Crustal Seismicity in the Back-Arc Region of the Southern Central Andes from Historic to Modern Times

Item Type text; Electronic Dissertation

Authors Alvarado, Patricia Monica

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 04/10/2021 04:27:26

Link to Item http://hdl.handle.net/10150/195537 CRUSTAL SEISMICITY IN THE BACK-ARC REGION OF THE SOUTHERN CENTRAL ANDES FROM HISTORIC TO MODERN TIMES

Patricia Mónica Alvarado

A Dissertation Submitted to the Faculty of the DEPARTMENT OF GEOSCIENCES In Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY In the Graduate College THE UNIVERSITY OF ARIZONA

2006

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Dissertation Committee, we certify that we have read the dissertation

prepared by PATRICIA MONICA ALVARADO

entitled CRUSTAL SEISMICITY IN THE BACK-ARC REGION OF THE SOUTHERN CENTRAL ANDES FROM HISTORIC TO MODERN TIMES and recommend that it be accepted as fulfilling the dissertation requirement for the

Degree of Doctor of Philosophy

______Date: April 14, 2006 Dr. Susan Beck

______Date: April 14, 2006 Dr. George Zandt

______Date: April 14, 2006 Dr. Clement Chase

______Date: April 14, 2006 Dr. Roy Johnson

______Date: April 14, 2006 Dr. Richard Bennett

Final approval and acceptance of this dissertation is contingent upon the candidate’s submission of the final copies of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

______Date: Date: April 14, 2006 Dissertation Director: Dr. Susan Beck

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the copyright holder.

SIGNED: Patricia M. Alvarado

ACKNOWLEDGEMENTS

I would like to thank first and foremost my advisor Susan Beck (head of the Geosciences Department) for her support, critical advice, assistance and encouragement to work in the CHARGE Andean project. I would also like to thank George Zandt for his openness for questions and guidance in my scientific project and other professors in my committee, Clem Chase, Roy Johnson and Rick Bennett, for showing interest in my research.

Several people offered me help during my studies not only limited to my scientific work. I express my gratitude to my friend Robert Fromm, who unexpectedly passed away in July 2004, for his generosity, open mind and great moments. I also thank Megan Anderson, Arda Ozacar, Hersh Gilbert, Berk Biryol, Keith Koper, Carlos Cintra, Stephen Sorenson, Facundo Fuentes, Lara Wagner, the GSAT (Susan and George’s) group, Randy Richardson, Anne Chase, Robert Butler and Terry Wallace.

I acknowledge several Institutions and Awards that made possible this research and the CHARGE project: National Science Foundation, PASSCAL-IRIS, University of Arizona, P.E.O., ChevronTexaco, Galileo Circle, PERISHIP-2004 (U.S.A.); Universidad Nacional de San Juan, Instituto Nacional de Prevención Sísmica, Fundación Antorchas (Argentina); Universidad de Chile (Chile); O.A.S.

My special thanks to my husband and daughter for their unconditional support, companion, and above all, for their love. Several friends have contributed to make our staying in Tucson an unforgettable experience. They will be in our memories forever. I apologize for not mentioning everyone that we were fortunate enough to meet. I will cover some of them: Pamela and Wayne Kappel, for making us feel part of their family during this long time studying abroad. Verónica and Fernando Barra and their kids, for their kindness and support, sometimes during difficult times. Eneida Lima, for her friendship, and the contact to meet the soccer group, which made my whole family to enjoy every Sunday morning at the Mitchell park. Lionel Meza and D’Londa Morrison for giving us the privilege of having their kids playing with our daughter every single week since the first day we came to the U.S.A. Pablo Guerenstein and his family for their help, especially at the beginning.

Finally, to my family and friends in Argentina and Chile. The long hours spent in e-mails and phone-calls were a tremendous support in the distance. I express particular thanks to Tony Monfret, Brígida and Eduardo Machuca, Alberto Sáez, Mario Pardo, Sergio Barrientos, Diana Comte, Mariana Aguiar, Marcelo Millán, Adriana Valenzuela, Ricardo Uliarte and Jorge Sisterna.

DEDICATION

To my husband Mauro Sáez, whose patience, optimism and conviction made possible to leave our home in Argentina for coming to the U.S.A. to know more about earth sciences.

To our daughter, Antonella, for showing to her that there is no end to a woman’s growth and ... for her future.

To the memory of Robert Fromm, whom I dreamed of sharing a bright future in South America ...

TABLE OF CONTENTS

ABSTRACT...... 9

INTRODUCTION...... 11

I. Historic Seismicity...... 13

II. Modern Crustal Seismicity...... 14

III. Crustal Structure...... 15

PRESENT STUDY...... 16

REFERENCES...... 18

APPENDIX A: Permission for Reproduction from Elsevier ...... 24

Source Characterization of the San Juan (Argentina) Crustal Earthquakes of 15 January

1944 (Mw 7.0) and 11 June 1952 (Mw 6.8). (Reprint)...... 26

Abstract...... 27

1. Introduction and Regional Tectonics...... 27

2. Data and Methods...... 31

3. Results and Discussion...... 33

3.1. The 15 January 1944 Earthquake...... 33

3.2. The 11 June 1952 Earthquake...... 38

4. Conclusions...... 40

Acknowledgements...... 41

References...... 41

APPENDIX B: Permission for Reproduction from Blackwell...... 44 7

TABLE OF CONTENTS - Continued

Crustal Deformation in the South-Central Andes Backarc Terranes as Viewed from

Regional Broad-Band Seismic Waveform Modelling. (Reprint)...... 47

Summary...... 48

1 Introduction...... 48

2 Regional Tectonic Setting...... 49

3 Data and Regional Waveform Modelling...... 51

4 Sensitivity to Crustal Models...... 52

5 Results and Discussion...... 56

5.1 Western Sierras Pampeanas-Sierra Pie de Palo and Valle Fértil Megashear...58

5.2 Eastern Sierras Pampeanas...... 61

5.3 Mendoza...... 62

5.4 High Cordillera...... 62

6 Conclusions...... 63

Acknowledgements ...... 63

References ...... 64

APPENDIX C: Crustal Structure of the South-Central Andes Cordillera and Backarc

Region from Regional Waveform Modeling. (Manuscript)...... 67

Abstract...... 68

1. Introduction, Tectonic Setting and Terrane Overview...... 70

2. Previous Regional Seismic Studies...... 72

3. Methods...... 73 8

TABLE OF CONTENTS - Continued

3.1. Regional Seismic Moment Tensor Inversion Data...... 74

3.2. Regional Forward Broadband Modeling...... 75

4. Regional Waveform Sensitivity Tests...... 76

4.1. Sensitivity to Crustal Thickness...... 77

4.2. Sensitivity to Average Crustal P-Wave Velocity...... 77

4.3. Sensitivity to Crustal P- to S-Wave Velocity Ratio...... 78

4.4. Sensitivity to Mantle Parameters...... 78

5. Grid-Search Forward Modeling to Determine Crustal Structure...... 79

6. Results and Discussion...... 84

6.1. High Cordillera and Active Arc...... 86

6.2. Eastern Sierras Pampeanas...... 88

6.3. Precordillera and Western Sierras Pampeanas...... 88

7. Comparison of Crustal Structure with Regional Crustal Seismicity...... 91

7.1. Observations...... 91

7.2. Discussion ...... 94

8. Conclusions...... 97

Acknowledgements...... 99

References...... 101

Table and Figure Captions ...... 112

Table 1...... 119

Figures...... 120 9

ABSTRACT

The western margin of South America between 30ºS and 36ºS is seismically active. While the largest magnitude earthquakes are the interplate subduction zone events, the historically most devastating earthquakes have been the moderate-to-large magnitude earthquakes with depths < 35 km in the Andean back-arc. This region is characterized by accreted terranes later reactivated during Mesozoic extensional processes. Crustal seismicity in the back-arc is related to the thin-skinned Precordillera

(PC) fold-thrust belt and the thick-skinned Sierras Pampeanas (SP) basement-cored uplifts overlying the flat slab segment. South of 33ºS, the active volcanic arc above the normally dipping subducted plate is also seismically active at crustal depths. In this study we combined historical and regional broadband seismic data to characterize moderate-to- large earthquakes and the crustal structure in this region. We have digitized and modeled teleseismic records of the 1944 and 1952 San Juan, Argentina PC earthquakes. Both events have shallow source depths, short duration of the source time functions with a thrusting focal solution for the 1944 (Mw 7.0) earthquake and a major strike-slip component in the 1952 (Mw 6.8) earthquake solution. By modeling regional broadband waveforms collected during the CHile-ARgentina Geophysical Experiment (CHARGE) during 2000 and 2002 we constrained the seismic moment tensor and improved focal depths for 27 crustal (3.5 < Mw < 5.1) earthquakes. We found predominantly thrust-fault focal mechanisms and focal depths of 10-26 km for earthquakes over the flat slab region; the eastern SP and active arc have earthquakes with strike-slip focal mechanisms and shallower depths. We used these same earthquakes to determine the crustal structure 10

using raypaths that sample different geologic terranes. Our results indicate high Vp, low

Vs for the northern Cordillera, PC and western SP thicker crust; low Vp, low Vs and a thinner crust beneath the arc (south of 33°S) consistent with a mafic composition and partial melt. The eastern SP basement shows low Vp, low Vs and thinner crust consistent with a more quartz-rich composition. These differences have an important control on the present day Andean earthquake deformation and the high seismic hazard posed in this region.

INTRODUCTION

The convergence between the oceanic Nazca plate and the continental South

American plate at a rate of 6.3 cm/yr and an azimuth of 79.5º (Kendrick et al., 2003), generates earthquakes at both slab (~100 km) and crustal (< 35 km) depths between 30ºS and 36ºS (Barazangi and Isacks, 1976; Cahill and Isacks, 1992; Smalley et al., 1993;

Pardo et al., 2002; Salazar, 2005; Alvarado et al., 2005; Anderson et al., 2006). However the most damaging earthquakes are the moderate-to-large magnitude events in the

Andean back-arc crust related to the Precordillera and Sierras Pampeanas over the flat slab subduction segment at about 31ºS (INPRES, 2006; Kadinsky-Cade, 1985; Langer and Bollinger, 1988; Langer and Hartzell, 1996; Alvarado and Beck, 2006). Crustal seismicity in the backarc since 1900 is responsible for three times larger seismic energy release than the shallow seismicity in the active volcanic arc (Kay et al., 1991) and the back-arc region (south of 33ºS) above the normal dipping subducting slab (Anderson et al., 2006; Gutscher et al., 2000).

The western margin of south central South America has a long accretionary history with two main tectonic-magmatic events: the Pampean orogeny at ~530-520 Ma

(Rapela et al., 1998), and the Famatinian orogeny at ~490-460 Ma (Ramos, 2004; Rapela and Pankhurst, 2001). Some authors (e.g. Ramos 2004) relate the collision of the composite Cuyania terrane with the second event. The Cuyania basement is composed of two terranes with Grenville ages, the Precordillera and the westernmost Sierras

Pampeanas, which do not show evidence of Cambrian-Ordovician magmatism between them related to active subduction (e.g. Ramos, 2004). More recently the allochthonous 12

Chilenia terrane collided on the western margin at ~400 Ma (Davis et al., 1999) (Fig. 1 of

Appendix C). The origin and evolution of these terranes, however, are still debated (Dalla

Salda et al., 1992; Aceñolaza et al., 2002; Thomas and Astini, 2003; Finney et al., 2003,

2004; Galindo et al., 2004; Ramos, 2004). In addition, Late Paleozoic and Mesozoic rifting events are well-documented in the region (Franzese and Spalletti, 2001; Franzese et al., 2003). The Andean compression has been taken up along pre-existing weak zones causing, in many cases, the inversion of ancient extensional faults associated with the uplift of the Sierras Pampeanas (Ramos et al., 2002).

The western Sierras Pampeanas have a mafic-ultramafic composition with abundance of crystalline calcites, amphiboles, and scarce granitic bodies (Caminos,

1979). In contrast, the eastern Sierras Pampeanas exhibit schists and gneisses of a mainly sedimentary origin, and granitic rocks, some of them reaching batholithic dimensions with abundant migmatites and granulites (Rapela et al., 1998).

This dissertation consists of three parts all related to crustal earthquakes located in arc and back-arc region of the Andes between 30°S and 36°S. The main goal is to characterize this seismicity and the crustal structure of the Andean arc and back-arc region in order to evaluate what controls the Paleozoic accreted terranes and their sutures have on the present Andean deformation. This analysis is important for the assessment of the region’s seismic hazards.

I. Historic Seismicity

The Argentinean historical record indicates that the 1861 and 1944 crustal earthquakes over the flat slab subduction zone were the most devastating earthquakes in

Argentina (INPRES, 2006) (Fig. 2 of Appendix B). The 1861 earthquake destroyed the city of Mendoza, killing one third of the city's population at that time (RJEHM, 1938).

Damage was caused not only by the strong ground motion but also the overflow of a ditch that crosses the city in a north-south direction and fires that raged immediately after the earthquake (Alexander, 1998). Our understanding of this event is only based on seismic intensities and historic reports since no seismograms are available for this devastating earthquake. In contrast, there are a large number of seismic paper-records for the 1944 and 1952 earthquakes that can be analyzed to extract useful quantitative information. The 1944 event devastated 80% of downtown San Juan and caused ~5,000 deaths in a population of ~80,000 people. For this reason the 1944 earthquake is considered the largest natural disaster in the Argentinean history. Its study is important to determine the type of seismic source, duration, size, depth, and the possible relationship to active faults in the region. The most recent large earthquake in the vicinity of San Juan killed 65 people in 1977. This Mw 7.5 event is the only large crustal earthquake studied in any detail using local, regional and teleseismic records (INPRES, 2006). Its source was located in the Western Sierras Pampeanas and was composed of two shocks associated with blind faults, separated by 64 km and 20 s (Kadinsky-Cade, 1985; Langer and

Bollinger, 1988; Langer and Hartzell, 1996). In this research the 1944 and 1952 earthquakes are investigated using seismic paper-records and historical reports. 14

II. Modern Crustal Seismicity

The largest events that have occurred in the region since 1963 have been studied using the global seismic network. Earthquakes of magnitudes over 5.0 are well recorded at teleseismic distances, but small-to-moderate earthquakes, barely recorded at these distances, are more abundant, and can provide important details to understanding the active tectonics of a region. Permanent short-period seismic networks of Argentina and

Chile, recording over the last 20 years, have improved detection, location, and some first- motion focal mechanism determinations. In the last ten years regional broadband seismic experiments have revealed some extraordinary details of seismic sources for small-to- moderate earthquakes around the world that are not well recorded by the global networks.

In this study we analyzed seismic broadband data recorded by the seismic network of the CHile-ARgentina Geophysical Experiment (CHARGE). This array deployed 22

PASSCAL broadband instruments between 30ºS and 36ºS from the coast of Chile to central Argentina during 18 months of 2000 and 2002 (Fig. 1 of Appendix B). We constrained the seismic moment tensor and source depths for 27 crustal earthquakes in the Andean backarc and the high Cordillera. The results of this study are compared with other CHARGE studies (Fromm et al., 2004; Gilbert et al., 2006), previous seismic determinations from the PANDA experiment (Regnier et al., 1992; Smalley et al., 1993) and GPS results (Brooks et al., 2003).

III. Crustal Structure

The Andean region between 30ºS and 36ºS shows important along strike variations in the geometry of the subducted Nazca plate (Barazangi and Isacks, 1976;

Cahill and Isacks, 1992; Anderson et al., 2006), the style of deformation of the upper plate (Barazangi and Isacks, 1976; Jordan and Allmendinger, 1986) and the generation of crustal earthquakes. One of the possible factors controlling nucleation of crustal earthquakes in the continental plate could be the contrasting terrane inheritance, associated with differences in crustal composition, or the accretion, rifting and re- accretion processes since Paleozoic (Ramos et al., 2002).

By using digital broadband regional waveforms we performed a detailed analysis of the crustal structure between single earthquake-CHARGE station paths. We develop a regional seismic velocity model between 30ºS and 36ºS. This model maps out differences in the geologic terranes and correlates with the uneven generation of crustal earthquakes.

PRESENT STUDY

This study of the crustal seismicity between 30ºS and 36ºS is developed in three appendix of this dissertation. They consist of two articles already published and one manuscript for submission to Geophysical Journal International. A brief description of the results is given below.

Appendix A includes the study of the 1944 and 1952 San Juan, Argentina earthquakes modeling teleseismic data and using first-motion local and regional data.

Both events are related to the interaction between the Precordillera and the western

Sierras Pampeanas. Our best results indicate that the 15 January 1944 earthquake has a thrust fault focal mechanism whereas the 11 June 1952 event exhibits a major strike-slip component in its solution. Both events show focal depths < 12 km and simple source time functions with one pulse of moment release of duration of 10 s and 8 s, respectively. The results for the 1944 earthquake are consistent with the parameters observed along a coseismic surface rupture of 6-8 km long in the frontal part of the eastern Precordillera.

Appendix B shows the analysis of the crustal seismicity of the Andes cordillera and its back-arc region using regional broadband CHARGE waveforms. The seismic moment tensor and focal depth are obtained for 27 (3.5

In the eastern Sierras Pampeanas the earthquakes have both reverse and strike-slip fault focal mechanisms and shallower depths between 5 and 18 km. Better matches between 17

synthetic and observed waveforms, for seismic raypaths traveling through the crust of the eastern Sierras Pampeanas, result using a low P-wave velocity (Vp=6.0-6.2 km/s), a high

S-wave velocity (Vp/Vs<1.70) and a thin crustal layer (th=27-35 km). In the western

Sierras Pampeanas we constrained a crust of Vp=6.2-6.4 km/s, Vp/Vs>1.80 and th=45-55 km.

Appendix C presents the results after investigating the seismic velocity structure for single station-earthquake paths. The behavior of regional broadband waveforms in the

Andes cordillera and its back-arc region is explored by making synthetic seismograms for crustal earthquakes in Appendix B, and comparing them to the broadband CHARGE waveforms. Our results indicate that the seismic waveforms are much more sensitive to the crustal parameters (P-wave velocity, S-wave velocity and thickness) than to the mantle parameters. The best crustal models indicate Vp of 6.4-6.6 km/s, Vp/Vs of 1.80-

1.85 and thickness of 50-60 km in the high Cordillera north of 33°S. The active arc is best represented by a crust of Vp of 5.8-6.0 km/s, Vp/Vs ~1.80 and thickness of 40-45 km. The Cuyania terrane exhibits a more complex crustal structure with average Vp of

6.4 km/s, Vp/Vs of 1.80-1.85 and total thickness of ~50 km. The Pampia terrane has a thinner single-layer crust of 30-40 km with Vp of 5.8-6.0 km/s and Vp/Vs < 1.70. These differences in crustal P- and S-wave velocities and thickness are consistent with the contrast in the compositional properties of the exposed terranes and partial melt beneath the active volcanic arc. Moreover, these properties reflect differences in the weakness of the crust of the Sierras Pampeanas basement that generate different levels of crustal seismicity in the region. 18

REFERENCES

Aceñolaza, F.G., Miller, H. and Toselli, A.J., 2002, Proterozoic-Early Paleozoic

evolution in western South America, a discussion, Tectonophysics 354, 121-137.

Alexander, A.J., 1998, La fotografía en Mendoza: desde sus orígenes al terremoto de

1861. In: Las ruinas de San Francisco, Arqueología e Historia, Mendoza, Argentina,

295-313.

Alvarado, P., Beck, S., Zandt, G., Araujo, M. and Triep, E., 2005, Crustal deformation in

the south-central Andes back-arc terranes as viewed from regional broadband

seismic waveform modelling, Geophysical Journal International 163, 580–598.

Alvarado, P. and Beck, S., 2006, Source characterization of the San Juan (Argentina)

crustal earthquakes of 15 January 1944 (Mw 7.0) and 11 June 1952 (Mw 6.8), Earth

and Planetary Science Letters 243(3-4), 615-631.

Anderson, M., Alvarado, P., Zandt, G. and Beck, S., 2006, Geometry and brittle

deformation of the subducting Nazca plate, Central Chile and Argentina, submitted

for publication to Geophysical Journal International.

Barazangi, M. and Isacks, B., 1976, Spatial distribution of earthquakes and subduction of

the Nazca plate beneath South America, Geology 4, 686–692.

Brooks, B.A., Bevis, M., Smalley, R.(Jr.), Kendrick, E., Manceda, R., Lauría, E.,

Maturana, R. and Araujo, M., 2003, Crustal motion in the southern Andes (26°–

36°S): do the Andes behave like a microplate?, Geochemestry Geophysics

Geosystems 4, 1–14. 19

Cahill, T. and Isacks, B.L., 1992, Seismicity and shape of the subducted Nazca plate,

Journal of Geophysical Research 97, 17 503–17 529.

Caminos, R., 1979, Sierras Pampeanas Noroccidentales, Salta, Tucumán, Catamarca, La

Rioja y San Juan. II Simposio de Geología Regional Argentina, Academia Nacional

de Ciencias, 1: 225-292, Córdoba, Argentina.

Dalla Salda, L.H., Dalziel, I.W.D., Cingolani, C., and Varela, R., 1992, Did the Taconic

Appalachians continue into southern South America?: Geology 20, 1059-1062.

Davis, J.S., Roeske, S.M. and McClelland, W.C., 1999, Closing the ocean between the

Precordillera terrane and Chilenia: Early Devonian ophiolite emplacement and

deformation in the SW Precordillera, In: Ramos, V.A. and Keppie, D. (Eds.)

Laurentia-Gondwana connections before Pangea Geological Society American

Special Paper 336, 115-138, U.S.A.

Finney, S., Gleason, J., Gehrels, G., Peralta, S. and Aceñolaza, G., 2003, Early

Gondwanan connection for the Argentine Precordillera terrane, Earth and Planetary

Science Letters 205(3-4), 349-359.

Finney, S., Gleason, J., Gehrels, G., Peralta, S. and Aceñolaza, G., 2004, Corrigendum to

"Early Gondwanan connection for the Argentine Precordillera terrane": [Earth

Planet. Sci. Lett. 205 (2004) 349–359], Earth and Planetary Science Letters, 219(3-

4), pp. 413.

Franzese, J.R. and Spalletti, L.A., 2001, Late Triassic–early Jurassic continental

extension in southwestern Gondwana: tectonic segmentation and pre-break-up

rifting, Journal of South American Earth Sciences 14(3), 257-270. 20

Franzese, J., Spalletti, L. Gómez-Pérez, I. and Macdonald, D., 2003, Tectonic and

paleoenvironmental evolution of Mesozoic sedimentary basins along the Andean

foothills of Argentina (32°–54°S), Journal of South American Earth Sciences 2003

16(1), 81-90.

Fromm, R., Zandt, G. and Beck, S.L., 2004, Crustal thickness beneath the Andes and

Sierras Pampeanas at 30°S inferred from Pn apparent phase velocities, Geophysical

Research Letters 31, doi: L006625, doi:10.1029/2003GL019231.

Gilbert, H., Beck, S. and Zandt, G., 2006, Lithospheric and upper mantle structure of

central Chile and Argentina, Geophysical Journal International 165(1), 383-398.

Gutscher, M.A., Spakman, W., Bijwaard, H. and Engdahl, R., 2000, Geodynamics of flat

subduction: Seismicity and tomographic constraints from the Andean margin,

Tectonics 19, 814-833.

INPRES, 2006, Listado de sismos históricos de Argentina (http://www.inpres.gov.ar),

online catalogue.

Kadinsky-Cade, K., 1985, Seismotectonic of the Chilean margin and the 1977 Caucete

earthquake of western Argentina, PhD thesis, Cornell University, Ithaca, New York,

U.S.A., p. 253 (in English).

Kay, S., Mpodozis, C., Ramos, V. and Munizaga, F., 1991, Magma source variations for

mid-late Tertiary magmatic rocks associated with a shallowing subduction zone and

a thickening crust in the central Andes (28° to 33°S), in Andean magmatism and its

tectonic setting, Geol. Soc. Am. Spec. Pap. 265, 113-137, eds. Harmon, R.S. &

Rapela, C.W., Boulder, CO, USA. 21

Kendrick, E., Bevis, M., Smalley, R.(Jr), Brooks, B.A., Barriga, R., Lauría, E., Souto,

L.P., 2003, The Nazca-South America Euler vector and its rate of change, Journal of

South American Earth Sciences 16, 125-131.

Langer, C.J. and Bollinger, G.A., 1988, Aftershocks of the western Argentina (Caucete)

earthquake of 23 November 1977: some tectonic implications, Tectonophysics 148,

131–146.

Langer, C.J. and Hartzell, S., 1996, Rupture distribution of the 1977 western Argentina

earthquake, Physics of the Earth and Planetary Interiors 94, 121–132.

Pardo, M., Comte, D. and Monfret, T., 2002, Seismotectonic and stress distribution in the

central Chile subduction zone, Journal of South American Earth Sciences 15, 11-22.

Ramos, V.A., 1994, Terranes of southern Gondwanaland and their control in the Andean

structure (30°-33°S latitude). In: Reutter, K.J., Scheuber, E. and Wigger, P.J. (Eds.):

Tectonics of the Southern Central Andes: 249-261, Springer-Verlag, New York,

U.S.A.

Ramos, V.A., 2004, Cuyania, an exotic block to Gondwana: review of a historical

success and the present problems, Gondwana Research 7(4), 1009-1026.

Ramos, V.A., Cristallini, E.O. and Pérez, D.J., 2002, The Pampean flat-slab of the central

Andes, Journal of South American Earth Sciences 15, 59-78.

Rapela, C.W., Pankhurst, R.J., Casquet, C., Baldo, E., Saavedra, J., Galindo, C. and

Fanning, C.M., 1998, The Pampean orogeny of the southern proto-Andes: Cambrian

continental collision in the Sierras de Córdoba, in: The proto-Andean margin of 22

Gondwana, Pankhurst, R.J. and Rapela, C.W. (Eds.), Geological Society, London,

Special Publications 142, pp. 181-217.

Rapela, C.W. and Pankhurst, R., 2001, The final assembly of southwestern Gondwana: a

proto-Pacific perspective. XI Congreso Latinoamericano de Geología and III

Congreso Uruguayo de Geología, Abstracts, 031. Montevideo, Uruguay. CD edition.

Regnier, M., Chatelain, J.L., Smalley, R.(Jr.), Chiu, J.M., Isacks, B. and Araujo, M.,

1992, Seismotectonics of Sierra Pie de Palo, a basement block upflit in the Andean

foreland of Argentina, Bulletin of the Seismological Society of America 82, 2549-

2571.

Regnier, M., Chatelain, J.L., Smalley, R.(Jr.), Chiu, J.M., Isacks, B. and Araujo, M.,

1994, Crustal thickness variation in the Andean foreland, Argentina from converted

waves, Bulletin of the Seismological Society of America 84, 1097-1111.

RJEHM (Revista de la Junta de Estudios Históricos de Mendoza), El terremoto del 20 de

marzo de 1861. Volumen X, Mendoza, Argentina, 1938.

Salazar, P., 2005, Análisis del campo de esfuerzos en la zona de subducción bajo Chile

Central (30º-34ºS). M.Sc. thesis, University of Chile, Santiago, Chile, p. 196 (in

Spanish).

Smalley, R.(Jr.), Pujol, J., Regnier, M., Chiu, J.M., Chatelain, J.L., Isacks, B.L., Araujo,

M. and Puebla, N., 1993, Basement seismicity beneath the Andean Precordillera

thin-skinned thrust belt and implications for crustal and lithospheric behavior,

Tectonics 12, 63–76. 23

Thomas, W.A. and Astini, R., 2003, Ordovician accretion of the Argentine Precordillera

terrane to Gondwana: a review, Journal of South American Earth Sciences 16, 67-

79.

APPENDIX A: Permission for Reproduction from Elsevier

The following is a copy of an e-mail conversation in order to get permission from Elsevier for publication of the Earth and Planetary Science Letters article "Source characterization of the San Juan (Argentina) crustal earthquakes of 15 January 1944 (Mw 7.0) and 11 June 1952 (Mw 6.8)” by Alvarado P. and Beck S., (2006) 243, 615-631. Doi:10.1016/j.epsl.2006.01.015

Dear Ms Alvarado,

We hereby grant you permission to reproduce the material detailed below in your thesis at no charge subject to the following conditions:

1. If any part of the material to be used (for example, figures) has appeared in our publication with credit or acknowledgement to another source, permission must also be sought from that source. If such permission is not obtained then that material may not be included in your publication/copies.

2. Suitable acknowledgment to the source must be made, either as a footnote or in a reference list at the end of your publication, as follows: "Reprinted from Publication title, Vol number, Author(s), Title of article, Pages No., Copyright (Year), with permission from Elsevier".

3. Reproduction of this material is confined to the purpose for which permission is hereby given.

4. This permission is granted for non-exclusive world English rights only. For other languages please reapply separately for each one required. Permission excludes use in an electronic form. Should you have a specific electronic project in mind please reapply for permission.

5. This includes permission for UMI to supply single copies, on demand, of the complete thesis. Should your thesis be published commercially, please reapply for permission.

Yours sincerely,

Natalie David Rights Assistant

Your future requests will be handled more quickly if you complete the online form at www.elsevier.com/permissions -----Original Message-----

Dear Elvira Nieuwerf:

I am writing you to obtain permission from Elsevier for including a copy of an entire journal EPSL article in my Ph.D. dissertation at the University of Arizona. The correct reference of the paper is "Source characterization of the San Juan (Argentina) crustal earthquakes of 15 January 1944 (Mw 7.0) and 11 June 1952 (Mw 6.8)", Alvarado P. and Beck S., Earth and Planetary Science Letters, 2006 Vol. 243, 615 - 631. Doi:10.1016/j.epsl.2006.01.015.

The University of Arizona requires that this permission extends to microfilming and publication by University Microfilms Incorporated (UMI) and that the copyright owners (Elsevier) are aware that UMI may sell, on demand, single copies of the thesis, dissertation or document, including the copyrighted materials, for scholarly purposes.

I would greatly appreciate your help to obtain this permission.

Thank you for your time,

Best regards,

Patricia Alvarado

Source Characterization of the San Juan (Argentina) Crustal Earthquakes of 15 January 1944 (Mw 7.0) and 11 June 1952 (Mw 6.8), Alvarado P. and Beck S., (2006) 243, 615-631. Doi:10.1016/j.epsl.2006.01.015

Reprint from Earth and Planetary Science Letters, with permission from Elsevier.

The original version of the paper Earth and Planetary Science Letters is available from the Elsevier online delivery service, accessible via the journal's website at http:// www.elsevier.com

APPENDIX B: Permission for Reproduction from Blackwell

The following is a copy of an e-mail conversation in order to get permission for publication of the paper "Crustal deformation in the south-central Andes back-arc terranes as viewed from regional broadband seismic waveform modelling" by Alvarado et al., 2005. Doi:10.111/j.1365-246X.200502759.x

Dear Patricia Alvarado,

Thank you for your e-mail request. In publishing with Blackwell, you retain the following rights to re-use the article as long as you do not sell or reproduce the Article or any part of it for commercial purposes (i.e. for monetary gain on your own account or on that of a third party, or for indirect financial gain by a commercial entity). These rights apply without needing to seek permission from Blackwell Publishing.

Prior to acceptance: We ask that as part of the publishing process you acknowledge that the Article has been submitted to the Journal. You will not prejudice acceptance if you use the unpublished Article, in form and content as submitted for publication in the Journal, in the following ways: you may share print or electronic copies of the Article with colleagues; you may post an electronic version of the Article on your own personal website, on your employer's website/repository and on free public servers in your subject area.

After acceptance: Provided that you give appropriate acknowledgement to the Journal, the society if relevant and Blackwell Publishing, and full bibliographic reference for the Article when it is published, you may use the accepted version of the Article in the format as originally submitted for publication in the Journal, and updated to include any amendments made after peer review, in the following ways: you may share print or electronic copies of the Article with colleagues; you may use all or part of the Article and abstract, without revision or modification, in personal compilations or other publications of your own work; you may use the Article within your employer's institution or company for educational or research purposes, including use in course packs; you may post an electronic version of the Article on your own personal website, on your employer's website/repository and on free public servers in your subject area. (For some journals there is a requirement that posting of the Article online does not take place until a specified minimum period has elapsed.) Electronic versions of the accepted Article must include a link to the published version of the Article together with the following text: ‘The definitive version is available at www.blackwell-synergy.com'.

Please note that you are not permitted to post the Blackwell Publishing PDF version of the Article online.

All requests by third parties to re-use the Article in whole or in part will be handled by Blackwell Publishing. Any permission fees will be retained by the Journal. All requests to adapt substantial parts of the Article in another publication (including publication by Blackwell Publishing) will be subject to your approval (which is deemed to be given if we have not heard from you within 4 weeks of your approval being sought by us writing to you at your last notified address). Please address any queries to [email protected]

Kind regards,

Laura Wilson. Permissions Dept. Blackwell Publishing

Permission requests can now be sent to [email protected]

Blackwell is committed to creating a culture of value and respect for all of our staff. We expect to work in an environment where there are high standards of behaviour and achievement. We maintain a culture which operates within accepted boundaries of professional behaviour and performance. -----Original Message-----

Dear Zoë Ellams:

Let me thank your email response to my message asking for permission for using Blackwell material for publication.

As I said, long time ago I obtained permission from Blackwell to include a copy of an entire journal GJI article in my PhD dissertation at the University of Arizona. The reference of the article is "Crustal deformation in the south-central Andes back-arc terranes as viewed from regional broadband seismic waveform modelling", Alvarado P., Beck S., Zandt G., Araujo M. & Triep E., Geophysical Journal International, 2005, volume 163, pages 580-598, doi:10.111/j.1365-246X.200502759.x However, the University of Arizona requires that this permission should extend to microfilming and publication by University Microfilms Incorporated (UMI) and that the copyright owners (Blackwell) are aware that UMI may sell, on demand, single copies of the thesis, dissertation or document, including the copyrighted materials, for scholarly purposes. Could you confirm if this specific permission is also granted? If not, how might I go in order to get it? Thank you for your time, Best regards,

Patricia Alvarado 46

-----Original Message-----

Dear Dr Alvarado,

Thank you for your email request. Permission is granted for you to use the material below for your thesis subject to the usual acknowledgements and on the understanding that you will reapply for permission if you wish to distribute or publish your thesis commercially.

Good luck!

Best Wishes Zoë

Zoë Ellams (Miss) Permissions Co-ordinator Blackwell Publishing

All future permission requests should be sent to mailto: [email protected]

Dear Blackwell Production Editor:

I would like to ask permission to use a copy of the color paper version entitled "Crustal deformation in the south-central Andes back-arc terranes as viewed from regional broadband seismic waveform modelling", Alvarado P., Beck S., Zandt G., Araujo M. & Triep E., Geophysical Journal International, 2005, volume 163, pages 580-598, as an appendix of my Ph.D. Thesis Dissertation. This paper, whose I am the first author, reflects one chapter of my research as a graduate student at the University of Arizona, U.S.A. I will make the proper statement that the original version of the Geophysical Journal International paper is available from the Blackwell Synergy online delivery service, accessible via the journal's website at http:// www.blackwellpublishing.com/journals/GJI or http://www.blackwell-synergy.com. I will appreciate your understanding,

Sincerely,

Patricia Alvarado

Crustal Deformation in the South-Central Andes Backarc Terranes as Viewed from Regional Broad-Band Seismic Waveform Modelling, Alvarado P., Beck S., Zandt G., Araujo M. and Triep E., (2005) 163, 580-598. Doi:10.111/j.1365-246X.200502759.x

Reprint from Geophysical Journal International.

The original version of the Geophysical Journal International paper is available from the Blackwell Synergy online delivery service, accessible via the journal's website at http:// www.blackwellpublishing.com/journals/GJI or http://www.blackwell-synergy.com

APPENDIX C: Crustal Structure of the South-Central Andes Cordillera and

Backarc Region from Regional Waveform Modeling, Alvarado P., Beck, S. and

Zandt G. (2006).

Manuscript for submission to Geophysical Journal International. 68

Crustal Structure of the South-Central Andes Cordillera and Backarc Region from

Regional Waveform Modeling

Alvarado, P.1,2, Beck, S.1 and Zandt, G.1

1 Department of Geosciences, University of Arizona, 1040 E 4th St., Tucson, AZ 85721,

U.S.A.

2 Departmento de Geofísica y Astronomía, Facultad de Ciencias Exactas, Físicas y

Naturales, Universidad Nacional de San Juan, Meglioli 1160 S, San Juan, 5400,

Argentina.

Abstract

We investigate the crustal structure in the Andes cordillera and its backarc region using regional broadband waveforms from crustal earthquakes. We consider seismic waveforms recorded at regional distances by the CHile-ARgentina Geophysical

Experiment (CHARGE) during 2000 to 2002 and utilize previous seismic moment tensor inversion results. For each single station-earthquake pair, we fixed the source parameters and performed forward waveform modeling using raypaths that sample the crust of the highest elevation cordillera and the accreted terranes in the backarc region. We performed a sensitivity test that indicates synthetic seismograms are most sensitive to crustal parameters and less sensitive to mantle parameters. Thus we carried out a grid 69

search around crustal thickness, P-wave velocity (Vp) and P- to S-wave velocity ratio

(Vp/Vs), fixing mantle parameters. We evaluate this waveform analysis by estimating an average correlation coefficient between observed and synthetic data for the three broadband components. We identify all acceptable crustal models that correspond to high correlation coefficients that provide best overall seismogram fits for the data and synthetic waveforms filtered between 10 and 80 s. Our results indicate along strike variations in the crustal structure for the north-south high cordillera with higher P-wave velocity and thickness in the northern segment (north of 33°S), and persistently high

Vp/Vs ratio (> 1.85) in both segments. This is consistent with a colder mafic composition for the northern segment and a region of crustal thickening and shortening above the flat slab region. In contrast, the currently volcanic arc results agree with a warmer region consistent with active volcanism presumably of an intermediate- to mafic composition. A distinctive feature in the backarc region is the marked contrast between the seismic properties of the Cuyania and Pampia terranes. The Cuyania terrane, composed of mafic- ultramafic rocks, exhibits high Vp, high Vp/Vs and a thicker crust versus the thinner more quartz-rich crust of the eastern Sierras Pampeanas associated with low Vp and low

Vp/Vs. These differences may have some effect on the mechanism that unevenly generates crustal seismicity in the upper 30 km for this active compressional region.

Thus, mapping of the seismic properties can help us determine what controls the

Paleozoic terranes and their sutures may have on the present Andean deformation.

1. Introduction, Tectonic Setting and Terrane Overview

The highest (> 6000 m) elevations along the Andean cordillera lie above the transition in the subduction zone geometry where the Nazca slab changes its dip from near horizontal to ~30° (normal) (Fig. 1) (Barazangi and Isacks, 1976; Cahill and Isacks,

1992; Anderson et al., 2006). Both the convergence between the Nazca and South

American plates, at a rate of 6.3 cm/yr and azimuth of ~80º (Kendrick et al., 2003), and the subduction geometry are first order factors producing north-south variations in the volcanism and crustal seismicity (Jordan et al., 1983). The modern volcanic arc runs along the normally dipping slab south of 33°S (Stern, 2004). North of this latitude, however, volcanism was active until the most pronounced flattening of the subducted plate at ~8-10 Ma (Kay et al., 1991). Continental crustal seismic activity is very intense above the flat slab segment (Gutscher et al., 2000), a region characterized by the

Cambrian-Ordovician carbonate thin-skinned Precordillera fold-thrust belt (Baldis and

Bordonaro, 1981) and the thick-skinned Sierras Pampeanas basement uplifts, considered a modern example to interpret the Laramide deformation in North America (Jordan and

Allmendinger, 1986) (Fig. 1). South of ~33°S, the frecuency of crustal earthquakes in the high Cordillera is also high (Barrientos et al., 2004).

The south-central Andes backarc region is a mosaic of accreted terranes (Fig. 1) with two main collisions against the Gondwana margin. The Pampia terrane collided at

~530 Ma (Rapela, 2000), and the composite Cuyania terrane collided at ~460 Ma

(Ramos, 2004). The Cuyania terrane is composed of different terranes: the Precordillera and the westernmost Sierras Pampeanas, neither of which show evidence of Cambrian- 71

Ordovician magmatism related to active subduction (e.g. Ramos, 2004). In addition, basement rocks of a Grenville (1000-1200 Ma) age in the Precordillera (Kay et al., 1996) and in this sector of the Western Sierras Pampeanas (WSP) (McDonough et al., 1993) seem to be part of the same basement structure with an intra-Grenville suture (Vujovich and Kay, 1998; Vujovich et al., 2004) (Fig. 1). Yet no consensus exists regarding their time of amalgamation or individual provenances (e.g. Aceñolaza et al., 2002; Thomas and Astini, 2003; Finney et al., 2003, 2004; Galindo et al., 2004; Dalla Salda et al.,

1992). Additional U-Pb Grenville age observations at ~37°S suggest the Cuyania terrane extends to the south (Sato et al., 2004; Ramos, 2004). The most recent accretionary event produced the onset of the allochthonous Chilenia terrane in the west at ~400 Ma (Davis et al., 1999). The superimposed rifting events in Late Paleozoic and Mesozoic caused extensive felsic magmatism and extensional faulting, sometimes developing sedimentary backarc basins along sutures of the Paleozoic terranes (Franzese and Spalletti, 2001;

Franzese et al., 2003). The Andean compression inverted these previous extensional structures associated with the uplift of the Sierras Pampeanas (Ramos, 1994; Ramos et al., 2002).

The contrasting terrane inheritance may be responsible for the different crustal composition and tectonic evolution of the eastern and western terranes. The predominant felsic quartz-rich character of the Eastern Sierras Pampeanas (ESP) is linked to the collision of a micro-continent, the Pampia terrane (Rapela et al., 1998), whereas the dominant mafic-ultramafic composition in the west is associated with oceanic fragments 72

of a crust arc/back-arc setting for the WSP (Vujovich and Kay, 1998; Ramos et al., 2000) and a passive-margin platform for the Precordillera (Thomas and Astini, 2003).

Given the increasing interest in relating provenance and composition of the crust with the evolution of active margins, we have analyzed broadband seismic waveforms from regional crustal earthquakes in the central-western part of Argentina using a forward modeling technique to investigate the characteristics and sensitivity of the waveforms to seismic parameters in the crust and upper mantle. Investigating the seismic properties along single station-source pairs in each terrane enable us to compare and contrast the heterogeneous crust of this region and explore the implications for Andean tectonic processes.

2. Previous Regional Seismic Studies

Recent regional studies using broadband data from the CHile ARgentina

Geophysical Experiment (CHARGE) have constrained a thick (~65 km) crust beneath the high Cordillera, which thins slightly to ~55 km in the Precordillera and WSP along the northern CHARGE transect (Fig. 1; Fromm et al., 2004). This region has high P-wave velocity (Vp) of 6.4 km/s (Alvarado et al., 2005) and high P- to S-wave velocity ratio

(Vp/Vs) > 1.80 (Gilbert et al., 2006). In addition, a complex multilayer structure with probable eclogitization in the lower levels characterizes the western terrane crust, producing a weak teleseismic receiver function Moho converted phase arrival (Gilbert et al., 2006; Calkins et al., 2006). In contrast, the same studies show that the easternmost 73

terranes have a higher-amplitude and single Moho signal defining a much thinner (~35 km) crust with lower Vp of 6.0 km/s and low Vp/Vs < 1.70. Local studies using the

PANDA array had previously suggested a thick (> 50 km) crust in the WSP and

Precordillera near San Juan (Regnier et al., 1994). Seismic reflection studies in the basins around station JUAN in the WSP have shown reflectors at 18, 30 and 38 km within the crust (Fig. 1) (Snyder et al., 1990; Comínguez and Ramos, 1991; Zapata and

Allmendinger, 1996) and a transitional acoustic boundary at the base of a ~52-km crust

(Zapata, 1998). At ~36°S mid-crustal receiver function arrivals along the southern

CHARGE transect are consistent with thicker crust related to the active arc, which extends into the backarc region (Fig. 1) (Gilbert et al., 2006).

We explore the crustal structure in between the two CHARGE transects and within the north-south trending terranes. Although we selected mainly north-south earthquake-receiver paths to conduct a grid-search study, we occasionally considered other geometries to test particular problems. For example, we use the only east-west available path in the Río de la Plata craton (Fig. 1) to study the crust in this terrane.

3. Methods

In the following two sections we describe the data used in the regional forward broadband waveform modeling and the technique applied in this study.

3.1. Regional Seismic Moment Tensor Inversion Data

A previous study constrained the seismic moment tensor (SMT) for 27 crustal earthquakes, with magnitudes Mw between 3.5 and 5.1, recorded at regional distances out

to 700 km on the portable broadband CHARGE stations (Alvarado et al., 2005), using an

inversion technique from Randall et al. (1995). This technique consists of modeling

multiple three-component full-waveforms with good azimuthal coverage, for a fixed

hypocenter using an average 1D crustal model. A series of sensitivity tests including a

range of epicentral distances in the inversion were run to evaluate errors related to

earthquake mislocations. For each earthquake, a SMT inversion at a series of trial

hypocentral crustal depths was performed. In addition, several average crustal models

were also tested to observe variations in the SMT inversion results for selected events.

Alvarado et al. (2005) showed that SMT inversion solutions are very robust for the focal

mechanism, and the best fit between observed and synthetic waveforms occurs for the

best hypocentral depth. Varying the average crustal structure caused no significant

variations in the SMT results for the major dislocation and the best hypocentral depth

(Alvarado et al., 2005).

In this study, we consider those same crustal events. Fixing the previously

determined source parameter estimates (SMT for the major dislocation and the modeled

hypocentral depth) for a particular crustal earthquake, we performed a more detailed

analysis of the crustal structure between the event and each CHARGE station. By

investigating a number of event-station paths sampling different terranes we develop a

regional crustal velocity model. 75

3.2. Regional Forward Broadband Modeling

We have used twelve crustal events in the Andean high Cordillera and backarc region from Alvarado et al. (2005) ranging in size from Mw = 4.5 to 5.1 and broadband

CHARGE data recorded at epicentral distances < 400 km (Figs. 1 and 2). We refer to these earthquakes by their year and Julian day of occurrence. These moderate earthquakes have similar epicentral locations determined by the local network of INPRES operating in the region, relocation techniques using the CHARGE data (Anderson et al.,

2006), and body wave tomographic CHARGE studies (Wagner et al., 2005). The events display reverse or strike-slip focal mechanisms and focal depths from 3 to 26 km (Table 1 and Fig. 1).

The broadband waveforms of these crustal earthquakes display clear direct P- wave (Pg) and head P-wave (Pn) arrivals, followed by reflections and conversions between the surface and the Moho. The most prominent arrivals in the waveforms are the surface waves.

We use the forward modeling technique of Randall (1994), which consists of calculating the three-component full waveforms at regional distances using a layered velocity structure. The most reliable structure between the earthquake and a single station should produce the best correlation of the synthetic to the observed data.

As a first step we conducted a series of sensitivity tests to recognize which parameters of the crust and upper mantle have a stronger effect on the waveforms and whether there are any trade offs by using combinations of them. Then, based on the different crustal properties observed in the CHARGE region, we performed a grid search 76

for seismic velocity parameters (average crustal P- and S-wave velocities and thickness) that produce the best correlation between synthetic and observed data for the three- component broadband records at regional distances. We use this information to map regional crustal differences.

4. Regional Waveform Sensitivity Tests

We have calculated synthetic three-component seismograms for the largest (Mw

5.1) crustal earthquake during the CHARGE period, event 01-285 (Fig. 1), to investigate the sensitivity of the complete waveforms to crustal thickness, average crustal Vp, average crustal Vs, upper mantle Vp and upper mantle Vs. We fixed the seismic source at

5 km depth with a focal mechanism of strike 36°, dip 75° and rake 198° (Alvarado et al.,

2005). We used an epicentral distance and azimuth of 310 km and 192°, respectively, to simulate the recording conditions at station MAUL (Figs. 1 and 2). We used the TauP- toolkit software (Crotwell et al., 1999) to create our simple seismic models and to predict arrivals and their traveltimes. We observed the Pg, Pn, the Moho reflection Pmp, Sg, Sn,

Love and Rayleigh waves as clear arrivals. In addition, there are several phases between the P- and S-wavetrains, such as the phase PmpSmp with long period energy but smaller amplitudes at regional distances. Lower case “p” denotes the traditional representation for an upward traveling P-wave. Overall, the waveforms in this study are dominated by well- developed surface waves which are very sensitive to crustal parameters, as shown below.

4.1. Sensitivity to Crustal Thickness

We started testing sensitivity of the waveforms to crustal thickness using a crustal layer-over-mantle halfspace model. We fixed the crustal Vp at 6.4 km/s, mantle Vp at

8.15 km/s and Vp/Vs at 1.80, and calculated three-component full waveforms for crustal thicknesses varying from 30 to 75 km in steps of 5 km. The fixed crustal parameters are consistent with previous observations in this region (Fromm et al., 2004; Alvarado et al.,

2005; Gilbert et al., 2006). For simplicity, Fig. 3 shows the results of this sensitivity analysis for five different thicknesses on the radial seismic displacement component, which is affected by both P- and S-waves. In this and following waveform figures, all seismograms are aligned on the first P-wave arrival. We observe a notable variation in the amplitudes of the P-waves and the surface waves, which show an increase in amplitude for an increase in the crustal thickness on all three components. There are also variations in timing and amplitude of the phases arriving between the P- and S-waves.

4.2. Sensitivity to Average Crustal P-Wave Velocity

We explore the sensitivity of the waveforms to variations in crustal Vp. We fixed the crustal thickness at 50 km, used mantle parameters as before and varied the average crustal Vp from 5.6 to 6.6 km/s in steps of 0.2 km/s (Fig. 4). Since this will also introduce variations in Vp/Vs for the crust, we fixed Vp/Vs at 1.80 calculating the associated S- wave velocity in each model (Vs varying from 3.11 to 3.67 km/s). The most direct impact on waveforms of variations in crustal Vp is a variation in timing and amplitudes of the

Rayleigh waves and the intermediate arrivals between P and S waves (Fig. 4). Increasing 78

the crustal Vp decreases the traveltime and decreases the amplitude between peaks and troughs for the waves. This is to be expected, because the contrast between crustal and mantle velocities is diminishing.

4.3. Sensitivity to Crustal P- to S-Wave Velocity Ratio

We analyze the sensitivity of waveforms to the crustal Vp/Vs by initially fixing

Vp at 6.4 km/s and allowing Vp/Vs to vary from 1.60 to 1.90. We also fixed the crustal thickness at 50 km and mantle parameters as before. We repeated this test for different fixed crustal Vp calculating each time the corresponding Vs to produce the same Vp/Vs range explored. As expected, increasing Vp/Vs produces a significant delay in the S waves and surface waves (Fig. 5). On the filtered records the large amplitude arrival increases for a greater Vp/Vs on the tangential component. The effect is also more visible on the radial than on the vertical component. This makes sense, as the tangential and radial components are dominated by S waves (Fig. 5).

4.4. Sensitivity to Mantle Parameters

We ran two different tests allowing first variations in the mantle P-wave velocity and later in the mantle Vp/Vs ratio. We started fixing the average crustal Vp at 6.4 km/s, crustal and mantle Vp/Vs at 1.80 and crustal thickness at 50 km and computed synthetic waveforms for a series of mantle Vp varying between 7.6 and 8.4 km/s. The second test consisted of fixing the crustal parameters as described above, fixing the mantle Vp at

8.15 km/s and calculating synthetic seismograms for Vp/Vs in the mantle varying from 79

1.65 to 1.90. This range in Vp/Vs corresponds to variations in the mantle S-wave velocity from 4.23 to 4.94 km/s. Neither variations in Vp nor Vp/Vs for the mantle significantly modify the regional calculated waveforms (Figs. 6 and 7). This result is very useful in our grid search approach. Because waveforms are much less sensitive to the mantle parameters, these parameters can be fixed.

5. Grid-Search Forward Modeling to Determine Crustal Structure

The sensitivity test results indicate that regional broadband waveforms from crustal earthquakes are much more sensitive to parameters in the crust than in the upper mantle. However there is a trade off among the crustal parameters themselves (thickness,

Vp and Vs). We find that these waveforms are more sensitive to variations in crustal thickness than in Vp and Vs (compare Figs. 3, 4 and 5). Since variations in the mantle parameters do not produce significant variations in the waveforms, we fixed Vp at 8.15 km/s and Vp/Vs at 1.80 in the upper mantle for all tested models. These estimations are in agreement with global observations (Kennett and Engdahl, 1991) and other CHARGE determinations (Fromm et al., 2004; Wagner et al., 2005).

In order to refine the grid search around the crustal parameters we simultaneously tested variations of the crustal thickness, Vp and Vp/Vs using sensitivity analysis to identify the appropriate crustal parameters producing the best matches between synthetic and observed seismograms. For every single path, we tested more than 500 crustal models consisting of a layer-over-halfspace with fixed mantle (halfspace) parameters. 80

Our method investigated crustal thickness variations of 20 to 75 km with a grid spacing of 5 km, crustal velocities Vp of 5.4 to 7.0 km/s with a grid spacing of 0.2 km/s, and crustal velocities Vs that produce Vp/Vs of 1.60 to 1.90 with a grid spacing of 0.05. In order to avoid circular reasoning, we varied the crustal thickness and Vp within the specified ranges using a fixed Vp/Vs. We then repeated the same process for a different

Vp/Vs, and so on, until completing the test for all Vp/Vs values. For each combination of crustal parameters, we calculated synthetic broadband full waveforms at a given regional epicentral distance and azimuth using the reflectivity code of Randall (1994). The observed and synthetic data were then filtered with a Butterworth bandpass filter from intermediate to long periods. In general, we applied a bandpass filter of 10 to 80 s for earthquakes with Mw > 5, 10 to 50 s for 4 < Mw < 5 and 10 to 35 s for Mw ≤ 4. We did not test smaller period wavelengths because they become more sensitive to small-scale lateral heterogeneities, which are not resolvable by our method, and are often contaminated with noise. After filtering, we interpolated synthetic and observed data using the same sample rate. Finally, we compared every (vertical, radial and tangential) synthetic seismogram with the corresponding observed seismic component aligned at the P-wave first arrival by estimating their cross-correlation coefficient and taking an average over the three components. The correlation coefficient varies between -1 for totally uncorrelated data to

1 for perfectly correlated data. One single map of averaged correlation coefficients for crustal thickness as a function of Vp was constructed for each (fixed) Vp/Vs tested. We determined the maximum in each map and carefully examined its corresponding 81

seismogram matches. We discuss in detail the results for the pairs 02-005/BARD and 01-

352/PICH to illustrate the grid search analysis.

Fig. 8 shows an example of the grid-search results around crustal thickness and

Vp for event 02-005 and station BARD (see the seismic path in Fig. 2). Each point in the map is the correlation coefficient between synthetic and observed data, averaged over the three-component waveforms, which were bandpass filtered between 10 and 50 s and compared over 120 s of record length. In this example, we fixed Vp/Vs at 1.80 in the crust and mantle parameters remained fixed as before. The maximum correlation (0.93) over this grid is considered to be the best layer model, with a of Vp of 6.2 km/s and thickness of 35 km. Table 1 compiles these data, together with the results for a range of correlation coefficients corresponding to waveform matches that differ by less than 5 % of the maximum correlation coefficient obtained for each component (Fig. 8b). The inspection of the waveform matches around the maximum averaged correlation coefficient enables us to discriminate high quality results (shown in Table 1) from estimations of less correlated data (compare waveform matches in Figs. 8b, c and d). The acceptable waveform matches outline the broadness of the best crustal model solution and gives some uncertainty (Table 1). In a few cases the map of correlation has a tendency to define a secondary maximum peak in the contouring of average correlation coefficients for the same event-station path, which we are also reporting in Table 1. For example, Fig. 8a shows the occurrence of a secondary peak with average correlation coefficient of 0.77 around thicknesses of 47-57 km and Vp of 6.0-6.2 km/s. Although this coefficient may not be related to acceptable fits we still include it in Table 1 to show the 82

behavior around the maximum average correlation coefficient. This enables us to better identify regions of single well-defined maximum peaks.

Another example is shown in Fig. 9 for event 01-352 and station PICH located at an epicentral distance of 208 km (Fig. 2). A set of forward synthetic three-component waveforms is calculated at a series of crustal Vp/Vs, varying crustal Vp and crustal thickness. Mantle parameters remained fixed as before. We use a bandpass filter between

10 and 50 s and a time window of 72 s after the first P-wave arrival to estimate the cross- correlation between synthetic and observed waveforms. Shown in Fig. 9a is the map of correlation coefficients between forward modeled synthetic waveforms, using a fixed crustal Vp/Vs of 1.70, and observed waveforms, averaged among the three components.

A similar map was created for each of the six different Vp/Vs values explored for this seismic path. A Vp/Vs ratio of 1.70 produces the overall maximum correlation coefficient in the entire grid-search process. The corresponding best layer-over-halfspace model occurs for a crustal thickness of 45 km and a crustal Vp of 5.8 km/s, with an average correlation coefficient of 0.94 (red dot in Fig. 9). For comparison, forward-modeled waveforms corresponding to relatively smaller maximum average correlation coefficients, determined for the six different crustal Vp/Vs explored in the grid-search, are also shown (Fig. 9b). Although all models (last column in Fig. 9b) define a maximum correlation between synthetic and observed seismograms for thicknesses varying from 30 to 45 km, the best overall matches indicate a thickness of 45 km, a low crustal Vp of 5.8 km/s and a low Vp/Vs of 1.70 (Fig. 9 and Table 1). 83

The last example (Fig. 10) shows the grid search results for two different regions using the same event (01-285) but different stations (NEGR and LENA). The three- component synthetic waveforms were calculated for the pairs 01-285/NEGR and 01-

285/LENA (Fig. 2) separately, for combinations of crustal parameters (thickness, Vp and

Vp/Vs) varying within the range of all reasonable estimates and maintaining mantle parameters fixed as before. For each thickness-Vp-Vp/Vs model, the synthetic waveforms were compared to observed data using a bandpass filter of 10 to 80 s and a time record window of 140 s for the pair 01-285/NEGR. The same grid search was carried out for the pair 01-285/LENA using a bandpass filter between 15 and 50 s and a time record window of 50 s (Fig. 10c and d). We calculated the correlation coefficient for the respective displacement components and took the average among the three components. The best single-layer model for the northern seismic path using station

NEGR has an average correlation coefficient of 0.93, with a crustal thickness of 60 km, a crustal Vp of 6.6 km/s and a Vp/Vs of 1.85 (Fig. 10a and b, Table 1). The best results for the pair 01-285/LENA occur for an average correlation coefficient of 0.91 with a crustal thickness of 45 km, a Vp of 5.8 km/s and a Vp/Vs of 1.80 (Fig. 10c and d, Table 1).

Other possible combinations of crustal parameters along the Cordillera are also acceptable for both examples, as reported in Table 1.

Using the same analysis as used for the single station-event pairs 02-005/BARD,

01-352/PICH, 01-285/NEGR and 01-285/LENA we constrained the best layer-over- halfspace model for a total of 31 different paths (Fig. 2 and 11). In some cases one of the 84

observed seismic components was noisy but we were able to use the remaining components during the grid search analysis (Table 1).

6. Results and Discussion

Our results for Vp and Vp/Vs in the crust are summarized in Fig. 11. Both maps show an interpolation of the results listed in Table 1. We considered one data point every one degree along each seismic path. For example, a seismic path corresponding to an epicentral distance of 300 km is represented by three Vp (or Vp/Vs) data points with a spacing of ~100 km. On the basis of previous CHARGE studies (e.g. Gilbert et al., 2006), we note that the listed thicknesses in Table 1 do not necessarily imply Moho depths. This is the case for seismic paths in the WSP and the Precordillera as we discuss in section

6.3. If more than one thickness is listed in Table 1 for a particular event-station path (e.g.

01-352/JUAN), we consider the Vp and Vp/Vs estimations for the best layer model associated with the overall maximum averaged correlation coefficient, defined by the first listed thickness.

The map in Fig. 12a shows our crustal thickness results using data points as in

Fig. 11. In addition, this map includes previous regional crustal thickness results compiled by Gilbert et al. (2006) and described in Section 2, as well as recent local determinations beneath station JUAN by Calkins et al. (2006) and teleseismic observations of pmP phases at 31°S and 69°W beneath the Precordillera by McGlashan et al. (2006). 85

A common feature observed in the region is the expected larger thickness beneath the high Cordillera and its decrease in the eastern terranes that support lower elevations

(Fig. 12a). This is the best estimated parameter in the grid-search. To the north of 33°S

we found a greater thickness beneath the high Cordillera than to the south (Figs. 10 and

12a). This thick crust is also clearly observed along paths in the Precordillera fold-thrust

belt (Fig. 2 and Table 1). This is consistent with an overthickened crust due to tectonic

shortening, where the slab subducts horizontally as suggested by Allmendinger et al.

(1990).

Another feature observed from the numerous single event-station paths is the

different character of the best layer-over-halfspace models between different regions. The

eastern terranes and high Cordillera show one single crustal layer, perhaps in isostatic

compensation with their elevations (Figs. 2 and 12a). If velocity discontinuities exist

within the crust of these regions, they are relatively minor in comparison with the crust-

mantle boundary for our surface wave modeling. The only results available for a

comparison in both regions are from receiver function analysis, which show a very

defined Moho signal for the ESP (beneath stations LLAN and PICH, Fig. 2) (Gilbert et

al., 2006). In contrast, the WSP and Precordillera terranes exhibit a more complex

structure that makes it difficult to identify one single crustal layer model (e.g. 01-

127/PACH).

We have compared our results with empirical data from a compilation of a wide

variety of lithologies in active continental margins by Brocher (2005) (Fig. 13). This

compilation of Vp and Vp/Vs, based on independent estimations of Vp and Vs, 86

corresponds to specific lithologies from laboratory and field measurements, borehole, seismic vertical profiles and seismic tomography. Our constraints are shown by rectangles in Fig. 13. We note that Vs determinations are more important than Vp to place our constraints within mafic/felsic limits and, some of the expected variation of the elastic constants with depth within the crust. We discuss each region below.

6.1. High Cordillera and Active Arc

North of 33ºS, the best observed-synthetic waveform matches were achieved using a crustal layer of 50-60 km thickness with average Vp of 6.4-6.6 km/s and Vp/Vs of ~1.85 (Figs. 10a, 11 and 12a). These results are consistent with an intermediate to mafic crustal composition (Christensen and Mooney, 1995; Brocher, 2005) (Fig. 13).

To the south, results along the active volcanic arc indicate also a high Vp/Vs ratio

(> 1.80), with a much lower Vp (5.8-6.0 km/s) (Fig. 10c and d, 11 and 12a). Our grid search results suggest a crustal layer model of ~45 km along the active arc.

The magmatic fluid source above the normal subduction segment south of ~33°S could decrease the P- and, still more dramatically, the S-wave velocities. In contrast, the results in the currently amagmatic northern segment are less straightforward and require more discussion. Comparing this region with its adjacent southern segment in the high

Cordillera, the high Vp/Vs ratio could be in part explained by the greater Vp (> 6.4 km/s) values but also by the low Vs. Magmatism in this northern part of the high Cordillera has been interpreted to cease at 6-5 Ma as the subducted slab flattened to the east (Kay et al.,

1991). Significant periods of andesitic and dacitic volcanism have occurred between 27 87

and 8 Ma (Kay et al., 1999; Bissig et al., 2003), and silicic volcanism at ~5 Ma (Ramos et al., 1989) with mineralization intruding granitiod basement rocks. Moreover, studies of rhyolitic rocks showing negative Eu anomalies in the northern segment have revealed that localized melting continued until ~2 Ma (Bissig et al., 2002). Thus, recent partial melting is associated with the economic mineralization shown in Figs. 1 and 2 in the northern segment of the high Cordillera. Different tectonic-magmatic mechanisms have been proposed to explain these melts, such as flat-slab derived melting (Gutscher et al.,

2000b), secondary melting within the crust (Kay and Mpodozis, 2002), or high-pressure melting and assimilation in the lower crust (Bissig et al., 2003).

The remarkable contrast in Vp between the northern and active arc segments is consistent with a much cooler northern region, although some possible localized melt in the lower crust cannot be ruled out, as shown by the low S-wave regional velocities.

Although different average Vp estimations characterize the northern high cordillera and the active arc, it is likely that they have similar mafic compositions. As shown in Fig. 13, both regions correlate with the lithologies along the mafic line. According to Brocher

(2005), the mafic line represents mafic and calcium-rich rocks, gabbros which have a Vs

0.2 to 0.3 km/s below general empirical Vp versus Vp/Vs trends. These results taken together indicate a relatively larger percentage of partial melt along the active arc. As laboratory measurements show, the effect of partial melt dramatically decreases Vp and increases Vp/Vs ratio independently of the melt composition or confining pressure

(Makovsky and Klemperer, 1999). The region immediately east of the active arc shows 88

an increase in both Vp and Vs which might be related to a region of relatively cooler backarc volcanism.

Another interesting observation is the thicker (more than 15 to 20 km) crust observed in the northern segment compared to the active arc. The thickening of an initially ~40-km thick crust in the northern segment started around 27 Ma (Kay and

Mpodozis, 2002). Thus, the flattening of the slab seems to be the best candidate to account for thickening and shortening in this region.

6.2. Eastern Sierras Pampeanas

The average crustal model in the Pampia terrane basement indicates low values of

Vp of 5.8-6.0 km/s, a layer of 35-45 km thickness and low Vp/Vs of 1.65-1.70 (E in Fig.

13). Both the Pampia and Rio de la Plata terranes have a more felsic quartz-rich composition (Rapela et al., 1998) compared to the western basement terranes. Our results are consistent with some granite-gneiss crystalline lithologies, as predicted by Brocher’s

(2005) data (Fig. 13) and in good agreement with the exposed rocks in this region

(Rapela et al., 1998). These results show a contrast with the properties of the western terranes as we discuss in next section.

6.3. Precordillera and Western Sierras Pampeanas

Our results for the Precordillera and WSP region using a simple one crustal layer indicate high Vp (~6.4 km/s), high Vp/Vs (1.80-1.85) and thickness of 35-40 km (Figs.

11 and 12a). 89

Comparing the Cuyania terrane with the Cordillera region above the flat slab segment, we observe a slightly lower Vp and similar Vp/Vs along the Cuyania terrane

(Fig. 11). Our determinations for the Cuyania terrane lie in the region of metagraywackes and mafic rocks, yet are close to the mafic-line lithologies (symbol W in Fig. 13). We note that good fits are achieved using a very simple seismic velocity structure with a crustal thickness of 35 to 40 km and a halfspace (with mantle parameters) immediately below (e.g. 02-117/USPA; 02-107/JUAN). Based on receiver function, Pn and multiple broadband waveform inversions that also used the CHARGE data (Gilbert et al., 2006;

Calkins et al., 2006; Alvarado et al., 2005; Fromm et al., 2004), local converted at the

Moho phases (Regnier et al., 1992) and teleseismic observations of the pmP phase beneath the Precordillera (McGlashan et al., 2006), a Moho depth of ~50-55 km is determined for this region (Fig. 12a). For this reason, a densification in the lower crust has been suggested by Gilbert et al. (2006), Calkins et al., (2006) and Snyder et al.,

(1990), which would also reconcile seismic observations and gravimetric data (e.g.

Smalley and Introcaso, 2003). Therefore, this region is likely characterized by crustal models consisting of more than one single layer.

In light of previous CHARGE studies suggesting a gradational increase of Vp and

Vs between 40 and 55 km depth in the lower crust for the Cuyania terrane (Calkins et al.,

2006), we have tested for the presence of a higher velocity lower crust. We used four different paths. One example is shown for the pair 02-117/USPA in Fig. 14. We note that using the receiver function model (Model_2 (RF), Fig. 14) from Calkins et al. (2006) to calculate synthetic waveforms diminishes the quality of the fits compared to our single 90

(Model_1) layer model. We note that the receiver function study is only sensitive to discontinuities and gradients in the velocity model but not to absolute velocities. Hence, a slight modification to the upper crustal Vs (Model_3) related to a decrease in Vp/Vs gives better fits. Thus, if a more realistic model with increasing Vp in the lower crust is taken into account, a different structure in the upper crust is also needed which can be represented by our simple average models. These results suggest that if this lower crust exists, it must have intermediate (between crust and upper mantle) parameters. This conclusion is based on observations that forward modeling exhibits good fit to the observed waveforms when using single layer-over-halfspace models simulating a Moho depth of ~35-40 km (Table 1 and Fig. 14). We also note that using higher frequencies, which are more sensitive to crustal parameters, might be useful to explore details of the lower crust. However, for the magnitude sizes of the earthquakes and epicentral distances in this study, we are not able to use higher frequency data due to the noise content.

Further work is needed to determine more detailed crustal models.

The Pampia and Cuyania terranes show different average P-wave and S-wave velocities. These differences might account for variations in their elastic parameters in a very simplistic model (Fig. 13, Brocher, 2005). However seismic velocities can be also affected by other possible factors such as fluid content, presence of sediments or local heterogeneities, degree of fracture, temperature variations, among others. In any case our results are the first independently quantifiable seismic parameters at a regional scale for comparison across the terranes. Although it is difficult to correlate the velocity variations 91

with one single responsible factor, we suggest that the implications might be controlling intraplate crustal seismicity nucleation.

7. Comparison of Crustal Structure with Regional Crustal Seismicity

We observed a higher rate of seismicity and moment release in the Cuyania terrane, and a high Vp/Vs ratio and evidence for a distinct two layer crust with a possible partially ecologitized lower crust. In an attempt to explain the crustal earthquake distribution and the crustal structure of the Cuyania terrane we discuss several possible factors including plate convergence rate, heat flow, down-going plate geometry, composition, and increased strain (as shown by the GPS measurements).

7.1. Observations

Seismicity over the last 20 years in the region of study is not biased by the occurrence of any large (magnitude > 6.4) earthquake. It is also the best local seismicity detected by local networks in Chile and Argentina in terms of station coverage. This seismicity shows that the Andean crustal earthquake distribution is highly dense in the backarc region near San Juan and Mendoza (Fig. 12b). The region above the flat slab generates more earthquakes and of a higher magnitude compared to adjacent segments, where the slab starts dipping more steeply (Gutscher et al., 2000a).

Over the last 100 years, three large damaging crustal earthquakes of magnitudes

Mw 7.0, 6.8 and 7.5 have occurred in the Precordillera and the WSP (Alvarado and Beck, 92

2006; Langer and Hartzell, 1996; Kadinsky-Cade, 1985). Local and regional studies show a predominance of thrust focal mechanisms with focal depths between 15 and 35 km for moderate seismicity (Smalley et al., 1993; Regnier et al., 1994; Alvarado et al., 2005;

Salazar, 2005).

This seismic activity contrasts with that of the region located immediately to the east of the active Sierra Valle Fértil Fault, coincident with the suture between the

Cuyania and Pampia terranes (Ramos, 1994) (Figs. 1 and 12b). The occurrence of crustal earthquakes in the Pampia terrane with magnitudes over 4.5 is very rare, although historical records point out one M ~6.0 event in the south, near San Luis (Fig. 12b)

(INPRES, 2006; Costa and Vita-Finzi, 1996). Seismicity in the Rio de la Plata craton is also lower than the crustal seismicity to the west in the Cuyania terrane.

An important distribution of crustal earthquakes occurs along the Sierra Valle

Fértil Fault (Fig. 12b). Studies along this major structure, based on seismic reflection data, ophiolite, and mylonite rock analysis, indicate this thrust fault is dipping to the east with a décollement zone at about 38 km depth (Snyder et al., 1990; Galindo et al., 2004).

This major structure has been affected by extensional Triassic to Jurassic deformation and later contractional deformation during the recent Andean compression (Ramos,

1994). The Sierra Valle Fértil Fault has likely generated several earthquakes along its northern termination detected by the global network (Chinn and Isacks, 1983). It also showed some reactivation in its central part during the 1977 earthquake (Kadinsky-Cade et al., 1985). This same reverse fault showed aligned events in its southern part detected by regional networks (Alvarado et al., 2005). 93

Further west, the Principal and Frontal Cordilleras above the flat slab segment have also had significantly fewer earthquakes of magnitudes > 3.5 during the last 20 years and no record of large-sized magnitude seismicity in the last 100 years (NEIC catalog), when compared with the eastern terranes.

Based on the intra-Andean deformation, regional GPS studies by Brooks et al.

(2003) have suggested a microplate behavior for the Andes relating larger crustal earthquake deformation to its orogenic front, which is more likely represented by the

Sierra Valle Fértil Fault at this latitude.

Crustal seismicity south of ~33°S is very intense in the Andes cordillera

(Barrientos et al., 2004). The larger crustal earthquakes along the main cordillera show a different style of deformation. These earthquakes have predominantly strike-slip focal mechanism solutions and shallower focal depths (< 15 km) (Piderit, 1961; Barrientos et al., 2004; Alvarado et al., 2005).

The backarc region to the south of ~33°S has been active in modern (Fig. 12b) and historical times, with several records of damaging crustal earthquakes located in the southern extension of the Precordillera in 1861 and 1985 (INPRES, 2006). This seismicity shows thrust fault focal mechanisms with depths of 10 to 26 km (Alvarado et al., 2005), although some normal fault first-motion focal mechanisms have been reported for small-to-moderate earthquakes using regional networks (Salazar, 2005).

7.2. Discussion

GPS studies show a systematically similar convergence vector along the western

South America boundary (Kendrick et al., 2003), where 50 to 90 % of the convergence is accommodated by permanent (earthquake) deformation (Brooks et al., 2003;

Allmendinger et al., 2005). Hence, the variations in crustal seismicity cannot be due to variations in plate convergence rate.

A colder rigid crust might be another source for explaining the more brittle behavior of the crust above the flat slab. Although sparse, heat flow measurements have shown a dramatic decrease in the WSP region with estimates of < 50 mW/m2 compared to its eastern region of 40 to 100 mW/m2 determinations (Hamza and Muñoz, 1996).

Although this could be a factor, a simple inspection shows that crustal seismicity extends into the active arc south of 33°S, which is associated with a higher heat flow of > 100 mW/m2 (Hamza and Muñoz, 1996).

New slab contours by Anderson et al. (2006) determined from local intermediate- depth CHARGE-relocated seismicity show a correlation between the flat slab geometry and the overriding region with a larger number of crustal earthquakes, of higher magnitudes and predominantly thrust fault focal mechanisms (Fig. 12b) (Kadinsky-Cade,

1985; Regnier et al., 1994; Alvarado et al., 2005; Salazar, 2005; Alvarado and Beck,

2006). For the same region, GPS velocity vectors in the South American upper plate are overall consistently oriented along the plate convergence direction and abruptly decay to the east of the Sierra Valle Fértil Fault (Brooks et al. 2003). So far, a simple mechanism that efficiently transfers shear stresses from the flat slab to the upper plate remains 95

enigmatic. Several authors (e.g. Gutscher et al., 2000; Yañez et al., 2002) suggest the subducting Juan Fernández Ridge is a factor that controls or enhances a more buoyant slab and hence, increases the coupling and refrigeration between the oceanic and continental plates. In addition, we note that crustal earthquakes are also observed near

Mendoza, where the slab inclines (Fig. 12b).

Our study delimits the Cuyania terrane above the flat slab as a region of higher

Vp/Vs when compared to the Pampia terrane. This is in agreement with other broadband seismic studies by Gilbert et al. (2006) and Calkins et al. (2006). Our grid-search results indicate that Vp, but mainly the lower Vs in the crust, is responsible for the high Vp/Vs ratio. Because the crust is overthickened in this part, Gilbert et al. (2006), Calkins et al.

(2006) and Snyder et al. (1990) have suggested a partial eclogitization of the lower crust which reconciles isostatic and seismic observations. This implies an increase in density, as well as in Vs, in the lower crust, which results in a lower contrast with mantle observations from tomographic studies (Wagner et al., 2005). We note that the high level of crustal seismicity (< 35 km) occurs above the high density lower crust. Our low Vs estimations related to high Vp/Vs ratios of 1.80-1.85 are likely representative of the upper crust. These values correlate with a mafic composition (Christensen, 1996) and the rocks exposed in this terrane. Laboratory measurements indicate that mafic crust is much stronger than felsic crust (e.g. Rutter and Brodie, 1992), which contrasts with the higher seismic activity in the Cuyania terrane. Taking this together it is possible that pre-existing structures might have weakened the basement through inherited foliation, schistosity, mylonitic zones, stratified foliation, or basement fabrics, which would directly lead to the 96

lower Vs in this terrane. Seismic reflection studies near San Juan have revealed such structural features (Zapata and Allmendinger, 1996; Comínguez and Ramos, 1991;

Snyder et al., 1990). According to Ramos et al. (2002) a series of rifts, with normal faults parallel to major discontinuites and associated with alkaline basalt flows, developed on the hanging-wall of the sutures between the Chilenia and Cuyania terranes and the

Cuyania and Pampia terranes. Thus, these accretionary and extensional pre-Andean processes created anisotropies that later controlled the resulting structures by inverting ancient normal faults. Examples of reactivation of extensional basement structures are also observed in the northern Sierras Pampeanas, where the development of large Andean faults are responsible for range uplifts having the same dip as the dominant fabric

(Cristallini et al., 2003).

We believe that a different mechanism may explain the crustal seismicity along the active arc south of ~33°S. Although GPS data in the backarc are scarce and uncertainties are large, an important decay of the GPS velocity vectors, accompanied by a slightly clockwise rotation, is observed when compared to forearc measurements, relative to cratonic stable eastern areas (Brooks et al., 2003; Allmendinger et al., 2005). In this case slip partitioning of the plate convergence between the Nazca and South America plates might be responsible for the crustal seismicity along the high cordillera, as suggested by Alvarado et al. (2005). This partitioning is consistent with the right-lateral strike-slip focal mechanisms, shallow focal depths (Barrientos et al., 2004; Alvarado et al., 2005) and north-south orientation of vertical faults in the region (Charrier et al., 2004;

Barrientos et al., 2004). In this case, crustal faults parallel to the trench may control the 97

ascending magmatic activity for the volcanoes in the arc. However the active arc and its type of volcanism are mainly the result of perpendicular convergence in a “normal” subduction zone. An alternative mechanism for the region south of ~33°S could be accommodating deformation between a transitional region of higher shortening in the north to lower shortening in the south (Rivera and Cembrano, 2000). This scenario predicts approximately east-west fault structures aligned more perpendicular to the trench. Studies around the Tupungatito volcano (the northernmost historically active stratovolcano in the arc) show an east-west distribution with 12 Holocene craters

(Gonzalez-Ferrán, 1995). Perhaps a combination of these processes is responsible for the seismic deformation observed in this segment. We note that our results are consistent with this type of strike-slip deformation generated at very shallow depths (< 5 km). Our results are also in agreement with a more intermediate (andesitic) composition, which can account for brittle behavior in the upper crust, as clearly observed in the hypocenter distribution in this area (Barrientos et al., 2004; Alvarado et al. 2005).

8. Conclusions

We have used a variety of earthquake-receiver geometries to analyze the crustal structure between the two CHARGE transects at 30° and 36°S. We have estimated the best simple crustal layer-over-halfspace model for each seismic path associated with different accreted terranes. Our results indicate crustal parameters of Vp of 6.4-6.6 km/s,

Vp/Vs of 1.80-1.85 and thickness of 50-60 km in the high Cordillera north of 33°S. The 98

active arc is best represented by a crust of Vp of 5.8-6.0 km/s, Vp/Vs ~1.80 and thickness of 40-45 km. The Cuyania terrane exhibits a more complex structure with Vp of 6.4 km/s,

Vp/Vs of 1.80-1.85 in the upper 40 km crust. The Pampia terrane has a thinner crust of

30-40 km with Vp of 5.8-6.0 km/s and Vp/Vs < 1.70.

We find evidence for an intermediate to mafic composition along the high

Cordillera, although conditions for a much cooler region dominate the northern thicker segment. The active arc is consistent with a region of comparatively significant amounts of melts.

An important contrast between the seismic properties of the Pampia and Cuyania terranes is evidenced by differences in crustal Vp, Vp/Vs ratio and structural complexity.

This correlates with the differences observed on the exposed rocks and the probable differences in the terrane provenances: a more mafic-ultramafic composition dominates the Cuyania terrane (e.g. Ramos, 2004) versus the more felsic rich-quartz composition of the Pampia terrane (e.g. Rapela et al., 1998).

The differences in the backarc terranes, evidenced by the crustal seismic parameters in this study, might be another factor controlling the nucleation of present-day seismicity. The Cuyania terrane has generated larger numbers of earthquakes of higher magnitudes from historic to modern times. Further east, the Pampia terrane and further west, the High Cordillera, are comparatively seismically quiet for moderate-sized events.

Although a greater seismic energy release from crustal earthquakes is directly linked to the Cuyania basement overriding the elongated flat slab and the Juan Fernández Ridge path, an important amount of seismic energy release occurs in the crust of the Cuyania 99

terrane where the slab inclines. We postulate that the western terranes with higher Vp/Vs might have more structures to reactivate, based on the complex history of accretion, rifting and re-accretion of Paleozoic terranes. These terranes show ancient faults and sutures across the middle and upper crust with mylonitic and structural grain and high deformation zones (Ramos, 1994), which provide pre-existing zones of weakness in the crust. This is also consistent with the region of probable partial eclogitization in the lower crust (Gilbert et al., 2006), which would also decrease its strength. This by no means excludes the possibility of the occurrence of large earthquakes in the eastern terranes, which are also under compression, although a magnitude larger than ~6.0 has not been reported for this region in historic times.

Acknowledgements

We acknowledge National Science Foundation grants EAR-9811878 and EAR-

0510966, which supported the CHARGE project (www.geo.arizona.edu/CHARGE). The instruments used in the field program were provided by the PASSCAL facility of the

Incorporated Research Institutions for Seismology (IRIS) through the PASSCAL

Instrument Center at New Mexico Tech. Data collected during this experiment will be available through the IRIS Data Management Center. The facilities of the IRIS

Consortium are supported by the NSF under Cooperative Agreement EAR-0004370. We thank the support of the Universidad Nacional de San Juan and INPRES in Argentina, and the Universidad de Chile. Processing of the seismic data was performed using 100

Seismic Analysis Code SAC2000 software (Goldstein et al., 1999). Maps were generated using Generic Mapping Tools software. We also used seismological data retrieved from the NEIC (USGS) and International Seismological Centre. We are grateful for the review of earlier manuscripts of this paper by Clem Chase, Roy Johnson, Rick Bennett and

Christine Gans.

101

References

Aceñolaza, F.G., Miller, H., and Toselli, A.J., 2002, Proterozoic-Early Paleozoic

evolution in western South America, a discussion, Tectonophysics 354, 121-137.

Allmendinger, R.W., Figueroa, D., Snyder, D., Beer, J., Mpodozis, C., and Isacks, B.L.,

1990, Foreland shortening and crustal balancing in the Andes at 30°S Latitude,

Tectonics 9, 789-809.

Allmendinger, R.W, Smalley, R.(Jr.), Bevis, M., Caprio, H., and Brooks, B., 2005,

Bending the Bolivian orocline in real time, Geology 33(11), 905-908.

Alvarado, P., Beck, S., Zandt, G., Araujo, M. and Triep, E., 2005, Crustal deformation in

the south-central Andes back-arc terranes as viewed from regional broadband seismic

waveform modelling, Geophys. J. Int. 163, 580-598.

Alvarado, P. and Beck, S., 2006, Source Characterization of the San Juan (Argentina)

Crustal Earthquakes of 15 January 1944 (Mw 7.0) and 11 June 1952 (Mw 6.8), Earth

and Planetary Science Letters 243, 615-631. Doi:10.1016/j.epsl.2006.01.015.

Anderson, M., Alvarado, P., Beck, S. and Zandt, G., 2006, Geometry and brittle

deformation of the subducting Nazca plate, Central Chile and Argentina submitted to

Geophys. J. Int.

Baldis, B. and Bordonaro, O., 1981, Evolución de las facies carbonáticas en la cuenca

Cámbrica de la Precordillera de San Juan, VIII Argentinean Geological Congress 2,

385-397, San Luis, Argentina.

Barazangi, M. and Isacks, B., 1976, Spatial distribution of earthquakes and subduction of

the Nazca plate beneath South America, Geology 4, 686-692. 102

Barrientos, S., Vera, E., Alvarado, P. and Monfret, T., 2004, Crustal seismicity in Central

Chile, J. South Am. Earth Sci. 16, 759-768.

Bissig, T., Clark, A.H. and Lee, J.K.W., 2002, Cerro de Vidrio rhyolitic dome: evidence

for Late Pliocene volcanism in the central Andean flat-slab region, Lama-Veladero

district, 29°20’S, San Juan Province, Argentina, J. South Am. Earth Sci. 15(5), 571-

576.

Bissig, T., Clark, A.H., Lee, J.K.W., and Quadt, A., 2003, Petrogenetic and

metallogenetic responses to Miocene slab flattening: new constraints from the El

Indio-Pascua Au–Ag–Cu belt, Chile/Argentina, Mineralium Deposita 38(7), 844-862.

Brocher, T., 2005, Empirical relations between elastic wavespeeds and density in the

earth’s crust, Bulletin of the Seismological Society of America 95(6), 2081-2092.

Brooks, B.A., Bevis, M., Smalley, R.(Jr.), Kendrick, E., Manceda, R., Lauría, E.,

Maturana, R. and Araujo, M., 2003, Crustal motion in the southern Andes (26°–

36°S): do the Andes behave like a microplate?, Geochem. Geophys. Geosyst. 4, 1-14.

Cahill, T. and Isacks, B., 1992, Seismicity and shape of the subducted Nazca plate, J.

Geophys. Res. 97, 17503–17529.

Calkins, J.A., Zandt, G., Gilbert, H. and Beck, S.L., 2006, Crustal images from San Juan,

Argentina, obtained using high frequency local event receiver functions, Geophys.

Res. Let. 33(7) L07309.

Charrier, R., Farías, M., Comte, D. and Pardo, M., 2004, Active tectonic in the Southern

Central Andes, a recent example: the 28 August, 2004 shallow Mw=6.5 earthquake, 103

Eos Trans. American Geophysical Union, 85(47), Fall meeting abstracts, Abstract

S43C-1015, San Francisco, U.S.A.

Chinn, D.S. and Isacks, B.L., 1983, Accurate source depths and focal mechanisms of

shallow earthquakes in western South America and New Hebrides Island Arc.

Tectonics 2, 529-563.

Christensen, N.I., 1996, Poisson’s ratio and crustal seismology, J. Geophys. Res. 101,

3139-3156.

Christensen, N. I. and Mooney, W. D., 1995, Seismic velocity structure and composition

of the continental crust: a global view, J. Geophys. Res. 100, 9761–9788.

Comínguez, A.H. and Ramos, V.A., 1991, La estructura profunda entre Precordillera y

Sierras Pampeanas de la Argentina: Evidencias de la sísmica de reflexión profunda,

Revista Geológica de Chile 18(1), 3-14.

Costa, C.H. and Vita-Finzi, C., 1996, Late Holocene faulting in the southeast Sierras

Pampeanas of Argentina, Geology 24, 1127-1130.

Cristallini, E.O., Comínguez, A.H., Ramos, V.A. and Mercerat, E.D., 2003, Basement

double-wedge thrusting in the northern Sierras Pampeanas of Argentina (27°S) –

Constraints from deep seismic reflection, in Thrust tectonics and hydrocarbon

systems, McClay, C.R. (Ed.) American Association of Petroleum Geologists memoir

82, pp. 1-26.

Crotwell, H.P., Owens, T.J., and Ritsema, J., 1999, The TauP toolkit: flexible seismic

travel-time and ray-path utilities, Seismological Research Letters 70, 154-160. 104

Dalla Salda, L.H., Dalziel, I.W.D., Cingolani, C., and Varela, R., 1992, Did the Taconic

Appalachians continue into southern South America?: Geology 20, 1059-1062.

Davis, J.S., Roeske, S.M. and McClelland, W.C., 1999, Closing the ocean between the

Precordillera terrane and Chilenia: Early Devonian ophiolite emplacement and

deformation in the SW Precordillera, in Laurentia-Gondwana connections before

Pangea, Ramos, V.A. and Keppie, D. (Eds.), Geological Society American Special

Paper, 336: 115-138, U.S.A.

Finney, S., Gleason, J., Gehrels, G., Peralta, S. and Aceñolaza, G., 2003, Early

Gondwanan connection for the Argentine Precordillera terrane, Earth and Planetary

Science Letters 205(3-4), 349-359.

Finney, S., Gleason, J., Gehrels, G., Peralta, S. and Aceñolaza, G., 2004, Corrigendum to

"Early Gondwanan connection for the Argentine Precordillera terrane": [Earth Planet.

Sci. Lett. 205 (2004) 349–359], Earth and Planetary Science Letters 219(3-4), pp.

413.

Franzese, J.R. and Spalletti, L.A., 2001, Late Triassic–early Jurassic continental

extension in southwestern Gondwana: tectonic segmentation and pre-break-up rifting,

J. South Am. Earth Sci. 14(3), 257-270.

Franzese, J., Spalletti, L., Gómez-Pérez, I. and Macdonald, D., 2003, Tectonic and

paleoenvironmental evolution of Mesozoic sedimentary basins along the Andean

foothills of Argentina (32°–54°S), J. South Am. Earth Sci. 16(1), 81-90. 105

Fromm, R., Zandt, G. and Beck, S.L., 2004, Crustal thickness beneath the Andes and

Sierras Pampeanas at 30°S inferred from Pn apparent phase velocities, Geophys. Res.

Lett. 31, L006625, doi:10.1029/2003GL019231.

Galindo, C., Casquet, C., Rapela, C., Pankhurst, R.J., Baldo, E. and Saavedra, J., 2004,

Sr, C and O isotope geochemistry and stratigraphy of Precambrian and lower

Paleozoic carbonate sequences from the Western Sierras Pampeanas of Argentina:

tectonic implications, Precambrian Research 131(1-2), 55-71.

Gilbert, H., Beck, S. and Zandt, G., 2006, Lithosperic and upper mantle structure of

Central Chile and Argentina, Geophys. J. Int. 165, 383-398.

Goldstein, P., Dodge, D. and Firpo, M., 1999, SAC2000: Signal processing and analysis

tools for seismologists and engineers, UCRL-JC-135963. Invited contribution to the

IASPEI International Handbook of Earthquake and Engineering Seismology, 1999.

Gonzalez-Ferrán, O., 1995, Volcanes de Chile. Instituto Geográfico Militar, Santiago,

Chile, 635 pp.

Gutscher, M.A., Spakman, W., Bijwaard, H. and Engdahl, R., 2000, Geodynamics of flat

subduction: Seismicity and tomographic constraints from the Andean margin,

Tectonics 19, 814-833.

Gutscher, M.A., Maury, R.C., Eissen, J.-P. and Bourdon, E., 2000, Can slab melting be

caused by flat subduction?, Geology 28(6), 535-538.

Hamza, V.M. and Muñoz, M., 1996, Heat flow map of South America, Geothermics

25(6), 599-646.

INPRES, 2006, Listado de terremotos históricos de Argentina. On-line catalogue. 106

Jordan, T.E., Isacks, B.L., Allmendinger, R.W., Brewer, J.A., Ramos, V.A. and Ando,

C.J., 1983, Andean tectonics related to the geometry od the subducted Nazca plate,

Geol. Soc. Am. Bull. 94, 341-361.

Jordan, T.E. and Allmendinger, R.W., 1986, The Sierras Pampeanas of Argentina: A

modern analogue of Rocky Mountain foreland deformation, Am. J. Sci. 286, 737-764.

Kadinsky-Cade, K., 1985, Seismotectonic of the Chilean margin and the 1977 Caucete

earthquake of western Argentina, PhD thesis, Cornell University, Ithaca, New York,

U.S.A., p. 253 (in English).

Kay, S., Mpodozis, C., Ramos, V. and Munizaga, F., 1991, Magma source variations for

mid-late Tertiary magmatic rocks associated with a shallowing subduction zone and a

thickening crust in the central Andes (28° to 33°S), in Andean magmatism and its

tectonic setting, Harmon, R.S. and Rapela, C.W., (Eds.), Geol. Soc. Am. Spec. Pap.

265, 113-137, Boulder, CO, USA.

Kay, S.M., Orrel, S. and Abruzzi, J.M., 1996, Zircon and whole rock Nd-Pb isotopic

evidence for a Grenville age and Laurentian origin for the basement of Precordillera

terrane in Argentina, Journal of Geology 104, 637-648.

Kay, S., Mpodozis, C. and Coira, B., 1999, Neogene magmatism, tectonism and mineral

deposits of the Central Andes (22 - 33°S latitude), in Geology and ore deposits of the

Central Andes, Skinner, B.J. (Ed.), Society of Economic Geologists special

Publication No. 7, College Station, Texas, USA, pp. 27-59. 107

Kay, S.M. and Mpodozis, C., 2002, Magmatism as a probe to the Neogene shallowing of

the Nazca plate beneath the modern Chilean flat slab, J. South Am. Earth Sci.15, 39-

59.

Kendrick, E., Bevis, M., Smalley, R.(Jr.), Brooks, B.A., Barriga, R., Lauría, E., Souto,

L.P., 2003, The Nazca-South America Euler vector and its rate of change, J. South

Am. Earth Sci. 16, 125-131.

Kennett, B.L.N. and Engdahl, E.R., 1991, Traveltimes for global earthquake location and

phase identification, Geophys. J. Int. 122, 429-465.

Langer, C.J. and Hartzell, S., 1996, Rupture distribution of the 1977 western Argentina

earthquake, Phys. Earth Planet. Inter. 94, 121–132.

Makovsky, Y. and Klemperer, S.L., 1999, Measuring the seismic properties of Tibetan

bright spots: evidence for free aqueous fluids in the tibetan middle crust, Journal of

Geophysical Research 104, 10,795-10,825.

McDonough, M.R., Ramos, V.A., Isachsen, C.E., Bowring, S.A. and Vujovich, G.I.,

1993, Edades preliminares de circones del basamento de la Sierra de Pie de Palo,

Sierras Pampeanas Occidentales de San Juan: Sus implicancias para el

supercontinente Proterozoico de Rodinia, XII Argentinean Geological Congress and

II Congreso de Exploración de Hidrocarburos, 3: 340-342, Mendoza, Argentina.

McGlashan, N.A., Brown, L.D. and Kay, S., 2006, Crustal thickness in the Andes from

teleseismically recorded depth-phase precursors, GSA meeting, Paper 12-9, Backbone

of the Americas, Mendoza, Argentina, April, 2006. 108

Piderit, E., 1961, Estudio de los sismos del Cajón del Maipo del año 1958. Civil

Engineering memoir, Universidad de Chile, Santiago, Chile, p. 125 (in Spanish).

Ramos, V.A., 1994, Terranes of southern Gondwanaland and their control in the Andean

structure (30°-33°S latitude), in Tectonics of the Southern Central Andes, Reutter,

K.J., Scheuber, E. and Wigger, P.J. (Eds.), pp. 249-261, Springer-Verlag, New York,

U.S.A.

Ramos, V.A., 2004, Cuyania, an exotic block to Gondwana: review of a historical

success and the present problems, Gondwana Research 7(4), 1009-1026.

Ramos, V.A., Kay, S.M., Page, R.N. and Munizaga, F., 1989, La ignimbrita Vacas

Heladas y el cese del volcanismo en el Valle del Cura, Provincia de San Juan, Revista

Asociación Geológica Argentina 44, 336–352.

Ramos, V.A., Escayola, M., Mutti, D.I. and Vujovich, G.I., 2000, Proterozoic-early

Paleozoic ophiolites of the Andean basement of southern South America, in

Ophiolites and oceanic crust: New insights from field studies and the ocean drilling

program, Dilek., Y., Moores, E.M., Elthon, D. and Nicolas, A. (Eds.), Geological

Society, London, Special Publications 142, 181-217.

Ramos, V.A., Cristallini, E.O. and Pérez, D.J., 2002, The Pampean flat-slab of the central

Andes, J. South Am. Earth Sci. 15, 59-78.

Randall, G.E., 1994, Efficient calculation of complete differential seismograms for

laterally homogeneous earth models, Geophys. J. Int.118, 245-254. 109

Randall, G.E., Ammon, C.J. and Owens, T.J., 1995, Moment-tensor estimation using

regional seismograms from a Tibetan Plateau portable network deployment, Geophys.

Res. Letters 22, 1665-1668.

Rapela, C., 2000, The Sierras Pampeanas of Argentina: Paleozoic building of the

southern proto-Andes, in Tectonic evolution of South America, Cordani, U.G.,

Milani, E.J., Thomaz-Filho, A. and Campos, D.A. (Eds.), pp. 381-387.

Rapela, C.W., Pankhurst, R.J., Casquet, C., Baldo, E., Saavedra, J., Galindo, C. and

Fanning, C.M., 1998, The Pampean orogeny of the southern proto-Andes: Cambrian

continental collision in the Sierras de Córdoba, in The proto-Andean margin of

Gondwana, Pankhurst, R.J. and Rapela, C.W. (Eds.), Geological Society, London,

Special Publications 142, pp. 181-217.

Regnier, M., Chatelain, J.L., Smalley, R.(Jr.), Chiu, J.M., Isacks, B., and Araujo, M.,

1992, Seismotectonics of Sierra Pie de Palo, a basement block upflit in the Andean

foreland of Argentina, Bull. Seism. Soc. Am. 82, 2549-2571.

Regnier, M., Chatelain, J.L., Smalley, R.(Jr.), Chiu, J.M., Isacks, B. and Araujo, M.,

1994, Crustal thickness variation in the Andean foreland, Argentina from converted

waves, Bull. Seism. Soc. Am 84, 1097-1111.

Rivera, O. and Cembrano, J., 2000, Modelo de deformación de cuencas volcano

tectónicas in zonas de transferencia oblicua a la cadena andina: el caso de las cuencas

oligo-miocenas de Chile central y su relación con estructuras NWW-NW (33°00’-

34°30’S). IX Chilean Geological Congress, 2, 631-636, Puerto Varas, Chile. 110

Rutter, E.H. and Brodie, K., 1992, Rheology of the lower crust, in D.M. Fountain, R.J.

Arculus and R.W. Kay (eds.), The Continental Lower Crust, Elsevier, pp. 201–267.

Salazar, P., 2005, Análisis del campo de esfuerzos en la zona de subducción bajo Chile

Central (30º-34ºS). M.Sc. thesis, University of Chile, Santiago, Chile, p. 196 (in

Spanish).

Sato, A.M., Tickyj, H., Llambías, E.J., Stipp Basei, M.A. and González, P.D., 2004, Las

Matras block, central Argentina (37°S-67°W): the southernmost Cuyania terrane and

its relationship with the Famatinian orogeny, Gondwana Research 7(4), 1077-1087.

Smalley, R.(Jr.) and Introcaso, A., 2003, Estructura de la corteza y del manto superior en

el antepaís andino de San Juan (Argentina), Academia Nacional de Ciencias Exactas,

Físicas y Naturales (Argentina), 53, p. 13-25.

Snyder, D.B., Ramos, V.A. and Allmendinger, R.W., 1990, Thick-skinned deformation

observed on deep seismic reflection profiles in western Argentina, Tectonics 9, 773-

788.

Stern, C.R., 2004, Active Andean volcanism: its geologic and tectonic setting, Rev. Geol.

Chile 31(2), 161-206.

Thomas, W.A. and Astini, R., 2003, Ordovician accretion of the Argentine Precordillera

terrane to Gondwana: a review, J. South Am. Earth Sci. 16, 67-79.

Vujovich, G.I. and Kay, S.M., 1998, A Laurentian? Grenville-age oceanic arc/back-arc

terrane in the Sierra Pie de Palo, Western Sierras Pampeanas, Argentina, in The

proto-Andean margin of Gondwana, Pankhurst, R.J. and Rapela, C.W. (Eds.),

Geological Society, London, Special Publications 142, pp. 159-179. 111

Vujovich, G.I., van Staal, C.R., and Davis, B., 2004, Age constrains on the tectonic

evolution and provenance of the Pie de Palo complex, Cuyania composite terrane, and

the Famatinian orogeny in the Sierra Pie de Palo, San Juan, Argentina, Gondwana

Research 7(4), 1041-1056.

Wagner, L.S., Beck, S. and Zandt, G., 2005, Upper mantle structure in the south central

Chilean subduction zone (30 degrees to 36 degrees S), Journal of Geophysical

Research 110, doi:10.1029/2004JB003238.

Yañez, G.A., Cembrano, J., Pardo, M., Ranero, C. and Selles, D., 2002, The Challenger-

Juan Fernández-Maipo major tectonic transition of the Nazca-Andean subduction

system at 33-34ºS: Geodynamic evidence and implications, J. S. Am. Earth Sci. 15,

23-38.

Zapata, T.R., 1998, Crustal structure of the Andean thrust front at 30ºS latitude from

shallow and deep seismic reflection profiles, Argentina, J. S. Am. Earth Sci. 11, 131-

151.

Zapata, T.R. and Allmendinger, R.W., 1996, Thrust-front zone of the Precordillera,

Argentina: A thick-skinned triangle zone, AAPG Bulletin 3, 359 – 381.

112

Table and Figure Captions

Table 1. Grid search results for different earthquake-single CHARGE-station pairs used in this study (Figs. 1 and 2). The source parameters in the first column include year-

Julian day of occurrence, station name, magnitude (Mw) and focal depth from a previous regional moment tensor inversion analysis (Alvarado et al., 2005). Distance refers to the respective epicentral distance; Periods, to the corner periods used in the bandpass filtering; Correlation Coefficient, to the maximum average correlation between observed and synthetic data among the three-component waveforms for the best model; Thickness,

Vp and Vp/Vs, to the best layer-over-halfspace model that produces the best fits and hence the maximum correlation. The last three columns also show a range of crustal parameters in parentheses for a 5 % of less correlated data around the best model estimated simultaneously in the three components. A star after the station name indicates that only two seismic components were used.

Fig. 1. Location map of the region of study showing crustal earthquakes in the Andes cordillera and its backarc region during the CHARGE experiment, their focal mechanisms in lower hemisphere projection constrained from regional moment tensor inversion by Alvarado et al. (2005), the CHARGE seismic network, plate convergence velocity from Kendrick et al. (2003), slab contours in km from Cahill and Isacks (1992), sutures, terranes, tectonic provinces, Quaternary volcanism and mineralized areas based 113

on Ramos (2004), Ramos et al. (2002), Stern (2004), Kay and Mpodozis (2001) and

Bissig et al. (2002).

Fig. 2. Earthquake-station paths studied here with forward broadband waveform

modeling at near-regional distances (100 – 350 km). The best layer thickness results for

each event-station pair using a layer-over-halfspace model are shown along the raypaths

(see Table 1 for other crustal parameters). Tectonic provinces, terranes, sutures and

mineralization areas as in Fig. 1. The inset shows elevation contours every 1 km.

Fig. 3. Sensitivity of synthetic broadband waveforms to variations in crustal thickness

using a layer-over-halfspace model. For simplicity we show the radial component

computed for event 01-285 at a distance and azimuth of 310 km and 192°, respectively.

Fixed seismic parameters are crustal Vp (6.4 km/s), crustal Vp/Vs (1.80), mantle Vp

(8.15 km/s) and mantle Vp/Vs (1.80). a) Broadband seismograms. b) Long-period

seismograms using a bandpass filter between 10 and 80 s.

Fig. 4. Sensitivity of the broadband waveforms to average crustal P-wave velocity (Vp)

calculated at the same distance and azimuth as in Fig. 3. We assumed a 50-km thick

crustal layer of Vp/Vs of 1.80. Mantle-halfspace parameters fixed as in Fig. 3. a)

Broadband radial-component seismograms. b) Synthetic waveforms using a 10 - 80 s bandpass filter.

114

Fig. 5. Sensitivity of the broadband waveforms to crustal S-wave velocity (Vs) calculated

at the same distance and azimuth as in Fig. 3. We used a 50 km thick crust of Vp = 6.4

km/s. Mantle parameters fixed as in Fig. 3. a) Broadband radial-component seismograms. b) Synthetic waveforms using a 10 - 80 s bandpass filter.

Fig. 6. Sensitivity of the broadband waveforms to mantle Vp calculated at the same distance and azimuth as in Fig. 3. We used a layer of 50 km thickness of Vp = 6.4 km/s and Vp/Vs = 1.80 over a mantle-halfspace of Vp/Vs = 1.80. a) Broadband radial- component seismograms. b) Synthetic waveforms using a 10 - 80 s bandpass filter.

Fig. 7. Sensitivity of the broadband waveforms to mantle Vp/Vs calculated at the same distance and azimuth as in Fig. 3. We used the same crustal layer parameters as in Fig. 6.

Mantle-halfspace Vp was fixed at 8.15 km/s. a) Broadband radial-component seismograms. b) Synthetic waveforms using a 10 - 80 s bandpass filter.

Fig. 8. a) Example of a map of the average correlation coefficient over the three- component waveforms between synthetic and observed data for the event-station pair 02-

005/BARD (see Fig. 2 for location). Sixty combinations of crustal thickness and Vp were tested with forward waveform modeling for seven different crustal Vp/Vs. In addition mantle parameters Vp and Vp/Vs remained fixed during the entire search at 8.15 km/s and 1.80, respectively. In this case, a selected grid search using a fixed crustal Vp/Vs at

1.80 is shown. A bandpass filter between 10 and 50 s was used. The red dot indicates the 115

maximum correlation for the best layer-model consisting of a thickness of 35 km, a Vp of

6.2 km/s and a Vp/Vs of 1.80. b) Synthetic (red line) and observed (black line) seismic displacements for the best layer-model in figure (a) corresponding to a maximum average correlation coefficient of 0.93. The number near the waveforms indicates the correlation percentage for each seismic component. c) Waveform matches for a layer-model with thickness of 55 km, Vp of 6.2 km/s and Vp/Vs of 1.80, corresponding to an average correlation coefficient of 0.75 (a star). d) Waveform matches for a layer-model with thickness of 45 km, Vp of 6.4 km/s and Vp/Vs of 1.80 corresponding to an average correlation coefficient of 0.44 (a square).

Fig. 9. a) Example map of the maximum correlation coefficient over the three component waveforms for event (Mw 4.5) 01-352 and station PICH (Fig. 2). We used a bandpass filter between 10 and 50 s. The best layer-over-halfspace model (red dot) occurs for a Vp of 5.8 km/s, a thickness of 45 km and Vp/Vs of 1.70, with a maximum correlation coefficient of 0.94. b) Synthetic three-component forward-modeled waveforms for the

maximum correlation results of each grid-search analysis around crustal thickness (in

km) and crustal Vp (in km/s) using different crustal Vp/Vs. The average correlation

coefficients for each best layer-over-halfspace model resulting from the respective grid

search are indicated to the right. Numbers to the left above each trace indicate the

percentage of correlation between the corresponding forward synthetic seismogram (red)

and the observed data (black). Note that the best overall results occur for a model with

crustal thickness of 45 km, Vp of 5.8 km/s and crustal Vp/Vs of 1.70. 116

Fig. 10. Comparison of the grid search results along the high Cordillera. a) Map of the maximum correlation coefficient over the three component waveforms for event (Mw 5.1)

01-285 and station NEGR (Fig. 2). We used a bandpass filter between 10 and 80 s. The best layer-over-halfspace model (red dot) occurs for a Vp of 6.6 km/s, a thickness of 60 km and a Vp/Vs of 1.85, with a maximum correlation coefficient of 0.93. b) Synthetic

(red) three-component forward-modeled waveforms for the best combination of crustal parameters (red dot in Fig. a) compared to observed seismic displacements (black). The percentage of correlation for each component between synthetic and observed waveforms is shown to the left above the traces. c) Map of the maximum correlation coefficient over the three component waveforms for the same event using station LENA (Fig. 2). We used a bandpass filter between 15 and 50 s. The best layer-over-halfspace model (blue triangle) occurs for a Vp of 5.8 km/s, a thickness of 45 km and a Vp/Vs of 1.80, with a maximum correlation coefficient of 0.91. d) Synthetic (cyan) three-component forward- modeled waveforms for the best combination of crustal parameters (triangle Fig. c) compared to observed seismic displacements (black). The percentage of correlation for each component between synthetic and observed waveforms is shown to the left above the traces.

Fig. 11. a) Map of P-wave velocity results from the best-correlated waveforms in the grid search analysis (Table 1). Vp-contours represent interpolated Vp-data points, assigned every 1-degree to the studied seismic paths (Fig. 2). Slab contours are from Anderson et 117

al. (2006). Other symbols as in Fig. 1. b) Same as (a) for Vp/Vs ratio. Note that this map

shows variations of Vs determined from the best layer model along each studied path.

Fig. 12. a) Map of crustal thickness results from the best-correlated waveforms in the

grid search analysis (Table 1), using the same criteria to assign thickness-data points as in

Fig. 11, and estimations by McGlashan et al. (2006), Gilbert et al. (2006), Calkins et al.

(2006), Fromm et al. (2004) and Regnier et al. (1992) (Table 1). Other symbols as in Fig.

1. b) Moderate crustal seismicity of depths < 50 km from 1986 to 2006, reported by the

local networks in Argentina (INPRES) and Chile (SSN), as indicated in the NEIC

(USGS) and ISC catalogs. The histogram shows the seismic energy (in 106 J) released in the upper 70 km of the high Cordillera and backarc region from 1900-1963 (red) and

1963-1995 (pink), from Gutscher et al. (2000a). Slab contours from Anderson et al.

(2006) and terranes boundaries from Ramos et al. (2002).

Fig. 13. Vp/Vs as a function of Vp for common lithologies modified from Brocher

(2005). The mafic line delineates calcium-rich rocks, gabbros, and mafic rocks with Vs

0.2-0.3 km/s below the general trend of empirical fits. Note the separation between different elastic constants (dashed purple line). Ellipses encircle data provided by single references in Brocher (2005). Numbers in parentheses link specific lithologies: (9): sedimentary rocks; (10): Salinian granite; (11): metagraywackes and mafic rocks; (12): metamorphic rocks. Rectangles represent each region in this study: E: Pampia (Eastern

Sierras Pampeanas); S: South Volcanic Zone; H: High Cordillera north of 33ºS; W: 118

Cuyania (Western Sierras Pampeanas and Precordillera). See Brocher (2005) for complete references.

Fig. 14. Testing for a gradational velocity increase in the lower crust of the Cuyania terrane. (a) P- and S-wave velocities as a function of depth. Model_1 represents the best layer-over-halfspace model from our grid-search. Model_2 (RF) is the crustal model determined by Calkins et al. (2006) from local receiver functions recorded at station

JUAN (Fig. 1). Model_3 has a modification to Model_2 for the upper crustal Vs. (b)

Forward synthetic waveforms calculated for each crustal model in (a) compared with observed seismograms using the event/station pair 02-117/USPA (Fig. 2). Numbers to the left above each seismic component are the percentage of cross-correlation between synthetic and observed data. The averaged correlation coefficient over the three seismic components using each crustal model is indicated below the waveform matches.

119

Table 1

Event/Station Distance Periods Correlation Thickness Vp Vp/Vs Mw/Depth(km) (km) (s) Coefficient (km) (km/s) 01-127/PACH 314 12 – 30 0.86 35 (35 – 43) 6.4 (6.2 – 6.4) 1.85 (1.80 – 1.85) 4.1/14 12 – 30 0.75 55 – 60 6.2 – 6.4 1.80 – 1.85 01-138/HUER 174 15 – 30 0.71 30 (20 – 35) 6.0 (5.9 – 6.0) 1.70 (1.65 – 1.70) 4.1/5 01-138/LLAN 167 15 – 30 0.95 25 5.9 – 6.0 1.60 4.3/7 15 – 30 0.95 40 (35 – 40) 6.4 (6.4 – 6.6) 1.65 (1.60 – 1.65) 01-285/NEGR 330 10 – 80 0.93 60 (50 – 64) 6.6 (6.4 – 6.6) 1.85 (1.80-1.85) 5.1/5 01-285/MAUL 310 10 – 50 0.68 45 (45 – 50) 5.8 (5.8 – 6.0) 1.85 (1.80 – 1.85) 01-285/LENA 214 15 – 50 0.90 45 (42 – 48) 5.8 (5.6 – 5.8) 1.80 (1.75 – 1.80) 01-285/PENA 349 15 – 50 0.94 45(43 – 46) 6.4 (6.4 – 6.6) 1.80 (1.80 – 1.85)

0.92 55 – 58 6.6 1.85 01-285/BARD* 276 10 – 50 0.79 45 (40 – 47) 6.2 (6.1 – 6.2) 1.80 01-285/PACH* 355 15 – 60 0.91 45 (45 – 50) 6.6 (6.5 – 6.8) 1.90 (1.85 – 1.90) 15 – 60 0.90 60 (55 – 60) 6.8 1.90 01-316/PENA 350 15 – 30 0.96 35 (32 – 37) 6.0 (5.9 – 6.1) 1.80 (1.80 – 1.85) 4.1/13 01- 175 15 – 50 0.89 35 (23 – 37) 6.4 (6.2 – 6.4) 1.85 (1.85 – 1.90) 348/RINC(*) 15 – 50 0.88 43 (42 – 44) 6.2 (6.2 – 6.4) 1.85 (1.85 – 1.90) 01-349/LLAN 121 15 – 30 0.93 30 (27 – 33) 5.7 – 5.8 1.75 4.0/20 01-349/PICH 237 15 – 30 0.95 30 (27 – 33) 5.8 1.70 (1.70 – 1.75) 01-352/LLAN 10 – 50 0.93 35 (33 – 37) 6.0 (6.0 – 6.2) 1.70 (1.70 – 1.75) 4.5/17 01-352/PICH 208 10 – 50 0.94 45 (40 – 45) 5.8 (5.8 – 5.9) 1.70 01-352/JUAN 142 15 – 50 0.94 35 (25 – 30) 6.2 – 6.5 1.80 – 1.85 0.93 (55 - 62) 6.4 – 6.6 1.80 00-356/USPA 308 10 – 50 0.93 50 6.4 1.80 4.0/14 0.88 60 (55 - 63) 6.2 1.80 02-005/PENA 321 10 – 50 0.90 40 (40 – 50) 6.2 (6.2 – 6.4) 1.80 (1.80 – 1.85) 02-005/BARD 286 10 – 50 0.93 35 (30 – 37) 6.2 (6.2 – 6.4) 1.80 (1.80 – 1.85) 0.77 50 – 55 6.2 – 6.4 1.80 – 1.85 02-074/PICH 15 – 30 0.94 30 (30 – 35) 6.2 (6.1 – 6.2) 1.70 02-107/JUAN 148 15 – 40 0.95 40 (35 – 40) 6.2 (6.0 – 6.2) 1.80 (1.80 – 1.85) 4.2/25 02-107/USPA 259 15 – 50 0.94 40 (38 – 53) 6.4 1.80 (1.80 – 1.85) 02-117/USPA 217 15 – 50 0.97 50 (45 – 50) 6.2 1.80 (1.75 – 1.80) 5.1/21 0.96 40 6.4 1.75 – 1.80 0.96 55 – 60 6.4 – 6.6 1.75 – 1.80 02-117/HEDI 15 – 50 0.92 35 (35 – 39) 6.0 (5.8 – 6.2) 1.85 120

Figures

Figure 1

121

Figure 2

122

Figure 3 123

Figure 4 124

Figure 5 125

Figure 6 126

Figure 7 127

Figure 8 128

Figure 9

129

Figure 10

130

Figure 11 131

Figure 12 132

Figure 13

133

Figure 14