TY - JOUR
AU - Kahl,, Maren
AB - Abstract An understanding of destructive historic eruptions has important implications for the assessment of active plumbing systems and the processes that might precede future hazardous eruptions. At Mount Etna (Sicily, Italy), magma production and eruption frequency have increased dramatically since 1970, however, the recent eruptions are considerably less voluminous than those of the 17th century, which occurred at greater intervals. Seventeenth century activity culminated in the 1669 flank eruption, the most voluminous and destructive in Etna’s recorded history, marking the beginning of a new eruptive period. In this study, we examine trace element zoning patterns recorded in clinopyroxene (lava hosted microcrysts: 0·5–1 mm, lava hosted macrocrysts: 1–5 mm and scoria hosted megacrysts: >5 mm) to reconstruct magma dynamics leading up to the 1669 eruption. The clinopyroxene data are considered alongside previous studies of olivine and plagioclase to present an updated conceptual model for the plumbing system, providing a better understanding of magmatic processes in the lead up to hazardous volcanism. Petrological observations in combination with laser ablation ICP-MS mapping reveal sharp compositional zoning of clinopyroxene, not seen in major element transects. Trace element data, including Cr, Zr, Ni and rare earth elements, show that core, mantle and rim regions originated in distinct magmatic environments. Chromium-rich cores (up to 1080 ppm Cr) are in disequilibrium with the glassy-microcrystalline host groundmass and indicate crystal inheritance from a primitive magma source. Oscillatory zoning in the mantle of the crystals suggests a sustained period of magma replenishment and crystallization. Finally, ubiquitous Cr-rich (170–220 ppm) rims host many large melt inclusions, suggesting a final recharge event inducing relatively rapid crystal growth and eruption. Temperatures of 1120–1160 ± 27°C and pressures of 300–600 ± 200 MPa calculated for the three magmatic environments based on clinopyroxene composition at 2 wt % H2O place most of the clinopyroxene crystallization at more than 10 km depth. Measuring the consistent thickness of crystal rims (219 ± 33 µm) and assuming growth at a low degree of undercooling (10−8 cm/s), we calculate that the eruption triggering magma recharge invaded the plumbing system less than a month before eruption onset, in agreement with historical accounts of pre-eruptive seismicity. Notably, Cr enrichment in the recharge magma was not coupled with increases in MgO content. We therefore propose that a cryptic recharge with similar composition to the resident melt may have tipped the system to erupt, and that the volume of recharge rather than composition or temperature acted as the primary trigger. Finally, LA-ICP-MS maps of clinopyroxene from the previous eruption of Mount Etna (1651–53) revealed strikingly similar compositional zonation to that of 1669, supporting the notion that magmatic storage environments, associated with voluminous 17th century activity, were long-lived. INTRODUCTION Mount Etna is a basaltic stratovolcano located on the eastern coast of Sicily, Italy, standing at over 3300 m high and covering an area of about 1250 km2 (Viccaro & Cristofolini, 2008). It is currently one of the most active volcanoes in the world, having seen an increase in eruption frequency since the early 1970s; however, Etna’s activity has waxed and waned throughout its recorded history (Hughes et al., 1990; Branca & Del Carlo, 2004). Periods of high activity present a significant physical and economic hazard to the 25% of Sicily’s population who live on the slopes of the volcano. Flank eruptions emerging close to population centres, such as the city of Catania (whose metropolitan population exceeds one million inhabitants), threaten agriculture and infrastructure, and can potentially disrupt air traffic (Andronico & Lodato, 2005; Barsotti et al., 2010; Bonaccorso et al., 2011). While a permanent monitoring system, installed in the early 1980s (Allard et al., 2006), has given significant insight into the present state of the plumbing system, less is known about the processes which drove Etna’s major historic eruptions, or long-term eruption cycles, which may aid in predicting the direction of future activity (Kahl et al., 2015). The magnitude, character and frequency of Etnean eruptions is strongly influenced by its magmatic source and plumbing system geometry, which have varied significantly through time (Ubide & Kamber, 2018). Understanding and recognizing the characteristic processes that led to Mount Etna’s most severe historic events and the signs that preceded them is therefore crucial, not only to present day hazard assessment, but also to the prediction of future activity. This study focuses on the infamous 1669 eruption of Mount Etna, considered the most destructive eruption in its recorded history and the culmination of a period of heightened activity in the 17th century (Branca et al., 2013b). We investigate trace element zoning patterns recorded in clinopyroxene, in combination with thermobarometric constrains on pre-eruptive crystallization and previous data on plagioclase and olivine populations, to reconstruct the lead up to this landmark eruption. Eruptive history and cyclicity Activity in the Mount Etna region has undergone several changes since the onset of magmatism at approximately 600 ka (Branca et al., 2013a). However, stabilisation of the plumbing system and the initiation of stratovolcano building began only at 60 ka with the formation of Ellittico volcano (60–15 ka; Romano & Sturiale, 1982) and most of the edifice we see today was formed during the current Mongibello phase (15 ka–present). Despite the identification of phases and even decadal cycles (Behncke & Neri, 2003) in Etna’s activity, little is known about its long-term eruptive patterns (Kahl et al., 2017). Observations and recording of eruptive behaviour at Mount Etna did not begin until the 17th century, therefore the relationship, if any, between periods of heightened activity is poorly understood (Branca & Del Carlo, 2004). The 17th century was a period of unusually high magma output at Mount Etna (total output volume: 3 km3; Hughes et al., 1990), during which 8 major flank eruptions occurred (1607, 1610, 1614–24, 1630, 1634–36, 1643, 1646–47, 1651–53, 1669; for a summary of these eruptions see table 1 of Kahl et al., 2017) with high effusion rates (1607–1669 average effusion rate: 1·19 m3/s). The 1669 eruption and associated caldera collapse marked a catastrophic end to this period of activity and a clear transition to shorter-lived eruptions of low output (average effusion rate 1670–1755: 0·02 m3/s), which were dominated by more mafic products (Corsaro & Cristofolini, 1996; Branca & Del Carlo, 2004; Branca & Carlo, 2005; Branca et al., 2013b). A similar pattern of reduced magmatic production following periods of voluminous activity has been noted in pre-historic times (e.g. A.D. 300–450 and 950–1060; Tanguy et al., 2007). Eruptions during the 17th century were not only long lasting (up to 10 years), but the lava and tephra erupted were more evolved (mugearites) than typical Etnean products (hawaiites) (Corsaro & Cristofolini, 1996; Viccaro & Cristofolini, 2008; Nicotra & Viccaro, 2012). The lavas of this period are locally named ‘cicirara’ because of the chickpea-like appearance of the plagioclase megacrysts (up to 1·5 cm) they host (Nicotra & Viccaro, 2012; Lanzafame et al., 2013). It has been suggested that the large crystal sizes and evolved compositions of ‘cicirara lavas’ formed in a long-lived, shallow magma reservoir, with crystal residence times exceeding those of recent Etnean magmas (Armienti et al., 1997; Viccaro et al., 2016). Although plagioclase crystallization is typical in the shallow plumbing system where degassing readily occurs, storage of the large magma volumes erupted during the 17th century would have required a plumbing system geometry unlike that we see today (Cristofolini & Romano, 1982; Murru et al., 1999; Aloisi et al., 2002; Patanè et al., 2006). Experimental and petrological evidence, however, place the onset of phenocryst nucleation at high pressures and temperatures (>12 km depth), suggesting that significant magma storage could have occurred in the deeper plumbing system (Armienti et al., 2013; Vetere et al., 2015; Mollo et al., 2015a). Interestingly, the final phases of Elittico volcano (14·1 ± 26 ka; De Rita et al., 1991) were also characterized by the eruption of ‘cicirara lavas’ and culminated in a major caldera collapse (De Rita et al., 1991; Coltelli et al., 2000). Conceivably, periods of protracted magma accumulation and high output could represent the final phases in long-term Etnean cycles (Behncke & Neri, 2003). A climactic and comprehensive emptying of a large, vertically extensive magma storage system could result in summit instability, caldera collapse and subsequent disruption of the plumbing system geometry (Nicotra & Viccaro, 2012). This line of thinking places Etna’s current activity in a new long-medium term eruptive cycle, which began post-1669, and has yet to reach its conclusion (Behncke & Neri, 2003; Clocchiatti et al., 2004; Branca & Carlo, 2005). Activity at Mount Etna has increased steadily following the 1669 eruption, with a dramatic increase in magma production beginning in the early 1970s that continues today (Behncke & Neri, 2003). A rise in the frequency of flank eruptions has been accompanied by a number of petrological and geochemical changes in the erupted products. Pronounced enrichment in alkali and fluid mobile elements, continuing a trend that began in the 17th century, has been attributed to a change in the dominant melt producing mantle source (Tonarini et al., 2001; Viccaro & Cristofolini, 2008). Furthermore, increasing radiogenic-Sr and volatile content since 1971 reflects the arrival of a new and distinct primitive magma to the plumbing system (Clocchiatti et al., 2004; Armienti et al., 2007, 2013; Kamenetsky et al., 2007; Ubide & Kamber, 2018), the signature of which continues to invade and replace the resident central conduit magma (Clocchiatti et al., 2004). This ‘invading’ magma is thought to have bypassed the central conduits in 2001 and 2002–03 during ‘eccentric’ or ‘deep-dyke fed’ eruptions. The absence of large plagioclase phenocrysts in the erupted products and minimal pre-eruptive seismicity (2-hr seismic crisis preceded the 2002 activity; Neri et al., 2005), suggested that these volatile rich melts rose rapidly from depth undergoing little degassing upon ascent (Clocchiatti et al., 2004; Neri et al., 2005). Frequent, low volume eruptions seen today, compared to large volume eruptions of the 17th century occurring at greater intervals (up to 16 years), point to a lack of ‘cicirara’ lava-producing magma storage regions in Mount Etna’s current ‘steady-state’ plumbing system (Branca & Carlo, 2005). Seismological and ground deformation data confirm the absence of a sufficiently large magma reservoir at shallow levels (Patanè et al., 2003). Rather, petrological studies infer several smaller regions of storage and degassing-driven crystallization at shallow levels (Viccaro et al., 2010, 2016; Giacomoni et al., 2014, 2016; Kahl et al., 2017). In addition, clinopyroxene barometry places much crystallization at approximately 10 km depth (Armienti et al., 2007; Ubide & Kamber, 2018; Ubide et al., 2019), in agreement with geophysical studies showing a high velocity body at 9–16 km (Patanè et al., 2003). Less commonly, olivine and clinopyroxene crystal cargoes record deeper mafic crystallization (>10–20 km) (Mollo et al., 2015b; Giacomoni et al., 2016; Ubide & Kamber, 2018). So, although eruptions of a 17th century scale have not yet recurred, the continued rate of magma supply to the plumbing system certainly highlights the possibility of a return to such conditions. The compositional zoning record Detailed study of eruptive products from Mount Etna’s past destructive events is a means of understanding the relationship between periods of past heightened activity and the heightened activity we see today. Mount Etna’s magmas are typically trachybasaltic (hawaiites and mugearites), and host a consistent mineral assemblage of clinopyroxene, plagioclase, olivine and titanomagnetite (D’Orazio et al., 1998; Corsaro et al., 2009; Ubide & Kamber, 2018). The compositional homogeneity of the eruptive products does not reflect Etna’s dynamic and complex open-plumbing system, where diverse melt compositions are mixed and buffered prior to eruption (Ferlito et al., 2012; Mollo et al., 2015b). Mineral zoning on the other hand records complex crystallization histories and has the potential to resolve the processes that led to periods of high-output and destruction (Davidson et al., 2007; Ginibre et al., 2007; Ubide & Kamber, 2018). Using mineral data from destructive historic eruptions to identify key magmatic processes and associated signs of unrest, we can aim to recognize the onset and development of similar processes in the future as well as inform the interpretation of Etna’s current and future mineral record (Blundy & Cashman, 2008). Clinopyroxene, in particular, is an early liquidus phase at Mount Etna and provides a uniquely extensive crystallization record across a wide range of pressures and water contents (Armienti et al., 2007, 2013; Giacomoni et al., 2016). Slow elemental diffusion in pyroxene relative to olivine also ensures that this record remains largely unaltered (Müller et al., 2013; Bouvet de Maisonneuve et al., 2016). Traditionally, petrographic observation and major elements analysis have been employed to identify processes such as recharge and mixing; however, in the absence of major element variation, trace element zonation can record more cryptic processes (Ginibre et al., 2007). For a given magma source, enrichments in compatible elements Cr, Ni, and corresponding depletion in incompatible elements (rare earth elements (REE) and high field strength elements, such as Zr) highlight the arrival of primitive melt into a resident magmatic system (Ubide & Kamber, 2018), a process well known to be an efficient trigger of eruptions (Sparks et al., 1977; Bergantz et al., 2015) and a crucial factor in growing and sustaining magma bodies (Reubi & Blundy, 2009). The general behaviour of incompatible elements can be somewhat modified at Mount Etna as a result of its unusual geodynamic setting. Melting of an enriched mantle source can produce melts enriched in HFSE (e.g. Zr), LILE and in LREE relative to HREE (Viccaro & Cristofolini, 2008; Kahl et al., 2015). However, assuming that two melts are derived from the same source, a relative depletion of these incompatible elements coupled with enrichment in compatible transition metals will be observed in the more primitive member. This study We investigate compositional zoning patterns recorded in clinopyroxene crystals from the landmark 1669 eruption. Our study includes individual megacrysts (>5 mm) sourced from the Monti Rossi scoria cone as well as macrocrysts (1–5 mm) and microcrysts (0·5–1 mm) hosted in associated lava flows. We examine the distribution of major and trace elements in the crystals using a combination of electron microprobe (EPMA) transects and laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) mapping. LA-ICP-MS mapping allows us to visually and quantitatively resolve subtle trace element zoning patterns, thus giving new insights into pre-eruptive magmatic processes (Ubide et al., 2015). This high-resolution approach is coupled with petrographic observations and geochemistry of handpicked groundmass. Our findings are considered alongside detailed analysis of plagioclase (Nicotra & Viccaro, 2012) and olivine (Kahl et al., 2017) from previous studies to present a comprehensive conceptual model of pre-eruptive plumbing system dynamics in 1669. THE 1669 ERUPTION: PRECURSORS AND PRODUCTS According to historical accounts, pre-eruptive seismicity began on 25 February 1669, 15 days before eruption onset, and was felt in Catania. Seismicity intensified from 6 pm on 8 March when earthquakes shook the towns of Nicolosi and Pedara. Finally, on 10 March an even stronger earthquake destroyed seventy buildings and caused widespread ground failure in Nicolosi, just northwest of which the eruption began the following day (Portoghese, 1869). The eruption itself has been described as ‘an archetype of the most hazardous expected eruption on the densely populated flanks of the volcano’ (Mulas et al., 2016). An anomalously large volume of magma was emitted (607 ± 105 × 106m3) at high average effusion rates (58 ± 10 m3/s), setting the 1669 eruption apart from any other eruption of Mount Etna in recorded history. It emerged at low elevation on the southern flanks of the volcano (site of the Monti Rossi scoria cone) and emitted, over the course of 4 months, effusive lava flows that reached a record length of 17 km (Branca et al., 2013b; Mulas et al., 2016). Many small villages (Nicolosi, Pedara, Malpasso, Mascalucia and Gravina) as well as the southwest portion of Catania were destroyed by lava flows and earthquakes (Branca et al., 2013b; 2015). The eruption began on the morning of the 11th of March with several vents forming along a series of extensive N–S fissures throughout the day. The main vent, which eventually formed the Monti Rossi scoria cone, opened at 6:30 pm that same day, 775–850 m above sea level, and the first lava flow began outpouring that night. During the first 2–3 days of the eruption (11–13 March), effusion rates were extremely high (160 to 225 m3/s) and almost 90% of the total erupted tephra was produced (Borelli, 1670; Mulas et al., 2016). Most of the Monti Rossi scoria cone was built during this early, explosive stage, and deposits were the most primitive and olivine-rich (beds A1 and A2; Mulas et al., 2016) of the entire eruption. Subsequent tephra was more evolved (beds A3 to C2; Mulas et al., 2016) and erupted at significantly lower effusion rates (30 m3/s) before a return to slightly more primitive products towards the end of the eruption (Unit D; Mulas et al., 2016). Two distinct eruptive stages have been noted within the lava products, though the extent of the compositional gap remains a subject of debate. Corsaro et al. (1996) distinguished between more primitive lavas erupted during the early stages of the eruption between 11–20 March (SET 1) and more evolved lavas erupted later between 21 March and 11 July (SET 2); the geochemical differences have since been attributed to variability in mineral accumulation between samples (Kahl et al., 2017). Kahl et al. (2017) did, however, note a difference in the composition of olivine rims found in SET 1 (M5 environment; Fo50–59) and SET 2 (M3 environment; Fo65–69) samples. This was attributed to the crystallization of olivine in two compositionally distinct dykes shortly before eruption. In the same samples, olivine cores (M1 environment; Fo75–78 & mm1 environment; Fo73–75) showed no significant difference in composition between SET 1 and SET 2, indicating that crystallization occurred in a shared magmatic environment prior to dyke intrusion and rim crystallization. The samples discussed in this paper are classified as SET 1 and SET 2 in accordance with the eruption dates defined by Corsaro et al. (1996) regardless of composition. In total, volcanic activity lasted for 122 days, coming to an end on 11 July (Borelli, 1670). SAMPLES AND METHODS Sample collection Single megacrysts of clinopyroxene occur at the crater rim of the Monti Rossi scoria cone (site of the 1669 eruption), weathered from the scoria fragments. The megacrysts were described in early studies by Washington & Merwin (1921) and Downes (1974), and belong to SET 1, erupted from 14–19 March (Unit MR2; Mulas et al., 2016). We studied 19 clinopyroxene megacrysts (>5 mm) as well as 12 lava-hosted macrocrysts (1–5 mm) and microcrysts (0·5–1 mm) from two lava flows and a lava bomb. The lava flow samples were collected from both SET 1 and SET 2 lava fronts (samples 1–6 and 2–4; Kahl et al., 2017). The exact eruption date of the lava bomb sample (MtRs; Kahl et al., 2017) is not known, although it is inferred to belong to SET 2 based on a characteristic set of olivine compositions (core–rim profiles) observed in other SET 2 lava samples (Kahl et al., 2017). The locations of all samples are provided in Fig. 1 and global positioning system (GPS) coordinates are listed in Supplementary Data Table S1 (supplementary data are available for downloading at http://www.petrology.oxfordjournals.org). In our study, particular focus is given to megacrysts and macrocrysts as they hold an extensive crystallization record. Fig. 1. Open in new tabDownload slide (a) Digital map of Mount Etna’s southern flank (from Google Earth), adapted from Kahl et al. (2017), showing the location of the 1669 eruption products and source cone in relation to the summit craters and the city of Catania. (b) Colour-coded map adapted from Kahl et al. (2017) showing the distribution of the 1669 lava flows and their eruption dates. The locations of the samples used in this study are marked by stars. SET 1 (yellow and orange) and SET 2 (pink and purple) lava samples were collected from lava flows (Kahl et al., 2017), while the megacrysts and MtRs bomb sample were collected on the rim of the Monti Rossi scoria cone. Fig. 1. Open in new tabDownload slide (a) Digital map of Mount Etna’s southern flank (from Google Earth), adapted from Kahl et al. (2017), showing the location of the 1669 eruption products and source cone in relation to the summit craters and the city of Catania. (b) Colour-coded map adapted from Kahl et al. (2017) showing the distribution of the 1669 lava flows and their eruption dates. The locations of the samples used in this study are marked by stars. SET 1 (yellow and orange) and SET 2 (pink and purple) lava samples were collected from lava flows (Kahl et al., 2017), while the megacrysts and MtRs bomb sample were collected on the rim of the Monti Rossi scoria cone. Sample preparation The Monti Rossi megacrysts were repeatedly rinsed in beakers of Milli–Q water, 3·5% HCl and ethanol in an ultrasonic bath, and dried overnight at 40°C to remove any organic matter, altered material or groundmass. After cleaning, the crystals were photographed to record the surface texture and morphology (Fig. 2) and mounted in resin, lying on their c-axes. The mounts were cut in half and re-mounted, such that the section perpendicular to the c-axis could be observed at the polished surface through the centre of the crystal, minimising the effects of hourglass sector zoning observed in the crystals (Downes, 1974; Duncan & Preston, 1980; Ubide et al., 2019). The bomb and lava flow samples were studied on 30 μm-thick polished thin sections. The samples were scanned with a Leica DM6000M automated microscope at the University of Queensland; thin sections were scanned in plane- and cross-polarised light, while mounted megacrysts were scanned in reflected light. Fig. 2. Open in new tabDownload slide Photograph of black, shiny, euhedral clinopyroxene megacrysts (5–14 mm along the c-axis) selected for mounting, showing columnar, stubby, and twinned examples. Crystallographic orientation is provided on the top left corner. White circles highlight olivine inclusions which protrude from the surface of the host crystals. Fig. 2. Open in new tabDownload slide Photograph of black, shiny, euhedral clinopyroxene megacrysts (5–14 mm along the c-axis) selected for mounting, showing columnar, stubby, and twinned examples. Crystallographic orientation is provided on the top left corner. White circles highlight olivine inclusions which protrude from the surface of the host crystals. In-situ major element analysis Major element compositions of clinopyroxene were determined by electron probe microanalysis (EPMA) using a JEOL JXA-8200 instrument at the Centre for Microscopy and Microanalysis, The University of Queensland, Australia. The EPMA uses a conventional W electron source and has five high-resolution wavelength detector spectrometers for quantitative X-ray analysis. Analyses were performed on carbon-coated polished samples under high vacuum conditions using an accelerating voltage of 15 kV and an electron beam current of 15 nA, with a beam diameter of 2 μm. Elemental counting times were 30 seconds on the peak and 5 seconds on each of two background positions for all elements. Corrections for inter-elemental effects were made using a ZAF (Z, atomic number; A, absorption; F, fluorescence) procedure. Calibration of the major and minor elements utilised orthoclase (K), albite (Na), wollastonite (Si, Ca), kyanite (Al), hematite (Fe), chromite (Cr), spessartine (Mn), F-apatite (P), rutile (Ti), P-140 Olivine (Mg) and Ni-Olivine (Ni). Springwater Olivine, Kakanui Augite and Lake Co Feldspar were used as quality monitors (Jarosewich et al., 1980). Spot analyses defined transects across the crystals with varying spot separation (15–100 μm) depending on the crystal size. Transects were conducted from core to rim of all crystals (Figs 3 and 4). Results were screened based on stoichiometry and analytical totals and the final dataset includes 484 analyses (Supplementary Data Table S2). Fig. 3. Open in new tabDownload slide Major element variations (oxide wt %) between cores, mantles and rims in clinopyroxene megacrysts and lava hosted macrocryst (see simplified illustrations (e) at the bottom for context). Clinopyroxene mantles and rims show a strong compositional overlap for all elements. In contrast, cores are higher in MgO and SiO2 than mantles and rims, and lower in TiO2, FeOt and Al2O3, which is set apart by a compositional jump. In general, megacryst and macrocryst regions are in close compositional agreement. Fig. 3. Open in new tabDownload slide Major element variations (oxide wt %) between cores, mantles and rims in clinopyroxene megacrysts and lava hosted macrocryst (see simplified illustrations (e) at the bottom for context). Clinopyroxene mantles and rims show a strong compositional overlap for all elements. In contrast, cores are higher in MgO and SiO2 than mantles and rims, and lower in TiO2, FeOt and Al2O3, which is set apart by a compositional jump. In general, megacryst and macrocryst regions are in close compositional agreement. Fig. 4. Open in new tabDownload slide Petrography, trace element maps and major element transects. Left: each of the panels (a)–(d) show a reflected light scan of a clinopyroxene megacryst taken perpendicular to the C-axis. Melt inclusions (MI) and plagioclase (Pl) hosted within the crystals are labelled on the scans and highlighted by white boxes. A BSE image of an enlarged rim-hosted melt inclusion, the composition of which was used as an input for thermobarometry, is shown as an inset in panel (d). A laser ablation ICP-MS trace element map showing chromium concentration on a colour scale of 0–200 ppm is superimposed on each crystal scan (crystal (b) was previously imaged by Petrus et al., 2017). All of the crystals show three well-defined regions of zoning (core, mantle and rim). Core regions vary between crystals, however, all are strongly resorbed. Cores (a), (b) and (c) contain zones of high (>400 ppm) and low Cr (<110 ppm), while core (d) is low in Cr. Mantle regions in all crystals show oscillatory zoning. In crystals (b) and (c) the oscillatory-zoned pattern is interrupted by high-Cr zones greater than 120 ppm (mantle enrichment). Two such enrichments are highlighted in crystal (b), and their Cr concentrations are quantified in ppm. Dissolution textures (dissol) associated with zones of high Cr in crystals (b) and (d) are also marked by white circles. Finally, rim zones show Cr levels in excess of 200 ppm, higher than those of the mantle, yet significantly lower than high-Cr core zones. The Cr-rich rims of crystals (a), (c) and (d) are followed by a sharp, thin decrease in Cr at the crystal edge. This feature is absent in crystal (b). Quantified Cr concentrations (ppm) from representative zones of the core, mantle and rim are labelled on crystal (a). The black arrows shown on each laser ablation map mark the locations of major element transects, obtained with electron microprobe. Right: major element transects corresponding to maps (a)–(d) show variations in major oxides Al2O3 and MgO across the crystals. For comparison with trace element zonation, each transect is subdivided into corresponding core, mantle and rim regions. Cores are shown in red, mantles in blue and rims in yellow. Major elements are uniform across the mantles and rims apart from a small uptick in Al2O3 at the outer rim. Core regions are rich in MgO and depleted in Al2O3 relative to mantles and rims. Fig. 4. Open in new tabDownload slide Petrography, trace element maps and major element transects. Left: each of the panels (a)–(d) show a reflected light scan of a clinopyroxene megacryst taken perpendicular to the C-axis. Melt inclusions (MI) and plagioclase (Pl) hosted within the crystals are labelled on the scans and highlighted by white boxes. A BSE image of an enlarged rim-hosted melt inclusion, the composition of which was used as an input for thermobarometry, is shown as an inset in panel (d). A laser ablation ICP-MS trace element map showing chromium concentration on a colour scale of 0–200 ppm is superimposed on each crystal scan (crystal (b) was previously imaged by Petrus et al., 2017). All of the crystals show three well-defined regions of zoning (core, mantle and rim). Core regions vary between crystals, however, all are strongly resorbed. Cores (a), (b) and (c) contain zones of high (>400 ppm) and low Cr (<110 ppm), while core (d) is low in Cr. Mantle regions in all crystals show oscillatory zoning. In crystals (b) and (c) the oscillatory-zoned pattern is interrupted by high-Cr zones greater than 120 ppm (mantle enrichment). Two such enrichments are highlighted in crystal (b), and their Cr concentrations are quantified in ppm. Dissolution textures (dissol) associated with zones of high Cr in crystals (b) and (d) are also marked by white circles. Finally, rim zones show Cr levels in excess of 200 ppm, higher than those of the mantle, yet significantly lower than high-Cr core zones. The Cr-rich rims of crystals (a), (c) and (d) are followed by a sharp, thin decrease in Cr at the crystal edge. This feature is absent in crystal (b). Quantified Cr concentrations (ppm) from representative zones of the core, mantle and rim are labelled on crystal (a). The black arrows shown on each laser ablation map mark the locations of major element transects, obtained with electron microprobe. Right: major element transects corresponding to maps (a)–(d) show variations in major oxides Al2O3 and MgO across the crystals. For comparison with trace element zonation, each transect is subdivided into corresponding core, mantle and rim regions. Cores are shown in red, mantles in blue and rims in yellow. Major elements are uniform across the mantles and rims apart from a small uptick in Al2O3 at the outer rim. Core regions are rich in MgO and depleted in Al2O3 relative to mantles and rims. Melt inclusions identified within the mantles and rims of the Monti Rossi clinopyroxene megacrysts were analysed at the iCRAG labs at Trinity College Dublin using a TIGER MIRA3 Field Emission-Scanning Electron Microscope (FE-SEM) equipped with two Oxford X-Max 150 mm2 EDS detectors, running Oxford AZtec X-ray microanalysis software. Analytical conditions and standardisation methods followed those described in Ubide & Kamber (2018). A rim-hosted melt inclusion composition (Supplementary Data Table S3), representative of the melt from which the clinopyroxene rims crystallized, was used as melt input for the thermobarometric calculations described below. Mantle-hosted melt inclusions are typically smaller (<20µm) than those hosted in crystal rims (up to 100 µm), where we focused the analytical work. In-situ trace element analysis: LA-ICP-MS mapping Trace element zoning patterns of clinopyroxene were investigated using laser ablation-inductively coupled plasma mass spectrometry (LA-ICP-MS) mapping. Selected mapping areas included the centre, intermediate part and rim of all crystals, including those analysed previously by electron microprobe. This approach allowed for direct comparison of major and trace element variations. Mapping experiments were carried out at The University of Queensland Centre for Geoanalytical Mass Spectrometry, Radiogenic Isotope Facility (UQ RIF-lab), following the rastering method developed by Ubide et al. (2015). We used an ASI RESOlution 193 nm excimer UV ArF laser ablation system with a dual-volume Laurin Technic ablation cell and GeoStar Norris software, coupled to a Thermo iCap RQ quadruple mass spectrometer with Qtegra software. For efficient transport and to aid ionisation, Ar make-up gas with a trace amount of N2 was added to the ultra-pure He in which ablation was performed. A spot size of 20 × 20 μm, repetition rate of 10 Hz, fluence of 3 J/cm2 and translation speed of 20 μm/s were used for all experiments. To obtain maps for a large number of trace elements without compromising spatial resolution, two separate experiments were carried out on all crystals using identical laser parameters (Ubide et al., 2019). The first measured Li, Ga, Sc, V, Cr, Ni, Zr, Sr, La and Nd with a total dwell cycle of 115 ms, while the second measured rare earth elements (REE) with a dwell cycle of 200 ms. Maps were built using the data reduction software Iolite (Paton et al., 2011) v2·5 in quantitative mode. The instrument was tuned with ablation lines on NIST 612 glass reference standard. NIST 610 glass reference material was used as a calibration standard and the calcium concentrations (15·2–15·7 wt % Ca) obtained by EPMA for the different crystals were used as internal standard. Accuracy and precision were monitored using glass reference materials BCR-2G and BIR-1G (http://georem.mpch-mainz.gwdg.de/), bracketed before and after each mapping experiment and treated as unknowns. Accuracy was typically better than 10% for Ga, Sr, Zr and all REE and better than 20% for Li, Sc, V, Cr and Ni. Long-term reproducibility was better than 5% for all elements apart from Li, Sc, Cr and Ni, which were typically better than 20%. The results of trace element analysis are shown in Figs 4–6. Fig. 5. Open in new tabDownload slide Rare earth element distribution in clinopyroxene megacrysts. (a) Normalized REE (normalized to chondrite; McDonough & Sun, 1995) plot showing the range of REE values measured in both the mantles and rims of clinopyroxene megacrysts. Quantitative concentrations were extracted from LA-ICP-MS maps using the Iolite plugin Monocle (Petrus et al., 2017). Mantle values are represented by a blue polygon and rim values by an empty polygon with a yellow outline. Overall, REE are more enriched in megacryst mantles than rims, though the highest values in the rims overlap with the lowest values in the mantles. The overlap is most pronounced in the heavier REE. The clinopyroxene REE composition in equilibrium with the 1669 groundmass (from the MtRs bomb sample), calculated using cpx-melt partition coefficients specific to Mount Etna hawaiites (VBE-72; D’Orazio et al., 1998), is represented by a dashed red line. The resulting REE pattern falls almost entirely within the range of values measured in megacryst rims. (b) LA-ICP-MS map showing Lanthanum (La) on a colour scale of 0–50 ppm. Core and rim zones are depleted in La and correspond to zones of high Cr shown in Fig. 6a. Fig. 5. Open in new tabDownload slide Rare earth element distribution in clinopyroxene megacrysts. (a) Normalized REE (normalized to chondrite; McDonough & Sun, 1995) plot showing the range of REE values measured in both the mantles and rims of clinopyroxene megacrysts. Quantitative concentrations were extracted from LA-ICP-MS maps using the Iolite plugin Monocle (Petrus et al., 2017). Mantle values are represented by a blue polygon and rim values by an empty polygon with a yellow outline. Overall, REE are more enriched in megacryst mantles than rims, though the highest values in the rims overlap with the lowest values in the mantles. The overlap is most pronounced in the heavier REE. The clinopyroxene REE composition in equilibrium with the 1669 groundmass (from the MtRs bomb sample), calculated using cpx-melt partition coefficients specific to Mount Etna hawaiites (VBE-72; D’Orazio et al., 1998), is represented by a dashed red line. The resulting REE pattern falls almost entirely within the range of values measured in megacryst rims. (b) LA-ICP-MS map showing Lanthanum (La) on a colour scale of 0–50 ppm. Core and rim zones are depleted in La and correspond to zones of high Cr shown in Fig. 6a. Fig. 6. Open in new tabDownload slide LA-ICP-MS chromium maps of clinopyroxene megacrysts (a)–(d) seen in Fig. 4, on a colour scale of 0–500 ppm, highlighting mantle enrichments (core-type) and high-Cr core zones. High-Cr blobs observed in all crystals are inclusions of titanomagnetite. Fig. 6. Open in new tabDownload slide LA-ICP-MS chromium maps of clinopyroxene megacrysts (a)–(d) seen in Fig. 4, on a colour scale of 0–500 ppm, highlighting mantle enrichments (core-type) and high-Cr core zones. High-Cr blobs observed in all crystals are inclusions of titanomagnetite. Extracting compositional information from LA-ICP-MS maps The Iolite plugin ‘Monocle’ (Petrus et al., 2017) was used to extract quantitative data from the trace element maps. Monocle allows polygons to be drawn on visually identified regions of interest (ROI) in each crystal map (Supplementary Data Fig. S1). Selected signal pools are averaged from compositionally homogenous regions to provide high precision compositional data. Chromium zonation observed in the maps was used as a guide to draw polygons on unique compositional regions/zones for all elements, avoiding inclusions. Groundmass geochemistry The Monti Rossi bomb (MtRs; Kahl et al., 2017) was selected for bulk analysis of the groundmass fraction. A jaw crusher was used to crush the sample into 1 mm fragments from which the groundmass fraction was separated by removing macrocrysts (>1 mm) using a set of tweezers, with the aid of a light microscope. Major and trace element analyses were carried out in the Environmental Geochemistry Laboratory at The University of Queensland. For major element analysis, 0·1 g of powdered material was fused with 0·4 g of lithium metaborate flux in a Katanax Automatic Fluxer at ∼1000°C. The resulting glass bead was dissolved in 100 ml of 5% nitric acid and analysed on a Perkin Elmer Optima 8300DV Inductively Coupled Optical Emission Spectrometer (ICP-OES). A multi-acid open beaker hotplate digestion was used to prepare samples for trace element analysis, which was carried out on an Agilent 7900 Inductively-Coupled Plasma-Mass Spectrometer (ICP-MS). A more detailed description of the method is provided in Crossingham et al. (2018). The groundmass composition is reported in Supplementary Data Table S3. Clinopyroxene thermobarometry Clinopyroxene crystallization temperatures were calculated using the melt-dependent, pressure-independent thermometer of Putirka et al. (1996; Eq. T1), the results of which are shown in Fig. 7. A representative rim-hosted melt inclusion was selected as liquid input composition (Mg# 42, where Mg# = 100 x Mg / (Mg + Fe) on a molar basis and Fe represents total iron as Fe2+; Supplementary Data Table S3). The selected melt inclusion is completely glassy, without microlites or gas bubbles, as shown in the back-scattered electron image in the inset in Fig. 4, panel (d). We used the Fe–Mg exchange coefficient kD(Fe–Mg)cpx-melt as a test for equilibrium between clinopyroxene compositions and the melt. We note that whilst many of the mantle and rim compositions are in equilibrium with the input melt (kD(Fe–Mg)cpx-melt 0·28 ± 0·08, Putirka 2008; Fig. 8), core compositions are in equilibrium with a more Mg-rich liquid. Therefore, we treat core temperature results with caution. Regardless, to overcome the partial disequilibrium between mineral and melt compositions, we calculated pressure estimates with the clinopyroxene-only, temperature-dependent barometer of Putirka (2008; Eq. 32 b), with Eq. T1 (Putirka et al., 1996) as input temperature and assuming a melt H2O content of 2 wt % (Armienti et al., 2007, 2013). An analysis of the efficacy of clinopyroxene thermobarometry in basaltic systems concluded that this combination of equations recover experimental pressure estimates most accurately (Hammer et al., 2016). Further, Mollo et al. (2010) noted that for volatile-rich, alkaline compositions (such as those at Mount Etna), single-clinopyroxene barometers can outperform liquid-based models. Fig. 7. Open in new tabDownload slide Clinopyroxene thermobarometry. (a) Plot showing pressure and temperature estimates for the crystallization of 1669 megacrysts and macrocrysts, in the context of Mount Etna’s crustal stratigraphy. Crystallization of cores, mantles and rims occurred mostly in the upper crystalline basement and felsic granulite lithologies below the overlying sedimentary cover at pressures of 300–600 ± 200 MPa (corresponding to 11–20 ± 7 km bsl) and temperatures of 1120–1160 ± 27°C. The 1669 P–T estimates are compared with those from recent eruptions (1974, 2002–03 and 2014; Ubide & Kamber, 2018) (grey circles) which define a more extensive polybaric trend that extends to sub-Moho depths and records higher temperatures of 1150–1250°C. Crustal densities and stratigraphy are after Corsaro & Pompilio (2004) and Moho depth is after Hirn et al. (1997). Barometry constraints from other 1669 minerals ( in natural and experimental samples; Pl, plagioclase; Nicotra & Viccaro, 2012; Lanzafame et al., 2013; Ol, olivine; Kahl et al., 2015; Vetere et al., 2015) are also provided. (b) Frequency histograms highlight the largely overlapping P–T fields calculated for 1669 core, mantle and rim regions (megacrysts and macrocrysts), revealing only subtle differences between them. For instance, clinopyroxene rims and cores crystallize at slightly lower pressures than their surrounding mantles, but all results agree within the error of the calibrations. Fig. 7. Open in new tabDownload slide Clinopyroxene thermobarometry. (a) Plot showing pressure and temperature estimates for the crystallization of 1669 megacrysts and macrocrysts, in the context of Mount Etna’s crustal stratigraphy. Crystallization of cores, mantles and rims occurred mostly in the upper crystalline basement and felsic granulite lithologies below the overlying sedimentary cover at pressures of 300–600 ± 200 MPa (corresponding to 11–20 ± 7 km bsl) and temperatures of 1120–1160 ± 27°C. The 1669 P–T estimates are compared with those from recent eruptions (1974, 2002–03 and 2014; Ubide & Kamber, 2018) (grey circles) which define a more extensive polybaric trend that extends to sub-Moho depths and records higher temperatures of 1150–1250°C. Crustal densities and stratigraphy are after Corsaro & Pompilio (2004) and Moho depth is after Hirn et al. (1997). Barometry constraints from other 1669 minerals ( in natural and experimental samples; Pl, plagioclase; Nicotra & Viccaro, 2012; Lanzafame et al., 2013; Ol, olivine; Kahl et al., 2015; Vetere et al., 2015) are also provided. (b) Frequency histograms highlight the largely overlapping P–T fields calculated for 1669 core, mantle and rim regions (megacrysts and macrocrysts), revealing only subtle differences between them. For instance, clinopyroxene rims and cores crystallize at slightly lower pressures than their surrounding mantles, but all results agree within the error of the calibrations. Fig. 8. Open in new tabDownload slide Calculating melt in equilibrium with clinopyroxene and olivine regions. The magnesium number (Mg#) values of clinopyroxene cores, mantles and rims (top) and olivine cores and rims (bottom) are plotted against the Mg# of a representative rim-hosted melt inclusion. The range of Mg# measured in all rim-hosted melt inclusions is represented by a black line in panel (d). Equilibrium curves for clinopyroxene (⁠ kDCpx/liq = 0·28 ± 0·08; Putirka, 2008) and olivine (⁠ kDOl/liq = 0·3 ± 0·03; Roeder & Emslie, 1970) indicate which crystal regions are in equilibrium with the ascribed melt composition. Arrows pointing to coloured boxes on the x-axes show the calculated equilibrium melt for the range of compositions measured in each mineral region. Two different populations of olivine cores (M1 and mm1; Kahl et al., 2017) and rims (M3 and M5; Kahl et al., 2017) are shown. Olivine and clinopyroxene cores are the most primitive in composition and are in equilibrium with a similar range of melt compositions (liquid Mg# 45–50). Clinopyroxene mantles (liquid Mg# 41–47) and rims (liquid Mg# 42–47) are more evolved and are themselves in equilibrium with a similar range of melt compositions, which closely match the melt inclusion composition (Mg# 42). According to Fe–Mg exchange, both populations of olivine rim are in equilibrium with melts more evolved than the clinopyroxene hosted melt inclusion, however, M3 olivine rims (liquid Mg# 36–40) agree within error with those in equilibrium with clinopyroxene mantles and rims and are in close equilibrium with the erupted groundmass. Melts in equilibrium with M5 olivine rims (liquid Mg# 24–30) are significantly more evolved than those calculated for any other crystal zone. The brackets shown on clinopyroxene data represent the range for which the equilibrium melt was calculated. The highest Mg numbers measured in clinopyroxene mantle regions are considered the result of mantle enrichments of core-type melt and were not considered for the calculation of equilibrium liquids. Fig. 8. Open in new tabDownload slide Calculating melt in equilibrium with clinopyroxene and olivine regions. The magnesium number (Mg#) values of clinopyroxene cores, mantles and rims (top) and olivine cores and rims (bottom) are plotted against the Mg# of a representative rim-hosted melt inclusion. The range of Mg# measured in all rim-hosted melt inclusions is represented by a black line in panel (d). Equilibrium curves for clinopyroxene (⁠ kDCpx/liq = 0·28 ± 0·08; Putirka, 2008) and olivine (⁠ kDOl/liq = 0·3 ± 0·03; Roeder & Emslie, 1970) indicate which crystal regions are in equilibrium with the ascribed melt composition. Arrows pointing to coloured boxes on the x-axes show the calculated equilibrium melt for the range of compositions measured in each mineral region. Two different populations of olivine cores (M1 and mm1; Kahl et al., 2017) and rims (M3 and M5; Kahl et al., 2017) are shown. Olivine and clinopyroxene cores are the most primitive in composition and are in equilibrium with a similar range of melt compositions (liquid Mg# 45–50). Clinopyroxene mantles (liquid Mg# 41–47) and rims (liquid Mg# 42–47) are more evolved and are themselves in equilibrium with a similar range of melt compositions, which closely match the melt inclusion composition (Mg# 42). According to Fe–Mg exchange, both populations of olivine rim are in equilibrium with melts more evolved than the clinopyroxene hosted melt inclusion, however, M3 olivine rims (liquid Mg# 36–40) agree within error with those in equilibrium with clinopyroxene mantles and rims and are in close equilibrium with the erupted groundmass. Melts in equilibrium with M5 olivine rims (liquid Mg# 24–30) are significantly more evolved than those calculated for any other crystal zone. The brackets shown on clinopyroxene data represent the range for which the equilibrium melt was calculated. The highest Mg numbers measured in clinopyroxene mantle regions are considered the result of mantle enrichments of core-type melt and were not considered for the calculation of equilibrium liquids. Clinopyroxene clock Timescales from magma recharge to eruption were calculated for both SET 1 and SET 2 crystals following the method of Ubide & Kamber (2018). The time taken for a magma to reach the surface following an eruption-triggering magma recharge was calculated based on the assumption that time elapsed is equal to the thickness of the growth rim divided by a given growth rate. Based on the euhedral morphology of the crystal rims we assume that growth was continuous following recharge and took place at relatively low degrees of undercooling, and hence constant growth rate (Vona & Romano, 2013). The clinopyroxene rims, to which this growth rate is applied, are host to abundant melt inclusions, the entrapment of which requires growth rates of at least 10−8–10−7 cm/s (Baker, 2008). A growth rate of 10−8 cm/s was calculated for Etnean clinopyroxene based on experiments at low degrees of undercooling (<20°C) (Orlando et al., 2008). Although the crystallization experiments of Orlando et al., (2008) were carried out on more recent Etnean lavas (1991–1993), we note that the 10−8 cm/s clinopyroxene growth rate is supported by crystal size distribution studies at Mount Etna (Armienti et al., 1994; 2013). A very thin and euhedral Cr-poor outer rim, present in most crystals, is also included in timescale calculations. This outer rim may have crystallized during magma ascent (see Discussion) or emplacement, and potentially at higher degrees of undercooling. We consider the effect of increasing undercooling conditions on timescale estimates by applying two separate growth rates of 10−7 cm/s and 10−8 cm/s to the Cr-poor outer rim (Supplementary Data Table S4). We consider the timescales obtained at a growth rate of 10−8 cm/s as maxima, and we use the timescales obtained with growth rates of 10−7 cm/s in Fig. 9 and for further discussion. Nonetheless, we note that Cr-poor outer rims are significantly thinner and have limited contribution to timescale estimates compared with Cr-rich inner rims (Fig. 4a,c,d). Fig. 9. Open in new tabDownload slide Schematic illustration of the plumbing system conditions preceding the 1669 eruption. Progressive stages of melt accumulation, crystallization and magma movement preceding the eruption of SET 1 and SET 2 products are represented by numbers 1–3. Each stage is accompanied by a labeled illustration of the crystal regions (olivine 6-sided crystals and clinopyroxene 8-sided crystals) produced during that stage. (a) Stage 1 (red) represents the migration of ‘core-type’ melt into the deep and shallow plumbing system. Primitive core-type melt formed clinopyroxene cores in the deep plumbing system and corresponding M1 olivine cores (71% of olivine cores) in the shallow plumbing system. Stage 2 (yellow–blue) depicts the formation of melt in equilibrium with clinopyroxene mantles. The circular inset (yellow–blue gradient) shows a transition from a ‘rim-type’ melt to a ‘mantle-type’ melt following storage, convection and crystallization. Clinopyroxene mantles crystallized in the deep plumbing system, nucleating new macrocrysts and overgrowing pre-existing clinopyroxene cores to generate megacrysts. Similar to stage 1, clinopyroxene mantle-type melt also migrated to the shallow plumbing system, forming mm1 olivine cores (21% of olivine cores). (b) Finally, stage 3 (yellow) represents a voluminous, cryptic recharge of ‘rim-type’ melt into the ‘mantle-melt’ storage region. Cr-rich clinopyroxene rims were formed at depth, magmatic overpressure was reached, and melt migrated to the shallower plumbing system, mixing with olivine dominated magmas before eruption onset. Olivine rims depicted in stage 3 represent invasions of melt, which intruded the M1 and mm1 environments in the months prior to the eruption-triggering recharge event. M3 olivine rims may be the result of progressive input of clinopyroxene mantle-type melt into the shallow plumbing system. In contrast, M5 olivine rims are extremely evolved and are likely unrelated. The migration of melt from the deep to the shallow plumbing system probably occurred along discrete feeder conduits, and is represented by vertical black lines connecting lenses of magma. Fig. 9. Open in new tabDownload slide Schematic illustration of the plumbing system conditions preceding the 1669 eruption. Progressive stages of melt accumulation, crystallization and magma movement preceding the eruption of SET 1 and SET 2 products are represented by numbers 1–3. Each stage is accompanied by a labeled illustration of the crystal regions (olivine 6-sided crystals and clinopyroxene 8-sided crystals) produced during that stage. (a) Stage 1 (red) represents the migration of ‘core-type’ melt into the deep and shallow plumbing system. Primitive core-type melt formed clinopyroxene cores in the deep plumbing system and corresponding M1 olivine cores (71% of olivine cores) in the shallow plumbing system. Stage 2 (yellow–blue) depicts the formation of melt in equilibrium with clinopyroxene mantles. The circular inset (yellow–blue gradient) shows a transition from a ‘rim-type’ melt to a ‘mantle-type’ melt following storage, convection and crystallization. Clinopyroxene mantles crystallized in the deep plumbing system, nucleating new macrocrysts and overgrowing pre-existing clinopyroxene cores to generate megacrysts. Similar to stage 1, clinopyroxene mantle-type melt also migrated to the shallow plumbing system, forming mm1 olivine cores (21% of olivine cores). (b) Finally, stage 3 (yellow) represents a voluminous, cryptic recharge of ‘rim-type’ melt into the ‘mantle-melt’ storage region. Cr-rich clinopyroxene rims were formed at depth, magmatic overpressure was reached, and melt migrated to the shallower plumbing system, mixing with olivine dominated magmas before eruption onset. Olivine rims depicted in stage 3 represent invasions of melt, which intruded the M1 and mm1 environments in the months prior to the eruption-triggering recharge event. M3 olivine rims may be the result of progressive input of clinopyroxene mantle-type melt into the shallow plumbing system. In contrast, M5 olivine rims are extremely evolved and are likely unrelated. The migration of melt from the deep to the shallow plumbing system probably occurred along discrete feeder conduits, and is represented by vertical black lines connecting lenses of magma. LA-ICP-MS trace element maps highlighted the crystal rims, which were measured in Adobe Illustrator. The thickness of each rim was measured on the {100}, {010} and {110} crystal planes, but only those taken on the {100} planes were considered, to avoid the effect of differential growth rate between crystal faces. In any case, all the megacrysts were studied on basal cuts looking down the C-axis, which removed any effects of very fast growth along the C-axis. Special caution was applied to measure the thickness of slow growing crystal faces in lava-hosted clinopyroxenes, and the average thickness and standard deviations of both SET 1 and SET 2 rims were calculated based on multiple measurements of each sample. RESULTS Petrography Megacrysts The Monti Rossi clinopyroxene megacrysts range in size from 5–14 mm. They are black and shiny, most of them perfectly euhedral, following the augite crystal form, and slightly to significantly elongated in the c-axis direction. They occur as single crystals and glomerocrysts. The single crystals have both columnar and stubby forms and examples of simple and penetration twinning are common. Olivine inclusions protrude from the surface (Fig. 2), but are not commonly found in the interior of the crystals. Megacryst scans in reflected light, looking down the c-axis (basal sections), reveal an abundance of inclusions. Plagioclase, titanomagnetite and large melt inclusions (up to 100 µm at the crystal rims; Fig. 4d inset) were observed in all 19 crystals, while olivine was confirmed in only one. Melt and plagioclase inclusions are typically elongated and aligned parallel to the euhedral crystal zones. Alignment of inclusions is most common in crystal cores and rims and corresponds with zones of chromium enrichment (see trace element description below; Fig. 4). Plagioclase inclusions commonly act as points of weakness along which the crystal is fractured (Fig. 4b). Lava-hosted macrocrysts and groundmass SET 1 and SET 2 lava flow samples used in this study are porphyritic (Porphyritic index (PI): 48%; see Supplementary Data Table S5) with a glassy–microcrystalline groundmass. The mineral assemblage includes clinopyroxene (9%), plagioclase (34%), olivine (4%) and titanomagnetite (1%). Examples of large macrocrysts (3–5 mm) are only seen in clinopyroxene (4%) and plagioclase (4%). The size range and abundance of each phase is listed in Supplementary Data Table S5 (point counted). A comprehensive petrographic description of 1669 lava flow samples can be found in Kahl et al. (2017). We note that significant variation in porphyricity (PI: 37–52%; Corsaro et al., 1996) and relative mineral proportions have been observed in other products from this eruption. Major elements Major element electron microprobe transects were placed to include all zones of each crystal (core to rim). The data highlighted clear differences between the composition of crystal cores and surrounding mantles/rims; however, mantles and rims were not easily distinguished. In fact, definite distinctions between core, mantle and rim (as shown in Figs 3 and 4) were retrospectively applied to major element data following subsequent trace element analysis. The major element concentrations of megacrysts and lava-hosted macrocrysts plot in equivalent compositional fields (Fig. 3). Major elements reveal some variability between individual clinopyroxene cores, however, some shared compositional characteristics clearly distinguish cores from their surrounding mantles and rims. Cores typically show lower Al2O3, TiO2 and total iron as FeO (Fig. 3a,b,c) than mantles and rims, but are higher in MgO (Mg# 75–78) and SiO2 (Fig. 3d). In some cases, these differences are marked by a gradual trend and in others, a large compositional jump (Fig. 4 transects). A more detailed examination of major element transects and trace element maps of individual crystals (below) reveals additional nuances within core regions. The differences in major element composition observed between crystal cores and mantles are not observed between mantles and rims. The range of Mg# measured in Cr-rich rims (Mg# 72–76) is similar to the range of Mg# measures in Cr-poor oscillatory-zoned mantles (Mg# 71–76) (Supplementary Data Fig. S3). Some very gradual evolutionary trends are observed from mantles to rims (MgO and SiO2 decrease while TiO2 and FeO increase), however, the compositions of all major elements in both regions largely overlap (Fig. 3). A very slight decrease in FeO and TiO2 concentration, combined with an uptick in Al2O3, occurs at the Cr-poor outer rim (Fig. 4 transects). Trace elements LA-ICP-MS trace element mapping of clinopyroxene reveals sharp and complex compositional and textural zonation that is not apparent in the major element data. According to these maps, megacrysts and lava-hosted macrocrysts can be clearly divided into three major regions; irregular core, oscillatory-zoned mantle and euhedral rim (Fig. 4a), defined based on the compositional and textural patterns of compatible (Cr and Ni) and incompatible (REE and Zr) elements. The level of Cr-enrichment, and corresponding REE depletion, within a zone is considered indicative of the mafic character of the melt from which that zone crystallized (Ubide & Kamber, 2018). Patterns of Cr-enrichment and REE depletion are mirrored by Ni and Zr, respectively (Supplementary Data Fig. S2). Quantitative data for all measured elements, extracted from LA-ICP-MS maps, are available in Supplementary Data Tables S6 and S7. Trace element maps of lava flow-hosted clinopyroxene macro- and microcrysts from two stages of the eruption (SET 1 and SET 2) show that the identified compositional regions are common to almost all crystals (lava flow crystals and scoria hosted megacrysts). The largest lava-hosted macrocrysts (3–5 mm) contain a core, mantle and rim region, although more common macrocrysts (1–3 mm) contain only a mantle and rim region, and small crystals (<1 mm) contain only a rim region (Fig. 3e). Core regions, only found in the largest crystals (>3 mm) are defined by anhedral habits and strong variability in composition, within and between crystals, hosting both Cr-rich and Cr-poor zones (Fig. 4c). Mantles, in contrast, comprise multiple zones that oscillate mildly in composition but overall are depleted in Cr and enriched in REE relative to the other regions. Unlike the cores and mantles, rims are typically composed of only two zones, a thick inner rim enriched in Cr and depleted in REE (Fig. 5) followed by a very thin outer rim depleted in Cr (Fig. 4a,c,d) and enriched in REE. The textural and compositional features of each region are expanded on below. Cores Zoning patterns are notably different in every core-region and range from anhedral (Fig. 4b) to subhedral (Fig. 4a,d) in shape. However, all are composed of a combination of two distinct compositional groups: Low-Cr 49–110 ppm (blue–red in maps; Fig. 4) and high-Cr 400–1080 ppm (yellow in maps). Some cores contain only high-Cr zones (Fig. 4b), some only low (Fig. 4d) and others a mixture of both (Fig. 4a,c). The high-Cr zones of the megacryst core regions are more primitive than those found elsewhere in the crystals (Figs 4 and 6), and are sometimes associated with irregular surfaces and melt inclusions (Fig. 4b ‘dissol’). Such features are typical of a change in magmatic conditions, often induced by an input of new magma (Tsuchiyama, 1985; Ginibre et al., 2007). Interestingly, those core zones that are depleted in Cr also appear to be enriched in MgO and depleted in Al2O3 (Fig. 4d core); so, while individual Cr-rich recharge events were certainly coupled with an increase in Mg#, the early magmatic environment, which these magma recharges intruded, appears to have been fundamentally distinct from that in which the surrounding mantles and rims were formed. Mantles Mantle regions make up the largest portion of each crystal. They are characterized by oscillatory zoning, undulating between two compositions: 85–110 ppm Cr (red) and 49–70 ppm Cr (blue). In some crystals, this pattern is interrupted by isolated zones of higher Cr followed by a return to oscillatory zoning (Figs 4b,c and 6b,c). These isolated zones, varying in composition from 120–700 ppm Cr, are referred to here as ‘mantle-enrichments’ and like the Cr-rich cores, can be associated with irregular surfaces (Fig. 4b) that suggest dissolution due to disequilibrium (Tsuchiyama, 1985; Ginibre et al., 2007; Chen & Zhang, 2009). Almost all zones of the mantle (except for mantle-enrichments) are Cr-depleted (<110 ppm) relative to the rims. This depletion is matched by an equivalent enrichment in REE (Fig. 5). Further, the REE compositional range of lava flow hosted clinopyroxene mantles and rims are almost identical to their megacryst equivalents (Supplementary Data Fig. S4). Rims Chromium-rich rims (120–260 ppm) were found in 15 of the 16 mapped crystals. Rim regions are referred to as ‘Cr-rich’ in relation to the mantle region only, considering the significantly higher Cr-levels found in the cores. Just as Cr-levels in the rims are enriched relative to mantles, it follows that REE are depleted accordingly. Figure 5 shows that maximum and minimum REE values measured in megacryst mantles and rims plot in largely exclusive compositional fields and that rim compositions of all crystals are closest to equilibrium with the REE concentrations of the erupted groundmass (clinopyroxene compositions in equilibrium with the erupted groundmass were calculated using partition coefficients specific to Mount Etna lavas; D’Orazio et al., 1998) (Fig. 5a). The small overlap in composition between mantles and rims may be explained by REE-depleted ‘mantle-enrichments’. Finally, trace element maps also reveal a sharp decrease in Cr at the outer rim of most crystals (Fig. 4a,c,d) with the exception of one crystal (Fig. 4b) whose rim may have been removed during sample preparation. Any further discussion of clinopyroxene rims refers to the Cr-rich inner rim unless otherwise specified. Only two of 16 analysed crystals did not conform to the typical core, mantle, rim compositional distinctions and contained zones extremely enriched in Cr (up to 2000 ppm; Supplementary Data Figs S3 and S5). These crystals are interpreted as antecrysts, entrained from another magmatic environment, possibly during fluidisation and erosion of the surrounding magma mush prior to eruption (Bergantz et al., 2015; Kahl et al., 2017), and are not considered further in this study. DISCUSSION Conditions of crystallization The three compositional regions (cores, mantles and rims) identified above, based on marked differences in trace elements (and major elements in the core), are interpreted as having formed in distinct magmatic environments. To further characterize the intensive parameters (temperature and pressure) of each of these environments, clinopyroxene major element thermobarometry was applied to each region. The temperatures, pressures and associated depths at which each of the three regions crystallized are crucial to understand the spatial and temporal interactions between them. The data reported here are interpreted considering the error of the calibrations and in the context of published petrological and experimental data. Both megacrysts and macrocrysts indicate that crystallization of clinopyroxene cores and mantles occurred at pressures of 300–600 ± 200 MPa (Fig. 7b), with a peak of data between 400 and 500 MPa, corresponding to depths of 11–20 ± 7 km below sea level (Fig. 7a). This range agrees with the onset of clinopyroxene (>20 km: Armienti et al., 2013; Giacomoni et al., 2016) and plagioclase nucleation ( 10–15 km; Viccaro et al., 2010; Nicotra & Viccaro, 2012; Lanzafame et al., 2013), and with the boundary between the crystalline basement and the sedimentary cover at Mt Etna (9.9 km; Corsaro & Pompilio, 2004) (Fig. 7a). Changes in crustal density affect magma buoyancy and are considered natural areas for magma stagnation and sill emplacement (Putirka & Condit, 2003; Menand, 2008; Armienti et al., 2013). Thus, the crystalline-sedimentary transition might have enabled magmas of variable compositions to accumulate, as described for the recent activity of the volcano (Corsaro et al., 2009; Ubide & Kamber, 2018). We note that the temperatures at which clinopyroxene crystallization occurred in 1669 (1120–1160 ± 27°C) are substantially lower than those observed in present day magmas (1974–2014: 1150–1250 ± 27°C; Fig. 7 and Ubide & Kamber, 2018), calculated using the same methodology. The accumulation of magma in melt pockets, stored at various crustal levels within a mush system, could have incubated the anomalously large volume of magma (607 ± 105 x 106 m3; Branca et al., 2013b) erupted in 1669. Protracted crystallization in distinct, long-lived magma storage zones could also account for the unusually large size, euhedral habits and oscillatory zoned nature of both clinopyroxene and plagioclase erupted throughout the 17th century (Landi et al., 2019). The largest crystals may be a direct result of growth around crystal ‘cores’ transported from an early magma into ‘mantle-forming’ conditions where crystallization was sustained until the arrival of ‘rim’ melt. The compositional regions identified in clinopyroxene reveal a complex pre-eruptive history, involving three distinct magmatic environments. A detailed discussion of each of these environments is presented below. Megacryst cores Core regions record the earliest stages of crystallization in the pre-1669 plumbing system. Compositions rich in Cr (400–1000 ppm) and MgO (Mg# 75–78) indicate that crystal cores formed in magma more primitive than that of the surrounding mantles and rims. Cr-rich, recharge-related magmatic input appears to have been a common feature of this early environment. However, MgO enrichment and Al2O3 depletion in megacryst cores applies to both the Cr-rich and Cr-poor zones. In fact, Al2O3 depletion is more pronounced in the Cr-poor zones, suggesting a fundamental difference in crystallization conditions. Crystallization experiments of 17th century ‘cicirara’ lavas from Mount Etna by Vetere et al. (2015) show that with decreasing H2O and/or increasing CO2 levels in the melt, the composition of clinopyroxene is modified towards higher MgO and lower Al2O3 concentrations. Considering differences of >1 wt % Al2O3 and MgO in clinopyroxene cores compared with mantle/rims, we suggest that core compositions might have crystallized under lower H2O saturation or higher CO2 fluxing conditions. The clinopyroxene core-mantle boundary is marked by an abrupt change in zoning pattern and irregular surfaces, which may result from crystal dissolution (core resorption) (Fig. 4b ‘dissol’), following an abrupt shift in magmatic conditions. Cores with variable crystallization histories may have been assembled in the new ‘mantle’ environment, providing a seed around which megacrysts subsequently formed. Mantles Clinopyroxene mantle regions are consistently depleted in Cr and enriched in REE relative to cores and rims (Figs 4 and 5), as typical after extensive crystallization (Vetere et al., 2015). Though generally defined by chromium concentrations less than 110 ppm, mantle regions are interrupted at irregular intervals by isolated Cr-rich zones (Figs 4b,c and 6b,c), which stand out from the low-Cr, oscillatory-zoned pattern observed throughout (blue–red; Fig. 4). Some of these enrichments are subtle, showing Cr concentrations matching those of the Cr-rich rim (Cr >120 ppm) (Fig. 3b), while others show concentrations equivalent to those found in crystal cores (300–1000 ppm Cr) (Figs 4b,d and 5). In both instances, these enrichments are coupled with dissolution surfaces, a feature associated with an abrupt change in crystallization conditions, often as a result of the mixing of two different magmas (‘petrological cannibalism’ Cashman & Blundy, 2013; Perugini et al., 2005). We suggest that Cr-rich zones may reflect episodic mafic input into the ‘mantle-environment’ during the period of storage preceding eruption. Compositional homogeneity (particularly in major elements) across the clinopyroxene mantles and rims (Figs 3 and 4) suggests that the ‘mantle-environment’ was primarily supplied by input of rim-type melt. The observed variability in the intensity of mantle-enrichments may reflect the location of a particular crystal within the reservoir at the time of recharge and therefore the degree to which mixing is recorded. A crystal positioned further from the source of magmatic input may record a somewhat diluted chemical signature compared with those adjacent to it, and lack any additional evidence of interaction (i.e. resorption texture) (Cashman & Blundy, 2013). Such variability should be expected as thorough mixing of a magmatic reservoir is difficult to achieve (Schleicher et al. 2016). Nonetheless, the remarkable regularity of compositional oscillations suggests that some variability in physiochemical conditions (pressure, temperature, oxygen fugacity; Singer et al., 1995; Ginibre et al., 2002; Nicotra & Viccaro, 2012; Moussallam et al., 2015) is periodically encountered, perhaps as convective mixing proceeds in the recharge effected area (‘mixing cavity’; Bergantz et al., 2015; Landi et al., 2019). Following recharges, the magma probably underwent a period of convection and crystallization, repeatedly evolving from rim-type compositions to mantle-type compositions, a process thought also to favour the formation of megacryts (Landi et al., 2019). Similar small-scale compositional oscillations and disequilibrium textures have been observed in plagioclase megacrysts throughout the 17th century eruption products (Nicotra & Viccaro, 2012), suggesting that clinopyroxene and plagioclase growth were subject to similar processes of recharge and mixing. Rims: the eruption trigger Chromium-rich rims, comprising most of the outermost region of mega and macrocrysts, as well as entire microcrysts (0·5–1 mm) (found in 15 of 16 mapped crystals; Fig. 4e), indicate that a late-stage recharge event occurred prior to eruption. Magma recharge is known to be an efficient trigger of volcanic eruptions (Sparks et al., 1977, Reubi & Blundy, 2009, Bergantz et al., 2015), particularly when a melt of primitive composition adds temperature and volatiles to the system (Métrich et al., 2004, Spilliaert et al., 2006, Kamenetsky et al., 2007). In this case, however, the final input of rim-forming melt (Cr-rich inner rim) differed only slightly in both composition (Fig. 4) and temperature (Fig. 7) from the clinopyroxene mantle-forming environment, which it intruded. We therefore suggest that ‘cryptic’ recharge (Humphreys et al., 2006; Nicotra & Viccaro, 2012) of the mantle environment with rim-type magma may have acted as the primary eruption trigger. The addition of this new magma, distinguishable from the mantle environment by its trace element chemistry only, may have added volume or volatiles (as opposed altering composition or temperature), generating the overpressure required to push the system beyond a critical threshold. In the case of ‘mantle enrichments’ described above, an eruption was probably not triggered, but volume and volatiles were added to the growing ‘mantle-forming’ environment. Episodic input may have primed the system for eruption before a final recharge event (Cr-rich inner rim), with sufficient volume and duration, generated critical overpressure and propagation of dykes towards the surface (Degruyter et al., 2016). An input of sufficient velocity would fluidise the entire mush system (‘mixing bowl’; Bergantz et al., 2015), converting an otherwise non-eruptible mush into eruptible magma (Cashman et al., 2017). A thin, Cr-poor outer-rim, may have crystallized upon ascent, at increased undercooling (higher Al2O–TiO2 contents; Mollo et al., 2010), though low enough still to produce a euhedral crystal surface. Reconstructing the 1669 plumbing system: integrating pyroxene, olivine and plagioclase records Interestingly, clinopyroxene rims commonly host small crystals of olivine, whereas only few olivine inclusions are found in the cores-mantles of clinopyroxene megacrysts (1 out of 19 samples; also noted in early petrographic descriptions by Washington & Merwin, 1921). The crystallization of olivine occurs at the expense of clinopyroxene with decreasing pressure, as the exsolution of water enlarges the olivine stability field (Sisson & Grove, 1993). This effect was highlighted by experimental crystallization of ‘cicirara’ magma, which shows no olivine in water under-saturated conditions (> 400 MPa) (Vetere et al., 2015), but successfully reproduces natural olivine compositions at pressures < 400 MPa in water-saturated conditions (Métrich & Rutherford, 1998). Indeed, the analysis and thermodynamic modelling of 1669 olivine populations showed that the majority of olivine crystallization occurred during storage and transport in shallow regions of the plumbing system, below 300 MPa and mostly below 100 MPa (Kahl et al., 2017), well above the main clinopyroxene crystallization level (Fig. 7a). In contrast, plagioclase inclusions (up to 1 mm), hosted in the clinopyroxene megacrysts, suggest some contemporaneous crystallization of clinopyroxene and plagioclase, although megacrystic plagioclase crystals, typical of the ‘cicirara’ lavas erupted throughout the 17th century, have been interpreted as resulting from extended residence times in a quietly degassing magma mush (Armienti et al., 1997; Nicotra & Viccaro, 2012). So, while some plagioclase crystallization certainly occurred at depth (patchy cores; Nicotra & Viccaro, 2012) the majority of plagioclase probably crystallized at slightly shallower crustal levels than clinopyroxene, above 5–6 km depth, where H2O saturation is inferred for historic Etnean magmas (Viccaro et al., 2010). Interestingly, textural and compositional features observed in ‘cicirara’ plagioclase, erupted during the 1651–53 eruption of Mount Etna (Nicotra & Viccaro, 2012), resemble closely those recorded in 1669 clinopyroxene and suggest crystallization in similar magmatic environments. Nicotra & Viccaro (2012) describe prominent sieve textures close to the rim of the crystals (50–100 µm inside the crystal rims) that surround oscillatory zoned mantles and patchy resorbed cores. These sieve textures, while coupled to a sharp increase in anorthite content, record relatively constant FeO concentrations, and are thought to represent ‘cryptic mixing’ (Nicotra & Viccaro, 2012), akin to that observed in 1669 clinopyroxene. Whilst plagioclase compositional data were not acquired in the context of the present study, the striking similarity in the textural characteristics of 1669 plagioclases relative to 1651–53 crystals suggests that similar ‘cryptic’ recharge events may have been recorded in clinopyroxene and plagioclase over decades of eruptive activity (see section: ‘Implications for Mount Etna’s eruptive cycles’). Unlike clinopyroxene and plagioclase, 1669 olivine crystals show normal zoning (Type IV and VII; Kahl et al., 2017). Notably, the compositional range of most olivine cores (M1 environment Fo75–78; Kahl et al., 2017), found in 71% of the olivine population, is in equilibrium with melts more primitive than those of clinopyroxene mantles, but matches closely the equilibrium melts of clinopyroxene cores (liquid Mg# 45–51; Fig. 8a,b) suggesting crystallization from a similar magma. A second olivine population (mm1 environment: Fo73–75; Kahl et al., 2017) makes up 21% of olivine cores, and is in closer equilibrium with the melts of clinopyroxene mantles. We hypothesise that both clinopyroxene mantle and core-forming melts may have risen and pooled at shallow levels, where they produced olivine-dominated mushes. Magmas saturated in CO2 and containing high levels of H2O (3%) can achieve positive buoyancy up to 2·2 km bsl at Mount Etna (Corsaro & Pompilio, 2004). We speculate that primitive core-producing melts were more volatile-rich than mantle-producing melts, allowing them to migrate preferentially to the shallow plumbing system (Fig. 9a), where clinopyroxene crystallization is supressed by H2O saturation (Sisson & Grove, 1993) and subaphyric magmas are produced (Armienti et al., 1997). It is likely that core (M1) and mantle-type (mm1) melts were mixed to varying degrees in the shallow plumbing system and that, like clinopyroxene cores, olivine cores once showed complex zoning patterns. However, the occurrence of a compositional plateaux in olivine crystal cores indicates that any original zoning was diffusively reset (Kahl et al., 2017) following residence at magmatic temperatures (Costa & Morgan, 2011; Bouvet de Maisonneuve et al., 2016). While we find no difference in the compositions of SET 1 and SET 2 clinopyroxene rims, SET 1 (M5: Fo50–59) and SET 2 (M3: Fo65–69) olivine rims (from the same samples as clinopyroxene) are different from one another. Furthermore, melts in equilibrium with olivine rims (group M5: liquid Mg# 24–30, M3: liquid Mg# 36–40; Kahl et al., 2017) are more evolved than those in equilibrium with clinopyroxene rims (Mg# 42–47), suggesting that neither olivine population records the eruption triggering recharge event (Fig. 9b) observed in clinopyroxene. Rather, as described in Kahl et al., (2017), olivine rims may record an earlier and more varied history of melt migration to the shallow plumbing system, beginning as early as 1·5 years before eruption onset. Both M5 and M3 rims are interpreted as having crystallized under polybaric conditions (50–300 MPa) in distinct feeder conduits, forming around olivine cores upon intrusion into the resident M1 and mm1 environments. The extremely evolved compositions of M5 rims clearly derive from a unique compositional melt, which is not observed in the clinopyroxene crystal cargo. While this melt certainly contributed to the pressurisation of shallow plumbing system in the lead up to the 1669 eruption, its overall contribution to the bulk melt is not considered hugely consequential. Olivine M3 rim compositions on the other hand, though slightly more evolved than the liquids in equilibrium with clinopyroxene, are in equilibrium with the erupted groundmass and are more closely related to clinopyroxene mantles and rims (Fig. 8e). These rims may reflect the episodic arrival of clinopyroxene mantle-type magma to the shallow plumbing system, recording a process during which mm1 olivine cores were likely formed as well. Magma conduits such as those described by Kahl et al. (2017) may act as ephemeral connections between the deep and the shallow plumbing system. This decoupling of clinopyroxene and olivine rim populations supports the idea of a tiered plumbing system in which the crystallization of clinopyroxene and olivine are, for the most part, occurring in separate regions. An eruption triggering recharge event invading the deep plumbing system, where clinopyroxene dominated melts resided, would crystallize the observed Cr-rich clinopyroxene rims. This magma would then ascend into the shallow plumbing system, via a network of dykes or conduits, mixing with resident, olivine dominated mushes (Fig. 9a,b). In this case, the recharged melt, having limited interaction with olivine prior to eruption, may not crystallize a recharge rim, leaving the event unrecorded in olivine crystals. Timescales from recharge to eruption The movement of magma in a plumbing system is often seismically detectable, however, many more seismic events occur at volcanoes than those which precede major eruptions (Sparks, 2003; Martin et al., 2008). Although seismicity at Mount Etna has been intensely monitored since the 1970s (INGV-Sezione di Catania, 2002), it remains difficult to interpret which signals will result in eruption. An understanding of the magmatic processes that preceded major historical events can provide crucial context for interpreting these signals. No eruption equal in scale and destruction to 1669 has occurred in the era of monitoring, therefore we must rely on a combination of historical accounts and petrological and geochemical observations to inform the forecasting and hazard assessment for such an event. Here, we calculate timescales from recharge to eruption using the thickness of Cr-rich clinopyroxene rims and integrate our results with historical accounts of pre-eruptive seismicity and diffusion chronometry timescales for SET 1 and SET 2 olivine (Kahl et al., 2017). Timescales for the Cr-rich inner rims were calculated based on crystal growth rates (10−8 cm/s), at low degrees of undercooling (<20°C) implied from euhedral habits, as detailed in the methods section. Considering the relatively rapid growth informed by large melt inclusions hosted within the megacryst inner rims (Kouchi et al., 1983; Baker, 2008 and references therein), the growth rates applied to this zone can be considered conservative and the resulting timescales maxima. Assuming instantaneous growth of the Cr-rich recharge rim (Petrone et al., 2018), these timescales reflect the onset of recharge, magma mobilisation and perhaps initial ascent. In contrast, a faster growth rate of 10−7 cm/s was applied to the (much thinner) Cr-poor outer rims, since an uptick in Al2O3 at the outer rim of the megacrysts (Fig. 4 transects) suggests that the growth of outer rims may have occurred at an increased cooling rate (Mollo et al., 2010), perhaps during magma ascent or lava emplacement, where abundant volatile degassing results in increased undercooling and crystallization (Kouchi et al., 1983; Blundy & Cashman, 2005). The reported values combine timescales calculated for both the inner and outer rim. Nine of the 16 mapped crystals (SET 1 = 5 and SET 2 = 4) showed a clear {100} form that allowed measurements across the slowest growing face, and thicknesses measured for macrocrysts and megacrysts were very consistent across the eruptive products (rim measurements reported in Supplementary Data Table S4). In the absence of an exact eruption date, the timescales calculated were subtracted from the first reported date of eruption of each sample type (SET 1 and SET 2) and were placed on our suggested timeline of events accordingly (Fig. 10). Fig. 10. Open in new tabDownload slide Timeline of events preceding the 1669 eruption, including timescales of recharge to eruption and historical accounts of seismic activity (read from left to right). Yellow boxes show timescales of recharge to eruption calculated based on the thickness of clinopyroxene growth rims. The width of the yellow boxes represents the standard deviation of crystal rim thickness (and therefore timescales) measured on the {100} planes of each group of samples. The last day of growth for each crystal rim was taken to be the first possible day that sample was erupted: 14 March for SET 1 megacrysts and macrocrysts (Kahl et al., 2017; M. Mulas, personal communication); 25 March for SET 2 macrocrysts (Kahl et al., 2017). However, the solid red lines represent the range of days during which each sample could have been erupted. Earthquake symbols mark the occurrence of major earthquakes on 25 February and 8 March 1669, reported in historic accounts of the eruption (Portoghese, 1869; Mulas et al., 2016), which roughly coincide with our calculated timescales. All days and events to the left-hand side of the dashed red line occurred before eruption onset on 11 March, and all dates to the right of the dashed red line after it. Two black lines (SET 1 and SET 2) below the main timeline highlight the range of eruption timescales calculated by olivine diffusion chronometry (Kahl et al., 2017). Fig. 10. Open in new tabDownload slide Timeline of events preceding the 1669 eruption, including timescales of recharge to eruption and historical accounts of seismic activity (read from left to right). Yellow boxes show timescales of recharge to eruption calculated based on the thickness of clinopyroxene growth rims. The width of the yellow boxes represents the standard deviation of crystal rim thickness (and therefore timescales) measured on the {100} planes of each group of samples. The last day of growth for each crystal rim was taken to be the first possible day that sample was erupted: 14 March for SET 1 megacrysts and macrocrysts (Kahl et al., 2017; M. Mulas, personal communication); 25 March for SET 2 macrocrysts (Kahl et al., 2017). However, the solid red lines represent the range of days during which each sample could have been erupted. Earthquake symbols mark the occurrence of major earthquakes on 25 February and 8 March 1669, reported in historic accounts of the eruption (Portoghese, 1869; Mulas et al., 2016), which roughly coincide with our calculated timescales. All days and events to the left-hand side of the dashed red line occurred before eruption onset on 11 March, and all dates to the right of the dashed red line after it. Two black lines (SET 1 and SET 2) below the main timeline highlight the range of eruption timescales calculated by olivine diffusion chronometry (Kahl et al., 2017). SET 1 lavas (14–25 March) and megacrysts (14–19 March; unit MR2; Mulas et al., 2016) were the first erupted and have a total clinopyroxene rim thickness (Cr-rich inner and Cr-poor outer) of 219 ± 33 µm, placing the initiation of an eruption-triggering recharge event 21 ± 2 days before eruption onset. Historical accounts indicate that a major earthquake occurred on 25 February 1669, two weeks prior to the eruption, and was felt in Catania (Portoghese, 1869; Branca et al., 2015; Mulas et al., 2016). Based on our constraints on the depth of clinopyroxene crystallization, this earthquake may have been caused by movement of the recharge magma as it invaded the deep plumbing system (11–20 ± 7 km bsl). Increasing seismicity reported during the following two weeks could be related to continued magmatic input at depth. A second bout of major earthquakes occurred from the 8–10 March damaging the town of Nicolosi (Portoghese, 1869; Mulas et al., 2016). These shallower, more destructive earthquakes were probably the result of magma moving from the deep to shallow plumbing system and finally to the surface where eruption occurred, next to Nicolosi, on 11 March (Figs 1 and 10). SET 2 lavas began erupting 10 days later, on 21 March, and record thinner rims (131·3 ± 37·8 µm) corresponding to significantly shorter timescales of 11 ± 2 days. The comparatively slower mobilisation and migration of SET 1 lavas to the surface may have cleared the way for more rapid transport and shorter mixing times of subsequent SET 2 lavas. According to our suggested timeline, Mount Etna’s most destructive and voluminous eruption could be triggered by a recharge event occurring less than a month before onset. Timescales calculated by olivine diffusion chronometry (Kahl et al., 2017) yield significantly longer and more variable results than those calculated using clinopyroxene growth rims. SET 1 olivine timescales show a peak of data at 100 days (max >120 days) and SET 2 show a peak at 40 days (max 120 days) (Fig. 11). However, as described above, olivine and clinopyroxene rims likely do not record the same recharge event. Rather, olivine rims may record progressive input of melt into the shallow plumbing system beginning as early as October 1667 (M5 olivine rims) (Kahl et al., 2017). The longest timescales recorded by SET 2 (M3 olivine rims) indicate that intrusion into the M1 and mm1 core environments began in November 1668, increasing in December and remaining high until the weeks prior to eruption on 11 March 1669. Progressive input may indeed have pressurised the shallow plumbing system before the recharge-driven mobilisation of large melt volumes at depth eventually tipped the system to erupt. In addition, we note that olivine timescale calculations were modelled at temperatures of 1070 ± 20°C, significantly lower than the temperatures of clinopyroxene crystallization reported in this study. As Fe–Mg diffusion is slower at lower temperatures, an adjustment of the diffusion modelling parameters could modify olivine timescales to lower values. For example, at 1120°C (the lower limit of clinopyroxene crystallization temperature), olivine timescales would be reduced by a factor of 0·4 resulting in a median SET 1 timescale of 34 days (average clinopyroxene timescale: 23 days) and a median SET 2 timescale of 12 days (average clinopyroxene timescale: 12 days), in much closer agreement with clinopyroxene timescales. However, given differences in clinopyroxene and olivine rim composition and crystallization conditions (shallower and likely colder for olivine than for clinopyroxene), as well as variability in olivine timescale results, we favour the idea of progressive magmatic input to the olivine-dominated shallow plumbing system. Fig. 11. Open in new tabDownload slide Frequency histogram showing the distribution of SET 1 (blue) and SET 2 (orange) olivine diffusion chronometry timescales from Kahl et al. (2017). SET 1 results show a peak of data at 100 days and a maximum timescale of more than 120 days. SET 2 results have a peak of data at 40 days and a maximum timescale of 120 days. For the purpose of comparison, the olivine data presented here (n = 37) represents only samples from which clinopyroxene timescale results were also calculated. The complete olivine timescales dataset in Kahl et al. (2017) includes additional samples and data points (n = 150), but is largely coincident with the presented selection. Fig. 11. Open in new tabDownload slide Frequency histogram showing the distribution of SET 1 (blue) and SET 2 (orange) olivine diffusion chronometry timescales from Kahl et al. (2017). SET 1 results show a peak of data at 100 days and a maximum timescale of more than 120 days. SET 2 results have a peak of data at 40 days and a maximum timescale of 120 days. For the purpose of comparison, the olivine data presented here (n = 37) represents only samples from which clinopyroxene timescale results were also calculated. The complete olivine timescales dataset in Kahl et al. (2017) includes additional samples and data points (n = 150), but is largely coincident with the presented selection. Implications for the 17th century enhanced activity Throughout the 17th century (1600–1669), the average effusion rate and volume of erupted magma was unusually high at Mount Etna. The eruption products emitted in 1651–53 were some of the most extensive in the volcano’s history (24·97 ± 3·75 km2; Hughes et al., 1990), second only to the subsequent 1669 eruption (39·74 ± 4·0 km2; Hughes et al., 1990). Trace element zoning patterns and compositions measured in clinopyroxene macrocrysts from the 1651–53 eruption are strikingly similar to those observed in 1669 (Fig. 12a). Chromium-rich rims (125–320 ppm) surround Cr-poor (35–95 ppm), oscillatory-zoned mantle regions (Fig. 12a), which are similarly indistinguishable in terms of major element composition. Thermobarometry also places the mantles and rims of the 1651–53 eruption within the range of crystallization conditions defined for 1669 crystals, at pressures of 300–600 ± 200 MPa and temperatures of 1120–1160 ± 27°C. This relationship confirms not only that a compositionally consistent body of magma may have been sustained for at least 18 years (1651–1669), but also that both eruptions were triggered by a similar, cryptic recharge event. In contrast, very low Cr concentrations (<20 ppm) were measured in mantles/rims of clinopyroxenes from the eruption of 1566 (Fig. 12b), an indication that the resident magma at this time was different to that of 1651–53 and 1669. An investigation of samples from other 17th century eruptions (1607, 1610, 1614–24, 1634–38, 1643 and 1646–47; Branca & Carlo, 2005) would be required to better constrain the nature of the source of this peculiar activity, as well as its first occurrence and development through time. Fig. 12. Open in new tabDownload slide Pre-1669 clinopyroxene Cr maps. (a) LA-ICP-MS map of a clinopyroxene macrocryst from the 1651 eruption of Mount Etna showing Cr concentration on a colour scale of 0–200 ppm. (b) LA-ICP-MS map of a clinopyroxene microcryst from the 1566 eruption of Mount Etna showing Cr concentration on a colour scale of 0–20 ppm. Fig. 12. Open in new tabDownload slide Pre-1669 clinopyroxene Cr maps. (a) LA-ICP-MS map of a clinopyroxene macrocryst from the 1651 eruption of Mount Etna showing Cr concentration on a colour scale of 0–200 ppm. (b) LA-ICP-MS map of a clinopyroxene microcryst from the 1566 eruption of Mount Etna showing Cr concentration on a colour scale of 0–20 ppm. That being said, the unparalleled effusion rates of the 1669 eruption, as well as its emergence low on the flanks of the volcano, set it apart from all others of the 17th century and thereafter (Branca et al., 2013b). We speculate that increased mixing with a primitive ‘core-like’ magma may have led to increasing effusion rates and explosivity, particularly in the early stages of the eruption, during which less evolved products were erupted (tephra beds A1 and A2; Mulas et al., 2016). Furthermore, local extensional tectonics may have facilitated a lateral and comprehensive draining of large magma volumes at a low elevation, eventually leading to summit instability and caldera formation (Sartorius von Waltershausen, 1880; Nicotra & Viccaro, 2012). Such an event could disrupt magma refilling cycles and set the stage for a new eruptive cycle. Implications for Mount Etna’s eruptive cycles A significant increase in the number and explosivity of eruptions at Mount Etna over the last 45 years (Tanguy et al., 1997; Métrich et al., 2004; Spilliaert et al., 2006; Kamenetsky et al., 2007; Ubide & Kamber, 2018) has drawn some comparisons between recent eruptions and the enhanced activity of the 17th century (Behncke & Neri, 2003). For instance, the timing of present-day pre-eruptive magma transfer resembles that of 1669. Olivine timescales reported for recent Etnean eruptions (1991–2008; 2011–2013; Kahl et al., 2015; Giuffrida & Viccaro, 2017) record a large range of timescales, the majority of which occur in the 6 months prior to eruption onset. However, LA-ICP-MS mapping of clinopyroxene from a similar suite of eruptions (1974–2014) also revealed recharge generated Cr-enrichments at the crystal rims which, as for 1669, record recharge-to-eruption timescales significantly shorter than olivine (approximately 2 weeks; Ubide & Kamber, 2018). In addition, geophysical (geodetic, seismic, infrasonic) and petrological (olivine and plagioclase) evidence from the 2016 activity at the volcano showed that ascent of mafic magma from depth drove the eruption of magma residing in shallower reservoirs in less than a month (Cannata et al., 2018). This observed pattern, common to both eruption cycles, indicates that mafic recharge continues to act as an eruption triggering mechanism in the present-day plumbing system and is likewise preceded by progressive migration of melt to the shallow plumbing system in the months leading up to eruption. However, we also find significant differences between eruption types and triggering mechanisms in these two cycles. While present day activity is producing relatively small magma volumes of increasingly mafic character on an almost yearly basis (‘steady-state’ activity), the 17th century saw extremely voluminous eruptions of some of Mount Etna’s most evolved magmas and at much greater time intervals (up to 16 years; Fig. 13). So, while the estimated average emission rate from 1971–2010 (0·8 m3/s; Harris et al., 2011) was close to that of the 17th century (1·19 m3/s) (Hughes et al., 1990; Branca & Carlo, 2005), the current eruptions are less voluminous and consequently less destructive. Furthermore, the crystallization of clinopyroxene in today’s magmas occurs at significantly higher temperatures (1150–1250 ± 27° C; Fig. 7), and Cr concentrations at the rim are one order of magnitude higher (>1000 ppm Cr-rims; Ubide & Kamber, 2018), than those observed in the 1669 clinopyroxenes studied here. In addition to a post-1669 change in plumbing system geometry (Andronico et al., 2005), the observed disparity in temperature and composition between eruptive periods probably reflects a shift in the dominant melt producing magmatic source, which has been recognized by a major jump in 87Sr/86Sr isotopes in both mineral and whole rock compositions (Armienti et al., 2007; Viccaro & Cristofolini, 2008; Corsaro & Métrich, 2016). Fig. 13. Open in new tabDownload slide Comparison of eruption frequency and volume of 17th century and present day activity at Mount Etna. (a) Volume and duration of flank eruptions between the years 1600 and 1680. (b) Volume and duration of flank eruptions between the years 1970–2014. Solid red lines represent the duration over which a single eruption occurred and grey bars represent the total volume of erupted material throughout that eruption, plotted as a single bar at the end of the eruption period. Eruptions from 1600–1680 were less frequent but more voluminous than those between 1970 and 2014. Fig. 13. Open in new tabDownload slide Comparison of eruption frequency and volume of 17th century and present day activity at Mount Etna. (a) Volume and duration of flank eruptions between the years 1600 and 1680. (b) Volume and duration of flank eruptions between the years 1970–2014. Solid red lines represent the duration over which a single eruption occurred and grey bars represent the total volume of erupted material throughout that eruption, plotted as a single bar at the end of the eruption period. Eruptions from 1600–1680 were less frequent but more voluminous than those between 1970 and 2014. While the Cr concentrations observed in today’s primitive melts (1974–2014) are comparable to those of the 1669 core-producing melts, 17th century primitive volatile-rich magma was probably diluted by the larger volume of more evolved melt being produced and incubated. Primitive magmas stagnating in the crust would have cooled, crystallized and buffered to more evolved compositions. Unlike in the 17th century, more primitive magmas now rise rapidly from depth, where protracted storage is limited, and undergo little differentiation, crystallization or cooling upon ascent (Behncke & Neri, 2003; Andronico et al., 2005). In fact, diffusion modelling of Sr zonation in plagioclase has shown that maximum crystal residence times have been decreasing steadily from the 17th Century (1607–1892: 98-year average residence) to the present (1983–2013: 43-year average residence) (Viccaro et al., 2016). The same authors concluded that shorter residence times and more frequent eruptions, compared with the 17th century, may be the result of an increasing extensional tectonic influence in the upper 10 km of crust overlying Mount Etna. Persistent activity over multiple decades further implies steady-state behaviour, in which the rate of magma replenishment is in close equilibrium with the rate of eruption, rather than prolonged storage at crustal levels (Clocchiatti et al., 2004; Bonaccorso & Calvari, 2013; Petrone et al., 2018). Our high-resolution study of clinopyroxene presents an updated model of the storage conditions, magma movement and recharge to eruption timescales prior to Etna’s voluminous 1669 eruption. A long interval between eruptions (16 years in the case of 1669) was likely accompanied by protracted storage of large bodies of magma, at depths of more than 10 km in the Etna crust, also resulting in a protracted period of megacryst-producing crystallization. This large volume eruption of catastrophic consequence was likely triggered by a cryptic recharge event involving two compositionally similar melts distinguishable only by subtle differences in trace elements, and leading to mush mobilisation within as little as c.20 days. Identical zoning patterns observed in the earlier 1651–1653 eruption suggest that similar processes may have led to voluminous eruptions throughout the 17th Century. Though Mount Etna is highly active today and eruptions are triggered by magma recharge, the extended intervals of magma accumulation and crystallization required to produce 17th century-style eruptions do not appear to be a feature of the current plumbing system. It is unclear whether we will see a return to a 17th century style of activity in the future. However, an understanding of Mount Etna’s past plumbing system will aid future investigation into the long-term evolution, short-term cyclicity and hazard assessment of one of the most active volcanoes on Earth. CONCLUSIONS Clinopyroxene compositional zoning revealed complex two-tier pre-eruptive storage involving distinct magmatic environments in the lead up to Mount Etna’s most destructive historic eruption. The magmatic processes and eruption triggering mechanism identified in 1669 mineral products provide insights into the conditions and signals that might precede future voluminous and long-lasting eruptions as well as a benchmark to compare the current and future mineral record. LA-ICP-MS trace element maps yielded three unique compositional regions within clinopyroxene crystals: cores, mantles and rims. Early cores (MgO-, SiO2- and typically Cr-rich) formed in a primitive magmatic environment, while clinopyroxene mantles and rims crystallized from more evolved melts distinguishable only at the trace element level. In the lead-up to the 1669 eruption, large volumes of melt accumulated in the deep plumbing system (300–600±200 MPa), where protracted storage and crystallization produced an evolved ‘mantle-type’ magma composition hosting clinopyroxene mega- and macrocrysts. The large body of magma was sustained by episodic recharge with two distinct melts, recognized by core- and rim-like Cr-enrichments in the mantle regions. The largest of the erupted macrocrysts (3–5 mm) and megacrysts (>5 mm) nucleated on early formed primitive clinopyroxene cores. The primitive melt from which these cores formed may also have migrated to the shallow plumbing system forming olivine-dominated mushes. We suggest a two-tier system of magma storage and crystallization in which clinopyroxene and plagioclase-dominated magmas of the deep- to intermediate plumbing system were mixed with more primitive olivine-dominated magmas in the shallow plumbing system in the days/weeks prior to eruption. The compositional zoning patterns recognized in the 1669 eruption are also recognized in clinopyroxene from the 1651–53 eruption, suggesting that large volumes of melt could have been present in the plumbing system at least 18 years before the catastrophic 1669 eruption. The storage and eruption of magma throughout the 17th century may have been subject to similar pre-eruptive processes. The voluminous eruptions of 1669 and 1651–53 were likely mobilised by ‘cryptic’ mixing of two magmas with similar compositions. The eruption-triggering recharge event was identified through sharp enrichments in Cr and depletions in REE at the crystal rims, not coupled with changes in major element composition. Minimal changes in the composition and temperature of the magma following recharge suggests that an increase in melt volume or volatile content may have induced overpressure and eruption. According to our timescale calculations, the mobilisation of melt from recharge occurred less than a month prior to eruption. The current ‘steady-state’ of Mount Etna’s plumbing system, in which magma production is matched by frequent output, does not appear to be consistent with the processes that lead to voluminous eruptions of 17th century-style. However, given the current rate of supply, a return to such conditions is certainly possible. SUPPLEMENTARY DATA Supplementary data are available at Journal of Petrology online. ACKNOWLEDGEMENTS We would like to acknowledge the facilities and staff of the Australian Microscopy & Microanalysis Research Facility at the Centre for Microscopy and Microanalysis, The University of Queensland. We also acknowledge the facilities and staff of the Radiogenic Isotope Facility at the University of Queensland. We thank Marietjie Mostert for her assistance with bulk geochemical analysis, Gang Xia for his assistance with sample preparation, Tracey Crossingham for her assistance with EPMA and Balz Kamber for access to the geochemical facilities at Trinity College Dublin. We also wish to thank Maurizio Mulas for insightful discussions regarding the chronology of the 1669 eruption and John Caulfield for his help and support throughout this research. We are grateful to George Bergantz, Marco Viccaro and an anonymous reviewer for their constructive comments, which undoubtedly improved the manuscript, and we acknowledge efficient editorial handling by Georg Zellmer. Special thanks to Hugh Magee who spent his summer holiday collecting megacrysts. FUNDING This work was supported by The University of Queensland (ECR UQECR1717581 and MRFF RM2016000555 grants to TU). REFERENCES Allard P. , Behncke B., D'Amico S., Neri M., Gambino S. ( 2006 ). Mount Etna 1993–2005: anatomy of an evolving eruptive cycle . Earth-Science Reviews 78 , 85 – 114 . Google Scholar Crossref Search ADS WorldCat Aloisi M. , Cocina O., Neri G., Orecchio B., Privitera E. ( 2002 ). Seismic tomography of the crust underneath the Etna volcano, Sicily . Physics of the Earth and Planetary Interiors 134 , 139 – 155 . Google Scholar Crossref Search ADS WorldCat Andronico D. , Branca S., Calvari S., Burton M., Caltabiano T., Corsaro R. A., Del Carlo P., Garfì G., Lodato L., Miraglia L., Murè F., Neri M., Pecora E., Pompilio M., Salerno G., Spampinato L. ( 2005 ). A multi-disciplinary study of the 2002–03 Etna eruption: insights into a complex plumbing system . Bulletin of Volcanology 67 , 314 – 330 . Google Scholar Crossref Search ADS WorldCat Andronico D. , Lodato L. ( 2005 ). Effusive activity at Mount Etna Volcano (Italy) during the 20th century: a contribution to volcanic hazard assessment . Natural Hazards 36 , 407 – 443 . Google Scholar Crossref Search ADS WorldCat Armienti P. , Innocenti F., D'Orazio M. ( 2007 ). Mount Etna Pyroxene as Tracer of Petrogenetic Processes and Dynamics of the Feeding System. Geological Society of America Special Paper 418, 265 – 276 . Armienti P. , Pareschi M., Innocenti F., Pompilio M. ( 1994 ). Effects of magma storage and ascent on the kinetics of crystal growth . Contributions to Mineralogy and Petrology 115 , 402 – 414 . Google Scholar Crossref Search ADS WorldCat Armienti P. , Pareschi M. T., Pompilio M. ( 1997 ). Lava textures and timescales of magma storage at Mt. Etna (Italy) . Acta Vulcanol 9 , 1 – 5 . OpenURL Placeholder Text WorldCat Armienti P. , Perinelli C., Putirka K. D. ( 2013 ). A new model to estimate deep-level magma ascent rates, with applications to Mt. Etna (Sicily, Italy) . Journal of Petrology 54 , 795 – 813 . Google Scholar Crossref Search ADS WorldCat Barsotti S. , Andronico D., Neri A., Del Carlo P., Baxter P. J., Aspinall W. P., Hincks T. ( 2010 ). Quantitative assessment of volcanic ash hazards for health and infrastructure at Mt. Etna (Italy) by numerical simulation . Journal of Volcanology and Geothermal Research 192 , 85 – 96 . Google Scholar Crossref Search ADS WorldCat Baker D. R. ( 2008 ). The fidelity of melt inclusions as records of melt composition . Contributions to Mineralogy and Petrology 156 , 377 – 395 . Google Scholar Crossref Search ADS WorldCat Behncke B. , Neri M. ( 2003 ). Cycles and trends in the recent eruptive behaviour of Mount Etna (Italy) . Canadian Journal of Earth Sciences 40 , 1405 – 1411 . Google Scholar Crossref Search ADS WorldCat Bergantz G. W. , Schleicher J. M., Burgisser A. ( 2015 ). Open-system dynamics and mixing in magma mushes . Nature Geoscience 8 , 793 – 796 . Google Scholar Crossref Search ADS WorldCat Blundy J. , Cashman K. ( 2005 ). Rapid decompression-driven crystallization recorded by melt inclusions from Mount St. Helens volcano . Geology 33 , 793 – 796 . Google Scholar Crossref Search ADS WorldCat Blundy J. , Cashman K. ( 2008 ). Petrologic reconstruction of magmatic system variables and processes . Reviews in Mineralogy and Geochemistry 69 , 179 – 239 . Google Scholar Crossref Search ADS WorldCat Bonaccorso A. , Bonforte A., Calvari S., Del Negro C., Di Grazia G., Ganci G., Neri M., Vicari A., Boschi E. ( 2011 ). The initial phases of the 2008–2009 Mount Etna eruption: a multidisciplinary approach for hazard assessment . Journal of Geophysical Research 116 , B03203. OpenURL Placeholder Text WorldCat Bonaccorso A. , Calvari S. ( 2013 ). Major effusive eruptions and recent lava fountains: balance between expected and erupted magma volumes at Etna volcano . Geophysical Research Letters 40 , 6069 – 6073 . Google Scholar Crossref Search ADS WorldCat Borelli G. A. ( 1670 ). Storia e Meteorologia dell'eruzione dell'Etna del 1669. Catania: Gioenia. ISBN: 88-09-02254-8. Bouvet de Maisonneuve C. , Costa F., Huber C., Vonlanthen P., Bachmann O., Dungan M. A. ( 2016 ). How do olivines record magmatic events? Insights from major and trace element zoning . Contributions to Mineralogy and Petrology 171 , 56 . Google Scholar Crossref Search ADS WorldCat Branca S. , Azzaro R., De Beni E., Chester D., Duncan A. ( 2015 ). Impacts of the 1669 eruption and the 1693 earthquakes on the Etna Region (Eastern Sicily, Italy): an example of recovery and response of a small area to extreme events . Journal of Volcanology and Geothermal Research 303 , 25 – 40 . Google Scholar Crossref Search ADS WorldCat Branca S. , Carlo P. D. ( 2005 ). Types of eruptions of Etna volcano AD 1670–2003: implications for short-term eruptive behaviour . Bulletin of Volcanology 67 , 732 – 742 . Google Scholar Crossref Search ADS WorldCat Branca S. , Del Carlo P. ( 2004 ). Eruptions of Mt. Etna during the past 3200 Years: a revised compilation integrating the historical and stratigraphic records. In: Bonaccorso A., Calvari S., Coltelli M., Del Negro C., Falsaperla S. (eds) Mt. Etna: Volcano Laboratory. Geophysical Monograph 143 . Washington, DC: American Geophysical Union , pp. 1 – 27 . Branca S. , Coltelli M., Groppelli G. ( 2013 a). Geological Evolution of Etna Volcano. Mt. Etna : Volcano Laboratory. Washington, DC : American Geophysical Union, pp. 49 – 63 . Branca S. , De Beni E., Proietti C. ( 2013 b). The large and destructive 1669 AD eruption at Etna volcano: reconstruction of the lava flow field evolution and effusion rate trend . Bulletin of Volcanology 75 , 694 . Google Scholar Crossref Search ADS WorldCat Cannata A. , Grazia G. D., Giuffrida M., Gresta S., Palano M., Sciotto M., Viccaro M., Zuccarello F. ( 2018 ). Space‐time evolution of magma storage and transfer at Mt. Etna Volcano (Italy): the 2015–2016 Reawakening of Voragine Crater . Geochemistry, Geophysics, Geosystems 19 , 471 – 495 . Google Scholar Crossref Search ADS WorldCat Cashman K. V. , Sparks R. S. J. & Blundy J. D. ( 2017 ). Vertically extensive and unstable magmatic systems: A unified view of igneous processes . Science 355 , eaag3055 . Google Scholar Crossref Search ADS PubMed WorldCat Cashman K. , Blundy J. ( 2013 ). Petrological cannibalism: the chemical and textural consequences of incremental magma body growth . Contributions to Mineralogy and Petrology 166 , 703 – 729 . Google Scholar Crossref Search ADS WorldCat Chen Y. , Zhang Y. ( 2009 ). Clinopyroxene dissolution in basaltic melt . Geochimica et Cosmochimica Acta 73 , 5730 – 5747 . Google Scholar Crossref Search ADS WorldCat Clocchiatti R. , Condomines M., Guénot N., Tanguy J.-C. ( 2004 ). Magma changes at Mount Etna: the 2001 and 2002–2003 eruptions . Earth and Planetary Science Letters 226 , 397 – 414 . Google Scholar Crossref Search ADS WorldCat Coltelli M. , Del Carlo P., Vezzoli L. ( 2000 ). Stratigraphic constraints for explosive activity in the past 100 ka at Etna Volcano, Italy . International Journal of Earth Sciences 89 , 665 – 677 . Google Scholar Crossref Search ADS WorldCat Corsaro R. A. , Cristofolini R. ( 1996 ). Origin and differentiation of recent basaltic magmas from Mount Etna . Mineralogy and Petrology 57 , 1 – 21 . Google Scholar Crossref Search ADS WorldCat Corsaro R. A. , Cristofolini R., Patanè L. ( 1996 ). The 1669 eruption at Mount Etna: chronology, petrology and geochemistry, with inferences on the magma sources and ascent mechanisms . Bulletin of Volcanology 58 , 348 – 358 . Google Scholar Crossref Search ADS WorldCat Corsaro R. A. , Métrich N. ( 2016 ). Chemical heterogeneity of Mt. Etna magmas in the last 15ka. Inferences on their mantle sources . Lithos 252–253 , 123 – 134 . Google Scholar Crossref Search ADS WorldCat Corsaro R. A. , Métrich N., Allard P., Andronico D., Miraglia L., Fourmentraux C. ( 2009 ). The 1974 flank eruption of Mount Etna: an archetype for deep dike-fed eruptions at basaltic volcanoes and a milestone in Etna's recent history . Journal of Geophysical Research 114 , B07204. OpenURL Placeholder Text WorldCat Corsaro R. A. , Pompilio M. ( 2004 ). Buoyancy-controlled eruption of magmas at Mt Etna . Terra Nova 16 , 16 – 22 . Google Scholar Crossref Search ADS WorldCat Costa F. , Morgan D. ( 2011 ). Time constraints from chemical equilibration in magmatic crystals. In: Dosseto A., Turner S. P., Van Orman J. A. (eds) Timescales of Magmatic Processes: From Core to Atmosphere . Chichester : Wiley, pp. 125 – 159 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Cristofolini, R. & Romano, R. (1982). Petrologic textures of the Etnean volcanic rocks. Memorie della Società Geologica Italiana 23, 99–115. Crossingham T. J. , Ubide T., Vasconcelos P. M., Knesel K. M. & Mallmann G. ( 2018 ). Temporal constraints on magma generation and differentiation in a continental volcano: Buckland, eastern Australia . Lithos 302-303 , 341 – 358 . Google Scholar Crossref Search ADS WorldCat Davidson J. P. , Morgan D. J., Charlier B. L. A., Harlou R., Hora J. M. ( 2007 ). Microsampling and isotopic analysis of igneous rocks: implications for the study of magmatic systems . Annual Review of Earth and Planetary Sciences 35 , 273 – 311 . Google Scholar Crossref Search ADS WorldCat Degruyter W. , Huber C., Bachmann O., Cooper K. M., Kent A. J. R. ( 2016 ). Magma reservoir response to transient recharge events: the case of Santorini volcano (Greece) . Geology 44 , 23 – 26 . Google Scholar Crossref Search ADS WorldCat De Rita D. , Frazzetta G., Romano R. ( 1991 ). The Biancavilla-Montalto ingnimbrite (Etna, Sicily) . Bulletin of Volcanology 53 , 121 – 131 . Google Scholar Crossref Search ADS WorldCat D’Orazio M. , Armienti P., Cerretini S. ( 1998 ). Phenocryst/matrix trace-element partition coefficients for hawaiite-trachyte lavas from the Ellittico volcanic sequence (Mt. Etna, Sicily, Italy) . Mineralogy and Petrology 64 , 65 – 88 . Google Scholar Crossref Search ADS WorldCat Downes M. J. ( 1974 ). Sector and oscillatory zoning in calcic augites from M. Etna, Sicily . Contributions to Mineralogy and Petrology 47 , 187 – 196 . Google Scholar Crossref Search ADS WorldCat Duncan, A. M. & Preston, R. M. F. (1980). Chemical variation of clinopyroxene phenocrysts from the trachybasaltic lavas of Mount Etna, Sicily. Mineralogical Magazine 43: 765–770. Ferlito C. , Viccaro M., Nicotra E., Cristofolini R. ( 2012 ). Regimes of magma recharge and their control on the eruptive behaviour during the period 2001–2005 at Mt. Etna volcano . Bulletin of Volcanology 74 , 533 – 543 . Google Scholar Crossref Search ADS WorldCat Giacomoni P. , Coltorti M., Bryce J., Fahnestock M., Guitreau M. ( 2016 ). Mt. Etna plumbing system revealed by combined textural, compositional, and thermobarometric studies in clinopyroxenes . Contributions to Mineralogy and Petrology 171 , 1 – 15 . Google Scholar Crossref Search ADS WorldCat Giacomoni P. P. , Ferlito C., Coltorti M., Bonadiman C., Lanzafame G. ( 2014 ). Plagioclase as archive of magma ascent dynamics on “open conduit” volcanoes: the 2001–2006 eruptive period at Mt. Etna . Earth-Science Reviews 138 , 371 – 393 . Google Scholar Crossref Search ADS WorldCat Ginibre C. , Kronz A., Wörner G. ( 2002 ). High-resolution quantitative imaging of plagioclase composition using accumulated backscattered electron images: new constraints on oscillatory zoning . Contributions to Mineralogy and Petrology 142 , 436 – 448 . Google Scholar Crossref Search ADS WorldCat Ginibre C. , Worner G., Kronz A. ( 2007 ). Crystal zoning as an archive for magma evolution . Elements 3 , 261 – 266 . Google Scholar Crossref Search ADS WorldCat Giuffrida M. , Viccaro M. ( 2017 ). Three years (2011–2013) of eruptive activity at Mt. Etna: working modes and timescales of the modern volcano plumbing system from micro-analytical studies of crystals . Earth-Science Reviews 171 , 289 – 322 . Google Scholar Crossref Search ADS WorldCat Hammer J. , Jacob S., Welsch B., Hellebrand E., Sinton J. ( 2016 ). Clinopyroxene in postshield Haleakala ankaramite: 1. Efficacy of thermobarometry . Contributions to Mineralogy and Petrology 171 , 7 . Google Scholar Crossref Search ADS WorldCat Harris A. , Steffke A., Calvari S., Spampinato L. ( 2011 ). Thirty years of satellite-derived lava discharge rates at Etna: Implications for steady volumetric output . Journal of Geophysical Research 116 , B08204. OpenURL Placeholder Text WorldCat Hirn A. , Nicolich R., Gallart J., Laigle M., Cernobori L. & ETNASEIS Scientific Group. ( 1997 ). Roots of Etna volcano in faults of great earthquakes . Earth and Planetary Science Letters 148 , 171 – 191 . Google Scholar Crossref Search ADS WorldCat Hughes J. W. , Guest J. E., Duncan A. M. ( 1990 ). Changing styles of effusive eruption on Mount Etna since AD 1600. In: Ryan M. P. (ed.) Magma Transport and Storage . New York: Wiley , 385 – 405 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Humphreys M. C. S. , Blundy J. D., Sparks R. S. J. ( 2006 ). Magma evolution and open-system processes at Shiveluch Volcano: Insights from phenocryst zoning . Journal of Petrology 47 , 2303 – 2334 . Google Scholar Crossref Search ADS WorldCat Jarosewich E. , Nelen J. A., Norberg J. A. ( 1980 ). Reference samples for electron microprobe analysis . Geostandards and Geoanalytical Research 4 , 43 – 47 . Google Scholar Crossref Search ADS WorldCat Kahl M. , Chakraborty S., Costa F., Pompilio M. ( 2011 ). Dynamic plumbing system beneath volcanoes revealed by kinetic modeling, and the connection to monitoring data: an example from Mt. Etna . Earth and Planetary Science Letters 308 , 11 – 22 . Google Scholar Crossref Search ADS WorldCat Kahl M. , Chakraborty S., Pompilio M., Costa F. ( 2015 ). Constraints on the nature and evolution of the magma plumbing system of Mt. Etna Volcano (1991–2008) from a combined thermodynamic and kinetic modelling of the compositional record of minerals . Journal of Petrology 56 , 2025 – 2068 . Google Scholar Crossref Search ADS WorldCat Kahl M. , Viccaro M., Ubide T., Morgan D. J., Dingwell D. B. ( 2017 ). A branched magma feeder system during the 1669 Eruption of Mt Etna: evidence from a time-integrated study of zoned olivine phenocryst populations . Journal of Petrology 58 , 443 – 472 . Google Scholar Crossref Search ADS WorldCat Kamenetsky V. S. , Pompilio M., Metrich N., Sobolev A. V., Kuzmin D. V., Thomas R. ( 2007 ). Arrival of extremely volatile-rich high-Mg magmas changes explosivity of Mount Etna . Geology 35 , 255 – 258 . Google Scholar Crossref Search ADS WorldCat Kouchi A. , Sugawara Y., Kashima K., Sunagawa I. ( 1983 ). Laboratory growth of sector zoned clinopyroxenes in the system CaMgSi2O6-CaTiAl2O6 . Contributions to Mineralogy and Petrology 83 , 177 – 184 . Google Scholar Crossref Search ADS WorldCat Landi P. , La Felice S., Petrelli M., Vezzoli L. M., Principe C. ( 2019 ). Deciphering textural and chemical zoning of K-feldspar megacrysts from Mt. Amiata Volcano (Southern Tuscany, Italy): insights into the petrogenesis and abnormal crystal growth . Lithos 324–325 , 569 – 583 . Google Scholar Crossref Search ADS WorldCat Lanzafame G. , Mollo S., Iezzi G., Ferlito C., Ventura G. ( 2013 ). Unraveling the solidification path of a pahoehoe “cicirara” lava from Mount Etna volcano . Bulletin of Volcanology 75 , 703 . Google Scholar Crossref Search ADS WorldCat Martin V. M. , Morgan D. J., Jerram D. A., Caddick M. J., Prior D. J., Davidson J. P. ( 2008 ). Bang! Month-scale eruption triggering at Santorini volcano . Science 321 , 1178 – 1178 . Google Scholar Crossref Search ADS PubMed WorldCat McDonough W. F. , Sun C. C. ( 1995 ). Composition of the Earth . Chemical Geology 120 , 223 – 253 . Google Scholar Crossref Search ADS WorldCat Menand T. ( 2008 ). The mechanics and dynamics of sills in layered elastic rocks and their implications for the growth of laccoliths and other igneous complexes . Earth and Planetary Science Letters 267 , 93 – 99 . Google Scholar Crossref Search ADS WorldCat Métrich N. , Allard P., Spilliaert N., Andronico D., Burton M. ( 2004 ). 2001 flank eruption of the alkali- and volatile-rich primitive basalt responsible for Mount Etna's evolution in the last three decades . Earth and Planetary Science Letters 228 , 1 – 17 . Google Scholar Crossref Search ADS WorldCat Métrich N. , Rutherford M. J. ( 1998 ). Low pressure crystallization paths of H2O-saturated Basaltic-Hawaiitic melts from Mt Etna: implications for Open-System Degassing of Basaltic Volcanoes . Geochimica et Cosmochimica Acta 62 , 1195 – 1205 . Google Scholar Crossref Search ADS WorldCat Mollo S. , Del Gaudio P., Ventura G., Iezzi G., Scarlato P. ( 2010 ). Dependence of clinopyroxene composition on cooling rate in basaltic magmas: implications for thermobarometry . Lithos 118 , 302 – 312 . Google Scholar Crossref Search ADS WorldCat Mollo S. , Giacomoni P. P., Andronico D., Scarlato P. ( 2015 a). Clinopyroxene and titanomagnetite cation redistributions at Mt. Etna volcano (Sicily, Italy): footprints of the final solidification history of lava fountains and lava flows . Chemical Geology 406 , 45 – 54 . Google Scholar Crossref Search ADS WorldCat Mollo S. , Giacomoni P. P., Coltorti M., Ferlito C., Iezzi G., Scarlato P. ( 2015 b). Reconstruction of magmatic variables governing recent Etnean eruptions: constraints from mineral chemistry and P–T–fO2–H2O modeling . Lithos 212 , 311 – 320 . Google Scholar Crossref Search ADS WorldCat Moussallam Y. , Oppenheimer C., Scaillet B., Buisman I., Kimball C., Dunbar N., Burgisser A., Ian Schipper C., Andújar J., Kyle P. ( 2015 ). Megacrystals track magma convection between reservoir and surface . Earth and Planetary Science Letters 413 , 1 – 12 . Google Scholar Crossref Search ADS WorldCat Mulas M. , Cioni R., Andronico D., Mundula F. ( 2016 ). The explosive activity of the 1669 Monti Rossi eruption at Mt. Etna (Italy) . Journal of Volcanology and Geothermal Research 328 , 115 – 133 . Google Scholar Crossref Search ADS WorldCat Müller T. , Dohmen R., Becker H. W., ter Heege J. H., Chakraborty S. ( 2013 ). Fe–Mg interdiffusion rates in clinopyroxene: experimental data and implications for Fe–Mg exchange geothermometers . Contributions to Mineralogy and Petrology 166 , 1563 – 1576 . Google Scholar Crossref Search ADS WorldCat Murru M. , Montuori C., Wyss M., Privitera E. ( 1999 ). The locations of magma chambers at Mt. Etna, Italy, mapped by b-values . Geophysical Research Letters 26 , 2553 – 2556 . Google Scholar Crossref Search ADS WorldCat Neri M. , Acocella V., Behncke B., Maiolino V., Ursino A., Velardita R. ( 2005 ). Contrasting triggering mechanisms of the 2001 and 2002–2003 eruptions of Mount Etna (Italy) . Journal of Volcanology and Geothermal Research 144 , 235 – 255 . Google Scholar Crossref Search ADS WorldCat Nicotra E. , Viccaro M. ( 2012 ). Unusual magma storage conditions at Mt. Etna (Southern Italy) as evidenced by plagioclase megacryst-bearing lavas: implications for the plumbing system geometry and summit caldera collapse . Bulletin of Volcanology 74 , 795 – 815 . Google Scholar Crossref Search ADS WorldCat Orlando A. , D'Orazio M., Armienti P., Borrini D. ( 2008 ). Experimental determination of plagioclase and clinopyroxene crystal growth rates in an anhydrous trachybasalt from Mt Etna (Italy) . European Journal of Mineralogy 20 , 653 – 664 . Google Scholar Crossref Search ADS WorldCat Patanè D. , Barberi G., Cocina O., De Gori P., Chiarabba C. ( 2006 ). Time-resolved seismic tomography detects magma intrusions at Mount Etna . Science (New York, N.Y.) 313 , 821 – 823 . Google Scholar Crossref Search ADS PubMed WorldCat Patanè D. , De Gori P., Chiarabba C., Bonaccorso A. ( 2003 ). Magma ascent and the pressurization of Mount Etna's Volcanic System . Science (New York, N.Y.) 299 , 2061 – 2063 . Google Scholar Crossref Search ADS PubMed WorldCat Paton C. , Hellstrom J., Paul B., Woodhead J., Hergt J. ( 2011 ). Iolite: freeware for the visualisation and processing of mass spectrometric data . Journal of Analytical Atomic Spectrometry 26 , 2508 – 2518 . Google Scholar Crossref Search ADS WorldCat Perugini D. , Poli G., Valentini L. ( 2005 ). Strange attractors in plagioclase oscillatory zoning: petrological implications . Contributions to Mineralogy and Petrology 149 , 482 – 497 . Google Scholar Crossref Search ADS WorldCat Petrone C. M. , Braschi E., Francalanci L., Casalini M., Tommasini S. ( 2018 ). Rapid mixing and short storage timescale in the magma dynamics of a steady-state volcano . Earth and Planetary Science Letters 492 , 206 – 221 . Google Scholar Crossref Search ADS WorldCat Petrus J. A. , Chew D. M., Leybourne M. I., Kamber B. S. ( 2017 ). A new approach to laser-ablation inductively-coupled-plasma mass-spectrometry (LA-ICP-MS) using the flexible map interrogation tool ‘Monocle’ . Chemical Geology 463 , 76 – 93 . Google Scholar Crossref Search ADS WorldCat Portoghese B. ( 1869 ). Storia della eruzione dell'Etna del 1669 per lo pretore ritirato senza soldo. Catania: Tipografia Vincenzo Nettiniero. Putirka K. , Condit C. D. ( 2003 ). Cross section of a magma conduit system at the margin of the Colorado Plateau . Geology 31 , 701 – 704 . Google Scholar Crossref Search ADS WorldCat Putirka K. , Johnson M., Kinzler R., Longhi J., Walker D. ( 1996 ). Thermobarometry of mafic igneous rocks based on clinopyroxene-liquid equilibria, 0–30 kbar . Contributions to Mineralogy and Petrology 123 , 92 – 108 . Google Scholar Crossref Search ADS WorldCat Putirka K. D. ( 2008 ). Thermometers and barometers for volcanic systems . Minerals, Inclusions and Volcanic Processes 69 , 61 – 120 . Google Scholar Crossref Search ADS WorldCat Reubi O. , Blundy J. ( 2009 ). A dearth of intermediate melts at subduction zone volcanoes and the petrogenesis of arc andesites . Nature 461 , 1269 – 1273 . Google Scholar Crossref Search ADS PubMed WorldCat Roeder P. L. , Emslie R. F. ( 1970 ). Olivine-liquid equilibrium . Contributions to Mineralogy and Petrology 29 , 275 – 289 . Google Scholar Crossref Search ADS WorldCat Romano R. , Sturiale C. ( 1982 ). Geologia del versante sud-orientale etneo F° 270 IV (NO, NE, SO, SE) . Bollettino- Societa Geologica Italiana 100 , 15 – 40 . OpenURL Placeholder Text WorldCat Sartorius von Waltershausen W. ( 1880 ). Der Aetna. Pub. Von Lasaulx, W. Engelman, II vol, Leipzig.. Schleicher J. M. , Bergantz G. W., Breidenthal R. E., Burgisser A. ( 2016 ). Time scales of crystal mixing in magma mushes . Geophysical Research Letters 43 , 1543 – 1550 . Google Scholar Crossref Search ADS WorldCat Singer B. S. , Dungan M. A., Layne G. D. ( 1995 ). Textures and Sr, Ba, Mg, Fe, K, and Ti compositional profiles in volcanic plagioclase: clues to the dynamics of calc-alkaline magma chambers . American Mineralogist 80 , 776 – 798 . Google Scholar Crossref Search ADS WorldCat Sisson T. W. , Grove T. L. ( 1993 ). Temperatures and H2O contents of low-MgO high-alumina basalts . Contributions to Mineralogy and Petrology 113 , 167 – 184 . Google Scholar Crossref Search ADS WorldCat Sparks R. S. J. ( 2003 ). Forecasting volcanic eruptions . Earth and Planetary Science Letters 210 , 1 – 15 . Google Scholar Crossref Search ADS WorldCat Sparks S. R. J. , Sigurdsson H., Wilson L. ( 1977 ). Magma mixing: a mechanism for triggering acid explosive eruptions . Nature 267 , 315 – 318 . Google Scholar Crossref Search ADS WorldCat Spilliaert N. , Allard P., Métrich N., Sobolev A. V. ( 2006 ). Melt inclusion record of the conditions of ascent, degassing, and extrusion of volatile‐rich alkali basalt during the powerful 2002 flank eruption of Mount Etna (Italy) . Journal of Geophysical Research 111 , B0420. OpenURL Placeholder Text WorldCat Tanguy J.-C. , Condomines M., Kieffer G. ( 1997 ). Evolution of the Mount Etna magma: constraints on the present feeding system and eruptive mechanism . Journal of Volcanology and Geothermal Research 75 , 221 – 250 . Google Scholar Crossref Search ADS WorldCat Tanguy J.-C. , Condomines M., Le Goff M., Chillemi V., La Delfa S., Patanè G. ( 2007 ). Mount Etna eruptions of the last 2,750 years: revised chronology and location through archeomagnetic and 226Ra-230Th dating . Bulletin of Volcanology 70 , 55 – 83 . Google Scholar Crossref Search ADS WorldCat Tonarini S. , Armienti P., D’Orazio M., Innocenti F. ( 2001 ). Subduction-like fluids in the genesis of Mt. Etna magmas: evidence from boron isotopes and fluid mobile elements . Earth and Planetary Science Letters 192 , 471 – 483 . Google Scholar Crossref Search ADS WorldCat Tsuchiyama A. ( 1985 ). Dissolution kinetics of plagioclase in the melt of the system diopside-albite-anorthite, and origin of dusty plagioclase in andesite . Contributions to Mineralogy and Petrology 89 , 1 – 16 . Google Scholar Crossref Search ADS WorldCat Ubide T. , Kamber B. S. ( 2018 ). Volcanic crystals as time capsules of eruption history . Nature Communications 9 , 326 . Google Scholar Crossref Search ADS PubMed WorldCat Ubide T. , McKenna C. A., Chew D. M., Kamber B. S. ( 2015 ). High-resolution LA-ICP-MS trace element mapping of igneous minerals: in search of magma histories . Chemical Geology 409 , 157 – 168 . Google Scholar Crossref Search ADS WorldCat Ubide T. , Mollo S., Zhao J-X., Nazzari M., Scarlato P. ( 2019 ). Sector-zoned clinopyroxene as a recorder of magma history, eruption triggers, and ascent rates. Geochimica et Cosmochimica Acta 251, 265 – 283 . Vetere F. , Mollo S., Giacomoni P. P., Iezzi G., Coltorti M., Ferlito C., Holtz F., Perugini D., Scarlato P. ( 2015 ). Experimental constraints on the origin of pahoehoe “cicirara” lavas at Mt. Etna Volcano (Sicily, Italy) . Bulletin of Volcanology 77 , 44 . Google Scholar Crossref Search ADS WorldCat Viccaro M. , Barca D., Bohrson W. A., D'Oriano C., Giuffrida M., Nicotra E., Pitcher B. W. ( 2016 ). Crystal residence times from trace element zoning in plagioclase reveal changes in magma transfer dynamics at Mt. Etna during the last 400 years . Lithos 248–251 , 309 – 323 . Google Scholar Crossref Search ADS WorldCat Viccaro M. , Cristofolini R. ( 2008 ). Nature of mantle heterogeneity and its role in the short-term geochemical and volcanological evolution of Mt. Etna (Italy) . Lithos 105 , 272 – 288 . Google Scholar Crossref Search ADS WorldCat Viccaro M. , Giacomoni P. P., Ferlito C., Cristofolini R. ( 2010 ). Dynamics of magma supply at Mt. Etna volcano (Southern Italy) as revealed by textural and compositional features of plagioclase phenocrysts . Lithos 116 , 77 – 91 . Google Scholar Crossref Search ADS WorldCat Vona A. , Romano C. ( 2013 ). The effects of undercooling and deformation rates on the crystallization kinetics of Stromboli and Etna basalts . Contributions to Mineralogy and Petrology 166 , 491 – 509 . Google Scholar Crossref Search ADS WorldCat Washington H. S. , Merwin H. E. ( 1921 ). Note on augite from Vesuvius and Etna . American Journal of Science Series 5 Vol. 1 , 20 – 30 . Google Scholar Crossref Search ADS WorldCat © The Author(s) 2020. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - The Lead-up to Mount Etna’s Most Destructive Historic Eruption (1669). Cryptic Recharge Recorded in Clinopyroxene
JF - Journal of Petrology
DO - 10.1093/petrology/egaa025
DA - 2002-09-01
UR - https://www.deepdyve.com/lp/oxford-university-press/the-lead-up-to-mount-etna-s-most-destructive-historic-eruption-1669-6fRp85bk6B
SP - 1
VL - Advance Article
IS - 
DP - DeepDyve
ER -