1 UNIVERSIDAD NACIONAL AUTÓNOMA DE MÉXICO POSGRADO EN CIENCIAS DE LA TIERRA CENTRO DE GEOCIENCIAS VULCANOLOGIA ORIGEN Y EVOLUCIÓN DE LA CALDERA DE ILOPANGO, EL SALVADOR (CENTROAMÉRICA): UN SUPERVOLCÁN ACTIVO CON MÚLTIPLES ERUPCIONES EXPLOSIVAS CUATERNARIAS Tesis con base en artículos científicos para optar por el grado de: DOCTOR EN CIENCIAS DE LA TIERRA PRESENTA IVAN SUÑÉ PUCHOL Director de Tesis: Dr. Gerardo de Jesús Aguirre Díaz Juriquilla, Querétaro, Agosto 2019 UNAM – Dirección General de Bibliotecas Tesis Digitales Restricciones de uso DERECHOS RESERVADOS © PROHIBIDA SU REPRODUCCIÓN TOTAL O PARCIAL Todo el material contenido en esta tesis esta protegido por la Ley Federal del Derecho de Autor (LFDA) de los Estados Unidos Mexicanos (México). El uso de imágenes, fragmentos de videos, y demás material que sea objeto de protección de los derechos de autor, será exclusivamente para fines educativos e informativos y deberá citar la fuente donde la obtuvo mencionando el autor o autores. Cualquier uso distinto como el lucro, reproducción, edición o modificación, será perseguido y sancionado por el respectivo titular de los Derechos de Autor. 2 Declaratoria de ética Declaro conocer el Código de Ética de la Universidad Nacional Autónoma de México, plasmado en la Legislación Universitaria. Con base en las definiciones de integridad y honestidad ahí especificadas, aseguro mediante mi firma al calce que el presente trabajo es original y enteramente de mi autoría. Todas las citas de, o referencias a, la obra de otros autores aparecen debida y adecuadamente señaladas, así como acreditadas mediante los recursos editoriales convencionales. 3 Dedicatoria: A toda la gran familia italo-catalana que estamos construyendo con el amor de mi vida… A mis padres Joan y Marisa, por darme fuerza y confianza siempre, gracias por todo… A mi hermano Xavi, la cunyi Nina y al nouvingut Victor, per compendrem i estar allí encara que hagi de ser massa per Skype… Ai suoceri, fratelloni, cugini, zie e zii, grazie a tutti per volermi tanto bene… A toda mi familia y amigos de México, El Salvador, Francia, Estados Unidos, Italia, Catalunya y España, gracias a todos por estar en mi vida y hacer posible toda esta aventura… E sopratutto a te amore mio, grazie per essere la donna piú spettacolare del mondo mundial…ho hem tornat a fer, junts, com sempre and forever…. 4 Agradecimientos Primero que nada quiero agradecerle al Dr. Gerardo Aguirre todo lo que ha hecho por mi durante esta importante etapa de mi vida. Le doy las gracias por haber conseguido sacar lo mejor de mí y por la paciencia que ha necesitado para hacerlo. Gracias por su asesoramiento y por apoyar mi trabajo de campo y de laboratorio con el financiamiento de su proyecto de Ilopango “CONACYT-CB nº 240447. Gracias por la estancia en EUA, los congresos y los artículos que hemos conseguido. Agradezco también a mi comité tutoral durante este doctorado: a la Dra. Lucia Capra, quien ha estado siempre disponible para aconsejarme y participar en todas las reuniones semestrales, y al Dr. Pablo Dávila Harris, quien a parte de asesorarme los 365 días del año, ha estado mil horas conmigo en el campo, me ha enseñado infinidad de cosas de valor incalculable y siempre ha estado listo para escucharme, como el buen amigo en el que se ha convertido para mí. Quiero dar las gracias también al Consejo Nacional de Ciencias y Tecnología - CONACyT, por haberme concedido una de sus becas para estudiar doctorado en el programa de Posgrado de la UNAM, y por haber financiado el proyecto del Dr. Aguirre sobre la caldera de Ilopango (240447). Mis más sincera gratitud a todo el personal del Ministerio de Medio Ambiente y Recursos Naturales – MARN y a la Policia Nacional Civil – PNC de El Salvador, porqué sin su apoyo en la logística y seguridad no se hubiera podido realizar el arduo trabajo de campo que se necesito para conseguir los objetivos de esta investigación. Gracias mi papá y mamá Salvadoreños Guayo y Guadalupe, a los hermanos Paco y Miguel, a la pequeña Emilia, a los tíos que me adoptaron para hacerme sentir como en casa en Sta. Tecla, a Luis, Walter, Demetrio, Fran, Douglas, Manuel, Celina, Victor, Willian, Fito, Jacky, Celia, J.A. Reyes, Don Angel, a todos, muchas gracias por la ayuda y el cariño. Al mismo tiempo quiero agradecer a las personas que hicieron posible mi estancia en el Laboratorio Geocronológico en la Oregon State University de Corvallis (EUA). Muchas gracias a mi amigo el Dr. Dan Miggins por confiar en mi y brindarme esa oportunidad única. Gracias también al Dr. Anthony Koppers por permitirme usar uno de los más avanzados laboratorios de Ar/Ar del mundo. Gracias a Victoria, Anna, Seiko, Shan, Wolfang, James, Jake y a todos los demás colegas de la OSU. 5 Estoy muy agradecido con todos los co-autores que han participado en el trabajo y publicación de los artículos que conforman esta tesis, como mis grandes amigos Pierre Lacan, Dario Pedrazzi, Antonio Costa y Carlos Ortega, quienes me han ayudado en campo, en laboratorio y trabajo de gabinete, así como en la construcción y redacción de los artículos científicos. A mis compañeros y amigos del CGEO que me ayudaron durante todos estos años y he podido compartir experiencias de vida irrepetibles, tanto a nivel académico como a nivel personal. Gracias especialmente a Gonzalo, que se animó a acompañarme en camioneta hasta San Salvador desde Querétaro para recoger las muestras de roca del Ilopango, gracias a Rorra, a Maria Isabel, Jaime, Berlaine, Diego, Gaby, Myrna, Mattia, Saul, Fito, Vania, Kurt, Eliseo, Paola, Paco, Walter, Rosario, Erik, y un largo etc., gracias por los asaditos, los partidos de fut, las fiestas y los follones que nos alegran la vida. Gracias también a todos los académicos y a toda la administración del CGEO, a los técnicos de laboratorio que me han ayudado, al Dr. Alex Iriondo, Manuel, Juan, Azucena, Blanca, Armando, Mariano, Carlos Mendoza, Dr. Juan Pablo Bernal, gracias a todos por haber hecho posible mi maestría y mi doctorado aquí con ustedes, y de una forma tan alegre, feliz e inolvidable. Quiero agradecer también a los doctores … por formar parte del comité de ésta tesis de Doctorado y por ayudarme a mejorar el trabajo. Quiero agradecer también a todas las personas que he conocido en estos años en México y agradecerles, gracias a ti Migue y a toda tu familia Trujillo, gracias Cristian, Elvia, Carla, y a toda vuestra familia también, gracias a todos de todo corazón por tratarme como uno más de la familia. Gracias al grupo de viejetes del Padel con quien me lo pasaba tan bien jugando y riendo un rato, gracias a mi familia catalana de Quretaro: la Ester, Ferran, Joan, Aina, Jordi, Mireya, Abel amigo, Dario, Aldo, Berta y Arnau, Victor, Josep, Joan, Miquel, Oscar, gracies a tots per ser-hi. Por último, quiero volver a agradecer a mi maravillosa familia, la qual he tenido tan lejos durante estos años, pero que gracias a Skype y Watsapp no se hace tan larga la distancia ni tan fuerte la anyoransa. Gracias a mis amigos de toda la vida, a la colla de Batea, al trenet de Masalio, ai amici di Perugia, al Edu, la Vane, a la Fanny, lo Mallo, a l’Agustí y especialmente gracias a mi esposa, al amor de mi vida, la persona más importante que tengo y con quien comparto esta aventura que es vivir, te amo Angela. 6 Índice Resumen …………………………………………………………… 7 Abstract ……………………………………………………………. 9 1. Introducción……………………………………………………. 11 1.1. Marco teórico: calderas y supererupciones explosivas……. 13 1.2. Marco geológico: tectónica, geodinámica y volcanismo en El Salvador……………………………………………… 17 1.3. Motivación, hipótesis y objetivos………………………….. 22 1.4. Contenido de la tesis……………………………………….. 24 2. Metodología…………………………………………………….. 26 2.1. Recopilación bibliográfica y análisis del terreno................... 26 2.2. Trabajo de campo: levantamiento estratigráfico, mapeo y muestreo………………………………………………….. 28 2.3. Envío de muestras y análisis en laboratorio: geocronología, geoquímica y petrografía…………………………………… 30 2.4. Trabajo de gabinete: elaboración de mapas, digitalización de series estratigráficas y modelado numérico……………… 35 3. Artículo 1: Descripción de las primeras erupciones formadoras de extensas ignimbritas por la caldera de Ilopango, una estructura vulcano-tectónica tipo graben/pull-apart……………..................... 37 4. Artículo 2: Revisión estratigráfica de toda la secuencia de la caldera de Ilopango y estimación del periodo de recurrencia para grandes erupciones explosivas…………………………………………….. 57 5. Artículo 3: Estudio vulcano-estratigráfico de la erupción Tierra Blanca Joven (TBJ): caracterización física del mayor evento Holoceno en Centro América……………………………………… 78 6. Discusiones y trabajos futuros…………………………………….. 101 7. Conclusiones………………………………………………………. 106 8. Referencias………………………………………………………… 108 7 Resumen La caldera de Ilopango es una estructura vulcano-tectónica con actividad en el Cuaternario, que causó grandes erupciones explosivas y afectando la parte central de El Salvador. Se originó hace 1.78 Ma, cuando el primer evento de colapso formó una caldera tipo graben, que dispersó una ignimbrita ~350 km3 con un estilo eruptivo “boiling-over”. Esta primera erupción formó la Ignimbrita Olocuilta, y cubre aproximadamente 2,000 km2 de territorio con espesores de hasta 120 m. La cartografía geológica y levantamiento estratigráfico realizada en este estudio, complementada con correlación geoquímica y petrográfica, indican que la caldera de Ilopango produjo como mínimo 13 erupciones explosivas hasta la actualidad. La de mayor volumen fue la Ignimbrita Olocuilta. Las ignimbritas posteriores fueron de menor magnitud, pero todas con volúmenes superiores a 1 km3 y VEI > 6. Cuatro de estas erupciones han ocurrido tan solo en los últimos 57 ka, siendo la última de todas la más conocida y mejor estudiada hasta el momento, la Tierra Blanca Joven (TBJ). Con esta investigación se ha definido por primera vez la historia volcánica explosiva de la caldera de Ilopango y la distribución de los depósitos piroclásticos asociados, así como un modelo sobre el posible origen vulcano-tectónico de la caldera y su evolución geológica. La caldera de Ilopango es parte del Arco Volcánico de Centroamérica, el cual está asociado al magmatismo generado por la subducción de la placa de Cocos por debajo de la placa del Caribe. La caldera mide 17 x 13 km, está parcialmente ocupada por el lago de Ilopango y se localiza a lo largo de la Zona de Falla de El Salvador (ZFES), una franja estrecha y alargada de fallas laterales derechas, conectadas por cuencas “pull- apart”. Concretamente, Ilopango se encuentra dentro de una de estas cuencas: el Pull- Apart de San Salvador. El origen y la evolución de la caldera está estrechamente ligado al desarrollo de éste pull-apart. Una prueba de esta posible relación vulcano-tectónica entre el Ilopango y las fallas regionales es la última erupción de la caldera, cuando se emplazó un domo intracratérico en el centro del lago (las Islas Quemadas) justo después de un terremoto tectónico ocurrido a finales del 1879. El estudio geocronológico realizado en las 13 tobas identificadas, utilizando los métodos U/Pb, 238U/230Th y Ar39/Ar40, ha aportado nueva información para reconstruir la historia volcánica de la caldera de Ilopango. Acorde a la nomenclatura moderna en estratigrafía volcánica, se subdividió el Grupo Ilopango en tres formaciones: la Fm. Comalapa (tres erupciones, 1.78 – 1.34 Ma), la Fm. Altavista (6 erupciones, 918 – 257 8 ka), y la Fm. Tierras Blancas (4 erupciones, <57ka). Los depósitos producidos durante estas erupciones caldéricas son el producto de magmas calcoalcalinos ricos en Si y K (riodacitas), con plagioclasa, hornblenda y piroxeno. Los periodos de recurrencia son muy variables, desde más de 220 ka para las primeras erupciones, a ~100 ka en las erupciones intermedias, y ~20 ka para las últimas erupciones de la caldera. Destacan los largos tiempos de quietud entre las tres formaciones de hasta 400 ka, los cuales pueden estar relacionados a cambios en la tectónica regional de la ZFES. El trabajo de vulcanología física realizado sobre el depósito de la TBJ, la última erupción explosiva de la caldera de Ilopango ocurrida hace tan solo unos 1500 años, detalla el tipo de procesos eruptivos que se originan en este volcán. El hidrovolcanismo producido por la interacción del magma con el agua del lago de Ilopango, determinó la intensa explosividad y la alta fragmentación durante la erupción de la TBJ, hasta tal punto que formó una nuve coignimbritica que envió ceniza hasta los 45 km de altitud. Los más de 30 km3 DRE eyectados a la atmosfera durante esta erupción fueron catastróficos para las comunidades Mayas que vivían en la región. Estudios geológicos completos en calderas volcano-tectónicas y multiepisódicas como la de Ilopango, son indispensables para evaluaciones futuras del peligro volcánico. Esta investigación describe la historia volcánica de la caldera y da una idea del potencial destructivo de la misma en caso de una nueva erupción. Los mapas de distribución obtenidos, así como la recurrencia estimada para las erupciones explosivas del Ilopango, brindan una primera aproximación a la amenaza que representa esta caldera, sobre todo para el área metropolitana de San Salvador y en general para Centroamérica y el sur de México. 9 Abstract The Ilopango caldera is an active volcano-tectonic structure, which caused large explosive eruptions during the Quaternary affecting the central part of El Salvador. The first caldera collapse was at 1.78 Ma, and formed a graben-type caldera, producing a large ignimbrite-forming eruption of ~ 350 km3. This first eruption formed the Olocuilta Ignimbrite, and covers approximately 2,000 km2 with thicknesses up to 120 m. The geological cartography carried out in this study, complemented with geochemical and petrographic correlation, indicate that the Ilopango caldera produced at least 13 explosive eruptions until present day. The largest volume was that of the Olocuilta Ignimbrite. The later ignimbrites were of smaller magnitude, but all with volumes larger than to 1 km3 and VEI> 6. Four of these eruptions have occurred in the last 57 ka, being the best known and most studied so far, as the last Tierra Blanca Joven (TBJ). This research has defined for the first time the explosive volcanic history of the Ilopango caldera and the distribution of the associated pyroclastic deposits, as well as a model on the possible vulcano-tectonic origin of the caldera and its geological evolution. Ilopango caldera is part of the Central America Volcanic Arc, which is associated to the magmatism generated by the subduction of the Cocos plate beneath the Caribbean plate. The caldera’s size is 17 x 13 km, and is partially occupied by the Ilopango Lake. It is located along the El Salvador Fault Zone (ESFZ), a narrow and elongated strip of right lateral-faults, connected by pull-apart basins. Specifically, Ilopango is located within one of these basins: the Pull-Apart of San Salvador. The origin and evolution of the caldera is closely linked to the development of this pull- apart. A proof of this possible volcano- tectonic relationship with Ilopango and the regional faults is the last eruption of the caldera, when an intra-crateric dome was emplaced in the centre of the lake just after a tectonic earthquake occurred at the end of 1879 (the eruption of the Islas Quemadas, which means Burned Islands). The geochronological study conducted on the 13 identified tuffs, using the U-Pb, 238U/230Th and Ar39/Ar40 methods, has provided new information to reconstruct the volcanic history of the Ilopango caldera. According to the modern nomenclature in volcanic stratigraphy, the Ilopango Group was subdivided into three formations: the Comalapa Fm. (three eruptions, 1.78 - 1.34 Ma), the Altavista Fm. (6 eruptions, 918 - 257 ka), and the Tierras Blancas Fm. (4 eruptions, <57ka). The deposits formed during these calderic eruptions are the product of calc-alkaline magmas rich in Si and K (riodacites), with plagioclase, hornblende and pyroxene. The periods of recurrence are variable, from 10 more than 220 ka for the first eruptions, to ~ 100 ka in the intermediate eruptions, and ~ 20 ka for the last eruptions of the caldera. The long periods of quiescence between the three formations of up to 400 ka stand out, which may be related to changes in the regional tectonics of the ESFZ. The physical volcanology work carried out on the TBJ deposit, the last explosive eruption of the Ilopango caldera that occurred only about 1500 years ago, describe the details of the eruptive processes originate in this volcano. The hidrovolcanism produced by the interaction of the magma with the water of the Ilopango Lake, determined the intense explosiveness and the high fragmentation during the eruption of the TBJ, to such an extent that the coignimbritic cloud reached more than 45 km of altitude. The more than 30 km3 DRE ejected into the atmosphere during this eruption were catastrophic for the Mayan communities that lived in the region. Complete geological studies in these multi-episodic volcano-tectonic calderas, such as Ilopango, are indispensable for future assessments of volcanic hazards. This research work describes the volcanic history of the caldera and provides an idea of the destructive potential of Ilopango in a probable future eruption. The distribution maps obtained, as well as the estimated recurrence for the explosive eruptions of Ilopango, provide a first approximation to the threat represented by this caldera, especially for the metropolitan area of San Salvador and in general for Central America and the south of Mexico. 11 1 Introducción La caldera de Ilopango (CI) se localiza en la parte central de El Salvador, a menos de 10 km de la capital San Salvador, que con más de dos millones de habitantes, es la ciudad más poblada de Centroamérica (Fig. 1). La CI tiene una forma romboédrica, de 17 km de largo por 13 km de ancho, la cual está parcialmente rellenada por el lago de Ilopango, de casi 300 metros de profundidad. Esta caldera está activa y su último evento volcánico fue una erupción efusiva de un domo dacítico emplazado en el centro del lago, que formó las Islas Quemadas en 1879-1880 (Golombek y Carr, 1978; Richer et al., 2004). La última erupción explosiva de la CI ocurrió hace tan solo unos 1500 años y se conoce como la Tierra Blanca Joven (TBJ), la cual produjo un extenso depósito piroclástico blanco rio-dacítico (Williams y Meyer-Abich, 1955; Rose et al., 1999; Hernández, 2004; Kutterolf et al., 2008; Dull et al., 2001; Saxby et al., 2016; Aguirre- Díaz et al., 2017; Pedrazzi et al., 2018). La TBJ, con espesores de hasta 60 m de potencia y un volumen de emisión estimado de ∼ 84 km3 de material piroclástico (Dull et al., 2010), fue una de las erupciones más catastróficas del Holoceno en Centroamérica, cuyos efectos fueron devastadores para los asentamientos del imperio Maya ubicados en las cercanías de la caldera de Ilopango (Sheets, 1979; Dull et al., 2001). Aparte de la TBJ, la CI ha provocado otras tres grandes erupciones explosivas en los últimos 57,000 años (57 ka): la TB4, TB3 y TB2 (TB’s; Rose et al., 1999; Hernández, 2004; Kutterolf et al., 2008). Otras ignimbritas más antiguas, estratigráficamente inferiores a las TB’s, fueron reportadas en la zona por Hernandez et al., (2004). Las calderas de colapso, como la de Ilopango, se forman por el hundimiento de bloques corticales dentro de una cámara magmática somera (Smith y Bailey, 1968; Lipman, 1997, 2000; Gottsmann y Martí, 2008). La subsidencia se produce a lo largo de fallas co-eruptivas, pero en algunos casos, las calderas utilizan fallas regionales preexistentes como discontinuidades corticales para colapsar, como sucedió durante el Oligoceno en la graben-caldera de Bolaños, en la Sierra Madre Occidental, México (Aguirre-Díaz y Labarthe- Hernández, 2003; Aguirre-Díaz et al., 2008). Algo así sucedió en el pasado, y sucede actualmente en la caldera de Ilopango, la cual fue formada en la Zona de Falla de El Salvador (ZFES, Martínez-Díaz, 2004), una franja 12 estrecha que cruza todo el país paralelamente al Arco Volcánico de El Salvador (AVES). Williams y Meyer-Abich (1955) describieron por primera vez la CI como una estructura vulcano-tectónica, la cual habría generado varias erupciones a lo largo de su actividad. Trabajos gravimétricos recientes de Saxby et al. (2016) confirman que la caldera de Ilopango tiene un control tectónico, donde las fallas verticales de la ZFES funcionan como una vía preferencial de extrusión de magma hasta la superficie (Tikoff y St. Blanquat, 1997). Fig. 1: Localización de la caldera de Ilopango (CI) junto al Área Metropolitana de San Salvador (AMSS), la cual se ubica entre ésta caldera y el volcán activo de San Salvador (VSS). Destaca la última colada de lava emitida por el VSS en 1917 (de color negro al nor- noreste del volcán). El punto amarillo en el centro del lago Ilopango representa la ubicación de las Islas Quemadas, producto de la última erupción de la caldera de Ilopango (1879- 1880). En este trabajo proponemos que Ilopango ha tenido grandes erupciones formadoras de extensas ignimbritas, asociadas a diferentes episodios de colapso caldérico (similar al complejo de Platoro en las montañas de San Juan, EUA, Lipman et al., 1996), y todas estas fases estarían relacionadas con la actividad y evolución de las fallas preexistentes de la ZFEZ (Soefield, 2004). La geomorfología del borde topográfico de la caldera de Ilopango, con varias bahías semicirculares, podría ser una evidencia de los múltiples eventos del colapso caldérico (Lexa et al., 2011). Durante esta investigación, se identificaron y caracterizaron todas estas erupciones y sus 13 respectivos depósitos generadas por la CI desde su formación, hace ~1.78 Ma, hasta la actualidad. 1.1. Marco teórico: calderas y supererupciones explosivas asociadas Las calderas de colapso son las estructuras geológicas más peligrosas y las que han provocado las erupciones más voluminosas y catastróficas que se conocen en la Tierra, como por ejemplo la erupción de la caldera de Toba (Sumatra, Indonesia), cuya supererupción ocurrida hace 74 ka estuvo a punto de extinguir a los Homo Sapiens (Francis 1983; Rampino y Self, 1992). Estas superestructuras volcánicas se forman como consecuencia del colapso o subsidencia de grandes bloques corticales dentro de una cámara magmática somera, los cuales provocan la eyección explosiva y rápida de grandes volúmenes de material piroclástico a la atmosfera (Smith y Bailey, 1968; Druitt y Sparks, 1984; Cas y Wright, 1987; Lipman, 1997; Marti et al., 1994; Cole et al., 2005; Gottsmann y Martí, 2008). El resultado morfológico de este colapso caldérico es una depresión en el terreno de grandes dimensiones, rodeada por fallas sin-eruptivas que la bordean por donde se produce el hundimiento de los bloques, dejando a la vista la pared interna de la caldera y sobresaliendo el borde caldérico (Fig. 2). Los productos generados en este tipo de erupciones son mayormente depósitos de caída de pómez y ceniza, flujos piroclásticos formadores de ignimbritas, gases y emplazamientos de domos y coladas de lava post-colapso (Lipman 1984, 2000). Fig. 2: Diagrama de una caldera de colapso tipo Pistón en 3D (Cole et al., 2005), donde se muestra esquemáticamente la estructura de estos complejos volcánicos. 14 Estas estructuras se encuentran en prácticamente todos los ambientes volcánicos de la tierra (Geyer y Martí, 2008; Fig. 3), como zonas de subducción (Toba; Francis 1983; Ilopango; Simkin y Siebert, 1994; Carr et al., 2007), rifft (Etiopia; Acocella et al., 2002) y hot – spots, tanto oceánicos (Las Cañadas, Islas Canarias, España; Schmincke, 1967; Marti y Gudmundsson, 2000), como continentales (Yellowstone, USA; Hildreth et al., 1984). La forma, génesis y el tamaño de una caldera es variable: algunas son circulares y con una sola fase de colapso (por ejemplo Reporoa, en Nueva Zelanda; Nairn et al., 1994; Beresford and Cole, 2000), mientras que otras son más alargadas, con tamaños que van de los pocos kilómetros hasta decenas de kilómetros (como Toba o La Garita, Colorado, USA; Lipmann 2000) y algunas forman parte de complejos caldéricos con varias fases de colapso (como el complejo Platoro, en San Juan Mountains; Lipman et al., 1996). Otras calderas están estrechamente ligadas a la tectónica, donde las fallas regionales condicionan desde la formación hasta el tipo de actividad volcánica de estas calderas (como la graben caldera de Bolaños, en la Sierra Madre Occidental de México, Jalisco; Aguirre- Díaz and Labarthe-Hernández, 2003; Aguirre-Díaz et al., 2008). Fig. 3: Mapa mundial con la localización (circulo negro) de las calderas de colapso conocidas (modificado de Geyer y Martí, 2008). Las erupciones en calderas pueden ser explosivas o efusivas; sin embargo, las más voluminosas y violentas son las explosivas, que pueden emitir grandes cantidades de material piroclástico a la atmosfera en cuestión de horas a pocos días (por ejemplo, Newhall y Dzurisin, 1988; Lipman, 1997; Cole et al., 2005; Gottsmann y Martí, 2008; 15 Costa et al., 2014). Las calderas se pueden clasificar por su composición (basálticas, peralcalinas, andesitico-daciticas o rioliticas; Cole et al., 2005) o también por su estilo y grado de subsidencia (Lipman 1995, 1997, 2000; Acocella 2007, 2008; Fig. 4): 1) Downsag es el primer paso en el colapso gravitacional de una caldera, con una subsidencia limitada donde las fallas anulares del borde no se han formado todavía o no han atravesado por completo el techo de la caldera, plegando y fracturando los materiales de la corteza litosférica (calderas basálticas no-explosivas como las de Hawaii o las Galapagos, Walker 1988); 2) Trap-door es el siguiente paso de madurez en el proceso de colapso caldérico, en donde solamente se ha hundido un sector del techo de la cámara mediante la formación de parte de las fallas inversas que conformarán el anillo del borde (p. ej. Valles caldera, Nuevo México, USA; Heiken et al., 1986); 3) Se conoce como Plate/piston cuando el colapso caldérico es producido por la subsidencia de un único bloque dentro de la cámara magmática, hundido gravitacionalmente una vez desarrolladas por completo las fallas periféricas del borde caldérico (Crater Lake, Oregón, USA; Bacon 1983); 4) Una caldera con avanzado grado de subsidencia también puede colapsar en estilo Piecemeal, en donde el techo de la cámara magmática se rompe en diferentes bloques (p. ej. Glencoe Caldera; Moore and Kokelaar, 1997, 1998); 5) las calderas tipo Funnel mayormente se relacionan a calderas pequeñas (< 2 – 4 km) que al colapsar adquieren una geometría de embudo por medio de un conducto central, con fallas internas empinadas y de fallas de borde ausentes (Guayabo caldera, Costa Rica; Hallinan 1993). Fig. 4: Modelos de los cinco estilos de colapso propuestos por Lipman (1997) Las calderas de colapso también se pueden clasificar dependiendo de su ubicación y relación con el contexto geológico regional donde se emplazó. La clasificación de Aguirre-Diaz (2008), define como caldera Somital, aquellas que se formaron en la parte alta de grandes estratovolcanes y asociadas a pequeños volúmenes emitidos de material 16 piroclástico (por ejemplo, Crater Lake en USA, el Somma – Vesubio en Italia o la caldera de San Pedro en volcán Temascalcingo, Estado de México; Roldán – Quintana y Aguirre – Díaz, 2006). Esta misma clasificación define como calderas Clásicas las que se forman en terrenos llanos, sin necesidad de la existencia de algún tipo de edificio volcánico anterior o estructuras tectónicas regionales (como por ejemplo Long Valley en USA o Los Humeros en el Estado de Puebla, México). El último tipo de caldera en la clasificación de Aguirre (2008), son las calderas tipo Graben, que son aquellas donde el techo de la cámara magmática colapsa a lo largo de fallas tectónicas regionales preexistentes en el terreno por donde se producen grandes erupciones fisurales (Fig. 5). Ejemplos de este tipo son varias calderas de Sierra Madre Occidental de México, como por ejemplo, el graben caldera de Bolaños (Aguirre-Díaz y Labarthe-Hernandez, 2003), o la misma caldera de Ilopango (El Salvador) como se define en este trabajo. Fig. 5: Modelo esquemático que muestra el mecanismo de erupción fisural en las calderas del tipo Graben, en que los flujos piroclásticos salen a través de las fallas normales regionales (Aguirre-Díaz y Labarthe-Hernandez, 2003). Las calderas de colapso de composición silícica son las únicas estructuras volcánicas capaces de provocar supererupciones explosivas, catalogadas como eventos que superan volúmenes emitidos de 450 km3 DRE (Dense Rock Equivalent; Sparks et al., 2005; Self 2006), las cuales pueden llegar a VEI = 8 (el máximo Indice de Explosividad Volcánico; Newhal y Self, 17 1982). Este tipo de erupciones, aunque infrecuentes (una cada ~ 50,000 años), tienen un efecto catastrófico en las regiones circundantes a la caldera debido a lo destructivo que son sus productos (espesas caídas de pómez y enormes flujos piroclásticos), pudiendo incluso afectar a la estabilidad y el desarrollo de sociedades como ocurrió en la erupción de Santorini (Grecia, 1639-1530 a.C), la cual provocó el declive de la civilización Minoica (Nincovich y Hezzen, 1965; Sparks, 1979; Sigurdsson et al., 2006). Las supererupciones llegan a tener afectaciones globales, generando inviernos volcánicos que enfrían el clima terrestre. Inviernos volcánicos como el que ocurrió en la última gran erupción del Pinatubo en 1991 (Hansen et al. 1996), que sin llegar a niveles de supererupción, la ceniza fina y aerosoles dispersados en la estratosfera, provocaron absorción de la radiación solar y la consecuente bajada de temperatura (Self y Blake, 2008). Mucho más drásticos fueron los efectos de la supererupción de Toba, cuyas afectaciones climáticas pudieron acelerar y detonar el inicio de la última gran glaciación que ha sufrido el planeta Tierra (Rampino y Self, 1993). Pero no todo son desventajas con las calderas volcánicas: son estructuras de alto interés económico, ya que la alta actividad geotérmica que producen las convierte en potenciales fuentes de energía renovable (p.ej. la Zona Volcánica de Taupo, Nueva Zelanda; Bibby et al., 1995). Además, son lugares favorables para la formación de depósitos minerales (Guillou – Frottier et al., 2000; Stix et al., 2003). En los últimos 25 años y gracias a la combinación entre el mapeo en calderas erosionadas (Lipman 1984, 1995, 2000a; Branney and Kokelaar, 1994; Aguirre – Díaz et al., 2003; 2007), al trabajo geodésico (Dvorak and Dzurisin, 1997) y al modelado analógico experimental (Marti et al., 1994; Acocella 2007, 2008), se ha adelantado mucho en el conocimiento y entendimiento de los procesos de formación y dinámica del colapso de las calderas volcánicas. 1.2. Marco geológico: tectónica, geodinámica y volcanismo en El Salvador La caldera de Ilopango se ubica en la República de El Salvador y es uno de los volcanes más activos del Arco Volcánico Centroamericano (AVC), el cual se extiende desde la frontera México-Guatemala hasta Costa Rica y forma parte del Anillo de Fuego del Pacífico (Simkin y Siebert, 1994; Carr et al., 2007; Smithsonian Global Volcanism Program, 2013). El territorio salvadoreño se encuentra en la costa Pacífica de la placa del Caribe. A unos 200 km al norte de San Salvador, se encuentra el límite con la placa de 18 Norteamérica, donde las fallas laterales izquierdas del sistema Polochic-Motagua-Swan Islands están desplazando la placa del Caribe hacia el Este a una velocidad promedio de 8 mm/año (Fig. 6a; Agostini et al., 2006; DeMets, 2001; Funk et al., 2009; Guzmán- Speziale et al., 2005). Fig. 6: A) Contexto geodinámico de El Salvador con las diferentes placas tectónicas que conforman Centroamérica. La Zona de Falla de El Salvador (ZFES) y la Depresión de Nicaragua (DN), asociadas a la subducción de la placa de Cocos por debajo de la placa Caribe, transcurren paralelamente al Arco Volcánico de Centroamérica (AVC). FPMSI: Fallas Polochic-Motagua-Swan Islands. B) Esquema 3D donde se muestra el desplazamiento lateral de la cuña trasarco entre la Trinchera Mesoamericana y el AVC, formando a su vez el pull-apart del golfo de Fonseca que separa El Salvador de Nicaragua (modificado de Alvarado et al., 2011). El magmatismo de la caldera de Ilopango y del AVC en general, está relacionado con la subducción oblicua de Cocos por debajo de la placa del Caribe (LaFemina et al., 2009), desarrollada a lo largo de la trinchera mesoamericana a una velocidad promedio de 73 - 85 mm / año (Dixon, 1993; Mann, 2007). Esta subducción oblicua, junto con el efecto cremallera producido en el punto triple de Guatemala (Authemayou et al., 2011) y el desacople de la interfaz provocada 19 por el roll-back del slab de Cocos por debajo de la placa Caribe (Alonso-Henar, et al., 2015; 2017) habrían provocado el inicio del deslizamiento de la cuña forearc (o antearco) hacia el NW, paralela a la trinchera mesoamericana y con una velocidad actual de 8-14 mm/año con respecto a la placa Caribe (Fig. 6; Turner et al., 2007). La combinación de estos dos movimientos relativos: 1) el movimiento hacia el Este de la placa del Caribe que provoca una extensión de trasarco, y 2) el desplazamiento hacia el Noroeste del antearco centroamericano, son aparentemente los responsables de la deformación trans-tensional a lo largo del AVC por medio de una serie de fallas laterales derechas en-echelon y conectadas por cuencas pull-apart (Montero y Dewey 1982; DeMets 2001, La Femina et al., 2002, Corti et al., 2005; Agostini et al., 2006, Turner et al., 2007; Funk et al., 2009; Canora et al., 2014), conocido como la Zona de Falla de El Salvador (ZFES, Martínez-Díaz, 2004). Uno de estos pull-apart es el del Golfo de Fonseca, que transfiere la deformación desde la ZFES hacia la Depresión de Nicaragua (Alvarado et al., 2011, Fig. 6b). La CI se encuentra dentro de una de estas cuencas tectónicas desarrollada en la parte central del país y conocida como el Pull-Apart de San Salvador (Garibaldi, et al., 2016), el cual se delimita al norte por la falla de Guaycume, la falla de la Cordillera del Bálsamo al sur, la falla de Zapotitan al Oeste y la de San Vicente al Este (Fig. 7). Esta última falla, con fuerte componente lateral derecho, está afectando la CI en su flanco oriental y es la causante del último gran terremoto de Mw 6.7 producido en la ESFZ en febrero de 2001 (Alvarenga et al. , 2001; Martínez Díaz et al., 2004; Canora et al., 2012; Fig. 7). Contextos tectónicos regionales como el que aporta el ZFES, el cual incluye fallas de desgarre, zonas de cizalla y cuencas pull-apart, favorecen el emplazamiento de centros magmáticos en niveles superficiales de la corteza y sirven como vías iniciales para el ascenso de magma (Hutton y Reavy, 1992). Litológicamente, El Salvador está compuesto principalmente por rocas volcánicas de edades que van desde el Paleógeno superior hasta el presente (Donelly et al., 1990, Rose et al., 1999). Según el mapa geológico escala 1: 500,000 de El Salvador (Weber et al., 1974), el basamento está compuesto principalmente de calizas Jurásico- Cretácicas de la Formación Metapán, expuestas esporádicamente en la parte Noroeste del país y cubriendo menos del 5% de la superficie total de El Salvador. En el Oligoceno empezó el volcanismo, el cual se concentró en la parte más 20 septentrional del país formando la Montaña Fronteriza: una cordillera formada por los remanentes de los volcanes más antiguos de El Salvador y que delimita de forma natural, la frontera política del este país con Honduras (Fig. 7). Los productos de estos primeros volcanes que se emplazaron hace ~30 Ma, van desde ignimbritas silícicas hasta efusivas básicas a intermedias-acidas que constituyen las formaciones geológicas de Morazán y Chalatenango. Fig. 7: Modelo Digital de Elevacion (DEM) de El Salvador donde se indican las fallas de la Zona de Falla de El Salvador (ZFES) y los principales volcanes del Arco Volcánico de El Salvador (AVES): Caldera de Coatepeque: CC, Volcán Santa Ana: VSA, Volcán San Salvador: VSS, Volcán San Vicente: VSV, Caldera Carboneras: CCa, Complejo Volcánico Berlín: CVB y Volcán San Miguel: VSM. El cuadrado negro nos indica la zona de estudio de esta investigación alrededor de la caldera de Ilopango (CI), la cual se localiza dentro del Pull-Apart de San Salvador, rodeado por la Falla San Vicente (FSV), la Falla Gaycume (FGy), la Falla de la Cordillera del Bálsamo (FCB) y la Falla del Zapotitán (FZp). Desde el Mioceno, la actividad volcánica migró hacia el sur y se acercó a la Trinchera Mesoamericana, como consecuencia del roll-back de la placa subducida de Cocos (Carr 1976; Weinberg 1992, Alvarado et al, 2011, Alonso-Henar et al., 2015; 2017). El vulcanismo desde mediados del Mioceno hasta el Plioceno formó grandes estratovolcanes basálticos y andesíticos justo al sur del actual Arco Volcánico de El 21 Salvador (AVES); como el volcán Panchinmalco, el Jayaque o el mismo antiguo volcán de Ilopango, los cuales pertenecen a la Formación Bálsamo (Weber et al., 1974; Lexa et al., 2011). Durante el Plioceno-Pleistoceno, estos edificios volcánicos colapsaron gravitacionalmente asociados a la formación y evolución de la ZFES, formando las calderas de Plan de Renderos, Jayaque, la Carboneras y la CI (Fig. 8), cuyos productos piroclásticos fueron rellenando las depresiones tectónicas desarrolladas a lo largo de la ZFES y corresponden a la Formación Cuscatlán (Weber et al., 1974). Los relictos de estos grandes estratovolcanes antiguos forman la Cordillera de Bálsamo (Fig. 8, Williams y Meyer-Abich, 1955). Fig. 8: Área de estudio con la caldera de Ilopango (CI) ubicada dentro del Pull-Apart de San Salvador (PASS) y rodeada por la mayor zona urbana del país. Volcán Panchinmalco: VP, volcán Jayaque: VJ, volcán de Ilopango: VI, volcán San Vicente: VSV, volcán Guasapa: VG, caldera de Plan de Renderos: CPR, caldera Jayaque: CJ, caldera la Carboneras: CCa, Cordillera de Bálsamo: CdB, falla San Vicente: FSV, falla Guaycume: FGy, Pull-Apart rio Lempa: PARL y ERL: embalse Rio Lempa. 22 Actualmente el AVES comprende 21 volcanes activos, tres de los cuales han hecho erupción en el último siglo, el Santa Ana, el San Salvador y el San Miguel (Fig. 7; Siebert and Simkin, 2002). Los productos de estos volcanes, así como los de la caldera de Coatepeque y los materiales más recientes de la CI (TB4, TB3, TB2, TBJ y domos efusivos), abarcan desde el Pleistoceno superior hasta el Holoceno y forman parte de la Formación San Salvador (Weber et al., 1974; Reynolds, 1987; CEL, 1992; Rose et al., 1999; Mann, 2007; Hernández, 2004, 2010; Kutterolf et al., 2008; Lexa et al., 2011). El marco tectónico de El Salvador es una pieza clave para la descripción de los procesos magmáticos y geoquímicos en la región. La mayoría de los volcanes basáltico- andesíticos activos de la zona se concentran dentro de las cuencas pull-apart, mientras que las calderas, de composiciones más silícicas, se localizan mayormente en la traza de las principales fallas laterales derechas del ESFS (Garibaldi et al., 2016), como es el caso de la caldera de Ilopango (Martínez-Díaz et al., 2004; Mann, 2007; Lexa et al., 2011; Saxby et al., 2016). Tal como ocurre en el AVC, en el AVES los volcanes se agrupan en diferentes centros eruptivos compuestos por varios conductos por donde se emplazaron volcanes, domos, conos y calderas. Los productos volcánicos emitidos a lo largo del AVES constituyen una asociación típicamente calcoalcalina de subducción, con composiciones que van desde basálticas a riolíticas (Carr et al., 2007). 1.3. Motivación, hipótesis y objetivos El principal motivo para investigar la CI es que se trata de una estructura vulcano-tectónica activa y peligrosa para la población salvadoreña, de la cual no se habían estudiado con detalle las diferentes fases eruptivas. Prácticamente todos los estudios geológicos previos realizados sobre la caldera de Ilopango, se habían centrado en la última erupción explosiva TBJ. Por ahora existen pocas publicaciones que reporten detalles sobre las erupciones previas (como la TB4, TB3 y TB2), y no hay ningún trabajo vulcanológico que caracterice las ignimbritas relacionadas a las etapas de formación de la caldera. Con esta falta de información se desconocía el origen y la evolución de la caldera, así como la relación entre la actividad volcánica de la caldera y la actividad tectónica de las estructuras regionales de la ZFES, las cuales propongo son el mecanismo disparador de las erupciones de una caldera tipo pull-apart/graben. La hipótesis de este trabajo es que la caldera de Ilopango pudo haber provocado 23 más de una docena de grandes erupciones explosivas por colapso en los últimos 2 Ma, muchas de ellas siendo incluso mayores en magnitud que la última erupción TBJ, a juzgar por los depósitos que afloran en sus alrededores. Es por eso que definir las múltiples fases volcánicas generadas desde su inicio, cuantificar la extensión y el volumen de los productos asociados, fecharlos para poder establecer los tiempos de recurrencia y determinar sus características físico-químicas, aportarían los datos geológicos necesarios para conocer mejor la naturaleza de la caldera de Ilopango y evaluar el riesgo que supondrían futuras erupciones paroxismales similares a la reciente TBJ. Cabe señalar que, a pesar de que la TBJ es la unidad mejor estudiada de la caldera, antes de este estudio no se habían realizado mapas de isopacas e isopletas con suficientes puntos de medición y secciones estratigráficas, para así poder determinar de manera más confiable la distribución espacial, espesores y volumen de los productos piroclásticos asociados. Es por eso que en esta investigación se realizó un trabajo sistemático de la TBJ, para así poder definir su alcance más allá de las fronteras de El Salvador. Sin duda alguna, la caldera de Ilopango es uno de los volcanes más complejos y peligrosos del AVES, siendo la erupción TBJ uno de los eventos más grandes del Holoceno y finales del Pleistoceno en la zona de Centroamérica. Si ocurriera una nueva erupción tipo TB’s hoy en día, afectaría de manera fatal a la sociedad salvadoreña, pudiendo perturbar en gran medida a la población de Centroamérica y México. Objetivos específicos: • Identificar el número de erupciones volcánicas explosivas de la caldera de Ilopango, para completar la historia volcánica junto con las últimas 4 erupciones piroclásticas conocidas como las Tierras Blancas (TB’s < 57 ka). • Caracterizar cada erupción explosiva identificada y sus depósitos asociados, determinar la distribución de los flujos piroclásticos, la composición química y mineralógica, sus edades, los estilos eruptivos y mecanismos de emplazamiento, etc. • Calcular volúmenes de material eyectado de todas las unidades posibles y determinar la dispersión de la pómez /ceniza de caída de la TBJ mediante modelado numérico. • Estimar el periodo de recurrencia y una evaluación preliminar de la peligrosidad volcánica que representa la caldera. 24 • Determinar el origen y la evolución vulcano-tectónica de la caldera. • Publicar los resultados en revistas científicas, así como presentarlos en foros nacionales e internacionales. 1.4. Contenido de la tesis En el Capítulo 1 se presenta una introducción de la tesis y el estado del arte de la caldera de Ilopango, además de los objetivos y motivos de la investigación. El Capítulo 2 describe la metodología utilizada, tanto para la recolección de datos de campo, incluyendo mapeo geológico, levantamiento estratigráfico y muestreo, como en el procesamiento, análisis e interpretación de datos de laboratorio, incluyendo los análisis geoquímicos, geocronológicos, petrográficos, y modelado numérico. En el último apartado de este capítulo se explica cómo se presentaron los resultados de estos análisis. En el Capítulo 3 se presenta el trabajo geológico realizado para caracterizar las tres primeras grandes erupciones explosivas de la caldera de Ilopango. El análisis detallado y desde una perspectiva multidisciplinaria de las extensas ignimbritas formadas por estas tres primeras erupciones de la CI nos permite proponer un modelo conceptual sobre la formación y evolución de esta caldera desde un punto de vista vulcano-tectónico. En este capítulo se presentan datos (geocronológicos, petrológicos, estratigráficos, estructurales, etc.) que soportan las primeras interpretaciones de otros autores en que se sugiere que la actividad de caldera de Ilopango está estrechamente ligada a la actividad tectónica regional ZFES. El primer colapso de la CI fue del tipo graben caldera en un contexto extensional a lo largo del trasarco salvadoreño, y posteriormente, a medida que la ZFES evolucionaba hacia la provincia transtensiva que es hoy en día, fue colapsando como una pull-apart caldera vinculada al sistema de fallas laterales derechas (strike-slip) de la ZFES. Las primeras erupciones fueron las de mayor volumen de toda la historia volcánica de esta caldera poligenética, pudiendo llegar hasta los valores de las supererupciones (~350 km3 DRE, Suñe-Puchol et al., 2019). Químicamente estos depósitos provienen de magmas calcoalcalinos muy evolucionados típicos de zonas de subducción, riolítas con alto contenido en potasio, con periodos de recurrencia largos de hasta 220 ka entre erupciones. 25 En el Capítulo 4 se presenta una revisión completa de la estratigrafía volcánica de la caldera Ilopango. Se propone una nueva nomenclatura de las unidades siguiendo las recomendaciones de Marti et al. (2018). Se define el Grupo Ilopango, subdividido en tres formaciones: 1) la Formación Comalapa (1.785 – 1.34 Ma), que incluye las tres primeras ignimbritas, presentadas en el capítulo anterior; 2) la Formación Altavista (918 – 257 ka), que incluye seis depósitos piroclásticos recién identificados y caracterizados, los cuales se presentan en este capítulo; y 3) la Formación Tierras Blancas (últimos 57 ka), que incluye a las últimas 4 erupciones explosivas, y las únicas que se habían documentado antes de esta tesis. Además, en el Capítulo 4 se presentan los datos detallados sobre las 6 erupciones de la Fm. Altavista, incluyendo datos estratigráficos, químicos, físicos, y geocronológicos, que nos permiten estimar la magnitud de cada erupción, los estilos eruptivos y procesos volcánicos, así como elaborar una historia volcánica detallada de la CI, así como los periodos de recurrencia de las grandes erupciones explosivas. En el Capítulo 5 se presenta un trabajo centrado en la última erupción explosiva de la CI, la unidad conocida como Tierra Blanca Joven (TBJ), la cual cubre la parte central de El Salvador con gruesas capas de ceniza. En este capítulo se presenta los datos estratigráficos y vulcanológicos de los depósitos asociados a la TBJ, utilizados para describir las fases y los procesos eruptivos ocurridos a lo largo de esta gran erupción, así como para estimar el volumen aproximado (> 30 km3 DRE) y modelar la dispersión de las cenizas y la distribución de los flujos piroclásticos. Durante este trabajo se encontraron las cenizas de la TBJ a distancias mayores de 120 km, tanto hacia el sureste (Golfo de Fonseca, frontera con Nicaragua y Honduras), como hacia el noroeste (en las pirámides de Tazumal, en la localidad de Chalchuapa). Para cerrar la tesis, en el capítulo 6 se presenta una discusión integrando los resultados obtenidos y planteando futuros trabajos para, una vez establecida la geología general de caldera, se pueda abordar el tema del peligro volcánico bajo el que vive la población salvadoreña. El Capítulo 7 presenta las conclusiones de este estudio involucrando el trabajo geológico, volcánico y estructural realizado en la CI. 26 2. Metodología Para la realización del estudio geológico, vulcanológico y estructural llevado a cabo durante esta tesis sobre la caldera de Ilopango y sus productos eruptivos, la metodología utilizada se basó en trabajo de campo, análisis de laboratorio y posterior interpretación de resultados. Para ello, y antes de ir campo, se recopiló toda la bibliografía disponible de la zona. Los datos obtenidos, tanto de campo como de laboratorio se integraron en mapas y manuscritos, productos elaborados en una fase de gabinete para posterior publicación. En este capítulo se presentan todas las técnicas usadas y el procedimiento del trabajo subdivididos en 4 apartados donde se abarcan las metodologías en teledetección, campo, laboratorio y gabinete. 2.1. Recopilación bibliográfica y análisis del terreno por teledetección En primer lugar, se realizó un trabajo exhaustivo para recopilar todos los datos geológicos, geoquímicos, geocronológicos, estructurales y de vulcanología en general publicados previamente sobre la CI. Se leyeron los primeros reportes sobre geología hechos en El Salvador como los de Williams y Meyer – Abich (1955) o Weyl (1957), para posteriormente centrarnos en los depósitos del Ilopango como es la TBJ (Hart y Steen – McIntyre, 1983; Vallance  y  Houghton,  1988;  Dull  et  al.,  2010), el resto de las TB’s (Rose et al., 1999; Kutterolf et al., 2008;) y de las ignimbritas más antiguas emitidas por esta caldera (Hernández, 2004; Lexa et al., 2011). Se examinaron otros trabajos más enfocados en la estructura de la caldera de Ilopango, como la gravimetría de Saxby et al. (2016), los modelos analógicos de las strike-slip caldera de Holohan et al. (2008) y el estudio estructural en el Pull-Apart de San Salvador de Garibaldi et al. (2016). Además, se inspeccionó la extensa literatura existente que trata sobre el contexto geodinámico y tectónico de Centroamérica y El Salvador en general (De Mets, 2001; Martínez – Díaz et al., 2004; Corti et al., 2005; Funk et al., 2006; Carr et al., 2007; Alvarado et al., 2011; Alonso-Henar et al., 2015, 2017), para comprender mejor lo que está sucediendo actualmente en la zona de Ilopango y poder plantear hipótesis sobre el origen de esta caldera desde una perspectiva vulcano-tectónica. 27 Una vez familiarizados con la caldera de Ilopango y considerando todos los trabajos existentes hasta la fecha, delimitamos la zona de estudio en base al mapa geológico 1:500,000, realizado por la Misión alemana en los años 70 (Weber et al., 1974). Gracias a ese documento, pudimos acotar preliminarmente la extensión de los productos piroclásticos antiguos de la caldera de Ilopango. Es importante mencionar que en ese trabajo pionero de cartografía que abarcó todo el territorio salvadoreño, no se identificaron el número de erupciones explosivas que tuvo la CI previamente a las 4 TB’s. Los productos piroclásticos más antiguos de la caldera de Ilopango se clasificaron en un mismo miembro, el “c1”, dentro de la Formación Cuscatlán (colores beige en el mapa de la Fig. 9). Esta formación incluye todos los depósitos desde la formación de la CI hasta la erupción TB4, emitida hace ~57 ka por Ilopango y que sirve de marcador para el inicio de la formación San Salvador “S3” (en amarillo, Fig. 9). Fig. 9: Mapa geológico de la zona de estudio sobre el Modelo de Elevación Digital de 10 m (modificado de Weber et al., 1974). El mapa geológico de Weber et al., (1974), originalmente analógico, se rasterizó y se georeferenció en un Sistema de Información Geográfica libre como es el QGis v. Las Palmas. Posteriormente se sobrepuso al Modelo de Elevación Digital terrestre de El Salvador, de 10 metros de precisión (DEM, Fig. 9), para así poder observar 28 conjuntamente la geomorfología de la zona y las formaciones litoestratigráficas. De esta manera elaboramos un primer mapa base donde poder organizar las salidas de campo y clasificar los datos ordenadamente. Concretamente el área de estudio delimitada ocupa unos 3,000 km3 de territorio, localizada en la parte central de El Salvador. Por otro lado, utilizamos imágenes de satélite para determinar afloramientos potencialmente útiles, así como para identificar las principales vías de acceso y comunicación. Esto se hizo antes y durante el trabajo de campo, por medio de una Tablet con GPS para estar localizados en todo momento y hacer más efectivo el tiempo en el campo. 2.2. Trabajo de campo: levantamiento estratigráfico, mapeo y muestreo Una vez establecida la zona de estudio y la posible distribución de los productos piroclásticos de la CI (Fig. 9), se procedió a organizar el trabajo de campo. El objetivo principal de esta fase del trabajo era realizar el mapeo geológico de la zona de estudio al completo y el levantamiento estratigráfico de todos los productos de interés de la caldera, para que una vez determinado el número de depósitos/erupciones y su distribución espacial mediante correlación, se pudiese proceder al muestreo ordenado de todas las unidades reconocidas y su posterior caracterización en laboratorio (geocronología, geoquímica, petrografía, etc.). Para ello se programaron hasta 8 campañas de trabajo de campo, que duraron entre 2 a 5 semanas cada una, y se extendieron desde 2015 a 2017. Aún con las dificultades que representa hacer campo en una zona tropical y muy cubierta por vegetación como es El Salvador, encontramos los suficientes afloramientos útiles para completar los propósitos de esta investigación. En total se realizaron alrededor de 350 puntos de observación, de los cuales se levantaron unas 200 secciones estratigráficas y se muestrearon todas las unidades litoestratigráficas necesarias para completar el estudio: desde las 13 ignimbritas emitidas por el Ilopango, hasta algunas de sus lavas y otras unidades procedentes de otros centros eruptivos (como el Volcán San Salvador, la caldera Carboneras o otras lavas viejas de la cordillera del Bálsamo). Cabe destacar que para realizar el trabajo de campo fue necesaria la colaboración logística por parte del Ministerio de Medio Ambiente y Recursos Naturales – MARN de El Salvador (vehículos y gente conocedora de la zona) y de la Policía Nacional Civil - PNC, quien nos aportó seguridad a la hora de trabajar en lugares conflictivos. Parte de estas campañas de campo se dedicaron también para trabajar otros aspectos del proyecto de investigación dirigido por el Dr. Gerardo Aguirre, titulado “Peligrosidad para México de super-erupciones originadas en Centroamérica: El caso de 29 la caldera de Ilopango, El Salvador, y su influencia en el declive del Imperio Maya”. Por ejemplo, conjuntamente con el Dr. Dario Pedrazzi y otros miembros del equipo, se realizó un trabajo sistemático de vulcanología física sobre la TBJ, con el objetivo de describir a detalle los procesos volcánicos de ésta última erupción explosiva ocurrida hace unos 1500 años (ver capítulo 5 para más detalles). Además, se visitaron excavaciones en sitios arqueológicos como las pirámides de Tazumal o la de San Andrés (Fig. 10a), donde se pudo observar directamente la relación entre las cenizas de la TBJ y estas construcciones Mayas. Fig. 10: a) Excavación al pie de la pirámide principal del sitio arqueológico de San Andrés, y b) muestreo de fluidos hidrotermales en el fondo del lago Ilopango (foto de David Alfaro). 30 En la última campaña de campo, y gracias a la disponibilidad de José Bairés y David Alfaro (buzos profesionales y expertos conocedores del Lago de Ilopango), nos sumergirnos a muestrear fluidos hidrotermales que emanan de forma continuada desde el fondo del lago de Ilopango (Fig. 10d). Todos estos datos de campo se recopilaron en varios formatos digitales: desde reportes de campo en documentos Word, tablas Excel, soporte en el Google Earth, etc. Las coordenadas geográficas se grabaron con un GPS portátil tipo Garmin, en sistema de proyección UTM (Datum: D_WGS_1984, zona 16P). 2.3. Envío de muestras y análisis en laboratorio Las muestras recolectadas en campo se organizaron y clasificaron en las instalaciones del Observatorio Ambiental de El Salvador. Des de allí se enviaron las muestras más importantes y urgentes por paquetería hasta Querétaro (UPS y Correos de El Salvador), y así poder empezar los primeros análisis en los laboratorios del Centro de Geociencias (CGEO). El resto de muestras se fueron acumulando en un almacén del MARN hasta que nosotros mismos las fuimos a buscar por vía terrestre. Con una camioneta de la UNAM-CGEO, atravesamos las fronteras de Guatemala hasta llegar a San Salvador, donde cargamos unas 30 cajas de muestras de rocas y cenizas para llevárnoslas de regreso con nosotros hasta el campus Juriquilla. Una vez en Querétaro, la mayoría de muestras fueron analizadas en los laboratorios del CGEO (geocronología por circones, geoquímica de roca total, laminación para observación petrográfica, etc.), pero otras muestras fueron enviadas por DHL al campus de la Oregon State University (OSU) de Corvallis (EUA), en donde se analizaron para conseguir fechamientos por el método de 39Ar/40Ar. Otras muestras de carbón se enviaron al laboratorio comercial Beta Analitic de Texas (EUA), para el fechamiento por el método del 14C. También se enviaron muestras de vidrio de la TBJ a la Universidad de Oxford (UK) para realizar análisis por microsonda electrónica. A continuación se explican todos los métodos del trabajo de laboratorio utilizados en este estudio. Geocronología:  U/Pb,   238 U/ 230 Th  y  Ar 39 /Ar 40     Un aporte destacable de esta tesis de doctorado es el estudio geocronológico realizado sobre los productos piroclásticos de la CI. Gracias a los fechamientos llevados a cabo usando métodos radiométricos como son el U/Pb, 238U/230Th y Ar39/Ar40, se ha podido descifrar, por primera vez, una historia volcánica más completa de la que se 31 conocía para esta caldera activa. En cuanto a la técnica del U/Pb y 238U/230Th, el mineral utilizado para medir esta relación isotópica es el Circón (ZrSiO4). Los circones son cristales muy resistentes a la intemperización, y por ello conforman un sistema cristalino prácticamente cerrado desde su formación. Es por eso que resultan muy útiles para medir la edad de las rocas que los contienen por medio del conteo isotópico U/Pb, o 238U/230Th para rocas más jóvenes de 350 ka. En este caso, los circones se separaron a partir de fragmentos juveniles de pómez de cada ignimbrita o depósito de caída asociado a las diferentes erupciones de la caldera. Este trabajo se realizó en los laboratorios de Molienda y Separación de Minerales del CGEO, en donde se removieron los líticos pegados a los clastos de pómez para evitar fechar circones heredados. Luego se trituraron las pómez y se tamizaron hasta la fracción 74-44 µm (las dimensiones de la mayoría de circones de origen volcánico se encuentran en ese rango). Esa fracción pulverizada se lavó y se bateó para concentrar los minerales pesados (Fig. 11a). Después de secar ese concentrado a 70º C en el horno, se pasó la muestra por el Electroimán Frantz para separar los minerales magnéticos (Fig. 11b). La porción de minerales no magnéticos, reducida notablemente después de todo el proceso de separado, se introduce en una placa de Petri y se pican los circones uno a uno con la ayuda de una lupa binocular. Una vez montados los suficientes circones en una probeta con resina endurecida (unos 50 para rocas volcánicas), se comprobó su pureza por catodoluminiscencia en el ELM-3R (Marshall, 1988) y luego fueron analizados por el técnico del Laboratorio de Estudios Isotópicos-LEI del CGEO, el Dr. Carlos Ortega Obregón, quien me supervisó y asesoró en todo el proceso de separación de circones. Para los fechamientos U/Pb se utilizó el espectrómetro Thermo ICAP Qc (LA-ICP-MS), con una resolución de 193 nm (Solari et al., 2010), y para los análisis U-Th el Thermo Neptune plus (Bernal et al. (2014). Una vez analizados los circones, se redujeron los datos usando el Isoplot software (v. 3. 7) para los análisis U-Pb (Ludwig, 2008) y el IntCal09 software para calibrar las edades U- Th (Reimer et al., 2009). Ver capítulos 3 y 4 para más detalles y ejemplos de edades obtenidas en circones. 32 Fig. 11: a) Batea de plástico en espiral para lavar y concentrar mineral pesado, y b) Electroimán Frantz para separar minerales magnéticos. Con la intención de complementar el estudio geocronológico y reforzar las edades obtenidas por el método U-Th-Pb, se realizó una estancia académica de 5 meses en el Laboratorio de Geocronología de la Oregon State University - OSU (Corvallis, EUA) para obtener edades 39Ar/40Ar de toda la secuencia piroclástica de la CI. En esta técnica, al igual que en el caso de los circones, se trituraron clastos de pómez juvenil para poder separar cristales apropiados en este tipo de fechamientos. Aunque los sanidinos son la fase mineral preferencial para este tipo de análisis en rocas jóvenes debido a su alto contenido en K y Ar, lamentablemente no se encontraron este tipo de cristales en ninguna unidad del Ilopango, así que básicamente se usaron plagioclasas y hornblendas. La pómez triturada se tamizó y se lavó la fracción de 500-177 µm. Usando un ultrasonido se removió todo el polvo pegado al concentrado mineral. Los feldespatos fueron separados magnéticamente de los anfíboles y piroxenos usando un Frantz (como el de la Fig. 11), y luego fueron bañados en ácidos (HF, HN03 y HCL) para eliminar restos de vidrio pegados a los cristales (Koppers et al., 2011). Una vez secos se seleccionaron los mejores cristales (por tamaño, forma y pureza) y se empaquetaron para ser irradiados en el reactor nuclear TRIGA CLICIT de la Oregon State University conjuntamente con los sanidinos Fish Canyon Tuff de edad conocida (28.201 ± 0.023 Ma, 1σ; Kuiper et al. 2008), que sirvieron como estándares. Los concentrados minerales fueron analizados con la técnica “Incremental Heating” debido a que en cristales individuales de plagioclasa o de hornblenda no hay suficiente argón radiogénico para obtener resultados confiables (Rose et al., 1999). Los analisis se efectuaron con el espectrómetro de masas ARGUS-VI (Fig. 12), que con su láser de CO2 y su multicolector de gases es capaz de medir los cinco isotopos de Ar simultáneamente (el 33 36, 37, 38, 39 y 40). Para conseguir la mayor precisión posible en cada fechamiento, se efectuaron un gran numero de pasos o “heating steps” (hasta 22-23 por muestra), en donde se median cada vez los 5 gases del argón subliminados durante el calentamiento del concentrado mineral por láser. Las edades se obtuvieron al reducir los datos con el software ArArCALC v2.5.1 (Koppers, 2002). Todo el proceso fue supervisado por el Dr. Dan Miggins, gerente del laboratorio del OSU. Para más detalles de esta técnica y ejemplos de edades 39Ar/40Ar, ver capítulos 3 y 4 de esta tesis. Fig. 12: Espectrómetro ARGUS VI con laser de CO2 y línea de extracción con multicolector (isótopos 36, 37, 38, 39 y 40 de Ar). Geoquímica  y  Petrografía     Con el propósito de caracterizar químicamente y soportar la correlación estratigráfica de los productos piroclásticos de la CI, se llevaron a cabo análisis de roca total utilizando clastos de pómez juvenil, previamente secada en el horno a 80º C y limpiada cuidadosamente a mano para evitar contaminación. Las muestras se pulverizaron manualmente en un mortero de Ágata hasta homogenizar todo a 74 µm, usando una malla plástica del 200 nueva para cada muestra. Los elementos mayores fueron medidos por la Dr. Patricia Girón mediante el método de Fluorescencia de Rayos X en el Departamento de Geoquímica del Instituto de Geología de la UNAM (Ciudad de México), usando el espectrómetro X RIGAKU ZSX Primus II. Las Tierras Raras 34 (REE) y los elementos traza los midió la M. en C. Ofelia Pérez Arvizu en el Laboratorio de Estudios Isotópicos y utilizando el cuarto ultralimpio del CGEO. Los métodos y manejo de las muestras se describen en Bernal y Lozano-Santacruz (2005). Todos estos datos se ilustran en graficas tipo TAS (Total-Alkali-Silica, LeBas et al., 1986), Spider- multielements (tierras raras normalizadas al MORB, Sun and McDonough, 1989) y Harker (donde se plotea la concentración de varios elementos químicos en frente del SiO2, como es el K; Pecerillo and Taylor, 1976). Para más detalles sobre estos análisis ver capítulos 3 y 4. En cuanto a los análisis químicos en la TBJ, además de hacer los de roca total, se efectuaron otros estudios para conocer la composición del vidrio y así poder cuantificar la abundancia en elementos como el S-2, el F- y el Cl- emitidos durante esa erupción y así evaluar su impacto ambiental. Estos nuevos análisis los realizó la Dra. Victoria Smith en el laboratorio de Arqueología e Historia del Arte de la Universidad de Oxford con el método de microsonda EPMA (wavelength-dispersive electron probe microanalysis). La sonda de electrones utilizada fue calibrada para cada elemento utilizando un mineral bien conocido como estándar, y se verificó este calibrado analizando el vidrio de referencia MPI-DING (Jochum et al., 2006). Para más detalles de este método ver capítulo 5. En el taller de laminación del CGEO y bajo la supervisión del técnico Juan Vázquez, se prepararon láminas delgadas para observación petrográfica de los productos de la CI, y así poder caracterizar las texturas y la mineralogía de los depósitos volcánicos del Ilopango. Con una cámara incorporada a un microscopio del Laboratorio de Petrografía del CGEO, y con la colaboración del Dr. Alexander Iriondo, se sacaron microfotografías de todas las unidades de interés. Otros  análisis  de  laboratorio:  Granulometría,  14C  y  dendrocronología     Como parte del estudio de vulcanología física sobre la TBJ, se realizaron análisis granulométricos en las muestras de pómez y ceniza recolectadas en campo con Dario Pedrazzi. Parte de estos análisis se efectuaron en los laboratorios del Observatorio Ambiental del MARN-El Salvador, donde se tamizaron en seco más de 141 muestras proximales, medias y distales, para así obtener la distribución del tamaño de grano y el porcentaje de componentes (rango de apertura de tamices desde −6 Φ a 3 Φ, donde Φ = ‒ log2 ; d es el diámetro en mm). En el Laboratorio de Vulcanología Física del CGEO 35 se realizó la otra parte del análisis granulométrico. Con la ayuda de un MicroTec Analisette22 Fritsch, se hizo el tamizado húmedo, separando las fracciones de ceniza más fina (4 Φ a >10 Φ). El porcentaje de peso de cada fracción tamizada se calculó y se ploteó en curvas acumulativas (ver detalles en capitulo 5). Se hizo el conteo de los juveniles entre -5 a 0 Φ con la ayuda de una microscopio binocular, se fotografiaron y se identificaron los diferentes componentes. De esta manera, gracias al mapeo, al levantamiento estratigráfico, a estos resultados granulométricos y a los modelos numéricos posteriores que se presentan en la siguiente sección y en el capítulo 5, se pudieron determinar parámetros físicos de la erupción de la TBJ como son la altura de la columna eruptiva, la tasa de emisión, duración y dispersión de cenizas. En el Laboratorio de Vulcanología Física del CGEO también se prepararon muestras de carbono que se encontraron incluidos en el depósito de la TBJ. Éstas se enviaron al laboratorio comercial de Beta Analitic (calidad ISO 17025) en Texas (EUA), para el fechamiento de esa erupción mediante el método de 14C y AMS (Accelerator Mass Spectrometry). 2.4. Trabajo de gabinete: elaboración de mapas, digitalización de series estratigráficas, modelado numérico y estimación de parámetros físicos En una etapa posterior al campo y los análisis de laboratorio, e integrando todos los datos obtenidos previamente, se generaron productos relevantes que se aportan en esta investigación y que forman parte de publicaciones en congresos y revistas indexadas. Estos son, por ejemplo, nuevos mapas georeferenciados en Qgis v. Las Palmas, donde se despliega la distribución espacial de todas las unidades piroclásticas identificadas de la CI, o series estratigráficas digitalizadas con el programa Adobe Ilustrator CS6, que además incluyen datos geocronológicos, estructurales y granulométricos (ver ejemplos en los capítulos 3, 4 y 5). En esta etapa de gabinete también se calcularon los volúmenes de los flujos piroclásticos del Ilopango por medio del método de Triangulación Deleuriana de Macedonio and Pareschi (1991), el cual utiliza el área ocupada por las ignimbritas, el espesor de la capa medida en diferentes puntos bien distribuidos espacialmente y el nivel 0 en los bordes del flujo, para así estimar un volumen de piroclastos por interpolación (ver capítulo 3). Estos volúmenes, los cuales se obtuvieron con la ayuda del Dr. Antonio Costa (miembro del proyecto Ilopango), después fueron transformados a Dense Rock Equivalent (DRE) siguiendo la metodología de Quane and Russell (2005). Para estimar el volumen de las caídas de 36 tefra de la erupción TBJ, el Dr. Dario Pedrazzi (miembro del proyecto Ilopango) reconstruyó los mapas de isopacas a partir de los datos recolectados en campo y los resultados de análisis granulométricos. Posteriormente se utilizaron métodos basados en adelgazamiento logarítmico de los depósitos (Bonadonna y Costa, 2012; 2013). Para modelar la dispersión de las cenizas de esta erupción TBJ el Dr. Antonio Costa utilizó el software HAZMAP (Macedonio et al., 2005). A partir de la inversa del volumen se pudo estimar la altura de la columna (Costa et al., 2009) y la tasa eruptiva (Mastin et al., 2009). Ver capitulo 5 para más detalles. 37 3. Descripción de las primeras erupciones formadoras de extensas ignimbritas por la caldera de Ilopango, una estructura vulcano-tectónica tipo graben/pull-apart en El Salvador. Artículo: Suñe-Puchol, I., Aguirre-Díaz, G.J., Dávila-Harris, P., Miggins, D.P., Pedrazzi, D., Costa, A., Ortega-Obregón, C., Lacan, P., Hernández, W., Gutiérrez, E., 2019. The Ilopango caldera complex , El Salvador    : Origin and early ignimbrite-forming eruptions of a graben / pull-apart caldera structure. J. Volcanol. Geotherm. Res. 371, 1– 19. doi:10.1016/j.jvolgeores.2018.12.004 Contribuciones individuales de los autores: ØØ Ivan Suñé Puchol: diseño y organización del estudio, trabajo de campo y de laboratorio, procesamiento, análisis e interpretación de datos, redacción del artículo. ØØ Gerardo Aguirre Díaz: financiamiento, concepción y plan de trabajo, supervisión, trabajo de campo, interpretación de los datos, revisión del artículo. ØØ Pablo Dávila Harris: diseño y supervisión del estudio, trabajo de campo, interpretación de datos y corrección del artículo. ØØ Dan Miggins: supervisión en laboratorio y en procesamiento e interpretación de fechamientos Ar/Ar, corrección del artículo. ØØ Dario Pedrazzi: trabajo de campo, interpretación de datos y corrección del artículo. ØØ Antonio Costa: trabajo de campo y cálculo de volumen de la ignimbritas mediante modelos numéricos. ØØ Carlos Ortega Obregón: supervisión en laboratorio y fechamientos U-Pb, procesamiento e interpretación de datos geocronológicos. ØØ Pierre Lacan: trabajo de campo e interpretación de datos vulcano-tectónicos. ØØ Walter Hernández: trabajo de campo e interpretación de la geología de la zona. ØØ Eduardo Gutiérrez: apoyo logístico y trabajo de campo. 38 Journal of Volcanology and Geothermal Research 371 (2019) 1–19 The Ilopango caldera complex, El Salvador: Origin and early ignimbrite-forming eruptions of a graben/pull-apart caldera structure Ivan Suñe-Puchol a,, Gerardo J. Aguirre-Díaz a, Pablo Dávila-Harris b, Daniel P. Miggins c, Dario Pedrazzi d, Antonio Costa e, Carlos Ortega-Obregón a, Pierre Lacan a, Walter Hernández f, Eduardo Gutiérrez f a Centro  de  Geociencias,  Universidad  Nacional  Autónoma  de  México,  Blvd.  Juriquilla  3001,  Campus  UNAM-­‐Juriquilla,  Querétaro,  76230,  Mexico   b División  de  Geociencias  Aplicadas,  IPICYT,  San  Luis  Potosí  78216,  Mexico   c College  of  Earth,  Ocean  and  Atmospheric  Sciences,  Oregon  State  University,  104  CEOAS  Administration  Building,  101  SW  26th  St,  Corvallis,  OR  97331,  USA   d ICTJA,  CSIC,  Group  of  Volcanology,  SIMGEO  UB-­‐CSIC,  Institute  of  Earth  Sciences  Jaume  Almera,  Lluis  Sole  i  Sabaris  s/n,  08028  Barcelona,  Spain   e Istituto  Nazionale  di  Geofisica  e  Vulcanologia,  INGV-­‐Bologna,  Via  Donato  Creti,  12,  40100  Bologna,  Italy   f Gerencia  de  Geología  del  Observatorio  Ambiental,  Ministerio  de  Medio  Ambiente  y  Recursos  Naturales  MARN,  San  Salvador  76230,  El  Salvador       a r t i c l e i n f o Article  history:   Received 3 August 2018 Received in revised form 4 December 2018 Accepted 6 December 2018 Available online 17 December 2018 Keywords:   Central America Volcanic Arc Tectono-volcanism El Salvador Fault Zone Fissure eruption Hydromagmatism a b s t r a c t The Ilopango caldera is located in the central part of El Salvador, within the right-lateral El Salvador Fault System (ESFZ) and adjacent to the capital city of San Salvador. The caldera has a polygonal shape of 17 × 13 km and hosts an intra-caldera lake. Ilopango caldera had multiple collapse eruptions that formed widespread and voluminous silicic ignimbrites. Volcanic activity of the caldera has been controlled by strike-slip faults of the ESFZ. In this work we present the geological characteristics of the first three ignimbrite-forming eruptions of Ilopango caldera, pro- viding an interpretation of the origin and initial stages of the volcanic evolution of this caldera complex. An initial extensional regime of the ESFZ possibly developed a graben at or near the actual Ilopango caldera, where the graben's master faults worked as fissure vents during the first caldera collapse. The Olocuilta Ignimbrite was emplaced at 1.785 ± 0.01 Ma BP, with a Dense Rock Equivalent (DRE) volume N 50 km3 (probably ~300 km3). The ESFZ stress gradually changed from extensive to transtensive, inducing the second collapse associated with a pull-apart caldera, producing the Colima Ignimbrite at 1.56 ± 0.01 Ma BP, with a DRE volume of N11 km3. The transtensive regime increased along the ESFZ, producing the third collapse in the pull-apart graben caldera apparently affected by the newly formed strike-slip San Vicente Fault. This phase corresponds to the ex- plosive eruption that formed the Apopa Ignimbrite at ~1.34 Ma BP, with N9 km3 DRE volume. The latter ignim- brite marks a change in the eruptive style producing hydromagmatic pyroclastic flows followed by a dense ignimbrite with coignimbrite lithic breccias. These features suggest the involvement of water that could come from a paleoIlopango lake within the caldera depression associated with the second caldera collapse at 1.56 Ma BP. Ilopango is thus a multistage caldera system associated with the largest explosive events registered in El Salvador so far. © 2018 Elsevier B.V. All rights reserved. 1. Introduction Collapse calderas are formed by the subsidence of crustal blocks along bounding faults into a shallow chamber during the fast evacuation of magma (Smith and Bailey, 1968; Druitt and Sparks, 1984; Lipman, 1997, 2000; Gottsmann and Martí, 2008). Caldera eruptions can be ex- plosive or effusive; but, the most catastrophic and voluminous are the explosive ones that can erupt massively large volumes of pyroclastic material within hours to few days (e.g., Newhall and Dzurisin, 1988; Lipman, 1997; Cole et al., 2005; Gottsmann and Martí, 2008; Costa et al., 2014; Costa and Marti, 2016). Silicic collapse calderas are Corresponding author. E-­‐mail  address:  ivansp@geociencias.unam.mx (I. Suñe-Puchol). associated with infrequent but catastrophic explosive supereruptions and are considered a major geohazard (e.g., Toba at ca. 74 ka, Francis et al., 1983; Rampino and Self, 1993; Self, 2006). In collapse calderas, the subsidence of the chamber roof is produced by syn-collapse faults formed during the collapse phase (e.g., Sparks et al., 1985; Marti et al., 1994; Smith and Braile, 1994; Bacon and Lanphere, 2006; Acocella, 2007). In some cases, collapse calderas used pre-existing tectonic faults as cortical discontinuities to collapse, such as calderas associated with the Sierra Madre Occidental in Mexico (Aguirre-Díaz and Labarthe-Hernández, 2003; Aguirre-Díaz et al., 2008; Gottsmann et al., 2009; Sunye-Puchol et al., 2017), the Cañas Dulces caldera in Costa Rica (Molina et al., 2014), as well as several cal- deras within the Ethiopian Rift (Acocella et al., 2002; Robertson et al., 2015). The Ilopango caldera (IC) apparently corresponds to this last https://doi.org/10.1016/j.jvolgeores.2018.12.004 0377-0273/© 2018 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Journal of Volcanology and Geothermal Research jour nal homepage: www.elsevier. com/ locate/ jvolgeores 39 2 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19     type, which is known as pull-apart/graben calderas (Aguirre-Díaz, 2008; Aguirre-Díaz et al., 2008) and related to extensional and/or trans-tensional tectonic regimes. Williams and Meyer-Abich (1955) first described IC as a volcano-tectonic depression with several erup- tions. Recently, Saxby et al. (2016) interpreted Ilopango as a strike-slip caldera, in the same sense as Aguirre-Díaz and Martí (2015), where the vertical faults of the El Salvador Fault Zone (ESFZ) worked as a pref- erential pathway for magma ascension to the surface (Tikoff and de Saint Blanquat, 1997). Other examples with reports of association be- tween large calderas and pull-apart basins are from Las Sierras- Masaya volcanic complex in Nicaragua (Girard and van Wyk de Vries, 2005; Holohan et al., 2008) and the Toba caldera in Indonesia, within the Great Sumatran Fault Zone (Bellier and Sébrier, 1994). The aim of this work is to study the origin of IC and the first ignimbrite forming eruptions and clarify the relationships between volcanism and the re- gional transtensive tectonics. IC is located in the central part of El Salvador, next to the capital city of San Salvador (Fig. 1, index map) and it has a rectangular- rhombohedral shape with 17 × 13 km. The caldera hosts the Ilopango Lake, which is 231 m deep, has an area of 70.5 km2 and contains ~12 km3 of water (Sánchez-Esquivel, 2016). The last volcanic event of the caldera is marked by the Islas Quemadas eruption at 1879–1880 (Fig. 2) that formed an intra-lake lava dome (Golombek and Carr, 1978; Richer et al., 2004). The youngest ignimbrite eruption formed by IC has been the historic Tierra Blanca Joven (TBJ) eruption (Williams and Meyer-Abich, 1955; Rose et al., 1999; Hernández, 2004; Kutterolf et al., 2008; Dull et al., 2001; Saxby et al., 2016; Aguirre-Díaz et al., 2017; Pedrazzi et al., 2018), approximately 1500 years ago and that apparently devastated the Mayan civilization in the region (Dull et al., 2010). Prior to TBJ eruption, there were 3 large explosive erup- tions during the last 57 ka: TB4, TB3 and TB2 (Rose et al., 1999; Hernández, 2004; Hernández et al., 2010; Kutterolf et al., 2008). In this work, we show that IC had multiple large-volume ignimbrite eruptions associated with collapse episodes (similar to the Platoro com- plex in San Juan Mountains; Lipman et al., 1996), and all of them were related to pre-existing regional structure faults associated with the ESFZ. This paper focuses on the ignimbrites of the three earliest volcanic eruptions related to IC, which have not been previously described and interpreted. These are the Olocuilta Ignimbrite (OI), Colima Ignimbrite (CoI), and Apopa Ignimbrite (ApI). For these three newly mapped ig- nimbrite sheets, we present their stratigraphic and chemical character- istics as well as a robust geochronological framework, and their link with the coeval tectonics of the region. 2. Tectonic and geologic setting of Ilopango caldera (IC) 2.1. Central  America  tectonic   framework     IC is the largest active volcano in El Salvador, a Central American country that lies on the Pacific shore of the Caribbean plate (Fig. 1). About 200 km to the north of San Salvador City is the boundary with the North American plate, represented by the left strike-slip system of Polochic-Motagua-Swan Island Fault (PMSIF), which is the source of the eastward displacement of the Caribbean plate at an average speed Fig. 1. Index map of main volcanoes and faults of El Salvador indicating location of Ilopango caldera and the study area shown in Fig. 2 (inset map). ESFZ: El Salvador Fault Zone, ESVF: El Salvador Volcanic Front, SS: San Salvador City, SAV: Santa Ana Volcano, CC: Coatepeque Caldera, SSV: San Salvador volcano, IC: Ilopango caldera, SVV: San Vicente Volcano, BVC: Berlin Volcano Complex, and SMV: San Miguel Volcano. Red arrows indicate the actual kinematic within pull-apart basins along ESFZ (after Staller et al., 2016). The inset map shows the conti- nental-scale location of El Salvador in Central America and its regional tectonic setting indicating the main fault systems along the Central American Volcanic Arc - CAVA (after Funk et al., 2009). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 40 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19   3 Fig. 2. Index map of the area covered in this study at central El Salvador, showing the distribution of the first three ignimbrites of Ilopango caldera (IC), Olocuilta Ignimbrite (OI, red), Colima Ignimbrite (CoI, blue) and Apopa Ignimbrite (ApI, green). Other features mentioned in main text are also indicated, including sampling location sites (blue stars, sample number with pre- fix ILO-), Las Pavas Lava (LPL, purple), main tectonic structures and other volcanoes and calderas. IVC: Ilopango Volcanic Complex, JC: Jayaque Caldera, JV: Jayaque Volcano, PV: Panchinmalco Volcano, PRC: Planes de Renderos Caldera, SJD: San Jacinto Dome, SSV: San Salvador volcano, SVV: San Vicente Volcano, CaC: Carbonera caldera, GV: Guasapa Volcano, SSPA: San Salvador Pull-Apart, BaMR: Balsamo Mountain Range, BoMR: Border Mountain Range, LRPA: Lempa River Pull-Apart, SVF: San Vicente Fault, GYF: Guaycume Fault; LRWR: Lempa River Water Reservoir. In yellow is the San Salvador Metropolitan Area (SSMA, ~3 million population). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) of 8 mm/year (Fig. 1; DeMets, 2001; Guzmán-Speziale et al., 2005; Agostini et al., 2006; Funk et al., 2009). Ilopango is a caldera of the Cen- tral American Volcanic Arc (CAVA), which extends from Guatemala to Costa Rica (e.g., Simkin and Siebert, 1994; Carr et al., 2007; Sitemap et al., 2014). The CAVA magmatism is related to the oblique subduction of Cocos under Caribbean plate, developed along the Mesoamerican trench since the early Tertiary until the present day, at an average speed of 73–85 mm/year (Fig. 1; Dixon, 1993; Mann, 2007; DeMets, 2001). Oblique subduction (LaFemina et al., 2009), together with a zip- per effect produced at the triple junction in Guatemala (Authemayou et al., 2011), and the decoupled interface initiated by the rollback of the Cocos slab (Alonso-Henar et al., 2015, 2017), might be the causes of the forearc sliver displacement of the El Salvador and Nicaragua to the NW (8–14 mm/year; Fig. 1; Corti et al., 2005; Turner et al., 2007; Alvarado et al., 2011). The eastward motion of Caribbean plate and the northwestward dis- placement of the Central American forearc are responsible for the transtensional deformation along the CAVA in a series of en-echelon right-stepping, dextral faults connected by pull-apart basins along fault step-overs (Montero and Dewey, 1982; DeMets, 2001, LaFemina et al., 2009; Corti et al., 2005; Agostini et al., 2006, Turner et al., 2007; Funk et al., 2009; Canora et al., 2014), known as El Salvador Fault Zone (ESFZ; Martínez-Díaz et al., 2004). The ESFZ crosses all of El Salvador with a WNW-ESE trend (Figs. 1 and 2), and results in the opening of the San Salvador Pull-Apart (Garibaldi et al., 2016), where IC is emplaced (Figs. 1 and 2). Carr (1976) mentions a pure extensional re- gime previous to the actual transtensional regime along the ESFZ, in- duced by the rollback of the Cocos slab started during the Pliocene (Canora et al., 2014; Alonso-Henar et al., 2015, 2017). Weinberg (1992) proposed a similar geodynamic evolution for the Nicaragua gra- ben, where the present-day deformation is controlled by a right-lateral transtensional regime, localized along the Nicaragua volcanic front and superimposed to a previous rift. The last fault-slip offset of ESFZ occurred during a Mw 6.7 earth- quake on February 13 of 2001, indicating a dominant transcurrent stress field along the San Vicente Fault (Fig. 2; Martínez-Díaz et al., 2004; Alvarenga et al., 2001; Canora et al., 2012). This structure, which is af- fecting the IC volcanism (Saxby et al., 2016), has totally displaced an 41 4 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19     older caldera located 12 km to the east: the Carbonera caldera (Fig. 2; Canora et al., 2014). Since the formation of Carbonera caldera about 2.1 Ma ago (Rotolo and Castorina, 1998), strike-slip tectonics along the San Vicente Fault has been active with dextral sense and an average slip rate of ~5 ± 0.5 mm/yr, (Canora et al., 2014). This regional tectonic framework including strike-slip faults, shear zones, pull-apart basins, releasing bends and step-overs, favored the initial pathways for ascen- sion of magma and the emplacement of magmatic centers at shallow levels of the crust (Hutton and Reavy, 1992). 2.2. Geologic  setting  of  Ilopango   caldera     El Salvador is predominantly made up of volcanic rocks with ages ranging from the Eocene to the present (Donnelly et al., 1990; Rose et al., 1999). The basement is mainly composed of Jurassic-Cretaceous limestone, represented by the Metapán Formation (Weber et al., 1974), which is sporadically exposed in the NW part of the country. During the Oligocene, volcanism formed the Border Mountain Range at the northernmost part of the country and corresponds to the Morazán and Chalatenango formations. This range marks the geograph- ical border between El Salvador and Honduras (Fig. 2). Since the Mio- cene, volcanic activity migrated southward and approached the Mesoamerican trench. Volcanism from mid-Miocene to Pliocene formed large basaltic and andesitic stratovolcanoes just south of the ac- tual El Salvador Volcanic Front, such as Panchinmalco Volcano, Jayaque Volcano and the oldest sequence of the Ilopango Volcanic Complex, rep- resented by the Balsamo Formation (Lexa et al., 2011; Fig. 2). The cores of these eroded volcanoes form the Balsamo Mountain Range (Figs. 1 and 2; Williams and Meyer-Abich, 1955). During the Pliocene- Pleistocene, several volcanic edifices collapsed gravitationally appar- ently associated with the ESFZ structures, forming the calderas of Plan de Renderos and Jayaque both belonging to the Balsamo formation (Lexa et al., 2011), and the Carbonera caldera and early volcanic rocks of the IC that form part of the Cuscatlán Formation (Fig. 2; Weber et al., 1974). The currently active volcanic belt forms the El Salvador Volcanic Front, which is part of the CAVA (Fig. 1; Carr et al., 2007). Products asso- ciated with these volcanic centers from late Pleistocene to Holocene comprise the San Salvador Formation (Weber et al., 1974; Reynolds, 1987; Lexa et al., 2011). The Coatepeque Caldera deposits and late Ilopango Tierras Blancas (TB4, TB3, TB2 and TBJ) are part of this last for- mation (CEL, 1992; Rose et al., 1999; Hernández, 2004; Hernández et al., 2010; Kutterolf et al., 2008). 3. Methodology 3.1. Field  work     Geologic work was based initially on the El Salvador 1: 500,000 geo- logical map of Weber et al. (1974), as well as on previous maps and stra- tigraphy reported by other authors (Hernández, 2008; Lexa et al., 2011; Garibaldi et al., 2016). Through this data and several fieldwork cam- paigns between 2015 and 2017, a new stratigraphic framework and a new geologic map for the Ilopango caldera complex and its surround- ings were built. This new map covers around 3000 km2 of the central part of El Salvador, from Border Mountain Range at north, to the Pacific coast at the south, and from the Lempa River Pull-Apart basin at the east, to Jayaque Caldera at the west (Fig. 2). Prior to, and during fieldwork, outcrops were identified via satellite images to determine the available access. Once identified and reached, exposures were described geologically, measured for stratigraphic sec- tions, and samples were collected. Despite the difficulties of fieldwork in the El Salvador tropical area, we found the necessary outcrops to ful- fill the goals of this study. A Digital Elevation Model (DEM) with 10 m precision was used as a topographic base map. Nearly 85 stratigraphic sections were measured, and several samples were collected for 40Ar/39Ar and U-Pb dating, for petrographic analysis in thin sections, and for major and trace element analyses. Geographic coordinates were recorded with a portable GPS (UTM projection system, Datum: D_WGS_1984, zone 16P). Samples were organized and classified in the Observatorio Ambiental of the Ministerio de Medio Ambiente y Recursos Naturales (MARN) of El Salvador, packed and sent to the Centro de Geociencias (CGEO) of the Universidad Nacional Autónoma de México (UNAM) in Querétaro, Mexico. Results of the geochronologic, petrographic and geochemical analyses are described in Section 4. Prior to sample preparation for both chemical and geochronological analysis, samples were inspected for purity and freshness with the binocular mi- croscope, and then in thin sections for petrographic analysis. 3.2. Geochronology  techniques     Zircon crystals were separated and dated by U-Pb isotopic analysis at the Laboratorio de Estudios Isotópicos (LEI) of CGEO-UNAM, using a 193 nm Resolution M50 laser ablation inductively coupled plasma mass spectrometer (LA-ICP-MS) Thermo ICAP Qc following the method described in Solari et al. (2010) and Ortega-Obregón et al. (2013). Pum- ice fragments were separated from the different units, removing altered surfaces and then crushed and sieved to fractions of 74–44 μm. This frac- tion was then washed using a plastic pan to concentrate the heavy crys- tals from the lighter glass. Mineral concentrates were magnetically separated by means of a Frantz Isodynamic Separator to further concen- trate the zircons. Representative zircons were handpicked under a bin- ocular microscope and checked for purity and zoning using an ELM-3R luminoscope by cathodoluminescence, before and after the ICP-MS laser ablation analysis. About 50 zircon grains per sample were selected in order to obtain a statistically representative age. Samples for 40Ar/39Ar dating were sent directly to the Oregon State University, Corvallis, U.S.A., for analysis in the OSU Argon Geochronol- ogy Laboratory. As with zircons, clean pumice clasts were separated and crushed to obtain mineral concentrations. Crushed samples were sieved to 500–177 μm fractions and washed with deionized water using an ultrasonic bath to remove dust. Feldspars were magnetically separated from amphiboles and pyroxenes, and all phases were acid- leached following methods of Koppers et al. (2011). Sanidine or ortho- clase is the preferred mineral phase in 40Ar/39Ar dating of young tephras. Unfortunately, no high-potassium feldspar was found in any of the IC samples. The only available phases to date were plagioclase and hornblende (ultra-pure separates picked clean of melt inclusion- rich crystals) and groundmass concentrations (lava samples). Incre- mental heating technique on bulk samples was employed to examine each phase analyzed. We used concentrates of the minerals since indi- vidual grains do not yield enough radiogenic argon for accurate single- crystal total fusion analyses (Rose et al., 1999). To achieve the highest possible precision in the 40Ar/39Ar age determinations, a large number of heating steps (22–33 heating steps) were carried out for each sample (Koppers et al., 2011). Age plateaus were chosen including as many con- tiguous and concordant step ages as possible. Before analyzing a sample, and after every three heating steps, system blanks were measured. In this way, after the first 7 to 14 low temperature steps, an adequate amount of discordant gases was released from alterations remaining and atmospheric contamination. To get the appropriate eruption timing it was necessary to recalculate the ages using the Kuiper et al. (2008) age for the Fish Canyon Tuff as flux monitor, reducing the data with the ArArCALC v2.5.1 software from Koppers (2002). All 40Ar/39Ar age er- rors reported here are 2σ. More details of the techniques employed in the dating process are provided in Appendix B and in Koppers et al. (2011). 3.3. Geochemistry     Chemical analyses were performed on whole-rock pumices that were previously dried in an oven at 90 °C for 24 h and hand-cleaned 42 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19   5 by removing the weathered surface under careful clean conditions to avoid contamination. Samples were crushed and grinded by hand using an agate ceramic mortar, until a homogenous powder with −200 mesh size was obtained using new plastic mesh for research stan- dards in each sample. Major, trace, and Rare Earth-Elements (REE) were obtained. Major and some trace elements were measured by means of X-Ray Fluorescence (XRF) technique in the laboratory of Instituto de Geología, UNAM (Mexico), using a X RIGAKU ZSX Primus II spectrome- ter. REE and other trace element analyses were obtained by ICP-MS in the LEI of CGEO. Method of sample handling and analyses techniques are described in Bernal and Lozano-Santacruz (2005). 4. Results 4.1. Stratigraphy  and  distribution  of  the  early  Ilopango  caldera  ignimbrites     The systematic geologic mapping and volcanic stratigraphy carried out on IC deposits has allowed for the identification and characteriza- tion of the three lowermost explosive eruptions from IC. They comprise broadly distributed Quaternary pyroclastic deposits, from oldest to youngest: Olocuilta, Colima and Apopa ignimbrites. They consist of widespread ignimbrite sheets covering the central portion of El Salvador (Fig. 2). 4.1.1. Olocuilta  Ignimbrite  (OI)   The first volcanic eruption from IC occurred at 1.785 ± 0.006 Ma BP (Ar/Ar age, Section 4.3 of geochronology results) and produced the high-K rhyolitic OI (Table 1). This pyroclastic deposit consists mostly of large pumice (40% vesicles), and lithics (15–25% total volume) in an ash matrix. In general, the pumices contain phenocrysts of plagioclase, quartz, pyroxenes and Fe-Ti oxides with accessory zircon and apatite. OI overlies a sequence of volcano-clastic sediments deposited over the Zaragoza Ignimbrite (Weyl, 1957), which is an older andesitic-dacitic tuff from Jayaque Caldera (Fig. 3). OI is light red to pink due to the alter- ation of the ferromagnesian components. At ILO-73, near the town of Olocuilta, the contact between the top of OI and the base of CoI is ex- posed (Figs. 2, 3 and 4d). OI shows a basal thin layer of pumice-lapilli fall between 30 and 25 cm overlaid by tens of meters of a massive ignimbrite unit (Fig. 4a), forming a thick deposit observable at distances of up to 30 km from the caldera (e.g., at ILO-184, Fig. 2). Diffuse internal stratifi- cation is marked by lenses of pumice (PlensL, Figs. 3 and 4e). The ignim- brite shows a 1–3 m thick basal pumice breccia (PmBr, Fig. 3), containing pumice up to 20 cm in diameter. The middle zone of the OI is a lithic-rich, massive lapilli tuff (mLTl, Fig. 3), at the medial to proxi- mal facies on the northern flank of IC, with lithics of 15 cm (ILO-251, ILO-10 and ILO-219, Fig. 2). At the top of OI and in particular in the southern outcrops, it is welded 8–10 m thick, with columnar jointing (Fig. 4b). Welding reaches from partially (eutaxitic texture, with pumice lapilli collapse, ILO-64; Figs. 2, 4f and 6a), to densely welded (with fiammes, ILO-96; Figs. 2 and 6b), being IV–V of Quane and Russell (2005) rank. The welded top of OI acted as a resistant caprock that protected the unconsolidated lower zones of this ignimbrite from ero- sion and formed an extensive plateau that dominates the geomorphol- ogy of the Balsamo Mountain Range (Figs. 2 and 4c). Perennial streams have incised deeply this plateau reaching the basement rocks (lavas and volcanic breccias of the Balsamo formation). Some of these canyons ex- pose entire stratigraphic sections of the OI up to 100 m thick such as ex- posures in ILO-192 (Figs. 2 and 4c). Above the welded level, there is an additional 25 m of non-welded ignimbrite that is indurated by vapor- phase silicification (ILO-180; Fig. 2), indicating a total thickness of ~125 m for OI. Table 1 Whole-rock chemistry of OI, CoI and ApI pumices. Sample ILO-199 ILO-313 ILO-21 ILO 36-A ILO-64 ILO-184 ILO-36-B ILO-82 ILO-73 ILO-134 ILO-78 Unit OI OI OI OI OI OI CoI CoI CoI ApI ApI SiO2 74.9 76.3 75.0 74.6 75.9 74.8 71.9 72.5 72.2 71.9 71.8 TiO2 0.13 0.11 0.13 0.16 0.13 0.12 0.35 0.35 0.39 0.30 0.29 Al2O3 14.9 13.0 14.1 14.9 13.7 14.8 14.9 14.5 15.3 15.3 14.3 FeO* 1.48 1.27 1.54 1.71 1.42 1.70 2.45 2.36 2.76 3.09 3.19 MnO 0.08 0.06 0.06 0.06 0.06 0.06 0.09 0.10 0.10 0.11 0.10 MgO 0.28 0.30 0.41 0.34 0.28 0.32 0.51 0.55 0.55 0.60 0.83 CaO 1.14 1.12 1.21 1.14 1.11 1.16 1.63 1.54 1.54 2.20 2.85 Na2O 2.38 2.06 2.30 2.43 2.14 2.87 2.64 2.66 2.56 3.53 3.71 K2O 4.69 5.72 5.22 4.60 5.22 4.05 5.30 5.35 4.57 2.87 2.84 P2O5 0.03 0.03 0.03 0.03 0.03 0.03 0.06 0.07 0.07 0.08 0.09 LOI 3.27 2.56 4.58 2.22 3.87 4.24 4.25 3.89 5.34 4.19 5.54 SUM 99.9 99.9 100.1 99.8 99.9 99.8 100.1 99.9 99.9 99.9 99.7 Rb 87 107 106 94 105 85 126 117 125 48 43 Sr 114 111 123 119 114 117 180 160 179 208 190 Ba 1212 1146 1297 1200 1322 1242 1163 1211 1172 1100 956 Y 17 19 17 15 15 18 37 32 36 19 11 Zr 104 89 114 123 106 102 301 273 323 210 92 Nb 5 5 4 3 2 5 7 5 6 5 3 V 5 4 8 11 19 6 17 12 15 15 23 Cr 0 0 b3 4 9 0 0 b3 b3 1 b3 Co 2 2 5 8 3 3 4 7 8 5 8 Ni 7 7 10 10 10 8 7 9 10 7 8 Cu 5 7 20 13 12 12 7 10 10 19 10 Zn 29 13 35 30 34 28 51 54 58 54 43 Th 7 6 8 7 8 7 7 8 10 3 3 Pb 11 7 8 7 9 9 14 12 15 8 5 X (m) 274,665 275,498 272,823 256,099 272,901 316,801 256,099 269,885 270,915 304,297 264,763 Y (m) 1,501,091 1,493,097 1,495,122 1,493,793 1,494,947 1,520,480 1,493,793 1,554,548 1,500,438 1,516,540 1,526,998 OI: Olocuilta Ignimbrite, CoI: Colima Ignimbrite, ApI: Apopa Ignimbrite. Major elements in wt% and trace elements in ppm. Samples analyzed by X-Ray Fluorescence in the Insituto de Geología (UNAM) by Patricia Girón. Coordinates in WGS84 system (zone 16P). FeO*-total iron; LOI: Lost of ignition. 43 6 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19       Fig. 3. Composite stratigraphic section of the early Ilopango caldera products, showing lithofacies type (following nomenclature of Branney and Kokelaar, 2002 and Brown and Branney, 2004), interpretation and age for each deposit. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 4.1.2. Colima  Ignimbrite  (CoI)   The second large explosive eruption from IC occurred at 1.57 ± 0.01 Ma BP (Ar/Ar age, see Section 4.3 of geochronology results) produc- ing the CoI, which is a high-K rhyolitic unit (Table 1) with plagioclase, hornblende, and accessory minerals of apatite, zircon and Fe-Ti oxides. This ignimbrite, like the previous one, starts with a thin layer of pumice-lapilli fallout overlying the paleosol developed at the top of OI (e.g. ILO-73 and ILO-36; Figs. 2, 3, 4d and 5b). Above this basal layer, there is a sequence of pyroclastic density currents (PDC) deposits com- posed mostly of a pale grey ash matrix with large white and highly ve- siculated pumice (50 vol%), and minor amounts of lithics. This ignimbrite traveled distances of up to 40 km to the north of the caldera, reaching the Lempa River Water Reservoir (ILO-82; Figs. 2 and 5a). In proximal facies of the caldera it is very difficult to find outcrops where the CoI stratigraphic section is completely exposed due to the heavy vegetation and because this unit is covered by younger Ilopango tuffs. However, in medial facies like at the ILO-150 (Fig. 2), there are at least 19 m of this deposit are observable. There is no evidence of lithic breccias or welded zones in the CoI and, in general, it is an unconsolidated tuff easily eroded by the tropical rains of El Salvador's weather, mainly on the steep slopes of the southern flank of IC, where the ignimbrite is sporadically exposed. The CoI is rel- atively massive, with large pumices concentration zones, generally at the base and middle part of the unit forming pumiceous breccias (pmBr in Fig. 3, pumice of ~20 cm). Within these breccias there are pieces of intrusive cognate blocks among the components (e.g., at ILO- 101; Figs. 2 and 5c). These are vesiculated fragments of crystalline igne- ous rock with 50% total volume in crystals with a non-fragmented 44 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19   7 Fig. 4. Field and sample photographs of OI. a) Quarry at ILO-64 (Fig. 2) exposing 50 m of OI (note geologist for scale). b) Top of OI showing moderate to dense welding with columnar jointing (ILO-64, Fig. 2). c) View to the south from ILO-92 (Fig. 2), where the Tapalhuaca canyon cuts the plateau formed by the welded OI and exposes at least 100 m of this ignimbrite. d) Basal pumice fall (mpL) of the CoI above OI at ILO-73 (Fig. 2). e) Pumice massive Breccia (PmBr) indicating PDC pulses within OI. f) Hand specimen of OI with eutaxitic texture sampled at ILO-64 (Rank IV of welding intensity scale, Quane and Russell, 2005; Fig. 2). vesiculated glass matrix (Fig. 6c). Polished thin sections of these en- claves show glomeroporphyritic texture with plagioclase and horn- blende (Fig. 6d), and poikilitic texture with small hornblende within plagioclase large phenocrysts (Fig. 6e). We infer that these cognate blocks are crystal  mush  clasts, which correspond to broken pieces of the subcaldera magma chamber wall. 4.1.3. Apopa  Ignimbrite  (ApI)   The third large explosive eruption associated with IC produced the ApI at ~1.34 Ma BP (age of a overlaying lava, see Section 4.3). This ignim- brite is a medium-K rhyolitic tuff (see Table 1) with plagioclase and hornblende as principal phases and minor amounts of biotite and clinopyroxene, with accessory zircon, apatite and Fe-Ti oxides. At the base of ApI there is a sequence of sub-parallel dilute PDC deposits (llbsT, surge deposits; Fig. 3) mostly composed of fine, yellow-white ash and accretionary lapilli, and small fragmented clasts of pumice (ILO-175; Figs. 2, 5d and e). The matrix is highly fragmented ash tuff, with sparse complete glass shards, and fine interstitial ash-dust and tiny fragments of the mineral assemblage (Fig. 6f). These fine-grained deposits of the ApI eruption resemble soft clay, probably due to the con- tent of zeolites or other alteration minerals. They reach up to 5 m in thickness, with diffuse cross-lamination and load structures (flames  in ILO-327, Figs. 2 and 5g). Directly above these laminated deposits there are: 1) a thin layer of orange pumice-lapilli (mpL), rich in crystals (10–15 vol%), with mostly hornblende and plagioclase, and 2) a sequence of denser lithic-rich PDCs deposits (mLT, Fig. 5d). These PDCs of the upper part of the ApI show a diffuse internal stratification with several pumice-rich horizons (plensL) and a lithic breccia layer (mlBr). This breccia can be correlated in several proximal outcrops to the north of IC and it is also observed 45 8 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19       Fig. 5. Field and samples photographs of CoI and ApI. a) CoI at ILO-82 (Fig. 2) mostly composed of white pumice and pale grey glass shards matrix. b) Contact of CoI and OI at ILO-36 (Fig. 2). c) Hand-size sample of a piece of crystal  mush  found in the pumice breccia of CoI at ILO-101 (Fig. 2): this clast is an intrusive cognate lithic from the magma chamber wall. d) Complete sequence of ApI above a paleosol at ILO-175 (Fig. 2) showing dilute hydromagmatic PDC deposits at the base (llbsT), a pumice fall layer in the middle (mpL) and denser PDC deposits at the top (mLT). e) Detail of the basal dilute PDCs deposits of ApI showed in the previous photo, where stands out the accretionary lapilli and internal lamination. f) Proximal facies of ApI along the Ilopango Lake shore (ILO-340, see Fig. 2) showing a coignimbritic lithic lag breccia zone. g) Load and flame structures within the dilute PDC deposits at the base of ApI at ILO-327 (Fig. 2). also along the shore of the Ilopango Lake, where angular lithics can be as large as 50 cm in diameter (ILO-145, ILO-225 and ILO-340, Figs. 2 and 5f). The ApI is the first IC ignimbrite that contains a heterolithological lithic breccia, with lava and plutonic lithic clasts, interpreted as a coignimbrite lithic lag breccia (ILO-89; Figs. 2 and 5g). The ApI reaches as far as the Lempa River Water Reservoir (ILO-82; Fig. 2), throughout a valley to the north for ~40 km. The accumulation of hot material filling this valley caused the incipiently welding of the top of ApI, with adhe- sion between clasts but remaining undeformed (rank II, Quane and Russell, 2005). At ILO-102, located about 20 km to the NW of IC (Fig. 2), this welded top level protected the underlying and unconsoli- dated tuff of ApI from erosion. At this site the ignimbrite reaches about 14 m of thickness. On the southern facies, as with the CoI, the ApI is sporadically exposed due to erosion. At some locations, 46 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19   9 Fig. 6. Microscopic characteristics of the OI, CoI and ApI. a) Phenocryst of Quartz (Qrtz) within the welded zone of OI, showing resorption embayment and glassy shards matrix (thin section polished from Fig. 4f). b) Microtexture of the dense welded OI at ILO-96 site (Rank V, Quane and Russell, 2005; Fig. 2) showing collapsed glass shards and fiammes (in yellow), around rigid plagioclase phenocrysts. c) Interlocking texture of the crystal mush piece from CoI (Fig. 5c), with vesiculated and non-fragmented glass matrix (Gl). The mineralogy of this cognate includes hornblende (Hbl) and plagioclase (Plag). d) Glomeroporphydic texture presented in the same crystal mush piece (grouped plagioclase and hornblende crystals). e) Poikilitic texture in the CoI crystal mush piece (small hornblende in large plagioclase crystals). f) Thin section of the basal surges deposits of ApI with very fragmented glass shards matrix (Gl), crystals of clinopyroxene (Cpx), biotite (Bi), plagioclase (Plag), hornblende (Hbl) and Fe-Ti oxides (Mgt), and lithics (Li). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 7. Geochemical classification of OI, CoI and ApI. a) Total-Alkali-Silica (TAS) diagram with classification of LeBas et al. (1986) showing that the three ignimbrites are subalkaline rhyolites, but each with a well-defined cluster in the plot. b) SiO2-K2O plot with classification of Peccerillo and Taylor (1976) showing that OI and CoI are in the High-K calcalkaline suite, and ApI in the Medium-K suite. 47 10 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19     (e.g., ILO-175, Fig. 2), the ApI is only covered by a consolidated lahar de- posit that protected this ignimbrite from fluvial erosion (Fig. 5d). 4.2. Geochemistry     Whole-rock chemical analyses were carried out on pumice frag- ments collected from the three ignimbrites (Table 1) in order to obtain the geochemical characteristics of the earliest volcanic products of IC. The pyroclastic products of the IC are all subalkaline rhyolites with high SiO2 content (Fig. 7a), and high-K2O values for the first two erup- tions (OI and CoI), and medium content for the third one (ApI, Fig. 7b). Harker variation diagrams (Fig. 8) show evolved magmas for the three ignimbrites, with OI as the most evolved, which is the richest in SiO2 and the poorest in total FeO and MgO. From OI to CoI to ApI there is a tendency for a slightly less evolved magma. The three ignimbrites show the typical calcalkaline nature of subduction type magmas along continental margins, as elsewhere in CAVA (Arculus and Curran, 1972; Morris and Ryan, 2004; Murphy, 2007). Abundances of the Rare Earth Elements (REE) and other key trace el- ements are displayed in a spider  diagram (Fig. 9), in which the elements are arranged from left to right in order of increasing compatibility, and values normalized to Mid-Ocean Ridge Basalt values (MORB, Sun and McDonough, 1989). This diagram confirms the comagmatic origin for the three ignimbrites, with some features that can be used for correla- tion purposes. For example, the CoI generally has higher concentrations of REE than the other two ignimbrites and a notable negative anomaly in the Sr and Eu. OI also has a negative anomaly in Sr but not in Eu. Fi- nally, ApI has no Eu anomaly, and a positive Sr anomaly (Fig. 9). Nega- tive Eu anomalies probably reflect plagioclase fractionation (Wilson, 1989). Plagioclase is the most abundant phenocryst in calcalkaline magmas, typically dominated over alkali feldspar (Murphy, 2007), as occurs in the Ilopango ignimbrites here described. 4.3. Geochronology     4.3.1. U/Pb  dating  in  zircon   Four samples from the three first IC ignimbrites were collected for radiogenic U-Pb zircon analyses, with a single sample from OI (ILO-64, Figs. 2 and 10a), two from CoI (ILO-73 and ILO-82, Figs. 2, 10b and c), and one from ApI (ILO-78 site, Fig. 2). Analytical data are summarized in Appendix A. The U-Pb zircon ages were obtained from the intersec- tion of the Concordia curve with the normal isochron (Fig. 10a, b and c) using the Isoplot software (v. 3.7) and methods of Ludwig (2008). To obtain the best possible results, we discarded the most discordant zircons (largest ellipsoids in the isochron diagram). U/Pb zircon age for OI yielded 1.64 ± 0.19 Ma (1.45–1.83 Ma range). The large error is caused probably because most zircons analyzed were too small (b80 μm), which have xenocrystic cores with concentric zoning (Fig. 10d and e), and due the limitations of the U-Pb method in such young zir- cons (all the errors of U/Pb reported in this study refer to 2σ  standard deviation). The most probable age of the crystallization of this zircon age is close to 1.8 Ma. The two samples of zircon analyzed for the CoI were collected from the proximal facies to the south of IC (ILO-73, Fig. 2) and at 40 km to the north of IC (ILO-82, Fig. 2). The zircon ages are 1.53 ± 0.28 Ma and 1.55 ± 0.12 Ma, for southern and northern sites respectively, supporting the correlation of this ignimbrite between the proximal and distal facies (Fig. 10b and c). No age for ApI could be obtained pos- sibly because of the low content in U and Pb for this ignimbrite (see the relative concentrations in the spider diagram of Fig. 9). Most ApI zircons were inherited from an old basement of around 70 Ma. Zircon analyses in such young products as the IC ignimbrites yield imprecise results, so it was necessary to use another geochronological method trying to better constrain the timing of each ignimbrite. We thus estimated 40Ar/39Ar ages. 4.3.2. 40Ar/39Ar  dating   Five new 40Ar/39Ar ages (Table 2 and Fig. 11) were obtained by in- cremental heating methods using the ARGUS-VI mass spectrometer. Five high-purity plagioclase concentrates were analyzed for the OI (ILO-64 site; Fig. 2), which were then stacked to produce a new pla- teau age of 1.785 ± 0.006 Ma (Fig. 11a). This high-precision age of OI is within the analytical error of previous reported ages for this unit (1.77 ± 0.4 Ma, 1σ; Lexa et al., 2011) and within the U-Pb zircon age reported above (1.64 ± 0.19 Ma 2σ). The CoI samples were ana- lyzed by plagioclase and hornblende separates from the same site Fig. 8. Harker variation plots of OI, CoI and ApI. Note well-defined groups of each ignimbrite. Arrows indicate possible evolving trends (symbols as in Fig. 8). 48 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19   11 Fig. 9. Spider diagrams of OI, CoI and ApI normalized to MORB (Sun and McDonough, 1989) showing the REE abundances. (ILO-73, Fig. 2), yielding 1.57 and 1.56 ± 0.01 Ma, respectively (Fig. 11b and c). This result agrees with the U-Pb zircon ages of this work (Fig. 10b and c). No 40Ar/39Ar ages were obtained for the ApI because the hornblende and plagioclase yielded too much atmospheric argon to obtain reliable ages. Since these analyses were unsuccessful, we sampled the unit overlying ApI: Las Pavas Lava (LPL, Fig. 3), a post-caldera dome emplaced just after the eruption of ApI (ILO-43 site, Fig. 2). Plagioclase and a groundmass separates were analyzed for this lava (Fig. 11d and e), yielding ages of 1.34 ± 0.02 Ma and 1.22 ± 0.02 Ma respectively. The LPL groundmass yielded a slightly younger age than the plagioclase age due to excess atmospheric argon. There is no well-developed Fig. 10. U-Pb zircon age concordia and isochron diagrams of OI and CoI including images of the analyzed zircons by means of LA-ICPMS. ApI results were discarded (see Section 4.3.1). a) OI (ILO-64). b) CoI (ILO-82). c) CoI sample (ILO-73). d) Cathodoluminescence image of the analyzed zircons from OI's ILO-64 sample. e) Scanning Electron Microscope (SEM) image of zircons from OI's ILO-64 sample after laser ablation ICP-MS analyses. 49 12 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19     Table 2 Summary of incremental heating 40Ar/39Ar analyses on early Ilopango caldera products. Sample information Age spectrum Total fusion Inverse isochron analyses Unit Sample X (m) Y (m) Material Age ± 2σ   39Ar K/Ca MSWD n N Age ± 2σ   K/Ca Age ± 2σ   40/36 MSWD (Ma) (%) (Ma) (Ma) intercept Olocuilta Ign. ILO-63 271,353 1,497,223 Plagioclase 1.785 ± 0.006 86.37 0.072 1.10 87 124 1.792 ± 0.06 0.073 1787 ± 0,006 291.80 ± 3.46 1.05 Colima Ign. ILO-73-P 270,915 1,500,438 Plagioclase 1.57 ± 0.01 100 0.058 2.77 22 22 1.56 ± 0.01 0.058 1,58 ± 0,01 276,71 ± 8,18 1.47 Colima Ign. ILO-73-H 270,915 1,500,438 Hornblende 1.56 ± 0.01 100 0.058 1.38 22 25 1.55 ± 0.01 0.058 1.57 ± 0.02 294.05 ± 3.21 1.39 Las Pavas Lava Las Pavas Lava ILO-43-P 288,141 1,518,462 Plagioclase 1.34 ± 0.02 92 0.030 1.39 19 28 1.37 ± 0.01 0.028 1.33 ± 0.04 303.55 ± 1.42 K/Ca values are calculated as weighted means for the age spectra or as total fusion K/Ca values by combining the gas analyses. Both the number of steps (n) included in the age plateau and isochron calculations and the total number of incremental heating steps (N) have been listed. MSWD values for the age plateaus and inverse isochrons are calculated using n − 1 and n − 2 degrees of freedom, respectively. All samples from this study where monitored against FCT-FM sanidine (28.03 ± 0.18 Ma) as calibrated by Kuiper et al. (2008). Reported errors on the 40Ar/39Ar ages are at the 95% confidence level (2σ) including 0.3–0.4% standard deviation in the J-value. paleosol at the top of this ignimbrite that contacts directly with the base of LPL, so the most probably age for the ApI is close to 1.34 Ma. Tables with all the analytical data of these Ar/Ar measurements are given in Appendix C. 4.4. Stratigraphic  correlation     Stratigraphic descriptions, geological mapping and analytical results have allowed us to make a spatial and temporal correlation for the first 3 Fig. 11. 40Ar/39Ar age spectra showing high-resolution incremental heating steps for OI, CoI and LPL. Results of ApI were discarded and instead LPL was used to constrain the age of ApI (see Section 4.3.2). The weighted average age or plateau age (±2σ) is indicated. a) 5 stacked runs of plagioclases with different crystal sizes from OI. b) 1 plagioclase run from CoI. c) 1 hornblende run from the same CoI sample shown in b). d) 1 plagioclase run of LPL. e) 1 groundmass run of the same LPL sample of d). 21.24 ILO-43-G 288,141 1,518,462 Groundmass 1.22 ± 0.03 41 0.345 1.97 11 33 1.27 ± 0.01 0.345 1.22 ± 0.04 295.36 ± 1.40 2.18 50 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19   13 ignimbrites of IC (Fig. 12). Four profiles (A, B, C and D) containing repre- sentative measured stratigraphic sections (logs) show the thickness and distribution of the early IC ignimbrites, as well as the lithofacies changes from proximal to distal deposits from the caldera. 5. Discussion 5.1. Volcanic  phases  of  early  ignimbrite-­‐forming  eruptions  of  Ilopango   caldera     OI represents the first and largest explosive eruption of IC, which began with a brief eruption column that deposited a thin layer of pumice-lapilli fall. This column collapsed apparently by depressurization due to the sudden opening and widening of the vent, from a central vent to a larger fissure vent possibly related to the exten- sional tectonics of the area (e.g., Aguirre-Díaz and Labarthe-Hernández, 2003; Costa et al., 2011 and Section 5.3). The eruption then changed to extrusion of dense PDCs from the wider and larger fissure vent probably with a boiling-over style (e.g., Pacheco-Hoyos et al., 2018), as indicated by the pumice-rich nature of the OI and the long runout distances (Roche et al., 2016), which suggest lower explosivity and fragmentation of magma during eruption. At least 74 km3 of pyroclastic material equiv- alent to 50 km3 DRE came out of the vent (Section 5.2), filling the paleotopographic lows and covering a total area of about 3000 km2, with runouts of at least 40 km. These PDCs were deposited continuously forming one single cooling unit, the massive OI. To the south of the Fig. 12. Representative stratigraphic sections of OI, CoI and ApI overlying the Balsamo Formation materials, including lavas, volcanic breccias and dacitic-andesitic ignimbrites (note inset with index map showing location of each log site). a) Representative logs to the south of IC. b) Representative logs to the north of IC (Tierra Blanca Joven-TBJ is the youngest explosive event of IC at about 1500 BP). c) Representative logs to the northeast of IC (the 1.34 Ma LPL is covering ApI in ILO-43 with no paleosol between). d) Representative logs to the east of IC (note the San Vicente strike-slip Fault affecting the Carbonera caldera rim). 51 14 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  371  (2019)  1–19               Fig. 12 (continued). caldera this ignimbrite reached a thickness of N100 m (ILO-192, Figs. 2 and 4c), forming a medium to intensely welded level and building a pla- teau that stands out in the geomorphology: the Balsamo Mountain Range (Figs. 2 and 4c). To the north, this ignimbrite is only slightly welded (ILO-251; Fig. 2), probably because OI in this lithofacies (Fig. 12b) is very rich in lava lithics that tend to cool the PDC deposit more rapidly (Marti et al., 1991). Lithic-rich proximal facies of the OI can be interpreted as coignimbritic lithic lag breccias (Branney and ¡g IProtilo e -C' I .ll..Q:.lll """""- """'-'- !>. ~ -- -- " "' ' , - , -mg- - - ~ - , , .' @l • " ~ ," ,..< ~, - ~ , ". , - -- --, ... ~In:!, -, @ ~, ® @> @ [ , ® ~, 3 o ® ,~ • ® 3 - H - , - ~ n:: Ji .. , -- ~ ...... l!&.l.H. ........ [QJ 1001 """-'- ~ --~ UllWbo_" O _"O E ~ • @ , • • l,opiINf .. ..,... iil Pl>mko bf 1.34 :1 0.02 mLT Ma(Sui\e- Puchol.2019) 66 108 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119       Fig. 7. Stratigraphy and representative outcrops of ILO-20 (Fig. 2). a) Schematic stratigraphic section of this site. b) ILO-20 site at roadcut in Panamericana Highway on the NW flank of IC (note the person to scale, yellow circle). c) Sketch of ILO-20 outcrop including paleosols between ignimbrites and the regional faults affecting the deposits of Altavista Formation (note the person to scale, yellow circle). d) Oblique view of ILO-20 outcrop showing the terraces of the Staircase site. e) Detail photo of pumice fall at the base of CorI with two interbedded surges and cross-stratified lapilli tuff (xsLT) deposited from dilute PDCs. f) Detail of the pumice fall at the base of ManI with hydromagmatic overlying surge PDCs deposit, and g) detail photo of accretionary lapilli from the xsLT of Manigua Ignimbrite (ManI). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 67 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119   109 • "'T pl€ns xsU @) m,' m" 'N, ® Acuetionary I ~ p ilti h Paleo\.Ol , . A(tu~ l .oi l ILO-8l mLT '"' 'N, • • o • mU Ar/Aranalysis Zi rcom¡ analy~i~ Ch"mical analy\i~ Petrography secUan Lithafacie~ type (se<'! figure 4) lliU2 '" (@ nT ' NO Farmatian boundary Member boumlary Unit boundary ~ lIthic breccla D Lapillituff II1II lapiHi falllaye! Cros5-stra tification e;:¡ ploduce.:! by hydfDmagm a t i ~m lJ.!UlQ m" • C® • IIsLT @) mLT 'N, B'l ~ i ~ II , , , ., i ! , ! mCT • ® IIsLT reworked C§!) mLT 'N, 68 110 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119       Fig. 9. Micro-photographs of polished thin sections of pumice clasts from the Altavista Formation tuffs of Ilopango caldera. a) Image in crossed Nichols of pumices from Cojutepeque Ignimbrite, with phenocrysts of plagioclase (plag) and hornblende (hbl) in a vesicular texture. b) The same micro-photograph that in a) but in parallel light. c) Pyramidal section of a hornblende phenocryst from San Juan Fall showing glomeroporphydic texture together with plagioclase crystals. d) Same image than c) in crossed Nichols. e) Tabular clinopyroxene (cpx) of egirine from San Juan Fall (image in crossed Nichols). f) The same egirine of e) in parallel light. 4.3. Geochronology  of  the  Altavista  Formation     A complete set of new isotopic ages of the Altavista Formation is pro- vided in this study in order to have, for the first time, a more detailed volcanic history of IC. 4.3.1. Results  of  U/Pb  and  U/Th  analysis   Six samples of zircon-separates that encompass the entire strati- graphic section of Altavista Formation were prepared to determine the age of each tuff-forming eruption. The first five members of the Altavista Formation were analyzed by U/Pb methods and only the last member was done by U/Th. Ages of the first five deposits were calculated with the intersection of the Concordia curve and the normal isochron (Figs. 12a, b, d and e), using the Isoplot software v. 3. 7, and following methods of Ludwig (2008). For ManI, the age was obtained as the aver- age of the non-inherited zircons ages (Fig. 11c). SoI age (Figs. 12f) was calculated using IntCal09 software (Reimer et al., 2009), to calibrate U/ Th ages (Bernal et al., 2014). The first member of the Altavista Formation, CojI, is dated at 960 ± 750 ka (Fig. 12a). The large error is caused probably because most Fig. 8. Representative stratigraphic sections of the six members of Altavista Formation. a) S\\N profile along the Panamerican Highway at NW of IC and b) W-E profile along the same road on the NE flank of the caldera. ILO localities and stratigraphic sections are shown in Fig. 2. 69 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119   111 Table 2 Whole-rock chemistry of the Altavista Formation tuffs of Ilopango caldera. Sample ILO-20-P ILO-20-E ILO-89-C ILO-20-F ILO-132-A ILO-144-A Unit SoI(TB5) CorI SJF (TB7) ManI DeI (TB9) CojI (TB6) (TB8) (TB10) SiO2 70.4 72.2 69.0 75.0 74.6 70.8 TiO2 0.28 0.23 0.40 0.25 0.21 0.33 Al2O3 18.5 17.5 17.8 14.8 13.9 16.2 FeO* 3.63 2.83 4.42 2.09 2.31 3.18 MnO 0.11 0.09 0.13 0.09 0.09 0.09 MgO 0.58 0.61 0.77 0.69 0.65 0.74 CaO 1.50 1.52 2.52 1.58 2.00 1.98 Na2O 1.27 1.02 2.52 1.54 2.89 2.42 K2O 3.48 3.99 2.51 3.94 3.29 4.31 P2O5 0.04 0.04 0.05 0.03 0.04 0.03 LOI 7.29 8.62 5.29 6.80 3.81 5.08 SUM 99.8 100.1 100.0 100.2 99.9 100.1 Rb 69 115 47 74 66 132 Sr 139 242 245 152 216 202 Ba 1009 864 1121 1147 1171 1091 Y 20 34 15 13 15 17 Zr 151 301 137 159 110 193 Nb 4 14 4 3 4 4 V 27 31 32 21 14 29 Cr b3 12 b3 b3 b3 b3 Co 5 8 7 6 b4 8 Ni 7 13 8 8 7 10 Cu 23 11 15 13 8 20 Zn 48 58 44 41 31 45 Th 6 19 4 6 6 10 Pb 9 22 9 8 8 11 X (m) 266,586 266,586 292,473 266,586 287,416 289,231 Y (m) 1,518,661 1,518,661 1,517,803 1,518,661 1,518,785 1,518,825 OI: Olocuilta Ignimbrite, CoI: Colima Ignimbrite, ApI: Apopa Ignimbrite. Major elements in wt% and trace elements in ppm. Samples analyzed by X-Ray Fluorescence in the Instituto de Geología (UNAM) by Patricia Girón. Coordinates in WGS84 system (zone 16P). FeO*-total iron; LOI: lost of ignition. analyzed zircons were too small (b80 μm), with xenocrystic cores and concentric zoning, and also due to the limitations of the U\\Pb method in such young zircons (all the errors of U/Pb reported in this study refer to 2σ  standard deviation). To obtain the best possible results, we discarded the most discordant zircons (largest ellipsoids in the isochron diagram, Fig. 12). The second member of this formation, the widespread DeI, is dated at 830 ± 140 ka (Fig. 12b). The third member, ManI, is dated at 750 ± 96 ka (Fig. 12c). The fourth member of this formation, SJF, yielded an U/Pb age of 590 ± 330 ka (Fig. 12d). The fifth member, CorI, yielded an age of 499 ± 73 ka (Fig. 12d). SoI, the youngest member of Altavista Formation, yielded an U/Th age of 257 ± 33 ka (Fig. 12f). All zircons age data are compiled in Appendix C. 4.3.2. Results  of  40Ar/39Ar  analyses   Five new 40Ar/39Ar ages were performed to obtain a second set of re- sults for comparison and to get more precise age constraints (Table 3 and Fig. 13). Four high-purity plagioclase and one hornblende separates were analyzed. Plagioclase of CojI (sampled at ILO-144, Fig. 3) yielded a flat plateau with an age of 918.8 ± 17.4 ka (Fig. 13a). This 40Ar/39Ar age for CojI improves the U\\Pb results presented above showing a smaller error. Plagioclase and hornblende separates of DeI did not yield mean- ingful 40Ar/39Ar ages due to the large abundance of atmospheric argon; thus, the U\\Pb age is the best one we have for this member, al- though with a large analytical error (Fig. 12b). The plagioclase concen- trate analyzed for ManI pumice yielded a 40Ar/39Ar age of 768.3 ± 49.4 ka. The age spectrum was discordant whose shape is classic “shaddle-shaped” type (Fig. 13c), which probably represents a mix of excess and atmospheric argon. This 40Ar/39Ar age agrees with the U\\Pb results obtained with zircons (Fig. 11c). SJF's age was calculated on hornblende phenocrysts because the plagioclase contained numer- ous melt inclusions throughout 99% of the grains, yielding a 40Ar/39Ar plateau age of 626.0 ± 75.1 ka (Fig. 13d). CoI's age was obtained from the analysis of a plagioclase separate, yielding a plateau age of 553.0 ± 16.6 ka (Fig. 13e). The 40Ar/39Ar analyses of this fifth member constrain better the age obtained by U\\Pb (Fig. 12e). SoI's age includes only few small plagioclase crystals, which resulted in low quantities of gas and higher uncertainties for many of the heating steps. We chose the upper heating steps that were the most concordant that yielded a weighted mean age of 335 ± 302.1 ka (Fig. 13f). A list of all 40Ar/39Ar age (plateau, mini-plateau, total fusion, normal and inverse isochron ages), as well as MSWD's and K/Ca ratios are given in Table 3. 5. Discussion 5.1. Volcanic  phases  and  eruptive  styles  of  the  Altavista  Formation  tuffs     The style and eruption dynamics within the Altavista Formation are very diversified. Eruption processes for each one of the six members of Fig. 10. Major element chemical classification plots for the Altavista Formation tuffs. a) Total Alkali Silica (TAS) diagram (after Le Bas et al., 1986). b) K20 vs SiO2 diagram (after Peccerillo and Taylor, 1976), where is remarkable the High-K content of the Altavista Formation. See Table 2 for chemical data. 70 112 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119       Fig. 11. N-MORB normalized multi-element diagram including of REE abundances of the Altavista Formation tuffs (normalizing values of Sun and McDonough, 1989). this formation were interpreted from their deposits (Fig. 15) and are de- scribed below. 1. Cojutepeque Ignimbrite (CojI; ~918 ka). The vent of the first erup- tion of the Altavista Formation that produced the CojI was local- ized in the NE sector of the caldera, inferred by the distribution of the associated ignimbrite (Fig. 3). The CojI eruption started with a Plinian or Sub-Plinian column (Fig. 14a) that deposited a 2 m thick pumice fall layer up to 2 km to the NE of the caldera rim (Fig. 5f). Later, as the sharp contact between this pumice- fallout and the overlying cross-bedded PDC deposit indicates (Fig. 5c), a drastic change in the eruptive style occurred (Fig. 14b). The initial eruptive column collapsed, probably due to the widening of the conduit or by an increase of the mass eruption rate (MER; Costa et al., 2018). This MER increasing could be produced by the beginning of a caldera collapse, which would start the rapid evacuation of magma from the subcaldera magma chamber. The presence of accretionary lapilli (Fisher and Schmincke, 1984), suggests a water-rich environment, likely resulting for the interaction of magma with water from the Ilopango caldera lake. But later, the eruption returned to a mag- matic style (Fig. 14c), as inferred by the presence of a massive tuff deposited at the top of CojI, formed by a series of lithic-rich PDCs deposits without accretionary lapilli. In fact, the caldera col- lapse onset likely coincides with the appearance of a lithic-rich PDC. This shift from hydromagmatic to magmatic style could be caused by a high magma/water ratio due to an increase of the magma eruption rate (Wohletz, 1986; Wohletz et al., 2013), prob- ably caused by the caldera collapse. This collapse may have been as trap-door type (Aguirre-Díaz, 2008), due the limited distribu- tion of the associated ignimbrite only present in the NE sector of the caldera (Fig. 3). This interpretation matches with the models of Aguirre-Díaz and Martí (2015); Aguirre-Díaz et al. (2016) and Saxby et al. (2016), who suggest this kind of collapse mechanism in the past and current IC. 2. Delgado Ignimbrite (DeI; ~830 ka). The second eruption of Altavista Formation apparently started with hydromagmatic explosions. This early activity was inferred from the parallel-stratified, fine-rich de- posit with accretionary lapilli at the base of DeI (Fig. 6e), probably produced by very dilute PDC's (Fig. 14d). Later on, the hydromagmatic eruption produced highly turbulent PDCs distrib- uted in all directions (Fig. 14e), depositing a widespread cross- stratified tuff, which is richer in solid components and thus appar- ently derived from denser PDCs than those that formed during the initial blasts (Figs. 3 and 6f). The last phase of this eruption appar- ently marks a possible change to a magmatic eruptive style, indicated by a lithic-rich level within a massive ignimbrite deposited on the top of this member (Fig. 5e). This kind of lithic-rich zones can be interpreted as a coignimbritic lag breccia suggesting the moment of a caldera collapse (Fig. 14f). Roof collapse of the magma chamber causes a massive and rapid evacuation of magma (Folch and Martí, 2009) that makes less effective the magma-water interactions neces- sary to carry out hydromagmatic explosions. The caldera collapse could have been the cause of the generation of dry magmatic PDCs at the end of the DeI eruption. The result is a massive and widely dis- tributed ignimbrite sheet (DeI). 3. Manigua Ignimbrite (ManI; ~768 ka). The eruption associated to ManI began as purely magmatic, producing likely an eruptive col- umn (Fig. 14g) that deposited a 30 cm thick layer of pumice fallout in the NW sector of IC (Fig. 7f). Later, the eruption became hydromagmatic, probably due to the input of water from the paleo- Ilopango lake into the vent, drastically increasing the explosivity of the eruption. The resulting product is a sequence of dilute and cross-bedded PDCs deposits, ~8 m thick, composed of ash matrix and accretionary lapilli, with some thin lens-shape layers rich in pumice fragments (pLens, Fig. 7), which serve as cross- stratification markers. This eruption apparently was laterally di- rected towards the W-NW flank of the caldera, since the ignimbrite is only observable in this sector of IC (Fig. 3). ManI eruption could be also linked to a partial collapse event, probably with a trap-door collapse style (Fig. 14h) due to the limited distribution of ManI in the eastern sector of the caldera (Fig. 3). 4. San Juan Fall (SJF; ~625 ka). This member was produced completely by a single eruptive column, probably Plinian or Sub-Plinian type (e.g., Bonadonna and Costa, 2013; Fig. 14i), which formed a thick pumice fallout of 5–6 m (Figs. 6d). This eruption was most likely purely magmatic, as no evidence of hydromagmatic processes were found in field exposures. This observation suggests that the vent of the SJF eruption was located outside the paleo-Ilopango lake, and probably in the northeast sector of IC because SJF is only observable in this direction, near Cojutepeque City (ILO-43, 89 and 316; Fig. 3). This eruption may not have produced a caldera collapse, because there is no associated widespread ignimbrite or other evidence of caldera roof subsidence. As in Las Pavas Lava and other IC products, 71 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119   113 Fig. 12. Results of zircons analyses from the members of Altavista Formation. a) Concordia diagrams with normal isochron to calculate the ages of CojI, and b) DeI by U/Pb methodology. c) Average ages of the most concordant zircons from ManI (U/Pb). d) Concordia diagrams with normal isochron to calculate the ages of SJF, and e) CorI respectively (by U/Pb). f) Age calculation by U/Th of SoI (the blue spheres are zircons in secular equilibrium, see Bernal et al., 2014). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) SJF may have erupted from the outer margin of the caldera, probably from vents controlled by the faults of the pull-apart system that have been linked to previous caldera eruptions of the Comalapa and Altavista Formations (Saxby et al., 2016; Suñe-Puchol et al., 2019). 5. Cortez Ignimbrite (CorI; ~553 ka). The eruptive processes associated to CorI are very similar to those of the previous ManI member. This eruption started with an eruptive column that deposited a thin layer of pumice fallout (b1 m; Fig. 7e). The vent of this eruption may be localized at the northwest sector of the caldera (Fig. 14j), in- ferred by the distribution of the associated pumice deposit, which is absent at the base of CorI on the northeast flank of IC. The initial erup- tive column was suddenly interrupted by hydromagmatic explosions, which deposited a fine-ash tuff with accretionary lapilli derived probably from dilute PDCs (Fig. 14k). We suggest that this eruption was linked to a caldera collapse event based on the rela- tively large distribution of CorI, which it is still preserved at both flanks of IC (ILO-87, ILO-221 and ILO-20 in the NW sector, and in ILO-89 in the NE sector; Fig. 3). 6. Soyapango Ignimbrite (SoI; ~257 ka). The last eruption of the Altavista Formation also started with a sustained eruptive column in the NW sector of IC (Fig. 14l), which deposited N2 m of pumice fall- out as far as 5 km from the caldera's topographic margin (Figs. 7 and 14l). This phase depressurized the subcaldera magma chamber and could have caused another caldera collapse. In this case, the possible collapse may have been complete along the caldera structure. The 72 114 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119     Table 3 Summary of Incremental Heating 40Ar/39Ar Analyses on Altavista Formation tuffs of Ilopango Caldera. Sample information Age spectrum Total fusion Inverse isochron analyses Unit Sample X (m) Y (m) Material Age ± 2σ   (Ka) 39Ar (%) K/Ca MSWD n N Age ± 2σ   (Ka) K/Ca Age ± 2σ   (Ka) 40/36 intercept MSWD Soyapango Ign. ILO-20-P 266,586 1,518,661 Plagioclase 335.4 ± 302.1 33.74 0.0232 0.81 5 18 2858.0 ± 512.3 0.0234 509.9 ± 341.8 294.53 ± 1.44 0.86 Cortez Ign. San Juan Fall ILO-20-E ILO-89-C 266,586 292,473 1,518,661 1,517,803 Plagioclase Hornblende 553.0 ± 16.6 626.0 ± 92.40 99.94 0.0428 0.0246 0.92 0.33 15 24 20 25 538.0 ± 20.2 599.8 ± 88.6 0.0445 0.0247 559.5 ± 23.3 679.0 ± 295.06 ± 1.21 291.83 ± 1.11 0.18 Manigua Ign. ILO-20-F 266,586 1,518,661 Plagioclase 75.1 768.3 ± 55.76 0.0327 1.32 9 24 691.2 ± 0.0340 89.4 768.2 ± 3.49 297.93 ± 2.14 Cojutepeque ILO-144-A 289,231 1,518,825 Plagioclase 49.4 918.8 ± 76.10 0.0261 0.46 13 24 190.7 907.2 ± 15.7 0.0264 87.6 910.2 ± 2.16 316.77 ± 0.46 Ign. 17.4 32.7 69.32 K/Ca values are calculated as weighted means for the age spectra or as total fusion K/Ca values by combining the gas analyses. Both the number of steps (n) included in the age plateau and isochron calculations and the total number of incremental heating steps (N) have been listed. MSWD values for the age plateaus and inverse isochrons are calculated using n-1 and n-2 degrees of freedom, respectively. All samples from this study where monitored against FCT-FM sanidine (28.03 ± 0.18 Ma) as calibrated by Kuiper et al. (2008). Reported errors on the 40Ar/39Ar ages are at the 95% confidence level (2σ) including 0.3–0.4% standard deviation in the J-value. Fig. 13. High-resolution incremental heating 40Ar/39Ar age spectra for the members of Altavista Formation. The weighted average age (± 2σ) is shown above a black bar that indicates the heating steps used in the calculation of a) CojI, b) ManI, c) SJF, d) CorI, and e) SoI ages. 73 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119   115 eruption style changed from vertical column to low pyroclastic fountaining and radial ejection of dense PDCs (Fig. 14m), overflowing the crater rim and forming a widespread ignimbrite, ~5 m thick, and observed up to 10 km from the caldera margin. This ignimbrite is rich in pumice lenses and lithics (heterolitologic). At the final stage, this eruption became slightly hydromagmatic as suggested by the presence of fine-ash particles (higher magma fragmentation) and accretionary lapilli in detriment of larger compo- nents, such as pumices and coarser ash matrix, as well as the change from a massive deposit to cross-stratified laminated deposit, al- though still with pumice-rich lenses. This member crops out on both northern sectors of the caldera (NW and NE), but as in CorI, the pumice-fall layer at the base of the SoI is not evident around the northeast of IC. Fig. 14. Sketch (not to scale) illustrating the eruptive processes that formed the deposits of the Altavista Formation, since ~918 to ~257 ka BP. a) Eruptive column and associated pumice fallout at the beginning of Cojutepeque Ignimbrite (CojI), followed by b) dilute PDCs generated from hydromagmatic explosions, and c) denser PDCs formed during a possible partial collapse with trap-door style. d) “Blast-type” explosions at the beginning of Delgado Ignimbrite (DeI), followed by e) dilute and turbulent hydromagmatic PDCs and by f) total caldera collapse that generated a lithic-rich ignimbrite. g) Brief eruptive column at the beginning of the Manigua Ignimbrite (ManI) followed by h) hydromagmatic and dilute PDCs erupted from a partial collapse at the east. i) Eruptive column that deposited the fallout unit of San Juan Fall (SJF). j) Initial phase of eruptive column with intercalated blast-type explosions at the initial stage of Cortez Ignimbrite (CorI) eruption, followed by k) caldera collapse and massive generation of dilute PDCs. l) Eruptive column during the first phase of Soyapango Ignimbrite (SoI)’s eruption followed by m) lithic-rich PDCs produced by caldera collapse. o W-E Q) W-E Cojl-91S ka BP COjl - 918kaBP @ W-E ® W-E Del -830 ka BP Cojl-918 ka BP ~ M~ Manl-768kaBP Manl -768 ka SP 74 116 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119       Fig. 14 (continued). 5.2. Volume  and  eruption  magnitude  of  the  Altavista  Formation  tuffs     Deposit volumes were calculated applying the Delaunay triangu- lation method (Macedonio and Pareschi, 1991), by using the thick- nesses measured in the field. All data was imported, managed and processed into an attribute table within ArcGIS 10.2 by ESRI ©, fol- lowing methods described in Pitcher et al. (2017). Because of the poor preservation of the Altavista Formation deposits, mainly due to erosion, the corresponding volumes are underestimated. The Delgado Ignimbrite (DeI) was the best member for volume estima- tion due to its widespread distribution and preservation (Fig. 3). A minimum volume of 5 km3 DRE was estimated for the DeI. This value corresponds to a magnitude of 6 (Pyle, 2000). This estimation ignores distal ash-fall facies, caldera fill and eroded material over the Balsamo Mountain Range (BMR, southern flank of IC, Fig. 3). Con- sidering that DeI deposits are thicker and more widespread than the TBJ ignimbrite (see Fig. 3), DeI may have had a volume similar to TBJ or even N40 km3 DRE, reaching a magnitude of 7. For the other five members of the Altavista Formation, we have established a minimum and conservative estimate, using the thick- nesses and distribution of sparse incomplete outcrops, resulting in the order of 1 to 5 km3 DRE volume for each, which could reach around 30 km3, as those reported for the recent Tierras Blancas tuff-eruptions (CEL, 1992), if the eroded parts are also considered in the estimate. 5.3. Recurrence  period  of  the  Ilopango  Caldera  explosive  eruptions     The IC has generated at least thirteen explosive eruptions during the Quaternary, most of them ignimbrite-forming events. According to Suñe-Puchol et al. (2019), the first three explosive phases, grouped as the Comalapa Formation, occurred in a time span of 440 ka, from 1.78 Ma (1st eruption) to 1.34 Ma (3rd eruption). Those three ignimbrite-forming eruptions had a similar recurrence period of around 220 ka (blue dashed square in Fig. 15). The 40Ar/39Ar ages presented in this study indicate a long quiescence pe- riod between the Comalapa Formation and the first eruption of the Altavista Formation (the CojI eruption, at ~918 ka). This volcanic hi- atus was as large as ~ 420 ka (Fig. 15). Furthermore, the recurrence period of explosive eruptions within the Altavista Formation is quite shorter than the return time of the first eruptions. This lasted about 100 ka (Fig. 15), which is practically half the time of the fre- quency of the Comalapa Formation eruptions. The chronogram of Fig. 15 shows all the IC major explosive erup- tions recorded from the stratigraphy and geochronology, by inte- grating the data of Suñe-Puchol et al. (2019) for the Comalapa 75 g . 1 5 . C h ro n o g ram o f Ilo p an g o cald era ig n im b rite-fo rm in g eru p tio n s d u rin g its activ ity , fro m th e fi rst ig n im b rite at 1.785 M a, to th e last o n e, T ierra B lan ca Jo v en ig n im b rite at ab o u t 1.5 k a. T h e ch ro n o g ram h ig h lig h ts th e lo n g p erio d o f q u iescen ce betw een C om alap a an d A ltav ista Form ation s, an d also th e lo n g h iatu s in v olcan ic activ ity betw een A ltav ista an d T ierras B lan cas Form ation s. T h e recu rren ce p erio d w ith in th e C o m alap a F o rm atio n eru p tio n s (~ 220 k a, b lu e d ash ed rectan g le) are tw ice lo n g er th an th e A ltav ista F o rm atio n eru p tio n s (~ 100 k a, g reen d ash ed rectan g le), an d m u ch h ig h er th an th e T ierras B lan cas Form ation eru p tion s (~ 20 k a, red d ash ed rectan g le). T h e fu ch sia d ash ed ellip so id h ig h lig h ts an an o m alo u s lo n g v o lcan ic h iatu s (~ 300 k a) w ith in th e A ltav ista F o rm atio n , w h ich m ay in d icate th at m o re eru p tio n s co u ld h av e h ap p en ed b etw een C o rtez an d So y ap an g o eru p tio n s (C o rI an d , So I). (F o r in terp retatio n o f th e referen ces to co lo u r in th is fi g u re leg en d , th e read er is referred to th e w eb v ersio n o f th is article.) Eroptions Zircons "'~ 40Ar/39Ar "'~ Mo.t IIkely ." ~ :¡> o ª N ~ O> • . ~ " Comalapa Formatlon Altavlsta Format lo n of Ilopango Group Olocu ilta Ign. Colima Ign. Apopa Ign. Cojutepeque Delgado 18n. MMligua Ign. Jan l uan Cortel 18n. Soyapango 18n. (01) (col) (ApI ) Ign.(Cojl) (Del) (Manl) Fall (5Jf ) (Cor1 ) (501) 1.64 ±0.19 1.55 ± O.1Z ------- M. M. 960 ± 750 ka 830 ± 140 750±96ka 590 ± 330 499:1: 73 I:.a 257:1:33 ka k. .. 1.785 ±0.006 1.57 ± 0.01 > 1.36 ± M. M. O.OZ Ma 9 18 .8:1: 17.4 ka 768.3 ± 49 .4 626.0 :1: 553.0 :1: 16.6 335 :1: 302. 1 ----- k. 75.1 ka .. .. 1.785 M~ 1.57 M ~ 1.34 Mil 918 kiI 830 ka no .. 625 Ka 553 ka 257 ka (X) ( - ) (*) (" ) (.:. ) (O) (- ) (+ ) (~) ~-------------------. ~--------------------------. CqJI DeI MaDI SJ\!' CorI , Bol , 1 l' • 1 I 1 1 1 1 1 1 1 , ... 1 I 1 11 1 '" 1 1 10, 11 ,..., , .... , I 11 1 1 f \ "'P"' r ---- .;.. ;- --~ '"' 1 1' 1 1 1 1 1 I II I ~ I ,' , . , I 1 I II 1 1 1 I 1 1 : ' 11 1 1 I 1 1 I I I ~ 1 1 1 I 1 I : ' \ j$i 11 1 I 1 ~ 1 l--l9r' 1 1+'1 1 OI CoI ApI TIerras Blancas Fm . TB4, Tlla, 'DL d'''' lBl (re 's) , , <: 571o:.a (.0.) ~ --- , TB'. , '"T""J4 "1 ' . 1 , . ' 11 : ...... 11 1 1 I 1 1, 1 , , I I ' r 1, • It(Ma) , 1.8 1.6 1.4 1.2 0.8 0.6 DA 0:2 ~--------------------------~ , O, ~- - _. 76 118 I.  Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119     Formation, the data of the present work for the Altavista Formation tuffs, and published data elsewhere for the Tierras Blancas Formation (Rose et al., 1999; Kutterolf et al., 2008). Eruption ages of U\\Pb, U\\Th and 40Ar/39Ar from our database are positioned along a time- scale bar with the corresponding analytical error bars. This chrono- gram shows a regular periodicity along the Altavista Formation eruptions, except for the time span between CorI (~553 ka) and SoI (~257 ka), which had a hiatus in the activity of ~300 ka (Fig. 15). In addition, it is evident another relatively long period of volcanic inac- tivity of ~200 ka between the Altavista and Tierras Blancas forma- tions of IC (Fig. 15). The most notorious observation from this chronologic chart is the relatively large quiescence periods between each formation, in the order of 200–400 ka each. The meaning of these quiet periods is still not clear, but we assume that the tectonic factor and the magma input rate are the two principal reasons. During the whole IC activity, from 1.78 to present, there was also intense tectonic ac- tivity in the area, and a seismogenically active period related to the ESFZ (Garibaldi et al., 2016). As previously reported in Aguirre-Díaz et al. (2016), Saxby et al. (2016), and Suñe-Puchol et al. (2019), the IC volcanic activity was in close relationship with the transtensive tectonic regime of the region, and IC is the result of several episodes of pull-apart-related caldera collapses that were linked to the strike- slip faults of the ESFZ. IC volcanism, since the initial eruptions to the last one of TBJ (1.5 ka), was, and is tied to the Neogene activity of the regional faults of the cen- tral El Salvador region. The recurrence of large explosive eruptions could be linked to tectonic slips and the corresponding displacements be- tween blocks, and probably, these tectonic events were the triggers for large explosive eruptions at IC, if the magma of the subcaldera cham- ber was close to eruptive conditions. 6. Conclusions Ilopango Caldera (IC) has had at least 13 large Quaternary explosive eruptions, mostly ignimbrite-forming events, which were grouped into three formations, Comalapa Formation (1.78–1.32 Ma), Altavista For- mation (918–257 ka) and Tierras Blancas Formation (b57 ka). The Altavista Formation, presented in this study for the first time, consists of ryhodacitic pyroclastic deposits that are the product of six explosive eruptions, including eruptive columns and dense and dilute PDCs of magmatic and hydromagmatic type. The largest member of the Altavista Formation is the Delgado Ignimbrite (DeI), with a mini- mum DRE volume N 30 km3 if eroded deposits are considered. The re- currence period of the Altavista Formation eruptions is ~100 ka, except between the two last events (553 ka CorI and 257 ka SoI). This is half the time of the Comalapa Formation explosive eruptions recur- rence period (~220 ka). A chronostratigraphic chart of all IC eruptions since the initial large Olocuilta ignimbrite (1.78 Ma; N150 km3 DRE) to the latest TBJ (~1.5 ka; 30 km3 DRE) shows large quiescence gaps of nearly 200 to 400 ka between major eruptive members. These long quiet gaps and in- tense activity within each formation are apparently linked with the transtensive tectonic activity of the ESFZ. IC is the result of several cal- dera collapse episodes, each for a major ignimbrite eruption, which were probably triggered by tectonic events, although this complex rela- tion needs to be investigate further. Supplementary data to this article can be found online at https://doi. org/10.1016/j.jvolgeores.2019.02.011. AcknowledgeThents This study was financed by CONACYT-CB grant 240447 to GJAD. We appreciate the logistical support of the Ministerio de Medio Ambiente y Recursos Naturales – MARN, and of the Policia Nacional Civil – PNC, of El Salvador. We thank the doctoral scholarship grant to the first author from CONACYT-Mexico. We acknowledge to Anthony Koppers for his help and support in the Ar/Ar geochronology laboratory at Oregon State University (OSU). We want to be grateful to Juan Vazquez for thin sec- tion elaboration, to Lozano Santacruz and Patricia Girón for XRF analy- ses, and Ofelia Perez for the REE and trace elements analyses. We also want to thank Victor Noll for his rappel training to sample the ignim- brites of Ilopango caldera. This manuscript was greatly improved by comments and suggestions from the editors Joan Martí and James Gard- ner, and also from two experts reviewers. References Agostini, S., Corti, G., Doglioni, C., Carminati, E., Innocenti, F., Tonarini, S., Manetti, P., Di Vincenzo, G., Montanari, D., 2006. Tectonic and magmatic evolution of the active vol- canic front in El Salvador: insight into the Berlín and Ahuachapán geothermal areas. Geothermics 35, 368–408. https://doi.org/10.1016/j.geothermics.2006.05.003. Aguirre-Díaz, G.J., 2008, Types of collapse calderas: Collapse calderas Workshop 19–25 October 2008, Querétaro, Mexico “Reconstructing the Evolution of Collapse Calderas: Magma Storage, Mobilization and Eruption”. IOP Conference Series: Earth and Envi- ronmental Science, 3 012021 (5pp) doi:https://doi.org/10.1088/1755-1307/3/1/ 012021 (http://www.iop.org/EJ/toc/1755-1315/3/1 Aguirre-Díaz, G.J., Martí, J., 2015. Graben Calderas: Examples from Mexico, Central America, and the Andes: 26th IUGG General Assembly, Prague, Czech Republic, June 22–July 2, 2015. Abstracts. https://www.czech-in.org/cm/IUGG/. Aguirre-Díaz, G.J., Suñe-Puchol, I., Dávila-Harris, P., Pedrazzi, D., Hernández, W., 2016. The Ilopango Caldera: A Long-Lived Caldera System of Central America: 6th International Workshop on Collapse Calderas, IAVCEI, Hokkaido, Japan; June 2016 (Abstracts). Aguirre-Díaz, J.G., Suñe-Puchol, I., Davila-Harris, P., Pedrazzi, D., Hernandez, W., Gutierrez, E., 2017. Volcanic History of the Ilopango Caldera, Central American Volcanic Arc. 113th Annual Meeting. GSA Cordilleran Section, Hawaii, USA, p. 2017. Alonso-Henar, J., Álvarez-Gõmez, J.A., Martinez-Díaz, J.J., 2017. Neogene-quaternary evo- lution from transpressional to transtesional tectonics in Northern Central America controlled by cocos: Caribbean subduction coupling change. J. Iber. Geol. https:// doi.org/10.1006/s415113-017-0034-2. Arculus, R.J., Curran, E.B., 1972. The genesis of the calc-alkaline rock suite. Earth Planet. Sci. Lett. 15, 255–262. Bernal, J.P., Lozano-Santacruz, R., 2005. Characterization of a new set of eight geochemical reference materials for XRF major and trace element analysis. Revista Mexicana de Ciencias Geológicas 22 (3), 329–344. Bernal, J.P., Solari, L.A., Gómez-Tuena, A., Ortega-Obregón, C., Mori, L., Vega- González, M., Espinosa-Arbeláez, D.G., 2014. In-situ230Th/U dating of Quater- nary zircons using LA-MCICPMS. Quat. Geochronol. 23, 46–55. https://doi.org/ 10.1016/j.quageo.2014.06.003. Bonadonna, C., Costa, A., 2013. Plume height, volume, and classification of explosive vol- canic eruptions based on the Weibull function. Bull. Volcanol. 75 (8), 1–19. Branney, M.J., Kokelaar, P., 2002. Pyroclastic density currents and the sedimentation of ig- nimbrites: Geol Soc London. Memoirs 27 (152 pp). Canora, C., Villamor, P., Martínez-Díaz, J.J., Berryman, K.R., Álvarez-Gómez, J.A., Capote, R., Hernández, W., 2012. Paleoseismic analysis of the San Vicente segment of the El Salvador Fault Zone, El Salvador, Central America. Geol. Acta 10, 103–123. https:// doi.org/10.1344/105.000001700. CEL (Comision Ejecutiva Hidroelectrica del Rio Lempa), 1992. Desarrollo de los Recursos Geotermicos del Area Centro-Occidental de El Salvador. Prefactibilidad Geotermica del Area de Coatepeque. Reconocimiento Geotermico. Informe Final. Internal report. Costa, A., Suzuki, Y.J., Koyaguchi, T., 2018. Understanding the plume dynamics of explo- sive super-eruptions. Nat. Commun. 9, 654. https://doi.org/10.1038/s41467-018- 02901-0. DeMets, C., 2001. A new estimate for present-day Cocos-Caribbean plate motion: Implica- tions for slip along the Central American volcanic arc. Geophys. Res. Lett. 28, 4043–4046. https://doi.org/10.1029/2001GL013518. Dixon, T.H., 1993. GPS measurement of relative motion of the Cocos and Caribbean plates and strain accumulation across the Middle America Trench. Geophys. Res. Lett. 20, 2167–2170. https://doi.org/10.1029/93GL02415. Donnelly, T.W., Horne, G.S., Finch, R.C., Lopez-Ramos, E., 1990. Northern Central America: the Maya and Chortis blocks. In: Dengo, G., Case, J.E. (Eds.), The Geology of North America: The Caribbean Region, pp. 37–76. Dull, R.A., Southon, J.R. and Sheets, P., 2001. Volcanism, Ecology and Culture: A Reassess- ment of the Voclán Ilopango TBJ eruption in the Southern Maya Realm, v. 12, No. 1, pp. 25–44. Dull, R.A., Southon, J.R., Kutterolf, S., Freundt, a., Wahl, D., Sheets, P., 2010. Did the TBJ Ilopango eruption cause the AD 536 event? AGU fall Meet. Abstr. 1, 2370. Fisher, R.V., Schmincke, H.-U., 1984. Pyroclastic Rocks: Berlin. Springer Verlag (472 p). Folch, A., Martí, J., 2009. Time-dependent chamber and vent conditions during explosive caldera-forming eruptions. Earth Planet. Sci. Lett. https://doi.org/10.1016/j. epsl.2009.01.035. Funk, J., Mann, P., McIntosh, K., Stephens, J., 2009. Cenozoic tectonics of the Nicaraguan depression, Nicaragua, and Median Trough, El Salvador, based on seismic-reflection profiling and remote-sensing data. Bull. Geol. Soc. Am. 121, 1491–1521. https://doi. org/10.1130/B26428.1. Garibaldi, N., Tikoff, B., Hernández, W., 2016. Neotectonic deformation within an exten- sional stepover in El Salvador magmatic arc, Central America: Implication for the in- teraction of arc magmatism and deformation. Tectonophysics 693, 327–339. https:// doi.org/10.1016/j.tecto.2016.05.015. 77 I. Suñe-­‐Puchol  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  374  (2019)  100–119   119 Golombek, M.P., Carr, M.J., 1978. Tidal triggering of seismic and volcanic phenomena dur- ing the 1879-1880 eruption of Islas Quemadas volcano in El Salvador, Central America. J. Volcanol. Geotherm. Res. 3, 299–307. https://doi.org/10.1016/0377-0273 (78)90040-9. Hernández, E.W., 2004, Características geotécnicas y vulcanológicas de las tefras de Tierra Blanca Joven de Ilopango, El Salvador: Tesis de maestría, Univ. Politécnica de El Salvador, San Salvador, 115 p. Hernández, W., 2008, Aspectos geológicos que influyen en las aguas subterráneas y en la respuesta sísmica del Area Metropolitana de San Salvador. Revista ASIA No. 162. 29- 41p. Hutton, D.H.W., Reavy, R.J., 1992. Strike-slip tectonics and granite petrogenesis. Tectonics 11, 960–967. Koppers, A.A.P., 2002. ArArCALC—Software for 40Ar/39Ar age calculations. Comput. Geosci. 28 (5), 605–619. https://doi.org/10.1016/S0098-3004(01)00095-4. Koppers, A.P., Gowen, M.D., Colwell, L.E., Gee, J.S., Lonsdale, P.F., Mahoney, J.J., Duncan, R.A., 2011. New 40Ar/39Ar age progression for the Louisville hot spot trail and impli- cations for inter–hot spot motion. Geochem. Geophys. Geosyst. 12: Q0AM02. doi: https://doi.org/10.1029/2011GC003804. Kuiper, K.F., Deino, A., Hilgen, F.J., Krijgsman, W., Renne, P.R., Wijbrans, J.R., 2008. Syn- chronizing rock clocks of Earth history. Science 320, 500–504. https://doi.org/ 10.1126/science.1154339. Kutterolf, S., Freundt, A., Schacht, U., Bürk, D., Harders, R., Mörz, T., Peréz, W., 2008. Pacific offshore record of plinian arc volcanism in Central America: 3. Application to forearc geology. Geochem. Geophys. Geosyst. 9. https://doi.org/10.1029/2007GC001826. LeBas, et al., 1986. A chemical classification of volcanic rocks based on the total alkali- silica diagram. J. Petrol. 27, 745–750. Lexa, J., Šebesta, J., Chavez, J.A., Hernández, W., Pécskay, Z., 2011. Geology and volcanic evolution in the southern part of the San Salvador Metropolitan Area. J. Geosci. 56, 105–140. https://doi.org/10.3190/jgeosci.088. Ludwig, K. 2008. Manual for Isoplot 3.7. Berkeley Geochronology Center, Special Publica- tion 4, rev. p. 77. Macedonio, G., Pareschi, M.T., 1991. An algorithm for the triangulation of arbitrarily dis- tributed points: applications to volume estimate and terrain fitting. Comput. Geosci. 17, 859–874. https://doi.org/10.1016/0098-3004(91)90086-S. Mann, P., 2007. Overview of the tectonic history of northern Central America. In: Mann, P. (Ed.), Geologic and Tectonic Development of the Caribbean Plate Boundary in North- ern Central America. 428. Geological Society of America Special Papers, pp. 1–19. Marshall, D.J., 1988. Cathodoluminescence of Geological Materials. Unwin Hyman, p. 181. Martí, J., Groppelli, G., & da Silveira, A. B., 2018. Volcanic stratigraphy: a review. J. Volcanol. Geotherm. Res. Volume 357, 15 May 2018, Pages 68–91. Martínez-Díaz, J.J., Álvarez-Gómez, J.A., Benito, B., Hernández, D., 2004. Triggering of de- structive earthquakes in El Salvador. Geology 32, 65–68. https://doi.org/10.1130/ G20089.1. Morris, J.D., and Ryan, J.G, 2004, Subduction zone processes and implications for changing composition of the upper and lower mantle, in Treatise on Geochemistry, ed., H.D. Holland and K.K. Turekian, Elsevier-Pergamon, Oxford, p. 451–470. Murphy, J.B., 2007. Igneous Rock Associations 8. Arc Magmatism II: Geochemical and Iso- topic Characteristics. Geosci. Can. 34, 1. Ortega-Obregón, C., Solari, L., Gómez-Tuena, A., Elías-Herrera, M., Ortega-Gutiérrez, F., Macías-Romo, C., 2013. Permian–Carboniferous arc magmatism in southern Mexico: U–Pb dating, trace element and Hf isotopic evidence on zircons of earliest subduction beneath the western margin of Gondwana. Int. J. Earth Sci. 103 (5), 1287–1300. https://doi.org/10.1007/s00531-013-0933-1. Peccerillo, A., Taylor, S.R., 1976. Geochemistry of Eocene calcalkaline volcanic rocks from the Kastamonu area, Northern Turkey. Contrib. Mineral. Petrol. 58, 63–81. Pedrazzi, D., Sunyé-Puchol, I., Aguirre-Díaz, G.J., Costa, A., Davila-Harris, P., Hernández, W., Gutierrez, E., 2018. The Ilopango Tierra Blanca Joven (TBJ) eruption, El Salvador: volcano-stratigraphy of a major Holocene event of Central America and hazards im- plications EGU General Assembly 2018. Vol. 20, EGU2018–8455. Pitcher, B. W., Kent, A.J.R, Grunder, A.L. and Duncan, R.A., Frequency and volumes of ignimbrite eruptions following the Late Neogene initiation of the Central Oregon High Cascades. J Volcanol Geotherm Res 339: 1–22, doi:10.1016/j. jvolgeores.2017.04.019. Pyle, D.M., 2000. The sizes of volcanic eruptions. In: Sigurdsson, H., Houghton, B., McNutt, S.R., Rymer, H., Stix, J. (Eds.), Encyclopedia of Volcanoes. Academic Press, London, UK, pp. 263–269. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., Weyhenmeyer, C.E., 2009. IntCal09 and Marine09 radiocar- bon age calibration curves, 0e50,000 years cal BP. Radiocarbon 51 (4), 1111e1150. Reynolds, J.H., 1987. Timing and sources of Neogene and Quaternary volcanism in south– central Guatemala. J. Volcanol. Geotherm. Res. 33, 9–22. Rose, W.I., Conway, F.M., Pullinger, C.R., Deino, A., McIntosh, K., 1999. An improved age framework for late Quaternary silicic eruptions in northern Central America. Bull. Volcanol. 61, 106–120. https://doi.org/10.1007/s004450050266. Saxby, J., Gottsmann, J., Cashman, K., Gutiérrez, E., 2016. Magma storage in a strike-slip caldera. Nat. Commun. 7, 12295. https://doi.org/10.1038/ncomms12295. Sheets, P.D., 1979. Environmental and cultural effects of the Ilopango eruption in Central America. Volcan. Act. Hum. Ecol. 525–564, ill. https://doi.org/10.1016/B978-0-12- 639120-6.50022-4. Sofield, D., 2004. Eruptive history and volcanic hazards of Volcan San Salvador. Geol. Soc. Am. Spec. Pap. 375, 147–158. 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 Oceanic Basins. 42. Geological Society of London Special Publication, pp. 313–345. Suñe-Puchol, I., Aguirre-Díaz, G.J., Dávila-Harris, P., Miggins, D.P., Pedrazzi, D., Costa, A., Ortega-Obregón, C., Lacan, P., Hernández, W., Gutiérrez, E., 2019. The Ilopango caldera complex, El Salvador: origin and early ignimbrite-forming eruptions of a graben/pull- apart caldera structure. J. Volcanol. Geotherm. Res. 371, 1–19. https://doi.org/ 10.1016/j.jvolgeores.2018.12.004. Weber HS, Wiesemann G, Wittek indt H (1974), Mapa Geológico de la República de El Salvador/Geologische Übersichtskarte der Republik El Salvador 1: 500,000 (after geo- logical maps 1:100000 – 1967-74) Bundesanstalt für Geowissenschaften und Rohstoffe, Hannover; Bundesanstalt fu¨r Bodenforschung, Hannover. Williams, H., Meyer-Abich, H., 1955. Volcanism in the southern part of El Salvador, with particular reference to the collapse basins of Lakes Coatepeque and Ilopango. Geol. Sci. Univ. Calif. Publ. 32 (64 p). Wohletz, K.H., 1986. Explosive magma-water interactions: Thermodynamics, explosion mechanisms, and field studies. Bull. Volcanol. 48 (5), 245–264. Wohletz, K.H., Zimanowski, B., Büttner, R., 2013. Magma-water Interactions. Modeling volcanic processes. Cambridge University Press, New York, pp. 230–257. 78 5. Estudio vulcano-estratigráfico de la erupción Tierra Blanca Joven (TBJ): caracterización física del mayor evento Holoceno en Centro America. Artículo: Pedrazzi, D., Suñe-Puchol, I., Aguirre-Díaz, G., Costa, A., Smith, V. C., Poret, M., Dávila-Harris, P., Miggins, D. P., Hernández, W. and Gutiérrez, E., 2019. The Ilopango Tierra Blanca Joven (TBJ) eruption, El Salvador: volcano-stratigraphy and physical characterization of the major Holocene event of Central America. J. Volcanol. Geotherm. Res. Contribuciones individuales de los autores: ØØ Dario Pedrazzi: diseño y organización del estudio, trabajo de campo, de laboratorio, modelado numérico, procesamiento, análisis e interpretación de datos, redacción del artículo. ØØ Ivan Suñé Puchol: organización y trabajo de campo, análisis en laboratorio, interpretación de datos, apoyo en la redacción del artículo. ØØ Gerardo Aguirre Díaz: financiamiento, supervisión, trabajo de campo, interpretación de los datos, revisión del artículo. ØØ Antonio Costa: estimación de parámetros físicos eruptivos. ØØ Victoria C. Smith: análisis químicos del vidrio de la TBJ ØØ Matthieu Poret: Modelado numérico para dispersión de ceniza ØØ Pablo Dávila Harris: trabajo de campo, interpretación de datos y corrección del artículo. ØØ Dan Miggins: análisis de laboratorio, interpretación de datos y corrección del artículo. ØØ Walter Hernández: trabajo de campo e interpretación de la geología de la zona. ØØ Eduardo Gutiérrez: apoyo logístico y trabajo de campo. 79 Journal of Volcanology and Geothermal Research 377 (2019) 81–102 The Ilopango Tierra Blanca Joven (TBJ) eruption, El Salvador: Volcano-stratigraphy and physical characterization of the major Holocene event of Central America Dario Pedrazzi a,, Ivan Sunye-Puchol b, Gerardo Aguirre-Díaz b, Antonio Costa c, Victoria C. Smith d, Matthieu Poret c, Pablo Dávila-Harris e, Daniel P. Miggins f, Walter Hernández g, Eduardo Gutiérrez g a ICTJA,  CSIC,  Group  of  Volcanology,  SIMGEO  UB-­‐CSIC,  Institute  of  Earth  Sciences  Jaume  Almera,  Lluis  Sole  i  Sabaris  s/n,  08028  Barcelona,  Spain   b Centro  de  Geociencias,  Universidad  Nacional  Autónoma  de  México,  Blvd.  Juriquilla  3001,  Campus  UNAM,  Querétaro  76230,  Mexico   c Istituto  Nazionale  di  Geofisica  e  Vulcanologia,  INGV-­‐Bologna,  Via  Donato  Creti,  12,  40100  Bologna,  Italy   d Research  Laboratory  for  Archaeology  and  the  History  of  Art,  University  of  Oxford,  1-­‐2  South  Parks  Road,  Oxford  OX1  3TG,  UK   e División  de  Geociencias  Aplicadas,  IPICYT,  San  Luis  Potosí  78216,  Mexico   f College  of  Earth,  Ocean  and  Atmospheric  Sciences,  Oregon  State  University,  104  CEOAS  Administration  Building,  101  SW  26th  St,  Corvallis,  OR  97331,  United  States  of  America   g Gerencia  de  Geología  del  Observatorio  Ambiental,  Ministerio  de  Medio  Ambiente  y  Recursos  Naturales  MARN,  San  Salvador  76230,  El  Salvador       a r t i c l e i n f o Article  history:   Received 29 August 2018 Received in revised form 7 March 2019 Accepted 8 March 2019 Available online 16 March 2019 Keywords:   Pyroclastic Density Currents Co-ignimbrite Tephra fallout Tephra dispersal modelling Ilopango caldera a b s t r a c t The Ilopango caldera is the source of the large Tierra Blanca Joven (TBJ) eruption that occurred about 1.5 ka years ago, between ca. AD270 and AD535. The eruption dispersed volcanic ash over much of the present territory of El Salvador, and pyroclastic density currents (PDCs) extended 40 km from the volcano. In this study, we document the physical characteristics of the deposits from all over El Salvador to further constrain the eruption processes and the intensity and magnitude of the different phases of the eruption. The succession of deposits generated by the TBJ eruption is made of 8 units. The eruption started with PDCs of hydromagmatic origin (Unit A0), followed by fallout deposits (Units A and B) that are b15 cm thick and exposed in sections close to the Ilopango caldera (within 10–15 km). The eruption, then, transitioned into a regime that generated further PDCs (Units C– F), these range from dilute to dense and they filled the depressions near the Ilopango caldera with thicknesses up to 70 m. Deposits from the co-ignimbrite plume (Unit G) are the most widespread, the deposits are found in Guatemala, Honduras, Nicaragua, Costa Rica and the Pacific Ocean and cm-thick across El Salvador. Modelling of the deposits suggests that column heights were 29 km and 7 km for the first two fallout phases, and that the co-ignimbrite phoenix plume rose up to 49 km. Volumes estimated for the fallout units are 0.15, 0.8 and 16 km3 dense rock equivalent (DRE) for Unit A, B and G respectively. The PDCs deposits volumes were estimated to be ~0.5, ~3.3, ~0.3 and ~9.1 km3 DRE for Units C, D, E and F, respectively. The combined volume of TBJ deposits is ~30 km3 DRE (~58 km3 bulk rock), indicating that it was one of largest Holocene eruptions from Central America. This eruption occurred while Mayan populations were living in the region and it would have had a significant im- pact on the areas within tens of kilometres of the vent for many years to decades after the eruption. © 2019 Elsevier B.V. All rights reserved. 1. Introduction Large caldera volcanoes pose a significant hazard to populations that surround them. In order to understand the likelihood and type of fur- ther activity it is key that the deposits of previous eruptions are well studied. This study focuses on the thick deposits of the Tierra Blanca Joven (TBJ) eruption from Ilopango Caldera, El Salvador. Ilopango Caldera (IC; Fig. 1), is a 13 by 17 km volcano-tectonic struc- ture filled by an intra-caldera lake (Mann et al., 2004), recently Corresponding author. E-­‐mail  address:  dpedrazzi@ictja.csic.es (D. Pedrazzi). interpreted as a strike-slip caldera by Saxby et al. (2016). The IC belongs to the San Salvador Extensional Step-over in the central part of the country (SSES; Fig. 1b; Garibaldi et al., 2016), which is in turn part of the El Salvador Fault Zone-ESFZ (Montero and Dewey, 1982; Siebert and Simkin, 2002; LaFemina et al., 2009; Corti et al., 2005; Turner Henry et al., 2007). The IC was formed and shaped by various erup- tions, and older (pre-57 ka) pyroclastic deposits are related to previ- ous caldera collapse episodes (Lexa et al., 2011; Aguirre-Díaz et al., 2017; Suñe-Puchol et al., 2019a, 2019b). There are only a few publi- cations that detail the eruptions in the last 57 ka, i.e. the TB4, TB3 and TB2 eruptions (Rose et al., 1999; Kutterolf et al., 2008a, 2008b; Hernández, 2004; Hernández et al., 2012; Mann et al., 2004) and some recent studies have been carried out on the pre-57 ka https://doi.org/10.1016/j.jvolgeores.2019.03.006 0377-0273/© 2019 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Journal of Volcanology and Geothermal Research jour nal homepage: www.elsevier. com/ locate/ jvolgeores 80 82 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102       Fig. 1. a) Geological setting of northern Central America; ESFZ: El Salvador Fault Zone; GF: Gulf of Fonseca; IG: Ipala Graben; JF: Jalpatagua Fault; ND: Nicaraguan Depression; PF: Polochic fault; b) simplified geological map showing all the major geological formations of El Salvador (Hernández, 2004). CC: Coatepeque Caldera; CG: Cerrón Grande dump; I: Izalco Volcano; IC: Ilopango Caldera; LO: Laguna Olomega; SA: Santa Ana Volcano; SM: San Miguel Volcano; SS: San Salvador Volcano; SV: San Vicente Volcano; c) Google Earth image of Ilopango caldera (IC) (US Depth of State Geographer 2018); SSMA: San Salvador Metropolitan area; IQ: Islas Quemadas. ignimbrites of Ilopango (Hernández, 2004; Hernández et al., 2010; Lexa et al., 2011; Aguirre-Díaz et al., 2017; Suñe-Puchol et al., 2019a, 2019b). More studies focused respectively on the eruption of a dacitic dome that formed the Islas Quemadas in Ilopango Lake (IQ; Fig. 1c) in 1879 (Richer et al., 2004), and a subaquatic eruption in this lake (Mann et al., 2004). 81 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   83 The last large explosive eruption of Ilopango volcano was the TBJ (Tierra  Blanca  Joven  – white young earth), which is estimated to have erupted ~30 km3 DRE of magma about 1.5 ka years ago, between AD270 and AD535 (Dull et al., 2001, 2010). The TBJ was a cataclysmic eruption (Rolo et al., 2004) and is considered to be the largest in Central America since the ca. 84 ka Los Chocoyos-Guatemala eruption (Dull et al., 2010). Outside of the zone of devastation by the TBJ eruption, there was a much larger area of prolonged depopulation (10– 150 years) following the TBJ eruption (Dull et al., 2001). Presently, the area around IC is densely populated with about 3,000,000 people living within 30 km of the caldera. The population den- sity during most of the late Holocene in El Salvador has been the greatest of any mainland country in the Americas (Daugherty, 1969; Denevan, 1992; Lovell and Lutz, 1995; Wilkie and Guadalupe Ortega, 1997). Since the last eruption was in AD1879, IC is still considered active and, it poses a major risk for El Salvador and neighbouring countries. In order to con- tribute to the hazard assessment at IC, we conducted a detailed field map- ping to further investigate the TBJ deposits with the aim of building on the previous work and accurately reconstructing the eruption sequence. There have been several publications about the TBJ eruption de- posits. They were first documented by Williams and Meyer-Abich (1955) and called “white earth” due to their peculiar white colour, al- though they were thought to originate from San Salvador Volcano. Fur- ther studies of IC deposits were carried out by the German Geological Mission (MGA) whilst they completed the 1:500,000 scale El Salvador Geological Map (Weber et al., 1974). They defined IC and divided the proximal deposits into Units s4 (TBJ deposits) and s3'a (TB4, TB3 and TB2 eruptions) as part of the San Salvador Formation. Later, Hart (1981) worked on the detailed stratigraphy of the TBJ deposits and identified two important eruptive stages; T1 and T2, whose products are subdivided into six units and associated with different eruptive phases. Subsequently, Hart and Steen-McIntyre (1983) described the stratigraphy and distribution of the TBJ tephra and Vallance and Houghton (1998) revised the stratigraphy of Hart and Steen-McIntyre (1983) and labelled the stratigraphic units, characterizing them litho- logically and refining associated eruptive processes. Recent works on TBJ by Hernández (2004) identified new ignimbrites (Alpha, Beta, and Grey) and detailed the characteristics of each unit in more detail. Despite all this studies, a detailed stratigraphic survey including mapping and reconstruction of eruptive dynamics was still lacking. This study presents new field descriptions, petrographic observations, major element glass geochemistry, granulometric data for the TBJ de- posits, and uses these data to further understand transport/depositional mechanisms and the corresponding eruption dynamics of the TBJ erup- tion. Moreover, the physical parameters of the eruption were deter- mined, including the total erupted mass, the height of the eruptive columns, the emission rate and, above all, reconstruct the distribution of the TBJ deposit using models and field observations. In particular, the stratigraphic and granulometric data obtained in the field were used to model the distribution of the TBJ tephra, including the disper- sion of the finest ash that covered vast areas (thousands of km2). 2. Geological setting 2.1. Central  America  and  El  Salvador  geodynamic  and  geology     El Salvador is located in North Central America, on the Pacific margin of the Caribbean Plate (Fig. 1a). To the north, this plate interacts with the North American plate with a relative velocity between plates of 19 mm/year (DeMets et al., 2000; Guzmán-Speziale et al., 2005; Funk and Mann, 2009). Towards the west of El Salvador, the relatively young Cocos Plate (b25 Ma; Protti et al., 1995; Barckhausen et al., 2001) subducts towards the NE under the Caribbean plate along the Middle America Trench, at a speed of 73–85 mm/year (Dixon, 1993; DeMets, 2001). The highest rate of continental tectonic deformation in El Salvador oc- curs in the El Salvador Fault Zone (ESFZ), a narrow E-W zone of right lateral faulting connected by pull-aparts, that extends for N150 km (Martínez-Díaz et al., 2004; Fig. 1a) from Guatemala, where it is known as the Jalpatagua Fault (JF), to the Nicaragua Depression (ND) (Canora et al., 2012). These faults are sub-parallel and affect volcanic products of Pleistocene-Holocene age (Corti et al., 2005). Geological and seismological analyses suggest that ESFZ is not laterally continuous and it has been subdivided into different sections (Martínez-Díaz et al., 2004; Corti et al., 2005). The chain of volcanoes along the Central American Volcanic Arc (CAVA; Fig. 1a) has been developing since the Tertiary (DeMets, 2001; Mann, 2007; Carr et al., 2007) and is part of the Pacific Ring of Fire (Simkin and Siebert, 1994; Carr et al., 2007; Saxby et al., 2016). The CAVA extends for N1000 km from the southeast of Mexico to the central valley of Costa Rica and defines an abrupt continental volcanic front lo- cated between 165 and 190 km from the Middle America Trench (Fig. 1a). Volcanoes of Panama are excluded from the CAVA as they are associated with the subduction of the Nazca Plate below the Carib- bean, which makes them distinct in composition and activity relative to those in the CAVA (Carr et al., 2007). Volcanism of the Volcanic Arc of El Salvador (VAES) constitutes one of the most active segments of the CAVA. VAES includes 21 active volca- noes, three of which have erupted in the last century: Santa Ana-SA, Izalco-I, San Salvador-SS and San Miguel-SM (Fig. 1b; Siebert and Simkin, 2002). Deposits from these volcanoes, together with volcanic rocks of ages ranging from the Cenozoic to the present, constitute most of the geology of El Salvador (F. ig. 1b) 2.2. Ilopango  caldera     The IC (Fig. 1c) is located b10 km from San Salvador City and it forms part of the same eruptive lineament as the San Salvador and San Vicente volcanoes (Fig. 1b). IC is located directly above faults in the San Salvador and San Vicente ESFZ segments within the San Salvador Pull-Apart (SSPA; Garibaldi et al., 2016), which is a tectonic structure-oriented NW-SE, with right trans-tensive dynamics, parallel to the Mesoameri- can trench. The transforming faults of the graben/pull-apart seem to control the morphology of IC, its formation and its volcanic eruptions (Sofield, 2004; Suñe-Puchol et al., 2019a), as described for other Graben Calderas (Aguirre-Díaz, 2008). Several authors, in their study of volca- nism in southern El Salvador, noticed that the IC was a volcanic- tectonic depression controlled by the faults of an ancient graben (Williams and Meyer-Abich, 1955; Golombek and Carr, 1978; Hutton and Reavy, 1992; Sofield, 2004; Aguirre-Díaz and Martí, 2015; Aguirre-Díaz et al., 2016, 2017). Recently, Saxby et al. (2016) interpreted IC as a strike-slip caldera. IC was the result of several col- lapses associated to large explosive ignimbrite-forming eruptions (Suñe-Puchol et al., 2019a, 2019b) as previously suggested by Williams and Meyer-Abich (1955). The topographic edge of IC has sev- eral semicircular bays (Fig. 1c), which are evidence for multiple collapse events (Lexa et al., 2011). 3. Methods Field mapping was carried out over an area of about 20,000 km2 across El Salvador to reconstruct the stratigraphy of the TBJ deposits and the stratigraphic relationships with other eruptive deposits. The characteristics of the deposits were recorded including grading, colour, sorting, apparent component content (juvenile and lithic fragments), and primary sedimentary structures. The nomenclature used in this study for the bed thickness, grain size and sorting of the pyroclastic de- posits follows that proposed by Sohn and Chough (1989). The classifica- tion of the primary volcaniclastic deposits follows White and Houghton (2006) and the nomenclature for volcanic stratigraphy is based on Martí et al. (2018), adopting the same criteria as Suñe-Puchol et al. (2019a, 2019b) for the previous Ilopango eruptions. A total of 82 stratigraphic sections were measured, but we focus here on 21 outcrops that we 82 84 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102     consider representative of the whole succession, and its spatial varia- tions and preservation of deposits. The geographical coordinates of the locations, stratigraphic sections and sampling points were recorded using a portable Garmin Dakota-20 GPS (precision of ∼3 m) and quoted on the UTM projection Datum: D_WGS_1984, zone 16 N. All this local information is reported in Sup- plementary Material 1. All the georeferenced data were managed and processed using the open source software Quantum GIS (Las Palmas; https://www.qgis.org/en/site/). Thicknesses of the deposits and specific units were measured to cre- ate a database (see Supplementary Material 1) for tephra dispersal sim- ulations (Macedonio et al., 2005). Tephra dispersal from virtual sources in an eruption column was simulated using the HAZMAP model, which solves equations for advection, diffusion and sedimentation of tephra particles in two dimensions (Macedonio et al., 2005). We followed an approach similar to Matthews et al. (2012) but used the Total Grain Size Distributions (TGSDs) (Bonadonna and Houghton, 2005) phases determined through the Voronoi Tessellation method, that we esti- mated for the different phases using data collected in this study. The granulometry data used to generate the TGSDs are available in Supple- mentary Material 2. Isopach maps were generated by modelling the ash deposition in terms of mass loading (kg/m2) and these were con- verted into thicknesses using a bulk density of 1000 kg/m3. In addition to the volumes, the solution of the inverse problem (Costa et al., 2009; Matthews et al., 2012) allowed us to estimate column heights, from which, by using the results of Mastin et al. (2009) and Bonadonna and Costa (2013), we assessed the corresponding Mass Eruption Rates (MER) for each unit. The volume estimations of the PDCs units were de- termined using the Delaunay triangulation method (Macedonio and Pareschi, 1991) that is particularly suitable for the reconstruction of vol- ume between geological horizons and the interpolation of bivariate data, when function values are available at irregularly-spaced data points, as in the case of geological outcrops. A binocular microscope was used to determine the main petro- graphic and textural characteristics of the juvenile components. In addi- tion, petrographic analyses were carried out in order to identify the mineralogy and general composition of the studied deposits. Thin sec- tions were produced at Wagner Petrographic LLC, a professional com- pany of Lindon, Utah (USA). Granulometric analyses were performed at the MARN (Ministerio de Medio Ambiente y Recursos Naturales) facilities of El Salvador Govern- ment and the Physical Volcanology Laboratory of Centro de Geociencias, Universidad Nacional Autonoma de Mexico (UNAM) in Juriquilla- Querétaro (Mexico). Representative levels of each stratigraphic unit were sampled and analysed (141 samples in total; Fig. 2 and Supplemen- tary material 2) for grain-size distribution and componentry. Grain-size analysis were performed by dry sieving at 1 phi (Φ) intervals through sieves with aperture sizes ranging from 64 to 0.25 mm (−6 Φ  to 3 Φ, where Φ  = − log2d  with d  is the diameter in mm) and by wet sieving through a MicroTec Analisette22 Fritsch from 0.125 mm to b0.01 mm (4 Φ  to N10 Φ). The weight percentages of the sieved fractions were cal- culated and then plotted as cumulative curves to give grain-size distribu- tion. All data from grain-size analysis are reported in Supplementary Materials 2, 3 and 4. The proportion of juveniles from −5 Φ  to 0 Φ  was defined by hand picking and from 0 Φ  to 2 Φ  using a binocular micro- scope and image analysis techniques (e.g. ImageJ software; https:// imagej.nih.gov/ij/). This point-counting method allows identifying the different components of each particle-size class using binocular micro- scope pictures. Modal proportions of juvenile pumice and accidental lithic fragments are reported in Supplementary Material 5. Whole rock pumice geochemical analyses for major elements, trace and rare earth-elements (REE) (Table 2) were measured at the CGEO LEI laboratory (trace and REE, with an ICP-MS) and at Instituto de Geología of UNAM (major and trace elements, X RIGAKU  ZSX  Primus  II  spectrom- eter), following standard sample preparation and analytical techniques (Bernal and Lozano-Santacruz, 2005). Electron probe X-ray microanalysis for mineralogy was performed using a JEOL JXA-8230 electron microprobe at the Scientific and Techno- logical Centers (Universitat de Barcelona). Wavelength-dispersive anal- yses of silicates were conducted using a 20 kV accelerating voltage and 15 nA current and with a focused beam. Glasses were analysed using a 6 nA current with a defocused 5–10 μm spot. Counting times were 10 s peak and 10 s background. A range of natural and synthetic stan- dards was used for calibration. The correction model XPP was used to convert X-ray intensity ratios into concentrations. Data are included in Supplementary Material 6. The major element compositions of the matrix glass of the TBJ were determined using wavelength-dispersive electron probe microanalysis (EPMA) in the Research Laboratory for Archaeology and the History of Art (RLAHA) at the University of Oxford. Analyses were carried out on samples from all units, A to G, and distal deposits located up to 130 km from the caldera. The EPMA of the TBJ glasses were acquired using an accelerating voltage of 15 kV, beam current of 6 nA, and 10- μm-diameter beam. The count times on peak were: 30 s for Si, Al, Fe, Ca, K and Ti; 50 s for Cl and Mn; 60 s for P; and 12 s for Na, and back- ground counts were collected for the same amount of time but split to positions either side of the peak. The PAP absorption correction method was used for quantification and the oxide compositions quoted assume stoichiometry. The electron probe was calibrated for each element using well-characterized mineral standards, which was verified by analysing MPI-DING reference glasses (Jochum Klaus et al., 2006). These MPI- DING glasses were used as secondary standards during each analytical run, and this data is included in the Supplementary Material 7 as they demonstrate the accuracy and precision of the TBJ datasets. All the glass analyses presented have been normalized to 100% to account for variable hydration and allow different samples to be compared, and all the raw compositional data can be accessed in Table 3. 4. Characteristics of the pyroclastic succession Proximal TBJ member products (0–10 km from the caldera) are ex- posed inside and close to the caldera with a maximum observed thick- ness of ∼60 m (Supplementary Material 1). The TBJ member can be divided in 8 units that were labelled alphabetically from base to top (A0-G; Fig. 2). Due to differences in dispersal patterns, lateral facies var- iations and surface erosion, the complete stratigraphy was recon- structed from a large number of individual outcrops. Simplified stratigraphic logs of 21 localities are shown in Fig. 2. The TBJ member consists of initial pumice lapilli-supported grain deposits and later of several units made of a coarse and fine ash, matrix-supported massive deposits with pumice lapilli and lithics interbedded with laminated levels of lapilli (i.e. ILO 18 and ILO22; Fig. 2). All these deposits were mapped across several dozens of km from the caldera rim. The medial succession can be observed up to 30–40 km from the caldera rim, where the best exposures are found on the southern slopes of IC (i.e. ILO 8 and ILO130; Fig. 2). The last unit, which comprises massive fine- grained deposits, is observed in medial exposures and distal ones that are N100 km from the caldera (i.e. ILO289 and ILO302; Fig. 2). Deposits from the TBJ eruption are characterized as being white soft and easily erodible, generating “badlands” type scarps (Šebesta, 2007). Most of the San Salvador Metropolitan Area (Fig. 1c) has been built on the TBJ tephra deposits. 4.1. Unit  A0     A0 is the first unit in the TBJ succession of deposits (stratigraphic log 22 in Fig. 2), which is observed in medial (10–40 km from the vent) out- crops mainly to the south of the caldera. Thickness ranges from 2 to 4 cm (Supplementary Material 1) and the deposits are characterized by poorly-sorted, thinly bedded or laminated, moist beds of rounded dense, glassy coarse and fine pumice ash with accidental lithic frag- ments. The deposit usually rests directly upon a paleosol or older, 83 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   85 1 '¡;J • , ! ! - . . • '¡.' : i -. -. o " • e J , ! J e h ~ • ~ ¡ , , so so 40 30 30 30 E? a En 10 l l Í l l 10 10 o o 33210123456789000) | "232101234507 0890510) | 432101234567 8910>10 SRAINSIZE (0 UNITS) GRAN SIZE (0 UNITS) SRAINSIZE (0 UNITS) 11018 (E) 1L022:5(E) 11.039: (F) 40 40 “0 30 20 30 Eso Ez Ez 5 5 s 10 10 10 o o o 33210123450 7080900)| "as210123450709050) | "432401234607 8.9105 RAN SIZE (0 UNITS) GRANSIZE (0 UNITS) GRAN SIZE (0 UNITS) 1LOt8-3(D) 1022-3815 (C) 11032: (F) 10 40 40 30 30 so E E E E Ezo Es 5 s 5 10 10 10 o 4321012345678 910210] o 4321012345678 910>0| o 4-3-2-1012345678-910>0| SRAIN SIZE (0 UNITS) SRAIN SIZE (0 UNITS) RAIN SIZE (0 UNITS) 1LO18-2(C) 1L022-3 (D) 1LO28-2 (F) 40 40 40 30 30 30 E z E pa En En 10 10 10 | | | | o o o 4321012345678 910>0] 43-2-10123 45.67 8 9100] 4321012345678 9100] RAIN SIZE (0 UNITS) GRAIN SIZE (0 UNITS) GRAIN SIZE (9 UNITS) 1LO184 (B) 1L022-28IS (8) 10281 (E) “0 40 40 30 30 30 E E E E En En 10. 1015 10 o o o 4323012345678 91010] 4-3-2-10123 45 6 7 8 910>0] 43-2-10123 456 7 8 910>0] GRAIN SIZE (0 UNITS) RAIN SIZE (9 UNITS) GRAN SIZE (9 UNITS) ILO8-2 (D) 1LO22-2 (A) 1LO22-8 (F) “0 7o 40 50 30 E so E ao E Ezo E Sao E E 30: s 40 20 10 10 o o o 432101234567 8 910>0] 4321012345878 9100] 432101234567 8 9100] GRAÍN SIZE (0 UNITS) RAIN SIZE (0 UNITS) GRAIN SIZE (0 UNITS) 1LO8=1(C) 1L022-1(A0) 1LO22-7 (F) 40 40 “0 30 30 2 Eso Eso Es S s 5 o 432-101234567 8 91010] 'GRAIN SIZE (0 UNITS) 0 432101234567 8 910>10] 'GRAIN SIZE (0 UNITS) o 43-2101234567 8 910210] 'GRAIN SIZE (0 UNITS) 84 86 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102     b ¡L01B_S (F) IL022.fi (F) IL038· ' (8) ~ ~ " ~ ~ ~ l~ :~ ~ I II , - ~~ , .111 I " 11 .llh, " 1. O O -4-3·2·10 I~J45~ I e 910>10 ... ·3·2· 0123456 78910>10 -4 -3 -2-10123456 78910>10 G AI SlZe (4) ITS) GRIIIN SlZE (4) HIT$) G l SIZE (4) ITSI IL018-4 ( ) ILQ22_S(E) ILOll_' ( ) " " ~ ~ ~ ~ l~ l l~ , 1 ,~ , " " .1 I "11.11 111 .... , •• ... ·3·2-10123 4 56 le91O>10 ... ·3·2·10123455 1891 >10 -4 -3-2 -1 1 2 3 4 5 678910>10 GRAlN SIZE (4) UNITS) IN SIZE 14> "IIT5) AlN sl e (<11 ITSl Il la_) (O) IlQ 2·381$ e) IL 2-' ( ) " " .. ~ ~ ~ • " ~~I ~! ~~ , I, ! ~ , -~ -~ , II.llb. , .hllfl " " ... 1 11 •. ... ·3·2·1 23456 7 8 9 10>10 -4-3·2 -' 23456 1891 >10 67 \1 91Q>1 G AIN lZ (01) ll" 1 GRfJN SIZE (4) UtlrTS) GRAlN IZE (<1> NITSI I 0 8·2 (e) Il 2-3 (O) I 028.2( ) " " '" ~ ~ ~ " " " -~ -~ -~ , , .l •.• JII , .,;!!, ' #~ = " " .... 1 111.111 •. " 11 •. " ... ·3·2·10 1 23 4 56 71191 10 ... ·3-.2·1(}123.5 18910>10 8 9 1 >10 GRAlN S~ {10 8910 1 "·3·2·1 01 2 3456 1 e 9 10>10 AA.IN IZ (4) NlfS) GAA.IN lZ (4) ITS) IN lZ (4) ITS) I 08·2 ( ) I 022·2 ( ) I 0 24( ) 10 76910>10 -4·3·2·1 0 12 345ti 7 a 9 10>10 AlN I (4) i S) G A.IN I E (4) ITS) RAlN IZ (4) lfS) I oa·, (e) I 2·' (AO) I on·, ( ) '" .. '" ~ ~ ~ ¡ , ¡ -~ ~ -~ , -~ , _11 1.1111. .JQ"ih:·, " " " .. o -4·3·2·101~3 4 56 7891O~10 -4·3·2·101~3 4 5 6 7 8 9 10>10 7 a 9 10>10 G l sw:: (4) iTS) G l sw:: (4) Ll ITS) GRAlN SIZE (4) LlNITS) 30: 30 40: E E 30 20 20 dE Ss 5 52 e 10; 10 Do o 0 - 01234567 8U910>10| 43-2-1012345678910>10] 4321012345678 6910>1 GRAIN SIZE (9 UNITS) GRAN SIZE (9 UNITS) GRAIN SIZE (0 UNITS) 1LO113:1 (6) ILO169 (F) 1LO172-5 (G) <0 40 30 30: 30: 30: > E E 10 10 10 0 0 o 4-3-210123456 7 8 910>10] 432101234567 86910>10] 4321012345678 0910>11 GRAIN SIZE (0 UNITS) GRAN SIZE (0 UNITS) GRAIN SIZE (0 UNITS) 1L.049-4 (F) 1LO165-3 (G) 1LO172-4 (F) 40 40 40 30: 30 30 E o E 10 10 10 il | | o o 4 012345678 910>10| 4-3-2-101234567868910>10] 4321012345678 910>1 RAIN SIZE (0 UNITS) GRAN SIZE (0 UNITS) GRAIN SIZE (0 UNITS) 1LO49-3 (D) ILO165-2 (F) 1L0172-3 (D) 40 40 40 30. 30 30 Eso Ez Ez E 5 $ 10 10 10 o o o 4324012345678 910>10] 4321012345678 6910>10] 4-3-2-101234 5678 910>1 'GRAIN SIZE (0 UNITS) GRAIN SIZE (0 UNITS) GRAIN SIZE (9 UNITS) 1LO49-2 (C) 1LO165-1 (D) 1L0172-2 (B) 40 40 40; 30 30 30. as 5 z Es Eso 2 5 s s 0 a3Z101234567 8 910%0] 0332101234567 8 910>0] : 4321012345678 9100 GRAIN SIZE (9 UNITS) GRAIN SIZE (0 UNITS) GRAIN SIZE (9 UNITS) j 1LO49-1 (B) 1L0130-3 (6) 1LO1724 (A) 40 “0 70 60: 30: 30 50. E z, Eso Ezo =20 E 5 5 50 10 10: 20 10 0 0 0 4321012345878 91010] 4321012345678 91010] 4321012345873 91020] GRAIN SIZE (0 UNITS) GRAIN SIZE (9 UNITS) GRAIN SIZE (9 UNITS) 1L046-1 (F) 1LO130-2 (F) 1LO169-C (F) 40 40 40 30: 30: 30: En on En o 4321012345678 910>10] 'GRAIN SIZE (0 UNITS) 321012345878 9100 GRAIN SIZE (0 UNITS) 0 4321012345678 910>10] 'GRAIN SIZE (0 UNITS) 85 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   87 e IL0130.1 (F) IL016M(G) IL0247-1 (A) " '" ~ '" '" " l ¡ i 30 ," -ro ~ 20 , " " " " -4-3·2·1 1 J 4!i~ 7 8 910>10 -4 -3 -2 - 23456 " 7 a 910>' -4-3·2 · 12345 6r~910>1 GRAlN SIZE (<1> UNITS) Af.l IZ€ (01) NrtS) GRAIN SIZE (<1> UNITSI I 0 13·1 (G) IL0169·1 ( ) I 0 72-5( ) ., ., ., '" '" '" ¿" ¡ ¿" -ro • , , " " " " " " -4 . .2.' -4-3 ·2-10123456 7 8 9 10'10 -4-3·2·10 345 67~9 1 0>1 AlN l (<1> ITS) AtI I E $ ITS) AJN I (IP ITSI I 0 -4 ( ) I I&5·3 ( ) I 0 72-4( ) " " " '" '" '" !20 • !2(I • ~ 20 • " " " " " " -4·3·2 _1 1 2 3 4 56 18 10'10 -4 ·3_ _ 0123 4 58 7 B 9 lO" -4·3· · 123 4 5 67$9 > 1 G AlN IZE (01) I13) IN I (01) ITSI AAIN I (10 "'-3·2,'OI234~ 6 7 10'10 -4-3-2-1 123 4 56 7 8 \1 10~10 R IIN I E " NITSl RA I E 41IJ:11ITS) RAlN IZE (41 I1'5) Il0 -1 ( ) Il 130-3 (G) Il01 -1 ( ) .. .. " "'" '" ~ i" <" i40 • ~ ~'" '" " " " , " , -4-3_2_1 0 1 2 34 56 7 /j 9 1 0~10 -4-3·2-101 2 345 8 789 10>10 -4-3· · 0 12 345 6739 10>10 AAlN lZE (" ITSI RAl)! lZ (41IJ:11ITS) RAlN lZ (41 ITS) I 6-1 ( ) I l 0·2 ( ) I 0 69-C( ) .. '" .. '" '" '" l" • é2Q ~ l2Q • '" " " " " " -4 _'. .1 o 1 2 3 4 5 6 169 10 -4·3-2- 0123 5 678910>10 -4- · · 12345 81 a 910>10 GA l l e " ITS) GRA>'1 5lZE (41IJ:11IT5) GRAlN s¡ze (41 UNITS) 30 30 Eso Ez s 5 10 10 ol o 432101234507881020) 432101234587 89100 RAIN SIZE (0 UNITS) GRAN SIZE (0 UNITS) 1L0286-1 (F) 1LO293-2 (F) “0 40 30 30 Eso En |S s 10 10 o o 4321012345678 910»10 432101234567 8 91090] SRAIN SIZE (8 UNITS) GRAN SIZE (0 UNITS) 1L0247-6 (F) 1L0293-1 (D) 40, “0 30 30 Ezo Eo s E 10 10 o o 4324012345678 910%] 4321012345678 910:0] GRAN SIZE (9 UNITS) GRAN SIZE (GUNS) 1LO247.5 (E) 1LO291-3(C) “0 40 30 30 zo Eso 5 E 0 4321012345678 91010] , o 432101234567 8 910>0| GRAIN SIZE (9 UNITS) GRAN SIZE (0 UNITS) 1L.0247-4 (D) 1LO291-2 (8) 40 40 30 30 E Es 20 10 10 0 o 232-101234567 8 9100] 4321012345567 8 9100] 'GRAIN SIZE (0 UNITS) GRAN SIZE (0 UNITS) 1.0247-3 (C) 1L.0291-1 (A) an Ae mn Al HISTOGRAM LEGEND so 30 [vo conooneney z E Es Es Mo s E 1 w "++ o 0 da2ro 1 3 aso rodun| Casio i231soroim mm GRAIN SIZE (9 UNITS) GRAIN SIZE (0 UNITS) 1L0247-2 (B) 1L.0289-4 (G) 1L.0302-1 (6) 40 “0 40; 30 30: 30 20 20 20! 0 4321012345678 91010] 'GRAIN SIZE (0 UNITS) E o 432101234567 8 91020 GRAN SIZE (9 UNITS) 4324012345678 910>10 RAIN SIZE (0 UNITS) 86 88 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102     d (L0289-' (G) IL0296-' (F) ~ ~ ~ ~ ,~ , l20 ,.. .. • • -4.J·~·IO '~34~C 1 e 910>1 "' - ~·~·10'l3~~Cl & 9 10>10 G AlN I ttJ NITS) GRAtI SIZE (4) UNITS) I 286·' ( ) I 0 3-2 ( ) .. .. ~ ~ • •-~ • -~ •.. .. • • ..01-3_2,,0123456 789, .1 -4-3·2,'O123~56 78 >,0 GRAON SllE (~ UNITS) GRAtI SIlE (O:> (.INflS) I 41-6( ) I 293·' (O) .. .. ~ ~ l;ro • 1 -4-3·2_'O ' 23~56 7891 >10 AlN lZ ttJ ITS) (; .Ul IZE (<11 ITSI IL0247-5 (E) ILOZ9I-3(C) .. .. ~ .. ;!20 , e«l ,.. .. • • -4-.3·2· 0123456 189 10>1 -4· ·2 ·,0123456 76 , 0>10 RAlN SIZE (41 ITS) AtI Il (O:> l I S) I 02 1-4 (O) IL O~I . 2(B) .. .. ~ ~ 'ro •-ro • •.. .. • • -4 -J -2 - 1 3 4 56 7 8 9 10>10 -4 -3 -2-1 01234567 8 910>10 G l I ttJ IT$) AN ll (
  • 10 ..01 -3 -2 -,01 4567 8 6 10>10 AIN Il N> ITS) RAN IZ (10 -4-3·2- I OI~34!>E 76 10'10 -4-3·2 -IO l ~3 4 56 7Sg l 0>IO AAN Stz (<11 "..I Sj GRAlN Sl (<11 NITSj 87 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   89 weathered pre-caldera lavas. At the outcrop scale, there are lateral var- iations in the thickness and number of beds, with pinch and swell struc- tures and locally erosive basal contacts (Fig. 3a,b). 4.2. Unit  A     Unit A (stratigraphic logs 22, 172, 247, 291 in Fig. 2) outcrops in dif- ferent points around IC, but mainly in the eastern and southern sectors at medial locations. It shows thicknesses from 3 to 14 cm (Supplemen- tary Material 1) and is characterized by massive well-sorted thin to me- dium coarse angular pumice ash beds (Fig. 3a,b) with ash-sized lithic fragments. A planar contact separates it from the underlying Unit A0. 4.3. Unit  B     Unit B (stratigraphic logs 18, 22, 38, 49, 172, 247, 291 in Fig. 2) is characterized by moderately sorted, massive thin beds of angular pum- ice lapilli and lithics with no ash (Fig. 3c). Thicknesses vary from 1 to ∼5 cm (Supplementary Material 1). This deposit shows sometimes yel- lowish colour due to the pigmentation and cementing of iron oxides by contact with the underlying paleosol. It appears in several outcrops at proximal and medial locations. 4.4. Unit  C     Unit C (stratigraphic logs 8, 18, 22, 49, 247 in Fig. 2) is only preserved at a few outcrops in proximal and medial locations. It has a peculiar grey-yellowish colour (Fig. 3a) and is a well-sorted, matrix-supported deposit with light stratification of pumice fragments with scattered ac- cretionary lapilli and hydrothermally altered lithics. Observed thick- nesses range from a few cm up to 10 m in some depressions (Supplementary Material 1). 4.5. Unit  D     Well-sorted, massive, lithic-poor ash rich deposit (Fig. 3d). Unit D outcrops at proximal and medial locations (stratigraphic log 8, 18, 22, 28, 38, 49, 172, 247, 291, 293 in Fig. 2). The intermediate and distal (N40 km from the caldera) facies of this unit are quite unconsolidated with a fine ash matrix and dispersed pumice juvenile fragments (Fig. 3ei) and with slight variations between one horizon and another. At proximal locations the deposits are more cemented with a coarse ash matrix and containing beds that show a strong enrichment of millimetric accretionary lapilli (Fig. 3eii). At some outcrops, the deposit shows planar stratification. The maximum measured thickness of the Unit D is about 8 m (Supplementary Material 1). 4.6. Unit  E     Unit E consists of doublets of thin to medium thick massive and lam- inated beds of rounded lapilli and coarse ash pumice (Fig. 3d,f). The unit outcrops at proximal and medial locations from the caldera (strati- graphic logs 8, 18, 22, 28, 49, 172, 247, 293 in Fig. 2). It represents a good stratigraphic marker of the TBJ eruption and to differentiate be- tween Units D and F (Fig. 3d). The massive deposits are light coloured and composed of unconsolidated thick ash with pumice thin lapilli and lithics. The laminated deposits constitute very fine, well-sorted ash, that is light brown and dark brown when wet. It is commonly quite consolidated and rich in glass fragments and crystals. Locally, these deposits show folding that is characteristic of soft sediments (Fig. 3g). The maximum measured thickness is 1 m (Supplementary Material 1). 4.7. Unit  F     Unit F is composed of chaotic, massive, poorly-sorted, non-welded, light-coloured to light beige (Fig. 3d) with thickness up to about 60–70 m thick (Supplementary Material 1). Unit F outcrops at both proximal and medial locations (stratigraphic logs 8, 18, 22, 28, 32, 33, 46, 49, 51, 130, 165, 169, 172, 247, 286, 293 in Fig. 2) and found up to 40 km from the caldera. To the north, the deposits extend away from the caldera for at least ∼35 km and outcrop close to Cerrón Grande (Fig. 2). To the west, deposits cover part of San Salvador Volcano (Fig. 2), reaching a maximum height of 930 m (1,740 m a.s.l.). Deposits were also found close to the Municipality of Colón (Fig. 2), where they achieve a distance of ∼40 km. Towards the southern part (Balsamo Cor- dillera; Fig. 2), deposits outcrop along the old channels of rivers and streams reaching distances of N30 km. East of IC, Unit F was recognized up to 30–35 km away, close to the San Vicente Volcano (Fig. 2). The de- posits in the proximal outcrops show a coarse ash matrix with abundant centimeter- and decimeter-sized pumice and lithic fragments (Fig. 3h,i). Visibly mingled pumice with dark to light grey bands within the white pumice are found in unit F at very proximal sites within the caldera, e.g. ILO-32 (Fig. 3h). The abundance of mingled clasts at this site is ~5–10% and the clasts range from around 5 to 20 cm in length. Some decimeter-sized lithic-rich beds are observed close to the cal- dera edge (Fig. 3j). Medial outcrops show the same massive, lithic-rich deposits with a fine ash matrix, and lithic and juvenile pumice up to few centimetres in size (Fig. 3k). Most of the outcrops show a lower layer with higher particle concentrations. Degassing pipes are seen in this unit at some outcrops (Fig. 3l). In some cases, Unit F is found directly above Unit D or with a reworked lower part (Fig. 3m). 4.8. Unit  G     It is an unconsolidated, massive, well-sorted, coarse to fine ash de- posit with millimeter-sized accretionary lapilli (Fig. 3n). In some out- crops, a slight stratification is observed, with a transitional contact with Unit F below. Deposits were described mainly at medial and distal outcrops (stratigraphic logs 22, 46, 49, 113, 130, 165, 169, 172, 289, 302 in Fig. 2) and found up to 100 km from the vent (Fig. 3o). Maximum measured thicknesses are ~6 m (Supplementary Material 1). 5. Physical paraTheters 5.1. Grain-­‐size  distribution     Data from Supplementary Material 3 was plotted in Supplementary Material 4 in order to show variation of TBJ grain-size at proximal (0–10 km), medial (10–40 km) and distal locations (N40 km) from IC. Data include Medium Diameter (MdΦ), Sorting (σΦ) and Skewness (αΦ) parameters (Supplementary Material 4a–f) as well as F1 [wt% b1 mm diameter (0Φ)] and F2 [wt% b1/16 mm diameter (4Φ)] (Supple- mentary Material 4 g-i). Granulometric data of the local distributions characterized up to phi 10 were used to reconstruct the Total Grain Size Distributions (Fig. 4). Figures j-ac of Supplementary Material 4 illustrate the grain-size dis- tribution of the TBJ samples depending on distance from the caldera. Both A0 and A samples show a bimodal trend. Conversely, samples from Unit A are characterized by a unimodal trend. No proximal and dis- tal samples were found for both Units A0 and A. Only two samples from Fig. 2. Stratigraphic logs of TBJ succession of deposits and their locations. Granulometric analysis and lithics content are shown as well. The TBJ eruption can be divided in 8 units from base to top; Units A0 to G. Stratigraphic logs are arranged from west (left) to east (right) and from south to north, and cover most of the El Salvador. Inset figure show the locations respectively of samples and outcrops of Figs. 2 and 3. 88 90 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102     Unit B were collected at proximal locations, and only one of the two samples shows a unimodal trend. Medial samples from Unit B seem no show a clear relationship between distance and grain size trend similarly to the only sample from a distal outcrop that only shows a slight shift to finer classes. Two samples from Unit C at proximal loca- tions show a polymodal trend, similarly to the ones at medial outcrops. Fig. 3. Field photographs of the TBJ units with views of details. a) Units A0-D resting on a paleosol, see the scraper for scale; b) features of Units A0 and A: the former is characterized by poorly-sorted thinly, laminated beds of rounded pumice lapilli and coarse ash, and the latter by lithic-rich, massive, well-sorted thin to medium coarse angular pumice ash beds; c) Unit B, it shows massive thin beds of angular pumice lapilli with no ash; d) Units D, E, F. Unit D is an ash rich deposit whilst Unit F is characterized by a coarse ash matrix with abundant centimeter- and decimeter-size pumice and lithic fragments. Unit E has laminated beds; e) photographs of Unit D showing ei) ash matrix and dispersed pumice juvenile fragments with slight variations between one horizon and another and eii) strong enrichment of millimeter-size accretionary lapilli. Photographs of the characteristics feature of Unit E; f) doublets of thin to medium thick massive and laminated beds lapilli and coarse ash pumice; g) soft-sediments deformation structures: folding; Unit F with h) coarse ash matrix with abundant centimeter-size and decimeter-size pumice and lithic fragments at proximal outcrops, in the inset figure a mingled pumice is shown as well i) chaotic massive poorly-sorted, non- welded, light-coloured deposits; j) some decimeter-size lithic-rich levels; k) at distal outcrop; l) degassing pipes; m) reworked (RW) lower part of Unit F; n) Unit G, unconsolidated massive ash deposits with dispersed accretionary lapilli (AL); o) distal outcrops of Unit G reach thicknesses of 40–50 cm at Tazumal Archaeological Site (Chalchuapa). Outcrops numbers are shown in yellow in the inset in Fig. 2. Yellow dotted lines divide different units of the TBJ Member. White dotted lines outline details of the field picture. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 89 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   91 Fig. 3 (continued). Only one sample was collected from distal outcrops showing a shift to- wards finer classes. However, unlike Unit C, samples from Unit D at proximal and medial outcrops have a clear polymodal trend. Therefore not a clear relationship between distance and grain size trend was ob- served for sample from this unit. Only one distal sample from Unit D in- dicates a shift towards finer classes similar to the samples from Unit C. Two proximal samples from Unit E show different tendencies with a unimodal trend but towards coarser and finer classes. The same is ob- served at medial distance. Only one distal outcrop from Unit E was found in the field. It shows a clear shift towards finer classes. Proximal and medial samples from Unit F show a polymodal trend with coarser classes being more representative. Distal samples from Unit F seem to show a slight bimodal trend without any substantial change in the granulometrical distribution. Only one sample from Unit G was col- lected at one proximal outcrop. Medial and distal samples from Unit G are characterized by a bimodal trend. 5.2. Componentry  analysis     Componentry of individual beds is presented in Fig. 2 and Supple- mentary Material 5. The modal proportions of juvenile pumice and acci- dental lithic fragments (mafic clasts and pre-TBJ eruption ignimbrites) 90 92 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102           Fig. 4. Total Grain Size Distributions of fallout units (A, B and G). For the sake of comparison TGSDs associated to the co-ignimbrite phase of the Campanian Ignimbrite are also reported (Marti et al., 2016). are given for each grain-size fraction (or class) until 2Φ  and their distri- bution among grain-size fractions, as well as units is not constant. Unit A0, which is only present at few scattered medial outcrops, has a lithic content of ~8–8.5%. The following Unit A shows variable values from ~10–11% up to ~22–23% at medial locations. Unit B, at medial locations, shows values between ~15 and ~19% up to 28%. At distal outcrops, lithics are ~12%. Unit C at proximal outcrops contains total lithic values of ~9%. Medial outcrops are characterized by lithic values of ~5–8.5%. Unit D shows a constant lithic content from proximal to distal outcrops with values ~1–4%. Unit E shows values comprised between ~8% and ~16% al- though several samples show a considerable decrease with only lithics of ~3%. Unit F at proximal outcrops shows values ~15% of lithics whilst at medial outcrops values are generally around 5–15%. Unit G is charac- terized by lithics values at medial and distal outcrops of ~1–3%. 5.3. Product  distribution  and  volume  of  the  different  eruptive  phases     The distribution of outcrops and the thickness data (reported in the Supplementary Material 1) from each unit is shown in Fig. 5. Combining these field observations and dispersal models for each phase, we esti- mate the corresponding mass of erupted material (in terms of DRE) and intensity (in terms of discharge rate). Concerning the fallout units, which includes Units A and B from sustained eruption columns, and G from a co-ignimbrite plume, we computed the tephra transport and sedimentation by solving an inverse problem (Pfeiffer et al., 2005; Costa et al., 2009) using the tephra dis- persal model Hazmap (Macedonio et al., 2005). The results are summa- rized in Table 1, where the Total Erupted Mass (TEM), the column height, maximum wind intensity, and other physical parameters are re- ported for the different units. Furthermore, for Unit A we estimated a TEM of ~3.5 × 1011 kg (i.e. 0.15 km3 DRE assuming a constant magma density of 2300 kg/m3), and an eruptive column height of ~29 km, cor- responding MER of b~108 kg/s (Bonadonna and Costa, 2013). TEM for Unit B is of ~2 × 1012 kg (i.e. 0.8 km3 DRE), with an eruptive column height of ~7 km, corresponding MER of ~105–106 kg/s (Bonadonna and Costa, 2013). For the fallout unit G from the co-ignimbrite column, we adopted a first order approach similar to Matthews et al. (2012). Results of the inverse problem for the co-ignimbrite phase suggest a TEM of ~4 × 1013 kg (i.e. 16 km3 DRE) with a co-ignimbrite plume that reached a height of ~49 km (corresponding to a MER of ~1010 kg/s). For the co-ignimbrite plume the source of ash is not “point source” but rise from all the surface of ignimbrite sheet, which can have a radius N 30–50 km (Costa et al., 2018). For this reason, the va- lidity of the tephra dispersal model, which assumes virtual sources along an eruption column, is not fully appropriate for points at distances smaller than 30–50 km and simulation results should be considered simply as model extrapolations. However, in our case most of the avail- able outcrops were at larger distances (see Supplementary Material 1). The individual grain-size distributions of the samples of each unit at several locations (Fig. 2 and Supplementary Material 2) were used to generate the TGSDs (Total Grain Size Distributions) reported in Fig. 4. These TGSDs were estimated using the Voronoi tessellation method of Bonadonna and Houghton (2005). For the sake of comparison, the vol- umes of Units A, B, and G were also assessed by adopting empirical inte- grations of the deposit thinning (Bonadonna and Costa, 2012). The dispersal of the different units as isopachs is shown in Fig. 5. From these maps, we can see that Units A (Fig. 5a) and B (Fig. 5b) were mainly dispersed to the west and west-south-west areas, respec- tively. In contrast, Unit G (Fig. 5g) was dispersed towards the south by weak winds. Taking into account that PDC of Unit F had a runout distance of ~50 km (Fig. 5f), from the results of Costa et al. (2018) we can estimate a MER of order of 1010 kg/s, which is consistent with the value estimated for the co-ignimbrite phase (Unit G) on the basis of the height of the co- ignimbrite plume (see Table 1). The volume of PDC Units C, D, and F were calculated using the Delaunay triangulation method (Macedonio and Pareschi, 1991), which is, as mentioned in the Methods Section, suitable for assessing the volume between geological horizons from irregularly-spaced data points. We obtained the following volume estimations: 1. ~0.7 km3 (i.e. ~0.5 km3 DRE) for Unit C; 2. ~5.0 km3 (i.e. ~3.3 km3 DRE) for Unit D; 3. ~0.5 km3 (i.e. ~0.3 km3 DRE) for Unit E; 4. ~14 km3 (i.e. ~9.1 km3 DRE) for Unit F. DRE volumes were calculated using an assumed deposit density of ~1500 kg/m3 (Quane and Russell, 2005) and a magma density of 2300 kg/m3. These volumes indicate that 30 km3 of magma was ejected during the TBJ eruption. 6. Petrography, geocheThistry and glass coThpositions of the TBJ deposits Pumice clasts from the TBJ units are moderately crystal-rich (up to 10–15%) and highly vesicular. Mineralogy assemblage consists of 70–75% euhedral to subhedral plagioclase (andesine and labradorite; Figs. 6a–d and Fig. 7a), about 20% of magnesio-hornblende (Figs. 6a,b, e,f and 7b), and 10 vol% of crystal content is made of pyroxene (Figs. 6g,h and 7c,d), Fe-Ti oxides and apatite. Plagioclase crystals often have sieve-textured cores and contain apatite inclusions, Fe-Ti ox- ides and clinopyroxene (Fig. 6a–d). The hornblende crystals (Fig. 6e, f) have pristine rims with abundant inclusions of apatite (Fig. 6a) and orthopyroxene. Whole-rock compositions of the TBJ pumices are dacitic to rhyolitic (Fig. 8a and Table 2), and glass compositions are typically rhyolitic with the exception of mingled pumices found in the upper sequence (Unit F; see above) that extend to basalt (Fig. 8a). The glass composi- tions were determined for individual shards using an electron Fig. 5. Distribution maps (in cm) of each unit of TBJ eruption: a) Unit A, b) Unit B, C) Unit C, d) Unit D, E) Unit E, F) Unit F, g) and gi) Unit G. Latitude and longitude are expressed in degreees. Thicknesses reported in Fig. c, d, e, f only refer to the main outcrops. 1 91 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   93 92 94 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102     Table 1 Summary of the physical parameters of the deposits from the TBJ eruption. Physical parameters Unit A0 Unit A Unit B Unit C Unit D Unit E Unit F Unit G Total TBJ Bulk Volume (km 3 ) <0.1 0.35 1.84 0.7 5 0.5 14 36.80 69.35 "+ caldera filling" Total Erupted Mass (kg) <0.1 3.5×10 11 2×10 12 1.2×10 12 8.2×10 12 0.7×10 12 2.3×10 13 4×10 13 7.5×10 13 "+ caldera filling" DRE volume (km 3 ) <0.1 0.15 0.8 0.5 3.3 0.3 9.1 16 30 "+ caldera filling" Mass Eruption Rate (kg/s) ~10 9 ~10 8 ~ 10 5 -10 6 ~10 9 ~10 9 ~10 9 ~10 10 ~10 10 - Runout PDC (km) 20 - - 25 25 20 50 - - Column Height (km) - 29 7 - - - - 49 - Magnitude - - - - - - - - 6.8 microprobe from samples through the entire succession of deposits, and from both proximal and distal sites. Excluding the rare mingled clasts in Unit F, other deposits display homogenous, rhyolitic major element compositions with SiO2 = 75.3–78.1 wt%, Al2O3 = 11.9–13.8 wt%, Total FeO = 0.99–1.53 wt%, MgO = 0.12–0.33 wt%, CaO = 0.9–1.6 wt %, NaO2 = 3.78–4.88 wt% and K2O = 2.38–3.37 wt% (n    = 239; Table 3; Fig. 8a–d). The darkest material within the mingled pumice is basaltic and ranges down to 48.63 wt% SiO2, 7.91 wt% Al2O3, 12.42 wt % Total FeO, 12.03 wt% MgO, and 15.02 wt% CaO (Table 3; Fig. 8a–d). These grey bands are heterogenous in composition and extend from the least evolved composition to SiO2 concentrations up to 68.5 wt%. The whole-rock XRF data plot between this dacitic composition and the dominant rhyolite (Fig. 8a–d). 7. Discussion The volume of material erupted during the TBJ eruption was ~58 km3 of bulk rock, equivalent to ~30 km3 DRE of magma and corre- sponding to a magnitude of 6.8 (Pyle, 2000) (Table 1). Eight units can be identified in the deposits that provide evidence for distinct eruptive styles. The sedimentological and lithological characteristics of these de- posits suggest that the TBJ eruption included phases associated with pure magmatic activity and those characterized by magma–water inter- action, which are also seen in older intra-caldera deposits (Mann et al., 2004; Suñe-Puchol et al., 2019a, 2019b). Paleosols separate the TBJ from previous eruption deposits at several outcrops (Fig. 2). The repose pe- riod before the TBJ was of a sufficient length for this pedogenesis to occur, and the caldera was probably quiescent for around 8 ka, i.e. since TB2 (Kutterolf et al., 2008a, 2008b). Unit A0 (b0.1 km3 total DRE volume - Table 1) represents the onset of the TBJ eruption. The field characteristics (Fig. 3a,b) and granulometric analysis (poorly sorted deposit, positive grain-size skew- ness values and a bimodal trend; Supplementary material 4b, e, j) suggest that this unit was deposited by dilute PDCs (Branney and Kokelaar, 2002; Dellino et al., 2004a, 2004b; Brand and White, 2007; Brand and Clarke, 2009). The high proportion of mafic lithic fragments is consistent with explosive excavation of the conduit and vent (Fig. 9a), as described in other studies e.g. Vesuvius, Italy (Barberi et al., 1989) and the AD1630 eruption of Furnas volcano, San Miguel, Azores (Cole et al., 1995). These surge clouds had a high momentum as they travelled at least up to 15–20 km from the vent. The deposits show similar field characteristics to the ones of the Layer LM1 from the Lower Member of the Neapolitan Yellow Tuff that represented the onset of the eruption (Orsi et al., 1992). Grain size and componentry (fine-grained deposits; Supplementary Material 4h and high mafic lithic content - Fig. 2), as well as ash deposits suggest that there was magma- water interaction (Self and Sparks, 1978; Barberi et al., 1989; Houghton and Schmincke, 1989; Houghton and Smith, 1993; Cole et al., 1995; Dellino and La Volpe, 1995; De Rita et al., 2002). The opening phases of volcanic eruptions present favourable conditions for magma-water interaction, similar to other case studies such as the Minoan, Santorini Island, Greece, AD79 Vesuvius, Italy (Cioni et al., 2000), Etna 122 BCE, Italy (Coltelli et al., 1998), and Tarawera AD1886, New Zealand (Houghton et al., 2004) eruptions. The explosive eruptions that formed Unit A (Fig. 9b) produced an eruptive column that rose to 29 km (Table 1) and it spread mainly west- wards in the proximal and medial area. Field evidence (Fig. 3a,b) and granulometric data (well-sorted deposit and a unimodal trend; Supple- mentary material 4b, e and k) of samples are consistent with a tephra fallout deposit (0.15 km3 total DRE volume - Table 1). Unit A was most likely hydromagmatic, due to the high lithic content (Fig. 2) and fine grain size at medial locations (Supplementary Material 4h) and a distribution mainly to the south of the caldera (Fig. 5a). Passing from di- lute PDCs of Unit A0 to fallout deposits of Unit A is probably related to changes in magma-water mass ratio, which has been observed at sev- eral historical hydromagmatic eruptions, e.g. Kilauea volcano, Hawaii, AD1790 (McPhie et al., 1990) or Capelinhos (1957–1958) in Faial, Azores (Cole et al., 2001). Concerning the first two phases (A0 and A), the magma-water mass ratio promoted a more or less high explosive efficiency, from wet PDCs and fallout deposits towards drier lapilli fall (Unit B), so the magmatic fragmentation became progressively more dominant. Then, the erup- tion entered a magmatic fall-dominated phase (Fig. 9c) that formed Unit B (Fig. 3c), which is characterized by highly vesiculated juvenile products released through a ~7-km-high column (Table 1) with a grain-supported deposit mainly oriented southwestwards from the source (Fig. 5b). This eruption phase produced a coarse, generally me- dium sorted (Supplementary Material 4a,b,d,e), pumice fall deposit with a 0.22 km3 total DRE volume (Table 1). General drier conditions can be related to any factors such as, for example, the variations in magma flux or availability of water in the system, or in some cases, some batches of magma can reach the surface without explosive inter- action with water, similarly to maar-diatreme eruptions (Valentine and White, 2012). Similar activity was observed for the C11 deposits of Caldeira Volcano, Faial Island, Azores (Pimentel et al., 2015). The eruption was characterized at the beginning by a series of hydromagmatic eruptions with fallout and PDCs deposits and a subse- quent more dominant magmatic fragmentation, due to the rapid draining of magma from the conduit, with the establishment of a sub- Plinian column. The increase in the dispersal area and grain size features in the deposits (Supplementary Material 4g, h, i, l, m, n) indicates steady growth of the eruption column. The column reached its climax without major fluctuations, as there are internal bedding features and the de- posits lack normal or inverse grading. This was probably facilitated by the gradual stabilization of the conduit walls associated with increasing vent diameter and magma discharge rate. 93 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   95 Fig. 6. Petrographic features of the TBJ eruption products: parallel and crossed polarized nichols: a) and b) mineralogy assemblage with euhedral to subhedral plagioclase and hornblende with apatite inclusions; c) and d) detailed picture of plagioclase with pyroxene and oxide inclusions; e) and f) euhedral hornblende; g) and h) subhedral pyroxene with apatite inclusions. Unit C (0.5 km3 total DRE volume - Table 1) represents an abrupt change in the eruption dynamics (Fig. 9d). This well-sorted (Supple- mentary Material 4a–c), massive, lithic-poor and ash-rich deposit (Sup- plementary Material 4d–f and g–i), with few dispersed pumice fragments and accretionary lapilli indicate deposition from PDCs (Fig. 3a) that flowed mainly to the south-east part of the IC (Fig. 5c). These dynamics were probably due to the shift of the vent location and a subsequent interaction of magma with external water that led to an enhanced magma fragmentation, as well as a greater explosivity of the eruption that contributed to the generation of fine ash (Supplementary Material 4o–q). The stratigraphic posi- tion of these hydromagmatic deposits immediately above the mag- matic deposits suggests a subsequent access of the lake water to the column of rising magma. However, we cannot discount the role of hydrothermal and groundwater in the hydromagmatic episode that lead to the emplacement of Unit C. The presence of 94 96 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102       Fig. 7. Microprobe data of a) feldspars (classification of Smith and Brown, 1988); b) amphiboles (classification of Leake et al., 1997); and c) sodium and d) calcium, magnesium, iron pyroxenes (classification of Morimoto, 1989) diagrams. hydrothermally altered lithic fragments suggests the occurrence of an extensive hydrothermal system within the caldera at the time of the eruption (Saxby et al., 2016). Unit D (3.3 km3 total DRE volume - Table 1) shows similar field char- acteristics (Fig. 3d,e) and granulometric data (Supplementary Material 4a–c and r–t) to the previous unit C (Fig. 3d), and suggest it was emplaced from PDCs of hydromagmatic origin (Fig. 9d). The hydrother- mally altered lithic fragments observed in Unit C are not recognized in the Unit D, so the ongoing magma-water interaction was most likely fuelled by surface water. A shallow lake seems to have been present in the IC at ≥43.670 ka years ago as proposed by Mann et al. (2004) al- though the last study of Suñe-Puchol et al. (2009a) suggests the pres- ence of a paleolake already at 1.5 Ma. As proposed by Aravena et al. (2018), natural aquifers appear unlikely to be sources of enough water to significantly affect the eruptive dynamics of an event with high mass discharge rate; conversely, evidence for magma-water interaction are probably related to the involvement of surface water or the injection of groundwater by high-magnitude collapse mechanisms. The same type of activity was also reported for Taal caldera lake, Philippines in 1991 (Delmelle and Bernard, 2000), the hydromagmatic eruption of Ki- lauea Volcano, Hawaii, in 1970 (Mastin, 1997), or the Nari Caldera at Ulleung Island, Korea (Kim et al., 2014). Changes from dry to wet condi- tions in such eruptions were also observed for the Askja 1875 eruption, Iceland (Sparks et al., 1981; Carey et al., 2010) and the AD232 Taupo eruption, New Zealand (Houghton et al., 2000). The absence of any fall deposits at the base of Units C and D rules out the possibility of a sustained eruptive column phase (Fig. 5d). During the course of the eruption, there was another change in the eruptive dynamics, with a switch to drier conditions (Fig. 9e). Unit E (0.3 km3 total DRE volume - Table 1) was deposited by alternation of di- lute PDCs and fallout, which is based on plane-parallel and low-angle cross laminations and grain-supported layers without traction struc- tures (Fig. 3f; Chough and Sohn, 1990; De Rosa et al., 1992; Dellino et al., 2004b; Solgevik et al., 2007), alternation of well and poorly sorted deposits (Supplementary Material 4 a–c) of ash and lapilli (Supplemen- tary Material 4 g–i), and a clear polymodal trend of the grain size distri- bution (Supplementary Material 4u–w). Soft sediment folding (Fig. 3g) might indicate that some of the layers were deposited wet as consequence of magma-water interaction, thus characterizing the whole unit as alternation of dry and wet deposits that were de- posited around the IC (Fig. 5e). At this time, due to structural faults that characterize IC, the magma might have had interaction with the almost empty Ilopango Lake after Unit D phase, thus allowing an intermittent magma-water interaction with the formation of short-lived columns and lateral blast. It is important to consider how, not only a change in the water- magma ratio might have led to the emplacement of fallout and PDCs de- posits, but also the scaled depth (ratio between depth of explosion and energy) can have huge effects on deposit characteristics, grain size and deposit morphology (see Taddeucci et al., 2013; Graettinger et al., 2014, 2015; Valentine et al., 2014, 2015; Sonder et al., 2015). As suggested in Graettinger et al. (2015), when scaled depth is constant, the crater fo- cuses the jet and results in decreasing overall volumes of coarse ejecta and the potential occurrence of fine-grained dilute density current de- posits. Progressively increasing scaled depth results in an overall de- crease in ejecta volume to the point where the explosion is confined and no ejecta are produced. A progressive decrease in scaled depth will result in an increase in ejecta volume and in the grain size of ejecta 95 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   97 Fig. 8. a) Plot of TBJ juvenile samples (i.e. pumice clasts) and mingled pumices of Unit F in the TAS (SiO2-Na2O+ K2O) classification diagram of Le Bas et al. (1986). Glass compositions from the entire composition succession of deposits, and from both proximal and distal sites: b) CaO vs FeO; c) and d) SiO2 vs Al2O3. deposits and low occurrence of fine-grained dilute density currents as the jet is larger than the previous crater and therefore does not exhibit significant focusing. The final phase (Fig. 9f) of the eruption was marked by a dramatic change in eruptive style with deposition of chaotic, massive, poorly- sorted (Supplementary Material 4a–c), non-welded dry thick PDC de- posits (Fig. 3h, i). The lag-breccia deposits of Unit F are observed only close to the caldera topographic edge (Fig. 3j). This might be related to the strong control exerted by the paleotopography on facies architec- ture as observed, for example, for the Abrigo Ignimbrite in Tenerife, Ca- nary Islands (Pittari et al., 2006) or the Acatlán ignimbrite, Mexico (Branney and Kokelaar, 1997). This is a lithic-rich ignimbrite that repre- sents continued clearing from fissure vents along the main bounding caldera faults (Fig. 9f). The sharp, erosive lower contact with underlying units, coarse, up to meter-sized lithic clasts and juveniles in a poorly sorted matrix (Fig. 3h–k), together with granulometric analyses (Sup- plementary Material 4g–i and x–z), suggest eruptive dynamics that were dominated by vigorous and prolonged pyroclastic fountaining that produced sustained quasi-steady PDCs, as the eruption waxed and stabilized. Both basal high-particle concentrations in the PDCs and the long runout distances were maintained because of the continuous supply of dense currents at the vent (Roche et al., 2016). These deposits formed an ignimbrite sheet, Unit F (9.3 km3 total DRE volume-Table 1) that reached the sea on southern sectors of the caldera and was wide- spread around IC (Fig. 5f). At this point, the increase in the magma erup- tion rate could have been produced by the start of the caldera collapse, which would have commenced the rapid evacuation of magma from the sub-caldera magma chamber, leading to a subsequent inefficient magma-water interaction during F eruptive phase. Similar mechanisms from wet to drier conditions were also observed during the Neapolitan Yellow Tuff eruption (Orsi et al., 1992). The mingled pumice clasts that extend to basaltic compositions are also found in deposits from this phase of the eruption suggesting that additional melts were erupted. Since these distinctive less evolved compositions are restricted to the clasts in the very proximal outcrops it implies that the erupted volume of this melt was incredibly small. It is quite common for additional melts to be erupted during caldera formation (cf. Smith et al., 2016). As for Units C and D, no fallout layers were recognized at the base of Unit F, thus, suggesting that an initial buoyant Plinian eruption column- building phase was not produced. This feature is similar to other ignim- brites such as Campanian (Marti et al., 2016) and Ora in Italy (Willcock et al., 2013), or Huichapan in Mexico (Pacheco-Hoyos et al., 2018). The occurrence of fines-poor elutriation pipes (Fig. 3l) indicates that follow- ing deposition, vigorous gas escape occurred elutriating fines. These pipes are interpreted as evidence of rapid emplacement involving parti- cle segregation and vigorous, post emplacement fluid (dusty gas) es- cape (Branney and Kokelaar, 2002), thus suggesting that at the time of deposition Unit F deposits were hot. Unit G (Fig. 3n) represents the final co-ignimbritic deposit of the TBJ eruption (Fig. 9g). Deposits were found at medial and distal locations that are N100 km from the caldera (Fig. 3o). This unit is made of moder- ately to poorly sorted (Supplementary Material 4a–c) ash (Supplemen- tary Material 4g–i) with a clear bimodality grain-size distribution trend (Supplementary Material 4aa–ac) that highlights the significance of ash aggregation processes in the transport and deposition. The absence of Plinian pumice fall deposits preceding the dense PDC deposits of TBJ is a typical characteristic of graben-type calderas as Ilopango (Aguirre-Díaz and Martí, 2015; Aguirre-Díaz et al., 2016, 2017; Saxby et al., 2016; Suñe-Puchol et al., 2019a) or fissure ignimbrite eruptions related to local/regional faults (Aguirre-Díaz and Labarthe- 98 tedraza et al. Joumal o Volcarology and Geothermal Research 377 (2019) 81-102 TableZ Whole rockanalyses of representative TB] samples. Sample 110322 1O125T TLOTOEA LO30ZT LORA [rasUnt — Fíbase) — Fíbase) Fíbase) 6 S ES Apulo — S.Amton Masahuat Oratorio LaUnin Santafiena [Distance — Proximal— Medal Media Dial Dista [Lattude — N13*42.504' N12*22.620' Ni3"48.202 N13"10:203' N13"24.080' longitude WB9*0S 305 WS9'02510 W8G"02301 WB7"54 421 Wes-24.580" [Major and minor elements (onde, we) [sio a73s 70305 08831 7O38 70081 no: 008 030 03% 020 025] [aos 14.693 10.828 tm0 taza 1:38] [Fez0s 3475 201 342 2481 2408] [uno 0122 0.108 A] 0:01 [go 1218 07 0985 0587 0.57] [eso 3453 2528 334 19 201 [aso 437 420 4205 300 354] lo. 2:08 2403 212 2021 2500] [e:0s Das 0.087, Dos 0.0 0.067] [rotal 00.975 90.008 E] lor 249 222 208 32 322| Frace siemens (6pm ls 14 15 e 20 20] ES 1 1 1 1 1 le 43 za a 20 31 3 o o o o al [sc 5 3 7 2 a ln o o o 0 al lv 40 z 5 20 2 or 3 3 3 3 2 [ca 5 4 4 3 al [nr 2 2 3 2 2 [cu a 5 e 14 ñ Zn a ss “ 40 4 [62 14 13 13 14 14 [ro 37 48 z ES ss] [se 208 242 204 191 219 Y 7 17 18 17 17 lr 144 140 140 130 140 [no 3 4 3 4 a| lo 2 2 2 2 2 [Sn 1 1 1 3 3 [so 1 1 1 1 1 [0 2 2 2 3 al [ba ser 1 ara 12m 1100] la 12 13 10 18 a [ce 24 2 19 2 20] er 3 2 3 3 a [na % 13 “ 2 xa [Sm 3 3 2 3 a EN 1 1 1 1 1 [ro o o o a al [a 3 3 2 3 a [oy 3 3 2 3 a lo 1 1 1 1 1 ler 2 2 z 2 2 [vo 3 2 2 2 2 la o o o 0 al lar 4 4 4 4 a| [ra o D o 0 al |w o 1 o 1 1 ln o o o a al [ro o 7 a a al [To 3 3 2 4 a| lu 1 2 1 2 2 Samples analysed by XRay Fluorescene in the Insituto de Geología (UNAM) by Patricia Girón. Coordínats in WGS34 system (zone 169). 101: Los ofigpiton Hemández,2003; Aguirre-Díaz et al, 2008). This is due to the significant control of tectonic stress on mass discharge rate (Costa et al., 2011; Costa and Martí, 2016), with graben-type calderas tending to generate large MER larger that are too high to sustain a Plinian column (see Costa etal, 2018). TheTB) deposits highlight that a single eruption can produce a com- plex sequence of enuption styles and depositional processes. The magni- tude of this eruption means that Mayan populations living in the region would have been considerably affected (Dull et al, 2001; Hernández, 2004; Hernández et al 2015). The human populations directly affected o z 6 ti Ta ra De NE ro ze s Nt ys a7 e0 E ] 6 05 1 Es po o — Co mo s: Pr oa Me rr os or Me es ao sr w e e we ra ac or an 10 03 0 00 11. 4 o 15 8 1 124 45 as o 30 00 002 005 020 ero s sa e oz va so 017 see s Ni p. oz na 02 a Es tn ge dI Me FA MI LI A FM De e no = Pr ol Pr oa , 12 se Ne em Nt s4 2s os uo 1. 005 017 42 28 00 017 Wie rse 0o7 ve yos zes We go s0 > we go sz s ws era sao s na aa ns 005 02 450 002 va Do Nr ra sa 7o vi ne rn a g or Es 0 o se om os os Y ES mE mi do E S mE md o ES em Em pd o Ut LaS Un L aS Ln L es s Ut . La sa 10 83 Do ir as 017 132 4 010 om 127 43 es o Nr aa oz a Wie eoa gos — w esr eac or mo sz zs om 019 122 4% 280 00 018 120 8 261 120 4 ora s 19 00 017 ase s mo s as mo ra s a p e 90 0 o 1z2 z Aco so) Me da py EPM A o l i n d i v i d u a l gla ss sha rds a cq ui re d at 15 kV an d 6 nA us in g a 10 um d ef oc us ed b eam . D ata a re no rm al iz ed to 100 % t o a cc ou nt fo r v ar ia bl e hy dr at io n and facilit ate co mp ar is on . NE s3 02 83 N1 3I O2 ES > N1 3I Oz 8 va g o4: os' Com saz a — Co al ve d mo nz a At a) Rep res ent ati ve gla ss ana lys es of t he T B) eru pti on uni t. Tab le 3 96 II ~ .. ot 1)t 1 « ok..,.,q a d CIn' ... M l ) Y p!r>. - il0-3'·' L - ' 28- ' ilO_'-'" "-0-30,_, -303_' ~ ~ ,-, ( .... ) ( ..... ) O O - ~. s """'" . ....... .., ~ -- s.. .. ..... -- p",--, - w_ - ~ - 13'42-"'>1' 13']2.02&' """,,'.5'O' W,.'02.30 " "r M. 1' "''''4."" ,, _ n ............... ( OO" . ..... ) ~, . .34 _'" ""."" 0 .45 70 .""' m , 0 .4"" _"'" .. " D.""7 .25 ~ , , ... " "_112!1 ....... ".7" ... "'" '-A H 75 ,_." 3 .'" HJ ' .4.' ~ .' 2 _' 0 0 .11' .- . 10' - '.21 ~ O_O" .~ o .• ,,, .""7 ~ Hr.3 '_528 1 1\ ,.. '"~ 4 .' " _203 4.' 05 .... ', '" 'P ' .' 88 ' _4" ' .122 2.62 ' ,., oA o.,,, • = O.' M •• .""7 ,~ ..... " "'.8 8 "." .... 03' ."' ,m H . '" ,. " . 2 T~_ ..... (pp ) , .. " • ,. ,. .. , , , , , • " " .. " "o • • • • • • , , , , , " • • • • • • .. " " ,. , , , , , , ~ , • • , • .. , , , , , " • • • .. .. • " ~ ~ .. "~ .. " " .. .. ~ " .. " .. .. • •• ,., ~ .. , '" • " " " " " " ... ... ... .. '" ~ , • , • • ~ , • , • , ~ , , , , , .. , , , , , " , • , , , .. ., 11\ ... m, .. - " " .. .. "" " , .. .. , " , , , , , • " " .. " "- , , • , , ' " , , , , , " • • • • • ~ , , , , , ~ , , , , , • , , , , , " , , , • , .. , , , • , '" • • • • • .. • • • • • " • • • • • w • , • , , " • • • • • .. • , • • • h , , , • • • , , , • • s-. I .. .. ~ ." x~ FhDn:~ 1Ir ri n e «>Iop.I p.¡o.M) ." ""';ci;r irón. ( .. :.-.ti ... "" GS84 ,>"", oe 61'). !DI: '-oli",i1i . em:ndez, 2 03; J\guir~..[)í.tz t aL. 2lXI:I). llis ue to tlle signifi ....t otrol F t o ic ll'"SS 00 as.<; ll.l.I'ge te QSt.1 t .11., 0 1; sta.100 .1rtl, 16), itil rnn--ty c.a ras ooi g 10 nerate e ER Larger tll.1t.1~ o Ili ll USLlin.1 l .1n r n $el' oSLl. t al, 018). e T" posit$ll l li llt il r. .1 le pti n .an r ..::e .1 m · l x sequl'oce of r ptí o styles.100 :lsiti oal Sl'"$. he agni· e f ilis pti o eans .lt ay.1n ul.1tions t lle ::m ould llave en r si erably .1 fect d ull t .11, 001; eroo.ooez, 04; He-oo. I'z t.11., 15). Ile Ilu an op.JI.1 ons l\"CtJy .1lfeclrd - ji ! L .. ~ 1 I . ~ H ~ 1 I ! 1 lB " 1 ! t~ H ! , , '1 ' s ... ~ ~ ¡ 1 I s., I ~ I I - ' I! L! i i~; ; ! ¡ 1!;::::'¡¡: l';t~.t $1!1!I ¡¡~"~"~l1~""" P~ ~ !: ~ H I 1; ~H ¡ ¡¡lH, 97 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   99 Fig. 9. Sketch (not to scale) illustrating the evolution of the TBJ eruption: a) rise of magma and interaction with a shallow aquifer or water lake and formation of the directional dilute PDCs that spread mainly southward (Unit A0); fallout phases represented by b) hydromagmatic Unit A and c) magmatic Unit B; d) PDCs of hydromagmatic origin, due to a renewed magma- water interaction, with formations of Unit C and Unit D; e) PDCs and fallout deposits from the transitional Unit E due to the alternation of dry and wet phases; f) main phase of the TBJ eruption with deposition of Unit F by dense PDCs associated to the caldera collapse; g) co-ignimbrite deposits. by the TBJ eruption would have been those living in the territory within 50 km of the IC. However, the indirect effects on social, economic, and political systems probably affected a much wider area of Mesoamerica (Dull et al., 2001). It has also been suggested that the sulphate peak, typ- ically associated with volcanic eruptions, in the both Greenland and Antarctic ice cores at 539–540 CE could be associated with the TBJ erup- tion (Sigl et al., 2015). These peaks are associated with the H2SO4 aero- sols that are injected into the high atmosphere during large volcanic eruptions, which increase the albedo and potentially produce a volcanic winter period (Robock, 2000). However, the date of the eruption has not been sufficiently resolved to establish if these sulphate peaks in the polar ice cores are in fact associated with the TBJ eruption as the 14C dates fall on a plateau in the radiocarbon calibration curve (e.g., Reimer et al., 2013), which results in an imprecise eruption range of AD270-AD400 (Lohse et al., 2018) to AD440-550 (Dull et al., 2010). The examination of this eruption sheds light on a number of impor- tant implications for hazard assessment when considered within the framework of the volcanism associated with IC and Country of El Salvador. The detailed study of the TBJ eruption together with the ones of Suñe-Puchol et al. (2019a, 2019b) about the older eruptions of IC, represent the first and necessary step towards improved volcanic hazard assessments for the region. These are essential to mitigate volca- nic risk for the large number of communities, including the City of San Salvador, that are expanding around this active volcano. 8. Conclusion In this study, we conducted a detailed stratigraphic and lithological study of the dacitic pumice Tierra Blanca Joven (TBJ) deposit. The TBJ is the last explosive eruption of Ilopango Caldera, representing a singu- lar eruptive episode and constitutes the last eruptive cycle of the Tierra 98 100 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102     Blanca sequence that starts with the TB4 eruption deposit. The TBJ erup- tion erupted ~58 km3 of bulk volume rock or ~30 km3 DRE of magma, corresponding to a 6.8 magnitude eruption. The eruption was characterized by eight phases (A0-G) with distinct eruptive styles without major pauses in between. The eruption started with dilute PDCs followed by two fallout phases that left only few cm of deposits, found mainly close to the IC. Subsequently, dense and dilute PDCs of hydromagmatic and magmatic origin filled the depressions near the Ilopango Lake. Deposits thicknesses are up to 70 m and reached dis- tances of at least 40 km from the vent, covering the area where the city of San Salvador is now located. Finally, coignimbritic ash deposits of the last stage of the eruption were found all over El Salvador with significant thicknesses, and also found dispersed into neighbouring countries. The TBJ was a cataclysmic event and is considered to be one of the largest Quaternary eruptions in Central America. TBJ eruptive products would have considerably affected the Mayan populations living in Sal- vadorian and nearby territories at that time. Consequently, long- and short-term hazard assessments for IC should take into account all possi- ble scenarios including those described for the TBJ eruption. Supplementary data to this article can be found online at https://doi. org/10.1016/j.jvolgeores.2019.03.006. AcknowledgeThents This study was financed by CONACYT-CB grant 240447 to GJAD and logistically supported by MARN-El Salvador and PNC-El Salvador. We thank Caterina Muñoz Torres, Academic Technician of CGEO and the students Karina Rodríguez García and Katia Jasso Torres for their help during the survey. This manuscript was greatly improved by comments and suggestions from the chief-editor Joan Martí and an anonymous reviewer. References Aguirre-Díaz, G.J., 2008. Types of collapse calderas. Collapse Calderas Workshop 19–25 October 2008, Querétaro, Mexico “Reconstructing the Evolution of Collapse Calderas: Magma Storage, Mobilization and Eruption”. IOP Conference Series: Earth Environ. Sci., 3 012021 (5pp). Aguirre-Díaz, G.J., Labarthe-Hernández, G., 2003. Fissure ignimbrites: Fissure-source ori- gin for voluminous ignimbrites of the Sierra Madre Occidental and its relationship with Basin and Range faulting. Geology 31 (9), 773–776. Aguirre-Díaz, G.J. and Martí, J., 2015. Graben calderas: examples from Mexico, Central America, and the Andes. 26th IUGG General Assembly, Prague, Czech Republic, June 22-July 2, 2015. Aguirre-Díaz, G.J., Labarthe-Hernández, G., Tristán-González, M., Nieto-Obregón, J., Gutiérrez-Palomares, I., 2008. The ignimbrite flare-up and graben calderas of the Si- erra Madre Occidental, Mexico. Dev. Volcanol. 10, 143–180. Aguirre-Díaz, G.J., Suñe-Puchol, I., Dávila-Harris, P., Pedrazzi, D., Hernández, W. and Gu- tiérrez, E., 2016. Volcanic history of the Ilopango caldera, Central American Volcanic Arc: Cities on Volcanoes 9, IAVCEI; Puerto Varas, Chile, (November 2016). Aguirre-Díaz, J.G., Suñe-Puchol, I., Davila-Harris, P., Pedrazzi, D., Hernandez, W. and Gutierrez, E., 2017 Volcanic history of the Ilopango caldera, Central American Volcanic Arc. 113th Annual Meeting, 2017 GSA Cordilleran Section, Hawaii, USA. Aravena, A., Vitturi, M., Cioni, R., Neri, A., 2018. Physical constraints for effective magma- water interaction along volcanic conduits during silicic explosive eruptions. Geology 46 (10), 867–870. Barberi, F., Cioni, R., Rosi, M., Santacroce, R., Sbrana, A., Vecci, R., 1989. Magmatic and phreatomagmatic phases in explosive eruptions of Vesuvius as deduced by grain- size and component analysis of the pyroclastic deposits. J. Volcanol. Geotherm. Res. 38 (3), 287–307. Barckhausen, U., Ranero Cesar, R., Huene, R., Cande Steven, C., Roeser Hans, A., 2001. Re- vised tectonic boundaries in the Cocos Plate off Costa Rica: Implications for the seg- mentation of the convergent margin and for plate tectonic models. J. Geophys. Res. Solid Earth 106 (B9), 19207–19220. Bernal, J.B., Lozano-Santacruz, R., 2005. Characterization of a new set of eight geochemical reference materials for XRF major and trace element analysis. Revista Mexicana de Ciencias Geológicas 22 (3), 329–344. Bonadonna, C., Costa, A., 2012. Estimating the volume of tephra deposits: a new simple strategy. Geology 40 (5), 415–418. Bonadonna, C., Costa, A., 2013. Plume height, volume, and classification of explosive vol- canic eruptions based on the Weibull function. Bull. Volcanol. 75 (8), 742. Bonadonna, C., Houghton, B.F., 2005. Total grain-size distribution and volume of tephra- fall deposits. Bull. Volcanol. 67 (5), 441–456. Brand, B.D., Clarke, A.B., 2009. The architecture, eruptive history, and evolution of the Table Rock complex, Oregon: from a Surtseyan to an energetic maar eruption. J. Volcanol. Geotherm. Res. 180 (2), 203–224. Brand, B.D., White, C.M., 2007. Origin and stratigraphy of phreatomagmatic deposits at the Pleistocene Sinker Butte Volcano, Western Snake River Plain, Idaho. J. Volcanol. Geotherm. Res. 160 (3), 319–339. Branney, M.J., Kokelaar, P., 1997. Giant bed from a sustained catastrophic density current flowing over topography: Acatlan ignimbrite, Mexico. Geology 25 (2), 115–118. Branney, M.J., Kokelaar, P., 2002. Pyroclastic Density Currents and the Sedimentation of Ignimbrites (Geological Society of London). Canora, C., Villamor, P., Martínez-Díaz, J.J., Berryman, K.R., Álvarez-Gómez, J.A., Capote, R., Hernández, W., 2012. Paleoseismic analysis of the San Vicente seg- ment of the El Salvador Fault Zone, EL Salvador, Central America. Geol. Acta 10, 103–123. Carey, R.J., Houghton, B.F., Thordarson, T., 2010. Tephra dispersal and eruption dynamics of wet and dry phases of the 1875 eruption of Askja Volcano, Iceland. Bull. Volcanol. 72 (3), 259–278. Carr, M.J., Patino, L.C. and Feigenson, M.D., 2007. Petrology and geochemistry of lavas. Central America: geology, resources and hazards. 1 565–577. Chough, S.K., Sohn, Y.K., 1990. Depositional mechanics and sequences of base surges, Songaksan tuff ring, Cheju Island, Korea. Sedimentology 37 (6), 1115–1135. Cioni, R., Gurioli, L., Sbrana, A., Vougioukalakis, G., 2000. Precursory phenomena and de- structive events related to the late Bronze Age Minoan (Thera, Greece) and AD 79 (Vesuvius, Italy) Plinian eruptions; inferences from the stratigraphy in the archaeo- logical areas. Geol. Soc. Lond., Spec. Publ. 171 (1), 123–141. Cole, P.D., Queiroz, G., Wallenstein, N., Gaspar, J.L., Duncan, A.M., Guest, J.E., 1995. An his- toric subplinian/phreatomagmatic eruption: the 1630 AD eruption of Furnas volcano, São Miguel, Azores. J. Volcanol. Geotherm. Res. 69 (1–2), 117–135. Cole, P.D., Guest, J.E., Duncan, A.M., Pacheco, J.-M., 2001. Capelinhos 1957–1958, Faial, Azores: deposits formed by an emergent surtseyan eruption. Bull. Volcanol. 63 (2–3), 204. Coltelli, M., Del Carlo, P., Vezzoli, L., 1998. Discovery of a Plinian basaltic eruption of Roman age at Etna volcano, Italy. Geology 26 (12), 1095–1098. Corti, G., Carminati, E., Mazzarini, F., Garcia, M.O., 2005. Active strike-slip faulting in El Salvador, Central America. Geology 33 (12), 989–992. Costa, A., Martí, J., 2016. Stress Field Control during Large Caldera-Forming Eruptions. Front. Earth Sci. 4 (92). Costa, A., Dell'Erba, F., Di Vito, M.A., Isaia, R., Macedonio, G., Orsi, G., Pfeiffer, T., 2009. Tephra fallout hazard assessment at the Campi Flegrei caldera (Italy). Bull. Volcanol. 71 (3), 259. Costa, A., Gottsmann, J., Melnik, O., Sparks, R.S.J., 2011. A stress-controlled mechanism for the intensity of very large magnitude explosive eruptions. Earth Planet. Sci. Lett. 310 (1–2), 161–166. Costa, A., Suzuki, J.Y., Koyaguchi, T., 2018. Understanding the plume dynamics of explo- sive super-eruptions. Nat. Commun. 9 (1), 654. Daugherty, H.E., 1969 Man-Induced Ecological Change in El Salvador. Ph.D. dissertation, University of California, Los Angeles. University Microfilms, Ann Arbor. De Rita, D., Giordano, G., Esposito, A., Fabbri, M., Rodani, S., 2002. Large volume phreatomagmatic ignimbrites from the Colli Albani volcano (Middle Pleistocene, Italy). J. Volcanol. Geotherm. Res. 118 (1–2), 77–98. De Rosa, R., Frazzetta, G., La Volpe, L., 1992. An approach for investigating the depositional mechanism of fine-grained surge deposits. The example of the dry surge deposits at “La Fossa di Vulcano”. J. Volcanol. Geotherm. Res. 51 (4), 305–321. Dellino, P., La Volpe, L., 1995. Fragmentation versus transportation mechanisms in the py- roclastic sequence of Monte Pilato-Rocche Rosse (Lipari, Italy). J. Volcanol. Geotherm. Res. 64 (3–4), 211–231. Dellino, P., Isaia, R., La Volpe, L., Orsi, G., 2004a. Interaction between particles transported by fallout and surge in the deposits of the Agnano–Monte Spina eruption (Campi Flegrei, Southern Italy). J. Volcanol. Geotherm. Res. 133 (1), 193–210. Dellino, P., Isaia, R., Veneruso, M., 2004b. Turbulent boundary layer shear flows as an ap- proximation of base surges at Campi Flegrei (Southern Italy). J. Volcanol. Geotherm. Res. 133 (1), 211–228. Delmelle, P., Bernard, A., 2000. Downstream composition changes of acidic volcanic wa- ters discharged into the Banyupahit stream, Ijen caldera, Indonesia. J. Volcanol. Geotherm. Res. 97 (1), 55–75. DeMets, C., 2001. A new estimate for present-day Cocos-Caribbean Plate motion: Implica- tions for slip along the central American Volcanic Arc. Geophys. Res. Lett. 28 (21), 4043–4046. DeMets, C., Jansma, P.E., Mattioli, G.S., Dixon, T.H., Farina, F., Bilham, R., Calais, E., Mann, P., 2000. GPS geodetic constraints on Caribbean-North America Plate Motion. Geophys. Res. Lett. 27 (3), 437–440. Denevan, W.M., 1992. Introduction. In: The Native Population of the Americas in 1492. 2nd (Ed.), Edited by William M. Denevan. University of Wisconsin Press, Madison, pp. xv–xli. Dixon, T.H., 1993. GPS measurement of relative motion of the Cocos and Caribbean Plates and strain accumulation across the Middle America Trench. Geophys. Res. Lett. 20 (20), 2167–2170. Dull, R.A., Southon, J.R., Sheets, P., 2001. Volcanism, ecology and culture: a reassessment of the Volcán Ilopango TBJ eruption in the southern Maya realm. Lat. Am. Antiq. 12 (1), 25–44. Dull, R., Southon, J., Kutterolf, S., Freundt, A., Wahl, D., Sheets, P., 2010. Did the Ilopango TBJ Eruption Cause the AD 536 Event. AGU Fall meeting, Abstracts. Funk, J., Mann, P., 2009. Cenozoic tectonics of the Nicaraguan depression, Nicaragua, and median Trough, El Salvador, based on seismic-reflection profiling and remote-sensing dataCenozoic tectonics of the Nicaraguan depression. GSA Bull. 121 (11−12), 1491–1521. 99 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102   101 Garibaldi, N., Tikoff, B., Hernández, W., 2016. Neotectonic deformation within an ex- tensional stepover in El Salvador magmatic arc, Central America: Implication for the interaction of arc magmatism and deformation. Tectonophysics 693, 327–339. Golombek, M.P., Carr, M.J., 1978. Tidal triggering of seismic and volcanic phenomena dur- ing the 1879–1880 eruption of Islas Quemadas volcano in El Salvador, Central America. J. Volcanol. Geotherm. Res. 3 (3–4), 299–307. Graettinger, A.H., Valentine, G.A., Sonder, I., Ross, P.S., White, J.D.L., Taddeucci, J., 2014. Maar-diatreme geometry and deposits: Subsurface blast experiments with variable explosion depth. Geochem. Geophys. Geosyst. 15 (3), 740–764. Graettinger, A.H., Valentine, G.A., Sonder, I., Ross, P.-S., White, J.L., 2015. Facies distribution of ejecta in analog tephra rings from experiments with single and multiple subsurface explosions. Bull. Volcanol. 77 (8), 66. Guzmán-Speziale, M., Valdés-González, C., Molina, E., Gómez, J.M., 2005. Seismic activity along the Central America volcanic arc: is it related to subduction of the Cocos plate? Tectonophysics 400 (1–4), 241–254. Hart, W.J.E., 1981. The Panchimalco tephra, El Salvador. New Brunswick, New Jersey, Rutgers University, Central America. M.S. thesis (101 p). Hart, W.J.E., Steen-McIntyre, V., 1983. Tierra Blanca Joven Tephra from the AD 260 erup- tion of Ilopango caldera. Archeology and Volcanism in Central America—The Zapotitán Valley of El Salvador, pp. 15–34. Hernández, E.W., 2004. Características geotécnicas y vulcanológicas de las tefras de Tierra Blanca Joven de Ilopango, El Salvador. Tesis de maestría, Univ. Politécnica de El Salvador, San Salvador (115 p). Hernández, E.W., Ferrés, D., Delgado-Granados, H., Pullinger, C. and Gutiérrez de Henríquez, E., 2010. The last 40 ka eruptive cycle of Ilopango caldera deposits: a settlement for the San Salvador Metropolitan Area (El Salvador). A. Abstract volume, International Conference Cities on Volcanoes, Fundación Canaria ITER, Tenerife: pp 56. Hernández, W., Delgado-Granados, H., Nieto, J., 2012. La erupción Tierra Blanca 4 (TB4), Caldera de Ilopango, El Salvador. Ministerio de Medio Ambiente y Recursos Naturales, Informe inédito (36 p). Hernández, W., Aguirre-Díaz, G. and Ayala, P., 2015. La erupción Tierra Blanca Joven y la diáspora de los Mayas. Museo Nacional de Antropología David J. Guzmán, V Congreso Centroamericano de Arqueología en El Salvador. José Erquicia y Shione Shibata editores. 238p. ISBN 978-99961-958-0-8: 227-237. Houghton, B.F., Schmincke, H.-U., 1989. Rothenberg scoria cone, East Eifel: a complex Strombolian and phreatomagmatic volcano. Bull. Volcanol. 52 (1), 28–48. Houghton, B.F., Smith, R.T., 1993. Recycling of magmatic clasts during explosive erup- tions: estimating the true juvenile content of phreatomagmatic volcanic deposits. Bull. Volcanol. 55 (6), 414–420. Houghton, B.F., Smith, R.T., Gilbert, J.S., 2000. Phreatoplinian eruptions. Encyclopedia of Volcanoes, pp. 513–525. Houghton, B.F., Wilson, C.J.N., Del Carlo, P., Coltelli, M., Sable, J.E., Carey, R.J., 2004. The in- fluence of conduit processes on changes in style of basaltic Plinian eruptions: Tarawera 1886 and Etna 122 BC. J. Volcanol. Geotherm. Res. 137 (1–3), 1–14. Hutton, D.H.W., Reavy, R.J., 1992. Strike-slip tectonics and granite petrogenesis. Tectonics 11 (5), 960–967. Jochum Klaus, P., Stoll, B., Herwig, K., Willbold, M., Hofmann, A.W., Amini, M., Aarburg, S., Abouchami, W., Hellebrand, E., Mocek, B., Raczek, I., Stracke, A., Alard, O., Bouman, C., Becker, S., Dücking, M., Brätz, H., Klemd, R., de Bruin, D., Canil, D., Cornell, D., de Hoog, C.J., Dalpé, C., Danyushevsky, L., Eisenhauer, A., Gao, Y., Snow, J.E., Groschopf, N., Günther, D., Latkoczy, C., Guillong, M., Hauri, E.H., Höfer, H.E., Lahaye, Y., Horz, K., Jacob, D.E., Kasemann, S.A., Kent, A.J.R., Ludwig, T., Zack, T., Mason, P.R.D., Meixner, A., Rosner, M., Misawa, K., Nash, B.P., Pfänder, J., Premo, W.R., Sun, W.D., Tiepolo, M., Vannucci, R., Vennemann, T., Wayne, D., Woodhead, J.D., 2006. MPI-DING refer- ence glasses for in situ microanalysis: New reference values for element concentra- tions and isotope ratios. Geochem. Geophys. Geosyst. 7 (2). Kim, G.B., Cronin, S.J., Yoon, W.S., Sohn, Y.K., 2014. Post 19 ka BP eruptive history of Ulleung Island, Korea, inferred from an intra-caldera pyroclastic sequence. Bull. Volcanol. 76 (4), 802. Kutterolf, S., Freundt, A., Perez, W., Mörz, T., Schacht, U., Wehrmann, H. and Schmincke, H.U., 2008a. Pacific offshore record of plinian arc volcanism in Central America: 1. Along-arc correlations. Geochemistry, Geophysics, Geosystems, 9(2). Kutterolf, S., Freundt, A. and Perez, W., 2008b. Pacific offshore record of plinian arc volca- nism in Central America: 2. Tephra volumes and erupted masses. Geochemistry, Geo- physics, Geosystems, 9(2). LaFemina, P., Dixon Timothy, H., Govers, R., Norabuena, E., Turner, H., Saballos, A., Mattioli, G., Protti, M., Strauch, W., 2009. Fore-arc motion and Cocos Ridge collision in Central America. Geochem. Geophys. Geosyst. 10 (5). Le Bas, M.J., Le Maitre, R.W., Streckeisen, A., Zanettin, B., Rocks, I.S.o.t.S.o.I., 1986. A chem- ical classification of volcanic rocks based on the total alkali-silica diagram. J. Petrol. 27 (3), 745–750. Leake, B.E., Woolley, A.R., Arps, C.E., Birch, W.D., Gilbert, M.C., Grice, J.D., Hawthorne, F.C., Kato, A., Kisch, H.J., Krivovichev, V.G., 1997. Report. Nomenclature of amphiboles: re- port of the subcommittee on amphiboles of the international mineralogical associa- tion commission on new minerals and mineral names. Mineral. Mag. 61 (2), 295– 321. Lexa, J., Sebesta, J., Chávez, J.A., Hernández, W., Pecskay, Z., 2011. Geology and volcanic evolution in the southern part of the San Salvador Metropolitan Area. J. Geosci. 56 (1), 106–140. Lohse, J.C., Hamilton, W.D., Brenner, M., Curtis, J., Inomata, T., Morgan, M., Cardona, K., Aoyama, K., Yonenobu, H., 2018. Late Holocene volcanic activity and environmental change in Highland Guatemala. Quat. Sci. Rev. 191, 378–392. Lovell, W.G. and Lutz, C.H., 1995 Demography and Empire. Dellplain Latin American Stud- ies, Number 33. Westview Press, Boulder. Macedonio, G., Pareschi, M.T., 1991. An algorithm for the triangulation of arbitrarily dis- tributed points: applications to volume estimate and terrain fitting. Comput. Geosci. 17 (7), 859–874. Macedonio, G., Costa, A., Longo, A., 2005. A computer model for volcanic ash fallout and assessment of subsequent hazard. Comput. Geosci. 31 (7), 837–845. Mann, P., 2007. Overview of the tectonic history of northern Central America. Geol. Soc. Am. Spec. Pap. 428, 1–19. Mann, C.P., Stix, J., Vallance, J.W., Richer, M., 2004. Subaqueous intracaldera volcanism, Ilopango Caldera, El Salvador, Central America. In: Rose, W.I., Bommer, J.J., López, D.L., Carr, M.J., Major, J.J. (Eds.), Natural Hazards in El Salvador: Boulder. Colorado, Geological Society of America Special Paper vol. 375, pp. 159–174. Marti, A., Folch, A., Costa, A., Engwell, S., 2016. Reconstructing the plinian and co- ignimbrite sources of large volcanic eruptions: a novel approach for the Campanian Ignimbrite. Sci. Rep. 6, 21220. Martí, J., Groppelli, G., Brum da Silveira, A., 2018. Volcanic stratigraphy: a review. J. Volcanol. Geotherm. Res. 357, 68–91. Martínez-Díaz, J.J., Álvarez-Gómez, J.A., Benito, B., Hernández, D., 2004. Triggering of de- structive earthquakes in El Salvador. Geology 32 (1), 65–68. Mastin, L.G., 1997. Evidence for water influx from a caldera lake during the explosive hydromagmatic eruption of 1790, Kilauea volcano, Hawaii. J. Geophys. Res. Solid Earth 102 (B9), 20093–20109. Mastin, L.G., Guffanti, M., Servranckx, R., Webley, P., Barsotti, S., Dean, K., Durant, A., Ewert, J.W., Neri, A., Rose, W.I., Schneider, D., Siebert, L., Stunder, B., Swanson, G., Tupper, A., Volentik, A., Waythomas, C.F., 2009. A multidisciplinary effort to assign re- alistic source parameters to models of volcanic ash-cloud transport and dispersion during eruptions. J. Volcanol. Geotherm. Res. 186 (1), 10–21. Matthews, N.E., Smith, V.C., Costa, A., Durant, A.J., Pyle, D.M., Pearce, N.J.G., 2012. Ultra- distal tephra deposits from super-eruptions: examples from Toba, Indonesia and Taupo Volcanic Zone, New Zealand. Quat. Int. 258, 54–79. McPhie, J., Walker, G.P.L., Christiansen, R.L., 1990. Phreatomagmatic and phreatic fall and surge deposits from explosions at Kilauea volcano, Hawaii, 1790 AD: Keanakakoi Ash Member. Bull. Volcanol. 52 (5), 334–354. Montero, W., Dewey, J., 1982. Shallow-focus seismicity, composite focal mechanism, and tectonics of the Valle Central of Costa Rica. Bull. Seismol. Soc. Am. 72 (5), 1611–1626. Morimoto, N., 1989. Nomenclature of pyroxenes. Mineral. J. 14 (5), 198–221. Orsi, G., D'Antonio, M., Vita, S.d., Gallo, G., 1992. The Neapolitan Yellow Tuff, a large- magnitude trachytic phreatoplinian eruption: eruptive dynamics, magma with- drawal and caldera collapse. J. Volcanol. Geotherm. Res. 53 (1), 275–287. Pacheco-Hoyos, J.G., Aguirre-Díaz, G.J., Dávila-Harris, P., 2018. Boiling-over dense pyro- clastic density currents during the formation of the ~100km3 Huichapan ignimbrite in Central Mexico: Stratigraphic and lithofacies analysis. J. Volcanol. Geotherm. Res. 349, 268–282. Pfeiffer, T., Costa, A., Macedonio, G., 2005. A model for the numerical simulation of tephra fall deposits. J. Volcanol. Geotherm. Res. 140 (4), 273–294. Pimentel, A., Pacheco, J., Self, S., 2015. The~1000-years BP explosive eruption of Caldeira Volcano (Faial, Azores): the first stage of incremental caldera formation. Bull. Volcanol. 77 (5), 42. Pittari, A., Cas, R.A.F., Edgar, C.J., Nichols, H.J., Wolff, J.A., Marti, J., 2006. The influence of palaeotopography on facies architecture and pyroclastic flow processes of a lithic- rich ignimbrite in a high gradient setting: the Abrigo Ignimbrite, Tenerife, Canary Islands. J. Volcanol. Geotherm. Res. 152 (3–4), 273–315. Protti, M., Guendel, F. and McNally, K., 1995. Correlation between the age of the subducting Cocos plate and the geometry of the Wadati-Benioff zone under Nicaragua and Costa Rica. Special Papers Geological Society of America: 309–309. Pyle, D.M., 2000. Sizes of volcanic eruptions. In the Encyclopaedia of Volcanoes, (Sigurds- son, H.; Houghton, B.; McNutt, SR; Rymer, H.; Stix, J., editors). Academic Press, London. Quane, S.L., Russell, J.K., 2005. Ranking welding intensity in pyroclastic deposits. Bull. Volcanol. 67 (2), 129–143. Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, C.E., Cheng, H., Edwards, R.L., Friedrich, M., 2013. IntCal13 and Marine13 radiocarbon age calibra- tion curves 0–50,000 years cal BP. Radiocarbon 55 (4), 1869–1887. Richer, M., Mann, C.P. and Stix, J., 2004. Mafic magma injection triggers eruption at Ilopango Caldera, El Salvador, Central America. Spec. Pap. 375 Nat. Hazards El Salvador 175–190. Robock, A., 2000. Volcanic eruptions and climate. Rev. Geophys. 38 (2), 191–219. Roche, O., Buesch, D.C., Valentine, G.A., 2016. Slow-moving and far-travelled dense pyro- clastic flows during the Peach Spring super-eruption. Nat. Commun. 7, 10890. Rolo, R., Bommer, J.J., Houghton, B.F., Vallance, J.W., Berdousis, P., Mavrommati, C. and Murphy, W., 2004. Geologic and engineering characterization of Tierra Blanca pyro- clastic ash deposits. Special Papers Geological Society of America: 55–68. Rose, W.I., Conway, F.M., Pullinger, C.R., Deino, A., McIntosh, W.C., 1999. An improved age framework for late Quaternary silicic eruptions in northern Central America. Bull. Volcanol. 61 (1), 106–120. Saxby, J., Gottsmann, J., Cashman, K., Gutiérrez, E., 2016. Magma storage in a strike-slip caldera. Nat. Commun. 7, 12295. Šebesta, J., 2007. Geomorfología del AMSS y su relación con los movimientos de ladera. Open file report, Czech Geological Survey Prague Oficina de Planificación del Área Metropolitana de San Salvador (OPAMSS): pp 1–40. Self, S., Sparks, R.S.J., 1978. Characteristics of widespread pyroclastic deposits formed by the interaction of silicic magma and water. Bull. Volcanol. 41 (3), 196. Siebert, L., Simkin, T., 2002. Volcanoes of the World: an Illustrated Catalogue of Holocene Volcanoes and their Eruptions. Smithsonian Institution, Global Volcanism Program Digital Information Series, GVP-3. http://www.volcano.si.edu/world/. 100 102 D.  Pedrazzi  et  al.  /  Journal  of  Volcanology  and  Geothermal  Research  377  (2019)  81–102     Sigl, M., Winstrup, M., McConnell, J.R., Welten, K.C., Plunkett, G., Ludlow, F., Büntgen, U., Caffee, M., Chellman, N., Dahl-Jensen, D., 2015. Timing and climate forcing of volcanic eruptions for the past 2,500 years. Nature 523 (7562), 543. Simkin, T., Siebert, L., 1994. Volcanoes of the. World2nd Ed. Geoscience Press, Tucson. Smith, J.V., Brown, W.L., 1988. Spectroscopy—IR, Raman, NMR, NQR, EPR, NGR (Mössbauer), XAS, EXAFS, ESCA, XPS, Feldspar Minerals. Springer 244–267. Smith, V.C., Isaia, R., Engwell, S.L., Albert, P.G., 2016. Tephra dispersal during the Campa- nian Ignimbrite (Italy) eruption: implications for ultra-distal ash transport during the large caldera-forming eruption. Bull. Volcanol. 78 (6), 45. Sofield, D., 2004. Eruptive history and volcanic hazards of Volcan San Salvador. Geol. Soc. Am. Spec. Pap. 375, 147–158. Sohn, Y.K., Chough, S.K., 1989. Depositional processes of the Suwolbong tuff ring, Cheju Island (Korea). Sedimentology 36 (5), 837–855. Solgevik, H., Mattsson, H.B., Hermelin, O., 2007. Growth of an emergent tuff cone: Frag- mentation and depositional processes recorded in the Capelas tuff cone, São Miguel, Azores. J. Volcanol. Geotherm. Res. 159 (1), 246–266. Sonder, I., Graettinger, A.H., Valentine, G.A., 2015. Scaling multiblast craters: General ap- proach and application to volcanic craters. J. Geophys. Res. Solid Earth 120 (9), 6141–6158. Sparks, R.S.J., Wilson, L., Sigurdsson, H., 1981. The pyroclastic deposits of the 1875 erup- tion of Askja, Iceland. Philosophical transactions of the Royal Society of London. Series A, Math. Phys. Sci. 299 (1447), 241–273. Suñe-Puchol, I., Aguirre-Díaz, G.J., Dávila-Harris, P., Miggins, D.P., Pedrazzi, D., Costa, A., Ortega-Obregón, C., Lacan, P., Hernández, W., Gutiérrez, E., 2019a. The Ilopango cal- dera complex, El Salvador: Origin and early ignimbrite-forming eruptions of a graben/pull-apart caldera structure. J. Volcanol. Geotherm. Res. 371, 1–19. Suñe-Puchol, I., Aguirre-Díaz, G.J., Pedrazzi, D., Dávila-Harris, P., Miggins, D.P., Costa, A., Ortega-Obregón, C., Lacan, P., Gutierrez, E., Hernández, W., 2019b. Stratigraphic revi- sion of the complete eruptive sequence and recurrence of large explosive eruptions. The Ilopango Caldera Complex. J. Volcanol. Geotherm. Res. 374, pp. 100–119 (El Salvador). Taddeucci, J., Valentine, G.A., Sonder, I., White, J.D.L., Ross, P.S., Scarlato, P., 2013. The effect of pre-existing craters on the initial development of explosive volcanic eruptions: an experimental investigation. Geophys. Res. Lett. 40 (3), 507–510. Turner Henry, L., LaFemina, P., Saballos, A., Mattioli Glen, S., Jansma Pamela, E., Dixon, T., 2007. Kinematics of the Nicaraguan forearc from GPS geodesy. Geophys. Res. Lett. 34 (2). Valentine, G.A., White, J.D.L., 2012. Revised conceptual model for maar-diatremes: Sub- surface processes, energetics, and eruptive products. Geology 40 (12), 1111–1114. Valentine, G.A., Graettinger, A.H., Sonder, I., 2014. Explosion depths for phreatomagmatic eruptions. Geophys. Res. Lett. 41 (9), 3045–3051. Valentine, G.A., Graettinger, A.H., Macorps, É., Ross, P.-S., White, J.D.L., Döhring, E., Sonder, I., 2015. Experiments with vertically and laterally migrating subsurface explosions with applications to the geology of phreatomagmatic and hydrothermal explosion craters and diatremes. Bull. Volcanol. 77 (3), 15. Vallance, J., Houghton, B., 1998. The AD 260 Eruption at Lake Ilopango. A complex explo- sive eruption through a caldera lake. National Science Foundation, Research Proposal, El Salvador. Weber, H.S., Wiesemann, G. and Wittekindt, H., 1974. Mapa Geológico de la República de El Salvador/Geologische Übersichtskarte der Republik El Salvador 1: 500,000 (after geological maps 1:100000 – 1967-74) Bundesanstalt für Geowissenschaften und Rohstoffe, Hannover; Bundesanstalt fu¨r Bodenforschung, (Hannover). White, J.D.L., Houghton, B.F., 2006. Primary volcaniclastic rocks. Geology 34 (8), 677–680. Wilkie, J.W. and Guadalupe Ortega, J., 1997 Statistical Abstracts of LatinAmerica. 33, the Regents of the University of California, Los Angeles. Willcock, M.A.W., Cas, R.A.F., Giordano, G., Morelli, C., 2013. The eruption, pyroclastic flow behaviour, and caldera in-filling processes of the extremely large volume (N1290 km3), intra- to extra-caldera, Permian Ora (Ignimbrite) Formation, Southern Alps, Italy. J. Volcanol. Geotherm. Res. 265, 102–126. Williams, H., Meyer-Abich, H., 1955. Volcanism in the Southern Part of El Salvador: With Particular Reference to the Collapse Basins of Lakes Coatepeque and Ilopango. Univer- sity of California Press. 101 6. Discusiones y trabajos futuros La caldera de Ilopango, considerada como poligenética, ha sufrido múltiples colapsos caldéricos a lo largo de su historia que han provocado grandes erupciones explosivas. Por ahora se han identificado 13 erupciones piroclásticas cuaternarias emitidas por esta caldera, las cuales están casi todas relacionadas a colapsos caldéricos totales o parciales. Todos estos materiales piroclásticos forman parte del Grupo Ilopango, introducido en este trabajo siguiendo la metodología actual de estratigrafía volcánica publicada por Martí et al. (2018). El Grupo Ilopango se subdivide en tres Formaciones, de la más antigua a la más reciente son: 1) Formación Comalapa (1.78 – 1.32 Ma); 2) Formación Altavista (918 – 257 ka), y 3) Formación Tierras Blancas (< 57 ka). Cada formación del Grupo Ilopango está constituida por varios miembros, formados por los depósitos asociados a diferentes erupciones explosivas. Estas erupciones han sido muy variadas, desde columnas eruptivas que han depositado espesas capas de pómez, a flujos piroclásticos densos y diluidos, tanto de origen magmático como hidromagmático. La Fm. Comalapa la constituyen las tres primeras grandes erupciones explosivas de la caldera de Ilopango, las cuales han generado las ignimbritas más voluminosas emitidas por esta caldera. Éstas son: 1) la Ignimbrita soldada de Olocuilta (OI, 1.785 ± 0.006 Ma), que cubrió un área de ~3,000 km2 con un volumen mínimo de 50 km3 DRE, pero que podría llegar a considerarse como el producto de una Supererupción (más de 350 km3 DRE; Miller y Wark, 2008) si tenemos en cuenta la ceniza coignimbrítica distal, el relleno caldérico, el material erosionado y los piroclastos que llegaron al océano. 2) La Ignimbrita Colima (CoI, 1.56 ± 0.01 Ma) y 3) la Ignimbrita Apopa (ApI~1.34 Ma), fueron de menor magnitud que la primera erupción, pero mucho mayores que la TBJ (> 40 km3), pudiendo llegar a superar el VEI 7 (100 km3 de tefra emitida). Estas dos últimas erupciones, marcan un cambio en los procesos eruptivos provocados por la interacción del magma con el agua del lago de Ilopango formado a raíz de los dos primeros colapsos de la caldera. La interacción hidrovolcánica genera un aumento en la explosividad de la erupción y la fragmentación de los piroclásticos, lo que produce depósitos mucho menos consolidados como los de la Colima y la Apopa, provocando una mayor susceptibilidad a la erosión y complicando una estimación más real de su volumen original. 102 Los depósitos de las siguientes seis erupciones explosivas forman parte de la Fm. Altavista. Sus miembros son: 1) la Ignimbrita Cojutepeque (CojI, 918.8 ± 17.4 ka), 2) la Ignimbrita Delgado (DeI, 830 ± 140 ka), 3) la Ignimbrita Manigua (ManI, 768.3 ± 49.4 ka), 4) la caída de pómez San Juan (SJF, 625.0 ± 75.1 ka), 5) la Ignimbrita Cortez (CorI, 553.0 ± 16.6 ka), y 6) la Ignimbrita Soyapango (SoI, 257 ± 33 ka). Estas tobas están muy mal preservadas y tienen muy poca continuidad lateral debido a la poca consolidación de los materiales piroclásticos de origen mayormente hidrovolcánico. Los depósitos de la Fm. Altavista solamente se exhiben en el flanco norte de la caldera, sobretodo en cortes profundos de carretera. Aun así, se ha podido hacer una estimación aproximada del volumen de estas tobas, las cuales tienen todas de 1-5 km3 DRE como mínimo, siendo la ignimbrita Delgado la mayor de todas, la cual podría llegar a superar los 40 km3 DRE si consideramos los depósitos erosionados. Los depósitos de las cuatro erupciones explosivas más recientes de la caldera de Ilopango (TB4, TB3, TB2 y TBJ), constituyen la Fm. Tierras Blancas (TB’s), siendo las únicas que se habían identificado y caracterizado antes de esta investigación. El estudio geocronológico realizado para fechar todas las otras erupciones explosivas más viejas que las TB’s (< 57 ka; Rose et al., 1999), ha permitido detallar una historia volcánica más completa de la caldera de Ilopango. El diagrama de la Fig. 13 ilustra las edades de estas erupciones explosivas antiguas. Como se aprecia en este cronograma, las erupciones de la Fm. Comalapa tienen un tiempo de recurrencia el doble de largo que las erupciones de la Fm. Altavista. Mientras las tres primeras Ignimbritas (OI, CoI y ApI; recuadro punteado azul de la Fig. 13), tienen hiatos de hasta 220 ka entre erupciones, las siguientes seis erupciones (CojI, DeI, ManI, SJF, CorI y SoI; recuadro punteado verde de la Fig. 13) tienen periodos de inactividad inter-eruptivos sobre los 100 ka. Mucho más corto es el periodo de retorno para las erupciones explosivas de la Fm. Tierras Blancas (recuadro punteado rojo de la Fig. 13). Además, destacan los periodos largos de inactividad que hay entre las diferentes formaciones, con más de 400 ka de quietud entre la Fm. Comalapa y la Altavista (elipsoide punteada naranja de la Fig.13), o los 200 ka entre la Altavista y las Tierras Blancas (elipsoide punteada amarilla de la Fig. 13). 103 Fig. 13: Cronograma de las erupciones explosivas de la caldera de Ilopango, desde la formación de la primera ignimbrita hace 1.785 Ma (OI), hasta las últimas tobas (TB’s). El cronograma resalta el largo período de inactividad entre las Formaciones Comalapa y Altavista (elipsoide de puntos naranjas), y también el largo hiato volcánica entre las Formaciones de Altavista y Tierras Blancas (elipsoide de puntos amarillos). El período de recurrencia dentro de las erupciones de la Formación Comalapa (~ 220 ka, rectángulo de puntos azules) es dos veces más largo que las erupciones de la Formación Altavista (~ 100 ka, rectángulo de puntos verdes), y mucho más alto que las erupciones de la Formación Tierras Blancas (~20ka). El elipsoide discontinuo fucsia representa un hiato volcánico >200 ka dentro de las erupciones de la Formación Altavista, entre las ignimbritas de Cortez y Soyapango (CorI y, SoI). Estos periodos de recurrencia similares para las erupciones de una misma formación, pero tan variables entre erupciones de diferentes formaciones, así como los largos intervalos de inactividad volcánica que hay entre formaciones, podrían ser causados por cambios en el régimen tectónico a lo largo de la Zona de Falla de El Salvador (ZFES), y más concretamente al origen y evolución del Pull-Apart de San Salvador. Parece ser que la génesis de la caldera de Ilopango estuvo estrechamente ligada a la tectónica regional de El Salvador, colapsando primero en una caldera tipo graben durante un periodo extensional en la ZFES, para posteriormente transformarse en una caldera tipo pull-apart al mismo tiempo que la ZFES evolucionaba hacia un régimen más transtensivo, el cual continúa hasta el presente, en donde la actividad volcánica sigue vinculada a las fallas regionales, tal y como lo demuestra la erupción de las Islas Quemadas en 1879, emplazadas justo después de un terremoto tectónico provocado por la falla San Vicente. En los últimos años varios autores han realizado trabajos sobre las Tierras Blancas, sobretodo en la TBJ, la última gran erupción explosiva de la caldera de 104 Ilopango. Durante esta investigación y en colaboración con otros participantes del proyecto Ilopango, se condujo un trabajo estratigráfico detallado y un análisis granulométrico del depósito piroclástico TBJ, a partir del cual se pudieron identificar ocho fases distintas de la erupción, caracterizadas por estilos y procesos eruptivos diversos, cada uno con sus propias características y sin pausas mayores de actividad entre ellas. La erupción TBJ inició con una explosión tipo “blast”, seguida de dos fases de caída de tefra producidas por columnas eruptivas. Posteriormente la erupción se desarrolló con una interacción del magma con el agua del lago Ilopango y la formación de flujos piroclásticos (PDCs) densos y diluidos del tipo hidrovolcánico, dando lugar al colapso caldérico y la consecuente evacuación rápida de la ignimbrita paroximal. Como última fase, la ceniza fina elutriada durante la erupción de los PDCs en forma de nube coignimbritica se deposita, distribuyendose a lo largo y ancho de todo el país, con alturas estimadas de hasta 45 km sobre la superficie terrestre. El depósito de la TBJ llegó a acumular hasta 70 m de espesor en las cercanias de la caldera y nuevos cálculos en desarrollo por el equipo del Proyecto Ilopango estiman un volumen mínimo de ~58 km3 de tefras expulsados durante esta erupción (~30 km3 DRE). En colaboración con los participantes del proyecto se elaboraron nuevos mapas de isopacas en base a la estratigrafía, distribución de los depósitos y datos granulométricos detallados, mediate los cuales se calcularon parámetros físicos como la altura de la columna, la duración de la erupción, y la tasa de emisión. La erupción TBJ fue cataclismica y está considerada la mayor erupción explosiva del Holoceno en Centroamérica. Todos los productos piroclásticos de la CI son calcoalcalinos ricos en sílice, de composición riolítica a dacítica, con contenido medio-alto en potasio, típico de los magmas producidos en zonas de subducción en márgenes continentales, como es el caso del Arco Volcánico Centroamericano. La mineralogía que presentan son mayormente plagioclasas y hornblendas, con presencia de piroxenos, con cuarzo para las unidades más riolíticas y con pequeñas cantidades de biotita en algunos depósitos. Todas las ignimbritas y caídas de pómez presentan minerales accesorios como los circones, apatitos o óxidos de Fe-Mg. Las unidades de CI no contienen ningún sanidino. 105 Trabajos futuros Todavía siguen en curso otros trabajos de este proyecto de investigación, como por ejemplo el estudio tefro-estratigráfico sobre los depósitos de las últimas erupciones explosivas de la caldera de Ilopango (las TB’s de la Fm. Tierras Blancas). Por ahora se ha realizado ya el mapeo geológico y el levantamiento estratigráfico en campo para caracterizar estas erupciones, y se han descrito con más detalle y elaborado estimaciones de magnitud y volumen para conocer el alcance de estas erupciones. Falta terminar el estudio geocronológico por medio de técnicas 238U/230Th y Ar39/Ar40, y en particular de radiocarbono en la unidad TBJ. De esta manera se podrían integrar estos nuevos datos más precisos con las edades de los paleoterremotos registrados a lo largo de la traza de la falla San Vicente (Canora et al., 2012), y hacer una reevaluación del peligro considerando esta relación vulcano-tectónica. La de San Vicente es una falla principal del Pull-Apart de San Salvador que afecta la caldera de Ilopango y que parece estar controlando la actividad actual de este volcán, es por eso que creemos que se debe considerar a estas dos estructuras (falla + caldera) como un único sistema natural. Otro trabajo específico es el que se está realizando sobre la TBJ, la erupción explosiva más reciente de la caldera y ocurrida hace tan solo ~1500 años (Dull et al., 2010). Una vez determinados los procesos eruptivos, la distribución y el volumen de sus productos, así como otros parámetros físicos de la erupción, el enfoque de la investigación se centra ahora en evaluar en el impacto que tuvo esa erupción sobre las poblaciones Mayas contemporáneas y en el medio ambiente. 106 7. Conclusiones • La caldera de Ilopango se originó hace 1.785 ± 0.006 Ma y ha tenido una formación cíclica, con múltiples colapsos caldéricos asociados a las estructuras tectónicas de la Zona de Falla de El Salvador. Primero colapsó en una caldera tipo graben, que evolucionó a una caldera tipo pull-apart, régimen que continua hasta el presente con un aumento paulatino de la componente lateral. • A lo largo de su historia, la caldera de Ilopango ha generado al menos 13 erupciones explosivas de gran magnitud, con un volumen mínimo por evento de 1-5 km3 DRE, y pudiendo llegar hasta los 350 km3 DRE para la primera ignimbrita (VEI ~ 8, parámetros de supererupción), la cual se soldó y ocupó un área de ~3,000 km2. Se propone una nueva nomenclatura estratigráfica donde todas las tobas de la caldera de Ilopango se han incluyen en el Grupo Ilopango, el cual comprende las formaciones Comalapa, Altavista y Tierras Blancas. • La Fm. Comalapa (1.78 – 1.34 Ma) está formada por los miembros Olocuilta, Colima y Apopa. La Fm. Altavista (918 – 257 ka) por los miembros Cojutepeque, Delgado, Manigua, San Juan, Cortez y Soyapango. La Fm. Tierras Blancas (<57ka) por lo miembros TB4, TB3, TB2 y TBJ. Las erupciones que formaron todos estos miembros fueron muy diversas: columnas eruptivas puramente magmáticas, flujos piroclásticos diluidos de origen hidromagmático (interacción con el lago de Ilopango) o PDC’s magmáticos eyectados sostenida y radialmente desde la caldera. • Los periodos de recurrencia para grandes erupciones explosivas en la caldera de Ilopango son variables, con ~220 ka durante las primeras tres erupciones (Fm. Comalapa), ~100 ka para las 6 erupciones intermedias (Fm. Altavista), y ~20 ka para las últimas 4 erupciones (Fm. Tierras Blancas). Entre las erupciones de las diferentes formaciones, hay intervalos de inactividad muy largos, hasta 400 ka entre la Fm. Comalapa y la Fm. Altavista, o 200 ka entre Fm. Altavista y Fm. Tierras Blancas. Esta distribución de la actividad volcánica en el tiempo, podría estar 107 determinada por cambios en el contexto tectónico regional, el cual controla el ascenso de magma a la superficie con las fallas profundas strike-slip y las cuencas pull-apart de la ZFES. • Los productos piroclásticos de la caldera de Ilopango son riolitas y riodacitas calcoalcalinas, con contenidos medio-alto en K, ricas en plagioclasa y hornblenda, con biotita ocasional y piroxeno. No contiene sanidino. Características químicas típicas de magmas de subducción en márgenes continentales, como es el caso del Arco Volcánico de Centroamérica. • La última erupción explosiva de la caldera de Ilopango generó un extenso depósito blanco de pómez y ceniza (la TBJ) que cubrió todo El Salvador y parte de los países vecinos de Guatemala, Honduras y Nicaragua. La TBJ fue una gran erupción hidrovolcánica con 8 fases diferentes (columnas eruptivas y flujos piroclásticos), con un volumen eyectado de hasta 30 km3 DRE, y dispersión de cenizas hasta los 45 km de altura que quedaron por años en la estratosfera pudiendo afectar el clima global terrestre. Con dataciones dendrocronológicas aún por terminar, se sabe que la TBJ ocurrió hace unos 1500 años, la cual afectó catastróficamente a las poblaciones Mayas de la zona y es considerada la mayor erupción del Holoceno en Centroamérica. Si hoy en día ocurriera una erupción similar a las TB’s, provocaría un desastre incalculable para la población metropolitana de San Salvador que con más de 3 millones de personas, es el núcleo urbano más habitado de Centroamérica. • La presente investigación sobre la caldera de Ilopango presentada en esta tesis de doctorado, ha dado como resultado la publicación de tres artículos publicados en revistas indexadas, los cuales representan en los Capítulos 3, 4 y 5. 108 8. Referencias Acocella, V., Korme, T., Salvini, F., Funiciello, R., 2002. Elliptic calderas in the Ethiopian Rift: the control of pre-existing structures. J. Volcanol. Geotherm. Res. 119, 189– 203. Acocella, V. 2007. Understanding caldera structure and development: an overview of analogue models compared to natural calderas. Earth-Sciences Reviews, 85, 125– 160. Acocella, V. 2008. Structural development of calderas: a synthesis from analogue experiments. In: Agostini, S., Corti, G., Doglioni, C., Carminati, E., Innocenti, F., Tonarini, S., Manetti, P., Di Vincenzo, G., Montanari, D., 2006. Tectonic and magmatic evolution of the active volcanic front in El Salvador: insight into the Berlín and Ahuachapán geothermal areas. Geothermics 35, 368–408. doi:10.1016/j.geothermics.2006.05.003 Aguirre-Díaz, G.J., Labarthe-Hernández, G., 2003. Fissure ignimbrites: fissure source origin for voluminous ignimbrites of the Sierra Madre Occidental and its relationship with Basin and Range faulting. Geology, 31, 773–776. Aguirre-Díaz, G.J., Labarthe-Hernández, G., Tristán-González, M., Nieto-Obregón, J., Gutiérrez- Palomares, I., 2007. Graben-calderas. Volcano-tectonic explosive collapse structures of the Sierra Madre Occidental, Mexico. European Geosciences Union Annual Meeting at Viena, Geophysical Research Abstracts, 9, 04704. Aguirre-Díaz, G.J., Labarthe-Hernández, G., Tristán-González, M., Nieto-Obregón, J., Gutiérrez-Palomares, I., 2008. Ignimbrite Flare-up and graben-calderas of the Sierra Madre Occidental, Mexico: in J. Gottsmann, and J. Martí, eds., “Caldera Volcanism: Analysis, Modelling and Response”, Developments in Volcanology 10, Elsevier, Amsterdam, 492 p., (143-180 pp.), ISBN 978-0-444-53165-0, doi: 10.1016/S1871-644X(07)00004-6. Aguirre-Díaz, J.G., Suñe-Puchol, I., Davila-Harris, P., Pedrazzi, D., Hernandez, W. and Gutierrez, E., 2017. Volcanic history of the Ilopango caldera, Central American 109 Volcanic Arc. 113th Annual Meeting, 2017 GSA Cordilleran Section, Hawaii, USA. Alonso-Henar, J., Schreurs, G., Martinez-Díaz, J.J., Álvarez-Gõmez, J.A., Villamor, P., 2015. Neotectonic development of the El Salvador Fault Zone and implications for deformation in the Central America Volcanic Arc: Insights from 4-D analog modeling experiments. Tectonics 34, 133–151. doi:10.1002/2014TC003723 Alonso-Henar, J., Álvarez-Gõmez, J.A. and Martinez-Díaz, J.J., 2017. Neogene- quaternary evolution from transpressional to transtesional tectonics in Northern Central America controlled by cocos: Caribbean subduction coupling change. J Iber Geol. DOI 10.1006/s415113-017-0034-2. Alvarado, D., DeMets, C., Tikoff, B., Hernandez, D., Wawrzyniec, T.F., Pullinger, C., Mattioli, G., Turner, H.L., Rodriguez, M., Correa-Mora, F., 2011. Forearc motion and deformation between El Salvador and Nicaragua: GPS, seismic, structural, and paleomagnetic observations. Lithosphere 3, 3–21. doi:10.1130/L108.1 Alvarenga, E.R., Hernández, D.A., Hernández-Flores, D.A., 2001, Cronología de sismos destructivos en El Salvador: Servicio Geológico Nacional, Área de Sismología. Arqueología Mexicana, 2010, Culturas prehispánicas de México, Edición Especial 34, pp. 44-49 Authemayou, C., Brocard, G., Teyssier, C., Simon-Labric, T., Guttiérrez, A., Chiquín, E.N., Morn, S., 2011. The Caribbean-North America-Cocos Triple Junction and the dynamics of the Polochic-Motagua fault systems: Pull-up and zipper models. Tectonics 30, 1–23. doi:10.1029/2010TC002814 Bacon, C.R., 1983. Eruptive history of Mount Mazama and Crater Lake Caldera, Cascade Range, USA. J. Volcanol. Geotherm.Res. 18, 57–115. Beresford, S.W., Cole, J.W., 2000. Kaingaroa Ignimbrite, Taupo Volcanic Zone, New Zealand: evidence for asymmetric caldera subsidence of the Reporoa Caldera. N.Z. J. Geol. Geophys 43, 471– 481. Bernal, J.P. and Lozano-Santacruz, R., 2005. Characterization of a new set of eight geochemical reference materials for XRF major and trace element analysis. Revista Mexicana de Ciencias Geológicas, 22(3): 329-344. 110 Bernal, J.P., Solari, L.A., Gómez-Tuena, A., Ortega-Obregón, C., Mori, L., Vega- González, M., Espinosa-Arbeláez, D.G., 2014. In-situ230Th/U dating of Quaternary zircons using LA-MCICPMS. Quat. Geochronol. 23, 46–55. doi:10.1016/j.quageo.2014.06.003 Bibby, H.M., Caldwell, T.G., Davey, F.J., Webb, T.H., 1995. Geophysical evidence on the structure of the Taupo Volcanic Zone and its hydrothermal circulation. J. Volcanol. Geotherm. Res. 68, 29–58. Bonadonna, C. and Costa, A., 2012. Estimating the volume of tephra deposits: A new simple strategy. Geology, 40(5): 415-418. Bonadonna, C. and Costa, A., 2013. Plume height, volume, and classification of explosive volcanic eruptions based on the Weibull function. Bulletin of Volcanology, 75(8): 742. Branney MJ, Kokelaar P (1994) Volcanotectonic faulting, softstate deformation, and rheomorphism of tuffs during development of a piecemeal caldera, English Lake District. Geol Soc Am Bull 106: 507–530 Canora, C., Villamor, P., Martínez-Díaz, J.J., Berryman, K.R., Álvarez-Gómez, J.A., Capote, R. y Hernández, W., 2012, Paleoseismic analysis of the San Vicente segment of the El Salvador Fault Zone, EL Salvador, Central America. Geologica Acta, Vol. 10, Nº 2, DOI: 10. 1344/105.000001700 Canora, C., Martínez-Díaz, J.J., Villamor, P., Staller, A., Berryman, K., Álvarez- Gómez, J.A., Capote, R., Diaz,M., 2014. Structural evolution of the El Salvador Fault Zone: an evolving fault system within a volcanic arc. J. Iber. Geol. 40, 471– 488. Carr, M.J., 1976. Underthrusting and Quaternary faulting in northen Central America. Geol. Soc. Am. Bull. 87, 825–829. doi:10.1130/0016- 7606(1976)87<825:UAQFIN>2.0.CO;2 Carr, M., Patiño, L., Feingenson, M., 2007. Chapter 22: Petrology and geochemistry of lavas. Cent. Am. Geol. Resour. hazards 1–26. Cas, R.A.F. and Wright, J.V., 1987, Volcanic Successions. Modern and ancient. Allen & Unwin, London, 528 p. CEL (Comision Ejecutiva Hidroelectrica del Rio Lempa), 1992. Desarrollo de los 111 Recursos Geotermicos del Area Centro-Occidental de El Salvador. Prefactibilidad Geotermica del Area de Coatepeque. Reconocimiento Geotermico. Informe Final. Internal report. Cole, J.W., Milner, D.M., and Spinks, K.D., 2005. Calderas and caldera structures: a review. Earth Science Reviews, v. 69, p. 1–96. Corti, G., Carminati, E., Mazzarini, F., Garcia, M.O., 2005. Active strike-slip faulting in El Salvador, Central America. Geology, 33, 989-992. Costa, A., Dell’Erba, F., Di Vito, M.A., Isaia, R., Macedonio, G., Orsi, G. and Pfeiffer, T., 2009. Tephra fallout hazard assessment at the Campi Flegrei caldera (Italy). Bulletin of Volcanology, 71(3): 259. Costa A., Smith V., Macedonio G., Matthews N., 2014. The magnitude and impact of the Youngest Toba Tuff super-eruption, Front. Earth Sci., 2, 16, doi:10.3389/feart.2014.00016. DeMets, C., 2001. A new estimate for present-day Cocos-Caribbean plate motion:Implications for slip along the Central American volcanic arc. Geophys. Res. Lett. 28, 4043–4046. doi:10.1029/2001GL013518 de Silva, S.L., and Gosnold, W.A., 2007. Episodic construction of batholiths: Insights from the spatiotemporal development of an ignimbrite flare-up: Journal of Volcanology and Geothermal Research, v. 167, p. 320–335, doi: 10.1016/j.jvolgeores.2007.07.015 Dixon, T.H., 1993, GPS measurement of relative motion of the Cocos and Caribbean plates and strain accu- mulation across the Middle America Trench: Geo- physical Research Letters, v. 20, p. 2167–2170, doi: 10.1029/93GL02415. Donnelly T.W., Horne G.S., Finch R.C., Lopez-Ramos E., 1990. Northern Central America: the Maya and Chortis blacks. In: Dengo G, Case JE (eds) The geology of North America: the Caribbean region, pp :37–76 Dull, R.A., Southon, J.R. and Sheets, P., 2001. Volcanism, Ecology and Culture: A Reassessment of the Voclán Ilopango TBJ eruption in the Southern Maya Realm, v. 12, No. 1, pp. 25-44. Dull, R.A., Southon, J.R., Kutterolf, S., Freundt, a., Wahl, D., Sheets, P., 2010. Did the 112 TBJ Ilopango eruption cause the AD 536 event? AGU Fall Meet. Abstr. 1, 2370. Druitt, T.H. and Sparks, R.S.J., 1984, On the formation of calderas during ignimbrite eruptions. Nature, v. 310, p. 679–681. Dvorak, J.J., Dzurisin, D., 1997. Volcano geodesy: the search for magma reservoirs and the formation of eruptive vents. Rev. Geophys. 35, 343–384. Francis, P.W., O’Callaghan, L., Kretzschmar, G.A., Thorpe, R.S., Sparks, R.S.J., Page, R.N., 1983. The Cerro Galan ignimbrite. Nature 301, 51– 53. Funk, J., Mann, P., McIntosh, K., Stephens, J., 2009. Cenozoic tectonics of the Nicaraguan depression, Nicaragua, and Median Trough, El salvador, based on seismic-reflection profiling and remote-sensing data. Bull. Geol. Soc. Am. 121, 1491–1521. doi:10.1130/B26428.1 Garibaldi, N., Tikoff, B., Hernández, W., 2016. Neotectonic deformation within an extensional stepover in El Salvador magmatic arc, Central America: Implication for the interaction of arc magmatism and deformation. Tectonophysics 693, 327– 339. doi:10.1016/j.tecto.2016.05.015 Geyer, A. & Martí, J. 2008. The new worldwide collapse caldera database (CCDB): a tool for studying and understanding caldera processes. Journal of Volcanology and Geothermal Research, 175, 334–354. Golombek, M.P., Carr, M.J., 1978. Tidal triggering of seismic and volcanic phenomena during the 1879-1880 eruption of Islas Quemadas volcano in El Salvador, Central America. J. Volcanol. Geotherm. Res. 3, 299–307. doi:10.1016/0377- 0273(78)90040-9 Gottsmann, J. and Martí, J., 2008, Caldera Volcanism: Analysis, Modelling and Response, Developments in Volcanology 10, Elsevier, Amsterdam, 492 p. Grube, N. (ed.) 2006, Los mayas. Una civilización milenaria: Tandem Verlag GmbH: Ullmann & Könemann. ISBN 978-3-8331-1959-0. Guillou - Frottier, L., Burov, E. B. & Milé SI, J.-P. 2000. Genetic links between ash- flow calderas and associated ore deposits as revealed by large scale thermo- mechanical modeling. Journal of Volcanology and Geothermal Research, 102, 339–361 Guzmán-Speziale, M., Valdés-González, C., Molina, E., Gómez, J.M., 2005. Seismic 113 activity along the Central America volcanic arc: Is it related to subduction of the Cocos plate? Tectonophysics 400, 241–254. doi:10.1016/j.tecto.2005.03.006 Hansen, J., Sato, M., Ruedy, R. y 27 otros, 1996. A Pinatubo climate modelling investigation. In: Fiocco G, Fua D, Visconti G (eds) The Mount Pinatubo Eruption: Effects on the Atmosphere and Climate. NATO Advanced Study Institute Series 14, Springer-Verlag Heidelberg, pp 233-272 Hallinan, S., 1993. Nonchaotic collapse at funnel calderas: gravity study of the ring fractures at Guayabo caldera, Costa Rica. Geology 21,367–370. Hart, W.J., and Steen McIntyre, V., (1983). Tierra Blanca Joven Tephra from the AD 260 Eruption of Ilopango Caldera: in Archeology and Volcanism in Central America. University of Texas Press, Austin, Texas, p. 14-34. Heiken, G., Goff, F., Stix, J., Tamanyu, S., Shafiqullah, M., Garcia, S., 1986. Intracaldera volcanic activity, Toledo Caldera and embayment, Jemez Mountains, New Mexico. J. Geophys. Res. 91B, 1799– 1815. Hernández, E.W., 2004. Características geotécnicas y vulcanológicas de las tefras de Tierra Blanca Joven de Ilopango, El Salvador: Tesis de maestría, Univ. Politécnica de El Salvador, San Salvador, 115 p. Hildreth, W., Christiansen, R.L., O’Neil, J.R., 1984. Catastrophic isotopic modification of rhyolitic magma at times of caldera subsidence, Yellowstone plateau volcanic field. J. Geophys. Res. 89B, 8339–8369. Holohan, E.P., Van Wyk de Vries, B., Troll, V.R., 2008. Analogue models of caldera collapse in strike-slip tectonic regimes. Bull. Volcanol. 70, 773–796. doi:10.1007/s00445-007-0166-x Hutton D.H.W., Reavy R.J., 1992. Strike-slip tectonics and granite petrogenesis. Tectonics 11:960–967 Jochum Klaus, P., Stoll, B., Herwig, K., Willbold, M., Hofmann, A.W., Amini, M., Aarburg, S., Abouchami, W., Hellebrand, E., Mocek, B., Raczek, I., Stracke, A., Alard, O., Bouman, C., Becker, S., Dücking, M., Brätz, H., Klemd, R., de Bruin, D., Canil, D., Cornell, D., de Hoog, C.J., Dalpé, C., Danyushevsky, L., Eisenhauer, A., Gao, Y., Snow, J.E., Groschopf, N., Günther, D., Latkoczy, C., Guillong, M., Hauri, E.H., Höfer, H.E., Lahaye, Y., Horz, K., Jacob, D.E., Kasemann, S.A., Kent, A.J.R., Ludwig, T., Zack, T., Mason, P.R.D., Meixner, A., 114 Rosner, M., Misawa, K., Nash, B.P., Pfänder, J., Premo, W.R., Sun, W.D., Tiepolo, M., Vannucci, R., Vennemann, T., Wayne, D. and Woodhead, J.D., 2006. MPI-DING reference glasses for in situ microanalysis: New reference values for element concentrations and isotope ratios. Geochemistry, Geophysics, Geosystems, 7(2). Koppers, A. A. P., 2002. ArArCALC—Software for 40Ar/39Ar age calculations, Comput. Geosci., 28(5), 605–619, doi:10.1016/S0098-3004(01)00095-4. Koppers, A.A.P., Gowen, M.D., Colwell, L.E., Gee, J.S., Lonsdale, P.F., Mahoney, J.J., Duncan, R.A., 2011. New 40Ar/39Ar age progression for the Louisville hot spot trail and implications for inter–hot spot motion. Geochem. Geophys. Geosyst. 12:Q0AM02. http:// dx.doi.org/10.1029/2011GC003804. Kuiper, K.F., Deino, A., Hilgen, F.J., Krijgsman, W., Renne, P.R., Wijbrans, J.R., 2008. Synchronizing rock clocks of Earth history. Science 320:500–504. http://dx.doi.org/10. 1126/science.1154339. Kutterolf, S., Freundt, A., Schacht, U., Bürk, D., Harders, R., Mörz, T., Peréz, W., 2008. Pacific offshore record of plinian arc volcanism in Central America: 3. Application to forearc geology. Geochemistry, Geophys. Geosystems 9. doi:10.1029/2007GC001826 La Femina, P., Dixon, T., and Strauch, W., 2002, Bookshelf faulting in Nicaragua: Geology, v. 30, p. 751–754, doi: 10.1130/0091- 7613(2002)030<0751:BFIN>2.0.CO;2. LaFemina, P., Dixon, T.H., Govers, R., Norabuena, E., Turner, H., Saballos, A., Mattioli, G., Protti, M., Strauch, W., 2009. Fore-arc motion and Cocos Ridge collision in Central America. Geochemistry, Geophys. Geosystems 10. doi:10.1029/2008GC002181 LeBas et al. 1986. A chemical classification of volcanic rocks based on the total alkali- silica diagram: Journal of Petrology, v. 27, p. 745-750. Lexa, J., Šebesta, J., Chavez, J.A., Hernández, W., Pécskay, Z., 2011. Geology and volcanic evolution in the southern part of the San Salvador Metropolitan Area. J. Geosci. 56, 105–140. doi:10.3190/jgeosci.088 Lipman PW (1984) The roots of ash-flow calderas in North America: windows into the tops of granitic batholiths. J Geophys Res 89 :8801–8841 115 Lipman, P.W., 1995. Subsidence of ash-flow calderas; role of magma chamber geometry. IUGG Gen. Assem. 21 (Week A), 452. Lipman, P.W., Dungan, M.A., Brown, L.L., Deino, A., 1996. Recurrent eruption and subsidence at the Platoro caldera complex, southeastern San Juan volcanic field, Colorado: New tales from old tuffs. Bull. Geol. Soc. Am. 108, 1039–1055. doi:10.1130/0016-7606(1996)108<1039:REASAT>2.3.CO;2 Lipman, P.W., 1997. Subsidence of ash-flow calderas: relation to caldera size and magma-chamber geometry. Bull. Volcanol. 59, 198–218. doi:10.1007/s004450050186 Lipman, P.W., 2000. Calderas. In: Sigurdsson, H. (Ed.), Encyclopedia of Volcanoes. Academic Press, San Francisco, pp. 643– 662. Lipman, P. and Mullineaux, D., 1981, The 1980 eruptions of Mount St. Helens, Washington. U.S. geological survey professional paper 1250, 844 p; Ludwig, K. 2008. Manual for Isoplot 3.7. Berkeley Geochronology Center, Special Publication 4, rev. p. 77 Macedonio, G. and Pareschi, M.T., 1991. An algorithm for the triangulation of arbitrarily distributed points: Applications to volume estimate and terrain fitting. Comput. Geosci. 17, 859–874. doi:10.1016/0098-3004(91)90086-S Macedonio, G., Costa, A. and Longo, A., 2005. A computer model for volcanic ash fallout and assessment of subsequent hazard. Computers & Geosciences, 31(7): 837-845. Mann, P., 2007. Overview of the tectonic history of northern Central America: In: Mann, P., ed., Geologic and Tectonic Development of the Caribbean Plate Boundary in Northern Central America, Geological Society of America Special Papers 428, pp. 1–19. Marshall, D. J. 1988. Cathodoluminiscence of geological materials. Unwin Hyman, 181 pp. Marti J., Ablay G. J., Redshaw L. T., Sparks R. S. J., 1994. Experimental studies of collapse calderas. J Geol Soc Lond 151 :919–929 Marti, J., Gudmundsson, A., 2000. The Las Canadas caldera (Tenerife, Canary Islands): an overlapping collapse caldera generated by magma-chamber migration. J. 116 Volcanol. Geo- therm. Res. 103, 161–173. Martí, J., Groppelli, G., & da Silveira, A. B., 2018. Volcanic stratigraphy: a review. Journal of Volcanology and Geothermal Research. Volume  357, 15 May 2018, Pages 68-91. Martínez-Díaz, J.J., Álvarez-Gómez, J.A., Benito, B., Hernández, D., 2004. Triggering of destructive earthquakes in El Salvador. Geology 32, 65–68. doi:10.1130/G20089.1 Mastin, L.G., Guffanti, M., Servranckx, R., Webley, P., Barsotti, S., Dean, K., Durant, A., Ewert, J.W., Neri, A., Rose, W.I., Schneider, D., Siebert, L., Stunder, B., Swanson, G., Tupper, A., Volentik, A. and Waythomas, C.F., 2009. A multidisciplinary effort to assign realistic source parameters to models of volcanic ash-cloud transport and dispersion during eruptions. Journal of Volcanology and Geothermal Research, 186(1): 10-21. Miles, D H, 1997 The interpretation, presentation, and use of tree-ring dates, Vernacular Architect, 28, 40-56 Moore, I., Kokelaar, P., 1997. Tectonic influences in piecemeal caldera collapse and Glencoe volcano, Scotland. J. Geol. Soc. (Lond) 154, 765– 768. Moore, I., Kokelaar, P., 1998. Tectonically controlled piecemeal caldera collapse: a case study of Glencoe volcano, Scotland. Geol. Soc. Am. Bull. 110, 1446–1466. Montero, W., and Dewey, J., 1982. Shallow-focus seismicity, composite focal mechanisms, and tectonics of the Valle Central of Costa Rica: Bulletin of the Seismological Society of America, v. 72, p. 1611–1626. Nairn, I.A., Wood, C.P., Bailey, R.A., 1994. The Reporoa caldera, Taupo volcanic zone: source of the Kaingaroa Ignimbrites. Bull. Volcanol. 56, 529– 537. Newhall, C.G., Dzurisin, D., 1988. Historical unrest at large calderas of the world. U.S. Geological Survey. 1109 pp. Newhall, C.G., Self, S., 1982. The volcanic explosivity index (VEI) an estimate of explosive magnitude for historical volcanism. Journal of Geophysical Research: Oceans 87: 1231–1238 http://dx.doi.org/10.1029/JC087iC02p01231. Ninkovich, D. Y Heezen, B., C., 1965. Santorini tephra. Colston Pap. (Bristol University), 17, 413. 117 Peccerillo, A., and Taylor, S. R., 1976. Geochemistry of Eocene calcalkaline volcanic rocks from the Kastamonu area, Northern Turkey. Contrib. Mineral. Petrol., 58:63-81. Pedrazzi, D., Sunyé-Puchol, I., Aguirre-Díaz, G. J., Costa, A., Davila-Harris, P., Hernández, W. and Gutierrez, E., 2018. The Ilopango Tierra Blanca Joven (TBJ) eruption, El Salvador: volcano-stratigraphy of a major Holocene event of Central America and hazards implications EGU General Assembly 2018, Vol. 20, EGU2018-8455. Pratt, S.E., 2012, Eruption of El Salvador’s Ilopango explains…A.D. 536 global cooling: Earth Magazine, February 25, 2012 issue, 5 p. Quane, S.L., Russell, J.K., 2005. Ranking welding intensity in pyroclastic deposits. Bull. Volcanol. 67, 129–143. doi:10.1007/s00445-004-0367-5 Rampino, M.R., Self, S., 1993. Climate-Volcanism Feedback and the Toba Eruption of ∼74,000 Years Ago. Quat. Res. doi:10.1006/qres.1993.1081 Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., Weyhenmeyer, C.E., 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0e50,000 years cal BP. Radiocarbon 51 (4), 1111e1150. Reynolds J., H., 1987. Timing and sources of Neogene and Quaternary volcanism in south–central Guatemala. J Volcanol Geotherm Res 33: 9–22 Richer, M., Mann, C.P., Stix, J., 2004. Mafic magma injection triggers eruption at Ilopango Caldera, El Salvador, Central America. Spec. Pap. 375 Nat. Hazards El Salvador 175–190. doi:10.1130/0-8137-2375-2.175 Rose, W.I., Conway, F.M., Pullinger, C.R., Deino, A., McIntosh, W.C., 1999. An improved age framework for late Quaternary silicic eruptions in northern Central America. Bull. Volcanol. 61, 106–120. doi:10.1007/s004450050266 Saxby, J., Gottsmann, J., Cashman, K., Gutiérrez, E., 2016. Magma storage in a strike- 118 slip caldera. Nat. Commun. 7, 12295. doi:10.1038/ncomms12295 Schmincke, H.-U., 1967. Cone sheet warm, resurgence of Tejeda caldera, and early geologic history of Gran Canaria. Bull. Volcanol. 31, 153–162. Sheets, P.D., 1979. Environmental and cultural effects of the Ilopango eruption in Central America. Volcan. Act. Hum. Ecol. 525–564, ill. doi:http://dx.doi.org/10.1016/B978-0-12-639120-6.50022-4 Self, S., 2006. The effects and consequences of very large explosive volcánic eruptions. Phil. Trans. R. Soc. A, 364, 2073-2097. Doi: 10.1098/rsta.2006.1814. Self, S. y Blake, S., 2008. Consequences of Explosive Supereruptions. Elements, 4, 41- 46. DOI: 10.2113/GSELEMENTS.4.1.41 Sigurdsson, H., et al., 2006. Marine Investigations of Greece’s Santorini Volcanic Field. Eos 87 (34): 337–348. Siebert, L., Simkin, T., 2002. Volcanoes of the World: an Illustrated Catalogue of Holocene Volcanoes and their Eruptions. Smithsonian Institution, Global Volcanism Program Digital Information Series, GVP-3, (http://www.volcano.si.edu/world/). Simkin, T., and Siebert, L., 1994. Volcanoes of the World, 2nd ed., Geoscience Press, Tucson Smith, R.L., 1979. Ash-flow magmatism: Geological Society of America Special Paper 1749, 180, p. 5–27. Smith, R.L., Bailey, R.A., 1968. Resurgent caldrons. Mem. Geol. Soc. Am. 116, 613– 662. Smithsonian Global Volcanism Program, 2013, Natural Museum of Natural History, Smithsonian Institution, Washington, U.S. http://www.volcano.si.edu. Sofield, D., 2004. Eruptive history and volcanic hazards of Volcan San Salvador. Geol. Soc. Am. Spec. Pap. 375, 147–158 Solari, L.A., Gómez-Tuena, A., Bernal, J.P., Pérez-Arvízu, O., and Tanner, M., 2010. U-Pb zircon geochronology with an integrated LA-ICP-MS microanalytical workstation: Achievements in precision and accuracy: Geostandards and Geoanalytical Research, v. 34, p. 5–18. 119 Sparks, R., S., J., 1979. The Santorini eruption and its consequences. Endeavour, New Series, Volume 3, No. 1. Sparks, R., S., J., Self, S., Grattan, J.,P., Oppenheimer, C., Pyle, D.,M. y Rymer, H., 2005. Supereruptions: global effects and future threats. Report of a Geological Society of London Working Group, The Geological Society, London, 24 pp Sun S.S, and McDonough W.F., 1989. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes, in Saunders A.D., and Norry, M.J.,eds ., Magmatism in the Oceanic Basins: Geological Society of London Special Publication 42, p. 313-345. Suñe-Puchol, I., Aguirre-Díaz, G.J., Dávila-Harris, P., Miggins, D.P., Pedrazzi, D., Costa, A., Ortega-Obregón, C., Lacan, P., Hernández, W., Gutiérrez, E., 2019. The Ilopango caldera complex , El Salvador    : Origin and early ignimbrite-forming eruptions of a graben / pull-apart caldera structure. J. Volcanol. Geotherm. Res. 371, 1–19. doi:10.1016/j.jvolgeores.2018.12.004 Stix, J., Kennedy, B. et al. 2003. Caldera-forming processes and the origin of submarine volcano- genic massive sulfide deposits. Geology, 31, 375 – 378. Tikoff, B., de Saint Blanquat, 1997. Transpressional shearing and strike-slip partitioning inthe Late Cretaceous Sierra Nevada magmatic arc, California. Tectonics 16, 442–459. Turner, H.L., La Femina, P., Saballos, A., Mattioli, G., Jansma, P., and Dixon, T., 2007. Kinematics of the Nicaraguan forearc from GPS geodesy: Geophysical Research Letters, v. 34, 5 p. Vallance J., and Houghton B. (1998). The AD 260 eruption at Lake Ilopango, El Salvador: A complex explosive eruption through a caldera lake. National Science Foundation, Research Proposal. Walker GPL (1988) Three Hawaiian calderas: an origin through loading by shallow intrusions? J Geophys Res 93:14773–14784 Weber H. S., Wiesemann G., Wittek indt H., 1974. Mapa Geológico de la República de El Salvador / Geologische Übersichtskarte der Republik El Salvador 1 : 500,000 (after geological maps 1:100000 – 1967-74) Bundesanstalt für Geowissenschaften und Rohstoffe, Hannover ; Bundesanstalt fu¨r Bodenforschung, Hannover. 120 Weinberg, R. E., 1992. Neotectonic development of western Nicaragua: Tectonics, 11(5), 1010–1017. Weyl, R., 1957. Las Tobas Fundidas de la Cadena Costera. Universidad de Kiel, Alemania, DC 5551.2 Williams, H., Meyer-Abich, H., 1955. Volcanism in the southern part of El Salvador, with particular reference to the collapse basins of Lakes Coatepeque and Ilopango. Univ. Calif. Publ. Geol. Sci. 32, 64 p. 121