Short and Long Term Volcano Instability Studies at Concepción Volcano, Nicaragua Jose Armando Saballos University of South Florida, [email protected]

Short and Long Term Volcano Instability Studies at Concepción Volcano, Nicaragua Jose Armando Saballos University of South Florida, Asaballo@Mail.Usf.Edu

University of South Florida Scholar Commons

Graduate Theses and Dissertations Graduate School

January 2013 Short and Long Term Volcano Instability Studies at Concepción Volcano, Nicaragua Jose Armando Saballos University of South Florida, [email protected]

Follow this and additional works at: http://scholarcommons.usf.edu/etd Part of the Geology Commons, and the Geophysics and Seismology Commons

Scholar Commons Citation Saballos, Jose Armando, "Short and Long Term Volcano Instability Studies at Concepción Volcano, Nicaragua" (2013). Graduate Theses and Dissertations. http://scholarcommons.usf.edu/etd/4757

This Dissertation is brought to you for free and open access by the Graduate School at Scholar Commons. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Scholar Commons. For more information, please contact [email protected]. Short and Long Term Volcano Instability Studies at Concepcion´ Volcano, Nicaragua

Jose´ A. Saballos

A dissertation submitted in partial fulﬁllment of the requirements for the degree of Doctor of Philosophy Department of Geology College of Arts and Sciences University of South Florida

Co-Major Professor: Charles B. Connor, Ph.D. Co-Major Professor: Rocco Malservisi, Ph.D. Sarah E. Kruse, Ph.D. Peter La Femina, Ph.D. Timothy H. Dixon, Ph.D.

Date of Approval: May 1, 2013

Keywords: Concepcion´ volcano, residual Bouguer anomaly, GPS baseline, volcano spreading, time series analysis, Nicaragua

To my son Armando Antonio, my wife Edith,

my parents Armando & Mar´ıa,

my sisters Carolina & Jacqueline, my brother Pabhel my nieces Helen, Martha, Mar´ıa del Pilar & Cristina. ACKNOWLEDGMENTS

I would like to thank my advisors Professor Charles B. Connor and Professor Rocco Malservisi, and the rest of my PhD committee for their great patience and guidance throughout my program at USF. Field assistance from Koji, Phable, Chuck, Laura, Lara, Sean, Adam, and Ometepe Ex- pedition people is highly appreciated. Dual-frequency GPS equipment borrowed from UNAVCO for several campaigns, and the differential GPS kit from CIGEO-UNAN-Managua made possible the success of my gravity and geodetic studies at Concepcion.´ I beneﬁted from invaluable lo- gistic support from Angelica,´ Allan, Hoffman, Antonio, Virginia, Emilio, Julio, Wilfried, Martha

N. and Rosario from Direccion´ de Geof´ısica de INETER. I also want to thank my lab mates and grad-colleagues: Leah, Mikel, Ana, Alain, Sophie, John P., Heather, Wayne, Mel, Koji, John O.,

Sylvain, Aurelie, Catie, Jacob and Ophelia. The co-authors of my research papers are thanked for their suggestions, ideas, and patience. Mike Ramsey from Pittsburgh University provided some

ASTER scenes, and the other ASTER imageries were acquired through the GEO-GRID project

(http://www.geogrid.org /en/index.html). Access to the Landsat data was possible through the Glo- vis server (http://glovis. usgs.gov/). Financial support from the USF Geology Department made possible my staying in the PhD program and in the US. Funding for ﬁeld research was also provided by the USF Geology Department and the Institute for the Study of the Latin America and the

Caribbean. Many other people in the Geology Department encouraged me to keep going and made my life easier with paperwork, and many other things, and I will be in debt to them, especially

Judy, Mandy, Mary, Lynn, Paul B. and Connie. Gracias a toda mi familia alla´ en Nicaragua, te amo hijo, tambien´ te amo mi nena. TABLE OF CONTENTS

LIST OF TABLES iii

LIST OF FIGURES iv

ABSTRACT vii

CHAPTER 1 INTRODUCTION 1

CHAPTER 2 GRAVITY AND GPS OF CONCEPCION´ VOLCANO, NICARAGUA 3 2.1 Introduction 3 2.2 Tectonic and Geological setting 4 2.3 Concepcion´ volcano 8 2.4 Gravity 10 2.4.1 Data and processing 10 2.4.2 Estimate of the bulk density of Concepcion´ volcano 11 2.4.3 Description of gravity anomalies 13 2.4.4 Modeling gravity data with GROWTH2.0 15 2.5 Geodetic GPS time series 18 2.5.1 Data and processing 18 2.5.2 GPS results 22 2.6 Discussion and Conclusions 26

CHAPTER 3 RELATIVELY SHORT-TERM CORRELATION AMONG DEFORMA- TION, DEGASSING AND SEISMICITY: A CASE STUDY FROM CON- CEPCION´ VOLCANO, NICARAGUA 32 3.1 Introduction 32 3.2 Concepcion´ volcano 33 3.2.1 2010 eruption 34 3.3 Data collection and analysis 35 3.3.1 Geodetic GPS data 35 3.3.2 Real–time seismic amplitude data 35 3.3.3 Remotely-sensed SO2 data 37 3.3.4 Data gaps 38 3.4 Results 38 3.5 Modeling 44 3.6 Discussion and Conclusions 47

i CHAPTER 4 LAHAR DELINEATION FROM SATELLITE DATA AT CONCEPCION´ VOLCANO 50 4.1 Introduction 50 4.2 Concepcion´ volcano 52 4.3 Slope stability of Concepcion´ volcano 54 4.3.1 Slope aspect of Concepcion´ volcano 56 4.4 Remote sensing of lahars 57 4.5 Satellite data and methodology 59 4.6 Results 61 4.7 Discussion 65 4.8 Conclusions 70 4.9 Recommendations 71

REFERENCES 73

APPENDICES 84 Appendix A Supplementary ﬁgures for Chapter 2 85 Appendix B Concepcion’s´ volcano gravity data 98 Appendix C PERL script to compute the complete terrain correction for gravity data 108 Appendix D Supplementary ﬁgures for Chapter 3 116

ABOUT THE AUTHOR End Page

ii LIST OF TABLES

Table 2.1 List of input/initial parameters used in GROWTH2.0 (Camacho et al., 2011). Table 2.2 summarized main results. 17

Table 2.2 List of output parameters from GROWTH2.0 (Camacho et al., 2011) using as input the data shown in Table 2.1. 18

Table 2.3 Concepcion´ volcano’s GPS stations’ coordinates in the WGS-84 reference ellipsoid. 24

Table 3.1 Parameters of the open pipe model (Bonaccorso and Davis, 1999) used to explain the geodetic GPS deformation observed at Concepcion´ volcano during the 2010 erupting phase. 47

Table 4.1 Reported lahars at Concepcion´ volcano since last decade 56

Table 4.2 Planimetric area inundated by lahars at Concepcion´ volcano. 66

Table B.1 Gravity data gathered at Concepcion´ volcano, Nicaragua, during campaigns carried out between 2007 and 2010. 98

iii LIST OF FIGURES

Figure 2.1 Location of Ometepe Island on the western side of Lake Nicaragua. 6

Figure 2.2 Residual gravity anomaly interpolated using the cubic spline algorithm draped atop of a shaded relief map of Ometepe Island. 14

Figure 2.3 3-D inversion of the gravity data shown in Figure 2.2 using the GROWTH2.0 model developed by Camacho et al., (2011). 19

Figure 2.4 3-D display of the location of the total anomalous center of masses obtained from the gravity data inversion using GROWTH2.0. 20

Figure 2.5 CON1-SINT baseline changes running NW-SE across the volcano shown in the inset. 25

Figure 2.6 MOYO-SABA baseline change running W-NE across the volcano. 26

Figure 3.1 Shaded relief map of Concepcion´ volcano. 36

Figure 3.2 Time series of the 2010 and 2011 data used in this study. 40

Figure 3.3 Periodograms. 42

Figure 3.4 Cross-Correlograms among all time series. 43

Figure 3.5 Schematic representation of the open pipe model, after Bonaccorso and Davis (1999). 46

Figure 4.1 Geometric relationships used by Iverson et al. (1998) to derive the semi- empirical relationships of a debris-ﬂow’s. 53

Figure 4.2 Shaded relief map of Concepcion´ volcano showing main towns. 55

Figure 4.3 Slope map of Concepcion´ volcano generated from a 20-m-resolution digital elevation model produced in 2004. 58

Figure 4.4 Comparison of different approaches to highlight debris ﬂows on Concepcion´ using the same Landsat 7 image acquired on January 27, 2001. 62

Figure 4.5 False color composite Landsat image, acquired on January 27, 2001 produced by the band combination R: band 4, G: ARVI. 64

iv Figure 4.6 Temporal evolution of the lahar deposits heading down to La Flor town from the western slopes of Concepcion´ volcano. 65

Figure 4.7 Temporal evolution of the lahar deposits of the ﬂows threatening Los Ramos town southeastern side of Concepcion´ volcano. 67

Figure 4.8 Temporal evolution of the lahar deposits north of San Jose´ del Sur town SSW slopes of Concepcion´ volcano. 68

Figure 4.9 Time series of inundated areas by lahar deposits extracted from satellite data. 68

Figure A.1 Simple Bouguer anomaly map computed using a density of 2500 kg m−3 used in gravitational spreading models (e.g. Borgia and van Wyk de Vries, (2003)). 85

Figure A.2 Simple Bouguer anomaly map computed using a density of 1539 kg m−3 provided with the 1-D Nettleton (Nettleton, 1939) and Parasnis’(Parasnis, 1997) methods. 86

Figure A.3 3-D inversion of the gravity data shown in Figure 2.2 using the GROWTH2.0 model developed by Camacho et al., (2011). 87

Figure A.4 3-D inversion of the gravity data shown in Figure 2.2 using the GROWTH2.0 model developed by Camacho et al., (2011). 88

Figure A.5 Time series for the three components of station CON1 for all the campaigns. 89

Figure A.6 Time series for the three components of station COS2. 90

Figure A.7 Time series for the three components of station MOYO. 91

Figure A.8 Time series for the three components of station SABA. 92

Figure A.9 Time series for the three components of station SINZ. 93

Figure A.10 Time series changes for the Longitude component of all stations around Con- cepcion´ volcano during April through July 2010. 94

Figure A.11 Time series changes for the Latitude component of all stations around Con- cepcion´ volcano during April through July 2010. 95

Figure A.12 Time series changes for the vertical component of all stations around Con- cepcion´ volcano during April through July 2010. 96

Figure A.13 Temporal changes of Latitude component of GPS station MANA between January 2007 and August 2010. 97

Figure A.14 Temporal changes of Longitude component of GPS station MANA between January 2007 and August 2010. 97

Figure D.1 Periodograms for GPS baseline changes during 2010 and 2011. 116

v Figure D.2 Periodograms for RSAM daily average during 2010 and 2011. 117

Figure D.3 Periodograms for SO2 daily average during 2010 and 2011. 118

Figure D.4 Scatter plots of the data used in this study. 119

vi ABSTRACT

Concepcion´ is the most active composite volcano in Nicaragua, and is located on Ometepe

Island, within Lake Nicaragua. Moderate to small volcanic explosions with a volcanic explosivity index (VEI) of 1–2 have been characteristic of this volcano during the last four decades. Although its current activity is not violent, its volcanic deposits reveal stages of violent activity involving

Plinian and sub-Plinian eruptions that deposited vast amounts of volcanic tephra in the Atlantic

Ocean. These observations, together with the 31,000 people living on the island, make Concepcion´

volcano an important target for volcanological research.

My research focuses on the investigation of the stability of the volcano ediﬁce of Concepcion,´

using geophysical data such as gravity, geodetic global positioning system (GPS), sulphur dioxide

(SO2) ﬂux, real-time seismic amplitude (RSAM), and satellite remotely-sensed data. The integra- tion of these data sets provides information about the short-term behavior of Concepcion,´ and some

insights into the volcano’s long-term behavior.

This study has provided, for the ﬁrst time, information about the shallow dynamics of Con-

cepcion´ on time scales of days to weeks. I furnish evidence that this volcano is not gravitationally

spreading in a continuous fashion as previously thought, that its bulk average density is comparable

to that of a pile of gravel, that the volcano ediﬁce is composed of two major distinctive lithologies,

that the deformation ﬁeld around the volcano is recoverable in a matter of days, and that the defor-

mation source is located in the shallow crust. This source is also degassing through the relatively

open magmatic conduit. There are, however, several remaining questions. Although the volcano is

not spreading continuously there is the possibility that gravitational spreading may be taking place

in a stick-slip fashion. This has important implications for slope stability of the volcano, and the

associated hazards. The factors inﬂuencing the long term slope stability of the volcano are still

vii not fully resolved, but internal volcanic processes and anthropogenic disturbances appear to be the major factors.

viii CHAPTER 1

INTRODUCTION

Volcanoes and volcanic eruptions are beautiful natural phenomena. Unfortunately, volcanoes also jeopardize the integrity of their surrounding landscape, and threaten human lives. Because of the fertile soils volcanoes produce, people settle around volcanoes to farm the land, often not knowing the hazards their volcano poses to them, and in some cases not even realizing they are living next to a volcano, rather than just a “mountain”.

The end product of volcanology is the hazard mitigation and prevention of disasters from processes associated to both active and inactive volcanoes. To reach that goal we ﬁrst need to understand what the volcano did in the past, what the volcano is doing now, and then we will be able to build models to attempt to foresee what the volcano may do in the future within certain levels of conﬁdence.

A single scientiﬁc technique is by no means capable of revealing all the processes operating in an active volcano. This is why volcanology is multidisciplinary. Advances in science and tech- nology have dramatically improved the way we look at volcanic processes, allowing us to greatly improve the physical and numerical models in volcanology at all levels.

Through the research presented in this dissertation, I contribute to the understanding of an active volcano (Concepcion)´ located on Ometepe Island, within lake Nicaragua, hosting more than

31,000 inhabitants. This research provides valuable information for volcanologists and decision- makers involved in volcanic crisis, and also demonstrates that there is still a lot to be done to clearly understand the complex dynamics of this volcano.

My research focuses on the interpretation of geophysical data (gravity and geodetic global positioning system, GPS) gathered at Concepcion´ volcano, Nicaragua, since the last decade. It begins by describing results derived from gravity and GPS data to look at the structure of the

1 volcano ediﬁce, and assess its short term surface response to the processes operating in its shallow plumbing system. Then, this dissertation describes in more detail what monitoring techniques such as GPS, real-time seismic amplitude, and sulphur dioxide emission detect when the volcano is in a moderate erupting phase, and what they record when the volcano is not erupting. A joint interpretation of these data sets is given together with a conceptual model that depicts a possible scenario (in a very simpliﬁed way) of the dynamics of the Concepcion’s´ plumbing system. Finally, this work ends by providing a chronology of the mass wasting processes triggered by rainfall during the last three decades at Concepcion´ volcano, which shed light on the picture of intermediate and long term stability of this volcano.

The pronoun “we” used in Chapters 2 and 3 reﬂects co-authors contribution on those manuscripts submitted for publication and in preparation. Chapter 2 is in press in the Geological Society of

America special paper volume entitled Understanding open-vent volcanism and related hazards, edited by W. I. Rose, J. L. Palma, H. Delgado Granados and N. Varley. Chapter 3 will be soon submitted to Bulletin of Volcanology, and chapter 4 is part of a manuscript in preparation that will be submitted to a scientiﬁc journal after lahar numerical modelling is incorporated.

2 CHAPTER 2

GRAVITY AND GPS OF CONCEPCION´ VOLCANO, NICARAGUA

2.1 Introduction

Volcanoes evolve depending on the nature of eruptions and eruptive products through time, the gravitational load of the volcano, and its interaction with the crust (Borgia, 1994; Borgia et al., 2000; Merle and Borgia, 1996; van Wyk de Vries and Borgia, 1996; van Wyk de Vries and

Matela, 1998). As the edifice of a volcano grows, the load of this edifice on the crust may cause the volcano to spread radially, and the resulting deformation affects hazards (e.g., potential for flank collapse and the location of potential flank eruptive vents). In this paper we use gravity and global positioning system (GPS) measurements collected at Concepcion´ volcano, Nicaragua to improve knowledge of its structure, and to investigate the possibility that the volcano is laterally spreading under its own load. Concepcion,´ along with Maderas volcano, forms Ometepe Island in Lake

Nicaragua (Figure 2.1). This small composite volcano is built upon sedimentary rocks and uncon- solidated sediments in the lake, and has been frequently cited as an example of a spreading volcano

(Merle and Borgia, 1996; Borgia et al., 2000; Borgia and van Wyk de Vries, 2003; Delcamp et al.,

2008). This potential spreading clearly has important implications for how volcanic hazards are assessed in populated areas on the island and possibly for all those living on the shores of Lake

Nicaragua.

Bulk density is a crucial parameter contributing to the potential for a volcano ediﬁce to spread

(Merle and Borgia, 1996; Borgia et al., 2000). Essentially, the higher the density of the volcano ediﬁce the more likely this load, added to the crust in a relatively short time, will cause isostatic adjustments. Isostatic adjustment, in turn, may enhance instability of slopes of the volcano or the response of the ediﬁce to subsequent intrusions (Borgia, 1994; Borgia and van Wyk de Vries,

3 2003). Gravity surveys are widely used to study the structure of volcanoes (Yokoyama, 1963;

Budetta et al., 1983; Rymer and Brown, 1986; Brown et al., 1987; Connor and Williams, 1990;

Rout et al., 1993; Camacho et al., 1997; Afﬂeck et al., 2001). In particular, gravity data provide direct information about volcano density (Finn and Williams, 1982; Brown et al., 1991; Deplus et al., 1995; Schulz et al., 2005; Cassidy et al., 2007), which is of fundamental importance for determining the potential for a volcano ediﬁce to spread laterally under its own load (Merle and Borgia,

1996; Borgia et al., 2000; Jordan et al., 2009). Similarly, GPS is one of the main tools for the study of volcano deformation (Dixon, 1991; Murray et al., 2000; Blewitt, 2007; Dzurisin, 2007), and offers the most direct evidence of whether a volcano is spreading and at what rate. Therefore, we undertook gravity investigations at Concepcion´ with the goal of investigating the structure and bulk density of the volcano, and GPS investigations in order to investigate the deformation (speciﬁcally the change in baseline length between geodetic GPS stations) over time.

In the following we describe the geologic and tectonic setting of Concepcion´ volcano with special reference to volcano spreading models. Then, we describe our new gravity map of Con- cepcion´ volcano and how these data are used to estimate bulk density of the ediﬁce and to reﬁne our understanding of the major structural features of the volcano. GPS results from campaigns in 2001–

2010 are presented and these position data analyzed on long-term (yearly) and short term (daily) timescales. If permanent deformation associated with gravitational spreading is occurring at a high rate, then lengthening of the baselines between GPS stations positioned around the volcano should be observed on a long-time scale. Instead, deformation is observed on short timescales (daily), and is associated with episodes of volcanic unrest, and is recoverable (resulting in elongation and shortening of baselines repeatedly).

2.2 Tectonic and Geological setting

Concepcion´ volcano is an ∼1600 m-high composite volcano (Figures 2.1, and 2.2). Concepcion´

is part of the Central American Volcanic Arc, the result of Cocos plate subduction beneath the

Caribbean plate. The volcanic arc is located within the Nicaraguan Depression, an Oligocene

- Pliocene half-graben that extends ∼500 km in length from northern Costa Rica to the Gulf of

4 Fonseca and has a width ranging from 40 km (northwest) to 70 km (southeast) (McBirney and

Williams 1965; Weinberg, 1992; Funk et al., 2009). Although there have been several different theories as to the tectonic signiﬁcance, geomorphology and history of the Nicaraguan Depression, we are here primarily interested in the stratigraphy of the Nicaragua Depression and the current tectonics of the margin at the latitude of Concepcion´ volcano.

The Nicaraguan Depression is ﬂanked to the northeast by the late Oligocene to middle Miocene volcanic arc (El Coyol Group volcanics; Ehrenborg, (1996)) and Nicaraguan highlands composed of pre-Mesozoic crust (Sundblad et al., 1991). The volcanic arc migrated to the Nicaragua De- pression during the middle Miocene (e.g., Ehrenborg, (1996); Plank et al., (2002); Carr et al.,

(2007); Saginor et al., (2011)) and resides there today. The depression is a half-graben with down- to-the-northeast displacement along margin-parallel (i.e., northwest-trending) normal faults along the southwestern margin. On the western edge of Lake Nicaragua, the boundary is composed of late Cretaceous to Miocene sedimentary and volcaniclastic rocks of the Sandino Basin that were folded and uplifted, and then faulted (Weyl, 1980; Weinberg, 1992; Ranero et al., 2000; Funk et al., 2009). The Pliocene El Salto Formation unconformably overlies the top of the Sandino Basin sequence (i.e., the middle Miocene El Fraile Formation, which is interbedded with the Tamarindo

Group volcanics). The upper section of the El Salto Formation is interbedded with the Pleistocene

Las Sierras Group volcanics, which represent the initiation of the ‘modern’ volcanic arc (Weyl,

1980).

The stratigraphic thicknesses of the Late Cretaceous to Pliocene stratigraphy are best estimated from well logs from offshore and onshore wells (Ranero et al., 2000). The thicknesses of these formations are dramatically thinner than reported by Weyl (1980 and references therein) based on mapping studies. For example, the Sandino Basin sequence is reported as ∼10 km in thickness by Weyl, (1980); however, the maximum thickness as measured in wells is 4.2 km (Ranero et al., 2000). The Pliocene El Salto Formation is thicker (>700 m) in offshore wells (Argonaut-

1 and Corvina-1 wells; Ranero et al., (2000)) than has been mapped onshore (100 m thickness;

Weyl, 1980). The Las Sierras Group has been mapped at a thickness of 680 m (Weyl, 1980).

Unfortunately, the thickness of Quaternary alluvium and lacustrine sediments in Lake Nicaragua

5 0 Honduras 0 0 0

8 CON1

2 ALT 1 LCO MOY Concepción SABA MOYO volcano SINT 0 0

0 ESQ COS1 2

7 SJS Panamá 2 1 JMFZ COS2 Maderas volcano 0

0 Quaternary volcanic deposits and 0

4 sediments 6 2

1 Rivas Formation: Late Cretaceous - Late Paleocene Brito Formation: Late Paleocene - Early Eocene 0 0

0 Normal fault 6

6 GPS site 5 Lake Nicaragua 2

1 Volcanic vent

SRFZ 0 0 0 8 4 2 1

0 5 10 15 20 Km 640000 650000 660000 670000 680000

Figure 2.1. Location of Ometepe Island on the western side of Lake Nicaragua. Concepcion´ volcano forms the NW part of the island. Names of major towns have been abbreviated as follows: LCO-La Concepcion,´ MOY-Moyogalpa, ESQ-Esquipulas, SJS-San Jose´ del Sur, ALT- Altagracia. Black squares depict the location of GPS monuments with their respective 4-character code. Red circles represent volcanic vents. Faults within the lake are from Funk et al., (2009). Inset: Location map of Concepcion´ volcano within Lake Nicaragua in Central America. Black triangles show the location of major Quaternary volcanoes along the Central America volcanic front. are not known. Various researchers have estimated or assumed the thicknesses of these deposits, with estimates ranging from ∼100 m to >1000 m (Swain, 1966; Elming and Rasmussen, 1997;

Borgia and van Wyk de Vries, 2003). The only direct measurement of Quaternary deposits in the

Nicaragua Depression is from the M16 exploratory drill hole at the Momotombo geothermal field, approximately 140 km NW of Concepcion´ volcano, that suggests 150 m of Quaternary sediment and ignimbrite that unconformably overlies Tertiary ignimbrites (van Wyk de Vries, 1993). We know of no wells in the Nicaraguan Depression that have reported well logs. Recent seismic reflection data collected in Lake Nicaragua indicate >20 m of sediment (Funk et al., 2009), but could not confirm total thickness of the sediments due to rapid attenuation of seismic waves.

The neotectonics of the Central American margin have been the focus of recent studies to investigate the kinematics of deformation of the forearc and volcanic arc. Geodetic studies indicate that the forearc is migrating from central Costa Rica to Guatemala at rates of 8–17 mm yr−1

(Norabuena et al., 2004; Correa et al., 2009; La Femina et al., 2009; Alvarado et al., 2011). Al- though it is debated how this forearc motion is accommodated in Nicaragua, it is clear that there are three active structural trends, 1) northwest-trending right-lateral strike-slip and normal faults;

2) northeast-trending left-lateral strike-slip and normal faults; and 3) north-trending normal faults and volcanic alignments.

Recent seismic reﬂection data collected in Lake Nicaragua imaged three faults that vertically displace the lake bottom: the San Ramon, Jesus´ Mar´ıa and Morrito fault zones (Funk et al., 2009).

The San Ramon fault zone, SRFZ, is a northwest-trending, down-to-the northeast, dip-slip normal fault that is located at the southwestern end of Maderas volcano and extends 25 km to the southeast

(Funk et al., 2009). The Jesus´ Mar´ıa fault zone (JMFZ; Figure 2.1) is a northeast-trending, irreg- ularly surfaced topographic high with ∼4-7 m vertical scarps (Funk et al., 2009). This structure is located directly southwest and aligned with an extended peninsula of Concepcion´ volcano, and a broader topographic high between Ometepe Island and the western edge of Lake Nicaragua (Funk et al., 2009). The Morrito fault zone is a 20 km long by 5 km wide northeast-trending depression with 5–7 m scarps located northeast of Ometepe Island (Funk et al., 2009).

7 In addition to these fault zones, a magnetic anomaly along the southeastern shore of Lake

Nicaragua has been interpreted qualitatively as the Lake Nicaragua fault zone (Funk et al., 2009), a northeast dipping depression bounding normal fault. Funk et al., (2009) propose that these faults are reactivated as dextral oblique-slip faults in the current kinematic regime to accommodate forearc motion, however, no historical seismicity is known to be associated with these fault systems.

Instead, there have been two recent earthquakes located on northeast-trending left-lateral strike slip faults northwest and southeast of Ometepe Island; the 1987 Mw 6.1 Lake Nicaragua and 2005

Mw 6.3 Ometepe earthquakes, respectively.

2.3 Concepcion´ volcano

The volcano is predominantly composed of clinopyroxene–plagioclase bearing andesite rocks found low and high on the slopes exposed in gorges. Basalts also occur and are found mostly on lateral vents and in some summit eruptions. A dacite pumice deposit is found at the north base of the volcano (McBirney and Williams 1965; van Wyk de Vries, 1993; Borgia and van Wyk de Vries,

2003). Lahars, pyroclastic units, and lava ﬂows are all common products of Concepcion´ (McBirney and Williams 1965; van Wyk de Vries, 1993; Borgia and van Wyk de Vries, 2003). Van Wyk de

Vries (1993) and Borgia and van Wyk de Vries, (2003) provide a detailed magmatic evolution of the volcanic system. Stratigraphically older magmas erupted from Concepcion´ are termed the

Quebrada Grande stage and are low-alumina (∼16 wt % Al2O3) and high-magnesium (∼8 wt %) basalts. Following the Quebrada Grande stage, magmas evolved to a more silica- and alumina-rich composition, culminating in the dacite amphibole-bearing Tierra Blanca tephra deposits (∼63-

66 wt % SiO2) (van Wyk de Vries, 1993; Borgia and van Wyk de Vries, 2003). The Tierra Blanca tephra is widespread throughout the region and is found in cores collected off-shore, in the Paciﬁc

Ocean. The Tierra Blanca tephra is dated at ∼19 ka (Kutterolf et al., 2008) and marks the most explosive (Plinian-style) activity known to have occurred at Concepcion.´ Following the Tierra

Blanca eruption, most tephra and lava compositions have been 50-55 wt % SiO2, although the top sections of the tephra erupted in 1957 had ∼62 wt % SiO2 (Borgia and van Wyk de Vries, 2003),

8 and eruptions from lateral vents generally have less silicic compositions than central vent eruptions

(Figure 2.1) (van Wyk de Vries, 1993).

A geodynamic model of Concepcion´ proposed by Borgia and van Wyk de Vries, (2003) suggests that spreading of the volcano ediﬁce began after the Tierra Blanca eruption, during the El

Mogote phase, when the volume of newly erupted products became sufﬁcient for the load of the ediﬁce to initiate spreading on the underlying ductile lake sediments. In their model this spreading involves a thicker portion of the sedimentary section, and even the magma chamber, over time.

As the increased load of the volcano ediﬁce drives spreading, deformation with spreading should be continuous (Borgia et al., 2000), but may be episodic if the Maxwell relaxation time is small compared to the characteristic spreading time (Borgia and van Wyk de Vries, 2003). Characteristic spreading time, T, is given by Borgia et al., (2000) as:

3µL2 T = 2 (2.1) ρgHvHd

where µ is the viscosity of the deforming ductile layer of thickness Hd beneath the volcano

of radius L, bulk density ρ and height Hv, and g is gravitational acceleration. Thus, a low bulk density and thin ductile layer will result in longer characteristic spreading time than a high bulk

density and thick ductile layer, for a volcano of a given geometry. Although the viscosity of the

ductile substratum layer can vary by several orders of magnitude, at Concepcion´ it is assumed to

be low because it is composed of soft clayey sediments (van Wyk de Vries, 1993; Borgia and van

Wyk de Vries, 2003). Van Wyk de Vries and Matela (1998) proposed a viscosity in the range of

1015−18 Pa· s.

N–S trending ﬁssures and aligned eruptive vents on the volcano ediﬁce may be a manifestation

of overall E–W extension associated with spreading (Borgia and van Wyk de Vries, 2003), although

N–S trending ﬁssures and vent alignments are ubiquitous on volcanoes in this part of the arc and

are generally attributed to extension within the dextral shear zone that deﬁnes Central America

fore-arc motion relative to the Caribbean, without requiring volcano spreading.

9 The most recent comparatively large-volume effusive eruption of Concepcion´ was in 1957

(volcanic explosivity index, VEI 2), and a minor effusive eruption took place in 1986 (VEI 1)

(McBirney and Williams 1965; Siebert et al., 2011). Ash explosions from small to moderate size occur frequently, characterized by VEI of 1 and 2 (Siebert et al., 2011). This activity is sufﬁcient to maintain a summit pyroclastic cone, degradation of which results in seasonal lahars that reach populated and agricultural areas at the base of the volcano.

2.4 Gravity

2.4.1 Data and processing

A total of 206 gravity readings were collected on and around Concepcion´ volcano during four different campaigns between 2007 and 2010, covering an area of approximately 18 km × 12 km on

Ometepe Island (Figure 2.2). The 2007 gravity campaign was carried out with a G-58 LaCoste &

Romberg instrument. The 2008 – 2010 campaigns were conducted with a Burris gravity meter (B-

38). During all surveys the observation points were positioned with a differential global positioning system instrument to achieve absolute vertical accuracies of at least 10 cm. The observed data were reduced using the standard procedure described by LaFehr, (1991) and Nowell, (1999), by applying corrections for: solid earth tide, instrument drift, geographical latitude, free-air, Bouguer slab (simple Bouguer correction or Bullard A correction), spherical cap (Bullard B correction), and terrain (complete Bouguer, or Bullard C correction). All gravity data are provided in Appendix B.

Terrain corrections were computed in three steps using a 20-m digital elevation model (DEM), provided by the Instituto Nicaraguense¨ de Estudios Territoriales (INETER). These three steps are based on the distance from the gravity observation point to points on the DEM grid (a copy of this code is provided in Appendix C). The inner zone correction accounts for topographic variation within Hammer’s zone C, 53.3 m, and was computed using the quarter-wedge method described by

Nowell, (1999), an improved version of the power-law approximation method of Campbell, (1980).

The intermediate zone correction was performed for DEM grid points that fall between Hammer’s zone D, >53.3 m, and outer radius of Hammer’s zone K, 9903 m. This terrain correction was done using the simpliﬁed gravity attraction of a prism approximated as an annular ring, described by

10 Kane, (1962). The far-ﬁeld terrain correction was performed for distances >9903 m and up to the extent of the input DEM, including all of the area of Ometepe Island. This far-ﬁeld correction was carried out by means of the vertical line mass approximation described by Blais and Ferland,

(1984), which is the approximation of the gravity attraction due to a prism in the far-ﬁeld. The lack of precise bathymetry data from Lake Nicaragua prevented implementation of a correction for the gravitational attraction from the water lake layer, but this effect must be small compared to the island terrain correction as lake depth is shallow, around 20 m in maximum depth in the region around Ometepe Island, and 43 m at its maximum reported depth in the central part of the lake

(Swain, 1966). The maximum terrain correction obtained using this method was 5.27 mGal, for a station located on the highest point along the ridge on the SW ﬂank of the volcano and located very near the San Jose´ del Sur, SJS, gully (Figure 2.1). Only points located on the slopes of the volcano have terrain corrections > 1 mGal.

A residual gravity anomaly was computed by subtracting the complete Bouguer anomaly (the anomaly obtained after the application of the terrain correction) from an assumed regional trend, estimated by ﬁtting a plane to the complete Bouguer anomaly map using the generalized least- squared method (Figure 2.2).

2.4.2 Estimate of the bulk density of Concepcion´ volcano

A primary goal of collection of gravity data on Ometepe Island is to estimate the bulk density of Concepcion´ volcano. Initially, the bulk density of the volcano ediﬁce was estimated using the

Nettleton (Nettleton, 1939) and Parasnis (Parasnis, 1997) methods, using gravity stations along two proﬁles that cross the volcano ediﬁce. These methods are based on the observation that complete

Bouguer gravity anomalies should be minimally correlated with topography in geologically homogeneous terrains. Ometepe Island is constructed predominantly of tephras and lavas in an area where lake sediments are dominated by volcaniclastic material (Swain, 1966), so this assumption is reasonable. The Nettleton and Parasnis methods yield a bulk average density of the volcano of

1439 ± 67 kg m−3 and 1539 ± 89 kg m−3 on each proﬁle. However, the resulting Bouguer anomaly maps are too positively correlated with topography (such densities produce Bouguer anomalies of

11 78 mGal in amplitude, -28 – 50 mGal, see Appendix A), indicating that these densities are lower than the actual bulk density of the terrain.

To improve the density estimate, we extended the Nettleton’s method from 1-D to a 2-D grid by encompassing the entire volcano ediﬁce (including many more observed points) and computing the correlation between the topography and Bouguer gravity anomaly, for a series of bulk densities.

The most likely bulk density of the volcano minimizes this correlation. The density thus obtained was 1764 kg m−3. We do not know the uncertainty in the model due to the real density hetero- geneities below the targeted area (i.e., the volcano edifice), but from the analysis of regression we estimate uncertainty to be ± 111 kg m−3. Thus, we deem the latter value as the uncertainty’s lower bounds for the bulk average density estimation. This density gives rise to a Bouguer anomaly amplitude of 49 mGal, -14 – 35 mGal. This bulk density is significantly less than values used in previous volcano spreading models (Borgia and van Wyk de Vries, 2003) but is consistent with the density of other volcanoes dominated by the accumulation of pyroclasts (Minakami, 1941; Brown et al., 1991; Affleck et al., 2001; Cassidy et al., 2007). Using a bulk density of 2500 kg m−3 (e.g.,

Borgia and van Wyk de Vries, 2003) produces a Bouguer anomaly map that is negatively correlated with topography (this yields a Bouguer anomaly of 94 mGal in amplitude, -44 – 54 mGal), indicating that 2500 kg m−3 is an overestimate of the density (see maps in Appendix A).

If we model the volcano ediﬁce of Concepcion´ as a frustum of a cone with a height of h

(1570 m, elevation difference between volcano’s maximum elevation an the elevation of its base), a radius of upper base r (150 m), and a radius of lower base R (3500 m), and using the bulk average density we have estimated from the gravity data (ρ = 1764 kg m−3), we can get an estimation of the mass of the volcano by the product of the density and volume (i.e., Mv = ρV). Thus, the formula we need to compute the mass of the volcano, Mv, is:

ρπh M = (R2 + rR + r2) (2.2) v 3

12 Plugging in the corresponding values we get a mass for Concepcion´ volcano of about 0.37 x

14 10 kg. Later, in section 2.6, we compare this mass, Mv, to the total anomalous mass below our surveyed gravity area that will be estimated using a 3-D inversion model.

2.4.3 Description of gravity anomalies

The distribution of gravity anomalies clearly reveals additional structure within the ediﬁce of

Concepcion´ volcano and on Ometepe Island (Figure 2.2). First, the gravity map of the volcano is dominated by a gravity low in the upper edifice of the volcano and comparatively high gravity values low on the edifice. The transition between the high and low values correlates with a slight break-in slope visible on the flanks of the volcano produced by the Tierra Blanca explosive phase, and the subsequent addition of pyroclastic material (Figure 2.2C). The low gravity values (< 10 mGal, Figure 2.2A) found on the upper slopes of the volcano are interpreted to reflect the low density pyroclastic cone that forms the summit area of the volcano. In contrast, the lower slopes of the volcano have a slightly higher density, resulting in a positive gravity anomaly.

Sharp gradients in the residual gravity anomaly map (Figure 2.2A) delineate faults that are partially and/or completely buried by young volcanic products. The low residual gravity anomaly

< -7 mGal trending N40◦W (parallel to the subduction zone in Nicaragua) on the NE side of

Concepcion´ (Figure 2.2A), may represent a fault, possibly an extension of the mapped faults on

Maderas volcano with about the same strike, and matches the fault partially mapped NNE of Con- cepcion´ by van Wyk de Vries (1993). If this anomaly is indeed associated with a normal fault, its hanging-wall block should be down-to-the-NE. The other low residual gravity anomaly SW of Concepcion´ volcano is more complex in geometry (Figure 2.2) and may be related to possible structural discontinuities of volcanic products within mudstones, which may also be associated with an extension of the SRFZ and JMFZ (Funk et al., 2009).

The profile P1P2 shown in Figure 2.2C suggests two main features. The first one is that the residual gravity anomaly is not flat as assumed in the model for estimating the bulk average density.

This does not imply that our volcano density estimation is wrong, since it minimizes the correlation between the Bouguer anomaly and the topography. But, this indicates that there are signiﬁcant

13 A 645000 B Elevation (m) C 1000 1200 1400 1600 200 400 600 800

P1 0 0040 0080 00 20 40 16000 14000 12000 10000 8000 6000 4000 2000 0 P 1 Topography Residual gravity Residual 1276000 1276000 Distancealong profile (m)

Residual Anomaly (mGal) P2 −21 −14 −70 7 14 21

660000

645000 14 Residual Anomaly (mGal) High: 24

Lake Nicaragua P −5 10 15 20 2 0 5 1264000 1264000 Low: -22 Residual gravity (mGal)

0 2.5 5 7.5 10 Km

Figure 2.2. (A) Residual gravity anomaly interpolated using the cubic spline algorithm draped atop of a shaded relief map of Ometepe Island. The residual gravity anomaly was computed by subtracting a linear regional trend from the complete Bouguer anomaly calculated using a density of 1764 ± 111 kg m−3, see text for details. Black circles represent gravity observation points. White circles represent gravity observation points used to computed the bulk-average density of the volcano by means of the 1-D Nettleton (Nettleton, 1939) and Parasnis’ (Parasnis, 1997) methods. The residual gravity anomaly map reveal concealed geologic structures of the volcano, and possible faults on the northeastern and southwestern sides. (B) Point color-map of residual gravity anomaly on Concepcion´ volcano. (C) Profile P1P2 comparing topography and residual gravity anomaly, see text for discussion. inhomogeneities (small spatial scale in the order of tens of meters) within the area used to compute the bulk average density, i.e., the volcano edifice and its nearby surroundings. The second important feature is that we can better appreciate the variations of the residual gravity anomaly with regard to the topography, mainly at the volcano edifice. The gravity residual anomalies < 10 mGal (Figure

2.2A) mark the transition from the high to low gravity anomalies and occur at elevations between

800 and 1000 m along the volcano proﬁle shown in Figure 2.2C.

2.4.4 Modeling gravity data with GROWTH2.0

We cannot build a direct subsurface structural density model of our study area from the gravity data since there is no detailed geologic information below our study area (i.e., stratigraphic data hundreds to thousands of meters below Concepcion´ volcano from cores, or other geophysical observations). Thus, a non-subjective inversion model is the best approach to model our gravity data for Concepcion´ volcano and investigate its subsurface structure. One of such a model is

GROWTH2.0 developed by Camacho et al., (2011). This is a non-subjective and almost automatic three-dimensional (3-D) model, which is based on the gravity inversion growing bodies model previously developed by Camacho et al., (2000, 2002). GROWTH2.0 is a non-linear inversion method that looks for the geometrical properties of the anomalous bodies ﬁtting the gravity observations

(Camacho et al., 2011). The major assumptions in GROWTH2.0 are that the subsurface anomalous structure is characterized by a prescribed density contrast (positive and negative values are accepted simultaneously), that the gravity data are unevenly spatially sampled and imprecise values, with a

Gaussian uncertainty, that the model parameters are also subjected to Gaussian uncertainties, and that the data contain a linear regional trend (Camacho et al., 2000, 2002, 2011).

The determination of the geometry of the anomalous bodies is carry out by dividing (seed- ing) the subsurface volume into discrete 3-D prismatic elements (cells) that are systematically and step-by-step tested for possibilities of model growth, in which “increased” prisms correspond to previously prescribed positive values of density contrast, “decreased” prisms correspond to negative values, and unchanged prisms correspond to null values. The problem of the non-uniqueness solution in the gravity data inversion is treated by GROWTH2.0 through a least square minimiza-

15 tion of the residuals (model ﬁtness), and the l2 minimization of the total anomalous mass (model smoothness; Camacho et al., 2000, 2002, 2011). Mathematically, the minimization algorithm of

GROWTH2.0 is expressed as (Camacho et al., 2011):

T −1 2 T −1 ν QD ν + λ f m QM m = min (2.3)

where ν is the matrix vector of residuals; QD is the covariance matrix of gravity data uncertainties, λ is a positive dimensionless factor to balance model ﬁtness and smoothness (by default λ =

30 in GROWTH2.0, for higher values of λ the inversion yields a simple model with a poor fit to the data, and for lower values the inversion produces a complex model of the subsurface structure with very good fit to the observations; an optimal value for λ yields a value of ' 0 for autocorrelation of residuals); f is a scale factor for fitting the modeled anomalies (initially each cell is assigned with a value of f ≥ 1, then during a subsequent iteration the value of f of a given cell decreases depending on its contribution to the model fitness, finally the growth criteria of the cell stops when

f reaches a value of ' 1); m is the matrix of the prescribed density contrast values; and QM is the uncertainty model parameters’ matrix (Camacho et al., 2011). Readers are referred to Camacho et al., (2011; and references therein) for further details of the GROWTH2.0 model.

We carried out inversion of our gravity data gathered at Concepcion´ volcano using the cited

GROWTH2.0 model (Camacho et al., 2011). We ran the model several times varying the input parameters. In Table 2.1 we have summarized three sets of input parameters that, in general, embrace the majority of the solutions that converge, and are also appropriate in our geological context. These set of parameters also yielded low values (close to 0) for the autocorrelation, a parameter related with the quality of the inversion based on the ratio between short distances among gravity observation points and data “noise” (Camacho et al., 2011), shown in Table 2.2. The homogeneity parameter used as part of the input to GROWTH2.0 (Table 2.1) is deﬁned by Camacho et al., (2011) as the factor controlling the homogeneity distribution of the density within and at the boundaries of the modeled anomalous bodies. High values of homogeneity (very close to, or

16 Table 2.1. List of input/initial parameters used in GROWTH2.0 (Camacho et al., 2011). Table 2.2 summarized main results. Set Depth to MBa Mean CLb No. of cells Initial DC rangec Homogeneityd (m) (m) (kg/m3) 1 -4883 601 8045 -400 – 400 0.5 2 -3000 500 9430 -300 – 300 0.2 3 -4000 600 7542 -400 – 400 0.6 aModel bottom (depth is negative). bInitial cells’ length of subsurface elements. cInitial density contrast range to start the model. dThis parameter takes values between 0 and 1, and deﬁnes density transition across anomalous bodies. See text for details. equal to 1) produce anomalous bodies with smoothed and diffuse boundaries, with extreme density values at the center of the bodies that decrease towards the edges (Camacho et al., 2011). A low homogeneity value (very close to, or equal to 0) yields anomalous bodies with abrupt boundaries, with homogeneous geometries, and with a single density value (Camacho et al., 2011).

In Table 2.2 we present the main outputs from the corresponding input parameters shown in

Table 2.1. obtained from GROWTH2.0. The maximum range of density contrast that best ﬁt the inversion model is -440 to 410 kg/m3, and the anomalous mass is around 1 x 1014 kg. Figure 2.3 depicts these results. In Appendix A (Figures A.3 and A.4), we show the ﬁgures resulting from the other two sets of parameters summarized in Tables 2.1 and 2.2. The three selected outputs from

GROWTH2.0 share great similarities. They represent Concepcion´ volcano as a cone dominated by a negative density contrast in its central and upper part. While its lower slopes and deeper parts are characterized by a positive density contrast anomaly, which implies that the lower slopes are denser than the upper part of the volcano (Figure 2.3). A signiﬁcant low density anomaly zone is located between the NE side of the volcano and the shoreline, trending NW-SE (Figure 2.3A). Whereas, on the SW side of the study area (Figure 2.3A) there is another low density anomaly zone, although more diffuse than that on the NE side. The cross-sections drawn by GROWTH2.0 and presented in Figures 2.3B and C show that the negative anomaly in the upper part of the volcano extends up to 1 km depth, while the higher density anomaly (that seems to compose the lower slopes of the volcano) surrounds the lower density anomaly, and extends up to depth of around 4.5 km. The

17 Table 2.2. List of output parameters from GROWTH2.0 (Camacho et al., 2011) using as input the data shown in Table 2.1. Set Model fitted DCa Balance factor, λb Auto-correlationc TA massd (kg/m3) (x 1014 kg) 1 -440 – 410 36 -0.05 1.16 2 -440 – 410 44 0.05 1.02 3 -370 – 310 36 -0.05 1.16 Location of center of total anomalous masse Set Easting (m) Northing (m) Depth (m) (m) (m) 1 650900 1276129 -2159 2 650010 1275368 -1697 3 650600 1275880 -2057 aDensity contrast range fitted by the model. bOptimal value of the balance factor found by the model. cSee text for explanation. dTotal Anomalous mass, i.e., cumulative of excess and deficit masses. eThe center of mass is virtually located below the volcano whose crater location is at 650500 (Easting), and 1275550 (Northing), see Figure 2.4.

other two models are very similar to that of Figure 2.3, but in those the higher density anomaly extends to greater depths below the volcano (see Appendix A).

2.5 Geodetic GPS time series

2.5.1 Data and processing

To assess the current deformation of Concepcion´ volcano’s ediﬁce, dual-frequency GPS data were collected at 6 different points on the volcano between 2001 and 2010. The ﬁrst 2 sites (CON1 and COS1, Figure 2.1) were installed during a campaign in late 2001. Three more sites (COS2,

MOYO, and SABA) were installed during the 2008 campaign (Figure 2.1). A water well was constructed in late 2009 a few meters from the COS1 site. Because of the potential for displacements associated with water pumping from the new well, in April 2010 a new GPS site (SINT, ∼2.5 km northeast of COS1, Figure 2.1) was installed to replace COS1. A week of data were collected

18 N W E

W E

650500 B Density contrast (kg/m3)

0 5 Km

S A 1275550 0 N S

0 -5Km -3 km C

Figure 2.3. 3-D inversion of the gravity data shown in Figure 2.2 using the GROWTH2.0 model developed by Camacho et al., (2011). The input parameters used in the modeling correspond to the ﬁrst set of parameters summarized in Table 2.1. (A) Structural density map at 1 km depth (-1000 m) of Concepcion´ volcano and its surroundings. (B) East-West cross-section of structural density map. (C) North-South cross-section of structural density map. simultaneously between the two sites to perform a tie and construct a common time series between

COS1 and SINT.

Data were collected in campaign mode (EGPS) in 2001, 2005, 2007, and nearly on an annual basis since 2008. Unfortunately, the two receivers used during the 2007 campaign had problems with ﬁrmware and provided results that are not reliable and with a large formal error in the measured position, a problem noted also for other sites where the instruments were utilized (H. Turner,

2007, personal communication). For completeness we present these data in the graph, but they are not used in the analysis. During all campaigns the geodetic markers were observed by mounting the antenna on ﬁxed height spike mounts to reduce uncertainties in vertical displacement. All the

19 VE: 2x1

< N

x 106

set 2

−1700

−1800

−1900 set 3

Depth (m) −2000 set 1 6.55 −2100 VE: 15x1 6.5 5 6.45 x 10 1.28 1.278 1.276 1.274 1.272 6.4 6 1.27 x 10 Easting (m) Northing (m)

Figure 2.4. 3-D display of the location of the total anomalous center of masses obtained from the gravity data inversion using GROWTH2.0. VE stands for vertical exaggeration. The set labels refers to the set of parameters listed in Tables 2.1 and 2.2. The center of the total anomalous masses (solid circles, Table 2.1) are located to depth between -1700 and -2200 m, and almost directly below the volcano. The total anomalous mass is ∼1 x 1014 kg (Table 2.2), and the mass of the volcano is ∼0.4 x 1014 kg (section 2.4.2), see text for details.

20 observations lasted at least 2 full UTC days to diminish the noise in the measured positions (in particular for effects due to troposphere, tide, and signal multi-path). The success rate in collecting the full day was signiﬁcant and we made an effort to analyze only ﬁles with at least 20 hours of continuous data collection. Every other epoch that for various reasons (e.g., power supply failure) had fewer data was discarded.

All GPS data were processed to produce non-ﬁducial daily solutions with the Jet Propul- sion Laboratory (JPL)’s software GIPSY-OASIS 6.1.2 using the standard precise point-positioning

(PPP) analysis strategy described by Zumberge et al., (1997) with non-ﬁducial orbits, clocks, and

Earth orientation parameters provided by JPL (ftp://sideshow.jpl.nasa.gov/pub/JPL GPS Prod- ucts/Final/). To improve the PPP of each daily solution we resolved phase ambiguity using the single receiver phase ambiguity resolution algorithm of Bertiger et al., (2010). The ﬁducial-free daily solutions were transformed into the International Terrestrial Reference Frame 2008 (ITRF08,

Altamimi et al., (2011)) through a seven-parameter transformation using parameters provided by

JPL. The ﬁnal product of the analysis consists of time series of the three components (E–W, N–S, and up–down) of the daily position for each site with respect to the ITRF08 reference frame.

The horizontal components of the daily solutions of our GPS sites at Concepcion´ were detrended using the data from the Continuously Operating Reference Stations MANA, located in

Managua, Nicaragua (∼95 km away from our network). This step was not implemented for the geodetic baseline computation as it does not require it. This approach will enhance the volcanic signal by minimizing most tectonic effects. The MANA site is about the same distance from the subduction trench as our network and is within the same tectonic block (Demets, 2001). The long term velocity of the MANA site was computed by applying the same PPP analysis previously described for EGPS sites to daily observations from the continuous site. To avoid co- and post- seis- mic effects related to the October 2004 earthquake in the Nicaraguan part of the subduction zone, only data after December 2006 were analyzed. Long term velocities of MANA in the North and

East direction with respect to ITRF08 were computed by linear ﬁtting of the latitude and longitude time series, respectively (9.5 ± 0.4 mm yr−1 for the north direction, and 7.8 ± 0.7 mm yr−1 for the east direction, velocity uncertainties were computed using the Hackl et al., (2011) algorithm). We

21 used these velocities to detrend the Concepcion’s´ network time series (see supplemental material provided in Appendix A for the resulting time series).

To better understand the deformation of the Concepcion´ volcano and to increase the signal to noise ratio we mainly analyze the change in length of two baselines across the volcanic ediﬁce.

The ﬁrst baseline runs in a NW-SE direction on the northeastern side of the volcanic ediﬁce and connects the two oldest sites CON1 and SINT (after tying COS1 with SINT, Figure 2.1 and 2.5).

The mean length of the baseline is 5,915.765 m and has a time span of observations of 8.625 yr. The time series for the CON1-SINT baseline length is presented in Figure 2.5. The second baseline runs across the volcano along an East-West direction and connects the sites MOYO and SABA (Figure

2.1 and 2.6). The mean baseline length is 11,146.439 m and it covers 1.882 yr; this time series is plotted in Figure 2.6. Both baselines cover time with unusual seismic activity near Concepcion´

volcano as reported by INETER (INETER, 2002c; INETER, 2002b; INETER, 2002a; INETER,

2002e; INETER, 2002d; INETER, 2003b; INETER, 2005b; INETER, 2005a; INETER, 2008a)

and documented in the scientiﬁc literature by French et al., (2010). Such periods are indicated

by light-gray vertical bars in Figures 2.5, 2.6 and those in the supplemental material (Appendix

A). They also cover time of volcanic activity consisting of the emission of volcanic ash through

small to moderate explosions (VEI 6 2), which have been indicated by dark-gray vertical bars in Figures 2.5, 2.6 and in the supplemental material. In particular the MOYO-SABA baseline time

series begins in mid August 2008 (Figure 2.6), 19 days after June 30 when a seismic event was

recorded at the Concepcion´ seismic station and the Civil Defense reported two ash explosions from

Concepcion´ volcano (INETER, 2008a).

2.5.2 GPS results

Table 2.3 presents the six stations’ coordinates we have observed around Concepcion´ volcano

as part of this study. We have subtracted from our GPS daily positions the effects of the surface

deformation (coseismic offsets) caused by the Mw 6.3 August 3, 2005 earthquake whose epicen-

ter was located about 10 km south of Maderas volcano. We used the fault geometry parameters

22 provided by French et al., (2010), and modeled as a rectangular dislocation embedded in a elastic half-space (Okada, 1985). These results are shown in Table 2.3.

Overall, the two baselines do not suggest long-term permanent deformation, at least within the detection limits of our data, described below. The maximum amplitude of baseline length change is observed in the CON1-SINT baseline between 2001 and 2005 period (Figure 2.5B), which is of about 40 mm. A weighted least-square line (weighted with regard to the error position) for this time window, 37 days, gives a velocity of 81.89 ± 147.01 mm/yr (Figure 2.5B). In principle, this could be compatible with episodic spreading of the volcanic edifice but it is very difficult to attribute it to a particular phenomenon (or phenomena) due to the large temporal gap and the significant seismic activity between the two GPS observation campaigns, and the July – September 2005 ash explosions from Concepcion´ volcano. In addition to the large Mw 6.8 October 9 2004 subduction earthquake that should not have affected our network, INETER (INETER, 2002c; INETER, 2002b;

INETER, 2002a; INETER, 2002e; INETER, 2002d; INETER, 2003b) documented seismic activity at Concepcion´ between June and October 2002, and April and June through August 2003 consisting of low-frequency volcanic tremor, and low magnitude earthquakes (i.e., ML ≤ 2.7, INETER,

2002a; INETER, 2003b). Furthermore, during the campaigns with longer period of observation

(CON1-SINT 2005 and 2010), the data contain large scatter, covering a range of deformation as large as the 2001-2005 signal (Figure 2.5). During the 2005-2010 period, within the scattering of the data the average length of the two baselines does not apparently change.

A similar situation was observed during the 2010 campaign when the volcano was erupting ash and emitting gas from March (INETER, 2010a) through May of that year. It is interesting to note that the only observations made during quite periods (2001 and 2009) seem to have a narrower distribution of the observed baseline length (note that these observations also correspond to short observations of only 2 full UTC days). A closer look at the data during the 2010 campaign (Figure

2.5C) indicates that the scattering of the data is not likely related to a purely stochastic processes, but instead reveals a very high temporal correlation. The data show periods of high frequency inﬂation- and deﬂation-like elastic deformation (∼30 mm in amplitude) in May 2010. A number of possible processes can cause this type of deformation, including volcanic or hydrothermal activity,

23 Table 2.3. Concepcion´ volcano’s GPS stations’ coordinates in the WGS-84 reference ellipsoid. Coseismic offsetsb Site ID Latitude Longitude HAEa North East Vertical (◦N) (◦E) (m) (mm) (mm) (mm) CON1 11.565 -85.625 261 -6.1 -4.7 -0.6 COS1 11.514 -85.603 238 -7.5 -7.1 -0.9 COS2 11.489 -85.649 58.4 3.6 -14.7 3.2 MOYO 11.538 -85.689 105.9 3.0 -10.5 1.4 SABA 11.548 -85.588 145.2 -12.4 -3.7 -2.3 SINT 11.529 -85.585 162.2 -13.1 -4.5 -2.7 aHeight above reference ellipsoid. bOkada’s (1985) rectangular dislocation in an elastic half-space model was used to compute the surface displacements. Fault geometry parameters were those from French et al., (2010), namely: epicenter Lat. = 11.34◦ N, epicenter Lon. = -85.50◦ E, fault centroid depth = 12 km, strike = 60◦, dip = 90◦, fault length = 16 km, fault width = 10 km, and slip = 0.9 m.

shrink and swell of the material below the GPS sites, and accumulation of volcanic ash on the antenna. We note that the volcano was erupting ash and gas during March through May in 2010, and so suggests that this deformation signal is directly related to volcanic activity. This non- permanent deformation is signiﬁcantly larger than the maximum deformation observed during the full study period and can lead to large changes in baseline length in the span of a few days. For this time period between late January and late July 2010, 182 days, a velocity of 5.54 ± 5.54 mm/yr is obtained from a weighted least-square straight line ﬁtting. While for the whole CON1-SINT baseline change time series, 8.6 yr from December 2001 to July 2010, the rate of change of this baseline is of 3.34 ± 0.81 mm/yr.

The time series of baseline change between MOYO and SABA stations (Figure 2.6) forms almost a E-W oriented baseline across the volcano (Figure 2.1, and inset in Figure 2.6). This time series begins in mid August, 2008, 19 days after a seismic event was recorded on the Concepcion’s´

seismic station, and Civil Defense reported two ash explosions from Concepcion´ volcano the same

day, on July 30 (INETER, 2008a). The scatter of the data seems to be less in periods with no

volcanic activity. This baseline shows an amplitude of almost 15 mm (a rate of change of -0.48 +-

0.48 mm/yr is estimated by ﬁtting a least-square straight line, Figure 2.6), which is about a half of

the amplitude of the change observed in the CON1-SINT baseline. MOYO station is about twice

24 A JAN 02 JAN 03 JAN 04 JAN 05 JAN 06 JAN 07 JAN 08 JAN 09 JAN 10 40 CON1-SINT 40

30 30 20 20

10 10

0 0 640000 650000 > −10 −10 N ____

1280000 " 1280000 −20 −20 CON1

velocity = 3.34 +− 0.81 mm/yr "

−30 −30 " " SINT Change in−40 baseline length (mm) −40 Change in baseline length (mm) " JAN 02 JAN 03 JAN 04 JAN 05 JAN 06 JAN 07 JAN 08 JAN 09 JAN 10

0 1 2 3 4 5 Km 1270000 640000 650000

B August 2005 C 31 7 14 21 28 4 JAN 10 FEB 10 MAR 10 APR 10 MAY 10 JUN 10 JUL 10 5 5 CON1-SINT 15 CON1-SINT 15 0 0 10 10 25 −5 −5 −10 −10 5 5 −15 −15 0 0

−20 −20 −5 −5

−25 −25 −10 −10 −30 velocity = 81.89 +− 147.01 mm/yr −30 −15 −15 −35 −35 velocity = 5.54 +− 5.54 mm/yr −20 −20 −40 −40 Change in baseline length (mm) Change in baseline length (mm) 31 7 14 21 28 4 JAN 10 FEB 10 MAR 10 APR 10 MAY 10 JUN 10 JUL 10 August 2005

Figure 2.5. CON1-SINT baseline changes running NW-SE across the volcano shown in the inset. Light-gray vertical bars indicate periods of unusual seismic activity on Concepcion´ volcano and/or near Ometepe Island, while dark-gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. Velocities were estimated by ﬁtting a weighted least-square straight line. (A) Full time series that includes data from all campaigns. Data from the 2007 campaign, gray squares, have been excluded from our analysis due to receivers’ ﬁrmware problems at that time. (B) Detail of the baseline changes during the 2005 campaign. (C) Detail of the baseline changes during the 2010 campaign. 640000 650000 > N ____ OCT 08 JAN 09 APR 091280000 JUL 09 " OCT 09 JAN1280000 10 APR 10 JUL 10

10 " 10 SABA " MOYO-SABA MOYO "

" 5 5

0 1 2 3 4 5 Km 1270000 0 640000 650000 0

−5 −5

−10 −10

velocity = −0.48 +− 0.48 mm/yr −15 −15 Change in baseline length (mm) Change in baseline length (mm) OCT 08 JAN 09 APR 09 JUL 09 OCT 09 JAN 10 APR 10 JUL 10

Figure 2.6. MOYO-SABA baseline change running W-NE across the volcano. Light-gray vertical bars indicate periods of unusual seismic activity on Concepcion´ volcano and/or near Ometepe Island, while dark-gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. Inset shows the location of MOYO and SABA sites. The rate of change (velocity) of this baseline was estimated by a weighted least-square ﬁtting. as far from the volcano than the other stations involved (Figure 2.1 and 2.6) leading to a baseline length twice longer than CON1-SINT baseline. Because we do not have baselines among other station-pairs, we cannot investigate whether the amplitude of the variations in baselines are also related to the distance from the volcano, or if it is simply a casual relationship. We did not compute baselines among other GPS station-pairs because the overlapping observation time were too short.

The change with regard to the ﬁrst measurement in all components for each of the ﬁve GPS sites around Concepcion´ volcano are included in Appendix A. The CON1 site shows the largest scatter during August 2005 and March–July campaigns in 2010, which were periods of comparatively intense seismic activity and volcanic ash eruption, and is the same pattern shown by the change in baseline from CON1 to SINT described above.

2.6 Discussion and Conclusions

New gravity data collected at Concepcion´ volcano allow us to estimate the bulk density of the volcanic ediﬁce to be 1764 ± 111 kg m−3. This value is signiﬁcantly lower than densities

26 of 2500 – 2700 kg m−3, previously assumed in spreading models for this volcano (van Wyk de

Vries and Borgia, 1996; van Wyk de Vries and Matela, 1998; Borgia et al., 2000; Borgia and van Wyk de Vries, 2003). A higher density favors gravitational spreading because the volcano will represent a bigger load resting upon weak lake sediments. Conversely, a lower density means less potential energy is available to deform and to induce flow of the substratum, and produce induced-gravitational spreading of the volcanic edifice. Geodetic GPS data do not show evidence of continuous spreading of the volcano over the time period of our observations. Nonetheless, because the density of a volcano is not the only factor controlling spreading (proper stress field condition, thickness and viscosity of the underlying substratum, magmatic intrusions, etc., see e.g., van Wyk de Vries and Matela, (1998), also play an important role), we do not rule out the occurrence of volcano spreading at Concepcion´ as an episodic or transient event possibly driven by magma intrusion, as already suggested by Borgia and van Wyk de Vries, (2003), or by the interplay with other factors such as regional and/or local seismicity.

Our longest time series baseline change that spans 8.6 yr, oriented NW-SE on the NE side of the volcano (Figure 2.5), indicates a rate of change of 3.34 ± 0.81 mm/yr. While the rate of change

of the baseline across the volcano in an E–W orientation shows a rate of change of -0.48 ± 0.48

mm/yr in 1.9 yr (Figure 2.6). This, together with the results described in the previous section,

suggests that the factors driven the baseline changes at Concepcion´ may be of different magnitude

at different locations. On the other hand, the non-corrected for the August 3, 2005 coseismic

offsets CON1-SINT baseline change between 2001 and 2010 (8.6 yr) has a rate of change of 3.01

± 0.82 mm/yr, which is slightly lower than the corrected version. This opens the possibility that

surface deformation induced by fault activity may be one of the factors inﬂuencing to the long-term

increase of the CON1-SINT baseline length. Could aseismic slip on nearby faults to Concepcion´

volcano have contributed to the increase in baseline length of CON1-SINT between 2001 and

2005 (Figure 2.5) that eventually ruptured on August 3, 2005 (main shock of Mw 6.3)? A more

dense spatial coverage of surface deformation measurement devices are required to investigate this

hypothesis further.

27 The deformation (tilting and faulting) of the deposits from the Tierra Blanca destructive phase, and younger deposits inclusive, mapped by van Wyk de Vries, (1993) and Borgia and van Wyk de Vries, (2003) opens the possibility that volcano spreading was at work at some time since that eruption. In a recent study based on experimental modeling, Galland, (2012) demonstrates that the complex pattern of surface deformation and its evolution in volcanic areas are strongly related with the shape of the intrusion and its subsequent evolution. Concepcion´ volcano lies next to a shear zone (the boundary of the Central American forearc sliver with the relatively stable Caribbean plate (Demets, 2001; Turner et al., 2007; La Femina et al., 2009)), and other fault zones within

Lake Nicaragua (Funk et al., 2009). In addition, the gravity map, the relationships between intrusions and surface deformation reproduced by Galland, (2012), and the volcanic activity from

Concepcion´ (indicating the presence of magma bodies below the volcano), all indicate the possibility that the deformation of the recent and older deposits from Concepcion´ may not necessarily be related to spreading of the volcano, but rather to a complex interplay of volcano-tectonic processes and features.

The gravity data also provide new information about the structure of the volcano, revealing two distinct lithologies within the ediﬁce, which we infer represents a change to a dominantly low-energy pyroclastic activity during the rapid reconstruction phase following the Tierra Blanca eruption, which corresponds to the second building phase of the volcano according to Borgia and van Wyk de Vries, (2003). Instead of a spreading-related thrust anticline, as pointed out by Del- camp et al., (2008), we infer that the linear residual gravity anomaly trending ∼N40◦W (Figure

2.2) on the northeastern side of the volcano is associated with a regional normal fault with its down-thrown block on the NE, partially buried by volcanic products, and whose northernmost end was mapped by van Wyk de Vries, (1993). If this is the case, this fault is associated with the

135◦–striking dextral transtensional fault zone below Maderas volcano described by Mathieu et al.,

(2011). This fault must be taken into account in the seismic hazard assessments of the Ometepe

Island and should be monitor to see if it presents any current activity. This fault may also have a component of right-lateral strike-slip displacement produced by the accommodation of the oblique convergence at the subduction zone in Nicaragua, similar to those found and described by Funk

28 et al., (2009), and possibly is a manifestation of tectonic escape, a product of the collision of the

Cocos Ridge in central Costa Rica (La Femina et al., 2009). Further research on both volcanoes may help differentiate between these models and better delineate the fault zone.

The low gravity anomalies found in the southwestern side of our study area (Figure 2.2) have a more complicated geometry that cannot be clearly discerned by our gravity survey due to our relatively low resolution sampling coverage in that area. However, surface faulting and structural discontinuities mapped within this area by van Wyk de Vries (1993); and Borgia and van Wyk de

Vries, (2003) are to a great extent correlated with the gravity anomalies. There is the possibility that these anomalies may be an extension of the SRFZ and/or JMFZ (Figure 2.1; (Funk et al., 2009)).

Both gravity data (this study) and geologic mapping (previous studies) show that the southwestern side of Concepcion´ volcano (Moyogalpa, Esquipulas, and Los Angeles, Figures 2.1 and 2.2) is the locality of small-scale complex structural features (e.g. local/regional faulting). More detailed geophysical study is required to elucidate the nature of these features, and thus estimate the hazard they represent to the people living in the island.

On the other hand, 3-D inversion of the gravity data seems to corroborate the statement that the upper part of the volcano is composed of lighter material than the rest of the volcano ediﬁce, Figures

2.2 and 2.3. This inversion also supports the existence of a possible fault on the NE side of the volcano, which is reﬂected in the residual gravity anomaly described in section 2.4.3. The inversion of the gravity anomaly yields a total cumulative anomalous mass of ∼1 x 1014 kg whose center of mass is located around 2 km below the volcano, while the volcano ediﬁce has an estimated mass of

∼0.4 x 1014 kg, using a bulk average density of 1764 kg m−3. The volcano’s mass represents ∼40

% of the total cumulative anomalous mass, indicating that most of the anomalous mass is concealed below the volcano.

When the GPS network was observed more regularly, as in 2005 and 2010, the volcanic edifice undergoes significant high-frequency recoverable deformation (and possibly noise in the signal is also affecting the measurements) that can lead to changes of the baselines up to 3 cm within a few days. This non-permanent high-frequency signal poses a significant problem in the evaluation of the long-term deformation of the volcano using episodic observations, as misleading interpretations

29 may result from samples collected for a short time and at large intervals. In our case, the amplitude of the high-frequency signal is comparable, if not larger than, the largest deformation measured, indicating that the trend observed between 2001 and 2005 may reﬂect poor temporal resolution. It is clear that to fully understand the deformation of the volcanic ediﬁce will be achieved only by deploying permanent or semi-permanent observation stations.

Borgia and van Wyk de Vries, (2003) carried out two GPS campaigns with dual-frequency instruments, the ﬁrst one in November 1994 and the second in May 1997. Their network consisted of

20 stations spaced between 2 and 5 km apart, and were located around the volcano’s base and half- way up the cone. A central station was used to measure the displacements of all other points. For baselines shorter than 10 km they used a “rapid static technique”, rather than a “static technique”, which is the technique that yields the highest accuracy (Dzurisin, 2007) and is the one we have used in our study. In general, the rapid static technique produces data uncertainty on the order of three times the error of static methods (Dzurisin, 2007), or approximately 1.5–2 cm on Ometepe Island.

Nevertheless, Borgia and van Wyk de Vries, (2003) interpret their results, especially the pattern of deformation, to be consistent with radial spreading of the volcano. However, this interpretation is complicated by several factors, including the data collection methods used, the possibility of base station motion during their survey, and the nature of high frequency movement we identify in our data sets. The 2001–2010 time series of baseline length indicate possible permanent change in baseline length over this total period of < 40 mm, or approximately 4 mm per year, a value similar to or less than short term recoverable (non-permanent) changes in baseline length. For example, short term recoverable deformation of 30 mm is observed in the time series in May, 2010 (Figure

2.5).

Based on the characteristic behavior shown by Concepcion´ volcano and herein described, we believe that Concepcion´ is predominantly an open vent volcano with a shallow magmatic reservoir, and whose upper conduit is eventually over-pressurized leading to small to moderate (VEI 1-2) ash and gas explosions. This is reﬂected in the relatively small amplitude variations in the GPS baselines, the short term recoverable deformation.

30 In conclusion, based on our observations we consider that Concepcion´ volcano is composed of two distinct lithologies (a low density upper cone surrounded by a denser structure that is perhaps the remnant of violent past eruptions), has a bulk density of 1764±111 kg m−3 and is not currently undergoing continuous gravitational spreading. A baseline oriented E-W across the volcano shows a zero length variation during almost 2 yr period of observations with episodic GPS campaigns style. Spreading is perhaps taking place in a different orientation across the volcano in an episodic fashion and is non-permanent, driven by factors other than gravitational loading (e.g., magmatic intrusion), although surface deformation induced by fault slip (and possibly strain accumulation by aseismic fault slip) may be inﬂuencing the increase in a baseline oriented NW-SE on the NE side of the volcano, which presents a rate of change of ∼3.34 +- 0.81 mm/yr during 8.6 yr of episodic

GPS observations.

31 CHAPTER 3

RELATIVELY SHORT-TERM CORRELATION AMONG DEFORMATION, DEGASSING AND SEISMICITY: A CASE STUDY FROM CONCEPCION´ VOLCANO, NICARAGUA

3.1 Introduction

During volcanic activity of 2010, surface deformation around Concepcion´ volcano, Nicaragua, was characterized by semi-periodic changes in baseline lengths measured across the volcano of up to ∼4 cm with period of several days to weeks. Modeling the source of this deformation may lead to improved understanding of how conduit processes govern eruption frequency and intensity in open magma systems, like Concepcion´ (e.g., van Wyk de Vries, 1993; Borgia and van Wyk de Vries,

2003; Saballos et al., 2013). Insight into the form of this model can be gained by cross-correlating time series of GPS measurements, real-time seismic amplitude measurements (RSAM), and SO2 ﬂux data.

Joint interpretation of surface deformation, SO2 ﬂux, and seismic data for periods of weeks to months provide a more integral view of volcanic processes than the analysis of any single data set (Casadevall et al., 1981; McGee and Sutton, 1994; Voight et al., 1999; Watson et al., 2000;

Williams-Jones et al., 2001; Olmos et al., 2007). With new advances in instrumentation, it has become routine to gather multi-variable measurements at high sampling rates (up to 1 Hz), allowing the study of short-term variations (seconds to hours) of the dynamics of magmatic systems associated with individual eruptions or episodes of volcano unrest (Fischer et al., 2002; Gotts- mann et al., 2007; Dalton et al., 2010; Kazahaya et al., 2011; Nadeau et al., 2011; Ichihara et al.,

2012). In this study we show that the comparison of daily observations of GPS geodetic measurements, RSAM, and SO2 ﬂux of intermediate length (∼2 mo sampling period during 2010) also exhibit highly correlated behavior during one eruptive period at Concepcion´ volcano. In contrast,

32 during a non-eruptive period (∼2.5 mo sampling period during 2011) of this volcano, signiﬁcant cross-correlations among these variables are not present. Thus, we suggest that these features provide important insights into volcanic processes on timescales of days to weeks during unrest at Concepcion´ volcano, and develop our model to attempt to explain magma movement on these timescales.

We also note that changes in baseline length, RSAM, and SO2 ﬂux were not of particularly large magnitude during the eruptive phase compared to the non-eruptive phase. That is, although anomalous behavior is observed in all three times series during the eruption, these anomalies become most clear by comparing the time series. This result is perhaps not surprising for an open-vent volcano.

However, this observation reinforces the idea that a continuous monitoring program of several geophysical time series, and near-realtime cross-correlation of these time series, might improve future responses to eruptive crises at Concepcion.´

3.2 Concepcion´ volcano

Concepcion´ is currently the most active composite volcano in Nicaragua, and is predominantly composed of clinopyroxene–plagioclase bearing andesite rocks (McBirney and Williams, 1965; van Wyk de Vries, 1993; Borgia and van Wyk de Vries, 2003). During the last four decades volcanic activity at Concepcion´ has been characterized by small to moderate explosions (VEI 1-2) of tephra and gas (van Wyk de Vries, 1993; Borgia and van Wyk de Vries, 2003; Siebert et al.,

2011). We note, however, that eruptive activity at Concepcion´ has not always been so mild. Plinian and sub-Plinian tephra fallout deposits are found in the late Pleistocene and Holocene stratigraphic record (McBirney and Williams, 1965; Borgia and van Wyk de Vries, 2003; Kutterolf at al., 2008).

Furthermore, past effusive eruptive activity has produced voluminous ﬂank lava ﬂows. The greatest concern for a large eruption of Concepcion´ lies in the effective evacuation of the island’s ∼31,000 inhabitants (INIFOM, 2001). In addition to hazards directly related to the volcanic activity, hazards associated with the remobilization of tephra fallout as lahars during subsequent rainfall are of great concern. These lahars, although small in volume, have damaged infrastructure and inundated agricultural areas (INETER, 2008b).

33 3.2.1 2010 eruption

According to reports from the Instituto Nicaraguense¨ de Estudios Territoriales (INETER), on

March 7, 2010 a new eruptive phase began that consisted of ash- and gas-laden explosions that rose up to 1 km above the crater (INETER, 2010b). From March 8 to March 17, INETER personnel observed and counted the number of visible ash and gas explosions that occurred during daylight.

They registered an average of 30 explosions/day, with a minimum of 14 explosions/day on March

8, and a maximum of 43 explosions/day on March 13 (INETER, 2010b). Fewer daily explosions were observed through March 24, then explosions became much more sporadic and the volcano was less regularly observed (INETER, 2010b). The Washington Volcanic Ash Advisory Center issued ﬁve reports of volcanic ash from Concepcion´ volcano detected by satellite. Two reports were issued on March 8 followed by single reports on March 9, March 12, and March 13, 2010

(http://www.ssd.noaa.gov/VAAC/ARCH10/archive. html#CONC).

Reports from INETER mention that ash explosions continued until mid–April (INETER, 2010a), and island residents reported eruptive activity until mid-May, 2010. We observed low intensity

(VEI 1) explosions during April 1 and 15 that deposited a very ﬁne ash layer (< 1 mm thick) on the volcano’s WSW ﬂank and in nearby communities. In general, during this period eruptions were characterized by an initial explosion, quickly followed by two or three additional explosions of smaller intensity at irregular time intervals. At the time of writing (March 2013) no new unrest have been reported from Concepcion´ volcano since mid–May, 2010.

Average seismic tremor frequency at Concepcion´ rose from a background of 0.5 Hz before

March 7 2010 to an average frequency of ∼3 Hz in the following two weeks. RSAM maximum daily average, 94, occurred on March 20 (INETER, 2010b). During the same period, INETER staff made mobile SO2 measurements with a Differential Optical Absorption Spectrometer, DOAS (e.g., Galle et al., 2003), along the main road on the N and W side of Concepcion´ volcano obtaining

SO2 daily ﬂuxes averages on March 9, and 13 to 15 of 133, 91, 507 and 127 tons/day, respectively (INETER, 2010b).

34 3.3 Data collection and analysis

3.3.1 Geodetic GPS data

We collected dual-frequency GPS data at two sites simultaneously during April 8 – June 25

2010, and February 4 – April 16 2011. These sites are located on the northern (CON1) and southeastern (SINT) ﬂanks of Concepcion´ volcano. Data collected at these two sites were used to compute a daily geodetic baseline across the northeastern side of the volcano (Figure 3.1).

GPS data were processed to produce non-fiducial daily solutions with the Jet Propulsion Lab- oratory (JPL)’s software GIPSY-OASIS 6.1.2 using the standard precise point-positioning (PPP) analysis strategy described by Zumberge et al., (1997) with non-fiducial orbits, clocks, and Earth orientation parameters provided by JPL (ftp://sideshow.jpl. nasa.gov / pub/JPL GPS Products /Fi- nal/) and FES2004 ocean loading (http://froste.oso. chalmers.se /loading/ /index.html). Ambiguity resolution was solved using the single receiver phase ambiguity resolution algorithm of Bertiger et al., (2010). The fiducial-free daily solutions were transformed into the International Terrestrial

Reference Frame 2008 (ITRF08, Altamimi et al., (2011)) through a seven-parameter transformation using parameters provided by JPL. The ﬁnal product of the analysis consists of time series of the three components (E–W, N–S, and up–down) of the daily position for each site with respect to the ITRF08 reference frame. Daily baseline length computation between CON1 and SINT sites

(Figure 3.2) were also performed with the GIPSY-OASIS software. During the full period, the average distance of the two sites was 5,915.765 m with a daily formal error of the order of 3 mm.

3.3.2 Real–time seismic amplitude data

RSAM data were collected using a broadband seismometer co-located with the CON1 GPS site. RSAM counts consist of the average of the absolute seismic signal’s amplitude calculated at a ten minute interval (Endo and Murray, 1991), which we averaged over a 24 hr interval to get

RSAM daily counts. In early 2011, a new seismic instrument was set up with a different gain setting at CON1 site (V. Tenorio, personal communication, 2011) producing a background level reading about ten times larger than the 2010 counts. To better compare the RSAM 2010 and 2011

35 648000

656000 N 1280000 X" CON1

! M o r r o 1276000 1276000 " ! SINT Jap ó n

14˚

1272000 0 5Km 1272000 Nicaragua 12˚ 656000 Costa Rica

10˚ Volcán Concepción

8˚ km 0 500

6˚ 270˚ 272˚ 274˚ 276˚ 278˚ 280˚

Figure 3.1. Shaded relief map of Concepcion´ volcano. Black squares depict the location of the dual-frequency GPS sites. The white “X” on the black square at CON1 shows the location of the only continuously operating seismic station at Concepcion.´ White circles show the location of the SO2 mini-DOAS instruments. Inset: Regional location map of Concepcion´ volcano, Nicaragua, in Central America. Black triangles show the location of major Quaternary volcanoes along the Central America volcanic front.

36 time series we divided the most recent RSAM counts by ten. This step does not affect the cross- correlation of the RSAM data with the other variables in any way.

3.3.3 Remotely-sensed SO2 data

SO2 measurements were made by the Differential Optical Absorption Spectroscopy, DOAS, technique (Platt and Stutz, 2008) using ground–fixed instruments developed under the Network for Observation of Volcanic and Atmospheric Change, NOVAC, project (Galle et al., 2010). One mini-DOAS instrument was installed in late April 2010 at a site ∼6 km southwest of the volcano’s crater (Japon,´ Figure 3.1). A second instrument was installed in March 2011 ∼2.5 km west of the volcano’s crater (Morro, Figure 3.1). The volcanic plume is scanned using a rotating mirror that redirects the light to a telescope with a focal length of 75 mm, which couples the transmitted light with two 16 mm optical fibers. The stray light is reduced using a Hoya U330 optical filter mounted on the telescopes. The optical fibers transfer the light to an Ocean Optics S2000 spectrometer. The rotation is made by a stepper motor with a resolution of 3.6 degree/step and covers 180 degrees along the horizon (Galle et al., 2010). The plume direction is calculated by analyzing the scan and determining the angle that corresponds to the center of mass of the plume. Given the characteristics of the plume from Concepcion´ (it predominantly bends over after leaving the volcano’s crater), the plume height is assumed to be the same as the volcano’s height (1600 m a.s.l). The plume speed is assumed to be the same as the wind speed at the plume height, and was obtained from the Global

Forecast System models data from the National Ocean and Atmospheric Administration. The mini-

DOAS instruments run during daylight; typically from 6: 00 to 18: 00 local time (GMT-6, thus within a single 24-hr GPS day). The maximum number of measurements within a day were 35, with a daily average of 5.

To be consistent with the temporal resolution of GPS data, we averaged the observations to daily ﬂuxes. The SO2 data were post-processed using the algorithm described by Platt and Stutz, (2008). We computed the percentage by which the standard deviation of the total measurements, carried out during a single day, corresponded to the average value of those measurements. We found that the standard deviation was around 34% of the daily average observed value (error bars

37 for SO2 time series in Figure 3.2). In nearly all cases, uncertainties estimated for a single SO2 measurement are smaller than 50% and greatest uncertainty is associated with change in weather conditions during the day.

3.3.4 Data gaps

In order to compute cross-correlation and periodograms, gaps in the data, days on which no measurements were obtained, were ﬁlled using the nearest neighbor, cubic splines, or piecewise cubic Hermite interpolation methods, depending on which interpolation method produced the closest variance to the original data set. For the 2010 period, there are no GPS data gaps. In 2011 there were ﬁve missing days (February 16, March 13, March 31, April 3, and April 13, open circles in

Figure 3.2) mainly related to missing GPS epochs in the daily data ﬁle due to power supply failure.

RSAM time series contain three missing days in 2010 (April 11, and May 10 and 11), and none in 2011. Gaps in the RSAM data were mainly related to problems with the telemetry system (V.

Tenorio, personal communication, 2011). 2010 SO2 data have gaps on April 30, and May 1, 8, 9, and 24 to 26. One daily gap occurred in 2011 (April 7). Gaps in the SO2 data were due to failure to detect the gas plume (i.e., clouds between the instruments and the volcanic plume were too dense) and/or to failure of the automatic instruments to resolve the plume geometry, a required parameter to compute the SO2 ﬂux. During 2010, only one mini-DOAS instrument was in operation, whereas two were during 2011. This fact accounts for the larger number of missing days in the SO2 data during 2010, as opposed to only one missing day in the time series in 2011.

3.4 Results

Three time series corresponding to each data type were correlated for GPS baseline, RSAM, and SO2 ﬂux during 2010 (April 8 through June 25 for GPS and RSAM cross-correlation, and April

28 through June 24 for SO2 with the GPS and RSAM data sets, Figure 3.2) and 2011 (February 4 through April 16). Time series for 2010 period were collected during the ﬁnal phase of eruptive activity.

38 For 2010, the change in GPS geodetic baseline length between CON1 and SINT sites (Fig- ure 3.1) has a total amplitude variation of 29 mm (-16 to 13 mm). This variation appears to be time-correlated and occurs progressively over periods of days to weeks, resulting in the broad ﬂuc- tuations about the mean baseline length (0 mm, Figure 3.2A). The maximum GPS baseline change took place on May 14, 2010 four days after island residents reported an explosion, the last explosion reported during the 2010 activity. On May 16 the baseline length returned to its average value, as does the RSAM (71 RSAM units), while the SO2 was slightly above its mean value (266 tons/day; Figure 3.2A). The pattern of variation of GPS and RSAM are very similar with a sharp

increase between May 5 and 14, and a sharp decrease in the GPS data between May 24 and 31,

2010 and a coeval increase in the RSAM signal (Figure 3.2A). The SO2 emissions began to increase on May 5 to reach its maximum (643 tons/day, or 7.44 kg/s) on May 10, the day of the last

reported explosion, then fall to background levels (∼200 tons/day) around May 17. Anomalously high values for change in baseline length, RSAM, and SO2 emissions occur within a time window between May 5 and 14 and are nearly in phase.

May 24 – 27, 2010, the GPS baseline contracts, then returns to average length by May 31

(Figure 3.2A). RSAM is elevated during this period, while the SO2 emissions remained at near average values (Figure 3.2A).

Overall, the daily average of SO2 ﬂux appears to correlate positively with positive changes in GPS baseline length. Fluxes larger than 300 tons/day correspond all to positive changes (extension)

in GPS baseline (Figure 3.2A). A similar pattern is revealed by comparing SO2 and RSAM in that

all SO2 ﬂuxes >300 tons/day occur when daily RSAM values were >77 units.

These relationships between SO2 and GPS, and SO2 and RSAM are not present in the 2011 time series data (Figure 3.2B). The amplitude in GPS baseline change in 2011 was 18 mm (-12 to

6 mm), which is 11 mm less than the 2010 counterpart (Figure 3.2A) and is characterized by ran-

dom variation about the mean, as opposed to the smoother variations observed during anomalous

episodes in 2010. In a similar way, the RSAM data collected in 2011 are characterized by random

variation about the mean. During the 2011 time series, the maximum daily average of SO2 ﬂuxes

39 April 2010 May 2010 Jun 2010 February 2011 March 2011 April 2011 A 8 11 15 18 22 25 29 2 6 9 13 16 20 23 27 30 3 6 10 13 17 20 24 B 4 7 11 14 18 21 25 28 4 7 11 14 18 21 25 28 1 4 8 11 15 15 15 10 10 5 5 0 0 −5 −5 −10 −10 −15 −15 −20 −20 GPS baseline change (mm) GPS baseline −25 100 change (mm) GPS baseline −25 100

80 80

60 60

40 40

MEP MCP 20 20 RSAM daily RSAM daily average RSAM daily RSAM daily average

800 0 800 0 700

40 700 600 600 500 500 400 400 (tons/day) (tons/day)

2 300

SO 200

SO 200 100 100 0 0 8 11 15 18 22 25 29 2 6 9 13 16 20 23 27 30 3 6 10 13 17 20 24 4 7 11 14 18 21 25 28 4 7 11 14 18 21 25 28 1 4 8 11 15 April 2010 May 2010 Jun 2010 February 2011 March 2011 April 2011

Figure 3.2. Time series of the 2010 and 2011 data used in this study. MEP stands for major extension phase in the GPS signal, while MCP major contraction phase shown in the GPS data. Closed circles represent observed data, while open circles interpolated data, see text for explanation. GPS baseline daily changes with regard to the first day on the time series. Error bars reflect formal error computed for a full observation day composed of measurements every 30 seconds. RSAM daily average of measurements every 10 minutes. SO2 daily fluxes. Error bars represent 34% uncertainty of the observed value for a full day composed of varying number of measurements, 5 per day in average, and up to 35 measurements maximum. (A) 2010 period when the volcano was in its most recent eruptive phase that begun on March 7 and extended up to around the end of May. (B) The 2011 data correspond a period when the volcano was not erupting. is 52% smaller (maximum daily flux was 309 tons/day, or 3.58 kg/s, Figure 3.2B), about 50% less than the maximum observed values during eruptive activity in 2010.

Additional structure in the time series is revealed by computing their periodograms (Figures

3.3A-C). Signiﬁcant autocorrelation exists in all three times series collected during the eruptive episode in 2010. The relatively broad spectrum of each series conﬁrms our qualitative interpretation that signal is found in the time series during the eruptive period, characterized by autocorrelation on timescales of 3 – 10 days, roughly, based on the decay observed in the amplitude of the periodograms at longer periods. In contrast, the periodograms for 2011 time series collected during a non-eruptive period are structureless; these periodograms are characterized by uniformity low amplitudes at each period. That is, during the 2010 eruption, variation in GPS baseline length,

RSAM and SO2 emissions contained trends over time that persisted for several days to about one week, whereas in 2011 the times series did not contain these trends.

Cross-correlograms provide further insight into the structure of the time series during the 2010 eruption and the 2011 non-eruptive period. The 2010 time series are characterized by correlation coefﬁcients that are signiﬁcantly different from zero (Figures 3.4A, C, E), whereas the 2011 non- eruptive period show poor or no correlation, except at lags 0 and 1 (Figures 3.4B,D, F). That is, for time series collected during the non-eruptive period in 2011, there is a slight correlation between

GPS baseline and RSAM on any given day, for example, but this correlation does not extend to subsequent days. During the 2010 eruptive period, signiﬁcant cross-correlation persists to lags of about ±5 days.

The largest correlation coefficients are observed between 2010 GPS baseline and SO2 emission at lag -3 days (Figure 3.4C) and SO2 emission and RSAM at lag zero days (Figure 3.4E). The maximum correlation coefficient at -3 days suggests that GPS baseline changes have a tendency to lead changes in SO2 emission during the 2010 eruptive period, although this tendency is quite subtle and may not be significant.

41 150 A GPS data, 2010 GPS data, 2011 100

|Power spectral density| spectral |Power 0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 600 Period (days) B RSAM data, 2010 RSAM data, 2011 400

200

|Power spectral density| spectral |Power 0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 4 x 10 4 C 3

2 SO data, 2010 2 1 SO data, 2011 2

|Power spectral density| spectral |Power 0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Period (days)

Figure 3.3. Periodograms. All data sets were linearly detrended before computing their periodograms, and then were smoothed by a Hanning window (e.g., Chatfield, 1996). (A) The 2010 GPS time series has significant autocorrelation with maximum at about 5-day period. (B) The 2010 RSAM data and (C) The 2010 SO2 reveal similar broad autocorrelation with maxima between 5 and 7 days. The periodograms of the 2011 data sets uniformity low amplitude power spectra compared to their 2010 counterparts, thus they do not have any meaningful autocorrelation and are best characterized by random fluctuation around mean values.

42 0.7 0.7 0.5 A 0.5 B 0.3 0.3 0.1 0.1 −0.1 −0.1 −0.3 −0.3 −0.5 GPS vs RSAM −0.5 GPS vs RSAM

rs−orlto coeff. Cross−Correlation −0.7 −0.7 −20 −15 −10 −5 0 5 10 15 20 −20 −15 −10 −5 0 5 10 15 20 0.7 Lag (days) 0.7 Lag (days) 0.5 C 0.5 D 0.3 0.3 0.1 0.1 −0.1 −0.1 −0.3 −0.3 −0.5 GPS vs SO GPS vs SO 2 −0.5 2

rs−orlto coeff. Cross−Correlation −0.7 −0.7 −20 −15 −10 −5 0 5 10 15 20 −20 −15 −10 −5 0 5 10 15 20 0.7 0.7

43 Lag (days) Lag (days) 0.5 E 0.5 F 0.3 0.3 0.1 0.1 −0.1 −0.1 −0.3 −0.3 −0.5 SO vs RSAM −0.5 SO vs RSAM 2 2

rs−orlto coeff. Cross−Correlation −0.7 −0.7 −20 −15 −10 −5 0 5 10 15 20 −20 −15 −10 −5 0 5 10 15 20 Lag(days) Lag(days)

Figure 3.4. Cross-Correlograms among all time series.√ Dashed horizontal gray lines represent the 95% conﬁdence interval, above or below which the correlation is statistically signiﬁcant (±2/ N, where N is the number of observed points). All individual data sets were linearly detrended before they were cross-correlated. (a) Cross-correlogram between the GPS and RSAM data for 2010. Positive correlation at positive lags means RSAM leads GPS baseline changes; positive correlation at negative lags means GPS baseline leads RSAM changes. (b) Cross- correlogram between the GPS and RSAM data for 2011; (c) CGPS and SO2 data for 2010; (d) GPS and SO2 data for 2011; (e) SO2 and RSAM data for 2010; (f) SO2 and RSAM data for 2011. 3.5 Modeling

We have shown that during the 2010 eruptive phase of Concepcion´ volcano, March to May

2010, the GPS baseline lengthens when the volcano was erupting ash and gases, and contracts when

the explosions ceased, Figure 3.2A. These changes in GPS baseline length were accompanied by

subtle but signiﬁcant changes in RSAM and SO2 emission. An important feature of the observed change in GPS baseline is that the average change returns to zero after these short-term (a few

days long) extension and contraction phases, implying that the volcano experienced a recoverable

deformation. Thus, GPS baseline extension represents inﬂation of the volcano ediﬁce (May 5 –

13, 2010; Figure 3.2), and baseline contraction represents deﬂation (May 23 – 29, 2010; Figure

3.2), compared to average conditions. This type of surface deformation has been observed during

small to moderate eruptions at other volcanoes, such as Mount Etna, Mount Saint Helens, Merapi,

Montserrat, Unzen, and others (Bonaccorso and Davis, 1999; Nishimura, 2009; and references

therein).

A simple model that explains these observations at Concepcion´ volcano is that of a magma

column connected to a shallow magma reservoir that is able to rise along the volcano conduit

(assumed to be cylindrical in shape) that is open at its top. As magma moves up toward the surface

a constant cylindrical dislocation of the conduit walls is produced because they are subjected to a

uniform pressure change (Figure 3.5), resulting in an inﬂation of the volcano and positive geodetic

baseline change measured across the ediﬁce. During an explosion, the pressure in the volcano

conduit is reduced, causing the walls to contract inward, and hence a volcano deﬂation measured

as a geodetic baseline contraction. This open pipe model is described in detail by Bonaccorso

and Davis (1999). The model relates the surface displacements, of any given point on the ground

surface, induced by the upward displacement (from depth c2 and c1, Figure 3.5) of a cylindrical magma pipe of radius α embedded in an uniform elastic half-space, which produces a uniform dislocation of the vertical conduit walls, b (Bonaccorso and Davis, 1999; Figure 3.5).

44 The equations to compute the horizontal and vertical displacements are as follow (Bonaccorso and Davis, 1999; Lisowski, 2007):

3 3 2 2 αbx c1 2c1(1 + ν) c2(1 + 2ν) + 2c2(x + y )(1 + ν) uE = 2 2 3 − + 3 2(x + y ) R1 R1 R2

3 3 2 2 αby c1 2c1(1 + ν) c2(1 + 2ν) + 2c2(x + y )(1 + ν) uN = 2 2 3 − + 3 2(x + y ) R1 R1 R2

2 2 2 αb c1 2ν 2νR2 − c2 uv = − 3 − + 3 (3.1) 2 R1 R1 R2

where uE and uN are the horizontal surface displacements in the easting and northing directions,

respectively; uv is the vertical surface displacement; ν is the Poisson’s ratio (taken to be equal to 0.25); x and y are the horizontal coordinates of the point of interest on the ground surface (e.g.,

2 2 2 2 2 2 2 2 easting and northing UTM coordinates); R1 = x + y + c1; and R2 = x + y + c2. To estimate the baseline changes for the periods of interest (i.e., May 5 – 13, and May 23 –

29 2010) using the open pipe model (Bonaccorso and Davis, 1999) we computed first the three components of the surface displacements with equation (1). Then, through a forward modeling approach we carried out a Chi-square goodness of fit to compared the open pipe model displacements with the observed ones provided by the GIPSY-OASIS software. The set of parameters given the best fit (the closest displacements from the model to the observations) were used to re-compute the

GPS baseline changes. The results are shown in Table 1, and essentially, they show that oscillation of the depth to the top of the magma column on order of 200 m can explain the observed changes in the GPS baseline for the periods of interest.

45 α

Level 2 c1

Level 1 c2

Figure 3.5. Schematic representation of the open pipe model, after Bonaccorso and Davis (1999), that we use to model the geodetic GPS deformation observed at Concepcion´ volcano during the inﬂation phase from May 5 to 14, 2010, and the deﬂation phase between May 24 and 31, 2010 (Figure 3.2). See text for details.

46 Table 3.1. Parameters of the open pipe model (Bonaccorso and Davis, 1999) used to explain the geodetic GPS deformation observed at Concepcion´ volcano during the 2010 erupting phase. Inﬂation phase between May 5 and 14, 2010. Dislocation Depth pipe’s top Depth pipe’s bottom pipe’s radius b (m) c1 (m) c2 (m) α (m) Average 4.8 378 1844 31 1σ 1.2 72 164 7 Deﬂation phase between May 23 and 27, 2010. Dislocation Depth pipe’s top Depth pipe’s bottom pipe’s radius b (m) c1 (m) c2 (m) α (m) Average 4.8 585 1838 43 1σ 1.2 129 191 7

3.6 Discussion and Conclusions

The correlations we have found in the 2010 time series (Figure 3.3 and 3.4) suggest that the changes in the GPS baseline, RSAM daily counts, and SO2 emissions during the eruptive phase may have shared a common source. Although the signal is weak in each time series, the cross- correlation is signiﬁcant. This interpretation is reinforced by the lack of such correlations observed in the time series collected in the 2011 non-eruptive period.

The 2010 time series, however, contains gaps - days during which no data were collected. This is a common problem with such time series collected on active volcanoes, and the signiﬁcance of these data gaps and the interpolation methods used to ﬁll the gaps to our interpretation warrants investigation. To test this, we replaced all the interpolated points with the minimum observed values

(i.e., 27 tons/day daily flux of SO2), and recomputed the cross-correlation. Even with this extreme interpolation method, significant cross-correlation among the time series persists. The main difference was a shift in the lag number of the significant cross-correlation, which we suggest indicates that the -3 day lag between SO2 emissions and GPS baseline changes may not be a significant feature of the time series. One addition test was performed to assess the robustness of our interpretation of the time series. As RSAM and SO2 values are always positive, we also computed the cross-correlation for the 2010 data sets with the absolute value of the GPS baseline changes. Again, significant cross-correlation among the time series persisted, although the correlation coefficient at zero lag slightly decreased.

47 As the 2011 time series were gathered during the dry season, we deem that scrubbing effects of

SO2 by interactions with hydrothermal ﬂuids are minimal, if it is taking place at all (e.g., Symonds et al., 2001). Part of the 2010 observations were made during the rainy season (May and June,

Figure 3.2A). However, SO2 remains correlated with GPS baseline changes and RSAM after the onset of the rainy season. Thus, if SO2 scrubbing is taking place it is possible that is not playing

a major role in changing the SO2 emission during our observation periods. Only a longer time

series, spanning several seasonal changes, is likely to quantify how much SO2 may be scrubbed by hydrothermal processes under Concepcion.´

The mild but frequent activity at Concepcion´ volcano during the last four decades is consistent

with the presence of a small-volume, shallow, and relatively open magmatic system (van Wyk

de Vries, 1993; Borgia and van Wyk de Vries, 2003). We were unable to observe the onset of

2010 eruptive activity on March 7. The ﬁrst set of anomalies observed between May 5 and 14,

2010 in the GPS time series (increase in baseline length of up to 12 mm), RSAM daily counts (a

maximum daily average increase of 89 units, 10 units lower than the maximum observed during

the onset of the eruption between March 7 and 9), and SO2 emission (a peak on the average daily ﬂux of 643 tons/day, more than %50 above background emission, Figure 3.2A) can be explained by

ascent of a magma column in the relatively open pipe. Modeling results suggest that the measured

12 mm extension between the two GPS stations is consistent with the bottom of the magma column

reaching a shallow magma reservoir at about 1.8 km depth, and its top ascending to about 370 m

depth during that period. The magma column that moved upward toward the surface (∼30 – 40 m

in radius) produced a cylindrical dislocation of the conduit walls of around 5 m (Table 1, Figure

3.5), which was observed as an inﬂation of the volcano ediﬁce. Increased SO2 emission during

this period is consistent with this model, through increase exsolution of SO2 from the ascending magma. RSAM counts during this time frame are also consistent with this model.

A few days later, May 24 – 31, the descending magma column reduced the pressure in the

volcano conduit, causing the walls to contract, which is inferred from the GPS record of deﬂation

(about 17 mm shortening of the GPS baseline). This deﬂation was accompanied by anomalous

seismic activity, reﬂected in daily RSAM counts of ∼80 units. No anomalous SO2 emission was

48 observed during this period (ﬂuxes < 200 tons/day), consistent with the model of magma withdrawal. The modeling suggests that for this event, the top of the magma column descended to a depth of almost 600 m. This magma withdrawal event was associated with the end of the volcanic crisis. The results of the open pipe model, showing a shallow magma reservoir below Concepcion´ volcano may be located at a depth of about 1.8 km, are consistent with the modeled depth range of the source of a negative gravity anomaly (∼1.7 – 2.1 km depth) seated below this volcano derived from the 3-D inversion of the Concepcion´ gravity data (Saballos, 2012).

Overall, these results suggest that much can be gained through a more robust multidisciplinary approach (i.e., the gathering of continuous geophysical data on dense networks), although we note that the comparatively small dimension of Ometepe Island imposes some constraints on spatial coverage of these networks. Such a network would, in turn, furnish vital information (e.g., magma intrusion linked to enhanced unrest visible at the surface, and retreat of magma linked to waning or end of volcanic activity) to decision–makers during volcanic crises. Because the magmatic system is relatively open and of small-volume, a single data set may not provide enough information to understand the processes taking place during intrusions, and some important details of the system dynamics may be overlooked.

49 CHAPTER 4

LAHAR DELINEATION FROM SATELLITE DATA AT CONCEPCION´ VOLCANO

4.1 Introduction

Lahars (volcanic debris ﬂows) are one of the deadliest volcanic phenomena, which have killed approximately 50,000 people in the circum–Paciﬁc region during historical times (Skermer and vanDine, 2005). The most recent infamous lahars in Latin America claimed the lives of ∼25,000 people in Armero, Colombia, from Nevado del Ru´ız volcano in 1985 (Lowe et al., 1986); 2,500 people in Chinandega, Nicaragua, from Casitas volcano in 1998 (Kerle and Oppenheimer, 2002); and 400 people in Panabaj, Guatemala, from Toliman´ volcano in 2005 (Connor et al., 2006).

Lahar generation is the product of a complex interplay of volcanic and climatic processes (Lav- igne, 1999). Lahars can be generated during volcanic eruptions (syn–eruptive or primary lahars) by incorporation of ejecta with water (Mastin and Witter, 2000), but can also occur long after eruptions (during periods of quiescence, post–eruptive or secondary lahars) by mobilization of volcanic tephra material during heavy rainfall, natural dam failures, or melting of snow and ice (Mastin and Witter, 2000; Vallance, 2005). Lahars can move as slowly as 1.3 m/s (∼5 km/h) or as fast as

40 m/s (144 km/h) with an incredible capacity of traveling very long distances from their sources

(Fisher et al., 1997; Scott et al., 2001; Vallance, 2005). A prehistoric lahar that descended from

Cotopaxi volcano, Ecuador, reached the Paciﬁc Ocean traveling more than 300 km and covering at least 440 km2 (Vallance, 2000).

Flow–routing algorithms are used to model the pathways of mass and energy through the landscape (Pelletier, 2008). There is no unique ﬂow–routing method for lahars because different con- stituents move through the landscape in different ways (Pelletier, 2008). Flow–routing models are the key tools to runout prediction of natural hazards, and so to determine the affected areas and

50 ﬂow intensity parameters, which are essential elements for producing hazard maps (Rickenmann,

2005). Traditional methods of assessing the hazard zones associated to gravity–driven ﬂows on volcanoes are based on reviews of historical records and ﬁeldwork to identify the limits of their deposits (Iverson et al., 1998; Stevens et al., 2003). Predictions of future hazard zones are then based on interpolation and extrapolation of known data, and researchers’ own experience (Iverson et al., 1998; Stevens et al., 2003). However, these techniques are less useful at volcanoes where little data on past activity are available, in which cases physical dynamic models and empirical– statistical approaches become more useful as predictive tools (Iverson et al., 1998; Stevens et al.,

2003; Rickenmann, 2005). Computer flow modeling is a powerful technique to delineate areas at volcanoes vulnerable to lahar hazards, and is achieved by calculating the predicted flow path and deposit inundation over a representation of the local topography in the form of a digital elevation model (Stevens et al., 2003). Digital elevation models (DEMs) are the core of flow–routing modeling, and are numerical representations of topographic maps. DEMs can be generated in several ways (Meisels et al., 1995; Stevens et al., 2003; Pelletier, 2008). For a full and detailed explanation of the different empirical relationships for debris flows the reader is referred to Rickenmann,

(1999); Rickenmann, (2005); and Pelletier, (2008).

Iverson et al., (1998) derived two semi–empirical equations to delineate lahar hazard zones in valleys that head on volcano ﬂanks. The rationale of the technique derives from scaling analysis of generic lahar paths and statistical analyses of 27 lahar paths at nine volcanoes, Figure 4.1. The two semi-empirical equations are (Iverson et al., 1998):

A = 0.05V 2/3 (4.1)

B = 200V 2/3 (4.2)

51 Where A is the valley cross-sectional area in m2, B is the planimetric area of the inundated valley in m2, and V is the lahar volume in m3. As the logarithmic relationship implies, these parameters are related by one order of magnitude in natural systems (Iverson et al., 1998).

To estimate the volume of the lahars from Concepcion´ volcano that have occurred during the last three decades, the planimetric area inundated by lahar deposits were ﬁrst extracted from remotely sensed satellite data (Landsat 2, 3, 5 and 7; and the Advanced Spaceborne Thermal Emis- sion and Reﬂection Radiometer (ASTER) sensors), and combined with ground truth data available for one lahar deposit (Los Ramos). Then, the volume of the deposits were estimated using Eq 4.2.

4.2 Concepcion´ volcano

The volcano is predominantly composed of clinopyroxene–plagioclase bearing andesite rocks found low and high on the slopes exposed in gorges (McBirney and Williams, 1965; van Wyk de

Vries, 1993; Borgia and van Wyk de Vries, 2003). Lahars, pyroclastic units, and lava ﬂows are all common products of Concepcion´ (McBirney and Williams, 1965; van Wyk de Vries, 1993; Borgia and van Wyk de Vries, 2003). Concepcion´ and the currently inactive Maderas volcano together form Ometepe Island, with approximately 31,000 inhabitants (INIFOM, 2001). Ometepe Island is located in Lake Nicaragua also known as Lake Cocibolca (Figure 4.2), in western Nicaragua.

Concepcion´ is the most active composite volcano in Nicaragua, characterized by effusive activity of basalt and basalt–andesite lava ﬂows, and more recently by small to moderate explosive activity. A large number of tephra explosions, VEI of 1 and 2, have been reported to have occurred during the last 100 years at Concepcion´ volcano (∼30) through the Global Volcanism Program and INETER websites, and lahars at Concepcion´ began to be systematically reported during the last decade, and appear to have increased in frequency, Table 4.1.

Debris ﬂows from Concepcion´ volcano have ruined crops, cut roadways, threatened the population and properties, and eroded or covered fertile soil, but fortunately have not caused casualties so far. However, there are a number of small towns, including schools and hotels, built upon prehistoric debris–ﬂow deposits mainly SSW from the volcano base all the the way to the shore lake, which represent a direct risk to about a half of the total population of the island, and an indirect

52 (A) volcano

(B) 9 10

8 Volcanic debris flows 10 Non-volcanic debris flows

7 10 ) 2

6 10

5 10 Planimetric Area (m 4 10

3 10

2 10 0 2 4 6 8 10 10 10 10 10 10 10 Volume (m3)

Figure 4.1. (a) Geometric relationships used by Iverson et al. (1998) to derive the semi-empirical relationships of a debris-flow’s channel cross-sectional area (A), and planimetric inundated area (B) to the debris-flow volume, (V); after Iverson et al. (1998). (b) Log-log plot of volume versus planimetric inundated area by debris flows from Iverson et al. (1998). The dashed lines represent the 95% confidence intervals of the regression given by equation 4.2.

53 risk of many more thousand people (and infrastructure) living in nearby cities and towns. Thus, it is extremely important to study the volcano instability of Concepcion,´ and its related volcanic hazards, in order to mitigate the effects of any potential major lahar from this volcano.

4.3 Slope stability of Concepcion´ volcano

In the 19th century, before the beginning of the reactivation of Concepcion´ volcano in 1883

(Siebert et al., 2011), vegetation on its flanks were abundant, and began to slowly decline as the volcanic activity increased in frequency. Erosion was accelerated by deforestation mainly from tobacco growers who used the wood from the flanks of the volcano to fuel ovens to treat tobacco leaves during the following decades (INIFOM, 2001). Thus, I believe that volcanic activity and anthropogenic disturbances taking place since the last century at Concepcion´ volcano should be the major factors that have undermined its slope stability. At Concepcion´ unstable material is mobilized by rainfall, and this happens when sufficient water saturates (or nearly saturates) the unstable soil mass increasing normal stress beyond the limits of the resistance imposed by shearing stresses (e.g., Iverson et al., 1997). Debris flows triggered by rainfall involve a number of physical processes operating on disparate timescales (Iverson, 2000).

The degree at which volcano activity is affecting the stability of Concepcion´ is not known yet, but I will briefly describe the most likely processes. Edifice volcano instability associated with geothermal systems develops when there is extensive rock dissolution and hydrothermal mineral alteration produced by highly acidic hot fluids, which are the result of the interaction of magmatic gases with groundwater (Lopez and Williams, 1993; Watters et al., 2000). Rock dissolution implies that the rock mass is weakened or softened leading to clay mineral formation and precipitation that may reach basal contacts, which ultimately favors mass sliding. Enhanced hydrothermal activity and strain within a saturated porous medium due to magmatic intrusions can increase pore water pressure, which in turn reduces the normal effective stress, increasing the potential for slope failures

(e.g., Voight and Elsworth, 1997). There are other effects associated with magmatic intrusions that have been associated with collapse of volcano slopes that I will not describe here since these are of a much bigger scale than the current laharic activity at Concepcion.´ I have observed hydrothermal

54 648000 SAM SJN ± 656000

1280000 CHI LAF ALT

LAC

SAB MOY

ESQ LOR 1272000 1272000

SJS 656000 0 5 Km Honduras 14˚

Nicaragua 12˚

Costa Rica

10˚ Volcán Concepción Panamá 8˚ km 0 500

6˚ 270˚ 272˚ 274˚ 276˚ 278˚ 280˚

Figure 4.2. Shaded relief map of Concepcion´ volcano showing main towns: San Marcos (SAM), La Chirca (CHI), San Jose´ del Norte (SJN), Altagracia (ALT), La Sabana´ (SAB), La Union´ (LAU), Los Ramos (LOR), San Jose´ del Sur (SJS), Esquipulas (ESQ), Japon´ (JAP), Moyogalpa (MOY), La Concepcion´ (LAC), an La Flor (LAF). Inset: Regional location map of Concepcion´ volcano, Nicaragua, in Central America. Black triangles show the location of major Quaternary volcanoes along the Central America volcanic front. Dashed-arrows indicate average flow direction of most significant debris-flows during occurrence.

55 Table 4.1. Reported lahars at Concepcion´ volcano since last decade Date Locationa Avg. thickness (m)b Reference 2002/10/28 La Chirca 1.5 INETER, (2002d) 2002/10/31 La Chirca 0.5 INETER, (2002d) 2003/11/25 San Marcos 1 INETER, (2003a) 2003/11/25 San Jose´ del Norte 2 INETER, (2003a) 2003/11/25 Japon´ 0.6 INETER, (2003a) 2004/05/21 La Chirca – INETER, (2004) 2005/05/18 La Flor – INETER, (2005d) 2005/05/18 San Jose´ del Norte 2 INETER, (2005d) 2005/06/29 La Concepcion´ – INETER, (2005c) 2005/06/29 La Chirca – INETER, (2005c) 2005/06/29 La Flor – INETER, (2005c) 2006/09/19 San Marcos – INETER, (2006) 2008/10/04 Los Ramos 5 INETER, (2008b) 2008/10/04 La Union´ – INETER, (2008b) 2008/10/05 Urbaite – INETER, (2008b) aSee Figure 4.2 for geographic location. bReported planimetric inundated area’s average thickness deposit.

altered rocks and soil mass high on the NW slope of Concepcion.´ These zone is the source region of the debris ﬂow that heads down to La Flor town, Figure 4.2.

Deforestation drastically decreases slope stability since trees provide mechanical stabilization through the roots (root cohesion to the soil enhances the shearing strength of the soil), and remove soil moisture trough transpiration reducing pore water pressure (Sidle, 1992; Nilaweera and Nuta- laya, 1999; Ammann et al., 2009). As tree roots decay in the soil mass, gaps in the interlocking root system of neighboring individual trees begin to form, increasing the likelihood for soil slides to occur (Sidle and Ochiai, 2006; Ammann et al., 2009). The weight of trees on soils, however, may not always be beneﬁcial, since this weight increases the normal and downhill force components, favoring slope failures as well (Sidle and Ochiai, 2006; Ammann et al., 2009).

4.3.1 Slope aspect of Concepcion´ volcano

To have a general idea of the slope conditions of Concepcion´ volcano as of 2004 I made a slope map from a 20-m-resolution DEM, constructed in 2004, and provided by Instituto Nicaraguense¨ de

56 Estudios Territoriales (INETER, Figure 4.3). Light-gray to white areas (slopes > 53◦, Figure 4.3) near the SW side of the summit, and at mid elevations on the NNW slopes, mark the locations of the deepest and largest gullies on the volcanic ediﬁce. A close look of the slope map of Concepcion´

(Figure 4.3) shows three distinct regions. The ﬁrst comprises the lower ﬂanks with slopes < 10◦ at

elevations < 300 m; the second with slopes between 11◦ and 27◦ at intermediate elevations (300 to

850–900 m); and a third region corresponding to the upper part of the cone with slopes > 28◦ that

goes up the summit at an elevation of about 1600 m. The second region matches to a great extent

the high residual gravity anomaly observed by Saballos et al., (2013), see also chapter 2 of this

work, distinguishing the denser lower ﬂanks of the volcano, believed to be the remnants of a series

of Plinian eruptions (Borgia and van Wyk de Vries, 2003) that occurred at Concepcion´ about 19 ka

(Kutterolf et al., 2008), from the upper lighter (third slope region, weaker and easier erodable) cone

that is associated to a low residual gravity anomaly, see chapter 2.

4.4 Remote sensing of lahars

A brief description of the main remote sensing techniques used to study debris ﬂows is given

here. The reader is referred to the bibliography cited in this section for further details. Currently,

there are more than 25 non-classiﬁed optical and radar sensors on board Earth orbiting satellites that

are useful in the monitoring and study of natural phenomena, including landslides and lahars (Kerle

and Oppenheimer, 2002; Pack, 2005). Ground or spatial resolution of these data range from 1 m

(IKONOS-2) to 1.1 km (NOAA-AVHRR); spectral resolutions go from the optical spectrum (∼0.4

– 0.7 µm) to microwaves (Radar X-band: 3 cm, C-band: 6 cm, L-band: 24 cm); and temporal resolutions (satellite revisit to the same location) from 8 to 16 days for high ground resolution

sensors, up to multiple overpasses per day for lower spatial resolution sensors (Flynn et al., 2001;

Kerle and Oppenheimer, 2002; Pack, 2005). Airborne (e.g., aerial cameras and Lidar) and ground

based (e.g., digital cameras and Lidar) remote sensing devices are also of great value in the study

of landslides and debris ﬂows, which in general offer a higher spatial resolution than satellite data

(Pack, 2005). Traditionally, optical remote sensing data have been the most common type of data

used in the study of lahar deposits, but near-infrared (NIR, ∼0.75 – 1.40 µm) data are extremely

57 (A)

Slope (deg) 0 - 1 2 - 3 1280000 P1 4 - 6 7 - 9 10 - 13 14 - 16 17 - 20 21 - 23 24 - 27 28 - 30

1275000 31 - 33 34 - 36

37 - 39 P2 40 - 42 43 - 44 45 - 47 48 - 50 51 - 52 53 - 55 1270000 ± 0 5 Km 56 - 58

640000 645000 650000 655000

1600 Residual gravity 20 (B) Topography 1400 15 1200 1000 10 800 5 600 0

Elevation (m) 400

200 −5

0 gravity (mGal) Residual 0 2000 4000 6000 8000 10000 12000 14000 16000 Distance along profile (m)

Figure 4.3. (A) Slope map of Concepcion´ volcano generated from a 20-m-resolution digital elevation model produced in 2004, provided by Instituto Nicaraguense¨ de Estudios Territoriales. See text for a detailed description. (B) Proﬁle P1P2 comparing topography and residual gravity anomaly, see chapter 2.

58 useful when lahars destroy vegetation cover (Francis and Wells, 1988; Kerle and Oppenheimer,

2002; Davila et al., 2007). A number of vegetation index algorithms have been developed since the late 1970s to characterize changes and conditions of vegetation cover traditionally using NIR and optical-red (∼0.625 – 0.740 µm) bands (e.g., Tucker, 1979; Huete, 1988; Kaufman and Tenre,´ 1992) that can be applied to study debris-ﬂow destruction and deposits. Also, digital elevation models (DEMs) derived from radar and Lidar data, and sensors with stereoscopy capabilities (e.g.,

ASTER) have been widely used to model debris ﬂows, develop more accurate hazard maps, and to study their deposits (Pack, 2005; Hubbard et al., 2007; Huggel et al., 2008; Vianello et al., 2009).

4.5 Satellite data and methodology

Because most recent lahars (within the last few decades) at Concepcion´ have been relatively small, moderate and high spatial resolution remote sensing data are the most suitable to study these lahars. The largest lahar deposits from Concepcion´ (i.e., La Flor and Los Ramos, Figure 4.4 and

4.5) are visible on ASTER and Landsat images, but smaller deposits are not. All ASTER and

Landsat images used in this study were already orthorectiﬁed. The 2008 Los Ramos lahar deposit was much larger in extent and volume than the 2009 deposit produced by a lahar that descended from the eastern ﬂank of Concepcion.´ Both the 2008 and 2009 Los Ramos lahar deposits are too small to be visible in a 30 m x 30 m ground resolution Landsat image or in a 15 m x 15 m ground resolution ASTER image, but the 2008 deposit is partially seen on a “jpg” sample off-nadir panchromatic QuickBird image acquired on March 05, 2011 and downloaded from the Digital

Globe website (http://www.digitalglobe.com/). An original off-nadir panchromatic QuickBird image has a ground resolution of 0.72 m, but this sample image is of lower resolution than that (actual ground resolution is not specified in the website), but of higher resolution than a Landsat or ASTER image. To digitize the Los Ramos lahar deposit on this QuickBird sample image, I georeferenced it, and then overlaid a set of differential GPS ground control points collected in the field on this deposit. The image showed part of the deposit where GPS data were not available due to vegetation cover blocking the GPS signal during the field campaign. The distal phase of the deposit is not seen on the QuickBird sample image, but I was able to gather GPS data along the edges of the deposit

59 at that location since there were no high trees blocking the view of the sky. This is what I refer to as a hybrid approach in which ground control points from GPS observations measured on the lahar deposit were used to assist the digitizing process on a georeferenced image.

I tried different approaches to extract information from satellite imagery (i.e., planimetric area inundated by lahars). First, I tried enhancing the contrast manually of the different bands for each single image to see how well lahar deposits can be distinguished from the background. Band

4 data (0.76 – 0.90 µm, NIR spectral region) of Landsat 7, and its equivalent band 3N (nadir view 0.78 – 0.86 µm) of the ASTER instrument, provided the best results for any single band. However, one of the major shortcomings of working with a single band was that lahar deposits, lahar channels (ravines or gullies), and any other deforested area (e.g., recent lava ﬂows) all look the same. Thus, I decided to try two different vegetation index techniques, which involve the mathematical combination of at least two bands to see if better results were obtained.

The Normalized Difference Vegetation Index (NDVI) was introduced by Tucker, (1979) to estimate the cover of green vegetation:

NIR − Red NDVI = (4.3) NIR + Red

where NIR, as stated above, is the spectral band in the near-infrared region of the electromagnetic region (∼0.75 – 1.40 µm), and Red is the spectral band in the red region (∼0.625 – 0.740

µm) of the electromagnetic spectrum. The major pitfalls of the NDVI is its high sensitivity to soil bakcground, atmosphere, and sun-angle conditions during image acquisition (e.g., Huete, 1988;

Kaufman and Tenre,´ 1992).

The Atmospherically Resistance Vegetation Index (ARVI) proposed by Kaufman and Tenre´

(1992) used three wavebands to try to minimize the atmospheric effects by incorporating the “blue”

band in the vegetation index technique, and is given by:

60 NIR − Red + γ(Blue − Red) ARVI = (4.4) NIR + Red − γ(Blue − Red)

where Blue is the band associated with blue color (∼0.450 – 0.495 µm), and γ (0 < γ ≤ 2) is a factor related to the aerosol type present in the atmosphere.

4.6 Results

Figure 4.4 displays a comparison of the results obtained by the techniques described in the previous section. To “isolate” the features of interest (i.e., lahar deposits), I assigned white colors to those unwanted features by means of a semi-unsupervised classification. The NDVI (Eq 4.1) is able to discern oldest from younger deposits when there is enough vegetation difference between these two features. For instance, younger deposits (lahar scars, gullies, and lava flows inclusive) shown as dark brown in Figure 4.4B are vegetation free features, while older scars and lava flows that have a minor vegetation cover are shown in yellow. The ARVI algorithm (Eq 4.2) did not yielded better results than the NDVI algorithm, and was technically identical than those from the single band approach. ARVI results improved by using a value of γ ∼0.1, see Figure 4.4. In general, all three techniques (NIR band, NDVI and ARVI) produce very similar results, and none of them are able to distinguish between a lahar deposit and scars (i.e., barren soil and gullies) left by the debris flow (or from a fire) on the topography or from lava flows.

A slight improvement can be gained in the distinction between the true lahar deposit and the background by doing a traditional false color composite band combination like the one shown in

Figure 4.4D, in which the Landsat 7 image is displayed as R: Band 4 (0.76 – 0.90 µm), G: Band

5 (1.55 – 1.75 µm), and B: Band 7 (2.08 – 2.35 µm). A band combination using the results of the algorithm previously described also provides a good contrast between lahar deposits and their backgrounds. This is shown in Figure 4.5 where lahar deposits are in blue, and some lahar scars are in pink and purple. Although some other features that lack vegetation, like lava ﬂows and gullies, are also shown in blue, lahar deposits are expected to be on the lower ﬂanks of the volcano. The

61 648000 648000 656000 656000 1280000 1280000

0 5 km 1272000 1272000 1272000 1272000 A 656000 656000 B

648000 648000 656000 656000 1280000 1280000 1272000 1272000 1272000 1272000

C 656000 D 656000

Figure 4.4. Comparison of different approaches to highlight debris flows on Concepcion´ using the same Landsat 7 image acquired on January 27, 2001. (A) Band 4. Deforested areas by debris flows (WSW) and lava flows (NE and ESE) look the same. (B) NDVI (Tucker, 1979). Most deforested areas are shown in dark brown, while less severe deforested areas are in yellow. (C) ARVI (Kaufman and Tenre,´ 1992). Best results were obtained using a γ value of 0.1, see Eq 4.2. Explanations are provided in the text. (D) False color composite (R: band 4, G: band 5, B: band 7). Most deforested areas appear in dark purple.

62 San Jose´ del Sur lahar actually starts a little high on the SW slopes as depicted by the yellow lines in Figure 4.5. This is due to a break in the slope caused by an apparent collapse of part of the slope in that area.

I proceeded in the same way with all the available images in order to digitize lahars near La

Flor, San Jose´ del Sur, and Los Ramos communities (Figure 4.2 and 4.5) in order to get an estimate of the planimetric area inundated by the deposits of these lahars. To get a rough estimate of the uncertainty (the minimum possible value) associated with the measurements of the planimetric area inundated by a lahar from a satellite image, I assumed that I have completely encompassed the lahar deposit (by digitizing a polygon around the deposit), and that at the margins the deposit covers only 50% of the image pixels. So, in this way, the uncertainty can be computed by the number of pixels along the perimeter of the digitized polygon (lahar deposit) times 50% of the pixel area, which depends on the spatial resolution of the satellite image. The volume of the lahar deposits were then computed using one of the Iverson et al., (1998) empirical-statistical equations, namely Eq 4.2. These results are presented in Table 4.2.

The chronology of the digitized lahar deposits for La Flor, Los Ramos and San Jose´ del Sur communities are shown in Figures 4.6, 4.7, and 4.8, respectively, see also Table 4.2. La Flor lahar inundated planimetric area increased between 1986 and 1987, from 1.58 x 105 to about 3.29 x 105

m2 (Figure 4.6, Table 4.2). The inundated planimetric area of the lahars in the zone of La Flor

during 2000 and 2003 are comparatively small. The deposit of a subsequent lahar seen on the

ASTER image of 2010 was relatively large, covering a similar surface area to that occurred in 1987

(Figure 4.6).

The ﬁrst lahar at Los Ramos that took place in October 2008 (INETER, 2008b), was 3.5 times

larger than the second one that occurred in the same place almost a year later (Figure 4.7, Table

4.2). On the other hand, the San Jose´ del Sur lahar zone seems to be most active one at Concepcion,´

at least from the available data set that dates back to 1978, but it does not seem to be progressively

increasing, rather the deposits in this lahar zone are slightly below 6 x 105 m2 (Figure 4.8). One possible explanation for this may be that this volume represents the maximum volume of material, on the volcano’s slope, the can resist failure from rainfall.

63 (A)

La Flor lahar

San José del Sur (B) ±

1280000 ± 1273000

1275000 Los Ramos 1272000

0 500 m

651000 652000

1270000 0 5 Km

645000 650000 655000

Figure 4.5. (A) False color composite Landsat image, acquired on January 27, 2001 produced by the band combination R: band 4, G: ARVI (see Eq 4.2), and B: NDVI (see Eq 4.1), and individually shown in Figure 4.4. Lahar deposits are highlighted by the yellow lines within the zoomed boxes. (B) Los Ramos 2008 and 2009 lahar deposits partially visible on a “jpg” sample image downloaded from the Digital Globe website. The digitizing of this lahar deposits were complemented with GPS data gathered in the ﬁeld.

64 648000 La Flor lahar Legend

1278000 1986-04-02 1278000 1987-10-14 2000-01-27 2003-01-10 2010-12-08

0 1,000 Meters 648000

Figure 4.6. Temporal evolution of the lahar deposits heading down to La Flor town from the western slopes of Concepcion´ volcano. See Figure 4.2 for geographic location of La Flor town, and Table 4.2 for an estimation of the area and volume of the deposits. Inset depicts the approximate location of the lahar.

A time series of the computed lahar volumes reported in Table 4.2 is shown in Figure 4.9. The time series for the San Jose´ del Sur lahar zone is the longest one with 9 lahar events unevenly spaced in time. Due to cloud cover during the time of image acquisition we could not digitize the

La Flor lahar deposits during 1990 and 1998. The San Jose´ del Sur lahar deposit is almost double in size than the largest La Flor lahar deposit, and almost ﬁve times larger than the Los Ramos deposit.

4.7 Discussion

The factors that destabilize slopes of volcanoes are very complex (e.g., Sidle, 1992; Voight and

Elsworth, 1997; Vallance, 2005). At Concepcion´ volcano, during the last few decades rainfall is the chief factor that triggers lahars by mobilization of loose volcanic material on the slopes of the volcano, which have become unstable and vulnerable to erosion mainly due to volcanic activity and anthropogenic disturbances. Reported laharic activity at Concepcion´ volcano has increased during

65 Table 4.2. Planimetric area inundated by lahars at Concepcion´ volcano. Datea Location Areab E. Areac Volumed E. Volumee Sensor (x105 m2) (x105 m2) (x103 m3) (x103 m3) 02/04/1978 San Jose´ 2.21 1.42 36.7 35.4 Landsat 3 del Sur 02/04/1986 San Jose´ 4.13 0.86 94.0 29.3 Landsat 5 del Sur 02/04/1986 La Flor 1.58 0.56 22.1 11.8 Landsat 5 14/10/1987 San Jose´ 1.31 0.57 16.7 10.9 Landsat 5 del Sur 14/10/1987 La Flor 3.29 1.10 66.8 33.5 Landsat 5 23/01/1990 San Jose´ 5.73 1.29 153.4 51.8 Landsat 5 del Sur 15/12/1998 San Jose´ 0.28 0.27 1.6 2.4 Landsat 5 del Sur 27/01/2000 San Jose´ 4.61 1.18 110.6 42.5 Landsat 7 del Sur 27/01/2000 La Flor 0.66 0.25 6.0 3.4 Landsat 7 10/01/2003 San Jose´ 5.94 0.91 161.7 37.2 ASTER del Sur 10/01/2003 La Flor 0.87 0.18 9.0 2.8 ASTER 31/01/2005 San Jose´ 4.21 0.60 96.5 20.6 ASTER del Sur 08/12/2010 San Jose´ 5.64 0.66 149.9 26.3 ASTER del Sur 08/12/2010 La Flor 2.52 0.47 44.7 12.5 ASTER 15/12/2008 Los Ramos 1.18 – 14.4 – Hybrid f 08/12/2010 Los Ramos 0.51 0.15 4.1 1.8 ASTER aDate of satellite image acquisition, except for ﬁeld campaign in Los Ramos during December 15, 2008. bPlanimetric inundated area (B) encircle by the digitized polygon along the edges of the lahar deposit. cError estimated in the planimetric area. See text for details. dComputed using Eq 4.2. eError in volume, estimated by error propagation. f Digitizing of this lahar deposit was done with auxiliary ground-truth GPS data.

66 652000 1273000 1273000

Legend

2008-12-15 2010-12-08

Los Ramos lahar 1272000 1272000

0 500 Meters

652000

Figure 4.7. Temporal evolution of the lahar deposits of the ﬂows threatening Los Ramos town southeastern side of Concepcion´ volcano. See Figure 4.2 for geographic location of the town, and Table 4.2 for an estimation of the area and volume of the deposits. Inset depicts the approximate location of the lahar.

67 (A) 648000 650000 (B) 648000

0 1,000 Meters

San José del Sur lahar

Legend

1978-04-02 1986-04-02 Legend

1987-10-14 1998-12-15 1990-01-23 2000-01-27 1273000 1273000 2003-01-10 1273000 1273000 2005-01-31 0 1,000 Meters 2010-12-08

648000 650000 648000

Figure 4.8. Temporal evolution of the lahar deposits north of San Jose´ del Sur town SSW slopes of Concepcion´ volcano. See Figure 4.2 for geographic location of the town, and Table 4.2 for an estimation of the area and volume of the deposits. Inset depicts the approximate location of the lahar. (A) Lahar chronology from 1978 to 1990. (B) Lahar chronology from 1998 to 2010.

6 6 )

2 Los Ramos lahar La Flor lahar m 5 5 San José del Sur lahar 5 Hurricane Joan 4 Hurricane Mitch 4

3 3

2 2

1 1 Lahardeposit area (x10 0 0 1978/01/10 1980/10/06 1983/07/03 1986/03/29 1988/12/23 1991/09/19 1994/06/15 1997/03/11 1999/12/06 2002/09/01 2005/05/28 2008/02/22 2010/11/18

Figure 4.9. Time series of inundated areas by lahar deposits extracted from satellite data. The plot seems to show that generally a small lahar is followed by a larger one, or vice versa. It is possible that Hurricane Joan triggered a large lahar at San Jose´ del Sur in 1988, but Hurricane Mitch seemed not to have triggered any lahars at Concepcion.´

68 the last decade (Table 4.1.). The lack of reported lahars in previous decades does not mean they did not occur because reporting is inconsistent, as it has been shown in this present work through the analysis of several satellite images dating back to 1978.

The lahar deposits that are most clearly detected with Landsat and ASTER satellite data are those near the San Jose´ del Sur and La Flor communities (Figure 4.2, 4.5, and 4.7) because they have the largest planimetric inundated areas. However, there are other lahars near La Chirca and La

Sabana (Figure 4.2, Table 4.1) that ﬂew through relatively vegetated areas and their channel ﬂows and deposits are too narrow as to be seen in these satellite data, but are known to have occurred as reported by INETER (Table 4.1).

The San Jose´ del Sur lahars are comparable in size to some reported at Mt. St. Helens and Mt.

Hood in 1980 (Gallino and Pierson, 1984; Major and Voight, 1986). La Flor lahars are comparable to those reported by Rodolfo, (1989) at Mayon volcano in the Philippines triggered by a typhoon in 1985. Hurricane Joan hit predominantly the Caribbean coast of Nicaragua in October 1988, and it has been reported that this hurricane triggered a “mudﬂow” (lahar) at Concepcion´ (Devoli et al.,

2007; and references therein). Thus, we attribute the increase in the size of the lahar deposits, observed in the data set available in this study, from the early 1980s to 1990 at Concepcion´ volcano

(Figure 4.9) to be due to the intense rainfall from the Hurricane Joan. However, the infamous

Hurrricane Mitch that badly affected the north-eastern and north-western sides of Nicaragua (and a large portion of Central America) in late October 1998 does not seem to have triggered any signiﬁcant debris ﬂows at Concepcion´ volcano, since there were no reports of such an event. This is supported by the satellite data of this study, which do not show an increase in lahar size deposits

(Figure 4.9) even in an image acquired immediately after that this hurricane hit Nicaragua.

I have computed lahar volumes using the planimetric inundated area empirical-statistical relationship of Iverson et al. (1998). These volumes can be used to simulate lahar ﬂows surging down the slopes of Concepcion´ volcano by ﬂow routing models like LAHARZ (Schilling, 1998), and more sophisticated numerical models such as Flo-2D (O’Brien et al., 1993), TITAN2D (Pa- tra et al., 2005; Patra et al., 2006) and VolcFlow (Kelfoun and Druitt, 2005; Kelfoun et al., 2009).

69 These simulations can then be translated into lahar hazard maps that are valuable tools for decision- makers like civil defense and local authorities.

The lahar volumes computed using Eq 4.2 and presented in Table 4.2 allow me to get a prelim- inary estimates of the erosion rate of the slopes of Concepcion´ volcano during the period 1978 to

2012, which is 29 x 103 m3 yr−1 (29 km3 Myr−1). This should be taken as a conservative minimum short term erosion rate, since lahar deposits at La Chirca and La Sabana (and other places) have not been taken into account, although they are very small (INETER, 2008b). Thouret, (1999) reported short term erosion rates for Pinatubo volcano after its massive eruption in 1991 of 136 x 103 – 219 x 103 km3 Myr−1, which are signiﬁcantly higher than our current estimation for Concepcion´ volcano. This is because the amount of material available to be eroded after the eruption of Pinatubo was huge. In contrast, tephra deposition on the ﬂanks from eruptions at Concepcion´ volcano are very small. High erosion rates due to the removal of “fresh” tephra products on active volcanoes tend to decline in a relatively short time, in a matter of few years, if there is no addition of new materials (Thouret, 1999), but it may be different for other volcanoes on different geological and geographical settings.

The predominantly tropical climate of Nicaragua, are unfavorable for the preservation of the small volume tephra deposits erupted from Concepcion´ volcano in the last four decades. Thus, it is difﬁcult to carry out a reliable volcano mass balance for Concepcion.´ However, the fact that the lahar channels are progressively growing through time, and that new lahar zones are developing on the slopes of the volcano, may suggest that currently erosion by lahars is more signiﬁcant than mass addition to the volcano by eruptions.

4.8 Conclusions

I have shown that moderate spatial resolution satellite data is of signiﬁcant importance in the extraction of inundated areas by debris ﬂows (those covering a planimetric area ≥104 m2), and mainly for past events that lack enough information or ground-truth data. Nonetheless, caution must be taken when mapping lahar deposits in this way because lahar scars and channels (and non- vegetated areas) may be misinterpreted as part of the actual deposit, leading to an overestimation

70 of area inundated and/or volume. There is not a straightforward approach to clearly and deﬁnitely

“isolate” the lahar deposits from other features that responsd in a similar fashion within the visible and near-infrared region of the electromagnetic spectrum. Thus, ground-truth data (e.g., comple- mentary data from GPS measurements) is essential to constraint the actual limits of the debris ﬂow deposit.

A time series of the inundated areas by lahar deposits, around Concepcion´ volcano, extracted from satellite data suggests that generally a small lahar is followed by a larger one, or vice versa.

The satellite set available for this study shows that the San Jose´ del Sur lahar zone (Figure 4.2) is currently the most active one at Concepcion,´ and is also producing the more voluminous lahars (in the order of 105m3).

New lahar zones are developing on the eastern and north-eastern slopes of Concepcion´ volcano. As today, the deposits and channels of this debris ﬂows are too small to be detected by moderate ground resolution satellite images (e.g., Landsat and ASTER), but are visible in high ground resolution satellite images.

4.9 Recommendations

A lahar monitoring and a robust educational programs should be prioritized for Ometepe Island, focused in the prevention and mitigation of volcanic disasters associated with debris ﬂows.

Although the real factors producing the instability of the slopes of Concepcion´ volcano are yet not fully known, I propose to use the results of this study as a starting point for the elaboration of lahar hazard maps through numerical simulations. What is known is that at Concepcion´ debris

ﬂows are triggered by rainfall, so I propose that a lahar monitoring program should be focused on rainfall-triggered lahars. At different volcanoes lahars are triggered by different amount of rainfall, so it is important to look for the rainfall threshold at which lahars are triggered at Concepcion.´

This can be achieved by at least one rain gauge recording data continuously. Under intense rain or

high cloud cover conditions, it is difﬁcult to see the onset of a lahar, thus seismic signals associated

to the lahar must be careful ﬁltered out from the seismograms. In this way a precise timing of

the amount of rainfall and the lahar initiation can be determined. Other hydraulic parameters are

71 also very important in a more precise determination of the factors leading to the lahar formations, so I would recommend the installation of piezometers on the areas susceptible to failure. This devices measure the water content of the soil, and can be attached to data-loggers and in turn to a telemetry system to get readings in a real-time basis. Soil-saturated conditions are important pieces of information to model debris ﬂows initiation and their dynamics, too. Other techniques to monitor lahars require a lot of maintenance and I think they will be worth to try in a later stage, after a more deep understanding of the genesis of lahar is achieved for Concepcion´ volcano.

72 REFERENCES

Afﬂeck, D. K., Cassidy, J., and Locke, C. A. (2001). Te Pouhawaiki volcano and pre-volcanic topography in central Auckland: Volcanological and hydrogeological implications. New Zealand Journal of Geology and Geophysics, 44:313–321.

Altamimi, Z., Collilieux, X., and Metivier,´ L. (2011). ITRF2008: An improved solution of the international terrestrial reference frame. Journal of Geodesy, 85:457–473, doi:10.1007/s00190– 011–0444–4.

Alvarado, D., DeMets, C., Tikoff, B., Hernandez,´ D., Wawrzyniec, T. F., Pullinger, C., Mattioli, G., Turner, H. L., Rodriguez, M., and Correa-Mora, F. (2011). Forearc motion and deformation between El Salvador and Nicaragua: GPS, seismic, structural, and paleomagnetic observations. Lithosphere, 3:3–21.

Ammann, M., Boll,¨ A., Rickli, C., Speck, T., and Holdenrieder, O. (2009). Signiﬁcance of tree root decomposition for shallow landslides. For. Snow Landsc. Res, 82(1):79–94.

Bertiger, W., Desai, S. D., Haines, B., Harvey, N., Moore, A. W., Owen, S., and Weiss, J. P. (2010). Single receiver phase ambiguity resolution with GPS data. Journal of Geodesy, 84:327–337, doi:10.1007/s00190–010–0371–9.

Blais, J. A. R. and Ferland, R. (1984). Optimization in gravimetric terrain corrections. Canadian Journal of Earth Sciences, 21:505–570.

Blewitt, G. (2007). GPS and space based geodetic methods. In Herring, T., editor, Treatise on Geophysics, volume 3, pages 351–390. Academic Press.

Bonaccorso, A. and Davis, P. (1999). Models of ground deformation from vertical volcanic con- duits with application to eruptions of Mount St. Journal of Geophysical Research, 104(B5):10– 531.

Borgia, A. (1994). Dynamic basis of volcanic spreading. Journal of Geophysical Research, 99:17791–17804.

Borgia, A., Delaney, P. T., and Denlinger, R. P. (2000). Spreading volcanoes. Annual Reviews in Planetary Sciences, 28:539–570.

Borgia, A. and van Wyk de Vries, B. (2003). The volcano-tectonic evolution of Concepcion,´ Nicaragua. Bulletin of Volcanology, 65:248–266.

Brown, G. C., Everett, S. P., Rymer, H., McGarvie, D. W., and Foster, I. (1991). New light on caldera evolution – Askja, Iceland. Geology, 19:352–355.

73 Brown, G. C., Rymer, H., and Thorpe, R. S. (1987). The evolution of andesite volcano structures: New evidence from gravity studies in Costa Rica. Earth and Planetary Science Letters, 82:323– 334. Budetta, G., Nunziata, C., and Rapolla, A. (1983). A gravity study of the island of Vulcano, Tyrrhenian Sea, Italy. Bulletin of Volcanology, 46:183–192. Camacho, A., Fernandez,´ J., and Gottsmann, J. (2011). The 3-D gravity inversion package GROWTH2. 0 and its application to Tenerife Island, Spain. Computers & Geosciences. Camacho, A., Montesinos, F., and Vieira, R. (2000). Gravity inversion by means of growing bodies. Geophysics, 65(1):95–101. Camacho, A., Montesinos, F., and Vieira, R. (2002). A 3-D gravity inversion tool based on explo- ration of model possibilities. Computers & geosciences, 28(2):191–204. Camacho, A. G., Montesinos, F. G., and Vieira, R. (1997). A three-dimensional gravity inversion applied to Sao Miguel Island (Azores). Journal of Geophysical Research, 102:7705–7715. Campbell, D. L. (1980). Gravity terrain corrections for stations on a uniform slope – a power law approximation. Geophysics, 45:109–112. Carr, M. J., Saginor, I., Alvarado, G. E., Bolge, L. L., Lindsay, F. N., Milidakis, K., Turrin, B. D., Feigenson, M. D., and Swisher, C. C. (2007). Element fluxes from the volcanic front of Nicaragua and Costa Rica. Geochemistry Geophysics Geosystems, 8:Q06001. Casadevall, T. J., Johnston, D. A., Harris, D. M., Rose, W. I., Malinconico, L. L., Stoiber, R. E., Bornhorst, T. J., Williams, S. N., Woodruff, L., and Thompson, J. (1981). SO2 emission rates at Mt. St. Helens from March 29 through December, 1980. In Lipman, P. W. and Mullineaux, D. R., editors, The 1980 Eruption of Mt. St. Helens, Washington, pages 193–207. US Geological Survey Profesional Paper 1250, USA. Cassidy, J., France, S. J., and Locke, C. A. (2007). Gravity and magnetic investigations of maar volcanoes, Auckland volcanic field, New Zealand. Journal of Volcanology and Geothermal Research, 159:153–163. Chatfield, C. (1996). The Analysis of Time–series: An Introduction. Chapmann & Hall/CRC, Florida, USA, 5 edition. Connor, C. B., Connor, L., and Sheridan, M. (2006). Assessment of October 2005 debris flows at Panabaj, Guatemala, and recommendations for hazard mitigation. OXFAM GB report, Univer- sity of South Florida, Tampa, USA. Connor, C. B. and Williams, S. N. (1990). Interpretation of gravity anomalies at Masaya caldera complex, Nicaragua. In Transactions 12th Caribbean Conference, St. Croix, U.S. Virgin Islands, pages 495–502. Miami Geological Society. Correa-Mora, F., DeMets, C., Alvarado, D., Turner, H. L., Mattioli, G., Hernandez, D., Pullinger, C., Rodriguez, M., and Tenorio, C. (2009). Gps-derived coupling estimates for the Central America subduction zone and volcanic arc faults: El Salvador, Honduras, and Nicaragua. Geo- physical Journal International, 179:1279–1291.

74 Dalton, M. P., Waite, G. P., Watson, I. M., and Nadeau, P. A. (2010). Multiparameter quantiﬁcation of gas release during weak Strombolian eruptions at Pacaya Volcano, Guatemala. Geophysical Research Letters, 37(L09303):doi:10.1029/2010GL042617.

Davila, N., Capra, L., Gavilanes-Ruiz, J., Varley, N., Norini, G., and Vazquez, A. (2007). Recent lahars at Volcan´ de Colima (Mexico): drainage variation and spectral classiﬁcation. Journal of Volcanology and Geothermal Research, 165(3-4):127–141.

Delcamp, A., van Wyk de Vries, B., and James, M. R. (2008). The inﬂuence of ediﬁce slope and substrata on volcano spreading. Journal of Volcanology and Geothermal Research, 177:925– 943.

Delgado, H. and Navarro, M. (2002). Mapa de amenaza volcanica´ del volcan´ Concepcion.´ Tech- nical report, Instituto de Geof´ısica, UNAM Mexico; and Instituto Nicaragense¨ de Estudios Ter- ritoriales.

DeMets, C. (2001). A new estimate for present–day Cocos–Caribbean Plate motion: Implications for slip along the Central American Volcanic Arc. Geophysical Research Letters, 28:4043–4046, doi:10.1029/2001GL013518.

Deplus, C., Bonvalot, S., Dahain, D., Diament, M., Harjono, H., and Dubois, J. (1995). Inner structure of the Krakatau volcano complex (Indonesia) from gravity and bathymetry data. Journal of Volcanology and Geothermal Research, 64:23–52.

Devoli, G., Morales, A., and Høeg, K. (2007). Historical landslides in Nicaragua–collection and analysis of data. Landslides, 4(1):5–18.

Dixon, T. H. (1991). An introduction to the global positioning system and some geological appli- cations. Review of Geophysics, 29:249–276.

Dzurisin, D. (2007). The global positioning system: a multipurpose tool. In Dzurisin, D., editor, Volcano Deformation: Geodetic Monitoring Techniques, pages 111–152. Praxis Publishing Ltd, Chichester, UK.

Ehrenborg, J. (1996). A new stratigraphy for the tertiary volcanic rocks of the Nicaraguan hihg- lands. Geological Society of America Bulletin, 108:830–842.

Elming, S. and Rasmussen, T. (1997). Results of magnetotelluric and gravimetric measurements in western Nicaragua, Central America. Geophysical Journal International, 128:647–658.

Endo, E. T. and Murray, T. (1991). Real–time seismic amplitude measurement (RSAM): a volcano monitoring and prediction tool. Bulletin of Volcanology, 53:533–545.

Finn, C. and Williams, D. L. (1982). Gravity evidence for a shallow intrusion under Medicine Lake volcano, California. Geology, 10:503–507.

Fischer, T. P., Roggensack, K., and Kyle, P. R. (2002). Open and almost shut case for explosive eruptions; vent processes determined by SO2 emission rates at Karymsky Volcano, Kamchatka. Geology, 30:1059–1062.

75 Fisher, R. V., Heiken, G., and Hulen, J. B. (1997). Volcanoes: crucibles of change. Princeton University Press, Princeton, New Jerssey.

Flynn, L., Harris, A., and Wright, R. (2001). Improved identiﬁcation of volcanic features using Landsat 7 ETM+. Remote Sensing of Environment, 78(1):180–193.

Francis, P. W. and Wells, G. L. (1988). Landsat thematic mapper observations of debris avalanche deposits in the central andes. Bulletin of Volcanology, 50:258–278. 10.1007/BF01047488.

French, S. W., Warren, L. M., Fischer, K. M., Abers, G. A., Strauch, W., Protti, J. M., , and Gonzalez, V. (2010). Constraints on upper plate deformation in the Nicaraguan subduction zone from earthquake relocation and directivity analysis. Geochemistry Geophysics Geosystems, 11:Q03S20, doi:10.1029/2009GC002841.

Funk, J., Mann, P., McIntosh, K., and Stephens, J. (2009). Cenozoic tectonics of the Nicaraguan depression, Nicaragua, and Median Trough, El Salvador, based on seismic-reﬂection proﬁling and remote-sensing data. Geological Society of America Bulletin, 121:1491–1521.

Galland, O. (2012). Experimental modelling of ground deformation associated with shallow magma intrusions. Earth and Planetary Science Letters, 317–318:145 – 156, doi:10.1016/j.epsl.2011.10.017.

Galle, B., Johansson, M., Rivera, C., Zhang, Y., Kihlman, M., Kern, C., Lehmann, T., Platt, U., Arellano, S., and Hidalgo, S. (2010). Network for Observation of Volcanic and Atmospheric Change (NOVAC)–A global network for volcanic gas monitoring: Network layout and instrument description. Journal of Geophysical Research, 115(D05304):doi:10.1029/2009JD011823.

Galle, B., Oppenheimer, C., Geyer, A., McGonigle, A. J., Edmonds, M., and Horrocks, L. (2003). A miniaturised ultraviolet spectrometer for remote sensing of SO2 ﬂuxes: a new tool for volcano surveillance. Journal of Volcanology and Geothermal Research, 119(14):241 – 254.

Gallino, G. and Pierson, T. (1984). The 1980 Polallie Creek debris flow and subsequent dam-break flood, East Fork Hood River basin, Oregon. Open file report 84-578, US Geological Survey.

Gottsmann, J., Carniel, R., Coppo, N., Wooller, L., Hautmann, S., and Rymer, H. (2007). Oscillations in hydrothermal systems as a source of periodic unrest at caldera volcanoes: Multiparameter insights from Nisyros, Greece. Geophysical Research Letters, 34(L07307):doi:10.1029/2007GL029594.

Hackl, M., Malservisi, R., Ugentobler, U., and Wonnacott, R. (2011). Estimation of velocity uncertainties from GPS time series: Examples from the analysis of the South African Trignet network. Journal of Geophysical Research, 116(B11404):doi:10.1029/2010JB008142.

Hubbard, B., Sheridan, M., Carrasco-Nu´nez,˜ G., D´ıaz-Castellon,´ R., and Rodr´ıguez, S. (2007). Comparative lahar hazard mapping at Volcan Citlaltepetl,´ Mexico using SRTM, ASTER and DTED-1 digital topographic data. Journal of Volcanology and Geothermal Research, 160(1):99– 124.

Huete, A. (1988). A soil-adjusted vegetation index (SAVI). Remote sensing of environment, 25(3):295–309.

76 Huggel, C., Schneider, D., Miranda, P., Delgado Granados, H., and Ka¨ab,¨ A. (2008). Evaluation of ASTER and SRTM DEM data for lahar modeling: A case study on lahars from Popocatepetl´ Volcano, Mexico. Journal of Volcanology and Geothermal Research, 170(1):99–110.

Ichihara, M., Takeo, M., Yokoo, A., Oikawa, J., and Ohminato, T. (2012). Monitoring volcanic activity using correlation patterns between infrasound and ground motion. Geophysical Research Letters, 39(4):doi:10.1029/2011GL050542.

INETER (2002a). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, August, 2002. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2002/08/index0208.htm.

INETER (2002b). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, July, 2002. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2002/07/index0207.htm.

INETER (2002c). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, June, 2002. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2002/06/index0206.htm.

INETER (2002d). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, October, 2002. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2002/09/index0209.htm.

INETER (2002e). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, September, 2002. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2002/08/index0208.htm.

INETER (2003a). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, November, 2003. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2003/11/index0311.htm.

INETER (2003b). Bolet´ın mensual: Sismos y Volcanes de Nicaragua: Catalogo´ anual, 2003. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2003/anual/index2003.htm.

INETER (2004). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, Mayo 2004. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2004/05/index0405.htm.

INETER (2005a). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, August 2005. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2005/08/index0508.htm.

INETER (2005b). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, July 2005. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2005/07/index0507.htm.

INETER (2005c). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, Junio 2005. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2005/06/index0506.htm.

INETER (2005d). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, Mayo 2005. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2005/05/index0505.htm.

INETER (2006). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, Septiembre 2006. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2006/09/index0609.htm.

INETER (2008a). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, July, 2008. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2008/07/index0807.htm.

77 INETER (2008b). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, Octubre, 2008. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2008/10/index0810.htm.

INETER (2010a). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, Abril, 2010. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2010/04/index1004.htm.

INETER (2010b). Bolet´ın mensual: Sismos y Volcanes de Nicaragua, March, 2010. Managua. http://webserver2.ineter.gob.ni/geoﬁsica/boletin/2010/03/index1003.htm.

INIFOM (2001). Caracterizaciones Municipales de Nicaragua. Techni- cal report, Instituto Nicaragense¨ de Fomento Municipal. available at http://www.inifom.gob.ni/municipios/municipios.html.

Iverson, R. et al. (2000). Landslide triggering by rain inﬁltration. Water Resources Research, 36(7):1897–1910.

Iverson, R., Reid, M., and LaHusen, R. (1997). Debris-ﬂow mobilization from landslides 1. Annual Review of Earth and Planetary Sciences, 25(1):85–138.

Iverson, R., Schilling, S., and Vallance, J. (1998). Objective delineation of lahar-inundation hazard zones. Geological Society of America Bulletin, 110(8):972–984.

Jordan, T. A., Ferraccioli, F., Jones, P. C., Smellie, J. L., Ghidella, M., and Corr, H. (2009). Air- borne gravity reveals interior of Antarctic volcano. Physics of the Earth and Planetary Interiors, 175:127–136.

Kane, M. F. (1962). A comprehensive system of terrain corrections using a digital computer. Geophysics, 27:455–462.

Kaufman, Y. and Tanre,´ D. (1992). Atmospherically resistant vegetation index (ARVI) for EOS- MODIS. Geoscience and Remote Sensing, IEEE Transactions on, 30(2):261–270.

Kazahaya, R., Mori, T., Takeo, M., Ohminato, T., Urabe, T., and Maeda, Y. (2011). Relation between single very-long-period pulses and volcanic gas emissions at Mt. Asama, Japan. Geo- physical Research Letters, 38(L11307):doi:10.1029/2011GL047555.

Kelfoun, K. and Druitt, T. H. (2005). Numerical modeling of the emplacement of Socompa rock avalanche, Chile. Journal of Geophysical Research, 110:B12202.

Kelfoun, K., Samaniego, P., Palacios, P., and Barba, D. (2009). Testing the suitability of frictional behaviour for pyroclastic ﬂow simulation by comparison with a well-constrained eruption at Tungurahua volcano (Ecuador). Bulletin of volcanology, 71(9):1057–1075.

Kerle, N. and Oppenheimer, C. (2002). Satellite remote sensing as a tool in lahar disaster management. Disasters, 26(2):140–160.

Kutterolf, S., Freundt, A., Perez,´ W., amd U. Schacht, T. M., Wehrmann, H., and Schmincke, H.- U. (2008). Paciﬁc offshore record of plinian arc volcanism in Central America: 1. along-arc correlations. Geochemistry Geophysics Geosystems, 9:Q02S01, doi:10.1029/2007GC001631.

78 La Femina, P. C., Dixon, T. H., Govers, R., Norabuena, E., Turner, H., Saballos, A., Mattioli, G., Protti, M., and Strauch, W. (2009). Forearc motion and Cocos Ridge collision in Central America. Geochemistry Geophysics Geosystems, 10:Q05S14, doi: 10.1029/2008GC002181.

LaFehr, T. R. (1991). Standardization in gravity reduction. Geophysics, 56:1170–1178.

Lavigne, F. (1999). Lahar hazard micro-zonation and risk assessment in Yogyakarta city, Indonesia. GeoJournal, 49:173–183.

Lisowski, M. (2007). Analytical volcano deformation source models. In Dzurisin, D., editor, Volcano Deformation: Geodetic Monitoring Techniques, pages 279–304. Praxis Publishing Ltd, Chichester, UK.

Lopez,´ D. and Williams, S. (1993). Catastrophic volcanic collapse: relation to hydrothermal processes. Science, 260(5115):1794–1796.

Lowe, D. R., Williams, S. N., Leigh, H., Connor, C. B., Gemmell, J. B., and Stoiber, R. E. (1986). Lahars initiated by the 13 November 1985 eruption of Nevado del Ru´ız, Colombia. Nature, 324:51–53.

Major, J. and Voight, B. (1986). Sedimentology and clast orientations of the 18 may 1980 southwest-ﬂank lahars, Mount St. Helens, Washington. Journal of Sedimentary Research, 56(5):691–705.

Mastin, L. and Witter, J. (2000). The hazards of eruptions through lakes and seawater. Journal of Volcanology and Geothermal Research, 97(1–4):195–214.

Mathieu, L., van Wyk de Vries, B., Pilato, M., and Troll, V. R. (2011). The interaction between volcanoes and strike-slip, transtensional and transpressional fault zones: Analogue models and natural examples. Journal of Structural Geology, 33(5):898 – 906.

McBirney, A. and Williams, H. (1965). Volcanic history of Nicaragua. University of California Publications in Geological Sciences, 55:1–65.

McGee, K. A. and Sutton, A. J. (1994). Eruptive activity at Mount St Helens, Washington, USA, 1984-1988: a gas geochemestry perspective. Bulletin of Volcanology, 56:435–446.

Meisels, A., Raizman, S., and Karnieli, A. (1995). Skeletonizing a DEM into a drainage network. Computers & Geosciences, 21(1):187–196.

Merle, O. and Borgia, A. (1996). Scaled experiments of volcanic spreading. Journal of Geophysical Research, 101:13805–13817.

Minakami, T. (1941). Mean density of volcano Asama. Bulletin of the Earthquake Research Institute, 20:40–59.

Murray, J. B., Rymer, H., and Locke, C. A. (2000). Ground deformation, gravity, and magnetics. In Sigurdsson, H., Houghton, B., McNutt, S. R., Rymer, H., and Stix, J., editors, Encyclopedia of Volcanology, pages 1121–1140. Academic Press.

79 Nadeau, P. A., Palma, J. L., and Waite, G. P. (2011). Linking volcanic tremor, degassing, and eruption dynamics via so2 imaging. Geophysical Research Letters, 38(L01304):doi:10.1029/2010GL045820.

Nettleton, L. L. (1939). Determination of density for reduction of gravimeter observations. Geo- physics, 4:176–183.

Nilaweera, N. S. and Nutalaya, P. (1999). Role of tree roots in slope stabilisation. Bulletin of Engineering Geology and the Environment, 57:337–342. 10.1007/s100640050056.

Nishimura, T. (2009). Ground deformation caused by magma ascent in an open conduit. Journal of Volcanology and Geothermal Research, 187(3):178–192.

Norabuena, E., Dixon, T. H., Schwartz, S., DeShon, H., Newman, A., Protti, M., Gonzalez, V., Dorman, L., Flueh, E. R., Lundgren, P., Pollitz, F., and Sampson, D. (2004). Geodetic and seismic constraints on some seismogenic zone processes in Costa Rica. Journal of Geophysical Research, 109:B11403, doi: 10.1029/2003JB002931.

Nowell, D. A. G. (1999). Gravity terrain corrections – an overview. Journal of Applied Geophysics, 42:117–134.

O’Brien, J. S., Julien, P. Y. and Fullerton, W. T. (1993). Two-dimensional water ﬂood and mudﬂow simulation. Journal of Hydraulic Engineering, 119:244 – 261.

Okada, Y. (1985). Surface deformation due to shear and tensile faults in a half-space. Bulletin of the Seismological Society of America, 75(4):1135–1154.

Olmos, R., Barrancos, J., Rivera, C., Barahona, F., Lopez,´ D., Henriquez, B., Hernandez,´ A., Benitez, E., Hernandez,´ P., Perez,´ N., and Galle, B. (2007). Anomalous emissions of SO2 during the recent eruption of Santa Ana Volcano, El Salvador, Central America. Pure and Applied Geophysics, 164:2489 – 2506. 10.1007/s00024-007-0276-6.

Pack, R. (2005). Application of airborne and spaceborne remote sensing methods. In Jakob, M. and Hungr, O., editors, Debris–ﬂow Hazards and Related Phenomena, pages 275–289. Praxis. Springer Berlin Heidelberg, Chichester, UK.

Parasnis, D. S. (1997). Principles of Applied Geophysics. Chapman and Hall, London, 5 edition.

Patra, A., Bauer, A., Nichita, C., Pitman, E., Sheridan, M., Bursik, M., Rupp, B., Webber, A., Stin- ton, A., Namikawa, L., et al. (2005). Parallel adaptive numerical simulation of dry avalanches over natural terrain. Journal of Volcanology and Geothermal Research, 139(1):1–21.

Patra, A., Nichita, C., Bauer, A., Pitman, E., Bursik, M., and Sheridan, M. (2006). Parallel adaptive discontinuous galerkin approximation for thin layer avalanche modeling. Computers & Geosciences, 32(7):912–926.

Pelletier, J. (2008). Quantitative modeling of earth surface processes, volume 1. Cambridge University Press.

Plank, T., Blazer, V., and Carr, M. (2002). Nicaraguan volcanoes record paleoceanographic changes accompanying closure of the Panama gateway. Geology, 30:1087–1090.

80 Platt, U. and Stutz, J. (2008). Differential Optical Absorption Spectroscopy: Principles and Ap- plications. Springer-Verlag Berlin Heidelberg, Germany, 1 edition.

Ranero, C. R., von Huene, R., Flueh, E., Duarte, M., Baca, D., and McIntosh, K. (2000). A cross section of the convergent paciﬁc margin of Nicaragua. Tectonics, 19:335–357.

Rickenmann, D. (1999). Empirical relationships for debris ﬂows. Natural Hazards, 19:47–77. 10.1023/A:1008064220727.

Rickenmann, D. (2005). Runout prediction methods. In Jakob, M., Hungr, O., and Rickenmann, D., editors, Debris–ﬂow Hazards and Related Phenomena, Springer Praxis Books, pages 305–324. Springer Berlin Heidelberg. 10.1007/3-540-27129-5 13.

Rodolfo, K. (1989). Origin and early evolution of lahar channel at Mabinit, Mayon Volcano, Philippines. Geological Society of America Bulletin, 101(3):414–426.

Rout, D. J., Cassidy, J., Locke, C. A., and Smith, I. E. M. (1993). Geophysical evidence for temporal and structural relationships within the monogenetic basalt volcanoes of the Auckland volcanic ﬁeld, northern New Zealand. Journal of Volcanology and Geothermal Research, 57:71– 83.

Rymer, H. and Brown, G. C. (1986). Gravity ﬁelds and the interpretation of volcanic structures: geological discrimination and temporal evolution. Journal of Volcanology and Geothermal Re- search, 27:229–254.

Saballos, J. A., Malservisi, R., Connor, C., Femina, P. L., and Wetmore, P. (2013). Gravity and geodesy at Concepcion´ volcano, Nicaragua. In Rose, W. I., editor, Open Vent Volcanoes. Geo- logical Society of America. In Press.

Saginor, I., Gazel, E., Carr, M., Swisher, C. C., and Turrin, B. (2011). New Pliocene–Pleistocene 40Ar/39Ar ages ﬁll in temporal gaps in the Nicaraguan volcanic record. Journal of Volcanology and Geothermal Research, 202:143–152.

Schilling, S. (1998). LAHARZ: GIS programs for automated mapping of lahar-inundation hazard zones. Open ﬁle report 96–178, US Geological Survey, Vancouver, Washington U.S.A.

Schulz, R., Buness, H., Gabriel, G., Pucher, R., Rolf, C., Wiederhold, H., and Wonik, T. (2005). Detailed investigation of preserved maar structures by combined geophysical surveys. Bulletin of Volcanology, 68:95–106.

Scott, K. M., Mac´ıas, J. L., Naranjo, J. A., Rodr´ıguez, S., and McGeehin, J. P. (2001). Catastrophic debris ﬂows transformed from landslides in volcanic terrains: mobility, hazard assessment, and mitigation strategies. Profesional Paper 1630, US Geological Survey, Reston, Virginia.

Sidle, R. (1992). A theoretical model of the effects of timber harvesting on slope stability. Water Resources Research, 28(7):1897–1910.

Sidle, R. and Ochiai, H. (2006). Landslides: processes, prediction, and land use. American Geophysical Union. Water Resources Monograph 18.

81 Siebert, L., Simkin, T., and Kimberly, P. (2011). Volcanoes of the World. University of California Press, Berkeley, 3 edition.

Skermer, N. A. and VanDine, D. F. (2005). Debris ﬂows in history. In Jakob, M. and Hungr, O., editors, Debris–ﬂow Hazards and Related Phenomena, pages 25–51. Praxis. Springer Berlin Heidelberg, Chichester, UK.

Stevens, N., Manville, V., and Heron, D. (2003). The sensitivity of a volcanic ﬂow model to digital elevation model accuracy: experiments with digitised map contours and interferometric SAR at Ruapehu and Taranaki volcanoes, New Zealand. Journal of Volcanology and Geothermal Research, 119(1–4):89–105.

Sundblad, K., Cumming, G. L., and Krstic, D. (1991). Lead isotope evidence for the formation of epithermal gold quartz veins in the Chortis block, Nicaragua. Economic Geology, 86:944–959.

Swain, F. M. (1966). Bottom sediments of Lake Nicaragua and Lake Managua, western Nicaragua. Journal of Sedimentary Research, 36:522–540.

Symonds, R., Gerlach, T., and Reed, M. (2001). Magmatic gas scrubbing: implications for volcano monitoring. Journal of Volcanology and Geothermal Research, 108(1-4):303–341.

Thouret, J. (1999). Volcanic geomorphology–an overview. Earth-science reviews, 47(1-2):95–131.

Tucker, C. (1979). Red and photographic infrared linear combinations for monitoring vegetation. Remote sensing of Environment, 8(2):127–150.

Turner, H. L., LaFemina, P., Saballos, A., Mattioli, G. S., Jansma, P. E., and Dixon, T. (2007). Kine- matics of the Nicaraguan forearc from GPS geodesy. Geophysical Research Letters, 34:L02302, doi:10.1029/2006GL027586.

Vallance, J. (2000). Lahars. In Sigurdsson, H., Houghton, B., McNutt, S. R., Rymer, H., and Stix, J., editors, Encyclopedia of Volcanology, pages 601–616. Academic Press.

Vallance, J. W. (2005). Volcanic debris ﬂows. In Jakob, M. and Hungr, O., editors, Debris–ﬂow Hazards and Related Phenomena, pages 247–274. Praxis. Springer Berlin Heidelberg, Chich- ester, UK.

Vallance, J. W., Schilling, S. P., Devoli, G., and Howell, M. M. (2001). Lahar hazards at Con- cepcion´ volcano, Nicaragua. Open ﬁle report 01–457, US Geological Survey, Vancouver, Wash- ington U.S.A.

van Wyk de Vries, B. (1993). Tectonics and magma evolution of Nicaraguan volcanic systems. Ph.d. dissertation, Open University, Milton Keynes.

van Wyk de Vries, B. and Borgia, A. (1996). The role of basement in volcano deformation. In McGuire, W. J., Jones, A. P., and Neuberg, J., editors, Volcano Instability on the Earth and Other Planets, pages 95–110. Geological Society of London.

van Wyk de Vries, B. and Matela, R. (1998). Styles of volcano-induced deformation: numerical models of substratum ﬂexure, spreading and extrusion. Journal of Volcanology and Geothermal Research, 81:1–18.

82 Vianello, A., Cavalli, M., and Tarolli, P. (2009). LiDAR-derived slopes for headwater channel network analysis. Catena, 76(2):97–106.

Voight, B. and Elsworth, D. (1997). Failure of volcano slopes. Geotechnique, 47(1):1–31.

Voight, B., Sparks, R. S. J., Miller, A. D., Stewart, R. C., Hoblitt, R. P., Clarke, A., Ewart, J., Aspinall, W. P., Baptie, B., Calder, E. S., Cole, P., Druitt, T. H., Hartford, C., Herd, R. A., Jackson, P., Lejeune, A. M., Lockhart, A. B., Loughlin, S. C., Luckett, R., Lynch, L., Norton, G. E., Robertson, R., Watson, I. M., Watts, R., and Young, S. R. (1999). Magma ﬂow instability and cyclic activity at Soufriere` Hills Volcano, Montserrat, British West Indies. Science, 283(5405):1138 – 1142.

Watson, I., Oppenheimer, C., Voight, B., Francis, P., Clarke, A., Stix, J., Miller, A., Pyle, D., Burton, M., Young, S., Norton, G., Loughlin, S., and Darroux, B. (2000). The relationship between degassing and ground deformation at Soufriere` Hills Volcano, Montserrat. Journal of Volcanology and Geothermal Research, 98(1 – 4):117 – 126.

Watters, R., Zimbelman, D., Bowman, S., and Crowley, J. (2000). Rock mass strength assessment and signiﬁcance to ediﬁce stability, mount rainier and mount hood, cascade range volcanoes. Pure and Applied Geophysics, 157(6-8):957–976.

Weinberg, R. F. (1992). Neotectonic development of western Nicaragua. Tectonics, 11:1010–1017.

Weyl, R. (1980). Geology of Central America. Beitrage¨ zur Regionalen Geologie der Erde, Berlin, 2 edition.

Williams-Jones, G., Stix, J., Heiligmann, M., Barquero, J., Fernandez,´ E., and Gonzalez,´ E. (2001). A model of degassing and seismicity at Arenal Volcano, Costa Rica. Journal of volcanology and geothermal research, 108(1-4):121–139.

Yokoyama, I. (1963). Structure of caldera and gravity anomaly. Bulletin of Volcanology, 26:67–72.

Zumberge, J. F., Heﬂin, M. B., Jefferson, D. C., Watkins, M. M., and Webb, F. H. (1997). Precise point positioning for the efﬁcient and robust analysis of GPS data from large networks. Journal of Geophysical Research, 102:5005–5017.

83 APPENDICES

84 Appendix A Supplementary ﬁgures for Chapter 2

Figure A.1. Simple Bouguer anomaly map computed using a density of 2500 kg m−3 used in gravitational spreading models (e.g. Borgia and van Wyk de Vries, (2003)). Note the narrow steep negative ring-like gradient formed around the volcanic cone. This is due to an overcorrection caused by a density that is higher than the bulk average terrain density.

85 Appendix A (Continued)

Figure A.2. Simple Bouguer anomaly map computed using a density of 1539 kg m−3 provided with the 1-D Nettleton (Nettleton, 1939) and Parasnis’(Parasnis, 1997) methods. Although this density does a better job in removing the effect of topography from the gravity anomaly than a density of 2500 kg m−3, there is still some positive correlation between the volcanic topography and the Bouguer anomaly caused by a lower density than the best bulk average density that represents the terrain.

86 Appendix A (Continued)

N W E

E B W Density contrast 3

650500 (kg/m )

0 5 Km

S A 1275550 0 N S

0 -5Km -2 km C

Figure A.3. 3-D inversion of the gravity data shown in Figure 2.2 using the GROWTH2.0 model developed by Camacho et al., (2011). The input parameters used in the modelling correspond to the second set of parameters summarized in Table 2.1. (A) Structural density map at 1 km depth (-1000 m) of Concepcion´ volcano and its surroundings. (B) East-West cross-section of structural density map. (C) North-South cross-section of structural density map.

87 Appendix A (Continued)

N W E

W E B 650500 Density contrast (kg/m3)

0 5 Km

S A 1275550 0 N S

0 -5 Km -3 km C

Figure A.4. 3-D inversion of the gravity data shown in Figure 2.2 using the GROWTH2.0 model developed by Camacho et al., (2011). The input parameters used in the modelling correspond to the third set of parameters summarized in Table 2.1. (A) Structural density map at 1 km depth (-1000 m) of Concepcion´ volcano and its surroundings. (B) East-West cross-section of structural density map. (C) North-South cross-section of structural density map.

88 Appendix A (Continued)

JAN 02 JAN 03 JAN 04 JAN 05 JAN 06 JAN 07 JAN 08 JAN 09 JAN 10

10 CON1

−5 Latitude (mm) −10

−15

CON1 30

0 Longitude (mm) −10

−20

80 CON1

Height (mm) −40

−80

JAN 02 JAN 03 JAN 04 JAN 05 JAN 06 JAN 07 JAN 08 JAN 09 JAN 10

Figure A.5. Time series for the three components of station CON1 for all the campaigns. Light- gray vertical bars indicate periods of unusual seismic activity on Concepcion´ volcano and/or near Ometepe Island, while dark-gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. CON1 is located slightly NW from Concepion,´ see Figure 1.

89 Appendix A (Continued)

OCT 08 JAN 09 APR 09 JUL 09 OCT 09 JAN 10 APR 10 JUL 10

5 5

0 0

−5 −5 COS2 −10 −10 Latitude (mm) Latitude (mm) −15 −15

−20 −20 5 5

COS2

0 0

−5 −5 Longitude (mm) Longitude (mm)

−10 −10

COS2 150 150

100 100

Height (mm) 50 50 Height (mm)

0 0

OCT 08 JAN 09 APR 09 JUL 09 OCT 09 JAN 10 APR 10 JUL 10

Figure A.6. Time series for the three components of station COS2. Light-gray vertical bars indicate periods of unusual seismic activity on Concepcion´ volcano and/or near Ometepe Island, while dark- gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. COS2 site, located slightly SW of the volcano, in the San Jose´ del Sur town (Figure 1), shows the largest vertical variation through time, an increase, of ∼13–16 cm in 1.932 yrs.

90 Appendix A (Continued)

OCT 08 JAN 09 APR 09 JUL 09 OCT 09 JAN 10 APR 10 JUL 10

MOYO 10

0 Latitude (mm)

−5

15 MOYO

0 Longitude (mm)

−5

0 Height (mm) −20 MOYO

−40

OCT 08 JAN 09 APR 09 JUL 09 OCT 09 JAN 10 APR 10 JUL 10

Figure A.7. Time series for the three components of station MOYO. Light-gray vertical bars indicate periods of unusual seismic activity on Concepcion´ volcano and/or near Ometepe Island, while dark-gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. MOYO is located of the western side of the volcano, near the town of Moyogalpa, see Figure 1.

91 Appendix A (Continued)

OCT 08 JAN 09 APR 09 JUL 09 OCT 09 JAN 10 APR 10 JUL 10

5 SABA

−5 Latitude (mm)

−10

15 SABA 10

0 Longitude (mm)

−5

10 SABA

−10

−20 Height (mm)

−30

−40

OCT 08 JAN 09 APR 09 JUL 09 OCT 09 JAN 10 APR 10 JUL 10

Figure A.8. Time series for the three components of station SABA. Light-gray vertical bars indicate periods of unusual seismic activity on Concepcion´ volcano and/or near Ometepe Island, while dark- gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. SABA is located in the northeast from Concepcion´ volcano in the community of La Sabana, see Figure 1.

92 Appendix A (Continued)

JAN 02 JAN 03 JAN 04 JAN 05 JAN 06 JAN 07 JAN 08 JAN 09 JAN 10

0 0 SINT

−25 −25

−50 −50 Latitude (mm) Latitude (mm)

−75 −75

10 10 SINT

0 0

−10 −10

−20 −20 Longitude (mm) Longitude (mm)

−30 −30

SINT 50 50

0 0 Height (mm) Height (mm)

−50 −50

JAN 02 JAN 03 JAN 04 JAN 05 JAN 06 JAN 07 JAN 08 JAN 09 JAN 10

Figure A.9. Time series for the three components of station SINZ. Light-gray vertical bars indicate periods of unusual seismic activity on Concepcion´ volcano and/or near Ometepe Island, while dark- gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. SINT is located southeast from Concepcion´ volcano, in the Sintiope town, see Figure 1. In April 2010 the GPS site COS1, see Figure 1, was tied to the new site SINT because a water well pump was built a few meters from COS1 in early 2010.

93 Appendix A (Continued)

CON1 COS2 SABA SINT MOYO

March April May 2010 June July August 2010 21 28 4 11 18 25 2 9 16 23 30 6 13 20 27 4 11 18 25 1 8

40 40

30 30

20 20

10 10 Longitude (mm) Longitude (mm)

0 0

−10 −10

21 28 4 11 18 25 2 9 16 23 30 6 13 20 27 4 11 18 25 1 8 March April May 2010 June July August 2010

Figure A.10. Time series changes for the Longitude component of all stations around Concepcion´ volcano during April through July 2010. Dark-gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. Time series have been offset for better visualization.

94 Appendix A (Continued)

CON1 COS2 SABA SINT MOYO

March April May 2010 June July August 2010 21 28 4 11 18 25 2 9 16 23 30 6 13 20 27 4 11 18 25 1 8

40 40

30 30

20 20

10 10 Latitude (mm) Latitude (mm) 0 0

−10 −10

−20 −20 21 28 4 11 18 25 2 9 16 23 30 6 13 20 27 4 11 18 25 1 8 March April May 2010 June July August 2010

Figure A.11. Time series changes for the Latitude component of all stations around Concepcion´ volcano during April through July 2010. Dark-gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. Time series have been offset for better visualization.

95 Appendix A (Continued)

CON1 COS2 SABA SINT MOYO

March April May 2010 June July August 2010 21 28 4 11 18 25 2 9 16 23 30 6 13 20 27 4 11 18 25 1 8 200 200

160 160

120 120

80 80

Height (mm) 40 40 Height (mm)

0 0

−40 −40

21 28 4 11 18 25 2 9 16 23 30 6 13 20 27 4 11 18 25 1 8 March April May 2010 June July August 2010

Figure A.12. Time series changes for the vertical component of all stations around Concepcion´ volcano during April through July 2010. Dark-gray vertical bars represent periods of volcanic activity characterized by ash and gas explosions. Time series have been offset for better visualization.

96 Appendix A (Continued)

JAN 07 JUL 07 JAN 08 JUL 08 JAN 09 JUL 09 JAN 10 JUL 10 JAN 11 JUL 11 JAN 12

70 70 MANA North component velocity 9.5 +- 0.4 mm/yr 56 56 Mean Latitude: 12° 8' 56.186" N

42 42

28 28

14 14 Latitude change (mm) Latitude change (mm)

0 0

JAN 07 JUL 07 JAN 08 JUL 08 JAN 09 JUL 09 JAN 10 JUL 10 JAN 11 JUL 11 JAN 12

Figure A.13. Temporal changes of Latitude component of GPS station MANA between January 2007 and August 2010. This continuous site is part of International GPS Service, IGS, network, and located inside the INETER’s facilities in Managua, Nicaragua. Red line corresponds to best linear ﬁt. Velocity uncertainties were computed using the Hackl et al., ( 2011) algorithm.

JAN 07 JUL 07 JAN 08 JUL 08 JAN 09 JUL 09 JAN 10 JUL 10 JAN 11 JUL 11 JAN 12

48 MANA East component velocity 7.8 +- 0.7 mm/yr 48 40 Mean Longitude: 86° 14' 56.372" W 40 32 32

24 24

16 16

8 8

0 0 Longitude change (mm) Longitude change (mm) −8 −8

JAN 07 JUL 07 JAN 08 JUL 08 JAN 09 JUL 09 JAN 10 JUL 10 JAN 11 JUL 11 JAN 12

Figure A.14. Temporal changes of Longitude component of GPS station MANA between January 2007 and August 2010. This continuous site is part of International GPS Service, IGS, network, and located inside the INETER’s facilities in Managua, Nicaragua. Red line corresponds to best linear ﬁt. Velocity uncertainties were computed using the Hackl et al., ( 2011) algorithm.

97 Appendix B Concepcion’s´ volcano gravity data

Table B.1: Gravity data gathered at Concepcion´ volcano,

Nicaragua, during campaigns carried out between 2007 and 2010.

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

(m) (m) (m) (mGal) (mGal) (mGal) (mGal) (mGal)

642913 1275747 105.925 1589.965 0.000 0.000 0.376 0.000

643630 1275619 158.678 1578.540 4.902 0.930 1.341 -0.340

644681 1275489 140.387 1588.745 9.501 6.906 7.189 3.535

645837 1274976 198.482 1581.769 20.625 13.657 14.258 8.805

646244 1274827 229.666 1577.131 25.662 16.348 17.065 10.947

646256 1275207 231.220 1578.301 27.191 17.760 18.228 11.706

643286 1276242 145.650 1582.836 4.978 1.986 2.425 0.809

641641 1275978 38.107 1605.268 -5.707 -0.598 -0.269 1.669

644788 1276357 141.198 1591.418 12.149 9.492 9.691 4.955

645210 1277272 146.492 1595.737 17.811 14.756 15.140 8.646

645329 1276476 154.159 1594.440 19.134 15.502 15.924 9.988

645475 1276229 164.102 1593.246 21.089 16.708 16.967 10.985

644429 1278116 123.104 1597.601 12.186 10.892 11.116 5.339

643839 1279292 82.325 1608.611 10.232 12.009 12.161 6.389

644117 1278848 98.139 1604.368 11.012 11.599 12.754 6.869

644893 1279168 99.602 1605.777 12.771 13.247 13.326 5.570

644473 1274814 125.135 1588.290 4.553 3.106 3.417 0.854

643794 1274073 99.862 1589.148 -2.156 -1.699 -1.532 -1.996

642769 1273951 76.140 1592.990 -5.599 -3.355 -2.786 -1.077

644188 1273830 127.357 1582.405 -0.333 -1.947 -1.518 -2.527

Continued on Next Page. . .

98 Appendix B (Continued)

Table B.1 – Continued

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

645277 1273977 141.234 1583.780 5.279 2.620 3.047 -0.286

645969 1273403 148.365 1583.388 7.272 4.076 4.562 0.418

646562 1273852 203.235 1580.839 21.520 14.194 14.484 8.705

643753 1273070 100.114 1584.683 -6.224 -5.786 -5.360 -4.739

642694 1272782 95.061 1582.966 -9.409 -8.591 -7.454 -4.426

641649 1272448 50.269 1592.642 -13.455 -9.262 -8.570 -3.119

652319 1273172 238.047 1589.819 41.464 31.520 32.017 15.404

653752 1274041 172.654 1607.606 38.785 33.761 34.233 13.886

653787 1274993 176.419 1607.043 39.082 33.774 34.328 12.958

653954 1273189 141.881 1612.601 34.551 31.843 32.345 12.445

654412 1272199 118.701 1609.772 24.881 23.918 25.067 5.242

655217 1273671 114.211 1612.886 26.140 25.516 26.351 3.444

653415 1272173 130.944 1605.552 24.449 22.565 23.376 5.570

654626 1273988 143.128 1611.556 33.637 30.835 31.112 9.069

652239 1272659 160.640 1600.205 28.115 23.995 24.404 8.464

652754 1271268 69.320 1617.875 18.035 20.792 21.512 5.934

653186 1270613 39.296 1624.024 15.123 20.143 20.583 4.796

652359 1270725 60.386 1616.705 14.280 17.710 18.237 3.991

651412 1270076 42.158 1612.330 4.484 9.288 9.909 -1.793

650697 1270751 54.765 1613.126 8.957 12.811 12.954 2.007

650015 1271178 102.493 1602.630 13.060 13.318 14.039 4.028

641870 1276813 38.704 1608.238 -2.819 2.246 2.610 3.254

642258 1276789 54.044 1605.001 -1.312 2.596 3.035 2.928

Continued on Next Page. . .

99 Appendix B (Continued)

Table B.1 – Continued

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

643394 1277469 111.085 1595.530 6.610 6.221 6.870 3.810

646422 1277209 202.133 1592.392 31.664 24.421 24.715 15.859

646232 1278864 133.110 1605.009 22.445 20.397 20.560 10.430

646507 1279488 110.087 1609.302 19.431 19.118 19.403 8.098

642733 1278137 59.677 1608.877 3.873 7.357 7.902 5.496

642471 1278964 41.488 1615.808 4.926 9.781 10.179 7.471

641496 1274501 50.616 1598.321 -8.322 -4.156 -3.616 0.088

641898 1274093 59.439 1595.725 -8.065 -4.563 -3.995 -0.687

642676 1274429 74.194 1594.593 -4.749 -2.358 -1.821 -0.405

649911 1278800 261.012 1585.262 42.206 30.534 31.095 13.670

649818 1279881 153.087 1607.107 30.386 26.835 27.617 9.297

649394 1278884 223.552 1594.105 39.457 30.602 30.955 14.481

649024 1280046 136.761 1608.379 26.565 24.243 24.363 7.466

647881 1281528 38.755 1625.467 12.925 17.986 18.348 2.255

647354 1280866 54.480 1620.729 13.252 17.128 17.419 3.042

648131 1280908 57.651 1621.678 15.167 18.804 18.923 2.951

646144 1280984 42.965 1622.293 11.224 15.967 16.365 4.290

645430 1280722 41.786 1622.500 11.150 15.983 16.485 6.100

645610 1279857 102.983 1606.956 14.774 14.996 16.028 6.148

644632 1279521 88.176 1608.707 12.061 13.398 13.884 6.296

647005 1279559 104.647 1611.441 19.868 19.964 20.049 7.678

648408 1278660 219.691 1592.516 36.747 28.183 28.356 14.078

648111 1279528 132.930 1607.062 24.231 22.197 22.575 8.022

Continued on Next Page. . .

100 Appendix B (Continued)

Table B.1 – Continued

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

650271 1278714 270.696 1583.122 43.083 30.683 31.483 13.424

650327 1280356 115.180 1613.526 24.951 24.254 24.319 4.507

650228 1281713 38.020 1626.057 13.229 18.345 18.636 -2.336

650789 1279590 150.529 1606.071 28.653 25.295 25.527 5.557

650825 1281527 37.810 1626.192 13.359 18.491 18.728 -3.251

651385 1280974 66.545 1618.909 15.123 18.089 18.259 -4.287

651318 1280043 112.597 1613.325 24.052 23.550 23.682 2.201

651710 1278901 195.227 1599.639 36.240 29.517 29.659 8.536

652057 1279185 149.131 1607.957 30.237 26.983 27.310 5.208

652328 1279523 112.380 1613.161 23.987 23.501 23.559 0.577

652253 1281038 72.731 1612.673 10.776 13.277 13.850 -10.496

651456 1281704 38.307 1622.873 10.137 15.231 15.507 -7.912

653233 1278845 99.901 1613.274 20.463 20.917 21.238 -2.875

652801 1278248 160.215 1605.040 31.040 26.952 28.527 5.874

653498 1279700 74.840 1612.934 12.114 14.456 14.485 -11.013

654042 1279048 83.182 1611.247 13.210 14.923 15.187 -10.748

655058 1279416 56.990 1614.218 7.978 11.664 11.867 -16.468

653960 1281249 40.047 1616.784 4.730 9.693 10.205 -17.767

653992 1276957 155.090 1602.190 27.018 23.316 23.565 -0.178

653064 1276721 258.213 1584.687 41.429 29.967 30.744 9.092

655070 1277593 95.907 1608.777 15.132 15.886 16.176 -10.360

655258 1276020 108.420 1609.460 20.178 19.990 20.474 -4.865

656070 1275199 72.781 1616.198 16.175 18.672 18.971 -7.170

Continued on Next Page. . .

101 Appendix B (Continued)

Table B.1 – Continued

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

655964 1274534 78.770 1616.394 18.432 20.477 21.148 -4.116

655526 1275349 107.596 1610.858 21.536 21.410 21.800 -3.404

657696 1272613 45.825 1626.500 18.978 23.506 23.814 -2.994

656820 1273910 64.052 1619.446 17.139 20.293 22.366 -3.987

651499 1272529 177.142 1594.645 27.691 22.329 22.997 8.667

654277 1274959 142.781 1614.219 35.883 33.108 33.204 10.888

654985 1276794 120.984 1603.720 18.071 16.937 17.199 -8.368

655010 1278431 74.779 1612.348 11.913 14.259 14.481 -12.772

653990 1280460 51.830 1615.196 7.031 11.106 11.233 -16.010

656647 1279134 38.831 1615.364 3.608 8.662 9.018 -22.213

647387 1270445 58.266 1596.438 -6.553 -2.962 -2.323 -6.345

646833 1271511 95.110 1589.802 -2.153 -1.339 -1.260 -5.240

647639 1269651 39.100 1598.514 -10.141 -5.106 -4.686 -8.417

648505 1270025 54.514 1599.178 -4.837 -0.964 -0.426 -6.264

649696 1270922 104.407 1598.582 9.684 9.799 10.007 0.891

650655 1271567 131.114 1600.110 19.252 17.355 18.219 6.540

647915 1271509 126.049 1589.625 7.222 5.706 6.134 -0.008

645415 1271396 80.280 1588.394 -8.103 -6.172 -5.369 -6.398

646159 1270421 61.728 1592.115 -9.799 -6.470 -5.799 -7.341

645634 1270673 65.365 1589.852 -11.020 -7.964 -7.268 -8.012

644801 1270898 64.839 1586.911 -14.195 -11.100 -10.333 -9.635

644090 1269352 38.042 1589.827 -19.059 -13.945 -13.621 -9.956

643371 1271091 70.931 1584.328 -14.958 -12.322 -12.003 -8.638

Continued on Next Page. . .

102 Appendix B (Continued)

Table B.1 – Continued

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

642328 1271172 70.007 1584.757 -14.840 -12.135 -11.603 -6.233

646589 1277167 212.413 1591.155 33.614 25.598 26.301 17.152

647474 1277059 297.922 1573.621 42.515 28.068 28.712 17.902

647832 1276928 390.459 1554.130 51.637 30.236 33.307 21.913

648253 1276619 600.936 1502.596 65.193 27.997 29.398 17.470

648838 1276259 904.131 1425.723 82.067 22.174 23.231 10.494

649071 1276172 988.414 1404.448 86.850 20.660 24.222 11.106

649504 1275934 1195.661 1349.615 96.105 14.449 16.386 2.641

649689 1275870 1311.171 1317.136 99.326 9.064 11.171 -2.880

654195 1277286 132.529 1606.946 24.705 22.701 23.167 -1.312

653552 1276794 200.131 1595.773 34.559 27.467 27.894 5.194

652722 1276558 315.535 1571.989 46.481 30.710 31.062 10.257

652199 1276348 517.477 1523.349 60.261 29.325 32.022 12.474

651927 1276374 641.580 1493.015 68.241 27.999 30.251 11.221

651470 1276382 877.125 1433.725 81.690 23.816 26.156 8.031

651087 1276267 1105.023 1374.698 93.086 18.190 20.571 3.329

650791 1276094 1329.744 1311.206 99.062 7.417 10.250 -6.229

642054 1276312 50.620 1604.202 -3.017 1.150 1.612 2.389

642825 1277534 65.194 1605.870 2.762 5.830 6.204 4.218

643751 1277432 130.432 1592.163 9.228 7.382 8.081 4.345

644710 1277695 133.897 1595.798 13.849 11.742 12.456 6.539

645458 1278102 141.585 1600.686 20.980 18.295 18.802 10.981

646174 1278762 136.546 1604.242 22.771 20.465 20.541 10.628

Continued on Next Page. . .

103 Appendix B (Continued)

Table B.1 – Continued

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

646871 1279279 116.008 1610.132 22.156 21.396 21.570 9.746

647693 1279580 111.993 1610.698 21.386 20.929 21.062 7.293

648654 1279579 145.432 1606.873 27.885 24.910 25.153 9.464

649679 1279305 191.111 1600.379 35.581 29.168 29.374 11.908

650527 1279565 160.635 1604.874 30.584 26.465 26.708 7.285

651470 1279572 148.532 1605.758 27.729 24.521 24.696 3.382

652459 1279507 109.587 1613.492 23.461 23.185 23.394 0.167

653221 1279083 93.663 1612.815 18.003 18.926 19.110 -5.217

654384 1279010 83.391 1612.088 14.128 15.825 16.479 -10.103

655035 1278452 74.268 1612.854 12.255 14.640 14.891 -12.434

655104 1277798 91.092 1610.377 15.180 16.297 16.694 -10.114

654986 1276808 121.389 1604.263 18.735 17.570 17.860 -7.722

655266 1275949 107.170 1610.891 21.246 21.152 21.528 -3.757

654820 1275103 122.784 1616.851 32.296 31.026 31.550 8.004

642154 1275518 61.640 1599.114 -4.450 -1.114 -0.897 0.474

641693 1274716 50.569 1598.823 -7.903 -3.733 -3.317 -0.222

641244 1273743 46.589 1596.644 -11.001 -6.531 -6.106 -1.141

641028 1272857 44.640 1594.862 -13.103 -8.486 -8.108 -1.824

641502 1271958 53.574 1590.391 -14.530 -10.586 -10.354 -4.119

642084 1271394 63.220 1587.144 -14.619 -11.402 -10.931 -5.296

643246 1271051 74.935 1583.998 -14.039 -11.705 -10.983 -7.328

644156 1270931 76.033 1583.823 -13.837 -11.586 -10.460 -8.506

645137 1270793 62.781 1588.620 -13.088 -9.838 -9.253 -9.124

Continued on Next Page. . .

104 Appendix B (Continued)

Table B.1 – Continued

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

645994 1270524 67.298 1591.515 -8.713 -5.803 -4.792 -6.106

647431 1270062 49.779 1597.669 -7.820 -3.590 -2.934 -6.661

648460 1270066 55.295 1600.025 -3.762 0.052 0.702 -5.087

649264 1270361 94.799 1594.899 3.214 4.052 4.264 -3.426

650015 1271176 103.202 1602.630 13.279 13.485 14.305 4.297

650981 1271914 147.885 1599.511 23.720 20.561 21.159 8.479

652067 1273050 214.720 1593.338 37.820 29.630 30.239 14.252

652942 1273572 220.145 1596.149 42.140 33.542 34.234 15.976

653695 1274056 177.736 1606.313 39.056 33.650 34.185 13.936

654604 1274926 130.870 1616.448 34.446 32.567 32.684 9.748

650023 1278477 314.573 1576.688 50.272 34.573 34.932 17.606

649793 1278190 417.957 1551.537 57.133 33.667 34.972 18.393

649903 1277780 569.667 1515.270 67.842 32.991 33.781 17.392

645908 1274484 187.059 1586.424 21.910 15.801 16.569 11.466

646052 1273972 183.461 1585.696 20.234 14.396 15.495 10.617

646524 1273918 210.365 1583.178 26.039 18.177 18.901 13.132

646997 1274043 259.213 1575.810 33.713 22.176 24.873 18.033

647586 1274552 352.700 1559.193 45.798 27.234 29.744 21.217

647894 1274524 393.456 1550.621 49.819 28.193 31.250 22.135

648485 1274761 546.264 1516.595 62.902 29.805 31.946 21.412

649100 1274976 793.274 1457.959 80.475 28.874 34.493 22.514

640845 1274192 41.655 1602.394 -6.917 -2.075 -1.611 3.704

640440 1273708 38.881 1601.535 -8.479 -3.428 -2.983 3.626

Continued on Next Page. . .

105 Appendix B (Continued)

Table B.1 – Continued

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

641153 1273889 46.878 1600.776 -6.827 -2.378 -1.842 3.160

639678 1272584 38.268 1600.381 -9.464 -4.367 -3.987 5.270

641915 1272041 55.157 1594.076 -10.383 -6.558 -5.961 -0.635

642140 1272076 61.462 1589.569 -12.954 -9.605 -8.600 -3.759

642737 1272214 98.542 1580.712 -10.408 -9.852 -7.991 -4.483

642724 1271820 81.336 1583.638 -12.668 -10.816 -8.954 -5.024

642199 1271745 59.380 1589.229 -13.831 -10.325 -9.661 -4.606

648389 1272059 163.277 1587.878 16.793 12.475 13.434 5.794

648689 1272962 244.662 1575.685 29.440 18.997 19.488 10.346

648760 1273693 344.361 1557.201 41.505 23.567 24.947 14.931

648773 1274013 410.686 1542.998 47.679 24.758 26.112 15.750

648920 1274244 466.839 1530.713 52.659 25.522 26.490 15.602

642775 1271698 68.205 1586.270 -14.051 -11.210 -10.650 -6.701

643326 1272180 84.868 1584.047 -11.284 -9.697 -9.101 -6.736

644286 1272332 90.152 1585.341 -8.406 -7.218 -6.543 -6.248

643807 1273113 101.761 1584.583 -5.829 -5.516 -5.086 -4.615

644311 1273696 144.993 1577.486 0.235 -2.707 -1.192 -2.314

645585 1273936 152.678 1582.705 7.750 4.230 4.807 0.899

646053 1273726 166.147 1583.262 12.533 7.998 8.595 3.959

646308 1273481 164.033 1584.805 13.501 9.126 9.564 4.664

642664 1274217 74.511 1593.780 -5.396 -3.030 -2.370 -0.718

652090 1273441 278.896 1580.770 44.942 31.925 33.133 16.710

651427 1273525 309.475 1571.303 44.889 29.573 30.077 14.895

Continued on Next Page. . .

106 Appendix B (Continued)

Table B.1 – Continued

Easting Northing Elevation Obs. g.a Free-airb Bullard Bc Bullard Cd Residuale

651210 1273981 454.333 1537.190 55.358 29.160 31.431 16.227

651137 1274268 581.557 1505.984 63.346 27.603 29.776 14.433 aObserved gravity corrected for earth solid tide and instrumental drift (mGal). bFree-air gravity anomaly (mGal). cSimple Bouguer anomaly (mGal). dComplete Bouguer anomaly, terrain corrected (mGal). eResidual gravity anomaly left after removal of the regional ﬁeld (mGal).

107 Appendix C PERL script to compute the complete terrain correction for gravity data

## PERL SCRIPT TO COMPUTE THE TERRAIN CORRECTION OF GRAVITY DATA

## BY J. A. SABALLOS, USF-GEOLOGY-VOLCANO GROUP

## The terrain correction (TC) is implemented in three stages,

## described below:

## 1] Inner TC up to Hammer’s zone C, 53.3 m using the quarter-wedge

## method described by Nowel (1999, JAG), which is an improved

## version of the Campbell (1980, Geophysics) technique.

## 2] Intermediate TC from Hammer’s zone D up to a desired distance,

## e.g Hammer’s zone K, but it’s up to the user. This stage is done

## using the simplified gravity attraction of a prism approximated as

## an annular ring, described by Kane (1962, Geophysics). This

## method computes the TC much faster than the Plouff (1976) or

## Nagy (1966) methods that consist of up to 24 terms per prism,

## while Kane (1962) is just a single equation with 4 terms.

## 3] Far-field or Outer TC, for distances beyond the the Intermediate

## TC up to the extent of the input DEM. This stage is done using

## the vertical line mass described by Blais & Ferland (1984, CJES),

## which is the simplified approximation of the gravity attraction of a

## prism in the far-field

##########################################################

## INPUT DEM format: It must be a 3-column file like this:

108 Appendix C (Continued)

## Column 1: X-UTM coordinate in meters.

## Column 2: Y-UTM coordinate in meters.

## Column 3: elevation in meters.

## INPUT Gravity data file: It is a 4-column file format like this

## Column 1: X-UTM station coordinate in meters.

## Column 2: Y-UTM station coordinate in meters.

## Column 3: Bouguer anomaly to be terrain corrected in milli Gals.

## Column 4: satation elevation in meters.

##########################################################

#!/usr/bin/perl

######## NAME OF INPUT DEM FILE IN THE NEXT LINE #########

$in_dem="dem.xyz"; # Replace ’dem.xyz’ by your input DEM file name.

####### NAME OF INPUT Gravity FILE IN THE NEXT LINE #########

$in_grav="gravity.dat"; #GRAVITY FILE to be terrain corrected (mGals)

### NAME OF THE OUTPUT GRIDDED FILE IN THE NEXT LINE ####

$out_cor="terrain_corrections.dat"; # Output file containing the TCs

## FORMAT OF THIS OUTPUT FILE:

## Column 1: X-UTM COORDINATE in meters

## Column 2: Y-UTM COORDINATE in meters

## Column 3: Complete Bouguer anomaly, Bullard-C, i.e., gravity

## input data + terrain corrections (mGals)

109 Appendix C (Continued)

## Column 4: Terrain Correction only in mGals

##########################################################

############### DEFINING GLOBAL CONSTANTS ##############

##########################################################

$rho=1764; ## TERRAIN’S DENSITY IN kg/m^3 #

$near=53.3; # Maximum extent of the inner terrain zone (m) #

$far=10E+3; # Distance considered to be the far field for outer TC (m)

$gamma=6.672E-11; ## Universal Grav. Const. in m^3/(Kg s^2)

$pi=atan2(1,1)*4;

#########################################################

######## THE SCRIPT IS GONNA TAKE IT FROM HERE #########

######################################################### use Math::Trig;

$dem_ct=0; print"Getting DEM data...\n"; open(DEM,"$in_dem") || die "WRONG input DEM file name: $in_dem";

## GETTING THE DEM DATA ONLY ## while()

{

($dx, $dy, $dz)=split " ", $_;

$x_dem[$dem_ct]=$dx;

$y_dem[$dem_ct]=$dy;

if($dz <= 0)

{

110 Appendix C (Continued)

$z_dem[$dem_ct]=0;## CHANGE THIS ACCORDINGLY ##

}

if($dz > 0)

{

$z_dem[$dem_ct]=$dz;

}

$dem_ct++;

} close(DEM); print"DEM data read successfully\n"; print"the DEM contains $dem_ct points\n";

## LET’S COMPUTE THE DEM "X & Y" RESOLUTION ##

$resx=$resy=0;

$ix=1; # Inner counter in X-direction DEM #

$iy=1; # Inner counter in Y-direction DEM #

# For "X" # until($resx != 0)

{

$resx = $x_dem[$ix] - $x_dem[$ix-1];

if($resx < 0) { $resx = $resx*(-1);}

$ix++;

}

# For "Y" # until($resy != 0)

{

111 Appendix C (Continued)

$resy = $y_dem[$iy] - $y_dem[$iy-1];

if($resy < 0) { $resy = $resy*(-1);}

$iy++;

} print"DEM X-resolution at element $ix-1: $resx (m)\n"; print"DEM Y-resolution at element $iy-1: $resy (m)\n";

#########################################################

##### Getting gravity station data & terrain-uncorrected gravity

## INPUT FILE FORMAT PER COLUMN:

## Column 1: X-UTM (m)

## Column 2: Y-UTM (m)

## Column 3: Elevation (m)

## Column 4: Terrain-uncorrected gravity data (mGals)

######################################################### open(GRAV,"$in_grav") || die "WRONG GRAVITY file name: $in_grav";

$gra_ct=0; print"\nGetting gravity data now...\n"; while()

{

($gx, $gy, $gg, $gz, $ds)=split " ", $_;

$x_st[$gra_ct]=$gx; # Grav. station X-coord (m) #

$y_st[$gra_ct]=$gy; # Grav. station Y-coord (m) #

$g_unc[$gra_ct]=$gg;# Bullard B anomaly #

$z_st[$gra_ct]=$gz; # Grav. station elevation #

$site[$gra_ct]=$ds; # Grav. station ID #

$gra_ct++;

112 Appendix C (Continued)

} close(GRAV); print"A total of $gra_ct gravity data points were read successfully.\n";

## print"Computing the terrain corrections, please be patient...\n";

##### COMPUTING THE TERRAIN CORRECTION ######### for($i3=0; $i3 < $gra_ct; $i3++)

{

$sum_1=$sum_2=$sum_3=0; # For inner , Far & intermediate TCs

for($ct4=0;$ct4 < ($dem_ct-1); $ct4++) ## DEM COUNTER ##

{

$deltax2 = ($x_st[$i3]-$x_dem[$ct4])**2;

$deltay2 = ($y_st[$i3]-$y_dem[$ct4])**2;

$H_rad=sqrt($deltax2 + $deltay2); #2-D distance#

if( ($H_rad <= $near) && ($H_rad > 0) )## INNER TC ##

{

$dist_3D=sqrt(($H_rad**2) +(($z_st[$i3]-$z_dem[$ct4])**2));

$g_inner=0.5*$pi*$gamma*$rho*$H_rad*($dist_3D-$H_rad)/

$dist_3D;

$sum_1+=$g_inner;

}# if, $R <= $near #

if ($H_rad >= $far) # Far distance outer TC #

{

113 Appendix C (Continued)

$xarea=$resy*$resx; # xs-sect. area of the DEM prism #

$d3=sqrt((($H_rad**2) + ($z_st[$i3]-$z_dem[$ct4])**2));

$far_c=$gamma*$rho*$xarea*((1/$H_rad)-(1/$d3));

$sum_3+=$far_c; # Terrain correction for far distances #

}

if (($H_rad < $far) && ($H_rad > $near)) # Interm. outer TC #

{

$side=sqrt(($resy**2)+($resx**2));

$R_minus = $H_rad-(0.63*$side);

$R_plus = $H_rad+(0.63*$side);

$hgt = $z_st[$i3]-$z_dem[$ct4]; # elevation diff. #

$numerator1 = (1.26*$side) + sqrt(($R_minus**2)+

($hgt**2)) - sqrt(($R_plus**2) +

($hgt**2));

$g_int = $gamma*$rho*$side*$numerator1*1E+5/(1.26*$H_rad);

$sum_2+=$g_int; # Intermmediate TC in mGals #

} ## if, ($H_rad < $far) && ($H_rad > $near) ##

} # ’ct4’ for ##

$inner_cor[$i3]= $sum_1*1E+5; # Near-field TC #; ## INNER TC ##

$inter_cor[$i3]= $gamma*$rho*$sum_2*1E+5;# INTERMEDIATE TC

to# Conversion mGals #

$outer_cor[$i3]= $sum_3*1E+5; # Far-field TC ## FAR OUTER TC

} # for($i3 #

##############################################################

################### GENERATING OUTPUT FILE##################

114 Appendix C (Continued)

############################################################## open(CORR,">$out_cor"); print CORR "#X-UTM Y-UTM Complete_B_anomaly(mGal) inner_TC(mGal) interm_TC(mGal) Outer_TC(mGal) site\n"; for($ct3=0;$ct3 < $gra_ct; $ct3++) ## GRAV. STATIONS COUNTER ##

{

## Complete CORRECTED Bouguer Anomaly ##

$BC_cor[$ct3]=$g_unc[$ct3] + $outer_cor[$ct3] + $inter_cor[$ct3] +

$inner_cor[$ct3];

print CORR "$x_st[$ct3] $y_st[$ct3] $BC_cor[$ct3] $inner_cor[$ct3]

$inter_cor[$ct3]

$outer_cor[$ct3] $site[$ct3]\n";

} close(CORR); print"Terrain correction computations are DONE!!\n"; print"The output file you’ve been waiting for is: $out_cor\n\n";

##################### END OF SCRIPT ##########################

115 Appendix D Supplementary ﬁgures for Chapter 3

200 GPS data, 2010 150 GPS data, 2011

100

|Power spectral density| 0 0 5 10 15 20 25 30 35 40 45 Period (days)

150 GPS data, 2010 GPS data, 2011 100

|Power spectral density| 0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Period (days)

Figure D.1. Periodograms for GPS baseline changes during 2010 and 2011. Upper panel: Long periodogram. Bottom panel: Fifteen-day periodogram to better appreciate the short term periodicity.

116 Appendix D (Continued)

600 RSAM data, 2010 RSAM data, 2011 400

200

|Power spectral density| 0 0 5 10 15 20 25 30 35 40 45 Period (days)

600 RSAM data, 2010 RSAM data, 2011 400

200

|Power spectral density| 0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Period (days)

Figure D.2. Periodograms for RSAM daily average during 2010 and 2011. Upper panel: Long periodogram. Bottom panel: Fifteen-day periodogram to better appreciate the short term periodicity.

117 Appendix D (Continued)

4 x 10 6 SO data, 2010 2 SO data, 2011 4 2

|Power spectral density| 0 0 5 10 15 20 25 30 35 40 45 Period (days)

4 x 10 4

2 SO data, 2010 2 1 SO data, 2011 2

|Power spectral density| 0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Period (days)

Figure D.3. Periodograms for SO2 daily average during 2010 and 2011. Upper panel: Long periodogram. Bottom panel: Fifteen-day periodogram to better appreciate the short term periodicity.

118 Appendix D (Continued)

100 100 80 80 60 60 40 40 2010 2011 20 20 RSAM daily RSAM daily average RSAM daily average −20 −15 −10 −5 0 5 10 15 −20 −15 −10 −5 0 5 10 15 A GPS baseline change (mm) B GPS baseline change (mm)

600 600

400 400 (tons/day) (tons/day) 2 200 2 200 SO 2010 SO 2011 0 0 −20 −15 −10 −5 0 5 10 15 −20 −15 −10 −5 0 5 10 15 C GPS baseline change (mm) D GPS baseline change (mm)

100 100 80 80 60 60 40 40 2010 2011 20 20 RSAM daily RSAM daily average RSAM daily average 0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700 SO (tons/day) SO (tons/day) E 2 F 2

Figure D.4. Scatter plots of the data used in this study. For clarity error bars were not plotted. All the 2011 data go from February 4 through April 16, a period when the volcano was not erupting. (A) GPS baseline changes versus RSAM daily average. The 2010 data go from April 08 through June 25. The largest positive GPS baseline changes corresponds to the largest RSAM values, those >83 RSAM units. Amplitude in GPS baseline change is 29 mm. (B) GPS baseline changes versus RSAM daily average for the 2011 period. Amplitude in GPS baseline change is 18 mm (when the volcano was not erupting), which is 11 mm less than the amplitude observed in 2010 during the final phase of the volcano’s most recent erupting activity. There is not any clear correspondence between the 2011 GPS and RSAM data. (C) GPS baseline changes versus SO2 fluxes. These data go from April 28 through June 24 2010. All SO2 fluxes > 350 tons/day correspond to positive values of GPS baseline changes. (D) GPS baseline changes versus SO2 fluxes for the 2011 period. SO2 fluxes decreased by 52% with regard to the 2010 period. (E) SO2 versus RSAM daily average. The 2010 data go from April 28 through June 24. SO2 fluxes > 300 tons/day are associated with RSAM values > 70 units. (F) SO2 versus RSAM daily average for the 2011 period. There is not any clear correspondence between the 2011 SO2 and RSAM data.

119 ABOUT THE AUTHOR

Jose´ Armando Saballos was born in Belen,´ Rivas, Nicaragua. He received his Bachelor in

Physics, with a minor in Geophysics from Universidad Nacional Autonoma´ de Nicaragua, Man- agua, in 2002. Later that year he got a scholarship from the Department of International Develop- ment of the UK to undertake a Master’s in Environmental Monitoring, Modeling and Management of the Environmental change at King’s College London, University of London, and graduated in

2003. Later in 2007 he joined the Department of Geology at the University of South Florida

(USF) to began a PhD under the supervision of Professor Charles B. Connor and Professor Rocco

Malservisi on the application of Geophysics techniques to study short– and long–term behavior of

Concepcion´ volcano in Nicaragua, while working as a teaching assistance in the Geology Depart- ment at USF. He participated on several ﬁeld trips and conferences in Nicaragua, Costa Rica and the USA. Mr. Saballos has published several papers in international journals and was awarded two mini–grants from the USF Institute for the Study of Latin America and the Caribbean (2009 and

2010) two support part of his ﬁeld trips.