Hazards and Monitoring of Volcanic Activity 2 E-Book

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 490

Veröffentlichungsjahr: 2022

Beschreibung

The impact of natural disasters has become an important and ever-growing preoccupation for modern societies. Volcanic eruptions are particularly feared due to their devastating local, regional or global effects. Relevant scientific expertise that aims to evaluate the hazards of volcanic activity and monitor and predict eruptions has progressively developed since the start of the 20th century. The further development of fundamental knowledge and technological advances over this period have allowed scientific capabilities in this field to evolve. Hazards and Monitoring of Volcanic Activity groups a number of available techniques and approaches to render them easily accessible to teachers, researchers and students. This volume sets out different surveillance methods, starting with those most frequently used: seismic surveillance and deformation. It then examines surveillance by remote sensing from ground, air and space, methods that exemplify one of the most spectacular advances in this field in recent times.

Leseprobe

Table of Contents

Cover

Title Page

Copyright

Foreword

Preface

List of Abbreviations

1 Seismic Monitoring of Volcanoes and Eruption Forecasting

1.1. Introduction

1.2. Instrumentation and seismic networks

1.3. Types of seismic-volcanic events

1.4. Volcanic seismicity

1.5. Processing of seismic-volcanic signals

1.6. Network data analysis

1.7. Forecasting eruptions

1.8. The FFM method

1.9. Case studies

1.10. Conclusion

1.11. References

2 Monitoring Volcano Deformation

2.1. Introduction

2.2. Phenomena at the origin of deformation

2.3. Deformation measurement techniques

2.4. Adequacy between deformation measurements and monitoring

2.5. Contributions and limitations of deformation modeling to the study of volcanoes

2.6. Perspectives: from operational monitoring to physical and predictive models

2.7. References

3 Volcano Monitoring by Remote Sensing

3.1. Introduction

3.2. Basic concepts in remote sensing

3.3. Main available operating systems

3.4. Monitoring volcanic ash plumes

3.5. Monitoring volcanic SO

2

plumes

3.6. Lava flow monitoring

3.7. Future developments

3.8. References

4 Volcano Remote Sensing with Ground-Based Techniques

4.1. Introduction

4.2. Basic concepts in ground-based remote sensing relying on electromagnetic waves

4.3. Ground remote sensing for the detection of volcanic gas

4.4. Ground-based IR remote sensing

4.5. Other volcano remote sensing methods with ground-based techniques

4.6. Future developments

4.7. References

List of Authors

Index

End User License Agreement

List of Tables

Chapter 1

Table 1.1.

Correspondence between the terminologies used for the classification ...

Table 1.2.

Factors making it easier or more difficult to forecast volcanic erupt...

Chapter 2

Table 2.1.

Link between volcanic displacements detected by InSAR and eruptions, ...

Chapter 3

Table 3.1.

List of main available sensors and operational services

Table 3.2.

List of main UV/VIS/IR sensors used for ash detection

Table 3.3.

UV/IR embedded sensors for satellites

Chapter 4

Table 4.1.

Explosive plume height at Etna (Scollo et al. 2014), and mass eruptio...

Table 4.2.

Physical characteristics of fumarole field temperatures

Table 4.3.

Main types of disdrometers used for monitoring ash fallout and their ...

Table 4.4.

Lightning detection networks. According to Behnke and McNutt (2014)

Guide

Cover

Table of Contents

Title Page

Copyright

Foreword

Preface

List of Abbreviations

Begin Reading

List of Authors

Index

End User License Agreement

Pages

v

iii

iv

xi

xii

xiii

xv

xvi

xvii

xviii

xix

xx

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

303

304

305

306

307

308

SCIENCES

Geoscience, Field Director – Yves Lagabrielle

Lithosphere–Asthenosphere Interactions, Subject Head – René Maury

Hazards and Monitoring of Volcanic Activity 2

Seismology, Deformation and Remote Sensing

Coordinated by

Jean-François Lénat

First published 2022 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms and licenses issued by the CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the undermentioned address:

ISTE Ltd

27-37 St George’s Road

London SW19 4EU

UK

www.iste.co.uk

John Wiley & Sons, Inc.

111 River Street

Hoboken, NJ 07030

USA

www.wiley.com

© ISTE Ltd 2022

The rights of Jean-François Lénat to be identified as the author of this work have been asserted by him in accordance with the Copyright, Designs and Patents Act 1988.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s), contributor(s) or editor(s) and do not necessarily reflect the views of ISTE Group.

Library of Congress Control Number: 2022934927

British Library Cataloguing-in-Publication Data

A CIP record for this book is available from the British Library

ISBN 978-1-78945-045-3

ERC code:

PE10 Earth System Science

PE10_5 Geology, tectonics, volcanology

PE10_7 Physics of earth’s interior, seismology, volcanology

Foreword

Claude JAUPART

Institut de physique du globe de Paris, Université de Paris, Académie des sciences, France

Volcanoes are fascinating because of the beautiful landscapes they form and their superb eruptions. Many books have documented the most catastrophic eruptions, the different eruptive regimes and the physical mechanisms involved. Others have popularized tales of adventurers getting close to volcanic explosions and flows at the risk of their lives. In comparison, the discreet and dedicated work of volcanologists monitoring volcanoes has remained little known. The methods and techniques they use have improved greatly in recent decades and now allow them to predict eruptions with very few limitations. These advances are to be presented and explained in the three volumes of this book.

Volcanoes are built on top of a vast network of underground plumbing, and it is within this network that eruptions occur. Near the surface, the permeable rocks are gorged with water that circulates, heats up at depth and vaporizes under widely varying conditions. The result is a myriad of manifestations, including fumarole fields with changing flows, small earthquakes and ground deformation. This high background noise makes monitoring difficult. An eruption is preceded by the setting in motion of magma from one or more reservoirs located several kilometers underground. This onset is often very discrete and the associated signals are not easily distinguished from the background noise. Once the magma is near the surface, the signs are numerous and leave no room for doubt, but things can move very quickly and it is often too late to evacuate the area. Simply recognizing that an eruption is imminent is not enough: one must also assess its intensity and regime. Sometimes an eruption can even occur without magma and take the form of phreatic explosions, where the water contained in the surface rocks of the volcano vaporizes explosively. The volcanologist’s work does not stop when the eruption begins; they must follow it over time and be able to distinguish between a temporary stop and its true end.

Faced with these multiple challenges, volcanologists have adopted methods that can be divided into two broad categories. The first is the historical study that reconstructs the past eruptions of a volcano and the time intervals between them. Knowing that an eruption covers the deposits of those that preceded it and destroys most of them, establishing a reliable chronology and estimating the volumes ejected rely on particular sampling strategies and frequent round trips between the field and the laboratory. New dating methods had to be developed to determine the ages of deposits older than a few tens of thousands of years. The second category covers all the physical and chemical methods used to determine the deep structure of a volcano and to locate the perturbations that it hosts. Rocks are difficult to penetrate and do not allow us to observe the reservoirs and conduits that feed eruptions. The information we obtain is indirect and often ambiguous. For example, small earthquakes are recorded, but they can be caused by the opening of cracks in a hydrothermal system or by magma that moves or by isolated landslides in places that are not easily accessible. An area of abnormal electrical conductivity can be detected, but it may be weathered rock or rock with water-filled fractures. Volcanologists have improved the uncertainty by combining several methods and have added to their toolbox over the years.

Remarkable progress has been made in the last four decades. Previously, the equipment available was limited to a few heavy and unwieldy devices designed for larger scale studies. Measurements have become much more precise, the number of sensors has increased enormously and the mathematical techniques of analysis have been refined. Well-instrumented observatories have been installed on many active volcanoes and in particular on the three volcanoes of the French national territory in Guadeloupe, Martinique and Reunion Island. Nowadays, a “typical” observatory maintains more than a hundred sensors of all kinds. The last two decades have seen the advent of very efficient satellite tools. However, it should not be inferred that volcanology has become the business of pure measurement experts or laboratory researchers. Knowledge of the special features of each volcano is needed to advance. The active volcano of Santorini in the Cyclades, which grows in the middle of a large caldera, is not monitored in the same way as Piton de la Fournaise on Reunion Island, which rises nearly 7 km above the sea floor and grows on the flanks of the ancient Piton des Neiges. It would be absurd to sample the deposits and install sensors randomly or only in easily accessible locations. Every volcanologist, whether geologist, geophysicist or geochemist, has studied their volcanoes for several years. This patient work has rarely been described. In this book, specialists from all the major disciplines of volcanology share their work and their discoveries. They explain how they decipher and interpret their measurements. One is likely to be surprised by the weakness of the signals detected, which can only be measured with sophisticated instruments, in comparison with the enormity of the eruptive phenomena. But it is thanks to these signals that we are able to travel to the very heart of volcanoes.

Preface

Jean-François LÉNAT

Laboratoire Magmas et Volcans, CNRS, IRD, OPGC, Université Clermont Auvergne, Clermont-Ferrand, France

The impact of natural disasters has become a major concern of our modern societies. Volcanic eruptions, although statistically less deadly and causing less damage than earthquakes or certain atmospheric phenomena, can have devastating local or global effects.

The methods used to determine hazards related to volcanic activity and to monitor the latter are part of many Earth science curricula, both at master and thesis levels.

There are many publications in these areas, but the information is fragmented, requiring teachers to consult a large number of documents to develop their teaching. The aim of this book is to provide them with a single resource, written by specialists, on the methods of monitoring and determining hazards.

The subject is vast, which has led us to present it in three volumes. The first is devoted to geological and historical approaches. The next two are devoted to monitoring methods. The aim of each chapter is not to be encyclopedic. Rather, the intention is to provide the reader with the basic fundamentals of each of the topics covered. On the other hand, each author has taken care to provide bibliographic references that will allow readers to find the detailed information they may need.

This book deals with a scientific field that is constantly evolving. The progress in scientific concepts, approaches, observations and techniques has been spectacular during the last decades. There is no reason why this dynamic should slow down in the future. A logical consequence is that updates should be made periodically to avoid obsolescence of such a book. We therefore hope that it will be useful in the present period and that future editions will enable it to retain its value over time.

March 2022

List of Abbreviations

AI

Ash-Index

AVTIS

All-weather Volcano Topography Imaging Sensor

BLNSS

Base Level Noise Seismic Spectrum

BT

Brightness Temperature

BTD

Brightness Temperature Difference

CAPPI

Constant Altitude PPI

COMET

Center for the Observation and Modeling of Earthquakes, Volcanoes and Tectonics

COSPEC

Correlation Spectroscopy

CRF

Continual Radio-Frequency

DIAL

Differential Absorption Lidar

DOAS

Differential Optical Absorption Spectroscopy

DU

Dobson Unit

EASA

European Aviation Safety Agency

EDM

Electronic Distance Measurement

ENGLN

Earth Networks Global Lightning Network

EUR/NAT

European and North Atlantic

FFM

Failure Forecast Method

FMCW

Frequency Modulated Continuous Wave

FOV

Field of View

FPA

Focal-Plane Array

FTIR

Fourier Transform Infrared Spectroscopy

GEO

Geostationary Earth Orbit

GNSS

Global Navigation Satellite System

GPRI

Gamma Portable Radar Interferometer

GPS

Global Positioning System

IASI

Infrared Atmospheric Sounder Interferometer

IATA

International Air Transport Association

IFOV

Instantaneous FOV

INGV

Instituto Nazionale di Geofisica e Vulcanologia

(Italian National Institute of Geophysics and Volcanology)

InSAR

Interferometric Synthetic Aperture Radar

LEO

Low Earth Orbit

LF

Linear Fit

LIDAR

Light Detection and Ranging

LP

Long-Period

LPM

Laser Precipitation Monitor

LTA

Long-Term Average

MER

Mass Eruption Rate

MIR

Mid-Infrared

MRR

Micro Rain Radar

MSG

Meteosat Second Generation

NIR

Near Infrared

NOVAC

Network for Observation of Volcanic and Atmospheric Change

NTI

Normalized Thermal Index

OMI

Ozone Monitoring Instrument

OVPF

Observatoire volcanologique du Piton de la Fournaise

(Volcanological Observatory of Piton de la Fournaise)

PIT

Pixel-Integrated Temperature

PPI

Plan Position Indicator

PPP

Precise Point Positioning

Radar

Radio Detection and Ranging

RSAM

Real-time Seismic Amplitude Measurement

RSEM

Real-time Seismic Energy Measurement

SARA

Seismic Amplitude Ratio Analysis

SEVIRI

Spinning Enhanced Visible and Infared Imager

SPAC

Spatial Autocorrelation

SSAM

Real-time Seismic Spectral Amplitude Measurement

STA

Short-Term Average

STFT

Short-Term Fourier Transform

TGSD

Total Grain Size Distribution

TIR

Thermal Infrared

TOMS

Total Ozone Mapping Spectrometer

TROPOMI

Tropospheric Monitoring Instrument

ULP

Ultra Long-Period

USGS

United States Geological Survey

UV

Ultraviolet

VAA

Volcanic Ash Advisories

VAAC

Volcanic Ash Advisory Center

VAG

Volcanic Ash Graphics

VIS

Visible

VLP or VLF

Very-Long-Period (or Very-Low-Frequency)

VOLDORAD

Volcano Doppler Radar

VT

Volcano-Tectonic earthquakes

WWLLN

World Wide Lightning Location Network

1Seismic Monitoring of Volcanoes and Eruption Forecasting

Philippe LESAGE

ISTerre, Université Savoie Mont Blanc, Université Grenoble Alpes, CNRS, IRD, Université Gustave Eiffel, France

1.1. Introduction

Many volcanologists consider seismology the backbone of volcano monitoring. Chapter 4 from Volume 1 describes its development since the end of the 19th century. Basic seismic volcano monitoring can be performed with relatively few physical resources, and easily detect early precursory signals of an eruption. It is the most widespread method used by observatories, as it is rare for an eruption not to be preceded by an increase in seismic activity within and nearby the volcano. This is due to the fact that in order to reach the surface, magma and other volcanic fluids must clear a passage through layers of solid rock, fracturing them and causing the recently opened cavities to vibrate. This results in various types of seismic signals, whose study provides information about the location and migration of magma and other fluids, as well as their physical properties.

As a first step, seismic observations from monitoring networks are used for determining a volcano’s status and assessing the probability of a potential eruption. Then, the same data can be analyzed in further detail to have a better grasp of these phenomena. These observations constitute one of the main sources of information on the physical processes involved in magmatic and hydrothermal systems and the structure of volcanic edifices. In this way, the two aspects of volcano seismology – monitoring and research – become complementary: data from observatories make it possible to advance fundamental studies and, in return, these contribute to the improvement of monitoring methods and the interpretation of observations. As with classical seismology, the purpose is to study seismic sources and wave propagation within structures. However, due to the specificity of phenomena provoking seismic vibrations inside volcanoes, to the complexity of such environments and of monitoring goals, volcano seismology has developed or adapted distinctive methods for analyzing the seismic signals observed, modeling their sources, studying the propagation of waves in highly heterogeneous structures and identifying the precursory signals of volcanic eruptions.

The main goal of this chapter is to discuss the state of the art on the seismic activity of volcanoes and the methods for monitoring and forecasting eruptions. It will review other approaches which have been developed for more fundamental studies likely to be applied to volcano monitoring and to the interpretation of observations.

1.2. Instrumentation and seismic networks

As in many other fields, in recent decades volcano seismology has benefited from remarkable improvements in instrumental techniques and measurement systems. Until the 1990s, sensors used on volcanoes were mainly short-period seismometers, data transmissions were analog and recordings were made on paper (either with ink or smoke) or on magnetic tape. Measurement systems had low sensitivities and dynamics, limited bandwidth, recordings were often saturated, their processing was burdensome and measurements of arrival times inaccurate. The development of broadband seismometers and the generalization of digital measurement, transmission, recording and processing systems have profoundly changed the seismic monitoring of volcanic activity, although in certain cases analog techniques are still used.

1.2.1. Measurement systems

A measurement system is made up of a set of elements through which information is transferred. The first element is a sensor which produces an analog signal, more or less proportional to a physical value. A seismometer, for example, is sensitive to vibrations and generates electrical voltage. This voltage is amplified and filtered into the signal conditioning element. This signal can be transmitted analogically and then digitized, or be digitized and then digitally transmitted (Lesage et al. 1995). In the first case, transmission is done by frequency modulated VHF radio. Each radio channel can transmit up to eight multiplexed seismic signals, with limited bandwidth and dynamic range (40–60 dB). Thus, all signals are transmitted – eventually via repeaters – to an observatory where they are digitized together, and then recorded on a storage device (hard disk, CD, etc.). In this case, a single clock, controlled by GPS, makes it possible to accurately date the signals of all the stations following a common time reference. In the second case, the conditioned signal is digitized on the spot and samples dated using a GPS clock and ultimately recorded in situ. All of these operations are integrated into a digital seismic station. This can include or be connected to a digital data transmission system. Digital telemetry includes error correction algorithms. Terrestrial (Wi-Fi type) systems are generally bidirectional, which makes it possible to control stations remotely. Satellite transmissions are implemented for stations located in isolated areas far from observatories. The digital data received at the observatory are then stored for subsequent analysis. If the system is analog and the transmission is temporarily interrupted, data are lost, whereas a digital system equipped with a buffer memory can transmit data once the connection is reestablished. One essential feature of a measurement system is the number of bits involved in the digitization process. After using analog telemetry with limited dynamics, a 12- or 16-bit digitization is sufficient. In stations including a broadband High Dynamic Range sensor, the digitizer often works on 24 bits. Another important feature of a measurement system is the sampling frequency. The frequency of seismic signals observed on volcanoes is generally lower than 50 Hz. In order to comply with the sampling theorem which states that the sampling frequency should be at least twice the signal’s maximum frequency, a value of 100 Hz is generally used. This corresponds to a sampling interval of 0.01 s, which is small enough to accurately pick up the arrival times of seismic waves.

1.2.2. Sensors

The most widespread sensors for monitoring volcanoes are seismometers, instruments delivering signals which are proportional to ground velocity. The main characteristic which differentiates the types of seismometers is their bandwidth and, in particular, their lower cutoff frequency. “Short-period” seismometers have lower cutoff frequencies of 1, 2 or 4 Hz. They are passive mechanical sensors and, therefore, do not need to be powered by an electrical source. They have the advantage of being robust, reliable and inexpensive, although their dynamic – the ratio being the maximum and minimum detectable signals – is limited (a few thousand, around ≃70 dB). Until the 1990s, volcano monitoring and study networks relied on “short-period” seismometers with one vertical component supplemented by a few three-component sensors in order to better detect S-waves, or to study wave polarization. Seismometers with broadened frequency band appeared in the 1990s. They use “short-period” seismometers and they include an electronic amplifier-filter system which makes it possible to divide the lower cutoff frequency by a factor of 5–10. For example, starting with a 2 Hz geophone, one obtains a seismometer with a 0.2 Hz cutoff frequency (Lennartz LE-3Dlite). On equal bandwidth, these instruments are less expensive and smaller in size than non-broadened sensors.

The main advancement in the field of instrumental seismology has been the development of “broadband” seismometers. These instruments were first used on volcanoes in the 1990s (Neuberg et al. 1994) and have made it possible to bring to light numerous phenomena producing seismic waves at frequencies below 1 Hz. They are based on small high-frequency seismometers, but have a feedback loop which keeps the seismometer’s mass fixed. The device measures the current intensity flowing into this loop. Their cutoff frequency reaches very low values, corresponding to periods of 30, 60 or 120 s. Their sensitivity and dynamics are very high (some 106, i.e. ≃130 dB). Most of these seismometers measure the three components of ground motion (Güralp CMG-40T, Nanometrics Trillium, Kinemetrics MBB-2). Many broadband seismometer models include a digitizer, a GPS receiver and a Wi-Fi type digital transmission system, making their configuration and integration into a network relatively easy (Güralp 6TD, Nanometrics Meridian). Some versions have been designed to be installed several meters deep in boreholes, which can greatly reduce the level of seismic noise and thus gain high sensitivity (Geotech KS-54000, Kinemetrics STS-5A, etc.). Despite having a significantly higher price than passive sensors, as well as greater complexity and susceptibility to failure, broadband seismometers have become essential components in seismic monitoring networks. Under certain conditions, these instruments can also be used as tiltmeters (Battaglia et al. 2000). The instrumental response in amplitude of two short-period seismometers and a broadband seismometer are compared in Figure 1.1.

Volcanoes also generate acoustic waves which propagate into the atmosphere, especially during explosions. These signals cover a wide spectral range, mainly in the infrasound range. Seismic networks are being increasingly supplemented by acoustic sensors in order to detect and study such signals. Acoustic sensors can be highly sensitive and very broadband microbarometers (Martec MB2005; Chaparral Physics Model 50), such as the instruments used in the Treaty on the Non-Proliferation of Nuclear Weapons (NPT) monitoring network, or microphones. More recently, infrasound sensors have expressly been developed for applications in volcanology. They are based on piezoresistive (MEMS) Differential Pressure Transducers capable of operating with frequencies down to a few mHz (Marcillo et al. 2012; Grangeon and Lesage 2019). These low-cost sensors are often grouped to form arrays which can improve device sensitivity and determine the direction of propagation and, therefore, the origin of detected acoustic waves (Johnson and Ripepe 2011).

Figure 1.1.Instrumental response. Comparison of the response curves of short-period seismometers L4C and L22 (low cutoff frequencies of 1 and 2 Hz, respectively) and the broadband sensor CMG40T (lower cutoff frequency corresponding to a period of 60 s). For a color version of this figure, see www.iste.co.uk/lenat/hazards2.zip

1.2.3. Monitoring network

The design of a monitoring network must take into account many factors as well as technical, logistical and budgetary constraints (Wassermann 2012). The quality of observations will depend on the choice of the network’s geometry and the station sites. A single, well-located station makes it possible to monitor a dormant volcano in order to detect a possible reactivation phase. About one third of volcanoes currently monitored around the world are equipped with only one station. This device must be quickly completed as soon as the first signs of reactivation are detected. A minimum of four stations is necessary to be able to approximately locate seismic sources.

However, the widely held opinion is that at least six stations are required to set up a quality network, some of which should be equipped with three-component seismometers in order to identify the arrival of S-waves and constrain source depth. Depending on budgetary possibilities, a network may include a variable proportion of “short-period” stations and “broadband” stations. The reader will find more detailed recommendations on the setting up of seismic networks depending on monitoring goals in Moran et al. (2008a), Wassermann (2012) and Thompson (2015).

The spatial distribution of a volcano seismic network must include stations at different distances from the crater or caldera’s center, and avoid too-large azimuthal gaps. This distribution ensures a better accuracy of hypocenter locations. While stations near the crater can detect low amplitude earthquakes, remote stations are useful for identifying regional tectonic earthquakes. Remote stations can also become essential in case strong eruptions destroy nearby stations or saturate them. In addition, each station must be within the direct visual range of at least another station, a relay or the observatory when the remote transmission of data is performed by terrestrial radio. Site choice must not only take into account noise levels, accessibility, the risks of vandalism, rockfall or flooding (Wassermann 2012), but also possible site effects which amplify seismic vibrations at certain frequencies (Mora et al. 2001). The quality of the facilities is also crucial (see Figure 1.2) and should ensure suitable equipment protection against often difficult environmental conditions (sun, rain, snow, wind, humidity, temperature, lightning, dust, ashes and rockfall, animals, etc.). Seismometers are generally buried in order to reduce the noise level and for them to be thermally insulated. It is advisable to oversize the power supply (photovoltaic panels and batteries) in order to avoid any kind of power cut.

Volcanological observatories are the places where data are received, saved, analyzed and interpreted (see Figure 1.3). Depending on the country, an observatory may be responsible for monitoring a single or a large number of volcanoes. In the latter case, some volcanoes can be very far from the observatory. Depending on the activity and risk levels of monitored volcanoes, the observatory can organize a permanent watch: daily and nightly. The first processing and visualization of these data need to be performed in near real-time. The data flow to be processed in an observatory is profuse, since recordings are made in continuous mode. They have to be stored on a long-term basis with a view to later in-depth analysis and possible availability for specialists via databases (IRIS, RESIF, ORFEUS, etc.). Observatories have thus entered the “Big Data” era and automated processing techniques are being increasingly used. Specialized software is used for performing routine tasks (see Box 1.1).

Figure 1.2.Seismic monitoring station of Popocatépetl volcano, Mexico, located at an altitude of 4,200 m. Instruments are protected by the hard shelter. Data are transmitted to the observatory via the visible antenna in the picture (photo courtesy of Cenapred)

Figure 1.3.Monitoring room of Pasto Volcanological and Seismological Observatory, Colombian Geological Survey. The screen wall displays the main information and signals from volcanoes in southern Colombia, including Galeras volcano. Operators keep a permanent watch and carry out routine data processing (photo courtesy of SGC-OVSP)

Box 1.1.Software, tools and websites used in volcano seismology

Seismic data acquisition and processing systems

Seisan: http://seisan.info/.

Earthworm: http://folkworm.ceri.memphis.edu/ew-doc/.

SeisComP3: https://www.seiscomp3.org/.

Antelope: https://brtt.com/software/.

Seismogram display and analysis

Swarm: https://volcanoes.usgs.gov/software/swarm/index.shtml.

SAC: https://ds.iris.edu/ds/nodes/dmc/software/downloads/sac/.

PITSA: https://seiscode.iris.washington.edu/projects/pitsa.

Waves: https://www.src.com.au/downloads/waves/.

Management of monitoring data and networks

Webobs: https://ipgp.github.io/webobs/.

Lakiy (Cadena-Ibarra and Meneses-Muñoz 2018).

Seismic signal processing

Integrated software include many seismic array processing tools.

Geopsy: http://www.geopsy.org/.

Interactive Matlab software for the analysis of seismo-volcanic signals.

Seismo_volcanalysis (Lesage 2009): [email protected].

Identification of families of similar earthquakes (multiplets).

GISMO (Matlab): https://github.com/geoscience-community-codes/GISMO/wiki/What-is-GISMO%3F.

REDPy (Python): https://github.com/ahotovec/REDPy.

Calculation of seismic noise correlation and velocity variations

MSNoise: http://www.msnoise.org/.

Seismological code library in Python

ObsPy: https://github.com/obspy/obspy/wiki.

Automatic arrival time picking

APASVO: https://pypi.org/project/APASVO/ or https://github.com/cageo/Romero-2016.

Location of hypocenters

Hypo71, hypoinverse, hypoellipse, hypocenter: http://seis.geus.net/software/seisan/node65.html.

NonLinLoc: http://alomax.free.fr/nlloc/.

HypoDD: https://www.ldeo.columbia.edu/~felixw/hypoDD.html.

Seismic event detection and location: https://backtrackbb.readthedocs.io/en/latest/.

Databases on volcanoes and their activity

Global Volcanism Program Database: http://volcano.si.edu/index.cfm.

Online resources for volcanology research and risk prevention: https://vhub.org/.

Database on volcanic activity and volcanic hazard and vulnerability.

Volcano Global Risk Identification and Analysis Project (VOGRIPA) website: http://www.bgs.ac.uk/vogripa/index.cfm.

Database on Volcanic Unrest: www.wovodat.org.

World Organization of Volcanological Observatories (WOVO): https://wovo.iavceivolcano.org/.

Seismological data (seismograms)

IRIS: https://www.iris.edu/.

ORFEUS: https://www.orfeus-eu.org/.

RESIF: http://seismology.resif.fr/.

1.3. Types of seismic-volcanic events

1.3.1. Introduction

“Volcanic earthquake” makes reference to an earthquake taking place within or nearby a volcanic structure (distance <40 km) and generated by its magmatic or hydro-thermal activity.

There are a large number of phenomena which can produce seismic vibrations and seismic waves on Earth. Whereas in a tectonic context, the predominant mechanism is fault rupture or slip, in volcanoes one can witness a great diversity of seismic events, triggered by different physical processes. This diversity is specific to the seismic activity of volcanoes. The identification, classification and interpretation of these various types of seismic events are among the central issues in volcano seismology. In this section, we will describe the main types of seismic-volcanic events, their temporal and spectral characteristics, as well as the various source mechanisms that have been proposed for each of them.

Table 1.1.Correspondence between the terminologies used for the classification of seismic-volcanic events. Based on McNutt (1996) and McNutt and Roman (2015)

Minakami (1960, 1974)

Latter (1979)/Latter et al. (1989)

McNutt (1996)

Chouet (1996a, 1996b)

Other names

A-type

A-type, HF, volcano-tectonic earthquake

High-frequency (HF)

Volcano-tectonic (VT)

Short-period earthquake, VT-A, VT-B

B-type

B-type, LF, volcanic earthquake

Low-frequency (LF)

Long-period (LP)

Long-coda,

tornillo

, micro-earthquake

–

C-type, medium-frequency earthquake

Hybrid

Hybrid

Multiphase (MP), mixed-frequency

Explosion quake

E-type, broadband eruption earthquake

Explosion quake

Explosion quake

–

Volcanic tremor

Volcanic tremor

Volcanic tremor

Volcanic tremor

Harmonic, spasmodic, intermittent, eruptive

There are several classifications of seismic-volcanic events (McNutt 1996). The first classification, established by Minakami (1960, 1974), was based on phenomenological criteria. Then other classifications, based on temporal and spectral characteristics of signals (Latter 1979; Latter et al. 1989) or based on different mechanisms acting at the seismic source (Chouet 1996a, 1996b), were proposed. In addition to the main types of events defined in these classifications, various terms are used locally to describe events with particular characteristics. Table 1.1 presents the main types of earthquakes defined by the main classifications, as well as the correspondences between them.

1.3.2. Volcano-tectonic earthquakes (VT, A-type, high-frequency, HF)

Figure 1.4.Piton de la Fournaise volcano-tectonic (VT) earthquake (April 1, 2007 at 08 h 29 m 28 UT, station BOR, vertical component). The figure shows the seismogram, the spectrogram and the spectrum of the signal delineated by the blue dashed lines. Above the seismogram, enlargement of the arrival of the P-wave showing its impulsive character. For a color version of this figure, see www.iste.co.uk/lenat/hazards2.zip

This type of seismic event is produced by fault ruptures associated with variations in the stress field. Focal mechanisms are generally single double-couples, but sometimes have a volumetric component. Tensile rupture can also occur due to cooling and magma solidification, or due to the emptying of cavities. Seismograms are characterized by an impulsive first arrival of P-waves and by visible S-waves. Their spectra are wide and can include energy up to 10–15 Hz or higher (see Figure 1.4). This upper limit depends on wave attenuation during propagation, higher frequencies being attenuated more strongly than lower frequencies. As a result, at a great distance from the source, a volcano-tectonic (VT) earthquake can be confused with a “long-period” (LP) earthquake. Some observatories draw a distinction between deep (>2 km) and shallow (<2 km) earthquakes by calling them VT-A and VT-B earthquakes, respectively. This type of event is similar to tectonic earthquakes. They can therefore be studied with all the tools and methods used for tectonic seismology (hypocenter location, focal mechanisms, magnitude, seismic moment calculation, b-factor, etc.).

1.3.3. Long-period earthquakes (LP, B-type, low-frequency, LF)

“Long-period” earthquakes are often characterized by the arrival of emergent P-waves, the lack of distinct S-waves and a relatively narrow spectrum (see Figure 1.5). Energy is generally found in the 0.5–6 Hz band, with a frequent concentration between 2 and 3 Hz. Most spectra contain dominant peaks. For some volcanoes, LP earthquake subcategories have been defined. The most common and remarkable is the “tornillo”, whose seismogram looks like a screw (tornillo in Spanish). These are long-coda events, whose amplitude steadily decreases over time, and whose spectrum contains a single or small number of dominant peaks (see Figure 1.6).

LP events have small magnitudes and their sources are generally close to the surface (<2 km). However, deep LP earthquakes (10–60 km) have been detected on volcanoes such as Kilauea (Aki and Koyanagi 1981) or Pinatubo (White 1996) during their reactivation phases. In this case, they were associated with magma movements in the crust and/or magmatic chamber supply phenomena. Deep LP earthquakes are difficult to observe but can be considered as early precursory signs of eruption. In addition, immediately before some eruptions, a strong increase in LP seismic activity can be observed. When the intervals between events become shorter than the duration of earthquakes, these cannot be differentiated from one another. They merge into continuous vibrations called “volcanic tremor” and share the same spectral characteristics (see section 1.3.7).

Figure 1.5.Popocatépetl “long-period” (LP) earthquake (November 21, 2002 at 00:59 UT, station PPP, vertical component). The magnification of the signal onset shows the emergent character of the first arrival. For a color version of this figure, see www.iste.co.uk/lenat/hazards2.zip

Figure 1.6.LP tornillo event in Misti, Peru. For a color version of this figure, see www.iste.co.uk/lenat/hazards2.zip

LP events are generally interpreted as phenomena involving the presence of fluids within the structure. They are therefore specific to volcanoes. The main source mechanisms proposed in the literature are as follows:

– the resonance of a fluid-filled cavity (magma, gas, water), the cavity possibly being a conduit or magmatic reservoir, or a fracture (Chouet 1996a, 1996b, 2003);

– the brittle rupture of magma in a conduit or dome (Neuberg et al. 2006);

– a slow-rupture in unconsolidated shallow volcanic materials (Bean et al. 2013).

Resonance phenomena explain the presence of dominant peaks in LP earthquake spectra, each peak possibly being associated with a vibration mode within the resonant cavity. The frequency of these modes and their quality factor – which describes the velocity at which the vibration is attenuated – depend on the cavity’s geometrical characteristics, as well as on the mechanical properties of the fluid and of the solid host rock. Wave velocity to consider when calculating the frequency is that of waves propagating at the solid-liquid interface (tube or fracture waves) (Ferrazzini and Aki 1987). Although various resonance triggering mechanisms have been proposed, it is difficult to know which one is involved in each type of event. Among these mechanisms, one can mention the opening or closing of a valve in a pressure conduit (Waite et al. 2008), an unstable fluid flow (Morrissey and Chouet 1997) or the interaction between magma and hydrothermal fluids, which can generate boiling or an explosion at depth (Matoza and Chouet 2010; White and McCausland 2019). Strombolian or Vulcanian explosions can produce signals similar to those of LP earthquakes, suggesting continuity between the two phenomena.

Magma is a complex fluid whose rheological behavior depends on composition, gas and solid crystal content, as well as pressure, temperature and deformation conditions. When magma is sufficiently degassed (i.e. at shallow levels), and suffers a high rate of shear deformation, it can experience a ductile-brittle transition (Tuffen et al. 2003). These conditions can be reached near conduit walls during magma ascent. After fracture, the material can heal and experience a repeated fracture, triggering crack resonances and, consequently, an LP event.

1.3.4. Hybrid earthquakes (C-type, multiphase, MP)

As the name suggests, hybrid earthquakes combine characteristics from VT and LP events. While the P-wave onset is impulsive, the S-wave arrival cannot be clearly identified. The spectrum at the start of the seismogram contains high frequencies like a VT event, whereas the coda is low frequency, resembling that of an LP event (see Figure 1.7). It is sometimes difficult to tell the difference between an LP earthquake and a hybrid event because the duration and amplitude of the high-frequency section are variable. Some authors consider that an earthquake is only hybrid if it presents mixed first-arrival polarities, that is to say, positive and negative arrivals depending on the station (Lahr et al. 1994; Chouet and Matoza 2013). This characteristic is consistent with a failure mechanism inside or near the conduit, which makes it resonate. Other researchers explain hybrid earthquakes as a result of dome growth-related shallow sources (Miller et al. 1998).

Figure 1.7.Hybrid earthquake of the Spurr volcano, United States, and its spectrum. Modified from Power et al. (2002)

1.3.5. Explosions

An explosion is the sudden release of overpressurized gas and solid or liquid material (magma, rock blocks, ashes). Depending on the amount of gas involved, initial overpressure, fragmentation mechanisms, emission’s duration and intensity, and characteristics of the eruptive vent, there is a broad diversity of volcanic explosions. The seismic signals produced by explosions also have varied features, in particular, their amplitude and duration. They have broadband spectra, sometimes with very low-frequency oscillations (see VLP events below), and sometimes with high-frequency energy. Seismic records of weak or moderate explosions can be similar to LP earthquakes, especially those produced by explosions at depth. Explosions are nonetheless recognizable by the acoustic waves they produce and which propagate into the atmosphere. These sound or infrasound waves can be detected using acoustic sensors or seismic sensors, thanks to the coupling of the atmosphere with the ground surface. The infrasound waves produced by large explosions propagate across thousands of kilometers, making them detectable and easy to locate by global networks, such as the international monitoring system from the Comprehensive Nuclear-Test-Ban Treaty (CTBT) (Johnson and Ripepe 2011; Fee and Matoza 2013).

1.3.6. Very long-period (VLP or VLF) and ultra-long-period (ULP) events

Thanks to the use of broadband seismometers, very low-frequency seismic signals have been detected on some volcanoes. They are called “very-long-period” (VLP) events when they are in the frequency range of 0.01–0.5 Hz, or “ultra-long-period” (ULP) events for frequencies below 0.01 Hz, corresponding to periods of several minutes (Fontaine et al. 2019). High-frequency seismic waves – in particular those generated by explosions – are often superimposed onto VLP or ULP events. It is therefore necessary to filter records so as to clearly identify them (see Figure 1.8). Since these signals have large wavelengths, they undergo little distortion during propagation and that makes them observable in the near-field, thus making it easier to model their source. For example, the ground particle motion associated with the first arrival points in the direction of the seismic ray thus indicates the source’s direction. When the waveforms recorded by a properly configured seismic network are available, one can also calculate a seismic source model thanks to the inversion of the seismic moment tensor (Legrand et al. 2000; Chouet et al. 2003). This provides information on the source location and type (dike, sill, vertical force), as well as on the fluid volume involved (Chouet 1996).

Figure 1.8.Aso volcano VLP event, Japan (September 15, 1994 at 10:20 UT). a) Original seismogram and b) filtered between 10 and 30 s. Modified from Legrand et al. (2000)

ULP events can be associated with the pressurization of a shallow reservoir or a caldera formation, such as that of Dolomieu on Piton de la Fournaise in 2007 (Fontaine et al. 2019), Kilauea in Hawaii in 2018 (Neal et al. 2019) or an underwater volcano near Mayotte in 2018 (Cesca et al. 2020). VLPs are generated by fluid transfers in the magmatic or hydrothermal system. This can be the passage of hydrothermal fluids through opening and closing fractures, or the injection of magma and gas episodically forming a hydraulic plug and provoking pressure variations in the conduit (Ohminato et al. 1998). Other possible mechanisms are the coalescence of a bubble layer followed by the ascent of a gas pocket in the conduit, or the sudden depressurization of a magmatic column during a vulcanian or plinian explosion (Arciniega-Ceballos et al. 1999).

1.3.7. Volcanic tremor

A volcanic tremor is a vibration which can last for minutes, days or even become almost permanent. The event’s onset is emergent, its amplitude is constant or fluctuating and its spectral content is similar to that of LP earthquakes. Some authors classify LP events and volcanic tremors under the same class of low-frequency seismicity. Tremors are typically related to the seismic activity of volcanoes, although non-volcanic tremors also exist. Several categories have been established so as to account for the diversity of the temporal and spectral characteristics of volcanic tremors.

The spectra of harmonic tremors contain one or more evenly spaced dominant peaks. The frequencies of the “harmonics” are equal to integer multiples of the fundamental frequency. The number of harmonics may reach up to 12, as for example at Semeru volcano, Indonesia (Schlindwein et al. 1995), or even 30 at Lascar volcano, Chile (Hellweg 2000). The peak frequencies are not always stable and one can observe glidings or sudden variations in the fundamental frequency and its harmonics as on Arenal volcano, Costa Rica (see Figure 1.9) (Lesage et al. 2006). Conversely, the spectra of spasmodic tremors do not show dominant peaks and the time signals appear to be almost random. These characteristics are close to those of the eruptive tremor, directly associated with volcanic eruptions. Their spectra are broadband and contain high frequencies, provided that the recording station is close enough to the active crater (Cosentino et al. 1989). McNutt and Nishimura (2008) showed correlations between vibration amplitude and vent radius. Sometimes one can also establish relationships between this amplitude and the eruption plume’s height. One deep tremor was identified 40 km beneath Kilauea, Hawaii, by Aki and Koyanagi (1981) and attributed to deep magma transfer. This magma flux appeared as stationary and was independent from eruptions on the surface. In some cases, especially in hydrothermal systems, the vibration is intermittent and called banded tremor due to the seismograms’ appearance plotted on a drum recorder. Finally, “drum-beats” can be considered as an intermediate category between LP earthquakes and tremors. These are series of LP-type events identically repeated at longer or shorter and more or less regular intervals for hours or days (see Figure 1.10) (Iverson et al. 2006).

Figure 1.9.Harmonic tremor of Arenal volcano, Costa Rica (May 2, 1997 at 12:27 UT). Seismogram, spectrogram and amplitude spectrum of the section delineated by the vertical dashed bars. For a color version of this figure, see www.iste.co.uk/lenat/hazards2.zip

Although one or more interpretative models have been proposed for each tremor category, it is often difficult to determine which one best describes the physical process of a given event. As with LP events, all these models involve fluids but the tremor’s long duration requires triggering mechanisms to last as long as the tremor itself. Resonance phenomena in fractures or fluid-filled conduits are often mentioned to account for the presence of dominant peaks in the spectra. Nevertheless, having many spectral harmonics, harmonic tremors cannot result from simple resonances as those from an organ pipe. The most convincing interpretation is that of a “Dirac comb” effect which takes place when an impulse repeats itself at highly regular T-period intervals. The signal can be seen as the convolution product of the impulse by a T-period Dirac comb, whereas the spectrum is the product of the impulse’s Fourier transform by a 1/T period Dirac comb in the frequency domain. The number of harmonics directly depends on the stability of the repetition period. According to this principle, scientists have proposed a model comprising a fracture in the magma plug letting gas flow through at regular intervals (Lesage et al. 2006). Stability is produced by a feedback effect relying on the resonance of the underlying