The Distribution and Origin of Radon, CO 2 and SO2 Gases ... · The Distribution and Origin of...

The Distribution and Origin of Radon, CO2 and SO2 Gases atArenal Volcano, Costa Rica

par

Glyn Williams-JonesDépartement de géologie

Faculté des Arts et des Sciences

Mémoire présenté à la Faculté des études supérieuresen vue de l’obtention du grade de

Maître ès sciences (M.Sc.)

Avril 1996Université de Montréal

Glyn Williams-Jones - MCMXCVII

Université de MontréalFaculté des études supérieures

Ce mémoire intitulé

The Distribution and Origin of Radon, CO2 and SO2 Gases atArenal Volcano, Costa Rica

présenté par:

Glyn Williams-Jones

a été évalué par un jury composé des personnes suivantes:

Président-rapporteur: Dr. Walter E. Trzcienski

Membre du jury: Dr. Hélène Gaonac’h

Membre du jury: Dr. John Stix

Frontispiece

Arenal

To all those that made this thesis possible.

ii

Abstract

Volcanic gases are one of several important indicators used to better understand

and forecast volcanic activity. However, direct sampling of these gases is often

dangerous or impossible due to the high level of activity and the common inaccessibility

of the crater areas of many volcanoes. Indirect methods such as the study of soil gases or

the use of remote sensing techniques are thus required. Soil gases such as radon and

carbon dioxide have been shown to correlate well with variations in volcanic activity.

Similarly, the remote sensing of gases such as sulphur dioxide has proven significant in

the geochemical characterisation of both passively and actively degassing volcanoes.

Techniques such as these can now provide important clues to the behaviour and future

activity of the volcano.

This thesis investigates the degassing of Arenal volcano. A small stratovolcano

in northwestern Costa Rica, Arenal is one of the most active volcanoes in Central

America, having been in continuous eruption since its reactivation in July 1968.

Estimates, using petrologic and remote sensing techniques, are made of the quantity of

SO2 emitted from Arenal since 1968 and are related to a degassing model for the

volcano. Observed spatial and temporal patterns of soil and plume gases are correlated

to eruptive and seismic activity, and the origin and transport of these gases at Arenal is

discussed. Measurements of seismicity, radon, CO2 and SO2 gas were made as (1) the

results could be compared to other volcanoes where similar measurements have been

made, (2) it was comparatively simple to measure radon, CO2, and SO2, and (3) these

gases are believed to respond to changes in activity and the stress-state of the volcano.

iii

Resumen

Los gases volcánicos son uno de los varios indicadores importantes usados para

entender mejor y pronosticar la actividad volcánica. Sin embargo, el muestreo directo de

éstos gases es frecuentemente peligroso o imposible a causa del alto nivel de actividad y

la usual inaccesibilidad del área de cráter de muchos volcanes. Los métodos indirectos

tal como el estudio de gases de suelo o el uso de técnicas de teledetección son

necesarios. Se observó que los gases de suelo tal como radon y el dióxido de carbono

son apropiados para correlacionar bien las variaciones de la actividad volcánica.

Igualmente, los gases estudiados a distancia tal como: dióxido de azufre, ha probado ser

importante en la caracterización geoquímica pasiva y activa de volcanes con emisiones

gaseosas. Técnicas como éstas pueden proveer ahora, pistas importantes del

comportamiento y actividad futura del volcán.

Esta tésis investiga la emisión gaseosa del volcán Arenal, un pequeño

estratovolcán en el noroeste de Costa Rica. Arenal, es uno de los volcanes más activos

en Centroamérica, con erupción continua desde su reactivación en Julio de 1968. La

estimación, por medio de técnicas petrológicas y de teledetección, se hizo teniendo en

cuenta la cantidad de gas SO2 emitidas del volcán desde 1968, y son relativas a un

modelo de emanaciones gaseosas para Arenal. Los modelos espaciales y temporales de

los gases de pluma y suelo observados, se correlacionan con la actividad sísmica y

eruptiva. También, el orígen y el transporte de éstos gases en Arenal se discute en esta

tésis.

Las medidas de sismicidad y medidas particulares de los gases radon, CO2 y SO2

se hicieron con motivo de que: (1) los resultados podrían compararse a otros volcanes

iv

donde medidas similares son disponibles, (2) los gases radon, CO2, y SO2 son

comparativamente simples de medir, y (3) se piensa que éstos gases responden a

cambios en la actividad volcánica y el estado de tensión del volcán.

v

Résumé

L’Arenal est un stratovolcan situé à 10.463°N 84.703°O dans le nord-ouest du

Costa Rica en Amérique Centrale. Arenal est le volcan costaricien le plus petit, avec un

volume de 15 km3, mais aussi le volcan le plus actif du pays. Il consiste d’un édifice

avec des flancs pentus formé de deux cratères sommitaux (C and D), et il est verdoyant

sur les flancs nord, sud, et est. Un grand champ de laves jeunes, mis en place depuis

1968, couvre le flanc ouest. Arenal est situé entre deux massifs, la Cordillera de

Guanacaste au sud-est et la Cordillera Central au nord-ouest. L’ensemble de ces deux

cordillères forment la chaîne volcanique qui comprend l’arc du Costa Rica. Environ

trois kilomètres au sud de l’Arenal on retrouve le volcan dormant de Cerro Chato.

Le matin du 29 juillet, 1968, après 10 heures d’activité sismique intense, l’Arenal

est entré en éruption de façon explosive, il a continué son activité éruptive durant une

période de trois jours, tuant ainsi 78 personnes et dévastant une région de 12 kilomètres

carrés sur le flanc ouest. L’explosion initiale était suivie par des colonnes d’éruptions

pliniennes, des coulées pyroclastiques et des bombes et blocs éjectés de façon balistique.

Trois nouveaux cratères (A, B, C) ont été formés durant cette période, avec une

orientation approximative est-ouest sur le flanc ouest du volcan. Un autre épisode

explosif a commencé le 17 juin 1975, avec l’emplacement d’une coulée pyroclastique de

cendres et blocs le long de la vallée Rio Tabacon. Ce dépôt provient de la formation des

nuées ardentes produites par des avalanches d’une coulée de lave provenant du cratère C.

Jusqu’en juin 1984, l’Arenal a continué son activité fumerolienne forte avec l’extrusion

de lave bloqueuse de type aa. Cette date marque une augmentation de l’activité à

Arenal, avec le début d’éruptions de cendres et de grandes coulées pyroclastiques. Cette

vi

activité se change en phase éruptive strombolienne produisant des tephras et des laves de

composition basaltique-andésitique. De nos jours le volcan Arenal est encore dans cette

phase strombolienne.

Les objectifs de ce mémoire sont, premièrement l’évaluation de la distribution

des gaz de sol sur les flancs du volcan Arenal, deuxièmement l’estimation de la quantité

du gaz SO2 dégagé par le volcan Arenal depuis 1968, en utilisant des techniques

pétrologique et télédétectée, troisièmement la tentative de trouver des patrons spatial et

temporel observés dans les gaz de sol et les fumerolles durant l’activité éruptive et

sismique, et finalement une discussion sur l’origine et le transport des gaz pour le volcan

Arenal. Pour réussir à ces objectifs, des mesures de sismicité, et des mesures de gaz de

radon, de CO2 et de SO2 ont été effectuées avec l’idée de comparer les résultats à

d’autres volcans où des mesures semblables ont été faites. De plus, il était relativement

simple de mesurer le radon, le CO2, et le SO2, et il est suggéré que ces gaz répondent

bien aux changements de l’activité volcanique et à l’état de stress du volcan.

Des mesures des concentrations de radon et CO2 ont montré des maxima

seulement près des failles possibles et sur les flancs inférieurs du volcan. Les données

de δ13C ont aussi été les plus lourdes sur les flancs inférieurs et près de ces failles

possibles. Il y a peu d’expression de la structure en surface du volcan Arenal parce qu’il

est couvert de laves récentes. Le niveau d’activité élevé de l’Arenal rend difficile les

corrélations entre l’activité sismique et les fluctuations des gaz de sol. Par contre, ces

fluctuations peuvent être expliquées par des variations dans la pression atmosphérique.

Ces observations impliquent que les concentrations des gaz de sol sont influencées

principalement par le niveau de développement du sol. Ensuite, le dégazage diffus des

vii

gaz magmatiques profonds sur les flancs supérieurs des volcans est négligeable dû à la

faible perméabilité du sol causée par le couvert des roches volcaniques jeunes.

Finalement, l’augmentation du dégazage magmatique sur les flancs inférieurs est le

résultat d’une augmentation de la fracturation des laves plus âgées.

Les gaz de sol mesurés pour deux autres stratovolcans actifs associés à la

subduction (le Poás, Costa Rica et le Galeras, Colombie) ressemblent bien à ce qui a été

observé à l’Arenal. Les concentrations de radon étaient maximales seulement près des

zones de failles, des zones d’activité sismique, près des cratères et fumerolles, et sur les

flancs inférieurs de ces volcans. Les valeurs les plus negatives de δ13C ce trouvaient

près des fumerolles à l’intérieur des cratères actifs, près des failles et sur les flancs

inférieurs. Ces observations impliquent que les failles majeures peuvent canaliser les

gaz profonds vers la surface seulement s’ils ont une expression superficielle. Des

volcans, comme ceux étudiés ici, réagissent comme des bouchons dans la croûte

continentale, limitant le dégazage aux fumerolles, failles, et flancs inférieurs fracturés.

L’utilisation de la télédétection des gaz pour l’étude des volcans actifs est encore

jeune. Le seul gaz qui peut être télédétecté de façon routinière est le dioxyde de soufre,

en utilisant la spectroscopie de corrélation dans la région ultraviolette du spectre

électromagnétique. Par l’intégration de ces données avec d’autres données

géophysiques, les changements temporaux dans les flux de SO2 peuvent suggérer des

comportements et activités dans l’avenir du volcan.

Arenal a dégagé un minimum de 1.31 x 106 tonnes métriques de SO2 depuis sa

réactivation en 1968. Les émissions du volcan ont continué avec une production

moyenne quotidienne de 130 ± 60 t/d SO2. Par sa forte activité, Arenal montre des

viii

cycles de diminution du flux de SO2 et de sismicité avant les éruptions. Suite à ces

événements, le flux de SO2 et l’activité sismique augmente. Ces fluctuations montrent

aussi une corrélation distincte avec les marées terrestres. En effet, lors du maximum de

marée, nous observons une diminution de l’activité éruptive, ce qui coïncide avec

l’augmentation des événements tremor. Arenal a probablement une chambre

magmatique profonde ou mi-croûte se comportant comme un système ouvert. Par

contre, le système ouvert est périodiquement bloqué près de la surface par la

cristallisation du magma dans le conduit et le développement subséquent d’une zone

étanche. Ceci cause la surpression du conduit et la destruction explosive éventuelle de

cette zone.

Les gaz volcaniques sont un outil pour mieux comprendre et prédire l’activité

volcanique. Donc, la télédétection des gaz de sol ou de la colonne est critique dans la

caractérisation géochimique, car elle est une méthode sécuritaire et efficace pour les

volcans actifs qui sont souvent inaccessibles.

ix

Acknowledgements

I would like to thank my supervisor, Prof. John Stix, for his constant enthusiasm,

encouragement and support, as well as his ability to put up with my terrible jokes.

Thanks to Martin Heiligmann for stimulating discussion and constructive criticism, his

help in Costa Rica, and having to submit (graciously, most often) to really early

mornings...wackey wackey! Many thanks to Isabel Lépine for help with translating, final

production, and many other things. Thanks also to the many other friends at the

Université de Montréal for accepting la Tête Carrée into their midst over these many

years, Jean-Marc Séguin, Chantal Bilodeau, Stephan Lamarche, Alain Legault, Kazuko

Saruwatari, Sandrine Caderon, André Lafferière, and Alex Beaulieu. I’m grateful to

Jorge Barquero, Erik Fernández, Eliazar Duarte, and Eduardo Malavassi of OVSICORI

for welcoming the gringo with open arms - Pura Vida! Thanks also to Vilma Barboza

(OVSICORI) for help with 1995 and 1996 seismic data. Special thanks to Ray Hoff and

Andrew Sheppard (Environment Canada) for the use of their COSPEC in 1996 and Niki

Stevens (Reading Univ.) and Mark Davis (Open Univ.) for their enthusiastic help during

the second field season in Costa Rica. Also many thanks to Neil Arner and Barbara

Sherwood Lollar at the University of Toronto, without whom my isotope work would

have been a nightmare! Thanks to Hazel Rymer (Open Univ.) for tidal gravity data

which turned out to be crucial, to Glen Poirier (McGill Univ.) for help and patience with

the microprobe analyses, and to Bill Melson (Smithsonian Inst.) for seismic data from

1995.

This research was supported by grants to John Stix from the Natural Sciences and

Engineering Research Council of Canada (NSERC), les Fonds pour la Formation de

x

Chercheurs et l’Aide à la Recherche (FCAR), and the Université de Montréal. Thanks

finally and most of all, to my parents and brother, I would not have made it here without

you.

xi

Table Of Contents

ABSTRACT__________________________________________________________ iiRESUMEN __________________________________________________________ iiiRÉSUMÉ_____________________________________________________________vACKNOWLEDGEMENTS________________________________________________ ixLIST OF FIGURES ___________________________________________________ xiiiLIST OF TABLES ___________________________________________________ xviiPREFACE _________________________________________________________ xviii

GENERAL INTRODUCTION ___________________________________________ 1INTRODUCTION _______________________________________________________2OBJECTIVES _________________________________________________________3LOCATION___________________________________________________________3GEOLOGICAL SETTING_________________________________________________5

REGIONAL GEOLOGY ___________________________________________ 5LOCAL GEOLOGY ______________________________________________ 9

VOLCANIC ACTIVITY _________________________________________________11VOLCANIC ACTIVITY PRIOR TO 1968______________________________ 11VOLCANIC ACTIVITY DURING AND AFTER 1968 _____________________ 11

REFERENCES________________________________________________________16

CHAPTER I __________________________________________________________ 20ABSTRACT__________________________________________________________21INTRODUCTION ______________________________________________________22FACTORS AFFECTING SOIL GAS CONCENTRATIONS AND DISTRIBUTIONS________23METHODOLOGY _____________________________________________________24

STATION LOCATIONS __________________________________________ 24SOIL GAS MEASUREMENTS _____________________________________ 24SEISMIC MEASUREMENTS ______________________________________ 31

RESULTS ___________________________________________________________31RADON _____________________________________________________ 31RADON EMANATING POTENTIAL _________________________________ 32CARBON DIOXIDE_____________________________________________ 38CARBON ISOTOPES IN CO2 SOIL GAS ______________________________ 38CARBON DIOXIDE FLUX ________________________________________ 49SOIL GAS TIME SERIES _________________________________________ 52

RADON ___________________________________________________ 52CARBON DIOXIDE ___________________________________________ 54

SEISMICITY __________________________________________________ 55ATMOSPHERIC VARIATIONS_____________________________________ 58

DISCUSSION_________________________________________________________58ORIGIN OF RADON AND CARBON DIOXIDE _________________________ 58

xii

SOIL GAS, SOIL DEVELOPMENT, AND ELEVATION ___________________ 62RELATIONSHIP TO PRESSURE CHANGE ____________________________ 67RELATIONSHIP TO SEISMICITY ___________________________________ 69

CONCLUSIONS_______________________________________________________70REFERENCES________________________________________________________72

CHAPTER II _________________________________________________________ 74ABSTRACT__________________________________________________________75INTRODUCTION ______________________________________________________76METHODOLOGY _____________________________________________________76RESULTS ___________________________________________________________81DISCUSSION_________________________________________________________82CONCLUSIONS_______________________________________________________85REFERENCES________________________________________________________86

CHAPTER III ________________________________________________________ 88ABSTRACT__________________________________________________________89INTRODUCTION ______________________________________________________90METHODOLOGY _____________________________________________________94

SO2 FLUX ___________________________________________________ 94SO2 FLUX ERRORS ____________________________________________ 98SEISMICITY _________________________________________________ 102

RESULTS __________________________________________________________102SO2 FLUX __________________________________________________ 102SO2 BUDGETS AT ARENAL _____________________________________ 111

COSPEC/PLUME TRACKER ESTIMATES__________________________ 111PETROLOGICAL ESTIMATES ___________________________________ 111

SEISMIC DATA ______________________________________________ 114DISCUSSION________________________________________________________119

CONDUIT OPENING AND CLOSING _______________________________ 119THE SULPHUR BUDGET OF ARENAL______________________________ 122THE OPEN NATURE OF ARENAL_________________________________ 124

CONCLUSIONS______________________________________________________125REFERENCES_______________________________________________________127

CONCLUSIONS _____________________________________________________ 131GENERAL CONCLUSIONS _____________________________________________132RECOMMENDATIONS FOR FUTURE WORK________________________________134

APPENDIX _________________________________________________________ 136

xiii

List Of Figures

Figure I-1 Geographic map of Costa Rica showing the location of Arenalvolcano 4

Figure I-2 Topographic map of Arenal volcano showing the lava fieldsemplaced since 1968 6

Figure I-3 Geological map of Costa Rica. After Mora (1983), Alvarado(1984), and Borgia (1988) 7

Figure I-4 Geological map of Arenal volcano. After Borgia (1988) 10

Figure I-5 Blocks from the blocky ash flow emplaced on the western flankof Arenal in July 1968. A 1992 blocky-aa lava flow is seen inthe background 13

Figure I-6 White cloud on left of image represents a small pyroclastic flowcaused by the collapse of part of the blocky aa flow (a levee isin evidence to the upper right of the cloud). View is of thewestern flank 15

Figure I-7 A small strombolian eruption towards the western flank. CraterC is to left, crater D, to the right. View from the ArenalVolcano Observatory, south of the volcano 15

Figure 1.1 Topographic map of Arenal volcano showing the location of Rnand CO2 soil gas stations (stars). The locations of seismic stationsare shown (red triangles). Contours are every 100 m 26

Figure 1.2 PVC tube (left) used in measuring radon and an aluminium tube(right) used in the measurement of CO2 soil gas. Both tubes areburied to a depth of ~75 cm 27

Figure 1.3 (a) Average Rn (pCi/l) versus elevation of stations (metres).There is an approximately negative linear trend. (b) Topographicmap of Arenal volcano showing concentrations of Rn (pCi/l)soil gas. There is a tendency towards increasing radon withdistance from the summit 35

Figure 1.4 Radon (pCi/l, Table 1.4) versus radon emanating potential (RnERaCin pCi/kg, Table 1.2). Stations with elevated radon generally haveelevated RnERaC 37

Figure 1.5 (a) Average CO2 (%) versus elevation of stations (m). There is a

xiv

general trend of increasing concentration with decrease in elevation.(b) Topographic map of Arenal volcano showing the concentrationsof CO2 (%) soil gas. There is a tendency towards increasing CO2with distance from the summit 42

Figure 1.6 (a) Average Rn (pCi/l) versus average CO2 (%). Note the generally positive correlation for the N, S and W lines. Stations from the NEand E lines are anomalous. (b) Average Rn (pCi/l) normalised byRnERaC (pCi/kg) versus average CO2 (%). Stations from the NEline are anomalous 43

Figure 1.7 (a) Average δ13C (‰) in CO2 soil gas versus elevation of stations(metres). Note the negative correlation (r = -0.73). (b) Topographicmap of Arenal volcano showing concentrations of δ13C (‰) in CO2

soil gas. Note the tendency towards heavier δ13C with distancefrom the summit 45

Figure 1.8 (a) δ13C (‰) in CO2 soil gas versus CO2 (%). Note the positivecorrelation. (b) Average δ13C (‰) in CO2 soil gas versus averageRn (pCi/l). (c) Average δ13C (‰) in CO2 soil gas versus averageRn (pCi/l) normalised by RnERaC (pCi/kg). Stations from the NEline are anomalous 47

Figure 1.9 (a) δ13C (‰) in CO2 soil gas versus organic δ13C (‰) in soil.Samples are for the NE line which are stations furthest from thesummit

50

Figure 1.10 (a) CO2 (%) soil gas versus CO2 flux (mg/m2⋅min). Note thegrouping of points where samples with elevated CO2 have low CO2

flux. (b) δ13C (‰) versus CO2 flux (mg/m2⋅min). Note that stationswith low flux have heavier δ13C 51

Figure 1.11 Plot of radon and CO2 concentrations versus time for the E, S, N,and W lines. Solid lines and symbols are radon values, dashed andclear symbols are CO2 concentrations. Note the common peaks onMarch 22 and April 6, 1995 53

Figure 1.12 (a) Number of volcanic eruptions per day and (b) total hours oftremor per day between March 23 and April 16, 1995 57

Figure 1.13 Fluctuations in radon concentration (pCi/l) and number of eruptionsversus time. Note the increase in average number of eruptionscoinciding with a peak in Rn concentration (April 6, 1995). Thehistogram represents the average number of eruptions between Rn

xv

measurement periods 59

Figure 1.14 Daily atmospheric pressure (mbar) fluctuations measured at theFortuna base station (~250 m). Note the relative minima onApril 6, 1995 61

Figure 1.15 View from the western flank of Arenal towards Lago Arenal. Largeboulders and blocks are part of the 1968 pyroclastic deposit. Theremains of the once dense jungle (a few trees) are visible 64

Figure 2.1 Topographic map of Arenal volcano showing average (a) radonconcentrations in pCi/l, (b) CO2 concentrations in volume %, and(c) δ13C values expressed as ‰. Contour interval of 100 m 77

Figure 2.2 Topographic map of Poás volcano showing average (a) radonconcentrations in pCi/l, (b) CO2 concentrations in volume %, and(c) δ13C values expressed as ‰. Contour interval of 500 m 78

Figure 2.3 Topographic map of Galeras volcano showing average (a) radonconcentrations in pCi/l, (b) CO2 concentrations in volume %, and(c) δ13C values expressed as ‰. Contour interval of 400 m 79

Figure 3.1 Geographic map of Costa Rica showing the location ofArenal volcano 91

Figure 3.2 Topographic map of Arenal volcano showing the locations ofseismometer stations (red triangle) and tilt stations (green house).Craters A and B are now buried by lava flows emplaced since1968. Contours are every 100 m 92

Figure 3.3 Plume Tracker and COSPEC ultraviolet spectrometers. The right-angle light tube of the Plume Tracker is visible in the top picture,while the bottom picture shows the control panel for the COSPECIV. Black and red cables connect to an analogue chart recorderand portable computer 95

Figure 3.4 Tilt station maintained by the Departamento de Geología of theInstituto Costaricense de Electricidad (ICE), located on the upperwestern flank of Arenal volcano. A portable seismometer wasburied approximately 50 cm below the surface, to the right of thedoor of the green hut 103

Figure 3.5 SO2 flux versus time for (a) February 28, (b) March 4, (c) March 5,1995 at Arenal. Eruptions are shown by inverted arrows 109

xvi

Figure 3.6 SO2 flux and eruption amplitude and SO2 flux and eruption durationversus time for (a),(b) February 29; (c),(d) March 1; (e),(f) March 5;(g),(h) March 6; (i),(j) March 8, 1996. Amplitude is in digital units,duration in seconds, and SO2 flux in metric tonnes per day 110

Figure 3.7 Fluctuation of gravity due to Earth tides between (a) February 29to (i) March 8, 1996. Gravity data is in microgals. Note that tremorgenerally begins at or just past high tide on March 5, 6, and 8 118

xvii

List Of Tables

Table 1.1 Radon soil gas from Arenal volcano for 1995-1995 34

Table 1.2 Radon emanating potential for soils and lava/debris at Arenalvolcano 36

Table 1.3 CO2 soil gas from Arenal volcano for 1995-1996 39

Table 1.4 Average CO2, radon and δ13C from Arenal volcano for 1995-1996 41

Table 1.5 δ13C in CO2 soil gas from Arenal volcano for 1995-1996 44

Table 1.6 Organic δ13C in soil from the NE line of Arenal volcano 48

Table 1.7 CO2 flux from Arenal volcano in 1995 48

Table 1.8 Seismic data from March 23 to April 16, 1995 at Arenal volcano 56

Table 1.9 Correlation coefficients for Rn vs. CO2 and Rn and CO2 vs.pressure at Arenal volcano 68

Table 2.1 Average Rn, CO2, and δ13C for Arenal, Poás, and Galerasvolcanoes 80

Table 3.1 SO2 concentrations of gas calibration cells for SO2 remote sensinginstruments 97

Table 3.2 Error calculation for SO2 measurements 99

Table 3.3 Windspeed measurements taken 4 km west (elev. ~550 m) of Arenalvolcano 101

Table 3.4 Daily SO2 flux for Arenal volcano from 1982 to 1996 104

Table 3.5 Average daily SO2 flux for Arenal volcano with and withouteruptions 106

Table 3.6 Chemical analyses of melt inclusions and matrix glasses from the1968 surge deposit and 1984 and 1992 lavas at Arenal volcano 112

Table 3.7 Predicted daily maximums of tidal gravity at Arenal volcano 117

xviii

Preface

This thesis consists of five chapters, the third and fourth of which are in

manuscript form, and are intended for submission to refereed journals. My thesis

advisor, Dr. John Stix, is second author on both manuscripts. His role in the preparation

of the manuscripts consisted of critical evaluation of the data and my interpretations

presented therein, as well as editorial suggestions regarding organisation of the text.

GENERAL INTRODUCTION

General Introduction______________________

2

Introduction

OLCANIC gases are one of several important indicators used to better

understand and forecast volcanic activity. However, direct sampling of these

gases is often dangerous or impossible due to the high level of activity and the common

inaccessibility of the crater areas of many volcanoes. Indirect methods such as the study

of soil gases or the use of remote sensing techniques are thus required. Soil gases such

as radon or carbon dioxide have been shown to correlate well with variations in volcanic

activity (cf. Allard et al., 1991; Brantley and Koepenick, 1995; Heiligmann et al., 1997).

Similarly, the remote sensing of gases such as sulphur dioxide has proven significant in

the geochemical characterisation of both passively and actively degassing volcanoes (cf.

Casadevall et al., 1981; Stoiber et al., 1986; Zapata et al., 1997). Techniques such as

these can now provide important clues to the behaviour and future activity of a volcano.

Arenal volcano, a small stratovolcano in northwestern Costa Rica, is one of the

most active volcanoes in Central America, having been in continuous eruption since its

reactivation in July 1968. Until the onset of this new activity, very little was known

about the geology or volcanic activity of Arenal. The geology of Arenal was

comprehensively described and mapped by Malavassi (1979). This work was followed

by that of Alvarado (1984), and Borgia et al. (1988) who continued the study of the

structural, stratigraphic and petrological evolution of the Arenal-Chato system.

Minakami et al. (1969), Melson and Saenz (1968, 1973, 1977), Fudali and Melson

(1972), and Saenz (1977) investigated the 1968 eruptive activity and estimated many of

the physical characteristics (e.g., volume, energy and released pressure), while Matumoto

and Umana (1976) and Van der Bilt et al. (1976) described the 1975 eruptive events.

V


3

Bennett and Raccichini (1977), Borgia et al. (1983), and Cigolini et al. (1984) studied

the dynamics and structure of lava flows while Wadge (1983) and Reagan et al. (1987)

investigated the origins of lava extruded between 1968 and 1986 and determined that

significant chemical changes in the magma had occurred since 1968. Casadevall et al.

(1984) made some of the first COSPEC sulphur dioxide measurements and particle

studies in Arenal’s volcanic plume.

Objectives

This thesis was undertaken to (1) evaluate soil gas distribution on the flanks of

Arenal volcano, (2) estimate, using petrologic and remote sensing techniques, the

quantity of SO2 emitted from Arenal since 1968, (3) attempt to correlate the observed

spatial and temporal patterns of soil and plume gases to eruptive and seismic activity,

and (4) discuss the origin and the transport of these gases at Arenal. In order to achieve

these objectives, I measured radon, CO2 and SO2 gas because (1) the results can be

compared to other volcanoes where similar measurements have been made, (2) it is

comparatively simple to measure radon, CO2, and SO2, and (3) these gases are believed

to respond to changes in volcanic activity and the stress-state of the volcano.

Location

Arenal is a stratovolcano located at 10.463°N 84.703°W in northwestern Costa

Rica, 90 km northwest of the capital, San José (Figure I-1). The volcano is a steep sided

Kil

omet

res

010

050

SA

NJO

SÉ

SA

NJO

SÉ

NIC

AR

AG

UA

PAN

AM

A

Pun

tare

nas

Nic

oyaL

iber

ia

1000

'o

1000

'o

Are

nal

volc

ano

Are

nal

volc

ano

San

Isid

roS

anIs

idro G

olfi

to

NO

RT

HPA

CIF

ICO

CE

AN

NO

RT

HPA

CIF

ICO

CE

AN

Pue

rto

Lim

ónP

uert

oL

imón

Cab

oG

raci

asa

Dio

sC

abo

Gra

cias

aD

ios

CA

RIB

BE

AN

SE

AC

AR

IBB

EA

NS

EA

8400

'o

8400

'o

General Introduction_____________________

4

Figure I-1: Geographic map of Costa Rica showing the location of A r e n a lv o l c a n o .


5

edifice vegetated to the north, east and south. The volcano consists of two summit

craters (C and D). A large field of young (post-1968) lava covers the western flank

(Figure I-2). Arenal is situated between two massifs, the Cordillera de Guanacaste (SE)

and the Cordillera Central (NW) which together form the volcanic chain that makes up

the Costa Rican Arc (Figure I-3, Stoiber and Carr, 1973). Approximately three

kilometres south of Arenal lies the small truncated and dormant volcano, Cerro Chato

(Figure I-2).

Geological Setting

Regional Geology

Arenal volcano is situated in the Central American volcanic chain, on the

boundary between the northern and central Costa Rican segments (Stoiber and Carr,

1973). The Cocos Plate is being subducted under the Caribbean Plate along the

Mesoamerican trench northeast of Arenal (Figure I-3). Costa Rica consists of six

principal geological provinces paralleling the Mesoamerican trench: 1) the Cretaceous to

Middle Tertiary ophiolitic suite; 2) Tertiary basins; 3) Tertiary volcanic ranges; 4) Active

Quaternary volcanic ranges; 5) Intra-arc basins; and 6) the Caribbean coastal plain

(Mora, 1983; Alvarado, 1984; Borgia et al., 1988).

An ophiolitic suite is found in the Nicoya Complex, which is comprised of

cherts, graywackes, tholeiitic pillow lavas and basaltic agglomerates. It is intruded by

gabbroic, diabasic, and dioritic rocks. The Tertiary basin is composed of sediments of

mainly marine origin, intercalated with volcaniclastic deposits. The Tertiary volcanic


6

Figure I-2: Topographic map of Arenal volcano showing the lava fieldsemplaced since 1968.

Kilometres

0 10050

Cocos Plate

Cretaceous to Middle Tertiary ophiolitic suite

Tertiary basins

Tertiary volcanic range

Active Quaternary volcanic range

Intra-arc basins

Caribbean coastal plain

Middle America Trench

CaribbeanPlate

Arenal-Chatovolcanic system

Irazú

Turrialba

Submerged contacts

Contacts betwen major geologicalprovinces

Major volcanoesCinder cones

Arenal-Chato volcanic system

Historically active volcanoes

�

Poás Barva

Orosi

Tenorio

Miravalles

Rincón dela Vieja


7

Figure I-3: Geological map of Costa Rica. After Mora (1983), Alvarado (1984)and Borgia (1988).


8

range is made up of a block-faulted horst, the Sierra de Tilaran y Abangares, extending

southeast into the Montes del Aguate. Composed of andesitic and basaltic flows,

volcanic agglomerates and tuffs, this range is part of the late Miocene-early Pliocene

Aguacate Volcanic Group (Dengo, 1962; Malavassi, 1979; Weyl, 1980). The

Quaternary Volcanic ranges comprise the Cordillera de Guanacaste to the northwest and

Cordillera Central to the southeast. The Cordillera de Guanacaste contains five

stratovolcanoes (Orosi, Rincón de la Vieja, Miravalles, Tenorio and Arenal), eruptive

products of which have compositions that vary from basaltic andesite (Melson and

Saenz, 1973) to andesite (Dengo, 1962; Pichler and Weyl, 1973). Large-scale rhyolitic

and dacitic tuffs crop out on the southwest flank of the Cordillera and overlie part of the

Aguacate Volcanic Group and Nicoya Complex (Dengo, 1962). The Cordillera Central

consists of four stratovolcanoes (Poás, Barva, Irazu, and Turrialba), deposits of which

have compositions that vary from basalt to dacite and andesite (Pichler and Weyl, 1973).

Extensive mudflows and volcanic ash deposits are exposed on the southwest side of the

Cordillera (Williams, 1952). Intra-arc basins comprise the Arenal graben to the

northwest and the Valle Central to the southeast. The Valle Central basement consists of

slightly folded Oligocene and Miocene marine sediments, overlain by tuffs, lavas and

ignimbrite sheets of the Cordillera Central (Weyl, 1980). Finally, the Caribbean coastal

plain is a sedimentary basin of Early Tertiary age composed of river alluvium and lahar

deposits from the volcanoes of the Cordillera Central. Some small Quaternary cinder

cones also are found in the coastal plains near Tortuguero (Weyl, 1980).


9

Local Geology

Arenal has a volume of only 15 km3 (Carr, 1984) and is the smallest but most

active of seven historically active Costa Rican volcanoes. The tectonic setting of the

volcano is disputed, with some authors suggesting that Arenal overlies a tear in the

subducting Cocos plate (Carr et al., 1979; Carr, 1984; Burbach et al., 1984) and others

believing there is a smooth transition in the orientation of the Wadati-Benioff zone,

thought to lie 150 km below Arenal (Guendel et al., 1984; Reagan et al., 1987). The

small truncated and dormant volcano, Cerro Chato, lies approximately three kilometres

southeast of Arenal (Figure I-4). Arenal is most likely directly tapping a lower to mid-

crustal magma chamber, possibly located at a discontinuity which lies at a depth of 22

km (Matumoto et al., 1977; Wadge, 1983; Reagan et al., 1987).

Three stages of differing magma compositions at Arenal are believed to coincide

with variations in eruptive activity. Stage-1 zoned magmas likely resided in the magma

chamber prior to the 1968 eruption. A new magma intruded into the chamber resulting

in the ejection of the stage-1 magma in July 1968. It subsequently mixed with the more

mafic parts of stage 1 to produce stage-2 magmas. Stage-3 magmas (mid-1974 to

present) are the product of continued mixing and fractional crystallisation along the

walls of the conduit and chamber. Each change in stage appears to correlate with a

variation in the cumulative volume of extruded material (Reagan et al., 1987).

The rocks around Arenal range in age from Pliocene-Pleistocene to Holocene and

are divided into five lithologies: 1) undivided Pliocene-Pleistocene volcanics, 2) Chato

lava flows, 3) Arenal lava flows, 4) undivided tephra from Arenal and Chato, and 5)

utA

C1

96

8

utA

C1

97

5

uT

AC

utA

C

uP

Pv

LA

uT

AC

utA

C1

23

5

Co

nta

ct

Riv

ers

utA

C-

un

div

ided

tep

hra

fro

mA

ren

alan

dC

hat

ou

tAC

dat

e-

12

35

,1

96

8&

19

75

py

rocl

asti

cfl

ow

s

Py

rocl

asti

cfl

ow

Sed

imen

tary

dep

osi

ts

Lak

es

A1

-p

ost

19

68

A2

-1

70

0to

18

00

A.D

.A

3,

A4

-~

10

80

A.D

.L

A-

22

0B

.C.

Are

nal

lav

as

>1

55

0B

.C.

Ch

ato

lav

as

Un

div

ided

Pli

oce

ne-

Ple

isto

cen

ev

olc

anic

s

uP

Pv

Kilo

met

res

02

1M

odif

ied

afte

rB

org

iaet

al.(1

988)

A4

Vo

lcan

Ch

ato sd

Vo

lcan

Are

nal

A3

uP

Pv

Rio

Are

nal

Lag

oA

ren

al

uP

Pv

A2

A1

E

AB

CD

A2

A4

A4

Cra

ter


11

Figure I-4: Geological map of Arenal volcano. After Borgia (1988).


11

sedimentary deposits (Figure I-4, Malavassi, 1979; Borgia et al., 1988). The oldest rocks

are from the Venado Formation, consisting of Miocene continental shelf deposits

(Malavassi and Madrigal, 1970; Obando, 1986). These are overlain by Miocene-

Pliocene deposits from the Aguacate Volcanic Group (Dengo, 1968), which are in turn

overlain by local Holocene alluvium deposits. These alluvium deposits are typically

found in the Lago Arenal area and along the margins of major rivers in the region

(Malavassi, 1979).

Volcanic Activity

Volcanic Activity Prior to 1968

Prior to 1965, no research had been conducted on Arenal volcano, and it was not

even mentioned in the Catalogue of Active Volcanoes of the World (Mooser et al., 1959).

However, based on tephrochronology and radiocarbon dating, Arenal previously has

erupted in 1750 ± 50 A.D., 1525 ± 20 A.D., 1080 ± 50 A.D., 220 ± 75 B.C., and 900 ±

150 B.C. (Simkin and Siebert, 1994). Arenal formed from the base of Chato’s edifice,

with pyroclastic and lava flows being deposited mainly to the northeast between the

Monterrey Hills to the north and Chato to the southeast. To the west, Lago Arenal was

formed by the damming effect of the volcanic deposits on the local drainage system.

Since the construction of a hydroelectric dam, the area of the lake has increased

dramatically to within 10 km of the summit of Arenal (Figure I-4).

Volcanic Activity During and After 1968

Premonitory manifestations of volcanic activity began in 1965 with the release of


12

colourless gas on the northeast flank of Arenal, the drying out of the Cedeño lagoon,

formation of a new hot spring, and an increase in the temperature of Rio Tabacón (Avila,

1978; Alvarado and Barquero, 1987; Barquero et al., 1992). On the morning of July 29,

1968, after 10 hours of intense seismic activity, Arenal erupted explosively and

continued to erupt over a period of three days, killing 78 people and devastating an area

12 square kilometres. The activity started with a lethal lateral blast that levelled the

densely forested western flank and destroyed the village of Pueblo Nuevo, 6 km west of

the summit. The initial blast was followed by plinian eruption columns, pyroclastic

flows, and ballistically ejected blocks and bombs (Figure I-5). Three new craters (A, B,

C) were formed during this time, with an approximately east-west orientation on the

western flank of the volcano (Figure I-4). The largest, crater C (1100 m), was the source

of all the major explosions (Melson and Saenz, 1973). These events were followed by

three days of relative calm consisting of minor ash and fumarolic activity. A fumarolic

phase began August 10 and continued to September 14. Activity was noted at all of the

craters, with the most intense activity at crater C. From September 14 to September 19,

renewed explosions consisting of low-energy, low-volume ejection of scoriaceous to

pumiceous andesite were observed (Melson and Saenz, 1973). This was followed by a

period of lava effusion from crater A (Sept. 19, 1968 to the end of 1973) consisting of

blocky lava which descended into the Quebrada Tabacón valley. Another explosive

phase started on June 17, 1975, with the emplacement of a blocky ash flow along the Rio

Tabacon valley. This deposit resulted from the formation of nuées ardentes produced by

avalanching of a lava flow being extruded from crater C (Figure I-4, Malavassi, 1979).


13

Figure I-5: Blocks from the blocky ash flow emplaced on the western flankof Arenal in July 1968. A 1992 blocky-aa lava flow is seen in thebackground.


14

This eruptive phase was followed by upslope migration of crater C to within close

proximity of crater D (Van der Bilt et al., 1976; Matumoto and Umana, 1976; Malavassi,

1979; Cigolini et al., 1984). Strong fumarolic activity and extrusion of blocky aa lava

followed, continuing until June 1984. This date marked an increase in activity at Arenal,

with the beginning of eruptions consisting of ash and large pyroclastics (Figure I-6).

This activity led to a Strombolian eruptive phase which produced basaltic-andesite

tephra and lava and continues at the time of the writing of this thesis (April 1997,

Figures I-4, I-7).


15

Figure I-6: White cloud on left of image represents a small pyroclastic flowcaused by the collapse of part of the blocky aa flow (a levee is in evidence tothe upper right of the cloud). View is of the western flank.

Figure I-7: A small strombolian eruption towards the western flank. CraterC is to the left, crater D to the right. View from the Arenal VolcanoObservatory, south of the volcano.


16

References

Alvarado, G. E., 1984. Aspectos petrologicos-geologicos de los volcanes y unidadeslavicas del Cenozoico Superior de Costa Rica. Thesis, Esc. Centroam. Geol.,Univ. Costa Rica, San Jose, Costa Rica, 183 pp.

Alvarado, G. E. and Barquero, R., 1987. Las señales sismicas del volcan Arenal(Costa Rica) y su relacion con las fases eruptivas (1968-1986). Cienc. Tec., 2:19-35.

Avila, G., 1978. Investigación y vigilancia del volcan Arenal, Alajuela, Costa Rica.Instituto Costaricense de Electricidad (ICE), Informe interno, San José, CostaRica, 40 pp.

Barquero, R., Alvarado, G. E., and Matumoto, T., 1992. Arenal Volcano (Costa Rica)premonitory seismicity. In: R. Gasparini, R. Scarpa, and K. Aki (Editors),Volcanic Seismology. Springer-Verlag, New York, pp. 84-96.

Bennett, F. D. and Raccichini, S., 1977. Las erupciones del Volcan Arenal, Costa Rica.Rev. Geograf. Amer. Cent., 5-6: 7-35.

Borgia, A., Casertano, L., and Cigolini, C., 1983. Estructura y dinámica de los flujosde lava del Arenal. Bol. Vulcanol., 14: 79-80.

Borgia, A., Poore, C., Carr, M. J., Melson, W. G., and Alvarado G.E., 1988.Structural, stratigraphic, and petrographic aspects of the Arenal-Chato volcanicsystem, Costa Rica: Evolution of a young stratovolcanic province. Bull.Volcanol., 50: 86-105.

Burbach, G. V., Frohlich, C., Pennington, W. D., and Matumoto, T., 1984.Seismicity and tectonics of the subducted Cocos plate. J. Geophys. Res., 89:7719-7735.

Brantley, S. L. and Koepenick, K. W., 1995. Measured carbon dioxide emissions fromOldoinyo Lengai and the skewed distribution of passive volcanic fluxes.Geology, 23: 933-936.

Carr, M. J., 1984. Symmetrical and segmented variation of physical and geochemicalcharacteristics of the Central American Volcanic Front. J. Volcanol. Geotherm.Res., 20: 231-252.

Carr, M. J., Rose Jr., W. I., and Mayfield, D. G., 1979. Relation of compositions tovolcano size and structure in El Salvador. J. Volcanol. Geotherm. Res., 5: 387-401.

Casadevall, T. J., Johnson, D. A., Harris, D. A., Rose Jr., W. I., Malinconico Jr., L.


17

L., Stoiber, R. E., Bornhorst, T. J., Williams, S. N., Woodruff, L., andThompson, J. M., 1981. SO2 emission rates at Mount St. Helens from March 29through December, 1980. In: P. W. Lipman and D. R. Mullineaux (Editors), The1980 Eruptions of Mount St. Helens. U.S. Geol. Surv. Prof. Pap., 1250, pp. 193-200.

Casadevall, T. J., Rose Jr., W. I., Fuller, W. H., Hunt, W. H., Hart, M. A., Moyers,J. L., Woods, D. C., Chuan, R. L., and Friend, J. P., 1984. Sulfur dioxide andparticles in quiescent volcanic plumes from Poás, Arenal, and Colima volcanoes,Costa Rica and Mexico. J. Geophys. Res., 89: 9633-9641.

Cigolini, C., Borgia, A., and Casertano, L., 1984. Intra-crater activity, aa-block lava,viscosity and flow dynamics: Arenal Volcano, Costa Rica. J. Volcanol.Geotherm. Res., 20: 155-176.

Dengo, G., 1968. Estructura geologica, historia tectonica y morfologia de la AmericaCentral. Centro Regional de Ayuda Tecnica (AID), Mexico DF, Mexico, 52 pp.

Dengo, G., 1962. Estudio Geológico de la Region de Guanacaste. Instituto GeograficoNacional, San Jose, Costa Rica, 112 pp.

Fudali, R. F. and Melson, W. G., 1972. Ejecta velocities, magma chamber pressure andkinetic energy associated with the eruption of Arenal volcano. Bull. Volcanol.,35: 383-401.

Guendel, F., McNally, K. C., Lower, J., Malavassi, E., and Saenz, R., 1984. Newevidence regarding subduction mechanisms near southern terminus of the MiddleAmerica Trench, Costa Rica. EOS Trans. Am. Geophys. Union, 65: 998.

Heiligmann, M., Stix, J., Williams-Jones, G., Sherwood Lollar, B., and Garzón V.,G., 1997. Distal degassing of radon and carbon dioxide on Galeras volcano,Colombia. J. Volcanol. Geotherm. Res., 77: 267-284.

Malavassi, E., 1979. Geology and petrology of Arenal Volcano, Costa Rica. M.Sc.Thesis, Department of Geology and Geophysics, University of Hawaii at Manoa,U.S.A., 111 pp.

Malavassi, V. and Madrigal, R., 1970. Reconocimiento geólogico de la zona norte deCosta Rica. Direccion Geologia, Minas y Petrolio, Informe Tecnico y NotaGeologica, Costa Rica, 9 (38): 12 pp.

Matumoto, T., Ohtake, M., Latham, G., and Umana, J., 1977. Crustal structure insouthern Central America. Bull. Seismol. Soc. Am., 67: 121-134.

Matumoto, T. and Umana, J. E., 1976. Informe sobre la erupción del volcan Arenaloccurida el 17 de Junio de 1975. Rev. Geofis. Inst. Panama Geogr. Hist., 5: 299-315.


18

Melson, W. G. and Saenz, R., 1968. The 1968 eruption of Arenal Volcano: preliminarysummary of field and laboratory studies. Smithsonian Center for Short-LivedPhenomena, Report 7-1968, 35 pp.

Melson, W. G. and Saenz, R., 1977. Las erupciones del Volcan Arenal, Costa Rica, enJulio de 1968. Rev. Geograf. Amer. Cent., 5-6: 55-148.

Melson, W. G. and Saenz, R., 1973. Volume, energy and cyclicity of eruptions ofArenal Volcano, Costa Rica. Bull. Volcanol., 37: 416-437.

Minakami, T., Utibori, S., and Hiraga, S., 1969. The 1968 eruption of VolcanoArenal, Costa Rica. Tokyo Univ. Earthquake Res. Inst. Bull., 47: 783-802.

Mooser, F., Meyer-Abich, H., and McBirney, A. R., 1959. Catalogue of the ActiveVolcanoes of the World including Solfatara Fields, Part 6, Central America.International Association of Volcanology, Napoli, Italy, 114 pp.

Mora, S., 1983. Una revisión y actualización de la clasificación morfotectónica de CostaRica. Bol. Vulcanol., 13: 18-36.

Obando, A. L. G., 1986. Estratigrafia de la formación Venado y rocas sobreyacentes(Micocebo Reciente) provincia de Alajuela, Costa Rica. Rev. Geograf. Amer.Cent., 5: 73-104.

Pichler, H. and Weyl, R., 1973. Petrochemical aspects of Central Americanmagmatism. Geol. Rund., 62: 357-396.

Reagan, M. K., Gill, J. B., Malavassi, E., and Garcia, M. O., 1987. Changes inmagma composition at Arenal Volcano, Costa Rica, 1968-1985: Real-timemonitoring of open-system differentiation. Bull. Volcanol., 49: 415-434.

Saenz, R., 1977. Erupción del volcan Arenal en 1968. Rev. Geograf. Amer. Cent., 5-6:149-188.

Simkin, T. and Siebert, L., 1994. Volcanoes of the World. Geoscience Press, Tucson,349 pp.

Stoiber, R. E. and Carr, M. J., 1973. Quaternary volcanic and tectonic segmentation ofCentral America. Bull. Volcanol., 37: 304-325.

Stoiber, R. E., Williams, S. N., and Huebert, B. J., 1986. Sulfur and halogen gases atMasaya caldera complex, Nicaragua: Total flux and variations with time. J.Geophys. Res., 91: 12215-12231.

Van der Bilt, H., Pangiagua, S., and Avila, G., 1976. Informe de la actividad delvolcan Arenal iniciada el 17 de Junio de 1975. Rev. Geofis. Inst. Panama Geogr.Hist., 5: 295-298.


19

Wadge, G., 1983. The magma budget of Volcán Arenal, Costa Rica from 1968 to 1980.J. Volcanol. Geotherm. Res., 19: 281-302.

Weyl, R., 1980. Geology of Central America. Gebrüder Borntraeger, Berlin, 371 pp.

Zapata, J. A., Calvache V., M. L., Cortés J., G. P., Fischer, T. P., Garzon V., G.,Gómez M., D., Narvaez M., L., Ordoñez V., M., Ortega E., A., Stix, J.,Torres C., R., and Williams, S. N., 1997. SO2 fluxes from Galeras Volcano,Colombia, 1989-1995: Progressive degassing and conduit obstruction of aDecade Volcano. J. Volcanol. Geotherm. Res., 77: 195-208.

CHAPTER I

DIFFUSE DEGASSING AT ARENAL VOLCANO,COSTA RICA

___________________

Chapter I - Diffuse Degassing at Arenal Volcano_____________________________________________

21

Abstract

Radon, CO2, and δ13C in soil gas have been measured at Arenal volcano, Costa

Rica. Rn and CO2 concentrations ranged from 1.2 to 69 pCi/l and from 0.01 to 9.62%,

respectively. Soil gases reach maxima near possible faults and on the lower flanks of the

volcanoes. The δ13C values, which varied between -10.7 ‰ and -30.8 ‰, were heaviest

on the lower slopes and close to possible fault lines. There are few surface-penetrating

structures on Arenal, as they are covered by young lavas. Arenal’s high level of activity

makes correlations between seismic activity and soil gas fluctuation difficult, if not

impossible. Rather, these fluctuations may be explained by variations in atmospheric

pressure. The trends of increasing soil gas concentrations and heavier isotope values

with distance from the summit suggest that (1) the soil gas concentrations are strongly

influenced by the level of development of the soil; (2) diffuse degassing of deep,

magmatic gas on the upper flanks of the volcanoes is negligible due to low permeability

from the cover of young volcanic rocks; and (3) increased magmatic degassing on the

lower flanks is the result of greater fracturing in the older lavas. Volcanoes such as

Arenal act as plugs in the continental crust, limiting degassing to fumaroles, faults, and

the fractured lower flanks.


22

Introduction

HE high level of activity at volcanoes such as Arenal and the general

inaccessibility of the crater area make the direct sampling of gas from crater

fumaroles highly problematic. By contrast, soil gases may be sampled safely at a certain

distance from the crater over extended periods of time. Many studies have shown the

association of soil gas variations and geologic activity (e.g., Thomas et al., 1986;

Baubron et al., 1991). Thus, in this chapter I present the results of a study of soil gas

CO2 and 222Rn concentrations and carbon isotope analyses of CO2 from Arenal volcano,

Costa Rica. The purpose of this chapter is to (1) evaluate the distribution of soil gas on

the flanks of Arenal, (2) correlate the observed spatial and temporal variations with

volcanic activity, and (3) investigate the origin and transport of these gases at Arenal.

222Radon is a radioactive noble gas with a half-life of 3.82 days which may be

emitted from any rock, soil or water that contains uranium or radium (Tilsley, 1992). It

is produced from the 238U decay series, with 226Ra being its immediate parental isotope.

Another isotope of radon, 220Rn, also referred to as thoron with a half-life of 55 s, is

produced from the Th decay chain. 226Ra, which is soluble in water, may be transported

significant distances before decay. This may result in the emplacement of large amounts

of radium near the surface, which can significantly increase 222Rn concentrations

(Tilsley, 1992). 222Rn also may be transported significant distances by water, by

advection and by mobile gases such as CO2 (Ozima and Podosek, 1983).

The importance of CO2 degassing on volcanoes was first recognised by

Carbonnelle and Zettwoog (1982) and Carbonnelle et al. (1985). After H2O, CO2 is the

T


23

most abundant species in volcanic gas. Its low solubility in silicate melts at low to

medium pressure allows for early exsolution, making CO2 a potentially good tracer gas

for the study of sub-surface magma degassing (Baubron et al., 1991). In the process of

diffuse magma degassing, the preferential removal of reactive gas species (e.g., SO2)

from the gas phase allows CO2 and inert gases (e.g., He) to reach the surface (Allard et

al., 1991). Volcanic CO2 has many different sources, among them the mantle, organic

processes, thermal metamorphism of carbonate rocks, and the mixing of gas from these

different sources (Irwin and Barnes, 1980).

Factors Affecting Soil Gas Concentrations and Distributions

Soil gas concentrations may be affected by processes that cause the development

of stress regimes in the ground or change the pore spaces and volume of cracks and

fissures. Processes such as (1) climatic variations (wind, rain, temperature, and soil

humidity), (2) atmospheric pressure variations, (3) deformation of a volcanic edifice, (4)

volcanic and volcanic seismic activity, and (5) tectonic seismic activity all may have

significant effects on soil gas concentrations (Heiligmann, 1997).

The nature and development of the soil at various elevations also may have a

significant effect on the radon, CO2 concentrations, and CO2 flux values. At higher

elevations, the soil often consists mainly of unconsolidated pyroclastic material, while

the soils become progressively more developed at lower elevations, i.e., the amount of

clay and organic material increases. This can affect concentrations and fluxes, since the

more developed soils are better able to retain moisture which leads to increased sealing

of the ground and the subsequent build-up of gas. The less consolidated material will


24

likely dry out faster after periods of precipitation, having efficiently removed the

humidity by percolation. This allows the gas to escape relatively easily. The more

organic-rich soils also may lead to increased concentrations of CO2 due to the bacterial

production of CO2 and decomposition of organic material (Hinkle, 1990).

Methodology

Station Locations

Due to Arenal’s relatively small size, most of the volcano is easily accessible by

car and by foot. The upper flanks, however, are generally inaccessible due to the

extremely high level of activity of the volcano. Four radial lines of 20 stations were

installed on the north, south, west and east flanks of the volcano, while another five

stations were installed around part of the base on the north and northeast sides. Labelled

for their geographic locations, the stations N-1 to N-5, S-1 to S-4, W-1 to W-5, E-1 to E-

5, and NE-1 to NE-5 are shown on Figure 1.1. The stations, which range in elevation

from 338 m to 855 m, were sampled for CO2 and 222Rn on a weekly basis over a period

of two months in 1995. The N, S, E, and W stations were sampled for carbon isotopes

twice during 1995, while the N, E, and NE lines were again sampled during the 1996

field season.

Soil Gas Measurements

222Rn was measured using the E-Perm technique developed by Rad-Elec Inc.

(Kotrappa et al., 1988; Kotrappa and Stieff, 1992) which consists of an electrostatically

charged Teflon disk (an electret) attached to an ion chamber of known volume. The disk


25

is placed at the bottom of a one metre long, 7.62 cm diameter PVC tube buried ~75 cm

in the ground for a period of approximately one week (Figure 1.2). The radioactive

decay of radon in the chamber ionises the air, producing negative ions which contact the

positively charged Teflon disk, resulting in a decrease of the electret voltage. Based on

the voltage drop, chamber volume, exposure time and ambient pressure, the

concentration of Rn may be calculated:

( )[ ] .RnV VCF T

Mi f=−

⋅

− ⋅0 120 (1.1)

where [Rn] is the concentration of radon in pCi/l; Vi and Vf are the initial and final

voltages, respectively; T is the period of exposure in days; 0.120 is a calibration constant

in units of pCi/l per µRad/hour; and M is the ambient gamma radiation in µR/h. For a

blue short-term electret in an L chamber, CF is a calibration factor calculated by:

CFV Vi f= + ⋅

+

0 2613 0 0001386

2. . (1.2)

where 0.2613 and 0.0001386 are calibration constants in units of pCi/l per µR/h (Rad

Elec Inc., 1993).

Instrumental error can be attributed to three sources. The first (%E1) is due to

imperfections in the instruments such as uncertainties in chamber volume and electret

thicknesses. This error has been experimentally measured at approximately 5%

(Kotrappa et al., 1990). The second source of error (%E2) is the uncertainty in voltage

measurement which amounts to an error of 1.4 volts between the initial and final

Chapter I - Diffuse Degassing at Arenal Volcano___________________________________________

25

Figure 1.1: Topographic map of Arenal volcano showing the location of Rnand CO soil gas stations (stars). The locations of seismic stations are shown

(red triangles). Contours ae every 100 m.2


27

Figure 1.2: PVC tube (left) used in measuring radon and an aluminium tube(right) used in the measurement of CO soil gas. Both tubes are buried ~75

cm.2


28

measurements (the square root of the sum of the squares of a 1 volt error per reading),

thus:

%.

EV Vi f

2100 14

=⋅

−

(1.3)

The third source of error (%E3) is due to the uncertainty in background gamma

radiation. These uncertainties have been estimated at 0.1-0.2 pCi/l; thus:

%.

[ ]E

Rn3

100 01=

⋅

(1.4)

The total error can be determined by taking the square root of the sum of the squares of

the three sources of error:

ET E E E= + +% % %1 2 32 2 2 (1.5)

The 226Radium emanating potential (RnERaC) of the ground was measured by

taking soil samples from the bottom (~75 cm) of the 222Rn holes. These samples were

kept in tightly sealed plastic bags in order to retain the soil humidity until they could be

analysed. In the laboratory, 20 to 30 g of soil were placed in a ceramic petri dish and

subsequently exposed to a short-term E-PERM electret in a sealed 3.74 L glass bottle for

a period of 11 days. The radon emanating potential of 226Ra then was calculated using

the following formula (Rad Elec Inc., 1993):

( )RnERaCRn M

Exp TT

=⋅

−− − ⋅

⋅

⋅

374

11 01813

01813

1000. ([ ] / )

( ( . )).

(1.6)

where 3.74 is the volume (L) of the glass jar; [Rn] is the 222Rn concentration (pCi/l); M

is the mass of soil (g) and T is the exposure time (days). The RnERaC, expressed in

pCi/kg, represents the ability of the soil to produce radon gas. Thus, rather than a


29

measure of 226Ra concentration, the RnERaC is a measure of radon-emitting 226Ra. This

radon is typically produced near the edges of the soil grains, escaping into the pore

spaces and subsequently travelling to the electret where it is measured.

For CO2 soil gas measurements, a metre-long, one centimetre-diameter

aluminium tube, with 5-6 small perforations cut along the bottom 10 cm, was placed

next to the PVC tube to the same depth. CO2 in the soil diffused into the tube, where it

was measured periodically by an infrared gas analyser (ADC LFG-20 Landfill Gas

Analyser). CO2 and CH4 were analysed by non-dispersive infrared absorption, while O2

was measured by an electrochemical cell. The ADC LFG-20 Landfill Gas Analyser has

measurement ranges for CO2 and CH4 of 0-10% and 10-100%, with corresponding

precisions of 0.5% and 3%, respectively. Only the first measurement range (0-10%) was

used due to the comparatively low concentrations of CO2 in the soil. O2 measurements

have a precision of ±0.4% on a scale of 0-25%. CO2 concentrations were corrected for

altitude using the following formula:

([CO2])·(Cf) (1.7)

where [CO2] is the volume percent CO2 measured at the site and Cf is the correction

factor calculated using:

Cf = 1 + (∆Elev.)⋅(2.678 x 10-4) (1.8)

where ∆Elev. is the difference in elevation between the station and the base station in

Fortuna (~250 m), and 2.678 x 10-4 is a factor based on the linear regression of

measurements of gas standards of known concentration taken at different elevations

(Heiligmann et al., 1997).

Diffuse CO2 flux was measured using a technique developed by Moore and


30

Roulet (1991). The technique consists of burying a 15 L chamber, with the chamber

opening towards the soil, ~75 cm in the ground. Tygon tubing pierces the top of the

chamber and allows for the measurement, by connection to the infrared gas analyser, of

CO2 concentrations over a period of 3-4 hours. Measurements were taken every 15

minutes for the first 1.5 hours and subsequently every 30 minutes for the remaining

period. By plotting CO2 concentration versus time and taking the slope of the initial

linear segment of the curve, a flux measurement (mg/m2·min) was calculated using the

following formula:

( ) ( ) ( ) ( ) ( )FluxCO

tppm

MM mg g V

AairCO air=

⋅ ⋅

⋅ ⋅ ⋅ ⋅

⋅−∆

∆2 610

11000

12

δ (1.9)

where ∆CO2 is the difference between initial and final CO2 concentrations (ppm); ∆t is

the elapsed time (min); Mair is the molecular weight of air (28.964 g/mole); MCO2 is the

molecular weight of CO2 (44 g/mole); V is the chamber volume (14.9 L); A is the area of

the chamber bottom (0.0519 m2); and δair is the density of air (g/L) calculated using

[(P/T)⋅(0.3483677)] where P is the pressure in mbar and T is the temperature in K.

Carbon isotope (13C/12C) analyses of soil gases were performed at the University

of Toronto using a Finnigan MAT 252 gas source mass spectrometer (MS) linked to a

Varian 3400 gas chromatograph (GC) equipped with a capillary column (GC-C-IRMS).

The interface of the GC and MS consists of a combustion oven containing Cu-Ni-Pt

trimetal. CO2 in the samples was separated by the GC at 27ºC. Three aliquots of each

sample were injected into the GC-C-IRMS system during δ13C analysis, with the average

δ13C for these three analyses reported. Accuracy and reproducibility of the isotopic

analyses are both < 0.1‰. δ13C analyses of the organic soil component were made of


31

samples from stations NE-1 to NE-5. The samples were initially sun-dried in the field

and transported in aluminium foil containers to the laboratory. They subsequently

underwent hand picking and removal of organic and lithic material. The soils were then

heated in an oven at 60°C for 2-3 days to remove any remaining humidity. They were

then crushed and decarbonated in 1 N HCl. The decarbonated samples were then rinsed

with deionised water and dried at 60°C. A subsample of 50-500 mg of soil was placed

in quartz tubes containing pellets of Cu and CuO catalyst. These were heated at 850°C

for 1 hour, followed by another hour at 600-650°C, and subsequently allowed to cool.

The resulting gas was cryogenically purified on a vacuum line and analysed on a

Finnigan MAT 252 mass spectrometer. All carbon isotope ratios are expressed as ‰

(per mil) difference from a Peedee Belemnite (PDB) standard.

Seismic Measurements

Seismic data of were collected in 1995 from a permanent seismographic station,

located 4 km east of summit, which continuously monitors the volcano (Figure 1.1).

This station has a short-period vertical seismometer (1 Hz) that is telemetered by

telephone line to the Observatorio Volcanológico y Sismológico de Costa Rica

(OVSICORI) of the Universidad Nacional in Heredia. Seismic data were limited in

coverage due to technical difficulties with the instrument in 1995.

Results

Radon

Observed 222Rn concentrations on Arenal range from 1.2 to 69 pCi/l with an error


32

from <1 to 4 pCi/l (Table 1.1). There is also a general inverse correlation between Rn

and altitude (Figure 1.3a). On the northern line, for example, station N-1 has an average

222Rn value of 6 pCi/l while station N-5, the furthest from the summit, has an average

value of 36 pCi/l (Figure 1.3b). The western and southern lines also have values which

increase outwards from the summit, with high stations showing low concentrations (e.g.,

W-1: 8 pCi/l; S-1: 8 pCi/l) and lower stations showing comparatively high values (e.g.,

W-5: 27 pCi/l; S-4: 32 pCi/l). The eastern line, however, is anomalous because it does

not show the tendency towards increasing radon with decreased elevation. Stations on

the northeastern line, which are the furthest from the crater and generally oriented

concentrically with the volcano, have the highest 222Rn values (e.g., NE-2: 62 pCi/l; NE-

5: 57 pCi/l).

Radon Emanating Potential

The 226Ra emanating potentials (RnERaC) for the 25 stations range from 60 to

288 pCi/kg. Generally, the RnERaC values are highest at stations with high radon (e.g.,

E-2, E-3 and E-4) and lowest at stations with low radon concentrations (e.g., W-1, N-1,

and N-3) (Table 1.2, Figure 1.4). The radon emanating potential also was measured for

crushed samples (35 mesh or 500 µm) of a 1968 pyroclastic debris flow and a 1992 lava

with a resulting RnERaC varying between 66 and 140 pCi/kg (Table 1.2).


33

Table 1.1 Radon soil gas from Arenal volcano for 1995-1996

Station Date Rn Error Station Date Rn Error

(d/m/y) (pCi/L) (pCi/L) (pCi/L) (pCi/L)

E-1 06/03/95 15 2 N-3 13/03/95 2

E-1 13/03/95 19 1 N-3 21/03/95 4

E-1 21/03/95 21 1 N-3 26/03/95 1

E-1 26/03/95 22 1 N-3 05/04/95 4

E-1 05/04/95 23 1 N-4 01/03/95 39 2

E-1 14/04/95 20 1 N-4 06/03/95 37 2

E-2 06/03/95 48 3 N-4 13/03/95 40 2

E-2 13/03/95 48 2 N-4 21/03/95 62 3

E-2 21/03/95 60 3 N-4 26/03/95 33 2

E-2 26/03/95 49 3 N-4 05/04/95 45 2

E-2 05/04/95 67 4 N-4 14/04/95 31 2

E-2 14/04/95 58 3 N-5 01/03/95 33 2

E-3 06/03/95 37 3 N-5 06/03/95 35 2

E-3 13/03/95 39 2 N-5 13/03/95 35 2

E-3 21/03/95 41 2 N-5 21/03/95 39 2

E-3 26/03/95 41 2 N-5 26/03/95 35 2

E-3 05/04/95 61 3 N-5 05/04/95 45 2

E-3 14/04/95 34 2 N-5 14/04/95 31 2

E-4 06/03/95 40 3 NE-1 09/03/96 17 1

E-4 13/03/95 34 2 NE-1 09/03/96 18 1

E-4 21/03/95 38 2 NE-2 09/03/96 57 3

E-4 26/03/95 29 2 NE-2 09/03/96 69 4

E-4 05/04/95 44 2 NE-3 04/03/96 30 2

E-4 14/04/95 32 2 NE-3 04/03/96 44 2

E-5 06/03/95 27 3 NE-4 08/03/96 48 2

E-5 13/03/95 20 1 NE-4 08/03/96 47 2

E-5 21/03/95 34 2 NE-5 08/03/96 56 3

E-5 26/03/95 30 2 NE-5 08/03/96 58 3

E-5 05/04/95 38 2 S-1 26/02/95 15 2

E-5 14/04/95 52 3 S-1 03/03/95 7 1

N-1 01/03/95 7 1 S-1 07/03/95 6 1

N-1 06/03/95 2 1 S-1 12/03/95 5 1

N-1 13/03/95 2 S-1 20/03/95 10 1

N-1 21/03/95 6 1 S-1 25/03/95 7 1

N-1 05/04/95 12 1 S-1 03/04/95 13 1

N-2 01/03/95 8 2 S-1 12/04/95 4

N-2 13/03/95 3 S-2 26/02/95 29 2

N-2 21/03/95 6 1 S-2 03/03/95 34 2

N-2 05/04/95 8 1 S-2 07/03/95 42 2

N-2 14/04/95 3 S-2 12/03/95 37 2

N-3 01/03/95 4 1 S-2 20/03/95 44 2

N-3 06/03/95 2 1 S-2 25/03/95 45 2


34

Table 1.1 Continued

Station Date Rn Error Station Date Rn Error

(pCi/L) (pCi/L) (pCi/L) (pCi/L)

S-2 04/04/95 47 2 W-4 12/03/95 5 1

S-2 12/04/95 49 3 W-4 20/03/95 7 1

S-3 26/02/95 23 2 W-4 25/03/95 7 1

S-3 03/03/95 17 1 W-4 04/04/95 11 1

S-3 07/03/95 19 1 W-4 12/04/95 54 3

S-3 12/03/95 18 1 W-5 26/02/95 24 2

S-3 20/03/95 20 1 W-5 04/03/95 23 1

S-3 25/03/95 18 1 W-5 07/03/95 25 2

S-3 04/04/95 21 1 W-5 12/03/95 23 1

S-3 12/04/95 15 1 W-5 20/03/95 33 2

S-4 26/02/95 28 2 W-5 25/03/95 23 1

S-4 03/03/95 28 2 W-5 03/04/95 26 1

S-4 07/03/95 34 2 W-5 12/04/95 40 2

S-4 12/03/95 31 2

S-4 20/03/95 32 2

S-4 25/03/95 32 2

S-4 04/04/95 38 2

S-4 12/04/95 35 2

W-1 26/02/95 10 1

W-1 03/03/95 7 1

W-1 07/03/95 11 1

W-1 12/03/95 4 1

W-1.5 04/04/95 10 1

W-2 26/02/95 11 1

W-2 03/03/95 7 1

W-2 07/03/95 7 1

W-2 12/03/95 6 1

W-2 20/03/95 9 1

W-2 25/03/95 6 1

W-2 12/04/95 6

W-3 26/02/95 11 1

W-3 03/03/95 7 1

W-3 07/03/95 4 1

W-3 12/03/95 5 1

W-3 20/03/95 5 1

W-3 25/03/95 4 1

W-3 04/04/95 8 1

W-3 12/04/95 4 1

W-4 26/02/95 9 1

W-4 03/03/95 7 1

W-4 07/03/95 6 1

Samples were taken at ~0.75 m depth

A

B

Average Rn vs Altitude

0

10

20

30

40

50

60

70

300 400 500 600 700 800 900

Altitude (m)

Rn

(pC

i\l)

E

N

NE

S

W


35

Figure 1.3: (a) Average Rn (pCi/l) versus elevation of stations (metres).There is an approximately negative linear trend. (B) Topographic map ofArenal volcano showing concentrations of Rn (pCi/l) soil gas. There is atendency towards increasing radon with distance from the summit.


36

Table 1.2 Radon emanating potential for soils and lava

/debris flow at Arenal volcano

Station Date Rn* RnERaC

(pCi/L) (pCi/kg)

E-1 29/01/96 0.60 196

E-2 25/06/96 0.73 237

E-3 08/08/96 0.80 263

E-4 25/06/96 0.88 288

E-5 29/01/96 0.62 203

N-1 08/08/96 0.23 74

N-2 25/06/96 0.45 123

N-3 25/06/96 0.36 119

N-4 29/01/96 0.61 192

N-5 29/01/96 0.64 205

NE-1 08/08/96 0.29 92

NE-2 08/08/96 0.31 101

NE-3 08/08/96 0.32 103

NE-4 08/08/96 0.19 60

NE-5 08/08/96 0.39 130

S-1 08/08/96 0.32 104

S-2 29/01/96 0.46 142

S-3 25/06/96 0.46 151

S-4 29/01/96 0.57 180

S-4 08/08/96 0.39 127

W-1 25/06/96 0.38 109

W-2 25/06/96 1.02 280

W-3 29/01/96 0.53 156

W-4 25/06/96 0.51 155

W-5 29/01/96 0.57 180

Lava/Debris flow (35 mesh)

1968a 04/09/96 0.24 71

1968b 04/09/96 0.29 83

1992b 04/09/96 0.23 66

1992c 23/09/96 0.42 140

*Radon concentrations are measured in the laboratory during

RnERaC measurements and are not related to field measurements.

Samples were taken at ~0.75 m depth


37

Figure 1.4: Radon (pCi/l, Table 1.4) versus radon emanating potential(RnERaC in pCi/kg, Table 1.2). Stations with elevated radon generally haveelevated RnERaC.

Rn

vs

Rn

ER

aC

0

10

20

30

40

50

60

70

05

01

00

15

02

00

25

03

00

Rn

ER

aC

(pC

i/k

g)

Rn(pCi/L)

E N NE

S W


38

Carbon Dioxide

CO2 concentrations vary from 0.01 to 9.6% (Table 1.3). On the northern line, for

example, station N-1 has an average CO2 concentration of 0.60% while station N-5, the

furthest from the summit, has an average 3.4% (Table 1.4). The western and southern

lines also have CO2 values which increase away from the summit, with high stations

showing low values (e.g., W-1: 0.04%; S-1: 0.96%) and stations on lower flanks

showing comparatively high values (e.g., W-5: 7.9%; S-4: 2.6%). There is a general

trend towards increasing concentration with decreased elevation (Figure 1.5a). Stations

on the northeastern line furthest from the crater have the highest CO2 concentrations,

with values ranging from 2.9 % at NE-2 to 9.6 % at NE-1. The stations on this line are

all approximately at the same elevation, with a difference of only 170 m between the

highest and lowest stations (Table 1.3, Figure 1.5b). A plot of average Rn versus

average CO2, shows a positive correlation for the N, S, and W lines (Figure 1.6). Lines

E and NE are anomalous, with the NE stations showing a strong negative correlation. If

one normalises radon by the RnERaC, in order to remove the effect of soil-produced Rn,

the negative correlation is degraded but nonetheless present (Figure 1.6b). By contrast,

line E shows a weak positive correlation.

Carbon Isotopes in CO2 Soil Gas

The δ13C values of CO2 vary from -10.7 to -30.8‰ (Table 1.5). There appears to

be a strong inverse relationship between elevation and δ13C (correlation coeff. r = -0.73,

Figure 1.7a).


39

Table 1.3 CO2 soil gas from Arenal volcano for 1995-1996

Station Date CO2 (%) Station Date CO2 (%)

E-1 06/03/95 2.45 N-2 06/03/95 0.69

E-1 13/03/95 2.63 N-2 13/03/95 0.73

E-1 21/03/95 2.81 N-2 21/03/95 0.82

E-1 26/03/95 2.85 N-2 26/03/95 0.81

E-1 05/04/95 3.48 N-2 05/04/95 0.96

E-1 14/04/95 3.08 N-2 14/04/95 0.73

E-1 04/03/96 2.29 N-2 04/03/96 0.49

E-2 06/03/95 2.10 N-3 01/03/95 1.68

E-2 13/03/95 2.06 N-3 06/03/95 1.66

E-2 21/03/95 2.13 N-3 13/03/95 1.68

E-2 26/03/95 2.06 N-3 21/03/95 1.95

E-2 05/04/95 2.44 N-3 26/03/95 1.86

E-2 14/04/95 2.06 N-3 05/04/95 2.44

E-2 04/03/96 2.22 N-3 14/04/95 1.51

E-3 06/03/95 6.15 N-3 04/03/96 1.17

E-3 13/03/95 5.98 N-4 01/03/95 6.35

E-3 21/03/95 6.17 N-4 06/03/95 6.39

E-3 26/03/95 5.70 N-4 13/03/95 6.52

E-3 05/04/95 7.06 N-4 21/03/95 7.38

E-3 14/04/95 6.76 N-4 26/03/95 7.54

E-3 04/03/96 7.25 N-4 05/04/95 7.99

E-4 06/03/95 4.62 N-4 14/04/95 6.77

E-4 13/03/95 4.25 N-4 04/03/96 8.27

E-4 21/03/95 4.34 N-5 01/03/95 3.40

E-4 26/03/95 3.73 N-5 06/03/95 3.26

E-4 05/04/95 5.26 N-5 13/03/95 2.96

E-4 14/04/95 4.40 N-5 21/03/95 3.37

E-4 04/03/96 5.44 N-5 26/03/95 3.41

E-5 06/03/95 7.03 N-5 05/04/95 4.05

E-5 13/03/95 6.83 N-5 14/04/95 3.14

E-5 21/03/95 7.19 NE-1 09/03/96 9.62

E-5 26/03/95 7.06 NE-2 09/03/96 2.86

E-5 05/04/95 8.17 NE-3 04/03/96 7.53

E-5 14/04/95 7.19 NE-4 08/03/96 5.57

E-5 04/03/96 7.64 NE-5 08/03/96 4.74

N-1 01/03/95 0.57 S-1 20/02/95 0.92

N-1 06/03/95 0.48 S-1 26/02/95 0.88

N-1 13/03/95 0.57 S-1 03/03/95 1.06

N-1 21/03/95 0.80 S-1 07/03/95 1.20

N-1 26/03/95 0.78 S-1 12/03/95 0.80

N-1 05/04/95 0.85 S-1 25/03/95 0.70

N-1 04/03/96 0.15 S-1 03/04/95 1.07

N-2 01/03/95 0.60 S-1 12/04/95 1.07


40

Table 1.3 Continued

Station Date CO2 (%) Station Date CO2 (%)

S-2 26/02/95 4.48 W-3 25/03/95 0.67

S-2 03/03/95 5.36 W-3 04/04/95 0.59

S-2 07/03/95 6.22 W-3 12/04/95 0.72

S-2 12/03/95 6.88 W-4 26/02/95 0.62

S-2 20/03/95 8.18 W-4 03/03/95 0.88

S-2 25/03/95 7.64 W-4 07/03/95 0.76

S-2 04/04/95 8.88 W-4 12/03/95 0.80

S-2 12/04/95 8.18 W-4 20/03/95 1.12

S-3 26/02/95 1.73 W-4 25/03/95 1.01

S-3 03/03/95 1.94 W-4 04/04/95 1.23

S-3 07/03/95 2.08 W-4 12/04/95 1.18

S-3 12/03/95 1.66 W-5 26/02/95 7.67

S-3 20/03/95 2.01 W-5 04/03/95 7.57

S-3 25/03/95 2.04 W-5 07/03/95 7.23

S-3 04/04/95 2.10 W-5 12/03/95 7.38

S-3 12/04/95 2.04 W-5 20/03/95 8.07

S-4 26/02/95 2.01 W-5 25/03/95 7.73

S-4 03/03/95 2.37 W-5 03/04/95 8.60

S-4 07/03/95 2.48 W-5 12/04/95 8.59

S-4 12/03/95 2.61

S-4 20/03/95 2.83

S-4 25/03/95 2.70

S-4 04/04/95 2.83

S-4 12/04/95 3.06

W-1 26/02/95 0.01

W-1 03/03/95 0.03

W-1 07/03/95 0.06

W-1 12/03/95 0.06

W-1.5 25/03/95 0.06

W-1.5 04/04/95 0.10

W-2 26/02/95 1.18

W-2 03/03/95 1.35

W-2 07/03/95 1.26

W-2 12/03/95 1.06

W-2 20/03/95 1.21

W-2 25/03/95 0.78

W-2 04/04/95 1.36

W-2 12/04/95 0.82

W-3 26/02/95 0.52

W-3 03/03/95 0.50

W-3 07/03/95 0.41

W-3 12/03/95 0.59

W-3 20/03/95 0.76

Error less than ± 0.05%. Samples were taken at ~0.75 m depth


41

Table 1.4 Average CO2, Radon and � 13C from Arenal

volcano for 1995 & 1996.

Station Altitude Avg. CO2 Avg. Rn Avg. � 13C

(m) (%) (pCi/L) (‰)

E-1 695 2.80 20 -25.23

E-2 600 2.15 55 -24.49

E-3 490 6.44 42 -19.76

E-4 405 4.58 36 -17.78

E-5 360 7.30 33 -14.00

N-1 855 0.60 6 -25.58

N-2 730 0.73 16 -26.10

N-3 655 1.75 3 -23.73

N-4 645 7.15 41 -14.18

N-5 505 3.37 36 -23.92

NE-1 338 9.62 18 -10.78

NE-2 360 2.86 63 -22.74

NE-3 360 7.53 37 -14.95

NE-4 508 5.57 47 -14.56

NE-5 500 4.74 57 -25.40

S-1 830 0.96 8 -30.39

S-2 740 6.98 41 -26.09

S-3 595 1.95 19 -25.73

S-4 730 2.61 32 -25.64

W-1 805 0.04 8

W-1.5 770 0.08 10

W-2 710 1.13 7 -25.32

W-3 645 0.61 6 -21.86

W-4 605 0.95 13 -20.92

W-5 575 7.85 27 -20.73

Average CO vs. Altitude2

0

1

2

3

4

5

6

7

8

9

10

A

B

300 400 500 600 700 800 900

Altitude (m)

CO

(%)

2

E

N

NE

S

W


42

Figure 1.5: (a) Average CO (%) versus elevation of stations (m). There is a

general trend of increasing concentration with decrease in elevation. (B).Topographic map of Arenal volcano showing the concentrations of

2

CO (%) soil

gas. There is a tendency towards increasing CO with distance from the summit.2

2

Average Rn vs. Average CO2

A

B

0

10

20

30

40

50

60

70

0 1 2 3 4 5 6 7 8 9 10

CO (%)2

Rn

(pC

i/l)

E

N

NE

S

W

E

N

NE

S

W

Rn/RnERaC vs. CO2

0

0.1

0.2

0.3

0.4

0.5

0.6

0 1 2 3 4 5 6 7 8 9 10

CO (%)2

Rn

/Rn

ER

aC


43

Figure 1.6: (a) Average Rn (pCi/l) versus average CO (%). Note the generally

positive correlation for the N, S and W lines. Stations from the NE and E lines areanomalous. (b) Average

2

Rn (pCi/l) normalised by RnERaC (pCi/kg) versus average CO

(%). Stations from the NE line are anomalous.2


44

Table 1.5 � 13C in CO2 soil gas from Arenal volcano for 1995-1996

Station Date CO2 � 13C Std. Dev. Station Date CO2 � 13

C Std. Dev.

(d/m/y) (%) (‰) (‰) (d/m/y) (%) (‰) (‰)

E-1 13/03/95 2.35 -25.56 0.06 S-4 12/03/95 2.31 -25.82 0.22

E-1 14/04/95 2.75 -25.43 0.04 S-4 12/04/95 2.71 -25.45 0.02

E-1 04/03/96 2.05 -24.69 0.08 W-2 12/03/95 0.94 -25.78 0.13

E-2 13/03/95 1.88 -24.53 0.08 W-2 12/04/95 0.73 -24.85 0.38

E-2 14/04/95 1.88 -24.61 0.03 W-3 12/03/95 0.53 -22.38 0.30

E-2 04/03/96 2.03 -24.33 0.05 W-3 12/04/95 0.65 -21.35 0.07

E-3 13/03/95 5.62 -19.70 0.06 W-4 12/03/95 0.73 -21.28 0.11

E-3 14/04/95 6.35 -19.46 0.08 W-4 12/04/95 1.08 -20.57 0.02

E-3 04/03/96 6.81 -20.11 0.09 W-5 12/03/95 6.79 -20.69 0.09

E-4 13/03/95 4.08 -18.26 0.03 W-5 12/04/95 7.90 -20.76 0.06

E-4 14/04/95 4.22 -17.51 0.07

E-4 04/03/96 5.22 -17.57 0.07

E-5 13/03/95 6.63 -13.79 0.06

E-5 14/04/95 6.98 -13.39 0.03

E-5 04/03/96 7.42 -14.82 0.04

N-1 13/03/95 0.49 -25.71 0.18

N-1 05/04/95 0.73 -25.46 0.22

N-2 13/03/95 0.65 -26.56 0.10

N-2 14/04/95 0.65 -25.61 0.11

N-2 04/03/96 0.43 -26.14 0.28

N-3 13/03/95 1.52 -24.05 0.10

N-3 14/04/95 1.36 -23.11 0.06

N-3 04/03/96 1.06 -24.05 0.22

N-4 13/03/95 5.90 -14.34 0.04

N-4 14/04/95 6.12 -14.31 0.04

N-4 04/03/96 7.48 -13.90 0.03

N-5 13/03/95 2.77 -23.83 0.20

N-5 14/04/95 2.94 -24.01 0.04

NE-1 09/03/96 9.40 -10.78 0.03

NE-2 09/03/96 2.78 -22.74 0.04

NE-3 04/03/96 7.31 -14.95 0.14

NE-4 08/03/96 5.21 -14.56 0.13

NE-5 08/03/96 4.44 -25.40 0.07

S-1 12/03/95 0.69 -30.78 0.11

S-1 12/04/95 0.93 -30.00 0.19

S-2 12/03/95 6.08 -26.08 0.09

S-2 12/04/95 7.23 -26.11 0.09

S-3 12/03/95 1.52 -25.71 0.09

S-3 12/04/95 1.87 -25.75 0.00

A

B

Average C vs. Altitude� 1 3

-35

-30

-25

-20

-15

-10

300 400 500 600 700 800 900

Altitude (m)

�� C

(‰)

E

N

NE

S

W


45

Figure 1.7: (a) Average C (per mil) in versus elevation of stations

(m). Note the negative correlation (r = -0.73). (b) Topographic map of Arenal

volcano showing the concentrations of

� 13CO soil gas

C (per mil) in CO soil gas Note the

C

2

2��

13

13

.

tendency towards heavier with distance from the summit.


46

There is also a good positive correlation (r = 0.76) between δ13C and CO2 concentrations

(Figure 1.8a). By contrast, only lines E and NE show a negative correlation the

correlation between δ13C and Rn (Figure 1.8b). This is also the case for the plot of δ13C

against Rn normalised by RnERaC, where line NE has a good negative correlation and

line E, a weak positive correlation (Figure 1.8c).

On the eastern line, station E-1 has an average δ13C value of -25.5‰ while

station E-5, the furthest from the summit, has an average -13.6‰ (Table 1.4, Figure

1.7b). The western and southern lines also have increased values outwards from the

summit, with high stations showing light values (e.g., W-2: -25.3‰; S-1: -30.4‰) and

stations on the lower flanks showing heavier values (e.g., W-5: -20.7‰). The northern

line also shows a trend of increasing δ13C with distance from the active crater, with the

lower stations having slightly heavier values (e.g., N-3: -23.7‰; N-5: -23.9‰) than the

higher stations (e.g., N-1: -25.6‰; N-2: -26.1‰). Station N-4, however, is anomalously

heavy with a δ13C value of -14.3‰. Radon and CO2 concentrations are also

anomalously high at N-4. Stations on the northeastern line are furthest from the crater

and have highly variable δ13C, with values ranging from -25.4‰ at NE-5 to -10.8‰ at

NE-1.

δ13C measurements were made of organic matter in soil samples from four of the

five stations on the northeastern line. Carbonated (i.e., carbonate-bearing) samples had

values ranging from -23.2‰ at NE-1 to -26.2 for NE-5. Decarbonated samples (i.e.,

carbonates were removed from the sample) showed very similar values, with -22.9‰ at

NE-1 to -26.3‰ for NE-5. There was an average difference of only -0.067‰ between

δ13C for the carbonated and decarbonated samples (Table 1.6). There also appears to be

� 13C vs. Rn/RnERaC

-26.00

-30.00

-22.00

-18.00

-14.00

-10.00

0 0.1 0.2 0.3 0.4 0.5 0.6

Rn/RnERaC

�13 C

(‰)

E

N

NE

S

W

�13 C

(‰) E

N

NE

S

W

Average C vs. Average Rn� 13

-35

-30

-25

-20

-15

-10

0 10 20 30 40 50 60 70

Rn (pCi/l)

�13 C

(‰)

E

N

NE

S

W

Average C vs. CO� 13

2

-35

-30

-25

-20

-15

-10

0 1 2 3 4 5 6 7 8 9 10

CO (%)2

A

B

C


47

Figure 1.8: (a) C (per mil) versus CO (%). Note the positive correlation. (b)

Average

� 13

2

� �13 13C (per mil) in (c) C (per

mil) in CO soil gas versus average Rn (pCi/l) normalised by RnERaC (pCi/kg). Stations

from the NE line are anomalous.2

CO soil gas versus average Rn (pCi/l). Average2


48

Table 1.6 Organic � 13C in soil from the NE line of Arenal volcano

Station Carbonated Std. Dev. Decarbonated Std. Dev. � (Carb-Decarb)

� 13C (‰) (‰) � 13

C (‰) (‰) (‰)

NE-1 -23.20 0.07 -22.90 0.01 -0.31

NE-2 -24.98 0.00 -24.84 0.07 -0.15

NE-3 -24.79 0.01 -24.88 0.06 0.09

NE-5 -26.23 0.07 -26.32 0.01 0.09

all values ± 0.2 ‰ vs. PDB

Table 1.7 CO2 flux from Arenal volcano in

1995

Station Date CO2 CO2 Flux

(d/m/y) (%) (mg/m2·min)

E-2 11/04/95 2.03 0.243

E-3 10/04/95 6.50 0.012

E-4 10/04/95 5.00 0.024

E-5 09/04/95 8.30 0.008

N-2 15/04/95 0.55 0.660

N-4 16/04/95 5.60 0.030

S-1 06/04/95 0.72 0.285

S-2 07/04/95 8.00 0.041

S-3 08/04/95 1.80 0.092

W-2 29/03/95 0.84 0.061

W-3 29/03/95 0.69 0.212

W-5 28/03/95 7.15 0.064

W-5 28/03/95 7.15 0.017


49

a correlation between δ13C in CO2 soil gas and soil organic δ13C (Figure 1.9). For

example, station NE-5 has both the lightest soil gas δ13C (-25.4‰) and lightest soil

organic δ13C (-26.2‰). Stations NE-2 and NE-3 have relatively intermediate values for

this line (e.g., NE-2 soil gas δ13C: -22.7‰, NE-2 organic δ13C: -24.9‰). Station NE-1

has the heaviest soil gas δ13C (-10.8‰) and heaviest soil organic δ13C (-23.2‰).

Carbon Dioxide Flux

CO2 flux values range from 0 to 0.66 mg/m2⋅min (Table 1.7). Due to time

constraints in the field, only partial flux coverage of the volcano is available. When CO2

concentration is plotted versus CO2 flux, two groups of values are apparent (Figure

1.10). The first group consists of stations comparatively high on the volcano (N-2, W-2,

W-3, S-1, S-3, E-2) with high and variable flux values ranging from 0.06 to 0.66

mg/m2⋅min. The second group consists of lower lying stations (E-3, E-4, E-5, N-4, S-2)

with low flux values ranging from 0.0008 to 0.064 mg/m2⋅min. The stations on the

upper part of the volcano also have lower CO2 concentrations (0.55 to 2.0%), lower Rn

and lighter δ13C, while stations at lower elevation have higher CO2 (5.0 to 8.3%), higher

Rn and heavier δ13C. It has been shown that the isotopic composition of soil CO2 is

strongly influenced by soil respiration rates (Cerling et al., 1991). This becomes

apparent if one plots δ13C versus CO2 flux. It is also apparent that there is a wide range

of δ13C (-26.1 to -14.2 ‰) at stations with low CO2 flux (0-0.01 mg/m2⋅min) The

stations with the heaviest δ13C have the lowest flux (Figure 1.10b).

The elevated CO2 at lower elevations may be due in part to the decomposition of

�13C

inso

ilvs.

�13C

inC

O2

-27

-26

-25

-24

-23

-22

-26

-24

-22

-20

-18

-16

-14

-12

-10

��C

inC

O(‰

)2

��Cinsoil(‰)

Car

bo

nat

ed

Dec

arb

on

ated


50

Figure 1.9: (a) C (per mil) .

Samples are for the NE line stations furthest from the summit.

� 13in CO soil gas versus organic C (per mil) in soil2 � 13

-35

-30

-25

-20

-15

-10

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

CO vs CO Flux2 2

C vs CO Flux� 13

2

CO

(%)

2C

(‰)

�13

0

1

2

3

4

5

6

7

8

9

A

B

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7

CO Flux (mg/m min)2

2

CO Flux (mg/m min)2

2

E

N

S

W

E

N

S

W


51

Figure 1.10: (a)

C (per mil)

.

� 13

CO (%) soil gas versus CO flux (mg/m min). Note the grouping of

points where samples with elevated CO have low CO flux. (b) CO

flux (mg/m min). Note that stations with low flux have heavier

2 2

2 2 2

2

2

versus

C� 13


52

organic material as well as to bacterial production (Hinkle, 1990). One would therefore

expect lighter carbon isotope values for the lower stations. The inverse is true, with the

lower stations having the heaviest values. The organic component can therefore not be

the main cause of the elevated CO2 levels.

Soil Gas Time Series

Radon and CO2 soil gas samples were plotted over time in order to investigate

fluctuations which may correlate with seismic or climatic variations.

Radon

Line E radon values are somewhat variable over time, with an overall increase

from March 7, 1995 to April 6, 1995 (e.g., E-2: 48 to 67 pCi/l). Except for E-5, the

values subsequently drop off (e.g., E-2: 58 pCi/l, Figure 1.11a). Line S radon values are

also quite variable, with only S-2 and S-4 showing a progressive increase in

concentration (e.g., S-4: 28 to 38 pCi/l) to a maximum on April 4 and subsequent decline

to 35 pCi/l on April 12, 1995 (Table 1.1, Figure 1.11b). Stations S-1 and S-3 do not

increase progressively, but rather fluctuate at fairly low levels. As with line E, stations

N-3 and N-4 radon values show a prominent peak at March 22 and a smaller peak on

April 6. Stations N-3 and N-4 also show the most variability of the northern line (Figure

1.11c). Radon concentrations at stations W-4 and W-5 show a marginal increase over

time. However, concentrations for stations W-1 through W-4 are approaching the

reliable limits of detection (~ 10 pCi/l), and thus any fluctuations may be instrumental in

nature (Figure 1.11d).


53

FIGURE 1.11

Plot of radon and CO2 concentrations versus time for the E, S, N, and W lines. Solid

lines and symbols are radon values, dashed and clear symbols are CO2 concentrations.

Note the common peaks on March 22 and April 6, 1995.


54

There are clearly some similarities in the fluctuation with time of radon values

amongst the four lines of stations. Lines E and S show the April 6 peak, while the peak

on March 22 is less evident. The inverse is true for the N line, where the April 6 peak is

weak or missing, while the March 22 peak is clearer. The W line is the least useful line

due to the comparatively low values for all stations except W-5.

Carbon Dioxide

As with radon, line E CO2 values show a small increase to April 6 and

subsequent decline (Table 1.3, Figure 1. 11a). CO2 values for S-1 and S-3 vary little,

whereas station S-2 shows a marked increase in concentration (e.g., 4.5% to 8.9%) on

April 4 and subsequent decline (e.g., 8.2%) on April 12, 1995 (Figure 1. 11b). Line N

CO2 values show little variation, although there is a small but visible peak on April 6

(Figure 1. 11c). Stations N-1 and N-2 are less clear due to their low concentrations.

Line W CO2 values are comparatively low (W-1: 0.01% to W-4: 1.2%) except for the

most distal station W-5, which shows a steady increase from 7.6% on February 2 to

8.6% on April 12, 1995 (Figure 1. 11d).

As with the radon time series, there are some similarities in the fluctuation of

CO2 concentrations between the four lines of stations. Line E, N, and station S-2 show

the April 6 peak clearly while that on March 22 is less evident. Little can be said for the

W line as the majority of the stations show comparatively low concentrations. All four

lines show very little fluctuation in the CO2 concentrations at the higher stations, which

may in part be due to their low concentrations.


55

Seismicity

Due to technical difficulties with the seismometers maintained by OVSICORI,

seismic data is incomplete for the sampling period in question, with data available only

from March 23, 1995 to April 16, 1995 (Table 1.8). Arenal’s high level of activity can

be seen from the number of daily eruptions which range from 4 to 26, with an average of

13 (Figure 1.12a). The highest number of eruptions (26) was recorded on April 16,

while the fewest eruptions (4) occurred on April 13. The total hours of daily tremor

varied from 0 to 21 hours, with an average of 12 hrs (Figure 1.12b). The maximum daily

tremor (21 hrs) was recorded on April 9, while no tremor was recorded on April 13. The

largest eruption for this period (amplitude: 123 digital units) occurred on April 10, while

the smallest eruption (amplitude: 4 digital units) occurred on March 29. The maximum

eruption duration of 210 s was noted for an eruption on April 15, while a minimum

duration of 14 s was recorded for an eruption on April 10 (W. Melson, personal

communication, 1995).

Although both the daily number of eruptions and daily tremor vary greatly over

the period of measurement, there may be a possible correlation with soil gas fluctuations.

For April 6, 1995, there was a relatively small number of eruptions and relatively

elevated tremor (Figure 1.12). This may be significant, as both radon and CO2 soil gases

on line E peak on this date (Figure 1.11a). The subsequent decline in soil gas

concentrations after April 6 could relate to the increase in the number of daily eruptions

from April 8 to April 11, 1995 (Figure 1.12a). As radon measurements are made on a


56

Table 1.8 Seismic data from March 23 to April 16, 1995 at Arenal

volcano.

Date Number of Hours of Avg. Amplitude Avg. Duration

Eruptions Tremor (digital units) (s)

23/03/95 5 5.00 52 38

24/03/95 7 19.00 39 34

25/03/95 8 12.00 30 30

26/03/95 18 13.75 32 34

27/03/95 16 10.50 23 35

28/03/95 15 8.50 30 33

29/03/95 22 9.50 25 31

30/03/95 10 13.00 32 30

31/03/95 17 9.00 18 35

01/04/95 13 14.50 21 30

02/04/95 9 15.50 21 24

03/04/95 12 17.50 28 26

04/04/95 20 9.50 37 32

05/04/95 13 14.50 51 34

06/04/95 9 14.50 41 30

07/04/95 10 20.50 33 34

08/04/95 7 16.50 36 33

09/04/95 15 21.00 39 32

10/04/95 18 3.25 41 34

11/04/95 23 11.00 34 34

12/04/95 18 2.50 20 30

13/04/95 4 0.00 10 23

14/04/95 8 15.00 10 37

15/04/95 11 13.00 40 61

16/04/95 26 13.00 27 54

Data from V. Barboza, Departamento de Sismologia, OVSICORI.

A

B

0

5

10

15

20

25

Date

Ho

urs

of

Tre

mo

rp

erd

ay

Eruptions per dayMarch 23 to April 16, 1995

Hours of Tremor per dayMarch 23 to April 16, 1995

0

5

10

15

20

25

302

3/0

3/9

5

24

/03

/95

25

/03

/95

26

/03

/95

27

/03

/95

28

/03

/95

29

/03

/95

30

/03

/95

31

/03

/95

01

/04

/95

02

/04

/95

03

/04

/95

04

/04

/95

05

/04

/95

06

/04

/95

07

/04

/95

08

/04

/95

09

/04

/95

10

/04

/95

11

/04

/95

12

/04

/95

13

/04

/95

14

/04

/95

15

/04

/95

16

/04

/95

23

/03

/95

24

/03

/95

25

/03

/95

26

/03

/95

27

/03

/95

28

/03

/95

29

/03

/95

30

/03

/95

31

/03

/95

01

/04

/95

02

/04

/95

03

/04

/95

04

/04

/95

05

/04

/95

06

/04

/95

07

/04

/95

08

/04

/95

09

/04

/95

10

/04

/95

11

/04

/95

12

/04

/95

13

/04

/95

14

/04

/95

15

/04

/95

16

/04

/95

Date

Nu

mb

ero

fE

rup

tio

ns


57

Figure 1.12: a) Number of volcanic eruptions per day and b) total hours of tremorper day between March 23 and April 16, 1995.


58

weekly basis, the average number of daily eruptions over the same period was taken in

order to facilitate comparisons. Thus for line E, which most clearly shows the April 6

peak, it becomes apparent that there is an increased number of eruptions for the period

before April 6 (Figure 1.13).

Atmospheric Variations

Over a two month period in 1995, atmospheric pressure measurements were

made twice daily at a base station in Fortuna (elev.: ~250 m). There was a gradual but

variable decrease in atmospheric pressure from 1018 mbar on February 22, 1995 to a low

of 1005 mbar on April 10, 1995, and subsequent increase to 1012 mbar on April 16,

1995 (Figure 1.14).

There should be only a minimal temperature effect, as there was very little

temperature variation during these two months and only limited variation in Costa Rica

in general. Due to Arenal’s distance from large populated centres, there are no rainfall

data available. However, the sampling period (March to April, 1995 and February-

March, 1996) was towards the end of the Costa Rican dry season.

Discussion

Origin of Radon and Carbon Dioxide

Radon is unlikely to originate from a magma chamber, since its half-life of 3.8

days is too short to allow for travel from even a shallow magma chamber or conduit.

Radon may escape diffusely through soil from a mean depth of only 2 m, making it

Rn

an

dN

um

ber

of

Eru

pti

on

sv

s.T

ime

Lin

eE

0

10

20

30

40

50

60

70

02

/03

/95

07

/03

/95

12

/03

/95

17

/03

/95

22

/03

/95

27

/03

/95

01

/04

/95

06

/04

/95

11

/04

/95

16

/04

/95

Da

te

Rn(pCi/L)

0246810

12

14

16

18

20

NumberofEruptions

E-1

E-2

E-3

E-4

E-5

Nu

mb

ero

fE

rup

tio

ns


59

Figure 1.13: Fluctuations in radon concentration (pCi/l) and number oferuptions versus time. Note the increase in average number of eruptionscoinciding with a peak in Rn concentration (April 6, 1995). The histogramrepresents the average number of eruptions between Rn measurements periods.


59

FIGURE 1.13

Fluctuations in radon concentration (pCi/l) and number of eruptions versus time. Note

the increase in average number of eruptions coinciding with a peak in Rn concentration

(April 6, 1995). The histogram represents the average number of eruptions between Rn

measurement periods.


60

difficult, if not impossible, to migrate from a conduit or chamber in such a short time

period (Graustein and Turekian, 1990; Appleby and Oldfield, 1992). Radon may also

travel significant distances by convection processes in hydrothermal systems or along

structures (Connor et al., 1996). However, there is little structure or hydrothermal

activity on Arenal, making convective transport of radon unlikely. Increased radon flux

may be found near structural weaknesses such as faults, since the distance over which

radon can travel may be enhanced by advective transport along a pressure gradient

(Tilsley, 1992). However, unlike larger and older edifices such as Poás and Galeras

(Charland et al., 1997; Heiligmann et al., 1997), there are few faults on Arenal that show

surface expression (Malavassi, 1979; Borgia et al., 1988). This is because they are

covered by recent lava flows, reducing the structural influence. The E line may,

however, be affected by a fault structures on the southern and eastern flanks of the

volcano (Figure 1.1). It is possible that this fault, mapped by Malavassi (1979) and

Borgia et al. (1988), runs between Arenal and Cerro Chato. Such a structure, should it

exist, might explain the anomalous behaviour of line E with respect to the other lines.

Other than the NE line, E stations have the highest radon and CO2 and heaviest δ13C on

the volcano.

On the rest of the volcano, high radon values are typically associated with soil

development at lower elevations, as shown by the high radon emanating potentials of the

soil (Figure 1.4). For example, station S-1 has an RnERaC of 104 pCi/kg and an average

radon concentration of 8 pCi/l, while S-4 has an RnERaC of 180 pCi/kg and an average

radon value of 32 pCi/l (Tables 1.1, 1.2). Similar trends are observed on the N line

Dail

lyA

tmosp

her

icP

ress

ure

vari

ati

on

s

1000

1005

1010

1015

1020

22/02/95

24/02/95

26/02/95

28/02/95

02/03/95

04/03/95

06/03/95

08/03/95

12/03/95

14/03/95

20/03/95

23/03/95

25/03/95

27/03/95

30/03/95

01/04/95

03/04/95

05/04/95

07/04/95

09/04/95

11/04/95

13/04/95

15/04/95

Date

Pressure(mbar)


61

Figure 1.14: Daily atmospheric pressure (mbar) fluctuations measured at theFortuna base station (~250 m). Note the relative minima on April 6, 1995.


62

where N-1 has an RnERaC of 74 pCi/kg and radon concentration of 6 pCi/l, whereas N-5

has an RnERaC of 205 pCi/kg and an average radon value of 36 pCi/l (Figure 1.4). Thus

the radon emanating potential of the soil may be partly responsible for the radon values

measured here.

However, the degree of soil development cannot explain the general correlation

between radon and carbon isotopes. Lines N, S, and W show a general tendency towards

increasing radon concentrations with heavier δ13C values (Figure 1.8b). The eastern line

is again anomalous, as is the northeastern line. These relations also are observed for a

plot of radon versus CO2, where lines E and NE are anomalous, while the remaining

lines show a general increase of CO2 concentrations with increasing radon (Figure 1.6a).

These observations suggest that variations in radon, CO2 and δ13C may be caused by a

similar mechanism.

Soil Gas, Soil Development, and Elevation

As stated above, levels of 222Rn, CO2, and δ13C generally appear to increase with

distance from the summit (Figures 1.3, 1.5, and 1.7). The apparent correlation with

altitude is likely due to the direct correlation between soil type with elevation. As has

been described previously, the upper flanks of the volcano consist mainly of

unconsolidated pyroclastic material while those on the lower flanks have more

developed clay-rich soil. Thus, although the highest station is only half-way up the flank

(855m) it nevertheless represents the soil type of the uppermost, inaccessible flanks

(>855 m) of the volcano.

Line W, which is situated in the devastated zone of the 1968 eruption, shows an


63

increase in 222Rn and CO2 concentrations and δ13C with distance from the crater.

Stations W-1 through W-4 have extremely low 222Rn and CO2 concentrations, which

may also be due to the fact that they are all situated in loose unconsolidated 1968

pyroclastic deposits (Figure 1.15). This may allow for the possibility of wind breathing

which occurs when wind increases the pressure on the ground, forcing air into the soil,

and diluting the soil gas. It is clearly the case for W-1 where CO2 concentrations are

essentially atmospheric. Only W-5, with a partially developed soil due to its proximity

to Lago Arenal, shows elevated levels of 222Rn and CO2.

Radon, carbon dioxide, and carbon isotopes for the southern line also appear to

follow this trend, in a general sense at least. There is, however, a polarity in the values,

with stations S-1 and S-3 having lower 222Rn and CO2 concentrations than S-2 and S-4

(Table 1.4). This again may be explained by the degree of soil development, i.e., the

degree of permeability and porosity of the substrate. Station S-1 is situated in

unconsolidated volcanic debris while S-3 is in a gravel-rich soil in close proximity to a

stream. S-2 and S-4, on the other hand, are situated in better-developed, organic and

clay-rich soils. As discussed above, this type of soil acts as a low permeability layer,

which allows for the concentration of radon and CO2 below the surface. The organic

material also may enhance CO2 output from bacterial and organic decay processes.

Similarly, the 222Rn and CO2 concentrations and δ13C for the N line exhibit the

highest values in the more distal stations N-3 to N-5. N-4 is clearly anomalous, with

high CO2 (7.2%), radon (41 pCi/l), and heavy δ13C (-14.0‰), implying the possible


64

Figure 1.15: View from the western flank of Arenal towards Lago Arenal. Largeboulders and blocks are part of the 1968 pyroclastic deposit. The remains of theonce dense jungle (a few trees) are visible.


65

presence of a structural weakness such as fault. This would facilitate the transport of

more 226Ra close to the surface and allow for higher radon concentrations. Similarly,

magmatic or deep CO2 could travel along the structure and concentrate near the surface.

A structural weakness below the surface may also explain the anomalous

behaviour of line E. Eastern stations have some of the highest radon and CO2

concentrations and heaviest δ13C on the volcano. The eastern side of the volcano has

been affected only slightly by recent volcanic activity and is consequently densely

forested. The soil has therefore had more time to develop. This also may allow for the

greater flow of soil gases from depth with a more magmatic component.

It is unclear, however, why the NE line behaves anomalously for plots of δ13C vs.

Rn and RnERaC. In both cases, the NE line shows a good negative correlation.

Although there is no obvious pattern or trend within the line, it should be noted that in

contrast to the four other lines, the NE line does not radiate down the flank of the

volcano. It is possible that, with the increased distance from the crater and subsequent

increase in fracturing, there may be significant local variations in levels of

fracturing/faulting which are not apparent on the surface.

Although stations high on the edifice have elevated CO2 fluxes, their CO2

concentrations are comparatively low (Figure 1.5). The inverse is true for stations on the

lower flanks. Again this also may be explained by pore closure in the soil due to

elevated humidity, as well as soil development and porosity/permeability of the

substrate. Poorly developed soils such as the unconsolidated pyroclastics of the western

flanks will be susceptible to wind breathing (D. Thomas, personal communication to M.

Heiligmann, 1996). Increased pressure on the ground, resulting from wind blowing


66

around the summit, may force air into the soil. This will create local circulation cells

and result in increased CO2 flux, while at the same time diluting the subsurface CO2. In

the unconsolidated pyroclastics of W-3, for example, there is elevated CO2 flux (0.21

mg/m2⋅min) but low CO2 concentration (0.69%). On the other hand, W-5 has a very low

flux (0.017 mg/m2⋅min) but comparatively high CO2 concentration (7.15%, Table 1.7,

Figure 1.10).

There is an apparent negative correlation for both CO2 and Rn concentrations

with elevation (r CO2: -0.66; Rn: -0.58, Figures 1.3a, 1.5a) . This may be explained by

the relationship between elevation and soil development, where the elevated stations

(except for the eastern line) are situated in unconsolidated volcanic debris while stations

lower on the volcano show better developed clay- and organic-bearing soils.

Allard et al. (1991) have suggested that Mt. Etna is covered by a dome of

magmatic CO2 from crater and diffuse flank degassing, especially concentrated near

faults and radial dykes. The heavy δ13C values (~-3.0‰ to ~0‰) along these structures

and at lower elevations are interpreted as due to the influence of marine carbonates

(Allard et al., 1991). However, the country rock surrounding Arenal is composed of

undivided Pliocene-Pleistocene volcanics with no surface exposure of carbonate rocks in

evidence (Borgia et al., 1988). Arenal δ13C values are highly negative near the summit,

with heavier values in the more distal areas of the volcano. CO2 and Rn concentrations

also are elevated in the more developed soils of the lower flanks. Only in these more

distal regions does there appear to be a significant magmatic component (e.g., NE-1:

δ13C = -10.8‰, Figure 1.7b). Thus, rather than a dome of magmatic CO2, it appears

more likely that the recent lavas of Arenal act as impermeable layers which plug


67

magmatic CO2.

Relationship to Pressure Change

As shown above, there is a gradual decrease in atmospheric pressure over time

(Figure 1.14). If this pressure variation is correlated with the variations of CO2 over

time, 68% of stations show a negative correlation (r > -0.5, Table 1.9). W-1, for

example, has a correlation coefficient of -0.95 with pressure, S-2 a -0.85 correlation

coefficient, and E-1 a -0.81 correlation. Line E also shows a good correlation between

CO2 and Rn versus time (r = 0.4 to 0.92), with most stations peaking on April 6, 1995.

Similar radon - pressure trends are visible on line E, with correlation coefficients (except

E-4) ranging from -0.74 to -0.53.

A possible explanation for these variations with pressure may be seen by again

looking at the degree of soil development. For example, as seen above, line S is

polarised with respect to 222Rn and CO2 concentrations, with S-2 and S-4 stations

showing higher soil gas values than S-1 and S-3 (Table 1.4). A similar polarisation

becomes apparent when looking at pressure correlations. Stations S-2 and S-4, situated

in well developed soils, show high correlations for CO2 (S-2: -0.85, S-4: -0.82) and Rn

with pressure (S-2: -0.93, S-4: -0.81). S-1 and S-3, on the other hand, are situated in

loose unconsolidated volcanics and gravel-rich soil and have quite low correlations for

CO2 (S-1: -0.06) and Rn (S-1: 0.37, S-3: 0.38) with pressure (Table 1.9). Thus, as the


68

Table 1.9 Correlation coefficients for Rn vs. CO2 and Rn and CO2

vs. pressure at Arenal volcano

Station Rn vs CO2 Rn vs P CO2 vs P

E-1 0.85 -0.74 -0.81

E-2 0.79 -0.53 -0.49

E-3 0.55 -0.57 -0.42

E-4 0.92 0.07 -0.14

E-5 0.42 -0.56 -0.66

N-1 0.78 -0.88 -0.76

N-2 0.13 -0.14 -0.70

N-3 0.48 0.09 -0.54

N-4 0.36 0.25 -0.79

N-5 0.82 -0.39 -0.63

S-1 0.37 -0.06

S-2 0.82 -0.93 -0.85

S-3 0.65 0.38 -0.73

S-4 0.78 -0.81 -0.82

W-1 0.30 0.10 -0.95

W-2 0.44 0.81 0.37

W-3 -0.59 0.68 -0.36

W-4 -0.72 -0.30 -0.73

W-5 -0.29 -0.26 -0.43


69

atmospheric pressure drops, the concentrations of radon and carbon dioxide in the better

developed soils of S-2 and S-4 increase. This suggests that elevated atmospheric

pressure may enhance the sealing capacity of these already impermeable soils and allow

for increased concentrations of soil gas beneath the surface. When the pressure drops,

this sealing effect is reduced and more soil gases are allowed to escape (Schery and

Petschek, 1983). On the other hand, there is little if any correlation with atmospheric

pressure for stations in the unconsolidated volcanics/gravels (Table 1.9). This may be

due to the fact that the soil gas concentrations are so low as to be barely affected by any

variation in atmospheric pressure and/or by the poor sealing of the unconsolidated

materials.

Relationship to Seismicity

Due to only a partial coverage by seismic data, there is very little that can be said

with respect to correlation with soil gas fluctuations over time. There may be a partial

correlation with radon fluctuations. The peak in radon concentrations seen on April 6,

1995, for the eastern line appears to coincide with an average increase in the number of

eruptions prior to April 6. This is followed by a decrease in radon concentrations, as

well as a decrease in the average number of eruptions (Figure 1.13). However, due to

the limited seismic coverage, this correlation should not be taken as being representative.

Furthermore, as has been discussed above, the eastern line is anomalous with respect to

the other lines on the volcano, and thus this correlation may not be characteristic of the

other stations.


70

Conclusions

This study of CO2 and radon soil gas at Arenal volcano suggests that soil gases

may allow for a better understanding of volcanic behaviour. Some important

conclusions from this work are the following:

1. Correlations between soil gas concentrations and seismic data are difficult, due in part

to limited seismic coverage, but more importantly due to the high level of activity at

Arenal.

2. Variations with time for radon and CO2 soil gases are due in large part to changes in

atmospheric pressure over time.

3. Rn and CO2 soil gases from the upper flanks of Arenal are unlikely to originate from

deep magmatic gases, but rather emanate from shallow surface sources, as the radon

half-life is too short and transport process too slow.

4. The diffuse soil gases are generally unable to penetrate the young lavas which cover

and seal the upper flanks of Arenal. Only on the lower flanks, where young lavas do not

crop out, is there significant gas flow from depth. This is shown by the increased CO2

concentrations and heavier δ13C values at distance from the crater.

5. The degree of soil development and the porosity/permeability of the substrate also

strongly influences, the concentrations of CO2 and radon soil gas at Arenal.

Unconsolidated volcanic soils on the upper flanks of the volcano have relatively low

RnERaC and subsequently low radon values. These soils are also more apt to rapidly

dissipate any precipitation, thus limiting sealing effects. This is in contrast to the more

clay- and organic-rich soils of the lower flanks, which retain humidity and increase

sealing. This results in a lower permeability in the better developed soils and permits the


71

accumulation of soil gas below the surface.

6. Arenal acts as a volcanic plug, sealing shallow levels of the continental crust and

limiting deep gas flux to faults and the fractured lower flanks of the volcano.


72

References

Allard, P., Carbonnelle, J., Dajlevic, D., Le Bronec, J., Morel, P., Robe, M. C.,Maurenas, J. M., Faivre-Pierret, R., and Martin, D., 1991. Eruptive anddiffuse emissions of CO2 from Mount Etna. Nature, 351: 387-391.

Appleby, P. G. and Oldfield, F., 1992. Application of lead-210 to sedimentationstudies. In: M. Ivanovich and R. S. Harmon (Editors), Uranium-seriesDisequilibrium: Applications to Earth, Marine, and Environmental Sciences.Clarendon Press, Oxford, pp. 731-778.

Baubron, J.-C., Allard, P., Sabroux, J.-C., Tedesco, D., and Toutain, J.-P., 1991.Soil gas emanations as precursory indicators of volcanic eruptions. J. Geol. Soc.London, 148: 571-576.

Borgia, A., Poore, C., Carr, M. J., Melson, W. G., and Alvarado G. E., 1988.Structural, stratigraphic, and petrographic aspects of the Arenal-Chato volcanicsystem, Costa Rica: Evolution of a young stratovolcanic province. Bull.Volcanol., 50: 86-105.

Carbonnelle, J., Dajlevic, D., Le Bronec, J., More, P., Obert, J.-C., and Zettwoog,P., 1985. Etna: composantes sommitales et pariétales des émissions de gazcarbonique. Bull. PIRSEV-CNRS

Carbonnelle, J. and Zettwoog, P., 1982. Local and scattered emissions from activevolcanoes: Methodology and latest results on Etna and Stromboli. Bull.PIRPSEV-CNRS

Cerling, T. E., Solomon, D. K., Quade, J., and Bowman, J. R., 1991. On the isotopiccomposition of carbon in soil carbon dioxide. Geochim. Cosmochim. Acta, 55:3403-3405.

Charland, A., Stix, J., Barquero, J., Fernández, E., Williams-Jones, G., Barboza,V., and Sherwood Lollar, B., 1997. Controls on diffuse degassing of radon andCO2 at Poás volcano, Costa Rica. Bull. Volcanol., in review.

Connor, C., Hill, B., LaFemina, P., Navarro, M., and Conway, M., 1996. Soil 222Rnduring the initial phase of the June-August 1995 eruption of Cerro Negro,Nicaragua. Journal of Volcanology and Geothermal Research, 73: 119-127.

Graustein, W. C. and Turekian, K., 1990. Radon fluxes from soils to the atmospheremeasured by 210Pb-226Ra disequilibrium in soils. J. Geophys. Res., 17: 841-844.

Heiligmann, M., 1997. Soil gases at Galeras volcano, Colombia, and their utility ineruption prediction. M.Sc. Thesis, Département de Géologie, Université deMontréal, Montréal, Canada, 114 pp.


73


Hinkle, M. E., 1990. Factors affecting concentrations of helium and carbon dioxide insoil gases. In: E. M. Durance (Editor), Geochemistry of Gaseous Elements andCompounds. Theophrastus Publications SA, Athens, pp. 421-447.

Irwin, P. W. and Barnes, I., 1980. Tectonic relations of carbon dioxide discharge andearthquakes. J. Geophys. Res., 85: 3115-3121.

Kotrappa, P., Dempsey, J. C., Hickey, J. R., and Stieff, L. R., 1988. An electretpassive environmental 222Rn monitor based on ionization measurement. HealthPhys., 54: 47-56.

Kotrappa, P. and Stieff, L. R., 1992. Elevation correction factors for E-PERM radonmonitors. Health Phys., 62: 82-86.

Kotrappa, P., Dempsey, J. C., Rasey, R. W., and Stieff, L. R., 1990. A practical E-Perm (electret passive environmental monitor) system for indoor radonmeasurement. Health Phys., 58: 461-467.

Malavassi, E., 1979. Geology and petrology of Arenal Volcano, Costa Rica. M.Sc.Thesis, Department of Geology and Geophysics, University of Hawaii at Manoa,U.S.A, 111 pp.

Moore, T. R. and Roulet, N. T., 1991. A comparison of dynamic and static chambersfor methane emission measurements from subarctic fens. Atmosphere-Oceans,29: 102-109.

Ozima, M. and Podosek, F. A., 1983. Noble Gas Geochemistry. Cambridge UniversityPress, Cambridge, Australia, 367 pp.

Rad Elec Inc., 1993. E-PERMR system manual. Rad Elec Inc., Virginia

Schery, S. D. and Petschek, A. G., 1983. Exhalation of radon and thoron: the questionof the effect of thermal gradients in soil. Earth Planet. Sci. Lett., 64: 56-60.

Thomas, D. M., Cuff, K. E., and Cox, M. E., 1986. The association between groundgas radon variations and geologic activity in Hawaii. J. Geophys. Res., 91:12186-12198.

Tilsley, J. E., 1992. Radon: Sources, hazards and control. Geosci. Can., 19: 163-167.

CHAPTER II

A MODEL OF DIFFUSE DEGASSING AT THREE SUBDUCTION-RELATEDVOLCANOES

GLYN WILLIAMS-JONES

AND

JOHN STIX

Département de géologie,Université de Montréal,

Montréal, Québec,Canada

Chapter II - Diffuse Degassing on Arenal, Poás, & Galeras_____________________________________________________

75

Abstract

Radon, CO2, and δ13C in soil gas have been measured at three active subduction-

related stratovolcanoes (Arenal and Poás, Costa Rica; Galeras, Colombia). Rn values

reach maxima only near fault zones, areas of seismic activity, and on the lower flanks of

the volcanoes. The heaviest δ13C values were found near fumaroles in the active craters,

close to faults, and on the lower slopes of the volcanoes. These observations suggest

that (1) major faults can channel deep gases to the surface only if the faults have surface

expression; (2) diffuse degassing of deep, magmatic gas on the upper flanks of the

volcanoes is negligible due to low permeability from the cover of young volcanic rocks;

and (3) increased magmatic degassing on the lower flanks is the result of greater

fracturing in the older lavas. These results are in contrast to findings for Mount Etna

where a broad dome of magmatic CO2 has been recognised over much of the edifice

(Allard et al., 1991). Volcanoes, such as those studied here, act as plugs in the

continental crust, limiting degassing to fumaroles, faults, and the fractured lower flanks.


76

Introduction

RENAL (1,657 m) is a ~5 km-diameter stratovolcano situated in Costa Rica

(10.463°N, 84.703°W) 90 km northwest of the capital San José and has been in

continuous activity since 1968, with frequent lava flows and strombolian-vulcanian

eruptions (VEI = 3, Figure 2.1). Poás (2708 m) is a stratovolcano ~14 km in diameter

located approximately 35 km northwest of San José (10.20°N, 84.233°W) and is

bordered by the Rio Desague and the Rio Toro faults (Figure 2.2). The central crater,

Laguna Caliente, has been active since the late 19th century. Galeras (4200 m) is a ~25

km-diameter stratovolcano in southern Colombia (1.22°N, 77.37°W). The volcanic

complex is intersected by the regional Romeral-Buesaca fault system which trends

northeast-southwest (Figure 2.3). Its most recent activity has been marked by explosive

eruptions in May 1989, lava dome emplacement in late 1991, and six vulcanian

eruptions in 1992-1993 (Stix et al., 1993).

Methodology

Diffuse degassing of Rn and CO2 was studied at 25 representative stations on

Arenal, 16 stations on Poás, and 30 stations on Galeras, between 1994 and 1996 (Table

2.1) (Heiligmann et al., 1997; Charland et al., 1997). Radon soil gas was sampled using

the E-PERM technique (Kotrappa et al., 1988), CO2 in soil gas was measured using an

ADC LFG-20 Landfill infrared gas analyser, and carbon stable isotopes were measured

in CO2, collected with 25 ml vials, by gas chromatograph combustion - isotope ratio

mass spectrometry (GC-C-IRMS). Rn data had average reproducibility of ~6%, CO2

A


77

FIGURE 2.1

Topographic map of Arenal volcano showing average (a) radon concentrations in pCi/l,

(b) CO2 concentrations in volume %, and (c) δ13C values expressed as ‰. Contour

interval of 100 m.


78

FIGURE 2.2

Topographic map of Poás volcano showing average (a) radon concentrations in pCi/l,

(b) CO2 concentrations in volume %, and (c) δ13C values expressed as ‰. Contour

interval of 500 m.


79

FIGURE 2.3

Topographic map of Galeras volcano showing average (a) radon concentrations in pCi/l,

CO2 concentrations in volume %, and (c) δ13C values expressed as ‰. Contour interval

of 400 m.


80


81

measurements had instrumental precisions of between 0.5% and 3% depending on

concentration, and δ13C reproducibility was better than 0.1‰.

Results

Radon values differed substantially among the three volcanoes, from 3 to 60

pCi/l for Arenal (Figure 2.1a), 7 to 1150 pCi/l for Poás (Figure 2.2a), and 10 to 1380

pCi/l for Galeras (Figure 2.3a). The values vary with the elevation and structure of the

volcanoes. At Arenal, radon concentrations increase towards the lower flanks (up to 60

pCi/l; no data is available for the crater area). Poás displays low values (7 to 130 pCi/l)

in the vicinity of the active crater and higher values on its flanks (360 pCi/l) and near

faults (60 to 1150 pCi/l). At Galeras, values also are high near faults and zones where

swarms of high-frequency earthquakes were recorded in 1995 and 1996 (520 to 1380

pCi/l). Fairly high radon concentrations also were observed near fumaroles on the outer

flanks of the active cone within the caldera (~330 pCi/l).

CO2 concentrations vary from 0.04 to 10.2% at Arenal (Figure 2.1b), <0.1 to

16% at Poás (Figure 2.2b), and 0.0 to 12.6% at Galeras (Figure 2.3b). On Arenal, the

concentrations are low on the upper flanks (0.04 to 2.8%) and higher on the lower flanks

(1.1 to 9.6%). At Poás, low CO2 values are found in the summit area (0.01 to 3.4%) and

higher values are found along faults (3.2 to 16%). This contrasts with Galeras where

CO2 concentrations are more variable and commonly higher on the volcano (up to

12.6%) than near faults (1.2 to 3.2%).

δ13C values range from -10.8 to -30.4‰ at Arenal (Figure 2.1c), -6.2 to -26.0‰

at Poás (Figure 2.2c), and -8.5 to -23.2‰ at Galeras (Figure 2.3c), respectively. On


82

Arenal, δ13C values are generally heaviest on the lower flanks (-10.8 to -25.7‰) and

lighter at higher elevations (-25.5 to -30.4‰). As with CO2 concentrations, δ13C values

at Poás are generally lower in the summit area (-19.5 to -24.7‰) and higher along faults

and on the lower flanks. The heaviest δ13C values are found near the Rio Toro fault zone

(-12.7 to -18.4‰) and at the Dome 2 station (-6.2‰) which is situated near fumaroles in

the active crater (Table 2.1) (Charland et al., 1997). The other summit stations at Poás

have δ13C values lighter than -19.5‰, suggesting that the magmatic component is absent

or insignificant. At Galeras δ13C in CO2 heavier than -15‰ is recorded only near active

crater fumaroles (e.g., Chavas: -7.9‰), faults (e.g., SHC: -15.1‰), areas of seismicity

(e.g., Sismo5: -8.5‰) (Heiligmann et al., 1997). All other areas on Galeras have no

significant deep CO2 component.

Discussion

Radon values vary by up to two orders of magnitude from the crater/summit area

to the active faults on Galeras and Poás. Although high radon concentrations on these

volcanoes are associated with faults and areas of seismic activity, the slow transport

velocities and short half-life of radon suggest that deep radon probably does not reach

the surface directly. Rather, increased concentrations are likely due to (1) faster

advective transport of radon produced near the surface, (2) greater availability of radon

at shallow levels due to superficial fracturing (Thomas et al., 1986), and (3) gases and

aqueous solutions rising through faults and depositing minerals rich in potassium and Rn

precursors near the surface (Nishimura and Katsura, 1990).

It has been suggested that Mt. Etna is covered by a dome of magmatic CO2 from


83

crater and diffuse flank degassing, especially concentrated near faults and radial dykes

(Allard et al., 1991). The heavy δ13C values along these structures and at lower

elevations are interpreted as due to the influence of marine carbonates superimposed on

magmatic δ13C values (Allard et al., 1991). However, our observations made on Arenal,

Poás and Galeras, indicate that there is no significant magmatic CO2 being diffusely

released on the flanks of these volcanoes. Instead, deep CO2 is released from the active

craters, in areas of seismic activity, and along fault zones that intersect the edifices and

have surface expression, i.e., are not covered by recent volcanic rocks. Recent work on

Oldoinyo Lengai (Brantley and Koepenick, 1995), however, shows that ~75% of the CO2

flux comes from 7 crater vents, with less than 2% of the total flux from the flanks

(Koepenick et al., 1996).

Using SO2 fluxes measured by COSPEC and appropriate CO2/SO2 ratios

(Galeras SO2 0.0042 x 1012 mol/yr (Zapata et al., 1997), molar CO2/SO2 2.9 (Fischer et

al., 1997); Poás SO2 0.00034 x 1012 mol/yr (Rowe et al., 1995), molar CO2/SO2 2.2

(Williams et al., 1992); Arenal SO2 0.000729 x 1012 mol/yr, molar CO2/SO2 4.2

(Williams et al., 1992)), we calculate crater CO2 fluxes of 0.0084 x 1012, 0.00075 x 1012,

and 0.0031 x 1012 mol/yr for Galeras, Poás, and Arenal, respectively. These values are

2-3 orders of magnitude smaller than at Mt. Etna (Allard et al., 1991; Brantley and

Koepenick, 1995).

Why does Arenal have significantly lower Rn values than Galeras and Poás? We

propose three possible explanations. (1) Arenal is less than 3000 years old (Borgia et al.,

1988), while Poás is ~1 Ma (Prosser and Carr, 1987) and Galeras is >1.1 Ma (Calvache

V. et al., 1997). Owing to its youth, Arenal is less fractured and faulted than Poás and


84

Galeras, resulting in less surface area for radon production at Arenal. (2) The older

fractured edifices of Galeras and Poás also tend to have more mature hydrothermal

systems (Fischer et al., 1997; Rowe et al., 1995) which may increase leaching of the

rock, thereby flushing radon into a carrier gas such as CO2 and promoting its transport to

the surface (Yoshikawa et al., 1990). Aqueous fluids also may help in transporting

radon-parent elements such as Ra and U. Significantly, soil gas near Chavas fumarole

on Galeras shows elevated levels of both Rn and magmatic CO2 (~330 pCi/l Rn, 16.2%

CO2, δ13C -7.9‰). (3) Large regional structures also may have an effect by controlling

surficial radon distributions. Central America is divided into seven tectonic segments

mirrored by different styles of volcanism (Stoiber and Carr, 1973). Arenal is situated

~80 km from the nearest segment break, while Poás lies ~40 km northwest of a segment

break running through the Irazu-Turrialba volcanic complexes. Poás also is associated

with major structures such as those manifested by its north-south alignment of cones and

the Rio Torro and Rio Sarapiqui faults. Galeras lies on a major tectonic break, the

Guairapungo Fracture, which runs northwest-southeast through the northern Andes, and

is intersected by the north-south trending Interandean Valley (Hall and Wood, 1985).

These major structures may thus facilitate the transport and remobilisation of the

parental isotope radium from wallrock and hydrothermal sources and result in elevated

Rn concentrations.


85

Conclusions

The three mechanisms discussed above all may affect flank degassing to variable

degrees. The radon and δ13C evidence presented here suggest that diffuse gases are

unable to penetrate young lavas on the flanks of these volcanoes. We propose that

volcanoes such as Arenal, Poás, and Galeras act as volcanic plugs which seal shallow

levels of the continental crust, limiting deep gas flux to fumaroles, faults, and fractured

lower flanks of the volcanoes. Paradoxically, volcanoes which are responsible for

significant output of magmatic gas through the central conduit also act as barriers to gas

flow on their upper flanks. We envisage a concentric zoning of gas flow: an inner zone

in the active crater where strong degassing occurs, an intermediate zone on the upper

flanks where gas flow is impeded, and an outer, fractured zone where deep gas can again

reach the surface.


86

References


Borgia, A., Poore, C., Carr, M. J., Melson, W. G., and Alvarado G.E., 1988.Structural, stratigraphic, and petrographic aspects of the Arenal-Chato volcanicsystem, Costa Rica: Evolution of a young stratovolcanic province. Bull.Volcanol., 50: 86-105.


Calvache V., M. L., Cortés J., G. P., and Williams, S. N., 1997. Stratigraphy andchronology of Galeras Volcanic Complex, Colombia. J. Volcanol. Geotherm.Res., 77: 5-20.

Charland, A., Stix, J., Barquero, J., Fernández, E., Williams-Jones, G., Barboza,V., and Sherwood Lollar, B., 1997. Controls on diffuse degassing of radon andCO2 at Poás volcano, Costa Rica. Bull. Volcanol., in review.

Fischer, T. P., Sturchio, N. C., Stix, J., Arehart, G. B., Counce, D., and Williams, S.N., 1997. The chemical and isotopic composition of fumarolic gases and springdischarges from Galeras Volcano, Colombia. J. Volcanol. Geotherm. Res., 77:229-254.

Hall, M. L. and Wood, C. A., 1985. Volcano-tectonic segmentation of the northernAndes. Geology, 13: 203-207.


Koepenick, K. W., Brantley, S. L., Thompson, J. M., Rowe, G. L., Nyblade, A. A.,and Moshy, C., 1996. Volatile emissions from the crater and flank of OldoinyoLengai volcano, Tanzania. J. Geophys. Res., 101: 13819-13830.

Kotrappa, P., Dempsey, J. C., Hickey, J. R., and Stieff, L. R., 1988. An electretpassive environmental 222Rn monitor based on ionization measurement. HealthPhys., 54: 47-56.

Nishimura, S. and Katsura, I., 1990. Radon in soil gas: Applications in explorationand earthquake prediction. In: E. M. Durance (Editor), Geochemistry of GaseousElements and Compounds. Theophrastus Publications SA, Athens, pp. 497-533.


87

Prosser, J. T. and Carr, M. J., 1987. Poás volcano, Costa Rica: Geology of the summitregion and spatial and temporal variations among the most recent lavas. J.Volcanol. Geotherm. Res., 33: 131-146.

Rowe Jr., G. L., Brantley, S. L., Fernández, J. F., and Borgia, A., 1995. Thechemical and hydrologic structure of Poás Volcano, Costa Rica. J. Volcanol.Geotherm. Res., 64: 233-267.

Stix, J., Zapata G., J. A., Calvache V., M., Cortes J., G. P., Gómez M., D., NarvaezM., L., Ordoñez V., M., Ortega E., A., Torres C., R., and Williams, S. N.,1993. A model of degassing at Galeras Volcano, Colombia, 1988-1993. Geology,21: 963-967.


Thomas, D. M., Cuff, K. E., and Cox, M. E., 1986. The association between groundgas radon variations and geologic activity in Hawaii. J. Geophys. Res., 91:12186-12198.

Williams, S. N., Schaefer, S. J., Calvache V., M. L., and Lopez, D., 1992. Globalcarbon dioxide emission to the atmosphere by volcanoes. Geochim. Cosmochim.Acta, 56: 1765-1770.

Yoshikawa, H., Endo, K., and Nakahara, H., 1990. 220Rn and 222Rn in volcanic gas.In: E. M. Durance (Eds.), Geochemistry of Gaseous Elements and Compounds.Theophrastus Publications SA, Athens, pp. 149-161.


CHAPTER III

A MODEL OF DEGASSING AND SEISMICITY AT ARENAL VOLCANO,COSTA RICA

_________________________

GLYN WILLIAMS-JONES

AND

JOHN STIX

Département de géologie,Université de Montréal,

Montréal, Québec,Canada

Chapter III - SO2 and Seismicity: A Degassing Model_________________________________________________

89

Abstract

Arenal volcano is the most active volcano in Costa Rica and has emitted a

minimum of 1.3 Mt of SO2 and an estimated 4.5 x 108 m3 of lava since its lethal

reactivation in July 1968. Gas emissions from the volcano have been both by passive

degassing and explosive eruptions, with passive degassing being dominant. Based on

COSPEC measurements made during 1982, 1995, and 1996, the average daily output is

128 ± 62 t/d SO2 emitted from Arenal. Arenal is extremely active and shows repeated

cycles of decreasing SO2 flux and tremor prior to eruptions. Following eruptions, SO2

flux and tremor levels increase. These fluctuations show a distinct correlation with

Earth tides, with decreased explosive activity and increased tremor coinciding with the

peak of high tide. The cyclic nature of explosive activity also may be caused by

corresponding fluctuations in the extrusion rate of lava. At high extrusion rates, the lava

of a non-explosive vent may overflow into an explosive vent, temporarily blocking the

conduit. Arenal is likely tapping a deep to mid crustal magma chamber and, unlike

many volcanoes, there is little difference between petrological (0.61 Mt since 1968) and

COSPEC SO2 estimates (1.3 Mt), suggesting that Arenal is being continuously supplied

by fresh magma. However, the open system is periodically blocked near the surface due

to crystallisation of magma in the conduit and/or minor variations in extrusion rate. This

results in the development of a seal, leading to the overpressurisation of the conduit and

eventual explosive destruction of the seal.


90

Introduction

RENAL is a 1,657 m high conical stratovolcano situated in northern Costa Rica

(10.463°N, 84.703°W), 90 km northwest of the capital San José (Figure 3.1). It

has been in continuous activity since July 29, 1968, when it reactivated with a plinian

explosive phase. Aa to blocky lava flows of basaltic andesite have been extruded nearly

continuously since September 19, 1968 (Cigolini et al., 1984). Arenal progressed into a

second major eruptive phase in June 1975, which included numerous Merapi- and

Soufrière-type nuées ardentes. The volcano entered an intense strombolian phase in

1984, with increased eruptions of ash, lapilli, and blocks which continue to the present

day. The frequency of eruptions was observed to be approximately 30 minutes in 1984

(Van der Laat and Carr, 1989), while a similar eruptive frequency was noted by us in

1995 and 1996. Activity at Arenal has been accompanied by gas emissions since 1968.

Current activity includes continued lava extrusion and numerous ash emissions, some

ascending to over 1 km above the active crater C (Figure 3.2), with small infrequent

pyroclastic flows travelling down the northwest flanks (Fernández et al., 1996a). Bombs

and blocks have been ejected ballistically to 1,100 m elevation (Fernández et al., 1996b).

Field observations indicate that crater C is likely divided into a least 2-3 vents from

which lava, gas and ash are emitted separately. The summit (crater C) continues to grow

at a rate of ~5 m/yr (Fernández et al., 1996a).

Arenal is the smallest but most active of seven historically active Costa Rican

volcanoes. The volcano, which has a volume of only 15 km3 (Carr, 1984), is situated

between two massifs, the Cordillera de Guanacaste (SE) and the Cordillera Central

A


91

FIGURE 3.1

Geographic map of Costa Rica showing the location of Arenal volcano.


92

FIGURE 3.2

Topographic map of Arenal volcano showing the locations of seismometer stations (red

triangle) and inclinometer stations (green house). Craters A and B are now buried by

lava flows emplaced since 1968. Contours are every 100 m.


93

(NW) which form the volcanic chain of the Costa Rican Arc (Stoiber and Carr, 1973). A

small truncated and likely extinct volcano, Cerro Chato, lies approximately three

kilometres south of Arenal (Figure 3.2). Arenal is most likely tapping a lower to mid-

crustal magma chamber possibly located at a discontinuity 22 km below the surface

(Matumoto et al., 1977; Wadge, 1983; Reagan et al., 1987).

Three stages of differing magma compositions at Arenal are believed to coincide

with variations in eruptive activity (Reagan et al., 1987). Stage-1 zoned magmas likely

resided in the magma chamber prior to the 1968 eruption. A new magma intruded into

the chamber in July 1968, resulting in the plinian eruption and ejection of the stage-1

magma. It subsequently mixed with the more mafic parts of stage 1 to produce stage-2

magmas. Stage-3 magmas (mid-1974 to present) are the product of continued mixing

and fractional crystallisation along the walls of the conduit and chamber. Each change in

stage appears to correlate with a variation in the cumulative volume of extruded material

(Reagan et al., 1987).

The extended duration and high level of activity of Arenal are enigmatic and

require further study. The periodic explosions and fluctuations in tremor and SO2 flux

raise numerous questions as to the eruptive mechanisms which control the volcano.

Arenal’s near-continuous extrusion of lava, in conjunction with explosive activity, also

raises the question of how magma travels to the surface. Due to the high level of activity

at the volcano and the inaccessibility of the crater area, we used remote sensing

techniques to study volcanic gases at Arenal. Ultraviolet correlation spectrometry

(COSPEC) has been used to study volcanoes since the early 1970’s and is ideal for the

measurement of SO2 at active volcanoes such as Arenal. In this article, we present


94

results from two field seasons on Arenal, in which 151 SO2 flux measurements were

made, currently the largest data set for this volcano. These data are compared with

seismological measurements which we made in 1996 in order to develop a degassing

model that can explain the cyclic variations in seismicity and SO2 flux seen at Arenal.

The problems of SO2 flux measurements also are discussed. COSPEC and petrological

estimates of the total SO2 emitted since 1968, as well as annual rates of CO2 and SO2

flux, are used to show the impact that young volcanoes such as Arenal have on the

troposphere.

Methodology

SO2 Flux

The concentration of SO2 in the volcanic plume was measured using a Plume

Tracker (1995) and a COSPEC (1996). Solar ultraviolet radiation passes through a

Cassegrain telescope, connected to the instrument, and is focused onto a diffraction

grating which separates the individual wavelengths (Figure 3.3). This radiation then

passes through a correlator disc that isolates the UV radiation values into wavelengths

where there is positive and negative absorption by SO2. This UV radiation is then

measured by a photomultiplier which converts the amount of radiation into voltage. The

ratio of these two sets of radiation (positively and negatively absorbed) is known when

there is no SO2 present and compared to the drop in voltage when SO2 is present. The

difference in the ratio is proportional to the amount of SO2 in the field of view of the

instrument. Calibrations are made by placing gas cells with known concentrations of


95

FIGURE 3.3

Plume Tracker and COSPEC ultraviolet spectrometers. The right-angle light tube of the

Plume Tracker is visible in the top picture, while the bottom picture shows the control

panel for the COSPEC IV. Black and red cables connect to an analogue chart recorder

and portable computer.


96

SO2 into the optical path of the instrument (Stoiber et al., 1983).

The instruments were connected to a portable computer and chart recorder and

transported beneath the volcanic plume, during which time the SO2 signal was digitally

recorded every 1-2 s (Figure 3.3). The instruments were driven below the column at an

approximately constant speed (e.g., 20 kph) and approximately perpendicular to the

column. A gas-cell calibration was made before and after each traverse. Comparatively

elevated SO2 concentrations in the plume necessitated the use of high-concentration

calibration cells (300 ppm⋅m and 339.2 ppm⋅m for Plume Tracker and COSPEC,

respectively, Table 3.1). The digital data were then processed and graphed using

commercial spreadsheet software. From the graph, the beginning and end of the plume

transect may be deduced, and consequently the flux may be calculated. A chart recorder

was used to gather analogue data as a backup and in instances where digital data were

lacking (e.g., during battery changes of the computer). Windspeed measurements were

typically made in the morning prior to the start of SO2 flux measurements (~1000 hrs

local time), at midday, and in the afternoon at the end of flux measurements (~1600 hrs).

There is a relatively steady easterly trade wind that blows over Arenal and westward out

over Lago Arenal. Thus, windspeeds were measured on the western flank of the volcano

at an elevation of ~550 m using a handheld anemometer. An individual traverse below

the gas plume was divided into segments in order to correct for the deviation from

perpendicularity of the traverse with respect to the column. The SO2 flux for each

segment was calculated and summed to determine the total SO2 flux for a given traverse.

The SO2 flux in metric tonnes per day (F) was calculated using the following equation:


97


98

F = (cosθ)·(dcol)·(νwind)·(0.00023)·([SO2]col) (3.1)

where θ (º) is the deviation from perpendicularity of the segment of road with respect to

the gas column; dcol is the width (m) of a particular segment determined using the vehicle

speed, time and distance travelled in the column; νwind is the average windspeed (ms-1)

measured at ground level 2-3 times per day; 0.00023 is a factor to convert ppm⋅m3⋅s-1

into metric tonnes per day; and [SO2]col is the path-length concentration of SO2 (ppm·m)

in the column. The concentration of SO2 was calculated from:

[ ] ( )SOPP

Ccol

avg

calcal2 =

⋅ (3.2)

where Pcal is the peak height of the calibration gas cell in arbitrary units; Pavg is the

average peak height for the segment; and Ccal is the concentration of the calibration gas

cell in ppm⋅m (Table 3.1).

SO2 Flux Errors

Variations in measured SO2 fluxes may be due, in part, to fluctuations in

windspeed and direction, changes in cloud cover, and change in sun angle, resulting in

variable amounts of solar ultraviolet radiation. The opacity of an eruptive plume also

varies due to changes in ash content which will increase the absorption of ultraviolet

radiation. Instrumental uncertainties include instrument calibration (± 2%), digital chart

reading error (± 2%), varying car speed (± 5%), and windspeed measurement (± 0-60%)

(Table 3.2, Casadevall et al., 1981; Stoiber et al., 1983).

The error in windspeed measurement is due to the fact that measurements were

made at the base of Arenal (elev.: ~550 m) and thus do not represent windspeeds at the


99


100

summit of the volcano where degassing is taking place. As there is approximately 1-1.5

km difference in altitude between the ground and the plume, the plume velocity may be

up to 1.5 to 3 times greater than that measured on the ground (Willett and Sanders,

1959). Windspeed measurements also were affected by instrumentation errors. Wind

measurements made in 1995 have an average standard deviation of ~30%, as the digital

anemometer that was used gave readings that varied continually with local wind gusts

(Table 3.3). The operator was thus forced to estimate the average range of speeds at the

time of measurement. The 1996 wind measurements used a fully mechanical

anemometer that allowed for the integration of windspeed over an interval of one

minute, thus eliminating the need for estimates on the part of the operator.

Consequently, the standard deviation was ~14%, or about half that of the previous year

(Table 3.3). Total average errors of ~31% and ~15% for the SO2 flux were calculated

for 1995 and 1996, respectively (Table 3.2). The errors are similar to those calculated by

Stoiber et al. (1983).

Generally, 10 to 15 traverses were made per day, with individual traverses lasting

about 20 minutes. Plume Tracker/COSPEC measurements were typically made to the

west and southwest of the volcano, at a distance of between 4 and 4.5 km from the

crater. Plume widths varied between 2 and 6 km but were typically 2 to 3 km. In order

to minimise the errors arising from variable amounts of ultraviolet radiation,

measurements were made only when the sun was at a relatively high angle, from ~0900

hrs until ~1600 hrs local time.


101


102

Seismicity

Seismic data were collected using a Personal Seismograph PS-1 having a

frequency range of 0.2 to 30 Hz. In the low-gain mode used in this study, the instrument

was able to detect a vertical acceleration of 0.93 µg and ground displacement of 231 nm

at 1 Hz. At 40 Hz, a vertical acceleration of 2.92 µg and ground displacement of 0.45

nm was detectable. The PS-1 was placed at ~750 m elevation on the western flank of the

volcano (Figure 3.2) near an inclinometer maintained by the Departemento de Geologia

of the Instituto Costaricense de Electricidad (ICE) (Figure 3.4). The seismometer was

buried approximately 50 cm below the surface to reduce the effect of wind, and was

connected to a portable computer for the collection of digital data. The seismic data later

were analysed using commercial software packages.

Results

SO2 Flux

Sulphur dioxide flux measurements were collected during the months of

February-April 1995 and February-March 1996 (Table 3.4). The 1995 data consisted of

11 days of Plume Tracker measurements, with average SO2 flux of 109 ± 61 tonnes/day

(Table 3.5). The overall SO2 flux varied between 47 ± 23 and 202 ± 107 t/d and

between 22 (a single measurement) and 193 ± 120 t/d when eruption-related SO2 was

excluded. Eruption-related SO2, which is the additional sulphur dioxide from an

explosive eruption, was distinguished from passive degassing of the volcano in order to

study trends that were not influenced by eruptions. A maximum explosive value for the

1995 field season of 367 t/d


103

FIGURE 3.4

Inclinometer station maintained by the Departemento de Geologia of the Instituto

Costaricense de Electricidad (ICE), located on the upper western flank of Arenal

volcano. A portable seismometer was buried approximately 50 cm below the surface, to

the right of the door of the green hut.


104


105


106


107

was measured on March 30, 1995, while a minimum of 18 t/d also was seen on the same

day.

The 1996 field season consisted of 6 days of COSPEC measurements, with an

average SO2 flux of 162 ± 58 t/d (Table 3.5). The overall SO2 flux varied between 110 ±

46 and 259 ± 83 t/d, and between 87 ± 17 and 186 ± 39 t/d when eruption-related SO2

was excluded. Standard deviations varied between 14 and 86 t/d. A maximum value for

the 1996 field season of 360 t/d was measured on March 8, while a minimum of 41 t/d

was seen on March 5 (Table 3.4). Maximum values for both years are quite similar, with

the average daily flux for 1996 (162 t/d) only slightly higher than the 1995 flux (109 t/d).

Similarly, the non-eruptive passive SO2 flux is also marginally higher in 1996 (128 t/d)

than in 1995 (95 t/d), however any differences are well within the error of the

measurements. However, if one takes into account the fact that windspeed at the plume

height may be as much as 3 times greater than at ground level, a maximum average value

of 327 t/d and 486 t/d is obtained for 1995 and 1996, respectively. Maximum non-

eruptive passive SO2 flux was therefore 285 t/d (1995) and 384 t/d (1996). The 1995-

1996 data are similar to the 8 measurements made by Casadevall et al. (1984) in 1982

which had average SO2 flux of 198 ± 41 t/d and are 2-3 times greater than the flux of

~50 t/d measured by Stoiber et al. (1982) in 1982 (Tables 3.4, 3.5).

Due to the high level and frequency of eruptive activity at Arenal, it is difficult to

obtain the statistical data necessary to prove systematic decreases in SO2 levels prior to

an eruption. There are, however, some instances where SO2 levels appear to decrease

progressively prior to an explosive eruption. On February 28, 1995, SO2 flux dropped

from 96 ± 30 t/d to 60 ± 19 t/d before an eruption at 1025 hrs local time. The flux


108

subsequently rose to 203 ± 63 t/d immediately after the eruption (Figure 3.5a). Although

this flux clearly contains an eruptive component, the marked contrast before and after the

eruption is nevertheless significant as it suggests that sealing of the conduit is taking

place. On March 4, SO2 flux decreased from 147 ± 46 t/d prior to an eruption at 1243

hrs to 34± 11 t/d after the eruption. In the afternoon of the same day, SO2 fluxes

dropped from 124 ± 38 t/d immediately after an eruption at 1444 hrs to 59 ± 18 t/d

(Figure 3.5b). On March 5, SO2 flux increased from 55 ± 17 t/d before an eruption at

1204 hrs to 114 ± 35 t/d after the eruption. In the same afternoon, SO2 rose to 152 ± 47

t/d from 40 ± 12 t/d (Figure 3.5c). No eruption was noted at that time, but the increase

in SO2 flux may be explained by a discrete gas release that went unrecorded.

Similar fluctuations in SO2 flux were seen during the 1996 field season. In the

afternoon of March 5, for example, SO2 flux decreased from 117 ± 17 t/d to 41 ± 6 t/d

prior to an eruption at 1417 hrs. The flux subsequently increased to 103 ± 15 t/d after

the eruption (Figure 3.6e,f). March 8 also showed these cyclical fluctuations. SO2 was

seen to drop from 360 ± 54 t/d (likely from an unrecorded eruption) to 114 ± 17 t/d

before an eruption at 1219 hrs and subsequently climbed to 242 ± 36 t/d immediately

afterwards (Figure 3.6i,j). Although more data is necessary to fully characterise the

eruptive nature of the volcano, these repetitive fluctuations raise the possibility that

Arenal may undergo cyclical opening and closing of the conduit, resulting in variable

pressurisation and leading eventually to explosive eruptions.


109

FIGURE 3.5

SO2 flux versus time for a) February 28, b) March 4, c) March 5, 1995 at Arenal.

Eruptions are shown by inverted arrows.


110

FIGURE 3.6

SO2 flux and eruption amplitude and SO2 flux and eruption duration versus time for a,b)

February 29; c,d) March 1; e,f) March 5; g,h) March 6; i,j) March 8, 1996. Amplitude is

in digital units, duration in seconds, and SO2 flux in metric tonnes per day.


111

SO2 Budgets at Arenal

COSPEC/Plume Tracker Estimates

An estimate for the total SO2 emission from Arenal may be made by taking the

average of measured SO2 flux and extrapolating back to 1968. There is very little

published SO2 flux data for Arenal, with only 8 airborne COSPEC measurements made

in February 1982 (Casadevall et al., 1984) and some ground-based measurements in

November, 1982 (Stoiber et al., 1982). At the time of these measurements, the gas

originated mainly from fumaroles near crater C (Cheminée et al., 1981; Casadevall et al.,

1984). 151 measurements were made by us at Arenal between 1995 and 1996, resulting

in an average of 128 ± 59 t/d of SO2 gas, likely originating from crater C (Table 3.5). If

one takes a weighted average of the three data sets, a mean daily flux of 128 ± 62 t/d SO2

is calculated. This leads to an estimate of ~1.31 Mt SO2 emitted since 1968, which is a

lower limit for the following reasons: (1) significant quantities of SO2 likely were

emitted explosively during the initial 1968 eruptive episode; (2) our windspeeds are

clearly minimums due to the difference in altitude between the column height and place

of measurement; (3) these data also neglect sulphur released in the form of H2S.

Petrological Estimates

Estimates of total SO2 released also were made using melt inclusion data from

samples of the 1968 surge deposit and a 1992 lava flow (Table 3.6). We assumed that

the non-degassed pre-eruption melt was the only source of sulphur, and that melt

inclusions trapped in plagioclase and pyroxene crystals represent the non-degassed


112


113

sulphur content of the magma. The coexisting matrix glass was assumed to represent

sulphur contents of the degassed melt after eruption. The difference, which is the

quantity of SO2 released, can be determined from the difference in the sulphur contents

of the melt inclusions and matrix glass. This petrological SO2 emission (ESO2 in Mt) can

be calculated for Arenal basaltic andesites using the following equation (Gerlach and

McGee, 1994):

ESO2 = (2 x 10-15) ⋅∆Sm⋅ ρm⋅ φm ⋅V (3.3)

where the constant is a conversion factor for S (ppm) into SO2 (Mt); ∆Sm is the S lost

from the melt during eruption in ppm, determined by the difference (332 ppm) in mean S

contents of 7 melt inclusions (356 ppm) and 8 matrix glasses (<24 ppm); ρm is the

basaltic andesite melt density, assumed to be 2700 kg⋅m-3; φm is the melt volume fraction

of 0.5, estimated from thin sections of 1968, 1984 and 1992 surge deposit/lava; and V is

the volume of magma extruded between 1968 and 1996. As there is currently no

available data for extrusion rates after 1985, a rough estimate of the total volume of lava

extruded to date may be obtained by assuming that the current rate of lava extrusion is

similar to that between 1973 and 1985 (9.3 x 106 m3yr-1; Wadge, 1983; Reagan et al.,

1987). Thus, by adding the volume of lava extruded from 1968 to 1985 (3.5 x 108 m3) to

the volume of lava extruded from 1985 to 1996, (9.3 x 106 m3yr-1 times 11 years = 1.0 x

108 m3), we estimate that a total lava volume of 4.5 x 108 m3 has been extruded since

1968. The resulting ESO2 is approximately 0.40 Mt of SO2, approximately 3.3 times less

that the 1.31 Mt estimated using COSPEC data.. If one assumes that the maximum

sulphur in melt inclusions (avg.: 671 ppm S; px92-15d, px92-15e, Table 3.6) represents

samples from least degassed magma, the petrologically estimated mass of SO2 emitted


114

since 1968 is 0.815 Mt. This is only 1.6 times less than the mass estimated from

COSPEC data. Even using maximum windspeeds, three times that of ground

measurements, a total of 3.92 Mt is obtained, still only 4.8 times less than petrological

estimates.

Using SO2 fluxes measured by COSPEC and appropriate CO2/SO2 ratios (Arenal

SO2 128 t/d, CO2/SO2 6.4, Williams et al. (1992)), we calculate crater CO2 fluxes of 820

t/d for Arenal. This flux is 2-3 orders of magnitude smaller than at Mt. Etna (Allard et

al., 1991; Brantley and Koepenick, 1995).

Seismic Data

Seismic measurements were made in conjunction with COSPEC measurements

during 6 days in 1996. A total of 63 eruptions were measured over the 6 day period

between ~1000 hrs and ~1600 hrs local time. The durations of these events varied

between ~5 s and ~90 s, while amplitudes ranged from 120 to 4400 digital units.

Numerous periods of tremor also were recorded between eruptions, with durations

ranging from 14 s up to a period of continuous tremor lasting >6600 s on March 2, 1996.

FFT analyses of the tremor signal revealed that frequencies typically varied between 2

and 5 Hz. Harmonic tremor with frequencies of ~4.5 Hz were also common. These

tremors fall in the class of intermediate frequency tremor which is believed to be

associated with strong degassing. Low frequency tremor at <3 Hz also has been noted

and may represent conduit resonance, gas fluctuation, or degassing (Barquero et al.,

1992).

The relationship between tremor and eruptive events varies somewhat over the


115

period of measurement; nevertheless, some interesting correlations are apparent. A

relative decrease in seismicity was seen before eruptions on February 29 (1450, 1537 hrs

local time, Figure 3.6a,b) and March 1, 1996 (1520, 1529 hrs, Figure 3.6c,d). Many

eruptions also were followed by tremor events on February 29 (1043, 1257, 1342, 1450

hrs).

Chugs are locomotive-like sounds caused by repetitive gas emissions (Melson,

1989). Such chugs frequently were accompanied by harmonic tremor, likely indicating

the presence of an open conduit (Barboza and Melson, 1990). We also have noted this

correlation, frequently in the early part of a day. For example, on March 2, a general

decrease or dissipation of both chugs and tremor was observed prior to eruptions at

1129, 1159 and 1214 hrs. This was also the case on February 29 (1450, 1537 hrs) and

March 1 (1520 and 1529 hrs). March 1 was also noted for the comparatively low level

of explosions and the comparatively high number of chugs and tremor events (Figure

3.6c,d).

We have noted that there was a higher frequency of eruptions in the mornings

compared to the afternoons. This also was remarked upon briefly by Barboza and

Melson (1990). For example, on March 5, there were 11 eruptions between 1000 and

1300 hrs, yet only 3 between 1300 and 1650 hrs. The morning eruptions on March 5

also were characterised by pairs of eruptions (8 of the 11 morning eruptions) with less

than 10 minutes separating them (Figure 3.6e,f). The amplitudes of the morning

eruptions (except for March 6) also were greater and more variable than those of the

afternoon. The subsequent decrease in afternoon eruption occurrences and amplitudes

generally coincided with the peak in predicted tidal gravity values. The predicted tidal


116

gravity showed an increase from 83.6 µgals at 0930 on February 29 to 165.7 µgals at

1230 on March 6 (Table 3.7, Figures 3.6, 3.7). Maximum gravity values then decreased

to 155.6 µgals at 1330 on March 8, the end of field measurements. It should be noted

that the gravity values were calculated on the half hour.

This variation coincides with changes in the presence and nature of tremor

(Barboza and Melson, 1990). Prior to ~1326 hrs on March 5, there was very little

chugging or tremor activity; however, after this point significant tremor commenced and

lasted throughout the afternoon until the cessation of measurements at ~1600 hrs.

Similarly, on March 6 and 8, tremor started after ~1328 and ~1357 hrs, respectively, and

was accompanied by a decreased frequency of eruptions. The afternoon tremor on

March 5 and 6 also began just after the onset of high tide (Figure 3.7f,g), whereas the

tremor of March 8, began immediately after the peak of high tide (Figure 3.7i).

The presence of small-amplitude, high-frequency (15-17 Hz) events prior to

eruptions was observed on March 6. This coincided with the highest daily SO2 average

(259 ± 83 t/d) of the field season. Excluding the eruption-related SO2 flux, March 6 still

has the second highest daily value of 162 t/d (single value). Such high-frequency events

may be related to rock fracturing (Anderson, 1978). The high-frequency events and high

SO2 fluxes together suggest that there may have been increased intrusion of

comparatively gas-rich magma into the conduit.


117


118

FIGURE 3.7

Fluctuation of gravity due to Earth tides between a) February 29 to i) March 8, 1996.

Gravity data is in microgals. Note that tremor generally begins at or just past high tide of

March 5, 6, and 8.


119

Discussion

Conduit Opening and Closing

The seismic data presented above appear to point to the repetitive opening and

closing of the shallow conduit system. The frequent periods of chugs and tremor, as well

as Arenal’s continuous activity since 1968, indicate that the volcano is generally

behaving as an open system. However, on a daily scale, Arenal appears to go through

cyclic changes between a closed and an open system.

The higher frequency of eruptions in the morning (e.g., February 29, March 5, 6,

and 8, 1996) and the observed sequence of paired eruptions in the morning (e.g., March

5 and 6, 1996) suggest that Arenal behaves as a comparatively closed system at these

times. The relative quiescence of seismic activity before eruptions on the mornings of

February 29, March 1, and March 6 may further indicate closure of the conduit prior to

the eruptions (Figure 3.6). The paired eruptions of March 5 also suggest that the system

is relatively closed. The first eruption of the pair may partially open the conduit with

comparatively little release of gas pressure. The second, larger-amplitude event will then

destructively open the conduit. These paired eruptions (seen on February 29, March 1, 2,

5, and 6) disappear with decreased eruptive activity in the afternoon, which coincides

with the onset of tremor. In contrast to these days, March 1 shows a distinct lack of

eruptive activity and greater tremor and chugging. This suggests that the volcano was

behaving as an open system throughout the day on March 1.

The decrease in frequency of eruption and increase in tremor activity also

coincide with the arrival of high tide (Figure 3.7). The correlation between volcanic


120

activity and Earth tides has been recognised for some time (Eggers and Decker, 1969;

Hamilton, 1973; Johnston and Mauk, 1972; Mauk and Johnston, 1973; Golombek et al.,

1978; Mauk, 1979), as has a correlation between Earth tides and volcanic SO2 emissions

(Stoiber et al., 1986; Connor et al., 1988). The increased lunar attraction causes the flow

of water towards the locus of maximum attraction, resulting in high tide. This

mechanism also may affect volcanic activity. Barboza and Melson (1990) suggest that

the inverse correlation between tremor activity and eruptions may be due to the rise of

magma in the conduit. As the lunar attraction increases to a maximum, comparatively

hot fresh magma may be pulled towards the surface, resulting in a decrease in surface

crystallisation and consequent decrease in sealing of the conduit. Thus, there is less

pressurisation of the conduit and a resulting decrease in explosive eruptions. Similarly,

degassing will occur more easily, evidenced by the increased chugs and tremor that we

observed in the afternoons. Stoiber et al. (1986) noted that at Masaya caldera, bursts of

gas from the lava lake were twice as likely to occur during the maxima or minimum of

Earth tides. The close relationship between the reduced frequency of eruptions, the

beginning of tremor and the high tide, for three days at least, suggests that Arenal’s

eruptive activity may be sensitive to relatively small changes in the confining pressure.

In a generally open system, these small changes may be sufficient to shift the activity

from a relatively closed system to a more open one or vice-versa.

There may also be a relationship between magma supply/extrusion rates and

periods between explosive eruptions. It was noted that changes in the extrusive activity

from the northern vent of crater C could affect the explosive activity of the southern vent

(M. Davis, personal communication, 1996). If extrusion rates are high, lava may be


121

forced into the southern conduit, causing a temporary blockage. This would allow for

overpressurisation and explosive eruption.

Our COSPEC measurements (e.g., March 4, 1995, March 8, 1996) show

fluctuations in SO2 flux, suggesting that the conduit may undergo cycles of gradual

sealing leading to overpressurisation and eventual eruption. Progressive decreases in

SO2 flux suggest a decrease in the conduit opening and resulting increase in conduit

pressure. In many instances, this increase in pressure continues until a critical limit is

reached, at which point the closure is destroyed by an explosive event. This often may

involve the release of relatively large ash columns and the ballistic ejection of bombs

and blocks. There are, however, instances where there is gas release without ash

emission.

Similar cycles of progressive decrease in SO2 flux have also been noted at

volcanoes such as Galeras (Zapata et al., 1997) and Nevado del Ruiz (Williams et al.,

1990). During the 1992-1993 eruptive period at Galeras, three of the largest eruptions

were preceded by low levels of degassing and immediately followed by intense long-

period seismicity and relatively high SO2 fluxes. Between 1985 and 1987, there were at

least three cases where the SO2 fluxes at Nevado de Ruiz decreased significantly prior to

eruptions (Williams et al., 1990). Arenal’s high level of activity also may allow cycles

of sealing and pressurisation of the conduit leading to eruptions, but on an extremely

short time scale. In contrast to volcanoes such as Galeras or Popocatepétl, sealing and

overpressurisation on Arenal generally occur over a matter of minutes to hours.


122

The Sulphur Budget of Arenal

The SO2 (0.047 Tg⋅yr-1) and CO2 estimates (0.301 Tg⋅yr-1) from Arenal are only a

relatively small fraction (SO2: 0.40%, CO2: 3.9%) of the annual global output of erupting

volcanoes (Stoiber et al., 1987; Williams et al., 1992). Galeras, with at least 1.9 Mt SO2

emitted over a seven year period, represents approximately 1.4% of annual volcanic

output, while CO2 estimates of 2.6 Mt at Galeras represent only 0.57% of total annual

volcanic CO2 flux (Zapata et al., 1997; Williams et al., 1992). Over 1,000 COSPEC

measurements from Mount St. Helens between 1980 and 1988 give a total SO2 emission

of 2 Mt (Gerlach and McGee, 1994). The 1963 eruption of Gunung Agung (Bali,

Indonesia) also released significant amounts into the atmosphere, with an estimated 2.5

Mt of SO2 emitted (Self and King, 1996). Although Arenal’s annual SO2 output is

small, its total output is nevertheless quite similar to those of Mt. St. Helens, Galeras,

Redoubt, and Gunung Agung. but merely over a longer time scale. On a short time

scale, small volcanoes such as Arenal may not have a significant impact on global

volcanic output. However, over long periods of time, continuously active volcanoes

such as Arenal may emit significant quantities of SO2 into the troposphere. These are

comparable to volcanoes which exhibit vigorous degassing but over shorter time scales.

The difference in SO2 estimates between Plume Tracker/COSPEC and

petrological methods necessitates a source for this excess sulphur. Various mechanisms

have been proposed to explain these discrepancies. At El Chichón, for example, excess

sulphur may have been derived from magmatic anhydrite and an S-rich vapour phase

(Luhr et al., 1984). At Nevado del Ruiz, anhydrite (Fournelle, 1990) and sulphur-rich

vapour phases (Sigurdsson et al., 1990) also were shown to be important. Excess


123

sulphur also may derive from the syn-eruptive degassing of sulphur in the melt of a non-

erupted convecting magma. Convection cells would allow for the continued upward

cycling of undegassed magma and consequent downward movement of degassed magma

(Casadevall et al., 1983; Andres et al., 1991; Kazahaya et al., 1994). Excess sulphur also

may arise from the degassing of mixed or commingled intrusions of basaltic magma.

The exsolution and upward migration of less soluble species such as CO2 also may

transport more soluble species such as SO2 to the surface (Andres et al., 1991). SO2 also

might be directly absorbed by the hydrothermal system (Williams et al., 1990).

Extensive hydrothermal systems also may act to seal a volcano and allow for the

accumulation of an independent vapour phase. This could then lead to the release of a

large sulphur-rich gas bubble. This is unlikely at Arenal, however, as it does not have an

extensively developed hydrothermal system.

Unlike Láscar and Lonquimay volcanoes in Chile (Andres et al., 1991) where

petrological estimates were 50-100 times less than COSPEC estimates, petrological

estimates for Arenal are only 1.6 to 4.8 times less than COSPEC values. The melt

inclusions analysed in our study may be samples of melt that had already undergone

some degassing. Thus, these melt inclusions may not represent pristine undegassed

melt. Whatever the cause of excess sulphur, it is of much less importance at Arenal.

This small difference does, however, suggest that Arenal is not being supplied by an

isolated, slowly degassing body of magma. Rather, it is more likely that Arenal is an

open system which is being continuously intruded by fresh magma. There may also be

convection in the conduit, allowing for the continued upward cycling of undegassed

magma. In contrast, volcanoes such as Láscar, Lonquimay, and Pinatubo have


124

undergone extensive sulphur degassing from the melt within an isolated slowly cooling

magma chamber, with the result that there are significant discrepancies between

COSPEC and petrological emission estimates.

The Open Nature of Arenal

According to Stoiber et al. (1986), an overall steady decrease in volcanic SO2

flux with time suggests that a single batch of magma is progressively degassing, without

the influx of new gas-rich magma. This appears to be the case for Masaya caldera in

Nicaragua, Galeras and Nevado del Ruiz in Colombia, and Mount St. Helens in the

United States, which show rate decay constants of 0.04 yr-1, 0.27 yr-1, 0.38 yr-1 and 1.41

yr-1, respectively (Stoiber et al., 1986; Zapata et al., 1997; Williams et al., 1990). In

contrast to Galeras which clearly has acted as a relatively closed system (Zapata, 1997),

SO2 fluxes at Arenal actually have increased slightly over time (Table 3.4). Unlike

Galeras or even Masaya, which receive episodic (decades) supply of small shallow

magma batches, Arenal appears to have a continuous input of magma from depth.

Rather than being supplied from a stagnant shallow chamber or small body of magma

that progressively degasses, magma beneath Arenal may reside in a chamber which itself

is open to replenishment. Based on observed geochemical changes in extruded lavas,

modelling by Reagan et al. (1987) concluded that the Arenal magma has undergone three

compositional stages prior to and after the 1968 reactivation. Changes in composition of

stage 3 magmas (1974-present) also indicate continued influx of magma along with

crystal removal. The chamber is probably located in the middle to lower crust (Reagan

et al., 1987), from which a series of conduits or fractures, opened by magmatic pressure,


125

rise to a kilometre beneath the volcano. The final kilometre to few hundred metres

consist of a conduit or conduits which open and close on short time scales, that result in

frequent strombolian eruptions and extrusion of lava (Wadge, 1983).

Extrusion of lava also appears to be independent of these strombolian eruptive

cycles, which suggests that there may be a complex system of conduits or fractures just

below the surface. Low frequency volcanic tremors (2-4 Hz) seen before and after

eruptions may represent the oscillation of magma or an organ-pipe effect in these

conduits (Matumoto and Umana, 1976; Wadge, 1983; Alvarado et al., 1988; Barquero et

al., 1992).

Conclusions

Due to the high level of activity at Arenal, collection of a large set of SO2 fluxes

between eruptions was difficult. However, it is nevertheless possible to observe cyclical

variations in SO2 fluxes before and after eruptions. When one compares the gas flux to

the seismic data showing declines in tremor prior to eruptions, it becomes apparent that

there is a repetitive cycle of activity. Correlations between seismic activity and Earth

tides suggest that an extremely open system such as Arenal may be quite sensitive to

minor variations in confining pressures, changing from a relatively closed system to a

comparatively open system over the space of minutes to hours. The small difference

between petrological and COSPEC SO2 flux suggests that Arenal is being continuously

supplied by fresh magma. The cycle of explosive eruptions may be explained by

repeated closure of the conduit(s) due to crystallisation of the magma, leading to

overpressure and explosive destruction of the magma cap. While Arenal may have only


126

a small influence in terms of the annual global volcanic input of SO2 and CO2 to the

atmosphere, the continuous activity of the volcano has nevertheless contributed at least

1.3 Mt of SO2 to the troposphere since 1968, comparable to volcanoes such as Mt. St.

Helens, Nevado del Ruiz, and Galeras. Arenal’s high level of activity allows for the

study of multiple cycles of conduit opening and closing and thus is an excellent tool for

better understanding the manner in which an open-system volcano degasses.


127

References


Alvarado, G. E., Matumoto, T. E., Borgia, A., and Barquero, R., 1988. Sintesisgeovulcanologica del Arenal (Costa Rica): 20 años de continua actividad eruptiva(1968-1988). Bol. Obs. Vulcanol. Arenal, 1: 1-55.

Anderson, O. L., 1978. The role of magma vapors in volcanic tremors and rapideruptions. Bull. Volcanol., 41: 341-353.

Andres, R. J., Rose Jr., W. I., Kyle, P. R., DeSilva, S., Francis, P., Gardeweg, M.,and Moreno Roa, H., 1991. Excessive sulfur dioxide emissions from Chileanvolcanoes. J. Volcanol. Geotherm. Res., 46: 323-329.

Barboza, V. and Melson, W. G., 1990. Correlacion entre las señales sismicas y lossonidos de las erupciones del Volcán Arenal. Bol. Vulcanol., 21: 8-12.

Barquero, R., Alvarado, G. E., and Matumoto, T., 1992. Arenal Volcano (Costa Rica)premonitory seismicity. In: R. Gasparini, R. Scarpa, and K. Aki (Editors),Volcanic Seismology. Springer-Verlag, New York, pp. 84-96.


Carr, M. J., 1984. Symmetrical and segmented variation of physical and geochemicalcharacteristics of the Central American Volcanic Front. J. Volcanol. Geotherm.Res., 20: 231-252.

Casadevall, T. J., Johnson, D. A., Harris, D. A., Rose Jr., W. I., Malinconico Jr., L.L., Stoiber, R. E., Bornhorst, T. J., Williams, S. N., Woodruff, L., andThompson, J. M., 1981. SO2 emission rates at Mount St. Helens from March 29through December, 1980. In: P. W. Lipman and D. R. Mullineaux (Editors), The1980 Eruptions of Mount St. Helens. U.S. Geol. Surv. Prof. Pap., 1250, pp. 193-200.

Casadevall, T. J., Rose Jr., W. I., Fuller, W. H., Hunt, W. H., Hart, M. A., Moyers,J. L., Woods, D. C., Chuan, R. L., and Friend, J. P., 1984. Sulfur dioxide andparticles in quiescent volcanic plumes from Poás, Arenal, and Colima volcanoes,Costa Rica and Mexico. J. Geophys. Res., 89: 9633-9641.

Casadevall, T. J., Rose Jr., W. I., Gerlach, T. M., Greenland, L. P., Ewert. J.,Wunderman, R., and Symonds, R., 1983. Gas emissions and the eruptions of


128

Mount St. Helens through 1982. Science, 221: 1383-1385.

Cheminée, J. L., Delorme, H., Barquero, J., Avila, G., Malavassi, E., and Güendel,F., 1981. Algunos aspectos fisicos y quimicos de la actividad de los volcanesPoás y Arenal. Bol. Vulcanol., 11: 12-16.

Cigolini, C., Borgia, A., and Casertano, L., 1984. Intra-crater activity, aa-block lava,viscosity and flow dynamics: Arenal Volcano, Costa Rica. J. Volcanol.Geotherm. Res., 20: 155-176.

Connor, C. B., Stoiber, R. E., and Malinconico Jr., L. L., 1988. Variation in sulfurdioxide emissions related to Earth tides, Halemaumau crater, Kilauea volcano,Hawaii. J. Geophys. Res., 93: 14867-14871.

Eggers, A. A. and Decker, R. W., 1969. Frequency of historic volcanic eruptions. EOSTrans. Am. Geophys. Union, 50: 343.

Fernández, E., Duarte, E., Barboza, V., Van der Laat, R., Hernandez, E., Martinez,M., Saenz, R., Soto, G. J., and Arias, J. F., 1996a. Arenal. GVN Bull., 21 (9):5.

Fernández, E., Duarte, E., Barboza, V., Van der Laat, R., Hernandez, E., Martinez,M., Saenz, R., Soto, G. J., Alvarado, G. E., and Arias, J. F., 1996b. Arenal.GVN Bull., 21 (6): 8-9.

Fournelle, J., 1990. Anhydrite in Nevado del Ruiz November 1985 pumice: Relevanceto the sulfur problem. J. Volcanol. Geotherm. Res., 42: 189-201.

Gerlach, T. M. and McGee, K. A., 1994. Total sulfur dioxide emissions and pre-eruption vapor-saturated magma at Mount St. Helens, 1980-88. Geophys. Res.Lett., 21: 2833-2836.

Golombek, M. P., and Carr, M. J., 1978. Tidal triggering of seismic and volcanicphenomena during the 1879-1880 eruption of Isles Quemadas volcano in ElSalvador, Central America. J. Volcanol. Geotherm. Res., 3: 299-307.

Hamilton, W. L., 1973. Tidal cycles and volcanic eruptions: Fortnightly to 19 yearlyperiods. J. Geophys. Res., 78: 3363-3375.

Johnston, M. J. and Mauk, F. J., 1972. Earth tides and the triggering of eruptions fromMt. Stromboli, Italy. Nature, 239: 266-267.

Kazahaya, K., Shinohara, H., and Saito, G., 1994. Excessive degassing of Izu-Oshimavolcano: magma convection in a conduit. Bull. Volcanol., 56: 207-216.

Luhr, J. F., Carmichael, I. S. E., and Varekamp, J. C., 1984. The 1982 eruptions ofEl Chichón Volcano, Chiapas, Mexico: Mineralogy and petrology of the


129

anhydrite-bearing pumices. J. Volcanol. Geotherm. Res., 23: 69-108.

Matumoto, T., Ohtake, M., Latham, G., and Umana, J., 1977. Crustal structure insouthern Central America. Bull. Seismol. Soc. Am., 67: 121-134.

Matumoto, T. and Umana, J. E., 1976. Informe sobre la erupción del volcan Arenaloccurida el 17 de Junio de 1975. Rev. Geofis. Inst. Panama Geogr. Hist., 5: 299-315.

Mauk, F. J., 1979. The triggering of activity at Soufrière de St. Vincent, April 1979, bysolid Earth tides. EOS Trans. Am. Geophys. Union, 60: 833.

Mauk, F. J. and Johnston, M. J., 1973. On the triggering of volcanic eruptions byEarth tides. J. Geophys. Res., 78: 3356-3362.

Melson, W. G., 1989. Las erupciones del Volcán Arenal 1 al 13 de Abril de 1989. Bol.Vulcanol., 20: 15-22.

Reagan, M. K., Gill, J. B., Malavassi, E., and Garcia, M. O., 1987. Changes inmagma composition at Arenal Volcano, Costa Rica, 1968-1985: Real-timemonitoring of open-system differentiation. Bull. Volcanol., 49: 415-434.

Self, S. and King, A. J., 1996. Petrology and sulfur and chlorine emissions of the 1963eruption of Gunung Agung, Bali, Indonesia. Bull. Volcanol., 58: 263-285.

Sigurdsson, H., Carey, S. N., Palais, J. M., and Devine, J., 1990. Pre-eruptioncompositional gradients and mixing of andesite and dacite magma erupted fromNevado del Ruiz Volcano, Colombia in 1985. J. Volcanol. Geotherm. Res., 41:127-151.


Stoiber, R. E., Malinconico, Jr. L. L., and Williams, S. N., 1983. Use of thecorrelation spectrometer at volcanoes. In: H. Tazieff and J. C. Sabroux (Editors),Forecasting Volcanic Events. Elsevier, New York, pp. 424-444.

Stoiber, R. E., Williams, S. N., and Huebert, B. J., 1986. Sulfur and halogen gases atMasaya caldera complex, Nicaragua: Total flux and variations with time. J.Geophys. Res., 91: 12215-12231.

Stoiber, R. E., Williams, S. N., and Huebert, B. J., 1987. Annual contribution ofsulfur dioxide to the atmosphere by volcanoes. J. Volcanol. Geotherm. Res., 33:1-8.

Stoiber, R. E., Williams, S. N., Naslund, H. R., Connor, C. B., Prosser, J. T.,Gemmell, J. B., Malavassi, E., and Barquero, J., 1982. Costa Rica: Activity at


130

4 volcanoes summarized. SEAN Bull., 7 (11): 6-7.

Van der Laat, R. and Carr, M., 1989. Arenal. In: L. McClelland, T. Simkin, M.Nielsen E. Summers, and T. C. Stein (Editors), Global Volcanism 1975-1985:The First Decade of Reports from the Smithsonian Institution's Scientific EventAlert Network (SEAN). Prentice Hall, Inc., New Jersey, pp. 510-511.

Wadge, G., 1983. The magma budget of Volcán Arenal, Costa Rica from 1968 to 1980.J. Volcanol. Geotherm. Res., 19: 281-302.

Willett, H. C. and Sanders, F. 1959. Descriptive Meteorology. 2nd Ed. Academic PressInc., New York, 355 pp.

Williams, S. N., Schaefer, S. J., Calvache V., M. L., and Lopez, D., 1992. Globalcarbon dioxide emission to the atmosphere by volcanoes. Geochim. Cosmochim.Acta, 56: 1765-1770.

Williams, S. N., Sturchio, N. C., Calvache V., M. L., Mendez F., R., Londono C., A.,and Garcia P., N., 1990. Sulfur dioxide from Nevado del Ruiz Volcano,Colombia: Total flux and isotopic constraints on its origin. J. Volcanol.Geotherm. Res., 42: 53-68.


CONCLUSIONS

Conclusions______________

132

General Conclusions

HIS study of seismicity, CO2 and radon soil gas, and SO2 at Arenal has provided a

better understanding of volcanic behaviour. Some important conclusions from

this work are as follows:

1. Correlations between soil gas concentrations and seismic data are difficult to

establish, due in part to limited seismic coverage, but more importantly, due to the high

level of activity at Arenal.

2. Temporal variations in radon and CO2 soil gas concentrations are due in large part to

changes in atmospheric pressure over time.

3. Rn and CO2 soil gases from the upper flanks of Arenal are unlikely to originate from

deep levels, but rather come from shallow surface sources, as the radon half-life is too

short and the transport process too slow.

4. The diffuse soil gases are generally unable to penetrate the young lavas which cover

and seal the upper flanks of Arenal. Only on the lower flanks, where young lavas do not

crop out, is there any gas flow from deeper levels. This is evident from the increased

CO2 concentrations and heavier δ13C values at greater distances from the crater.

5. The degree of soil development and permeability of the substrate also strongly

influences the concentrations of CO2 and radon soil gas at Arenal. Unconsolidated

volcanic soils on the upper flanks of the volcano have relatively low RnERaC and

consequently low radon values. These soils are also more apt to rapidly dissipate any

precipitation, thus limiting sealing effects. This is in contrast to the more clay- and

organic-rich soils of the lower flanks, which retain humidity and increase sealing. This

results in a lower permeability in the better developed soils and permits the accumulation

T


133

of soil gas below the surface.

6. Volcanoes such as Arenal, Poás, and Galeras act as volcanic plugs which seal shallow

levels of the continental crust, limiting deep gas flux to fumaroles, faults, and fractured

lower flanks of the volcanoes.

7. The high level of activity at Arenal makes collection of a large set of SO2 flux

measurements between eruptions difficult. It was, nevertheless, possible to see cyclical

variations in SO2 fluxes before and after eruptions. When one compares the SO2 results

to the seismic data showing declines in tremor prior to eruptions, it becomes apparent

that there is a repetitive cycle of activity.

8. Correlations between seismic activity and Earth tides suggest that an extremely open

system such as Arenal may be quite sensitive to minor variations in confining pressures,

changing from a relatively closed system to one that is open over the space of minutes to

hours.

9. The small difference between petrological and COSPEC SO2 output suggests that

Arenal is being continuously supplied by fresh magma. The cycle of explosive eruptions

may be explained by repeated sealing of the conduit(s) due to cooling of the magma.

Subsequent gas pressure increases may then lead to explosive destruction of the magma

cap.

10. Arenal exerts a small influence in terms of annual global volcanic input of SO2 and

CO2 to the atmosphere. However, the continuous and high activity of the volcano has

nevertheless contributed at least 1.3 Mt of SO2 to the troposphere since 1968. Arenal’s

high level of activity permits us to the study multiple cycles of conduit closing and

overpressurisation and is thus an excellent means for better understanding the manner by


134

which an open-system volcano degasses.

Recommendations for Future Work

During this study, only the lower part of the edifice was surveyed by soil gas

measurements. Should current activity decrease to a point where access to the upper

flanks is possible, systematic sampling in closer proximity to the crater would greatly

improve the degassing hypothesis. Stable carbon isotopes should be studied in more

detail at Arenal, specifically in proximity to areas of structural weakness and with

increased distance from the summit. This also will aid in clarifying the relationship

between the capping effect of young lavas on the upper flanks and their subsequent

breakdown and fracturing on the lower flanks.

Year-round monitoring of SO2 flux in conjunction with the emplacement of a

more extensive seismic network on the volcano are necessary. This would allow for

more detailed analysis of the correlations between flux and seismic fluctuations.

Volcanic seismic activity may cause soil gas anomalies on Arenal. Additional research

is necessary to better understand the short term fluctuations due to seismic fluctuations.

Should explosive activity decrease substantially, emplacement of a small

meteorological station near the summit would greatly increase the accuracy of windspeed

measurements and thus decrease the uncertainty of SO2 flux measurements. While the

current strombolian activity continues, a small helium balloon and/or theodolite should

be used to better constrain the actual windspeed of the plume.

A more detailed study of recent extrusion rates and volumes of lava emplaced

since 1985 is required. In conjunction with more detailed analyses of concentrations of


135

sulphur in melt inclusions, this allow will for better petrological estimates of the total

SO2 emitted since reactivation of the volcano. This will also lead to a better

understanding of the presence or lack of excess sulphur in open-system volcanoes.

APPENDIX

Appendix A______________

A-1

Appendix A______________

A-2

Appendix A______________

A-3

Appendix A______________

A-4

Appendix A______________

A-5

Appendix A______________

A-6

Appendix A______________

A-7

Appendix A______________

A-8

Appendix A______________

A-9

The Distribution and Origin of Radon, CO 2 and SO2 Gases ... · The Distribution and Origin of...

Documents

Transcript of The Distribution and Origin of Radon, CO 2 and SO2 Gases ... · The Distribution and Origin of...