I ••
11
Terrestrial Fluids, Earthquakes and Volcanoes: The Hiroshi Wakita Volume III Edited by Nemesio M. Pérez Sergio Gurrieri Chi-Yu King Yuri Taran
Birkhäuser Basel · Boston · Berlin
Reprint from Pure and Applied Geophysics (PAGEOPH), Volume 165 (2008) No. 1 Editors: Nemesio M. Pérez Environmental Research Division Instituto Tecnológico y de Energias Renovables Polígono Industrial de Granadilla s/n 38611 Granadilla, Tenerife Canary Islands Spain e-mail:
[email protected]
Sergio Gurrieri Istituto Nazionale di Geofisica e Vulcanologia Sezione di Palermo V. Ugo La Malfa, 153 90146 Palermo Italy e-mail:
[email protected]
Chi-Yu King Earthquake Prediction Research, Inc 381 Hawthorne Ave. Los Altos, CA 94022 USA e-mail:
[email protected]
Yuri Taran Volcanology Department Institute of Geophysics UNAM 3000, Av. Universidad Mexico D.F., 04510 Mexico e-mail: taran@geofisica.unam.mx
Library of Congress Control Number: 2006043001 Bibliographic information published by Die Deutsche Bibliothek: Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the Internet at
ISBN 978-3-7643-8737-2 Birkhäuser Verlag AG, Basel · Boston · Berlin This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in other ways, and storage in data banks. For any kind of use permission of the copyright owner must be obtained.
© 2008 Birkhäuser Verlag AG Basel · Boston · Berlin P.O. Box 133, CH-4010 Basel, Switzerland Part of Springer Science+Business Media Printed on acid-free paper produced from chlorine-free pulp. TCF ∞ Printed in Germany ISBN 978-3-7643-8737-2 987654321
e-ISBN 978-3-7643-8738-9 www.birkhauser.ch
PURE AND APPLIED GEOPHYSICS Vol. 165, No. 1, 2008
Contents 1
Introduction N. M. Pe´rez, S. Gurrieri, C.-Y. King, Y. Taran
5
Spatial and Temporal Changes of Groundwater Level Induced by Thrust Faulting Y. Chia, J. J. Chiu, Y.-H. Chiang, T.-P. Lee, C.-W. Liu
17
Geochemical Monitoring of Geothermal Waters (2002–2004) along the North Anatolian Fault Zone, Turkey: Spatial and Temporal Variations and Relationship to Seismic Activity S. Su€er, N. Gu€lec¸, H. Mutlu, D. R. Hilton, C. C¸ifter, M. Sayin
45
Coupling Between Seismic Activity and Hydrogeochemistry at the Shillong Plateau, Northeastern India A. Skelton, L. Claesson, G. Chakrapani, C. Mahanta, J. Routh, M. Mo¨rth, P. Khanna
63
Radon Changes Associated with the Earthquake Sequence in June 2000 in the South Iceland Seismic Zone P. Einarsson, P. Theodo´rsson, A´. R. Hjartardo´ttir, G. I. Guðjo´nsson
75
CO2 Degassing over Seismic Areas: The Role of Mechanochemical Production at the Study Case of Central Apennines F. Italiano, G. Martinelli, P. Plescia
95
Changes in the Diffuse CO2 Emission and Relation to Seismic Activity in and around El Hierro, Canary Islands E. Padro´n, G. Melia´n, R. Marrero, D. Nolasco, J. Barrancos, G. Padilla, P. A. Herna´ndez, N. M. Pe´rez
115
SO2 Emission from Active Volcanoes Measured Simultaneously by COSPEC and mini-DOAS J. Barrancos, J. I. Rosello´, D. Calvo, E. Padro´n, G. Melia´n, P. A. Herna´ndez, N. M. Pe´rez, M. M. Milla´n, B. Galle
135
Underground Temperature Measurements as a Tool for Volcanic Activity Monitoring in the Island of Tenerife, Canary Islands A. Eff-Darwich, J. Coello, R. Vin˜as, V. Soler, M. C. Martin-Luis, I. Farrujia, M. L. Quesada, J. de la Nuez
147
Carbon Dioxide Discharged through the Las Can˜adas Aquifer, Tenerife, Canary Islands R. Marrero, D. L. Lo´pez, P. A. Herna´ndez, N. M. Pe´rez
173
List of Publications
Hiroshi Wakita was born in Nishinomiya (Hyogo, Japan) on September 29, 1936, and is actually an Emeritus Professor at The University of Tokyo (Japan). He graduated with B.Sc., M.Sc. and Ph.D. from Gakushuin University (Tokyo) in 1962, 1964 and 1968, respectively. He was a Researcher at the Japan Atomic Energy Research Institution, Tokyo (1964–68), Research Associate at Oregon State University, Corvallis, USA (1968– 71), Research Associate at the University of Tokyo (1971–77), Lecture at The University of Tokyo (1977–78), Associate Professor at The University of Tokyo (1978–86), Professor at The University of Tokyo (1986–97), and Professor at the Gakushuin Women’s College, Tokyo (1998–2007). He was also the Director of the Laboratory for Earthquake Chemistry at The University of Tokyo (1988–97), Associate Editor of Applied Geochemistry (1992–96), President of the Geochemical Society of Japan (1992–93), and Vice-president of the Geochemistry Research Association (1996). He received the Miyake Prize for his contributions on the field of geochemistry from the Geochemistry Research Association in 1989.
Pure appl. geophys. 165 (2008) 1–3 0033–4553/08/010001–3 DOI 10.1007/s00024-007-0295-3
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
Introduction
Terrestrial Fluids, Earthquakes and Volcanoes: The Hiroshi Wakita Volume III is a special publication to honor Professor Hiroshi Wakita for his scientific contributions to science, which have been closely linked with one of the major objectives of the 2008 International Year for the Earth Planet. Reducing natural risks in active tectonic and volcanic environments by searching for and detecting early warning signatures related to earthquakes and volcanic eruptions has been a major research goal for Hiroshi Wakita. The volume III consists of nine original papers written by researchers from Taiwan, Italy, Turkey, Iceland, USA, Sweden, India and Spain dealing with various aspects of the role of terrestrial fluids in earthquake and volcanic processes, which reflect Prof. Wakita’s wide scope of research interests. The volumes I and II consist of 17 and 10 original contributions which were published in Pure and Applied Geophysics on May 2006 and December 2007, respectively. These Pure and Applied Geophysics Hiroshi Wakita volumes should be useful for active researchers in the subject field, and graduate students who wish to become acquainted with them. Professor Wakita founded the Laboratory for Earthquake Chemistry in April 1978 with the aim of establishing a scientific base for earthquake prediction by means of geochemical studies, and served as its director from 1988 until his retirement from the university in 1997. He has made the laboratory a leading world center for the study of earthquakes and volcanic activities by means of geochemical and hydrological methods. Together with his research team and numerous foreign guest researchers whom he attracted, he has made many significant contributions in the above-mentioned scientific fields of interest. This achievement is a testimony for not only his scientific talent, but also his enthusiasm, his open-mindedness, and his drive in obtaining both human and financial support. The nine contributions of this volume III are arranged into two groups. The first group of five papers deals with the movement and signatures of terrestrial fluids related to earthquakes and active tectonic regions. The paper by CHIA et al. describes the observed changes of groundwater level induced by thrust faulting during the Mw 7.6, 1999 Chi-Chi earthquake recorded in 276 monitoring wells in Taiwan. Most of the observed coseismic falls appeared near the seismogenic fault as well as other active faults, while coseismic rises prevailed removed from the fault. The following paper by SU¨ER et al. presents the results of the 2002–2004 geochemical monitoring period of terrestrial fluids in geothermal fields located along an 800-km long E-W transect of the North Anatolian
2
N. M. Pe´rez et al.
Pure appl. geophys.,
Fault Zone (NAFZ) in Turkey, in order to both characterize the chemical nature of the individual fields and identify possible temporal variations associated with localized seismic activity. The paper by SKELTON et al. describes transient hydrogeochemical anomalies observed in a granite-hosted aquifer, which is located at a depth of 110-m, north of the Shillong Plateau, Assam, India. Their onsets preceded moderate earthquakes on December 9, 2004 (MW = 5.3) and February 15, 2005 (MW = 5.0), respectively, 206 and 213 km from the aquifer. The observation of these two hydrogeochemical events with the only two MW C 5 earthquakes in the study area argues in favor of causeand-effect seismic-hydrogeochemical coupling. The paper by EINARSSON et al. describes the observed radon changes in geothermal waters from drill holes related to an earthquake sequence at the transform plate boundary in South Iceland, which included two magnitude 6.5 earthquakes in June 2000. The authors emphasize that these radon anomalies were large and unusual if compared to a 17-year history of radon monitoring in this area. The paper by ITALIANO et al. provides field observations and new experimental data for the potential of the unexpected additional CO2 gas source production by mechanical energy applied to carbonate rocks in active tectonic regions beside mantlederived CO2 or CO2 produced by thermometamorphism. Data collected during the seismic crisis which struck the Central Apennines in 1997–1998 have shown an enhanced CO2 flux not associated with the presence of mantle or thermometamorphic-derived fluids. This earthquake-tectonic-related paper is then followed by four additional contributions dealing with observations related to volcanic processes. The paper by PADRo´N et al. describes the continuous monitoring of diffuse CO2 emission at El Hierro volcanic system, Canary Islands (Spain) and the observed geochemical anomalies before the occurrence of low magnitude seismic events in and around the volcanic island in 2004. The authors applied the material Failure Forecast Method (FFM) on the diffuse CO2 emission data to forecast successfully the first seismic event that took place in El Hierro in 2004. The following volcano-related paper by BARRANCOS et al. provides additional and recent SO2 emission data from eight active volcanoes: Santa Ana (El Salvador), San Cristo´bal and Masaya (Nicaragua), Arenal and Poa´s (Costa Rica), Tungurahua (Ecuador), Sierra Negra (Gala´pagos) and Etna (Italy). This paper also describes a comparison of SO2 emission measurements by COSPEC and mini-DOAS showing that most of the observed relative differences were lower than 10%. The paper by EFF-DARWICH et al. describes the spatial distribution of groundwater temperatures in Tenerife (Canary Islands) thanks to the vast network of *1.500 subhorizontal tunnels which provide most of the water resources for the island. Geological, hydrological and volcanological characteristics seem to be responsible for the actual groundwater temperature spatial distribution which has been characterized during a quiescent period, in order to detect changes in heat flow related to volcanic activity. The last volcano-related paper is also related to the groundwater system at Tenerife, Canary Island (Spain). The paper by MARRERO et al. provides an estimation of the water mass balance and the CO2 budget in Las Can˜adas’ aquifer; the largest aquifer on the island. The relatively high dissolved inorganic carbon
Vol. 165, 2008
Introduction
3
content in the groundwaters explains the ability of this aquifer to dissolve and transfer magmatic CO2, even during quiescence periods. The guest editorial team would like to thank all the contributors, and reviewers involved, who are listed below: R.M. Azzala, Werner Balderer, Alain Bernard, Emily Brodsky, Giorgio Capasso, Carlo Cardellini, Yeeping Chia, Antonio Eff-Darwich, Williams C. Evans, Cinzia Federico, Fausto Grassa, Jens Heinicke, Pedro A. Herna´ndez, David Hilton, George Igarashi, Kohei Kazahaya, Naoji Koizumi, Paolo Madonia, Rayco Marrero, Norio Matsumoto, Agnes Mazot, Eleazar Padro´n, Antonio Paonita, J. W. Rudnicki, Francesco Sortino, Jean-Paul Toutain, Nick Varley, Giuseppe Vilardo and Vivek Walia. Special thanks are due to Kenneth McGee, who served as co-guest editor in The Hiroshi Wakita volume I, for his support of this special issue, to Pedro A. Herna´ndez for his great assistance to the Guest-Editorial team, and to Renata Dmowska, without whose marvellous and tremendous support help the third special volume would not have been possible.
Nemesio M. Pe´rez Environmental Research Division Instituto Tecnolo´ gico y de Energı´as Renovables (ITER) Tenerife, Canary Islands Spain Sergio Gurrieri Istituto Nazionale di Geofisica e Vulcanologia, V. Ugo La Malfa 153 - 90146 Palermo Italy
Chi-Yu King Earthquake Prediction Research, Inc. 381 Hawthorne Ave. Los Altos, CA 94022 USA Yuri Taran Institute of Geophysics Universidad Nacional Auto´noma de Me´xico (UNAM) Mexico D.F 04510 Mexico
Pure appl. geophys. 165 (2008) 5-16
' Birkhauser Verlag, Basel, 2008
0033^553/08/010005-1 2
D , , , . ^ ^ » H Ar^r^li^r l
r-^^,^l.woi..o
’ ^^^^ ^" ^ Applied Geophysics
DOT 10.1007/S00024-007-0293-5
Spatial and Temporal Changes of Groundwater Leve l Induced by Thrust Faulting YEEPIN G CHIA / JESSIE J. CHIU,^ YI-HSUA N CHIANG / TSAI-PIN G L E E / and CHEN-WUM G LIU"^
Abstract Changes of groundwater level, ranging from a fall of 11.10 m to a rise of 7.42 m, induced by thrust faulting during the 1999 M ^ 7.6, Chi-Chi earthquake have been recorded in 276 monitoring wells in Taiwan. Most coseismic falls appeared near the seismogenic fault as well as other active faults, while coseismic rises prevailed away from the fault. Coseismic groundwater level rises and falls correlated fairly well with hypocentral distance in the vicinity of the thrust fault. W e found a major difference of coseismic changes in wells of different depths at most multiple-wel l stations. The recovery process of coseismic groundwater level changes is associated with the confining condition of the aquifer. Cross-formational flow is likely to play an important role in groundwater level changes after the earthquake. In the hanging wall of the thrust fault, an abnormal decline of groundwater level was observed immediately before the earthquake. The underlying mechanism of the unique preseismic change warrants further investigation. Key words: Groundwater, Chi-Chi earthquake. Thrust fault, Coseismic, Postseismic, Preseismic.
1. Introduction Water level in a well-confine d aquifer could be sensitive to crustal strain 1967). Field observations have shown a correlation between the estimated tectonic strains and the coseismic changes of well water level during the 1974 Izu-HantoOki earthquake (WAKITA , 1975). Variations of groundwater level in seismic regions have been used to monitor crustal deformation and to search for an earthquake precursor (BREDEHOEFT ,
(BAKU N and LINDH , 1985; KISSI N et al., 1996).
Coseismic groundwater level changes have been reported in many places around the world (MONTGOMER Y and MANGA , 2003), howeve r most changes are either sparsely distributed or concentrated in a few spots. Postseismic changes may reflect the subsurface flow in response to coseismic changes or permeability changes (ROELOFFS, 1998; WAN G et al., 2004). The observation of the flow process, particularly after small changes, is often impeded by pumping or other hydrologic factors. Preseismic changes are seldom
^ Department of Geosciences, National Taiwan University, Taipei 106, Taiwan. E-mail: [email protected] ^ Atomi c Energy Council, Yonghe , Taipei 234, Taiwan. ^ Department of Bioenvironmen t System s Engineering, National Taiwan University, Taipei 106, Taiwan.
6
Y . Chia et al.
Pure appl. geophys.,
reported (ROELOFFS and QUILTY , 1997; KOIZUM I and TSUKUDA , 1999), and the supporting evidence and underlying mechanism of their relations to fault deformation or earthquakes are not clear (KIN G et al, 2000). Whil e studies pertaining to the distribution and process of earthquake-related hydrologic phenomena are hampered by limited data, preliminary clues have been obtained by examining groundwater level changes recorded by a dense monitoring well network in the vicinity of a thrust fault ruptured during a large earthquake in Taiwan. Here we use monitoring records before, during, and after the earthquake to enhance our understanding of the spatial and temporal distribution as well as the possible mechanisms of groundwater level changes induced by thrust faulting.
2. Earthquake and Monitoring Wells On 21 Septembe r 1999, an earthquake of Mw 7.6 occurred near the town of Chi-Chi in central Taiwan at 1:47 a.m. local time. The hypocentral depth was estimated to be 10 km (SHIN et al., 2000). The best fitting focal mechanism has a nodal plane with a strike of 5 , a dip of 34 and a rake of 65 (CHAN G et al, 2000; KA O and CHEN, 2000). A s shown in Figure 1, widespread surface rupture resulted from thrusting along the Chelungpu fault extended approximately 100 km in the north-south direction (ANGELIE R et al, 2003). The hanging wall is on the east side of the thrust fault. Field investigations and GPS data indicated that the hanging wall move d as much as 10.1 m laterally and 8 m vertically. In contrast, up to 1.5-m lateral displacement and 0.26-m vertical displacement were observed in the footwall (Y u et al., 2001). In the coastal plain of Taiwan, the second generation network of monitoring wells had been installed since 1992 for improving groundwater resource management. A t the time of the Chi-Chi earthquake, 377 monitoring wells were operational in Taiwan. In the vicinity of the seismogenic fault, all wells, except one, are located in the footwall (west side) of the fault. Groundwater level is recorded by the digital data logger at one-hour interval. Som e wells equipped with the analog data logger also provide continuous records. Al l of the monitoring wells were screened in highly permeable sand or gravel layers. An y changes of groundwater level in the aquifer can, therefore, be quickly reflected by changes of water level in the monitoring well.
3. Spatial Distribution of Coseismic Groundwater Changes Coseismic changes of groundwater level in central western Taiwan due to the Chi-Chi earthquake have been discussed by CHIA et al. (2001) and WAN G et al. (2001). The mechanisms of these coseismic changes have been discussed by LEE et al. (2002), WAN G et al. (2003), KOIZUM I et al. (2004) and LA I et al. (2004). Tw o types of coseismic changes of water level were observed: oscillatory changes and persistent changes.
Vol . 165, 2008
Changes of Groundwater Leve l Induced by Faulting
Figure 1 Map of Taiwan showing the spatial distribution of 276 wells located at 126 monitoring stations where coseismic groundwater level rises and falls were observed during the M ^ 7.6 Chi-Chi earthquake on Septembe r 21, 1999.
Oscillatory coseismic changes, recorded on analog records, were the response of water column in the well and pore pressure in the aquifer to passing earthquake waves (COOPER et al., 1965; Liu et ai, 1989). The amplitude of oscillatory changes diminished shortly after the earthquake. Persistent coseismic changes have been proposed to be the
8
Y . Chia et al.
Pure appl. geophys.,
response of formation fluid pressure to shear strain due to the redistribution of stress field resulted from fault movement (MuiR-Woo d and KING , 1993; GE and STOVER , 2000). It is analogous to the response of pore pressure in soils to shear strain under an undrained triaxial test (WANG , 1997). Of the 377 monitoring wells, persistent coseismic groundwater level changes were observed in 276 wells during the Chi-Chi earthquake. Figure 1 illustrates spatial distribution of coseismic groundwater level rises and falls based on hourly records of 276 wells at 126 monitoring stations in the coastal plain of Taiwan. Of those, 203 wells at 80 stations are located within 50 km from the thrust fault. Included are a coseismic rise of 7.42 m and a coseismic fall of 11.09 m (Figs. 2a and 2b), the largest changes ever documented. It is noticed that, in the footwall of the seismogenic fault in central western Taiwan, coseismic falls were primarily observed at the stations near the ruptured segment. In contrast, coseismic rises prevailed at the stations away from the fault. A t some stations in the transition area, both coseismic rises and falls were observed in wells of different depths. Similar distribution patterns were also observed in the footwall of the unruptured segment of the thrust fault in southwestern Taiwan as well as another active fault in southern Taiwan. There is a general trend in the variation of the magnitude of coseismic water level change with hypocentral distance, as shown in Figure 3. The
Figure 2 The largest coseismic rise and coseismic fall of groundwater level observed during the earthquake. Hypocentral distances are denoted on top of each panel, (a) Hourly records (solid line) of the HW 2 well showing a coseismic rise of 7.42 m from 1 a.m. to 2 a.m., while analog records (dotted line) showing a rise of 8.5 m at 1:47 a.m. (b) Hourly records of the JS2 well showing a coseismic fall of 11.09 m.
Vol . 165, 2008
Changes of Groundwater Leve l Induced by Faulting
9
Figure 3 Coseismic groundwater level rise and fall versus hypocentral distance. Here coseismic changes in unconfined aquifers were excluded because they failed to represent the actual changes.
equation of the best-fit regression curve for the scatter plot of coseismic fall versus hypocentral distance up to 70 km is log h = 6.48^.4 4 log d, where h is the magnitude of the coseismic change in meters and d is the hypocentral distance in kilometers. The squared correlation coefficien t (R^ ) of 0.83 indicates a good correlation between coseismic fall and distance from the hypocenter in the vicinity of the thrust fault. The coseismic fall becomes considerably small beyond a hypocentral distance of 70 km . The equation of the best-fit regression curve for the plot of coseismic rise against hypocentral distance is log h = 4.09-2.39 log d, with a squared correlation coefficien t of 0.53. This suggests a moderate correlation between coseismic rise and hypocentral distance. A s shown in Figure 3, large rises were observed primarily at stations located between 30 km and 60 km from the hypocenter. Nevertheless, the magnitude of coseismic rise at these multiple-wel l stations changes drastically with depth. Whil e the rise diminishes gradually beyond 60 km , moderate rises were observed in six wells beyond 130 km . Al l of these six wells are located in northern Taiwan, implyin g that the redistributed stress induced by thrust faulting may concentrate in certain areas far from the seismogenic fault. Seismi c shaking of the saturated porous medium provides another possible way of explaining the occurrence of coseismic rise at a distance from the epicenter (MONTGOMER Y and MANGA , 2003).
4. Postseismic Changes of Groundwater Level The recovery of groundwater level in response to coseismic changes began immediately after the earthquake in most wells. The recovery process in an aquifer was controlled primarily by groundwater flow, and the recovery rate of coseismic change
10
Y . Chia et al.
Pure appl. geophys.,
varied with tlie confining condition of an aquifer. In a confined aquifer, tlie recovery process of coseismic change could last for several week s to a few months (Fig. 4a). The coseismic change in a confined aquifer, defined as the water level change from 1 a.m. to
Figure 4 Temporal changes of groundwater levels under various confining conditions after the earthquake, (a) Recovery of water level in the JL3 well approximately 5 months after a coseismic rise in a confined aquifer, (b) Recovery of water level in the TC2 well approximately 16 hours after a coseismic fall in a partially confined aquifer, (c) Recovery of water level in the SL l well within one hour after a coseismic rise in an unconfined aquifer, (d) Analog records at HS l showing recovery in approximately 15 minutes after coseismic rise in an unconfined aquifer.
Vol . 165, 2008
Changes of Groundwater Leve l Induced by Faulting
11
2 a.m. Septembe r 21 on hourly records, is close to the actual coseismic change at 1:47 a.m. on analog records (Fig. 2a). On the contrary, the recovery rate of groundwater level in an unconfined or partially confined aquifer is considerably faster. For instance, the TC station is located in a groundwater recharge area. The shallow well, TCI , was installed in an unconfined aquifer while the deep well, TC2, was installed in an aquifer confined partially by the overlying silty layer. The coseismic change in TC2 took 16 hours to recover to a steady level (Fig. 4b). In an unconfined aquifer, such as SLl , a pulse-like water level change is typically observed on hourly records, indicating the recovery process was completed within one hour after the coseismic change (Fig. 4c). On analog records, however, most actual coseismic changes at 1:47 a.m. in the unconfined aquifer were recovered within tens of minutes after the mainshock (Fig. 4d). Consequently, only a fractional or unnoticeable water level change could be observed at 2 a.m. on hourly records. A s the recovery rate of coseismic change depends on the confining condition of an aquifer, the variation pattern of a coseismic water level change and its recovery process could be a potential criterion for characterizing the degree of aquifer confinement . Most monitoring stations are composed of 2 to 5 wells screened at different depths, ranging from 14 m to 300 m. Records of these multiple-wel l stations revealed clues to the variation of coseismic change in the vertical direction as well as to the occurrence of cross-formational flow in the subsurface after the earthquake. They also provide field evidence for interpreting earthquake-related groundwater level anomalies. First, the magnitude of coseismic change fluctuated in wells at different depths. For instance, the coseismic rise at the Y L station, as shown in Figure 5a, varies from 6.55 m in YL l at the depth of 69 m to 1.2 m in YL 4 at 198 m. The DZ station recorded a coseismic fall of 9 cm at DZ2, but a coseismic rise of 28 cm and 14 cm at DZ l and DZ4, respectively (Fig. 5c). Such a difference in the vertical direction provides a possible explanation for difficultie s in matching coseismic groundwater level changes with volumetric strain changes estimated from fault displacement based on simple dislocation models (IGARASH I and WAKITA , 1991; QUILT Y and ROELOFFS, 1997; GRECKSC H et al, 1999). Second, postseismic anomalous changes of groundwater level at some multiple-wel l stations revealed the occurrence of cross-formational flow in the subsurface after the earthquake. Regional groundwater flow is considered a steady-state condition before the earthquake. The variation of coseismic changes at different locations and depths, however, initiated a transient flow in both lateral and vertical directions after the earthquake. For instance, the permeable layers tapped by YL 2 and YL 3 are hydraulically interconnected as evidenced in Figure 5a by the overlap of their groundwater levels before the earthquake. It is noted that the coseismic rise was 6.46 m in YL2 , but only 4.12 m in YL3 . The hydraulic gradient induced by different coseismic changes drove groundwater flow downward, resulting in a decline of water level in YL 2 and a rise in
12
Y . Chia et al.
Pure appl. geophys.,
Figure 5 Temporal change of groundwater levels in different depths at three multiple-wel l stations, (a) The Y L station. The coseismic rise fluctuates in wells of different depths. The water levels at YL 3 and YL 4 continued to rise for a few days after the earthquake, (b) The LY station. Afte r a coseismic fall of 4.62 m and a postseismic fall of 1.80 m, the water level at LY l approached that of LY2 . (c) The DZ station. The cosiesmic rise appeared in DZ2 while coseismic falls were observed in DZ l and DZ4. The coseismic change in DZ3 cannot be identified.
YL 3 until they came close again four days after the earthquake. The continued rise of groundwater level after a coseismic rise, similar to YL3 , was also observed in YL 4 and several other wells.
Vol . 165, 2008
Changes of Groundwater Leve l Induced by Faulting
13
The confining conditions of some aquifers near the surface rupture zone may have been changed by the earthquake. For instance, both LY l and LY2 , located 5 km from the ruptured segment, were artesian wells before the earthquake (Fig. 5b). Instead of a gradual rise after the coseismic fall, the groundwater levels declined further for several months after the earthquake. This phenomenon implies that subsurface fracturing may have been developed through the artesian aquifer during the earthquake. Crossformational flow could be generated when the confinemen t of the aquifer was breached, resulting in a rapid dissipation of pore water pressure after the earthquake (ROJSTACZE R et al, 1995; KIN G et al, 1999). Both wells were no longer artesian after the earthquake.
5. Preseismic Changes of Groundwater Level Several abnormal changes of groundwater level, observed from tens of minutes to a few months before the Chi-Chi earthquake, have been identified. Further examination indicated that these changes were likely to be associated with local pumping, rainfall, or improper data management, except a rapid decline of groundwater level immediately before the earthquake in the SL l well (Fig. 4c). The SL l well, located approximately 1.5 km east of the ruptured fault, is the only monitoring well installed in the hanging wall. It was placed in a shallow unconfined aquifer. The water level in SL l had not been interfered by pumping since the beginning of monitoring in 1997 until the end of 2000. A t the beginning of each recession (falling) stage during this period, the water level usually declined slowly with the rate of decline increasing gradually or remaining steady. The decline had never exceeded 3 cm during the first 3 hours in all recession stages, except for the one right before the Chi-Chi earthquake (Fig. 6). It is observed that, after rainfalls from Septembe r 15 to 19, 1999, the water level in SL l rose gradually on Septembe r 19 and 20, as shown in Figure 7, but declined 4 cm abruptly over a 3-hour period right before a 14-cm pulse-like coseismic rise. By
Figure 6 Water-level decline in the SL l well during the first 3 hours in recession stages from 1997 to 2000. The largest decline, 4 cm, appeared right before the Chi-Chi earthquake.
14
Y . Chia et al.
Pure appl. geophys.,
Figure 7 Ten-day records of groundwater level in the SL l well showing a 4-cm decline from 22:00 Septembe r 20 to 1:00 Septembe r 21 immediately before a 14-cm coseismic rise at 2:00 Septembe r 21 (also shown in Fig. 4c). The decline continued after the earthquake and the rate of decline decreased gradually when the pulse-like coseismic rise and its recovery were ignored.
ignoring coseismic change and its recovery, the rate of water level decline decreased gradually during the beginning of the recession starting from 22:00 Septembe r 20. The 4-cm decline of groundwater level immediately before the earthquake and the unique decreasing rate of decline at the beginning of the recession stage, therefore, becomes a unique hydrologic anomaly in SLl . Whethe r the decline is caused by dilatational deformation in the hanging wall right before the earthquake remains to be investigated.
6. Conclusions The comprehensive data recorded at monitoring well stations during the M^ 7.6 Chi-Chi earthquake provide a preliminary framework of regional distribution of coseismic groundwater level changes. The spatial distribution of observed coseismic change may reflect the redistribution of stress field in the shallow subsurface induced by the displacement of seismogenic fault. Such a complex distribution of coseismic changes is not expected to be consistent with volumetric strains calculated from a simple dislocation model. It is desirable to develop a more sophisticated model, taking into consideration aquifer characteristics and other geologic conditions, to describe coseismic changes induced by fault displacement. Postseismic groundwater level changes are essentially the subsurface hydrologic responses to various coseismic changes. Records from multiple-wel l stations revealed the importance of vertical confinemen t and cross-formational flow in the water level recovery after coseismic changes, providing a basis for interpreting anomalous changes of groundwater level after the earthquake. Moreover, the coseismic change and the postseismic recovery process could be considered a natural hydraulic testing for characterizing the confining condition or estimating hydraulic parameters of aquifers.
Vol . 165, 2008
Changes of Groundwater Leve l Induced by Faulting
15
A n anomalous decline of groundwater level was observed in the hanging wall immediately before the earthquake, although the underlying mechanism of the phenomenon is not clear. A s large earthquakes occur frequently in Taiwan and the monitoring well network continues to expand, more earthquake-related water level records will be available to facilitate further investigations.
Acknowledgements W e gratefully acknowledge access to monitoring records and hydrogeologic infor› mation of the Water Resources Agenc y and the Central Geological Survey of Taiwan. This work is supported by the National Science Council of Taiwan (NSC-94-2116-M002 011) and the Water Resources Agenc y (MOEAWRA0950187) .
REFERENCE S
ANGELIER , J., LEE, J. C , CHU, H . T. , and Hu, J. C. (2003), Reconstruction of fault slip of the September 21st, 1999, Taiwan Earthquake in the asphalted surface of a car park, and co-seismic slip partitioning, J. Struct. Geol. 25, 345-350. BAKU N W . H . and LINDH , A . G. (1985), The Parkfield, California Earthquake prediction experiment, Science 229, 619-624. BREDEHOEFT , J. D. (1967), Response of well-aquifer systems to earth tides, J. Geophys. Res. 72, 3075-3087. CHAN G C.H., Wu , Y.M. , SHIN , T . C , and WAN G C.Y . (2000), Relocation of the 1999 Chi-Chi earthquake in Taiwan, Terrestrial, Atmos . and Ocean. Sci. 11, 581-590. CHIA , Y.P. , WANG , Y.S. , CHIU, J.J., and Liu, C.W . (2001), Changes of groundwater level due to the 1999 Chi-Chi earthquake in the Choshui River alluvial fan in Taiwan, Bull. Seismol . Soc. Am . 91, 1062-1068. COOPER, H . H., Jr., BREDEHOEFT, J. D., PAPADOPULOS , L S. , and BENNETT , R . R . (1965), The response of well-
aquifer systems to seismic waves, J. Geophys. Res. 70, 3915-3926. GE, S . and STOVER , S . C . (2000), Hydrodynamic response to strike- and dip-slip faulting in a half-space, J. Geophys. Res. 105, 25513-25524. GRECKSCH , G., ROTH , F., and KUMPEL , H.J. (1999), Co-seismic well-level changes due to the 1992 Roermond earthquake compared to static deformation of half-space solutions, Geophys. J. Int. 138, 470-478. IGARASHI , G . and WAKITA , H . (1991), Tidal responses and earthquake-related changed in the water level of deep wells, J. Geophys. Res. 96, 4269^278 . KAO , H . and CHEN, W . P . (2000), The Chi-Chi earthquake sequence: Active, out-of-sequence thrust faulting in Taiwan, Science 288, 2346-2349. KING , C.-Y. , AZUMA , S. , IGARASHI , G., OHNO, M . , SAITO , H., and WAKITA , H . (1999), Earthquake related water-
level changes at 16 closely clustered wells in Tono, Central Japan, J. Geophys. Res. 104, 13073-13082. KING , C.-Y. , AZUMA , S., OHNO, M . , ASAI , Y. , HE, P., KTTAGAWA , Y. , IGARASHI , G., and WAKITA , H . (2000), In
search of earthquake precursors in the water-level data of 16 closely clustered wells at Tono, Japan, Geophys. J. Int. 143, 469^77 . KissiN, I.G., BELIKOV , V.M. , and ISHANKULIEV , G.A . (1996), Short-term groundwater level variations in a seismic region as an indicator of the geodynamic regime, Tectonophysic s 265, 313-326. KOIZUMI , N . and TSUKUDA , E . (1999), Preseismic changes in groundwater level and volumetric strain associated with earthquake swarms off the east coast of the hu Peninsula, Japan, Geophys. Res. Lett. 26, 3509-3512. KOIZUMI , N . , LAI , W . - C , KTTAGAWA , Y. , and MATSUMOTO , N . (2004), Comment on "Coseismic hydrological changes associated with dislocation of the September 21, 1999 Chi-Chi earthquake, Taiwan" by Min LEE et al, Geophys. Res. Lett., 31, L13603, doi: 10.1029/2004GL019897.
16
Y . Chia et al.
Pure appl. geophys.,
LAI , W . - C , KOIZUMI , N . , MATSUMOTO , N . , KITAGAWA , Y. , LIN , C.-W. , SHIEH, C.-L. , and LEE, Y . - P . (2004), Ejfects
of seismic ground motion and geological setting on the coseismic groundwater level changes caused by the 1999 Chi-Chi earthquake, Taiwan, Earth Planets Space 56, 873-880. LEE, M. , LIU, T.K. , MA , K.F., and CHANG , Y.M . (2002), Coseismic hydrological changes with dislocation of the September 21, 1999 Chi-Chi earthquake, Taiwan, Geophys. Res. Lett. 29, doi: 10.1029/2002GL015116. LIU, L.B. , ROELOFFS, E., and ZHENG , X . Y . (1989), Seismically-induced water level fluctuations in the Wall well, Beijing, China, J. Geophys. Res. 94, 9453-9462. MONTGOMERY , D . R . and MANGA , M . (2003), Streamflow and water well responses to earthquakes. Science 300, 2047-2049. MuiR-WooD , R. and KING , G . C . P. (1993), Hydrologic signatures of earthquake strain, J. Geophys. Res. 98, 22035-22068. QuiLTY , E.G. and ROELOFFS, E.A . (1997), Water-level changes in response to the 20 December 1994 earthquake near Parkfield, California, Bull. Seismol . Soc. Am . 87, 310-317. ROELOFFS, E.A . (1998), Persistent water level changes in a well near Parkfield, California, due to local and distant earthquake, J. Geophys. Res. 103, 869-889. ROELOFFS, E.A . and Quilty, E. (1997), Water level and strain changes preceding and following the August 4, 1985 Kettleman Hills, California, Earthquake, Pure Appl . Geophys. 149, 21-60. RojSTACZER , S., WOLF , S. , and MICHEL , R . (1995), Permeability enhancement in the shallow crust as a cause of earthquake-induced hydrologic changes. Nature 373, 237-239. SHIN , T.C., Kuo, K.W. , LEE, W.H.K. , TENG , T.L. , and TSAI , Y . B . (2000), A preliminary report on the 1999 Chi-Chi (Taiwan) earthquake, Seismol . Res. Lett. 71, 23-29. WAKITA , H . (1975), Water wells as possible indicators of tectonic strain. Science 189, 553-555. WANG , C . Y. , CHENG, L . H., CHIN, C. V. , and Yu , S. B . (2001), Coseismic hydrologic response of an alluvial fan to the 1999 Chi-Chi earthquake, Taiwan, Geology 29, 831-834. WANG , C.Y. , DREGER , D.S., WANG , C.H., MAYERI , D., and BERRYMAN , J. G. (2003), Field relations among
coseismic ground motion, water level change and liquefaction for the 1999 Chi-Chi (M^ = 7.5) earthquake, Taiwan, Geophys. Res. Lett. 30, doi: 10.1029/2003GL017601. WANG , C . Y. , WANG , C . H., and Kuo, C. H. (2004), Temporal change in groundwater level following the 1999 (M^ = 7.5) Chi-Chi earthquake, Taiwan, Geofluids 4, 210-220. WANG , H . F. (1997), Effects of deviatoric stress on undrained pore pressure response to fault slip, J. Geophys. Res. 102, 17943-17950. Yu , S.B. , Kuo, L.C., Hsu, Y.J. , Su, H.H., Liu, C.C, Hou, C.S., LEE, J.F., LAI , T . C , LIU, C . C , LIU, C.L., TSENG ,
T.F., TsAi , C. S., and SHIN , T.C. (2001), Preseismic deformation and coseismic displacements associated with the 1999 Chi-Chi, Taiwan earthquake. Bull. Seismol . Soc. Am . 91, 995-1012. (Receive d June 1, 2007, revised Novembe r 12, 2007, accepted Novembe r 17, 2007) PubHshed Online First: February 1, 2008
To access this journal online: www.birkhauser.ch/pageoph
Pure appl. geophys. 165 (2008) 17–43 0033–4553/08/010017–27 DOI 10.1007/s00024-007-0294-4
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
Geochemical Monitoring of Geothermal Waters (2002–2004) along the North Anatolian Fault Zone, Turkey: Spatial and Temporal Variations and Relationship to Seismic Activity SELIN SU¨ER,1 NILGU¨N GU¨LEC¸,1 HALIM MUTLU,2 DAVID R. HILTON,3 CANDAN C¸IFTER,4 and MESUT SAYIN4
Abstract—A total of nine geothermal fields located along an 800-km long E-W transect of the North Anatolian Fault Zone (NAFZ), Turkey were monitored for three years (2002–2004 inclusive; 3-sampling periods per year) to investigate any possible relationship between seismic activity and temporal variations in the chemistry and isotope characteristics of waters in the fields. The geothermal fields monitored in the study were, from west to east, Yalova, Efteni, Bolu, Mudurnu, Seben, Kurs¸ unlu-C ¸ ankırı, Hamamo¨zu¨, Go¨zlek and Res¸ adiye. The chemical (major anion-cation contents) and isotopic (18O/16O, D/H, 3H) compositions of hot and cold waters of the geothermal sites were determined in order to both characterize the chemical nature of the individual fields and identify possible temporal variations associated with localized seismic activity. The geothermal waters associated with the NAFZ are dominantly Na-HCO3, whereas the cold waters are of the Ca-HCO3 type. The oxygen- and hydrogen-isotope compositions reveal that the hot waters are meteoric in origin as are their cold water counterparts. However, the lower d18O, dD and 3H contents of the hot waters point to the fact that they are older than the cold waters, and that their host aquifers are recharged from higher altitudes with virtually no input from recent (post-bomb) precipitation. Although no major earthquakes (e.g., with M C 5) were recorded along the NAFZ during the course of the monitoring period, variations in the chemical and isotopic compositions of some waters were observed. Indeed, the timing of the chemical/isotopic changes seems to correlate with the occurrence of seismic activity of moderate magnitude (3 < M < 5) close to the sampling sites. In this respect, Cl, 3H and Ca seem to be the most sensitive tracers of seismically-induced crustal perturbations, and the Yalova and Efteni fields appear to be the key localities where the effects of seismic activity on the geothermal fluids are most pronounced over the monitoring period. The present study has produced a ‘baseline’ database for future studies directed at characterizing the effects of moderate-major earthquakes on the composition of geothermal waters along the NAFZ. Future work involving longer monitoring periods with more frequent sampling intervals should lead to a better understanding of the underlying mechanism(s) producing the observed chemical and isotopic variations. Key words: North Anatolian Fault Zone, geothermal fluid, seismic activity, geochemical monitoring, Turkey. 1 Department of Geological Engineering, Middle East Technical University, 06531 Ankara, Turkey. E-mail: [email protected]; [email protected] 2 Department of Geological Engineering, Eskis¸ ehir Osmangazi University, 26480 Eskis¸ ehir, Turkey E-mail: [email protected] 3 Fluids and Volatiles Lab., Geosciences Research Division, Scripps Inst. of Oceanography, UCSD, La Jolla, CA 92093-0244, USA. E-mail: [email protected] 4 Department of Technical Research and Quality Control, General Directorate of Turkish State Hydraulic Works, TR-06100 Ankara, Turkey. E-mail: [email protected]; [email protected]
18
S. Su¨er et al.
Pure appl. geophys.,
1. Introduction In seismically-active areas, periodic or continuous monitoring of the chemistry of terrestrial fluids is an approach increasingly utilized for understanding both the mechanisms inducing earthquakes and the associated response in the affected region of the crust. Variations in the chemical and isotopic composition of terrestrial fluids are considered to reflect sub-surface physical and chemical processes, such as fluid mixing, micro-fracturing and associated permeability changes (KING, 1986; THOMAS, 1988; ZHANG, 1994; IGARASHI and WAKITA, 1995; SILVER and WAKITA, 1996; WAKITA, 1996;TOUTAIN and BAUBRON, 1999; KING and IGARASHI, 2002; KING et al., 2006). A number of geochemical tracers have been used to investigate such subsurface processes, with both dissolved and free-gas phase samples showing great sensitivity in recording responses associated with seismic activity. For example, monitoring studies of terrestrial fluids in earthquake-prone regions have revealed significant changes in 222Rn (HAUKSSON, 1981; TENG and SUN, 1986; WAKITA et al., 1989; VIRK and SINGH, 1993; IGARASHI et al., 1995; PE´REZ et al., 1998; DAS et al., 2005; WALIA et al., 2006) and H2 emissions (WAKITA et al., 1980; SATO et al., 1986), variations in helium and carbon isotope ratios (3He/4He and 13C/12C) (SANO et al., 1986; 1998; SOREY et al., 1993; ITALIANO and MARTINELLI, 2001; BRA¨UER et al., 2003), fluctuations of diffuse CO2 emission (SALAZAR et al., 2002; PE´REZ et al., 2003; PE´REZ and HERNA´NDEZ, 2007; PADRO´N et al., this volume), as well as changes in gas ratios, such as He/Ar, He/222Rn, N2/Ar, CH4/Ar and CO2/He (SUGISAKI, 1978; KAWABE, 1984; SUGISAKI and SUGIURA, 1986; SUGISAKI et al., 1996; HILTON, 1996; VIRK et al., 2001). In addition to the variations observed by targeting gases, significant variations have also been recorded (either as pre- or post-seismic signals) in the chemical and isotopic composition of water phase samples. Variations in groundwater level/discharge and hydrogeochemistry were observed related to the 1995 Kobe earthquake (M 7.2), and were attributed to mixing of waters with different chloride and tritium contents (TSUNOGAI and WAKITA, 1995; KING et al., 1995; SANO et al., 1998). TOUTAIN et al. (1997) recorded a *36% increase (above background values) in the Cl content of a mineral water prior to a Pyrenean earthquake (M 5.2) in France. PE´REZ et al. (1998) also registered a *10% increase in the Cl content of a mineral water prior to earthquakes (M 4.2) in the NW Iberian Peninsula. NISHIZAWA et al. (1998) reported a two-fold increase in both Cl and SO4 contents in Yugano hot spring waters, a few days after the onset of the 1995 seismic swarm (M < 3 to 4.8) in the Izu Peninsula, central Japan. Likewise, SONG et al. (2006) detected increases in the Cl and SO4 contents (up to 138% and 278% of mean values, respectively) in hot and artesian springs of the Kuantzeling area (west-central Taiwan) prior to a September 1999 earthquake (M 7.3). FAVARA et al. (2001) and ITALIANO et al. (2004) observed temporal variations in the chemical composition and temperature of thermal waters and dissolved gases in the Umbria region. CLAESSON et al. (2004) reported short-term preseismic anomalies of Cu, Mn, Zn, Fe and Cr in the groundwater prior to a magnitude 5.8 earthquake in the Tjo¨rnes
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
19
fracture zone, north Iceland. All these changes were attributed to subsurface mixing processes induced by stress/strain changes associated with localized earthquake activity. In terms of the isotope systematics of the water phase, measurements of dD and d18O concentrations of groundwaters in seismically-active regions are potentially useful for searching earthquake precursors and in elucidating mechanisms operative during earthquakes. O’NEIL and KING (1981) observed an increase in the dD content of groundwaters while the d18O content remained constant during a magnitude 5.7 earthquake at the Oroville Dam in 1975 and a series of events near SAN JUAN BAUTISTA in 1980. This increase suggested that H2O may have either decomposed or reacted to form molecular H2 at depth. The study of BOLOGNESI (2000) reported a large increase in d18O values in geothermal reservoir waters at Vulcano Island, Italy, following seismic activity that occurred near the island. The most prominent increase in d18O was a change from +1.0 ± 0.5% to 3.4 ± 0.5% after a major earthquake (M: 5.5) in April 1978. The response was attributed to earthquake-induced increases in the contribution of hightemperature, d18O-rich magmatic condensates to the geothermal reservoir. Stable isotope variations in groundwaters were also observed before and after the magnitude 5.8 earthquake in the Tjo¨rnes fracture zone, north Iceland, suggesting stress-induced source mixing and leakage of fluid from an external (hotter) basalt-hosted source reservoir, where fluid-rock interaction was more rapid (CLAESSON et al., 2004). In relation to the present work, BALDERER et al. (2002) reported results of sampling campaigns before and after the devastating 1999 earthquakes along the North Anatolian Fault Zone, and pointed out variations in both chemical (Ca, K, Na, NO3, SO4 and Cl) and isotopic (d18O and dD) compositions of the thermal and mineral waters in the Kuzuluk, Bursa and Yalova/Gemlik areas. They attributed the variations in chemical compositions to the mobilization of deep-seated brines in response to the seismic activity and their mixing with waters of surficial origin. On the other hand, variations in d18O and dD values were interpreted as resulting from isotopic exchange with CO2 and H2S gases, respectively, and/or mixing with groundwaters of shallow and/or deep origin with differing recharge conditions. In this study we extend the work of BALDERER et al. (2002) by targeting a total of nine geothermal localities located along the North Anatolian Fault Zone (NAFZ) in an attempt to identify anomalies in the chemical and isotopic composition of fluids which may be related to regional/localized seismic activity. The present approach involves a 3-year monitoring program with 3 sampling periods per year covering 2002 (March-JulyOctober), 2003 (April-July-October) and 2004 (April-June-October). The geothermal fields studied in this project are located along an 800-km long transect of the NAFZ and are, from west to east, Yalova, Efteni, Bolu, Mudurnu, Seben, Kurs¸ unlu-C¸ankırı, Hamamo¨zu¨, Go¨zlek and Res¸ adiye (Fig. 1). Both hot and cold waters were utilized and analyzed for their chemical (major anion-cation contents) and isotopic (18O/16O, D/H, and 3H) compositions.
20
S. Su¨er et al.
Pure appl. geophys.,
Figure 1 Locations of the sampled geothermal fields and the recent seismic activities along the NAFZ (after C ¸ EMEN et al., 2000).
2. NAFZ: Tectonic Setting and Recent Seismicity The North Anatolian Fault Zone (NAFZ), comprising one of the major geologic structures in Turkey, developed during the Neotectonic period in response to intra-continental convergence following the Late Miocene collision of the Arabian promontory with Eurasia (MCKENZIE, 1972; DEWEY and S¸ ENGo¨R, 1979; S¸ENGo¨R et al., 1985; BARKA, 1992). This fault zone is a 1500-km long, few hundred meters to 40 km wide, dextral strike-slip fault system which forms an intracontinental transform boundary between the Eurasian Plate in the north and the Anatolian Plate in the south (KOc¸YIg˘IT et al., 1999a, b). It consists of shorter subparallel fault strands locally displaying an anastomosing fault pattern along much of its length. The NAFZ extends from eastern Anatolia in the east to the Aegean Sea in the west. The fault zone bifurcates into two major strands just east of the Sea of Marmara. The northern strand traverses, and the southern strand bounds, the southern margin of the Sea of Marmara (Fig. 1). The age of dextral motion and the total offset along the fault zone are controversial, covering the range of Middle Miocene - Early Pliocene and 20–85 km, respectively (see BOZKURT, 2001 for a review). The other structural members of the Neotectonic period that
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
21
developed along with the NAFZ are the Bitlis Suture Zone (BSZ), East Anatolian Fault Zone (EAFZ) and the Western Anatolian Graben System (WAGS) (Fig. 1). The NAFZ is known worldwide as one of the great strike-slip fault zones that host many medium to large-scale earthquakes. The most recent destructive effects of the _ NAFZ were experienced in 1999 in Izmit (M 7.4; 17 August, 1999) and Du¨zce (M 7.2; 12 November, 1999) provinces, resulting in more than 25,000 fatalities in the region. The August and November earthquakes occurred, respectively, at depths of 11 km (with two successive shocks in 45 seconds at the centers of Go¨lcu¨k and Arifiye) and 10 km (epicenter Du¨zce), creating a total length of surface rupture of 177 km from Go¨lcu¨k eastwards (KOc¸YIG˘IT et al., 1999a, b; C ¸ EMEN et al., 2000) (Fig. 1). During the monitoring period of the present study (2002–2004 inclusive) the magnitudes of the earthquakes that occurred along the NAFZ were all less than 5.0, with the seismic activity mainly concentrated in the western and central segments of the fault zone (Fig. 1). The most recent earthquakes recorded along the NAFZ during the monitoring period with magnitudes greater than 3.0 include: the Sea of Marmara (M 4.7—23 March, 2002), Armutlu-Yalova (M 3.1—3 July, 2002 and M 3.1—13 July, 2002), Yıg˘ılca-Du¨zce (M 3.1—14 July, 2002, M 3.1—11 October, 2002, M 4.0—25 July, 2003, M 3.6—15 June, 2004, M 4.8—2 July, 2004), Kaynas¸ lı-Du¨zce (M 3.2—1 July, 2003 and M (3.1 and 4.0)—25 July, 2003), Mudurnu (M 3.4—1 November, 2002), C ¸ ınarcık-Yalova (M 3.5—22 July, 2003), and Bolu (M 3.5—7 August, 2003, M 4.7—14 April, 2004, M 4.6—23 June, 2004).
3. Regional Geology and Hydrogeologic Outline The basement rocks in the geothermal areas of this study are represented by Paleozoic metamorphics (schists and marbles). These units are unconformably overlain by Upper Jurassic-Lower Cretaceous limestones and Upper Cretaceous flysch consisting of intercalations of limestone, conglomerate, marl, sandstone, claystone and siltstone. Products of widespread volcanic activity, extending from Late Cretaceous to Miocene and consisting of basaltic-andesitic lava flows, tuffs and agglomerates, are observed either intercalated with, or overlying, the Upper Cretaceous flysch. In turn, these units are overlain by Neogene clastics and lacustrine limestones (S¸ AHINCI, 1970; CANIK, 1972; ¨ ZCAN and U ¨ NAY, 1978; MU¨FTu¨OG˘LU and AKıNCı, 1989; ERZENOG˘LU, 1989; KOC¸AK, 1974; O MTA, 1996). Plio-Quaternary fluviatile deposits form the youngest units of the succession and are accompanied, in the eastern-central segment of the NAFZ, by Plio-Quaternary volcanics of limited areal extent (TATAR et al., 1996). Numerous hot-springs emerge at geothermal fields located along the fault and associated fracture zones. The temperature of the hot-springs ranges between 30C and 74C, with flow rates between 0.1 and 5.6 l/s. Except for the Yalova and Kurs¸ unlu fields, where the primary reservoirs are comprised of Tertiary volcanics, the geothermal reservoir rocks are mainly
S. Su¨er et al.
22
Pure appl. geophys.,
Mesozoic limestones (KOC¸AK, 1974; MTA, 1996). Impervious clayey levels of the flysch facies and the Neogene sequence frequently act as cap rocks to the geothermal systems.
4. Sampling and Analytical Techniques Sampling was performed over a total of nine periods, comprising three discrete sampling intervals per year (March-July-October 2002, April-July-October 2003 and April-June-October 2004). This schedule was maintained for all geothermal fields (Yalova, Efteni, Mudurnu, Seben, Bolu, Kurs¸ unlu, Hamamo¨zu¨, Go¨zlek) with the exception of Res¸ adiye (where sampling ceased in October 2002). Both natural springs and production wells were utilized during sampling (Table 1). Cold water sampling commenced at most locations in 2003 in order to better evaluate possible proximal subsurface processes, such as mixing and boiling, which may contribute to compositional variations. The Res¸ adiye field could be sampled only in the year 2002 since the hotspring source dried-up due to excessive usage for balneological purposes. Furthermore, in April 2004, after sampling in Efteni and Bolu fields on 12 and 13 April, respectively, a second sampling campaign was mounted for both fields following the seismic activity of 14 April at Bolu Town Center (M 4.7). Samples were collected for major anion-cation contents, stable isotope ratios (18O/16O and D/H) and tritium (3H) contents using polyethylene bottles. The anioncation chemistry, tritium, 18O/16O and D/H analyses of water samples were carried out Table 1 Sample numbers, sample types and geographic coordinates of the geothermal fields Locality
Sample no.
Latitude
Efteni
1a-hot 1b-cold 2a-hot 2c-cold 3a-hot 3b-cold 4a-hot 4d-cold 5a-hot 5d-cold 6a-hot 6b-cold 7a-hot 7b-cold 8a-hot 8b-cold 9a-hot 9b-mineral 9c-cold
40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40
Yalova Bolu Mudurnu Seben Hamamo¨zu¨ Go¨zlek Res¸ adiye Kurs¸ unlu
a
450 450 360 370 410 410 270 270 190 190 460 460 330 330 230 230 490 490 500
4200 4200 1600 1400 1000 0500 3300 3500 3100 3100 5900 5900 0800 0800 3100 3100 3400 5700 0400
a
Longitude N N N N N N N N N N N N N N N N N N N
031 031 029 029 031 031 031 031 031 031 035 035 035 035 037 037 033 033 033
010 010 100 130 370 370 140 140 320 320 010 010 400 400 200 200 100 100 100
4100 4100 1200 6800 1600 0500 2700 2600 2400 2400 2400 2400 3600 3600 1500 1500 3200 4300 4300
E E E E E E E E E E E E E E E E E E E
Sample type
Spring Spring Spring Spring Production well Spring Production well Spring Spring Spring Production well Spring Production well Spring Spring Spring Production well Mineral spring Spring
Numbers given beside production wells represent production depths, A: Artesian, P: Pumping.
(P)-83 m (A)-80 m
(A) (A)-800 m
(A)-165 m
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
23
at the laboratories of the Turkish State Hydraulic Works (DSI). Anion-cation analyses were performed by conventional methods (titration for Ca, Mg, HCO3, CO3 and Cl, flame photometry for Na and K, and spectrophotometry for SO4). Tritium contents were determined by a liquid scintillation counting system (Packard Tri Carb 2260 XL) after electrolytic enrichment (ALTAY and C ¸ IFTER, 1996). d18O and dD ratios were measured on a Micromass 602C mass spectrometer, using the standard water equilibration (EPSTEIN and MAYEDA, 1953) and zinc reduction (COLEMAN et al., 1982) methods, respectively.
5. Results The results of the chemical and isotopic analyses, as well as the in-situ temperature, pH and EC measurements, are given for all sampling periods in Table 2. This dataset includes measurements on both the hot and cold water samples. As seen in Table 2, temperatures range between 37.4–72.6C for the hot waters, and 6.7–20.5C for the cold waters. The hot waters are slightly acidic to slightly alkaline in character, with pH values ranging between 5.9 and 8.0. In contrast, the cold waters are relatively more alkaline, with pH values from 6.5 to 8.8. The TDS values (Total Dissolved Solid content) are much higher for the hot waters (ranging up to 11181 mg/l) compared to those for the cold waters (ranging up to 996 mg/l). Among all samples, Kurs¸ unlu 9a hot water (collected from a production well) has the highest TDS value which may be due, in part at least, to steam phase separation which has resulted in calcite scaling in the well. Regarding the stable isotope compositions (expressed with respect to V-SMOW), the d18O values of the hot waters range between -13.4% and -8.4%, while the cold waters have values ranging from -12.7% to -8.1%. The dD values, on the other hand, range between -98.3% and -64.2% for the hot, and between -86.1% and -54.7% for the cold waters. Tritium (3H) contents of the hot waters lie between 0–12.55 TU, with the majority being less than 5 TU. For the cold waters, the range is 3.00–15.70 TU, although most of the samples have concentrations above 8 TU. 5.1. Hydrogeochemical Facies The hydrogeochemical facies of the waters, defined by the dominant cation-anion pairs, are represented as pie diagrams in Figure 2. As observed, the hot waters are mostly Na-HCO3 type waters (Efteni, Seben, Go¨zlek, Res¸ adiye and Kurs¸ unlu samples), except for the Ca-HCO3 type at Bolu and Mudurnu, and the Na-SO4 type at Yalova (Fig. 2). The only geothermal field which displays a mixed-character water is Hamamo¨zu¨ since none of its individual ions exceeds 50% of the total composition. The cold waters at this locality, on the other hand, display Ca and/or Mg-HCO3 character.
S. Su¨er et al.
24
Pure appl. geophys.,
Table 2 Results of chemical analyses Locality
T
Efteni-1a 23/3/02 42.3 8/7/02 41.4 27/10/02 43.4 07/4/03 43.0 09/7/03 43.1 15/10/03 43.3 12/4/04 43.6 14/4/04 43.4 28/6/04 44.7 11/10/04 44.4 average 43.2 r 1.0 VC% 2.3 Efteni-1b 07/4/03 11.2 09/7/03 20.0 15/10/03 12.6 12/4/04 18.5 28/6/04 19.8 11/10/04 20.0 average 17.0 r 4.0 VC% 23.7 Yalova-2a 24/3/02 60.1 09/7/02 60.6 28/10/02 56.9 08/4/03 64.3 10/7/03 65.0 16/10/03 63.4 13/4/04 65.7 28/6/04 53.8 12/10/04 60.6 average 61.2 r 3.9 VC% 6.4 Yalova-2c 08/4/03 13.5 10/7/03 20.0 16/10/03 18.2 13/4/04 14.4 28/6/04 17.1 12/10/04 20.2 average 17.2 r 2.8 VC% 16.3 Bolu-3a 24/3/02 41.6 09/7/02 41.6
pH
EC
HCO3
Cl
SO4
Na
K
Ca
Mg
TDS
d18O
6.5 7.4 7.7 6.2 6.3 6.4 6.2 6.2 6.3 6.3 6.5 0.5 8.0
3090 3020 3050 3070 3010 3080 3053 38 1
1835 1793 1702 1704 1765 1580 1787 1756 1608 1560 1709 96 6
197 209 206 201 196 198 202 206 195 192 200 5 3
1 1 4 5 5 5 4 4 4 2 3 2 44
450 520 298 308 347 232 339 329 302 263 339 86 25
18 18 13 13 12 11 12 12 12 13 13 3 19
164 104 141 219 63 86 170 168 53 223 139 61 44
135 148 163 109 194 206 140 143 193 105 154 35 23
2800 2793 2526 2557 2582 2318 2654 2617 2365 2358 2557 171 7
-11.1 -11.2 -11.3 -11.7 -10.8 -11.6 -11.5 -11.0 -11.8 -10.5 -11.3 0.4 3.7
-83.1 0.00 ± -83.5 0.05 ± -80.0 2.20 ± -81.8 0.00 ± -81.9 4.85 ± -79.5 4.55 ± -86.0 0.00 ± -87.5 4.00 ± -84.7 5.80 ± -81.5 0.90 ± -82.9 2.24 2.6 2.35 3.1 105.00
1.65 1.55 1.70 1.60 1.95 1.90 2.05 1.90 1.90 1.85
8.8 8.8 8.6 8.5 8.3 8.5 8.6 0.2 2.4
195 208 225 253 209 229 22 10
134 145 120 133 13 10
9 4 8 7 3 39
7 6 8 7 1 20
2 0 2 0 1 0 2 0 1 0 33 33
14 11 7 11 4 33
23 23 23 23 0 1
190 191 168 183 13 7
-11.1 -12.1 -11.7 -11.5 -11.6 -10.8 -11.5 0.4 3.9
-80.9 -78.6 -80.8 -83.4 -78.1 -78.8 -80.1 2.0 2.5
3.00 ± 4.90 ± 5.65 ± 6.75 ± 4.50 ± 4.45 ± 4.88 1.26 25.87
1.60 2.00 1.90 2.20 1.90 1.95
7.9 7.8 7.7 7.4 7.6 7.7 7.3 7.8 7.5 7.6 0.2 2.6
1922 1902 1913 1916 1923 1912 1917 6 0
48 28 31 43 43 81 43 49 39 45 15 34
101 69 96 97 97 95 104 94 94 94 10 11
875 950 816 750 764 820 770 810 824 820 62 8
280 6 330 6 241 5 243 5 209 4 290 4 269 4 299 4 199 5 262 5 43 1 16 18
151 2 154 0 162 16 157 7 199 2 154 7 158 0 142 0 162 38 160 8 16 12 10 152
1463 1537 1365 1301 1316 1451 1349 1398 1361 1393 77 6
-11.3 -11.4 -11.7 -11.1 -11.7 -11.2 -11.6 -11.4 -10.5 -11.3 0.4 3.2
-75.0 8.90 ± -75.4 12.55 ± -72.2 3.30 ± -75.0 3.70 ± -77.3 0.05 ± -73.8 0.00 ± -80.3 1.10 ± -77.1 0.00 ± -76.2 1.90 ± -75.8 3.50 2.3 4.41 3.0 126.11
1.90 2.00 1.75 1.70 1.80 0.60 2.05 1.90 1.90
7.4 7.6 7.8 7.5 7.1 7.6 7.5 0.3 3.4
600 557 574 572 606 584 19 3
332 298 320 300 322 262 306 25 8
14 16 15 14 14 14 14 1 5
38 28 30 34 29 86 41 23 55
15 2 11 1 37 2 13 2 12 1 16 2 17 2 10 0 57 26
513 461 488 467 488 497 486 19 4
-8.8 -8.8 -9.0 -9.6 -9.8 -8.4 -9.1 0.5 5.8
-60.4 -54.7 -57.3 -61.0 -61.8 -59.1 -59.0 2.7 4.5
1.80 2.10 0.90 2.35 1.85 1.85
7.0 6.2
-
817 512
7 625 5 675
105 102 75 100 106 108 99 12 13
8 5 10 4 4 7 6 2 38
dD
52 18 344 112 1975 -11.9 -88.8 55 17 304 53 1622 -12.2 -87.7
3
H
7.10 ± 8.95 ± 8.10 ± 12.35 ± 5.90 ± 6.10 ± 8.08 2.39 29.60
9.75 ± 1.90 4.65 ± 1.80
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
25
Table 2 (Contd.) Locality
T
28/10/02 43.0 08/4/03 42.5 10/7/03 43.1 16/10/03 41.8 13/4/04 37.4 14/4/04 40.8 27/6/04 43.4 12/10/04 44.5 average 42.0 r 1.9 VC% 4.6 Bolu-3b 08/4/03 6.7 10/7/03 13.7 16/10/03 11.6 13/4/04 19.7 27/6/04 18.1 12/10/04 16.9 average 14.6 r 5.4 VC% 36.7 Mudurnu-4a 25/3/02 38.3 10/7/02 38.3 29/10/02 44.3 09/4/03 39.1 11/7/03 39.2 17/10/03 39.3 14/4/04 39.4 27/6/04 39.9 13/10/04 40.1 average 39.8 r 1.8 VC% 4.5 Mudurnu-4d 25/3/02 11.3 10/7/02 15.4 29/10/02 16.9 09/4/03 12.2 11/7/03 15.7 17/10/03 17.4 14/4/04 13.2 27/6/04 15.5 13/10/04 18.7 average 15.1 r 2.5 VC% 16.3
HCO3
Cl
SO4 Na
K
Ca
Mg
TDS
d18O
dD
1980 1961 1960 1957 1957 1957 0 0
787 955 786 663 793 788 808 648 756 120 16
11 10 16 9 14 14 11 14 11 4 32
373 326 325 482 499 451 298 412 447 127 29
47 22 33 38 40 41 39 46 41 10 23
18 18 14 16 16 16 15 17 16 1 8
328 360 335 356 369 361 273 351 338 30 9
27 41 20 15 27 30 48 16 39 29 74
1589 1731 1528 1579 1758 1702 1493 1504 1648 148 9
-8.6 -11.4 -12.1 -11.5 -12.4 -11.8 -12.5 -10.6 -11.5 1.2 10.2
-64.2 -82.7 -84.0 -79.4 -87.3 -89.1 -82.3 -82.1 -82.8 7.3 8.8
3.35 ± 3.30 ± 1.00 ± 1.85 ± 2.05 ± 3.65 ± 0.00 ± 0.90 ± 3.05 2.76 90.47
1.75 1.65 1.85 0.70 2.00 1.70 1.70 1.80
7.5 7.7 7.7 7.9 8.4 8.4 8.0 0.4 5.1
78 86 84 89 90 88 3 4
25 41 51 34 49 46 41 10 24
3 5 4 9 4 7 5 2 46
15 8 10 9 5 4 8 4 49
4 3 1 4 4 5 3 1 39
2 2 3 2 2 3 2 0 18
11 1 12 2 12 5 9 4 11 3 10 4 11 3 1 1 11 46
60 72 85 70 77 78 74 9 12
-11.6 -12.4 -12.2 -12.5 -12.0 -10.8 -11.9 0.6 5.4
-79.1 -79.2 -75.1 -84.5 -82.8 -75.7 -79.4 3.7 4.7
15.20 ± 11.95 ± 8.40 ± 11.20 ± 7.30 ± 8.70 ± 10.46 2.92 27.90
2.00 2.15 0.90 2.30 1.90 2.05
6.3 6.3 6.2 6.1 6.3 6.3 6.2 6.2 6.1 6.2 0.1 1.0
1162 1151 1162 1152 1148 1150 1150 2 0
767 787 689 780 744 695 753 726 634 731 50 7
1 3 7 11 13 8 11 8 12 8 4 49
30 10 34 32 29 31 25 27 21 27 7 28
30 34 23 46 16 5 20 18 23 24 12 48
9 8 6 7 5 7 6 6 7 7 1 18
24 43 46 43 70 56 25 66 26 44 17 39
1030 1066 949 1080 1010 950 1039 976 896 1000 61 6
-11.8 -12.2 -12.2 -11.7 -12.1 -12.0 -12.1 -12.7 -11.1 -12.0 0.4 3.7
-83.3 6.00 ± -88.0 3.90 ± -81.6 0.00 ± -87.3 0.65 ± -82.9 0.00 ± -84.3 0.25 ± -86.9 2.30 ± -86.0 0.00 ± -82.1 0.30 ± -84.7 1.49 2.4 2.15 2.8 144.68
1.80 1.85 1.60 1.65 1.75 0.60 2.00 1.80 1.90
7.3 7.1 6.9 7.0 7.1 7.1 7.2 7.0 7.0 7.1 0.1 1.6
791 743 808 743 797 839 793 48 6
415 432 476 464 495 461 518 432 462 35 8
8 5 9 9 12 11 7 13 9 3 28
72 50 35 36 37 24 28 75 45 19 43
24 28 18 17 13 15 14 19 18 5 29
4 108 32 5 130 16 4 82 46 4 97 37 2 69 63 3 114 25 4 76 58 4 99 47 4 97 40 1 21 16 21 21 40
662 666 669 662 692 653 704 689 674 18 3
-10.5 -11.4 -11.6 -12.3 -11.1 -11.1 -11.1 -10.9 -10.5 -11.2 0.6 5.0
-80.7 -82.7 -82.0 -82.6 -76.7 -74.5 -83.4 -79.1 -81.9 -80.4 3.1 3.8
2.05 1.95 1.70 1.85 2.05 0.80 2.20 1.90 1.95
pH
EC
6.1 6.1 5.9 6.3 6.2 6.0 6.2 6.1 6.2 0.3 4.7
168 182 144 161 133 147 199 126 174 159 24 15
3
H
12.46 ± 9.30 ± 6.35 ± 4.95 ± 8.15 ± 5.20 ± 5.25 ± 5.80 ± 4.55 ± 6.89 2.61 37.95
S. Su¨er et al.
26
Pure appl. geophys.,
Table 2 (Contd.) Locality
T
Seben-5a 25/3/02 70.1 10/7/02 69.9 29/10/02 72.3 09/4/03 70.7 11/7/03 71.8 17/10/03 72.3 14/4/04 71.6 27/6/04 72.6 13/10/04 71.5 average 71.4 r 1.0 VC% 1.4 Seben-5d 25/3/02 13.1 10/7/02 29/10/02 15.1 09/4/03 14.5 11/7/03 14.5 17/10/03 15.7 14/4/04 14.4 27/6/04 16.5 13/10/04 16.2 average 15.0 r 1.1 VC% 7.4 Hamamo¨zu¨-6a 26/3/02 41.3 11/7/02 41.1 30/10/02 42.6 10/4/03 42.1 12/7/03 42.7 18/10/03 42.5 15/4/04 42.5 25/6/04 43.6 14/10/04 44.4 average 42.5 r 1.0 VC% 2.4 Hamamo¨zu¨-6b 30/10/02 15.8 10/4/03 10.6 12/7/03 19.3 18/10/03 17.4 15/4/04 12.7 25/6/04 19.5 14/10/04 19.0 average 16.3 r 3.5 VC% 21.4
HCO3 Cl SO4
Na
K
Ca
Mg
TDS
d18O
2210 2180 2200 2190 2180 2190 2187 6 0
1201 1171 1109 1221 1202 1194 1294 1098 1186 63 5
80 138 60 98 65 138 67 57 62 55 66 101 64 48 25 94 61 91 16 36 26 39
530 427 441 408 434 427 486 363 439 50 11
40 31 32 32 28 30 29 31 32 4 12
58 40 51 57 46 51 38 38 47 8 17
0 19 6 21 21 20 3 19 14 9 66
2047 1847 1842 1863 1848 1889 1962 1667 1871 109 6
-12.4 -12.3 -11.7 -11.8 -12.5 -12.7 -12.3 -12.3 -10.9 -12.1 0.6 4.7
7.3 7.3 7.2 7.5 7.4 7.4 7.4 7.5 7.4 0.1 1.2
1076 1070 1073 1128 1107 1079 1105 25 2
332 329 353 331 625 331 356 272 366 108 29
12 11 14 16 13 20 12 16 14 3 21
400 291 249 259 37 342 240 307 266 107 40
100 62 13 53 66 64 57 63 60 24 40
6 6 6 5 6 6 5 6 6 0 9
106 112 114 103 98 132 77 84 103 17 17
40 40 59 44 42 44 59 56 48 8 18
996 -8.7 -69.3 850 -8.3 -68.7 806 -8.5 -66.8 810 -8.9 -67.2 886 -9.2 -67.8 938 -10.9 -75.4 804 -8.6 -66.6 803 -8.1 -65.4 862 -8.9 -68.4 73 0.9 3.1 8 9.6 4.5
7.8 7.7 7.2 7.0 7.3 7.3 7.2 7.4 7.3 7.4 0.3 3.4
516 505 516 514 513 514 514 1 0
234 221 235 244 244 244 234 254 231 238 10 4
36 39 33 35 36 33 36 35 37 36 2 5
15 28 25 29 24 25 23 26 18 24 4 19
49 8 62 8 44 6 50 6 38 5 41 6 42 5 39 1 45 5 46 6 7 2 16 35
35 37 39 39 48 30 39 36 41 38 5 13
16 13 17 16 17 24 18 26 16 18 4 23
394 408 399 418 413 404 397 417 393 405 10 2
-12.4 -12.3 -11.9 -11.7 -12.5 -11.9 -12.0 -12.3 -11.3 -12.0 0.4 3.1
-87.2 -88.2 -92.3 -85.3 -85.3 -83.7 -90.5 -85.1 -89.1 -87.4 2.8 3.3
2.10 ± 1.75 ± 1.15 ± 0.55 ± 1.40 ± 0.05 ± 0.75 ± 1.50 ± 0.60 ± 1.09 0.66 60.02
1.65 1.55 1.65 1.50 1.75 0.60 2.10 1.80 1.80
7.6 7.3 7.4 7.8 7.4 7.5 7.5 7.5 0.2 2.1
636 505 609 602 611 626 613 12 2
351 340 330 311 345 321 320 331 15 4
9 21 12 9 11 8 11 11 4 38
31 38 46 41 34 35 32 37 5 15
17 2 13 2 11 1 28 2 14 2 12 4 15 2 16 2 6 1 39 42
84 98 45 66 93 66 97 79 20 26
24 17 47 20 18 27 11 23 12 50
517 528 491 478 515 474 488 499 21 4
-10.7 -11.8 -10.2 -9.1 -9.5 -9.7 -9.6 -10.1 0.9 9.0
-78.3 -73.8 -70.8 -66.3 -82.1 -75.8 -74.4 -74.5 5.1 6.8
9.65 ± 10.55 ± 12.50 ± 10.00 ± 10.45 ± 12.25 ± 10.45 ± 10.84 1.10 10.15
1.80 1.90 1.90 0.95 2.25 2.10 2.15
pH
EC
6.7 6.9 6.5 6.4 6.6 6.8 6.5 6.4 6.5 6.6 0.2 2.6
dD
3
H
-92.3 0.00 ± -91.4 0.85 ± -90.0 0.00 ± -88.4 0.80 ± -86.5 0.00 ± -85.8 0.40 ± -89.5 0.00 ± -91.1 0.00 ± -86.0 1.65 ± -89.0 0.41 2.5 0.58 2.8 141.94 8.40 ± 10.20 ± 10.20 ± 11.20 ± 9.70 ± 8.30 ± 9.20 ± 10.30 ± 9.69 1.00 10.33
1.60 1.50 1.60 1.55 1.75 0.60 2.00 1.80 1.95
1.80 1.90 1.90 1.80 1.00 2.25 2.00 2.00
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
27
Table 2 (Contd.) Locality
T
Go¨zlek-7a 27/3/02 38.4 11/7/02 38.3 31/10/02 39.3 10/4/03 39.6 12/7/03 38.9 18/10/03 38.8 15/4/04 38.8 25/6/04 39.3 14/10/04 40.1 average 39.1 r 0.6 VC% 1.5 Go¨zlek-7b 31/10/02 16.1 10/4/03 16.3 12/7/03 17.2 18/10/03 19.2 15/4/04 15.8 25/6/04 20.5 14/10/04 19.6 average 17.8 r 1.9 VC% 10.7 Res¸ adiye-8a 27/3/02 41.3 12/7/02 41.1 average 41.2 r 0.1 VC% 0.3 Res¸ adiye-8b 27/3/02 12.5 12/7/02 17.4 31/10/02 15.9 average 15.3 r 2.5 VC% 16.4 Kurs¸ unlu-9a 28/3/02 55.9 13/7/02 55.5 01/11/02 57.4 11/4/03 57.4 12/7/03 57.9 19/10/03 58.1 16/4/04 56.0 26/6/04 58.2 15/10/04 59.9 average 57.4
pH
EC
HCO3
Cl
SO4
Na
K
Ca Mg
TDS
d18O
dD
3
H
7.9 8.0 7.6 7.4 8.0 7.8 7.8 7.7 7.7 7.8 0.2 2.3
497 488 497 496 494 495 495 1 0
251 248 238 226 258 245 250 267 256 249 12 5
14 13 14 17 15 14 16 14 15 15 1 8
5 40 37 35 35 34 30 33 28 31 10 33
78 89 67 78 68 73 77 92 71 77 9 11
6 6 4 5 4 4 4 3 4 5 1 21
17 18 22 18 32 17 19 19 19 20 5 23
5 6 11 6 9 12 9 5 10 8 3 34
376 420 393 384 420 400 405 432 404 404 18 4
-13.3 -13.0 -12.5 -13.4 -12.3 -13.4 -12.8 -13.0 -11.8 -12.8 0.6 4.3
-95.5 -95.3 -96.7 -92.8 -91.3 -90.1 -98.3 -94.2 -93.9 -94.2 2.6 2.7
7.9 7.5 7.5 7.8 7.8 7.7 7.7 7.7 0.2 2.0
1063 1157 1255 1293 1242 1304 1280 33 3
363 440 429 411 438 336 403 43 11
43 54 65 64 54 60 57 8 14
228 200 200 248 212 290 230 35 15
47 50 38 53 45 56 48 6 13
4 4 3 4 4 4 4 0 10
80 110 68 131 81 133 101 28 28
70 64 95 58 85 65 73 14 20
834 921 898 969 919 945 914 46 5
-10.7 -10.1 -9.9 -9.9 -10.5 -9.6 -9.4 -10.0 0.5 4.7
-80.1 11.20 ± 1.80 -74.2 9.00 ± 1.80 -75.2 9.96 ± 1.85 -74.9 9.50 ± 0.95 -82.2 7.95 ± 2.15 -76.7 11.40 ± 2.05 -73.6 10.45 ± 2.00 -76.7 9.92 3.3 1.23 4.3 12.35
786 40 821 195 804 118 25 110 3 93
760 840 800 57 7
2 55 9 125 9 46 7 75 4 43 60 57
20 42 25 29 12 40
6.4 6.4 6.4 0.0 0.3
- 1805 - 1829 - 1817 17 1
7.8 6.5 7.6 7.3 0.7 9.4
481 481 -
180 238 235 218 32 15
7.2 7.1 6.9 6.9 7.1 7.1 6.6 7.0 7.0 7.0
11300 11100 11030 11270 11070 11140 11160
6100 4906 5612 6002 6809 6184 5984 5942
49 280 85 54 328 92 52 304 89 4 34 5 7 11 6
2.10 ± 2.45 ± 0.55 ± 1.40 ± 2.12 ± 0.25 ± 0.00 ± 1.80 ± 0.55 ± 1.25 0.92 73.88
1.65 1.65 1.60 1.55 1.65 0.60 1.90 1.80 1.80
3805 -12.9 -93.3 4160 -12.6 -92.5 3982 -12.7 -92.9 251 0.2 0.6 6 1.9 0.6
3.05 ± 1.70 2.10 ± 1.65 2.58 0.67 26.09
19 30 23 24 5 21
320 494 381 398 88 22
-12.7 -11.6 -11.9 -12.1 0.6 4.7
-86.1 -84.5 -82.2 -84.3 1.9 2.3
9.20 ± 1.80 9.25 ± 1.70 7.55 ± 1.80 8.67 0.97 11.16
- - 756 52 2650 44 73 28 773 46 2090 219 113 11 797 58 2445 220 39 22 762 103 2478 225 52 58 813 108 2490 832 108 21 758 48 2630 241 15 31 767 52 2450 47 68 0 775 67 2462 261 67 24
9703 8158 9192 9680 11181 9907 9368 9598
-8.5 -8.5 -8.4 -8.7 -8.5 -8.6 -8.8 -8.6 -7.5 -8.5
-88.5 -88.5 -87.2 -87.5 -89.3 -90.6 -94.9 -84.2 -83.8 -88.3
0.35 ± 0.00 ± 0.00 ± 1.35 ± 0.00 ± 0.40 ± 0.00 ± 0.00 ± 0.26
3 3 2 3 1 22
40 48 41 43 4 10
1.70 1.55 1.65 1.60 1.60 0.60 1.95 1.80
S. Su¨er et al.
28
Pure appl. geophys.,
Table 2 (Contd.) Locality
T
r 1.4 VC% 2.4 Kurs¸ unlu-9b 28/3/02 12.6 13/7/02 14.7 01/11/02 15.3 11/4/03 10.6 average 13.3 r 2.1 VC% 16.1 Kurs¸ unlu-9c 28/3/02 9 13/7/02 14.9 01/11/02 13.3 11/4/03 11.9 12/7/03 12.5 19/10/03 13.4 16/4/04 11.0 26/6/04 14.1 15/10/04 14.5 average 13.2 r 1.3 VC% 10.1
pH
EC
0.2 2.6
101 1
HCO3 Cl SO4
Na
K
581 10
22 27 3 40
184 266 7 102
6.5 6.4 6.5 2660 1476 6.5 2440 1429 6.4 2550 1452 0.0 156 34 0.2 6 2
85 33 82 37 84 35 2 3 2 8
458 431 445 19 4
8.2 7.8 8.1 7.7 7.9 8.0 8.0 8.0 8.1 8.0 0.2 1.9
22 2 7 6 4 8 5 9 8 6 78
416 406 427 437 474 455 455 19 4
241 240 247 256 250 271 275 245 253 13 5
7 5 10 11 12 14 6 9 9 3 33
8 13 7 6 31 6 5 9 11 9 80
Ca
Mg
TDS
d18O
36 53
18 74
905 9
0.4 4.7
dD
66 72 64 62 57 75 66 65 66 6 9
8 4 10 15 3 11 15 12 10 5 48
353 337 346 355 356 387 372 349 357 16 4
-10.4 -10.0 -10.0 -9.8 -9.3 -10.3 -9.0 -9.8 -9.5 -9.8 0.4 4.4
H
3.3 0.47 3.8 179.48
- -11.3 -86.3 - -9.7 -83.5 43 71 32 2199 -10.2 -77.7 45 130 1 2154 -10.7 -77.9 44 101 16 2177 -10.5 -81.3 1 42 22 32 0.7 4.2 2 42 136 1 6.5 5.2 1 1 1 0 1 1 0 0 1 0 60
3
-76.9 -81.5 -77.8 -74.9 -75.0 -78.6 -79.5 -73.5 -78.8 -77.4 2.6 3.3
7.25 ± 9.85 ± 9.65 ± 11.85 ± 9.65 1.88 19.52
1.80 1.70 1.80 1.90
15.70 ± 13.75 ± 13.25 ± 14.75 ± 13.85 ± 14.25 ± 10.80 ± 14.45 ± 13.85 1.44 10.36
2.00 1.90 1.85 2.00 2.05 1.10 2.10 2.05
Temperature in C, Electrical Conductivity (EC) in mmho/cm, major ion concentrations and total dissolved solid contents in mg/l. d18O and dD expressed in % with respect to V-SMOW. The analytical errors associated with d18O and dD are 0.1 % and 1 % , respectively. Tritium concentrations given as Tritium Units (TU) together with the associated analytical errors. r is the standard deviation, VC is the variation coefficient (= {standard deviation / mean} x 100)
The bicarbonate character of most waters seems to be compatible with the dissolution of reservoir rocks that are dominated by Mesozoic limestones, whereas the dominancy of Na cation in the hot waters can be attributed to ion exchange (of the waters) with the overlying sediments, including the impermeable clayey levels. As an exception to the dominantly bicarbonate nature of the NAFZ waters, the Yalova thermal waters display sulphate character. In general, sulphate-rich waters can originate from either i) dissolution of evaporites containing gypsum (as previously suggested by EISENLOHR (1995) for the sulphate waters of Armutlu located in close proximity to Yalova), or ii) oxidation of sulphide-bearing minerals (contained in Mesozoic-Paleozoic rocks). Alternatively, the sulphate character of the Yalova thermal waters might be genetically connected to young organic accumulations in Izmit Bay (located to the north of Yalova) and may reflect the release of sulphur (entrapped in these organic-rich sediments) induced by proximal seismic activity.
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
29
Figure 2 a) Pie diagrams of the hot water samples, b) Pie diagrams of the cold water samples (average data of all sampling periods are taken, TDS contents are not reflected in the pie diameters of the diagrams).
5.2. Stable Isotope Compositions The oxygen- and hydrogen-isotope compositions of the waters are presented in the d18O vs. dD diagram in Figure 3, where the data plotted represent the mean values of the measurements throughout the monitoring period. As can be seen from the figure, both hot and cold waters lie close to the Global Meteoric Water Line (CRAIG, 1961). This observation reveals that the hot waters, as well as the cold waters, are overwhelmingly meteoric in origin. However, the hot water sample (no. 9a) from the Kurs¸ unlu field has a
30
S. Su¨er et al.
Pure appl. geophys.,
Figure 3 d18O vs. dD diagram of the sampled waters (data represent the average of nine sampling periods, GMWL: Global Meteoric Water Line (CRAIG, 1961), MMWL: Mediterranean Meteoric Water Line (GAT and CARMI, 1970)).
relatively high d18O value which we interpret as reflecting the effects of (i) water-rock exchange, and/or (ii) calcite scaling in the production well from which the sample was collected. The spatial distributions of the oxygen- and hydrogen- isotope compositions of the water samples are depicted in Figures 4 and 5 as d18O vs. locality and dD vs. locality diagrams, respectively. As observed, the hot waters in almost all fields show more negative (lower) values of d18O and dD than the cold waters, and this situation persists for all sampling periods. This is a rather interesting feature as hot waters might be predicted to have higher d18O values than cold waters owing to intense water-rock interaction at high temperatures. The only explanation for the low values recorded for the hot waters in this study seems to be their recharge characteristics which must occur at higher altitudes (compared to the altitudes where cold water aquifers are recharged). However, owing to the low number of cold waters sampled in the study, no attempt has been made here to estimate the recharge altitudes of the various waters: this would require sampling of several cold waters discharging from varying elevations.
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
31
Figure 4 d18O vs. locality diagrams for all sampling periods (filled symbols show hot, empty symbols show cold waters, gray-filled symbol shows the mineral water sample of Kurs¸ unlu field).
32
S. Su¨er et al. Pure appl. geophys.,
Figure 5 dD vs. locality diagrams for all sampling periods (symbols are same as Fig. 4).
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
33
5.3. Tritium Contents The spatial distribution of the tritium measurements are shown in Figure 6 as plots of tritium concentrations (expressed as Tritium units, TU) vs. sampling localities. Inspection of Figure 6 reveals that, for each field, the tritium concentrations of the hot waters are lower (majority < 5 TU) than those of the cold waters (majority > 8 TU). Given that (i) the level of tritium in the atmosphere was 5–25 TU in the 1950s, while it increased to about 2900 TU in 1963 following the start of nuclear testing in 1960 (IAEA, 1992), and ii) tritium has a half life of 12.26 years, the rather low tritium concentration levels of the hot waters in this study suggest that their aquifers were recharged by precipitation exceeding approximatly 50 years. The cold water aquifers, on the other hand, appear to have been recharged with a component of relatively younger precipitation.
6. Discussion Temporal variations were recorded in both the chemical and isotopic compositions of the water samples as well as in their temperature, pH and EC values during the course of the monitoring program. 6.1. Temperature and pH Variations Examination of Table 2 reveals a temporal variation in temperature, with the amplitude of fluctuations (expressed by variation coefficient, VC) ranging between 7.4 (Seben) and 36.7 (Bolu) for the cold waters, and between 0.3 (Res¸ adiye) and 6.4 (Yalova) for the hot waters. The observation that fluctuations are higher in the cold waters points to the fact that the waters from shallow aquifers either reflect admixture with recentlyrecharged waters or that they have been considerably more affected by seasonal changes in the ambient temperature. In contrast, the pH variations appear to be site-dependent rather than affected by seasonal effects: In some localities the amount of fluctuations in the hot waters is higher than those in their cold water companions, whereas, at other localities, the situation is reversed (Table 2). The highest pH variation in the hot waters was recorded in the Efteni field (8% of the mean), and the lowest variation was recorded in the Res¸ adiye field (0.3% of the mean) where the variation in the cold waters is the highest (9.4% of the mean) among all the fields. 6.2. Chemical Variations As regards chemical compositions, an important feature to note is that the extent of observed variation is essentially dependent on the levels of ionic concentrations (the lower the concentration, the higher the variation coefficient). In this respect, the variations are higher in cold waters due to their lower TDS contents. However,
34
S. Su¨er et al. Pure appl. geophys.,
Figure 6 Tritium concentrations vs. locality diagrams for all sampling periods (symbols are same as Fig. 4).
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
35
considerable fluctuations are also observed in the hot waters of Efteni and Yalova which are located along the western-central segment of the NAFZ. The most striking features, in this respect, can be summarized as follows: 6.2.1. Efteni geothermal field. In the July 2002 sampling period, there is about 1.3-fold decrease in the Ca content (compared to the mean of all the periods) for the Efteni hot water. This decrease is accompanied by an increase in the Cl content and pH value, the latter reaching its maximum in the October 2002 period (Fig. 7). Given that any pH increase in waters is accompanied by a decrease in calcite dissolution, and that other studies in progress on Efteni waters reveal synchronous increases in d13C values and CO2/3He ratios of the geothermal fluids (DE LEEUW et al., in prep.), these variations likely reflect disturbance in the hydrothermal system via degassing. This leads to (i) an increase in the CO2/3He ratio as well as the d13C values of the residual waters (as He and 12C partition into the gas phase in preference to CO2 and 13C, respectively) and (ii) an increase in pH-inducing calcite precipitation and hence a decrease in the Ca content. The occurrence of travertine deposition around the Efteni hot-spring is consistent with these observations. The recorded variations appear to correlate with the 14 July, 2002 Yıg˘ılca-Du¨zce (M 3.1) and the 15 July, 2002 Yıg˘ılca-Du¨zce (M 2.8) earthquakes (Fig. 7). In this respect, degassing in the hydrothermal system might have been induced by a decrease in regional stress/strain levels related to the afore-mentioned earthquakes. In April 2003, the pH value at Efteni seems to have returned to the mean value observed prior to the July 2002 events: This was accompanied by an increase in Ca content. The July 2003 and June 2004 periods at Efteni also deserve attention as there seems to be about 0.5-fold decrease in Ca concentration paralleling the two-fold increase in the Mg content of the hot water (compared to the mean values). Given that Mg is the dominant cation of the cold water in the Efteni field (Fig. 2 and Table 2), the variations observed in these 2 periods suggest mixing between hot waters and shallow, cold waters. It is important to note that (i) the period of mixing in July 2003 appears to follow the July 1, 2003 Kaynas¸ lı-Du¨zce (M 3.2), and/or precede the July 25, 2003 Yıg˘ılca-Du¨zce (M 3.1 and 4.0) earthquakes, and (ii) the mixing in June 2004 correlates with the June 15, 2004 Yıg˘ılca Du¨zce (M 3.6), June 23, 2004 Bolu (M 4.6) and/or the July 2, 2004 Yıg˘ılca-Du¨zce (M 4.8) earthquakes (Fig. 7). Interestingly, however, no significant change was recorded in the ionic concentrations (and/or in temperature-pH) of the Efteni water prior to or following the nearby Bolu earthquake on April 14, 2004 (M 4.7). 6.2.2. Yalova geothermal resort. For the Yalova hot water, compared to the mean of all the sampling periods, there is about a 0.7-fold decrease in Cl and accompanying 1.2-fold increase in SO4 contents in the July 2002 period (Fig. 7). Since Cl is considered to be a conservative constituent of the hot waters, the decrease in the Cl content points to a hot-cold water mixing process, and seems to be correlated with the seismic activity which occurred on the 3rd and/or 13th of July, 2002 in Armutlu-Yalova (M 3.1). Given that the
36
S. Su¨er et al.
Pure appl. geophys.,
Figure 7 Temporal variations of some selected parameters for the Efteni and Yalova hot waters (solid vertical lines: timing of the nearby earthquakes, Apr-I(04) and Apr-II(04) stand for sampling campaigns before and after the 14/4/2004 Bolu earthquake).
sampling date in Yalova was the 9th of July, it is difficult to decide, however, whether the above mentioned Cl decrease is the precursor of the 13th of July, 2002 seismic activity or if it is the post-earthquake response to the 3rd of July, 2002 activity. Regarding the increase in SO4, however, mixing cannot be a viable mechanism as the dominant anion of the cold waters in the Yalova field is HCO3. This increase may be due to the possible release of sulphur (entrapped in the organic sediments) induced by the seismic activity of this period.
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
37
Another important variation in the Yalova hot water concerns the drastic increase in the Mg content in the October 2004 period. Although the sampling date in this period (12/10/ 2004) correlates with the October 10th, 2004 C¸ınarcık-Yalova earthquake (M 3.0), there is no other significant variation recorded in other parameters. This leads to the suspicion that the variation may be an analytical artifact unrelated to any seismic activity. 6.3. Tritium Variations As in the case of chemical variations, the variation coefficient is higher in those waters which have lower levels of tritium concentration. In this respect, hot waters display considerably higher tritium concentration variations than cold waters. The variations in tritium concentrations, however, require particular caution due to high analytical errors (about ± 2 TU) resulting from the high background level in the counter. Nevertheless, significant variations beyond the limits of analytical error are observed in the Bolu, Mudurnu and Yalova fields, in the March and July 2002 sampling periods (Fig. 8). Although the anomalies in Bolu (March 2002) and Mudurnu (March and July 2002) appear to correlate with the March 23rd, 2002 Sea of Marmara (M 4.7), July 14th, 2002 Yıg˘ılca-Bolu (M 3.1), and the July 15th, 2002 Yıg˘ılca-Du¨zce (M 2.8) earthquakes, there is no concomitant variation, in these periods, in the chemical composition of the hot waters in the relevant fields. At Yalova, in the July 2002 period, a significant tritium increase (outside the limits of analytical errors) was recorded (Fig. 8) accompanying the decrease in Cl content (Fig. 7) and supporting the idea of hot-cold water mixing (section 6.2) that can be correlated with the 3rd and/or the13th of July 2002 Armutlu-Yalova earthquake (M 3.1). Apart from July 2002, a high tritium content in the Yalova hot water sample was also detected in March 2002. Although not coupled with any significant variation in anion-cation contents, this
Figure 8 Temporal variations of tritium contents (TU) for the Yalova, Bolu and Mudurnu hot waters (Apr-I(04) and Apr-II(04) stand for sampling campaigns before and after the 14/4/2004 Bolu earthquake).
38
S. Su¨er et al.
Pure appl. geophys.,
tritium anomaly correlates with the seismic activity that occurred in the Sea of Marmara (M 4.7) on the same day as sampling (23 March, 2002). It is particularly important to note that in March 2002, an anomalous value of +5.79% was recorded in the Yalova geothermal field in d13C values of dissolved CO2—this value contrasted with significantly lower (< 0%) values, for all other sampling periods (DE LEEUW et al., in prep.). In the April 2004 sampling period, a significant increase was detected in the TU contents of Efteni (outside the limits of analytical error, 4-fold increase) and Bolu (within the limits of analytical error) in the second sampling campaign performed following the April 14, 2004 earthquake in Bolu (M 4.7). At Mudurnu, on the other hand, an apparent increase in the tritium content was detected on the same day as the April 14, 2004 Bolu earthquake. Although the afore-mentioned fields are located along different strands of the NAFZ (Mudurnu located on the southern strand, and Efteni and Bolu fields located on the northern strand), their response to the seismic activity was the same. Therefore, the increase observed in the tritium contents of these fields can reflect an immediate response probably related to a hot-cold water mixing process triggered by the earthquake in Bolu. 6.4. d18O – dD Variations Temporal variations in d18O and dD values are less than 5% of the mean for hot waters, and up to 10% of the mean for cold waters. This probably reflects the effects of seasonal variations (in the amount of precipitation and/or evaporation) which are more pronounced in cold waters emanating from shallow aquifers. The most prominent variations in hot waters are observed in the Bolu field where a nearly 1.5-fold increase
Figure 9 Temporal variations of d18O and dD values for the Efteni and Bolu waters (Apr-I(04) and Apr-II(04) stand for sampling campaigns before and after the 14/4/2004 Bolu earthquake).
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
39
was recorded in the October 2002 period for both d18O and dD values (Fig. 9). One possible means of producing variations in the d18O and dD values of waters is isotopic exchange with CO2 and H2S gases, respectively, as has been previously proposed by BALDERER et al. (2002) for the increase in d18O and dD values of Kuzuluk mineral water and the Bursa thermal spring. Given that the isotopic variations in Bolu in October 2002 correlate with the October 21st, 2002 Bolu earthquake (M 2.6), seismicity-induced gas emissions could be a possible consequence. In fact, in the Bolu field during the October 2002 period, there seems to be a slight decrease in pH and increase in HCO3 values (as expected from CO2 dissolution in water) supporting the link between CO2 emission and observed d18O variation. Regarding the dD variation, on the other hand, although it was not monitored, any possible H2S gas release is anticipated to increase the SO4 content of the waters. During the October 2002 period, however, about a two-fold decrease was recorded in SO4 content of the Bolu hot water compared to the preceding period. Another alternative means to increase d18O and dD values is evaporation. Such a process could have resulted from possible adiabatic boiling associated with a pressure decrease due to either (i) a disturbance in the regional stress/strain distribution (possibly seismicity-induced), or (ii) excessive pumping in the well from which the sample was collected. This latter suggestion is difficult to evaluate, however, as pumping records are unavailable. The Efteni field also deserves attention in regard to d18O – dD variations: There is a slight (about 3–4% above the mean value) decrease in dD values of both hot and cold waters in the April 2004 period. The sampling dates in this field are 12 and 14 April, 2004, where the second sampling was performed immediately after the seismic activity which occurred on April 14, 2004. Considered in this framework, dD values appear to have started decreasing before the earthquake and increased to their original values after the earthquake. However, no accompanying variations were recorded in d18O values, and it is difficult to assess the mechanism responsible for dD variations.
7. Conclusions
1. The geothermal waters along the NAFZ are mostly Na-HCO3 in character with the exceptions of Na-SO4 type waters (at Yalova) and Ca-HCO3 type waters (at Bolu and Mudurnu). In contrast, the cold waters are mostly of Ca-HCO3 type. While the dominant HCO3 character in the hot and the cold waters seems to be compatible with the dissolution of reservoir rocks that are dominated by Mesozoic limestones, ion exchange with the overlying sediments is probably responsible for the dominancy of Na cation in the hot waters. 2. The oxygen-and hydrogen- isotope compositions suggest a meteoric origin for both hot and cold waters. The high d18O value observed in Kurs¸ unlu hot water, as an exception, is probably related to more extensive water-rock interaction and/or the
40
S. Su¨er et al.
Pure appl. geophys.,
scaling problem observed in the production well from which samples were collected, either higher recharge altitudes for the aquifers of the former compared to the latter, or different climatic conditions during infiltration. 3. Tritium contents of the cold waters are higher than those of the hot waters, revealing that the cold water aquifers are recharged by more recent precipitation. The low TU contents of the hot waters, on the other hand, suggest deep circulating hot waters with long residence times (> 50 years). 4. The monitoring program covering a total of nine sampling periods over three years has revealed temporal variations in both chemical and isotopic compositions of the geothermal waters. Some of these variations appear to correlate with seismic activity occurring close in time and space to the sampling sites. These variations probably reflect the effects of disturbances in the hydrologic/hydrogeochemical system that might have been induced by changes in the regional stress/strain distribution. In this respect, Cl, tritium and Ca appear to be the most sensitive geochemical parameters, and Yalova and Efteni are the key localities for further monitoring studies. 5. Continuing monitoring of water compositions, coupled with gas chemistry (e.g., CO2/ He, and d13C) should lead to a better understanding of their relationship to seismic activities.
Acknowledgements _ ¨ BITAK This study was supported by TU (YDABAG-100Y097) and NSF (Grant no. ¨ zdemir and Hu¨seyin Sendir for EAR-0229508) projects. We would like to thank Yavuz O ¨ zcan Eyu¨pog˘lu for her assistance their support during the field work, and Sabahat O during tritium analyses. We are grateful for helpful comments and constructive reviews by N. Pe´rez and W. Balderer, which improved our manuscript.
REFERENCES _¸ leri Genel Mu¨du¨rlu¨g˘u¨, Yayın No. ALTAY, T. and C¸IFTER, C. (1996), Trityum laboratuvarı el kitabı. Devlet Su Is IZ-903 (in Turkish). _ ¸ ınarcık-Izmit BALDERER, W., LEUENBERGER, F., SUNER, F., YALC¸ıN, T., and STICHLER, W. (2002), Effects of C August 17, 1999 earthquake on the composition of thermal and mineral waters as revealed by chemical and isotope investigation, Geofisica Internacional 41 (4), 385–391. BARKA, A.A. (1992), The North Anatolian Fault Zone, Annales Tecton. 6, 164–195. BOLOGNESI, L. (2000), Earthquake-induced variations in the composition of the water in the geothermal reservoir at Vulcano Island, Italy, J. Volcanol. Geoth. Res. 99, 139–150. BOZKURT, E. (2001), Neotectonics of Turkey - A Synthesis, Geodinamica Acta 14, 3–30. BRA¨UER, K., KA¨MPF, H., STRAUCH, G., and WEISE, S.M. (2003), Isotopic evidence (3He/4He, d13C(CO2)) of fluidtriggered intraplate seismicity, J. Geophy. Res. 108 (B2), 2070, doi:10.1029/2002JB002077. CANIK, B. (1972), Du¨zce-Derdin hamamı civarının jeoloji ve hidrojeoloji etu¨du¨, MTA Report 6359, Ankara (in Turkish).
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
41
¨ ZAKSOY, V., and ERKMEN, C. (2000), Turkish earthquakes reveal C ¸ EMEN, I., Go¨KTEN, E., VAROL, B., KıLıc¸, R., O dynamics of fracturing along a major strike-slip fault zone, EOS, Transact. Am. Geophys. Union, 81/28. CLAESSON, L., SKELTON, A., GRAHAM, C., DIETL, C., MO¨RTH, M., Torssander, P., and Kockum, I. (2004), Hydrogeochemical changes before and after a major earthquake, Geology 32, 641–644. COLEMAN, M.L., SHEPPARD, T.J., DURHAM, J.J., ROUSE, J.E., and MOORE, G.R. (1982), Reduction of water with zinc for hydrogen isotope analysis, Anal. Chem. 54, 993–995. CRAIG, H. (1961), Isotopic variations in meteoric waters, Science 133, 1702–1703. DAS, N.K., BHANDARI, R.K., GHOSE, D., SEN, P., and SINHA B. (2005), Anomalous fluctuation of radon, gamma dose and helium emanating from thermal spring prior to earthquake, Curr. Sci. 89, 1399–1404. DE LEEUW, G.A.M., HILTON, D.R., GU¨LEc¸, N., and MUTLU, H., He-C relationships in geothermal waters from the North Anatolian Fault Zone (NAFZ), Turkey (in prep). DEWEY, J.F. and S¸ ENGo¨R, A.M.C. (1979), Aegean Sea and surrounding regions: Complex multiplate and continuum tectonics in a convergent zone, Geol. Soc. Am. Bull. Part I, 90, 84–92. EISENLOHR, Th. (1995), Die Thermalwa¨sser der Armutlu-Halbinsel (NW-Tu¨rkei) und deren Beziehung zu Geologie und aktiver Tektonik, Diss. ETH-Zu¨rich, Nr. 11340 (1995). EPSTEIN, S. and MAYEDA, T. (1953), Variation of 18O content of waters from natural sources, Geochim. et Cosmochim. Acta 4 (5), 213–224. ERZENOG˘LU, Z. (1989), Amasya Gu¨mu¨s¸ hacıko¨y Hamamo¨zu¨ (Arkutbey) kaplıcası H-1 no’lu sıcaksu sondajı kuyu bitirme raporu, MTA Report 8714, Ankara (in Turkish). FAVARA, R., ITALIANO, F., and MARTINELLI, G. (2001), Earthquake-induced chemical changes in the thermal waters of the Umbria region during the 1997–1998 seismic swarm, Terra Nova 13, 227–233. GAT, J.R. and CARMI, I. (1970), Evolution in the isotopic composition of atmospheric waters in the Mediterranean Sea area, J. Geophys. Res. 75, 3039–3048. HAUKSSON, E. (1981), Radon content of ground water as an earthquake precursor: Evaluation of worldwide data and physics basis, J. Geophys. Res. 86, 9397–9410. HILTON, D.R. (1996), The helium and carbon isotope systematics of a continental geothermal system: Results from monitoring studies at Long Valley Caldera (California, U.S.A.), Chem. Geol. 127, 269–295. IAEA (1992), Statistical treatment of data on environmental isotopes in precipitation. Tech. Rep. Series 331, Vienna. IGARASHI, G., SAEKI, T., TAKAHATA, N., SUMIKAWA, K., Tasaka, S., SASAKI, Y., and SANO, Y. (1995), Groundwater radon anomaly before the Kobe earthquake in Japan, Science 269, 60–61. IGARASHI, G. and WAKITA, H. (1995), Geochemical and hydrological observations for earthquake prediction in Japan, J. Phys. Earth 43, 585–598. ITALIANO, F. and MARTINELLI, G. (2001), Anomalies of mantle-derived helium during the 1997–1998 seismic swarm of Umbria-Marche, Italy, Geophysi. Res. Lett. 28, 839–842. ITALIANO, F., MARTINELLI, G., and RIZZO, A. (2004), Geochemical evidence of seismogenic-induced anomalies in the dissolved gases of thermal waters: A case study of Umbria (Central Apennines, Italy) both during and after the 1997–1998 seismic swarm, Geochem. Geophys. Geosyst. 5, 11, doi:10.1029/2004GC000720. KAWABE, I. (1984), Anomalous changes of CH4/Ar ratio insubsurface gas bubbles as seismo-geochemical precursors at Matsuyama, Japan, Pure Appl. Geophys. 122, 194–214. KING, C.-Y. (1986), Gas geochemistry applied to earthquake prediction: An overview, J. Geophys. Res. 91, 12269–12281. KING, C.-Y., KOIZUMI, N., and KITAGAWA, Y. (1995), Hydro-geochemical anomalies and the 1995 Kobe earthquake, Science 269, 38–39. KING, C.-Y., and IGARASHI, G. (2002), Earthquake-related hydrologic and geochemical changes, International Handbook of Earthquake and Engineering Seismology 81A, 637–645. KING, C.-Y., ZHANG, W., and ZHANG, Z. (2006), Earthquake-induced groundwater and gas changes. Puer Appl. Geophys. Topical Volume Terrestrial Fluids, Earthquakes and Volcanoes: The Hiroshi Wakita Volume I, 633– 646. ¸ avundur (Kurs¸ unlu-C ¸ ankırı) kaplıca bo¨lgesinin hidrojeoloji etu¨du¨ raporu, MTA Report KOc¸AK, A. (1974), C 6373, Ankara (in Turkish). ¨ ZACAR, A., and TEKSO¨Z, B. (1999a), Neotectonic framework of Turkey: A KOc¸YIG˘IT, A., BOZKURT, E., CIHAN, M., O special emphasis on the 17 August 1999 Go¨lcu¨k- Arifiye earthquake (NE Marmara, Turkey), Proc. Internat. Con. Earthquake Hazard and Risk in the Mediterranean Region, Near East University, Cyprus.
42
S. Su¨er et al.
Pure appl. geophys.,
¨ ZACAR, A., and TEKSO¨Z, B. (1999b), 12 Kasım 1999 Dag˘dibi (Du¨zceKOc¸YIG˘IT, A., BOZKURT, E., CIHAN, M., O Bolu) depremi, Geological Reconnaissance Report, Department of Geological Engineering, Middle East Technical University, Ankara (in Turkish). MCKENZIE, D.P. (1972), Active tectonics of the Mediterranean region, Geophys. J. R. Astron. Soc 30, 109–185. MTA (1996), Geothermal inventory of Turkey, General Directorate of Turkish Mineral Research and Exploration (MTA), Ankara, 480 pp (in Turkish). MU¨FTU¨OG˘LU, E. and AKıNCı, K. (1989), Bolu-Seben Keseno¨zu¨ kaplıcası ve c¸evresi jeotermal enerji aramaları sismik etu¨d raporu, MTA Report 8856, Ankara (in Turkish). NISHIZAWA, S., IGARASHI, G., SANO, Y., SHOTO, E., TSAKA, S., and SASAKI, Y. (1998), Radon, Cl and SO4 anomalies in hot-spring water associated with the 1995 earthquake swarm off the east coast of the Izu Peninsula, Central Japan, Appl Geochem. 13, 89–94. O’NEIL, J.R. and KING, C.-Y. (1981), Variations in stable-isotope ratios of ground waters in seismically active regions of California, Geophys. Res. Lett. 8, 429–432. ¨ ZCAN, M. and U ¨ NAY, T. (1978), Bolu-Seben-Keseno¨zu¨ kaplıcası ve c¸evresi jeotermal enerji aramaları rezistivite O etu¨du¨ raporu, MTA Report 8269, Ankara (in Turkish). PADRO´N, E., MELIA´N, G., MARRERO, R., NOLASCO, D., BARRANCOS, J., HERNA´NDEZ, P.A., and PE´REZ, N.M. (2007), Changes on diffuse CO2 emission and relation to seismic activity in and around El Hierro, Canary Islands, Spain. Pure Appl. Geophys. Topical Volume, Terrestrial Fluids, Earthquakes and Volcanoes: The Hiroshi Wakita Volume II, in press. PE´REZ, N.M., SALAZAR, J.M.L., IGARASHI, G., WAKITA, H., and NAKAI, S. (1998), Precursory geochemical signatures of recent earthquakes in Galicia, Spain, Internat Workshop on Frontiers in Monitoring Science and Technology for Earthquake Environments, Tono Geoscience Center, JNC, Japan, November, 16–19, 1998. PE´REZ, N.M., SALAZAR, J.M.L., HERNA´NDEZ, P.A., MELIA´N, G., HERRERA, M., and STRAUCH, W. (2003), Precursory diffuse CO2 emission signature of the 2002 magnitude 4.2 Managua earthquake, Nicaragua, 8th Workshop on Volcanic Gases-CCVG 2003 Abstracts, 58. PE´REZ, N.M. and HERNA´NDEZ, P.A. (2007), Earthquake forecasting research in active volcanic areas by means of diffuse CO2 emission studies. In Geochemical Precursors for Earthquakes, MacMillan Advan. Res. Ser. 94–103. S¸ AHINCI, A. (1970), Mudurnu kaplıcası hidrojeoloji etu¨d raporu, MTA Report 4367, Ankara (in Turkish). SALAZAR, J.M.L., PE´REZ, N.M., HERNA´NDEZ, P.A., SORIANO, T., BARAHONA, F., OLMOS, R., CARTAGENA, R., LO´PEZ, D., LIMA, N., MELIA´N, G., GALINDO, I., PADRO´N, E., SUMINO, H., and NOTSU, K. (2002), Precursory diffuse carbon dioxide degassing signature related to a 5.1 magnitude earthquake In El Salvador, Central America, Earth Plan. Sci. Lett. 205, 81–89. SANO, Y., NAKAMURA, Y., WAKITA, H., NOTSU, K., and KOBAYASHI, Y. (1986), 3He/4He ratio anomalies associated with the 1984 western Nagano earthquake: Possibly induced by a diapiric magma, J. Geophys. Res. 91 / B12, 12291–12295. SANO, Y., TAKAHATA, N., IGARASHI, G., KOIZUMI, N., and STURCHIO, N.C. (1998), Helium degassing related to the Kobe earthquake, Chem. Geol. 150, 171–179. SATO, M., SUTTON, A.J., MCGEE, K.A., and RUSSEL-ROBINSON, S. (1986), Monitoring of hydrogen along the San Andreas and Calaveras faults in Central California in 1980–1984, J. Geophys. Res. 91 (B12), 12315–12326. S¸ ENGO¨R, A.M.C., GO¨RU¨R, N., and S¸ AROG˘LU, F. (1985), Strike-slip faulting and related basin formation in zones of tectonic escape: Turkey as a case study, The Society of Economic Paleontologists and Mineralogists, Special Publication No. 37, 227–264. SILVER, P. and WAKITA, H. (1996), A search for earthquake precursors, Science 273, 77–78. SONG, S.R., KU, W.Y., CHEN, Y.L., LIU, C.M., CHEN, H.F., CHA, P.S., CHEN, Y.G., YANG, T.F., CHEN, C.H., LIU, T.K., and LEE, M. (2006), Hydrogeochemical anomalies in the springs of the Chiayi area in west-central Taiwan as possible precursors to earthquakes, Pure Appl. Geophys. 163, 675–691. SOREY, M.L., KENNEDY, B.M., EVANS, W.C., FARRAR, C.D., and SUEMNICHT, G.A. (1993), Helium isotope and gas discharge variations associated with crustal unrest in Long Valley Caldera, California, 1989–1992, J. Geophys. Res. 98 (B9), 15871–15889. SUGISAKI, R. (1978), Changing He/Ar and N2/Ar ratio of fault air may be earthquake precursors, Nature 275, 209–211. SUGISAKI, R., and SUGIURA, T. (1986), Gas anomalies at three mineral springs and a fumarole before an inland earthquake, Central Japan, J. Geophys. Res. 91 (B12), 12296–12304.
Vol. 165, 2008
Monitoring of Geothermal Waters along NAFZ, Turkey
43
SUGISAKI, R., ITO, K., NAGAMINE, K., and KAWABE, I. (1996), Gas geochemical changes at mineral springs associated with the 1995 southern Hyogo earthquake (M = 7,2), Japan, Earth Plan. Sci. Lett. 139, 239–249. TATAR, O., PIPER, J.D.A., PARK, R.G., and GU¨RSOY, H. (1996), Paleomagnetic study of block rotations in the Niksar overlap region of the North Anatolian Fault Zone, Central Turkey, Tectonophysics 244, 251–266. TENG, T. L., and SUN, L.F. (1986), Research on groundwater radon as a fluid phase precursor to earthquakes, J. Geophys. Res. 91 (B12), 12305–12313. THOMAS, D. (1988), Geochemical precursors to seismic activity, Pure Appl. Geophys. 126, 241–266. TOUTAIN, J.P., and BAUBRON, J.C. (1999), Gas geochemistry and seismotectonics: A Review, Tectonophysics 304, 1–27. TOUTAIN, J.P., MUNOZ, M., POITRASSON, F., and LIENARD, A.C. (1997), Spring water chloride ion anomaly prior to a M = 5.2 Pyrenean earthquake, Earth Plan. Sci. Lett. 149, 113–119. TSUNOGAI, U. and WAKITA, H. (1995), Precursory chemical changes in groundwater: Kobe earthquake, Japan, Science 269, 61–63. VIRK, H.S., and SINGH, B. (1993), Radon anomalies in soil-gas and groundwater as earthquake precursor phenomenon, Tectonophysics 227, 215–224. VIRK, H.S., WALIA, V., and KUMAR, N. (2001), Helium/Radon precursory anomalies of Chamoli earthquake, Garhwal Himalaya, India, J. Geodyn. 31, 201–210. WAKITA, H., NAKAMURA, Y., KITA, J., FUJII, N., and NOTSU, K. (1980), Hydrogen release: New indication of fault activity, Science 210, 188–190. WAKITA, H., IGARASHI, G., NAKAMURA, Y., and NOTSU, K. (1989), Coseismic radon changes in groundwater, Geophys. Res. Lett. 16, 417–420. WAKITA, H. (1996), Geochemical challenge to earthquake prediction, Proc. Natl. Acad. Sci., USA 93, 3781– 3786. WALIA, V., VIRK, H.S., and BAJWA, B.S. (2006), Radon precursory signals for some earthquakes of magnitude > 5 occurred in NW Himalaya: an overview, Pure and Appl. Geophys. Topical Volume, Terrestrial Fluids, Earthquakes and Volcanoes: The Hiroshi Wakita Volume I, 711–722. ZHANG, W. (1994), Research on hydrogeochemical precursors of earthquakes, J. Earth Prediction Res. 3, 170– 182. (Received October 2, 2007, revised October 11, 2007, accepted October 17, 2007)
To access this journal online: www.birkhauser.ch/pageoph
Pure appl. geophys. 165 (2008) 45–61 0033–4553/08/010045–17 DOI 10.1007/s00024-007-0288-2
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
Coupling Between Seismic Activity and Hydrogeochemistry at the Shillong Plateau, Northeastern India ALASDAIR SKELTON,1 LILLEMOR CLAESSON,1 GOVINDA CHAKRAPANI,2 CHANDAN MAHANTA,3 JOYANTO ROUTH,1 MAGNUS MO¨RTH,1 and PARAM KHANNA4
Abstract—Transient hydrogeochemical anomalies were detected in a granite-hosted aquifer, which is located at a depth of 110 m, north of the Shillong Plateau, Assam, India, where groundwater chemistry is mainly buffered by feldspar alteration to kaolinite. Their onsets preceded moderate earthquakes on December 9, 2004 (MW = 5.3) and February 15, 2005 (MW = 5.0), respectively, 206 and 213 km from the aquifer. The ratios [Na+K]/Si, Na/K and [Na+K]/Ca, conductivity, alkalinity and chloride concentration began increasing 3–5 weeks before the MW = 5.3 earthquake. By comparison with field, experimental and theoretical studies, we interpret a transient switchover between source aquifers, which induced an influx of groundwater from a second aquifer, where groundwater chemistry was dominantly buffered by the alteration of feldspar to smectite. This could have occurred in response to fracturing of a hydrological barrier. The ratio Ba/Sr began decreasing 3–6 days before the MW = 5.0 earthquake. We interpret a transient switchover to anorthite dissolution caused by exposure of fresh plagioclase to groundwater interaction. This could have been induced by microfracturing, locally within the main aquifer. By comparison with experimental studies of feldspar dissolution, we interpret that hydrogeochemical recovery was facilitated by groundwater interaction and clay mineralization, which could have been coupled with fracture sealing. The coincidence in timing of these two hydrogeochemical events with the only two MW C 5 earthquakes in the study area argues in favor of cause-and-effect seismichydrogeochemical coupling. However, reasons for ambiguity include the lack of similar hydrogeochemical anomalies coupled with smaller seismic events near the monitoring station, the >200 km length scale of inferred seismic-hydrogeochemical coupling, and the potential for far-field effects related to the Great Sumatra– Andaman Islands Earthquake of December 26, 2004. Key words: Hydrogeochemistry, seismic-hydrogeochemical coupling, water-rock interaction, Shillong Plateau, India.
1. Introduction Changes in groundwater chemistry have been reported before and after several earthquakes (ULOMOV and MAVASHEV, 1971; IGARASHI et al., 1995; TSUNOGAI and WAKITA, 1995; CLAESSON et al., 2004; SILVER and WAKITA, 1996). These include pre-seismic 1 2 3 4
Department of Geology and Geochemistry, Stockholm University, 106 91 Stockholm, Sweden. Department of Earth Sciences, Indian Institute of Technology, Roorkee 247667, India. Department of Civil Engineering, Indian Institute of Technology, Guwahati 781039, India. Wadia Institute of Himalayan Geology, Dehra Dun, India.
46
A. Skelton et al.
Pure appl. geophys.,
changes in the concentrations of Rn, CO2, CH4, N2, H2, He, Cl, F, SO4, Fe, Cr, Mn, Zn and Cu and post-sesimic changes in the concentrations of B, Ca, K, Li, Mo, Na, Rb, S, Si, Sr, Cl, and SO4 (CLAESSON et al., 2004). Such hydrogeochemical changes are often detected far (>100 km) from the earthquake focus (SILVER and WAKITA, 1996). These have been related to a range of processes, but the most widely-accepted are (1) the exposure of fresh mineral surfaces to groundwater interaction and (2) the mixing or switching of source aquifers, induced by pre-seismic damage (SCHOLZ et al., 1973; THOMAS, 1988). However, the mechanism whereby this can occur so far from an earthquake’s focus remains poorly understood (WYSS, 1997; GELLER et al., 1997). Here, we combine (1) hydrogeochemical monitoring, (2) petrographic observation of the granite, which hosts the aquifer, (3) reaction stoichiometry and (4) theoretical considerations, imposed by the critical earthquake model (BUFE and VARNES, 1993; SORNETTE and SAMMIS, 1995; JAUMe´ and SYKES, 1999; JOHANSEN et al., 2000; Zo¨LLER and HAINZL, 2002), to test the hypothesis that hydrogeochemical changes detected in a granite-hosted aquifer, near the Shillong Plateau, northeastern India, were coupled with the only two MW C 5 earthquakes in the region during two years of sampling.
2. Study Area The Brahamaputra valley and the northern Bengal Basin, which flank the 1.6–2 km elevated Shillong Plateau in northeastern India are both densely populated and strongly affected by earthquakes (Fig. 1). The region, which is subjected to frequent M C 5 earthquakes, suffered major loss of life and infrastructural damage caused by the great 1897 Assam earthquake. The Shillong Plateau is interpreted as a ‘pop-up’ structure, which is bounded to the south by a northward-dipping reverse fault (the Dawki fault), and to the north by an inferred buried southward-dipping reverse fault (the Oldham fault) (BILHAM and ENGLAND, 2001; MITRA et al., 2005).
3. Hydrogeochemical Monitoring From October 1, 2004 to December 1, 2005, 171 groundwater samples were collected, on average every 2–3 days at a commercial bottling plant (Silver Drop). The groundwater source is a granite-hosted aquifer at a depth of 110 m located at latitude 26 120 24.6600 and longitude 91 410 27.2800 . Groundwater was pumped from this aquifer at a constant rate of 5000 liters/hour during operational hours (daytime only). The mean pH and temperature of the groundwater, reported at the plant, was 7.7 and 25C, respectively. This location is on the northern flank of the Shillong Plateau and close to the subterranean culmination of the inferred Oldham fault. We extended this time series to include the period December 2003 to September 2004, by analyzing 10 groundwater samples, which had been bottled at the plant for commercial purposes. Groundwater samples were shipped to Stockholm
Vol. 165, 2008
Coupling Between Seismic Activity and Hydrogeochemistry
47
Figure 1 Location map (modified from SRIVASTAVA and SINHA, 2004) showing major tectonic features, regional geology, the location of the hydrogeochemical monitoring station (open triangle), north of the Shillong plateau, and the epicenters of the MW = 5.3 earthquake on December 9, 2004 (206 km from Silver Drop), the MW = 5.0 earthquake on February 15, 2005 (213 km from Silver Drop), the mb = 3.9 on January 12, 2004 (29 km from Silver Drop) and the mb = 4.2 September 27, 2004 (52 km from Silver Drop).
University and analyzed for pH, conductivity, alkalinity (by titration), cations: Al, As, B, Ba, Be, Ca, Cd, Ce, Co, Cr, Cu, Eu, Fe, K , La, Li, Mg, Mn, Mo, Na, Ni, P, Pb, Rb, Sc, Si, Sr, Ti, V, Y, Yb, Zn and Zr (using an Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES, Varian Vista Pro Ax) and anions: Cl and SO4 (by ion chromatography, Dionex DX-300). Ion balance errors, computed using PHREEQCI were less than 5% for >70% of the data and less than 10% for >97% of the data. Two years of data were collected to rule out hydrogeochemical variation related to seasonal rainfall variation in the region. This is relatively systematic, ranging from a monthly average of
48
A. Skelton et al.
Pure appl. geophys.,
4 mm in December to 239 mm in July. Unfortunately, we were unable to obtain local meteorological data.
4. Seismic Activity During the study, two MW C 5.0 earthquakes, listed in the HARVARD CMT catalogue (http://www.globalcmt.org/), occurred within a 2 9 2 rectangle, centered on Silver Drop (Fig. 1). These earthquakes occurred south of the Shillong Plateau. On December 9, 2004, a MW = 5.3 earthquake occurred at latitude 24 390 3600 and longitude 92 430 1200 , and, on February 15, 2005, a MW = 5.0 earthquake occurred at latitude 24 310 1200 and longitude 92 360 3600 . The distances between the epicenters of these earthquakes and the sampled aquifer were 206 km and 213 km, respectively. In this study, we aim to test the hypothesis that hydrogeochemical changes detected Silver Drop (see below) were coupled with these two earthquakes. Before proceeding, we must therefore consider both larger events occurring outside our ‘‘study area’’ and smaller events occurring near to Silver Drop. The epicenters of the next nearest MW C 5 and MW C 6 earthquakes listed in the HARVARD CMT catalogue, which occurred during the study were 366 km (MW = 5.2 on March 25, 2005) and 904 km (MW = 6.0 on March 27, 2004) from Silver Drop, respectively. These events were probably too distant to have generated crustal strain at Silver Drop, similar to that generated by the MW C 5 earthquakes on December 9, 2004 and February 15, 2005. No MW C 7 earthquakes occurred within 2000 km of Silver Drop during the study. However, similar crustal strains could have been generated both by the MW = 9.0 (HARVARD CMT) Sumatra–Andaman Islands earthquake on December 26, 2004 (2585 km from Silver Drop) (cf. HILL et al., 1993), and by two of the 26 smaller events (mb < 5) listed for the study area (Table 1) in the NEIC/PDE catalogue (http://earthquake.usgs.gov/regional/neic/). These events occurred on January 12, 2004 (mb = 3.9), 29 km from Silver Drop, and on September 27, 2004 (mb = 4.2), 52 km from Silver Drop.
5. Results Only Al, Ba, Ca, K, Mg, Na, Si, Sr, SO4 and Cl occurred in concentrations that significantly exceeded the limits of analytical detection. There was no temporal variation in the concentrations of Si (34.24 ± 0.05 ppm), Mg (5.84 ± 0.02 ppm), Al (6.5 ± 1.2 ppb) and SO4 (2.28 ± 0.04 ppm). Si data are plotted in Figure 2a. pH remained approximately constant at 7.71 ± 0.05. In November 2004 (3–5 weeks before the MW = 5.3 earthquake), the concentrations of Na and K began increasing from their respective baseline values of 15.7 ± 0.3 ppm and 1.32 ± 0.01 ppm. Respective maxima of 28–37 ppm and 1.47–1.48 ppm were reached in
Vol. 165, 2008
Coupling Between Seismic Activity and Hydrogeochemistry
49
Table 1 Earthquakes listed in the NEIC/PDE (http://earthquake.usgs.gov/regional/neic/) and HARVARD CMT (http:// www.globalcmt.org/) catalogues for a 2 9 2 rectangle centered on Silver Drop. Body-wave magnitudes (mb) are given for entries from the NEIC/PDE catalogue (normal text) and moment magnitudes (MW) are given for entries from the HARVARD CMT catalogue (italic text). The final entry is for the Great Sumatra–Andaman Islands Earthquake. Entries for the December 9, 2004 and February 15, 2005 earthquakes for both NEIC/PDE and HARVARD CMT catalogues are listed (bold text) Date
Longitude
Latitude
Depth
mb/MW
Distance
Catalogue
2/12/03 6/12/03 12/1/04 12/5/04 13/7/04 4/8/04 9/8/04 27/9/04 2/11/04 5/11/04 12/11/04 24/11/04 7/12/04 9/12/04 9/12/04 21/1/05 21/1/05 15/2/05 15/2/05 27/2/05 11/3/05 3/5/05 29/5/05 24/6/05 17/7/05 12/9/05 11/11/05 12/12/05 27/12/05 31/12/05 26/12/04
90.39 90.25 91.9 93.44 92.78 90.26 91.8 92.15 92.58 93.71 92.08 90.94 92.43 92.54 92.72 92.66 92.72 92.52 92.61 91.55 90.49 91.06 92.42 93.14 93.39 90.53 93.04 92.31 93.99 90.38 94.26
25.79 25.6 26.36 25.23 26.21 25.92 27.58 26.13 26.44 24.09 27.27 27.33 24.41 24.76 24.66 27.42 27.44 24.55 24.52 25.37 27.34 25.76 26.93 26.48 26.41 25.88 25.46 25.96 24.78 25.73 3.09
23 26 38 57 71 61 16 37 48 56 45 10 72 34 39.4 52 55 35 27.2 24 52 33 56 54 26 35 51 35 86 35 28.6
3.8 4.4 3.9 4.5 4.2 4.2 4.1 4.2 4.2 4.7 4 4 4.3 5.5 5.3 3.5 3.7 5.1 5 4.2 4.4 4.3 4.3 4.3 4.8 4.2 3.7 4.3 4.7 4.2 9.0
152 174 29 223 121 162 153 52 102 325 126 150 216 186 206 173 178 206 213 94 183 86 114 164 190 134 171 74 301 155 2585
NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE HARVARD CMT NEIC/PDE NEIC/PDE NEIC/PDE HARVARD CMT NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE NEIC/PDE HARVARD CMT
January 2005 and the concentrations of Na and K returned to their respective baseline values thereafter (Figs. 2b, c). Similar behavior was seen for alkalinity, conductivity and Cl concentration (Figs. 2g–i). Between February 9 (six days before the MW = 5.0 earthquake) and February 17, 2005 (two days after the earthquake), the concentrations of Ca and Sr increased and the concentration of Ba decreased from respective baseline values of 18.6 ± 0.5 ppm, 90.7 ± 0.4 ppb and 20.4 ± 0.2 ppb, to respective maxima/minima of 28.8 ppm, 138 ppb and 7.4 ppb (Figs. 2d–f). Their concentrations returned to baseline values thereafter.
50
A. Skelton et al.
Pure appl. geophys.,
Vol. 165, 2008
Coupling Between Seismic Activity and Hydrogeochemistry
51
b
Figure 2 Time series for (a) Na, (b) K, (c) Ca, (d) Sr, (e) Ba, (f) Si, (g) total alkalinity, (h) conductivity and (i) Cl. The timing of (1) the MW = 5.3 earthquake on December 9, 2004 (206 km from Silver Drop) and the MW = 5.0 earthquake on February 15, 2005 (213 km from Silver Drop) are shown by solid vertical lines; (2) the mb = 3.9 on January 12, 2004 (29 km from Silver Drop) and the mb = 4.2 September 27, 2004 (52 km from Silver Drop) are shown by dotted vertical lines; and (3) the MW = 9.0 Sumatra–Andaman Islands earthquake on December 26, 2004 (2585 km from Silver Drop) is shown by a dashed vertical line.
6. Water-rock Interaction GARRELS and MACKENZIE (1967) showed that granite alteration by interaction with groundwater and therefore groundwater chemistry is depth-dependant. They showed that the chemistry of shallow circulating groundwater is buffered by feldspar alteration to kaolinite, whereas deeper groundwater is buffered by feldspar alteration to smectite. To test whether such behavior can explain hydrogeochemical changes observed in this study, we conducted a petrographic study of altered granite which was collected near the bottling plant. This granite contained 46.6 ± 4.5 modal % quartz, 40.6 ± 4.4 modal % microcline with minor lamellae of albite (Ab>99), 11.8 ± 2.9 modal % plagioclase (zoned from An16 to An3) with patches of K-feldspar and 1.0 ± 0.9 modal % biotite. Petrographic examination and electron microprobe analysis of the granite showed that (i) plagioclase is extensively altered to smectite along narrow fracture surfaces, which exhibit preferred crystallographic orientation, and (ii) plagioclase and microcline are locally altered to aggregates of clay minerals (kaolinite + smectite ± gibbsite) and magnetite along wider non-crystallographic fractures and at grain edges (Fig. 3). Quartz is unaltered and biotite shows only minor alteration to clay minerals along cleavage planes. We interpret intracrystalline alteration of feldspar to smectite by groundwater interaction at deeper levels, overprinted by intercrystalline alteration of feldspar to kaolinite by groundwater interaction at shallower levels. This is similar to the depthdependent alteration of feldspar reported by GARRELS and MACKENZIE (1967).
7. Reaction Stoichiometry The alteration of feldspar to kaolinite and smectite caused by interaction with groundwater can occur by the (simplified) reactions: 2ðNa;KÞAlSi3 O8 þ 2CO2 þ 11H2 O ! Al2 Si2 O5 ðOHÞ4 þ 2ðNa;KÞþ þ 2HCO 3 þ 4H4 SiO4 ðalbite;microclineÞ
ðkaoliniteÞ
½1
52
A. Skelton et al.
Pure appl. geophys.,
Figure 3 Thin section image viewed in polarized light (a) showing preferential alteration of plagioclase adjacent to a fracture in granite collected near the bottling plant. Image size: ca. 5 9 3 mm. Back-scattered electron (BSE) image (b) showing alteration of plagioclase to smectite along narrow, sometimes crystallographically-aligned fractures and kaolinite along wider fractures and at grain edges.
2ðNa;KÞAlSi3 O8 þ 2CO2 þ 6H2 O ! Al2 Si4 O10 ðOHÞ2 þ 2ðNa;KÞþ þ 2HCO 3 þ 2H4 SiO4 ðalbite;microclineÞ
ðsmectiteÞ
½2 CaAl2 Si2 O8 þ 2CO2 þ 2H4 SiO4 ! Al2 Si4 O10 ðOHÞ2 þ Ca2þ þ 2HCO 3 þ 2H2 O ðanorthiteÞ
ðsmectiteÞ
½3
The stoichiometry of reactions [1–3] has been simplified by ignoring (1) structurallybound H2O in kaolinite and smectite, and (2) Na, Ca and Mg in smectite. Because reactions [1] and [2] release [Na + K] and H4SiO4 into solution in the molar ratios 1:2 and 1:1, respectively (see GARRELS and MACKENZIE, 1967) and reaction [3] releases Ca, but no
Vol. 165, 2008
Coupling Between Seismic Activity and Hydrogeochemistry
53
H4SiO4 into solution and based on the assumption that feldspar alteration was a dominant control of groundwater chemistry, we studied their time-dependent activities by plotting temporal variations of [Na+K]/Si, Na/K, Ba/Sr and [Na+K]/Ca ratios (Figs. 4a–d). Electron microprobe analysis indicates that smectite in the granite contained small amounts of Na2O (1.2–1.4 wt. %) and CaO (0.8–1.0 wt. %), but negligible MgO. This would reduce the molar ratio of [Na + K] and H4SiO4 released into solution by reaction [2] to * 0.9:1. Feldspar may also have been altered by reaction with hydrochloric acid, but this reaction was volumetrically minor (alkalinity & 20 9 molar Cl), yielding similar stoichiometric ratios for [Na+K]/Si, Na/K, Ba/Sr and [Na+K]/Ca. Finally, the reason for examining the Ba/Sr ratio is that this ratio has been shown to be a particularly sensitive indicator of the relative inputs from K- and Ca-bearing minerals (LAND et al., 2000). Because Ba substitutes for K in microcline and Sr substitutes for Ca in the anorthitic component of plagioclase, the Ba/Sr ratio was used to compare the dissolution rates of microcline by reaction [1] and anorthitic plagioclase by reaction [3].
8. Hydrogeochemical Events
In November 2004 (3–5 weeks before the MW = 5.3 earthquake), the molar [Na+K]/Si ratio increased from its baseline value of 0.58 ± 0.01, reaching a maximum of 1.0–1.4 in January 2005 (Fig. 4a). This is consistent with a switchover between source aquifers, with groundwater in the main aquifer dominantly buffered by reaction [1] and groundwater in the second aquifer dominantly buffered by reaction [2]. This could occur if (for example) the second aquifer was at a deeper level or if the residence time for groundwater in this aquifer was longer. This switchover can be visualized on a plot of molar Na+/Ca2+ against molar HCO–3/H4SiO4 after GARRELS (1967) (Fig. 5). The interpretation of an open system switchover between groundwater aquifers is further supported by the following observations: (1) The concentration maximum for [Na + K] was not mirrored by a concentration minimum for Si (Figs. 2a-c). This would have been expected for a closed system switchover between reactions [1] and [2] (cf., GARRELS and MACKENZIE, 1967). (2) Clay mineral aggregates occupy fractures in plagioclase and microcline (Fig. 3). These could record switchover(s) between chemically-distinct groundwater aquifers during clay mineral growth, but they might alternatively record simultaneous growth of two or more clay minerals. Possible mechanisms which could cause the interpreted switchover between groundwater aquifers include (1) fracturing of a hydrological barrier between two aquifers, and (2) a change in the relative pressures between two aquifers (THOMAS, 1988). The [Na+K]/Si maximum coincided with a Na/K maximum of 30–40 (compared with a baseline of 20.1 ± 0.3, Fig. 4b), which is consistent with experimental and theoretical
54
A. Skelton et al.
Pure appl. geophys.,
Vol. 165, 2008
Coupling Between Seismic Activity and Hydrogeochemistry
55
b
Figure 4 Groundwater molar [Na+K]/Si, Na/K, Ba/Sr and [Na+K]/Ca ratios. The timing of (1) the MW = 5.3 earthquake on December 9, 2004 (206 km from Silver Drop) and the MW = 5.0 earthquake on February 15, 2005 (213 km from Silver Drop) are shown by solid vertical lines; (2) the mb = 3.9 on January 12, 2004 (29 km from Silver Drop) and the mb = 4.2 September 27, 2004 (52 km from Silver Drop) are shown by dotted vertical lines; and (3) the MW = 9.0 Sumatra–Andaman Islands earthquake on December 26, 2004 (2585 km from Silver Drop) is shown by a dashed vertical line. The wide dot/dashed horizontal lines at (Na+K)/Si = 0.5 and 1 dashed lines are the predicted groundwater molar [Na+K]/Si ratio buffered by reactions [1] and [2], respectively. The narrow dotted horizontal lines denote the 2r limits for all data excluding November 2004 – February 2005. The solid curve is the best-fit to the critical earthquake model for equations [4]. Inset shows the molar Ba/Sr ratio in February, 2005 and the MW = 5.0 earthquake.
studies, showing that sodic feldspar dissolves more rapidly than potassic feldspar (BUSENBERG and CLEMENCY, 1976; LASAGA, 1984; STILLINGS and BRANTLEY, 1995). The Cl maximum (Fig. 2i) with no corresponding maximum in SO4 concentration as reported by (e.g.) TSUNOGAI and WAKITA (1995) might relate to the rapidity of SO4 adsorption on clay minerals. In February 2005, the [Na + K]/Si ratio began decreasing. We suggest that this hydrogeochemical ‘‘recovery’’ records effective isolation of the main aquifer from the second aquifer. Possible mechanisms which could cause its effective isolation include (1)
Figure 5 Plot of molar Na+/Ca2+ against molar HCO-3/H4SiO4 after GARRELS (1967). The inferred ‘‘switchover’’ between source aquifers, with groundwater in the main aquifer dominantly buffered by reaction [1] and groundwater in the second aquifer dominantly buffered by reaction [2] is highlighted.
56
A. Skelton et al.
Pure appl. geophys.,
clogging of fractures by newly-formed clay minerals or (2) readjustment of the relative pressures between the aquifers. Renewed buffering of groundwater chemistry by reaction [1] in the main aquifer during or after its isolation could explain the progressive recovery of the [Na + K]/Si ratio towards1:2. Figure 4a shows that recovery is initially rapid, becoming progressively slower until some steady state is approached. This is of interest, because experimental investigations of the artificial weathering of feldspar record sequential nonparabolic, parabolic and linear kinetic stages (e.g., BUSENBERG and CLEMENCY, 1976). Parabolic kinetics was initially attributed to diffusion across a residual layer formed by non-stoichiometric dissolution. However, this was later shown to be an experimental artifact relating to the initial dissolution of fine and ultrafine particles, produced during sample preparation (HOLDEN and BERNER, 1979). However, because production of fine and ultrafine particles may also occur during fracturing, we suggest that this experimental ‘‘artifact’’ might nevertheless provide a reasonable representation of the hydrogeochemical response to fracturing or microfracturing in a natural aquifer. Furthermore, we note that parabolic kinetics could also reflect rate-limiting diffusion across a layer of newly-formed clay minerals. Based on this analysis, we reach the tentative conclusion that hydrogeochemical recovery in the main aquifer at Silver Drop (Fig. 2) is facilitated by progressive fracture healing related to water-rock interaction and consequent clay mineralization, following initial fracturing of a hydrological barrier between two aquifers. Between February 9 (six days before the MW = 5.0 earthquake) and February 17, 2005 (two days after this earthquake), the Ba/Sr ratio decreased from its baseline of 0.144 ± 0.001, to a minimum value of 0.045. We interpret the transient influence of reaction [3] on the chemistry of the sampled groundwater. This increases the release rates of Ca and Sr (Figs. 2d and e). Figure 2f shows a corresponding decrease in the release rate of Ba. We interpret this as closed system behavior: Reaction [3] consumes CO2 from the system, lowering its activity and slowing the forwards progress of reaction [1] and the release rate of Ba (substituting for K). The lack of a corresponding minimum for K (Fig. 2c) reflects its lower sensitivity to water-rock interaction, compared with Ba (LAND et al., 2000). We suggest that reaction [3] was induced by the exposure of unaltered plagioclase to groundwater. This could be caused by local microfracturing in the main aquifer, and would induce a transient stage during which (anorthitic) plagioclase would be preferentially dissolved (BUSENBERG and CLEMENCY, 1976; STILLINGS and BRANTLEY, 1995). Figure 3 shows preferential dissolution of plagioclase in the granite, which hosts the studied aquifer. The Ba/Sr ratio recovered to its baseline by April 2005. We interpret that this recovery occurred because unaltered anorthite in direct contact with groundwater was exhausted (perhaps due to coating by newly-formed clay minerals) and the release of Ca and Sr by reaction [3] ceased. The availability of CO2 and consequently the release rate of Ba increased. We observe a similar pattern, with rapid initial recovery, becoming progressively slower, until some steady state is approached. We conclude that this event involved closed system water-rock interaction within the main aquifer. Both events are seen on Figure 4d, which shows temporal variation of [Na+K]/Ca. In November 2004, the [Na+K]/Ca ratio increased from its baseline value of 1.54 ± 0.03, to
Vol. 165, 2008
Coupling Between Seismic Activity and Hydrogeochemistry
57
a maximum of 2.5–3.8 in January 2005. We interpret that this maximum corresponds to switching between aquifers, resulting in an influx of groundwater from a second aquifer, where groundwater chemistry is buffered by reaction [2]. The [Na+K]/Ca ratio reached a minimum of 0.59 on February 12, 2005. We interpret that this minimum records the transient influence of reaction [3] on the chemistry of groundwater in the main aquifer.
9. Seismic-hydrogeochemical Coupling Coincidence of timing is not proof that a hydrogeochemical shift is caused by an earthquake. Nevertheless, we consider the observation that the respective onsets of two distinct hydrogeochemical shifts preceded the only two MW C 5 earthquakes to occur during two years of sampling, convincing evidence of a cause-and-effect association. As such, it is tempting to relate these hydrogeochemical shifts to pre-seismic damage (causing fracturing in the aquifer region). However, elastic half-space models (OKATA, 1992) indicate that the permanent static strain created by a moderate earthquake (which is a probable upper limit for preseismic strain (JOHNSTON et al., 1987)) will be diminishingly small at a distance of 200 km from its focus. On the other hand, precursory hydrogeochemical and hydrological anomalies are often reported at similar distances from earthquake foci, usually confined within the same fault system (SILVER and WAKITA, 1996; ROELOFFS, 1988). This paradox is addressed by the ‘‘critical earthquake model’’ (BUFE and VARNES, 1993; SORNETTE and SAMMIS, 1995; JAUME´ and SYKES, 1999; JOHANSEN et al., 2000; Zo¨LLER and HAINZL, 2002). In this model, a large earthquake is viewed as the culmination of a sequence of seismic cycles, which occur at increasingly large scales within a volume of crust with lateral dimensions, several times greater than the length of the seismic rupture (see also: JOHANSEN et al., 2000; KNOPOFF et al., 1996; ALLe`GRE et al., 1982). For a moderate earthquake, the average radius of this region may still be less than 200 km (WYSS, 1979). However, we note the existence of a pathway of structural weakness, linking the Dawki and (inferred) Oldham faults (BILHAM and ENGLAND, 2001; MITRA et al., 2005) (Fig. 6), which might extend the affected region (cf., ROELOFFS, 1988). The critical earthquake model predicts that cumulative damage before an earthquake follows a power law of the form (BUFE and VARNES, 1993): D A þ Bðtc tÞz ;
½4
where A and B are empirically-determined constants, 0 < z < 1, t is time and the earthquake occurs at time, tc. D is cumulative damage before the earthquake. Both cumulative earthquake frequency (SORNETTE and SAMMIS, 1995) and groundwater ion concentrations (JOHANSEN et al., 2000) have been used as proxies for D. Here, we use preseismic [Na + K]/Si as a proxy for D and determine its goodness-of-fit for equation [4] (Fig. 4a). The best-fit value of tc was December 9, 2004 (± 1 day) (R2 = 0.78). We have not incorporated the log-periodic corrections introduced by SORNETTE and SAMMIS (1995) for two reasons. Firstly, our dataset is insufficient to support the additional variables
58
A. Skelton et al.
Pure appl. geophys.,
Figure 6 Schematic cross section of the Shillong ‘pop-up’ structure (modified from MITRA et al., 2005). The pathway of structural weakness between the epicenter of the MW C 5 earthquakes which occurred on December 12, 2004 and February 15, 2005, south of the Shillong Plateau and the hydrogeochemical monitoring station, links between the Dawki and Oldham faults along the base of the crust.
required (cf. MAIN et al., 1999). Secondly, the validity of these corrections has been questioned (JAUME´ and SYKES, 1999). The applicability of the critical earthquake model to smaller earthquakes (M < 6.5) has also been questioned (Zo¨LLER and HAINZL, 2002). On the other hand, potentially useful results have been obtained for earthquakes as small as mb = 3.5 (BREHM and BRAILE, 1998). The goodness-of-fit of the critical earthquake model to our data and the accuracy with which tc is estimated supports the interpretation of a causal association between the MW = 5.3 earthquake on December 9, 2004 and the first hydrogeochemical anomaly. This would imply that this anomaly occurred in response to a switchover between source aquifers triggered by pre-seismic fracturing. However, we cannot unequivocally exclude far-field effects related to the Great Sumatra–Andaman Islands Earthquake on December 26, 2004 (cf. HILL et al., 1993). On the other hand, precursory changes 2000 km from the epicenter seem unlikely. The lack of any measurable hydrogeochemical signal coupled with the two earthquakes which occurred nearest to Silver Drop on January 12, 2004 (mb = 3.9) and September 27, 2004 (mb = 4.2) weakens the argument for coupling between seismicity and hydrogeochemistry. The resolution of the Ba/Sr data is insufficient to perform a similar analysis for the second hydrogeochemical anomaly. However, the shorter duration of this anomaly and the closed system hydrogeochemical behavior are consistent with weaker pre-seismic damage associated with the second and smaller earthquake. This would imply that this anomaly occurred in response to limited pre-seismic microfracturing, locally within the main aquifer. On the other hand, inferred coupling between the second hydrogeochemical anomaly and the smaller earthquake on February 15, 2005, raises further doubts as to why other events did not yield measurable hydrogeochemical signals.
Vol. 165, 2008
Coupling Between Seismic Activity and Hydrogeochemistry
59
10. Conclusion We conclude that (1) prior to November 2004, groundwater chemistry in the main aquifer was dominantly buffered by the alteration of feldspar to kaolinite, (2) the first hydrogeochemical anomaly, likely recorded a transient switchover between source aquifers, which induced an influx of groundwater from a second aquifer, where groundwater chemistry was dominantly buffered by the alteration of feldspar to smectite, (3) this could have been coupled with the MW = 5.3 earthquake on December 9, 2004, (4) the second hydrogeochemical anomaly likely recorded a transient switchover to anorthite dissolution induced by exposure of fresh plagioclase to groundwater interaction, (5) this could have been caused by microfracturing in the main aquifer coupled with the MW = 5.0 earthquake on February 15, 2005, and (6) subsequent hydrogeochemical recovery was facilitated by groundwater interaction and clay mineralization, which could have been coupled with post-seismic fracture sealing. The main argument in support of seismic-hydrogeochemical coupling is the coincidence in timing of two hydrogeochemical events with two MW C 5 earthquakes. Reasons for ambiguity include the lack of similar hydrogeochemical anomalies temporally coupled with other seismic events, the >200-km length scale of inferred seismichydrogeochemical coupling, and the potential for far-field effects related to the Great Sumatra–Andaman Islands Earthquake of December 26, 2004. The hydrogeochemical anomalies reported in this study meet some of the validation criteria of the IASPEI (International Association of Seismology and Physics of the Earth’s Interior) subcommission on earthquake prediction (WYSS, 1991, 1997) in that a relation to pre-seismic stress and that some dependence on distance from the earthquake foci is inferred (Table 1). However, hydrogeochemical data was collected from only one site, and even although the hydrogeochemical anomalies are recorded using several instrumental methods the reported anomalies ([Na+K]/Si, Ba/Sr, conductivity, alkalinity) are not truly independent of one another. We thus suggest that to resolve these and similar ambiguities, and to confirm a cause-and-effect association between seismicity and hydrogeochemistry, requires extended time series hydrogeochemical monitoring, conducted simultaneously at several sites, preferably in several seismically-active regions. With respect to site selection, we note that the first hydrogeochemical anomaly detected at Silver Drop caused a measurable change in conductivity and is thus amenable to low cost online monitoring.
Acknowledgements Colin Graham and Heiko Woith are thanked for constructive scientific input to this manuscript. Klara Hajnal and Heike Siegmund are thanked for analytical work. Pedro A. Herna´ndez Pe´rez and an anonymous reviewer are thanked for constructive comments which improved an earlier version of this manuscript. This research was supported by VR SIDA and Carl Tryggers Stiftelse.
60
A. Skelton et al.
Pure appl. geophys.,
REFERENCES ALLE`GRE, C.J., LE MOUEL, J. L., and PROVOST, A. (1982), Scaling rules in rock fracture and possible implications for earthquake prediction, Nature 297, 47–49. BILHAM, R. and ENGLAND, P. (2001), Plateau ‘pop-up’ in the great 1897 Assam earthquake, Nature 410, 806– 809. BUFE, C.G. and VARNES, D. J. (1993), Predictive modeling of the seismic cycle in the Greater San Francisco Bay Region, J. Geophys. Res. 98, 9871–9983. BUSENBERG, E. and CLEMENCY, C.V. (1976), The dissolution kinetics of feldspars at 25C and 1 atm CO2 partial pressure, Geochim. Cosmochim. Acta 40, 41–49. CLAESSON, L., SKELTON, A., GRAHAM, G., DIETL, C., Mo¨RTH, M., TORSSANDER, P., and KOCKUM, I. (2004), Hydrogeochemical changes before and after a major earthquake, Geology 32, 641–644. GARRELS, R.M. Genesis of some ground waters from igneous rocks. In Researches in Geochemistry, Vol. II (ed. P.H. Abelson) (Wiley & Sons, New York 1967), pp. 405–420. GARRELS, R.M. and MACKENZIE, F.T. (1967), Origin of the chemical composition of some springs and lakes, Adv. Chem. Ser. 67, 222–242. GELLER, R.J., JACKSON, D.D., KAGAN, Y.Y., and MULARGIA, F. (1997), Earthquakes cannot be predicted, Science 275, 1616. HILL, D.P. et al. (1993), Seismicity remotely triggered by the magnitude 7.3 Landers, California earthquake, Science 260, 1617–1623. HOLDEN, G.R. and BERNER, R.A. (1979), Mechanisms of feldspar weathering – I. Experimental studies, Geochim. Cosmochim. Acta 43, 1161–1171. IGARASHI, G., SAEKI, S., TAKAHATA, N., SUMIKAWAY, K., TASAKA, S., SASAKI, Y., TAKAHASHI, M., and SANO, Y. (1995), Ground-water radon anomaly before the kobe earthquake in japan, Science 269, 60–61. JAUME´, S.C. and SYKES, L.R. (1999), Evolving towards a critical point: A review of accelerating seismic moment/ energy release prior to large and great earthquakes, J. Appl. Geophys. 155, 279–305. JOHANSEN, A., SALEUR, H., and SORNETTE, D. (2000), New evidence of earthquake precursory phenomena in the 17 January 1995 Kobe earthquake, Japan, Eur. Phys. J. B 15, 551–555. JOHNSTON, M.J.S., LINDE, A.T., GLADWIN, M.T., and BORCHERDT, R.D. (1987), Fault failure with moderate earthquakes, Tectonophys. 144, 189–206. KNOPOFF, L., LEVSHINA, T., KEILIS-BOROK, V. I., and MATTONI, C. (1996), Increased long-range intermediatemagnitude earthquake activity prior to strong earthquakes in California, J. Geophys. Res. 101, 5779–5796. ¨ HLANDER, B. (2000), Ba/Sr, Ca/Sr and 87Sr/86Sr ratios in soil water LAND, M., INGRI, J., ANDERSSON, P.S., and O and groundwater: Implications for relative contributions to stream water discharge, Appl. Geochem. 15, 311– 325. LASAGA, A.C. (1984), Chemical kinetics of water-rock interaction, J. Geophys. Res. 89, 4009–4025. MAIN, I.G., LEONARD, T., PAPASOULIOTIS, O., HATTON, C.G., and MEREDITH, P.G., (1999), One slope or two? Detecting statistically significant breaks of slope in geophysical data, with application to fracture scaling relationships, Geophys. Res. Lett. 26, 2801–2804. MITRA, S., PRIESTLEY, K., BHATTACHARYA, A. K., and Gaur, V.R. (2005), Crustal structure and earthquake focal depths beneath northeastern India and southern Tibet, Geophys. J. Int. 160, 227–248. OKADA, Y. (1992), Internal deformation due to shear and tensile faults in a half-space, Bull. Seismol. Soc. Am. 82, 1018–1040. ROELEFFS, E.A. (1988), Hydrological precursors to earthquakes: A review, Pure Appl. Geophys. 126, 177–209. SCHOLZ, C.H., SYKES, L.R., and AGGARWAL, Y.P. (1973), Earthquake prediction: A Physical Basis, Science 181, 803–810. SILVER, P.G. and WAKITA, H. (1996), A search for earthquake precursors, Science, 273, 77–78. SORNETTE, D. and SAMMIS, C.G. (1995), Complex critical exponents from renormalization group theory of earthquakes: Implications for earthquake predictions, J. Phys. Int. France 5, 607–619. SRIVASTAVA, R.K. and SINHA, A.K. (2004), Early Cretaceous Sung Valley ultramafic-alkaline-carbonatite complex, Shillong Plateau, Northeastern India: Petrological and genetic significance, Mineral. Petrol. 80, 241–263. STILLINGS, L.L. and BRANTLEY, S.L. (1995), Feldspar dissolution at 25C and pH 3: Reaction stoichiometry and the effect of cations, Geochim. Cosmochim. Acta 59, 1483–1496.
Vol. 165, 2008
Coupling Between Seismic Activity and Hydrogeochemistry
61
THOMAS, D. (1988), Geochemical precursors to seismic activity, Pure Appl. Geophys. 126, 241–266. TSUNOGAI, U. and WAKITA, H. (1995), Precursory chemical changes in groundwater: Kobe earthquake, Japan, Science 269, 61–63. ULOMOV, V.I. and MAVASHEV, B.Z. (1971), The Tashkent Earthquake of 26 April, Tashkent, Akad. Nauk Uzbek. SSR, FAN, 188–192. WYSS, M. (1979), Estimating the maximum expectable magnitude of earthquakes from fault dimensions, Geology 7, 336–340. WYSS, M, Evaluation of Proposed Earthquake Precursors (ed. M. Wyss), (AGU, Washington DC, 1991), 94 pp. WYSS, M. (1997), Second round evaluation of proposed earthquake precursors, Pure Appl. Geophys. 149, 3–16. Zo¨LLER, G., and HAINZL, S. (2002), A systematic spatiotemporal test of the critical point hypothesis for large earthquakes, Geophys. Res. Lett. 29, 1558. (Received February 13, 2007, revised November 10, 2007, accepted December 10, 2007) Published Online First: February 1, 2008
To access this journal online: www.birkhauser.ch/pageoph
Pure appl. geophys. 165 (2008) 63–74 0033–4553/08/010063–12 DOI 10.1007/s00024-007-0292-6
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
Radon Changes Associated with the Earthquake Sequence in June 2000 in the South Iceland Seismic Zone ´ STA RUT HJARTARDO´TTIR,1 and PA´LL EINARSSON,1 PA´LL THEODO´RSSON,2 A GUðO´N I GUðJO´NSSON2
Abstract—An earthquake sequence at the transform plate boundary in South Iceland, that included two magnitude 6.5 earthquakes in June 2000, was anticipated on the basis of a centuries-long seismicity pattern in the area. A program of radon monitoring in geothermal water from drill holes, initiated in 1999, rendered distinct and consistent variations in radon in association with these events. All seven sampling stations in a 50 9 30 km zone covering the epicentral area showed a consistent pattern. Four types of change could be identified: 1) Preseismic decrease of radon. Anomalously low values were measured 101–167 days before the earthquakes. 2) Preseismic increase. Spikes appear in the time series at six stations 40–144 days prior to the earthquakes. These anomalies were large and unusual if compared to a 17-years history of radon monitoring in this area. 3) Coseismic step, most likely related to the coseismic change in groundwater pressure observed over the entire area. 4) Postseismic return of the radon values to the preseismic level about three months later, also concurrent with groundwater pressure changes. Key words: South Iceland Seismic Zone, radon, earthquake precursor, co-scismic changes.
1. Introduction Various studies have shown that an increase in the concentration of radon in groundwater is an earthquake precursor, see e.g., reviews by HAUKSSON (1981) and KING (1985), and more recent studies by WAKITA (1996), ROELOFFS (1999), TRIQUE et al. (1999), and ZMAZEK et al. (2002). Even though numerous examples of premonitory radon anomalies have been identified and described in the literature, statistical analysis of the relationship between radon and earthquakes has been difficult because of the lack of long-time series from a network of recording stations in active seismic or volcanic areas. A program of radon monitoring was initiated for this purpose in the plate boundary areas of Iceland in 1977 (HAUKSSON and GODDARD, 1981; HAUKSSON, 1981).
1
Institute of Earth Sciences, University of Iceland, Sturlugata 7, 101 Reykjavı´k, Iceland. E-mail: [email protected] 2 Science Institute, University of Iceland, Dunhaga 3, 107 Reykjavı´k, Iceland.
64
P. Einarsson et al.
Pure appl. geophys.,
A network of 11 sampling stations was established, nine of which were located within or near the South Iceland Seismic Zone; a transform zone where the most damaging historical earthquakes in Iceland have occurred. Judged from the time pattern of previous events a sequence of large earthquakes was expected in this zone with high probability (STEFA´NSSON et al., 1993; EINARSSON et al., 1981). This labor-intensive radon program gave promising results (Jo´NSSON and EINARSSON, 1996) but was discontinued in 1993 because of declining funding and deteriorating instruments. Radon monitoring was resumed in 1999 with new and improved instruments and time-saving sample preparation (THEODo´RSSON, 1996; GUDJo´NSSON and THEODo´RSSON, 2000) during which the sample preparation time was reduced from 3 hours to less than 10 minutes. The expected earthquakes occurred in June 2000 (EINARSSON et al., 2000; STEFa´NSSON et al., 2000) within the network of monitoring stations and after one year of radon monitoring. The earthquake sequence occurred along a 90-km stretch of the plate boundary and contained two magnitude 6.5 (Mx) earthquakes (on June 17 and 21) and several ´ RNADo´TTIR et al., 2004). This paper magnitude 5+ events (see e.g., PEDERSEN et al., 2003; A documents the radon time series and reports on the radon changes detected in association with these events.
2. The South Iceland Seismic Zone The South Iceland Seismic Zone is a transform-type plate boundary; a branch of the mid-Atlantic plate boundary that crosses Iceland (Fig. 1). Plate divergence in the southern part of Iceland is accommodated by two sub-parallel rift zones: the Western and the Eastern Volcanic Zones. The gap between them is bridged in the south, near 64N, by a zone of high seismic activity, the South Iceland Seismic Zone, which takes up the transform motion between the Reykjanes Ridge and the Eastern Volcanic Zone (EINARSSON, 1991). The two rift zones and the transform demarcate a block or a microplate, the Hreppar microplate. It has been argued that rifting is dying out in the Western Rift Zone, and is being taken over by the Eastern Rift Zone, or that the partition of rifting between the rift zones may be uneven and changes with time (SIGMUNDSSON et al., 1995). The South Iceland Seismic Zone has been defined by destruction areas of historical earthquakes, Holocene surface ruptures and instrumentally determined epicenters. It is oriented E-W and is 10–15 km wide. Destruction areas of individual earthquakes and surface faulting (Fig. 2) show, however, that each event is associated with faulting on N-S striking planes, perpendicular to the main zone. The overall left-lateral transform motion along the zone, i.e., between the Hreppar microplate to the north and the Eurasia plate to the south, thus appears to be accommodated by right-lateral faulting on many parallel, transverse faults and counter-clockwise rotation of the blocks between them, ‘‘bookshelf faulting’’ (EINARSSON et al., 1981).
Vol. 165, 2008
Radon Changes and the South Iceland Earthquakes of June 2000
65
Figure 1 The active plate boundary of Iceland passes near the center of the Iceland hotspot. The radon program is centered on the transform zone of South Iceland, SISZ. The Western and Eastern Volcanic Rift Zones are marked with W and E, respectively. Arrows show the direction of relative plate movements across the plate boundaries. The rate is 18.5 to 19.5 mm/year.
Earthquakes in South Iceland tend to occur in major sequences in which most of the zone is affected. These sequences last from a few days to about three years. Each sequence typically begins with a magnitude 7 event in the eastern part of the zone, followed by smaller events farther west. Sequences of this type occurred in 1896, 1784, 1732–34, 1630–33, 1389–91, 1339 and 1294. Apart from the historic gap between 1391 and 1630, the sequences thus occur at intervals that range between 45 and 112 years (EINARSSON et al., 1981), and it has been argued that a complete strain release of the whole zone is accomplished in about 140 years (STEFA´NSSON and HALLDo´RSSON, 1988). The lengthy time since the last sequence led to a long-term forecast published in 1985 (EINARSSON, 1985), later refined by STEFA´NSSON et al. (1993), of a major earthquake sequence within the next decades. The original forecast gave a 80% probability for the occurrence of a major earthquake sequence within the next 25 years (i.e., within the 1985–2010 time window). The magnitude of the first event was estimated in the range 6.3–7.5 and the most likely location was given in the eastern part of the seismic zone. In the refined version the magnitude range was reduced to 6.3–7.0 and two seismic gaps were identified, at 20.3W and 20.7W. The forecast was fulfilled in June 2000 when
66
P. Einarsson et al.
Pure appl. geophys.,
Figure 2 The location of the radon sampling stations within the South Iceland Seismic Zone shown with triangles. Thin lines show Holocene surface fractures, formed in earthquakes before 2000. The two long, thick lines show the source faults of the two earthquakes of June 17 and 21, 2000 as delineated by aftershocks; the former one on the easternmost fault, the latter on the western fault. The sense of faulting was right-lateral strike-slip on both faults. Elevation contours are at 50 m intervals.
two magnitude 6.5 events occurred in the zone; one at 20.37W and the other at 20.71W.
3. The 2000 Earthquake Sequence The earthquakes of 2000 were the largest in the zone since 1896 and 1912. They occurred within the SIL-network of seismographs operated by the Icelandic Meteorological Office (see e.g., website http://www.vedur.is/, STEFA´NSSON et al. 1993). The sequence began on June 17 at 15:40 with a magnitude 6.5 event in the eastern part of the zone (Fig. 2). This immediately triggered a flurry of activity along at least a 90-km-long stretch of the plate boundary to the west, apparently triggered by the passing S waves from the first event. Among them was an event with an anomalously low seismic radiation but a ´ RNADo´TTIR et al., 2004). An earthquake of moment equivalent to a magnitude 5.9 (A magnitude 5.7 (mb) followed two minutes later on a small, parallel fault, about 3–4 km to the west of the first shock. An event of magnitude 4.9 (mb) then occurred on the Reykjanes Peninsula, 90 km to the west, about 5 minutes after the first shock. Several other significant shocks also occurred this day along this segment of the plate boundary (PAGLI et al., 2003). A second mainshock similar in magnitude to the first event occurred about 20 km west of the first one on June 21 at 00:51. It was clearly preceded by a clustering of ´ RNADo´TTIR microearthquakes along the eventual source fault (STEFA´NSSON et al., 2000). A et al. (2003, 2004) present evidence that triggering played a large role in the occurrence of
Vol. 165, 2008
Radon Changes and the South Iceland Earthquakes of June 2000
67
events within the sequence, both dynamic triggering by the passing waves and triggering by regional changes in the Coulomb stress due to faulting. The aftershock distribution, moment tensor inversions, distribution of surface faulting, and modeling of surface deformation measured by GPS and InSAR confirm that the mainshocks of the sequence occurred on N-S striking faults, transverse to the zone itself (CLIFTON and EINARSSON, 2005; ´ RNADo´TTIR et al., 2001; PEDERSEN et al., 2003). The sense of faulting was right-lateral A strike-slip conforming to the model of ‘‘bookshelf faulting’’ for the South Iceland Seismic Zone. The two mainshocks occurred on pre-existing faults and were accompanied by surface ruptures consisting primarily of en-e´chelon tension gashes and push-up structures (CLIFTON and EINARSSON, 2005). The main zones of rupture were about 15 km long and fractured the crust down to 10 km depth. The surface faults coincided with the epicentral distributions of aftershocks. Fault displacements were of the order of 0.1–1 m. Faulting along conjugate, left-lateral strike-slip faults also occurred, but was less pronounced than that of the main rupture zones. The maximum fault displacement at depth, determined by modeling of geodetic data, was 2–2.5 m. The source faults of the two largest earthquakes are shown in Figure 2. Large hydrological changes were observed in a wide area surrounding the seismically active zone. Pressure changes in boreholes followed a regular pattern conforming with crustal stress changes (BJo¨RNSSON et al., 2001; Jo´NSSON et al., 2003). Pressure decreased in areas to the NE and SW of the epicenters but increased in the quadrants to the NW and SE. These changes were large, but were reversed and equilibrated in less than three months. A post-earthquake crustal deformation signal was detected by InSAR that correlates with the water pressure changes (Jo´NSSON et al., 2003).
4. Previous Radon Studies in South Iceland The relationship between radon and earthquakes has been studied in this area since 1977, when the first equipment for this purpose was installed. The instruments were operated until 1993. The radon monitoring network consisted of up to 9 stations. Samples (0.6 l) of geothermal water were collected from drill holes every few weeks and sent to the laboratory for radon analysis. The resulting time series varied in length from 3 to 16 years. Many earthquake-related radon anomalies were identified (Jo´NSSON and EINARSSON, 1996). They are represented by both positive and negative excursions from the mean values, and occur mostly prior to the seismic events, i.e., within a few weeks. For a statistical analysis of the anomalies and comparison with the seismicity time series, significant earthquakes were selected according to the criteria of HAUKSSON and GODDARD (1981): M 2:4 log D 0:43
and
M 2;
ð1Þ
where M is the magnitude and D is the distance to a radon monitoring station. Thus 98 independent seismic events were selected. They were in the magnitude range 2–5.8. The main conclusions were as follows:
68
P. Einarsson et al.
Pure appl. geophys.,
1. 2. 3. 4.
Radon anomalies were observed before 30 of the significant events. 35% of all observed anomalies were related to seismicity. 80% of the anomalies observed before earthquakes were positive. If a positive anomaly is detected at one station, the probability of a significant earthquake occurring afterwards is 38%. 5. Some sampling sites were found to be more sensitive than others. The sensitivity appears to depend on local geological conditions. 6. A few radon anomalies appeared to be related to eruptive activity of the neighboring Hekla volcano.
5. Revival of the Radon Monitoring A new radon program was initiated in 1999 using a new time-saving technique and an instrument developed at our institute. It is based on a novel liquid scintillation technique where counting only Bi-218/Po-218 pulse pairs gives high sensitivity with a simple construction (THEODo´RSSON, 1996; GUDJo´NSSON and THEODo´RSSON, 2000). Scintillations other than those associated with radon-222 are thus excluded. Samples of 200 ml are taken from the geothermal drill holes at the sampling sites (Table 1) and analyzed in the laboratory. About 60% of the radon from the water samples is transferred to a scintillator in a 15 ml liquid scintillation counting vial by circulating and bubbling air for four minutes between the two liquids. The scintillation liquid is mineral oil. The scintillations are subsequently counted in a laboratory-made automatic sample changer. 226 counts per hour correspond to 1 Bq/l of radon in the water. The system represents a significant progress in the radon measuring technique where high sensitivity is needed. The technique also saves considerable time compared to previous procedures. Sample preparation time was reduced from 3 hours to less than 10 minutes. Sampling from geothermal wells in the South Iceland Seismic Zone began a year before the destructive earthquakes of June 2000 occurred. Water samples were taken
Table 1 Location, depth and temperature of geothermal holes sampled for radon Station
Latitude
Longitude
Depth, m
Teperature, C
Bakki ¨ xnalækur O Selfoss, hole 13 Hlemmiskeið Flu´ðir, hole 5 Kalda´rholt Laugaland, hole 3
6356.6 6359.1 6356.8 6400.6 6407.7 6400.2 6355.0
2116.6 2111.3 2057.5 2033.2 2019.5 2028.8 2025.0
886 953 500 85 321 38 1100
120 100 & 85 66 94 62 100
Vol. 165, 2008
Radon Changes and the South Iceland Earthquakes of June 2000
69
about twice a week from geothermal drill holes at seven sites (Fig. 2, Table 1) and sent to the Science Institute, University of Iceland, for analysis. These holes, ranging in depth between 38 m and 1100 m, were mostly the same as used in the earlier radon program. The June 2000 earthquakes originated within our sampling network. The nearest station, Laugaland, is only 2 km away from the source fault of the first earthquake (Fig. 2).
6. The Radon Changes The radon time series for the seven monitoring stations are shown in Figure 3. The raw data are plotted as individual points connected by thin lines. The numbers denote counts of disintegrations per hour in the sample. The thick, broken line shows a sevenpoint running average of the data. The vertical bars give the time of significant earthquakes in the area. The height of the bar reflects the ‘‘excess magnitude’’ of the event, ME, defined on the basis of the criteria of equation (1): ME ¼ M ð2:4 log D 0:43Þ:
ð2Þ
Largest probabilities of radon anomalies are expected prior to events with positive excess magnitude. The scale in the plots is arbitrary and the relative height of the bars is only presumed to reflect the likelihood that the respective earthquake is associated with a radon excursion at that particular station. The radon variations form a distinct pattern that can be related to the earthquakes. A typical behavior is seen at the station Hlemmiskeid. The radon values prior to the earthquakes of June 2000 are relatively stable, varying by less than ± 50% around the mean. The largest deviations are positive, spike-like excursions that occur 59 and 115 days before the earthquakes. These are the highest values measured at this station during the two years of operation. Several of the lowest values are measured in a limited time interval 125–167 days before June 17. A pronounced coseismic step is observed at this station. The mean value drops by about 50% at the time of the earthquakes. About three months later the mean value returns to the pre-earthquake level. The main features of the Hlemmiskeið time series can be identified in the other time ¨ xnalækur series has a stable level in the series as well, except at Kalda´rholt. The O beginning, low values 124–138 days prior to the earthquakes, and a peak value 114 days before them. At Selfoss the same behavior is observed, but subdued. Low values occur 124–142 days before, and high values 114 and 144 days before the events. At Bakki anomalous values are observed, but the background level of radon is very low. Therefore the negative deviations are more difficult to identify. The values are apparently low in the time interval 117–134 days before the earthquakes, and 3 out of the 4 highest values are recorded 57, 50, and 43 days before June 17. At Flu´ðir the negative excursions are not prominent, although a large peak is seen 54 days before the earthquakes. The Laugaland series has well developed negative deviations 101–149
70
P. Einarsson et al.
Pure appl. geophys.,
prior to the events but positive spikes are not seen, except possibly a high value 58 days before the events. The Kalda´rholt time series is quite irregular and does not display the same pattern as the other stations. Kalda´rholt is located quite close to the source fault of the June 17 earthquake. ¨ xnalækur, Flu´ðir, Bakki, and In summary, the preseismic spikes are observed at O ¨ xnalækur, Selfoss, Laugaland, and possibly at Laugaland. The preseismic low is seen at O ¨ xnalækur, Laugaland, and possibly subdued at Bakki. The coseismic drop is observed at O at Selfoss. At Kalda´rholt and Flu´ðir the sampling was discontinued because of the coseismic pressure drop in the geothermal system. The radon values had returned to normal at all stations about three months after the earthquakes. The pressure in the geothermal systems also returned to normal at about this time.
7. Discussion The changes in radon concentration at our stations occur on a regional scale as shown by the similarity between the time series in Figure 3. The distance between the stations Selfoss and Laugaland is 27 km. This argues for a common cause of the changes. A meteorological cause is considered highly unlikely. The samples are taken from deep geothermal boreholes. No evidence for correlation with precipitation has been found. Seasonal effects are also considered unlikely. Radon time series are available for the same boreholes for the time period 1977–1993. Seasonal effects were only seen at one of the stations (Bakki) and only for a part of the observation period (Jo´NSSON, 1994). All the boreholes are producing geothermal holes, used for house heating. The production is at a maximum during the winter months. Therefore, if a seasonal effect is responsible the radon concentration is expected to be low in the winter and high in the summer. This is not consistent with the changes observed in 2000 (Fig. 3). The most robust result of this study is the demonstration of the coseismic drop in radon concentration and its postseismic return to previous values. Because of its temporal correlation with observed changes in groundwater pressure, we suggest that there is a causal relationship between these parameters. We note, however, that the radon concentration at all stations dropped during the earthquakes, regardless of whether they were located within areas of increasing or decreasing water pressure. It would seem that both increasing and decreasing groundwater pressure leads to a decrease in radon flow from the crustal rocks. This may not be as unreasonable as it seems at first sight. BJo¨RNSSON et al. (2001) and Jo´NSSON et al. (2003) point out that the pressure variations show a spatial pattern consistent with stress changes due to the faulting during the earthquakes. Pressure increases in areas where coseismic volumetric compression occurs in the crustal rocks. Pressure decreases where volumetric expansion occurs. The volumetric changes and water pressure are related through the porosity of the rock which in this case is to a large degree due to fractures. High water pressure is thus caused by the closure of cracks. Since the
Vol. 165, 2008
Radon Changes and the South Iceland Earthquakes of June 2000 J
1200
F M
A M
J
J
F M
A M
J
J
A S
O
N D
O
N D
Öxnalækur
900 J
A S
71
600
2000
Hlemmiskeið
300
1500
0
1000
5000
Kaldárholt
4000
500
3000 0
2000 1000
1200
Laugaland
0
900
6000
Flúðir
600
4000 300
2000
0
0
800
Selfoss 600
300
400
200
200
100
Bakki
0 0
50
100
150
200
N day
250
300
350
0 0
50
100
150 200 N day
250
300
350
Figure 3 Time series of radon activity at the seven sampling stations. The activity is given in counts per hour. 226 counts per hour correspond to 1 Bq/l of radon in the water. The data points are connected by a thin line. The dashed lines show a seven-point running average of the data. Vertical bars with stars on top give the time of significant earthquakes in the area. Their heights are proportional to the excess magnitude of the events, as defined by equation (2) and depend on the earthquake magnitude and epicentral distance to the radon station.
release of radon into the water takes place across fracture walls, the closure of fractures leads to reduced radon concentrations in the water. In the areas where coseismic dilatation takes place, increasing pore volume leads to a drop in pressure which inhibits the flow of water out of the rock. This will also lead to a reduced flux of radon out of the crustal rocks. Chemical components or physical parameters other than radon have not been monitored at our stations. However, a multi-parameter approach is highly desirable for a more meaningful interpretation of the causes of the changes as shown by the work of CLAESSON et al. (2004, 2007) in North Iceland.
72
P. Einarsson et al.
Pure appl. geophys.,
8. Conclusions Four different patterns can be identified in the radon time series within the South Iceland Seismic Zone in association with the earthquake sequence of June 2000: 1. Pre-seismic decrease of radon. Anomalously low values were measured in the period 101–167 days before the earthquakes. 2. Preseismic increase. Positive spikes appear in the time series 40–144 days prior to the earthquakes. 3. Coseismic step. The radon values decrease at the time of the first earthquake. This is most likely related to the coseismic change in groundwater pressure observed over the whole area. 4. Postseismic return to preseismic levels about three months after the earthquakes, probably also linked with the pressure equilibration in the geothermal systems. In view of the positive results of the project, we are developing and testing a new, automatic radon instrument, Auto-Radon, based on the same design that continuously monitors the radon concentration in the geothermal groundwater (THEODo´RSSON and GUDJo´NSSON, 2003; Jo´NSSON et al., 2007). The instruments are located at the drill hole stations, measuring radon four times a day.
Acknowledgements The radon programs in South Iceland have been supported by grants from several agencies, including the Icelandic Research Council, the SEISMIS Project, and the European Union under the projects PRENLAB and PREPARED. Numerous persons have participated in this radon project and measurements. We particular like to mention Gı´sli Jo´nsson and the attendants of the sampling sites in South Iceland, Guðlaugur ´ lafur O ´ lafsson, Valdimar Þorsteinsson, Vilhja´lmur Eirı´ksson, Sveinsson, Stefa´n O Guðru´n Magnu´sdo´ttir, Olgeir Engilbertsson, and Hannes Bjarnason.
REFERENCES ´ RNADo´TTIR, Th., HREINSDo´TTIR, S., GUðMUNDSSON, G., EINARSSON, P., HEINERT, M., and Vo¨LKSEN, C. (2001), A Crustal deformation Measured by GPS in the South Iceland Seismic Zone due to two Large Earthquakes in June 2000, Geophys. Res. Lett. 28, 4031–4033. ´ RNADo´TTIR, Th., Jo´NSSON, S., PEDERSEN, R., and GUDMUNDSSON, G. (2003), Coulomb stress changes in the South A Iceland Seismic Zone due to two large earthquakes in June 2000, Geophys. Res. Lett. 30, doi:10.1029/ 2002GL016495, no. 5. ´ RNADo´TTIR, Th., GEIRSSON, H., and EINARSSON, P. (2004), Coseismic stress changes and crustal deformation on A the Reykjanes Peninsula due to triggered earthquakes on June 17, 2000, J. Geophys. Res. 109, B09307, doi:10.1029/2004JB003130.
Vol. 165, 2008
Radon Changes and the South Iceland Earthquakes of June 2000
73
´ . G., SÆMUNDSSON, K., and EINARSSON, E. M. (2001), Pressure changes in Icelandic BJo¨RNSSON, G., FLo´VENZ, O geothermal reservoirs associated with two large earthquakes in June 2000. In Proceedings to Twenty-Sixth Workshop on Geothermal Reservoir Engineering, Stanford University. CLAESSON, L., SKELTON, A., GRAHAM, C., DIETL, C., Mo¨RTH, M., TORSSANDER, P., KOCKUM, I. (2004), Hydrogeochemical changes before and after a major earthquake, Geology 32, 641–644. CLAESSON, L., SKELTON, A., GRAHAM, C., and Mo¨RTH, M. (in review 2007), The timescale and mechanism of fault sealing and water-rock interaction after an earthquake, Geofluids. CLIFTON, A. and EINARSSON, P. (2005), Styles of surface rupture accompanying the June 17 and 21, 2000 earthquakes in the South Iceland Seismic Zone, Tectonophysics 396, 141–159. EINARSSON, P. (1985), Jarðskja´lftaspa´r (Earthquake prediction, in Icelandic with English summary), Na´ttu´rufræðingurinn 55, 9–28. EINARSSON, P. (1991), Earthquakes and present-day tectonism in Iceland, Tectonophysics 189, 261–279. EINARSSON, P., BJo¨RNSSON, S., FOULGER, G., STEFA´NSSON, R., and SKAFTADo´TTIR, Th. (1981), Seismicity pattern in the South Iceland Seismic Zone. In Earthquake Prediction - An International Review (eds. D. Simpson and P. Richards), Am. Geophys. Union, Maurice Ewing Series 4, 141–151. EINARSSON, P., CLIFTON, A., SIGMUNDSSON, F., and SIGBJo¨RNSSON, R. (2000), The South Iceland earthquakes of 2000: Tectonic environment and effects, Am. Geophys. Union, Fall Meeting, San Francisco, EOS 81, 890. GUDJONSSON, G. I. and THEODo´RSSON, P. (2000), A compact automatic low-level liquid scintillation system for Radon in water by pulse pair counting, Appl. Radiation and Isotopes 53, 377–380. HAUKSSON, E. (1981), Radon content of groundwater as an earthquake precursor: evaluation of worldwide data and physical basis, J. Geophys. Res. 86, 9397–9410. HAUKSSON, E. and GODDARD, J. (1981), Radon earthquake precursor studies in Iceland, J. Geophys. Res. 86, 7037–7054. Jo´NSSON, S. (1994), Radonmælingar a´ Suðurlandi (Radon measurements in South Iceland, in Icelandic), University of Iceland, Faculty of Science, BScThesis, 214 pp. Jo´NSSON, S. and EINARSSON, P. (1996), Radon anomalies and earthquakes in the South Iceland Seismic Zone 1977–1993. In Seismology in Europe (ed. Thorkelsson, B. et al.), European Seismol. Commission, Reykjavı´k, pp. 247–252. Jo´NSSON, S., SEGALL, P., PEDERSEN, R., and BJo¨RNSSON, G. (2003), Post-earthquake ground movements correlated to pore-pressure transients, Nature 424, 179–183. JONSSON, G., THEODORSSON, P., and SIGURDSSON, K. (2007), Auto-radon — a new automatic liquid scintillation system for monitoring Radon in water and air. In: Chalupnik S., Schonhofer, F., and Noakes J, eds. LSC 2005, Advances in Liquid Scintillation Spectrometry, in press. KING, C.-Y. (1985), Gas geochemistry applied to earthquake prediction: An overview, J. Geophys. Res. 91, 12,269–12,281. PAGLI, C., PEDERSEN, R., SIGMUNDSSON, F., and FEIGL, K. L. (2003), Triggered Seismicity on June 17, 2000 on the Reykjanes Peninsula, SW-Iceland Captured by Radar Interferometry, Geophys. Res. Lett. 30, 1273, 10.1029/ 2002GL-015310. ´ RNADo´TTIR, Th., SIGMUNDSSON, F., and FEIGL, K. L. (2003), Fault slip distribution of PEDERSEN, R., Jo´NSSON, S., A two June 2000 Mw 6.4 earthquakes in South Iceland estimated from joint inversion of InSAR and GPS measurements, Earth Planet. Sci. Lett. 213, 487–502. ROELOFFS, E. (1999), Radon and rock deformation, Nature 339, 104–105. SIGMUNDSSON, F., EINARSSON, P., BILHAM, R., and STURKELL (1995), Rift-transform kinematics in South Iceland: Deformation from global positioning system measurements, 1986 to 1992, J. Geophys. Res. 100, 6235–6248. STEFa´NSSON, R., BO¨ðVARSSON, R., SLUNGA, R., EINARSSON, P., JAKOBSDo´TTIR, S., BUNGUM, H., GREGERSEN, S., HAVSKOV, J., HJELME, J., and KORHONEN, H. (1993), Earthquake prediction research in the South Iceland Seismic Zone and the SIL Project, Bull. Seismol. Soc. Am. 83, 696–716. STEFa´NSSON, R. and HALLDo´RSSON, P. (1988), Strain release and strain build-up in the South Iceland Seismic Zone, Tectonophysics 152, 267–276. ´ RNADo´TTIR, Th., BJo¨RNSSON, G., GUðMUNDSSON, G. B., HALLDo´RSSON, P. (2000), The two large STEFa´NSSON, R., A earthquakes in the South Iceland Seismic Zone in June 2000. A basis for earthquake prediction research, Am. Geophys. Union, Fall Meeting, San Francisco, EOS 81, 890. THEODo´RSSON, P. (1996), Improved automatic Radon monitoring in groundwater, In Seismology in Europe (eds. Thorkelsson, B. et al.), European Seismological Commission, Reykjavı´k, pp. 253–257.
74
P. Einarsson et al.
Pure appl. geophys.,
THEODo´RSSON, P. and GUDJONSSON, G. I. (2003), A simple and sensitive liquid scintillation counting system for continuous monitoring of Radon in water, Advances in Liquid Scintillation Spectrometry, 249–252. TRIQUE, M., RICHON, P., PERRIER, F., and AVOUAC, J. P. (1999), Radon emanation and electrical potential variations associated with transient deformation near Reservoir Lakes, Nature 399, 137–140. WAKITA, H. (1996), Geochemical challenge to earthquake prediction, Proc. National Acad. of Sciences, USA, Vol. 93, No. 9 (Apr. 30, 1996), pp. 3781–3786. ZMAZEK, B., ITALIANO, F., ZIVCIC, M., VAUPOTIC, J., KOBAL, I., and MARTINELLI, G. (2002), Geochemical monitoring of thermal waters in Slovenia: Relationships to seismic activity, Appl. Radiat. Isot. 57, 919–930. (Received February 15, 2007, revised October 22, 2007, accepted October 24, 2007) Published Online First: February 1, 2008
To access this journal online: www.birkhauser.ch/pageoph
Pure appl. geophys. 165 (2008) 75–94 0033–4553/08/010075–20 DOI 10.1007/s00024-007-0291-7
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
CO2 Degassing over Seismic Areas: The Role of Mechanochemical Production at the Study Case of Central Apennines F. ITALIANO,1 G. MARTINELLI,2 and P. PLESCIA3
Abstract—Field observations coupled with experimental results show that CO2 can be produced by mechanical energy applied to carbonate rocks becoming an unexpected additional gas source besides that degassed from the mantle or produced by thermometamorphism. The evidence that a large amount of carbon dioxide associated with radiogenic-type helium (R/Ra as low as 0.01–0.08) is released through continental areas, denotes the absence of a contribution from the mantle or from mantle-derived fluids. Data collected during the seismic crisis which struck the Central Apennines in 1997–98 have shown an enhanced CO2 flux not associated with the presence of mantle or thermometamorphic-derived fluids. On the other hand, new experimental results highlight the possibility of producing CO2 by mechanical energy that acts on the calcite crystalline lattice. While the CO2 released over the geothermal areas (e.g., Larderello Geothermal Field) is obviously derived by mantlederived activities, this is not the case of the huge amount of CO2 released over the seismically active areas where the presence mantle-derived products is ruled out. We propose that mechanical energy, e.g., released during seismic events, microseismicity or creeping processes is a possible additional energy source able to produce CO2 and thus could explain the presence of CO2 degassing over tectonic areas where the influence of the mantle is low.
1. Introduction Apart from the water vapor that in high energy magmatic and geothermal systems is by far the most abundant gaseous specie, CO2 is the main residual component when temperatures drop and the water vapor condenses, and is the most common gas released from colder systems over tectonically active areas, where CO2-dominated mofettes are the typical gaseous manifestation. CO2 is also globally generated by the decomposition of organic matter or by microbial activity. The areas where CO2 degasses at the global scale have been outlined by IRWIN and BARNES (1980) who listed the recognized CO2-degassing prone areas as related to active tectonics. The active tectonic area of the Central Apennines is characterized by a widely distributed CO2 gas emission which is released both at seismically-active areas and via geothermal activity. A number of hypotheses 1
INGV Istituto Nazionale Geofisica e Vulcanologia, Sezione di Palermo, Italy. E-mail: [email protected] 2 ARPA Environmental Protection Agency of Emilia Romagna, Via Amendola 2, 42100 Reggio Emilia, Italy. 3 CNR Istituto Studio Materiali Nanostrutturati, Montelibretti RM, Italy.
76
F. Italiano et al.
Pure appl. geophys.,
have been proposed about its origin: PANICHI and TONGIORGI (1976) proposed an origin of CO2 from carbonate hydrolysis based upon both carbon isotope composition and geotemperature estimates. GIANELLI (1985) proposed an origin of the geothermal CO2 due to metamorphic process. MARINI and CHIODINI (1994) and CHIODINI et al. (1995) proposed that CO2 discharging in Central Italy originated by mixing of a mantle and a thermometamorphic component, while MINISSALE (2004) suggested that the CO2 is of mantle origin, considering the correlation between the anomalously high CO2 fluxes and the minimum depth of the Moho. Whereas the various postulated generation mechanisms are probably all presently active along the Apennine chain, to a lesser or greater extent, all of them require a thermal energy source which is available from the mantle or from mantle-derived products. The geothermal area of Larderello is located at the western sector of the Central Apennines and the released CO2 is obviously produced by the above-mentioned mechanisms. However, the Central Apennines are well known because of the active tectonic structures that generate moderate earthquakes (e.g., 1694, Me = 6.9; 1857, Me = 7.0; 1915, Me = 7.1; 1962, Me = 6.1; 1980, Me = 6.7, BOSCHI et al., 1995) and release huge amounts of CO2. Since the released CO2 is associated with a helium component marked by a typical radiogenic signature (ITALIANO et al., 2001; MINISSALE, 2004), it is difficult to believe that, although the two areas are close each other, the main source for the released CO2 over those seismic zones, is the same as for the Larderello geothermal area. The absence of geochemical evidences of mantle-derived products implies also that an additional thermal energy source from the mantle acting as the engine for thermometamorphic processes is not available. Following the experimental results about CO2 production, we provide in this paper an explanation for the presence of CO2 in continental seismic areas starting from the observation that: a) the 1997–1998 seismic crisis (Umbria Region; Central Apennines) induced an increased degassing of CO2 (HEINICKE et al., 2000; ITALIANO et al., 2001; CARACAUSI et al., 2005) that carried radiogenic-type helium; b) there is no evidence of He and CO2 decoupling so CO2 cannot be derived from a mantle source; and c) our experimental results of carbonate rock milling have shown the occurrence of CO2 production from mechanical processes.
2. The Field Work The field work started in October 1997 and is still ongoing. The three main dry vents of the seismic area of the Umbria region (Central Apennines) were repeatedly sampled and their gas output was evaluated during and after the seismic period. The collected samples were analyzed in the laboratory for their chemical composition and the isotopic composition of both carbon and helium. Details of the field sampling and analytical methods are given in Appendix 1.
Vol. 165, 2008
CO2 Degassing over Seismic Areas
77
Figure 1 displays the location of the CO2-dominated venting gases (Umbertide – UB; Montecastello di Vibio – VB; San Faustino – SF; Caprese Michelangelo – CM), besides other CO2-rich sources (bubbling and dissolved gases, CO2 exploited wells) of Central
Figure 1 Simplified map of the main seismogenic sources (top) crossing the Apennine chain (after VALENSISE and PANTOSTI, 2001). The location of the sampling sites is reported on the heat flow map of the investigated area of the Central Apennines (modified after CATALDI et al., 1995). The numbers show the sampling sites’ location as listed in Table 1.
78
F. Italiano et al.
Pure appl. geophys.,
Apennines. Table 1 lists the main geochemical features of the CO2-dominated gases from the Central Apennines. Hereafter, some information on the venting sites are summarized: • The Caprese Michelangelo (CM) degassing area is located to the north of the seismic area and characterized by the presence of several natural degassing vents and a well drilled to a depth of about 5000 m (HEINICKE et al., 2006). The main CO2 reservoir is located at a depth of 3700 m, at a static pressure of 70 MPa and a temperature of 120C. • The Montecastello di Vibio (VB) gas emission is located within Miocenic sediments and is characterized by a CO2 flow rate of about 1 m3/sec. Flow rate appears to be independent of seasonal variations. • The S. Faustino (SF) CO2-dominated gas emission occurs close to Massa Martana. Its flow rate is about 1 m3/sec and is characterized by constancy in time and apparently no sensitivity to seasonal variations. It is also a naturally-sparkling water exploited for bottling. • The Umbertide (UB) CO2-dominated gas emission derives from drilling for hydrocarbon exploration up to a depth of about 4800 m. It is currently the strongest CO2 emission located in a very dangerous depression close to the Umbertide town (Umbria Region, Central Apennines). The venting gas is characterized by a flow rate of about 3 m3/sec. The helium isotopic composition in the samples from UB and CM displays the lowest ratios in the range of 0.026–0.034 Ra (Ra = atmospheric ratio = 1.39 9 10-6) and 0.03–0.05 Ra at UB and CM vents and the well, respectively. Vents sampled at MV and SF have 3He/4He ratio ranging between 0.12–0.18 Ra and 0.24–0.28 Ra, respectively higher by an order of magnitude with respect to the 3He/4He ratio measured at the UB site. The helium concentration averages 50 ppmv at the CM gas vents and 440 ppmv at the CM well; 14.2 ± 1.5 (1r) ppmv at MV, 8.5 ± 0.7 ppmv at SF and 44 ± 8 ppmv at UB.
3. The Fluids Geochemistry and the CO2 Output during the Seismic Crisis The last seismic crisis that hit the Central Apennines had a strong impact on Italian society because it killed and injured people, and destroyed a large number of buildings including masterpieces of Italian cultural heritage. The area was struck by hundreds of earthquakes, starting on September 4, 1997, with a Ml = 4.4 foreshock, followed by two events with Ml = 5.6 and 5.8 on September 26 (MORELLI et al., 2000). Information on the earthquakes during the entire seismic sequence is available on web data bases (C.S.I., 1.1; 2003). The hypocenters of the entire seismic sequence were located at 5–10 km depth in the shallow crust (AMATO et al., 1998), except for one on March 26, 1998 (Mw 5.3), located at a depth of 51 km
CO2 Degassing over Seismic Areas
Vol. 165, 2008
79
Table 1 Main geochemical features for gas sources located along the Central Apennine chain. CO2 concentration, carbon and helium isotopic composition ðexpressed as R/RaÞ are reported. Samples labelled with * are from drilled wells. Geographical references in UTM coordinates. Data source: aÞ This work; bÞ MINISSALE ð2004Þ Site # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 40 41 42 43 44 45 46 47 48 49
Site name
Latitude
Longitude
% CO2
Asciano Acqua Bolle Caprese Michelangelo * Baccanella Piersanti Bruciano Borboi Vagliagli Acqua Borra Il Bagno * Caggiolo * Rimaggio * Grotte S. Stefano San Faustino Umbertine Montecastello di Vibio Stifone Triponzo Parrano Fersenone Fontecchio Rapolano Torrite di Siena Venturina Pienza Bagni S. Filippo S.Albino Zancona Roselle Selvena Torre Alfina Saturnia Pereta S.Martino Fiora Strada Ferento Muralto Bagnaccio Tuscania Monterozzi Terme Cotilia Solfatara Nepi Vaiano Borgo Pantano Caldara Palidoro Tivoli Lavinio Cava di selci
4851292 4842116 4837423 4836167 4834130 4782165 4818901 4814551 4801192 4817118 4798615 4817989 4709216 4709677 4800803 4746322 4708090 4743827 4750135 4749531 4676447 4799675 4786346 4774391 4773372 4757952 4751219 4750256 4746821 4740856 4737198 4728674 4728234 4727675 4709875 4707201 4705009 4703754 4702096 4693510 4678936 4678053 4670899 4663899 4650026 4647013 4595671 4504251
136601 183740 256719 153970 130131 653941 153440 205176 210303 716793 727027 717845 264826 300994 281548 283748 294277 331380 264029 276943 384897 224437 233189 141376 229191 229238 244494 217022 184431 224012 250303 214311 201312 222162 259556 284651 259182 244956 225945 335135 277171 261742 232195 260205 267870 311168 299133 520819
65.6 94.8 94.8 96.4 99.1 91.3 96 94.4 99.2 99.8 97.2 96.3 98.5 94 93 92 23.7 56.2 43.2 48.1 18 99.3 93.1 95.2 94.8 96.1 99.4 94.9 27.3 90.2 98.8 34.7 75.3 99.2 97.9 97.3 99.3 97.5 98.4 95.8 97.6 98.3 98 98 97.7 91.6 94.2 98.1
d13C CO2
R/Ra
Source
-10.3 -6.5 -4.1 -7.0 -6.7 -5.1 -9.3 -5.1 -5.9 -8.0 n.a. -7.2 0.1 -1.7 -3.5 -0.1- + 2.1 -2.5 -4.4 n.d. n.d. n.d. -7.5 -3.8 -13.4 -3.6 -3.3 -5.2 -4.6 -9.5 -3.3 1.0 -6.3 -6.2 0.1 -0.3 -0.9 -1.9 n.a. -0.1 -2.1 -1.3 -0.2 -2.1 -2.4 -1.8 -3.5 -0.5 n.a.
0.09 0.02 0.03 0.04 0.14 2.06 0.07 0.08 0.17 0.03 0.28 0.03 0.03 0.24 0.02 0.14 0.13 0.02 0.08 0.06 0.31 0.13 0.11 0.82 0.21 0.13 0.08 0.45 0.07 0.4 0.4 0.4 0.89 0.69 0.56 0.64 0.54 0.43 0.36 0.11 0.26 0.45 0.31 0.24 0.18 0.6 0.22 1.46
b b a b b b b b b a a a a a a a a a a a a b b b b b a b b b a b b b b b b b b a b b b b b b b a
80
F. Italiano et al.
Pure appl. geophys.,
Table 2 CO2 output from degassing areas, wells and vents of Central Apennines. Gas output data from wells refer to production data. The wells drain gas from shallow ð200–700 m deepÞ and pressurized reservoirs Site Geoth. systems Tuscany Grotte S. Stefano Caggiolo S. Faustino Vibio Stifone S. Albino Umbertide Il Bagno Grotte S. Stefano Caggiolo Rimaggio
Source type Geothermal gases Drilled well Drilled well Venting and bubbling gas Venting and bubbling gas Springs Drilled well Venting and bubbling gas Drilled well Drilled well Drilled well Drilled well
CO2 t/d
R/Ra
Ref.
39.8 64.8 4.8 7.0 10.0 1.0 38.4 20.0 3.6 61 6 6
2.80 0.60 0.30 0.23 0.13 0.10 0.08 0.03 0.03 0.03 0.03 0.03
CHIODINI et al. (1999) This work This work ITALIANO et al. (2001) ITALIANO et al. (2001) CHIODINI et al. (1999) This work ITALIANO et al. (2001) This work This work This work This work
(MORELLI et al., 2000). Focal mechanisms of the main shocks which occurred in 1997 (as well as 1979 and 1984), also confirmed by stress indicators (MONTONE et al., 1997), highlighted active extension processes in the NE-SW and E-W directions (FREPOLI and AMATO, 1997; AMATO et al., 1997). The entire seismogenic process caused crustal deformation with a maximum horizontal co-seismic displacement of 14 ± 1.8 cm and a maximum co-seismic subsidence of 24 ± 3 cm detected by means of SAR differential interferometry and GPS data (STRAMONDO et al., 1999), while a post-seismic long-term deformation process was detected by means of levelling data (BASILI and MEGHRAOUI, 2001). Starting from the beginning of the seismic crisis (September 26, 1997), samples of thermal waters (Bagni di Triponzo, Parrano and Stifone) and gas vents (Montecastello
Figure 2 Schematic representation of the experimental system. The jar with the ring mill is washed by the GC-grade inert gas (argon), then the milling process starts and the mechanochemically-produced gas is collected inside the preevacuated sampling bottle. It is possible to collect several samples simply by switching the three-way valve to replace the bottle.
Vol. 165, 2008
CO2 Degassing over Seismic Areas
81
di Vibio, Umbertide, San Faustino) were collected on a weekly basis during the period October 1997 to July 1998 and at longer intervals thereafter (twice a week, then monthly and then seasonal during 2001–2002). The data collected during a seismically quiet period (1999–2007) has allowed us to identify the background values for some geochemical parameters that could characterize the study area. Some previously published results have pointed out that: a) The thermal waters are CaSO4-enriched due to their circulation in the deep evaporitic basement and have changed their mixing proportions with cold waters typically equilibrated in shallow carbonate systems (FAVARA et al., 2001); during the seismic period, deep-originating waters contaminated shallow aquifers exploited for human purposes (ITALIANO et al., 2004); b) the gases dissolved in the thermal waters, CO2-dominated with the presence of CH4 and an excess of N2 in respect to the atmosphere, displayed various influxes of both deep-originating and atmospheric-derived components (ITALIANO et al., 2004); the 3He/4He ratios of the venting gases showed that although the region is located in a typical crustal environment, a slight contribution of a mantle-derived component could be detected (ITALIANO et al., 2001); the CO2 flux had shown modifications during the occurrence of the seismic crisis (HEINICKE et al., 2000; ITALIANO et al., 2001). A significant fluctuation of the 3He/4He ratio was observed at the MV site in concomitance with the strongest seismic events of 1997 and with the March 26, 1998 deep event. As an example, Figure 3 displays the results of the long-term monitoring at the gas vent of Umbertide. The data are useful for evaluating the temporal variations of CO2 and He contents in addition to 3He/4He ratios. The earthquakes recorded in the area are also reported (starting from the foreshock of the seismic crisis) and filtered as to be higher than Ml 2.5 and to have occurred within a radius of 40 km to the respect of the sampling site. The geochemical variations record changes in crustal permeability due to deformations associated with the subduction processes that characterize the Apennine chain. The most relevant geochemical changes were apparently linked to 5 events characterized by Ml C 5.6 and limited to the period September 1997 to April 1998. Conversely, the long-term geochemical trends indicate the persistence of long-lasting degassing processes which have no relationships with the seismic shocks. The CO2 released over the area affected by the seismic crisis was probably already stored in relatively shallow reservoirs whose existence is demonstrated by the large number of wells drilled for CO2 exploitation up to a depth of 5000 m (e.g., Caprese Michelangelo-CM and Umbertide – UB sites). Field measurements and observations carried out during the seismic crisis revealed that the broadly higher CO2 degassing occurred over the entire seismic area and induced a pH-drop in several springs with an increase in flow rate of about 50% at all the main gas vents (ITALIANO et al., 2001). All of the released CO2-dominated gases carried helium with a typical radiogenic signature. Only during the 1997–1998 seismic crisis were small additions of mantle-derived helium detected in the venting gases (ITALIANO et al., 2001) and
82
F. Italiano et al.
Pure appl. geophys.,
Figure 3 Temporal variations of CO2 (vol. %) and He (ppm vol.) concentration and 3He/4He ratio (R/Ra) values at the venting site of Umbertide. The earthquakes which occurred within a range of 40 km and with a magnitude above 2.5 are reported.
interpreted as due to normal faulting that temporarily enhanced the vertical permeability, allowing deep-originated fluids to mix with the shallow/crustal-derived main components.
CO2 Degassing over Seismic Areas
Vol. 165, 2008
83
4. Laboratory CO2 Production from Mechanical Energy The origin of CO2 in the absence of a degassing mantle or magmatic products is an already faced problem since early 1960s. Experimental results and theoretical approaches provided data and models about the possibility of generating CO2 as a result of hydrolysis processes (KISSIN and PAKHOMOV, 1969) or through decarbonation metamorphic processes (GIANELLI, 1985). All of the proposed models require a thermal energy source: KISSIN and PAKHOMOV (1969, 1975) deal with experimental water-rock interactions at boiling temperatures up to 250C; GIANELLI (1985) proposed theoretical results obtained in a PT range of 500–2000 bar (hydrostatic) and 350–520C with specific mineral assemblages. All of the previous results provide ways to produce CO2 in a geothermal environment, when enough thermal energy is available at a depth in the range of 2–8 km (lithostatic pressure, assuming a rock density of 2.6 g/cm3). The investigated area of Central Apennines is characterized by low heat flux with a geothermal gradient of about 30C/km. As such all of the proposed thermometamorphic temperatures can be available in the natural environment only at depths in the range of 11–17 km, namely at lithostatic pressures of 3–4.5 kbar, well above that in equilibrium for CO2 generation. The experimental activity aimed to produce CO2 by mechanochemical energy was developed in the laboratory using a commercial ring mill, modified to improve its gastightness and to allow the gas sample collection from the jar. Figure 2 schematically shows the experimental apparatus. A weighted amount of pure calcite powder was milled and the powder after the milling procedure was also carefully recovered and weighted. The produced gas was analyzed by gas chromatography. All of the experimental conditions are given in Table 3, while Table 4 lists the analytical results of the collected gas phase. Details of the experimental procedures are given in Appendix 2. The experimental results provide the information that mechanical energy is able to destroy the crystalline lattice allowing the calcite dissociation (CaCO3 = CaO + CO2). The analytical results reported on Table 4 show that the gas collected from the jar is always a mixture between air and an additional component mainly composed of CO2, where H2, CO and methane are also present in variable concentrations. We always adopted the same experimental procedure: to load the jar with the sample (5 g), wash the system by pure argon (washing time = 10 min.), milling (20 min.) and gas sampling. Finally we recovered the milled powder. The variable gas compositions we found Table 3 Results from the experiments on synthetic matter and natural rock ðCalcare Massiccio; Umbria region, ItalyÞ. The table displays the initial weight ðI.W.Þ, the final weight ðF.W.Þ after milling and the milling time Exp #
Composition
I.W. g
F.W. g
Grinding time minutes
A B C
Synthetic CaCO3 Natural CaCO3 Synthetic CaCO3 + Clay
5 5 5
2.5 n.d 3.9
20 10 20
84
F. Italiano et al.
Pure appl. geophys.,
Table 4 Analytical results of the collected gases. The sampling interval shows how many minutes after the beginning of the milling process the sampling bottle was opened ðfirst numberÞ and how long the sampling lasted ðsecond numberÞ. < = below the detection limit Exp. #
Sample #
H2 ppm
O2 vol.%
N2 vol.%
CO ppm
CH4 ppm
CO2 vol. %
Sampling interval
A A B B C C
1 2 5 6 9 10
< 177 72 142 < 428
19.5 19.3 0.4 5.9 19.9 4.3
78.3 78.0 2.0 23.6 77.2 17.7
< < 31 58 122 1.3
2.4 25 31 64 2.1 58
0.07 0.05 0.46 0.55 0.13 10.40
5–15 1–20 1–4 6–15 5–5 10–10
(Table 4) seem to be related to the different carbonate products (Table 3). The amount of CO2 produced by pure, synthetic calcite, is the lowest we measured, ranging from a little higher to double of the atmospheric content. The CO2 increased by one order of magnitude with the second set of experiments, when natural samples were milled (limestone from Central Italy known as Calcare Massiccio). The highest CO2 content, up to 10.4 vol%, was measured in gas samples collected when kaolinite was added to the synthetic calcite. To gain a quantitative estimation of the CO2 produced during the experiments, the gas analyses were recalculated and compared to the estimated volume of the milling system (3000 ml). Since the sampled gas was in an argon matrix, we restored the CO2 concentration assuming the total gas volume as composed of oxygen, nitrogen and carbon dioxide (Table 5). The restored CO2 concentration (Table 5) was then calculated by Ar removal and the amount of produced CO2, expressed as moles in Table 5, was estimated scaling the volume of the experimental system to the restored CO2 concentration. The recalculated CO2 concentrations show the presence of a significant amount of CO2 in most of the experiments, with concentrations up to 32%. The theoretical amount of CO2 produced by the mechanochemical process can be quantitatively estimated considering a total dissociation of the solid CaCO3; starting from Table 5 Recalculated CO2 data. Starting from the CO2 concentration data ðTable 4Þ. The table lists the recalculated data considering: The total gas concentration ð1Þ namely O2 + N2 + CO2 concentration; the Ar concentration ð100% - O2 + N2 + CO2Þ; the restored CO2 concentration is then calculated from the CO2 content ðTable 3Þ after Ar removal. The volume of CO2 is estimated considering the volume of the experimental apparatus ð3000 mlÞ and the CO2 concentration Sample #
Total gas (1)
Ar vol. %
Restored [CO2]
Volume moles CO2
1 2 5 6 9 10
97.84 97.37 2.93 30.03 97.28 32.41
2.16 2.63 97.07 69.97 2.72 67.59
0.07 0.05 15.70 1.83 0.14 32.09
9.29 1.04 2.60 2.84 2.35 4.80
9 9 9 9 9 9
10-5 10-4 10-2 10-3 10-4 10-2
Volume ml CO2 2.08 2.34 581.47 63.59 5.25 1074.40
Vol. 165, 2008
CO2 Degassing over Seismic Areas
85
5 grams (5 9 10-2 moles) of solid matter, the process might produce the same number of moles of both CO2 and CaO. The maximum volume of CO2 is thus fixed in 1120 ml. Table 5 shows the calculated amount of CO2 produced by the various experiments. The maximum amount (sample 10, Table 5) is close to the theoretical value. The powders from experiment A resulted to be 2.5 and 2 g (samples 1 and 2, respectively) after the weight, showing that a large amount of gas was missed during the experiments. Similar results came from the experiment C, sample 9: a final weight of 3.9 g (Table 4) should provide a CO2 amount of 115.4 ml, well above the measured 5.25 ml, showing that a better gas-tight system has to be developed. Although the experimental apparatus needs improvements and the experimental conditions can be considered distantly to approach the natural systems, the evaluation of the experimental results show that the production of CO2 is possible without any thermal energy as the primary energy source, and data from the induced pressure, had shown that they fit some conditions already determined on natural faults. In fact, the evaluation of pressure induced on samples by grinding was made by the analysis of residual stress on CaCO3 crystals, using the method of two peaks (MARTINELLI and PLESCIA, 2004). The mean stress was calculated to be: ðr1 þ r2 Þ ¼ ðE=rÞ Dd=d; where (r1 + r2) are the main components of the stress, E is the Young module, r is the Poisson ratio, d is the interplanar distance at normal condition and Dd is the difference induced by the stress on the interplanar distance. The maximum value of residual stress was about 6.2 kbar and the mean value of 2.2 kbar. The numerical simulations of the stress in a fault system (ZOBAK et al., 1987), provided a shear stress of the order of 2 kbar, assuming a Mohr-Coulomb relation to calculate the normal stress in a fault system at lithostatic pressure (3 kbar at a depth of 10 km). A more detailed tensorial analysis provided a depth-averaged shear strength for faults in the brittle continental crust in the range of 1.5 kbar for thrust faulting, 0.6 kbar for strike-slip faulting, and 0.3 kbar for normal faulting (HICKMAN, 1991).
5. Discussion To infer the provenance of a natural gas mixture, a useful approach is to couple the isotopic features of carbon and helium. The various CO2 sources of a natural environment are marked by a different d13C ratio that is considered a powerful tool to investigate the origin of CO2 in a gas mixture (e.g., FAURE, 1977; JAVOY et al., 1986; ROLLINSON, 1993); the d13C data listed in Table 1, however, show that the isotopic composition of carbon deriving from a mantle-derived source (geothermal area of Larderello, R/Ra in the range of 0.89–2.06) and from other, radiogenic type sources (R/Ra in the range of 0.03–0.2), span over the same wide range of d13C values (between +1% and -7%). Such a wide range of isotopic composition highlights that gases coming from the same geochemical
86
F. Italiano et al.
Pure appl. geophys.,
environment do not identify any typical source. Since modifications of the d13C values are related to a wide spectrum of possibilities (e.g., isotopic fractionation, isotopic equilibrium, mixing of different sources etc.) the isotopic composition of carbon from CO2 does not provide useful information to constrain the origin of the gases released in the Central Apennines. In contrast, helium comes from three different and well distinguishable sources (mantle, crust and air), each marked by different isotopic ratios (e.g., OZIMA and PODOSEK, 1983). Moreover, due to its chemical inertness, the 3He/4He ratio is only sensitive to mass fractionation and mixing processes. It is well documented that helium isotopic ratios can be useful for evaluating a variety of geophysical and geological environments (MAMYRIN and Tolstikhin, 1984; OZIMA and PODOSEK, 1983). For example, in subduction zones, there is a clear geographical contrast in the 3He/4He ratio between lower values in the frontal arc and values higher in the volcanic arc (e.g., Southern Italy; SANO et al., 1989). In continental settings, it has long been established that the presence of mantle-He correlates well with tectonic and magmatic activity (POLYAK and TOLSTIKHIN, 1985), and that the mantle-He input occurs in areas with young volcanism (O’NIONS and OXBURGH, 1983; MARTY et al., 1989). As such, He-isotope ratios have been used in several tectonically active regions to identify mantle-derived products intruded at shallow levels in the crust (e.g., ITALIANO et al., 2000). The fluids released through the seismic and geothermal areas of the Central Apennines are a CO2-dominated mixture containing helium with isotopic ratios ranging from a typical radiogenic of 0.03 Ra to a mantle-derived signature of 2.8 Ra (Table 1). The presence of the mantle, or mantle-derived products, at shallow levels may account for both higher helium isotope values and CO2 generation either by direct degassing or by thermometamorphic processes. However, although this happens in the volcanic and geothermal peri-Tyrrhenian areas, there is no clear evidence of mantlederived products in other CO2-dominated gases released over the seismic areas of the Central Apennine chain. The existence of the small, Pleistocene, volcanic structures of San Venanzo (STOPPA and SFORNA, 1995) in the vicinity of Montecastello di Vibio and Colle Fabbri (STOPPA, 1988) in the vicinity of Massa Martana gives clues about possible tectonic movements that might have induced temporary, large increases in the vertical permeability, allowing the passage of tiny amounts of mantle-derived volatiles. This short-lived effect might allow some magmas to intrude to shallower crustal layers and to be erupted, giving rise to small volcanic edifices (at a reduced scale, this is the case of the 3He anomalies recorded during the 1997–1998 seismic crisis; ITALIANO et al., 2001). Considering the geochemical features of the fluids circulating at present in the area (e.g., Parrano and Fersenone springs with CO2-dominated dissolved gases and He isotope ratio of 0.08 and 0.06 Ra, respectively) we cannot consider those magmatic products as the origin of the presentlyreleased CO2, neither in terms of gas availability nor in terms of thermal energy. The evidence of the absence of a shallow thermal energy source fits also with the general map of the heat flow (HF) over the Italian peninsula (CATALDI et al., 1995) that highlights the
Vol. 165, 2008
CO2 Degassing over Seismic Areas
87
Figure 4 He vs. 4He for the three sampled vents of Umbertide (UB), San Faustino (SF) and Montecastello di Vibio (VB) are shown (modified after ITALIANO et al., 2001). The plotted data originate from the long-term geochemical monitoring and show that the isotopic ratio is almost constant with the time even on a ten-years scale. The CO2 flux is reported in tons per day units. 3
low HF values along the Apennine chain which is a ‘‘cold’’ area. Figure 4 shows the calculated 3He and 4He contents for various mixing proportions between the ‘‘local’’ mantle-derived source and the radiogenic source. The 3He/4He ratio of 4.48 9 10-6 is the highest value measured in the Central Apennines (Larderello geothermal field), and adopted as the ‘‘local’’ mantle-derived end-member for our calculations. The mixing lines are for pure radiogenic component and for the addition of 1, 2, 4 and 10% of mantlederived helium. Using those end-members, the UB site may have an addition of 1% of mantle-derived helium at most, while gases from SF display the highest mantle-derived helium content, reaching a value close to 10%. Significantly, the amount of degassed CO2 is lower where the helium isotopic ratio is higher. Moreover, scaling the helium concentration with the 3He/4He isotopic ratio for all the sampling sites, we could calculate that, although the measured helium isotopic ratios differ by an order of magnitude, 3He content is similar at all the sites (UB = 1.8 ± 0.4 9 10-6, MV = 2.8 ± 0.3 9 10-6 and SF = 3.1 ± 0.3 9 10-6 lmol/mol). It is also worth noting that the higher the helium concentration, the lower the 3He/4He isotopic ratio. The abovementioned observations suggest that similar amounts of 3He are supplied on a regional scale by the deep mantle-derived source, and 4He is added in different proportions at shallow depth together with CO2.
88
F. Italiano et al.
Pure appl. geophys.,
Figure 5 XRD of ‘‘as is’’ sample (natural limestone, black spectrum) and of the recovered powder (red spectrum) after 20’ grinding. The calcite peaks disappeared and the red line shows an amorphous powder marked by a higher background and the presence of only quartz as a crystalline product.
The occurrence of travertine rocks in the vicinity of many CO2 gas emissions (BARBIER and FANELLI, 1976; MINISSALE et al., 2002), highlights how the CO2 ascent is a massive and long-lasting phenomenon. The CO2 flux from the investigated vents (Table 2) can be considered as a massive degassing if compared, for example, to the CO2 degassing rate from the fumarolic field of the volcanic island of Vulcano (Aeolian Islands) estimated to be in the range of 50–200 tons per day as a function of the magmatic activity (ITALIANO et al., 1994). A possibility to support such a large degassing rate over a sedimentary area might be a shallow-depth magmatic intrusion as reported for the gas emission of Mefite d’Ansanto (Southern Apennines, ITALIANO et al., 2000) where helium isotope ratios reveal a clear mantle-derived signature (2.54 Ra). In those conditions, a CO2 production from any mantle-derived contribution lacks, and different or additional production mechanisms have to be hypothesized. Among them, the mechanochemical process is proposed as the energy feeder for CO2 production and degassing, for which experimental results fit with observations from the natural environment.
6. Conclusions The huge amount of CO2 released along the seismic area of the Central Apennines is considered to be originated by processes involving mantle degassing and/or thermometamorphism. Even though this is true for the geothermal areas (e.g., Larderello geothermal field), it does not look like the case for the seismic areas mainly located in the Umbria and Marche regions.
Vol. 165, 2008
CO2 Degassing over Seismic Areas
89
An unavailable thermal (mantle or mantle-derived) source along the Central Apennines, considered also a ‘‘cold’’ area on the basis of the low HF values besides the isotopic signature of helium that denotes a typical radiogenic (crustal) origin of the circulating fluids, rules out the origin of CO2 from mantle-derived or thermometamorphic processes. The mechanochemical production of CO2 appears by the process that is able to supply enough CO2 by a diverse source from the mantle and to be a comporary fit of the collected experimental and field data. We propose that over the investigated area, mantle-derived fluids are available at the main geothermal area of Larderello (Tuscany Region, western sector). The high heat flux denotes also the availability of thermal energy capable of producing CO2 from thermometamorphism that mixes to mantle-derived CO2 as already noted by other authors. The fluids from that area move through the tectonic discontinuities and can be spread throughout the Central Apennines. The mechanical energy available as the friction produced by fault movement and/or by deformation can be the other energy source able to produce CO2 and to release the radiogenic products trapped in rocks. Temporal modifications of the gas composition, including modifications of the helium isotopic composition, are interpreted as modifications of the mixing ratios between shallow and deep components.
Acknowledgments The authors are grateful to the Ph.D. student Sonia Pizzullo who supported the laboratory work proposed here and is still producing new data on the mechanochemical production of greenhouse gases. The authors wish to acknowledge D. Hilton, M. Kurtz, J. Heinicke, N. Perez, O. Vaselli, H. Woith, G. Yuce and M. Zimmer for their interesting and extensive discussions pertaining on this new topic during the Ninth International Conference on Gas Geochemistry, ICGG9, Taipei Conference. The paper benefited from the revision by David Hilton and Jens Heinicke, who greatly improved the earlier version of the manuscript. The work was partially supported by funds from the INGV-DPC-V5 project, RU9-Italiano.
Appendix 1 The gas samples were collected in pyrex flasks with valves at both ends. The bubbling gases were collected using a funnel to carry the bubbles toward the sampling bottle. The whole sampling apparatus, including the sampling bottle, was submerged and filled with water to prevent any air contamination. The venting gases were collected using a pipe or a funnel, depending on the site conditions. A long tygon tube connected the pipe to the sampling system made of a three-way valve joining together the sampling bottle and a syringe. The sample was collected after a cleaning procedure to purge the sampling bottle
90
F. Italiano et al.
Pure appl. geophys.,
of air. The gas was slowly aspired by the syringe and then pushed hardly to the bottle several times. The cleaning procedure was repeated until the aspired gas volume was 10 times larger than the volume of the sampling bottle. In the laboratory, the chemical analyses, the carbon isotopic ratios and the helium isotopic ratios were determined by gas chromatography and mass spectrometry, respectively, using gas aliquots extracted from the same sampling bottle. Chemical analyses were carried out using a Perkin Elmer 8500 gas chromatograph. An aliquot of 0.3 cm3 at a known pressure was admitted to the allmetal purification UHV line for helium isotope analysis. Helium was purified following three steps in isolated sections of the system: First, the reactive gases were absorbed into a charcoal trap held at 77 K; then two SAES getters worked simultaneously to absorb residual nitrogen at 250C and hydrogen at room temperature, and finally, after measuring the 4He/20Ne ratio and checking for residual 40Ar on an in-line quadrupole mass spectrometer, was to completely separate helium from neon by a charcoal trap held at 40 K. The isotopic analyses of the purified helium fraction were performed by a modified static vacuum mass spectrometer (GVI5400TFT) that allows for the simultaneous detection of 3He and 4He-ion beams, thereby causing the 3He/4He measurement error to drop to very low values. Typical uncertainties in the range of low-3He (radiogenic) samples are within ±5%. The gas output estimates were carried out following the simple method of water displacement. The gas was captured by a funnel or a specific device able to convey all the released gas to a container filled with water. The time the gas took to displace the water out of the container provided the output of the venting gas. For wide gas emissions, all the venting surface was covered by the adopted device. The uncertainty of the method (in the range of ±10%) was estimated by repeated measurements on the same vent.
Appendix 2 For the simulation of friction and mechanical compression over a large surface, we used a ring mill, which is the most efficient grinding system available at present for laboratory work (PLESCIA et al., 2003; MARTINELLI and PLESCIA 2004). The ring mill consists of a reinforced jar to contain the material and a grinding body made of a steel ring and solid steel roller. The ring mill operates basically through nonhydrostatic compression impact on the material particles and through friction rotation forced by the grinding body on the particles and between the particles and the jar walls. We account for some experiments carried out using a weighted amount of solid matter (5 grams) milled for 10–20 minutes (see Table 3). The system was connected to a bottle of high-pure argon to purge the jar before milling and to a pre-evacuated sampling bottle. An inlet and an outlet hole on the jar cover (Fig. 2) were allowed to flush before the beginning of the milling process to minimize possible reaction due to the presence of air. Then a pre-evacuated (internal pressure = 10-4 mbar) pyrex gas sampling bottle was connected at the outlet hole and
Vol. 165, 2008
CO2 Degassing over Seismic Areas
91
gas samples were collected at fixed time intervals during the 20’ milling (5, 10, 15 and 20 minutes after beginning the milling). Table 3 displays the experimental conditions for every single laboratory experiment. The collected gases were analyzed for their chemical composition. Chemical analyses were performed by a Perkin Elmer gas chromatograph equipped with a 4-m Carboxen 1000 column, using Ar as carrier and a double detector HWD and FID plus a methanizer. The detection limits are: 2 ppm vol. for He and H2, 1 ppm vol. for CH4 and CO2 and 500 ppm vol. for N2 and O2. The analytical errors are ± 5% for He and ± 3% for the other gaseous species. The analytical results are listed in Table 4. At the end of the milling process the pulverized sample was carefully collected by cleaning the jar, and all the grinding parts and the powder were weighted to evaluate the weight loss. XRD and TGA analyses were carried out before and after the milling process to evaluate the main composition of the solid matter. The solid matter milled during our experiments consisted of a pure synthetic calcium carbonate for analyses, a natural limestone and a calcium carbonate-kaolinitic clay mixture. The maximum theoretical amount of CO2 produced by the process can be calculated considering a total dissociation of the solid CaCO3, therefore starting from 5 grams (5 9 10-2 moles) of solid matter, the process might produce the same number of moles of both CO2 and CaO. The maximum volume of CO2 is thus fixed in 1120 ml, however the amount of milled powder (namely pure CaO) we extracted after the process is lower than the theoretical weight by 10%, probably because of manual errors in collecting all of the powder from the jar. Since the boundary conditions of the experiment were far to match a close system, losses of the produced gases have to be considered. Although the adopted system had many technical limitations, all the gas chromatographic analyses show the presence of CO2 concentration anomalies. In particular the slight CO2 anomaly found in the first experiment (synthetic CaCO3), becomes a significant anomaly (Table 4) with the second experiment (natural limestone). The last experiment (C; Tables 3 and 4) was carried out with some improvements to minimize the leaks from the jar, and the collected CO2 approached the theoretical value. The gas-chromatographic analyses of the gases generated during the milling of a mixture 2.5 + 2.5 grams of CaCO3 + Kaolynite (‘‘C’’ experiment, Table 3) display the largest CO2 concentration besides a significant amount of H2 and CH4. Although the considerations of the release of H2 and carbon from the steel jar are still valid, dehydroxylation processes of the clay fraction provide a large amount of H+ and OH- ions that can take part to the chemical reactions. The protonization event coincides with dehydroxylation of the clay during grinding. It has been clearly demonstrated by different authors (AGLIETTI et al., 1986; PLESCIA et al., 2003; KAMEDA et al., 2004) that phyllosilicates are particularly sensitive to mechanochemical action, which easily leads to the elimination of the OH groups and to a completely amorphous structure. The final gas composition is closely related to the mineralogical composition of the milled material, allowing the production of CH4, CO,
92
F. Italiano et al.
Pure appl. geophys.,
H2 besides the large amount of CO2. The produced gas phase exhibits a complex chemical composition that, however, has already been observed as H2 and hydrocarbon generation during grinding processes (e.g., AGLIETTI et al., 1986; KHOMENKO and LANGER, 1999; KAMEDA et al., 2004). Figure 3 shows a notable difference between the XRD spectra of the as-is and the milled samples, where the calcite peak is clearly present before the milling, while quartz is the only crystalline phase in the milled sample. A notable difference can be observed between gas samples collected a short time after the beginning of the milling process and gas samples collected at the end of the process. Good results were also obtained for the natural carbonate sample (Calcare Massiccio from the Central Apennines): we had the lowest air contamination and a reasonable amount of CO2 in the range of about 0.5% vol. Even though this paper focuses attention on the mechanochemical production of CO2 from pure calcium carbonates, it is valuable to note the significant H2 and CH4 generation during the milling process of CaCO3Kaolinite mixtures.
REFERENCES AGLIETTI, E.F., LOPEZ, P.J.M, and PEREIRA, E. (1986), Mechanochemical effects in kaolinite grinding, Int. J. Min. Proc. 16, 135–146. AMATO, A. and 18 others (1998), The 1997 Umbria-Marche, Italy, earthquake sequence: A first look at the main shocks and aftershocks, Geophys. Res. Lett. 25, 15, 2861–2864. BARBIER, E. and FANELLI, M. (1976), Main fractures of Italy from Earts satellite images and correlations with thermal springs, volcanoes and earthquakes. In (Aoki, H. and Iizuka, S., Eds.). BASILI, R. and MEGHRAOUI, M. (2001), Coseismic and postseismic displacement related with the 1997 earthquake sequence in Umbria-Marche (Central Italy), Geophys. Res. Lett. 28, 14, 2695–2698. BOSCHI, E., GUIDOBONI, E., FERRARI, G., MARIOTTI, D., VALENSISE, G., and GASPERINI, P. (2000), Catalogue of Strong Italian Earthquakes from 461 BC to 1197- Introductory text and CDrom. Annali di Geofisica 43, 4, 609–868. CARACAUSI, A., ITALIANO, F., MARTINELLI, G., PAONITA, A., and RIZZO, A. (2005), Long-term geochemical monitoring and extensive/compressive phenomena: Case study of the Umbria region (Central Apennines, Italy), Annals of Geophys. 48, 1, 43–53. CATALDI, R., MONGELLI, F., SQUARCI, P., TAFFI, L., ZITO, G., and CALORE, C. (1995), Geothermal ranking of Italian territory. Geothermics 24, 115–129. CATALOGO DELLA SISMICITa` ITALIANA (2003), CSI 1.1 1981–2002, http://legacy.ingv.it/CSI/. CERLING, T., QUADE, J., YANG, W., and BOREMAN, J. (1989), Soil and paleosols as ecologic and paleoecologic indicators, Nature 341, 138–139. CHIODINI, G., FRONDINI, F., and PONZIANI, F. (1995), Deep structures and carbon dioxide degassing in Central Italy, Geothermics 24, 81–94. CHIODINI, G., FRONDINI, F., KERRIK, D.M., ROGIE, J., PARELLO, F., PERUZZI, L., ZANZARI, A.R. (1999), Quantification of deep CO2 fluxes from central Italy. Examples of carbon balance for regional aquifers and of soil diffuse degassing. Chem. Geol. 159, 205–222. FAURE, G., Principles of Isotope Geology (J. Wiley, New York 1977). FAVARA, R., ITALIANO, F., and MARTINELLI, G. (2001), Earthquake-induced chemical changes in thermal waters of Umbria region during the 1997–1998 seismic swarm, Terra Nova, 13-3, 227–233. FREPOLI, A. and AMATO, A. (1997), Contemporaneous extension and compression in the Northern Apennines from earthquake fault-plane solutions, Geophys. J. Int. 129, 368–388.
Vol. 165, 2008
CO2 Degassing over Seismic Areas
93
GIANELLI,, G. (1985), On the origin of geothermal CO2 by metamorphic processes, Boll. Soc. Geol. Ital. 104, 575–584. HEINICKE, J., ITALIANO, F., LAPENNA, V., MARTINELLI, G., NUCCIO, P.M. (2000), Coseismic geochemical variations in some gas emissions of Umbria region, Central Italy, Phys. Chem. Earth 25, 289–293. HEINICKE, J., BRAUN, T., BURGASSI, P., ITALIANO, F., and MARTINELLI, G. (2006), Gas flux anomalies in seismogenic zones in the Upper Tiber Valley, Central, Geophys. J. Int. 167, 794–806. HICKMAN, S. (1991), Stress in the lithosphere and the strength of active faults, Rev. Geophys. 29, 759–775. IRWIN, W.P. and BARNES, I. (1980), Tectonic relations of carbon dioxide discharges and earthquakes, J. Geophys. Res. 85 (B6), 3115–3121. ITALIANO, F., NUCCIO, P.M., and PECORAINO, G. (1994), Fumarolic gas output at La Fossa di Vulcano Crater, Acta Vulcanolog. 6, 39–40. ITALIANO, F., MARTELLI, M., MARTINELLI, G., and NUCCIO, P.M. (2000), Geochemical evidences of melt intrusions along lithospheric faults of Irpinian Apennines (Southern Italy): Geodynamic and seismogenetic implications, J. Geophys. Res. 105, B6, 13569–13578. ITALIANO,, F., MARTINELLI, G., and NUCCIO, P. M. (2001), Anomalies of mantle-derived helium during the 1997– 1998 seismic swarm of Umbria-Marche, Italy, Geophys. Res. Lett. 28, 5, 839–842. ITALIANO, F., MARTINELLI, G., and RIZZO, A. (2004), Seismogenic-induced variations in the dissolved gases of the thermal waters of the Umbria region ðCentral Apennines, ItalyÞ during and after the 1997–1998 seismic swarm. G-Cubed 5, 11, doi:10.1029/2004GC000720. JAVOY, M., PINEAU, F., and DELORME, H. (1986), Carbon and nitrogen isotopes in the mantle, Chem. Geol. 57, 41–62. KAMEDA, J., SARUWATARI, K., and TANAKA, H. (2004), H2 generation during grinding of kaolinite, J. Colloid and Interface Sci. 275, 225–286. KANAMORI, H. (1994), Mechanics of earthquakes, Ann. Rev. Earth Planet. Sci. 22, 207–237. KHOMENKO, V.M. and Langer, K. (1999), Aliphatic hydrocarbons in structural channels of cordierite: A first evidence from polarized single-crystal IR absorption spectroscopy, Am. Min. 84, 1181–1185. KISSIN, I.G. and PAKHOMOV, S.I. (1967), The possibility of carbon dioxide generation at depth at moderately low temperature, Dokl. Akad. Nauk SSSR 174, 451–454. KISSIN, I.G. and PAKHOMOV, S.I. (1969), A contribution to the geochemistry of carbon dioxide in deep zones of underground hydrosphere. Geokhimiya, 4, 450–471 (in Russian). KISSIN, I.G. and PAKHOMOV, S.I. (1975), Some features of the geochemistry of thermal water in platform areas from experimental data. Proceedings of the Grenoble Symposium, August 1975, IAHS publication 119, 7–15. MAMYRIN, B. A. and TOLSTIKHIN, I. N. (1981), Helium Isotopes in Nature (Energoizdat, Moscow) [in Russian]. MARINI, L. and CHIODINI, G. (1994), The role of carbon dioxide in the carbonate-evaporite geothermal systems of Tuscany and Latium ðItalyÞ, Acta Vulcanol. 5, 95–104. MARTINELLI, G. and PLESCIA P. (2004), Mechanochemical dissociation of calcium carbonate: laboratory data and relation to natural emissions of CO2, Phys. Earth Planet. Int. 142, 3–4, 205–214. MARTY, B., JAMBON, A., and SANO, Y. (1989), Helium isotopes and CO2 in volcanic gases of Japan. Chem. Geol, 76, 25–40. MINISSALE, A. (1991), Thermal springs in Italy: Their relation to recent tectonics, Appl. Geochem. 6, 201–212. MINISSALE, A., (2004), Origin, transport and discharge of CO2 in Central Italy. Earth Sci. Rev. 66, 89–141. MINISSALE, A., EVANS, W., MAGRO, G., and VASELLI, O. (1997a), Multiple source components in gas manifestations from northcentral Italy, Chem. Geol. 142, 175–192. MINISSALE, A., KERRICK, D.M., MAGRO, G., MURRELL, M.T., PALADINI, M., RIHS, S., STURCHIO, N.C., TASSI, F., and VASELLI, O. (2002), Geochemistry of Quaternary travertines in the region north of Rome ðItalyÞ : structural, hydrologic and paleoclimatic implications, Earth Planet. Sci. Lett. 203 709–728. MONTONE, P., AMATO, A., FREPOLI, A., MARIUCCI, M.T., and CESARO, M. (1997), Crustal stress regime in Italy, Annali di Geofisica XL, 3, 741–757. MORELLI,, A., EKSTROM, G., and OLIVERI, M. (2000), Source properties of the 1997–1998 Central Italy earthquake sequence from inversion of long-period and broad-band seismograms, J. Seismol. 4, 365–375. O’NIONS, R.K. and OXBURGH, E.R. (1983), Heat and helium in the Earth, Nature 306, 429–431. OZIMA, M. and PODOSEK, F.A. Noble Gas Geochemistry. (Cambridge University Press, Cambridge, 1983) 286 pp. PANICHI, C., and TONGIORGI, E. (1976), Carbon isotopic composition of CO2 from springs, fumaroles, mofettes and travertines of central and southern Italy: A preliminary prospection method of geothermal areas, Proc.
94
F. Italiano et al.
Pure appl. geophys.,
2nd U.N. Symp. on the Develop. and Use of Geotherm. Energy, San Francisco, USA., 20–29 May 1975, pp. 815–825. PLESCIA, P., GIZZI, D., BENEDETTI, S., CAMILUCCI, L., FANIZZA, C., and PAGLIETTI, F. (2003), Mechanochemical treatment to recycling asbestos containing waste, Waste Managem. 23, 209–218. POLYAK, B.G., TOLSTIKHIN, I.N. (1985), Isotopic composition of the Earth’s helium and the problem of tectogenesis. Chem. Geol. 52, 9–33. ROLLINSON, H., Using Geochemical Data (Longman Group, London 1993). SANO, Y., WAKITA, H., ITALIANO, F., and NUCCIO, P.M. (1989), Helium isotopes and tectonics in southern Italy, Geophys. Res. Lett. 16, 6, 511–514. STOPPA, F. (1988), L’eurimite di Colle Fabbri ðSpoletoÞ: un litotipo ad affinita` carbonatitica in Italia, Boll. Soc. Geol. It. 107, 239–248. STOPPA, F. and SFORNA, S. (1995), Geological map of the San Venanzo volcano ðCentral ItalyÞ: Explanatory notes, Acta Vulcanologica 7, 85–91. STRAMONDO et al. (1999). ZOBACK, M.L., ZOBACK, V., MOUNT, J., EATON, J., and HEALY et al. (1987), New evidence of the state of stress of the San Andreas fault zone, Sci. 238, 1105–1111. (Received October 16, 2007, revised October 28, 2007, accepted October 30, 2007)
To access this journal online: www.birkhauser.ch/pageoph
Pure appl. geophys. 165 (2008) 95–114 0033–4553/08/010095–20 DOI 10.1007/s00024-007-0281-9
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
Changes in the Diffuse CO2 Emission and Relation to Seismic Activity in and around El Hierro, Canary Islands ELEAZAR PADRO´N, GLADYS MELIA´N, RAYCO MARRERO, DA´CIL NOLASCO, JOSE´ BARRANCOS, GERMA´N PADILLA, PEDRO A. HERNA´NDEZ, and NEMESIO M. PE´REZ
Abstract—Significant changes in the diffuse emission of carbon dioxide were recorded in a geochemical station located at El Hierro, Canary Islands, before the occurrence of several seismic events during 2004. Two precursory CO2 efflux increases started thirteen and nine days before two seismic events of magnitude 2.3 and 1.7, which took place near El Hierro Island, Canary Islands, on March 23 and April 15, reaching a maximun value of 51.1 and 46.2 g m-2 d-1, respectively, five and eight days before the two seismic events. Other similar increases started thirteen and five days before the occurrence of two seismic events of magnitude 1.3 and 1.5 which took place on October 15 and 21 respectively, reaching the maximum values four and one day before the earthquakes. These changes were not related to variations in atmospheric or soil parameters. The Material Failure Forecast Method (FFM), which analyzes the rate of precursory phenomena, was successfully applied to forecast the first seismic event that took place in El Hierro Island in 2004. Key words: El Hierro Island, precursors, Material Failure Forecast Method, diffuse degassing, carbon dioxide.
1. Introduction El Hierro Island (278 km2) is one of the youngest and the southwesternmost of the Canary Islands and rises 4000 m above the sea floor (Fig. 1). The main characteristics of El Hierro consist of a truncated trihedron shape and three convergent ridges of volcanic cones. The older subaerial rocks of El Hierro have been dated at 1.12 Ma (GUILLOU et al., 1996) and there is only one questionable report of a single volcanic eruption in El Hierro Island in the last 500 years, Lomo Negro volcano, in 1793 (HERNA´NDEZ PACHECO, 1982). The volcanic evolution of El Hierro can be divided into three successive volcanic edifices: Tin˜or volcano, El Golfo volcano and Rift Volcanism (GUILLOU et al., 1996; CARRACEDO et al., 1997). The island has been covered in the last 37 ka by lavas erupted by the last stage of the volcanic evolution and deep embayment has been produced by giant landslides between the three rift zones, being the most recent El Golfo failure on the
Enviromental Research Division, Instituto Tecnolo´gico y de Energı´as Renovables (ITER), 38611 Granadilla, S/C de Tenerife, Spain. E-mail: [email protected]
96
E. Padro´n et al.
Pure appl. geophys.,
northwest flank of El Hierro, which occurred approximately 15 ka ago (MASSON, 1996; GEE et al., 2001). Since fumarolic activity is absent at the surface environment of El Hierro, the study of the evolution of diffuse CO2 emissions becomes an ideal geochemical tool for monitoring its volcanic activity (CHIODINI et al., 1998; HERNA´NDEZ et al., 2001a, b, c, 2003, 2006; SHIMOIKE et al., 2002; FRONDINI et al., 2004; NOTSU et al., 2005, 2006; GRANIERI et al., 2006). CO2 is, after water vapor, the major gas species in basaltic magmas (BARNES et al., 1988), and it is a good geochemical tracer of subsurface magma degassing, since its low solubility in silicate melts at low and moderate pressure (GERLACH and GRAEBER, 1985). Natural emissions of CO2 have different sources: mantle, carbonate metamorphism, descomposition of organic matter and biological activity (IRWIN and BARNES, 1980) and active faults favor gas leaks because they are preferential paths for crustal and subcrustal gases (IRWIN and BARNES, 1980; SUGISAKI et al., 1983; KLUSMAN, 1993; GIAMMANCO et al., 1998; KING, 1996; KING et al., 2006). Areas with high CO2 discharges can indicate high pore pressure at depth and might be a tool to identify potential seismic regions (ROJSTACZER et al., 1995; CASTAGNOLO et al., 2001; SPICAK and HORALEK, 2001). Relatively high CO2 fluxes correlate with areas that show deep fractures or faults with emissions of CO2 from magmatic reservoirs or decarbonation processes (TOUTAIN and BAUBRON, 1999) and increases of diffuse CO2 emissions related to seismic events and volcanic activity have been reported (HERNA´NDEZ et al., 2001b; ROGIE et al., 2001; SALAZAR et al., 2002; CARAPEZZA et al., 2004; PE´REZ et al., 2005). In order to improve the volcanic surveillance program of El Hierro Island and to provide a multidisciplinary approach, a continuous geochemical station to measure CO2 efflux was installed on September 2003 in Llanos de Guille´n, the interception center of the three volcanic-rift zones of the island, with the aim of detecting changes in the diffuse emission of CO2 related to the seismic or volcanic activity. Monitoring of CO2 efflux has demonstrated to be a useful tool to forecast precursory signals of volcanic eruptions and seismic events. HERNA´NDEZ et al., (2001b) reported an increase from 120 to 240 t/d on the CO2 efflux six months before the volcanic eruption of the Usu volcano, Japan, which occurred in 2000. CARAPEZZA et al. (2004) observed a significant increase of nearly double the maximum CO2 efflux values measured previously by an automatic geochemical station one week before the 2002 Stromboli eruption, Italy. SALAZAR et al. (2002) observed anomalous changes in the diffuse emission of carbon dioxide before some of the aftershocks of the 13 February 2001 El Salvador earthquake. PE´REZ and HERNA´NDEZ, 2005 and PE´REZ et al., 2006 have reported significant increases in a CO2 efflux time series prior to seismic events, as the increase observed from approximately 16 g m-2 d-1 to 270 g m-2 d-1, in the carbon dioxide efflux values measured in an automatic geochemical station nine days before the January 2002 short temp unrest occurred at San Miguel volcano, El Salvdor. In the last 15 years, the Instituto Geogra´fico Nacional (IGN) has reported the occurrence of several seismic events in and around El Hierro Island. Figure 2 shows the number of earthquakes registered in and around El Hierro Island since 1993. An
Vol. 165, 2008
Changes in the Diffuse CO2 Emission
97
anomalous increase in the seismic activity occurred in 2004. Unfortunately, no mechanism information is available for these earthquakes due to the characteristics of the seimic network of IGN in the Canary Islands.
2. Procedures and Methods The automatic geochemical station (EHI01) to measure the CO2 efflux was installed at Llanos de Guille´n, in the center of El Hierro Island (Latitude: N 27 420 58.2@; Longitude: W 18 010 8.8@) on September 25, 2003. Previous CO2 efflux surveys covering the entire island indicated that the selected location for the automatic station shows one of the highest CO2 efflux values measured in El Hierro Island (MART´ıNEZZUBIETA, 2001; PADRo´N et al., 2006). Moreover, the place is located at the interception center of the three volcanic rifts of the El Hierro Island. The station measures on an hourly basis the CO2 and H2S efflux, the CO2 and H2S air concentrations, the soil water content and temperature and the atmospheric parameters: wind speed and direction, air temperature and humidity and barometric pressure. The meteorological parameters together with the air CO2 concentration are measured 1 m above the ground and the soil water content and soil temperature are measured 40-cm deep, and recorded contemporaneously with CO2 efflux. On October 5, 2004, a rain gauge was also installed in the geochemical station. Both CO2 and H2S diffuse fluxes are estimated according to the accumulation chamber method (PARKINSON, 1981) by means of a nondispersive specrtophotometer (LICOR Li-820) with a 2000 ppm span cell and a ¨ GER Polytron II, respectively. The geochemical station is powered by a solar cell DRA panel and a battery. Each CO2 and H2S efflux measurement starts when the open side of the chamber is placed onto a fixed collar in the soil surface. A pump allows the air contained in the chamber to circulate through the NDIR spectrophotometer and then back into the chamber. To verify the performances and the reliability of this method, several calibration tests were made in the laboratory and the accuracy was estimated to be ±10%. Each hour the station also measures the soil temperature and water content and the meteorological parameters. All the data are stored on flash memory and radio-telemetered to ITER.
3. Results and Discussion During 2004 a total of thirteen seismic events were registered by the seismic network of IGN in and around El Hierro Island. The locations of these seismic events are shown in Figure 3. A time series of the total 6,385 measured data of CO2 efflux, wind speed, soil water content and temperature, air humidity and temperature and barometric pressure during 2004 is shown in Figure 4. A 48-hour moving average is also plotted for CO2 efflux, wind speed, air humidity and temperature and barometric pressure time series.
E. Padro´n et al.
98
Pure appl. geophys.,
Due to instrumental and telemetry problems, the time series has a 27.1% of missing data, with the main lag of data occurring between June 8 to July 29. Table 1 summarizes the results of the total recorded data. The CO2 efflux ranged between nondetectable values to 53.1 g m-2 d-1, with an average value of 12.5 g m-2 d-1. The detection limit of the automatic station has been estimated to be 0.5 g m-2 d-1. During the period of study, the H2S efflux values were always below the detection limit of the instrument (<1.5 g m-2 d-1). Inspection of the CO2 efflux time series shows four main relatively anomalous increases in the degassing rate. The first increase (A in Fig. 4) began approximately on March 10 and reached a maximum value of 52.1 g m-2 d-1 on March 18, five days before a seismic event of magnitude 2.3 occurred near El Hierro Island (Fig. 5). The earthquake was located 24-km deep and its epicenter at 12.6 km from EHI01. The second increase (B in Fig. 4) started approximately on April 6, 2004 and reached a maximum value of 46.3 g m-2 d-1 on April 7 (Fig. 5). The third increase (C in Fig. 4) in the CO2 efflux was recorded on October 5 and reached a maximum value of 43.8 g m-2 d-1 on October 14, followed by a seismic event of magnitude 1.3 which occurred on October 15 (Fig. 6). The last increase (D in Fig. 4) started on October 16, 2004 and reached a maximum value of 43.3 g m-2 d-1 on October 21, the same day as the occurrence of a seismic event magnitude 1.5 near El Hierro (Fig. 6). Between April 15 and May 25, a total of ten seismic events occurred inside and near El Hierro Island, with magnitudes between 1.0 and 2.2. However, there was no significant increase in the CO2 emission prior to these seismic events. The observed increases in the CO2 efflux at EHI01 seem uncorrelated with significant changes in any of the other parameters recorded by the automatic geochemical station. Short-temp. CO2 efflux changes driven by meteorological fluctuations have been reported Table 1 Statistical summary of the variables measured by the automatic geochemical station EHI0 in El Hierro during 2004
-2
-1
CO2 efflux (g m d ) H2S efflux (mg m-2 d-1) Air Humidity (%) Air Temperature (C) Barometric Pressure (HPa) Pumping flow (cm3/min) Soil temperature (C) Soil water content (%) Wind direction (N) Wind Speed (m/s) Pluviometry (mm/h) Air CO2 concentration (ppm) Air H2S concentration (ppm)
Mean
Maximum
Minimum
Median
12.5 103.4 64.4 12.6 889.6 994.9 15.7 18.8 164.7 1.8 0.2 353.4 n.d.
53.1 557.0 99.6 33.4 898.3 1385.1 25.6 54.8 359.0 9.2 41.4 1108.6 0.4
n.d. n.d. 6.7 2.2 861.5 0 10.4 6.9 0 0 0 146.3 n.d.
8.9 47.2 78.1 11.9 890.3 743.9 13.7 18.3 125.0 1.4 0 347.0 n.d.
Non detected values (n.d.) were bellow the detection limit of the instrument.
Vol. 165, 2008
Changes in the Diffuse CO2 Emission
99
Figure 1 Geographical localization of El Hierro in the Canary Islands.
in other volcanic systems (SALAZAR et al., 2000, 2002; PADRo´N et al., 2001; ROGIE et al., 2001; GRANIERI et al., 2003). Semidiurnal fluctuations inversely correlated with the barometric pressure variation and small increases correlated with soil water content have
Figure 2 Evolution of the seismicity registered by Instituto Geogra´fico Nacional since 1993 in and around El Hierro Island.
100
E. Padro´n et al.
Pure appl. geophys.,
Figure 3 Location of the seismic activity during 2004 in and around El Hierro Island.
been observed in the CO2 efflux time series at the station site. Significant changes with the wind speed have not been observed, probably due to the relatively low wind speeds recorded at the station place. However, the observed changes in the CO2 efflux cannot be explained in terms of such meteorological fluctuations.
Vol. 165, 2008
Changes in the Diffuse CO2 Emission
A
B
101
C D
Figure 4 Time series of the measured CO2 efflux and soil and meteorological parameters recorded during 2004 at the geochemical station, El Hierro Island. Moving average of 48 h is also displayed. A, B, C and D dots lines indicate the start of the four significant CO2 efflux increases recorded.
3.1. Filtering the Automatic Station Data In order to check the temporal variability and its possible dependence with external variables, the soil CO2 efflux was differenced one time to obtain a substantially greater stationary time series. Figure 7 shows the Fast Fourier Transform of the resulting time
102
E. Padro´n et al.
Pure appl. geophys.,
Figure 5 Time series of the measured CO2 efflux with a 48 h moving average and barometric pressure at the geochemical station together with the seismic activity during the first part of 2004, El Hierro Island. Arrows indicate the occurrence of seismic events.
series, with the typical diurnal and semidiurnal cycles (12 and 24 h-periods, respectively). Spectral coherences between the soil CO2 efflux and barometric pressure and air temperature were also observed, yielding significant peaks at 12 and 24 h (Fig. 8). These coherences suggest that short-time fluctuations in the diffuse CO2 emission in the observation site are partially driven by meteorological parameters. The observed fluctuations on the 24 h and 12 h periods shown in Figure 8 indicate that diurnal and
Vol. 165, 2008
Changes in the Diffuse CO2 Emission
103
Figure 6 Time series of the measured CO2 efflux with a 48 h moving average and barometric pressure at the geochemical station together with the seismic activity during the second part of 2004, El Hierro Island. Arrows indicate the occurrence of seismic events.
semidiurnal fluctuations in the CO2 efflux time series are partially explained in terms of air temperature and barometric pressure fluctuations. Multivariate Regression Analysis (MRA) was also used to isolate the response of the CO2 efflux to the externally measured variables. MRA is used to predict the dependent variable value as a function of relevant explanatory variables. In this case, we used as external variables the barometric pressure, wind speed and direction, air temperature and relative humidity, soil temperature and water content and power supply
104
E. Padro´n et al.
Pure appl. geophys.,
Figure 7 Fast Fourier Transform of the resulting CO2 efflux time series recorded during 2004 at the geochemical station, El Hierro Island.
voltage. MRA is useful to delineate the relations between the CO2 efflux and external factors measured on the geochemical station site. The analysis built a linear model in which the dependent variable is the CO2 efflux and the independent variables are the external factors measured at the station. Results of MRA provide an understanding of the percentage of the variability in the dependent variable which is explained by the selected set of independent variables. For the selected independent variables, the square of the multiple regression coefficients was 0.19, which means that about 19% of the variability in the estimated soil CO2 efflux was explained by the regression model. The statistical significance level of the regression model (p value) showed a value lower than 0.0001 for all the variables less wind speed, indicating a highly significant model. Figure 9 shows the original time series predicted and the residuals data. An inspection of the residual time series shows the four increases of the original CO2 efflux time series, indicating that these increases are not explained by changes in the external parameters measured in EHI01. The maximum values given in the predicted time series are related to meteorological parameters and changes in the soil properties, such as soil water content increases. 3.2. Forecasting the Seismic Events Since the observed changes in the CO2 efflux always occurred a few days or hours before the seismic events, an attempt to forecast the seismicity was done by means of the Material Failure Forecast Method (FFM) (VOIGHT, 1988, 1989; CORNELIUS and VOIGHT, 1995). VOIGHT (1988, 1989) proposed the FFM, based on his work during the Mount St. Helens eruption in 1981. The method uses the rock failure as a fundamental cause for
Vol. 165, 2008
Changes in the Diffuse CO2 Emission
105
(A)
(B)
Figure 8 Spectral coherences calculated using a 41 elements Tukey-Hamming window for diffuse CO2 emission rate and (A) air temperature and (B) barometric pressure.
most precursory activity, which can lead to a volcanic eruption or seismic event, and is based on the empirical failure material model: The rate of change of some observable parameters before a volcanic eruption or a seismic event follows the following equation:
106
E. Padro´n et al.
Pure appl. geophys.,
Figure 9 MRA results for the 2004 CO2 efflux data recorded at the geochemical station of El Hierro with a 48 h moving average.
d2 X dt2
a dX ¼A ; dt
ð1Þ
where X is any characteristic parameter (strain, rotation, traslation, seismic energy released,…) and A and a are two different constants. The method finally describes terminal failure of rocks, metals, etc. If a time series of a precursor seems to be representative of a
Vol. 165, 2008
Changes in the Diffuse CO2 Emission
107
solution for equation (1), it would be possible to determine the time of the event extrapolating the descending section of the inverse precursor time series to the time axis (CORNELIUS and VOIGHT, 1995). The time of a volcanic eruption or a seismic event is derived from the time of failure implied by the accelerating rate of the measured precursor. VOIGHT (1988) applied this method successfully to forecast volcanic eruptions (Mount St. Helens, USA, September 1981 and March 1982, and Bezymianny, Russia, April 1960), and for large landslides (Mount Toc, Italy, 1963). The method has also been applied by other authors on other volcanic systems (YAMAOKA, 1993; DE LA CRUZ-REINA and REYESDA´VILA, 2001; ENDO and MURRAY, 1991; ORT´ıZ et al., 2003). The application of the FFM method to the CO2 efflux time series was firstly reported by PE´REZ et al. (2005). PE´REZ and HERNA´NDEZ (2007) successfully applied the FFM to a precursory CO2 efflux increase prior to a seismic event of magnitude 2.7 at Tenerife, Canary Islands. Similar results can be obtained applying the FFM to the data reported by HERNA´NDEZ et al. (2001b), where the authors measured the total diffuse emission of CO2 released by the Usu volcano in September 1998 and September 1999, showing an increase from 120 to 340 t/d. The results obtained by the application of the FFM method for the possible eruption date agree excellently with March 2000, when the Usu eruption took place. For this case, the release of volcanic gases and mainly carbon dioxide, seems to control the eruption process. In the case of CO2 efflux, the term of the equation (1) is the change with time of the diffuse degassing rate. The FFM graphical technique is based on an inverse representation of the characteristic parameter rate ð1=UCO2 Þ versus time. The volcanic eruption or seismic event time is found by simple extrapolation of the data set or linear fit towards the time axis. At this time the parameter would reach an infinite value. In this work the FFM method was only successfully applied to the first anomalous CO2 emission rate increase. To filter the external atmospheric influence in the time series, we used the difference between the predicted values by the MRA model and the observed (real) values of the CO2 efflux time series. These differences can be negative if the data predicted by the MRA model is higher than the real value. The fluctuations in the residual data are essentially produced by variables which are not being measured in the automatic station; such as the deep contribution from the volcanic-tectonic environment of the island. This deep contribution in the negative residual data is negligible because it is considered to be completely explained in terms of the atmospheric and the soil parameters that have been monitored. Inspection of the residual data shows that the increase started on March 12th at 9:00 hours (Figure 10). The negative values were excluded from the analysis since they are completely explained in terms of the atmospheric parameters. Figure 10(B) depicts the inverse rate of CO2 efflux data during the selected period toward the time axis. A least-squareslinear fit is extended toward the time axis showing an interception approximately on March 23 at 18:00 h, which means only three hours of delay with the occurrence of the magnitude 2.3 seismic event. These results suggest that the residually selected period of the CO2 efflux time series obtained after MRA can be taken as a solution of the differential equation (1). The upper and lower 95% confidence limits of the linear fit are
E. Padro´n et al.
108
Pure appl. geophys.,
(A)
(B)
Figure 10 Application of the FFM for the forecasting of the seismic events which occurred in and around El Hierro Island during 2004. (A) Residuals of the MRA for the March 8–23 period. (B) Inverse rate of the CO2 efflux data during the selected time period toward the time axis. The upper and lower 95% confidence limits of the linear fit are also shown. Arrows indicate the interception of the linear fit and 95% confidence limits.
also shown in Figure 10(B). The interception of the confidence limits with the time axis allows us to compute a 95% confidence time interval of the interception between March 22, at 9:30 h and March 25, at 21:00 h.
Vol. 165, 2008
Changes in the Diffuse CO2 Emission
109
3.3. Relation between Seismic Events and the Precursory CO2 Efflux Signals Figure 11(A) shows a plot of the time lag between the start of the anomalous increase in the CO2 efflux and the occurrence of the seismic events versus the epicentral distances. An inverse correlation between both parameters is observed. Figure 11(B) offers a positive correlation between the maximum CO2 efflux value measured at each precursory signal and the ratio Magnitude/Epicentral Distance (M/Ed) for each related seismic event. These results agree with the idea that the geochemical station should be more sensitive for closer and higher magnitude seismic events. TOUTAIN and BAUBRON (1999) did not observe a similar result when they studied the correlation between the amplitude of different geochemical precursors and the epicentral distance of earthquakes. A positive correlation between the time lag (difference between the start of the CO2 efflux increase and the occurrence of the seismic events) and the ratio M/Ed is also observed (Fig. 11 (C)). In the case of shorter epicentral distances and higher magnitude earthquakes, the increase in the CO2 efflux should start earlier. However, no correlation was found between the duration of the precursory geochemical signal and the epicentral distance neither seismic magnitude. On the contrary, TOUTAIN and BAUBRON (1999) found a positive correlation between maximum duration of long-term anomalies and both magnitudes and epicentral distances. Strain-induced vertical fluids flow might be the origin of these anomalous increases in the CO2 efflux, as was proposed by KING (1978), for the soil radon anomalies found in the San Andreas fault. It has been widely accepted that fluids play an important role in the faulting mechanism (SHI and WANG, 1984/85; SPICAK and HORALEK, 2001). Crustal discontinuities with relatively high vertical permeability are preferential paths for crustal and subcrustal gases to escape towards the surface. This fluid circulation could explain the release of seismic energy through the occurrence of seismicity some days or even hours after these observed increases in the CO2 efflux. The deformation of the crust in the tectonic generation of an earthquake may force the fluids contained in the pores and fractures to move to different locations, sometimes increasing the gas concentration and its flux in a sensitive point (KING et al., 2006). This movement of chemical compounds may originate anomalous concentrations of that chemical specie, as was observed by TSUNOGAI and WAKITA (1995) for the disastrous magnitude 7.2 Kobe earthquake in 1995. The distance from the earthquake to the sensitive point has a direct relation with the time lag between the gas anomaly reaching the sensitive point and the quake, as can be observed in Figures 11(A) and (C) in the case described here. The greater deformation generated during the earthquake and the closer to the sensitive point, the greater this gas emission anomaly would be (Fig. 11B). Increasing the high pore pressure confined within a seismogenic zone can also enhance the stress concentration beneath a seismogenic layer. Similar phenomena were described by SALAZAR et al. (2002) at San Vicente volcano, El Salvador, Central America, where a significant increase in CO2 efflux rate was mainly driven by strain changes prior to a 5.1 magnitude earthquake which occurred
Time Lag (days) Maximum CO2 efflux value (g m-2 d-1)
Time Lag (days)
110 E. Padro´n et al.
(A)
(B)
(C)
Pure appl. geophys.,
Vol. 165, 2008
Changes in the Diffuse CO2 Emission
111
b
Figure 11 (A) Plot of the time lag between the start of the anomalous increase in the CO2 efflux and the occurrence of the seismic event, versus the epicentral distances, (B) maximum CO2 efflux value measured at each anomalous increase versus the ratio magnitude/epicentral distance for each related seismic event and (C) time lag between the start of the anomalous increase in the CO2 efflux and the occurrence of the seismic event versus the ratio magnitude/epicentral distance.
25 km away from the monitoring point and by PE´REZ et al. (2006) where a significant CO2 efflux increase was observed in an automatic geochemical station nine days before a short-time unrest occurred at San Miguel volcano, El Salvador. However, remarks on the observed relation between seismicity and diffuse degassing at El Hierro island must be made because there are unsuccessfully explained facts regarding the results shown here: — The absence of diffuse degassing precursors for the earthquakes occurred between April 15 and May 25. The short temporal window data at EHI01 geochemical station and the lack of information about the triggering mechanism of the earthquakes make it difficult to reach a satisfactory explanation regarding this absence of precursors. One possible explanation would result if the second increment of CO2 registered at the geochemical station is considered as a precursor of this seismic swarm which occurred in May. It is important to note that at this moment there are not enough available data that allow us to build satisfactory relationships between diffuse degassing precursors and the volcano-tectonic environment of El Hierro Island. — Although the results shown here indicate that diffuse degassing studies are promising tools for seismic monitoring studies, additional and more extensive studies are needed to validate the application of the FFM method to forecast a seismic event with the diffuse degassing studies. In fact, there are only such good results for the first earthquake of 2004 at El Hierro. The relatively low magnitude of the few earthquakes registered at El Hierro Island could be related to this unsuccessful approach.
4. Conclusions Four significant increases in the diffuse CO2 emission rate were observed few hours before various seismic events which occurred during 2004 in and around El Hierro Island. The results obtained after the application of MRA to the time series which were recorded at the automatic geochemical station EHI01, and the successful application of the FFM model to forecast the first seismic event, together with the correlation observed between the four anomalous increases with the magnitudes and epicentral distances of the seismic events, seem to indicate a close relationship between the diffuse CO2 efflux and the seismicity which occurred in and around El Hierro Island. Although short-temp fluctuations in the diffuse CO2 emission at the observation site are partially driven by meteorological parameters, the main CO2 efflux changes were not driven by fluctuations
112
E. Padro´n et al.
Pure appl. geophys.,
of meteorological variables such as wind speed or barometric pressure and seem clearly to be associated with fluid pressure fluctuations in the volcanic system. Intrusion of fluids and its migration through porous rocks might cause changes in pore pressure and trigger the seismicity. These results demonstrated the potential of applying continuous monitoring of soil CO2 efflux to improve and optimize the detection of early warning signals of future seismic events at El Hierro as well as in other active volcanic systems. Further observations are needed to verify the existence of a close relationship between the diffuse CO2 emission rate and the occurrence of earthquakes.
Acknowledgements Funds provided by the projects ALERTA and ALERTA II (financially supported by INTERREG IIIB Azores-Canaries-Madeira), Direccio´n General de Universidades e Investigacio´n of the Canary Islands Government under the project PI2001/025, Cabildo Insular de El Hierro and Canary Islands Government supported this work.
REFERENCES BARNES, I., EVANS, W.C., and WHITE, D. (1988), The role of mantle CO2 in volcanism, Appl. Geochem. 3, 281– 285. CARAPEZZA, M.L., INGUAGGIATO, S., BRUSCA, L., and LONGO, M. (2004), Geochemical precursors of the activity of an open-conduit volcano: The Stromboli 2002–2003 eruptive events, Geophys. Res. Lett. 31, L07620, doi:10.1029/2004GL019614. CASTAGNOLO, D., GAETA, F.S., DE NATALE, G., PELUSO, F., MASTROLORENZO, G., TROISE, C., PINGUE, F., and MITA, D.G. (2001), Campi Flegrei unrest episodes and possible evolution towards critical phenomena, J. Volcan. Geotherm. Res. 109, 13–40. CARRACEDO, J.C., DAY, S., GUILLOU, H., and PE´REZ TORRADO, F. (1997), El Hierro Geological Excursio´n Handbook. Estacio´n Volcanologica de Canarias and Universidad de Las Palmas, Tenerife/Gran Canaria, 43 pp. CHIODINI, G., CIONI R., GUIDI M., RACO B., and MARINI L. (1998), Soil CO2 flux measurements in volcanic and geothermal areas, Appl. Geochem. 13, 543–552. CORNELIUS, R. and VOIGHT, B. (1995), Graphical and PC-software analysis of volcano eruption precursors according to the Materials Failure Forecasting Method (FFM), J. Volcanol. Geotherm. Res. 64, 295–320. DE LA CRUZ-REINA, S. and REYES-DA´VILA, G. (2001), A model to describe precursory material-failure phenomena: Application to short-term forecasting at Colima volcano, Mexico, Bull. Volcanol. 63, 297–308. ENDO, T.E. and MURRAY T. (1991), Real-time seismic amplitude measurement RSAM. A volcano monitoring and prediction tool, Bull. Volcanol. 53, 533–54. FRONDINI, F., CHIODINI, G., CALIRO, S., CARDELINI, C., GRANIERI, D., and VENTURA, G. (2004), Diffuse CO2 degassing at Vesuvio, Italy, Bull. Volcanol. 66, 642-651. GEE, M.J.R., MASSON, D.G., WATTS, A.B., and MITCHELL, N.C. (2001), Offshore continuation of volcanic rift zones, El Hierro, Canary Islands, J. Volcanol. Geotherm. Res 105, 107–119. GERLACH, T.M. and GRAEBER, E.J. (1985), Volatile budget of Kilauea volcano, 313, no. 6000, 273–277. GIAMMANCO, S., GURRIERI, S., and VALENZA, M. (1998), Anomalous soil CO2 degassing in relation to faults and eruptive fissures on Mount Etna (Sicily, Italy), Bull. Volcanol. 60, 4, 252–259, doi 10.1007/s004450050231. GRANIERI, D., CHIODINI, G., MARZOCCHI, W., and AVINO, R. (2003), Continuous monitoring of CO2 soil diffuse degassing at Phlegraean Fields (Italy): Influence of environmental and volcanic parameters, Earth Planet. Sci. Lett. 212, 167–179.
Vol. 165, 2008
Changes in the Diffuse CO2 Emission
113
GRANIERI, D., CARAPEZZA, M.L., CHIODINI, G., AVINO, R., CALIRO, S., RANALDI, M., RICCI, T., and TARCHINI, L. (2006), Correlated increase in CO2 fumarolic content and diffuse emission from La Fossa crater (Vulcano, Italy): Evidence of volcanic unrest or increasing gas release from a stationary deep magma body?, Geophys. Res. Lett. 33, L13316, doi:10.1029/2006GL026460. GUILLOU, H., CARRACEDO, J.C., PE´REZ-TORRADO, F., and RODR´ıGUEZ BADIOLA, E. (1996), K-Ar ages and magnetic stratigraphy of a hotspot-induced, fast-grown oceanic island: El Hierro, Canary Islands, J. Volcanol. Geotherm. Res. 73, 141–155. HERNA´NDEZ, P.A., NATALE, G., TSUNOMORI, F., SUGIYAMA, K., ITO, T., NOTSU, K., OKADA, H., and PE´REZ, N.M. (2001a), Preliminary results of diffuse emissions of CO2 and soil gas pressure gradient measurements at Tarumae volcano, Japan, Bull. Volcanol. Soc. Jpn. 46, 121– 125. HERNA´NDEZ, P.A., NOTSU, K., SALAZAR, J.M., MORI, T., NATALE, G., OKADA, H., VIRGILI, G., SHIMOIKE, Y., SATO, M., and PE´REZ, N.M. (2001b), Carbon dioxide degassing by advective flow from Usu volcano, Japan, Science 292, 83–86. HERNA´NDEZ, P.A., SALAZAR, J.M., SHIMOIKE, Y., MORI, T., NOTSU, K., and PE´REZ, N.M. (2001c), Diffuse emission of CO2 from Miyakejima volcano, Japan, Chem. Geology 177, 175–185. HERNA´NDEZ, P.A., NOTSU, K., TSURUMI, M., MORI, T., OHNO, M., SHIMOIKE, Y., SALAZAR, J., and PE´REZ, N.M. (2003), Carbon dioxide emissions from soils at Hakkoda, north Japan, J. Geophys. Res. 108 (B4), 2210. HERNA´NDEZ, P. A., NOTSU, K., OKADA, H., MORI, T., SATO, M., BARAHONA, F., and PE´REZ, N.M. (2006), Diffuse emission of CO2 from Showa-Shinzan, Hokkaido, Japan: A sign of volcanic dome degassing, Pure Appl. Geophys. 163, 869–881. HERNA´NDEZ PACHECO (1982), Sobre una posible erupcio´n en 1793 en la isla de El Hierro, Canarias (On a posible eruption in 1793 in the island of El Hierro, Canary Islands), Estudios Geol. 38, 15–25. IRWIN, W.P. and BARNES, I. (1980), Tectonic relations of carbon dioxide discharges and earthquakes, J. Geophys. Res. 85, 3115–3121. KING, C.Y. (1978), Radon emanation on San Andreas fault, Nature 271, 516–519. KING, C.Y., KING, B.S., EVANS, W.C., and ZHANG, W. (1996), Spatial radon anomalies on active faults in California, Appl. Geochem. 11, 497–510. KING, C.Y., ZHANG, W., and ZHANG, Z. (2006), Earthquake-induced groundwater and gas changes, Pure Appl. Geophys. 163, 633–645. KLUSMAN, R.W., Soil Gas and Related Methods for Natural Resource Exploration (John Wiley and Son, New York 1993). MART´ıNEZ-ZUBIETA, A. (2001), Desgasificacio´n difusa de dio´xido de carbono, rado´n y metano en El Hierro, Islas Canarias, Master Thesis. Universidad de La Laguna, Spain, 82 pp. MASSON, D.G. (1996), Catastrophic collapse of the volcanic island of El Hierro 15 ka ago and the history of landslides in the Canary Islands, Geology 14, 231–234. NOTSU, K., SUGIYAMA, K., HOSOE, M., UEMURA, A., SHIMOIKE, Y., TSUNOMORI, F., SUMINO, H., YAMAMOTO, J, MORI, T., and HERNA´NDEZ, P.A. (2005), Diffuse CO2 efflux from Iwojima volcano, Izu-Ogasawara arc, Japan, J. Volcanol. Geotherm. Res. 139, 147–161. NOTSU, K., MORI, T., CHANCHAH DO VALE, S., KAGI, H., and ITO, T. (2006). Monitoring quiescent volcanoes by diffuse CO2 degassing: case study of Mt. Fuji, Japan, Pure Appl. Geophys. 163, 825–835. ORTIZ, R., MORENO, H., GARC´ıA, A., FUENTEALBA, G., ASTIZ, M., PEN˜A, P., SA´NCHEZ, N. and TA´RRAGA, M. (2003), Villarrica volcano (Chile): Characteristics of the volcanic tremor and forecasting of small explosions by means of a material failure method, J. Volcanol. Geotherm. Res. 128, 247–259. PADRo´N, E., SALAZAR, J.M.L., HERNA´NDEZ, P.A., and PE´REZ, N.M. (2001), Continuous monitoring of diffuse CO2 degassing from Cumbre Vieja volcano, La Palma, Canary Islands, EOS Transactions, Am. Geophys. Union 82, F1132. PADRo´N, E., PLATA, A., SANDOVAL, J., GONZA´LEZ, G., MELIA´N, G., HERNA´NDEZ, P.A. and PE´REZ, N.M. (2006), Temporal evolution of the diffuse carbon dioxide emission from Cumbre Vieja and El Hierro volcanic systems, Canary Islands, 5a Asamblea Hispano.Portuguesa de Geodesia y Geofı´sica, Sevilla, Spain, January 30th– February 3rd. PARKINSON K. J. (1981), An improved method for measuring soil respiration in the field, J. Appl. Ecol. 18, 221– 228. PE´REZ, N.M. and HERNA´NDEZ, P.A. (2005), Earthquake monitoring and prediction research in active volcanic areas by means of diffuse CO2 emission studies, Geophys. Res. Abst. 7, 10152.
114
E. Padro´n et al.
Pure appl. geophys.,
PE´REZ, N.M. and HERNA´NDEZ, N.M. (2007), Earthquake forecasting research in active volcanic areas by means of diffuse CO2 emission studies. In Geochemical Precursors for Earthquakes (eds Sen, P. and Das, N.K.) (Macmillan India Ltd.), pp 94–103. PE´REZ, N.M., HERNA´NDEZ, P.A., and ITER VOLCANO RESEARCH TEAM (2005), Premonitory geochemical and geophysical signatures of volcanic unrest at Tenerife, Canary Islands, 9th Gas Workshop of IAVCEI Commission on the Chemistry of Volcanic Gases. Southern Italy, May 1–10. ´ N, G., SALAZAR, P., PE´REZ N.M., HERNA´NDEZ P.A., PADRo´N, E., CARTAGENA, R., OLMOS, R., BARAHONA, F., MELIA and Lo´PEZ, D.L. (2006), Anomalous diffuse CO2 emission prior to the January 2002 short-term unrest at San Miguel volcano, El Salvador, Central America, Pure Appl. Geophys. 163, 883–896. ROGIE, J.D., KERRICK, D.M., SOREY, M.L., CHIODINI, G., and GALLOWAY, D.L. (2001), Dynamics of carbon dioxide emission at Mammoth Mountain, California, Earth Planet. Sci. Lett. 188, 535–541. ROJSTACZER, S., WOLF, S., and MICHEL, R. (1995), Permeability enhancement in the shallow crust as a cause of earthquake-induced hydrological changes, Nature 373, 237–239. SALAZAR, J.M.L., PE´REZ, N.M., HERNA´NDEZ, P.A., SORIANO, T., BARAHONA, F., OLMOS, R., CARTAGENA, R., Lo´PEZ, D.L., LIMA, R.N., MELIA´N, G., GALINDO, I., PADRo´N, E., SUMINO, H., and NOTSU, K. (2002), Precursory diffuse carbon dioxide signature related to a 5.1 magnitude earthquake in El Salvador, Central America, Earth Planet. Sci. Lett. 205, 81–89. SALAZAR, J.M.L., PE´REZ, N.M., and HERNA´NDEZ, P.A. (2000), Secular variations of soil CO2 flux levels at the summit cone of Teide volcano, Tenerife, Canary Islands, EOS Transactions, Am. Geophys. Union 81, F1317. SHI, X.J. and WANG, C.J. (1984/85), Instability on a weakening fault, Pure Appl. Geophys. 122, 478–491. SHIMOIKE, Y., KAZAHAYA, K., and SHINOHARA, H. (2002), Soil gas emisio´n of volcanic CO2 at Satsuma-Iwojima volcano, Japan. Earth Planets Space. 54, 239– 247. SPICAK, A. and HORALEK (2001), Possible role of fluids in the process of earthquake swarm generation in the West Bohemia/Vogtland seismoactive region, Tectonophysics 336, 151–161. SUGISAKI, R., IDO, M., TAKEDA, H., ISOBE, Y., HAYASHI, Y., NAKAMURA, N., SATAKE, H., and MIZUTANI, Y. (1983), Origin of hydrogen and carbon dioxide in fault gases and its relation to fault activity, J. Geology. 91, 239– 258. TOUTAIN, J.C. and BAUBRON, J.C. (1999), Gas geochemistry and seismotectonics: a review, Tectonophysics 301, 1–27. TSUNOGAI, U. and WAKITA, H. (1995), Precursory chemical changes in ground water: Kobe earthquake, Japan, Science 269, 5220, 61–63. VOIGHT, B. (1988), A method for prediction of volcanic eruptions, Nature 332, 10, 125–130. VOIGHT, B. (1989), A relation to describe rate-dependent material failure, Science 243, 200–203. YAMAOKA, K. (1993), State of stress and magma movement inferred from seismic activity during 1986–1987 eruption of Izu-Oshima volcano, Workshop on Volcanic Disaster Prevention, Panel on Volcanic Disaster Prevention under Japan-U.S. Science and Technology Agreement, 8–12 March 1992, Menlo Park, CA, pp. 16–18. (Received January 31, 2007, revised June 25, 2007, accepted June 27, 2007) Published Online First: February 1, 2008
To access this journal online: www.birkhauser.ch/pageoph
Pure appl. geophys. 165 (2008) 115–133 0033–4553/08/010115–19 DOI 10.1007/s00024-007-0290-8
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
SO2 Emission from Active Volcanoes Measured Simultaneously by COSPEC and mini-DOAS JOSE´ BARRANCOS,1 JOSE´ I. ROSELLO´,2 DAVID CALVO,1 ELEAZAR PADRO´N,1 GLADYS MELIA´N,1 PEDRO A. HERNA´NDEZ,1 NEMESIO M. PE´REZ,1 MILLA´N M. MILLA´N,2 and BO GALLE3
Abstract—We measured SO2 emission rate from six volcanoes in Latin America (Santa Ana, El Salvador; San Cristo´bal and Masaya, Nicaragua; Arenal and Poa´s, Costa Rica; Tungurahua and Sierra Negra, Ecuador) and from Mt. Etna, Italy, using two different remote sensing techniques: COSPEC (COrrelation SPECtrometer) and miniDOAS (miniaturized Differential Optical Absorption Spectroscopy). One of the goals of this study was to evaluate the differences in SO2 emission rates obtained by these two methods. The observed average SO2 emission rates measured during this study were 2688 td-1 from Tungurahua in July 2006, 2375 td-1 in September 2005 and 480 td-1 in February 2006 from Santa Ana, 1200 td-1 in May 2005 from Etna, 955 td-1 in March 2006 and 1165 td-1 in December 2006 from Masaya, 5400 td-1 of March 7, 2006 and 265 td-1 in March 2006 from San Cristobal, 113 td-1 in April 2006 from Arenal, 104 td-1 in April 2006 from Poa´s and 11 td-1 in July 2006 from Sierra Negra volcano. Most of the observed relative differences of SO2 emission measurements from COSPEC and miniDOAS were lower than 10%. Key words: COSPEC, mini-DOAS, SO2 fluxes, volcanoes.
1. Introduction Magma degassing is a very important process affecting the eruption style and the evolution of magmas in the crust. Among volcanic gases, H2O is the most abundant followed by CO2 and SO2, HCl, HF, H2, S2, H2S, CO, etc. (SYMONDS et al., 1994; GIGGENBACH, 1996). In recent years, scientists have made significant progress in monitoring and interpreting the significance of volcanic gas release (CHIODINI et al., 1993, TARAN et al., 2002 and SYMONDS et al., 2001). Monitoring of volcanic gas emissions has been performed conventionally by the analysis of fumarolic gas samples using ‘‘Giggenbach bottles’’ (GIGGENBACH and GOGUEL, 1989, GIGGENBACH, 1996, and GERLACH et al., 1986). This 1 Environmental Research Division, Instituto Tecnolo´gico y de Energı´as Renovables (ITER), 38611 Granadilla, S/C de Tenerife, Spain. E-mail: [email protected] 2 Fundacio´n Centro de Estudios Ambientales del Mediterra´neo, C/Charles R. Darwin, 14, 46980 Paterna,Valencia, Spain. 3 Department of Radio and Space Science, Chalmers University of Technology, Go¨teborg SE-412 96, Sweden.
116
J. Barrancos et al.
Pure appl. geophys.,
technique represents a potential hazard because it requires accessing directly to the fumaroles, steam vents, heated pools, etc. Remote sensing techniques considerably reduce the danger of the direct sampling for the geochemical monitoring of a volcano because it is not necessary to access directly to the gas emission point and allow the measurement of more representative integrated plume compositions (STOIBER et al., 1983; VANDAELE et al., 1994; EDMONDS et al., 2003a; GALLE et al., 2002; BURTON et al., 2004; MCGONIGLE et al., 2005; BOBROWSKI et al., 2006; MORI and BURTON, 2006; WILLIAMS-JONES et al., 2006). A complete temporal series of SO2 emissions can provide important information about the chemistry and dynamics of magma (EDMONDS et al., 2003a), the permeability and pressures in the inner conduit, and the role of the hydrothermal system (EDMONDS et al., 2003b). The emission rate of SO2 measured with COSPEC has been regarded as a basic monitoring parameter of degassing and eruptive activities (ANDRES et al., 1993). Some interesting examples are: Pinatubo, 1991 (HOFF, 1992; DAAG et al., 1996); Montserrat, 1995 (YOUNG et al., 1998); and Kilauea, 1979-present (SUTTON et al., 2001). COSPEC is still being used for monitoring SO2 fluxes in several volcanoes around the word, i.e., Miyakejima (Japan), Etna (Italy), Arenal (Costa Rica), Pacaya (Guatemala), Tungurahua (Ecuador). Recently, other spectrometric techniques have been incorporated to monitor the chemical composition and fluxes of volcanic plumes, i.e., Differential Optical Absorption systems (DOAS, miniDOAS, MAX-DOAS, IDOAS). In the present work, we report the results of SO2 emission measurements from several active volcanoes during the last 2 years by means of simultaneously using COSPEC and mini-DOAS in mobile-terrestrial position.
2. Geological Settings of Selected Volcanoes Figure 1 shows the location maps of the volcanoes selected for this study. The selected volcanoes for this study are Etna (Italy), Santa Ana (El Salvador), Masaya (Nicaragua), San Cristobal (Nicaragua), Arenal (Costa Rica), Poas (Costa Rica), Tungurahua (Ecuador) and Sierra Negra, Gala´pagos (Ecuador). During this study, volcanic eruptions occurred at Santa Ana (El Salvador), San Cristo´bal (Nicaragua), and Tungurahua (Ecuador) volcanoes. Etna (37.75N, 15.00E and 3323 m a.s.l.) is a stratovolcano located on the Italian Island of Sicily, which extends over 1200 km2. Etna is a young volcano that has been growing during the last 500,000 years. The SO2 emission rates from Mt. Etna depend mainly on its volcanic activity level, and the emission rates have been estimated to range between 1000 and 2000 td-1 for inter-eruptive periods, and more than 5000 td-1 during eruptive periods (BRUNO et al., 2001; PUGNAGHI et al., 2002, 2006). Santa Ana volcano (13.85N, 89.63W and 2381 m a.s.l.) is located 40 km west of San Salvador, El Salvador. It is a massive stratovolcano and contains an acid lake at its summit crater (0.5 km diameter) with an average surface temperature between 18–20C. This explosion crater was formed during the most recent eruption which took place in
Vol. 165, 2008
SO2 Emission from Active Volcanoes
117
Figure 1 Geographical location of the active volcano systems: (a) Santa Ana (El Salvador), San Cristo´bal, Masaya (Nicaragua), Arenal, Poa´s (Costa Rica), (b) Tungurahua (Ecuador) (c) Sierra Negra (Gala´pagos) and (d) Etna (Italy).
1904. Hot springs, gas bubbling and intense fumarolic emissions are observed along the shoreline of this crater lake. A volcanic plume, usually driven by the NE trades, can be seen rising to 500 m from the summit crater of the Santa Ana volcano (OLMOS et al., 2007 and BERNARD et al., 2004). Previous studies of SO2 emission from the crater of Santa Ana reported an average emission rate of 140 td-1 in February 2001 and January 2002 (RODR´ıGUEZ et al., 2004). Masaya (11.98N, 86.16W and 635 m a.s.l.), is a large basaltic shield volcano located about 25 km south of Managua, Nicaragua. It is composed of a group of calderas and craters. Inside Las Sierras caldera lies the Masaya volcano, basically a shallow shield formed from basaltic lavas and tephras. The Masaya caldera was formed about 2.5 Ky ago by an 8 km3 ignimbrite eruption (WILLIAMS, 1983). Inside this caldera two cones, Masaya and Nindiri cones have grown. The main degassing occurs from the Masaya pit crater, located centrally within the Masaya cone, and which is now vegetated but probably was formed after the late sixteenth century. Since 1993, the volcanic activity has changed, becoming more active with a continuous degassing process and some explosive
118
J. Barrancos et al.
Pure appl. geophys.,
episodes. Recent studies of SO2 emission at Masaya reported an average value of 800 td-1 (MATHER et al., 2006; DUFFELL et al., 2003; DELMELLE et al., 1999). San Cristo´bal (12.70N, 87.00W and 1745 m a.s.l.) is a stratovolcano with a crater of 5009600 m located about 100 km Northwest of Managua (Nicaragua). It is the youngest volcano of a volcanic complex with the same name, which forms part of the Maribios volcanic range. Volcanic activity at San Cristo´bal volcano is characterized by a strong fumarolic degassing, occasional explosions and incandescence at the crater. Recently a scanning-DOAS system has been installed to monitor the SO2 flux from the volcano. Studies carried out prior to 2004 reported an average SO2 emission rate of about 790 td-1 (MATHER et al., 2006). Arenal volcano (10.46N, 84.70W and 1633 m a.s.l.) is located in the northern part of Costa Rica. It is a young and very active stratovolcano belonging to the Arenal-Chato system, which is a 12-km-long volcanic system aligned in a SE-NW direction between two lines of active volcanoes, the Cordillera Central in the south and the Cordillera de Guanacaste in the north. It was reactivated by the eruption of 1968, and since then shows a continuous degassing process that extends to the present. Recent studies have reported average SO2 emission rates about 130 td-1 in March 1995 and March 1996 (WILLIAMSJONES et al., 2001) and 180 td-1 in March 2001 (ZIMMER et al., 2004). Poa´s (10.20N, 84.23W and 2709 m a.s.l.) is a basaltic-andesite stratovolcano, located about 30 km north of San Jose´, Costa Rica. It has three craters aligned along a N-S line and a crateric, acidic lake, very active with a shallow hydrothermal system. The most important previous reported activity occurred in March 24, 2006, with steam explosions that ejected mud and ballistics outside the crater rim. Previous studies reported average SO2 emission rates of 8.3 td-1 in March 2001 (ZIMMER et al., 2004; CASADEVALL et al., 1984). Tungurahua volcano (1.47S, 78.44W and 5023 m a.s.l.) is a steep-sided volcano located on the eastern Cordillera of the Ecuadorian Andes, Ecuador and about 125 km south of Ecuador’s capital, Quito. Currently, Tungurahua is one of the most active volcanoes in Ecuador, showing constant activity since 1999. After the beginning of a high volcanic activity with the occurrence of several volcanic explosions in October 1999, which produced a major ash out-fall, the activity has continued until May 2006, when the activity has increased and culminated in two violent eruptions on July 14 and August 16, 2006. The August 16th eruption was accompanied by a 10-km high ash cloud which later spread over an area of 740 by 180 km and, by pyroclastic flows resulting in seven deaths and destruction of several hamlets and roads on the western and northwestern slopes of the volcano. Previous studies estimated a typical SO2 emission rate lower than 300 td-1 over 2005 and 1000 td-1 according to the July 2006 monthly report of the Instituto Geofı´sico-EPN. Sierra Negra (0.83S, 91.17W and 1124 m a.s.l.) is one of the most important volcanoes in the Galapagos Islands, a ‘‘hot spot’’ basaltic archipelago, and is located at Isabela Island, the biggest of the Galapagos chain. The main feature of Sierra Negra volcano is the large caldera of 7 km910 km and 110 m deep, larger than any other caldera on Galapagos Islands (REYNOLDS et al., 1995; MUNRO and ROLAND, 1996).
Vol. 165, 2008
SO2 Emission from Active Volcanoes
119
Fumarolic activity and a small volcanic plume occurred at Mina Azufral, located at the west margin of an intracaldera horst bounded by a sinuous fault zone to the north, west, and south (REYNOLDS et al. 1995; GOFF et al., 2000). 3. Methodology 3.1. COSPEC V The COSPEC (see scheme on Fig. 2a) is a ‘mask correlation’ spectrometer designed for vertical or slant column measurements of SO2 and NO2 concentrations, using zenith sky-scattered sunlight (MOFFAT and MILLA´N, 1971; MILLA´N and HOFF, 1978). COSPEC V has been during recent years the reference instrument for the remote surveillance of volcanic plumes. The spectral radiant power coming from the zenith sky is collected by the telescope and focused onto an entrance slit located at the entrance of a polychromator. The radiation is dispersed by the diffraction grating and focused over the exit plane of the polychromator. A rotating disc is situated on the focal plane, and the radiation is transmitted selectively through four masks engraved sequentially on the disc. Each mask consists of an array of slits in the form of circular arcs, etched on an aluminum layer deposited on the quartz disc (MOFFAT and MILLA´N, 1971; MILLA´N and HOFF, 1978). The masks are separated into two pairs. Each pair has one mask whose slits correspond to some of the transmission maxima in the spectrum of the target gas (peak mask) and another mask whose slits correspond to some of the transmission minima in the spectrum of the target gas (trough mask). Behind the disc there is a photodetector, a photomultiplier (PMT) in the case of the COSPEC which measures the radiation. The PMT signal is processed to generate the COSPEC signal. The parameters of each mask are calculated taking into account the available backgrounds and their expected variations, the variations in the absorption coefficient wavelengths of the target gas, the expected range of concentrations to be measured, and the optical characteristics of the instrument (optical aberrations, diffraction effects, stray light, etc.) The COSPEC has an Automatic Gain Control (AGC) that fulfills the purpose of compensating for multiplicative background radiance fluctuations. The calibration is performed in situ by placing cells containing known amounts (ppmm) of gas in the path of the radiance dispersed by the instrument (‘high and low’) (NEWCOMB and MILLA´N, 1970; MOFFAT and MILLA´N, 1971). In our case, the COSPEC ‘high’ is 356 ppmm and ‘low’ is 92 ppmm. The best signal-to-noise ratio in the SO2 remote measurements is due to the optical and electronic COSPEC design. Some improvements have been made to the COSPEC electronics and optics to improve the signal-to-noise ratio of the instrument and permit its extended operation under lower sun elevation angles (MILLA´N and ROSELLO, 2005).
120
J. Barrancos et al.
Pure appl. geophys.,
Figure 2 Scheme of a) COSPEC (WEIBRING et al., 1998) and b) miniDOAS (GALLE et al., 2002).
In the present comparison the COSPEC has been used for mobile measurements by placing it on a moving platform (such as a car, boat, helicopter or aircraft) and traversing beneath, and roughly perpendicularly to, a volcanic plume; totally integrated SO2 concentration cross sections are thus obtained. These values are typically multiplied by estimated plume velocities to yield total source emission rates (e.g., in td-1 or Kg/s).
Vol. 165, 2008
SO2 Emission from Active Volcanoes
121
3.2. miniDOAS This instrument is based on an Ocean Optics USB2000 spectrometer and collects the light caught by the telescope and guided by the optical fiber towards the spectrometer, according to Figure 2b. This detector, which can be operated at ambient temperature, is a CCD-array of 2048 elements (13 lm (width) 9 200 lm (height)) treated for enhanced sensitivity below the 360 nm resolution, a 50 lm slit and spectral resolution of *0.6 nm over the wavelength range of 245–380 nm (GALLE et al., 2002). Although the reading noise would have to diminish as it increases the time of exposition, the residual noise does not decrease below a peak-to-peak value of 0.1–0.2%. The instrument is reduced in dimensions (89964934 mm) and weighs only 0.2 kg. In addition, it consumes only 1 W by USB port from the laptop, which also supports the data transfer (GALLE et al., 2002). A comparison of the specifications of both remote instruments is made in Table 1. For each traverse: ‘high’ and ‘low’ were made with COSPEC, and a ‘dark’ spectrum was collected before entering or after leaving the plume by blocking the telescope entrance for miniDOAS (this dark spectrum was subsequently subtracted from the skyspectra in order to correct for electronic offset and dark current). The measures were made as perpendicular to the plume as was possible from a car. The position was fixed by the GPS and registered at any moment with a frequency of five seconds. The software used for miniDOAS data acquisition was developed by the Technological University of Chalmers. Drawing at every moment the time vs. the GPS position of the vehicle, the SO2 measures obtained, and the wind speed (we assume that the pen moves at the same speed), and integrating the whole route, we obtain the value of the SO2 flux equation: Z ð1Þ /SO2 ðt d1 Þ ¼ 0:00023 SO2 ðppmÞ X V sin h; R
where 0.00023 constant arises from changes in units [(density of SO2 gas: 2.86910-3 gppm-1m3; STP correction factor: 273/293) ppmm3s-1 2.86910-3 gppm-1m-3 Table 1 Physical characteristics of both instruments
Weight Power Wavelength coverage Dimensions Minimum detection limit Field of view
COSPEC V
Mini-DOAS
20.5 kg 8 W on 12/24 V DC & 23 W on 15/230 V AC 300–315 nm, Resolution 0.2 nm
0.4 kg 1 W powered from USB port of laptop 245–400 nm (303–313 nm used to fit SO2), Resolution 0.6 nm 89964934 mm3 2.5 ppmm with 3 s. integration time (S.D.) 20 mrad
102953928 cm3 2.5 ppmm with 1 s. integration time 10930 mrad
122
J. Barrancos et al.
Pure appl. geophys.,
Table 2 a
SO2 emission data from seven active volcanoes. Number of measurements, bdifference between COSPEC and miniDOAS and nm means not measured Volcano
Etna Etna Etna Etna Etna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna SantaAna San Cristo´bal San Cristo´bal San Cristo´bal Masaya Masaya Masaya Masaya Masaya Masaya Masaya Arenal Arenal Poa´s Poa´s Poa´s Poa´s Sierra Negra Tungurahua Tungurahua Tungurahua Tungurahua
Date (dd/mm/ aaa)
COSPEC a
n Average (tons/ day)
Max (tons/ day)
b
miniDOAS Min (tons/ day)
09/05/2006 10/05/2006 11/05/2006 12/05/2006 13/05/2006 18/09/2005 20/09/2005 21/09/2005 22/09/2005 24/09/2005 25/09/2005 26/09/2005 27/09/2005 20/02/2006 21/02/2006 22/02/2006 23/02/2006 24/02/2006 26/02/2006 27/02/2006 28/02/2006 07/03/2006
3 4 nm nm nm 5 3 nm 3 4 8 3 3 nm 4 nm 4 nm 10 4 1 4
1360 2311
1460 2410
1310 2100
1472 1626
1780 1922
1025 1390
4624 2151 1502 3136 736
5987 3324 2495 3315 809
3563 1140 997 2867 687
679
810
405
414
641
292
400 396 354 5247
602 427 354 7240
11/03/2006
11
317
12/03/2006
4
a
n Average (tons/ day)
Max (tons/ day)
Min (tons/ day)
Difference (%)
319 346 354 2754
5 8 9 8 10 7 4 2 4 6 7 3 3 2 5 6 6 8 10 4 1 3
1177 1976 831 656 900 1409 1786 3558 4481 2418 1374 3344 768 297 579 787 396 564 391 418 365 5627
1709 2622 1237 1060 1285 1606 2082 4876 5625 3395 2158 3642 851 337 685 935 677 830 521 448 365 7208
839 1407 570 302 548 1123 1477 2240 3645 1338 932 2904 700 256 374 644 269 456 281 381 365 2645
486
212
12
316
470
181
0.32
224
283
157
4
201
284
128
10.15
13/03/2006 14/03/2006 15/03/2006 16/03/2006 06/12/2006 08/12/2006 10/12/2006 25/04/2006 26/04/2006 21/04/2006 24/04/2006 27/04/2006 30/04/2006 10/07/2006
7 1183 10 842 7 1018 7 1031 nm nm nm 4 154 nm nm nm nm nm nm
1952 1333 1318 1585
727 510 732 740
7 965 15 819 nm 5 1098 15 1458 6 990 14 1046 18 121 8 106 2 99 5 160 2 86 2 90 1 11
1374 1438
701 394
18.46 2.75
1605 2213 1379 1888 237 156 123 220 98 110 11
894 495 425 573 33 62 76 112 74 70 11
-6.55
21/07/2006 22/07/2006 23/07/2006 25/07/2006
nm nm nm nm
1000 2819 3191 5100
379 1263 3191 1956
225
89
3 8 1 6
604 1809 3191 3727
4.31 -9.79 3.08 -12.44 8.53 -6.63 -4.38 14.76 4.40 2.30 -5.53 -3.16 -7.25
21.48
Vol. 165, 2008
SO2 Emission from Active Volcanoes
123
Table 2 (Contd.) Volcano
Tungurahua Tungurahua Tungurahua Tungurahua Tungurahua
Date (dd/mm/ aaa)
26/07/2006 27/07/2006 28/07/2006 29/07/2006 30/07/2006
COSPEC a
n Average (tons/ day)
nm nm nm nm nm
Max (tons/ day)
b
miniDOAS Min (tons/ day)
a
n Average (tons/ day)
11 4 2 8 6
3094 4601 2374 1241 3569
Max (tons/ day) 5576 5448 2645 1598 5257
Min (tons/ day)
Difference (%)
1940 4005 2104 976 1602
910-6 tg-1 9 86400sd-1]. The SO2 flux is expressed in td-1, the wind speed (V) in m/s, the displacement (X) in meters, and h is the angle between V and X vectors (STOIBER et al., 1983). The wind data for the flux calculation throughout was taken from: http://www.arl.noaa.gov/ready/amet.html and measurements were made in stable meteorological conditions (clear skies) between 9:30 and 15:30 in local time (UTC).
4. Results and Discussions The field campaign was carried out with both instruments from May 2005 to December 2006. This work presents the results of the 399 SO2 emission measurements performed at different volcanoes (see Fig. 1). A total of 286 measurements were made with miniDOAS and 113 with COSPEC (given its lower portability). The instruments were mounted on a car (except for the case of Sierra Negra where measurements were performed by walking), and traverses under the plume were made through different ways. The results of these campaigns are summarized as follows: 4.1. Etna Volcano, May 2005 SO2 emissions from Etna were measured using the COSPEC along the road nearest the coast. The SO2 fluxes obtained during these campaigns are shown in Table 2. Using both instruments, estimations for Etna volcano emission rate during the measurement period were computed in an average value of 1200 td-1. From the values obtained, the maximum and minimum daily averages were 2600 and 300 td-1. The data of SO2 emission from Mt. Etna are not included in the comparative on Figure 8 nor in Table 2 because they were not simultaneous measurements.
124
J. Barrancos et al.
Pure appl. geophys.,
4.2. Santa Ana Volcano, September 2005 and February 2006 Measurements were carried out just before the October 1st, 2005 eruption, allowing estimation of a considerably larger SO2 emission rate than those estimated for intereruptive periods, namely between 30 and 280 td-1 (RODR´ıGUEZ et al., 2004; MATHER et al., 2006). The values are typical of the eruptive phases in which they were taken. For the pre-eruptive period, the average value was 2400 td-1, while the maximum was higher than 6000 td-1 and the minimum was 700 td-1. In the second period, the mean values were 480 td-1 and the standard deviation was lower than 164 td-1. The resulting values are shown in Table 2, and the temporal evolution is shown in Figures 3a and 3b. During these two campaigns, SO2 fluxes were measured simultaneously with COSPEC and miniDOAS 51 times. The results obtained by the COSPEC and miniDOAS allow us to make an intercomparison of the SO2 fluxes measured by both instruments. By graphing the SO2 flux values from the COSPEC against those from the miniDOAS and fitting them by linear regression through the zero point, we were able to visualize the SO2 emission differences between the two instruments. This result is shown in Figure 3c. The relative difference was around 1%, with some anomalous values of 15%. Also in the comparison, the slope (&1) and the correlation coefficient (0.99) indicate a good correlation between the measurements of both instruments. 4.3. Masaya Volcano, March and December 2006 This campaign was made in collaboration with INETER (Instituto Nacional de Estudios Territoriales, Nicaragua). COSPEC was only available for use during the first period, while miniDOAS was used in both phases. Both phases corresponded to intereruptive periods of relative volcanic calm, which explains why the fluxes were very similar with no great oscillations. In the first campaign, the mean values were 955 td-1, with a maximum of 1952 and a minimum of 400 td-1 for a total of 58 measurements, 22 of them performed simultaneously. In the second campaign, the mean values were 1165 td-1, with a maximum of 2200 td-1 and a minimum of 425 td-1 for a total of 35 measurements. Estimations in previous measurements made before 1997 computed around 800 td-1 of the SO2 emission rate (MATHER et al., 2006). The temporal evolution and the summary of the results are shown in Figures 4a and 4b and Table 2. The relative differences between both instruments, Table 2 and Figure 4c, show a mean value lower than 2%, and a slope approximately equal to 1, with an excellent correlation coefficient (0.97). 4.4. San Cristo´bal Volcano, March 2006 This campaign also was made in collaboration with INETER and coincided with the March 7 eruptive episode which involved substantial gas and ash emissions. The data
Vol. 165, 2008
SO2 Emission from Active Volcanoes
125
Figure 3 Temporal evolution of SO2 emission from Santa Ana volcano: a) September, 2005; b) February, 2006 and c) SO2 by COSPEC vs by miniDOAS.
obtained in our work are shown in Table 2 and Figure 5a. During the eruptive episode of March 7, 2006, the estimated SO2 mean rate was 5400 td-1; subsequently, the rate decreased to 270 td-1 four days later. These values of the SO2 emission differ slightly from the average value since 1997 of 690 td-1 reported by MATHER et al. (2006). On March 7, seven measurements were taken, 3 of them were carried out by both COSPEC and miniDOAS simultaneously. During the entire campaign, 38 measurements, 15
126
J. Barrancos et al.
Pure appl. geophys.,
Figure 4 Temporal evolution of SO2 emission from Masaya volcano: a) March, 2006; b) December, 2006 and c) SO2 by COSPEC vs. by miniDOAS.
simultaneous, were made. This set of measurements also allowed us to obtain the relative differences between the flux values obtained by both instruments. Figure 5b shows the linear fit of points for fluxes lower than 500 td-1. An excellent agreement was also found and similar fluxes were computed for both instruments, with a slope and regression coefficient equal to 1.
Vol. 165, 2008
SO2 Emission from Active Volcanoes
127
Figure 5 a) Temporal evolution of SO2 emission and b) SO2 by COSPEC vs. by miniDOAS from San Cristo´bal volcano.
4.5. Arenal and Poa´s Volcanoes, April 2006 This work was carried out in collaboration with the University of Costa Rica and the Costa Rican Institute of Electricity. In the case of Arenal volcano, a small plume from the summit crater is usually visible, however explosions of ash and gas also occur with certain regularity. The SO2 fluxes have shown a very irregular behavior, and strongly depend on the volcanic activity, frequently manifested by explosions. Nonetheless, we were able to make 26 points for Arenal. The daily values are shown in Table 2. For Arenal volcano, the SO2 emission rates ranged between 50 to 190 td-1, with a mean of 115 td-1 and a standard deviation of 45 td-1. This is very similar to 130 ± 60 td-1 obtained by WILLIAMS-JONES in March 1995 and March 1996 (WILLIAM-JONES et al., 2001) and 180 td-1 in March 2001 (ZIMMER et al., 2004). For Poa´s volcano (11 measurements), the SO2 emission rates ranged between 50 and 135 td-1 with a mean of 105 td-1 and a standard deviation of 35 td-1. This rate is very different to 8.3 td-1 in March 2001 by ZIMMER et al. (2004). In this campaign, simultaneous measurements could be carried out only at the Arenal volcano; adverse weather conditions did not allow a similar comparison at Poa´s volcano.
128
J. Barrancos et al.
Pure appl. geophys.,
Figure 6 SO2 by COSPEC vs. by miniDOAS for Arenal volcano.
Due to logistical problems, only four simultaneous measurements were possible at Arenal volcano. Comparing the SO2 fluxes from the two instruments for this volcano, and fitting the points, it is possible to see the adjustment in Figure 6. In this case, due to the limited number of points available, the agreement between COSPEC and mini-DOAS was not as good as the results shown before; the obtained slope was 1,05 with a regression coefficient of 0.97. 4.6. Tungurahua, July 2006 This work was carried out in collaboration with USFQ (University of San Francisco, Quito) and IG-EPN (Instituto Geofı´sico de la Escuela Polite´cnica Nacional). The measurement period extends from July 21 to the 30, in which, only miniDOAS was used. The rates observed during this study were much higher and typical of eruptive periods. The daily rates observed, with only one exception, were above 1000 td-1, with an average of 2690 td-1 and a maximum of 5600 td-1. The temporal evolution of these values is shown in Figure 7. 4.7. Sierra Negra, July 2006 The SO2 emission rate shown at this work was measured around Los Azufres fumarolic field, the largest fumarolic area in the Gala´pagos Islands, located on the west side of Sierra Negra caldera. The measurements were done by miniDOAS walking towards the N-E direction, perpendicularl to the plume direction. Due to the weather conditions only one traverse could be carried out and 11 td-1 were computed. 4.8. Comparison between COSPEC and miniDOAS An overall comparison between the results obtained by both instruments was made, including all the obtained values computed where COSPEC and mini-DOAS were used
Vol. 165, 2008
SO2 Emission from Active Volcanoes
129
Figure 7 Temporal evolution of SO2 emission from Tungurahua volcano.
simultaneously. The plot of all these values and the linear fit is shown on Figure 8. We conclude that, for a set with a significant number of points, both instruments measured very similar fluxes of SO2. An excellent agreement between the results obtained by both instruments was found: the linear fit shows a slope of 1.00 and a correlation coefficient of 0.99. Similar results were obtained by HORTON et al., (2005), at Kilauea, Hawaii, in the comparative COSPEC vs. FLYSPEC between March 2002 and February 2003. The observed relative differences were lower than the ones observed by ELIAS et al., (2006) (8%) also at Kilauea, Hawaii.
5. Conclusions We studied the SO2 emission from seven volcanoes located in different geological and tectonic environments, and with different levels of volcanic activity by means of
Figure 8 SO2 of COSPEC vs. miniDOAS from all volcanoes.
130
J. Barrancos et al.
Pure appl. geophys.,
COSPEC and miniDOAS techniques from September 2005 until December 2006. The mean SO2 emission rates measured at these volcanic systems were: 1200 td-1 for Etna volcano in May 2005; 2375 td-1 in September 2005 and 480 td-1 on February 2006 for Santa Ana volcano; between 955 td-1 March 2005 and 1165 td-1 December 2006 for Masaya volcano; 5400 td-1 during the volcanic eruption of March 7, 2006, and 265 td-1 during the following days for San Cristo´bal volcano; 104 td-1 and 113 tons/d in April 2006 for Poa´s and Arenal volcanoes, respectively; 11 td-1 and 2688 td-1 in July 2006 for Sierra Negra and Tungurahua (during the volcanic eruption of July 14) volcanoes. Those volcanoes with a stable visible plume during the field campaigns showed slight variations in the SO2 emission rate. In the case of San Cristo´bal, Santa Ana and Tungurahua volcanoes, which experienced eruptive phases during the measurements, relatively high pulses of SO2 were observed, with a higher variability in the SO2 emission rates. According to Table 2, the differences between the SO2 fluxes measured with COSPEC and miniDOAS are about 2%. The excellent correlation observed between both instruments (R2 ¼ 0.99) demonstrates that both instruments measure similar SO2 flux values. The combined use of remote sensors like COSPEC and miniDOAS with other techniques like the one described by SHINOHARA (2005), will allow estimation of the emission rates of other gas species in volcanic plumes and create a better understanding of the dynamics of volcanic degassing phenomena and the role of volcanic gases during periods of volcanic unrest.
Acknowledgements This research was supported by the Cabildo Insular de Tenerife, the Spanish-AIDAgency (AECI), and the projects CGL2005-07509-CLI and CGL2004-22023-E financed by the Spanish Ministry of Science and Education (MEC), ALERTA and ALERTA II financed by the EU Programme INTERREG III B Ac¸ores-Madeira-Canarias. We also thank the support of UES (Universidad de El Salvador, with special thanks to Rodolfo Olmos), INETER (Instituto Nicaragu¨ense de Estudios Territoriales, especially to Wilfried Strauch), UCR (Universidad de Costa Rica, particular thanks to Mario Ferna´ndez, Rau´l Mora and Carlos Ramı´rez), ICE (Instituto Costarricense de Electricidad, notable thanks to Guillermo Alvarado), USFQ (University of San Francisco, Quito, Ecuador, special thanks to Theofilos Toulkeridis) and IG-EPN (Instituto Geofı´sico, Escuela Polite´cnica Nacional, Ecuador, particular thanks to Hugo Yepes). Also we are grateful to DORSIVA project and NOAA for the wind data. We are thankful to Carlos Rudamas and Claudia Rivera for their assistance and constructive discussions during the measurements in Santa Ana and Masaya volcanoes.
Vol. 165, 2008
SO2 Emission from Active Volcanoes
131
REFERENCES ANDRES, R.J., ROSE, W.I., STOIBER, R.E., WILLIAMS, S.N., MAT´ıAS, O., and MORALES, R. (1993), A summary of sulfur dioxide emission rate measurements from Guatemalan volcanoes. Bull. Volcanol. 55, 379–388. BERNARD, A., ESCOBAR, C. D., MAZOT, A., and GUTIERREZ, R. E. (2004), The acid volcanic lake of Santa Ana volcano, El Salvador. In: ROSE W. I., BOMMER J. J., LOPEZ D. L., CARR M. J., MAJOR J. J. (eds), Natural Hazards in El Salvador. Geol. Soc. Amer. Spec. Pap, 375: 121–133. BOBROWSKI, N., Ho¨NNINGER, G., LOHBERGER, F., and PLATT, U. (2006), IDOAS: A new monitoring technique to study the 2D distribution of volcanic gas emissions. J. Volcanol. Geotherm. Res. 110, 137–153. BRUNO, N., CALTABIANO, T., GIAMMANCO, S., and ROMANO, R. (2001), Degassing of SO2 and CO2 at Mount Etna (Sicily) as an indicador of pre-eruptive ascent and shallow emplacement of magma. J. Volcanol. Geotherm. Res. 150, 329–338. BURTON, M., CALTABIANO, T., SALERNO, G., MURE, F., and CONDARELLI, D. (2004), Automatic measurements of SO2 flux on Stromboli and Mt. Etna using a network of scanning UV spectrometers, paper presented at 1st EGU Meeting. Eur. Geosci. Union, Nice, France. CASADEVALL, T., ROSE, W.I., FULLER, W., HUNT, W., HART, M., MOYERS, J., WOODS, D., CHUAN, R., and FRIEND, J. (1984), Sulfur dioxide and particles in volcanic plumes from Poa´s, Arenal and Colima Volcanoes, Costa Rica and Mexico. J. Volcanol. Geotherm. Res. 89, 9633–9641. CHIODINI, G., CIONI, R., and MARINI, L. (1993), Reactions governing the chemistry of crater fumaroles from Vulcan Island, Italy, and implications for volcanic surveillance. Applied Geochemistry. 8, 357–371. DAAG, A.S., TUBIANOSA, B.S., NEWHALL, C.G., TUN˜GOL, N.M., JAVIER, D., DOLAN, M.T., DE LOS REYES, P.J., ARBOLEDA, R.A., MARTINEZ, M.L., and REGALADO, T.M. (1996), Monitoring sulfur dioxide emission at Mount Pinatubo. In: NEWHALL C.G., PUNONGBAYAN R.S. (Eds.), Fire and Mud: Eruptions and Lahars of Mount Pinatubo, Philippines. Philippine Institute of Volcanology and Seismology, Quezon City. University of Washington Press, Seattle, WA, 409–434. DELMELLE, P., BAXTER, P., BAULIEU, A., BURTON, M., FRANCIS, P., GARCIA-ALAVAREZ, J., HORROCKS, L., NAVARRO, M., OPPENHEIMER, C., ROTHERY, D., RYMER, H., ST-AMAND, K., STIX, J., STRAUCH, W., and WILLIAMS-JONES, G. (1999), Origin, effects of Masaya volcano’s continued unrest probed in Nicaragua. EOS, Transactions, American Geophysical Union 80, 575–581. DUFFELL, H. J., OPPENHEIMER, C., PYLE, D. M., GALLE, B., MCGONIGLE, A.J.S., and BURTON, M. R. (2003), Changes in gas composition prior to a minor explosive eruption at Masaya volcano, Nicaragua. J. Volcanol. Geotherm. Res. 126, 327–339. EDMONDS, M., HERD, R.A., GALLE, B., and OPPENHEIMER, C. (2003a), Automated, high-time resolution measurements of SO2 flux at Soufriere Hills volcano, Montserrat. Bull. Volcanol. 65, 578–586. EDMONDS, M., OPPENHEIMER, C., PYLE, D.M., HERD, R.A., and THOMPSON, G. (2003b), SO2 Emissions from Soufriere Hills Volcano and their relationship to conduit permeability, hydrothermal interaction and degassing regime. J. Volcanol. Geotherm. Res. 124, 33–43. ELIAS, T., SUTTON, A.J., OPPENHEIMER, C., HORTON, K.A., GARBEIL, H., TSANEV, V, MCGONIGLE, A., and WILLIAMSJONES, G. (2006), Comparison of COSPEC and two miniature ultraviolet spectrometer system for SO2 measurements using scattered sunlight. Bull. Volcanol. 68, 313–322. GALLE, B., OPPENHEIMER, C., GEYER, A., MCGONIGLE, A., EDMONDS, M., and HORROCKS, L.A. (2002), A miniaturized ultraviolet spectrometer for remote sensing of SO2 fluxes: a new tool for volcano surveillance. J. Volcanol. Geotherm. Res. 119, 241–254. GERLACH, T. M. and CASADEVALL, T. J. (1986), Fumarole emissions at Mt. St Helens volcano, June 1980 to October 1981: degassing of a magma-hydrothermal system. J. Volcanol. Geotherm. Res. 28, 141–160. GIGGENBACH, W.F. (1996), Chemical composition of volcanic gases. In Monitoring and Mitigation of Volcano Hazards, Scarpa and Tilling (Eds.). Springer-Verlag, Berlin, 221–256. GIGGENBACH, W.F. and GOGUEL, R.L. (1989), Collection and analysis of geothermal and volcanic water and gas discharges. Chem. Div. DSIR Rept. CD 2401, 81. GOFF, F., MCMURTRY, G. M., COUNCE, D., SIMAC, J. A., ROLDA´N-MANZO, A. R., and HILTON, D. R. (2000), Contrasting hydrothermal activity at Sierra Negra and Alcedo volcanoes, Galapagos Archipelago, Ecuador. Bull. Volcanol. 62, 34–52. HOFF, R.M. (1992), Diferential SO2 column measurements of the Mt. Pinatubo volcanic plume. Geophys. Res. Lett. 19, 175–178.
132
J. Barrancos et al.
Pure appl. geophys.,
HORTON, K.A., WILLIAMS-JONES, G., GARBEIL, H., ELIAS, T., SUTTON, J., MOUGINIS-MARK, P., PORTER, J.N., and CLEGG, S. (2005), Real-time measurement of volcanic SO2 emissions: validation of a new UV correlation spectrometer (FLYSPEC). Bull. Volcanol. 26, 323–327. MATHER, T.A., PYLE, D.M., TSANEV, V.I., MCGONIGLE, A.J.S., OPPENHEIMER, C., and ALLEN, A.G. (2006), A reassessment of current volcanic emissions from the Central American arc with specific examples from Nicaragua. J. Volcanol. Geotherm. Res. 149, 297–311. MCGONIGLE, A. J. S., HILTON, D. R., FISCHER, T. P., and OPPENHEIMER, C. (2005), Plume velocity determination for volcanic SO2 flux measurements. Geophys. Res. Lett., 32, L11302, doi:10.1029/2005GL022470. MILLA´N, M.M. and ROSELLO, J.I. (2005), Development of a new front optics and electronics to improve COSPEC V measurements. IAVCEI Comission on the Chemistry of Volcanic Gases 9th Gas Workshop. MILLA´N, M.M. and HOFF, R.M. (1978), Remote sensing of air pollutants by correlation spectroscopyinstrumental response characteristics. Atmospheric Environment. Vol 12, pp 853–864. MOFFAT, A.J. and MILLA´N, M.M. (1971), The application of optical correlation techniques to the remote sensing of SO2 plumes using skylight. Atmospheric Environment. Vol 5, pp 677–690. Monthly summary (2006), Actividad del volca´n Tungurahua. Escuela Polite´cnica Nacional. Departamento de Geofı´sica. Instituto Geofı´sico de Ecuador. . MORI, T. and BURTON, M. (2006), The SO2 camera: A simple, fast and cheap method for ground-based imaging of SO2 in volcanic plumes. Geophys. Res. Lett., 33, L24804, doi:10.1029/2006GL027916. MUNRO, D.C. and ROWLAND, S.K. (1996), Caldera morphology in the western Galapagos and implications for volcano eruptive behavior and mechanisms of caldera formation. J. Volcanol. Geotherm. Res. 72, 85–100. NEWCOMB, G. and MILLA´N, M.M. (1970), Theory, applications and results of the long-line correlation spectrometer. IEEE Trans. Geosci. Electron., 8, 149–157. OLMOS, R., BARRANCOS, J., RIVERA, C., BARAHONA, B., Lo´PEZ, D., BENANCIO HENRIQUEZ, HERNa´NDEZ, A., EFRAIN BENITEZ, HERNA´NDEZ, PEDRO A. , NEMESIO M. Pe´REZ., and GALLE B. (2007), Anomalous emissions of SO2 during the recent eruption of Santa Ana Volcano, El Salvador, Central America. Special Issue HIROSHI WAKITA vol. II Pure and Applied Geophysics (PAGEOPH), this volume. PUGNAGHI, S., GANGALE, G., CORRADINI, S., and BUONGIORNO, M.F. (2006), Mt. Etna sulfur dioxide flux monitoring using ASTER-TIR data and atmospheric observations. J. Volcanol. Geotherm. Res. 152, 74–90. PUGNAGHI, S., TEGGI, S., CORRADINI, S., BUONGIORNO, M.F., MERUCCI, L., and BOGLIOLO, M.P. (2002), Estimation of SO2 abundance in the eruption plume of Mt. Etna using two MIVIS thermal infrared channels: a case study from the Sicily-1997 Campaign. Bull. Volcanol. 64, 328–337. REYNOLDS, R.W., GEIST, D., and KURZ, M.D. (1995), Physical volcanology and structural development of Sierra Negra volcano, Isabela island, Galapagos archipelago. Geol. Soc. Am. Bull. 107, 1398–1410. RODR´ıGUEZ, L.A., MATTHEW WATSON I., WILLIAM I. ROSE, YVONNE K. BRANAN, GREGG J.S. BLUTH, GUSTAVO CHIGNA, OTONIEL MAT´ıAS, ESCOBAR, D., SIMON A. CARN., and TOBIAS P. FISCHER (2004), SO2 emissions to the atmosphere from active volcanoes in Guatemala and El Salvador, 1999–2002. J. Volcanol. Geotherm. Res. 138, 325–344. SHINOHARA, H. (2005), A new technique to estimate volcanic gas composition: plume measurements with a portable multi-sensor system. J. Volcanol. Geotherm. Res. 143, 319–333. STOIBER, R. E., MALINCONICO, L. L., and WILLIAMS, S. N. (1983), Use of the correlation spectrometer at volcanoes. In Forecasting volcanic events (ed. H. TAZIEFF and J. C. SABROUX). Elsevier. 425–444. SUTTON, A.J., ELIAS, T., GERLACH, T.M., and STOKES, J.B. (2001), Implications for eruptive processes as indicated bysulfur dioxide emissions from Kilauea Volcano, Hawaii, 1979–1997. J. Volcanol. Geotherm. Res. 108, 283–302. SYMONDS, R.B., GERLACH, T.M., and REED, M.H. (2001), Magmatic gas scrubbing: implications for volcano monitoring. J. Volcanol. Geotherm. Res. 108, 303–341. SYMONDS, R. B., ROSE, W. I., BLUTH, G. J. S., and GERLACH, T. M. (1994), Volcanic-gas studies: Methods, Results, and applications, in Volatiles in Magmas, Rev. in Mineral., vol. 30, edited by M. R. CARROLL and J. R. HOLLOWAY, pp. 355–374, Mineral. Soc. of Am., Washington, D.C. TARAN, Y., GAVILANES, J.C., and CORTES, A. (2002), Chemical and isotopic composition of fumarolic gases and the SO2 flux from Volcan de Colima, Mexico, between the 1994 and 1998 eruptions. J. Volcan. Geotherm. Res. 117(1), 105–119. VANDAELE, A. C., SIMON, P. C., GUILMOT, J. M., CARLEER, M., and COLIN R. (1994), SO2 absorption cross-section measurement in the UV using a Fourier transform spectrometer. J. Geophys. Res., 99, 25,599–25,605.
Vol. 165, 2008
SO2 Emission from Active Volcanoes
133
WEIBRING, P., EDNER, H., SVANBERG, S., CECCHI, G., PANTANI, L., FERRARA, R., and CALTABIANO, T. (1998), Monitoring of volcanic sulphur dioxide emissions using differential absorption lidar (DIAL), differential optical absorption spectroscopy (DOAS), and correlation spectroscopy (COSPEC). Appl. Phys. B 67, 419–426. WILLIAMS-JONES, G., HORTON, K. A., ELIAS, T., GARBEIL, H., MOUGINIS-MARK, P. J., SUTTON, A. J., and HARRIS A. J. L. (2006), Accurately measuring volcanic plume velocity with multiple UV spectrometers. Bull. Volcanol. 68, 328– 332. WILLIAMS-JONES, G., STIX, J., HEILIGMANN, M., BARQUERO, J., FERNANDEZ, E., and GONZa´LEZ E.D. (2001), A model of degassing and seismicity at Arenal Volcano, Costa Rica. J. Volcanol. Geotherm. Res. 108, 121–139. WILLIAMS, S.N. (1983), Plinian airfall of basaltic composition. Geology. 11, 211–214. YOUNG S., FRANCIS P.W., BARCLAY J., CASADEVALL T.J., GARDNER C.A., DARROUX B., DAVIES M.A., DELMELLE P., NORTON G.E., MACIEJEWSKI A.J.H., OPPENHEIMER C., STIX J., and WATSON I.M. (1998), Monitoring SO2 emission at the Soufrie’re Hills volcano: implications for changes in eruptive conditions. Geophys. Res. Lett. 25, 3681. ZIMMER, M. M., FISCHER, T. P., HILTON, D. R., ALVARADO, G. E., SHARP, Z. D., and WALKER, J. A. (2004), Nitrogen systematics and gas fluxes of subduction zones: Insights from Costa Rica arc volatiles. Geochem. Geophys. Geosyst., 5, Q05J11, doi:10.1029/2003GC000651. (Received October 1, 2007, revised October 19, 2007, accepted October 29, 2007) Published Online First: February 1, 2008
To access this journal online: www.birkhauser.ch/pageoph
Pure appl. geophys. 165 (2008) 135–145 0033–4553/08/010135–11 DOI 10.1007/s00024-007-0284-6
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
Underground Temperature Measurements as a Tool for Volcanic Activity Monitoring in the Island of Tenerife, Canary Islands A. EFF-DARWICH,1 J. COELLO,1 R. VIN˜AS,1 V. SOLER,2 M. C. MARTIN-LUIS,1 I. FARRUJIA,3 M. L. QUESADA,1 and J. DE LA NUEZ1
Abstract—The spatial distribution of groundwater temperatures in the volcanic island of Tenerife, Canary Islands, has been inferred through measurements of water temperatures collected in the vast network of wells and subhorizontal tunnels, locally called ‘‘galleries,’’ which constitutes the main water supply of the island. The spatial coverage of the network of galleries allows us to reach from depth almost any geological feature of the island. The complex spatial distribution of temperatures in the interior of Tenerife is the result of the complex geological evolution of the island. Groundwater temperatures are greatly affected by groundwater flow and are considerably warmer in those galleries located in areas where water circulation is reduced due to the low permeability of materials and/or to the low infiltration rate of cooling meteoric water. In this sense, groundwater temperature should be characterized in quiescent conditions (background level), in order to facilitate monitoring changes in heat flow, such as those induced by ascending gases expected with an increase in volcanic activity. Key words: Tenerife, groundwater, volcanic eruption, thermal precursors.
1. Introduction The implementation of a volcanic activity monitoring network in the island of Tenerife, Canary Islands (Spain) has to take into account the complex geological evolution of the island. In this sense, recent eruptive activity includes a sub-plinian eruption that took place approximately 2000 years ago in the central part of the island, as well as six strombolian eruptions in the last 300 years that were scattered over a large portion of the island. This dispersion in the frequency and location of eruptive activity makes it difficult to run a monitoring network and hence, simple and robust observational methodologies have to be considered. Groundwater monitoring reveals as a promising option in Tenerife, since the saturated zone could be reached through more than 1000 galleries and more than 400 1 Departamento de Edafologı´a y Geologı´a, Universidad de La Laguna, Av. Astrofı´sico Francisco Sa´nchez s/n., La Laguna, 38206 Tenerife, Spain. E-mail: [email protected] 2 IPNA-CSIC, Volcanological Station of the Canary Islands, Tenerife, Spain. 3 Consejo Insular de Aguas, Cabildo Insular de Tenerife, Tenerife, Spain.
136
A. Eff-Darwich et al.
Pure appl. geophys.,
wells. Galleries are horizontal tunnels, usually several kilometres long depending on the excavated distance needed to reach the general saturation level. Water flows through one or several outcrops (that may be separated up to several hundreds of meters), being transported to the entrance into pipes or channels. The location of the entrance of these tunnels varies from a few meters above sea level to approximately 1800 meters. Galleries are optimal locations to study the geodynamical processes occurring in the interior of the island. On one hand, they penetrate the different stratigraphical units that conform the interior of the volcanic edifice, providing valuable information on the location and evolution of the general water table, the regional water balance, the groundwater flow pattern and the geology of the island (CUSTODIO, 1987; NAVARRO, 1991). On the other hand, galleries are very stable locations that are ideal to monitor gaseous dynamics, since the effect of most atmospheric variables is negligible. Air-temperature variations are very small and the only recordable environmental parameters that may affect gas flows are barometric pressure and temperature differences between the air inside and outside the galleries (MART´ıN et al., 2002; EFF-DARWICH et al., 2002). Systematic monitoring of groundwater in active volcanic regions is one of the tools used in early detection of volcanic eruptions. Indeed, groundwater could trap the main components of fluids released from magma, providing information on temporal and spatial changes in heat and gas transfer within the volcanic edifice (TEDESCO, 1995; FEDERICO et al., 2002). It is expected that magma ascent towards the surface modifies the temperature distribution inside the volcanic edifice, as the result of heat transfer between magmatic gases, cracks and groundwater. In this sense, it may occur that temperature in wells and springs could change during the early stages of volcanic eruptions. This has been reported in many examples in the literature, e.g., temperature changes in crater lakes (SIGURDSSON, 1977; BADRUDIN, 1994), in hot springs (SATO et al., 1992), wells and springs (BONFANTI et al., 1996; MART´ıN DEL POZZO et al., 2002). YAMASHINA and MATSUSHIMA (1999) also reported variations on ground temperature outside the geothermal area at Unzen volcano prior to the phreatic eruption in November 1990. In the case of volcanic islands, the connection between groundwater and hot magmatic fluids is poorly understood, since the available volcanologic and hydrologic data are limited (VIOLETTE et al., 1997). In the case of Tenerife, the interaction between volcanic activity and groundwater is represented by fumarolic gaseous emissions (ALBERT et al., 1989) and by the presence of thermal and hydrochemical anomalies (i.e., BRAVO et al., 1976; CUSTODIO, 1987; NAVARRO, 1991; FARRUJIA et al., 1994). In this work, we have obtained the spatial distribution of temperatures in the groundwater system of the Island of Tenerife. We attempted to show how temperature is a proxy for the dynamics of the groundwater system, as well as for volcanic activity in Tenerife. In this sense, this work provides valuable information for setting up an instrumental network (e.g., temperature and radon) in galleries and wells that could become an essential part of any early detection system of volcanic activity in the island.
Vol. 165, 2008
Underground Temperature Measurements
137
2. Geology and Hydrology of Tenerife Tenerife is the largest island (2034 km2) of the Canarian Archipelago and one of the largest oceanic islands in the world. It is located between latitudes 28–29N and longitudes 16–17W, 280 km away from the African coast. The construction of the island has been brought about by the accumulation, during the last millions of years, of volcanic materials with different compositions. The stratigraphic units conforming the emerged part of Tenerife may be placed in three groups (Fig. 1). The first group is conformed by the Old Basaltic Edifices or Shield Edifices, dated from 12 Ma to approximately 3.3 Ma (GUILLOU et al., 2004). The second group, the Central Volcanic Edifice (compressing the so-called Can˜adas Edifice and Teide-Pico Viejo Complex), has been dated from 3.5 Ma to present (ANCOCHEA et al., 1999; MART´ı et al., 1994; HUERTAS et al., 2002). The most representative structures of the Central Volcanic Edifice are a large elliptical depression measuring 16 9 9 km2, known as Las Can˜adas Caldera and, in the northern sector of the
Figure 1 Simplified geological map of Tenerife, including the main volcanic edifices and recorded historical eruptions (stars), namely Siete Fuentes (1), Fasnia (2), Arafo (3), Arenas Negras (4), Chahorra (5) and Chinyero (6).
138
A. Eff-Darwich et al.
Pure appl. geophys.,
caldera, the Teide-Pico Viejo strato-volcano. Finally, the third group is conformed by three basaltic ridges or Dorsal Edifices (NE, NW and S ridges), that overlap in time with the Central Volcanic Edifice. These ridges converge at the center of the island where the strato-volcano Teide-Pico Viejo is located (Fig. 1). The growth of the island has not been homogeneous, occasionally concentrating volcanic activity in some areas where there has been an excessive vertical accumulation of materials. This accumulation has induced gravitational instabilities that have led to giant landslides and to the generation of gravitational depressions (valleys), being the Icod, La Orotava and Gu¨´ımar valleys the largest (CANTAGREL et al., 1999; ABLAY and HURLIMANN, 2000; WALTER et al., 2005). A portion of the Central Volcanic Edifice did not collapse during the formation of the Icod and La Orotava valleys, conforming the present Tigaiga massif. Recorded eruptive activity has consisted of six strombolian eruptions (CABRERA and HERNa´NDEZ-PACHECO, 1987), namely Siete Fuentes (1704), Fasnia (1705), Arafo (1705), Arenas Negras (1706), Chahorra (1798) and Chinyero (1909). The last three eruptions occurred in the NW ridge, the most active area of the island together with El Teide-Pico Viejo Complex for the last 50,000 years (CARRACEDO et al., 2003). The hydrological behavior of the island is defined by three different aspects that characterize the insular edifice, namely the stratigraphic accumulation of materials, the dorsal ridges and the gravitational depressions (NAVARRO, 1991). Indeed, the deeper the volcanic edifices are located in the stratigraphic sequence, the more compact and altered they are and hence, the less permeable. In this sense, the permeabilities in the Shield Massifs and Can˜adas Edifice are gradually reduced with increasing depth until they become nearly impervious, whereas the more recent edifices are less compact and altered and hence, they constitute highly permeable units. The dorsal ridges play an important role in the hydrodynamics of the island. In the central strip of the ridges, where the dike swarm density is greater, the open fractures attenuate the differences in the original permeabilities of the different stratigraphic units, significantly increasing vertical interconnection. Dikes and fractures are usually parallel to the ridges, facilitating longitudinal flow and considerably reducing transversal flow. This induces a super-elevation of the saturated zone and the transversal profile of the water table becomes stepped. The valleys are formed by an impervious basement (debris avalanche deposit), known as ‘‘Mortalo´n’’ (CANTAGREL et al., 1999), covered by high permeability post-landslide lavas, where the saturated zone is located. Infiltrated water cannot be retained by the lava fill-in and circulates down to the basement, where it begins to flow towards the sea.
3. Observational Methodology The Water Council Board (hereafter WCB) of the local Government, the Cabildo de Tenerife, manages a large database, consisting of physico-chemical measurements of water samples collected in more than 1000 galleries and wells spread out all over the
Vol. 165, 2008
Underground Temperature Measurements
139
Figure 2 Map of Tenerife indicating the location of the galleries (solid lines) and wells (filled circles) that constitute the main water exploitation system of the island. Contour lines for the position of the general water table are plotted every 200 meters.
island. We took from this database 390 measurements of groundwater temperatures collected from 1998 to 2004. Moreover, we also used 91 water temperature measurements collected by COELLO (1976) at water outcrops in galleries between 1969 and 1973. Neither galleries nor wells have reached the saturated zone in the central part of Las Can˜adas Caldera. Hence, information on the temperature of the aquifer is provided by two scientific boreholes drilled in this area by the WCB. The Water Council has also measured the location of water outcrops in galleries and water level in wells, obtaining in this way the shape of the groundwater system (see Fig. 2), that at large scale retains the pyramidal-like shape of the island. WCB data were collected at the entrance of the galleries and hence this may not exactly reflect the actual conditions at the saturation level, because of changes in temperature during the transport of water into the pipes or channels. Comparison of the data sets revealed differences in the temperatures measured of up to 4C at a given sampling point. This prevented us from performing a quantitative analysis of the data and hence, a qualitative study of the temperature distribution has been carried out.
4. Results The distribution of groundwater temperature is highly heterogeneous, as illustrated in Figure 3. The mean temperature for groundwater is 23C, being 21.6C and 24.4C for galleries and wells located in the northern and southern slopes of the island, respectively. Four different aspects have to be considered when analyzing the spatial distribution of
140
A. Eff-Darwich et al.
Pure appl. geophys.,
Figure 3 Distribution of temperatures for the groundwater system of the island of Tenerife. Contour lines, plotted every 200 meters, indicate the position of the general water table. Filled circles indicate both the location of wells and the deepest part of the galleries.
temperatures, namely meteorology, permeability of volcanic materials, water flow from Las Can˜adas Caldera and volcanic activity. Rainfall varies from less than 200 mm/year at low altitudes in the southern part, up to 1000 mm/year at altitudes between 1000 and 1500 m, specially in the NE ridge. Some snow percolates El Teide-Pico Viejo and Las Can˜adas Caldera, helping to recharge the aquifer in the central part of the island. The larger precipitation rate in the northern part of the island is a consequence of the N and NE wet trade winds. Larger permeabilities are found in the northern slope of the island, namely the lava fill-in of La Orotava and Icod valleys, as well as the NE and NW ridges. Those areas located in the southern flank of Las Can˜adas Caldera are well defined geologically by the presence above the saturated zone of a thick and extensive pile-up of scarcely permeable phonolitic lavas. Groundwater is nearly stagnant and the cooling effect of infiltrated meteoric water is insignificant due to the presence of this low-permeable layer, as well as to the low precipitations recorded in the area. Cold meteoric water collected in the very permeable lavas of the Las Can˜adas Caldera easily flows down through the lava fill-in of La Orotava and Icod valleys, as well as through the NW ridge and that part of the Tigaiga massif in direct contact to Las Can˜adas Caldera, acting as a refrigerating agent. The effect of this flow from Las Can˜adas aquifer is also illustrated in Figure 4, where it is represented by the distribution of HCO3 content in water. These data were also taken from the WCB database. Groundwater in direct contact to Las Can˜adas Caldera presents substantial values for the content of HCO3 and relatively cold temperatures, reflecting the conditions found in the Caldera aquifer, namely cold temperature (approximately 15C) and high contents of CO2
Vol. 165, 2008
Underground Temperature Measurements
141
Figure 4 Distribution of HCO3 content in water for the groundwater system of the island of Tenerife. Contour lines, plotted every 200 meters, indicate the position of the general water table. Filled circles indicate both the location of wells and the deepest part of the galleries.
dissolved in water (SOLER et al., 2004). Since there are no sedimentary carbonates in Tenerife, HCO3 originates in the dissolution of volcanic CO2 in water (CUSTODIO, 1987). The warmer groundwaters of the southern slope of the island are not refrigerated by the Caldera aquifer, since the southern wall of the caldera behaves as an impermeable barrier (NAVARRO, 1991). Volcanic activity could also play an important role in the temperature distribution of groundwater. It has been postulated that some anomalously high water temperature areas may be associated with the cooling of shallow magma chambers that fed historical eruptions (VALENT´ıN et al., 1990) as occurs in the areas closer to Fasnia, Siete Fuentes, Chinyero and Arafo volcanoes. This idea is supported by the fact that groundwater located in this areas presents high temperatures and geo-chemical anomalies. Heat upflow may also be induced by convective ascent of hot endogenous gases from a deep source through preferential paths, such as dike swarms or fracture systems. The existence of a large-scale upflow of gases in the central part of the island has already been proposed by BRAVO et al. (1976) and PE´REZ et al. (1996), based on measurements of the concentration of CO2 in the air of the galleries and the ratio of 3He/4He, respectively. There are some areas where the density of galleries is large enough to study the dependence of groundwater temperature with depth and altitude. These areas are shown in Figure 3 as circles and labelled as a, b and c. Area a is located in the southern slope of the island, where both precipitation rates and permeabilities are low. The dependence of temperatures with depth below topographic surface and altitude of the measuring point above sea level (panel a of Figs. 5 and 6) shows no significant trend. In the case of area b, located in the NE ridge closer to the Gu¨imar valley, we found a significant increment of temperature with depth and altitude (panel b of Figs. 5 and 6), being more evident in the
142
A. Eff-Darwich et al.
Pure appl. geophys.,
case of altitude. The better match between temperature and altitude may be a geometrical effect caused by rapid variations in the altitude of the topographic surface relative to the altitude of the sampling points. The variation of temperature with depth is approximately 0.02C/m, being closer to the value of the standard geothermal gradient, 0.03C/m. Area c is located in the Tigaiga Massif, however it is not in direct contact with the Las Can˜adas Caldera aquifer. Groundwater reserves are negligible when compared to other areas. In this sense, water flow does not play an important role in controlling the vertical distribution of temperatures and hence, an increase in temperature with depth of approximately 0.05C/m is observed (panel c of Figs. 5 and 6). This is larger than the standard geothermal gradient, however the lack of precision in the temperature measurements prevents us from making further conclusions. In summary, it has been shown that the combination of larger precipitation rates, larger permeabilities and the cooling effect of the cold flow from Las Can˜adas aquifer may explain the lower groundwater temperatures found in the northern slope of Tenerife, relative to the southern slope. It is still not clear whether recent volcanic eruptions or large scale heat upflow or a combination of both could explain the temperature distribution in the southern slope of the island. It is evident that more precise measurements are necessary to better understand the thermal regime of the groundwater system in Tenerife, in particular the temperature gradient in Tigaiga, the temperature
Figure 5 Panels a), b) and c) represent the distribution of groundwater temperature relative to altitude above sea level of the temperature measuring points for the areas a, b and c that are shown in Figure 3.
Vol. 165, 2008
Underground Temperature Measurements
143
Figure 6 Panels a), b) and c) represent the distribution of groundwater temperature relative to depth below topographic surface for the areas a, b and c that are shown in Figure 3.
distribution in the southern slope of the island or some high temperature spots in the Orotava valley.
5. Conclusions We have carried out a qualitative analysis of the regional distribution of groundwater temperature for the volcanic island of Tenerife. This analysis was thought as a first step for setting up a network of water temperature monitoring stations that could become part of an early detection system of volcanic activity. Groundwater temperature is greatly affected by groundwater flow, being warmer those galleries located in areas where water circulation is reduced. The cold-water aquifer in Las Can˜adas Caldera is connected to highly permeable areas, such as the NW ridge, La Orotava and Icod valleys in the North, but it is disconnected from the warmer areas of the southern slope of the island, since the southern wall of the Can˜adas Caldera acts as an impermeable barrier to water flow. Although the origin and spatial distribution of heat flow remains uncertain, it seems reasonable to characterize the spatial and temporal evolution of groundwater temperatures. In this way, it will be possible to define the temperature distribution in quiescent conditions (background level) from where changes in the heat flow could be monitored,
144
A. Eff-Darwich et al.
Pure appl. geophys.,
such as those induced by the ascending gases expected with an increase in volcanic activity. During the elaboration of this work, Prof. Juan Coello passed away after a fatal car accident. Undoubtedly, he has been one of the most prominent geologist in the Canary Islands and his work greatly improved our understanding on the structure and evolution of this volcanic archipelago.
Acknowledgements AED thanks the Spanish Ministerio de Educacio´n y Ciencia for the contract under the programme ‘Ramo´n y Cajal’. We also thank the Water Council Board of the Cabildo de Tenerife for providing the data and for their comments and suggestions.
REFERENCES ABLAY, G. and HU¨RLIMANN, M. (2000), Evolution of north flank of Tenerife by recurrent giant landslide processes, J. Volcanol. Geoth. Res. 103, 135–159. ALBERT, J.F., Dı´EZ GIL, J.L., VALENT´ıN, A., GARC´ıA DE LA NOCEDA, C., and ARAN˜A, V. (1989), El sistema fumaroliano del Teide. In Los volcanes y la caldera del Parque Nacional del Teide (Tenerife, Islas Canarias), Aran˜a, V. y Coello, J. Ed. ICONA, Madrid. 347–358 pp. ANCOCHEA, E., HUERTAS, M.J., CANTAGREL, J.M., COELLO, J., Fu´STER, J.M., ARNAUD, N., and IBARROLA, E. (1999), Evolution of the Can˜adas edifice and its implications for the origin of the Can˜adas Caldera (Tenerife, Canary Islands), J. Volcanol. Geoth. Res. 88, 177–199. BADRUDIN, M. (1994), Kelut volcano monitoring: Hazards, mitigation and changes in water chemistry prior to the 1990 eruption, Geochem. J. 28, 233–241. ´ ALESSANDRO, W., DONGARR, G., PARELLO, F., VALENZA, M. (1996), Medium-term anomalies in BONFANTI, P., D groundwater temperature before 1991–1993 Mt. Etna eruption, J. Volcanol. Geoth. Res. 73, 303–308. BRAVO, T., COELLO, J., and BRAVO, J. (1976), Areas de emanaciones gaseosas y anomalı´as te´rmicas en la provincia de Santa Cruz de Tenerife (Islas Canarias), II Asamb. Nac. Geod. Geof., 2235–2244. CABRERA, M.P. and HERNA´NDEZ-PACHECO, A. (1987), Las erupciones histo´ricas de Tenerife (Canarias) en sus aspectos vulcanolo´gico, petrolo´gico y geoquı´mico, Rev. Mat. Proc. Geol. V, 143–182. CANTAGREL, J.M., ARNAUD, N.O., ANCOCHEA, E., FU´STER, J.M., and HUERTAS, M.J. (1999), Repeated debris avalanches on Tenerife and genesis of Las Can˜adas caldera wall (Canary Islands), Geology 27, 739–742. CARRACEDO, J.C., PATERNE, M., GUILLOU, H., PE´REZ TORRADO, F.J., PARIS, R., RODR´ıGUEZ BADIOLA, E., and HANSEN, A. (2003), Dataciones radiome´tricas (14C y K/Ar) del Teide y del rift noroeste, Tenerife, Islas Canarias, Estudios Geol. 59, 15–29. COELLO, J. (1976), Las series volca´nicas en subsuelos de Tenerife, Estudios Geol. 29, 491–512. CUSTODIO, E. (1987), Hydrogeochemistry of Tenerife Island. In Simposio Canarias 2000, Tenerife, Spain. EFF-DARWICH, A., MART´ıN LUIS, M.C., QUESADA, M.L., DE LA NUEZ, J., and COELLO, J. (2002), Variations on the concentration of 222Rn in the subsurface of the volcanic island of Tenerife, Canary Islands, Geophys. Res. Lett. 29, 26. FARRUJIA, I., DELGADO, P., and BETHENCOURT, J. (1994), Calidad y contaminacio´n de las aguas subterra´neas de Tenerife en el marco de la planificacio´n hidrolo´gica, Congr. Ana´lisis y Evolucio´n de la Contaminacio´n de las Aguas Subterra´neas, Alcala´ de Henares (Madrid), Tomo II, 397–416 pp. FEDERICO, C., AIUPPA, A., ALLARD, P., BELLOMO, S., JEAN-BAPTISTE, P., PARELLO, F., and VALENZA, M. (2002), Magma-derived gas influx and water-rock interactions in the volcanic aquifer of Mt. Vesuvius, Italy, Geochim. Cosmochim. Acta 66, 963–981.
Vol. 165, 2008
Underground Temperature Measurements
145
GUILLOU, H., CARRACEDO, J.C., PARIS, R., and PE´REZ TORRADO, F.J. (2004), Implications for the early shield-stage evolution of Tenerife from K/Ar ages and magnetic stratigraphy, Earth Planet. Sci. Lett. 222, 599–614. HUERTAS, M.J., ARNAUD, N.O., ANCOCHEA, E., CANTAGREL, J.M., and FU´STER, J.M. (2002), 40Ar/39Ar stratigraphy of pyroclastic units from the Can˜adas Volcanic Edifice (Tenerife, Canary Islands) and their bearing on the structural evolution, J. Volcanol. Geoth. Res. 115, 351–365. MART´ı, J., MITJAVILA, J., and ARAN˜A, V. (1994), Stratigraphy, structure and geochronology of the Can˜adas Caldera (Tenerife, Canary Islands), Geol. Mag. 131(6), 715–727. MARTIN-DEL POZZO, A.L., ACEVES, F., ESPINASA, R., AGUAYO, A., INGUAGGIATO, S., MORALES, P., and CIENFUEGOS, E. (2002), Influence of volcanic activity on spring water chemistry at Popocatepetl Volcano, Mexico, Chem. Geol. 190, 207. MART´ıN, M.C., QUESADA, M.L., EFF-DARWICH, A., DE LA NUEZ, J., COELLO, J., AHIJADO, A., CASILLAS, R., and SOLER, V. (2002), A new strategy to measure radon in an active volcanic island (Tenerife, Canary Islands), Environmental Geology 43, 72–78. MEZCUA, J., BUFORN, E., UD´ıAS, A., and RUEDA, J. (1992), Seismotectonics of the Canary Islands, Tectonophysics, 208, 447–452. NAVARRO, J.M. (1991), Plan Hidrolo´gico de Tenerife. Cabildo Insular de Tenerife. PE´REZ, N.M., NAKAI, S., WAKITA, H., HERNA´NDEZ, P.A., and SALAZAR, J.M. (1996), Helium-3 emission in and around Teide volcano, Tenerife, Canary Islands, Spain, Geophysical Rese. Lett., 23, 3531. SATO, T., WAKITA, H., NOTSU, K., and IGARASHI, G. (1992), Anomalous hot spring water changes: Possible precursors of the 1989 volcanic eruption off the east coast of the Izu Peninsula, Geochem. J. 26, 73–83. SIGURDSSON, H. (1977), Chemistry of the crater lake during the 1971–72 Soufrie`re eruption, J. Volcanol, Geoth. Res. 2, 165–186. SOLER, V., CASTRO, J.A., VIN˜AS, R.T., EFF-DARWICH, A., SA´NCHEZ, S., HILLARIE-MARCEL, C., FARRUJIA, I., COELLO, J., DE LA NUEZ, J., MART´ıN, M.C., QUESADA, M.L., and SANTANA, E. (2004), High CO2 levels in boreholes at El Teide volcano complex (Tenerife, Canary Islands): Implications for volcanic activity monitoring, Pure Appl. Geophys, 161, 1519–1532. TEDESCO, D., Monitoring fluids and gases at active volcanoes, Monitoring active volcanoes (ed, McGuire, B., Kilburn, C., and Murray, J.) (UCL, London (1995)) pp. 315–345. VALENT´ıN, A., ALBERT-BELTRA´N, J.F., and Dı´EZ, J.L. (1990), Geochemical and geothermal constraints on magma bodies associated with historic activity, Tenerife, J. Volcanol. Geotherm. Rese. 44, 251–264. VIOLETTE, S., LEDOUX, E., GOBLET, P., and CARBONNEL, J.P. (1997), Hydrologic and thermal modelling of an active volcano: The ‘‘Piton de la Fournaise’’, Reunion Island, Journal of Hydrology 191, 1–4, 37–63. WALTER, T.R., TROLL, V.R., CAILLEAU, B., BELOUSOV, A, SCHMINCKE, H.U., AMELUNG, F., and v.d. Bogaard, P. (2005), Rift zone reorganization through flank instability in ocean island volcanoes: An example from Tenerife, Canary Islands, Bull. Volcanol. 67, 281–291. YAMASHINA, K. and MATSUSHIMA, T. (1999), Ground temperature change observed at Unzen Volcano associated with the 1990–1995 eruption, J. Volcanol. Geotherm. Res. 89, 65–71 . (Received January 31, 2006, revised February 28, 2007, accepted March 15, 2007) Published Online First: February 1, 2008
To access this journal online: www.birkhauser.ch/pageoph
Pure appl. geophys. 165 (2008) 147–172 0033–4553/08/010147–26 DOI 10.1007/s00024-007-0287-3
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
Carbon Dioxide Discharged through the Las Can˜adas Aquifer, Tenerife, Canary Islands RAYCO MARRERO,1 DINA L. LO´PEZ,2 PEDRO A. HERNA´NDEZ,1 and NEMESIO M. PE´REZ1
Abstract—Carbon dioxide is one of the first gases to escape the magmatic environment due to its low solubility in basaltic magmas at low pressures. The exsolved CO2 gas migrates towards the surface through rock fractures and high permeability paths. If an aquifer is located between the magmatic environment and the surface, a fraction of the CO2 emitted is dissolved in the aquifer. In this paper, an estimation of the water mass balance and the CO2 budget in Las Can˜adas aquifer, Tenerife, Canary Islands, is presented. Magmatic CO2 is transported by groundwater and discharged through man-made sub-horizontal drains or galleries that exist in this island, and by the flow of groundwater discharged laterally towards other aquifers or to the ocean. In addition, the pCO2 at the gallery mouth (or entrance) and at the gallery bottom (internal and deepest discharge point where the gallery starts) are calculated and mapped. The total CO2 advectively transported by groundwater is estimated to range from 143 to 211 t CO2 d-1. Considering that the diffuse soil emission of CO2 for the same area is 437 t d-1, the diffuse/dissolved CO2 flux ratio varies between 2 and 3. The high dissolved inorganic carbon content of groundwater explains the ability of this low temperature hydrothermal water to dissolve and transfer magmatic CO2 at volcanoes, even during quiescence periods. Key words: Tenerife, volcanic aquifer, carbon dioxide, groundwater.
1. Introduction During periods of non-eruptive activity, volcanic-hydrothermal systems release to the atmosphere large fluxes of CO2, either as diffuse emissions through the soil-air interface (CHIODINI et al., 1996; HERNA´NDEZ et al., 1998) or as CO2 transported by hot water and discharged advectively as hot springs, fumaroles, or steaming ground. However, there is another component of the output of CO2 from volcanic systems that deserves consideration: the output of CO2 dissolved in discharging cold water in volcanic aquifers. Recent studies indicate that calculations of the CO2 budget in volcanic systems that do not consider this component of the CO2 discharged are seriously underestimated (CHIODINI et al., 1999; CHIODINI and FRONDINI, 2001; EVANS et al., 2002; GAMBARDELLA et al., 2004). In this paper we present results for the estimation of the CO2 budget of Las 1 Environmental Research Division, Instituto Tecnolo´gico y de Energı´as Renovables (ITER), 38611 Granadilla, S/C de Tenerife, Spain. E-mail: [email protected] 2 Department of Geological Sciences, 316 Clippinger Laboratories, Ohio University, Athens, OH 45701, USA.
148
R. Marrero et al.
Pure appl. geophys.,
Can˜adas aquifer, Tenerife, Canary Islands, considering the gases released at the soil-air interface as well as the CO2 transported by groundwater and discharged through manmade sub-horizontal drains or galleries that exist in this island, and by the flow of groundwater discharged laterally towards other aquifers or the ocean. Carbon dioxide is one of the first gases to escape the magmatic environment due to its low solubility in basaltic magmas at low pressures (STOLPER and HOLLOWAY, 1988). The exsolved CO2 gas migrates towards the surface through rock fractures and permeable paths. If an aquifer is located between the magmatic environment and the surface, a fraction of the CO2 emitted is dissolved and assimilated into the aquifer. The CO2 accumulated into the aquifer flows up within the vadose zone until it reaches the soil-air interface. If the partial pressure of CO2 (hereafter pCO2) in the aquifer is higher than the total pressure, CO2 bubbles form and move up. If the partial pressure of CO2 is lower than the total pressure, CO2 moves up due to diffusion. The dissolved CO2 is transported by groundwater away from the volcanic source. Waters with high alkalinity at many exploited hydrothermal systems around the world are providing heat and energy throughout large volumes of extracted groundwater during long periods of time (OKADA et al., 2000; PADRo´N et al., 2003; WHITE et al., 2005), and then releasing high fluxes of CO2 to the atmosphere. High pCO2 values are found close to the source of this gas. As the water circulates within the matrix rock, pH increases and the pCO2 decreases. For volcanic aquifers that are not located within carbonate rocks, zones with an abnormally high pCO2 suggest either a better connection between the aquifer and the source of gases due to high permeability pathways within the rocks or to a close location of the source with respect to the measuring point. Las Can˜adas aquifer is a good site for hydrogeological studies due to the existence of dozens of galleries constructed to reach the saturated zone at different depths and elevations (Fig. 1) to exploit the aquifer. The chemical composition of groundwater at Las Can˜adas aquifer is sodium bicarbonate-rich and shows relatively high contents of total dissolved solids ranging from 1.6 to more than 2.5 g/L (NAVARRO, 1994). Most of the solutes in the groundwater are derived from a significant water-rock interaction due to the input of deep-seated gases from Teide volcanic-hydrothermal system (ALBERT-BELTRAN et al., 1990; VALENTIN et al., 1990; NAVARRO, 1994; PE´REZ et al., 1996) that provide aggressiveness to the water, enhancing rock dissolution and alteration. The diffuse soil CO2 degassing at Las Can˜adas Caldera and Teide-Pico Viejo volcanic complex has been studied (HERNA´NDEZ et al., 1997). According to this study, the main structure releasing CO2 at Las Can˜adas Caldera is the summit cone of Teide volcano where fumarolic activity occurs. Other anomalous levels of diffuse CO2 were identified along the NW rift zone. The fluxes of CO2 measured at Las Can˜adas and its surroundings and at Teide volcano averaged 1.9 g m-2 d-1 and 690 g m-2 d-1, respectively (HERNA´NDEZ et al., 1997). Both CO2 fluxes represent a total CO2 flux of 563 t d-1 for the 197.9 km2 of the central part of the Tenerife Island (Fig. 1).
Vol. 165, 2008
Carbon Dioxide Discharge
149
Figure 1 Las Can˜adas aquifer, Tenerife, Canary Islands. Stars represent the gallery mouths considered in this work and the dark lines the trace of each gallery. Dotted ellipse represents the surface area of soil diffuse degassing studies reported in HERNA´NDEZ et al. (1997).
Abundant chemical data are available of the water from the galleries that have been filled by the Consejo Insular de Aguas de Tenerife (Tenerife Island Water Agency, hereafter CIA). This information, their hydrological setting, and the sub-horizontal drillings reaching the Las Can˜adas aquifer at different levels and positions are ideal for estimating the levels of subsurface degassing of CO2. In addition, it makes identification possible of those areas with anomalously high levels of total dissolved inorganic carbon (hereafter DIC) and pCO2 in the groundwaters of this aquifer.
2. Geological and Hydrogeological Setting of Study Area Tenerife (2034 km2) is the largest island of the Canarian archipelago, which is the only active volcanic region in Spain. Three main volcanic rift-zones (NE, NW, and N-S) occur in the island. The recent historical eruptions have occurred along those rifts. The most recent eruption at Tenerife (Chinyero volcano, Fig. 1) occurred in 1909 AD at the NW rift-zone. The intersection of the three rifts occurs at Las Can˜adas Caldera (Fig. 1). This Caldera is located at the central part of Tenerife and forms a large volcanic depression (16 9 9 km). Several hypotheses have been proposed to explain the origin of
150
R. Marrero et al.
Pure appl. geophys.,
this caldera including erosion, explosion, single or multiple vertical collapse and giant landslide (BRAVO, 1962; FUSTER et al., 1968; MACFARLANE and RIDLEY, 1968; ARAN˜A, 1971; RIDLEY, 1971; NAVARRO and COELLO, 1989; ANCOCHEA et al., 1990; MARTI et al., 1997; ABLAY and MARTI, 2000; MARTI and GUDMUNSSON, 2000; POUS et al., 2002). The most spectacular geological feature at Las Can˜adas Caldera is the Teide-Pico Viejo stratovolcanic complex (elevation 3718 m.a.s.l.) at the northern side of the caldera. Constant fumarolic degassing at a temperature of 85C occurs at the summit cone of Teide volcano. Two types of deposits can be distinguished within the caldera: The Pre-Can˜adas deposits formed before the formation of the caldera, and the Post-Can˜adas deposits produced by the eruptions of Teide-Pico Viejo volcanic complex during the last 0.17 million years (MARTI et al., 1994). The post-Caldera deposits are mainly basaltic, trachytic and phonolytic in nature, revealing mixtures of deep basaltic magmas with other more developed magmas (ARAN˜A et al., 1989). In addition to the Teide-Pico Viejo stratovolcanic complex, whose last eruption took place on 1798 AD (Chahorra Eruption, Fig. 1), an intense network of phonolytic dikes and different monogenetic cone alignments are visible at the surface of Las Can˜adas Caldera. This caldera is opened to the sea at its north side due to collapses that have occurred in that side of the island (ABLAY and HU¨RLIMANN, 2000). The post-Can˜adas deposits represent continuous layers of almost non-altered lava flows and fall deposits with relatively high permeability. These volcanic deposits form a good hydrogeological reservoir contrasting with the low permeability of the basement rocks and the low transversal permeability of the injected dykes, possibly forming the largest groundwater reserve of the island. Magnetoteluric studies of the caldera (POUS et al., 2002) and measurements of the water level in two boreholes (S-1 and S-2 in Fig. 1) drilled by the CIA in the rocks filling the caldera (FARRUJIA et al., 2001a, b) show that the phreatic level has different elevations within the amphitheater of the caldera. S-1 presents a higher water table elevation, about 40 m above the phreatic surface at S-2. These changes in elevation probably have been produced by the geological characteristics of the caldera filling, the network of dykes, and the impact of the different galleries exploiting the aquifer. Previous hydrogeological studies carried out at Las Can˜adas Caldera suggest that groundwater flows mainly to the north, partially diverted by the presence of the TeidePico Viejo volcanic complex (NAVARRO, 1994). This conclusion is supported by the spatial distribution of the total dissolved solids (TDS) as well as other hydrochemical parameters (CUSTODIO et al., 1987; SALAZAR et al., 1997).
3. Methodology Several components of the CO2 released by this volcanic complex have been considered in the CO2 budget: The diffuse CO2 soil discharges already reported in HERNA´NDEZ et al. (1997), the CO2 dissolved in the groundwater discharged by the
Vol. 165, 2008
Carbon Dioxide Discharge
151
galleries that intercept Las Can˜adas aquifer (controlled discharges), and the CO2 dissolved in the groundwater discharged laterally towards other neighboring aquifers and galleries that intercept the aquifer but that cannot be measured and is largely unknown. This last component of the CO2 discharges will be called ‘‘non-controlled CO2 discharges’’. As water is the important carrier phase for the transport of CO2, knowledge of the water balance for the aquifer is essential. In this paper, a discussion of the data base is presented first, followed by an estimation of the water mass balance, and then the CO2 balance. In addition, the pCO2 at the external gallery mouth and at the gallery bottom (internal and deepest discharge point where the gallery starts) are calculated and mapped. 3.1. Data Base Water samples of two boreholes (505 and 404 m depth) drilled at Las Can˜adas Caldera and 37 galleries have been selected for this study (Fig. 1). The galleries were selected based on hydrogeological factors such as a clear connection with Las Can˜adas aquifer and water chemistry data reflecting a low ion balance error (Table 1). The position of the galleries with respect to the aquifer allows the identification of four sectors (Fig. 1): [1] Icod - La Guancha Valley or N sector: the high magnitude of the discharges and the chemical composition of groundwaters in this sector (high TDS Na-HCO3 waters) suggests that a high fraction of the Las Can˜adas water is discharging in that direction (MARRERO, 2004). [2] Head of La Orotava Valley or NE sector: high flows of water are discharging to the NE due to a high hydraulic gradient between Las Can˜adas Caldera and La Orotava Valley, and to several galleries intercepting Las Can˜adas aquifer in that direction. [3] Vilaflor—Adeje or S sector: several galleries are discharging water from Las Can˜adas. It should be noted that according to the excavating history of these galleries, they did not discharge water until they intercepted the materials filling the caldera (NAVARRO, 1994). [4] Boca Tauce—NW Ridge: the galleries in this sector drain water transported throughout permeable fractures parallel to dykes located along the NW ridge and that penetrate the caldera aquifer. Flow data are recorded and reported by the gallery’s managers and the CIA. Data base selected for this study includes data from 1991 to 2001, prior to the recent seismic activity started on 2004 at the northwestern part of Tenerife Island (GOTTSMANN et al., 2006; ALMENDROS et al., 2007; MARRERO et al., submitted). Physical-chemical data are reported in Table 1. Note that the data used for this period were mostly collected by CHIODINI (1993) and the CIA. The water sampling was performed during a relatively long time (1991–2001) for the characterization of the aquifer in quiescence periods. Some
152
R. Marrero et al.
Pure appl. geophys.,
Table 1 Chemical composition of waters discharged at galleries that intercept Las Can˜adas aquifer. T: Temperature (C); E.C.: Electrical Conductivity (lS cm-1); Si as meq l-1 SiO2; Alk: alkalinity as mg l-1 HCO3; Ca, Mg, Na, K, SO4, Cl, NO3, and F as meq l-1; IBE: Ion Balance Error (as %); (*) Borehole; (#) Samples without temperature data (water temperatures recorded at a different sampling event or at neighboring galleries was assumed for PHREEQC simulations). Italic: Samples without cations concentrations, Na was assumed to equilibrate anions in those samples for PHREEQC simulations. Data source: CIA and CHIODINI (1993) Sample name Agujero del agua Almagre (El) Barranco de ´ nimas las A Barranco de Vergara Bilbao Canal (La) (1) Canuto Caramujo Nuevo Cumbre (La) Encarnacio´n y Santa ´ rsula U Fuente de Pedro (2) Fuente Frı´a Gotera (La) Hondura (La) (1) Hoya de la Len˜a Hoya del Cedro Hoya del Pino Junquillo (El) Lomo Colorado Longueras (Las) Luz de Guı´a Madre (La) Monte Frı´o Nia´gara (El) Partido (El) Pinalete (El) Pinalito (El) Porvenir (El) Revento´n (El) Rı´o Bermejo Roque Caramujo (1) Salto de Chen˜eme
Legend
T
pH E.C.
Si
Alk
Ca
Mg
Na
K
SO4
Cl
NO3
-
-
-
-
-
-
-
-
-
1
15.3# 8.61
2 3
25.0 8.50 3760 1.07 1966 0.65 14.2 22.6# 8.23 1728 -
4
17.0
631
F
IBE
-
-
20.42 2.54 0.75 4.97 0.08 -
0.08 -
0.33 -
7.90 1960 1.00 1207 0.69
5.02 14.51 1.54 0.86 0.54 0.11
0.30
0.34
5 6 7 8
19.0 26.0 12.0 22.0#
7.85 1117 1290 7.23 1697 1.43 1155 6.25 8.1 351 0.22 138 0.45 8.60 540 0.82 315 0.60
21.15 7.75 5.39 1.15 0.80 0.33 0.64 2.18 0.24 0.31 0.35 0.93 3.76 0.39 0.37 0.34
0.07 0.15 0.09
0.01 0.33 0.05
0.97 1.53 2.91
9 10
25.0 27.0
8.50 2210 1.00 1259 1.60 7.65 142 1.00 60 0.04
3.70 14.55 1.81 0.72 0.65 0.15 0.17 1.10 0.05 0.11 0.13 0.08
0.14 0.04
1.46 0.54
11
15.0
8.30 1680 1.10 1091 0.18
5.44 12.21 1.55 0.55 0.51 0.14
0.27
0.03
12 13 14
18.0 7.4 146 0.83 53 0.12 18.0# 8.60 1182 0.72 625 0.28 17.0 7.80 1936 1.07 1205 0.63
0.34 0.96 0.10 0.15 0.47 0.13 1.70 8.44 1.10 0.86 0.61 0.08 4.72 15.55 1.18 1.98 0.67 0.08
0.02 0.33 0.17
3.90 2.65 1.26
15
21.7
7.41 2320 1.18 1226 3.90
7.17 12.52 1.38 2.20 1.50 0.20
0.04
1.86
16
11.0
6.70 2415 1.02 1577 1.40
4.67 18.96 1.80 0.40 0.76 0.11
0.52
1.51
17 18 19
20.5 18.0 37.0
7.3 1316 1.53 854 2.99 7.60 1822 0.90 1303 0.55 8.50 3280 1.47 1591 0.23
3.46 9.96 1.25 1.98 0.76 0.08 5.25 16.36 1.52 0.70 0.80 0.21 0.38 34.72 1.67 8.41 2.15 0.00
0.02 0.01 0.08
2.37 1.19 0.22
20
17.0# 8.20 1856 1.07 1087 1.04
4.21 14.09 1.44 1.98 0.65 0.08
0.21
0.09
1.59 4.20 3.95 1.52 2.96 0.52 2.05 2.13
0.03 3.63 0.02 1.47 0.26 3.13 0.01 1.67 0.28 5.49 0.11 2.00 0.16 11.27 0.36 4.05
#
#
21 22 23 24 25 26 27 28 29 30 31
24.0 17.0 18.0# 29.0 18.7 18.0 18.0# 17.0 17.0# 26.7 17.0
8.10 7.60 8.70 6.50 7.80 8.60 6.50 8.60 8.30 8.40 8.40
32
16.0
6.37
654 2050 1461 745 904 1254 1284 1018 1561 498 590
1.23 334 1.36 1.10 994 0.92 0.43 867 0.21 1.20 357 1.54 1025 0.40 756 0.70 2.17 814 1.02 0.40 701 0.81 0.43 1079 547 0.82 390 0.92
541 1.15
293 0.58
1.26
2.48 14.40 13.05 3.00 17.68 7.75 11.85 6.48 20.38 9.49 3.54
0.70 1.28 1.03 0.77 1.54 0.35 0.98 0.31
0.50 2.37 1.68 0.36 0.90 0.45 0.61 1.79 0.32
0.54 1.42 0.90 0.24 0.88 0.74 0.34 0.49 0.83 0.52 0.30
0.04 0.08 0.09 0.14 0.13 0.04 0.20 0.07 0.11
3.74 0.60 0.90 0.35 0.07
0.02
0.33
Vol. 165, 2008
Carbon Dioxide Discharge
153
Table 1 (Contd.) Sample name
Legend
Salto del Fronto´n Santa Teresa Tamuja Trinidad (La) Vergara 2 S-1 MM * S -2 EP *
33 34 35 36 37 38 39
T
23.0 19.0# 20.0 17.0 17.0 18.0 18.0
pH E.C.
8.10 8.60 7.90 7.80 8.40 6.53 6.69
1242 1200 976 770 1650 2210 1418
Si
Alk
Ca
Mg
Na
K
SO4
0.97 728 0.48 2.31 9.15 1.12 0.74 0.43 764 0.24 2.63 9.10 1.00 1.24 0.43 580 4.43 2.82 3.74 1.17 1.50 0.28 234 2.00 1.60 3.89 0.35 1.30 0.97 1014 0.70 4.24 12.28 1.61 0.97 0.85 1347 1.18 7.84 13.54 2.40 0.71 0.85 844 0.50 4.03 9.44 1.43 0.56
Cl
NO3
F
IBE
0.66 0.75 1.10 2.96 0.65 1.63 0.45
0.13 0.14 0.06 0.20 0.15 0.28 0.21
0.25 0.27 0.02 0.01 0.34 0.07 0.20
2.53 6.98 0.08 2.94 0.23 0.38 0.48
variations in water composition such as seasonal effects should be expected. Unfortunately, all the galleries have not been sampled at the same time. This fact limited the availability of the data. However, if the dissolved inorganic carbon (DIC) determined at the gallery mouths is plotted for the different years (Fig. 2), most of the galleries do not present significant variations. That suggests that these sparse data can be used to analyze the conditions of the aquifer before the present cycle of volcanic activity without introducing significant errors in the results.
500
1991-2001 1994 1997
450
DIC (mg l -1 C)
400 350 300 250 200 150 100 50
TAMUJA
VERGARA 2
SALTO DE LFRONTON
SALTO DE CHEÑEME
RIO BERMEJO
ROQUE CARAMUJO 1
PINALITO (EL)
MADRE (LA)
NIAGARA (EL)
LUZ DE GUIA
LONGUERAS (LAS)
JUNQUILLO (EL)
LOMO COLORADO
HOYA DEL CEDRO
HOYA DE LA LEÑA
GOTERA (LA)
HONDURA (LA) (1)
FUENTE DE PEDRO 2
CUMBRE (LA)
ENCARNACION Y SANTA URSULA
CARAMUJO NUEVO
BARRANCO DE VERGARA
ALMAGRE (EL)
BARRANCO DE LAS ANIMAS
AGUJERO DEL AGUA
0
Figure 2 Comparison between the dissolved inorganic carbon in the galleries during 1991–2001, 1994 and 1997. Most galleries show small differences in DIC concentrations.
154
R. Marrero et al.
Pure appl. geophys.,
3.2. Water Mass Balance The water input to Las Can˜adas aquifer (hereafter UH2Or; in l s-1) takes place only through rainfall and snow infiltration because the high permeability of the rocks filling the recharge zone and the moderate precipitation do not allow for the formation of superficial water bodies such as rivers or lakes. The outputs of water from the aquifer are: 1) The controlled discharges through the galleries (hereafter UH2Oc; in l s-1) of the different sectors (Fig. 1) and 2) the non-controlled discharges towards neighboring aquifers (hereafter UH2Onc; in l s-1). Flow data are available for the galleries included in controlled discharges while for other galleries included in the non-controlled discharge, flow and chemical data are not available. Direct observations of the water level in Las Can˜adas volcanic aquifer in boreholes S-1 and S-2 during the period 1994–1999 (FARRUJIA et al., 2001b) suggest that the water table has descended during that time period. This water descent must be considered in the water balance of the aquifer. The equation describing the water mass balance for Las Can˜adas is as follows: UH2 Or þ UH2 Owtd ¼ UH2 Oc þ UH2 Onc:
ð1Þ
The total water drained by the galleries accounts for 1193 l s-1 (UH2Oc in Table 3). In order to obtain the non-controlled water flow (UH2Onc), the flow of water that produces the descent in the water table (hereafter UH2Owtd; in l s-1) must be calculated. This term is calculated using the following equation: UH2 Owtd ¼ Aq m Rwtd;
ð2Þ
2
where Aq (in m ) represent the aquifer surface area, m represents the drainable porosity of the aquifer; and Rwtd is the rate of water table descent (in m y-1). 3.3. CO2 Mass Balance For the mass balance of CO2, several assumptions are made: 1) There are no carbonate rocks present within Las Can˜adas Caldera, only volcanic rocks; then the majority of the CO2 transported advectively will be assumed to have volcanic origin; 2) as the water infiltrates, it equilibrates with the soil CO2 present in the soils; 3) as the water circulates within the rocks in the saturated zone, CO2 could be added as gases released from the magmatic environment, and when the water reaches the saturation index of carbonates, these minerals can precipitate from the water. However, with the data available there is no way to distinguish the carbonate minerals inputs and outputs to this system. Assuming steady-state conditions (inputs equals to outputs), the CO2 mass balance for the aquifer can be written as: UCO2 t ¼ UCO2 c þ UCO2 nc þ UCO2 s - UCO2 rain;
ð3Þ
where UCO2t (in t d-1) is the total CO2 degassing from the volcanic-hydrothermal system of the Teide volcano; UCO2c (in t d-1) is the controlled CO2 flux discharges of the 37
Vol. 165, 2008
Carbon Dioxide Discharge
155
galleries; UCO2nc (in t d-1) is the non-controlled CO2 flux discharges; UCO2s (in t d-1) is the diffuse soil CO2 emissions; and UCO2rain (in t d-1) is the CO2 flux from rain water.
4. Results 4.1. Water Balance Infiltration studies by the CIA have estimated an average annual recharge of 280 mm (PLAN HIDROLO´GICO INSULAR, 1992; FARRUJIA et al., 2001b, 2004). Considering that the surface area of Las Can˜adas is approximately 144 106 m2, the recharge is 41 hm3 y-1 or 1300 l s-1 (NAVARRO, 1994). Las Can˜adas depression is filled with relatively young phonolytes, trachytes and basalt layers of almost non-altered lava flows and fall deposits that are likely to preserve a high initial porosity. According to CUSTODIO and LLAMAS (2001), vesiculated basalts and pyroclasts have an effective porosity ranging from 5 to 20%. The higher limit for this range is assumed to correspond to the vesiculated basaltic flows. The average descent observed during the 1994–1999 period in the S-1 borehole located in the central region of the caldera (Fig. 1) can be used to calculate the rate of water table descent (FARRUJIA et al., 2001b) as 0.42 m y-1. Using the previous values for the different terms of Equation (2), UH2Owtd is 384 l s-1. Equation (1) then gives 491 l s-1 for UH2Onc. Finally the total UH2O discharge (UH2Onc + UH2Oc) is 1684 l s-1 (53.1 hm3 y-1). This value is equal to the recharge plus the annual volume of water taken from storage that produces the water table descent. It should be noted that the non-controlled discharge calculated in this way is uncertain. The uncertainties in this calculation of the non-controlled discharges (UH2Onc) include: 1) Errors in the calculation of the recharge (UH2Or) due to errors in the determination of the different terms of the superficial water balance (rainfall, evapotranspiration, and infiltration), and 2) errors in the calculation of UH2Owtd because we have considered only the descent in borehole S-1 that provides incomplete information about the evolution of the entire aquifer. The descent in borehole S-2 has not been considered because this borehole is too close to several producing galleries (Fig. 1) affecting the recorded variation in depth (2.29 m y-1 average between 1995–1999, FARRUJIA et al., 2001a,b) 4.2. CO2 Mass Balance 4.2.1. Controlled CO2 flux discharges (UCO2c). This flux was evaluated using the following equation: UCO2 ci ¼ DIC UH2 Oci:
ð4Þ
The water discharge at each gallery (UH2Oci) is known because the CIA keeps good records of these discharges (CIA, Data Base). The chemical data reported in Table 1 was used to determine the total DIC for each sample using the aqueous speciation model
156
R. Marrero et al.
Pure appl. geophys.,
PHREEQC (PARKHURST, 1995). However, the majority of the samples in Table 1 were taken at the gallery mouth instead of the gallery bottom. Variations of total DIC occurring along the gallery channel can be due to two processes: CO2 degassing from the water to the air, and precipitation of carbonate minerals because the waters are still supersaturated with respect to these minerals at the gallery mouth. Water degassing processes occur along the length of the gallery from the bottom to the discharge point at the gallery mouth. Precipitation of calcite (CaCO3) within the channel has been reported (COELLO, 1973). Towards the interior of the gallery, the sudden release of CO2 has been reported as soon as a dry gallery has water leaking from the walls (COELLO, 1973). These degassing and precipitation processes indicate that the DIC found at the gallery bottom is considerable higher than the value at the gallery mouth where the groundwater makes contact with the atmosphere. For a few galleries, chemical compositions of the waters at the gallery mouth and bottom have been determined (Table 2). Calculation of the difference between CO2 concentrations as well as the sum of Ca + Mg concentration difference at those two points shows that carbonate precipitation could be responsible for only 0.2 to 23% of the CO2 lost. Further, the Ca concentration does not show inverse correlation with pH (Table 1). These points argue against substantial subsurface precipitation of calcite during groundwater transport along the gallery channel. The main process responsible for the change in CO2 concentrations is water degassing rather than carbonate precipitation. The composition of the air within the gallery is a function of the rate of degassing and the velocity of air renovation in the gallery, which is determined by variations in atmospheric pressure. The calculation of the total CO2 released by each gallery requires the restoration of the DIC to the conditions at the gallery bottom. Table 2 Galleries with two sampling points along the channel used to calculate the empirical CO2 degassing factor, as well as their flow rate (UH2O), Dissolved inorganic carbon (DIC) and controlled CO2 flux discharges (UCO2c) at each sample point, distance between samples (L), difference between DIC at the bottom and mouth at each gallery (DCO2), difference between Ca + Mg concentration at the bottom and mouth at each gallery (D(Ca + Mg)), and percent difference in concentration of D(Ca + Mg) DCO-1 2 between the bottom and mouth at each gallery produced by solid precipitation. * = No data available Sample name
Almagre (El) Nia´gara (El) Lomo Colorado Fuente Frı´a
Analysis date
1995 1995 1988 1988 1991 1988 1984 1973 Luz de Guı´a 1998 1994
UH2Oc (l s-1)
77.0 42.2 8.0 10.0 4.6
DCO2 DIC (mol l-1 (meq CO2) l-1)
1.347 1.854 3.067 7.999 1.034 1.257 4.431 9.887 2.407 3.580
E-03 23.05 E-03 E-04 22.42 E-04 E-03 10.14 E-03 E-05 2.48 E-05 E-04 5.33 E-04
D(Ca + Mg) (meq l-1)
100 (DCa + Mg) DCO2 -1 (% meq meq-1)
UCO2c (kg d-1)
L (m)
5.37
23.31
3500 0.96
0.05
0.22
0.06
0.59
0.22
8.87
*
*
8925 12284 1118 2917 715 869 38 85 96 142
CO2 degassing factor F (kg m-1 d-1)
2975 0.60 4000 0.04 1010 0.05 2950 0.02
Vol. 165, 2008
Carbon Dioxide Discharge
157
Since few galleries have been sampled at points located at different horizontal distances within the tunnel (Table 2), it is not possible to use an analytical or numerical model to restore the DIC of the water. In addition, some of the galleries have complex layouts (e.g., galleries having two branches or multiple water additions), which makes difficult the application of analytical models to find the DIC in the gallery bottom. However, two factors that are likely to affect the degassing process are the length of the gallery and fluid velocity or discharge. Table 2 shows values of gallery length, water discharge at the gallerie (UH2Oci), change in total DIC, and a parameter defined as the degassing factor F (mass of CO2 degassed per unit length per day in kg m-1 d-1). The five galleries having two sampling points along them and simple channel morphology and water discharge were used to investigate how the degassing factor depends on the discharge. The degassing factor F was plotted versus water discharge (Table 2, thick line in Figs. 3a and 3b). A good correlation was observed between F and UH2Oci, with r2 = 0.98, which is statistically significant at the 99% confidence level when the test of significance of the correlation coefficient is applied (SWANS and SANDILANDS, 1995). The equation of the best fitted line is: F ¼ 0:0127 UH2 Oci:
ð5Þ
Mass transfer theory at the liquid-gas interphase (THIBODEAUX, 1996) can be used to understand the meaning of this linear behavior of the CO2 degassing factor. Using mass transfer theory, a series of curves was constructed and compared with our data (see Appendix for the development of the degassing model). The factors that define the magnitude of F are the initial DIC (DIC1), the number of equivalents of cations (cat), the width and depth of the channel (w and h), the total length of the gallery, the water discharge (UH2Oci), and the concentration of CO2 in the atmosphere of the gallery (qA1). A few measurements of the pCO2 in the atmosphere of the galleries have been reported (ALBERT-BELTRAN et al., 1990). These values range from 0.01 atm to 0.1 atm. The average cations in the galleries is 0.02 eq l-1 and the maximum around 0.04 eq l-1, the length ranges from 1643 m to 5058 m, the average DIC is 0.03 mol l-1 with values as high as 0.04 mol l-1. These condition ranges have been used to model the degassing factor as a function of water discharge. The width and depth of gallery channels range from around 25 to 50 cm. The theoretical modeled curves are illustrated in Figures 3a and 3b. The linear trend for the degassing factor versus the water discharge falls clearly within the modeled curves, suggesting that water discharge is the controlling factor in water degassing. Finally, to restore the dissolved CO2 flux discharged at the gallery bottom (#UCO2ci, kg d-1) the following equation is used: #
UCO2 ci ¼ UCO2 ci þ ðF LÞ:
ð6Þ
UCO2ci is the flux of CO2 at the sampling point and L (m) is the distance between the sampling point and the gallery bottom.
158
A
R. Marrero et al.
Pure appl. geophys.,
3.0
pCO2 0.01 atm
2.5
NS
Y = 0.0127 X R2 = 0.9798
F (kg d-1 m-1)
2.0
IO AT
L .02;
300
0; S
25 x
25
0
;C
DIC
6 O.O
1.5
Las Cañadas CO2 degassing factor
1.0
DIC O.O
6; CATIO
NS 0.04
; S 25 x
; L 3000
25
S 25 x 25 NS 0.02; L 3000; DIC O.O3; CATIO x 50 0.02; L 3000; S 50 S ION CAT 3; DIC O.O 25 S 0.02; L 5000; S 25 x DIC O.O3; CATION
0.5
0.0 0
10
20
30
40
50
60
70
80
90
100
-1 H2O (l s )
B
3.0
pCO2 0.11 atm 2.5
Y = 0.0127 X 2 R = 0.9798
F (kg d-1 m-1)
2.0
D
.O IC O
6; C
O ATI
NS
5x
;S2
000
L3 .02;
25
0
1.5
Las Cañadas CO2 degassing factor
1.0
DIC O.O
0.5 DIC O.O3; CAT
0; S 50 x 50
IONS 0.02; L 300
NS 0.04;
6; CATIO
L 3000; S
25 x 25
S 0.02; L 3000; S 25 x
DIC O.O3; CATION
DIC O.O3; CATIONS 0.02; L 5000; S
25
25 x 25
0.0 0
10
20
30
40
50 H2O
60
70
80
90
100
(l s-1)
Figure 3 (A) Plot of diffuse degassing factor vs. UH2Oc at pCO2 0.01 atm and (B) Plot of diffuse degassing factor vs. UH2Oc at pCO2 0.11 atm. DIC as mol l-1; cations as eq l-1; L length of the gallery in m; S section of channels in cm2.
With the #UCO2ci from Equation (6) it is possible to find the DIC at the gallery bottom (#DIC) using the equation: #
DIC ¼ UH2 Oci1
#
UCO2 ci:
ð7Þ
Results for the DIC at the gallery mouth and the gallery bottom are presented in Table 3. Note that DIC discharges at the gallery mouth are considerably lower than the restored values computed at the gallery bottom, especially for the galleries with higher water discharge.
Vol. 165, 2008
Carbon Dioxide Discharge
159
Table 3 Dissolved inorganic carbon, partial pressure of CO2, and controlled discharge of CO2 (UCO2c) at the gallery mouth and at the gallery bottom, as calculated with the use of the degassing factor. DIC UCO2c pCO2 (mol (106 mol (atm) l-1) y-1) GALLERY MOUTH
# DIC UCO2c #pCO2 (atm) (mol (106 mol y-1) l-1) GALLERY BOTTOM
Legend (Table 1 and Fig. 1)
Distance from sample to depth zone (m)
UH2Oc (l s-1)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
3000 2500 4949 3000 2500 1665 2600 3670 2825 3662 2000 1820 4567 4543 4222 3510 3834 3350 5058 3190 3300 3814 4356 2700 3830 4163 1643 2820 2040 4200 4000 252 4700 3300 3072 3600 2256 – –
1.3 75.1 13.3 318.4 16.5 11.2 3.0 6.5 21.3 8.1 6.2 10.0 3.3 29.7 42.0 237.2 15.7 65.0 6.1 5.0 4.5 4.0 12.4 28.2 10.4 8.0 11.9 2.0 20.0 2.9 35.9 2.5 21.4 13.3 10.1 6.7 103.6 – –
– 1.65E-04 – 5.29E-04 5.62E-04 2.09E-03 4.64E-05 2.65E-05 1.37E-04 4.68E-05 1.94E-04 8.36E-05 5.43E-05 6.66E-04 1.52E-03 1.21E-02 1.45E-03 1.12E-03 1.26E-04 2.36E-04 8.88E-05 8.81E-04 5.62E-05 3.65E-03 5.75E-04 6.41E-05 8.09E-03 6.14E-05 1.91E-04 2.67E-05 5.45E-05 4.73E-03 1.91E-04 6.41E-05 2.48E-04 1.38E-04 1.38E-04 1.46E-02 6.16E-03
7.47E-03 3.06E-02 1.87E-02 2.01E-02 2.16E-02 2.09E-02 2.29E-03 5.00E-03 2.01E-02 1.02E-03 1.77E-02 9.50E-04 9.98E-03 2.03E-02 2.15E-02 3.80E-02 1.54E-02 2.24E-02 2.52E-02 1.77E-02 5.47E-03 1.71E-02 1.36E-02 9.50E-03 1.73E-02 1.20E-02 2.14E-02 1.12E-02 1.77E-02 8.66E-03 6.29E-03 9.53E-03 1.20E-02 1.22E-02 9.62E-03 3.95E-03 1.63E-02 3.66E-02 2.00E-02
0.31 72.50 7.85 202.09 11.24 7.40 0.22 1.02 13.53 0.26 3.47 0.30 1.03 19.01 28.49 283.95 7.62 45.90 4.87 2.82 0.78 2.16 5.33 8.44 5.68 3.03 8.02 0.70 11.15 0.80 7.11 0.75 8.06 5.10 3.07 0.83 53.36 – –
0.0010 0.0048 0.0070 0.0122 0.0138 0.0625 0.0009 0.0007 0.0040 0.0014 0.0042 0.0020 0.0013 0.0154 0.0404 0.2319 0.0371 0.0267 0.0050 0.0054 0.0025 0.0203 0.0013 0.1187 0.0140 0.0015 0.1926 0.0014 0.0044 0.0008 0.0013 0.1059 0.0052 0.0016 0.0063 0.0032 0.0032 0.3465 0.1466
1.74E-02 3.89E-02 3.51E-02 3.01E-02 2.99E-02 2.65E-02 1.09E-02 1.72E-02 2.95E-02 1.32E-02 2.44E-02 6.98E-03 2.51E-02 3.53E-02 3.55E-02 4.96E-02 2.81E-02 3.35E-02 4.20E-02 2.83E-02 1.64E-02 2.98E-02 2.81E-02 1.84E-02 3.00E-02 2.58E-02 2.69E-02 2.05E-02 2.44E-02 2.26E-02 1.95E-02 1.04E-02 2.75E-02 2.31E-02 1.98E-02 1.59E-02 2.38E-02 – –
TOTAL
–
1193
–
–
838
–
–
CCO2 (mol l-1)
#
0.71 92.12 14.73 301.92 15.56 9.35 1.03 3.51 19.83 3.37 4.76 2.20 2.60 33.13 47.02 370.98 13.92 68.67 8.11 4.50 2.34 3.75 10.98 16.38 9.84 6.51 10.06 1.29 15.42 2.09 22.10 0.81 18.55 9.69 6.32 3.36 77.79 – – 1235
0.0024 0.0081 0.0131 0.0219 0.0250 0.0936 0.0047 0.0027 0.0061 0.0195 0.0070 0.0153 0.0037 0.0321 0.0807 0.3465 0.0788 0.0479 0.0105 0.0105 0.0084 0.0419 0.0033 0.2457 0.0280 0.0038 0.2856 0.0030 0.0072 0.0061 0.0046 0.1228 0.0139 0.0035 0.0150 0.0144 0.0056 – – –
In this model it is assumed that the galleries transport water from the deeper zone or bottom throughout an open channel (as it happens in the majority of the galleries) to the gallery mouth. However, some galleries present water leaking from the walls at different points instead of only the deeper point. Other galleries have total or partial tubing of the channel, thus limiting water degassing. In many galleries, it is impossible to get accurate information about water contributions along the channel and the existence of tubed
160
R. Marrero et al.
Pure appl. geophys.,
sections because of limited access. These limitations suggest that instead of a single value for the total CO2 discharged by the measured galleries UCO2c, it is better to determine a range given by the two extreme cases: 1) Assuming no degassing and DIC at the gallery mouths equal to DIC at the bottom, and 2) considering the DIC restored at the bottom using the CO2 degassing factor. For the first case, the UCO2c found using Equation (4) is 838 106 mol y-1. For the second case, #UCO2c evaluated using Equation (6) is 1235 106 mol y-1. These values indicate a flux of CO2 discharged throughout the 37 galleries between 101 and 149 t d-1. 4.2.2. Non-controlled CO2 flux discharges (UCO2nc). The water discharged from Las Can˜adas aquifer to neighboring aquifers and non-measured galleries contains DIC acquired within Las Can˜adas. The total water discharged by the non-controlled discharges (UH2Onc) and the concentration of DIC in that water can be used in Equations (4) and (7) to evaluate the flux of CO2 associated to that water output. This value can be obtained using a DIC value assumed as the weighted average of the DIC values at the gallery mouth with respect to the water discharges at each gallery, or as the weighted average of the DIC values at the gallery bottom. The weighted average DIC for the gallery mouth and the gallery bottom was 0.022 and 0.033 mol l-1, respectively. These values give a range of 42 to 62 t CO2 d-1 for the advective fluxes of CO2 transported by the noncontrolled discharges. Finally, the total advective groundwater transport of CO2 leaving the aquifer gives values ranging from 143 (101 + 42) to 211 (149 + 62) t CO2 d-1 for the two extreme degassing cases. 4.2.3. Diffuse soil CO2 emissions (UCO2s). Since 1997, several studies have been carried out at Las Can˜adas to determine the diffuse soil degassing of this volcanic-hydrothermal system, especially at the summit cone of Teide volcano (e.g., HERNA´NDEZ et al., 2000). Measurements of diffuse emission of carbon dioxide were carried out by using the accumulation chamber method (PARKINSON, 1981; BAUBRON et al., 1991; CHIODINI et al., 1996). However, only the 1997 survey was comprehensive and included the Las Can˜adas Caldera as well as the summit cone of Teide volcano (HERNA´NDEZ et al., 1997). The total diffuse soil emissions of CO2 were estimated as 563 t CO2 d-1 for an area of 197.9 km2 (this value includes Las Can˜adas Caldera and surrounding area, see Fig. 1), with 101 t CO2 d-1 corresponding to the summit cone of Teide volcano and 462 t CO2 d-1 to the soils of the caldera and surrounding area. The following years, the studies of soil diffuse degassing were done only in the summit cone of the Teide volcano that has an area of only 0.53 km2 and the highest fluxes and concentrations of CO2 in the soils (HERNA´NDEZ et al., 2000). The fluxes of CO2 in soils obtained at the summit cone of Teide volcano were 101, 97, 20, 380, 73 and 69 t d-1 for 1997, 1999, 2000, 2001, 2003 and 2004 respectively. If the lowest and highest numbers (years 2000 and 2001) are not considered, the estimated CO2 flux for 1997 (101 t CO2 d-1) can be assumed as a representative
Vol. 165, 2008
Carbon Dioxide Discharge
161
value of the quiescent stage of Teide volcano. HERNA´NDEZ et al. (1997) obtained a total flux of 462 t CO2 d-1 involving Las Can˜adas Caldera and its surroundings for an area of 197.9 km2. For an area of 144 km2 that corresponds to only the area within Las Can˜adas Caldera, the average total flux found by HERNA´NDEZ et al. (1997) for this area is 336 t CO2 d-1. Finally, the total diffuse soil degassing UCO2s can be assumed as 437 t CO2 d-1 (336 t CO2 d-1 from the caldera floor and 101 t CO2 d-1 from the Teide cone). Previous investigations in boreholes S-1 and S-2 (Fig. 1) in Las Can˜adas (SOLER et al., 2004) have noted that the pCO2 just above the saturated zone in these boreholes is high (e.g., 0.3 atm in S-1), as it is the concentration of H2CO3 in the aquifer (e.g., 502 mg l-1 in the deeper measured point). According to the authors of this paper, these high values of pCO2 contrast with the low concentrations in the soils measured by HERNA´NDEZ et al. (2000) (0.0029 atm). However, it is possible to compare the CO2 flux predicted by the data collected in the boreholes with the flux determined by HERNA´NDEZ et al. (2000). For the mass transfer process of CO2 from the groundwater to the vadose zone, the following equation (THIBODEAUX, 1996; CARON et al., 1998) can be used: F ¼ KL qCO2 soils qCO2 ; ð8Þ where F is the flux of CO2 upward from the groundwater, qCO2 soils (g m-3) is the concentration or pCO2 in the soils away from the boundary between the interstitial air and the water table, and qCO2 (g m-3) is the concentration or pCO2 at the water table, and KL (meter hour-1) is the mass transfer coefficient between water and interstitial air in the porous media. The concentration of CO2 measured close to the water table and within the vadose zoneqCO2 (0.3 atm) also can be obtained using Henry’s law and the measured concentration of H2CO3 in the water (502 mg l-1). For the mass transfer coefficient (KL), CARON et al. (1998) have done experiments to find the mass transfer coefficient of CO2 from sandy soils to moving groundwater. They found a value of 1.9 10-4 m h-1 for waters with pH equal to 6.4 and 3.1 10-4 m h-1 for waters with pH equal to 6.1. Using the value for the higher pH (closer to the pHs observed at Las Can˜adas, Table 1), and the pCO2 found in S-1 and the soils, we derive a flux of 2.1 10-6 t CO2 d-1 m-2. Taking the area of Las Can˜adas as 144 km2, a total flux of 303 t CO2 d-1 is obtained. This value is comparable with the value obtained by HERNA´NDEZ et al. (1997b). We need to consider the limitations of this approach: 1) We have used only one point within the caldera for our calculation; it is possible that there is variable groundwater composition within the caldera, and 2) the mass transfer coefficient that we have used could be slightly different at the pH of Las Can˜adas waters. However, these results suggest that the values reported by HERNA´NDEZ et al. (2000) likely represent the average diffuse soil degassing discharge of CO2 at Las Can˜adas during quiescence periods. 4.2.4. CO2 flux from rain water (UCO2rain). In order to verify that the contributions of CO2 from external sources at Las Can˜adas aquifer are negligible and that the main component is endogenous CO2 (CUSTODIO et al., 1987; SOLER et al., 2004), we have
162
R. Marrero et al.
Pure appl. geophys.,
calculated how much CO2 could come from rain water, assuming chemical equilibrium between the air and rain water. The total concentration of CO2 dissolved in the rain water was calculated simulating equilibrium between pure water and atmospheric CO2, 1 atm and temperature equal to 15C (air temperature annual average at Las Can˜adas Caldera). The total concentration of CO2 dissolved in the rain water was found as 1.675 10-5 mol l-1. The total concentration of CO2 dissolved in the rain water multiplied by the recharge UH2Or gives the flux of CO2 in the recharge. This value was 0.082 t CO2 d-1. This low value indicates that the dissolved CO2 reaching the aquifer due to the infiltration of rain water is negligible when it is compared with the internal contributions of the volcano to the flowing water (between 143 and 211 t CO2 d-1). 4.3. Partial Pressure of CO2 (pCO2) at Las Can˜adas Aquifer The pCO2 (atm) in the groundwaters of Las Can˜adas aquifer can be used to identify the zones of this system that are better connected with the Teide magmatic system. The pCO2 for the different water samples collected at the gallery mouths was determined. The pCO2 at the gallery mouth was calculated using the activity H2CO3 (aH2CO3) obtained with the speciation and reaction path modeling program PHREEQC (PARKHURST, 1995) and the Henry’s constant for CO2 (H): pCO2 ¼ aH2 CO3 ðHÞ1 :
ð9Þ
For these calculations, all the analyses available for the period before the reactivation of the system (1991–2001) were used, without considering if they were inside (Table 3) or outside the Las Can˜adas aquifer (Table 4). In this way, variations of pCO2 in and around the aquifer could be observed. Several anomalous zones can be observed in Figure 4, the most important are the boreholes S-1 and S-2 (0.35 and 0.15 atm, respectively). Galleries numbers 16, 27, and 78 have high pCO2 compared to other galleries. It should be noted that water in these galleries is transported within closed pipes in several sections of their length, limiting the release of CO2 to the air and the decrease in pCO2. However, in other galleries with high pCO2, such as 24, which is located at the western end of the Las Can˜adas Caldera (Fig. 1), water is transported by means of open channels, suggesting that the levels of pCO2 within the aquifer are even higher. On the other hand, the pCO2 at the gallery bottoms was calculated only for those galleries that seem to be discharging water from Las Can˜adas aquifer to avoid interference of other nearby volcanic centers along the NE and NW rift zones. The pCO2 at the gallery bottom was calculated using the #DIC obtained with Equation (7) and the following equation, according to the definition of DIC as the sum of carbonate species: #
1 2 1 pCO2 ¼# DIC ½KCO2 ½1 þ K1 ðaþ þ K2 K1 ðaþ HÞ H Þ :
ð10Þ
As the water pH at the gallery bottom is not available, the pH at the gallery mouth was used and assumed to have small variations (around 0.5 pH units difference is observed at
Vol. 165, 2008
Carbon Dioxide Discharge
163
Table 4 Partial pressure of CO2 of different sampling points at the gallery mouth at central Tenerife Island Sample name
Legend (Fig. 4)
Sample name
pCO2 (atm)
Abandonada (La) Acaymo Acevin˜o (1) Aguas del Valle Ancon de Juan Marrero
40 41 42 43 44
0.0001 0.0001 0.0002 0.0103 0.0010
Angeles (Los) Arguayo o Mollero (El) Bebederos (Los) Begon˜a Cerca de la Fortuna (La) Cercado de la Vin˜a Chajan˜a Chamoco Chupadero (El) Cuevas Viejas Dieciseis de Mayo Durazno (El) Fuente Bella o Fuente del Valle Gambuezo de Tamadaya Goteras (Las) Hondura de Fasnia Honduras de Luchon Jurado (El)
45 46 47 48 49 50 51 52 53 54 55 56 57
0.0106 0.0158 0.0007 0.0028 0.0038 0.0030 0.0109 0.0042 0.0245 0.0157 0.0028 0.0051 0.0012
58 59 60 61 62
0.0093 0.0007 0.0010 0.0022 0.0016
Lomo del Quicio Majada (La) Mayatos (Los) Milagro (El) (2) Mozas (Las) o Tamaimo Oportunidad (La) Piedrita (La) Quince de Septiembre Ranas (Las) Rebosadero (El) Reina (La) Rio de la Fuente (1) Rio de la Plata Saltadero de Sosa San Fernando (1) San Fernando (3) San Isidro (1) San Jose o Aguas de San Jose Santa Margarita Sauces (Los) Senor del Valle (El) Sorpresa (La) Topo y Chija
Legend (Fig. 4)
pCO2 (atm)
64 65 66 67 68
0.0001 0.0048 0.0070 0.0009 0.0006
69 70 71 72 73 74 75 76 77 78 79 80 81
0.0034 0.0070 0.0044 0.0034 0.0051 0.0006 0.0003 0.0060 0.0005 0.1027 0.0713 0.0130 0.0001
82 83 84 85 86
0.0014 0.0535 0.0157 0.0071 0.0027
galleries sampled in both points). This approximation produces underestimation of the pCO2 at the galleries bottom because pH increases as CO2 is degassed from the solution. pCO2 values at the gallery mouth and at the gallery bottom are presented in Table 3 and a distribution map of pCO2 values at the gallery bottom is presented in Figure 5. The larger differences are observed in the galleries of higher discharge. This can be explained knowing that the restored DIC at the gallery bottom depends strongly on the water discharge. Higher water discharge implies higher degassification and the correction for the pCO2 is also greater. However, Figure 5 shows that the anomalies are located at both boreholes and the galleries of the southwest sector, which seem to have good connection with the Teide volcanic-hydrothermal system.
5. Discussion The Las Can˜adas volcanic aquifer total water discharge (UH2O = UH2Onc + UH2Oc) is 1684 l s-1 (53.1 hm3 y-1). This value is higher than aquifer total discharge
164
R. Marrero et al.
Pure appl. geophys.,
3150000
3140000 34 23 28 5 3 26 25 13 33 11 19 37 4
29 7 20
3130000
62
84 61 49
79
NW
46 70 66
14
Rift
17
15
16
3120000
82
38
10 12
64 67
86
3110000
2
Teide Volcano
36 32 3518 21 30 24 22 47 6
27
9 1 8 31 39
59
48 40
0.40
ft Ri 77 NE74 81 52
0.30
42 76 65 75 41 45 50 78 60 53 51 85 83 63 58 73 54 80 44
56 69
N-S Rift
UTM (m)
68
0.20 0.15 0.10 pCO2 (atm)
72 71 55
0.05 0.02 0.01
5743
0.00
ATLANTIC OCEAN
0
5
10
KILOMETERS 3100000 310000
320000
330000
340000
350000
360000
370000
380000
UTM (m) Figure 4 Partial pressure of CO2 at gallery mouths at central Tenerife Island. Galleries inside and outside Las Can˜adas Caldera are considered. Dark dots represent the sample point (gallery mouths and borehole). Gridding method is Natural Neighbor.
calculated in the Island Hydrologic Plan of Tenerife for 1985 and 2000 (1021 l s-1 and 945 l s-1, respectively) for the same region. It is assumed that the recharge estimated by other authors (1300 l s-1, BRAOJOS et al., 1997) is correct, as well as the effective drainable porosity assumed for this aquifer (CUSTODIO and LLAMAS, 2001). However, considering that the water table is descending (as is observed in S-1 and S-2, FARRUJIA et al., 2004), the discharges from the aquifer cannot be lower than the recharge plus an additional volume of water draining from the aquifer storage (384 l s-1) , which is the basic assumption in the water balance approach. The total emission of CO2 from the volcanic-hydrothermal system of Teide (UCO2t) obtained through Equation (3) ranges from 640 (211 + 437) to 572 (143 + 437) t CO2 d-1. Between 33 and 25% of this UCO2t is discharged laterally through the groundwater of Las Can˜adas aquifer. This large DIC flux demonstrates the ability of cold groundwater to absorb and transfer magmatic CO2, even at volcanoes during quiescence periods (EVANS et al., 2002). Comparison of the CO2 discharged by the groundwater of Las Can˜adas aquifer (11.8 to 17.5 108 mol y-1) with other volcanic aquifers of the world (Table 5) shows that the fluxes are similar in magnitude to the aquifers of Vesuvio in Italy, and Mammoth
Vol. 165, 2008
Carbon Dioxide Discharge
165
3140000
UTM (m)
3135000
3130000
3125000
3120000 310000
320000
330000
340000
350000
360000
UTM (m) Figure 5 Partial pressure of CO2 at Las Can˜adas aquifer at gallery bottom. Only galleries intercepting Las Can˜adas aquifer are considered here. Dark dots represent the sample points (gallery mouths and borehole). Gridding method is Natural Neighbor.
Mountain in USA, but lower than Etna volcano aquifer. Etna is a volcano with higher and more frequent activity (EVANS et al., 2002; GAMBARDELLA et al., 2004). Hydrothermal systems in carbonate rocks such as Albani Hills (GAMBARDELLA et al., 2004), in Italy, also present higher CO2 groundwater fluxes. However, if the specific CO2 flux uCO2ad (mass of CO2 per unit time per unit area) is calculated, Las Can˜adas aquifer has the highest CO2 emission rate per km2 after Mammoth Mountain. In previous investigations, HERNA´NDEZ et al. (2000) found a concentration of CO2 in the soil atmosphere of Las Can˜adas of only 2900 ppmV (0.0029 atm). In comparison, SOLER et al. (2004) found high concentrations of CO2 accumulated above the water table.
Table 5 Surface area (S), groundwater flow (total UH2O discharged), total CO2 flux discharged by advection (UCO2ad = UCO2c + UCO2nc), and CO2 flux discharged per unit area (Specific uCO2ad) at Las Can˜adas aquifer and other volcanic aquifers of the world. (*) Gallery mouth; (#) Gallery bottom; Data sources: (a) GAMBARDELLA et al. (2004); (b) EVANS et al. (2002) Aquifer Las Can˜adas*, Spain Las Can˜adas#, Spain Vesubioa, Italy Etnaa, Italy Albani Hillsa, Italy Mammoth Mountainb, USA
S (km2)
Total UH2O (1010 l y-1)
144
5.31
153 1322 1516 25
5.05 69 43.8 2.5
Total CO2ad (108 mol y-1)
Specific uCO2ad (106 mol y-1 km-2)
11.8 17.5 9.6 104.0 38.5 4.6
8.2 12.2 6.3 7.87 2.54 18.2
166
R. Marrero et al.
Pure appl. geophys.,
These low concentrations of CO2 at the soils of Las Can˜adas Caldera do not mean low CO2 fluxes at the large soil-air interface. High permeability materials such as those present at Las Can˜adas Caldera cannot retain and store large concentrations of CO2. Consequently, CO2 concentrations can be low. In comparison, low permeability soils could retain CO2 due to low fluid velocities and accumulation of gas. The results of this research show that between 67 and 75% of the total CO2 emitted by Teide volcano during quiescence periods reaches the air-soil interface. The global rate of CO2 emission of subaerial and submarine volcanoes is approximately 400,000 t d-1 (GERLACH, 1991), then total degassing from the volcanichydrothermal system of Teide volcano (UCO2t) supplies about 0.19–0.18% of the global CO2 discharge. The pCO2 values for the gallery mouth as well as for the gallery bottom clearly show the anomalies of high concentration centered at Teide volcano and at the southwest side of Las Can˜adas Caldera. In addition, most of the geochemical and geophysical evidence of the seismo-volcanic crisis at Tenerife Island in 2004–2005 was located at the southwest side of Las Can˜adas Caldera (GOTTSMANN et al., 2006; ALMENDROS et al., 2007; MARRERO et al., submitted). This behavior suggests that this area represents a good connection zone with the volcanic-hydrothermal system of Teide volcano.
6. Conclusions Volcanologists started studying the flux of volcanic gases such as SO2 and CO2 in the plumes, as important indicators of magmatic activity (e.g., STOIBER and WILLIAMS, 1990). During the last decade, the importance of diffuse soil degassing throughout the volcanic edifices to the atmosphere has been stated (BAUBRON et al., 1991; ALLARD et al., 1991; ALLARD, 1992; CHIODINI et al., 1996; CHIODINI and FRONDINI 2001). In our work, and the recent work in other volcanoes of the world (HERNA´NDEZ et al., 1998; 2001, 2003, 2006; NOTSU et al., 2005; PEREZ et al., 2004, 2006), a more complete picture of how volcanoes degas is emerging. It is becoming evident that for the gas budget of volcanoes, the advective transport of gases by flowing groundwater is also important. At Teide volcano, most of the dissolved CO2 is released to the unsaturated zone, nonetheless a significant fraction of the CO2 remains in solution and it is transported away from the volcano by the groundwater flow. It should be noted that the capacity of volcanic aquifers to dissolve and trap CO2 should depend on the relative size of the aquifers and magmatic source, and the magnitude of the recharge. A volcanic aquifer in a region with high rainfall and recharge, and with high permeability rocks can have more water available for the dissolution reactions and could trap a higher fraction of emitted CO2 than a volcanic aquifer in an arid climate and with a similar magmatic source.
Vol. 165, 2008
Carbon Dioxide Discharge
167
At the same time, the results of our research suggest that monitoring of groundwater CO2 can be important for volcanic forecasting and monitoring. A significant increase in this parameter may be related to the abnormal input of magmatic CO2 released by a magma body moving to a shallower level. However, these studies should be complemented with other studies such as C isotopic compositions, and the concentration of other gases and chemical species such as sulfate, chloride, radon, and helium, which can also indicate magmatic inputs.
Acknowledgements We are grateful to G.Chiodini and CIA to provide some of these data, and E. Custodio for his useful comments. Nemesio M. Pe´rez and Pedro A. Herna´ndez thank G. Chiodini for providing us the idea of doing this research work. This work was partially supported by the Cabildo Insular de Tenerife and Gobierno de Canarias under the project Vigilancia Volca´nica de Tenerife, as well as by the Interreg IIIB Azores-Madeira-Canarias community initiative under the projects Alerta and Alerta II.
Appendix 1: CO2 Degassing Model along an Open Gallery Channel In this paper we have assumed that water velocity or discharge is important for water degassing processes. This assumption can be verified generating a family of theoretical curves that relate the CO2 degassing factor F (mass of CO2 degassed per unit length per day in kg m-1 d-1) and water discharge (UH2Oc; l s-1). To generate those curves, a computer program was written to simulate the CO2 degassing of water along a channel of length L (m), using the degassing mass transfer equations for the plug flow model and for the completely mixed model (THIBODEAUX, 1996). The complete length of the channel was divided in segments of length Dx (m) and the water output from one segment was the water input for the next segment. The following equations describe the fraction of gas Fp and Fm that remains in the water channel at the end of each segment according to the plug flow model and the completely mixed model, respectively: 0
Fp ¼ expð1 KA2 A=UH2 OciÞ
and
Fm ¼
1þ
1 K0 A2
1 A=UH2 Oci
ðaÞ 0
A (m2) is the area in contact with the air, UH2Oci (l s-1) is the water discharge, and 1 KA2 is the mass transfer coefficient of CO2 between the water and the air. The area A is given by the width of the channel and the length of the segment. The width and depth of gallery channels range from around 25 to 50 cm. For the mass transfer coefficient we have used the equation presented in THIBODEAUX (1996) for the surface renewal theory. In this theory, parcels of fluid make in contact with air at the interface and transfer the chemical
168
R. Marrero et al.
Pure appl. geophys.,
to the air; these parcels are replaced by other parcels at the interface as they sink into the water due to the water movement. Water velocity (vx; in m s-1) and depth (h; in m) of the channel are important factors determining the mass transfer coefficient according to the next equation: DA2 vx 1=2 DA2 UH2 Oci 1=2 1 0 KA2 ¼ ¼ ; ðbÞ h wh where DA2 is the molecular diffusivity of CO2 in water (1.77 E-05 cm2 s-1 at 20C, CRANK, 1976), and w (m) is the width of the channel. At the end of each segment the new H2CO3 in the water qA2 is calculated using the equation (THIBODEAUX, 1996): F¼
qA2 qA2 ; qA21 qA2
ðcÞ
where F (kg d-1 m-1) is either Fp or Fm, qA21 is the concentration of H2CO3 at the inflow side of the segment and qA2 is the concentration at the outflow or the end of the segment. qA2 is the equilibrium concentration of H2CO3 at the very interface which is defined by Henry law: qA2 ¼
qA1 ; KCO2
ðdÞ
where qA1 is the concentration or pCO2 in the local air, and KCO2 is the Henry constant for CO2. The new DIC2 at the end of the modeled channel segment is then: DIC2 ¼ DIC1 ðqA21 qA2 Þ:
ðeÞ
The pH in the water determines the concentration of carbonate species present in the water (DREVER, 1997). As water degasses, the system re-equilibrates by changing the pH and concentration of the other carbonate species (assuming minerals do not precipitate). For that reason, at the end of each small channel segment along the gallery, the concentration of H2CO3 in the water is corrected to take into consideration this re-equilibration process. The pH is determined by the concentration of bicarbonate according to the dissociation equation for H2CO3 to form bicarbonate. The corrected concentration of H2CO3 at the end of the segment is given by the equation: 1=2 ðcat DIC2 Þ þ ðcat DIC2 Þ2 4 KK21 cat2 ; ðfÞ qA2 ¼ 2 where cat is the number of equivalents of cations compensating the bicarbonate, and K1 and K2 are the first and second dissociation constants for dissolved CO2. The average CO2 transferred per unit length to the atmosphere along the complete channel was evaluated computing segment by segment the transferred CO2 to the local atmosphere, summing it and dividing it by the total length of the gallery. The factors
Vol. 165, 2008
Carbon Dioxide Discharge
169
that define the magnitude of the F are the initial DIC (DIC1), the number of equivalents of cations (cat), the width and depth of the channel (w and h), the total length of the gallery, the water discharge (UH2Oci), and the concentration of CO2 in the atmosphere of the gallery (qA1). A few measurements of the pCO2 in the atmosphere of the galleries have been reported (ALBERT-BELTRAN et al., 1990). These values range from 0.01 atm to 0.11 atm. The average cations in the galleries is 0.02 eq l-1 and the maximum around 0.04 eq l-1, the length ranges from 1643 m to 5058 m, the average DIC is 0.03 mol l-1 with values as high as 0.04 mol l-1. These condition ranges have been used to model the degassing factor as a function of water discharge. The length of each modeled cell or segment was 10 m. The results for the two degassification models (plug flow model and completely mixed model) were not significantly different. The results for the range of conditions observed at Las Can˜adas aquifer are presented in Figures 3a and 3b.
REFERENCES ABLAY, G.J. and HU¨RLIMANN, M. (2000), Evolution of the north flank of Tenerife by recurrent giant landslides, J. Volcanol. Geotherm. Res. 103, 135–159. ABLAY, G.J. and MARTI, J. (2000), Stratigraphy, structure, and volcanic evolution of the Pico Teide-Pico Viejo formation, Tenerife, Canary Islands, J. Volcanol. Geotherm. Res. 103, 175–208. ALMENDROS, J., IBA´N˜EZ, J.M., CARMONA, E., and ZANDOMENEGHI, D. (2007), Array analyses of volcanic earthquakes and tremor recorded at Las Can˜adas caldera (Tenerife Island, Spain) during the 2004 seismic activation of Teide volcano, J. Volcanol. Geotherm. Res. 160, 285–299. ALBERT-BELTRAN, J.F., ARAN˜A, V., DIEZ, J.L., and VALENTIN, A. (1990), Physical-chemical conditions of the Teide volcanic system (Tenerife, Canary Islands), J. Volcanol. Geotherm. Res. 43, 321–332. ALLARD, P., CARBONELLE, J., DAJLEVIC, D., LE BRONEC, J., MOREL, P., ROBE, M.C., MAURENAS, J.M., FAIVREPIERRET, R., MARTINS, D., SABROUX, J.C., and ZETTWOOG, P. (1991), Eruptive and diffuse emissions of CO2 from Mount Etna, Nature 351, 387–391. ALLARD, P. (1992), Diffuse degassing of carbon dioxide through volcanic systems: Observed facts and implications, Rept. Geol. Surv. Japan 279, 7–11. ANCOCHEA, E., FUSTER, J.M., IBARROLA, E., CENDRERO, A., COELLO, J., HERNAN, F., CANTAGREL, J.M., and JAMONA, C. (1990), Volcanic evolution of the island of Tenerife (Canary Islands) in the Light of new K-Ar data, J. Volcanol. Geotherm. Res. 44, 231–249. ARAN˜A, V. (1971), Litologı´a y estructura del edificio Can˜adas, Tenerife (Islas Canarias), Estudios Geolo´gicos 27, 95–135. ARAN˜A, V., APARICIO, A., GARCIA-CACHO, L., and GARCIA-GARCIA, R., El complejo volca´nico del Teide-Pico Viejo. In Los Volcanes y La Caldera del Parque Nacional del Teide (Tenerife, Islas Canarias) (eds, Coello, J., and Aran˜a, V.) (ICONA, Madrid 1989) pp. 101–126. BAUBRON, J.C., MATHIEU, R., and MIELE, G. (1991), Measurement of gas flows from soil in volcanic areas: The accumulation method. Intl Conf. in Active Volcanoes and Risk Mitigation, Napoli, Italy. BOTTINGA, Y. and JAVOY, M. (1989), MORB degassing: evolution of CO2, Earth Planet. Sci. Lett. 95, 215–225. BRAVO, T. (1962), El circo de las Can˜adas y sus dependencias, Bol. R. Soc. Esp. Hist. Nat. 60, 93–108. BRAOJOS, J.J., Definicio´n de la recarga a trave´s del balance hı´drico en las Islas Canarias Occidentales. In La evaluacio´n de la recarga a los acuı´feros en la planificacio´n hidrolo´gica (AIH, Las Palmas de Gran Canaria 1997) pp. 267–277. CARON, F., MANNI, G., and WORKMAN, W. (1998), A large-scale laboratory experiment to determine the mass transfer of CO2 from a sandy soil to moving groundwater, J. Geochem. Explor. 64, 111–125.
170
R. Marrero et al.
Pure appl. geophys.,
CHIODINI G. (1993). Idrogeochimica della zona Sud e della zona della Can˜adas, isola di Tenerife. Internal report. CHIODINI, G., FRONDINI, F., and RACO, B. (1996) Diffuse emission of CO2 from the Fossa crater, Vulcano Island (Italy), Bull. Volcanol. 58, 41–50. CHIODINI, G., FRONDINI, F., KERRICK, D.M., ROGIE, J., PARELLO, F., PERUZZI, L., and ZANZARI, A.R. (1999), Quantification of deep CO2 fluxes from Central Italy. Examples of carbon balance for regional aquifers and of soil diffuse degassing, Chem. Geol. 159, 205–222. CHIODINI, G. and FRONDINI, F. (2001), Carbon dioxide degassing from the Albani Hills volcanic region, Central Italy, Chem. Geol. 177, 67–83. COELLO, J. (1973), Las series volca´nicas en subsuelos de Tenerife, Estudios Geolo´gicos 30, 491–512. CRANK, J., The Mathematics of Difusio´n (Oxford University Press, 1976). CUSTODIO, E., HOPPE, J., HOYOS-LIMo´N, A., JIME´NEZ, J., PLATA, A., and UDLUFT, P., Aportaciones al conocimiento hidrogeolo´gico de Tenerife utilizando iso´topos ambientales. In Hidrogeologı´a y Recursos Hidra´ulicos 11 (ed. AEH) (1987), pp. 263–280. CUSTODIO, E., Hidrogeochemistry of Tenerife Island. In Revista Espan˜ola de Hidrogeologı´a 3 (ed. AEH) (1988), pp. 1–19. CUSTODIO, E. and LLAMAS, M.R., Hidrologı´a Subterra´nea (Omega, Barcelona 2001). DREVER, J. I., The Geochemistry of Natural Waters. (Prentice-Hall, New Jersey 1997). EVANS, W.C., SOREY, M.L., COOK, A.C., MACK KENNEDY B., SHUSTER, D.L., COLVARD, E.M., WHITE, L.D., and HUEBNER, M.A. (2002), Tracing and quantifying magmatic carbon discharge in cold groundwaters: Lessons learned from Mammoth Mountain, USA, J. Volcanol. Geotherm. Res. 114, 291–312. FARRAR, C., SOREY, M., EVANS, W., HOWLE, J., KERR, B., KENNEDY, B., KING, C., and SOUTHON, J. (1995), Forest-killing diffuse CO2 emission at Mammoth Mountain as a sign of magmatic unrest, Nature 376, 675–678. FARRUJIA, I., BRAOJOS, J. J., and FERNA´NDEZ, J. (2001a), Estacio´n de adquisicio´n de datos del sondeo de Montan˜a Majua. Las Can˜adas del Teide, 7th Symposium de Hidrogeol. de la AEH, Murcia, Spain, pp. 673–685. FARRUJIA, I., FERNA´NDEZ, J., LO´PEZ, L.M., and GONZA´LEZ, S. (2001b), Ejecucio´n de dos sondeos profundos en Las Can˜adas del Teide, 7th Symposium de Hidrogeol. de la AEH, Murcia, Spain, pp. 661–673. FARRUJIA, I., VELASCO, J.L., FERNA´NDEZ, J., and MARTI´N, M.C. (2004), Evolucio´n del nivel frea´tico en la mitad oriental del acuı´fero de las Can˜adas del Teide, 8th Symposium de Hidrogeol. de la AEH, Zaragoza, Spain, pp. 131–142. FUSTER, J.M., ARAN˜A, V., BRANDLE, J.L., NAVARRO, J.M., ALONSO, U., and APARICIO, A., Geologı´a y Volcanologı´a de las Islas Canarias. Tenerife (Inst. Lucas Mallada, Madrid 1968). GOTTSMANN, J., WOOLLER, L., MARTI´, J., FERNA´NDEZ, J., CAMACHO, A.G., GONZA´LEZ, P.J., GARCI´A, A., and RYMER, H. (2006), New evidence for the reawakening of Teide volcano, Geophys. Res. Lett. 33, L20311. GAMBARDELLA, B., CARDELLINI, C., CHIODINI, G., FRONDINI, F., MARINI, L., OTTONELLO, M., and ZUCCOLINI, V. (2004), Fluxes of deep CO2 in the volcanic areas of central-southern Italy, J. Volcanol. Geotherm. Res. 136, 31–52. GERLACH, T.M. (1991), Present-day CO2 emissions from volcanoes, EOS (Trans. Am. Geophys. Union) 72, 249– 251. HERNA´NDEZ, P.A., CHIODINI, G., SALAZAR, J.M., and PE´REZ, N.M. (1997), Diffuse emission of CO2 flux from Can˜adas Caldera, Tenerife, Canary Islands, Spain, EOS (Trans. Am. Geophys. Union) 78, 778. HERNA´NDEZ, P.A., PE´REZ, N.M., SALAZAR, J.M., NAKAI, S., NOTSU, K., and WAKITA, H. (1998), Diffuse emission of carbon dioxide, methane, and helium-3 from Teide volcano, Tenerife, Canary Islands, Geophys. Res. Lett. 25 (17), 3311–3314. HERNA´NDEZ, P.A., PE´REZ, N., SALAZAR, J., SATO, M., NOTSU, K., and WAKITA, H. (2000), Soil gas CO2, CH4, and H2 distribution in and around Las Can˜adas Caldera, Tenerife, Canary Islands, Spain, J. Volcanol. Geotherm. Res. 103, 425–438. HERNA´NDEZ, P.A., NOTSU, K., SALAZAR, J.M., MORI, T., NATALE, G., OKADA, H., VIRGILI, G., SHIMOIKE, Y., SATO, M., and PE´REZ, N.M. (2001), Carbon dioxide degassing by advective flow from Usu Volcano, Japan, Science 292, 83–86. HERNA´NDEZ, P.A., NOTSU, K., TSURUMI, M., MORI, T., OHNO, M., SHIMOIKE, Y., SALAZAR, J.M., and PE´REZ, N.M. (2003), Carbon dioxide emissions from soils at Hakkoda, North Japan, J. Geophys. Res. 108, 6–1 to 6–10.
Vol. 165, 2008
Carbon Dioxide Discharge
171
HERNA´NDEZ P.A., NOTSU K., OKADA, H., MORI, T., SATO, M., BARAHONA, F., and PE´REZ, N. (2006), Diffuse emission of CO2 from Showa-Shinzan, Hokkaido, Japan: A sign of volcanic dome degassing, Pure Appl. Geophys. 163, 869–881. MACFARLANE, D.J. and RIDLEY, W.I. (1968), An interpretation of gravity data for Tenerife, Canary Islands, Earth Planet. Sci. Lett. 4, 481–486. MARRERO, R. (2004), Estudio Hidrogeoquı´mico del Acuı´fero de Las Can˜adas del Teide, Tenerife, DEA, Universidad de La Laguna. MARRERO, R., MELIA´N, G., PADRO´N, E., BARRANCOS, J., CALVO, D., NOLASCO, D., PADILLA, G., Lo´PEZ, D.L., HERNA´NDEZ, P.A., and PE´REZ, N.M., Physical-chemical hydrological changes related to the recent volcanic unrest at Tenerife, Canary Islands, J. Volcanol. Geotherm. Res., submitted. MART´ı, J., MITJAVILA, J., and ARAN˜A, V. (1994), Stratigraphy, structure and geochronology of the Las Can˜adas Caldera (Tenerife, Canary Islands), Geol. Magazine 131, 715–727. MART´ı, J., HURLIMANN, M., ABLAY, G., and GUDMUNDSSON, A. (1997), Vertical and lateral collapses on Tenerife (Canary Islands) and other volcanic ocean islands, Geology 25, 879–882. MART´ı, J. and GUDMUNSSON, A. (2000), The Las Can˜adas Caldera (Tenerife, Canary Islands): An overlapping collapse caldera generated by magma-chamber migration, J. Volcanol. Geotherm. Res. 103, 106–173. NAVARRO, J.M. and COELLO, J. (1989), Depressions originated by landslide processes in Tenerife, In ESF Meeting on Canarian Volcanism, Lanzarote, Spain, pp. 150–152. NAVARRO, J.M., Geologı´a e hidrogeologı´a del Parque Nacional del Teide (Subdireccio´n General de Espacios Naturales, Santa Cruz de Tenerife 1994). NOTSU, K., SUGIYAMA, K., HOSOE, M., UEMURA, A., SHIMOIKE, Y., TSUNOMORI, F., SUMINO, H., YAMAMOTO, J., MORI, T., and HERNA´NDEZ, P.A. (2005), Diffuse CO2 efflux from Iwojima volcano, Izu-Ogasawara arc, Japan, J. Volcanol. Geotherm. Res. 139, 147–161. OKADA, H., YASUDA, Y., YAGI, M., and KAI, K. (2000), Geology and fluid chemistry of the Fushime geothermal field, Kyushu, Japan, Geothermics 29, 279–311. PADRo´N, E., Lo´PEZ, D. L., MAGAN˜A, M.I., MARRERO, R., and PE´REZ, N.M. (2003), Diffuse degassing and relation to structural flow paths at Ahuachapan Geothermal Field, El Salvador, Geothermal Resources Council Transactions 27, 12–15. PARKHURST, D.L. (1995), User´s guide to PHREEQC - A computer program for speciation, reaction path, advective-transport, and inverse geochemical calculations, U.S. Geol. Surv. Water Resour. Invest. Rep. 95-4227, 1–143. PARKINSON, K.J. (1981), An improved method for measuring soil respiration in the field, J. Appl. Ecol. 18, 221– 228. PE´REZ, N.M., NAKAI, S., WAKITA, H., HERNA´NDEZ, P.A., and SALAZAR, J.M. (1996), Helium-3 emission in and around Teide volcano, Tenerife, Canary Islands, Spain, Geophys. Res. Lett. 23, 3531–3534. PE´REZ, N.M., SALAZAR, J.M.L., HERNA´NDEZ, P.A., SORIANO T., Lo´PEZ K., and NOTSU, K. (2004), Diffuse CO2 and 222 Rn degassing from San Salvador volcano, El Salvador, Central America, Bull. Geolog. Soc. Am. 375, 227– 236. PE´REZ, N.M., HERNA´NDEZ, P.A., PADRO´N, E., CARTAGENA, R., OLMOS, R., BARAHONA, F., MELIA´N, G., SALAZAR, P., and Lo´PEZ, D.L. (2006), Anomalous diffuse CO2 emission prior to the January 2002 short-term unrest at San Miguel Volcano, El Salvador, Central America, Pure Appl. Geophys. 163, 883–896. PLAN HIDROLO´GICO INSULAR DE TENERIFE (1992). Cabildo Insular de Tenerife. POUS, J., HEISE, W., SCHNEGG, P., MUN˜OZ, G., MARTI´, J., and SORIANO, C. (2002), Magnetotelluric study of the Las Can˜adas Caldera (Tenerife, Canary Islands): Structural and hydrogeological implications, Earth Planet. Sci. Lett. 204, 249–263. RIDLEY, W.I. (1971), The origin of some collapse structures in the Canary Islands, Geol. Magazine 108 (6), 477–484. SALAZAR, J.M., NAVARRO, J.M., PE´REZ, N.M., CHIODINI, G., and HERNA´NDEZ, P.A. (1997), Subsurface diffuse emission of carbon dioxide from Teide volcano by means of hydrological studies. Tenerife. Canary Islands, EOS (Trans. Am. Geophys. Union) 78, 778. SOLER, V., CASTRO-ALMAZA´N, J.A., VIN˜AS, R.T., EFF-DARWICH, A., SA´NCHEZ-MORAL, S., HILLAIRE-MARCEL, C., FARRUJIA, I., and COELLO, J. (2004), High CO2 levels in boreholes at El Teide volcano complex (Tenerife, Canary Islands): Implications for volcanic activity monitoring, Pure Appl. Geophys. 161, 1519– 1532.
172
R. Marrero et al.
Pure appl. geophys.,
STOIBER, R.E. and WILLIAMS, S.N. (1990), Monitoring active volcanoes and mitigating volcanic hazards; the case for including simple approaches, J. Volcanol. Geotherm. Res. 42, 129–149. STOLPER, E. and HOLLOWAY, J.R. (1988), Experimental determination of the solubility of carbon dioxide in molten basalt at low pressure, Earth Planet. Sci. Lett. 8, 397–408. SWAN, A.R.H. and SANDILANDS, M., Introduction to Geological Data Analysis, (Blackwell Science, 1995) 446 pp. THIBODEAUX, L.J., Environmental Chemodynamics, Movement of Chemicals in Air, Water, and Soil (John Wiley and Sons Inc., 1996). VALENTI´N, A., ALBERT-BELTRA´N, J.F., and DI´EZ, J.L. (1990), Geochemical and geothermal constraints on magma bodies associated with historic activity, Tenerife (Canary Islands), J. Volcanol. Geotherm. Res. 44, 251–264. WHITE, P.A. and HUNT, T.M. (2005), Simple modeling of the effects of exploitation on hot springs, Geyser Valley, Wairakei, New Zealand, Geothermics 34, 187–207. (Received February 9, 2006, revised June 15, 2007, accepted June 28, 2007) Published Online First: February 1, 2008
To access this journal online: www.birkhauser.ch/pageoph
Pure appl. geophys. 165 (2008) 173–180 0033–4553/08/010173–8 DOI 10.1007/s00024-007-0296-2
Birkha¨user Verlag, Basel, 2008
Pure and Applied Geophysics
Publication List HIROSHI WAKITA
WAKITA, H. and KIGOSHI, K. (1964) Activation analysis of Thorium in shell, J. Chem. Soc. Japan 85, 476–480 (in Japanese). WAKITA, H., NAGASAWA, H., UYEDA, S., and KUNO, H. (1967) Uranium and thorium contents in ultrabasic rocks, Earth Planet. Sci. Lett. 2, 377–381. NAGASAWA, H. and WAKITA, H. (1967) Neutron activation analysis of potassium in ultrabasic rocks, Geochem. J. 1, 149–154. WAKITA, H., NAGASAWA, H., UYEDA, S., and KUNO, H. (1967) Uranium, thorium and potassium contents of possible mantle materials, Geochem. J. 1, 183–198. NAGASAWA, H. and WAKITA, H. (1968) Partition of uranium and thorium between augite and host lavas, Geochim. Cosmochim. Acta 32, 917–921. ONUMA, N., HIGUCHI, H., WAKITA, H., and NAGASAWA, H. (1968) Trace element partition between two pyroxenes and the host lava, Earth Planet. Sci. Lett. 5, 47–51. NAGAWASA, N., WAKITA, H., HIGUCHI, H., and ONUMA, N. (1969) Rare earths in peridotite nodules: An explanation of the genetic relationship between basalt and peridotite nodules, Earth Planet. Sci. Lett. 5, 377– 381. SCHMITT, R.A., WAKITA, H., and REY, P. (1970) Abundances of 30 elements in lunar rocks, soil and core samples, Science 167, 512–515. SCHMITT, R.A., LINN, T.A. Jr., and WAKITA, H. (1970) The determination of fourteen common elements in rocks via sequential instrumental activation analysis, Radiochim. Acta 13, 200–212. REY, P., WAKITA, H., and SCHMITT, R.A. (1970) Radiochemical neutron activation analysis of indium, cadmium, yttrium and the 14 rare earth elements in rocks, Anal. Chim. Acta 51, 163–178. GOLES, G.G., RANDLE, K., OSAWA, M., SCHMITT, R.A., WAKITA, H., EHMANN, W.D., MORGAN, J.W. (1970) Elemental abundances by instrumental activation analyses in chips from 27 lunar rocks, Proc. of the Apollo 11 Lunar Science Conf. (LEVINSON, A. A. ed.), vol. 2, 1165–1176 (Pergamon Press, New York). WAKITA, H., SCHMITT, R.A., and REY, P. (1970) Elemental abundances of major, minor and trace elements in Apollo 11 lunar rocks, soil and core samples, Proc. of the Apollo 11 Lunar Science Conf. (A. A. LEVINSON, ed.), vol. 2, 1685–1717 (Pergamon Press, New York). WAKITA, H. and SCHMITT, R.A. (1970) Rare earth and other elemental abundances in the Allende meteorite, Nature 277, 478–479. WAKITA, H. and SCHMITT, R.A. (1970) Elemental abundances in seven fragments from lunar rock 12013, Earth Planet. Sci. Lett. 9, 169–176. WAKITA, H. and SCHMITT, R.A. (1970) Lunar anorthosites: Rare-earth and other elemental abundances, Science 170, 969–974. WAKITA, H. and SCHMITT, R.A. (1971) Bulk elemental composition of Apollo 12 samples: Five igneous and one breccia rocks and four soils, Proc. of the Second Lunar Science Conf. (A. A. LEVINSON, ed.), vol. 2, 1231– 1236 (MIT Press, Massachusetts). WAKITA, H., REY, P., and SCHMITT, R.A. (1971) Abundances of the 14 rare-earth elements and 12 other trace elements in Apollo 12 samples: Five igneous and one breccia rocks and four soils, Proc. of the Second Lunar Science Conf. (A. A. LEVINSON, ed.), vol. 2, 1319–1329 (MIT Press, Massachusetts).
174
Publication List
Pure appl. geophys.,
SHOWALTER, D.L., WAKITA, H., SMITH, R.H., SCHMITT, R.A., GILLUM, D.E., and EHMANN, W.D. (1972) Chemical composition of sawdust from lunar rock 12013 and comparison of a Java tektite with the rock, Science 175, 170–172. GILLUM, D.E., EHMANN, W.D., WAKITA, H., and SCHMITT, R.A. (1972) Bulk and rare earth abundances in the Luna-16 soil levels A and D, Earth Planet. Sci. Lett. 13, 444–449. LAUL, J.C., WAKITA, H., SHOWALTER, D.L., BOYNTON, W.V., and SCHMITT, R.A. (1972) Bulk, rare earth and other trace elements in Apollo 14 and 15 and Luna 16 samples, Proc. of the Third Lunar Science Conf. (D. HEYMANN, ed.), vol. 2, 1181–1200 (MIT Press, Massachusetts). SHOWALTER, D.L., WAKITA, H., and SCHMITT, R.A. (1972) Rare earth and other abundances in the Murchison carbonaceous meteorite, Meteoritics 7, 295–301. NOGUCHI, M. and WAKITA, H. (1973) Measurements of low-level radon for the determination of radium in marine carbonates, Geochem. J. 7, 81–88. OSBORN, T.W., WARREN, R.G., SMITH, R.H., WAKITA, H., ZELLMER, D.L., and SCHMITT, R.A. (1974) Elemental composition of individual chondrules from carbonaceous chondrites, including Allende, Geochim. Cosmochim. Acta 38, 1359–1378. WAKITA, H., LAUL, J.C., and SCHMITT, R.A. (1975) Some thoughts on the origin of lunar ANT-KREEP and mare basalts, Geochem. J. 9, 25–41. WAKITA, H. (1975) Water wells as possible indicators of tectonic strain, Science 189, 553–555. NOGUCHI, M., WAKITA, H., and OYO BUTURI, (1975) Low-level radon measurement with a-b coincidence methods, Japan Soc. Appl. Phys. 44, 979–983 (in Japanese with English abstract). WAKITA, H. and HAMAGUCHI, H. (1976) Neutron flux gradient of the VT-2 hole of the JRR-2 reactor, Radioisotopes 25, 499–503. WAKITA, H., NOTSU, K., NAKAMURA, Y., MORIOKA, M., and NOGUCHI, M. (1976) Ground upheaval observed in the lower Tamagawa area and variation of radon concentration in groundwater, Zisin (J. Seismol. Soc. Japan) 29, 71–81 (in Japanese with English abstract). NOGUCHI, M. and WAKITA, H. (1977) A method for continuous measurement of radon in groundwater for earthquake prediction, J. Geophys. Res. 82, 1353–1357. KANNO, H., WAKITA, H., and HAMAGUCHI, H. (1977) Distribution of trace rare earth element ions in the crystallization process of alums, Bull. Chem. Soc. Japan 51, 1557–1558 (1977). WAKITA, H., FUJII, N., MATSUO, S., NOTSU, K., NAGAO, K., and TAKAOKA, N. (1978) Helium spots: Caused by a diapiric magma from the upper mantle, Science 200, 430–432. WAKITA, H. and SCHMITT, R.A. (1978) Cadmium B-M, Handbook of Geochemistry (K. H. WEDEPOHL, ex. ed.), vol. II/5 (Springer-Verlag, Berlin). SCHMITT, R.A. and WAKITA, H. (1978) Indium C-III, Handbook of Geochemistry (K. H. WEDEPOHL, ex. ed.), vol. II/5, (Springer-Verlag, Berlin). WAKITA, H. (1978) Earthquake prediction and geochemical studies in China, Chinese Geophys. 1, 443–457. WAKITA, H. (1979) Earthquake prediction by geochemical techniques, Rec. Progr. Nat. Sci. Japan 4, 69–75. ONUMA, N., NOTSU, K., NISHIDA, N., WAKITA, H., TAKEDA, H., NAGASAWA, H., OUYANG ZIYUAN, and WANG DAODE, (1979) Major and trace element compositions of two Chinese meteorites, Nantan iron and Anlung chondrite, Geochimica 1, 52–60. MINAI, Y., TAKEDA, M., WAKITA, H., and TOMINAGA, T. (1978) A Mossbauer study on alteration of a deep-sea basalt, Radiochem. Radioanal. Lett. 36, 199–206. DANON, J., SCORZELLI, R.B., AZEVEDO, I.S., ALBERTSEN, J.F., KNUDSEN, J.M., ROY-POULSEN, N.O., MINAI, Y., WAKITA, H., and TOMINAGA, T. (1979) Iron-nickel alloy superstructures in the mineral Josephinite, Radiochem. Radioanal. Lett. 38, 339–342. WAKITA, H., NAKAMURA, Y., NOTSU, K., NOGUCHI, M., and ASADA, T. (1980) Radon anomaly: A possible precursor of the 1978 Izu-Oshima-kinkai earthquake, Science 207, 882–883. NAGAO, K., TAKAOKA, N., WAKITA, H., MATSUO, S., and FUJII, N. (1980) Isotopic compositions of rare gases in the Matsushiro earthquake fault region, Geochem. J. 14, 6–69. WAKITA, H., NAKAMURA, Y., KITA, I., FUJII, N., and NOTSU, K. (1980) Hydrogen release: New indicator of fault activity, Science 210, 188–190. KITA, I., MATSUO, S., WAKITA, H., and NAKAMURA, Y. (1980) D/H ratios of H2 in soil gases as an indicator of fault movements, Geochem. J. 14, 317–320.
Vol. 165, 2008
Publication List
175
NOTSU, K., ABIKO, T., and WAKITA, H. (1980) Coseismic temperature changes of well water related to volcanic activities of Usu Volcano, J. Phys. Earth 28, 617–624. WAKITA, H. (1981) Precursory changes in groundwater prior to the 1978 Izu-Oshima-kinkai earthquake, Earthquake Prediction, An International Review (D. E. SIMPSON and P. G. RICHARD, eds.), Maurice Ewing Series, vol. 4, pp. 527–532 (Amer. Geophys. Union, 198 Washington D.C.). NAKAMURA, Y. and WAKITA, H. (1982) Terrestrial heat flow around the aseismic front of the Japanese Island Arc, Tectonophys. 81, T25–T36. TAKEHANA, Y., KOBAYASHI, Y., WAKITA, H., and NAKAMURA, Y. (1982) Emission of hydrogen along the Neguro fault, Median Tectonic Line, Zisin (J. Seismol. Soc. Japan) 35, 103–115 (in Japanese with English abstract). KITA, I., MATSUO, S., and WAKITA, H. (1982) H2 generation by reaction between H2O and crushed rock: An experimental study on H2 degassing from the active fault zone, J. Geophys. Res. 87, 10789–10795. SANO, Y., TOMINAGA, T., NAKAMURA, Y., and WAKITA, H. (1982) 3He/4He ratios of methane-rich natural gases in Japan, Geochem. J. 16, 237–245. SANO, Y., TOMINAGA, T., and WAKITA, H. (1982) Elemental and isotopic abundances of rare gases in natural gases obtained by a quadrupole mass spectrometer, Geochem. J. 16, 279–286. WAKITA, H. (1982) Changes in groundwater level and chemical composition, Earthquake Prediction Techniques (T. ASADA, ed.), pp. 175–216 (University of Tokyo Press, Tokyo). WAKITA, H. and SANO, Y. (1983) 3He/4He ratios in CH4-rich natural gases suggest magmatic origin, Nature 305, 792–794. NOTSU, K., ABIKO, T., and WAKITA, H. (1983) Radon concentration changes in groundwater related to the volcanic activities of Usu Volcano, Bull. Volcanol. Soc. Japan 28, 305–308. NAKAMURA, Y., WAKITA, H., KANAZAWA, T., NOGUCHI, M., and FUJII, N. (1983) An approach to surveillance of active faults in sea area via c-ray measurements, Zisin, (J. Seismol. Soc. Japan) 36, 469–472 (in Japanese with English abstract). SANO, Y., NAKAMURA, Y., WAKITA, H., URABE, A., and TOMINAGA, T. (1984) 3He emission related to volcanic activity, Science 224, 150–151. WAKITA, H. (1984) Groundwater observations for earthquake prediction in Japan, Proc. of the International Symposium on Continental Seismicity and Earthquake Prediction, pp. 494–500 (Seismol. Press, Beijing). EBIHARA, M., NAKAMURA, Y., WAKITA, H., KURASAWA, H., and KONDA, T. (1984) Trace element composition of Tertiary volcanic rocks of northeast Japan, Geochem. J. 18, 287–295. SANO, Y., NAKAMURA, Y., and WAKITA, H. (1985) Areal distribution of 3He/4He ratios in the Tohoku district northeastern Japan, Chem. Geol. 42, 1–8. URABE, A., TOMINAGA, T., NAKAMURA, Y., and WAKITA, H. (1985) Chemical compositions of natural gases in Japan, Geochem. J. 19, 11–25. NAKAMURA, Y. and WAKITA, H. (1985) Precise temperature measurement of groundwater for earthquake prediction study, Pure Appl. Geophys. 122, 164–174. SANO, Y., URABE, A., WAKITA, H., CHIBA, H., and SAKAI, H. (1985) Chemical and isotopic compositions of gases in geothermal fluids in Iceland, Geochem. J. 19, 135–148. WAKITA, H., NAKAMURA, Y., and SANO, Y. (1985) Groundwater radon variations reflecting changes in regional stress fields, Proc. of the U.S.-Japan Seminar on Practical Approaches to Earthquake Prediction and Warrning, Earthq. Predict. Res. 3, 545–557. SANO, Y. and WAKITA, H. (1985) Geographical distribution of the 3He/4He ratios in Japan: Implications for arc tectonics and incipient magmatism, J. Geophys. Res. 90, 8729–8741. SANO, Y., TOYODA, K., and WAKITA, H. (1985) 3He/4He ratios of marine ferromanganese nodules, Nature 317, 518–521. SANO, Y. and WAKITA, H. (1985) 3He/4He ratios of gases dissolved in pore water Leg 87, sites 583 and 584, Init. Repts. DSDP, 87, pp. 861–864, Washington (U.S. Govt. Printing Office). WAKITA, H., SANO, Y., FUJII, N., and TAKEUCHI, A. (1985) 3He/4He ratios of gases in deep sea sediments, Legs 89 and 90, Init. Repts. DSDP, 90, pp. 1261–1263, Washington (U.S. Govt. Printing Office). SANO, Y., NAKAMURA, Y., WAKITA, H., NOTSU, K., and KOBAYASHI, Y. (1986) 3He/4He anomalies associated with earthquake: Possibly induced by a diapiric magma, J. Geophys. Res. 91, 12291–2295. SANO, Y., WAKITA, H., and HUANG, C-W. (1986) Helium flux in a continental land area estimated from 3He/4He ratio in northern Taiwan, Nature 323, 55–57.
176
Publication List
Pure appl. geophys.,
SANO, Y., NAKAMURA, Y., WAKITA, H., and ISHII, T. (1986) Light noble gases in basalt glasses from Mariana Trough, Geochim. Cosmochim. Acta 50, 2429–2432. WAKITA, H., NAKAMURA, Y., and SANO, Y. (1986) Background fluctuation in groundwater radon observation, J. Phys. Earth 34, S81–S89. SHOWALTER, D.L., WAKITA, H., SMITH, R.H., and SCHMITT, R.A. (1987) Abundances of 14 rare earth elements and 12 other major, minor and trace elements in the Allende Meteorite Reference Sample by neutron activation analysis, The Allende Meteorite Reference Sample (E. JAROSEWICH, R. S. CLARKE, Jr., and J. N. BARROWS, eds.) Smithonian Contributions to the Earth Sciences 27, pp. 40–42 (Smithonian Institutions Press, Washington, D.C.). SANO, Y., WAKITA, H., and GIGGENBACH, W.F. (1987) Island arc tectonics of New Zealand manifested in helium isotope ratios, Geochim. Cosmochim. Acta 51, 1855–1860. WAKITA, H., SANO, Y., and MIZOUE, M. (1987) High 3He emanation and seismic swarms observed in a non volcanic, forearc region, J. Geophys. Res. 92, 12539–12546. SANO, Y., WAKITA, H., OHSUMI, T., and KUSAKABE, M. (1987) Helium isotope evidence for magmatic gases in lake Nyos, Cameroon, Geophys. Res. Lett. 14, 1035–1038. WILLIAMS, S.N., SANO, Y., and WAKITA, H. (1987) Helium-3 emission from Nevado Del volcano, Colombia, Geophys. Res. Lett. 14, 1039–041. WAKITA, H., NAKAMURA, Y., and SANO, Y. (1988) Short-term and intermediate-term geochemical precursors, Pure Appl. Geophys. 126, 267–278. SANO, Y. and WAKITA, H. (1988) Precise measurement of helium isotopes in terrestrial gases, Bull. Chem. Soc. Jpn. 61, 1153–1157. NOTSU, K., WAKITA, H., and NAKAMURA, Y. (1988) Strontium isotope composition of oil-field and gas-field waters in Japan, Appl. Geochem. 3, 173–176. SANO, Y., NAKAMURA, Y., NOTSU, K., and WAKITA, H. (1988) Influence of volcanic eruptions on helium isotope ratios in hydro-thermal systems, Geochim. Cosmochim. Acta 52, 1305–1308. NOTSU, K., NAKAMURA, Y., WAKITA, H., ARAKAWA, Y., and KOBAYASHI, Y. (1988) Geochemistry of lavas and ejecta of the 1986 eruption of Izu-Oshima volcano, Bull. Volcanol. Soc. Japan 33, S265–S270 (in Japanese with English abstract). WAKITA, H., NOTSU, K., NAKAMURA, Y., and SANO, Y. (1988) Temporal variation in geochemical parameters of gas from steam well and hot spring water associated with the 1986 eruption of Izu-Oshima volcano, Bull. Volcanol. Soc. Japan 33, S285–S289 (in Japanese with Englishi abstract). SANO, Y., WAKITA, H., and XU SHENG, (1988) Atmospheric helium isotope ratio, Geochem. J. 22, 177–181. SANO, Y. and WAKITA, H. (1988) Helium isotope ratio and heat discharge rate in the Hokkaido Island, northeast Japan, Geochem. J. 22, 293–303. SANO, Y., WAKITA, H., WAKINO, K., MURATA, M., YAMAMOTO, H., and MATSUDA, H. (1989) Helium isotope ratio measurement by a single focusing mass spectrometer with large incident and exit angle, Int. J. Mass Spectrum. Ion Processes 91, 69–77. WAKITA, H., IGARASHI, G., NAKAMURA, Y., SANO, Y., and NOTSU, K. (1989) Coseismic radon changes in groundwater, Geophys. Res. Lett. 16, 417–420. SANO, Y., WAKITA, H., ITALIANO, F., and NUCCIO, M.P. (1989) Helium isotopes and tectonics in southern Italy, Geophys. Res. Lett. 16, 511–514. SANO, Y., WAKITA, H., ISHIBASHI, J., GAMO, T., and SAKAI, H. (1989) Helium isotope ratios in Japan Sea water, Chikyukagaku (Geochemistry) 23, 61–67 (in Japanese with English abstract). SANO, Y., WAKITA, H., MAKIDE, Y., and TOMINAGA, T. (1989) A ten-year decrease in the atmospheric helium isotope ratio possibly caused by human activity, Geophys. Res. Lett. 16, 1371–1374. MIZOUE, M., SANO, Y., and WAKITA, H. (1989) Wakayama seismic swarms and high 3He emanation observed in non-volcanic forearc region, Proc. of China-Japan Symp. on Earthquake Prediction (D. GUOYU, ed.), pp. 101– 116 (Seismological Press, Beijing). WAKITA, H., SANO, Y., URABE, A., and NAKAMURA, Y. (1990) Origin of methane-rich natural gas in Japan: Formation of gas fields due to large-scale submarine volcanism, Appl. Geochem. 5, 263–278. IGARASHI, G. and WAKITA, H. (1990) Groundwater radon anomalies associated with earthquakes, Tectonophysics 180, 237–254.
Vol. 165, 2008
Publication List
177
SAKAI, H., GAMO, T., KIM, E-S., TSUTSUMI, M., TANAKA, T., ISHIBASHI, J., WAKITA, H., YAMANO, M., and OOMORI, T. (1990) Venting of carbon dioxide-rich fluid and hydrate formation in mid-Okinawa trough backarc basin, Science 248, 1093–1096. SANO, Y., WAKITA, H., and WILLIAMS, S.N. (1990) Helium-isotope systematics at Nevado del Ruiz volcano, Columbia: Implications for the volcanic hydrothermal system, J. Volcanol. Geotherm. Res. 42, 41–52. TEDESCO, D., ALLARD, P., SANO, Y., WAKITA, H., and PECE, R. (1990) Helium-3 in subaerical and submarine fumaroles of Campi Flegrei caldera, Italy, Geochim. Cosmochim. Acta 54, 1105–1116. IGARASHI, G., WAKITA, H., and NOTSU, K. (1990) Groundwater observations at KSM site in Northeast Japan, a most sensitive site to earthquake occurrence, Tohoku Geophys. J. (Sci. Rept. Tohoku Univ., Ser. 5) 33, 165– 175. SAKAI, H., GAMO, T., KIM, E.-S., SHITASHIMA, K., YANAGISAWA, F., TSUTSUMI, M., ISHIBASHI, J., SANO, Y., WAKITA, H., TANAKA, T., MATSUMOTO, T., NAGANUMA, T., and MITSUZAWA, K. (1990) Unique chemistry of the hydrothermal solution in the mid-Okinawa trough backarc basin, Geophys. Res. Lett. 17, 2133–2136. MINAI, Y., ISHII, T., NAKAMURA, Y., WAKITA, H., and TOMINAGA, T. (1990) Neutron activation analysis of altered oceanic tholeiite: Variation of lanthanide concentrations with degree of alteration, J. Radioanal. Nucl. Chem. Lett. 146, 375–384. NOTSU, K., WAKITA, H., and IGARASHI, G. (1991) Precursory changes in fumarolic gas temperature associated with a recent submarine eruption near Izu-Oshima volcano, Japan, Geophys. Res. Lett. 18, 191–193. SANO, Y., WAKITA, H., MAKIDE, Y., and TOMINAGA, T. (1991) Reply to Lupton and Graham, Geophys. Res. Lett. 18, 486–488. WAKITA, H., IGARASHI, G., and NOTSU, K. (1991) An anomalous radon decrease in groundwater prior to an M6.0 earthquake: A possible precursor? Geophys. Res. Lett. 18, 629–632. IGARASHI, G. and WAKITA, H. (1991) Tidal responses and earthquake-related changes in the water level of deep wells, J. Geophys. Res. 96, 4269–4278. NOTSU, K., WAKITA, H., IGARASHI, G., and SATO, T. (1991) Hydrological and geochemical changes related to the 1989 seismic and volcanic activities off the Izu Peninsula, J. Phys. Earth 39, 245–254. YABUKI, T., KANAZAWA, T., and WAKITA, H. (1991) Anomalous measurements in Oshima volcano associated with the off Ito submarine eruption revealed from GPS measurements, J. Phys. Earth 39, 155–164. IGARASHI, G., WAKITAI, H., and SATO, T. (1992) Precursory and coseismic anomalies in well water levels observed for the February 2, 1992 Tokyo Bay earthquake, Geophys. Res. Lett. 19, 1583–1586. LESNIAK, P.M., SAKAI, H., ISHIBASHI, J., and WAKITA, H., (1992) Mantle helium signal in the compressed feature, Outer Carpathians, Poland, Proc. of the 7th International Symp. on Water-Rock Interaction WRI-7 (Y. K. KHARAKA and A. S. MALST, eds.), pp. 959–962, (Rotterdam). SATO, T., WAKITA, H., NOTSU, K., and IGARASHI, G. (1992) Anomalous hot spring water changes: Possible precursors of the 1989 volcanic eruption off the east coast of the Izu Peninsula, Geochem. J. 26, 73–83. SANO, Y., SAKAMOTO, M., ISHIBASHI, J., WAKITA, H., and MATSUMOTO, R. (1992) Helium isotope ratios of pore gases in deep-sea sediments, Leg 128, Proc. ODP, Sci. Results, 127/128, Pt.1: College Station, TX (Ocean Drilling Program), 747–751. SAKAMOTO, M., SANO, Y., and WAKITA, H. (1992) 3He/4He ratio distribution in and around the Hakone volcano, Geochem. J. 26, 189–195. SANO, Y., MASUDA, M., TAKAHATA, N., WAKITA, H., ZENG, Y., NOJIRI, Y., MUKAI, H., and BANDO, H. (1992) Carbon isotope measurement of natural and anthropogenic methane, Chikyukagaku (Geochemistry) 26, 105– 114 (in Japanese with English abstract). SANO, Y., URABE, A., WAKITA, H., and WUSHIKI, H. (1993) Origin of hydrogen-nitrogen gas seeps, Oman, Appl. Geochem. 8, 1–8. GIGGENBACH, W.F., SANO, Y., and WAKITA, H. (1993) Isotopic composition of helium, and CO2 and CH4 contents in gases produced along the New Zealand part of convergent plate boundary, Geochim. Cosmochim. Acta 57, 3427–3455. MORI, T., NOTSU, K., TOHJIMA, Y., and WAKITA, H. (1993) Remote detection of HCl and SO2 in volcanic gas from Unzen volcano, Japan, Geophys. Res. Lett. 20, 1355–1358. IGARASHI, G., TOHJIMA, Y., and WAKITA, H. (1993) Time-variable response characteristics of groundwater radon to earthquakes, Geophys. Res. Lett. 20, 1807–1810. TOHJIMA, Y., WAKITA, H. (1993) Estimation of methane discharge from a plume: A case of landfill, Geophys. Res. Lett. 20, 2067–2070.
178
Publication List
Pure appl. geophys.,
NAKAI, S., Xu, S., WAKITA, H., FUJII, N., NAGAO, K., ORIHASHI, Y., and WANG, X. (1993) K-Ar ages of young volcanic rocks from Tengchong area, west Yunnan, China, J. CHEN, and Z. Liao, Bull. Volcanol. Soc. Japan 38, 167–171. NOTSU, K., MORI, T., IGARASHI, G., TOHJIMA, Y., and WAKITA, H. (1993) Infrared spectral radiometer: a new tool for remote measurement of SO2 of volcanic gas, Geochem. J. 27, 361–366. ONO, A., SANO, Y., WAKITA, H., and GIGGENBACH, W.F. (1993) Carbon isotopes of methane and carbon dioxide in hydrothermal gases of Japan, Geochem. J. 27, 287–295. TOHJIMA, Y. and WAKITA, H. (1994) Development of a continuous measurement system and areal distribution of methane in some source areas, Appl. Geochem. 9, 141–146. TOHJIMA, Y., IGARASHI, G., and WAKITA, H. (1994) Variations in groundwater flow rate at the KSM well, a most sensitive well to earthquake occurrences, J. Phys. Earth 42, 221–235. TSUNOGAI, U., ISHIBASHI, J., WAKITA, H., GAMO, T., WATANABE, K., KAJIMURA, T., KANAYAMA, S., and SAKAI, H. (1994) Peculiar features of Suiyo seamount hydrothermal fluids, Izu-Bonin Arc: Differences from subaerial volcanism, Earth Planet. Sci. Lett. 126, 289–301. ISHIBASHI, J., WAKITA, H., NOJIRI, Y., GRIMAUD, D., JEAN-BAPTISTE, P., GAMO, T., AUZRNDE, J.-M., and URABE, T. (1994) Helium and carbon geochemistry of hydrothermal fluids from the North Fiji Basin spreading ridge (Southwest Pacific), Earth Planet. Sci. Lett. 128, 183–197. PE´REZ, N.M., WAKITA, H., NAKAI, S., SANO, Y., and WILLIAMS, S.N. (1994) 3He/4He isotope ratios in volcanichydrothermal discharges from the Canary Islands, Spain: Implications on the origin of the volcanic activity, Mineral. Magazine 58A, 709–710. XU, S., NAKAI, S., WAKITA, H., WANG, X., and CHEN, J. (1994) Helium isotopic compositions in Quaternary volcanic geothermal area near Indo-Eurasian collisional margin at Tengchong, China, In Noble Gas Geochemistry and Cosmochemistry (J. MATSUDA, ed.), pp. 305–313 (Terra Sci. Publ. Co., Tokyo). SANO, Y., GAMO, T., NOTSU, K., and WAKITA, H. (1995) Secular variations of carbon and helium isotopes at IzuOshima volcano, Japan, J. Volcanol. Geotherm. Res. 64, 83–94. ISHIBASHI, J., SANO, Y., WAKITA, H., GAMO, T., TSUTSUMI, M., and SAKAI, H. (1995) Helium and carbon geochemistry of hydrothermal fluids from the Mid-Okinawa Trough Back Arc Basin, southwest of Japan, Chem. Geol. 123, 1–15. MORI, T., NOTSU, K., TOHJIMA, Y., WAKITA, H., NUCCIO, P.M., and ITALIANO, F. (1995) Remote detection of fumarolic gas chemistry at Vulcano, Italy, using an FT-IR spectral radiometer, Earth Planet. Sci. Lett. 134, 219–224. TSUNOGAI, U. and WAKITA, H. (1995) Precursory chemical changes in ground water: Kobe earthquake, Japan, Science 269, 61–63. IGARASHI, G. and WAKITA, H. (1995) Geochemical and hydrological observaitons for earthquake prediction in Japan, J. Phys. Earth 43, 585–598. HIGUCHI, T., IGARASHI, G., TOHJIMA, Y., and WAKITA, H. (1995) Time series analysis of groundwater radon using stochastic differential equations, J. Phys. Earth 43, 117–130. XU, S., NAKAI, S., WAKITA, H., XU, Y., and WANG, X. (1995) Helium isotope compositions in sedimentary basins in China, Appl. Geochem. 10, 463–475. XU, S., NAKAI, S., WAKITA, H., and WANG, X. (1995) Mantle-derived noble gases in natural gases in Songliao basin, China, Geochim. Cosmochim. Acta 59, 4675–4683. TSUNOGAI, U., ISHIBASHI, J., WAKITA, H., GAMO, T., MASUZAWA, T., NAKATSUKA, T., NOJIRI, Y., and NAKAMURA, T. (1996) Fresh water seepage and pore water recycling on the seafloor: Sagami Trough subduction zone, Japan, Earth Planet. Sci. Lett. 138, 157–168. WAKITA, H. (1996) Geochemical challenge to earthquake prediction, Proc. Natl. Acad. Sci. USA, 93, 3781– 3786. SILVER, P. and WAKITA, H. (1996) A search for earthquake precursors, Science 273, 77–78. JOHANSEN, A., SORNETTE, D., WAKITA, H., TSUNOGAI, U., NEWMAN, W.I., and SALEUR, H. (1996) Discrete scaling in earthquake precursory phenomena: Evidence in the Kobe earthquake, Japan, J. Phys. I France 6, 1391–1402. PE´REZ, N.M., NAKAI, S., WAKITA, H., HERNA´NDEZ, P.A., and SALAZAR, J.M. (1996) Helium-3 emission in and around Teide volcano, Tenerife, Canary Islands, Spain, Geophys. Res. Lett. 24, 3531–3534. XU, S., NAKAI, S., WAKITA, H., WANG, X., and FENG, X. (1996) Chemical and isotopic evidences for shortdistance mantle-crustal mixing processes, Geothermics, 179–192.
Vol. 165, 2008
Publication List
179
PE´REZ, N. M., NAKAI, S., WAKITA, H., ALBERT-BELTRA´N, J. F., and REDONDO, R., (1996), Preliminary results on 3 He/4He isotopic ratios in terrestrial fluids from Iberian peninsula: seismotectonic and neotectonic implications. Geogaceta 20 (4), 830–833. PE´REZ, N. M., WAKITA, H., LOLOK, D., PATIA, H., TALAI, B., and MCKEE, C. O. (1996), Anomalous soil gas CO2 concentrations and relation to seismic activity at Rabaul caldera, Papua New Guinea. Geogaceta 20 (4), 1000–1003. OHNO, M. and WAKITA, H. (1996) Coseismic radon changes of the 1995 Hygo-ken Nanbu earthquake, J. Phys. Earth 44, 391–395. TSUNOGAI, U. and WAKITA, H. (1996) Anomalous changes in ground water chemistry: Possible precursors of the 1995 Hyogo-ken Nanbu earthquake, Japan, J. Phys. Earth 44, 381–390. XU, S., NAKAI, S., WAKITA, H., XU, Y.-C., and WANG, X. (1997) Carbon isotopes of hydrocarbons and carbon dioxide in natural gases in China, J. Southeast Asian Earth Sci. 15, 89–101. OHNO, M., WAKITA, H., and KANJO, K. (1997) A water well sensitive to seismic waves, Geophys. Res. Lett. 24, 691–694. ISHIBASHI, J., WAKITA, H., OKAMURA, K., NAKAYAMA, E., FEELY, R.A., LEBON, G.T., BAKER, E.T., and MARUMO, K. (1997) Hydrothermal methane and manganese variation in the plume over the superfast-spreading southern East Pacific Rise, Geochim. Cosmochim. Acta 61, 485–500. NAKAI, S., WAKITA, H., NUCCIO, M.O., and ITALIANO, F. (1997) MORB-type neon in an enriched mantle beneath Etna, Sicily, Earth Planet. Sci. Lett. 153, 57–66. SHIMOIKE, Y., NOTSU, K., and WAKITA, H. (1997) Development of a continuous monitoring system of volcanic gas chemistry and its application to the observation at Izu-Oshima volcano, Chikyukagaku (Geochemistry) 31, 111–118 (in Japanese with English abstract). LESNIAK, P.M., SAKAI, H., ISHIBASHI, J., and WAKITA, H. (1997) Mantle helium signal in the West Carpathians, Poland, Geochem. J. 31, 383–394. TOHJIMA, Y., WAKITA, H., MAKSYUTOV, S., MACHIDA, T., INOUE, G., VINNICHENKO, N., and KHATTATOV, V. (1997) Distribution of tropospheric methane over Siberia in July 1993, J. Geophys. Res. 102, 25371– 25382. HERNA´NDEZ, P.A., PE´REZ, N.M., SALAZAR, J.M., NAKAI, S., NOTSU, K., and WAKITA, H. (1998) Diffuse emission of carbon dioxide, methane, and helium-3 from Teide volcano, Tenerife, Canary Islands, Geophys. Res. Lett. 25, 3311–3314. XE, S., NAGAO, K., UTO, K., WAKITA, H., NAKAI, S., and LIU, C. (1998) He, Sr and Nd isotopes of mantle-derived xenoliths in volcanic rocks of NE China, J. Asian Earth Sci. 16, 547–556. TSUNOGAI, U., ISHIBASHI, J., WAKITA, H., and GAMO, T. (1998) Methane-rich plumes in the Suruga Trough (Japan) and their carbon isotopic characterization, Earth Planet. Sci. Lett. 160, 97–105. WAKITA, H. (1998) Radon observation for earthquake prediction, in Radon and Throne in the Human Environment, (A. KATASE and M. SHIMO, eds.), Proc. of the 7th Tohwa University International Symp., pp. 124–130 (Singapore, World Scientific Publ. Co.). KING, C.-Y., AZUMA, S., IGARASHI, G., OHNO, M., SAITO, H., WAKITA, H. (1999) Earthquake-related water-level changes at 16 closely clustered wells, in Tono, Central Japan, J. Geophys. Res. 104, 13073–13082. NISHIMURA, R., TSUNOGAI, U., ISHIBASHI, J., WAKITA, H., and NOJIRI, Y. (1999) Origin of 13C-enriched methane in the crater lake Towada, Japan, Geochem. J. 33, 277–283. OHNO, M., SATO, T., NOTSU, K., WAKITA, H., and OZAWA, K. (1999) Groundwater-level changes in response to bursts of seismic activity off the Izu Peninsula, Japan, Geophys. Res. Lett. 26, 2501–2504. SATO, M., TSUNOGAI, U., ISHIBASHI, J., NOTSU, K., and WAKITA, H. (1999) Carbon isotope measurement of extremely low amounts of CH4: Application to volcanic gases from Satsuma-Iwojima, Japan, Analytical Sciences 15, 513–516. SATO, M., MORI, T., NOTSU, K., and WAKITA, H. (1999) Carbon and helium isotopic composition of fumarolic gases and hot spring gases from Kirishima volcanic area, Bull. Volcanol. Soc. Japan 44, 279–283 (in Japanese with English abstract). WAKITA, H. (1999) Geochemical features of natural gas in Japan, The Search for Deep Gas (M. J. WHITICAR and E. FABER), pp. 127–141, ‘‘Geologisches Jahrbuch’’ Reihe D, Heft 107, Hannover. YOKOYAMA, T., NAKAI, S., and WAKITA, H. (2000) Helium and carbon isotopic compositions of hot spring gases in the Tibetan Plateau, J. Volcanol. Geotherm. Res, 88, 99–107.
180
Publication List
Pure appl. geophys.,
HERNA´NDEZ, P., PE´REZ, N.M., SALAZAR, J.M., SATO, M., NOTSU, K., and WAKITA, H. (2000) Soil gas CO2, CH4 and H2 distribution in and around Canada’s caldera, Tenerife, Canary Islands, Spain, J. Volcanol. Geotherm. Res. 103, 425–438. KING, C.-Y., AZUMA, S., OHNO, M., ASAI, Y., HE, P., KITAGAWA, Y., IGARASHI, G., and WAKITA, H. (2000) In search of earthquake precursors in the water-level data of 16 closely clustered wells at Tono, Japan, Geophys. J. Int. 143, 469–477. NOTSU, K., NAKAI, S., IGARASHI, G., MORI, T., Suzuki, M., and WAKITA, H. (2001) Spatial distribution and temporal variation of 3He/4He in hot spring gas release from Unzen volcano area, Japan, J. Volcanol. Geotherm. Res. 111, 89–98. ISHIBASHI, J., SATO, M., SANO, Y., WAKITA, H., GAMO, T., and SHANKS III, W.C. (2002) Helium and carbon gas geochemistry of pore fluids from the sediment-rich hydrothermal system in Escanada Trough, Appl. Geochem. 17, 1457–1466. SUMINO, H., NOTSU, K., NAKAI, S., SATO, M., NAGAO, K., HOSOE, M., and WAKITA, H. (2004) Noble gas and carbon isotopes of fumarolic gas from Iwojima volcano, Izu-Ogasawara arc, Japan: Implications for the origin of unusual arc magmatism, Chem. Geol. 209, 153–173. HERNA´NDEZ, P.A., PA´rez, N.M., SALAZAR, J.M., REIMER, M., NOSU, K., and WAKITA, H. (2004) Radon and helium in soil gases at Canadas caldera, Tenerife, Canary Islands, Spain, J. Volcanol. Geotherm. Res. 131, 59–76. OHNO, M., SATO, T., NOTSU, K., WAKITA, H., and OZAWA, K. (2006) Groundwater-level changes due to pressure gradient induced by nearby earthquakes off Izu Peninsula, 1999, Pure Appl. Geophys. 163, 647–655. WAKITA, H. Volcanic events and their social and economic impacts, J. Volcanol. Geotherm. Res. (to be published). PE´REZ, N. M., HERNA´NDEZ, P. A., IGARASHI, G., TRUJILLO, I. NAKAI, S., SUMINO, H., and WAKITA, H., Searching and detecting earthquake geochemical precursors in CO2-rich groundwaters from Galicia, Spain, Geochemical Journal (to be published).