DSpace at VNU: Water geochemistry and soil gas survey at Ungaran geothermalfield, central Java, Indonesia

DSpace at VNU: Water geochemistry and soil gas survey at Ungaran geothermalfield, central Java, Indonesia tài liệu, giáo...

Water geochemistry and soil gas survey at Ungaran geothermal field,

central Java, Indonesia

Nguyen Kim Phuong1,2,⁎ , Agung Harijoko3, Ryuichi Itoi1, Yamashiro Unoki1

1

Department of Earth Resource Engineering, Faculty of Engineering, Kyushu University, 744 Motooka, Nishiku, Fukuoka 819-0395, Japan

2 Department of GeoEnvironment, Faculty of Geology and Petroleum Engineering, Ho Chi Minh City University of Technology, 268 Ly Thuong Kiet street, Ward 14, District 10,

Ho Chi Minh City, Viet Nam

3

Department of Geological Engineering, Faculty of Engineering, Gadjah Mada University, Jl Grafika 2, Yogyakarta (55281), Indonesia

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 7 September 2011

Accepted 3 April 2012

Available online 11 April 2012

Keywords:

Soil gas

Water geochemistry

Ungaran geothermal field

Indonesia

A soil gas survey for radon (Rn), thoron (Tn), CO2, and mercury (Hg), and the chemical analysis of hot spring waters, were undertaken in the Ungaran geothermalfield, Central Java, Indonesia The results of soil gas surveys indicate fault systems trending NNE–SSW and WNW–ESE Particularly high CO2concentrations (>20%), and high

Hg concentrations were detected in vicinity of the fumaroles Emanometries of Rn, Tn and CO2also conclusively identified the presence of a fracture zone for the migration of geothermal fluid The Hg results infer that the up-flow zone of high temperature geothermal fluids maybe located in the north of fumaroles in the Gedongsongo area (near the collapse wall) Chemistry of thermal springs in the up-flow zone are acid (pH=4) and show a Ca–Mg–SO4composition The thermal waters are mainly Ca–Mg–HCO3and Ca–(Na)–SO4–HCO3types near the fumarolic area and are mixed Na–(Ca)–Cl–(HCO3) waters in the south east of Gedongsongo Theδ18O (between

−5.3 and −8.2‰) and δ (between −39 and −52‰) indicate that the waters are essentially meteoric in origin A conceptual hydro-geochemical model of the Gedongsongo thermal waters based on the soil gas, isotope and chemical analytical results, was constructed

© 2012 Elsevier B.V All rights reserved

1 Introduction

Geothermal exploration began in Indonesia in 1970 More than

200 geothermal systems with significant active surface

manifesta-tions occur throughout Indonesia Most of the geothermal systems

in Indonesia that have surface manifestations withfluid discharges

at boiling temperature occur in areas with Quaternary volcanism

and active volcanoes, along well-defined volcanic arcs (Sudarman

et al., 2000; Hochstein and Sudarman, 2008) During the 1980s the

Pertamina Company explored a number of areas characterized by

significant fumarolic activity on the flanks of inactive Holocene

volca-noes At Ungaran, Pertamina Co conducted geophysical surveys using

several geophysical methods (resistivity, magnetotellurics, and

aero-magnetic surveys) between 1985 and 1990 at Ungaran (Budiardjo

et al., 1989)

The Ungaran volcano is located in the Central Java province about

30 km southwest of Semarang, Indonesia (Fig 1), and is an undeveloped

geothermal prospect There are two active fumaroles on the southern

flank of the dormant Ungaran volcano, about 2 km SW from its summit The Gedongsongo area, on the southernflank of Ungaran volcano, is characterized by the presence of thermal manifestations such as fuma-roles, hot springs, acidic mud pools and hydrothermal alteration zones Four wells were drilled down to a depth of 500 m around the Gedongsongo area These wells showed slightly anomalous tempera-ture (about 47 °C at 300 m depth) and temperatempera-ture-gradient values near the bottom (Hochstein and Sudarman, 2008) A low resistivity anomaly, about 30Ωm, associated with a deep geothermal reservoir was inferred to occur beneath the Ungaran summit (Budiardjo et al.,

1989) Further exploration was not conducted because about 90% of the prospect area is located in a protected forest with restricted access The soil gas method, i.e the measurement of gas concentrations in the soil, has been used to estimate the location of heat sources and their areal extension in geothermal prospecting (Varekamp and Buseck, 1984, 1986; Bertrami et al., 1990; Finlayson, 1992; Hernández

et al., 2000) The gases (Rn, Tn, CO2, and Hg) are assumed to be released from active geothermal systems at depth and then ascend through overlying rock formations and/or fracture zones The high mobility of these gases makes them ideal pathfinders for concealed natural re-sources, as the gases produced and accumulated in geothermal reser-voirs can escape to the surface by migration along fractures and faults (Koga, 1982; Fridman, 1990) Soil gas surveys (Rn, Tn, CO2and Hg) have regularly been used for the exploration of geothermal reservoirs (Varekamp and Buseck, 1984, 1986; Corazza et al., 1993), and for

⁎ Corresponding author at: Department of Earth Resources Engineering, Faculty of

Engineering, Kyushu University, 744, Motooka, Nishiku, Fukuoka 819-0395, Japan.

Tel./fax: +81 92 802 3345.

E-mail addresses: nkphuong6280@yahoo.com , nkphuong@mine.kyushu-u.ac.jp

(N.K Phuong).

0377-0273/$ – see front matter © 2012 Elsevier B.V All rights reserved.

Contents lists available atSciVerse ScienceDirect

Journal of Volcanology and Geothermal Research

j o u r n a l h o m e p a g e : w w w e l s e v i e r c o m / l o c a t e / j v o l g e o r e s

detecting and delineating faults and fractures (Gregory and Durrance,

1985; Toutain and Baubron, 1999; Guerra and Lombardi, 2001; Walia

et al., 2005, 2008; Yang et al., 2005)

The enrichment of radon (Rn), carbon dioxide (CO2) and mercury

(Hg) in soil or soil gas has been observed in several geothermal areas

(e.g.,Koga, 1982; Varekamp and Buseck, 1983; Chuaviroj et al., 1987;

Lescinsky et al., 1987; Klusman, 1993; Murray, 1997) Radon is a

ra-dioactive noble gas that is soluble in water and decays by alpha

emis-sion The presence of Rn in geothermal areas is a function of the

porosity and fracture distribution of the rocks in between the deep

geothermal source and the surface, i.e., the pathway for uprising

fluids (Koga, 1988) Radon gas surveys are widely used to monitor

seismic activity and to detect the locations of fractures and faults

(Walia et al., 2005; Yang et al., 2005)

Soil and soil gas surveys of Hg have been successfully used as

geo-thermal exploration techniques.Van Kooten (1987),Lescinsky et al

(1987)and Murray (1997)all found broad Hg anomalies outlining

high-temperature thermal activity zones Soil Hg surveys has also

been used to locate faults in volcanic and geothermal regions

Buseck, 1983) In the sub-surface, Hg is strongly partitioned into the

ascending vapor and is transported to the surface as elemental Hg

This vapor is absorbed onto organic matter and clay minerals in the

shallow, low-temperature soil horizons, producing elevated (above

10 ppm) concentrations of Hg (Nicholson, 1993) Mercury is

absorbed by the soil in anomalous concentrations relative to the

sur-rounding areas (Lescinsky et al., 1987; Van Kooten, 1987) Mercury

levels in soil are the result of accumulation and loss processes; conse-quently, soil gas mercury is a reliable indicator of geothermalfluid at depth (Koga, 1988)

The present study aimed to i) delineate the up-flow zone of high temperature geothermalfluids at depth, fractures below the surface,

in the Ungaran geothermalfield using a soil gas survey; ii) character-ize chemical characteristics of hot spring waters in the Ungaran geo-thermalfield; and iii) develop a hydro-geochemical conceptual model

of thermal water in the Ungaran geothermalfield

2 Geological setting Geothermal areas in Central Java, including the Ungaran volcano, are located in the Quaternary Volcanic Belt (Solo Zone) (Fig 1) This belt is located between the North Serayu Mountains and the Kendeng Zone and contains numerous Quaternary eruptive centers, including Dieng, Sindoro, Sumbing, Ungaran, Soropati, Telomoyo, Merapi, Muria, and Lawu (Van Bemmelen, 1970; Thanden et al., 1996) (Fig 1) A structural analysis of this area revealed that the Ungaran volcanic system is primarily controlled by the Ungaran collapse struc-ture that runs from west to southeast of the Ungaran volcano Old vol-canic rocks of the pre-caldera formation are controlled by northwest– southwest and southeast–southwest fault systems The post-caldera volcanic rocks, however, do not seem to be structurally controlled

by the regional faulting system (Budiardjo et al., 1997) The pre-caldera volcanic rocks and the Tertiary marine sedimentary rocks

Fig 1 Location of the Ungaran volcano (solid red circle).

are inferred to be the main geothermal reservoir rocks (Budiardjo et

al., 1997)

Van Bemmelen (1970)noted that there is a gradual development of

volcanism along the transverse fault from north to south, starting in the

north with the Oldest or Proto-Ungaran in the Lower Pleistocene and

ending in the south with the very active Merapi volcano (Fig 1) Two

generation of the Ungaran volcano (2050 m) were observed because

of gravitational collapse The Oldest Ungaran deposits resulted from

submarine activity Its basement is transitional beds, in which the facies

changes from marine into fresh water deposits consisting of coarse

polymictic conglomerates of the Lower Damar Beds After magma

broke through the crust, the Oldest Ungaran volcano originated at the

eastern end of the crest The coarse volcanic breccias of the Middle

Damar Beds, and the coarse conglomerates, tuff-sandstones and

black-clay of the Upper Damar Beds occur at the northern foot of the Oldest

Ungaran volcano In the Upper Pleistocene, volcanic activity was

wide-spread In the eastern part of the northern Serayu Range, volcanic

activity built up the Old Ungaran volcano, which is the second

genera-tion of Ungaran volcano The breccias at its northern foot form the

Notopuro Beds, which cover the breccias of the Oldest Ungaran in the

Damar Beds with an angular unconformity After the early Pleistocene

phase of volcanic growth, volcanic activity continued until the Holocene,

building up the Young Ungaran volcano, which consists of pyroclastic

flow deposits, pyroclastic lava and alluvial deposits

The Ungaran geothermal system is associated with the Upper

Quaternary volcanism of the Ungaran volcano The volcanic rocks

are rich in alkali metals and are classified as trachyandesite to

trachy-basaltic andesite, primarily containing plagioclase, sanidine and

cris-tobalite (Budiardjo et al., 1997; Kohno et al., 2005) Gedongsongo is

the main geothermal area on the southernflank of the Ungaran

volca-no (Fig 2) The Gedongsongo area is characterized by the presence of

fumaroles (90–110 °C), neutral pH bicarbonate warm/hot springs and

diluted steam heated hot spring (22–80 °C) with underground

tem-peratures of 20 °C to 82 °C measured from 1 m depth According to

Budiardjo et al (1997), the composition of thermal spring waters at

Gedongsongo can be divided into two water types The hot water

around the fumarolic area originates as a steam heated meteoric

water characterized by low chloride content (similar to local surface

water), high sulfate content (up to 1000 ppm), and low pH (up to

5) while neutral bicarbonate or chloride waters are located at

the other areas Based on the analysis of soil and rocks samples

col-lected around the Ungaran volcano,Kohno et al (2005)concluded

that quartz, halloysite and alunite are the main secondary minerals found in the hydrothermal alteration zones Quartz is formed by the alteration of cristobalite from Ungaran rocks, while halloysite and alunite are minerals formed by alteration s due to acidic and low tem-perature hydrothermal waters

3 Sampling and measurement The study area is focused on the Gedongsongo area (Fig 2), which

is the main geothermal prospect in the Ungaran area, located in the southern part of the Ungaran volcano Water samples (UW-1 to UW-7A and UW-7B) were collected around the Gedongsongo area where comprises a volcanic complex terrain at an altitude ranging from 1200 to 2000 m above sea level Others samples (8A, UW-8B and UW9), however, were collected at Kendalisodo area (approxi-mately 8 km far from the Gedongsongo area) with altitude at 600 m above sea level

3.1 Chemical analysis of water Water sampling was complemented by in situ measurement of pH, temperature and conductivity The water samples werefiltered through 0.45μm membrane filters prior to storage in sterile polyethylene bottles (HDPE) Samples for cation (Li, NH4, Na, K, Mg and Ca) and silica (SiO2) analyses were collected in plastic bottles that had been acidified with

1 mL of concentrated HCl Filtered, un-acidified samples were collected for anions (F, Cl, HCO3, SO4) analysis All water analyses were conducted

at Kyushu University using standard methods Cations and anions were analyzed using ion chromatography (Dionex ICS-90) while boron (B) was analyzed using ICP-AES (Vista-MPX) SiO2contents was deter-mined by colorimetry and analyzed using a digital spectrophotometer (Hitachi U-1100) (APHA, 2005), while HCO3was analyzed by titration with 0.1 M HCl The analytical error for techniques was≤5%

An aliquot of the water samples (20 mL) was collected and stored

in sterile polyethylene bottles (HDPE) for stable isotope analysis Water isotopes (δ18O and δD) were determined using the CO2–H2 equilibration method (Epstein and Mayeda, 1953) Then, the isotope ratios were measured using the DELTA Plus mass-spectrometer at Fukuoka University, Japan These internal standards were calibrated using international reference materials V-SMOW and SLAP with analytical precisions of ±0.1‰ for δ18O and ±1‰ for δD

3.2 Soil gas measurement

Soil gas surveys for Rn, Tn, CO2, mercury in soil gas (Hgsoil-gas) and

in soil samples (Hgsoil) were conducted in an area approximately

1.3 km north to south by 1.5 km west to east (Fig 3) The distance

be-tween measurement points varied from 50 to 150 m Soil gases were

collected from a depth of 60 cm using a steel pipe (5 cm in diameter)

inserted into the ground

3.2.1 Radon measurement

The Rn and Tn concentrations were measured with a radon

detec-tor (RD-200, EDA Instruments Co Ltd.) The soil gas was circulated

through the detector with an electrical pump for 10 s, replacing the

air in the detection cell The Rn concentration was measured by an

α-scintillation radon counter with the soil gas pumped directly into

a scintillation chamber When the α-particles produced during

radon decay impact the ZnS(Ag) layer in the scintillation counter,

an energy pulse is created in the form of photons, measured by a

photo-multiplier and a counter As both Rn and Tn decay by means

of α-emission, the concentrations of Rn and Tn were calculated

from three counts in each minute obtained for three sequential

minutes

3.2.2 CO2measurement

To measure the CO2concentration, 100 mL of soil gas was sampled

from the stainless steel probe inserted into the ground using a

stain-less steel syringe, and the CO2concentration was measured using an

SA-type gas detector tube (Komyo-Kitagawa Instruments Co Ltd.)

This gas detector works on the principles of chemical reaction and

physical absorption and has ±1% analytical precision As the gas is

entered into the detector tube, a constant color is produced, which

varies in the length of discolored layer due to the reaction between

the reagent and the CO The CO concentration can then be obtained

directly by reading from the measuring scale on the tube or using a concentration chart

3.2.3 Mercury in soil gas (Hgsoil-gas) and in soil (Hgsoil) The Hg concentration was measured by the gold wire method, which indicates both the Hg in soil gas (Hgsoil-gas) in the hole, and the concentration in the ascending gas The Hgsoilmeasurement rep-resents the concentration of Hg absorbed onto the surface of soil par-ticle To measure Hg in soil gas, a pure gold wire (10 cm long, 1 mm diameter, 1.5 g weight and 3.16 cm2 in the effective surface) was left in the hole for 7 days after completing the CO2 measurement (Koga, 1982, 1988) After a week, the gold wire was removed from the hole and stored in a tightly sealed glass tube

Soil samples were collected at 0.6 m depth in the hole and sealed

in plastic bags The soil samples were then air-dried at room temper-ature for two weeks and ground with a mortar and a pestle after re-moving rock fragments and plant roots

The mercury concentrations were determined in the laboratory

by the cold vapor atomic absorption method using mercury analyzer SP-3 (Nippon Instruments Co., Japan) This equipment uses the heating-vaporization (700 °C) technique to liberate mercury present

in the sample

4 Results and discussions 4.1 Water chemistry and stable isotope compositions The results of the water analyses are given inTable 1, and show that the water temperatures ranged from 18 °C (UW 6) to 56 °C (UW 3), while pH values were in the range 3.45–7.87 The SiO2 con-tents of the thermal waters ranged from 47 to 219 mg/L, while the

EC values were generally between 36 and 561μS/cm, except for rela-tively high values in UW 8A and 8B (up to 5300μS/cm) Other major

Fig 3 Location of soil gas and water samples around the fumarolic area (UTM coordination system).

elements range from 2 to 746 mg/L for Na, from 3.54 to 278 mg/L for

Ca while concentrations of Mg are lower (b126 mg/L) The waters

contain relatively low K concentrations (1.18–47.11 mg/L) For anions,

the HCO3concentration is relatively high (39–1824 mg/L) followed by

SO4(b246 mg/L) Chloride concentration is rather low except UW-8A and 8B are relatively high (about 1000 mg/L)

The chemical compositions of the water samples are plotted on the Cl–SO4–HCO3 (Giggenbach, 1988) and Na–SO4–Mg diagrams shown inFig 4 The UW 1 and UW 2 samples are classified as acid-sulfate waters, with high concentrations of SO4 (247 mg/L and

136 mg/L, respectively), but low concentrations of F (b0.25 mg/L) and Cl (below 1.5 mg/L) (Table 1) suggests that UW 1 and UW 2 have been steam heated, absorbing a gas phase enriched in S-bearing compounds The SO4 enrichment can be explained by the

O2-driven oxidation of H2S to H2SO4 in oxygenated near surface groundwater (Henley and Stewart, 1983; Tassi et al., 2010; Joseph

et al., 2011) Differences from the above samples, most samples are HCO3-dominated water, mostly Ca–HCO3or Ca–Mg–HCO3type (UW

3, 4, 5, 7A, 7B and 9) while UW 8A and 8B are of the Na–HCO3–Cl or

Na–Ca–Cl–HCO3type with much higher Na, Ca, HCO3,Cl and B con-centrations than the other samples (Table 1) Water–rock interaction should be a source of sodium and chloride in UW 8A and 8B The minerals in the volcanic rocks primarily consist of plagioclase, sanidine, and cristobalite with some biotite and hornblende (Kohno

et al., 2005).Table 2shows the molar ratios of some of the major components of thermal waters in the study area Increases in the Na/Cl and K/Cl ratios in thermal waters are likely to reflect reactions with feldspar or clay minerals These ratios can therefore be used as

an independent indicator of residence time Thermal waters often fol-low a longer, deeper, regionalflow path than non-thermal waters, and thus have much higher Na/Cl and K/Cl ratios than non-thermal waters (Han et al., 2010) The Na/K ratio is controlled by temperature dependent mineral–fluid equilibria (Koga, 1988; Gemini and Tarcan,

2002) The ratios of Na/K are large for all water samples, indicating

Table 1

Chemical composition (in mg/L) and δ 18 O and δD values of water samples in the Ungaran geothermal field.

ID Temp pH EC HCO 3 − F − Cl − SO 4 − SiO 2 Li + Na + NH 4+ K + Mg 2

+

Ca 2 + B δD δ 18 O

UW-1 21.9 3.45 561 50 0.12 1.8 247 58 1.6 34.3 0.43 15.0 13.4 42.5 1.22 −47 −7.9 UW-2 40.0 5.36 368 59 0.21 1.2 136 109 1.1 25.3 0.64 8.6 10.3 32.6 0.79 −49 −7.9 UW-3 56.0 6.10 333 200 0.13 0.8 31.8 86 0.3 14.1 0.45 7.9 15.1 37.1 0.61 −50 −8.0 UW-4 32.2 6.00 297 465 0.05 0.8 2.6 82 0.1 10.7 0.49 5.5 14.7 35.9 0.65 −51 −8.2 UW-5 a

n.a 6.31 36 100 0.02 0.7 3.5 23 0.1 2.3 0.02 1.2 0.7 3.5 0.36 −51 −8.2 UW-6 18.0 5.42 177 107 0.01 0.7 50.3 51 0.2 6.8 0.04 3.1 5.6 18.2 0.42 −50 −8.0 UW-7A 50.0 6.10 491 501 0.06 0.8 3.4 93 0.5 11.8 0.65 6.4 19.8 38.9 0.58 −51 −8.1 UW-7B 25.0 5.90 164 496 0.05 0.8 2.9 89 0.4 12.3 0.53 6.0 17.6 40.7 0.52 −52 −8.2 UW-8A 35.2 6.84 4580 1732 0.06 998 0.2 92 4.4 700 16.1 44.2 117.7 217.3 15.9 −39 −5.3 UW-8B 38.1 6.78 5210 1824 0.06 1088 0.1 95 5.6 746 18.0 47.1 126.0 278.4 19.7 −40 −5.3 UW-9 a 23.8 7.87 513 351 0.07 7.2 4.4 51 0.5 23.2 0.29 2.4 26.9 62.1 0.15 −39 −6.1 n.a.: not analyzed.

a

River water.

Fig 4 Chemical compositions (a) Cl–SO 4 –HCO 3 and (b) SO 4 –Mg–Na of thermal waters

field.

Table 2 Molar ratios of some major components of water samples in the Ungaran geothermal field.

ID Temp Na/Cl K/Cl Ca/Cl Na/K Na/Ca HCO 3 /SO 4 Cl/B (°C)

UW-1 21.9 28.68 7.42 20.4 15.8 0.32 2.17 0.46 UW-2 40.0 33.82 6.80 25.0 29.5 0.68 2.25 0.44 UW-3 56.0 28.36 9.37 43.0 152 9.91 2.61 0.38 UW-4 32.2 21.90 6.57 42.1 358 280 2.83 0.35 UW-5 a

– 5.40 1.63 4.7 87.9 44.58 1.79 0.56 UW-6 18.0 16.06 4.39 24.8 96.0 3.36 2.12 0.47 UW-7A 50.0 21.78 6.94 41.4 349 231.9 2.28 0.44 UW-7B 25.0 23.11 6.62 44.1 353 274 2.09 0.48 UW-8A 35.2 1.08 0.04 0.2 26.9 5.6 13,629 19.1 UW-8B 38.1 1.04 0.04 0.2 26.8 4.66 28,706 16.8 UW-9 a 23.8 4.96 0.3 7.6 16.7 0.65 124.4 14.6 n.a.: not analyzed.

a

that the temperature of geothermal reservoir is not probably too high.

This is in agreement with general increase in Na/K ratios of thermal

water with decreasing reservoir temperature (Ellis and Mahon,

1967; Koga, 1988; Cortecci et al., 2005) Based on relatively low

Na/K ratios (b15,Table 2) water of springs at the Gedongsongo area

that have reached the surface rapidly and are therefore associated

with up-flow structures or permeable zones while higher Na/K ratios

(>15,Table 2) are indicative of lateralflows which may undergone

near-surface reactions and conductive cooling (Nicholson, 1993;

Cortecci et al., 2005; Di Napoli et al., 2009) Similarly, high Na/Ca

ra-tios are also indicative of direct feeding from a geothermal reservoir

and less groundwater contribution, while the HCO3/SO4ratio can be

used as an indicator offlow direction (Table 2) The Na/Ca values

for deep well thermal waters are very high (>50), while for cold

groundwater this ratio is around 0.25 Low Na/K and Na/Ca ratios

are found in thermal waters in the north of the fumarole (UW 1 and

UW 2) while to the south of the fumarole, the thermal waters have

high Na/K and Na/Ca ratios and increasing HCO3/SO4ratios (Table 2)

Therefore, we can infer that thermal waters in the north of the fumarole

are associated with up-flow zones, while thermal waters to the south of

the fumarole are associated with lateralflow

The behavior of conservative components useful in the delineation

of formation processes of waters, involving Cl, Li and B, is investigated

inFig 5 As pointed out above, Li is the alkali element least affected by secondary absorption processes Li is also released during water–rock interactions and remains largely in solution (Giggenbach and Soto, 1992; Mainza, 2006; Tassi et al., 2010) Boron and Cl−are not readily incorporated into secondary, alteration minerals, so they can be considered conservative chemical species (Seyfried et al., 1984; Nicholson, 1993; Tassi et al., 2010) Boron may have several origins

It may be leached from sedimentary rocks; due to its volatility in high temperature steam, it may also be introduced with any high temperature vapor phase absorbed into water Moreover, the B con-tent of thermalfluids is likely during the early heating up stages Therefore,fluids from older hydrothermal systems can be expected

to be depleted in B while the converse holds for younger hydrother-mal systems (Mainza, 2006) It is, however, striking both Cl and B is adding to the Li containing solutions in proportions close to those in crustal rocks For the UW 8A and 8B, it can be elucidated that dissolu-tion of an averaged andesitic–rhyolitic rock, followed by exchange with secondary minerals or interaction with gases (Fig 5) Moreover, water rock interaction can be postulated by Cl/B ratio Ellis and Mahon (1967)found that in areas where andesitic or rhyolitic rocks predominate, Cl/B ratios are often between 10 and 30 The Cl/B ratios

Fig 5 Binary diagram of Li vs Cl and B vs Cl.

Table 3 Estimated temperature (in °C) for thermal water in the Ungaran geothermal filed using silica geothermometers.

ID Measured

temperature

Estimated temperature (°C)

(°C)

T Qza T Qzb T Cc T d

UW 1 21.9 109 109 80 110

UW 2 40.0 142 137 116 143

UW 3 56.0 129 126 101 129

UW 4 32.2 127 124 99 127

UW 6 18.0 102 103 73 103

UW 7A 50.0 133 129 106 133

UW 7B 25.0 131 127 103 131

UW 8A 35.2 132 129 105 132

UW 8B 38.1 134 130 107 134

UW 9* 23.8 102 103 72 103

a

Quartz — no steam loss from Fournier (1983)

b Quartz — maximum steam loss at 100 °C from Fournier (1983)

c Chalcedony from Fournier (1983)

d

From Fournier and Potter (1982)

Fig 7 Na–K–Mg ternary diagram for thermal waters in the Ungaran geothermal field

of the UW 8A and 8B are from 14 to 19 (Table 2), thus, it is suggested

that high concentration of Na, Cl, Li and B is originated from water–

rock interaction

The results of the stable isotope analysis in the Gedongsongo area

are given inTable 1and plotted in theδD vs δ18O diagram (Fig 6)

Stable isotope compositions of meteoric water from coastal Jakarta

(Gat and Gonfiantini, 1981) were used as reference data (δD=8.05

δ18O + 16.48) (Fig 6) This line does not deviate significantly from

the global meteoric water line defined byCraig (1961) InFig 6, all

samples from Gedongsongo plot along the meteoric water line,

sug-gesting that the thermal waters are of meteoric origin Compared to

the others samples, increase inδD values of UW 8A, 8B and UW 9 are results of altitude affect The UW 8A, B and 9 were located at area whose altitude (about 600 m a.s.l.) is relatively lower than the Gedongsongo area Moreover, the thermal waters from UW 8A and 8B show positively shift ofδ18O which caused by a reaction with rock The estimated reservoir temperatures for the Ungaran geother-malfield using silica geothermometers (Fournier and Potter, 1982; Fournier, 1983) are listed inTable 3 Chalcedony geothermometers indicate lower temperatures (72 °C–116 °C) than quartz geother-mometers (102 °C–142 °C) The Na–K–Mg1/2 triangle proposed by

Giggenbach (1988) is shown in Fig 7 All of the thermal waters

Table 4

Soil gas concentrations of Rn, Tn, CO 2 and Hg soil-gas and Hg soil in the Gedongsongo area.

ID Temp at 60 cm depth Rn Tn CO 2 Rn/Tn Hg soil-gas Hg soil

from the Gedongsongo area are classified as immature waters

(locat-ed to the Mg apex), so the use of chemical geothermometers for

esti-mating subsurface temperatures is not appropriate for this system

The use of silica (quartz-no steam loss) geothermometers may

there-fore be acceptable for estimating reservoir temperatures of the

Ungaran geothermalfield However, estimating reservoir

tempera-tures are rather low because maybe part of SiO2precipitated during

storage

4.2 Soil gas survey

4.2.1 Statistical interpretation of soil gas data

The soil gas measurements for all samples were conducted within

one or two days to minimize the influence of changes in

meteorolog-ical conditions on the soil gas compositions.Table 4shows the soil

gas concentrations of all samples collected in the Gedongsongo area

(Fig 3)

Threshold values, used to recognize anomalous concentrations in

the soil gas data, were calculated using the geometric mean plus

one standard deviation (Lepeltier, 1969; Klusman and Landress,

1979; Varekamp and Buseck, 1983; Lescinsky et al., 1987; Klusman,

1993) Samples with concentrations above this threshold are

consid-ered anomalous In geochemical exploration, cumulative frequency

diagrams are used for the determination of low (background) range,

anomalous samples or recognition of multiple populations in

log-normally distributed data (Lepeltier, 1969; Varekamp and Buseck, 1983; Lescinsky et al., 1987; Klusman, 1993) Individual populations were separated by visual assessment using the procedure outlined

byLepeltier (1969) The geometric mean, m, was read at the 50th per-centile; and the coefficient of deviation, σ, representing the spread in the data, is the logarithm of the ratio of the value one standard deviation from the geometric mean over the geometric mean Thus, soil gas data from the study area are classified into three populations

as low or background (I), high (II), and anomalous values (III) as shown in cumulative frequency diagrams (Fig 8andTable 5) High soil gas concentrations were found over broad areas, while anoma-lous concentrations were identified in the north of the fumarole in the Gedongsongo area Areas located at the east and south of the fu-marole had generally low soil gas concentrations

4.2.2 Spatial distribution of soil gas data and recognition of trace

of fault/fracture

To characterize the study area, the spatial distribution of soil gas data was interpreted using contour maps (Figs 9, 10 and 11) that were produced using the kriging method with interpolation based

on a linear variogram model provided by the Surfer software

Fig 9shows contour maps of the Rn concentration and the Rn and

Tn ratio The Rn results show that the high and anomalous concentra-tions occur in the northern parts of the surveyed area (200 m from the fumarole) (Fig 9a andTable 5) Anomalously high radon values

Fig 8 Cumulative frequency diagrams of Rn (a), Tn (b), CO 2 (c) Hg soil-gas (d) and Hg soil (e) in the Gedongsongo area Three populations: low (I); high (II); and anomalous (III) are shown.

Table 5

Distribution of soil gas into three populations (low, high and anomalous).

Classification Radon Thoron CO 2 Hg soil-gas Hg soil

High (σbconcentrationbσ+m) 123–221 320–770 7.7–13.1 37–71 4.3–6.0 Anomalous (σ+mbconcentrationbσ+2 m) 221–318 770–1266 13.1–20 71–105 6.0–7.8

could be indicative of enhanced permeability where Rn-222 rapidly

mi-grates to the surface before disintegrating into daughter products

How-ever, the release of radon is dependent on other factors, including the

degree of rock fractures and the ability of the ground water to permeate through such rocks Percolating ground water transports radon from fractured porous rocks by preventing diffusion As described above, the radon will partition into the steam phase and be transported to higher elevations through permeable zones A NNE–SSW alignment of

Rn anomalies, associated with the fault in this area, can be observed Other areas west of the fumarole especially on a WNW–ESE alignment also have relatively high Rn values, while low Rn values are observed

in most of the remaining survey area

Many studies have been published on the feasibility of using Rn and CO2measurements to detect active structures such as fractures and faults (King, 1980; Koga, 1988; Etiope and Lombardi, 1995; Giammanco, et al., 1998; Fu et al., 2005; Yang et al., 2005; Lan et al.,

2007) Faults favor gas transport because they increase rock permeabil-ity, helping the gas ascending to the surface Furthermore, gases from a deep source can migrate upward through faults where the gasflow is driven by advection As Tn has a short half-life, 55 s, its concentration

in counts per minute (cpm) decreases quickly during 3 min of sequen-tial measurement However, Rn has a half-life of 3.8 days and can be transported in fractures for a considerable distance To detect the pres-ence of fracture or fault systems connecting the deep zone to the sur-face, the Rn/Tn concentration ratio can be a suitable indicator.Fig 9b shows the Rn/Tn ratio contour map High Rn/Tn ratios (>0.4) are found mainly about 200 m to the north and 250–300 m to the south east of the fumaroles Zones with a high Rn/Tn ratio not only indicate

Fig 9 Contour map of (a) radon concentration and (b) radon to thoron concentration

ratio.

Fig 11 Contour maps of (a) Hg soil-gas and (b) Hg soil concentration.

the presence of a fault/fracture zone but also indicate the extending of

fault/fracture from deep zone to the surface

Fig 10is a contour map of the CO2concentration The CO2map

clearly shows a close relationship with the Rn, and Rn/Tn maps It is

relevant that in the Gedongsongo, CO2 values greater than 20% are

considered anomalous The high and anomalous CO2values (>10%)

were detected around the fumaroles and 250–300 m to south of

the fumarole while low CO2 values are mainly located in the east

(Fig 10) Zones with anomalous and high CO2concentrations trend

from NNE–SSW and WNW–ESE, and these can be postulated as fault

locations It is interesting to note that areas with high Rn/Tn ratios and

anomalous CO2concentration occur at the same location This

align-ment, trending NNE–SSW and WNW–ESE (Figs 9b and10) can be

pos-tulated as a fault in this area The NNE–SSW alignment of soil gas

anomalies agrees with topographic and geologic data of a fault zone

(Van Bemmelen, 1970; Thanden et al., 1996), while the WNW–ESE

alignment may be another fault zones, unknown prior to the soil gas

survey

The combined CO2concentrations and soil temperatures (Table 4)

are correlated to the north of the fumarole This observation may

identify the displacement of a high temperature heat source or an

up-flow zone of high temperature geothermal fluid (Lan et al.,

2007) The interpretation of Hg survey results will provide more

evi-dence on this potential up-flow zone

4.2.3 Mercury as feature of geothermal activities

The Hg concentrations have a wide range from 1.1 to 142.3 ng and

0.01 to 21.6 ppm for Hg Hgsoil-gasand Hgsoil, respectively.Fig 11a shows

a contour map of Hgsoil-gas The high concentration zone represents

Hgsoil-gasconcentrations above 40 ng, and extends from 0.7 km north

to south by 1 km west to east from the fumarole This supports the

sug-gestion that geothermal activity is widespread around the fumarole

zone Absorbed Hg on the soil surface is difficult to desorb except after

heating at high temperatures Thus, the resulting mercury anomalies

are related to the temperature of the ascending geothermalfluids,

which act as a carrier while also providing the migration pathways

anomalies occur when geothermalfluids escape from a deep reservoir

and migrate to shallow levels Anomalous Hgsoil-gasvalues were found

in the north of the fumarole which consistent with the location of the

anomalous Rn, Tn, CO2and temperature values The chemistry of spring

waters in this area is acid-sulfate type, indicating that a component of

thesefluids has condensed from H2S-rich vapor phase Mercury is

strongly partitioned into the vapor phase, so steam escaping towards

the surface will be enriched in Hg, and acid hot-springs and fumaroles

will display strong Hg anomalies We, therefore, postulate that

upwell-ing and subsequent boilupwell-ing occurs beneath the area to the north of the

fumarole The soil gas contours appear to define the NNE–SSW and

WNW–ESE alignments, which are thought to represent fault/fracture

zones Non-thermal waters or a mixture of thermal and non-thermal

waters are found in the south of fumarole and this area coincides with

low Hg concentrations

Mercury enrichment in soils is a dynamic process, as

re-volatilization and biogenic uptake with subsequent re-volatilization

the soil A steady-state will occur after an initial period of

non-equilibrium The continuous Hg loss process means that old and current

thermal activity can be distinguished (Koga, 1982, 1988; Varekamp and

Buseck, 1983) The Hg results in soil and in soil gas are not always in

good agreement because Hgsoil-gasindicates current geothermal activity

while Hgsoilshows the history of geothermal activity up to the present

(Koga, 1982, 1988).Fig 11b shows high Hgsoilconcentrations are

locat-ed in the east of the fumarolic area, and we infer that was an ancient

geothermal zone Varekamp and Buseck (1983) have documented

cases where active geothermal zones are enriched in Hgsoil-gasbut lack

Hg

4.3 Conceptual hydro-geochemical model of the Ungaran geothermal field

From this study, a hydro-geochemical model of the Gedongsongo thermal waters has been developed and is shown inFig 12 This model shows the up-flow of high temperature geothermal fluid located in the north of the fumarole The anomalous Hg values occur in the vapor-dominated part of the system, which is explained

by the strong partitioning of Hg into the vapor phase during boiling Deep geothermalfluids are present below this area, and circulate through the fault and fracture zones with some portion of thefluids discharging at the surface Based on the geochemical and isotopic data, the thermal springs at Gedongsongo are of meteoric origin The meteoric waters percolate through the fault systems into the mixing zone where they are heated by deep geothermalfluids and as-cend to the surface along the NNE–SSW and WNW–ESE fault In the shallow zone, CO2and H2S rich steam rises off the thermal waters, which can lead to formation of sulfate-rich waters, while some of the ascending thermal waters mix with cold water and rise along the WNW–ESE fault The areas with low soil gas concentrations show that the boundary of the limited geothermal system is in the northern section of the fumarolefield at the Gedongsongo area

5 Conclusions The chemistry of the thermal waters discharged in the Gedongsongo area indicates steam that heated acid-sulfate waters (Ca–(Na)–Mg–

SO4) are present in the north of the fumarole while mixed bicarbonate (Ca–Mg–HCO3) and bicarbonate–chloride (Na–HCO3–Cl or Na–Ca–Cl– HCO3) water is present in the south and southeast of the Gedongsongo area The compositions of the thermal waters reveal that they have not reached chemical equilibrium with the host rocks The reservoir tem-peratures estimated using silica geothermometers is from102 °C to

142 °C

A soil gas survey was also conducted at the Ungaran geothermal field and has provided an overview of soil gas distribution patterns produced by the underlying system The soil gas contour maps show that, Rn, and CO are reliable and sensitive indicators for tracing

Fig 12 Simplified conceptual hydro-geochemical model of the Ungaran geothermal field.