This chapter serves as a general introduction to the research presented in this thesis. First, the aims and objectives of the research are introduced, followed by a review of the current state of knowledge on the dynamics and petrological diversity of volcanic plumbing systems.

Subsequently, a summary of previous research on Nevado de Toluca volcano, which serves as a case study for much of the work, is presented. During the course of this PhD, two consecutive field studies have been carried out in Mexico with are briefly described. Finally, the structure of this thesis is laid out.

1.1 Aims and objectives

The risk emanating from volcanic hazards to mankind is today higher than at any time in

recorded history, as close to 800 million people live within radius of 100 km to an active volcano (Brown et al., 2015). Consequently, the principle goal of volcanology is to improve forecasting abilities, which requires a mechanic understanding of the dynamical processes and rates

operating in magmatic plumbing systems (Sparks and Cashman, 2017; Edmonds et al., 2019).

Despite large advances over the last 30 years in the understanding of the architecture and compositional diversity of subvolcanic magma plumbing systems (recently summarized by Sparks et al., (2019)), many facets of how volcanoes work remain enigmatic.

This thesis explores different aspects of the rates, pace and chemical diversity in

subvolcanic reservoirs based on case studies at Nevado de Toluca volcano in the central Mexican Highland (Macias and Arce, 2019). The two main aims of the presented research are: 1)

Developing new tools to infer critical magmatic system variables based on geological archives, and 2) improve forecasting abilities for volcanic eruptions by linking petrological records to geophysical models. The concrete objectives which are addressed in this thesis are:

(I) To investigate which processes dominantly generate the erupted magma geochemistry over the long-term history of the volcano and constrain how the composition of rocks and minerals varies with time.

(II) Evaluate the current physico-chemical state and eruptive potential of Nevado de Toluca by further advancing a novel inversion technique based on zircon

geochronology and thermal modelling (Caricchi et al., 2016).

(III) To determine how fast long-dormant volcanoes, such as Nevado de Toluca, can transition into unrest and eruption, and to examine differences and similarities in the processes that led to different eruptions.

(IV) Constrain the controls on the focus and range of chemical composition erupted by volcanoes over time.

1.2 Dynamics and chemical diversity of crustal magmatic systems

Hidden from direct observation, magmas move through and accumulate in the Earth’s crust at largely unknown rates (White et al., 2006). However, the pace of magma input and storage is anticipated to modulate the variety of chemical compositions that give rise to the diversity of igneous rocks (Kent, 2013; Wörner et al., 2017; Till et al., 2019), the pressurization of

subvolcanic reservoirs initiating volcanic eruptions (Jellinek and DePaolo, 2003; Gregg et al., 2013; Degruyter and Huber, 2014), and the temporal evolution of the physico-chemical state of magmatic systems (Caricchi and Blundy, 2015). Understanding these mass and energy fluxes is therefore important for a range of problems in Earth Science, including the differentiation of the Earth’s crust (Ducea et al., 2017), the formation of magmatic ore deposits (Chelle-Mishou et al., 2017) and the inner workings of volcanoes (Caricchi et al., 2014a), which is the primary focus of this thesis.

Igneous plumbing systems are at present envisioned as highly complex networks of sub-vertical dykes, horizontal sills and irregularly shaped magma bodies (Gudmundsson et al., 2014;

Tibaldi, 2015; Cruden et al., 2017; Breitkreuz et al., 2018; Magee et al., 2018), which are

transiently interconnected with each other during short episodes of instability and reorganization that may culminate in volcanic eruptions (Sparks et al., 2018). Field relations, petrological and geophysical evidence suggests that the anatomy of such structures may extend through the entire Earth crust (Ruprecht and Plank, 2013; Christopher et al., 2015; Cashman et al., 2017), yet a shallow magma storage region at pressures of 2±0.5 kbar is evident for most volcanic systems, which may reflect a favorable magma accumulation depth due to crustal rheology and the pressure dependent exsolution of volatile elements from melts (Menand, 2011; Huber et al., 2019).

While not universally accepted (Adam et al., 2016), most would agree that the compositional diversity of magmas is generally obtained at deep to mid crustal levels, during incremental assembly of magma batches into larger reservoirs over prolonged periods of time (Annen and Sparks, 2002, Dufek and Bergantz, 2005; Annen et al., 2006). Pulsed magma

injection has been extensively recorded in petrological archives, such as crystal zonation textures and chemistry (Singer et al., 1995; Tepley III et al., 2000; Cashman and Blundy, 2003;

Humprehys et al., 2006; Ginibre et al., 2007; Ruprecht and Wörner, 2007; Smith et al., 2009;

Kent et al., 2010) and is evident from geophysical and geodetic signals, observed during unrest

2014; Hickey et al., 2016). Open system heat and/or mass exchange (i.e. mixing/mingling processes) between recharging magmas and pre-existing reservoirs are therefore widely recognized as one of the major agents in driving the geochemical diversity of volcanic rocks (Eichelberger et al., 2000; Reubi and Blundy, 2009; Kent, 2013; Laumonier et al., 2014). Partial fusion or complete digestion of mostly deep crustal rocks by intruding magmas (Hildreth and Moorbath, 1988; Martin and Sigmarsson, 2007; Solano et al., 2012; Clemens and Stevens, 2016), gravity driven fractionation of crystals and melt (Grove et al., 2003; Lee and Bachmann, 2014;

Nandedkar et al., 2014; Ulmer et al., 2018; Müntener and Ulmer, 2018), as well as the

segregation of magmatic liquids (Bachmann and Bergantz, 2004; Dufek and Bachmann, 2010;

Solano et al., 2012; Floess et al., 2019; Holness, 2018) e.g. by reactive porous melt flow (Spiegelman and Kelemen, 2003; Lissenberg and MacLeod, 2016; Jackson et al., 2018), are all believed to contribute to the compositional diversity of magmas, but their relative importance is still widely debated. While the chemical diversity observed in genetically related rock suits at individual volcanoes reflects magma differentiation processes, some system erupt magmas covering a wide spectrum of geochemical compositions, other volcanoes produce magmas with monotonous chemistry throughout their entire lifespan. Such differences may be attributed to physical filtering processes, such as density or viscosity barriers, that may prevent certain compositions from being erupted (Stolper and Walker, 1980; Marsh, 1981; Pinel and Jaupart, 2000). Such barriers might be overcome by hybridization of mafic and silicic compositions during recharge into crustal magma reservoirs (Kent et al., 2010; Kent, 2013). Alternatively, differences in the variety of erupted magma compositions may reflect local differences in the architecture of magmatic systems and rates of magma throughput (Wörner et al., 2017).

Volcanic plumbing systems also undergo substantial time evolution (Annen et al., 2006;

Melekhova et al., 2013; Annen et al., 2015; Caricchi and Blundy, 2015). Over the last 30 years, detailed mapping and dating of volcanic successions have revealed life-cycles of hundreds of thousands of years up to several million years for composite volcanoes in Arc settings (Hildreth and Lanphere, 1994; Frey et al., 2004; Bacon and Lanphere, 2006; Jicha et al., 2006; Hildreth, 2007; Hora et al., 2007; Singer et al., 2008; Gertisser and Keller, 2003; Thouret et al., 2005;

Muir et al., 2015; Godoy et al., 2018). Numerical models, simulating the large scale thermal and mechanical evolution of magmatic reservoirs, as well as mirco-scale geochemical and

petrological records, show that such systems tend to homogenize in time and spend most of their lifespans in a highly crystallized state (e.g. Bachmann and Bergantz, 2008; Annen, 2009; Huber et al., 2011; Allan et al., 2013). This is corroborated by crystallization age distributions of zircon crystals, which typically record protracted silicic melt presence on the order of 104-106 years

(Kent and Cooper, 2018; Cooper, 2019 and references therein. These long timescales are contrasted by estimates from modelling of diffusive elemental relaxation between adjacent crystal zonation boundaries that indicate years to millennial duration of associated processes, such as magma recharge (e.g. Costa et al., 2019). Taken together, the diversity of chemical compositions erupted by volcanoes and the large range of timescales derived by radiometric dating and kinetic modelling, point towards a multitude of nested processes (Spera and Bohrson, 2018) that need to be deciphered in order to improve mitigation of the impact of volcanic


1.3 Nevado de Toluca, Trans Mexican Volcanic Belt

Nevado de Toluca (19 06 30N; 99 45 30 W; 4680 m above sea level), also known as volcán Xinantéctal, is a long-lived stratovolcano located in the central Mexican highland about 80 km southwest of Mexico City (Macias and Arce, 2019). It is part of the Trans Mexican Volcanic Belt, a 1000 km long east-west trending volcanic arc of Miocene to Holocene age that developed in response to the subduction of the oceanic Cocos and Rivera plates beneath the North

American continental lithosphere at different angles (Pardo and Suárez, 1995; Ferrari et al., 2012). The volcanic edifice was built from various short silicic lava flows and domes with intercalated pyroclastic deposits, with the most prominent feature of the volcanoes morphology being a 2.5 ×1.5 km2 horseshoe shaped crater that opens towards the East in direction in the direction of the City of Toluca (Fig. 1.1, Fig. 1.2; Norini et al., 2004).

Detailed stratigraphic and radiometric dating has shown that volcanic activity at Nevado de Toluca started in the Early Pleistocene at 1.5 Ma and was continuous since then, frequently with long periods of dormancy of several thousand years between eruptions (Bloomfield and Valastro, 1974; Garcia-Palomo, 2002; Arce et al., 2003, 2005; Capra et al., 2006; Torres-Orozco et al., 2017). The interplay of three fault systems that intersect beneath the volcano potentially influenced the location and eruptive history of the volcano (Garcia-Palomo et al., 2000; Bellotti et al., 2006), especially several cone destruction episodes by sector collapses that gave rise to a set of widely dispersed debris avalanche deposits around the volcano (Capra and Macias, 2000;

Caballero and Capra, 2011). The volcanos history can be divided into several stages based on geomorphology and age relations (Fig. 1.1b; Torres-Orozco et al., 2017). During the early ‘Old

(Fig. 1.3a, b) between 1.5 and 0.27 Ma in the periphery of the main volcano. Only a few

monogenetic systems have been dated in the area (Bloomfield, 1975; Arce et al., 2013a; Torres-Orozco et al., 2017), but available data point towards a long age range between at least 0.86 Ma to 8.5 ka. The younger eruptive history can be dived into the 0.57 – 0.009 Ma ‘Recent Nevado’

stage, mostly characterized by silicic lava effusion in the crater area, and a sequence of

pyroclastic deposits ‘Young PD stage’ spanning ca. 0.114 to 0.003 Ma (Macias et al., 1997; Arce et al., 2006).

The young pyroclastic deposits (Fig. 1.3c-f) record a complex series of dome

destruction events that are preserved in block and ash flow deposits (Garcia-Palomo et al., 2002), as well as by a sequence of at least three Plinian eruptions (Arce et al., 2003, 2005, 2006; Capra et al., 2006). The oldest and least voluminous of these Plinian events is the 0.85 km3 DRE Lower Toluca Pumice (LTP) that was dated at 26 ka calibrated 14C years BP (Capra et al., 2006). In the following 13 ka activity at Nevado de Toluca was again dominated by dome formation and destruction, notably at about 16500 cal. 14C years BP when a massive block and ash flow of 0.5 km3 was produced (Arce et al., 2006). Renewed activity commenced at approximately 14

thousand years ago when a Plinian eruption of 1.8 km3 DRE gave rise to the Middle Toluca Pumice (MTP) that was distributed mainly to the north-east of the volcano (Arce et al., 2005).

The so far largest and last Plinian-type eruption from Nevado de Toluca, the 8 km3 DRE Upper Toluca Pumice (UTP), was erupted at about 12.5 ka and blanketed the area today occupied by Mexico City within the 10 cm isopach (Arce et al., 2003), demonstrating the large potential hazards emerging from this volcano. Following this event, the dacite dome ‘El Ombligo’ (Fig.

1.2c) was emplaced inside the crater at about 9.5 ka (Arce et al., 2003; Bernal et al., 2014). The last eruptive activity at Nevado de Toluca was described as a phreatomagmatic surge and dated by Macias et al. (1997) with a calibrated 14C age of 3768 years BP.

Nevado de Toluca is a remarkably uniform volcano in terms of its compositional diversity, which is restricted to minor andesite and subalkaline dacite throughout its lifetime (Torres-Orozco et al., 2017). Several detailed petrological studies have been conducted on the Plinian deposits (Arce et al., 2003, 2005, 2006; Martínez-Serrano et al., 2004; Capra et al., 2006;

Smith et al., 2009 Arce et al., 2013), which are summarized here as these eruptions are a major focus of this study. Petrographic analysis of the Plinian pumice showed that the UTP and MTP deposits are identical in terms of their mineral phase assemblage plag>opx>hbl>>ilm+mt+bt.

The LTP rocks, however, lack orthopyroxene as a major phenocryst phase. Notably, this eruption also differs from the preceding events by the presence of schist fragments that are not found in

the MTP and UTP, possibly reflecting different fragmentation or reservoir depth (Arce et al., 2013). Estimates of pre-eruptive conditions (P-T-X-fO2) for all three eruptions have been obtained by experimental phase equilibria and Fe-Ti oxide thermometry (Arce et al., 2006, 2013).

Figure 1.1. Topography and geology of Nevado de Toluca a) Satellite image and digital elevation model (DEM) of the Nevado de Toluca area (source: Google Earth). B) Geological map overlaying the DEM showing the major eruptive stages of the volcano (modified from Torres-Orozco et al., 2017).

Relatively high pre-eruptive temperature conditions have been estimated for the LTP oxide pairs ranging between 861–870°C, which are about 10 to 20°C higher than estimates for the UTP and MTP magmas (Arce et al., 2013). Experimental constraints on storage pressures are similar for the UTP and LTP (between 150 and 200 MPa) at water saturated conditions and are with 200 to 300 MPa slightly higher for the MTP magma (Arce et al., 2006, 2013). This shallow part of the magmatic system evolves as an open system, where fresh pulses of dacitic and

andesitic magma enter the preexisting reservoirs (Smith et al., 2009). Martínez-Serrano et al., (2004) used isotopic (Sr, Nd, Pb) and trace element evidence to show that even though Nevado de Toluca sits on thick continental crust of about 50 km, crustal melting plays only a marginal role in silicic magma genesis. Interestingly, trace element patterns of Nevado de Toluca and the neighboring monogenetics show an adakite-like signatures. The wealth of information from field, geochronologic and petrological studies obtained on Nevado de Toluca makes it an ideal target for more advanced studies to understand magma dynamics in volcanic arcs.

Two consecutive field campaigns have been carried out at Nevado de Toluca during the course of this study in August 2016 and January 2018. The principle goal of this fieldwork was to collect samples representative of the entire eruptive history of the volcano and surrounding monogenetic cones. Further, we focused on stratigraphic relations and sampling of young pyroclastic deposits. A total of 134 samples were collected, of which 97 were analyzed for this thesis. A description of the collected samples and their locations is attached in the electronic supplementary materials (Table E2.1).

Figure 1.2. Selected photographs of Nevado de Tolucas’ crater area from various perspectives.

a) View from south-west showing the main crater domes U31 and U30 and the peripheral dacite dome U27. b) Crater of Nevado de Toluca and Cerro Gordo dome seen from west. c) View into the crater looking southwards. The interior of the crater holds the rounded ‘Ombligo’ dome and two lakes: Laguna de la Luna and Laguna del Sol.

Figure 1.3. Selected outcrop photographs showing the range of eruptive styles and stages of Nevado de Toluca. a) Typical peripheral dacitic lava dome. b) Lava dome U27 in vicinity of the crater. c) Block-and-ash flow deposit. d) Sequence of wet surges and block-and-ash flow

deposits. e) Massive pyroclastic density current deposit (WQ) underlying a fallout layer (line above persons’ heads). e) Plinian fallout (upper coarse white layer), massive and dilute PDCs with cross bedding of the Upper Toluca Pumice eruption.

1.4 Structure of this thesis

This PhD thesis is organized into six main chapters, a bibliography of the referenced literature, and an appendix of supplementary materials to all main research chapters (2-5). Data tables are presented in the electronic supplement to this thesis. In addition, three published articles of the author that are not a direct part of this work are included as a supplement. Chapters 2-5 have been submitted for publication to scientific journals and are either published or currently under peer-review.

Chapter 2 investigates the long-term petrological history of Nevado de Toluca volcano by detailed analysis of bulk-rock and mineral chemistry. A range of thermobarometric

calculations are presented to constrain intensive magmatic system variables. Based upon results from thermal modelling, the life cycle of Nevado de Toluca and the petrologic monotony of Mexican stratovolcanoes are discussed.

Chapter 3 integrates the zircon crystallization and trace element record of Nevado de Toluca with thermal modelling of pulsed upper crustal magma assembly. Matching synthetic and natural zircon age populations and temperature distributions, magma fluxes and

extrusive:intrusive (E:I) ratio are constrained. These results allow to characterize the physico-chemical state and size of the subvolcanic reservoir currently residing beneath Nevado de Toluca.

Chapter 4 provides detailed mineral chemistry for a sequence of Plinian eruptions from Nevado de Toluca. Using the systematics of mineral zoning pattern, a recurrent pattern of magma recharge prior to these events is identified. Using diffusion chronometry on plagioclase and orthopyroxene, the timescale of these processes are investigated.

Chapter 5 explores the question why some volcanoes erupt a large range of chemical compositions and others, like Nevado de Toluca, produce the same monotonous magma chemistry throughout their lifespan. Using a novel integration of thermal and petrological modelling, this study shows that the thermal structure of the crust and magma recharge rates modulate the diversity and temporal evolution of magma bodies.

Chapter 6 provides a general discussion and synopsis of the presented research and recommends directions for future research.




The petrologic diversity of volcanic rocks reflects the dynamics of magma reservoirs and the temporal evolution of magma chemistry can provide valuable information for hazard assessment.

While some stratovolcanoes monotonously produce intermediate magmas (55-68 wt.% SiO2), dominantly erupted magma types (e.g. basaltic andesite, andesite or dacite) frequently differ even between neighboring volcanoes. If such differences arise due to thermal maturation processes over time or are predetermined by other properties of magmatic systems remains poorly understood. This study helps to elucidate the underlying factors modulating the chemistry of the magma preferentially erupted by Nevado de Toluca volcano in Central Mexico. We present a new dataset of bulk-rock and mineral chemistry spanning the entire 1.5 Million years of the volcanos’ eruptive history. The results reveal that Nevado de Toluca dacites and minor andesite originate in a stable configuration of pre-eruptive processes and plumbing system architecture by hybridization between an upper crustal silicic mush and deeper sourced basaltic andesite magmas. Yet, a subtle trend towards increasing silica content with time (2 wt.% in 1.5 Ma) and episodicity in magma hybridization conditions are observed. We use thermal

simulations of pulsed magma injection to probe the controlling variables on the temporal variation and compositional mode of magma geochemistry. The results show that the subtle temporal trend towards increasing bulk-rock SiO2 content is plausibly explained by slightly

Gregor Weber1, Luca Caricchi1, José Luis Arce2

1 Department of Earth Sciences, University of Geneva, Geneva, Switzerland.

2 Instituto de Geología, Universidad Nacional Autónoma de México, Coyoacán, México.

Submitted to Frontiers in Earth Sciences on 18 May 2020.

dropping recharge rates and continued upper crustal reservoir growth. Our modelling also shows that the dominant composition of eruptible magmas (“petrologic mode”) can shift as a function of magma flux, extrusive:intrusive ratio and temperature of the recharge magma. A comparison of SiO2 whole rock distributions for monotonous Mexican stratovolcanoes and their peripheral cones shows that their petrologic modes vary in concert, indicating that the recharge magma chemistry or temperature is a major control on the preferentially erupted magma composition for these volcanoes.

2.1 Introduction

Over the last 30 years, combined geological mapping, dating and geochemical analyses have revealed that the range of erupted magma chemistry differ significantly for individual volcanic centers (Hildreth and Lanphere, 1994; Singer et al., 1997; Gertisser and Keller, 2003; Frey et al., 2004; Thouret et al., 2005; Bacon and Lanphere, 2006; Hildreth et al., 2007; Hora et al., 2007;

Over the last 30 years, combined geological mapping, dating and geochemical analyses have revealed that the range of erupted magma chemistry differ significantly for individual volcanic centers (Hildreth and Lanphere, 1994; Singer et al., 1997; Gertisser and Keller, 2003; Frey et al., 2004; Thouret et al., 2005; Bacon and Lanphere, 2006; Hildreth et al., 2007; Hora et al., 2007;

Dans le document Magma fluxes, timescales and petrological diversity in volcanic plumbing systems: new perspectives from Nevado de Toluca (Mexico) (Page 27-192)