Electromagnetic monitoring of Miyake-jima Volcano, Izu-

Bonin Arc, Japan: a preliminary report, J. Geomagn. Geo-

electr. 49 (1997) 12931316.

[21] J. Zlotnicki, Y. Sasai, P. Yvetot, Y. Nishida, M. Uyeshima, F.

Fauquet, H. Utada, H. Takahashi, G. Donnadieu, Resistivity

and self-potential changes associated with volcanic activity:

the July 8, 2000 Miyake-jima eruption (Japan), Earth Planet.

Sci. Lett. 205 (2003) 139154.

[22] B. Nourbehecht, Irreversible thermodynamic effects in inho-

mogeneous media and their applications in certain geoelectric

problems, Ph.D. thesis, M.I.T., Massachusetts, 1963.

[23] M.J.S. Johnston, J.D. Byerlee, D. Lockner, Rapid fluid dis-

ruption: a source for self-potential anomalies on volcanoes, J.

Geophys. Res. 106 (2001) 43274335.

[24] R.F. Corwin, D.B. Hoover, The self-potential method in geo-

thermal exploration, Geophysics 44 (1979) 226245.

http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


13/13

[25] W.R. Sill, Self-potential modeling from primary flows, Geo-

physics 48 (1983) 7689.

[26] T. Ishido, H. Mizutani, Experimental and theoretical basis of

electrokinetic phenomena in rockwater systems and its appli-

cation to geophysics, J. Geophys. Res. 86 (1981) 17631775.

[27] A. Dey, H.F. Morrison, Resistivity modeling for arbitrarily

shaped three dimensional structures, Geophysics 44 (1979)

753780.

[28] K. Kanda, S. Mori, Self-potential anomaly of Satsuma-Iwojima

volcano, Earth Planets Space 54 (2002) 231237.

[29] T. Tosha, N. Matsushima, T. Ishido, Zeta potential measured

for an intact granite sample at temperatures to 200 8C, Geo-

phys. Res. Lett. (2003), doi:10.1029/2002GL016608.

[30] T. Ishido, J.W. Pritchett, Numerical simulation of electrokinetic

potentials associated with subsurface fluid flow, J. Geophys.

Res. 104 (1999) 1524715259.

[31] G.R. Olhoeft, Electrical properties of granite with implications

for the lower crust, J. Geophys. Res. 86 (1981) 931 936.[32] D.B. Fitterman, W.D. Stanley, R.J. Bisdorf, Electrical structure

of Newberry volcano, Oregon, J. Geophys. Res. 93 (1988)

1011910134.

[33] H. Utada, H. Shimomura, Resistivity structure of Izu-

Oshima Volcano revealed by the ELF-VLF magnetotelluric

method, J. Geomagn. Geoelectr. 42 (1990) 169194.

[34] Y. Ogawa, S. Takakura, CSAMT measurements across the

1986 C Craters of Izu-Oshima Island, Japan, J. Geomagn.

Geoelectr. 42 (1990) 211 224.

[35] Y. Ogawa, N. Matsushima, H. Oshima, S. Takakura, M.

Utsugi, K. Hirano, M. Igarashi, T. Doi, A resistivity cross-

section of Usu volcano, Hokkaido, Japan, by audiomagneto-

tellurics soundings, Earth Planets Space 50 (1998) 339349.

[36] N. Matsushima, H. Oshima, Y. Ogawa, S. Takakura, H. Satoh,

M. Utsugi, Y. Nishida, Magma prospecting in Usu volcano,

Hokkaido, Japan, using magnetotelluric soundings, J. Volca-

nol. Geotherm. Res. 109 (2001) 263277.

[37] J. Zlotnicki, G. Boudan, J.P. Viode, J.F. Delarue, A. Mille, F.

Bruere, Hydrothermal circulation beneath Mount Pelee in-

ferred by self potential surveying. Structural and tectonic

implications, J. Volcanol. Geotherm. Res. 84 (1998) 7391.

http://-/?-http://-/?-http://-/?-

Documents

MT Study of Mt Fugi Hydro Thermal System AizawaEPSL05