Chemical and Isotopic Studies of Monogenetic Volcanic Fields: Implications for Petrogenesis and Mantle Source Heterogeneity

Home , Burton (crater), Ceboruco, Dacite, Lambert (Martian crater), Potrillo volcanic field, Trachybasalt

MIAMI UNIVERSITY The Graduate School

Certificate for Approving the Dissertation

We hereby approve the Dissertation

Christine Rasoazanamparany

Candidate for the Degree

DOCTOR OF PHILOSOPHY

______Elisabeth Widom, Director

______William K. Hart, Reader

______Mike R. Brudzinski, Reader

______Marie-Noelle Guilbaud, Reader

______Hong Wang, Graduate School Representative

ABSTRACT

CHEMICAL AND ISOTOPIC STUDIES OF MONOGENETIC VOLCANIC FIELDS: IMPLICATIONS FOR PETROGENESIS AND MANTLE SOURCE HETEROGENEITY

by Christine Rasoazanamparany

The primary goal of this dissertation was to investigate the petrogenetic processes operating in young, monogenetic volcanic systems in diverse tectonic settings, through detailed field studies, elemental analysis, and Sr-Nd-Pb-Hf-Os-O isotopic compositions. The targeted study areas include the Lunar Crater Volcanic Field, Nevada, an area of relatively recent volcanism within the Basin and Range province; and the Michoacán and Sierra Chichinautzin Volcanic Fields in the Trans-Mexican Volcanic Belt, which are linked to modern subduction. In these studies, key questions include (1) the role of crustal assimilation vs. mantle source enrichment in producing chemical and isotopic heterogeneity in the eruptive products, (2) the origin of the mantle heterogeneity, and (3) the cause of spatial-temporal variability in the sources of magmatism. In all three studies it was shown that there is significant compositional variability within individual volcanoes and/or across the volcanic field that cannot be attributed to assimilation of crust during magmatic differentiation, but instead is attributed to mantle source heterogeneity. In the first study, which focused on the Lunar Crater Volcanic Field, it was further shown that the mantle heterogeneity is formed by ancient crustal recycling plus contribution from hydrous fluid related to subsequent subduction. The second study focused on the 1759-1774 eruption of Jorullo volcano in the Michoacán-Guanajuato Volcanic Field. There the observed temporal-compositional variation is attributed to a combination of a complex magmatic plumbing system that leads to variable degrees of magmatic fraction, along with concomitant changes in the mantle source over time. The source heterogeneity was produced by a single mantle source fluxed by two distinct subduction components dominated by terrigenous sediment-derived fluid. The third study focused on the origin of closely associated low-Nb, calc-alkaline and high-Nb, alkaline/transitional magmas within the Sierra Chichinautzin Volcanic Field. The low- and high-Nb magmas were shown to be genetically related, with the former produced primarily by addition of sediment-derived hydrous fluid, and the high-Nb magmas generated by melting of pyroxenite-veined mantle formed by sediment melt-mantle reaction. These studies demonstrate the importance of detailed field-based and geochemical analysis of small, monogenetic eruptions for revealing processes of mantle metasomatism and crust-mantle evolution.

CHEMICAL AND ISOTOPIC STUDIES OF MONOGENETIC VOLCANIC FIELDS: IMPLICATIONS FOR PETROGENESIS AND MANTLE SOURCE HETEROGENEITY

A DISSERTATION

Presented to the Faculty of

Miami University in partial

fulfillment of the requirements

for the degree of

Doctor of Philosophy

Department of Geology & Environmental Earth Science

Christine Rasoazanamparany

The Graduate School Miami University Oxford, Ohio

2015

Dissertation Director: Elisabeth Widom, Ph.D.

TABLE OF CONTENTS

LIST OF TABLES v

LIST OF FIGURES vi

ACKNOWLEDGMENT viii

CHAPTER 1: INTRODUCTION Introduction 1 References 6

CHAPTER 2: ORIGIN OF CHEMICAL AND ISOTOPIC HETEROGENEITY IN A MAFIC, MONOGENETIC VOLCANIC FIELD: A CASE STUDY OF THE LUNAR CRATER VOLCANIC FIELD, NEVADA Abstract 9 Body Text 11 References 32 Appendix 72

CHAPTER 3: TEMPORAL AND COMPOSITIONAL EVOLUTION OF JORULLO VOLCANO, MEXICO: IMPLICATIONS FOR MAGMATIC PROCESSES ASSOCIATED WITH A MONOGENETIC ERUPTION Abstract 76 Body Text 77 References 100 Appendix 139

CHAPTER 4: PETROGENESIS OF MAFIC MAGMAS IN THE SIERRA CHICHINAUTZIN VOLCENIC FIELD, MEXICO: CONSTRAINTS FROM OSMIUM ISOTOPE SYSTEMATICS Abstract 150

iii Body text 151 References 171 Appendix 213

CHAPTER 5: SUMMARY Summary 214 References 219

iv LIST OF TABLES

CHAPTER 2: ORIGIN OF CHEMICAL AND ISOTOPIC HETEROGENEITY IN A MAFIC, MONOGENETIC VOLCANIC FIELD: A CASE STUDY OF THE LUNAR CRATER VOLCANIC FIELD, NEVADA

1. Whole rock analyses of the northern LCVF 48

CHAPTER 3: TEMPORAL AND COMPOSITIONAL EVOLUTION OF JORULLO VOLCANO, MEXICO: IMPLICATIONS FOR MAGMATIC PROCESSES ASSOCIATED WITH A MONOGENETIC ERUPTION

1. Major and Trace Element and Isotopic Data 109

CHAPTER 4: PETROGENESIS OF MAFIC MAGMAS IN THE SIERRA CHICHINAUTZIN VOLCENIC FIELD, MEXICO: CONSTRAINTS FROM OSMIUM ISOTOPE SYSTEMATICS

1. Whole Rock Major and Trace Element and Isotopic Data SCVF 183

v LIST OF FIGURES

CHAPTER 2: ORIGIN OF CHEMICAL AND ISOTOPIC HETEROGENEITY IN A MAFIC, MONOGENETIC VOLCANIC FIELD: A CASE STUDY OF THE LUNAR CRATER VOLCANIC FIELD, NEVADA

1. Regional Map of the Lunar Crater Volcanic Field 52 2. Major and Trace Element Variation Diagrams 54 3. Primitive Mantle Normalized Trace Element Diagrams 56

4. Strontium-Nd, ƐNd-ƐHf and Sr-Pb Isotopes Diagrams 58 5. Strontium-Nd-Pb Isotope Diagrams 60 6. Osmium Concentrations vs. Os Isotope Diagram, AFC Model 62 7. Osmium Isotope versus Nb/U Diagram 64 8. Lead-Sr and Pb-Os Isotope Diagrams, Mixing Models 66 9. Trace Element Modeling Diagrams 68 10. Cartoon Diagram of Petrogenetic Model 70

CHAPTER 3: TEMPORAL AND COMPOSITIONAL EVOLUTION OF JORULLO VOLCANO, MEXICO: IMPLICATIONS FOR MAGMATIC PROCESSES ASSOCIATED WITH A MONOGENETIC ERUPTION

1. Regional Tectonic Map 113 2. Geologic Map of Jorullo 115 3. MgO versus Eruptive Phase and Relative Stratigraphic Height 117 4. Major Element versus MgO Diagrams 119 5. Trace Element versus MgO Diagrams 121 6. N-MORB Normalized Trace Element Diagrams 123 7. Isotope Variation Diagrams 125 8. Neodymium-Sr and Pb-Sr Isotope Diagrams 127 9. Osmium Isotope versus Ni Diagram 129 10. Strontium-Pb and Nd-Hf Isotope Diagrams, Mixing Models 131

vi 11. Isotopic Models 133 12. Trace Element Models 135 13. Cartoon Diagram of Petrogenetic Model 137

CHAPTER 4: PETROGENESIS OF MAFIC MAGMAS IN THE SIERRA CHICHINAUTZIN VOLCENIC FIELD, MEXICO: CONSTRAINTS FROM OSMIUM ISOTOPE SYSTEMATICS

1. Tectonic Map of the Trans-Mexican Volcanic Belt 185 2. Geologic Map of SCVF 187 3. Major Element versus MgO Diagrams 189 4. Trace Element versus MgO Diagrams 191 5. Trace Element Variation Diagrams 193 6. N-MORB Normalized Trace Element Diagrams 195 7. Isotope Variation Diagrams 197 8. Osmium Isotope versus MgO, Ni, and Os Diagrams 199 9. Osmium-Pb-Sr Isotope versus Ce/Pb and Ba/Zr Diagrams 201 10. Osmium-Pb Isotope Crustal Assimilation Model 203 11. Oxygen Isotope versus Ni 205 12. FC3MS versus MgO Diagram 207 13. Osmium-Pb Isotope Source Mixing Model 209 14. Cartoon Diagram of Petrogenetic Model 211

vii ACKNOWLEDGEMENTS

First and foremost, I would like to express my deep sense of gratitude and appreciation to my advisor Dr. Elisabeth Widom for her supervision with invaluable help and advice during my PhD program. I learned a lot from her not only in academia area but in other spheres of life also. I acknowledge the time and energy that she has poured out on my behalf from the very beginning of this study up to the completion of the thesis. Dr. Widom has been an important part of my professional development; and she helped me grow as real scientist.

I would like to thank all my committee members: Drs. William K. Hart, Mike Brudzinski, Marie-Noelle Guilbaud and Hong Yang for their valuable input. I would also like to thank the other member of my oral defense committee Dr. Hailiang Dong for his time and insightful questions. This dissertation could not have been completed without the help and contributions from collaborators. I would like to thank Drs. Greg Valentine, Eugene Smith, Claus Siebe, Joaquim Cortés, Mike Spicuzza, John Valley, Sergio Salinas and Marie-Noelle Guilbaud.

I would also like to thank Dave Kuentz for his assistance in lab and helped train me on the TIMS. I would also like to thank John Morton for his assistance in the lab. I would also like to thank all of the members of Dr. Elisabeth Widom and William Hart research group for all of the scientific discussions, humors and entertainment during my stay at Miami University. I would like to thank my family for their love, support and encouragement during the final stages of this PhD. Most of all, I thank the almighty God who has given me the strength and courage to carry out this work. Thank you for letting me through all the difficulties. Thank you Lord.

This work was supported by the National Science Foundation (NSF EAR#1016042 for Lunar Crater Volcanic Field and EAR#1019798 for Mexico) awarded to Elisabeth Widom, and the Geological Society of America awarded to Christine Rasoazanamparany.

viii CHAPTER 1

Introduction

Small-volume eruptions (<1 km3) produced by single episodes of volcanic activity, without subsequent eruptions, are referred to as monogenetic volcanoes (Connor and Conway, 2000). Monogenetic volcanoes often occur in high concentrations, comprising monogenetic volcanic fields. Monogenetic volcanoes commonly occur as cinder cones, maars, tuff cones tuff rings, shield volcanoes and lava domes (Valentine and Perry, 2007). The duration of an individual eruption within the field is typically shorter than that of composite volcanoes (e.g. stratovolcanoes), in the orders of days to years (Connor and Conway, 2000), however the volcanic field as a whole may span millions of years. Although monogenetic volcanoes are often thought of as posing less of a hazard than stratovolcanoes, they exhibit a wide range of eruptive styles from relatively non- explosive (e.g. effusion of lava flows) to highly explosive (e.g. violent strombolian), and as such, they can pose significant hazards. The hazards associated with monogenetic volcanism are generally localized. However, because of their occurrence in some heavily populated urban areas, such as Auckland, New Zealand or Mexico City, Mexico, their associated volcanic risks can be very high. In addition, because the location of volcanic activity within the volcanic field tends to shift over time unpredictably, they may pose hazards to important infrastructure (Connor et al., 2009). Numerous studies have focused on the temporal and spatial distributions of monogenetic volcanoes, and the probability of future activity within a monogenetic volcanic field, with the aim of forecasting potential future volcanic hazards (e.g. Connor et al., 2000, Valentine and Perry, 2006). Recent studies have shown furthermore that the eruptive style and scale of individual volcanoes could be related to the characteristics of their mantle source (Valentine and Perry, 2007). Constraining the magmatic processes operating in volcanic fields is therefore important. From a societal perspective, understanding the link between magmatic processes and eruption style is crucial for predicting potential volcanic hazards and for risk assessment and management.

Monogenetic volcanic fields occur in every known tectonic setting, and very often produce near primitive basaltic magmas (i.e. Ni > 200 ppm and Cr > 500 ppm) that suggest rapid ascent from the mantle source to the surface (Rutherford, 2008). Because of their small volume and primitive character, they are considered to ascend more or less directly from their mantle source with little or no interaction with the wall-rock. Thus, they have the potential to reveal information about sources and processes of magma generation as well as the magmatic processes occurring during ascent and emplacement of mantle-derived magmas. As such, mafic monogenetic volcanic systems can improve our understanding of the processes involved in the origin of basaltic volcanism in general. Monogenetic volcanism is often characterized by significant compositional variations within individual eruptions and within given volcanic fields (Reiners, 2002; Siebe et al., 2004; Brenna et al., 2012; Cousens et al., 2013; Gazel et al., 2012; Rasoazanamparany et al., 2014). Ongoing debate regarding the origin of this compositional variation focuses mostly on the relative roles of (1) assimilation of lithospheric mantle or continental crust during ascent and emplacement and (2) heterogeneity in the mantle source due to mantle metasomatic processes. The latter involves infiltration of the mantle and reaction with small volume melts or fluids, whether from modern or ancient subduction systems, or in some cases, small volume asthenospheric or plume-derived melts. A related issue is the question of whether monogenetic systems develop sustained magma chambers, thus facilitating crustal contamination in the crust, or whether magmas erupt more-or-less directly from their mantle sources regions. Other key questions regarding the petrogenesis in monogenetic volcanic systems include: (1) the nature and scale of the heterogeneous mantle source and (2) the role of pyroxenite veins in producing the heterogeneity. Melting of pyroxenitic and pyroxenite/peridotite hybrid mantle has been invoked to explain chemical and isotopic heterogeneity in both intraplate and subduction-related mafic monogentic eruptions, and also to explain the observation that monogentic eruptions often become progressively more silicic with time (Reiners, 2002; Straub et al., 2012). Such temporal-compositional relationships could be explained in some cases by variable degrees of fractional crystallization of a single parental magma, but in many cases isotopic data indicate that

2 the compositional changes may be linked to compositionally distinct mantle sources feeding the magmatism. Furthermore, questions remain about the origin of closely linked calc-alkaline and alkaline/transitional magmas, which can occur together in a single volcanic field, and sometimes within a single eruption. The ultimate objective of this dissertation is to address these fundamental questions in order to better understand the petrogenetic processes that govern the evolution of monogenetic volcanic systems. To this end, we have combined detailed field studies of monogenetic volcanic regions with geochemical and isotopic analysis, including whole-rock major and trace element abundances and Sr, Nd, Pb, Hf, Os and mineral O isotopic compositions, to evaluate the processes involved in magma genesis and evolution. In the second chapter of this dissertation, we investigate the Lunar Crater Volcanic Field located in central Nevada, USA, as a case study to investigate (1) the compositional variations within individual, closely spaced monogenetic centers as well as the encompassing volcanic field, (2) the relative importance of crustal assimilation and the role of mantle heterogeneity including the potential role of crustal recycling in producing mantle heterogeneity and (3) the spatial and temporal variability of the sources of magmatism in the monogenetic volcanic field. In this project, we focus on the most recent volcanism in the northern part of the volcanic field. The LCVF is one of the Quaternary monogenetic volcanic fields within the Basin and Range extensional setting, and contains more than 100 volcanic centers. Quaternary volcanism within the Basin and Range is mostly concentrated along the margins, but the LCVF is isolated in the central part of the basin. Numerous studies have indicated that extension plays an important role in magma generation within the Basin and Range province (e.g. Farmer et al., 1989) but it not clear why extension would focus most of the volcanism around the margins, rather than throughout the interior of the Basin and Range. The potential role of enriched and/or wet mantle in generating this volcanism is not well understood. This study area provides an ideal opportunity to address the petrogenesis of monogenetic volcanoes and the scale of heterogeneity beneath the mantle source region, as there are numerous young, well- exposed mafic lava flows, tephra fall deposits and mantle xenoliths that can provide key information related to magma genesis. Chapter 2 has already been published in Chemical Geology as Rasoazanamparany et al. (2014), “Origin of chemical and isotopic

3 heterogeneity in a mafic, monogenetic volcanic field: A case study of the Lunar Crater Volcanic Field, Nevada”, Chemical Geology 397, 76–93. The study of the petrogenesis of Jorullo volcano and its satellites cones in the third chapter provides a unique opportunity to evaluate the origin of temporal-compositional variation within individual monogenetic centers, as well as magmatic processes operating in subduction systems in general. Jorullo volcano erupted from 1759 to 1774 and is located within the Michoacán Guanajuato Volcanic Field (MGVF), which represents the volcanic front of the Trans-Mexican Volcanic Belt (TMVB). The latter is believed to represent the volcanic arc associated with the subduction of Cocos and Riviera plates beneath North American plate along the Middle American Trench (MAT). Because of its location within the volcanic front, Jorullo volcano and its satellites cones represent a key area to study the relationship between central Mexican volcanism and subduction of the Cocos plate. We use the elemental and isotopic data (Sr-Nd-Pb-Hf-Os and oxygen isotopic data) of Jorullo lavas and tephra combined with field observations to address the following aspects of the petrogenesis: 1) the influence of crustal assimilation; 2) the types and nature of the subduction components being added to the mantle wedge (sediment versus oceanic crust; fluid versus melt); and 3) the cause of the systematic compositional variations with time in the Jorullo eruption. Results from studies of individual eruptive centers can be used to evaluate the processes of magma generation on a volcanic field scale, with implications for monogenetic volcanism on a global scale. Chapter 3 comprises a paper almost ready to submit for publication to Chemical Geology. The petrogenesis of mafic magmas in the Sierra Chichinautzin Volcanic Field (SCVF) forms the fourth chapter of the dissertation. The SCVF is located on the southern boundary of the TMVB and comprises over 220 Quaternary volcanoes. Most volcanic activity in the SCVF is relatively young <40 ka (Siebe et al., 2004a), suggesting that the volcanic field is highly active. Because of its location close to the densely populated cities of Mexico, understanding the eruptive style and frequency of eruptions in the area is critical. In addition, this study area represents a key location to document the role of subduction zone processes in the origin of the Mexican volcanism, and to investigate the cause of the unusual co-occurrence of low-Nb, calc-alkaline and high-Nb, alkaline/transitional magmas. The origin of the two compositionally distinct magmas has

4 been the subject of extensive petrological and geochemical studies (e.g. Luhr 1997, Wallace and Carmichael, 1999; Straub et al., 2011,2012, 2015). Although the calc- alkaline and intraplate alkaline magmas erupt in close proximity and sometimes from the same volcanic vent, it has been suggested that they come from distinct mantle domains (Luhr, 1997), in which the intra-plate alkaline basalts are produced by partial melting of a convecting upper mantle uncontaminated by subduction components, and the sub- alkaline rocks derived by melting of sub-arc mantle wedge that has been metasomatized by slab-derived materials (Luhr 1997, Wallace and Carmichael, 1999; Luhr et al. 2006; Cai et al., 2014). Other studies have suggested that alkaline and calc-alkaline series magmas are derived from silica-poor and silica rich pyroxenitic sources, respectively (Straub et al., 2013). In this study we use major and trace element data and Sr-Nd-Pb-Hf- Os and oxygen isotopic data to evaluate the relative importance of lithospheric contamination vs. source-related enrichment in generating the compositional variations observed within the volcanic field as well as the genetic relationship between low-Nb and high-Nb magmas within the volcanic field.

References

Brenna, M., Cronin, S.J., Smith, I.E.M., Maas, R., Sohn, Y.K., 2012. How small-volume basaltic magmatic systems develop: a case study from the Jeju Island Volcanic Field, Korea. Journal of Petrology 53, 985-1018.

Connor, C.B., Conway, F.M., 2000. Basaltic volcanic fields. In: Sigurdsson, H. (Ed.), Encyclo-pedia of Volcanoes. Academic Press, New York, pp. 331–343.

Connor, C.B., Stamatakos, J.A., Ferrill, D.A., Hill, B.E., Ofoegbu, G.I., Conway, F.M., 2000. Volcanic hazards at the proposed Yucca Mountain, Nevada, high-level radioactive waste repository. Journal of Geophysical Research, 105, 417-432.

Cousens, B., Wetmore, S., Henry, C.D., 2013. The Pliocene-Quaternary Buffalo Valley volcanic Field, Nevada: Post-extensional, intraplate magmatism in the north-central Great Basin, USA. Journal of Volcanology and Geothermal Research 268, 17-35.

Farmer, G.L., Perry, F.V., Semken, S., Crowe, B., Curtis, D., DePaolo, D.J., 1989. Isotopic evidence on the structure of origin of subcontinental lithosphere mantle in southern Nevada. Journal of Geophysical Research 94(B6), 7885-7898.

Gazel, E., Plank, T., Forsyth, D.W., Bendersky, C., Lee, C-T.A., and Hauri, E.H., 2012. Lithospheric versus asthenospheric mantle sources at the Big Pine Volcanic Field, California. Geochemistry Geophysics Geosystems, 13, 25 pp. doi:10.1029/2012GC004060.

Rasoazanamparany, C., Widom, E., Valentine, G.A., Smith, E.I., Cortés, J.A., Kuentz, D., Johnsen, R., 2014. Origin of chemical and isotopic heterogeneity in a mafic, monogenetic volcanic field: A case study of the Lunar Crater Volcanic Field, Nevada. Chemical Geology 397, 18 March 2015, Pages 76–93

Reiners, P.W., 2002. Temporal-compositional trends in intraplate basalt eruptions: Implications for mantle heterogeneity and melting processes. Geochemistry Geophysics Geosystems 3, 30 pp. doi:10.1029/2001GC000250.

Siebe, C., Rodriguez-Lara V., Schaaf, P., Abrams, M., 2004a. Radiocarbon ages of Holocene Pelado, Guespalapa and Chichinautzin scoria cones, south of Mexico City: implications for archaeology and future hazards. Bull. Volcanol. 66 203-225.

Straub, S.M., Gómez-Tuena, A., Zellmer, G.F., Espinasa-Pereña, R., Stuart, F.M., Langmuir, C.H., Martin-del Pozzo, A.L., Mesko, G.T., 2012. The Processes of Melt Differentiation in Arc Volcanic Rock: Insights from OIB-type Arc Magmas in the Central Mexican Volcanic Belt. J. Petrol. 0, 1-42.

Straub, S.M., Gómez-Tuena, A., Stuart, F.M., Zellmer, G.F., Espinasa-Pereña, R., Cai, Y., and Iizuka, Y., 2011, Formation of hybrid arc andesites beneath thick continental crust. Earth Planet. Sci. Lett. 303, 337–347.

Straub, S.M., Gómez-Tuena, A., Zellmer, G.F., Espinasa-Pereña, R., Stuart, F.M., Cai, Y., Langmuir, C.H., Martin-Del Pozzo, A., Mesko, G.T., 2013a. The processes of melt differentiation in arc volcanic rocks; insights from OIB-type arc magmas in the central Mexican Volcanic Belt. Journal of Petrology 54, 665-701.

Valentine, G.A., Perry, F.V., 2007. Tectonically controlled, time-predictable basaltic volcanism from a lithospheric mantle source (central Basin and Range Province, USA). Earth and Planetary Science Letters 261, 201-216.

Wallace, and P.J., Carmichael, I.S.E., 1999. Quaternary volcanism near the Valley of Mexico: implications for subduction zone magmatism and the effects of crustal thickness variations on primitive magma compositions. Contributions to Mineral Petrololgy, 135: 291-314.

CHAPTER 2

Origin of Chemical and Isotopic Heterogeneity in a Mafic, Monogenetic Volcanic Field: A Case Study of the Lunar Crater Volcanic Field, Nevada

This paper has been published in the Elsevier Journal Chemical Geology as: Rasoazanamparany, C., Widom, E., Valentine, G.A., Smith, E.I., Cortés, J.A., Kuentz, D., Johnsen, R., 2015. Origin of chemical and isotopic heterogeneity in a mafic, monogenetic volcanic field: A case study of the Lunar Crater Volcanic Field, Nevada. Chemical Geology 397, Pages 76–93. doi:10.1016/j.chemgeo.2015.01.004

Abstract:

Major and trace element geochemistry and Sr, Nd, Pb, Hf and Os isotope signatures of basaltic lavas and tephra from volcanic centers in the northern Lunar Crater Volcanic Field (LCVF), Nevada, provide insight into the nature of their mantle sources and the role of lithospheric contamination versus source-related enrichment in producing compositional variations in basaltic monogenetic volcanic fields. Three of the studied eruptive centers (Hi Desert and Mizpah, ~620-740 ka; and Giggle Springs, <80 ka) are located within ~500 m of each other; the Marcath volcano (~35-38 ka) and Easy Chair (140 ka), two of the youngest eruptive centers in the field, are located ~6 and 12 km southwest of these cones, respectively. Isotopic studies of the volcanic rocks show a limited range in 143Nd/144Nd and 176Hf/177Hf, but significant heterogeneity in 87Sr/86Sr, 206Pb/204Pb and 187Os/188Os. The older (>140 ka) Hi Desert, Mizpah, proto-Easy Chair and several unnamed flows exhibit Nb-Ta enrichment, Rb, Cs and K depletion, and high 206Pb/204Pb but low 87Sr/86Sr. In contrast, the younger (<140 ka) Giggle Springs, Easy Chair and Marcath lavas have high Ba, Rb and Cs and lower 206Pb/204Pb and higher 87Sr/86Sr. The lavas produce a well-defined negative correlation between Sr and Pb isotopes, attributed to mixing of heterogeneous mantle sources. The geochemical and isotopic signatures of the older Hi Desert, Mizpah, proto-Easy Chair and unnamed lavas are consistent with derivation from a mantle source with a component of ancient recycled oceanic crust. In contrast, the relatively high Ba, Rb and Cs coupled with lower 206Pb/204Pb and higher 87Sr/86Sr of the younger Giggle Springs, Easy Chair and Marcath lavas are consistent with derivation from a similar, but fluid-enriched, mantle source. Mixing calculations indicate that incorporation of ~18% of 0.8 Ga recycled oceanic crust into depleted mantle can explain the trace element and isotopic signatures of the older group end member. Subsequent addition to this source of minor (<1%) hydrous fluid derived from subducted oceanic crust could account for the chemical and isotopic compositions of the younger group end member. Variable degrees of mixing between these two mantle end members can generate the full range of isotopic compositions observed in the northern LCVF sample suite, as well as within single eruptions. Our data indicate that the mantle source region in the LCVF is characterized by chemical and isotopic heterogeneity that manifests itself over a very small spatial scale (< 500 m) and

9 within the time frame of a single monogenetic eruption. Similar processes may explain the geochemical and isotopic heterogeneities observed in other mafic monogenetic volcanic fields, the evidence for which may be preferentially preserved where small degrees of melting and rapid source to surface transport prevail.

Keywords: monogenetic volcanism, Lunar Crater Volcanic Field, mantle heterogeneity, radiogenic isotopes

1. Introduction

Monogenetic volcanic fields consist of scattered small volcanoes that can occur as scoria cones, tuff cones or maars, depending on their eruptive style (Connor and Conway, 2000; Martin and Németh, 2006; Valentine et al., 2006; Valentine and Perry, 2007; Valentine and Gregg, 2008), and occur in a variety of tectonic settings, including extensional environments (e.g. Basin and Range) and subduction zones (e.g. Trans- Mexican Volcanic Belt). Such volcanic fields represent a common expression of continental volcanism, and can be active over periods of thousands to millions of years despite the fact that individual eruptions lasts only days to tens of years. Several studies have demonstrated that individual or closely spaced monogenetic eruptions can show significant compositional variation (e.g. Johnson et al., 2008; Luhr and Carmichael, 1985; Reiners, 2002). However, the origin of the chemical and isotopic heterogeneity remains controversial, and has been variously attributed to shallow level crustal assimilation (Chesley et al., 2002; Lassiter and Luhr, 2001; Righter et al., 2002), mantle source heterogeneity (Blondes et al., 2008; Reiners, 2002; Siebe et al., 2004; Smith et al., 1999), or some combination of these two processes (Borg et al., 2000). For example, studies of monogenetic volcanic fields including the Sierra Chichinautzin Volcanic Field (SCVF), Mexico and the Big Pine Volcanic Field (BPVF), California have attributed compositional variation to heterogeneous mantle sources (Gazel et al., 2012; Siebe et al., 2004). In contrast, Chesley et al. (2002) and Lassiter and Luhr (2001) concluded that the compositional variations observed in the Michoacán-Guanajuato Volcanic Field and the western Mexican Volcanic belt resulted from crustal assimilation. It also remains unclear whether monogenetic volcanoes develop prolonged magma storage zones in the crust, thus facilitating crustal contamination, or whether the magmas ascend more or less directly from their respective mantle source regions. A prior study of the Southwestern Nevada Volcanic Field (SNVF) suggested that monogenetic magmas ascended quickly through the crust (Valentine and Perry, 2007), but it is not clear if this is unique to that field. Furthermore, questions remain about the nature and scale of the potentially heterogeneous mantle source regions, and the role of pyroxenite veins in producing the heterogeneity. A number of studies have noted that the upper mantle may contain a large fraction of pyroxenite-rich veins (Hirschmann and Stolper, 1996; Mukasa et al., 1991;

Pearson et al., 1993), which could play an important role in the origin of continental alkali-basalts (Allègre and Turcotte, 1986; Carlson and Nowell, 2001; Hirschmann and Stolper, 1996; Leeman and Harry, 1993; Reiners, 2002; Prinzhofer et al., 1989; Zindler et al., 1984). In order to better understand the evolution of mafic monogenetic volcanic systems, it is important to perform detailed studies of both individual monogenetic centers as well as the encompassing volcanic fields, and to constrain the spatial and temporal extents of the compositional variations in addition to the nature of the chemical and isotopic signatures. This study focuses on the Lunar Crater Volcanic Field (LCVF) in central Nevada, one of the Quaternary monogenetic volcanic fields within the Basin and Range extensional setting, and is part of a larger collaborative project aimed at understanding the spatial and temporal scales of variations in magma source, evolution, ascent, and eruptive processes, and the links between these (e.g., Cortés et al., 2015). In the present study, the chemical and isotopic compositions (Sr-Nd-Pb-Hf-Os) of suites of volcanic rocks from individual eruptive centers (Marcath and Easy Chair), as well as samples from a closely-spaced cluster of volcanic centers are used to investigate the relative importance of (1) upper and/or lower crustal assimilation, (2) the role of mantle heterogeneity, including the potential role of crustal recycling in producing mantle heterogeneity, and (3) the spatial and temporal variability of the sources of magmatism in the monogenetic volcanic field.

2. Geologic Setting of the Lunar Crater Volcanic Field

The Lunar Crater Volcanic Field, central Nevada, covers an area of ~1,000 km2 (Bergman 1982) and contains more than one hundred vents (Hintz and Valentine, 2012; Valentine and Cortés, 2013) that are distributed over two mountain ranges, the Reveille Range and Pancake Range (Figure 1). Previous studies have shown that basaltic volcanism commenced in the Pliocene in the Reveille Range, and has generally shifted northward to the Pancake Range with time (Foland and Bergman, 1992). The early eruptions (~5.9 to 5 Ma) consist of alkali-basalt, followed by trachytic eruptions (~4.4 to 4.2 Ma) and lastly, basaltic eruptions (4.6 Ma to 38 ka; Naumann, 1991). The most recent activity occurred in the northern part of the volcanic field as recently as 35-38 ka ago

(Heizler, 2013; Shepard et al., 1995) suggesting that the LCVF represents an active volcanic field. The LCVF magmas were erupted through relatively thin crust (30-33 km), with a lithospheric thickness beneath the volcanic field estimated from elastic models and seismic anisotropy data to be 65-70 km (Jones and Phinney, 1998). The crust is composed of Proterozoic crystalline basement rocks overlain by a sequence of Paleozoic carbonate and clastic sediments (Menzies, 1989), which are in turn overlain by Oligocene and Miocene ash flow tuffs that range in composition from dacite to rhyolite (Ekren et al., 1972). The volcanic field is part of the Basin and Range province, which has been dominated by an extensional regime since the Oligocene (Atwater and Stock, 1998; Christiansen and Yeats, 1992; Eaton, 1982; Glazner and Bartley, 1984; Zoback et al., 1981). Pliocene through Quaternary volcanism in the Basin and Range occurs mostly along its margins; however, the LCVF is isolated in the central part of the basin. Previous studies have noted that the LCVF parallels the NNE-trending normal faults of the Basin and Range Province, and have suggested that the volcanism may be associated with upwelling of asthenospheric mantle in response to Basin and Range lithospheric extension (Foland and Bergman, 1992; Lum et al., 1989). In detail, structural analysis indicates that feeder dikes for Quaternary volcanoes in the LCVF are aligned parallel to the maximum horizontal compressive stress that has been active during the Quaternary, as well as being parallel to Quaternary-age faults (Tadini et al., 2014). On a broader regional scale, the relationship between Basin and Range basaltic magmatism and extension remains unclear. Numerous studies have indicated that mafic volcanism progresses from lithospheric to asthenospheric sources with time, indicating that extension may play an important role in magma generation (Farmer et al., 1989; Kempton et al., 1991), but it is not clear why this mechanism should focus most of the volcanism around the margins, rather than throughout the interior of the Basin and Range. Other mechanisms, such as mantle upwelling due to lithospheric delamination (e.g. West et al., 2009), asthenospheric convection at a lithospheric keel edge (Stickney 2004) and shear-driven upwelling (Conrad et al., 2010) have also been proposed to generate the recent mafic volcanism in the Basin and Range.

3. Sample Locations, Ages and Petrography

Companion papers by Valentine and Cortés (2013) and Cortés et al. (2015) discuss in detail the field geology, age and petrography of the northern LCVF. The present study focuses on the volcanic centers in the northern part of the volcanic field including three recently named volcanic centers that we refer to as the Hi Desert, Mizpah, and Giggle Springs volcanoes (Cortés et al., 2015), several yet unnamed lava flows, and Marcath and Easy Chair volcanoes. Hi Desert, Mizpah and Giggle Springs are located within ~500 m of each other, and were selected to test for compositional variability in closely spaced volcanoes that erupted over a span of time. The Marcath volcano, located ~6 km SW of these cones and representing the most recent eruption in the volcanic field, is extremely well preserved and allowed for detailed sampling to test for compositional variability within a single monogenetic event. Easy Chair, located ~6 km SE of Marcath, was subject to a similar detailed sampling campaign that demonstrated compositional variability throughout the duration of the eruption; furthermore, evidence was found for an older eruption at the same location, referred to as proto-Easy Chair (Valentine and Cortés, 2013; Cortés et al., 2015). Recent 40Ar/39Ar age determinations yield ages of 840±3 ka for proto-Easy Chair, between 620-740 ka for Mizpah, 582±6 ka for an unnamed lava flow, 140±5 ka for Easy Chair, <81±5 ka for Giggle Springs, and 35±7 ka for Marcath (Heizler, 2013). Shepard et al. (1995) determined a 36Cl and 10Be age of 38.1±9.7 ka for the Marcath lavas. Age data are not available for the Hi Desert basalts; however, Giggle Springs lavas were emplaced around and bank up against the Hi Desert cone remnants, thus Hi Desert is clearly older than Giggle Springs. We collected 33 lava and tephra samples throughout the northern LCVF (Figure 1). All of the lavas are porphyritic and contain 5 to 20 % subhedral to euhedral phenocrysts that rarely exceed 3 mm in size (Cortés et al., 2015). Phenocryst phase assemblages are predominantly olivine + clinopyroxene + plagioclase in the Mizpah and Hi Desert samples, and olivine + clinopyroxene or olivine + plagioclase in the Giggle Springs, Easy Chair and Marcath flows (Valentine and Cortés, 2013; Cortés et al., 2015). Groundmass consists mainly of plagioclase and ferromagnesian minerals in all samples (Cortés et al., 2015). Large megacrysts (>1 cm) of olivine, clinopyroxene (Ti-dioside and Cr-diopside) and plagioclase (anorthite to bytownite) are common in the Marcath, Giggle Springs and

Easy Chair lava flows, and also occur as loose crystals in the related tephra deposits. Megacrysts of amphibole (Mg-hastingite) and spinel (pleonaste) also have been reported by Bergman et al. (1982) in the Marcath lavas. Marcath and Easy Chair lavas also contain ultramafic nodules of varying lithologies including harzburgites, dunites, wehrlites, lherzolites and clinopyroxenites (Bergman, 1982; Valentine and Cortés, 2013). No megacrysts or mantle xenoliths are found in the Mizpah flows, but the Hi Desert flows contain abundant megacrysts of plagioclase, ranging from 1-5 cm.

4. Analytical Techniques

Fresh lava samples were processed for chemical and isotopic analysis after removal of inclusions and megacrysts. Details of sample preparation for chemical analysis, and major and trace element analysis techniques, are provided in Cortés et al. (2015). Concentrations of major elements were obtained by X-ray fluorescence spectrometry (XRF) in the Department of Geoscience, University of Nevada Las Vegas. Trace element concentrations were determined by ICP-MS at Actlabs in Ancaster, Ontario, Canada. For isotopic analyses, samples were chipped in an alumina jaw crusher, and the chips were leached with 4N HCl to remove any potential surficial carbonate deposits. The chips were then handpicked under a binocular microscope to remove any remaining visible vesicle-filling material before being crushed and powdered in a high purity alumina shatterbox. Sample powders were leached in 2N HCl for 20 minutes in an ultrasonic bath and the residues rinsed multiple times with 18 mega-Ω H2O, following procedures of Snyder et al. (2004). One unleached sample powder (LC10-13) was also analyzed to compare with the leached residue of the same sample. Pb, Sr and Nd were all separated from the same sample aliquot of ~0.1g of leached powder, following conventional procedures described in detail by Snyder (2005). Sr, Nd, and Pb isotopes were measured by thermal ionization mass spectrometry (TIMS) on a Thermo-Finnigan Triton at Miami University. The measured 87Sr/86Sr was corrected for mass fractionation using 86Sr/88Sr =0.1194. 143Nd/144Nd was corrected for mass fractionation using 146Nd/144Nd= 0.7219. External precision based on long-term, two standard deviation (2 SD) reproducibility of the NBS 987 and La Jolla standards are ±0.000015 and ±0.000007 for 87Sr/88Sr and 143Nd/144Nd respectively. 206Pb/204Pb,

207Pb/204Pb, 208Pb/204Pb were corrected for mass fractionation by 0.10% per amu. Errors on 206Pb/204Pb, 207Pb/204Pb, and 208Pb/204Pb were 0.014, 0.019, and 0.063, respectively based on long-term 2 SD external reproducibility of the NBS 981 standard. No significant difference was found between the Nd and Pb isotope ratios of the leached and unleached runs of sample LC10-13. However, 87Sr/86Sr decreased slightly from 0.70324 (unleached) to 0.70321 after leaching, likely due to dissolution and removal of secondary carbonates and confirming the importance of the leaching step for all other samples. For Hf isotopic analysis, a separate aliquot of ~150 mg of sample powder was dissolved, and Hf was separated and purified for isotopic analysis following the procedures of Connelly et al. (2006). Hf isotopes were measured using a Nu Instruments Nu Plasma multi-collector (MC)-ICP-MS at the Department of Terrestrial Magnetism (DTM), Carnegie Institution of Washington. Isotopic measurements of samples were interspersed with those of the JM 475 standard, and the samples were normalized to a 176Hf/177Hf value for JM 475 of 0.282160. The average measured 176Hf/177Hf ratio for JM 475 during the period of data collection was 0.282151, with a 2 SD external reproducibility of ± 0.000009 to 0.000011. Os isotopic compositions and concentrations by isotope dissolution were determined together on a single sample aliquot. Approximately 1 g of leached sample powder was digested and equilibrated with 190Os spike in reverse aqua regia by reacting in a sealed Carius tube at 230⁰ C for 2 days, following the procedures described by Shirey and Walker (1995). Os was then extracted from the reverse aqua regia through two distillation steps, including a macro-distillation based on the procedures of Nägler and Frei (1997) and described in detail in Yu (2011), followed by a micro-distillation following the procedure of Roy-Barman and Allègre (1995). Purified samples were loaded onto single Pt filaments with a Ba-hydroxide overcoat, and analyzed by negative thermal ionization on the Thermo-Finnigan Triton at Miami University, utilizing a peak hopping method on a single secondary electron multiplier (SEM). Os isotope ratios were corrected for mass fractionation using 192Os/188Os = 3.0826, and corrected for oxygen isotopic composition using the values of Nier (1950). The external reproducibility for 187Os/188Os measurements based on the long-term reproducibility of a NIST Os solution standard is

±0.0002. However, internal errors for 187Os/188Os measurements of the samples were variable and ranged from 0.03-0.1%.

5.Results

5.1.Major and Trace Element Geochemistry

Volcanic rocks from the northern LCVF range in composition from basanite and basalt to trachybasalt, and are compositionally variable both across the volcanic field as well as within individual volcanic centers (e.g. Marcath and Easy Chair). Relatively primitive melt compositions (~10-11% MgO, 220-560 ppm Cr, 120-330 ppm Ni) are associated with several of the eruptive centers including Marcath, Giggle Springs, Hi

Desert, and one unnamed flow. Coherent variation for all LCVF lavas on a plot of Al2O3 versus MgO (Figure 2a) is consistent with variable degrees of fractionation of similar primitive parental magmas. However, the lavas from Giggle Springs, Easy Chair and

Marcath are more alkaline, with slightly lower average SiO2 (~44-45%) and higher total alkalis (K2O+Na2O >5%) than those from the other eruptive centers (most ~45-47% SiO2 and K2O+Na2O ≤5%) (Figure 2b). These differences are independent of degree of fractionation, as the former group of lavas exhibits higher K2O and Na2O at any given MgO (Figure 2c,d). One older sample from the Easy Chair locality, referred to as proto- Easy Chair, has a similar composition to the Hi Desert, Mizpah and unnamed lavas flows. Together, these data suggest that the older LCVF lavas (those from Hi Desert, Mizpah, the unnamed flows and proto-Easy Chair, all >140 ka) are compositionally distinct from the younger LCVF lavas of Giggle Springs, Easy Chair and Marcath (all ≤140 ka). For ease of discussion, the two compositional groups will be referred to hereafter as the GEM (Giggle Springs, Easy Chair, Marcath) and HiMUp (Hi Desert, Mizpah, Unnamed flows and proto-Easy Chair) groups. Trace element abundances and ratios also vary substantially across the volcanic field and within individual eruptive centers (Figure 2e-h). As for the major elements, the LCVF lavas fall into two distinct trace element groups, the GEM lavas displaying higher Ba, U, La/Yb and Zr/Hf at a given MgO than those of the HiMUp lavas (Figure 2e-g). All of the samples have trace element abundances and patterns (Figure 3) similar to those

17 observed in ocean island basalts (OIB), with enrichments relative to primitive mantle in moderately and highly incompatible elements, and enrichments in light rare earth elements (LREE) relative to heavy rare earth elements (HREE). However, there are subtle but significant compositional differences between the GEM and HiMUp lavas. The HiMUp lavas display trace element patterns that are broadly similar to those of HIMU- type basalts with significant enrichment of Nb and Ta, and depletion in large ion lithophile elements (LILE) such as Rb, Cs, Ba and K (Figure 3a). The GEM lavas exhibit similar trace element patterns, but overall are more enriched in highly incompatible trace elements including Rb, Cs and Ba, characteristics similar to those of EM-type OIB (Figure 3b).

5.2.Isotope Systematics

Isotopic data for the LCVF rocks are provided in Table 1. The GEM lavas exhibit minor Sr and Pb isotopic variability between one another and within a given volcanic center. In contrast, the HiMUp lavas display considerable ranges of Sr and Pb isotopic compositions and are isotopically distinct from those of the GEM lavas. Nonetheless, all northern LCVF volcanic rocks are characterized by similar Nd and Hf isotopic compositions. Overall, the Sr and Nd isotopic compositions of the northern LCVF samples are similar to those of other lavas from the Basin and Range, including those of the Geronimo Volcanic Field (southern Arizona), the Mojave Desert (California) and the Potrillo Volcanic field (New Mexico) (Kempton et al., 1991; Figure 4a). In detail, the isotope systematics of the northern LCVF lavas are somewhat complex, with essentially constant Nd and Hf isotope ratios (Figure 4b), but significant differences in Sr and Pb isotopes (Figure 4c). In εHf - εNd space, the samples fall slightly below the mantle array, and within the field of HIMU-type OIB such as those of St. Helena and Cook-Austral basalts (Figure 4b). Furthermore, all samples plot near the Northern Hemisphere Reference Line (NHRL) of Hart (1984) in both 207Pb/204Pb-206Pb/204Pb and 208Pb/204Pb- 206Pb/204Pb (Figure 5c, 5d). However, in Sr-Pb isotope space, the GEM lavas are isotopically distinct from the HiMUp lavas, the former having higher 87Sr/86Sr ratios (0.70353-0.70360), but lower 206Pb/204Pb ratios (19.170 -19.258) than the latter (87Sr/86Sr

= 0.70307-0.70328 and 206Pb/204Pb = 19.368-19.556) (Figure 5a, 5b). Together, the northern LCVF volcanic rocks produce a well-defined negative correlation between 87Sr/86Sr and 206Pb/204Pb (Figure 4c). In general, on a variety of multi-isotope diagrams (Figures 4 and 5), the GEM lavas fall within the range of isotopic compositions observed for EM-type OIB, whereas the HiMUp group samples fall within the field of FOZO as recently redefined by Stracke et al. (2005). However, on a 208Pb/204Pb versus 206Pb/204Pb diagram, the GEM lavas are displaced to slightly lower 208Pb/204Pb at a given 206Pb/204Pb compared to typical EM-like OIB, and trend towards the depleted MORB field. The northern LCVF volcanic rocks also display a wide range in Os isotopic compositions, with 187Os/188Os ranging from 0.1336-0.1678 in the GEM lavas and 0.1339-0.6107 in the HiMUp lavas (Figure 6). These values are significantly more radiogenic than depleted MORB mantle (DMM) and primitive upper mantle (PUM) (0.1275 and 0.1296 respectively; Snow and Reisberg, 1995; Meisel et al., 1996), however, most samples fall within the range inferred for nominally assimilation-free basalts (187Os/188Os ≤0.15) (e.g. Chesley et al., 2004; Lassiter and Hauri, 1998; Widom et al., 1999). Os concentrations of the LCVF basalts are also extremely variable, with the GEM lavas ranging from 4- 241 ppt Os, and the HiMUp lavas from 2- 40 ppt Os. These concentrations fall within the range of typical OIB and MORB (Reisberg et al., 1991; Roy-Barman and Allègre, 1995; Widom, 1997) and most western US basalts including Columbia River Flood basalts and those from the Oregon Plateau (e.g. Chesley and Ruiz, 1998; Hart el al., 1997). There are no clear correlations between 187Os/188Os and indices of differentiation such as MgO, Ni or Os concentration, although the most radiogenic 187Os/188Os samples tend to have the lowest Os (Figure 6), and the least radiogenic samples tend to have the highest Os concentrations. No correlations exist between 187Os/188Os and any of the other isotopes.

6. Discussion

The LCVF lavas exhibit a coherent trend on a plot of Al2O3 versus MgO (Figure 2a), potentially consistent with their derivation from similar parental magmas via similar fractional crystallization paths. However, slightly lower average SiO2 in GEM relative to HiMUp lavas suggests that the former may have been derived by deeper mantle melting

(e.g. Klein and Langmuir, 1987). Higher K2O, Na2O, Ba, U and La/Yb at a given MgO in the GEM lavas (Figure 2c-g) further suggests that they were derived either by lower degrees of melting or from a more alkaline, incompatible element enriched source. The positive correlation between La/Yb and Zr/Hf (Figure 2g) suggests the latter, as Zr/Hf ratios should not be significantly fractionated by variable degrees of partial melting due to their nearly identical ionic radii and charge (Hofmann, 1997). The distinct groupings of the LCVF lavas in terms of major and trace element compositions are consistent with the Sr and Pb isotope variations (Figure 4), which also require distinct sources. Together, the major element, trace element and isotopic data therefore define two end-member geochemical groups within the northern Lunar Crater Volcanic Field: the first includes the older lavas that comprise the HiMUp group, and the second include the younger (≤140 Ka) lavas that comprise the GEM group. The variable but enriched trace element and isotopic signatures can potentially be explained by several mechanisms including (1) crustal assimilation by mantle-derived melts, (2) melting or assimilation of metasomatized sub-continental lithospheric mantle, or (3) recycling into the mantle of oceanic crust and/or sediment. Furthermore, the well- defined trend in Sr-Pb isotope space is suggestive of two-component mixing, and could be attributed either to lithospheric assimilation or mixing of two mantle components (Figure 4c). Previous studies have interpreted chemical heterogeneities of LCVF magmas to result from a heterogeneous enriched mantle source (e.g. Bergman et al., 1981, 1982; Lum 1986). However, these studies did not address how the source diversity may arise. Later studies (e.g. Dickson, 1997; Foland and Bergman, 1992) revised previous interpretations and suggested that the trace element and isotopic compositions of the older LCVF magmas reflect lithospheric contamination. We use the new elemental and isotopic data to further evaluate the potential effects of magma-lithospheric interaction as well as the potential role of crustal recycling in the petrogenesis of the LCVF magmas.

6.1. Role of Crustal Assimilation

The negative correlation between 87Sr/86Sr and 206Pb/204Pb isotopes is consistent with two-component mixing, and could reflect a role for significant lower crustal assimilation in the petrogenesis of the GEM magmas (Figure 4c). Lower crustal granulite and eclogite

20 xenoliths from localities within the Basin and Range (e.g. the Geronimo Volcanic Field) are extremely heterogeneous but most have very high 87Sr/86Sr and low 206Pb/204Pb (Kempton et al., 1990), making them suitable end-member compositions for the GEM lavas from the northern LCVF (Figure 5a). However, almost all of these lower crustal xenoliths have positive 8/4Pb signatures, and therefore assimilation of such crust is unlikely to produce Pb isotopic signatures of LCVF GEM lavas (Figure 5d) or Basin and Range lavas in general (Kempton et al., 1990). Os isotopes are highly sensitive tracers of crustal assimilation due to the extremely radiogenic Os isotope signatures in crustal rocks relative to the mantle (Chesley et al., 2002; Widom et al.,1999), and thus can help further assess the potential role of lower crustal assimilation in the petrogenesis of the GEM lavas. Figure 6 demonstrates that the most evolved samples with Os concentrations <10 ppt and 187Os/188Os > 0.15, which includes a subset of the GEM samples, may have experienced up to ~2-2.5% lower crustal assimilation. Nevertheless, the lack of correlation of other isotopes with 187Os/188Os suggests that the incompatible radiogenic isotopes (Sr, Nd, Pb and Hf) are not significantly impacted by lower crustal assimilation. Furthermore, the scattered but positive correlation of Os isotopes with Nb/U (Figure 7) is precisely opposite that expected from assimilation of either upper or lower crust, both of which are typically characterized by highly radiogenic Os (Esperança et al., 1997; Esser and Turekian, 1993; Saal et al., 1998) but low Nb/U (Rudnick and Gao, 2003). Rather, the Nb/U ratios in the lavas, although variable, are within the range of normal MORB and OIB (47±10; Hofmann 1997), indicating that the lava trace element signatures also are primarily controlled by their mantle source compositions without significant modification by assimilation of crust. Hence, we interpret the variations in trace elements and the Sr-Nd- Pb and Hf isotope ratios to reflect mantle source compositions. In the following discussion, we therefore consider mantle sources and processes that could explain the observed geochemical and isotopic variations in the northern LCVF volcanic rocks.

6.2. Role of Metasomatized Subcontinental Lithospheric Mantle

Metasomatism of lithospheric mantle can generate OIB-like trace element signatures and significant isotopic heterogeneities within continental basalts (e.g. Erlank et al.,

1987; Niu and O’Hara, 2003; Pilet et al., 2010). The regions of northern Nevada and Utah are likely underlain by Proterozoic lithospheric mantle, with Nd model ages ranging from 2 to 2.3 Ga (Bennett and DePaolo, 1987). Low-angle subduction of the Farallon plate beneath much of the western United States during the Laramide orogeny likely resulted in hydration of the North America lithosphere (Coney and Reynold, 1977; Riter and Smith, 1996; Smith et al., 1999) and the enriched trace element and isotopic signatures of some of the Basin and Range alkali basalts including the Death Valley- Pancake Range have been attributed to partial melting of such metasomatized lithospheric mantle (Beard and Johnson, 1997; Bradshaw et al. 1993; Farmer et al., 1989; Kempton et al., 1991; Yogodzinski and Smith, 1995). However, in the case of the northern LCVF lavas, pressures and temperatures of melt generation based on the orthopyroxene-liquid geobarometer of Putirka (2008) suggest that melting beneath LCVF occurs at depths of ~60-130 km, within the asthenosphere (Cortés et al., 2015). Wang et al. (2002) present similar findings based on whole rock major element and REE chemistry, and suggest that alkaline basalts from the Central Basin and Range including the LVCF are sourced in the asthenosphere at depths >120 km, and that melting stops at the base of the lithosphere. It is possible the mantle beneath the LCVF experienced thermal conversion of lithosphere into asthenosphere (Beard and Johnson, 1997), resulting in the local asthenosphere inheriting geochemical signatures of the bulk lithosphere. The Nd and Pb isotopic signatures of the LCVF lavas, however, are difficult to reconcile with expected lithospheric mantle compositions. The geochemical signatures of the Tahoe-Truckee Volcanic Field lavas, western Nevada are generally interpreted to represent the chemical signatures of the lithospheric mantle beneath the area (e.g. Cousens et al., 2011, 2012). These lavas have low 143Nd/144Nd and significantly positive Δ7/4 and Δ8/4 that are distinct from the LCVF lavas (Figure 5b-d), which have much higher 143Nd/144Nd, Δ7/4 values that fall on the NHRL, and Δ8/4 values that fall below and parallel to the NHRL. Taken together, these findings suggest that the LCVF magmas are neither sourced in, nor contain a significant component of, the lithospheric mantle. Below, we explore alternative models for deriving the enriched chemical and isotopic compositions from an asthenospheric mantle source.

6.3. Origin of the HiMUp Source

Trace element and Sr-Pb isotope ratios of the HiMUp lavas are very similar to those of HIMU/FOZO-type basalts. HIMU- and FOZO-type OIB signatures are commonly attributed to a component of recycled oceanic crust in their mantle source (Chauvel et al., 1992; Weaver, 1991). Although the net effect of the subduction processes on the elemental budget of basaltic crust and sediments is not particularly well constrained, experimental studies have shown that subduction dewatering can remove a substantial fraction of the fluid-mobile element budget from the slab (Brenan et al., 1995; Chauvel et al., 1995; Kessel et al., 2005; Kogiso et al., 1997; Schmidt and Poli, 1998) resulting in depletion in LILE but a relative enrichment in the residue of water insoluble elements such as Nb and Ta. Incorporation of such dehydrated crust into the mantle therefore can potentially explain the depletion in Rb, Cs, Pb, and Ba and the enrichment of the immobile elements Nb and Ta observed in the HiMUp lavas. Although the K-depletion in continental basalts is often attributed to the presence of amphibole and/or phlogopite in the source during partial melting (e.g. Panter et al., 2006), these phases are not stable in asthenospheric mantle (Niida and Green, 1999) and thus the depletion in K also may be linked to the involvement of dehydrated oceanic crust in the source of HiMUp lavas. Recycling of ancient oceanic crust can also provide an explanation for the high 206Pb/204Pb but relatively low 87Sr/86Sr in the HiMUp lavas. Because of the large difference in element mobility, the preferential loss of Rb and Pb relative to U, Th and Sr during subduction dewatering leads to high U/Pb but low Rb/Sr in the dehydrated oceanic crust. Long-term evolution of such a high U/Pb and low Rb/Sr component produces high 206Pb/204Pb but similar or slightly low 87Sr/86Sr ratios compared to that of MORB (Stracke et al., 2005), consistent with the Sr and Pb isotopic signatures of the HiMUp lavas (Figure 5).

The negative εHf signatures in the HiMUp lavas can also be explained by ancient recycled crust. Oceanic crust, produced by mantle melting with residual garnet, will have low Lu/Hf relative to Sm/Nd and will evolve to relatively unradiogenic 176Hf/177Hf and negative εHf over time (Chauvel et al., 2009; Hanan et al., 2004; Patchett et al., 1984; Salters and White, 1998; Stracke et al., 2005). Recycling of such crust would produce a mantle reservoir that falls below the mantle array. Indeed, the LCVF volcanic rocks all

23 plot slightly below the MORB-OIB array in εHf-εNd space, and fall well within the field of

HIMU basalts. The positive εNd but negative εHf values of the LCVF magmas fall within the range of late Cenozoic basalts in the western United States that have been interpreted to have been derived from an asthenospheric source plus a lithospheric pyroxenite contribution (Beard and Johnson, 1997), but as indicated previously, the melting depths and Pb isotopic compositions of the LCVF lavas favor an asthenospheric source for the pyroxenitic component. Taken together, the trace element and Sr-Nd-Pb-Hf isotope data are all consistent with the LCVF HiMUp magmas being derived from an asthenospheric mantle source broadly similar to HIMU OIB and containing a component of ancient recycled, dewatered oceanic crust.

6.4. Origin of the GEM Source

Volcanic rocks from the GEM lavas are characterized by higher 87Sr/86Sr but lower 206Pb/204Pb compared to the HiMUp lavas. In addition, they are characterized by greater enrichment of LILE such as Rb, Cs, Ba, U and K. These signatures could be generated by subduction of oceanic crust plus sediments followed by long-term storage in the mantle, as many sediments are characterized by high LILE and low U/Pb (Weaver, 1991). However, recycled sediment is also characterized by high Th/U, which would evolve to high 208Pb/204Pb for a given 206Pb/204Pb, and thus produce positive 8/4Pb signatures as observed in EM-type OIB. In contrast, the GEM lavas have 8/4Pb signatures that fall below the mantle array, arguing against the involvement of ancient sediment or sediment- derived fluids in their source. Furthermore, as noted previously, both the GEM and HiMUp lavas have identical Hf and Nd isotopic compositions, indicating that they share a common component in their mantle sources, and arguing against the addition of significant Hf or Nd from any additional source. Addition of hydrous fluid to a source similar to that of the HiMUp lavas appears to be most consistent with the lack of variation in 176Hf/177Hf, as Hf and other HFSE (e.g. Nb, Ta) have very low solubility in aqueous fluid (Kogiso et al., 1997; , Martin et al., 2011; Tatsumi and Kogiso, 1997) but are easily mobilized by silicate melts (e.g. Kessel et al., 2005). We therefore suggest that the elevated Rb, Cs, Ba, U and K in

24 the GEM lavas compared to the HiMUp lavas might be attributed to contribution of hydrous fluid from a subducted slab. The negative 8/4Pb of the GEM lavas compared to typical EM-type OIB, and the lower 206Pb/204Pb and 208Pb/204Pb relative to the HiMUp lavas, could be explained by addition to the GEM mantle source of fluid derived from subduction of hydrated oceanic crust.

6.5. Isotope and Trace Element Mixing Models

Based on the above considerations, we infer that the LCVF magmas are sourced by a heterogeneous mantle modified by subduction processes including incorporation of ancient recycled oceanic crust ± hydrous fluid derived from subducted altered oceanic crust, to produce the GEM and HiMUp chemical and isotopic signatures, respectively. The negative linear correlation of 87Sr/86Sr vs. 206Pb/204Pb suggests that mixtures of these two mantle components could reproduce the entire compositional range of the LCVF (Figure 8; modeling parameters in Appendix 1). We chose as model end-members a depleted mantle source with trace element compositions similar to those of modern MORB mantle (e.g. Salters and Stracke, 2004), and subducted oceanic crust assumed to have the chemical composition of average N-MORB but modified by seawater-alteration and sub-arc dehydration processes. The trace element compositions of both recycled, dehydrated oceanic crust and hydrous fluid were calculated applying the element mobilities of Kogiso et al. (1997). Because oceanic crust and hydrous fluid have very high Pb and Sr concentrations compared to the mantle, the Sr and Pb isotopic compositions of the mixture are strongly influenced by those of the recycled components. Modeling results indicate that mixing of approximately 18% of 0.8 Ga recycled oceanic crust with ~82% depleted MORB mantle (DMM) can reproduce the Pb (206Pb/204Pb, 207Pb/204Pb and 208Pb/204Pb) and Sr isotopic compositions of the HiMUp end-member. Incorporation of <1% hydrous fluid derived from oceanic crust into the source of the HiMUp lavas can reproduce the Pb and Sr isotopic signatures of the GEM lavas (Figure 8). Variable degrees of mixing between these two mantle end-members can generate the full range of isotopic compositions observed in the northern LCVF sample suite. Within the HiMUp sample suite (e.g. Hi Desert and Mizpah), isotopic variations indicate a minor but variable contribution from the fluid modified source. Small isotopic variations within

25 single eruptions of the GEM group (e.g. Marcath and Easy Chair) also indicate a variable but substantially greater contribution from the fluid modified source. The mixing models can also explain the suprachondritic 187Os/188Os signatures of the most primitive LCVF samples (Figures 6, 8b). Because basaltic crust has low Os but high Re contents (e.g. Schiano et al., 1997; Hauri and Hart 1993, 1997; Kogiso et al., 2004), ancient recycled oceanic crust would be expected to develop extremely radiogenic 187Os/188Os. Assuming an average 187Re/188Os in the recycled crust of 50 (e.g. Widom et al., 1999), 0.8 Ga old recycled oceanic crust would develop 187Os/188Os of ~0.76. A mixture containing 18% of such recycled oceanic crust (Os = 100 ppt, e.g. Widom et al., 1999) and 82% depleted MORB mantle (Os = 3 ppb, 187Os/188Os of 0.1275; Snow and Reisberg, 1995) can reproduce the least radiogenic 187Os/188Os signatures of the LCVF lavas. Although the solubility of Os during slab dehydration is still in question (e.g. Becker, 2000; Brandon et al. 1996; Widom, 2011), the high solubility of Pb in slab- derived fluids dictates that such fluid would have a high Pb/Os ratio, resulting in a significant change in Pb but minimal change in Os isotope ratios between the fluid- present and fluid-absent sources of the GEM and HiMUp lavas, respectively. The distinct Pb isotopic compositions but relatively constant 187Os/188Os in the most primitive samples from both LCVF groups (0.1336-0.1339) is consistent with radiogenic Os contributed by recycled oceanic crust, but a negligible contribution from the slab-derived fluid (Figure 8b). We have also investigated whether the trace element systematics of the LCVF lavas can be adequately explained by the above model involving mixing of 82% DMM and 18% recycled oceanic crust. Similar to the isotopic modeling, the depleted mantle end- member is assumed to have trace element compositions identical to those of present-day DMM, and the subducted oceanic crust is assumed to have a chemical composition of average normal MORB (N-MORB) but modified by seawater-alteration and subduction processes. The element mobilities of Kogiso et al. (1997) were applied to calculate the trace element compositions of both the dehydrated oceanic crust and the hydrous fluid. Both non-modal batch and non-modal fractional melting of the hybrid mantle source (i.e. 82% DMM and 18% ROC) replicate the overall incompatible element patterns of the HiMUp lavas (Figure 9). We find the best agreement between observed and modeled

26 trace element compositions for 0.5% melting of a hybridized mantle with Ol:Opx:Cpx:Grt = 45:25:22:8. Similarly, the incompatible element pattern produced by batch melting of the same mixture, to which <1% hydrous fluid derived from oceanic crust has been added, closely resembles the pattern observed in the GEM lavas. In this model, ~0.7% partial melting of the hydrated hybrid mantle source is required to generate the overall abundances of the incompatible elements of the GEM lavas. The slightly higher degree of partial melting in the generation of the GEM lavas compared to the HiMUp lavas is consistent with the addition of a hydrous fluid component to the source of the former. The general model of a common mantle source beneath the LCVF, composed of depleted MORB mantle with a component of ancient recycled oceanic crust to which hydrous fluid has been variably added, is therefore capable of explaining both the isotope characteristics and the trace element patterns of LCVF magmas.

6.6. Implications for Melt Volumes, Magma Ascent and Storage

Some studies have suggested that Basin and Range basalts are produced by substantially higher degrees of melting than that inferred from our models above. For example, Wang et al. (2002) suggested based on major element systematics that 5-15% partial melting is required to produce Basin and Range basalts, including those of Lunar Crater. However, Beard and Johnson (1997) argued for low degrees of partial melting (~1%) of a hybrid mantle source to produce the range of trace element abundances in central Basin and Range basalts, again including Lunar Crater. Most U-series isotopic studies also suggest that continental basalts are produced by low degrees of melting (<1%). For example, Asmerom et al. (2000) have demonstrated that the Pa-Th-U isotopic data for Pinacate volcanic field, Mexico, and Potrillo volcanic field in the southern Rio Grande Rift, USA, require very low degree (~0.4%) partial melting of a garnet-bearing source. The two latter estimates of degree of melting are similar to those in the present study. Although a detailed discussion of mechanisms of melt transport is beyond the scope of this paper, recent studies have shown that small volumes of melt can segregate from the mantle and readily ascend to the surface through dike or crack propagations. For example, Spera and Fowler (2009) suggest that small volume melts may segregate from

27 the mantle residue and begin to ascend to the surface through vertical to sub-vertical cracks when the magma pressure is slightly above the mean normal stress σn (σ1 +σ2 +σ3). The rate of magma ascent depends on the melt viscosity, crack sizes and the pressure gradient. Similarly, Valentine and Perry (2007) suggest that small volume partial melts existing in the lithosphere and uppermost asthenosphere can collect in localized domains in response to local deformation such as shearing. The melts that are collected in these domains may achieve sufficient interconnectivity that a buoyant pressure head can develop and trigger dike propagation. Tadini et al. (2014) discuss how the spatial distribution of vent clusters within the northern LCVF might also be consistent with a deformation-driven focusing process. An important aspect of this magma transport mechanism is that fractional crystallization can occur along the crack margins as the melt rises through the cold lithosphere (Spera and Fowler, 2009), which could explain the variable major element compositions of the lavas observed within each given eruptive center (Figure 2) without the requirement for significant stalling or ponding of magmas during ascent, and consistent with the evidence for limited crustal assimilation as well as the presence of mantle xenoliths in some of the lavas (e.g. Marcath and Easy Chair). 6.7. Implications for the Origin and Evolution of Volcanism in the Lunar Crater Volcanic Field

The apparent presence of two chemically and isotopically distinct mantle domains and the close geographic proximity of the resulting eruptive products indicate that the mantle beneath the LCVF region is characterized by small-scale chemical and isotopic heterogeneity. A tentative observation from this study is that the eruption of magmas with distinct trace element and isotopic signatures may be primarily a function of time rather than space, as the GEM group comprises the youngest eruptive centers (all <140 ka) and the HiMUp group the oldest volcanic centers (>140 ka). Of particular note is that the lavas from Easy Chair and proto-Easy chair have isotopic signatures characteristic, respectively, of the GEM and HiMUp endmembers, despite having erupted in essentially the same geographic location. However, our results also demonstrate that single eruptions (e.g. Marcath, Hi Desert) can show moderate to substantial degrees of heterogeneity consistent with variable degrees of mixing between these two distinct endmember mantle

28 sources (Figures 4 and 5), and suggesting that both mantle “flavors” are present in the melting region beneath single eruptive centers. The slightly lower SiO2 at a given MgO (Figure 2) as well as calculated pressures of 2.3-3.2 GPa and ~3-3.5 GPa in the GEM and HiMUp group lavas, respectively (Cortés et al., 2015; E.I. Smith, unpublished data), indicate that depth of melting may have increased with time, further suggesting a vertical stratification in mantle source composition. Figure 10 summarizes the proposed scenario for the development of the GEM and HiMUp mantle sources in time and space, based on the observed elemental and isotopic variations in the LCVF, and the models developed in the present study. The cause of melting beneath the LCVF still remains uncertain. The LCVF is underlain by relatively thin lithospheric mantle (60-75 km) and is located just southeast of the high-velocity cold mantle cylinder that is interpreted as evidence of lithospheric drip beneath the Great Basin (West et al., 2009). The LCVF magmatism may be due to upwelling and melting of hot asthenosphere in response to this lithospheric downwelling (West et al., 2009) or to extension and lithospheric thinning (Jones, 1987; Wernicke et al., 2008). If adiabatic decompression melting is the main mechanism for melt generation beneath the LCVF, the thickness of the overlying lithosphere dictates the average depth of melting of a mantle column (Humphreys and Niu, 2009), which will be deeper when the lithosphere is thicker. In the Buffalo Valley volcanic field, which is just 200 km north of the LCVF and also lies on thinned lithosphere, Cousens et al. (2013) likewise observed an increase in depth of source melting with time, which they interpret to indicate cessation of the lithospheric thinning followed by slow cooling and re-thickening of the lithosphere such that more recent melting occurs at greater depth. A similar scenario could also explain the apparent increase in depth of melt generation over time beneath the LCVF. Mafic, monogenetic volcanic fields frequently exhibit chemical and isotopic heterogeneity, with evidence for mixing of multiple geochemically distinct mantle components, regardless of their tectonic setting. Global examples include the Trans- Mexican Volcanic Belt (e.g. Luhr, 1997; Verma and Hasenaka, 2004), the Auckland Volcanic Field (e.g. McGee et al., 2013), Jeju Island, Korea (Brenna et al., 2012), and the Harrat Ash Shaam, Jordon (e.g. Shaw et al., 2003), among many others. Many of these

29 exhibit broadly similar behaviors to the LCVF in terms of low degrees of partial melting (e.g. McGee et al., 2013), increasing depth of melting with time (e.g. Brenna et al., 2012; Cousins et al., 2013), and rapid source to surface transport without significant magma ponding (e.g. Brenna et al., 2012; McGee et al., 2013), but major element fractionation trends that may be attributed to crystallization along the conduit (e.g. Brenna et al., 2012). Although the detailed geochemical and isotopic signatures vary among monogenetic volcanic fields globally, many involve heterogeneous asthenospheric sources ± lithospheric contributions. Although the processes by which such heterogeneities are generated likely differ depending on tectonic setting, the evidence for these processes may be preferentially preserved in these types of volcanic systems where small degrees of melting and rapid source to surface transport prevail; such processes may be more easily masked in other types of volcanic systems. Detailed studies of mafic, monogenetic volcanic fields are therefore key to unraveling processes of mantle source enrichment and heterogeneity in time and space.

Acknowledgments We are grateful to Rick Carlson for helping with the Hf isotopic analyses, and to Amanda Hintz, Peter Johnson, Jamal Amin, Dayana Schonwalder, Patrick Whelley, Mai Sas and Hugo Belmontes for their assistance during the fieldwork. We also thank Vince Salters and Mary Reid for their very thorough and helpful comments on the manuscript. This study was supported by NSF EAR grants #1016042 (E. Widom), #1016100 (G. Valentine and J. Cortés), and #1016104 (E. Smith).

References

Allègre, C.J., Turcotte, D.L., 1986. Implications of a two-component marble-cake mantle. Nature, 323, 123–127.

Atwater, T., Stock, J., 1998. Pacific North American plate tectonics of the Neogene southwestern United States: An update, International Geology Review, 40, 5, 375-402.

Asmerom, Y., Cheng, H., Thomas, R., Hirschmann, M., Edwards, R.L., 2000. Melting of the Earth’s lithospheric mantle inferred from protactinium-thorium-uranium isotopic data. Nature 406, 293-296.

Bach W., Peucker-Ehrenbrink, B., Hart, S.R, Blusztajn, J.S., 2003. Geochemistry of hydrothermally altered oceanic crust: DSDP/ODP Hole 504B – Implications for seawater-crust exchange budgets and Sr- and Pb-isotopic evolution of the mantle. Geochemistry Geophysics Geosystems 4, 29 pp. doi:10.1029/2002GC000419.

Beard, B. L., Johnson, C. M., 1997. Hafnium isotope evidence for the origin of Cenozoic basaltic lavas from the southwestern United States. Journal of Geophysical Research 102, 20, 149- 20,178.

Becker, H., 2000. Re-Os fractionation in eclogites and blueschists and the implications for recycling of oceanic crust into the mantle. Earth and Planetary Science Letters 177, 287-300.

Bennett, V.C., DePaolo, D.J., 1987. Proterozoic crustal history of the western United States as determined by neodymium isotopic mapping. Geological Society of America Bulletin 99, 674- 685.

Bergman, S.C., 1982. Petrogenetic aspects of the alkali basaltic lavas and included megacrysts and nodules from the Lunar Crater Volcanic Field, Nevada, U.S.A. Ph.D. dissertation, Ohio State University, Columbus, Ohio.

Bergman, S.C., Foland, K.A., Spera, F.J., 1981. On the origin of an amphibole-rich vein in a peridotite inclusion from the Lunar Crater Volcanic Field, Nevada, U.S.A. Earth and Planetary Science Letters 56, 343-361.

Blondes, M.S., Reiners, P.W., Duccea, M.N., Singer, B.S., Chesley, J., 2008. Temporal- compositional trends over short and long time-scales in basalts of the Big Pine Volcanic Field, California. Earth and Planetary Science Letters 269, 140-154.

Borg, L.E., Brandon, A.D., Clynne, M.A., Walker, R.J., 2000. Re-Os isotopic systematics of primitive lavas from the Lassen region of the Cascade arc, California. Earth and Planetary Science Letters 177, 301-317.

Bradshaw, T.K., Hawkesworth, C.J., Gallagher, K., 1993. Basaltic volcanism in the southern Basin and Range: No role for a mantle plume. Earth and Planetary Science Letters 116, 45-62.

Brandon, A.D., Creaser, R.A., Shirey, S.B., 1996. Osmium recycling in subduction zones. Science 272, 861-863.

Brenan, J.M., Shaw, H.F., Ryerson, F.J., Phinney, D.L., 1995. Mineral-aqueous fluid partitioning of trace-elements at 900⁰C and 2.0 GPa: Constraints on the trace-element chemistry of mantle and deep-crustal fluids. Geochimica et Cosmochimica Acta 59, 3331-3350.

Carlson, R.W., Nowell, G.M., 2001. Olivine-poor sources for mantle-derived magmas: Os and Hf isotopic evidence from potassic magmas of the Colorado Plateau. Geochemistry Geophysics Geosystems 2, 2000GC000128

Chauvel, C., Goldstein, S.L., Hofmann, A.W., 1995. Hydration and dehydration of oceanic crust controls Pb evolution of the mantle. Chemical Geology 126, 65–75.

Chauvel, C., Hofmann, A.W., Vidal, P., 1992. HIMU-EM: The French Polynesian connection. Earth and Planetary Science Letters 110, 99-119.

Chauvel, C., Marini, J-C., Plank, T., Ludden, J.N., 2009. Hf-Nd input flux in the Izu-Mariana subduction zone and recycling of subducted material in the mantle. Geochemistry, Geophysics, Geosystems 10, 23 pp. doi:1029/2008GC002101.

Chesley J., Ruiz, J., Righter, K., Ferrari, L., Gómez-Tuena, A., 2002. Source contamination versus assimilation: an example from the Trans-Mexican Volcanic Arc. Earth and Planetary Science Letters 195, 211-221.

Chesley, J.T., Ruiz, J., 1998. Crust-mantle interaction in large igneous provinces. Implications from the Re–Os isotope systematics of the Columbia River flood basalts. Earth and Planetary Science Letters 154, 1–11.

Chesley, J.T., Righter, K., Ruiz, J., 2004. Large-scale mantle metasomatism: a Re-Os perspective. Earth and Planetary Science Letters 219, 49-60.

Christiansen, R.L., Yeats, R.S., 1992. Post-Laramide geology of the western U.S. Cordillera, in Burchfiel, B. C., Lipman, P. W., and Zoback, M. L., The Cordilleran Orogen: conterminous U. S.: The Geology of North America, Volume G-3, Decade of North American Geology, Geological Society of America, Boulder, p. 261-406.

Coney, P.J., Reynolds, S.J., 1977. Flattening of the Farallon slab. Nature 270, 403-406.

Connelly, J.N., Ulfbeck, D.G., Thrane, K., Bizzarro, M., Housh, T., 2006. A method for purifying Lu and Hf for analyses by MC-ICP-MS using TODGA resin. Chemical Geology 233, 126-136.

Connor, C.B., Conway, F.M., 2000. Basaltic volcanic fields. In: Sigurdsson, H. (Ed.), Encyclopedia of Volcanoes, Academic Press, New York, pp. 331-343.

Conrad, C.P, Wu, B., Smith, E.I., Bianco, T.A., Tibbetts, A., 2010. Shear-driven upwelling induced by lateral viscosity variations and asthenospheric shear: A mechanism for intraplate volcanism. Physics of the Earth and Planetary Interiors 178, 162–175.

Cortés, J.A., Smith, E.I., Valentine, G.A., Johnsen, R., Rasoazanamparany, C., Widom, E., Sas, M., Ruth, D., 2015. Intrinsic conditions of magmas genesis at the Lunar Crater Volcanic Field (Nevada): Implications for internal plumbing and magma ascent. American Mineralogist 100.

Cousens, B.L., Henry, C.D., Harvey, B.J. 2011. Secular variations in magmatism during a continental arc to post-arc transition: Plio-Pleistocene volcanism in the Lake Tahoe/Truckee area, Northern Sierra Nevada, California. Lithos 123, 225-242.

Cousens, B., Henry, C.D., Gupta, V., 2012. Distinct mantle sources for Pliocene-Quaternary volcanism beneath the modern Sierra Nevada and adjacent Great Basin, northern California and western Nevada, USA. Geosphere 8, 562-580.

Dickson, L.D., 1997. Volcanology and geochemistry of Pliocene and Quaternary basalts on Citadel Mountain, Lunar Crater Volcanic Field, Pancake Range, Nevada. M.S. Thesis, University of Nevada Las Vegas.

Eaton, G.P, 1982. The Basin and Range Province: Origin and tectonic significance. Annual Review of Earth and Planetary Sciences10, 409-440.

Ekren, E.B., Snyder, R.P., Dixon, G.L. 1972. Geologic map of the Lunar Crater quadrangle, Nye County, Nevada. U.S. Geological Survey Miscellaneous Geologic Investigations Map I-700.

Erlank, A., Waters, F., Hawkesworth, C., Haggerty, S., Allsop, H., Rickard, R., Menzies, M., 1987. Evidence for mantle metasomatism in peridotite nodules from the Kimberly pipes, South Africa. In Mantle Metasomatism, Menzies, M. and Hawkesworth, C., eds., 221-312. Academic Press, Geology Series.

Esperança, S., Carlson, R.W., Shirey, S.B., Smith, D., 1997. Dating crust-mantle separation; Re- Os isotopic study of mafic xenoliths from central Arizona. Geology, 25, 651–654.

Esser, B.K., Turekian, K.K., 1993. The osmium isotopic composition of the continental crust. Geochimica et Cosmochimica Acta, 57, 3093-3104.

Foland, K.A., Bergman, S.C., 1992. Temporal and spatial distribution of basaltic volcanism in the Pancake and Reveille Ranges north of Yucca Mountain. Proceedings of the International Nuclear Waste Symposium, v. 2, American Nuclear Society and American Society of Civil Engineers, 2366-2371.

Glazner, A.F., Bartley, J. M., 1984. Timing and tectonic setting of Tertiary low-angle normal faulting and associated magmatism in the southwestern United States. Tectonics 3, 385–396.

Hanan, B.B., Blichert-Toft, J., Pyle, D.G., Christie, D.M., 2004. Contrasting origins of the upper mantle revealed by hafnium and lead isotopes from the Southeast Indian Ridge. Nature 432, 91– 94.

Hart, S.R., 1984. A large-scale isotope anomaly in the Southern Hemisphere mantle. Nature 309, 753–757.

Hart, W.K., Carlson, R.W., Shirey, S.B., 1997. Radiogenic Os in primitive basalts from the northwestern USA: Implications for petrogenesis. Earth and Planetary Science Letters 150, 103 - 116.

Hauri, E.H., Hart, S.R., 1997. Rhenium abundance and systematics in oceanic basalts. Chemical Geology 139, 185-205.

Hauri, E.H., Hart, S.R., 1993. Re-Os isotope systematics of HIMU and EMII oceanic island basalts from the South Pacific Ocean. Earth and Planetary Science Letters 114, 353-371.

Heizler, M.T., 2013. 40Ar/39Ar geochronology results for Lunar Crater Volcanic Field basalts, Nevada. New Mexico Geochronological Research Laboratory Report NMGRL-IR-792. https://vhub.org/resources/2503.

Hintz, A.R., Valentine, G.A., 2012. Complex plumbing of monogenetic scoria cones: New insights from the Lunar Crater Volcanic Field (Nevada, USA). Journal of Volcanology and Geothermal Research 239-240: 19-32.

Hirschmann, M.M., Stolper, E.M., 1996. A possible role for garnet pyroxenite in the origin of the ‘‘garnet signature’’ in MORB. Contributions to Mineralogy and Petrology 124, 185–208.

Hofmann, A.W., 1997. Mantle geochemistry: the message from oceanic volcanism. Nature 385, 219-229.

Humphreys, E.R., Niu, Y., 2009. On the composition of ocean island basalts (OIB): The effects of lithospheric thickness variation and mantle metasomatism. Lithos 112, 118- 136.

Johnson, E.R., Wallace, P.J., Cashman, K.V., Delgado Granados, H., Kent, A.J.R., 2008. Magmatic volatile contents and degassing-induced crystallization at Volcán Jorullo, Mexico: implications for melt evolution and the plumbing systems of monogenetic volcanoes. Earth and Planetary Science Letters 269, 478-487.

Jones, C.H., Phinney, R.A., 1998. Seismic structure of the lithosphere from teleseismic converted arrivals observed at small arrays I the southern Sierra, Nevada and vicinity, California. Journal of Geophysical Research 103, 10065-10090.

Jones, C.H., 1987. Is extension in Death Valley accommodated by thinning of the mantle lithosphere beneath the Sierra Nevada, California? Tectonics 6, 449–473.

Kempton, P.D., Harmon, R.S., Hawkesworth, C.J., Moorbath, S., 1990. Petrology and geochemistry of lower crustal granulites from the Geronimo Volcanic Field, southeastern Arizona. Geochimica et Cosmochimica Acta 54, 3401-3426.

Kempton, P.D., Fitton, J.G., Hawkesworth, C.J., Ormerod, D.S., 1991. Isotopic and trace element constraints on the composition and evolution of the lithosphere beneath the southwestern United States. Journal of Geophysical Research 96, 13713–13735.

Klein, E.M., Langmuir, C.H., 1987. Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness. Journal of Geophysical Research 92, 8089-8115.

Kessel, R., Schmidt, M.W., Ulmer, P., Pettke, T., 2005. Trace element signature of subduction- zone fluids, melts and supercritical liquids at 120-180 km depth. Nature 437, 724-727.

Kogiso, T., Tatsumi, Y., Nakano, S., 1997. Trace element transport during dehydration processes in the subducted oceanic crust, 1, Experiments and implications for the origin of ocean island basalts. Earth and Planetary Science Letters 148, 193-205.

Kogiso, T., Hirschmann, M.M., Reiners, P.W., 2004. Length scales of mantle heterogeneities and their relationship to ocean island basalt geochemistry. Geochimica et Cosmochimica Acta 68, 345-360.

Lassiter, J.C, Hauri, E.H, 1998. Osmium-isotope variations in Hawaiian lavas. Evidence for recycled oceanic lithosphere in the Hawaiian plume, Earth and Planetary Science Letters 164, 483-496.

Lassiter, J.C, Luhr, J.F, 2001. Osmium abundance and isotope variations in mafic Mexican volcanic rocks: Evidence for crustal contamination and constraints on the geochemical behavior of osmium during partial melting and fractional crystallization. Geochemistry, Geophysics and Geosystems 2, 2000GC000116.

Leeman, W.P., Harry, D.L., 1993. A binary source model for extension-related magmatism in the Great Basin, Western North America. Science 262, 1550–1554.

Luhr, J.F., Carmichael, I.S.E., 1985. Jorullo Volcano, Michoacán, Mexico (1759–1774): The earliest stages of fractionation in calc-alkaline magmas. Contributions to Mineralogy and Petrology 90, 142–161.

Luhr, J.F., 1997. Extensional tectonics and the diverse primitive volcanic rocks in the western Mexican Volcanic Belt. The Canadian Mineralogist 35, 473–500.

Lum, C.C.L., 1986. Aspects of petrogenesis of alkali basalts from the Lunar Crater Volcanic Field, Nevada. M.S. thesis, Ohio State University, Columbus, Ohio.

Lum, C.C.L., Leeman, W.P., Foland, K.A., Kargel, J.A., Fitton, J.G., 1989. Isotopic Variations in Continental Basalts as Indicators of Mantle Heterogeneity: Examples from the Western U.S. Cordillera. Journal of Geophysical Research 94, 7871-7884.

Martin, L.A.J., Wood, B.J., Turner, S., Rushmer, T., 2011. Experimental measurements of trace element partitioning between lawsonite, zoisite and fluid and their implication for the composition of arc magmas. Journal of Petrology 52, 1049-1075.

Martin, U., Németh, K., 2006. How Strombolian is a “Strombolian” scoria cone? Some irregularities in scoria cone architecture from the Transmexican Volcanic Belt, near Volcan Ceboruco, (Mexico) and Al Haruj (Libya). Journal of Volcanology and Geothermal Research 155, 104-118.

McGee, L.E., Smith, I.E.M., Millet, M-A., Handley, H.K., Lindsay, J.M., 2013. Asthenospheric control on melting processes in a monogenetic basaltic system: a case study of the Auckland Volcanic Field, New Zealand. Journal of Petrology 54, 2125-2153.

Meisel, T., Walker, R.J., Morgan, J.W., 1996. The osmium isotopic composition of the Earth`s primitive upper mantle. Nature 383, 517–520.

Meisel, T., Walker, R.J., Irving, A.J., Lorand, J-P., 2001. Osmium isotopic compositions of mantle xenoliths: a global perspective. Geochimica et Cosmochimica Acta 65, 1311-1323.

Menzies, M.A., 1989. Cratonic, circumcratonic and oceanic mantle domains beneath the western United States. Journal of Geophysical Research 94, 7899-7915.

Mukasa, S.B., Shervais, J.W., Wilshire, H.G., Nielson, J.E., 1991. Intrinsic Nd, Pb, and Sr isotopic heterogeneities exhibited by the Lherz alpine peridotite massif, French Pyrenees. Journal of Petrology, Special Volume 117–134.

Nägler, Th.F., Frei, R., 1997. “Plug in” Os distillation. Schweizerische Mineralogische und Petrographische Mitteilungen 77, 123-127.

Naumann, T.R., Smith, E.I., Shafiqullah, M., Damon, P.E., 1991. New K-Ar ages for Pliocene mafic to intermediate volcanic rocks in the Reveille Range, Nevada. Isochron West, 57, 12-16.

Nier, A.O., 1950. A redetermination of the relative abundances of the isotopes of carbon, nitrogen, oxygen, argon, and potassium. Physical Revue Letters, Physical review, and Reviews of Modern Physics 77, 789-793.

Niida, K., Green, D.H., 1999. Stability and chemical composition of pargasiatic amphibole in MORB pyrolite under upper mantle conditions. Contributions to Mineralogy and Petrology 135, 18-40.

Niu, Y., O'Hara, M.J., 2003. Origin of ocean island basalts: A new perspective from petrology, geochemistry, and mineral physics considerations. Journal of Geophysical Research 108, 2209, doi:10.1029/2002JB002048.

Panter, K.S., Blusztajn, J., Hart, S.R., Kyle, P.R., Esser, R., McIntosh, W.C., 2006. The origin of HIMU in the SW Pacific: Evidence from intraplate volcanism in southern New Zealand and subantarctic islands. Journal of Petrology 47, 1673-1704

Patchett, P.J., White, W.M., Feldmann, H., Kielinczuck, S., Hofmann, A.W., 1984. Hafnium/rare earth element fractionation in the sedimentary system and crustal recycling into the Earth’s mantle. Earth and Planetary Science Letters 69, 365-378.

Pearson, D.G., Davies, G.R., Nixon, P.H., 1993. Geochemical constraints on the petrogenesis of diamond facies pyroxenites from the Beni Bousera peridotite massif, North Morocco. Journal of Petrology 34, 125–172..

Peucker-Ehrenbrink, B., Ravizza, G., 2000. The marine osmium record. Terra Nova 12, 205-219.

Pilet, S., Ulmer, P., Villiger, S., 2010. Liquid line of descent of a basanitic liquid at 1.5 GPa: constraints on the formation of metasomatic veins. Contributions to Mineralogy and Petrology 159, 621–643.

Prinzhofer, A., Lewin, E., Allègre, C.J., 1989. Stochastic melting of the marble cake mantle: evidence from local study of the East Pacific Rise. Earth and Planetary Science Letters 92, 189– 206.

Putirka, K.D., 2008. Thermometers and barometers for volcanic systems. Reviews in Mineralogy and Geochemistry 69, 61-120.

Reisberg, L.C., Allègre, C.J., Luck, J-M., 1991. The Re-Os systematics of the Ronda ultramafic complex of southern Spain. Earth and Planetary Science Letters 105, 196-213.

Righter, K., Chesley, J.T., Ruiz, J., 2002. Genesis of primitive arc-type basalt: Constraints from Re, Os and Cl on the depth of melting and the role of fluids. Geology 30, 619–622.

Riter, J.C.A., Smith, D., 1996. Xenolith constraints on the thermal history of the mantle below the Colorado Plateau. Geology 24, 267-270.

Roy-Barman, M., Allègre C.J., 1995. 187Os/186Os in oceanic island basalts: tracing oceanic crust recycling in the mantle. Earth and Planetary Science Letters 129, 145–161.

Rudnick, R.L., Gao, S., 2003. Composition of the Continental Crust. Treatise on Geochemistry 3, 1-64.

Saal, A.E., Rudnick, R.L., Ravizza, G.E., Hart, S.R., 1998. Re–Os isotope evidence for the composition, formation and age of the lower continental crust. Nature 393, 58–61.

Salters, V.J.M., White, W.M., 1998. Hf isotope constraints on mantle evolution. Chemical Geology 145, 447-460.

Salters, V.J.M., Stracke, A., 2004. Composition of the depleted mantle. Geochemistry Geophysics Geosystems 5, 27 pp. doi:10.1029/2003GC000597.

Siebe, C., Rodriguez-Lara, V., Schaaf, P., Abrams, M., 2004. Geochemistry, Sr-Nd isotope composition, and tectonic setting of Holocene Pelado, Guespalapa and Chichinautzin scoria cones, south of Mexico City. Journal of Volcanology and Geothermal Research 130, 197-226.

Shaw, J.E., Baker, J.A., Menzies, M.A., Thirlwell, M.F., Ibrahim, K.M., 2003. Petrogenesis of the largest intraplate volcanic field on the Arabian plate (Jordan): a mixed lithosphere- asthenosphere source activated by lithospheric extension. Journal of Petrology 44, 1657-1679

Schmidt, M.W., Poli, S., 1998. Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation. Earth and Planetary Science Letters 163, 361-379.

Schiano, P., Birck, J.L., Allegre, C.J., 1997. Osmium-strontium-neodymium-lead isotopic covariations in mid-oceanic ridge basalt glasses. Earth and Planetary Science Letters 150, 363- 379.

Shirey, S.B., Walker, R.J., 1998. The Re-Os isotope system in cosmochemistry and high- temperature geochemistry. Annual Review of Earth and Planetary Sciences 26, 423–500.

Shirey, S.B., Walker, R.J., 1995. Carius tube digestion for low-blank rhenium-osmium analysis. Analytical Chemistry 67, 2136-2141.

Shepard, M.K., Arvidson, R.E., Caffee, M., Finkel, R., Harris, L., 1995. Cosmogenic exposure ages of basalt flows: Lunar Crater volcanic field, Nevada. Geology 23, 21-24.

Smith, E.I., Sánchez, A., Walker, J.D., Wang, K., 1999. Geochemistry of mafic magmas in the Hurricane volcanic field, Utah: Implication for small and large-scale chemical variability of the mantle. The Journal of Geology 107, 433-448.

Snow, J.E., Reisberg, L., 1995. Os isotopic systematics of the MORB mantle: results from altered abyssal peridotites. Earth and Planetary Science Letters 133, 411-421.

Snyder, D.C., Widom, E., Pietruszka, A.J., Carlson, R.W., 2004. The role of open-system processes in the development of silicic magma chambers: a chemical and isotopic investigation of the Fogo A trachyte, São Miguel, Azores. Journal of Petrology 45, 723-738.

Snyder, D.C., 2005. Processes and time scales of differentiation in silicic magma chambers: chemical and isotopic investigations. PhD dissertation, Miami University, 227p.

Spera, F.J., Fowler, S.J., 2009. Conceptual Model for small-volume alkali basalt petrogenesis: implications for volcanic hazards at the proposed Yucca Mountain Nuclear Waste Repository, in (editors) Connor, C.B., Chapman, N.A., Connor, L.J., Volcanic and tectonic Hazard Assessment for Nuclear Facilities, Cambridge University Press, Cambridge, UK.

Staudigel, H., Davies, G.R., Hart, S.R., Marchant, K.M., Smith, B.M., 1995. Large scale isotopic Sr, Nd and O isotopic anatomy of altered oceanic crust DSDP/ODP sites 417/418. Earth and Planetary Science Letters 130, 169-185.

Stickney, E.K., 2004. The volcanology and petrogenesis of the northern Lunar Crater Volcanic Field, Nye County, Nevada. M.S. thesis, University of Nevada-Las Vegas.

Stracke, A., Bizimis, M., Salters, V.J.M., 2003. Recycling oceanic crust: Quantitative constraints. Geochemistry Geophysics Geosystems 4, 33 pp. doi:10.1029/2001GC000223.

Stracke, A., Hofmann, A.W., Hart, S.R., 2005. FOZO, HIMU and the rest of the mantle zoo. Geochemistry Geophysics Geosystems 6, 20 pp. doi:10.1029/2004GC00824.

Sun, S.S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. In: Saunders, A.D., Norry, M.J. (Eds.), Magmatism in the ocean basins. Geological Society Special Publication, 313-345.

Tadini A., Bonali, F.L., Corazzato, C., Cortés, J.A., Tibaldi, A., Valentine, G.A., 2014. Spatial distribution and structural analysis of vents in the Lunar Crater Volcanic Field (Nevada, USA) Bulletin of Volcanology 76, Article No. 877.

Tatsumi, Y., Kogiso, T., 1997. Trace element transport during dehydration processes in the subducted oceanic crust: 2, Origin of chemical and physical characteristics in arc magmatism. Earth and Planetary Science Letters 148, 207-221.

Valentine, G.A., Gregg, T.K.P., 2008. Continental basaltic volcanoes – processes and problems. Journal of Volcanology and Geothermal Research 177, 857-873.

Valentine, G.A., Perry, F.V., Krier, D., Keating, G.N., Kelley, R.E., Cogbill, A.H., 2006. Small- volume basaltic volcanoes: Eruptive products and processes, and post-eruptive geomorphic evolution in Crater Flat (Pleistocene), southern Nevada. Geological Society of America Bulletin 118, 1313-1330.

Valentine, G.A., Cortés, J.A., 2013. Time and space variations in magmatic and phreatomagmatic eruptive processes at Easy Chair (Lunar Crater Volcanic Field, Nevada, USA). Bulletin of Volcanology, 75-752, doi:10.1007/s00445-013-0752-z.

Verma, S.P., Hasenaka, T., 2004. Sr, Nd, and Pb isotopic and trace element geochemical constraints for a veined-mantle source of magmas in the Michoacán-Guanajuato volcanic field, west-central Mexican Volcanic Belt. Geochemical Journal 38, 43-65.

Wang, K., Plank, T., Walker, J.D., Smith, E.I., 2002. A mantle melting profile across the Basin and Range, SW USA. Journal of Geophysical Research 107, 2017. doi:10.1029/2001JB000209

Weaver, B.L., 1991. The origin of oceanic island basalt end-member compositions: trace element and isotopic constraints. Earth and Planetary Science Letters 104, 381–397.

Wernicke, B., Davis, J.L., Niemi, N.A., Luffi, P., Bisnath, S., 2008. Active megadetachement beneath the western United States. Journal of Geophysical Research 113, B11409, 26 pp. doi:10.1029/2007JB005375.

West, J.D., Fouch, M.J., Roth, J.B., Elkins-Tanton, L.T., 2009. Vertical mantle flow associated with a lithospheric drip beneath the Great Basin. Nature Geoscience 2, 439–444.

Widom, E., 2011. Recognizing recycle osmium. Geology 39, 1087-1088.

Widom, E., 1997. Sources of ocean island basalts: a review of the osmium isotope evidence. Physica A: Statistical mechanics and its applications 244, 484–496.

Widom, E., Hoernle, K.A., Shirey, S.B., Schmincke, H.U., 1999. Os isotope systematics in the Canary Islands and Madeira: lithospheric contamination and mantle plume sources. Journal of Petrology 40, 279-296.

Yogodzinski, G.M., Smith, E.I., 1995. Isotopic domains and the area of interest for volcanic hazard assessment in the Yucca Mountain area: Eos (Transactions of the American Geophysical Union) 76, p. 669.

Yu, H., 2011. Li, Hf and Os isotope systematics of Azores basalts and a new microwave digestion method for Os isotopic analysis. PhD Dissertation, Miami University, 183p.

Zindler, A., Staudigel, H., Batiza, R., 1984. Isotope and trace element geochemistry of young Pacific seamounts. Implications for the scale of upper mantle heterogeneity. Earth and Planetary Science Letters 70, 175–195.

Zoback, M.L., Anderson, R.E., Thompson, G.A., 1981. Cenozoic Evolution of the State of Stress and Style of Tectonism of the Basin and Range Province of the Western United States. Philosophical Transactions of the Royal Society of London, Series A, 300, 407-434.

Figure 1. Regional map showing the location of the Lunar Crater Volcanic Field and the volcanoes of the Pancake Range sampled in this study, including the Marcath, Giggle Springs, Easy Chair, proto-Easy Chair, unnamed lava flows, Mizpah, and Hi Desert cones. Deposits associated with each cone and the respective sample localities are also shown (modified after Cortés et al., 2015 and Valentine and Cortés, 2013).

Figure 2. Major and trace element variation diagrams for the lavas from the northern LCVF. Lavas from Hi Desert, Mizpah, Unnamed flows and proto-Easy Chair (the HiMUp group) are shown in shades of red; lavas from Giggle Springs, Easy Chair and Marcath (the GEM group) are shown in shades of blue. Lavas from different volcanoes are shown with different shapes and/or colors, with the exception of the unnamed cones, which are all designated by dark pink diamonds.

Figure 3. Primitive mantle normalized trace element patterns for representative lavas from (a) the GEM group and (b) the HiMUp group. Normalizing values are from Sun and McDonough (1989). Average trace element abundances for HIMU basalts from Tubuai, St. Helena, and EMI basalts from Tristan da Cunha are from Stracke et al. (2003) and were compiled from the GEOROC database (http://georoc.mpch-mainz.gwdg.de).

Figure 4. (a) 87Sr/86Sr vs. 143Nd/144Nd plot of the LCVF volcanic rocks in comparison to other Cenozoic alkali-basalts from western United States, and the lower crustal Proterozoic granulites and eclogites from Geronimo Volcanic Field, southeastern Arizona (Kempton et al.1991). (b) εNd versus εHf diagram comparing the alkali-basalts of the northern LCVF to different types of OIB.

Box showing the Kaapvaal peridotites and Canadian Nikos peridotites represents a range of ε Hf and εNd characteristic of cratonic subcontinental lithospheric mantle. Fields for Basin and Range data are from Beard and Johnson (1997); Fields for the Tahoe-Truckee Volcanic Field (TTVF) are from Cousens et al. (2011); MORB and OIB data are from Stracke et al. (2003) and were compiled from the LDEO petrological database (htt://petdb.ldeo.columbia.edu/petdb) and GEOROC databases (http://georoc.mpch-mainz.gwdg.de). c) 206Pb/204Pb vs. 87Sr/86Sr diagram for the northern LCVF volcanic rocks. Field for the Proterozoic lower crustal eclogites and granulites from Kempton et al. (1990). Errors are within the symbol size except where ±2σ error bars are presented.

Figure 5. Plots of (a) 87Sr/86Sr, (b) 143Nd/144Nd, (c) 207Pb/204Pb and (d) 208Pb /204Pb versus 206Pb/204Pb showing samples from the northern LCVF. Basin and Range fields are compiled from Beard and Johnson (1997), Kempton et al. (1991), and Bradshaw et al. (1993); fields for the Tahoe-Truckee Volcanic Field (TTVF) are from Cousens et al. (2011). MORB and OIB fields including HIMU, EM1, EM2 and MORB are from the GEOROC database. Data from the Cook- Austral Islands, St. Helena and the recently redefined FOZO field are also shown (Stracke et al., 2005). Error bars are ±2σ for Pb isotopes; for Nd and Sr isotopes errors are within the symbol size.

Figure 6. 187Os/188Os vs. Os for LCVF lavas (>2 ppt Os samples only). AFC calculations are shown for depleted mantle-derived melt and upper and lower crustal rocks. Four AFC model curves are shown, two each illustrating lower crustal assimilation (solid lines) and upper crustal assimilation (dashed lines) by depleted mantle derived melts. The 187Os/188Os isotopic composition of the parental melt was chosen to represent that of an average depleted MORB mantle (e.g. Snow and Reisberg, 1995); the Os concentration is representative of a partial melt of peridotite, similar to the most primitive OIB. Os isotopic compositions of lower crustal xenoliths are from Esperança et al. (1997), upper crust compositions are from Oligocene tuff analyzed in this study, and the average upper crust is from Peucker-Ehrenbrink and Ravizza (2000). Typical

DOs values between 10-20 (Chesley and Ruiz, 1998; Lassiter and Luhr, 2001) were used for all AFC models; r (rate of assimilation to rate of crystallization) varies from 0.1 to 0.99. The compositions of the end-members are as follows: DMM derived melt: 187Os/188Os =0.127 (Meisel et al., 2001); Os = 350 ppt (Lassiter and Luhr, 2001); Upper crust: 187Os/188Os = 0.677- 1.2, Os = 22-30 ppt (this study and Peucker-Ehrenbrink and Ravizza, 2000); Lower crust: 187Os/188Os = 0.2-1.5, Os = 7-194 ppt (Esperança et al., 1997). Tick marks represent percent crustal assimilation.

Figure 7. 187Os/188Os vs. Nb/U plot of the samples from the northern LCVF. 187Os/188Os data for primitive mantle and DMM are from Meisel et al. (1996, 2001). Nb/U values for upper and lower crust are from Rudnick and Gao (2003). Error bars are within the size of symbols.

Figure 8. a) 206Pb/204Pb vs. 87Sr/86Sr diagram showing calculated mixing trends involving depleted MORB mantle (DMM), 0.8 Ga recycled oceanic crust and hydrous fluid. Modeling parameters are given in Appendix A and described in the text. The HiMUp mantle end-member can be modeled using a mixture of 18% 0.8 Ga recycled oceanic crust and 82% DMM; addition of <1% hydrous fluid into the HiMUp mantle source can produce the EM-like end-member. Tick marks indicate percentage of recycled oceanic crust and hydrous fluid. b) 187Os/188Os vs. 206Pb/204Pb diagram showing that the least radiogenic Os isotopic compositions of the HiMUp and GEM lavas are consistent with the model derived to explain the Pb and Sr isotope systematics; the more radiogenic signatures in each group are likely the result of minor lower crustal assimilation. The compositions of the end-members are as follows: Parental magma: 187Os/188Os =0.1275 (sample LC10-19), Os = 240 ppt (sample LC10-19), 206Pb/204Pb = 19.19, Pb =5; ROC: 187Os/188Os = 0.76, Os = 100 ppt (Widom et al., 1999), 206Pb/204Pb = 21.512, Pb= 0.06 ppm; Fluid: 187Os/188Os =1.4 (Borg et al., 2000), Os = 5 ppt, 206Pb/204Pb = 18.7 (Staudigel et al., 1995), Pb =5 ppm; Lower crust and AFC parameters: 187Os/188Os = 0.2-1.5, Os = 7-194 ppt (Esperança et al., 1997), 206Pb/204Pb = 19-19.4 (Kempton et al., 1990), Pb= 2.8 (Rudnick and

Gao, 2003), DOs = 15 (Chesley and Ruiz, 1998; Lassiter and Luhr, 2001), r = 0.1. Tick marks represent percent crustal assimilation.

Figure 9. Trace element modeling results for the HiMUp and GEM lavas and their respective sources. Modeling parameters are given in Appendix B. a) Trace element patterns of the HiMUp source and melt. The compositions of the HiMUp lavas are best reproduced by 0.5% non-modal batch melting of a mixture of 18% recycled oceanic crust and 82% DMM. b) Estimates of the trace element abundances of the GEM source and melt. Approximately 0.7% melting of the HiMUp reservoir with a minor amount (<1%) of hydrous fluid produces trace element patterns similar to those of GEM group lavas.

Figure 10. Schematic representation of the proposed model. First, ancient subducted oceanic crust was recycled into the asthenospheric mantle. This reservoir remained isolated for approximately 0.8 Ga and evolved to produce the HiMUp isotopic values. Low degree partial melting (0.5%) of this hybrid mantle at a depth of ~80 km produced the HiMUp trace element signatures. The GEM lavas were produced by slightly higher degree (~0.7%) and deeper melting of a similar HiMUp reservoir to which minor subduction-related hydrous fluid was added. The fluid addition could relate to subduction of the Farallon plate, although the timing of the fluid addition cannot be constrained. Variable degrees of mixing of the HiMUp and GEM sources can explain the range of isotopic signatures of the northern LCVF lavas. A potential systematic change with time in the depth and composition of the mantle source that is tapped in the production of LCVF magmas is illustrated.

Appendix A: Trace element and isotopic compositions of end-members used in the isotopic mixing models

Dehydrated Recycled Depleted Mantle MORB MORB MORB Fluid (present day) (present day) (0.8 Ga ago) (present-day)

Sr (ppm) 8 113 67c 67 3100 Rb (ppm) 0.06 1.2 0.2c 0.2 U (ppm) 0.002 0.07 0.069c 0.06 Pb (ppm) 0.02 0.3 0.06c 0.06 5 Os(ppt) 3100 5-900 100 100 5 87Rb/86Sr 0.054 0.054 0.054a 0.054 238U/204Pb 8.2 8.2 37 37 87Sr/86Sr 0.7027 0.7027 0.70266b 0.70328 0.7045 d 206Pb/204Pb 18.24 18.24 16.627 21.512 18.7d 187 188 Os/ Os 0.1275 0.1275 0.76 1.4

a Assumes that Rb gained through seawater alteration is removed from the oceanic crust during subduction. b Isotopic composition after seawater alteration using a 20-25% exchange rate as suggested by Staudigel et al. (1995). c Trace element abundances estimated based on the element mobilities of Kogiso et al. (1997) d Average Pacific MORB isotopic compositions. Determination of the composition of the recycled oceanic crust and slab hydrous fluid: In this model, the present-day isotopic values for the ancient recycled oceanic crust were calculated as a function of the age, the initial isotopic composition at the time of recycling, as well as the parent/daughter ratios Rb/Sr, U/Pb and Th/Pb of the dehydrated crust following the method of Stracke et al. (2003). The initial isotopic composition of the recycled oceanic crust was assumed to be similar to that of altered MORB at the time of recycling, and was back-calculated from the present day isotopic composition of altered N-MORB. Different parts of the subducting slab (e.g. altered basaltic crust, fresh basalt, gabbro and a slab composite) were tested to better characterize the potential contaminant, but among these, only altered basaltic crust could produce an appropriate 87Rb/86Sr to evolve to the measured 87Sr/86Sr of the northern LCVF lavas. The Pb isotopic evolution of the recycled crust was calculated using a µ value inferred from the measured U/Pb ratios of the older group samples (µ ranging from 33 to 40 with an average of

~37), which should represent the maximum possible 238U/204Pb for the mantle source. This µ value, combined with the measured 206Pb/204Pb and 207Pb/204Pb ratios of the most extreme HiMUp samples dictates the age of the recycled crust. For the slab-derived fluid component, we used 87Sr/86Sr and 206Pb/204Pb isotopic values for present day altered Pacific MORB (Staudigel et al., 1995). The trace element compositions of both recycled oceanic crust and hydrous fluid were calculated applying the element mobilities of Kogiso et al. (1997). For the Os isotopic compositions of the recycled oceanic crust, a chondritic initial 187Os/188Os of 0.09 (Shirey and Walker, 1998) and an average 187Re/188Os in the recycled crust of 50 (e.g. Widom et al., 1999) were used to determine the present day 187Os/188Os ratios of a 0.8 Ga old recycled oceanic crust. The hydrous fluid was assumed to have the present day Os isotopic values of Pacific N-MORB i.e. 187Os/188Os of 1.4 (Borg et al., 2000).

Appendix B. Trace element characteristics of the end-members used in trace element modeling.

Element Element Dehydrated Hydrous (ppm) DMMa MORBb mobilityc MORBd Fluide Bulk Df Cs (ppb) 0.73 14 51% 6.90 479 0.0003 Rb 0.06 1.26 63% 0.47 53 0.0003 Ba 0.63 14 53% 6.59 485 0.0001 Th 0.01 0.19 38% 0.12 4.74 0.0032 U (ppb) 2.98 71 76% 17 3602 0.0033 Nb 0.14 3.51 4% 3.37 9.35 0.0019 Ta 0.01 0.19 4% 0.18 0.51 0.0019 La 0.17 3.90 56% 1.71 145 0.0118 Ce 0.56 12 51% 5.88 408 0.0164 Pb 0.03 0.30 85% 0.07 28 0.0163 Nd 0.52 11 31% 7.71 231 0.0436 Sr 7.64 113 41% 67 3094 0.0286 Zr 6.25 104 0% 104 0.00 0.0512 Hf 0.16 2.97 0% 2.97 0.00 0.0648 Sm 0.22 3.75 14% 3.23 35 0.0854 Eu 0.09 1.34 8% 1.23 7.12 0.1291 Ti 623 9682 6% 9101 38728 0.1263 Gd 0.34 5.08 5% 4.82 17 0.1949 Dy 0.48 6.30 4% 6.05 17 0.2636 Y 3.94 36 2% 35 48 0.3461 Er 0.35 4.14 6% 3.89 16 0.3840 Yb 0.39 3.90 1% 3.86 2.60 0.6363 Lu 0.06 0.59 1% 0.58 0.39 0.6858

a Concentration of elements in depleted mantle from Salter and Stracke et al. (2004). b Concentration of MORB entering subduction zone, assuming that 87Sr/86Sr ratios are affected by alteration to 1600m in oceanic crust, whereas other geochemical indicators, particularly the mobile elements (e.g. Rb, U and Ba), are only minimally affected below 800m (e.g. Bach and Peucker-Ehrenbrink, 2003). c Element mobility: Percentage of element removed during sub-arc dehydration processes. d Dehydrated MORB: composition of MORB obtained after extraction of hydrous fluid. e Hydrous Fluid: composition of a slab derived hydrous fluid calculated based on the original MORB concentration and the element mobility assuming 1.5 wt% H2O in the MORB. Bulk partition coefficient D calculated using Ol:Opx:Cpx:Grt = 45:25:22:8.

CHAPTER 3

Temporal and Compositional Evolution of Jorullo Volcano, Mexico: Implications for Magmatic Processes Associated with a Monogenetic Eruption

Abstract:

The 1759-1774 eruption of the Jorullo volcano in the Michoacán Guanajuato Volcanic field (MGVF), Mexico, produced lavas that range in composition from basalt to basaltic andesite. We have conducted new major and trace element and isotopic studies (whole rock Sr-Nd-Pb-Hf-Os, and O isotopes in olivine separates) of the Jorullo lavas and tephras spanning the duration and compositional range of the eruption, to further constrain the potential roles of mantle source heterogeneity, subduction-related metasomatism, and crustal assimilation in the petrogenesis of the Jorullo magmas. This study presents the first Hf, high precision Pb and comprehensive oxygen isotope measurements for Jorullo volcanic rocks. All samples have arc-like trace element patterns with enrichments in large ion lithophile elements (e.g. Ba, Rb, and Pb) and depletions in fluid immobile elements (e.g. Nb, Ta). In addition, the samples show variations in 87Sr/86Sr (0.7038-0.7040), 143Nd/144Nd (0.51280-0.51285), 176Hf/177Hf (0.28297-0.28300), 206Pb/204Pb (18.62-18.66), 207Pb/204Pb (15.57-15.59) and 208Pb/204Pb (38.34-38.43). Osmium isotope signatures are, with one exception, more radiogenic than the depleted and primitive mantle 187 188 18 ( Os/ Os = 0.1231-0.1616). Oxygen isotope signatures of olivine phenocrysts (δ OSMOW ~5.62-5.97‰) show limited variation, but are isotopically heavier than normal mantle olivine. The samples define two geochemical groups: high-MgO samples with higher 87Sr/86Sr, lower 143Nd/144Nd and 176Hf/177Hf, and a positive correlation of Sr and Pb isotopes; and low-MgO samples displaying lower 87Sr/86Sr but higher 143Nd/144Nd and 176Hf/177Hf than the former group, and a negative correlation of Sr and Pb isotopes. The high-MgO group comprises most of the early tephra and lavas, whereas the low-MgO group includes most of the late tephra and lavas. These compositional variations are inconsistent with shallow level contamination, but rather are interpreted to reflect mantle source heterogeneity. Trace element and isotopic signatures are consistent with North Mexican Extensional Province (NMEP) mantle metasomatised by subduction components composed of sediment- and oceanic crust-derived hydrous fluid. The temporal-compositional variations observed in Jorullo magmas are inferred to result from a combination of variable degrees of fractional crystallization of magmas produced by tapping a progressively less metasomatised mantle source that is vertically and/or laterally heterogeneous.

1. Introduction

Small-volume, mafic magmatic systems (<1 km3) produced by single episodes of volcanic activity are referred to as monogenetic volcanic fields (Connor and Conway, 2000). They tend to occur as cinder cones, maars, tuff cones and lava domes. The duration of an individual eruption is typically shorter than that of composite volcanoes (e.g. stratovolcanoes), lasting from several days to years (Connor and Conway, 2000), but the volcanic field as a whole may span millions of years. Because of their small volume and mafic character, they are often considered to ascend very rapidly from their mantle source, and thus provide a key to understanding the magmatic processes that are often masked in larger magmatic systems. Nonetheless, they often show significant compositional variation within individual eruptions as well as among closely spaced eruptive centers. The origin of this compositional variation has been the source of considerable debate. Some studies suggest that the compositional variation reflects heterogeneity in the mantle source (e.g. Reiner et al., 2002; Siebe et al., 2004; Blondes et al., 2008), whereas others suggest that it is the result of assimilation of lithospheric mantle or crust during ascent (Lassiter and Luhr, 2001; Chesley et al., 2002, Siebe et al., 2004). Furthermore, questions remain about the origin of source heterogeneity including the role of subduction derived fluid and melts. Distinguishing subduction signatures from shallow level crustal assimilation poses a challenge, as both processes can produce very similar chemical and isotopic signatures, and both processes may work in concert to produce compositional variations and temporal-compositional trends observed in some basaltic monogenetic eruptions and volcanic fields (Wallace and Carmichael, 1999; Chesley et al., 2002; Siebe et al., 2004). A related issue is the question of whether monogenetic systems develop sustained magma chambers, or whether magmas erupt more-or- less directly from their mantle sources. The study of the petrogenesis of Jorullo volcano and its satellites cones provides a unique opportunity to evaluate the origin of compositional variation within individual monogenetic centers, as well as magmatic processes operating in subduction systems in general. Jorullo volcano is located within the Michoacán Guanajuato Volcanic Field (MGVF), which represents the volcanic front of the Trans-Mexican Volcanic Belt (TMVB). Previous workers have analyzed Jorullo samples for major and select trace elements (e.g. Luhr and Carmichael 1985, Lassiter and Luhr, 2001), and Sr-Nd-Pb and Os isotopes (e.g. Lassiter and Luhr, 2001; Chesley et al., 2002;

Verma and Hasenaka, 2004), and (for one sample only) oxygen isotopes (Johnson et al., 2009), but none covered the full duration or compositional range of the eruption. This study presents the first Hf isotope and high precision Pb isotope data for Jorullo samples, as well as an extensive new data set of Sr, Nd, Os and oxygen isotope analyses spanning the duration of the Jorullo eruption. We use the elemental and isotopic data for the Jorullo lavas and tephra to address the following aspects of their petrogenesis: 1) the influence of crustal assimilation; 2) the types and nature of the subduction components being added to the mantle wedge (sediment versus oceanic crust; fluid versus melt); and 3) the cause of the systematic compositional variations with time in the Jorullo eruption.

2. Tectonic Setting

The TMVB is an active volcanic arc that trends in an E-W direction, and is approximately 200 km wide and 1,000 km long. It contains more than 8,000 eruptive vents including stratovolcanoes, calderas, domes and monogenetic cinder cones. On the basis of available geochronological data, volcanic activity within the belt has occurred since the Miocene (Ferrari et al., 1999). Magmatism is generally accepted to be associated with the subduction of the Cocos and Rivera Plates beneath the North American Plate along the Middle America Trench (Fig.1). The oblique orientation of the TMVB relative to the Middle American Trench is attributed to near-horizontal subduction of the Cocos plate beneath central Mexico between 100 and 250 km from the trench, and steep subduction of the Rivera plate to the west (e.g. Pardo and Suarez, 1995). Ferrrari (2004) proposed that volcanism has migrated trenchward at a rate of ~10 km/Ma over the past 2 Ma, due to rollback of the subducted Cocos plate. Volcanism within the belt is compositionally diverse, as both calc-alkaline and intra-plate-type magmas have erupted across the volcanic belt, and in many places they are coeval. The calc-alkaline signatures are often considered to be derived from partial melting of the subarc mantle modified by slab melts or slab dehydration (e.g. Siebe et al., 2004; Blatter et al., 2010; Cai et al., 2014; Straub et al., 2015). In contrast, the OIB-type magmas have been variously interpreted to represent melts of the unmodified subarc asthenospheric mantle (e.g. Luhr, 1997; Siebe et al., 2004; Luhr et al., 2006), melts of pyroxenitic mantle produced by subduction fluid/melt infiltration (Straub et al., 2008, 2015), or a deep mantle plume (Márquez et al., 1999b; Verma, 2000). Despite the controversy of

78 the origin of the OIB-type magmas, clear subduction signatures are present in most samples and calc-alkaline magmas are volumetrically dominant compared to the OIB-type magmas. The MGVF represents one of several distinct monogenetic volcanic fields in the TMVB, and is thought to represent the volcanic front of the TMVB (Hasenaka and Carmichael, 1985). Based on hypocenter relocation of local earthquakes, the volcanic field is thought to lie >80 km above the subducted Cocos plate (Pardo and Suárez, 1995), and as such, has been regarded as a key area to study the relationship between central Mexican volcanism and subduction of the Cocos plate (e.g. Verma et al., 2002). The MGVF contains more than 1,000 eruptive centers, most of which are located about 200-300 km from the trench (Hasenaka and Carmichael, 1985). Two historic eruptions have occurred within the MGVF, including Jorullo (1759-1774) and Paricutin (1943-1952) volcanoes. The volcanic field consists mostly of monogenetic scoria cones, lava domes, shield volcanoes and maars. In addition, both alkaline and calc-alkaline compositions have erupted throughout the field (Luhr, 1997, 2006). The origin of these compositionally distinct magmas has been the subject of extensive petrological and geochemical studies. It has been suggested that the coexistence of calc-alkaline and intraplate alkaline volcanism in close proximity implies that they come from distinct domains in the mantle (Luhr, 1997) in which the intra-plate alkaline basalts are produced by partial melting of a convecting upper mantle uncontaminated by subduction components, whereas the calc-alkaline rocks are derived by melting of subarc mantle wedge that has been metasomatised by slab-derived materials (Luhr, 1997). Some studies have also invoked the involvement of crustal contamination during petrogenesis (Lassiter and Luhr, 2001; Chesley et al., 2002). Other studies (e.g. Verma et al., 2004) have attributed the origin of the MGVF calc-alkaline rocks in general to partial melting of veined metasomatised mantle sources and suggested that the MGVF might not be related to the subduction of the Cocos plate, but rather to the ongoing rifting processes within the TMVB (e.g. Verma et al., 2004). In summary, despite the abundance of trace element and isotopic data available, the origin of the magmatism and the petrogenetic processes operating within the MGVF remain inconclusive.

3. Jorullo Eruption and Samples

Jorullo volcano and its satellite cones (Volcán del Norte, an unnamed cone, Volcán de Enmedio, and Volcán del Sur) erupted over a fifteen-year period from 1759 to 1774, and produced pyroclastic deposits and lava flows totaling a volume of approximately 2 km3 (Luhr and Carmichael, 1985). Historical accounts reported that the early part of the eruption (from 1759 to ~1764) consisted of pyroclastic activity and covered the surrounding area with wet ash fall deposits (Luhr and Carmichael, 1985). The effusion of most of the lava flows occurred between 1760 and 1766 (Gadow, 1930). It was also reported that the pyroclastic activity continued throughout the eruption, and ash fall layers blanket most of the lava flows (Gadow, 1930; Luhr and Carmichael, 1985). Luhr and Carmichael (1985) were the first to document the geology and the compositional evolution of Jorullo lavas over time. They subdivided the lava flows into three stages: early, middle and late. They noted that the Jorullo lavas became progressively more silica-rich over the course of the eruption. This compositional variation was attributed primarily to fractional crystallization of olivine-plagioclase-augite and minor spinel at variable depths within the lower crust to upper mantle (Luhr and Carmichael, 1985). More recent field mapping has revealed evidence for at least seven distinct eruptive phases. Based on the new mapping, the earliest lava flows (stages 1 and 2) are thought to have been emitted from multiple vents along a fissure to the southwest of the main Jorullo cone. Phase three corresponds to the collapse of the South, Middle and North cones and emplacement of related deposits, and lavas from phases four to seven were emitted primarily from the main cone. Pyroclastic activity occurred throughout the duration of the eruption, but mostly between stages three and four (after collapse of satellite cones), with minor explosive activity after stage seven, emitted primarily from the main cone. We sampled lavas from all seven phases of the eruptive sequence in order to investigate geochemical and isotopic variations through time during the eruption (Fig. 2). In addition, we excavated two sites located <1 km to the south and southeast of the main cone, and sampled tephra deposits that encompass the full range of explosive activity at Jorullo. The lowermost part of the tephra sequence was sampled to the southeast of the main cone and the upper and middle parts of the sequence were recovered just south of the cone. The tephra deposits largely preserve their original stratigraphy and consist mostly of alternating layers of interbedded fine ash and lapilli. Several samples of the local basement rock (the La

Huacana granitoid), which occurs as xenoliths in the Jorullo eruptive products, were also collected.

4. Sample Preparation and Analytical Techniques

Mineral compositions of olivine and oxide inclusions were determined by wavelength- dispersive X-ray emission microanalysis using a CAMECA SX50 electron microprobe at the Department of Geological Sciences at Indiana University. Conditions of measurement were 15 keV accelerating voltage, 20 nA beam current, ~1 μm beam size and peak counting time of 20 s. The accuracy of the analyses were monitored using an international standard (San Carlos olivine, USNM 1113122/444). For whole rock major element, trace element and isotopic analyses, we used fresh samples free of xenoliths and alteration. Samples were first cut into thin slabs, ground with silicon- carbide sand paper to remove any metal traces from the rock saw, and cleaned rigorously with 18

MΩ H2O in an ultrasonic bath. The samples were dried in an oven overnight at 110⁰ C, then crushed and powdered in a high purity alumina shatterbox. For the Jorullo lavas and tephra, major elements and a subset of trace elements (Ni, Cr, Sc, V, Ga, Cu and Zn) were analyzed by X-ray fluorescence spectrometry (XRF) at the Geoanalytical Laboratory at Washington State University, and all other trace element measurements were conducted by inductively coupled plasma mass spectrometry (ICP-MS), also at Washington State University. Details of the sample preparation and analytical procedures are provided on the Geoanalytical Laboratory website http://environment.wsu.edu/facilities/geolab/). For the basement granitoid samples, major element data are from the UNAM dataset (analysis performed by Activation Laboratories, Ancaster, Canada), and trace element data were obtained at Miami University by ICP-MS, following the procedures described in Yu (2011). Sr, Nd and a subset of the Pb isotope analyses were performed at Miami University.

Approximately 100 mg of sample powder was dissolved in HF-HNO3, and separations of Pb, Sr and Nd from a given sample aliquot were performed by ion exchange chromatography following standard procedures described in detail in Snyder et al. (2005). Isotopic measurements were performed on a Thermo-Finnigan Triton thermal-ionization mass spectrometer (TIMS) at Miami University. Measured 87Sr/86Sr and 143Nd/144Nd isotopic ratios by TIMS were corrected for mass

81 fractionation by normalizing to 86Sr/88Sr =0.1194 and 146Nd/144Nd= 0.7219. A mass fractionation correction of 0.10% per amu was used for 206Pb/204Pb, 207Pb/204Pb, 208Pb/204Pb measured by TIMS. External precision based on long-term two standard deviation (2 SD) reproducibility of the NBS 987 and La Jolla reference materials gave ±0.000015 and ±0.000007 for 87Sr/88Sr and 143Nd/144Nd respectively. The TIMS 2 SD external reproducibility of the NBS 981 standard reference material on 206Pb/204Pb, 207Pb/204Pb, and 208Pb/204Pb are ±0.014, ±0.019, and ±0.063, respectively. Higher precision Pb isotopic measurements were obtained for most samples on separate sample dissolutions. These measurements were performed on a Nu Plasma multi- collector inductively coupled plasma mass spectrometer (MC-ICP-MS) at the Department of Terrestrial and Magnetism (DTM), Carnegie Institution of Washington, using 205Tl/203Tl ratios to correct for instrumental mass bias. Measured 204Hg/202Hg was used to correct for the isobaric interference of 204Hg on 204Pb. The 2 SD external reproducibility of the NBS 981 standard reference material during the analytical campaign was ±0.0018 for 206Pb/204Pb, ±0.0021 for 207Pb/204Pb and ±0.0059 for 208Pb/204Pb. For Hf isotopic analysis, chemical separation of Hf was conducted at Miami University, and isotopic measurements were performed by MC-ICP-MS using the Nu Plasma at DTM. Hf dissolution and purification methods followed the procedures of Yu (2011) and Connelly et al. (2006), described briefly below. Approximately 150 mg of powder was dissolved, and Hf was then purified by a two-step column method utilizing first BioRad AG50W-X8 resin and then Eichrom TODGA resin (50-100um). The first column removes major elements and rare earth elements (REE), and the high field strength elements (HFSE) are eluted. The second purification column separates Hf from the other HFSE. Over the course of the MC-ICP-MS runs, regular analysis of standard reference material JM 475 was performed to correct for mass bias. All samples were normalized to a 176Hf/177Hf value for JM 475 of 0.282160. The average measured 176Hf/177Hf for JM 475 during the period of data collection was 0.282151, with a reproducibility of ±0.000007. Os chemistry and mass spectrometry were performed at Miami University. Os separations were carried out using the Carius tube digestion method of Shirey and Walker (1995). Os purification was accomplished through two distillation steps, including a macro-distillation following the procedures of Yu (2011) modified from the method of Nägler and Frei (1997), and a micro-distillation using the procedure of Roy-Barman (1993) and Roy-Barman and Allègre

(1995). Os isotopic ratios and concentrations were measured by negative ion TIMS (N-TIMS). Samples were loaded onto single Pt filaments with a Ba-hydroxide activator, and measured as - OsO3 molecular ions by peak hopping on a single secondary electron multiplier. All measured Os isotope ratios were corrected for mass fractionation using 192Os/188Os = 3.0826, and corrected for oxygen isotopic composition using the values of Nier (1950). The external reproducibility for 187Os/188Os measurements based on the long-term reproducibility of a NIST Os solution standard is ±0.0002. However, internal errors for 187Os/188Os measurements of the samples were variable and ranged from 0.03-0.15%. Olivine separates for oxygen isotope measurements were prepared at Miami University, and were free of cracks and alteration, and contained <1% oxide inclusions. Olivines were separated from their host lavas and tephras by crushing and sieving to a fraction of ~200 µm and handpicked from the sieved fraction under a binocular microscope. The handpicked grains were further crushed with an alumina mortar and pestle and passed through a Frantz magnetic separator to remove crystal fragments with high concentrations of oxide inclusions. To ensure removal of mineral fractions with oxide inclusions or coatings of matrix, the separates were further handpicked under a binocular microscope. The separates were then cleaned three times in an ultrasonic bath with 18.2 MΩ H2O. Oxygen isotope measurements were performed at the University of Wisconsin on a Finnigan MAT 251 mass spectrometer, following the methods outlined in Valley et al. (1995). Approximately 2 to 3 mg of olivine was loaded into the source chamber, which was pre-fluorinated prior to analysis. During the analyses, focusing and monitoring of the laser was conducted to avoid sample cross contamination. The Gore Mountain garnet standard UWG-2 was run with the samples to monitor the precision and accuracy of the analyses. The reproducibility of the UWG-2 garnet standard obtained during the period of analyses was 5.79±0.12‰ (2 SD). The δ18O values of the samples are reported relative to V- SMOW.

5. Results

5.1 Petrography and mineral chemistry

The petrography and mineralogy of the Jorullo lavas, tephra and granitoid inclusions have been described in detail (Luhr and Carmichael, 1985; Guilbaud et al., 2011). Jorullo lavas analyzed in this study are porphyric with euhedral and skeletal phenocrysts of olivine, and microphenocrysts of olivine, plagioclase, and augite. Groundmass textures are subophiotic to interstitial with plagioclase, augite, olivine and spinel. Plagioclase occurs exclusively as groundmass phases. Spinel inclusions are present in most olivine phenocrysts. Olivine phenocrysts are mostly normally zoned with core to rim Fo contents ranging from 90 to 80 in the early lavas, and 87 to 79 in the late lavas (Appendix 1). Inverse zoning was not observed in the olivines analyzed. The Fo contents of the microphencrysts are similar to those of the rims of the phenocrysts. Most of the olivines have compositions in or close to equilibrium

FeMg with their host, using K D of 0.3 ± 0.03 of Roeder and Emslie (1970).

5.2 Major and trace elements

Major and trace element data from this study are presented in Table 1. Chemical analyses for other Jorullo samples previously analyzed at Activation Laboratories, Ancaster (UNAM dataset) are provided in Appendix 2. The Jorullo tephra and lavas range in composition from relatively primitive basalt to basaltic andesite, with SiO2 contents between 52 and 55% and MgO ranging from 4.7 to 10.3%. Tephra samples encompass the full compositional range of the eruption, and mostly fall within the range of the lavas. As observed by Luhr and Carmichael (1985), the

Jorullo lavas generally became richer in SiO2 as the eruption progressed, although based on the new mapping (Guilbaud et al., 2011) the lavas appear to reflect a somewhat more complex compositional evolution in SiO2 and MgO with time (Fig. 3a). Likewise, the tephra sequences do not exhibit a strictly monotonic variation in SiO2 or MgO content with stratigraphic height (Fig. 3b), especially among those with high MgO contents (>8%). Nonetheless, the new data are in general agreement with the previous findings that the early tephra and lavas are relatively primitive, and the late eruptive products more evolved.

Clear correlations between major oxides are also apparent (Fig. 4). SiO2, Al2O3, Na2O,

K2O increase with decreasing MgO, whereas CaO and FeO(t) decrease with decreasing MgO. In contrast, TiO2 shows variable behavior: the samples with MgO >8% tend to have lower TiO2 contents (0.7 - 0.8%) and show negative correlation with MgO, whereas the samples with MgO

84 contents <8% exhibit higher TiO2 (0.92 - 0.98%) and display positive correlation with MgO. Samples with high MgO (>8%) are also characterized by relatively high Ni (~200 - 270 ppm) and Cr (~500 - 585 ppm), whereas samples with low MgO contents (<8%) extend to lower Ni (~46 - 150 ppm) and Cr (~85 - 290 ppm). For simplicity, the two sample groups are henceforth referred to as the high-MgO (>8%) and low-MgO (<8%) groups. Trace elements also exhibit correlations with MgO. Incompatible elements such as La, Ba, and Sr, exhibit negative correlations with MgO, as does La/Yb (Fig. 5a-f), whereas compatible trace elements such as Ni and Cr (and Sc, not shown) exhibit positive correlations with MgO. All samples have low Ce/Pb (<10) and Nb/U (<20) relative to MORB and OIB (Fig. 5g, h), but similar to those of island arc basalts and continental crust. As a whole, the Jorullo samples do not display any clear correlation between Ce/Pb or Nb/U and MgO, although the low-MgO group samples exhibit a slight decrease in Ce/Pb and Nb/U with decreasing MgO. All Jorullo samples are marked by slight enrichment in light rare earth elements (LREE) but depletion in heavy rare earth elements (HREE) relative to normal mid-oceanic basalts (N-MORB). In addition, on a MORB-normalized multi-element diagram (Fig. 6), the samples exhibit typical arc trace element patterns with enrichments of fluid-mobile large ion lithophile elements (LILE) such as Cs, Rb, K, Ba, Pb and Sr, and depletions in high-field-strength elements (HFSE) such as Nb and Ta. However, the samples exhibit little to no anomaly in Hf concentration (Hf/Hf* ~0.96 - 1.04, where Hf/Hf* = HfN/[NdN+SmN)/2]; Wade et al., 2005).

5.3 Isotopes

The Sr-Nd-Pb-Hf-Os and oxygen isotope compositions of the Jorullo samples are listed in Table 1. Jorullo samples display significant ranges in 87Sr/86Sr, 143Nd/144Nd, 176Hf/177Hf and Pb isotopes (Fig. 7), but exhibit a relatively limited range in 187Os/188Os, and have essentially 18 87 86 143 144 uniform  Oolivine (Fig. 8). Sr/ Sr ranges from 0.70376 - 0.70402, Nd/ Nd from 0.512804 - 0.512853, and 176Hf/177Hf from 0.282969 - 0.282999. Pb isotope ratios vary as follows: 206Pb/204Pb = 18.618 - 18.659, 207Pb/204Pb = 15.572 - 15.591 and 208Pb/204Pb = 38.336 - 38.425. 187Os/188Os ratios range from 0.1231 - 0.1616 over a range in Os concentrations from 6 - 173 ppt, but samples with >10 ppt Os range only from 0.1231 - 0.1407. All samples except JOR-1290E are characterized by more radiogenic 187Os/188Os than depleted MORB mantle and primitive

85 upper mantle values. Likewise, olivine δ18O values (full range 5.70-6.02‰; most ~5.8 ± 0.1‰) are all higher than the oxygen isotope compositions typical of depleted MORB mantle olivine (5.2 ± 0.2‰; e.g. Mattey et al., 1994; Eiler, 2001). In general, the Sr, Nd, Pb, Os and oxygen isotopic compositions are similar to those determined in previous studies (e.g. Lassiter and Luhr, 2001; Chesley et al., 2002; Verma and Hasenaka, 2004; Johnson et al., 2009), and fall within the isotopic ranges of the MGVF and other volcanic rocks from the volcanic front of the TMVB. On 207Pb/204Pb vs. 206Pb/204Pb and 208Pb/204Pb vs. 206Pb/204Pb diagrams (Fig. 7), the Jorullo samples plot above the North Hemisphere Reference Line (NHRL; Hart 1984). They are compositionally intermediate between local subducting terrigenous and pelagic sediments from DSDP Site 487, but significantly more radiogenic than the subducting oceanic crust and East Pacific Rise (EPR) basalts. All samples fall along the terrestrial mantle-crust array in Hf-Nd isotope space, and are characterized by lower Hf and Nd isotope signatures than DMM and basalts from DSDP site 487, but higher Hf and Nd isotope signatures than the local subducting terrigenous and pelagic sediments. The Jorullo samples define an overall negative correlation between 87Sr/86Sr and 143Nd/144Nd (Fig. 9a), although in detail, the high-MgO group forms a slightly steeper array than that of the low-MgO samples. Both trend toward the compositions of the local granitic basement and Mexican lower crustal xenoliths. However, the high-MgO samples have the most radiogenic 87Sr/86Sr and lowest 143Nd/144Nd, while the low-MgO samples are offset towards lower 87Sr/86Sr and higher 143Nd/144Nd isotopic compositions. Distinctions between the low- and high-MgO groups can also be observed on a 206Pb/204Pb versus 87Sr/86Sr diagram (Fig. 9b), in which they form a negative and positive correlation, respectively. However, one low-MgO sample (JOR- 0766), which represents the final stage of eruption at Jorullo, has similar trace element and isotopic characteristics to those of high-MgO group samples. None of the isotope systems correlate with indices of differentiation such as MgO or Ni, and no correlation is observed between 187Os/188Os and the other isotope systems.

6. Discussion

On the basis of the new major element, trace element and Sr-Nd-Pb-Hf isotope data presented in this study, the two sample groups defined above (high-MgO and low-MgO groups)

86 can be recognized as distinct geochemical groups. The high-MgO group comprises the early tephra and lavas and is characterized by more primitive (Ni >200 ppm and Cr >500 ppm) compositions, and higher 87Sr/86Sr but lower 143Nd/144Nd and 176Hf/177Hf isotope ratios. The low-MgO group includes the late tephra and lavas and is characterized by more evolved compositions (Ni <200 ppm, Cr <500 ppm), lower 87Sr/86Sr and higher 143Nd/144Nd and 176Hf/177Hf ratios than the high-MgO samples. The Jorullo samples show a broad negative correlation in 87Sr/86Sr vs. 143Nd/144Nd (Fig. 9a), with the high-MgO group forming an array with a steeper slope than the low-MgO group. Similarly, on the 206Pb/204Pb vs. 87Sr/86Sr diagram (Fig. 9b), the high-MgO and low-MgO samples form positive and negative correlations, respectively. Possible models to explain the temporal-compositional changes and the observed isotopic variations of the eruptive products include: 1) fractional crystallization with variable crustal assimilation; and 2) tapping of a heterogeneous mantle source in which different components predominate throughout the eruption.

6.1 Fractional crystallization

Major and trace element trends of Jorullo tephra and lavas are generally compatible with closed system fractionation. Marked increases in SiO2 and Al2O3 and the decrease in Ni and Cr with decreasing MgO (Fig. 4) are consistent with fractionation of olivine. The general decrease in CaO/Al2O3, CaO, and Sc concentrations (not shown) with decreasing MgO are further indicative of clinopyroxene fractionation. In addition, the change from increasing to decreasing

TiO2 in the high- and low-MgO groups, respectively, indicates that oxides became important in the fractionating mineral assemblage at MgO <8%. Although Luhr and Carmichael (1985) proposed fractionation of olivine + clinopyroxene + plagioclase + minor spinel to explain the major and trace element variations of the products of Jorullo, the lack of decrease in Al2O3 or Sr with decreasing MgO (Figs. 4, 5), and the lack of negative Eu anomalies (Fig. 6), suggest that plagioclase was not a major fractionating phase, in accord with the petrographic observation that plagioclase phenocrysts are absent in most of the samples. Johnson et al. (2008) further suggested a role for amphibole fractionation based on major element modeling. Although amphibole is observed in the phenocryst assemblage only in the very most evolved Jorullo lavas (e.g. JOR-0766), it may have been segregated from residual magmas at depth as has been

87 commonly proposed for arc magmas (“cryptic fractionation”, e.g. Davidson et al., 2007; Johnson et al., 2008), and/or it may be a relatively late stage addition to the fractionating mineral assemblage. The increasing La/Yb with decreasing MgO (Fig. 5f; caused by increasing La with essentially constant Yb) is consistent with the relative Kd’s for amphibole, which preferentially incorporates HREE relative to LREE (Blundy and Wood, 2003; Davidson et al., 2007). Crystallization of amphibole from the low-MgO group magmas would indicate fractionation at mid-crustal depths (Davidson et al., 2007; Johnson et al., 2008). The increase in highly incompatible element abundances such as La, Ba, Nb, Zr, Rb with decreasing MgO is consistent with closed system fractional crystallization of one or more similar parental magmas, although variable degrees of partial melting of a common source or assimilation-fractional crystallization (AFC) might alternatively explain these trends. Variable degrees of partial melting can fractionate elements with different compatibilities, such as La/Yb, thus the increase in La/Yb with decreasing MgO (Fig. 5f) could potentially be explained by such a mechanism. However, the variable Ce/Pb and Nb/U cannot be explained by this mechanism due to the identical partition coefficients during mantle melting for Ce-Pb and Nb-U, respectively (Hofmann et al., 1986). Fractionation of amphibole, which preferentially incorporates Pb relative to Ce and Nb relative to U (Romick et al., 1992; GERM database http://earthref.org/GERM/) could explain the slight decrease in Ce/Pb and Nb/U with decreasing MgO among the low-MgO group samples. Nevertheless, the very low Ce/Pb and Nb/U for both the low- and high-MgO group Jorullo samples relative to MORB and OIB (Figs. 5g, h) requires crustal involvement in the petrogenesis of the Jorullo magmas, whether by shallow level contamination or by crustal recycling into the mantle.

6.2 Crustal assimilation

Previous geochemical and isotopic studies have proposed respectively that either upper or lower crustal contamination is an important process in the petrogenesis of the MGVF volcanic rocks including Jorullo (e.g. Lassister and Luhr, 2001; Chesley et al., 2002; Johnson et al., 2008). Likewise, several aspects of our new data could potentially be indicative of crustal assimilation by mantle melts, including the negative correlation of 87Sr/86Sr and 143Nd/144Nd, low Ce/Pb and 187 188 18 87 86 Nb/U ratios, suprachondritic Os/ Os ratios, and high δ Oolivine. The higher Sr/ Sr and

88 lower 143Nd/144Nd in the high-MgO samples relative to the low-MgO samples are, however, opposite that expected by progressive assimilation-fractional crystallization (AFC) of a single evolving magma. Although this could reflect significant crustal contamination of the high-MgO samples, and lesser contamination of the low-MgO samples, the high-MgO samples with Sr and Nd isotopic signatures closest to those of the crustal basement comprise the most primitive magmas observed in Jorullo (e.g. Ni >200 ppm, Cr >500 ppm), which is inconsistent with these magmas having undergone extensive differentiation or assimilation. In addition, within the high- MgO group, no correlations are observed between indices of fractionation (e.g. MgO, Ni) and potential indices of contamination (e.g. Ce/Pb, Nb/U, or Sr-Nd isotopes), as would be expected if AFC processes were important in their petrogenesis. Although Johnson et al. (2008) inferred based on major and trace element (e.g. La, Zr and Y) systematics of olivine melt inclusions that shallow assimilation of the La Huacana granitic bedrock played an important role in the compositional variations of Jorullo magmas, the Jorullo samples do not trend towards the local granitic crust in Nd and Hf isotope space, but rather towards lower 176Hf/177Hf signatures (Fig. 7). Prior studies of the MGVF have suggested alternatively that lower crustal assimilation may play a role in their petrogenesis, based on a correlation of 187Os/188Os with indices of fractionation (e.g. MgO and Ni; Lassiter and Luhr, 2001; Chesley et al., 2002). For Jorullo samples, no correlation between 187Os/188Os and MgO or Ni is observed, either within the sample suite as a whole or within the high-MgO or low-MgO subgroups (Fig. 8a). Rather, the Jorullo samples have relatively low and constant 187Os/188Os compared to other MGVF samples, despite the similarly large range in Ni contents (46 - 270 ppm). AFC modeling using the assimilant composition and all other parameters from the Chesley et al. (2002) study fail to reproduce the essentially constant 187Os/188Os of the Jorullo samples including those with low Ni (Fig. 8a). The 18 δ Oolivine signatures of the Jorullo samples also are essentially homogenous within the 2 SD measurement error (with only one exception, 5.82 ± 0.12‰; Fig. 8b), further arguing against a significant role for lower (or upper) crustal assimilation as the cause for the variability in chemical and radiogenic isotope compositions. Likewise, the Jorullo samples lie essentially parallel to the MORB-OIB array on Hf-Nd isotope space, inconsistent with significant assimilation of the low Hf (deviation from the MORB-OIB array) signatures of Mexican lower

89 crust (Fig. 7). Therefore, we consider the chemical and isotopic variations more likely to be a result of mantle source heterogeneity, for which we explore the petrogenetic implications below.

6.3 Mantle source heterogeneity

Jorullo lavas and tephra define two distinct trends in 87Sr/86Sr - 206Pb/204Pb space (Fig. 9b), with the high-MgO samples forming a positive trend and the low-MgO samples a negative trend, intersecting at low 206Pb/204Pb and intermediate 87Sr/86Sr. The two trends are also evident in 87Sr/86Sr - 143Nd/144Nd space (Fig. 9a), in which the high-MgO group forms a negative trend with a slightly steeper slope than that of the low-MgO group. These observed trends indicate that at least three components contributed to Jorullo magma petrogenesis. Trace element characteristics of both the low-MgO and high-MgO sample groups, including LILE enrichments and HFSE depletions, strongly suggest the involvement of subduction-related hydrous fluid in their 18 petrogenesis. Heavy δ OOlivine in Jorullo (5.82 ± 0.12‰) compared to that of the depleted mantle (5.2 ± 0.2‰; Mattey et al., 1994; Eiler, 2001) suggests significant involvement of sediment and/or uppermost (low-temperature) altered oceanic crust in the Jorullo mantle. We therefore explore below whether the heterogeneous trace element and radiogenic isotope signatures in the Jorullo magmas could be generated by addition of subduction components to a homogeneous pre-subduction mantle wedge. The pre-subduction mantle composition beneath the TMVB is most often interpreted to be similar to the mantle source of the mafic alkaline volcanism that occurs behind the arc in the Northern Mexican Extensional Province (referred to as NMEP-type mantle by Luhr et al., 2006). The NMEP, known also as the Eastern Alkaline Province or Mexican Basin and Range, is a broad extensional region that extends from near the USA-Mexico border southward to 21° N latitude, where it intersects the TMVB (Luhr et al., 2006). The NMEP back-arc mantle is thought to advect into the sub-arc environment by slab-induced convection (e.g. Siebe et al., 2004). Calc- alkaline magmas, volumetrically dominant throughout the TMVB, as well as more rare lamprophyre magmas, are interpreted to be derived from partial melting of subduction modified NMEP mantle (Luhr, 1997; Wallace and Carmichael, 1999; Siebe et al., 2004; Luhr et al., 2006). Although the Jorullo samples all have higher 87Sr/86Sr than NMEP, the low-MgO sample group trends towards NMEP compositions in Sr-Pb and Sr-Nd isotope space (Fig. 9), suggesting that

90 the source of the Jorullo samples could potentially be produced by addition of subduction components to the NMEP source. The two trends defined by the low-MgO and high-MgO groups could therefore be explained by input of one subduction component (SC1) characterized by lower 206Pb/204Pb but higher 87Sr/86Sr into the NMEP mantle to generate the source of the low-MgO group, and subsequent addition of another subduction component (SC2) with higher 87Sr/86Sr and 206Pb/204Pb to the already metasomatised NMEP mantle to produce the source of the high-MgO group. In order to investigate which subduction components and processes might best produce the range of trace element and isotopic compositions observed in Jorullo, we have conducted modeling of the Jorullo trace element and isotope data using the inferred composition of the NMEP mantle source and the compositions of subducting sediment and altered basaltic oceanic crust (AOC) dredged from the DSDP site 487, assumed to represent the materials being subducted beneath the TMVB (e.g. Gómez-Tuena et al., 2003; Gómez-Tuena et al., 2007, Cai et al., 2014). The NMEP mantle source is assumed to have a depleted mantle (DMM) trace element composition, and isotopic signatures within the range of basalts from the NMEP (Appendix 3). The sediment column from DSDP site 487 includes ~105 m of primarily hemipelagic or terrigenous sediment, overlying ~70 m of late Miocene to Pliocene pelagic sediment, which in turn overlies ~15 Ma basaltic oceanic crust (Verma, 2000). The isotopic compositions of these potential subduction components are highly variable, and the AOC and sediments generally bracket the range of compositions of the Jorullo samples (Fig. 7). In the modeling presented here, we have estimated the compositions of possible components that could be released from the subducting Cocos plate using the isotopic and elemental compositions of locally subducting materials, as well as experimentally determined mobilities for trace elements during subduction dewatering and melting. The chemical and isotopic compositions of the sediment and basaltic oceanic crust have been reported by Verma et al., 2000, Gómez-Tuena et al., 2003; Lagatta, 2003, Gómez-Tuena et al., 2007, and Cai et al., 2014. The trace element composition of oceanic crust-derived fluid was calculated by using the mobility data of Kessel et al. (2005) at 700⁰ C and 4 GPa, which are appropriate for dehydration of metabasalt, and for sediment-derived fluid and melt we have used the D values at 650⁰ C and 900⁰ C at 2 GPa from the datasets of Johnson and Plank (1999). The total extraction of hydrous fluid during dehydration was assumed to be 1.5 wt.% in both oceanic crust and sediment, based on the results of high-pressure dehydration

91 experiments on amphibolite (Tatsumi and Kogiso, 1997). All parameters used in the model are given in Appendix 3. Simple mass balance, assuming an NMEP mantle wedge with a normal δ18O of 5.2‰, and the isotopically heaviest potential subduction component (sediment with δ18O of 20‰; e.g. Kolodny et al., 1976; Muehlenbachs, 1986;), requires that ≥3% subduction-derived oxygen be added to the Jorullo mantle source to generate the lowest measured δ18O of 5.7‰. A likely maximum amount of subduction component can be estimated assuming addition of pure AOC- derived fluid from the low-temperature altered portion of the subducting upper oceanic crust, which has an average δ18O signature of 10‰ (Staudigel et al., 1995); this would require that no more than 13% subduction fluid be added to the Jorullo mantle source. In order to further evaluate the amount and relative importance of the potential subducted materials that might contribute to the source of the Jorullo magmas, we first consider the isotopic signatures of the most fluid mobile elements, Sr and Pb. Because the AOC from DSDP site 487 is rather unradiogenic in 87Sr/86Sr (~0.7032), a high proportion of sediment relative to AOC is required to produce a subduction component sufficiently high in 87Sr/86Sr to generate the radiogenic signatures of the low-MgO group Jorullo samples, and especially those of the high- MgO group samples. Strontium isotopes do not, however, distinguish between sediment types, as the subducting terrigenous and pelagic sediment are both comparably radiogenic in 87Sr/86Sr (Fig. 7). Pb isotope signatures, on the other hand, will be strongly controlled by the type of sediment involved, as the terrigenous sediment is significantly more radiogenic than the Jorullo samples, and the pelagic sediment substantially less so. Figure 10a shows that in order to reproduce the Jorullo low-Mg group sample trend in Sr-Pb isotope space, mixing between an NMEP mantle source and fluid derived from ≥78% sediment (90:10 terrigenous to pelagic) and ≤22% AOC is required to produce the radiogenic Sr isotope signatures, and accommodate addition of ≥4% subduction component (i.e. ≥3% sediment). If the sediment component were dominated by pelagic rather than terrigenous sediment, then an even greater fraction of the subduction component would have to be sediment-derived to produce the Sr and Pb isotope signatures of the low-MgO group samples. However, in the latter scenario, <4% subduction component (i.e. <3% sediment) can be accommodated, which would be at odds with the minimum subduction component required to reproduce the heavy δ18O signatures.

The Hf-Nd isotope systematics are also particularly sensitive to the relative proportions of terrigneous and pelagic sediment, hence can provide further constraints on the composition of the subduction component (Fig. 10b). The relatively steep trend of the Jorullo samples (approximately parallel to the MORB-OIB array), combined with the strongly concave downward curvature for mixing between NMEP and subduction fluid or melt, clearly requires a sediment component that is dominated by terrigenous sediment with a minimal contribution of pelagic sediment, in agreement with the Sr-Pb isotope systematics. Together, the δ18O and the Sr-Pb-Nd-Hf isotope systematics of the low-MgO group Jorullo samples are well reproduced by mixing between an NMEP mantle source and 4% of a subduction component SC1 that is characterized by ~80% sediment-derived fluid (90:10 terrigenous:pelagic) and ~20% AOC- derived fluid (Fig. 11). For the high-MgO group samples, the situation is more complex, because mixing involving unmodified NMEP mantle wedge, sediment and altered oceanic crust cannot directly produce their trend in Sr-Pb isotope space. However, input of a second slab component (SC2) consisting of higher 87Sr/86Sr and 206Pb/204Pb to the already metasomatised mantle source of the low-MgO samples can explain the high-MgO group. In this case, a subduction component with a slightly higher fraction of terrigeneous relative to pelagic sediment (95:5) compared to that of the low- MgO mantle source (90:10), but again comprising ~80% sediment- and 20% slab-derived fluid, can explain the distinct trends in Sr-Pb isotope space (Fig. 11a). Based on our modeling results, addition of ~1% of SC2 to the most metasomatised part of the mantle source of the low-MgO group can produce the range of Sr, Nd, Pb, Hf isotopic compositions of the high-MgO Jorullo group with a minimal increase in the δ18O values (Fig. 11). The mixing scenario developed above based on oxygen and radiogenic isotope systematics also produces a good fit to most of the trace element data (Fig. 12). In this model, we assume an initially unmetasomatised NMEP mantle source with typical upper mantle mineralogy of 55% olivine, 32% orthopyroxene, 10% clinopyroxene, and 3% garnet (e.g. Carter, 1970), and depleted mantle trace element abundances. The presence of garnet in the source region is indicated by the high Dy/Yb (1.8 to 2) values of the Jorullo samples relative to N-MORB (Dy/Yb~1.5). The trace element patterns of the low-MgO group samples can be reproduced by ~9% partial melting of the SC1-metasomatised NMEP mantle source. However, as shown in Fig 12a, this model (dictated by the oxygen isotope requirement of ≥4% subduction fluid addition) predicts higher

Cs, Rb and Pb concentrations than observed. Likewise, the three-component mixing model involving NMEP and composite subduction components SC1 and SC2 also produces acceptable fits to most of the trace element data. In general, the trace element abundances of the high-MgO magmas can be generated by ~15% partial melting of the SC1- and SC2-metasomatised NMEP mantle (Fig.12b), although the model again predicts higher Cs, Rb and Pb concentrations than observed. Similar issues for Rb (Johnson et al., 2009) and Ba, U and Pb (Gómez-Tuena, 2007) have been noted previously in other studies of the TMVB. Straub et al. (2015) recently proposed that the recycled crustal component may be controlled by subduction erosion of the upper plate, in which there are known crustal lithologies (e.g. granodiorite and biotite gneiss) with relatively low Rb, Ba and Pb concentrations, although the terrigenous sediment from site DSDP 487 is also most likely derived from the crystalline basement of the Mexican margin (Plank and Langmuir, 1998). Given the uncertainties in the experimentally determined partition coefficients, the discrepancies in Cs, Rb and Pb concentrations in the present model relative to the Jorullo data could indicate that the partition coefficients (Dslab/fluid) utilized in the models are not appropriate; the Jorullo Cs and Rb data are best approximated if the Dslab/fluid is ~3-fold higher than those reported by Johnson and Plank (1999) at 2 GPa and 700º C. It is nevertheless notable that the three elements that exhibit significant discrepancy between the model and the Jorullo data are among those that are generally considered to be the most highly fluid mobile elements (e.g. Weaver, 1991; Tatsumi and Kogiso, 1997). Therefore, it is likely that a fraction of the sediment budget of Cs, Rb and Pb would be lost during low temperature, low pressure dewatering in the forearc (e.g. Brocher et al., 2003; Schmidt and Poli, 1998). An alternative possibility is that the assumption of ≥3% sediment addition in the subduction component added to the mantle wedge beneath Jorullo, based on measured heavy δ18O signatures in the Jorullo olivines, is incorrect. In Appendix 4 (Appendix Figs. 1 and 2) we demonstrate that a similar model, in which the NMEP mantle is fluxed by a lesser amount (~1%) of SC1 and SC2, followed by lower degrees of partial melting (~3.5% and 6%, versus 9% and 15%) can also explain the chemical and radiogenic isotopic signatures of Jorullo samples, without the discrepancies in Cs, Rb and Pb concentrations. Such a model may be applicable if the δ18O signature of the pre-subduction mantle wedge beneath Jorullo is already isotopically heavy, as has been proposed for the Cascades. Martin et al. (2011) postulated that elevated δ18O

94 values (to 6.08‰) in olivines from high magnesium olivine tholeiites and basaltic andesites from Mt. Shasta and Medicine Lake volcanoes might be attributed to a slab window that allows flow of previously enriched ancient forearc mantle into the mantle wedge beneath Cascades (e.g. Zandt and Humphreys, 2008). Such a process could also potentially explain the elevated δ18O observed in the Jorullo magmas. There is evidence for the Rivera-Cocos slab tear extending behind and to the eastern edge of the MGVF (e.g. Ferrari et al., 2004), and perhaps more directly relevant, a slab tear has been proposed along the projection of the Orozco fracture zone, just to 18 the east of Jorullo (Dougherty et al., 2012). However, Johnson et al. (2009) reported δ Oolivine of 5.5‰ for the Hoya Alvarez alkaline magma, located behind the MGVF volcanic front and minimally affected by subduction processes (<1.4% H2O); if such mantle comprises the pre- subduction mantle wedge beneath Jorullo, then this argues against a significantly heavy preexisting δ18O signature. We therefore favor a model in which the oxygen isotope signatures as well as the radiogenic isotope and trace element signatures of the Jorullo magmas are dominated by the current subduction, with initial dewatering of sediment (and partial loss of sediment Cs, Rb and Pb) in the forearc, followed by continued sediment dewatering and dehydration of the AOC and mantle wedge serpentinite beneath Jorullo. We note that aspects this model are in good agreement with those derived for the MGVF by independent means. The obtained range in degree of partial melting is in good agreement with the results of Johnson et al. (2009), which were based on volatile and trace element compositions of melt inclusions, and the total slab contribution to the mantle wedge (~4-5%) is broadly consistent with the water contents of Jorullo magmas (5-6%) and other MGVF melt inclusions (Johnson et al., 2009), as well as for estimates of subduction contributions to arcs globally (e.g. McCulloch and Gamble, 1991; Elliot et al., 1997).

6.4 Slab dehydration versus slab melting beneath MGV: implications from thermal structure and geochemical data

The geochemical models presented here suggest that mantle metasomatism is dominated by sediment-derived hydrous fluid rather than altered oceanic lithosphere-derived fluid, and that no slab melt component is required. Although many subduction zones are characterized by oceanic-

95 crust dominated fluid (e.g. Elliot et al., 1997), the geometry and thermal structure of the slab beneath the MGVF may favor sediment dewatering as the dominant contributor to the subduction component. Jorullo volcano is located approximately 240 km from the trench, which, based on seismic evidence, corresponds to a depth to slab of ~50-80 km (Pardo and Suaréz, 1995; Peréz-Campos et al., 2008; Kim et al., 2012). This slab depth, with an associated slab surface temperature of ~600-700 ºC (Johnson et al., 2009) would lead to complete sediment dewatering, but is shallower than the depth (~120-130 km) at which dehydration of the oceanic crust would be complete (Johnson et al., 2009; Manea and Manea, 2010), consistent with a lessor contribution of oceanic crust-derived fluid to the subduction component beneath Jorullo. Although some previous studies have suggested that slab melting occurs beneath the TMVB (e.g. Gomez-Tueña et al., 2007; Cai et al., 2014), our trace element and isotopic model for the Jorullo volcanic rocks is consistent with slab dehydration and does not require any slab melt. Based on the Johnson et al. (2009) estimated slab surface temperature of ~600-700 ⁰C, neither sediment nor oceanic crust melting would be predicted (e.g. Johnson and Plank, 1999; Johnson et al., 2009). Mantle wedge temperatures above the volcanic front, estimated to be 1200-1300 ⁰C, are 40 to 50 ⁰C below the dry peridotite solidus (Johnson et al., 2009), but fluid fluxing from the slab and breakdown of mantle wedge serpentinite at these pressure and temperature conditions would induce melting (Johnson et al., 2009).

6.5 Physical model & chronology of metasomatism, magma generation, and eruption

Based on our mixing models, we suggest that the Jorullo magmas are derived from melting of a heterogeneous mantle source produced by subduction modification of an NMEP-type mantle wedge. The origin of the mantle sources of the low and high-MgO groups can be explained by a multi-step, three component mixing process as illustrated in Fig 13. Initially, mantle previously melted to produce the intraplate alkaline basalts of the NMEP was advected into the sub-arc environment by corner flow. This convecting NMEP upper mantle was then infiltrated by a H2O- rich component (SC1) derived through dehydration of the subducting Cocos plate (terrigenous sediment >> pelagic sediment >> AOC), which produced the mantle source of the low-MgO Jorullo samples.

Influx of a second subduction component (SC2), which differed only by a slightly higher contribution of terrigenous to pelagic sediment (95% vs. 90%), was subsequently added to the metasomatised source of the low-MgO group samples to produce the source of the high-MgO group samples. The minor difference in the relative proportions of the sediment types in SC1 versus SC2 could be attributed to heterogeneity in the sediment pile and/or to incomplete homogenization of a heterogeneous sediment-derived fluid. Alternatively, Patino et al. (2000) proposed that a horst and graben structure in the subducting plate could lead to variability in sediment proportions because the upper terrigenous unit may be scraped off a horst into an adjacent graben during subduction, resulting in a higher proportion of terrigenous sediment in the graben structure. Regardless, it is envisioned that this H2O-rich SC2 component infiltrated the already metasomatised mantle, forming the mantle source for the high-MgO group. Estimates of the depth of melt extraction using the silica-activity barometer of Putirka (2008) indicates that both the high-MgO and low-MgO group melts were extracted from similar depths (~1.4-1.6 GPa; Appendix 5), suggesting that the mantle source region is heterogeneous on a small spatial scale, laterally and/or vertically, beneath Jorullo. The temporal-compositional variation identified within the prolonged monogenetic eruption of Jorullo can therefore be attributed in part to changes in mantle source over time, and can be summarized as follows. Melting of the NMEP mantle metasomatised by both SC1 and

SC2, which was presumably the more metasomatised and water-rich mantle wedge region, produced the early high-MgO group. Previous studies (Luhr and Carmichael, 1985; Johnson et al., 2008) suggest that the early melts rose from depth and stalled at the lower crust-upper mantle boundary where they underwent initial fractionation; however, based on the major and minor element data, the early erupted magmas are relatively primitive, suggesting minimal storage prior to eruption. Nevertheless, the variations in MgO with stratigraphic height in the tephra sequence (Fig. 3b) indicate that the high-MgO group magmas may have been produced in at least two batches, the first of which underwent minor fractionation prior to a recharge event involving a second pulse of mafic magma generation from the same mantle source. The new mapping of the Jorullo area (Guilbaud et al., 2011) suggests that these early lava flows were emitted from fissure vents rather than solely from the main Jorullo cone. Evidence from major and trace element data suggest that some of the early melt underwent extensive fractionation before eruption, likely in a shallow magma chamber, and produced the

97 very last eruptive phase at Jorullo (the amphibole-bearing sample JOR 0766). This late phase of the eruption had major and trace element characteristics of the low-MgO group, but isotopic signatures similar to those of the high-MgO group, arguing for a shared mantle source and early magma generation despite its late eruption. Exhaustion of the more metasomatised, water-rich mantle lead to melting of the less metasomatised mantle (that received only SC1) with time, producing the parental magmas for the low-MgO group. The late Jorullo melts rose from depth, and likely stalled in a magma chamber above or around that of JOR-0766, where they underwent extensive fractionation. Previous studies have suggested that a shallow magma chamber developed as the eruption of Jorullo progressed. The evidence for amphibole fractionation in the low-MgO group magmas suggests that the fractionation occurred in part at mid-crustal depths, but melt inclusion volatile data suggest that pre-eruptive crystallization continued to very shallow crustal depths, possibly <1 km (Johnson et al., 2008). Nevertheless, the isotopic and trace element data obtained in this study indicate that stalling and fractionation in a shallow magmatic plumbing system did not promote significant crustal assimilation. These low-MgO melts erupted mainly from the main Jorullo cone, as did the final stage of the eruption, which consisted of the eruption of the earlier formed JOR-0766 magma. Together, the detailed field and geochemical studies of the complete Jorullo eruptive cycle have provided new insight into the origin of the chemical heterogeneity observed in this single monogenetic eruption. Our results suggest that the mantle is heterogeneous on a small spatial scale beneath the volcano, and that sequential tapping of this variably hydrated mantle, produced by at least two distinct slab-derived fluid fluxes, led to the production of two distinct parental magmas. Although previous models for monogenetic eruptions that become more silicic with time have invoked progressive tapping of pyroxenite-veined mantle (e.g. Reiners 2002; Straub et al., 2011) to explain correlations between major element composition and isotopic indicators of source, our results suggest that distinct magma batches with similar initial major and trace element compositions evolved separately within a complex magmatic plumbing system. This process led to the apparently fortuitous relationship in which sequentially more evolved magmas erupted with time from a progressively less metasomatised source. Only the final eruption of the JOR-0766, which links the source of the most evolved, latest erupted magma to that of the least

98 evolved, earliest erupted magmas, allows us to distinguish between these models for compositionally zoned monogenetic eruptions.

Acknowledgments Thanks are owed to Dave Kuentz for assistance in the field and TIMS analyses at Miami University, and to Rick Carlson for assistance with the high-precision Pb and Hf isotope measurements at DTM. This work was supported by NSF EAR grant # 1019798 awarded to Widom and a GSA student grant awarded to Rasoazanamparany. Field and laboratory costs of Salinas, Valdés, Guilbaud, and Siebe were defrayed from projects funded by the Consejo Nacional de Ciencia y Tecnología (CONACyT-167231 and 152294) and the Dirección General de Asuntos del Personal Académico UNAM-DGAPA IN-101915 and 105615 granted to C. Siebe and M.-N. Guilbaud. The inhabitants and authorities of la Huacana and Mata de Plátano are thanked for their friendliness and support.

References

Blatter, D.L., Hammersley, L., 2010. Impact of the Orozco Fracture Zone on the central Mexican Volcanic Belt. J. Volcanol. Geotherm. Res. 197, 67–84.

Blondes, M.S., Reiners, P.W., Ducea, M.N., Singer, B.S., Chesley, J., 2008. Temporal compositional trends over short and long time-scales in basalts of the Big Pine Volcanic Field, California. Earth Planet. Sci. Lett. 269, 140-154.

Blundy, J., Wood, B., 2003. Partitioning of trace elements between crystals and melts. Earth Planet. Sci. Lett. 210, 383–397.

Brenan, J., Shaw, H., Ryerson, F., Phinney, D. 1995. Mineral-aqueous fluid partitioning of trace element at 900 oC and 2.0 GPa: constraints on the trace element chemistry of mantle and deep crustal fluids. Geochim. Cosmochim. Acta 59, 3331-3350.

Cai, Y., LaGatta, A., Golden, S.L., Langmuir, C.H., Gómez-Tuena, A., Pozzo, A.L.M., Carrasco- Nuñez, G., 2014. Hafnium isotope evidence for slab melt contributions in the Central Mexican Volcanic Belt and implications for slab melting in hot and cold slab arcs. Chem. Geol. 377, 45–55.

Chesley J., Ruiz J., Righter K., Ferrari L., Gómez-Tuena A., 2002. Source contamination versus assimilation: an example from the Trans-Mexican Volcanic Arc. Earth Planet. Sci. Lett. 195, 211–221.

Connelly, J.N., Ulfbeck, D.G., Thrane, K., Bizzarro, M., Housh, T., 2006. A method for purifying Lu and Hf for analyses by MC-ICP-MS using TODGA resin. Chem. Geol. 233, 126– 136.

Connor, C.B., Conway, F.M., 2000. Basaltic volcanic fields. In: Sigurdsson, H. (Ed.), Encyclo- pedia of Volcanoes. Academic Press, New York, pp. 331–343.

100

Davidson, J. P., Turner, S., Handley, H., Macpherson, C. Dosseto, A., 2007. Amphibole ‘sponge’ in arc crust? Geology 35, 787-790.

Dougherty, S.L., Clayton, R.W. Helmberger, D.V. 2012. Seismic structure in central Mexico: implications for fragmentation of the subducted Cocos plate. Journal of Geophysical Research 117, B09316, doi: 10.1029/2012JB009528.

Eiler, J.M., Schiano, P., Valley, J.W., Kita., Stolper, E.M., 2001. Oxygen-isotope and trace element constraints on the origins of silica-rich melts in the subarc mantle. Geochem. Geophys. Geosyst., 8, 9, doi:10.1029/2006GC001503.

Elliot, T., Plank, T., Zindler, A., White, W., Bourdon B., 1997. Element transport from subducted slab to juvenile crust at the Mariana Arc. J. Geophys. Res. 102 14991-15019.

Ferrari, L., López-Martínez, M. , Aguirre-Díaz, G., Carrasco-Núñez, G., 1999. Space-time patterns of Cenozoic arc volcanism in central Mexico: From the Sierra Madre Occidental to the Mexican Volcanic Belt, Geology 27, 303–306.

Ferrari, L., 2004. Slab detachment control on mafic volcanic pulse and mantle heterogeneity in central Mexico. Geology 32, 77-80.

Gadow, H., 1930. Jorullo: The history of the volcano Jorullo and the reclamation of the devastated district by animals and plants. London: Cambridge University Press. p. 101.

Gómez-Tuena, A., LaGatta A, Langmuir, C., Goldstein S., Ortega-Gutiérrez, F., Carrasco- Núñez, G., 2003. Temporal control of subduction magmatism in the Eastern Trans-Mexican Volcanic Belt: mantle sources, slab contributions and crustal contamination. Geochem. Geophys. Geosyst. 4, 9. http://doi:10.1029/2003GC000524

101

Gómez-Tuena, A., Langmuir, C.H., Goldstein, S.L., Straub, S.M., Ortega-Gutiérrez, F., 2007. Geochemical Evidence for Slab Melting in the Trans-Mexican Volcanic Belt. J. Petrol. 48, 537- 562.

Guilbaud, M-N, Siebe, C., Layer, P., Salinas, S., Castro-Govea, R., Garduño-Monroy, V-H., Le Corvec, N., 2011. Geology, geochronology, and tectonic setting of the Jorullo Volcano region, Michoacán, México. J. Volcanol. Geotherm. Res., 201, 97-112.

Hart, S., 1984. A large-scale isotope anomaly in the Southern Hemisphere mantle, Nature, 309, 753–757.

Hasenaka, T., Carmichael, I.S.E., 1985. The cinder cones of Michoacan-Guanajuato, central Mexico: their age, volume and distribution, and magma discharge rate. J. Volc. Geotherm. Res. 25, 105-124.

Hofmann, A. W., Jochum, K. P., Seufert, M. White, W. M., 1986. Nb and Pb in oceanic basalts: New constraints on mantle evolution, Earth Planet. Sci. Lett. 79, 33-45.

Johnson, E.R., Wallace, P.J., Cashman, K.V., Delgado-Granados, H., Kent, A.J.R., 2008. Magmatic volatile contents and degassinginduced crystallization at Volca¤n Jorullo, Mexico: Implications for melt evolution and the plumbing systems of monogenetic volcanoes. Earth Planet. Sci. Lett. 269, 477-486.

Johnson, E.R., Wallace, P.J., Granados, H.D., Manea, V.C., Kent, A.J.R., Bindeman, I.N., Donegan, C.S., 2009. Subduction-related volatile recycling and magma generation beneath central Mexico: insights from melt inclusions, oxygen isotopes and geodynamic models. J. Petrol. 50, 1729-1764.

Johnson, M.C., Plank, T., 1999. Dehydration and melting experiments constrain the fate of subducted sediments. Geochem. Geophys. Geosyst. 1, 1999GC000014.

102

Kessel, R., Schmidt, M.W., Ulmer, P., Pettke, T., 2005. Trace element signature of subduction- zone fluids, melts and supercritical liquids at 120-180 km depth. Nature 437, 724-727.

Kim, Y., Miller, M.S., Pearce, F., Clayton, R.W., 2012 Seismic imaging of the Cocos plate subduction zone system in central Mexico. Geochem. Geophys. Geosyst. 13, 7, doi:10.1029/2012GC004033

Kolodny, Y., and Epstein, S., 1976. Stable isotope geochemistry of deep sea cherts. Geochim. Cosmochim. Acta, 40, 1195-1209.

LaGatta, A.B, 2003. Arc Magma Genesis in the Eastern Mexican Volcanic Belt. PhD thesis Columbia University, New York.

Lassiter, J., Luhr, J., 2001. Osmium abundance and isotope variations in mafic Mexican volcanic rocks: Evidence for crustal contamination and constraints on the geochemical behavior of osmium during partial melting and fractional crystallization, Geochem. Geophys. Geosyst. 2 (2000GC000116)

Luhr, J., Carmichael, I., 1985. Jorullo Volcano, Michoacán, Mexico (1759–1774): the earliest stages of fractionation in calc-alkaline magmas. Contrib. Mineral. Petrol. 90, 142–161. Luhr, J., 1997. Extensional tectonics and the diverse primitive volcanic rocks in the western Mexican Volcanic Belt. Can. Mineral., 35, 473–500.

Luhr, J.F., Kimberly, P., Siebert, L., Aranda-Gómez, J.J., Housh, T.B., Kysar Mattietti, G., 2006. México's Quaternary volcanic rocks: insights from the MEXPET petrological and geochemical database. Geol. Soc. Am. Spec. Pap. 402, 1-44.

Manea, V.C., Manea, M., 2010. Flat-slab thermal structure and evolution beneath Central Mexico. Pure Appl. Geophys. 168, 1475-1487.

103

Márquez, A., Oyarzún, R., Doblás, M. Verma, S.P., 1999. Alkalic (ocean-island basalt type) and calc-alkalic volcanism in the Mexican volcanic belt: A case for plume-related magmatism and propagating rifting at an active margin? Geology 27, 51-54.

Martin, E., Bindeman, I.N., Grove, T.L., 2011. The origin of high-Mg magmas in Mt Shasta and Medicine Lake volcanoes, Cascade Arc (California): higher and lower than mantle oxygen isotope signatures attributed to current and past subduction. Contrib. Mineral. Petrol. 162, 945- 960.

Mattey, D., Lowry, D., Macpherson, C., 1994. Oxygen isotope composition of mantle peridotite, Earth Planet. Sci. Lett., 128, 231–241.

McCulloch, M.T., Gamble, J.A., 1991. Geochemical and geodynamical constraints on subduction zone magmatism. Earth and Planetary Science Letters 102, 358-374.

Meisel, T., Walker, R.J., Morgan, J.W., 1996. The osmium isotopic composition of the primitive upper mantle. Nature 383, 517–520.

Muehlenbachs, K., 1986. Alteration of the oceanic crust and the 18O history of seawater. In: Valley, J. W., Taylor, H. P. & O’Neil, J. R. (eds) Stable Isotopes in High Temperature Geological Processes, Vol. 16 . Washington, DC: Mineralogical Society of America, 425–444

Nägler, Th.F., Frei, R., 1997. “Plug in” Os distillation. Schweiz. Mineral. Petrogr. Mitt. 77, 123– 127.

Nier, A.O., 1950. A redetermination of the relative abundances of the isotopes of carbon, nitrogen, oxygen, argon and potassium. Phys. Rev. 77, 789–793.

Pardo, M., and Suárez, G., 1995. Shape of the subducted Rivera and Cocos plate in southern Mexico: Seismic and tectonic implications, J. Geophys. Res. 100, 12357–12373.

104

Patino, L.C., Carr, M.J., Feigensen, M.D., 2000. Local and regional variations in Central American arc lavas controlled by variations in subducted sediment input. Contrib. Mineral. Petrol. 138, 265-283.

Pérez-Campos, X., Kim, Y., Husker, A. , Davis, P. M. , Clayton, R. W. , Iglesias, A., Pacheco, J. F. , Singh, S. K., Manea, V. C., Gurnis, M., 2008. Horizontal subduction and truncation of the Cocos Plate beneath central Mexico, Geophys. Res. Lett., 35, L18303, doi:10.1029/2008GL035127.

Plank, T., Langmuir, C.H., 1998. The chemical composition of subducting sediment and its consequences for the crust and mantle. Chemical Geology 145, 325-394.

Putirka, K.D., 2008. Thermometers and barometers for volcanic systems. Reviews in Mineralogy and Geochemistry 69, 61-120.

Reiners, P.W., 2002.Temporal-compositional trends in intraplate basalt eruptions: implications for mantle heterogeneity and melting processes. Geochem. Geophys. Geosyst. 3, doi:10.1029/2001GC000250.

Roeder, P. E., Emslie, R. F.,1970. Olivine-liquid equilibrium. Contrib. Mineral. Petrol. 29, 275- 289.

Romick, J. D., Kay, S. M., Kay, R.W., 1992. The influence of amphibole fractionation on the evolution of calc-alkaline andesite and dacite tephra from the central Aleutians, Alaska.Contributions to Mineralogy and Petrology 112:101-118.

Roy-Barman, M., 1993. Mesure du raport 187Os/188Os dans les basalts et les peridotites: contribution a la systematique 187Re-188Os dans le manteau, PhD thesis, Université de Paris.

105

Roy-Barman, M., Allègre, C.J., 1995. 187Os/186Os in oceanic island basalts: tracing oceanic crust recycling in the mantle. Earth Planet. Sci. Lett. 129, 145–161.

Schaaf, P., Heinrich, W., Besch, T., 1994. Composition and Sm–Nd isotopic data of the lower crust beneath San-Luis-Potosi, Central Mexico — evidence from a granulitefacies xenolith suite. Chem. Geol. 118, 63–84.

Siebe, C., Rodríguez-Lara, V., Schaaf, P., Abrams, M., 2004. Geochemistry, Sr-Nd isotope composition, and tectonic setting of Holocene Pelado, Guespalapa and Chichinautzin scoria cones, south of Mexico City. Journal of Volcanology and Geothermal Research 130, 197- 226.

Shirey, S.B., Walker, R.J., 1995. Carius tube digestion for low-blank rhenium–osmium analysis, Anal. Chem. 67, 2136–2141.

Schmidt, M.W., Poli, S., 1998. Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation. Earth Planet. Sci. Lett., 163, 361-379.

Snow, J.E., Reisberg, L., 1995. Os isotopic systematics of the MORB mantle: results from 1062 altered abyssal peridotites. Earth Planet. Sci. Lett. 133, 411–421.

Snyder, D.C., 2005. Processes and time scales of differentiation in silicic magma chambers: chemical and isotopic investigations. (PhD dissertation), Miami University (227p).

Staudigel, H., Davies, G. R., Hart, S. R., Marchant, K. M, Smith, B. M., 1995. Large scale isotopic Sr, Nd and O isotopic anatomy of altered oceanic crust: DSDP/ODP sites 417/418. Earth Planet. Sci. Lett. 130, 169–185.

Straub, S.M., Gómez-Tuena, A., Stuart, F.M., Zellmer, G.F., Espinasa-Pereña, R., Cai, Y., Iizuka, Y., 2011, Formation of hybrid arc andesites beneath thick continental crust. Earth Planet. Sci. Lett. 303, 337–347.

106

Straub , S.M., Gómez-Tuuena, A., Zellmer, G.F., Espinasa-Pereña, R., Stuart, F.M., Langmuir, C.H., Martin-del Pozzo, A.L., Mesko, G.T., 2012. The processes of melt differentiation in arc volcanic rock: Insights from OIB-type arc magmas in the Central Mexican Volcanic Belt. J. Petrol. 54, 665-7011.

Straub, S.M., Gómez-Tuena, A., Bindeman, I.N., Bolge, L.L., Brandl, P.A., Espinasa-Pereña, R., Solari, L., Stuart, F.M., Vannucchi, P., Zellmer, G.F., 2015. Crustal recycling by subduction erosion in the central Mexican Volcanic Belt. Geochimica et Cosmochimica Acta 166, 29-52.

Sun, S.S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes, Magmatism in the ocean basins, Geological Society Special Publication 42. pp. 313–345.

Tatsumi, Y., and Kogido, T., 1997. Trace element transport during dehydration processes in the subducted oceanic crust: 2. Origin of chemical and physical characteristics in arc magmatism, Earth Planet. Sci. Lett., 148, 207-221.

Valley, J. W., Kitchen, N., Kohn, M. J., Niendorff, C. R., Spicuzza, M. J. 1995. Strategies for high precision oxygen isotope analysis by laser fluorination. Geochimica et Cosmochimica Acta 59, 5223–5231.

Verma, S.P, 2000. Geochemistry of the subducting Cocos plate and the origin of subduction- unrelated mafic volcanism at the front of the central Mexican Volcanic Belt. In Delgado- Granados H, Aguirre-Díaz G, Stock J (Eds.). Cenozoic Tectonics and Volcanism of Mexico Special Papers Geological Society of America 334, 1–28.

Verma, S., 2002. Absence of Cocos Plate subduction-related basic volcanism in southern Mexico: a unique case on Earth? Geology 30, 1095-1098.

107

Verma, S.P., Hasenaka, T., 2004. Sr, Nd, and Pb isotopic and trace element geochemical constraints for a veined-mantle source of magmas in the Michoacán–Guanajuato volcanic field, west-central Mexican Volcanic Belt. Geochem. J. 38, 43–65.

Wallace, P.J., Carmichael, I.S.E., 1999. Quaternary volcanism near the Valley of Mexico: implications for subduction zone magmatism and the effects of crustal thickness variations on primitive magma compositions. Contrib. Mineral. Petrol. 135, 291-314.

Weaver, B., 1991. The origin of ocean island basalt endmember compositions: trace element and isotopic constraints. Earth Planet. Sci. Lett. 104, 381–397.

Yu, H., 2011. Li, Hf and Os isotope systematics of Azores basalts and a new microwave digestion method for Os isotopic analysis. PhD Dissertation, Miami University, 183p.

Zandt, G., Humphreys, E., 2008. Toroidal mantle flow through the western US slab window. Geology 36, 295–298.

108

109

110

111

112

Figure 1. Simplified regional tectonic map of the Trans-Mexican Volcanic Belt (yellow), modified after Siebe et al. (2004). The MGVF region and the location of Jorullo volcano are also indicated.

114

Figure 2. Geologic map of Jorullo and associated deposits. Eruptive phases are numbered from I (earliest) to VII (latest). Sample locations for the lava and tephras are indicated as red and orange dots respectively.

115

116

Figure 3. Temporal-compositional variations in Jorullo magmas. a) Variation in MgO content of the lavas throughout the eruption. b) Variation in MgO contents of the tephra with relative stratigraphic height. Diamond symbols represent lava samples; filled circle symbols represent Jorullo tephra. Two geochemical groups can be recognized within the Jorullo samples: high- MgO (>8 wt %) and low-MgO (<8 wt %) groups. Pink symbols represent the high-MgO group samples, and blue symbols represent the low-MgO group samples. The green diamond represents the last material erupted at Jorullo (bomb sample JOR-0766). Black crosses represent additional Jorullo samples analyzed for major and trace elements (UNAM database), the data for which are included in Appendix 2. These color & symbol designations are used for all following figures.

117

118

Figure 4. Whole rock major element vs. MgO (wt. %) for the Jorullo samples analyzed in this study. Samples from Luhr and Carmichael (1985) (indicated by black dots) and the additional samples from Appendix 2 (black crosses) are shown for comparison.

119

120

Figure 5. Incompatible trace element abundances vs. MgO (wt %) for the Jorullo samples analyzed in this study. Samples from Luhr and Carmichael (1985) and the additional samples from Appendix 2 are shown for comparison.

121

122

Figure 6. Trace element patterns of the Jorullo samples, normalized to normal mid-oceanic ridge basalt (N-MORB). Normalizing values are those of Sun and McDonough (1989).

123

124

Figure 7. Isotope variation diagrams for the Jorullo samples. a) Sr-Nd isotope diagram, b & c) Pb isotopic variation diagrams, d) Hf-Nd isotope diagram. Basement granitoids analyzed in this study (black squares) and literature data for the TMVB (open gray dots; Gómez et al., 2003; Luhr et al., 2006) are shown for comparison. Also shown are: terrigenous sediment (blue crosses) and pelagic sediment (open orange crosses) from DSDP site 487 (LaGatta, 2003), and the average altered oceanic crust (open pink cross) from DSDP site 487 (Verma, 2000). Mantle xenoliths and lower crustal xenoliths (stars and “Y”s, respectively) are compiled from Schaaf et al. (2004). Continental basement drilled at Sites 493 (granodiorite) and sites 489 (Biotite gneiss) are from Straub et al., (2015)

125

126

Figure 8. a) 143Nd/144Nd vs. 87Sr/86Sr and 206Pb/204Pb vs. 87Sr/86Sr for the Jorullo samples. Also shown are fields for the Northern Mexican Extensional Province of Luhr et al. (2006). Upper crust and lower crustal xenolith fields are compiled from Schaaf et al. (2004).

127

128

Figure 9. a) 187Os/188Os vs. Ni for Jorullo samples and other mafic samples from the MGVF (black triangles, data from Chesley et al., 2002). The mixing curves represent the AFC models of Chesley et al. (2002) and are based on the following parameters: parental magma: Os = 250 ppt, Ni = 300 ppm and 187Os/188Os = 0.135; lower crust: Os = 50 ppt, Ni = 100 ppm and 187Os/188Os =

0.6. Bulk DNi = 5. Two AFC trends are shown for bulk DOs of 10 and 20. Tick marks represent mass fraction of assimilant in 2% increments. b) δ18O vs. Ni contents of the Jorullo magmas.

129

130

Figure 10. a) 87Sr/86Sr vs. 206Pb/204Pb showing mixing between NMEP mantle and potential subduction components with variable mixtures of AOC and pelagic and terrigenous sediment- derived fluids. Mixing trajectory between terrigenous and pelagic sediment is also shown. A relatively high proportion of sediment to AOC (~80:20) is required to produce a subduction 87 86 component with high Sr/ Sr required by SC1. Only the AOC-sediment mixture that is dominated by terrigenous sediment (90:10 terrigenous:pelagic) produces the required Sr and Pb isotopic signatures with >4% subduction component added to the NMEP, as required by the 18 heavy δ Oolivine signatures; greater contributions of pelagic sediment dictate lower fractions of subduction component in the mixture. b) 143Nd/144Nd vs. 176Hf/177Hf diagram further illustrating the influence of terrigenous sediment in the petrogenesis of Jorullo magmas. The hypothetical NMEP mantle is represented by the green circle, and its isotopic composition was chosen to fall along an extension of the low-MgO samples data trend in Sr-Nd and Sr-Pb space and within the NMEP field (Figure 9). The dashed and solid curves in the diagrams indicate mixing between sediment-derived fluid (with variable proportions of terrigenous and pelagic sediment) and AOC-derived fluids, which comprise the subduction component added to the hypothetical NMEP mantle source. Tick marks show the relative proportions of the mixing components. The Jorullo samples are reproduced in Hf-Nd isotope space with the same parameters as for Sr-Pb isotopes

131

132

Figure 11. Isotopic models to account for the compositional variation at Jorullo, involving an

NMEP mantle wedge and two slab components SC1 and SC2. The subducted sediment and AOC compositions are from DSDP site 487, and the associated fluid compositions are derived using experimentally determined partition coefficients as described in the text. Model parameters are provided in Appendix 3. The stars represent the two different slab components, SC1 and SC2, that best reproduce the low-MgO and high-MgO group Jorullo magmas for a) 206Pb/204Pb vs. 87Sr/86Sr; b) 143Nd/144Nd vs. 87Sr/86Sr; c) 176Hf/177Hf vs. 143Nd/144Nd; and d) δ18O vs. 206Pb/204Pb.

The models depicted in figures a-d represent NMEP plus 4% subduction fluid SC1 derived from 80% sediment (90:10 terrigenous:pelagic) and 20% AOC; and a subsequent addition of 1.2% subduction fluid SC2 derived from 80% sediment (95:5 terrigenous:pelagic) and 20% AOC.

133

134

Figure 12. Trace element modeling results showing calculated mantle source compositions for NMEP plus subduction component SC1 ± SC2 (thin solid lines), and the magma derived by partial melting of the respective sources (dashed lines). a) The trace element pattern of the low- MgO group can be closely reproduced by 9% partial melting of an NMEP mantle source metasomatised by a subduction component (SC1) as apportioned in Figure 11. b) The trace element pattern of the high-MgO group can be closely matched by ~15% partial melting of an

NMEP mantle source metasomatised by slab component SC1 and SC2, also as apportioned in Figure 11. Note that in both models Cs, Rb and Pb are predicted by the model to be in somewhat higher abundance than that observed in the Jorullo samples. See text and Appendix 3 for additional explanation.

135

136

Figure 13. Schematic representation of the formation of the mantle sources of the high- and low- MgO group samples, and the evolution with time of the high- and low-MgO magmas, leading to the documented compositional-temporal variations in the Jorullo eruption. 1) Melting of NMEP mantle metasomatised by SC1 and SC2 to produce the early high-MgO group and their eruption along the rift between the Jorullo main cone and satellite cones. The early melts rose from depth and may have ponded briefly at the lower crust-upper mantle boundary before eruption. 2) Storage of some residual high-MgO melt at a shallow crustal level to produce JOR-0766. 3)

Melting of the less fertile mantle (i.e. NMEP + SC1) as time progressed, producing the parental low-MgO group magma. The low-MgO melts rose from depth, and likely stalled in a mid-crustal depth magma chamber above or around the chamber of JOR-0766, where amphibole fractionation occurred. The low-MgO melts erupted from the main cone. 4) The very last product of the Jorullo eruption, represented by sample JOR-0766, erupted through the main cone.

137

138

139

140

141

142

143

144

Electronic Appendix 3. Table 1. Compositions and partition coefficients used in modeling

Element 2Terrigenous 2Pelagic 3Mobility 1Mantle 2AOC 4Dsed/fluid 5Dsed/Melt (ppm) Sed. Sed. AOC Cs 0.001 8.62 4.69 0.01 0.98 2.34 0.79 Rb 0.058 114 63 3 0.92 1.32 0.42 Ba 0.630 1154 5716 2 0.92 1.04 0.48 Th 0.008 7.58 5.29 0 0.63 4.13 0.82 U 0.003 3.62 2.45 0 0.61 3.06 0.62 Nb 0.140 11 7.56 1 0.15 2.99 1.23 Ta 0.008 0.82 0.44 0 0.08 2.72 1.23 La 0.165 24 64 1 0.75 1.7 1.52 Ce 0.563 49 49 3 0.58 1.56 1.3 Pb 0.025 31 58 1 0.89 0.94 Nd 0.518 23 61 4 0.22 1.44 1.46 Sr 7.640 164 339 61 0.85 0.91 1.23 Zr 6.250 95 128 35 0.08 Hf 0.164 2.45 2.37 1.20 0.10 0.7 Sm 0.218 4.84 12 1.80 0.09 1.61 1.62 Eu 0.089 1.16 3.48 0.70 0.05 1.56 1.74 Ti 623 0.06 Gd 0.339 4.37 15 3 0.02 2.02 1.66 Dy 0.478 4.06 14 3.70 0.01 2.6 1.72 Y 3.940 24 89 2.30 0.01 2.93 1.68 Er 0.348 2.33 8.08 2.40 0.004 3.17 1.69 Yb 0.388 2.29 7.45 0.003 3.66 1.68 Lu 0.062 0.36 1.21 0.40 0.003 3.85 1.87 87Sr/86Sr 0.70330 0.70844 0.70849 0.70320 143Nd/144Nd 0.512870 0.512506 0.512537 0.513100 206Pb/204Pb 19.400 18.801 18.504 18.220 176Hf/177Hf 0.282997 0.282783 0.282940 0.283100 δ18O (‰) 5.2 20 25 12 1 Trace elements from average depleted mantle (DMM) of Salters & Stracke, 2004. 2 Average trace element and isotopic compositions of terrigenous, pelagic and altered oceanic crust determined for DSDP sites 487 (Lagatta, 2003 and Verma, 2000). 3 Average trace element mobility, 700 ⁰C, 4 GPa dataset of Kessel et al. (2005a). 4 Partition coefficients 650 ⁰C, 2 GPa dataset of Johnson & Plank (1999); Hf Kd assumed equal to Zr. 5 Partition coefficients 900 ⁰C, 2 GPa dataset of Johnson & Plank (1999).

145

Electronic Appendix 4

Generating Jorullo mantle source heterogeneity: an alternate model

If we do not consider the oxygen isotope results to require addition of at least 3-4% slab component to the modern mantle wedge, but rather assume a starting mantle wedge with heavy oxygen isotope signatures, then an alternate model with substantially less slab contribution can be evaluated. Using the same approach as that described in the text for the oxygen isotope- constrained model, we find that the Sr-Nd-Pb-Hf isotope and trace element compositions of the

Jorullo samples can be alternatively explained with addition of as little as ~1% each of SC1 and

SC2, followed by lower degrees of partial melting of the metasomatised mantle (~3.5% and 6%, versus 9% and 15), without the discrepancies in Cs, Rb and Pb concentrations.

The models shown in Appendix 4 Figs. 1 and 2a are the result of adding ~0.9% subduction component SC1 to the NMEP mantle source followed by ~3.5 % partial melting, which reproduces the low-MgO group samples. This model is very similar to the oxygen isotope- constrained model, again with ~21% altered oceanic crust-derived fluid and a 90:10 terrigenous:pelagic sediment component, but in this case 71% sediment-derived hydrous fluid with ~8% sediment melt, the latter required in order to reproduce the slope in Nd- Sr isotope space. The Sr-Nd-Pb-Hf isotopic signatures of the high-MgO group magmas can likewise be modeled by addition to the previously SC1-metasomatised mantle source of ~1% subduction component SC2 (Appendix 4 Fig. 1). In this case, SC2 consists of 19% altered oceanic crust- derived fluid, ~73% sediment-derived hydrous fluid and ~8% sediment melt, with the sediment a 95:5 terrigenous:pelagic mix as for the oxygen isotope-constrained model. The trace element abundances of the high-MgO magmas can be generated by ~6% partial melting of this doubly- modified NMEP mantle (Appendix 4 Fig. 2b). Note that the oxygen isotope signatures cannot be matched by this model unless the mantle wedge is considered to be ancient forearc mantle that had previously been fluxed with a sediment-derived component to produce its heavy 18O signature prior to the modern subduction processes that control the radiogenic isotope systematics.

146

147

148

CHAPTER 4

Petrogenesis of Mafic Magmas in the Sierra Chichinautzin Volcanic Field, Mexico: Constraints from Os Isotope Systematics

149

Abstract:

We present new major and trace element abundances and Sr, Nd, Pb, Hf, Os and O isotopic compositions for high-MgO magmas from the Sierra Chichinautzin Volcanic Field (SCVF) to evaluate the origin of the compositional variations observed within the volcanic field. The primitive magmas within the volcanic field define two geochemical groups: high-Nb (>15ppm), generally alkaline/transitional magmas, and low-Nb (<15 ppm), calc-alkaline magmas. The high-

Nb group displays lower SiO2 but higher TiO2, FeO, P2O5, Y and Zr at a given MgO content relative to the low-Nb samples. The two geochemical groups overlap in their ranges of Sr-Nd- Pb-Hf-O isotopic compositions; however the high-Nb samples have significantly more radiogenic 187Os/188Os than the low-Nb samples. In addition, the high-Nb and low-Nb groups together define a positive trend with concave upward curvature in 187Os/188Os-206Pb/204Pb space, which cannot be attributed to lithospheric contamination, but instead reflects mixing between two compositionally distinct mantle sources. The low-Nb group can be explained by addition to a depleted mantle wedge of hydrous fluid derived from dehydration of trench sediment and upper altered oceanic crust. In contrast, major and trace element data as well as Os and Pb isotopic data suggest the involvement of a pyroxenite component in the mantle source of the high-Nb group samples. Our models suggest that the pyroxenite may be formed by melt-mantle reaction involving a slab-derived sediment melt, likely caused by fluxing of previously dewatered sediment with a slab-derived fluid beneath the volcanic front. Mixing between the two sources, potentially as melts, can produce the compositional spectrum of the SCVF monogenetic field. This study suggests a genetic relationship between the high-Nb and low-Nb magmas found in close proximity in the SCVF.

150

1. Introduction

Monogenetic volcanic systems are characterized by small volume magmas that are predominantly basaltic in composition (Connor and Conway, 2000). They occur in all tectonic settings and often produce near primitive magmas (i.e. Ni > 200 ppm and Cr > 500 ppm) that suggest rapid ascent from the mantle source to the surface (Rutherford, 2008). Because of their mafic character, monogenetic systems have the potential to reveal information about sources and processes of magma generation as well as the magmatic processes occurring during ascent and emplacement of mantle-derived magmas. As such, mafic monogenetic volcanic systems can improve our understanding of the processes involved in the origin of basaltic volcanism in general. Monogenetic volcanism is often characterized by significant compositional variations within individual eruptions and within given volcanic fields (Siebe et al., 2004b; Rasoazanamparany et al., 2015). Ongoing debate regarding the origin of this compositional variation focuses mostly on the relative roles of (1) assimilation of lithospheric mantle or continental crust during ascent and emplacement and (2) heterogeneity in the mantle source due to subduction fluid/melt fluxing of a mantle wedge or other mantle metasomatic processes. Related key questions include the origin of closely linked low-Nb, calc-alkaline and high-Nb, or “OIB-type” alkaline/transitional magmas, which can occur together in a single volcanic field and sometimes within a single eruption. Melting of pyroxenitic and pyroxenite/peridotite hybrid mantle has been proposed to explain geochemical and isotopic heterogeneity in both intraplate and subduction-related mafic monogenetic eruptions, and also to explain the observation that monogenetic eruptions often become progressively more silicic with time (Reiners, 2002; Straub et al., 2013a). Such temporal-compositional relationships could be explained in some cases by variable degrees of fractional crystallization or assimilation-fractional crystallization (AFC) of a single parental magma, but in many cases isotopic data indicate that the compositional changes may be linked to compositionally distinct mantle sources feeding the magmatism. Direct evidence for lithologically heterogeneous mantle, including pyroxenite veins within peridotite host mantle, is abundant in mantle xenolith suites from a range of tectonic settings (Hirschmann and Stopler, 1996; Pearson and Nowell, 2004; Berly et al., 2006) as well as ultramafic massifs from around the globe (Pearson and Nowell, 2004; Berly et al., 2006). Although the abundance of pyroxenites in the upper mantle is not well constrained, estimates

151 from exposed ultramafic massifs suggest that the mantle may contain as much as 2 to 5% of a variety of pyroxene-rich lithologies including pyroxenites (websterites, clinopyroxenites and orthopyroxenites) and eclogites in the form of veins and lenses ranging in size from the centimeter to meter scale (e.g. Reiners, 2002). Although there is some debate as to the origin of these pyroxene-rich lithologies, they may be formed by recycling of oceanic crust and, perhaps in many cases, melt-mantle reaction (e.g. Sobolev et al., 2005; Herzberg, 2011; Straub et al., 2013a, 2015). Subduction systems, in which slab-derived fluids and/or melts percolate through the overlying mantle wedge, may be particularly susceptible to the formation of pyroxenitic veined mantle (e.g. Kepezhinskas et al., 1987; Berly et al., 2006). However, their importance in magma production in arc settings is still in question, because it is often difficult to distinguish subduction-related mantle metasomism from crustal assimilation, as subducted materials (e.g. sediment and altered oceanic crust) and crustal rocks can have similar elemental and isotopic signatures. In this contribution, we present new whole-rock major and trace element abundances and Sr, Nd, Pb, Hf, Os and O isotopic compositions for a suite of relatively primitive, young (<25,000 year old) mafic rocks from the Sierra Chichinatuzin Volcanic Field and vicinity in the Trans-Mexican Volcanic Belt to (1) evaluate the relative importance of lithospheric contamination vs. source-related enrichment in generating the compositional variations observed within the volcanic field, (2) to evaluate the genetic relationship between low-Nb and high-Nb magmas within the field, and (3) to evaluate the potential role of pyroxenite in magma generation beneath the SCVF.

2. Tectonic setting and magmatism

The 1,000 km long and 20-200 km wide Trans-Mexican Volcanic Belt (TMVB) is an E-W oriented volcanic province that crosses central Mexico from the Pacific Ocean to the Gulf of Mexico. Volcanism in the TMVB is generally thought to be associated with the subduction of the Rivera and Cocos plates beneath the North American plate along the Middle American trench. Volcanic activity has been ongoing since the Miocene and has produced more than 8,000 eruptive centers including domes, stratovolcanoes, and monogenetic cinder cones (Ferrari et al., 1999). The volcanic belt erupts basaltic to dacitic magmas with both calc-alkaline and

152 alkaline/transitional affinities. The calc-alkaline basalts to dacites comprise the most abundant magmas in the volcanic belt and exhibit typical subduction signatures with enrichment in large- ion lithophile elements (LILE) relative to the high field strength elements (HFSE). The more rare alkaline/transitional magmas make up less than 5% by volume of the erupted magmas. These magmas exhibit trace element and isotopic signatures similar to those of intraplate basalts such as ocean island basalts (OIB), and are often referred to as “high-Nb” magmas (Wallace and Carmichael, 1999; Cai et al., 2014; Straub et al., 2015). The Sierra Chichinautzin Volcanic Field (SCVF) is located in the central part of the TMVB, just south of Mexico City, and represents the volcanic front of the volcanic belt in this region (Fig. 1). The SCVF marks the southern limit of the Mexico Basin and the volcanic field is underlain by an active E-W transtensional fault system (Mazzarini et al., 2010) and bounded by the Popocatépetl stratovolcano to the east and the Nevado de Toluca stratovolcano to the west. The basement beneath the SCVF is composed of a folded, ~3 km thick Cretaceous calcareous sequence and a half kilometer thick sequence of Eocene-Oligocene volcanoclastics (calcareous conglomerates with dacitic and andesitic clasts), lava flows, sandstones, volcanic siltstones and lacustrine deposits (Velasco-Tapia and Verma, 2013). The Laramide orogeny is believed to be responsible for the folding of the Cretaceous basement rocks (Fries, 1960). The Cretaceous calcareous sequence itself was intruded by 50±10 Ma granitoid dykes (De Cserna et al., 1974) and is composed of massive limestone with black chert lenses, gypsum beds, bedded limestones, and greywacke interbedded with limonite and shale (Velasco-Tapia and Verma, 2013). The SCVF hosts more than 220 monogenetic volcanoes, mainly scoria cones that are 2 distributed over an area of nearly 2,500 km . Paleomagnetic studies indicate that the vast majority of the volcanic centers are younger than 780 ka (Mooser et al., 1974; Bloomfield, 1975; Siebe et al., 2004a). Previous studies have shown that the volcanic field comprises predominantly calc-alkaline basalt to dacite with typical arc geochemical signatures, but high-Nb magmas are not uncommon in the SCVF, and are often erupted in close spatial association with the more dominant calc-alkaline magmas; in several cases they are known to have been produced together with calc-alkaline magmas during the course of single eruptions (e.g. in the development of the Texcal flow and the Chichinautzin Volcano; Straub et al., 2013a; 2015). Wallace and Carmichael (1999) suggested that the most primitive calc-alkaline magmas in the SCVF were formed by melting of a depleted mantle source that had been enriched in fluid

153 mobile elements derived from the subducting slab. In contrast, they suggested that the high-Nb basalts were produced by melting of a mantle with variable compositions, ranging from highly depleted source to mantle similar to the source of intraplate OIB-type magmas. More recently, Cai et al. (2014) and Straub et al. (2011; 2013) have suggested, respectively, that the high-Nb magmas may be related to slab melting and/or the formation of pyroxenite-veined mantle.

3. Samples and analytical techniques

Eight lava samples that represent some of the most primitive lavas within the SCVF were selected for this study, including samples from Xitle (1,670 yr BP; Siebe, 2004a), Chichinautzin (1,835 yr BP; Siebe et al., 2004a), Cilcuayo (>14,000 yr BP) and Pelagatos (<14,000 yr BP; Guilbaud et al., 2009; Agustín-Flores et al., 2011), and Xochitlán (Schaaf et al., 2005) scoria cones (Fig. 2). One mafic sample from Popocatépetl, from a mega-block in a debris avalanche deposit dated at ca. 23,000 yr BP (Siebe and Macías, 2006) was also included in this study. Three crustal rocks, which represent potential assimilants, were also included: a quartz-rich cumulate from Chichinautzin volcano, and from the Tepotzlán area ~30 km south of Mexico City, an andesite fragment contained in a Tertiary lahar deposit and a sample of the Cretaceous carbonate basement (Cosky, 2010).

Whole-rock major element (SiO2, TiO2, Al2O3, FeO, MnO, MgO, CaO, Na2O, K2O and

P2O5) and a subset of trace element ( Ni, Cr, V, Ga, Cu, and Zn) abundances were determined by X-ray fluorescence (XRF) spectrometer at Washington State University, and abundances of rare earth element (REE) and other trace elements were determined by inductively coupled plasma mass spectrometry (ICP-MS) at Washington State University. Strontium, Nd, Pb and Os isotopic compositions were previously determined by Cosky (2010) using thermal ionization mass spectrometry (TIMS) on a Thermo-Finnigan Triton at Miami University. Details of the chemical separation procedures for Sr, Pb and Nd can be found in Snyder (2005). The measured 87Sr/86Sr was corrected for mass fractionation using 86Sr/88Sr = 0.1194. 143Nd/144Nd was corrected for fractionation using 146Nd/144Nd = 0.7219. External precision based on long-term two standard deviation (2 SD) reproducibility of the NBS 987 and La Jolla standards are ±0.000015 and ±0.000007 for 87Sr/88Sr and 143Nd/144Nd respectively. Lead isotopes (206Pb/204Pb, 207Pb/204Pb, 208Pb/204Pb) were corrected for mass fractionation by 0.10% per amu. Errors on 206Pb/204Pb,

154

207Pb/204Pb, and 208Pb/204Pb were ±0.014, ±0.019, and ±0.063, respectively based on long-term 2 SD external reproducibility of the NBS 981 standard. Osmium isotopes were measured as the - negative molecular ions of OsO3 and corrected for oxygen isotopes based on the oxygen isotopic composition of Nier (1950). Mass fractionation correction for Os was based on 192Os/188Os = 3.0826. Long-term 2 SD external reproducibility of Os isotopic ratios measured on an in-house NIST Os standard is ± 0.0002. Hafnium chemistry involved dissolution of separate aliquots of sample powder. A total of

150 mg of leached powder was weighed and dissolved in Savillex beakers using the HF-HNO3 dissolution procedure of Connelly et al. (2006). Upon full dissolution, the samples were dried and re-equilibrated with 0.5N HCl, at which point the samples were ready for chemical separation. The chemical separation of Hf was completed in two steps following the method of Connelly et al. (2006). The first step consisted of a cation exchange column using BioRad AG50W-X8 resin in which high field strength elements (HFSE) were separated from the matrix. The second purification step used a column with Eichrom TODGA resin (50-100µm) in which Hf and Zr were subsequently separated from the other HFSE. Isotopic analyses were done on a Multi-Collector Inductively Coupled Plasma-Mass Spectrometer (MC-ICP-MS) using a VG P-54 at the Department of Terrestrial Magnestism (DTM) of the Carnegie Institution of Washington. Isotopic measurements of samples were interspersed with those of the JM 475 standard, and the samples were normalized to a 176Hf/177Hf value for JM 475 of 0.282160. The average measured 176Hf/177Hf ratio for JM 475 during the period of data collection was 0.282151, with a 2 SD external reproducibility of ± 0.000014. Olivines were separated from their host lavas by crushing and sieving to a fraction of ~200 µm followed by handpicking from the sieved fraction under a binocular microscope. The handpicked grains were further crushed with an alumina mortar and pestle and passed through a Frantz magnetic separator to remove crystal fragments with high concentrations of oxide inclusions. To ensure removal of mineral fractions with oxide inclusions or coatings of matrix, the separates were further handpicked under a binocular microscope. The separates were then cleaned three times in an ultrasonic bath with 18.2 MΩ H2O. Oxygen isotopes were measured by laser fluorination using a MAT 251 mass spectrometer at the University of Wisconsin. Approximately 2 to 3 mg of olivine was loaded into the pre-fluorinated source chamber. Two measurements were performed for each sample and the average δ18O values are reported in Table

155

1. The Gore Mountain garnet standard UWG-2 was run with the samples to monitor the precision and accuracy of the analyses. The average δ18O obtained for the UWG-2 standard on a given day was used to normalize the samples relative to the accepted value of 5.80‰. The reproducibility of the UWG-2 garnet standard obtained during the period of analyses was 5.79±0.12‰ (2 SD). The δ18O values of the samples are reported relative to V-SMOW.

4. Results

4.1. Major and trace elements

Major and trace element data for the SCVF samples are provided in Table 1. The samples included in this study span a compositional range from basalt to basaltic andesite with MgO contents ranging from 10.2 – 7.7 wt. %, and exhibiting low-Nb (<15 ppm) calc-alkaline to high- Nb (>15 ppm) alkaline/transitional affinities, respectively. Major element compositions fall within the range observed previously in the SCVF, and the TMVB in general. Variation diagrams for the samples from this study, together with previously published data for other mafic monogenetic cones within the SCVF, are shown in Fig. 3 as a function of wt. % MgO. The samples from this study, although from different volcanic centers and therefore not expected to be consanguineous, are in accord with the relatively systematic variations of major element abundances versus MgO observed in the TMVB and SCVF. In general, with decreasing MgO,

CaO decreases, whereas SiO2, Al2O3, Na2O, K2O and P2O5 abundances increase. Notably, the high-Nb samples display systematically lower SiO2 and higher TiO2, FeO and P2O5 at a given MgO content relative to the low-Nb samples. Although the samples from this study exhibit some degree of scatter, abundances of the compatible minor elements Cr and Ni, are also positively correlated with MgO, consistent with control by olivine fractionation. However, the samples analyzed in this study comprise some of the highest Cr, Ni and MgO contents found in the SCVF, and several are within the compositional range of primitive mantle melts. Variations of trace elements with MgO are illustrated in Fig 4. Among the samples from this study, as well as those from the SCVF as a whole, incompatible trace elements generally lack systematic variations with MgO. However, strong positive correlations exist between most incompatible trace elements (Fig 5). Notably, on incompatible trace element variations diagrams

156 involving high field strength elements (HFSE), the high-Nb group samples can be distinguished from the low-Nb group samples, with the former typically displaying higher abundances of elements such as Zr, Y, Hf and Ta (and to some extent Th, U, and REE) at a given MgO content (Fig. 4). The high-Nb group samples also exhibit higher ratios of Ce/Pb and Nb/U. Although the high-Nb samples display higher Ce/Pb (>10) and Nb/U (~20) than the low-Nb samples, these ratios are lower in all of the SCVF samples compared to those of mid-ocean ridge basalts (MORB) and OIB for which Ce/Pb = 25 and Nb/U = 47 ± 10 (Hofmann, 1986). The trace element patterns of the samples normalized to N-MORB (Fig. 6) are broadly similar to those of typical subduction zones magmas, with enrichment of fluid-mobile LILE (e.g. Cs, Rb, Ba, Pb, Sr and K) and depletion in fluid-immobile HFSE (e.g. Nb and Ta). In all samples, the HREE are slightly depleted compared to normal MORB (N-MORB). Relative to the low-Nb group samples, the high-Nb group samples exhibit less pronounced Nb-depletions and higher concentrations of REE, but concentrations of most LILE (e.g. Ba, Pb, K and Sr) are similar in both the high-Nb and low-Nb groups.

4.2. Sr, Nd, Pb, Hf, Os and oxygen isotopes

The 87Sr/86Sr isotope ratios of the samples analyzed in this study display a significant range from ~0.70361 to 0.70419 (Table 1; Cosky, 2010). The corresponding 143Nd/144Nd ratios range from 0.512814 to 0.512932 and exhibit a broad negative correlation with 87Sr/86Sr. These data fall within the range of Sr and Nd isotope ratios previously determined for the SCVF (Luhr et al., 2006; Gomez-Tueña et al., 2007) and between the compositions of EPR-MORB and local subducting sediments (Fig. 7a). In general, the high Nb samples in this study have slightly higher 143Nd/144Nd ratios and lower 87Sr/86Sr than the low-Nb samples, consistent with the relationship found in the SCVF as a whole (Cai et al., 2014; Straub et al., 2015). 206Pb/204Pb ratios for the samples in this study range from 18.634 to 18.759 and fall within the range of the sparse published Pb isotope data for the TMVB and SCVF (Cai et al., 2014; Straub et al., 2015). 207Pb/204Pb ratios show slight variation ranging from 15.572 to 15.616, and 208Pb/204Pb ratios vary from 38.363 to 38.532. In plots of 207Pb/204Pb and 208Pb/204Pb versus 206Pb/204Pb, the SCVF samples all plot well above the Northern Hemisphere Reference Line (NHRL), with the high-Nb samples extending to more radiogenic Pb isotope signatures than

157 those of low-Nb samples (Figs. 7b, c). All of the samples are significantly more radiogenic than the EPR-MORB and the DSDP site 487 altered oceanic crust, but fall between the fields for terrigenous and pelagic sediments dredged from the DSDP site 487. 176Hf/177Hf isotope ratios of the samples from this study vary from 0.282944 to 0.283030. These values overlap with published data for other SCVF low- and high-Nb samples (Cai et al., 2014; Straub et al., 2015). On a 176Hf/177Hf vs. 143Nd/144Nd diagram (Fig. 7d), the samples define a positive correlation and falls within the MORB-OIB field, with the low-Nb samples displaced to slightly higher Hf isotopic ratios relative to the high-Nb samples at a given 143Nd/144Nd value. The Os concentration and 187Os/188Os ratios of the samples in this study span a large range and vary from 12 to 150 ppt and 0.1329 to 0.2608, respectively. Os behaves as a strongly compatible element, and exhibits a positive correlation with indices of differentiation including MgO and Ni (Fig. 8a, b). As observed for many basaltic systems, samples with low Os concentration (< 30 ppt; e.g. Reisberg et al., 1993; Widom and Shirey, 1996; Lassiter and Luhr, 2001) are substantially more radiogenic than those with higher Os concentrations (Fig. 8c). The high-Nb samples in this study are characterized by low Os (and Ni) concentrations, and exhibit the most radiogenic 187Os/188Os signatures. Nevertheless, all samples, including those with very high Os concentrations (>> 100 ppt), possess significantly more radiogenic Os isotope signatures than depleted MORB mantle (DMM) or primitive upper mantle (PUM) (Fig. 8). 187Os/188Os also exhibits a positive correlation with some trace element ratios including Ce/Pb, La/Nb and Ba/Zr (Fig. 9), as do 206Pb/204Pb and 87Sr/86Sr. Notably, 187Os/188Os exhibits a well-defined positive correlation with 206Pb/204Pb, with the high-Nb group samples substantially more radiogenic in both isotope systems (Fig. 10). The SCVF samples in this study display a relatively restricted compositional range in 18 δ Oolivine from 5.84-6.46 ± 0.09‰, within the range recently reported for other samples from the 18 SCVF (~5.6-7.8‰; Straub et al., 2015). No clear correlation is observed between δ Oolivine and whole-rock MgO or Ni (Fig. 11). Nevertheless, all samples have δ18O signatures significantly higher than the normal range for mantle olivine (5.2 ± 0.2‰), and slightly heavier on average than olivines from the Michoacán-Guanajuato Volcanic Field (MGVF), which range from ~5.6- 6.0‰ (Johnson et al., 2009; Rasoazanamparany et al., in prep) with an average for Jorullo volcano of 5.82 ± 0.12‰ (Rasoazanamparany et al., in prep). Consistent with the results of

158

18 Straub et al. (2015), there is no clear distinction in δ Oolivine between the low-Nb and high-Nb group samples in this study (Fig. 11).

5. Discussion

The samples selected for this study are all relatively mafic, with MgO contents ≥ 7.7 wt. %, and they comprise some of the most primitive within the volcanic field, with Ni and Cr contents as high as ~250-300 ppm and ~500 ppm, respectively (Fig. 3). Among these mafic samples, the suite spans a wide compositional range including the low-Nb, calc-alkaline rocks and high-Nb, alkaline/transitional-type rocks previously recognized in the SCVF. Major element compositional distinctions are apparent among these samples; most notably, the high-Nb samples display higher TiO2, FeO, and P2O2, and lower SiO2 than the low-Nb samples at a given MgO content (Fig. 3). These features could be indicative of different depths and degrees of partial melting, in which the high-Nb samples are extracted from smaller degrees of melting at greater depth than the low-Nb samples (e.g. Klein and Langmuir, 1987). However, variable Ce/Pb and Nb/U ratios among samples from the two groups argue against differences in depth and degree of melting being the sole cause of the compositional differences, as Ce-Pb and Nb-U cannot be fractionated by variable degrees of melting (e.g. Hofmann, 1986). Rather, the higher Ce/Pb and Nb/U in the high-Nb relative to the low-Nb samples (Fig. 4) could indicate that the two magma groups emanate from lithologically and/or compositionally distinct mantle source regions, or that they have been differentially modified by shallow magmatic processes. All samples have Ce/Pb and Nb/U ratios significantly lower than MORB and OIB, suggesting the involvement of continental crust in their petrogenesis, either as a crustal assimilant during AFC, or as a subducted crustal component introduced into their mantle source region. Furthermore, the high- Nb and low-Nb samples together define a positive trend and hyperbolic curve on a plot of 187Os/188Os and 206Pb/204Pb (Fig. 10), suggestive of two component mixing, which again could potentially be attributed to crustal contamination or mixing of two compositionally distinct mantle sources.

5.1. The role of crustal contamination in the origin of the mafic SCVF magmas

159

Numerous studies have documented the important role of crustal assimilation in the petrogenesis of arc magmas (e.g. Hildreth & Moorbath, 1988; Thirlwall et al., 1996; Chesley et al., 2002; Davidson et al., 2007). The samples analyzed in this study comprise the most mafic rocks within the SCVF, some of which are close to primitive mantle compositions (Cr = 500 ppm, Ni = 200 ppm), suggesting that they have not experienced significant fractionation, and implying limited crustal storage time and limited opportunity for wall-rock assimilation to occur. Nonetheless, even primitive magmas could be affected to some degree by shallow magmatic processes as they ascend through the crust. Several geochemical characteristics of the SCVF samples appear to be consistent with contamination of a primary magma at crustal depths: (1) the elevated incompatible trace element abundances relative to MORB including LILE and LREE, (2) the low Ce/Pb and Nb/U relative to MORB and OIB, (3) the radiogenic Os-isotopic composition of the samples, which are much more radiogenic than MORB-mantle and primitive upper mantle, and (4) the negative correlations of 187Os/188Os and 206Pb/204Pb with MgO. Trace element ratios such as Ce/Pb and Nb/U are effective indicators of crustal assimilation because they are minimally affected by crystal fractionation or variable degrees of partial melting of the mantle, but they can be significantly modified by involvement of continental crust or sediment due to the much lower ratios in the crust. An inverse correlation between Ce/Pb and Nb/U with indices of differentiation such as MgO or Ni would be expected if crustal assimilation were responsible for the geochemical signatures of the SCVF magmas. However, no such correlation is observed. The Os isotope system is also a highly sensitive indicator of crustal assimilation, due to the extreme difference in Os isotopic composition between continental rocks and mantle-derived magma. Even minor (a few % or less) crustal contamination can increase the 187Os/188Os of a mantle-derived magma (e.g. Reisberg et al., 1993; Widom and Shirey, 1996). However, the Os isotope data, combined with trace element and Pb isotope data, are incompatible with magma contamination at crustal depths. Both upper and lower crust have relatively low Ce/Pb ratios (~5; Rudnick and Gao, 2003), thus assimilation at any crustal level should produce a negative correlation between 187Os/188Os and Ce/Pb. The positive correlation observed for the SCVF samples (Fig. 9) is thus inconsistent with either upper or lower crustal assimilation. Likewise, the sense of curvature of the data trend on the 187Os/188Os and 206Pb/204Pb trend (concave upwards) is inconsistent with crustal assimilation, but rather supports a model in

160 which the radiogenic Os and Pb are produced by source enrichment (see Fig. 10 caption for model details).

5.2. Origin of the low-Nb samples

As discussed above, the relatively primitive SCVF volcanic rocks in this study appear to be unaffected by significant upper or lower crustal contamination. Thus, their chemical and isotopic signatures most likely reflect crustal input to their mantle source regions. Previous studies have suggested that the pre-subduction mantle wedge in the TMVB may be similar to the mantle source of mafic, back-arc volcanism characteristic of the North Mexican Extensional province (Luhr et al., 2006). The NMEP back-arc mantle is interpreted to advect into the sub-arc environment by slab-induced convection (e.g. Siebe et al., 2004a). The high-Nb magmas within the SCVF have been interpreted previously to represent partial melting of unmodified or slightly modified NMEP-type mantle (Luhr 1997; Luhr et al., 2006), whereas the calc-alkaline magmas are typically interpreted to be derived from partial melting of this residual mantle after removal of the high-Nb magmas and subsequent enrichment through addition of slab melt components (Siebe et al., 2004a). Because of the high mobility of LILE relative to HFSE in hydrous fluid, the enrichment in LILE (e.g. Pb, Ba, Rb, Cs and K) and depletion in HFSE (Nb, Ta, Hf) observed in the low-Nb samples (Fig. 6) can be explained by addition to their mantle source of aqueous fluid from the subducted sediment or altered oceanic crust. The higher 87Sr/86Sr but lower 206Pb/204Pb, Nb/U and Ce/Pb ratios of the low-Nb relative to the high-Nb samples may indicate that the mantle source of the low-Nb samples received more fluid relative to that of the high-Nb source. Similar primitive, low-Nb, calc-alkaline magmas from Jorullo Volcano in the MGVF (~200 km east of the SCVF), have recently been interpreted to reflect addition to an NMEP mantle wedge of primarily sediment-derived hydrous fluid to an NMEP mantle wedge (Rasoazanamparany et al., in prep). That study demonstrated that ≥4% subduction-related fluid added to the Jorullo mantle source was required to generate the lowest measured δ18O of 5.7‰, and that subducted terrigenous sediment dominated the flux of hydrous fluid and controls the overall geochemical signatures of the Jorullo magmas (Rasoazanamparany et al., in prep). The low-Nb SCVF samples share similar geochemical characteristics to those of the primitive high-

161

MgO group samples of Jorullo volcano, suggesting that they may tap a similar mantle wedge composition. However, the elevated δ18O values (> 5.85‰) in the SCVF low-Nb samples relative to those of Jorullo require a larger subduction contribution to the SCVF mantle source. As argued for the Jorullo samples, the radiogenic 87Sr/86Sr signatures of the low-Nb samples require a high proportion of sediment relative to altered oceanic crust (AOC) in the subduction component, whereas the Pb, Hf and Nd isotopic signatures require relatively high proportions of terrigneous to pelagic sediment (Rasoazanamparany et al., in prep). For the SCVF, our simple mixing calculations between an NMEP mantle and subducted sediment and basalt from the DSDP site 487 suggests that addition of ≥5% hydrous fluid from sediment (85% sediment, 90:10 terrigenous:pelagic) and AOC (~15%) is sufficient to explain the Sr, Nd, and Pb isotopic compositions of the low-Nb samples. The negative correlation of Ce/Pb with 87Sr/86Sr and the positive correlation with 206Pb/204Pb (Fig. 9) is qualitatively consistent with this model. Mixing between NMEP mantle and the Cocos plate sediment and AOC derived fluid might also explain the slightly radiogenic 187Os/188Os relative to MORB mantle (Fig. 8), although the behavior of Re and Os during mantle metasomatism remains debatable. Some studies argue that the high Os concentrations in the mantle (~3ppb) compared to those of potential metasomatic agents preclude modification of the Os isotopic compositions of the mantle during metasomatic events (e.g. Walker et al., 1989), and that metasomatic fluid or melts would have to possess Os abundances two to four orders of magnitude higher than the subducting source rocks in order to alter the Os isotopic compositions of the mantle (Chesley et al., 2004). Nonetheless, radiogenic 187Os/188Os of up to 0.17 has been reported for mantle xenoliths in subduction settings (Brandon et al., 1996; Widom et al., 2003). Recent experimental studies have shown that a high ƒO2 Cl- rich fluid can mobilize Os (Xiong and Wood, 2000). The infiltration of chlorine rich slab-derived fluid can potentially scavenge and deplete the mantle Os under oxidized conditions, such that the Os-poor mantle can be easily contaminated by subsequent infiltration of fluid or silicate rich melt with radiogenic 187Os/188Os signatures (Brandon et al., 1996; Widom et al., 2003; Widom, 2011). Thus, the slightly elevated 187Os/188Os ratios (<0.15) in the low-Nb might be attributed to overprinting of mantle osmium by chlorine rich hydrous fluids. Wallace and Carmichael (1999) and Straub et al. (2008) reported bulk rock oxygen fugacities ranging between -1 and +1.5 relative to QFM suggesting that the mantle is slightly oxidized. Blatter and Carmichael (1998) and Carmichael et al. (1996) reported evidence of high oxygen fugacity (e.g. QFM +1.5 to +2.4)

162 in eastern and western TMVB respectively, suggesting that the mantle beneath the TMVB could be highly oxidized in places, thus supporting Os mobility in slab-derived fluid. Lassiter and Luhr (2001) also noted that the calc-alkaline series within the TMVB generally have high oxygen fugacities ( -0.2 to +2.5) relative to the alkaline intraplate magmas (-0.8 to +1.7).

5.3. Origin of the high-Nb samples

As illustrated in Fig. 6, the trace element patterns for the high and low-Nb samples are similar, the primary difference being that the high-Nb samples are characterized by higher HFSE and REE and less depletion in Nb relative to the low-Nb samples. In addition, the high-Nb samples exhibit higher TiO2, P2O5, and FeO but lower SiO2 contents at a given MgO relative to the low-Nb group. Relatively high TiO2 and FeO, but lower SiO2 could be attributed to incorporation of mafic lithologies such as pyroxenites in the source of the high-Nb magmas, depending on the mineralogy of the pyroxenite (Hirschmann et al., 2003; Kogiso et al., 2003; Humphreys & Niu, 2009; Herzberg, 2011). Yang and Zhou (2013) demonstrated that most major element characteristics of experimentally produced peridotite and pyroxenite melts are very similar to one another, hence distinguishing between them by conventional means (e.g. comparison of major element abundances or ratios) can be problematic. However, they developed an empirical parameter, referred to as FC3MS and defined as (FeO/CaO) - 3 x MgO/SiO2, that is relatively insensitive to variability in depth and degree of partial melting, and that distinguishes experimental melts of pyroxenite from those of peridotite. To further investigate the potential involvement of pyroxenite in the source of the high-Nb samples, we have employed the FC3MS parameter of Yang and Zhou (2013). Fig. 12 shows a comparison of the low- and high-Nb samples from the SCVF on the FC3MS diagram. Although the low-Nb SCVF samples fall at values of FC3MS < 0.6 and plot within the peridotite-derived melt field, the high-Nb samples from this study fall at or above this value, with one sample falling in the pyroxenite field and the other at the transition between the pyroxenite and peridotite fields. These results are therefore consistent with a pyroxenitic and/or hybrid pyroxenite-peridotite source for these high-Nb magmas. It is important to note, however, that many high-Nb basalts from the SCVF literature span the range of FC3MS values from the peridotite to the pyroxenite fields; this may indicate diverse origins for SCVF

163 high-Nb magmas. In contrast, the pyroxenite field is essentially devoid of low-Nb samples, suggesting that only a predominantly peridotitic source is involved in their petrogenesis. Straub et al. (2008) have furthermore demonstrated the presence of high-Ni olivines within the high-Nb SCVF lavas, which they interpret as evidence for a pyroxenite source, following the suggestion of Sobolev et al. (2005; 2007) that anomalously high Ni melts and olivines result from melting of pyroxenite produced by reaction of recycled crust-derived melt and surrounding peridotite. Recent studies have further demonstrated a link between radiogenic Os isotope compositions and pyroxenite melting in OIB from Iceland (Sobolev et al., 2008) and the Hawaiian and Canary Islands (Sobolev et al., 2005; Herzberg, 2011). The radiogenic Os and Os-Pb isotope systematics of the SCVF samples in this study can likewise be explained by the involvement of a pyroxenite-rich component in their mantle source. Both low-Nb and high-Nb samples have distinctly higher 187Os/188Os than those of depleted mantle and primitive upper mantle, thus consistent with addition of crustal components to their mantle sources. The hyperbolic curvature on 187Os/188Os vs. 206Pb/204Pb (Fig. 13) suggests two-component mixing in which one end-member is characterized by more radiogenic 187Os/188Os (>.261), higher 206Pb/204Pb (≥ 18.75) and higher Os/Pb, and a second component characterized by less radiogenic 187Os/188Os (< 0.13), lower 206Pb/204Pb (~18.63) and lower Os/Pb. As illustrated in Fig. 13, the high-Nb samples extend toward the high 187Os/188Os and 206Pb/204Pb mixing end-member, whereas the low-Nb samples extend toward the less radiogenic Os and lower 206Pb/204Pb end-member. Both subducting oceanic crust and sediment can have very radiogenic 187Os/188Os, and could potentially provide a source for the radiogenic Os in the high-Nb samples. The oceanic crust beneath the SCVF region is estimated to be 23 Ma. Using the average 187Re/188Os (~4455) of the East Pacific Rise (EPR) MORB (Schiano et al., 1997) and an initial 187Os/188Os of 0.125, the subducted oceanic crust beneath the SCVF yields a calculated 187Os/188Os of 1.76. However, the relatively low 206Pb/204Pb (~18.2) precludes significant contribution of the subducted oceanic crust in the source of the high-Nb samples. In contrast, sediment (particularly terrigenous sediment), may provide a suitable source for both the radiogenic 187Os/188Os and 206Pb/204Pb. The elevated Nb and Ta content of the high-Nb relative to the low-Nb samples could indicate the involvement of sediment melt, as Nb and other HFSE have low solubilities in aqueous fluid (e.g. Kogiso et al., 1997) but are easily mobilized by silicate melts (e.g. Kessel et al., 2005). Nonetheless, the similar LILE abundances (e.g. Rb, Ba

164 and Pb) in the high and low-Nb groups indicate the contributions of hydrous fluid in addition to sediment melt in the source of the high-Nb samples. Although early studies indicated that the Re-Os isotope system should be relatively immune to mantle metasomatic processes, given the very high concentrations of Os in mantle peridotite relative to metasomatic hydrous fluid or silicate melts (e.g. Walker et al., 1989; Chesley et al., 1998), numerous recent studies have documented disturbance of the Re-Os isotope system in the mantle beneath ancient continents and active subduction zones, indicating that at least in some circumstances mantle Os isotope signatures can be overprinted (e.g. Brandon et al., 1996, McInnes et al., 1999; Widom et al., 2003; Griffin et al., 2004). A pertinent question that arises is whether sediment melt could influence the Os isotopic signatures of the surrounding mantle source of the high-Nb samples. One mechanism by which mantle peridotite 187Os/188Os signatures can be modified is by sulfide metasomatism. Sulfides are volumetrically insignificant in the mantle (0.1 wt%, e.g. Luguet et al., 2003), however, they dominate the mantle Re and Os budget, and exert a strong control on the Re and Os abundances and isotopic signatures of mantle-derived melts (e.g. Martin et al., 1993; Hart and Ravizza, 1996; Righter et al., 2000). A number of studies have suggested that interaction of sulfur-saturated silicate melt or fluid with mantle peridotite can result in precipitation of secondary sulfides (Luguet et al., 2008; Harvey et al., 2010). These secondary sulfides are often found along grain boundaries within mantle xenoliths and are characterized by radiogenic 187Os/188Os isotopic compositions (up to 1.7; Griffin et al., 2004). The behavior of sulfur during sediment melting remains largely unknown. However, a recent experimental study of pelitic sediment melting shows that elevated sulfur contents (up to 1 wt %) can be achieved in sediment melts at 2-3 GPa, 700-950⁰C (Prouteau and Scaillet, 2012). Due to the low solidus temperature of sulfides, they tend to melt first relative to the silicate minerals during mantle melting (Ballhaus et al., 2006). Because sulfides have extremely high 4 5 KdOs (~10 -10 , e.g. Hart and Ravizza, 1996) they will dominate the bulk-rock Os isotopic compositions. For example, in a study of highly metasomatized mantle peridotite from the Kaapvaal craton, Luguet et al. (2008) demonstrated that metasomatized mantle peridotite that has been enriched in secondary base metal sulfide can have both high Os abundances (0.41ppb) and radiogenic Os (up to 0.92 for Beni Bousera pyroxenite), even with very minor metasomatic sulfide addition (0.035 to 0.12 %). Such mantle can generate primitive melts with radiogenic

165

187Os/188Os. Sulfide inclusions in mafic phenocrysts from central TMVB lavas are relatively common (Blatter and Carmichael, 2001), and have been reported in some olivines from the high- Nb alkaline basalts of the SCVF, consistent with melting of relatively low oxygen fugacity mantle (Straub et al., 2008). In summary, involvement of pyroxenite veins produced by sediment melt-peridotite reaction can potentially explain not only the radiogenic 187Os/188Os and 206Pb/204Pb of the high- Nb samples, but also the enrichment in Nb (and other HFSE) relative to those of low-Nb samples. Our proposed model incorporates melting and reaction of sediment melts with the NMEP mantle source in the genesis of the high-Nb samples. This model differs from that of Straub et al. (2015) for the petrogenesis of the high-Nb samples. In their model, high-Nb samples are produced by partial melting of silica-poor pyroxenite veins that formed as a result of reaction of oceanic basalt-derived melt with the surrounding mantle peridotite. However, the positive correlation between 187Os/188Os and 206Pb/204Pb argues against a significant role for melt from the subducting oceanic crust, as that is relatively unradiogenic in Pb.

5.4. Os-Pb isotope systematics: a two component mixing model

Based on the above discussion, we propose that the low-Nb magmas are sourced by an NMEP mantle metasomatized by hydrous fluid, mostly derived from subducted sediments (SED: AOC ≈ 85:15), whereas the high-Nb magmas are derived from mixture of pyroxenite and surrounding mantle peridotite. The pyroxenite is considered to have formed by percolation of sediment melt through mantle peridotite previously fluxed by hydrous fluids to form the source of the low-Nb basalts. The hyperbolic trend in 187Os/188Os and 206Pb/204Pb indicates that the compositional spectrum of the high-Nb and low-Nb samples can be explained by mixing between melts generated from this pyroxenite-veined source and the hydrous fluid metasomatized mantle source that is the dominant source for the low-Nb, calc-alkaline magmas. We have conducted modeling of the observed 187Os/188Os - 206Pb/204Pb relationship using the inferred composition of the NMEP mantle source and the compositions of subducting sediment and altered basaltic oceanic crust (AOC) dredged from the DSDP site 487, representing the materials being subducted beneath the TMVB (e.g. Gomez-Tueña et al., 2003; Gomez-Tueña et al., 2007, Cai et al., 2014). The NMEP mantle source is assumed to have a depleted mantle

166

(DMM) trace element and 187Os/188Os isotopic composition, and Pb isotopic signatures within the range of basalts from the NMEP. In the model, we use the trace element and isotopic compositions of the sediment and basaltic oceanic crust from DSDP sites 487 reported by Verma et al., 2000, Gomez-Tueña et al., 2003; Lagatta, 2003, Gomez-Tueña et al., 2007, and Cai et al., 2014. To estimate the trace element composition of hydrous fluid derived from dehydration of oceanic crust we use the mobility data set of Kessel et al. (2005) for dehydration of metabasalt at 700⁰ C and 4 GPa. Trace element compositions of sediment-derived fluid and melt were calculated using the D values at 650⁰ C and 900⁰ C at 2 GPa from the datasets of Johnson and Plank (1999). The fraction of fluid released during dehydration was assumed to be 1.5 wt.% in both oceanic crust and sediment, similar to the fraction of hydrous fluid obtained from high-pressure dehydration experiments on amphibolite (Tatsumi and Kogiso, 1997. The Os abundances in slab fluid are estimated using the mobility of Suzuki et al. (2000). All parameters used in the model are given in Appendix 1. As shown in Fig. 13, higher Pb concentrations in the subducted sediment and altered oceanic crust than in the mantle produce concave upward mixing curves, reflecting large changes in 206Pb/204Pb relative to 187Os/188Os. Consistent with the Sr-Nd isotope data, the 187Os/188Os and 206Pb/204Pb of the low-Nb group can be reproduced by addition of 5 to 10% hydrous fluid to the same NMEP mantle. For the origin of the high-Nb samples, we infer melting of subducted sediment induced by fluid fluxing from the underlying altered oceanic ± serpentinized mantle (e.g. Johnson et al., 2009). In this model, the slab component (i.e. sediment melt and hydrous fluid) infiltrates and reacts with the surrounding peridotite to produce pyroxene-rich mantle (e.g. Tatsumi and Hanyu, 2003). In our model calculations, we assume that the NMEP mantle has the Os and Pb abundances and isotopic compositions of modern upper mantle. The average Os abundances and 187Os/188Os ratios for marine sediment (e.g.Peucker-Ehrenbrink and Ravizza, 2000) were used for the sediment components. The 187Os/188Os ratios of the altered oceanic crust was estimated based on the average 187Re/188Os (~4455) of the East Pacific Rise (EPR) MORB and its estimated age beneath the volcanic front (23 Ma). Our quantitative calculations for mixing between an NMEP mantle and melt plus fluid (from subducted sediment, and basalt and serpentinized mantle, respectively) from the DSDP site 487 suggests that a mixture of 80% pyroxenite veins (produced by sediment melt-mantle reaction) and 20% peridotite can reproduce the Os and Pb isotopic compositions of the high-Nb samples. High degrees of melting of

167 pyroxenite relative to peridotite (e.g. 20-60% melt of pyroxenite to 0.1-5% melt of peridotite) are predicted for a given pressure-temperature melting regime (Reiners, 2002). Lower overall degrees of melting of cooler and/or less water-rich hybrid mantle will thus favor the contribution of pyroxenite melt, and will dominate the geochemical and isotopic signatures of the resulting erupted magmas. The compositional spectrum from high-Nb to low-Nb magmas erupted in the SCVF can thus be explained by mixing between melts generated from this pyroxenite-veined source and the hydrous fluid metasomatized mantle source that is the dominant source for the low-Nb, calc-alkaline magmas.

5.5. Slab dehydration and melting beneath SCVF

Recent seismic studies of the Mexican subduction zone suggest a slab depth of 120 km beneath Popocatépetl, located just to the west of the SCVF (Manea and Manea, 2010). The young age of the subducted Cocos plate beneath the TMVB, and the steep dip of the slab beneath the SCVF, have led many studies to suggest that the slab beneath the SCVF is hot enough to induce slab melting (e.g. Gomez-Tueña et al., 2007; Cai et al., 2014). However, Johnson et al. (2009) have estimated the thermal structure of the subducting Cocos plate based on the present day slab geometry, and have suggested that surface temperatures of the slab beneath SCVF are ~600 ºC at 3 GPa, probably insufficient for sediment and oceanic crust melting to occur. Rather, fluids derived from dehydration of sediment, which will occur primarily within the forearc, would react with the overlying mantle wedge to form serpentinite, and subsequent down- dragging of the hydrated mantle wedge by corner flow can transfer this H2O to greater depths (Tatsumi and Eggins, 1995). Manea and Manea (2010) suggested that significant fluid stored in the serpentinized mantle wedge is released at 100 to 120 km depths (e.g. Grove et al., 2006). Our results suggest that the chemical and isotopic compositions of the low-Nb, calc-alkaline magmas in this study reflect hydrous fluid metasomatism of the mantle wedge, which may reflect this process. In addition, deep release of fluid from the altered oceanic crust may serve as a flux that can induce melting in the overlying dehydrated sediment (Johnson et al., 2009). The Os-Pb isotope systematics of the high-Nb samples appear to require a component of sediment melt, which may be produced in this manner. Subsequent melt-mantle reaction between silica-rich sediment melts

168 and peridotitic mantle wedge will result in pyroxenite veins that may vary significantly from silica-deficient to silica-excess depending on the melt:rock ratio (e.g. Reiners, 2002; Sobolev et al., 2005; Gerbode and Dasgupta, 2010; Straub et al., 2015). Modification of the mantle during this reaction process is likely to mobilize mantle sulfides (e.g. Luguet et al., 2008; Griffin et al., 2004; Pearson et al., 1998), the primary hosts for Os in silicate rocks, allowing the relatively low-Os sediment melts to overprint the Os isotopic signature of the mantle to produce a radiogenic Os pyroxenite. Evidence for this process has been reported in mantle xenolith suites from beneath modern subduction zones in the Papua New Guinea (Mclnnes et al., 1999). The heavy 18O signatures in the SCVF olivines (this study and Straub et al., 2015) combined with relatively low water content in melt inclusions from the high-Nb Xitle volcano (~1%; Johnson et al., 2009), are further consistent with sediment melt as an important metasomatic agent in the source of the high-Nb SCVF magmas.

6. Implications for petrogenesis

The co-occurrence of the high-Nb and low-Nb magmas within the SCVF indicates that the mantle beneath the area is highly heterogeneous. The relationship observed between 187Os/188Os and 206Pb/204Pb, and the linear correlation between most highly incompatible trace elements (Fig 5) suggest that the low-Nb and high-Nb samples can be related by two component mixing, possibly mixing of melts. Integrating the geodynamic model of the Mexican subduction zone and the chemical and isotopic data obtained in this study, we propose the following scenario for the production of the SCVF magmas (Fig. 14): The NMEP mantle was dragged beneath the volcanic front of the TMVB by corner flow (Luhr et al., 2006). This convecting, depleted NMEP mantle was metasomatized in the forearc region by a H2O-rich fluid derived through dehydration of the subducting Cocos plate, mostly from dehydration of terrigenous sediment with a minor contribution from pelagic sediment and the underlying altered oceanic crust, which formed the mantle source of the low-Nb, calc-alkaline lavas. This metasomatised mantle wedge was then dragged down where it received a second slab component consisting primarily of sediment melt derived by fluxing of the previously dehydrated sediment by deep fluid release from the underlying AOC ± serpentinized mantle. These silica-rich melts infiltrated and reacted with the surrounding metasomatized mantle to form pyroxenite-veined peridotite, which resulted in

169 deposition of metasomatic sulfides with radiogenic Os, producing the source of the high-Nb group magmas. The compositional variations observed in the SCVF reflects mixing between these variably metasomatised mantle sources and/or melts derived therefrom. Future studies to investigate the existence and prevalence of highly radiogenic Os in more primitive, higher-Nb magmas within the volcanic field will be critical to further evaluate the importance of pyroxenite veins in the mantle source of the high-Nb basalts in the SCVF, and perhaps elsewhere in the TMVB.

170

References

Agustín-Flores , J., Siebe, C., Guilbaud, M.-N., 2011. Geology and geochemistry of Pelagatos, Cerro del Agua, and Dos Cerros monogenetic volcanoes in the Sierra Chichinautzin Volcanic Field, south of México City. Journal of Volcanology and Geothermal Research 201, 143–162

Ballhaus, C., Bockrath, C., Wohlgemuth-Ueberwasser, C., Laurenz, V., Berndt, J., 2006. Fractionation of the noble metals by physical processes. Contributions to Mineralogy and Petrology 152, 667-684.

Berly, Thomas J., Hermann, J., Arculus, Richard J., Lapierre, H., 2006. Supra-subduction Zone Pyroxenites from San Jorge and Santa Isabel (Solomon Islands). Journal of Petrology 47-8, 1531-1555.

Blatter , D. L., Carmichael, I.S.E., 1998. Hornblende peridotite xenoliths from central Mexico reveal the highly oxidized nature of subarc upper mantle. Geology 26, 1035-1038 Bloomfield, K., 1975. A late quaternary monogenetic field in central Mexico. Geologic Research, 64: 476-497.

Blatter , D. L., and Carmichael, I.S.E., 2001. Hydrous phases equilibria of a Mexican high-silica andesite: a candidate for a mantle origin? Geochemica, Cosmochimica, Acta 65, 4043-4065.

Bloomfield, K., 1975. A late quaternary monogenetic field in central Mexico. Geologic Research, 64: 476-497.

Brandon, A.D., Creaser, R.A., Shirey, S.B., Carlson, R.W., 1996. Osmium recycling in subduction zones. Science, 272, 861-864.

Cai, Y., LaGatta, A., Golden, S.L., Langmuir, C.H., Gomez-Tuena, A., Pozzo, A.L.M., Carrasco- Nunez, G., 2014. Hafnium isotope evidence for slab melt contributions in the Central

171

Mexican Volcanic Belt and implications for slab melting in hot and cold slab arcs. Chem. Geol. 377, 45–55.

Carmichael, I.S.E, Langer, R.A., and Luhr, J.F., 1996. Quaternary minettes and associated volcanic rocks of Mascota, western Mexio: A consequence of plate extension above a subduction modified mantle wedge. Contributions to Mineralogy and petrology 124, 302-333

Chesley, J., Righter, K., Ruiz, J., 2004. Large-scale mantle metasomatism: a Re-Os perspective. Earth and Planetary Science Letters 219, 49-60.

Chesley, J., Ruiz, J., 1998. Crust-mantle interaction in large igneous provinces: implications from the Re-Os isotope systematic of the Columbia River flood basalts. Earth and Planetary Science Letters 154, 1-11.

Chesley J., Ruiz J., Righter K., Ferrari L., Gomez-Tuena A., 2002. Source contamination versus assimilation: an example from the Trans-Mexican Volcanic Arc. Earth Planet. Sci. Lett. 195, 211–221.

Connelly, J.N., Ulfbeck, D.G., Thrane, K., Bizzarro, M., Housh, T., 2006. A method for purifying Lu and Hf for analyses by MC-ICP-MS using TODGA resin. Chem. Geol. 233, 126– 136.

Connor, C.B., Conway, F.M., 2000. Basaltic volcanic fields. In: Sigurdsson, H. (Ed.), Encyclopedia of Volcanoes, Academic Press, New York, pp. 331-343.

Cosky, B.W. 2010. Os-Sr-Nd-Pb isotopic and trace element study of magmatic processes within the Sierra Del Chichinautzin volcanic field, Trans-mexican volcanic belt. Master of Science, Miami University, Geology.

172

Davidson, J. P., Turner, S., Handley, H., Macpherson, C. Dosseto, A., 2007. Amphibole ‘sponge’ in arc crust? Geology 35, 787-790.

De Cserna, Z., Fries, C., Rincón-Orta, C., Silver, L.T., Westley, H., Solorio-Munguia, J., Schmitter-Villava, E., 1974. Datos geocronométricos terciarios de los Estados de México y Guerrero: Boletín Asociación Mexicana de Geólogos Petroleros, 26, 263-273.

Ferrari, L., Rosas-Elguera, J., 1999. Alkalic (OIB type) and calc-alkaline volcanism in the Mexican Volcanic Belt: a case for plume related magmatism and propagating rifting at an active margin? Comment to the article by Marquez, A., Oyarzun R., Doblas, M., Verma, S.P. Geology, 27, 1055-1056.

Fries, C. 1960. Geologia de Estado de Morelos y partes adyacentes de Mexico y Guerrero, region central meridional de Mexico. Boletín del Instituto de Geología, UNAM 60, 1–236.

Gerbode, C. and Dasgupta, R., 2010. Carbonate-fluxed melting of MORB-like pyroxenite at 2.9 GPa and genesis of HIMU ocean island basalts. Journal of Petrology 51, 2067-2088.

Gómez-Tuena, A., Langmuir, C.H., Goldstein, S.L., Straub, S.M., Ortega-Gutiérrez, F., 2007. Geochemical Evidence for Slab Melting in the Trans-Mexican Volcanic Belt. J. Petrol. 48, 537- 562.

Griffin, W.L., Graham, W.S., O’Reilly, S.Y., Pearson, N.J., 2004. Lithosphere evolution beneath the Kaapvaal Craton: Re-Os systematics of sulfides in mantle derived peridotites. Chemical geology 208, 89-118.

173

Grove, T.L., Chatterjee, N., Parman, S.W., Médard E., 2006. The influence of H2O on mantle wedge melting. Earth and Planetary Science Letters 249, 74-89.

Guilbaud, M.-N., Siebe, C. & Agustin-Flores, J. (2009). Eruptive style of the young high-Mg basaltic-andesite Pelagatos scoria cone, southeast of Mexico City. Bulletin ofVolcanology 71,859-880, doi: 10.10071s00445-009-0271-0.

Harvey, J., Gannoun, A., Burton, K.W., Schiano, P., Rogers, N.W., Alard, O., 2010. Unravelling the effects of melt depletion and secondary infiltration on mantle Re-Os isotopes beneath the French massif central. Geochemica et Cosmochimica Acta 74, 293-320.

Hart, S.R., Ravizza, G.E., 1996. Os partitioning between phases in lherzolite and basalt, in Earth Processes: Reading the isotopic code, Geophysi. Monogr. Ser., vol 95, edited by A. Basu and S.Hart, 123-134, AGU, Washington, D.C.

Herzberg, C. (2011). Identification of source lithology in the Hawaiian and Canary Islands: implications for origina. Journal of Petrology 52, 113-146.

Hildreth, W., Moorbath, S., 1988. Crustal contributions to arc magmatism in the Andes of Central Chile. Contributions to Mineralogy and Petrology 98, 455-489.

Hirschmann, M.M., Kogiso, T., Baker, M.B., Stopler, E.M., 2003. Alkalic magmas generated by partial melting of garnet pyroxenite. Geology 31, 481-484.

Hirschmann, M.M., Stopler, E.M., 1996. A possible role for garnet pyroxenite in the origin of the garnet signature in MORB. Contributions to Mineralogy and Petrology 124, 185-208

Hofmann, A. W., Jochum, K. P., Seufert, M. White, W. M., 1986. Nb and Pb in oceanic basalts: New constraints on mantle evolution, Earth Planet. Sci. Lett. 79, 33-45.

174

Humphreys, E.R., Niu, Y., 2009. On the composition of ocean island basalts (OIB): The effects of lithospheric thickness variation and mantle metasomatism. Lithos 112, 118-136. Johnson, M.C. and Plank, T., 1999. Dehydration and melting experiments constrain the fate of subducted sediments. Geochem. Geophys. Geosyst. 1, 1999GC000014.

Kepezhinskas. K.B, Kepezhinskas. V.V, Zaitsev, N.S. 1987. Evolution of the Earth`s crust of Mongolia during the Precambrian-Cambrian. Nauka, Moscow, 226pp.

Kessel, R., Schmidt, M.W., Ulmer, P., Pettke, T., 2005. Trace element signature of subduction- zone fluids, melts and supercritical liquids at 120-180 km depth. Nature 437, 724-727.

Klein, E.M., Langmuir, C.H., 1987. Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness. Journal of Geophysical Research 92, 8089-8115.

Kogiso T., Hirschmann, M.M., Frost, D.J., 2003. High-pressure partial melting of garnet pyroxenite: possible mafic lithologies in the source of ocean island basalts. Earth Planet. Sci. Lett. 216, 603-617.

LaGatta, A.B, 2003. Arc Magma Genesis in the Eastern Mexican Volcanic Belt. Columbia University.

175

Luhr, J.F., 1997. Extensional tectonics and the diverse primitive volcanic rocks in the western Mexican Volcanic Belt. The Canadian Mineralogist 35, 473–500.

Luguet, A., Graham P.D., Nowell, G.M., Dreher, S.T., Coggon, J.A., Spetsius, Z.V., Parman, S.W., 2008. Enriched Pt-Re-Os isotope systematics in plume lavas explained by metasomatic sulfides. Science 319, 453

Luguet, A., Lorand, J.-P., Seyler, M., 2003. Sulfide petrology and highly siderophile element geochemistry of abyssal peridotites: a coupled study of samples from the Kane Fracture zone (45⁰W 23⁰20”N, MARK area, Atlantic Ocean) Geochemica et Cosmochimica Acta 67, 1553- 1570.

Martin, C.E., Wasserburg, G.J., Papanastassiou, D.A., Peach, C.L., 1993. Os isotopic composition of sulfide globules from MORB, EOS Trans. AGU 74, 121.

Manea, V.C., Manea, M., 2010. Flat-Slab Thermal Structure and Evolution Beneath Central Mexico. Pure Appl. Geophys. 168, 1475-1487.

Mazzarini, F., Ferrari, L., Isola, I., 2010. Self-similar clustering of cinder cones and crust thickness in the Michoacan-Guanajuato and Sierrade Chichinautzin volcanic fields Trans- Mexican Volcanic Belt. Tectonophysics 486, 55-64.

McInnes, B. I. A., McBride, J. S., Evans, N. J., Lambert, D. D., Andrew, A. S., 1999. Osmium isotope and noble metal recycling in subduction zones: implications for the metallogenesis of porphyry–epithermal Cu–Au deposits. Science 286, 512–516.

176

Meisel, T., Walker, R.J., Morgan, J.W., 1996. The osmium isotopic composition of the primitive upper mantle. Nature 383, 517–520.

Mooser, F., Nairn, A., Negendank, J., 1974. Paleomagnetic investigations of the Tertiary and Quaternary igneous rocks. Geologic Research, 63: 451-483.

Nier, A.O., 1950. A redetermination of the relative abundances of the isotopes of carbon, nitrogen, oxygen, argon and potassium. Phys. Rev. 77, 789–793.

Pearson, D.G., Nowell, G.M., 2004. Re–Os and Lu–Hf isotope constraints on the origin and age of pyroxenites from the Beni Bousera peridotite massif: implications for mixed peridotite– pyroxenite mantle sources. Journal of Petrology 45(2):439-455.

Pearson, D.G., Shirey, S.B., Harris, J.W., Carlson, R.W. 1998. Sulfide inclusions in diamonds from the Kofiefontein kimberlite, S.Africe: constraints on diamond ages and mantle Re-Os systematics. Earth Planetary Science Letters 160 311-326

Peucker-Ehrenbrink, B., Ravizza, G., 2000. The marine osmium record. Terra Nova 12, 205-219.

Prouteau, G., Scaillet, B., 2012. Experimental constraints on sulphur behavior in subduction zones: implications for TTG and adakite production and the global sulphur cycle since the Archean. Journal of Petrology, 202, 1-31.

Rasoazanamparany, C., Widom, E., Siebe, C., Guilbaud, M.-N., Spicuzza, M.J., Valley, J.W., Valdez, G., Salinas, S. in prep. Temporal and compositional evolution of Jorullo volcano, Mexico: Implications for magmatic processes associated with a monogenetic eruption

Rasoazanamparany, C., Widom, E., Valentine, G.A., Smith, E.I., Cortés, J.A., Kuentz, D., Johnsen, R., 2015. Origin of chemical and isotopic heterogeneity in a mafic, monogenetic volcanic field: A case study of the Lunar Crater Volcanic Field, Nevada. Chemical Geology 397, Pages 76–93

177

Reiners, P. W., 2002. Temporal-compositional trends in intraplate basalt eruptions: Implications for mantle heterogeneity and melting processes. Geochem. Geophys. Geosys. 3, 10.1029/2001GC000250.

Reisberg, L., Zindler, A., Marcantonio, F., White, W., Wyman, D., Weaver, B., 1993. Os isotope systematics in ocean island basalts. Earth and Planetary Science Letters 129, 149–167.

Righter, K., Walker, R.J., Warren, P.W., 2000. The origin and significance of highly siderophile elements in the lunar and terrestrial mantles, in: R.M. Canup, K. Righter (Eds), Origin of the Earth and Moon, University of Arizona Press, Tucson, AZ, 2000, 291-322.

Rudnick, R. and Gao, S., 2003. Treatise on Geochemistry: Composition of the Continental Crust, Elsevier Ltd., Chapter 3, 1-64.

Rutherford, M.J., 2008. Magma ascent rates. In: Putirka, K. D. and Tepley, F.J., III (eds) Minerals, Inclusions and Volcanic Processes. American Society of Mineralogy and Geochemical Society, Reviews in Mineralogy and Geochemistry 69, 241-271.

Schaaf, P., Stimac, J., Siebe, C., Macias, J., 2005. Geochemical evidence for mantle origin and crustal processes in volcanic rocks from Popocatepetl and surrounding monogenetic volcanoes, Central Mexico. Journal of Petrology, 46, 1243-1282.

Schiano, P., Birck, J.-L., Allégre, C.J., 1997. Osmium-strontium-neodymium-lead isotopic covariations in mid-ocean ridge basalt glasses and the heterogeneity of the upper mantle. Earth and Planetary Science Letters 150, 363-379.

178

Siebe, C., Rodríguez-Lara, V., Schaaf, P., Abrams, M., 2004b. Geochemistry, Sr-Nd isotope composition, and tectonic setting of Holocene Pelado, Guespalapa and Chichinautzin scoria cones, south of Mexico City. Journal of Volcanology and Geothermal Research 130, 197- 226.

Siebe, C., and Macias, J.L., 2006. Volcanic hazards in the Mexico City metropolitan area from eruptions at Popocatepetl, Nevado de Toluca, and Jocotitlan stratovolcanoes and monogenetic scoria cones in the Sierra Chichinautzin Volcanic Field. Geological Society of America, Special Paper #402: 253-329.

Snow, J.E., Reisberg, L., 1995. Os isotopic systematics of the MORB mantle: results from 1062 altered abyssal peridotites. Earth Planet. Sci. Lett. 133, 411–421.

Snyder, D.C., 2005. Processes and time scales of differentiation in silicic magma chambers: chemical and isotopic investigations. (PhD dissertation), Miami University (227p).

Sobolev, A.V., Hofmann, A.W., Sobolev, S.V., Nikogosian, I.K., 2005. An olivine-free mantle source of Hawaiian shield basalts. Nature 434(7033):590-597.

Sobolev, A.V., Hofmann, A.W., Kuzmin, D.V., Yaxley, G.M., Arndt, N.T., Chung, S.-L., Danyushevsky, L.V., Elliot, T., Frey, F.A., Garcia, M.O., Gurenko, A.A., Kamenetsky, V.S., Kerr, A.C., Krivolutskaya, N.A., Matvienkov, V.V., Nikogosian, I.K., Rocholl, A., Sigurdsson, I.A., Sushchevskaya, N.M., Teklay, M., 2007. The amount of recycled crust in sources of mantle-derived melts. Science, 316, 412-417.

Sobolev, A.V., Hofmann, A.W., Brugmann, G., Batanova, V.G. and Kuzmin, D.V. (2008). A quantitative link between recycling and osmium isotopes. Science 321, 536.

179

Straub, S.M., LaGatta, A.B., Martin-Del Pozzo, A.L. Langmuir, C.H., 2008. Evidence from high-Ni olivines for a hybridized peridotite/pyroxenite source for orogenic andesites from the central Mexican volcanic belt.

Straub, S.M., Gomez-Tuena, A., Stuart, F.M., Zellmer, G.F., Espinasa-Perena, R., Cai, Y., and Iizuka, Y., 2011, Formation of hybrid arc andesites beneath thick continental crust. Earth Planet. Sci. Lett. 303, 337–347.

Straub, S.M., Gomez-Tuena, A., Zellmer, G.F., Espinasa-Perena, R., Stuart, F.M., Cai, Y., Langmuir, C.H., Martin-Del Pozzo, A., Mesko, G.T., 2013a. The processes of melt differentiation in arc volcanic rocks; insights from OIB-type arc magmas in the central Mexican Volcanic Belt. Journal of Petrology 54, 665-701.

Straub, S.M., Gómez-Tuena, A., Bindeman, I.N., Bolge, L.L., Brandl, P.A., Espinasa-Perena, R., Solari, L., Stuart, F.M., Vannucchi, P., Zellmer, G.F., 2015. Crustal recycling by subduction erosion in the central Mexican Volcanic Belt. Geochimica et Cosmochimica Acta 166, 29-52.

Suzuki, K., Aizawa, Y., Tatsumi, Y., 2000. Osmium transport during dehydration processes in the subducted slab: experiments and implications for the Os isotopic compositions of arc Magmas. Journal of Conference Abstracts 5, 970.

Tatsumi, W., Eggins, S., 1995. Subduction Zone Magmatism. Blackwell Sci., Malden, Mass. Vinnik, L.P., 1977. Detection of waves converted from P to SV in the mantle. Phys. Earth Planet. Int. 15, 39–45.

180

Tatsumi, W., and Hanyu, T., 2003. Geochemical modeling of dehydration and partial melting of subducting lithosphere: toward a comprehensive understanding of high-Mg andesite formation in the Setouchi volcanic belt, SW Japan. Geochemistry Geophyiscs Geosystems 4. Doi:10.1029/2003GC000530.

Thirlwall, M.F., Gee, M.A.M., Lowry, D., Mattey, D.P., Murton, B.J., Taylor, R.N., 2006. Low δ18O in the Icelandic mantle and its origins: evidence from Reykjanes Ridge and Icelandic lavas. Geochemica et Cosmochimica Acta, 70, 993-1019.

Thirwall M.F., Graham, A.M., Arculus, R.J., Harmon, R.S., Macpherson, C.G., 1996. Resolution of the effects of crustal assimilation, sediment subduction, and fluid transport in island arc magmas: Pb-Sr-Nd-O isotope geochemistry of Grenada, Lesser Antilles. Geochemica et Cosmochimica Acta 60 4785-4810.

Velasco-Tapia, F., Verma, S.P., 2013. Magmatic processes at the volcanic front of Central Mexican Volcanic Belt: Sierra de Chichinautzin Volcanic Field (Mexico). Turkish Journal of Earth Sciences, 22: 32-60.

Vervoort, J.D., Patchett, P.J., Albarède, F., Blichert-Toft, J., Rudnick, R., Downes, H., 2000. Hf- Nd isotopic evolution of the lower crust. Earth and Planetary Science Letters 181, 115-129.

181

Walker, R.J., Carlson, R.W., Shirey, S.B., Boyd, F.R., 1989. Os, Sr, Nd and Pb isotope systematics of southern African peridotite xenoliths: implications for the chemical evolution of subcontinental mantle. Geochimica et Cosmochimica Acta 53, 1583–1595.

Widom, E., 2011. Recognizing recycled osmium. Geology 39, 1087-1088.

Widom, E., Shirey, S.B., 1996. Os isotope systematics in the Azores: implications for mantle plume sources. Earth and Planetary Science Letters, 142: 451-465.

Widom, E., Kepeshinskas, P., Defant, M., 2003. The nature of metasomatism in the sub-arc mantle wedge: evidence from Re-Os isotopes in Kamchatka peridotite xenoliths. Chemical Geology 196, 283-306.

Xiong, Y., Wood, S.A., 2000. Experimental quantification of hydrothermal solubility of platinum-group elements with special reference to porphyry copper environments, Mineral. Petrol. 68, 1-28.

Yang, Zong-F. and Zhou, Jun-H., 2013. Can we identify source lithology of basalt? Scientific Reports 3, 1856. DOI: 10.1038/srep01856.

182

183

184

Figure 1. Tectonic map of the Trans-Mexican Volcanic Belt (yellow), modified after Siebe et al. (2004). The open box represents the location of the Sierra Chichinautzin Volcanic Field (SCVF).

185

186

Figure 2. Simplified geologic map of the SCVF with sample locations, select monogenetic cinder cones within the volcanic field, and the stratovolcanoes that surround the area. Sample locations are indicated by the sample numbers (after Cosky 2010).

187

Figure 2

188

Figure 3. Whole rock major elements vs. MgO (wt. %) for SCFV samples. Samples from this study shown as green diamonds (high-Nb) and purple circles (low-Nb). Literature data from Wallace and Carmichael (1999), Luhr et al. (2006), Straub et al. (2008, 2011, 2013, 2015) shown as gray diamonds and circles (high-Nb and low-Nb samples, respectively).

189

190

Figure 4. Trace elements vs. MgO (wt. %) for mafic SCVF samples (MgO > 7.5 wt. %). Symbols and literature data sources as in Figure 3.

191

192

Figure 5. Incompatible element-element plots for mafic SCVF samples (MgO > 7.5 wt. %). All symbols and data sources as in Fig. 3.

193

194

Figure 6. Multi-element diagram for selected SCVF samples including low- and high-Nb samples from this study and from Straub et al., 2015. Samples are normalized to normal mid- oceanic ridge basalt (NMORB) values of Sun and McDonough (1989).

195

196

Figure 7. Isotope variation diagrams for the SCVF samples including a) Sr-Nd isotopes, b) & c) Pb isotopes, and d) Hf-Nd isotopes. Literature data for the TMVB (gray symbols; from Gomez- Tueña et al., 2003; Luhr et al., 2006; Cai et al., 2014; Straub et al. 2008, 2011, 2013, 2015) are shown for comparison. Also shown are: terrigenous sediment (blue crosses) and pelagic sediment (open orange crosses) from DSDP site 487 (LaGatta, 2003), and the average altered oceanic crust (open pink cross) from DSDP site 487 (Verma, 2000). Mantle xenoliths and lower crustal xenoliths (stars and “Y”s, respectively) are compiled from Schaaf et al. (1994). Lower crust Hf-Nd data from Vervoort et al. (2000). Continental basement drilled at Sites 493 (granodiorite) and sites 489 (Biotite gneiss) are from Straub et al., (2015)

197

198

Figure 8. 187Os/188Os vs. a) MgO, b) Ni, and c) Os concentrations. Samples from Jorullo volcano (Rasoazanamparany et al., in prep) shown for comparison. DMM and PUM values for 187Os/188Os from Snow and Reisberg (1995) and Meisel et al. (1996).

199

200

Figure 9. 187Os/188Os, 206Pb/204Pb and 87Sr/86Sr versus Ce/Pb and Ba/Zr showing the positive correlation among samples from this study and literature samples (sources as in Fig. 3).

201

202

Figure 10. a) 187Os/188Os versus 206Pb/204Pb for the SCVF samples from this study. DMM and PUM values for 187Os/188Os shown for comparison (Snow and Reisberg, 1995; Meisel et al., 1996). b) Model curves indicate mixing trends for NMEP mantle with a radiogenic Os and Pb subduction component (dashed curve), and mantle-derived basalt with various crustal assimilants including local limestone basement, average upper crust, and average lower crust (solid curves). None of the mixing curves for crustal assimilation can reproduce the data.

203

204

Figure 11. Oxygen isotope vs. Ni concentration for SCVF samples from this study with literature data (gray symbols; data sources as in Fig. 3) and Jorullo volcano (Rasoazanamparany et al., in prep; gray field) shown for comparison.

205

206

Figure 12. FeO/CaO-3MgO/SiO2 (FC3MS) vs. MgO, modified after Yang and Zhou (2013). Dashed line indicates the boundary between pyroxenite and peridotite melts, which lie above and below the line, respectively.

207

208

Figure 13. Os versus Pb isotopic compositions. The shaded area represents the Os isotopic composition of the depleted mantle (DMM) from Snow and Reisberg (1995). The “Fluid” curve represents a mixture of an NMEP mantle source and fluid from trench sediment and altered oceanic crust. The “Melt” curve represents a mixture between an NMEP mantle source and a sediment melt produced by fluxing of sediment with fluid from AOC and serpentinized mantle. Sediment melt-mantle reaction produces a pyroxenite. The “Pyroxenite” mixing curve represents a mixture between the resulting pyroxenite and a hydrated NMEP mantle that has received ~5% hydrous fluid. Mixtures are based on the following parameters: NMEP: Os = 3000 ppt, Pb = 0.025 ppm and 187Os/188Os = 0.1275 and 206Pb/204Pb = 19.4; Terrigeneous sediment: Os = 30 ppt, Pb = 30 ppm and 187Os/188Os = 1.3, 206Pb/204Pb = 18.8, Pelagic sediment: Os = 30 ppt, Pb = 50 ppm and 187Os/188Os = 1, 206Pb/204Pb = 18.5; AOC: Os = 50 ppt, Pb = .069 ppm and 187Os/188Os = 1.76 based on 23Ma oceanic crust, 206Pb/204Pb = 18.22, serpentinized mantle: Os = 3000 ppt, Pb = .025 ppm and 187Os/188Os = .1275, 206Pb/204Pb = 18.2. Mobility of 13% is used for Os. Trench sediment and AOC compositions are from Lagatta (2003) and Verma (2000).

209

210

Figure 14. Cartoon showing magma genesis beneath the SCVF. The Northern Mexican Extensional Province mantle was dragged beneath the volcanic front of the TMVB by corner flow (Luhr et al., 2006). This convecting depleted NMEP mantle was metasomatized in the forearc region by a H2O-rich fluid derived through dehydration of the subducting Cocos plate, mostly from dehydration of terrigenous sediment with a minor contribution from pelagic sediment and the underlying altered oceanic crust, which formed the mantle source of the low- Nb, calc-alkaline magmas. This metasomatised mantle wedge is then dragged down where it receives a second slab component consisting primarily of terrigenous sediment melt derived as a result of fluid-fluxing of the previously dehydrated sediment during deep fluid release from the underlying AOC ± serpentinized mantle. Metasomatism of the hydrated NMEP mantle with this melt produces reaction pyroxenite veins, which are the source of the SCVF high-Nb magmas.

211

212

Electronic Appendix 1. Table 1. Compositions and partition coefficients used in modeling

2Terrigenous 2Pelagic 3Mobility 1Mantle Sed. Sed. 2AOC AOC 4Dsed/fluid 5Dsed/melt Pb 0.025 30.88 58.08 0.69 0.89 0.94 1.29 Nd 0.518 23.35 60.5 4 0.28 1.44 1.46 Sr 8 164 339 63 0.85 0.91 1.23 Os (ppt) 3500 30 30 50 0.13 87Sr/86Sr 0.703 0.7084 0.7085 0.7032 143Nd/144Nd 0.513 0.51251 0.51254 0.51310 206Pb/204Pb 19.400 18.801 18.504 18.220 187 188 Os/ Os 0.1275 1.3 1 1.76 1 Trace elements from average depleted mantle (DMM) of Salters & Stracke, 2004. 2 Average trace element and isotopic compositions of terrigenous, pelagic and altered oceanic crust determined for DSDP sites 487 (Lagatta, 2003 and Verma, 2000) 3 Average trace element mobility, 700 ⁰C, 4 GPa dataset of Kessel et al. (2005a) 4 Partition coefficients 650 ⁰C, 2 GPa dataset of Johnson & Plank (1999) 5 Partition coefficients 900 ⁰C, 2 GPa dataset of Johnson & Plank (1999)

213

CHAPTER 5

Summary

This chapter summarizes the key results and implications for the three studies that comprise this dissertation. Mafic monogenetic volcanic systems frequently exhibit chemical and isotopic heterogeneity both at the volcanic field scale as well as on the scale of a single eruption. Although in many cases single monogenetic eruptions become progressively more silica- rich as they progress, our studies show that these compositional variations, as well as those on the volcanic field scale, are generally inconsistent with shallow level contamination. Rather, in all three studies, the chemical heterogeneity is interpreted to reflect mantle source heterogeneity. Although the processes by which these heterogeneities (as well as those in monogenetic volcanic fields globally) are generated may differ depending on tectonic setting, the present studies show that there are important commonalties. Despite the distinct tectonic settings of the studies in this dissertation - the Basin and Range extensional province, and the Mexican convergent margin in two regions that differ in slab age, subduction rate and depth to slab - all three volcanic fields exhibit trace element and isotopic variations that are indicative of mixing of multiple geochemically distinct mantle sources that have resulted from recycling of crustal material into the mantle. These studies further demonstrate that the processes, and in some cases the timing, of the mantle enrichment events can be clearly preserved in mafic monogentic volcanic systems. It is likely that the small degrees of melting and rapid source to surface transport that prevail in these small, mafic volcanic systems help to preserve detailed evidence of their petrogenetic histories that is often masked or obliterated in more voluminous, silicic volcanic systems. Detailed studies of mafic monogenetic volcanic fields such as those presented here, in which field evidence serves as a foundation for geochemical and isotopic investigations, are therefore key to unraveling processes of mantle source enrichment and heterogeneity in time and space. In all three of these studies, quantitative or semi-quantitative models for the origin of the observed mantle source heterogeneity were developed based on trace element and

214 isotopic compositions of the volcanic rocks. A common feature of the model results in both the extensional and modern subduction settings is that subducted crustal components are required. The OIB-like signatures observed in the LCVF magmas, which occur in an extensional setting, require incorporation of ancient recycled oceanic crust (0.8 Ga) ± minor hydrous fluid derived from subducted altered oceanic crust in the mantle source region. In addition, in the LCVF study, exciting fundamental finding was that magmas with distinct chemical and isotopic compositions erupted essentially in a same geographic location, strongly suggesting that the mantle heterogeneity is on a scale less than a kilometer. Field investigations combined with chemical and isotopic compositional studies reveal that the compositional changes may be primarily a function of time rather than space. Younger volcanic rocks (<140 ka) seem to derive from EM like mantle source whereas older volcanic rocks (>140 ka) were generated by partial melting of pyroxenite veined mantle. Our results also demonstrate that the depth of melting increased with time, suggesting that the mantle source is vertically heterogeneous. Based on the observed trace element and isotopic compositions of the volcanic rocks, the broad compositional spectrum within the volcanic field can be explained by variable degrees of mixing between the two distinct mantle sources. Many monogenetic volcanic fields exhibit broadly similar behaviors to the LCVF, such as the Auckland Volcanic Field (McGee et al., 2013), Buffalo Valley volcanic field (Cousens et al., 2013) and Jeju Island (Brenna et al., 2012) monogenetic system, suggesting that these processes may be common in monogentic volcanic fields globally, in a range of tectonic settings. The heterogeneous geochemical and isotopic signatures in the modern TMVB subduction system further show complexity in the recycled materials and their fluxes to the mantle wedge. For example, the Jorullo eruption documents incorporation of hydrous fluid derived from both sediment and altered oceanic crust to the mantle wedge in multiple fluid fluxing events. The sequential tapping of this mantle wedge, metasomatized by at least two distinct slab-derived fluid fluxes, led to the production of two distinct parental magmas. Our results document, therefore, that the mantle source of individual eruptions can be highly heterogeneous on a small spatial scale. Furthermore, the Jorullo study indicates that mantle metasomatism is dominated by sediment-derived hydrous fluid rather than altered oceanic lithosphere-derived fluid, and that no slab melt

215 component is required, despite previous suggestions of slab melting in this area. Detailed field mapping allowed constraints on the variations with time in the Jorullo eruption, demonstrating that the early magmas are relative less evolved compare to the late magmas, but that they are also isotopically distinct. The progressive increase in SiO2 with time commonly found in monogenetic eruptions are often interpreted to result from progressive tapping a pyroxenite-veined mantle (e.g. Reiners 2002; Straub et al., 2011). Our results at Jorullo suggest that the temporal-compositional variation is indeed attributed in part to changes in mantle source over time, but that this is not the cause of the variation in silica content of the erupted magmas. Rather, evidence from major and trace element data suggest that two distinct magma batches, produced by melting of isotopically distinct mantle domains, evolved separately and to variable degrees. Because they had similar initial major and trace element compositions, there is the appearance of a single evolving magma body being erupted through time. Importantly, this study documented that some of the early-formed melt underwent extensive fractionation before erupting as the final eruptive product, providing strong evidence for the development of a magma chamber; this may be an important mechanism of fractionation beneath monogenetic systems in general, especially those in which there is a significant variation in silica content and other indices of fractionation. In the Sierra de Chichinautzin volcanic field, we have observed two distinct geochemical groups that also indicate heterogeneous mantle beneath the area. A key finding is that the co-existing low Nb, calc-alkaline and the high-Nb alkaline/transitional magmas can be genetically related to one another, despite previous suggestions of different pre-subduction mantle wedge compositions, with one an anomalous OIB-type mantle not commonly associated with subduction zones. Our data indicate that the mantle source of the low-Nb magmas is derived from incorporation of hydrous fluid derived mostly from sediment into the mantle wedge beneath the area, in essentially the identical process to that proposed for the MGVF. In contrast, the mantle source of the high-Nb magmas was formed by infiltration and reaction of silica-rich melts, most likely sediment melts, to the metasomatized mantle source of the low-Nb magmas, which resulted in the formation of pyroxenite-veined mantle. The compositional spectrum within the volcanic

216 field seems to require mixing between the two end-member mantle sources; hydrated mantle and pyroxenite veined mantle. In each of these studies, field constraints combined with major and trace element and Sr-Nd-Pb-Hf-Os ±oxygen isotopic analysis has allowed us to establish a detailed petrogenesis of the monogenetic system and to document the scale and origin of heterogeneity in the mantle. In the TMVB in particular, oxygen isotopes have allowed us to determine the minimum amount of slab component being added to the mantle wedge. In addition, Sr and Pb isotopic data have allowed us to identify the type of subducted materials (oceanic crust and/or sediment) and Nd, Pb and Hf allow the determination of the type of sediment involved (terrigenous vs. pelagic sediment). Major elements such as 187 188 TiO2, SiO2, trace element such Nb, Y and Zr and elevated Os/ Os and Pb isotopic data in the SCVF high-Nb magmas allow the identification of pyroxenite component in their source. However, in order to obtain more precise estimates of the compositions of the subducted components returned to the mantle (i.e. fluid melt, oceanic crust, sediment), a better understanding of subduction zone systems and processes are required. Understanding the slab geometry and the thermal structure of a given subduction zone is critical to evaluate whether the slab dehydrates or melts. Investigation of the pre-eruptive

H2O contents of arc magmas is complex, due to questions about the faithfulness of melt inclusions to preserve this information (Danyushevsky et al., 2002; Bucholz et al., 2013), but it is crucial for determining the nature (hydrous fluid vs. melt) of the subduction component being added to the mantle wedge and has important implications for the degree of partial melting in the mantle. Currently, a major source of uncertainty in quantitative modeling is the lack of accurate experimentally derived partition coefficients (D) for all trace elements, which is due at least in part to the technical difficulty of conducting experiments at temperatures and pressures relevant to subduction zones (Kessel et al., 2005). Currently, the uncertainties in the experimentally determined partition coefficients utilized in many models of subduction systems leave room for a range of interpretations regarding the nature of the recycled material (fluid versus melt) as well as the amount of subducted crust incorporated into the mantle. Furthermore, most experimental studies have not reported partition coefficients for the element Hf, which

217 some studies argue is completely immobile in hydrous fluid, and mobilized only in cases of sediment or slab melting (e.g. Tollstrup et al., 2005; Barry et al., 2006), but others argue for substantial fluid mobility (e.g. Woodhead et al., 2001). Hf isotopes could therefore be a key tracer to distinguish between these fundamental processes, and to help constrain the thermal structure of subduction zones, but its behavior is still not well enough documented experimentally (e.g. Johnson and Plank, 1999). Obtaining accurate partition coefficient is thus crucial for geochemical modeling and to improve our understanding of mass transfer from the slab to mantle wedge. The primary objective of this dissertation was to explore the petrogenesis of monogenetic systems to identify, and where possible to quantify, the processes that lead to mantle source heterogeneity through the use of field constraints and elemental and isotopic compositions. New insights have been gained, particularly regarding the scale of the heterogeneity in the mantle and the role of subduction components in the origin of source heterogeneity. These studies have improved our understanding of the nature and complexity of magmatic processes in the development of mafic monogenetic magmas in extensional and subduction settings. Nonetheless, the advances made here are only one piece of the puzzle, and future studies that integrate geophysical data to help assess thermal structure, water content and depth to slab will be important. Although the small scale of monogentic volcanic systems may test the limits of geophysical techniques, geophysical data might also be able to provide better constraints on the nature of the plumbing systems beneath individual monogenetic centers and volcanic fields, and the size of the partially molten regions and/or magma chambers.

218

References

Barry, T.L., Pearce, J.A., Leat, P.T., Millar, I.L., Roex, A.P., 2006. Hf isotope evidence for selective mobility of high-field-strenght elements in a subduction setting: South Sandwich Islands. Earth Planet. Sci. Lett. 252, 223-244.

Brenna, M., Cronin, S. J., Smith, I. E. M., Maas, R., Sohn, Y. K., 2012. How small- volume basaltic magmatic systems develop: a case study from the Jeju Island Volcanic Field, Korea. Journal of Petrology 53, 985-1018.

Cousens, B., Wetmore, S., Henry, D.C., 2013. The Pliocene-Quaternary Buffalo Valley volcanic Field, Nevada: Post-extensional, intraplate magmatism in the north-central Great Basin, USA. Journal of Volcanology and Geothermal Research 268, 17-35.

Kessel, R., Schmidt, M.W., Ulmer, P., Pettke, T., 2005. Trace element signature of subduction-zone fluids, melts and supercritical liquids at 120-180 km depth. Nature 437, 724-727.

Johnson, M.C. and Plank, T., 1999. Dehydration and melting experiments constrain the fate of subducted sediments. Geochem. Geophys. Geosyst. 1, 1999GC000014.

McGee, L.E., Smith, I.E.M., Millet, M.-A., Handley, H.K., Lindsay, J.M., 2013. Asthenospheric control on melting processes in a monogenetic basaltic system: a case study of the Auckland Volcanic Field, New Zealand. Journal of Petrology 54 (10), 2125- 2153, doi: 10.1093/petrology/egt043.

Reiners, P.W., 2002. Temporal-compositional trends in intraplate basalt eruptions: implications for mantle heterogeneity and melting processes. Geochemistry Geophysics Geosystems 3

219

Tollstrup, D.L., and Gill, J.B., 2005. Hafnium systematics of the Mariana arc: evidence for sediment melt and residual phases, Geology 33, 737–740.

Woodhead, J.D., Hergt, J.M., Davidson, J.P., Eggins, S.M., 2001. Hafnium isotope evidence for ‘conservative’ element mobility during subduction zone processes, Earth Planet. Sci. Lett. 192 331–346

220